Skip to main content

Full text of "Microsomal particles and protein synthesis; papers presented at the First Symposium of the Biophysical Society, at the Massachusetts Institute of Technology, Cambridge, February 5, 6, and 8, 1958"

See other formats

Microsomal Particles and 
Protein Synthesis 


Marine Biological Laboratory Library 

Woods Hole, Massachusetts 

Microsomal Particles and 
Protein Synthesis 

Papers presented at the First Symposium of the 
Biophysical Society ', at the Massachusetts Institute 
of Technology j Cambridge, February 5, 6, and 8, 1958 

W-, , , 




Biophysical Society 

Microsomal Particles and 
Protein Synthesis 


Published on behalf 
of the 






© Richard B. Roberts 1958 

Library of Congress Catalog Card Number 58-13658 




The formation of a new society provides an occasion for innovations in the 
forms of meetings and publications. The Program Committee of the Bio- 
physical Society attempted to seize this opportunity in arranging the 1958 meet- 
ing at Cambridge. In addition to the usual short contributed papers, review 
papers were scheduled for the three afternoon sessions to inform the member- 
ship of progress and problems in selected areas of biophysics. The contributed 
papers in these areas were allocated ample time for presentation and discus- 
sion, resulting in what might be called "contributed symposia." Those work- 
ing in the field had the advantages of a symposium type of meeting; those less 
familiar had an introduction provided by the review papers followed by an 
opportunity to participate in the detailed technical sessions. 

It also appeared desirable to alter the usual publication procedures. Custom- 
arily the complete proceedings of "invited symposia" are published but only 
the abstracts of short contributions. The twenty papers dealing with micro- 
somal particles normally would be published individually during the next two 
years, scattered through different issues of five or ten journals. Alternatively, 
it would be possible to publish a transcript of the symposium. This procedure 
would provide a more complete account than is given in the abstracts; the 
material would be in one volume; and it could be issued much more rapidly 
than individual contributions to journals. Both the Council of the Biophysical 
Society and a majority of the contributors favored publication of the symposium 

Transcripts, however, require an enormous editorial effort. It therefore 
seemed preferable to request the contributors to provide their own edited tran- 
scripts. These transcripts have been accepted with the understanding that pub- 
lication in this volume which records material presented at a meeting would 
not preclude later publication in the usual journals. 

The purpose of this volume is not to present well established theories or 
reviews of well known work. Rather, it is to publish new facts and new data 
while they are still fresh, useful, and possibly wrong. The relative costs of re- 
search time and publication make such a book worth purchasing if it saves 
no more than an hour of research time. The entire publication costs are justi- 
fied if the book can save a month in some research program. 


The Washington Academy of Sciences has agreed to act as publisher in the 
hope of demonstrating that this type of book can be brought out rapidly and 
that it does serve a useful function. 

We wish to thank the Office of Publications of Carnegie Institution of Wash- 
ington for editorial help in seeing the project through the press. 


Chairman, Monograph Committee 
Washington Academy of Sciences 


Chairman, Program Committee 
Biophysical Society 

R. B. ROBERTS, Editor 

Carnegie Institution of Washington 
Department of Terrestrial Magnetism 
5241 Broad Branch Road, Northwest 
Washington 15, D. C. 





The topic "Microsomal Particles and Protein Synthesis" seemed particularly 
appropriate for the first symposium of the Biophysical Society. The particles 
owe their recognition to the electron microscope and the ultracentrifuge. X-ray 
diffraction studies will undoubtedly contribute to the picture of the structure 
of the particles; radioactive tracers show the kinetics of formation of the par- 
ticles and their products; radiation experiments can give other evidence about 
the role of the particles in protein synthesis. Thus many of the special areas 
of competence of biophysicists are involved in the study of the particles. 

More important, however, is the timeliness of a symposium devoted to a 
discussion of these ubiquitous granules. For a number of years circumstantial 
evidence has accumulated which indicates that ribonucleic acid (RNA) is 
implicated in protein synthesis. More recently it has been recognized that a 
large part of the RNA occurs in the form of ribonucleoprotein (RNP) par- 
ticles. Particles of roughly the same size and composition have been isolated 
from sources differing as widely as rat liver, pea seedlings, and microorganisms. 
Accordingly there has developed a widespread faith that the particles are an 
important part of the machinery for protein synthesis. 

It must be admitted, however, that the evidence is entirely circumstantial. A 
number of arguments would come immediately to the mind of a lawyer de- 
fending the particles from the charge of protein synthesis. (1) In vivo experi- 
ments have shown incorporation of radioactive tracers which is initially higher 
in the microsome fraction than in the soluble proteins, but the kinetic data are 
not sufficiently complete to prove a precursor-product relationship. For exam- 
ple, it is possible that steady-state conditions do not prevail; there is no cer- 
tainty that both components draw on the same pool of amino acids; only the 
average of the soluble proteins is measured, whereas individual components 
might behave quite differently. (2) There are many cases both in complete and 
in cell-free systems where RNAase has been observed to inhibit protein syn- 
thesis. In many of these the addition of RNA (not RNP) is sufficient to 
restore the synthetic activity. (3) Cell-free systems showing unequivocal pro- 
tein synthesis use cells that are only partially disrupted, and the requirement 
for particles is not demonstrated. In those systems where the particles have 
been partly purified, the incorporation data are more suggestive of exchange 
than of true protein synthesis. (4) No mechanism has been suggested which 
shows how the structure of the particle is compatible with its function as the 
template for synthesis of long chains. It appears that the particles have not yet 



been proved guilty beyond all reasonable doubt. In the last few years, however, 
there have been marked advances in the study of the particles which promise 
to resolve these lingering doubts. Thus a symposium dealing with the particles 
and their function in living cells could hardly fail to bring forth new and 
exciting information. 

In this symposium a number of papers were concerned with methods of 
isolation, the size, the composition, and the stability of the particles. One strik- 
ing observation was that particles of roughly 80 S are found in a wide variety 
of materials; another area of agreement was in the requirement for magnesium 
to stabilize the particles. There was a consensus of opinion that carefully puri- 
fied particles have little enzymatic activity and that their RNA content is 40 
per cent or more. Several reports showed that the protein moieties of nucleo- 
protein have certain distinctive properties. Other studies explored the reasons 
for the variations in particle sizes that are observed both in vivo and in vitro. 
New kinetic data were presented which indicate that the protein of the par- 
ticles does not serve as precursor material for nonparticulate protein. The in- 
corporation of adenylamino acids was demonstrated in one study which also 
illuminated the need for caution in the interpretation of incorporation studies. 
Other papers reported less direct methods of approach to the understanding of 
the particles and their role in protein synthesis, such as studies of radiation 
effects and studies of incorporation of amino acid analogs. All together these 
reports provide a number of new facts that must be taken into account by any 
theory of protein and nucleic acid synthesis. 

During the course of the symposium a semantic difficulty became apparent. 
To some of the participants, microsomes mean the ribonucleoprotein particles 
of the microsome fraction contaminated by other protein and lipid material; 
to others, the microsomes consist of protein and lipid contaminated by particles. 
The phrase "microsomal particles" does not seem adequate, and "ribonucleo- 
protein particles of the microsome fraction" is much too awkward. During the 
meeting the word "ribosome" was suggested; this seems a very satisfactory 
name, and it has a pleasant sound. The present confusion would be eliminated 
if "ribosome" were adopted to designate ribonucleoprotein particles in the size 
range 20 to 100 S. 

The symposium provided to the participants an opportunity for comparing 
notes on methods and techniques and for exchange of views on the status of 
various problems. It undoubtedly affected the immediate research plans of a 
number of the participants. This volume is being published in the hope that 
it will extend some of these benefits to those who did not attend. 


1. Isolation and Characterization of Bacterial Nucleoprotein Particles — ■ 
William C. Gillchriest and Robert M. Boc/{ 1 

2. The Stabilization and Physical Characteristics of Purified Bacterial 
Ribonucleoprotein Particles — Jac\ Wagman and Weston R. Trawic\ 11 

3. Stability of Ribonucleoprotein Particles of Escherichia coli — Ellis T. 
Bolton, Bill H. Hoyer, and Daniel B. Ritter 18 

4. Biochemical Characterization and Electron-Microscopic Appearance 

of Microsome Fractions — David Garfin^el 22 

5. The Configurational Properties of Ribonucleic Acid Isolated from 
Microsomal Particles of Calf Liver — Benjamin D. Hall and Paul Doty 27 

6. Microsomes and Ribonucleoprotein Particles [Invited paper] — George 

E. Palade . 36 

7. The Influence of Conditions of Culture on Certain Soluble Macro- 
molecular Components of Escherichia coli — S. Dagley and J. Sy\es . 62 

8. Physicochemical and Metabolic Studies on Rat Liver Nucleoprotein — 
Mary L. Peter mann, Mary G. Hamilton, M. Earl Balis, Kumud 
Samarth, and Pauline Pecora 70 

9. Ultracentrifugal Studies of Microsomes from Starving, Nonproliferat- 
ing, and Proliferating Yeast — James K. Ashi\awa 76 

10. Fractionation of Escherichia coli for Kinetic Studies — Richard B. 
Roberts, Roy J. Britten, and Ellis T. Bolton 84 

11. Microsomal Structure and Hemoglobin Synthesis in the Rabbit Retic- 

ulocyte — Howard M. Dintzis, Henry Borsoo\, and Jerome Vinograd 95 

12. Effects of /7-Fluorophenylalanine on the Growth and Physiology of 
Yeast — G. N. Cohen, H. O. Halvorson, and S. Spiegelman . . . 100 

13. Enzymatic and Nonenzymatic Synthesis in Adenyl Tryptophan — 
Martin Karasel(, Paul Casteljranco, P. R. Krishnaswamy, and Alton 
Meister 109 

14. Participation of Adenyl Amino Acids in Amino Acid Incorporation 
into Proteins — Paul Casteljranco, Alton Meister, and Kivie Moldave 115 




15. The Synthesis of Hydroxyproline within Osteoblasts [Abstract] — 
Sylvia Fitton Jackson 121 

16. Studies on Amino Acid Incorporation in Bacteria Using Ionizing 
Radiation — Ellis Kempner and Ernest Pollard 123 

17. The Effect of X Rays on the Incorporation of Phosphorus and Sulfur 
into Escherichia coli — Ernest Pollard and Jane Kennedy .... 136 

18. Statistical Relations in the Amino Acid Order of Escherichia coli 
Protein — Harold J. Morowitz 147 

19. The Formation of Protomorphs — Eran\ T. McClure and Richard B. 
Roberts 151 

20. Structure of Microsomal Nucleoprotein Particles from Pea Seedlings 
-Paul 0. P. Ts'o 156 


Isolation and Characterization 
of Bacterial Nucleoprotein Particles 


Department of Biochemistry, University of Wisconsin 

Fractionation of the particulate matter from broken cells has long excited 
the biochemist. Lilienfeld [1] prepared nuclear and cytoplasmic fractions and 
studied the properties of a deoxynucleoprotein (DNP). Huiskamp [2] noted 
the influence of buffer salts on isolated DNP. The possibility of differential 
extraction of subcellular structures was investigated by Bensley and Hoerr [3]. 
The technique of purifying subcellular components has advanced rapidly 
through the efforts of Claude [4], Hogeboom and Schneider [5], and 
Anderson [6]. 

The fractionation of subcellular components offers the possibility of integrat- 
ing the fields of intracellular anatomy, cellular physiology, and biochemistry. 
Siekevitz [7] working with mitochondria and Palade [8] and Zamecnik [9] 
working with the microsomal fraction have begun this integration by equating 
isolated fractions to structures observed in the electron microscope. Our studies 
with the ribonucleoprotein of Azotobacter vinelandii have clearly demon- 
strated that progress in this integration of fields demands a detailed under- 
standing of the properties and stability of the subcellular particles. Previous 
studies [10] on the protein synthesis in cell-free extracts of A. vinelandii must 
now be reinterpreted in the light of our current understanding of the stability 
of bacterial ribonucleoprotein. 

In 1954 Palade and Porter [8a] demonstrated endoplasmic reticulum in ani- 
mal cells. Hodge, Martin, and Morton [11] in 1957 demonstrated similar struc- 
tures in plant cells, and Sacks [12] has found related structures in yeasts, higher 
molds, and algae, leaving, at the present time, only the bacteria without clearly 
demonstrated endoplasmic reticulum. Pochon [13] found structures in A. vine- 
landii which by staining and observation in the light microscope were identi- 
fied as nuclei. We have observed granularity in regions of thin sections through 



A. vinelandii examined in the electron microscope. Several speakers at this 
conference have referred to similar granularity in sections of Escherichia coli as 
ribonucleoprotein particles, but no clear identification of these granules as such 
has yet been accomplished. Ribonucleoprotein particles, similar in size and 
chemical composition to those from animal cells, yeast, and fungi [14], can, 
however, be prepared from bacteria [15]. Several papers in this volume describe 
the ribonucleoprotein of E. coli, and this paper will treat the preparation and 
properties of ribonucleoprotein particles from A. vinelandii. 


One of the aims of this work is to prepare subcellular structures having a 
useful correspondence to structures that existed in the intact cell. We have no 
single criterion to indicate when such a preparation has been accomplished, 
but we use as supporting evidence reproducibility of the product when pre- 
pared by several varying methods and we also invoke all the information about 
the stability of isolated components. The stability observations are described 
in a later section of this paper. Three different methods of cell rupture have 
been found to permit isolation of indistinguishable particles, provided that the 
cultures used were harvested at a similar stage of growth. 

The first method was physical grinding with number 320 mesh Carborundum 
which had been washed with hydrochloric acid and rinsed with distilled water 
until neutral. The cells had been previously washed with distilled water. The 
packed cell paste was ground with 4 parts by weight Carborundum for approxi- 
mately 15 minutes or until moist. An additional 2 parts of Carborundum was 
added, and the cells were ground for approximately 5 minutes more. Visible 
microscopic examination of the mixture revealed that approximately 95 per 
cent of the cells were ruptured by the grinding method. The ground cells were 
diluted with 8 times the original cell volume of the following buffer: 1.6 X 10" 3 
M K2HPO4, 0.4 X 10" 3 M KH2PO4, and 5 X 10" 3 M MgS0 4 . This will be re- 
ferred to as the RNP buffer. The supernate from centrifuging this mixture at 
500g for 30 minutes is referred to as the crude extract. 

The second method of cell breakage employed cells grown in the presence 
of 2 M glycerol. The cells were collected by low-speed centrifugation, and the 
pellet was diluted into 8 volumes of the RNP buffer to rupture and produce 
the crude extract. 

In the third method of cell breakage, A. vinelandii protoplasts were ruptured 
by osmotic shock. Weibull [16] showed that Bacillus megatherium, when 
treated with lysozyme in sucrose solutions of high osmotic pressure, changed 
to a spherical form which is readily ruptured by lowering the osmotic pressure. 
We have avoided the use of sucrose in view of our findings on the instability 
of the isolated particles in dilute sucrose solutions. The A. vinelandii cells 
were washed in the RNP buffer, and suspended in 1.5 XlO" 3 M EDTA at a 
dilution such that the optical density at 660 m\\ was approximately 0.75. These 
solutions had been previously osmotically adjusted with glycerol or with Carbo- 


wax "4000" to maintain the protoplasts. The turbid solutions were brought to 
13 ug/ml in crystalline egg white lysozyme, and the turbidity was observed until 
its rapid decrease ceased. The protoplasts were then collected by low-speed 
centrifugation, washed once in osmotically adjusted RNP buffer, collected 
again, and ruptured by osmotic shock upon dilution with 10 times the packed 
cell volume of RNP buffer to yield a crude extract. During the development 
of the protoplast procedure, both the formation of the protoplasts and their 
osmotic rupture upon dilution were followed in the visible and phase contrast 


The crude extract derived from any of the above three procedures is centri- 
fuged at 4900g for 30 minutes. The pellet that accumulates consists of cell debris 






(4,900g x 30min) 


(I05,400g x 60min) 



(8,700 g x 15 min) 


Fig. 1. Flow sheet for differential centrifugation of ribonucleoprotein from a crude 
extract of A. vinelandii prepared by I grinding, II osmotic shock, or III protoplastic osmotic 


and, with method I, some Carborundum that was not removed at lower speeds. 
The supernatant liquid from this step is now centrifuged at 105,00% for 60 
minutes. The pellet so obtained is solubilized in RNP buffer for approximately 
12 hours, and then centrifuged at 8700g for 15 minutes. The precipitate is dis- 
carded, and the supernatant liquid is examined in the analytical ultracentrifuge 
to determine the number of sedimenting components, their relative amounts, 
and their sedimentation coefficients. When the 86 S component is desired, the 
105,000^- and the 8700^ cycle is repeated until over 90 per cent of the area in the 
schlieren pattern is under the appropriate peak. 


When a crude extract is processed to the stage labeled "preparation" in 
figure 1, and is examined in the analytical ultracentrifuge, it is found to sedi- 
ment as a single peak of sedimentation coefficient 86 S. If, however, the same 
crude extract is carried to the same stage employing a buffer in which the mag- 
nesium concentration has been reduced to 10~ 3 M, the ultracentrifuge pattern 
now shows five significant components. Comparison of the schlieren and ultra- 
violet absorption photographs in the ultracentrifuge suggests that all these com- 
ponents contain nucleoprotein. The sedimentation coefficients extrapolated to 
zero concentration and corrected to 20° C are 86, 77, 58, 39, and 10 S. The 86, 
58, and 39 S components are usually found in largest amount. Figures 2 and 3 

Fig. 2. An electron micrograph of the edge of a droplet of RNP particles sprayed onto 
a collodion membrane and shadowed with uranium. Magnified 34,000 times. Taken on a 
Siemens Elmiskop I by Professor Paul Kaesbcrg. 


Fig. 3. An electron micrograph of a central portion of a sprayed droplet showing RNP 
particles magnified 170,000 times. The smallest particles are about 200 A in diameter, the 
largest about 250 A. From the shapes of their shadows it is estimated that their thicknesses 
are about 75 per cent as great as their diameters. Taken on a Siemens Elmiskop I by 
Professor Paul Kaesberg. 

are electron micrographs of a purified preparation of the 86 S particles which 
was diluted lOOOfold with distilled water and then quickly sprayed on a col- 
lodion membrane and air-dried. It is not yet known how the short exposure to 
distilled water will affect particle structure. The electron micrographs taken 
under these conditions show that at least two size classes are present, both of 
which are roughly spherical. The 86, 58, and 39 S particles all show small de- 
pendence of sedimentation coefficient on concentration, which also suggests 
that the particles are not markedly asymmetric. 

The particles appear to contain ribonucleic acid and protein and to be free 
of lipid and deoxy nucleic acid. The nucleic acid component has been sepa- 
rated and purified by detergent treatment [17], phenol [18], chloroform [19], 
and glacial acetic acid [20] extraction. The protein component when sepa- 
rated from the nucleic acid has been found to be insoluble in aqueous solutions 
unless prepared through a 67 per cent glacial acetic acid procedure. The pro- 
tein shows an ultraviolet absorption typical of a protein rich in tyrosine. We 
have derived only one major protein from the particle at this point. The de- 
rived protein appears to have only one type of N-terminal amino acid, which 
we have very tentatively identified as glycine. The number of protein subunits 
per particle has not yet been quantitatively determined, but the experiments to 
date suggest a large number. 


Ribose nucleic acid prepared from the particle by chloroform extraction of 
the protein shows markedly the hyperchromic effect characteristic of poly- 
merized nucleic acids. Immediately upon addition of the alkali the optical 
density at 260 mu increases 15 per cent. After incubation for 16 hours at 30° C 
with 0.5 N NaOH and adjusting the pH to 7.5, the final hyperchromic effect 
is found to be 39.1 per cent. The nucleotides (table 1) arising upon alkaline 
hydrolysis of the ribose nucleic acid have been chromatographed on Dowex-1- 
formate ion-exchange columns developed with gradient elution. The unknown 
nucleotide shows chromatographic behavior similar to that of the new ribonu- 
cleotide reported by Cohn, but its acid, alkaline, and neutral ultraviolet ab- 
sorption spectra are not identical to those of the fifth nucleotide which we have 
isolated from yeast. 

The 86 S particle has been examined for its stability as a function of salt, 
chelating agents, enzymatic attack, pH, and sucrose concentration (fig. 4) . The 
results of these studies were fed back into improvements in the preparative pro- 
cedure and are of utmost importance to the interpretation of labeled amino 
acid incorporation studies in the particulate fractions of A. vinelandii. If the 
86 S particle is suspended in a pH 7.05 buffer of 2xl0" 3 M phosphate, 10~ 3 
M MgSO-i, with NaCl added to a total ionic strength of 0.03, or is dialyzed 
against 2 X 10" 3 M K 2 HPC>4:KH 2 P04 (4:1) buffer, it dissociates to yield 58 and 
39 S components. 

Our early studies showed that the 58 and 39 S components could be returned 
to 10^ 3 M Mg ++ -containing solutions without re-forming the 86 S particle which 
had previously been stable to that environment. Encouraged by our discussions 
with Dr. Paul Ts'o at this conference, we explored further and found that in 
5 X 10" 3 M Mg ++ the 58 and 39 S components recombined to form the 86 S par- 
ticle, and that once formed this particle again was stable in 10" 3 Mg ++ . In all 
these studies the buffer also contained 2 X 10" 3 M potassium phosphate buffer 
of pH 7.05. 1 We also confirm Ts'o's observation that the area of the 39 S peak 
is about one-half that of the 58 S peak, which suggests that one small 39 S 
and one larger 58 S particle combine to form the 86 S particle. Upon addition 
of 0.01 M, pH. 7, ethylenediaminetetraacetic acid, the particles further dissociated 
to ribonucleoprotein of sedimentation coefficient less than 5 S (not extrapolated 
to zero concentration). The particles are also rapidly degraded to small frag- 
ments by ribonuclease but are not attacked by deoxyribonuclease. They are 
precipitated by pH below 6.5 or above 7.5. 

Attempts to use sucrose for certain stages of the purification led to the ob- 
servation that, if sucrose was added to the RNP buffer, the particles aggregated 
and were readily removed by low-speed centrifugation. Sucrose concentrations 
from 3 to 30 per cent were all found to have this effect. This finding necessi- 

1 Note added in proof: We recently reported at the 1958 meeting of the Federation of 
American Societies for Experimental Biology that a buffer 5xl0~ 3 M in MgO and ad- 
justed to p¥L 7.05 with cacodylic acid gives improved yield and excellent stability of the 
80 S class of RNP from yeast, E. coli, and A. vinelandii. 


TABLE 1. Analysis of Nucleic Acid Hydrolyzed with 0.5 AT NaOH 

for 16 Hours at 37° C 

Hydrolyzate was chromatographed on a Dowex-1-formate column 
with gradient elution. 





mole % 

mole % 
















Not reported 

* Whole cell data from 

Lombard and 

Chargaff [1 


tates re-evaluation of some studies [10] that have been carried out on A. vine- 
landii and raises the important question whether this phenomenon can occur 
in ribonucleoprotein from other sources. 

The ribonucleoprotein has been assayed for a large number of enzymatic 
activities. It appears free of nucleotide phosphatase activity, glucose-1-phos- 
phatase, oxidative phosphorylation enzymes, and electron-transport enzymes. 
A feeble glucose-6-phosphatase activity of 10 mM phosphate released per minute 



t — J 









Fig. 4. The preparation and stability of A. vinelandii nucleoprotein. Each analytical 
ultracentrifuge pattern is labeled with the stage of preparation (top pair) or the treatment 
to which a pure 86 S product was subjected. The salt and EDTA treatment are described 
in the text. The RNAse action occurred in less than 5 minutes at 1 part of enzyme per 
1000 parts of particle whereas the DNAse at the same concentration produced no change 
in 45 minutes. 


per gram of particle was detected, but it is at most a few per cent of that found 
in equal weights of liver microsome fractions. 


The study of A. vinelandii ribonucleoprotein has been more successful in 
posing interesting questions than in providing answers to previous questions. 
Can the sharp requirements of the nucleoprotein for divalent cations and for 
certain pW ranges be exploited to give information on the mode of combina- 
tion of nucleic acid and protein or combination between nucleoprotein subunits ? 
Does the marked difference in ribonuclease sensitivity of the plant ribonucleo- 
protein viruses and the bacterial ribonucleoprotein imply a different orientation 
or localization of nucleic acid and protein? The plant viruses now appear to 
have a protein coating with nucleic acid (or nucleoprotein) in an inner layer 
concentric with the protein coat and thus protected from ribonuclease attack. 
How different must the structure be to permit the rapid attack observed? 

Will the small subunits derivable by salt and EDTA treatment also yield in- 
formation on the mode of action or size of the functional nucleic acid? 

The striking effect of divalent cations on the physical state of nucleoprotein 
is now becoming recognized as a phenomenon common to many systems. 
Huiskamp [2] in 1901 noted that thymus nucleoprotein was precipitated by 
0.01 M calcium, barium, and magnesium salts and dissolved in excesses (0.1 M) 
of these same salts. He equated changes in physical properties of nucleoprotein 
solutions during dialysis to losses of divalent cations. He noted that heavy- 
metal divalent cations formed nucleoprotein precipitates that were difficult to 
dissolve. Korkes [22] and co-workers used a similar observation on manganese 
RNP to remove RNP from bacterial extracts. Carter and Hall [23] working 
in the laboratories of J. W. Williams noted that thymus nucleoprotein in 
sodium chloride solutions was a rodlike molecule but in calcium chloride solu- 
tions it became compact and showed no dependence of sedimentation rate on 
concentration. Wiberg and Neuman [24] have studied the binding of mag- 
nesium and calcium by RNA and DNA and find a region of concentration 
through which the number of equivalents bound changes rapidly. This con- 
centration range is the same as that which we find critical for nucleoprotein 
structural changes. 

The studies reported here, added to the work of Mazia [25] on the role of 
polyvalent cations in deoxynucleoprotein and nuclear structure and of Chao 
and Schachman [14a] in ribonucleoprotein stability, and to the many excellent 
contributions presented at the second annual Biophysics Conference, will help 
establish the basic rules for fractionation of subcellular particles in a reproducible 

The dependence of nucleoprotein structure upon divalent cation concentra- 
tion is striking enough for these ions to become of interest in consideration of 
the variables which dictate when a nucleic acid will be in the double helix, 


when it will split, when it is "soluble RNA," and when not. Re-evaluation of 
Brachet's [26] findings on the relation of ribonucleoprotein to growth phase 
of yeast will be warranted in the light of current concepts of the importance 
of buffer media. 

Whereas the divalent cations of the buffer medium play an important role 
in nucleoprotein structure, there appear to be mineral elements which are in- 
fluential in nucleic acid function and are not in free equilibrium with the 
buffer. Zittle [27] and Jungner [28] demonstrated that carefully isolated yeast 
RNA contained characteristic amounts of metal ions. Kihlman [29] and 
Mazia [25] have related structural integrity of chromosomes to metal content. 
Loring and Cooper [30] find that certain cations are important for nucleopro- 
tein stability of tobacco mosaic virus and hence for infectivity. Racker and 
Krimsky [31] have evidence that metal ions are involved in an animal virus. 
These are but a few of the many nucleoprotein-metal systems cited in the 

Thus it appears that for viruses, as well as for subcellular particles, elucida- 
tion of the structural and functional role of cations will be a fertile and chal- 
lenging frontier for those with pioneering instincts. 


Ribonucleoprotein particles of sedimentation coefficient S 2 o o = S6 have been 
isolated from A. vinelandii. The particles are free of lipid and DNA. They 
are stable at neutral pH, in low-ionic-strength solution when divalent cations 
are present; they are unstable in sucrose, in concentrated salts, and in the 
presence of ribonuclease. Nucleic acid derived from the particles contains an 
unidentified fifth base. The 86 S unit reversibly dissociates to particles of 58 
and 39 S when the Mg ++ concentration is lowered. 


We are pleased to acknowledge the generous aid of our colleagues. Profes- 
sor Paul Kaesberg conducted the electron-microscope studies, Miss Vatsala 
Thakur provided the N-terminal amino acid analyses, and Miss Fay Hoh col- 
laborated in studies on the protein derived from the particles. Financial support 
from the National Institutes of Health and Wisconsin Alumni Research Foun- 
dation is gratefully acknowledged. 


1. L. Lilienfeld, Z. physiol. Chem., 18, Ab. A. Claude, Advances in Protein 
473 (1894). Chem., 5, 423 (1949). 

2. W. Huiskamp, Z. physiol. Chem., 32, 5. W. C. Schneider and G. H. Hoge- 
145 (1901). boom, /. Biol. Chem. 183, 123 (1950). 

3. R. R. Bensley and N. L. Hoerr, Anat. 6. N. G. Anderson, Science, 121, 775 
Record, 60, 251 (1934). (1955). 

4a. A. Claude, Harvey Lectures, 43, 121 7. P. Siekevitz and M. L. Watson, /. Bio- 

(1948). phys. Biochem. CytoL, 2, no. 6, 653 (1956). 



8a. G. E. Palade and K. R. Porter, /. Ex- 
ptl Med., 100, 641 (1954). 

8b. G. E. Palade, /. Biophys. Biochem. 
Cytol., 2, 547 (1954). 

8c. G. E. Palade and K. L. Porter, /. Bio- 
phys. Biochem. Cytol, 3, 269 (1957). 

8^. G. E. Palade, /. Biophys. Biochem. 
Cytol, 2, no. 4, Suppl., 85 (1956). 

8<?. G. E. Palade and P. Siekevitz, /. Bio- 
phys. Biochem. Cytol, 2, 171 (1956). 

9. P. C. Zamecnik, Set. American, 5, 
118 (1958). 

10. D. P. Burma and R. H. Burris, 
/. Biol. Chem., 225, 287 (1957). 

11. A. J. Hodge, E. M. Martin, and 
R. K. Morton, /. Biophys. Biochem. Cytol, 
3, no. 1, 61 (1957). 

12. Irving B. Sacks, personal communi- 

13. J. Pochon, Y. T. Tchan, and T. L. 
Wang, Ann. inst. Pasteur, 74, 182 (1948). 

\4a. Fu-Chuan Chao and H. K. Schach- 
man, Arch. Biochem. Biophys., 61, 220 

\4b. Fu-Chuan Chao, Arch. Biochem. 
Biophys., 70, 426 (1957). 

15. H. K. Schachman, A. B. Pardee, and 
R. Y. Stanier, Arch. Biochem. Biophys., 38, 
245 (1952). 

16. C. Weibull, /. Bacteriol, 66, 688 

17. E. R. M. Kay and A. L. Dounce, 
/. Am. Chem. Soc, 75, 4041 (1953). 

18. G. Schramm, A Symposium on the 
Chemical Basis of Heredity (W. D. Mc- 
Elroy and Bentley Glass, eds.), p. 513, 

19. M. G. Sevag, D. B. Lackman, and 
J. Smolena, /. Biol. Chem., 124, 425 

20. H. Fraenkel-Conrat, Virology, 4, 1-4 

21. A. Lombard and E. Chargaff, Bio- 
chim. ct Biophys. Acta, 20, 285 (1956). 

22. S. Korkes, A. del Campino, I. C. 
Gunsales, and S. Ochoa, /. Biol. Chem., 
193, 721 (1951). 

23. R. O. Carter and }. L. Hall, Nature, 
144, 329 (1939). 

24. J. S. Wiberg and W. F. Neuman, 
Arch. Biochem. Biophys., 72, 66 (1957). 

25. D. Mazia, Proc. Natl. Acad. Sci. 
U. S., 40, 521 (1954). 

26. J. Brachet and R. Jener, Biercs et 
boissons,3, 422 (1942). 

27. C. A. Zittle, /. Biol Chem., 163, 111 

28. G. Jungner, Science, 113, 378 (1951). 

29. B. A. Kihlman, /. Biophys. Biochem. 
Cytol, 3, 363, 381 (1957). 

30. H. S. Loring and W. D. Cooper, 
/. Biol Chem., 211, 505 (1956). 

31. E. Racker and I. Krimsky, /. Ex- 
ptl Med., 85, 715 (1945). 


The Stabilization and Physical Characteristics of 
Purified Bacterial Ribonucleoprotein Particles 


U. S. Army Chemical Corps, Fort Detric\, Frederick., Maryland 

Principally through electron-microscope studies on thin sections of various 
species [1] it has been possible to show that a major part of the bacterial cyto- 
plasm consists of widely dispersed granules 100 to 200 A in diameter. In ultra- 
centrifugal analysis of aqueous extracts from disrupted bacteria, Schachman, 
Pardee, and Stanier [2] found one of the major components, with j 2 o,w = 40 S, 
to consist of roughly spherical particles, about 150 A in diameter, which con- 
tain much of the cytoplasmic ribonucleic acid (RNA). A variety of enzymatic 
functions have since been attributed to these particles [3], including the systems 
for electron transport, for oxidative phosphorylation, and for some of the re- 
actions of the tricarboxylic acid cycle. 

The present paper reports attempts to isolate the 40 S component, by suc- 
cessive differential centrifugation of extracts from Escherichia coli, and physi- 
cal measurements obtained with purified material. Particular attention is drawn 
to the observation that the stability of these particles is dependent upon a 
dialyzable substance in cell extracts which apparently inhibits decomposition. 


The extracts used were prepared from E. coli (ATCC 4157) grown 24 hours 
on nutrient agar at 37° C. After washing by alternate centrifugation and re- 
suspension, cells were disrupted by shaking with glass beads in a Mickle dis- 
integrator. The procedure consisted of shaking 6-ml volumes of suspensions 
(about 7xl0 10 cells/ml) with 5 g of beads (type 114 Minnesota Mining and 
Manufacturing Company) for 5-minute periods at 1700 cycles per minute. The 
extracts were then cleared of unbroken cells and debris by low-speed centrifu- 




gation. Potassium phosphate buffer, pH 7.0 and ionic strength 0.1, was used in 
the preparation of extracts. 

Preparative and analytical sedimentation were carried out in Spinco ultra- 
centrifuges Models L and E, respectively. A diffusion constant measurement 
was made by free diffusion in a Claesson cell [4]. Partial specific volume was 
determined by density measurements in a Lipkin pycnometer [5]. 


Decomposition Inhibitor in E. coli Extracts. It has been shown that the ex- 
tractive procedure described here yields solutions that are highly reproducible 
as determined by sedimentation and electrophoretic behavior [6]. The sedi- 
mentation diagram (see fig. 1) corresponds closely to those obtained previ- 
ously [2] by other methods of cell disruption. The rapidly sedimenting 40 S 
component is clearly resolvable, and it appears that a considerably purified prep- 
aration should be obtainable simply by successive differential centrifugation. 
In early fractionation attempts, pellets from solutions subjected to centrifugal 
fields about 100,000^ for 90 minutes (in the no. 30 rotor of the Model L Spinco) 
were found to be only partly resoluble, in agreement with the finding of 
Schachman et al. [2]. Moreover, the soluble material, as shown in figure 1, 
contained an unexpectedly large amount of more slowly sedimenting material 
along with a disappointingly low quantity of 40 S component. 

In a subsequent study to determine the effect of dialysis on the nature of 
E. coli extracts, an observation was made that proved to be an important step 
in this purification problem. As shown in figure 2, the 40 S component in 
dialyzed extracts was greatly reduced in concentration with a simultaneous in- 
crease in the quantity of more slowly sedimenting material. In additional ex- 
periments, however, the effect was found to be diminished as the dialysate- 
extract volume ratio was decreased. This suggested the presence, in those 
extracts, of a dialyzable material that functions as a stabilizer of the 40 S com- 
ponent. A substance with an analogous property has been reported by Peter- 
mann and Hamilton [7] in studies with rat liver homogenates. 


Pellet fraction 

Fig. 1. Sedimentation diagrams of an E. 
coli extract and the pellet fraction derived 
by a two-step centrifugation at 100,000^ for 
90 minutes. Patterns were recorded 7 min- 
utes after attainment of full field strength, 
about 250,00%. 


Extract dialyzed 
against buffer 

Fig. 2. The effect of dialysis on the sedi- 
mentation behavior of an extract from E. 
coli. Recording of diagrams took place 8 
minutes after full field strength was reached, 


The stabilizing effect of the dialyzable material, which is tentatively desig- 
nated "decomposition inhibitor" (DI), was further demonstrated as follows. 
After overnight dialysis of 1 volume of extract solution against 4 volumes of 
water, the dialysate was lyophilized and reconstituted to four-fifths of the vol- 
ume of extract used. This procedure yields a solution, which we shall refer 
to as buffer-DI, whose concentration of dialyzable material is roughly equal to 
that of the original extract. The stabilizing effect of the DI was then observed 
by comparing the sedimentation behavior of fresh extract solutions dialyzed 
against buffer and buffer-DI, respectively. As is shown in figure 3, the DI 
effectively prevented the disappearance of 40 S component. 

The nature of the stabilizing substance is not clear. An analysis of the di- 
alyzable material indicates the presence of peptides and nucleotides as well as 
trace amounts of magnesium, iron, and other metals usually found in bacterial 
extracts. Metal ions appear to be ruled out as the active substance, for the sta- 
bilizing effect is lost by heating the buffer-DI for 5 minutes at about 90° C. 

Purification of 40 S Component. It now appears that the early difficulties in 
fractionating the 40 S component were due to a procedure which separates that 
component from the stabilizer. By slightly modifying the original fractionation 
scheme, the 40 S component was prepared in a relatively high state of purity 
and stability. The pellets, after each of two successive centrifugation steps, were 
redissolved in previously prepared buffer-DI. Figure 4 shows a comparison of 
preparations, from a single batch of cells, obtained by this and the previous 
methods. The sedimentation patterns demonstrate the degree of purity with 
which it is possible to obtain the 40 S component, and illustrate the activity of 
the DI. 

Physicochemical Properties. An analysis of the fractionated material indicated 




Extract dialyzed against: 
Buffer Buffer-DI 

Fig. 3. Sedimentation diagrams illustrating the protective effect of the dialyzable stabi- 
lizer on the 40 S component in E. coli extracts. 



Crude extract 

Centrifugal! y fractionated RNP (40 S) component 

Without buffer-DI 

Using buffer-DI 
observed initially 

Using buffer-DI 
observed 4 days later 

Fig. 4. Sedimentation diagrams illustrating the influence of the stabilizing medium 
(buffer-DI) on the nature of the RNP fraction obtainable from E. coli extracts by differen- 
tial centrifugation. Diagrams were recorded about 5 minutes after full field strength was 
reached, 250,00%. 

that it consisted entirely of protein and ribose nucleic acid in a proportion about 
3 to 1 by weight. Since the material appears to be essentially monodisperse, it 
is tentatively concluded that the 40 S particles contain protein and nucleic acid 
in combination as a ribonucleoprotein (RNP) . 

Sedimentation, diffusion, and partial specific volume measurements were 
carried out with the RNP fraction, the more detailed data being obtained with 
water as the solvent. The results are summarized in table 1. Although, as is 
seen in figure 4, very small amounts of more slowly sedimenting material were 
present in these preparations, they are believed not to cause serious errors in 
the use of these data to determine particle characteristics. 

TABLE 1. Sedimentation and Diffusion Constants of Centrifugally Fractionated 
Ribonucleoprotein from Extracts of E. coli 

Prepa- tration 

ration Solvent mg/ml 

a 0.1 ionic strength 6.3 
phosphate buf- 
fer, pH 7.0 
b Water 8.0 




•*20, W 




46.9 * 

10 7 D 2(V 

cm- sec 



V 20 






* Value obtained by extrapolating \/s. 


The very slight concentration dependence of S20, w (with a value, at infinite 
dilution, of 46.9 S) is in agreement with the finding by Schachman et al. [2] that 
these particles appear to be roughly spherical in the electron microscope. On 
the assumption of a negligible concentration dependence for diffusion, the 
molecular weight calculated from the measured data is 1,240,000, and the fric- 
tional ratio, f/fo, is 1.16, corresponding to spheres, 137 A in diameter, hydrated 
to the extent of 0.37 ml/g of RNP. 

An electron micrograph, made using a water solution of RNP, is shown in 
figure 5. The particles appear to be uniformly spherical with a mean diameter 
about 152 A. This value is in good agreement with the sedimentation-diffusion 
figure, inasmuch as the probable errors with the electron microscope tend to 
yield too high a figure. 


It has been reported (e.g., by Chao and Schachman [8] and by Bolton, Hoyer, 
and Ritter, paper 3 of this volume) that more rapidly sedimenting components 
(60 S and 80 S) appear in extracts from microbial cells when magnesium ions 
are present in sufficient concentration in the extract media. This addition to 
the ionic environment is not required for the preservation of the 47 S particles 
dealt with in the present report. In the intact cell, the synthesis and dissocia- 
tion of RNP particles are probably controlled by several factors which may 
include magnesium and other ions. The action of ribonuclease, for example, 
is inhibited by a large number of substances [9] including bivalent cations and 
mononucleotides, all of which are apparently present in the dialyzable fraction 
of E. coli extracts. 

