bacterial growth and division - u-m personal world wide web server

1

Bacterial Growth and Division

Stephen CooperUniversity of Michigan Medical School, Ann Arbor, MI, USA

1 Growth of Bacterial Cultures 41.1 The Classical Growth Curve 41.2 Bacterial Growth at Different Growth Rates 51.3 The Shift-up 51.4 Classical Growth Pattern Is Merely a Series of Shift-ups and Shift-downs 6

2 The Bacterial Division Cycle 62.1 Cell Age and the Age Distribution during Balanced Growth 62.2 Aggregation Theory and the Control of Synthesis during

the Division Cycle 7

3 Synthesis of Cell Components during the Division Cycle 83.1 Cytoplasm Synthesis during the Division Cycle 83.1.1 Exponential Cytoplasm Synthesis during the Division Cycle 83.1.2 Variation in the Rate of Cytoplasm Increase 93.2 DNA Synthesis during the Division Cycle 103.2.1 Pattern of DNA Synthesis during the Division Cycle 103.2.2 Regulation of Initiation of DNA Replication 123.2.3 The Specific Origin of Replication 123.2.4 Protein Factors Controlling DNA Replication 133.2.5 Chromosome Organization and Bidirectional Replication 133.2.6 DnaA Protein Is Involved in Initiation of DNA Replication 133.3 Surface Synthesis during the Division Cycle 143.3.1 Peptidoglycan Synthesis during the Division Cycle 143.3.2 The Invagination Process 173.3.3 FtsZ Protein Is Involved in Septum Formation and Cell Division 173.3.4 Localization of the Division Site 17

2 Bacterial Growth and Division

4 Plasmid Replication during the Division Cycle 184.1 High-copy Plasmids 184.2 Low-copy plasmids 184.3 Minichromosome Replication during the Division Cycle 18

5 Segregation of Cell Components at Division 185.1 Segregation of Cytoplasm 185.2 Segregation of DNA 195.3 Segregation of the Cell Wall 20

6 Are There Events during the Division Cycle? 20

7 Do Checkpoints Control the Bacterial Cell Cycle? 21

8 Temporal Variation of the Bacterial Division Cycle 22

9 Alternative Bacterial Division Cycles 229.1 Caulobacter crescentus Division Cycle 229.2 Bacillus subtilis Division Cycle 229.3 Streptococcal Growth during the Division Cycle 23

10 Mass Increase Is the Driving Force of the Division Cycle 23

11 The Bacterial Growth Law during the Division Cycle 24

12 Experimental Analysis of the Bacterial Division Cycle 25

13 Genetic Analysis of the Bacterial Cell Cycle 25

14 Die Einheit der Zellbiologie 26

Bibliography 27Books and Reviews 27

Keywords

C-periodThe time for a round of DNA replication, from initiation at the origin to completion atthe terminus. In Escherichia coli, the C-period, for cells growing between 20- and60-min doubling times, is a constant of approximately 40 min.

Bacterial Growth and Division 3

D-periodThe time between termination of DNA replication and cell division. In Escherichia coli,the D-period, for cells growing between 20- and 60-min doubling times, is a constant ofapproximately 20 min.

Multiple ForksThe situation that occurs when initiation of DNA replication occurs on a molecule thathas not finished its previous round of replication; there are forks upon forked material.

CytoplasmThe collective name for the numerous components of the cell that are not associatedwith either the cell wall or the genome. It is composed of the small molecules such asions, metabolites, and cofactors, macromolecules such as soluble enzymes, tRNAs,mRNAs, and higher complexes such as ribosomes.

PeptidoglycanThe stress-bearing, presumably shape-maintaining, layer of the bacterial cell wall. It iscomposed of glycan chains cross-linked with amino acids.

Cell MembraneThe lipid-containing structures that lie adjacent to the peptidoglycan layer. Some cellshave a single membrane layer, while others have two different membranes, an outerand an inner membrane.

SegregationThe distribution of cell material from the dividing mother cell into the two daughtercells. Different cell components may have different segregation patterns.

