J. Mol. Biol. (1967) 24, 417427 Rate of DNA Synthesis during the Division of Escherichia coli B/r Cycle CEURLES E. HELMSTETTER Roswell Park Memorial Institute Buffalo, New York, U.S.A. (Received4 October1966) The rate of DNA synthesis during the division cycle of Eacherichia coli B/r has been measured in exponentially growing cultures. This was accomplished by pulse-labeling exponential phase cultures with [14C]thymidine and measuring the amount of label incorporated into cells of different ages in the cultures. The amount of label in cells of different ages was found by binding the bacteria to the surface of a membrane filter at the end of the labeling period and counting the radioactivity in the cells which were eluted continuously from the membrane. Since primarily new-born cells are eluted from a membrane-bound culture of E. coli B/r, the radioactivity per effluent cell during each generation of elution was a measure of the amount of [14C]thymidine incorporated into the oldest through the youngest cells in the exponential phase culture. This experimental approach was considered to be superior to those employing synchronized cells for these studies, because all manipulations were carried out after, rather than before, the incorporation took place. It was found that there is an abrupt increase in the rate of DNA synthesis during the division cycle of cells growing in glucose, glycerol, suocinate or acetate minimal medium which occurs at a cell age, in fractions of a generation, of 0.4, 0.8, O-9 and 0.9, respectively. In succinate- and acetate-grown cells, this increase is preceded by a pronounced minimum in the rate of [14C]thymidine incorporation during the last half of the division cycle. It is suggested that the period of reduced thymidine incorporation corresponds to a gap between rounds of DNA replication, and that the abrupt increase corresponds to the start of a new round of DNA replication. 1. Introduction Synchronously dividing populations of Esch&chia coli have frequently been used to study the relationship between DNA synthesis and the division cycle of this organism. However, serious reservations concerning this approach have been expressed; the principal one being that the procedures used to obtain synchrony might alter the pattern of DNA synthesis (Schaechter, Bentzon & Maalee, 1959; Abbo t Pardee, 1960; Maalee, 1962). These reservations appear well justified in view of the conflicting data which have been reported. In early studies on synchronized cultures of E. coli in glucose-salts medium, DNA synthesis was found to occur during only a portion of the division cycle (Barner & Cohen, 1956; Maruyama, 1956; DeLamater, 1956). However, in more recent experiments, in which synchronization procedures that were considered to have less effect on cellular metabolism were used, essentially continuous synthesis was found (Abbo & Pardee, 1960; Nagata, 1963; Cummings, 1965; Clark & Maalee, 1967). Abbo & Pardee (1960) presented evidence that the rate of DNA synthesis increases exponentially during the division cycle of synchronously 417 418 C. E. HELMSTETTER dividing E. coli B growing in glucose-salts medium. They obtained the synchronous population by a filtration procedure which was designed to collect the smallest cells from an exponential phase culture. More recently, Clark & Maalae (1967) found that the rate of DNA synthesis doubles in the middle of the division cycle during synchronous growth of E. coli B/r in glucose medium. In their experiments synchronous growth was obtained by the membrane-selection technique of Helmstetter & Cummings (1964), which selects the youngest cells from a growing culture. In this report an alternative method for studying the cell cycle with the membraneselection technique is described. This alternative approach is based on the idea that, since the effect of any synchronization treatment on cellular metabolism is uncertain, it would be preferable to study the kinetics of DNA synthesis in untreated, exponentialphase cells. This could be accomplished in theory by pulse-labeling an exponential phase culture with radioactive thymidine and measuring the incorporation of t,he label into cells of different ages. In order to perform such an experiment, a method for sorting out cells of different ages would be required. We have used the membraneselection technique for this purpose. In brief, the rate of DNA synthesis was measured as a function of cell age by: (1) pulse-labeling an exponentially growing culture of E. coli B/r with radioactive thymidine, (2) binding the