Synthesis of messenger ribonucleic acid and the assembly of polysomes in Bacillus megaterium infected with bacteriophage

J. Mol. Biol. (1966) 22,223-233

Synthesis of Messenger Ribonucleic Acid and the Assembly of Polysomes in Bacillus megaterium infected with Bacteriophage M. S~HAE~HTER~ mu

K. MU&-N

19thdepartment of Chemical Microbiology Department of Biochemisty University of Cambridge, England (Received15 December1965, and in revisedform 19 October1966) The kinetics of synthesis, association with ribosomes and breakdown of messenger RNA were studied in bacteriophage-infected Badh wwgatmiwn. During the late stages of infection, the synthesis of stable (ribosomal and transfer) RNA was greatly reduced, permitting preferential labelling of messenger RNA over long periods of time. Sedimentation analysis of lysates indicated that even after short periods of labelling, the bulk of the messenger RNA was associated with ribosomes, suggesting that the free, first synthesized end of messenger RNA was associated with ribosomes before the molecule is completed. The formation of average molecules of messenger RNA and their complete loading with ribosomes was estimated to take place within 30 to 120 seconds at 37’C. Messenger RNA’s labelled during pulses of various length were compared with regard to their stability after the addition of actmomycin D. It was found that messenger RNA labelled for short periods decayed faster than that labelled for longer periods. This tiding suggested that the portion of messenger RNA made most recently is most sensitive to breakdown.

1. Introduction Most experimental approaches to the problem of mRNA synthesis in bacteria are based on the belief that it has only a transient existence before being broken down. Thus, the proportion of radioactive precursors found in mRNA, as compared with that in total RNA, is greater after a short pulse than after longer periods of labelling. The ability of aotinomycin D to inhibit the synthesis of all kinds of RNA was used to study the kinetics of breakdown of mRNA in Bacillus eubtilti by Levinthal, Keynan & Higa (1962). They found that after addition of this antibiotic a substantial fraction of pulse-labelled ribonucleic acid decayed with a half-life of about one minute. Schaechter, Previc & Gillespie (1965) obtained similar results with Bacillus megaterium, but observed that the breakdown of polyribosomes to single ribosomes was three to four times slower than that of the pulse-labelled RNA. Independent studies with the same organism had shown that the decay of protein-synthesizing capacity after addition of actinomycin D had a half-life of about four minutes (McQuillen, 1963, unpublished work). We wondered whether these differences might reflect stages in the synthesis of mRNA and its association with ribosomes. t Permanent address: Department Boston, Mass., U.S.A.

of Microbiology, 223

Tufts University

School of Medioine,

224

M.

SCHAECHTER

AND

K. McQUILLEN

Normally growing bacteria accumulate label from radioactive RNA precursors in the stable components, ribosomal and transfer RNA, and this makes kinetic studies of mRNA metabolism difficult. On the other hand, B. nze@eriztm infected with a virulent bacteriophage synthesizes a much smaller amount of stable RNA components and presumably makes mainly mRNA. By avoiding the complicated RNA metabolism of normally growing cells, we have been able to examine the properties of the unstable RNA labelled for different lengths of time. We have studied the decay of such material after the addition of sctinomycin D, and its association with ribosomes.

2. Materials and Methods (a)

Bacteria

and

bnctwiophnge

Two strains of BaciEZus megaterium were used throughout, this work: strain KM, and st,rain KMTT, which required t,hymine and tryptophan for growth. The double mutant was kindly supplied by Dr J. T. Waohsman. Phage M4 nas obtained through the eourtcsy of Dr S. Brenner. Phage titrcs were det,ermined by the method of plaque counting. Purified phage stocks were prepared by low-speed contrifugation of lysates to remox-e debris. The pellet from a subsequent high-speed ccntrifugntion wn.s sllbjected t,o caesium chlorida equilibrium centrifugation in an S30 Beckman rotor at 30,000 rcv./min for 18 hr (Kaiser & Hogness, 1960). Fract.ions containing the phages were tlialysed against, distilled water. Purified phages were stable in dist,illrtl water in the presence of chloroform when stored in a refrigerator for periods of several weeks. Unless otherwise stated, all experiments

