Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Journal of General Microbiology (1 990), 136, 599-606. Printed in Great Britain 599 Review Article Cell growth and abrupt doubling of membrane proteins in Escherichia coli during the division cycle HERBERT E. KUBITSCHEK-~ Biological and Medical Research Diuision, Argonne National Laboratory, 9700 South Cass Auenue, Argonne, IL 60439, USA Regulation of cell growth is a major biological mystery. Even the kinetics of cell growth, although studied for many years, remain elusive. This state of affairs reflects, in large part, the fact that cell growth represents the interaction of a large set of complex and interrelated biochemical processes that allow the cell to survive and respond to a tremendous variety of challenges and stimuli. Thus, it might seem at first that studies of cell growth during the cellular division cycle may be of little help. Nevertheless, interest in cell growth patterns has continued to increase because such studies reveal the presence of major processes that contribute to the regulation of cell growth and may also provide clues to their biochemical identification. Methods and criteria Cell growth has been studied far more extensively in Escherichia coli than in any other prokaryotic or eukaryotic cell type. Several experimental approaches have been used, including studies of growth in individual cells, synchronized cultures, or exponential-phase populations. Because each of these approaches has produced a variety of results and conclusions for the kinetics of cell growth during the cell cycle in this micro-organism, no common consensus on the pattern of cell growth has developed, much less an understanding of its regulation. The slow progress in this area reflects both the complexity of growth and the technological difficulties in its study, as well as the involvement of semantic differences. For example, in these different studies, ‘growth’ has variously been defined and measured as increase in cell length, surface area, volume, mass, or content of cellular RNA and protein. The accumulated evidence now makes it possible to resolve some of the earlier conflicting results, especially as we have become more cognizant of the kinds of difficulties associated with these experiments. In addiTShortly after completing work on this paper, Dr Herbert E. Kubitschek died on 28 November 1989, at the age of 69. tion to reproducibility, there are three major requirements for experiments on cell growth. (1) Resolution must be sufficient to distinguish among the three major growth models that have been proposed - exponential, linear, or bilinear growth. For example, the experimental resolution must be far better than the largest difference, about 6%, between exponential and linear growth (Fig 1). (2) Growth characteristics of cells in synchronized cultures must remain unchanged from their values ii; steady-state or exponential phase cultures: that is9 growth should not be perturbed by procedures used tc synchronize cells. (3) Data analysis must concern an appropriate growth parameter. In particular, as discussed later, averages over measured values of divided and undivided cells in synchronous cultures are inappropriate because the two groups represent different generations. The earliest method for measuring cell growth was to observe individual cells by light microscopy as they progressed through the division cycle. Results were not very meaningful until the late 1950s, however, when the concept of maintaining balanced growth conditions was established (Campbell, 1957). Nevertheless, even when these conditions were adopted, very different patterns o i length increase were found for the single strain E. cdr B/r, including exponential (Schaechter et al., 1962), bilinear (Cullum & Vicente, 1978) and interrupted (Hoffman & Frank, 1965). These disparate patterns reflected insufficient optical resolution, approximately 0.4-0.5 ym, in the measurement of cells 1-2 pm in length. This lack of sufficient optical resolution led N . Nanninga, C. L. Woldringh and their colleagues at the University of Amsterdam to study cell growth and division by a technique capable of much higher resolution : electron micrography. However, with this technique, they could not follow the growth of individuai cells; consequently, cell growth kinetics were obtained with a method based upon analysis of steady-state population distributions of cell size, as pioneered by Collins & Richmond (1962). This method provided greater consistency for the measurement of the kinetics 0001-5987 Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sat, 17 Jun 2017 23:06:17 600 H . E. Kubitschek 2.0 1.5 --. E 1.0 0.5 No. of generations 1.0 Fig. 1. Cell growth models. During linear growth relative cell mass (rn/rno) increases at a constant rate, as shown by line A. Curve B represents exponentialgrowth, differing from A by less than 6%at any point during the cycle. Line C represents bilinear growth; the rate of growth is constant, doubling at some point within the cycle. Actually, linear growth is the particular case of bilinear growth for which the growth rate change occurs at the end of the division cycle. Curve D is the locus of rate change points for all bilinear curves, and