THE ANATOMY OF THE CELL ENVELOPE OF A MARINE VIBRIO
EXAMINED BY FREEZE-ETCH I NG

By

JACK T. CRAWFORD

A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL
FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA
1973

ACKNOWLEDGMENTS

I wish to express my appreciation to the following people whose
efforts made this study possible.

I wish to thank Dr. Max E. Tyler for his guidance, encouragement,
and patience during the course of this study.

I wish to thank Dr. Henry C. Aldrich for his considerable efforts
in my training in all aspects of electron microscopy, for the use of
the facilities of the Biological Ul trastructure Laboratory, and for his
encouragement and friendship.

I wish to thank Dr. Arnold S. Bleiv/eis for serving on my advisory
convnittee and for the use of his facilities in preparing the amino acid
anal yses.

I wish to thank Dr. Daniel Billen for serving on my advisory
commi ttee.

I wish to especially thank my wife, Kay, for her unfailing under-
standing and patience.

I I

ACKNOWLEDGHENTS , . .
LrST OF TABLES . . ,
LIST OF FIGURES , . ,
KEY TO SYMBOLS . . .

ABSTRACT

INTRODUCTION . . . ,

MATERIALS AND METHODS

RESULTS

DISCUSSION

LITERATURE CITED . .

BIOGRAPHICAL SKETCH .

TABLE OF CONTENTS

Page

' ' ° ii

iv

* ' V

ix

• ' X

1

13

25

89

93

105

1 1 1

LIST OF TABLES

Table
].
2.

Amino acid analysis of purified peptidoglycan

Determination of the specificity of antibody
label 1 i ng

Page
29

78

IV

LIST OF FIGURES

Figure Page

1. Longitudinal and cross sections of cells

fixed in gl utaraldehyde and OsOr 27

2. Cross section of a cell shov;ing the double-

track appearance of the membranes 27

3. Cross section of a cell showing an extra layer

inside the cytoplasmic membrane 27

k. Purified trypsi n-treated peptidoglycan negatively

stained v;i th potassium phosphotungstate .... 31

5. Purified LPS positively stained with uranyl

acetate Jk

6. Purified LPS positively stained with uranyl

acetate 3^

7. Freeze-etched cells suspended in complete

salts without glycerol 37

8. Freeze-etched cell suspended in complete

salts, showing the fracture face of

the cytoplasmic membrane 39

9. Unetched convex fracture of cells without

glycerol showing the smooth fracture

face of the outer membrane 39

10. Deep etched cell in complete salts shov;ing

the outer surface of the cell revealed

by etching k]

11. Freeze-etched cell showing the smooth

convex fracture face of the outer

membrane and the cell surface k]

12. Freeze-etched cell showing the concave

fracture face of the cytoplasmic

membrane ^3

Figure Page

13. Freeze-etched cell shov/ing the smooth

concave fracture face of the outer

membrane 43

14. Concave fracture of a cell showing a

surface with distinct subunit structure .... i+3

15. Freeze-etched cell showing a paracrystal 1 I ne

array inside the cytoplasmic membrane 46

16. Freeze-etched glycerol -treated cell showing

the concave fracture face of the

cytoplasmic membrane 48

17. Freeze-etched glycerol -treated cell showing

the convex fracture face of the

cytoplasmic membrane 48

18. Freeze-etched glycerol-treated cell showing

the smooth concave fracture face of

the outer membrane 50

19. Convex fracture of a cell showing the smooth

fracture face of the outer membrane 50

20. Freeze-etched glycerol-treated cell showing

the rough convex fracture face of the

rigid layer 52

21. Freeze-etched glycerol-treated cell showing

the rough convex fracture face of the

rigid layer 52

22. Freeze-etched glycerol-treated cell showing

the concave face of the globular layer 54

23. Freeze-etched glycerol-treated cell showing

the concave face of the globular layer 54

24. Concave fracture of a glycerol-treated cell

in which the globular layer is

incomplete 57

25. Cross fractured cell envelopes of three

cells showing the edges of the

cytoplasmic membrane, rigid layer, and

outer membrane 57

26. Freeze-etched cell shov-^ing the cross

fractured cell wall 57

VI

Figure Page

27. Complementary surfaces of a glycerol-

treated eel] observed by double-replica

technique, showing the fracture faces

of the cytoplasmic membrane 59

28. Complementary fracture faces of a glycerol -

treated cell shovjing the concave globular

layer and the convex rough layer 6I

29. Complementary fracture faces of a glycerol-

treated cell showing the smooth fracture

faces of the outer membrane 63

30. Thin section of Isolated cell envelopes prepared

by lysing bacteria in a French pressure cell . . 66

31. Thin section of Isolated cell envelopes prepared

by lysing bacteria in a French pressure cell . . 66

32. Isolated cell envelopes positively stained with

uranyl acetate 66

33. Freeze-etched isolated cell envelopes showing

the convex fracture face of the cytoplasmic

membrane 68

3^. Freeze-etched isolated cell envelopes shovjjng

the surface exposed by etching 68

35. Freeze-etched Isolated cell envelopes showing

the concave globular layer 68

36. Freeze-etched Isolated cell envelopes showing

the globular layer exposed by etching

alone 68

37. Freeze-etched cell walls. The walls were

prepared by lysing cells with Triton

x-100 71

38. Freeze-etched cell walls 71

39. Freeze-etched cell v;alls 71

^0. Freeze-etched cell walls 71

k] . Freeze-etched cell walls digested with

lysozyme, showing the fibrous concave

etch surface 7^

VI I

Figure Page

k2. Freeze-etched cell vjalls digested with

lysozyrne, shov;ing the fibrous concave

surface and the globular layer 7^

43. Freeze-etched cell walls digested with

lysozyrne, showing the fibrous concave

surface emerging from the ice, indicating

that it is an etch surface yk

kk. Freeze-etched isolated outer mer.brane material

showing the concave and convex smooth f'-acture

faces of the vesicles 76

kS, Fractured outer membrane material showing the

fracture face and the unfractured surface ... 76

US. Thin sectioned cell labelled with ferritin-

conjugated antiserum. This cell was fixed

following the antibody treatment 80

ky . Thin sectioned cell labelled with ferritin-
conjugated antiserum. This cell was
fixed in gl utaraldehyde prior to anti-
body treatment 8O

hS. Freeze-etched unfixed cell labelled with

ferr i t i n-conj ugated antiserum 83

kS. Freeze-etched unfixed cell labelled with

ferr i t i n-conj ugated antiserum 83

50. Freeze-etched unfixed cell labelled with

ferr i ti n-conj ugated antiserum showing

the fracture face of the outer

membrane 85

51. Freeze-etched unfixed cell labelled with

ferr i ti n-conj ugated antiserum showing

the fracture face of the outer

membrane 85

52. Freeze-etched glutaraldehyde fixed

ferritin labelled cell showing the

relationship of the smooth fracture face

and the ferritin labelled surface 87

53. Diagramatic representation of the cell

envelope of the marine vibrio as

determined by f reeze-etchi ng. The

location of the fracture planes are

illustrated 9^

VI I I

KEY TO SYMBOLS

CM cytoplasmic membrane

0 outer membrane

CM cytoplasmic membrane fracture face, convex

CM cytoplasmic membrane fracture face, concave

R rigid layer fracture face, convex

G globular layer fracture face, concave

0 outer membrane fracture face, convex

0 outer membrane fracture face, concave

0 sur actual outer surface of the outer membrane

0 sur actual inner surface of the outer membrane

cyto cytoplasm

^ direction of shadow

IX

Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy

THE ANATOMY OF THE CELL ENVELOPE OF A MARINE VIBRIO
EXAMINED BY FREEZE-ETCHI NG

By

Jack T. Crawford
December, 1973

Chairman: Max E. Tyler

Major Department: Microbiology

The structure of the cell envelope of the marine vibrio MW^tO was
examined by f reeze-etching and other techniques. In thin section the
organism was similar to other gram-negative bacteria. The cell enve-
lope appeared as two double-track layers. No intermediate dense layer
was observed.

Purified peptidoglycan was prepared from cells during the exponen-
tial phase of growth by treating cells with hot sodium dodecyl sulfate.
The peptidoglycan had a typical amino acid and amino sugar composition
and did not have a covalently linked lipoprotein. The material appeared
fibrous when observed in the electron microscope.

Purified 1 i popolysacchar ide (LPS) was extracted by the hot phenol-
water procedure. its composition and appearance was typical. The LPS
was used to determine the specificity of anti-LPS antiserum.

For f reeze-etching, eel 1 s were suspended in a salt solution con-
taining 0.22M NaCl, 0,026M MgCl , and O.OIM KCl. In some cases 20%
glycerol was added as a cryoprotect ive agent. The cell envelope of
this organism f reeze-f ractured in three planes. One fracture split the

cytoplasmic membrane along its fiydrophobic center revealing particle-
studded fracture faces similar to those seen in other bacteria. The
convex face was more densely covered with particles than the concave
face. Large particle-free areas were observed on both faces. The
appearance of the faces was the same with and without glycerol.

A second fracture revealed smooth particle-free concave and convex
faces. This fracture occurred primarily without glycerol, but was
occasionally seen in g 1 ycerol -treated cells.

The third fracture produced a rough convex face and a concave face
composed of subunits approximately 1 0 nm in diameter. In areas where
the subunit layer was incomplete it was observed that the subunits
were globular and were backed by a smooth surface. This fracture
occurred in cells f reeze-etched with glycerol, and occasionally the
globular surface was seen without glycerol.

The outer surface of the cell v-^as exposed by etching in prepara-
tions without glycerol. The surface appeared smooth or finely granular,
but may have been obscured by a thin eutectic layer.

Complementary replicas were prepared and it was demonstrated that
the three pairs of fracture faces were in fact apposed and v;ere produced
by three fractures.

Isolated cell envelopes were prepared by lysing cells in a French
pressure cell. The appearance of these envelopes when f reeze-etched
was similar to v/hole cells. It was observed, however, that the globular
surface could be exposed by etching alone indicating that it had sepa-
rated from the rough surface before freezing.

Crude outer membrane material was prepared by washing the cells

XI

with NaCl and sucrose solutions. This material f reeze-f ractured produc-
ing smooth concave and convex faces.

Specific rabbit anti-LPS antiserum was prepared and used to label
the LPS on the cell surface. This antibody was then labelled with
ferri t in-conjugated anti-rabbit immunoglobulin antiserum. In thin
sections these cells appeared coated with a band of ferritin along
the outer double-track, but separated from it by an electron transparent
space. For f reeze-etching the cells were suspended in 0.05M MgC1_.
The preparations were deep etched and the presence of the ferritin on
the etch face proved that the outer surface of the cell was revealed.
The smooth convex fracture face was immediately adjacent to this ferritin
coated surface.

It was concluded that the envelope fractured at three levels. One
fracture split the cytoplasmic membrane. A second split the outer
membrane along its hydrophobic center revealing smooth fracture faces.
The third fracture exposed the rough convex surface of the rigid layer
and a globular layer which separated the rigid layer from the outer
membrane.

XI 1

INTRODUCTION

The cell envelopes of gram-negative bacteria are multi layered
structures of complex composition. Numerous studies have been reported
which dealt v/ith the chemical composition, structure, and biosynthesis
of the various components of the envelope. The organization of these
components into the functional cell envelope has also been extensively
studied. Related work has focused on the antigenic specificity of the
envelope and its role in pathogenicity. The result has been the
formulation of a generally accepted, though not particularly detailed,
model of the gram-negative cell envelope (16).

In the study reported here the structure of the cell envelope of a
marine vibrio was examined, primarily by f reeze-etch ing . The purpose
of this investigation was first to compare the structure of this organism
with that of more commonly studied gram-negative bacteria such as
Escherichia coll. The second and main objective was to determine if
the technique of f reeze-etch ing could be used to obtain additional in-
formation on the structure of the cell envelope of gram-negative
bacteria in general.

The cell envelope of gram-negative bacteria consists of at least
three layers, the cytoplasmic membrane, the rigid layer, and the outer
layer or membrane. The latter tv-,'0 layers comprise the cell wall,
although this term is often used to designate the rigid layer alone.
The cytoplasmic membrane appears to be a typical membrane composed of

1

phospholipid and protein, and is comparable in structure and function
to the more easily studied cytoplasmic membrane of gram-positive
bacteria (54) .

The rigid layer is composed of pept idoglycan and associated pro-
teins. It is in the form of a "bagshaped macromolecule" which gives
the cell shape and strength {7h) . The outer membrane contains phos-
pholipid, 1 ipopolysacchar ide , and proteins or lipoproteins, and is
unique to gram-negative bacteria. The term "membrane" is used to denote
its appearance when thin-sectioned and viewed in the electron microscope,
and does not imply that it has the functions of other membranes.