were performed wit.h an estimated mukiplicity colony-forming

of 7 plaque-forming

units per bnctcrial

unit. (h) Cuztivation

The basic medium consisted of solution C (;1IcQuillcn & Roberts, 1954), 0.196 glucose, u,ith O.OOlq{, L-tryptophan ZOq/, sucrose, 0.5% Casamino acids. This was supplcmentcd and O.OOl”/o thymine for the mutant strain KaITT. Cultures ncre grown wit.h acrat,ion at 37°C and growth followed by ahsorbancy measurements at GO0 mp. In this mrdium, absorbancy doubled in 40 min. Experiments were carried out when cultures were growing exponentially and contained no more than 200 pg dry wright of cells/ml. (c) Lnbelling experiments Steady-state labelling with 3T was achieved bv growth for not less than t.wo genemtion times in media containing 1 (*c/ml. of carrier-free [32P]phospllate. [3H]Uridine stock solutions ofspeeific activities ranging from 7.6 to 22.0 c/m%1 were used. In some experiments, they were diluted with non-radioactive uridine. (d) Preparation of lysates of p7mage-infected cells Cultures were harvested by pouring themonanequalvolume ofcrushed, frozen, TBIK-20% sucrose (TRIK consists of 0.01 x-Tris; 0.01 x-magnesium acet,ate; 0.1 nz-KCl, pH 7.0, at) 0°C adjusted wit,h HCl) containing 200 pg crystalline egg white lysozyme/ml. Prcaparations were immediately centrifuged and wa,shcd with TMK-20% sucrose. The rpwtlting pellet consist,cd of protoplast,s which were resuspended in ThIK-%OO/, sucrose ant1 I>-scd by 1110 addit,ion of 0.496 sodium dcoxycholat,e (Schsechter, Previc & Gillespie, 1965). Lysatcs wore treated with 10 pg of elrd~rophorrtirnlly purified DP;aso (‘ll’orthington Corp.) per ml. for 5 to IO min at O’C. (0) Pxpawtion of nivA RNA dodecyl

was prepared from sulphate procedure

the above of Scherrer

DKase-treated lysatcs & Darnell (1962).

(f) Seclimentation

by t.he hot phenol-sodium

analysis

Lysates were diluted wit,h equal parts of TMK without layered directly on 5 ml. of 15 to 30 y0 (xv/v) sucrose gradients fuged at 39,000 rev./n% for 45 min. RNA solutions (0.1 ml.)

sucrose; 0.1 ml. of this was containing T&HI, and centriwere layered on 5 ml. of 4 to

PLATE I. Electron micrographs of bacteriophage M4 negatively stained with potassium phoepho t,ungstate. (a) Part of a typical field showing the high degree of purity of the preparation. x 40,000. (b) Higher magnification of the same preparation, showing icosahedral heads in which the individual protein subunits can be seen. X 200,000. The micrographs were taken with a Siemens Elmiskop 1 by Dr H. Home of the Agricultural Research Council Unit of Animal Physiology, Rabraham, Cambridgeshire, England.

( f,rr.i,ry ,, $2.;

MESSENGER 20% (w/v) sucrose centrifuged for 4 hr and filtered through photometer which Bray’s scintillation

RNA

AND

POLYSOMES

226

gradients containing 0.05 M-NaCl, 0.01 M-acetate buffer (pH 6.4), and at 37,000 rev./min. Fractions were either collected directly in 6% TCAt membrane filters, or passed through a continuous recording spectromeasured the absorbancy at 260 rnp, and then collected directly in fluid (1960).