shows the maximum deviations between bilinear growth and the other models. of cell length extension, which was exponential during the division cycle (Koppes et al., 1978a, b ; Kubitschek & Woldringh, 1983). Furthermore, if cell width were constant, these results implied that cell volumes should increase essentially exponentially. However, Trueba & Woldringh (1980) found that cell diameters did not remain constant but diminished by as much as 8% during part of the cycle, corresponding to a change in crosssectional area of about 15%. In this case, cell volume should increase less rapidly than exponentially. In support of this view, Kubitschek (1981) found that when both lengths and diameters were measured for individual cells, then cell volume appeared to increase bilinearly, i.e. with a step-up in growth rate during the cycle. The Collins-Richmond analysis, however, requires data of much greater precision than do more direct measurements of cell growth because the shape of the population distribution depends not only upon cell growth during the cycle, but also upon the distributions of cell size at birth and division. Analysis requires estimations of these contributions from birth and division, which must be either assumed or derived from distributions of deeply constricted cells. Because these different techniques are so strongly influenced by statistical variation, experimental errors can only be reduced by the labour-intensive approach of collecting far more data than is required by other methods (Koch, 1966; Kubitschek & Woldringh, 1983). The third method of determining cell growth kinetics uses synchronous cultures. Ideally, a synchronously growing cohort of cells of a given age or size is selected from a steady-state or exponential-phase culture, and the growth of this cohort is then used to represent the growth pattern of the average individual cell. The use of synchronized cultures, however, has led to observations of a variety of growth patterns for E. coli. For length extension alone, these patterns include, in addition to those discussed above, linear (Donachie et af., 1976; Pierucci, 1978), approximately exponential (Marr et af., 1969), or even a more complex response (Meyer et al., 1979). When cell volume increase was examined in synchronized cultures, the growth patterns were mainly of two kinds and appeared to depend upon the synchronizaton technique. Cell volume increase (already concluded to be approximately exponential from length measurements alone) was also approximately exponential when cells were synchronized by membrane elution or elutriation centrifugation and their sizes were measured micrographically (Olijhoek et af., 1982; Olijhoek, 1983). On the other hand, cell volume increase was found to be linear when cells were synchronized by sucrose gradient centrifugation and their sizes determined with a Coulter electronic counter analyser system (Kubitschek, 1968a, 1986, 1990). This apparent dependence upon methodology highlighted a requirement for consistency of experimental results with those from cultures in steadystate or exponential growth for adequate comparisons to be made. The second criterion, the absence of perturbation of cell growth, is commonly recognized, and tests for perturbation are usually performed. However, these tests are customarily based only upon the timing of increase in cell numbers (Nanninga & Woldringh, 1985; Kubitschek, 1987), and as such are incomplete because cell volumes may also be perturbed in synchronized cultures. Several observations support the existence of cell growth perturbation in cultures synchronized by membrane elution (Kubitschek, 1987, 1990). First, when cells of E. coli B/r were filtered on membranes, synthesis of protein D and other altered outer-membrane proteins was induced, changing the chemical composition of the outer membranes of these cells (Boyd & Holland, 1977). Second, buoyant densities of E. coli cells synchronized by membrane elution were significantly higher than those of cells in their unsynchronized parent culture (Kubitschek et al., 1983, 1984). Third, rates of increase of cell length and volume changed periodically during the cell cycle after membrane elution (Olijhoek & Nanninga, cited by Nanninga et al., 1982). Fourth, in slowly growing membrane-eluted cultures of E. coli B/rA, DNA replication was consistently initiated at or near the beginning of the cell cycle, with a large gap in synthesis for most of the final quarter or third of the cycle (Helmstetter & Pierucci, 1976; Koppes et al., 19786). In exponentialphase cultures, however, the gap in DNA synthesis occurred, instead, during the first part of the division Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sat, 17 Jun 2017 23:06:17 Cell growth and membrane protein doubling cycle as measured by DNA :DNA hybridization (Chandler et al., 1975), decay of incorporated 251 (Kubitschek & Newman, 1978), and flow cytometry (Skarsted et al., 1985). These comparisons provide very strong evidence for perturbation of cell growth by the membrane elution technique. Thus, although this technique has yielded fundamental information on DNA replication, one cannot conclude that it provides synchronized cultures with cell characteristics representative