In addition to this basic structure, certain bacteria have layers
external to the outer membrane. These extra layers are usually found
in halophilic, photosynthet i c , and other more unusual bacteria. There
are a number of excellent reviews on the structure and composition of
the gram-negative cell envelope (lA, 25, 27, 30, 3^, 35, ^7, 53).

Chemical composition and structure of envelope components. --Pept i -
doglycan has been isolated from a variety of gram-negative bacteria
by various methods, usually involving phenol, hot detergents, or
other h^rsh treatments, combined with enzymatic digestions. Pepti-
doglycan consists of a glycan backbone of alternating N-acetyl
glucosamine and N-acetyl muramic acid residues with a peptide moiety
linked to the carboxyl group of muramic acid. In E_. col i the peptide
is compo5ed of L-alanine, D-glutamic acid, meso-d iaminopimel ic acid,
and D-alanine. Some of the peptide chains are cross-linked from the
amino group of diaminopimel ic acid to the carboxyl group of D-alanine.

This basic structure appears to be universal for gram-negative
bacteria, although in many cases all that is known about the pepti-
doglycan is the amino acid content (56). The amount of pept idoglycan
varies from about ]0% of the cell wall to none in the case of certain
halophiles.

Braun and coworkers (10, 12, 13) have purified a pept idogl ycan-
lipoprotein complex from E^. col i and shown that the lipoprotein is
covalently linked to the carboxyl group of d iaminopime! ic acid. This
linkage is specifically split by trypsin allowing the lipoprotein to
be solubilized with hot sodium dodecyl sulfate (SDS). The amino acid
sequence of the protein has been determined, and it was found that the
lipid is covalently bound (8). The molecular weight of the lipoprotein
is about 10,000.

These workers have also reported that a similar lipoprotein is
attached to the pept idoglycan of several strains of Salmonel la and
Serratia marcescens, but is not found in Pseudomonas fluorescens or
Proteus mirabil is (II). They later found that if the pept idoglycan of
P. mirabi 1 is is isolated from stationary rather than exponential phase
cultures, a lipoprotein is attached (see ref. 36). No lipoprotein was
found in the marine pseudomonad studied by MacLeod's group (24).

Weidel et al . (73) studied meta 1 -shadowed preparations of isolated
rigid layers of E_. col i and found cell-shaped granular structures with
globular units on their surface. Digestion with proteolytic enzymes
removed these globules. Recently, Martin et_ aj_. (36) examined negatively
stained preparations of purified pept idog ! yean. All of the preparations
appeared as granular cell-shaped structures. Layers from E_. coli and
P. mirabi lis were covered with globular particles which were about 9 to

10 nm in diameter and about 20 nm apart. Preparations from Pseudomonas
aerug inosa showed considerably fewer particles, and layers from
Spirillum serpens were free of particles. The particles on the
Pseudomonas pept idog lycan were readily removed by proteolytic enzymes.

Although it would be convenient to associate these particles with
the lipoprotein studied by Braun, they are simply too large for a
molecule with a molecular weight of 10,000. It is possible, however,
that the lipoprotein molecules occur in groups and aggregate to form
larger units.

The composition and structure of the 1 ipopol ysaccharides (LPS)
of the outer membrane have also been extensively studied (3^). They
are readily isolated by the hot phenol-water extraction procedure (76)
and have been characterized in a wide variety of bacteria. The lipid
portion, lipid A, is covalently linked to a carbohydrate core contain-
ing the unusual sugars keto-deoxyoctanoic acid and glycero-D-mannoheptose.
Attached to the core are the 0-antigenic side chains, the composition
of which varies according to the strain of bacteria. In rough strains
the side chains are short or absent. When purified LPS is observed in
the electron microscope it appears to have the structure of membranes.
Shands et_ a_1_. (62) observed LPS positively stained with uranyl acetate
and found various forms, all of which had areas which appeared membrane-
like, that is as a trilaminar or double-track appearance. They also
found that the dimensions of the double-track were the same in LPS
from smooth and rough strains, indicating that the polysaccharide side
chains are not stained and are not seen in thin sections. dePetris
(16) observed that thin sectioned LPS also has a double-track appearance.

The cell envelope also contains a cons iderable amount of lipid,

with the major portion being phospholipid. The predominant fatty acids
are C]^ and C,o straight chain acids, and p-hydroxymyr i state is found
in the LPS. Sterols are absent.

In most descriptions of the outer membrane lipoproteins are listed
as a major component. The work on these molecules has been very vague
and no one has clearly shown that there are any true lipoproteins, that
is covalently linked lipid and protein, in the outer membrane.

Recently, the proteins of the cell envelope of E_. col i have been
studied using SDS-polyacry lamide gel electrophoresis (57, 58).
Schnaltman found from 20 to 30 bands of protein in cell envelope
extracts, and one major protein possessing an apparent molecular weight
of A^ijOOO. By using sucrose gradient cent ri fugat ion he was able to
partially separate the cell wall and cytoplasmic membrane. The major
protein band was found to be localized in the cell wall and accounted
for 70^ of the wall protein. Although further work as indicated that
the initial results may be oversimplified, the basic conclusion, that
there are major structural proteins, is still valid (Schnaitman,
personal communication). Comparable results have been obtained in
E. col i (31) and Salmonella typhimurium (^8).

Studies of proteins released from P_. aeruginosa by ethylened iamine-
tetraacetic acid (EDTA) treatment indicate that structural proteins
exist in the cell wall (68). The location of these proteins will be
discussed later.

Fine structure of the cell envelope. — When thin sections of gram-
negative bacteria are examined in the electron microscope, the cell
envelope appears to consist of a smooth inner membrane and a wavy outer
membrane. Both of these membranes have a double-track appearance and

measure about 7-5 nm in width. In early work these were the only
structures seen, but improved techniques have allov/ed the visualiza-
tion of the intermediate rigid layer (43). This general anatomy has
been observed in many gram-negative bacteria and no attempt will be
made to review this literature (27).

Aside from its double-track appearance, the cytoplasmic membrane
does not have any fine structure detectable in thin sections. Because
of the difficulty in separating it from the cell wall, the membrane has
not been thoroughly studied by negative staining, but available results
indicate that it does not have a subunit structure (55).

The appearance of the intermediate dense layer, or rigid layer,
varies according to the organism studied, the method of fixation, and
the method of staining (27). In some bacteria this layer is seldom or
never seen. This may be due to a thinner layer of pept idogl yean (24)
or possibly to the lack of a covalently bound lipoprotein. The rigid
layer generally follows the contours of the. cytoplasmic membrane. It
has a thickness of 3 to 8 nm.

Evidence that the pept idogl yean is associated with the dense layer
seen in thin sections is provided by lysozyme digestion. Cells which
have been treated with lysozyme and EDTA lack the intermediate layer,
and this treatment allows the outer layer to separate from the cyto-
plasmic membrane (16, 43). It is not known whether the metal stain
is localized in the pept idoglycan alone, or the dense layer represents
another layer which is solubilized when the pept idogl yean is digested.
Purified pept idog 1 yean "saeeuli" appear as a dense layer in thin
section, although they are thinner than the layer seen in whole cells
(16, 29).

The space between the rigid layer and the cytoplasmic membrane
is generally not stained, but is apparently not "empty" since the two
layers are never in contact. in Ni trosocyst i s oceanus globular material
was observed in this area (72).

There is also a space between the rigid layer and the outer mem-
brane. In some organisms the material in this space is stained and
the rigid layer seems to be associated with the outer layer (27).
dePetris (15, 16) showed that this material sometimes appears globular.
Digesting the cells with proteolytic enzymes eliminates this material
and allows the outer membrane to separate from the rigid layer.
Digestion of isolated E^. col i eel 1 envelopes with trypsin, a procedure
which cleaves the lipoprotein From the pept idoglycan , also caused the
envelope to separate, but in this case the rigid layer was not visible
in the thin sections (10).

The outer double-track layer is identical in appearance to typical
membranes and is believed to be a 1 ipid-prote in bi layer. The LPS and
protein are probably arranged in a mosaic fashion, with the lipid
portion of the LPS extending into a phospholipid bi layer. Extraction
of cells with hot phenol -water removes the LPS and such cells are
devoid of the outer membrane (5, 16). With a marine pseudomonad it is
possible to remove the outer layer by washing the cells (22, 23).
This procedure yields isolated membrane material which contains most
of the LPS of the eel Is.

The 0-antigenic side chains of the LPS extend out from the outer
layer and are not stained in routine procedures. In thin sections of
packed cells there is generally a clear space between the outer layers
of adjacent cells, but with rough strains the outer membranes appear
to fuse.

8

Shands (61) used ferri t in-label 1 ed antibody to directly demonstrate
the presence of the LPS in the outer membrane of E_. col i and Salmonella.
He observed that the ferritin molecules extended for considerable
distances from the outer double-track and concluded that there was a
large amount of unstained polysaccharide on the cell surface. Similar
results were obtained with Ve i 1 lonel la (37). It is not Icnown if the
LPS is also located on the inner surface of the outer membrane, but the
apparent area required for the side chains would make such an arrange-
ment unl ikel y.

Freeze-etching studies of gram-negative bacter ia . --In recent
years the new technique of freeze-etching has been used to study the
structure of cells (39i ^0). In this process a sample is flash frozen,
placed in a vacuum chamber, and fractured. ice is sublimed (etched)
from the fracture surface, and the surface is shadowed with evaporated
metal and replicated with evaporated carbon. The replica is cleaned
of adhering cellular material and is ready for examination in the
electron microscope. This procedure reveals the surface topography
of the fracture faces and the faces exposed by etching. A modification
of this method, double-recovery or complementary replica technique,
allows the replication of both surfaces produced when the specimen
is fractured (^2, 6?) ■ By observing the same cell in both replicas,
it is possible to directly demonstrate the association of the various
fracture faces.

One aspect of freeze-etching that caused considerable confusion
and controversy for a period of years is the concept of membrane
splitting. When membranes are f reeze-etched, two surfaces, one concave
and one convex, are produced. These surfaces were originally thought

to be the inner and outer surfaces of the membrane. Branton (6)
proposed an alternate interpretation, suggesting that membranes split
along an internal hydrophobic plane revealing two complementary surfaces
which do not represent the true surfaces of the membrane. Pinto da
SMva and Branton (49) proved this by labelling the outer surface of
red blood cells with ferritin and showing that the fracture face was
unlabel led and the labelled surface was only exposed by etching.
These results were verified by double-replica technique, by which it
was demonstrated that the convex and concave faces are produced by the
fracture of a single membrane in a particular cell {k]) .

The structure of the cell envelope of gram-negative bacteria has
been examined by f reeze-etching (50). in both gram-positive and gram-
negative bacteria the cytoplasmic membrane fractures, producing particle
covered convex and concave faces (^5, 65, 71)- The convex face is
studded with a large number of particles, often arranged in a netlike
array. The concave face is sparsely covered with particles and some-
times appears pitted. Between the particles the surface is smooth,
and occasionally particle-free areas are present {kk , 65). Fiil and
Branton (19) observed that in magnesium starved E^. col i the particles
are arranged in regular arrays and there are large particle-free areas.

The nature of these particles, which are also seen in a variety of
other f reeze-etched membranes, has not been clearly established, but
they probably represent proteins which are intercalated into the lipid
bi layer (7, Gh) . Membranes v/hich have little or no protein, such as
myelin sheath or artificial lipid bilayers, do not have particles on
their fracture faces (7).

The cell envelope of gram-negative bacteria fractures in at least

one other plane. Nanninga (kk) observed that the cell wall of glycerol-
treated E_. col i fractures revealing a rough irregular convex face and
a concave face which is composed of tightly packed flattened subunits
approximately 10 nm in diameter. He suggested that the subunits rep-
resented a protein layer superimposed on the peptidoglycan. In a later
study Van Gool and Nanninga (71) prepared complementary replicas and
proved that the two faces are apposed. By this time the idea of mem-
brane splitting had been established, and they concluded that this
fracture split the outer membrane in a manner analogous to the splitting
of the cytoplasmic membrane. This interpretation requires that the
subunit layer be superimposed on the inner surface of the outer half
of the outer membrane.

Gilleland et al . (26) denionstrated a similar fracture plane in
glycerol-treated P^. aerug inosa , and they also concluded that it was
located in the outer membrane. The concave face in P. aerug inosa
appears as a smooth surface partially covered with spherical units
which are somewhat smaller than the subunits seen in E_. col i .
Extraction of the cells with EDTA solubilizes proteins and yields
osmotically fragile cells. This treatment eliminates the spheres on
the concave fracture face. The protein can be restored to the cell
wall reversing the effect, P^. aeruginosa was also studied by Licl<feld
et_ a]_. (32). A similar fracture was observed and was again postulated
to be a fracture of the outer membrane.