Although the RNP component has been identified with a variety of enzymatic 
functions, it is not clear whether these are intrinsic or merely adsorbed. Elec- 
trophoretic data [6] show that this material, despite its apparent monodispersity 
in the ultracentrifuge and electron microscope, is heterogeneous, consisting of 
two or more components with a wide range of anodic mobilities at pW 7.0. 


We wish to acknowledge the assistance of Mrs. F. Elizabeth White in the 
growth and harvesting of the E. coli cultures and of Mr. Robert L. Sine in the 
preparation of the electron micrograph. 


A dialyzable substance in extracts from E. coli has been found to inhibit 
the decomposition of the RNP particles (previously referred to as the 40 S 
component). This observation has led to the preparation of the RNP fraction 
in a relatively high state of purity and stability. Physical measurements indi- 
cate that the RNP component consists of uniformly spherical particles with 
a molecular weight 1,240,000 and a diameter of 137 A. 



Fig. 5. Electron micrograph taken with a specimen of purified RNP. Shadow angle 
3:1, RCA electron microscope. Polystyrene latex particles, 0.188-micron diameter, were 
used as internal standard. 




1. J. R. G. Bradfield, "Organization of 
bacterial cytoplasm," from Bacterial Anat- 
omy, Cambridge University Press, 1956. 

2. H. K. Schachman, A. B. Pardee, and 
R. Y. Stanier, Arch. Biochem. Biophys., 38, 
245-260 (1952). 

3. M. Alexander, Bacteriol. Revs., 20, 
67-93 (1956). 

4. S. Claesson, Nature, 158, 834 (1946). 

5. Ace Glass Company, Vineland, New 

6. J. Wagman, E. Pollack, and E. J. 
Weneck, Arch. Biochem. Biophys., 73, 161- 
170 (1958). 

7. M. L. Petermann and M. G. Hamil- 
ton, /. Biophys. Biochem. CytoL, 1, 469- 
472 (1955). 

8. F.-C. Chao and H. K. Schachman, 
Arch. Biochem. Biophys., 61, 220 (1956). 

9. M. R. McDonald, in Methods in En- 
zymology (Colowick and Kaplan, eds.), 
vol. 2, pp. 433-434, Academic Press, New 
York, 1955. 


Stability of Ribonucleoprotein Particles 

of Escherichia coli 


Department of Terrestrial Magnetism- 
Carnegie Institution of Washington 


Rocky Mountain Laboratory 

National Institute of Allergy and Infectious Diseases 

U. S. Public Health Service, Hamilton, Montana 

Investigations concerned with the structure and function of ribonucleopro- 
teins of microorganisms require particle preparations that are representative, 
reproducible, and stable. This report presents some results of exploratory 
studies in which the analytical ultracentrifuge was used to assess the influence 
of various suspending media on the ribonucleoproteins of "Escherichia coli. 


E. coli, B (ATCC 11303) harvested during the exponential phase of growth 
in a glucose-salts culture medium * was used for all studies. The bacteria were 
washed and resuspended (25 mg dry weight of bacteria per milliliter) in ap- 
propriate buffer solutions and disrupted by means of a modified French pres- 

1 The composition of C medium and other culture conditions may be found in 
Roberts et al., Studies of Biosynthesis in Escherichia coli, Carnegie Inst. Wash. Publ. 607, 
Washington, D. C, 1955. 




sure cell 2 operated at approximately 10,000 psi. Break- 
age of the bacteria by this means is essentially com- 
plete. The resulting bacterial juices were examined in 
the analytical ultracentrifuge (Spinco, Model E) as 
soon as practicable (about 30 minutes after rupture), 
or after various periods of storage at 4° C. The 
centrifuge was routinely brought up to speed in 6 to 
7 minutes and held at about 60,000 rpm for the dura- 
tion of the run. 


Figures 1 to 4 are illustrative sedimentation dia- 
grams showing that the pattern of rapidly sediment- 
ing components varies in accord with the kind of 
suspending medium used. Figure 1 compares the 
sedimentation behavior of the components in extracts 
prepared from bacteria broken in 0.01 M Tris-0.004 M 
succinic acid-0.005 M magnesium acetate buffer (pH. 
7.6, "TSM"), in 0.01 M Tris-0.004 M succinic acid 
("TS"), or in TSM + 0.07 M phosphate (pH 7.6). 
Values spotted along the abscissa are approximate ap- 
parent sedimentation coefficients. It is evident from 
this comparison that more, and larger, components 
are observed when magnesium has been included in 
the buffer and also that phosphate abolishes the more 
rapidly sedimenting materials. Whether the effect of 
phosphate is specific or whether the result is due to an 
increased ionic strength of the medium is not known. 
The sharp spike characteristic of highly polymerized 
deoxynucleic acid (DNA) 3 is missing from these dia- 
grams, although it is readily observed in juices pre- 
pared by breaking E. coli as a result of lysozyme treat- 
ment and osmotic shock. In spite of this finding, 
three-quarters of the ultraviolet-absorbing substance 
and one-seventh of the protein of E. coli disrupted in 
the TSM medium may be sedimented in the prepara- 
tive rotor (100,000g, 90 minutes). 

Figure 2 shows that juices prepared by pressure 
cell disruption maintain constant sedimentation dia- 
grams for at least 20 hours. If, however, sodium 
ethylenediaminetetraacetate (EDTA, 0.1 M, pH. 7.6) is 
extracts, all components greater than about 20 S disappea 

TSM + 
07M PO, 


8 mm 60,000 RPM 

Fig. 1. Sedimentation 
diagrams of E. coli dis- 
rupted in various buffer 
solutions. The concentra- 
tion of the bacterial juices 
differed among the runs. 

added to the bacterial 
r. This occurs whether 

2 C. S. French and H. W. Milner, Methods in Enzymology I, Academic Press, p. 65. A 
similar device is marketed by the American Instrument Company, Silver Spring, Maryland. 

3 See, for example, the sedimentation diagrams reported by W. Gillchriest and R. Bock, 
S. Dagley and J. Sykes, and J. Wagman reported in the present volume. 




20 hours 

12 min 60,000 RPM 

Fig. 2. Influence of storage at 4° C. The suspending buffers (TSM, left; TS, right) 
also contained 0.25 M sucrose, although subsequent runs have shown that sucrose has no 
effect on the pattern of components. 

or not magnesium is included in the original suspending medium, as figure 3 
demonstrates. In addition, a markedly decreased (<10 per cent) ultraviolet 
absorption occurs in the 100,000^-1 hour pellets (preparative rotor) when 
EDTA has been added to the bacterial juices. Figure 4 shows that DNAase 
(2 ug/ml) has little, if any, effect upon the number and size of the rapidly 
sedimenting materials, whereas RNAase (approximately 10 ug/ml) removes 
these components. 


Pressure cell disruption of E. coli at pH 7.6 in magnesium-containing solu- 
tions of low ionic strength (e.g., 0.01 M Tris-succinate) releases high-molecu- 
lar-weight components which range from 20 to 80 S. These components "fall 
apart," i.e., become elements having sedimentation coefficients less than about 
20 S, when a chelater, EDTA, or the enzyme ribonuclease is allowed to act 
upon them. DNAase, sucrose, or cysteine exerts no apparent effect on either 
the number of components or their relative quantities. Nearly all (>80 per 
cent) of the ribonucleic acid and about one-seventh of the protein of E. coli 
can be sedimented in a preparative rotor under optimum conditions (TSM, 
100,000g, 90 minutes). No RNA and only a trivial amount of protein can be 
sedimented after EDTA or ribonuclease treatment. Hence, it may be con- 






6 mm 60,000 RPM 

Fig. 3. Effect of EDTA on the sedi- 
mentation diagrams of E. coli juice. The 
two lower diagrams are from preparations 
containing one-half as much material as 
those for the upper pattern. 

6 min 60,000 RPM 

Fig. 4. Effect of nucleases on sedimenta- 
tion diagrams. The lower pattern is from a, 
preparation one-half as concentrated as that 
of the upper diagram. 

eluded that the bulk of the high-molecular-weight components of E. coli is 
composed of ribonucleoproteins held together in a fashion in which divalent 
cation (s) (probably Mg ++ ) and the integrity of ribonucleic acid play important 
roles. Thus, in certain physical and chemical attributes the "ribosomes" (ribo- 
nucleoprotein particles) of E. coli resemble constitutive elements of the cyto- 
plasm of other bacteria, and also of yeast, plants, and mammals. 


Biochemical Characterization and 

Electron-Microscopic Appearance of 

Microsome Fractions 


Eldridge Reeves Johnson Foundation for Medical Physics 
University of Pennsylvania 

The electron-microscope studies on microsomes by Palade and Siekevitz [1] 
have resulted in the definition of three kinds of microsomes: granules of 150 A 
diameter; smooth-surfaced vesicles; and rough-surfaced vesicles which differ in 
appearance from the smooth-surfaced ones primarily by having the granules 
attached to them. Biochemical studies of microsomes have resulted in the iso- 
lation of two varieties of microsomes — the 150 A granules just mentioned, 
which are the principal subject of interest in this symposium, and which are 
rich in RNA but poor in lipid and cytochrome b 5 , and a particle isolated by 
Perm and Mackler [2] which is rich in cytochrome b 5 and lipid and poor in 
RNA. It will be shown here that there are at least three biochemically dis- 
tinct varieties of microsomes, correlated with those observed in the electron 

It is possible to fractionate mammalian liver microsomes (the work here 
described is with pig liver) so as to obtain, in addition to the microsomes as 
they are usually prepared, a small light fraction of microsomes which is usually 
found to contain about twice as much cytochrome b 5 and less than half as 
many ribonucleoprotein granules per unit biuret protein as the bulk of the 
microsomes, hereafter referred to as the bul\ fraction. This fractionation may 
be made, for instance, by centrifuging a concentrated (1 part liver to 2 parts 
0.25 M sucrose) homogenate, after the mitochondria have been removed, at 

1 Public Health Service Research Fellow of the National Cancer Institute, 1955-1957. 



70,000^ for 25 minutes. In addition to the bulk microsomal pellet, the light 
fraction is obtained as a suspension at the bottom of the centrifuge tube. In 
order to permit spectroscopic study, both fractions are washed with Ringer's 
solution, which removes the hemoglobin. It also washes out about 80 per cent 
of the RNA (although the ribonucleoprotein granules do not disintegrate and 
may still be isolated [3]), so that there are no accurate values for RNA con- 
centration. It is found, however, that the bulk fraction contains appreciable non- 
cytochrome heme, somewhat more than the cytochrome heme, whereas the light 
fraction contains very little. A detailed description will be published elsewhere. 

We have seen that the microsomes that centrifuge down last, the light frac- 
tion, are rich in cytochrome b 5 . Palade and Siekevitz [1] found that they were 
likely to be smooth-surfaced. This finding suggests a correlation between 
smooth-surfaced appearance and the presence of cytochrome b 5 . So does the 
fact that studies [4] in various tissues indicate that much of the cytochrome 
is present only where the electron microscope shows smooth-surfaced endo- 
plasmic reticulum. 

Drs. Ian R. Gibbons and T. F. Anderson kindly took photographs of one 
of these microsomal preparations with the electron microscope. Figure 1 shows 
views of the light and bulk fractions. It is seen that there is much more ma- 
terial of smooth-surfaced origin (free of the ribonucleoprotein granules) in the 
light fraction than in the bulk fraction. Counts of the numbers of smooth- and 
rough-surfaced microsomes, using unfixed preparations to avoid any enrich- 
ment of one microsomal type in the process of fixation, indicated that the per- 
centage of smooth-surfaced microsomes was proportional to the cytochrome con- 
centration (a light fraction which was twice as rich in cytochrome as the bulk 
fraction contained twice as many smooth-surfaced microsomes). Apparently 
cytochrome b 5 is localized in the smooth-surfaced microsomes and the rough- 
surfaced ones contain little of it. This observation is also in agreement with 
the fact that predominantly rough-surfaced microsomes can be prepared (from 
pancreas, for instance [4, 5]), and they contain little or no cytochrome. It 
would be desirable to confirm this by preparing pure smooth- or rough-surfaced 
microsomes and finding their cytochrome content (since the preparative method 
used enriches the smooth-surfaced microsomes only twofold, the resulting con- 
clusions regarding their properties should not be considered final). 

We are not limited to the electron microscope, but can also study the micro- 
somes by biochemical methods. The procedure used here is digestion with 
pancreatin, a mixture of digestive enzymes from the pancreas. After digestion, 
which is not complete, what is left of the microsomes is centrifuged down and 

Initially, two suspensions of light and bulk microsomes have equal con- 
centrations of cytochrome. Assuming that the smooth-surfaced microsomes 
contain nearly all the cytochrome, then their concentration is equal in the two 
suspensions. The protein concentrations are shown in table 1. The digestion 

Fig. 1. Electron micrographs of light (above) and bulk (below) fraction microsomes, 
fixed with osmium tetroxide and air-dried from distilled water. Magnification 20,000 X . 
The rough-surfaced microsomes may be identified by the little particles attached to them. 





Sedimentable Protein, mg/ml 

Before digestion 
After digestion 

Light Fraction 



Bulk Fraction 


for 10 days to 2 weeks is sufficient to solubilize nearly all the cytochrome; 
this is in fact the method of preparing cytochrome b 5 [6, 7]. The nucleopro- 
tein granules have practically disappeared in the process. Examination of the 
sedimentable protein content of these two suspensions shows that much more 
protein is left in the bulk fraction. Nearly all the nucleoprotein has been re- 
moved, and since the cytochrome contents were equal there should be the same 
amount of residue from the smooth-surfaced microsomes in both fractions. 
Apparently there is something present in the bulk fraction (of which there is 
much less in the light fraction) to account for the difference in protein con- 
tent. Since the other two forms were eliminated, this must be the rough-sur- 
faced microsomes. That there is a qualitative difference in the residues can be 
seen by looking at the centrifuged pellets. Both contain a transparent amber 
layer, but this is all that is left of the light microsomes, whereas the bulk-frac- 
tion pellet has below this transparent layer an opaque tan one, showing some 
signs of further layering. This bulk-fraction pellet still contains the noncyto- 
chrome heme originally present. The rough-surfaced microsomes, which have 
not previously been characterized biochemically, are therefore found to contain 
a tan pigment and the noncytochrome heme, and to be distinct from the smooth- 
surfaced ones. This finding is summarized in table 2. It should be kept in mind 
that the expression "rough-surfaced microsomes" means the vesicles themselves 



in Electron 


This Pai 


-tide Is 



Rich in 

Poor in 



Hamilton [3] 


Cytochrome b 5 , 



Penn and 
Mackler [2] 
(probably a 

Cytochrome b 5 , 



Empty circle 
or ellipse 
with small 
filled-in cir- 
cles attached 
to outside 

tan pigment 



without the attached granules. The fact that three varieties of microsomes have 
been defined is not intended to imply that any of these varieties of microsomes 
is itself homogeneous. 


1. G. E. Palade and P. Siekevitz, /. Bio- 
phys. Biochem. CytoL, 2, 171 (1956). 

2. N. Penn and B. Mackler, Federation 
Proc., 16, 232 (1957). 

3. M. L. Petermann and M. G. Hamil- 
ton, /. Biol. Chem., 224, 725 (1957). 

4. D. Garfinkel, unpublished experi- 

5. G. E. Palade and P. Siekevitz, /. Bio- 
p/iys. Biochem. CytoL, 2, 671 (1956). 

6. D. Garfinkel, Arch. Biochem. Bio- 
phys., 70, 111 (1957). 

7. P. Strittmatter and S. F. Velick, 
/. Biol. Chem., 221, 253 (1956). 


The Configurational Properties of 

Ribonucleic Acid Isolated from 
Microsomal Particles of Calf Liver 


Department of Chemistry, Harvard University 

Although ribonucleic acids (RNA) from many sources have been examined 
by physical methods within the last few years no clear and consistent picture 
of the configurational properties of RNA has materialized. Most studies of 
RNA have been complicated by spontaneous changes of molecular weight, ag- 
gregation under some conditions and degradation under others. In the work 
reported here we have avoided these complications by finding experimental con- 
ditions under which the RNA is stable and have then proceeded to establish 
its configurational properties in solution by means of several different physical 

The choice of microsomal particles from liver as our source of RNA was 
prompted by the particular importance that these ribonucleoprotein particles 
have assumed by virtue of their participation in protein synthesis [1, 2] and the 
fact that they can be isolated in pure form [3] before the preparation of the 
RNA itself. 

It is important to emphasize at the outset that our major emphasis in the 
work reported here has been on the configurational properties of stable RNA 
isolated from these particles. We defer until a later time a report on the molecu- 
lar weight and configuration of RNA within the microsomal particles and the 
relation of the work presented here to these properties. 


Preparation of Microsomal Particles. The procedure summarized below 
evolved from those used by Zamecnik et al. [1] and by Petermann and Hamil- 
ton [4] for the isolation of similar particles from rat liver. Calf liver was 




quickly frozen in Dry Ice within 3 minutes of slaughter, and thawed immedi- 
ately before proceeding with the preparation. The liver was thoroughly chopped 
while thawing and the cell walls were broken by blending in an Osterizer 
Blendor intermittently for 3 to 4 minutes at about half speed. The suspending 
medium for this operation was ice-cold 0.25 M sucrose (2 cc/g liver). 

The resulting suspension was centrifuged twice for 30 minutes at 1500^ in 
the cold. The supernatant solution, containing the microsomes, glycogen, and 
soluble liver proteins, was removed by pipet after each centrifugation. The 
microsomal particles were then separated from the lipoprotein portion of the 
microsomes by emulsifying the microsomes with sodium deoxycholate [1]. A 
pellet of microsomal particles can then be obtained by ultracentrifugation of 
the deoxycholate-treated microsome suspension. To the supernate from the 
second low-speed centrifugation, 1/9 volume 5 per cent sodium deoxycholate 
(in 0.05 M Tris buffer, pH 8.2) was added with stirring. Stirring was con- 
tinued for 15 minutes at 0° C. The microsomal particles were sedimented by 
centrifugation for 5 hours at 29,000 rpm in the no. 30 rotor of a Spinco model L 
ultracentrifuge. The dark red supernatant solution was removed from the 
microsomal-particle pellet by decantation. The pellet was used without further 
purification for preparing RNA. For studies on the microsomal particles, fur- 
ther centrifugation, both high- and low-speed, was employed to obtain micro- 
somal particles free from contaminating proteins. 

Properties of the Microsomal Particles. The degree of homogeneity of the 
particles is revealed by the sedimentation diagram shown in figure 1. Extra- 
polation of numerous measurements at various concentrations in the 0.025 M 

Fig. 1. Sedimentation diagram of microsomal particles in 0.025 M NaHCO s , 0.004 M 
MgCl 2 , pli 8.5. Picture taken at bar angle of 32° after 24 minutes at 27,690 rpm. ^ 20 = 78.7 
at this concentration (1.2 g/dl) ; ^° 20 = 81.3. 


NaHC0 3 containing 4 mM MgCl 2 (pH 8.5) yielded 81 for i° 2 o,w. The ratio of 
optical densities at 260 and 280 mp was 1.80 in this solvent. The ratio at 260 
and 230 mu was 1.32. Various mild treatments such as heating to 37° C cause 
the appearance of 50 S and 5 S components accompanied by increases in the 
optical density at 260 mp (resembling the denaturation of DNA) and conse- 
quently increases in the values of the optical density ratios reported above. 

Preparation of RNA from the Microsomal Particles. In order to obtain RNA 
of high purity from microsomal particles, two principal steps must of neces- 
sity be included in the procedure: (1) dissociation of the RNA from the pro- 
tein; (2) separation of denatured proteins and other contaminating substances 
from the nucleic acid. This procedure makes use of one anionic detergent, 
sodium lauryl sulfate, to disrupt the protein-nucleic acid complex (by denatur- 
ing the protein and displacing the nucleic acid from cationic groups on the 
protein), and another, sodium xylene sulfonate (Naxonate), to remove the 
denatured protein from solution [5]. 

The pellet of microsomal particles was suspended in 0.01 M versene, pH 7.0. 
In a typical preparation, beginning with 200 g of liver, the volume of the sus- 
pension was 100 ml. The suspension was brought to 20° C, and sufficient solid 
sodium lauryl sulfate was added (with stirring) to bring the concentration to 
4 per cent. Stirring was continued until a clear solution resulted; this was 
allowed to stand for 12 hours at 20° C. At the end of this time, the solution 
was cooled to 5° C, and to it were added 3 volumes of an ice-cold solution of 
0.2 M KCl, 0.01 M versene pll 7.0 containing 12 g sodium xylene sulfonate per 
100 cc. The pll of this mixture was reduced to 4.3 by dropwise addition of 6 TV 
acetic acid. After standing 15 minutes at 0°, the suspension was centrifuged 
for 30 minutes at 1500g- in the cold. The supernatant solution was decanted 
and brought to pH 7.0 by addition of 6 N NH 4 OH. After the solution was 
warmed to 20° C, RNA was precipitated with 2 volumes isopropyl alcohol. 
The precipitate was allowed to settle for 2 hours; then it was centrifuged down. 
The liquid was decanted, and the RNA was dissolved in 0.03 M sodium acetate 
solution. This solution was treated with Naxonate to complete the removal of 
protein. To it were added 3 volumes 40 per cent Naxonate; then the solution 
was stirred for 30 minutes. After cooling to 0° C, the pH was brought to 4.3 
and the solution was filtered through celite and sintered glass. 1 This treatment 
removed the protein-Naxonate complex, which is insoluble, leaving glycogen 
as the only nondialyzable impurity in the RNA. Glycogen may conveniently 
be removed by centrifugation for 20 minutes at 30,000^ (it forms a pellet). 
After the removal of glycogen, the RNA solution was dialyzed against 0.01 M 
KH2PO4-K2HPO4 (1:1) in order to remove ultraviolet-absorbing impurities. 

1 It has recently been found that a substantial part of the RNA (up to 85 per cent of the 
total) is lost in this step because of adsorption on the celite. Besides lowering the yield, 
this may have led to fractionation of the RNA, if the adsorption was selective. This step 
may be omitted, for the protein-Naxonate aggregates can be removed along with glycogen 
in the centrifugation at 30,000^. 




The RNA prepared in the manner described above exhibited values of sedi- 
mentation constant and intrinsic viscosity that depended very much on the 
solvent employed, and invariably these quantities slowly diminished with time. 
Choosing 0.01 M phosphate buffer at pYL 7.0 as solvent, typical preparations 
would have initially sedimentation constants of 9 and intrinsic viscosities of 0.6. 
The sedimentation pattern as observed in ultraviolet optics was rather broad 
and usually single peaked. The use of the above values in the Mandelkern- 
Flory equation [6] gave approximately 200,000 as the molecular weight. 

In studying the gradual decay in molecular weight it was observed that heat- 
ing to 60° C for a few minutes or the addition of KCNS to the extent of 0.2 M 
would cause a 30 per cent fall in molecular weight. This indicates that the 
polynucleotide chains of the original material were aggregated to a small but 
significant extent. 

When the RNA was heated to progressively higher temperatures in the 
0.01 M phosphate buffer the molecular weight was observed to continue falling 
up to temperatures of about 80° C. After exposure to 80° the molecular weight 
remained unchanged with time at room temperature, for prolonged periods at 
83° and for short periods at 95°. Thus a single exposure to about 80° C pro- 
duced a stabilization of the RNA. 

The changes which such heating induced in the sedimentation pattern are 
shown in figure 2, where the sedimenting boundaries observed with ultraviolet 
optics under identical conditions are shown for 0, 5, 15, and 45 minutes' heating 
at 83° C. It is seen that the effect of the heating is to lower and to narrow the 

I oo 


C /c 





Relotive concentration versus 
distance from center of rototion 

Solvent -OIM KH 2 P0 4 - K 2 HP0 4 , pH 7 
RNA concentration 38 if/cc. 

6.2 6 4 


6 6 

Fig. 2. Effect of heating upon RNA sedimentation boundaries after 21 minutes at 
59,780 rpm. 


sedimentation distribution; the change is nearly completed during the first 5 
minutes of heating. The intrinsic viscosity showed a similar behavior. Both 
these changes indicate a fall in apparent molecular weight to a limiting value. 
The sharpening of the sedimentation distribution suggests a sharpening of the 
molecular-weight distribution as well. Consequently the fall in molecular 
weight appears to result from the dissociation of an aggregate and not the hy- 
drolysis of phosphate ester bonds. 

In order to proceed with the investigation of RNA stabilized by heating, a 
standard procedure was adopted: solutions in 0.01 M phosphate buffer were 
heated at 83° C for 10 minutes and then cooled to room temperature for 


The intrinsic viscosity and sedimentation constant measured for a number 
of preparations are listed in table 1. Recent results, shown in the lower part 
of the table, have consistently given sedimentation constants near 7 S, in con- 
trast to variable and lower results obtained earlier, which are listed in the upper 
part of the table. RNA preparations having these higher and more consistent 
sedimentation values were obtained by improving the separation of the micro- 
somal particles from other cellular fractions (thereby reducing ribonuclease 
contamination) and by maintaining a temperature of 5° or less at all stages 
of the preparation in which RNA might be attacked. 

For a typical, improved preparation of stabilized RNA with s 2 j = 6.6 S and 
the intrinsic viscosity [•/]] =0.29, a molecular weight (weight average) of 106,000 
can be calculated from the Mandelkern-Flory equation [6] using 2.35 X 10 6 for 
(4> 1/3 /P) and 0.55 for the partial specific volume. Light-scattering measure- 
ments on the same sample yielded the same value for the weight-average 
molecular weight. 

TABLE 1. Sedimentation Constant, Intrinsic Viscosity, and Molecular 
Weight of RNA Preparations 

Measured in 0.01 M phosphate, pH 7.0, after heating to 83° for 10 minutes. 

*°25 M25 M s 3 w 

2.9 0.14 21,700 

4.1 0.17 40,700 
4.6 0.19 50,400 

5.0 ... 

6.2 0.26 93,000 
6.6 0.29 106,000 
6.6 ... 

7.1 ... 

7.4 0.27 120,000 

6.8 ±0.4 0.27 106,000 



The agreement between the light-scattering molecular weight and that cal- 
culated from sedimentation and viscosity data justifies the application of 
the Mandelkern-Flory relation to the lower-molecular-weight samples listed in 
the upper part of table 1. The molecular-weight dependence of these two quan- 
tities can then be examined. This is done in figure 3, where the logarithms of 
s° and [y\] are plotted against the logarithm molecular weight, yielding the 
linear relations 

/=2.1xl0" 2 M 0A9 
[n]=:6.2xl0- 4 M - 53 

This type of dependence is associated with homologous samples of linear, ran- 
domly coiled polymer chains. These exponents are close to the limiting value 
of 0.5 which is reached for chains having the maximum permissible extent of 
coiling [7], So high a degree of coiling is unexpected in a highly charged poly- 
electrolyte at the relatively low ionic strength used here and must be taken to 
indicate that the intrachain attractions are strong enough to overcome the ex- 
pansive electrostatic effect. 

Provided that RNA is a randomly coiled, single chain, we should expect the 
relatively tight coiling to give way to a much more expanded coil in the ab- 
sence of added electrolyte. This would be recognized by a much higher viscosity 
and the further increase in the reduced specific viscosity upon dilution with 
water, a behavior known as the electroviscous effect in When 
the RNA is transferred to aqueous solution (pH 5), its reduced specific 
viscosity at 0.6 g/dl is found to be 0.85 and to increase strikingly upon dilution. 
These results, shown in figure 4, clearly indicate the progressive expansion of the 





Fig. 3. Logarithmic dependence of the sedimentation constant and intrinsic viscosity of 
RNA upon the molecular weight. 



molecule as its counterions become further removed from it. Upon the addi- 
tion of salt the viscosity returns to a low value, showing the reversibility of 
the effect. Thus the solution properties of the RNA are found to be satisfac- 
torily correlated with a randomly coiled, single-chain structure. 


In view of the substantial nucleotide attractions that appear to be necessary 
to account for the relatively tight coiling of RNA in 0.01 M phosphate buffer 
we sought to examine the possibility that hydrogen bonds between pairs of 
nucleotides were present in a random fashion within each coiled RNA mole- 
cule. Since the break-up of hydrogen bonds between base pairs in the denatura- 
tion of deoxyribose nucleic acid is accompanied by a substantial rise in ultraviolet 
absorbance [8], it seemed reasonable to look for the same effect in RNA. The 
hydrogen bonds that may exist in RNA would be broken during the expansion 
that was seen to accompany the removal of salt from the RNA solution. Alter- 
natively such hydrogen bonds may be expected to be broken with increasing 
temperatures, as occurs with deoxyribose nucleic acid. We have combined these 
two hydrogen-bond-breaking effects by measuring the optical density of RNA 
solutions at 258 mu as a function of temperature at various ionic strengths. 

The results of this study are shown in figure 5. It is seen that in 0.01 M 




.600 - 


0.2 4 6 


Fig. 4. Reduced specific viscosity of 
RNA in water; dependence on concen- 


Fig. 5. Variation of the optical density of 
RNA solutions at 258 m« with temperature and 
ionic strength. 


phosphate buffer the optical density rises at once as the temperature is raised 
above room temperature and reaches a maximum value, 28 per cent higher, at 
about 80° C. At higher ionic strength the rise is similar but does not begin 
until a higher temperature is reached. At lower ionic strength the optical 
density at room temperature is already considerably above the lower limiting 
value. Thus at the reduced ionic strength, where viscosity measurements show 
that the molecule is partly expanded, these measurements indicate that the 
hydrogen bonding is correspondingly reduced. These observations are consist- 
ent with the initial hypothesis and support the view that the intrachain attrac- 
tions in RNA arise from hydrogen bonding between the purine and pyrimidine 
bases. The pairing would presumably be similar to that existing in DNA, but 
this fact carries no implication that the base pairs would be periodically 

Since the magnitude of the hypochromic effect (about 28 per cent) is more 
than half the total that results from the hydrolysis of RNA [9], the implica- 
tion is that a large fraction of the purine and pyrimidines participate in the 
pairing when the ionic strength is 0.01 M or more. The fraction of base pairs 
involved appears to be unchanged by the heating cycle. Indeed, the first heat- 
ing cycle of freshly prepared RNA shows the same results as successive cycles, 
in contrast with DNA, where the optical density never returns to the original 
value after the first heating. One would therefore conclude that there is no 
periodic arrangement of base pairs in RNA as there is in DNA. 

Finally, it is of interest to note that the heat treatment used to stabilize the 
RNA, heating to 83° C of a 0.01 M phosphate buffer solution, is precisely the 
treatment required to reach the maximal optical density and presumably break 
the hydrogen bonds. Consequently the drop in apparent molecular weight pro- 
duced by the heat treatment may have done nothing other than permit some 
aggregates of RNA molecules to be dissociated through the opening-up of the 
hydrogen bonds holding them together. It remains for future work to show 
whether or not this interchain bonding is a remnant of structural organization 
of RNA in the microsomal particle. 


We are deeply indebted to Dr. Norman S. Simmons, Atomic Energy Project, 
University of California at Los Angeles, for his helpful advice and discussions 
regarding the preparation of RNA. One of us wishes to thank also the Na- 
tional Science Foundation and the Union Carbon and Carbide Corporation for 
fellowship support during the course of this investigation. This investigation 
was supported by the National Institutes of Health (C2170) . 

Note Added in Proof 

Using aqueous phenol for removal of protein, we have isolated from calf-liver 
microsomal particles RNA of substantially higher molecular weight (~1 X 10°) 
than that obtained in detergent RNA preparations. This large RNA appears 



to be composed of subunits which are single-chain coils of the type described 
here. When it is heated to 85° in 0.02 M phosphate buffer, the molecular weight 
falls to 120,000, the intrinsic viscosity to 0.22, and the sedimentation constant 
to 7.5. 


1. J. W. Littlefield, E. B. Keller, J. Gross, 
and P. C. Zamecnik, /. Biol. Chem., 211 , 
111 (1955). 

2. P. C. Zamecnik, E. B. Keller, M. B. 
Hoagland, J. W. Littlefield, and R. B. 
Loftfield, Ciba Foundation Symposium on 
Ionizing Radiations and Cell Metabolism, 
pp. 161-168, Churchill, London, 1956. 

3. M. L. Petermann, N. A. Mizen, and 
M. G. Hamilton, Cancer Research, 16, 620 

4. M. L. Petermann and M. G. Hamil- 
ton, /. Biol. Chem., 224, 725 (1957). 

5. N. Simmons, private communication. 

6. L. Mandelkern and P. J. Flory, 
/. Chem. Phys., 20, 212 (1952). 

7. P. J. Flory, Principles of Polymer 
Chemistry, p. 622, Cornell University Press, 
Ithaca, N. Y., 1953. 

8. R. Thomas, Biochim. ct Biophys. 
Acta, 14, 231 (1954). 

9. B. Magasanik in The Nucleic Acids, 
vol. 1, p. 393, Academic Press, New York, 


Microsomes and Ribonucleoprotein Particles 


The Laboratories of the Rockefeller Institute for Medical Research 

I should like to present in the following pages a short history of the develop- 
ment of our present concepts on microsomes and ribonucleoprotein (RNP) 
particles. I consider the historical background of this field of research interesting 
in itself; moreover, I believe that its knowledge may throw some light on the 
actual relations between microsomes and RNP particles as well as on some basic 
principles of biological organization. 


The discovery of the microsomes was a by-product of work done on virus- 
induced tumors by Albert Claude at the Rockefeller Institute in the late 1930's. 
Trying to purify a tumor-inducing fraction obtained by differential centrifuga- 
tion from breis of Rous sarcomata, 1 Claude found as expected that the prepara- 
tion was rich in ribonucleic acid (RNA) [1], and was inactivated by various 
agents known to affect nucleoproteins and nucleic acids [2]. At the same time, 
however, he unexpectedly discovered that cell fractions, similar in their gross 
chemistry to the tumor-inducing preparations, could be isolated from chick 
embryos [3], and from a variety of tissues, adult as well as embryonic, and 
normal as well as tumorous [4, 5]. From the beginning Claude was convinced 
that these fractions consisted of pre-existing cell structures, not of cytoplasmic 
aggregates artificially produced by tissue grinding. After some hesitation, 2 he 
[4, 5] arrived at the conclusion that the structures involved were new cell 
components of widespread occurrence which had eluded detection by light 
microscopy because they were too close to, or below, the limit of resolution of 

1 Chicken tumor I. 

2 For a while he assumed that the fraction consisted of mitochondria or mitochondrial 



usual light optics. 3 In his early reports he described the new components as 
"small particles" or "small granules"; later he chose the term "microsomes" 
(small bodies) [6, 7], which met with favor and has remained in common use 
ever since. From his centrifugation data, 4 Claude calculated that the micro- 
somes measured ~ 50 to 200 m/* in diameter, and on the strength of his 
chemical analyses he defined them as "phospholipide-ribonucleoprotein com- 
plexes" [4, 5]. In this way and thus defined the microsomes entered the bio- 
chemical thinking of our times. 

Claude's discovery was followed by an extensive period of biochemical re- 
search which confirmed and greatly extended his findings. For practical rea- 
sons, the research effort was concentrated almost exclusively on liver, and as a 
result the voluminous literature thereby produced (for reviews see [8, 9, 10]) 
applies primarily to liver microsomes, not, as usually assumed, to microsomes in 
general. According to this literature, the dominant biochemical feature of the 
microsomal fraction is its high RNA content: ~ 40 to 50 per cent of the RNA 
of the tissue brei is usually recovered in the microsomes, together with ~ 15 per 
cent of its proteins. Consequently the microsomal RNA/protein ratio is high— 
until recently higher than that of any other cell fraction. It should be men- 
tioned, however, that, with all this concentration, the RNA does not represent 
more than ~ 10 per cent of the microsomal dry weight. Another apparently 
characteristic feature of the fraction is its large content and high concentration 
of phospholipides : ~ 50 per cent of the phospholipides of the tissue are re- 
covered in the microsomes. As far as biochemical activities are concerned, the 
microsomes are distinguished by high concentrations of diphosphopyridine 
nucleotide-cytochrome c reductase [11], cytochrome m or b- [12, 13], and glu- 
cose-6-phosphatase [14], and especially by their ability to incorporate labeled 
amino acids into their proteins both in vivo [15-19] and in vitro [20-22]. Ac- 
cording to current interpretations, the last property indicates that the micro- 
somes are, or contain, the sites of protein synthesis of the cytoplasm. 

In contrast with the active and diversified work on the biochemical aspects 
of the problem, research on the identity of the microsomes, or of their pre- 
cursors, inside the intact cell made little progress because of a number of tech- 
nical limitations. Of the instruments available for morphological investigation, 

3 The microsomes can be seen as distinct particles in the dark-field microscope [5], and 
as a shimmering mass of indistinct small bodies in the light microscope, especially under 
phase contrast optics. 

4 For Claude the microsomes were the fraction sedimented in 1 hour at 18,0(% from the 
supernatant of the "large granule" (mitochondrial) fraction. The medium he used in pre- 
paring tissue "extracts" was either water or dilute phosphate buffer or 0.15 M NaCl. When 
sucrose solutions were introduced in cell-fractionation procedures, the centrifugal force was 
increased to compensate for the higher density and viscosity of the new medium. At 
present the microsomes are usually separated by centrifuging a mitochondrial supernatant 
for 1 hour at ~ 100,00%, irrespective of the sucrose concentration in the suspending 
medium. Since this concentration varies from 0.25 to 0.88 M, the microsome fractions 
described in the literature are not strictly comparable to one another. 


one — the light microscope — did not have enough resolving power, and the other 
— the electron microscope, introduced in biological research at the time of the 
discovery of the microsomes — could provide the necessary resolution but pre- 
paratory techniques for the electron microscopy of biological specimens were 
still inadequate. As a result the microsomes remained until recently a cyto- 
chemical concept only, without a known structural counterpart in the organiza- 
tion of the intact cell. Many cytologists and cytochemists even doubted that 
these small bodies were derived from a pre-existing cell structure and were 
inclined to consider them artifacts due to cytoplasmic clumping or mitochon- 
drial fragmentation [23] during the homogenization of the tissue. 

Because of their characteristically high content of RNA, the microsomes were 
rather early correlated with the so-called "basophil substance" of the cytoplasm. 
Although the correlation was based on circumstantial evidence obtained on a 
limited number of cell types [7, 24], it was soon assumed to be generally valid, 
and consequently the microsomes and the basophil cytoplasm began to be re- 
garded as equivalent terms in two different technical vocabularies. An apparent 
convergence was thus effected between two lines of cytochemical research: an 
old line exemplified by the work of Brachet's (cf. [25]) and Cassperson's (cf. 
[26]) groups and aiming at the localization of nucleic acids in situ, and a new 
line based on cell-fractionation procedures. The correlation appeared to be 
further strengthened when the large body of circumstantial evidence accumu- 
lated by Brachet and Cassperson on the role of RNA in protein synthesis 
received full and repeated support from experiments showing that the micro- 
somes were the most active cell fraction in the incorporation of labeled amino 
acids into proteins. 

The lack of morphological information, the concentration of the work on 
liver, and the equation of the microsomes with the basophil substance of the 
cytoplasm had some unfavorable consequences on the development of general 
ideas in the field. It was assumed, for instance, without enough evidence, that 
entities similar or identical to Claude's microsomes existed in all cells, bacterial 
cells included, and it was also believed that the chemical composition of liver 
microsomes was representative for microsomes in general. As we shall see, both 
assumptions are now in need of revision. 