OriginA sequence on a chromosome or a plasmid where normal replication of DNA initiates.

TerminusA point or region on a chromosome or plasmid where a round of replicationterminates.

Bidirectional ReplicationReplication of DNA with two replication forks proceeding in opposite directions awayfrom the origin of replication.

GC-skewThe excess (or deficit) of G residues over C residues within a single strand of DNA.While G = C in a double helix, an individual strand may have an excess of G within aregion or an excess of C residues within a region.

FtsZ RingThe ring of synthesis of the pole region during pole formation.

mcb200300028


� A dividing bacterial cell must, on average, have precisely twice as much of everythingfound in a newborn cell. How does a bacterial cell ensure that all cell componentsare duplicated between divisions, and how does a cell ensure that it does not divideprior to duplicating all of its components? These questions have been investigated bystudying biosynthesis of cell components during the division cycle. Cell cytoplasmincreases uniformly and exponentially during the division cycle. There does notappear to be any cell-cycle-specific syntheses of cytoplasmic components during thedivision cycle of a well-studied bacterium such as E. coli. DNA replication is initiatedat a specific time in the division cycle when the cell mass per chromosome originreaches a particular value. Once initiated, DNA replicates at a constant rate such thatthe time between initiation and termination of replication is relatively independentof the growth rate. The time between termination of replication and cell divisionis also relatively constant. High-copy plasmids replicate throughout the divisioncycle and low-copy plasmids replicate at a particular time during the division cycleaccording to rules similar to that for initiation of chromosome replication. The cellsurface grows in response to the increase in cell mass so that the turgor pressureinside the cell, and the cell density, are constant during the division cycle. Atdivision, the components of the cell are segregated to the two daughter cells. Thecytoplasm segregates randomly, the DNA segregates stochastically but nonrandomly,and the peptidoglycan segregates in a manner consistent with the fixed location ofthe synthesized peptidoglycan. The key ‘‘events’’ during the division cycle are theinitiation and termination of DNA replication and the initiation and termination ofpole formation. There do not appear to be other cell-cycle-specific events or synthesesduring the division cycle that are related to the regulation of the division cycle.

1Growth of Bacterial Cultures

1.1The Classical Growth Curve

For almost a century (since Buchanan,1918), the classic growth curve has beenused as a description of the pattern ofbacterial growth. An overgrown culture isdiluted into fresh medium and at first thereis a ‘‘lag’’ phase, experimentally definedas a period before viable cell numberbegins to increase. Then cell numbersbegin to increase and the culture entersthe ‘‘log’’ phase. The log phase is theperiod of exponential increase in cell

number. Although the cells during thisperiod of steady state cell number increaseare growing exponentially, the cells aregenerally referred to as being in ‘‘log’’phase because of the usual plotting of cellnumber on semilogarithmic coordinates.During the log phase, plotting cell numberincrease on semilogarithmic paper yields astraight line that can be used to determinethe doubling time of a culture. Afterthe cells grow to higher concentrations,cell growth slows as the cells enter thestationary phase. Cell growth then ceases.Finally, if a culture is studied long enough,cells begin to lose viability or the ability toform colonies, and the culture enters thedeath phase.

mcb200300028


During these phases, there are changesin cell sizes. The initial cells in theovergrown culture are small, cell sizeincreases during the lag phase, cells arelargest during the log phase, and then cellsize decreases as cells enter the stationaryphase. This classical view of the way inwhich cells grow is really a special case ofthe phenomenon of changes in cell sizewith changes in growth rate.

1.2Bacterial Growth at Different Growth Rates

It is possible to grow a given bacterium incultures at a wide range of growth rates.For example, it is possible to grow E. coliwith a 300-min interdivision time withserine as the sole source of carbon, and alittle faster than 20-min interdivision timesin various complex broths available forbacterial culture. By varying the mediumcomposition, it appears that any possiblegrowth rate is allowable, so that there isa continuous range of allowable growthrates between the extremely slow and theextremely fast.