labeled cells to a membrane filter, and continuously (3) measuring from the the membrane radioactivity in as described the in new-born a previous cells report which were (Helmstetter eluted $ Cummings, 1964). Thus, the incorporation of radioactive thymidine into cells of different ages in the exponential-phase culture was determined by measurement of the radioactivity in their progeny. Although the utility of this approach for studying the cell cycle depends upon the accuracy with which new-born cells can be identified with their parents, the advantage of the method over synchronous growth studies is that the uncertainty as to the physiological state of the cells at the time of labeling has been eliminated. 2. Materials (a) Bacteria and Methods and growth conditions The organism used was E. coli B/r (ATCC 12407). The minimal salts medium contained 2 g NH,Cl, 6 g Na2HP04, 3 g KHzP04, 3 g NaCl and 0.25 g MgSO, in 1 liter of distilled, demineralized water. Glucose, glycerol, sodium succinste or sodium acetate was added at 1 g/liter for the appropriate experiments. Each experiment was performed on a loo-ml. culture which had incubated for approximately 17 hr at 37°C with shaking, and was at a concentration of 1 x 10s cells/ml. in exponential growth. After these cells were bound to the membrane filter, they were washed and eluted with medium which had previously supported growth to 1 x lOa cells/ml., and had been filtered through a grade HA Millipore filter (Millipore Filter Corp.). This filtered medium will hereafter be called conditioned medium. (b) Apparatus The procedure for continuous withdrawal of new-born cells from a culture attached to the surface of a membrane filter has been described previously (Helmstetter & Cummings, 1964). The cells were bound to a grade GS (0.22 p pore-size), 152-mm diameter Milliporo membrane filter which was clamped in a two-section porcelain filter apparatus (Carl Schleicher & Schuell Co., type PA 15). The support plate in this apparatus was replaced with a stainless-steel support screen (Millipore Filter Corp., cat. no. YY2214264), and the porcelain upper section with several rubber gaskets which were supplied with the apparatus. After filtering the cells onto the membrane, the assembly was inverted and placed on top of an additional porcelain bottom section to maintain constant temperature on the membrane. Conditioned medium was then poured into the top. The elution rate was DNA SYNTHESIS IN E. COLI 419 governed by connecting the top to a reservoir of conditioned medium through a model AL-4-E Sigmamotor pump (Sigmamotor, Inc.). The stems of both bottom sections had been cut off to facilitate handling and to permit the effluent to drop freely from the membrane. The entire procedure was performed in a 37°C Full View incubator (Precision Scientific Co.). (c) Experimental procedure [‘%]Thymidine (30 me/m-mole, Nuclear Research Chemicals, Inc.) was added to a IOO-ml. culture when it had reached a concentration of 1 x lOa cells/ml. in exponential growth. After brief exposure (1 to 2 min) to the label, the culture was poured onto the membrane and filtered under vacuum. Filtration was stopped when almost all of the fluid had passed through the filter, and the remaining portion was poured off. The cells, now bound to the membrane, were washed by passing 100 ml. of conditioned medium through the filter in the same manner. The beginning of the washing period was considered to be the termination of the labeling period and also the zero time of elution. The filter apparatus was then inverted and elution was begun by pouring conditioned medium into the top and starting the pump. The total time required for these operations was approximately 45 sec. Samples of the effluent were collected consecutively during constant time-intervals, and 0.5 ml. was removed for measurement of the bacterial concentration with a Coulter Counter model B as described previously (Helmstetter & Cummings, 1963). Each sample of the effluent was precipitated with cold trichloroacetic acid at 5% final concentration and kept in an ice bath for 30 min. The samples were then collected on 25mm diameter membrane filters, washed with cold 5% trichloroacetic acid containing 100 pg of unlabeled thymidinc/ ml. and dried. The radioactivity of the membrane was determined in Liquifluor (Nuclear Chicago Corp.) with a Nuclear Chicago liquid scintillation system. 