of exponentialphase growth, as stated recently (Cooper, 1988a). The third criterion is that the analysis of cell growth must concern the proper cohort in a synchronousculture. Cell growth during the division cycle has sometimes been interpreted as the increase in mean, modal or total cell volume in the synchronized culture (Scott et al., 1980; Olijhoek et al., 1982). Although each of these parameters represents growth of the average individual cell early in the cycle, each ceases to represent the average as soon as any cells divide, because the newly divided cells advance to the following generation and are no longer representative of the kinetics of cell growth of the first generation (Kubitschek, 1986). Because ages and volumes of newly divided cells are smaller than those of the remaining undivided cells, if both groups are averaged together, then average cell size and age will decrease as more and more cells divide near the end of the cycle. Furthermore, the mean cell volume cannot then double during the cycle, as occurs for undivided cells during steady-state or exponential phase growth. Thus, average individual cell growth is correctly represented only by the growth of the undivided cells alone, for which both age and volume continually increase in the same manner as they do in steady-stateor exponential-phase cultures. The misuse of mean, modal or total cell volume to describe cell growth is readily apparent from graphs of cell growth: these incorrect calculations yield a single, sinuous curve connecting the points of successive division cycles. In contrast, proper analysis of the growth of undivided cells provides distinct curves of volume increase for each generation. Even when the three criteria (resolution, absence of perturbation, and appropriate cohort) appear to be satisfied, experimental results may still differ for different techniques. To ascertain which set of results is correct, the guiding principle is again to compare the experimental results with those obtained in steady-state or exponential-phase cultures. If a direct comparison cannot be made, perturbation may be revealed by examining some alternative measurable characteristic. The timing of DNA replication of E. coli B/rA in slowly growing exponential phase cultures is an example. As observed by autoradiography of individual cells from an exponential-phaseculture, initiation of DNA replication starts very near the beginning of the division cycle and is 601 associated with a long D period (Koppes et al., 1978b). However, when measurements were made upon exponential-phase cultures by methods of DNA :DNA hybridization, by survival of cells exposed to decay of incorporated 251,or by flow cytometry, initiation was found to occur much later in the cycle, following a large gap in DNA synthesis and associated with much shorter D periods. Again, the concurrence obtained with the latter three widely different techniques strongly supports a large initial gap in DNA replication during the division cycle of this strain of bacterium. The source of the alternative kinetics observed with electron micrography is not yet identified. Possibly, the different kinetics concern the consequences of fixation. Although electron micrography provides excellent dimensional resolution, the cells are flattened and deformed in cross-section during specimen preparation (Woldringh et al., 1977). Also, although the fixation technique was carefully designed to match optical dimensions of cell length and width, the unusual pattern produced for initiation of DNA synthesis in E. coli B/rA continues to raise questions about the consequences of the fixation technique. These questions also apply to more recent electron micrographic results obtained for cells in steady-state cultures exposed to shifts in growth rate. In particular, in an experimental and analytical tour de force, Zaritsky et al. (1982) examined a number of models of cell growth for their agreement with the experimental results during a shift in growth rate. Their results supported a model of linear increase in cell surface area and a gradual increase in growth rate during the growth shift. As they pointed out, however, assumptions of constancy of cell mass at initiation and constant periods of chromosome replication might invalidate their conclusions. The three major criteria for E. coli cell-growth studies, as discussed above, appear to have been achieved for synchronous cultures by measurements of cell volume with a Coulter counter analyser system in cultures synchronized by selection of cells from velocity gradients. Results of tests for perturbation of growth were negative, with the same doubling periods for cell numbers and cell volumes in synchronous cultures and parent cultures, and the same sizes of cells at birth and at division (Kubitschek, 1986,1987,1990).The accuracy of the Coulter counter analyser system also was examined by comparing cell volumes observed with the counter to those obtained with a common biophysical technique, the measurement of excluded volume in packed cells. The results supported the same cell volumes within experimental errors (Kubitschek & Friske, 1986) and therefore establish that the counter analyser measured the cell volume within the outer membrane, as expected (rather than, say, that within the inner cell membrane). Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sat, 17 Jun 2017 23:06:17 H . E. Kubitschek Growth models: comparison and implications In the cultures just described, cell volumes increased linearly in E. coli B/rA at doubling tmes of 25 and 40 min, in 15.coli 15THU at doubling times of 25 and about iO5 min, and in E. coli 12 WP2s at a doubling time of about 105 min (Kubitschek, 1986, 1990). In two ways, these results effectively ruled out an exponential increase in cell volume during the cycle. Regressions fitted to the data had slopes and intercepts that were in significant disagreement with those predicted by the exponential growth model. In addition, the nonrandom distribution of residuals (i.e. differences between observed and theoretical values) invalidated the model of exponential growth. Bilinear growth was not completely ruled out, but in one case at least it was shown that if growth is bilinear, then growth rate doubling must occur very near cell division, within 0.03 generations, and the pattern would appear almost indistinguishable from linear growth (Kubitschek, 1986). These results also effectively ruled out an exponential increase in cell mass, the fundamental growth parameter. It has been shown that cell buoyant density in E. coli is essentially independent of age (Kubitschek et al., 1983); and thus cell mass, which is the product of cell volume times density, must remain proportional to volume. Studies of buoyant cell density and the criteria that must be satisfied for experimental determination thereof were reviewed earlier (Kubitschek, 1987). Despite this evidence against exponential growth, Cooper (1988a) argued that cell growth in E. coli must be exponential because protein and RNA are known to accumulate essentially exponentially in cells during the division cycle. That argument, however, neglects the contribution of soluble material in the cell pools to the total cell mass. In the eukaryote Schizosaccharomyces pornbe, for example, this ratio is approximately 15% during midcycle, as observed long ago by Mitchison (1957). Although the fraction of pool mass is only a few percent of the total in E. coli, the kinetics of cell growth depend strongly upon pool size variation during the cycle. Assuming that macromolecular synthesis increases exponentially during the division cycle, linear cell growth predicts that soluble pools will increase to midcycle maxima and then decrease to their initial values again at the end of the cell cycle, whereas exponential cell growth predicts the absence of such a midcycle maximum. The linear growth model was supported by the behaviour of cell pools for uracil, histidine and methionine in E. coli THU (Kubitschek & Pai, 1988). What is the implication of linear cell mass increase during the cycle? Clearly, constant rates of mass increase require constant rates of accumulation of compounds from the surrounding medium. To test that possibility for individual compounds, Kubitschek (1968 b) examined rates of accumulation of glucose, acetate, phosphate, sulphate, leucine, glycine and thymidine in synchronous cultures of E. coli. In every case, the relative constancy of the net uptake rates both supported the linear growth model and provided evidence against the increasing rates of uptake during the cycle predicted by the exponential model. Uptake of potassium in synchronous cultures also supported the linear growth model (Kubitschek et al., 1971). Accumulation rate experiments inherently have much greater resolution than measurements of cell mass or volume and consequently are more capable of discriminating between alternative cell growth models. Despite this advantage, I am aware of only one other attempt to measure uptake rate, rather than incorporation, of a protein or RNA precursor in E. coli, an experiment by Cooper (1988~).He treated cells with a pulse of radioactive leucine, then washed away the free radioactivity with warm medium before inverting the membrane and sampling at intervals. However, these samples could not have represented leucine uptake rates during the cell cycle as stated because unincorporated leucine would have almost immediatedly been dissipated by diffusion from the cells or have become incorporated into protein before the cells were eluted from the membrane. Thus, those results cannot reveal the sizes of the leucine pools later in cells of different ages or the fraction of the pool leucine that becomes incorporated into the cell at different ages. Models have been proposed for both exponential and linear growth. In his passive continuum model, Cooper (1979,1982) proposed that cell mass increase is essentially exponential because the increase in cell mass is the sum of the different macromolecular and micromolecular components of the cell and the rates of synthesis of these components are assumed to be essentially proportional to their masses. This proportionality has been found to hold, to good approximation, for RNA and protein synthesis. Cooper’s (1984) model also postulates that, at most, a very few ‘events’ occur during the cell cycle, and therefore that ‘cell division is the end of a process and the beginning of none’. That