Full text of "Keratin and keratinization; an essay in molecular biology"

See other formats

INTERNATIONAL SERIES OF MONOGRAPHS ON
PURE AND APPLIED BIOLOGY

Division: MODERN TRENDS IN PHYSIOLOGICAL SCIENCES

General Editors : P. Alexander and Z. M. Bacq

Volume 12

KERATIN

AND

KERATINIZATION

OTHER TITLES IN THE SERIES ON PURE AND
APPLIED BIOLOGY

Vol.

10.

Vol.

11.

Vol.

13.

Vol.

14.

Vol.

15.

MODERN TRENDS
IN PHYSIOLOGICAL SCIENCES DIVISION

Florkin — Unity and Diversity in Biochemistry

Brachet — The Biochemistry of Development

Gerebtzoff — Cholinesterases

Brouha — Physiology in Industry

Bacq and Alexander — Fundamentals of Radiobiology

Florkin (Ed.) — Aspects of the Origin of Life

Hollaender (Ed.) — Radiation Protection and Recovery

Kayser — The Physiology of Natural Hibernation

Francon — Progress in Microscopy

Charlier — Coronary Vasodilators

Gross — Oncogenic Viruses

Heath — Organophosphorus Poisons

Chantrenne — The Biosynthesis of Proteins

Rivera — Cilia, Ciliated Epithelium and Ciliary Activity

BIOCHEMISTRY DIVISION
Vol. 1 . Pitt-Rivers and Tata — The Thyroid Hormones
Vol. 2. Bush — The Chromatography of Steroids
Vol. 3. Engel — Physical Properties of Steroid Hormones

BOTANY DIVISION
Bor — Grasses of Burma, Ceylon, India and Pakistan
Turill (Ed.) — Vistas in Botany-Volume 1
Schultes — Native Orchids of Trinidad and Tobago
Cooke — Cork and the Cork Tree

PLANT PHYSIOLOGY DIVISION
Vol. 1. Sutcliffe — Mineral Salts Absorption in Plants
Vol. 2. Siegel — The Plant Cell Wall

ZOOLOGY DIVISION

Raven — An Outline of Developmental Physiology

Raven — Morphogenesis: The Analysis of Molluscan Development

Savory — Instinctive Living

Kerkut — Implications of Evolution

Tartar — The Biology of Stentor

Jenkin — Animal Hormones

Corliss — The Ciliated Protozoa

George — The Brain as a Computer

Arthur — Ticks and Disease

Raven — Oogenesis

Mann — Leeches {Hirudinea)

Vol.

10.

Vol.

11.

KERATIN

AND

KERATINIZATION

An Essay in Molecular Biology
E. H. Mercer, D.Sc, Ph.D.

Chester Beatty Research Institute:

Institute of Cancer Research,
Royal Cancer Hospital, London

PERGAMON PRESS

NEW YORK • OXFORD • LONDON • PARIS

1961

PERGAMON PRESS INC.
122 East 55th Street, New York, 22, N. Y.
1404 Neiv York Avenue N.W., Washington 5 D.C.

PERGAMON PRESS LTD.

Headington Hill Hall, Oxford

4 & 5 Fitzroy Square, London, W.l

PERGAMON PRESS S.A.R.L.
24 Rue des Ecoles, Paris, V.
PERGAMON PRESS G.m.b.H.
Kaiserstrasse 75, Frankfurt am Main

Library of Congress Card No. 60-53516

Set in Imprint 11 on 12 pt. and printed in Great Britain by
BELL AND BAIN LTD., GLASGOW, SCOTLAND

This book is dedicated to the memory of
W. T. ASTBURY

Contents

Preface

Acknowledgements

I Keratin and Molecular Biology

Macromolecules and biology

Orders of magnitude
Types of fibrous proteins and their classification
Birefringence
X-ray methods

The a, j8 and collagen patterns

Stabilization

Ecdysis
Distribution of the fundamental fibre-types

Some difficulties in defining a keratin

The significance of the variable amino acid composition
The fine structure of cells

Electron microscopy and cytology
The cell surface, its specializations and intercellular adhesion
The cell membrane

Cilia, flagella, etc

Surface invaginations

Specializations of opposed surfaces. Cell contacts
Desmosomes

Interdigitation of confronted membranes
The differentiation of surface organelles
Cytoplasmic structures
Particulates
Membrane systems
The nucleus
Differentia
The phylogeny of keratinization
II The Keratinized Tissues

Generalized histology of the vertebrate epidermis
Dermis and epidermis
The epidermis . .
The epidecmal family of cells

H^H^mUy

xiii
1
1
3
5

10
11
14
19
22
22
27
31
34
34
37
37
39
40
40
41
43
44
45
45
46
48
48
49
53
53
53
55
57

LIBRARY

^> V mat: / Q*^

MASS.

81317

CONTENTS

The differentiation of epidermal cells
Hard and soft keratins
The specialized appendages

Localized epidermal thickenings
Scales
Horns

The digital tips : claws, nails, hoofs
Feathers
Hairs

The phylogeny of hair
Other possibly keratinized structures
III Differentiation and Protein Synthesis
The cytology of keratinizing cells

The basal layer cells

Cell contacts during differentiation
The dermoepidermal junction
The development of basal membranes and their role in the
formation of epithelia
The differentiated layers and the variety of cell products
The epidermis

Tonofibrils . .

Keratohyalin

The hair follicle

Intracellular differentiation in the hair bulb

The feather follicle

The avian secreted keratins
The synthesis of protein in epidermal systems
Cytology of cells which form protein
Biochemistry of protein synthesis
Secondary and tertiary structures
Synthesis in retaining systems
Synthesis in fibre-forming systems
The supermolecular organization of fibrous tissues (tertiary
structure)
Macromolecular fibrous texture
Fibrogenesis

Collagen

Fibrous insulin
Silk fibroin — aggregation of the molecule after unfolding
Organization of fibrous tissues
Epidermal fibrils
IV The Growth of Epidermal Structures
The epidermis as a growing organ

CONTENTS

Mitosis in the basal layer
General theories of growth
Periodic growth and cyclic activity
Control of epidermal growth

Competition
Patterns of hair growth and control

Zig-zags, curls and crimps
Allometric growth

Molecular and Macromolecular Structure
The present status of the chemical structure of the keratins

End groups
Molecular structure

Methods of partial degradation

Non- destructive methods
X-ray diffraction
The low-angle pattern

The wide-angle pattern

The elastic properties and the structure of hair
Current crystallographic analysis

Pleated sheet configurations — Silk fibroin .

Helical configurations — The a-helix . .
Coiled coils and a-filaments

The organization of a-filaments into larger structures
The a-j8 transformation in terms of the a-helix
The non-crystalline fraction
Other methods of determining chain configuration

Optical rotation and rotary dispersion

Infra-red spectra and structure

Deuterium exchange

The cross-/? pattern

Other a-proteins

Some a- and jS-proteins of insect origin

Feather keratin

The Keratinization Process

The hard keratins

The development of orientation

The development of stability

Thiol and disulphide groups during keratinization

Nucleic acids and synthesis

Metabolic enzymes

Glycogen

Acid mucopolysaccharides and Schiff-reactive substances

Phosphatases

135
138
143
146
149
150
156
159
161
161
161
163
163

164
165
170
172
176
179
181
183
187
188
190
194
194
196
200
200
202
203
205
210
210
211
214
217
219
221
221
221
222

CONTENTS

Lipids

Water content
The fine histology of the hair follicle in the keratinization
zone

The cortex

The cuticle

The inner root sheath
Soft keratinization

The epidermis

Keratinization of horn
Follicular nutrition and the entrance of sulphur
Soluble products of partial keratinization

Extracts from the pre-keratinized zone

Soluble derivatives of keratinized tissues
Reduction of wool
Oxidizing agents
Thioglycollate extracts

Soluble derivatives of feather and other keratins
The location of the cystine residues
Filaments (a component)
Matrix (y component)
Physiochemical properties and keratinization

Salt linkages

Disulphide bonds

Hydrogen bonds

Molecular configuration in the supercontracted state

The setting of hairs
Cell membranes in keratinized tissues

The membranes and cellular adhesion

The morphology of the membranes of keratinized tissues

The fate of the intracellular apparatus during keratinization
The hair cuticle
The medulla
The residues remaining after the chemical extraction of

keratins
Uneven keratinization and its histological distribution
Keratinized cysts and epidermal tumours
Pigmentation

The pigment granule

The chemistry of melanization

References
Author Index
Subject Index

Preface

An explanation of the sub-title of this work may not be out-of-place
since it will explain at the same time the treatment of the main theme.
The term molecular biology has gained currency recently; its use implies
an attitude towards biology, an acceptance of the belief that biological
phenomena can be related to the interaction of the molecules found in
biological systems. In practice it means that the primary emphasis is
placed on the determination of molecular structure by any of the means
available.

In its pretensions molecular biology is a generalized science; it aims to
provide a common, integrating background to such special sciences as,
for example, biochemistry, histology and physiology, by rendering them
alike explicable in terms of molecular interaction. The name but not the
subject matter is new. In the pre-war period such writers as K. H. Meyer
and H. Mark, W. T. Astbury, J. D. Bernal, W. J. Schmidt, F. O. Schmitt
and A. Frey-Wyssling, to mention only a few, had the objective well in
mind. What is new, however, is the tempo of achievement in the post-
war years; detailed structures of key biological macromolecules have
been obtained and a complete coverage of cellular contents of all dimensions
is possible now by means of microscopy. These achievements have lifted
the dream of relating structure and function out of speculation to become
a problem capable of experimental investigation.

That it is still largely a project one realises quickly enough in setting out
to give an account of even a simple, relatively uniform group of tissues
such as the epidermis and its appendages. Thus, while this book attempts
to pose the problems of keratinization consistently in molecular terms
and to avoid concepts not stateable in such terms, it rather quickly
degenerates into an outline of unsolved problems. Keratinization will
be regarded as a development of certain primitive cellular traits adapted
to serve the end of providing a protective coating to a multicellular
organism. The traits emphasised are intercellular adhesion and the
proliferation and stabilization of cytoplasmic protein filaments. We shall
be concerned for the most part with the structure and synthesis of these
filaments, with the structure of cell membranes and the nature of inter-
cellular adhesion.

I am grateful to Professor A. Haddow, f.r.s., Director of the Chester
Beatty Research Institute, where I have found the opportunity to

xii PREFACE

continue this work and to associate with many stimulating colleagues.
My particular thanks are due to: Mr. M. S. C. Birbeck with whom I
have collaborated in much of the experimental work described here ; to
Mrs. Rex Dadd who prepared the typescript; to Mr. K. R. Moreman
and Mr. M. Docherty for their careful attention to the photographic
material; to Dr. P. Alexander for reading the manuscript and for his
encouragement; to Dr. K. M. Rudall for reading the page proofs and
for pointing out many shortcomings; to the staff of Pergamon Press
Ltd., in particular Mr. B. J. Adams, for their unfailing attention during
the preparation of the work for publication.

E. H. Mercer
1961

Acknowledgements

My thanks are due to the following who have very kindly provided me
with original photographs for plates: Mr. H. J. Woods (Plates 1 and 2),
Dr. K. M. Rudall (Plate 3), Mr. M. S. C. Birbeck (Plates 5B and 10B),
Dr. G. E. Rogers (Plate 16), Dr. I. Brodv (Plate 17) and Dr. I. Heiger
(Plate 22A).

The following publishers and societies have granted permission to
reproduce figures from their publications :
The Royal Society

(Proceedings of the Royal Society, Philosophical Transactions of the
Royal Society)
The Textile Institute

(Journal of the Textile Institute)
Rockefeller Institute for Medical Research

(Journal of Biophysical and Biochemical Cytology, Journal of General
Physiology)
Macmillan & Co.

(Nature)
Elsevier Publishing Co.

(Biochimica et Biophysica Acta)
North-Holland Publishing Co.

(Mechanical Properties of Fibres)

Cambridge University Press

(Journal of Endocrinology, An Introduction to Comparative Bio-
chemistry (3rd edition)

Athlone Press

(Lectures on the Scientific Basis of Medicine)

Society of Dyers and Colourists
(Fibrous Proteins)

Almquist and Wiksell

(Proceedings of the International Conference on Electron Microscopy,
Stockholm 1956)
Society Cosmetic Chemists, London

(Journal of the Society of Cosmetic Chemists)

XIV ACKNOWLEDGEMENTS

Radio Corporation of America

(Radio Corporation of America Scientific Instrument News)

New York Academy of Sciences

(Annals of the New York Academy of Sciences)

Academic Press Inc.

(Advances in Protein Chemistry, Journal of Ultrastructure Research)

Springer- Verlag, Berlin

(Proceedings of the 4th International Congress on Electron Microscopy)

C.S.I.R.O. Australia

(Australian Journal of Biological Sciences, Proceedings of the Inter -
national Wool Textile Conference, Australia 1956)

American Chemical Society

(Journal of the American Chemical Society)

Royal Society of Edinburgh

(Transactions of the Royal Society of Edinburgh)

Society for Experimental Biology

(Symposium of the Society for Experimental Biology)

CHAPTER 1

Keratin and Molecular Biology

Macromolecules and biology

Biological processes, whatever the organism, plant or animal, are
inseparably associated with macromolecules. Some of these form the
solid frameworks which support and protect organisms; others as more-
mobile particles effect the reactions with each other and with smaller
molecules which we recognize as ultimately characteristic of life; others
again form the material basis of inheritance. It is no more than the truth
to say that the way of life of an organism is determined by the nature of
the large molecules synthesized by its cells. The familiar difference
between the higher plants and animals furnishes an example as illustration.
Animals are able to move about or to move their parts because certain
of their cells have the ability to synthesize contractile muscle proteins.
Plant cells have largely lost this power; on the other hand they are
able to form and secrete quantities of rigid encrusting substances,
such as cellulose, which permit of a very different anatomy and mode
of life.

This statement is indeed equivalent to saying that animals have muscles
and plants have woody cell walls; but the emphasis is different. In
drawing attention in the first place to the macromolecular content of the
cell and the organism as the basic factor determining behaviour, we
are led to ask a particular kind of question, of which the following are
examples :

(a) What are these macromolecules and what is their molecular
structure ?

(b) Can we predict their biological function from their structure ?

(d) When do they appear in the course of embryonic development
(molecular ontogeny) and how do they influence this development ?

(e) How and when were the various molecular types evolved (molecular
phylogeny) and how has their appearance influenced the course of
evolution ?

Attempts to answer these questions are already engaging much attention
and undoubtedly theoretical biology of the future will base itself largely

2 KERATIN AND KERATINIZATION

on the answers arrived at (Perutz, 1959 and Schmitt, 1960). Even today,
when only partial and tentative answers can be given, the effort is made to
pose the problem in a form which envisages an answer in terms of the
physical chemistry of the molecular constituents of the system. Thus,
while it may be admitted at the outset that the treatment is foredoomed to
be incomplete, it is in this spirit that we shall approach the study of keratin
and keratinization. The discussion of keratin is not in itself an unfavour-
able theme for the purpose of systematizing parts of the already consider-
able amount of information on macromolecular biology since, largely as a
result of the work of Astbury, this protein and others closely related to it
have been shown to occupy a central position in any such discussion.

The keratins form a class of resistant, insoluble proteins found in the
vertebrate epidermis and its appendages: hairs, feathers, claws, horns,
etc., and in small amounts in certain of the internal epithelia. The name
keratinization is given to the process by which these tissues are rendered
tough and insoluble. Together with the dermis, which is in effect a closely-
knit fibrous meshwork of the protein collagen, the epidermis and its
derivatives constitute the protective integument. Everyone is familiar in a
general way with the properties of the keratins: their insolubility, their
toughness combined with elasticity, the enormous variety of forms
assumed. Our object will be to try to correlate these properties with their
molecular basis and to sketch (for that is all we can do at present) their
development. Attention will be mainly directed towards cellular and
subcellular structures for, although the macroscopic anatomy of these
tissues is of great variety and interest, it would carry us beyond the present
intention to attempt a detailed description. For convenience, the salient
facts of their anatomy will be recalled ; for the details reference may be
made to the specialized texts referred to on p. 80.

Current research emphasizes the role of two classes of macromolecules
in biology: the nucleic acids and the proteins. The genetic functions of
cells devolve on the nucleic acids which, it is currently believed, alone
exercise the power of self-replication and of controlling the replication
of other vital molecules (Crick, 1958); the proteins are components of
most of the other working parts of cells and of many extracellular deposits.
On the basis of their degree of aggregation two classes of proteins are
distinguished : the corpuscular proteins, that is proteins normally carrying
out their function in a particulate form in solution or adsorbed on surfaces.
Examples of this class are the respiratory proteins and most enzymes.
The other class is that of the structural proteins, which carry out their
function, often in part mechanical, by virtue of their property of forming
large aggregated, often fibrous masses. Among these we place, for example,
the contractile muscle proteins and the various reinforcing or protective
proteins including the keratins.

KERATIN AND MOLECULAR BIOLOGY 3

The special suitability of the proteins for these various roles is attributed
to their being high molecular weight, long chain polypeptides:

R R R

I I I
CH.CO.NH.CH.CO.NH.CH.CO.NH

.... CHR . CO . NH . . . . = amino acid residue
R = side chain
with a large number of reactive side chains which, as a result of the folding
of the main chain, may form a variety of surface patterns permitting
specific interactions with other molecules, large and small, and with other
proteins (Frey-Wyssling, 1953). The fibre-forming habit, common among
the structural proteins, is seen as a special example of protein-protein
interaction which can lead by association to macromolecular aggregates.
Here we are concerned more with the structural proteins, so named
because they are major constituents of the large-scale structures of cells
and tissues; in their mass they are readily visible in the light microscope
and obviously are related to the function of the organ in which they occur.
Resistant structural proteins are produced by introducing various cross-
linkages between the polypeptide chains.

It is a basic assumption that structure and function are interrelated and
for this reason the determination of structure is the primary concern of
macromolecular biology. Two other problems, which relate macro-
molecular structures immediately to main themes of biology, morpho-
genesis and differentiation, are the biosynthesis and the appearance of the
molecules in the course of individual development. Further, since
biological macromolecules were developed in the course of a long
biochemical evolution, there are phylogenetic as well as ontogenic aspects
to the problem of their adaptation to a biological role. In these terms then,
keratinization may be looked upon as a particular phenomenon character-
ized by the appearance in epidermal tissues of certain macromolecules
giving rise to gross structures capable of a protective function.

Orders of magnitude

The objects which are the concern of molecular biology range in size
from small molecules of diameter a few Angstrom units to massive materials
which can be examined with the unaided eye. It is helpful when attempting
to form a conception of the vast range of size involved to have some
visual aid to hand and for this purpose reference may be made to Fig. 1.

Orders of magnitude are indicated on the right by a logarithmic scale
and various levels of organization : molecular, macromolecular, cytological
and histological are distinguished. The formations at the higher levels
of organization are constructed from the smaller macromolecular and

4 KERATIN AND KERATINIZATION

molecular elements most easily studied by electron microscopy or X-ray
diffraction.

A brief description of these and other specialized techniques will be
given in the course of the text when it is necessary to make the discussion
clear, but for an adequate account of the methods reference must be made

Level of

Organization

Structure

Instrument

Dimensions

TISSUES

-100M

HISTOLOGICAL

I
G

CELLS

~ 10m

CYTOLOGICAL

ORGANELLES

c
o
p

(bacteria )

E
L

E
C

- lM

T
R
0

— . —

rv0.2/J

(viruses)

C
R

-0.1m = 1000A

MACRO-

R
A
Y

MOLECULES

MACRO-

MOLECULAR

MICRO-

F
F
R
A

T
I

— 100A

MOLECULAR

_ 10A

MOLECULES

physico-chemical

metho

Fig. 1. Levels of organization of biological structure and range of
instruments.

KERATIN AND MOLECULAR BIOLOGY 5

to the various texts available (Bunn, 1946; Hall, 1953; Schmidt, 1924;
and Oster and Pollister, 1955-56).

It will be seen from Fig. 1 that with present-day instrumentation we
possess the means to investigate all levels of biological significance. Thus
it is now feasible to hope to obtain the information necessary to permit of a
stepwise reconstruction of a tissue beginning with elements of a
molecular size. Already the results of comparative fine-structure studies
have indicated the nature of the basic structural elements found in all
cells and developed to different degrees in different types of cells. Its
particular endowment of these structural elements gives a cell its
characteristic cytology and, in turn, from the special association of these
cells arises the histology of the tissue. For example the keratinized tissues
themselves achieve their primary function as the toughened outposts of
the protective integument by virtue of an enhancement principally of two
structural features common to all Metazoan cells. In the first place all
cells contain greater or lesser amounts of structural fibrous proteins; in
keratinized cells there is a great increase in these proteins which are
subsequently subjected to chemical changes which stabilize them.
Secondly, the cells of most tissues must adhere and in the purely cellular
tissues, such as the epithelia, this means that the surfaces of the cells
themselves must stick together. In the specialized protective epithelia
including the keratinized tissues, clearly this intercellular adhesion must
be enhanced. Thus our discussion of keratinization will be found to be
concerned largely with these two aspects of the tissues : the nature of the
stabilized proteins within the cells and the manner in which the cells
themselves are held together. We shall accordingly begin by giving a
general survey of the types of fibrous proteins found in tissues and follow
this with an account of the fine structure of cells with emphasis on the
structural devices associated with their surfaces which may be concerned
with intercellular adhesion.

Types of fibrous proteins and their classification

The classification of the proteins has been made on the basis of several
grounds all more or less arbitrary. We have mentioned the convenience
of the division into structural and particulate proteins, which is a useful
distinction when one is concerned with insoluble proteins forming
structures extending far beyond the molecular level. Traditionally the
soluble, particulate proteins are classified on the basis of their solubilities
in various aqueous media, a means of distinction which dates back to a
time when the principal preoccupation of the protein chemist was the
separation of definite individuals from mixtures.

The basic chemical character of a macromolecule (protein, poly-
saccharide, nucleic acid) may be established by the chemical analysis of a

O KERATIN AND KERATINIZATION

sample after complete hydrolysis. In this way the familiar fibrous proteins,
collagens, keratins and silks were early shown to be distinct proteins,
characterized by the different amounts of amino acid residues forming
different patterns (Fig. 2). The common polysaccharides, cellulose and
chitin, yield glucose and acetoaminoglucose, respectively.

There have been numerous analyses made of keratins beginning with

==%-

Alanine
Valine

Leucine + isoleucir
Serine
Threonine
Phenylalanine
Tyrosine
Proline

3 § Hydroxyproline

Lysine
Arginine
Histidine

Tryptophane
Aspartlc
Glutamic

Amide N

— - — i $ Cystlne/2

Methionine

Fig. 2. Number of amino acid residues in 105g protein in three fibrous
proteins from Tristram's data (1953).

(1) Silk fibroin (s)

(2) Collagen (c)

(3) Wool (w)

§ Hydroxy-proline is only found in collagens and is regarded as " diag-
nostic."
If The cystine is given as " half cystine " (HS.CH.NH2COOH).

KERATIN AND MOLECULAR BIOLOGY 7

those made by Abderhalden and Voitinovici in 1907. Most of these are
today of little more than historical interest, but they served to sketch out the
broad outlines of the amino acid pattern of the group and to establish a
family unity. No purpose would be served here by reproducing these data.
Reference may be made to the paper by Simmonds (1955) in which a
table of all the results will be found.

Table 1 . Percentage Weights of Amino Acids
in Wool

Amino acid

c +

glycine

5-16

5-5

5-26

alanine

3-71

4-3

3-73

valine

4-96

5-7

5-78

leucine

7-63

8-9

7-69

tsoleucine

3-07

3-7

3-79

serine

9-04

9-9

7-15

threonine

6-55

5-56

6-58

phenylalanine

3-43

4-0

3-40

tyrosine

6-38

5-5

410

tryptophane

2-10

0-94

—

proline

7-28

6-8

6-58

hydroxyproline

—

lysine

2-82

3-3

3-15

hydroxylysine

0-68

—

arginine

10-49

9-8

9-03

histidine

0-90

1-2

—

aspartic acid

6-69

6-8

6-29

glutamic acid

14-98

14-5

12-8

ammonia

1-42

—

cystine and cysteine

11-3

10-3

11-0

methionine

0-69

0-56

0-55

* Simmonds (1955).

f Corfield and Robson (1955).

% Quoted by von Bergen (1954).

The estimations in J were made by microbiological methods which
may not be as accurate as those used by the other authors.

At the present time we possess reliable analyses by modern methods,
in which the greater part of the nitrogen of the proteins has been accounted
for, only of wool and feather. For purpose of reference we reproduce here
the analyses reported by Simmonds (1955) (Table 1 (a)), by Corfield

8 KERATIN AND KERATINIZATION

and Robson (1955) (Table 1(b)) and those quoted by von Bergen (1954)
(Table 1 (c)) all of samples of wool, and by Schroeder and Kay of
feathers (1955) (Table 2). Although it contains older and less reliable
data, Table 3, taken from Ward and Lundgren (1954), is valuable in that
it permits a comparison between several varieties of keratin. The question
of the significance of the differences between the figures given, even for an
apparently homogeneous material such as wool, will be discussed later
(p. 31).

Table 2. Amino Acid Composition of White Turkey Feather Parts and of
Goose Feather Barbs and Goose Down.*

(Values are in terms of g of amino acid per 100 g of moisture- and ash-free

material.)

Amino acid

Turkey

Goose

barbst

calamus

medulla

rachis

barbs

down

glycine

7-25

9-60

8-90

1014

8-38

7-26

alanine

4-01

7-12

5-89

7-66

4-10

3-96

valine

8-60

8-43

8-59

8-65

7-34

7-89

/joleucine

4-98

3-94

3-96

3-90

4-58

4-76

leucine

7-26

8-85

8-07

9-37

7-68

7-74

serine

12-90

15-09

12-37

14-09

12-53

12-38

threonine

4-68

4-73

4-35

4-51

4-94

5-42

phenylalanine

4-96

5-76

5-59

5-75

4-04

3-80

tyrosine

2-32

3-97

3-81

2-91

4-46

3-69

proline

10-50

10-98

10-87

10-97

10-05

9-81

lysine

1-23

0-98

1-32

0-88

1-30

1-41

arginine

6-44

6-69

6-57

6-18

6-04

6-52

histidine

0-39

0-59

0-78

0-34

0-44

0-33

aspartic acid

6-55

7-09

7-01

7-41

7-47

7-26

glutamic acid

9-08

8-74

8-60

8-84

8-99

9-06

ammonia

1-85

1-43

1-35

1-49

1-91

1-89

cystine

8-68

8-29

8-10

8-48

10-75

11-31

methionine

0-36

0-34

0-44

0-39

0-25

0-32

* Data taken from Schroeder and Kay (1955).

f For an explanation of the parts of a feather see p. 30.

Even when the chemical composition of a fibrous macromolecular
material is known, its detection and characterization in cells and tissues
may offer difficulties in routine histochemistry. For some materials
reliable histochemical tests have been developed (see p. 29); others
are recognized more or less negatively, simply by their fibrous character
and their intractable behaviour towards the usual stains and reagents.

KERATIN AND MOLECULAR BIOLOGY

«5

* ?

^4"

•7* ?" CP

^; Lp

IO £ w

CO «b

^ SO tH

CO CN

*- c

1 i

.32 H

sO -^^
CN <- qs O CN co CO 00 CN

U-) •>+■ ■+

-t- **

Os T-l sO

CO p

ts'S

4^.^- sb 4=Osb^CNco

sb CO CO

SO OS

lb CO ©

CN ~

WX»«

1-1

cn cooOfp9

O 00 CN

in t^-

u-, r^ t~»

CN Lp

w c

sb co cb sb lo o

4- 4 cn

r^ as

i^t^ 6

CO o

Feath
chick(
hard

O O Os CN •<*- co ■<*- co r* CO

(Ni-O

l l
cop

l I I

iflO cp

eo^-^-

i>> ^h cn r^ in cb r-» m 4 cb

64<n

i>, Os

sbA ©

sb o o

CS-f

-b io

„ 3

— o-U

= 0 H

l i

co p h> so
ib co ib co

r-c OS Cp

r^ -o
cc r-

r* cn ©

o
6

o co
cb o

C a-c 8

O O iri (T) 00 o

1^ sO o

■*■

r^-cN so

sb co ib CO 4 4

r~-

6 co ^-

in CN i^-i

c *> —

£ -r: ^3

■<n^

I 5x 8

00 *« t^» sC ip co sO co CN CN

00 T1 *>

r-- co

°? ■*■ ^P

ic.lflOO

i-nn Os CN LO t^ ■>}- co CO

sb sb co

t~» co

sb CN o

Os CN CN Osco °? sO sO

sOiflO

r-» cn

00 T ^

O O CO

so "-< ■*■ iocb,*'OOON

6»m

r»4

O co ■^i

T^-l

00 ~ o

7 i i TT i i 7

T l i

1 *!*

1 1 1

■a E"H

,2 3 w

in »-c O t— 00 m ■<$- r^ ■<*- co

t*-©CN

qs so

Os Os sO

■*

so r^- ip

EjCJS s

lb 4i ib 4 CN lb sb 4 CN 4

r^- r^cN

cb co

cc ^ 6

sboo

Os co © IT. + CA r< lfl O ^

ifliN'f

rp ©

sO CO ^

r~ r^

Cth4 sb 4 ib cb 4 4 ob

Os sO sO

ti. sb

6co ~

co O

"5 a

o sr-o

III 1 1 1 1 1 1 1

1 I

1 1 1

1 1

O o> u-

<N t-. O CN 't- O sO •- -1" co

rlvOO

■*- *7"

CN 00 CN r-

OT+

\Oiin lonir, tsMMui

r^ sb 4

sb cb

O^CNOO

T^-OO

."2

■S« 3

to P +- U

-si 1

1-11

Ker;
sour
type
stru

2s -Is si

g,-3|8!0£i:ag|.S8

C 4>

£■§8 8

^ a 3
3 «"&

s is

3 rtJC-xix;

^-c 5

3 c.S v §.S

o a 3 C"£ «

S is o. t- £ sH-

•- £3 S

C 03 w CS

KERATIN AND KERATI NIZAT ION

Birefringence

The presence of oriented material, either fibrous or membranous, is
most conveniently recognized by its optical properties using the polarizing
microscope (Schmidt, 1924). The method is extremely sensitive, but
yields no information concerning molecular structure.

All major biological fibres and, in particular, the keratinized epidermal
tissues are birefringent (double refracting) and the direction of maximum
refractive index coincides with the fibre axis as inferred from microscopic
inspection. The birefringence is quantitatively defined as:
At? = Vu - 7^

when r/n and r]± are the refractive indices in a direction parallel to the
fibre and at right-angles to it. In practice, usually a retardation R is
measured for a sample of material of thickness d and A17 = Rjd. Some
typical values for fibres are given in Table 4.

Table 4. Optical Constants of Hair and Related

Fibres: -qn Refractive Index Parallel to Fibre Axis;

■qL Refractive Index Normal to Fibre Axis

(Al7 = Vw ~ vD-

Fibre

V\\

wool (dry)*
fibroin (silk)f
cellulose (ramie)f
nylon (polyamide)f

1-5633
1-584
1-599
1-580

1-5494
1-529
1-532
1-520

0-0139$

0-057
0067
0-060

* Barnes (1933).
f Frey-Wyssling (1953).
I The value for wool taken as a typical, non-medullated, keratin with
a thin cuticle (i.e. consisting largely of oriented fibrous keratin) is
regarded as a maximum value. Other observers quote figures for A17
nearer to 0-01. Not infrequently lower values (<~ 0-008) can be measured,
but this may be presumed to be due to deterioration. Whether A77
approaches a quite definite figure, characteristic of pure, dry, undamaged
fibrous keratin has not been established.

The actual values of A77 are not greatly used for identifying fibre-types
since, being influenced by many factors (such as swelling, imbibition of
liquids and tension), they vary greatly and further, although the retardation
is easily observed and measured with a compensator, for microscopic
objects in cells the thickness is not so easily determined. Nevertheless
the polarizing microscope is a most sensitive instrument for the recognition
of oriented material and for detecting changes in orientation produced by
deformation, chemical effects, heat, etc. It can be particularly useful when
studying the degree of stabilization of an oriented material.

KERATIN AND MOLECULAR BIOLOGY

X-Ray Methods

With the growing hope of actually determining the molecular structure
of proteins by means of X-ray diffraction, or at least of obtaining certain
experimental criteria of structure even when this may not be determinable
in detail, the possibility arises of devising a rational system of classification
based on molecular structure. The simple procedures based on X-ray

Photographic plate

Fig. 3. The principal features of an X-ray diffraction apparatus.
T, X-ray generator consisting of a copper target on which impinge
high-speed electrons emitted in vacuo from the heated cathode E; W
is the window from which the rays emerge and are collimated into a
narrow beam at C to fall on the specimen mounted at S. The diffracted
X-rays are recorded on a flat photographic plate placed at right angles
to the beam at P. Other forms of plate may be used but the flat plate
is commonly employed in fibre studies. The diffracted rays within a
few degrees of the plate centre are referred to as the low-angle pattern ;
the more widely-scattered reflections constitute the wide-angle pattern
used to characterize the fibre-type.

diffraction have been of value for the rapid survey of material as is required
in comparative biochemistry replacing the laborious chemical methods.
Another advantage is that the material is not destroyed and need not
necessarily be isolated pure. Its biological applications were pioneered
by Astbury and developed in particular by Rudall. It has played a great
part in the study of keratins.

The first X-ray patterns of biological materials were in fact obtained
from natural fibres, among them hairs. The commonly-occurring fibrous
materials have each been shown to give characteristic X-ray patterns of

12 KERATIN AND KERATINIZATION

the fibre-type when examined by monochromatic X-rays (Astbury, 1933).
It is significant and, at first sight, surprising in view of the astronomical
numbers of stereoisomers possible for a polypeptide chain, that, judging
from X-ray patterns, very few fundamental arrangements of chains are
actually found in nature.

The material in a suitable fibrous form is mounted at right-angles to a
narrowly-collimated beam of X-rays (usually CuKa radiation) and the
diffraction pattern is recorded on a photographic plate as shown in Fig. 3.

Fig. 4. Illustrating Bragg's law for the reflection of an X-ray beam,
i is the incident beam and r the reflected beam. The crystal consists
of many parallel planes containing atoms which can scatter the incident
radiation. The set shown consists of planes a distance d apart. When
the crystal planes are so oriented with respect to the incident beam that
the angle 6 satisfies the Bragg equation : nA = 2d sin d, reinforcement of
the scattered radiation occurs giving a definite reflected beam. In any
natural fibre a great many crystallites having the appropriate direction
will occur.

The diffraction pattern is immediate evidence of a characteristic
arrangement of the atoms in the specimen, but we need not pause at this
point to discuss its interpretation (see Chapter 5). It is sufficient for our
present purposes to accept each pattern as a sort of "finger print" revealing
the presence of the fibre-type in question. The method is simple and,
when a positive result is obtained, reliable. It has the drawback that it
tends to overemphasize the crystalline fibrous components, which alone
give recognizable patterns, and to overlook non-crystalline components
which may contribute importantly to the properties of the material.
Patterns are recognized partly on sight, using standard examples such as
are reproduced in Plates 1 , 2 and 3 as guides. The positions of a limited
number of characteristic reflections are also measured and a corresponding
lattice spacing calculated from the Bragg equation. Crystals consist of a
large number of parallel planes containing atoms each of which may
scatter X-rays from an incident beam. It was shown by Bragg, that only
when certain geometrical conditions are satisfied, is a definite reflected
beam of X-rays obtained. This condition may be understood from Fig. 4.

KERATIN AND MOLECULAR BIOLOGY 13

The equation which must be satisfied by the angle 6 between the incident
radiation and a given set of planes is the Bragg law:

nX = 2d sin d

Where n is any whole number (the order of the reflection), d is the spacing
between the set of planes and A the wavelength of the X-rays. Thus,
from the photographic plate, the angle 6 is determined from the distance
of any reflection (" spot ") from the central spot due to the undeflected
beam, A is known from the target used (in biological work it is often copper

itHrHf.

i' 1 [ !'
_-ff-f-4-+4-4^

<i i l \T

A ' * V

Fig. 5. Idealized diagram of a typical " X-ray fibre photograph " taken
with the X-ray beam perpendicular to the fibre-axis and to the photo-
graphic film (fibre axis vertical). The reflections are seen to lie along a
series of hyperbolae referred to as layer lines. Redrawn from Astbury
and Bell (1939).

giving a wavelength of 1-54 A for copper K a- rays) and thus the distance d
between the planes can be calculated. Notice that the smaller d is, the
greater the distance the reflection occurs from the centre. For this reason
the smaller spacings < 10 A, arising from close-packed atoms and
characteristic of the molecular level of organization, are found at wide
angles ( > 10°) and constitute the characteristic " wide-angle patterns" '.

With single crystals of pure substances the practice is to rotate the
crystal in the beam to give all sets of planes in it a chance to pass through
the Bragg angle and thus register themselves photographically. In natural
high-polymeric materials in the fibrous form, this is not necessary because
these consist of innumerable small crystals (or crystallites) with one
principal direction parallel to the fibre axis, but randomly arranged in

14 KERATIN AND KERATINIZATION

other senses. Thus when the fibre is mounted at right-angles to the
beam (Fig. 3) every possible orientation will be present and reflections
from all sets of planes will be possible without rotating the fibre. Such
patterns are called " fibre-type patterns " (Astbury, 1933) and from them
the most important characteristic of a fibre, the distance along the axis at
which the molecular pattern repeats, can be immediately calculated. The
diffracted rays emerge from a fibre (or a rotated crystal) on a series of
cones about the fibre axis, and since these intersect a flat photographic
plate, mounted as in Fig. 3, in a series of hyperbolae, we find the spots
lying on these hyperbolae (Fig. 5) which are referred to as layer lines
(Bunn, 1946).

Fig. 6. Diagnostic reflections of the oc pattern.

The patterns given by biological fibres are not usually very sharp and as
detailed as those of well-formed crystals. This is in part due to the small-
ness of the crystallites and in part to their imperfect orientation which
has the effect of drawing each spot into an arc. These defects render a
strict crystallographic determination of structure well nigh impossible.
but in no way hinder the use of the patterns for recognition purposes.

The a, j8 and Collagen Patterns

The principal features of the main, fibre-type, wide-angle X-ray patterns,
which constantly recur in discussions on biomolecular structure, are
summarized below. The classification and characterization are largely due
to Astbury (Astbury and Bell, 1939).

(a) The oc-pattern (Plate 1A, Fig. 6). Type material, mammalian hair.
Characteristic features :

(i) the axial repeat spacing appears as a strong sharp meridional arc
corresponding to 5-1 A (5-05 — 5-15 AV

KERATIN AND MOLECULAR BIOLOGY

(ii) two very strong diffuse extensive areas of reflections symmetrically
disposed on the equator and centred about 10 A referred to as
" side spacings ";
(iii) a sharp meridional reflection at 1-5 A not usually recorded on flat

photographic plates with an arrangement such as that of Fig. 3 ;
(iv) a strong diffuse halo centred about 4-2 A forms a background to
the sharper fibre pattern.
A definite feature of a material giving an a-pattern is that by stretching,
pressing, or heating in water, it can be transformed into an isomer giving
a jS-type X-ray pattern (see below).

The number of proteins capable of giving an a-pattern is large and
includes, in addition to all types of mammalian keratins, actomyosin, the

Fig. 7. Diagnostic reflections of the ft pattern.

contractile muscle protein, the blood proteins, fibrinogen and its insoluble
form fibrin, bacterial flagella, etc. There is reason to suppose that most
intracellular fibrous proteins are of the a-type. Astbury refers to the group
as the kmf proteins (keratin-myosin-fibrinogen).

(b) The fi-pattern (Fig. 7 and Plate IB). Type materials, silk fibroin,
stretched hair. Characteristic features :
(i) axial repeat spacing about 3-5 A;

(ii) strong side-chain reflections (in the keratin-type) centred around
10 A resembling those of the a-form. In the simpler silks the
spacing may be much less (3 — 5 A) ;
(iii) strong broad symmetrically disposed spots on the equator at

4-5 A called the " back-bone spacing ".
(iv) a diffuse halo about 4-2 A apparently identical with that of a-

patterns, but less well developed in the well-crystalline silks.
The numerous silks are usually, but not always, of the /3-type.

16 KERATIN AND KERATINIZATION

(c) Avian fi-keratin pattern (Plate 2A). This pattern is related to the
normal /S-pattern in having most of its characteristics. It is distinct in
having a shorter (apparent) axial repeat spacing of 3-1 A and in being
generally far more elaborately developed. Type material, bird feather
calamus or rachis.

(d) The collagen pattern (Fig. 8). Type material, rat tail tendon.
Characteristic features :

(i) axial repeat ^ 2-8 A;

(ii) diffuse side-chain spacing at ~ 12 A. Unlike the corresponding
spacing at 10 A in the a-patterns, the side spacing of the collagens
is sensitive to hydration and may increase to 15 A and more in
swollen materials.

#••#

Fig. 8. Principle features of the collagen diagram.

On stretching, collagen fibrils usually break at about 10 per cent extension
without any change in the type of X-ray pattern yielded. The reflections
are considerably sharpened (Randall, 1953). Heating, swelling and other
destructive influences likewise do not change the nature of the pattern
although reducing its intensity. Collagen fibres are common in vertebrate
connective tissue and have a wide distribution elsewhere (see later).

Notes. Numerous other reflections are visible in all patterns and are
shown for a- and ^-patterns in Figs. 9 and 10; they have been variously
described and are indexed in Astbury and Bell (1939). They are usually
less strongly developed and may not be detectable. When the specimen
is poorly oriented the various reflections are drawn out into arcs and are
less readily recognized. The completely disoriented condition is not
uncommonly met with, particularly in artificial preparations and the
corresponding patterns may be recognized as follows :

(e) Disoriented oc-pattern. Type example, regenerated precipitates of

KERATIN AND MOLECULAR BIOLOGY 17

a-keratins. Only two obvious spacings can be recognized corresponding
to the 10 A side spacing and the original 4-2 A halo, which here appears
with a dense and sharper inner circumference due to the presence of the
definitive 5-1 A spacing.

040

'331

1 20 and 020

'll2

010

9 • o •

100 100

etc.

y
y

Fig. 9. An early attempt by Astbury and Bell (1939) to indicate the
many other reflections visible in an a-pattern. The " indexing" (see Bunn,
1946) in terms of an orthorhombic cell: a = 27 A, b = 10-3 A,
c = 9-8 A, would not be accepted by all. See also Fig. 68, p. 166.

(f) Disoriented ^-patterns (Fig. 11). Type example, boiled egg white.
The characteristic and often very sharp ring due to the 4-5 A backbone
spacing appears overlaid on the diffuse halo. The side-chain spacing
is sometimes less well defined.

(g) Non-crystalline protein pattern. Many proteins when dried yield a
very vague pattern consisting of two diffuse haloes centred around 10 A
and 4-2 A which is distinguishable from the unoriented a- and /^-patterns
by the absence of either a 5-1 A or 4-5 A reflection. A close inspection
at the inner edge of the outer halo may be necessary to recognize the
absence of the a-spacing.

It should be emphasized that these X-ray patterns are not indicative
of single specific proteins but of families of proteins which have in common

KERATIN AND KERATINIZATION

030

/220 020 N

<4I0

**' A

1400 |200# OOOI0 |

Fig. 10. The /3-pattern indexed in terms of an orthorhombic unit cell
a = 9-3 A, b = 6-66 A and c = 9-7 A. The 001 and 200 reflections are
those ascribed to reflections from the two main dimensions (side and
backbone) of the polypeptide chain. The assignment of indices shown
in Figs. 9 and 10 should be regarded as tentative only. Many of the weaker
reflections indicated in the drawings (Figs. 9 and 10) are often obscured
by the extensive halo pattern also present and are rarely measured
(Plate IB). Figures 9 and 10 have been redrawn from Astbury and
Bell (1939).

Fig. 11. Idealized X-ray diagram of disoriented /? and denatured
proteins.

• •••'•

Fibre axis

Plate 1 (Captions overleaf)

Plate 1

A. The a-type X-ray diffraction pattern. Material mohair, fibre to plate
distance 4 cm, Ka radiation. For a description see pp. 14 and 17.
The strong meridional arcs corresponding to 5-1 A on the vertical axis
and the very strong reflections ( ~ 10 A) on the equator are the two most
characteristic features. Strong spacings may be seen on the equator cor-
responding to 27-30 A and on the meridian a series of faint " long
spacings " can be seen. Fibre axis vertical.

B. The /3-type X-ray pattern given by stretched and set fibrous keratins
and other a-proteins (stretched Lincoln wool). The 5-1 A reflection of
o:-keratin is missing and a further arc (3-5 A) has appeared. On the
equator strong reflections at 4 -5^1 -6 A are present. Other reflections on
the layer lines may be made out.

Both fibre patterns are overlaid by a wide diffuse halo centred about
4-2 A.

Photographs kindly lent by Mr. H. J. Woods.

KM I

• •

Fibre axis

Plate 2 (Captions overleaf)

Plate 2

A. Pattern of sea gull quill, the most detailed yielded by any keratin. It
is of the /i-type (Fig. 7) as shown by the meridional arc at 3 '3 A and
by the strong spacings ( ~ 10 A and 4-5 A) on the equator. For a list of
the spacings see Tables 10 and 11.

B. The cross-/? pattern obtained from Lincoln wool by treatment with a
solution of urea containing bisulphite and restretching (see p. 200). The
4.6 A formerly on the equator now appears as an arc on the meridian.
See Fig. 84 and compare with Plate IB.

Figures kindly provided by Mr. H. J. Woods.

Fibre axis

Plate 3

A more elaborate version of the a-pattern obtained from the proteir
of the mantid ootheca, see p. 204.

Figure kindly provided by Dr. K. M. Rudall.

KERATIN AND MOLECULAR BIOLOGY 19

certain structural regularities arising from the configurations of their
main polypeptide chains. For example, a-type patterns, very nearly
identical at large angles, are given by a whole class of proteins, including
keratin, myosin, fibrin, etc., which, for this reason, is referred to
as the kmf etc. . . . family. The specific differences between proteins,
which determine their function, concern superficial chemical groups whose
presence usually affects little the main chain structures yielding the
patterns described above.

More recently it has been found that some fibres have a characteristic
fine structure visible in the electron microscope and this may be used to
identify them. Of particular value in this respect is the appearance of
some kinds of collagen fibril which typically display a longitudinal spacing
of 640 A. A recent demonstration of collagen in the neural tissues of
certain insects was based on their electron-microscopic appearance
(Gray, 1959 and Hess, 1958) and confirmed the earlier reports by Rudall
(1955) based on X-ray evidence.

It is currently believed now that the fundamental configurations of the
main chains of polypeptides are determined in advance by the stereo-
chemical characteristics of the component residues (see Chapter 5), which
permit a limited number of stable configurations, and that these are
spontaneously assumed in solution by free chains if circumstances permit
it. Since an almost unlimited variety of side-chain composition is com-
patible with the main-chain configurations, these stereochemical demands
place little limitation on the functional possibilities of proteins.

Stabilization

The fibrous macromolecules forming the protective coatings of organisms
are usually subjected to a stabilizing process which may take various
forms and be developed to varying degrees. As a result of this process
the protective layer is hardened and insolubilized. The simplest means
of effecting changes of this sort is by crystallization between the long
polymer chains. When the chains are of a simple character, or of such a
regular shape that they readily fit together, localized crystallization may
develop with the formation of crystallites, which, when stabilized by
sufficiently large energy of crystallization, virtually lock the chains together
and thus render the network insoluble. The effectiveness of this device
is apparent in such materials as silk or cellulose which are very insoluble ;
yet when sufficient of the hydroxy groups of the pyranose rings of cellulose
are methylated to prevent crystallite formation, readily soluble methyl-
celluloses result.

A further very common method of insolubilizing proteins is the chemical
process known as tanning (Gustavson, 1956) in which covalent chemical
cross-linkages between the polypeptide chains are introduced by a reaction

20 KERATIN AND KERATINIZATION

with an accessory molecule usually an aromatic polyphenol (Q) (Hackman,
1953 and 1959):

h i r~

Q Q Q

I I I

The cross-linked protein may, in turn, embed a fibrous meshwork of a
different character. For example, in the arthropod cuticle, a chitinous
meshwork of fibrils, themselves stabilized by crystallite formation, is
embedded in a tanned ;8-type protein (-s) named arthropodin (Fraenkel
and Rudall, 1947). The result is a very rigid exoskeleton whose inflexibility
has had much to do with the evolution of these animals.

Brown (1949a and b and 1950) has discussed the various chemical and
histochemical methods for recognizing the presence of a tanned protein.
There is no entirely satisfactory method. A hard, insoluble, darkened
material is usually presumed to be tanned. The actual bond between the
quinone and the polypeptide is still in dispute. Pryor's original proposal
(1940) that the link involved the amino end groups of the polypeptides:

HO - ^NHP
HO

NH.P. . .

[P . . . = polypeptide chain]

and an o- or />-diphenol is not now accepted. If the material is not dissolved
by keratinolytic agents, i.e. those breaking disulphide bonds and hydrogen
bonds (see pp. 234 and 236) but is dissolved by sodium hypochlorite
solution (an unknown reaction) it may be tanned. The demonstration of
diphenols in the secreting cells is strong support.

Hardness may also be influenced by the deposition in the protein matrix
of inorganic materials, usually calcium salts or silica. For example, the
crustaceans mix calcium salts with their chitinous cuticles. Probably for
reasons of weight, insect cuticles are usually free of calcium salts although
oxalates may be deposited in their egg cases (Rudall, 1955). The verte-
brates uniformly lay down calcium salts in association with their connective
tissue protein, collagen, to form bone as a rigid endoskeleton. Such
mineral deposits are rare in keratinized tissues. Pautard (in press) has
demonstrated by means of X-rays and by electron microscopy the presence
of apatite in " whalebone," which is not bone, as the name implies, but an
extensive horny proliferation of the oral epithelium in certain whales.
The enamel of teeth may also contain a keratinous component. See p. 78.

Although calcium has many important biological roles its concentration
within cells is normally lower than that in the surrounding fluids. Never-

KERATIN AND MOLECULAR BIOLOGY 21

theless in almost all animal and plant groups deposits of calcium salts may
be found intracellularly as well as extracellularly or on the cell membrane.
The salts appear usually in association with a protein matrix and it is
supposed that some spacial relationship exists between the salt molecules
(or ions) and sites on the surface of the protein molecules. In the verte-
brates the greater part of the calcium found in bone is always associated
with collagen (Fig. 12) and mucopolysaccharides. In this case, the
macromolecular form of the collagen is also important since, of the
several arrangements of the collagen molecules (see p. 128) which can be
formed in vitro, only that having the naturally occurring spacing of
640 A appears able to initiate salt deposition. The fact that cells con-
taining keratin seem rarely to accumulate calcium may also mean that
the special arrangements of surface groupings required is lacking in the
keratin molecule (Bachra et al., 1959).

A third method of hardening proteins is keratinization which resembles
tanning in that covalent cross-links are established between protein chains;
but these are of a special type, the sulphur bridges made possible by the
linking of cysteine residues in adjacent chains. Keratinized proteins are
also in part stabilized by crystallite formation as is shown in fact by the
existence of the a X-ray pattern.

~~r ~r

s s

I I

s s
! 1

— CH2— S— S— CH2—
the cystine " bridge " or disulphide link

All these methods of insolubilizing and toughening natural polymers
have their analogies in the chemistry of artificial polymers and much
understanding of the natural process has come from a study of these.

Valuable information concerning the stability of a fibrous system, which
is particularly relevant to our present interest in stabilized, protective
proteins, can be obtained directly by X-ray means simply by taking a
photograph after a tissue has been subjected to some disorienting influence
such as heating in water. If the structure has been disrupted by the
treatment, the fibre-type pattern of distinct spots or arcs is replaced by a
pattern of diffuse, circular haloes. A loss of birefringence may also be
detected by means of a polarizing microscope (Fig. 109, p. 213). The X-ray
pattern may also change its character. An a-keratin may frequently be
converted into a /?-type keratin by heating in water (Rudall, 1946) and

22 KERATIN AND KERATINIZATION

the new pattern may be that of disoriented jS-crystallites, which may be
oriented by stretching and then give an oriented jS-pattern.

Ecdysis. One consequence following from the fact that the chemical
changes involved in stabilization are irreversible, is that provision must be
made for the removal of the hardened layer to permit further growth. This
may take the form of moulting or casting when an entire covering of
feathers, scales, shell, etc., may be lost or, in the case of the keratinized
cellular epidermis, superficial cells may be constantly shed and a covering
of constant thickness maintained— a procedure which has the advantage of
not exposing its owner to a period of vulnerability such as follows the
moulting of a rigid covering. The factors governing ecdysis are obscure ;
they seem to be hormonal and are often geared to seasonal changes and,
in turn, to other cyclic hormonal-controlled activities, such as sexual
display, in which the external coverings may play a conspicuous part
(pp. 133 et seq.) (Turner, 1960).

Distribution of the fundamental fibre-types

There are two distributions to be considered : the first is the distribution
of the different macromolecular types among the parts of any particular
kind of organism, which raises questions of ontogeny; the other is the
sharing-out of molecular types between the entire range of organisms and
here there are problems of phylogeny.

The first of these distributions is linked with the problem of the
differentiation of organs during embryogenesis which is usually recognized
and defined by the appearance of the typical histology of the various
tissues as seen in the light microscope. These changes at a relatively
large-scale level are in fact partly the consequence of the appearance and
accumulation in the developing tissues of the characteristic macromolecules
under discussion. Differentiation may thus be described in terms of the
macromolecules on whose presence the future function of the tissue
depends.

In very generalized terms, the fertilized egg is biochemically and
structurally omnipotent; it is potentially capable of synthesizing all the
products later appearing in its descendant cells. As development proceeds,
these potentialities are shared out among the organ systems, each of which
finally makes a limited range of substances required for their special
function. Thus in the adult, many of the tissues may be characterized
by the fact that they contain a limited range of structural macromolecules
of which a few (perhaps only one) associated with the tissue's function,
greatly predominate in amount.

This is well illustrated by the sharing out of the fibre-forming potential-
ities in the vertebrates (Fig. 12). Certain groups of cells in the middle

KERATIN AND MOLECULAR BIOLOGY

layers of the embryo come to produce and to retain intracellularly the
contractile muscle protein (actomyosin) and thus become muscle cells;
others of a similar origin become fibrocytes and by secreting the fibrous
protein collagen help to build the connective tissues of the organism and
the lower layer of the integument. The superficial cells, on to which
devolves the special task of enclosing the whole system, commence to
differentiate early and lay down intracellular keratin.

The second distribution, that of the macromolecules among the different
phyla, is a fascinating problem with far-reaching implications for their

FlG. 12. Cross-section of an hypothetical vertebrate body to show dis-
tribution of fibrous proteins. The three major divisions are shown
schematically. E, the external cellular epidermis (ectoderm) containing
keratin fibrils K; the internal epithelia I (endoderm) are separated
by the middle layers (mesoderm) containing the musculature M and the
connective tissue (including dermis) containing collagen fibrils C.
Basement membranes BM (see p. 86) separate the epithelia from the
mesodermal layers.

evolutionary development. It would seem that the power to make protein
and/or polysaccharide materials which can be hardened by various
chemical devices is a primitive and persistent cell property (Fig. 16).
Organisms can call upon this property if and when necessary to form
integuments, egg cases and other hard parts. Thus no necessary cor-
respondence between phylogenetic relationships and the distribution
of hardened parts can be insisted on. Nevertheless in the event
some persistent trends exist.

24 KERATIN AND KERATINIZATION

The very convenient X-ray method, supplemented by chemical and
histological data, has permitted a fairly extensive survey of the types of
structural proteins and of their distribution among the various groups of
animals and plants. The results of this survey show that in structural
matters, as in biochemical matters in general, organisms are conservative
for, in the whole course of evolution, only the very limited number of
basic macromolecular structures described above have appeared and have
been adapted to the necessary variety of uses by introducing variations in
the side-chain composition. Given the limited range of biological polymers :
proteins, polysaccharides and nucleic acids, the number of solutions at
the molecular level to the problem of forming a protective integument is
seen to be severely limited. In a similar way the chemical devices available
to render proteins and polysaccharides more stable and tough are also
limited by the chemical possibilities of these polymers.

There has, nevertheless, been a degree of biochemical evolution in the
usage of fibre-types and the several great branches of living organisms are
endowed to different degrees with the various possibilities. Cellulose is,
for example, the typical structural support in plants, although a material
morphologically very similar (but chemically distinct) is found in the tunics
of the tunicates— creatures (protochordates) a long way from plants
and related to our vertebrate ancestors, and perhaps even in vertebrates
(Jeffery and Cruise and Keech, 1959; Cruise and Jeffery, 1959). Higher
plants lack the fibrous proteins associated with movement in animals.
Insects and crustaceans typically contain chitin in their hard parts
(Rudall, 1949 and Lotmar and Picken, 1950). They possess contractile
muscle proteins, but apparently little collagen (see Fig. 13).

The vertebrates, a comparatively uniform group, have most of their
protein fibres in common: all the land-dwelling forms harden their
epidermis with a keratinized protein and support the epidermis with a
dermis containing collagen. One remarkable difference distinguishes the
reptiles and birds on the one hand from the mammals on the other. The
mammals, in this sense at least the more conservative, have taken over
and keratinized an a-type protein probably similar to that occurring
primitively in cells and still found in the partly-keratinized skins of lower
vertebrates; in birds and reptiles a similar a-keratin is found in the softer
regions of the skin, but in their more characteristic hard parts, feathers,
scales and claws, an entirely different keratin of the 0-type, referred to as
feather keratin, is found. This discovery by Astbury and Marwick (1932)
was one of the early triumphs of the X-ray method of detecting and
classifying proteins.

The production of proteins having a particular type of polypeptide
configuration is undoubtedly a consequence of a genetically-controlled

KERATIN AND MOLECULAR BIOLOGY

activity in the cells forming the protein. Thus the appearance of a new
type of protein, such as feather keratin, would seem to imply a mutation,
which in this case must have occurred in the reptilian stem-line after the
mammal-like reptiles had branched off, but before the birds separated
from the main stream (Heilmann, 1926) (Fig. 13). An appropriate new
protein appearing in the epidermal cells would presumably be subjected
to the same processes of keratinization as its predecessor. In the instance
of feather keratin, its utilization by reptiles, the earlier forms, would seem

REPTILES, BIRDS

-<?}- Collaqen
-@- Chitin
-0- Keratin

Fig. 13. Distribution of collagen, chitin and keratin shown on a con-
ventional phylogenetic tree.

to be a simple replacement of the commoner a-type keratin conferring no
advantage obvious to us. In birds the situation is very different. The
/3-type molecule, as will be shown in Chapter 5, is inextensible and inflexible
in contrast to the (extensible) a-type molecule (also the basis of muscle)
and forms an admirable structural foundation for the economical con-
struction of feathers which must be both light and stiff. The successful
invasion of the air is thus partly based on the exploitation of this new type
of epidermal keratin.

With a sufficient knowledge of the distribution of types of macro-
molecules among existing animals and plants, combined with a knowledge
of the actual evolutionary descent of organisms, it would seem possible

26 KERATIN AND KERATINIZATION

to construct a geneological table for the major macromolecular species.
As soon as this is attempted, however, it becomes plain that, as stated
above, in its grand features biochemical evolution has been small relative
to structural. The major phyla are distinguished certainly by their
characteristic spectra of macromolecules (Fig. 13) but very clearly this

Fig. 14. The surface specializations of cells and their role in tissue
building and in the formation of an integument, (a) Generalized cell;
(b) isolated cell surrounded by a layer of mucin which may form a
pellicle ; (c) isolated cell with surface covered with motile cilia ; (d, f ) one
theory of the formation of a multicellular organism. Failure of cells
to separate after division owing to the adhesions of the superficial coats;
(e, f) alternative view of the origin of multicellularity, the cellularization
of a large cell (Hadzi, 1954). This procedure is followed in the subdivision
of the egg. (g) and (h) The two types of protective integument : (g) the
extracellular cuticle or pellicle secreted by the surface epithelium. This
must be shed to permit growth ; (h) an integument formed by the surface
layer of cells themselves which have synthesized a fibrous intercellular
material; this is the case in the keratinized tissues.

sharing-out of molecular types took place at a time anteceding anything
of which we have fossil evidence. The structural evolution of which we
have evidence has represented a shuffling of possibilities presented by a
more-or-less constant molecular endowment. Nothing may give a more
vivid impression of the unity of life than the recognition throughout its
manifestations of similar molecular species and of similar biochemical
devices based upon them (Florkin, 1960).

Since the distribution of macromolecular types occurred so early, a
consideration of this distribution and its relation to function should offer
a promising pathway to investigate the obscure field of primordial life and

KERATIN AND MOLECULAR BIOLOGY 27

its dispersion into phyla. What is of particular interest here, since it is
our object to discuss the integument, is the possibility that the molecular
basis, both for the appearance of multicellularity and for the separation of a
limited number of well-defined types of organism, is to be sought for in
the nature of the substances present on the surface of cells (Fig. 14).
These substances determine in the first place the intercellular adhesion,
the essential basis of the existence of cells in colonies (see Fig. 14(d),
(e) and (f)) and, in the second place, the types of material from which
superficial cells construct their protective layers, is correlated with the
divergent lines of evolution in such a way as to suggest the choice represents
a major cause of this dispersion.

The chemical processes and materials responsible for toughening tissues
are also distributed in a phylogenetically significant way. For example,
among plants we normally find the fibrous, polysaccharide cellulose
embedded in various encrusting substances such as pectins and lignins;
among the arthropods, we find the polysaccharide, chitin, embedded in a
tanned /2-type protein (Richards, 1951). Tanning is more or less univer-
sally distributed among animals and thus must be judged the more
primitive device for cross-linking protein chains. According to Mason
(1955) its wide distribution is related to that of the ubiquitous enzyme
systems, phenolases, which catalyse the formation of o-quinones from
phenols. Tanning seems to be the rule in invertebrates; in vertebrates
the process is found mainly in one kind of cell, the pigment producing
melanocyte, where pigment granules are darkened and hardened by the
formation of tanned melanoproteins (see p. 276). A fundamental difference
in body plan with related mechanical consequences thus arises between
the invertebrates, with their rigid, tanned exoskeleton on the one hand,
and the vertebrates, with a rigid endoskeleton of collagen and calcium
salts and a more flexible keratinized epidermis on the other. Some further
consideration of the phylogeny of keratinization itself among the vertebrates
will be given on p. 49.

The relationship, between certain fundamental structural " inventions "
and the molecular bases on which they rest, is set out in tabular form in
Fig. 16.

Some difficulties in denning a keratin

The definition of a keratin assumed above is that it is a hardened and
insolubilized protein found within the epidermal cells of vertebrates.
This definition covers almost all the proteins which will be discussed in
this work, but it certainly does not cover all those which might, on one
ground or another, be considered as having a claim to the name. There-
fore it is desirable to consider alternative definitions based on some
characteristic molecular structure or chemical feature. The mammalian

KERATIN AND KERATINIZATION

3 «j g vj

■5 s •« a

—

u
u

c
U

<-'

—

•a I

£ 2

^ I

g 5

J 3 ft ©

c £
.2 §
~ >
'E o

-:£

<38
is

C u

C ca

•o 2

■h!

T3 g>8

I'll

H « H

-9 3

£ s

~o J2 -2 <■>

"ft 4, 3 M

11 > "U .£

jj "-p u £

3 £ '3 2

— ° 2 c

8S. ft8

£ c o

tH-O-^

« £'■£ o

.£ 8*1 &

5) «

I 1^
ft *7j ""

«— £

o « is

-3 <-» js

« t+-i 3
G |5 c

Saw,
3 a u

t* « J=
« Ji «J

■£T3 ?

•a s

°13 °

C <* fj
4> <U U
03 — ft 03

D U -rs C

JJ c •-

o "o ••* c

= £ .2 3>

■*§ e°
^£ II

2 h S

£8

KERATIN AND MOLECULAR BIOLOGY 29

keratins are all of the a-type and they owe their stability and insolubility
primarily to the covalent cross-links formed between their polypeptide
chains by the disulphide bridges of the amino acid cystine. The proteins
in the epidermal hard parts of reptiles and birds are also stabilized by
disulphide bonds; on the other hand, having /^-characteristics, they are
of a quite different molecular structure from the mammalian forms.
Mammalian keratins on being stretched can also assume the |8-form as a
kind of stereoisomer. Thus a definition of a keratin must include both
a- and /3-forms. Further in some special situations, e.g. the hair cuticle
(p. 265), a peculiar, very insoluble and highly cross-linked amorphous
keratin is found. Thus a more inclusive definition might be forced to
place less emphasis on the molecular character of the crystalline form and
define a keratin simply as a protein stabilized by disulphide cross-linkages.
Emphasis needs to be placed on the fact that the cystine cross-linkages
produce insolubility as well as stabilization or such soluble proteins as
insulin (12-5 per cent cystine) would be included.

Block's definition (Block and Boiling, 1950; Block, 1931; Block and
Vickery, 1931) sums up a traditional view in placing emphasis on the
insolubility: "A keratin is a protein which is resistant to digestion by
pepsin and trypsin, which is insoluble in dilute acids and alkalis, in
water and in organic solvents." He adds, however, a further criterion
based on the molecular ratio of the basic amino acids found by hydrolysis,
which is not now admissible (pp. 31 and 32).

The routine histological tests for " keratin " assume that its presence
is to be inferred from the presence of protein-bound disulphide bonds.
They are based either on the oxidation of the cystine bridge to produce
the very acidic — S03H group :

P— S— S— P -> 2P— S03H

which is then detected by the increased basophilic at low pH (Pearse's
method, 1951 and 1953); or by reduction of the bridge to sulphydryl
groups: h

P— S— S— P -* 2P— SH

and the detection of these by the nitroprusside test, the Prussian blue
test (Chevremont and Frederic, 1943) or most specifically by coupling them
to the Bennett reagent: l-(4-chloromercuriphenylazo)-naphthol-2:

OH
CI— Hg— </_ N— N=N—

" mercury orange," Bennett's reagent (Barrnett, 1953)
to yield a coloured dye.

30 KERATIN AND KERATINIZATION

According to this definition various fibrous proteins of the internal
epithelia of vertebrates (oesophagous, vagina, etc.) would be keratins and
are generally so called. A purely epidermal location cannot therefore be
insisted on. Certain extracellular exudates in birds, such as the horny
linings of the gizzard and proventriculus, are hardened by the presence
of cystine bonds and are therefore also keratins (Broussy, 1932). The
so-called " ovokeratin," which forms, along with mucin, the fibrous
membranes of birds' eggs, is a more ambiguous case. Some analyses
(Calvery, 1933) stress a resemblance to wool keratin in composition (cystine
content 3-7 per cent); on the other hand X-ray photographs suggest a
protein of the collagen type (Champetier and Faure-Fremiet, 1938). The
fibrils are quite unlike those of other keratins; they consist (as seen
electron microscopically) of a core of osmiophilic material enclosed
in a sheath of less-stained material (Mercer, unpublished. Plate 18B).
They may thus consist of two distinct substances of which one could
be cystine-stabilized.

With such a broad definition the keratins cannot be considered as
exclusively limited to the vertebrates since there are instances reported
of cystine-stabilized proteins of invertebrate origin (Brown, 1949).
Krishnam (1953, 1954) has shown in a scorpion that the proteins of the
epicuticle (analogous morphologically to the tanned structures of other
arthropods) are stabilized by disulphide bonds and give an X-ray pattern
not identical with either the a- or jS-patterns of keratins or of the /S-pattern
of arthropodin. The cuticle of Limulus is reported by Lafon (1943) to
be similar. Brown (1949, 1950) has cited other instances of invertebrate
proteins stabilized by disulphide bonding although she is not inclined to
call all these keratins. There is even some suggestion that quinone cross-
linking may occur along with cystine cross-linking in some instances, or
perhaps the quinone may link directly to the thiol of a cysteine residue
(Hughes, 1959). Thus tanning and keratinization may become inter-
changeable or even mixed as a means for the insolubilization and stabiliza-
tion of proteins.

It may be claimed also that the spindle fibre protein, which plays an
important role in the mitotic cycle in animal cells, is a keratin. Mazia and
Dan (1952) have succeeded in isolating the mitotic apparatus from sea-
urchin eggs and found that to redissolve the isolated fibres it was necessary
to reduce them with thioglycollic acid (see p. 240) or to use strongly-
alkaline or strong solutions of urea. The mitotic cycle, involving an
appearance and disappearance of the fibrils, could thus be based on a
" reversible keratinization " effected perhaps by an oxidation-reduction
cycle between the protein-bound SH groups and cellular glutathione.
The spindle fibres might be formed by an aggregation of particles effected
by disulphide bonds.

KERATIN AND MOLECULAR BIOLOGY 31

It is evident that, unless a very narrow definition based on distribution
is insisted on, the keratins are far from being a homogenous body of
proteins. Neither by morphology, by molecular structure nor by overall
amino acid composition can they be classed together. This is in great
contrast to the collagens, for example. These proteins are widely spread
yet always characterized by the same high angle X-ray pattern (Fig. 8),
usually, but not invariably, by a low-angle X-ray pattern and electron
microscopic appearance derived from the presence of a master period of
~ 640 A, and by an amino acid composition in which one residue in three
is always glycine (a feature known now to be necessarily associated with
the molecular structure which gives rise to the X-ray pattern) and quantities
of the amino acid hydroxyproline are always present (see Fig. 2).

In fact it may not be far from the truth to say that keratinization is a
fate which could befall any of a number of kinds of protein, provided they
contain enough cysteine or are mixed with a cysteine-rich accessory
(see p. 248) in a biochemical milieu where the cysteine can be oxidized to
cystine. Thus it is only the process of keratinization for which a distinction
is evident and the conditions for its occurrence seems to be found mainly
in epithelia.

The Significance of the Variable Amino Acid Composition

Before proceeding it is desirable to give some further consideration to
the question of the variable composition of these proteins as revealed by
the analyses quoted. How many keratins are there ? And are they unique
substances with a constant composition or not ? From what will be said
below concerning the morphology of the various epidermal tissues, it
will be clear that any keratinized tissue is a mixture of numerous
chemical species among which a variety of keratin predominates in
amount. The analysis of such tissue is not therefore the analysis of a
single definite chemical substance and we can envisage the possibility,
apparently confirmed by actual analysis, that the composition may vary
from animal to animal, and from time to time, and from site to site even in a
single animal. The keratinized tissues are in this respect far less constant
in composition than other tissues of the body. Block and Boiling (1950)
have shown that the total amino acid composition of most tissues is
remarkably constant. This finding probably reflects the constancy both of
the cytoplasmic apparatus and of the specialized cell products in the tissue.
Of the tissues studied by Block, the keratins showed by far the widest
spread in their amino acid pattern.

These early analyses of Block and Boiling are not as complete as those
now available, nevertheless for comparative purposes they form a basis of
comparison, since the same procedures were followed in each case. Block
believes, however, that the hard keratins (eukeratins in his nomenclature)

KERATIN AND KERATINIZATION

show a significant constancy in the molecular ratios of histidine to
lysine to arginine (1:4:12), and thought that this pointed to some
underlying structural feature that was quite characteristic of the group.
More recent and more accurate analyses appear to cast doubt on the
integral nature of the ratios. Their approximate values are, however,
useful in distinguishing analytically between the hard keratins (1:4:12)
and the soft keratins (1:4:4) (see p. 65) and between keratins and other
insoluble and ill-characterized proteins.

rrrrrrm merino 64S wool

■HB MERINO 70S WOOL

5CORRIEDALE 56 S WOOL

ALA AG A P AMIDE CYS 6LU GLY HIS iSOLEU LEU LYS MET PHE PRO SER THR TRY TYR VAL

Fig. 15. Amino acid composition of one hydrolysate from each of Merino
64's, Merino 70's and Corriedale 56's quality wools. From Simmonds
(1955). The figure shows in graphical form the analyses by Simmonds
of three different wools in which a considerable variation in composition
is apparent.

When the best of the more recent analyses are compared (see Tables
1 and 2) it seems clear that: (a) very considerable differences exist between
some keratins and others and (b) smaller but definite differences exist
between different samples of similar materials, e.g. between different wools
(Fig. 15). Differences in cystine content are frequently reported even when
other acids were not determined. In the case of big differences, e.g.
between wool (hair), feather and epidermis, this can only mean that we
are dealing with rather different proteins although all may be keratinized.
The difference between feather and the other tissues is also revealed by its
different X-ray pattern (p. 16). For the a-type keratins we have accurate
figures only for several wools and hairs. These reveal closer similarities,
but even apart from cystine, no exact identity. Is there a unique ac-
keratin? None of the analyses suffices to prove such a protein exists.

The thorough study of the amino acid composition of the protein found
in the various morphologically distinct parts of feather: calamus, rachis,

KERATIN AND MOLECULAR BIOLOGY 33

barbs and medulla (see Fig. 30, p. 70) made by Schroeder and Kay and
their associates (1955) shows that the various parts differ in composition
and that species differences also exist. We may conclude that feather
keratin, like wool keratin, also has no precisely fixed composition. The
two keratins, wool and feather, are, however, quite distinct. In particular,
feather keratin reveals itself as peculiar in having 10 per cent of the amino
acid proline, a circumstance which must profoundly affect the configuration
of the polypeptides. Forty per cent of residues are the small residues
glycine, alanine and serine. These features of the feather polypeptides
have influenced the model proposed by Krimm and Schor (p. 208).

There may be technical reasons for small variations since these analyses
are difficult to perform and, further, they are usually carried out on whole
tissues, not on isolated purified proteins. As commonly practised the
materials are usually simply extracted successively with aqueous and lipid
solvents and, without further fractionation, are submitted to various
degradative processes as a preliminary to analysis. Since the keratinized
fibrous protein itself usually forms the predominate constituent, the
analytical findings will give a good idea of its composition. However,
it would be a mistake to regard the findings as representing precisely the
composition of " keratin " and to draw far-reaching inferences from their
exact values. Ideally the several components of the tissue ought to be
separated, purified and individually analysed as is obligatory in the case of
the soluble proteins. This is a counsel of too much perfection when dealing
with insoluble hardened products and little effort has been made to comply
with it in the case of the keratins. Nevertheless, since keratin remains as
an insoluble residue, when the tissues containing it are simply digested with
trypsin, it is regrettable that analysts have not attempted at least this
degree of purification before beginning an elaborate and lengthy
investigation.

Undoubtedly methods of solubilizing the proteins of a keratinized
tissue and of extracting pure individual proteins will be perfected sooner
or later. When these have been attained, is it possible to anticipate that
definite individual " keratins " will be distinguished ? It is difficult to
answer this affirmatively. Rather from the biological point of view a
variable composition might well be expected, for it is clear that the
demands made upon the epidermis and its derivatives are widely variable
and, if an adaptive response is to arise, a variation in the nature of its
composition might be expected. The interesting idea, discussed by
Tristram (1953) and more fully by Colvin et al. (1954) and Fox (1953),
that proteins exhibit a certain " spread " in their composition, may well
be applicable to structural proteins of the keratin-type even if it proves
untenable in other instances. The fact that the replacement of a single
amino acid residue may impair the function of a haemoglobin molecule

34 KERATIN AND KERATINIZATION

(Ingram, 1957) may seem a serious objection to the universal application
of this proposal. On the other hand residues not involved in forming the
precisely-patterned topography of the active patches of a protein molecule
may perhaps be exchanged with greater impunity (Tristram, 1953). The
probability that the function of a keratin does not demand the same
detailed specificity of structure, as for example an enzyme, may be relevant
here. Keratinized tissues have a mechanical function and a role as water
barriers. The properties required are a certain insolubility and toughness
combined with elasticity. For this a precise sequence of amino acids does
not seem immediately in demand ; numbers of polymeric networks having
an appropriate balance of hydrophobic and hydrophilic side chains might
be envisaged with similar properties. Further, a case could be made out
for supposing that the ability of the germinal cells to differentiate into
cells producing proteins of a variable composition could be the basis of
adaptation. Normally, we could suppose, the pattern of synthesis is
dominated by the site, e.g. producing hair keratin, horn keratin, etc., in
special sites, but, if the synthetic mechanism were also capable of con-
tinuous adjustment to (say) mechanical demands, the system would be
adaptive as in the epidermis it seems to be.

The biochemical mechanism of such an " adaptive synthesis " would
have to be sought in a selective pressure brought to bear on the population
of RNA molecules which emerges from the nucleus during the course of
synthesis (see p. 1 10). Since the mechanism by which external influences are
fed back through the cytoplasm to influence nuclear activity is one of those
phenomena most in need of experimental elucidation, we can carry this
speculation no further.

The fine structure of cells

Electron Microscopy and Cytology

X-ray diffraction methods are rarely applicable to cell inclusions or to
surface structures except when components can be isolated in a suitable
form. Chemical analyses and indirect physicochemical methods have
proved of more value but at the present time most of our knowledge is
coming from electron microscopy. In many cases, as may be appreciated
from Fig. 1, this form of microscopy appears the only approach to such
minute and irregular detail.

In the last few years following the perfection in the early 1950's of
methods of fixing, embedding and sectioning of biological material for
the electron microscope, cytology has undergone a veritable revolution.
Today we possess a wealth of morphological material covering most cell
types expressed largely in terms of the membranes, particles and filaments
whose images appear in electron micrographs. The work of recognizing

KERATIN AND MOLECULAR BIOLOGY 35

the objects visible in the light microscope in terms of these new structures
has also made satisfactory progress. Since this information is not yet
common currency, it is advisable at this point to name and describe
the commonly-occurring fine structural units in terms of which our
later accounts of cell structure will be given. These descriptions will be
brief and are intended to serve simply as morphological definitions.
Fuller accounts of the function and structure of the units will be given
later.

To a degree at the present time pure morphology has out-run knowledge
of the chemistry and function of cell constituents. It is, for example, not
always possible to state with certainty the chemical nature of the materials
giving images in electron microscopes. To appreciate the special nature of
this problem it is necessary to consider briefly the preparative procedures
of electron microscopy. Biological material intended for sectioning is
first of all fixed, i.e. subjected to a chemical treatment which kills the cells
and converts (more or less effectively) certain of their constituents into
derivatives of greater physical and chemical stability. The fixed specimen
is then dehydrated, which adds further to its stabilization, embedded in a
polymer and cut into suitably thin sections. Certain of the chemicals
previously employed by light microscopists, such as osmium tetroxide
(Os04) and formaldehyde, have been found useful as fixatives by electron
microscopists (Palade, 1952) and others (potassium permanganate) (Luft,
1956) have been introduced. The most commonly used fixative is the
buffered (pH 7-8) solution of osmium tetroxide introduced to electron
microscopy by Palade and it is with the results obtained using it that we
are mostly concerned with here. Osmium is an element of high atomic
number and there is no question that much of the contrast of osmium
fixed material arises from the electron scattering produced by osmium
atoms present in the fixed specimen either as lower oxides or as compounds
with the organic matrix. Unfortunately for the prospects of a super-
histochemistry the nature of the reactions of Os04 with tissue components
is both obscure and hitherto little studied. Bahr (1954), by treating in
vitro a series of pure organic compounds with osmium tetroxide,
has shown that certain chemical groupings are able to react with osmium
tetroxide to produce coloured compounds. It is probably correct to
assume that the reaction of osmium tetroxide with a complex molecule
such as a protein represents the sum of the reaction with its component
reactive groups. Figure 17 taken from Bahr summarizes the known
reactivity with osmium tetroxide. Since we are interested primarily in a
sulphur-containing protein, it should be noted that the most reactive
groups in proteins are the sulphur-containing amino acids.

A further factor of great importance in producing contrast is the com-
pactness of the organic substrate. This influences the amount of osmium

36 KERATIN AND KERATIN IZATION

per cubic centimetre in the fixed state which is the principle factor increas-
ing the contrast. Experiments in vitro have established that objects which
exhibit the greatest contrast in osmium fixed material are: phospholipid
(or phospholipid plus protein) membranes, compact masses of reactive
proteins and lipid inclusions containing unsaturated fats. Nucleic acids,
mucopolysaccharides and polysaccharides show little increase in contrast,

Fig. 17. In vitro reactions of osmium tetroxide (Bahr, 1954).
Reaction zvith Os04 (appearance of black lower oxides).

— SH peptides, proteins, enzymes.

= = fats, waxes, lecithin, cerebrosides, vitamins, certain hormones,

bile acids and other biological substances containing a basic
sterol structure.

->■ N tert.bases co-ordination, tryptophane.

— NH2 in terminal positions and not salt-linked.
— S — sulphide sulphur, cystine, methionine.
OH

— CHO *n terminal positions and on certain carbon chain lengths.

Certain heterocyclic compounds.

Aromatic compounds with at least two hydroxyl groups in suitable positions:

plant material, tanning substances.

No reaction with Os04

— COOH acid group.
— CH2 — CH2 — paraffin chain.
— CO— NH— peptide bond.
— COO-...NH3+ salt link.

— HS03 sulphonic acid group.

R — <^ y — OH monohydroxy- (tyrosine).

-OH halide-substituted (di-iodotyrosine).
I
Carbohydrates — Sugars and their polymers such as starch, glycogen, pectin,
amino sugars, heparin, hyaluronic acid, lignin.
Nucleic acids — Nucleotides, ribose sugars, sugar phosphates.
Various forms of high and low polymer DNA and RNA.

although compact forms of nucleic acid (see below) are dense because of
the phosphate groups they contain. Neutral formaldehyde may give a
more complete fixation of proteins; it does not, however, lead to any
increase in density (i.e. no staining effect) and is thus not so popular as a
fixative since the image is less photogenic. The presence of proteins in
formaldehyde fixed material may be demonstrated by staining with

KERATIN AND MOLECULAR BIOLOGY 37

phosphotungstic acid or lead salts. Useful discussions of these problems
have been given by Palade (1952), Sjostrand (1956), Bahr (1954) and
Baker (1945, 1955).

The reliability of electron microscopic findings is assessed by comparing
them when possible, with those of the light microscopy or X-ray
diffraction (see Fig. 1), by the internal evidence of the micrographs,
which may suggest deleterious changes and, in the final analysis, by their
contribution to the understanding of the problem of the function of the
tissue or organism.

We shall treat first, and in greater detail, the cell surface and its
specializations, since these are of greater importance for our subsequent
discussions ; the structures found in the cytoplasm follow next and finally
the nucleus.

The Cell Surface, its Specializations and Intercellular
Adhesion

The keratinized epidermal tissues are cellular in the sense that the
amount of intercellular material is very small; the cells surfaces are
effectively in contact and the whole formation owes its coherence to
intercellular adhesion. This is in contrast to the mesodermal tissues where
the extracellular material greatly predominates (Fig. 12), and coherence
is due in the vertebrate to the meshwork of collagen and other fibres laid
down by the cells. The cell membrane, the nature of intercellular adhesion
and its modifications in cellular tissues, adapted to withstand mechanical
and chemical shock, must therefore be given special consideration in
relation to the total phenomenon of keratinization.

The cell membrane. The simplest unicellular organisms possess a
boundary, which separates the intracellular domain from the environ-
ment, with special properties of permeability and mechanical strength
(Fig. 14). This cell membrane or plasma membrane may well have been
the first and most primitive organelle of the cell, since its existence without
such a definite boundary is difficult to admit. Its earliest function was
probably in essentials what it is today: by virtue of its selective perme-
ability it retains, in the neighbourhood of the cellular apparatus, a higher
concentration of certain molecules than exists in the surrounding medium.
Accordingly the permeability properties of the plasma membrane have
been much studied and this work has been summarized by Davson and
Danielli (1952) and Danielli (1942). The membrane proves to have pre-
dominantly a lipid character, i.e. it is most permeable to substances
soluble in non-aqueous solvents, as would be required of a membrane
whose principal function is to act as a barrier to substances dissolved in
water. This finding is supported by actual chemical analysis of membranes,
such as that of the erythrocyte, which can be obtained largely free from

38 KERATIN AND KERATINIZATION

other constituents. Measurements of the surface tension of free cells
show this to be very low (0-1 dyn/cm) which would suggest that the
protein is present as an adsorbed surface layer. The chemical analysis
(Davson and Danielli, 1952) also shows that proteins are present.

These experimental findings suggest the model (Harvey and Danielli,
1938; Danielli, 1942; and Stoeckenius, 1959) shown in Fig. 18 (a) which
pictures the membrane as a continuous lipid-like layer covered on each

\AAAA/vVAVW\ P
99999999999999999

55666666666555555
AA/VWWWWV P

FlG. 18. Two interpretations of the nature of the plasma membrane.
Both agree that the membrane consists of a lipid layer covered with
protein layers as indicated at (a) the Danielli-Harvey model; (b)
shows a somewhat more detailed interpretation in which the lipid is
represented as a bimolecular leaflet of lipid molecules covered with
monolayers of non-lipid protein.

face by a layer of protein. Its thickness cannot be determined precisely by
permeability or impedance measurements, but an order of less than 100 A
is indicated. Direct measurements, using a special device, the leptoscope,
made by Waugh (1954) confirm this order of magnitude.