In the E. col i (71) and Pseudomonas (26) studies the cells were
suspended in glycerol as a cryoprotect ive agent. When these cells
were f reeze-etched without glycerol the convex fracture face was
smooth. Both groups suggested that the smooth surface was the same

11

as the rough convex face and its appearance was altered by glycerol.
Neither study mentioned a concave fracture face in cells without
glycerol. Similar results were observed in Nitrosomonas europaea (70).

DeVoe et al . (17) observed a smooth convex fracture face in a
marine pseudomonad f reeze-etched without glycerol. This surface was
also seen in spheroplasts produced by lysozyme digestion. No cell wall
fracture plane was present in cells which had been treated to remove
the outer membrane.

Nitrosomonas oceanus, an organism with a complex cell wall, was
examined by Watson and Remsen (72). They f reeze-etched the bacteria in
a salt solution without glycerol and observed a concave globular layer
similar to the one seen in E_. col i ■ According to their interpretation,
the corresponding convex layer is another globular layer obscured by a
thin rough layer. They concluded that the cell wall split at the
level of the pept idoglycan and the globular concave layer separated
the pept idogl yean from the outer membrane.

The outer surface of the cell envelope has been demonstrated by
etching of preparations without glycerol. The surface is often
smooth, but a fine subunit structure has been reported in E_. col i and
P_. aeruginosa {k , \3, 26). Large subunit structures have been observed
on the surfaces of a few bacteria, but these are clearly extra layers
external to the outer membrane (50).

In this study it will be shown that the cell envelope of the marine
vibrio MWi*0 is typical of gram-negative bacteria. When f reeze-etched
under a variety of conditions the cell envelope fractures in three
planes. The location of the fracture planes and the nature of the
fracture faces are discussed and a model is presented vjhich is consistent

12

with the current understanding of the freeze-etch process and the
structure of the cell envelope as determined by other techniques.

MATERIALS AND METHODS

Organism and maintenance of cultures. — The organism used in this
study was isolated in this laboratory from a sample of water from the
Waccasassa River estuary on the west coast of Florida and was designated
marine vibrio MW'40. It is a marine organism requiring added salts for
growth in ordinary culture media and lysing in distilled water. It is
one of nine strains studied phenotypica 1 1 y in a preliminary study
(unpublished results). It was found that the organism is a gram-negative
polarly flagellated slightly curved rod, which is oxidase positive,
fermentative on glucose, sensitive to the pteridine 0/129, and has a GC
ratio of approximately ^3%- It produces an insoluble blue-black extra-
cellular pigment which is probably indigoidine. Using the scheme of
Shewan (63) the organism would be classified as a Vibrio. The organism
has sheathed flagella and would be classified by Baumann as a member of
the genus Beneckea (l). It appears to be similar to the organism he
named Beneckea nigrapulchr i tuda (2) .

Stock cultures were maintained on plates of a medium containing
O.S% gelatin (Difco Laboratories, Detroit, Mich.) and 2% agar, in a
salt solution, hereafter designated "complete salts," consisting of
0.22M NaCl, 0.026M MgCl and O.OIM KCl.

Culture and harvest of cells. — Cells were cultured in a medium
containing O.S% trypticase (Baltimore Biological Laboratories, Baltimore,
Md.) and 0.1^ yeast extract (Difco) in complete salts. The trypticase

13

and yeast extract medium was prepared at twice tfie final concentration
and adjusted to pH 7-6 with KOH before mixing with an equal volume of
double strength complete salts.

For routine work cells were cultured in 50 ml of medium in a 250 ml
flask, or in 250 ml of medium in a 1000 ml flask, incubated on a rotary
shaker overnight at 25 C. Cells were harvested by centr ifugat ion at
3000 X g for 10 min. Except where otherwise noted, all centr if ugat ion
was perfomred at 1-^ C in a Sorvall RC-2B refrigerated centrifuge using
an SS-S't rotor (Ivan Sorvall, inc., Norwalk, Conn.).

Large quantities of cells v/ere cultured in 11 1 of medium in a

20 1 carboy incubated at 25 C with aeration until late exponential

9
phase of growth, giving an approximate cell concentration of 1.5 x 10

cells/ml. The culture was cooled to 0 C by addition of 't 1 of ice and

sufficient NaCl (50 g) to maintain the salt concentration. Cells were

harvested in a precooled Delaval cream separator (Delaval Separator Co.,

Poughkeepsie, N. Y.).

Preparation of pept idoglycan . — Cells from an 11 1 batch culture
were suspended to 200 ml in complete salts at 0 C. The suspension was
rapidly heated to 60 C by adding 1 1 of complete salts preheated to
70 C, and placing in a water bath at 60 C for 10 min. The cells were
lysed by addition of 200 ml of a 28^ (w/v) solution of Triton X-100.
The mixture was stirred for 10 min and then transferred to an ice-water
bath and cooled to about 4 C. When the temperature had fallen to 37 C,
500 ug of deoxyr ibonuclease (DNase) were added to reduce the viscosity.
The suspension was centrifuged at 10,000 x g for 90 min in a GSA rotor.

The pellets were v/ashed once in 1200 ml of 0. IM NaCl and divided
into two equal portions. One half of the material was resuspended to

15

200 ml in O.OIM tris (hydroxymethy I ) -ami no methane (Tris) buffer,

pff 8.2, containing 40 mg of trypsin, and shaken for 30 min at 25 C.

A 120 ml volume of a solution of sodium dodecyl sulfate (SDS) was added

to give a final concentration of k%. The mixture was shaken for 30 min

at 25 C, and then centrifuged at 27,000 x g for I hr at 20 C. The other

half of the material was suspended to 200 ml in O.OlM Tris buffer,

pH 7.2, without trypsin, mixed with SDS, and centrifuged as described.

The subsequent treatment was the same for both pellets.

The pellet was suspended in 50 ml of k% SDS and added slowly with
stirring to 250 ml of boiling kZ SDS. After boiling for 5 min the
suspension was cooled slightly, shaken for 2 hr, and lef t overn ight at
room temperature. The pept idog lycan was pelleted by centr i fugat ion at
27,000 X g for 1 hr at 20 C, and washed once in 0.02M NaHCO . It was
dialyzed for 48 hr against three changes of 0.02M NaHCO and two
changes of distilled water, and then lyophilized.

Peptidoglycan was also prepared by lysing cells in hot SDS. A
600 ml volume of culture was removed from the shaker, poured into 600 ml
of boiling kl SDS, and placed in a water bath at 60 C for 30 min. The
suspension was cooled to 20 C and centrifuged at 16,000 x g for 1 hr
in a GSA rotor. The pellet was treated with boiling SDS, dialyzed, and
lyophilized as previously described. This procedure produced only a
small quantity of material, but it appeared to be whiter and purer than
the peptidoglycan prepared from the batch cultures.

Preparation of LPS.— LPS was extracted by a modification of the
hot phenol-water procedure (76). As suggested by 0' Leary ejt_ aj^. (46)
a solution of MgCl^ was substituted for distilled water in this procedure.
Cells from an 11 1 batch culture were washed once in complete salts and

)6

resuspended in 200 ml of 0.05M MgCl^- The suspension was mixed with an
equal volume of 90-i phenol held in a v;ater bath at 60 C, and the mixture
was shaken vigorously for 30 min. The mixture v/as cooled to 20 C and
certrifuged at 1500 x g for 30 min. The top aqueous phase containing
the LPS was removed and saved, and the phenol phase and the interface
were reextracted with an equal volur^e of fresh 0.05M MgCl .

The aqueous phases were combined and centrifuged at 27,000 x g
to remove debris. The supernatant was dialyzed for 36 hr against
several changes of O.OIM MgCI. and then lyophilized.

The crude LPS was purified by a modification of the method of
Romeo et al . (52). The LPS, 2 mg of ribonuclease (RNase), and 50 ug
of DNase were suspended in 60 ml of a solution consisting of O.OlM
Trisacetate buffer, pH 7-5, O.lM NaCl, O.OIM MgCl , and O.OlM sodium
azide. The suspension was incubated with stirring for 2k hr at 37 C,
and then dialyzed for ^4 hr against 3 • of the same salt solution minus
the NaCl. A 1 mg quantity of protease was added to the suspension in
the dialysis bag, and it v;as incubated for 2^ hr at 37 C. The solution
was diluted to 100 ml with O.OIM MgCl and centrifuged at 96,000 x g
for 8 hr in a Beckman Model L-2 ul tracen tr i f uge using a Ti50 rotor
(Beckman Instruments, Inc., Palo Alto, Calif.). The pellets were
resuspended to kO nl in O.OIM MgCl 160 ml of 95% ethanol v-ere added,
and the mixture was stored overnight at 't C to allow complete precipita-
tion of the LPS. The LPS was collected by centr ifugat ion at 10,000 x g
for 30 min in a GSA rotor, resuspended in double-distilled water, and
lyoph i 1 ized .

Preparation of cell envelopes using the French pressure cells. —
Except where otherwise noted, all of the manipulations involved in

17

preparing cell envelopes, cell walls, and isolated outer membranes were
performed at 0-k C, and the final products were stored at 0 C until
processed for electron microscopy.

Cells from a 250 ml culture were suspended in 35 ml of 0.05M MgCl
containing 50 ug of RNase and 50 ug of DNase. The cells were lysed by
passing the suspension thru a French pressure cell (American Instrument
Co., Silver Springs, Md.) operated at a pressure of approximately
16,000 psi. The lysate was diluted with an equal volume of distilled
water and centrifuged at 17,000 x g for 1 hr. The pellet was re-
suspended in O.OIM MgCK and centrifuged at 3,000 x g for 15 min to remove
whole cells. The supernatant was decanted and centrifuged at 17,000 x g
for 1 hr to pellet the cell envelopes.

Preparation of cell walls using Triton X-1 00. --Cel Is from a 250 ml
culture were resuspended in 250 ml of complete salts containing 50 ug
of RNase and 50 ug of DNase. The cells were lysed by the addition of
5 ml of a 50% (w/v) solution of Triton X-100 and the mixture vias shaken
for 10 min. The cell walls were pelleted by centrifuging at 17,000 x
g for 1 hr, and washed once in a solution containing 0.05M MgCl_ and
O.OIM Tris buffer, pH 7.5- The pellet was resuspended in kO ml of the
Tris-Mg solution and divided into two portions. A 0.5 mg quantity of
lysozyme was added to one of the suspensions and both were incubated
for 5 rnin at 37 C in a water bath. The suspensions were cooled and
centrifuged at 17,000 x g for 1 hr.

Preparation of crude outer membranes. — Crude outer membranes were
released from the cells by a modification of the procedures of Forsberg
e^ a_l_. (22). Cells from a 250 ml culture were washed once with complete
salts and three times with 0.5M NaCl by resuspending them in 200 ml

18

volumes of the wash solution and centrifuging at 3000 x g for 10 min.
The final pellet was resuspended in 100 ml of 0.5M sucrose and placed
in a rotary shaker for 30 min at 20 C. The cells were pelleted by
centrifuging at 7500 x g for 15 min. The supernatant was carefully
removed and recentr i fuged to remove any remaining whole cells. MgCl
was added to the suspension to give a final concentration of O.OIM,
and the cell wall material was pelleted by centrifuging at 96,000 x g
for 2 hr. The pellets were resuspended, pooled in one tube, and
centrifuged as before.

Preparation of anti-LPS ant i sernm. -Thr.. antigen preparations
were used: (1) whole cells treated briefly in a blender to remove
flagella and suspended in complete salts; (ii) cell walls prepared in
Triton X-100 and suspended at a concentration of 1 mg/ml in 0.9?
NaCl; (iii) partially purified pept idoglycan prepared by treating
cell walls with warm SDS and suspended at a concentration of 1 mg/ml
in 0.9% NaCl.

New Zealand white rabbits were injected subcutaneousl y with 1 ml
of antigen. Four injections were given at two-week intervals, and
the animals were bled two v/eeks after the final injection. Serum
was heated at 56 C for 30 min to inactivate complement, sterilized
by Millipore filtration, and frozen.

The activity of the antisera was determined by the agglutination
of bacterial cells. Cells were washed and resuspended in complete salts
solution to a concentration of approximately 5 x 10^ cells/ml. 0.2 ml
of this suspension was mixed with 0.2 ml of serial dilutions of serum
and incubated at 30 C for 30 min. The degree of agglutination was
determined by microscopic examination.

19

Antiserum was adsorbed with LPS by mixing the serum with an equal
volume of complete salts containing purified LPS at a concentration
of 100 ug/ml . The mixture was incubated for 2 hr at 30 C and then
centrifuged at 27,000 x g for 15 min. The supernatant was carefully
removed and centrifuged again.