When bacteria are grown at a giventemperature in different media at anumber of different growth rates, a regularpattern of cell composition and cell sizeis observed. The faster a cell grows(richer medium) the larger the cell andthe more the DNA, RNA, protein, andother components per cell (Fig. 1). Whencell components are accurately measured,it is found that the material per cellincreases exponentially as a function ofthe reciprocal of the doubling time (i.e.

growth rate, as a shorter doubling timeindicates a faster growth rate).

1.3The Shift-up

During a shift-up from a slow growthmedium (e.g. minimal medium, in whichcells are small) to rapid growth in a richmedium (e.g. nutrient broth, in whichcells are larger), there is a regular patternof change in cell size and composition.Immediately after the shift-up, there isa rapid change in the rate of RNA andprotein synthesis to the new rate ofincrease. There is a slower change inthe rate of DNA synthesis. There is adelay before there is any change in therate of cell number increase. Because of

Fig. 1 (a) Composition of bacteria as afunction of growth rate. (b) Synthesis ofcell components during a shift-up fromminimal medium to faster growth inricher medium.

Doublings per hour

0 1 2

RNA orprotein

DNA

Time [hours]−1

Cell number

Rel

ativ

e am

ount

per

mL

Shift-up

1

10

(b)

1 2 3

RNA

Rel

ativ

e am

ount

per

cel

l

1

10

Interdivision time [min]60 30 20

Mass

DNA

Protein

(a)

mcb200300028


the dissociation of mass increase and cellnumber increase, the average cell sizeincreases during this first part of the shift-up. After a period of time, there is a sharpchange in the rate of cell increase from theold rate to the new rate. This occurs 60 minafter the shift-up (Fig. 1). Because there is alag in the change in the rate of cell division,while cytoplasm synthesis increases to thenew rate almost immediately, during theshift-up period there is an increase in cellsize producing the larger-sized cells in thefaster growth medium.

1.4Classical Growth Pattern Is Merely a Seriesof Shift-ups and Shift-downs

The classical growth pattern for a culturecan now be seen as a series of shift-ups andshift-downs. When an overgrown culture(nongrowing, and thus of essentially zerogrowth rate) containing small cells isdiluted into fresh medium, there is a‘‘shift-up’’ in growth rate. The massincreases immediately with no cell divisionoccurring. The ‘‘lag’’ phase is thus one ofa lag in cell number increase, but thereis essentially no lag when consideringmass increase. After the normal cell size isachieved for that medium, both the massand the cell number increase in parallel.The rate of cell increase is determined bythe medium supporting the more rapidgrowth. There is a constant larger-size cellduring exponential growth. The reverseoccurs as growth begins to slow down andthe cells approach the stationary phase.The rate of mass synthesis decreases first,followed later by a cessation in the increasein cell number. Because division continuesafter mass increase ceases, the average cellsize decreases in an overgrown culture.

2The Bacterial Division Cycle

2.1Cell Age and the Age Distribution duringBalanced Growth

By convention, a newborn cell has an ageof 0.0 and grows during the division cycleto divide at age 1.0. The cell ages during thedivision cycle are thus numbers between0.0 and 1.0 indicating where, in thedivision cycle, a cell is located at a particularinstant. A cell half-way between birth anddivision is age 0.5. Cell age, by definition,increases linearly during the division cycle.However, cells at the same absolute timeafter birth are only approximately the samecell-cycle age because of the variability inthe absolute time required for the divisioncycle of an individual cell.

In a growing culture, the distributionof cell ages is not uniform. There aretwice as many newborn cells as dividingcells. The age distribution is given by theformula N = 21−x, where x is the cellage between 0.0 and 1.0. This formulaindicates that there are twice as manynewborn cells (x = 0.0) as dividing cells(x = 1.0). If all cells grew with exactlythe same interdivision time, this formulawould give the age distribution exactly.However, because of cycle variability thereis a smoothing of the function and theactual distribution is an approximationof the theoretical distribution. Becauseof the age distribution, the propertiesof an exponentially growing culture areindependent of time. This means that allpatterns of synthesis within the divisioncycle give an exponentially increasingamount of material in a growing culture.This is because the age distribution isconstant during steady state, exponentialgrowth. Thus, whether something is

mcb200300028


made linearly, exponentially, or at aninstant during the division cycle, inan exponentially growing culture theamounts of all cell components increaseexponentially and in parallel. This isreferred to as balanced growth. Theobserved balanced growth is the result ofthe age distribution during growth beinginvariant.