3. Experimental (a) Description Design of the procedure Figure 1 is a schematic outline of the procedure for of different ages in an exponentially growing culture Step I (pulselabeling) o.gsogj Step to & mMembrane] the rate at which a macromolecule. cells For 1 It (binding) r get!Zatioy (elution) :Z,:fce”S C [I‘-achlnmt Step determining synthesize III Start of elution + FIG. 1. Outline of the procedure for determining the into cells of different ages in an exponentially growing 23 TtiT rate of incorporation culture. of a labeled molecule 420 C. E. HELMSTETTER simplicity, four representative cells from an exponential-phase culture are followed through the three steps of the procedure. The ages of the cells are shown as fractions of an interdivisional period. In step I the cells are pulse-labeled with a radioactive precursor of a hypothetical macromolecule, and the amount of label which is assumed to be incorporated during the pulse is indicated by the letter in each cell. The cells are then immediately bound to a nitrocellulose membrane by filtration in step II. After inverting the membrane, elution is begun by pouring minimal medium on top of the membrane. The subsequent growth of the four cells and their progeny during each generation time of elution is shown in step III. The division stages of the cells at four distinct times during each generation of continuous elution are shown beginning at the top and progressing clockwise. The arrows indicate the cells which would appear in the effluent, based on the observation that primarily new-born cells are eluted from a membrane-bound culture of E. co& B/r (Helmstetter & Cummings, 1964). It is assumed that the new-born cells elutetl are daughters which did not form part of the attachment of their parents to the membrane. The amount of label in each new-born cell is indicated. The letter “11” symbolizes the filial relationship between the eluted cell and the original bound cell; i.e., during the first generation of elution “n” equals one, during the second it equals two, etc. At the beginning of elution (twelve o’clock in step III), the new-born cell eluted is a daughter of the cell which wss at age 1.0 initially and it contains D/2 units of label. After 0.25 generation of elution, each cell has progressed 0.25 unit in age, and the new-born cell eluted from the membrane at this time is a daughter of the cell which was initially at age 0.75. It contains C/2 units of label. This sequence of events continues until all of the cells initially attached to the membrane have divided, and is then repeated during the second and succeeding generations of elution. In summary, the cells eluted from the membrane during each generation are progeny of the oldest through the youngest cells init,ially attached, and each of these contains half of the label in its pa,rent. Thus, the rate of incorporation of a labeled compound into cells of different ages in a population is dct,ermined by pulse-labeling the cells and measuring the amount of label in their progeny. This experimental procedure is quite similar to the radioautographic method for studying the period of DNA synthesis which was first described by Howard & Pelt (1953) and which has been used in various forms by a number of investigators. In their experiments, as in those to be described here, cells were labeled with a precursor of DNA, but the DNA synthetic period was analyzed by observing the subsequent appearance of labeled mitotic figures by radioautography rather than by withdrawing cells of a particular age from the culture. For clarity, it has been assumed that when a cell divides on the membrane one daughter is eluted and the other remains attached. This has been indicated by showing the cells attached to the membrane by one end. Actually most cells may be attached longit,udinally, and, as a result, both daughters may remain attached when some of the cells divide. Therefore, the number of new-born cells which remain attached may be greater than the number which are eluted; but this effect would not alter interpretation of the data presented in this report. It has also been assumed that the label is equally distributed between the daughter cells. This need not be the case and, in fact, the method could be used to determine the presence of oriented or non-equal partition of the label. (b) Theoretical considerations To facilitate interpretation of data obtained with this procedure, a hypothetical experiment will be described. First, the concentration of cells .in the effluent from a membrane-bound population of bacteria with a particular age distribution will be considered. Figure 2(a) shows the ideal distribution of cell ages in an exponentially growing culture. If the ages of the