Structures of these dimensions are within the range of the electron
microscope and, as mentioned above, our knowledge of membrane
structure has been greatly augmented recently by the use of this instru-
ment. Sections cut through the plasma membranes of a variety of fixed
cells have shown that its thickness is of the order of 70 A (± 10 A) and
that it frequently reveals a fine structure consisting of two dense outer
surfaces enclosing a less dense inner layer (Plate 4A) (Mercer 1959,
Robertson, 1959). This finding may be compatible with the structures of
Fig. 18 if we assume that the protein layers, P (and perhaps part of the

KERATIN AND MOLECULAR BIOLOGY

lipid), react with the fixatives (osmium tetroxide or potassium per-
manganate) more vigorously than the inner layer, L, and thus have
become " stained."

In some simple organisms the cell membrane is protected by a
secretion of slimes or mucins (Fig. 14(b)); in others this takes a more
definite form as an external pellicle (Fig. 14 (g)). In these secretions it is
perhaps permissible to trace the primitive forerunners of both the adhesive
intercellular cements and of the elaborate extracellular coats found in
some higher animals. According to Weiss (1933) cells freed from tissues
and cultivated in vitro surround themselves with a colloidal exudate.

Cilia, flagella, etc. Cilia and flagella are very similar in internal structure
although somewhat different in behaviour. They are long motile pro-
trusions of the membrane and in electron micrographs reveal a complicated

ft rt (TS fj

J U U 111 •;,...;•• L

m&tm

Id)

(e)

(f)

Fig. 19. Surface membrane specializations. See text.

internal structure (Fig. 19 (a)). (Bradfield, 1955; Fawcett and Porter,
1954; and Manton, 1952). In cross-section they are seen to consist
of the enveloping plasma membrane, nine pairs of peripheral
filaments, a central pair of filaments of a different character from the
nine and an amorphous ground substance. The peripheral filaments in
the protozoa are seen to end on a basal body within the cytoplasm which
consists of a small short cylinder apparently consisting of nine short
rodlets and which may control the organization of the bundle of filaments.
The centriole found in many cells of animals appears to consist of an
identical object and here too it may assist in organizing cytoplasmic
proteins (Porter, 1957; Grimstone, 1961).

40 KERATIN AND KERATINIZATION

The pattern of 9 + 1 pairs of filaments occurs with remarkable persist-
ance throughout the animal and vegetable kingdoms and modified versions
of such surface protusions are to be found in such unlikely sites as the
rods and cones of the vertebrate eye.

In electron micrographs of these organelles faint indications of a very
fine filamentous (diameter ~ 50 A) cytoplasmic component are common
(Fig. 19 (e) ). In ciliated epithelia (Fawcett, 1958), for example, and in
association with fine pseudopods in other situations, denser deposits
beneath the membranes are to be seen and may indicate a region of firmer
gelation of the cytoplasm which helps in maintaining the shape of the
surface protusions. Little is known of the detailed composition of such
filaments. Nevertheless, it is in such ill-defined fibrous proteins of the
cytoplasm that we may find primitive precursor of both keratin and the
contractile muscle proteins (Fig. 19 b and c).

Surface invaginations. These occur as long pleats of the plasma
membrane forming double surfaced membranes penetrating far into the
cell (Fig. 19 (c) ). They are particularly common in cells involving water
transport. Most elaborate examples are noted in the cells of the stomach
which secrete hydrochloric acid.

Temporary invaginations are commonly associated in free living cells
with the ingestion of solid material (phagocytosis) and liquids (pinocytosis).

Specializations of opposed surfaces. Cell contacts. In many tissues of
the multicellular organism the component cells are closely opposed and at
such surfaces of " contact " a variety of specializations has been observed.
Most of these seem to be associated with cell adhesion and in their totality
they form the devices by which a cellular tissue is held together.
Obviously (as mentioned above) in an external, protective, purely cellular
layer, such as the vertebrate epidermis, these devices become of great
importance; for while an extracellular cuticle may be effectively con-
tinuous and sufficiently strong to retain the enclosed cells, a cellular
tissue, however its cells may be hardened by intracellular deposits, will
be of little protective value unless the adhesion between the cells them-
selves is of an adequate strength. Certain experiments on the growth of
keratinizing cells in tissue culture (McLoughlin, 1959) suggest even that
the production of strong intercellular adhesion with the formation of a
stratified tissue is in itself an important factor in causing the cells to
keratinize.

In electron micrographs of fixed and sectioned tissues it is seen that
when two cells are in contact, their dense plasma membranes do not touch
or fuse. They remain separated by a space of about 120-200 A, which
appears light by contrast with the darker membranes (Fig. 20 and Plate
5), but which may sometimes be stained with electron-dense materials.
We may suppose that this space is occupied by a cellular exudate or

KERATIN AND MOLECULAR BIOLOGY

secretion, of low intrinsic density or poor affinity for the fixatives and
stains of current electron microscopy, which normally coats the surfaces
of the cells (Weiss, 1960).

A question of nomenclature arises here. The rather definite width of
the clear space between cells suggests that the surface coat itself has a
definite thickness and might well be included as an element of the plasma
membrane itself. This viewpoint is particularly cogent when specialized
developments of the cell membrane, such as those forming the myelin
sheath of nerve axons (Robertson, 1956 and 1957), are considered. How-
ever, for the purely morphological reason, that the double line bounding the

Fig. 20. The structure of the surface of contact between two cells A and
B (as seen electron microscopically in sections of tissues fixed in
osmium tetroxide or potassium permanganate). Mx and M2 are the two
plasma membranes which are each seen to consist of three sheets.
Between Mx and M2 is found an intercellular sheet (or cement) C.

cytoplasm is so obvious a feature and definable in all cells, we will refer
to it here as the plasma membrane. In this terminology the lighter space
is intercellular and, since the material occupying this space is ultimately
responsible for sticking the cells together, it may be referred to as an
intercellular cement. These definitions are illustrated in Fig. 20.

Desmosomes. In most epithelia even in the lowest of multicellular animals
opposed surfaces are studded with small dense areas called desmosomes
(Fig. 21). These are most common and conspicuous in stratified squamous
epithelia such as the vertebrate epidermis. They have been shown electron
microscopically (see Chapter 3) to consist of localized thickenings of the
membranes produced by the deposition of a layer of amorphous material
inside and outside the cell membrane. A similar layer is formed within
the opposing cell. In the epidermis these remain localized and appear to
resemble a pair of disks about 1 /n in diameter; in other situations, e.g.
the columnar epithelium of the intestine (Fig. 21 (a)), the desomosomes
may develop into bars or long bands running around the cell near the
free surface. There is evidence based on their behaviour when cells are

42 KERATIN AND KERATINIZATION

stretched (reviewed in Fawcett, 1958) that the desmosomes and terminal
bars are effective in holding the cells together and have been referred to as
attachment plaques. They are elaborately developed as the intercalated
disks in heart muscles where their role is again at least partly mechanical
in transmitting tension from cell to cell. Here the contractile fibrils enter
the desmosomal deposits when they reach the ends of the cells (Fig. 21 (e) ).

(d)

Fig. 21. Examples of desmosomes. In each case the desmosome is

shown as a pair of opposed thickenings of the cell membrane to which

may be attached small tufts of fibrils.

(a) The " terminal bar " type of desmosome is found in columnar or
cuboidal epithelia where it forms a long band running along near the free
edges of the cells. It could assist in preventing the separation of the
cells (Plate 5A).

(b) The simple plaque desmosome found in most squamous epithelia.
Extracellular sheets may be deposited between the plaques in some
situations (see page 92). This type of structure forms the traditional
" intercellular bridge."

(c) Desmosomes of type (b) may be regularly spread over the surfaces
of contact which may become more or less regularly corrugated to
produce a " tongue and groove " effect (Plates 6A and 12C).

(d) When a small duct passes between two cells desmosomes may
again form in the positions shown where they could function to contain
the duct.

(e) The very elaborate desmosomal development between cells in
heart muscle where it is recognized as an " intercalated disk." The
contractile muscle filaments end in the dense deposits at the cell
boundaries (Fawcett, 1958).

KERATIN AND MOLECULAR BIOLOGY 43

An important feature of the desmosome, which indeed justifies the
special attention given to it, is that it probably represents a point of fixed
(quasi-permanent) intercellular attachment. This permanence one deduces
from the observation that the structures in one half of the formation,
which must have taken some time to form, are mirrored closely by similar
structures in the opposing half. The two halves must thus have remained
opposed for some time. Elsewhere, over the shared surfaces of contact
where one observes interdigitating folds, it would seem that the surfaces,
although sticky, can slide laterally over each other, i.e. the intercellular
cement has the properties of a viscous liquid. One can readily visualize
that a range of viscosity is possible, depending on the degree of cross-
linking of the molecules between the surfaces. It is also possible that these
deposits may mark the sites where special forms of communication between
cells takes place, but definite evidence for this is wanting.

That fine intracellular fibrils, usually ending on desmosomes, occur in
many if not all cells has been long recognized (Schneider, 1902 and Schmidt,
1924) and recently Leblond and colleagues (Leblond et ai, 1960 and
Puchtler et al., 1958) by systematically applying a new staining technique
(successive treatment of fixed tissue with tannic acid, phosphomolybdic
acid and amido black: " TPA " staining) have demonstrated them in
many kinds of cells with exceptional clarity. The geometrical arrangement
of these fibrils and their attachment to studs (desmosomes) on the cell
membranes suggests a mechanical role in maintaining cell shape and
rigidity, i.e. they are literally " tonofibrils " (see p. 94). They stain as
basic proteins quite distinct from the extra-cellular collagen fibrils but
similar to the first formed fibrils in keratinizing systems. From a com-
parison of their location and density in several different cells, it is clear
that they occur in enhanced amounts in precisely the situations where
support is demanded. Epidermal cells and muscle cells show the most
marked development of TPA positive fibrils and in these cells their identity
with keratin and muscle fibrils respectively is obvious. It would seem
possible that all these fibrillar systems are composed of homologous fibrous
proteins and that chemical modifications have been developed to fit them
for special purposes. Keratinized fibrils, for example, are modified to
enhance their strength and stability; muscle fibres show an enhancement
of the latent contractility of the polypeptide chain.

Inter digitation of confronted membranes. In some epithelia the opposed
cell membranes, while remaining parallel, become greatly convoluted or
corrugated and a tongue and groove relation may develop, which has the
effect of greatly increasing the area of contact and presumably the adhesion.
Desmosomes usually form on such surfaces to add to the adhesion
(Fig. 21C and Plate 6A).

In keratinized tissues with their special requirements of strength it is to

44 KERATIN AND KERATINIZATION

be expected that these devices to increase intercellular adhesion will be
elaborately developed. In fact, as will be described later, the cell surfaces
become very convoluted, deeply interdigitated and heavily studded with
desmosomes. Further the intercellular cement becomes modified in its
solubility and chemical stability and forms dense intercellular sheets
between the surfaces.

The Differentiation of Surface Organelles

We referred above to those aspects of differentiation which arise from
the appearance within the cell of characteristic macromolecules. An
equally important feature is the appearance of the specialized surface
organelles which have just been described.

The factors bringing about differentiation are little understood and their
investigation is currently a major research preoccupation which will be
discussed later. Here we wish merely to refer to a certain antithesis which
exists between the specializations found on free surfaces and those on
bound surfaces. It seems adequately demonstrated experimentally, by the
failure of isolated cells to differentiate or to maintain differentiation
(Willmer, 1954) and by the appearance of differentiation when different
cell types are cultivated together (Moscona, 1952, 1956, 1957), that
differentiation results from the effect of one cell on another. For cells in
the interior of an organism the environment is either wholly cellular or
consists of solutions containing the products of other cells. On the other
hand, cells on the surface are uniquely situated in having at least one face
free from the immediate influence of other cells. Their environment on
this face resembles that of a free living cell and, in fact, the surface
differentia appearing here are identical with those found on free living
cells. These special responses to an external situation: the sprouting of
cilia, the secretion of mucins, the formation of a cuticle or intracellular
fibrils beneath the membrane (see Figs. 14 and 19), may be regarded as a
cell's free-surface " repertoire." They appear whenever the surface is
free and are repressed on surfaces in contact. Metazoan cells respond
to contact by adhering, which implies the secretion of the specialized
intercellular cements and the suppression of the free surface repertoire.
It may be argued that the property which above all others distinguishes
the multicellular organisms from the unicellular (see Fig. 14) is the
formation of intercellular adhesives, probably macromolecules among
which appeared to be mucopolysaccharides (p. 54) and proteins. Such
intercellular cements must have played their part in the evolution of the
metazoa and, in the life of each individual, the cells appear to pass from
an embryonic condition of poor adhesion to an adult in which, in many
organs, strong intercellular attachments are the rule (see Fig. 14). The
keratinized tissues carry this process to the extreme.

KERATIN AND MOLECULAR BIOLOGY

Cytoplasmic Structures
Particulates

(a) Mitochondria. These oval or elongate objects (diameter <~ \fx)
were recognized by light microscopists and characterized by various
staining reactions. Electron microscopically they appear as a small
vesicle enclosed by a double membrane, the inner membrane being cast
into characteristic folds. This internal structure may now be considered
as definitive (Fig. 22a). Mitochondria are extremely common in most
cells have been shown by biochemists (p. 115) to be the site of numerous
enzyme systems associated with cell metabolism.

Fig. 22. Intracellular membrane bounded organelles; (a) mitochondria;
(b) "particle covered" vesicles and membranes (see Plate 10A); (c)
" pleats " of surface (plasma) membrane (y-membrane pairs); (d) and
(e) parallel sheets of phospholipid-protein complexes; (e) " Golgi "
membranes.

(b) Smaller (0-5-0-2/x) membrane enclosed particles. A variety of small
bodies named microbodies, ultramitochondria, small vesicles, etc., scarcely
visible in the light microscope, are recognized as distinct on grounds of
size, absence of an internal system of membranes (of the mitochondrial
type) and the texture of their contents. Their nature is obscure and they
are certainly heterogeneous. Some may be virus inclusions, others sacs
of special enzymes (lysosomes), small secretory granules, etc. Some
vesicles contain phospholipids recognized by characteristic clusters of
concentric shells of membranes (see below).

46 KERATIN AND KERATINIZATION

(c) Small dense particles. Since the electron microscope can resolve
macromolecules of diameters as small as 20-30 A, the " ground sub-
stance " of cells often presents a fine particulate appearance. Conspicuous
among these fine particles on account of its size (100-200 A) and density
is a particle shown by Palade (1955) to contain ribonucleic acid (RNA).
These particles, ribosomes, may be free or associated with membranes to
form an important cell organelle, the basophilic reticulum described below.

Membrane systems

Membraneous systems are particularly well preserved and " stained "
by fixation in osmium solutions and in permanganate and are therefore
conspicuous in electron micrographs. Several organelles recognizable in
the light microscope are now known to be based on a skeleton of
membranes. The mitochondrion described above as a particulate may be
regarded as an example. Others less delimited in their extent are
described below.

Systems of particle covered membranes (ergastoplasm, basophilic reticulum).
The cytoplasm of many cells (see p. 108 et seq.) contains more-or-
less elaborately developed systems of membranes whose surfaces are
covered with small dense particles identical with those described above
(Fig. 22 (b))-. The membranes exhibit an intricate complexity of profiles
when examined in sections, which appear to be views of a geometrically
complex and intricately-interconnected reticulum of surfaces. This in effect
divides the cytoplasm into two parts: that inside the vesicular system
and that outside.

Systems of smooth surfaced membranes. The membranes here are
similar to those just described, but their surfaces are not associated with
particles. Some are simple " empty " vesicles of uncertain import. Two
rather more defined formations are usually described. The first, often
well developed in cells associated with the transfer of water, consists of
simple pairs of closely opposed parallel membranes which have been
traced back and shown to be deep pleats of the plasma membranes of the
cell (Fig. 22 (c)). The second is a variable yet always characteristic
stack of double membranes (flattened sacs) (Fig. 22 (d) and (e)) which
appears in all cells. It is usually assumed to correspond to the Golgi
complex of light microscopy, an organelle of uncertain function (p. 110)
(Grasse, 1956; Haguenau and Bernhard, 1955; Baker, 1955; Dalton and
Felix, 1956). Other rather more regular stacks of flattened sacs associated
with the nuclear membrane have also been described.

Concentric membranes, " whorls," myelinic forms. These formations may
vary from very perfectly-formed concentric shells of membranes with an
intermembrane spacing of the order of 40 A, through more open, con-
centric-shell formations, to stacks of parallel membranes closely related

KERATIN AND MOLECULAR BIOLOGY 47

to Golgi clusters. They may be imitated in vitro using preparations of
phospholipids extracted from cells (Stoeckenius, 1959; Mercer, 1960).
The myelin sheath of vertebrate nerve and the stacks of plates in the
retinal receptors are special cases of such structures (Sjostrand, 1956).

The similar morphology (see p. 37) of all these membranes and the
observations on the polymorphic possibilities of phospholipid membranes
in vitro, has led to the growing opinion that all intracellular biological
membranes have a common molecular basis in consisting of biomolecular
leaflets of phospholipids covered with layers of protein, as was proposed
many years ago for the plasma membrane itself and described above.
The different appearances and functions are thought to be determined
by the absorption on their surfaces of various macromolecules. Several
systems of nomenclature have already been proposed and, since some
confusion is possible, these will be outlined here. Sjostrand distinguishes
three types of membranes : (a) a-cytomembranes or membranes associated
with dense particles, (b) jS-cytomembranes, smooth surface membranes
found in the Golgi region and (c) y-cytomembranes, smooth flattened
invaginations of the cell membrane. Porter and Palade also recognize
these types, but prefer to regard all cytoplasmic membranes as portions of a
single membrane system which may become locally specialized for certain
functions, e.g. for protein synthesis by becoming associated with RNA
particles. The system of particle-studded (a-cytomembranes) is also iden-
tifiable with the basophilic ergastoplasm of Gamier (1897) (see Haguenau,
1958) and Bernhard and his associates (Bernhard et al., 1951 and 1954) are
inclined to refer to the entire membrane system as ergastoplasm. These
several proposals are set out in Fig. 23.

Some authors would go further in an attempt to unify the membrane
systems of cells under a single concept, by considering even the external
plasma membrane of the cell as part of this system. This view is implicit
in Ben Geren's views on the origin of the myelin sheath of nerve fibres
as an elaborate involution of the Schwann cell membrane and by
the work of Robertson (1959). In some cells, such as amoeba, the
formation of many vacuoles by invagination of the external membrane
is obvious, and all membranes retain the same fine structure (Mercer,
1959). A common molecular framework forming the basis of bio-
logical membranes is also envisaged by the Danielli and Harvey theory
of membrane structure already referred to on p. 38.

A certain lability of membrane structure is indicated by the profusion
of forms assumed by the cytoplasmic membranes in vivo and also by the
experimentally-produced breakdown of the reticulum and its re-formation
as smaller microsome vesicles (p. Ill) (Plate 10B). Bacterial membranes,
which are membranes of a widely- different origin, also possess a similar
property of reforming smaller vesicles on breaking up.

KERATIN AND KERATINIZATION

The nuclear membrane may be considered for the purposes of classifica-
tion among the cytoplasmic membranes. It consists of two surfaces,
whose distance apart is rather variable (500-1000 A), and which touch at
intervals to give the impression of a circular pore (Plate 4C). These pores
may form a regular pattern over the surfaces of some nuclei and are held
by some authors (Watson, 1954) to be genuine pathways permitting
nuclear-cytoplasmic interchanges at a macromolecular level.

Fig. 23. Nomenclature of cytoplasmic membranes.

Type

Sjostrand

Porter and
Palade

Bernard and
Haguenau

membranes associated
with dense particles
in Os04 fixed material

a-cytomembranes

basophilic

endoplasmic*

reticulum

ergastoplasm

smooth surfaced
membranes in Golgi
region

/3-cytomembranes

" smooth "

endoplasmic*

reticulum

Golgi
membranes

smooth surfaced
membranes linked to
cell surface

y-cytomembranes

—

phospholipid " liquid "

crystals

(inclusions)

- t

* Occasionally Porter and Palade prefer the less restricted term
" reticulum " to cover the entire system of membranes.

t For these the name S-cytomembranes has been proposed (Schulz
et al., 1958).

The nucleus

Within the volume delimited by the nuclear membrane there are no
membrane enclosed objects. One or more dense aggregates of particles
appear as the images of the basophilic nucleolus of light microscopy (see
p. 80). At the appropriate phases of cell division the nuclear membrane
dissolves and chromosomes may appear as denser aggregations of particulate
or finely-fibrous material.

Differentia

Under this name we gather a variety of cell products, which seem
distinct from the vital synthetic and respiratory machinery of the cell
and represent rather the end results of a specialized path of synthesis.
Their presence often gives the cell (and tissue) its characteristic appear-

KERATIN AND MOLECULAR BIOLOGY 49

ance and function. Examples are numerous secretory granules (enzymes
and hormones), pigment granules and fibrils of various sorts including
keratin.

It may seem significant that the number of basic structural elements
at the macromolecular level is so small. To some the situation would
seem to be a strong argument in favour of supposing a common descent
from an archetypal cell in which such devices as cilia and mitochondria
were already present. To others, and this may be the more austerely
biomolecular view, these resemblances in the primary organelles
indicate no more than that, given the limited molecular materials available
(proteins, phospholipids, polysaccharides, etc.), the number of structural
solutions to such problems as: enclosing whole cells, segregating intra-
cellular catalysts and a genetic apparatus, the provision of surface organs
of motility and a protective integument, is limited. The actual devices
found are effective and perhaps the only solutions to the problems.
While we can scarcely hope to discover how these structures came into
being, an experimental demonstration of their formation in vitro is
certainly conceivable as recent experiments on phospholipids show
(Stoeckenius, 1959; Mercer, 1960). In the same way the limited number
of fibre-types reflects the few possible ways of folding and packing poly-
peptide chains. These are limitations at the chemical level. Nevertheless
the tendency to produce certain molecular species and to use them in a
given situation is inherited and has phylogenetic significance (p. 22).

The phylogeny of keratinization

In the form found among existing land animals the epidermis is the
culmination of a long evolution on dry land, the steps of which can be
reconstructed from its histology among these animals (Romer, 1955 and
Young, 1950). Obviously this evolution commenced before the dry land
was invaded since many fishes deposit intracellular fibrils in their epidermal
cells and there are instances, as in the horny teeth of lampreys (Barrnett,
1953), of localized deposits of hardened protein among the lower verte-
brates. Also, from their almost universal distribution, there is every reason
to suppose that the group of cell responses associated with the surface
membrane, which were described in the previous section, can be traced
back to the earliest free-living cells. It is not without interest to try to
trace back the origin of a specialized epidermis to more remote beginnings,
to consider it as part of the more general problem of the evolution of
protective layers in organisms. What lends a particular interest to such
speculations is the possibility already mentioned that the events at the
surface of cells may have played decisive roles in initiating the various lines
of evolution.

We have pointed out that the beginnings of such supporting fibrillar

50 KERATIN AND KER ATINIZATION

deposits in the cytoplasm might be found in the cytoplasmic fibrils
beneath the plasma membranes in many generalized cells. Possibly too,
they share a common molecular ancestor with the contractile muscle
proteins and other intracellular fibrillar systems of the same basic molecular
type (see p. 22), but these views must remain conjectural in the present
state of our knowledge. Such cytoplasmic fibrous proteins are not strictly

WATER+ SALTS

SEMIPERMEABLE
MEMBRANE

IMPERMEABLE
CUTICLE

EXCRETORY
ORGAN

HYPOTONIC
URINE

Fig. 24. An osmoregulatory mechanism depending on the development

of a water impervious cuticle limiting input and output of water to

small areas. From Baldwin (1937) with permission.

keratin in the sense of a definition restricting this term to stabilized fibrils
although their role may have been protective. An accepted opinion
relates the early specializations of the outer layers to the problem of water
control rather than mechanical protection. It is supposed that the need
for such control arose in the first place with the colonization of fresh-
water habitats (Baldwin, 1937). Animals leaving the sea, where the osmotic
pressures inside and outside their cells were nearly equal, and entering
fresh water, where the salt concentration was less, would be faced with
stresses owing to the entry of water into their cells. They would therefore
be forced to evolve specialized organs to control the entry and egress of
water. At the same time by enclosing themselves in a waterproof coat and
thus limiting their water exchanges to a small area (Fig. 24), the work
required of the excretory organ would be lessened. At first, if we take
the condition in present-day fish as a clue, the epidermal layer was
supplemented by the secretion of a slimy mucilagenous layer ; this device
has, however, tended to lessen in importance as the degree of keratinization
of the cells themselves increased.

KERATIN AND MOLECULAR BIOLOGY 51

If we accept this hypothesis of the origin of the strengthened and water-
proofed integument, then we may regard its evolution in fresh water as a
fortunate pre-adaptation without which a subsequent colonization of the
more unfavourable environment, the dry land, could not have been
attempted. However this may be, it was on the land where the full
possibilities of keratinization were revealed and the astonishing variety of
modified skins, claws and scales were evolved from epidermal thickenings
culminating in the appearance of feathers, the distinguishing mark of
birds, and of hairs, equally characteristic of mammals. Since these
hardened parts may leave fossil imprints and since a sufficient variety of
animal types has survived until today, this later evolution is reasonably
well documented.

The greatest degree of keratinization was reached in reptiles; in birds
and mammals, with the elaboration of feathers and hairs, the thickness and
degree of keratinization of the epidermis itself lessens.

In most organisms the waterproofing properties of the toughened
framework formed by the structural macromolecules (proteins and/or
polysaccharides) are supplemented by the addition of lipid materials.

An increase in the degree of stabilization of the epidermal protein
itself appears to have occurred in higher vertebrates, if we may judge
from the skins of surviving types. An experimental measure of the degree
of stabilization may be obtained by observing the temperature at which
an oriented fibrous system contracts on heating due to the shortening of
its molecular chains by thermal agitation (see p. 255). Using this method
Rudall (1955) has shown more precisely that in the case of the newt,
Triturus, the stability of the skin was intermediate between that of a
film of myosin, an unstabilized muscle protein of the a-type, and human
leg stratum corneum. The temperature at which the X-ray pattern became
disoriented or was converted into a /J-type pattern was also deter-
mined by Rudall and again demonstrated an intermediate degree of
stabilization.

The amphibians, in this respect as in others, are " living fossils " and
preserve, in the changes which take place in their epidermal cells at
metamorphosis, a suggestion of the course of evolution of the keratinized
skin of the fully land-dwelling animals. The larval skin contains a variety
of cells, some ciliated, some secreting mucins and others containing
masses of fine fibrils, from which it is possible to infer that the factors
determining complete keratinization are not yet present. Histochemical
tests (Barrnett, 1953) based on the demonstration of cystine cross-linkages
as stabilizing elements (Hergersberg, 1957) show almost a complete
absence of true keratinization; after metamorphosis, when the animals
become capable of living out of water, the epidermis becomes keratinized.
The change is also provoked by thyroxin which causes premature meta-

52 KERATIN AND KERATINIZATION

morphosis. Some authors have reported free SH groups in the amphibian
stratum corneum.

Obviously the discovery of some method of hardening, reducing the
swelling and waterproofing the integument is a prerequisite for life in a
dry environment and, in this sense, the process of keratinization is rightly
described as one of the key biochemical discoveries on whose exploitation
the success of the vertebrates is based. In other land-dwelling forms
different macromolecules have been adapted to meet these same needs,
and, as mentioned above, the nature of the structural macromolecules found
in the integument places certain limitations on the evolution of a phylum.
A more fully-developed knowledge of molecular phylogeny may be able
to relate the mutations, which gave rise to the various molecular types to
the subsequent evolutionary development.

CHAPTER II

The Keratinized Tissues

Generalized histology of the vertebrate epidermis

Dermis and Epidermis

The integument of vertebrates consists of two quite distinct parts—
an epidermis and a dermis (Fig. 25) which together form a well-bonded
unit for the protection of the organism. The two parts are of very different
character and of different origin embryonically. The epidermis is entirely
cellular and its most characteristic products are retained within the cells
producing them; the dermis is primarily a fibrous meshwork in which
are distributed sparsely cells of mesenchymal (mesodermal) origin which
secrete a dense feltwork of fibrils of collagen and elastin. The inter-
fibrillar spaces contain, among other constituents, gelatinous muco-
polysaccharides. All these materials are found outside the cells in contrast
to the intracellular location of the fibrils of the epidermal tissues.

The dermis is continuous with the sub -cutaneous connective tissue
and with the other connective tissue of the body and has thus no definite
inner boundary such as the definite boundary separating it from the
epidermis. This surface separating the two is called the dermoepidermal
junction and has long been recognized, but its detailed structure and
special character were only revealed by electron microscopy. The con-
siderable early literature on the subject should be read in the light of these
newer findings which will be described in detail in the next chapter.
Membranes of a similar type seem to separate all the superficial epithelia
from the mesodermal tissues (BM in Plates 7, 9 and 23B).

Although conspicuous and predominant, collagen is not the only con-
stituent of the intercellular spaces of the dermis. By means of special
staining techniques numerous other substances can be demonstrated in
the light microscope and some have been extracted and partly characterized
chemically. It is customary to distinguish both formed, fibrillar elements
and an amorphous colloidal ground-substance. Collagen, except in a few
special situations, constitutes the bulk of the fibrous material. Less
conspicuous are elastin fibrils, as yet poorly-characterized but distinguished
from collagen by their microscopic appearance and their extensibility, and
other finer fibrils called reticulin.

54 KERATIN AND KERATINIZATION

The work of Meyer (1945, 1951 and 1957) in particular has led to the
recognition of the importance of a special class of polymers forming the
colloidal ground-substance of the intercellular spaces, the mucopoly-
saccharides. These are proteins linked to polysaccharides containing hexos-
amine or glucuronic acid. They are distinguished partly by the amount
and type of polysaccharide and by the presence or absence of sulphuric
acid, e.g. chondroitin sulphuric acid, contains sulphuric acid and hyaluronic
acid does not. Meyer (1957) distinguishes four or five mucopolysaccharides :
hyaluronic acid, chondroitin sulphates A B and C differing in their
optical rotary power [a]D (see p. 194). Heparin may also be classed
among them. The amount and types of mucopolysaccharides vary from
site to site which suggests some relation with the overlying tissue. Their
importance in tissue maintenance is sufficiently indicated by the dramatic
effects produced by cortisone, by the adrenocorticotrophic hormone
(ACTH) and by hyaluronidase. The two histochemical tests used to
distinguish mesenchymal elements: the periodic acid-Schiff test (PAS)
and metachromatic staining are often believed to stain mucopolysaccharides.

Metachromasy is the phenomenon of a change in colour of a dye on
becoming associated with a structure, an effect believed to be partly due
to the state of aggregation of the combined dye. The commonly used
dye, toluidine blue, is blue in dilute solutions and stains (basement
membranes) metachromatically to give a purple colour. The PAS test
is an application of the well-known test for aldehydes using a bleached
solution of basic fuchsin (Pearse, 1953). It may be used histochemically in
cases where a chemica ltreatment can cause the release of aldehydes from
polysaccharides or sugar-containing complexes. The Feulgen test (p. 80)
is based on the fact that mild acid hydrolysis liberates an aldehyde from
DNA which restores the colour of the dye. Other polysaccharide com-
plexes yield aldehydes after oxidization by periodic acid and can then be
stained by the Schiff reagent. Reticulin and basal membranes are
strongly PAS-positive. Formerly the basal membrane was on these
grounds said to contain a fine network of reticulin fibrils; probably it
contains an amorphous mucopolysaccharide as suggested by its appearance.
Leblond and his associates have cast doubt on the interpretation of these
tests by showing that extracts of pure components do not always give the
expected reactions.

The PAS-positive and metachromatically-staining substances are
invariably present in sites of rapid growth (p. 221) and seem therefore
associated with the process, but their role remains obscure. They are
strongly hydrophilic and could help to retain water and to form a viscous
gelatinous scaffolding in advance of more permanent formations. It also
seems likely that in a more condensed form they can function as com-
ponents of intercellular adhesives and the related basement membranes.

THE KERATINIZED TISSUES

Certain of the polysaccharide fibres from vertebrate skin are said to yield
an X-ray diffraction pattern like that of cellulose (p. 24).

The configuration of the boundary between the two layers is much
influenced by the presence of specialized appendages — hairs, scales, etc.,
formed by the epidermis. In the absence of these, it may be smooth and
run parallel to the external surface; when present they dip deeply into
the dermis. In mammals there is a marked development of small regularly-
placed dermal papillae and ridges which, being more prominent the

DESQUAMATED
CELL

<rr^>

> KERAT1N17ED
■) LAYERS

I <

•cs

I KERATINIZING
J ZONE

CELL MOVEMENT

DIFFERENTIATED
LAYERS

GERMINAL
LAWYER

EPIDERMIS

BASAL MEMBRANE
DERMIS

XXX

Fig. 25. A generalized, stratified, keratinized, squamous epithelium

resting on a basal membrane backed by a collagen-containing dermis.

In all keratinized tissues (except the thinnest skins) the several layers

shown on the right-hand side may be distinguished.

thicker the epidermis, are related probably to the nutrition of the superficial
cells. They have the effect of greatly increasing the area of contact
between the two formations and thus should facilitate the transfer of
nutrient from the vasculated dermis to the epidermis.

The Epidermis

Among the vertebrates the epidermis is always a stratified epithelium
consisting of a few to many layers of cells produced by the proliferation
of a single basal layer (Fig. 25). Epithelia, more-or-less regular and
compact layers of cells, are common tissues and usually cover the external
surfaces of an organism or its internal cavities. Their superficial location
appears to impose certain common histological features. The cells may

56 KERATIN AND KERATINIZATION

become progressively more specialized as they near the surface and may
show an internal polarization of structure. They are bounded below by
the basement membranes separating them from the connective tissues
which are vasculated. Being themselves avascular, nutrient materials
must reach them by passing from the vessels of the subjacent connective
tissue into the interstitial fluid and thence by diffusion across the basement
membrane. There may often be a constant loss of cells from the exposed
surface (squamous epithelia) which must be made good by the proliferative
activity of a germinal layer.

The keratinizing epithelium is normally squamous, its hardened
surface cells being shed either continuously or at intervals as a whole
in the form of a moult. In its simplest form it may appear to consist of a
single layer of living germinal cells and a thin cuticle of keratinized cells
as in the mouse. Usually more layers can be defined and the character of
the skin is influenced by the number of cells in each layer. An idealized
generalized, keratinizing epidermis might be said to consist of at least the
following layers: (a) a germinal layer of cells whose proliferation main-
tains the entire cell population; (b) a differentiating layer in which the
protein is synthesized and keratinized; and (c) the dead and hardened
layer (see Fig. 25). If the tissue is to be of constant average thickness, the
cells formed over a given time must equal those lost by exfoliation in the
same time. The thickness varies with the total number of cells in each
of the layers and structures with a variety of properties may be produced by
variations in the proportion of differentiating cells and hardened cells,
i.e. in the relative thicknesses of the softer still hydrated layers and the
tough, cornified and relatively drier cells. The ease of exfoliation deter-
mines the thickness of the homy layer and obviously a thick, hard layer
will result if the exfoliation is slowed down. Further by supposing
localized differences in rate of cell formation and loss, it is possible to
understand, in principle, the production of the various horny appendages.

It is sometimes said that keratinization is a degenerative phenomenon,
a consequence of poor nutrition, of desiccation or other deleterious
factors. That this is not true is shown very clearly by observations on cells
cultivated in vitro where conditions are under closer experimental control
(Fischer, 1924; Miszurski, 1937; Hardy, 1949; and Strangeways, 1931).
Skin cultivated in vitro readily undergoes keratinization with the pro-
duction of histologically-normall}' keratinized cells. The same sequence of
histochemical events as in vivo occurs and the product is also birefringent
(Litvac, 1939; Miszurski, 1937; and Hardy, 1949). Feathers have been
cultivated to a limited degree. Strangeways and Hardy (Strangeways,
1931; Hardy, 1949) also succeeded in growing hairs and noted even in
these abnormal conditions that histogenesis, up to a certain point, and
fibrillar orientation proceeded normally.

THE KERATINIZED TISSUES 57

Nevertheless keratinization in vitro does not take place in all types of
cell regardless of origin ; nor, as was shown very definitely by Miszurski
(1937), in epithelial cells themselves is it initiated or promoted by poor
nutritive conditions or low oxygen tension or lower temperature. The
phenomenon is properly to be regarded as the final stage of an intrinsic
differentiation of epithelial cells, as well adapted to the function of the
tissue as in, for example, the production of collagen by fibrocytes. This
is not to say, of course, that it cannot assume an abnormal, perhaps
degenerative form.

That the epithelial habit with the potentiality of keratinization is a
fundamental type of cell behaviour is shown by the fact that it is one of the
forms to which cells revert when cultured in vitro for some time. Willmer
(1954) describes three such cells forms: epitheliocyte, mechanocyte and
amoebocyte. When conditions are appropriate epitheliocytes adhere
laterally and grow to form flat unicellular sheets. The intercellular contacts
are of the types described above (p. 40); the cells may secrete muco-
polysaccharides (mucins) and may keratinize. Mechanocytes form open
meshworks, secrete collagen and a distinct type of mucopolysaccharide.
The mobile amoebocytes form no permanent associations with other cells.
Recent work (e.g. Puck, 1957) shows that the morphological appearances
may be deceptive but the biochemical differences are more persistent and
significant.

The epidermal family of cells

From the ectcderm of the embryo is developed not only the adult
epidermis but an entire family of cells which includes the keratinized
appendages and numerous glands. The " genealogical tree " of the
epidermal family is shown in Fig. 26. The entire population is produced
and maintained by the proliferation of the undifferentiated cells of the
germinal or Malpighian layer, which everywhere covers the outer surface
of the dermoepidermal membrane. These cells not only have the same
embryonic origin but, according to some authors (Montagna, 1956), retain
everywhere, even in the adult, the potentiality to differentiate into any of
the cell types found among their descendants and thus on the biochemical
level to produce any of a number of distinct chemical substances, such as
keratin (in several forms), sebum, mucin, etc.

The development of glands capable of secreting mucin or lipids is a
character of the epidermis as typical as and, phylogenetically speaking, of
earlier development than, its keratinizing potentialities. The glands may
take the form either of unicellular " glands," i.e. single cells discharging
their contents directly on to the surface, or of more elaborate multicellular
formations sunk into the dermis and communicating by a duct with the
surface.

KERATIN AND KERATINIZATION

The secreted mucins are viscous, shiny substances providing a protective,
slippery, extracellular sheath possibly related chemically to the more
permanent and condensed cuticles. They are ill-defined chemically,
containing a protein and a carbohydrate moiety, but the nature of the
association of the two is obscure (Meyer, 1957 and p. 54). The relation
between the mucoprotein and other types of protein which may be
synthesized by cells arising from the same germinal layer does not seem to

epg

ectoderm neural crest

melanocytes!

mesoderm

endoderm

internal

epithelia

(some keratin)

mucous
glands

sweat
glands

sebaceous
glands —
hairs*

horns

feathers*

adult
epidermis
digital

appendages —
claws, nails,
hoofs

dermis

Fig. 26. Genealogical tree of the epidermal family and related tissues.

* It is not implied that all types of appendage (including hair and
feathers) appear on a single skin.

t The pigment-forming melanocytes migrate into the dermis and
epidermis where, after attaching to the dermo-epidermal basement mem-
brane, they may pigment the growing epidermal-type cells (p. 276).

I This symbol is meant to indicate the special anatomical union
between dermis and epidermis.

have been explored. Owing to the extremely elongated nature of their
molecules and their high negative charge, they raise the viscosity of the
secretions and in this way lubricate the surfaces and protect them against
mechanical, chemical and perhaps bacteriological injury. The lipids
comprise a multitude of compounds having in common a solubility in
non-aqueous solvents. They may exhibit species specificity (Hilditch,
1949). The phospholipids are an essential constituent of most biological
membranes (p. 37 et seq.).

Mucin secretion is common in aqueous forms or on moist internal
surfaces of land forms; lipid secretion is found more among the dry

THE KERATINIZED TISSUES 59

skins of land forms where it serves to lubricate and improve the water-
proofing of the skin. It is difficult to suggest a reason for the extraordinary
variety of compounds found in these secretions.

Birds and reptiles have fewer cutaneous glands than mammals. The
heavily-keratinized epidermis of reptiles does not favour their development
nor for perhaps the opposite reason does the thin skin of birds. Birds
usually possess a uropygial gland opening in front of the tail, the secretion
of which, spread over the feathers during preening, helps to waterproof
the layer of feathers. Vitamin D is produced from its secretion by the
action of sunlight and plays a part in nutrition (Hotta, 1928 and 1929)
which seems to demonstrate a special reason for the presence of some lipid
molecules.

The mammalian skin is rich in number and has more varied types of
glands. The types generally present are : (1) mammary glands which give
the phylum its name, (2) sebaceous glands, and (3) sweat glands. The
sebaceous glands are usually associated with hair follicles and produce an
oily secretion which softens the skin and helps to lubricate the hair.
Mammary glands and sweat glands are histologically similar. The small
sweat glands or eccrine glands derive directly from the embryonic
epithelium, the apocrine (partly sebaceous, larger sweat glands) and the
sebaceous glands indirectly via the follicular epithelium of the outer root
sheath (p. 96). The latter glands are found usually in association with
hairs; the eccrine (sweat) on hair-free surfaces where, as in man, they
may play an important role in temperature control. Sebaceous and
apocrine glands possibly are related to similar glands in the hairless skins
of earlier vertebrates. The very active cutaneous mucinogenic glands of
cyclostomes are said to be partly holocrine and may be remotely related
(Rothman, 1954). Some cutaneous holocrine glands of reptiles, undoubtedly
derived from similar phylogenetic ancestors, are sac-like invaginations of
the epidermis producing fatty materials. Certain mandibular cloacal
glands of alligators are reported (quoted by Rothman) to produce lipids
and keratin simultaneously. The cell peripheries keratinize and the centre
produces lipids ; the whole is ultimately shed. This is a demonstration of
bifunctionality of epidermal cells. It would seem that epidermal cells can
still produce lipids, in addition to the structural phospholipid of cell
membranes (Table 4, p. 486, Rothman, 1954) indicating that the
potentiality is still present (see next section).

There is a sense in which the entire epidermal system may be regarded,
as Montagna has put it, as an immense holocrine gland. Most of the
" secretion," i.e. the keratinized material, is shed and lost, but certain
other constituents may be absorbed either by their producer, its young or
associates, and serve further physiological ends. This is most obviously
so among the mammals whose milk glands are elaborated sweat glands,

60 KERATIN AND KERATINIZATION

and whose persistent habits of licking and grooming must introduce many
substances of epidermal origin into the alimentary canal. Vitamin I)
production is an epidermal function. Among birds the preen gland has
been shown to be essential for the well-being of its possessors. An extra-
ordinary example is provided by the aquarium fish (Symphysodon discus)
whose young are nourished by the mucous secretion of the parents which
covers large areas of their bodies (Hildemann, 1959). Probably much
remains to be discovered in this field.

An important anatomical aspect of mammalian skin is the close
association of glands and hair follicles. Except in rodents, the central
primary follicles and most, but not all secondaries are associated with a
sebaceous gland which usually opens into its lumen. This rather constant
association of hair follicles and sebaceous glands suggests that the natural
functional unit of the mammalian skin is the follicle group and its glands,
the " pilosebaceous unit " (Montagna, 1956) (Figs. 34 and 35), which
together produce keratinized hairs and the means to lubricate and condition
them. In the dry scaly skins of reptiles or the glandless (with the exception
of the single preen gland near the tail) skin of birds, it is hard to find an
analogue of such a unit.

The Differentiation of Epidermal Cells

It is a problem of wide interest to determine what are the factors which
are responsible for the appearance of the variety of cell-types which we
have included in the " epidermal family." Information has been sought
principally on the factors determining early embryonic differentiation and
on those maintaining the stability of the various cell-types found at
different sites on the adult.

The epithelial character of the external cells of an embryo appears very
early in life — in effect already in the blastula — and we have mentioned the
hypothesis that this is due to the appearance of intercellular adhesion
which continues thereafter to play an important morphogenetic role.
A further discussion of the role of cellular adhesion in controlling cell
differentiation among the persistently embryonic cells of the germinal
layer of the epidermis will be given in Chapter 3. Granting that the
primary step in the differentiation of epidermal cells is the production of
the epithelial habit by intercellular adhesion, there is direct experimental
evidence provided by grafting to suggest that in the next phase the factors
responsible for localized specializations arise in the underlying mesoderm.
For example, using the embryonic chick, by grafting mesoderm from a
presumptive foot bud beneath wing ectoderm, one can cause the formation
of a claw instead of a wing feather (Cairns and Saunders, 1954). The
epidermal cells at this stage may be described as being sandwiched
between two environments: the external relatively-free space and the

THE KERATINIZED TISSUES 61

underlying mesodermal domain. In the first place the external situation
imposes a class of differentiations and secondly the mesodermal organiza-
tion further limits differentiation and gives rise to site-characteristic
developments. As these develop and call into being an appropriate dermal
organization to support them, consisting in part of fibrous collagenous
depositions and a blood supply, the situation is reversed. The epidermal
tissue becomes dominant and grafting now shows that, when sufficient
epidermal tissue is transferred, the site characteristics (skin, claws,
feathers, etc.) are now preserved (Cairns and Saunders, 1954).

xxxyxxx

000 era ipBM

1 o a m «*

Fig. 27. Diagrammatic representation of the interreaction between
epidermis and dermis in successive stages of the establishment of a
differentiated epidermis (r.h.s.). O, outside environment, I, internal
environment, BM, basal membrane. The arrows indicate the direction
of dominant influence and the shading differentiated cells.

From the numerous experiments (Zwilling, 1955; Waddington, rev.
1956; and McLoughlin, 1959) which show that the underlying mesen-
chymal tissue induces and maintains the different epidermal differentiations,
there is a suggestion that the basal membrane itself could be the important
factor. It appears to differ in thickness from site to site (200-600 A) and
there are differences in the types of mucopolysaccharides present. It is
easy to picture such a continuous layer of colloids of this type, with their
fixed network of charged sites, functioning as ion exchange resins (Meyer,
1956) and exerting a selective effect on the transfer of signal molecules
from the blood to the epithelial cell population. The idea is, however,
very insufficiently explored experimentally as yet.

If, at some risk, one attempts to summarize the findings in a field as
yet imperfectly explored experimentally and in rapid development, it
seems possible to distinguish several successive stages in the establishment
of the skin in which dominant control in the dermoepidermal partnership
swings successively from one member to the other as suggested by the
arrows in Fig. 27. In the earlier phase the superficial cells, responding to
their exposed position (see also p. 90), begin to stick together and thus

62 KERATIN AND KERATINIZATION

bring about the epithelial pattern. Their, as yet unstabilized, free, external
surfaces produce a selection of responses from what we have termed the
" cell's surface repertoire." The establishment of a definite surface layer
encloses the other cells in a different environment which diverts them
toward synthesizing different products (mesenchymal substances), which
in their turn react with the inner surfaces of the basal layer cells to form
the basal membrane and to induce local variations in the epidermal layer.
By their subsequent development these epidermal variations make return
demands for food and support on the underlying layers leading to the
building up of a dermal organization of fibrillar scaffolding and supply
vessels.

The molecular basis of these events is ill understood and their discussion
is often marred by vague concepts which in effect conceal our ignorance.
Further discussion of the problem will be found in Chapter 3.

There is much evidence, emphasized by Montagna, Chase and colleagues
(Montagna, 1956) and by Billingham (1958), to show that the determination
is not irrevocable and that the cells of the germinal layer itself remain
effectively multipotential. In many animals, hairs may normally become
differentiated from basal layer cells throughout life and Billingham cites
the extraordinary case of the hairy velvet covering the growing horns
of deer which is in its entirety reformed annually. Further, after losses due
to many injuries (wounds, burns, X-radiation) many epidermal elements
regenerate from the remaining basal layer cells or from cells of the outer
root sheath of the hair follicle (Montagna, 1956; Billingham, 1958).

Other evidence of persistent multipotency is provided by the epithelia
of many internal surfaces which may exhibit a cyclic metaplasia under
hormonal control with a well defined physiological function. These may
be of endodermal origin, but since they may be capable of keratin
formation, they are relevant here. The best-known example is that of
the vaginal epithelium which oscillates between mucin and keratin
production, cells of a contrasted cytology being produced successively
by the same basal layer (Nilsson, 1959) (Fig. 59). In other situations
keratin production oscillates with glycogen (Hinglais-Guillard, 1959).

Thus it would still seem that the course of differentiation followed by a
cell leaving the basal layer is determined by effects emanating from
neighbouring cells and from the underlying dermis. That in grafts cells
may retain the characteristics of their site of origin could be due to the
fact that large numbers of cells are transferred in a graft and, in effect,
carry with them their original environment. When a denuded area is
merely " seeded " with small groups of cells, the new growths are said to
be typical of the new site rather than the old (Montagna, 1956).

In many malignant tumours arising from epithelial surfaces (carcinomas)
the normal controls maintaining a stable differentiation seem relaxed and

THE KERATINIZED TISSUES 63

latent potentialities of the cells may find expression. Small nests of cells
with a well-defined, but aberrant histology, e.g. ciliated borders, mucin-
secreting surfaces, etc., may often be observed. That the potentiality of
producing both keratin and mucin can exist even in the same cell is
indicated by some observations of Gliicksmann and Cherry (1956) on
mixed carcinomas.

It would seem that the several varieties of epidermal cells are examples
of what Weiss (1950) has preferred to term cell modulations which are at
first reversible and which require for their maintenance the piesence of
other elements of the cellular community. Modulations are to be con-
trasted with the perhaps irreversible differentiations which accompany
embryogenesis and which divide the total cell community into several
major families of common descent. Support for such a view is given by
direct experimental evidence of cell metaplasia produced by relatively
simple chemicals. The work of Fell and Mellanby (1953) (Fell, 1957) and
their associates has established that vitamin A disposes the epidermis
towards mucin formation. Seven-day old chick embryo ecoderm cultivated
in vitro in a normal culture medium undergoes precocious keratinization,
the two-layered epithelium being replaced by a stratified layer. When
vitamin A is added to the medium (2000-3000 i.u. per 100 ml) keratiniza-
tion is prevented and a mucous secreting, often ciliated, epithelium appears.
The change is not stable, for when the layered epithelium was trans-
ferred back to a normal medium (i.e. lacking vitamin A) a typical mucous
membrane containing ciliated cells and mucous cells at first appeared, but
after a time this was replaced by a squamous keratinizing epithelium
forming beneath it.

Lasnitzki (1956) showed that the effect on mammalian skin (human
embryo) was essentially the same. Embryonic epidermis (3-4 months
foetus) in normal medium formed a typical squamous keratinizing
epithelium including a keratohyalin layer. In a medium containing
vitamin A several layers of large cuboidal cells appeared which contained
mucin-like materials. Older skin is less responsive but vitamin A sup-
pressed keratinization.

The vitamin A induced metaplasia was correlated with changes in the
uptake of sulphur detected by using radio-active sulphate. In explants of
skin treated with S35 (as sulphate), mucous secreting material was intensely
active; the keratinizing layers much less so. On the other hand, the
uptake of radioactive cystine was greater in the keratinizing epidermis
(Fell et al, 1954 and 1956. See also p. 264).

Weiss and James (1955) found that a brief exposure to vitamin A in
higher concentration produced the same effect as the continuous administra-
tion of lower doses used by Fell and co-workers, and concluded from this
that vitamin A acted as an inductive agent which switched the development

64 KERATIN AND KERATINIZATION

of the cells along an alternate pathway. A demonstration of induction by
" a crucial event of relatively short duration " is of some theoretical
importance, but Lasnitzki and Greenberg (Lasnitzki, 1958) have cast doubt
on the conditions of the experiment by demonstrating the persistence
of vitamin A in cultures treated as were those of Weiss and James.
They conclude that the action of vitamin A is due to its continuous
presence.

The effect of vitamin A is most apparent on the germinal layer cells as
might be expected, since these are " uncommitted," but cells in the
process of keratinization can still be deflected in their course by the
vitamin. The mucin forming cells once formed cannot, however, revert
when returned to a normal medium but are shed.

The variety of differentiations of the epidermis, its simplicity and its
experimental accessibility assure that in the future it will continue to
play a part in the investigation of the general problems of differentiation.

Hard and Soft Keratins

It is customary to distinguish between " soft " keratins (the epidermis
itself) and the " hard " keratins, hair, feather, horn, etc. The classification
was put on a well-defined basis by Giroud, Bulliard and Leblond (1934).
Primarily the distinction is based on the immediate sensation of hardness
or softness, and the fact that soft keratins (epidermis) desquamate while
the hard keratins (hair, nails, etc.) persist. These properties were shown
to be linked with other differences appearing in the course of keratinization
and with the chemical composition of the final product. These are
tabulated in Table 5.

Some of these distinctions, which seem obvious enough at first sight,
become less obvious on closer analysis. The difference between the two
types is, in fact, only relative and, if pathological material is admitted, a
continuous spectrum of tissues between hard and soft exists. Moreover,
the same germinal matrix can be made to produce a graded series of
tissues of various textures, as is shown by the development of callosities,
warts and corns, the thickened skin of Ichthyosis vulgaris and by some
effects following radiation (Chase, 1954). However, the classification is
useful as representing two extremes of the synthetic potentialities of
epidermal cells with obvious adaptive potentialities. Typically where the
site and function demand a squamous tissue, soft keratin develops;
where a persistent growth is required, we find hard keratins. We know as
yet, little of the underlying causes determining the type of keratin produced
at any site.

In general terms a structure is hard and coherent when its units are
hard and they do not fall apart. Translated into histological terms, this
means that the factors underlying the differences between hard and soft

The keratinized tissues

tissues are to be sought in : (a) the hardness and coherence of the inter-
cellular contents, and (b) the intercellular adhesion. The properties listed
in Table 5 give grounds for saying that in hard keratins the fibrous intra-
cellular protein is more plentiful, harder and more completely fused into

Table 5. Properties of " Hard " and " Soft " Keratins

Soft keratin

Hard keratin

soft and pliable*

tough and hard*

desquamating

permanent, non-
desquamating

in course of development
cells pass through a kerato-
hyalin layer (p. 95)

no keratohyalin phase

higher lipid content!

low lipid contentf

lower sulphur content
«3%)

higher sulphur content
(> 3%) and stronger thiol
reaction in course of
hardening (p. 217)

lower thermal stability

highe thermal stability

ratio of basic amino acids
histidine lysine and arginine
= 1:4:4 (31)J

ratio basic amino acids
1:4::12 +

less perfect ordering

better oriented

* This distinction is more apparent than real and is partly due to the
more massive character of the hard keratins.

t The lipids of the softer keratins may act as " plasticizer." When
extracted the materials become very tough and hard.

I This difference in the ratio of the basic amino acids was deduced
by Block on the basis of a large number of analyses. Block regards
the basic amino acids as forming a characteristic structural element in
proteins and he distinguishes sharply between the two classes of keratin
prefering to call the soft keratins " pseudo-keratins." Recent analysis
seems to show that the ratios are by no means exact integers.

continuous masses and that the adhesion between the cells is more complete
and more persistent. In the soft keratins the fibrils may be less completely
or less strongly fused. The fibrillar content of soft keratins partly derives
from keratohyalin (see p. 95); this substance appears deficient in the
covalent cross-linking responsible for cohesion (p. 234) and thus may more
readily break up. The thickness of the hardened layers (stratum lucidum

66 KERATIN AND KERATINIZATION

plus stratum corneum) depends on the coherence of the total formation
and such abnormally-thickened areas as corns show a thick clear layer —
the most thoroughly-bonded zone. Electron micrographs of desquamating
cells of human skin show a simultaneous separation between both cell
membranes and between the intercellular fibrillar contents in the body of
the cell. The spontaneous exfoliation could thus be due to a failure in
both components.

It is evident from Table 5 that the sequence of events in the formation
of a soft keratin differs somewhat from that in the formation of a hard.
It will be convenient to consider this problem later in Chapter 6, pp.
210-282.

The specialized appendages

Localized Epidermal Thickenings

The epidermis and the relative thickness of its various strata show
some characteristic local variations. In thin skins the intermediate layers
(Fig. 25) {stratum granulosum and stratum lucidum) may be absent,
the transition between the germinal and horny layer being quite abrupt.
This condition may have some significance in showing that a granular
phase (keratohyalin intermediary) may not be necessary for cornification
(p. 94); nevertheless such skins can often be provoked by appropriate
stimuli to assume a multilayered appearance and it is possible that, during
the actual growth phase, granules are formed.

The thickened areas of the epidermis have a functional purpose, e.g.
the horny pads of the digits, and become more thick with use. These
thickened areas are genetically determined and display thickening in the
embryo before being stimulated externally. What is also inherited is the
tendency of the cells in these areas to respond to friction or pressure by a
further proliferation, thus leading to an individually-adapted response.

Scales

Scales are specialized epidermal thickenings with a characteristically
patterned appearance which are strongly developed among reptiles where
they form a horny exoskeleton sometimes of considerable thickness
(Fig. 28). They are never separate, as in fishes, being simply localized
thickenings of an otherwise continuous epidermal layer. The scales of fish
are in fact quite distinct structures of dermal origin (p. 75) and are not
homologous to epidermal scales. Scales are also found on the legs of birds,
revealing their reptilian affinities, and among a few mammals such as
rodents, the scaly tails of rats being familiar. The well-developed covering
of the scaly ant-eater is said to be a secondary development.

Beaks and bills are horny developments of the jaw margins, often

THE KERATINIZED TISSUES

associated with a loss or reduction in teeth, and are typical of birds, turtles
and even some mammals (platypus).

Embryonically the scale, like the feather, appears as a small out-
pocketing of the epidermis containing a dermal papilla whose formation,
as a denser gathering of cells, precedes the actual proliferation of epidermal
cells and is thought to induce and to control the epidermal changes. The
flat upper surface of the outgrowth gives rise to the hardened scale (Fig. 28).
The lower surface which may be more or less overlapped by the upper
scaly surface and constitutes an inter-scale region, usually consists of a
softer more normal epidermis, and imparts some flexibility to the entire

.S (3

■I ec

Fig. 28. The reptilian scale structure. The scales S are thickened
epidermis and are not separate. The scaly layer S yields the /?-type X-ray
pattern and the interscale region I an a-pattern. m is the germinal matrix.

integument. As mentioned above, the keratin of reptilian scales is of the
)8-type which is rather inextensible. Rudall (1949) has found, however,
that in the softer, flexible inter-scale region an a-type keratin tends to
predominate. The production of two molecular types of keratin from
neighbouring cells arising from the same germinal epithelium poses some
interesting questions of differentiation (p. 104).

A continuous scaly epidermis may be so intensely hardened that it
cannot be shed by sloughing or simply worn off, but must be loosened
periodically and cast off as a unit. This is effected by a temporary cessation
of growth followed by the reformation of an entirely new horny layer
beneath the old.

Horns

The hollow horns of cattle, goats, sheep, etc., are horny sheaths covering
a bony core (Fig. 29 (a), (b)). They are not shed, but as they are worn
away they are renewed by the proliferation of a germinal layer. Such horns
do not branch although, owing to different rates of proliferation of the
germinal layers from one side of the horn to the other, they may grow in
graceful curves and spirals (Thompson, 1942).

The antlers of deer, etc., are not strictly speaking horns, being bony
growths forming beneath a covering of hairy skin, the velvet, which dies
and is rubbed off leaving the naked bony antler. Pronghorns of certain
antelopes are permanent bony antlers extended by a thimble-like sheath
of true horn, which in this case is shed periodically like a scale and renewed
without loss of the bony core (Fig. 29a).

KERATIN AND KERATINIZATION

(a)

(b)

(e)

(f)

Fig.

29. The various epidermal appendages containing hard keratin
shown black and softer varieties shown stippled.

(a) and (b) are two kinds of horns : (a) is the pronghorn consisting of a
" thimble " of true horn capping a " bony horn " covered by hairy
skin. The cap is shed annually; (b) is the true horn covered entirely
with horn keratin which is not shed. The antlers of deer are not true
keratinous horns but consist of bony growths at first covered by a hairy
skin, (c) The principal parts of a hoof which consists of an outer covering
of hard keratin (the unguis) and an inner subunguis of softer keratin.
This combination of a harder and softer keratin recurs in claws
(d, e and f). Claws of a carnivore, a bird and a rat. The wearing away
of the softer subunguis helps to maintain the sharp cutting edge of the
carnivore claw (Le Gros Clark, 1936). (g) The human nail in which
only the hard keratin layer remains.

THE KERATINIZED TISSUES

The Digital Tips: Claws, Nails, Hoofs

The fingers and toes of all vertebrates above the amphibians are re-
inforced by horny appendages which are adapted to the way of life of their
owners. The homology of all these structures will be apparent from
Fig. 29. Claws are structurally similar wherever found and consist of two
unequally-developed scale-like surfaces which meet over the end of the
digit (Fig. 29 (d), (e) and (f)). The dorsal surfaces are formed of a tougher
and more compact keratin than the lower or sole, and this contrivance
ensures that where the two meet, a sharp projecting cutting edge will be
formed. The orientation of the component cells may be different in the
two layers (Fig. 29 (f) ). The nails of primates are developed from claws
by flattening and losing the underlayer leaving only the compact layer
(Clark, 1936 and Horstmann, 1955). The sole is perhaps represented
by a small area beneath the projecting nail (Fig. 29 (g)). In hoofs, the sole
although softer forms a more important part of the weight-bearing surface,
and is encased in a sheath of more compact and harder horn (Fig. 29 (c)).
This use of keratins of different degrees of toughness to effect special
functional properties is an important aspect of keratinization which will
be returned to in Chapter 6.

Feathers

The structure of a typical feather is shown in Fig. 30. The simple
down-feather or plumule consists of a cylindrical quill opening into a tuft
of barbs and barbules; the filoplume is a fine hair-like feather; contour
or flattened flight feathers are more complex, consisting of a quill or
calamus and a shaft or rhachis which bears the barbs and barbules.

Feathers, the uniquely-distinguishing mark of birds, are believed on
embryological and paleontological grounds, to be homologous with
reptilian scales. As in the case of scales the first indication of the site of
a presumptive feather is in the gathering of dermal cells beneath the
epidermis which then projects to form a papilla containing dermal elements.
At this stage, feather and scale " germs " are much alike. Later the
whole formation sinks into the skin to form the follicle. Regarded in its
simplest ideal form, a feather is a hollow type of cornified epidermis
growing from a ring of germinal cells at the bottom of the follicle (Fig. 45,
Chapter 3). The development of a plumule (down feather) which consists
of a short cylindrical quill opening into a circle of soft barbs and smaller
barbules can be understood from Figs. 46 and 47, pp. 103-105. The
germinal layers at the base of the papilla first form a number of longitudinal
columns (seen in cross-section in Fig. 46) ; these separate and keratinize
each to form a barb. The basal part of the growing feather does not
separate into columns, remaining as a continuous cylinder to form the

KERATIN AND KERATINIZATION

quill. When the thin skin forming the sheath and covering the feather
during growth bursts, the feather opens.

The larger and more elaborate contour feathers in which the barbs
project from the sides of a shaft are formed in a more complicated way.
Up to a point development is similar to that of a plumule ; then the mid-
dorsal region of the germinal collar begins to proliferate more rapidly to

Fig. 30. The parts of a typical flight (contour) feather. CA calamus

a simple hollow cylinder, AR the rhachis, and B, barbs. The barbules

(not shown) lock the barbs together. X-ray patterns are usually obtained

from the calamus or the rhachis.

form the rhachis. As before, the calamus (or quill) grows as a continuous
cylinder and fails to split into barbs. This question is returned to in
Chapter 3.

Feathers are normally shed in an annual moult and replaced. The
first down or nestling feathers are replaced by juvenile feathers which
resemble true contour feathers growing from the same follicle. The same
follicle is thus capable of growing feathers of different kinds.

THE KERATINIZED TISSUES

Hairs

Most of our information concerning the properties of keratin comes from
a study of hair and wool which will thus, perforce, form the subject
matter of much of the discussion to follow. The reasons for this are
partly economic, funds for research being derived from the textile and
cosmetic industries, and partly experimental convenience, hairs being
easy to obtain in quantity, easy to purify, and their thread-like form
lends itself to many physical investigations.

Hairs are as characteristic of mammals as feathers are of birds. When
fully grown they consist of a tapering tip, a shaft and a base normally
embedded in the skin; in cross-section, a cuticle, cortex and medulla
may be distinguished (Fig. 31). A great deal of variety can be produced

Fig. 31. The parts of a hair seen in cross-section. The proportions may
vary greatly, and in fine hairs and wool the medulla may be missing.

with these simple themes, every species of mammal having its own hair
style by which it may be characterized.

The diameter and shape of a hair fibre and its parts changes from tip
to root. Often these features are characteristic and are employed in the
identification of the hair (Lochte, 1938; Hausmann, 1925; Wildman,
1955). Hairs normally cease growing after a more or less definite period,
thus achieving a genetically-determined length and, after a further period,
are shed usually as the result of a new growth. As with feathers the
hairs of successive generations from the same follicle may be different in
character.

In what is usually taken as the primitive condition, hairs slope backwards
from head to tail and are rather uniform in type and length. This simple
pattern is approximated to in rodents; in other mammals great variations
involving slope reversals, whorls, tufts, etc., are common. The direction
of slope defines the hair-stream which is closely involved with other
structures present (scales and glands) and with the organization of the
permal fibrils.

As with feathers, a great variety of hairs has been evolved to meet

72 KERATIN AND KERATINIZATION

peculiarities of the life of their possessors. Only certain broad features
distinguishing the main types can be mentioned here. In the classification
based on various sources given by Danforth (1932) hairs are primarily
divided on the basis of the presence or absence of erectile tissue surround-
ing their follicles. In effect this division amounts to typing hairs according
to their sensory role. The follicles containing erectile tissue are richly
enervated (vibrissae, tactile hairs, sinus hairs, whiskers, etc., are common
names) and their function is primarily sensory, the possibility of erecting
them stiffly adding to their sensitivity. This could represent a more
primitive function. Hair follicles without erectile tissue are usually also
enervated but their role becomes more purely defensive or protective.
Commonly the stiff, longer hairs (guard hairs) are distinguished from an
undercoat of finer, softer, and often more curly or crimpy hair, whose
function largely is that of heat insulation. Usually the coarser hairs
appear ontogenetically earlier than the finer and their follicles are said to
be primary; later-formed follicles are termed secondary (Fig. 35).

The much-studied human hair and sheep's wool are each exceptional
cases. The long human hair often called " terminal," a variety of guard hair,
is of limited distribution on the body which elsewhere is covered by a short
fine hair. Wool is largely made up of fine, crimpy secondary hairs and even
the primary hairs, although distinguishable, have also become fine and
crimped or curly.

The hair grows from a follicle which is an invagination of the epidermis
deep into the dermis (Figs. 40 and 43). Embryonically this forms
immediately as a downgrowth from the epidermis and is, in this sense,
in contrast with the early steps in the formation of a scale or a feather
and constitutes a reason for regarding hairs as having a different phylogeny
from these (see p. 73).

The fibrous properties of a hair reside in the cortex, a bundle of
longitudinally-aligned closely-adhering spindle shaped ( ~ IOOju, X 5 — 7/x)
keratinized cells (Lehmann, 1943, for illustrations). The keratinized
residues of these cells, when liberated from the fibre by enzymatic digestion
(p. 271) appear fibrous and are birefringent. They are similar, with
small differences in length, in all species. Woods (1938) showed that
cortical cells paralleled closely, in elastic behaviour, optical and diffraction
properties, the properties of the whole fibre.

The cuticle is of a contrasted construction. It consists of thin ( ~ 1/x)
sheet-like overlapping cells forming a protective sheath to the cortex.
Whereas the cortex is similar in most hairs, the cuticle is highly variable
and its features are much used in fibre identification. The thickness
varies owing to the degree of overlapping of the cells, or scales — one to
two in sheep, up to twenty or more in some fur hairs (Rudall, 1941 ;
Stoves, 1947). The degree of overlap of free margin, and the shape

THE KERATINIZED TISSUES 73

of the free margin, affect the external appearance and are features relied
on in identification. The free edges, projecting in a direction away from
the skin, make the fibre feel rougher when rubbed towards the skin. This
" directional friction " assists in keeping clean and in the grooming of the
hairy covering and in freeing it from tangles. However, when the hair is
cut from the skin the same property promotes tangling since each individual
hair then tends to creep persistently in a rootward direction when the
fibre mass is disturbed (Speakman and Stott, 1931). This tendency is
made use of in manufacturing felts from wool and fur, but is also the cause
of the shrinkage of woollens during washing.

The bulk of the keratinized feather also consists of long, spindle-shaped
cells very similar to the cortical cells of hairs. The surface layers are
covered with flattened, polygonal cells (Auber, 1955) perhaps analogous to
the flattened, cuticular scales of hairs but, unlike the hair cuticle cells, con-
taining within them a lattice-like network of fibrils.

The medulla is remarkably developed in the hair of certain animals, e.g.
rodents, where there seems to have been a considerable pressure in the
direction of producing a lighter, more bulky and stiffer hair for a given
weight of material. Medullary cells are often large and their fibrous
contents are concentrated peripherally against the cell wall producing a
cavity largely air-filled when the hair is dry. The mechanical problem
here (as with feathers too) is similar to that met with in constructional
engineering: to obtain maximum stiffness for a given expenditure of
material, and in fact many medullated hairs are reminiscent of girders.
In such cases, the pattern is genetically determined and is often of use in
identifying the hair.

In other types of hair only a feebler disposition towards medullary
formation is inherited; its actual manifestation depends upon the sizes
of the papilla and follicle and the nutrition of the growing hair. This is the
case among sheep where it assumes some economic importance. Rudall's
(1956) extensive survey of sheep follicles showed that the papillary
dimensions and shape control the appearance of the medulla (p. 150).

The Phylogeny of Hair

During the early heroic days of the application of the theory of evolution
to comparative anatomy, the phylogeny of so distinctive a mammalian
character as hair naturally attracted much attention and several theories
were advanced. That hairs are homologous at least remotely to feathers
and scales is obvious enough. Whereas the likeness between the feather
and the scale is close and is supported by their embryology, hairs are
sufficiently different to have led paleontologists to suppose that the actual
forerunner of the hair may have been some other organ of epidermal
origin. Some have found the precursor in teeth, others in cutaneous

74 KERATIN AND KERATINIZATION

sense organs, and others again in specialized scales. If the problem is
less discussed today, this is not so much through lack of intrinsic interest
as in the difficulty of obtaining further evidence which might bear upon
the question. Because of its importance some aspects of these discussions
will be summarized here.

The lining of the mouth is epidermal in origin and character, and the
formation of teeth, like that of scales, hairs and feathers, is another example
of dermoepidermal co-operation for the production of a superficial organ.

Fig. 32. Diagram of the early development of a tooth to show the
dermal and epidermal contribution to its structure. The tooth consists
of dentine, secreted by dermal cells of the dental papilla, capped by
enamel formed by ameloblasts of the enamel organ derived from the
basal layers of the epidermis. The keratinous constituent is found in
the enamel. Redrawn from Hyman (1947).

The earliest sign (Fig. 32) of an impending tooth is an epidermal pro-
liferation leading to an infolding of the epidermis, cf. the hair primordium,
to form the enamel organ. Beneath the enamel organ a mesodermal
papilla now forms and presses into the enamel organ to form a double-
walled cap. The cells of the enamel organ, the ameloblasts, are of epidermal
origin and secrete the hard enamel which caps the tooth; the cells of
the dermal papilla, or odontoblasts, secrete the dentine, the bulk of the
tooth. Thus the tooth, like the hair bulb and papilla, is in origin partly
epidermal and partly mesodermal.

THE KERATINIZED TISSUES 75

Teeth are certainly homologous with the placoid scales or dermal
denticles of elasmobranch fish (Romer, 1955). Embryonically, the forma-
tion of these scales follows a similar course although the denticles are said to
lack an enamel layer. There are thus similarities between teeth and scales
which are similarly distributed as widely over the body of primitive
fishes as are hairs in mammals. Brandt, in particular, was led to suppose
that, by a degeneration of the dermal element (dentine) and by a concomit-
ant increase in the epidermal contribution, a horny tooth capable of
evolving into a hair could be produced. He found support for this in
the existence of the genuinely epidermal keratinous teeth of the more
primitive cyclostomes (see Danforth, 1932). These teeth are horny caps
supported by a cartilagenous pad in an everted dermal papilla. There is
nevertheless little resemblance to true teeth, but some to scales and other
epidermal thickenings found in amphibians and higher types.

Another group of comparative morphologists has drawn attention to
the similarity in early embryogenesis between the hair primordia and
those of certain sense organs in fishes and amphibia and has suggested
that hairs have developed from these sense organs. A variant of this
theory would have that hairs descended from certain tactile spots on the
scales of reptiles.

The relationship between hair groups and scales pointed out many
years ago by De Meijere (see Noback, 1951) is suggestive in this con-
nexion. Hairs are never distributed uniformly or randomly over the
skin; there are regional variations in density and hair-type which are
as much a genetically determined morphological feature as any other
aspects of anatomy. Even in the hair-bearing areas, the hairs are not
randomly distributed, but are arranged in small, well-defined groups
(Figs. 33 and 35). De Meijere noted that the basic group seemed to be
three hairs with the larger in the centre. When scales were also present
(Fig. 34) the hairs emerged from the underside of the scales. This concept
of the " basic trio group " as a morphogenetic unit has been accepted as a
working hypothesis by most recent workers (Carter, 1943; Carter and
Clarke, 1957).

Studies of the course of embryonic appearance of hairs has partly
confirmed these views by showing that the first follicles to form (primary
follicles) are the central follicles of the trio group and that subsequent
follicles differentiate laterally to these to complete the group. The group
is not rigidly defined in numbers; some primaries remain solitary, others
have only a single lateral. Later other follicles, the secondaries, may
develop. The successive generations of follicles are related to the existence
in most hairy coats of the two distinct kinds of hairs : a coarser, longer
over-hair (guard hair) and a finer (often more woolly) under-hair. The
earlier developing follicles produce the over-hair and the later the fine

KERATIN AND KERATINIZATION

undercoat. Many of the differences between different furs and fleeces is
to be found in the relative development of the over- and under-hair.

When scales and hairs occur together, as on the tails of rodents, the
hair group develops in relation to the scale as mentioned above. The fact
that, when scales are absent, the hairs still form in groups suggests that the
ancestors of existing mammals may have had a scale associated with each
trio group and that in the course of evolution the scale has been lost,
leaving the hair group to mark the site. The argument is persuasive,

"'* • ■ m.
3

A
B

» ©®«®®
4

c © © ©

D © © ©
5

B .'•'.

c •:•■■

Fig. 33. Various arrangements of hair follicles illustrating the formation

of " trio groups " with the suggestion of a relation to an ancestral scale

distribution. Reproduced from De Meijere through Noback (1951).

but, when the origin of the hair-scale pattern itself is considered, we are
forced further into hypothesis. Two views have been advanced: (1) The
early mammals (or pro-mammals) may have been covered with fine
scales which diverged into two types, one of which continued to resemble
the reptilian form and the other became reduced in size and was converted
into hair. It may be significant here that the guard hairs of the primitive
platypus terminate in a flat spade resembling a scale. The follicle here at
first produces the spade and then turns over to producing a typical hair
shaft (Wildman and Hanby, 1938). (2) The hair precursors may have
been sense organs on the scales and later moved off into the softer inter-
scale regions where they became true hairs and commenced an independent
evolution. Alternatively the scale failed to appear, leaving the hair.

THE KERATINIZED TISSUES 77

It would seem that a better knowledge of the factors influencing the
relative rates of proliferation of dermal and epidermal elements when
these are co-operating to form an organ is necessary. The recent develop-
ments in the culture of hairs and feathers in vitro gives promise that this
may be obtained.

In the course of their development all these special structures are
necessarily related and more light is thrown on their relations by embryo-
logy than by an examination of the mature structures and their arrange-
ment. Fleischhauer (1953) has shown, for instance, that regional hair-

Fig. 34. The hypothetical scale-hair-gland complex. H-T the head-tail
line, S, scale; h, hair; g, gland; m, muscle. On the lower right-hand side
a view looking down on the unit. A further element not shown in this
diagram is the " hair disk " found in some cases behind the hair according
to Pinkus (1905).

streams can be detected before the hair germs appear. From spreads of
foetal human skin he concluded that the first hairs develop at roughly
fixed distances from each other in a quasi-hexagonal arrangement. New
anlagen appear between these when a critical distance is reached owing
to growth. This is a pattern which might be expected theoretically if we
regard these primordia as successful centres of proliferation which are
able to repress like developments in their immediate neighbourhood.
The two lateral hairs completing the " trio group " next appear in a line
at right angles to the main body axis. The three hairs already slope
backwards. The explanation usually offered for trio formation (above) is
that the hairs develop as though they were growing out from under
the free edge of a scale as indeed they do on the tails of rodents. There
is much to suggest in these facts that the scale, the hair group and its
associated glands together form a unit (Fig. 34) which in most mammals
has degenerated to the hair group and glands, the " pilosebaceous unit "

78 KERATIN AND KERATINIZATION

of Montagna. The " scale-hair-gland " unit is more general and
plausibly explains some aspects of the mammalian hair pattern, if we
suppose that the ontologically-earlier preparations are directed towards
the formation of the entire unit and that later the separate single com-
ponents come to develop to different degrees leading at last to the non-

FlG. 35. The basic trio group (11 '1') of primary hairs associated with
bilobed sebaceous glands (cf. Fig. 34) and the secondary hairs (22) which
appear later arising either de novo or from out-pocketing of the outer
root sheaths of the original primaries (Hardy and Lyne, 1956). The
entire cluster is marked off from its neighbours by bundles of collagen
fibres in the dermis. Redrawn from Carter (1943).

appearance of the scale. Hairs retain a clear sensory function as an
accessory mechanical lever for the stimulation of the nerve endings
associated with them.

Other possibly keratinized structures

A small amount ( ~ 2 per cent by weight) of a resistant, sulphur-
containing protein, usually referred to as a keratin, is found in the enamel
of teeth where it may contribute towards the bonding of this highly
crystalline, inorganic material (Scott, 1955). The enamel layer of a tooth
is a secretion of modified epidermal cells and thus any keratin it contains
would be a secreted protein. The formation of a tooth has already been
discussed on p. 74 and its components (partly dermal, partly epidermal)
may be seen in Fig. 32 which shows a section of a developing tooth.

The cystine content and basic amino acids of the organic matrix of
enamel have been determined (Battistone and Burnett, 1956; Hess, 1953;
and Block et al., 1949). The cystine is low for a keratin (0-2 per cent) and
the ratios of the basic amino acids show it to be a pseudokeratin as defined
by Block (see p. 31). Hydroxyproline, usually associated with collagen, is
reported by Hess et al. (1953) and by Battistone and Burnett (1956). The
X-ray pattern is suggestive of keratin rather than of collagen (Pautard,
1961).

THE KERATINIZED TISSUES 79

The " neurokeratin," a material remaining when nervous tissue is
exhaustively extracted, analysed by Block and found to contain cystine,
now appears to be simply a resistant product of decomposition. Histo-
chemists continue to describe a definite, cystine-containing " neurokeratin
network."

Keratin-like materials in extracellular situations are found in the
linings of the gizzards of gallinaceous birds (see pp. 30 and 107) and in
egg-shell membranes.

CHAPTER III

Differentiation and Protein Synthesis

In this chapter we shall discuss some general properties of epidermal
cells, such as differentiation and protein synthesis, which they share in
common with many other cells of the organism. Their more specialized
aspects relating to keratinization will be considered in a later chapter.

The cytology of keratinizing cells

A brief account of the histology of the epidermis and its derivatives has
already been given. For further details reference may be made to the
many standard texts (Maximov and Bloom, 1948; Horstmann, 1957;
Biedermann, 1926 and 1930; and Cowdry, 1932). The description which
follows is designed to draw attention to those features at the fine histo-
logical and macromolecular level which are particularly relevant to the
main themes to be discussed in later sections, or which, as a result of
recent electron microscopy, seem to require a redescription in terms some-
what different from those found in the classical works. The cytology of
the generalized basal layer cells and the fine structure of the dermo-
epidermal junction will be dealt with first. Then an account of the more
specialized structures of the hair and feather follicles will be given and
compared with that of the epidermis.

The Basal Layer Cells*

The intracellular equipment of the germinal cells in the basal layer is
similar wherever these are found, and consists of those elements which
are recognized as common necessities for the functioning of all types of
cell (p. 34) except those of bacteria and related small forms. The nucleus
is large and one or more nucleoli containing dense particles of the order of
120-200 A in diameter, may be found (Fig. 36 and Plates 4B and C,
7, 9, 10, 11 and 12). The nuclear contents which, except when the cell is
dividing, are diffuse and granular at all magnifications (as seen in electron
micrographs of thin sections), are strongly basic and stain positively with
the Feulgen technique for demonstrating desoxyribonucleic acid (DNA)
(Pearse, 1953). The nucleolus is Feulgen-negative, but gives positive tests
for ribonucleic acid (Montagna, 1956). The nuclear membrane consists of

* Also called Malpighian cells.

DIFFERENTIATION AND PROTEIN SYNTHESIS

two distinct sheets and is covered by small circular markings (Plate 4c).
Mitochondria are small and not particularly common in the cytoplasm
(Fig. 9). They posses a double-layered outer membrane enclosing an
inner chamber penetrated by what seem to be invaginations of the inner
membrane. According to Montagna (1956), mitochondria are often
difficult to see in the light microscope, but no problem arises in observing
them electron-microscopically. Numerous small vacuoles, often in
clusters, which may be identified with the Golgi apparatus (pp. 46-47)
are also visible in electron micrographs (Fig. 22e).

Fig. 36. The cytology of basal layer cells. BM, basal membrane;

N, nucleus; nu, nucleolus; m, mitochondrion; M, cell membrane;

D, desmosome; P, dense RNP particles.

The cytoplasm of the basal-layer cells is strongly basophilic and rich
in ribonucleic acid (Montagna, 1956 and Hardy, 1952). Electron-
microscopically the most striking feature is the presence of large numbers
of small (150-200 A diameter) dense particles (P in Plates 7 and 11). In
this respect these cells resemble many other kinds of rapidly-growing
cells: early embryonic cells in general, tumour cells, etc. Palade (1955)
has marshalled the evidence to show that the dense particles are in fact
the images of a ribonucleic acid-protein molecule (RNP). This matter is
discussed in detail on p. 108 et seq.

The cells rest on and are attached to a basement membrane (BM in
Plates 7, 9 and 14B) the detailed structure of which is considered below.
Partly as a result of this attachment on one face and of their close-packed
condition, the germinal cells assume a columnar form in some situations.
The basement membrane itself is supported by a dense feltwork of collagen
fibrils. The follicles of growing hairs and feathers penetrate below the
general level of the base of the epidermis and become enclosed in a basket-
like network of circular and longitudinal fibrils, which, however, do not

82 KERATIN AND KERATINIZATION

penetrate within the papillae of the follicles where the epidermal and
mesenchymal cells are separated by the single basement membrane.
The meshwork of collagen constitutes the " glassy membrane " visible in
the light microscope (Horstmann, 1957 and Montagna, 1956).

The one-sided attachment of the cells to the basal membrane also
establishes an intracellular polarity which is revealed by the often
asymmetrical arrangement of the cell contents. The small cluster of
vesicles (Fig. 36, p. 81) referred to as the Golgi complex, tends to lie
distal to the nucleus and mitochondria m may be more common nearer
the basal membrane.

The germinal cells, by their persistent cell division, maintain the
population of keratinizing cells. In this sense they conserve an embryonic
character which is emphasized by the generalized nature of their cell
contents. The cytoplasm of embryonic cells contains mitochondria, many
clusters of smooth-surfaced y-cytomembranes (p. 46), vacuoles containing
phospholipid and vast numbers of the small dense RNP particles. The
more specialized structures, such as the a-cytomembranes of secretory-
cells and the specialized inclusions, which later distinguish differentiated
cells, are rare. The similarity of this cytology to that of the germinal
cells of the skin is obvious. The surface of attachment of these latter cells
is, however, a specialized feature distinguishing them from the earliest
embryonic cells.