(g) Counting procedure All samples were counted in a Nuclear Chicago liquid-scintillation membrane filters were counted with toluen+POP-POPOP solution. (h) Chernimle Isotopically labelled compounds were obtained Amersham, England. Actinomycin D was obtained Company, Rahway, N.J., U.S.A.

system. Samples on

from the Radiochemical Centre, through the courtesy of Merck &

3. Results (a) Characterization of phuge iW4 Since this phage has not been extensively used in previous work, we list below some of its chemical, physical and biological characteristics. The phage was isolated and described by Friedman & Cowles (1953). Some biochemical investigations carried out by Crawford (1958) suggested that M4, like the T-even series of coliphages, requires new proteins for synthesis of phage DNA. This phage contains cytosine and not hydroxymethylcytosine in its DNA. After equilibrium centrifugation in a caesium chloride solution, M4 phages appeared as a narrow band of buoyant density 1.487 g/ cc, and this treatment did not substantially reduce the plaque-forming titre (unpublished results). After dialysis against distilled water, preparations were negatively stained with potassium phosphotungstate and examined by Dr R. Horne in a Siemens Elmiskop I. Plate I shows the morphology of the phage particles and the high degree of purity of the preparation. One-step growth curves performed at 37°C by infecting a culture with seven phages per bacterium showed that the latent period was 40 to 60 minutes, the burst size about 100 phages per bacterium, and that lysis of bacteria was complete. Using lysozyme (200 pg1m.l. for two minutes at 37°C) to induce premature lysis, intracellular phages were found after 25 to 30 minutes. At this time, the rate of RNA synthesis (as measured by the incorporation of tritiated uridine into TCA-precipitable, alkali-labile material) was 5 to 10% that of uninfected cells growing in the same medium. These properties, and the absence of turbid plaques, suggest that M4 is a virulent phage. (b) Xedimentution analysis of pulse-lubelled RNA RNA prepared from cells infected for 25 minutes was sedimented through sucrose gradients. As shown in Fig. l(a), RNA from cells labelled for ten seconds sedimented between the peaks corresponding to 16 s ribosomal RNA and 4 s soluble RNA. A small amount of labelled RNA, corresponding to 20% of the total, overlapped with these components. After a pulse and 300 seconds, the labelled RNA was found in a slightly broader band (Fig. l(b)), indicating that some ribosomal or transfer RNA might have been made and accumulated. However, in uninfected cells labelled for only 30 seconds, the profile of radioactive RNA followed the optical density profile much more closely (Fig. l(c)), indicating that the labelled material was quite different t Abbreviation

used: TCA, triohloroaoetio acid.

226

M. SCHAECHTER (a) Infected: 10-w

1 Bottom

5

AND

(b) Infected: 300-x

pulse

IO

K. MoQUILLEN

15

20

25 Top traction

Cc)Uninfected: 30-set

1 Bottom no.

5

IO

pulse

I5

20 Top

pulse

I:~:~~

5 2

,dC

‘s.* 5

IO

15

Bdttom

20 Top

Fraction

no.

FIG. 1. Sediment&ion of RNA purified from infected and uninfected B. mega&-hm. (a) A culture was infeoted with phage and 26 min later labelled with [3H]uridine for 10 sec. (b) The same TV (a) but labelled for 300 set 20 min after infection, (c) Au uninfected oulture wee labelled with [3H]uridine for 30 sec. RNA waa extraoted from each, and samples were layered on a 4 to 20% sucrose gradient containing 0.06 M-N&I, 0.01 M-acetate buffer (pH 6.4). These were centrifuged for 6 hr (a), and 4 hr (b) end (0). 8t 37,000 rev&in. -9 0-D. OBO n& - - 0 - - 0 - -, 3H radioactivity.

in sedimentation properties from that found in infected cells. After 300 seconds of l&belling, the accumulation of ribosomal RNA in the infected cultures was 10% or lese of that seen in infected cultures. This is whst would be expected if the infected cells synthesized mainly mRNA, whereas all classes of RNA were being made in the uninfected cultures (see Discussion). (c) Stability of pult&ubelled