Labelling of cells with fer r i t i n-conj'ugated antibody. — The in-
direct method was used to label the LPS on the cell surface with
ferr It in-conjugated antibody. Cells were suspended to a concentration
of approximately 9 x 10 cells/ml in a dilute medium containing 0.1^
trypticase in complete salts. A 30 ml volume of this suspension was
mixed with 5 ml of rabbit anti-LPS antiserum and incubated for 30 min
at 30 C. The cells were pelleted by centr i fugat ion at 12,000 x g for
10 min, washed three times with '*0 ml volumes of trypticase medium,
and resuspended in 10 ml of the same medium. To this was added 0.5 ml
of commercially prepared ferr i t in-conjugated IgG fraction of goat anti-
rabbit immunoglobulin (igA + IgG + IgM) (Cappel Laboratories, Inc.,
Downingtown, Pa.)- The suspension was incubated for 30 min at 30 C,
diluted to 40 ml with trypticase medium, and centrifuged at 3000 x g
for 10 min. The pellet was washed once in kO ml of complete salts,
and resuspended in 20 ml of complete salts. The suspension was divided
between two 15 ml glass centrifuge tubes and centrifuged at 3000 x g
for 10 min. One pellet was fixed for thin sectioning and the other was
resuspended in 10 ml of 0.05M MgCl and repel leted. It was then re-
suspended in several drops of the supernatant and transferred to a
plastic microtube. This tube was placed in a large tube and centrifuged
at low speed for several minutes. The microtube was cut off just above
the pellet and the cells were transferred to specimen holders for freeze-

20

etching. This procedure was necessary because of the very small size
of the pel let.

In some experiments, cells were fixed in glutaraldehyde before
antibody labelling. Cells from ^0 ml of culture were fixed for 1 hr
in 20 ml of kl glutaraldehyde in buffered salts (see section on thin
sectioning). A 20 ml volume of 0.5% trypticase medium was added and
the suspension was centrifuged at 3000 x g for 10 min. The cells were
resuspended in 40 ml of dilute trypticase medium and incubated for 1 hr,
This treatment was designed to eliminate any unreacted glutaraldehyde
and prevent non-specific uptake of antibody. The cells were pelleted,
resuspended in dilute trypticase medium, and labelled with antibody
as previously described.

Except for the incubations with antisera, all manipulations were
carried out at 0-^ C.

Thin sectioning. --Cel Is were fixed for thin sectioning in ^
glutaraldehyde in a buffered salts solution consisting of 0.22 M flaCl,
O.OIM MgCl^, O.OIM KCL, and 0.02M potassium phosphate buffer, pH 7.4.
The concentration of Mg ion was reduced from the concentration in
complete salts to prevent the formation of a precipitate with the
phosphate buffer.

A i» ml volume of the A% glutaraldehyde fixative was added to hO ml
of an actively growing culture and the suspension was set at 4 C for
30 min. Cells were harvested by centr if ugat ion resuspended in 10 ml
of fresh fixative, and set for 2 hr in the cold. Cells were pelleted,
washed three times in buffered salts, and pelleted again. The super-
natant was carefully removed and 2 ml of 2% OsO, in buffered salts was
added without resuspending the pellet. The cells were fixed in the

21

cold for 2 hr, and then washed with repeated changes of buffered salts
and finally with distilled water. The cells were repelleted as

necessary.

Cells were dehydrated thru a series of 25. 50. 75. 95. and 1 OOl
ethanol and placed in acetone for embedding in plastic. The pellet was
broken into s.all pieces and embedded in an Epon-Arald i te mixture by
the method of MoHenhauer (38).

Antibody labelled cells were fixed and embedded as described,
omitting the initial glutaral dehyde fixation in the medium. Cells
which were fixed with glutara Idehyde prior to antibody labelling were
postfixed with OsO^ and embedded as described.

isolated cell envelopes were fixed by resuspending in k% glutaralde-
hyde in phosphate buffered O.OIM MgCl^. postfixed in OsO^. and embedded

as described for whole cells.

Embedded material was sectioned with a diamond knife (DuPont de
Nemours and Co.) in a Porter-Blum MT-2 ul tram i crotome (Sorvall, Inc..
Norwalk. Conn.). Sections were poststained with lead citrate (50 and

uranyl acetate.

Positive and negative sta i ns . - 1 sol ated cell envelopes were
positively stained by mixing a suspension with 0.5°^ aqueous uranyl
■ acetate. A small drop of the mixture was applied to a formvar coated
grid and ii^iediately drawn off with filter paper- Purified LPS was
suspended in O.OIM MgCl^ and stained with uranyl acetate as described.
Purified peptidoglycan was suspended in distilled water and negatively
stained with potassium phosphotungstate. Grids were examined in the

electron microscope immediately.

Freeze-etching. -Glycerol -treated cells were prepared by mixing a

22

culture wTth an equal volume of kQ% (v/v) glycerol in complete salts
and incubating for 15 min at 't C. Other specimens were frozen in the
appropriate suspending medium without cryoprotect i ve agents. Pellets
of the various materials were resuspended in very small quantities of
the supernatant and transferred to specimen discs with a drawnout
Pasteur pfpette. Gold specimen discs with a central hole were routinely
used. For very small specimens, such as antibody labelled cells, gold
dfscs without a well were used. Specimens were frozen by plunging
them into liquid Freon 22 held at -150 C or a mixture of liquid and
sol id N at a temperature of -209 C.

Specimens were f reeze-etched according to the methods of Moor and
Muhlethaler (^O) using a Balzers BA 36OM freeze-etch apparatus (Balzers
High Vacuum Corp., Santa Ana, Calif.). The specimen was transferred to
the precooled (-150 C) stage and placed under a high vacuum. The
specimen temperature was raised to -100 C and it was fractured with a
cold microtome until a smooth surface was obtained. The surface was
etched for 0 to 2 min and then replicated.

The replicas were cleaned by floating them on S0% (v/v) Chlorox
overnight and rinsed on several changes of distilled water. Replicas
of ferritin labelled cells required an additional cleaning on 40% (w/v)
chromic acid. Replicas were picked up on uncoated 3OO or 400 mesh grids.

Double-replica freeze-etch ing . — Complementary replicas were prepared
by the methods of Muhlethaler et a 1 . (A2) . A suspension of glycerol-
treated cells was frozen in a sandwich of grids (28), Two 3 nun copper
discs were flared using a special press. Two nickel London finder grids

were simiSarly bent. One of the finder grids was placed on a copper
disc and a small quantity of vaseline was applied to the edges. The

23

second finder grid was placed on top of this and carefully adjusted so
that the corresponding grid squares were superimposed. The cell sus-
pension was applied to the grids, being sure that it penetrated through
the grid openings of both grids. The second copper disc was set on
top and the sandwich was clamped with a pair of forceps. The excess
liquid was removed with filter paper and the sandwich was plunged into
1 iqu id-sol id N. .

The sample was fractured in the Balzers apparatus using a special
hinged holder (custom designed and manufactured in this laboratory accord-
ing to the design of Muhlethaler et al. (^2)). The holder replaces the
usual specimen stage. After fracturing, the specimens were replicated
immediately without etching. The finder grids with the attached
replicas were carefully separated from the copper discs and allowed to
dry. The grids were placed in chloroform to remove the vaseline and
then removed and dried. Replicas were cleaned overnight in kO% chromic
acid. Distilled water was added to dilute the acid and the grids were
transferred to fresh water. They were then removed and dried. If
the manipulations were done carefully the replicas remained attached
to the finder grids. This facilitated finding the corresponding halves
of cells in the two replicas.

Electron microscopy . --A) 1 materials were examined with a Hitachi
HUll-C or HUl 1-E electron microscope (Hitachi, Ltd,, Tokyo, Japan)
operated at an accelerating voltage of 75 Kv . Photographs were taken
on DuPond Cronar film.

Amino acid anl ays i s . --Amino acids and amino sugars were determined
on a JEOL model JLC-5AH automated amino acid analyzer (JEOL U.S.A., Inc.,
Cranford, N. J.) according to Spackman ejt_ aj_. (66). Samples were
hydrolyzed in 4NHC1 for 11 hr at 105 C in sealed ampules.

2k

Reagents. --The following chemicals were used: glutaraldehyde, 8%
under N„ gas (Pol ysciences, Warrington, Pern.); OsO, (Engelhard
Industries, Newark, N. J.); and Triton X-100 (Sig.ma Chemical Co.,
St. Louis, Mo.). The following enzymes vyere obtained from Sigma:
r I bonuclease, pancreatic, Type 1-A; deoxyr i bonuclease, pancreatic:
lysozyme, eggwhite, Grade I; protease, fungal, Type V; trypsin,
pancreatic, Type ill. All other reagents were reagent grade.

RESULTS

Thin sections of whole cells. --In longitudinal sections the vibrios
appeared as slightly curved rods with a fine structure typical of gram-
negative bacteria (Fig. 1). Two double-track layers, the cytoplasmic
membrane and the outer membrane, were visible in the cell envelope
(Figs. 1, 2, and 3). The outer membrane was wavy, but in places appeared
to be closely associated with the cytoplasmic membrane (Fig. 2). The
fragmented appearance of the outer membrane was probably due to its
undulating contour, since it appeared as a double-track only when it
was perpendicular to the plane of the section.

In some cells there was a dense layer inside the cytoplasmic mem-
brane (Figs. 1 and 3). The nature of this layer is unknown. The area
between the inner and outer membranes was often fuzzy, but no organized
intermediate layer was visible. The dense layer seen in E. coll (16)
is seldom seen in marine bacteria, possibly because the pept idogi yean
is very th in (2^) .

Purified pept i dog 1 yean. — The method used to prepare pept i dogl yean
was similar to that of Braun and Rehn (10) and involved the treatment of
crude cell walls with hot SDS. Cells were heated to inactivate auto-
lytie enzymes and were lysed with Triton X-100 to minimize the fragmen-
tation of the pept idogl yean which occurs during lysis of cells by
mechanical means.

In order to determine if there was a lipoprotein covalently attached
to the pept idog lycan, two samples were prepared. One sample was digested

25

Figs. I, 2, and 3- Thin sections of cells fixed in g lutara Idehyde and

OsO, . Note the double layered structure of the
cell envelope.

Fig. I. Longitudinal and cross sections of cells,
(x 80,000).

Fig. 2. Cross section of a cell showing the double-track
appearance of the cytoplasmic and outer membranes
and close association of the two (arrows).
(x 212,000) .

Fig. 3. Cross section showing an extra layer (arrow) inside
the cytoplasmic membrane. (x 167,000).

27

•t^

■ 5\

28

with trypsin, the other was not. The results of the amino acid analysis
of these two samples were the same (Table 1), Both contained the usual
amino acids and amino sugars found in pept idogl yean from gram-negative
bacteria, that is, glucosamine, muramic acid, alanine, glutamic acid,
and d iaminopimel ic acid, in a ratio of approximately 1:1:2:1:1, Other
amino acids were present in very small amounts. Although the trypsin-
treated sample was slightly cleaner, the results indicated that there
was no lipoprotein attached to the pept idogl yean.

With the conditions used in the amino acid analysis neither muramic
acid and serine nor diaminopimel ic acid and methionine could be
separated. Since the amounts of the other amino acids not normally
found in pept idoglycan were low, it was unlikely that this affected
the results significantly.

It was possible that an autolytic enzyme was cleaving the lipo-
protein, especially during the 30 to kS minutes required to harvest the
cells. To minimize this possibility a small sample of pept idogl yean
was prepared by pouring an actively growing culture directly into hot
SDS. The amino acid content of this sample v/as the same as before, and
there was no evidence of a covalently linked lipoprotein (Table 1).

When negatively stained, the pept idoglycan appeared as cell shaped
fragments which were distinctly fibrous (Fig. ^) . The pept i dogl yean
was quite fragmented and did not resemble the finely granular "sacculi"
obtained from other bacteria. The preparation shown was treated with
trypsin, but the untreated material had the same appearance. No
particles or granules were seen in either preparation.

Purified 1 i popolysacchar i de. — The primary purpose of preparing LPS
was to obtain a purified antigen to test the specificity of antisera.

29

c

•— >

>>

JQ

c/1

•— 00

T5

U1 Q

(U

>-ai

*J

^

fO

4-1

*J o

o

O X

U1

(U

—

u

•^

T3

4-J

m

<Ni

LA

r>^

CNI

O

CO

i_

O

O

—

o

•

0)

' —

—

—

04

—

o

o

1/1
0)

"o

E
3.

c

TO
(J
>-

O)

o

a.

D.

■a

O
in
(/I

C

O

o

c

<

TO

B
O

T3
0)

I/)

a.

o

-3-

LA

r-~ -3-

CO —

CSI
CM

1/1
0)

"o

£
a.

>-

m

o

E

1/1

"o

E
3.

in
■a

I-

•—

D1

o

3

(t)

c

01

0

CO

O

c

c

•—

• — .