Given that between the birth of a celland its subsequent division, all of the cellcomponents double, we can ask, ‘‘At whatrate is material synthesized during thedivision cycle of an individual cell, and howdoes the cell ensure that there is a precisedoubling of cell components by the timethe cell divides?’’ Let us first analyze whichcomponents must be dealt with in termsof the division cycle.

2.2Aggregation Theory and the Control ofSynthesis during the Division Cycle

In economics, the aggregation problemis how to combine various sectors ofan economy. Should the figures forthe production of capital machinery becombined with those for the productionof consumer goods? Is paper producedfor boxes in the same economic categoryas stationery? It is difficult to treat eachitem in an economy individually; someaggregation is necessary in order tounderstand the whole system.

Now, consider the aggregation problemfor the analysis of the bacterial divisioncycle. How should one aggregate the dif-ferent cell components in order to achievean understanding of the biochemistry ofgrowth and division? Is there a uniquepattern of synthesis during the division cy-cle for each enzyme, or are there a limitednumber of patterns with different enzymesor molecules synthesized according to any

one of these patterns? Are there ways ofgrouping proteins or RNA molecules sothat one can consider classes of moleculesrather than individual molecular species?Should we consider the cell membrane as adifferent category from that of peptidogly-can? There are approximately a thousandproteins in the growing cell, and if eachprotein had a unique cell-cycle syntheticpattern, or if there were only a few en-zymes exhibiting any particular pattern,we would have an insuperable task de-scribing the biosynthesis of the cell duringthe division cycle.

Fortunately, one need consider onlythree categories of molecules, each ofwhich is synthesized with a unique pattern.The growth pattern of the cell is the sumof these three biosynthetic patterns. Thefirst category is the cytoplasm, which isthe entire accumulation of proteins, RNAmolecules, ribosomes, small molecules,water, and ions that make up the bulkof the bacterial cell. It is the materialenclosed within the cell surface that isnot the genome. The second category isthe genome, the one-dimensional linearDNA structure. The third category isthe cell surface, which encloses thecytoplasm and the genome. The surface iscomposed of peptidoglycan, membranes,and membrane-associated proteins andpolysaccharides. Everything in the cellfits into one of the three categories, andeach category has a different pattern ofsynthesis during the division cycle. Thesethree patterns are simple to understandas they can be derived from our currentknowledge of the principles involved in thebiosynthesis of the cytoplasm, the genome,and the cell surface.

The division cycle of bacteria in generalmay be best described by discussing thedivision cycle of the bacterium about whichmost is known, E. coli, and then describing

mcb200300028


other bacteria and their patterns of growthand division with respect to this archetypalpattern. The division cycle of E. coli willbe described, indicating how the variouscomponents of the cell are made anddistributed into two new daughter cells.

3Synthesis of Cell Components during theDivision Cycle

3.1Cytoplasm Synthesis during the DivisionCycle

3.1.1 Exponential Cytoplasm Synthesisduring the Division CycleConsider a unit volume of cytoplasm.It contains enzymes required for thebreakdown of nutrients, enzymes for theproduction of energy for biosynthesis,enzymes for the biosynthesis of low molec-ular weight precursors of macromolecules,and enzymes for the synthesis of macro-molecules. Each unit of cytoplasm pro-duces a small amount of new cytoplasmover a small interval of time. If the new cy-toplasm is indistinguishable from the old,and if the new cytoplasm acts to synthe-size more cytoplasm immediately, then thepattern of cytoplasm synthesis is exponen-tial. This is because after each interval ofsynthesis, the rate of synthesis increasesowing to the new increment of materialadded to the cytoplasm. This exponentialpattern is illustrated in Fig. 2 with both therate of synthesis and the amount of ma-terial increasing exponentially during thedivision cycle.