RNA

Cultures of strain KMTT were labelled with tritiated uridine for pulses of 30, 120 and 300 seconds terminating 25, 30 or 35 minutes after infection. Actinomyoin D (10 pg/ml.) was added and duplicate samples were alternately squirted into 5 y0 TCA and into O-5 N-NaOH. The latter were incubated at 37°C overnight, neutralized and brought to 5% TCA. Figure 2 shows the time-course of changes in the amount of radioactive material in the two series of samples. In this experiment, actinomycin D was added 30 minutes after infection. Radioactive material precipitable by TCA decreased, while the alkali-stable fraction increased. This increase was due to transfer of the label from the RNA into the cytosine of DNA, as shown by paper chromatography (unpublished experiments). The difference between the pairs of curves in Fig. 2 represents pulse-labelled RNA. A certain amount of this RNA did not decay in the presence of actinomycin D. This quantity, which varied from 0 for 30-second pulses, to 20% of the RNA in 300second pulses, was subtracted from the values for RNA at different times. The remaining values represent the labile component of the RNA, and are plotted as a

MESSENQER

.1

(a) 30-set

RNA

AND

pulse

POLYSOMES

(b) 120-set

0

2

4

pulse

6

8

IO

12

14

16

I8

2C1

Time (min)

I 0

.., 2

4

6

. 8 IO 12 Time (min)

14

., 16

I8

2C

FIG. 2. Fate of pulse-labelled

RNA after aotinomyoin D addition to phage-infeuted B. megotsriunr. Cultures were infected with phage and Isbelled with [*H]uridine for (a) 30 WC, (b) 120 BBO,and (c) 300 sea-up to 30 min of i&&ion. An excess (1 pg/ml. per min of labelling) of unlabelled was added and samples were squirted uridine wa8 added. At this time, actinomycin D (10 &ml.) into TCA (-a-e---) and NaOH et intervala, the letter being incubated at 37°C overnight, neutralized and treated with TCA (- - 0 - - 0 --). Both series were oounted on membrane l%ltera.

function of time in Fig. 3. By contrast, in analogous experiments with uninfected cells, 60% of the RNA labelled during a 300~second pulse was stable after addition of actinomycin D. RNA made during the 30 seconds decayed about twice as fast as the RNA made in 120 or 300 seconds. This was highly reproducible, and essentially the same results were obtained whether the antibiotic was added at 25, 30 or 35 minutes after the infection. Analogous experiments were carried out with uninfected cells. As shown in Table 1, the relationship between increasing pulse length and decay time was similar to that seen in infected cells. SOo/oof the RNA made in 10 to 15 seconds decayed in about 30 seconds. The time for comparable decay of RNA labelled for 120 seconds or more was about three times longer. Several experiments performed with [14C]uridine or [3H]adenosine gave similar results.

228

M. SCHAECHTER

AND

I I 4 6 Time (min)

I 2

Fro. 3. Decay

of pulse-labelled

RNA

K.

after

eddition

McQUILLEN

I 8

I IO

of actinomycin

D to phage-infected

B.

megaterium.

From the curves drawn in Fig. 2, differences at each point between the total TCA-insoluble radioactivity and the alkaline-stable, TCA-insoluble radioactive material, were calculated. In addition, the value for the alk8liJabile radioactive material remaining at the end of the experiment were subtracted from 8ll points. The points are thus corrected for stable RNA initially made and for DNA that is labelled before and after addition of the drug. The value at the time of addition of actinomycin D was taken aa 100%. (-m-m-) 300-set pulse; (-O-O-) 120-set pulse; (-*-a-) 30-set pulse.