bound cells were also distributed in this manner, and if elution proceeded as shown in Fig. 1, then ideally the concentration of cells in the effluent during the first two generations of elution would appear as shown in Fig. 2(b). Since the cells which are eluted during each generation are progeny of the oldest through the youngest cells initially bound to the membrane, the elution curve during each generation is the reverse of the age distribution curve. Again it has been assumed that when a bacterium divides on the membrane one daughter is eluted and the other remains attached. If less DNA SYNTHESIS than one-half of the new-born cells would be greater in each succeeding membrane would increase. E. COIL1 IN were eluted, generation, 421 the concentration since the number of cells in the of cells bound effluent to the i”\ll~~~ 0 0.50 Cell 0 I.0 050 age of a generation) (fraction Elution I.0 time (a) I.50 2.0 (generations) (b) FIG. 3. (a) Idealized age distribution in an exponential phase culture containing no dispersion in generation times of individual cells. (b) Theoretical concentration of cells in the effluent from a membrane-bound culture with an age distribution as shown in (a). $.” 2.0 2: 2s Fi? 2% m B I.0 A o 050 Cell (fraction of I.02 age 0 0.50 El ution I I.0 time I.50 2.0 (generations) a genemtion) (a) (b) FIG. 3. (a) Rate of synthesis of two hypothetical macromolecules a8 a function of cell age. (b) Theoretical radioactivity per cell in the effluent from a membrane-bound culture if it were pulse-labeled with radioactive precursors of the macromolecules described in (a). Next suppose that the rate of synthesis of two hypothetical macromolecules as a function of cell age is given by Fig. 3(a). That is, the rate of synthesis of molecule A doubles in the middle of the division cycle, whereas the rate of synthesis of molecule B increases exponentially during the division cycle. Figure 3(b) shows the results expected if this population were pulse-labeled with radioactive precursors of molecules A and B, and then analyzed with bhe elution procedure. Curve A in Fig. 3(b) shows the radioactivity per effluent cell due to incorporation of the precursor of molecule A. For the first half of the first generation of elution, the radioactivity per effluent cell is constant, since these cells would be progeny of cells which were between the age 1.0 and 0.50 at the time of pulselabeling. During the last half of the first generation of elution, the radioactivity per cell is one-half the initial value since these cells would be progeny of cells which were at age 0.50 through 0 initially. This same pattern is repeated in the second generation of elution, but at corresponding elution times the radioactivity per effluent cell is equal to one-half that of the first generation. Similarly, the radioactivity per effluent cell due to incorporation of the precursor of macromolecule B would be expected to decrease exponentially as is shown by curve B of Fig. 3(b). Consequently, the rate of synthesis of these macromolecules as a function of increasing cell age is shown from right to left in each generation of elution. C. E. 422 HELMSTETTER 4. Results (a) E&ion of cells from a membrane-bound population When an exponentially growing population of E. coli B/r is filtered onto the surface of a membrane filter, and the membrane is inverted and conditioned medium is passedthrough it, primarily new-born cells are eluted from the membrane (Helmstetter & Cummings, 1964). Figure 4 shows the number of bacteria in consecutive samplesof the effluent from membrane-bound cultures of E. coli B/r in minimal-salts medium containing glucose, glycerol, succinate or acetate as the carbon source. In this and the next Figure, the time scalesare arranged so that the positions of the peaks in the bacterial elution curves coincide, and the ordinates are graduated in logarithmic scalesto ease visualization of the periodicity in the curves. As noted I I I I 1 4.0 - 20 I.0 Glucose ;c 4.0 80 40 120 2.0 4) ‘-0 I.0 x 0 Glycerol ;c 0 100 50 2 z 4.0 -z v 20 Succinate Id 0 210 70 2.0 i I.0 Acetate 0.5 :;:-II_ 0 210 105 Elution time 315 (mln) Fro. 4. Number of cells in consecutive samples of the effluent from membrane-bound cultures of E’. coli B/r growing in minimal-salts medium containing glucose, glycerol, succinate or acetate as carbon source. In each experiment a loo-ml. culture containing 1 x 108 cells/ml. was filtered o&o the membrane. The rate of elution with the four media was 6.5 ml./min, 65 ml./min, 5 ml./min and 5 ml.,’ min, respectively. The