There are various opinions about the detailed course of the process
whereby the basal layer both maintains itself and supplies cells to form
the differentiated layers. A common view is that there is some asymmetry
in the division of a basal cell in the sense that two unlike cells result;
one, referred to as a " stem cell," remaining attached and preserving a
generalized character, the other free to move up and enter the stream of
differentiating cells. This cell may also be capable of further divisions.
Mitoses are to be seen among the matrix cells some distance from the
basal layer in the hair follicle and it has been maintained that dividing
cells are also to be seen well above the basal layer in the epidermis
(Thuringer, 1924). Most observers agree now that nearly all dividing cells
in epidermis are found in the basal layers and critical opinion (Hanson,
1947 and Leblond, 1951) holds that the earlier observations were unreliable
on the grounds that it is easy to be mistaken when examining oblique
sections. It seems more likely that the widely accepted view, that cell
division and cell differentiation are mutually exclusive, applies to the
epidermis. Cell division appears to cease in cells in which cytoplasmic
fibrils have commenced to accumulate, which, on the face of it, means that
the cells' synthetic activities have swung over from producing materials
needed for division to producing keratin precursors. The factors control-
ling mitosis are discussed in the next chapter.

*>•;

Plate 4 (Captions overleaf)

- LIBRARY }=*

\2/V MASS. >

Plate 4

A. Example of a plasma membrane at high resolution showing the ultra-
structure of this membrane. Material: Amoeba proteus, Os04 fixation.
Two dense lines ( ~ 20 A wide) are seen enclosing a third less-dense
layer ( /— ' 20 A) to form a sandwich-like surface about 60-70 A thick.
For interpretations see Fig. 18, p. 38.

B. The basement membrane as seen in the plantar skin of the rat. PTA
stain. M, basement membrane ; D, desmosome ; /, tuft of desmosomal
fibrils (tonofibrils) ; C, cross-sections of dermal collagen fibrils ; m, mito-
chondrion.

C. Section passing through the surface of a nucleus in an epidermal cell
of the 12-day-old chick embryo to reveal the " pores," the small circular
markings seen at O. PTA stain.

• siH— i ft

Plate 5 (Captions overleaf)

Plate 5

A. Examples of the surface of contact between cells. Figure from two
ectodermal cells A and B, in the surface of a 12-day-old chick embryo.
O, outer surface; D, desmosome; I, interdigitated membranes. Note
the two dense, parallel plasma membranes and the lighter deposit between
them revealing the intercellular cement. / is a tuft of fine fibrils (pre-
keratin) attached to the desmosome plates. The cytoplasm contains fine
dense particles but no particle-covered membranes. Stain: PTA.

B. Another example of an intercellular contact between two rat pancreas
cells A and B. Again the two contacting membranes appear as dense,
parallel lines separated by a dark material, the intercellular cement. /// is
a mitochondrion, and P are RNP granules attached to the system of
cytoplasmic membranes of these granular secreting cells. Stain: lead
hydroxide.

Micrograph kindly supplied by M. S. C. Birbeck.

Plate 6

A. The convoluted cell contacts in rat skin at high resolution to show the
multiple layers L deposited between the dense cell membranes or
desmosomes.

B. A further example of a complex membrane development in the
hardened cells of the Henle layer of the human hair. Between the dense
plasma membranes M is a single dense layer L. At T are to be seen
sections of two " ball and socket " joints formed by tongues of one cell
penetrating into the neighbouring cell.

"V * •• •

Plate 7

Portion of the basal layer of rat plantar epidermis and of the basal
membrane. N, large nucleus of basal layer cell; Nu, nucleolus; BM,
continuous basal membrane; C, bundles of dermal collagen fibrils
beneath the basal membrane; M, cell membranes; P, RNP particles and
G, Golgi cluster. The curious invagination of the nuclear membrane is
common in basal layer cells.

Plate 8

Xenopus tadpole tail epithelium. Near the end of the tail there are
only two layers of cells present and the mesodermal cavity has degener-
ated to the long gap A. No basement membrane lines the inner surfaces
of the cells in this site and at this time. Nil, nucleoleus ; N, nucleus ; m,
mitochondrion; L, phospholipid granule showing concentric whorls;
Mu, mucin pocket on the outer surface O of the cells; /, cytoplasmic
fibrils underlying the outer surface membrane. The inner surfaces of the
cells are poorly adhesive and pseudopods such as U project from the
unstabilized inner membrane ; at T the outer segment of the surface cells
is seen to be forming a close " adhesive contact " and incipient desmo-
somes D have appeared (see p. 90).

•;\V>V*

Plate 9

Portion of a basal layer cell of the epidermis of the tail of a frog tadpole
(Rana temporaria). Fixation: osmium tetroxide, stain : PTA. BM, basal
membrane; D, desmosomes; C, mesh-work of collagen fibrils which
support the basal membrane ; F, base of long process of dermal fibroblast,
which applies to and spreads over the surface of the collagen meshwork
presumably secreting further fibrils; M, mitochondrion; N, nucleus;
H, nuclear " pores "; v, small vesicles possibly arising from the pores;
and/, fibrils of cytoplasmic protein precursors of keratin.

DIFFERENTIATION AND PROTEIN SYNTHESIS 83

A further feature of the germinal cells, which is again indicative of their
unspecialized nature, is the behaviour of their plasma membranes. These
are seen (electron-microscopically in sections) as greatly-convoluted dense
lines which may often be separated by gaps of a variable width into which
may be thrust small surface protrusions. The pattern suggests that the
adhesion of the membranes to each other is not as strong as it later becomes

(P- 84).

In the epidermis, but less so in the basal layers of the hard keratins,
the cell surfaces are also studded with desmosomes (p. 41). The portion
of the cell surface facing the basal membrane is covered even more
extensively with dense, desmosomal-like deposits (Plate 14B). Between
these the plasma membrane appears more free and often small invagina-
tions are to be seen (" blebs ") which may well be associated with the
entrance of liquid since all the metabolites required by the epidermis
appear to enter through this layer (Pillai et al, 1960; Fasske et al, 1959).
Desmosomes have been described earlier (pp. 41-43) but, owing to
their importance in epidermal tissues, some further comment is required.
As observed electron-microscopically in sections, well-developed examples
appear as a pair of very dense, often planar (straight lines in sections)
deposits distributed over the cell surfaces. Examples may be seen in
Figs. 21, 36 and Plates 12C and 14B; their structure is shown dia-
grammatically in Fig. 21. They may be developed to varying degrees
ranging from a mere increase in the density of material immediately
adjacent to the cell membranes to large intracellular deposits in which
may be embedded tufts of filaments running into the cytoplasm (Plate 6)
and associated intercellular (extracellular) deposits. The two halves of a
desmosome are usually similar in degree of development. Since they
occur in situations where the transmission of mechanical tension from
cell to cell seems a reasonable supposition, most authorities think their
main purpose is to supplement cell adhesion and to form attachment for
fibrils, i.e. the desmosomes permit of an enhanced adhesion, the intra-
cellular deposits providing a sort of supporting backing for the fibrils
transmitting tension to other surfaces (p. 95).

The dense deposits on the membrane of the layer of cells facing the
dermis are in a sense " half-desmosomes " since, in the absence of a second
cell, one half is lacking. Nevertheless here too they seem to increase the
adhesion of the cell to the basal membrane and thus to the collagenous
meshwork beneath (Fawcett, 1958; Weiss, 1959).

Epidermal desmosomes are visible as small dots (diameter ~l/x) in
the light microscope. Thus the " desmosome " of the light microscopist
would include several distinct elements : the two cell membranes, the dense
intracellular deposits backing the membranes, the material between the
membranes and the terminations of fibrils in the dense bodies. They are

84 KERATIN AND KERATINIZATION

therefore not in any sense homogenous and histochemical tests merely
inform us of the presence of certain types of material without precisely
indicating its location. Wislocki (1951) demonstrated phospholipids and
Leblond (1951) describes a positive periodic- SchifT reaction indicating poly-
saccharides. Protein is undoubtedly present. Possibly the polysaccharide
is located intercellularly as the adhesive " cement " and the phospholipid
in the thickened cell membranes themselves (see also p. 94). By digesting
skin with various enzymes before fixation, Weiss (1958) was able to
" dissect " these basal membrane desmosomes. After pancreatic lipase
the dense surfaces lost their osmophilia indicating a high lipid content.
Amylase also freed the epidermal cell suggesting a polysaccharide
constituent.

Cell contacts during differentiation

The relation of the plasma membranes of cells in close contact in the
epithelial type of tissue of ecto- and endodermal origin has been described
in Chapter 1 (p. 40). Typically the two dense plasma membranes run
closely parallel to each other and are separated by a less-dense layer of
material usually of the order of 200 A thick. We shall refer to such a
contact as adhesive. Its strength may, of course, be supplemented by the
specialized organs of attachment, the desmosomes. In early embryonic
tissues generally, in many tissues of mesenchymal origin and in the germinal
layers of constantly proliferating tissues (epidermis, intestinal mucosa,
etc.), the adhesive type of contact is less extensively found. In these
tissues the cell surfaces are more convoluted, wider and less regular
intermembraneous spacings are to be seen (Fig. 37) and the cells seem
able to force small protrusions (microvilli, tubular pseudopods) into the
intercellular space. The poor adhesion is shown by the ease with which
embryonic material may be dispersed into single cells (Weiss, 1958;
Moscona, 1952, 1956 and 1957).

Thus it appears that during the progression from embryo to adult there
is, in certain organs, an enhancement of intercellular adhesion. A change
of this nature could have several consequences which may be factors in
influencing the maturation of the tissues taking place in this period. We
have given reasons (p. 44) for regarding differentiation as a phenomenon
arising from the action of one cell upon another, or in other words, of
intercellular communication. This communication must be effected either
by the transfer of samples of cell product from one cell to another, or by an
effect produced when cells come into contact. In either case a change in
the nature and activity of the cell membranes will influence communication.
The further possibility that intercellular adhesion can also play an active
role in moulding the shape of cells and guiding their movements is
suggested by the work of Holfreter (1947 and 1948), Weiss (1958),

DIFFERENTIATION AND PROTEIN SYNTHESIS 85

Abercrombie (1947 and 1948) and Moscona (loc. cit.) among many others
(De Haan, 1958). That some deterioration in cell adhesion is associated
with cancer has been advocated particularly by Coman (1954).

Certain essential notions concerning the activity of cell surfaces and
their behaviour when they are brought into contact are best derived from
time lapse films of living cells, such as those made by Weiss, and by
Abercrombie and Ambrose. From such observations one gains the
impression that many isolated cells behave essentially like unicellular
protozoa of the amoeboid type. Their surfaces are constantly thrusting

Fie. 37. The increase in intimacy of cell contact in passing from an
embryonic to an adult condition. A similar change occurs in the differen-
tiation of epidermal tissues.

forth protrusions which bring about the movement of the entire cell.
When like cells are brought into contact the motion of the surface ceases
at the point of contact, a phenomenon called contact inhibition by
Abercrombie (Abercrombie and Heaysman, 1953 and 1954), and the area
of contact may spread zipper-like further immobilizing the cells. On the
other hand, unlike cells on meeting do not inhibit each other's movements.
Thus, as is shown most clearly in the experiments by Moscona and Weiss,
in a mixture of like and unlike cells, the like cells, as a result of their
specific adhesion following random contact, will sort each other out
(Fig. 37). Under other circumstances, a zipper-like spread of contacts
may actually mould the cell formation into columnal epithelium (Fig. 44)
or an intricate interdigitated condition as seen in the notocord (Waddington,
1956). These specific intercellular adhesions seem to be effected by a
sticky exudate covering the cell surfaces which we have referred to as an
intercellular cement.

Little is known of the chemical nature of these cements and, further,
little concerning the structural devices involved can be learned from light
microscopy since the intercellular spacing is of the order of 200 A, far
below the resolving power of the light microscope. However, electron
micrographs of thin sections have established the existence of the thin

86 KERATIN AND KERATINIZATION

plasma membranes and of the less-dense layer of rather constant width
which separates them and which must represent in life the site of the
postulated intercellular cement. Evidently the intrinsic density (electron-
scattering power) of this material is low and, further, it does not react
with any of the common fixatives (osmium tetroxide permanganate and
formaldehyde) to produce a denser reaction product. After a treatment of
the fixed material with phosphotungstic acid and/or lead hydroxide
(Birbeck, 1959), it becomes visible (Plate 5). These findings merely
suggest that the intercellular layer (or exudate) is present in low absolute
concentration, and that it consists of chemical substances of an unreactive
character. Among the various suggestions compatible with these rather
negative requirements, is that it consists predominantly of a polysaccharide,
probably with a protein moiety (mucopolysaccharide) responsible for the
specificity. We may suppose that during the early stages of embryogenesis
it is secreted by or shed from the cell surfaces, each cell type producing
its own specific layer. See also the remarks made above concerning the
composition of the more specialized desmosome p. 84.

Since the epidermal cells undergo rapid differentiation when they
leave the germinal layer, we are able to find in the stream of cells (Fig. 42),
taking its origin in this layer and ending in the fully-keratinized layers, the
whole sequence of changes preserved at one time in the correct sequence.
It is this circumstance which further recommends the use of epidermal
tissues for the study of differentiation. The small volume in which the
changes occur makes the material ideal for electron microscopic study
which alone permits a visualization of the cell membranes themselves.
The results of the electron microscopic study of the developing epidermis
are given in the next two sections and of the developing hair follicle on
p. 95.

The Dermoepidermal Junction

The basal membrane is a structural feature which seems essential to
the establishment of an epithelium, for without it there seems no reason
why the intercellular adhesion postulated above should produce an
orderly, layered structure rather than a ball of interdigitating cells. Its
structure and formation therefore require special consideration.

Owing to the difficulty of resolving the fine details of the structure of
the dermoepidermal boundary, its nature has been much in dispute.
More recently, electron micrographs of a sufficient variety of tissues drawn
from amphibian, avian and mammalian sources have clarified the issue
and show that essentially the same structure is present in all these classes
of organisms (Weiss and Ferris, 1954; Porter, 1956; Jackson, 1954; Selby,
1955; and Mercer, 1958). Proceeding from within a basal-layer cell and
moving towards the dermis (Fig. 39), we encounter firstly the plasma

DIFFERENTIATION AND PROTEIN SYNTHESIS

Tail tip

Ectoderm

Stage 1 . Ectodermal cells showing
very convoluted surfaces indicative
of poor and impermanent adhesion.
No adhesion to or between meso-
dermal cells.

Stage 2. Formation of close contact
between those portions of the cell
membranes near to the outer surface.
Further in from the surface wide
gaps and convoluted surfaces persist.
Traces of intracellular fibrils.

Stage 3. Further differentiation of
surface cells. Appearance of (1)
mucin cells, (2) ciliated cells and
(3) definite deposits of intracellular
fibrils.

Stage 4. Epidermis increases in
thickness and basement membrane
appears. Cells in basal layer contain
both filaments and endoplasmic
reticulum.

Stage 5. Appearance of first colla-
gen fibrils beneath basement mem-
brane. Mesodermal cells increase
amounts of organized reticulum.

Stage 6. Rapid build-up of colla-
gen beneath basal membrane, in-
creased synthesis of intracellular
fibrils, fading of reticulum in surface
cells. Mesodermal cells recognizable
as fibroblasts F (much rough sur-
faced reticulum) and muscle cells
(little rough surfaced reticulum and
muscle fibres M).

Fig. 38. Shows the succession of cell types noted on passing from the
extreme tip of a tadpole tail (Xenopns and Rana) towards the head. The suc-
cession also represents stages in differentation from a poorly-differentiated
condition at the tip to a well-defined state of differentiation further head-
wards and may thus be regarded as a time sequence also. Six stages are
recognizable and are indicated (l)-(6) on the l.h.s. See Plate 8.

88 KERATIN AND KERATINIZATION

membrane of the cell about 70 A thick, the fine structure of which was
discussed above (p. 37); next we enter a lighter space (150-200 A wide)
which is continuous with the lighter space surrounding and separating the
cells of the overlying epidermis and is, presumably, of a similar nature.
Beneath this layer again we encounter another more dense, diffuse layer

Fig. 39. The fine structure of the basal membrane and an hypothesis to
explain its formation.

M represents the cell membrane of two contacting ectodermal
(epidermal) cells A and B. Their surfaces are covered with a sheet of
intercellular cement which functions as an adhesive (I). On the surfaces
facing the mesodermal space a diffuse, denser and thicker membrane
(BM) appears which defines the boundary between epidermis and
dermis and is named the basal membrane.

It is supposed that this membrane results from an interaction
between the fixed ectodermal exudate forming the intercellular cement
(I) and a more diffusible component emanating from the mesodermal
cells.

which is everywhere continuous (surrounding hair follicles and other
epidermal irregularities) and represents a quite definite morphological
surface separating the dermis and epidermis. We have reserved for this
particular layer the name basal membrane {BM, Fig. 39).

Membranes of this special character, to judge from the limited electron-
microscopic data yet available (Policard and Collett, 1959), appear to form
whenever two populations of cells which have followed sufficiently different
embryonic pathways, and become differently differentiated, are brought

DIFFERENTIATION AND PROTEIN SYNTHESIS

eratinlzed dome
of inner root sheath

mal papilla

%@P

Fig. 40. The hair cycle in the rat : (a) the commencement of growth in a

follicle containing a club hair; (b), (c) and (d), stages leading to full

growth; (e) and (f), cessation of growth. Reproduced by permission

(Johnson, 1957).

90 KERATIN AND KERATIN IZATI UN

again into physical contiguity. Thus they are found wherever mesodermal
tissues contact ecto- or endodermal tissues (Fig. 12, p. 23). They are
apparent after various fixation procedures and are more dense after the
preparations are stained with such " electron-dense ' ' materials as phospho-
tungstic acid. Their origin and nature will be discussed further below.
Beneath the basal membrane are found bundles of collagenous fibrils,
the characteristic product of the dermis (Fig. 25). These fibrils are found
more or less regularly arranged in different sites and in different animals.
They may form a well-organized orthogonal pattern in amphibian larval
skin (Plate 9) which is smooth and free from structural disturbances
caused by hair follicles (Weiss and Ferris, 1954 and Porter, 1956) or less
well-organized bundles in mammalian skins (Ottoson et al., 1953; Selby,
1955; and Mercer, 1958). For the anatomy of the collagen deposits, see
Horstmann, 1959 and Salecker, 1944.

The Development of Basal Membranes and their Role in the
Formation of Epithelia

Since the epidermis (ectoderm) must early take over the task of holding
the embryo together, it seems clear that the ectodermal cells develop
self-adhesion at an early stage in embryonic development. Thus the
study of the establishment of the early epidermis offers a means to
investigate these fundamental events. The early work of Holfreter
(1947) and the electron microscopy of Weiss and Ferris (1954) and Porter
(1956) showed that the larval amphibian skin was useful material for this
purpose. These electron-microscope studies have mostly concerned the
organization of the dermal collagen fibrils. The present writer has examined
earlier stages, commencing before the appearance of the membrane, in the
tails of tadpoles of Xenopus and Rana. In a Xenopus tadpole (Stage 19,
Nieuwkoop and Faber, 1956) and similarly-developed Rana tadpoles the
extreme tip of the tail consists only of a single layer of ectodermal cells
enclosing a cavity containing very few cells of mesodermal origin.

The cells are already showing the epithelia habit but, significantly they
are only closely adherent at their outside edges, the inner faces being widely
separated and highly convoluted (Plate 8). There is no basal membrane
and no dermal collagen. Later stages (developmental-wise) are found
nearer the head. Here the tail skin has acquired a defined epidermal
appearance, a basal membrane exists and the deposition of collagen has
commenced (Fig. 38) (p. 87). In these more advanced areas the dense
basal membrane (BM) everywhere follows the smooth undersurfaces
of the epidermal cells (Figs. 38 and 39) at a rather fixed distance. The
lighter layer (C) between it and the cell membrane (M) is continuous with
the lighter layer extending between the cells. It is probably missing on
the free surfaces facing the external environment where significantly may

DIFFERENTIATION AND PROTEIN SYNTHESIS 91

be found occasional bunches of cilia, finger-like protrusions or mucin-
containing pockets.

The basal membrane once established appears to provide a suitable
substratum for the crystallization of fibrils of collagen, the soluble precursor
of which (tropocollagen) is thought to be produced by the mesodermal
fibroblasts (Fig. 38 and Plate 9).

Essentially similar observations concerning the development of inter-
cellular adhesion between the epidermal cells of chicken embryos have
also been made by the present writer.

Thus the key steps in epidermal differentiation appear to be :

(a) The appearance of close contacts between cells which results in the
monolayer of surface cells acquiring a coherent epithelial character.
Differentiated features (intracellular filaments, cilia and mucin droplets)
appear on the outer surface about the same time and recall the normal
differentia found in cells with free surfaces (p. 44).

(b) The appearance of a basal membrane — a diffuse sheet underlying
the basal layer cells.

(c) At this stage the basal membrane cells are (to judge from their
fine cytology) still polyfunctional. They have the dispersed basophilic
reticulum of a secreting cell (collagen ?). Collagen in any event rapidly
accumulates beneath them; the mesodermal fibroblasts also spread out
over the sheet and appear to secrete directly against the collagen mesh-
work. By this stage the definitive histology is established. The explanation
of these events must be speculative at the present time but, with this
caution in mind, it is worth while proposing what are in effect working-
hypotheses as follows (Fig. 38):

(i) The first differentiations of superficial cells (cilia, mucin formation,
synthesis of intracellular fibrils) are similar to those noted in single cells
and they appear because the surfaces are free. While this explains nothing,
it refers the problem to a larger one not confined to the Metazoa. In these
early stages most of the cell surfaces are free and, to judge from their
convolutions, in active movement. They are little removed from free-
living cells, as is shown, indeed, by Holfreter's observations on cells,
liberated from amphibian embryos by reagents dissolving the weak inter-
cellular adhesive, which were able to assume an amoeboid habit and move
about.

(ii) Organized multicellular differentiation is initiated by intercellular
adhesion which commences between the contiguous portions of cell
membranes facing the external environment and from there travels
inwards. It is assumed that this is because the cells at this point start to
secrete an intercellular cement; their surfaces become sticky. This would
be the decisive metazoan feature.

KERATIN AND KERATINIZATION

(iii) The superficial cells (ectoderm) lead the way in differentiation and
it may be supposed that diffusible products from them influence the
underlying cells (Fig. 27). Rose (1952) has supposed that the appearance
of one form of product in a group of cells will suppress a similar appearance
in neighbouring cells and permit a second type of differentiation to arise.
Certainly, however this may be, the mesodermal cells henceforth are
enclosed in a bag of already-differentiated cells whose products could
influence them. Fibroblasts soon are recognizable by their content of

Fig. 41. Types of cell contact seen in the hair follicle. (A) the convoluted
cell membranes with irregular intercellular spacing seen in the undifferen-
tiated matrix. (B) localized contacts spreading to give uniform contact.

(D) and (E) Complete intercellular formations noted in the keratinizing

levels and higher (from Birbeck and Mercer, 1956 and 1957).

membranes covered with dense particles (RNP). At this stage (Fig. 38)
the two races of cells become separated by the appearance of the basal
membrane (BM). The peculiarities of this membrane are its fuzzy
character, its poor reaction wTith osmium tetroxide, its staining with
phosphotungstic acid (protein), its situation as a further sheet covering
the sheet of material already covering the ectodermal cell membranes and
which elsewhere acts as an intercellular adhesive. These features combine
to suggest that it is a reaction product (precipitate) formed between proteins
leaving the ectodermal cell and a more rapidly-diffusing substance
emanating from the mesodermal cells, as suggested by Fig. 39.

Henceforth the epithelium is established as a constantly-multiplying and
migrating cell population whose cell movement, in an inward direction, is

DIFFERENTIATION AND PROTEIN SYNTHESIS 93

constrained by this barrier formed beneath them. The basal membrane,
itself becomes the seat of further reinforcement in the form of collagen
fibrils. It is therefore proper to emphasize the importance of the basal
membrane in the architecture of the tissue; and further, one must
pinpoint, as the earlier crucial step in organ construction, the secretion
by the epithelial cells of a substance which can (a) link specifically to
sites on their surface and become an intercellular cement, and (b) when
the cells face the mesodermal space, react with mesodermal products to
form a basal membrane to which the cells are anchored.

A further important development which takes place at about the same
time as the formation of the basement membrane is the appearance of a
patterned arrangement of dense deposits over the surface of the cell facing
the basement membrane. These are the dermoepithelial desmosomes
(Fig. 36 and Plate 14B). They are not identical in structure with those
which form between epidermal cells (epidermal-epidermal desmosomes)
or in similar epithelia and which are illustrated in Fig. 21, p. 42 and
Plate 12C. That these structures are mechanically attached to the basal
membrane was shown by Weiss (1958). The deposit within the epidermal
cell often is double-layered (Weiss refers to them as " bobbins ") or even
multi-layered (Porter, 1956).

The reason for the formation of these localized deposits is not known.
They arise evidently in response to some external stimulus — another cell
or a basal layer. It could be, adapting an idea due to Weiss (1950), that
the localized external stimulus promotes a selective absorption of a cyto-
plasmic component on the plasma membrane which then provides the
site for further depositions. Alternatively, the cells response may be
due to the localized entry of some material from without which combines
with and precipitates an intracellular component. This is equivalent to
saying that the desmosomes mark the sites of porous spots. A section
running parallel to and through the neighbourhood of the basal membrane
(Fig. 39, Plate 14B) gives the impression of a rather regular arrangement
recalling that of the pores in the nuclear membrane (Plate 4C). The
concept of " membrane pores " is current as an explanation of specific
membrane permeabilities (Danielli, 1942). These are not thought of as
small " holes " but areas of modified porosity revealing themselves only
by the intracellular reaction provoked by an entering molecule.

That the ends of intracellular fibrils become attached to desmosomes
may be accounted for by assuming that the dense deposits present favour-
able sites for the initiating fibril growth. The importance of these attach-
ments for the ultimate mechanical function of the cells is obvious (see
also p. 95).

94 KERATIN AND KERATINIZATION

The differentiated layers and the variety of cell products

The Epidermis

The epidermis is an avascular, stratified, squamous epithelium of
variable thickness depending on the number and thickness of the cell
layers it contains (Fig. 25, p. 55). It necessarily contains a germinal layer
and may, as in the mouse, consist of only one other layer of horny cells
(Setala et al., 1960); in other animals and at other sites, and in some
pathological conditions, it may reach an extreme thickness of some
millimetres. Although the thin mouse skin does not reveal the usual
stratification, this does not mean that the intermediate stages of differentia-
tion are absent or that keratinization follows a different course. Cells
representative of the intermediate stages are in fact present, but insuffici-
ently common to constitute distinct layers (Gliicksmann, 1945). After
various treatments causing hyperplasia, and particularly after the applica-
tion of carcinogenic hydrocarbons (benzpyrene, methylcholanthene, etc.),
the thickened skin appears typically stratified.

When the mammalian epidermis reaches a moderate thickness (~ ten
cells deep) four layers can be distinguished: (a) the germinal or basal
layer (stratum germinativum); (b) the stratum granulosum; (c) stratum
lucidum; and (d) the stratum corneum (Fig. 25). The development of
these stages in differentiation seems inherent in the epidermis and depends
on the attainment of a sufficient thickness.

Even the basement layer cells contain loose bundles of dense filaments
(< 100 A diameter) (Plate 7) showing that synthesis has commenced
at this level (Plate 9). These filaments often sprout in bundles from
the sense plate-like desmosomes (D in Plate 4B) on the plasma membranes.
In the layer of cells immediately above the basal layer of the epidermis
there is an increase in the number of filaments ending on desmosomes,
making them more conspicuous and giving to the cell the appearance of
being covered with small short prickles. The layer is for this reason
referred to as " the prickle layer".

Tonofibrils. The fine birefringent fibrils visible in the light microscope
are often referred to as tonofibrils, particularly in the older literature.
The use of this term arises out of the view of the classical light histologists
who believed that the fibrils had a particular organization relating them to
the mechanical function of the tissue. The word " tonus " (Greek)
implies a " brace or support " and the word " tonofibril " expresses the
idea that the fibrils run from one face of the cell to another and are thus
capable of transmitting tension directly. Some authors, observing the
so-called " intercellular bridges," held further that the tonofibrillar system
ran continuously from one cell to another. This latter view is now un-
tenable since it is clear that, in several distinct cell types, fibrils certainly

DIFFERENTIATION AND PROTEIN SYNTHESIS 95

end on the desmosomes and do not pass across the cell membranes.
One of the best established examples of this is found in heart muscle
(mentioned above, p. 42) where parallel bundles of myofilaments are
seen to enter and fuse with elaborately- developed deposits of the
desmosome type which cover the cell membranes and are visible in the
light microscope as " intercalated disks".

The situation is not as clear in epithelial cells. Certainly many fibrils
end on desmosomes (Porter, 1956; Charles and Smiddy, 1957), but it
remains to be shown what proportion of fibrils have both ends attached.
In view of the extreme thinness of the sections used in electron microscopy
( < 500 A) proof of continuity is difficult to obtain. From a mechanical
point of view it is not necessary to assume that they all do since the whole
mass of fibrils is ultimately fused together during keratinization. There
is thus a use for both terms " fibrils and tonofibrils " and the latter may be
used when an author wishes to assert adherence to the view that the
fibrils run from one desmosome to another within the cells. In other
cases the less restrictive word " fibrils " will be used here.

Keratohyalin. In the lower layers of the epidermis the synthetic
activities of the cell appear to proceed directly to the formation of filaments
(F in Plate 9) and in this respect they resemble the cells of the hair cortex
to be discussed later. In the stratum granulosum of the epidermis,
however, a new product makes its appearance in the form of rounded
droplets which, accumulating, give the cells a granular appearance.
These droplets consist of a distinct substance called keratohyalin and a
great deal of study has been devoted to it, the main point at issue being
whether it is, or is not, to be regarded as a precursor of the fibrous keratin
of the stratum corneum (see also p. 228). The granules disappear in the
next cellular layer, the clear, glassy, highly-birefringent stratum lucidum*
The appearance of keratohyalin granules is one of the definitive
characteristics of the " soft keratinization " of the epidermis (see Table 5,
p. 65) to be contrasted with the keratinization of a hard keratin, such as
hair (next section) which proceeds without the formation of a granular
layer. In the squamous cells ot the inner root sheath of the hair follicle a
similar granular layer appears (Plates 22 and 23).

The Hair Follicle

The hair follicle forms embryonically as a down-growth of the basal
layer of the epidermis. The first sign, indicating a locally-increased rate
of cell division, is a cluster of smaller crowded cells above the basal

* Ranvier in 1879 introduced the term " eleiden " (from the Greek word for
" oil ") as a name for the clear glassy substance of the stratum lucidum. Although
this word occurs frequently in the older literature there does not seem to be any
need for it in the present description.

96 KERATIN AND KERATINIZATION

membrane, the primary hair germ. As proliferation continues, the bud
projects downwards and penetrates the dermis as a solid cord of cells.
Perhaps conditions here favour further division, the cells finding more food
and living space. The advancing tip calls forth a response from the dermis
shown by the approach of a number of dermal cells. The two kinds of
cells do not mix, since they are already separated by the basement
membrane. The advancing epidermal cells spread over the dermal
cluster to form a cap, the future hair bulb, and the mesodermal elements
within form the future papilla. These events are partly recapitulated each
time a new hair develops in a resting follicle (see Fig. 40).

This co-operative relation between dermal cells and epidermal has
impressed many observers; moreover, the morphogenetic control which
subsequently develops in the follicle is also dependent upon the continued
presence of the dermal component. This is shown perhaps most clearly
by Hardy's (1949, 1951) experiments on the formation of hair in tissue
culture and by equivalent experiments concerning the development of
feather germs (Lillie, 1942). In the absence of dermal remnants, true
follicles never develop in tissue culture; together, dermis and epidermis
can produce and maintain follicles with differentiating cells. No other
accessory structures, such as blood vessels or nerves, are required.
Moreover, the arrangement of the follicle in groups is similar to that of
skin growing in situ. This is an important demonstration of the morpho-
genetic competence of the skin and it emphasizes that the control of
differentiation in the epidermis is localized (see p. 146).

After the papilla has formed within the tip of the cells descending
from the basal layers of the epidermis, mitosis becomes restricted to the
lower half of the follicular bulb. The hair now sprouts up from this
matrix and penetrates the originally-solid plug of epidermal cells which
thus become the outer root sheath (Fig. 40). The advancing tip forms a
cone of cells continuous with those of Henle's layer of the inner root
sheath.

Typically follicular activity is cyclic, a growing phase is followed by a
resting phase (Montagna, 1956); Fig. 40 depicts, for example, the events
in the cycle of growth in rat skin. Cyclic growth is discussed in the next
chapter. In a condition of steady growth the hair follicle offers a remark-
ably compact example of organogenesis. From the mass of undifferentiated
cells in the lower half of the bulb (Fig. 42) arises a solid cylinder composed
of six concentric cylinders each consisting of cells which become different
in shape and which follow different paths in internal development. In
terms of their characteristic products of synthesis, the three external
cylinders, each a single layer of cells,* which comprise the inner root

* The layer of Huxley may consist of more than one layer in thickness, in
asymmetrical hairs (Rudall, 1956).

DIFFERENTIATION AND PROTEIN SYNTHESIS

(>7

sheath, and the central core, the medulla, may be classed together, since
they each form a peculiar protein called trichohyalin (probably very
similar to the keratohyalin of the skin). It forms as amorphous droplets
and is later converted into a fibrous form. Chemically it is distinct from
keratin and is unique in containing the amino acid citrulline (6 per cent

7 6 5 4 3 2 1

LAYERS FROM OUTER ROOT SHEATH

Fig. 42. Purely diagrammatic representation of the relations of the
several cell streams (1-7) in the hair follicle to the basal membrane M and
the outer root sheath to illustrate the " position " theory of differentiation.
The intercellular gaps are emphasized. C is a melanocyte, the papillary-
space is to the left. The dense dots in the inner root sheath cells (Huxley
and Henle layers and cuticle 2) are trichohyalin.

by weight) (Rogers, 1959). The other concentric cylinders form the
cuticle and cortex of the hair, each containing a distinct variety of keratin
(Fig. 42).

In respect of their developmental history, the cells of the bulb are
equipotential and could apparently proceed to synthesize any of the
epidermal products. The sole factor, which initially seems to distinguish

KERATIN AND KERATINIZATION

DEVELOPMENT OF THE HAIR CUTICLE

Fig. 43. A hypothesis of the differentiation of the hair cuticle in terms of
the development of intercellular adhesion.

(a) A germinal cell with poor intercellular adhesion and convoluted
(active) membranes.

(b) The smoothing out and immobilizing of membranes as a result of
the appearance of areas of close intercellular adhesion.

(c) and (d) Complete stabilization of membranes followed by a flatten-
ing and tilting due to the spread of adhesion " zipping " the cells together
(See also Fig. 44).

(e) and (f) the commencement of protein synthesis. (From Birbeck
and Mercer, 1957). See also Fig. 41.

DIFFERENTIATION AND PROTEIN SYNTHESIS

one from another, is their distance from the outer root sheath or from the
dermal environment (Fig. 42) which suggests again that some influence
from this direction controls differentiation. Differentiation is usually
recognized in the light microscope by a change in cell shape or by an
alteration in staining properties or cell texture which indicates the accum-
ulation of different chemical products. However, as stated earlier, electron-
microscopically the cell membranes (as defined above) can be clearly

Fig. 44. (a) to (c) The conversion of a flattened into a tall columnar
epithelium as a consequence of the spread of intercellular contact.
(Refer to Fig. 43.)

(d) The development of "tilting" in a cuticle, for example, owing to
the creeping of a cell tip due to enhanced adhesion. (Birbeck, 1957).

distinguished, and their behaviour provides an earlier indication that a
change is taking place (see Fig. 41).

Birbeck and the present writer (1956, 1957a and 1957b) have tried to
correlate the outset of differentiation with differential membrane adhesion
between the several presumptive layers of the hair. The course of events
revealed in a longitudinal section of a follicle is illustrated diagrammatically
in Figs. 42 and 43. The first cell layers to distinguish themselves from the
undifferentiated mass in the lower bulb are those of the cuticles. It may be
significant, from the point of view of the penetration of an inductive effect
from the dermis, that although these layers are of a variable distance from

100 KERATIN AND KERATINIZATION

the papilla, they are always the fourth or fifth cell layer from the surround-
ing basement membrane. Their cells acquire a cuboidal shape, which is
easily apparent, since their membranes have drawn together zipper- wise,
effacing surface irregularities and becoming somewhat denser. The same
events follow in the layers of Huxley and Henle; they occur several cell
diameters higher in the central mass of cells destined to become cortical
cells. Between these latter cells, openings are observed as high as the
papillary tip where they are occupied by processes of the melanocytes
(see p. 276).

While it must remain purely speculative in our present state of know-
ledge, it is possible to suppose that the cuticles lead the way in differen-
tiation because they, in some way dependent on their position relative to
the dermis, first begin to form and secrete their specific intercellular
cement. The possibility of developing a columnar epithelium by a zipper-
like spread of adhesive cell contacts was envisaged by Schmitt (1941)
(Fig. 44), and the change in shape of the cuticle cells looks very like an
illustration of Schmitt's hypothesis.

Intracellular differentiation in the hair bulb

The first visible signs of synthesis of intracellular products are observed
in the cells immediately in contact with the external root sheath layer.
These cells become Henle's layer, the most peripheral layer of the inner
root sheath (Fig. 42). Small trichohyalin droplets appear in the cyto-
plasm of cells at about the middle line of the bulb and rapidly grow in size.
Henle's layer hardens, becomes birefringent and clear suddenly at the
level of the constriction of the bulb. Its function at this level appears to
be largely mechanical; it provides a solid cylindrical support to carry
upwards the soft cells within. Its outer surface is said to slide over the
surface of the outer root sheath, which in this respect is a static structure
and does not move out with the rest of the follicle. The mechanics of
these movements are far from being clearly understood (Montagna, 1956;
Auber, 1950).

When first formed the trichohyalin droplets are isotropic ; at the level
where Henle's layer hardens, the cells suddenly become birefrin-
gent and the layer itself becomes clear and more difficult to stain.
Electron-microscopically fine filaments or ribbons can be seen extending
from the tips of the lenticular-shaped droplets of the amorphous precursor
(Plate 21). These condense to give a compact highly-oriented mass
which is the clear, glassy, birefringent layer visible in the light microscope.

The course of events in Huxley's layer and in the cuticle of the inner
root sheath is similar, but the tempo is slower, the cells remaining full.
Following Auber (1950) we may suppose that these cells form a firm but
tenacious vice to grip and support the softer tissues of the hair itself.

DIFFERENTIATION AND PROTEIN SYNTHESIS 101

The synthesis of fibrous keratin commences in the cells of the cortex in
the middle and upper bulb. Small wispy bundles of filaments are already
visible electron-microscopically in the cells of the middle bulb (Plate 12B).
There is no obvious accumulation of non-fibrous precursor as occurs in the
sheath cells; however, if follicles are fixed for 12 hr in buffered formal-
dehyde and stained with phosphotungstic acid, a procedure designed to
retain more completely the contents of the cells, the cytoplasm is seen to be
packed with a structureless protein which may suggest the existence of a
precursor in a soluble form.

Synthesis is delayed in the hair cuticle until above the bulb, when
droplets of an amorphous keratin separate in peculiar patterns packed
against the outer wall of the tilted cuticle cells (Fig. 43) (Plate 20).

The characteristic products of the hair follicle, fibrous keratin, cuticular
keratin and fibrous trichohyalin, are thus formed in distinctly different
ways which will be more fully discussed in a subsequent section (p. 223).
Returning for a moment to consider the cells of the epidermis, it would
seem that there the two methods of forming fibrils are actually to be found in
the same cell, although at different times, for filaments are built up directly
in the basal and prickle-cell layers and from keratohyalin in the stratum
granulosum. The conversion of the granular layer into the clear bi-
refringent stratum lucidum involves a transformation of the isotropic
keratohyalin granules into a fibrous form apparently analogous to that
found in the inner root sheath cells. The fibrils of both origins seem to
fuse into a common formation in the stratum corneum. This problem is
further discussed in Chapter 6.

The Feather Follicle

A brief account of the feather follicle has already been given (p. 69).
The feather grows out from a germinal matrix (Fig. 48) at the bottom of
the follicular shaft consisting of an ectodermal wall and a relatively large
mesodermal core called the feather pulp (Figs. 45, 46 and 47). The feather
cylinder itself comprises three layers: an external one forming a protective
sheath to the developing feather, a thicker middle layer from which the
feather itself is derived and an inner layer next to the pulp. The external
and internal layers may be likened to the root sheaths of the growing hair.
All three layers are produced by the proliferation of the generalized cells
of the matrix, called the collar, at the base of the follicle (Fig. 48).

In its growing phase the feather is a hollow, pointed cylinder set like a
cap on the core of mesoderm; it simply elongates owing to the addition of
cells by division at its lower end, the " collar ". Some details of this process
including barb formation are illustrated in the simplified drawing Fig. 45.
Here at (a) a portion of the growing cylinder is seen from its ventral side.
R is the rhachis found on the dorsal side, the barbs B, greatly reduced in

102

KERATIN AND KERATINIZATION

number, are indicated by their distance from the central rhachis. The
collar is at C. The barbs form a subdivision of the cylinder by clefts as may
be seen more clearly in (b) where following Lillie the cylinder is shown cut
along its ventral side, spread out flat, and viewed from inside the follicle.
Cell movement is everywhere vertical, i.e. parallel to the rhachis, and, in
the area of the collar giving rise to the rhachis, is continuous; in those

O 1 2 3 4

sNS^;

'////

V^V\

v///

— *

«—

R 5 b 7 8

Fig. 45. Illustrating rhachis and barb formation in the feather follicle,
(a) A short length of the feather cylinder adjacent to the collar, (b) the
cylinder opened out, (c) a section cut across (b). The rhachis R grows
steadily upwards. The bases of the barbs move inwards to meet the
rhachis as a result of the migration of the growing areas but cell movement
is always vertical.

regions giving rise to barbs, however, growth is restricted to the areas
shown in (c), separated from each other by clefts which thus give rise to a
series of ridges on the inner face of the cylinder. In nesting feathers the
ridges ran parallel to the rhachis (itself here only a barb) and the feather
when open appears as tube with a slotted end. The growing areas of the
collar are here stationary relative to the rhachis. In forming adult (contour)
feathers the discrete growing areas generating barb ridges migrate towards
the rhachis (Figs. 46 and 47) and ultimately meet it and join the barb to the
rhachis. Only the sites of growth migrate inwards, the movement of cells
remains vertical, and the effect is to produce the series of spiral grooves
inside the cylinder as shown in Fig. 45, each completed barb describing a

DIFFERENTIATION AND PROTEIN SYNTHESIS

103

Fig. 46. Diagrams of cross-sections of a feather to show the dorsal fusion
of barbs to form the rhachis (A-F) ; the fusion of barbs ventrally to form
the hyporhachis (D-F); the lateral fusion of rhachis and hyporhachis
to form the calamus (F) ; the structure of the calamus (G) and the tip of
the new feather forming within the base of the calamus of its predecessor
(H); ba, barb of aftershaft; bs, barb of shaft ; ca, calamus; h, hypo-
rhachis; i, intermediate cells (i.e. collar); p, pulp; ri, ridge; s', sheath
of old feather; s", sheath of new feather. (Redrawn from Hosker (1936).)

104 KERATIN AND KERATINIZATION

half spiral in the cylinder. During the formation of the continuous
cylinder of the calamus, which completes the feather, growth is continuous
around the entire ring of the collar.

This description follows the classical accounts due to Strong (1902) and
Davies (1889). More recently somewhat different views were advanced by
Lillie and Juhn (1932 and 1938) which envisaged an actual migration of
growing tissue tangentially along the collar to enter the mounting rhachis as
suggested by the drawing Fig. 47. Their view has been contested by
Hosker (1936) and 'Espinasse (1939) in particular, and it seems it cannot be
held in its extreme form. Lillie and Juhn make the point that the rhachis
appears an independent growth to which barb material secondarily
becomes attached (Lillie, 1942). Certainly surgical experiments prove that
the capacity to generate rhachis is a special differentiation of the dorsal
portion of the collar. Following removal of the ventral half of a follicle the
entire feather may be regenerated; on the other hand the removal of the
dorsal half leads to the regeneration of a feather lacking a rhachis. The
actual relevance of some of these data to Lillie and Juhn's theory is not
immediately apparent. An interesting discussion of these questions will be
found in Waddington's book (1952).

According to the concrescence theory the rhachis is formed by a process
of concrescence of the continually-growing right and left halves of the
collar, the levels from apex to base being formed successively (Fig. 47).
The forming barbs are carried along with the constantly-streaming halves
of the collar to their definitive positions at the sides of the shaft with con-
sequent change of orientation. As the series of barbs move dorsally (nos.
1-15, Fig. 47D), new barbs (nos. 16-25, Fig. 47E) take their origin in
the space thus provided at the ventral surface of the collar.

The cells of the germinal collar closely resemble in their cytological
features those of the hair bulb (Mercer, 1958). The basement membrane
is typical and the basal layer cells, which abut it, form a columnar-like
epithelium. They are strongly basophilic and their cytoplasm abounds in
clusters of small dense particles of the same kind as described in the
germinai cells of hair and skin. Differentiation becomes apparent in the
cell layers immediately above the basal layer with the appearance of wispy
filaments in the cytoplasm. Although these filaments are known to consist
of a jS-type keratin (p. 16), their appearance, their manner of formation and
the cytology of the cell, seem exactly similar to the cells of the hair follicle
forming filaments of an a-keratin (Plate 12A).

The production of the j3-keratin type of structure in the epidermis
of birds and reptiles poses some interesting questions, which have
been considered by Rudall (1949). Both a- and £-type keratins are
produced in these structures by cells which originally belonged to the same
primary ectoderm; the later development of two cell-types, distinguished

DIFFERENTIATION AND PROTEIN SYNTHESIS

105

Fig. 47. Diagrams to contrast the development of a feather according to
the fusion of barbs or classical theory (A, B, C) and the concrescence
theory of Lillie and Juhn (D, E, F). In the classical theory the whole
feather arises from a ring of embryonic cells (i.e. the collar) surrounding
the base of the feather germ. According to the concrescence theory the
barbs, nos. 1-9, arise from the collar and as more cells are added to
the barbs from the rapidly-dividing collar cells, it will follow that they will
gradually approach the mid-dorsal line, and fuse with the dorsal-most
barb or rhachis. This of necessity becomes broader and takes on the
definitive shape of the rhachis. b, barb; co, collar; r, rhachis; s, sheath;
v, ventral (from Hosker, 1936).

106 KERATIN AND KERATINIZATION

by the type of protein fibril contained, may be regarded as a modulation
although little is known about the stability of cell-type in the basal cells
supplying the different streams of cells. In the feather follicle the intermed-
iate cells (Fig. 48) produce feather keratin exclusively; the outermost layers
of the stratum comeum and the stratum cylindricum produce a-keratin with
little or no feather keratin. The several streams of cells advance in parallel
and the situation is not unlike that found in the hair follicle in which
several parallel streams also occur. In the hair follicle the weight of
evidence seemed to support the idea that the cells of the germinal matrix

Fig. 48. The germinal collar at the base of the feather follicle showing
the several cell streams (1, 2, 3, 4) arising from it : (1) is the layer adjacent
to the dermal papilla which gives rise to the medullary caps; (2) is the
feather proper; (3) the feather sheath; and (4) the epidermal lining of
the papilla. The distribution of protein types is shown on the left-hand
side.

were all similar and that their position determined their subsequent
development. In the feather system a more fundamental difference de-
velops: a difference in basic chain-type between the proteins produced
in the contiguous streams. No histochemical or electron-microscopical
feature distinguishes two classes of germinal cells. On his evidence Rudall
is unable to decide whether the factors for synthesizing the two kinds of
keratin are segregated into two classes of cells but considers the possibility
of cells producing both kinds of keratin in a mixed form.

In snake scales the hard horny outer layer is feather keratin and the less
compact inner layers give an a-pattern (Rudall, 1947). Here it would seem
that the same germinal matrix produces both types of protein in series.
Unfortunately, again a certain ambiguity remains as to whether precisely
the same cells in the matrix contribute to both layers. There seems good
reason to believe that growth rhythms (perhaps diurnal) occur in the
feather follicle. The regularity of the barb structure suggests this clearly,
the presence of growth bars and of the succession of bands of radio-
active sulphur deposited in the calamus (Liidicke, 1959) following injection
of radioactive sulphur compounds are further indication. The relation of
rhythmic growth to the whole phenomenon of feather formation has not
yet been explored in detail.

Plate 18 (Captions overleaf)

Plate 18

A. Section of gizzard of 17-day-old chicken embryo showing threads of
keratin S passing down the lumen L of the secreting epithelium. The
secretion gives a positive test for disulphide groups and appears to form
by the coalescence of fine filaments f.

B. Cross-section through an egg-shell membrane of the chicken egg. In
the light microscope the membrane is seen to consist of fibrils roughly
hexagonally arranged parallel to the egg surface. In e.m.gs. the fibrils are
cut in various directions and each is seen to consist of two components
(1 and 2); the inner component has reacted more strongly than the outer
with the osmium fixative and may thus be the keratin component (see
p. 107).

C. Part of a secreting cell.

1 NM

.y*

Plate 19 (Captions overleaf)

Plate 19

The cells secreting keratin to form the lining of the gizzard in a 17-day-old
chicken.

Top: S, secretory granules; L, lumen of glandular pocket; M, cell
membrane. The arrows indicate secretion escaping from the cells be-
tween the villi.

Bottom: A cell not filled with secretion granules in which particle-
covered membranes R, mitochondria M and Golgi apparatus can be seen;
S is the secretion in the lumen L. N, nucleus; NM, double nuclear
membrane.

differentiation and protein synthesis 107

The Avian Secreted Keratins

The secreted keratins, the horny lining of birds' gizzards and the egg-
shell membranes, have not been as extensively studied as their unusual
extracellular location would seem to warrant. We lack an exact knowledge
of their structure and mode of formation. Even the type of fibrous protein
found in egg-shell membranes is disputed. According to Champetier and
Faure-Fremiet (1938) the material gives a collagen pattern. On the other
hand, the chemical analysis of purified membranes reported by Calvery
(1933) shows it to be of the keratin type. The purification involved an
enzymatic digestion aimed at removing constituents less resistant than
keratin (mucin, collagen, etc.) which may account for the different findings.
The present writer has made electron micrographs of sections of hen's egg
membranes (Plate 18B) which revealed a felt- work of fibrils each of which
consists of two distinct components. By staining methods and the light
microscopy two separate fibrillar systems are usually distinguished (Moran
and Hale, 1936): mucin and keratin. It would appear that actually the
fibrils of one system (probably keratin from its stronger reaction with
osmium tetroxide) are enveloped by a deposit of a less reactive material
(mucin?).

No study has been made of the fine cytology of the cells whose secretion
coats the eggs with their membranes during their passage through the
oviduct.

The tough, thick lining of the gizzard, which protects the cellular
surfaces of this muscular mill against wear and tear, consists of coherent
sheets of protein mixed with mucin. The material gives a strong and
specific reaction for disulphide groups and is accepted as a keratin (Broussy,
1932). The layer is birefringent, but irregularly oriented. Longley (1950)
reports that it gives an X-ray pattern consisting of three haloes cor-
responding to spacings of 10, 4*7 and 3*7 A which would suggest it is a
/S-keratin (feather keratin?).

The layer is secreted by the cells of a columnar epithelium resting on a
typical basement membrane. The secretion collects as rounded liquid-like
inclusions in the apex of the cells and escapes to flow down a typical
glandular lumen to join the overlying material (Broussy, 1932 and Fell
(private communication)). An electron microscopic study of these cells in
21 and 17-day-old chicken embryos made by the writer shows their fine
cytology to resemble that of epidermal cells. RNP granules are common
but a basophilic (particle covered) reticulum is also present (Plate 19).
Clusters of smooth membranes (y-cytomembranes) were conspicuous.
A structureless basal membrane underlies the epithelium; there are few
desmosomes and no intracellular tonofibrils. In the body of the cytoplasm
and towards the cell apex the secretion appears as structureless rounded
droplets of various sizes (Plate 19) which stain strongly with PTA

108 KERATIN AND KERATINIZATION

indicative of protein rather than mucin. The free surfaces of the cells are
covered with stubby villi between which appear pockets containing
granules in the act of escaping. On leaving the cell (or even before it) the
granules commence a transformation. They lose their homogeneous
appearance, break into a cluster of coarse granules, then into finer granules
which appear to open into masses of short fine filaments (Plate 18 A) whose
coalescence produces thick, tapering rodlets and then the horny layer.

This sequence of changes seems at first sight unlike that found in the
formation of other keratins. However, there is a likeness to keratohyalin.
In each case an amorphous precursor is produced which accumulates as
droplets or granules. In the epidermis the keratohyalin granules transform
into fibrils within the cell; in the gizzard on the other hand, the secretion
escapes from the cell before the transformation into the fibrous form,
which then occurs extracellularly. The problem clearly requires closer
study. Since the horny lining is incompletely soluble in keratinolytic
solvents and is coloured, the possibility of tanning must be considered.
Mucins are produced by the same (or neighbouring) cells but the relation
between the two secretions is obscure.

The organic matrix of tooth enamel may be a keratin (p. 78) and recent
micrographs of Watson (1960) show that it is extracellular and that the
cytoplasm of the ameloblasts has the basophilio (particle covered)
membranes characteristic of secretory cells (see p. 113).

The synthesis of protein in epidermal systems

Cytology of Cells which form Protein

It is desirable to consider at first in greater detail the cytology and bio-
chemistry of proteogenic cells. Protein synthesis, in the sense of the
initial synthesis of a high molecular weight polypeptide having a specific
structure, occurs only within cells; certain other processes which, as we
shall see later, are important in the assembly and organization of fibrous
structures may, however, occur extracellularly. Epidermal cells as a class
are notably active in protein synthesis and, on account of their often simple
geometrical arrangement in such organs as skin or hair, they lend them-
selves to microscopic study. On the other hand, the number of cells
available is not usually large and this does not favour biochemical tech-
niques. Thus the wealth of morphological detail is offset by a lack of
biochemical information based directly on a study of epidermal cells
themselves and we are forced to accept as a working basis the general
conclusions of the course of synthesis worked out on more convenient
systems, such as the mammalian liver or micro-organisms. It would be
impossible to do more than present a sketch of this work which is develop-
ing rapidly at the present time.

DIFFERENTIATION AND PROTEIN SYNTHESIS 109

Glandular cells, such as those of the pancreas or thyroid, have a high
rate of protein synthesis and it has long been known that their cytoplasm
stains deeply with basic dyes (Caspersson, 1950 and Brachet, 1957). When
it was shown that this basophilia was due to nucleic acids a connexion
between these acids and synthesis seemed certain. There are two kinds of
nucleic acids distinguished, among other things, by the sugars they
contain: desoxyribonucleic acid (DNA) containing the sugar desoxy-
ribose, and ribonucleic acid (RNA) containing ribose. The view that
nucleic acids were somehow related to synthesis was argued some years ago
by Caspersson (1950) who used ultra-violet absorption methods to detect
them and Brachet (1957) who used specific dyes. Brachet more definitely
urged the participation of RNA. Since that time direct chemical analysis
has confirmed the presence of nucleic acid and an enormously diverse
amount of cytochemical evidence based on all sorts of organisms has been
accumulated to show that DNA is invariably present in the nucleus of all
cells, and that RNA is always present in the cytoplasm of proteogenic cells
(Brachet, 1957).

Caspersson at first suggested a scheme, based largely on the distribution
of ultra-violet absorption material, in which it was supposed that genes
(DNA) located on the chromosomes controlled the synthesis of histone-like
(basic) proteins which accumulated first as a nucleolus, and later passed
through the nuclear membrane and, entering the cytoplasm, provoked the
formation of RNA and the specific proteins. This view cannot now be
sustained in full. The presence of basic proteins is questioned and the
movements of RNA are thought to be different.

Current research is centred on the interrelations of the genetic DNA,
RNA and protein and both experiment and speculation are very active.
Almost all authors accept the view that genetic information is carried by
molecules of DNA (Crick, 1958) and that these molecules must therefore
be duplicated at each cell division. The DNA of the nucleus in its genetic
role is said to contain all the information to ensure its own replication and
to form the various other materials of the cell. It is a basic assumption (a
" central dogma " according to Crick) that only nucleic acids possess this
peculiar property of conserving information and using it to guide synthesis.

Isolated DNA has been shown by physicochemical methods and by
electron microscopy to be an extremely long molecule (several microns)
with a molecular weight of several million. A combination of X-ray
crystallography and chemical analysis shows that it consists of two helically-
intertwined chains (Watson and Crick, 1953) which are complementary in
shape to each other. If the two unwind and separate each might serve as a
" template " for the assembly of a new complement or for the formation
of other nucleic acids (RNA) as " copies." Current literature abounds
with hypothetical schemes for this replication process. In the chromosome

110

KERATIN AND KERATINIZ ATION

the nucleic acid is probably combined with basic proteins to form a
nucleoprotein, probably also a helix with the polypeptide chain closely
linked to the DNA helix (Wilkins et al, 1959).

Except in very general outline it must be admitted that nucleo-cyto-
plasmic relations remain very obscure. A current hypothesis is illustrated

Fig. 49. Principle cytological features of a cell synthesizing protein for
secretion. N, nucleus; Nu, nucleolus; m, mitochondrion; R, reticulum
(basophilic); P, an RNA-containing particle; G, Golgi apparatus; S,
secretory granule. The arrows A and B symbolize the exchange of
control between cell and environment.

in Fig. 49. Working copies in the form of RNA molecules are formed from
segments of the DNA molecules (master copies) of the chromosomes. The
gene is here pictured as a copiable segment of the DNA thread. These
RNA molecules fold up with protein into corpuscles and accumulate first
as a nucleolus (perhaps in this phase some nuclear synthesis occurs): they
then pass through the nuclear membrane either to join with a cytoplasmic
membrane to form part of the basophilic reticulum of a secretory cell

DIFFERENTIATION AND PROTEIN SYNTHESIS 111

or to form small clusters in a cell which retains its products (see
later).

The demand for a particular product of synthesis by other cells, or its
repression by cells already in production (Rose, 1952) which may be the
factor determining the course of differentiation by stabilizing a pattern of
synthesis, is symbolized by the feed-back arrows at A (environment to
cytoplasm) and B (cytoplasm outwards). The presence of soluble unused
products in the cytoplasm may ultimately suppress the production of the
RNA copies required for their formation: used products (i.e. secreted or
rendered insoluble or " wrapped up " in the cell) will not provoke this
kind of feed-back against themselves.

Epidermal cells may be freer of this kind of control than other cells (p.
146); their characteristic differentiated products seem to be produced
because they possess free surfaces which face the environment. The
various surface induced responses have been mentioned (p. 37), but it
must be admitted we have little idea of how the results are produced.

We approach firmer ground when the problem of RNA participation in
cytoplasmic synthesis is considered. The greater cytological detail made
possible by electron microscopy has, on the whole, supported and ex-
tended the views of earlier microscopists.

When thin sectioning for electron microscopy first developed, glandular
cells were immediately examined (see Haguenau, 1958, for review). The
RNA-containing nucleolus and the basophilic deposits against the nuclear
membrane appeared granular, the nuclear membrane itself double-
layered. A more important finding was that the basophilic areas of the
cytoplasm contained a system of membranes covered with small dense
particles of diameter 120-200 A (Plate 10A). Palade (1955) surveyed a
large number of cell-types and established the widespread distribution of
these particles. Essentially similar findings were reported by Sjostrand and
Hanzon (1954) and Bernhard (Bemhard et al, 1951; see also Haguenau,
1958).

Most of the authors referred to above, recognized at once that the system
of membranes-plus-particles must be related to protein synthesis and to be
the origin of the cell fragments called microsomes. Palade and Siekevitz
(1956) gave a clear proof of this by showing that on mechanical disin-
tegration of liver cells the membrane system broke down into microsomes
known to be rich in RNA and lipid and which consisted of smaller particle-
studded vesicles (Plate 10B). They further showed that after removal of
much lipid, the remaining material richer in RNA consisted largely of the
dense particles. Their participation in protein synthesis has been further
demonstrated by Zamecnik et al. (1956) and Simkin (1959) (Simkin and
Work, 1958) who showed that the particles were common in preparations
which most actively incorporated amino acids (see below).

112 KERATJN AND KERATINIZATION

The internal structure of the microsome particles is not known.
They contain both proteins and RNA but their X-ray diffraction patterns
show little resemblance to the patterns given either by RNA or by mixtures
of protein and RNA, which can only mean that the RNA is bound into the
particles in a form different from what it assumes when free. This may be
contrasted with the fact that the structure of isolated DNA is similar to its
structure in vivo where it also exists in a DNA-protein complex. The
microsomal particles resemble viruses in size, composition and in some
respects function (it has been suggested that viruses may be microsomal
particles gone wrong!), and the protein moiety dominates the structural
picture in some plant viruses.

After fixation by freeze-drying, which may be considered to introduce a
minimum of chemical change and a maximum of retention of material, the
particles are less visible among the protein adjacent to the membranes
(Hanzon et ai, 1959). This suggests that in their dense, compact form, as
seen after osmium fixation, the particles are an artifact. Their original
form may be a more diffuse and extended particle, a condition which might
render their function as linear templates more understandable.

Some details of a possible means by which the basophilic material, now
identified as the granular, RNA-containing material seen in micrographs,
might reach the cytoplasm were revealed by Watson (1954) who drew
attention to the existence of small circular markings (diameter 500 A) on
the double-layered nuclear envelope. He suggested these markings were
openings or " pores " in the double membrane which might permit
nucleocytoplasmic interchanges. The particles may pass through the
pores and, after associating with the external sheath of the nuclear mem-
brane, enter the cytoplasm together with it to form a typical particle-
studded membrane (Fig. 49).

The above description applies to glandular cells, i.e. to cells which
produce protein for secretion. The account given above of the cells of the
epidermis shows that protein synthesis may be associated with a different
type of cytology from that of secretory cells, as was first clearly pointed out
by Birbeck and Mercer (1957). The cells of the hair cortex, for example,
contain vast numbers of dense particles of the same size and appearance as
those noted in the pancreas and elsewhere ; but they are not associated with
a membrane system (Fig. 36) (Plate 11). They appear to be scattered in
small, often well-defined clusters, throughout the cytoplasmic space. The
cytoplasm is uniformly basophilic in these cells due to RNA, and there is
no reason to doubt that here, too, the particles contain RNA and participate
in the synthesis (see also p. 120). The cytoplasmic distribution of the two
kinds of nucleic acid in the hair follicle is admirably demonstrated in
Hardy's work illustrated in Fig. 90, p. 220. In the lower bulb, the germinal
region, the DNA of the nuclei is obvious; at higher levels where

DIFFERENTIATION AND PROTEIN SYNTHESIS

113

cytoplasmic synthesis is active, the diffuse cytoplasmic RNA increases.
(See also Braun-Falco (1958).)

The nuclear membrane of epidermal cells also is double-layered and
exhibits pores (Plate 4C). There is perhaps a difference in the mechanism
by which the RNA particles reach the cytoplasm since clearly they are not

Type of cell

Disposal of
protein

Lifetime of
synthetic phase

Cytoplasmic*
features

Examples

Secreting

passed through
cell membrane
usually as a
granule S

long, periodic

mitochondria m
Golgi cluster G,
reticulum R and
bound particles P

Pancreatic and
thyroid cells, etc.,
silk gland cells

Retaining

retained within
cell (/)

short, single phase

mitochondria m
Golgi cluster G
little reticulum, many
dense free particles P

Epidermal cells,
myogenic cells,
early embryonic
cells, anaplastic
tumours

Fig.

50. Comparative characteristics of cells forming protein (*see also
Fig. 49).

shed along with the external nuclear membrane to form a particle-studded
cytoplasmic membrane as pictured for glandular cells. They may simply
diffuse through the pores which are larger than particles or they may
leave in the small vesicles V to be seen in Plate 9.

There are many other cell-types with these same characteristics, and some,
along with examples of cells of glandular cells are mentioned in Fig. 50. It
will be seen that two classes of protein-forming cells are to be distinguished

114 KERATIN AND KERATINIZATION

by : (a) their differences in fine cytology and (b) by the manner in which
they dispose of their synthesized protein. In one case the protein simply
accumulates within the cell until, perhaps by a kind of mass action or
a simple physical exhaustion of space, synthesis is arrested. They appear
capable only of a single burst of activity. We shall term such cells retaining
cells (Fig. 50). In the other case, the product is discharged from the cell
after which another cycle of activity and secretion is initiated. Such cells
may continue their synthetic activities for a more or less prolonged period.
(Birbeck and Mercer, 1961).

Secreting cells possess the more complex cytology since, in addition to
their often elaborately-developed membrane system, they may possess
specialized devices for the temporary accumulation, transfer and for the
removal of protein from the cell. Since both types of cell synthesize
proteins, the capacity to do this must reside in their common structural
feature, the RNA particles of the cytoplasm. The membrane system is
therefore secondary and appears to be associated with the prolonged
activity and removal of secretion. Obviously for a cell to be able to
produce many times its weight of secretion, some elaboration of structure
is required to facilitate the entrance, transport and removal of material. An
organization analagous to the production line of a factory might be ex-
pected, and in fact, such cells are typically polarized. One portion of their
surface, the basal region, usually adjacent to the sources of raw material
(blood vessels) becomes specialized as an input area; the opposite aspect,
the apex, where secretion granules may collect and which usually abuts a
storage space or lumen, is the output area. These requirements are the
basis of the familiar histological pattern of glandular cells. Internally, as
Porter has emphasized (1954), a system of interconnected membranes is
admirably suited for the channelled diffusion of metabolites to and from
active sites and also for the collection of the products of synthesis.

According to Porter and Palade the total cytoplasmic membrane system
should be regarded as a unit, a definite cell organelle which may assume a
variety of forms : flattened interconnected sacs, cisternae, canaliculi and
isolated vesicles, for which the name endoplasmic reticulum is proposed.
The qualifying adjective " endoplasmic " was suggested by observations
on whole cells in tissue culture (Porter, 1954). Since subsequent obser-
vations on cells in sections have shown that membranes may spread
throughout the cytoplasm, the simpler name " reticulum " may finally be
adopted as more accurate. For nomenclature, see Fig. 23, p. 48.

Such a system of membranes enormously increases the internal surface
available as sites for catalysts and is, in fact, precisely the sort of " cyto-
skeleton " long demanded by biochemists (Peters, 1937) as a structural
support for an organized array of enzymes. The mitochondrion provides a
similar, even more compact, bundle of membrane-supported sites.

Plate 10 (Captions overleaf)

£°8 *3>\^

LIBRARY j=*j

Plate 10

A. Portion of cytoplasm of a rat pancreas cell. A typical secreting cell
(p. 110). R, reticulum of particle covered membranes; P, are RNP
particles; M, cell membranes; m, mitochondrion.

B. Microsomes from rat liver cells homogenized and fractionated by
centrifugation. Note many small, particle-covered vesicles derived from
the break-up of more extensively-developed, particle-covered membranes
in the original cells. V, vesicles, and P, particles.

Micrographs kindly supplied by M. S. C. Birbeck.

Plate 11

Portion of the cytoplasm of a cell in the bulb of the hair follicle before
filaments of keratin have commenced to accumulate. The picture is
typical of a retaining cell. There are small numbers of mitochondria,
(m), small vesicles, V, not obviously organized as a definite Golgi
apparatus, and vast numbers of dense (RNP) particles P which are
apparently free in the cytoplasm and not associated with membranes
(Fig. 50, p. 112).

Plate 12

A. Portion of a typical epidermal cell forming a hard keratin, in this case
from the feather follicle, showing the edge of a nucleus (l.h.s.), the
double nuclear membrane, NM, mitochondria, m, and/, fibrils of fibrous
keratin. The nucleus contains short lengths of filaments and the cyto-
plasm contains large numbers of dense particles P, not associated with
membranes.

B. Fibrils, F, of fibrous keratin in the bulb region of the human-hair
follicle. Each fibril is seen to be composed of fine filaments,/. Note the
dense particles, P, which crowd the cytoplasm.

C. A number of desmosomes distributed along a very convoluted contact
between two cells in the germinal layer of rat skin where the desmosomes
are not yet associated with cytoplasmic filaments. Each desmosome con-
sists of a paired thickening of the plasma membranes backed in each cell
by a dense amorphous deposit.

DIFFERENTIATION AND PROTEIN SYNTHESIS 115

Retaining cells do not possess a reticulum in association with RNA
particles, but numerous smaller vesicles with smooth-surfaced membranes,
i.e. membranes not associated with dense particles, are observed (Plate 11).
In secreting cells such vesicles also occur, often with an elongated flattened
profile, in compact clusters located adjacent to and distal to the nucleus
(Fig. 49). This location is that of the Golgi apparatus of the classical
histologists, who have long disputed its structure and function (Baker,
1955); very probably this group of smooth membranes is the electron
microscopic image of the Golgi apparatus. A certain unity of organization
and coherence is suggested by the fact that the clusters can be separated
from homogenized cells (Dalton and Felix, 1956), but the greater struct-
ural detail revealed by electron microscopy has not finally clarified their
function. No connexion with protein synthesis has been demonstrated,
though the opinion prevails that they are associated with the secretory
phase and probably it is the cell centre for membrane assembly when these
are required for special functions.

Biochemistry of Protein Synthesis

The combination of the results given by the various forms of microscopy
has yielded a sufficiently-detailed and usable picture of the intracellular
structures associated with at least one pathway of protein synthesis.
Unfortunately, in the present state of histochemistry, it is impossible to
investigate the composition and activity of the cellular organelles with the
same degree of resolution. For most biochemical work, a considerable
weight of any cell derivative is required and this has led to the need to
separate from large numbers of cells a sufficient weight of well-defined,
selected cellular constituents, such as nuclei or mitochondria for analysis.
The most commonly-employed procedure introduced by Claude (1938) is
that of homogenizing and fractioning a mass of cells, i.e. of rupturing the
cell membranes mechanically and fractionating the homogenate by
centrifugation. Most of the work on mammalian tissues has been carried
out on the liver, since this large organ is easily homogenized, and nothing
comparable has been attempted with an epidermal tissue.

When liver is homogenized mechanically and the product suspended in
a sucrose solution (0-25-O88M) four fractions are conventionally recog-
nized in a fractional centrifugation: (a) a nuclear fraction, (b) a mito-
chondrial fraction, (c) a microsome fraction, and (d) a supernatant. Such
fractions form the basis of most biochemical studies on the activity of the
intracellular elements recognized microscopically. The fractions are by no
means pure, nor do they consist necessarily of single components. It is
preferable to control their composition by electron-microscopical methods
(e.g. Palade and Siekevitz, 1956).

A study of the mitochondrial fraction has shown that many of the

116 KERATIN AND KERATINIZATION

enzymes associated with the oxidative degradation of sugars, etc., are
located here. The internal organization of the mitochondrion (Fig. 22a,
p. 45) seems admirably designed as a site for the organized array of these
enzymes which seem necessary to effect the sequence of reactions envisaged
in multi-enzyme reactions.

The mitochondrion is also a seat for the synthesis of low molecular
weight key substances such as adenosine triphosphate (ATP) which,
because of the so-called " high energy bonds " they contain, are able to
effect many biochemical reactions demanding an expenditure of energy.
Probably among these reactions is the formation of the peptide bonds
linking amino acids in the polypeptide chains in proteins.

Fischer's original suggestion that proteins were polypeptides containing
peptide bonds formed by the reaction :

RltCH(NH2).COOH + R2.CH(NH2)COOH

= R^HfNH^.CO.NHCHfR^COOH + H20

has been amply confirmed both by degradative and synthetic methods.
The free energy necessary for the synthesis of a peptide bond has been
determined from thermodynamic data on reactants yielding peptides or
compounds containing peptide links. It lies in the range of 2000-4000 cal/
mole. From the corresponding equilibrium constant, it may be calculated
that this value requires in equilibrium 99% of the material on the side of
hydrolysis. Obviously energy must be introduced into the system for
synthesis to approach completion.

A method of moving the equilibrium in the direction of synthesis by
selecting reactants which would yield insoluble products was devised by
Bergmann and Fraenkel-Conrat (1937). In the presence of a hydrolytic
enzyme such reactions will move in the direction of synthesis as the
product is removed from solution. At the present time a synthesis
catalysed by proteolytic enzymes is not thought probable. Nevertheless,
in the formation of an insoluble protein such as keratin, the reaction could
conceivably be assisted by the removal of the product in the form of
insoluble fibrils, etc. That peptide-bond synthesis requires energy has
been repeatedly emphasized on biochemical grounds (Borsook, 1955).
Siekevitz (1952) showed that the uptake of " tagged " amino acids in
homogenates is more closely linked to phosphorylation than to direct
oxidation. Zamecnik and Keller (1954) showed that incorporation of
amino acids into microsomes proceeds only in the presence of an ATP-
generating source. In the cell ATP generation is a function of the mito-
chondria and they are thus shown to be indirectly involved in peptide
formation. Borsook (1955) showed that for each peptide bond formed one
molecule of ATP is broken down.

Nevertheless the amount of energy involved is not large in comparison

DIFFERENTIATION AND PROTEIN SYNTHESIS 117

with that required for certain forms of mechanical work, e.g. that required
by muscles or for transferring substances across membranes against a
concentration gradient, and in fact one finds that the number of mito-
chondria in muscle cells or in the proximal tubular cells of the kidney is
much greater than that in protein-forming cells.

The observation that ATP is implicated in peptide-bond formation
suggests that an activated amino acid is probably the intermediate in the
synthesis. Haogland et al. (1957) have obtained evidence of enzymes
which could effect activation. Enzymes which catalyse carboxyl activation
of at least two amino acids (tryptophane and tyrosine) are now known, and
probably there is an enzyme for each amino acid, i.e. according to Crick
(1958) twenty in all. The product of the reaction between the amino acid,
the activating enzyme and ATP is an amino acid-adenosine mono-
phosphate anhydride and it is supposed that compounds of this type are
intermediates (see Fig. 51).

The reasonable expectation that the long peptide chains would be pro-
ceeded by the formation of short peptide sequences seems to be negatived
by several experiments. Small peptides are rarely found and ingested pep-
tides cannot be directly utilized, but are first broken down to amino acids.
There is, however, some confusing evidence on this point which cannot be
considered finally settled.

Many efforts have been spent tracing the fate of labelled amino acids
when these are injected into animals or fed to micro-organisms. The
newly-formed protein is customarily isolated as a trichloracetic acid-
soluble fraction from homogenized cells. If the liver, for example, is
examined very soon after the administration of a labelled acid, the isotope
is found predominately in the microsome fraction. Siekevitz (1952) found
that incorporation of radioactive amino acids also occurred in cell-free
homogenates and that the radioactive label was again predominantly in the
microsome fraction. Further, a microsome preparation is effective when
cell sap is added, or if even a partially-purified mixture or activating
enzymes providing an energy source (ATP or GTP) is also added.

The work of Zamecnik et al. (1956) shows further that the amino acid is
activated by the specific enzymes and ATP before incorporation and it is
probable that the activated acids are first transferred to a soluble RNA
(S-RNA). The transfer of this amino acid from soluble RNA to the
RNA particles requires the presence of guanine triphosphate. The
microsomal particles contain protein and the labelled acids are linked by
true peptide bonds. These steps are summarized in Fig. 51, p. 118.

Other experiments have shown that the material bound to microsomal
particles can be dissociated from it and appears in new protein. Rabinovitz
and Olson (1957) prepared reticulocytes containing radioactive leucine and
incubated the microsomes isolated from these with fresh sap. In the

118 KERATIN AND KERATINIZATION

presence of ATP the radioactivity was partially transferred to the haemo-
globin fraction. Hendler (1957), and Simkin and Work (1958) have
reported similar findings. Fig. 51, adapted from Stephenson et al. (1959),
summarizes the several stages of protein synthesis as outlined above.

1. Activation of Amino Acids

AA + ATP + E ^± [AA ~ AMP] - E + PP

2. Addition of Nucleotide End Group to S-RNA

- RNA + ATP + CTP «± [RNA - pCpCpA] + PP

3. Binding of Amino Acid to this RNA

AA + ATP + [RNA - pCpCpA] ^ [RNA - pCpCpA - AA] + PP

4. Transfer of Amino Acids to RNP Particles, Binding to RNA of
Particles Followed by Polymerization of Amino Acids

[RNA - pCpCpA - AA] + ATP + GTP + RNP

(Particles) + Soluble fraction (?)
— > [polypeptide chain — particle]

5. Release of Bound Polypeptide and Completion of Protein by Cross-
linking and Secondary Bonding

AA amino acids

E enzyme (many specific enzymes known)
PP pyrophosphate
CTP Cytosine triphosphate
S-RNA " soluble " RNA
pCpCpA nucleotide end group

] units linked as complex

Fig. 51. A current scheme showing possible steps in protein synthesis.

A special problem, arising in cells which are secreting their formed
protein, is the relation of the microsomal RNA particles to the system of
membranes found in these cells (see Fig. 49). Probably the microsomal
protein is released to the membranes and accumulates within them (Palade
and Siekevitz, 1956) as can often be noted in micrographs. In keratin-
forming cells and other retaining cells, membranes are not involved
and the newly-formed protein appears to be released directly into the cell
sap (Plates 9 and 11).

differentiation and protein synthesis 119

Secondary and Tertiary Structures

While a partial answer to the problem of peptide formation seems in
sight, a more formidable problem yet to be solved is that of specificity, i.e.
the formation of precise sequences of amino acids in the polypeptide chain
and the folding of these chains into equally precise three-dimensional
structures. Whether keratin, a protein whose function is a more or less
passive mechanical one, will actually prove to have a specificity as clearly
defined as an enzyme has not yet been proved by an actual determination
of an amino acid sequence. It is, however, antigenic (Pillemer et al., 1939
and 1938) and in well-crystalline forms (feather and porcupine quill) it
yields an X-ray pattern suggesting a complexity not inferior to that of
soluble proteins (Chapter 5).

Following a suggestion by Haurowitz and by others, it seems reasonable
to think that the final assembly of amino acids takes place in two steps:
(1) the formation of the definite polypeptide sequence (or sequences) on a
template and (2) the folding of the polypeptide to form a three-dimensional
molecule (Fig. 51, Steps 4 and 5). The attack on this problem is at the
moment largely speculative. Most writers assume that the sequence of
bases along a nucleic-acid helix somehow ultimately determines the
sequence of amino acids, and attempts to solve the problem, ranging from
biochemical experiments to abstract considerations based on coding theory,
are being made. The general feeling is that the microsomal RNA is the
most likely candidate for a template on which to assemble the amino acids
in the correct order and Crick has advanced further arguments to show
that an " adaptor molecule " is also necessary to hold the activated amino
acid on the template. In the absence of experimental evidence it is not
easy to carry this discussion further, but reference may be made to Crick's
article (1958).

The problem of the second step, the folding of the long polypeptide
chain into a specific configuration and the overall shaping of the molecule,
has been illuminated experimentally by the work on the lability of protein
configurations in solution and the dependence of both synthetic and
natural polypeptides on the interaction between the solvent and the chain.
This work will be returned to later (p. 194). What is important is that this
step seems to require no enzymatic or nucleic acid intervention; it
depends simply on the energy relations of the interactions between side
chains of the macromolecule and the molecules of the solvent or other
associated molecules Since the side-chain composition of a polypeptide
chain is determined in the primary act of synthesis (assembly on a tem-
plate) the configuration ultimately assumed by the molecule in a given
medium (cell sap) will be determined at the same time and by the same
means.

120 KERATIN AND KERATINIZATION

Experimentally, the fact that no special intracellular apparatus is
necessary to induce the polypeptide chains to adopt the specific foldings
which are responsible for the characteristic fibre patterns, is evident since
both proteins and synthetic polypeptides (Bamford et al., 1949, 1956) form
the structures spontaneously in solutions in vitro. Newer methods of
demonstrating the presence of specific structures in solution (p. 194) have
shown that only portions of the chains may adopt the folded form. This
probably also occurs in vivo and, surviving into the solid state, is probably
one of the sources of the non-crystalline fraction (Chapter 5).

Synthesis in Retaining Systems

Most of the evidence described above relating to protein synthesis has
been obtained from secretory cells; retaining cells such as epidermal cells
have been less studied. Their cytoplasm is, however, as rich in RNA and,
electron-microscopically, it is crowded with dense particles (Plates 7, 9,
11 and 12A and B) apparently identical with the RNA-containing particles
of secretory cells. It is perhaps desirable to mention the experimental
evidence which shows that in cells with this cytological pattern that
synthesis also involves RNA and up to a point is identical with the process
in secretory cells. Reticulocytes, which synthesize and accumulate
haemoglobin becoming erythrocytes, are retaining cells and in them the
course of synthesis seems to follow the lines indicated above. Bacterial
cells, although less easily classified, resemble in some respects the retaining
cells of higher organisms and in them protein synthesis also follows the
same course (Loftfield, 1957).

The same general picture of diffuse cytoplasmic basophilia in the light
microscope and of vast numbers of dense particles with a scanty develop-
ment of membranes, when seen in the electron microscope, is found in
the cells of certain anaplastic tumours. These again lend themselves
more easily to biochemical study and make it possible to confirm that
synthesis is associated with the particulate elements of the cytoplasm.
Littlefield and Keller (1957) for example, using ascites tumour cells
showed that the most rapid incorporation of radioactive leucine occurred
in the cell fraction containing the dense, ribonucleic acid-containing
particles.

There is then no reason to doubt that protein synthesis in retaining cells
in general follow the same course as in the more commonly studied secreting
cells. But whereas in secretory or glandular type cells protein synthesis
continues more or less indefinitely in a succession of cycles, the products
being (continuously or on demand) removed from the cell, in accumu-
lating cells there is a single phase, the product is retained within the cell
and synthesis comes to a halt as the cell fills, possibly as a result of mass
action or the simultaneous production of an inhibitor. Such cells then

DIFFERENTIATION AND PROTEIN SYNTHESIS 121

enter on a more-or-less prolonged life in which their activity is of a
different order. They perform some specialized function based upon the
properties and behaviour of their accumulated products, e.g. muscle cells
accumulate actomyosin and their subsequent functional behaviour is
based on the contractile properties of this substance; the reticulocyte
becomes the mature erythrocyte filled with haemoglobin and enters the
blood stream as an oxygen carrier; the cells of the stratum corneum harden
and form a protective layer to the organism. In this phase their RNA
content (dense particles) falls.

Synthesis in Fibre-forming Systems

No comparable biochemical studies have been made on cells whose
function is the formation of protein fibres. Nevertheless, a survey of the
cytology of such cells reveals their essential similarity to that of cells which
produce soluble, non-fibrous proteins (Mercer, 1958). The same dis-
tinction exists between cell systems secreting a precursor of the fibre,
which then forms extracellularly, and systems in which the fibres form
intracellular^. The cells secreting fibres have as before elaborately-
developed particle-studded membranes of which perhaps the most
developed examples are the cells of the silk gland of the silk worm (Bombyx
mori) (Mercer, 1958). The cells of the hair cortex are, of course, the
typical retaining cells. The dense particles found in all fibre-forming cells
have the same range of dimensions 120-200 A as those found in cells
forming soluble proteins, and may be identified on the same grounds as
particles containing RNA. Histochemical tests confirm this picture (see
p. 220).

Further, no special cytological features are found to be associated with
the production of fibres having the different basic types of molecular
structure indicated by X-ray analysis of fibres. The a-type proteins occur
intracellularly as keratin or myosin (Astbury, 1947) in retaining cells, and
extracellularly as fibrinogen (Bailey et ai, 1943). The secreted fibres
(collagen-type) and fibroin (|8-type), originate in cells having the familiar
pattern of membranes and particles of the secreting cell. Thus there
would seem every reason again to associate the RNA system only with the
original link-up of amino acid in polypeptide chains and to consider that
the specific folding determined by this sequence occurs after the release
from the RNA particle. The subtle, genetically-determined differences
between an RNA particle producing a collagen sequence and one pro-
ducing an a-keratin sequence remain unknown.

The novel feature about a cell, which produces a protein capable of
forming a fibre, is not then to be sought in the basic machinery used to
effect the synthesis which is common to all, but rather in the genetically-
determined instructions or information associated with the RNA particle.