^

a

0~l

o
o

CO

o

o

o
o

-a-

-3-

-3-

CSI

CO

-3-

o
CO

-3-

CTi

CO

O

O
O

UA

O — —

— O

CNI

o

LA

-3-

-3-

CA

CPy

o

CT^

CM

CA

-^

rA

o

T3

.—

■o

—

»—

(1)

O

O

4)

c

o

(0

E

C

_

ra

■—

>«

E

o

Q.

E

(D

o

—

D

O-O

TO

l/l

•—

E

C

c -o

1/1

o

E

fD

■ ■~

O

(J

CD

■M

c

'i o

o

u

1-

3

(T3

TJ fD

3

~~

3

^

■ —

•—

^—

CJ3

CJ

c

(U

10

JD

(D

L.

,

O

C

in

E

■M

3

c

3

"o

o

u

E

ra

4-1

l-

OJ

o

o

s:

(T3

\n

t-

4->

c

•

<0

c

TO

c

••^

•«■

cn

.

c

4-1

c

0)

o

>

c

O

cr

O

(U

•— .

sz

o

1/1

1-

4->

•

0)

(U

0)

1>

^

u

j:

in

E

Q.

4->

m

E

E

TO

0)

X

O

O

1_

4-)

&_

u

c

0)

o

14-

n-

(U

3

^

-^

-o

T)

TO

in

c

0)

OJ

4-1

■o

o

4-1

4-1

•—

TO

TO

XI

O

U1

i_

1-

• —

TO

D

TO

TO

(J

iJ

O.

Q.

TO

O

3

(1)

0)

c

^—

in

in

O

•—

<u

•—

E

<U

dJ

r^

TO

0)

j2

-Q

^E

u

•—

4-P

4.>

(D

E

o

o

Q.

JZ

TO

c

c

O

•M

in

c

o

O

-o

T3

O

^—

^^

E

1 —

3

3

3

TO

f—

f—

o

O

<

<J

<_)

«_)

C3

TO

JD

o

■D

<U

Fig. ^4. Purified tryps in-treated pept idoglycan negatively stained with
potassium phosphotungstate. Note the fibrous nature of the
layer. (x 22A,000).

31

32

but the general nature of the LPS v/as also of interest. The standard
method of preparing LPS from smooth strains of bacteria, phenol-v;a ter
partition, was used to obtain crude LPS, and it was further purified by
enzymatic treatments.

The results of preliminary work on the cha-nical analysis of the
LPS indicated that it was reasonably pure and similar to that found in
other bacteria. Keto-deoxyoctanoic acid and heptose v/ere present (18,
75). Amino acid analysis demonstrated three major components, probably
amino sugars, and small amounts of the various amino acids.

When uranyl acetate-stained preparations were viewed in the
electron microscope the LPS had the appearance of ribbons of discs
(Figs. 5 and 6). Similar results were reported by Shands et_ aj_. (62)
for LPS from Sal none 1 la. Only a few areas appeared to have a double-
track structure.

Freeze-etching of cells without glycerol . --When cells were freeze-
fractured in a salts solution without glycerol the cell envelope
fractured in two planes, splitting the cytoplasmic membrane and the cell
wall (Figs. y-l'*). The fracture faces of the cytoplasmic membrane were
similar to those seen in other bacteria. The convex face was a smooth
surface densely covered with particles, '4 to 8 nm in diameter, which
were sometimes arranged in a netlike array (Figs. 7-10). The concave
face was more sparsdy studded with particles (Figs. 7, 12, I3, and ]k).
In some preparations there were large circular areas devoid of particles
on both the convex and concave fracture faces (Figs. 7, 9, and 12).
There was no apparent correlation between the growth conditions of the
cells and the presence or absence of these particle-free areas.
Generally all or none of the replicas prepared from a single batch of

Fig. 5. Purified LPS positively stained witii uranyl acetate showing

ribbon structure and occasional areas (arrows) with double-track
appearance. (x 168,000).

Fig. 6. Purified LPS positively stained with uranyl acetate showing
ribbons and discs. (x 168,000).

^

'^'

%'"^«"*%

%

^iu

X

/

3A

^^

35

cells demonstrated this feature, v/hich indicates that it was a function
of the condition of the cells and not an artifact of the f reeze-etchi ng
process.

The cell envelope also fractured at another level, although with
much less frequency and generally only in small areas. This fracture
revealed two smooth faces and was probably splitting the outer layer.
The convex face was often seen as patches of cell wall material on the
cytoplasmic membrane (Figs. 7, 8 and 9). These patches appeared to
originate from the outer cell wall layer (Figs. 7 and 9). The corre-
sponding concave face was seen as holes in the cytoplasmic membrane
(Figs. 7, 12, and 13). The location of these patches did not appear
to be related to the particle-free areas on the cytoplasmic membrane.

The relationship between the smooth fracture face and the outer
cell wall layer was more clearly shown in etched preparations (Fig. II).
Etching of preparations containing considerable amounts of salts often
revealed a eutectic layer surrounding the cells (Fig. 8, and ref. 17).
In some cases the eutectic layer was very thin, and although very fine
surface detail may be obscured, the general structure of the surface
was seen. This was demonstrated by the presence of flagella on the
cell surface (Figs, 10 and 11). It was observed that the smooth convex
fracture face is revealed by fracturing away a thin piece of the outer
layer (Fig. 11).

Another concave fracture face was occasionally observed in cells
fractured without glycerol. This surface was composed of globular
subunits (Fig. 1 4) . The corresponding convex face was not demonstrated
in these preparations.

A more regular paracrystal 1 ine array of subunits was sometimes

Fig. 7. Freeze-etched cells suspended in complete salts without glycerol
This low magnification view shows concave and convex fracture
faces. (x 50,000).

37

Fig. 8. Freeze-etched cell suspended in conplete salts, showing the
particle-studded convex fracture face of the cytoplasmic
membrane and a patch of cell wall material. Etching has
revealed a eutectic layer which sometimes surrounds the cells,
(x 61 ,500).

Fig. 9. Unetched convex fracture of cells vnthout glycerol. Note the
smooth fracture face of outer membrane and relation of this
fracture to the outer surface (edge) of the cell. Large
particle-free areas of the cytoplasmic membrane face are shown
(arrcAv) . (x 82,000) .

39

Fig. 10. Deep etched cell in complete salts, shovnng the smooth or

slightly granular outer surface of the cell exposed by etching.
Note the structure of the fractured sheathed flagellum and
unidentified filaments. (x 83,200).

Fig. 11. Etched cell in complete salts showing the smooth convex

fracture face of the outer membrane and the true surface of
the cell. Note the thickness of the fracture edge (arrow),
(x U2,000).

k]

1/

flagellum 10

Fig. 12. Concave fracture of a cell in complete salts showing the

particle-studded fracture face of the cytoplasmic membrane,
and holes in the surface which reveal the smooth concave
fracture face of the outer membrane. Particle-free areas
(arrow) are also shown. (x 80,000).

Fig. 13. Concave fracture of a cell clearly showing the smooth fracture
face of the outer membrane. (x 68,000).

Fig. I^. Concave fracture of a cell showing a surface with distinct
subunit structure. (x 116,500).

43

seen (Fig. 15). Although it is not apparent in this figure, this
array is located on the inner surface of the cytoplasmic merr.brane and
possibly corresponds to the extra layer seen in thin sections (Fig. 3).

Freeze-etchinq of glycerol -treated cells. — The primary fracture
plane in glycerol -treated cells v;as the cytoplasmic membrane, and the
fracture faces were indistinguishable from those seen in cells without
glycerol (Figs. 16 and 17). The outer layer also fractured revealing
the two smooth faces previously described (Figs, 18 and 19). Frozen
solutions of glycerol and v.ater do not sublime significantly and there-
fore the outer surface of the cell could only be seen as an edge
(Fig. 19).

A third fracture commonly occurred which revealed a rough convex
surface and a concave surface composed of globular subunits (Figs. 20,
21, 22, and 23). This fracture plane has been demonstrated in other
bacteria and generally was thought to split the outer mem.brane (26, 32,
71). Our results suggest that a more plausible interpretation is that
the wall splits between the rigid layer and a globular protein layer
which separates the rigid layer from the outer membrane.

The convex surface of the rigid layer was generally seen only at
the ends of the fractured cell, with the envelope usually splitting at
the level of the cytoplasmic membrane (Fig. 20). The surface is rough
or granular with no clearly defined structure.

The concave globular layer appeared to be composed of globular
units approximately 10 nm in diameter (Figs. 22 and 23), In some areas
the subunits were in rows, but generally no ordered arrangement was
apparent. The layer was usually only partially visible, extending
from beneath the fractured cytoplasmic membrane. It was sometimes

Fig. 15. Freeze-etched cell in complete salts showing a paracrystal 1 ine
array (arrow) inside the cytoplasmic membrane, a. (x 60,000)
b. (x 167,000).

^

Fig. 16. Concave fracture of the cytoplasmic membrane in a glycerol-
treated cell. Cells were suspended in complete salts con-
taining 20°^ glycerol. The edge of the cell wall is visible
(arrow). (x 106,000) .

Fig. 17. Convex fracture of the cytoplasmic membrane in a glycerol-
treated cell. Note the netlike arrangement of the membrane
particles. (x 107,000).

k8

Fig. 18. Unetched fractured glycerol-treated cell showing the concave
fracture face of the outer membrane. Note the vvidth of the
shadow cast by the layer (arrow) indicating that it is quite
thin. (x 76,000) .

Fig. 19- Convex fracture of a glycerol-treated cell showing the smooth
fracture face of the outer mei-nbrane. In preparations with
glycerol the outer surface of the cell can only be seen as
an edge (arrow). (x 92,000).

50

*

r

/•I--

-i- V

^J

1-. > "^ ' ^

... *Vf' ;.

Figs. 2U and 21. Freeze-etched glycerol-treated cells showing the

rough convex fracture face of the rigid layer.

Fig. 20. (x 100,000).
Fig. 21. (x 157,500).

52

¥

m

'A

.^^/

Fig. 22. Concave fracture of a glycerol-treated cell. Note the material
apparently ice, on the surface of the globular layer.
(x 122,000).

Fig. 23. Concave fracture of the globular layer in a glycerol-treated
cell. a. (x 66,000), b. (x 167,000).

54

LL

23 b'

55

obscured by an intermediate layer which appeared to be ice (Fig. 22).
This suggests that the fracture faces were separated by fluid prior
to freezing.

In areas where the globular layer was incomplete it was apparent
that the layer was in fact composed of individual subunits, and that
the subunits were backed by a smooth surface (Fig. 2k). This smooth
surface was probably the true inner surface of the outer membrane as
opposed to the smooth fracture face of the outer membrane.

The cell envelope was also seen in profile in cross fractured
cells (Figs. 25 and 26). The inner and outer membranes were routinely
observed, and in some cases an intermediate layer was visible. The
membranes occasionally had a double-track appearance, but there is no
good explanation for this.

Using double-replica technique, complementary replicas of glycerol-
treated cells v;ere prepared. The matching fracture faces of individual
cells were located in both replicas. in the figures shov/n, one of the
negatives was inverted before printing so that the images would appear
superimposable rather than as mirror images. The results of the double-
replica work clearly demonstrated that the three pairs of fracture
faces previously described were complementary, and each pair was
produced by a single fracture (Figs, 27, 28, and 29). It was also
apparent that while a rather thin layer was fractured away to expose
the smooth convex face, a thick piece of wall vjas removed in exposing
the rough convex face. This suggests that the rough face is at a
lower level in the cell envelope.

Structure of isolated cell envelope fractions. --The study of
isolated cell envelopes and various fractions of envelopes by freeze-

Fig. 2k. Concave fracture of a gl ycerol -treated cell. The globular
layer is incomplete and the individual subunits (arrow) and
the smooth backing layer are visible. The smooth surface is
thought to be the true inner surface of the outer membrane,
a. (x 60,000), b. (x 157,000).

Fig. 25- Cross fractured cell envelopes of three adjacent cells show-
ing the edges of the cytoplasmic membrane, rigid layer, and
outer membrane. The membranes have a double-track appearance,
(x 132,000).

Fig. 26. Edge fractured cell showing the fracture surface of the

membrane and the cross fractured cell wall. (x 132,000).

57

Fig. 27. Complementary surfaces of a glycerol -treated cell observed by
double-replica technique. This pair of photographs clearly
shcK-/s that the nembrane faces are produced by fracturing.
PHB, poly-B-hydroxybutyrate granule. (x 80,000).

59

Fig. 28. Complementary fracture faces of a gl ycerol -treated cell

showing the relationship between the concave globular layer
and the convex rough faces. (x 75,000).

61

Fig. 29- Complementary fracture faces of a gl ycerol -treated cell

showing the relationship between the smooth concave and convex
fracture faces. (x 100,000).