sampling periods are indicated by the horizontal bars in the histograms, and were 4 min in glucose and glycerol, and 5 min in succinate and acetate medium. DNA SYNTHESIS IN E. COLI 423 above, the shape of the elution curves mainly reflected the age distribution of the bacteria initially bound to the membrane and the efficiency with which the nemborn cells were eluted from the membrane. The shapes of the curves are similar t,o those anticipated based on the theoretical exponential curve shown in Fig. 2(b), except for the effect of dispersion in division times and the increase in bacterial concentrations with elution time. The upward trend of the curves indicates that fewer than half of the new-born cells are eluted from the membrane, especially during the first generation of elution. (b) Incorporation of [‘“Cjthymidine into cells of different ages The rate of incorporation of [14C]thymidine into cellsof different agesin populations of E. coli B/r growing in glucose, glycerol, succinate or acetate medium was determined by the method described. Figure 5 shows the radioactivity per efiuent cell I 4.0: ,hL; I I I 2.0 I I I i I I I I I ;Succinate 1 iAcetate Elution I20 I I timetmin) Fro. 5. Radioactivity per cell in the effluent from membrane-bound cultures of E. coli B/r which had been pulse-labeled with [lW]thymidine. Cells growing in glucose, glycerol, succinate and acetate medium were exposed to [‘%]thymidine at 0.1 PC/ml. for 1 min, 0.1 PC/ml. for 2 min, 0.1 PC/ml. for 2 min and 0.15 +/ml. for 2 min, respectively. The cultures were bound to membrane filters as described in Materials and Methods and washed and eluted with conditioned medium containing the same carbon source. 424 C. E. HELMSTETTER during elution of the four membrane-bound cultures which had been pulse-labeled with [14C]thymidine. The vertical broken lines indicate approximately the elution time in generations. The generation times were determined by measuring the time interval between peaks of the bacterial elution curves obtained in each experiment. In glucose-grown cells, there was little change in the cts/min per effluent cell during elution of the progeny of those cells which were between the end and the middle of their division cycle during the pulse-labeling period. The radioactivity in the progeny of cells which were between the middle and the beginning of their division cycle was 1st Generation 1 I 0 2nd Generation Gly,cerol 4.0 Ia0 2 .-5 O 5 3.0 eQ) m P 2,o 2 60 4.0 )li- . 3.0 I.5 2-o I.0 I.0 0.5 0 (fractions 05 I.0 2 0 0 0.5 I*0 Cell age of a generation) Fm. 6. Rate of [‘Wlthymidine incorporation during the division cycle of E. coli B/r grown in glucose, glycerol, succinate and acetate minimal medium. This Figure is a replot of the cts/min/106 effluent cells in the first two generations of elution in Fig. 5 as a function of the age of the parents (first generation) and grandparents (second generation) of these cells at the end of the pulse-labeling period. The first generation in this Figure is a plot of the data between the first vertical dashed line and the origin in Fig. 5 (i.e., the first generation of elution in reverse order), and the second generation is data between the second and the first verticd dashed lines in Fig. 5. DNA SYNTHESIS IN E. COLI 423 also roughly constant, but equal to about O-6of the cts/minin the progeny of the older cells. Therefore, the rate of [14C]thymidine incorporation into cells which were in the last half of their division cycle was approximately double that into cells which were in the tlrst half of their division cycle. Figure 6 showsthe results of the sameexperiments, but with the radioactivity of the effluent cellsduring the first two generations of elution plotted as a function of the age of their ancestors at the time of pulselabeling. Therefore, this Figure showsthe rate of [14C]thymidine incorporation during the division cycle directly. The results show that the rate of thymidine incorporation increasesabruptly around the middle of the division cycle of glucose-grownE. coli B/r. Similarly, the results in Figs 5 and 6 indicate that in glycerol-grown cells the rate of incorporation of [l*C]thymidine increases abruptly between the middle and the end of the division cycle. The results with succinate- and acetate-grown cells indicate a period of reduced [14C]thymidine incorporation during the last half of the division cycle, followed by a rapid increase in incorporation just before the end of the cycle. Finally, it can be seen from Fig. 6 that the cts/min/106effluent cells in the second generation of elution were not equal to O-5of that in the first generation, as would be expected theoretically, but they were greater-about O-6 of the first generation. I cannot explain this discrepancy at present, and in the interpretation of these experiments I rely more on the shapesof the incorporation patterns than on the absolute values. 