122 KERATIN AND KERATINIZATION

These instructions determine the sequence of amino acids and this
sequence determines the sort of intermolecular interaction which follows.
Some types of sequence produce strong intramolecular chain association
and a corpuscular type molecule with weak forces of intermolecular
association; these may form proteins soluble as discrete particles. Other
sequences may lead to the stronger form of witeraiolecular association we
recognize as fibrous. The several means by which protein fibres are
synthesized are classified in Fig. 52.

micromolecules
(amino acids)

I*
soluble microprecursors
(activated amino acids)

condensed micromolecules

(polypeptide associated
with RNP particle)

I *
shed macromolecule
(non-fibrous precursor)

retaining-type cells

I
immediate intracellular
condensation in fibrous
form (fibrous hair keratin)

secretory-type
cells

temporary accumulation
of precursor followed
by a transformation into
the fibrous form
(keratohyalin)

extracellular conversion
into the fibrous form
(fibroin)

Fig. 52. Classification of types of protein fibre synthesis.
(*See also Figure 51.)

Epidermal cells provide examples of most of these methods of forming
fibres. In the hair cortex and in the feather, filaments seem to appear
directly without the accumulation of substantial amounts of macro-
molecular precursor ; in the inner root sheath of the hair follicle trichohy-
alin accumulates as a precursor and undergoes a transformation into
fibrils. Skin cells may employ both methods at different times (p. 228).
The hair-cuticle cells are peculiar in that the keratin appears as amorphous
droplets and condenses without changing into fibrils. These cells remain
non-birefringent and yield no defined X-ray pattern. Among birds,

DIFFERENTIATION AND PROTEIN SYNTHESIS 123

keratin may even be secreted. The cells lining the gizzard produce a horny
lining of keratin (p. 108) and the oviduct covers the egg with a mixed
mucin and keratin layer which forms the tough membranes immediately
underlying the shell (Plate 18B).

The supermolecular organization of fibrous tissues
(tertiary structure)

Macromolecular Fibrous Texture

We have up to this point been considering those aspects of the synthesis
of fibrous proteins which they share in common with the soluble proteins.
However, the special characteristic of fibres is the supermolecular aggre-
gates which they form whose dimensions may extend from the molecular
to the histological level and beyond. These aggregates often display a
remarkable and intricate structure at several levels and it is a further
problem to give an account of the genesis of this larger-scale organization.

Fibres usually perform a mechanical function; i.e. they transmit
tension, strengthen membranes or provide against impact, and their
organization within tissues is normally related to these functions. In fact
some of the most striking examples of biological adaptation are provided by
the fibrous tissues. It is more useful in this connexion to speak of a
fibrous texture rather than of fibres and thus to focus attention on the
microscopic rather than the macroscopic elements of structure.

In a purely geometrical sense we may list the possible varieties of
fibrous texture as in Fig. 53.

Organisms exploit all these possible arrangements and it is not usually
difficult to relate the textures to the mechanical function. The difficulty
is to account for the origin of the structure in molecular terms, particularly
in instances where it appears apparently in advance of function.

When studying a complex organism in its structural or functional
aspects, it is convenient to distinguish a hierarchy of levels of organiza-
tion. At each level new possibilities of organization are introduced
and the study of the structure as a whole is facilitated if the events at each
of the several levels can be considered separately. Fibrous tissues lend
themselves to this form of analysis. They exhibit, as Astbury has put it,
" patterns within patterns". In this sense we shall distinguish here three
levels: the molecular, the macromolecular and the microscopic level, which
happen also to be those which correspond to structures of orders of
magnitude most easily studied by X-ray crystallography, electron micro-
scopy and light microscopy respectively (Fig. 1) ; but, although convenient,
there is no significance in this correspondence, which is in any case inexact
and not lasting, since the domains of X-ray diffraction analysis and electron
microscopy increasingly overlap. The structures existing at each level are

124

KERATIN AND KERATINIZATION

produced by causes which we might group together generally as directive
activities operative at that level and the structures themselves are a record
of these activities from which we may hope to infer something of their
nature.

Texture

Possible function

Example

Tangles (three-
dimensional)

plugging
holes

blood clots
(fibrin)

Sheets with
rodlets in one
plane and disordered
in other senses

^ protection

/■ — •/ against
v) \ impact, etc.

epidermis

Parallel arrangement
in one plane

as membranes

occurs only
transiently

Sheets of (quasi)
orthogonally-
arranged rodlets
(multilayered)

Sheets of

hexagonally-arranged

rodlets

Parallel rodlets
in three dimensions
(various arrangements
in cross-section)

protective
membranes,
retaining
basket (allows
movement by
distorting mesh)

membrane
rigid against
distortion

transmission of
•»,.. mechanical
■'■'" tension

Fig. 53. Types of fibrous texture.

earthworm
cuticle (Reed and
Rudall, 1948),
amphibian dermis
(Plate 14B)

peritrophic
membrane of
insects (Mercer
and Day, 1952)

hairs, tendons,
muscles, etc.

An apt analogy is often made between fibres, yarns and fabrics on the
one hand, and molecules, fibrils and tissues on the other. In this vein we
may speak of the spinning and weaving of molecular yarns and fabrics
from the raw material provided by the primary synthesis at the molecular
level. This primary synthesis we have already discussed, our problem here
is the spinning and weaving of the yarns and fabrics which we shall term
fibrogenesis and fibrillar organization (Fig. 54). More precisely we wish to
consider firstly the formation of the long fibrils (or ribbons) which are
used as the constructional units of the tissue and, secondly, the forces
which organize (or weave) these fibrils into more complex formations and
stabilize them.

DIFFERENTIATION AND PROTEIN SYNTHESIS 125

In histological systems, unlike those found in the textile mill, " spinning
and weaving " usually occur together in time and space, i.e. the fibrils
appear precisely where and when they are required to enter into the
structure of the growing fabric. However, it may be possible to separate
the two processes artificially if the precursor of the fibrous system can be

Micromolecules
*

Macromolecules

Protofibrils
I

Fibrous textures

Stabilized tissues

Synthesis
Fibrogenesis
Organization
Stabilization

Fig. 54. Formation of secondary and tertiary fibrous structure
(*see also Fig. 51 for this step).

isolated, either before it has been converted into fibres or by reversing the
process of fibre formation. In this case it may be possible to study the
formation of fibrils in vitro and, by artificially orienting these in imitation of
the natural system, to gain some insight into the processes which must be
operating to orient them in vivo.

Fibrogenesis

Fibrogenesis (Fig. 54) is by definition the formation of an elementary
elongated unit from which the more complex formations are constructed.
An appropriate name for such a unit, which indicates that it is the first
fibrous unit in a hierarchy of such structures, is the protofibril. Electron
microscopy definitely established the existence of protofibrils in many
fibre tissues although their presence had earlier been inferred from X-ray
diffraction and polarized-light studies. Protofibrils form the structural
basis of some fibrous keratins; their appearance and condensation into
macrofibrils and more massive formations has been described in the hair and
feather follicle. Unfortunately, valuable though these morphological
observations are, in the case of keratin they cast little light on the physico-
chemical processes involved. Information concerning these processes
could certainly be obtained if it were possible to study the phenomenon

126 KERATIN AND KERATINIZATION

in vitro. To do this it would be necessary to isolate a precursor in a fonn
still capable of forming fibrils, either from the germinal tissues of the
epidermis or from the fibres themselves, by reversing the process of
fibrogenesis. Unfortunately, largely as a result of the chemical reactions
involved in keratinization (Chapter 6), it has proved impossible to redis-
solve keratin without gross modifications of its structure. Further, it has
also proved impossible to separate the unaltered precursor from the
germinal cells.

For these reasons our views on fibrogenesis in keratin must be based on
reasonable inferences drawn from a study of more tractable systems and
checked against a background of direct observation of cellular events.

Logically we may distinguish two methods of fibrogenesis : either (a) a
macromolecular precursor is formed first and subsequently aggregated to
form a fibril ; or (b) micro-units, e.g. amino acids are directly built into the
growing fibril and no macromolecular precursor is involved. In many
fibrous systems we know experimentally that (a) is the actual course
followed and it may well always be the case. No macromolecular pre-
cursor can be demonstrated for the fibrous hard keratins; but while
admitting that we do not know precisely how polypeptides are formed,
we know that it involves the participation of other large molecules, such as
RNA particles, and it is difficult on spatial grounds to see how these large
bodies — larger in diameter than the filaments themselves — can be brought
into position at the growing points of the filaments. Furthermore, no
close association of particles and fibrils is in fact observed in the cells of the
hair follicle (Plate 11). Thus it may be concluded that a soluble precursor
exists transiently. The considerable quantities of amorphous protein
demonstrable electron-microscopically after special staining may, in part,
represent this precursor.

In a general sense, we can anticipate that the nature of the aggregation
is likely to be much influenced by the shape of the precursor molecule,
" interaction profile " (Hodge, 1960), or by modifications in its shape
which accompany fibrogenesis. In some systems a more-or-less iso-
diametric molecule may simply aggregate without marked internal change,
and the process is then very similar to crystallization (see also Rees, 1951).
In others, preliminary modifications of structure precede aggregation as is
the case with the fibrinogen-fibrin system (Lorand, 1952; Lorand and
Middlebrook, 1952). Other unexpected, even bizarre events should not be
ruled out. For example, Rudall (1955-6), in attempting to trace out the
development of the fibrous ribbons of the egg case of a mantid, discovered
that lumps of precursor are first formed into vacuolated droplets which are
thinned, flattened and drawn out to yield the fibrous ribbons. It is
obviously necessary to treat each case as a special case, if this is at all
possible. We can only review very briefly a few examples of systems which

DIFFERENTIATION AND PROTEIN SYNTHESIS 127

have been extensively studied and draw what inferences we can about the
formation of keratin.

Collagen. By far the best understood case is that of collagen. As was
well demonstrated by Nageotte (1927), collagen is an admirable protein for
research on fibrogenesis since its solution in weak acids can very readily be
made to reform fibres. Today largely owing to the work of Schmitt and
Bear and their associates (1955 and 1960) and of Randall and Jackson
(1956) (Randall and Robinson, 1953) it has become the prototype model
for studies of morphogenesis at the macromolecular level. Not only may
it be readily dissolved and regenerated in fibrous form; it also yields,
when conditions are changed, a remarkable variety of fibrous fabrics some
of which are not found in nature (Hodge, 1960). The precursor is synthe-
sized by fibroblasts and secreted as a soluble molecule into the inter-
cellular space where it proceeds to form fibrils and fabrics which are
exquisitely adapted to the demands of the mechanical forces operative at
that point (see Plate 23B).

Soluble collagen, or tropocollagen, has been shown by light scattering
(Boedtker and Doty, 1956; Cohen, 1955) to have a long (3000 A) thin
(13-6 A) molecule composed of three helically coiled chains, strongly-
H-bonded, of molecular weight about 345,000 in Stainsby, 1958). This
long molecule is able to aggregate, principally by lateral adhesion, to
yield a variety of structures in addition to that normally found in native
collagen fibres and each is recognized electron-microscopically by its
banded structure. For recent summaries, see Bear (1952), Robinson
(1953), Randell et al. (1953) and Hodge (I960).

The possible arrangements of both intact and partially fragmented
molecules have been very fully worked out and are shown diagrammatically
in Fig. 55. It is characteristic of the collagen type of lateral aggregation of
long thin precursors that there should be large well-marked longitudinal
repeat spacings detectable by X-rays or electron microscopy, and an
absence of side spacings larger than that corresponding to the molecular
diameter. The existence, in the X-ray patterns (see Chapter 5) of well-
formed examples of fibrous keratin, of strong long-spacings on the
equator, i.e. side spacings (Tables 9 and 10, p. 167), and the absence of the
lower orders of the main longitudinal repeat pattern, suggests that the
collagen model is not immediately applicable to keratin. Electron-
microscopically, the keratin protofibrils are seen to be thicker (60 A) than
the collagen unit, and of indefinite length (at least 2000 A long), and no
marked longitudinal spacing is visible electron-microscopically, although
long spacings expressible as orders of a major spacing of 198 A (hair) and
98 A (feather) are found in X-ray diffraction patterns (Chapter 5).

The aggregation of filaments to form muscles seems in principle to be
very like that involved in collagen formation (Hodge, 1959 and 1960).

128 KERATIN AND KERATINIZATION

Fibrous insulin. An example of fibre formation from a more nearly
spherical type of molecule is provided by fibrous insulin (F-insulin) which
has been extensively investigated by Waugh (1954). Insulin molecules are
small and display a marked tendency to form small aggregates in solution.
When acid solutions are heated, the insulin separates out in the form of
long stiff fibrils (Farrant and Mercer, 1952); the suspension displays
strong birefringence and may gel. Oriented fibre-type X-ray patterns can
be obtained (Bear, 1955). The reaction is reversible in alkaline media and
soluble, biologically-active insulin may be regenerated. Since the insulin
molecule is cross-linked internally by disulphide bridges, it is unlikely to
be grossly distorted when entering the fibril. All these observations suggest
a simple aggregation of the insulin units.

NATIVE

Bill II ill Hi: ill ill

liHMlH|ill=//\ ^\

Fig. 55. Diagrammatic illustration of patterns of aggregation of tropo-

collagen macromolecules in native, FLS and SLS types. Polarization of

macromolecules indicated by arrow. (After Schmitt, 1958).

There is an important general geometrical principle, pointed out by
Crane (1950) that the structures which result from the successive addition
of asymmetrical units are always helices. The linear aggregate formed by
the addition of large molecules should conform to this principle, and
accordingly we may expect to find that many protofibrils are helices.
Pauling (1953) has described F-insulin as a helix of this type and it would
seem not unlikely that some of the coiled-coiled models proposed for
keratin, which appear as fine filaments in electron micrographs, have such
an origin (Chapter 5).

DIFFERENTIATION AND PROTEIN SYNTHESIS 129

Silk fibroin — aggregation of the molecule after unfolding. Another model
for fibrogenesis, suggested by its analog}' with spinning, supposes that
aggregation is preceded by an unfolding of the molecular chains composing
the precursor particle. This process seems probable in cases where fibre-
formation is produced by a mechanical process of extrusion and drawing,
as when artificial fibres are manufactured from viscous solutions of
cellulose derivatives or of dissolved keratin derivatives. Nevertheless, it
seems unlikely that it can occur generally in biological systems where
fibrils may appear and disappear reversibly with a slight change in
variables. Often a mechanical factor analogous to drawing is absent.
While drawing may orient fibrils once formed, i.e. in fibrillar organization,
it seems to play no role in fibrogenesis itself.

The formation of silk fibre (fibroin) by silkworms seems at first sight an
example of fibre formation by drawing. A soluble precursor of the fibrous
form is produced in cells, which are rich in RNA and contain an extra-
ordinary development of particles and membranes (Mercer, 1957); it is
stored as a strong viscous solution in a dilated portion of the silk gland,
and is converted into a thread by being extruded through a fine spinerette.
Nevertheless, if the contents of the silk gland are diluted with water and
allowed to stand for some hours, masses of fine fibrils separate spontane-
ously from the solution (Mercer, 1951c). Fibrogenesis thus again seems
to be a spontaneous phenomenon of aggregation requiring no mechanical
assistance, but the extrusion and drawing occurring during spinning are
responsible for the orientation of the protofibrils.

Organization of Fibrous Tissues

Assuming that the basic fibril has been formed, we have now to consider
the means by which this is used as a unit for the construction of higher-
ordered structures. The geometrical form of the structures is most easily
discovered with the electron microscope, although it may be deducible
from X-ray photographs or even with the light microscope. The problem
is to find the factors, mechanical or otherwise, which organize it. There is
often a relation between the mechanical function of the fibrous system and
its structure, which may provide clues.

Some examples of the possible arrangement of fine fibrils are shown
in Fig. 53. Tangled " brush heaps " arise in the absence of orienting
influences, such as in a fibrin clot. Fibres in which all the elementary
filaments are parallel have been subjected to an orienting influence either
during or after the formation of the filaments. The most obvious influence
is the shear due to flow, which probably initiates the orientation of silk
thread. Drawing after extrusion may improve the orientation.

Some of the more interesting structures are to be found among the
fibrous membranes. Collagen in skin, and certain cuticles are often found

130 KERATIN AND KERATINIZATION

arranged into parallel sets of fibrils forming approximately a right angle
with a similar set above and below it (Weiss and Ferris, 1954). The
ultimate orienting influence is clearly the surface of the animal, which
ensures that the whole formation lies parallel to it; the immediate con-
trolling influence is not so obvious (Plate 14B). In other instances the
geometry is even more complex and sometimes of surprising regularity.

Such sheets may be formed by exfoliation from a surface composed of
the cells that secrete the precursor. Here we can conceive of two kinds of
organizer: (a) a pattern or " die " on the cell surface or (b) the existing
pattern of the preceding sheet, which acts as a template for the assembly
of its successor. The peritrophic membrane lining the mid-gut of insects
could be an example of the first suggestion (Mercer and Day, 1952). Here
the secreting cells are covered with projecting microvilli (the brush border
of histology) with a cross-sectional diameter about the same size as the
holes in the membrane. We could imagine the filaments forming in the
grooves between the studs on the surface. Sections of the cell surface,
although showing the pattern of microvilli, and the layers of shed mem-
brane, have not yet provided an example of a membrane in the act of
formation; therefore decisive evidence is still wanting.

The collagen meshworks in skin, and in earthworms (Rudall and Reed,
1948) seem to form some distance from the cell surfaces, which are
covered with amorphous material. The " self-template " seems more
likely here. We can form a conception of how this could operate by
supposing that the upper surfaces of fibrils have " studs " on them which
fit into " holes " in other fibrils when these are laid across then at right
angles.

This discussion of fibrogenesis and organization has been limited to the
special case in which the elementary fibrous unit is a fine filament (or
ribbon). While this is applicable to many systems, we must expect that
other devices will be found such as Rudall has described in the secretion
of the colleterial gland of a mantid (p. 126).

Tactoids, familiar from their occurrence in tobacco mosaic virus (TM V)
solutions, have been proposed as fibre-forming elements (Bernal, 1940).
The rodlets of TMV are not unlike the protofibrils under consideration
here and it seems quite probable that, in some instances, e.g. the bundles
of fine filaments of keratin in hair cells, the same forces which maintain
tactoids are operating. Spherulites and sheaves, e.g. in F-insulin seem to
result from growth by aggregation from a single, or a group of centres in
the absence of external orienting forces. The filaments must also be
supposed to possess little lateral attraction.

As mentioned in Chapter I calcium salts may be deposited in associ-
ation with fibrils of collagen to form bone. The deposition seems to be
initiated at definite sites in the 640 A banded collagen fibril and the earliest

DIFFERENTIATION AND PROTEIN SYNTHESIS 131

crystals are randomly oriented (Jackson, 1954). As they grow an orientation
develops apparently directed by the oriented fibrillar matrix in which the
crystal forms. Only the naturally-occurring type of fibril with the 640 A
period seems able to initiate crystal deposition (Bachra et al., 1959).

Epidermal fibrils

Unfortunately it has not yet proved possible to obtain a soluble
precursor of keratin or keratohyalin which will produce fibrils spon-
taneously in vitro. Our information concerning fibrogenesis is thus limitep
to what can be obtained from the microscopy of the tissues themselves
combined with any applications of general principles we can infer from a
study of other fibre-forming systems such as those just described. From
what has been said it is evident that the formation of fibrous keratin has no
exact parallel in other systems. A complicating factor is that essentially
the same final system (compare Plates 16 and 17) appears to be arrived at
by two different courses : (a) in the epidermis partly through the formation
of a non-fibrous intermediate form, keratohyalin and (b) in the hard
keratins without the appearance of this intermediate form. The isolation
ol keratohyalin, its analysis and its behaviour in vitro would greatly help to
clear up this obscurity. Keratohyalin and trichohyalin after accumulating
as droplets of isotropic precursor are converted into the fibrous form
(Plate 21) in a manner which has about it something akin to crystalli-
zation. The orientation of the fibrous form of trichohyalin in the cells of
the inner root sheath of the hair follicle is strictly parallel to the axis of the
follicle. Here we may suspect that the slight shear affecting the cells of the
bulb as they approach the follicular constriction, which orientates the
elongated mitochondria and nuclei, also orientates the initial small
formations of fibrous trichohyalin and thus directs the subsequent massive
transformation. In the epidermis the transformed, fibrous keratohyalin of
the stratum lucidum runs approximately parallel to the stratum corneum,
at right angles to the prevailing fibrillar orientation in the germinal layer.
The cells at this level are already somewhat flattened and probably here
too the shear produced during the change in cell shape controls the
direction of orientation (Plate 22).

Many fine filaments are seen attached to desmosomes in the cells of the
lower layers, where they seem to provide suitable sites for initiating
fibrogenesis, and this attachment may help, by holding on to one end of a
tuft of filaments, to orient it when the cell is deformed (Charles and
Smiddy, 1957).

Rather less can be asserted about the origin of orientation in hair,
feather, horn and nails. No precursor accumulates and filaments, when
they appear, are already oriented. The particulate contents of the cells,
long nuclei and mitochondria, seem to be oriented by the flow due to cell

132 KERATIN AND KERATINIZATION

deformation in the hair follicle and it is in this polarized matrix that the
first filaments are assembled. Tentatively it would seem that the earliest-
formed filaments are oriented by the same flow and that the orientation of
subsequent deposits is determined in turn by these " seeds".

The opinion sometimes expressed that the narrowing neck of the follicle
acts like a spinerette, is certainly not correct except in the sense that it
supplies the initial orientation to the " seeds". Thereafter fibrillar growth
itself is oriented.

CHAPTER IV

The Growth of Epidermal Structures

The epidermis as a growing organ

The growth of epidermal structures is a subject of interest and im-
portance in itself; however, it gains a wider importance since, for reason
of the ease with which superficial changes can be observed, the epidermis
is often the tissue chosen for the investigation of the factors governing
growth in general. Obviously an account of such an enormous subject
would be impossible here. Nevertheless, the peculiar suitability of the
epidermal structures for these studies and the probability of their future
use, makes a limited discussion of some points desirable.

The sum total of the epidermis and its appendages constitutes an organ
of a quite definite morphological and functional character, as well defined
as that of the internal organs. Its pattern, in considerable detail, is charac-
teristic of the species, with usually a male and female variation under the
supplementary control of the sex hormones. The fact that it is a superficial
organ, with much easily-observed structural detail some of which, such as
feather or hair, is amenable to quantitative evaluation, is the reason for its
use by taxonomists, geneticists, experimental physiologists and others.
Its pathology provides valuable signs and symptoms of less-readily
observed disorders. The feather, hair or nail is in fact a permanent record
in chronological order of the synthetic events which led to its formation.

In the adult, the cells in most organs divide infrequently; growth has
practically ceased, and the residual divisions are probably those required
to make good " wear and tear." In certain situations, however, e.g in the
seminal vesicles, the intestinal mucosa, the bone marrow and in the epi-
dermis itself (Leblond and Storey, 1956), cell loss is a normal physiological
process and cell division continues as part of the normal activity of the
tissue. In exposed situations superficial cells are simply shed or scraped
off and obviously their loss must be made good. The entire range of
epidermal derivatives is maintained by the proliferation of the cells of the
continuous germinal layer underlying the whole system, a population of
apparently-uniform and interchangeable cells (p. 57). Thus the problem
of the growth of the whole formation resolves itself into the question of
what factors control cell division in the germinal layer and what determines
the course of differentiation of the cells after leaving this layer.

133

134 KERATIN AND KERATINIZATION

There is no difference here between the coherent hard keratins and the
soft, which spontaneously exfoliate. Both varieties are subject to wear
and more-or-less continuous growth is required for replacement. Nails,
claws and epidermal horns seem, however, to be continuously produced
irrespective of demand and their growth in excess of needs may even
become a nuisance. Feathers and hairs on the other hand have a quite
distinct unity and grow to a defined shape. Plucking is followed by re-
growth, but only in a remote sense can we speak of this as a renewal in
response to wear.

In the epidermis a steady state normally prevails in which cell loss is
balanced by cell replacement. The renewal time at any site is defined as the
time taken for the replacement of an amount of material equal to the
amount present in the layers above that site. It is also the time taken for a
cell to pass from the germinal layer to the surface where it is shed, and is
analogous to the growth period for hairs and feathers. Since growth may
not be continuous over short periods of time (see p. 135) the steady state
is only an average state maintained over a more-or-less extended period.
Leblond and others have determined the renewal time for a number of
proliferating tissues. Some examples are given in Table 6. The methods
employed are based on the direct counting of nuclei in division over an
extended period of time or on determining the number of nuclei arrested
in metaphase by colchicine in this time.

The existence of a definite equilibrium thickness of the epidermis
differing from site to site and of the definite shapes of feathers and hairs
shows at once that some sort of overall control must exist throughout the
epidermal system. In this respect the epidermis is no different from other
organs whose forms and cellular composition are also strictly controlled.
In fact the entire cellular community constantly maintains a state of
homeostasis in which its numbers and composition are kept in balance
with each other and with the environment.

The factors likely to affect the growth of a tissue have been sought both
by direct observation of normal growth and from the results of experi-
mental interference. In this way a large amount of information has been
gathered concerning the growth of whole animals and organs, which
although often of immediate practical value, is not easily related back to
the activity of the individual cells. A number of growth factors and of
hormones influencing growth are known, but their effects are invariably
complex when whole tissues are considered. For reviews, see Thomas
(1956).

Undoubtedly many hormones also affect the behaviour of epidermal
cells, but their action is complex and far from clearly defined. Oestrogens
definitely stimulate cell division according to Bullough (1953). The cyclic
changes of the vaginal epithelium, the cells of which oscillate mucin

THE GROWTH OT EPIDERMAL STRUCTURES

135

production and keratin production, are under the control of the sex
hormones (p. 144). The effects of various hormones on hair growth have
been described by Mohn (1958) and on feather by Lillie (1942).

The most obvious effect of the sex hormones is on hair and feathers.
The action of the male hormone appears to affect directly the length of the
growth period of certain follicles so that longer and stouter hairs (or

Table 6. Some Renewal Times for Epidermal Tissues.

Tissue

Animal

Renewal time
(days)

ear (Malpighian layer)*

mouse

abdomen (Malpighian layer)

human

100

forearm (Malpighian layer)

human

hypothmar (Malpighian layer

human

30-36

plus corneum)

foot pad (Malpighian layer)

guinea pig

40-50

plus corneum)

foot pad (Malpighian layer)

rat

19-1

* Taken from Leblond and Storey (1951 and 1956).

Tissue

Animal

Renewal time
(days)

skinf

guinea pig

ear

guinea pig

143

metatarsal pad

guinea pig

prepuce

guinea pig

tongue

guinea pig

8-4

t From Piatt (1960), other figures will be found in Hooper (1956),
Price (1958), Meyer et al. (1960) and Scheving (1959).

feathers) are produced, although the type may also be affected. That other
quite extraneous substances may act as stimulants is shown, for example,
by the marked effect of scarlet fever toxin (Heyningen, 1950).

Mitosis in the basal layer

Although very little that is not hypothetical can be said about the control
of the overall patterns of growth, there is better experimental evidence
concerning the mechanisms of short period fluctuations.

While the average rate of cell replacement in the epidermis is relatively
steady, Bullough has established the presence of diurnal cycles in the

136 KERATIN AND KERATINIZATION

mouse, and these probably occur elsewhere. The maximum epidermal
mitotic activity occurs during sleep and the minimum during muscular
exercise. Among humans the maximum of mitosis occurs at night, again
the period of rest.

Epidermal cells need a supply of energy for mitosis and division, for the
synthesis of their specialized products and for keratinization when this
occurs. Carbohydrates are the main source of energy and these are
probably supplied as glucose and stored as glycogen. Glycogen is not
found in the germinal layers, but may occur in the prickle cell layers
(Bradfield, 1951) and is stored in quantity in the outer root sheath of the
hair. The energy of the glucose probably becomes available in anoerobic
glycolysis through the agency of the tricarboxylic Krebs acid cyclic
(Bullough and Johnson, 1951 ; Bullough, 1952). Many of the intermediate
substrates of the Krebs cycle can be utilized by skin and have been found
in hair roots (Bullough, 1958). Rothman (1954) believes there may be other
pathways specific to skin. Bullough (1952) found that many of the inter-
mediate substrates of the Krebs cycle will support cell division, and was
thus led to suppose that mitosis required energy and could only occur
when the cells are able to absorb adequate amounts of carbohydrates and
oxygen. That is, the special necessities for mitosis are stored in some form
during antephase and are syphoned off when division commences. Were
this the case muscular activity may well lead to short supplies in the
epidermis and delay preparations for division. The diurnal cyclic activity
of the epidermal cells is thus seen as an indirect consequence of the cylic
muscular activity induced by diurnal fluctuations of light. The cycle is
absent in skin cultivated in vitro, thus clearly demonstrating its dependence
on extracellular factors. More recently Bullough and Laurence appear to
have abandoned this opinion (Bullough and Laurence, 1958). Further,
having found that, in the skin of starved rats in which the number of
mitoses is very much reduced there is a dramatic burst to several times the
normal number when skin is removed and transferred to saline (in absence
of oxygen and glucose), Bullough and Laurence (1961) conclude that
epidermal cells are always able to complete preparations for division, but
that some factor inhibits the process in early prophase. They give reasons
for believing that it is the high adrenalin levels associated with muscular
activity which inhibit these cells. Thus the epidermal rhythm is linked
with the rhythmic changes in adrenal activity (see p. 144).

Bullough's general conclusion, that glucose and its subsequent con-
version to yield energy is a critical factor controlling mitosis, has not been
accepted without question. In a series of papers Gelfant (1958, 1959a,
1959b) has confirmed the participation of glucose, but insists that an
adequate supply of glucose and oxygen alone will not stimulate mitosis in
intact mouse ear epidermis; the mitogenic factor may be the cutting. This

THE GROWTH OF EPIDERMAL STRUCTURES 137

finding would not necessarily clash with Bullough's present views in-
volving control of inhibitors (p. 149). However, Gelfant's work does imply
that some of Bullough's experiments were carried out in conditions that
were " sub-optimal " for mitosis and that his conclusions may not be valid
under the optimal conditions which may be assumed to prevail in vivo.

Cycles with a 24 hr period, ultimately linked to the diurnal fluctuations
in illumination (Reinberg and Ghata, 1957) are common (see p. 146). For
example in animals, the body temperature, a measure of muscular activity,
the glucose concentration of the blood, the concentrations of water,
glycogen, fat and protein of the liver all show such variations. The calci-
fication of teeth, a dermoepidermal function, is also cyclic.

There is no diurnal cycle in the hair follicle of Rodents— the only case
examined. To account for this relative immunity of the hair follicle from
the fluctuations caused by alternations of rest and activity, it is assumed
that the follicle has its own independent source of food supply and, in fact,
large amounts of glycogen are found in the cells of the outer root sheath
(Montagna, 1956; Hardy, 1952). Glycogen is reduced in amount or is
absent when there is no hair growth (Montagna, 1956; Montagna et al.
1952). Possibly when required, the glycogen is mobilized as glucose and
transported by the network of blood vessels surrounding the shaft of the
follicle to the bulb. In support of this it may be noted that the vascular
network of the follicle of growing hairs is remarkably developed (Durward
and Rudall, 1949 and 1958; Ryder, 1956). Ryder (1958) injected
radioactive glucose into mice and found that in 1 hr there was isotope in
the bulb, and also in the outer sheath where it increased over the next 24
hrs. The rapid uptake in the bulb could be due to the glucose which
provides energy for mitosis and the slower accumulation in the sheath to
the storage of glycogen.

No such detailed information exists concerning the other long-growing,
hard keratins, nails, claws, etc., but it is not unlikely that their continuous
growth is sustained in a similar way.

The question of the mitotic rate and location of mitoses in the epidermis
has occasioned much discussion which has to some extent been cleared up
by the realization that there are diurnal variations in the rate in the skins of
rats, mice and humans, which provided the bulk of the material (see also
p. 146). Further, on hair-bearing skins subject to cyclic variations in hair
growth, the activity of the epidermis is linked to that of the adjacent hair
follicles. In the mouse and rat, the skin thickens in the early phases of hair
growth and relapses again before hair growth ceases.

The mitotic rate is not the only factor which determines the thickness of
the total epidermis. Clearly this depends on the renewal time, the time a
cell takes to reach the surface and be shed. Ebling (1954) showed, for
example, that oestradiol while increasing the number of mitoses four

138 KERATIN AND KERATINIZATION

times actually decreased the total thickness of the whole layer. The rate
of differentiation and of exfoliation, about which less is known, influences
the thickness of the intermediate layers and of the horny layer, respectively.

General theories of growth

More recently attempts have been made to refer the problem of growth
to a more fundamental basis by examining directly the behavioural
patterns of the individual cells. These investigations take two forms: one
is experimental, the direct observation of cellular activity in tissue culture —
some of this work has been already referred to — the other is theoretical
and attempts to develop an adequate general theory, which would relate
the growth of the whole organism to the growth rates of the constituent
cells and, further, would be competent to infer the characteristic stability
and homeostasis from the properties and interaction of these cells. At the
present time only tentative solutions of this problem have been made, but
it seems worthwhile to mention them here, not only in an attempt to
inquire how far epidermal growth may be included within the scope of a
more general discussion but also because it seems that the future develop-
ment and experimental verification of these theories will involve further
experiments on the epidermis.

If the initial special events of cleavage and the blocking out of the early
embryo are omitted, and growth considered only after the point where the
embryo increases in weight, the bald facts demanding an explanation
according to Weiss and Kavenau (1957) are:

(a) The increase in number of cells.

(b) Their divergence into organ systems containing differentiated cells.

(c) The quantitative facts of the growth curve, i.e. the sigmoid shape of
the plot of total weight of cell mass against time (Fig. 56).

(d) The steady state which the organisms approach as the adult size
is achieved in which the various organs exist in equilibrium with each
other and with the external environment. In this state the several organs
have become specialized in the functions which they perform on behalf of
each other. The equilibrium is dynamic in the sense that it is maintained
by a constant intercellular communication to which considerations of a
cybernetical order are applicable.

Observations made on cells isolated and cultivated in the absence of
other cells show that in these conditions cells gradually cease to produce
their characteristic products and assume a more generalized character.
Further, when mixed cell populations are grown together differentiation
again takes place. These very general findings are sufficient to prove that
differentiation is maintained by restraints exerted by one cell type on
another, either by direct contact or by exchange of their products through
the medium of their common humoral pool (p. 61).

THE GROWTH OF EPIDERMAL STRUCTURES

139

Such a community of interest functioning in terms of an economy of
supply and demand might well maintain a constant ratio of cell-types while
permitting an unlimited growth of the total population. However, the fact
is ((c) above), that in all organisms the adult size is a rather well-defined
limit and the growth curve follows a characteristic sigmoid course (Fig. 56).

Fig. 56. The sigmoid growth curve. N is the number of individuals

(cells). The population N early passes through a phase of approximately

exponential growth, later a logistic law is a better fit.

It is evidently necessary to suppose that a further control mechanism
exists which maintains this size and is responsible for the shape of the
growth curve.

That many natural populations follow what is referred to as the logistic
law:

= eN-hN*

(Where N = number of individuals, e and h are constants) has long been
known. A population obeying such a law tends towards a limit W =
e/h. When h is negligible, dN/dt = e N and the population increases
exponentially, N = N0ea. This type of increase may be observed in
cellular populations during a limited period (the " log phase ") when
conditions are favourable, but sooner or later limiting factors appear, the
growth rate declines and a logistic law is more applicable. The problem is
usually to identify the limiting factors.

The sigmoid shape of the growth curve can be simulated in a formal
sense by a number of physicochemical models, e.g. by the autocatalytic
monomolecular reaction, by systems which make demands in proportion
to their mass (L3) and are able to accumulate (or lose) in proportion to
their surface areas (L2). Such physicochemical models have a broad

140 KERATIN AND KERATINIZATION

general validity (Rashevsky, 1948), but it is one of the consequences of
multicellularity that they cannot be applied in a simple direct form.

It seems more probable to many that each organ system itself produces
changes which automatically lead to its limiting its own proliferation. Such
a view is in harmony with modern theories of self-controlled mechanisms
which envisage control, in very general terms, as being effected by a " feed-
back " of information which introduces a limiting factor proportional to
the deviation from a norm.

Physiological evidence that the entire population of cells in an organ
controls its own size is obtained from a variety of experiments in which
part of the population is removed. Partial hepatectomy is followed by a
burst of mitotic activity in the remaining tissue leading to a restoration in
size. That the influence causing the mitotic wave is carried by the blood
is shown by experiments in which hepatectomy is practised on one of two
rats whose blood supplies have been joined (parabiotic union). Mitosis
occurs in the second undamaged liver. These effects might be ascribed to
a stimulating substance (wound hormone), but the loss of an inhibitor is
indicated by the observations that mitosis can be induced in normal livers
if the blood is diluted by saline, and that regeneration itself is slowed up by
increasing the plasma concentration (Glinos, 1958). Weiss (1955) has
reviewed experiments in which the removal of one member of a paired
organ, e.g. the kidney, is followed by an increase in size of the remaining
member. Perhaps the best demonstration of the existence of control by
inhibitor productions is found in the experiments of Bullough on growth
control in the epidermis itself which will be described in the next section.

Certainly numerous other factors, among them well-recognized hor-
mones, affect growth as is made clear in the reviews of Abercrombie (1957
and 1958) and Swann (1955, 1957 and 1958). Nevertheless, the possibility
exists that primarily control is based on hormones of the inhibitor-type and
that other hormones could operate by secondarily affecting the cells'
response to inhibition. These problems are returned to again below
(p. 146).

The logistic growth curve is obtained from the exponential-type curve
simply by the addition of a further negative term (see above) which here
might be regarded as the " negative feed-back " term. The total growth
curve may be considered as the sum of the separate organ sigmoids.
Several proposals of this sort have been made, for example, by Morales
and Kreautzer (1945) and by Sock and Morales (1945). Weiss (1955) has
attempted to give these concepts a more definitely-biological basis and
recently with Kavenau (1957) has obtained solutions of a growth equation
which are sufficiently precise to be put to a quantitative test. They suppose
that each specific cell reproduces itself by a mechanism in which certain key
compounds act as catalysts (templates). The growth rate is proportional

THE GROWTH OF EPIDERMAL STRUCTURES

141

to the concentration of these intracellular templates which constitute the
generative mass. This mass is being constantly converted into a non-
reproductive mass which is the differentiated product. These reactions are
further supposed to be accompanied by the formation of inhibitors (anti-
templates) which block the templates, and thus may potentially act as

HUTfll£NTj{~

l0m&sffS&.

Wkkl J

?30L/C loss

Fig. 57. Illustrating Weiss and Kavenau's model (1957) for a cell
showing control of growth by the production of an inhibitor. The cell
contents are divided into generative mass G and its product the differ-
entiated mass D. In this example the inhibitor I is supposed to be
produced by D and its action is fed back to control the processes of
growth and reproduction (reproduced by permission).

growth regulators by an intracellular negative feed-back (Fig. 57). In
order to effect a control over the whole distribution of a cell type, Weiss
and Kavenau assume (a) that the inhibitor molecules diffusing from the
cells enter the common humoral pool and thus reach other cells, (b) carry
a " tag " or " label " enabling them to be recognized by other cells of the
same type and (c) that they have a normal rate of degradation or loss
which, in equilibrium, balances their rate of production (Fig. 58).

These assumptions are biologically acceptable and sufficiently general
and simple to permit of mathematical expression. It is clear, without
attempting to formulate and solve the growth equations, that since they
contain negative feed-back terms they will lead to a system which will
automatically regulate its own size. Weiss and Kavenau set up differential
equations for their system and applied a solution of these to the case of the
growth of a chicken. The quantitative agreement between theoretical and
experimental data is surprisingly close.

One feature of systems stabilized by feed-back should be pointed out.
Since a delay in time occurs between the despatch of a signal from one part

142 KERATIN AND KERATINIZATION

of a system to the part controlled by the signal, the possibility of oscillations
arises (p. 143). Weiss and Kavenau believe these will arise, for example,
when organs regenerate after partial removal. These may be manifested
also at the cellular level and possibly account for some of the fluctuations
in cellular activity commonly noted (see also p. 148).

Fig. 58. Further aspects of Weiss' theory of intercellular control by the
exchange in inhibitor molecules. Two kinds of differentiated cells are
distinguished by their differentiated products, open circles and triangles.
Two kinds of inhibitor molecules, black circles and triangles, are released
into the common humoral pool. The rate of production of the two kinds
of products is controlled specifically by concentration of their specific
products in the pool (reproduced by permission).

In its present state of development the theory does not give a place to
interactions between different cells, which are probably in part effected by
the exchange of samples of the differentiated mass, and thus does not
attempt to account for the appearance of differentiation and its main-
tenance. A differentiating system can be devised, following Rose (1952) if
one supposes that the inception of a certain reaction in one group of cells

THE GROWTH OF EPIDERMAL STRUCTURES 143

suppresses this development in adjacent cells and permits a second
reaction to arise in these cells. The products of the first reaction diffusing
from the " dominant " group of cells thus " induce " a second and
different reaction in neighbours. Possible fine structural evidence of this
form of induction in the epidermis and dermis has been described above
(p. 90). It would not be difficult to generalize the growth equations by
introducing terms expressing the interaction between cells assuming that
the anti-templates (or secreted differentiating mass) in a dominant early-
maturing group of cells can suppress similar development in less-advanced
cells. See also Waddington (1948).

Also, no necessary place has been given to the fact that tissues are
organized in a cellular form. Certainly, although cell division introduces a
discontinuity in the output of a single cell, these irregularities would be
smoothed out when the output of a large non-synchronously-dividing
population is considered. Moreover, if adequate arrangements exist for
transport to and from sites of synthesis, the cellular habit does not in itself
seem essential for continued synthesis. For example, in insects relatively
enormous differentiated cells are common. Probably cell division is an
inherited act, originally developed to permit of replication and dissemi-
nation of the genetical apparatus, that occurs normally when the DNA is
duplicated and the cell has synthesized adequate amounts of the materials
required to provide the apparatus of division. These latter activities could
involve paths of synthesis distinct from those involved in the formation of
specialized products.

That all authorities do not yet accept the necessity of control by in-
hibitor production is evident from a recent discussion on the growth of
proliferating tissues (Price, 1958). Obviously the possibility that stability
is maintained throughout multicellular organisms by the circulation of
inhibitors is a conception of far-reaching consequences. It implies, in
effect, the existence of a whole system of hormones which has escaped
notice. The already-known hormones and other growth influencing agents
would seem to effect the sensitivity of the cells to the circulating inhibitors
or act to influence the dispersal and disappearance of these. It is highly
desirable that an attempt be made to isolate these postulated inhibitors and
that their mode of action on cells be determined. It is evident that the
theory must be regarded as unproven until some of the postulated in-
hibiting substances have been isolated and the mode of action on the cells
observed directly (Bertalanffy, 1960).

Periodic growth and cyclic activity

Even in the adult when the size has become more or less constant,
certain organs undergo a periodic fluctuation in size and activity. Con-
spicuous among these are the sexual organs and with their changes are

144

KERATIN AND KERATINIZATION

linked equally-marked changes in the entire endocrine system (Burrows,
1949; Bullough, 1951). The often striking periodic changes in the plum-
age of birds and the hairy covering of mammals are clearly linked both to
the sexual and seasonal cycles. In exhibiting these changes it seems likely
that the integument responds to the general endocrine situation, that is to
say, its periodicity arises indirectly from periodic changes in the concen-
tration of circulating hormones.

cm*

Fig. 59. The production of two cells of contrasted type from the same

germinal layer: upper l.h.s. a keratinizing cell, upper r.h.s. a mucin

forming cell. The metaplastic change may be effected by hormonic means

or by such additions as vitamin A (p. 63).

Obviously cyclic activity, which ranges from short period oscillations,
such as the heart beat, to the slow oscillations, which gear organisms to
the daily and annual changes in their physical environment, must be
regarded as one of the most important biological phenomena. The cyclic
changes may be reflected not only in size, but in cell function which is
revealed in cyclic histological and cytological changes (metaplasia). For
example, in the not uncommon metaplastic cycle between a mucin-
producing and a keratinizing epithelium, the same germinal layer gives
rise to two types of cells of contrasted cytology: (a) keratinizing cells
(l.h.s. Fig. 59) with many RNP granules and a poorly-developed system

THE GROWTH OF EPIDERMAL STRUCTURES 145

of cytomembranes; and (b) mucous cells with an elaborate system of
y-cytomembranes and few RNP particles (r.h.s. Fig. 59) (Burgos and
Wislocki, 1958; Nilsson, 1959; Schulz et al., 1958). Cyclic behaviour
of this character is in effect a cyclic change in differentiation. Like
differentiation in the more stable sense, it is a consequence of the inter-
action between one group of cells (organ) upon another. It was precisely
this interaction which was omitted for simplicity in the simplified theory
of growth outlined above (p. 138). When interaction is permitted, periodic
phenomena can arise spontaneously from the effects of intercellular
communication when there is a time lag between the transmission of a stimulus
from one organ and the return of a counter stimulus from the other. Such
questions are currently discussed under the heading of cybernetics.

t + T

Fig. 60. Illustrating the possibility of generating cyclic activity by a feed-
back link between two organs A and B. A signal from A stimulates the
development of B whose secretion returns with phase lag T to repress
the activity of A (see text).

The possibility of oscillations can be understood most easily from a
consideration of the ideal situation depicted in Fig. 60. We suppose two
organs A and B. The secretion SA of organ A stimulates the growth of
organ B, whose secretion SB is capable of inhibiting the activity of A.
Suppose the secretion SA reaches a threshold value at a time t and initi-
ates a growth phase in B, which after a further time T leads to the release
of SB by the organ B. This inhibitor SB in turn now acts on and suppresses
the action of A ; with the consequent fall in ^4's activity, the stimulus to
B falls off and its activity declines again, leading to a fall in the inhibitor
SB and the recommencement of activity in A. Thus cyclic activity is
set-up both in A and B with a difference in phase introduced by the time
lag T (Fig. 60).

In the language of cybernetics, we say that the two organs are controlled
by a feed-back linkage, and that the oscillation is made possible because
the feed-back signal from B to A is out of phase with that from A to B.
The feed-back from B to A is negative; that from A to B is positive.

As described, the feed-back signals (SA and SB) are " hormones " and
the lag between them is introduced by the time of maturation of a target
organ B. This illustration was chosen because it corresponds, in the type

146 KERATIN AND KERATINIZATION

of change and in the times likely to be involved (hours or days), with the
actual endocrine changes with which epidermal changes are linked. But
the argument is general : A and B may be parts of the same organ, SA
and SB may be nervous or mechanical signals and T may be very short.
Actual situations are never as simple as that of Fig. 60. Usually several
cyclic systems interact and at some point a " sensitive " element responsive
to the environment may introduce a signal which " gears " the entire
system to the diurnal and annual cycles.

To explain this further it is necessary to introduce another important
idea, that of adaptive oscillations. The frequency of the cycle A^±B (Fig.
60) is capable of great variation because of variations in the time of
maturation of the cells in B (and/or A). Thus when the cycle A ^± B is
linked to others C^*D, E^±F, etc., the possibility of coupled oscillations
with resonance arises, because the frequencies of the separate systems can
change until all have the same frequency (or multiples of this) when they
will resonate at this frequency with phase differences determined by the
time lags in the various feed-back loops. It is possible to say that the
appropriate matching frequencies evolve from the possible range of
frequencies by a kind of natural selection — the best-adapted frequency
survives and increases its amplitude. This mechanism in essentials was pro-
posed by Pringle (1951) in developing a theory of the activity of the brain.

By such a feed-back train a system of cycles could be " geared " to the
diurnal and annual astronomical cycles and an organism's total activity
adapted to the physical environment. In these complex events the in-
tegument seems, as far as is known, to follow the lead of the endocrine
system. It is a target organ of graded sensitivity, but there is little proven
evidence of its returning a control stimulus to the deeper tissues.
However, this may seem so largely because of our ignorance; the possi-
bility certainly exists that epidermal products can be fed back into the
organism, thus constituting yet another closed cycle. For example, the
grooming habits of both birds and mammals are so persistent that con-
siderable quantities of epidermal material must re-enter the organism
through the mouth (see p. 59). Further, from the wider viewpoint of the
enormous system of communication which constitutes ecology (Hutchin-
son, 1948), the integrative function of the integument as a signalling
system to predators, to congeners, and to sexual partners can only be
mentioned here, emphasized and left.

Control of epidermal growth

When we come to consider the epidermis in terms of such general
theories, we see at once, as has been emphasized already, that because of
its position on the outside of the cell system, it constitutes a special case
among the organs. Moreover, it is non-vascular, unevenly enervated

THE GROWTH OF EPIDERMAL STRUCTURES 147

(Arthur and Shelley, 1959) and its cells grow outwards. Its cells have
evidently only a limited possibility of communicating with each other
and the rest of the system. The existence of " fleece mosaics " in sheep
has recently assumed some importance as a proof of local autonomy. A
fleece is said to be mosaic when it comprises two distinct types of wool
grown on different areas of the same skin. The areas may be adjacent and
it is clear that the follicles concerned, although producing different types of
fibre, enjoy the same environment externally and internally, thus demon-
strating beyond doubt the over-riding control of local, non-systemic
factors on the kind and quantity of fibre formed. It is assumed that
mosaics arise from a somatic mutation, i.e. from a mutation occurring in a
cell subsequent to the first division of the egg which initiated development.
The change will be apparent only in the line of cells issuing from the
mutated cell. Thus an area of skin, producing aberrant type of wool, is
assumed to be a colony of cells arising from a single cell in which a somatic
mutation has occurred. The actual local histological factors have not yet
been fully explored, but the existence of the phenomenon proves the
genetic control of localized epidermal structures which in turn determine
the nature of the product quite independently of systemic factors.

There is in the epidermis a vertical integration but clearly only a
rather limited lateral one. By postulating the same intracellular
features, generative mass, differentiated mass (keratin, mucin, etc.) and
inhibitor production, the possibilities of control in a simple stratified
epithelium may be considered. Division is largely confined to the basal
layer and synthesis of specialized products takes place in more distal
layers. The inhibitor molecules, produced in the stream of outwardly-
moving cells during these later reactions, are largely lost when these cells
are shed, and can only feed-back to the germinal layer by back diffusion,
and only by crossing the dermoepidermal junction can they reach the
general circulation and thus be carried to distant parts of the system.

Certain possibilities may be made clear by considering a cell which has
just been produced by division. Since division has occurred we may
suppose that inhibition is minimal. The cell leaves the germinal layer and,
at a higher level, begins to differentiate, to synthesize both the differen-
tiated product and also the inhibitor. The following conditions may arise :

(a) Before sufficient inhibitor is produced or diffuses back, the cell in the
germinal layer again divides. This would be a condition permitting of
uncontrolled, continuous growth.

(b) Before division can occur again, sufficient inhibitor diffuses back to
prevent it. After a further time the concentration falls again (with the
decrease in synthetic activity in the differentiated layer and the decay of
inhibitor molecules) and division again occurs. We have here a condition
of periodic division under the control of the events in the differentiated

148

KERATIN AND KERATINIZATION

layer. The time elapsing between divisions will depend on the rate of
synthesis (i.e. also the rate of formation of inhibitor) in the differentiated
layers and the rate of loss or decay of inhibitor at the germinal level. Since
a certain lateral diffusion from any cell is possible, a synchrony could
develop in the germinal layers owing to the overlapping of the effects of
several adjacent cells. This synchronization may be favoured also by the

■• • ' ' ' ••'■"/ wound V ' ' ' -^

hummi!i^iii\iiiiL

THUD

rmnnm

wound

Fig. 61. Diagrams to show the regions of high epidermal mitotic activity
expected in undamaged ear epidermis opposite to an area 3 mm square
from which the epidermis and superficial dermis have been removed on
the assumption (a) that a stimulating " wound hormone " is produced by
damaged epidermis, and (b) that the concentration of epidermal inhibitor
is reduced in the neighbourhood of a wound (from Bullough and Laurence,
1960).

possibility that a general systemic stimulation reaching a group of cells, in
which the inhibitor concentration is already low, may cause them to enter
division together. The diurnal variations in mitotic activity noted by
Bullough, may be an instance of this type of control (p. 136).

(c) Stimulation of growth may be caused by activities which facilitate
the fall in inhibitor concentration. Experimentally, Pinkus (1951) found an
increased mitotic rate following the stripping off of the upper layers of the
skin with adhesive tape. Mechanical trauma, such as piercing or cutting

THE GROWTH OF EPIDERMAL STRUCTURES 149

the skin, are known to be sufficient to provoke regrowth in quiescent areas
of skin in rabbits and rodents (Slen, 1958). In animals such as sheep with
constantly-growing follicles, an increased blood supply has been thought
to increase the rate of growth of wool (Ferguson et al., 1949). Perhaps one
of the most characteristic properties of the epidermis, its adaptive response
to the effects of hard work, may be produced by the dissipation of inhibitor
resulting from friction and pressure. Thorium-X plaster acts similarly.

The epidermis in health fits its bearer snugly. Obviously some control
adjusts lateral growth so that the area of the covering increases or de-
creases with the volume contained. Possibly the factor here is mechanical,
a tension or compression arising from the expansion or contraction in
volume of the tissues beneath. Stretching would thin the covering layers
and reduce the amount of inhibitor diffusing back to the germinal cells
which would then divide and replace the overlying layers.

Recently by means of an ingenious experiment, Bullough and Laurence
(1960) claim to have shown that inhibition rather than stimulation is
the real growth controlling factor. Following a wound, mitosis and
growth are initiated in a limited area (approx. 1 mm wide) of the epidermis
surrounding the wound. This could be described, and usually was,
as the result of the liberation of a " wound hormone " stimulating
growth. Using the fact that the skin on a mouse's ear is less than 1 mm
thick, Bullough removed the epidermis on one face making a wound more
than 1 mm wide (Fig. 61) and observed the effect on the epidermal layer
on the other side of the ear. If the diffusion of a stimulant to a radius of the
order of 1 mm was the stimulating factor, a limited area of mitoses opposite
the edges of the wounds should be seen. If the removal of inhibitor was
the cause, mitoses should be seen over the whole area lying beneath the
wound, which was the condition actually observed (Fig. 61).

Competition

Another factor, which almost certainly plays a part in controlling
growth rates and through them morphogenesis, is competition between
cells and organs and parts of organs for some essential requirement for cell
growth. Here again data obtained from the observation of epidermal
growth have been extensively used to demonstrate the actual operation of
competition. We have mentioned that a competition between the pri-
mordia of follicles may account for their appearance in a hexagonal
pattern on certain skin areas (p. 77) and that inter-organ competition
within a " scale-hair-gland unit " may be appealed to for an explanation of
the suppression of some structures in favour of others.

The very extensive quantitative data concerning the growth rate and
dimensional properties of wool fibres available from Australian sources
have been used particularly by Fraser (Fraser and Short, 1960) in an

150 KERATIN AND KERATINIZATION

attempt to demonstrate quantitatively the effects of competition between
follicles, An interfollicular competition for nutrient substance is immedi-
ately suggested by the most striking feature which emerges from the com-
parative study of the fleeces of the many domesticated sheep : viz. the fact
that the denser the fleece (fibres per cm2) the finer and shorter are the
individual fibres. For quantitative data see Carter and Clarke (1957).
Fraser and Short (1960) have demonstrated also a negative correlation
between fibre size and the diameter and distance of adjacent fibres. See
also Ryder (1957). Fraser (1951) originally used the concept of com-
petition to explain the differences in the shape of the tip curl between
fibres formed by central and lateral primary follicles (Fig. 35), (p. 78). The
number of crimps in a periodic function of time (Norris, 1931, p. 156) and
a regular crimp of constant curvature will result if the growth rate is
constant. If the rate decreases, following on the initiation of adjacent new
follicles which compete for fibre forming substances, the curvature would
decrease and a sickle shaped tip would result, as is found on the primary
central fibre of a trio group.

Fraser's original proposals (1951) were clear-cut and offered plausible
suggestions relating growth rates, tip shapes and order of appearance of
wool fibres during development. The subsequent attempt to evaluate
these theories quantitatively by statistical analysis of the data has led to
some secondary elaboration which makes it difficult to test them ex-
haustively as once envisaged. Such concepts as : competition for space in
early development and variations in a genetically-determined " efficiency
of competition " on the one hand, and actual histological findings of
secondary follicles having different origins — directly from the epidermal
surface and by budding from existing follicles on the other — have intro-
duced complexities. Fraser and Short (1960) are now of the opinion that
more information concerning the performance of the individual follicle
must be obtained. In fact, Rudall (1956) (p. 73) has already shown that
it is the dimensional structure (diameter, surface area and height) of the
papilla which is most strongly correlated with the follicle output and the
dimensions of the fibre. Presumably it is these papillary dimensions which
are influenced directly by competition and other controlling factors.
Considerations of this sort, which promise to relate follicle output to the
activity of the actual cells covering the papillary surface by supplementing
the statistical approach, are likely to lead to a clearer understanding of the
growth process in general. See also Burns and Clarkson (1949).

Patterns of hair growth and control

Hair growth and replacement varies in different animals and is a species
characteristic; within a species the sex hormones in particular determine a
masculine and feminine pattern which disappears after gonadectomy. The

THE GROWTH OF EPIDERMAL STRUCTURES

151

same is true of birds' feathers. Since much is known about hair patterns
and the factors controlling them, we shall consider them in terms of the
theory developed above.

Three different types of follicular activity may be distinguished :
(a) Periodic activity with neighbouring follicles not in phase, i.e. each
follicle behaves independently with periods of growth followed by periods
of rest. This may be referred to as " mosaic growth " and probably
represents the basic pattern of isolated follicular activity.

Fig. 62. The growth waves on rat skin demonstrated by the dye-
absorption of growing follicles. Regeneration of hair at 3, 8, 30 and 35
days after shearing the area of A, B, C, D. Similarly-hatched areas are
those in which the hair is growing at the same time and in which the
alloxazine pigmentation would develop in response to administration of
the compound at about that time. (Reproduced with the kind permission
of Professor Haddow and the Editor of Nature).

(b) Continuous activity as found in sheep and on the human head. This
is a special case of (a) with exceptionally long growing phases and a short
resting phase.

(c) Periodic activity with neighbouring follicles in phase as is found in
rats and mice (Butcher, 1934; Haddow et al., 1945; Fraser and Nay,
1953). At any time most of the follicles are quiescent and activity is
confined to small areas which may form recognizable wave-fronts (Fig. 62).

Intensive studies have been made of hair growth in man, sheep, some
rodents and fur-bearing animals. Quasi-continuous activity (b) is rare:

152 KERATIN AND KERATINIZATION

in most animals hair of a well-defined species-characteristic, site-character-
istic length is produced during a growth phase and a more-or-less lengthy
resting phase follows (Chase, 1954 and 1955). Shedding may be continuous
or exhibit a seasonal dependence (Fig. 40, p. 89).

The pattern is essentially similar in the mouse and the rat. Important
dermal changes occurring (Durward and Rudall, 1949; Montagna, 1956,
Review) in correlation with the growth cycle are (a) an increased vascu-
larity beneath the growing area and (b) increased deposits of hypodermal
fat. It has been debated whether these changes cause or are caused by the
epidermal activity. Durward and Rudall proved the absence of nervous
control of the growth wave in the rat by severing all nervous connexions.
They demonstrated that the growth proceeded largely under local control
and independent of systemic control by exercising portions of skin
rotating them and grafting. The pattern of growth (and also that of
pigmentation) is quite unrelated to the distribution of nerves or blood
vessels. Butcher earlier had reached different conclusions but Durward
and Rudall believe that, by choosing to study young animals, he may have
failed to distinguish the characteristic behaviour of the mature skin.

Durward and Rudall conclude that, in very general terms, the growth
wave is a consequence of a " resting stage inertia and a stimulus provided
by neighbouring vascular activity." The vascular activity probably arises
in response to the demand of the cells in follicles already actively growing
and synthesizing keratin. Special modifying systemic factors such as
hormones or food supply (p. 135) certainly exist but they are assumed to
remain constant in these experiments.

Montagna and Chase, using the mouse, have reached the same con-
clusions. Chase (1954) in particular has tried to trace out in greater detail
the interconnexion of the various elements of the skin. He emphasizes the
morphological continuity of the basal layer of the epidermis, the external
root sheath and bulb of the hair follicle and the peripheral cells of the
sebaceous gland and their functional interdependence. These elements
act as a unit (the pilosebaceous unit) sometimes centred on a single follicle,
sometimes on a group of follicles and their associated glands as described
earlier (Fig. 63).

The control diagram, Fig. 63 taken from Chase and simplified, is an
attempt to bring out the morphological continuity and to indicate the
possible lines of communication which transmit the control from one unit
to another in the skin, and effect the integration of the whole.

Essential conditions required for prolonged growth are themselves
probably provided by the geometry of the follicle, a long thin cylinder
penetrating deeply into a potentially well-vasculated region, the dermis
and hypodermis, where growing cells may satisfy their demands for food
and where, at the same time the rate of accumulation of inhibiting

THE GROWTH OF EPIDERMAL STRUCTURES

153

molecules is reduced. In isolation from other follicles, a follicle might be
expected to continue in production until the accumulation of inhibiting
molecules reaches a critical concentration, when growth would cease. It
could recommence when the concentration fell again to a lower critical
threshold. Local histological pecularities probably control the rate of

Yield

Yield sloughed or outside
Periodic, not continuous, loss
Reversible process
Inhibition or neg feedback
Draining effect
Morphological continuity

fluence
Induction

Causing degeneration of
enclosed elements

Fig. 63. Chase's attempt to illustrate the factors which operate to
integrate the pilosebaceous unit (hair follicle plus sebaceous gland).

BE basal layer of epidermis.

PSC peripheral cells of sebaceous gland.

UEC upper, permanent outer root sheath.

DP dermal papilla.

CTS connective tissue sheath.

C corium.

A adipose layer.

accumulation and dissipation of the inhibitors ; and, as an important dis-
tinction from the general systemic control mechanism described above, it is
supposed that the inhibitor molecules which enter general circulation have
a short life and that, therefore, little control over distant sites exists. Since
there are no diurnal variations in growth rate, as is found in skin (p. 136)
we must assume an adequate supply of nutrition (p. 137).

Co-operative behaviour becomes possible according to Chase in a
population of follicles when the individuals are close enough {ca. 1 mm

154 KERATIN AND KERATINIZATION

apart) to share the same diffusion fields. It is in such conditions that
growth is transmitted in a wave-form. This view is essentially that of
Durward and Rudall if we equate their " resting stage inertia " with
" accumulation of inhibitor." In these terms the development of growth

Ht\\\\\\\\\\\

CROWING FOLLICLE

(b)

Fig. 64. (a) The generation of a growth wave in a population of follicles
whose growth is accompanied by the build-up of inhibitor molecules
(small dots) which on reaching a critical concentration cause growth to
regress. In the top line growth is advancing towards the right as follicles
recommence growth. After growth has persisted for sufficient time,
sufficient inhibitor accumulates to cause the follicle to cease growth
(bottom line). Growing follicles are indicated by a line with an enlarge-
ment at the end, and non-growing follicles as a shorter line.

(b) The lower illustration is an indication of how a typical growth wave,
commencing on the ventral surface of an animal, can travel dorsally.

waves can be explained as follows. We consider a population of resting
follicles and suppose that at A the concentration of inhibitor falls below
the threshold value permitting growth to recommence (Fig. 64). In
response to the demands of the growing cells, which initially include cells
of the adjacent epidermis, and outer root sheath, the vascularity beneath A
increases. If now the concentration of inhibitor in the neighbour of

THE GROWTH OF EPIDERMAL STRUCTURES

155

follicles adjacent to A is also approaching the critical value, and if the
follicles are sufficiently close to one another, the increased blood flow will
initiate growth in these follicles. By the same process further follicles
will be stimulated and growth will travel towards fiasa wave. Behind the
front, growth will continue until the inhibitor again accumulates to the
critical valve when the follicles will again enter the resting state together.
On a simple cylinder (Fig. 64b) (model of the body of an animal) if
growth commences along a ventral line, waves will travel dorsally on lines
parallel to the initial line as is observed. The extremities, the legs, ears,

dorsal

GRAFT

Fig. 65. The result of rotating an area of skin through an angle of 180°
and regrafting it as described by Ebling and Johnson (1959). When a
growth wave moving dorsally reaches the level AB, hair commences to
grow at the dorsal edge of the rotated graft (as if this were not displaced)
and travels ventrally. This is contrary to the predictions of a theory based
on simple control by accumulated inhibitor.

etc., pose special conditions which will break the uniformity of the pattern
but growth should still proceed on fronts. Inhibitor theory is thus able to
describe qualitatively the appearance of waves, but it cannot yet be
developed quantitatively.

Ebling and Johnson (1959) claim to throw doubt on this explanation.
They severed areas of rat skin entirely from their dermal connexions
and regrafted them after rotation through 180°. On such grafts the
growth wave normally proceeding in a ventral-dorsal direction (see Fig. 65)
commenced on the dorsal side of the graft and travelled downwards. That
is to say the follicles in the graft behaved exactly as they would have if they
had not been rotated, whereas the wave transmission theory would predict
that the wave on reaching AB would advance dorsalwards across the
graft. Ebling and Johnson infer that the follicles are actually independent
and the wave advances simply because there is a ventral dorsal gradient

156 KERATIN AND KERATINIZATION

in the sensitivity of the follicle to some periodic systematic stimulant
which is perhaps brought about by a graded difference in the rate at which
the growth inhibitors are dissipated. Whitely's (1958) findings are similar.
Periodic growth is also apparent in feathers, in the formation of growth
bars and perhaps daily variations also occur (Lillie and Wang, 1940;
Liidicke, 1959).

Zig-zags, Curls and Crimps

In the rat and mouse, the fine hairs are not straight, they divide into
short lengths separated by narrow nodes to form " zig-zags " (Dry, 1926
and 1928). This phenomenon seems to point to the existence of small
localized fluctuations in inhibitor concentration which affect more the
finer hairs, less deeply embedded, often more closely clustered, and able
to exert less command over supplies and/or with less chance to dissipate
inhibitor concentration than the stouter more deeply-seated primary hairs.

Zig-zags are probably related to crimps and curls. Histologically these
latter modifications are associated with curved follicles and it seems not
unlikely that the basic cyclic activity of the bulb is linked to cyclic changes
in a curved and asymmetric follicle which co-operate to stabilize the
emerging wave-form, i.e. the period of the bulb becomes related to the
time taken for the cells to pass through the curved tube formed by the
upper part of the follicle. The remarkable regularity of the emerging
wave-form would suggest that some kind of feed-back, possibly
mechanical, must integrate the entire follicular activity.

With stout hairs the simple existence of a follicle curved in the zone of
hardening (Chapter 6) would seem sufficient to produce a hair of more-or-
less constant curvature, i.e. a helix or simple curl, since the emerging hair
would retain the shape of the " mould " in which it was set (Fig. 66). The
curly locks usually found in the fleece or pelt are of extra-follicular origin
and seem to result from the tendency of the hairs after emergence to adhere
laterally to each other in clumps or bundles (Horio and Kondo, 1953).

Crimps are more nearly planar wave-forms and often of impressive
regularity (Norris, 1931). They develop typically in fine flexible hair and
there is a close correlation between the wavelength of the crimp and the
ease of bending the fibre (diameter) which is taken advantage of in the
practical method of judging fibre diameter by eye. The variable curvature,
which gives rise to " sickle tips," is discussed on p. 150.

The periodicities of crimped wool which comprise not only wave-
length but rhythmic changes in diameter, shape and chemical composition
(Mercer, 1954) force one to suppose that the various steps in its formation
are interconnected in such a way that information concerning the portion
of a wave already produced is fed back to the keratinizing and germinal
layers to control current production. We appeal here to a broad principle

THE GROWTH OF EPIDERMAL STRUCTURES

157

of cybernetics, that such a regularity of form in changing conditions could
only be maintained by information transfer. It remains, however, to
identify the links in the feed-back train.

The histology of the wool follicle has been much studied and a very com-
plete account given by Auber (1950). Problems relating to crimp for-
mation are discussed by Wildman (1932), by Rudall (1936) and also by

Fig. 66. Theory of crimp formation in a wool follicle (see text). The
following structural elements are distinguished:

(a) The point of emergence of the fibre from the skin. The fixed
orifice.

(b) The upper reaches of the follicle which contain a set and perman-
ently curved segment of fibre with radius of curvature R.

(c) The keratinization zone. The "mould" of curvature — R (opposite
to that in segment B) in which the growing fibre is " set " with a cur-
vature — R.

(d) The bulb, the origin of the rodlet of protein which enters the zone C.
For the purposes of clearer illustration, the geometry of the follicle has
been simplified and idealized. The situation depicted shows the follicle
in an extreme position in which C is shown in its extreme extent to the
right. O and P indicate the locations of the two types of keratin ortho-
and para- (see p. 273).

Auber (1950). We shall return in Chapter 6 to special questions relating
to keratinization in curved follicles. For our present purposes the structural
elements likely to be involved in crimp production and maintenance are:
the deflected bulb, the curved follicle and the (usually) asymmetric placing
of the hair structures within the outer root sheath (Fig. 114).

158

KERATIN AND KERAT I NI ZAT IO N

It can be ascertained by passing a wire, bent into the form of a sine
curve, through a short length of flexible tubing held at the end from which
the wire emerges, that the movement leads to an oscillation in the cur-
vatures of the free end. Thus in the case of the crimped fibre, the passage
of the hardened hair through the upper portions of the follicle (fixed end)
will tend to cause an oscillation in curvature of the softer, lower portions of
the follicle (the analogue of the free end) (Figs. 66 and 67). A feed-back

(a) (b) (c)

Fig. 67. Successive stages in the production of a full crimp wave.

(a) A segment 1 of curvature + R is passing along the neck B and
emerging through the fixed orifice. At the lower end of B a segment 2
of radius — R is about to enter B.

(b) The segment 2 has entered B and has deflected the neck B into the
opposite curvature — R by the time it is about to emerge. By deflecting
B into this shape it has deformed the setting zone C into the reverse
curvature -f Rso that the length of fibre 3 being hardened at this point
will now have the reversed curvature.

(c) Segment 3 has now entered the upper levels swinging them over to
the original shape + R and the effect on the lower levels is to reverse the
curvature there.

Thus the curvature in the " setting zone " is maintained in an out-of-
phase condition relative to the upper zone by the mechanical feed-back
link. Growth conditions on the inner face of the curved zone C of Fig.
66 and the adjacent zone of the bulb D are assumed to be less favourable
than on the outer face and to result in less rapid growth on that side.
Assuming further that growth in the bulb is also periodic, the rate, quality
(o or p) and size of outer root sheath also swing from side to side. This
oscillation may also resonate with the mechanical vibration in zones B
and C, the asymmetric growth giving rise to a curved segment to enter in
phase with the curvature of the zone C.

THE GROWTH OF EPIDERMAL STRUCTURES 159

to the germinal level could thus be effected through this mechanical link.
A control over growth in the bulb could be produced by the movement of
the bulb into regions differing in inhibitor content or simply by mechanical
deformation. Three periods are involved : (a) the period of oscillation pro-
duced by the curved fibre passing through the upper levels; (b) the time
for the fibre to pass through the zone of hardening; and (c) the period of
the growth cycle of the bulb. All these periods are variable within limits
and it is likely that the coupled oscillation, which evolves and is stabilized
by feed-back, results from a selected resonance between all three (p. 146).

The successful transfer of control from the curved fibre in the upper
reaches of the follicle back to the lower levels required that the fibre be
stiffened before it passes into the curved upper levels. A soft fibre would
not produce the postulated oscillation. In confirmation of this, Marston
(1946) has shown that in the event of incomplete keratinization, as occurs
in sheep deficient in copper, the wave-form is poorly developed and its
frequency (number of waves produced per unit time) is lower as would be
expected from a weaker fibre. The above explanation of crimp formation
has some features in common with that given by Auber (1950) and
Wildman (1932). However, Auber's assumption on which his explanation
rests must be rejected. He supposes that the a-structure of the fibrils is
produced by an actual contraction in length produced during hardening
and that the contraction is greater on the more highly-keratinized inner
face {para) of the fibre. There is ample experimental evidence (p. 21 1 et
seq.) that the a-structure is present in the fibrils as originally formed and
owes nothing to the subsequent chemical changes occurring during
keratinization.

A characteristic relationship between crimp form and tip shape in the
various classes of wool fibres was first described by Dry (1926 and 1928)
and more recently discussed by Fraser (1951) (p. 149) who has tried to
explain it in terms of interfollicular competition. Fraser develops the idea
of a competition between follicles for the materials needed for growth, and
suggests that their efficiency in this competition depends in part on the
time of origin of the follicle. In these terms he gives an explanation of the
formation of the first few curves at the tip of a wool fibre which precede
the establishment of the regular crimp form. The concepts of competition
and follicular efficiency seem to overlap and supplement those based on
inhibition; both ideas stand in need of further analysis and testing (see
also p. 149).

Allometric growth

When the different parts of an organism are compared it is usually found
that they grow at different rates and that the proportions of the organism
thus change as life continues. A formula which has often been found to

160 KERATIN AND KERATINIZATION

relate the amounts of growth in different parts is the allometric equation :

y = bxa
where x and y are the sizes of two parts and a and b are constants. The
equation gives expression to the idea that each part grows by self-multi-
plication (giving the exponential law p. 139), but that for reasons depending
on their " appetites", their command over or access to food supplies, etc.,
the exponential factors are not equal in the two separate growth equations
(Reeve and Huxley, 1945; Richards and Kavanagh, 1945).

That the law applies to epidermal growth vis a vis that of the whole
organism is suggested by the often noted fact that horns tend to be
relatively larger, the larger the animal, but no data really adequate to test
it exist. In any case much epidermal growth is strictly accretionary, i.e. a
fixed amount of growing tissue is constantly adding to a store of dead
material which itself, since it is no longer contributing to the growth,
should be subtracted from the measured size before testing the equation as
in the modified version proposed by Robb (see Reeve and Huxley, 1945) :
y = bxx + c, where c is the " dead weight". The existence of cyclic
growth, ecdysis and changes in relative rates due to changes in hormonal
patterns following crises such as puberty (equivalent to discontinuous
changes in a) are further complications.

CHAPTER V

Molecular and Macromolecular Structure

The present status of the chemical structure of the keratins

The chemical composition and constitution of a protein may be con-
sidered established when the following are known: (a) the number of
separate polypeptides composing the molecule and the nature of any
covalent cross-linkages uniting them ; (b) the amino acid sequence in each
of the polypeptides; (c) if a prosthetic group is present, its relation to the
polypeptide moiety.

Complete solutions to (a) and (b) are available for three proteins; for the
keratins, which are far more complex, there is no immediate prospect of
even partial solutions. There is no evidence, however, to show that there
is anything in the nature of a prosthetic group to complicate the position
further. Essential information concerning a protein is provided by a
knowledge of its total amino acid composition and its end-group com-
position, i.e. the groups which terminate the main polypeptide chains.
For many keratins we have adequate, although not complete, deter-
minations of both end-group and total amino acid composition from which
may be inferred a general picture of the overall chemical reactivity of the
molecular complex (Table 7). For purposes of reference, in Tables 1, 2 and
3 (Chapter I) will be found the amino acid composition from a number of
determinations as reported by various authorities. The most studied
material is wool and the data relating to it are considered to be of a high
order of accuracy although, it is clear that precise agreement between
independent analysts has not been reached. Table 7, devised by Ward and
Lundgren (1954), presents a summary of the amounts and kinds of chemical
groups to be found in several keratins as calculated from their amino acid
composition. Assuming all these are accessible to reagents the Table
enables a fair prediction to be made of chemical behaviour. Similar
Tables will be found in Tristram's extensive compilation (1953) from
which the data for Fig. 2 (Chapter I) were taken.

End Groups

Information concerning molecular weight, the minimum number of
polypeptides, the presence of branched or cyclic chains and amino acid
composition may be obtained by determining the amino acids which form

161

162

KERATIN AND KERATINIZATION

Table 7. Keratins as Chemically-Reactive Substances*
(gramme equivalents per 105 g of keratin).

Keratin

Reactive group

Epi-

Wool

Hair

Horn

Quill

Feather

dermis

free carboxyl

58-66

38-70

71-72

27-44

27-83

amide

79-98

carboxyl plus amide

137-164

122-154

152-153

185

105-122

110-166

phenolic hydroxyl

22-36

12-17

20-31

11-12

19-32

aliphatic hydroxyl

124-148

129-172

116

91-104

134-174

186

total basic

78-92

68-91

59-92

65-68

46-59

59-120

amino

20-24

13-21

16-25

7-12

21-47

aromatic nucleif

27-43

16-24

24-37

13-14

23-38

half-disulfide

92-114

138-150

101-131

67-80

57-68

19-32

oxidizablej

526-650

722-814

538-754

381^144

309-376

207-280

* Taken from Ward and Lundgren (1954).

t Aromatic nuclei available for coupling with diazonium salts : tyrosine
plus histidine.

X Oxidation capacity calculated as twice the number of oxygen atoms
rapidly consumed from performic acid.

Table 8. N-Terminal Amino Acids of Hair and Wool*

Material

Amino

acid

Lincoln

Romney

Merino

Wool

Human

wool

(fine)

hair

glycine

5-2

4-5

4-7

3-8

3-9

alanine

1-3

1-2

1-1

1-0

valine

2-4

1-0

4-0

serine

1-3

1-2

1-05

1-0

threonine

4-8

4-9

3-2

4-0

aspartic

0-6

0-5

glutamic

1-3

1-2

0-9

1-0

See references on page 163.

MOLECULAR AND MACROMOLECULAR STRUCTURE 163

the terminal residues of the chains. The end with a free — NH2 is referred
to as N-terminal; that with a free COOH as a C-terminal.

N-terminal groups are determined by reacting the protein with Sanger's
reagent dinitrofluorobenzene (DNF) which attaches to the free amino
group producing a substituted amino acid easily identified after hydrolysis
by its yellow colour. The C-terminal groups are found by reacting the
protein with hydrazine (N2H4) when the chain residues are converted into
hydrazides with the exception of the C-terminal acid.

Several workers have examined wool and are in good agreement as to
which acids form the end groups although estimates of numbers vary.
Blackburn (1950) reports the same groups for a N.Z. coarse wool as Kerr
and Godin (1959) for human hair and horse hair. The latter refrain from
citing their quantitative findings, accepting Thompson's (1957) criticism
that these methods are not reliable for materials such as solid keratins.

Middlebrook (1951) rounded off his figures to give the following twenty-
seven N-terminal groups: glycine 8, alanine 2, valine 4, serine 2, threonine
8, aspartic acid 1, and glutamic acid 2 in 106 g wool keratin. He calculates
a chain weight of the order of 60,000 and thus a total molecular weight
for " wool keratin " of 27 x 60,000 ~ 1,600,000 assuming homogeneity.
Estimates for chain weight from the figures in Table 8 vary from 50,000-
90,000 approximately.

The C-terminal groups of wool were also found by Blackburn and Lee
(1954) to be glycine, alanine, serine, threonine, aspartic and glutamic acid.
These same six residues were found by Kerr and Godin in hair. Alexander
and Smith (1956) have determined the end groups of three fractions a, 0
and y derived from wool oxidized by peracetic acid (see p. 238) and found
the same groups in somewhat different proportions.

The N-terminal end groups of a soluble derivative of feather keratin
have been determined by Woodin (1954a and b, 1955 and 1956). The
same groups as for wool and hair were found in very small quantities
indicating a deficiency of N-terminal groups which Woodin thinks may
mean that feather keratin is a cyclic polypeptide. Krimm and Schor
(Schor, 1958), who favour a much larger unit (see p. 208) than Woodin's
monomer (m.w. 10,000), think that in solublizing his feather, he hydrolyses
the peptide bonds linking the frequently-occurring proline residues and
thus produces shorter chains with proline end groups which are not then
detected by the procedure used.

Molecular structure

Methods of Partial Degradation

Amino acid analyses and other analytical procedures show that keratin
consists of polypeptides and that, if constituents other than amino acids
are present, they are there in very small quantities. Methods of partial

164 KERATIN AND KERATINIZATION

degradation are resorted to in the attempt to discover the existence of
macromolecular units of an intermediate range of size. The study of the
problem of how many polypeptides participate, the composition and length
of these chains and their sequence of amino acids has not progressed far.
The formidable difficulties of the problem are obvious. Most methods of
studying proteins have been developed for the soluble proteins and, to
apply these, we have first to develop methods of obtaining soluble deri-
vatives of keratin by procedures which will permit the structure of the
insoluble original to be inferred from that of the soluble derived fragments.
The solutions must be fractionated into their constituent polypeptides
and the location of the constituents in the original solid complex established.
Most of the work of this nature will be described in Chapter 6 and is
only mentioned here for reasons of formal completeness. Most workers,
accepting the view that the insolubility of keratin is due to two factors :
(a) the interchain cross-linking by the sulphur bridge of the cystine
residues, and (b) the presence of numerous hydrogen bonds, have at-
tempted, either in a single stage or in successive stages, to rupture both
classes of bonds and thus to obtain a preparation of the free, constituent
polypeptides (p. 233). A considerable number of preparations have in fact
been made, often with a view to obtaining a product of some commercial
value, but the few of these that have been examined in detail have not
proved promising sources for the extraction of pure proteins or their
derivatives. It is in the purification of these mixtures that a very substan-
tial contribution towards further progress is expected. For the present
we may summarize the results for wool as follows: all the materials
examined contain several polypeptide species probably differing in com-
position. There is some evidence to suggest that the derived polypeptides
can be grouped into two classes probably related to two distinct mor-
phological components, filaments and matrix, visible in electron micro-
graphs (Chapter 6). The proteins derived from the filaments have a
higher molecular weight (50,000-80,000) than those of the matrix
(~ 10,000); the latter contain a greater amount of the cystine residues.
None of the many components have been prepared in a state of complete
purity and little progress is possible at the moment towards forming a
detailed picture of the place of each in the original insoluble hair. In the
case of feather Woodin's 10,000 m.wt. unit has some claims to be regarded
as a definite monomer.

Non-destructive Methods
X-ray diffraction

Although the chemical analysis of a complex material of biological
origin may have only a limited value unless pure components are first
extracted, the X-ray analysis of the same material may bear immediate

MOLECULAR AND MACROMOLECULAR STRUCTURE 165

fruits if a single dominant component is sufficiently crystalline to
yield a recognizable diffraction pattern (Chapter I). The method, as it
were, cuts through the tangle of secondary structures and minor constit-
uents and yields immediate information concerning the arrangement of
the atoms in the crystalline regions. Further, since it leaves the material
unharmed the same sample may be used for other tests. Fortunately all
the keratinized mammalian tissues, as was discovered by Astbury and his
associates (see Chapter I), give substantially the same pattern, proving that
they contain crystallites of similar molecular structure. Feather keratin
gives a /S-type pattern, one of the most detailed yielded by any protein
fibre.

The formal analysis of the X-ray diffraction patterns of a well-crystalline
material can lead to the exact placing of the atoms (other than hydrogen)
in the structures and a growing number of organic compounds of biological
importance have been determined in this way. The proteins, whose
molecules may contain thousands of atoms, offer enormous difficulties
but great success has been achieved in recent years (Perutz, 1959).

The low-angle pattern. The total X-ray pattern yielded by biological
fibres is separated conventionally into two parts (Fig. 3, p. 11) described
as the wide-angle pattern and the low-angle pattern. The dividing line is
arbitrary, but reflections corresponding to a Bragg spacing of less then
20 A are referred to the wide-angle pattern, those corresponding to
longer spacings belong to the low-angle pattern. As described in Chapter
I, the number of wide-angle patterns is small — whole groups of fibrous
proteins being characterized by the same pattern. In contrast to this
simple situation among the wide-angle patterns, a considerable variety of
small-angle patterns are found even among proteins which are classified
together on account of their similar wide-angle patterns. These low-angle
X-ray patterns of the protein fibres are often of great complexity, and
indicate the existence of elaborate structures of macromolecular dimensions
which are as yet ill-understood. The relation of these structures to the
smaller molecular formations responsible for the large-angle patterns we
have just been considering, is also obscure. The " long spacings", which
may be either meridional or equatorial in placing in the X-ray patterns, can
be measured with some accuracy. If the reflection is sharp, a lattice
spacing corresponding to it is usually calculated by means of the Bragg
equation. When they are diffuse this procedure is difficult to apply; but
the totality of the pattern is characteristic of the protein and may undergo
changes when the material is chemically altered. In this respect the
low-angle pattern is more characteristic of a particular protein than are the
wide-angle patterns.

Sketches of two low-angle a-patterns are given in Figs. 68 and 69, and in
Table 9 are listed the long spacings of the keratins as measured by

166

KERATIN AND KERATIN IZATION

Macarthur (1943), Bear (1943) and Bear and Rugo (1951). Table 10 by
Schor (1958) summarizes the various reflections reported for the more
elaborate feather diagram. Discussion of this is postponed until later (p.
208). In a-keratin the longitudinal spacings can be regarded as orders
of a master spacing of 198 A and the equatorial spacings as orders of
80-90 A. In feather the main longitudinal spacing is usually given as 95 A
but Schor prefers 189 A.