is, cell growth between divisions is predicted to be simple and continuous (Cooper, 1988b). The model thus predicts that sudden transitions and new modes of regulation at the whole-cell level do not occur. In contrast, the model for linear cell growth is a transport model of growth regulation (Kubitschek, 1968a, 6). It postulates that increase in cell mass during steady-state growth is limited by and proportional to the net rate of transport of materials into the cell, which in turn is proportional to the number of functional transport Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sat, 17 Jun 2017 23:06:17 Cell growth and membrane protein doubling sites on the cell surface. Cell growth rate constancy requires that the (functional) number of these sites remain constant during the cell cycle and that the number be doubled at or near cell division, thus maintaining the same average numbers of sites on daughter cells. Two predictions of this latter model have now been supported. The first is that some cell proteins associated with the surface should double abruptly; abrupt doubling has been observed for a number of envelope proteins of E. coli (Ohki, 1979; Churchward & Holland, 1976; Boyd & Holland, 1979) and for binding proteins for lactose and maltose (Shen & Boos, 1973; Dietzel et al., 1978), as discussed below. Second, assuming that cellular macromolecular synthesis increases essentially exponentially during the division cycle, linear growth predicts that soluble pools of RNA and protein precursors should increase to midcycle maxima, then decrease again to their initial values by the end of the division cycle. Mesured values of pool sizes for uracil, histidine and methionine in E. coli THU supported these predicted pool kinetics (Kubitschek & Pai, 1988). Abrupt doubling of transport proteins Perhaps the most distinctive difference between the exponential and the linear growth models concerns the latter’s prediction of a set of transport proteins that double abruptly during the cell cycle. This difference will almost certainly distinguish between the two models and has already begun to do so, as described below. First, however, in opposition to that interpretation, it can be argued that although a number of candidates have been found, the evidence for the sharpness of the doubling is insufficient because these doublings occurred at many different times during the cell cycle and the relationships of some of these proteins to transport was not established. Furthermore, stepwise increases contrast with the results of Lutkenhaus et al. (1979), who observed by means of two-dimensional electrophoresis essentially continuous synthesis of 750 individual proteins of E. coli. For an even more precise estimate of synthesis rates, they also measured the radioactivity of some 30 of these proteins synthesized during the cell cycle and found that the variability of each was 15% or less. These results would negate the linear growth model unless the proteins measured are structural or ‘housekeeping’ proteins not involved in the regulation of cell growth. The transport sites model predicts that only a small fraction of the cell proteins should double abruptly. In agreement, most or all proteins observed to double abruptly are envelope or periplasmic proteins. Another objection to the evidence presented below, for 603 abrupt doubling of specific proteins during the cell cycle, might be that this evidence is obtained from experiments that have not been shown to satisfy the critical tests for absence of growth perturbation. Indeed, in some cases, data were obtained from cultures synchronized by membrane elution and, consequently, are probably perturbed. However, although the precise kinetics of abrupt doubling of specific proteins may not be established, there is no question about the occurrence of the phenomenon. In each study, the time interval required for doubling of envelope/periplasmic protein was much shorter than the cell generation time. The first membrane proteins found to increase abruptly in E. coli K12 were cytochrome b, and L-aglycerolphosphate dehydrogenase (Ohki, 1972). The two proteins appeared to double simultaneously during the cell cycle, but the cell age at doubling depended upon the method used for synchronization and occurred at ages of either 0.6 or 0.2 generations. This dependence of time of synthesis upon the synchronization method appears to be a common feature in measurement of abruptly doubling proteins and suggests that different methods of synchronization perturb the timing of doubling to different degrees. The following year, using nutritionally synchronized cultures, Shen & Boos (1973) found that the periplasmic galactose-binding protein was synthesized only during or near cell division. Within a few years, several other proteins were found to be produced in a cycle-dependent manner. These included envelope proteins [the a-receptor protein (Ryter et al., 1975) and protein D (Gudas et al., 1976; Boyd & Holland, 1977)], inner-membrane proteins [L-a-glycerolphosphate dehydrogenase (Ohki & Mitsui, 1974) and lactose permease (Ohki & Sato, 1975)], and another cytoplasmic protein [the maltose-binding protein (Dietzel et al., 1978)l. Churchward & Holland (1976) studied the