I shall turn now to another development, in a different field, which from the 
beginning seemed to have some connection with our main problem. A few 
years after the discovery of the microsomes, and again in conjunction with work 
done on Rous sarcomata, Porter, Claude, and Fullam \27] and Claude, Porter, 
and Pickels [28] succeeded in obtaining electron micrographs of thinly spread 
cells (avian fibroblasts) maintained in tissue culture. Below the limit of reso- 
lution of the light microscope and well within the range of calculated micro- 
somal sizes, they found a "lace-like" network of slightly higher density than the 


rest of the cytoplasm. Porter, continuing the study of this structure in cultured 
material, arrived at the conclusion that it consisted of vesicles and tubules 
interconnected in a continuous network. Since the network was restricted to 
the inner or endoplasmic region of the cytoplasm, he proposed the name endo- 
plasmic reticulum for the entire structure [29, 30]. The discovery of this net- 
work provided a likely candidate for the role of microsomal precursor within 
the intact cell, and indeed both Claude [31] and Porter [30] speculated that 
the microsomes might represent derivatives of the endoplasmic reticulum, but 
the assumption could not be verified at the time because the electron microscopy 
of cells and tissues was still in infancy. The examination of a liver cell, for 
instance, was a difficult and uncertain project; in fact, the first attempt [32] 
to identity "microsomes" in electron micrographs of sectioned hepatic cells 
yielded misleading results. 5 

In the early 1950's, however, a succession of technical improvements covering 
the whole series of preparative steps [33, 34] but affecting primarily microtomy 
[35, 36] made possible the examination of thin sections of practically all cell 
types in the electron microscope. With this spectacular breaking through the 
barriers of technical difficulties, the search for the intracellular equivalents of 
the microsomes finally became possible. But most electron microscopists en- 
gaged at that time in the study of the fine structure of the cell were not directly 
interested in the microsome problem; they were rather attracted by a related 
question: namely that of the structural substrate of cytoplasmic basophilia. 
Apparently it did not matter, because, as already mentioned, the two problems 
were expected to have a common solution. In the newly opened realm of fine 
cellular organization the microscopists found an unsuspected and, at the begin- 
ning, puzzling abundance of structures. After a few years spent in deciphering 
these findings and in arguing about various interpretations, 6 it became clear 
that the ground substance of the cytoplasm contained an extensive system of 
spaces, described as vesicles, tubules, and cisternae 7 (figs. 1 and 2), limited by 
a thin membrane (~ 7 m/*) and interconnected in a more or less continuous 

It was soon realized that the so-called ground substance of the cytoplasm was 
divided into two distinct phases by the existence of this internal membrane 
system: one represented by the content of the interconnected vesicles, the 
other by the surrounding cytoplasmic matrix. It was also observed that small, 
dense particles, ~ 15 m/x in diameter, appeared to be attached to the membrane 
(on the surface facing the cytoplasmic matrix) in certain parts of the reticulum 
(fig. 2) while other parts remained free. In addition, it was found that similar 
particles occurred apparently freely scattered throughout the cytoplasmic matrix 

r ' Masses of glycogen were apparently taken for microsomes. 

6 Representative samples of the various interpretations advanced can be found in refer- 
ences 37 to 39. 

7 The term designates flat, shallow vesicles which measure only 50 to 70 m(i in depth 
but reach into microns in the other two directions. 


(fig. 2). Work done in our laboratory established that the system corresponds 
to the endoplasmic reticulum of cultured cells [40], and subsequent observa- 
tions showed that a number of local differentiations occur within this con- 
tinuous network which appears to possess, for instance, a rough-surfaced part 
[40, 41], on account of the attached particles already described, and a smooth- 
surfaced [40, 41] or agranular [42] part free of such particles. Frequently the 
elements of the system, particularly its cisternae, show preferred orientation 
and seem to be disposed parallel to one another at more or less regular intervals, 
thus forming stacks or piles of various sizes (figs. 1 and 2). Finally, further w. vr k 
showed that the system varies characteristically from one cell type to another 
and that these variations affect the total volume of the system, the relative extent 
of its rough- and smooth-surfaced parts, as well as the extent of preferential 
orientation encountered within the system [38]. 

The Structural Substrate of Cytoplasmic Basophilia. With the limited infor- 
mation initially available, it was believed that the endoplasmic reticulum as a 
whole was the structural substrate of basophilia, but subsequent observations 
brought forward serious discrepancies between the distribution of the reticulum 
on one side and that of basophilia on the other. It was found, for instance, that 
there are cell types with an intensely basophil cytoplasm in which the endo- 
plasmic reticulum is poorly developed. Such cells, however, have a large popu- 
lation of small, dense particles, most of them freely scattered in the cytoplasm 
[43]. The erythroblasts and the undifferentiated cells of rapidly growing 
epithelia (epithelia of the intestinal crypts, stratum germinativum of the epi- 
dermis) belong to this category. A converse situation is encountered in mature 
leucocytes and in seminal epithelia (rat) whose acidophil cytoplasm contains 
a relatively well developed endoplasmic reticulum, most of it smooth-surfaced, 
but has only a few small particles, free or attached. 

In all the cases examined, the cytoplasmic component whose distribution 
matched best that of the affinity for basic dyes appeared to be represented by the 
small, dense particles. As a result of these findings, in 1953 I advanced [44, 43] 
the hypothesis that these particles, rather than the membranous material of the 
endoplasmic reticulum, contained most of the RNA of the cytoplasm, and that 
they were consequently the sought-for structural substrate of cytoplasmic baso- 
philia. The hypothesis rested upon the results of a broad survey of various cell 
types which covered a large number of "test specimens," i.e., cells known for 
the intensity and characteristic distribution of "basophil substance" in their 
cytoplasm, as well as cells known for their cytoplasmic acidophilia. The postu- 
late derived additional support from the fact that at that time small particles in 
the same size range and containing a large amount of RNA had already been 
isolated from yeast and bacterial cells [45] by Schachman et al., and from certain 
mammalian tissues, such as liver and spleen, by Petermann et. al. [46, 47]. 

From the beginning the hypothesis implied that the small particles and the 
endoplasmic reticulum represent two basically distinct components of the cyto- 
plasm which may exist and develop independently of each other. Their close 


and more or less extensive association was considered a secondary phenomenon 
that occurs at a relatively late stage in the evolution of cellular organization. 
The view, originally based on findings on undifferentiated or embryonic cells, 
subsequently received full support from electron-microscope studies of various 
bacteria [48, 49, 50] which revealed that bacterial protoplasm contains a large 
population of small, dense particles (usually smaller than those found in animal 
cytoplasm), but apparently no internal membranous system comparable to the 
endoplasmic reticulum [50]. 

A Correlated Morphological and Biochemical Analysis of Hepatic Micro- 
somes. The next task was to put the hypothesis to a test by trying to find out in 
what cell fractions the particles segregate during the differential centrifugation 
of tissue breis. Fraction chemistry and particle distribution were correlated by 
using duplicate pellets of the fractions under study : one pellet for biochemical 
analysis, and the other for electron microscopy after appropriate fixation, em- 
bedding, and sectioning. It was found both necessary and expeditious to fix the 
pellets in toto, and to cut them in such a way as to be able to survey them from 
top to bottom. With such precautions, the existence and the extent of inter- 
contamination among cell fractions could be easily detected and the presence of 
distinct layers in some pellets clearly demonstrated. 

The work, carried out in collaboration with Dr. Philip Siekevitz, started 
with an analysis of the microsomal fraction isolated from rat-liver breis [51]. 
We found that this fraction consists almost exclusively of closed vesicles limited 
by a dense continuous membrane, ~ 7 mp thick, and filled with a material of 
relatively low density. Most of these vesicles are derived from the rough-surfaced 
part of the endoplasmic reticulum as indicated by the small (~ 15 mp), dense 
particles attached to the outer surface of their limiting membrane. Smooth- 
surfaced vesicles are also present in the microsomal fraction, but their origin is 
more difficult to ascertain; they may represent fragments of the smooth-surfaced 
part of the reticulum or they may be derived from other sources (Golgi com- 
plex? 8 cell membrane?). In 0.88 M sucrose, the medium used in our experi- 
ments, the microsomal vesicles retained the flattened appearance of intracellular 
cisternae and reacted like osmometers to changes in the concentration of the 
medium: they swelled in hypotonic media. Treatment with versene (2 per cent 
in 0.88 M sucrose, for 60 minutes at 0° C) removed ~ 60 per cent of the micro- 
somal RNA and resulted in extensive loss of attached particles. Incubation in 
ribonuclease (0.5 mg/ml 0.88 M sucrose; 60 minutes at 37° C) caused RNA 
losses of ~ 85 per cent and produced a heavy agglutination of microsomal 
vesicles. During the incubation, the attached particles apparently were lost. 
Finally treatment with sodium deoxycholate (DOC) (0.5 per cent in 0.88 M 
sucrose, at pH 7.5) "solubilized" most of the protein and phospholipides of the 

8 According to our interpretation [38], the "Golgi complex" is a differentiated part of 
the endoplasmic reticulum. Other cytologists [52] consider this structure a distinct and 
independent cell organelle. 


microsomes, but left ~ 80 per cent of their RNA and ~ 15 per cent of their 
protein still in sedimentable form. The pellets obtained from DOC-treated 
microsomes consisted of small, dense particles ~ 10 to 15 rw* in diameter with 
a small admixture of vesicles. We interpreted the results as indicating that 
DOC "solubilizes" the membrane and content of the microsomal vesicles while 
affecting their attached particles to a lesser extent, and we inferred from these 
experiments that the small, dense particles consist of ribonucleoprotein and con- 
tain most of the microsomal RNA (~ 80 per cent), while the membrane and 
content account for most of the protein and almost all the phospholipides of 
the microsomes [51]. 

We soon found out that, contrary to assumptions then current, the final 
supernatant still contained numerous structured elements of membranous or 
particulate nature, and accordingly we made an attempt to separate these ele- 
ments by further centrifugation of the microsomal supernatant, the fraction 
usually considered the liquid phase of the cytoplasm or the "cell sap." We iso- 
lated two successive postmicrosomal fractions ° in the hope that one of them 
might consist mainly of free particles, but we found the corresponding pellets 
to be mixtures of smooth-surfaced vesicles (probably derived from the smooth- 
surfaced part of the reticulum), free particles, and amorphous material. The 
RNA content of these postmicrosomal fractions was low, and its concentration 
therein considerably lower than in the microsomes [51]. 

In the case of the liver we succeeded, therefore, in identifying the microsomes 
as fragments of the endoplasmic reticulum, derived primarily from its rough- 
surfaced part. We obtained evidence indicating that they are closed vesicles, 
and that most of their RNA is present as ribonucleoprotein in their attached 
particles. We did not succeed in isolating the free particles of the cytoplasmic 
matrix, nor did we obtain a "clean" preparation of smooth-surfaced vesicles. The 
information on the gross chemistry of microsomal membranes was indirect and 
relied on subtractions and many assumptions. 


As the next object for testing our hypothesis, we chose the pancreas of the 
guinea pig with the following considerations in mind. The exocrine cells, 
which form the bulk of the cell population of the gland, have an endoplasmic 
reticulum remarkable in its large volume and in the small extent of its smooth- 
surfaced part (figs. 1 and 2). In addition, their cytoplasmic matrix contains free 
particles in great numbers (fig. 2). With the new material, therefore, we stood 
a better chance to obtain a more homogeneous microsomal fraction and to 
separate free particles in a postmicrosomal fraction. The results met our expecta- 

9 The first postmicrosomal fraction (PMj) was obtained by centrifuging the microsomal 
supernatant for 2 hours at 105,000^; the second postmicrosomal fraction (PM 2 ), by 
centrifuging the supernatant of PM : for 16 hours at 105,000 g. 


tions. We found that pancreatic microsomes [53] are closed vesicles (figs. 4 
and 5) almost exclusively derived from the rough-surfaced part of the cells' 
voluminous reticula. They are, however, more labile structures than their 
hepatic counterparts: even when isolated in 0.88 M sucrose they enspherulate 
and do not retain the flattened cisternal form that the elements of the endo- 
plasmic reticulum have in situ. They are also more susceptible to various treat- 
ments, especially to deoxycholate, which, at 0.3 per cent final concentration, 
solubilizes ~ 85 per cent of their protein and ~ 40 per cent of their RNA. As 
in the case of the liver, the DOC-insoluble material consists of RNA particles 
(fig. 6) with a relatively high RNA/ protein ratio. The main difference be- 
tween hepatic and pancreatic microsomes concerns their phospholipide content: 
there is ~ 8 times less phospholipide in pancreatic microsomes than in their 
hepatic counterparts, and the phospholipide concentration in the microsomal 
fraction is equal to, or only slightly higher than, that in the original pancreatic 

Centrifugation of the microsomal supernatant resulted in the sedimentation of 
further material which was arbitrarily divided into two postmicrosomal frac- 
tions °( PMi and PM 2 ). We found that the corresponding pellets consist almost 
exclusively of small, dense particles ~ 15 m^ in diameter (fig. 7), with a mod- 
erate admixture of small vesicles in the PMi. Chemically both fractions were 
made up of ribonucleoproteins with a negligible amount of phospholipides. 
In the original fractionation scheme [53], the two postmicrosomal fractions 
contained RNA in comparable concentrations, but in a recent modification 
PMi has a noticeably higher RNA/protein ratio than PM 2 . 

The results obtained with the pancreas confirmed and extended our previous 
findings on liver. In both the microsomes were found to be vesicles derived 
either to a large extent [51], or almost exclusively [53], from the rough-surfaced 
part of the endoplasmic reticulum. We used a characteristic structural detail, 
namely the attached particles, to establish this derivation, and our conclusion 
appears reasonably valid; there is good general agreement between the basic 
morphological features of the microsomes and the fine structural details of the 
endoplasmic reticulum. There is, however, no agreement as far as general dimen- 
sions are concerned: the network is a continuous structure which may spread 
throughout the entire cytoplasm of the intact cell (~ 20 X ~ 30 X ~ 40 fi), 
whereas the microsomes are considerably smaller (0.05 to 0.3 n), corresponding 
in dimensions to the vesicles, tubules, and cisternae which by their interconnec- 
tions form the reticulum. Accordingly we must assume that the network is 
broken during tissue grinding, the resulting fragments being the microsomes. 
Since these fragments are closed vesicles, we are further obliged to postulate that 
the broken segments heal readily into closed structures or that the fragmenta- 
tion of the reticulum occurs by a generalized pinching-ofT process rather than 
by mechanical disruption [51, 53]. 



The view that the small, dense particles represent the structural substrate of 
cytoplasmic basophilia was clearly supported by the isolation of RNP particles 
from DOC-treated microsomes (fig. 6) and further strengthened by the finding 
that the pancreatic postmicrosomal fractions (fig. 7), assumed to represent the 
free particles of the cytoplasmic matrix, are also comprised of ribonucleopro- 
teins. It should be pointed out, however, that the RNP particles, free and at- 
tached, do not account for all the RNA of the cytoplasm. Leaving aside the 
controverted question of the RNA content of mitochondria, we found that the 
microsomes contain a certain amount of RNA, small in the liver but relatively 
large in the pancreas, which is solubilized by DOC treatment, under the condi- 
tions of our experiments, and whose structural connections are unknown. Fi- 
nally there is a relatively small amount of RNA (~ 10 per cent of the RNA of 
the cytoplasm) that remains in the supernatant of the last postmicrosomal frac- 
tion. Evidently we do not know to what extent our findings reflect the situa- 
tion inside the living cell, or to what extent they are affected by preparation 
artifacts. It can be argued, for instance, that attached particles can be detached 
during tissue grinding and fractionation, or that the sedimentation of RNA 
particles is still incomplete by the end of our last centrifugation. 

With such possibilities in mind, Dr. Siekevitz and I tried to find out in more 
recent work [54] whether there are functional differences among the various 
RNP preparations described. In agreement with reports on other tissues [55, 
56], we noted that the attached particles (DOC-insoluble microsomal material) 
are more active than the parental microsomes in the incorporation of labeled 
amino acids into proteins (presumably protein synthesis 10 ). In addition, we 
found that the activity mentioned is considerably higher in the attached particles 
than in the free RNP particles of the postmicrosomal fractions. The second 
postmicrosomal fraction, however, proved to be more active than any other 
particulate fraction in the incorporation of labeled adenine into RNA. In turn, 
its activity was greatly exceeded by that of the "soluble" RNA of the final 

Such metabolic differences, though not entirely excluding preparation arti- 
facts, render them more unlikely, and suggest that a variety of RNP's with dif- 
ferent structural connections and different functions exist within the animal 
cell. I should also add that by comparing microsomes obtained from starved 
and fed animals we observed a relatively large increase in microsomal pro- 
teolytic 11 and ribonuclease activity [57]. What seems to be of considerable 
interest is the finding that a sizable part of the enzymatic activities of the micro- 

10 In our experiments the incorporation was carried out in vivo; in those reported in 
references 55 and 56 the labeled amino acids were incorporated either in vivo or in vitro 
by whole microsomes. 

11 Due mainly to trypsinogen and chymotrypsinogen. These two proteases and the 
ribonuclease are enzymes synthesized on a large scale by the exocrine cells of the pancreas 
to be released in the intestine for the digestion of food. 


somes is found associated with their attached RNP particles. Further work will 
show whether these enzymes are newly synthesized proteins still attached to 
their sites of synthesis or enzymes released from other locations and absorbed 
on RNP particles during tissue grinding. 


It follows from the results thus far summarized that the cytoplasm contains 
more separable entities than were assumed a few years ago. It also follows that 
Claude's microsomes are relatively large and complex structures in which mem- 
branous and particulate components can be easily recognized. There is also a 
microsomal content which, though usually amorphous, may occur as formed 
granules ([57], cf. [58]) under certain conditions, thus increasing the com- 
plexity of the structure. The term "microsomal particles," used in many com- 
munications made at this meeting, can be properly applied to RNP particles 
attached to the surface of microsomal vesicles. As already indicated, such par- 
ticles can be detached by DOC treatment and subsequently collected in rela- 
tively clean preparations. It is doubtful, however, that the same term can be 
used to designate the free RNP particles isolated in postmicrosomal fractions 
and especially RNP particles separated from bacterial cells, in which there is 
no endoplasmic reticulum to start with, and from which no microsomes can be 
obtained. Pellets obtained from bacteria are composed of much smaller and 
simpler cell components which morphologically seem to correspond to the free 
RNP particles of the pancreatic cell. In general, a morphological label is justi- 
fied as long as there is no information on the chemistry and function of the 
structure involved. Here, however, we know that we are dealing with ribonu- 
cleoprotein particles, and consequently there is not too much sense in retaining 
a morphological label, especially after realizing that it is misleading. 

More than accurate terminology is involved in this argument. What is known 
so far about the fine structure of bacterial cells suggests that internal mem- 
branous systems, like the endoplasmic reticulum, are not necessary for the 
organization and function of a simple type of self-sustaining cell. Such mem- 
branous systems appear in more elaborate cell forms and could therefore be 
regarded as superstructures. We do not know what special problems are solved 
by their introduction, but we may wonder whether they are not connected with 
an increase in cell volume, subsequent difficulties in diffusion, and relative 
decrease in available surface. At higher levels of biological organization, similar 
problems are frequently encountered and usually solved by the invagination of 
surface structures and by the concomitant interiorization of part of the sur- 
rounding medium. There is therefore an important difference in organization 
between bacterial cells on one side, and animal and plant cells on the other: 
the superstructure that can be ground into microsomes appears only in the 
latter. Terms such as "microsomes" or "microsomal particles" of bacterial 
origin do not take this basic difference into account. 



During the past 5 years considerable progress has been made in the study of 
cell organization through the extensive use of electron microscopy and cell- 
fractionation procedures. Accordingly it is of interest to see how the hypothesis 
formulated in 1953 has fared through this period of rapid development. 

The assumption that the small particulate component of the cytoplasm is 
the structural substrate of basophilia has remained in good agreement with 
the large majority of the findings made on new and very numerous cell types 
of animal and plant origin (see for examples [59, 60]). The only exception so 
far encountered is represented by the heart muscle of the turtle [61], in which 
a slightly larger particulate was found in an acidophil cytoplasm that gave a 
positive test for glycogen. Consequently it was postulated that particulate glyco- 
gen might be mistaken for RNP particles in certain cell types, especially in 
muscle [61]. The actual isolation of this particulate material from various 
muscle fibers could settle the question, but so far it has not been accomplished. 
It should be pointed out, however, that under prevailing technical conditions a 
certain amount of confusion of the type suggested cannot be excluded. Because 
of the dimensions involved, we are examining only the gross morphology of 
the small particles, and in so doing we are not helped thus far by any charac- 
teristic detail of structure. If particles of different chemical composition and of 
different fractions happen to have the same general size and shape, we cannot 
avoid lumping them together in a common category. Morphological expressions 
that can be distinguished at the present level of practical resolution are un- 
doubtedly less numerous than functional characteristics or macromolecular 
species. It is exactly for this reason that morphological information should be 
supplemented, wherever possible, by biochemical and metabolic data. 

Ribonucleoprotein particles, with a sedimentation constant of 70 to 80 S and 
a calculated or measured diameter of 10 to 15 m/x, have been isolated from 
many new and old sources such as liver [62, 65], yeast [63], ascites cells [56], 
and pea seedlings [64]. They have been described in terms of their gross 
chemistry [62-65], biochemical activities [56], and physicochemical properties 
[62-65]. A perusal of this symposium shows that recently RNP particles have 
also been isolated from a variety of bacterial cells. In general, there is good 
agreement between these findings and our observations, but there is little or no 
information about the existence, frequency, and topography of the particles 
in situ. 

Finally it is still debated whether the microsomal RNA is mainly located in 
the attached particles or is also present in large amounts in the microsomal 
membranes as originally assumed by Kuff et al. [66]. As already mentioned, 
an exclusive RNA location in the attached particles cannot be claimed because 
the particles do not account for ~ 20 per cent of the microsomal RNA in the 
liver and for ~40 per cent in the pancreas. Recently Chauveau et al. [67] 
found that there is no good correlation between the frequency of vesicles with 


attached particles and the amount of RNA present in microsomal pellets. They 
indicate, however, that particles appear after treating predominantly mem- 
branous pellets with DOC, and they suggest that the particles could be masked 
by incorporation into the membrane under certain uncontrolled conditions. It 
is evident that we should know more about the behavior of these particles 
under various metabolic conditions, and about their reaction to various suspen- 
sion media used in tissue fractionation, before arriving at an understanding 
of these conflicting pieces of evidence. 


In conclusion it can be said that small (~ 15 m/*), dense particles have been 
found, either free or attached to the membrane of the endoplasmic reticulum, 
in practically all animal and plant cells thus far examined. Comparable parti- 
cles, usually unattached to membranous structures, exist in bacterial cells. The 
distribution of these particles is largely similar to that of cytoplasmic basophilia, 
by implication to that of cytoplasmic RNA. 

Ribonucleoprotein particles of relatively small size (10 to 20 mn and 40 to 
80 S) have been isolated by various procedures, such as ultracentrifugation and 
electrophoresis, from a variety of animal, plant, and bacterial sources. Evidence 
that the small, dense particles seen in the intact cell consist of ribonucleoprotein 
has been obtained for the liver (rat) and the pancreas (guinea pig). Accord- 
ingly the assumption that these particles are the structural substrate of basophilia 
has been verified for two cell types only. For all the others, it remains what 
it has been, a hypothesis to be tested by further work. Integrated studies pro- 
viding an adequate coverage from cell to pellets are evidently needed for more 
kinds of tissues. 

Although it appears that the RNP particles are cytoplasmic components of 
widespread occurrence, probably basic structural elements in the organization 
of the cell, many points in their history are still uncertain or controverted, and 
many pertinent questions remain unanswered. For instance, though there is 
good agreement about the presence of RNA in particles, there is still doubt 
about the presence or absence of RNA in membranous structures, primarily in 
the membrane of the endoplasmic reticulum. Considerable variation occurs, 
apparently connected with the methods of preparation, in the RNA content of 
these particles; accordingly one would like to know the procedure by which 
the situation in situ is more closely approximated. There is morphological, 
physicochemical, and metabolic diversity among these particles, but very little 
is known about the way in which the various differences are correlated, or 
about the significance of this diversity. In this respect one may wonder whether 
particles with different locations and activities represent distinct, fully developed 
cell organs, or whether they correspond to successive stages in the differentiation 
of a single or a few cell organs. What seems to be particularly disturbing at 
present is the meagerness of our information about the functional role of these 



particles. Without more knowledge on this aspect, progress is uncertain even 
if it takes the appearance of elegant physicochemical data. 


1. A. Claude, /. Exptl. Med., 66, 59 

2. A. Claude and A. Rothen, /. Exptl. 
Med., 71, 619 (1940). 

3. A. Claude, Proc. Soc. Exptl. Biol. 
Med., 39, 398 (1938). 

4. A. Claude, Science, 91, 77 (1940). 

5. A. Claude, Cold Spring Harbor 
Symposia Quant. Biol, 9, 263 (1941). 

6. A. Claude, Science, 97, 451 (1943). 

7. A. Claude, Frontiers in cytochemis- 
try, in Biol. Symposia, 10, Jaques Cattell 
Press, Lancaster, Pa., 1943. 

8. C. de Duve and J. Berthet, Intern. 
Rev. Cytol, 3, 225 (1954). 

9. G. H. Hogeboom and W. C. Schnei- 
der, The cytoplasm, in The Nucleic Acids, 
2 (E. Chargaff and J. N. Davidson, edi- 
tors), Academic Press, New York, 1955. 

10. G. H. Hogeboom, E. L. KufT, and 
W. C. Schneider, Intern. Rev. Cytol, 6, 
425 (1957). 

11. G. H. Hogeboom, /. Biol Chew.., 
177, 847 (1949). 

12. C. F. Strittmatter and E. G. Ball, 
/. Cellular Comp. Physiol, 43, 57 (1954). 

13. M. Bailie and R. K. Morton, Na- 
ture, 176, 111 (1955). 

14. H. G. Hers, J. Berthet, L. Berthet, 
and C. de Duve, Bull. soc. chim. biol, 
33, 21 (1951). 

15. H. Borsook, C. L. Deasy, A. }. 
Haagen Smit, G. Keighley, and P. H. 
Lowry, /. Biol Chem., 196, 669 (1952). 

16. T. Hultin, Exptl. Cell Research, 1, 
376, 599 (1950). 

17. N. D. Lee, J. T. Anderson, R. 
Miller, and R. H. Williams, /. Biol. 
Chem., 192, 733 (1951). 

18. E. P. Tyner, C. Heidelberger, and 
G. A. Le Page, Cancer Research, 13, 
186 (1953). 

19. R. M. S. Smellie, W. M. Mclndoe, 
and J. N. Davidson, Biochcm. et Biophys. 
Acta, 11, 559 (1953). 

20. P. Siekevitz, /. Biol. Chem., 195, 
549 (1952). 

21. V. Allfrey, M. N. Daly, and A. E. 
Mirsky, /. Gen. Physiol, 37, 157 (1953). 

22. P. C. Zamecnik and E. B. Keller, 
/. Biol Chem., 209, 337 (1954). 

23. D. E. Green, /. Cellular Comp. 
Physiol, 39, suppl, 2, 75 (1952). 

24. S. Brenner, S. African J. Med. Sci., 
12, 53 (1947). 

25. J. Brachet, Chemical Embryology, 
Interscience, New York, 1950. 

26. T. Cassperson, Cell Growth and 
Cell Function, Norton, New York, 1950. 

27. K. R. Porter, A. Claude, and E. F. 
Fullam, /. Exptl. Med., 81, 233 (1945). 

28. A. Claude, K. R. Porter, and E. G. 
Pickels, Cancer Research, 7, 421 (1947). 

29. K. R. Porter and F. L. Kallman, 
Ann. N. Y. Acad. Sci., 54, 882 (1952). 

30. K. R. Porter, /. Exptl Med., 97, 
727 (1953). 

31. A. Claude, Harvey Lectures, 48, 
121 (1947-1948). 

32. A. Claude and E. F. Fullam, /. 
Exptl. Med., 83, 499 (1946). 

33. S. B. Newman, E. Borysko, and M. 
Swerdlow, /. Research Natl. Bur. Stand- 
ards, 43, 183 (1949). 

34. G. E. Palade, /. Exptl. Med., 95, 285 

35. K. R. Porter and J. Blum, Anat. 
Record, 117, 685 (1953). 

36. F. S. Sjostrand, Experientia, 9, 114 

37. W. Bernhard, A. Cautier, and C. 
Rouillier, Arch. anat. micr. et morphol. 
exp.,43, 236 (1954). 

38. G. E. Palade, /. Biophys. Biochcm. 
Cytol, 2, suppl, 85 (1956). 

39. F. S. Sjostrand, Electron microscopy 
of cells and tissues, in Physical Tech- 
niques in Biological Research, III (G. 
Oster and A. W. Pollister, editors), Aca- 
demic Press, New York, 1956. 

40. G. E. Palade and K. R. Porter, /. 
Exptl. Med., 100, 641 (1954). 

41. G. E. Palade, /. Biophys. Biochem. 
Cytol, 1, 567 (1955). 



42. S. L. Palay and G. E. Palade, /. 
Biophys. Biochem. Cytol, 1, 69 (1955). 

43. G. E. Palade, /. Biophys. Biochem. 
Cytol., 1, 59 (1955). 

44. G. E. Palade, /. Applied Phys., 24, 
1419 (1953). 

45. H. K. Schachman, A. B. Pardee, 
and R. Y. Stanier, Arch. Biochem. Bio- 
phys., 38, 245 (1952). 

46. M. L. Petermann and M. G. Hamil- 
ton, Cancer Research, 12, 373 (1952). 

47. M. L. Petermann, N. A. Mizen, and 
M. G. Hamilton, Cancer Research, 13, 372 

48. G. B. Chapman and J. Hillier, /. 
Bacterial., 66, 362 (1953). 

49. A. Birch Andersen, O. Maal0e, and 
F. S. Sjostrand, Biochim. Biophys. Acta, 
12, 395 (1953). 

50. J. R. D. Bradfield, Organization of 
bacterial cytoplasm, in Bacterial Anatomy, 
University Press, Cambridge, 1956. 

51. G. E. Palade and P. Siekevitz, /. 
Biophys. Biochem. Cytol, 2, 171 (1956). 

52. A. J. Dalton and M. D. Felix, 
Symposia Soc. Exptl. Biol., 10, 148 (1957). 

53. G. E. Palade and P. Siekevitz, /. 
Biophys. Biochem. Cytol., 2, 671 (1956). 

54. P. Siekevitz and G. E. Palade, 
Federation Proc, 17, 311 (1958); /. Bio- 
phys. Biochem. Cytol., 4 (in press) (1958). 

55. J. W. Littlefield, E. B. Keller, J. 

Gross, and P. C. Zamecnik, /. Biol. Chem., 
217, 111 (1955). 

56. J. W. Littlefield and E. B. Keller, 
/. Biol. Chem., 224, 13 (1957). 

57. P. Siekevitz and G. E. Palade, /. 
Biophys. Biochem. Cytol., 4 (in press) 

58. G. E. Palade, /. Biophys. Biochem. 
Cytol., 2, 417 (1956). 

59. A. F. Howatson and A. W. Ham, 
Can. J. Biochem. Physiol., 35, 549 (1957). 

60. H. A. Lund, A. E. Vatter, and J. B. 
Hanson, /. Biophys. Biochem. Cytol., 4, 
87 (1958). 

61. D. W. Fawcett and C. C. Selby, 
/. Biophys. Biochem. Cytol, 4, 63 (1958). 

62. M. L. Petermann, M. G. Hamilton, 
and N. A. Mizen, Cancer Research, 14, 
360 (1954). 

63. F. C. Chao and FI. K. Schachman, 
Arch. Biochem. Biophys., 61, 220 (1956). 

64. P. O. Ts'o, J. Bonner, and J. Vino- 
grad, /. Biophys. Biochem. Cytol, 2, 451 

65. M. L. Petermann and M. G. Hamil- 
ton, /. Biol Chem., 224, 725 (1957). 

66. E. L. Kuff, G. H. Hogeboom, and 
A. J. Dalton, /. Biophys. Biochem. Cytol, 
2, 33 (1956). 

67. J. Chauveau, Y. Moule, and C. 
Rouillier, Exptl Cell Research, 13, 398 

Fig. 1. The electron micrograph shows part of two adjacent exocrine cells in the 
pancreas of a guinea pig. The apposed membranes of the two cells appear obliquely sec- 
tioned at cm. 

The basal region of one of these cells occupies the lower half of the figure and char- 
acteristically contains a few mitochondrial profiles (m) and numerous profiles of the 
endoplasmic reticulum (rs), which belong to the rough-surfaced type and show a certain 
amount of preferred orientation, i.e., are disposed in more or less parallel rows. 

The apical region of the cell, in the upper part of the figure, is occupied by a few circular 
profiles of zymogen granules (z) . Their dense content consists of stored digestive enzymes 
and enzyme precursors. Elements of the endoplasmic reticulum appear disposed at random 
among the zymogen granules. Part of the centrosphere region of the cell, with its char- 
acteristic clusters of smooth-surfaced vesicles, can be seen at cs. 

A region similar to the lower right quarter of this figure appears in figure 2 at a higher 

Specimen fixed for 2 hours at 0° C in 1 per cent Os0 4 in veronal acetate buffer, pU 7.6, 
containing ~5 per cent sucrose; embedded in n butyl methacrylate. 

Magnification: 24,000. 

Fig. 2. The micrograph shows at a high magnification a small field in the basal region 
of a pancreatic exocrine cell (rat). 

Parts of two mitochondrial profiles appear at m; the rest of the field is taken by numerous 
profiles of the endoplasmic reticulum (rs), most of which are of elongated form and appear 
disposed in parallel rows. In three dimensions many of these profiles correspond to rela- 
tively large but flat vesicles known as cisternae. 

The membrane limiting the cavities of the reticulum appears as a dense, fine line (n) 
whenever sectioned normally, and as a less dense, poorly outlined band (o) when cut 
obliquely. In a few places the section has opened small windows (/) in the wall of the 

The membrane of the endoplasmic reticulum separates two distinct phases in the 
cytoplasm: one is represented by the light, homogeneous material enclosed in the cavities 
of the system (c); the other, by the surrounding cytoplasmic matrix (mx). 

Numerous small, dense particles, — 150 A in diameter, appear attached to the outer 
surface of the membrane limiting the cavities of the endoplasmic reticulum (ap). In ad- 
dition to these attached particles, particles of comparable size and density occur apparently 
freely scattered in the cytoplasmic matrix (fp). Note, however, that many of these particles 
form short chains (arrows) anchored with one end among the attached particles. 

Fixation: 24 hours at room temperature in 1 per cent Os0 4 in acetate veronal buffer, 
pH 7.6. As a result of the long fixation, part of the cytoplasmic matrix has been extracted 
and the profiles of the endoplasmic reticulum and the small, dense particles appear in 
better contrast. 

Embedding: n butyl methacrylate. 

Magnification: 50,000. 









H : 

I _<> 

Fig. 3. The micrograph shows a small field in the basal region of a pancreatic exocrine 
cell which was damaged (cut open) during the trimming of the tissue block before fixation. 

A comparison with figure 2 indicates that in the damaged cell the profiles of the endo- 
plasmic reticulum have distended cavities and are predominantly circular. In three dimen- 
sions they correspond to spherical and oval vesicles. An enspherulation of this type is 
usually accompanied by a breaking-down of the system into a collection of isolated vesicles. 
Note that the vesicles are still aligned in more or less parallel rows. 

The limiting membrane of the vesicles and its attached particles appear in normal section 
at n, and in oblique section at o. In this case there are few free particles left in the 
cytoplasmic matrix. 

Fixation and embedding as for the specimen in figure 2. 

Magnification: 50,000. 

l ti. 

Fig. 4. Section through a microsome pellet isolated by differential centrifugation from a 
pancreatic brei (guinea pig) prepared in 0.88 M sucrose. 

The microsomes are small, closed vesicles limited by a thin membrane which bears small 
(~150 A), dense particles attached to its outer surface. The apparent heterogeneity of the 
microsomal fraction is due primarily to sectioning. Some vesicles are seen in median sec- 
tion {mv x ) and therefore clearly display their normally sectioned membrane and their 
attached particles. Other vesicles are cut medially (mv z ) and show obliquely sectioned, 
poorly defined membranes. Finally, in lateral sections (mv s ), the cavity of the microsomes 
cannot be seen, and their particle-studded membrane appears in full-face view. 

The structural details described indicate that the microsomes are derived (by fragmenta- 
tion) from the rough-surfaced part of the endoplasmic reticulum. A comparison with 
figure 2 suggests that the fragmentation occurs spontaneously when the cell membrane 
is ruptured during tissue grinding. 

Note that the microsomal content varies widely in density from light (rhv ± ) to medium 
(raz/ 4 ) and high (mi/ 5 ). A ruptured microsomal vesicle (mv G ) contains the equivalent of 
an intracisternal granule. 

Fixation: 2 hours at 0° C in 2 per cent Os0 4 in 30 per cent (0.88 M) sucrose. 

Embedding: n butyl methacrylate. 

Magnification: 72,000. 

Fig. 5. Small field in a microsomal pellet (guinea pig) prepared like die specimen in 
figure 4. 

A few microsomal vesicles appear in median {mv^), medial (mv 2 ), and lateral (mv 3 ) 
section. The particles attached to the outer surface of the limiting membrane (arrows) 
display their characteristic small size ( — 150 A) and high density. 

Fig. 6. Small field in a pellet obtained by differentially centrifuging a microsomal 
suspension (like the one in figure 5) treated with Na deoxycholate (0.1 per cent, pH 7.2). 

The pellet consists of dense particles, — 150 A in diameter, which frequently occur in 
chains (arrows) or in clusters. Their general morphology suggests that they are particles 
detached from the microsomes as a result of the solubilization of the microsomal membrane 
by deoxycholate. 

Both pellets were fixed and embedded like the specimen in figure 4. 
Magnification: 120,000 for both figures. 

Fig. 7. Pellet of a second postmicrosomal fraction obtained by differential centrifugation 
(16 hours at 105,000^) from a pancreatic brei (guinea pig) prepared in 0.88 M sucrose. 

The fraction consists of small (~150 A), dense particles which occur either isolated or 
in chains (arrows) . They are assumed to be the "free" particles of the cytoplasmic matrix. 

Fixation and embedding as for the specimen in figure 4. 

Magnification: 160,000. 



The Influence of Conditions of Culture on 
Certain Soluble Macromolecular Components 

of Escherichia colt 


Department of Biochemistry, University of Leeds, England 

Schachman, Pardee, and Stanier [1] used the analytical ultracentrifuge to 
examine soluble extracts prepared from various bacterial species disrupted by 
different methods, and they showed the presence in all extracts of three major 
components having sedimentation coefficients (uncorrected) of about 40, 29, 
and 5 S. Other workers (Siegel, Singer, and Wildman [2] and Billen and 
Volkin [3]) have obtained ultracentrifuge patterns in substantial agreement. 
For the strain of Escherichia coli we have used, a typical "basic" pattern is seen 
in figure 3a, where the boundaries, reading from left to right, sediment re- 
spectively at 40, 29, and 20 S followed by a large, slow-moving peak which, on 
centrifuging for longer periods, resolves into two peaks of 8 and 5 S. Although 
the ultracentrifuge has revealed a common pattern of macromolecules, how- 
ever, the ever-increasing range of enzymes shown to be induced in bacteria 
supports the assertion that they are "the most plastic of living material" (Ste- 
phenson [4]); and accordingly we have tried to find out whether changes in 
their environment have any influence upon the ultracentrifuge pattern. 

This work may be divided into two parts. First, we examined the basic ultra- 
centrifuge pattern for modifications that might result from growth of the cells 
on different sources of carbon; and we also determined rates of sedimentation 
of certain enzymes, some of which were induced by addition of substrates to 
cultures, in order to decide whether any of them appeared to be associated with 
macromolecules revealed by ultracentrifugal analysis. The second series of 
experiments was concerned with factors affecting the concentration of the 40 S 

component inside the living cells. 




E. coli cells were usually grown without aeration in media containing 0.13 M 
KH2PO4 brought to pH 7 with NaOH; growth in 0.01 M glucose was then 
limited entirely by exhaustion of the source of carbon, and changes in pH 
were negligible. The medium was completed by addition, per liter, of 0.2 g 
MgSOWHoO and either 1 g (NH 4 ) 2 S04 (for a "mineral salts" medium) or 
10 g Difco bactopeptone. Extracts were prepared from cells disintegrated in 
the Hughes [5] bacterial press without abrasive; soluble material was ex- 
tracted by stirring with 0.066 M phosphate buffer, pH 7, and cell debris was re- 
moved by centrifuging for 15 minutes at 102,000^. The protein content of the 
extracts was then determined by the biuret colorimetric method and adjusted 
to 10 mg protein/ml by addition of buffer. This solution could be stored at 
0° C for 24 hours before examination, but freezing and thawing caused altera- 
tions in ultracentrifuge patterns. When glucose was the source of carbon, 
irrespective of whether it was limiting for growth or supplied in excess, or 
whether (NH^SOi or peptone was the nitrogen source, cells in mid-logarith- 
mic phase gave patterns with boundaries that sedimented at about 40, 29, 20, 
and 8 S. 