Fig. 68. Composite chart of a-keratin diffraction pattern given by por-
cupine quill. Fibre axis vertical; plane plate. (To D = 3 A.) Taken
from Macarthur (1943).

A certain independence between the long spacing (macromolecular
level) and the short spacings (molecular level) is suggested by the fact that
the large-scale order may be destroyed without affecting the order at the
molecular level. This is commonly observed in the regenerated keratins,
i.e. materials which have been dissolved and reformed into fibres. These
may give either a- or j8- or cross ^-patterns at high angles (Astbury,
1947; Rudall, 1946 and 1952; Mercer, 1949a) indistinguishable from the
original patterns, but they lack long spacings proving that although
the small-scale structures in the a-crystallites have been regenerated the

MOLECULAR AND M ACROMOLECUL AR STRUCTURE

167

larger-scale structures have not reformed. On the other hand, experimental
procedures, such as the " heat-moisture treatment " of Bear and Rugo
(1951), also exist by which the small-scale molecular order may be de-
ranged while leaving the macromolecular pattern largely intact.

The long spacings revealed by diffraction (30-700 A) overlap those
which can be resolved in the electron microscope (50 A and longer) and in
some instances the two methods give results in good agreement. For

Table 9.

Small-Angle Diffractions of oc-Keratin Fibres!
Meridional and near meridional reflections.

African

porcupine

quill

Human hair

(A)

198

195

49*

196

49*

196

195

27-4

192

27-6

193

24-5

196

24-3

194

22-0

198

19-8

198

19-4

194

18-1

199

15-2

198

15-0

195

13-2

197

12-4

198

12-1

194

10-4

198

Equatorial reflections

«o

(?)

45*

47*

28f

29f

The d columns contain the measured spacings which, when
multiplied by the assigned k or h indices, yield the large fibril
period, b0, or the large transverse fundamental spacing, a0, of
80-90 A. The / columns indicate in rough fashion the relative
intensities of the diffractions.
* Overlaid with faint, poorly-oriented rings, probably due to

lipid.
t From Bear and Rugo (1951)

168

KERATIN AND KERATINIZATION

example, in the case of collagen which shows a well-marked long spacing
of 640 A, the electron-microscopic image reveals the same fundamental
period.

Yet even in this instance, in which a relatively-large amount of data
from many sources is available and the basic arrangements of the poly-
peptide chains as a triple chain helix are precisely known (see p. 127),
the nature of the major spacing and the many subspacings visible electron-
microscopically is still under discussion (Hodge, 1960).

;.'.; :'

■ *•■» •

♦

.v. *

■■m

*L'

. . .

•'•••••••* *

. •

*"• '.\-V;li

■;**

•

^::--' '• •'

Fig. 69. Drawings of the respective wide-angle and pinhole small-angle
patterns of clam ( Venus) muscle. The fibre-axis direction in all patterns
is vertical. M is the prominent composite meridional arc at 5'1 A, E the
equatorial diffraction at 9'6 A, these being the characteristic a-pattern
diffractions at wide angles referred to in the text. L in the left diagram
marks the small-angle series of diffractions which are indicated in greater
detail in the right figure. Adapted from Bear (1951).

With keratin the situation is much less satisfactory. Convincing
electron-microscopic observation of the longitudinal spacings is wanting.
Filaments can be observed both in growing hair and in disintegrated wool
which have diameters of the expected order of 60-100 A and the packed
array of these may be responsible for the lateral spacings (Birbeck and
Mercer, 1957a; Rogers, 1959) (Plates 15 and 16). Fibrils in disintegrated
wool (Farrant et al., 1947; Jeffrey, Sikorski and Woods, 1956) and in
extracted sections of skin (Porter, 1956) have a quasi-regular nodular
appearance which does not possess a periodicity of the expected order

MOLECULAR AND MACROMOECUL AR STRUCTURE

169

•3

47
23-4

10-4
9-1

6-26

4-93
4-42

3-54

3-29
3-08
2-94
2-74

00 OOt^iOCNtNt^

co O tN q\ ro On lo <N ©

,1c o\ i tin n nn

tN »-< n ■* tN ui O O

t-«ir)cibinir)rtOa>obr^t^vouTt't

T*- CN ^ T-. •rH ~ T*

ID © C". 00 "■- ■* ONOO-^--*

•O On ■<*- ninOvt m NO^t^
(^ t^o \oio-4--4- n nntNN

3
u

t>« r^ fj on-* o (^lDc^^cj^ooiD-^-cNoqNr^inir) t* V
© c<n On A6 r-» ^ u~, ^•&v-,^>c^cri<ri<r-. r4c<ic*\c< NcN

MrnOOcoOONtN(NOnOifiOOHOiNq<tNtNinn
6oboN>o6a>obt>t^h.6'-<o^ooc>ooo>ooNooooffva>oo

ononooooonooooooooooononooooooooooonoooooooooooo

OhO
00 00 00

tN'+COOiNt'OMONint^OiNOiOiNCOOnClXO'+OC
^.^r-.T-HT-CNCNCNCNcoc^CDTt-Tl-iDiDu-iirivOvO'O

ON Tj- O

r^ oo on

c
c

|^|S3S?«|||«rs,^SI-||

1 £ £

> > >

60
03

in oo *-i © on on t~» "">

j^ii/I vOCT^iDt^ nvfuOC 00 •"}- r?

-*• o t^ ■**- o r*» i> "■> oi>

(^ © SO -4" MM^cnco CN fS

CM Tt

on m o

MCNtH

CS CN CN

170 KERATIN AND KERATINIZATION

calculated from, the X-ray patterns (198 A). In thin sections even well-
formed materials, such as feather and porcupine quill, fail to show
unambiguous evidence of longitudinal spacings although the expected
values, 96 A and 198 A respectively, are well within the instrumental range.
This further emphasizes the different organization of the collagens and
keratins. The system of long equatorial side spacings could very likely
arise as scattering from the quasi-regular packing of filaments embedded
in a matrix of different scattering power (p. 247). Significantly, Fraser
et al. (1957 and 1959) have observed intensity changes in this group
of spacings when fibres are treated with osmium tetroxide in a manner
known by electron microscopy to lead to strong deposits of osmium
compounds in the matrix (p. 248). The 84 A lateral spacing probably
corresponding to the interlayer spacing of filament and matrix is strongly
enhanced.

Speculation on the nature of these long meridional spacings has tended
to take the form of either of two extreme theories which may be called (a)
long chain theories, and (b) corpuscular-aggregate theories. In the first
theory the larger fibril is pictured as being built up from many parallel
chains (or other thin linear elements) and the periodicities along the
macroformation arise from the repetition of structure in the basic chain.
This mode of construction seems to apply to collagen. Many of the electron-
microscopical images can be accounted for in terms of such a unit (Hodge,
1960). Solutions of both types have been proposed at one time or another
for the keratins. The elaborate equatorial reflections given by feather
suggested an aggregation of corpuscles to Astbury and Marwick (1932).
Macarthur (1943) considered the possibility of the long-chain model;
and the nodular appearance of the fibrils released from reduced wool by
enzymatic digestion suggested a linear aggregation of corpuscles to
Farrant et al. (1947). There is other evidence that many other fibres
are formed by the aggregation of particles (Astbury, 1949 and 1958;
Jeffrey et al, 1956).

The wide-angle patterns. The wide-angle patterns yielded by the fibrous
proteins (see Chapter 1) contain far less information in the crystallo-
graphic sense than those now available from crystalline proteins. Never-
theless, the first progress towards an understanding of the arrangement of
the polypeptide chains in the solid state came from a study of mammalian
hair and silk fibroin. This progress was made possible by an ingenious
integration of data derived from a variety of sources: chemical com-
position, X-ray diffraction, physicochemical and elastic behaviour. We
owe the development of these methods principally to Astbury (1933) and
Astbury and Woods (1933), and in spite of an enormous increase in the
precision of the crystallographic side of the work, the principles of the
methods remain essentially as devised then. See also Kendrew (1954).

[OLECULAR AND MACROMOLECULAR STRUCTURE

171

rp sp rp 00 Op

to i> «-h do to

« O

3£

omoooo 10

* g £ £

3 h

-a

do ^ \b ode i"

:ss«f? T

o>(So»hooi«

-111

172 KERATIN AND KERATINIZATION

The first and most essential experimental step was the discovery by
Astbury and Street (1931), that stretched hair gave a different pattern
from unstretched hair, a pattern of the j8-type similar in fundamentals to
that yielded by silk. Meyer (Meyer and Mark, 1930) had given a broadly
satisfactory interpretation of the structure of silk in terms of a model
structure composed of a bundle of parallel straight polypeptide chains
oriented parallel to the fibre axis. Astbury recognized that, allowing for
the differences in amino acid composition between the rather simple
fibroin and the more complex keratin containing large bulky amino acid
groups, the same type of model would account for the major features of
the /3-type keratin pattern. This close similarity between silk fibroin and
/?-keratin still forms the basis of present-day proposals for the structure
of the latter. The structure of fibroin is now known with great precision,
(Marsh, Pauling and Corey, 1955) and accordingly the proposed structures
for /?-keratin are formulated with equal precision and are thus more
capable of being experimentally checked. These will be described later.

Astbury's second step was the far-reaching proposal that the a-structure
developed from the shortening by folding of the straight polypeptide chains
of /3-keratin. This idea was suggested by a comparison with rubber, the
elasticity of which was currently being explained in terms of the coiling up
of the polyisoprene chains and by the proposal of Meyer and Mark (1930)
that other forms of " protoplasmic " elasticity, such as muscle contraction
might rest on chain coiling. In Astbury's hands this concept of polypep-
tide chain folding was extensively exploited and today forms the basis of
most ideas about protein structure and of polypeptide behaviour.

The elastic properties and the structure of hair

If a-keratin is a shortened form of /3-keratin, then it would seem possible
that the degree of folding of the polypeptide chains in a-keratin could be
inferred from the increase in length necessary to produce the a-/? trans-
formation. This possibility led Astbury and Woods (1933) to an extensive
analysis of the elastic properties of wool and hair.

A description of some aspects of the complex rheological behaviour of
hair and wool will be given later (p. 249). Here we are concerned with one
particular problem: the relation between molecular extensibility and the
change in length of the whole fibre, and the possibility of inferring the
extensibility of the molecular chains from the extension of the whole
fibre. For this purpose it is necessary to analyse in greater detail the
sequence of events which occur during the stretching of a hair. Fig. 70 is
an idealized version of a stress-strain curve for a wool fibre or hair as
described by Speakman (1928) and by Astbury and Woods (1933). It will
be seen that it appears to divide into several distinct steps :

OA (0-2%): In this section the curve is closely linear and is for this

MOLECULAR AND MACROMOLECUL AR STRUCTURE

173

reason referred to as the Hooke's law region. Here the fibre is behaving
like a conventional elastic solid and the axial molecular-lattice spacings,
as found by X-ray photographs, show a reversible increase of the same
order of extension as the fibre. The value of Young's modulus and the
amount of the Hookian extension decrease as the humidity and tempera-
ture are increased.

AB (~2-~20%): At the point A the fibre yields suddenly and
there is a rather rapid extension of the fibre to about 20%. During this
extension the X-ray photograph remains of the a-type but becomes less

20 30

% extension

Fig. 70. Idealized stress-strain curve for a keratin fibre^(wool or hair)
stretched in water at room temperature : OA (0-2%) Hooke's law region ;
AB (2-20%) extension of amorphous phase (phase I) a-fades and /?
appears; BC (20^4-5%) extension of crystalline phase (phase II); CD
(45%) extension of phase III non-crystalline (adapted from Astbury and
Woods, 1933).

perfect as extension increases. Dry fibres break usually at about 20%
extension. At B a shoulder develops showing that the fibre is becoming
more difficult to stretch; the a-pattern rapidly fades and ^-reflections
appear. This shows that in this region the a-crystalline phase is being
destroyed and that a new configuration is appearing.

CD (> ~ 45%): The fibre again becomes more easy to stretch. The
transformation of the a-crystallites into the jS-modification seems complete.

Extension rarely exceeds 60-70% in cold water, but is greater in hot

174 KERATIN AND KERATINIZATION

water or in steam. It is supposed that at these higher extensions a further
non-crystalline fraction must extend.

An important feature of these extensions and the X-ray changes that
accompany them is that (excluding the phenomena of set, p. 249) they are
reversible and that the type of X-ray pattern is closely connected with the
degree of extension.

Astbury and Woods suggested that the step-like nature of the curve
OABCD was due to the presence in the fibre of three " phases " of keratin
which differed in their ease of extensibility and which were effectively in
series. On stretching a fibre each extended in turn: at A the " transfor-
mation tension " of phase I was reached, at B phase II, etc. Since no
marked change in X-ray pattern occurs in the range AB or beyond C
these portions of the curve must correspond to the extension of non-
crystalline material (phases I and III); the range B-C where the a — /3
transformation occurs must be where the crystalline phase extends
(phase II).

Since the steps in the curve are not sharp it was assumed that the
separate phases were not simply in series but that, as a result of restrictions
exerted by one phase on another, they were also partly in parallel. With
the assumption of three extensible elastic elements, with different re-
sistance to extension, this model has sufficient variables to provide a fit
even for the infinitely varied responses of the wool fibre.

Woods (1938) met the possible criticism that, since hairs were histo-
logically complex, certain features of the step-curve might be due to the
extension of various histologically-recognizable components joined partly
in series and that the extension of these parts (including the crystalline
region) might not be the same as that of the whole fibre. He stretched
wool fibres to various percentage extensions, " set " them at these lengths
by steaming and, by means of tryptic digestion, isolated the stretched and
set, keratinized contents of the cortical cells. By plotting the percentage
increase in length of the whole fibre against that of the cell residues, he
showed that the cell contents parallelled very closely the extension of the
fibie as a whole. A slight lag was interpreted as due to either a loss of set
during the isolation of the cells or, as is most likely the case, to the extension
of the component which links the cells together, i.e. in current histological
terms, to the cell membranes and intercellular cement. It is thus possible
to say that the elastic properties of the fibre are mainly those of the
keratinized contents of the cortical cells.

In water at ordinary temperatures extensions of the order of 70%
are possible; at elevated temperatures (steam) or in solutions of dilute
caustic soda a limiting extension of the order of 100% was found by
Astbury and Woods (Fig. 71). They assumed from this that at 100%
extension all the polypeptide chains were fully extended. By making a

MOLECULAR AND MACROMOLECULAR STRUCTURE

175

further assumption that all parts of the fibre (and their component mole-
cules) were equally extensible, they were thus able to conclude that the
fully extended molecular chain (/3-form) was twice as long as the folded
chain (a-form). This was equivalent to assuming that the molecular
configurations in the non-crystalline fractions (phases I and III, Fig. 70)
were essentially similar to those prevailing in the crystalline phase II, but
for various steric reasons were unable to assume the perfection necessary
for crystallization.

Fig. 71. Creep curves for Cotsvvold wool under constant load in steam

(continuous line) and in a 1 % aqueous solution of caustic soda (broken

line). The timescale has been contacted by X2 for curve (a) and expanded

by X10 for curve (b) (Astbury and Woods, 1933).

On the basis of models of polypeptide chains and the assumption that
the a-form is half the length of the j8-form, Astbury and his associates
advanced successively two definite proposals for an a-keratin structure.
The first of these (Astbury and Woods, 1933) was shown by Neurath (1940)
to be too tightly folded to accommodate the side groups. The second
(Fig. 72) (Astbury and Bell, 1941) gives more ample opportunities of
intrachain H-bond formation, appeared to satisfy most of the require-
ments. Even without a completely detailed description of atomic
positions this broad concept of a folded polypeptide chain linked by a
variety of side chains to other main chains proved, in the hands of the

176

KERATIN AND KERATINIZATION

Leeds School, capable of co-ordinating a great deal of physicochemical
data. Recent discussions of the elastic properties of fibres are given by
Feughehnan (1959), Peters and Woods (1956), Skertchly and Woods
(1960).

The conclusion that the a-keratin chain was half as long as the ^-chain
was not universally accepted and other models were also proposed.
Notably Ambrose and Hanby (1949) and Zahn (1947 and 1949) proposed
a fold, which yielded the required axial periodicities and gave an a -> /J
extensibility of ~ 33%. This appeared to accord with the observation that
the actual transformation of the a-crystallites into the jS-form took place
between fibre extensions of 20-45% (phase II, Fig. 70) which could be

~Xfu%r

.t*o Cl+J> cL^ji o m >s

Fig. 72. The Astbury model for a-keratin: arrows represent the
direction of the main chain; $ represents a side chain pointing up from
the plane of the diagram ; O represents a side chain pointing down from
the plane of the diagram. From Astbury and Bell (1941).

interpreted to mean that the crystallites themselves were fully extended by
a change of length from 1*2 to ~ 1*5, i.e. of the order of 25-30%. Ambrose
and Hanby also claimed that the less-folded structure fitted in better with
their observations on the average orientation of H-bonds inferred from
absorption spectra of polarized infra-red radiation (p. 197). Other pro-
posals stemmed from Huggins (1943). It is to be noted that none of these
earlier models appears capable of yielding an axial periodicity of 1-5 A
which is now recognized as being fundamental to the a-structure (p. 182).
All yield the 5-1 A (axial) and 10 A side spacing.

Current crystallographic analysis

A decisive step in the approach to the problems of the structure of the
fibrous proteins was ushered in by the proposal of helical structures by

MOLECULAR AND M ACROMOLECULAR STRUCTURE 177

Pauling, Corey and Branson (1951). Up to this time it had seemed
difficult to propose structures which would be capable of quantitative test.
However, stimulated by the precise structures proposed by Pauling and
Corey, a new attack opened, with the result that at the present time
probable structures for most of the fibrous proteins are known.

These impressive developments stemmed in the first place from methods
which were in principle the same as those originally applied by Astbury
to the keratin problem. That is, from the study of a number of small and
simple compounds, such as amino acids and peptides, which permit of a
full crystallographic analysis, accurate information about the bond lengths
and angles found in unit structural elements occurring in proteins is
obtained. From these data models of the structural units are constructed,
and from these the probable conformations of peptide chains are inferred.
In this way Pauling and Corey, using the very accurate data relating to
bond lengths and angles which had accumulated in their hands by 1953,
predicted the existence of two helical structures, which could serve as
models of folded chains in fibrous proteins, and of other structures built
from extended chains which could form the basis of models for £-type
structures.

A later important development, increasing the purely crystallographic
element in this work, was the publication by Cochran, Crick and Vand
(1952) of a theory of X-ray diffraction from helical structures which greatly
facilitates the interpretation of the patterns. The most fruitful result of
this theory was the elucidation of the structure of DNA (p. 109). It has,
however, also allowed a detailed comparison to be made between the
observed X-ray pattern of several synthetic polypeptides, prepared as
simpler models of proteins, and that predicted on the basis of helices, with
the result that the structures of these polypeptides are no longer in doubt.

The basis of any model of the polypeptide chain is the amide unit and
originally Pauling, Corey and Branson (1951) formed their conclusions
concerning its structure on these three compounds: A^-acetyl glycine,
/3-glyclyglycine and diketo-piperazine. Since that time several more
glycylpeptides have been solved and also the tripeptide, glutathione.
These analyses combine to give as the probable value for the C — N
distance 1-32 A and show that the amide group is planar with the CO and
NH groups in the ^raws-configuration. These features now figure in all
models of polypeptides and proteins (Fig. 73).

The structural principles formulated by Pauling, Corey and their
associates (Pauling, Corey and Branson, 1951; Pauling and Corey, 1951a
and b; Pauling and Corey, 1953; Pauling, 1958) as necessary conditions
for the precise prediction of stable polypeptide configurations are as
follows :

(1) The amide group ( — NH— CO— ) (Fig. 73) is planar.

178 KERATIN AND KERATINIZATION

(2) The following bond distances and angles found in small peptides,
etc., will occur in polypeptides:

aC— C = 1-53 A

C— O = 1-24 A

C— N = 1-32 A

N— aC = 1-47 A (all ± 0-01 A)

N-

-aC-

-c

110°

-c-

-N

114°

-c-

-N

1-25°

C-

-N— aC

123°

These dimensions are illustrated in Fig. 73 taken from Pauling and
Corey. The planarity of the amide group and the shortening of the C — N
distance to 1*32 A is attributed to resonance.

(3) Hydrogen bonds NH — O will form where possible with an N — O
distance of 1*79 ± 0*1 A and the O lies close to the NH axis.

(4) The /raws-configuration of the amide group is significantly more
stable than the m-configuration.

Fig. 73. The planar amide group according to Pauling and Corey (1955).

This structural information was systematically employed to construct
polypeptide models and two types of structure were derived depending on
whether the attempt was made to form the H-bonds by intrachain or,
on the other hand, by interchain association. The first led to helical,
folded chains, the second to sheets of associated polypeptides.

Not all crystallographers accept the Pauling-Corey conditions. Huggins
(1958) in particular, points out that a departure of 30° from planarity of
the amide group does not involve a marked instability and might well

MOLECULAR AND MACROMOLECULAR STRUCTURE 179

exist if other structural or energetic requirements were better satisfied.
There would seem to be reason to expect that structural irregularities
arising from this and other causes will be found in the polypeptide chains
in the amorphous regions.

Pleated- Sheet Configurations — Silk fibroin

Figure 74 (a) is a view parallel to the planes of the planar amide groups in
a polypeptide in which all of the side chains (j8-carbons) project away from
the amide groups. The planes of the amides slant alternatively to right and
left and thus may be fitted on to a " pleated sheet " with the pleats running
out from the paper. Successive chains may be placed on the sheet in either

qq oq go

-OQ~ ~ ^ -Oo-OQ- ~ Aj}

^«£^

(a)

lb)

Fig. 74. The structure of silk fibroin: (a) view parallel to the planes of
the planar amide groups ; (b) the pleated sheet.

of two ways depending on whether adjacent chains are parallel or anti-
parallel in direction. By making small adjustments of the parameters of
the atoms in order to achieve near straight H-bonds, it was found that two
configurations had significantly different identity distances along the
polypeptide chains : 7*0 A for the antiparallel sheet and 6'5 A for the
parallel pleated sheet. Thus the two sheets might well serve as models for
silk fibroin (axial repeat distance 7 A) and /8-keratin ( ~6'5 A), respectively.
Among the fibrous proteins the silk fibroins are the most promising for

180 KERATIN AND KERATINIZATION

structural analysis. Chemically they are rather simple. Bombyx silk and
Tussah silk fibroin are alike in that glycine, alanine, serine and tyrosine
comprise over 90% of the amino acids. The classical studies of Meyer and
Mark and Astbury, which have been referred to above, made it clear that
in certain fibres, yielding the j3-type pattern with a fibre period of 7 A,
the polypeptide chains must be almost fully extended. The completely
extended chain is calculated to have a repeat spacing of 7*27 A which is
significantly greater than that actually found.

Marsh, Corey and Pauling (1955) have made a thorough comparison
between the diffraction data and the predictions based on the anti-parallel
sheet model and have established beyond question that this structure
exists in Bombyx silk. In their model the chains of fibroin consisting of (in
the main) alternate glycine and alanine residues, are arranged so that the
CH3 groups of the alanines stick out entirely on one side of the sheets.
Pairs of sheets are thus separated alternately by distances of 3*5 A and
5*6-7 A, spacings which figure prominently in the X-ray patterns. There
are minor points still disputed. Peptides comprising the bulk of the pro-
tein and containing, apparently, long sequences of glycyl-alanine pairs
(Lucas et al., 1956 and 1958) and sequences in which serine substitutes for
glycine or alanine residue (Waldschmidt-Leitz and Zeiss, 1955) have been
separated from digest of Bombyx fibroin. These glycyl-alanine clusters
probably join together to form the crystallites. The solution of War-
wicker (1954) is essentially the same.

There is no place in these models for the more bulky side chains, such
as that of tyrosine, which together constitute some 18% of the protein. It
is supposed therefore that they are inserted at intervals (regularly or at
random) in the chains, thus buckling the sheets and causing some of the
diffuse reflections always present. Since in the case of keratins the residues
with larger side chains form a far greater part of the protein, distortions
arising from such causes must be far more common and perhaps account
for the larger amount of diffuse scattering from these materials.

Tussah silk yields a /^-pattern distinct from Bombyx silk and is isomor-
phous with polyalanine (Bamford et al., 1954). It contains more alanine
than Bombyx and thus, assuming a similar antiparallel-sheet arrangement,
both faces must contain methyl groups and all pairs of sheets must be
similarly spaced at 5*3 A. Kay et al. (1956) have shown the existence of
" cores " of peptides containing long sequences of alanine.

No attempt has yet been made to discuss rigorously the diffraction data
of the ^3-keratins of stretched hair and wool in terms of the pleated-sheet
models and it would seem, in view of the small amount of information to
be obtained from the actual patterns themselves, that the attempt would
not be immediately fruitful. The general similarity of all these patterns
leaves little doubt, however, that in the crystalline regions of the materials

MOLECULAR AND MACROMOLECULAR STRUCTURE

181

yielding them, structures based on pleated sheets are present. There
are other silks with somewhat bulkier side-chain composition and hence
larger side spacings which approach more closely to the actual conditions
prevailing in the keratins. A knowledge of the particular peptide sequences
in keratins which participate in crystallite formation is still wanting.

Helical Configurations — The <x-helix

Helical structures were first proposed by Huggins (1943) and Taylor
(1941); those described by Pauling, Corey and Branson are, however,
more precisely specified and are thus capable of quantitative evaluation.

Fig. 75. Diagram of the oc-helix of Corey and Pauling (reproduced by
permission).

They discovered two helical (a and y) arrangements, and since that time
Low and Baybutt (1952) have described another. Only the a-helix has been
extensively studied since it does seem to provide the required structure.

182 KERATIN AND KERATINIZATION

The a-helix is based in the first place on geometrical requirements
deduced from the known structures of small peptides (see above) (Fig. 73).
When a model of a single polypeptide chain, built according to these
requirements, is formed into a spiral, it is found that one very compact
formation results, in which there are 3*6 amino acid residues per turn of
5*4 A so that each residue occupies 1*5 A of the length of the helix (Figs.
75 and 76). The stability of the helix is assured by the formation of
multiple internal H-bonds, and Donohue (1953) has shown that of the
several possible helices the a is the most stable.

DIMENSIONS OF AN or HELIX

Rise per
Residue

26° ^&rr

JL 3

J„

5TH
TURN

4TH
TURN

3 RD

TURN

2 ND
TURN

18 Residues
27 A

5.4 A Pitch 1 1 ST
3.6 Residues [TURN

Fig. 76. Diagram of an a-helix indicating the 1 *5 A rise per residue which

gives rise to the characteristic axial X-ray spacing and the relation between

the pitch 5-4 A (36 residues) and the 54 A spacing. (After Corey and

Pauling.)

The a-helix was put forward in the first place as a proposal. As a direct
experimental test for the presence of such a helix, Perutz (1951) pointed out
that such a structure should give a strong meridional reflection of 1*5 A
corresponding to the segment of the helix occupied by one residue (Fig.
76). He found that, in fact, this reflection was present in the patterns
given by hair, quill, muscle, various other proteins and polypeptides
(Perutz and Huxley, 1951). This reflection had also been recorded earlier
by Macarthur from porcupine quill (1943). Since no other of the chain
configurations proposed for the a-proteins can give a 1*5 A reflection, the

MOLECULAR AND MACROMOLECUL AR STRUCTURE 183

occurrence of this particular spacing has become diagnostic for the
a-helix. Astbury et al. (1959) proved that the 1-5 A reflection arose from
the same structure as the 5-1 A by showing that both reflections were
similarly altered by the small (2%) reversible extensions of fibres.

Very detailed comparisons have been made between the observed X-ray
patterns of certain synthetic polypeptides and that predicted from the
a-helix by Bamford et al. (1956), Yakel (1953) and Brown and Trotter
(1956) and they have demonstrated that the a-helix forms the structural
basis of these polypeptides. The patterns given by fibres of these materials
are often of remarkable perfection and far superior to those of the natural
fibrous proteins. It seems probable that the study of a synthetic polypeptide,
based more closely on the naturally-occurring sequence of residues in
the crystalline regions of the keratins may lead most easily to further
advances in the understanding of the natural structures.

When the patterns of the a-synthetic polypeptides (see Bamford, Elliott
and Hanby, 1956) are compared with those of keratin and muscle, the
resemblance is striking and leaves little doubt that the natural structures
are based on the helix. Differences are equally striking, and these have
been emphasized by Bamford and Hanby (1951). The main characteristics
of the a-pattern are the strong meridional arcs of 5*15 A and 1*5 A, and a
group of spacings centred around 10 A at or near the equator (Astbury and
Woods, 1933; Macarthur, 1943). When these observational facts are
compared with the pattern to be expected from a crystal of hexagonally-
packed (Fig. 78) a-helices parallel to the axis two difficulties appear:

(1) The a-helix would give a strong layer line at 54 A but the intensity
on the meridian would be zero. In fact we find the strong 5*15 A
arc on the axis.

(2) If the centre of the broad equatorial reflection (9-8 A) is taken as the
(10-0) reflection of a simple hexagonal lattice the calculated density
for a-keratin is too low.

Coiled coils and a-filaments

An attempt to resolve these difficulties by suggesting that the whole
helix (minor helix) might be twisted into a super helix (major helix) or
coiled coil, has been made by Pauling and Corey (1953a), and by Crick
(1952). Both proposals involve tilting the a-helices to form the super-helix
or a coiled coil (Fig. 77) with a pitch angle of about 18° giving a projection
on the axis with a periodic variation in density at intervals of 54 cos 18° =
5*1 A. According to Crick the reason for the deformation may be found in
the difficulty of packing the side chains projecting from the non-integral
helix. If a simplified model in which the side chains are represented by
knobs is taken and, on a piece of paper wrapped around it, the position of
the knobs is marked, a pattern is found on unwrapping into which a second

184

KERATIN AND KERATINIZATION

helix can only fit when it is tilted ^ 20° to the first. Crick suggests a
three-strand rope to give the 198 A longitudinal period but this does not
account for the strong 27 A reflection on the equator. The high density
(1*3 g/cm3) remains a difficulty.

Pitch of

large
helix

•^'.Pftch of
V .small helix

la)

(b)

Fig. 77. Illustrating the curving of the a-helix into a super helix (a) and
(b) and two possible combinations of super helical structures to give com-
pound helices (coiled coils), (c) the seven-strand cable and (d) a three-
strand coiled coil (Pauling and Corey, 1956. Reproduced with permission).

Pauling and Corey reached a similar conclusion on the grounds that an
a-helix composed of a repetition of different amino acids in a regular
pattern would not have a straight axis but one distorted into a large helix.
They suggest that a radius of 10 A for the a-helix would permit six

MOLECULAR AND MACROMOLECULAR STRUCTURE 185

molecules to twist around a seventh straight a-helix to form a " seven-
strand cable " (Fig. 78). Further, a closely-filled structure of hexagonally-
packed seven-strand cables with individual a-helices occupying the
interstices, would improve the fit to the density (Fig. 78). This solution
gives a good fit for the many strong reflections (Pauling and Corey, 1956)
but is very much a crystallographers' solution and lacks support from other
directions. A subfilament of three a-helices with irregularities about 27 A
apart (" segmented three-stranded cable ") is indicated by the most
recent analysis of the keratin diffraction data (Fraser and MacRae, 1961).

:...

,5, .< pj:, ■"%

•-•-•.. ..—••. &?&'■ / \ #i"i-"A- '" "••••• \

r}ks?A ?"y\ M&! \

; :V:rr-X> '""'v -: :.:.';v;'-c\ -s ••'

"<$*&*'

Fig. 78. A suggestion by Pauling and Corey for the macromolecular

structure of a-keratin in terms of hexagonally-packed seven-stranded

cables in the interstices of which are packed single helices (C) to correct

the density deficit.

The a-helix of diameter (c. 10 A) is a structural element rather on the
small side for current electron microscopy. With the concept of super-
helices we enter a domain of dimensions which should be accessible to
microscopy. The diameter of the seven-strand cable is 20-30 A. Such a
dimension has not yet been observed in cross-sections of keratin in hair
and wool. The actual cross-section of what seem to be elementary fila-
ments of keratin in hair and skin is of the order of 60 A (Figs. 79 and 80)
and for myosin a similar value is reported. These figures would seem to
demand much more than seven-component spirals, the width of the
cylindrical filament being approximately that of six a-helices suggesting a
rope of between twenty and thirty component helices as indicated in Fig. 79.

186 KERATIN AND KERATINIZATION

When models of structures of this size are attempted either by forming
concentric shells of a-helices or by arranging these on helical sheets,
difficulties are met in effectively filling the centre of the filament without
some disorder (Fig. 79). Perhaps an indication that the nature of the

helix =1

10 20 JO A

Fig. 79. Relation between the a-helix and the a-filament. The a-helix
(diameter 10 A) is the structural unit deduced from model building and
from the data afforded by actual X-ray patterns. The a-filament l.h.s.
(diameter <— ' 60 A) is the smallest structural unit observed electron-
microscopically. The filament could be a twisted yarn of a-filaments
with a prevailing angle relative to the axis of about 20° but the internal
arrangements are obscure.

2000-l0,000A|
I0A 60A £| I 0-2-1/*

(0) (b)

Fig. 80. Relative dimensions in cross-section of the molecular, macro-
molecular and histological units of hair keratin, (a) the a-helix of
diameter 10 A {molecular level); (b) the a-filament diameter ~ 60 A
whose internal structure in terms of a-helices is not known {macro-
molecular level) ; (c) the fibril visible in the light microscope composed of
large numbers of filaments {histological level). Cross-sections of fibrils
are to be seen in Figs. 102 and Plates 15 and 16.

packing alters (or that an element of disorder enters) is given by the fact
that the centres of the filaments seen electron-microscopically are some-
what more stained than their peripheries (Rogers, 1959). There seems no
structural principle that we can appeal to limit the number of helices in a
filament, yet the evidence is that the filaments are a uniform and definite
structural element.

MOLECULAR AND MACROMOLECULAR STRUCTURE 187

The flagellum protein of certain flagellated bacteria is of the a-type and
each single flagellum is a whip-like thread whose dimensions are of the
same order as those of an a-filament in hair of which it may be regarded
as a structural analogue. Burge (1960) has inferred from the X-ray pattern
of isolated flagella (see Fig. 86) that the hexagonally close-packed bundles
of a-helices must consist of only a small number (3-7) of helices. There
must then be several of such bundles in parallel to form a structure as
large as the whole flagellum ( ~ 100 A) suggesting again the existence of
sub-filaments within a main a-filament.

The Organization of <x-Filaments into Larger Structures

On passing from the molecular level to the macromolecular and histo-
logical levels we enter the domain (see Fig. 1) where microscopy is able to
provide definite answers to structural problems. While these matters will
be more fully considered in the next chapter it will be useful here to discuss
some geometrical aspects of the larger structures. It is perhaps advisable
to emphasize the succession of structures of increasing size with which
we are concerned. In hair, for example, a-helices ( ~ 10 A diameter),
a-filaments (60-80 A diameter) and fibrils (0-05-1/n diameter). These are
depicted diagrammatically in cross-section in Fig. 80.

Various and somewhat speculative schemes can be advanced to explain
the packing of the a-filaments as it is actually observed in the fibrils of hair
(Fig. 102, p. 247), wool and skin (Fig. 98, p. 225). The geometry of the
packing is very variable : in skin cells, an extreme case of irregular packing,
the filaments cluster in irregular-sized sheaves with no constant orientation
relative to each other but on the whole tending to lie in the plane of the
flattened cell (Fig. 99, p. 229). In hair, wool and quill, we find definite
macrofibrillar bundles of filaments oriented parallel to the fibre axis (Fig.
102, p. 247). In cross-section these show some variety of appearances
ranging from good hexagonal packing to spirals (see Plates 15 and 16)
the latter appearing more common in the middle stages of development of
the hair. Stacks of flat sheets may occur in the a-protein of the mantis
ootheca (see p. 205).

If an a-filament (diameter < 100 A), the constructional unit of these
formations, is not circularly symmetrical but is polarized as suggested in
Fig. 81, the energy conditions governing the recruiting of new members to
an already-formed aggregate might well favour the development of flat
sheets, Fig. 81 (b) or (a) helical sheets. The stacking of one sheet against
another could be governed by conditions of the sort discussed by Crick
in connexion with the association of the smaller units, the a-helices, to
form filaments. That is, we may assume that the filaments themselves also
have helically-fluted surfaces (or lines of special attraction) in this case
arising from their construction as multi-stranded cables of a-helices, and

188 KERATIN AND KERATINIZATION

that compact arrangements can result either from flat sheets lying over each
other at a suitable angle (Fig. 87) as Rudall has proposed for the ootheca
protein (p. 205), or as in hair, where they lie on the circumference of
concentric circles (or helically-wound sheets) with an appropriate increase
in the angle of pitch between successive layers.

The degree to which these initial aggregations would persist into the
finally-hardened state could easily depend on the speed with which the
final changes are effected. For, if we regard the cystine-rich matrix protein
(p. 248) as an interpolation between the filaments (or as an addition to

0GX3GX5

Fig. 81. Illustrating the lateral aggregation of rodlets (a-filaments)

to give (a) a spiral compound or (b) a sheet. The rodlets are here shown

in cross-section.

their surfaces) this could, by rapidly cross-linking, maintain an earlier
condition; on the other hand, were hardening more slow, it would facilitate
a rearrangement of the filaments into the most compact condition,
hexagonal packing. This may be pictured more plausibly by thinking of
the filaments as being separated by a viscous but fluid medium which both
reduces the close interfilamentous contacts responsible for the earlier
packing and also permits the movements of readjustment. Fraser et al.
have appropriately likened the matrix in keratin to the interfilamentous
water of other fibrous systems (1959). In skin we may suppose these
changes occur too rapidly to permit of much readjustment. Rogers (1959b)
has shown that hexagonal packing prevails in the more stable fraction of
wool (paracortex).

The a-/? transformation in terms of the a-helix

One of the commendable features of the a-helix is that, while it was
developed primarily to satisfy structural principles derived from a study
of simple peptides, etc., by crystallographic means, it has also, on being

MOLECULAR AND MACROMOLECULAR STRUCTURE

189

stretched into a straight-chain configuration, an extensibility of the right
order ( ~ 120%) to satisfy the requirements of the a-j8 transformation as
envisaged by Astbury (p. 174). No detailed proposal of the nature of
this transformation in terms of the helix has been given and there are
certainly formidable unsolved problems relating to side-chain movement
and of inter- and intra-chain cross-links to be considered. These become
increasingly difficult when multi-strand cables are introduced.

An inspection of models shows that when two helices of the same sense
are joined at more than one point uncoiling is physically impossible without

00 -helix?)
stretched

a -helix
(ijnstretched)

Fig. 82. Difficulties encountered in extending a bundle (a-filament) of
a-helices to yield a j8-fibril. The a-helices must be supposed to uncoil in
some manner and to reform to yield a jS-type structure in the stretched
filament. If rotation is restricted in any way this would seem geomet-
rically incompatible with the filament retaining its identity.

rupture. If segments in which the sense changes from right to left-hand are
permitted, as Lindley (1955) has proposed for insulin, taking advantage of
the bends introduced by proline residues, straightening may be possible.
In multi-stranded filaments, it would seem that the strand structure must
persist during extension and again permanent cross-links must be severely
restricted or the individual helices will lack the mobility to open up.
Those who are convinced that coiled helical structures exist in keratin may
find in this a further argument for excluding sulphur cross-linkages from
the helical (fibrous) phase.

Figure 82 has been constructed to emphasize these difficulties assuming
that the electron-microscopically visible filaments are bundles of a-helices.

190 KERATIN AND KERATINIZATION

Individual helices must have freedom to uncoil and the j9-form remains a
multi-stranded structure in which the jS-chains may be conceived to form
stacks of pleated sheets which are slightly twisted to form concentric,
closed cylindrical pleated sheets (|8-helices) such as has been recently
proposed for feather keratin (p. 209).

It would seem that the structure at the macromolecular level of the forms
of a-proteins found in hair is far from settled. The purely theoretical
discussion given by Lindley (1955) on model making and the packing of
helices shows that the possibilities are far from being exhausted. Lindley's
models are based strictly on the dimensional criteria of Corey and Pauling
but, by skillfully exploiting the possibilities of helices of left-hand and
right-hand sign and the discontinuities introduced by the presence of such
residues as proline and cystine, he showed that a number of unexpected
packings could be achieved.

The non-crystalline fraction

The structures we have been discussing up to this point are those of the
crystalline material occurring in a keratinized tissue, i.e. that part of the
tissue in which the molecules are sufficiently well arranged to yield a
definite X-ray diffraction pattern in the form of discrete spots. There is no
question but that a large part of these tissues is not in such a well-organized
form and that, moreover, many of the important properties of the materials,
such as their elastic and chemical behaviour towards mild reagents, is
influenced by this fraction. It is not very useful to regard the keratins
as perverse molecules which may some day be persuaded to assume a
perfectly crystalline form and give the crystallographers their chance.
Ordered crystalline regions certainly exist in the large masses of hardened
protein, and for these precise structures may be described; but equally
certainly disordered non-crystalline regions exist too, and these imperfect
regions must be regarded as an essential part of the whole formation since
they confer on it certain properties required for a performance of its
biological role. For these reasons it is necessary to consider as a separate
problem the type of structure which prevails in these regions and to
estimate what fraction it forms of the entire tissue. For this purpose there
are available in addition to X-ray diffraction techniques, various other
methods of a physicochemical nature. It should be emphasized that we
are not considering here the non-keratinous constituents (p. 270) but
rather that portion of the keratin itself which, since it does not contribute
to the fibre-type X-ray pattern, may be referred to as " non-crystalline",
a term more exact than " amorphous".

The amount of this material, usually expressed as a crystalline/amor-
phous ratio, may well vary from cell to cell and from tissue to tissue. The
estimates of its value as found in the literature are mostly for wool or hair

MOLECULAR AND MACROMOLECULAR STRUCTURE 191

and vary widely. The methods employed do not in theory all measure the
same quantity and even when the actual measurements themselves are
precise, their theoretical bases are much in dispute. An attempt has been
made to make sense out of this unsatisfactory situation by Alexander and
Hudson (1954) and reference may be made to their discussion since it
would carry us far beyond the present intention to attempt to review the
work here. The broad conclusion of the physicochemical work based on
the penetration of various molecules into fibres is that only a small pro-
portion of the fibre (10-30%) is inaccessible to small molecules and can in
this sense be regarded as crystalline. " Inaccessibility " is a concept not
necessarily equivalent to " crystallinity " in the X-ray sense, since
crystallites capable of giving a Bragg reflection may be partly or wholly
penetrated by some reagents.

From a suitable X-ray pattern an estimate of the crystalline/amorphous
ratio may also be calculated from measurements of the amount of radiation
scattered in the form of discrete reflections and that scattered into diffuse
reflections. In favourable cases, e.g. rubber, this theory has been used to
determine the ratio in the stretched and unstretched state, and has helped
to confirm that the long-range elasticity is based on the stretching of
randomly-coiled chains which may then crystallize when held in the
stretched condition. A similar study carried out on a suitable keratin fibre
would be of great value. However, the vague patterns render the project
well-nigh impossible since there is considerable overlap of diffuse and
discrete reflections.

An inspection of a typical a-pattern (Plate 1) reveals the presence of
considerable amounts of diffuse scattering due to amorphous material.
This is seen principally as a broad diffuse halo centred about an average
spacing of 4-5 A (Astbury and Street, 1931; Astbury and Woods, 1933).
No quantitative estimate of the amount of this reflection has been made.
Qualitatively it appears considerable and seems to suggest that more
than half of the fibre substance is amorphous. During the stretching of a
hair, the appearance of the pattern is little affected in the range of 0-20%
extension; after that certain reflections specifically associated with the
a-type structure fade, although there is no perceptible change in the
distribution and intensity of the diffuse scattering. When a fibre which has
been stretched beyond 50% is steamed, in order to induce recrystallization
of the /S-form, the /^-pattern appears against a background of diffuse
scattering very similar in intensity to that noted in the unstretched fibre.
This shows that stretching does not itself produce an increased amount of
the fibre substance in an ordered form (a or /S) and suggests that the same
well-ordered regions give rise both to the a- and the /S-pattern according to
their extension.

Measurements of the birefringence of hairs (Barnes, 1933; Mercer,

192 KERATIN AND KERATINIZATION

1949c) (p. 10) show that the original value is of the order of Nu — Nx =
0*01, and that it increases in the range of stretching from 0 to 40% sug-
gesting an increased ordering of the structure. It reaches its highest value
in fibres stretched to 40-50% extension and held under tension, and falls
when these are steamed to relax the tension and to produce the /2-forms
causing " set." From this it may be deduced that set involves a decreased
average molecular alignment parallel to the axis of main chains plus side
chains and that the /?-form itself is not necessarily more birefringent
than the a-form.

All methods combine to suggest that a very considerable fraction
of the material is in a somewhat disordered form. Such non-crystalline
regions could be pictured in various ways. Astbury and Woods (1933)
originally suggested that the non-crystalline regions approximate to the
crystalline in structure and thus that conclusions drawn from the dif-
fraction pattern of the crystallites are applicable broadly to the whole
keratin system. This assumption necessarily underlies their interpretation
of fibre elasticity (p. 174) since all " phases " must be capable of the same
ultimate extensibility of 100% if one is to infer molecular extensibility
from whole fibre extensibility. The sequence of events during extension
is still under investigation by X-rays. (Bendit, 1957 ; Skertchly and Woods,
1960).

Whatever the solution may be, it seems not unlikely that, as in the
fibroins, the poorly-organized regions will be associated with side chains
which are difficult to pack. Fibroin crystallites contain predominately the
shorter side chains which pack readily ; the acids with longer chains seem
to be excluded from these crystalline clusters and their interpolation
elsewhere impairs the chain alignment and thus introduces a non-
crystalline region. There are many more amino acids with bulky side
chains in keratin, a fact which might be associated with the lower crystalline/
amorphous ratio, and the crystalline regions might well be those where
clusters of the smaller acid residues (glycine, alanine, leucine, serine)
occur. The isolation from a natural keratin of peptides containing such
sequences, as has been analogously effected with fibroin, would be a
valuable indication. Large side chains do not in themselves preclude
crystallization. Among the synthetic polypeptides containing a-helices,
excellent ordering is possible even with large side chains since the homo-
genity favours packing even when the side chains are long. The helices are
spaced further apart in this event (Bamford et al., 1956). Certain natural
silks contain crystalline regions with long side chains (Warwicker, 1959).

The presence of certain residues in the crystalline region may be doubted
on other grounds than size. Model building shows that most residues have
no special importance in the sense that their side chains do not influence
the possible configurations of the main chain. Four residues — those

MOLECULAR AND MACROMOLECULAR STRUCTURE 193

derived from proline and hydroxyproline, cystine and glycine — have,
however, special consequences. Proline actually constrains the shape of
the polypeptide chain by introducing a " bend." Thus a chain containing
a proline residue cannot be straight and an a-helix cannot be maintained
at segments where a proline or hydroxyproline residue is inserted;
although Lindley (1955) has pointed out that, in conjunction with appro-
priate neighbours, it can change the sense of the helix from r.h. to l.h., a
feature which could be of importance. Glycine is unique in having no side
chain. It is a " space saver " and may permit a more compact packing of
adjacent chains. It is common in collagen (a frequency of 1 in 3) and its
smallness plays a special part in the construction of the triple helix assumed
by that protein. Cystine, by virtue of its power to form intra- or inter-chain
cross-links, would reduce drastically the mobility of any polypeptide
system.

It is significant that most chemical reactions with wool and hair have
little effect on the X-ray diffraction pattern unless they are of the type that
leads to a dissolution of the H-bonds. Reactions not markedly affecting
the pattern include : absorption of water, neutralization of acid or basic
side chains, nitration, iodination, reduction and substitution of the cystine
bridges. It would seem then that many residues are either excluded from
or are to be found only on the surfaces of the crystallites.

The picture originally (1933) presented by Astbury and Woods of the
crystalline and amorphous regions, stretching partly in series and partly
in parallel when the fibre is stretched, still commands general acceptance
in spite of a considerable evolution of views concerning the actual <x-
structure itself. There have been other proposals based mainly on
different interpretations of the course of events in stretching (see p. 176).
For example, Peters and Speakman (1949) and Burte and Halsey (1947)
have envisaged the possibility, even at zero extension, of an equilibrium
between portions of the molecular chains in the a- and ^-configurations in
the non-crystalline region and have developed this concept to give a
quantitative description of limited aspects of the elastic behaviour of
swollen wool. The existence of any considerable fraction of the molecules
in an extended form at zero extension would invalidate the basic assump-
tion of Astbury and Woods that ultimate extensibility of the whole fibre
(100%) is also that of the component molecular chains. According to
Elliott (p. 199) there is evidence from infra-red absorption spectra of
/3-configurations in unstretched hairs.

In terms of the " matrix-plus-filament " model developed in Chapter 6
a certain amount of the diffuse scattering of X-rays must be produced also
by that fraction of the material described as y-keratin, which is considered
to exist outside of, and to envelop the fibrillar component proper. The
polypeptides in this region may be pictured as being so heavily cross-linked

194 KERATIN AND KERATINIZATION

by the frequently-occurring disulphide bridges (a frequency of from 1 in 2
to 1 in 3) that crystallization is impossible and that, as in liquids, a
diffraction pattern of diffuse haloes is produced simply because certain
interatomic spacings occur with a higher frequency. This " liquid-like
order " of the non-crystalline fraction is to be distinguished from an
unoriented ^-arrangement. The /3-form is a well-defined structure
maintained by H-bonds of a perfectly definite length. The y-structure is
pictured as being structurally amorphous since the packing of the chains
in either a- or /?-forms is hindered by either cross-links or awkwardly-
shaped side chains. A certain amount of the non-crystalline material
within filaments themselves may also be pictured either in this form or as
a disordered a-structure, since all a-type protein patterns are characterized
by the same broad halo of average spacing of 4*5 A irrespective of their
fine histological structure. This is the case for natural fibres, keratinized or
not, and also for regenerated protein fibres devoid of histological or fine
structure. When no crystalline material is present, as in some regenerated
fibres, only the broad haloes appear. It seems probable that, to a degree
limited by the steric hindrance of side chains and cross-links, short
segments in the a-fraction approximate more-or-less closely to the a-helix
favoured on energetic grounds.

The radial distribution patterns to be expected from the unoriented
a-helix have been calculated by Donohue (1954) and some comparison with
the experimental scattering curves made by Arndt and Riley (1955), but
according to Kendrew and Perutz (1957) the radical distribution function
is not a sensitive test of configuration.

There appears to be some connexion between crystallinity and density.
The densities of horn, wool and porcupine quill (1-28, 1-302 and 1*32)
are in order of the degree of crystallinity; but human hair, far less
crystalline than quill, has much the same density (1*317). This discrepancy
may rise from the rather higher cystine content of the hair (Fraser and
MacRae, 1957).

Other methods of determining chain configuration

Optical Rotation and Rotary Dispersion

During the last few years there has been a renewed interest in the
measurement of optical rotation [a]A and its dependence on wavelength
(rotary dispersion) as an added tool for the evaluation of configurational
changes and for estimating the " helical content " of protein preparations.
While little work has been attempted on keratin preparations, a good deal
of information has been gathered concerning proteins in general and of
various well-characterized a-type fibrous proteins which is relevant to the
keratin problem.

MOLECULAR AND MACROMOLECUL AR STRUCTURE 195

The theoretical basis for the interpretation of the measurements is still
developing (Leach, 1959) but with the increase in data sound, partly-
empirical methods now exist. The specific rotation [a] is defined as:

[a]i ~ CpL

where a is the angle of rotation of the plane of polarized light, C the
concentration in grammes of solute in 100 milliequivalents of solvent, p
the density, and L the path length. The change of [c
to as dispersion and is described by the classical equation of Drude :

l-ja - A2 _ A2c

(k and Ac constants of the system).

The Drude equation is followed at high wavelengths in solutions of pro-
teins; at low wavelengths the dispersion becomes " anomalous " and the
anomalous contribution is thought to be supplied by the portion of the
molecular chains folded into helices.

Anomalous dispersion is usually described by a modification of the
Drude equation due to MofRtt (1956):

/100 n* + 2\ \ aQ\l

(A2 -A*)5

M0 is the residue weight, a0,b0 and A0 are constants. The second term in
square brackets represents the anomalous contribution due to the helical
structure. It is customary to plot l/[a] against A2 to test the fit to the
Drude equation which is usually good for protein solutions at longer wave-
lengths and allows an estimate of Ac and k. Various methods exist for
estimating the amount of folded material. In general globular proteins in
aqueous solution have specific rotations (Na Z)-line) of the order of from
—25° to —80° which increases to —80° to —120° when unfolded by urea.
Assuming that the change in [a]^ on unfolding is a linear function of the
number of residues unfolded, an estimate of the percentage of folded form
is possible if the values of [<x]D in the fully helical and fully unfolded
conditions are known. These values have been obtained from polymers
and certain proteins where independent means of knowing the helical
content exist.

The estimated helical content of most globular proteins is only 15-50%.
For certain a-type fibrous-muscle proteins (Review by Leach, 1959) the
figure is higher (50-90%). This high figure is thought to be due to their
low content of proline residues and disulphide bonds which restrict the
formation of the a-helix.

196 KERATIN AND KERATINIZATION

Woods (1959) found that the " low sulphur extracts " of wool resemble
the synthetic polypeptides in their optical behaviour. In 8 M urea the
molecule is in a completely random form with [a]D = — 105°; in some
organic solvents [a]D is approximately zero, i.e. the molecule is largely
helical. In aqueous solution [a]^ is of the order of —60° which, according
to some forms of calculation, would mean a helical content of 30-40%.
Undoubtedly the helical content of a keratin will be restricted by its
proline and cystine content. When there is more than 8% proline dis-
tributed along a polypeptide chain (wool keratin 9*5% and feather keratin
10%) it is possible that there will be no segments long enough to be stable
in the helical configuration. Very high proline contents may favour
another structure such as that proposed for feather by Krimm and Schor
(p. 208).

Infra-red Spectra and Structure

Infra-red spectra arise from changes in the vibrational energy ot mole-
cules produced by the absorption of infra-red radiation. Their value in
structural studies is due to the fact that the absorption effects the move-
ment of nuclei in the field of the interatomic binding forces, and the
examination of large numbers of substances of known structure has shown
that certain frequencies are associated with particular valency bonds
whose presence in other compounds of unknown structure may thus be
deduced from their spectra. Fortunately for protein studies, bonds of
hydrogen atoms with other atoms give characteristic absorptions and their
study provides a method of investigating associations between groups
which are mediated by H-bonds.

When the molecules are oriented as in crystals or in fibres it is also
possible to obtain information concerning the direction of some valency
bonds by using polarized infra-red radiation (Ambrose and Elliott, 1951
and 1952). The absorption coefficient is proportional to the square of the
cosine of the angle between the E- vector of the radiation and the direction
of the rate of change of the dipole moment of a normal mode of vibration
of the molecule. If the bond associated with this mode of vibration is
already known, it may be possible to infer its direction in the fibre (see
Fig. 83).

The bands in infra-red spectra are usually given as wave-numbers,
rather than frequencies, where the wave-number is the reciprocal of the
wavelength in centimetres (cm-1); and the spectra are presented graphi-
cally by plotting wave-number against optical density:

intensity of incident radiation
s 10 intensity of transmitted radiation
The ratio of the optical density, measured first with the electric vector

MOLECULAR AND MACROMOLECULAR STRUCTURE

197

parallel to the fibre and then at right angles, is called the dichroic ratio.
Infra-red spectra have proved fruitful in the study of the synthetic poly-
peptides particularly in the hands of the Courtauld group and reference
may be made to their book (Bamford et at. 1956).

The frequencies associated with the CO and NH groups may show small
variations depending on whether they are H-bonded in an inter- or
intra-molecular configuration. The wave-number associated with a group

i-5r

0 5

1500 1600 1700

Wav e number (cm77)

Fig. 83. Infra-red spectra of a-keratin (porcupine quill), upper curves,
jS-keratin (swan quill), lower curves. Full line, light polarized perpen-
dicular to fibre axis; broken curve, light polarized parallel to fibre axis.
(Ambrose and Elliott, 1952. Reproduced by permission.)

bonded in a j8- configuration is somewhat lower than in the other forms.
These relationships have been particularly studied by Ambrose and
Elliott (1951 and 1952) who deduced a correlation between structure and
absorption from a study of certain synthetic polypeptides known on X-ray
grounds to exist in more than one form. In Table 12 some key frequencies
as determined by them are given. For example, poly-1-glutamicbenzyl
ester shows amide absorption as a single band at 1658 cm-1 and the X-ray
diffraction photograph shows it to be in the a-form. When the methyl

198 KERATIN AND KERATINIZATION

ester (normally giving two bands at 1658 cm"1 and 1628 cm"1) is cast from
formic acid, a pure /3-form results giving only a band at 1629 cm-1.

Generalizing from these observations the Courtauld group proposed
that as a simple test an a-configuration could be diagnosed, for example,
by a carbonyl stretching frequency round about 1665-1660 cm-1 as
opposed to the corresponding bond of the /^-configuration at about 1630-1.
Not all authorities accepted this test (Darmon and Sutherland, 1949;
Sutherland, 1952) and recent reports by Elliott, Hanby and Malcolm
(1954, 1956 and 1958) show that these workers themselves have abandoned
its rigid application.

Table 12. Wave-Numbers (cm-1) of Infra-Red
Absorption Bonds in a- and /^-Configurations.

CO stretching
NH deformation
NH stretching

a-configuration

1650-1660
1540-1550
3290-3305

/^-configuration

1630 (J.)*
1520-1526 (||)
3287-3301 (J_)

* The dichroic sense in the jS-form is indicated by the symbols (]|)
parallel and ( J_) perpendicular to the fibre axis.

Recently developments (p. 194) have shown that a-helices can be
diagnosed more certainly by observations on the dispersion of the optical
rotation, and Elliott, Hanby and Malcolm, having measured the infra-
red absorption spectra and the optical rotation of several polypep-
tides and protein films, have come to the conclusion that the carbonyl bond
near 1660 cm-1 is not necessarily associated with the a-helix. They
describe several polypeptide preparations (alkaline salts of poly-L-glutamic
acid and copolymers of L- and D-polyalanine) and a soluble fibroin in
which the dispersion of the optical rotation shows conclusively that
a-helices are absent, but which also have the carbonyl absorption bond
near 1660 cm-1 earlier proposed as a test for the presence of the a-form.
The present position would seem to be that the /^-configuration shows a
characteristic absorption at 1630 cm-1 and that almost any departure from
this ranging from an a-helix to a random coil may show a bond at 1660
cm-1.

The spectra of proteins of course show many other bands to be expected
from our knowledge of chemical composition of the proteins, but identi-
fication is not always easy (Bellamy, 1954). The measurement of dichroic
ratios is valuable in excluding some proposed structures. The dichroic
ratios found are small and, when these are found in specimens giving a

MOLECULAR AND MACROMO LECU L AH STRUCTURE 199

well-oriented X-ray pattern, a low crystalline amorphous ratio is indicated.
The spectra in this case may provide data concerning chain configuration
in the important amorphous phase which is not accessible to X-ray
methods. In particular the a- and ^-configurations may be recognized by
their distinct dichroic effects as demonstrated in several polypeptides.
Ambrose and Elliott (1951) examined sections of elephant tail hair (a-
keratin) and obtained spectra such as shown in Fig. 83. The NH and CO
directions seem to be more parallel to the fibre axes (parallel dichroism),
although the dichroic ratio is very low. Robinson and Ambrose (1952)
claimed that this evidence excludes the Astbury-Bell model.

A structure based on the a-helix would have a high parallel dichroism,
and the low figure found is therefore only possible if we assume a large
admixture of amorphous material. Parker (1955) modified the amorphous
regions by replacing the H atoms by D atoms and thus displaced the
(OH and NH) bands to much lower wave-numbers enabling the dichroism
of the crystallite to be observed alone. The parallel dichroism rose from
1*5 to 4*5 showing more certainly that the low value usually obtained is
due to the large fraction of amorphous material and lending support to
structures of the a-helix type.

Elliott (1952) has also concluded that there are significant amounts of
|8-keratin in the amorphous phase of unstretched hair and of the a-form
in hair stretched 100%. This deduction is now uncertain. It was held that
these observations show that care must be exercised when attempting to
correlate molecular and whole-fibre extensions. The chief value of these
infra-red observations remains now that they allow some investigation of
the often large amounts of non-crystalline constituents not revealed by
X-ray diffraction and indicate whether or not this is in the /3-form with
some measure of its orientation.

The important fact that reversible a-/? transformations can be obtained
from purely synthetic polypeptides was established by the Courtauld
group using indications provided by infra-red absorption spectra (Bam-
ford et ai, 1956). For examples, fibres of a-poly-L-alanine stretched in
steam give a highly-oriented /3-pattern and somewhat similar results are
obtained with poly-a-amino-w-butyric acid. Reversible conversions from
one form to another are also obtained by immersing specimens in various
swelling solvents. Concentrated formic acid often produces a /3-form and
chloroform or w-cressol may reconvert to the a-form.

formic acid

m-cresol

The conversions are reversible and there is no loss or degradation of
material. This demonstration of the influence of the solvent on the

200

KERATIN AND KERATINIZATION

configuration assumed by the polypeptide chains is of considerable interest
from the viewpoint of the biosynthesis (Chapter III) of the various poly-
peptide configurations.

Deuterium Exchange

A further method of estimating helical content (or at least the content of
tightly packed chain segments) is that of deuterium exchange which is based
on the finding that, when a protein is dissolved in deuterium oxide (D20),
certain of the hydrogen atoms participating in hydrogen bonding are
exchanged with deuterium atoms and are given off again at varying rates
when the protein is returned to water. Thus one can distinguish instan-
taneous, rapidly and slowly exchangeable hydrogens. The slowly
exchangeable hydrogens are interpreted by Linderstrom-Lang (1955) to
be the hydrogen atoms in the tight helically-folded segments and thus the
helical content can be estimated by the proportion of these to the total
H-bond content. The method is being actively applied. Fraser and
MacRae (1958) have shown that in fibrous keratin the irreplaceable H-
bonds (inaccessible) were in the crystalline phase, i.e. presumably the
a-helices.

Fig. 84. Characteristic reflections of the parallel-/? (||) and the cross-/?

(X/?) patterns. The " backbone spacing " 4-65 A in the ||/? pattern occurs

on the equator; in the cross-/? it is found on the meridian.

The cross-/? pattern

In X-ray photographs of denatured proteins, such as the muscle
proteins and egg white (Astbury et ai, 1935) and some contracted keratins
(epidermin (Rudall, 1946), wool (Mercer, 1949a; Peacock, 1959)),
there may be found a /?-type pattern in which the (4-65 A) reflection usually
associated with interchain CO . . . NH linkages is not in its normal place
on the equator (see Plate 2B) but is on the meridian (Fig. 84). To
distinguish this pattern from the usual ^-pattern, or parallel-/? (|| /?), it is
referred to, by Rudall and his associates, as the cross-/? pattern (X/?).

The stretching of boiled egg white at first produces an oriented pattern

MOLECULAR AND M ACROMOLECU L AR STRUCTURE

201

which can only mean that the crystallites (or orientable units) contain
bundles of polypeptide chains running at right angles to the long axis of
the crystallites. While the actual molecular arrangements are still regarded
as uncertain, a proposal by Rudall (1946), to the effect that large transverse
folds (Fig. 85) such as were originally proposed by Astbury for the
supercontracted state, would be likely to lead to the cross-/? pattern, forms
the current working hypothesis.

The discovery that highly-oriented " natural " cross-/? patterns are given
by the egg stalk of the green lace- wing fly Chrysopa (Parker and Rudall,
1957) places this pattern on a more definite basis. The polypeptide chains

Fig. 85. Illustrating Rudall's proposal for a cross-/? crystallite which
could give rise to reflections from the 4-65 A spacing on the meridian. The
chains are extensively folded with their side spacings parallel to the fibre
axis. The crystallite is supposed to be longer in the side-chain direction
and stretching converts it directly to the parallel-/? form without rotating
the crystallite.

here undoubtedly run at right angles to the length of the fibre and are
folded since a conversion to the ||/? form takes place on stretching. Long
spacings relatable to the transverse folds disappear on stretching.

Astbury holds that a supercontraction resulting from a folding from the
a-form into a shorter configuration, such as that proposed in the cross-/?
structure, occurs in muscle contraction although it appears most clearly in
contracted muscle when this is produced by heating. Strength is given to
this opinion by a recent discovery (Astbury et al, 1955) that the cross-/?
configuration normally accompanies the a-pattern in bacterial flagella (see
Fig. 86) which appear from their helical shape in life to be in a state of
equilibrium between a contracted and an extended configuration.

An entirely different proposal to explain the occurrence of the spacing
of 4-7 A on the meridian in the cross-/? pattern has been made recently by
Zubay (1959). He points out that an arc of spacing 4-7 A passing over the
meridian could be produced by small microcrystals formed by a-helices

202

KERATIN* AM) KKRATIN I ZATION

which had been released from the coiled-coil form of the normal a-structure
by the swelling and warming required to produce the pattern. If the
crystals are small and sufficiently disoriented, the two off-meridional spots
of the first row line (from the 54 A spacing) will spread and produce a
single arc across the meridian which will then appear meridional arising

Fig. 86. Composite fibre diagram of X-ray reflections from Proteus

vulgaris and Bacillus subtilis. The a-characteristics are indicated by

reflection 47 and the cross-/? by 48. For the other spacings listed, see

Astbury et al. (1955). Reproduced by permission.

from a 4*7 A spacing. He likens the pattern to the completely-disoriented
halo pattern with spacings of 4-7 and 10-1 A (see p. 16) given by many
unoriented proteins. No trace of the 1-5 A spacing occurs in the X/3.

Other a-proteins

The a-proteins are among the commonest structural proteins to be
found both intra- and extra-cellularly. There is now no doubt that the
a-helix perhaps in a distorted form is to be found in the compact cor-
puscular molecules (Perutz, 1959). The significance of this would

MOLECULAR AND M ACROMOLECULAR STRUCTURE 203

seem, to be largely stereochemical : a polypeptide chain tends to assume
the tightly H-bonded helix unless other overriding side-chain inter-
action or interference is present to prevent it. When sufficiently long
runs of suitable residues exist the helix may form and may associate with
other helical segments to produce an a-crystallite which may be detected
by X-ray means. While many such forms may thus yield the definitive
wide-angle X-ray pattern they may differ greatly in their secondary and
tertiary structures. Some of these arrangements, interesting as special
problems, may not be immediately relevant to the keratin problem;
others, on the other hand, suggest structural possibilities which may be
found in the keratins themselves.

An important example of an a-protein, important because of its origin
and the absence of complicating histological structures, is the material of
the bacterial flagellum to which Astbury has given the name " flagellin "
(1955). Bacterial flagella are thin ( ~ 100 A) whip-like threads attached
to bacteria and connected with mobility.

Astbury, Beighton and Weibull (1955) have described the rich X-ray
pattern obtainable from preparations of flagella from Proteus vulgaris and
Bacillus subtilis (Fig. 86). It is clearly of the a-type with added reflections
indicating an admixture of the cross-/? configuration. Heating produces a
/3-form which may be oriented by pressing. The diagnostic 5*1 and 1*5 A
reflections are present. The numerous meridional reflections appear as
orders of a master spacing of 410 A which is similar to that observed in
muscle. The thickness of flagella would suggest that they are structural
analogues of the a-filaments of hair keratin. Their poor thermal stability
and composition show that they are not keratinized and they could there-
fore be related to the primitive non-stabilized a-type proteins of the
cytoplasm. Although they are located extracellularly, they are not simple
extrusions of the cell wall or capsule. They in fact take their origin from
a small basal granule located beneath the protoplast membrane to which
they remain attached even when the cell wall has been removed. Thus
they remain in direct communication with the cytoplasm and it is per-
missible then to regard them in this special sense as cytoplasmic filaments.

Some a- and jS-proteins of insect origin

Keratinization is not known in insects; their sclerotized proteins
are normally hardened by tanning with aromatic phenols (Chapter I).
The various dermal glands are, however, able to produce some variety
of protein-types among which are forms of a-type proteins whose study
has added importantly to our knowledge of molecular configurations and
their interconversions. Silk glands in Bombyx are labial glands and
their secretion while in the gland, according to infra-red evidence, has
random coil features. It transforms into the insoluble £-form during

204 KERATIN AND KERATINIZATION

spinning. The detailed structure of /3-fibroin illustrates clearly how the
insolubilization is produced by multiple H-bonding between the carbonyl
and imino groups of the peptide bonds. Nevertheless the transformation
is not fully understood; there is evidence both for an unfolding of the
chains effected by stretching and for an aggregation process which does
not obviously involve chain unfolding (p. 129).

Other glands, the colleterial glands, accessory to the female genital
system, produce various fibrous secretions, which serve to support or
protect the eggs. One example, the egg-bearing stalk of Chrysopa, is
important as a naturally-occurring example of a cross-/3 system (see p. 201),
i.e. the polypeptide chains in a /S-form are oriented at right angles
to direction of stretching of the fibre. Such a configuration is not un-
common in tissue proteins under experimental conditions (see p. 200) but
never achieves the perfection of orientation found in these egg stalks.
Another extremely interesting structure based on an a-form was discovered
by Rudall (1955-6) in the egg case of a mantis which forms from the
secretion of the colleterial glands. The dried case is tanned and insoluble.
However, when freshly formed it may be dispersed by tryptic digestion
and is seen to be composed of masses of long, very thin ribbons (5Qa X
1-2//) which (electron-microscopically) are crossed by regularly-spaced lines
(120 A apart) making an angle of about 20° with the length of the ribbon.

When first secreted from the colleterial gland the protein is a viscous
mass of globules which change first into long fibrils and then into ribbons.
The air-dried secretion before this change gives a somewhat diffuse X-ray
pattern but the definitive a-spacings : 1*5 A and 5*18 A are present. The
usual, strong, equatorial reflections at 10 A are missing, the lateral spacing
being represented by diffractions near 14*5 and 8*3 A. This pattern could
be given by rods of diameter 16*5 A packed in hexagonal array. This same
secretion after changing into ribbons yields a remarkably different pattern :
the main wide-angle meridional spacings at 1*5 A and 5*1 A are present,
but photographs taken at right angles to the length of the ribbons and
parallel to their face are of the a-type, and the pattern is dominated by a set
of row lines which are orders of 17*5-1 8*5 A (depending on hydration
(Plate 3A)).

Parallel to the long axis of the ribbons the diffractions from planes
perpendicular to the surface of the ribbons contain one set indicating a
period of 31 or 62 A, leading to a picture of rods 10-3 A in diameter in
hexagonal packing (Fig. 87).

On stretching the ribbons in water a £-form with axial periodicity of
3*33 A is produced which returns to the a-form on releasing the tension.
The /^-structure is still double-oriented, i.e. the " backbone spacing "
(4*7 A) is oriented perpendicular to the ribbon surfaces, which would seem
to imply that the a ±5: /? transformation is effected with the maintenance of

MOLECULAR AND MACROMOLECUL AR STRUCTURE

205

the double orientation. The devising of a molecular structure capable of
accomplishing this change is a difficult problem.

Rudall has offered two interpretations which, however, he finds only
partially satisfactory. The first considers straight a-helices of diameter
10*3 A hexagonally packed. Spacings of 31 A and 17*8 A at right angles
are thus achieved (Fig. 87). The a ±? j8 transformation presents no special
difficulty since the helices are straight. The diagonal lines crossing the
ribbons (as seen electron-microscopically) are accounted for by supposing
a superficial layer of fibrils making an angle of 20° lying across the main
sheets. Crick's argument suggests a good fit would be achieved between

31 A

A B

Fig. 87. Proposals for the packing of a-helices to give rise to the spacings

observed in X-ray patterns fron mantis ootheca (Rudall, 1955-56) and

to the diagonal lines observed electron-microscopically crossing ribbons

of the material at an angle of 20°.

(a) hexagonal packing of straight a-helices of diameter 10*3 A;

(b) two superposed layers of a-helices inclined at about 20°. Repro-
duced by permission.

fibrils placed together in this way (p. 183). The other interpretation
is based on a two-stranded rope of a-helices. A layer of such ropes,
in which the pitch of the supercoils is chosen to be 180-190 A, has a
pattern of well-marked grooves running parallel across the layer making
an angle of 20° with its length. The similarity to the pattern seen in the
ribbons is close, but the grooves in the model are separated by 30 A while
those in the ribbons are separated by about 120 A indicating an axial
period of 360 A or double that of the model. A model involving coiled
helices also meets some difficulty in explaining how a double-oriented
condition can be maintained during extension to the j8-form.

Feather keratin

The X-ray patterns yielded by feathers are much more detailed than
those given by any mammalian keratin (Plate 2A). The best patterns are
given by the calamus and rachis (Fig. 30, p. 70) and Rudall has shown

206 KERATIN AND KERATINIZATION

that essentially the same basic features are exhibited by patterns from
many hard parts of birds and reptiles (1949). The feather pattern is
perhaps more promising for detailed analysis than any other given by a
keratin, but a complete solution of the structure has not yet been given.

According to Astbury and Marwick (1932) the wide-angle pattern is
typically jS with the definitive axial repeat of 6*2 (2 X 3*1) A and complex
side spacings centred around 10 and 4*5 A (Plate 2A). Feathers may be
stretched by about 7% of their length before breaking. This does not alter
the nature of the pattern, the various longitudinal spacings increasing by
the same order; this behaviour contrasts with that of the a-keratins. The
axial long spacings have been regarded as orders of a main spacing of
95 A, cf. in a-keratin 198 A. The existence of definite and strong lateral
spacings at low angles distinguished the feather pattern sharply in type
from the pattern of collagen. Collagen fibrils give no lateral reflections
indicative of structures of greater thickness than the basic monomeric
filament. It was the lateral spacings of the feather pattern which early led
to the idea that in this material we were concerned with what were
essentially long crystals of precisely-constructed protein molecules not
essentially different in the detailed nature of their internal structure from
molecules of soluble proteins (Astbury and Marwick, 1932). This idea
persists in the more recent attempts to elucidate the structure made by
Bear and Rugo (1951).

In these attempts, Bear and Rugo (1951), while not proposing a solution
of the small-scale structure, have drawn attention to the implication at the
macromolecular level of the characteristic manner in which the pattern
degenerates when feather is subjected to the disintegrative action of water
and heat (" heat-moisture treatment "). After a prolonged treatment many
of the details of the pattern become blurred and fade and are replaced by a
much simpler pattern (" net-pattern ") which can be derived from a net of
the type shown in Fig. 88. What seems to have happened is that the heat-
moisture treatment has disturbed the fine-scale structural order to the
point that it is no longer capable of coherent reflections leaving only the
large particles (now effectively internally amorphous) centred about the
nodal points of the net to dominate the scattering. Two arrangements of
these large particles (macromolecules), are envisaged by Bear and Rugo,
and it will be seen that the major axial-repeating distance of 95 A may be
referred to the length of two or four of the participating particles.

These proposals illustrate very clearly the tendency shown in connexion
with the structure of protein fibrils, and already referred to above (p. 165),
to separate the X-ray diffraction pattern into two parts : (a) that given by
the small-scale molecular spacings (i.e. smaller than 20 A), in this case the
/3-pattern; and (b) that to be referred to the ordered packing of macro-
molecular units.

MOLECULAR AND M ACROMOLECULAR STRUCTURE

207

Evidence for the existence of large macromolecular units of this kind
may be sought in the products which result when fibres are dissolved.
For feather, as for wool and hair (Chapter VI) unfortunately the solubili-
zation process is drastic and destructive of larger units. Ward, High and
Lundgren (1946) using a detergent and reducing agents have dissolved
feather and found a molecular weight of a detergent-feather complex of the
order of 30-40,000. They concluded also that the particle was elongate and

Fig. 88. A proposal due to Bear and Rugo (1951) for the fibrillar structure
of feather keratin. Two packings of ellipsoidal molecules which would
account for the pattern of long spacings remaining after successive
treatments with hot water have disorganised the wide-angle pattern, are
illustrated. An example of an " aggregation of macromolecular particles
model ". Reproduced by permission.

the mixture polydisperse. Probably the molecules were unfolded by the
process. If formed into a sphere, a molecule of the molecular weight found
by Ward et at. would have been of a size as envisaged by Bear and Rugo.
More recently Woodin (1954a and b) using a reducing solution containing
urea obtained an electrophoretically-homogenous material. Osmotic
pressure measurements, viscosity, sedimentation rate and light scattering
concurred to give a molecular weight of the order of 10,000 and showed
that the particle was rather asymmetric (Woodin, 1955). Rougvie (1954)
found the same particle weight in an extract of feather oxidized by peracetic
acid (see also p. 163).

The infra-red absorption spectra of feather has provided information
about the orientation of hydrogen bonds. The CO stretching mode

208 KERATIN AND KERATINIZATION

at 1650 cm-1 reveals a perpendicular dichroism. The NH stretching
mode at 3315 cm-1 is also perpendicularly dichroic. By exchanging the
H of the OH and NH groups in the accessible (non-crystalline) regions
with deuterium, Parker (1955) removed their absorption from these
regions and showed that the perpendicular dichroism of the well-oriented
region rose to 4*8.

The absorption bond of the CO stretching frequency has a double peak
(Fig. 83) and, on the basis of their empirical rule (see p. 198), Ambrose
and Elliott (1952) concluded that this meant that a mixture of a- and
j8-forms were present. We have already indicated (p. 198) that this deduction
would not necessarily follow from our present-day views of the meaning
of these frequency shifts.

Recently a new examination of the diffraction data has been made by
Krimm and Schor (1956) who, while confirming that the structure is of a
/?-type, consider that none of the hitherto-proposed arrangements is
correct. On the grounds of the difficulty of fitting in all the axial reflections
in Tables 10 and 11 as orders of 94*5 A they consider that the true spacing
is 2 X 94*5 = 189 A. The 3*07 A spacing usually regarded as meridional
is stated by them to be an off-meridional 3*15 A layer line spacing. A
meridional reflection at 2*9 A is interpreted as the amino acid repeat (or
rise) and there are thus sixty-four residues in the master period (189 A).

They have proposed a model which might be described as a modified
^S-helix which consists of ten polypeptides, each of sixty-four residues,
wrapped helically around a cylinder (Fig. 89). The helices are right-
handed and every eighth residue is a prolyl, the side chain being on the
inside of the cylinder. The non-proline sequences form a modified
j8-pleated sheet. The ten chains aggregate coaxially by hydrogen bonding,
the prolyl residues of neighbouring chains coming out at about the same
level, and the strong 23*4 A meridional reflection arises from the planes
containing these residues (see Fig. 89). The R groups project both inside
and outside the cylinder. The particles of molecular weight ^ 10,000
isolated by Woodin and which appeared to have no (or very few) end
groups, they believe arise by a fission of the cylinder adjacent to the prolyl
planes thus producing prolyl end groups which are not estimated by the
usual methods.

A fairly satisfactory prediction of the strong equatorial reflections at 33*5
1 1*2 and 55 A (Schor's figures) was obtained by assuming that compound
cylinders, each consisting of seven unit cylinders, were placed in hexagonal
array.

It will be seen that the Schor-Krimm model is essentially a return to the
idea that the large axial spacings represent distances over which a sequence
of amino acids is repeated. In this particular instance the prolyl residues
are supposed to recur at intervals of ten residues. What little chemical

MOLECULAR AND MACROMOLECULAR STRUCTURE

209

evidence is available is not incompatible with this, the frequency of
proline by analysis being 1 in 10, but direct evidence on the form of
isolated peptides is still wanting. It would not seem too difficult to observe
electron-microscopically the hexagonal packing of the compound cylinders
(diameter 67 A) and filaments of a diameter less than 100 A have in fact
been observed by the author (Mercer, 1956). One interesting result of this
work is the introduction of the idea of 8-helices.

Fig. 89. Illustrating the proposal of Krimm and Schor for a feather
keratin structure. The drawing shows a projection on an 8 A cylinder of
one turn of their cylindrical unit X = propyl residue, O = non-propyl.
The master repeat period is 189 A and the 94-5 A is regarded as a
pseudo repeat unit (taken from Schor's thesis (1958). Reproduced by
permission.

In a recent comment on feather keratin Fraser and MacRae (1959) reject
the proposals of Krimm and Schor as being inadequate to account for the
intensity distribution of the diffraction pattern and re-assert that some
variant of the older ^S-type structure will prove the correct solution. Since
none of these discussions attempts a strictly crystallographic treatment it is
difficult to assess their merits, but it is evident that a final solution of the
structure of feather keratin has yet to be proposed.

CHAPTER VI

The Keratinization Process

In this chapter we shall discuss in detail the changes by which the
living and growing epithelial tissue is converted into a lifeless, tough,
insoluble, translucent, fibrous substance. These changes, occurring in a
more-or-less clearly-defined zone which follows the zone of differentiation
and growth (Fig. 25, p. 55) are characteristic and broadly similar in all
instances; by them we recognize and define a keratinized or cornified
tissue. In thin or quiescent skins it may not be possible to distinguish a
distinct zone of keratinization ; in many thicker skins, or in the matrix of
the hard keratins, even more subdivisions suggest themselves.

The literature concerning the staining of keratinized tissues is large and
many writers have described progressive changes in staining properties of
cells during their keratinization. A good historical account with particular
reference to wool and hair will be found in Auber's monograph (1950).
Recent reviews have been given by Montagna (1956) and Braun-Falco
(1958). There is no point in further summarizing these accounts; rather,
we shall by selection and omission attempt to co-ordinate the findings into
a coherent view of the process.

For all its usefulness in making evident histological detail, much of this
work is disappointing in contributing little towards elucidating the process
itself because most dyes have little specificity and their reactivity is poorly
understood. Histochemical demonstrations of definite substances (or
groups) are more rare and more useful. It is fortunate that reliable histo-
chemical methods exist for demonstrating thiol groups (SH).

Since there is some difference between the sequence of events in the
formation of a hard keratin and a soft keratin, it will be better to describe
these separately, commencing with a typical hard keratin, hair.

The hard keratins

These include all the coherent appendages: hairs, feathers, nails,
claws, horns, etc. Most of the data relating to the hardening of this type of
keratin comes again from the hair follicle, but the fewer observations made
on other examples (horn, nails, and feathers) are sufficient to show that
the changes seen in the hair follicle are general and justify the use we
propose to make of it for purposes of illustration. For economy of pre-
sentation it will be an advantage to summarize the data in a uniform

210

THE KERATINIZATION PROCESS

211

diagrammatic form in one place. The hair follicle, the best-studied and
most "diagrammatic" tissue, will be used for this purpose and is shown in
the series of Figs. 90, 91 and 97. The illustrations in most cases refer to
events in the presumptive cortical cells and each diagram summarizes the
findings relative to some property or particular activity. An inspection of
these diagrams shows clearly that in the hair follicle several distinct stages
in keratin formation, separated in time and in space, may be distinguished.
These are indicated in Fig. 90 :

A, zone of cell division (germinal matrix).

B, zone of differentiation and cell growth.

C, zone of fibril formation.

Dl , ... f(i)

„ }, keratimzation zone {;...•
e y l(u)

F, keratinized zone.

Analogous levels can usually be distinguished in the other tissues.

The diagrams have been devised in terms of the several interrelated
themes which can be distinguished in the process of keratin formation:

( 1 ) General cellular phenomena : (a) nucleic acid metabolism and protein
synthesis (Fig. 97b); (b) cell metabolism (Figs. 97a and d).

(2) Phenomena peculiar to Keratinisation: (a) orientation (Fig. 91);
(b) stabilization (Figs. 91, 92 and 97c).

The Development of Orientation (Fig. 91)

The fibrous contents of the cells of the hard keratins are usually well
oriented and, since this orientation is related to the mechanical function,
its development is of special interest. In the upper regions of the bulb
(Chapter III) the cells of the presumptive cortex elongate and there is a
marked increase in the number of oriented fibrils within the cells. The
orientation is most conveniently observed by means of the polarizing
microscope (Schmidt, 1924) and its quantitative development may be
measured by means of a compensator (Mercer, 1949b). Figure 91 (r.h.s.)
shows the growth of birefringence in a follicle plucked from the human
head and the l.h.s. shows the development of birefringence in relation to
the anatomy of the follicle. The important feature is the rapid rise at the
constriction of the bulb to a value which is almost equal to that of the final
hair.