5. Discussion The rate of thymidine incorporation into cells of different ages in exponentially growing cultures of E. coli has been determined. This was accomplished by pulselabeling the cultures with [14C]thymidine and measuring the radioactivity in the new-born cells released from the cultures after they were bound to the surface of membrane filters. The rate of incorporation of [14C]thymidine into cells of different ages was considered a measure of the rate of DNA synthesis during the bacterial division cycle. It must be noted that the influence of relevant intracellular pools on the incorporation of [14C]thymidine during the pulse-labeling period and during the initial elution period has not been determined. However, Clark & Maaloe (1967) measuredthe time course of thymidine incorporation into bacterial samplestaken at various times during synchronous growth, and they found linear incorporation curves which extrapolated to 30 secondson the time axis in each case. This result suggeststhat there is little, if any, significant change in pool-size during the division cycle of E. coli B/r. It was found that the rate of DNA synthesis increasesabruptly at a time in the division cycle which is characteristic of the carbon source in the medium in which the cells were grown. We will consider that the age at which the rate of DNA synthesis increasesin the average cell can be estimated by determining the age which correspondsto the mid-point of the increase in radioactivity per eflluent cell in the second generation in Fig. 5 or Fig. 6. The secondgeneration was chosenbecausethe first and third generations are incomplete. On this basis,the cell age in fractions of a generation at which the rate of DNA synthesis increases is 0.4 in glucose-, 0.8 in glycerol-, O-9 in succinate- and O-9 in acetate-grown cells. Thus, the increase occurs about 24 minutes before division in glucose-grown cells and about 10 minutes before division in glycerol-, succinate- and acetate-grown cells. The results in glucose-grown cells support the finding of Clark & Maalee (1967) that the rate of DNA synthesis doubles 426 C. E. HELMSTETTER around mid-cycle during synchronous growth of E. coli rather than increasing exponentially during the cycle as reported by Abbo & Pardee (1960). DNA synthesis was found to occur throughout the division cycle of cells growing in glucose minimal medium. This is in agreement with previous experiments which demonstrated essentially continuous DNS synthesis in E. coli growing exponentially in glucose minimal medium by the use of radioautography following brief exposure to [3H]thymidine or [3H]thymine (Schaechter, Bentzon & Maaloe, 1959; Lark 8: Lark, 1965) and by measuring the loss of viability after incorporation of [3H]thymine (Pachler, Koch & Schaechter, 1965). The results also indicate that the abrupt increase in thymidine incorporation near the end of the division cycle of glycerol-, succinateand acetate-grown cells is preceded by a period of lesser incorporation with respect to the first half of the cycle. The decrease in thymidine incorporation was very slight in glycerol-grown cells, and more pronounced in succinnte- and acetate-grown cells. This period of reduced incorporation presumably corresponds to the interval between rounds of DNA replication in E. coli suggested by Maalee & Hanawalt (1961) and Maalee (1961), and observed in the radioautographic experiments of Lark (1966). Lark found a period devoid of DNA synthesis in slow-growing cells by briefly exposing cultures of E. coli ETto [3H]thymine and measuring by radioautography the fraction of cells which incorporated the label. The results reported here show the same basic effect of slow growth except that the gap in DNA synthesis in E. coli B/r is evident at faster growth rates in comparison t)o Lark’s findings with E. coli 15’l!