It was only when cell division had ceased for several hours that modifica- 
tions could be seen, namely, a reduction in the size of the 40 S peak and the 
appearance of a small 13 S peak. By contrast, when cells grown on glucose 
were incubated in a lactose growth medium, a 13 S boundary was observed 
for actively dividing cells. When lactose was utilized as sole carbon source in 
a mineral salts medium there was a lag of 80 minutes before cell division 
began, but the (3-galactosidase activity of the culture was increasing in this 
period and a trace of the 13 S component appeared in that time. When pep- 
tone was the nitrogen source, 3-galactosidase was synthesized much faster and 
the lag preceding cell division was only 20 minutes. 

Ultracentrifuge patterns for extracts from these cells are shown in figure 1, 
where, since photographs were taken at 32 minutes, the 40 S boundaries have 
already sedimented. After incubation with lactose for 100 minutes, peaks of 
29, 20, and 13 S are well defined. Extracts from cells at this stage of growth 
on glucose show only 29 and 20 S boundaries in addition to 8 S after centrifug- 
ing for 32 minutes, the trace of 13 S component in the initial stationary phase 
culture being lost during cell division. When cells were adapted to utilize galac- 
tose and D-xylose as sources of carbon, boundaries that sedimented at about 
13 S also appeared; and they were visible in extracts containing the induced 
enzyme citratase (Dagley and Sykes [6]). 

A correlation may therefore be suggested between the appearance of 13 S 
components and the induction of certain enzymes, including those that may 
be developed by glucose-grown cells as they remain in the stationary phase in 
the presence of accumulated products of metabolism. The significance of such 
a correlation, however, is not evident. It is possible, for example, that 13 S 



Fig. 1. Patterns at 32 minutes and 187,000^ for extracts from E. coli during growth in 
lactose medium. Cells were harvested from: (a) a stationary-phase glucose mineral salts 
culture; (b) lactose medium when in early logarithmic phase, 30 minutes after transfer 
from (a); (c) 100 minutes after transfer; (d) after 160 minutes. Sedimentation is to the 
left, and visible peaks have sedimentation coefficients of 29, 20, 13, and 8 S. In each pattern 
the 40 S boundary has already sedimented. 

components might have some general significance in enzyme induction proc- 
esses, or, alternatively, that the actual enzyme molecules are synthesized in 
amounts sufficient to affect the ultracentrifuge patterns and that, by coincidence, 
the molecules of those we have studied are all of the size (and shape) to sedi- 
ment at about 13 S. 

Data tending to favor the second suggestion were obtained by following the 
sedimentation of enzyme activities in extracts. For these measurements, the 
rotor was allowed to come to rest after a field of 187,000g had been maintained 
for a definite time interval, and the cell was carefully removed. By means of 
a syringe it was possible to withdraw almost the whole of the supernatant 
without disturbing the pellet deposited in the cell. The activity remaining in 
the supernatant was then determined, and the sequence of operations was re- 
peated for a different length of centrifugation so that a graph could be con- 
structed to relate duration of spin to supernatant activity. 

The enzymes assayed were arginine, lysine, and glutamate decarboxylases 
(Gale [7]) for extracts from cells grown in media containing 2 per cent glucose 
and 1 per cent peptone with addition of the corresponding amino acids; cit- 
ratase using the media and methods of Dagley and Sykes |6); and the 3-galac- 
tosidase (Lederberg [8]) of cells grown at the expense of lactose. The consti- 
tutive enzymes of the TCA cycle, malic and isocitric dehydrogenases, fumarase 
and aconitase, present in extracts of cells grown in mineral salts media with 
limiting glucose, were also assayed by the spectrophotometric methods used by 
Englesberg and colleagues [9, 10]. 

In figure 2 it is seen that centrifuging for about 90 minutes removed all the 
citratase, (3-galactosidase, and glutamic decarboxylase of extracts. This is the 
time taken to sediment the 13 S boundary. Lysine and arginine decarboxylases 
sedimented faster, at about the speed of the 20 S boundary. Malic and isocitric 
dehydrogenase activity moved much more slowly, and fumarase and aconitase 
appreciably slower than 13 S. The behavior of the four TCA-cycle enzymes 
is in agreement with the view that they were present in these extracts as indi- 
vidual molecules, since an approximate molecular weight of 40,000 has been 



assigned to malic dehydrogenase (Wolfe and Nielands [11]), of 64,000 to iso- 
citric dehydrogenase (Dixon and Moyle [12]), and 204,000 to fumarase (Mas- 
sey [13]; Cecil and Ogston [14]). If an estimate of 700,000 is taken as the 
molecular weight of (3-galactosidase (Cohn [15]), the enzyme is certainly no 
smaller than the 13 S component and its synthesis during adaptation to lactose 
might well result in an addition to the pattern in this region. If this is also 
true for the other induced enzymes it is surprising that they appear to be syn- 
thesized in such quantity and that the molecules are so large that they sediment 
between 13 and 20 S. 

It is sometimes stated that a number of enzymes are located in the 40 S 
component, but this is not so for the activities we have investigated. In this 
connection, observations on another system studied in these laboratories are 
relevant (Callely, Dagley, and Hodgson [16]). Extracts have been prepared 
from a vibrio which catalyze the oxidation of octanoate and other fatty acids 
to acetate, apparently by the reactions of the fatty acid spiral (Lynen and 
Ochoa [17]). If there are associations of related biochemical activities upon the 
40 S particle analogous to those present in various particles from higher organ- 
isms, they might well be sought here; but in fact the component may be re- 
moved from these extracts with little diminution of the over-all rate of oxidation. 










30 60 

Time of cenfrifuging, minutes 


Fig. 2. Sedimentation of enzymes at 187,000^. Experimental points for: (3-galactosidase, 
solid circle; citratase, square; glutamic decarboxylase, open circle. Arrow a shows time 
when no lysine or arginine decarboxylase remained; and b, c, d, and e show levels of super- 
natant activities after 90 minutes for fumarase, aconitase, isocitric dehydrogenase, and malic 
dehydrogenase, respectively. 



Variations in the concentration of the 40 S component due to cultural changes 
were far greater than for 13 S. Cells grown in peptone contained higher con- 
centrations of the 40 S component than those at a corresponding phase of 
growth in mineral salts media, and for both types of medium the concentration 
increased during cell division and decreased in the stationary phase. Since the 
40 S component carries most of the RNA of E. coli (Schachman et al. [1]), these 
observations agree with Wade and Morgan [18], who found a higher RNA 
content for dividing than for resting cells; but we have never examined extracts 
of cells harvested from stationary-phase cultures that were entirely devoid of 
40, 29, or 20 S components although we adopted the same extraction procedures 
as these authors in several experiments. Cells that had remained 2 hours in a 
stationary culture were not distinguishable from those late in logarithmic 
growth; and a significant reduction in the 40 S component could only be seen 
after 8 hours. 

The most striking effect observed was the disappearance of the 40 S boundary 
when cells were resuspended in phosphate buffer of the same strength (0.13 M) 
as that in the growth medium from which they were harvested. Apparently 
no permanent damage is done to the cells by this treatment because, on addi- 
tion to the buffer of glucose, Mg ++ and (NH4) 2SO4 or peptone, growth occurs 
with little or no lag period. These observations are illustrated in figure 3 with 
patterns of extracts from cells grown to the stationary phase in mineral salts 
medium (a), which were then transferred to peptone medium and examined 
in mid-logarithmic phase (b) and at the end of this phase (c). The augmen- 
tation of 40 S during cell division in peptone is evident; and, in (c), 29 S has 
also increased. Cells that gave pattern (c) were incubated for 30 minutes in 
0.13 M phosphate, and most of the 40 S disappeared to leave two small peaks 
of 43 and 40 S (d) ; on further incubation only traces of material remained 
that sedimented in this region (e). Peptone and glucose were now added to 
the suspended cells; growth resumed, and (/) gives the pattern after 1 hour. 

Some of the conditions for stabilizing the 40 S component within whole cells 
were investigated. Dr. A. Tissieres informed us that its disappearance de- 
pended on the concentration of phosphate in the buffer used to suspend the 
cells, and we have found a considerable diminution of the rate of loss of 40 S 
when the phosphate is reduced from 0.13 M to 0.066 M. It does not appear that 
the effect of orthophosphate is specific, however, since disappearance of the 
40 S component has been shown for whole cells resuspended in potassium 
chloride of the same ionic strength as 0.13 M phosphate. The importance of 
magnesium ions is shown by the results of figure 4. If the growth medium is 
complete except for addition of Mg ++ , 40 S is lost (fig. 4£). On addition of 
Mg ++ , 40 S is stabilized and cell division may begin; but Mg ++ may stabilize this 
component even in cells that are not able to divide because the source of energy 
has been omitted (fig. 4<r). Wade and Morgan [18] have shown an association 






ifiB a I P»l _■ ; I ■ *»*..■ .w,.,^..-.,Jl 




Fig. 3. Patterns after 16 minutes' centrifuging at 187,000g for extracts from cells, in 
stationary-phase glucose mineral salts culture (a); in mid-logarithmic phase after transfer 
to glucose peptone medium (b) ; in late logarithmic phase in this medium (c) ; after incu- 
bation with 0.13 M phosphate buffer, pH 7, (d) for 30 minutes and (<?) for 120 minutes. 
The volume of buffer for resuspension was the same as that of the culture from which the 
cells were taken. On addition of glucose and peptone to the suspension the cells grew, and 
pattern (/) was given after 1 hour. 

of magnesium ions with an ultracentrifuge fraction that contains the 40 S 
component, and, drawing attention to the observation of Webb [19] that bac- 
teria may synthesize protein but may not divide when there is a deficiency of 
magnesium, they make the stimulating suggestion that the RNA in this frac- 
tion is directly concerned with the cell-division process. 

Without further study we cannot conclude that the stability of the 40 S com- 
ponent in whole cells is determined solely by the concentrations of Mg ++ and 
other inorganic ions in the medium. Thus it appears that when cells are grown 
in a rich peptone medium they are less readily depleted of 40 S than when 
grown in mineral salts; and, in our studies of 3-galactosidase development dur- 
ing the lag in cell division in a lactose mineral salts medium, we observed a 
reduction in this component although the magnesium content of the culture 
was normal. It is therefore possible that other factors concerned with the 
metabolism of the cell may also control the stabilization or degradation of the 
40 S macromolecule. 

When a cell-free extract was diluted, or the buffer concentration used in its 



(a) (b) (c) 

Fig. 4. Patterns, after 16 minutes' centrifuging at 187,000^, for extracts from cells grown 
in a glucose mineral salts medium and then resuspended in solutions containing 1 g 
(NH 4 ) 2 S0 4 per liter of 0.13 M phosphate buffer, pH 7, with no further addition (a); 
addition of 0.01 M glucose (b); addition of 0.2 g MgS0 4 '7H 2 per liter (c). Cells were 
incubated for 90 minutes in the same volume of solution as that from which they were 

preparation was increased, the proportion of 40 S relative to the other compo- 
nents was reduced. The patterns shown in figure 5 suggest a split of 40 S 
macromolecules into 29 and 20 S on dilution, favored by high phosphate con- 
centration. Thus, no 40 S boundary was visible in 0.2 M phosphate: the lead- 
ing peaks in the extract which contained 7.5 mg protein/ml sedimented at 
measured (uncorrected) speeds of 27 and 20 S respectively; and from their in- 

1, j H 


M i 




Fig. 5. Patterns, after 16 minutes' centrifuging at 187,000g-, for extracts prepared in 
three strengths of phosphate buffer. Extract concentrations were: 10 mg protein/ml (a); 
7.5 mg protein/ml (b). Sedimentation to the left; in 0.066 M and 0.04 M phosphate, 
boundaries of 40, 29, 20, and 8 S are visible. The three photographs of either scries were 
taken at the same schlieren angle; but an adjustment was made between series to provide 
comparable areas for the 8 S peak. 



creased areas it appears that the components forming these boundaries may 
contain the material that appears as 40 S at lower phosphate concentrations. In 
0.066 M phosphate, the leading boundary sedimented at 40 S followed by 26 S 
and 19 S at 10 mg protein/ml, but there was dissociation of the 40 S component 
at 7.5 mg protein/ml. In 0.04 M phosphate, 40 S was stable at both concentra- 
tions of extract. It is of interest that concentrations of phosphate above 0.06 M, 
but not below, effect a loss of the 40 S components whether they are inside the 
cells or in extracts isolated from them. This supports the evidence of Roberts 
et al. [20] that the phosphate concentration inside E. coli does not differ greatly 
from that outside. The final results of disintegration of the 40 S component, 
however, are not the same in whole cells as in extracts, for 29 and 20 S com- 
ponents appear to be the main disintegration products in extracts, whereas in 
whole cells these peaks are not strongly augmented. It is possible that the ini- 
tial split of 40 S is to give 29 and 20 S components and that in whole cells the 
process goes further to produce lower-molecular-weight diffusible material. 

We are grateful to the Medical Research Council for their financial support 
of this work. 


1. H. K. Schachman, A. B. Pardee, and 
R. Y. Stanier, Arch. Biochem. Biophys., 
38, 245 (1952). 

2. A. Siegel, S. J. Singer, and S. G. Wild- 
man, Arch. Biochem. Biophys., 41, 278 

3. D. Billen and E. Volkin, /. BacterioL, 
67, 191 (1954). 

4. M. Stephenson, Bacterial Metabolism, 
3d ed., p. 311, Longmans, Green and Com- 
pany, London, 1949. 

5. D. E. Hughes, Brit. J. Exptl. Pathol., 
32, 97 (1951). 

6. S. Dagley and J. Sykes, Arch. Bio- 
chem. Biophys., 62, 338 (1956). 

7. E. F. Gale, Advances in Enzymol., 6, 
1 (1946). 

8. J. Lederberg, /. BacterioL, 60, 381 

9. E. Englesberg, J. B. Levy, and A. 
Gibor, /. BacterioL, 68, 178 (1954). 

10. E. Englesberg and J. B. Levy, /. Bac- 
terioL, 69,418 (1955). 

11. R. Wolfe and J. Nielands, /. Biol. 
Chcm., 221, 61 (1956). 

12. M. Dixon and J. Moyle, Biochem. ]., 
63, 548 (1956). 

13. V. Massev, Biochem. J., 51, 490 

14. R. Cecil and A. G. Ogston, Biochem. 
J., 51, 494 (1952). 

15. M. Cohn, BacterioL Revs., 21, 140 

16. A. G. Callely, S. Dagley, and B. 
Hodgson, Biochem. J., 66, 47P (1957). 

17. F. Lynen and S. Ochoa, Biochim. et 
Biophys. Acta, 12, 299 (1953). 

18. H. E. Wade and D. M. Morgan, 
Biochem. /., 65, 321 (1957). 

19. M. Webb, /. Gen. Microbiol., 3, 410 

20. R. B. Roberts, P. H. Abelson, D. B. 
Covvie, E. T. Bolton, and R. J. Britten, 
Studies of Biosynthesis in Escherichia coli, 
Carnegie Inst. Wash. Publ. 607, Washing- 
ton, D. C, 1955. 


Physicochemical and Metabolic Studies on 
Rat Liver Ribonucleoprotein 1 



Sloan-Kettering Institute for Cancer Research 
and Sloan-Kettering Division, Cornell University Medical College 

The presence in uninfected tissues of particles with sedimentation constants 
of about 75 S has long been known to virologists [1]. Similar macromolecules 
have been found in tumors (Kahler and Bryan [2]). We first observed them 
in spleen [3], and later in liver, pancreas, and various tumors [4]. They are 
now known to be ribonucleoproteins (RNP) [5]. Ultracentrifugal patterns of 
these RNPs show a number of boundaries. Their sedimentation coefficients 
are so strongly dependent on concentration that we have denoted them by 
letters, such as B, C, and E, rather than identifying them by their sedimenta- 
tion constants at infinite dilution. B has a sedimentation constant of 78 S, C of 
about 62 S, and E of 46 S, corresponding roughly with the S80, S60, and S40 
boundaries found in nucleoproteins from microorganisms. 

In general, where we find a large amount of nucleoprotein B, as in pancreas 
and normal liver [4], the electron microscopists find most of the granules at- 
tached to endoplasmic reticulum [6] ; but where we find increased amounts of 
C and E, as in liver tumors [4], the electron microscopists find granules not 
attached to reticulum [7]. We therefore wanted to see whether a large micro- 
some fraction, containing only RNP bound to endoplasmic reticulum, con- 
tained only B, while a small microsome or ultramicrosome fraction containing 
free RNP particles was rich in C and E. 

1 The authors wish to acknowledge the assistance of the Atomic Energy Commission 
(Contract no. AT (30-1)— 910), and the National Cancer Institute of the United States 
Public Health Service (grants nos. C-2329 and CY-3190). 

2 Visiting research fellow. 




We prepared such fractions by differential centrifugation at 78,000g in 0.79 
M sucrose. The whole microsome fraction was centrifuged for 5 hours. The 
large-microsome fraction was centrifuged for only 50 minutes, in order to sedi- 
ment only large fragments of endoplasmic reticulum. From this supernatant 
the small-microsome fraction was sedimented for 5 hours. Each fraction was 
washed by resedimentation in 0.79 M sucrose containing 5 X 10~ 4 M K2HPO4 
and KH2PO4, and 5 X 10" 4 M MgCl 2 to preserve the RNP [8], and finally was 
suspended in this same buffer without the sucrose, or in water, to a RNP con- 
centration of about 5 mg/ml. For ultracentrifugal analysis each sample was 
diluted with 0.2 volume of a fivefold concentrated buffer. Two sets of analyses 
carried out in 0.1 M KHCOs are shown in figure 1. When a mere trace of 
magnesium is present, as in the upper row, the RNP from the large micro- 
somes is chiefly B, with only small amounts of C and E. The small microsomes, 
however, do show considerable amounts of C and E. With 0.0024 M magne- 
sium, in the bottom row, all preparations show less C and E, but the same rela- 
tionship is retained — there is more C and E in the small microsomes than in 
the large ones. The whole microsome fraction falls in between, as one would 
expect. Varying the pH or the concentrations of monobasic and dibasic ions 
gives an assortment of patterns, but there is always more C and E in the ex- 
tracts of the small microsomes. 

Whether any of the RNP in the small-microsome fraction is really "free" in 
the liver cell, however, is difficult to determine. In extracts made in water or 
the dilute phosphate-magnesium buffer, at pH 7, no RNP boundaries are seen 
in the ultracentrifuge. They appear when the pH is raised to 8.0 or the ionic 
strength is increased. Since the state of the particles is so dependent on their 
ionic environment, they will have to be studied in buffers which approximate 






♦ i ♦ 

k^/_jLV ^J^J 


pH KHCO3 Phos. MgCI 2 
M M M 

8.0 O.IO 






8.0 O.IO 0.0008 0.0024 

Fig. 1. Ultracentrifugal patterns of RNP extracted from microsomal fractions. The 
pictures were taken after 14 minutes at 37,020 rpm. 



Fig. 2. Electrophoretic patterns of purified RNP in 0.10 M KHC0 3 , pH 8.2, contain- 
ing 0.001 M MgCl 2 . Pictures taken after 60 minutes at 5.2 volts/cm. The left-hand pat- 
terns show RNP that had been washed three times. The right-hand patterns show the 
same material after removal of the ferritin. (See text.) 

the ionic composition of the liver cell before any conclusions can be drawn. 3 
Most of our recent work has been carried out on purified RNP [9]. The 
microsomes are disrupted with deoxycholate, and the RNP is purified by alter- 
nate cycles of high- and low-speed centrifugation. Recently we have made some 
improvements in the procedure. The livers do not have to be perfused. The 
addition of penicillin (100 units/ml) to all the solutions increases the stability. 4 
Instead of washing with calf-liver dialysate we now use the potassium phos- 
phate-magnesium chloride buffer; its pW is about 7. The washed RNP still 
shows about 5 per cent of ferritin on electrophoretic analysis (fig. 2). This can 
be removed by precipitating the RNP with 0.005 M barium acetate and redis- 
solving it by dialysis against 5 X 10" 4 M K2HPO4, KH 2 P0 4 , and MgS0 4 (fig. 2) . 
We have carried out extensive studies of the stability of purified RNP. Like 
the crude microsomal extracts, the purified RNP is very sensitive to pH, ionic 
strength, and dibasic and monobasic cations. Figure 3 shows the effect of mag- 
nesium in the presence of 0.1 M NaHCOs. Fresh RNP looks like the third 

3 Recently we have prepared large- and small-microsome fractions by sedimentation in 
0.20 M sucrose, and resuspended them in water at pH 7.1 without washing. On ultracen- 
trifugal analysis sizable C and E boundaries were observed in the small-microsome frac- 
tion, although none were apparent in the large-microsome fraction. 

4 The purification of the RNP requires about 3 days; although the preparation is kept 
cold, unless penicillin or sterile technique is used the RNP has an odor like that of spoiled 
meat. This may be due to cadaverine produced from N-terminal lysine by bacterial lysine 
decarboxylase; after treatment of the RNP with dinitrofluorobenzene and acid hydrolysis, 
the only DNP amino acid detectable is lysine. 


picture down, chiefly B. After dialysis overnight against bicarbonate contain- 
ing 0.005 M MgCl 2 the RNP is unchanged. With less magnesium, as shown 
in the two upper pictures, it dissociates; with more magnesium, as shown in 
the bottom row, the B boundary gets smaller, because the RNP has begun to 
aggregate and precipitate out. 

In the purified RNP we found 40 per cent RNA on a dry-weight basis [9]. 
Similar values have been found for purified RNPs from yeast [10] and pea 
seedlings [11], again on a dry-weight basis. When RNA is compared with 
protein nitrogen, on the other hand, we and many other workers find values 
for RNA of 50 to 60 per cent. Some of this discrepancy is probably caused by 
lipid, which contributes more to the dry weight than to the nitrogen. Some of 
it, however, seems to occur because hot 5 per cent trichloroacetic acid extracts 
some other nitrogen besides that in the RNA. With 10, 15, and 20 per cent 
trichloroacetic acid we still find extra nitrogen in the extracts. 

A B E S P H NaHC0 3 Phos. MgCI 2 

+ + + + M MM' 


82 O.IO O.OOI 

8.2 O.IO O.OOI 0.0005 

8.2 0.10 0.001 0.0050 

8.2 0.10 0.001 0.0100 

Fig. 3. The effect of magnesium on the stability of purified RNP. Ultracentrifugal 
patterns were obtained after 14 minutes at 37,020 rpm. 

We have carried out a number of metabolic studies on purified RNP con- 
taining about 4 per cent ferritin. Figure 4 shows the incorporation of glycine- 
1-C 14 into adenine and peptide glycine. The upper row shows the time course 
of the incorporation in normal male rats. Fifteen minutes after the injection 
of the isotope, by tail vein, the specific activity of the acid-soluble adenine of 
the liver was quite high, whereas that of the adenine in the purified RNP was 
still very low, and, although it increased very slowly with time, it was still low 
at 17 hours. The specific activity of the liver nucleoprotein glycine was highest 


















I J - W 















V 4 i 


400 o 



2000 z> 

1600 cr 










The incorporation of glycine-1-C 14 into the adenine and glycine of rat liver RNP, 
and into "supernatant" and serum proteins (see text). 

Fig. 4. 

at 15 minutes and then dropped off. The glycine of the liver supernatant pro- 
teins, plus the microsomal lipoproteins soluble in deoxycholate, reached a maxi- 
mum between 1 and 2 hours, and then fell. The glycine of the serum proteins 
increased very slowly, and was highest at 17 hours. Similar results have been 
obtained with labeled adenine and methionine as tracers. 

The lower row shows the results obtained on rats pretreated with bovine 
growth hormone, 410 [\g per rat per day, for 2 weeks. There was no effect on 
the RNP adenine, but the effects on the three protein fractions were marked. 
Everything seemed to have been speeded up. The nucleoprotein glycine ac- 
tivity dropped more rapidly; the liver supernatant activity rose more steeply; 
and the serum proteins apparently reached a maximum at some time before 17 
hours, and then declined again. 

These results illustrate, first, the lack of correspondence in the extents of 
isotope incorporation into the RNA and the total protein glycine of the RNP, 
and, second, the effect of growth hormone, which speeded up the incorporation 
of glycine into protein without any change in the rate at which the particle 
RNA became labeled. From these observations we conclude that, at least in 
rat liver, the total RNA of the RNP cannot be functioning as an active template. 




1. A. R. Taylor, D. G. Sharp, and B. 
Woodhall, Science, 97, 226 (1943). 

2. H. Kahler and W. R. Bryan, /. Natl. 
Cancer Inst., 4, 37 (1943-1944). 

3. M. L. Petermann and M. G. Hamil- 
ton, Cancer Research, 12, 373 (1952). 

4. M. L. Petermann, N. A. Mizen, and 
M. G. Hamilton, Cancer Research, 16, 620 

5. M. L. Petermann, M. G. Hamilton, 
and N. A. Mizen, Cancer Research, 14, 
360 (1954). 

6. G. E. Palade, /. Biophys. Biochem. 
Cytol, 1, 59 (1955). 

7. A. F. Howatson and A. W. Ham, 
Cancer Research, 15, 62 (1955). 

8. F.-C. Chao, Arch. Biochem. Biophys., 
70, 426 (1957). 

9. M. L. Petermann and M. G. Hamil- 
ton, /. Biol. Chem., 224, 725 (1957). 

10. F.-C. Chao and H. K. Schachman, 
Arch. Biochem. Biophys., 61, 220 (1956). 

11. P. O. P. T'so, J. Bonner, and J. Vino- 
grad, /. Biophys. Biochem. Cytol., 2, 451 


Ultracentrifugal Studies of Microsomes from 

Starving, Nonproliferating, and 

Proliferating Yeast 


Dontier Laboratory of Biophysics and Medical Physics 
University of California, Berkeley 

In the decade since Claude's successful isolation of microsomes from liver 
and other tissue homogenates by differential centrifugation [1, 2], considerable 
progress has been made toward elucidating their biochemical and morphological 
characteristics. The microsomal "ribonucleoprotein" particles isolated from bac- 
terial cells and from plant and animal tissues are rich in ribonucleic acid [3-10]. 
In electron-microscopic observations these microsomal particles, either isolated or 
in intact plant and animal cells, appear as spherical particles with diameters 
ranging from 100 to 400 A. They occur either bound to the endoplasmic reticu- 
lum or freely dispersed in the cytoplasmic matrix [7-14]. 

There are interesting studies indicating that microsomes may be actively in- 
volved in protein and lipid synthesis [5, 15, 16]. In vivo [17-20] and in vitro 
studies [15, 20, 21] have shown that labeled amino acids are preferentially in- 
corporated into the microsomal proteins, suggesting synthetic activity. 

Since these microsome particles appear to be functional organelles in both 
plant and animal cells, it is highly probable that their physicochemical proper- 
ties will be altered by varying the physiological state of the cell. Preliminary 
studies of microsomes isolated from starving, nonproliferating, and proliferat- 
ing yeast cells have shown that several new ribonucleoprotein particles appear 
during cell division [22, 23]. This paper presents further evidence thereon. 


Haploid yeast cells (Saccharomyces cerevisiae strain S. C. 7) aerobically cul- 
tured for 24 hours at 30° C in yeast extract and dextrose ( YED 1 : 2 per cent) 




were harvested and washed twice with sterile distilled water. Aliquots of cells 
were then suspended in nitrogen-deficient medium (4 per cent dextrose : M/GO 
KH 2 PO4:3/60 M NaHsPOi) and aerated for 48 to 11 hours at 30° C. During 
this nitrogen starvation, the buffered dextrose medium in some of the cultures 
was changed every 12 hours. At the end of starvation, the cells were centrifuged 
and rewashed twice with sterile distilled water. Approximately 6 gram aliquots 
of wet cells were suspended in 2-liter flasks containing 1500 ml YED (1:2 per 
cent) and induced to grow aerobically at 30° for varying time intervals. The 
growth curve of the cells and the corresponding bud counts for the first 26 
hours are shown in figure 1. The yeast cells were then harvested at different 
times, and microsomal particles were isolated according to Wolfe [23]. Cells 
hand-ground with 100-grid Carborundum in mortar and pestle at 0° C were 
extracted with several volumes of solvent (0.00125 M KH2PO4-K2HPO4 3:7, 







"i r 



8 12 16 20 24 28 

Fig. 1. Relationship of budding to growth of yeast cells, (a) Growth curve of 48-hour 
nitrogen-starved yeast cells aerobically cultured at 30° C in YED (1:2 per cent), (b) Curve 
correlating percentage of visible buds to corresponding growth stages. 



0.001 M MgCl 2 , and 0.01 M KC1). After vacuum nitration through celite filter- 
aid, the microsomes were sedimented from the filtrate by ultracentrifugation in 
the preparative ultracentrifuge (Spinco Model L) at 114,000g (40,000 rpm) for 
1 hour. The translucent microsomal pellets isolated from yeast cells in different 
growth stages were carefully redissolved in the buffered solvent and analyzed 
in the analytical ultracentrifuge (Spinco Model E) at 200,000^. Figure 2 shows 




Fig. 2. Sedimentation photographs of microsome particles from yeast-cell extracts 
6 minutes after up to speed (UTS) at 200,000,? in the analytical ultracentrifuge (Spinco 
Model E). (a) Top: Cell extract from 48-hour nitrogen-starved cells corresponding to the 
inoculum cells at time 0. Bottom: Extract from nitrogen-starved cells 1.5 hours after cells 
given utilizable nitrogen, (b) Top: After 3 hours. Bottom: After 5 hours, (c) Top: 
10 X concentrated extract from nitrogen-starved cells 7 hours after cells given utilizable 
nitrogen. Bottom: Extract from nitrogen-starved cells 10 hours after cells given utilizable 
nitrogen, (el) Top: After 26 hours. Bottom: After 52 hours. 



the changes in the sedimentation pattern which indicate the appearance of new 
microsomal particles [22, 23]. During the lag phase when less than 3 per cent 
of the cells have visible buds, only the 80 S component (a peak in fig. 3) is pres- 
ent in the ultracentrifuge pattern of cytoplasmic extract. As the cells begin to 
proliferate and enter the log phase of growth (40 to 50 per cent visible buds), 
four new microsomal components (3, y, 5, and £ peaks in fig. 3) appear in the 
cell extracts. From log growth phase through stationary phase, the only change 
in the sedimentation pattern appears to be in the ratio of these components. 
Cell starvation is first evidenced by a decrease in the stability of the micro- 
somes, which is indicated by the presence of polymerized material that sedi- 







Fig. 3. Sedimentation photographs of microsomes isolated from 48-hour stationary-phase 
cells 6 minutes after UTS at 200,0(%. (a) Yeast extract suspended in phosphate buffer 
(0.00125 M KH 2 P0 4 -K 2 HP0 4 3:7, 0.001 M MgCU, and 0.01 M KCl) at 4° C for 1 day. (b) 
Yeast extract suspended" in phosphate buffer at 4° C for 2 days, (c) After 7 days, (d) 
After 10 days. (<?) After 13 days. (/) After 27 days. 





Fig. 4. Sedimentation photographs of isolated 80 S particles taken at 0, 2, and 4 minutes 
after UTS at 200,0C% in the analytical ultracentrifuge. Particles moving from left to 
right, (a) 80 S particles from 24-hour stationary-phase cells observed immediately after 
isolation, (b) 80 S particles, isolated from 24-hour stationary cells, suspended in phosphate 
buffer and kept at room temperature for 2 days, (c) 80 S particles isolated from cells 
nitrogen-starved for 72 hours suspended in phosphate buffer and kept at room temperature 
for 2 days. 


ments as a rapidly spreading fore peak [23] before the centrifuge is up to speed 
(UTS). The growing cell components disappear as starvation progresses. 

The stability of microsome components isolated from log phase, stationary 
phase, and starving cells and kept at 4° C or at room temperature in buffered 
solvent was also studied. There was no detectable difference in stability between 
particles isolated from log phase and those isolated from stationary phase. 
Figure 3 shows the degradation of microsomes isolated from 48-hour stationary- 
phase cells and suspended in phosphate buffer at 4° C. 

Since only the 80 S component is present in old stationary-phase cells and in 
starving cells, the stability of this component isolated from proliferating, non- 
proliferating, and starving cells, and dissolved in buffered solvent, was com- 
pared. Figure ^a shows the sedimentation pattern at 0, 2, and 4 minutes after 
UTS at 200,00% of the 80 S particles isolated from 24-hour stationary-phase 
cells. The sedimentation pattern of this component from 72-hour nitrogen- 
starved cells is similar, except for the broad, rapidly spreading, and sedimenting 
fore peak. As shown in figure %, the 80 S component isolated from log-phase 
and stationary-phase cells gave rise to the 60 S and 40 S component [24] when 
kept at room temperature for 2 days. On the other hand, the 80 S component 
isolated from cells nitrogen-starved for 72 hours became degraded by forming 
rapidly sedimenting aggregates. See figures 4c and 5. 


When starved cells are given utilizable nitrogen, the growth curve shows a 
characteristic lag phase corresponding to the degree of starvation. During this 
phase the 80 S component appears to be degraded and reconstituted. As shown 
in figure 2, a decrease in the 80 S component is noted 1.5 hours after starved 
cells are given utilizable nitrogen. Two hours after inoculation, the microsomal 
concentration appears approximately equal to that of the inoculum cells. At 3 
hours, new microsomal components appear in the cytoplasmic extract. The cells 
are now entering the log growth phase, during which the microsome concen- 
tration reaches a maximum. This condition is followed by a gradual quantita- 
tive change in the microsomal components as the cells pass through the sta- 
tionary phase. 

Chao and Schachman [8, 24] have shown that, in vitro, altering the ionic 
environment of the solvent will dissociate or aggregate the 80 S component. It 
would therefore be interesting to ascertain whether a similar mechanism is re- 
sponsible for changes in the microsomal components in respiring cells. 

Since the 80 S particles from only starving cells of low viability are degraded 
by forming rapidly sedimenting aggregates (figures 4c and 5), there appears 
to be a correlation between call viability and the chemical state of this particle. 

Furthermore, if microsomes are involved in the synthesis of proteins and lipids 
[5, 15, 16], the changes observed in the microsomal components with cell growth 
could be a mechanism that controls their synthetic activities. 






Fig. 5. Sedimentation photographs at UTS (200,0(%) of 80 S particles isolated from 
cells nitrogen-starved for 72 hours and suspended in phosphate buffer at 4° C. (a) After 
1 day. (b) After 2 days, (c) After 7 days, (d) After 11 days. 


Ultracentrifugal studies of yeast-cell extracts have shown a correlation be- 
tween the physiological states of the cells and the stability and appearance or 
disappearance of microsomal particles from the cytoplasm. 


This work was done in the Biophysics group of Donner Laboratory and was 
supported by contract with the Atomic Energy Commission. I wish to thank 
Dr. Cornelius A. Tobias for his interest in this work. 




1. A. Claude, Science, 97, 451 (1943). 

2. A. Claude, Harvey Lectures, 48, 121 

3. C. P. Barnum and R. A. Huseby, 
Arch. Biochem., 19, 17 (1948). 

4. G. E. Palade and P. Siekevitz, Fed- 
eration Proc, 14, 262 (1955). 

5. J. W. Littlefield, E. B. Keller, J. Gross, 
and P. C. Zamecnik, /. Biol. Chew., 217 , 
111 (1955). 

6. H. K. Schachman, A. B. Pardee, and 
R. Y. Stanier, Arch. Biochem. Biophys., 38, 
245 (1952). 

7. P. O. P. Ts'o, J. Bonner, and J. Vino- 
grad, /. Biophys. Biochem. Cytol., 2, 451 

8. F. Chao and H. K. Schachman, Arch. 
Biochem. Biophys., 61, 220 (1956). 

9. G. E. Palade and P. Siekevitz, /. Bio- 
phys. Biochem. Cytol. , 2, 171 (1956). 

10. G. E. Palade and P. Siekevitz, /. Bio- 
phys. Biochem. Cytol, 2, 671 (1956). 

11. G. E. Palade, /. Biophys. Biochem. 
Cytol, 1, 59 (1955). 

12. G. E. Palade and K. R. Porter, /. Ex- 
ptl.Med., 100, 641 (1954). 

13. K. R. Porter, /. Exptl. Med., 97, 727 

14. D. B. Slautterback, Exptl Cell Re- 
search, 5, 173 (1953). 

15. P. Siekevitz, /. Biol. Chem., 195, 
549 (1952). 

16. H. P. Klein, /. Bacteriol, 73, 530 

17. T. Hultin, Exptl Cell Research, 1, 
376, 599 (1950). 

18. N. D. Lee, J. T. Anderson, R. Miller, 
and R. H. Williams, /. Biol. Chem., 192, 
733 (1951). 

19. R. M. S. Smellie, W. M. Mclndoe, 
and J. N. Davidson, Biochim. et Biophys. 
Acta, 11, 559 (1953). 

20. V. Allfrey, M. N. Daly, and A. E. 
Mirsky, /. Gen. Physiol, 37, 157 (1953). 

21. P. C. Zamecnik and E. B. Keller, 
/. Biol Chem., 209, 337 (1954). 

22. R. G. Wolfe, UCRL-2553, April 

23. R. G. Wolfe, Arch. Biochem. Bio- 
phys., 63, 100 (1956). 

24. F. Chao, Arch. Biochem. Biophys., 
70, 426 (1957). 


Fractionation of Escherichia coli for 
Kinetic Studies 


Department of Terrestrial Magnetism 
Carnegie Institution of Washington 

A single cell of the bacterium Escherichia coli contains roughly 10,000 ribo- 
somes (microsomal particles). If the cells are broken open and their contents 
are examined, the analytical centrifuge shows a series of peaks with sedimenta- 
tion constants of roughly 20, 30, 40, 60, and 80 S [1-8]. The existence of these 
particles in such variety and in such large numbers immediately provokes a 
number of questions. Is the rapid growth rate of E. coli a consequence of the 
high proportion (25 per cent) of the cellular material that is organized into ribo- 
somes? In other words, are ribosomes the sites of protein synthesis in E. coli? 
If so, what size of particle is active in protein synthesis ? How are the particles 
themselves synthesized? Do the different sizes of particles represent different 
stages in the growth of a particle ? 

Eventually answers will be found for these questions, but not easily. It will 
be necessary to know the composition of the particles— the composition of the 
individual classes of particles, not just the composition of a pellet containing an 
assortment of particles plus other material. Also the kinetics of isotope incor- 
poration will have to be studied. It will not be sufficient simply to deal with 
the microsomal fraction, the 100,000g- pellet; rather, the individual groups of 
particles will have to be sorted out and measured. Suppose, for example, that 
one size of particle is the precursor of another. Kinetic measurements will show 
this clearly if the two groups can be resolved; kinetic measurements of both 
groups lumped together in a pellet will show nothing. 

Chemical fractionation of E. coli gives good separation between the different 
classes of compounds, and it is easy to show by kinetic measurements of the 
incorporation of radioactive compounds that the small molecules serve as 




precursors of the large ones. In contrast, the simple separation into cell wall, 
microsome, and soluble fractions is not sufficient to reveal clearly any precursors 
or products among the macromolecules. A further fractionation of the micro- 
some pellet is required. 

Pellets of somewhat greater homogeneity can be obtained by choosing an 
appropriate centrifuging schedule. The material that sediments in 15 minutes at 
100,000g- is richer in the large particles than the pellet obtained by centrifuging 
down (2 hours at 100,000^) material which stayed in suspension during three 
successive 15-minute periods at 100,000^. The composition of the pellet also 
varies; the early pellet contains nearly twice the lipid and protein per unit of 
nucleic acid. This approach, however, shows no promise of giving adequate 

A somewhat better fractionation can be obtained by using the swinging 
bucket head for the Spinco Model L centrifuge. Microsome pellets are resus- 
pended and layered on top of a sucrose gradient. After a period of centrifuga- 
tion, layers are taken off with a pipet. This technique is adequate to show 
marked differences in the distributions, depending on the initial material. Fig- 
ure 1 shows one curve for a resuspended pellet composed mostly of large (80 S) 
particles; another for the smaller particles (20 to 40 S) that result if magnesium 
is lacking [8]; and a third, for the nonsedimenting material. The analytical 

7 9 

Fraction no 

Fig. 1. Fractionation of particle preparations using the swinging bucket centrifuge. 
Five-tenths milliliter of suspension is placed on top of 4.5 ml sucrose gradient in the cen- 
trifuge tube. After 45 minutes at 100,000^, 0.3-ml fractions are taken off from the top with 
a pipet. 



centrifuge shows that the bottom layers are rich in the heavy particles and lack 
the light particles, whereas the top layers show the opposite distribution. 