synthesis of 29 envelope proteins in cultures synchronized by the membrane elution technique. They found that rates of synthesis of the great majority of these proteins increased stepwise and that these steps occurred early in the cycle, (However, one envelope protein, the D protein of M i 76000, was synthesized periodically near the time of division.) To be a prediction of the linear growth model, it would seem that abrupt doublings in rates of synthesis of these surface proteins would be limited by the numbers of functional transport sites. Ohki (1979) identified 11 proteins, each of which doubled 10-20 min before cell division, and also described their cellular locations ; these proteins were primarily inner-membrane proteins, but a few were outer-membrane or cytoplasmic. Ohki assigned these proteins to cell-cycle control A. He assigned the maltoseand galactose-binding proteins and other periplasmic proteins to cell-cycle control B because these proteins Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sat, 17 Jun 2017 23:06:17 604 H . E. Kubitschek were synthesized 11 min later. His studies clearly demonstrated that there may be more than one cell-cycle control mechanism that produces abrupt protein doubling; they suggest, however, no more than two. Boyd & Holland (1979) also found that the rate of synthesis of outer-membrane proteins, and in particular of a 36.5 kDa porin, doubled during the final part of the cell cycle. Doubling of phospholipid synthesis, observed by Pierucci (1979) and Pierucci et al. (1981), was confirmed by Joseleau-Petit et al. (1984), and again later by the same group (Joseleau-Petit et al., 1987), using the relatively new technique of automatic synchronization (Kepes & Kepes, 1980). The confirmation also provides support for this synchrony technique, which uses a forcing method rather than selection to obtain synchronization. The abrupt synthesis of the 1-receptor protein was confirmed by Ohki & Nikaido (cited in Ohki, 1979), and later, by a very different method, by Vos-Scheperkepter et al. (1984), who showed that the 1amB protein also doubled before division. There is little or no evidence that abrupt doubling is related to DNA replication per se. When DNA synthesis was blocked, rates of synthesis of total and outermembrane and of cytoplasmic proteins continued to increase in parallel (Boyd & Holland, 1979). Nor is there evidence that the control for abrupt doubling operates at the level of gene expression: while total RNA and trp mRNA synthesis continued at almost normal rates at a nonpermissive temperature in the cell-cycle control mutant tsC42, the mutant failed to synthesize lac mRNA at the nonpermissive temperature (Ohki, 1979). However, Benner et al. (1985) provided evidence that the control of the maltose transport system is at the transcriptional level. Furthermore, results for doubling of the rate of phospholipid synthesis are compatible with models in which some event is linked to replication termination or nucleoid segregation (Joseleau-Petit et al., 1987).Thus, the mechanism(s)for cell-cycle regulation of abrupt doubling of selected proteins have not yet been definitively identified. The isolation of a cell-cycle control mutant by Ohki & Mitsui (1974) is an important step in the elucidation of this mystery. The defect of the tsC42 mutant (divE42)is phase specific. The divE42 gene functions only about 20 min before cell division and is essential for cell growth and membrane synthesis (Ohki & Sato, 1975). When shifted to 42"C, the mutant increased in protein by a factor of about 1.8 before ceasing growth, and at this stage the cells were large and homogeneous in size. When returned to a permissive temperature, cells divided synchronously after an immediate burst of synthesis of particular enzymes, such as succinate dehydrogenase. In synchronous cultures, the cell cycle stage at which mutant cells were arrested coincided with the time of synthesis of the particular proteins whose synthesis was controlled by divE42, leading Ohki (1979) to propose that a defect in the divE42 gene is involved in both cell cycle regulation of specific protein synthesis and of cell division. The defect produced by this mutation was later identified as a single alteration of the D stem of the isoacceptor serine t R N A y (Tamura et al., 1983). Finally, these results for the divE42 gene suggest that the abrupt doubling of specific proteins in the absence of gene doubling may also have a role in cell division. Other evidence has recently been obtained by A. Robin & D. Joseleau-Petit (personal communication), who found that theftsZ gene was transcribed bilinearly, but with a doubling in rate around the time of initiation of DNA synthesis rather than near the end of the cell cycle; they used an f t s Z : :lac2 fusion carried on a A prophage in cultures under automatic synchronization. Thus, the phenomenon of abrupt doubling of proteins near the end of the cell cycle may offer crucial information on the development of cell division, as well as on the regulation of cell growth. I thank Fred J. Stevens for his discussion and comments on the development of this paper. This work was supported by the US Department of Energy, Office of Health and Environmental Research, under