63

7

I

V

6^

etching offers several theoretical advantages over using whole cells.
Since the fractions are of less complex composition, it should be
easier to correlate a particular fracture face v;ith the cell wall
component which is being studied. In whole cells only those surfaces
which are natural fracture sites can be seen. Using isolated cell
envelopes it should be possible to observe other surfaces, such as
the inner surface of the cytoplasmic membrane, by merely etching away
the ice. In the case of marine bacteria there is an additional
advantage. Cell envelopes can be frozen in distilled water or dilute
salt solutions thereby reducing the tendency to form eutectic layers
which obscure etch surfaces.

Complete cell envelopes were prepared by lysing the bacteria In
a French pressure cell. Examination of thin sections revealed that a
variety of different structures viere produced in this lysis process
(Figs. 30 and SO- The majority of the envelopes v/ere double-membrane
structures which were either open C-shaped fragments or closed vesicles,
Single membrane vesicles were also seen and it was not possible to
determine if these were formed from the cytoplasmic membrane or the
outer membrane. As in whole cells, the rigid layer v/as not visible.

The various structures were also seen in uranyl acetate stained
envelopes (Fig. 32). Similar results were obtained with potassium
phosphotungstate and ammonium molybdate negative stains. No subunit
structure was observed in any of these preparations.

Freeze-etchi ng of these complete envelopes yielded little
additional information. The cytoplasmic membrane fracture faces and
the globular layer were observed, but a variety of other surfaces
could not be identified (Figs. 33 to 36). Many of the envelope frag-

Figs. 30 and 31. Thin sections of isolated cell envelopes prepared by

lysing bacteria in a French pressure cell. (x 127,000)

Fig. 32. Isolated cell envelopes positively stained with uranyl
acetate, showing double membrane fragments of various
sizes. (x 151,000).

66

f^igs. 33, 3'*, 35, and 36. Freeze-etched isolated cell envelopes.

Envelopes were prepared by lysing bacteria
in a French pressure cell and suspended in
O.OIM MgCl .

Fig- 33. This fragment shows a typical cytoplasmic

membrane fracture face and the outer surface
exposed by etching. (x 116,000).

Fig. 34. An unfractured fragment exposed by etching,
(x 58,500).

Fig. 35- A concave fracture exposing the globular
layer. (x 58,500).

Fig. 36- This fragment shows the globular layer
apparently unfractured and exposed by
etching alone. (x 157,500).

68

4

69

ments v.'ere unfractured and revealed by etching (Fig. 35). Tiie enve-
lopes were suspended in 0.01M MgCK rather than in complete salts to
minimize the formation of eutectic layers. The true outer surface of
the envelopes appeared smooth or finely granular. The granularity
may represent the true structure of the surface or it may be an artifact
of very low angle shadowing since it is only seen in areas where the
surface is sloping away from the direction of shadow.

The true inner concave surface of the cytoplasmic membrane should
be revealed by etching but v;as not recognizable. Surprisingly, the
concave globular layer was exposed by etching (Fig. 36). This Indicates
that this layer separated from the rough surface during the preparation
of the envelopes, and therefore could not be an internal fracture
surface of a membrane.

Cell walls were prepared by lysing cells with Triton X-100.
Schnaitman (59) has shown that in the presence of magnesium this deter-
gent solubilizes the cytoplasmic membrane leaving only slightly altered
cell walls. The f reeze-etched appearance of these cell walls was very
complex and difficult to interpret. During preparation the walls appar-
ently packed together and flattened out, and possibly turned inside
out, making it difficult to recognize concave and convex surfaces.

A variety of surfaces were observed (Figs. 37 to kO) . As expected,
no typical cytoplasmic membrane fracture faces were seen. Many of the
surfaces were composed of large circular structures which may have been
flattened vesicles (Fig. 38). Since the cytoplasmic membrane Is absent,
the inner surface of the rigid layer should be exposed. This surface
was seen superimposed on the globular layer (Figs. 39 and kO) .

Figs. 37i 38, 39i and 40. Freeze-etched cell v.'alls. The walls were

prepared by lysing cells with Triton X-100
and suspended in O.OlH MgCl . (x 90,000).

Fig. 37- This fragment has a smooth surface and shows

circular structures (arrows) of unknown nature,

Fig. 38. Cell walls shov-^ing large circular structures
which are probably flattened vesicles.

Figs. 39 and kO. Fractured cell walls showing the globular

layer and adjacent layers (arrows). it is
not apparent whether these are concave or
convex surfaces.

71

72

When these walls were partially digested with lysozyrr,e they had
a different appearance. Concave surfaces v;ere observed which had a
distinct fibrous appearance (Figs. ^+1 , k2, and k3) . These surfaces
were exposed by etching. When this surface was fractured away the
globular layer was revealed (Figs. k2 and 43). This fibrous layer was
not seen in any other type of preparation and probably represents the
partially digested pept i dogl yean .

Forsberg e_t aj_. developed a procedure which allows the removal of
the outer membrane from a marine pseudomonad (22, 23). Application of
this washing procedure to the marine vibrio caused the release of outer
membrane material, but did not markedly affect the viability of the
culture. The solubilized material was collected by ul tracentri fugation,
and the pellet obtained was a clear gelatinous material unlike the
white, easily resuspended cell wall material produced in the other
procedures.

Freeze-etching of this material revealed variously shaped vesicles
which were seen in cross section and as concave and convex fracture
faces (Figs, hk and ^+5) . Slight etching exposed an edge around the
convex faces demonstrating that these faces are produced by fracturing.
Both of the fracture faces were smooth and appeared identical to the
smooth faces seen in whole cells. A few patches of globular layer were
also observed in these preparations (Fig. kk) . These layers were
generally larger than the membrane vesicles and appeared flatter and
more cell shaped, suggesting that they v;ere part of a larger fragment
of cell wall or had some innate structure which prevented them from
forming vesicles.

Figs. ^I, ^2, and h3 . Freeze-etched cell walls prepared by lysing cells

with Triton X-lOO and partially digesting with
1 ysozyme .

Fig. M. Note the fibrous nature of the concave surface

(arrows) and the patch of ice which indicates that
this is an etch surface. (x 66,000).

Fig. k2. This fracture shov/s the relationship of the

globular layer and the concave fibrous layer.
(x 66,000).

Fig. k3- In this area the fibrous material is emerging

from the ice background (arrow) indicating that
this is an etch surface. (x 100,000).

Ih

Fig. kk. Freeze-etched isolated outer membrane material suspended in

O.OIM MgCl«. Cross fractured vesicles and the smooth concave
and convex fracture faces are shown. Etching has revealed
the outer surface of a vesicle (arrow). A large fragment of
the globular layer is also shown. (x 86,000).

Fig. kS- Fractured outer membrane vesicles showing the fracture face
and unfractured surface (arrow). (x 135,000).

76

>v TV*'^-^'**' ^"^

77

Localization of LPS v/ith ferritin conjugated antibody. --Three
different antigen preparations v/ere used to produce antisera. The
activities of the antisera were determined by agglutination of whole
cells. The highest titer was produced by injecting whole cells, but
it was found that adsorbing this antiserum v/ith purified LPS would
reduce the agglutination activity by only 50%. Repeatedly adsorbing
with LPS had no further effect. Only the antiserum produced by in-
jecting partially purified cell walls was specific for LPS. This anti-
serum was of lower titer than the others, but its activity could be
completely adsorbed with LPS and it was used for all labelling experi-
ments.

Cells were labelled by the indirect method using anti-LPS anti-
serum followed by ferr i ti n-conj ugated ant i -robbi t immunoglobulin anti-
serum. The specificity of the labelling was determined by treating
cells with anti-LPS antiserum or LPS adsorbed anti-LPS antiserum
followed by ferr i ti n-conj ugated antiserum. Cells v/ere also treated
with ferri ti n-conj ugated antiserum alone. The cells were examined in
the electron microscope without staining and the degree of labelling
was determined. The results are shov/n in Table 2. Neither unfixed nor
glutaraldehyde fixed cells were labelled by ferr i ti n-conj ugated anti-
serum alone, and they were only slightly labelled by LPS adsorbed anti-
serum. Using unadsorbed antiserum the fixed cells were more heavily
labelled than the unfixed cells. This may have been caused by loss of
labelled LPS from the cell surface during the washing procedure.

Thin sections of labelled cells revealed that the ferritin v/as
localized in a band external to the outer double-track layer (Figs.
^S and 47). in cells v;hich were not fixed before 1 abel 1 i ng, the ferritin

78

Table 2. Determination of the specificity of antibody labelling.

Cells Treatment Result

Unfixed cells Anti-LPS then ferri tin-conjugated Labelled

ant i serum

Unfixed cells LPS adsorbed anti-LPS Unlabelled

then ferr it in-conjugated antiserum

Unfixed cells Ferr i t in-conjugated antiserum only Unlabelled

Glutaraldehyde Anti-LPS then ferr i t in-conjugated Labelled
fixed cells antiserum

Glutaraldehyde LPS adsorbed anti-LPS then Unlabelled

fixed cells ferri t in -conjugated antiserum

Glutaraldehyde Ferr i t i n-conjugated antiserum only Unlabelled
fixed eel 1 s

Figs. ^6 and kl . Thin sectioned cells labelled with ferritin. The cells

were treated with rabbit anti-LPS antiserum, washed,
and then labelled with ferr i t in-conjugated goat anti-
rabbit antiserum.

Fig. it6. This cell was fixed following the antibody treatments.
(x 80,000).

Fig. ^7. This cell was fixed in g lutara Idehyde before antibody
treatments. Note the distance between the ferritin
molecules and the outer double-track. (x 167,000).

80

-CM

-ferritin

■<rt

J:

-.m-:

• - I"

y%

I

46.

:^

47

81

was limited to small areas. The labelling was heavier and more uniform
in fixed cells. In all cases the ferritin v;as a considerable distance
from the outer double-track. This is similar to the results of Shands
(61), and was probably due to the fact that the polysaccharide side
chains extend out from the cell surface and are not stained by the
heavy metal stain. The use of indirect labelling puts two antibody
molecules betvyeen the antigen and the ferritin molecules which could
account for a 20 to 50 nm space.

For f reeze-etchi ng, ferritin labelled cells were suspended in
0.05M MgCl . It was found that this solution would maintain the
integrity of the cells and still allow the deep etching needed to
expose the cell surface. The presence of ferritin confirmed that the
outer surface of the cells v/as revealed by etching. It was difficult
to compare the appearance of the ferritin in thin sections and in
f reeze-etchi ng. All of the particles in f reeze-etched cells were lying
close to the surface of the cell. It is possible that during the
freezing process the ferritin molecules are excluded from the growing
ice crystals thus packing them down onto the cell surface.

With unfixed cells the ferritin molecules were clustered in
patches on the cell surface (Figs. k8 to 50- The envelope fractured
normally revealing the convex cytoplasmic membrane and outer membrane
fracture faces (Figs. 50 and 51).

Glutaraldehyde fixed cells were heavily labelled (Fig. 52). The
ferritin was evenly distributed on the cell surface and many individual
ferritin molecules were seen. In unfixed cells the LPS was freely
mobile in the membrane and v/as probably drawn together by the divalent

Figs. ^8 and ^9- Freeze-etched unfixed cells labelled with ferritin and

suspended in O.OIM MgCl_, The LPS on the cell surface
was labelled indirectly with ferr i t in-conjugated anti-
body. The cell surface was exposed by deep etching.
The ferritin molecules are clustered in patches.
(x 92,000).

83

Figs. 50 and 51- Convex fractured unfixed cell labelled with ferritin.

Note the relationship of the smooth fracture face to
the ferritin labelled surface. (x 92,000).

85

Fig. 52. Ferritin labelled glutara Idehyde fixed cell. The ferritin

molecules are more evenly distributed than in unfixed cells.
Note the relationship of the smooth convex fracture face and
the ferritin covered surface exposed by etching. The
fracture edge is quite thin (arrov/) . (x 157,000).

87

ferritiiv-

52

88

antibody. Glutaraldehyde fixation stabilized the membrane and pre-
vented the LPS from moving. The relationship between the smooth
fracture face and the ferritin labelled outer surface v/as clearly
shown by these experiments (Fig. 52).

DISCUSSION

It was shown in this study that the cell envelope of the marine
vibrio MW40 f reeze-f ractures in three planes. The relationships be-
tween the various fracture faces were proven by double-replica tech-
nique. It has been firmly established in previous studies (65, 71)
that one fracture splits the cytoplasmic membrane, and this interpreta-
tion is consistent with the results of this study. Lickfeld e_t al .
(32) recently offered an alternate explanation, but it appears to be
untenable.

Localizing the other two fracture planes is more difficult. As
described in the Results section, these fractures are interpreted as
follows. One fracture splits the outer membrane revealing smooth
convex and concave faces. The second fracture splits the cell wall
at an intermediate level exposing a rough convex face and a concave
face which is composed of globular subunits superimposed on a smooth
surface.