Quite a different type of fractionation results from chromatography on col- 
umns of diethylaminoethyl cellulose (DEAE) [9, 10]. Extremely high reso- 
lution can be achieved giving a separation of various proteins as shown in 
figure 2. Nucleoprotein appears as a prominent peak in the elution diagram 
of the total cell juice but not in the diagram obtained with the 100,000^ super- 
natant fluid (fig. 3). The corresponding ultraviolet diagrams show that there 
are in fact two nucleoprotein peaks: the first peak consists of nucleoprotein of 
high molecular weight which can be spun down in the centrifuge; the second 

40 60 80 100 

Fraction number 

Fig. 2. Cell suspension washed and broken as described by Bolton et al. [8] ; 0.5 g wet 
weight of cell juice adsorbed on DEAE column (1 cm 2 X 20 cm) and eluted with con- 
centration gradient to 0.7 M of NaCl in tris-succinate buffer plus magnesium. Lower 
curve, total protein indicated by Folin reaction; upper curve, assay for activity of three 
different enzymes. One-milliliter samples collected in fraction collector. 



^Totol C«M eitroct 


\ .,100,000 g SN 


Eluting fluid, ml 


Fig. 3. Elution patterns of total cell juice and supernatant fluid of 100,000^ 2-hour spin. 
Upper curve, optical density at 254, indicating nucleic acid concentration; lower curve, 
S 35 radioactivity, indicating protein. Note nucleoprotein peak which is missing in 100,00% 


is partly nucleoprotein and partly due to free DNA and RNA which still re- 
main in the 100,000^ supernatant fluid. 

The elution pattern is not sensitive to the size of the particles. The same pat- 
tern is obtained whether the microsome pellet is composed mostly of the large 
(80 S) particles or of the smaller 20 to 40 S particles that result from magne- 
sium deficiency. Compare figures 4# and b. 

Microsome pellets when resuspended and analyzed on the column show the 
nucleoprotein peak together with a quantity of other protein which depends on 
the method of preparation (fig. 4). A part of the contamination of the micro- 
some pellet is due to small bits of cell wall, and another part is due to nonpar- 
ticulate protein. In 2 hours at 100,000^ roughly 70 per cent of the 3-galac- 
tosidase is sedimented. See also Dagley and Sykes [5, 11]. Accordingly the 
least-contaminated preparations of ribosomes are those obtained by resuspend- 
ing a microsome pellet and centrifuging again in the swinging bucket head 
(fig. 4c). 

Unfortunately the column cannot be used to prepare purified ribosomes be- 
cause the material eluted from the column is quite different from that origi- 
nally adsorbed. When the fractions containing the nucleoprotein peak are 



2 I- 



O I 


TS Microsome pellet 

L_ i ^r .^r- -. aa-g t =i 










pellet j 

- i (b) 




_. / / 
• / 






1 1 1 

i .i/i 

1 l l T^f^^O,-^-! 

TSM Microsome pellet 

3 5 7 9 II 13 15 17 19 21 23 25 27 29 31 33 35 
Column fraction no. 

Fig. 4. Elution patterns of microsome pellets, a, 100,0C% 2-hour pellet without mag- 
nesium present; b, same with magnesium present; c, microsome pellet resuspended and 
fractionated with swinging bucket centrifuge. 



centrifuged (100,000^, 2 hours), a colorless glassy pellet is formed which con- 
tains approximately 65 per cent of the protein and nucleic acid. This pellet 
resuspends easily and completely. The analytical centrifuge shows that it con- 
tains peaks in the 20 to 40 S region, whereas the 80 S peak was most promi- 
nent in the original material. The ratio of nucleic acid to protein in this pellet 
(measured by optical density at 260 mu and S 35 ) is twice that of the starting 
material, and the elution pattern obtained when the pellet is rerun on a DEAE 
column is very different (fig. 5). 

These changes appear to be caused by the column material and not by the 
salt of the eluting fluid. Ribosomes exposed to molar NaCl show a reduction 
in size but no change in composition or elution pattern. 

The fractionation and analysis procedures outlined above are beginning to 
yield some useful information about the composition and function of the ribo- 
somes. The purified ribosomes are markedly different from the microsome 
pellet. For example, the microsome pellet contains considerable phospholipid 
(table 1); the ribosomes, little if any. Moreover, the nucleic acid to protein ratio 
is somewhat variable in the crude microsome pellet, but the purified ribosomes 




"l \ 


— T \ 1 \ 



/ \ i .\- 


"V 1 \ 


T \»J \ 


i , -\- 


— N-P Peak 

-,' ; \ 


1 , ^B 


* 1 

1 ' , ,J T \ 

» •' 

' "*Sa ■' \ -V 


' * *"\ 



J "\^ 


' »\- 


! "*~ ^Sw*" 


r n ^"^^T 




-r-f—f-i i-h-4— +-=m--i i "I i — i— -i — 

16 20 24 

Column fraction no. 

28 32 36 40 

Fig. 5. 
change to 

Nucleoprotein peak of elution pattern spun down and rechromatographed. Note 
elution pattern like that of nucleic acid. 


TABLE 1. Chemical Fractionation of coli Components 

Cell Wall 



Whole Cell 

30,000^ pellet 

100,0(% pellet 

100,00% SN 

Small molecules 





















Total 100 20 30 50 

obtained from the swinging bucket give a constant ratio indicating two amino 
acids per nucleotide (NA/P = 60/40 measured by absorption at 260 mu and 
Folin [12] test for protein). 

The protein of ribosomes differs from other proteins of the cell. Purified 
ribosomes were obtained from cells grown with C 14 glucose as the sole carbon 
source. The protein, after hydrolysis and chromatography, showed an amino 
acid distribution in which glutamic acid, alanine, glycine, and lysine were pro- 
portionately higher than in the whole cell, whereas methionine and aspartic acid 
were lower. In this protein neither cysteine nor cystine seems to be present. 

The absence of cystine can best be shown by growing the cells in the presence 
of S 35 Oi to label cystine and methionine. After hydrolysis and chromatography 
the radioactivity of methionine and cystine can be measured. In the protein of 
the whole cell there is approximately twice as much methionine as cystine [13]. 
In ribosomes purified in the swinging bucket centrifuge this ratio is 10:1. In 
nucleoprotein eluted from the column and sedimented the ratio is greater than 

Alternatively the lack of cystine can be demonstrated without hydrolysis and 
chromatography. Cells containing S 35 cystine and S 32 methionine were grown 
by adding S 35 Oi and S 32 methionine to the medium. To prevent even a slight 
leakage of S 35 into methionine, a methionine-requiring mutant was used [13]. 
The sulfur radioactivity per unit protein of the nucleoprotein (obtained by col- 
umn analysis of a microsome pellet and sedimentation of the nucleoprotein frac- 
tion of the eluate) was 50 times lower than that of the whole cell. Since the 
usual occurrence of cystine is only 1 per 60 residues, its occurrence in the nucleo- 
protein is less than 1 per 3000. 

Kinetic studies of the fractions obtained from the column are also in progress. 
S 35 has been used to follow incorporation into protein. Exponentially growing 
cells were exposed to the tracer for varying periods of time and then broken 
and their constituents analyzed. The specific radioactivity of the protein frac- 
tions was measured by TCA-precipitable S 35 and Folin reaction color. When the 
cells are exposed to the tracer for a prolonged period (steady state) the specific 
radioactivity varies throughout the elution pattern by a factor of roughly 3, 
being lowest in the nucleoprotein fraction. These variations are simply due to 
variations in the sulfur content. Other cells were grown for three generations in 



a nonradioactive medium after exposure to the tracer. In this treatment any 
intermediates which have a rapid turnover should lose their radioactivity. The 
resulting "persistent pattern" was entirely similar to the "steady-state pattern," 
and no protein components could be identified as intermediates. 

Finally, cells were exposed to the tracer for short periods. After a 4-minute 
exposure the resulting "pulse pattern" was similar to the "steady-state pattern" 
except that the radioactivity of the nucleoprotein peak was only one-half of that 
expected from the "steady-state" pattern. A similar result was obtained with 
cells exposed for 4 minutes to a mixture of C 14 -labeled amino acids. 

Similar experiments carried out with P 32 Oi give much more striking results. 
Figure 6 shows the macromolecular region of the elution patterns obtained with 
cells exposed to the tracer for increasing periods of time. The radioactivity ap- 
pears first in a quite distinct fraction of the elution pattern, passing through at 
a later time to the other regions. In the steady-state and persistent patterns the 
phosphorus radioactivity was proportional to the optical density (at 260 mu). 
Thus the DEAE column is capable of resolving the nucleic acid and nucleo- 
protein into fractions that seem to be precursors and products. Similar kinetic 
differences were also observed by Creaser, who used ECTEOLA columns [9] 
to analyze alcohol-extracted nucleic acid [14]. 

The analysis of these data runs into a number of complications. The leading 
peak is composed solely of RNP, but the secondary peak is an unresolved mix- 
ture of RNP, RNA, and DNA. Furthermore, the pool of low-molecular-weight 
precursors to RNA is large and may or may not be in equilibrium with the 
smaller pool of DNA precursors [14]. 

A rough analysis can be made on the basis of several simplifying assump- 
tions. Assume first that the low-molecular-weight precursors of the macromole- 


I 1/2 mm 


24 min 


Fraction number 

Fig. 6. Elution patterns of cell extracts after growing cells were exposed to P 32 4 for 
times indicated. Only a small region of the elution pattern is shown. 



cules have the average specific radioactivity of the TCA-soluble pool. Second, 
since the persistent and steady-state runs show all the macromolecular ultra- 
violet-absorbing material to be uniformly labeled, assume that it is the end 
product and that its specific radioactivity is that of the nucleoprotein. Analyzed 
on this basis the data point to an intermediate containing roughly 10 per cent 
of the nucleic acid. 

Some other characteristics of this intermediate have been determined. Very 
short exposures to the tracer were used to prepare cell juices which were shown 
by column analysis to have most of the P 32 in the intermediate and little in the 
end products. Most of the low-molecular-weight materials were removed by 
washing the cells with water before breaking. This material was analyzed in 
the swinging bucket centrifuge. The results (fig. 7) show that the TCA-pre- 
cipitable radioactivity sediments at less than half the rate of the ultraviolet- 
absorbing material, which is mostly in 80 S ribosomes. Incubation with 
RNAase showed the usual rate of release of nucleotides. 










1 1 1 

• P* c/s 









o Optical density X = 260 — 




o\i\« » 


\ • 








1 1 1 


Fraction no. 

Fig. 7. Growing cells were exposed to P 32 for 3 minutes, then broken, and the micro- 
some pellet was analyzed in the swinging bucket centrifuge. Note that the maximum 
radioactivity does not correspond to the maximum of the ultraviolet absorption. Cf. figure 1. 

Those findings, together with its column elution pattern, suggest that the 
intermediate is RNA of high molecular weight, either free or associated with 
less protein than the bulk of the nucleoprotein. It should be emphasized that 
neither lipids nor fragments of cell wall or cell membrane are eluted from the 
column, and it is observed that a large part of the P 32 incorporated in short ex- 
posures is irreversibly bound to the column. An important part of the kinetics 
may thereby be missed. 


To interpret the detailed workings of the cell in terms of its structural com- 
ponents, fractionation procedures are needed to separate those components. The 


procedures outlined above are only a step toward the needed resolution, but they 
have already given indications that: 

1. The 80 S ribosomes are composed of nucleoprotein of approximately two 
amino acids per nucleotide. 

2. This composition is unaltered in the smaller disintegration products re- 
sulting from magnesium deficiency. 

3. Adsorption followed by elution from the DEAE column causes disintegra- 
tion into smaller particles of different composition containing approximately 
one amino acid per nucleotide. 

4. The protein of the ribosomes is a special protein or at least a special class 
of proteins lacking cystine and cysteine. It is therefore doubtful that any of the 
enzymes that have been reported in the microsome pellet are actually in the 
ribosome fraction. 

5. The protein of the ribosomes is most certainly not precursor to the non- 
particulate proteins. Such a relationship is ruled out by the data on the com- 
position and on the kinetics of formation. 

6. Incorporation of amino acids, sulfur, and phosphorus into nucleoprotein of 
the ribosomes shows a kinetic delay which indicates that the ribosomes have a 
macromolecular precursor. 

7. This precursor has properties suggestive of nucleic acid or nucleoprotein 
of a low protein content. 

These findings when checked and verified will be useful in providing further 
conditions that must be met by any theory of protein synthesis. The low initial 
specific radioactivity found in the ribosome fraction differs markedly from the 
high initial specific radioactivity found in the deoxycholate-insoluble part of the 
microsome fraction of rat liver [15]. A partial explanation for the difference 
may be that the nucleoprotein has been stripped clean of adhering newly formed 
protein by the column; it is not a complete explanation, however, because the 
nonparticulate protein of the microsome fraction did not have a high initial 
specific radioactivity. More likely, the difference arises from the difference in 
the growth rates. If the ribosomes furnish the templates for protein synthesis, 
and if chains of 150 amino acid residues are produced by the ribosomes, then 
each of the 10,000 ribosomes of a coli cell must produce one polypeptide chain 
per 10 seconds to give the observed rate of protein synthesis. If one polypeptide 
chain adheres to each ribosome, after 4 minutes' exposure to the tracer only 1/24 
of the newly formed polypeptide chains would be found still adhering to the 
particles. To show kinetic effects in protein synthesis with these rapidly grow- 
ing organisms it will be necessary to use much shorter exposures to the tracer. 

The synthesis of the particles themselves appears to be a distinctly different 
process, as it proceeds at a more leisurely rate. Even after 24 minutes there are 
still marked departures from the steady-state distribution. These findings are 
compatible with, but certainly do not prove, the idea that the smaller particles 
observed in the cell juice are not simply bits broken off during disruption of 



the cell, but that they have biological significance and that they may represent 
stages in the growth cycle of the particles. 


1. S. E. Luria, M. Delbruck, and T. 
Anderson, /. BacterioL, 46, 57 (1943). 

2. H. K. Schachman, A. B. Pardee, and 
R. Y. Stanier, Arch. Biochcm. Biophys., 38, 
245 (1952). 

3. A. Siegel, S. J. Singer, and S. G. 
Wildman, Arch. Biochcm. Biophys., 41, 
278 (1952). 

4. D. Billen and E. Volkin, /. BacterioL, 
67, 191 (1954). 

5. S. Dagley and J. Sykes, Arch. Bio- 
chcm. Biophys., 63, 338 (1956). 

6. H. E. Wade and D. M. Morgan, Bio- 
chcm. I, 65, 321 (1957). 

7. A. B. Pardee, K. Paigen, and L. S. 
Prestidge, Biochim. et Biophys. Acta, 23, 
162 (1957). 

8. E. T. Bolton, B. H. Hoyer, and D. B. 
Ritter, paper 3 of this volume. 

9. E. A. Peterson and H. A. Sober, 
/. Am. Chem. Soc, 78, 751 (1956). 

10. Report of the Biophysics Section, De- 
partment of Terrestrial Magnetism, Car- 
negie Inst. Wash. Year Boo\ 56, p. 118, 
Washington, D. C, 1957. 

11. S. Dagley, paper 7 of this volume. 

12. O. H. Lowry, N. J. Rosebrough, A. 
Furr, A. Lewis, and R. }. Randall, /. Biol. 
Chem., 193, 265 (1951). 

13. R. B. Roberts, P. H. Abelson, D. B. 
Cowie, E. T. Bolton, and R. J. Britten, 
Studies of Biosynthesis in Escherichia coli, 
Carnegie Inst. Wash. Publ. 607, Washing- 
ton, D. C, 1955. 

14. Report of the Biophysics Section, 
Department of Terrestrial Magnetism, 
Carnegie Inst. Wash. Year Boo\ 55, p. 110, 
Washington, D. C, 1956. 

15. W. Littlefield, E. B. Keller, J. Gross, 
and P. C. Zamecnik, /. Biol. Chem., 217 , 
111 (1955). See also J. Brachet, Biochemi- 
cal Cytology, pp. 4M7, 240-286, 296-355, 
Academic Press, New York, 1957. 


Microsomal Structure and Hemoglobin 
Synthesis in the Rabbit Reticulocyte 


Gates and Crellin Laboratories of Chemistry 

and Kercl^hoff Laboratories of Biology 

California Institute of Technology 

A great deal of evidence has now accumulated which suggests very strongly 
that microsomal particles are somehow connected with the process of protein 
synthesis [1, 2, 3]. Because of their high content of ribonucleic acid which 
might act as a coding template, it has become fashionable to postulate that these 
particles are the actual sites of protein assembly from activated single amino 
acids. To date, however, no evidence has been put forth which could be called 
direct proof that such is indeed the case. 

Nevertheless, the hypothesis is so attractive that it seems worth while to pro- 
ceed on the assumption that it is true and to investigate the detailed structure 
of microsomal particles and the relation of this structure to the protein which 
the particles are supposedly synthesizing. Such a study would have as its object 
an understanding of the molecular nature of the microsomal particle, the molec- 
ular structure of the growing peptide chains of the protein being synthesized, 
and, if possible, the interrelation between the two. 

To best carry out such a study, one would like to find a system consisting of 
free floating cells of a single type, actively engaged in the synthesis of predomi- 
nantly a single type of protein molecule. Fortunately these desirable properties 
are exhibited by mammalian reticulocytes. Such cells can be made in quantity 
if rabbits are made anemic by daily injections of phenylhydrazine. After a 
week of such injections, reticulocytes (immature red cells) account for 80 to 90 
per cent of the red cells present in the blood. These cells are actively producing 
hemoglobin and will continue to do so for many hours if suspended in the 



proper incubation medium [4, 5]. During such incubation, hemoglobin ac- 
counts for over 90 per cent of the protein produced. 

The microsomal fraction of rabbit reticulocytes has been shown to be very 
active in incorporating radioactive amino acids [6]. The following work repre- 
sents a beginning step toward an understanding of the relationship between 
microsomal structure and hemoglobin synthesis in the sense described above. 


Preliminary experiments on isolation of the microsomal fraction showed that 
the predominant component had a sedimentation coefficient of about 80 S. Vari- 
ations in the method of breaking the cells and in the buffer used for isolating 
the microsomal particles were explored before the following standard procedure 
was developed. The reticulocytes were frozen, thawed, and mixed with 3 vol- 
umes of cold buffer containing 0.14 M KCl, 0.001 M MgCl 2 , and 0.01 M tris- 
chloride, pH 7.2. Cell walls and debris were spun out at low speed. Micro- 
somes were then pelleted at 100,000g for 3 hours. The red pellet thus obtained 
was then twice redissolved and respun in 100 volumes of buffer, giving a light 
amber-colored pellet after the third centrifugation. 

Microsomal particles prepared in this way were found to give three compo- 
nents in the ultracentrif uge : 82 per cent of 78 S, 9 per cent of 120 S, and 9 per 
cent of 50 to 60 S. On electrophoresis in the same buffer the preparation showed 
a negative charge, and migrated with only slight skewing in the descending 
limb and a splitting into two components in the ascending limb. 

The intrinsic viscosity was found to be 0.08 dl/g in the buffer, and the partial 
specific volume was found to be 0.63 ml/g. 

From these numbers it may be calculated that, if the particles are spherical, 
the hydration is 2.6 g of water per gram of anhydrous particle, a very high 
value indeed. Taking as a model a highly hydrated sphere, one calculates a 
frictional coefficient of 1.72 and a molecular weight of 4.1 XlO 6 . 

That such a model cannot be very far wrong was indicated by light-scatter- 
ing measurements on the same preparations. These showed a molecular weight 
of 4 XlO 6 and a measured dissymmetry of 1.08 (45°/135°) with light of the 
mercury blue line. Since a dissymmetry of 1.06 is to be expected for a hydrated 
sphere of this molecular weight (diameter, 340 A), the axial ratio cannot be far 
from unity, and shapes such as rods and random coils are definitely excluded. 

The protein/RNA ratio for these particles was found to be almost unity on 
a weight basis. No lipid could be extracted. If the RNA from these particles 
is banded in an equilibrium density gradient [7] in cesium formate, the molec- 
ular weight of the RNA is found to be approximately 500,000. This is con- 
sistent with a small integral number of RNA molecules per microsomal par- 
ticle. If the above molecular weights are accepted, this integral number is 4. 




In the following experiments rabbit reticulocytes were centrifuged several 
times in buffer, suspended at 37° in a medium containing amino acids, iron, 
and other materials necessary to ensure optimal hemoglobin syntheses. A given 
C 14 labeled amino acid was added, and the living cells were incubated for a 
definite time period; then several volumes of ice-cold saline were added and the 
cells centrifuged several times from cold saline to remove extracellular label. 
The cells were then frozen, and microsomes were prepared as described above. 

When the cells were incubated for various time intervals with carboxyl C 14 
labeled leucine (15,500 cpm/mg) the resultant specific activities of protein in 
microsomes and in hemoglobin were as shown in table 1. 

It would appear that a steady-state concentration of labeled amino acids in 
the microsomes is reached within approximately 10 minutes or less. 








Specific Activity, 


cpm/mg protein 















Rate of Activity 

Increase in 



The labeled amino acid in the purified microsomes is present in some tightly 
bound form, as shown by the facts that (1) it is not removed at all by dissolv- 
ing the microsomes in buffer saturated with nonlabeled leucine and (2) little, 
if any, count is removed by extraction with trichloroacetic acid. 

However, the label is rapidly turned over in the living cell, as shown by the 
fact that if cells are labeled for 15 minutes in medium containing radioleucine, 
and then placed in medium containing nonlabeled leucine for 15 minutes, 90 
per cent of the label is removed from the microsomes. This shows that most 
of the label present in the microsomal particle is in a dynamic state, and is not 
a permanent part of the microsome structure. 

In order to obtain information concerning the amino acid composition of the 
transient material in the microsome, experiments were conducted using various 
labeled amino acids in the incubation medium. Radioleucine was always run 
as a control, and the molar ratios of other amino acids to leucine were deter- 
mined in the hemoglobin and microsomes after 15 minutes of incubation. The 
results are summarized in table 2, where the labeled amino acid ratios are com- 
pared with the total amino acid ratios determined by microbiological assay. It 
can be seen that the leucine-histidine ratio of transient material in the micro- 
some is compatible with the supposition that this material is hemoglobin pro- 
tein and not microsome structural protein. The leucine-phenylalanine ratio is 


not very informative, and the measured leucine-arginine ratio, which lies almost 
exactly half way between microsomal protein and hemoglobin, is very ambigu- 
ous. These results are compatible with the assumption that the transient ma- 
terial in the microsome is largely pre-hemoglobin. More evidence is needed to 
confirm this assumption, however. 

TABLE 2. Molar Ratios of 


to V; 

inous Amino Acids 

















Phenylalanine 2.5 









Boiling microsomes with 66 per cent ethanol extracted about 25 per cent of the 
radioactive material, whose free amino nitrogen increased greatly on renuxing 
with 6 N HC1, suggesting that this material is peptide in nature. The specific 
activity of the material was 5 to 10 times that of the unextracted microsomal 
protein. Both the extractable and nonextractable radioactive materials were 
transient; i.e., the counts were removed on incubating cells in nonlabeled amino 
acid for 15 minutes. 

One would like to conclude from the above extraction data that the extracted 
material is richer than the whole microsome in growing peptide chains of short 
length, whereas the unextractable material represents growing peptide chains 
which are too long to dissolve in 66 per cent ethanol. The short chains would 
presumably represent the earliest stages of hemoglobin formation in the micro- 
some. Further purification and characterization are necessary to prove this 


The above data lead to some interesting results if the following assumptions 
(or approximations) are made: (1) hemoglobin is the only protein being made 
in rabbit reticulocytes; (2) all microsomal particles are equally active in syn- 
thesizing hemoglobin; (3) all the transient label in the microsome is pre-hemo- 
globin. The steady-state label level of the microsomal particle, the specific ac- 
tivity of leucine used, together with the facts that 12 per cent of the protein is 
leucine and one-half of the microsome is protein, lead to the conclusion that 
0.05 per cent of the mass of the microsomal particle is pre-hemoglobin, i.e., grow- 
ing peptide chain. Since the molecular weight of the microsomal particle is 
4,000,000 as shown above, this means that the average weight of growing chain 
per particle is 2000. In a random population of growing chains the average 
weight might be expected to be about one-half of the finished chain weight. If 
all the growing chain per particle is in one piece, this leads to a value of 4000 
for the finished weight of polypeptide chain made per microsomal particle, a 
value reasonably close to the weight of one-fourth of a hemoglobin molecule, 


i.e., one polypeptide chain. If assumption 2 above is incorrect, and only a frac- 
tion of the microsomal particles is functional, the agreement is even better. 
From the rate of incorporation of label into finished hemoglobin molecules and 
the concentrations in the living cells of hemoglobin (15 per cent) and micro- 
somes (0.5 per cent), one may calculate that to account for the production of 
new hemoglobin each microsomal particle must make one-quarter of a hemo- 
globin in 1.5 minutes. 


The above data lead to the picture of a microsomal particle as an almost spher- 
ical sponge-like structure of anhydrous molecular weight 4,000,000 and diame- 
ter 340 A. One-half of the mass is represented by ribonucleic acid which ap- 
pears to be present as four strands of molecular weight 500,000. The half of 
the microsomal particle which is protein appears to be almost entirely (99.9 per 
cent) structural in nature; i.e., it is not transient protein precursor. 

Woven into this sponge-like structure in some way is a very small amount 
(0.05 per cent by weight) of transient protein precursor. Taken together with 
the observed rate of hemoglobin production, this amount of precursor is com- 
patible with the conclusion that one microsomal particle makes one polypeptide 
chain of hemoglobin in approximately 1 minute. 

Contribution No. 2338 


1. H. Borsook, C. L. D:asy, A. J. 4. J. Kruh and H. Borsook, /. Biol. 

Haagen-Smit, G. Keighley, and P. H. Chem., 220, 905 (1956). 

Lowy, /. Biol. Chan., 187, 839 (1950). 5. H. Borsook, E. H. Fischer, and G. 

2. J. W. Littlefield, E. B. Keller, J. Gross, Keighley, /. Biol. Chan., 229, 1059 (1957). 

and P. C. Zamecnik, /. Biol. Chan., 217, 6. M. Rabinovitz and M. E. Olson, 

111 (1955). Exptl. Cell Research. 10, 747 (1956). 

3. J. W. Littlefield and E. B. Keller, 7. M. Meselson, F. W. Stahl, and J. 

/. Biol. Chan., 224, 13 (1957). Vinograd, Proc. Natl. Acad. Sci. U. S., 43, 

581 (1957). 


Effects of ^-Fluorophenylalanine on the 
Growth and Physiology of Yeast 1 


Department of Bacteriology, University of Wisconsin 

Department of Bacteriology, University of Illinois 

Halvorson and Spiegelman [1] examined a series of amino acid analogs for 
their ability to inhibit growth of yeast, to deplete the "free amino acid pool," 
and to synthesize protein. When exponentially growing yeast is washed and 
resuspended in a nitrogen-free buffer in the presence of glucose, the free amino 
acids are rapidly incorporated into proteins, thus depleting the amino acid pool. 
In the presence of high concentrations of ^-fluorophenylalanine (10 -2 M to 
2X10" 2 M), this depletion is severely inhibited, as shown by chromatographic 
study of the pool components or by analysis of the glutamic acid content of the 
pool. This inhibition was interpreted as follows: The presence of an amino 
acid analog prevents the incorporation not only of its natural homolog but of all 
the other amino acids as well. 

On the other hand, Munier and Cohen [2] found that /7-fluorophenylalanine, 
when added to an exponentially growing culture of Escherichia coli, caused a 
linear growth. During that period the differential rate of incorporation of 
valine or S 35 

(Avaline or AS 35 ) /Amass 

was the same as in a control culture, although incorporation was somewhat 
slower than during exponential growth. Radioactive amino acids were formed 

1 This investigation was aided by a grant (G-4258) from the Division of Biological and 
Medical Sciences of the National Science Foundation. 



from radioactive glucose and incorporated in the presence of the analog during 
the linear growth. In addition, /7-fluorophenylalanine was incorporated (Munier 
and Cohen [2], [3]) to a great extent (up to 200 m|jmoles/g dry weight). In- 
duced (3-galactosidase was synthesized at the same differential rate as in a con- 
trol culture without analog. Because of the contradictory results of these two 
studies, it was decided to re-examine the effects of /7-fluorophenylalanine on 


Saccharomyces italicus Y1225 was used in these experiments. The cultures 
were grown in synthetic medium (Halvorson and Spiegelman [1]) in Erlen- 
meyer flasks which were shaken at 30° C. During the exponential phase of 
growth, the protein content per cell was found to be proportional to the opti- 
cal density. Therefore, for reasons of convenience, growth was followed by 
measurements of the optical density in a Beckman DU Spectrophotometer at 
600 mu. Under these conditions 1.00 O D = 772 ug dry wt./ml. 

Viable counts were determined by plating appropriate dilutions on dextrose 
broth agar medium. 

The cells were centrifuged, washed, and fractionated for isotope distribution 
as previously described (Halvorson [4]). The components of the protein hy- 
drolysates were identified by radioautography. For radioactivity measurements, 
aliquots were evaporated to dryness on stainless-steel planchets and counted in 
a gas flow counter, and the radioactivities were corrected to infinite thinness. 

The photomicrographs were taken on a 35-mm microfile film with a 100 
apochromatic objective, N. A. = 1.30. Magnification on the film was 1250 X. 
3-C 14 -DL-phenylalanine (Phe) (2.1 mc/mmole), 4-4'-C 14 -DL-valine (1.33 mc/ 
mmole), and 3-C 14 -DL-p-fluorophenylalanine (/>-FPhe) (2.35mc/mmole) were 
obtained from the Commissariat a l'Energie Atomique, France. Uniformly 
labeled glucose (2.4 mc/mmole) was obtained from the Fisher Scientific Com- 
pany, and carrier-free S 35 sulfate from the Oak Ridge National Laboratory. 


Effects of p-Fluorophenylalanine and $-2 Thienylalanine on the Growth of 
Yeast. Figure 1 shows that, upon addition of 10 2 M /7-FPhe or $-2 thienyl- 
alanine (Thiala) to an exponentially growing culture of yeast, the mass in- 
creases at a rate which is linear with time, as previously observed with E. coli. 
In this particular experiment, from the time of addition to the termination of 
growth, the mass increased 4 times in the presence of /7-FPhe and 2.7 times in 
the presence of Thiala. Under these conditions, although no component of the 
medium was limiting for growth, there was no increase in viable count in 
the presence of the antagonists. The increase in optical density in the presence 
of /7-FPhe (4 times) can be attributed to an increase in cell size (fig. 2). As- 
suming that yeast cells are ellipsoids of revolution, the average cell volume in- 
creased approximately 3.6 times in the presence of p-FPhe. 



H-q Dry Weight /ml 






120 240 

Time (min) 


Fig. 1. Effects of p-fluorophenylalanine and (3-2-thienylalanine on the growth of yeast. 

Synthesis of Cell Material during Linear Growth. Washed cells from an 
exponentially growing culture were placed in two flasks, one containing C 14 
glucose, and the other containing C 14 glucose and /?-FPhe (10~ 2 M final 

Optical density was followed throughout the experiment, and 10-ml samples 
taken at intervals were placed in precooled centrifuge tubes (0° C). The 
cells in the control culture were allowed to increase 2.6 times, and those in 
the linear culture 1.9 times. The samples were fractionated (Halvorson [4]), 
and the radioactivity of the hot-TCA-soluble fraction (containing the nucleic 
acids) and of the protein fraction (hot-TCA-insoluble) was determined. The 
differential rates of synthesis of total hot-TCA-soluble material and of total pro- 
tein were then plotted. The differential rate of synthesis of a component X is 
expressed by AX/AM, M being the increase in mass during the time necessary 

Fig. 2. Effect of /7-fluorophenylalanine on the size of yeast cells. Pictures taken initially 
(.i), after an optical density increase of 2.7 times (B), and 5.4 times (C). 



to obtain the increase AX; this expression is independent of the rate of growth. 
Figures 3 and 4 show that the differential rates of synthesis of hot-TCA-soluble 
material and of protein are the same in the presence and absence of p-FPhe. 

/ig Hot TCA Soluble C/m 

H-t Protein C/m 

o control 
• p-FPhe 

100 200 

eg Dry Weight/ml 

100 200 

CO Dry Weight /ml 

Fig. 3. Differential rate of synthesis of 
total hot-TCA-soluble material in the pres- 

Fig. 4. Differential rate of synthesis of 
proteins in the presence or absence of 

ence or absence of p-fluorophenylalanine. See p-fluorophenylalanine. See text for details. 
text for details. 

Incorporation of C li -p-Eluorophenylalanine in Exponentially Growing Yeast. 
Radioactive /7-FPhe was added to two exponentially growing cultures at final 
concentrations of 3.87 X 10 4 M and 1.04 X 10" 3 M, and the cultures were allowed 
to grow linearly (300 minutes) until their mass increased by 3.6 times. Samples 
were taken at given intervals, and the radioactivity of the protein fraction was 
determined. The differential rate of incorporation of /7-FPhe was calculated, 
and from the slopes of the straight lines obtained (fig. 5) the content of /7-FPhe 
in the proteins was found. As in E. coli, increasing the external concentrations 
of p-FPhe increases the analog content of the proteins. The incorporation is 
far from negligible, reaching 12 per cent of the normal phenylalanine content 
for a concentration of the analog of 1.03 XlO" 3 M (31.2 umoles /7-FPhe/g dry 
wt.). Munier (personal communication) and Kerridge (quoted in Gale and 
McQuillen [5]) have also shown that /7-FPhe is incorporated into yeast proteins. 
The relative amount of phenylalanine in yeast was determined from its differ- 
ential rate of incorporation. In two experiments, where the C 14 phenylalanine 
concentration was 3.12 X 10" 5 M and 6.25 X 10~ 5 M respectively, identical differen- 
tial rates of incorporation were observed. 

In all experiments, radioautograms of the acid hydrolysates of the protein 
fractions were made. Both for cells grown on C 14 -Phe and for those grown on 
C 14 -/7-FPhe, the radioactivity of the protein hydrolysates was identified exclu- 
sively with the added isotope. 



m/i moles Protein p-FPhe /ml 

3 87-10" M 

100 150 

eg Dry Weight/ ml 


0. 3 



Ht) Protein S/ml 

o control 
• p-FPhe 

20 30 

Time ( mln) 


Fig. 5. Differential rate of incorporation Fig. 6. Effect of p-fluorophenylalanine 

of p-fluorophenylalanine. See text for details. on S 35 sulfate incorporation into proteins of 

resting yeast cells. See text for details. 

Incorporation of p-Fluorophenylalanine under "Resting" Conditions. Since 
the experiments of Halvorson and Spiegelman [1] were conducted with "rest- 
ing" cells, it was interesting to know whether incorporation of p-FPhe occurred 
also under these conditions. Exponentially growing yeast was washed twice 
with buffer and resuspended in phosphate-succinate buffer, pH 4.7, with glu- 
cose as energy source, but without an exogenous nitrogen source. The sus- 
pension was divided in two flasks to which C 14 Phe and C 14 p-FPhe were 
respectively added (final concentration: 6.25x10 3 M). In both flasks the in- 
corporation of radioisotope was linear with time and ceased after 50 minutes' 
incubation. At this time, 31.8 mumoles of Phe had been incorporated, whereas 
19.4 mumoles of p-FPhe had been incorporated in the proteins per milliliter 
of culture. Thus, the analog was incorporated at 61 per cent the extent of Phe 

Incorporation of Sulfur from Radiosulfate into Yeast Proteins under "Rest- 
ing" Conditions. An exponential culture from synthetic medium was cen- 
trifuged, washed, resuspended in nitrogen-free medium, and divided in two 
flasks containing S 35 with and without 0.01 M p-FPhe. The incorporation of 
S 35 into proteins was linear with time over 40 minutes with identical slopes of 
incorporation in the two flasks (fig. 6). 

Incorporation of Radiovaline under "Resting" Conditions in Presence or 
Absence of p-FPhe. Yeast was grown in broth, centrifuged while in the ex- 



ponential phase, washed twice, and resuspended in phosphate-succinate buffer 
with glucose but without an exogenous source of nitrogen. Radioactive valine 
was added at time, and the suspension was shaken for 8 minutes at 30° C. 
The suspension was then centrifuged, washed twice in the cold, and divided 
into two flasks in the same medium without valine, and with and without 
10" 2 M p-FPhe. Valine incorporation proceeded at the same rate in the two 
flasks (fig. 7) ; as shown previously (Halvorson and Cohen [6]), the valine pool 
size is sufficient under these conditions for an unchanged rate of valine incor- 
poration. These experiments show that, under growing or "resting" condi- 
tions, p-FPhe does not inhibit protein synthesis. It was then interesting to 
find out why, under the conditions of Halvorson and Spiegelman [1], the pools 
were not depleted in the presence of p-FPhe. 

m/imoles Protein Valine/ml 
I4 r 

"iMfl Protein S/ ml 


» NH, 


10 12 

Time (mln) 

Fig. 7. Valine incorporation into the pro- 
teins of resting yeast cells. See text for 

a p-FPhe 

100 200 


Fig. 8. Effect of exogenous nitrogen on 
the incorporation of S 35 into proteins in ni- 
trogen-starved cells. See text for details. 

The Effect of Nitrogenous Compounds on the Incorporation of Sulfur into 
Proteins in Nitrogen-Starved Cells. A culture growing exponentially in broth 
was centrifuged, resuspended with glucose in a phosphate-succinate buffer, and 
then shaken for 18 hours (original dry weight: 1.55 mg/ml). These starved 
cells were then washed and resuspended in nitrogen-free, phosphate-succinate 
buffer with glucose and S 35 sulfate with or without 10~ 2 M NH 4 C1, 10 -2 M Phe, 
or 10" 2 M p-FPhe. The curves of figure 8 show that /7-FPhe and Phe can act as 
nitrogen sources for S incorporation, probably providing — NH2 groups through 
transamination to carbon acceptors. 

Effect of p-Fluorophenylalanine on Protein Degradation. In the absence of 
an exogenous source of nitrogen and energy, the degradation of cellular proteins 
leads to elevated pool levels (Halvorson [7]). Since high amino acid pool levels 



are observed in the presence of p-FPhe (Halvorson and Spiegelman [1]), a 
direct test of its effect on the rate of protein degradation was undertaken. 

Cells were grown overnight in synthetic medium in the presence of radio- 
active valine or Phe. They were then centrifuged, washed, resuspended in fresh 
synthetic medium containing glucose, and shaken for 210 minutes in order to 
diminish any pool radioactivity that might have been present. The cells were 
again centrifuged, washed twice, and resuspended to a density of 3 mg dry 
wt./ml in two flasks containing buffer with or without p-FPhe 10~ 2 M. Protein 
degradation was followed by the appearance of radioactivity in the soluble pool 
(cold-TCA-soluble radioactivity). The results in table 1 show elevated pool 
levels in the presence of p-FPhe. Since the previous experiments show that 
p-FPhe does not inhibit the amino acid incorporation observed in the presence 
of an exogenous source of energy, these results indicate that the antagonist ac- 
celerates the rate of protein degradation. 

TABLE 1. Effect of p-Fluorophenylalanine on Protein Breakdown 

Radioactivity Released f 

Incubation time,* 

Growth Supplement 


C 14 valine X 





C 11 phenylalanine § 










* Incubated aerobically in phosphate buffer, pH 4.5, at 30° C with or without 0.01 M 
t Increase in cpm of cold-TCA-soluble fraction/ml incubation mixture. 
t 19,860 cpm protein/ml incubation mixture. 
§ 81,320 cpm protein/ml incubation mixture. 


A reanalysis of the effects of /7-FPhe on the growth of yeast shows a strong 
parallelism with its effects on E. coli. A linear rather than exponential rate 
of growth is seen in the absence of cell division and without decreasing the dif- 
ferential rates of synthesis of protein and hot-TCA-soluble material or total 
carbon incorporation. Furthermore, in resting yeast cells, the incorporation of 
endogenous amino acids is not influenced by the presence of p-FPhe. The previ- 
ously observed inhibition of a-glucosidase synthesis by p-FPhe (Halvorson and 
Spiegelman [1]) may therefore represent another example of inactive enzyme 
synthesis (Cohen and Munier [8]). In contrast to the results found with E. coli, 
however, p-FPhe is capable of completely suppressing a-glucosidase induction 
in resting yeast cells only when added simultaneously with the inducer (Hal- 
vorson and Jackson [9]). When p-FPhe was added at various times after the 


inducer, induced synthesis became more and more refractory to the inhibitor. 