(d) Sedimentation an&y&

of lysates

Cultures were labelled with tritiated uridine for lo- to 300~second pulses, all of which terminated 25 minutes after the infection, and lysates were centrifuged through sucrose gradients. Profiles from five such analyses are shown in Fig. 4. We estimated that at least 80% of the radioactivity was associated with ribosomes and polyribosomes. The “specific radioactivities” of the pulse-labelled material were computed by taking the 3H/32P ratios, where the 3H-labelled material was synthesized during the pulse, and the 32P-labelled material represents stable RNA (Fig. 5). At all times the 70 s fraction had the lowest specific radioactivity, whereas the highest specific radioactivity at early times was associated with the 80 to 150 s region. Later, there was a progressive increase and equalization of specific radioactivities in all polyribosome components.

4. Discussion The RNA metabolism of M4-infected B. megaterium cultures is very different from that of uninfected cells. Two sets of observations indicate that RNA made towards the end of the eclipse period includes little or no ribosomal and transfer RNA. First, analysis by sucrose gradient centrifugation of RNA isolated from cells pulse-labelled with uridine showed that most of the labelled material was lighter than 16 s ribosomal RNA, and heavier than 4 s soluble RNA (Fig. 1). The bulk of the labelled material

MESSENGER

RNA

AND

229

POLYSOMES

TABLE 1 Decay of pulse-labelled

Pulse length @=I

RNA

Radioactive precm%ors

in uninfected

B. megaterium

Time for decay of labile RNA @W to 50%

to 20%

10 10 10

0.45 0.57 0.76

1.0 2.1 1.8

15 15 16

0-60 0.75 0.78

1.7 1.9 2.0

30 30

O-78 0.88

2.2 2.5

45

1.05

3.4

60 60 60

1.20 1.26 1.40

8.0 3-2 4.7

120 120 120 120

1.45 I.53 1.52 2.00

3.6 4.9 5.1 6.0

180

1.56

7.0

300 300 300

1.45 1.60 1.80

4.4 5.2 5-6

Growing cultures of B. mgate&m were pulse-labelled for the times shown with the radioactive precursors indicated. U&belled precursors were added in excess (1 pg/ml. per min of the l&belling period). Actinomycin D (10 pg/ml.) was added to terminate the pulse. After subtracting the stable RNA remaining after 20 min treatment with the drug, the times required for decay to 60% and 20% of the labile RNA were computed.

M. SCHAECHTER

250

AND

I

1

I

(a) IO-set

K. MoQUILLEN

pulse

200. x .e .z (cl 30-set

::

(d) l20-SK

pulse

4 IOOOa”

- 5 1000

g

800-

-4 800.

600.

-3 600-

puke

-

-2 400-

I.

1

I 5

IO

I.5

20

25

Bottom

1 Bottom

5

30

1~

I

8 IO

TOP Fraction

no.

IO

20

I5

15

SoL.om5

25

20

25

30 TOP

30 TOP

Fraction

no.

Wa. 4. Sucrose gradient analyses of lysatea from phage-infected B. mqateriuna. Cultures were grown for three generations in the presence of [3aP]phosphate, infected with phage and labelled with [sH]uridine for (a) 10 see, (b) 20 WC, (c) 30 sec. (d) 120 BBC,and (e) 300 see-up to 25 min of infection. Lyeates were prepared and sedimented as indioated in the text. Fractions were collected, preoipitated with cold 6% TCA and counted on membrane fltere. The specific radioaotivity of the:lP was the came in each experiment,“that of the [3H’juridine was the eeme --0 --) 3H radioin (e), (b) and (c), but wae reduced byj1/50 in (d) and by l/ZOO in (e). (--0 activity; (-a-*-) 3aP radioaotivity.

MESSENGER

RNA

AND

I

I 280 Sedimentation FIG.