While polarization microscopy provides the simplest method of de-
tecting the existence of an oriented structure, the interpretation of the
results is not without ambiguity. It is useful to be able to distinguish
between intrinsic birefringence, i.e. birefringence due to an oriented
molecular structure, and form birefringence, which arises simply from a

212

KERATIN AND K ER AT IN I Z ATI ON

keratinized zone

zone of keratinization

zone of fibrillation

zone of differentiation

and growth

germinal matrix

Fig. 90. The location diagram for the series of figures (Figs. 90-97)
illustrating the development of keratiaization in the human hair follicle.
On the left-hand side are indicated the several zones A-F into which the
cortex is divided. In the growing follicle (centre) (1), (2) and (3) indicate
the cortex (and cuticle), the inner root sheath and the outer root sheath,
respectively. The cell shapes are those of the presumptive cortical cells.
On the r.h.s. is shown a " club root " or non-growing follicle to indicate
the extent to which the growing tissues may be resorbed (see also Fig. 40).

THE KERATINIZATION PROCESS

213

parallel arrangement of long, thin rodlets, with or without internal
anisotropy, when immersed in a medium of a different refractive index
(Schmidt, 1924; Schmitt and Bear, 1939). Experimentally the dis-
tinction is made by observing the change in the birefringence due to form

Fig. 91. The development of orientation and its stability. X-ray patterns
obtainable from the different levels are shown diagrammatically. The
a-pattern is obtained immediately above the bulb in zone D but on
warming the unstabilized fibre is disoriented and yields an unoriented
/3-pattern. In the zones E and F the pattern is stable to heating below
100° C. On the right-hand side is shown the birefringence (An) in the
cortex and the levels (zone D shown clear) where it is destroyed by heat.

when fluids of different refractive index are allowed to penetrate the object.
Unfortunately this method is inapplicable to the hard keratins, such as
hair, since by the very process of keratinization the material has become
impervious to liquids which do not cause far-reaching disintegration
(Barnes, 1933; Mercer, 1949). For this reason the clearest understanding
of orientation phenomena comes from X-ray diffraction.

214 KERATIN AND KERATIN IZATION

The appearance of an oriented structure at the molecular level can be
demonstrated readily by X-ray diffraction (see Chapters I and V), and the
method has the advantage that the type of structure (in the molecular
sense) is also demonstrated. From the bulb alone only a diffuse, uncharac-
teristic diffraction pattern of two rings (Fig. 91) can be obtained.
Immediately above the bulb, however, an oriented a-type pattern, apparently
identical to that of the final hair, is obtained. This finding shows beyond
question that at this level there is a considerable amount of oriented
protein present with essentially the same crystalline molecular organization
(a-type) as in mature hair.

This important point, that the appearance of the typical a-structure
precedes keratinization was demonstrated originally by Derksen, Heringa
and Weidinger (1937), using the thickened epidermis of cow's lip, a more
amenable material than hair follicles, and by Giroud and Champetier
(1936) using the " chestnut " of the horse. Sections cut at different levels
up to the fully-hardened layers gave the same a-pattern. Similar experi-
ments on a cow's nose were later carried out by Rudall (1946) and by the
present writer using the hair root (1949b). After heating in water the
lower unstabilized layers become disoriented.

In their totality these experiments prove that the filaments, which can
be seen to form in the cells below the keratinizing zone, already possess
the typical a-structure, and further that, whatever chemical reactions may
go on in the later stages of keratinization, they in no way affect the arrange-
ment of molecules in the crystalline regions.

The older histologists, whose work is summarized by Biedermann
(1926), recognized that the oriented structures and birefringence arose
from fibrils and deduced the existence of smaller invisible anisotropic units
from their polarization studies. Various schemes of fibrillar architecture
(see Biedermann) were developed which were substantially correct. For
this type of research the modern electron microscope is now more con-
venient (see p. 223) but the polarizing microscope is still much used.

The Development of Stability

Fully-hardened hair has a high stability towards many chemical reagents
and physical conditions and, accordingly, the development of keratinization
may be assessed in terms of the action of any of these influences. Owing
to its linear arrangement the plucked hair root is a very convenient object
on which to make such tests, and the results of several are depicted in Figs.
91 and 92.

All these tests agree in showing that the cortex of human head hair is
fully stabilized at a level about one-third of the total length above the bulb.
Further, it shows that the keratinization zone itself which extends from the

THE KERATINIZATION PROCESS 215

constriction to the fully stabilized level may be divided into two distinct
parts :

(a) A lower zone (Fig. 90, zone D) in which the synthesis and orientation
of the fibrous protein itself seems largely complete, but in which the
structure is poorly stabilized.

(b) An upper zone (Fig. 90, zone E) in which stability rapidly develops.

(b)

Fig. 92. Histochemical methods of demonstrating the unconsolidated
zone D of the pre-keratin. Birefringent regions are shown in black.

(a) Destruction of orientation in zone D by heating for 30 sec at 90° C.

(b) Result of tryptic digestion of a follicle for some hours. D is removed,
the inner root sheath is resistant above the level B and the earlier harden-
ing of the fibre cuticle is apparent, (c) Destruction of DR due to the
penetration of dilute acid, (d) Result of treatment with dilute alkali
which acts destructively even on the partly-hardened regions, zone E.

From Mercer (1949b).

These zones are well distinguished by the tests illustrated in Fig. 92.
Briefly, many reagents and treatments dissolve only the lower zone (D,
Fig. 90) of weaker stability; more violent treatment is required to derange
the subsequent levels (Fig. 92d). The exact range of these levels is not the
same in all types of follicle, but Hardy's work on the mouse follicle shows
that the sequence of changes is the same.

In experiments on plucked hair roots the present writer (1949b) showed
that the cortical orientation (observed as double refraction) was destroyed

216

KERATIN AND KERATINIZATION

Fig. 93. Progressive dispersion of the unconsolidated pre-keratin by
saturated urea. Above D the partly-hardened protein swells only and to
a diminishing degree as the hardening progresses. Birefringent cortex
shown black. The time taken for the whole action is from 5 to 10 min
(from Mercer, 1949d).

2000/X

Fig. 94. Swelling (percentage of original area on ordinate) of a follicle in

a, water, b, saturated urea as a function of the distance from the base of

the papilla (Mercer, 1949b).

THE KERATINIZATION PROCESS

217

in zone D by heating for 30 sec at 90°C or by soaking in dilute hydrochloric
acid. Trypsin readily digests the same layers A, B, C and D. Alkaline
reagents are, however, able to destroy the birefringence in the lower parts
of zone E (Fig. 92). Perhaps the most spectacular and instructive effects
follow the application of saturated urea to the base of a plucked hair
follicle (Fig. 93). The lower zone D swells rapidly and is dissolved; the
upper half E swells only, and to a diminishing degree, as the follicle is
ascended (Fig. 94).

(a)

(b)

(c)

Fig. 95. The location of the keratinization zone ( + SH) in three hard
keratins, (a) the nail, (b) the claw and (c) a horn (edge only shown).
Positive SH shown black. Redrawn from Giroud and Bulliard (1930).

Thiol and Disulphide Groups during Keratinization (Fig. 97(c))

The most important single observation relevant to the chemistry of
keratinization is that the keratinizing layers give a positive reaction for
thiol (SH) groups and that this reaction fades as the tissues harden. In
their important comparative study of numerous keratinizing tissues
Giroud and Bulliard (1930) established the existence of an SH-positive
layer in every case and distinguished between the hard and soft keratins
by the intensity of the reaction. Fig. 95 has been redrawn from their work.
Giroud and Bulliard used the nitroprusside reagent for thiol groups. Since
that time more permanent and specific reagents have been introduced
(Chevremont et al., 1943; Rudall, 1952; Barrnett, 1953; Barrnett and
Seligman, 1952) and the use of these has confirmed the earlier work. The
conclusion to be drawn from these results is that the fixed SH, i.e. that
resistant to washing by being joined to a protein framework, disappears in
the course of keratinization and that therefore one of the reactions under-
lying keratinization is the oxidation of the SH groups to produce cystine
bridges: (— S— S— )

2— SH + 0 = — S— S— + H20
When the SH groups are first blocked by alkylation, and the tissue reduced.

218

KERATIN AND KERATI NIZAT ION

the oxidation may be reversed and the distribution of new SH groups will
be that of the former disulphide bonds. This distribution is found to be
complementary to the SH (Hardy, 1952; Rudall, 1952) as may be seen
from Fig. 97(c).

In some pathological conditions oxidation may be incomplete, and a
positive thiol reaction persists into the normally-hardened layers. A
significant case of this was described by Marston (1946), in sheep reared
on a diet deficient in copper (p. 159). This element seems to play a role

Fig. 96. The positive SH zones in the growing feather follicle (after
Giroud and Bulliard). Compare with figure of the hair follicle (Fig. 90).

as a coenzyme in the oxidation of thiol groups (Flesch, 1949); when
deficient in copper the positive- SH reaction persists along the greater
part of the follicular shaft, the emerging wool fibre is less thoroughly
stabilized and the crimp has a longer wavelength (see also p. 159).

Thus the outstanding fact relating to the progress of keratinization is
that it occurs substantially after fibril formation (zones D and E) and in
two steps. The two steps can be satisfactorily correlated with (a) a primary
stabilization effected by hydrogen bending (zone D) and rather readily
disorganized (Fig. 90D) followed by (b) a consolidation of the primary
structure by the progressive introduction of cystine bridges resulting from
the oxidative linking of cysteine (SH) side chains. These chemical
changes are accomplished without detectable change in the a-type structure
which is established and stabilized by hydrogen bonds in the first-formed
fibrils. The sudden increase in stability, particularly apparent in the tests
depicted in Figs. 92 and 93 (level (E)), seems to coincide with the appear-
ance of a dense material between the filaments seen in the electron
microscope and described on p. 224.

THE KERATINIZATION PROCESS 219

It seems not unlikely that other chemical modifications occur con-
comitantly. The changes in the affinity of the proteins in zones C, D and
E for acidic and basic dyes described by Auber (1950) Odland (1953) and
Montagna (1956) suggest that other side chain modifications are occurring.
According to Montagna's summary, the first-formed fibrils are basophilic,
i.e. they take up basic dyes in a pH range 4-6; as the material matures it
loses its power to bind basic dyes at lower pH values although it still binds
acid dyes. Evidently some acid groups are lost or modified. This could be
due to the conversion of some acid groups (COOH) to amides (CONH2).
For example, in wool the effective acid and basic groups are closely
equivalent when allowance is made for the acids in amide form. Table 13
is adapted from Simmonds (1954 and 1955).

Table 13.

(gramme equivalents amino acids in 105

g wool.)

glutamic acid 101-8
aspartic acid 50-3
tyrosine OH 35-3

lysine 19-3
arginine 60-2
histidine 5-7

187-4
amides — 88-6

unknown 71

total 98-8

total 92-3

The amides may also contribute to the stability by participating in
hydrogen-bond formation. A more complete investigation aiming at the
localization of amino acid residues other than cysteine and cystine is
needed. Ryder has described preliminary tests for several (1959). Gillespie
et al. (1960) point out that analysis of the fibrillar (a-component, see p. 240)
and the y-component, which may enter during the progress of keratinization
(see p. 224), differ in their content of basic and acidic amino acids. The
matrix proteins (y-keratins) (p. 248) are more basic than the fibrillar
a-proteins.

Nucleic Acids and Synthesis (Fig. 97 (b))

The dividing cells of the matrix (zone A) have prominent Feulgen-
positive nuclei and a basophilic cytoplasm containing RNA which is said
to increase as the cells differentiate (zone B) and commence to synthesize
protein. The pattern of the nucleic acids : well- developed DNA-containing
nuclei, prominent RNA nucleoli and a strongly RNA-positive cytoplasm
are compatible with current views of the role of these acids in synthesis as
outlined in a previous chapter.

220 KERATIN AND KERATINIZATION

The disappearance of the nucleic acid from the keratinized hair poses
some problems. A nuclear residue persists (see p. 263) but is not Feulgen-
positive. It would seem that, even in the upper levels of zone E where it
would be thought that the dehydrated and moribund cells would be
incapable of further activity, some reactions, which mobilize and remove
nucleic acids, can still take place. Perhaps this operation is of value in the
hair follicle in that valuable materials (phosphorus) are resorbed. Bolliger
and Gross (1952 and 1956) have, however, reported ample quantities of
nucleic acid breakdown products in hairs and other keratins.

£^

K^=J

c^ (F^

(a) (b) (c) (d)

Fig. 97. The results of some histochemical tests on the human-hair
follicle: (a) mucopolysaccharides and/or alkaline phosphatase; (b) the
nucleic acids [D = DNA, R = RNA1 ; (c) thiol (SH) and disulphide
sulphur (S2); (d) glycogen. In the shaft of the cortex the black areas are
the fully-stabilized hair, and the stippled are the pre-keratin = positive
thiol regions.

Another unsolved problem is raised by the recent results of Fell,
Mellanby and Pelc (1954 and 1956) which revealed an apparent localization
of radioactive sulphur in nuclei of the cells of the oesophageal epithelium
after injection of labelled crystine. This would seem to imply that some
reaction involved in the addition of cystine to the cytoplasmic proteins has
its onset in the nuclei. Pelc (1958 and 1959) believes that cystine is ex-
changed against some DNA constituent which is itself broken down. Such
a view would link the final disappearance of the DNA with the late
synthetic processes (see also p. 264). However, the actual finding in these
experiments is the localization of sulphur in nuclei and this could equally
well be explained by an association between nuclear RNA and a sulphur-
containing protein, a view which would accord better with the other views
of the participation of RNA in synthesis.

the keratinization process 221

Metabolic Enzymes

Certain enzymes concerned with basic cellular metabolism are necess-
arily present in all cells and show a diffuse cytoplasmic distribution.
Rogers (1953) applied the Nadi reagent as a test for cytochrome oxidase to
wool roots and found a strong reaction in the bulb extending to the proxi-
mal third of the shaft. Montagna (1956) reports similar findings. Dehy-
drogenase activity has been demonstrated by means of tetrazolium salts
which are reduced by the enzymes to yield purple granules (see Montagna,
1956, for review). Rogers specifically demonstrated succinic, /?-glycero-
phosporic, lactic and malic dehydrogenases.

The oxidases and dehydrogenases are associated with energy-supplying
reactions and their presence in the hair bulb could also be associated with
mitosis which Bullough (p. 136) has shown requires energy. The dehy-
drogenases of the shaft could also be concerned with the energy-consuming
process of transport of metabolites from the bulb or of the glycogen from
the outer root sheath. A cyclic variation in these enzymes in the epidermis
during the hair cycle is described by Carruthers et al. (1959). Presumably
also, the energy required for protein synthesis is obtained through similar
reactions. Since many of these enzymes are now known to be located in
mitochondria, their presence could also be inferred from the mitochondria
visible in electron micrographs of these cells.

Glycogen (Fig. 97 (d))

The most conspicuous deposits of glycogen are in the o.r.s. cells. In the
middle third of the follicle the cells are virtually filled with it. The possible
role of this glycogen as an energy store, which buffers the follicular
system against the fluctuations in glucose content of the systemic blood
supply and thus enables a steady rate of mitosis to be maintained in the
bulb, has been mentioned in Chapter IV. When the hair follicles of rodents
are quiescent no glycogen is present, it increases rapidly when growth
recommences (Montagna et ai, 1951). In good agreement Ryder (1958)
found that radioactive glucose rapidly (1 hr) entered the bulb where it was
presumably being utilized directly to sustain mitosis and later was localized
(probably as stored glycogen) in the outer root sheath. Bradfield (1951)
believed that epidermal cells stored glycogen while in the basal layer and
carried it outwards where it supplied energy for protein synthesis in the
outer layers.

Acid Mucopolysaccharides and Schiff-reactive
Substances (Fig. 97 (a))
Metachromatically-staining substances (p. 54) are present in the
dermal papilla (Sylven, 1950 and 1951; Montagna, 1956) and Schiff-
positive material seems to occur in much the same situation (Leblond,

222 KERATIN AND KERATINIZATION

1951). Both materials increase during the growing phase and decrease
during rest. Metachromatic staining usually indicates acid muco-
polysaccharides and is common in growing mesenchymal tissues else-
where but its exact role is not clear. Sylven's suggestion that the sul-
phur of polysaccharide was transferred to the growing hair through the
papilla and was a source of sulphur for keratinization seems to be refuted
by Ryder's observations to be described later (p. 232).

Phosphatases (Fig. 97 (a))

Moog (1946) has summarized the evidence to show that there is
commonly an association between alkaline phosphatase and transport of
materials. The distribution of these enzymes in the skin and hair follicle
(Fell and Danielli, 1943; Montagna, 1956) is such as to suggest that they
could be concerned with the transport of metabolites. They occur together
with the mucopolysaccharides. From their location it would seem that
phosphatases are not directly associated with the processes of keratinization
itself. See also Braun-Falco (1958).

Lipids (not Illustrated)

According to Montagna (1956) lipids stainable by means of Sudan dyes
are not strongly developed in the lower follicle and even less can be
demonstrated in the upper levels. This is in contrast to the epidermis, but
it seems to show that lipids in themselves have little to do with keratiniz-
ation at least in the hard keratins. In the epidermis also they are probably
accessory, serving to plasticize and waterproof the formation. Contrary
opinions have been expressed (see Rothman, 1954).

Water content

The water content of the cells of the lower bulb is probably of the order
of 90%; that of the keratinized hair about 30%. There is clearly a
considerable dehydration in the course of the formation and consolidation
of the fibre, but it has not been measured with any accuracy. A very
similar dehydration occurs in the stratum lucidum and corneum of the
epidermis. In the hair follicle the most marked loss of water accompanies
the rapid synthesis and coalescence of fibrils in the upper bulb and
coincides with the narrowing of the lumen and the rise in birefringence.
It would seem that, with the condensation of the cytoplasmic proteins into
compact fibrous masses, many hydrophilic end groups are either modified
chemically or are utilized in the formation of various bonds internal to the
fibrils and thus release the water molecules previously bound by them.
The " free " water may then leave the cells for osmotic reasons. When
H-bonds are broken in reduced hairs a marked swelling and hydration
occurs, which is probably a partial reversal of processes occurring in the
follicle in zones B to C.

THF KERATINIZATION PROCESS 223

The fine histology of the hair follicle in the keratinization
zone

In the hair follicle may be found examples of the formation of both hard
and soft keratins and for this reason the detailed study of the fine structure
of its various layers provides an opportunity to compare and contrast the
two modes of development.

As indicated in Chapter III, the separate cell streams, which form the
inner root sheath, the cuticle, cortex and medulla of the hair, are clearly
differentiated above the middle of the bulb ; synthesis and stabilization
follow distinct courses in each. They will be described separately.

The Cortex

Interest here chiefly concerns the further development and consolidation
of fibrous keratin. The fine parallel filaments which first appeared as
wispy bundles in the cells of the presumptive cortex at a level a little below
the tip of the papilla (Fig. 90, zone B) rapidly accumulate in the upper
regions of the bulb until, at the level of the constriction of the follicle, the
cells appear almost full. The length of the follicle from its constriction to
the level where the cortex achieves its definitive form has been called the
zone of keratinization (D and E, Fig. 90).

The fine filaments (diameter ~ 60 A) (Plate 12B) also described as
the protofibrils (Chapter III) are at first individually distinct and form
small clusters which rapidly grow in size. When in the upper bulb, these
aggregates become a few tenths of a micron in width, they can be seen in the
light microscope and are then described as fibrils (Plates 13, 15 and
16). It is at this same level, a point immediately following the follicular
constriction, that birefringence rapidly rises and reaches almost the
value of the final fibre and an oriented a-type X-ray pattern (Fig. 91),
indistinguishable from that of the fibre, can be obtained. These obser-
vations show again that the synthesis of the basic fibrous structure is
virtually complete at this level and that the chemical changes, which follow
and produce hardening and stabilization of the formation, must take place
outside the crystalline regions, and probably outside the filaments them-
selves if we take the further step of identifying the filaments with the
" X-ray crystallite".

Some support for the idea that changes external to the filament occur
comes from changes in the electron-microscopic appearance of the bundles
of filaments as they advance through the zone of keratinization. At first
separate filaments are seen in cross-section as clusters of dense dots and
their arrangement is rather irregular. Then areas of better order (quasi-
hexagonal packing) appear in which the filaments now appear relatively
light against a darker matrix (Plates 15 and 16). It is as if a new and more

224 KERATIN AND KERATINIZATION

dense component was forming between the filaments and causing them to
pack more compactly (Birbeck and Mercer, 1957; Mercer, 1958). At the
level of the constriction (Fig. 90) and immediately above it, it is possible to
distinguish a population of fairly distinct fibrils (diameters in the range
005-02^) each consisting of masses of fine, light filaments in quasi-
hexagonal array, which thus constitute a rather definite level of organi-
zation of the fibrous keratin (Fig. 80 and Plate 13); then, as the cells
advance, these fibrils rapidly fuse laterally, to produce progressively large,
irregularly-shaped aggregates, and finally an almost complete fusion into a
solid mass in which only residual, irregularly-dispersed gaps remain as
evidence of the earlier existing interfibrillar spaces (Fig. 16). It should be
noted that whereas the filaments appear to be perfectly definite structural
elements, all having the same diameter, the fibrils, which are aggregates of
filaments (Figs. 80 and 102), are not of uniform size although they cluster
around an average diameter.

In the later stages of consolidation, the hexagonal arrangement of the
filaments within the fibrils becomes distorted, leading to the appearance
in cross-section of spirals or fingerprint-like whorls. It is possible that
each fibril becomes slightly twisted on its long axis (Plates 14A, 15 and
16).

At the histological level (light microscopy), the course of keratinization
in the other hard mammalian keratins seems essentially similar to that of
hair. The fine histology has not yet been studied, but there would seem to
be every reason to think it will also prove similar to that of hair. In
feathers the formation and condensation of fine filaments follows similar
lines (Mercer, 1957). There is nothing in the electron-microscopic appear-
ance of the feather cells to show that a /8-type rather than an a-type protein
is being formed.

The Cuticle

The presumptive cuticle cells, which form a single layer as they leave
the matrix of the bulb, tilt sharply as they approach the bulb constriction
and achieve their fully-tilted and overlapping condition at the level of the
constriction where Henle's layer turns birefringent. Here they still contain
little or no keratin and their contents are no longer symmetrically disposed.
Nuclei are usually found towards the basal pole of each flattened cell and a
well-developed system of vesicles (Golgi apparatus) lies largely apical to it.
The flattened cell is almost vertical and its two surfaces now face towards
very different environments, the internal towards the hair cortex, the
external towards the sheaths and follicular surroundings. In such circum-
stances the cell contents develop a stratification parallel to the cell mem-
branes (Plate 20A).

Keratin appears as small (~300 A) rounded, dense droplets and moves

THE KERATINIZATION PROCESS

225

towards the peripheral surface of each cell where it collects in curiously
patterned groups (Birbeck and Mercer, 1957). The aggregation continues
and with the closure of the gaps a compact layer packs against the cell wall.
The cellular apparatus comes to occupy the inner part of each cell and the
stratification thus produced persists in the final cell state. Cross-sections
of cuticle cells (Plate 20A) show these two layers of distinct texture, but
reveal also that within the keratin itself there is also some stratification
(Sikorski and Simpson, 1959). When reduced and stained by osmium
tetroxide or metal salts, the layer adjacent to the external-facing membrane

Fig. 98. Schematic drawing of two cells in the stratum corneum of the
guinea-pig showing the characteristic keratin pattern with bundles of
filaments (F) embedded in an interfilamentous substance (IF). The cell
boundary consists of a fairly opaque, broad inner zone (Z) and an opaque,
fine outer membrane (PM). The intercellular space is filled with a
substance of a fairly low opacity (I). In this substance opaque, homogen-
eous, diffusely outlined bodies (IB) are observed. From Brody (1959).

is more dense (probably a higher cystine content). Further, the cell
membrane facing the cuticle of the inner root sheath is the more dense and
conspicuous. This stratified structure explains many of the properties and
reactions of cuticles (see p. 265).

The keratin of the cuticle is not fibrous in texture; only traces of fibrils
form and there is no transformation into a fibrous form as occurs to the
rather similar droplets of trichohyalin in the inner root sheath cells (p.
226). Consolidation seems to be a simple coalescence of the (perhaps
sticky) droplets to form the very coherent, continuous layer immediately

226 KERATIN AND KERATINIZATION

beneath the cell membrane. The reasons for the delay in synthesis and for
the different type of keratin produced are not known. It is possible to
speculate that the denser and closer opposed membranes of the cuticular
formation deprive the cells of some essential factor, and that synthesis
proceeds under limiting conditions which tend to produce amorphous
peptides richer in cysteine. Similar conditions may prevail in the upper
keratinization zone of the cortex where an increase in cystine also occurs.
A sufficiently-high frequency of cystine residues in a polypeptide would
seem to preclude both the possibility of assuming the regular folding
necessary for crystallization and of long-range extensibility.

The birefringence, which can be observed in cross-sections of the cuticle
cells, particularly in swollen hairs, may be due either to the fact that the
molecular chains of the keratin are compressed into one plane or to the
compression of the largely membranous cell apparatus (Schmidt, 1925).

The Inner Root Sheath (i.r.s.)

The characteristic product of the inner root sheath is trichohyalin, sl
substance as yet imperfectly characterized. The name is due to Vomer
(see historical review in Auber) (1950) who wished to emphasize by it the
clear structureless character of the granules and a distinction from similar
granules (keratohyalin) found in the skin. It is a protein and when in the
fibrous form is strongly birefringent and yields an a-type X-ray pattern.
It is quite distinct chemically from fibrous keratin, notably in being
deficient in cystine. Rogers (1959) has shown further that it is peculiar in
containing the amino acid citrulline. The similar keratohyalin occurs in
the cells of the epidermis proper where its appearance, as strongly staining
granules, gives rise to the name, stratum granulosum. In spite of some
differences in stainability (see Rothman, p. 376 (1954)) the two substances
seem essentially similar in nature. In this sense the inner root sheath is a
soft keratin ; and, like the epidermis itself, it desquamates half-way along
the follicle. It is this event which actually frees the advancing hair shaft
from its enveloping sheaths.

The synthesis of trichohyalin commences first and proceeds further in
the cells of the layer of Henle, the outer-most layer of the i.r.s. contiguous
with the outer root sheath. Small dense granules of trichohyalin appear
free in the cytoplasm and grow rapidly. They have no internal structure,
are not birefringent, stain heavily (in the electron-microscopic sense) with
phosphotungstic acid and with basic dyes as judged by light microscopy.
The fact that the granules absorb basic dyes, as does nucleic acid, has led
some to think that they contain nucleic acid. This is not now thought to
be the case since digestion with ribonuclease has only a small effect on the
basophilia (Leuchtenberger and Lund, 1951) and electron-microscopically
the trichohyalin droplets are free from the small dense particles usually

Plate 13

Cross-section of the cortex of a human hair in the keratinizing zone
of the follicle to show the arrangement of fibrils. N, nucleus ; M (arrows),
cell membranes bounding the cortical cells; p, pigment granules; P,
particulate cytoplasmic residue. Higher magnifications of fibrillar
keratin appear in Plates 14A, 15 and 16.

■ r >. *^

> V

" *¥

•i>-

B
Plate 14 (Captions on facing page)

Plate 14

A. Cross-section of human-hair follicle in the pre-keratin zone to show
the whorl-like arrangement of the fine filaments in the fibrils, F. The fine
filaments may show a hexagonal packing (H) in the centre of the fibrils
which changes towards the fibril periphery to resemble a spiral (S).
Notice the filaments are embedded in a denser ground substance. For an
enlargement showing this detail, see Plates 15 and 16. At Cu appears a
portion of a cuticle cell containing irregular lumps of cuticular keratin
devoid of fine structure.

B. Section running approximately parallel to the basement membrane in
the epidermis of a frog tadpole. Centrally it passes into the epidermal cell
and whorls of "keratin" fibrils, /, are seen; nearer the edges of the
picture the section passes obliquely through the basal membrane, BM.
Note the pattern of dense desmosomal studs D and the finer granular
deposits a. Surrounding this are seen the dermal collagen fibrils in quasi-
orthogonal array. For a section at right angles to this, see Plate 9. PTA
stain.

Plate 15

High resolution electron micrograph of several fibrils as noted in
cross-section of the hair cortex in the pre-keratinous zone of the human-
hair follicle. Compare with Plate 16.

A fibril, F, is seen to be composed of fine filaments, / (a-filaments)
about 60 A wide which are embedded in a denser ground substance to
produce a characteristic pattern. For other forms, see Plates 16 and 17.
A surface of contact between two cells winds across the section and
may be seen in detail at C. In the intercellular spaces S much cyto-
plasmic debris persists. Taken from Birbeck and Mercer (1957).

Plate 16

Cross-section of fibrous keratin in the cortex of the fully-hardened
wool fibre. Portions of several fibrils are seen with an almost effaced
interfibrillar line at (I). At R are seen residues of cytoplasmic debris.
The material consists of close packed filaments embedded in a ground
substance (matrix), which is denser after reaction with the osmium fixative
than the filaments themselves. At H an almost perfect hexagonal packing
is visible. For an interpretation of the pattern see Figures in text. Com-
pare with the less well-oriented material found in skin keratin shown in
Plate 17. Magnification X 100,000.

Photograph kindly lent by Dr. G. E. Rogers.

Plate 17

Section of the fibrous keratin in the keratinized cells of the guinea-pig
epidermis. The pattern of light filaments embedded in a dense matrix is
in essentials identical with that of the well-oriented hair shown in Plates
15 and 16, but the orientation is here less perfect and filaments seem
to run in bundles in various directions.
Photograph kindly lent by Dr. I. Brody.

Plate 20

A. Cross-section of edge of hair cuticle and the inner-root sheath of a
human hair in the hair follicle. Same level as Plate 14A. Sections of three
cuticle cells {Cu 1, 2 and 3) appear in the upper half and within each cell
may be distinguished the cytoplasmic material largely particulate, and
the lumps of amorphous keratin, K, which are packing against the outer
membrane, M, limiting each cell. Outside the cuticle (lower half of picture)
may be seen the cuticle of the inner root sheath Cs and a Huxley-layer
cell Hit. In each of these may be seen sections of granular trichohyalin,
H, and transformed material, T (fibres seen in longitudinal section, see
Plate 20B).

B. Higher magnification of the transformed trichohyalin (T of 20A).
There seem to be here sections of filaments and of ribbons (or coalesced
filaments).

The keratinization process 227

regarded as containing RNA although the adjacent cytoplasm itself is
densely packed with these.

All the cells of the i.r.s. produce trichohyalin, but its appearance is the
more delayed the further the cells are from the outer root sheath, i.e. it
appears last and least abundantly in the cuticle of the i.r.s.

The most characteristic property of trichohyalin (as also of keratohy-
alin) is its ability to be transformed into a compact fibrous modification.
This change is shown vividly in the light microscopy by sudden appearance
of birefringence in the cells of Henle at the level of the constriction in the
follicle. At the same time these cells become more difficult to stain and
the tubular sheath formed by Henle's layer becomes relatively insoluble in
strong solutions of urea.

Electron-microscopically the change is equally sudden and remarkable.
It is associated with the appearance of bundles of fine filaments which seem
to " stream " out of the tips of the now lenticular-shaped droplets of
trichohyalin (Plate 21). Birbeck and the writer (Birbeck and Mercer, 1957 ;
Mercer, 1958) consider that this is evidence of a transformation of the
accumulated reserves of amorphous trichohyalin and is in some ways
analogous to other instances, e.g. that of actin, in which a fibrous modifica-
tion develops from a non-fibrous precursor (see Chapter III, p. 127). The
filaments appear in well-ordered bundles strictly parallel to the axis of the
follicle, i.e. to the long axis of the extended cells. Electron micrographs,
such as that shown in Plate 21, suggest a likeness to crystallization, the
long thin filaments seeming to move out from the droplets as though they
were being continuously formed at their surface.

In cross-sections the bundles of trichohyalin are ambiguous in appear-
ance; both dots, i.e. sections of filaments, and wavy lines, sections of
sheets or ribbons, are found (Plate 20B).

After the transformation, Henle's layer forms a strongly-coherent
sheath not dispersed in urea solutions. There is, however, no sign that the
filaments of trichohyalin cross from cell to cell. They continue to the cell
membrane and seem to end in close contact with this structure. A very
similar picture is found in muscle cells where the filaments also end on
extensive desmosomes (p. 42). The actual seat of intercellular adhesion
must still be sought in the intercellular cement between the cells. This
question has been discussed earlier (p. 84).

The function of the transformation of the trichohyalin into a hard
fibrous tube at the level of the follicular constriction seems to be to provide
a solid retaining support to carry the still soft tissues contained within it
towards the surface of the skin. The somewhat slower transformation in
the inner layers of the sheath may permit the layers to remain plastic and
to form a more snug attachment between the hair and its sheaths (Auber,
1950).

228 KERATIN AND KERATINIZATION

Soft keratinization

The Epidermis

Matoltsy (1958) gives as constituents of stratum corneum of human
epidermis: 65% insoluble keratin, 10% soluble protein, 10% dialysable
material (largely amino acids), 7-9 % lipid and 5 % cell membranes. The
tissue is clearly more heterogeneous than a hard keratin and gives the
impression, from its content of soluble protein and low molecular weight
substances, that the hardening process is incomplete. On the other hand,
a comparison of Plates 16 and 17 shows that, apart from the orientation
of the filaments, the fine structure of soft keratin is very similar to that of
hard keratin.

There is a large amount of histochemical work on the epidermis. Like
that on hair its significance is marred by the lack of specificity of the means
employed. The significant tests (summarized in Montagna (1956)) are
essentially those already described for hair and their meaning for the
understanding of keratinization is much the same. A good deal of attention
has been paid to elucidating the nature of keratohyalin granules but
without complete success (see p. 226). Staining with fluorescent dyes
(Jarrett et al., 1959) demonstrates vividly several phases of keratinization,
but the interpretation of the effects is not obvious.

The distinct histological feature of epidermal (soft) keratinization, when
several layers are fully developed, is indeed the occurrence of a granular
layer due to the deposition of granules of keratohyalin. This substance is
in essentials identical with trichohyalin of the inner root sheath of the hair
follicle but whereas in the i.r.s. trichohyalin is apparently produced alone
and its transformation provides the entire fibrous contents of the hardened
cells, the course of formation of the fibrous component of the epidermal
cells is more complex and not fully understood. Filaments, as far as can be
ascertained, of the same type as those of the cortex of hair follicle, are
already to be found in small amounts in the basal layer cells (Plates 7 and
9). Here these are directed predominantly at right angles to the basal
layer and a weak birefringence in this direction is detectable (Montagna,
1956; Biedermann, 1926). In the stratum granulosum keratohyalin
granules appear as dense amorphous bodies apparently quite independently
of the earlier-formed filaments (Plates 22 and 23).

The relation between keratohyalin granules and the fibrous keratin of
the stratum corneum has for long been a subject for dispute and it cannot
be said that electron microscopy has finally settled the points at issue.
There are two main points of view current, both of long standing although
it is not always easy to interpret the views of the classical microscopists
immediately in modern terms. The first view is that keratohyalin " mixes
with " or " spreads over " the fibrils (tonofibrils) already present to yield

THE KERATINIZATION PROCESS

229

the cornified keratin which thus may be regarded as a mixture. This view
is supported in part by the superb micrographs of Brody (1959) (Plate 17)
which show that the keratin of the epidermis has essentially the same fine
structure as the hair cortex: poorly-stained filaments embedded in a
strongly-stained matrix. The filaments here are not strictly parallel as in

iiiii

Fig. 99. Schematic drawing of the upper part of the epidermis (guinea-
pig) taken from Brody (1959). The fully-keratinized cells contain the
typical keratin pattern of unstained filaments embedded in a stained
amorphous matrix as illustrated in the electron micrographs (Plate 17).
In the stratum spinosum S tonofibrils predominate, in the granular layer
G filaments and granules K of keratohyalin are present. The curious
association of granules and filaments is suggested at T. In the trans-
itional cell, which follows, the cytoplasm is more condensed, the cell
boundaries more dense and keratohyalin predominates. The dense
particles P still persist.

hair cortical keratin; they cluster in somewhat wavy bundles lying more
or less in the plane of the flattened cell. Brody suggests from the juxta-
position of granules and filaments in the stratum granulosum that the
keratohyalin granules " incorporate " the filaments (Fig. 99). He thus
derives the interfilamentous cement from the keratohyalin and on the
analogy with the work on the cortex supposes that this material must be
rich in cystine. He cites in support the finding of Chevremont and

230 KERATIN AND KERATINIZATION

Frederic (1943) of a high concentration of SH groups in the granules. On
the other hand, most other observers seem to agree that there is little or no
sulphydril or disulphide in the granules (Van Scott and Flesch, 1954;
Montagna, 1956; Eisen et al.).

The present writer (Mercer, 1958) has supported the view that the
keratohyalin granules are a direct precursor of the fibrous keratin and that
something of the same type of transformation occurs as was found in the
Henle layer of the inner root sheath of the hair follicle (see Plate 22).
Images showing filaments apparently emerging from granules may be
found. The amounts of fibrillar material in the cell of the granular layer also
appear to be too small to provide the amounts visible in the immediately
adjacent cells of the stratum corneum. Thus it seems that this material can
come only from the accumulated granules. A conversion of the granules
into fibrils running roughly parallel to the skin in the flattened cells could
account for the sudden rise in birefringence (with its positive axis parallel
to the skin) and for the glassy appearance of these layers {stratum lucidum)
which stongly resembles Henle's layer above the transformation level. It
would seem that here the two potentialities of synthesis of the basal
epidermal cells, which become separated into two distinct cell lines (the
cortex and the i.r.s.) in the hair follicle, occur together in the epidermal cell
but manifest themselves at different times and levels.

Thus (in this view) the final fibrillar contents of the cells of the stratum
corneum would seem to be derived from two sources: (a) a small early
contribution of fibrils analogous to the cortical filaments of hair and
appearing like them without a precursor, and (b) a larger amount produced
later by the transformation of the non-fibrous precursor, keratohyalin.
The relation of these two kinds of fibrils to each other and the reasons why
the cells switch from one form of production to another are not known.

When in the fibrous form, trichohyalin yields an a-type X-ray pattern
and its sulphur content is low (Rogers, 1959a). These findings also apply
to the keratohyalin of skin, if we accept the claims that the granules contain
little or no S. Amounts of SH and disulphide are demonstrable in the
cytoplasm of epidermal cells at all levels with a stronger band near the
clear layer. Van Scott and Flesch (1954) report that there is little increase
in the total S in passing from the germinal to the horny layer which might
agree with the picture presented above if we assume that the earlier formed
fibrils become stabilized by disulphide cross-linking but that keratohyalin
on forming fibrils contributes little further disulphide. By measuring the
specific absorption of X-rays, Engstrom and Lindstrom (1947) showed that
the concentration (S per cm3) was many times higher in the stratum
corneum. This increase is undoubtedly largely due to cell dessication but
may also indicate an absolute increase in S content.

Rothman (1954) has proposed the idea that the cell inclusions may

THE KERATINIZATION PROCESS 231

dissolve and reform as the final keratin but this is not provable by
microscopy. There is some evidence in electron micrographs that the
early-formed tononbrils partly disappear in the granular layer.

A property of the fibrous form of trichohyalin of the hair follicle,
probably connected with the fact that it is not stabilized by disulphide
cross-linking, is its tendency to " fall apart " during the disintegration of
the inner root sheath. This same property plays a similar role in the des-
quamation of the epidermis.

One further puzzling feature of epidermal keratinization is that kerato-
hyalin is not invariably present. Many thin skins (e.g. birds) keratinize
normally without a granular layer and it is found in a variable degree
elsewhere. The view that it represents an alternative pathway of synthesis
and may make a contribution to an independently-formed system of fibrils
does something to explain these facts. The scaly (hard) keratin of rat tails
forms without a granular layer, whereas adjacent perifollicular skin is
softer, more flexible and has a granular layer. According to Jarrett and
Spearman (1961) treatment (externally) with vitamin A causes the appear-
ance of a granular layer and a softer keratin in the scale regions. Epidermal
cells differ evidently in the effect on their keratinization of vitamin A with
a range of responses included in the sequence :

... . Vit A f , Vit A

hard keratin > softer keratin -> mucin

(no keratohyalin) (keratohyalin)

See also p. 63 et seq.

X-ray photographs, optical methods (Matoltsy, 1957) and electron
micrographs (Plate 17) all show that the arrangement of the filaments in
the stratum corneum is less perfect in soft keratins than in hard. It is also
likely then that this imperfect organization contributes to lowering the
stability of the formation. It is known in wool (Rogers, 1959b) that in cells
where the filaments are less perfectly aligned, the keratin is less stabilized.
In skin no fibrils analogous to those of the hair cortex form, the filaments
appearing to form simply loose inter-connecting bundles (see Fig. 99).

Recently Swanbeck (1959), on the basis of the scattering of X-rays at
low angles, has concluded the existence of scattering unit of diameter of
260 A, which, assuming that the scattering phenomena have been correctly
interpreted, would seem to imply a close association of the 100 A filaments
in small groups. This is not immediately apparent in micrographs.

Keratinization of Horn

The special interest of horn lies in the fact that from it one can obtain
massive samples particularly suitable for some kinds of experiment. The
SH reactivity was examined by Giroud and Bulliard (1930) (Fig. 95) and
in further detail by Rudall (1956) who established clearly that there was an

232 KERATIN AND KERATINIZATION

increase in the intensity of the SH reaction on passing from the inner
mucosum to the outer, suggesting that an increase in the amount of
cysteine occurs during the later stages of hardening. This was confirmed
by an actual chemical analysis of the several layers which in this material
can be separated mechanically and chemically. This conclusion is in good
accord with the suggestion to be developed later that cysteine-rich pep-
tides are added during hard keratinization.

Follicular nutrition and the entrance of sulphur

There seems little doubt on histological grounds that the greater part of
the material supplies for the growth of the hair are conveyed by the papilla
whose dimensions control the output of keratinized cells. This impression
is given a quantitative basis by Rudall's (1956) extensive survey of wool
follicles referred to on p. 150.

The elaborate vascularization of the middle region of the shaft (Dur-
ward and Rudall, 1949; Ryder, 1958), which fluctuates with the hair-
growth cycle, remains to be explained. Possibly it is involved in the
transport of glucose, mobilized from the glycogen of the outer-root sheath
to the papilla or it could be associated with keratinization since the
evidence (see below) is that the sulphur enters at this level.

At the biochemical level it has been shown by the use of radioactive
tracers that methionine and not cystine in the diet contributes its sulphur
to the hair of rats (du Vigneaud, 1947; Marston, 1946). It is thought that
methionine is converted to homocystine and then linked to 1-serine to
form the compound :

NH2

NH2
HOOC'

CH.CH2—

— S.CH2.CH S

COOH

This compound is split in vivo at X-X and 1-cysteine is produced, the S
having been transferred to the serine to give cysteine. However, in other
animals the sulphur metabolism may differ (Ryder, 1958).

The site of entry into the follicle would seem at first sight to be the zone
of keratinization since here for the first time SH can be detected (Figs. 95,
96 and 97). It would be possible on this histochemical evidence alone to
suppose that below this level the sulphur is present in a concealed (non
SH) form, but further work on the uptake of radioactive sulphur following
injection has dispelled uncertainties. Injection of radioactive cystine is
followed by the rapid appearance of radioactivity in the keratinization
zone, but not in the bulb (Ryder, 1958 ; Bern et al., 1955 and 1957). On the
other hand labelled carbon and phosphate compounds enter through the

THE KERATINIZATION PROCESS 233

bulb. This evidence shows clearly that sulphur compounds do not enter
through the papilla and advance up the follicle. They must enter in some
way through the walls of the follicle at the level of the keratinization zone.
The transport of sulphur compounds could be one of the functions of the
extensive vascular network surrounding the follicle. There is also evidence
of an absolute increase in labelled compounds in the upper levels of the
keratinizing levels according to Ryder (1958). Rudall, who separated
and analysed the several layers of the growing tissues of horn, established
the same fact (1955).

On account of its shape, the hair follicle enables sulphur absorption to
be separated from fibrillar growth, thus facilitating interpretation of the
phenomena. In the epidermis on the other hand, the sulphur compounds
must diffuse upwards from the basal layers into the keratinization layers,
and this may have some influence on the lower uptake. Comparable
experimental work has not been carried out on other keratinizing tissues.

Sylven (1950), commenting on the presence of acid sulphate-containing
mucopolysaccharides in the papilla (see also Montagna (1956)), believed
these might be active in transferring their sulphur to the growing hair. In
view of the demonstration that labelled sulphur-containing amino acids
enter at the level of keratinization while labelled phosphate enters the
papilla, this would now seem unlikely. These acid mucopolysaccharides
(substances staining metachromatically with thiazine dyes) are most obvious
during active hair growth and seem clearly connected with the proliferation
of the bulb, but their sulphur is certainly not transferred to the growing
cells at this level.

No definite intracellular structure has been yet shown to be associated
with the absorption of the sulphur acids or their change into cystine.
Enzymes are of course suspected but not isolated. " Microbodies " (Plate
22), single-walled dense bodies, are often common in the keratinizing zone
and may contain a special enzyme system.

Soluble products of partial keratinization

However the process of keratinization is viewed, it certainly involves in
the first place the synthesis of one or more polypeptides which sub-
sequently become stabilized by the formation of cystine bridges. Clearly
some insight into the nature of keratinization would be obtained if the
state of aggregation of the proteins in the cells at various levels in the
keratinization zone were known. The concept of the molecular weight of
keratin in the hardened tissue is itself meaningless, since the protein
is extensively united into large and indefinite heterogeneous formations
by covalent cross-linkages. It would, however, be valuable to know in
the first place the number, composition and molecular weight of the
primary polypeptide chains, i.e. chains containing only peptide links, which

234 KERATIN AND KERATINIZATION

presumably form the first stage of synthesis. And further, it is meaningful to
ask whether or not there exists a hierarchy of definite molecular association
of increasing molecular weight whose formation precedes the appearance
of the finest filaments which can be seen microscopically (diameter
~ 60 A).

There are two ways in which these problems have been attacked: (a)
extracts have been made of the growing tissues with the object of dissolving
the proteins before they become keratinized, and (b) attempts have been
made, starting with the hardened tissues, to reverse the keratinization
process and to obtain soluble macromolecular products from the hardened
tissue. This degradation could obviously be continued until small pep-
tides and amino acids were obtained. The full analysis of the complete
mixtures resulting from such a partial hydrolysis would clearly do much to
elucidate the amino acid sequence of the original polypeptides and it is
regrettable that, apart from the pioneering work of Consdon, Gordon,
Martin and Synge (Martin, 1946), little has been attempted. For the
present we are more concerned with the possible existence of high mole-
cular weight, intermediate polypeptides of a definite character.

From a consideration of the composition and reactions of proteins in
general, and of the keratins in particular, the following kinds of bonds
might be supposed to participate in the consolidation of an insoluble
protein :

(a) Hydrogen bonds, i.e. associations between neighbouring CO and NH
groups mediated by the hydrogen atom (see Chapter V).

(b) Salt bridges, i.e. salt-like linkages formed between acid groups
( — COOH) and amino groups (— NH2). Speakman (1934) has
amassed evidence to show these are effective in hair and wool. (For
a contrary opinion, see Jacobsen and Linderstram-Lang (1949).)

(d) Disulphide bridges (— S— S— ).

(e) Other bonds have been proposed, e.g. between phenolic OH groups
and acid groups (Alexander and Hudson, 1954) but are not known to
exist for certain.

All these bonds and perhaps others not yet discovered, may play a role in
stabilizing insoluble proteins; accordingly, the solvents used to effect a
solution or make extracts of keratinized tissues have been chosen because
of their specific effect on one or more of these bonds. In keratin the co-
valent disulphide bonds appear ultimately to prohibit solution and next in
importance on account of their number are the hydrogen bonds. These
appear particularly to influence the dry hardness and extensibility. The
histochemical experiments described above (Figs. 91 and 92) seem to show
that the H-bond sustains the structure in the early stages of keratinization

Plate 21

The level of transformation of granular trichohyalin into the fibrous
form as seen in Henle's layer of the human-hair follicle. On the l.h.s. (A)
is a light micrograph in which may be seen the edge of the hair cortex C
containing fibrils of keratin and pigment granules, the hair cuticle Cu,
the cuticle of the inner root sheath, Huxley's layer containing granular
trichohyalin H and Henle's layer He in which the sudden change from the
granular section G into the hyaline birefringent, coherent section B is
evident. An electron micrograph of the portion encircled (in an adjacent
section) is shown on the r.h.s. (B) where the fibrous strings, /, of trans-
formed keratohyalin may be seen extending from the tips of the elongated
granules G. Dense particles persist in the cytoplasm. PTA stain. Taken
from Birbeck and Mercer (1957).

Plate 22

The transitional zone between granular keratohyalin and fibrous
keratin as seen in the upper layers {stratum lucidum) of the plantar
epidermis of the rat. G, granule of keratohyalin with fibrils, /, which
appear run out of it in a manner analogous to those seen in Plate 21.
m, mitochondrion; b, microbodies ; D, desmosomes associated with the
convoluted cell membranes.

MMMHMM

c ;

i— i

3f x

Plate 23 (Captions overleaf)

Plate 23

A. Portion of granular and cornified layers of the plantar skin of a rat
showing keratohyalin granules; KH , keratohyalin ; C, " transformed "
keratohyalin ; T, microbodies ; B, cell membrane D.

B. Illustrating the development of the basal-layer cells of the epidermis
of a 12-day-old chicken embryo. C, collagen fibrils below the basement
membrane (here cut obliquely); F, fibroblast containing filaments; E,
epidermal cell; G, intercellular gap; X, intercellular exudate; m, mito-
chondrion (PTA stain).

»*£

Plate 24 (Captions overleaf)

Plate 24

A. A pair of coalescing " keratin pearls", cysts lined with a squamous,
stratified, keratinizing epithelium in a tongue of invasive epidermal tissue
on the skin of a mouse treated with benzpyrene. Light micrograph,
stain: haemotoxylin-eosin. Preparation made by Dr. I. Hieger.

B. Light micrograph of cross-sections of wool fibres to show the dis-
tribution of o- and p-type cells. The fibre has been oxidized with pera-
cetic acid and stained under acid conditions with methylene blue (Pearse's
technique (1951)). The enhanced basiphilia (the dark stain) indicates the
regions of higher sulphur content (para-type). Note that although the
ortho-para distribution is predominantly bilateral, para-ce\\s may be
found among the ortho.

THE KERATINIZATION PROCESS 23$

before the closure of the disulphide bonds. Hydrogen bonds are usually
weakened by the introduction of urea into the solvent. The thermodynamic
factor involved here is the heat change accompanying the transfer of the
H-bond between two peptides to a pair of urea molecules which appears to
favour the peptide-urea association. In a sense the urea molecules prise
apart the chains and destroy the secondary structures maintained by them.
Useful discussions of this problem will be found in articles by Ward and
Lundgren (1954) and by O'Donnell and Woods (1956).

Extracts from the Pre-keratinized Zone

The idea of extracting the proteins from a keratinizing tissue before
they harden is attractive since the extract might be expected to contain
soluble precursors. Rudall (1946 and 1952) showed, however, that the
buffered aqueous solvents commonly used in biochemical extractions
removed very little protein from skin. He found, however, that the ad-
dition of urea to the solution suffices to dissolve copious amounts of
protein from a thick skin such as a cow's nose. He named this protein
" epidermin".

When precipitated from solution by ammonium sulphate, epidermin
forms a voluminous, white, sticky, curd-like material, easily gathered
together and drawn into fibres. These fibres are somewhat elastic when
wet, are birefringent and yield an excellent a-type X-ray pattern. Stretch-
ing produces a fibre giving a /^-pattern. When heated in water (50-60°C)
oriented fibres contract, and the contracted material gives a disoriented
/3-pattern. Epidermin thus behaves as an unstabilized keratin, i.e. its basic
molecular framework is established but, in the absence of cross-linkages,
is readily disorganized.

Rudall (1946) and others (Derksen et al., 1937) have demonstrated the
increasing thermal stability of strips of tissue cut from successively higher
layers of epidermis. The effect of keratinization can be imitated closely by
cross-linking epidermin fibres with formaldehyde and benzoquinone
(Rudall, 1946).

Similar extracts may also be made from other keratinized tissues by
urea solutions. Here again the simple linear arrangement of the hair
follicle enables an exact location of the extracted protein to be determined
(Mercer, 1949b). When a plucked human head follicle bearing a papilla is
placed in concentrated urea solutions there is an immediate swelling
followed by dissolution of the lower half of the keratinization zone (Fig.
93). Above a rather definite level the precortex merely swells and to a
decreasing degree the higher the level (Fig. 94). The germinal tissues of
the bulb containing little protein are less affected; nuclei are not dissolved.
The relation of these events to other histochemical features is best seen in
the series of Figs. 91, 92 and 97. It will be noted that only the lower half

236 KERATIN AND KERATINIZATION

of the SH-positive zone dissolves which means that disulphide cross-
linking becomes effective at a rather definite point. Below this point
H-bonding is the major stabilizing element; above it is supplemented
progressively by disulphide cross-linking. The addition of reducing agents
(thioglycollic acid) to the urea carries the solution to a higher level.

A velocity-sedimentation analysis of epidermin solutions in the ultra-
centrifuge was made by Mercer and Olofsson (1951a) and revealed the
presence of several components (three or more) with sedimentation
constants ranging from 1 to 7. These components were reduced to one by
the addition of a reducer which suggests that the heavier components
were aggregates of the lighter, held together by disulphide linkages.
Epidermin would seem to be a mixture containing some complexes already
united by disulphide bonds.

Proteins have also been extracted from the follicle of the wool fibre using
the ingenious method of harvesting these in quantity devised by Ellis
(1948). Rogers (1959) has reported a detailed investigation on these
extracts and reached the conclusion that the larger part of the extract is a
fibrous material of low-sulphur content (< 2%) and that the remainder,
less well-defined, has the higher sulphur content ( >4%). A comparison
of the amino acid composition of whole wool and the fibrous low sulphur
component is given in Table 14 taken from Rogers.

Soluble Derivatives of Keratinized Tissues

This approach has attracted more attention for economic reasons. The
large quantities of keratinous materials going waste (hair, horns, feathers,
etc.) are a potential source of high molecular weight protein possibly of use
in the polymer industry (Jones and Mecham, 1943).

On theoretical grounds one might look for a series of soluble derivatives
of definite molecular size ; in practice, solvents of little specificity must be
used and polydisperse mixtures, difficult to analyse, are obtained. Two
approaches to rupturing the disulphide bond are open : reduction or oxid-
ation. By using various different pH values and by adding hydrogen bond
breakers, a variety of products can be obtained.

Reduction of wool. Sulphides which combine reducing properties with a
high pH have long been known as keratinolytic agents. Olofsson and
Gralen (1953) found that a sulphide extract of wool contained a mixture
of polypeptides of average weight of the order of 8000-10,000. Since all
disulphide bonds were broken, these polypeptides might be regarded as
primary although the high pH (11) may have produced some main chain
hydrolysis. Much earlier Goddard and Michaelis (1934 and 1935) used the
more specific reagent thioglycollic acid in alkaline solution, and obtained
solutions which contained two main components.

Other methods of extraction make use of the simultaneous action of

THE KERATINIZATION PROCESS

237

hydrogen-bond breakers, and reducing agents at a more moderate pH
(7-8). As reducing agents, thioglycollic acid, bisulphites, sulphides,
cyanides, mercaptoethanol, etc., have been used. By using a lower pH it
was hoped that peptide-bond hydrolysis and the conversion of the cystine
bridge into the more stable lanthionine bridge ( — CH2— S — ) could be
avoided.

Table 14. Comparison of Amino Acid Composition

of Wool and the Low-Sulphur Protein (a-

Fraction) of Wool Roots (Rogers (1959)).

Amino acid

a- Fraction

Whole wool

glycine

5-30*

5-58

alanine

4-76

4-04

valine

3-77

2-57

leucine

7-14

4-69

z'soleucine

2-89

1-62

serine

5-34

7-39

threonine

3-40

4-06

phenylalanine

2-00

1-76

tyrosine

2-37

2-54

proline

2-71

5-61

lysine

7-00

3-60

arginine

19-48

18-32

histidine

2-37

141

aspartic acid

6-54

4-64

glutamic acid

9-26

7-18

ammonia (amide N)

7-74

6-57

cystine

3-18

6-99

methionine

0-31

Amino acid N as % of total N.

About a quarter of the weight of wool dissolves in 5% sodium bisulphite
(pH8) and 10 M urea at 50°C in 24 hr (Jones and Mecham, 1934;
Mercer, 1949a). The extraction portion is now known to come from the
less keratinized orthocortex (see p. 268). When reprecipitated by the
addition of ammonium sulphate the extracted keratin derivative is obtained
as a white, sticky, coherent curd very closely resembling Rudall's epider-
min and Alexander's a-keratose (q.v.). Like these it can be drawn into
birefringent threads which give an a-type X-ray pattern. The attempts
made to estimate the molecular weight of extracts in urea are not wholly

238 KERATIN AND KERATINIZATION

satisfactory owing to a marked polydispersity and a tendency to aggregate
with time. Mercer and Olofsson (1951b) reported 84,000; Woods (1952)
estimated for a low molecular weight diffusible fraction 10,000 and for the
non-diffusible 50,000. For a recent review see Gillespie et al. (1960).

Jones and Mecham (1943) showed that many other keratins give soluble
extracts in solutions of urea containing reducing agents but with the
exception of feather (see below) little is known of the molecular character-
istics of the dissolved material. There is a marked resemblance between
these protein extracts and Rudall's epidermin. All precipitate as white,
sticky curds on the addition of ammonium sulphate and can be drawn into
fibres giving an a-type X-ray pattern.

Very similar extracts were obtained from feathers by Lundgren, Ward
and associates (Ward et al, 1946; Ward and Lundgren, 1954). More
recently Woodin (p. 163) has claimed that better defined " monodis-
perse " derivatives can be obtained by careful reduction with thioglycollate
at pH 1 1 which would seem to prove that in the case of feather at least a
definite macromolecular monomer can be isolated.

Oxidising reagents. Oxidation of the disulphide bonds with peracetic
acid was shown by Alexander and Earland (1950) to be an extremely
satisfactory method of obtaining a soluble derivative from keratins. The
oxidized keratin is readily soluble in dilute alkalis and can be thrown out
of solution on the addition of ammonium sulphate or acids as a white
coherent material, again resembling epidermin, which may be drawn into
threads yielding well-oriented a-type X-ray patterns.

Not all of the material extracted from oxidized wool can be so readily
precipitated from solution. A portion remains and has been designated
y-keratose by Alexander (Alexander and Hudson, 1954). Ultracentrifugal
analysis shows that the solution of oxidized wool contains two somewhat
polydispersed components, but the physicochemical characteristics of
these are still in dispute (O'Donnell and Woods, 1955). This particular
method of dissolving keratins has provided nevertheless one of the most
valuable insights into the nature of keratinization. It suggested to Alex-
ander the existence in a hard keratin of two main components of very
different character : one fibrous of high molecular weight, the other non-
fibrous and of a lower molecular weight. This second fraction has the
much higher sulphur content (see Table 15). The insoluble fraction,
called /3-keratose (less than 10%) has been shown to consist principally of
the membranes of the keratinized cells (Mercer, 1951d and 1953), see
p. 260.

Corfield, Robson and Skinner (1958) have determined the amino acid
composition of the oxidized keratin fractions (a, jS and y keratose) and their
results for a- and y-keratose are given in Table 16.

It will be seen that y-keratose contains much larger amounts of cystine

THE KERATINIZATION PROCESS

239

(cysteic acid), proline, serine and threonine and smaller amounts of
alanine, aspartic and glutamic acids, leucine, lysine and phenylalanine than
the original wool. The reverse is true of the a-fraction. These figures

Table 15. Protein Fractions Obtained from Oxidized
Wool. (Alexander and Hudson, 1954.)

Property

a-Keratose

y-Keratose

solubility

precipitated from
ammonia extract at
PHc4

not precipitated
atpH4

appearance

white sticky curds
of fibrous texture
birefringent as
thread

after drying,
yellow brittle
solid or powder

mol. weight

50,000-80,000 (?)

3000-10,000
(more poly-
disperse) (?)

chain length

X-ray diffraction

a-type pattern

ill-defined
perhaps j8-type

sulphur content*

2-4

2-90

2-5

1-88

1-96

6-lf

6-13 §

5-84**

4-72!

percentage weight \
of original sample J

60
56

30f
25**

* S content of original wool averages 3-5 %.
f Alexander and Hudson (1954).
X Earland and Knight (1956).
§ Alexander and Smith (1955).
** Corfield, Robson and Skinner (1958).

*[[ O'Donnell and Thompson, quoted Gillespie (1960) (performic
acid).

should be compared with those for " kerateine 2 " (Table 17) and those for
the " wool root extract " (Table 14). The higher sulphur content of
y-keratose shows it to be the more cross-linked and the higher proline is

240

KERATIN AND KERATINIZATION

of some structural significance in rendering the formation of regular helices
more difficult (p. 193).

Other oxidizing reagents which destroy cystine bridges (hydrogen
peroxide, chlorine, chlorine peroxide, etc.) have been less used and appear
less satisfactory. Das and Speakman (1950) demonstrated a variety of
polypeptides in extracts of wool oxidized by chlorine peroxide. The
molecular size was of a similar order to that found in sulphide extracts.

Table 16. Analysis of a- and )/-Keratose (Oxidized
Wool). (Corfield et al, 1958.)

(% of total nitrogen)

Amino acid

a-Keratose

y-Keratose

glycine

5-16

4-97

4- alanine

4-83

2-58

valine

3-98

4-15

4- leucine

7-30

2-55

t'soleucine

2-49

2-14

t serine

6-70

9-70

t threonine

3-45

746

4- phenylalanine

1-94

1-15

4- tyrosine

2-44

1-41

t proline

2-69

9-85

4- lysine

4-60

1-03

arginine

20-8

19-0

histidine

1-24

1-57

4- aspartic acid

6-25

1-79

4- glutamic acid

10-9

5-87

ammonia*

10-25

11-05

t cysteic acid

3-72

14-5

The arrows indicate an increase ( t ) or decrease
( 4- ) of a residue in the y-fraction relative to that of the
original wool.
* Too high according to Gillespie et al. (1960).

Performic acid, which may have a more specific action, has also attracted
some attention (Thompson et al., 1959; Gillespie et al., 1960).

Thiogly collate extracts. The original method of Goddard and Michaelis
(1934), who introduced the use of thioglycollic acid as the reducing agent,
also pointed to the existence of two different types of protein in extracts of
reduced keratin or kerateine. This method is undoubtedly one of the best
solubilization techniques since it i s in a sense reversible, insoluble disul-
phide cross-linked powders being obtained by oxidation of kerateines.

THE KERATINIZATION PROCESS

241

These products are amorphous, the larger structural elements character-
istic of the original material not being rebuilt.

This method has been much developed recently by Gillespie and
Lennox (1953 and 1955) who have made an extensive study of the elect-
rophoretic behaviour of several kerateines and their alkylated forms
(Gillespie, 1960). Their practice is to reduce in thioglycollic acid at pH
105 at 59°C to obtain several extracts (extracts A-E) at this pH, which
prove to be very heterogeneous, then to raise the pH to 12-3 when larger
amounts of an electrophoretically pure component " kerateine-2 " are
removed. The extracts are stabilized by blocking the SH groups by

GLYALA VAL LEU LEU SER THR MET

LYS ARC HIS ASP CLU NH2 CY5

Fig. 100. Comparison of the amino acid composition of a keratin de-
rivative (stippled) with that of merino wool from which it was extracted.
Figures represent percentages of total nitrogen. Reconstructed from data
given by Simmonds (1954).

alkylation. Simmonds and Stell (1956) have analysed these extracts and
demonstrated some striking differences in composition between whole
wool kerateine-2 and extract A. Extract A is characteristically higher in
cystine than kerateine-2. Figure 100 permits a visual comparison between
the composition of kerateine-2 and the wool from which it was obtained
and Table 17 shows the amino acid composition of the extract.

At 50°C 65% of wool can be dissolved in 0-1 M thioglycollate at pH 12-6
and seven components can be detected electrophoretically. The minor
components can be removed by five 20 min extractions at a lower pH
(10*5) leaving the residue from which the major component (41% of wool)
can be obtained by a further extraction at pH 12*3 (kerateine-2).

242

KERATIN AND KERATINIZATION

In these extractions a number of processes are involved, the rates of some
of which are diffusion controlled, the limiting step being either the ingress
of the reagent on the egress of the dissolved protein. These rates differ in
the two segments (o and p, see p. 268) of the fibre and probably the
o-segment contributes most of the extracted protein.

The analysis of kerateine-2 by Simmonds (Table 18) shows that relative
to the original wool it contains : increased amounts of the di-amino acids,

Table 17. Amino Acid Composition of Whole Wool and

Various Fractions. (Simmonds and Steel, 1956.)

(Amino acid-N as per cent total-N).

Amino acid

Extract A

Kerateine-2

Whole wool

glycine

12 09

5-27

5-80

alanine

2-75

4-22

3-51

valine

347

3-55

3-57

leucine

3-92

6-01

4-90

tsoleucine

1-96

2-24

1-97

serine

10-04

6-47

7-25

threonine

5-56

4.44

4-61

phenylalanine

2-40

1-72

1-75

tyrosine

4-49

2-46

2-97

proline

6-52

3-66

5-33

lysine

1-12

5-03

3-25

arginine

14-66

12-26

20-32

histidine

1-49

1-42

1'46

aspartic acid

2-91

5-68

4-24

glutamic acid

5-70

10-82

8-58

amide-N

5-79

11-56

7-46

cystine

13-76

4-33

7-93

leucine, lysine and ammonia N, and less of cystine, proline, serine and
tryptophane. Harrap's (1955) molecular weight determinations by means
of a surface balance suggest a magnitude of 30,000. Table 17 shows the
change in composition between the first extract A and the major component
"kerateine-2". It is evident from the analyses that a-keratose and kera-
teine-2 are not derived from precisely the same original fraction but there
appears to be some " overlap". There is thus reason to think that
numerous separate polypeptides exist in the original wool and that the
various solubilization procedures sample these differently.

THE KERATINIZATION PROCESS 243

Soluble derivatives of feather and other keratins. Feather contains less
cystine (6-8%) than hair (16-18%) or wool (11-13%) and is more readily
dissolved (Jones and Mecham, 1943). About 80% dissolves in 10 M urea,
with O'lM mercaptoethanol and 0*2 M lithium chloride and Ward et al.
(1946) estimated a molecular weight of 10,000. Ward, High and Lundgren
(1946) also examined the protein-detergent complex which is dissolved in
bisulphite and sodium alkylbenzenesulphonate. They found a number
average molecular weight of 40,000 (50% detergent). Woodin reports the

Table 18. Amino Acid Composition of a Keratin Derivative

Extracted from Merino 64's Quality Wool.

(From Simmonds (1958). Reproduced with permission.)

Wt./lOO g

Wt. residues/

No. of residues

Amino acid

dry protein

100 g

per m.w.
15,000

alanine

4-01

3-20

6-75 ~ 7

arginine

9-72

8-71

8-19 8

aspartic

7-81

6-75

8-81 9

amide

2-02

18-90 19

£ cystine

5-67

4-82

14-16 14

glutamic

18-79

16-49

19-16 19

glycine

4-14

3-15

8-29 8

histidine

0-77

0-68

0-75 1

woleucine

3-21

2-77

3-68 4

leucine

8-54

7-37

9-78 10

lysine

3-70

3-24

3-80 4

phenylalanine

3-86

3-44

2-78 3

proline

4-59

3-87

5-98 6

serine

7-62

6-33

10-89 11

threonine

6-59

5-59

8-26 8

tryptophane

0-79

0-72

0-58 1

tyrosine

4-50

4-05

3-73 4

valine

4-76

4-03

6-10 6

isolation of more definite units with molecular weights of the order of
10,000 and of a considerable asymmetry (1954a and b). See also p. 163.

Apart from the work of Jones and Mecham (1943) little effort has been
made to examine the other hard keratins.

There is still, in spite of recent efforts, particularly on the part of the
Australian group, much that remains to be explained in the results obtained
from the various solubilized keratin proteins. In their detailed study in
which viscosity, sedimentation and diffusion measurements were made on
both oxidized and reduced products, O'Donnell and Woods (1956a and b)

244 KERATIN AND KERATINIZATION

showed that all these preparations were more polydisperse and unstable
than had been previously described. They concluded that possibly the
solutions contain monomers in equilibrium with aggregates and that there
may be complicating factors in changes of shape and solvation under
various conditions. Further, in most cases some peptide bonds are broken
as well as disulphide bonds and special precautions must be taken to
prevent further hydrolysis with time. Moreover, most solutions show,
superimposed on these changes, others due to a progressive aggregation of
material with standing. They are not hopeful that the various effects due
to aggregation and disaggregation, hydrolysis, changes in shape and
solvation can be distinguished. In later articles, Woods (1952) and
Gillespie et al. (1960) take a somewhat gloomy view of the possibility of
further analysing these solutions, since apart from the actual experimental
difficulties, the theoretical interpretation of the results in such systems is
also obscure. Nevertheless when the whole of the results obtained from
solubilized wool keratin is reviewed it is clear that several significant
conclusions emerge:

(a) It is possible to extract from wool quantities (50-60% of weight) of
an a-type, fibre-forming protein which are low in cystine and have other
significant departures in composition from the original wool. This may be
termed the a-component ; its oxidized form is a-keratose with a molecular
weight of the order of 50,000-80,000. Kerateine-2 appears to be a related
product. The actual analyses in part support this view.

(b) In addition to the a-component, other less well-defined, probably
heterogeneous polypeptides, of smaller molecular weight (3000-10,000) and
higher in cystine, can be extracted. These are, in their reduced form,
contained in part in Lennox and Gillespie's A-E extracts and, when
oxidized, Alexander's y-keratose. These polypeptides show little tendency
to form fibres.

(c) These various polypeptides may be linked by oxidation of their thiol
groups to reform higher polymers and insoluble, partly-synthetic keratins.

(d) Even in the absence of the possibility of disulphide cross-linking they
still exhibit a marked potentiality for aggregation. This property probably
arises from the multiple possibilities of interaction between the rich side-
chain population of the keratin polypeptides, and may well be of impor-
tance in the formation of the initial fibrous aggregates in the follicle prior
to the formation of disulphide bonds.

The similarities between a-keratose, kerateine-2 and Lindley's (1947)
cetylsulphonic acid (CSA)-soluble extract on the one hand and y-keratose
and the CSA-insoluble fraction on the other, have also been pointed out by
Earland and Wiseman (1959) and are brought out in their table (Table 19).
The CSA method is effective although the disulphide bonds remain largely
intact.

THE KERATINIZATION PROCESS

245

Table 19.* The Amino Acid Composition of Hydrolysates of Fractions
from Merino 64's WooL.f

Amino acid

Whole wool

a-Keratose

CSA-

soluble

Kerateine-2

y-Keratose

CSA-
insoluble

alanine

3-51

4-12

4-83

4-0

4-22

2-58

3-3

amide-N

7-46

—

10-25

—

11-56

11-05

—

arginine

20-32

191

20-8

—

21-12

19-0

15-5

aspartic acid

4-24

4-38

6-25

—

5-68

1-79

—

cystine

7-93

—

3-72

5-20

4-33

14-5

20-30

glutamic acid

8-58

8-48

10-9

—

10-82

5-87

—

glycine

5-80

6-29

5-16

—

5-27

4-97

—

histidine

1-46

1-91

1-24

—

1-42

1-57

—

woleucine

1-97

2-44

2-49

—

2-24

2-14

—

leucine

4-90

5-85

7-30

9-2

6-01

2-55

3-6

lysine

3-25

3-92

4-60

—

5-03

1-03

—

phenylalanine

1-75

2-07

1-94

1-7

1-72

115

0-70

proline

5-33

5-05

2-69

1-0

3-66

9-85

7-6

serine

7-25

7-87

6-70

—

6-47

9-70

—

threonine

4-61

4-70

3-45

—

4-44

7-46

—

tyrosine

2-97

2-62

2-44

2-0

2-46

1-41

1-3

valine

3-57

4-16

3-98

4-0

3-55

4-15

3-5

* per cent N of total-N.

| Earland and Wiseman (1959).

The location of the cystine residues

The evidence, which has been reviewed up to this point, is adequate to
establish that during the later stages of the process of keratinization the
thiol groups (SH) of cysteine residues in some polypeptides are oxidized to
yield cystine bridges or disulphide cross-linkages ( — S — S — ). There are,
however, two different views as to the location of these cross-linkages. The
simpler view is that keratin is a single uniform protein of more-or-less
definite composition and that half cystine residues are distributed along
the component polypeptide chains in a definite manner as are the residues
(Fig. 101) of other proteins. In a physicochemical sense the resulting
cross-linked system is closely analogous to the artificial, three-dimensionally
cross-linked polymers, such as vulcanized rubber, to which it is often
likened. This picture gives an explanation of the major facts : the stability,
insolubility and the sensitivity to disulphide rupture. It is compatible
with the evidence that the sulphur enters after the establishment of the
fibrous structure, if we assume that the original polypeptides contain less
cysteine and that other residues (e.g. serine) are later converted into
cysteine residues by a topochemical reaction and then cross-linked. This
view of the location of the cystine is widely accepted and is adequate for
the interpretation of most physical and chemical experiments on intact
material.

The second view tries to take into account a number of other obser-
vations which suggest that keratin is not a uniform material at the

246

KERATIN AND KERATINIZATION

macromolecular level, that in fact it consists of two components : one of a
fibrous, relatively-crystalline character, the other less well-organized, and
containing the larger part of the cystine residues.

This view was implicit in the earlier X-ray work (Astbury and Woods,
1933 ; Astbury, 1943) which showed clearly that the crystalline fraction of
keratin fibres, as judged by the persistence of the a-pattern, was unaltered
by a variety of chemical treatments which involved the disulphide bridges
and the acid and basic side chains of the protein. It was recognized that

s-s-

■s s

[a]

-S-S-

(b)

Fig. 101. The classical view of the location of disulphide bonds (S2) as

cross-linkage between polypeptide chains. In (a) all the linkages are

shown as side chains. In (b) the possibility of the linkage occurring in

the main chain direction is indicated.

most of these reactions must take place in the " non-crystalline " regions.
The size of the various reflections of the X-ray diagram and the greater
strength of reflections arising from planes parallel to the fibre axis showed
that the crystallites were small (< 100 A wide), long and thin with their
long axis parallel to the fibre axis. They may well be identical with the
filaments seen in electron micrographs.

More direct, morphological evidence was later obtained by examining
fragments of disintegrated fibres. A system of filaments (microfibrils)
embedded in an amorphous matrix was proposed by Farrant, Rees and

THE KERATINIZATION PROCESS 247

Mercer (1947) to account for the appearance of fragments of reduced and
ethylated wool fibres produced by enzymatic digestion.

The analysis of " solutions " of keratin was not at first productive but
with the introduction of peracetic acid as the agent for breaking disulphide
bonds (p. 238) clear-cut results were obtained. The separation of a fibre-
forming a-component and a non-fibrous sulphur-rich y-component from
solutions of oxidized wool led Alexander and Hudson in their book (1954)
to propose unambiguously a crystallite-plus-matrix model for keratin.

:&: Matrix (^-keratin)
# Filament {a -keratin)

Fig. 102. The " filament plus matrix " model for a fibril of fibrous

keratin. The filaments consist of bundles of a-helices (see p. 183) and

are embedded in an amorphous matrix with a higher cystine content.

See also Fig. 98, p. 225. (Birbeck and Mercer, 1957.)

Strong, directly-morphological evidence from intact material was later
forthcoming when it became possible to examine electron-microscopically
(see p. 223) sections of developing hairs. Birbeck and Mercer (1957)
concluded from the pattern of osmium deposition in the hair fibrils (Plate

15) the existence of a system of filame.its embedded in a sulphur-rich
matrix (Fig. 102). This was confirmed in fully-hardened hair, wool (Plate

16) and quill by Rogers (1959). Brody (1959) (Plate 17) demonstrated a
similar pattern in the epidermis.

The existence of a filamentous system prior to, and distinct from, the
finally stabilized keratin, is demonstrated by X-ray methods (p. 211), by

248 KERATIN AND KERATINIZATION

histochemistry (p. 214), by chemical (p. 236) and mechanical separation
(p. 231), followed in some cases by analyses proving a low-sulphur content.
Alexander has also pointed out that the existence of several distinct
methods of producing supercontraction, which are discussed on p. 259 et
seq., also shows that keratin is divided into cystine-stabilized and hydrogen-
bond-stabilized fractions (1951).

A summary of the outstanding properties of the two components of what
may be termed the filament-plus-matrix model (Fig. 102) is as follows:

(a) Filaments (cc-component)

Dimensions. About 60-80 A diameter ; length indefinite, at least > 2000
A in wool and hair.

Internal structure. Composed of a-type fibrous protein (see Chapter 5)
(Figs. 77, 78 and 79).

Composition. Cystine content is less than whole keratin (Table 15).
Serine, threonine, proline lower, acidic residues higher than whole fibre.

Macromolecular composition. May contain a primary peptide of molecu-
lar weight between 50,000-80,000. Its relation to the filament is obscure.

(b) Matrix (y-coMPONENT)

Amorphous, shorter chain polypeptides having a higher content of
cystine, serine, threonine and proline.

Molecular structure. Probably not a-type, may be irregular. The ratios
of <x:y is of the order of 1:1 in wool and hair to judge from electron
micrographs. From the analysis of extracts of oxidized wool it would
appear that the a-component may amount to 60% of the total.

It seems reasonable to suppose that all the hard keratins will possess a
similar fine structure (Rudall, 1952). Fibrils essentially similar to those
noted in wool and hair have been demonstrated in feather both in the germ-
inal layers of the follicle and in the fully keratinized material (Mercer, 1958 ;
Mercer, unpub.). Porcupine quill possesses a most regular structure
(Rogers, 1959a and b). In the present state of our knowledge it is possible
to suppose that the y-matrix protein is an entirely new protein which the
cells of the tissue commence to make in the keratinization zone in response
to the altered conditions prevailing there and that it secondarily deposits
on the bundles of a-filaments to form a cementing matrix. Or it may be
thought to be synthesized directly on the filaments establishing an actual
peptide linkage with existing polypeptides. The first alternative would seem
the more probable since the two proteins are separable when the cystine
links are severed. It is suggestive that, in the hair cuticle cells at this same
level, a very similar amorphous high-sulphur keratin is also synthesized.

With the soft keratin of the epidermis the situation is by no means so
clear. This material, composed largely of transformed keratohyalin, has an

THE KERATINIZATION PROCESS 249

irregular fibrous texture (Fig. 99), in which a definite fibrillar and non-
fibrillar component can be discerned. Epidermin, the precursor of skin
keratin, shows several components in the analytical ultracentrifuge, none of
which can be related with certainty to the a- and y-components of hair.
Moreover, the lower sulphur content, the sudden nature of the change
from amorphous keratohyalin (or trichohyalin) granules into the fibrous
form and its consolidation as a resistant birefringent fibrous layer, all are
in some contrast with the consolidation of the prefabricated fibrils of
a-keratin in the hair. If a y-component exists it must be smaller in amount
and perhaps not so easily distinguished on account of the lesser regularity
of the structure. Brody (1959b) is of the opinion that a y-component is
derived from keratohyalin, but the evidence that the granules contain
cysteine is not good (see p. 230). Nevertheless it is possible to conclude
that in most instances the vertebrate keratins are duplex structures pro-
duced by embedding a primary system of filaments (usually a-type but,
as feathers and claws show, a /3-type is possible) in a matrix of short-chain
polypeptides rich in cysteine residues whose conversion into cystine resi-
dues stabilizes the formation.

The concept of keratinization as a process subsequent to a primary
process of fibril formation harmonizes very well with the broader view-
point which presents the a-type proteins as the common intracellular fibre
type which, by secondary modifications, is adapted to a variety of functions.

Physicochemical properties and keratinization

A great deal has been learned concerning the stabilizing bonds produced
during keratinization by the study of the dependence of the mechanical
and dimensional properties of the hardened tissue on the physicochemical
environment. For the study of these " mechanochemical " properties, to
adopt Speakman's useful expression, the tissue chosen needs to have a
convenient form, and in fact the greater part of these experiments have
been made on hair, wool, feather and horn. Since information concerning
the properties of wool and hair is of value to both the textile and the
cosmetic industries, the amount of work carried out is enormous and it
would be impossible to review it here. Reference may be made to the book
by Alexander and Hudson (1954), the review of Ward and Lundgren
(1954) and papers by Speakman (q.v.).

When stretched under well-defined conditions hairs yield characteristic
stress-strain curves (Fig. 70). Their dependence on temperature (Fig.
104) and water content (Fig. 103) shows that the effect of water and a rise
in temperature is to loosen those internal bonds which are opposing
extension and to reduce the work of extension. Various theoretical
attempts have been made to explain the shape of the stress-strain curve.
As already described in Chapter V (p. 172) the normal curve shows several

250

KERATIN AND KERATINIZATION

steps (Fig. 70) which according to the theory of Astbury and Woods (1933)
mark the successive extension of " phases " differing in the force required
to stretch them (see p. 174). American workers (Burte and Halsey, 1947),
with the complex elastic behaviour of polymers in mind, have been more
inclined to regard the curve as being that of a single, uniform cross-linked
polymer. No entirely satisfactory quantitative account of the whole range
of elastic behaviour has been given, but some success has been obtained in

r.h %

20/ / /
40/ / /
60/^ /

100^/

Fig. 103. Influence of relative humidity on the stress-strain curve of

wool at 22° D with constant rate of loading of 18g/min (Peters and

Woods, 1956).

the simpler problem, the description of the elastic behaviour of fibres in
which most of the internal molecular restraints have been removed
(Burte and Halsey, 1947; Peters and Speakman, 1949).

To assess numerically the effect of a chemical treatment on a keratin
fibre, Speakman (1934 and 1947) introduced what may be called the " 30%
work index", which is defined as the ratio of the work required to stretch
the treated fibre 30% to the work required to stretch it 30% before
treatment. The choice of 30% (or in some instances 20%) is based on
the experimental fact that the stretching of wool fibres in the range of

THE KERATINIZATION PROCESS

251

0-30% is completely reversible. Beyond this point the fibre may recover
its length after stretching, but it is thereafter easier to stretch, showing that
irreversible damage has occurred. The smaller the index the greater the
reaction between the fibre substance and the reagent.

>. 10

■S

fo°c
I zoy

1 A0/

80^""'^

100' ■

Fig. 104. Stress— strain curves of wool fibres in water at various tempera-
tures. From Peters and Woods (1956) by permission.

Salt Linkages

All the keratins contain a high proportion of residues of the diamino and
dicarboxylic acids, and in wool these are approximately equal in amount
(p. 219). In the neutral condition when both types of group are ionized:

P.COO- +H3N.P (P = polypeptide chain)

there is the possibility of electrostatic attraction between the two which
could contribute to the cohesion of the material. This electrostatic force is
referred to as a " salt-link". Experimentally the probable existence of such
salt link combinations is demonstrated by numerous experiments made by

252 KERATIN AND KERATINIZATION

Speakman and his collaborators. The dependence of the 30% index on
pH is shown in Fig. 106 (Speakman and Hirst, 1933; Speakman, 1934)
reveals a weakening of the fibre towards the extremes of pH suggestive of
"hydrolysis" of salt-linkages, that is a weakening of electrical attraction
following the discharge of ionized groups by combination with H+ or
OH-. Speakman and Hirst (1933), showed that the reduction in work in
acid solutions was proportional to the amount of acid combined with the
fibre, i.e. proportional to the number of salt-links put out of action. The
amount of acid or base bound as a function of pH is shown in Fig. 105 and
it may be compared to the work-reduction in Fig. 106.

There is room for some difference in opinion over the theoretical
interpretation of these findings or even whether the term " salt-link " is
really appropriate. Nevertheless, from the point of view of the effects of
keratinization, it is significant that the reduction in cohesion of the fibre
and its swelling (and its combination with acids and bases which is inter-
related with these) are minimal in a broad range of pH values including
neutrality. The material is in a sense " buffered " against environmental
changes in the range of variables which, biologically speaking, might well
be encountered by the integument.

Disulphide Bonds

In the literature dealing with the interpretation of the elastic properties
of wool the relative importance of disulphide bonds and hydrogen bonds
has been much debated. Speakman in particular insisted on the pre-
dominant role of the sulphur bridge; Alexander (1951), Elod and Zahn
(1944 and 1949) principally did much to direct attention to the importance
of hydrogen bonding. Speakman and collaborators (Speakman, 1934, 1936
and 1947) have demonstrated both the weakening of fibres when the
disulphide bond is broken and their recovery when it is reformed.