One feature of the effect of /7-FPhe on resting yeast cells requires special 
attention. Previous studies showed that, in the presence of /7-FPhe, the contents 
of the free amino acid pool remained essentially the same as in unstarved cells 
(Halvorson and Spiegelman [1]). On the basis of the present experiments, the 
nondisappearance of the pool in the presence of p-FPhe can be attributed to: 
(1) partial replenishment of the free amino acid pool from the nitrogen of 
/7-FPhe, and (2) an increased rate of protein degradation in the presence of 

The reversal of /7-FPhe inhibition by Phe can be related not only to a com- 
petition between these two amino acids for an accumulating system (Halvor- 
son and Cohen [6]) but also to its incorporation into proteins. Halvorson and 
Spiegelman [1] had derived from their experiments with /7-FPhe the conclu- 
sion that there were no intermediate precursors of induced maltozymase in 
yeast. Although amino acid antagonists may prove valuable for studies on 
protein synthesis, it would seem on the basis of these and other experiments 
(Cohen and Munier [8]) that conclusions derived from their use as a tool in 
the study of intermediates in protein synthesis are unwarranted. The de novo 
nature of induced enzyme synthesis has been established on other grounds, 
however (Hogness, Cohn, and Monod [10]; Rotman and Spiegelman [11]). 


We wish to express our appreciation to Dr. R. Munier for advice and assist- 
ance in the early phases of these experiments and to Dr. Pichat of the Com- 
missariat a l'Energie Atomique, France, for the synthesis of 3-C 14 -DL-/7-fluoro- 


The addition of 0.01 M /7-fluorophenylalanine to a growing culture of Sac- 
charomyces italicus Y1225 results in (1) an incorporation of the analog into 
cellular proteins, (2) a linear rather than an exponential rate of growth as a 
function of time, and (3) an inhibition of cell division. The antagonist does 
not influence the differential rates of synthesis of protein and hot-TCA-soluble 

Under "resting conditions," p-fluorophenylalanine does not inhibit either pro- 
tein synthesis or the utilization of the free amino acid pool. Elevated pools in 
the presence of the antagonist were attributed to pool replenishment from the 
nitrogen of antagonist and to an increased rate of protein degradation. 


1. H. O. Halvorson and S. Spiegelman, Incorporation d'analogues structuraux 
The inhibition of enzyme formation by d'amino acides dans les proteines bacterien- 
amino acid analogues, /. BacterioL, 64, 207 nes, Biochim. et Biophys. Acta, 21, 592- 
(1952). 593 (1956). 

2. R. L. Munier and G. N. Cohen, 3. R. L. Munier and G. N. Cohen, 



Incorporation d'analogues structuraux 
d'amino acides dans les proteines d'Esch- 
erichia coli, Ann. inst. Pasteur, in press. 

4. H. O. Halvorson, Studies on protein 
and nucleic acid turnover in growing cul- 
tures of yeast, Biochim. et Biophys. Acta, 
27, 267 (1958). 

5. E. F. Gale and K. McQuillen, Nitro- 
gen metabolism, Ann. Rev. Microbiol., 11, 
283 (1957). 

6. H. O. Halvorson and G. N. Cohen, 
Incorporation comparee des amino acides 
endogenes et exogenes dans les proteines 
de la levure, Ann. inst. Pasteur, in press. 

7. H. O. Halvorson, Intracellular protein 
and nucleic acid turnover in resting yeast 
cells, Biochim. et Biophys. Acta, 27 , 255 

8. G. N. Cohen and R. L. Munier, Effect 
des analogues structuraux d'amino acides 
sur la croissance, la synthase de proteines 
et la synthese d'enzymes, chez Escherichia 
coli, Ann. inst. Pasteur, in press. 

9. H. O. Halvorson and L. Jackson, The 
relation of ribose nucleic acid to the early 
stages of induced enzyme synthesis in 
yeast, /. Gen. Microbiol., 14, 26 (1956). 

10. D. Hogness, M. Colin, and J. Monod, 
Induced synthesis of beta-galactosidase in 
E. coli, Biochim. et Biophys. Acta, 16, 99 

11. B. Rotman and S. Spiegelman, On 
the origin of the carbon in the induced 
synthesis of beta-galactosidase in Escher- 
ichia coli, ]. Bacterial., 68, 419 (1954). 


Enzymatic and Nonenzymatic Synthesis in 

Adenyl Tryptophan 1 



Department of Biochemistry, Tufts University School of Medicine 

Acyl adenylates have been postulated as intermediates in the activation of 
acetate [1] and of fatty acids [2, 3], as well as in the synthesis of phenylacetyl- 
glutamine and hippurate [4]. Thus, the activation of acetate and of phenylace- 
tate may be represented as follows: 3 

Acetate + ATP ±± Acetyl-AMP + PP 
Phenylacetate + ATP ^± Phenylacetyl-AMP + PP 

The activation of amino acids, which has been observed with several enzyme 
preparations, appears to involve an analogous reaction [5, 6, 7, 8] : 

Amino acid + ATP *± Aminoacyl-AMP + PP 

This reaction has been followed by observing the formation of amino acid hy- 
droxamate when enzyme preparations are incubated with amino acid, mag- 
nesium ions, adenosine triphosphate, and high concentrations of hydroxylamine. 
The reaction has also been observed by determining the rate of exchange of 
radioactive inorganic pyrophosphate with adenosine triphosphate in the pres- 
ence of magnesium ions, amino acid, and enzyme. 

That the intermediate formed in the activation of amino acids is an aminoacyl 
adenylate of the type (fig. 1) postulated to occur in other systems [1-4] is sug- 
gested by several observations. For example, synthetic acyl adenylates are 

1 Supported in part by research grants from the National Science Foundation and the 
National Institutes of Health, Public Health Service. 

2 Postdoctorate fellow of the National Heart Institute, Public Health Service. 

3 Abbreviations: adenosine triphosphate, ATP; adenylic acid, AMP; pyrophosphate, PP. 




I || II 

- CH 2 -C-C - 0- P-O- Ribose -Adenine" 1 " 
I l 

NH 3 + 0~ 

Fig. 1. Tryptophanyl adenylate. 

known to react promptly with hydroxylamine to yield the corresponding hy- 
droxamates [1-4]. Furthermore, it has been reported that synthetic leucyl 
adenylate formed adenosine triphosphate when incubated with inorganic pyro- 
phosphate and an activating enzyme purified from Escherichia coli [8]; a simi- 
lar experiment has been carried out with methionyl adenylate and an activat- 
ing enzyme isolated from yeast [6]. An additional piece of evidence consistent 
with the formation of an anhydride linkage between the phosphoric acid 
group of adenylic acid and the carboxyl group of amino acids has been ob- 
tained in experiments with amino acids labeled with O 18 ; transfer of O 18 from 
the carboxyl group of the amino acid to adenylic acid was associated with 
enzymatic activation. The pyrophosphate formed did not contain appreciable 
quantities of O 18 [9]. 

The available data are consistent with the hypothesis that aminoacyl adenyl- 
ates are intermediates in the amino acid activation reaction, but the formation 
of such anhydrides has not yet been shown. Although there is as yet no experi- 
mental demonstration of the net synthesis of acyl adenylates in acetate or fatty 
acid oxidation, the respective enzyme systems are apparently able to utilize 
added synthetic acyl adenylate derivatives [1-3]. Previous inability to detect 
the formation of such intermediate anhydrides may be related to the instability 
of the anhydrides and perhaps also to the high affinity of the enzyme for the 
anhydride; accordingly the actual intermediate in these activation reactions 
may be enzyme-bound acyl adenylate. 

We have attempted to obtain evidence for the net synthesis of tryptophanyl 
adenylate by the tryptophan-activating enzyme of beef pancreas [7]. In these 
studies, we have used aminoacyl adenylates prepared as described in the fol- 
lowing paper by Castelfranco et al. [10]. Although tryptophanyl adenylate is 
hydrolyzed rapidly at pH 7.2 and 37° C, we have found that only about 10 per 
cent of the anhydride is hydrolyzed in 2 hours at pH 4.5 and 0°. Paper 
ionophoresis at pH 4.5 in 0.05 M ammonium formate buffer at 0° indicated 
that tryptophanyl adenylate was positively charged and moved with a greater 
mobility than tryptophan itself (fig. 2) . The positively charged band quenched 
the fluorescence of the paper under ultraviolet light and gave the ninhydrin 
color reaction. Elution of this material from the paper strip yielded an alkali- 
labile compound which formed a hydroxamic acid promptly on treatment with 
hydroxylamine. The hydroxamate was identified as tryptophan hydroxamate 
by paper chromatography in several solvent systems. Incubation of the eluted 
compound with the tryptophan-activating enzyme, magnesium chloride, inor- 






6 cm 

Fig. 2. Paper ionophoresis of tryptophan (TRY) and tryptophanyl adenylate (TRY- 
AMP) in 0.05 M ammonium formate buffer (pH 4.5) ; apparatus of Markham and 
Smith [11]. 

ganic pyrophosphate, and tris(hydroxymethyl) aminomethane buffer led to 
synthesis of adenosine triphosphate. Adenosine triphosphate was identified by 
coupling the reaction between radioactive inorganic pyrophosphate and trypto- 
phanyl adenylate with the phosphorylation of glucose by hexokinase. The phos- 
phate esters were separated by ethanol-barium salt fractionation, and radioac- 
tive glucose-6-phosphate was identified by paper chromatography in two solvent 

An experiment designed to demonstrate the net synthesis of tryptophanyl 
adenylate was carried out as follows. The total yield of tryptophan-activating 
enzyme obtained from 10 lb of beef pancreas (15 mg) was incubated with 
DL-tryptophan-3-C 14 , magnesium chloride, crystalline pyrophosphatase, adenosine 
triphosphate, and tris (hydroxy methyl) aminomethane buffer for 30 minutes 
at 37° C. At the end of the incubation period, 2 mg of synthetic tryptophanyl 
adenylate was added as carrier and the reaction mixture was lyophilized. The 
lyophilized reaction mixture was fractionated according to the scheme shown 
in figure 3. 

In the first step of this procedure, treatment with glacial acetic acid in the 

TRY-C 14 


ATP + 

Mg ++ 

+ Enzyme 




J — " 

Enzyme ] 


v Ether 




Paper electrophoresis 
(/>H 4.5 and 0°) 

Fig. 3. Scheme for the isolation of tryptophanyl adenylate (TRY-AMP) from enzymatic 
reaction mixtures. 


cold separated tryptophanyl adenylate from the enzyme which remained in 
the insoluble residue along with some magnesium chloride and tris(hydroxy- 
methyl) aminomethane buffer. Subsequent addition of ether to the acetic acid 
extract resulted in the precipitation of adenosine triphosphate and tryptophanyl 
adenylate, leaving radioactive tryptophan in the supernatant solution. The pre- 
cipitate, which contained both adenosine triphosphate and tryptophanyl adenyl- 
ate, was washed several times with cold ether-glacial acetic acid followed by 
several ether extractions in order to remove residual acetic acid. The precipi- 
tate was dissolved in a small amount of ammonium formate buffer (pW 4.5), 
and an aliquot of this material was analyzed ionophoretically. The paper 
ionophoretic separation yielded a major radioactive band which corresponded in 
mobility to authentic tryptophanyl adenylate. A smaller band, negatively 
charged, was detected but has not yet been identified. The positively charged 
band was eluted, treated with hydroxylamine, and chromatographed on paper 
in several solvent systems. Radioactive tryptophan hydroxamic acid was identi- 
fied in each system. 

A similar experiment carried out without the addition of carrier trypto- 
phanyl adenylate was also performed. Ionophoretic analysis again revealed evi- 
dence for the formation of tryptophanyl adenylate; however, somewhat less 
radioactivity was found in the tryptophanyl adenylate area. These findings sug- 
gested that there was incorporation of radioactive tryptophan into tryptophanyl 
adenylate in the experiment with carrier. In order to investigate this possi- 
bility directly, radioactive tryptophan was incubated with tryptophanyl adenyl- 
ate, enzyme, and magnesium ions. The tryptophan hydroxamate isolated 
from this reaction mixture (after the addition of hydroxylamine) contained 
appreciable radioactivity; the findings therefore suggest that an exchange be- 
tween tryptophan and tryptophanyl adenylate occurred : 

TRY-C 1 4 + TRY-AMP -» TRY-C 14 -AMP 4- TRY 

We have also found that the tryptophan-activating enzyme catalyzes the 
formation of adenosine triphosphate from inorganic pyrophosphate and a va- 
riety of a-aminoacyl adenylates. Thus, aminoacyl adenylates of l- and D-trypto- 
phan, l- and D-phenylalanine, L-isoleucine, L-glutamine, L-alanine, glycine, L-pro- 
line, L-valine, L-leucine, and L-tyrosine gave adenosine triphosphate in this 
system. It is of some interest that d- and L-tryptophanyl adenylate and d- and 
L-phenylalanyl adenylate were about equally active. Examination of the D-anhy- 
drides (after hydrolysis) by optically specific enzymatic methods [12] revealed 
that the optical purity of the amino acid moieties was greater than 99.5 per cent; 
it is therefore unlikely that the activity of the D-aminoacyl adenylate is due to 
the presence of adenylate derivatives of the corresponding enantiomorphs. Of 
the aminoacyl adenylates examined, only those of a-amino acids were active. 4 

4 We have recently found that L-tryptophanyl inosinate [10] is inactive, and that inosine 
triphosphate is not active in place of adenosine triphosphate in the forward reaction (reac- 
tion 1, table 2). 


Thus, carbobenzoxytryptophanyl adenylate, 3 _a l an yl adenylate, and benzoyl 
adenylate were inactive (table 1). Table 2 summarizes some of the reactions 
catalyzed by the tryptophan-activating enzyme preparation. Under the condi- 
tions employed, the enzyme catalyzed the formation of L-tryptophan hydroxa- 
mate, but not that of the d isomer. This is in striking contrast to the results ob- 
tained on the synthesis of adenosine triphosphate (table 1). Further studies of 
the specificity of the enzyme system are in progress. 

TABLE 1. Specificity of the Tryptophan-Activating Enzyme with Respect to Acyl 
Adenylate in Synthesis of Adenosine Triphosphate 

Acyl Adenylates 

A . 

I \ \ 

Active Inactive 

L-Tryptophan-AMP 3-Alanine-AMP 

D-Tryptophan-AMP Acetyl-AMP 

L-Phenylalanine-AMP Benzoyl-AMP 

D-Phenylalanine-AMP Carbobenzoxytryptophan-AMP 

L-Glutamine-AMP Phenylacetyl-AMP 


TABLE 2. Types of Reactions Catalyzed by Tryptophan-Activating Enzyme 


Reaction Type l 


TRY + ATP + NH 2 OH + 




l-TRY-AMP + dl-TRY-C 14 


Amino acid + ATP + NH 2 OH 


Amino acid-AMP + PP + 


with d- and l-TRY-C 14 are not yet complete. 





Note Added in Proof 

Novelli [Proc. Natl. Acad. Set. U. S. t 44, 86 (1958)] has very recently reported 
synthesis of adenosine triphosphate from pyrophosphate and several aminoacyl 
adenylates with this enzyme, and Berg {Federation Proc., 16, 152 (1957) ; per- 
sonal communication] has made similar observations with a methionine-acti- 
vating enzyme obtained from yeast. Rhodes and McElroy (personal communi- 
cation) have recently observed enzymatic synthesis of adenyl oxyluciferin by 



firefly luciferase; they have also obtained evidence for tight binding of this 
intermediate to the enzyme. 

Recent studies in our laboratory indicate that the affinity of the tryptophan- 
activating enzyme is greater for L-tryptophanyl adenylate than for D-trypto- 
phanyl adenylate and a number of the other ot-aminoacyl adenylates listed in 
table 1. 


1. P. Berg, /. Biol. Chen?., 222, 1015 

2. W. P. Jencks and F. Lipmann, /. Biol. 
Chem., 225, 207 (1957). 

3. H. S. Moyed and F. Lipmann, /. 
Bacterial., 73, 117 (1957). 

4. K. Moldave and A. Meister, /. Biol. 
Chem., 229,463 (1957). 

5. M. B. Hoagland, E. B. Keller, and 
P. C. Zamecnik, /. Biol. Chem., 218, 345 

6. P. Berg, /. Biol. Chem., 222, 1025 

7. E. W. Davie, V. V. Koningsberger, 

and F. Lipmann, Arch. Biochem. Bio- 
phys., 65, 21 (1956). 

8. J. A. DeMoss and G. D. Novelli, Bio- 
chim. et Biophys. Acta, 22, 49 (1956). 

9. M. B. Hoagland, P. C. Zamecnik, N. 
Sharon, F. Lipmann, M. P. Stulberg, and 
P. D. Boyer, Biochim. et Biophys. Acta, 26, 
215 (1957). 

10. P. Castelfranco, A. Meister, and 
K. Moldave, paper 14 of this volume. 

11. R. Markham and J. B. Smith, Bio- 
chem. J., 52, 552 (1952). 

12. A. Meister, L. Levintow, R. M. 
Kingsley, and J. P. Greenstein, /. Biol 
Chem., 192, 535 (1951). 


Participation of Adenyl Amino Acids in Amino 
Acid Incorporation into Proteins 1 


Department of Biochemistry, Tufts University School of Medicine 

It has been postulated that aminoacyl adenylates possessing the general struc- 
ture shown in figure 1 are formed in the enzymatic activation of amino acids 

i ii ii 

R-C-C-O-P- 0-Ribose-Adenine + 

I i 

NH 3 + 0" 

Fig. 1 

by adenosine triphosphate, and that such "activated amino acids" are inter- 
mediates in the incorporation of amino acids into microsomal proteins observed 
in cell-free systems [1, 2]. Whether such incorporation represents protein syn- 
thesis is not yet known, although this hypothesis has indeed been considered. 
Recent studies in our laboratory have been directed toward a better under- 
standing of the activation of amino acids and the possible role of aminoacyl 
adenylates in amino acid incorporation into protein. In the preceding paper 
by Karasek et al. [3], evidence for the net synthesis of tryptophanyl adenylate 
by a purified tryptophan-activating enzyme is described. These observations 
appear to give direct support to the idea that aminoacyl adenylates are the 
initial products of the activation reaction. The subsequent reactions of ami- 
noacyl adenylates are not yet clearly understood, although it has been sug- 
gested that the amino acid moieties of such anhydrides are transferred to pro- 
tein via specific acceptors, possibly ribonucleic acid [4]. 

1 Supported in part by research grants from the National Science Foundation and the 
National Institutes of Health, Public Health Service. 

2 Postdoctorate fellow of the National Heart Institute, Public Health Service. 



We began this work by attempting to prepare a number of aminoacyl adenyl- 
ates in a reasonable state of purity and in good yield in order to make possi- 
ble the synthesis of radioactive aminoacyl adenylates. It is not unusual in mod- 
ern biochemical research to synthesize and study compounds that are believed 
to be intermediates in biochemical reactions. In the present instance this ap- 
proach may suffer from a possible difficulty in that the intermediates may be 
bound to enzymes and therefore not be in equilibrium with an external source 
of intermediate. It has been observed, however, that chemically synthesized acyl 
adenylate derivatives are enzymatically active in systems that catalyze activa- 
tion of acetate [5] and fatty acids [6, 7] and the synthesis of phenylacetyl- 
glutamine and hippurate [8]. 

Two procedures have been described for the preparation of aminoacyl adenyl- 
ates. One of these [9] involves the condensation of the acid chloride of an 
amino acid with silver adenylate. DeMoss et al. [9] obtained leucyl adenylate 
in 9 per cent yield by this procedure. The other method involves condensa- 
tion of the free amino acid and adenylic acid in the presence of N,N'-dicyclo- 
hexylcarbodiimide, and precipitation of the product by addition of acetone [2]. 
In our hands, these methods suffered from shortcomings often encountered in 
the attempted synthesis of highly reactive molecules; thus, we obtained very 
low yields of products of very low purity, and experienced great difficulty in 
attempts at purification of the anhydrides. 

In an effort to solve these problems, we investigated a number of synthetic 
approaches, two of which have proved successful. The first consisted of con- 
densing an N-carbobenzoxyamino acid anhydride with adenylic acid in aque- 
ous pyridine; the N-carbobenzoxyaminoacyl adenylate was isolated, and the 
blocking group was removed by catalytic hydrogenation with palladium [10]. 

Subsequently, an alternative synthesis was developed which has proved to be 
more convenient; it will therefore be described here in greater detail. Equi- 
molar quantities of N-carbobenzoxyamino acid and adenylic acid were shaken 
in aqueous pyridine with an excess of N,N'-dicyclohexylcarbodiimide for sev- 
eral hours. N,N'-Dicyclohexylurea was removed by filtration, and the filtrate 
was treated with acetone to precipitate the product. Treatment with acetone 
removed unreacted N-carbobenzoxyamino acid, N,N'-dicyclohexylcarbodi- 
imide, and most of the pyridine and water. The precipitate was extracted 
with ethylene glycol monomethyl ether; the product is soluble, and adenylic 
acid is insoluble in this solvent. The N-carbobenzoxyaminoacyl adenylate was 
precipitated from the extract by addition of ether. After catalytic hydrogena- 
tion of the carbobenzoxy compound and removal of the catalyst, the super- 
natant solution was lyophilized. The free aminoacyl adenylate was obtained 
as a white powder. 

The yields varied from 40 to 80 per cent; for example, glycyl-l-C 14 -adenylate 
was obtained in 75 per cent yield, and the yield of DL-tryptophanyl-3-C 14 -adenyl- 
ate was 44 per cent. The final products are estimated to be 70 to 80 per cent 


pure on the basis of the hydroxamic acid-ferric chloride color reaction [11]. 
The major impurities consist of adenylic acid and amino acid formed by hy- 
drolysis of the anhydride during hydrogenation. This method has been suc- 
cessfully applied to the following amino acids: glycine, alanine, valine, leucine, 
isoleucine, (3-alanine, proline, phenylalanine, tyrosine, tryptophan, glutamine, 
asparagine, threonine, methionine, and serine. Studies on the remaining natural 
amino acids are in progress. 

Evidence for the proposed anhydride structure (fig. 1) includes the follow- 
ing: (a) hydrolysis in alkaline solution yields equivalent quantities of adenylic 
acid and amino acid; (b) reaction with hydroxylamine yields the correspond- 
ing amino acid hydroxamates, which have been identified by paper chromatog- 
raphy; and (c) paper ionophoretic study indicates that the aminoacyl adenyl- 
ates have a net positive charge at pW 4.5. The possibility that the carboxyl 
group of the amino acid may be linked to the adenylic acid moiety through 
a group (e.g., 6-amino group of adenine) other than the phosphoric acid group 
appears unlikely in view of the unusual reactivity of these compounds. 3 An 
additional property of a-aminoacyl adenylates which has proved of value in 
characterization is their reactivity in the presence of the tryptophan-activating 
enzyme and inorganic pyrophosphate to yield adenosine triphosphate [3]. 

a-Aminoacyl adenylates are very labile in aqueous solution at values of pH 
above 5.5. Thus, at pH 7.2 at 37° C, they exhibited half-lives of 5 to 10 minutes. 
On the other hand, carbobenzoxyaminoacyl adenylates suffered only about 10 
to 20 per cent hydrolysis in 2 hours at 37° at pH. 7.2. Acetyl adenylate and 
benzoyl adenylate exhibit stability of approximately the same order as carbo- 
benzoxyaminoacyl adenylates under these conditions. 

Preparation of glycyl-C 14 -adenylate and tryptophanyl-C 14 -adenylate made it 
possible to study incorporation of the respective amino acid moieties into pro- 
teins in systems previously employed for studies of amino acid incorporation. 
The enzyme preparation was obtained as described by Zamecnik and Keller 
[12]; it consisted of the supernatant solution (containing microsomes) ob- 
tained by centrifuging a 25 per cent rat liver homogenate at 12,000^. This 
preparation catalyzed the incorporation of amino acids into microsomal pro- 
teins in the presence of adenosine triphosphate and an adenosine triphosphate- 

3 Additional evidence for the proposed structure has recently been obtained. Thus, we 
have been able to convert carbobenzoxytryptophanyl adenylate with nitrous acid to the 
corresponding inosinic acid derivative. The latter compound has also been prepared by 
condensing inosinic acid with N-carbobenzoxytryptophan by the procedure described in the 
text for anhydrides of adenylic acid. 

Acylation of the hydroxyl groups of ribose appears to be excluded. Thus, carbobenzoxy- 
aminoacyl adenylates consumed theoretical quantities of periodate and, after reaction with 
periodate, reacted with hydroxylamine to give the corresponding carbobenzoxyamino acid 
hydroxamates. Paper ionophoretic study of the carbobenzoxyaminoacyl adenylates in borate 
and other buffers was also consistent with the presence of free ribose hydroxyl groups; 
the mobility of these compounds (and of adenylic acid) was greater in borate buffer than 
in tris(hydroxymethyl) aminomethane buffer at pH 9.1. 


generating system. Thus, in the system of Zamecnik and Keller [12], we 
observed 57 and 81 cpm/mg protein, respectively, with glycine-1-C 14 and trypto- 
phan-3-C 14 . As is indicated in table 1, incubation of the enzyme preparation 
with radioactive glycyl adenylate resulted in significant incorporation of isotope 
into the protein subsequently isolated. When the anhydride was hydrolyzed 
with alkali before study, significant incorporation was not observed. Further- 
more, equimolar concentrations of radioactive glycine plus adenylic acid did 
not lead to incorporation. It should be emphasized that the specific activities of 
the C 14 -aminoacyl adenylates were 5 per cent of the values for the free amino 
acids used by us in the system of Zamecnik and Keller. 

TABLE 1. Incorporation Studies 

Reaction Mixtures * cpm/mg 

Enzyme + Glycine-1-C 14 + Adenylate 0.1 7 

Enzyme + Glycyl-l-C 14 -adenylate 17.1 

Enzyme t + Glycyl-l-C 14 -adenylate 195. 

Enzyme t + Glycine-1-C 14 + Adenylate 1 .3 1 

* The reaction mixtures contained enzyme ( 1 ml) 
and glycyl-C 14 -adenylate (2.5 /zmoles; 3.6 X 10 5 cpm) 
in a final volume of 2.5 ml; incubated at 38° for 
30 minutes. Similar results were obtained when 
concentrations of aminoacyl adenylate from 10" 2 M 
to 10~ 6 M were employed. 

f Enzyme heated at 100° for 10 minutes. 

Although these results appeared to be consistent with the hypothesis that 
aminoacyl adenylates are intermediates in the incorporation of amino acids into 
proteins, further experiments have raised the possibility that such incorpora- 
tion may be explained in terms of nonenzymatic acylation of protein. Thus, 
it was found that, when the enzyme preparation was heated for 10 minutes at 
100° before incubation with C 14 -aminoacyl adenylate, the incorporation of 
isotope into protein was considerably greater than with the unheated enzyme 
preparation. In the experiments with heated enzyme, appreciable incorporation 
of isotope did not occur with hydrolyzed anhydride preparations. Similar re- 
sults have been obtained with tryptophanyl-C 14 -adenylate. 

With both heated and unheated enzyme preparations, the binding of the in- 
corporated amino acid to protein was quite stable and could be released only 
by the drastic acid hydrolysis required for the cleavage of peptide bonds. Thus, 
with glycine-l-C 14 -labeled protein (heated and unheated), the quantity of free 
amino acids and the percentage of isotope released as C 14 2 by ninhydrin in- 
creased in parallel fashion during hydrolysis with 6 N HC1 at 105° over a 
period of 16 hours. When the proteins labeled by incubation with glycyl-1-C 14 - 
adenylate and with tryptophanyl-3-C 1 '-adenylate were treated with l-fluoro-2,4- 
dinitrobenzene, followed by acid hydrolysis, dinitrophenylamino acid prepara- 
tions were obtained which contained more than 70 per cent of the radioactivity 


originally incorporated into the protein. Similar results were obtained with 
heated and unheated enzyme preparations. 

We have also found that, when glycyl-l-C 14 -adenylate is incubated with puri- 
fied rat-liver ribonucleic acid obtained by phenol extraction [13], considerable 
radioactivity remains associated with the ribonucleic acid preparation after ex- 
haustive dialysis and ethanol precipitation. Approximately 40 per cent of the 
radioactivity of the ribonucleic acid preparation was alkali-labile. When such 
ribonucleic acid preparations were incubated with unheated and heated pro- 
tein preparations, significant quantities of radioactivity were found in the protein 
subsequently isolated. Thus, incubation of 1 micromole of glycyl-l-C 14 -adenyl- 
ate with 50 mg of liver ribonucleic acid in 1 ml of water for 30 minutes at 
38° gave a ribonucleic acid preparation containing 3000 cpm (after dialysis 
and precipitation). When 20 mg of this C 14 -ribonucleic acid (1500 cpm) was 
incubated with enzyme preparation (table 1) for 30 minutes at 38°, approxi- 
mately 100 cpm was associated with the protein subsequently isolated. Similar 
results were obtained with heated enzyme preparation. 

The observed reactions of the aminoacyl adenylates with proteins are consist- 
ent with the reactivity expected of anhydrides of this type. A similar result might 
occur when proteins are treated with radioactive acetic anhydride. The reaction 
of the aminoacyl adenylates with the heated protein preparations would appear 
to be a nonenzymatic acylation reaction involving the free reactive groups of 
proteins. Heat denaturation of the protein would be expected to expose a greater 
number of amino groups to the action of the acylating agents. 4 The extent to 
which the labeling observed with unheated proteins may be due to an enzy- 
matic mechanism is not known. Although we believe that in the present studies 
transfer of the amino acid moieties from aminoacyl adenylates to proteins oc- 
curred largely by a nonenzymatic process, the possibility cannot be excluded 
that some of this transfer is enzymatically catalyzed. Perhaps nonenzymatic 
acylation of proteins may also take place to some extent when amino acids are 
incubated with adenosine triphosphate, an adenosine triphosphate-generating 
system, and a suitable enzyme preparation. The aminoacyl adenylates formed 
in the activation reaction [3] might be expected to react in such a manner. 
Thus, the high reactivity of aminoacyl adenylates with protein and ribonucleic 
acid may explain at least some of the reported [4, 12] incorporation phenomena. 
It must be emphasized that the concentrations of aminoacyl adenylates pre- 
sumably formed in the amino acid incorporation systems previously studied 
would be expected to be considerably lower than the concentrations of aminoacyl 
adenylates we have used. Much lower concentrations of aminoacyl adenylates 
than those used here must be employed to make a meaningful comparison 
of the labeling in the two systems. Such studies will require aminoacyl adenyl- 
ates of considerably higher specific radioactivity. Although the present in- 
vestigations raise the possibility that incorporation in cell-free systems into 

4 Porter [14] has reported that heat denaturation of several proteins increases the number 
of e-amino groups of lysine that can react with acylating agents. 


microsomal protein may be at least to some extent nonenzymatic, it is quite 
possible that physiological mechanisms exist for the controlled transfer of the 
amino acid moieties of aminoacyl adenylates. 

Note Added in Proof 

Zioudrou, Fujii, and Fruton have recently described the synthesis of C 14 - 
tyrosinyl adenylate and C 14 -glycyltyrosinyl adenylate by a procedure similar to 
ours. They observed labeling of heated and unheated rat-liver mitochondria 
by these compounds and by their N-carbobenzoxy derivatives. They have also 
concluded that the labeling of the mitochondria is due to nonenzymatic acyla- 
tion (personal communication from Dr. J. S. Fruton; Proc. Natl. Acad. Sci. 
U. S., in press). 

Further studies in our laboratory indicate that enzymatically synthesized 
tryptophanyl adenylate can acylate microsomal preparations and also other pro- 
teins (e.g., bovine serum albumin, ovalbumin) ; these experiments were carried 
out with systems containing pancreatic tryptophan-activating enzyme, ATP, 
magnesium ions, and acceptor protein. Labeling of ribonucleic acid prepara- 
tions was also observed by such systems. The recent findings of Berg and 
Ofengand [15] and of Schweet, Bovard, Allen, and Glassman [16] are consistent 
with the possibility that specific binding sites for amino acids exist on soluble 
ribonucleic acid molecules. Whether such specific binding of amino acids to 
ribonucleic acid can be obtained with chemically synthesized aminoacyl adenyl- 
ates remains to be determined. The present studies emphasize the importance 
of isolating specific proteins in experiments on protein biosynthesis; the recent 
report of Bates, Craddock, and Simpson [17] on the incorporation of valine into 
mitochondrial cytochrome c appears to be a significant step in this direction. 


1. M. B. Hoagland, E. B. Keller, and G. D. Novelli, Proc. Natl. Acad. Sci. U. S., 
P. C. Zamecnik, /. Biol. Chem. 218, 345 42, 325 (1956). 

(1956). 10. M. Bergmann and L. Zervas, Ber., 

2. P. Berg, Federation Proc, 16, 152 67, 1192 (1932). 

(1957). 11. F. Lipmann and L. C. Turtle, /. Biol. 

3. M. Karasek, P. Castelfranco, P. R. Chem., 159, 21 (1945). 
Krishnaswamy, and A. Meister, paper 13 12. P. C. Zamecnik and E. G. Keller, 
of this volume. /. Biol. Chem., 209, 337 (1954). 

4. M. B. Hoagland, P. C. Zamecnik, and 13. A. Gierer and G. Schramm, Nature, 
M. L. Stephenson, Biochim. et Biophys. 177, 702 (1956). 

Acta, 24, 215 (1957). 14. R. R. Porter, Biochim. et Biophys. 

5. P. Berg, /. Biol. Chem., 222, 1015 Acta, 2, 105 (1948). 

(1956). 15. P. Berg and E. J. Ofengand, Proc. 

6. W. P. Jencks and F. Lipmann, /. Biol. Natl. Acad. Sci. U. S.. 44, 78 (1958). 
Chem., 225, 207 (1957). 16. R. S. Schweet, F. C. Bovard, E. 

7. H. S. Moyed and F. Lipmann, /. Bac- Allen, and E. Glassman, Proc. Natl. Acad, 
teriol, 73, 117 (1957). Sci. U. S., 44, 173 (1958). 

8. K. Moldave and A. Meister, /. Biol. 17. H. M. Bates, V. M. Craddock, and 
Chem., 229, 463 (1957). M. V. Simpson, /. Am. Chem. Soc, 80, 

9. J. A. DeMoss, S. M. Genuth, and 1000 (1958). 


The Synthesis of Hydroxyproline 
within Osteoblasts 



Medical Research Council Biophysics Research Unit 
W heatstone Laboratory , King's College, London 

Biochemical and morphological methods are being used to study the stages 
of synthesis of intercellular material in active collagen-producing tissue cultures. 
The direct oxidation of proline already bound in peptide linkage may be an 
important step in the sequence of the synthetic processes which lead to the 
formation of collagen protein [Stetten, 1949]. In previous work it has been 
found that appreciable amounts of protein-bound hydroxyproline were formed 
during the first 24 hr of culture before the appearance of characteristic collagen 
fibrils [Fitton Jackson and Smith, 1957]. Free C 14 -L-proline was also readily 
incorporated into the proteins of the growing tissue, and as much as 20 per 
cent was converted to protein-bound C 14 -hydroxyproline [Smith and Fitton 
Jackson, 1957]. 

Cell fractionation studies have been made on similar tissue cultures in an 
attempt to establish whether the site of incorporation of free proline into the 
proteins of the cell was the same as that of the formation of the protein-bound 
hydroxyproline. The cultures were grown in contact with C 14 -L-proline for 
various times and subsequently homogenized and subjected to differential cen- 
trifugation in 0.88 M sucrose solution. The amount of labeled proline incor- 
porated and converted to hydroxyproline in the six isolated fractions was meas- 
ured. Observations were made in parallel on the morphology of the whole 
cells and the various cellular fractions by means of the electron microscope. 

Chemical analyses demonstrated the consistent presence of protein-bound 
hydroxyproline in the fractions of larger particle size (3000 A). The results 



also showed that, under the influence of the cells, free C 1 l -L-proline was most 
rapidly incorporated into the proteins of the supernatant (obtained after final 
centrifugation), and into the "small-granule" fraction of the cytoplasm (sedi- 
mented at 105,000^) ; these fractions also contained the greatest amount of 
ribose. Subsequently part of the labeled proline appeared as hydroxyproline in 
a large-granule fraction (~3000 A particle size) as well as in fractions of larger 
particle size; for example, after 21 hours of culture growth followed by contact 
with radioactive proline for 1 hr it was found that for every 100 residues of 
protein-bound proline 16.9 residues were hydroxylated in this fraction. With 
longer contact times the radioactivity of the larger-granule fractions increased 

The significance of these results in relation to the synthesis of collagen pro- 
tein by the osteoblasts and in the mechanism of the formation of typical collagen 
fibrils was discussed. 


M. R. Stetten, 1949, /. Biol Chew., 181, R. H. Smith and S. Fitton Jackson, 1957, 

31. /. Biophys. Biochem. CytoL, 3, 692. 

S. Fitton Jackson and R. H. Smith, 1957, 
/. Biophys. Biochem. CytoL, 3, 679. 


Studies on Amino Acid Incorporation in 
Bacteria Using Ionizing Radiation 


Biophysics Department, Yale University a 

In recent years a growing body of evidence supports the idea that ribonucleo- 
protein particles form at least one of the sites of protein synthesis [1]. Evidence 
regarding this highly interesting system can be obtained by working with cell- 
free systems of purified particles, and undoubtedly such evidence is valuable and 
convincing. The drastic destruction of cell organization which is involved, how- 
ever, leaves the question that perhaps the whole nature of amino acid incor- 
poration is not being observed, but only the part that can survive the disrup- 
tion of order in the cell. A method of study having the great advantage that 
the cell is intact, or very nearly so, throughout the whole process is the use of 
ionizing radiation as a powerful local disruptive agent. Such radiation is able 
to penetrate all parts of the cell; it acts only at single, nearly isolated points, and 
is wholly without action elsewhere. Such high-energy spot probes, or line 
probes, can be employed against the organization of the cell, and, from the 
effect on any particular part of the synthetic process, deductions can be made 
regarding the process itself. Under good conditions, information can be ob- 
tained on the following points: (a) the approximate size (within a factor of 
2 in volume) of the region concerned with synthesis and sensitive to radiation; 
(b) the approximate thickness (within a factor of 2 or 3), and hence the ap- 
proximate length; and (c) the sensitivity of the synthetic region to radiation 
(within a factor of 2) . These data can be compared with the sizes, thicknesses, 
and lengths of cellular elements that could take part in the synthetic process. 
This comparison can then be used as one more piece of evidence regarding the 
nature of the process. A start along this line of investigation was made by 

1 Aided by a grant from the John A. Hartford Foundation. 



Hutchinson, Morowitz, and Kempner [2]. If sources of radiation are available, 
the method is, relatively speaking, technically easy, and therefore attractive. It 
is necessary, however, to be aware of the uncertainties of interpretation, to be 
sure that misleading deductions have not been made. The use of ionizing 
radiation to study cellular processes has been under intensive study in this 
laboratory for several years [2, 3, 4, 5, 6], and therefore a summary of the find- 
ings seems worth while, so that the validity of the conclusions can be estimated. 

The two major classes of biological macromolecules, proteins and nucleic 
acids, appear to be very sensitive to ionizing radiation. An enzyme molecule 
loses its activity if a cluster of ions forms anywhere inside the molecule; an 
antigenic protein loses its ability to combine with antibody if such a cluster 
forms in a volume somewhat smaller than that of the protein. DNA, as trans- 
forming principle, loses its function if such a cluster forms within a unit of 
about 300,000 molecular weight. If irradiations are carried out in solution, re- 
action products can move around, and they may have marked inactivating 
power. Studies by Hutchinson [7] on yeast cells in various conditions of mois- 
ture indicate that in the cell such reaction products carry their effectiveness 
over a distance of only 30 A. All these effects can be modified by factors of 
about 2 by several environmental conditions, notably oxygen tension and degree 
of aggregation between protein molecules. Thus, until the final sorting out of 
cause and effect is accomplished, the statistical interpretation of radiation effects 
must be considered to be approximate only. Even so, it is valuable as an aid 
in studying an important, inaccessible process. To give some idea of the va- 
lidity of the conclusions drawn we reproduce here a diagram, prepared by 
W. R. Guild, showing the relation between the "target molecular weight" de- 
rived from the statistical radiation analysis of radiation inactivation and the 
accepted molecular weights. Since the diagram is a log-log plot, it should not 
be viewed over-optimistically, but the reason can be seen for the claim that a 
factor of 2 is normally all that is involved as error. 