231

POLYSOMES

5. Sucrose gradient

140 70 210 coefficient (Svedberg units)

analyeee of lysates from phage-infected

B. megaterium.

Speci& activities were derived from the data shown in Fig. 4 by taking the ratio of sH/3aP throughout the gradients. These were calculated for [3H]uridine pulses of 10, 20, 30, 120 and 300 sec. The specific activities have been normalized by taking a value of l-00 for the 70 B fraction.

sedimented between these components, in a manner similar to the mRNA of T4infected Escherichia coli (Asano, 1965). Second, the RNA synthesized during a 30-second pulse decayed completely in the presence of actinomycin D. 20% of the material labelled for 300 seconds was stable in the presence of this drug (Fig. 2). By contraat, in experiments with uninfected cells, 60% of the RNA labelled during a 300~eecond pulse was stable. In addition, the synthesis of ribosmes or ribommeprecursor particles was almost completely inhibited. In uninfected growing cells labelled with uridine for 30 seconds, about three-quarters of the label is found in material sedimenting at about 50 s (Schaechter et al., 1965) which probably corresponds to ribosome precursors or “neosomes” (McCarthy, Britten & Roberts, 1962). In infected cells, very little label was found in such material. Moreover, the specific radioactivity of 70 s ribosomes did not perceptibly change after labelling the cells for different times (Fig. 4). Since the synthesis of stable ribosomal and transfer RNA was greatly reduced, we believe that mRNA constituted the bulk of the RNA labelled over a short period of time during the late stages of phage infection. This system thus resembles Escherichia c&i infected with T-even phages (e.g. Nomura, Hall & Spiegelman, 1960). Our operational definition of mRNA is in terms of kinetics and sedimentation behaviour, rather than information content of the molecules, and does not take into account that different functional molecules may be of different lengths. On this basis, our results permit the following conclusions. (1) Ribosomes become rapidly attached to mRNA chains labelled during a short (10 seconds) pulse (Fig. 4). 16

232

M. SCHAECHTER

AND

K.

MoQTJILLEN

(2) This rapidly labelled mRNA is associated mainly with small polysomes (containing 1 to 3 ribosomes), although some is found in polysomes containing more than six ribosomes (Fig. 4). (3) The ratio of mRNA to ribosomes becomes constant for all polysomes after 30 to 120 seconds of labelling (Figs 4 and 5). This can be considered to be the mean time for the “maturation” of polysomes, i.e., the time required for the synthesis and complete loading with ribosomes of an average mRNA molecule. The constancy of the mRNA to ribosome ratio suggests that in phage-infected B. megaterium, as in yeast (Marcus, Bretthauer & Halvorson, 1965), the spacing of ribosomes on mRNA is, on the average, the same for polysomes of different size. (4) The time of labelling affects the kinetics of decay of mRNA in the presence of actinomycin D. Within 30 to 120 seconds the decay kinetics change from an initial rapid value to one which is slower and constant (Figs 2 and 3). We assume that most mRNA chains labelled during a ten-second pulse are in the process of being synthesized and are still attached to DNA. This is suggested by our :finding that, although the average time for the assembly of polysomes is 30 to 120 seconds, some label is found in polysomes containing more than six ribosomes after only ten seconds of labelling. Either the mRNA of these polysomes was in the course of being synthesized during the span of the pulse, or they represented a class that was assembled de novo in a time much shorter than the average. Similarly, relatively long times for the synthesis were reported for the mRNA of the tryptophan operon (several minutes at 37°C; Imamoto, Morikawa & Sato, 1965), and of p-galactosidase (2-5 minutes at 30°C; Leive, 1965a). Based on these estimates, our results indicate that the free, first-synthesized portions of the mRNA molecules are associated with ribosomes before the molecules are completed. This is in agreement with the hypothesis of Stent (1966), that translation of a nascent mRNA molecule is required for continuing its synthesis. However, our finding that mRNA labelled for short times preferentially associates with few ribosomes suggests that it may not be maximally loaded with ribosomes. This may be the consequence of a rapid attachment of the first ribosome or two, followed by a slow rate of movement along the mRNA molecule. The increase in decay time in the presence of actinomycin D with time of labelling may be due to one of two things: either, as Brenner (1965) hassuggested, some mRNA is normally destroyed before it can be translated, or, the presence of the drug induces the preferential breakdown of the most recently made portions of mRNA. Actinomycin D may induce the detachment of this portion of the molecules and render them spuriously sensitive to decay. This is a distinct possibility, since this drug appears to induce the detachment from DNA of RNA polymerase (Goldberg, Reich & Rabinowitz, 1963), which normally seems to associate the nascent RNA molecules to the DNA (Bremer & Konrad, 1964). The older portions of the molecules can then be expected to decay with normal kinetics since they are relatively protected by ribosomes. Stabilization of mRNA by attachment to ribosomes was shown by Byrne, Levin, Bladen & Nirenberg (1964) in vitro. Fan, Higa & Levinthal(1964) have shown that if protein synthesis is interrupted by anaerobiosis, or by the addition of various inhibitors, pulse-labelled RNA is protected in vivo from decay in the presence of actinomycin D; they interpret this protection as being due to the attachment of mRNA to ribosomes. Similar differences in the kinetics of decay of RNA labelled for various times were also observed by us in uninfected B. megaterium. Here the faster