-. One major difference is that Lark’s data indicate that the gap occurs during the first half of the division cycle, whereas the results described in this report show that it occurs during the last half of the cycle of E. coli Bjr. Unfortunately, this apparent difference cannot be invest,igated with the elution technique since E. coli 15T- does not bind properly to the membrane. If the period of reduced thymidine incorporation corresponds to a gap between rounds of DNA replication, then the time of the abrupt increase in the rate of incorporation at the end of this period corresponds to the time in the cell cycle when each new round of DNA replication commences. Clark & Maaloe (1967) investigated t,he time of initiation of DNA replication in glucose-, glycerol- and succinate-grown cells by measuring the extent of incorporation of thymidine in the presence of chloramphenicol by cells in different stages of synchronous growth. Since it was assumed that protein synthesis is required to initiate DNA synthesis, the time of synthesis of the protein required for initiation of DNA replication would be determined by their experiments. They found that replication is initiated in glucose cells at the same time that the rate of DNA sYynthesis doubles, and that initiation occurs at about the same time prior to division (20 to 25 minutes) in glycerol and succinate cells. This time is slightly earlier in the division cycle of glycerol and succinate cells than the time before division that a new round of DNA replication begins as reported here. This could mean that the protein required for initiation is synthesized a few minutes before a new round of DNA replication begins in slow-growing cells, but to be certain of this apparent time difference both experiments would have to be performed under the same conditions. The major advantage of the method presented here is that the labeling took place in an unaltered exponential-phase culture and all manipulations were carried out after, rather than before, the incorporation had occurred. In fact, the reason I have adopted this approach is that I have observed unusual and apparently distorted DNA SYNTHESIS IN E. 427 COLI patterns of thymidine and u&line incorporation into synchronous populations obtained from membrane-bound cells. These observations, along with a comparison between results obtained using both applications of the membrane-selection technique, will be presented elsewhere. With regard to the possible sources of error in the present technique, it should be noted that I have assumed that the attachment of log-phase cells to the surface of the membrane did not distort the sequence in which cells of different ages in the culture would have divided. In addition, it was assumed that there was no delay in cell division at the beginning of elution, and that new-born cells were &ted from the membrane at the moment of division. If a delay of either kind occurred, the time in the division cycle at which a round of DNA replication was found to begin in these experiments would be earlier than the observe& ime. Although we have no evidence to suggest the possibility of errors of this kind, further experiments with this method will be necessary to establish the validity of these assumptions. I wish to express my sincere thanks to Dr Ole Maalee ment during my visit in his laboratory, where many were developed. Special thanks are due to Dr Stephen suggestions from the very beginning of this investigation. Pierucci for many helpful discussions and to Evadne technical assistance. This investigation was supported by U.S. Public CA 08232 from the National Cancer Institute. for his hospitality and encourageof the ideas for this investigation Cooper for his numerous valuable I am also indebted to Dr Olga Chin and Eras Revelas for expert Health Service Research grant REFERENCES Abbo, F. E. & Pardee, A. B. (1960). Biochim. biophys. Acta, 39, 478. Barner, H. D. & Cohen, S. S. (1956). J. Bact. 72, 115. Clark, D. J. & Maalee, 0. (1967). J. Mol. Biol. 23, 99. Cummings, D. J. (1965). Biochim. biophys. Acta, 85, 341. DeLamater, E. D. (1956). Symp. Sot. Gen. Microbial. 6: 215. Hehnstetter, C. E. & Cummings, D. J. (1963). Proc. Nat. Acad. Sci., Wash. 50, 767. Helmstetter, C. E. & Cummings, D. J. (1964). Biochim. biophys. Acta, 82, 608. Howard, A. & Pelt, S. R. (1953). Heredity, 6 (Suppl.), 261. Lark, C. (1966). Biochim. biophyiya. Acta, 119, 517. Lark, K. G. & Lark, C. (1965). J. Mol. Biol. 13, 105. Maaloe, 0. (1961). Cold Spr. Harb. Symp. Quad. Biol. 26, 45. Maalee, 0. (1962). In The Bacteria, cd. by I. D. Gunsalus 8z R. Y. Stanier, vol. IV, New York: Academic Press. Maalee, 0. & Hanawalt, P. G. (1961). J. Mol. Biol. 3, 144. Maruyama, Y. (1956). J. Boxt. 72, 821. Nagata, T. (1963). Proc. Nat. Acad. Sci., Wash. 49, 551. Pachler, P. F., Koch, A. L. & Schaechter, M. (1965). J. Mol. Biol. 11, 650. Schaechter, M., Bentzon, M. W. & Maalee, 0. (1959). Nature, 183, 1207. p. 1.