The existence of disulphide bonds and their effect on fibre properties is
revealed most clearly by the series of elegant experiments of Harris and his
collaborators (Patterson et al., 1941; Harris and Brown, 1946) who used
thioglycollic acid to reduce cystine bridges:

P— S— S— P + 2 HS.CHo.COOH -> 2 P.SH + CH.,COOH

I
S2.CH2.COOH

and investigated not only the properties of the reduced fibres, but also those
of the fibres in which the cross-linkages had been rebuilt (a) or blocked (b)
by reacting the reduced thiols with alkylhalides :

2 P.SH + Br.CH2CH2Br -> P.S.CH2.CH2.S.P + 2 HBr (a)

P.SH + Br.CH2.CH3 -> P.S.CH2CH3 + HBr (b)

THE KERATINIZATION PROCESS

253

No oBded soli
o 0°

o 40°
e SIT

1 02

•\

o
o

^"=5^

**rx

«>

^%;\^

\ v \\

■ \v

1-0.

\ N

\
\

l 15

Fig. 105. Combination of wool with hydrochloric acid and with potas-
sium hydroxide as a function of pH and temperature in the absence of
added salt. (The dotted lines show the correction by Harris and
Rutherford for the reaction of alkali with cystine.) (Reproduced by
permission.)

X"*^,^

Z""5-

°s

Fig. 106. The effect of pH on the reduction in work to stretch a fibre by
30% for normal (O) and deaminated hair ( X ). (Speakman, 1934.) Peters
and Woods, 1956. (Reproduced by permission.)

254

KERATIN AND KERATINIZATION

The effect of the reduction of the disulphide bonds on the work of ex-
tension is shown in Fig. 107. Reduced wool or " blocked wool " (reaction
(b)) was found to be more easily stretched, more readily dissolved and less
resistant to enzymatic digestion than normal (Geiger et ai, 1941 and 1942).
Resistance was restored when new cross-linkages were introduced by
alkylation (reaction (a)).

% Cystine bonds broken

Fig. 107. Relationship between wet strength of wool fibres and the
number of disulphide bonds (Harris and Brown, 1946).

These reactions form the basis of current hair-waving treatments in
which the hair, softened by thioglycollate reduction, is deformed into the
desired shape and held there until cross-linkages have reformed by
oxidation (McDonough, 1952).

Hydrogen Bonds

Even when the greater part of the disulphide bonds are broken, the
strength of dry fibres is not greatly reduced and the fibre form may be
retained when the fibre is placed in solutions of pH< 9 at normal temper-
ature. Above this pH much of the keratin may enter solution (see p. 240).
The stabilization persisting after reduction is attributed to hydrogen
bonding, the presence of which is directly revealed by infra-red absorption
spectra (Chapter V, p. 196). When steps are taken to break the hydrogen
bonding in a previously-reduced fibre, a characteristic contraction in
length occurs. Following Astbury and Woods, this is usually called

THE KERATINIZATION PROCESS

255

super contraction since it may follow also on the ordinary recovery of length
occurring when stretched fibres are released (Fig. 108).

Supercontraction is a property that the keratin fibres share with other
systems containing long oriented molecular chains. The most probable
configuration of a long, free, flexible molecular chain in solution is a
random coil, for in this condition the entropy is a maximum. However,
less probable configurations may be assumed as a result of the molecule
interacting with other molecules or with other parts of itself. These inter-

;

^ t ; loot

^•90°

N \

\K ^A~ '80° \j

H\/ L* — ' TV

1 64"

^c^ | V

50°

laxation time

Fig. 108. Recovery of Cotswold wool in steam, after being held at 50%

extension in steam or hot water for given times of relaxation. Times to be

multiplied by 3 for the 40% (Astbury and Woods, 1933).

actions provide the " forces of crystallization " and thermodynamically the
reduction in internal energy resulting from these interactions opposes the
randomizing effects of entropy. A system of long, flexible molecules with
very weak interaction between the chains may constitute a rubber. The
form of the molecules and of the bulk specimen in the rubbery state is
largely controlled by the entropy factor, i.e. the contractile force opposing
extension arises from the tendency of the extended molecules to return to
a shorter, more probable configuration.

Among the protein fibres a large contribution to the internal energy
factor is provided by hydrogen bonding which can occur between the
numerous peptide groups (Fig. 109). In general terms, where steric
conditions are favourable and the close packing of portions of many chains
is possible, a crystallite stabilized by multiple H-bonds is formed. We may
say then that such systems are stabilized by " crystallite cross-links " to
distinguish the condition from the single-chain covalent linkages due to
cystine bridges. Some protein fibres, collagen and silk for example, owe
their insolubility and stability almost entirely to crystallite cross-linking.

256

KERATIN AND KERATINIZATION

The destruction of the crystallites in such a system, produced either by a
solvent able to penetrate the crystallites or by a rise in temperature suffi-
cient to " melt " them, releases the molecular chains which may then
assume a more probable, less oriented, and shorter, configuration. This
change may be visible as a change in shape of an oriented specimen, and
is accompanied by the loss of other signs of orientation, such as bi-
refringence and the fibre-type X-ray pattern (Fig. 109). The contracted

(a) (b) (c)

Fig. 109. To illustrate the relation between the X-ray pattern given by a
crystallite in an unswollen fibre (a); a fibre (b) in which swelling mole-
cules have penetrated only the amorphous phase and the X-ray pattern is
unaffected and (c) a fibre in which molecules have penetrated the crystal-
lite disorienting the molecular arrangement to give a non-oriented X-ray
pattern and causing a further shortening (supercontraction) of the fibre
length.

material may also acquire rubber-like properties. Perhaps the most
striking example of this is provided by collagen fibres which have less
interchain bonding than keratin and may shorten to as much as one-third
of their original length when heated above a well-defined temperature in
water. This "contraction temperature" is the point at which the crystallites
disperse or " melt"; it is not as well defined in the keratins. On cooling,
contracted collagen spontaneously resumes its original length and the crys-
tallites reform. Clearly the normal triple helical structure is adequately
stabilized by internal energy considerations at normal temperatures.

THE KERATINIZATION PROCESS 257

The temperature at which contraction commences is a useful measure
of the degree of stabilization of a fibre. Most hard keratins require a
temperature above 100°C in water. It is lowered by treatments which
reduce the degree of cross-linking (H bonds, covalent bonds or salt-
linkages) and the temperature difference is a measure of the reduction in
internal stabilization. For comparison of different keratins it is more
convenient experimentally to use solutions of substances (phenol, forma-
mide, etc.) which weaken the internal cohesion due to hydrogen bonding
and bring the contraction-temperature below 100°C. Elod and Zahn (1946
and 1949) and Stoves (1947) examined various hairs in this way and found
that the temperature required to initiate contraction increased with the
cystine content and the fibre diameter.

The careful analysis of the elastic behaviour of keratin fibres by Bull
(Bull and Gutman, 1944; Bull, 1945) and Woods (1946) has thrown
much light on the relation between the internal energy factors tending to
stabilize the oriented structure and the randomizing entropy factors. It is
shown in thermodynamics that, if the tension on a specimen is P and its
length L, then for reversible changes :

p= \di)~ T\bi)r UZ/r+ \Wl

Where U is the internal energy, S the entropy and T the temperature.

In these equations the tension P is expressed as the sum of the two
terms: one Pu depending on the internal energy changes and the other
Ps depending on the change in entropy when L is increased. Since the
latter can be measured by observing the temperature coefficient of the
tension at constant length (dPjdT)L, the two can be determined separately.
The dominant factor controlling the length was shown by Woods and
by Bull to be the internal energy term (Pu) ; even in swollen and relaxed
fibres the entropy term remains small but becomes more important in
supercontracted fibres in which the crystallites are dispersed and the
molecular arrangement more randomized.

The long-range reversible elasticity shown by hairs should not be
thought of as a characteristic of keratin per se. It is rather a characteristic
of the a-type proteins and is related to the a-type molecular structure. The
definitive feature of a keratin is the stabilization based on cystine cross-
linking and this chemical device may be used to stabilize proteins of a
different structure and quite different elastic potentialities (see p. 24).
Feather cells, for example, are similar in fine structure to the cortical cells
of hair and the fibrillar system is also stabilized by cystine cross-linking ;
the keratin is, however, of jS-type which permits of extensions of only a few
per cent. This lack of long range extensibility was one of the direct
indications which led Astbury and Marwick (1932) to the view that the

258 KERATIN AND KERATINIZATION

feather structure was based on more or less fully-extended chains. Some-
what surprisingly the well-oriented /^-configuration of feather does not
supercontract in solvents which both reduce disulphide bonds and break
H-bonds although it swells and partly dissolves. When regenerated in
fibre form, the amorphous threads in water are rubber-like (Ward et al.,
1946).

Meyer et al. (1949, 1952) have criticized the broad picture developed
by Astbury and Woods and asserted that the long range elasticity is really
rubber-like. This criticism seems to be based on a misconception. Meyer's
experiments were carried out on hairs which were strongly swollen and
reduced, i.e. on hairs in which special steps had been taken to reduce the
internal energy factors, and under these conditions the entropy term may
be expected to predominate. The theories of Astbury and Woods were, in
fact, developed to give an account of the elastic behaviour of fibres in which
the internal constraints were effective. In their theory the swollen fibre
with its reduced interchain bonding and a relative absence of covalent
cross-linking should be essentially rubber-like. In fact this condition has
been demonstrated experimentally by Meyer and Haselbach (1949).

Molecular Configuration in the Supercontracted State

The molecular mechanism involved in supercontraction is complex.
In terms of the earlier interpretation of Astbury (1933) the shortening in
length (about 30%) is due to a further folding of the polypeptide chains
into a shorter configuration and this view still prevails although no precise
picture of how an a-helix shortens has been proposed.

Strongly supercontracted fibres usually yield an unoriented £-type
X-ray pattern which is not very helpful. In some instances an oriented
a-pattern may persist in fibres shortened by 20%, showing that the
contraction is occurring in the non-crystalline phase ; in others all signs of
pattern may vanish. These findings suggest that the mechanism of
contraction may not always be the same. The numerous chemical treat-
ments which produce some degree of supercontraction have in common
only the property of reducing interchain bonding.

A classification of types of supercontraction is, however, possible on
structural grounds. The simplest situation is perhaps the most drastic,
e.g. boiling in solutions of bisulphite, which leads to a disoriented /3-type
of configuration (Astbury and Woods, 1933; Whewell and Woods, 1946).
The explanation here is that as a result of a chemical weakening of inter-
chain interaction, the main chains, acting as mobile individuals under the
influence of thermal agitation, are able to assume a shorter and more
random configuration. On drying these randomized chains may in part
recrystallize in the ^-configuration.

The second type of supercontraction is that which results from hydrogen

THE KERATINIZATION PROCESS 259

bond rupture and the dispersion of the crystallites without rupture of
disulphide bonds. In this form the X-ray a-pattern fades and the con-
tracted fibre shows no wide-angle pattern. Characteristically the con-
traction may be reversed by washing out the reagent; the crystallites
reform, the fibre resumes its original length and the a-pattern returns.
Examples: supercontraction in cuprammonium solution (Whewell and
Woods, 1946), lithium bromide (Alexander, 1951), phenols (Zahn, 1947)
and formamide (Elod and Zahn, 1944). A third type of contraction occurs
when the disulphide bond system is destroyed and when the conditions of
contraction are not too violent (<20% contraction). In this case the
a-pattern may still be elicited from the contracted fibre. Examples:
contractions in dilute caustic soda (Whewell and Woods, 1946) or after
peracetic acid oxidation (Alexander, 1951). When conditions are more
drastic (higher temperature or longer treatments) the crystallites may be
affected and the more strongly-contracted fibre gives a jS-pattern.

Astbury and Woods, Whewell and Woods and Alexander all recognized
that these results show that, in a sense, the keratin may be divided into two
parts which may be induced to contract more or less independently. This
separation can now be understood in terms of the filament-plus-matrix
model. Disulphide-bond destruction converts the matrix into a viscous
material which facilitates chain movement both in itself and in the filaments.
If a limited contraction of the non-crystalline chain segments occurs, the
fibre shortens and the a-crystallites persist. In the other case, when the
disulphide cross-linked matrix is intact while contraction is induced by
freeing the chains in the crystallites, a contraction results with crystallite
destruction which is reversible, because the over-ruling macromolecular
organization is preserved by the cross-linked matrix which envelops the
filaments.

Another form of supercontraction is that which leads to the formation of
a cross-jS configuration in the contracted fibre. This may be the case when
contraction is produced under mild conditions in reagents which both
loosen the matrix by reducing disulphide bonds and also disperse the
crystallites by rupturing hydrogen bonds, e.g. strong solutions of urea
containing bisulphite (Mercer, 1949a). See Fig. 84 and Plate 2B. The
cross-/3 pattern has already been discussed on p. 200 et seq.

The Setting of Hairs

The study of the important phenomenon of set, by which is meant the
more-or-less permanent retention of a deformed state, has shown that the
same factors which stabilize a fibre in its natural state are also those which
operate to maintain a stretched state or set. A hair which is stretched in
water, relaxed and dried will partly retain the stretched length. It quickly
returns to its original length in water (Fig. 108) and more rapidly in

260 KERATIN AND KERATINIZATION

solutions of high or low pH. Evidently this set is being maintained only
by salt-bonds and hydrogen bonds (Speakman, 1934; Woods, 1933). A
more permanent set is induced by relaxation at higher temperatures under
conditions in which the crystallites are transformed into the /3-modification
(Astbury and Woods, 1933; Woods, 1933). This type of set can be
released by strong solutions of urea (Rudall, 1946), and here therefore
H-bonds are obviously the factors stabilizing the jS-crystallites. Permanent
set is defined as a set which is not relaxed by prolonged steaming or by
solutions which rupture hydrogen bonds, i.e. it is a set which is sustained
by covalent cross-linkages analogous to the disulphide bonds which are
effective at the original length. These cross-linkages may be reformed
disulphide bonds (Speakman, 1933) or bridges introduced between
reduced disulphide bonds by dihalides (see p. 252) or linkages apparently
formed between COOH groups and amino groups in steamed fibres
(Speakman, 1933).

Taken together all these physicochemical methods provide semi-
quantitative measures of the contribution of the several cross-links:
salt-linkages, disulphide and hydrogen bonds, to the stabilization of the
keratinized fibres, which are in good agreement with conclusions reached
on other grounds.

Cell membranes in keratinized tissues

The Membranes and Cellular Adhesion

Up to this point we have concentrated attention on what happens to the
intracellular proteins during keratinization. Other constituents of the cells
also undergo changes during keratinization. The cell membranes, in
particular, play an important role in maintaining the hardened structure.
These membranes and their behaviour during the establishment of tissues
have been discussed already in Chapter III; that they are also important
after keratinization is proved by many experiments.

Electron microscopy has shown that the fibrils of keratinizing cells are
wholly intracellular and that no intercellular connexions composed of
fibrils cross from cell to cell binding the mass together. Such " bridges "
were often described in earlier works, but their true nature is now better
understood. For many histologists who feel that cells must be held to-
gether by " string " rather than "sealing wax", they had a strong fascin-
ation. In fact cells are stuck together, and for this reason the properties of
the adhering surfaces, and of the adhesive, are as important in maintaining
the whole formation as the hardened cell contents themselves. This is
shown very simply by digesting a tough keratin, such as hair, by means of an
enzyme (trypsin) which removes the membranes and cement. The
tensile strength falls rapidly (Elod and Zahn, 1946), and in a few days the
fibres drop apart. Examination of the residue by a variety of means

THE KERATINIZATION PROCESS 261

shows that the keratin itself is unchanged (Mercer et al., 1956); sections
examined electron-microscopically show that the bundles of keratinized
fibrils are of normal appearance and reveal that the components which have
been removed are the cell membrane and intercellular cement. The loss
in weight is of the order of only 10% (Elod and Zahn, 1946) but it repre-
sents the vital links connecting the chains of cells.

A further important characteristic of these altered membranes is that
their chemical character is complementary to that of the keratinized protein.
That is to say, chemical conditions which soften or dissolve keratin have
little effect on the membranes. The strength and weakness of keratin
itself lies largely in the disulphide bond which is peculiarly vulnerable to
reduction, hydrolysis and oxidation. However, the system of membranes
resists these actions to a far greater degree. It is found that, when keratin
is dissolved (by the methods described above) the insoluble residue
consists largely of membranes (Mercer, 1951 and 1953) (Fig. 112) (Lager-
malm et al., 1951).

Considering their biological origin the chemical resistance of these
membranes is remarkable. They are not dissolved by the following strong
reagents, which include both reducing agents and hydrogen bond breakers:
5 N caustic soda, 8 M urea, 8 M urea containing thioglycollic acid or
sodium bisulphite at pH 10, concentrated formic acid, 10% sodium
sulphide, and aqueous peracetic acid followed by 0*1 N alkali. On the
other hand, when not protected by being incorporated in a solid, intact
tissue, they are rapidly digested by proteolytic enzymes.

These properties show that, while a protein constituent is certainly
present, the resistance cannot be due entirely to hydrogen bonds or
disulphide bonds of the type found in keratin. It is perhaps permissible
to see that a certain biological advantage is gained by enclosing keratin in
small sacs which resist dissolution by precisely those reagents most
injurious to their contents.

The little known of the composition of biological membranes (p. 37)
does not help to explain the changes which could convert them into the
singularly-insoluble form they assume in keratinized tissues. That a
protein moiety is present is shown by the dissolution by proteolytic
enzyme; the several reports (Corfield et al, 1958) of amino acids found in
membrane hydrolysates confirm this. Matoltsy (1957) has reported finding
the following amino acids in membranes from human skin : glycine, valine,
leucine, woleucine, serine, threonine, aspartic and glutamic acids, arginine
lysine, histidine and methionine. Other analyses indicating protein have
been made on the particularly-toughened membrane obtained from
Allworden sacs (p. 267) on wool fibres. Corfield, Robson and Skinner
(1958) carried out a complete amino acid determination of the residue
remaining after oxidized wool is extracted with ammonia (referred to as

262 KERATIN AND KERATINIZATION

/3-keratose), which consists largely of membranes (p. 270) and found it to
resemble whole wool. The few small differences observed are not suffi-
cient to explain the insolubility. All these analyses must, however, be
treated with some reserve unless the absence of unremoved keratin in the
membrane preparations is demonstrated.

Since disulphide bonds and hydrogen bonds seem inadequate to explain
the insolubility (see above), suggestions have been made that some new
bond is present, e.g. a type of tanning linkage as found in melanin or the
insect cuticle. It may be significant that melanin granules have a similar
chemical resistance to the hardened membranes. Residues consisting
largely of membranes have been examined by X-ray diffraction and were
found to give only an unidentifiable pattern with some /^-characteristics
which, after such a chemical treatment, is of little significance.

The Morphology of the Membranes of Keratinized Tissues

The special nature of these membranes is made clear by examination of
their fine structure.

The intercellular spacing commonly found between cell membranes in
normal tissues is about 150 A wide (p. 41) and a spacing of this order is
found between cells in the upper bulb of the hair follicle and in all other
germinal tissues. As the cells of the presumptive cortex fill with fibrous
keratin, this spacing widens and may reach 400 A. At the same time the
outlines of the cells become more wavy, so that the surfaces themselves are
more interlocked (Birbeck and Mercer, 1957; Rogers, 1959a and b). In
material fixed in osmium tetroxide the membranes are dense and the
intermembrane material contains light and dark bands. In the final
keratinized cortex these relative differences in density still persist. These
observations prove that the amount of intercellular cement is increased
during keratinization and that it undergoes some chemical modification.

More elaborate changes associated with membranes are visible in the
cuticle and the inner-root sheath (Birbeck and Mercer, 1957a and b;
Rogers, 1959a and b). The intercellular space between the cuticle cells also
widens and in this instance several dense layers, separated by lighter layers,
appear transiently and later fuse to give a single broad dense layer. Be-
tween the various cells of the sheath, single broad sheets are deposited
(Plate 6B). Again while we have no idea of the chemistry behind these
appearances, they show that changes are going on in the intercellular
spaces and we may assume they they are related to strengthening the
adhesion.

Somewhat similar events occur in skin. The localized " studs " or
desmosomes have been described (p. 83). In the granular layers and
higher, the contacting cell surfaces may become extremely wave-like
producing a very considerable degree of quasi-regular interlocking. The

THE KERATINIZATION PROCESS 263

intercellular space widens (Fig. 98) and bands are deposited (Odland,
1958; Mercer, 1958; Horstmann and Knoop, 1958; Pillai, I960). When
a cell is finally shed the parting may occur either between the cells, i.e.
due to failure of the cement, or in the bulk of the keratin due to its
separating into fibrils (Plate 6A).

The purely mechanical consequences of the changes in the geometry of
the surfaces of contact of the cells deserve special emphasis. Not only is
the area of contact and therefore the total adhesion vastly increased by the
formation of wavy surfaces, but the interpenetrating crests may develop to
the point where they constitute veritable " press-studs", which after the
hardening of their contents, literally lock the cells together. The enhance-
ment of these surface irregularities in the hair may be a consequence
simply of the continued addition of filaments to centres near the surfaces
of the cells. In the epidermis the numerous desmosomes and their
associated fibrils appear to cause the buckling of the membranes (Plate 6A).

The Fate of the Intracellular Apparatus during
Keratinization

Much of the intracellular apparatus either consists of membranes or is
enclosed in membranes. We have mentioned above that there are reasons
for supposing that all these membranes have a similar basic constitution
and, it is of interest to note that, during keratinization of the cell, a portion
of the cellular apparatus undergoes changes in solubility similar to those
affecting the external cell membranes and is found along with these among
the resistant residues when the keratin is extracted (Figs. 110 and 112).

The exact fate of the cellular apparatus — nuclei, mitochondria, RNA,
etc., in keratinizing cells is obscure (see p. 220). In a structural sense the
nuclei can be followed through the keratinization zone and remnants can
be demonstrated in the emerging hair. They elongate probably passively
with the cell in the upper bulb, and are here Feulgen positive (Fig. 97b);
but this reaction fades during keratinization although a "structural residue "
is still visible electron microscopically. Spier and Van Caneghem (1957)
report increased DNA-ase activity in this zone, and presumably the DNA is
depolymerized. Bolliger and Gross (1952 and 1956) found many possible
low molecular weight breakdown products (pentoses, uric acid and other
purines) of nucleic acid in hair. When keratin is chemically extracted from
hair, a portion of the insoluble residue consists of long, thin, chemically-
modified remnants of nuclei (Mercer, 1953).

In the light microscope the nuclei are seen to shrink and grow more
dense and are said then to be " pycnotic." Electron micrographs show
first a thick gathering of dense material beneath the membrane and some
signs that material may be being shed into the cytoplasm.

Pycnosis is a degenerative condition recognized by an increase in both

264 KERATIN AND KERATINIZATION

the acid and base-binding capacity of nuclei accompanied by a decrease in
nuclear volume. It is supposed that the basic proteins (histones) normally
combined with the DNA become dissociated from it, each then becoming
more easily stainable, and that there is also a loss in water-binding power.
The total DNA content remains more or less constant.

According to Bern et al. (1954 and 1957) the nuclear changes during
keratinization are, however, different from this and are not simply de-
generative. The nuclear volume at first increases to be followed by a
collapse in the final stages of cell condensation. During the swelling stage
the nuclei are less stainable and the DNA content decreases. This fall in
DNA continues during the phase of active keratin synthesis. The histone
content seems to remain constant and these proteins in an altered form
may represent in part the "nuclear remnants" found in the keratinized cell.

A production of masses of keratohyalin by the nucleus or even the actual
dissolution of that body has been envisaged by Hinglais-Guillard (1959)
as a result of her study of keratinization in the cervical epithelium of
women. This tissue undergoes a cyclic change of functional activity in
phase with the other sexual tissues (see p. 144), the cells oscillating
between a keratinizing condition and one in which glycogen is accumulated.
In the keratinizing phase the nuclei of the superficial cells reveal a curious
clumping of their contents and rather similar lumps of material are to be
seen in the adjacent cytoplasm.

These observations are not incompatible with the possibility that the
nucleus is playing a different synthetic role in the later stages of keratin
formation. Evidence, given below (p. 268), suggests that a peculiar
cystine-rich protein may be formed at this stage and its formation may
involve the nucleus. The observation of Fell and Pelc, already referred to
(pp. 63 and 220), that radioactive sulphur compounds on injection appear
first in nuclei may also mean that sulphur-containing amino acids are built
into protein precursors in the nucleus. It is known that a small amount of
protein synthesis occurs in nuclei in other sites and that it involves energy
transfers following pathways not involving the enzymes normally partici-
pating in the cytoplasm (Allfrey et al., 1953 and 1957). The situation in the
keratinizing cell is admittedly peculiar — it is largely cut off from supplies of
metabolites and its life as a synthetic unit is drawing to a close — it would
not be surprising then if special mechanisms were called into play. For
example, it is compatible with the little evidence we have to suppose that
here the DNA molecules participate at first-hand (i.e. not through the
intermediary of RNA) in the synthesis of proteins and are themselves
consumed in the process.

The RNA granules of the cytoplasm persist into the keratinization zone
and are lost from view between the masses of condensing fibrils. Histo-
chemically also the cells of the bulb are strongly RNA positive, and the

THE KERATINIZATION PROCESS 265

reaction disappears in the lower part of the keratinizing zone (Fig. 97 (b)).
We have seen that protein synthesis is substantially complete at this same
level.

Mitochondria and various vacuoles (Golgi apparatus) are also lost sight
of as distinct structures between the condensing fibrils. Their remnants
persist as modified membranes and seem responsible for various gaps
which are to be seen in the otherwise almost uniformly-fused mass of
keratin fibrils (Plates 15, 16 and 17). It is not known what enzymes
are involved in the oxidative closure of the sulphur bridges, although
copper is concerned as it is with the oxidases in melanin formation (p. 279).

The hair cuticle

The peculiar type of amorphous keratin of the hair cuticle cells seems to
possess a limited extensibility, since these cells usually part company or
split when the hair is stretched more than 50% (Lehmann, 1943), and when
a many-layered cuticle is present as in fur hairs (Stoves, 1947 and 1943) the
elastic behaviour of the whole hair is considerably modified. Owing to its
fibrous texture the hair cortex may fray and split; against this tendency
the cuticle forms a retaining sheath whose laminated structure is adapted
to this end. The keratin of the cuticle is, in fact, so different from that of
the cortex as to require separate consideration; it represents yet another
distinct product of epidermal cells.

Proceeding inwards from the surface of the fibre (Fig. 110) we encounter
(a) a thickened and altered external cell membrane, which is strongly
attached, through the intermediary of an intercellular sheet, to the mem-
branes of contiguous cells. The whole of this formation of external
membranes forms a strong, chemically resistant skin of such distinct
character that it is referred to as an epicuticle (Lindberg, 1949). (b) Beneath
the epicuticle lies the layer of keratin, its compact amorphous structure
being responsible for much of the mechanical protection. This may be
called an exocnticle. It is not uniform in texture. A layer immediately
beneath the epicuticle stains more deeply after reduction and fixation and
thus seems more cross-linked (Sikorski, 1960). (c) A layer of modified
cellular residue (p. 270) (3 in Fig. 110) (endocuticle) lies next. Its high
resistance to keratinolytic solvents (see pp. 261 and 268) supplements that
of the exocuticle and thus increases the total protection against chemical
action.

Cuticular keratin is distinct from the other products of the cells of the
hair follicle in the following respects.

(a) It is not fibrous. It has rather, during its formative stages (Plate
20A), the appearance of a viscous liquid condensing to give a coherent
amorphous lamella closely adhering to the external cell membrane (Fig.

266

KERATIN AND KERATINIZATION

110). The X-ray pattern (Rudall, 1941) shows no orientation and no
crystallite formation.

(b) It has a higher sulphur content than whole hair (Geiger, 1944). It
forms somewhat later than the mass of fibrils in the cortex, in fact accumu-
lating most rapidly during the later keratinization of the cortex. We have
given reasons above to suppose that at this level the cortical cells have

iLi'iniiiiiiii:::::

(a)

QTrrnrrfi

i i ■
(b)

iMrrmmfi

1+3+4+5

(d)

Fig. 110. The lamellar structure of the hair cuticle cell and the results of
selective removal of components.

(a) 1 . is the resistant external cell membrane or epicuticle ;

2. the layer of amorphous keratin or exocuticle;

3. the inner layer of altered cellular residue;

4. the nuclear residue;

5. the inner cell membrane less altered than 1.

(b) The epicuticular membrane as released by bursting AllwSrden sacs.

(d) The " non-keratinous " residue remaining after removal of the
keratin (2) by oxidization and extraction.

swung over to the synthesis of a cystine-rich, amorphous protein, y-keratin.
It is possible that the cuticle cells are producing a very similar protein
which here, in the absence of fibrils, simply condenses as an amorphous
heavily cross-linked mass.

From the viewpoint of cellular differentiation, it is interesting to find

THE KERATINIZATION PROCESS 267

that, before commencing definitely to form " droplets " of cuticular
keratin, the cuticle cells form small amounts of fibrillar material (Birbeck
and Mercer, 1957).

The laminated structure and the differing chemical nature of the several
component layers are responsible for some peculiar reactions. When hairs
are immersed in acidified chlorine water, which attacks keratin with the
production of low molecular weight osmotically-active substances, the
resistant surface membranes are inflated into bubbles by the water, which
enters the cell (Allworden's reaction) (Fig. 111). The bubbles may develop

Fig. 111. Explanation of the appearance of bubbles on the surface of
hairs immersed in chlorine water (Allworden's reaction). The external
cell membranes are modified chemically and form a very resistant layer,
the epicuticle. When chlorine penetrates it and oxidizes the proteins
within the cell, lower molecular weight soluble compounds are produced.
The membrane is dilated by the entry of water owing to the high osmotic
pressure within the cell.

to an extraordinary degree on hairs with elaborate frilly cuticular scales
such as bat hair (Muller 1939). The study of this phenomenon first led
Miiller to recognize the existence of specialized resistant surface membrane.
Lindberg (1949) obtained clear pictures of the membrane by mechanically
abrading wool fibres covered with well-developed bubbles. Bromine
water is less active. Bubbles form, but the skin is thicker and seems to
consist of the surface membrane and the more resistant portion of the
exocuticle. A variety of other sheaths and tubules can also be obtained
from the lamellar cuticular structures of hair which differ principally in
the amount of keratin remaining attached to the epicuticle (Mercer et al.,
1949; Manogue and Moss, 1953; Lagermalm et ai, 1951).

268 KERATIN AND KERATTNIZATTON

The surprisingly hydrophobic character of the intact surfaces of hairs and
feathers, even after thorough removal of lipids, is presumably due to the
chemical inertness of the modified external cell membrane. When this is
mechanically or chemically damaged the material becomes wetable and
dyes and other large molecules more readily penetrate (Mercer et al., 1949).

Fig. 112. Illustrating the resistant residues which remain after a wool
fibre, oxidized by means of peracetic acid, is extracted with ammonia.
The cell membrances and nuclear residues of the cortical cells are at M
and R, respectively. The external membrane E of the cuticle cells
encloses the swollen fibre. In the paracortical segment P some keratin
remnants may persist; O is the orthocortex (see text).

The medulla

As is apparant from the description given by Auber (1950) (Fig. 113), the
changes taking place in the medulla of hairs are complex. Keratohyalin
granules appear in the differentiated cells and electron micrographs of
rodent hairs show that they change into a fibrous form as elsewhere. The
total amount synthesized is inadequate to fill the cell cavities and much of
the transformed material simply condenses against the cell membranes.
During desiccation intracellular gaps appear (Fig. 113c) and the final result
may be a rather open girder-like structure, light but stiff.

Chemically the medulla resists alkalis and keratinolytic reagents and
may be isolated (undoubtedly altered) by digesting away the keratinized
cortex and cuticle. It thus resembles the altered membranes of these
structures rather than their keratinized contents and morphologically this
seems to arise from a close fusion of fibrous keratohyalin and membranes.
That much of the protein of the medulla is of the trichohyalin type is
supported by analyses, reported by Rogers (1959a), of the medullary cells

THE KERATINIZATION PROCESS

269

Fig. 113. Structure of medulla. Diagrams to illustrate the detailed
formation and structure of the medulla in hair, a, b and c are three
stages in the obliteration of a cell and the condensation of its contents.
A-G illustrates the full sequence of cell forms in which the appearance
and assimilation of trichohyalin (see p. 268) may be followed. Re-
produced from Auber (1950) by permission. The transformation of
trichohyalin into fibrils and its fusion with the cell membranes is indicated.

270

KERATIN AND KERATINIZATIOM

of rabbit fur (Table 20). The cystine content is very low, citruiline is
present and half the side chains are either acidic or basic. It thus appears
quite distinct from other hard keratins and resembles trichohyalin.

Table 20. Composition of Rabbit Fur

and Medulla.*

(N as per cent of total-N.)

a- Amino acid

Fur

Medulla

glycine

4-74

1-42

alanine

2-50

2-78

valine

2-21

1-18

leucine

4-22

7-95

tsoleucine

1-53

1-00

serine

5-38

1-38

threonine

3-02

0-86

phenylalanine

1-38

1-45

tyrosine

1-55

0-80

tryptophane

—

proline

4.49

0-85

citrullinef

0-35

5-2f

lysine

4-64

15-84

arginine

17-60

11-72

histidine

3-81

1-38

aspartic acid

3-29

2-42

glutamic acid

6-84

21-7

cystine

9-11

methionine

—

ammonia

1-32

7-02

* Taken from Rogers (1959a).

t The presence of citruiline may prove to be diagnostic of tricho-
hyalin.

The residues remaining after the chemical extraction of
keratins

Much has been learned concerning the chemical nature of the con-
stituents of keratinized tissues by scrutinizing the residues which remain
after subjecting the tissue to the treatments described above designed to
remove definite components. For example, the extraction of the proteins,
which become soluble after complete disulphide-bond destruction, enables

THE KERATINIZATION PROCESS 271

a useful distinction to be made between keratinized and non-keratinized
components, and to establish what fraction of the tissue actually consists
of keratins.

The most satisfactory procedure is to oxidize the disulphide bonds with
peracetic acid and to extract thoroughly the oxidized material with dilute
ammonia. Thioglycollates at high pH (11-12) have also been used, but
effect a less complete extraction. The undissolved residue contains the
following components recognizable in the light microscope :

(a) The ill-defined and obscurely altered remnants of the cell nuclei.

(b) Many small particles distributed throughout the cytoplasmic space
which are the remnants of the various cell organelles, mitochondria
and Golgi membranes.

(c) The external membranes of the cells which are still held together in
a foam-like formation by the intercellular deposits.

Similar residues remain after extraction by other reagents and the findings
are the same for all types of keratinized tissues (Mercer, 1953; Matoltsy,
1958). In some cases, when the cells assume a specialized form, as in the
hair cuticle discussed earlier, the intercellular remnants may separate
during growth from the keratinized protein and thus form a distinct layer
within the cell which is visible (Fig. 110 (d)) after extraction.

A picture which is almost the " negative image " of the above is obtained
when the material is digested with a proteolytic enzyme (trypsin or papain)
which removes everything except the keratinized proteins. The pro-
cedure is not very satisfactory with the soft keratin, epidermis; with all the
hard keratins (hair, horn, nails, etc.) after prolonged digestion, a clear
delineation of the resistant protein results. The tissue falls apart, since the
cell membranes are removed, and the residue is often referred to as
" cells", which they resemble superficially in shape, although they consist,
in fact, only of the keratinized protein, all other components visible in the
intact state being removed (p. 261). Such preparations should form the
material for analysis of " keratin". More prolonged digestion removes
part of the keratin itself, suggesting that this material is not entirely
uniform in composition. This suggestion, that the " degree of keratini-
zation " may be variable at a microscopic level, is compatible with the idea
that keratinization is a process which takes place gradually and may remain
incomplete and patchy.

A " dissection " of the cuticle cell, as effected by various methods of
extraction, illustrated in Fig. 110, shows clearly the differing properties of
its several layers.

The recognition that hardened tissues contain several components
differing in resistance to chemical attack was the basis of an earlier system
of classification, introduced by Unna (1926) in which keratins were to be
distinguished by the proportions they contained of " keratin A", insoluble

272 KERATIN AND KERATINIZATION

in nitric acid or in sulphuric acid plus hydrogen peroxide, and " keratin
B " which was soluble. The system was little used although it gave rise to
the more useful distinction between hard and soft keratins. Nevertheless,
it would be valuable when classifying keratins to have measures of two
factors, which really were at the root of Unna's ideas : (a) the amount of
keratinized protein present in a tissue relative to non-keratinized material,
and (b) the degree to which the keratin is insolubilized. From the stand-
point of our present-day knowledge, these could be usefully defined as
follows :

D .• f i s ^ i weight of tissue soluble in 0*1 N NH3

Ratio of keratins to total r .... c „, - . on/

= after oxidizing for 24 hr in 2% peracetic

acid

weight

original weight

Ratio of easily soluble . , r . , , , n , ,

, • ^ , J , , , weight of material extracted by 8 M urea

keratin to keratin soluble = «„»» !• i «• -it- i TT -,

• , ,-rn lt^ 0*2 M thioglycolhc acid adjusted to pH 7

weight of peracetic acid oxidized material
soluble in 0-1 N NH3

In the choice of a solvent for the " softer " keratin in the second ratio,
there is, of course, an arbitrary element and other solvents could be used.
In textile circles, for example, the " alkaline solubility," is assessed, i.e. the
amount of fibre soluble in 0*1 N caustic soda at 65 °C in 1 hr, and this is
found useful as a measure of " damage", where this is defined as a deteriora-
tion in the insolubility or stability of the keratin. After allowing for
custom and the practical aspects, it would seem nevertheless, better to use
a solvent, such as thioglycolhc acid and urea, whose action is based on
differences in the degree of H-bonding and disulphide cross-linking which
are the chemical bases of insolubilisation. The experiments of Lees and
Elsworth (1955), Jones and Mecham (1943) and of Ward and Lundgren
(1954) on dissolving keratins show that they differ significantly when their
solubilities are compared by solvents such as proposed above.

Uneven keratinization and its histological distribution

The variations in keratinization, shown by such tests as " alkaline
solubility " or " urea-bisulphite " solubility or by histological stains,
which reveal directly the distribution of disulphide concentration (Plate 24B)
in a tissue, may arise in either of two ways. Firstly, since in a hard keratin
the consolidation process takes place after the synthesis and organization
of the protein into a fibrous structure and requires a certain time, it could
easily fail to achieve completion for accidental or systemic reasons. Secondly,

THE KERATINIZATION PROCESS 273

soft keratin containing transformed keratohyalin may be " mixed " in
various proportions with a fibrous keratin of the hard type. These
differences in degree of keratinization produce differences in physical and
chemical properties having obvious functional value and this has not been
overlooked in the evolution of the various epidermal appendages. It takes
perhaps its most interesting form when the function of a particular organ
is found to depend on differences in keratinization in its several histological
parts. The case of the hair follicle has been discussed above. There we find
the soft cells of the root sheaths desquamating and freeing the hair ; the
tough impact-resisting cuticle protecting the fibrous cortex; the light open
framework of the medulla providing a rigid girder-like internal skeleton.

The epidermis itself varies in thickness and toughness from site to site
and the distribution is clearly related to the special demands made on each
site. The thickened epidermis with its marked pattern of papillary ridges
found on the palmar surface of the hands and feet is genetically determined
as is also the ability to respond to mechanical friction and pressure
by further thickening. On the histological level, hard and softer regions in
the papillary ridges are said to have the effect of enhancing mechanically
the sensitivity to touch (Cauna, 1954).

In the terminal appendages (claws, hoofs, etc.) two parts, the unguis
and the softer subunguis, can normally be distinguished (Fig. 29, p. 68)
and the different rates of wear of these two parts has much to do with the
maintenance of the shape and the functional efficiency of the parts. For
example, in the ungulate hoof the hard cylinder of the unguis largely
surrounds the softer subunguis which wears the more rapidly and thus
maintains a more or less flattened, load-bearing surface. The sharp cutting
or piercing tip of a claw is maintained by a difference between the hardness
and orientation of the two layers of which the claw is composed. These
two layers, the superficial stratum and the deep stratum (unguis and
subunguis) may be seen in Figs. 29 d, e and f, and it is obvious how the
sharp cutting edge of the claw is maintained automatically by the more
rapid wearing of the deeper layer.

Hair also contains parts varying in keratinization in the cortex. An
elaborate relationship is found between morphology, physical behaviour
and variations in keratinization in fine crimpy or curly hair or wool
(Mercer, 1954). This type of hair, already discussed in Chapter IV, p. 156,
forms waves which are very nearly uniplanar and it is found that, in terms
of stabilization and of chemical reactivity, these fibres, like many of the
larger appendages, are also bilateral (Horio and Kondo, 1953). They con-
sist in fact of two hemi-cylinders differing in stability which are twisted
together in such a way that the helix of the two hemi-cylinders is always
in phase with the crimp wave and that the outside face of the fibre is
always less keratinized than the inside face (Horio and Kondo, 1953;

274 KERATIN AND KERATINIZATION

Mercer, 1953 and 1954; Fraser et ai, 1954) (Fig. 114). A bilateral fibre
has thus a built-in tendency to maintain its curled shape and the functional
value of this in maintaining the insulating properties of the fleece of the
animal, which depend on the air trapped by the mass of crimpy fibres, is
obvious.

OUTER ROOT-SHEATH

HENLES LAYER,

KERATINIZED
HUXLEYS LAYER.
I CONTAINING TRICHOHYALIN
j 'CUTICLE. PRE-KERATINIZED

FIBRE CUTICLE

INNER
>ROOT-
SHEATH

CORTEX

HUXLEY'S LAYER
PREKERATINIZED

ROOT-SHEATH CUTICLE
PREKERATINIZED.

----- OUTER ROOT- SHEATH

dS0£&m&fc V _ HENL E'S LAYER. }

Pg^'AlVBMl.4U I HUXLEY'S LAYER. (ROOT SHEATH

L ! S^Hf J HMHi "~ " I " KERA TIN I ZED.

mttr as.- — ;- CUTICLE, KERATINIZED.)

"fibre cuticle

CORTEX

Fig. 114. The development of keratinization in an asymmetrical wool
follicle leading to the formation of ortho and para segments (FG and HI)
(see Fig. 112). From Auber (1950) by permission. The sections pass
through the early (D) and late (E) levels of the keratinizing zone. Notice
the asymmetrical root sheaths.

Bilateral fibres of this type are widely distributed in the woolly coat
(secondary hair) of many animals, and are particularly marked in the sheep.
They are less well-defined in coarse hair and in the case of some forms of
crimpy negro hair are difficult to demonstrate (Spearman and Barnicot,
1960). The relation between external form and internal structure revealed
in this phenomena poses some interesting problems in keratinization which

THE KERATINIZATION PROCESS 275

are only partly understood and are more properly viewed in the wider
context of the general factors controlling the differentiation of the many
cell subtypes produced from the epidermal germinal layer.

The problem of crimp formation has already been discussed in Chapter
IV and it is obvious that the factors governing keratinization must be related
to the other morphogenetic factors involved. Auber (1950) and Rudall
(1936) discovered that keratinization is bilateral in follicles producing
crimpy fibres, i.e. the hardening of the fibre begins at a lower level on the
side which will emerge as the more keratinized (Fig. 114). Auber showed
that in such follicles the hair shaft was asymmetrically placed within the
cylinder of the outer root sheath and that hardening commenced on the
side of the hair nearest the thinner part of the sheath. This immediately
suggested that something supplied from beyond the sheath was required
for keratinization and could penetrate (or escape) most readily on the thin
side. There is ample other evidence implying a close dermo-epidermal
co-operation in controlling epidermal differentiation (p. 61). By partly
dissolving plucked follicles it has been shown that the difference between
the two sides extends below the keratinization zone, i.e. that the cell
types are committed early in their course. This again illustrates the
interlocking of the events at various depths in the follicle (p. 156).

The chemical basis of the difference between these variants of keratin is
not yet clear even in a case as well studied as wool. There is evidence from
histochemistry (Dusenbury and Menkart, 1956 (Plate 24B) : Menkart and
Coe, 1958) and from the analysis of resistant residues to suggest (but not
prove) that in wool the resistant fraction (para-type) has a higher cystine
content and perhaps differs in other respects. Simmonds (1958) could,
however, find no difference between high- and low-crimped material.
Rogers (1959b) found that the packing of the a-filaments in the ^>ara-cells
was hexagonal (Plate 16) and that whorls occurred more often on the
ortho-side (less resistant). Nevertheless whorls are no less common in
human hair which on a basis of its resistance and sulphur content is para-
type (Birbeck and Mercer, 1957).

Keratinized cysts and epidermal tumours

Tumours and cysts arising from epidermal cells are not uncommon and
may exhibit interesting, if abnormal aspects of keratinization that deserve
notice here. The skin of the mouse is also one of the commonest test
objects of experimental carcinogenesis, and figured largely in the classical
work of Kennaway and his colleagues which led to the isolation of definite
carcinogenic compounds from tar (Kennaway, 1955; Ludford, 1925).

It seems now established that the tumours produced by benzpyrene and
other carcinogenic hydrocarbons take their origin from the rather undiffer-
entiated basal cells of the upper outer root sheath of the hair follicle

276

KERATIN AND KERATINIZATION

(Wolbach, 1951; Borum, 1954) and that the effectiveness of the carci-
nogen depends on the phase of the hair cycle. When applied to a skin
area containing resting follicles the effect is small ; however, when the hairs
are growing the carcinogen seems able to penetrate the skin via the hair
follicles and produces a more profound effect. If the hair papilla cells are
killed by the applied chemical, normal follicle reformation becomes
impossible and the cyclic regenerative changes associated with growth
waves (p. 150) leads to proliferation of the outer root sheath cells with the
production of deeply-seated keratinizing cysts or tubes at the centres of
which the keratinizing cells, unable to exfoliate in the normal way, build up
to form keratin pearls (see Plate 24A) (Gliicksmann, 1945). The continued
growth forms at first a benign papilloma. According to Wolbach such
cells are still responsive to the stimulus of the growth waves passing over
the neighbouring normal skin. Genuinely malignant tumours may
ultimately develop from such papillomas.

In the so-called " hairless " mouse mutant after the first wave of hair
growth, follicles fail to reform normally and very similar keratinizing cysts
may form beneath the skin from the upper portion of the root sheath
(Gnineberg, 1952).

Pigmentation

Most epidermal derivatives are pigmented, and the great variety of
integumental colours and pattern which can be produced is of immense
importance in the life of animals. The various colours (" structural
colours", which result from the diffraction of light by regular structures,

Table 21. Epidermal Pigments.

Property

Melanin

Pheomelanin

colour

brown black

yellow (red)

shape

oval to round

round and smaller

dimensions

0-1-3/x

chemical type

protein tanned with
melanin polymer

alkali solubility

almost insoluble

soluble

precursor

tyrosine

tyrosine and tryptophane

excepted) are derived solely from combinations of black, brown or yellow
pigment in the form of granules combined possibly with a red non-granular
pigment. Table 21, adapted from Fitzpatrick, Brunet and Kukita (1958)
summarizes the main facts relating to granular pigments.

Pigment granules are the exclusive product of pigment-forming cells
called melanocytes which in the adult animal are found among the basal

THE KERATINIZATION PROCESS 277

layer cells of the keratinizing system. According to Medawar (1953) the
melanocytes comprise from 5 to 15% of the total cell population of the
germinal layers of the epidermis. The epidermis thus really consists of two
entirely different classes of cells the members of which have distinct
morphologies, functions and embryonic origins. Keratinizing cells arise
embryonically from the ectoderm; melanocytes have, however, been
traced back largely to the neural crest (Rawles, 1947) (see also Niu, 1959)
and they enter the epidermis only after this has been clearly differentiated.
They are an amoeboid type of cell with several long arborescent processes
called dentrites, and are perhaps best referred to as dentritic cells. In their
dispersion from their site of origin they seem impelled by a mutual
repulsion which leads them ultimately to colonize the dermis and epidermis
and there to adopt a dispersed distribution, each cell occupying a small
domain determined by the extreme reach of its dentritic processes. They
also accumulate densely in a few other sites, such as the pigmented layer
of the eye. Each epidermal melanocyte pigments the small group of
keratinizing cells within reach of its dentrites. The granules of pigment
are formed in the perikaryon of the cell, pass along the processes and enter
the keratin cells. Owing to their situation in the basal layer attached to the
basal membrane, pigmentation occurs before the formation of keratin and
the subsequently-formed fibrils may lead to an orientation of the granules.
The presence of melanocytes seems in no way essential to the well-being
of the keratinizing system since some epithelia naturally lack melanocytes
and others may be deprived of them, accidently or by experiment, without
appearing to be at a disadvantage.

The distribution of pigment cells is under genetic control and, since
changes in the integument are easy to observe, much attention has been
given to the genetics of skin and hair pigmentation. The value in terms of
natural selection of pigmentary patterns is obvious; but for all that, little is
known of the underlying causes determining the distribution of pigment
cells (Du Shane, 1944).

According to Billingham and Medawar (Billingham, 1948; Billingham
and Medawar, 1948 and 1953; Billingham, 1958) not all dentritic cells
produce pigment. White skin patches are said to contain a full complement
of dentritic cells although special means are required to demonstrate these,
since they contain no pigment. The skin of white human beings is said to
contain as many melanocytes as that of negroes, for example. It would seem
that melanocytes differ in their response to the influences which provoke
pigmentary activity. This would be an inherited difference distinguishing
different sub-races of melanocytes even on a single skin. Some never fail
to begin production once they reach the epidermis; others may be pro-
voked into activity by exposure to actinic radiation; others normally
remain latent. The hormonal balance can also cause changes in activity.

278 KERATIN AND KERATINIZATION

There would appear to be ample opportunity for these variations to arise
when the long train of enzymatically-calalysed reactions involved in pig-
ment formation is considered (p. 279).

In their grafting experiments on guinea-pig skin, Billingham and
Medawar (1948) have demonstrated an actual diffusion of pigmentary
activity from a black graft into a white skin. They proposed an explanation
which is of some theoretical importance. They consider that the latent
melanocytes of the white skin are actually " infected " by the neighbouring
black melanocytes by a process which is related to the normal method of
transferring pigment from melanocyte to a keratinizing cell. They have
demonstrated microscopically that anastomoses between processes of
neighbouring melanocytes do actually occur and suppose that through such
contacts one melanocyte may transfer to another a sample of its cyto-
plasmic apparatus. Electron micrographs of the tips of dentritic processes
found in keratinizing cells show that in fact samples of cytoplasmic
membranes and particles are transferred along with pigment granules
(Birbeck et al, 1956).

Such a transference of a cytoplasmic element capable of permanently
modifying the activity of an acceptor cell, if it proves not merely a peculia-
rity of the system of dentritic cells, could be of importance in the normal
developmental history of cell lines. It would imply first that inherited
differences between cells could have a cytoplasmic as well as a nuclear
basis, and second that a population of cells could establish and maintain
uniformity among their members by cytoplasmic exchanges. The actual
mechanics of the transfer of pigment from melanocyte to epidermal cell
needs further elucidation. Two broad possibilities may be envisaged: the
keratin cell may be actively penetrated by the tip of the melanocyte process
or alternatively, it may play the active role and phagocytize the tip. The
latter process receives some support from electron micrographs which
show small processes of the keratinizing cell in various stages of enveloping
the pigmented tips.

The Pigment Granule

The actual pigment granule contains both protein and pigment and is
often referred to as a melanoprotein, although almost nothing is known of
the nature of the linkage between the two components. It may be formu-
lated as a bipolymer in which polypeptides alternate with melanin, or
it may be a tanned protein, i.e. a network of protein chains cross-linked by
melanin polymer or other phenols. Whatever its structure, the granules
may exhibit remarkable chemical stability as is proved by the fact that their
separation from keratinized tissues by violent chemical destruction of the
keratin seems to leave them morphologically intact. Electron micrographs
by Birbeck demonstrate clearly the existence of a regular framework of

THE KERATINIZATION PROCESS 279

protein in the pre-melanin granules on which the melanin later poly-
merizes (Fig. 115 (h) and (i)) (Birbeck and Barnicot, 1949).

The similarity between the types of chemical resistance exhibited by
melanin granules and the cell membranes of the keratinized tissue has been
mentioned above. It is certainly possible that a similar tanning reaction
has cross-linked the proteins of the membranes and the granules — a
possibility of some phylogenic interest. Recently, tanned membranes
having similar solubility properties have been described by Jones (1958).

(a) (b) <QgH^ (c) (d)

• •

• ••

•

• £<®

• •

(f)

©fa

(e) v" (9) M

Fig. 115. Varieties of melanin granules. Redrawn to scale from authors

cited. The examples are chosen to show the extreme range of size and

type. X 10,000.

(a) Human hair white male (Birbeck, Mercer and Barnicot, 1956).

(b) Human hair negro male (Birbeck, Cuckow and Barnicot, 1955).

(d) Black alpaca (Laxer and Whewell).

(e) Retinal eye pigment (Birbeck, private communication).

(f) Harding-Passey melanoma from section of tumour by the writer.

(g) Squid ink (Birbeck, private communication).

(h) and (i) Human-head hair with indications of internal structure as
seen in sections of melanocytes.

The granules vary in size from 0- 1 to 3/x and for this reason their size and
shape has been much studied electron-microscopically. This may be done
either on isolated granules or in sections of fixed tissue. Fig. 115 shows
outline drawings of a number of types of granule and will give some idea
of the range of size encountered. Sections of granules often suggest a clear
separation between layers of melanin deposits and those of a lighter
material, presumably protein.

The Chemistry of Melanization

Although the protein moiety and its relation to the melanin polymer
has been little studied, much attention has been paid to the biochemical
steps in the formation of the melanin itself. Melanins, quite apart from

280 KERATIN AND KERATINIZATION

their occurrence as pigment granules, are very widely distributed in nature,
occurring in plants, in vertebrates and invertebrates. They are formed by
the action of copper-containing oxidases, known as phenolases, that catalyse
the oxidation of mono- and di-hydric phenols to o-quinones (Mason, 1953
and 1955). Phylogenetically the chemical reactions involved in melani-
zation developed before keratinization and quite independently of it.
Mason has discussed the wide distribution and varied applications of the
" phenolase system " in the different phyla. It constitutes an excellent
example of a simple biochemical system, which catalyses essentially the
same reactions in plants and highly-organized animals, although with an
increased specificity towards substrates as the phylogenetic tree is ascended,
and which thus finds expression in widely- different characters at different
levels. For example, it is responsible for the browning of plant tissues, the
hardening of cuticle of arthropods and the pigmentation of chordates. In
the higher animals the site of melanin formation has become limited to the
pigment cell or melanocyte.

The biochemical evidence has been reviewed recently by Mason (1955)
and Fitzpatrick, Brunet and Kukita (1958). Briefly, the amino acid
tyrosine has been shown to be the precursor of the insoluble pigment. The
phenolase, tyrosinase, converts tyrosine to DOPA (3 : 4-dihydroxy-
phenylalanine) and to " DOPA quinone", which becomes 5 : 6-dihydroxy-
indole, the immediate precursor or monomer of the large polymer molecule
melanin, which may then be linked to a protein (Fig. 1 16).

The darkening of a tissue when treated with DOPA (dihydroxypheny-
lalanine), Bloch's DOPA-oxidase reaction, has long been used to demon-
strate cytologically the sites of melanin formation (Block, 1921), but it is
now less favoured, since it may be non-specifically oxidized to melanin.
Tyrosine is preferred as substrate although it may fail to demonstrate
tyrosinase when the latter is in low concentration. Histochemically
tyrosine activity has been demonstrated in autoradiographs by using
radioactive tyrosine and C14 (see Fitzpatrick et al., 1958).

The small quantities of melanized material from mammalian sources
have hindered biochemical research, but by taking advantage of the larger
quantities of material available in pigmented tumours (melanomas) this
may be overcome. The granules of the Harding-Passey melanoma appear
as aggregates of fine dense granules perhaps incompletely supplied with a
protein framework (Mercer, unpublished) (Fig. 115(f)).

Something of the cytological structure associated with these syntheses
has been revealed by the electron microscope and by the histochemical
location of tyrosinase activity. The cell is roughly polarized in a manner
similar to glandular cells (see p. 110) with a limited basophilic reticulum
at the end proximal to the attachment to the basement membrane (Birbeck
et al., 1956). Here presumably the protein of the granule is synthesized.

THE KERATINIZATION PROCESS 281

Distal to the nucleus is a clear region relatively free of dense melanized
particles with large numbers of small, rounded vacuoles (Golgi-type
vacuoles) which contain variable amounts of denser material often arranged
in concentric shells. The vacuoles on the periphery of this region contain
additions of very dense material identifiable as melanin.

On both histological and electron-microscopical grounds there is now
little doubt that the intracellular sites for the formation of melanin are these

Fig. 116. Possible courses in the synthesis of melanins and melanoproteins. *

o-diphenols > o-quinones > simple polymers

(melanins)
(poorly-organized)

+
proteins (organized framework)

bi-polymersf
melanoprotein

* From Mason (1955) with modifications.

t Morphological evidence would indicate that this is a bipolymer
whose components are each of a macromolecular size.

small vesicles. An older opinion, based on similarity of size and staining
properties of granules and mitochondria, held that the melanogenic vesicle
is really a mitochondrion. Since the granules contain, in addition to
tyrosinase at least two other enzymes, cytochrome oxidase and succinic
dehydrogenase usually located in mitochondria, some support is given for
this view. Morphologically, however, the two organelles are quite distinct
although in their remote origins both may possibly be traced back to
similar vesicular formations. Recently mitochondria and premelanin
granules, characterized by containing tyrosinase and not mitochondrial
enzymes, have been separately isolated from melanoma homogenates
(Baker et al., 1960).

282 KERATIN AND KERATINIZATION

The appearance of granules in early stages of formation suggests that a
rather regular protein framework (Fig. 115 (h)) is laid down first and is
subsequently melanized by a polymerization occurring in its interstices
(Birbeck and Barnicot, 1959). The melanoprotein thus seems to be a well-
ordered particle although little of this is apparent after melanization is
complete. Melanocytes in albino hair possess large numbers of small
vacuoles containing the protein framework, but no deposition of dense
material occurs. In these cases, in the absence of tyrosinase, the melanin
polymer is not formed.

Some interest attaches to the control of melanocyte activity in the hair
root which is correlated with the hair growth cycle (Montagna, 1956) and
with the development of various pigmentary patterns both in hair and in
feathers. Montagna's observation that melanocyte activity ceases shortly
before hair growth ceases is cited as evidence of the accumulation of a
general growth inhibitor (see Chapter IV) affecting first the more sensitive
melanocytes.

The yellow-red pigment (pheomelanin) also occurs in granular form,
but less is known about the metabolic pathways leading to its formation.
The ultimate precursor appears to be the amino acid tryptophane.
Reference may be made to the review by Fitzpatrick et al. (1958).

References

Abercrombie, M. (1957) Symp. Soc. exp. Biol. 11, 235.

and Heaysman, J. E. M. (1953) Exp. Cell. Res. 5, 111 ; (1954) 6, 293.

A Symposium on the Chemical Basis of Development, p. 318. Ed. by McElroy,

W. D. and Glass, B. Johns Hopkins Press, 1958.
Alexander, P. (1951) Ann. N.Y. Acad. Sci. 53, 653.

and Earland, C. (1950) Nature, Lond. 166, 396.

and Hudson, R. F. Wool: its Chemistry and Physics. Chapman & Hall,

London, 1954.
— - and Smith, L. F. (1955) Proc. Inter. Wool Text Conf. Austr. Vol. B., p. 56;

C.S.I.R.O., Melbourne, 1956.
Allfrey, V. G., Daly, M. M. and Mirsky, A. E. (1953) J. gen. Physiol. 37, 157.

Mirsky, A. E. and Osawa, S. (1957) J. gen. Physiol. 40, 451.

Ambrose, E. J. and Elliott, A. (1951a) Proc. Roy. Soc. A205, 47.

and Hanby, W. E. (1949) Nature, Lond. 163, 483.

and Elliott, A. (1952) Proc. Roy. Soc. A206, 206.

Arndt, U. W. and Riley, D. P. (1955) Phil. Trans. A247, 409.
Arthur, R. P. and Shelley, W. B. (1959) J. invest. Dermat. 32, 397.
Assheton, R. (1896) Quart. J. micros. Sci. 38, 465.

Astbury, W. T. and Street, A. (1931) Phil. Trans. A230, 75.

and Marwick, T. C. (1932) Nature, Lond. 130, 309.

and Woods, H. J. (1933a) Phil. Trans. A232, 333.

Fundamentals of Fibre Structure. Oxford U.P., London, 1933b.

(1947) Proc. Roy. Soc. B134, 303 (Croonian Lecture).

(1943) Advanc. Enzymol. 3, 63.

. Dickinson, S. and Bailey, K. (1935) BiochemJ. 29, 2351.

and Bell, F. O. (1939) Tabulae Biologica XVII. part 1, 90.

and Dickinson, S. (1940) Proc. Roy. Soc. B129, 307.

and Bell, F. O. (1941) Nature, Lond. \41, 696.

Dalgleish, C. E., Darmon, S. E. and Sutherland, G. B. B. M. (1948)

Nature, Lond. 162, 596.

(1949) Exp. Cell Res. Suppl. 1. 234.

Beighton, E. and Weibull, C. (1955) Soc. exp. Biol. Symp. 9, 282.

(1955) Proc. Int. Wool Text. Conf. Austr. Vol. B., p.204; C.S.I.R.O., Mel-

bourne, 1956.

(1958) Disc. Faraday Soc. 25, 80.

Beighton, E. and Parker, K. D. (1959) Biophys. Biochim Acta 35, 17.

Auber, L. (1950) Trans. Roy. Soc, Edinb. 62, 191.

(1955) Nature, Lond. 176, 1218.

Bachra, B. N., Sobel, A. E. and Stanford, J. W. (1959) Arch. Bioch. Biophys.

84, 79.
Bahr, G. F. (1954) Exp. Cell Res. 7, 457.

and Moberger, G. (1954) Exp. Cell Res. 6, 506.

283

284 KERATIN AND KERATINIZATION

Bailey, K., Astbury, W. T. and Rudall, K. M. (1943) Nature, Lond. 151, 716.
Baker, J. R. Cytological Technique. 2nd Ed. Methuen, London, 1945.

- (1955)7. Roy. micros. Soc. 74, 217.

(1959) J. Roy. micros. Soc. 77, 116.

Baker, R. V., Birbeck, M. S. C, Blaschko, H., Fitzpatrick, T. B., and Seiji, M.

(1960) Nature, Lond. 187, 392.
Baldwin, E. An Introduction to Comparative Biochemistry. 3rd Ed. Rep. Cambridge

University Press, 1949.

Perspectives in Biochemistry, p. 99. Ed. by J. Needham and D. E. Green,

Cambridge University Press, London, 1937.
Bamford, C. H. et al. (1953) Proc. Roy. Soc. B141, 49.
- — - Elliott, A. and Hanby, W. E. Synthetic Polypeptides. Academic Press,

New York, 1956.

Brown, L., Elliott, A., Hardy, W. E. and Trotter, I. F. (1954) Nature,

Lond. 173, 27.

Hanby, W. E. and Happey, F. (1949) Nature, Lond. 164, 138.

and Hanby, W. E. (1951) Nature, Lond. 168, 340.

Barker, S. G. Wool Quality, H.M. Stationery Office, 1933.

Barnes, R. J. Studies in the Optical Properties of Wool Hair and related Fibres, Ph.D.

Thesis, Leeds, 1933.
Barnicot, N. A., Birbeck, M. S. C. and Cuckow, F. W. (1955) Ann. Human

Genet. 19, 231.
Barrnett, R. J. (1953)7. Nat. Cancer Inst. 13, 905.

and Seligman, A. M. (1952) Science 116, 323.

Battistone, G. C. and Burnett, G. W. (1956) J. Dent. Res. 35, 260.
Bear, R. S. (1943) J. Amer. Chem. Soc. 65, 1784; (1944) Ibid. 66, 2043.

and Rugo, H. J. (1951) Trans. N.Y. Acad. Sci. 53, 627.

(1952) Advanc. protein Chem. 7, 69.

Bellamy, L. T. The Infra-red Spectra of Complex Molecules. Methuen, London,

1954.
Bendit, E. G. (1957) Nature, Lond. 179, 535.
Bennett, H. S. (1948) Anat. Rec. 100, 640.
von Bergen, W. Matthew's Textile Fibers. 6th Ed. Ed. by H. R. Mauersberger,

p. 484. John Wiley, New York, 1954.
Bergmann, M. and Fraenkel-Conrat, H. (1937) J. Biol. Chem. 119, 707.
Bern, H. A. (1954) Nature, Lond. 174, 509.

Harkness, D. R. and Blair, S. M. (1955) Proc. Nat. Acad. Sci., Wash. 41, 55.

Alfert, M. and Blair, S. M. (1957) J. Histochem. Cytochem. 5, 105.

Bernal, J. D. in The Cell and Protoplasm, p. 199. Ed. by F. R. Moulton, Wash.

Ass. Adv. Sci., 1940.
Bernhard, W., Gautier, A. and Rouiller, C. (1954) Arch d'Anat. micros, et de
Morph. exper. 43, 236.

Haguenau, F., Gautier, A. and Oberling, Ch. (1952) Z. f. Zellforsch. u.

mikros. Anat. 37, 281.

Gautier, A. and Oberling, Ch. (1951) C. R. Soc. Biol, Paris 145, 566.

Bertalanffy, L. von, in Fundamental Aspects of Normal and Malignant Growth,

p. 137. Ed. by W. N. Nowinski, Elsevier, Amsterdam, 1960.
Biedermann, W. (1926) Ergebn. Biol. 1, 342; (1928) Ibid. 3, 354; (1928) Ibid. 4,

361.
Billingham, R. E. (1948) J. Anat. 82, 93; 83, 109.
— - and Medawar, P. B. (1948) Heredity 2, 29.

REFERENCES 285

Billingham, R. E. and Medawar, P. B. (1953) Phil. Trans. B237, 151.

The Biology of Hair Growth. Ed. by W. Montagna and R. A. Ellis. Ch. 9,

p. 451, Academic Press, New York, 1958.
Birbeck, M. S. C. (1957) Private communication.

Mercer, E. H. and Barnicot, N. A. (1956a) Exp. Cell Res. 10, 505.

and Mercer, E. H. (1956b) Nature, Lond. 178, 985.

- and Barnicot, N. A. in Biology of Hair Growth, p. 239. Ed. by W. Montagna

and R. A. Ellis, Academic Press, New York, 1958.

and Barnicot, N. A. Pigment Cell Biology, p. 549. Ed. by M. Gordon,

Academic Press, New York, 1959.

and Mercer, E. H. (1957a) J. biophys. biochem. Cytol. 3, 203, 215, 227.

- and Mercer, E. H. (1957b) Proc. Int. Conf. Elect. Micros., Stockholm, 1956,

p. 156.

and Mercer, E. H. (1961) Nature, Lond. 189, 558.

Blackburn, S. (1950) Biochem. J. 47, 443.

and Lee, G. R. (1954) Biochem. J. 58, 227.

Block, B. (1921) Archiv. Dermatol, u. Syphilis 135, 77.
Block, R. J. (1931) J. J. Biol. Chem. 94, 647.

and Vickery, H. B. (1931) J. Biol. Chem. 93, 113.

- Horwitt, M. K. and Bolling, D. (1949) J. dental Res. 28, 518.

and Bolling, D. Amino Acid Composition of Proteins and Foods. Thomas,

Springfield, 111., 1950.
Boedtker, H. and Doty, P. (1956) J. Amer. Chem. Soc. 78, 4267.
Bolliger, A. and Gross, R. (1952) Aust. J. exp. Biol. med. Sci. 30, 181, 399, 395.

(1956) Aust. J. exp. Biol. med. Sci. 34, 219.

Borsook, H. Proc. 3rd Int. Congr. Biochem., Brussels, p. 92. Academic Press, New
York, 1955.

(1955) Advanc. Protein Chem. 8, 127.

Borum, K. (1954) Acta Pathol. Microbiol. Scand. 34, 521.
Brachet, J. Chemical Embryology. Interscience, New York, 1950.

Biochemical Cytology, Academic Press, New York, 1957.

Bradfield, J. R. G. (1951) Nature, Lond. 167, 40.

(1955) Soc. exp. Biol. Symp. 9, 306.

Braun-Falco, O. in Biology of Hair Growth. Ch. 4, p. 65. Academic Press, New

York, 1958.
Brody, I. (1959) .?. Ultrastruct. Res. 2, 482.
(1959)7. Ultrastruct. Res. 3, 84.

(1960) J. Ultrastruct. Res. 4, 264.

Broussy, J. (1932) Thesis No. 52, Universite de Montpellier, Facultede Medicine.
Brown, C. H. (1949) Quoted in L. E. R. Picken, Phil. Trans. B234, 1.

Ph.D. Thesis, Cambridge, 1949.

(1949) Exper. Cell Res. Suppl. 1, 351.

(1950) Nature, Lond. 165, 275.

(1950) Quart, y. micros. Sci. 91, 331.

Brown, L. and Trotter, I. F. (1956) Trans. Faraday Soc. 52, 537.
Bull, H. B. and Gutman, M. (1944) y. Amer. Chem. Soc. 66, 1253.

- (1945) y. Amer. Chem. Soc. 67, 533.
Bullough, W. S. (1946) Phil. Trans. B231, 453.

Vertebrate Sexual Cycles, Methuen, London, 1951.

and Johnson, M. (1951) Proc. Roy. Soc. B138, 562.

(1952) Biol. Rev. 27, 133.

286 KERATIN AND KERATINIZATION

Bullough, W. S. (1954) Exp. Cell Res. 7, 176; (1954) Ibid. 7, 186.

(1955) Exp. Cell Res. 9, 108.

and Laurence, E. B. (1957) Brit. J. exp. Path. 38, 278.

and Laurence, E. B. In Biology of Hair Growth, ch. 8. Ed. by W. Montagna

and R. A. Ellis, Academic Press, New York, 1958.

and Laurence, E. B. (1960) Proc. Roy. Soc. B151, 517.

- — and Laurence, E. B. (1961) Private communication.

Bunn, C. W. Chemical Crystallography, Clarendon Press, Oxford, 1946.

Burge, R. E. (1960) Proc. Roy. Soc. A, 260, 558.

Burgos, M. H. and Wislocki, G. B. (1958) Endrocrinology 63, 106.

Burns, M. and Clarkson, H. (1949) J. Agric. Sci. 39, 315.

Burrows, H. Biological Action of Sex Hormones, Cambridge University Press,

1949.
Burte, H. and Halsey, G. (1947) Text Res. J. 17, 456.
Butcher, E. O. (1934) Anat. Rec. 61, 5.

Cairns, J. M. and Saunders, J. W. (1954) J. exp. Zool. 127, 221.
Calvery, H. O. (1933) J. Biol. Chem. 100, 183.
Calvin, M. (1956) Amer. Scientist 44, 248.
Carruthers, C, Quevedo, W. C. and Woernley, D. L. (1959) Proc. Soc. exp.

Biol.,N.Y. 101, 374.

and Santzeff, V. (1953) Physiol. Rev. 33, 229.

Carter, H. B. (1943) Bull. C.S.I.R.O., Australia, No. 164, Melbourne.

and Clarke, W. H. (1957) Aust. J. agric. Res. 8, 91, 109.

Caspersson, T. O. Cell Growth and Cell Function. W. W. Norton, New York, 1950.

(1947) Syinp. Soc. exp. Biol. 1, 127.

Cauna, N. (1954) Anat. Rec. 119, 449.

Champetier, G. and Faure-Fremiet, E. (1938) C. R. Acad. Sci., Paris 207, 1133.
Charles, A. and Smiddy, M. B. (1957) J. Invest. Dermat. 29, 327.

(1959) Exper. Cell Res. 18, 138.

Chase, H. B. (1954) Physiol. Rev. 34, 113.

(1955) J. Soc. Cos. Chem. 6, 9.

(1959) Ann. N. Y. Acad. Sci. 83, 365.

Chevremont, M. and Frederic, J. (1943) Arch. Biol. Liege 54, 589.
Clark, W. E. Le Gros (1936) Proc. Zool. Soc. 106.
Claude, A. (1938) Proc. Soc. exp. biol. Med. 39, 398.

(1946) J. exp. Med. 84, 51.

Cockran, W., Crick, F. H. C. and Vand, V. (1952) Acta Cryst., Camb. 5, 581.
Colvin, J. R., Smith, D. B. and Cook, W. H. (1954) Chem. Rev. 54, 687.
Coman, C. R. (1954) Cancer Res. 14, 519.
Corey, R. B. and Wyckoff, R. W. G. (1936) J. Biol. Chem. 114, 407.

and Pauling, L. (1955) Proc. Int. Wool Text. Res. Conf. Austr. Vol. B.,

p. 249; C.S.I.R.O., Melbourne, 1956.
Corfield, M. C. and Robson, A. (1955) Biochem. J. 59, 62.

Robson, A. and Skinner, B. (1958) Biochem. J. 68, 348.

Cowdry, E. V. Special Cytology. Hoeber, New York, 1932.
Crane, H. R. (1950) Scientific Monthly 70, 376.
Crick, F. H. C. (1952) Nature, Lond. 170, 882.

(1953) Acta Cryst., Camb. 6, 685, 689.

(1958) Symp. Soc. exp. Biol. 12, 138.

and Watson, J. D. (1954) Proc. Roy. Soc. A223, 80.

Cruise, A. J. and Jeffery, J. W. (1959) Nature, Lond. 183, 677.

REFERENCES 287

Cruise, A. J. and Keech, M. K. (1959) J. Roy. micros. Soc. 78, 95.

Dalton, A. J. and Felix, M. D. (1956) J. biophys. biochem. Cytol. 2 Suppl., 79.

■ — and Felix, M. D. (1957) Symp. Soc. exp. Biol. 10, 148.

Danforth, C. H. Special Cytology. Sect. 2, p. 41. Ed. by E. V. Cowdry, Hoeber
New York, 1932.

Danielli, J. F. in Cytology and Cell physiology, p. 1. Ed. by G. H. Bourne,
Oxford University Press, 1942.

Darmon, S. E. and Sutherland, G. B. B. M. (1949) Nature, Lond. 164, 440.

Davies, H. R. (1889) Morph. Jb. 15, 560.

Davson, H. and Danielli, J. F. The Permeability of Natural Membranes. Cam-
bridge University Press, London and New York, 1952.

Das, D. B. and Speakman, J. B. (1950) J. Soc. Dy. Col. 66, 583.

Day, M. F. (1951) Aust. J. Sci. Res. B4, 42.

Derksen, J. C, Heringa, G. C. and Weidinger, A. (1937) Acta ne'erl. morph.
norm, et path. 1, 31.

Donohue, J. (1953) Proc. Nat. Acad. Sci., Wash. 39, 470.

(1954) Proc. Nat. Acad. Sci., Wash. 40, 377.

Doty, P. and Yang, T. J. (1956) J. Amer. Chem. Soc. 78, 498.

(1956) Proc. Nat. Acad. Sci., Wash. 42, 791.

Dry, F. W. (1926)7. Genet. 16, 287; (1928) J. Genet. 20, 131.

Durward, A. and Rudall, K. M. (1949)7. Anat. 83, 327.

in The Biology of Hair Growth. Ch. 9, p. 189. Academic Press, New York

1958.
Dusenbury, J. H. and Menkart, J. (1955) Proc. Int. Wool Text. Cont. Austr.

Vol. F, p. 142. C.S.I.R.O., Melbourne, 1956.
Du Shane, G. P. (1944) Quart. Rev. Biol. 19, 98.

Earland, C. and Wiseman, A. (1959) Biochim. Biophys. Acta 36, 273.
Ebling, F. J. (1953)7. Endocrinol. 9, 31, 32.

(1954)7. Endocrinol. 10, 147.

and Johnson, E. (1959)7. Embryol. Exper. Med. 7, 417.

Eisen, A. Z., Montagna, W. and Chase, H. B. (1953) J. Nat. Cancer Inst. 14, 341
Elliott, A. (1952) Text. Res. J. 22, 783.

(1953) Proc. Roy. Soc. A221, 104.

(1954) Proc. Roy. Soc. A226, 408.

Hanby, W. E. and Malcolm, B. R. (1954) Brit. J. App. Phys. 5, 377.

■ Hanby, W. E. and Malcolm, B. R. (1956) Nature, Lond. 178, 1170.

Hanby, W. E. and Malcolm, B. R. (1958) Disc. Faraday Soc. 25, 167.

Ellis, W. J. (1948) Nature, Lond. 162, 957.

Gillespie, J. M. and Lindley, H. (1950) Nature, Lond. 165, 545.

Elod, E. and Zahn, H. (1944) Kolloid-Z. 108, 94.

and Zahn, H. (1946) Melliand Textilber. 27, 68.

and Zahn, H. (1949) Kolloid-Z. 113, 157.

Engstrom, A. and Lindstrom, B. (1947) Experientia 3, 191.
'Espinasse, P. G. (1939) Proc. Zool. Soc. Lond. Ser. A. 109, 247.

Farrant, J. L., Rees, A. L. G. and Mercer, E. H. (1947) Nature, Lond. 159, 535.

and Mercer, E. H. (1952) Biochem. biophys. Acta 8, 355.

Fasske, E. and Themann, H. (1959) Z. Zellforsch. 49, 447.

Fawcett, D. W. Frontiers in Cytology, p. 19. Ed. by S. L. Palay, Yale University
Press, Newhaven, 1958.

and Porter, K. R. (1954) J. Morph. 94, 221.

288 KERATIN AND KERATINIZATION

Fell, H. B. and Canti, R. G. (1934) Proc. Roy. Soc. B116, 316.

(1957) Proc. Roy. Soc. B146, 242.

- and Danielli, J. F. (1943) Brit. jf. exp. Path. 24, 196.

and Mellanby, E. Sir (1953) J. Physiol. 119, 470.

- Mellanby, E. Sir and Pelc, S. R. (1954) Brit. med. J. 2, 611.

Mellanby, E. Sir and Pelc, S. R. (1956) J. Physiol. 134, 179.

Feughelman, M. (1959) Text. Res. J. 29, 223.

Fischer, A. (1924) J. exp. Med. 39, 585.

Fitzpatrick, T. B. and Kukita, A. J. (1956) J. Invest. Dermat. 26, 173.

Brunet, P. and Kukita, A. The Biology of Hair Growth, p. 255. Ed. by

W. Montagna and R. A. Ellis, Academic Press, New York, 1958.
Fleischauer, K. (1953) Z. f. Zellforsch. 38, 50, 328.
Flesch, P. (1949) Proc. Soc. exp. Biol. Med. 70, 79.

Florkin, M. Unity and Diversity in Biochemistry. Pergamon, Oxford, 1960.
Fox, S. W. (1953) Amer. Nat. 87, 253.

Fraenkel, G. and Rudall, K. M. (1947) Proc. Roy. Soc. B134, 111.
Fraser, A. S. (1951) Nature, Lond. 167, 202.

(1952) Austral. J. agric. Res. 3, 419.

and Short, B. F. (1952) Austal. jf. agric. Res. 3, 445.

(1955) Austal. jf. agric. Res. 6, 770.

and Nay, T. (1953) Aust. J. Biol. Sci. 6, 645-656; (1955) Ibid. 8, 428.

and Short, B. F. (1958) Aust. J. Biol. Sci. 11, 200.

and Short, B. F. The Biology of the Fleece. Animal Res. Lab. Tech. Paper

No. 3. C.S.I.R.O., Melbourne, 1960.
Fraser, R. D. B., Lindley, H. and Rogers, G. E. (1954) Biochim. biophys. Acta
13, 295.

and Rogers, G. E. (1955) Aust. J. Biol. Sci. 8, 289.

and MacRae, T. P. (1956) Text. Res. J. 26, 618.

and MacRae, T. P. (1957) Text. Res. Jf. 27, 867.

and MacRae, T. P. (1957) Nature, Lond. 179, 732.

and MacRae, T. P. (1958a) Biochim. biophys. Acta 29, 229.

and MacRae, T. P. (1958b) J. chem. Phys. 28, 1120.

and MacRae, T. P. (1959) J. mol. Biol. 1, 387, 1959.

- MacRae, T. P. and Rogers, G. E. (1959) Nature, Lond. 183, 592.

and MacRae, T. P. (1961) Nature, Lond. 189, 572.

Frey-Wyssling, A. Submicroscopic Morphology of Protoplasm. Elsevier, Amster-
dam, 1953.

Garnier, Ch. in Hagenau, F. The Ergastoplasm ; (1958) Int. Rev. Cytol.7, 425.
Geiger, W. B. (1944) J. res. Nat. Bur. Stand. 32, 127.

Patterson, W. I., Mizell, L. R. and Harris, M. (1941) J. res. Nat. Bur.

Stand. 27, 459.

Kobayashi, F. F. and Harris, M. (1942) J. res. Nat. Bur. Stand. 39, 381.

Gelfant, S. (1958) Exp. Cell Res. 15, 423, 451; (1959a) Ibid. 16, 527; (1959b)

Ibid. 18, 594.

(1960) Ann. N. Y. Acad. Sci. 90, 536.

Gillespie, J. M. (1958) Biochim. biophys. Acta 27, 225.

(1955) Proc. Int. Wool Text. Res. Cont. Austr. Vol. B., p. 35; C.S.I.R.O.,

Melbourne, 1956.

REFERENCES 289

Gillespie, J. M. (1960) Aust.J. biol. Sci. 13, 81.

and Lennox, F. G. (1953) Biochim. biophys. Acta 12, 481.

and Lennox, F. G. (1955) Austr. J. biol. Sci. 8, 97.

O'Donnell, I. J., Thompson, E. O. P. and Woods, E. F. (1960) Proc.

Wool Text. Res. Conf., Harrogate, 1960, J. Text. Inst. 51, T703.
Giroud, A. and Bulliard, H. (1930) Arch, de morph. 29, 1.

and Champetier, G. (1936) Bull. soc. chim. biol. 18, 656.

and Leblond, C. P. (1951) Ann. N. Y. Acad. Sci. 53, 613.

- Bulliard, H. and Leblond, C. P. (1934) Bull Histol. Tech. micr. 11, 129.
Glinos, A. D. The chemical Basis of development (Mech. of liver groivth), p. 813.

Ed. by W. D. McElroy and B. Glass, Baltimore, Johns Hopkins Press,
1958.
Glucksmann, A. (1945) Cancer Res. 5, 385.

and Cherry, C. P. (1956) Cancer 9, 971.

Goddard, D. R. and Michaelis, L. (1934) J. Biol. Chem. 106, 605.

and Michaelis, L. (1935) J. Biol. Chem. 112, 361.

Gordon, M. Pigment Cell Groivth. Academic Press, New York, 1953.

Grasse, P. P. (1956) C. R. Acad. Sci., Paris 242, 858.

Gray, E. G. (1959) Proc. Roy. Soc. B150, 233.

Grimstone, A. V. (1961) Biol. Rev. 36, 97.

Gross, J. (1956) J. biophys. biochem. Cytol. Suppl. 2, 261.

Highberger, J. H. and Schmitt, F. O. (1954) Proc. Nat. Acad. Sci., Wash.

40.

(1956) J. biophys. biochem. Cytol. Suppl. 2, 261.

Gruneberg, H. The Genetics of the Mouse, ch. 6. Martinus Nijhoff, The Hague,

1952.
Gustavson, K. H. The Chemistry and Reactivity of Collagen. Academic Press, New

York, 1956.

The Chemistry of Tanning Processes. Academic Press, New York, 1956.

Haan De, R. L. In The Chemical Basis of Development, p. 339. Ed. by W. D.

McElroy and B. Glass, The Johns Hopkins Press, Baltimore, 1958.
Hackman, R. H. (1953) Biochem. J. 54, 362, 367, 371.

(1959) Proc. 4th Int. Cong. Biochem. Vienna (1958). Ed. by O. Hoffmann-

OSTENHOF, 12, 48.

Haddow, A., Elson, L. A., Roe, E. M. F., Rudall, K. M. and Timmis, G. M.

(1945) Nature, Lond. 155, 379.
Hadzi, J. in De Beer, G. R. Evolution as a Process, p. 24. Ed. by J. Huxley, A. C.

Hardy and E. B. Ford. Allen & Unwin, London, 1954.
Haguenau, F. (1958) Int. Rev. Cytol. 7, 425.

Hall, C. E. Introduction to Electron Microscopy. McGraw-Hill, New York, 1953.
Hanson, J. (1947) J. Anat. 81, 174.

Hanzon, V., Hermodsson, L. H. and Toschi, G. (1959) J. ultrast. Res. 3, 216.
Hardy, M. H. (1949) Jour. Anat., Lond. 83, 364.

(1952) Amer. J. Anat. 90, 285.