The smooth fracture faces were seen primarily when the cells were
f reeze-etched without glycerol, but v;ere demonstrated in glycerol-
treated cells. The rough convex face was only seen i n glycerol -treated
cells, but the globular concave face was observed with and without
glycerol. Although the two fracture planes have not been clearly
demonstrated in the same cell, they have been observed in different
cells in the same replica. It can thus be concluded that while the

89

90

presence or absence of glycerol tends to determine the location of
the fracture, it does not alter the appearance of the fracture faces
as others have implied (26, 71). The action of glycerol as it affects
the fracturing process is not known.

If it is accepted that there are two distinct fracture planes in
the cell v/all, it must be decided where the fractures occur. It is
logical to assume that if the outer membrane is a lipid bilayer, it
will fracture along its hydrophobic center as has been demonstrated
for other membranes. In other studies only the globular-rough fracture
plane v/as observed and it was postulated that this represents the
splitting of the outer membrane (26, 32, 71). In view of our results
thi s seems unl i kely.

A survey of membrane fracture faces in a variety of natural and
artificial membranes reveals that although they may be studded with
particles, the portion of the face which associated with the lipid is
generally smooth. The smooth cell wall fracture faces would therefore
fit this description while the rough convex face would be atypical.

The globular layer could represent an unusual type of membrane
face, but the dimensions of the layer seen to preclude this. The sub-
units are 10 nn in diameter, and our results show that they are nearly
spherical. The outer double-track measures about 7-5 nni and could not
accommodate a layer of 10 nm globules. While the double-track does
not represent the entire thickness of the outer membrane, it does
reflect the size of the lipid portion between which the globular layer
is supposed to be located. A similar problem exists in explaining the
particles seen in most membranes, and it has been proposed that these
particles are intercalated into the lipid layer. The globular layer.

9]

however, is a continuous surface, and such an explanation v^ould require
that the entire outer half of the outer menbrane be composed of the
globular units. This seems to be unlikely, and the demonstration of
a smooth surface in areas where the subunits are missing tends to
disprove it.

The relative location of the two fracture planes can be inferred
from the thickness of the fracture edges, although it is difficult to
actually measure these edges. Only a thin layer is fractured away to
expose the smooth convex face, and this is consistent with the belief
that this fracture removes the outer half of the outer membrane. The
fracture edge down to the rough convex face is thicker, indicating that
it represents a layer in the middle of the cell envelope.

It was observed in this study that the globular layer is sometimes
obscured by a covering of ice. A similar result was obtained in
complementary replicas of E. col i . in which it was observed that the
suspending medium could intrude between the globular and rough faces
(71). This could not occur if these faces were the hydrophobic portion
of a membrane. The demonstration that in isolated cell envelopes the
globular layer can be exposed by etching alone clearly proves that
this face is not a membrane fracture face.

The observation of smooth fracture faces in isolated outer membrane
material also supports our interpretation. A similar study was re-
ported by Forge et aj_. (21) but the results and conclusions are very
vague and can not be directly compared to our results.

Little direct evidence has been offered in support of the alternate
interpretation. GI 1 leland et a_I^. (26) stated that in cells showing both
cross-fractured envelope and surface fractures, the rough convex face

92

could be seen to originate from the inner track of the outer double-
track, Lickfeld et aj_, (33) demonstrated that the cross-fractured
outer (T.embrane does have a double-track appearance in some areas,
although there is no plausible explanation for this. They also showed
that the cross-fractured cell vjal i is composed of two layers, that is,
the outer layer and the rigid layer (similar to Fig. 26), It appears
probable that the double-track layer observed by Gilleland et aj_. is
actually two layers, and the rough face corresponds to the inner of
these layers. This layer is probably the rigid layer.

Van Gool and Nanninga (71) observed that the globular-rough
fracture v;as probably located in the outer membrane since the rough
face was separated from the cytoplasmic membrane by an intermediate
layer which they believed was the rigid layer. Careful examination
of their results indicates that this layer is actually the unfractured
surface of the cytoplasmic membrane.

In the sane study it was stated that no cell wall fracture plane
v;as present in lysozym.e produced spheroplasts. It seems unlikely that
digestion of the pept i doglycan layer would eliminate the globular-
rough fracture plane if it was located in the outer membrane. DeVoe
et aj_' (^7) reported a smooth fracture face in spheroplasts of a marine
pseudomonad f reeze-etched without glycerol. There was no cell wall
fracture plane in cells vjhich lacked the outer membrane.

It therefore appears that the majority of the results of this and
other studies are compatible with our interpretation. From the freeze-
etching work a model of the cell envelope of the marine vibrio can be
proposed (Fig. 53). This illustrates the location of the fracture
planes and relationship of the various faces.

Fig. 53- Diagrammatic representation of the cell envelope of the marine
vibrio MW40 as determined by f reeze-etch ing . The locations
of the fracture planes are illustrated.

9^

95

The cytoplasmic membrane fractures exposing particle-studded faces.
The true surfaces of the membrane are only seen as edges. The particles
are believed to represent proteins which are a part of the membrane (7).
In bacteria there has been no direct evidence to substantiate this con-
clusion, and on one had demonstrated any relationship between the nature
of the particles and the physiological state of the cells. Large
particle-free areas were sometimes present on both fracture faces.
Similar areas have been reported in other studies {kk, 65). These
patches were larger and more numerous in cells v^hich were subjected to
the minimum amount of handling before freezing. This Indicates that
these areas are normally present in actively grov;ing cells and during
storage the membrane rearranges.

The second fracture separated the cell v/all exposing the rough
convex face of the rigid layer and a concave face composed of globular
subunits. In areas v;here the globular layer is incomplete a smooth
surface is revealed v/hich is probably the true inner surface of the
outer membrane. This fracture may occur in a hydrophobic plane formed
by the lipoprotein attached to the pept idoglycan and the protein layer
superimposed on the inside of the outer membrane. Such a hydrophobic
association was postulated by Schnaitman (60) .

The organism used In this study does not have a covalently linked
lipoprotein on the peptidoglycan , at least during exponential grovjth.
It is not knovyn whether the lipoprotein is entirely absent or present
but not covalently linked. In either case, this may account for the
fact that this fracture plane is not as prevalent as in E^. col i or
Pseudomonas. This fracture v/as not seen In a marine pseudomonad (I7)
which also lacks a lipoprotein {2k).

96

dePetris (16) and Thornley and Glauert (69) demonstrated a
globular layer between the rigid layer and the outer membrane, and
found that this material was eliminated by digesting the cells vnth
proteolytic enzymes. Fischman and Weinbaum (20) studied negatively
stained E. col i cell walls and observed a regular hexagonal array of
globular units. This array was removed by proteolytic enzymes. The
subunits were 13 to ]k nm in diameter which is somewhat larger than
the subunits seen in f reeze-etching.

Schnaitman (58) examined the protein content of the cell wall of
_E. col i and found a major protein with a molecular weight of approximate-
ly 'f4,000. The size and amount of this protein would make it a likely
prospect for the subunits of the globular layer. These protein are
quite hydrophobic and could be postulated to bind the rigid layer to
the outer membrane. The proteins released from _P. aerugi nosa by EDTA
extraction have been directly associated with the units seen in freeze-
etching (26). These units are probably comparable to the globular
subunits seen in this study.

In the marine pseudomonad (23) the entire outer membrane is
removed by treatments v^hich eliminate the magnesium ions from the cell
envelope. In this procedure a protein is also released. These results
are consistent v;i th our interpretation of the work with _P. aerugi naosa
and may indicate that one role of divalent cations is in binding the
globular layer to the rigid layer. E^. col i is not markedly affected
by chelating agents, and this binding may be of less importance in this
organism since it has a more continous protein layer and a much larger
number of lipoprotein molecules bound to its peptidoglycan (9).

The third fracture splits the outer membrane revealing smooth

97

fracture faces. It is thought that this membrane is very low in
enzymatic activity and this may be reflected by the absence of
particles on the fracture faces. The fracturing of the outer membrane
also indicates that it is probably a lipid bi layer with a hydrophobic
center as has been assumed.

The patches of cell wall material seen in some preparations were
also reported by Bayer and Remsen {k) . These patches may correspond
to the areas of adhesion between the cell wall and the cytoplasmic
membrane observed in E. col i (3).

The true outer surface of the outer membrane is exposed by etching.
This surface is smooth or finely granular. In other studies a subunit
structure was reported on the cell surface {k, 19, 26).

In summary, it appears that the technique of f reeze-etch i ng is a
valuable tool in studying the structure of the cell envelope. Although
further work is needed to clearly prove the location of the fracture
planes and new techniques must be developed to correlate the chemical
nature of the envelope components with the structures seen in freeze-
etching, this study has demonstrated several important points. The
location of the fracture plane in the cytoplasmic membrane and the
nature of the faces v-;as confirmed. The existence of a globular layer
in the cell wall was confirmed and this layer was localized in the
middle of the cell wall, probably forming a fracture plane with the
rigid layer. It v;as shown that the outer membrane fractures and the
fracture faces are free of particles. Although the work was not
completely successful, it was found that the technique could be
applied to Isolated cell envelope components. And finally It was
demonstrated that ferritin labelled antiserum could be used to localize

98

a component on the cell using f reeze-etchi ng. This technique may
prove very useful in localizing other envelope components which are
present in a less continuous layer on the cell surface.

LITERATURE CITED

1. Bauniann , P., L. Baumann, and M. Mandel. 1971- Taxonomy of marine
bacteria: The genus Beneckea . J. Bacteriol. 107:268-29^-

2. Baumann, P., L. Baumann, M. Mandel, and R. D. Allen. 1971-
Taxonomy of marine bacteria: Beneckea n igrapul chr i tuda sp. n.
J. Bacteriol. 1 08 : 1 38O- 1 383 .

3. Bayer, M. E. 1 968 . Areas of adhesion between wall and membrane
of Escherichia col i . J. Gen. Microbiol. 53:395-^04.

k. Bayer, M. E. , and C. C. Remsen. 1970. Structure of Escher ich ia
col i after f reeze-etching. J. Bacteriol. 101:30^-313-

5. Bladen, H. A., and S. E. Mergenhagen. 1964. U 1 trastructure of
Ve i 1 lonel la and morphological correlation of an outer membrane
with particles associated with endotoxic activity. J. Bacteriol.
88:1482-1492.

6. Branton, D. I966. Fracture faces of frozen membranes. Proc.
Nat. Acad. Sci. U.S.A. 55:1048-1056.

7. Branton, D. 1969- Membrane structure. Ann. Rev. Plant Physiol.
20:209-238.

8. Braun, V., and V. Bosch. 1972- Sequence of the mureIn-1 ipoprotein
and attachment site of the lipid. Eur. J. Biochem. 28:51-69.

9. Braun, V., H. Gnirke, U. Henning, and K. Rehn. 1973- Model for
the structure of the shape-maintaining layer of the Escher i chia
col i cell envelope. J. Bacteriol. 114:1264-1270.

10. Braun, V., and K. Rehn. 1969- Chemical characterization, spatial
distribution, and function of a lipoprotein (mure i n- 1 i poprote i n)
of the Escherichia col i cell wall. The specific effect of trypsin
on the membrane structure. Eur. J. Biochem. 10:426-438.

11. Braun, V., K. Rehn, and H. Wolff. 1970. Supramolecu lar structure
of the rigid layer of the cell v^a 1 1 of Salmonel la, Ser rat ia,
Proteus , and Pseudomonas fluorescens. Number of lipoprotein
molecules in a membrane layer. Biochemistry 9:5041-5049-

99

100

12. Braun, V. , and N. Sleglin. 1970. The covalent mureIn-1 ipoprote in
structure of the Escherichia col i eel 1 wall. Eur. J. Biochem.
13:336-3^6.

13- Braun, V., and H. Wolff. 1970. The murein-l ipoprotein linkage in
the cell wall of Escherichia col i . Eur. J. Biochem. l'*:387-39].

lA. Costerton, J. W. 1970. The structure and function of the cell
envelope of gram-negative bacteria. Rev. Can. Biol. 29:299-316.

15- de Petris, S. 1965. Ul trastructure of the cell wall of
Escherichia col i . J. Ultrastruc. Res. 12:247-262.

16. de Petris, S. 1967- Ul trastructure of the cell wall of
Escher ichia col i and chemical nature of its constituent layers.
J. Ultrastruc. Res. 19:^5-83.

17. DeVoe, I. W. , J. W. Costerton, and R. A. MacLeod. 1971.
Demonstration by f reeze-etch ing of a single cleavage plane in
the cell wall of a gram-negative bacterium. J. Bacteriol. 106:
659-671.