In order to gain the maximum information from irradiation studies, at least 
three, and preferably more, types of irradiation should be carried out: (1) Irra- 
diation by radiation sources very rich in fast electrons, as, for example, electrons 
themselves, of energy 0.5 Mev or more, or y-ray sources of energy in excess of 1 
Mev, where the secondary electrons due to Compton recoil and photoelectric ab- 
sorption have energies, in the main, in excess of 0.5 Mev. (2) Irradiation by 
heavy particles of variable rates of energy loss. Such particles have dense ioniza- 
tion, largely confined to tracks, and they give a different distribution of local en- 
ergy releases from fast electrons. Heavy particles of at least two energies should 
be used, to give a range of separation of energy releases. In our experiments we 
have employed cobalt 60 y radiation, deuterons of varied energies, and a par- 
ticles as bombarding agents. The results show that the uptake of methionine 
into the protein fraction (fraction insoluble in cold trichloroacetic acid) is re- 
tained unless very heavy irradiations are employed, and the sensitive region fits 
very well with a sphere of radius 130 A. For the uptake of proline into the 






1 1 

I0 6 

M 1/ 

1 y 1 1 





2.0 3 

^r • 


in 2 

/ 1 


1 1 


I0 3 10 I0 3 




Fig. 1. Comparison of accepted molecular weights and those determined by target 
analysis. We wish to thank W. R. Guild of our laboratory for permission to use this 

same fraction the same analysis cannot hold; instead, the best fit to the data 
is found for a long, thin, sensitive volume of radius roughly 11 A and length 
roughly 2.2 microns. Thus the methionine incorporation can very well be identi- 
fied with a process taking place in a microsomal particle, usually estimated as 
having a radius of 100 A [8], while the proline incorporation appears to impli- 
cate a whole chain, probably of nucleic acid, and may mean that the incorpora- 
tion of proline is dependent on the integrity of a system that binds together 
several microsomal particles. The experiments forming the basis for these con- 
clusions can now be described. 


Cultures of Escherichia coli B, maintained in this laboratory for a year, were 
grown with aeration at 37° C in an inorganic salt medium ("Minimal C 
medium" [9]) containing 5 g of glucose per liter. Aeration was stopped when 
the bacteria reached a concentration of approximately 5 X 10 s cells/ml as read 
in a Bausch and Lomb spectrophotometer. This is about the middle of the 


logarithmic-growth phase in this medium. The cells had a generation time 
of about 50 minutes. 

Cobalt 60 irradiation: Twenty-milliliter samples of bacteria were sealed in 
culture tubes and placed in a cobalt 60 source which delivered 380,000 r/hour. 
The temperature in the source was approximately 30° C. After irradiation, 
the tubes were placed unopened in a 37° C water bath, and allowed to rise to 
that temperature. 

Cyclotron bombardment : Bacterial cells were spun down in a Sorvall Model 
SS-1 centrifuge, and the pellet was resuspended at a concentration of 1 X 10 10 
cells/ml in minimal medium with no glucose added. One-tenth milliliter of 
this suspension was placed on fine-pore filters (Millipore Filter Corporation) 
and kept moist with a coarse filter backing containing distilled water. Irradia- 
tion was performed in air at 0° C [10]. After irradiation, the bacteria were 
resuspended in 10 ml of minimal medium with no glucose and allowed to 
come to 37° C. 

After irradiation and temperature equilibration, the bacterial suspension was 
added to an equal volume of minimal medium containing glucose which was 
aerated at 37° C. This incubation mixture contained 0.2 |jc of the radioisotope 
to be studied. L-Methionine-S 35 , 5.5 mc/g, and L-cystine-S 35 , 14.5 mc/g, were 
obtained from the Abbott Laboratories, Oak Ridge. L-Proline-C 14 , 8.9 mc/mM; 
L-leucine-C 14 , 7.95 mc/mM; and D-glucose-C 14 , 2.06 Mc/mg, were supplied by 
the Nuclear-Chicago Corporation. At various times during the incubation, 
2-ml samples were taken for the "whole cell" and "TCA-insoluble" fractions 
[2]. The filters on which these fractions were placed were then dried in air 
and counted under a thin window (less than 150 ug/cm 2 ) Geiger-Miiller coun- 
ter. Background was about 17 counts per minute. 

In order to provide the variety in distribution of ionization densities as men- 
tioned in the introduction, cyclotron bombardments were carried out with vari- 
ous thicknesses of aluminum absorber between the bacteria and the beam. 
Since the beam has a definite range, the amount of inactivation it produced 
varied with the absorbers, falling to zero when the range in absorber was ex- 
ceeded, so that no deuterons hit the bacteria. From these curves the equiva- 
lent absorption of each bacterial preparation could be measured and thus the 
effective energy of the deuterons hitting them estimated. Since the density of 
ionization varies with energy in a known way [5], the appropriate value for 
each bombardment can be determined. 


The effect of irradiation with cobalt 60 on the uptake of L-methionine is 
shown in figure 2. Three sets of curves are presented. The first, on the left, 
applies to the uptake of unirradiated E. coli. The counts per minute are plotted 
against time for two samplings: the whole cell, as represented by the upper 
line, and the fraction insoluble in cold trichloroacetic acid (TCA). The differ- 


o.) Unirradiated b.) 253,000 r 


760,000 r 

15 5 10 15 




Fig. 2. The effect of y radiation on methionine incorporation in E. coli. Except where 
otherwise marked in these figures, solid circles correspond to the label in the intact cell, 
and crosses refer to the cold-TCA-insoluble fraction. The difference between these two 
curves, the "pool," is seen to decrease with increasing dose of radiation. Background was 
20 cpm and has been subtracted. 

o) Unirradiated 

b.) 3 1 7, 000 r 

c.) 950,000 r 


15 5 10 




15 min 

Fig. 3. The effect of y radiation on proline incorporation. The cold-TCA-soluble "pool" 
rises in magnitude and then falls off with dose. 



ence between these two, following Britten, Roberts, and French [11] and Cohen 
and Rickenberg [12], is designated as the metabolic pool. In figure 2b, the 
same process is employed on bacteria that have received 253,000 roentgens of 
cobalt 60 irradiation. It can be seen that the total amount of radioactivity in- 
corporated in the TCA-insoluble fraction is reduced; the rate of incorporation 
is also less. The amount taken up by the whole cells is less, though not quite 
to the same extent as the reduction in the TCA-insoluble fraction. In figure 2c, 
still higher irradiation was employed, and the reductions in both fractions are 
still more apparent. 

Similar data are shown in figure 3 for the uptake of proline. There is a 
marked difference from the effects with methionine in that more irradiation is 
necessary to reduce the amount incorporated into any fraction, about twice the 
dose in the TCA-insoluble fraction and nearly 10 times in the whole-cell frac- 
tion. Because of the disparity in these effects it can be seen that the pool actu- 
ally rises after bombardment. 

For these two amino acids, with the exception of the proline pool, the amount 
of activity remaining seems to be a diminishing exponential function of the 
dose. The data scatter somewhat, but we have no real evidence in favor of a 
multiple-hit type of process, where the activity remains nearly constant and 
then rapidly falls. The per cent remaining activity for methionine incorpora- 
tion is plotted against dose in figure 4. For proline incorporation the effect 
of radiation is definitely less in the TCA-insoluble fraction than for methionine, 
and very markedly less in the whole-cell case. Such dose effect curves can be 
analyzed statistically in terms of an inactivation volume V, which is the sensi- 
tive region that must escape an ionization in order to retain the effect being 
measured. If the ionizations occur at a number / per unit volume, then the 
average number of ionizations occurring in the sensitive region is IV, and by 
the Poisson relation the probability that the region will escape is e~ Iv . Thus 
the natural logarithm of the ratio remaining to that in the unirradiated con- 
trol should be —IV. For 37 per cent remaining, the value of IV is unity. 

Table 1 summarizes the results of cobalt 60 studies. The first column gives 
the 37 per cent dose found from the survival curves. The second column gives 
the corresponding number of primary ionizations per cubic centimeter for such 
a dose. Column three, the sensitive volume, is the reciprocal of the value in 

TABLE 1. Summary of Incorporation Studies in the TCA-insoluble Fraction of E. coli 

Irradiated with Cobalt 60 y Rays 








Dose, r 

per cm 3 

cm 3 

Radius, A 

Leucine C 14 

0.36X10 6 

1.8 XlO 17 

5.6X10- 18 


Cystine S 35 


2.3 XlO 17 

4.3 XlO- 18 


Methionine S 35 

0.20 X10 6 

l.OxlO 17 

10.0 XlO" 18 


Proline C 14 


2.3 XlO 17 

4.3 XlO" 18 


Glucose C 14 


4.3 XlO 17 

2.3 XlO- 18 








^ 30 


° 10 

Whole Cell 




0.2 0.4 0.6 0.8 




Fig. 4. Survival of methionine incorporation with Y-ray dose. The ordinate is on a 
logarithmic scale. Pool fraction shows similar survival. 

the second column. The last column, the "equivalent spherical radius," is the 
value calculated assuming that the sensitive volume is the shape of a sphere; 
it is listed for comparative purposes only. 

Cyclotron irradiations produced similar effects. The data scatter even more, 
probably owing to the difficulty in securing irradiations involving the same 
time of exposure on the Millipore filter. There is quite clear evidence of a 
reduction in activity, as can be seen from figure 5, where the uptake of methi- 
onine is shown for control and two irradiated points. Uptake curves of the 
sulfur-labeled amino acids do not extrapolate to zero counts per minute at zero 
time, owing to adsorption of the label, and this correction is deducted in plot- 
ting survival curves. Figure 6 shows analogous curves for proline incorpora- 
tion, also after deuteron irradiation. These experiments were performed with- 
out added aluminum absorber in the cyclotron beam. 

Similar studies were carried out with various absorbing foils, and, from the 
uptake curves, curves relating the per cent uptake remaining to the number of 
deuterons per square centimeter used to irradiate were drawn for each foil 
thickness. In each case, the 37 per cent dose (in deuterons per square centi- 



oJ Unirrodiated 





10 2 10 2 

b.) 11.3 x 10 deuterons/cm c.) 33 x 10 deuterons / cm 






6 5 10 




15 min. 

Fig. 5. The effect of deuteron bombardment on methionine incorporation. Count rate 
at zero time is due to adsorption of the label. 





a.) Unirrac /•'/ b.) 8.5 x lo'° deuterons/cm 2 c.) 23.3 x I0 10 deuterons /cm 2 

•300 ^ _ 300r 

10 15 5 10 15 


10 15 min. 

Fig. 6. The effect of deuteron bombardment on proline incorporation. Here the pool 
decreases in absolute magnitude with dose. 



meter) was found. The statistical analysis is similar. If we assume that a 
deuteron is able to inactivate as long as it passes through the sensitive region, 
then the area of the sensitive region is all that matters. On that basis, if D is 
the number of deuterons per square centimeter and S is the area of the sensi- 
tive unit concerned with amino acid uptake, we see that the average number 
of deuterons per target is SD, and once again the probability of escape is e~ SD . 
Therefore at the value of D for 37 per cent remaining, SD = \. S is referred to 
as the "cross section." It must be remembered that a deuteron may not be per- 
fectly efficient in producing inactivation, and in particular, if the ionizations 
have a finite separation, then a thin target can be "straddled." This effect is 
apparently present for proline uptake. 

In figure 7, the cross section is plotted against the air equivalent of the foil 
thickness for the TCA-insoluble fraction containing C 14 proline. The maxi- 
mum of this curve is near 26 X 10" 12 cm 2 at a value of 5.5 cm air equivalent of 
foil. From the residual range of the beam, the value of the linear energy trans- 
fer (LET) at this cross section is found to be 400 electron volts per 100 A. 



1 2 





10 - 


- xl0"' 2 cm 2 






< i 


i i iV 





16 cm. 


Fig. 7. Measured cross section for proline incorporation into the TCA-insoluble portion 
of E. coli after deuteron bombardment as a function of added aluminum in the cyclotron 
beam; 12.5 cm is seen to be the range of the deuteron beam in air. 

Incorporation studies after a-particle bombardment are indicated in figure 8 
for proline. The survival curves for such studies were reasonably exponential, 
with a 37 per cent dose for the TCA-insoluble portion labeled with proline 
of about 2.7 X 10 10 a particles per cm 2 . 

The results of y-ray, deuteron, and a-particle irradiations are combined in 
the LET plot [5] in figure 9. The two relations, in terms of volume V and 
area S, reduce to the same expression when the density of ionization, or linear 
energy transfer, is low. Under such circumstances, if / is the number of pri- 
mary ionizations per unit length, l — Di, so that we obtain for the probability 
of escape e~ DiV , or S=iV. Thus the value of S close to the origin can be found 










15 5 10 15 5 10 I5min 


Fig. 8. The effect of a-particle irradiation on proline incorporation. The pool decreases 
with increasing dose. 




a so 


CI4 Proline 

TCA Insoluble Fraction 

S35 Methionine 

400 800 1200 



1600 e.v./IOOA 

Fig. 9. Measured cross sections for proline (crosses, dashed line) and methionine (solid 
circles, solid line) incorporation into the cold-TCA-insoluble fraction as a function of the 
rate of energy loss (LET). Slopes at the origin are calculated from y-ray experiments as 
explained in the text. 

by substituting the product of i and V as already found from irradiation with 
Y rays. Such a relation only applies close to the origin, but can be used to de- 
termine the initial slope. To a reasonable approximation, i is the energy loss 
per centimeter, in electron volts, divided by 110. 


The slopes at the origin (dashed for proline, solid for methionine) are cal- 
culated from the sensitive volumes found with y radiation. The points near 
400 ev/100 A are from the deuteron experiments as described above. The LET 
for a particles in our experimental arrangement was 1500 ev/100 A, and the 
cross sections found by this method are shown at that LET value. 


The analysis of the methionine-cold-TCA-insoluble fraction data indicates a 
spherical target. The radius calculated from the inactivation volume is 130 A; 
the radius calculated from the cross-sectional area is 127 A. The agreement is 
well within the uncertainties of this method of study. 

The proline data are somewhat more complex. The upward concavity near 
the origin indicates a rather complex response to radiation. It can be explained 
most simply by supposing that closely grouped ionizations are more effective 
than single ionizations, and the rapid change of slope suggests that this multi- 
plicity is small, probably a double ionization requirement [5]. There is no 
agreement between the area and volume determinations if a spherical target 
is assumed, no matter which ionization requirement (one, two, three, or four) 
is used. The only simple model that fits the data is a long, thin rod [10]. For 
a double ionization requirement, the volume and area determinations lead to 
a model of a rod of 11 A radius and 2.2 u length. The analysis based upon 
the model of a thin plate leads to substantially the same result, an extremely 
long plate of small cross-sectional area. 

Preliminary experiments with deuteron-bombarded E. coli cells indicate that 
the incorporation of glucose into the TCA-insoluble portion of these cells re- 
quires the integrity of a sphere of approximately 80 A radius. Since, in these 
experiments, glucose was the single carbon source available to the organisms, 
it seems reasonable to conclude that amino acids are produced in the cells in 
spheres of this size. Since the doses required to stop glucose incorporation into 
the cold-TCA-insoluble fraction are large compared with the doses required to 
stop exogenous methionine from being incorporated into the same fraction, it 
would appear that the formation of proteins from amino acids is not directly 
related to the incorporation of glucose. Further work on this very interesting 
result is clearly necessary. 

If the two radiation-sensitive spheres found for glucose and methionine in- 
corporation were assumed to have a density of 1.3, and to be sedimenting in a 
solution of viscosity 1.5 centipoises, then from 

d 2 = lSv)sV/(l-Vp) 

it can be calculated that the spheres associated with glucose incorporation 
(160 A diameter) should have a sedimentation constant of the order of 25 S, 
while the larger, methionine-associated sphere, 260 A in diameter, should have 
a sedimentation constant of about 77 S. Since particles of such size have been 


reported in extracts of E. coli [13], our work can be interpreted as support for 
the functional character of these particles. 

If the incorporation of exogenous amino acids into a TCA-insoluble form 
corresponds to protein synthesis, then the conclusion that at least two different 
amino acids are incorporated by two different structures within the cell logically 
leads to a model in which protein synthesis takes place in at least two steps. 
If the two units found by radiation were simultaneously required and both 
essential for protein synthesis, then methionine and proline would be expected 
to give the same radiation targets. The fact that targets of quite different char- 
acter are found indicates that there must be some difference between the two. 
It is clearly interesting to continue the studies to see whether groupings of 
types of target exist. 

Actually, that an ionization in a cellular unit which is probably composed 
of a dozen or so subunits should destroy its function is surprising. It seems 
certain that the effect of radiation on one part can precipitate a disruptive effect 
on the whole. This suggestion has already been made by Billen and Volkin 
[14]. Our data tend to support their conclusions. 


We wish to thank Messers P. Hanawalt, P. Schambra, and J. Lowry for 
assistance in running the cyclotron. We are also grateful to Drs. R. Roberts, 
D. Cowie, R. Britten, and E. Bolton for advice in rapid filtration procedures 
as well as for many stimulating discussions. 


Ionizing radiations of different character, y rays, deuterons, and a particles, 
were used to determine radiation targets for the incorporation of methionine, 
proline, and glucose into the cold-TCA-insoluble fraction of Escherichia coli. 
Spherical targets were found for methionine (260 A diameter) and glucose 
(160 A diameter) incorporation. The target for proline incorporation is a long, 
thin rod, 22 A in diameter and 2.2 u long. That the units associated with 
methionine and glucose correspond closely in size to microsomal particles found 
in cell debris is therefore evidence for the functional importance of these 

The size and shape of the proline-associated incorporation target appear to 
be those of a nucleic acid unit, although it is not possible to distinguish be- 
tween RNA and DNA by this method. 

The fact that the target determined for glucose incorporation is smaller than 
that found for the incorporation of the exogenous amino acid methionine indi- 
cates that glucose is not directly connected with the binding of incorporated 
amino acids into bacterial protein. 




LP. C. Zamecnik and E. B. Keller, 
/. Biol. Chem., 209, 337 (1954); J. W. 
Littlefield, E. B. Keller, J. Gross, and P. C. 
Zamecnik, /. Biol. Chem., 217, 111 (1955). 

2. F. Hutchinson, H. Morowitz, and E. 
Kempner, Science, 126, 310 (1957). 

3. E. C. Pollard, Advances in Biol, and 
Med. Phys.,3, 153 (1953). 

4. Conference on Ionizing Radiation 
and the Cell, L. F. Nims, Chairman, Ann. 
N. Y. Acad. Sci., 59, 467-664 (1955). 

5. E. C. Pollard, W. R. Guild, F. Hutch- 
inson, and R. B. Setlow, Progr. in Biophys. 
and Biophys. Chem., 5, 72 (1955). 

6. W. R. Guild and F. DeFilippes, Bio- 
chim. et Biophys. Acta, 26, 241 (1957). 

7. F. Hutchinson, Radiation Research, 7 , 
473 (1957). 

8. H. K. Schachman, A. B. Pardee, and 
R. Y. Stanier, Arch. Biochem. Biophys., 38, 
245 (1952). 

9. R. B. Roberts, D. B. Cowie, P. H. 
Abelson, E. T. Bolton, and R. J. Britten, 
Studies of Biosynthesis in Escherichia coli, 
Carnegie Inst. Wash. Publ. 607, Washing- 
ton, D. C, 1955. 

10. E. C. Pollard, J. Setlow, and E. 
Watts, Radiation Research, 8, 77 (1958). 

11. R. J. Britten, R. B. Roberts, and E. F. 
French, Proc. Natl. Acad. Sci. U. S., 41, 
863 (1955). 

12. G. N. Cohen and H. C. Rickenberg, 
Compt. rend., 240, 2086 (1955). 

13. Other papers this volume. 

14. D. Billen and E. Volkin, /. Bac- 
teriol, 67, 191 (1954). 


The Effect of X Rays on the Incorporation 
of Phosphorus and Sulfur into Escherichia colt 


Biophysics Department^ Yale University 

In the previous paper of this volume an account was given of preliminary 
studies on the effect of ionizing radiation on the incorporation of amino acids 
and glucose into a fraction of the bacterial cell that is not soluble in trichloro- 
acetic acid. It is hoped that a continuation of such probing into the synthetic 
processes in the cell by radiation will give some information on the nature of 
the processes themselves. Though the subject of keenest interest at the moment 
is undoubtedly the fate of an individual metabolite, and the way in which it 
becomes incorporated, it seemed to be important to know something of the 
way in which radiation affects the uptake of two much more generally utilized 
elements of the growth medium, phosphate and sulfate. The studies to be de- 
scribed were originally meant simply as a means of gaining general informa- 
tion about the effect of radiation on the cell metabolism, to ensure that no seri- 
ous discrepancies existed between the findings with one amino acid and the 
whole metabolic process of the cell. During the studies some results appeared 
that seem to indicate rather remarkable radiation sensitivities and are, more- 
over, of interest in themselves. For instance, the ability of the cell to incor- 
porate phosphate is remarkably sensitive to X radiation; the sulfate-incorpora- 
tion ability is also highly sensitive, though less so; and sulfide incorporation is 
definitely still less sensitive. The hypothesis that suggests itself is that the 
phosphorus incorporation is determined by relatively large and sensitive units, 
and that when these are damaged by radiation there is a proportionally smaller 
phosphorus uptake. For sulfur uptake the results are in accord with the idea 
that there is a considerable synthetic chain which presents sensitivity at a num- 

1 Aided by a grant from the United States Public Health Service. 



ber of loci. The results are only preliminary, but they do contribute to an 
emerging picture of cellular processes which is worth some consideration. 

Closely related to this work is a series of studies by Billen and Volkin [1], 
Billen and Lichstein [2], Billen, Stapleton, and Hollaender [3], and Billen [4] 
on the effect of X rays on several factors in E. coli. Their results will be 
discussed later. Some work by Labaw, Mosley, and Wyckoff [9] is also of in- 
terest, though directed at bacteriophage development. 


For these experiments the simplest and most available source of ionizing 
radiation, X rays, was used. X rays produce ionizations which are distributed 
along the tracks of secondary electrons, light particles that ionize relatively 
sparsely. In our arrangements, where 250-kv X rays filtered through 1 mm of 
aluminum were used, the average secondary-electron energy is in the neigh- 
borhood of 50 kv, and so produces ionizations which are generally separated 
by distances of over 1000 A. Since the electrons scatter readily, the net effect 
is to produce local energy releases, which average 110 ev each, at random 
throughout the bacterium. The number per unit volume depends on the dose 
of radiation, and for the purposes of this paper will be taken as 5 X lO^/cc/r, 
which assumes a density of a little over unity for the bacterium. 

Many previous studies, quoted in the preceding paper of this volume, have 
shown that the energy releases can remove biological potency from quite large 
biological molecules. In proteins it is thought that the reason is the migration 
of the positive charge left by ionization, along a covalently bonded structure 
until it reaches a weak point, where chemical action by water or oxygen can 
cause a destructive chemical change. Then ionization anywhere in the cova- 
lently bonded structure is destructive. For nucleic acids the best evidence sug- 
gests that either a break in the chain can result, or a cross linking of two chains 
can take place, both drastic events. In any event, the philosophy of these ex- 
periments is to estimate the chance of complete escape from any ionization at 
all, arguing that the chance of being ionized and escaping any effect is statisti- 
cally too low to be worth considering. Such estimates are, of course, tentative 
to some degree. Nevertheless, the sum of 10 years' work in our laboratory, to- 
gether with the findings of many others, renders them plausible. 

The structure that is being bombarded by these random energy releases is 
shown in the section electron micrograph in figure 1. The section is of two 
Escherichia coli bacteria, one of which has passed into a rather filamentous 
form, and the other of which is normal. The sectioning and microscopy are 
entirely due to L. Caro of our laboratory, and the illustration is presented by 
his permission. It shows clearly the presence of an inner region, almost cer- 
tainly the locale of the DNA, as has been shown by Caro, Van Tubergen, and 
Forro [5], and an outer region which contains microsomal particles, or "ribo- 
somes," and which is presumed to be the region of protein synthesis. Figure 2 



Fig. 1. An electron micrograph of a thin section of E. coli, taken by L. Caro. The 
ribosome granules can be seen in the body of the cell. The inner region, presumed to 
contain DNA, can be seen, and the structure of the cell wall is also apparent. 

is a schematic drawing of the escape process, in which one of the ribosomes 
is indicated as present in a region 1000 A square and 100 A deep, with the 
appropriate energy releases accompanying it. On the left is the effect produced 
by 10,000 r, and the microsomal particle is shown to scale with the single ioniza- 
tion which is to be expected accompanying it. According to the studies of 
Hutchinson [6], in a cell containing the normal proportion of water there can 


\0,000r \00,000r 














Fig. 2. A schematic representation of the statistical analysis of radiation damage. It is 
supposed that to escape damage a microsomal particle must escape an ionization in its 
structure, or even the diffusion of an active agent from ionization to its structure. The 
probability of escape can be seen to be clearly dependent on the particle size and the amount 
of ionization. 

be a migration by diffusion of the energy release to the extent of 30 A. This 
little spread of energy has been indicated by a dotted line around the ioniza- 
tion. It can be seen that the chance of escape at 10,000 r is very good, the cal- 
culated figure being 98 per cent. In the second region the microsomal particle 
has been given an increase of 30 A in radius to allow for the possibility of 
migration of energy, and now each ionization has to be considered as a single 
dot. The chance of escape is now less but still high; it is estimated at 75 per 
cent. With 1,000,000 r it is only 8 per cent, and with 3,000,000 r the chance of 
escape is very remote indeed, and it is expressed by the unlikely figure of 0.07 
per cent. 

The experimental method, therefore, is to expose bacterial cultures held in 
a static condition by being chilled to 4° C to X rays, to permit them to warm 
up, and then to add labeled phosphate or sulfate to them and observe the up- 
take of radioactivity as compared with the uptake of the unirradiated prepara- 
tion. The ratio of the two is then compared with the chance of escape, and 
deductions regarding the size of the particle are made from this probability. 


The experiments with phosphate are complicated by the fact that very com- 
monly phosphate buffer is used in minimal media. This means that a large 
excess of cold phosphate is present, rendering uptake somewhat difficult. 
Accordingly the procedure we evolved was to grow the bacteria overnight in 
the C minimal medium as described by Roberts, Abelson, Bolton, Britten, and 
Cowie [7] and then transfer for a 90-minute growth period to Tris buffered 
medium containing 80 mg of phosphate/1. The 90 minutes permitted the bac- 
teria to establish themselves satisfactorily in the new medium; they were 
then chilled and irradiated with a 250-kv X-ray machine at approximately 
3000 r/minute. Subsequent to irradiation both the control and the irradiated 
sample were warmed in a 37° bath and an appropriate number of counts of 
radioactive phosphate was added to each one. The bacteria were sampled at 



different intervals of time by means of a small syringe and filtered through an 
S and S filter. The filters, after being stored to dry, were counted in the 
ordinary way. 

For sulfate the same procedure was adopted because it was thought worth 
while to compare phosphate and sulfate uptake under similar conditions. For 
the uptake of sulfides we encountered considerable difficulty, as was pointed 
out by Cowie, in that the sulfide very readily formed sulfate and does not stay 
in its original form. We overcame this to some extent, though not always con- 
sistently, by having an excess of cold sulfide present and by working as rapidly 
as was consistent with the experiment. When fresh radioactive sulfide and 
fresh sodium sulfide were used the data indicated a very clear difference from 
those obtained with sulfate and could also be made to give consistent results 
though not quite to the same degree as sulfate uptake. 


The form of an uptake curve for phosphate is shown in figure 3. The ex- 
periments differ from those of the previous paper in that there is no limitation 
in the amount of tracer and so the bacteria can continue to incorporate phos- 
phorus throughout a whole period of 2 hours. It can be seen that for the con- 
trol and the slightly irradiated points there is an exponential uptake of P 32 but 
that at the higher doses the uptake becomes linear. By comparing the early 
slopes of the lines a survival ratio can be obtained which can then be plotted 






1200 -- 

800- - 

400" " 



20 40 60 80 100 



Fig. 3. The results of one individual experiment on phosphate uptake. The control and 
least-irradiated cultures show exponential uptake; the more heavily irradiated cultures show 
reduced linear uptake. 



against the dose. The change from logarithmic to linear increase is significant 
and probably means, as pointed out by Cohen [8], that some unit is being 
formed in the irradiated cell which is not suitable for further growth. The 
cell is then continuing to operate on its existing material and not making fresh. 

With radioactive sulfate the curves shown in figure 4 were obtained. These 
are the results of a single run, and they are very similar to the uptake curve 
for phosphate except that the sensitivity to ionizing radiation is not quite so 
great. In the course of study we noticed that the initial part of the curve did 
not seem to be as greatly affected by ionizing radiation as later stages, and 
accordingly a deliberate attempt to study this was made, with the results shown 
in figure 5. Much higher doses were employed, and it is apparent that the 
effect of quite high doses on sulfate uptake is to cause a breaking away from 
an initial line which occurs progressively earlier and earlier in the uptake proc- 
ess. The final slope of the uptake is diminished as found previously, but the 
breaking-away point seems to come later and later with lower and lower dose. 

This again seems to indicate that the effect of radiation is on some mecha- 
nism in the cell that prevents future development of means for further growth. 












/ 5,300 r 

X/ 43,000 r 


60 80 100 


120 WIN. 

Fig. 4. The results of one uptake experiment on sulfate. For low irradiations the ex- 
ponential character remains, but at higher doses the uptake becomes linear. The uptake is 
less sensitive than that of P0 4 . 






1000- - 




39,000 r 

x„— — 78,000 r 

_o Ill.OOOr 

60 80 100 




140 MIN. 

Fig. 5. The initial uptake of sulfate at higher doses. For 10 minutes, even at a dose of 
111,000 r, the uptake is normal. Thereafter, the rate of uptake breaks away from normal, 
at earlier times for higher doses. The implication is that some synthesis can go on with 
the original synthetic apparatus but that adequate provision for future synthesis is not made 
in irradiated cells. 

The initial processes in the cell seem to be capable of continuing, but they run 
out and the damage to the cell then becomes apparent. 

The incorporation of sulfide was studied in much the same way. The results 
again look very similar to those found for sulfate except that still higher doses 
were necessary to produce an effect, and also there was some indication that 
the inactivation instead of being linear was curved. This might mean that 
more than one process is involved in sulfide uptake. 

By way of check experiments, the number of colonies produced after irradi- 
ation was also studied and the effect of radiation on the optical density was 
measured. The optical density is that after a standard period of growth, usu- 
ally about 80 minutes. 

The results of all these experiments can be seen in figure 6, where curves for 
per cent remaining versus radiation dose in roentgens are plotted. It can be 
seen that by far the most sensitive factor is the formation of colonies and that 
this is followed, in order, by phosphate and sulfate uptake, optical density, 
and sulfide uptake. Lastly, the effect on the methionine uptake is very small 
indeed and can hardly be plotted on the graph. Analyzed in terms of the 
probability of escape as mentioned earlier, the data are presented in table 1. 






■ l 






■~" 50 lOOxlOOOr 


Fig. 6. Logarithmic dose response plots for colony counts, optical density, and phos- 
phate, sulfate, sulfide, and methionine uptakes. The relative sensitivities can be analyzed 
in terms of sensitive volumes, which are given in table 1. 

The most surprising result is the great sensitivity of phosphorus uptake. It 
is difficult to see how the relatively simple processes involving phosphate can 
be involved with such large target units. A repeat of the uptake, using C mini- 
mal medium, has been made, with results still indicating a great sensitivity, 
though perhaps not as great as in Tris. We are at a loss to explain the finding, 
and rather than speculate about the control of phosphorus uptake by large units 
of DNA we prefer to continue studies aimed at finding the nature of the dam- 
aged products. It should be pointed out that, with all three tracers, phosphate, 







37% Dose, r 

Volume, cc 

of Cell 

Colony formation 


4.X10- 16 


Optical density 

Initial slope 


2.4 xlO" 16 

1.5 XlO- 4 

Final slope 


4.7 xl0- 1T 

3 XlO" 5 

Phosphate uptake 


1.4xl0^ 1G 

9 XlO" 5 

Sulfate uptake 


8.7 XlO- 17 

6 XlO" 5 

Sulfide uptake 


5xl0- 17 

3 XlO" 5 

Methionine uptake 


io- iT 

6.5 XlO" 6 

sulfate, and sulfide, the initial uptake is unaltered. This finding was reported 
previously by Billen and his associates [1, 2, 3, 4], and we confirm it. 

For the sulfate and sulfide data a simple explanation can be advanced. It 
is best seen diagrammatically as in figure 7. We suppose that the synthesis of 
protein is determined by the presence of intact units of the size stated in the 
previous paper for methionine. However, the protein made can be essential 
for the enzymatic reduction of sulfate or combination of reduced sulfur with 
other elements to form intermediates, finally ending in methionine, cysteine, or 
glutathione. Each enzyme is a protein and probably requires methionine or 
cysteine, i.e., sulfur in the right form. The entire uptake of sulfur is thus sensi- 
tive in three ways: first as final incorporation into protein, second as requiring 
the availability of a suitable enzyme for a needed process, third in terms of the 
synthesis of such enzymes. The sensitivity of methionine incorporation then 
requires only the functioning of one ribosome. Sulfate will not be incorporated, 
however, if enzymes are inactivated (experience shows this inactivation to be 
relatively difficult, as each enzyme molecule is small), or if enzyme synthesis 
is stopped, or if protein synthesis is stopped. Thus a much larger inactivation 
volume for sulfate uptake can be predicted. From the ratio of the inactivation 
volume for sulfate uptake, after the initial stage is over, to that for methionine 
uptake, the number of steps involved can be estimated. From table 1 the steps 
from sulfate to methionine are 8.7, and from sulfide to methionine 5. The 
process is shown schematically in figure 7. We make no claim for correct inter- 
mediary biochemistry. If the explanation we propose is right it could be checked 
against known intermediary processes by means of competition techniques. 
This check is planned for future work. 

One question naturally asked concerns any change in the fractions in the 
cell after irradiation. We studied the four fractions: alcohol-soluble, cold-TCA- 
soluble, hot-TCA-soluble, and residue. The results are shown in table 2. We 
found the relative proportion of S 35 in the hot-TCA-soluble and residue frac- 
tions to depend somewhat on the time and temperature of exposure to hot TCA, 
and we do not regard the variations in the hot-TCA fraction as significant. 
Possibly the fall in the alcohol-soluble fraction after heavy irradiation in sulfate 




sq,gso 3 

<\aCT. METH S0 3 


R 2 SH 

R 2 SH^ 


s 2 o 3 



' f X ' 

;V > ^> 



Fig. 7. A schematic diagram to indicate the reason for the high sensitivity of sulfate 
uptake. A series of reduction and combination steps, requiring enzymes (indicated as 
rectangles), has to be complete before the activated amino acid can be incorporated. Each 
enzyme is made by a microsomal particle vulnerable to radiation. Hence the sensitivity for 
a nine-step process is nine times higher than for a single step. 

medium is a real effect. Otherwise the major conclusion is that no change in 
the broad distribution is observed. 

That the initial uptake is unchanged, but reduced uptake occurs sooner 
for more heavily irradiated cells, argues that a considerable amount of incor- 
poration can go with the damaged synthetic apparatus, but that an unbalance 
occurs as the cell develops. If the damaged ribosomes were making enzymati- 

TABLE 2. Cell Fractionation under Various Conditions 


Alcohol-soluble (lipids, protein) 

Cold-TCA-soluble (transient inter- 
mediates, glutathione) 

Hot-TCA-soluble (nucleic acid 

Residue (protein) 




50 kr 


300,000 r 























cally impotent protein, such would be the case. The linear rather than expo- 
nential uptake in damaged cells tends to argue in favor of this hypothesis. We 
intend to study specific cell fractions as a check. 


We wish to thank Stuart Hauser for assistance in the later stages, particularly 
for the data for figure 5. Discussions with Ellis Kempner have been most 


The uptake of P 32 phosphate and S 35 sulfate and sulfide by E. coli as in- 
fluenced by X rays has been studied. The initial uptake is not greatly changed 
by doses up to 100,000 r, but the rate of uptake is radically changed at later 
times. Expressing the ratio of the final rate of uptake to that of unirradiated 
controls in terms of dose we find a logarithmic relation. Analysis in terms of 
a sensitive volume gives large sensitivities as follows: 

Colony formation (for comparison) 

P0 4 uptake 

S0 4 uptake 

S uptake 

Methionine uptake (for comparison) 

4 X 10- 16 cc 

1.4 X10- 16 

9.5 xlO- 17 
5xl0- 17 

io- 17 

The surprisingly large sensitivity for phosphate may mean that intact DNA 
controls phosphorus uptake. The figures for sulfate, sulfide, and methionine 
can be explained in terms of a synthetic chain of microsomal particles involv- 
ing nine steps from S0 4 to protein and five from sulfide to protein. 


1. D. Billen and E. Volkin, Effect of 
X-rays on the macromolecular organiza- 
tion of E. coli B/r, /. Barter iol, 67, 191— 
197 (1954). 

2. D. Billen and H. C. Lichstein, Effect 
of X-radiation on the adaptive formation 
of formic hydrogenlyase, /. BacterioL, 63, 
553-555 (1952). 

3. D. Billen, G. E. Stapleton, and A. 
Hollaender, The effect of X-radiation on 
the respiration of E. coli, f. BacterioL, 65, 
131 (1953). 

4. D. Billen, Modification of the release 
of cellular constituents by irradiated E. 
coli, Arch. Biochem. Biophys., 67, 333— 
340 (1957). 

5. L. Caro, R. T. Van Tubergen, and F. 
Forro, Jr., Radioautography of sectioned 

bacteria, Biophysical Society Program Ab- 
stract R 12, 1958. 

6. F. Hutchinson, The distance that a 
radical formed by ionizing radiation can 
diffuse in a yeast cell, Radiation Re- 
search, 7, 473-483 (1957). 

7. R. B. Roberts, P. H. Abelson, D. B. 
Cowie, E. T. Bolton, and R. J. Britten, 
Studies of Biosynthesis in Escherichia coli, 
Carnegie Inst. Wash. Publ. 607, Washing- 
ton, D. C, 1955. 

8. G. N. Cohen, Synthase de proteines 
anormales chez Escherichia coli K-12 cul- 
tive en presence de L-valine, Ann. inst. 
Pasteur, 94, 15-30 (1958). 

9. L. W. Labaw, V. M. Mosley, and 
R. W. Wyckoff, Development of bacterio- 
phage in x-ray inactivated bacteria, /. 
BacterioL, 65, 330-336 (1953). 


Statistical Relations in the Amino Acid Order 
of Escherichia coli Protein 1 


Biophysics Department, Yale University 

Since proteins consist largely of chains of amino acids, a formal analogy is 
suggested between protein structure and written language, which consists of 
chains of letters. The analogy breaks down when we come to consider func- 
tion. For the function (meaning) of language is completely determined by the 
sequence of letters, whereas the function of proteins depends on the secondary 
structure (coiling) was well as on the sequence. (In poetry, secondary struc- 
ture is extremely important.) 

One characteristic of both proteins and language is the nonrandom frequency 
of occurrence of letters. Thus if we examine a long passage written in English 
we get a rank frequency distribution represented by figure 1, which also shows 
the rank frequency distribution of amino acids in Escherichia coli protein. Fur- 
ther examination of language reveals certain high-frequency pairs, triplets, and 
higher groupings of letters [1, 2]. 

A further feature shows up on inspection of amino acid composition of pro- 
tein. The nonrandom distribution of amino acids that is apparent in over-all 
collections of proteins arises from a similar nonrandomness in individual pro- 
teins. There are notable exceptions to these relations, however, particularly in 
structural proteins like silk and collagen. 

A further question suggests itself. In proteins, are there pairs, triplets, and 
higher amino acid sequences that occur with unexpectedly high or low fre- 
quencies ? That is, are there laws, similar to the laws of language, governing 
the ordering of amino acids in peptide chains ? These statistical relations could 
be the result of thermodynamic stability, of evolutionary selection, or of the 

1 This research aided by a grant from the United States Public Health Service. 

































e t aoni s r 

s n 





-c c 



5"s >*v. "> Q. 


I dcupfmwybgvkqx j z 

I lnn.-.^ 


Fig. 1. 

mode of synthesis of protein. Regardless of their cause, knowledge of them 
would place severe restrictions on our theories of synthetic cellular processes. 
In recent years a number of workers have tried to elucidate relations among 
amino acids by examining experimentally available sequences and frequencies 
[3]. These attempts have been thwarted by some serious difficulties. The 
known sequences are selected by the small size and ease of purification of the 
proteins involved. The proteins are thus often unrelated as to source or func- 
tion. The failure, in this limited range, to find constraints should therefore not 
be taken as the final word on the problem. 