Contract no. W-3 1-109ENG-38. References BENNER,D., MULLER,N . & Boos, W. (1985). Temperature-sensitive catabolite activator protein in Escherichia coli BUG6. Journal of Bacteriology 161, 347-352. BOYD,A. & HOLLAND, I. B. (1977). Protein d , an iron transport protein, is induced by filtration of cultures of Escherichia coli. FEBS Letters 76, 20-24. BOYD,A. & HOLLAND, I. B. (1979). Regulation of the synthesis of surface protein in the cell cycle of E. coli B/r. Cell 18, 287-296. CAMPBELL, A. (1957). Synchronization of cell division. Bacteriological Reviews 21, 263-272. CHANDLER, M., BIRD,R. E. & CARO,L. (1975). The replication time of the Escherichia coli K12 chromosome as a function of cell doubling time. Journal of Molecular Biology 94, 127-132. CHURCHWARD, G. G. & HOLLAND, 1. B. (1976). Envelope synthesis during the cell cycle in Escherichia coli B/r. Journal of Molecular Biology 105, 245-261. COLLINS, J. F. & RICHMOND, M. H. (1962). Rate of growth of Bacillus cereus between divisions. Journal of General Microbiology 28, 15-33. COOPER, S. (1979). A unifying model for the G 1 period of prokaryotes and eukaryotes. Nature, London 280, 17-19. COOPER, S. (1982). The continuum model : application to G 1-arrest and G(0). In Cell Growth pp. 315-336. Edited by C. Nicolini. New York: Plenum Publishing Corporation. COOPER, S. (1984). Application of the continuum model to the clock model of the cell cycle. In Cell Cycle Clocks, pp. 209-218. Edited by L. Edmunds. New York: Marcel Dekker. COOPER,S. (1988a). Leucine uptake and protein synthesis are exponential during the division cycle of Escherichia coli B/r. Journal of' Bacteriology 170, 436-438. Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sat, 17 Jun 2017 23:06:17 Cell growth and membrane protein doubling COOPER,S. (1988b). What is the bacterial growth law during the division cycle? Journal of Bacteriology 170, 5001-5005. CULLUM, J. & VICENTE, M. (1978). Cell growth and length distribution in Escherichia coli. Journal of Bacteriology 134, 330-337. DIETZEL,I., KOLB,V. & Boos, W. (1978). Pole cap formation in Escherichia coli following induction of the maltose binding protein. Archives of Microbiology 118, 207-218. DONACHIE, W. D., BEGG,K. J. & VICENTE, M. (1976). Cell length, cell growth, and cell division. Nature, London 264, 328-333. GUDAS,L. J., JAMES,R. & PARDEE,A. B. (1976). Evidence for the involvement of an outer membrane protein in DNA initiation. Journal of Biological Chemistry 251, 3470-3479. HELMSTETTER, C. E. & PIERUCCI, 0. (1976). DNA synthesis during the division cycle of three substrains of Escherichia coli B/r. Journal of Molecular Biology 102, 477-486. HOFFMAN, H. & FRANK,M. E. (1965). Time-lapse photomicrography of cell growth and division in Escherichia coli. Journal of Bacteriology 89, 2 12-216. JOSELEAU-PETIT, D., KEPES,F. & KEPES,A. (1984). Cyclic changes of the rate of phospholipid synthesis during synchronous growth of Escherichia coli. European Journal of Biochemistry 139, 60561 1. JOSELEAU-PETIT, D., KEPES,F., PEUTAT,L., D'ARI, R. & KEPES,A. (1 487). DNA replication initiation, doubling of rate of phospholipid synthesis, and cell division in Escherichia coli. Journal of Bacteriology 169, 3701-3706. KEPES,F. & KEPES,A. (1980). Synchronisation automatique de la croissance de Escherichia coli. Annales de Microbiologie 131A, 316. KOCH, A. L. (1966). Distribution of cell size in growing cultures of bacteria and the applicability of the Collins-Richmond principle. Journal of General Microbiology 45, 409-41 7. KOPPES,L. J. H., OVERBEEKE, N. & NANNINGA, N. (1978~).DNA replication pattern and cell wall growth in Escherichia coli PAT 84. Journal of Bacteriology 133, 1053-1061. C. L. & NANNINGA, N. (1978b). Size KOPPES,L. J. H., WOLDRINGH, variations and correlation of different cell cycle events in slowgrowing Escherichia coli. Journal of Bacteriology 134, 423433. KUBITSCHEK, H. E. (1968~).Linear cell growth in Escherichia coli. Biophysical Journal 8, 792-804. H. E. (1968b). Constancy of uptake during the cell cycle KUBITSCHEK, in Escherichia coli. Biophysical Journal 8, 1401-1412. KUBITSCHEK, H. E. (1981). Bilinear cell growth of Escherichia coli. Journal of Bacteriology 148, 730-733. KUBITSCHEK, H. E. (1986). Increase in cell mass during the division cycle of Escherichia coli B/rA. Journal of Bacteriology 168, 613618. KUBITSCHEK, H. E. (1987). Buoyant density variation during the cell cycle in microorganisms. CRC Critical Reviews in Microbiology 14, 73-97. KUBITSCHEK, H. E. (1990). Cell volume increase in Escherichiacoli after shifts to richer media. Journal of Bacteriology 172, 94-101. KUBITSCHEK, H. E. & FRISKE,J. A. (1986). Determination of bacterial cell volume with the Coulter counter. Journal of Bacteriology 168, 14661467. KUBITSCHEK, H. E. & NEWMAN, C. N. (1978). Chromosome implication during the division cycle in slowly growing, steady-state cultures of three Escherichia coli B/r strains. Journal of Bacteriology 136, 179190. KUBITSCHEK, H. E. & PAI,S. R. (1988). Variation in precursor pool size during the division cycle of Escherichia coli: further evidence for linear cell growth. Journal of Bacteriology 170, 43 1-435. KUBITSCHEK, H. E. & WOLDRINGH, C. L. (1983). Cell elongation and division probability