MESSENGER

RNA

AND

POLYSOMES

233

decay of RNA labelled

for very short times indicates that newly mude ribosomal RNA is also unstable in the presence of actinomycin D. Calculation of the kinetics of decay of labile RNA in E. cdi reported by Leive (1965b) also reveals these differences. Note added in proof: The early appearance of mRNA in small polysomes has also been found by Dresden & Hoagland (J. Biol. Cltem., in the press) in E. coli recovering from glucose starvation. We thank Miss Susan Foster for cheerful technical assistance. One of us [M. S.] holds a career development award from the National Institutes of Health, U.S. Public Health Service. This investigation was partly supported by grant AI 06103 of the National Institutes of Health, U.S. Public Health Service. REFERENCES Asano, K. (1966). J. Mol. Biol. 14, 71. Bray, G. A. (1960). Analyt. Biochem. 1, 279. Bremer, H. & Konrad, M. W. (1964). Proc. Nat. Acud. Sci., Wash. 51, 801. Brenner, S. (1966). Brit. Med. Bull. 21, 244. Byrne, R., Levin, J., Bladen, H. A. & Nirenberg, M. W. (1964). Proc. Not. Acud. Sci., Wash. 52, 140. Crawford, L. V. (1968). B&him. biophys. Acta, 28, 208. Fan, D. P., Higa, A. & Levinthal, C. (1964). J. Mol. BioZ. 8, 210. Friedman, M. & Cowles, P. B. (1953). J. Bact. 66, 379. Goldberg, I. H., Reich, E. & Rabmowitz, M. (1963). Nature, 199, 44. Imamoto, F., Morikawa, N. t Sato, K. (1966). J. Mol. BioZ. 13, 169. Kaiser, A. D. & Hogness, D. S. (1960). J. Mol. BioZ. 2, 392. Leive, L. (1965a). Biochem. Biophys. Res. Comm. 20, 321. Leive, L. (19653). J. Mol. BioZ. 13, 862. Levinthal, C., Keynan, A. & Higa, A. (1962). Proc. Nat. Ad. Sci., Wash. 48, 1631. McCarthy, B. J., Britten, R. J. & Roberts, R. B. (1962). Biophye. J. 2, 57. McQuillen, K. & Roberts, R. B. (1954). J. BioZ. Chem. 207, 81. Marcus, L., Bretthauer, R. & Halvorson, H. (1965). Science, 147, 615. Nomura, M., Hall, B. D. & Spiegehnan, S. (1960). J. Mol. BioZ. 12, 306. Schaechter, M., Previc, E. P. t Gillespie, M. D. (1966). J. MOE. BioZ. 12, 119 Scherrer, K. & Darnell, J. E. (1962). Biochem. Biophya. Res. Comm. 1, 486 Stent, G. S. (1966). Proc. Roy. Sot. B, 164, 181.