18. Dische, Z. 1953- Qualitative and quantitative colorimetric
determination of heptoses. J. Biol. Chem. 20'4:983-997.

19- Fiil, A., and D. Branton. 1969- Changes in the plasma membrane
of Escherichia col i during magnesium starvation. J. Bacteriol.
98:1320-1327.

20. Fischman, D. A., and G. Weinbaum. 1967. Hexagonal pattern in
cell walls of Escher ichia col i 8. Science 155:472-47'*.

21. Forge, A., J. W. Costerton, and K. Annkerr. 1973- Freeze-
etching and X-ray diffraction of the isolated double-track layer
from the cell wall of a gram-negative marine pseudomonad. J.
Bacteriol. 113:445-451.

22. Forsberg, C. W. , J. W. Costerton, and R. A. MacLeod. 1970.
Separation and localization of the cell wall layers of a gram-
negative bacterium. J. Bacteriol. 104:1338-1353.

23. Forsberg, C. W. , J. W. Costerton, and R. A. MacLeod. 1970.
Quantitation, chemical characteristics, and ul trastructure of the
three outer cell wall layers of a gram-negative bacterium. J.
Bacteriol. 104:1354-1368.

24. Forsberg, C. W. , M. K. Rayman, J. W. Costerton, and R. A. MacLeod.
1972. Isolation, characterization and ul trastructure of the

pept idoglycan layer of a marine pseudomonad. J. Bacteriol. 109:
895-905.

101

25- Ghuysen, J. M. 1968. Use of bacteriolytic enzymes in determina-
tion of wall structure and their role in cell metabolism.
Bacterid . Rev. 32:425-^64.

26. Gilleland, H. E., Jr., J. D. Stinnett, I. L. Roth, and R. G. Eagon.
1973. Freeze-etch study of Pseudomonas aeruginosa: Localization
within the cell wall of an ethylened iaminetetraacetate - extractable
component. J. Bacteriol. 113:417-432.

27. Glauert, A. M., and M. J. Thornley. I969. The topography of the
bacterial cell wall. Ann. Rev. Microbiol. 23:159-198.

28. Hess, W. M., R. L. Bair, and M. Neushul . 1972. Production and
cleaning of freeze-etch replicas v/hich shov; complementary surfaces
of fractured fungus spores and hyphae. Stain Tech. 47:249-255.

29. Hofshneider, P. H., and H. H. Martin. I968. Diversity of surface
layers in L-Forms of Proteus mirabi lis. J. Gen. Microbiol. 51:
23-33.

30. Horecker, B. L. 1966. The biosynthesis of bacterial polysaccharides.
Ann. Rev. Microbiol. 20:253-285.

31. Inouye, M. , and M. L. Yee. 1973- Homogeneity of envelope proteins
of Escherichia col i separated by gel electrophoresis in sodium
-odecyl sulfate. J. Bacteriol. 113:304-312.

32. Lickfeld, K. G., M. Achterrath, F. Hentrich, L. Kol eha inen-seveus ,
and A. Persson. 1972. Die Fe instrukturenn von Pseudomonas
aeruginosa in ihrer Deutung durch die Gef r ieratztechni k,

Ul tramikrotomie, und Kryou 1 t rami krotom ie. J. Ultrastruct. Res.
38:27-45.

33- Lickfeld, K. G., M. Achterrath, and F. Hentrich. 1972. The
interpretation of images of cross fractured frozen-etched and
shadowed membranes. J. Ultrastruct. Res. 38:279-287.

34. Luderitz, 0., M. Staub, and 0. Westphal. I966. Immunochemi st ry
of 0 and R antigens of Sa Imonel la and related Enterobacter iaceae .
Bacteriol. Rev. 20:192-255.

35. Martin, H. H. I966. Biochemistry of bacterial cell walls, Ann.
Rev. Biochem. 35, pt. 2:457-484.

36. Martin, H. H. , H. D. Heilmann, and H. J. Preusser. 1972. State
of the rigid-layer in cell walls of some gram-negative bacteria.
Arch. Mikrobiol. 83:332-346.

37- Mergenhagen, S. E., H. A. Bladen, and K. C. Hsu. I966. Electron
microscopic localization of endotoxic 1 i popolysacchar ide in gram-
negative organisms. Ann. N.Y. Acad. Sc i . 133:279-291.

102

38. Mollenhauer, H. H. 196^- Plastic embedding mixtures for use in
electron microscopy. Stain Technology 39:111~11'*-

39. Moor, H. 1969- Freeze-etching. Int. Rev. Cytology 25:391-^12.

kO. Moor, H., and K. Muhlethaler. 1963- Fine structure in frozen-
etched yeast cells. J. Cell. Biol. 17:609-628.

^1 . Muhlethaler, K. 1971. Studies on freeze-etching of cell membranes.
int. Rev. Cytology 31:1-22.

^42. Muhlethaler, K. , E. Wehrli, and H. Moor. 1970. Double fracturing
methods for freeze-etching. Proc. Septifeme Congres International
de Microscopie El ectron ique. Grenoble 1:4^9-^^50.

i»3, Murray, R. G. E. , P. P. Steed, and H. E. Elson. 1965- The
location of the mucopeptide in sections of the cell wall of
Escherichia coli and other gram-negative bacteria. Can. J.
Microbiol. 11:5^7-560.

kk. Nanninga, N. 1970. U 1 trastructure of the cell envelope of
Escher ichia col i B after f reeze-etch ing. J. Bacterid. 101:
297-303.

45- Nanninga, N. 1971- Uniqueness and localization of the fracture
plane in the plasma membrane of Baci 1 1 us subt i 1 lus. J. Cell.
Biol. ^49:564-570.

A6. O'Leary, G. P., J. D. Nelson, and R. A. MacLeod. 1972. Require-
ment for salts for the isolation of 1 i popolysacchar i des from a
marine pseudomonad. Can. J. Microbiol. l8:601-606.

47. Osborn, M. J. I969. Structure and biosynthesis of bacterial
cell wall. Ann. Rev. Biochem. 38:501-538.

48. Osborn, M. J., J. E. Gandn, E. Parisi, and J. Carson. 1972.
Mechanism of assembly of the outer membrane of Salmonel la
typhimurium. J. Biochem. 247:3962-3972.

49. Pinto da S i 1 va , P., and D. Branton. 1970. Membrane splitting in
freeze-etching. J. Cell. Biol. 45:598-605-

50. Remsen, C. C, and S. W. Watson. 1972. Freeze-etching of bacteria.
Int. Rev. Cytology 33:253-296.

51. Reynolds, E. S. I963. The use of lead citrate at high pH as an
electron-opaque stain in electron microscopy. J. Cell. Biol.
17:208-212.

52. Romeo, D. , A. Girand, and L. Rothfield. 1970. Reconst i tut ion of
a functional membrane enzyme system in a monomolecular film.

I. Formation of a mixed monolayer of 1 i popolysacchar ide and
phospholipid. J. Mol. Biol. 53:475-490.

103

53. Salton, M. R. J. ]SGk. The bacterial cell wall. Elsevier
Publishing Co., Amsterdam.

5^. Salton, M. R. J. 1967- Structure and function of bacterial cell
membranes. Ann. Rev. Microbiol. 2\ ■.k]'J-kk2.

55. Salton, M. R. J. 1971- The bacterial membrane, in "Biomembranes 1"
(L. A. Manson, ed.) p. 1-65, Plenum Press, New York.

56. Schlcifer, K. H. , and 0. Kandler. 1972. Pept i dogl yean types of
bacterial cell walls and their taxonomic implications. Bacteriol.
Rev. 3S:kO-]'kn.

57. Schnaitman, C. A. 1970. Examination of the protein composition
of the cell envelope of Escher ichia col i by pol yacry 1 am ide gel
electrophoresis. J. Bacteriol. 10^1:882-889.

58. Schnaitman, C. A. 1970. Protein composition of the cell wall
and cytoplasmic membrane of Escher ich ia col i . J. Bacteriol. 104:
890-901.

59- Schnaitman, C. A. 1971- Solubilization of the cytoplasmic

membrane of Escher ichia col i by Tr i ton X-1 00. J. Bacter iol . IO8:
545-552.

60. Schnaitman, C. A. 1971- Effect of ethylened iaminetetraacet ic
acid, Triton X-100, and lysozyme on the morphology and chemical
composition of isolated cell walls of Escher ichia col i . J.
Bacteriol. 108:553-563.

61. Shands, J. W. I965. Localization of somatic antigen on gram-
negative bacteria by electron microscopy. J. Bacteriol. 90:268-270.

62. Shands, J. W. , I. W. Graham, and K. Nath. 1967- The morphologic
structure of isolated bacterial 1 ipopol ysacchar ide. J. Mol . Biol.
25:15-21.

63. Shewan, J. M., G. Hobbs, and W. Hodgkiss. I960. A determinative
scheme for the identification of certain genera of gram-negative
bacteria, with special reference to the Pseudomonadaceae. J.
Appl. Bacteriol. 23:379-390.

64. Singer, S. J., and G. L. Nicolson. 1972. The fluid mosaic model
of the structure of cell membranes. Science 175:720-731-

65. Sleytr, V. B. 1970. Fracture faces in intact cells and protoplasts
of Bacillus stearothermophi 1 us - A study by conventional freeze-
etching and f reeze-etch ing of corresponding fracture moieties.
Protoplasma 71 :295-312.

66. Spackman, D. H. , W. H. Stein, and S. Moore. 1958. Automatic
recording apparatus for use in the chromatography of amino acids.
Anal. Chem. 30:1190-1206.

1 04

67- Steere, R. L. 1969- Freeze-etching and direct observation of
freezing damage. Cryobiology 6:137"150.

68. Stinnett, J. D., H. E. Gilleland, Jr., and R. G. Eagon. 1973-
Proteins released from cell envelopes of Pseudomonas aeruginosa
on exposure to ethylenediaminetetraacetate: Comparison with
dimethylformamide - extractable proteins. J. Bacteriol. ]\k:
399-^07.

69. Thornley, M. J., and A. M. Glauert. I968. Fine structure and
radiation resistance in Ac inetobacter : Studies on a resistant
strain. J. Cell. Sci. 3:273-29^-

70. Van Gool , A. P. 1972. Ul trastructure of Nitrosononas europaea
cells as revealed by freeze-etching. Arch. Mikrobiol. 182:120-
127.

71. Van Gool, A. P., and N. Nanninga. 1971. Fracture faces in the
cell envelope of Escherichia col i . J. Bacteriol. 108:^47^-^81.

72. Watson, S. W. , and C. C. Remsen. 1970. Cell envelope of
Ni trosocyst i s oceanus. J. Ultrastruc. Res. 33:1^8-160.

73. Weidel, W. , H. Frank, and H. H. Martin. I96O. The rigid layer

of the cell wall of Escherichia col i strain B. J. Gen. Microbiol.
22:158-166.

7^. Weidel, W. , and H. Pelzer. 196^. Bagshaped macromolecules - a

new outlook on bacterial cell walls. Advan. Enzymol . 26:192-232.

75. Weissbach, A., and J. Hurwitz. 1959- The formation of 2-keto-3-
deoxyheptonic acid in extracts of Escher ichia col i . i. identifica-
tion. J. Biol. Chem. 23^:705-709.

76. Westphal, 0., 0. Luderitz, and F. Bister. 1952. Extraction of
bacteria with phenol/water. Z. Naturfursch 7b: 1^8-155.

BIOGRAPHICAL SKETCH

Jack Thornton Crawford was born August 7, IS^'Sj in Hamilton, Ohio.
He graduated from Hamilton Taft High School in June, 1363- The follow-
ing fall he entered Bowling Green State University and graduated from
the Ohio State University in June, 1967, with a Bachelor of Science
degree in microbiology. In September, 1967, he entered the University
of Florida and held a graduate teaching ass istantship from September,
1967, until August, 1972, and a graduate teaching assi stantship from
September, 1972, to June, 1973-

He is a member of the American Society for Microbiology and the
Southeastern Branch of the American Society for Microbiology. At
present, he is a candidate for the Ph.D. degree in microbiology.

Mr. Crawford is married to the former Kay Woodten and they have
two sons - Brent and Gregory.

105

I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and Is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.

Max E, Tyler, Chai rm.an
Professor of Microbiology

I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy. ' = ~i:'- '.' :

rtv

C' [i(J^,.

'. c

Henry C. /Aldri ch

Associate Professor of Botany

I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.

C Arnold S. Bleiweis
Associate Professor of Microbiology

I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholar4Y presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.

Cl^/-^

el Bi 1 len
Professor of Radiation Biology

This dissertation was submitted to the Graduate Faculty of the College
of Agriculture and to the Graduate Council, and was accepted as partial
fulfillment of the requirements for the degree of Doctor of Philosophy.

December, I973

Qui*4J •

<^^

Dean , Co 1 lege of .jP(gri cul ture

u

Dean, Graduate School

I

\

4

.t)

'Y