An alternative method of investigating the ordering relations is to give up 
dealing with pure proteins and examine mixtures of related proteins in an effort 
to elucidate statistical relations. Thus we may inquire into the frequency dis- 
tribution of N-terminal amino acids, C-terminal amino acids, amino acids made 
N terminal by tryptic digestion, or any other experimentally available cut. We 
may look at the entire collection of proteins from a single cell type, or any 
subgrouping of proteins definable by some experimental procedure. It must be 
kept in mind that we are seeking general laws for given collections of proteins. 

The experimental procedures are derivable from the techniques of sequence 
analysis [4]. In particular, use is made of the fluorodinitrobenzene reaction and 
the Edman degradation [5]. The application of these techniques must be modi- 
fied, as we are dealing with a collection of proteins rather than a single protein. 

Since detailed experimental protocols are being published elsewhere, we shall 
only outline the method here [6]. E. coli cells are grown on a defined medium 
and randomly labeled with a single radioactive amino acid (leucine) or in the 
case of sulfur 35 label they are randomly labeled in three amino acids, cysteine, 
cystine, and methionine. The cells are then washed, and extracted with ethanol, 
ether, and hot trichloroacetic acid. The residual material containing most of 
the cell protein and little else is subjected to oxidation with performic acid. 
One portion of this material is subjected to fluorodinitrobenzene end group 
analysis or conversion to the phenylthiohydantoin derivatives; a second portion 
is digested to completion with trypsin and then subjected to the same end 
group analysis. Quantitative end group analysis is carried out by counting the 
spots on paper chromatograms that result from N-terminal amino acids and 
interior amino acids. 

We thus determine the frequency of occurrence of cysteine (including 
cystine), methionine, and leucine in their positions, N terminal and following 
the tryptic cut (which has a high degree of specificity for the carboxyl end of 
arginine and lysine). The results of these experiments are shown in table 1. 


Percentage of 
Amino Acid in 

Percentage of 
Amino Acid 



Amino Acid 


Tryptic Cut 




Cystine and cysteine 

Less than 0.2 


To interpret these results, let us first consider what results would have been 
obtained if there were no ordering rules. That is, suppose that each protein 
sequence was completely independent and any ordering was equally probable. 
If there were a large number of different protein species, on a statistical basis 
the amino acids would be expected to be randomly ordered. On the basis of 


average chain length as calculated from Van Slyke amino nitrogen determina- 
tions any amino acid should be N terminal 1.5 per cent of the time and should 
follow the tryptic cut about 14 per cent of the time. 

Failure to observe these ratios indicates one of two situations: either there 
are statistical rules of sequence, or a small number of protein species con- 
tributes a large fraction of the total protein mass. Recent results reported on 
the column separation of E. coli proteins [7] indicate that the protein mass 
is distributed with reasonable uniformity among a large number of proteins or 
protein groups. Thus, the results presented in table 1 suggest that statistical 
rules of sequence are operative in specifying E. coli proteins. 


1. H. F. Gaines, Cryptanalysis, Dover 5. Fraenkel-Conrat et al., Methods of 
Publications, Inc., 1956. Biochem. Anal, 2, 399 (1955). 

2. R. B. Roberts et al., Studies of Bio- g. H . J. Morowitz and M. Spaulding, 
synthesis in Escherichia coli, Carnegie Inst. Biochim. et Biophys. Acta, in press. 
Wash Publ. 607, Washington, D. C, 1955 ? Work reported in ^ yolume by 

3. G. Gamow et al., Advances in Biol. . 

and Med. Phys., 4, 23 (1956). workers at Carne g ie Institution of Wash- 

4. C. B. Anfinsen and R. R. Redfield, ington, Department of Terrestrial Mag- 
Advances in Protein Chem., 11, 1 (1956). netism. 


The Formation of Protomorphs 


Department of Terrestrial Magnetism 
Carnegie Institution of Washington 

Disrupted cells of Escherichia coli were suspended in a number of different 
solutions to find which ones were suitable for making stable preparations of 
ribosomes. Among those tested were some containing manganese and mag- 
nesium, because these ions had been found essential for incorporation of amino 
acids by cell-free systems [1, 2]. After standing for several hours these solu- 
tions became turbid and finally gave a white precipitate of unusual appearance. 

Examination of these solutions in the phase contrast microscope showed that 
the turbidity was caused almost entirely by the presence of large numbers of 
nearly spherical, highly refractile particles with diameters of 1 to 5 microns 
(figs. 1 and 2). 

The appearance of these cell-like particles in a solution that originally con- 
tained nothing visible in the microscope was quite surprising. The formation 
of large, stable aggregates with distinct boundaries from a fluid containing 
macromolecules in a homogeneous suspension seemed to illustrate a process 
which perhaps was important in the origin of life. Accordingly, we proceeded 
to investigate some of their properties. It was soon found that the particles con- 
tained protein, nucleic acid, and lipid in proportions typical of biological ma- 
terials. Because they are formed from protoplasm and have distinctive shape 
we refer to them as "protomorphs" to distinguish them from other particles 
or structures that exist in the living cell. 

The usual procedure for preparation of protomorphs is as follows: Harvest 
10 g wet weight of Escherichia coli cells growing in synthetic medium "C" [3]. 
Wash twice with tris(hydroxymethyl)aminomethane-succinate buffer, 0.01 M, 

1 Present address: Johns Hopkins University, Applied Physics Laboratory, 8121 Georgia 
Avenue, Silver Spring, Maryland. 


Fig. 1. Protomorphs in the solution from which they form. 

Fig. 2. Aggregated protomorphs scraped off glass surface. Phase contrast photomicro- 
graphs taken by W. R. Duryee. 



pH 7.6 (TS). Suspend with 10 ml TS, and break the cells by forcing the sus- 
pension through an orifice at 10,000 psi with a flow rate of roughly 2 ml/min. 
Add TS to bring the total volume to 50 ml, and centrifuge 15 minutes at 
40,000^. Discard the precipitate, which contains unbroken cells and large frag- 
ments of cell walls. Centrifuge again at 105,000^ for 15 minutes, and discard 
the precipitate, which contains smaller fragments of cell walls and a small pro- 
portion of the ribosomes. Dilute the supernatant fluid with TS to 100 ml, and 
add MgCL and M11CI2 to make it 0.005 M in each. In 5 to 48 hours the solu- 
tion will show turbidity because of protomorph formation. The entire proce- 
dure is carried out at 0° to 5 J C. 

The yield of this procedure is variable, as is the time required for formation 
of the protomorphs. Some of the sources of variability have been identified; 
others remain unknown. 

The addition of manganese is essential. No protomorphs were formed when 
Mn ++ was omitted even though adequate Mg ++ was present. Higher concen- 
trations of Mn" (0.01 M to 0.5 M) caused the formation of a precipitate of 
particles of irregular shape and widely variable size. No attempt was made to 
find other cations that might substitute for the Mn + \ The Mg ++ was not essen- 
tial but seemed to increase the yield. No difference was noted whether the 
magnesium was added before or after the cells were broken; thus there would 
appear to be no difference to protomorph formation whether the ribosomes 
were in the 80 S form or not (see paper 3 of this volume). 

No protomorphs were formed when the £>H of the solution was outside the 
limits 7 to 8; a pH of 7.5 seemed to be optimum. 

The concentration of orthophosphate affects the yield. When the cells are 
carefully washed in TS before breaking, the phosphate of the growth medium 
is removed and phosphate must be added to give a concentration of 10" 3 M. 
If higher quantities of phosphate are added the particles become less refractile 
in appearance and are dark, rough, and "hairy." At still higher concentra- 
tions of phosphate, precipitates are formed in manganese solutions lacking cel- 
lular material. Although these inorganic precipitates have only a slight re- 
semblance to the protomorphs, it is possible that the inorganic material provides 
a framework on which the organic material deposits. 

The concentration of the cellular material is important. When the usual pro- 
cedure was followed a twofold dilution of the cell extract prevented the forma- 
tion of protomorphs. The presence of cell-wall material was not important to 
the yield; the yield was the same whether or not the centrifugation steps to re- 
move cell walls and unbroken cells were omitted. When the wall material was 
present the protomorphs appeared less smooth, as if irregular fragments of wall 
had been incorporated. 

The constituent responsible for the sensitivity to the concentration is prob- 
ably deoxyribonucleic acid (DNA). The addition of DNAase invariably pre- 
vents the formation of protomorphs. The pressure cell routinely used to dis- 
rupt the cells also degrades DNA, as preparations of DNA lose their viscosity 


on being forced through the orifice of the pressure cell. Cell juices prepared 
with the pressure cell do not show the peak characteristic of DNA in their 
sedimentation diagram [4]. It is these preparations (made with the pressure 
cell) that will not give protomorphs when diluted. In contrast, preparations 
made by grinding the cells with alumina or by lysozyme treatment followed 
by osmotic shock (methods which preserve DNA) do yield protomorphs even 
at one-tenth the usual concentration. Finally, the addition of DNA to dilute 
pressure-cell preparations restores the yield. It seems quite likely that varia- 
tions in the pressure and in the conditions at the orifice during the disruption 
of the cells affect the quantity of intact DNA remaining and thereby influence 
the yield in an erratic way. 

The formation of protomorphs was photographed by Drs. B. Hoyer and 
N. Kramis, of the Rocky Mountain Laboratory, U. S. Public Health Service. 
Their time-lapse photomicrography shows that the growth of an individual 
protomorph from its first appearance to full size requires only a few minutes 
after a much longer induction period. There may well be a slow process of 
nucleation followed by a rapid process of growth. Neither fission nor fusion 
played any part in the growth process. 

Once formed, the protomorphs are stable. Unlike simple coacervate particles 
which exist only in a narrow pYL range and have a strong tendency to fuse 
or to dissolve, protomorphs can be handled like bacteria or yeast. There is no 
difficulty in centrifuging the photomorphs into a pellet (lOOOg-) and resuspend- 
ing in fresh media. They are quite stable in a number of ordinary media, and 
persist for weeks even though overgrown by bacterial contamination. They 
are not dissolved by short exposures to ammonia (1 M), 5 per cent trichloro- 
acetic acid (TCA), ethanol, or ether. They are dissolved in 0.01 M ethylenedi- 
aminetetraacetic acid to give a clear solution. 

On standing, glass vessels containing protomorph suspensions develop a 
white film over the surface. Microscopic examination of the material scraped 
from the glass indicates that the protomorphs have formed a moderately well 
packed monolayer on the surface (fig. 2). 

The organic components of the particles had roughly the proportions found 
in living tissues. They contain ultraviolet-absorbing material which hydrolyzes 
to yield the bases expected from ribonucleic acid. In addition the diphenyl- 
amine test [5] indicates that a small part (10 per cent) of the ultraviolet absorp- 
tion is due to DNA. The ratio of nucleic acid to protein (measured by the 
Folin reaction [6]) is 1/6 as compared to 1/4 in the bacterial juice. Paper 
chromatography shows the presence of lipid material. 

Incorporation experiments were carried out with thoroughly washed prepa- 
rations of protomorphs. Radioactive phosphate was incorporated at a constant, 
high rate for several hours. This process was not studied in any detail because 
all the radioactivity so incorporated could be extracted with cold TCA and 
there was no evidence of incorporation into macromolecules. 

The incorporation of radioactive amino acids was of more interest. The re- 



suits were erratic from one preparation to another, and the incorporation rate 
was only 1/1000 that of intact cells at the highest. Accordingly contamination 
by intact cells was a constant worry. The number of intact cells was estimated 
both by plate counts and by microscopic examination. The uptake observed 
was as much as 100 times that which could be attributed to the contaminants. 

There were also several qualitative features which distinguish the behavior 
of the protomorph preparation from that of intact cells. In the first place, the 
incorporation doubled if ATP was added or if the concentration of the amino 
acid mixture was doubled. These variations have little effect on incorporation 
by whole cells. Secondly, the distribution of incorporated radioactivity among 
the fractions soluble in cold TCA, alcohol, ether, and hot TCA was different 
from that obtained with whole cells [3]. Finally, the hot-TCA-insoluble ma- 
terial after hydrolysis yielded a pattern of radioactive amino acids different 
from the mixture supplied and different from what would be incorporated by 
whole cells. 

Accordingly we believe that the observed incorporation was in fact real, 
though not reproducible from day to day. Since these experiments were done 
we have learned of the activity of the cell-wall fraction in protein synthesis 
[7, 8]. In retrospect it seems quite likely that the variability in the synthetic 
capacity of the protomorphs may have been due to a variability in their content 
of cell-wall fragments. 

These protomorphs are of course very different from the ribosomes which 
are the subject of this symposium, but there may be a relationship between 
them. It is a common belief that the bacterial cell is not a homogeneous mix- 
ture of its various components; on the contrary, various lines of evidence indi- 
cate that it has a high degree of organization. Organization in turn implies 
the action of forces between the various constituents such as DNA, RNA, pro- 
tein, and ribosomes. It is possible that the aggregation of these cellular con- 
stituents into protomorphs may be another manifestation of the same forces 
which maintain organization in the living cell and may furnish a material in 
which the forces are more amenable to study. If so, the protomorphs may 
eventually contribute to our knowledge of how the ribosomes are organized 
within the cell. 


1. S. Spiegelman, A Symposium on the 
Chemical Basis of Heredity, p. 232, Johns 
Hopkins Press, 1957. 

2. B. Nisman, personal communication, 

3. R. B. Roberts, P. H. Abelson, D. B. 
Cowie, E. T. Bolton, and R. J. Britten, 
Studies of Biosynthesis in Escherichia coli, 
Carnegie Inst. Wash. Publ. 607, 1955. 

4. E. T. Bolton, B. H. Hoyer, and D. B. 
Ritter, paper 3 of this volume. 

5. Z. Dische, The Nucleic Acids, vol. I, 
chapter 9, Academic Press, New York, 

6. O. H. Lowry, N. J. Rosebrough, A. L. 
Farr, and R. J. Randall, /. Biol. Chem., 
193, 265 (1951). 

7. V. R. Srinivasan and S. Spiegelman, 
Bacterial. Proc, 58, 101 (1958). 

8. S. Ochoa, Spring Meeting, National 
Academy of Sciences, 1958. 


Structure of Microsomal Nucleoprotein 
Particles from Pea Seedlings 1 


Division of Biology, California Institute of Technology 

Microsomal nucleoprotein particles from pea seedlings have been isolated 
and characterized in our laboratories [1|. These particles have an RNA/pro- 
tein ratio of 4/6 and a molecular weight of 4 to 4.5 X 10 r> , and they appear in 
the electron microscope as oblate spheroids (fig. 1). The microsomal nucleopro- 
tein particles of pea seedlings thus seem to be similar to those of yeast [2], liver 
[3], and perhaps also to those of bacteria as reported in this meeting. 

The problem of the structure of the particles can be approached from at 
least three lines of inquiry. (1) How are the smaller units of nucleoprotein put 
together in the 80 S particles? (2) How are the RNA and protein put together 
in the nucleoprotein? (3) If protein synthesis takes place in the particle, what 
is the structural relation of the newly synthesized protein to the nucleoprotein 
that constitutes the particle? The present paper is principally concerned with 
the first of these questions. 

Two salient features of the approach and of the interpretation should first be 
noted. Experimentally, every precaution has been taken to assure that the 
dissociation agents employed do not exert hydrolytic action on covalent bonds. 
In interpreting the results of dissociation studies, it is assumed that all particles 
in the preparation have similar gross structure. The simple ultracentrifugal 
patterns of the dissociated particles, and the reversibility of certain dissociating 
processes, seem to support this assumption. It is entirely conceivable, however, 
that though in general features the subunits of all the particles may be very 
similar, the cohesive strength with which they stick together inside the particle 

1 This work was supported in part by grants Nos. Rg-3977 and Rg-5143 from the 
National Institutes of Health, United States Public Health Service. 




Fig. 1. Electron micrograph of a freeze-dry preparation of microsomal particles from pea 
seedlings. Magnification 41,000 X; thorium shadowed with an angle of 6.5:1. 

may vary during the different phases of cell growth and differentiation. Pre- 
liminary information suggesting that this is so has been reported in this meet- 
ing: information concerning rapidly growing cells, such as those of yeast and 
bacteria. Eventually, then, studies of structure of particles in the steady state 
will evolve into studies of structure of particles through a life cycle. 

Processes or agents that remove or substitute divalent ions from the particles 
promote dissociation of the particles. For instance, in a 0.8 per cent particle 
solution, in pH 6.5, 0.025 p, phosphate buffer, there is only 6 per cent dissocia- 



tion of 80 S particles into 60 S and 40 S units. In pH 7.5, 0.05 u, buffer, there 
is 50 per cent dissociation, and in pW 8.5 buffer, there is over 80 per cent dis- 
sociation (fig. 2a, b, c). Substantially, all the subunits in pH 7.5, 0.05 u, phos- 
phate buffer can be returned to the form of 80 S particles either by addition of 
magnesium (5 X 10" 4 M) or by titration of the pH back to pH 6.5 (fig. 2d, e) . 
Thus, the dissociation of 80 S particles to 60 S and 40 S units is a reversible 
one. Increase in pH or concentration of phosphate buffer promotes the disso- 
ciation by taking magnesium away from the particles. This interpretation is 







Fig. 2. Ultracentrifuge patterns of particles in 0.05 fi K-phosphate buffer as affected by 
pH. Speed 35,600 rpm. (a) In pH 6.5 buffer at 0° C for 13 hours. (/>) In pH 7.5 buffer 
at 0° C for 16 hours. (<:) In pH 8.5 buffer at 0° C for l'/ 2 hours, (d) Solution in pH 7.5, 
0° C for 2 hours, back-titrated with phosphoric acid to pH 6.5, stored at 0° C for 3 hours. 
(e) Solution in pH 7.5, 0° C for V/z hours. MgCl 2 added to final concentration of 
5 X 10~ 4 M, stored at 0°C for 3 hours. 



further substantiated by experiments on dialysis of the particles. When a par- 
ticle preparation is dialyzed against a buffer in which it is stable, dissociation 
to 60 S and 40 S units occurs (fig. 5a, d). This dissociation is caused by re- 
moval by dialysis of essential cofactor(s), since it can be prevented or reversed 
by dialyzing the particles against the deproteinized supernatant (fig. 3>b, e). 
The cofactor is present in the supernatant of the extract from which the par- 
ticles were originally isolated; it is completely heat stable and not absorbed on 
charcoal; finally, its protective and dissociation-reversing properties have been 
duplicated by magnesium chloride, 5xl0" 4 M (fig. 3c, /). Calcium chloride 

Fig. 3. Ultracentrifuge patterns of particles after dialysis. Speed 36,500 rpm. (a) Dialy- 
sis against 0.025 fi K-P0 4 , pH 6.5, 2-4° C, for 13 hours, (b) Supernatant, plus 0.025 fi 
K-P0 4 , pH 65, 2-4° C, for 14 hours, (c) 0.025 /x K-P0 4 , pH 6.5, 5 X 10" 4 M MgCl 2 , 
2-4° C, for 19 hours, (d) 0.025 ^ K-P0 4 , pH 6.5, 2-4° C, for 24 hours, (e) Solution first 
dialyzed in phosphate buffer to produce solution of (a), and the solution of (a) redialyzed 
against supernatant with 0.025 p K-P0 4 , pH 6.5, 2-4° C, 14 hours. (/) Solution first 
dialyzed in phosphate buffer to produce a solution similar to (a), and the solution of 75, 
60, and 40 S particles redialyzed against 0.025 p. K-P0 4 , pH 6.5, with 5 X 10' 4 M MgCl.,, 
2-4° C, for 14 hours. 



at the same concentration has also been found to have protective properties. 
Similar results have been obtained for yeast particles [4]. 

The dissociation of particles to smaller subunits is a relatively rapid process. 
Thus, essentially the same proportions between 80, 60, and 40 S components 
are obtained in runs on materials incubated at 0° C in phosphate buffer, pH 
7.5, 0.05 |j, for 2 minutes or for 16 hours. No evidence was found in these ex- 
periments to suggest that the 60 and 40 S components are formed by aggre- 
gation of smaller products. 

In concentrated potassium chloride solution (0.35 to 0.7 M), the particles 
also dissociate into 60 and 40 S (fig. Ab, e) subunits, and this process too is sub- 
stantially reversible (fig. 4&, c). Addition of magnesium chloride (0.015 to 

KCI 0.7M 







Fig. 4. Ultracentrifuge patterns of particles in KCI solutions, (a) In 0.7 M KCI, 24° C, 
for 30 minutes. Rotor speed 35,600 rpm. (/>) In 0.7 M KCI, 0° C, for 1 hour. Centrifuged 
at 7° C. Rotor speed 42,040 rpm. (c) Solution of (b) dialyzed in 0.025 fi K-P() 4 , pH 6.5, 
MgChl X 10" 3 M, 2-4° C, for 14 hours. Centrifuged at 6.8° C. Rotor speed 42,040 rpm. 
(d) In 0.7 M KCI with 0.015 M MgCh, 0° C, for 1 hour. Rotor speed 35,600 rpm. (<?) In 
0.35 M KCI, 0° C, for 1 hour. Rotor speed 35,600 rpm. 



0.1 M) partly suppresses but does not completely eliminate the dissociation (fig. 
4d). The effects of potassium chloride are, therefore, probably twofold. Potas- 
sium ions appear to replace magnesium within the particle. The further effect 
of potassium which cannot be suppressed by addition of magnesium is probably 
due to ionic effects upon charged groups. An additional complication sets in 
when the particles are exposed to potassium chloride solution at room tem- 
perature. As will be described in more detail later, the RNA in the dissociated 
particles is now hydrolyzed by the contaminating RNAase in the preparation 
with a rate eight times higher than the rate constant for attack of the RNA 
ot the nondissociated particles. The effect of the contaminating RNAase on 
the particles in 0.7 M KCl can be observed by comparing figures 4a and \b. 
There is much more material of very low sedimentation coefficient in the 
room-temperature runs. Furthermore, the area ratio of the 80 S : 60 S : 40 S 
components in the low-temperature run is 1.3 : 1.8 : 1.0, whereas in the run af- 
fected by RNAase it is about 1:1:3. The 40 S component, therefore, may be 
more resistant to RNAase than the others. 

Magnesium ions can be further removed by EDTA. At neutral pH, the sys- 
tem treated with EDTA aggregates at room temperature. Thus, this system 
can be analyzed only at low temperature or at alkaline pH. In pH 6.5 phos- 
phate buffer, and in the presence of 2.5 X 10~ 2 M EDTA, the particles disso- 
ciate to yield components of 40 S (64 per cent), 26 S (30 per cent) and 3 to 6 S 
(6 per cent) (fig. 5b). Lower concentrations of versene, such as 5 X 10" 3 M, 
yielded poor resolution of the 40 and 26 S components (fig. 5a). At higher 
pH (9.0), higher concentrations of EDTA (5x10"' M), and room tempera- 
ture, the particles dissociate into the 40 and 26 S units together with increased 
amounts of material of 3 to 6 S(fig. 5c). Dialysis to remove EDTA and to 
replenish magnesium causes aggregation of the system. It should be noted 


Fig. 5. Ultracentrifuge patterns of particles in EDTA. Speed 42,040 rpm. (a) In 
5 X 10- 3 M EDTA, pH 6.6, 0.025 [x, K-P0 4 0° C, for 1 hour. Centrifuged at 4.6-6.0° C. 
(b) In 2.5 X 10" 2 M EDTA, pH 6.6, 0.025 ^ K-P0 4 , 0° C, for 1 hour. Centrifuged at 
6° C. (c) In 5 X 10" 2 M EDTA, pH 9.0, 0.02 M K-CO s , 0° C, for 3 hours. Centrifuged 
at 20°C. 



that in this concentration of versene (0.05 M), and at pH 9.0, less than 1 per 
cent of the original magnesium should remain with the particles. 

The RNA/protein ratios of the dissociated components have been investi- 
gated. The 60 and 40 S components were shown to be attacked and precipi- 
tated by RNAase and protamine, indicating that they contain RNA. The 
problem was then approached more quantitatively by comparing the ultracen- 
trifuge patterns obtained by schlieren optics with those obtained by ultraviolet 
absorption optics which measures the sedimentation of nucleic acid. It should 
be noted that the concentration of particles in solutions employed for ultra- 
violet absorption optics is 30 to 50 times lower than that used for schlieren optics. 
In dilute solutions, the sedimentation coefficient will be higher and the reversi- 
ble dissociation reaction will proceed further toward completion. Figure 6a 
shows that all the nucleic acid moves as one component of 79 S in water. At 
pH 7.5, 0.05 u, phosphate (fig. 6b), the particles are dissociated into two ultra- 
violet-absorbing components of 37 and 59 S, those shown above to appear also 
in the schlieren pattern. The absence of the 80 S particles, which do appear in 
the schlieren pattern, in the ultraviolet absorption pattern in this phosphate 
buffer is to be attributed to the complete dissociation of the original particles 







Fig. 6. Ultracentrifuge ultraviolet absorption patterns of particles under dissociating and 
nondissociating conditions, (a) In water. Rotor speed 25, 980 rpm. (b) In 0.05 /x K-P0 4 , 
pH 7.5, 0° C, for 1 hour. Centrifuged at 7.7° C. Rotor speed 37,020 rpm. (c) In 0.025 M 
ETDA, pH 6.6, 0.025 p, K-P0 4 , 0° C, for 1 hour. Centrifuged at 6.2° C. Rotor speed 
42,040 rpm. (d) In 0.7 M KC1, 0° C, for 1 hour. Centrifuged at 8.5° C. Rotor speed 
37,020 rpm. 

TS'O 163 

into the 59 and 39 S components at the low concentration. The ratio of the 
amounts of ultraviolet-absorbing material, i.e., nucleic acid, contained in the 
59 and 39 S boundaries is similar to the ratio of amounts of material contained 
in the two boundaries as observed with schlieren optics. These ratios are ap- 
proximately 2.3 : 1. Treatment with EDTA results in nucleic acid-containing 
components of 24 and 44 S (fig. 6c) as were likewise observed by schlieren 
optics (fig. 5b). The ratio of the amounts of nucleic acid in the two compo- 
nents is similar to the ratio of the amounts of material contained in the two 
boundaries as observed in schlieren optics, namely, 2:1. The pattern of the 
particles incubated in 0.7 M KC1 and observed with absorption optics (fig. 6d) 
is, however, markedly different from that observed with schlieren optics (fig. 
4£). Components of 75 and 58 S which are observed with schlieren optics are 
not observed with absorption optics. In the latter case, however, a new bound- 
ary of component 23 S is present to the extent of about 40 per cent of the 43 S 
component. The probable explanation of this behavior is that, at the low con- 
centration necessary for the employment of ultraviolet optics, the 80 S and the 
59 S particles are further dissociated to an undetectable concentration and a new 
component of lower sedimentation appears. The 24 S component observed 
with ultraviolet optics in potassium chloride-treated preparations is presumably 
similar to or identical with the 25 S component observed in solutions treated 
with EDTA. Thus, the original particles and their dissociation products all 
have similar ratios of ultraviolet-absorbing material to mass. 

The intrinsic viscosity of microsomal particles in media which promote dis- 
sociation was studied in order to find out what over-all changes in frictional 
coefficient are attendant upon dissociation. These experiments were performed 
in versene and in potassium chloride solutions at low temperature since at 
higher temperature the system in versene aggregates and the system in potas- 
sium chloride shows evidence of ribonuclease action. Even though particle 
preparations were clarified twice by low-speed centrifugation, the possibility 
of the presence of small amounts of aggregates cannot be excluded. The in- 
trinsic viscosity of the microsomal particles in phosphate buffer was found to 
be 0.11 (100 cc/g) (fig. la), identical with the value found for 24° C. In water, 
the value was 0.12/100 cc/g, an increase which may be due to the electroviscous 
effect. In solutions of dissociated particles in versene at p¥L 6.8 and 9.0, and in 
potassium chloride (fig. lb, c, d), a slightly higher value of reduced viscosity 
was obtained at high concentration. The intrinsic viscosity of these solutions, 
however, was again found to be 0.11 to 0.12. Thus, no change in intrinsic 
viscosity is observed upon dissociation of the particles. The most reasonable 
explanation for this result is that the units of 60, 40, and 25 S have the same 
frictional coefficient as the 80 S particles. 

The above experiments on the dissociation of the 80 S particles can be sum- 
marized as follows: 

1. When magnesium and calcium are removed from microsomal particles 
(79 to 81 S), three ultracentrifugally identifiable components are obtained as 







O ,4 


~ 12 

P0 4 pH 6.5 


.~*~ .10 



0.7 M 


o^~- — ""° 












pH 68 














EDTA P H9.o 




m 9/cc 

12 16 


Fig. 7. Reduced viscosity of particles under dissociating and nondissociating conditions. 
(a) In 0.025 fx. P0 4 , pH 6.5, with 1 X 10~ 4 M MgCU. (£) In 0.7 M KCI, pH 6.5, 0.02 p. 
K-P0 4 , 0° C, for 3 hours, (c) In 0.03 M EDTA, pH 6.8, 0.02 /* K-P0 4 , 0° C, for 3 hours. 
(</) In 0.05 M EDTA, pH 9.0, 0.01 M Tris buffer, 0° C, for 3 hours. 

dissociation products. At infinite dilution these components have sedimenta- 
tion coefficients of 59 to 61 S, 40 to 43 S, and 25 to 27 S. 

2. All these units contain nucleic acid and probably have RNA/protein ratios 
similar to that of the 80 S particles. 

3. The mixtures, and perhaps the individual units, of 60, 40, and 25 S com- 
ponents have frictional coefficients similar to that of the 80 S particles. 

4. Under conditions in which only 60 and 40 S units are formed by dissocia- 
tion of 80 S particles, the two components occur in an amount ratio of 2.2-2.3 
to 1 respectively in both schlieren and absorption optics. Treatment with EDTA 
over a wide range of concentration of both EDTA and particles results in the 
formation of only 40 and 25 S units. These are present in an amount ratio 
of 2:1, indicating that the two units in the system are in a stable state. 

On the basis of the above observations, the molecular weights of the 60, 40, 
and 25 S components can be estimated to be 2.6 to 3.0 X 10 6 , 1.3 to 1.5 X 10 6 , and 
6.5 to 7.5 X 10\ respectively. The ratios between the molecular weight of the 
80 S particles and the molecular weights of the successively smaller units are 
then in the series 6:4:2:1. Chao has reached a similar conclusion for the 60 and 
40 S components of yeast particles [4|. Reports of Wagman, and of Tissieres 
and Watson, in this meeting also tend to support this formulation. Both groups 
find that the 40 S particle from Escherichia coli is spherical, with an RNA/pro- 
tein ratio close to unity, and a molecular weight of 1.3 to 1.8 million. Since the 
40 S particle is spherical, the difference between its sedimentation coefficient and 

TS'O 165 

the sedimentation coefficients of the 60 and 80 S particles of E. coli reported by 
Bolton et al. in this meeting is most probably due to differences in molecular 
weight. These would then form the series 3:2:1. In addition, particles of 25 S 
have been found as dissociation products of the 40 S particle of E. coli by 
Tissieres and Watson, this too having a spherical shape in the electron micro- 
scope, as observed by Hall. A 25 S component has also been reported by Chao 
[4] as one of the dissociation products of yeast particles. 

The following scheme is proposed to account for the experimental findings 
concerned with the dissociation of the 80 S particles. When a certain fraction of 
the magnesium ion is removed from the system (about 50 per cent as based on 
the binding constants of RNA and phosphate [5]), the 80 S particle dissociates 
reversibly to form a 40 S unit, one-third of the original particle, and a 60 S unit, 
two-thirds of the original particle. When larger amounts of magnesium ions 
are removed (over 95 per cent, by versene), not only does dissociation of the 
80 S particles to 40 and 60 S units occur, but in addition the 60 S unit also is 
degraded, possibly irreversibly, to form a 40 S unit and two 25 S units. The 
final result of such a dissociation should be a system containing two 40 S units 
and two 25 S units per original 80 S particle. This formulation, then, fits the 
experimental findings both as to the mass ratio of 2:1 between 40 and 60 S units 
and as to the mass ratio of 2:1 between 40 and 25 S components. 

The magnesium and calcium contents of the particle have been analyzed by 
flame spectrophotometry." There are 3.0 to 3.2 umoles of magnesium per 12 
umoles of RNA-phosphorus or per 10 mg dry weight of particles. All the mag- 
nesium appears to be extractable by 0.5 N TCA at 3 C. Assuming that cal- 
cium can also be completely extracted by TCA, then, again, per 12 umoles of 
RNA-phosphorus, there is 0.45 to 0.55 umole of calcium, about one-sixth of 
the amount of magnesium. Thus, the particles contain 3.5 to 3.7 umoles of 
magnesium and calcium per 12 umoles of phosphate, or 1 mole of divalent ions 
for each 3.3 ±0.2 umoles of phosphate. 

It is of interest to estimate the apportionment of these divalent ions between 
the RNA and the protein. Combining the data on molecular weight (4.5 X 10 c ), 
RNA content (40 per cent), and content of divalent ions per mole of phos- 
phorus, we may calculate that there are 1.5 XlO 3 moles of magnesium and cal- 
cium per mole of particle. Since the cation binding capacity of the microsomal 
protein is unknown, we introduce for comparison the results that may be cal- 
culated for bovine serum albumin, a protein with a high proportion of dicar- 
boxylic amino acids. According to Carr [6], there are 8 calcium binding sites 
per molecule at pW 7.4. The total amount of protein per mole of microsomal 
particles is equivalent to 36 moles of bovine serum albumin. This protein 
would therefore bind no more than 300 moles of calcium, and would account 
for less than 20 per cent of the divalent ions bound by the particles. We con- 

2 The technical assistance of Mr. Merck Robison, Carnation Company, is gratefully 


elude, therefore, that more than 80 per cent of the divalent ions are bound by 
RNA. Experiments with pancreatic ribonuclease further support this con- 
clusion. By the action of ribonuclease, a partial separation of protein and 
degraded nucleic acid of the particles is achieved. Most of the protein precipi- 
tates from solution, leaving behind a mixture of nucleotides and polynucleo- 
tides. The distribution of magnesium between supernatant and precipitate 
should provide information concerning the binding sites of the magnesium 
ions. It was found that, after RNAase had acted on the particles for 2 hours, 
85 per cent of the magnesium had been liberated into the supernatant, which 
contains 65 per cent of the nucleotide phosphate and only 10 to 15 per cent of 
the protein. If we assume that 80 per cent of the magnesium is associated with 
RNA in the particle, then half of the phosphate groups in the RNA exist in 
the form of magnesium salts. 

Attempts to study the protein of the particle have been made with RNAase. 
After 3 hours' incubation with 7 ug/ml of pancreatic RNAase in a 1 per cent 
solution of particles at 27° C, the solution becomes turbid. Between 85 and 
90 per cent of the total protein and 8 per cent of the total phosphorus are sedi- 
mentable by low-speed centrifugation, leaving 10 to 15 per cent of the protein 
and 90 per cent of the phosphorus in the form of mono- or oligonucleotides in 
the supernatant. That the protein aggregation is related to RNAase action and 
not to the presence of nucleotides was shown by experiments in which large 
amounts of 2,3'-phosphate nucleotides were added to particle preparations with- 
out aggregating effect. The protein aggregate is insoluble in buffer at pH 4 to 
11, in strong salt solution, in acid, in urea, or in performic acid, but does dissolve 
in alkali at pH 12 to 13 as well as in 80 per cent saturated guanidium chlo- 
ride at pH 8 to 9. If the guanidium ions are removed by dialysis or if the pH 
is lowered to 10 to 11, the protein again precipitates. 

The action of RNAase on the particles is enhanced by addition of potassium 
chloride or phosphate buffer and is suppressed by addition of magnesium. It 
was also found that there is a small amount of contaminating RNAase in our 
preparation of particles. Incubation of particles in phosphate buffer (0.05 u, 
pH 6.5) for only 2 hours at room temperature hydrolyzes 4 to 5 per cent of the 
RNA. This rate can be increased by 8 times by addition of potassium chloride 
(0.15 to 0.7 M), and it can be suppressed back to 1.5 to 2 times by addition of 
magnesium chloride (0.001 M) with the potassium chloride. 

The present studies shed little light on the nature of the binding between 
RNA and protein. Electrostatic forces cannot be the only ones involved, since 
at least half of the phosphate groups of the RNA are present in the form of 
salt linkages with magnesium and calcium. In addition, RNA does not sepa- 
rate from protein in the ultracentrifuge in the presence of 0.7 to 1.0 M KCl, 
nor can protein be separated from nucleic acid by alcohol precipitation in such 
solutions. Agents which denature proteins, such as phenol or detergent, have 
been successfully applied to separate RNA from the protein of the particle. 
The protein, however, aggregates as the RNA is removed, as shown by the 

TS'O 167 

experiment with RNAase. These findings, coupled with the fact that urea does 
demolish the particle, suggest that hydrogen bonding between RNA and pro- 
tein may be an important factor in holding the two constituents together. 

Microsomal particles have been isolated from sections of pea seedlings that 
have been incubated with C 14 -leucine for 1 hour. 3 The particles, after a series 
of centrifugations and dialyses, were shown to be void of contaminating free 
C 14 -leucine, by two methods. The specific activity (cpm/mg protein) of such 
particles was the same as that of particles which were washed further by hot 
TCA and 1 N NaOH. Moreover, acid hydrolysis of the microsomal protein 
after treatment with dinitrofluorobenzene yields over 98 per cent of the C 14 - 
leucine as free leucine, indicating that most of the C 14 -leucine is linked inside 
the protein molecule. With such a preparation of labeled 80 S particles, we 
wish to find answers to two specific questions. 

The first question is based on the idea that only about 1 to 5 per cent [7] of 
the protein in the particle is being actively synthesized in the particles. It has 
been suggested that this newly formed protein then passes into the cytoplasm. 
In the dissociation studies, it is found that after treatment with EDTA, pH 
6.5, most of the nucleoproteins aggregate and sediment out of solution, leaving 
about 6 per cent of protein in the supernatant. After the action of RNAase, 
there is also about 10 per cent of protein which stays in the solution with the 
nucleotides, while 90 per cent of the protein precipitates out of solution. When 
the labeled particles were subjected to the above two treatments, only a small 
percentage of the total counts of the labeled protein remained in the super- 
natant, while over 90 per cent of the counts precipitated with the aggregates. 
The supernatant proteins obtained from these two treatments were precipitated 
by TCA and were shown to have specific activities no higher than that of the 
aggregated nucleoproteins. This experiment suggests that substantial removal 
of magnesium and RNA does not liberate a large percentage of labeled proteins 
or proteins of very high specific activity from the isolated particles. 

The second question is based on the idea that RNA is a template [8]. Thus, 
is there a large amount of radioactive amino acids in the particle attached to 
RNA through covalent bonds? The RNA was separated from protein by 
phenol as well as by density gradient centrifugation in cesium chloride. In- 
variably, very few if any counts could be found in the RNA. Furthermore, 
the soluble nucleotide fraction of the supernatant after deproteinization by 
TCA, obtained by RNAase treatment of the particles, also contained very few 
counts. Therefore, if amino acids are attached to RNA through covalent bonds, 
the total amount of such material at a given time is likely to be very small. 

In summary, microsomal particles from pea seedlings consist of smaller units 
of nucleoproteins cemented together through linkage of magnesium ions and 
the phosphate group of RNA. RNA and the protein (s) in these units are held 

3 Experiments with C 14 -amino acids were performed in cooperation with Dr. Clifford 



tightly together, probably through hydrogen bonds. Removal of magnesium 
ions and hydrolysis of RNA from the particles labeled with C 14 -amino acids in 
the cell did not liberate labeled amino acid or protein (s) of high specific activity 
from the particles. 


1. P. O. P. Ts'o, J. Bonner, and J. Vino- 
grad, /. Biophys. Biochem. Cytol., 2, 451 

2. Fu-Chaun Chao and H. K. Schach- 
man, Arch. Biochem. Biophys., 61, 220 

3. M. E. Petermann and M. G. Hamil- 
ton, /. Biol. Chem., 224, 725 (1957). 

4. Fu-Chaun Chao, Arch. Biochem. Bio- 
phys., 70, 426 (1957). 

5. J. S. Wieberg and W. F. Neuman, 
Arch. Biochem. Biophys., 72, 66 (1957). 

6. C. W. Carr, Arch. Biochem. Biophys., 
43, 147 (1953). 

7. J. W. Littlefield, E. B. Keller, J. Gross, 
and P. C. Zamecnik, /. Biol. Chem., 217, 
111 (1955). 

8. H. Borsook, /. Cellular Comp. Phy- 
siol., 47, Supplement 1, 35 (1956).