(1951) Ann. N. Y. Acad. Sci. 53, 546.

and Lyne, A. G. (1956) Aust. J. Biol. Sci. 9, 423.

Harkness, D. R. and Bern, H. A. (1957) Acta Anatomica 31, 35.
Harrap, B. S. (1955) Aust. J. Biol. Sci. 8, 122.

290 KERATIN AND KERATINIZATION

Harris, M. and Brown, A. E. in Fibrous Proteins. Sym. Soc. Dyers and Col.,

Leeds, p. 203. Chorley & Pickersgill, Leeds, 1946.
Harvey, E. N. and Danielli, J. F. (1938) Biol. Rev. 13, 319.
Haurowitz, F. Chemistry and Biology of Proteins. Academic Press, New York,

1950.
Hausman, L. A. (1925) Amer. Nat. 59, 529.

Heilmann, G. The Origin of Birds. H. F. & C. Witherby, London, 1926
Hendler, R. W. (1957) Fed. Proc. 16, 194.
Hergersberg, H. (1957) Z. f. Zellf. 45, 569.
Hess, A. (1958) J. biophys. biochem. Cytol. 4, 731.

van Heyningen, W. E. Bacterial Toxins, p. 52. Blackwell, Oxford, 1950.
Hildemann, W. H. (1959) Amer. Nat. 93, 27.
Hilditch, T. P. Chemical Composition of Natural Fats. Chapman & Hall,

London, 1949.
Hinglais-Guillaud, N. (1959) Bull, du Cancer 46, 212.
Hoagland, M. B., Zamecnik, P. C. and Stephenson, M. L. (1957) Biochem.

biophys. Acta 24, 215.
Hodge, A. J. (1959) Rev. Mod. Phys. 31, 409.

\th Internat. Congress for Electron Microscopy, 1958. Vol. 2, 119. Springer-

Verlag, 1960.
Holfreter, J. (1947)^. morphol. 79, 27.

(1948) Ann. N. Y. Acad. Sci. 49, 708.

Hooper, C. E. S. (1956) J. Hist. Cytochem. 4, 531.
Horio, M. and Kondo, T. (1953) Text Res. J. 23, 373.

Horstmann, E. Handbuch der mikroskopischen Anatomie des Menschen. Haut und
Sinnesorgane. Dritter Band, Dritter Teil. Die Haut., p. 1-276. Springer-
Verlag, Berlin, 1957.

(1955) Z. Zellforsch. 41, 532.

and Knoop, A. (1958) Z. Zellforsch. 47, 348.

Hosker, A. (1936) Phil. Trans. B226, 143.

Hotta, K. (1928) Chinese J. Physiol. 2, 345; (1929) Ibid. 3, 171.
Huggins, M. L. (1943) Chem. Rev. 32, 195.

(1958) J. Polymer Sci. 30, 5.

Hughes, T. E. (1959) J. exp. Biol. 36, 363.

Hutchinson, G. E. (1948) Ann. N. Y. Acad. Sci. 50, 221.

Hyman, L. H. Comparative Vertebrate Anatomy. Chicago University Press, 1947.

Ingram, V. M. (1957) Nature, Lond. 180, 326.

Jackson, S. Fitton (1956) Proc. Roy. Soc. B144, 556.

Jacobsen, C. F. and Linderstrgm-Lang, K. (1949) Nature, Lond. 164, 411.

Jarrett, A., Spearman, R. I. and Hardy, J. A. (1959) Brit. J. Dermatol. 71, 277.

and Spearman, R. I. (1961) Private communication.

Jeffrey, G. M., Sikorski, J. and Woods, H. J. (1955) Proc. Int. Wool Text. Conf.

Aust. Vol. F., p. 130; C.S.I.R.O., Melbourne, 1956.
Jennings, M. A. and Florey, H. W. (1956) Quart. J. exp. Physiol. 41, 131.
Johnson, E. (1958) J. Endocrinol. 16, 337.
Jones, B. M. (1958) Proc. Roy. Soc. B149, 263.
Jones, C. B. and Mecham, D. K. (1943) Arch. Biochem. 2, 209; 3, 193.

REFERENCES 291

Karrer, H. E. (1960) J. biophys. biochem. Cytol. 7, 181.

Kay, L. M., Schroeder, W. A., Munger, N. and Burt, N. (1956) J. Amer. Chem.

Soc. 78, 2430.
Kendrew, J. C. Structure Proteins, I. The Proteins. Ed. by H. Neurath and K.

Bailey, Vol. II, Part B. Academic Press, New York, 1954.

and Perutz, M. F. (1957) Ann. Rev. Biochem. 26, 327.

Kennaway, E. Sir (1955) Brit. Med. J. 2, 1382.

Kerr, M. F. and Godin, C. (1959) Canad.J. Chem. 37, 11.

King, H. T. and Nicholls, J. E. (1932) Trans. Faraday Soc. 29, 272.

Krebs, H. A. (1953) Brit. Med. Bull. 9, 92.

Krimm, S. and Schor, R. J. (1956) J. Chem. Phys. 24, 922.

Krishnan, G. (1953) Quart. J. micr. Sci. 94, 11 ; (1954) Ibid. 95, 371.

Kukita, A. J. (1957) J. Invest. Dermatol. 28, 273.

Lafon, M. (1943) Bull. Inst. Oceanogr. Monaco 40, No. 850. Quoted by Krish-

nam, 1953.
Lagermalm, G., Philip, B. and Lindberg, J. (1951) Nature, Lond. 168, 1080.
Lasnitzki, I. (1956) Brit. jf. Cancer 10, 510.

(1958) Ann. Rev. Cytol. 7, 79.

Laxer, G. and Whewell, C. S. (1955) Proc. Int. Wool Text. Conf. Austr. Vol. F.,

p. 186; C.S.I.R.O., Melbourne, 1956.
Leach, S. J. (1959) Rev. pure appl. Chem. 9, 33.
Leblond, C. P. Dynamics of Proliferating Tissue. Ed. by D. Price, University of

Chicago Press, 1958.

(1951) Ann. N. Y. Acad. Sci. 53, 464.

Puchtler, H. and Clermont, Y. (1960) Nature, Lond. 186, 784.

and Storey, W. F. (1956) Physiol. Rev. 36, 255.

Lees, K. and Elsworth, F. F. (1954) Proc. Int. Wool Text. Conf. Austr. Vol. C,

p. 363; C.S.I.R.O., Melbourne, 1955.
Lehmann, E. (1941) Melliand Textilber. 22, 145.
Lerner, A. B., Fitzpatrick, T. B., Calkins, E. and Summerson, W. H. (1949)

J. Biol. Chem. 178, 185.
Leuchtenberger, C. and Lund, H. Z. (1951) Exp. Cell Res. 2, 150.
Lindberg, J., Philip, B. and Gralen, N. (1948) Nature, Lond. 162, 458.

(1949) Text. Res. J. 19, 43.

Lillie, R. F. and Wang, H. (1941) Physiol. Zool. 14, 103.

(1942) Biol. Rev. 17, 247.

and Juhn, M. (1932) Physiol. Zool. 5, 124; (1938) 11, 434.

and Wang, H. (1944) Physiol. Zool. 17, 1.

and Wang, H. (1940) Proc. Nat. Acad. Sci., Wash. 26, 67.

Linderstrom-Lang, K. Symposium on Peptide Chemistry. Special Publication 2,

The Chemical Society, London, 1958.

and Duspiva, F. (1936) C. R. Lab. Carlsberg Sir. Chim. 21, 53.

Lindley, H. (1947) Nature, Lond. 160, 190.

(1955) Proc. Int. Wool Text. Conf. Austr. Vol. B., p. 193; C.S.I.R.O., Mel-

bourne, 1956.
Littlefield, J. W. and Keller, E. B. (1957)^. Biol. Chem. 224, 13.
Litvac, A. (1939) Arch, d'anat. microsc. 35, 55.
Lochte, Th. Atlas der menschlichen und tierschen Haare. Paul Schops, Leipzig,

1938.

292 KERATIN AND KERATINIZATION

Loftfield, R. B. (1957) Prog. Biophys. Chem. 8, 347.

Longley, J. B. (1950) Ph.D. Thesis, Cambridge.

Lorand, L. and Middlebrook, W. R. (1952) Biochem. J. 52, 196.

(1952) Biochem. J. 52, 200.

Losee, F. L. and Hess, W. C. (1949) J. Dental Res. 28, 512.

Lotmar, W. and Picken, L. E. R. (1950) Experientia 6, 58.

Low, B. W. and Baybutt, R. B. (1952) J. Amer. Chem. Soc. 74, 5806.

Lucas, F., Shaw, J. T. B. and Smith, S. G. (1956) Nature, Lond. 178, 861.

Shaw, J. T. B. and Smith, S. G. (1958) Adv. Prot. Chem. 13, 107.

Ludford, R. J. (1925) Proc. Roy. Soc. B98, 557.

Ludicke, M. (1959) Naturwissenschaften 46, 658.
Luft, J. H. (1956) J. biophys. biochem. Cytol. 3, 799.
Lundgren, H. P. (1945) Text. Res. J. 15, 335.

MacArthur, I. (1943) Nature, Lond. 152, 38.

McDonough, E. G. (1952) Chem. Engng. News 30, 5254.

McLoughlin, C. B. Ph.D. Thesis, Cambridge, 1959.

Manogue, B. and Moss, M. S. (1953) Nature, Lond. 172, 806.

Manton, I. (1952) Symp. Soc. Exper. Biol. 6, 306.

Marsh, R. E., Pauling, L. and Corey, R. B. (1955) Biochem. biophys. Acta 16, 1.

Marston, H. in Fibrous Proteins. Symp. Soc. Dyers and Col. p. 207; Chorley &

Pickersgill, Leeds, 1946.
Martin, A. J. P. Fibrous Proteins. Symp. Soc. Dyers and Col. p. 1 ; Chorley &

Pickersgill, Leeds, 1946.
Mason, H. S. Pigment Cell Biology, p. 563. Ed. by M. Gordon, Academic Press,

New York, 1953.

(1955) Advanc. Enzymol. 16, 105.

Matoltsy, A. G. (1958) in The Biology of Hair Growth. Ch. 7, p. 135. Academic

Press, New York, 1958.
Maximow, A. A. and Bloom, W. A Textbook of Histology. 5th Ed. Saunders.

Philadelphia and London, 1948.
Mazia, D. (1955) Symp. Soc. Exper. Biol. 9, 335.

and Dan, K. (1952) Proc. Nat. Acad. Sci., Wash. 38, 826.

Medawar, P. B. (1953) Quart. J. micros. Sci. 94, 481.
Menefee, M. G. (1957) J. Ultrast. Res. 1, 49.

Menkart, J. and Coe, A. B. (1958) Text. Res. J. 28, 218.
Mercer, E. H. (1949a) Nature, Lond. 163, 18.

(1949b) Biochim. biophys. Acta 3, 161.

(1949c) Ph.D. Thesis, Leeds University.

(1949d) J. Text. Inst. 40, T640.

Lindberg, J. and Philip, Bo. (1949e) Text. Res. J. 19, 678.

and Olofsson, B. (1951a) J. Polymer Sci. 6, 261.

(1951b) J. Polymer Sci. 6, 671.

(1951c) Nature, Lond. 168, 792.

(195 Id) Nature, Lond. 168, 824.

and Day, M. F. (1952a) Biol. Bull. 103, 384.

(1952b) The Sci. Mon. 75, 280.

(1953) Text. Res. J. 23, 388.

(1954) Text. Res. J. 24, 39.

Farrant, J. L. and Rees, A. L. G. (1955) Proc. Int. Wool Text. Conf.Austr.

Vol. F., p. 120; C.S.I.R.O., Melbourne, 1956.

REFERENCES 293

Mercer, E. H. The Biology of Hair Growth, p. 91. Ed. by W. Montagna and
R. A. Ellis, Academic Press, New York, 1958.

(1959) Proc. Roy. Soc. B150, 216.

Electron Microscopy in Anatomy, p. 100. E. Arnold, London, 1961.

Meyer, J., Medak, H. and Weinmann, J. P. (1960) Growth 24, 29.
Meyer, K. (1945) Advanc. Protein Chem. 2, 249.

and Rapport, M. M. (1951) Science 113, 596.

(1957) Harvey Lectures 51, 88.

Meyer, K. H. and Haselbach, C. (1949) Nature, Lond. 164, 33.

and Mark, H. Der Aufbau der hochpolymeren organischen Naturstoffe, Acad.

Verlagages. Leipzig, 1930.

Wyk, Vander, A. J. A., Gonon, W. and Haselbach, C. (1952) Trans. Faraday

Soc. 48, 669.
Middlebrook, W. R. (1951) Biochim. biophys. Acta 7, 547.
Miszurski, B. (1937) Arch. exp. Zell. forsch. 20, 123.
Moffitt, W. (1956) J. Chem. Phys. 25, 467.
Mohn, M. P. The Biology of Hair Growth, p. 336. Ed. by W. Montagna and

R. A. Ellis, Academic Press, New York, 1958.
Montagna, W. (1950) Quart J. micros Sci. 91, 205.

Chase, H. B., Malone, J. D. and Melaragno, H. P. (1952) Quart. J.

Microbiol. Sci. 93, 241.

The Structure and Functions of Skin. Academic Press, New York, 1956.

Chase, H. B. and Hamilton, J. B. (1951)^. Invest. Derm. 17, 147.

Moog, F. (1946) Biol. Rev. 21, 41.

Morales, M. F. and Kreutzer, F. L. (1945) Btdl. Math. Biophys. 7, 15.
Moran, T. and Hale, H. P. (1936) J. exp. Biol. 13, 35.
Moscona, A. (1952) Exp. Cell Res. 3, 535.

(1956) Proc. Soc. exp. Biol. N. Y. 92, 410.

(1957) Proc. Nat. Acad. Sci., Wash. 43, 184.

(1957) Science 125, 598.

Muller, C. (1939) Z.f. Zellforsch. 29, 1.

Nageotte, J. (1927) C. R. Acad. Sci., Paris 184, 115.

Neurath, H. (1940) J. phys. Chem. 44, 296.

Nieuwkoop, P. D. and Faber, J. Normal Tables of Xenopus laevis. North Holland,

Amsterdam, 1956.
Nilsson, O. (1958) Exp. Cell Res. 14, 341.

(1959) J. Ultrastr. Res. 2, 331, 342.

Niu, M. C. Pigment Cell Biology, p. 37. Ed. by M. Gordon, Academic Press, New

York, 1959.
Noback, C. R. (1951) Ann. N. Y. Acad. Sci. 53, 627.
Norris, M. H. (1931)7 Text. Inst. 22, T17.

Odland, G. F. (1953) y. Invest. Dermat. 21, 305.

(1958), y. biophys. biochem. Cytol. 4, 529.

O'Donnell, I. J. and Woods, E. F. (1955a) Proc. Int. Wool Text. Conf. Austr.
Vol. B., p. 48; C.S.I.R.O., Melbourne, 1956a.

and Woods, E. F. (1956b) y. polymer Sci. 21, 397.

294 KERATIN AND KERATINIZATION

Olofsson B. and Gralen, N. (1947) Proc. 11th Inter. Conf. Chem. 5, 151,

London, 1953.
Oparin A. I. The Origin of Life on the Earth. 3rd Ed. Translated by A. Synge,

Oliver & Boyd, London, 1957.
Oster, G. and Pollister, A. W. Physical Techniques in Biological Research.

Academic Press, New York, 3 vols, 1955-6.
Ottoson, D., Sjostrand, F. S., Stenstrom, S. and Svaetichin, G. (1953) Acta

Physiol. Scand. 29, 611.

Palade, G. E. (1952) J. exp. Med. 95, 285.

(1955) J. biophys. biochem. Cytol. 1, 59, 567.

(1956) J. biophys. biochem. Cytol. 2, 417.

and Siekevitz, P. (1956) J. biophys. biochem. Cytol. 2, 171.

Parat, M. (1928) Arch. anat. microsc. 24, 73.

Parker, K. D. (1955) Biochim. biophys. Acta 17, 148.

and Rudall, K. M. (1957) Nature, Lond. 179, 905.

Paschkiss, K. E. (1958) Cancer Res. 18, 981.

Patterson, W. I., Geiger, W. B., Mizell, L. R. and Harris, M. (1941) J. Res.

Nat. Bur. Stand. 27, 89.
Pauling, L., Corey, R. B. and Branson, H. R. (1951) Proc. Nat. Acad. Set.,

Wash. 37, 205.

and Corey, R. B. (1951) Proc. Nat. Acad. Sci., Wash. 37, 261.

(1951) Proc. Nat. Acad. Sci., Wash. 37, 251.

(1953a) Nature, Lond. 171, 59.

(1953b) Disc. Faraday Soc. 13, 170.

Symposium on Protein Structure, p. 17. Ed. by A. Neuberger, Methuen,

London, 1958.
Pautard, F. G. E. (1958) Nature, Lond. 182, 788.

(1961) Arch, oral Biol. 3, 217.

Peacock, N. (1959) Biochim. biophys. Acta 32, 220.
Pearse, A. G. E. (1951) Quart. J. micros. Sci. 92, 393.

Histochemistry, Theoretical and Applied. Churchill, London, 1953.

Pelc, S. R. (1959) Exp. Cell Res. Suppl. 6, 97.

(1958) Exp. Cell Res. 6, 105.

Perutz, M. F. (1951) Nature, Lond. 168, 653.

and Huxley, H. E. (1951) Nature. Lond. 167, 1053.

(1959) Research 12, 8-9, 326-334.

Peters, L. in Fibrous Proteins, Soc. Dyers and Col., p. 138 ; Chorley & Pickersgill,
Leeds, 1946.

and Speakman, J. B. (1949)^. Soc. Dyers Col. 65, 63.

and Woods, H. J. in The Mechanical Properties of Textile Fibres, p. 153. Ed.

by R. Meredith, North Holland, Amsterdam, 1956.
Peters, R. A., Sir in Perspectives in Biochemistry, p. 36. Ed. by J. Needham and

D. E. Green, Cambridge University Press, 1937.
Philip, Bo., Lagermalm, G. and Gralen, N. (1951) Biochem. biophys. Acta 6,

497.
Pillai, P. A., Guenin, H.-A. and Gautier, A. (1960) Bull. Soc. Vaud. Sci. Nat.

67, 215.
Pillemer, L., Ecker, E. E. and Wells, J. R. (1938) Science 88, 16; (1939) J.

exp. med. 69, 191.

REFERENCES 295

Pinkus, F. Die normale Anatomie der Haut. Handbuch der Haut und Geschlechs-
kankeiten. Vol. 1, Part 1. Ed. by Jadassohn S. Karger, Berlin, 1927.

Pinkus, H. (1951) J. Invest. Dermatol. 16, 383.

Policard, A. and Collett, A. (1959) Rev. franc. Etud. din. biol. 4, 283.

Porter, K. R. (1954) Proceed, of Internat. Conf. Electr. Micros., London, p. 539.
Ed. by R. Ross, 1956.

(1957) Harvey Lectures 51, 175.

(1954) J. Histochem. Cytochem. 2, 346.

Price, D. (Editor) Dynamics of Proliferating Tissues. The University of Chicago

Press, Chicago, 1958.
Pringle, J. W. S. (1951) Behaviour 3, 174.
Pryor, M. G. M. (1940) Proc. Roy. Soc. B128, 378, 393.
Puchtler, H. and Leblond, C. P. (1958) Amer.J. Anat. 102, 1.
Puck, T. T. in Cellular Biology, Nucleic Acids and Viruses Special Publ. N. Y.

Acad. Sci. Vol. 5, p. 293, 1957.

Rabinovitz, M. and Olson, M. E. (1957) Fed. Proc. 16, 235.

Randall, J. T. (Editor) Nature and Structure of Collagen. Butterworths, 1953.

Ranvier, E. (1879) C. R. Acad. Sci., Paris 88, 1361.

Rashevsky, N. Mathematical Biophysics. Revised edition, University of Chicago

Press, 1948.
Rawles, M. (1947) Physiol. Rev. 28, 383.

Reed, R. and Rudall, K. M. (1948) Biochim. biophys. Acta 2, 7.
Rees, A. L. G. (1951) J. phys. coll. Chem. 55, 1340.
Reeve, E. C. R. and Huxley, J. S. in Essays on Growth and Form, p. 121. Ed. by

W. E. Le Gros Clark and P. B. Medawar, Clarendon Press, Oxford,

1945.
Reinberg, A. and Ghata, J. Rythmes et Cycles Biologiques. Presses Universitaires

de France, Paris, 1957.
Rich, A. and Crick, F. H. C. (1955) Nature, Lond. 176, 915.
Richards, A. G. The Integument of Arthropods. University of Minnesota Press,

Minneapolis, 1951.
Richards, O. W. and Kavanagh, A. J. in Essays on Grozvth and Form, p. 188. Ed.

by W. E. Le Gros Clark and P. B. Medawar, Clarendon Press, Oxford,

1945.
Robertson, J. D. (1956) J. biophys. biochem. Cytol. 2, 381; (1957) Ibid. 3, 1043;

(1958) Ibid. 4, 349.

(1957) Biochem. Society Symposium, No. 16, p. 1. Ed. by E. M. Crook,

Cambridge University Press, 1959.
Robinson, C. The Nature and Structure of Collagen. Butterworths, London, 1953.

and Ambrose, E. J. (1952) Trans. Faraday Soc. 48, 854.

Rogers, G. E. (1953) Quart. J. Micr. Sci. 94, 253.

(1959a) Ann. N. Y. Acad. Sci. 83, 378, 408.

(1959b) J. Ultrastr. Res. 2, 309.

Romer, A. S. The Vertebrate Body, 2nd Ed. Saunders, Philadelphia, 1955.

Rose, S. M. (1952) Amer. Nat. 86, 337.

Rothman, S. Physiology and Biochemistry of the Skin. University of Chicago Press,

1954.
Rougvie, M. A. Ph.D. Thesis, M. I. T., 1954, cited by A. M. Woodin (1955)

Nature, Lond. 176, 1117.

296 KERATIN AND KERATINIZATION

Rudall, K. M. Ph.D. Thesis, Leeds, 1936.

(1941) Proc. Leeds phil. lit. Soc. 4, 13.

in Fibrous Proteins. Svmph. Soc. Dyers and Col. Chorley & Pickersgill,

Leeds, p. 15, 1946.

(1947) Biochem. biophys. Acta 1, 549.

(1952) Advanc. Protein Chem. 7, 253.

(1953) (Discussion) Proc. Roy. Soc. B141, 39.

(1955) Symp. Soc. exp. Biol. 9, 49.

(1955) Proc. Int. Wool Text. Conf. Austr. Vol. F., p. 176; C.S.I.R.O., Mel-

bourne, 1956.

(1955-6) Led. Sci. Basis Med. V, 217. Athlone Press, London, 1956.

Ryder, M. L. (1955) Proc. Wool Text. Conf. Austr. Vol. F., p. 63, 1956.

(1956) J. Agric. Sci. 47, 129.

(1957) J. Agric. Sci. 49, 275.

in The Biology of Hair Growth. Ch. 14, p. 305, Academic Press, New York,

1958.

(1959) Quart, y. Microsc. Sci. 100, 1.

Salecker, J. (1944) Morphol. ydhrb. 88, 225.
Schellman, J. A. (1955) C. R. Carlsberg. Ser. chim. 29, 230.
Scheving, L. E. (1959) Anat. Rec. 135, 7.

Schmidt, W. J. Die Bausteine des Tierskorpers im polarisierten Lichte. Frederich
Cohen, Bonn, 1924.

(1925) Mikrokosmos 19, 65.

Schmitt, F. C, Gross, J. and Highberger, J. H. (1955) Symp. Soc. exp. Biol.

Fibrous Proteins 9, 148.
Schmitt, F. O. (1941) Growth 5, 1.

Advanc. Protein Chem. 1, 25. Ed. by M. L. Anson and J. T. Edsall,

Academic Press, New York, 1944.

- (1957) y. Cell Comp. Physiol. 49 (Suppl.), 85.

(1958) Proc. 4th Int. Cong. Electr. Micros. Vol. 2, p. 1. Springer- Verlag,

Berlin, 1960.

and Bear, R. S. (1939) Biol. Rev. 14, 27.

Schneider, K. C, Lehrbuch der vergleichen Histologic der Tiere. Fischer, Jena, 1902.
Schor, P. (1958) Ph.D. Thesis, Michigan, 1958; (1958) Diss. Abst. 19, No. 6, 1189.
Schroeder, W. A., Kay, L. M., Lewis, B. and Munger, N. (1955) y. Amer. chem.

Soc. 77, 3901.
Schulz, H. and de Paola, D. (1958) Z.f. Zellforsch. 49, 125.
Scott, D. B. (1955) Ann. N. Y. Acad. Sci. 60, 575.
van Scott, E. J. and Flesch, P. (1954) Arch. Derm. Syph. 70, 141.
Selby, C. C. (1955) y. biophys. biochem. Cytol. 1, 429.

(1957) y. Invest. Dermatol. 29, 131.

Setala, K., Merenmies, L., Stjernvall, L. and Nyholm, M. (1960) y. Nat.

Cancer Inst. 24, 329.
Siekevitz, P. (1952) y. Biol. Chem. 195, 549.

and Palade, G. E. (1958) y. biophys. biochem. Cytol. 4, 557.

Sikorski, J. (1958) Proc. 4th Int. Congr. Elec. Micros. Berlin. Springer- Verlag,
Berlin, 1960.

- and Simpson, W. S. (1959) y. Roy. micros. Soc. 78, 35.

REFERENCES 297

Simkin, J. L. and Work, T. S. The Biological Replication of Macromolecules,
p. 164, Cambridge, 1958; (1958) Symp. Soc. exp. Biol. 12, 164.

(1959) Ann. rev. Biochem. 28, 145.

Simmonds, D. H. (1954) Austr.J. biol. Sci. 7, 98.

(1955) Austr. J. biol. Sci. 8, 114.

(1958) Text. Res. J. 28, 314.

Simmonds, D. H. and Stell, I. G. (1955) Proc. Int. Wool Text. Conf. Austr

Vol. C, 75; C.S.I.R.O., Melbourne, 1956.
Sjostrand, F. S. and Hanzon, V. (1954) Exp. Cell Res. 7, 393.

Internat. Rev. Cytol. 5, 455. Eds. Bourne, G. H. and Danielli, J. F.,

New York, Academic Press, 1956.
Skertchley, A. and Woods, H. J. (1960) J. Text. Inst. 51, T517.
Slen, S. B. (1958) Nature, Lond. 182, 1322.

Sock, N. W. and Morales, M. F. (1945) Bull. Math. Biophys. 7, 15.
Speakman, J. B. (1928) Proc. Roy. Soc. B103, 377.

(1931) Proc. Roy. Soc. A132, 167.

and Hirst, M. (1933) Trans. Faraday Soc. 29, 148.

(1934).?. Soc. Dyers Col. 50, 34.

(1936) Nature, Lond. 138, 327.

(1947)^. Text. Inst. 38, T102.

and Stott, E. (1931) J. Text. Inst. 22T, 339.

(1955) Symp. Soc. experim. Biol. 9, 169.

Spearman, R. I. and Barnicot, N. A. (1960) Amer. J. Phys. Anthrop. 18, 91.
Spier, H. W. and van Caneghem, P. (1957) Arch. klin. u. exptl. Dermatol. 202, 120.
Stainsby, G. (Editor) Recent Advances in Gelatin and Glue Research. Pergamon

Press, New York, 1958.
Stephenson, M. L., Hecht, L. I., Littlefield, J. W., Loftfield, R. B. and

Zamecnik, P. C. Subcellular Particles. Ed. by T. Hayashi, Ronald Press,

New York, 1959.
Stoeckenius, W. (1959) J. biophys. biochem. Cytol. 5, 491.
Stoves, J. L. (1942) Analyst 67, 385.

(1943) Proc. Leeds phil. lit. Soc. 4, 84.

(1947) J. Soc. Dyers Col. 63, 65.

in Fibrous Proteins, Symp. Soc. Dyers and Col. Chorley & Pickersgill, Leeds

p. 58, 1946.
Strangeways, D. H. (1931) Arch. exp. Zellforsch. 11, 344.
Strong, R. M. (1902) Bidl. Mus. Comp. Zool. Harv. 40, 147.
Sutherland, G. B. B. M. (1952) Advanc. Protein Chem. 7, 291.
Swanbeck, G. (1959) J. Ultrastr. Res. 3, 51.
Swann, M. M. (1957) Cancer Res. 17, 727; (1958) Ibid. 18, 1118.
Sylven, B. (1941) Acta Chirurg. Scand. 86, Suppl. 66, p. 1.

(1950) Exper. Cell Res. 1, 582.

Taylor, H. S. (1941) Science 93, 465 (Report).

Thomas, J. A. (Editor) Les Facteurs de la Croissance Celhdaire. Masson, Paris,

1956.
Thompson, D'arcy W., On Growth and Form, 2nd ed. Cambridge University

Press, 1942.
Thompson, E. O. P. (1957) Austr.J. biol. Sci. 10, 225.
and O'Donnell, I. J. (1959) Austr. J. biol. Sci. 12, 282.

298 KERATIN AND KERATINIZATION

Thuringer, J. M. (1924) Anat. Rec. 28, 31.

Tristram, G. R. The Proteins. 1A, p. 220. Ed. by H. Neurath and K. Bailey,

Academic Press, New York, 1953.
Turner, C. D. General Endocrinology. Saunders, Philadelphia and London, 1960.

Unna, P. G. Histochemie der Haut. F. Deuticke, Leipzig, 1928.

Vigneaud, du V. Proc. XI Intern. Cong. Pure and App. Chem., 1947.

Waddington, C. H. (1948) Symp. Soc. exp. Biol. 2, 145.

The Epigenetics of Birds. Cambridge University Press, 1952.

Principles of Embryology. Allen & Unwin, London, 1956.

Waldschmidt-Leitz, E. and Zeiss, D. (1955) Z. physiol. Chem. 300, 49.
Ward, W., High, L. M. and Lundgren, H. P. (1946) J. Polymer Res. 1, 22.

and Lundgren, H. P. (1954) Advanc. Protein Chem. 9, 243.

Warwicker, J. O. (1954) Acta Cryst., Camb. 7, 565.

(1959) Nature, Lond. 184, BA25 (Rep. Conf. Inst. Phys. Leeds).

Watson, J. D. and Crick, F. H. C. (1953) Nature, Lond. 171, 737.
Watson, M. L. (1954) Biochim. biophys. Acta 15, 475.

(1960) J. biophys. biochem. Cytol. 7, 489.

and Avery, J. K. (1954) Amer. J. Anat. 95, 109.

Waugh, D. F. (1954) Advanc. Protein Res. 9, 325.

Weiss, L. (1960) Int. Rev. Cytol. 9, 187. Ed. by G. H. Bourne and J. F. Danielli.
Weiss, P. (1933) Amer. Nat. 67, 322.

(1945) J. exp. Zool. 100, 353.

(1950) Quart. Rev. Biol. 25, 177.

(1958) Int. Rev. Cytol. 7, 391. Ed. by G. H. Bourne and J. F. Danielli.

and Ferris, W. (1954) Proc. Nat. Acad. Set., Wash. 40, 528.

and James, R. (1955) Exp. Cell Res. Suppl. 3, 381.

Specificity in Growth Control. Ch. 10, p. 195. Ed. by E. G. Butler, Princeton

University Press, 1955.

and Kavenau, J. L. (1957) J. gen. Physiol. 41, 1.

Whewell, C. S. and Woods, H. J. in Fibrous Proteins. Symp. Soc. Dyers Col.

p. 50, Chorley & Pickersgill, Leeds, 1946.
Whiteley, H. T. (1958) J. Anat. 92, 593.
Wildman, A. B. (1955) Proc. Int. Wool Text. Conf. Austr. Vol. F., p. 157;

C.S.I.R.O., Melbourne, 1956.
(1932) Proc. Zool. Soc. Lond. 2, 257.

and Manby, J. (1938) Trans. Roy. Soc, Edinb. 59, 333.

Wilkins, M. H. F., Zubay, G. and Wilson, H. R. (1959) J. molec. Biol. 1,

179.
Willmer, E. N. Tissue Cidture. Methuen monographs. Revised, Methuen, London,

1954.
Wislocki, G. B. (1951) Anat. Rec. 109, 388.
Wolbach, S. B. (1951) Ann. N. Y. Acad. Sci. 53, 517.
Woodin, A. M. (1954a) Nature, Lond. 173, 823.

(1954b) Biochem. J. 57, 99.

(1955) Nature, Lond. 176, 1117.

(1956) Biochem. J. 63, 576.

REFERENCES 299

Woods, E. F. (1952) Austr. J. Sci. Res. 5A, 555.

(1959) Austr. J. biol. Sci. 12, 96.

(1959) Austr. J. chem. 12, 497.

Woods, H. J. (1938) Proc. Roy. Soc. B166, 76.

(1946) J. Colloid Sci. 1, 407.

Yakel, H. L. (1953) Acta Cryst., Camb. 6, 724.

Young, J. Z. (1950) The Life of Vertebrates. Oxford University Press, 1950.

Zahn, H. (1947) Naturforsch. 26, 104.

(1947) Z.f. Naturforsch. 26, 286.

(1949) Kolloid-Z. 113, 157.

Zamecnik, P. C. and Keller, E. B. (1954) J. Biol. Chem. 209, 337.

Keller, E. B., Littlefield, J. W., Hoagland, M. B. and Loftfield, R. B.

(1956) J. cell. comp. Physiol. 47, Suppl. 1, 81.
Zubay, G. (1959) y. molec Biol. 1, 186.
Zwilling, E. (1955) ^. exp. Zool. 128, 423.

Author Index

Abercrombie, M. 85, 140
Alexander, P. 163, 191, 234, 237,

238, 239, 247, 248, 249, 252, 259
Alfert, M. 264
Allfrey, V. G. 264
Ambrose, E. J. 85, 176, 196, 197, 199,

208
Arndt, U. W. 194
Arthur, R. P. 147
Astbury, W. T. 2, 11 et seq., 27, 121,

123, 165, 170, 172, 174, 175, 180, 183,

189, 191, 192, 193, 200, 201, 203,

206,246, 254, 257, 258
Auber, L. 73, 100, 157, 159, 210, 219,

226, 227, 268, 274, 275

Bachra, B. N. 131

Bahr, G. F. 35, 36 (Fig. 17), 37

Bailey, K. 121, 200

Baker, J. R. 37,55,115

Baker, R. V. 281

Baldwin, E. 50

Bamford, C. H. 120, 180, 183, 192,

197, 199
Barnes, R. J. 10, 191, 213
Barnicot, N. A. 274, 279, 280, 282
Barrnett, R. J. 29,49,51,217
Battistone, G. C. 78
Baybutt, R. B. 181
Bear, R. S. 127, 166, 167, 168, 206,

207, 213
Beighton, E. 201, 203
Bell, F. O. 13 et seq., 175, 176
Bellamy, L. T. 198
Bendit, E. G. 192
Bennett, H. S. 29
Bergen, von, W. 7
Bergmann, M. 116
Bern, H. A. 232, 264
Bernal, J. D. 130
Bernhard, W. 46,47,111

BlEDERMANN, W. 80, 214, 228

Billingham, R. E. 62, 276, 277
Birbeck, M. S. C. 86, 92, 99 (Fig. 44d),

112, 114, 168, 224, 227, 262, 267, 275

279, 280, 281, 282

Blackburn, S. 163

Blair, S. M. 264

Blaschko, H. 281

Block, B. 280

Block, R. J. 29, 31, 78

Bloom, W. 80

Boedtker, H. 127

Bolliger, A. 220, 263

Bolling, D. 29, 31

Borsook, H. 116

Borum, K. 276

Brachet, J. 109

Bradfield, J. R. G. 39, 136, 221

Branson, H. R. 176 et seq.

Braun-Falco, O. 113, 210, 222

Brody, I. 225 (Fig. 98), 229, 247, 248

Broussy, J. 30, 107

Brown, A. E. 252, 254

Brown, C. H. 20, 30

Brown, L. 180, 183

Brunet, P. 276, 280

Bull, H. B. 257

Bulliard, H. 64, 217, 231

Bullough, W. S. 134, 135, 136, 144

148, 149, 221
Bunn, C. W. 5, 14
Burge, R. E. 187
Burgos, M. H. 145
Burnett, G. W. 78
Burns, M. 150
Burrows, H. 144
Burt, N. 180
Burte, H. 193, 250
Butcher, E. O. 151, 152

Cairns, J. M. 60, 61

Calvery, H. O. 30

Caneghem, van, P. 263

Carruthers, C. 221

Carter, H. B. 75, 78 (Fig. 35), 150

Cassperson, T. O. 109

Cauna, N. 273

Champetier, C. 30, 107, 214

Charles, A. 95, 131

Chase, H. B. 62, 64, 151, 152

Cherry, C. P. 63

300

AUTHOR INDEX

301

Chevremont, M. 29, 217

Clark, W. E. Le Gros 69

Clarke, W. H. 75, 150

Clarkson, H. 150

Claude, A. 115

Cochran, W. 176

Coe, A. B. 275

Cohen, 127

Collett, A. 88

Colvin, J. R. 33

Coman, C. R. 85

Consdon, R. 234

Corey, R. B. 172, 176 et seq., 183

Corfield, M. C. 7, 238, 239, 261

Cowdry, E. V. 80

Crane, H. R. 128

Crick, F. H. C. 2,109,116,119,177,

183
Cruise, A. J. 24
Cuckow, F. W. 279

Dalton, A. J. 46, 115

Daly, M. M. 264

Dan, K. 30

Danforth, C. H. 72, 75

Danielli, J. F. 37, 38 (Fig. 18), 47,

93, 222
Daemon, S. E. 198
Das, D. B. 240
Davies, H. R. 104
Davson, H. 37
Day, M. F. 124, 130
Derksen, J. C. 214, 235
Dickinson, S. 200
Donohue, J. 182, 194
Doty, P. 127
Dry, F. W. 156, 159
Durward, A. 137, 152, 232
Dusenbury, J. H. 275

Earland, C. 239, 244

Ebling, F. J. 137, 155

Ecker, E. E. 119

Eisen, A. Z. 230

Elliott, A. 120, 180, 183, 193, 196,

197, 199, 208
Ellis, W. J. 236
Ellsworth, F. F. 272
Elod, E. 252, 257, 259, 260
Engstrom, A. 230
'Espinasse, P. G. 104

Faber, J. 90

Farrant, J. L. 128, 168, 170, 246, 261

Faure-Fremiet, E. 30, 107

Fawcett, D. W. 39, 40, 42, 83

Felix, M. D. 46, 115

Fell, H. B. 63, 107, 220, 222, 264

Ferris, W. 86, 90, 130

Feughelman, M. 176

Fischer, A. 56

Fitzpatrick, T. B. 276, 280, 282

Fleischhauer, K. 77

Flesch, P. 230

Florkin, M. 26

Fox, S. W. 33

Fraenkel, G. 20

Fraenkel-Conrat, H. 116

Fraser, A. S. 149, 150, 151, 159

Fraser, R. D. B. 170, 188, 194, 200,

209, 274
Frederic, J. 29
Frey-Wyssling, A. 3

Garnier, Ch. 47

Gautier, A. 83, 263

Geiger, W. B. 252, 254, 266

Gelfant, S. 136

Geren, Ben, B. 47

Ghata, J. 137

Gillespie, J. M. 219, 238, 241, 244

Giroud, A. 64, 214, 217, 231

Glinos, A. D. 140

Glucksmann, A. 63, 94, 276

Goddard, D. R. 236

Godin, C. 163

Gordon, A. H. 234

Gordon, M. 276

Gralen, N. 236

Grasse, P. P. 46

Gray, E. G. 19

Grimstone, A. V. 39

Gross, J. 127

Gross, R. 220, 263

Gruneberg, H. 276

Guenin, H. A. 83, 263

Gustavson, K. H. 19

Gutman, M. 257

Haan, de, R. L. 85
Hackman, R. H. 20

302

AUTHOR INDEX

HADDOW, A. 151

Hadzi, J. 26

Haguenau, F. 46, 47, 1 1 1

Hale, H. P. 107

Hall, C. E. 5

Halsey, G. 193, 250

Hanby, W. E. 76, 120, 176, 183, 198,

199
Hanson, J. 82
Hanzon, V. 111,112
Happey, F. 120
Hardy, J. A. 228
Hardy, M. H. 56, 78, 81, 96, 112,

137, 218
Hardy, W. E. 180
Harkness, D. R. 264
Harrap, B. S. 289
Harris, M. 252, 254
Harvey, E. N. 38, 47
Haselbach, C. 258
Haurowitz, F. 119
Hausman, L. A. 71
Heaysman, J. E. M. 85
Hecht, L. I. 118
Hendler, R. H. 118
Hergersberg, H. 51
Heringa, G. C. 214
Hermodsson, L. H. 112
Hess, A. 19, 78
Heyningen, van, W. E. 135
High, L. M. 207

HlLDEMANN, W. H. 60

Hinglais-Guillard, N. 62, 264

Hirst, M. 250

Hodge, A. J. 126, 127, 168, 170

HOLFRETER, J. 84, 90

Hoagland, M. B. 111,117,118

Hooper, C. E. S. 135

Horio, M. 273

Horstmann, E. 69, 80, 82, 90, 263

Hosker, A. 103 (Fig. 46), 104, 105

(Fig. 47)
Hotta, K. 59
Hudson, R. F. 191, 234, 239, 247,

249
Huggins, M. L. 176, 178, 181
Hughes, T. E. 30
Huxley, H. E. 182
Huxley, J. 160
Hyman, L. H. 74

Ingram, V. M. 34

Jackson, S. F. 86, 127, 131

Jacobsen, C. F. 234

James, R. 63

Jarrett, A. 228, 231

Jeffrey, G. M. 24, 168, 170

Johnson, E. 89 (Fig. 40), 115

Johnson, M. 136

Jones, B. M. 279

Jones, C. B. 236, 237, 243, 272

Juhn, M. 104

Kavanagh, A. J. 160
Kavenau, J. L. 138, 140
Kay, L. M. 8, 33, 180
Keech, M. K. 24
Keller, E. B. 116, 120
Kendrew, J. C. 170, 194
Kennaway, E. 275
Kerr, M. F. 163
Knoop, A. 263
Kondo, T. 273
Kreautzer, F. L. 140
Krimm, S. 33, 163, 208
Krishnam, G. 30
Kukita, A. J. 276, 280

Lafon, M. 30
Lagermalm, G. 261, 267
Lasnitzki, I. 63, 64
Laurence, E. B. 136, 137, 149
Laxer, G. 279
Leach, S. J. 195

Leblond, C. P. 43,54,64,82,84,133
Lee, G. R. 163
Lees, K. 272
Lehmann, E. 72, 265
Lennox, F. G. 241 et seq., 244
Lillie, R. F. 96, 104, 135
Lindberg, J. 261, 265, 267, 268
Linderstrom-Lang, K. 200, 234
Lindley, H. 189, 190, 244, 274
Lindstrom, B. 230

LlTTLEFIELD, J. W. Ill, 118, 120

Litvac, A. 56

Lochte, Th. 71

Loftfield, R. B. Ill, 118, 120

LONGLEY, J. B. 107

Lorand, L. 126

AUTHOR INDEX

303

Lotmar, W. 24

Low, B. W. 181

Lucas, F. 180

Ludford, R. J. 275

Luft, J. H. 35

Lundgren, H. P. 8, 9, 161, 162, 207,

235, 238, 243, 249, 272
Lyne, A. G. 78

Macarthur, I. 166, 170, 182, 183

McDonough, E. G. 254

McLoughlin, C. B. 40, 60

MacRae, T. P. 170, 188, 194, 200,
209

Malcolm, B. R. 198

Manogue, B. 267

Manton, I. 39

Mark, H. 172, 180

Marsh, R. E. 172

Marston, H. 159, 218, 232

Martin, A. J. P. 234

Marwick, T. C. 24, 170, 206, 257

Mason, H. S. 280, 281

Matoltsy, A. G. 228, 231, 261, 271

Maximov, A. A. 80

Mazia, D. 30

Mecham, D. K. 7, 236, 243, 272

Medak, H. 135

Medawar, P. B. 276, 277

Meijere, de 75, 76 (Fig. 33)

Mellanby, E. 63, 220

Menkart, J. 275

Mercer, E. H. 30, 38, 47, 49, 86, 90,
92, 99, 104, 107, 112, 114, 121, 124,
128, 129, 130, 156, 166, 168, 170,
191, 200, 209, 211, 216, 224, 227,
230, 235, 237, 238, 246, 248, 259,
261, 262, 263, 267, 268, 271, 273,
280

Meyer, J. 135

Meyer, K. 54, 58, 61

Meyer, K. H. 172, 180, 258

Michaelis, L. 236

MlDDLEBROOK, W. R. 126, 163

Mirsky, A. E. 264
Miszurski, B. 56, 57
Mizell, L. R. 294

MOBERGER, G. 36

Moffitt, W. 195

Mohn, M. P. 135

Montagna, W. 57, 59, 60, 62, 78, 81,

82, 96, 100, 137, 210, 219, 221, 222,

228, 282
Morales, M. F. 140
Moran, T. 107
Moscona, A. 44, 84, 85
Moss, M. S. 267
Muller, C. 267
Munger, N. 180

Nageotte, J. 127

Nay, T. 151

Neurath, H. 175

Nieuwkoop, P. D. 90

Nilsson, O. 62

Niu, M. C. 276

Noback, C. R. 75, 76 (Fig. 33)

Norris, M. H. 150, 156

Odland, G. F. 219, 263
O'Donnell, I. J. 219, 235, 238, 240,

243, 244
Olofsson, B. 236
Olson, M. E. 117
Oparin, A. I. 28
Osawa, S. 264
Oster, G. 5
Ottoson, D. 90

Palade, G. E. 35, 46, 81, 111, 114,

115, 118
Parker, K. D. 199, 201, 208
Patterson, W. I. 252
Pauling, L. 128, 172, 176 et seq., 181,

183
Pautard, F. G. E. 20, 79
Peacock, N. 200
Pearse, A. G. E. 29, 54, 80
Pelc, S. R. 220, 264
Perutz, M. F. 2, 165, 182, 194, 202
Peters, L. 193, 176, 250
Peters, R. A. 114
Philip, B. 261, 267, 268
Picken, L. E. R. 24
Pillai, P. A. 83, 263

PlLLEMER, L. 119
PlNKUS, F. 77

Pinkus, H. 148

304

AUTHOR INDEX

POLICARD, A. 88
POLLISTER, A. W. 5

Porter, K. R. 39, 47, 48, 86, 90, 93,

94, 114, 168
Price, D. 135, 143

PUCHLER, H. 43

Puck, T. T. 57
Quevedo. W. C. 221

PvABINOVITZ, M. 117

Randall, J. T. 127

Ranvier, E. 95

Rashevsky, N. 140

Rawles, M. 276

Reed, R. 124

Rees, A. L. G. 126, 168, 170, 246
261

Reeve, E. C. R. 160

Reinberg, A. 137

Richards, A. G. 27

Richards, O. W. 160

Riley, D. P. 194

Robb 160

Robertson, J. D. 38, 41, 47

Robinson, C. 127, 199

Robson, A. 7, 238, 261

Roe, E. 151

Rogers, G. E. 97, 168, 170, 186, 188,
221, 226, 230, 231, 236, 237, 247, 248,
262, 268, 274, 275

Romer, A. S. 49

Rose, S. M. 92, 111, 142

Rothman, S. 59, 136, 222, 230

Rougvie, M. A. 207

Rudall, K. M. 11, 19, 20, 21, 24, 51,
67, 72, 73, 104, 106, 121, 124, 126,
137, 150, 151, 152, 157, 166, 200,
204, 205, 214, 217, 218, 231, 232,
233, 235, 248, 260, 265, 275

Rugo, H. J. 166, 167, 206, 207

Ryder, M. L. 137,150, 21 9, 221 , 232,
233

Salecker, J. 90

Saunders, J. W. 60, 61

Scheving, L. E. 135

Schmidt, W. J. 5, 10, 43, 211, 213,

226

Schmitt, F. O. 2, 100, 127, 213
Schneider, K. C. 43

Schor, P. 33, 163, 166, 169 (Table

10), 208
Schroeder, W. A. 8, 33, 180
Schulz, H. 48, 145
Scott, D. B. 78
Scott, van, E. J. 230
Seiji, M. 281
Selby, C. C. 86, 90
Seligman, A. M. 217
SetXla, K. 94
Shaw, J. T. B. 180
Short, B. F. 149, 150
Siekevitz, P. 115, 116, 117, 118
Sikorski, J. 168, 170, 225, 265
Simkin, J. L. Ill, 118
Simmonds, D. H. 7, 32, 219, 241,

275
Simpson, W. S. 225
Sjostrand, F. S. 37, 47, 48, 90, 111
Skertchley, A. 176
Skinner, B. 238, 261
Slen, S. B. 149
Smiddy, M. B. 95, 131
Smith, L. F. 163, 239
Smith, S. G. 180
Sobel, A. E. 131
Sock, N. W. 140
Speakman, J. B. 73, 172, 193, 234.

240, 249 et seq., 252, 260
Spearman, R. I. 228, 231, 274
Spelley, A. 147
Spier, H. W. 263
Stainsby, G. 127
Stanford, J. W. 131
Stell, I. G. 241
Stenstrom, S. 90
Stephenson, M. L. 118
Stoeckenius, W. 38, 47, 48
Storey, W. F. 133
Stott, E. 73
Stoves, J. L. 72, 257, 265
Strangeways, D. H. 56
Street, A. 172, 191
Strong, R. M. 104
Sutherland, G. B. B. M. 198

SVAETICHIN, G. 90
SWANBECK, G. 231

Swann, M. M. 140
Sylven, B. 221, 233
Synge, R. L. M. 234

AUTHOR INDEX

305

Taylor, H. S. 181

Thomas, J. A. 134

Thompson, D. W. 67

Thompson, E. O. P. 219, 240, 243

Thuringer, J. M. 82

Timmis, G. M. 151

Tristram, G. R. 6 (Fig. 2), 33, 161

Trotter, I. F. 180, 183

Turner, C. D. 22

Unna, P. G. 271

Vand, V. 177
Vickery, H. B. 29

VlGNEAUD, DU, V. 232

Waddington, C. H. 60, 85, 104, 143

Waldschmidt-Leitz, E. 180

Ward, W. 8, 9, 161, 162, 207, 235,

238, 243, 249, 272
Warwicker, J. O. 180, 192
Watson, J. D. 109,112
Watson, M. L. 48, 108, 112
Waugh, D. F. 128
Weibull, C. 201, 203
Weidinger, A. 214
Weinmann, J. P. 135
Weiss, L. 41

Weiss, P. 39, 60, 63, 83, 84, 86, 90, 93,

130, 138, 140
Wells, J. R. 119
Whewell, C. S. 258, 259, 279
Whiteley, H. J. 116
Wildman, A. B. 71, 76, 157, 159
Wilkins, M. H. F. 110
Willmer, E. N. 44, 57
Wilson, H. R. 110
Wiseman, A. 244
Wislocki, G. B. 84
Woernley, D. L. 286
Wolback, S. B. 276
Woodin, A. M. 163, 207, 238
Woods, E. F. 196, 219, 235, 238, 243,

244
Woods, H. J. 72, 168, 170, 172, 174,

175, 176, 183, 191, 192, 193, 246,

255, 257, 258, 259, 260
Work, T. S. 111,118

Yakel, H. L.
Young, J. Z.

183
49

Zahn, H. 252, 257, 259, 260
Zamecnik, P. C. 111,116,117,118
Zeiss, D. 180
Zubay, G. 110, 201

ZWILLING, E. 60

Subject Index

Acid mucopolysaccharides in hair

follicle 220 (Fig. 97), 221
ACTH 54
Adaptation of epidermal thickenings

66
Adenosinetriphosphate (ATP) and

mitochondria 116
Aggregation of tropocollogen 1 28 (Fig.

55)
Allometric growth 159
AllwOrden reaction 267 (Fig. Ill)
Amide bond, structure of 7 (Fig. 73),

176
Amino acid composition of low S

extract of wool 237 (Table 14)

content of fibrous proteins (silk,

collagen and wool) 6 (Table
3)
Amphibians, keratinization in 51
Antigenicity of keratins 119
Antlers 67
Arthropodin 20
Astbury-Bell model for a-keratin 176

(Fig. 72)
Avian secreted keratins 30, 107

Bacterial fiagella, protein of 2, 201
Basal layer cells, division of 82

— of keratinizing tissues 80

— membranes and epithelia 53, 90

— in electron microscope 88 (Fig. 39)
Beaks and bills 66

Benzpyrene on epidermis 275

— in skin 94

Biochemical evolution 24, 26, 49
Biochemistry of protein synthesis 115

et seq.
Birefringence 10 (Table 4)

— and stability 21

— changes on stretching hairs 2, 191

— in hair follicle 213 (Fig. 91)

— in swollen cuticle 226

— intrinsic and form 211, 213

— of Stratum corneum 230
Bombyx mori, synthesis of silk by 121

— silk 180

— molecular structure of 1 79

Bonds consolidating keratins 21, 27,

234, 245, 249
Bragg's Law 12, 13

Calcium salts and collagen 20, 21, 130
Carcinogenesis tests using mouse skin

275
Carcinogenic hydrocarbons on skin

94, 275
Cell adhesion and cell membranes 40,

84, 260
in basal layer 82, 83 et seq.

— contacts 40 et seq. (Fig. 20)

— during differentiation 84

— in hair follicle 92 (Fig. 41)

— membranes as /2-keratose 43, 260

— in keratinized tissues 238

— of basal layer cells 83

— surface 37 et seq.

Cellular adhesion during differentiation
84, 90, 99

— in cancer cells 85

— in embryo 84

— behaviour in cine films 85

Cetyl sulphonic acid extract (CSA)
244

— of wool 245 (Table 19)

Chain configuration (a or /3) from infra-
red spectra 197 et seq.

Chemical difference between o- and
p-type keratin 275

— modifications during keratinization

(other than cystine formation) 219

— reactivity and chemical composition

of keratins 161, 162 (Table 7)

— structure of keratins, present status

161 et seq.
Chitin 20, 24
Chlorine peroxide as disulphide bond

oxidant 240
Chrysopa egg stalk 201, 204
Cilia and fiagella 39
Citrulline in inner root sheath 226

— in medulla 270

— in trichohyalin 97
Classification of hairs 72

300

SUBJECT INDEX

307

Claws 69

— cutting edge and uneven keratini-

zation 273
" Club " hair 89 (Fig. 40)
Coiled coils and a-helices 183, 184

(Fig. 77)
Collagen 2, 16, 19, 31, 53

— appearance in electron microscope

19, 21, 168

— as a secreted fibre 121

— fibre formation of 127
heat contraction of 256

— in insects 19

— meshworks 1 30

— molecular structure of 127
Colleterial glands in insects, proteins

of 5, 204

— structure of 205 (Fig. 87)
Comparative cytology of cells forming

protein 113 (Fig. 50)

Competition, as control factor in
growth 149, 159

Components of keratinized tissue re-
vealed by chemical extraction 271

Composition and properties of keratoses
239 (Table 15)

— of rabbit fur and medulla 270
(Table 20)

Concentric membranes, whorls, my-
elinic forms in cytoplasm 46

Configuration of polypeptides in syn-
thesis 119

Contact inhibition of cell movement
85

Control diagram of hair growth (Chase)
152, 153 (Fig. 63)

— of growth by inhibitors 140 et seq.
Cortex of hair 71,72,101

electron microscopy of 101, 211

et seq., 223
Cortical cells and elasticity of hairs 174

— in hair follicle 4, 223

— of hair and wool 72

— of wool, extensibility 174
Cortisone 54

Crimp, and bilateral structure of wool
273

— formation, theory of 157, 158

(Figs. 66, 67)

Crimp, in wool 156 et seq.
Cross-linkages of protein(s) 19 et seq.

— reformation by alkylation 3, 252
Cross-linking of epidermin (prekeratin)

235
Cross /3-pattern 200 et seq. (Fig. 84)

— explanation of 201

— from bacterial flagella 202

— from insect egg stalks 201

— from muscle 201
Crystalline/amorphous ratio 191
Crystallites, stabilizing effect of 19,

255

— and supercontraction 256

— and X-ray diffraction 12, 256
Crystallographic analysis of keratins

176 et seq.
Curls from curved follicles 156
Cuticle cells, in keratinization 224

of hair 72

fine structure of 225 (Plate 20A)

action of osmium tetroxide 225

— of hair 71,72,98,99,265

— extensibility of 265

— structure 266 (Fig. 110)
Cuticular keratin, characteristics 101,

225, 265
Cybernetics 145, 157, 158, 159
Cycles in mouse and human skin 89

(Fig. 40), 96, 136
Cyclic activity, generated by feed-back,
*141, 145 (Fig. 60)

— changes in vaginal epithelium 135,

— 136

— growth (rhythms) in feather follicle,
106

Cyclostomes, and mucinogenic glands
59

— teeth of 75

Cystine 21, 29, 214 et seq., 230

— cross-linkages, location of 234, 245

et seq., 252
absent from crystallites 246

— formation in follicle 232, 245

— in cuticular keratin 226

— reaction with osmium tetroxide
35, 36

Cysts, keratinized 275
Cytology, of cells synthesizing protein
108 etseq., 110 (Fig. 49)

308

SUBJECT INDEX

Cytology of melanocyte 280
Cytoplasmic structures 45 et seq.
Cytoskeleton 114

Degrees of keratinization, 64, 68, 272

— method to estimate 272
Density and crystallinity 194
Dental keratin 78, 108
Dentritic cells, 277, 278

Dermal changes during hair growth

152
Dermis 53

— configuration of 55, 81
Dermoepidermal junction 53, 86, 96

— formation of 88
Desmosomes (dermo-epidermal) 41,

42 (Fig. 21), 93

— composition 84

— in light microscope 83

— of basal layer cells 83
Desoxyribonucleic acid (DNA) 54,

80, 109, 177

— in hair follicle 219, 220 (Fig. 97)
Desquamation of epidermal cells 22,

134, 231
Deuterium exchange 200
Development of stability in hair follicle

214, 215 (Fig. 92), 216 (Figs. 93

and 94)
Dichroic ratios and amorphous phase

198, 199

— after deuterium substitution 199

— in infra-red 196

Dichroism of infra-red absorption and
structure of a-keratin 176, 196 et
seq.

Differentia in cytoplasm 48

Differentiated epidermal tissues 94
et seq.

Differentiation 44

— of epidermal cells 60, 90 et seq.

— of surface organelles 44
Difficulties in defining keratins 30

et seq.
Diffusion of pigment from grafts 278
Digital tips 69
Directional friction, and the hair

cuticle 73

— and felting 73

Disintegrated wool, electron microscopy
of 246

Dispersion of optical rotation and chain

configuration 195, 198
Distribution of disulphide groups in

hair follicle 218, 220 (Fig. 97)

— of fibre types in vertebrates 23,
25 (Fig. 12)

— of fundamental fibre types 22
et seq.

Disulphide bonds, and elasticity 252

— and strength of wool fibres 254
(Fig. 107)

— content of stratum corneum 230
Diurnal cycle(s) 136

— absence of in hair follicle 137

— of mitosis in epidermis 148
Dopa 280

Drude's equation 195

Earthworm cuticle 124,130
Ecdysis 22, 160

— horns and scales 67
Egg case of mantids 126

— shell membranes 107 (Plate 18B)
Elastic behaviour of wool and hair 172

— properties, and the structure of
hair 172 et seq.

— of wool and humidity 250 (Fig.

103)

and temperature 251 (Fig. 104)

and pH 253 (Fig. 105)

Electron microscopy and cytology
34 et seq.

— methods 35

— of cuticle 265 (Plate 20A)

— of fibrous keratin 246 (Plates 12,

13, 14A, 15, 16)

— of skin 94, 223 (Plates, 4B, 7, 8,

9, 17, 22)
" Eleidin " 95 (footnote)
Embryonic cells, generalized structure

of 81, 82
Enamel of teeth, keratin in 78, 108
End groups in keratins 161, 162

(Table 8), 163

— methods 163
Endocuticle of hair 265
Endoplasmic reticulum 46, 114

— nomenclature 48 (Fig. 23)
Entropy and fibre extension 257

— and molecular configuration 255

SUBJECT INDEX

309

Enzymatic digestion of cuticle 266
(Fig. 110)

— of hair 271

— of reduced wool 254
Epicuticle of hairs 265, 267
Epidermal family of cells 57 et seq.

(Fig. 26)

— fibrils, organization of 131, 132,
229 (Plate 17)

— glands 57 et seq.

— growth 133 et seq.

control of 146 et seq.

effect of inhibitors 147

stimulated by stripping, blood

supply and Thorium-X radi-
ation 149

— keratin, fine structure of 228 et seq.

(Plate 17)

— pigments 276 (Table 21)

— thickenings 66,134
Epidermin and reduced keratins 238

— preparation and properties 235

— sedimentation analysis of 236
Epidermis 55 (Fig. 25)

— fine structure of 94

of keratinizing layers 228,

229 (Fig. 99)

— idealized form 56

— local control of morphogenesis 96,

146
Epithelia, differentiation of 90 et seq.

— nutrition of 55, 56
Ergastoplasm 46, 48 (Table 23) (Plate

10 A)
Exocuticle of hair 265
Extensibility of feather 24, 257

— of hairs 174,175

— — molecular basis 172, 175
Extracellular keratins 30, 107

Feather(s) 69, 70 (Fig. 30)

— amino acids in 8 (Table 2)

— and scales, homologies 69

— cells, fine structure 104 (Plate 1 2 A)

— cyclic polypeptides 163

— extracts 238, 243

— fibrils 224

— follicle 101 et seq., 102 (Fig. 45),
103 (Fig. 46), 105 (Fig. 47)

— keratin 33, 34

Feather(s), keratin, evolution 24, 25

solutions 207

structure of 205 et seq., 207

(Fig. 88), 209 (Fig. 89)
X-ray diffraction data 169

(Table 10), 171 (Table 11)

— monomeric unit of 163, 164

— surface cells 73

— thiol groups in follicle of 218

(Fig. 96)
Feed-back, and oscillations 141, 142

— control of growth 140

— control in follicle 152, 153 (Fig.

63), 158, 159

— controls 141

— intracellular 141
Felting 73

Feulgen test 54, 220 (Fig. 97b)
Fibre forming properties of extracted

keratins 244
Fibrillar structures of epidermis 230

(Plates 17 and 22)
Fibrils of fibrous keratin 187

— formation and X-ray pattern 223

— internal structure 187, 188 (Fig.

81) (Plates 15 and 16)
Fibrinogen 121
Fibrinogen-fibrin system 126
Fibroblasts in dermis 92 (Plate 23 B)
Fibrogenesis 125 et seq.
Fibrous texture 123
Filament-plus-matrix model 193, 247

(Fig. 102), 248
Filaments, and fibrils in hair follicle

223 (Plates 13, 15 and 16)

— in a-proteins found electron micro-

scopically, 185
a-Filaments 183

— and a-fibrils, definition 186 (Fig.

80), 187

— characteristics 248

— structure 186

organization into larger units 187

Fine histology of hair follicle 95, 223
et seq.

— structure of cells 34 et seq.
Flagella of bacteria (X-ray pattern)

187, 202 (Fig. 86), 203
Flagellin 203

310

SUBJECT INDEX

Fleece mosaics 147

Fluorescent dyes and keratinization

228, 231
Folding of polypeptide chains 172

— elasticity of 175
Follicular activity 89, 96, 151

— nutrition 232

Gizzard, keratin of 107, 108

— of birds 30, Plates 18 and 19
Glandular cells, cytology of 109 et seq.,

110 (Fig. 49), (Plate 10A)
Glycogen in epidermis 136, 221

— in outer root-sheath of hair follicle

137, 221

and growth 221

Golgi apparatus 45 (Fig. 22e), 46, 115

— in basal layer cells 81, 82

— in cuticle cells in follicle 224
Grafting of epidermal tissues 60 et seq.

— of pigmented skin 278

— of skin, effect on hair growth 152,

155
Granular layer of epidermis 95 et seq.,

228 et seq. (Plates 22 and 23A)
Growth, factors influencing 1 34 et seq.

— general theories of 138 et seq.

— limiting factors in 140

— of epidermal structures 133

— theory of Weiss and Kavenau 140

et seq. (Fig. 57 and 58)

— waves, explanation in terms of

accumulation of inhibitors 154

(Fig. 64)

as graded response to stimulant

155
in hair 151 et seq.

Hair(s) 71 et seq. (Fig. 31)

— cuticle, development of 98 (Fig. 43)
fine structure in follicle 224

(Plate 20A)
structure of 265 et seq.

— cycles 96, 1 50 et seq.
in rat 89 (Fig. 40)

— elastic properties 249 et seq. (Figs.

103-107)

— follicle 72

diagrams of histochemical re-
actions in 211 et seq.

Hair(s), follicle, fine structure 95 et
seq., 223 etseq. (Plates 11, 12, 13,
14A, 15)

groups 75 et seq. (Figs. 33-35)

in tissue culture 96

— human, long spacings from 167
— ■ waving treatments 254

Heat-moisture treatment (Bear and
Rugo) 167

— of feather 206
Helical, aggregation 128

— configurations, the a-helix 181

— content by optical methods 194

et seq.

of various proteins 195

of wool extracts 196

— structures 176 et seq.
a-Helices 181 et seq. (Figs. 75, 76)

— extensibility 188, 189 (Fig. 82)

— a-filament and electron microscopy

185, 186 (Fig. 79)

— in synthetic polypeptides 183
/?-Helices 209 (Fig. 89)

Henle's Layer during hair growth 224

— in hair follicle 96, 97 (Fig. 42),

226 (Plates 20 and 21)

— properties of 227

Histology of hair and elastic behaviour
174

— of vertebrate epidermis 53 et seq.,

80 et seq., 228 et seq.
Holocrine glands 59
Homogenization of tissues 115 et seq.
Hoofs 69, 273
Hooke's law region in stress-strain

curve of hair 173
Horns 67, 68 (Fig. 29)

— and size of animal 160

— keratinization 2, 217, 231 (Fig. 95)
Huxley layer in hair follicle 97 (Fig.

42), 100
Hyaluronidase 54
Hydrogen bonds, and elastic properties

254

— and infra-red spectra 197

— and protein structure 178

— stabilization by 234
Hydroxyproline, in collagens 6

— in tooth enamel keratin 78

SUBJECT INDEX

311

Ichthyosis vulgaris 64
Inaccessibility and crystallinity 191
Infra-red spectra, and chain configur-
ation 197, 198 (Table 12)

— and structure 196

— of feather 208

— of keratins 197 (Fig. 83)
Inhibitors of growth 140 et seq.

— as hormones 143

— and growth hormones as mitotic
controls 148 (Fig. 61), 149

Inner root sheath of hair follicle, fine
structure of 95 et seq., 226 et seq.

Insect proteins (a and /? types) 203
et seq.

Insulin (fibrous) 29, 128

Interaction between dermis and epi-
dermis 61 (Fig. 27), 86 et seq., 87
(Fig. 38), 88 (Fig. 39)

Intercellular adhesion 40, 44, 81, 84
et seq., 90 et seq., 99 et seq., 98 (Fig.
43), 99 (Fig. 44)

— — and differentiation 60, 61, 90,

— apparatus during keratinization 263

— " bridges " (desmosomes) 42, 83,

94, 260

— cements 85

— in electron microscope 86

— contacts in hair follicle 99 (Fig.

44), 100
Interdigitation of membranes 43, 263
Interfollicular competition in sheep

149, 150
Internal energy and entropy 257
Intracellular differentiation in hair

bulb 100
Invertebrates, keratins of 30

— tanning in 27

Kerateines and keratoses 241 et seq.,
244

compositions compared 245

(Table 19)
Keratin, A and B (Unna) 2, 271

— chemical analyses of 6 (Fig. 2),
7 (Table 1) 8 (Table 2), 9 (Table
3), 161 (Tables 7, 8), 241 (Fig, 100),
242 (Table 17), 243 (Table 18), 245
(Table 19)

Keratin definition of 2, 21, 27 et seq.

— hard 64 et seq., 210

— hard and soft 64 et seq., 65 (Table

5), 134, 228, 272 et seq.
Keratin, hard and soft, in claws and
hoofs 69, 273

— histological tests for 29

— in hair cuticle cells 101 (Plate 20),

265

— in tooth enamel 78, 108

— " pearls " 275

a-Keratin pattern (Macarthur) 166
(Fig. 68)

— wide angle 15, 170 et seq.

— characteristic 1*5 A spacing of 15,

176, 182
/9-Keratin, as a pleated sheet 179, 180

— in feather follicle 69, 104

— in supercontracted fibres 258
Keratinization in bird skin 63, 231

— ■ in curved follicles 157

— in skin and scales of rodents 231

— in hair follicle 210 et seq.

— not a degenerative phenomenon 56

— process 210 et seq.

— — definition of effects 210

— zone 55 et seq. (Fig. 25), 56, 157

(Fig. 66), 211, 223

Keratinized cysts and epidermal tu-
mours 6, 275

Keratohyalin 65 (Table 5), 95, 228
et seq.

— and trichohyalin 226, 227

— conversion into fibrils 230 et seq.

(Plate 21)

— granules in stratum granulosum 228
Keratoses (a and y) 237, 238, 240

(Table 16)
k, m, f . . . proteins 15, 19
Krebs cycle intermediates and mitosis

136

Lanthionine 237

Levels of organization 3, 4 (Fig. 1), 10,

123 etseq., 125 (Fig. 54)
Lillie and Juhn's hypothesis 105

(Fig. 47)
Limulus 30

312

SUBJECT INDEX

Lipids in keratinizing tissues 59, 222
Liver, growth after hepatectomy 140
Location of cystine residues 245 et seq.
Logistic law of growth 139
Long spacings 11, 165 et seq.

— and short spacings, independence of
166

— in electron microscope 19, 167

— theories of origin 170

Low angle diffraction patterns 11,
165 et seq.,

— of keratins 166 (Fig. 68), 168

(Fig. 69)

Macrofibrils in fibrous keratin 123,

125, 187, 186 (Fig. 80), 224 (Plates

13, 14A, 15, 16)
Macromolecules and biology 1
Malpighian layers and cells 57, 80

et seq.
Mammary glands 59
Matrix, in electron micrographs 223,

224

— of fibrous keratin (y-component)

248

— plus filament model and X-ray

scattering 4, 193

— protein (y-keratin) as interfila-

mental cement 188, 223, 246
et seq.
Mechanical trauma as growth stimulant

149
Mechanochemical phenomena 249 et

seq.
Medulla of hair 71, 73, 97, 268 et seq.
Melanin 276 et seq.
Melanization, chemistry of 279, 280,

281
Melanocyte(s) 27

— activity and hair cycle 282

— and keratinizing cells 278

— and pigmentation 276 et seq.

— in hair follicle 97 (Fig. 42), 100
Membraneous systems of cytoplasm,

definition 46

— nomenclature of 47, 48 (Fig. 23)
(Plates 4, 5, 6, 7, 10)

Membrane(s) adhesion and morpho-
genesis 84 et seq.

in hair follicle 97 (Fig. 42),

98 (Fig. 43), 99

Membrane(s) as tubules derived from
cuticle of hair and wool 267 et seq.

— chemical character 261

— enclosed particulates 45

— in keratinized tissues 260

— morphology 262

Metabolic enzymes in keratinizing

tissues 221
Metachromasy 54, 221
Metaplasia 62, 63, 233

— cyclic 144 (Fig. 59)
Microbodies in keratinizing tissues

233 (Plates 22, 23 A)
Microsomes 1 1 1 et seq.
Mitochondria, definition 45

— in basal layer cells 81
Mitochondrial enzymes 115, 116, 221.
Mitosis in basal layer 82, 135 et seq.
Mitotic apparatus 30
Modulations 63

Molecular, and macromolecular struc-
tures 4, 119, 123, 161

— configuration in supercontracted
state 258 et seq.

— macromolecular and histological

units of fibrous keratin 186
(Fig. 80)

— structure of degraded keratins 163

et seq.
Mouse, diurnal cycles in epidermis
136

— ear epidermis, mitosis control in

149

— skin 94

Mucinogenic cells, cyclic changes in

145
Mucopolysaccharides 54, 58, 61, 84,

221
Muscle, X-ray pattern of 15, 168
Myelinic forms 46
Myosin (actomyosin) 15, 121, 168

Nails, nature of 69

— thiol reaction in 217 (Fig. 95)
Neurokeratin 79
Non-crystalline fraction 190 et seq.

— molecular structure in 192

— in silk 192
Non-keratinous residues 270 et seq.

— as ^-keratoses 238

SUBJECT INDEX

313

Non-keratinous residues, morphology of

266 (Fig. 110),268(Fig. 112)
Nuclear membrane 48,112

— in keratin cells (Plate 4C), 80, 81

(Fig. 36)

— pores 48, 112 (Plate 4C)
Nuclei, fate of in keratinizing tissues

4,263
Nucleic acids 109

— distribution in hair follicle 112,

219, 220 (Fig. 97)
Nucleoli 80, 81 (Fig. 36), (Plates 7, 8)
Nucleus 48, 80, 81 (Plates 7, 8, 12, 13)

— and protein synthesis 109 et seq.,

110 (Fig. 49), 264

— pores Plate 4C

Oestrogens and cell division in epider-
mis 134

Optical rotation and chain configuration
194 et seq.

Orders of magnitude 3, 4 (Table 1)

Organization of fibrous tissues 129
et seq.

Orientation in hair follicle by X-ray
diffraction 214

origin of 131

— of fibrils in epidermis 231 (Plates

17, 22)

— of fibrous keratin in hair follicle

210 (Fig. 90), 211
Ovokeratin 30, 107 (Plate 18B)
Ortho- and paracortex of wool fibres

268 (Fig. 112), 273
Osmium tetroxide, fixation by 35, 36

(Fig. 17), 112

— reaction with cystine 35, 36, 247

(Fig. 102)
Osmoregulation and cuticles 50 (Fig.

24)
Oxidation of keratins to keratoses, etc.

238 et seq.

— as histochemical test 29

Particle covered membranes (endo-
plasmic reticulum) 46

— in secretory cells 111 (Plate 10A)
PAS test 54

— on hair follicle 220 (Fig. 97), 221

Patterns of hair growth and control

150 et seq.
Peptides from silk 180

— structures of 7, 176

Peracetic acid as reagent for oxidizing

keratins 238
Periodic changes in plumage and pelt

144

— growth, and cyclic activity 143 et seq.

— of feathers 1 56 et seq.
Peritrophic membrane 124
Phenolases 280
Pheomelanin 276, 282
Phosphatases 222

Phospholipid membranes 35, 46, 49
et seq.

— myelinic forms 46

Phylogeny, and fibre type 22, 25
(Fig. 13)

— and keratinization 22, 49 et seq.

— of hair 72, 73

Physicochemical properties and keratin-
ization 249 et seq.

Pigment granules 276 et seq.
■ — chemical stability of 278

— internal structure of size 279 (Fig.

115)
Pigmentation 276
Pilosebaceous unit 60, 77
Plasma membrane 37 et seq.

— definition of 41
Platypus hair 67, 77

Pleated sheet configurations 179, 180

(Fig. 74)
Polymerization of keratin 234
Polypeptides, synthetic and a-structure

183
Porcupine quill, and 1*5 A spacing 182

— fine structure 247, 248

— X-ray pattern of 167 (Table 9)
Precursors of keratin 126,211
Prekeratin, properties of in germinal

layers 211 et seq.
Prekeratins 235

Prickle cells and desmosomes 94
Primary and seconday hair follicles 72,

75, 78 (Fig. 35), 150
Proliferating tissues 133

314

SUBJECT INDEX

Proline, and chain configuration 193,
196

— in feathers 163
Pronghorns 67

Prosthetic groups, absent in keratins

161
Protein(s), corpuscular and structural

— fibre synthesis, classification of 1 22

(Fig. 52)
- — • synthesis, cytology of 108, 116,
120, 121

in epidermal systems 108 et seq.,

131

in hair cortex 112

synthesis theories of 109 et seq.

a-proteins other than keratins 202

et seq.
Proteus vulgaris and Bacillus subtilis,

X-ray pattern of flagella 3, 202
Protofibrils 125
Pycnosis of nuclei 4, 263

Radio sulphur, incorporation into nuclei

in keratinizing cells 220
Rana, tail epithelium 90 et seq.
Rat and mouse hair, zig-zags in 156
Reduction of wool 236 et seq.
Regenerated keratins 166, 237, 238
Relation between a-helices, a-filaments

and fibrils 186 (Figs. 79, 80)
Relaxation of hairs and wool fibres 255
Renewal time(s) 1 34

— for epidermal tissues 135 (Table 6)
Residues remaining after chemical

extraction of keratins 270
Resistant membranes in keratinized

tissues 262, 270
Retaining cells, definition 112, 114,

120

— comparison with secreting cells

113 (Fig. 50)
Reticulin 53

Rhythmic activity, generated by feed-
back 141, 145 (Fig. 60)

— changes in shape and composition
of crimped wool fibres 156

Ribonucleic acid (RNA), and protein
synthesis 109 et seq.

Ribonucleic acid granules, fate of in
keratinizing tissues 264

— in basal layer cells 80, 81
as ribosomes 81

— in hair follicle 219, 220 (Fig. 97)
Ribosomes (RNA) (particles) 46, 111,

112

— in basal layer cells 81

— in hair cortex 112, (Plate 11)
Rubber-like properties, of hairs 255,

256

— of reduced hairs 258

Salt bridges, stabilization by 234

— linkages and elasticity 251, 253
(Fig. 106)

Scale-hair-gland complex 77 (Fig. 34),

78
Scale(s) 66, 67 (Fig. 28)

— embryology of 67

— of fish 66, 75

— of snakes 106

- — (reptilian) structure 67 (Fig. 28)

— types of keratin in 67, 104
Schmitt's hypothesis for development

of columnar epithelia 99 (Fig. 44),

100
Sebaceous glands 59
Secondary and tertiary structures in

proteins 119
Secreted keratins in birds 79, 107

(Plates 18 and 19)
Secretion of proteins 112 et seq.
Secretory cells, cytology of 110 (Fig.

49), 113, 114
Setting of hairs 259, 260
Seven-stranded cable for a-proteins 5,

184, (Fig. 78)
Shrinkage of woollens 73
Sickle tips of primary fibres 150, 159
Silk and /?-keratin patterns compared

172

— fibroin 121

— formation of fibrils 129

— structure of 179, 180

— glands 203

Soft keratinization 64, 95, 97, 228
et seq.

Soluble derivatives of keratinized tis-
sues 236 et seq.

SUBJECT INDEX

315

Soluble products, by oxidation of
keratins 238 et seq.

— by reduction of keratins 240 et seq.

— of partial keratinization 233
Specializations of opposed surfaces 40

et seq.
Specialized appendages of epidermis

66 et seq.
Specific rotation (definition) 195
Spherulites of insulin 130
Spirals and whorls of fibrils in hair

cortex 224 (Plate 14 A)
Stability, histochemical tests for in hair

follicle 214 et seq.
Stabilization 19 et seq., 24, 51, 241

et seq.
Stratum corneum, constituents of 228

— structure of 225 (Fig. 98), (Plate

17), 229 (Fig. 99)

— granulosum 95, 228, 229 (Fig.
99) (Plate 23A)

— hicidum 95, 230 (Plate 22)
Stress-strain curve(s), for hair 172

et seq., 249 et seq.

— for wool, interpretation of 193

— interpretation of by Astbury and

Woods 174
Stripping of epidermis as growth

stimulant 148
Structural colours 276
Structure of residues remaining after

extraction of keratins 271
a-Structure, structural proposals 175
Sulphide solutions of wool 236
Sulphur, content of keratoses 239

(Table 15)

— content of skin 230

and nuclear synthesis 264

— level of entry into follicle 232 et seq.
Sulphydril reagent (Bennett) 29

— - reactions in, hair follicle 220 (Fig.
97c)

claws and horn 217 (Fig. 95)

Supercontraction 255-8

— in copper solutions 259

— in lithium bromide 259

— in phenol, etc. 259
Supermolecular organization of fibrous

tissues 123
Surface " repertoire " of cells 62

Surface specializations of cells and
tissue formation 26 (Fig. 14), 86,
87 (Fig. 38), 90 et seq., 97 et seq.

Symphysodon discus and mucin secretion
60

Synthesis of fibres, steps in 125 (Fig.
54)

cytoplasmic pattern of 121

— ■ keratin in hair bulb 101 et seq.,
223 etseq. (Plates 11, 12)

— melanins and melanoproteins 281

(Fig. 116)

— peptide bonds 116 et seq.

— proteins 93 et seq., 115 et seq.

in fibre forming systems 121

in hair follicle 96 et seq. (Fig.

42)

in retaining cells 1 20 et seq.

in secreting cells 116

Synthetic polypeptides, compared with
a-proteins 183

— a-/? transformation in 199

Tactoids and fibres 130
Tanning 19 et seq., 27

— and resistant membranes in keratin-

ized tissues 262
Teeth and hair 74

— development of 74 (Fig. 32)
keratin in enamel of 78, 108

Theories of growth 138 et seq.
Thickness of epidermis 66,134
Thioglycollic acid as reducing agent

236, 240
Thiol (SH), and disulphide groups in

hair follicle 217 et seq., 220 (Fig.

97c)

— reactions, in horns, nails and claws

217 (Fig. 95)

in feather 218 (Fig. 96)

Thyroxin, metamorphosis and keratin-
ization 52
Tissue culture, and differentiation 44

— keratinization in 56, 57, 63, 64

— types of cells in 57
Tonofibrils 43, 94

" TPA " stain, after Leblond et al.

43
Transference of pigment to keratinizing
cells 278

316

SUBJECT INDEX

oc-/3 transformation, and fibre extension
172, 173

— and the a-helix 1 88

a-/? transformation, proposals of
Ambrose, Huggins, Hanby and Zahn
176

Trichohyalin 97, 226 et seq., 230

— an a-protein 230

— fine structure of 226

— in inner root sheath 100

— properties of 226 et seq.

— transformation in to fibrils 100

(Plate 21), 226
Trio groups of hair follicles 76 (Fig.

33), 78 (Fig. 35)
Triturus 51

Tropocollagen and collagen 127
Trypsin, action of on hair follicle 217
Tryptic digestion of hair 260
Tumours, epidermal 62, 63, 274
Tussah silk 180
Types of fibrous proteins 5

— of fibrous texture 124 (Fig. 53)

— of keratin in feather follicle 106
(Fig. 48)

Tyrosine, as precursor of melanin 280

— residues in fibroin 180

Uneven keratinization and its histology
272 et seq.

— and function of tissue 273
Unguis and subunguis 68, 273
Unoriented a-helices, X-ray pattern

from 194
Urea, action of saturated solution on
hair follicle 216 (Fig. 93), 217

— as solvent for prekeratins 235

— swelling of follicle by 216 (Fig. 94),

235
Uropygial gland in birds 59

Variable amino acid composition of

keratins 31 et seq.
— of wool 32 (Fig. 15)
Vascular supply to follicle 152, 232
Vertebrates, tanning in 27

Vitamin A, action on hard keratin in rat
scales 231

— up metaplaasia of epidermal trans-

plants 63 et seq.

— D 59, 60

Warts and corns 64, 66
Water content of hair bulb 222
Wide angle diffraction patterns 11,
13 et seq.

— interpretation of 170 et seq.
Wool, amino acids in 7 (Table 1), 9

(Table 3)

— fibre follicle and asymmetrical fibre
274 (Fig. 114)

— follicle and crimp 157, 158 (Fig.

66, 67), 273

— root extract 237 (Table 14)
Work index (30%), definition of 250

— to stretch wool fibres as function
ofpH 253 (Fig. 106)

Wound " hormones " 149

X-ray diffraction, and molecular struc-
ture of keratins 164 et seq.

— methods 1 1 (Fig. 3), 24, 165 et seq.

— patterns, of fibres 13 et seq., 164

et seq.

idealized type 13 (Fig. 5)

of supercontracted fibres 256

(Fig. 109), 257

— and chemical reactions 193

— of bacterial flagella 2 (Fig. 86),

201

— of collagen 16 (Fig. 8)

— of cuticle 265

— of egg case protein 204 (Plate 3)

— of feather 16 (Plate 2A), 169, 205

— of scales, etc. 206

— oc-type 14 (Fig. 6), 17 (Fig. 9)

(Plate 1A, 3), 165 et seq.
— /9-type 15 (Fig. 7), 18 (Fig. 10),
(Plate IB), 205

— study of hair follicle 214 (Fig.
91)

Xenopus, tail epithelium 87 (Fig. 38),
90 et seq.

Zig-zags, curls and crimps 156 et seq.

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Keratin and keratinization; an essay in molecular biology"

See other formats