during the growth cycle of Escherichia coli. Journal of Bacteriology 153, 1379-1 387. M. L. & SILVER,S. (1971). Potassium KUBITSCHEK, H. E., FREEDMAN, uptake in synchronous and synchronized cultures of Escherichia coli. Biophysical Journal 11, 787-797. KUBITSCHEK, H. E., BALDWIN, W. W. & GRAETZER, R. (1983). Buoyant density constancy during the cell cycle of Escherichia coli. Journal of Bacteriology 155, 1027-1032. H. E., BALDWIN, W.W., SCHROETER, S. J. & GRAETZER, KUBITSCHEK, 605 R. (1984). Independence of buoyant cell density and growth rate in Escherichia coli. Journal of Bacteriology 158, 296299. LUTKEMIAUS, J. F., MOORE,B. A., MASTERS,M. & DONACHIE, W. D. (1979). Individual proteins are synthesized continuously throughout the Escherichia coli cell cycle. Journal of Bacteriology 138, 352360. MARR,A. G., PAINTER,A. R. & NILSON,E. H. (1969). Growth and division of individual bacteria. Symposia of the Society .#or General Microbiology 19, 237-261. N. (1979). MEYER,M., DE JONG,M. A., DEMETS,R. & NANNINGA, Length growth of two Escherichia coli substrains. Journal of Bacteriology 138, 17-23. MITCHISON, J. M. (1957). The growth of single cells. I. Schizosaccharomyces pombe. Experimental Cell Research 13, 244-262. NANNINGA, N. & WOLDRINGH, C. L. (1985). Cell growth, genome duplication, and cell division. In Molecular Cytology of Escherichia coli, pp. 259-318. Edited by N. Nanninga. New York: Academic Press. C. L. &KOPPES,L. J. H. (1982). Growth NANNINGA, N., WOLDRINGH, and division of Escherichia coli. In Cell Growth, pp. 225-270. Edited by C. Nicololine, New York : Plenum Publishing Corporation. OHKI,M. (1972). Correlation between metabolism of phosphatidyl glycerol and membrane synthesis in Escherichia coli. Journal of Molecular Biology 68, 249-264. OHKI, M. (1979). The cell cycle-dependent synthesis of envelope proteins in Escherichiacoli. in Bacterial Outer Membranes :Biogenesis and Function, pp. 293-314. Edited by M. Inoue. New York: Wiley. OHKI,M. & MITSUI,H. (1974). Defective membrane synthesis in an E. coli mutant. Nature, London 252, 64-66. OHKI,M. & SATO,S. (1975). Regulation of expression of lac operon by a novel function essential for cell growth. Nature, London 253,654-656. OLIJHOEK,A. J. M. (1983). Cell growth and envelope synthesis in Escherichia coli. PhD Thesis, University of Amsterdam. OLIJHOEK, A. J. M., KLENCKE,S., PAS, E., NANNINGA,N. & U. (1982). Volume growth, murein synthesis, and murein SCHWARTZ, cross-linkage during the division cycle of Escherichia coli PA 3092. Journal of Bacteriology 152, 1248-1 254. PIERUCCI, 0. (1978). Dimensions of Escherichia coli at various growth rates : model for envelope growth. Journal of Bacteriologj, 135, 559574. PIERUCCI, 0. (1979). Phospholipid synthesis during the cell division cycle of Escherichia coli. Journal of Bacteriology 138, 453-460. O., MELZER,M., QUERINI, C., RICKERT, M. & KRAJEWSKI, PIERUCCI, C. (1981). Comparison among patterns of macromolecular synthesis in Escherichia coli B/r at growth rates of less and more than one doubling per hour at 37 "C. Journal of Bacteriology 158, 313316. H. & SCHWARTZ, M. (1975). Integration of the RYTER,A., SHUMAN, receptor for bacteriophage lambda in the outer membrane of Escherichia coli: coupling with cell division. Journal of Bacteriology 122, 295-301. SCHAECHTER, M., WILLIAMSON, J. P., HOOD,J. R., JR & KOCH,A. L. (1962). Growth, cell, and nuclear division in some bacteria. Journal of General Microbiology 29, 421434. SCOTT, R. I., GIBSON, J. F. & POOLE,R. K. (1980). Adenosine triphosphatase activity and its sensitivity to ruthenium red oscillate during the cell cycle of Escherichia coli K 12. Journal of General Microbiology 120, 183-198. SHEN,B. H. P. & Boos, W. (1973). Regulation of the B-methylgalactoside transport system and the galactose-binding protein by the cell cycle of Escherichia coli. Proceedings of the National Academy o j Sciences of the United States of America 70, 1481-1485. K., STEEN,H. B. & BOYE,E. (1985). Escherichia coli DNA SKARSTAD, distribution measured by flow cytometry and compared with theoretical computer simulation. Journal of Bacteriology 163, 661668. TAMURA,F., NISHIMURA,D. & OHKI, M. (1983). The E. coli divE mutation, which differentially inhibits synthesis of certain proteins, is in t R N A P . EMBO Journal 3, 1103-1107. TRUEBA, F. J. & WOLDRINGH, C. L. (1980). Changes in cell diameter during the division cycle of Escherichia coli. Journal of Bacteriology 142, 869-878. Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sat, 17 Jun 2017 23:06:17 606 H . E. Kubitschek VOS-SCHEPERKERTER, G. H., PAS,E., BRAKENHOFF, G. J., NANNINGA, N. & WITHOLT,B. (1984). Topography of insertion of the lamB protein into the outer membrane of Escherichia coli wildtype and laclamB cells. Journal of Bacteriology 159, 440-447 WOLDRINGH, c. L., DE JONG, M.A., VAN DEN BERG,w. & KOPPES,L. (1977). Morphological analysis of the division cycle of two Escherichiacoli substrains during slow growth. Journal of Bacteriology 131, 270-279. ZARITSKY,A., WOLDRINGH, C. L., GROVER,N. B., NAAMAN, J. & ROSENBERGER, R. F. (1982). Growth and form in bacteria. Comments on Molecular and Cellular Biophysics 1, 237-260. Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sat, 17 Jun 2017 23:06:17