The pattern for cytoplasm as a wholeis reflected in the synthesis of theindividual cytoplasmic components. Thereare no changes in the specific rate ofsynthesis or pattern of cytoplasm synthesisthat are related to any particular cell-cycle event. (Actually, there is a minorcaveat to this generalization, which istreated below.) No particular molecule ofthe cytoplasm is made differently fromany other molecule of the cytoplasmduring the division cycle. In addition,there is no variation, during the divisioncycle, in the relative concentration of anycomponent of the cytoplasm. All parts ofthe cytoplasm accumulate exponentiallyduring the division cycle, and this patternis unchanged even when the cell isdividing. At the instant of division, thecombined rate of cytoplasm synthesis inthe two new daughter cells is preciselyequal to the synthetic rate in the dividingmother cell.

One minor exception to the rule thatspecific proteins are made exponentiallyduring the division cycle regards the effectof momentary sequestration of DNA uponreplication. Specific regions of DNA aresequestered in the membrane at differenttimes during the cell cycle at the time oftheir replication. During this sequestrationperiod, the DNA is not available fortranscription. Thus, for all genes there

Fig. 2 Cytoplasm synthesis during thedivision cycle. Both the rate ofcytoplasm synthesis and the pattern ofaccumulation of cytoplasm areexponential during the division cycle.

1

2

Cell age

0 1

Rat

e of

cyt

opla

sm

synt

hesi

s [lo

g sc

ale] A

mount of cytoplasm

per cell[log scale]

mcb200300028


is a time during the division cycle atwhich gene expression is prohibited.This momentary cessation of transcriptionshould not obscure the more fundamentalpoint that there does not appear to beany specific control of gene expression forany gene during the division cycle that isrelated to the regulation of that gene duringthe division cycle. One should imagine thiscessation of gene expression as merely anecessary accommodation of transcriptionto the needs of the cell for DNA replication.It could be imagined that if the cellwere able to replicate DNA without anycessation of transcription, there would beno reason to expect that the cell wouldalter specific gene expression at any pointduring the division cycle.

The evidence for a smooth, continuous,and exponential increase in cytoplasmcomes from experiments that use themembrane-elution method, colloquiallyreferred to as the baby machine (seebelow). The pattern of DNA synthesisduring the division cycle, which wassubsequently shown to be correct by flow-cytometric and other analyses, was actuallydiscovered using the membrane-elutionmethod. Thus, the results from a methodcapable of producing an accurate analysisof biosynthesis during the division cyclehave confirmed the exponential synthesisof cytoplasm.

In addition to this experimental support,an evolutionary argument can be madefor an invariant rate of accumulation ofcytoplasm during the division cycle. Ifthe synthetic rate of an enzyme changedabruptly during the division cycle, therewould be a relative excess or deficiencyof that enzyme at some point during thedivision cycle; this would be an inefficientuse of resources. The optimal allocation ofresources is for each component to be at aconstant concentration during the division

cycle. Cells would not evolve controls thatmake them less efficient in cell production.The ideal pattern of cytoplasm synthesisis an invariant cytoplasm compositionduring the division cycle.

A third argument supporting exponen-tial cytoplasm synthesis is that only thispattern is explained in known biochemicalterms. The cycle-independent exponentialsynthesis of cytoplasm can be derivedfrom our current understanding of thebiochemistry of macromolecule synthesis.Enzymes make RNA and protein, whichmake ribosomes, which then lead to moreprotein synthesis. More proteins meanmore RNA polymerases, more catabolicand anabolic enzymes, and the continu-ously increasing ability of the cell to makemore and more cytoplasm. In contrast, ifnewly synthesized material were not activeor available for synthesis when made, butinstead was recruited for biosynthesis onlyat the instant of division or at some spe-cific cell-cycle age, one would have linearcytoplasm synthesis during the divisioncycle. There are no known mechanismsenabling the cell to distinguish betweennewly synthesized cytoplasm and old cyto-plasm, or which lead to cell-cycle-specificvariation in the synthetic rate of any par-ticular molecule during the division cycle.We therefore conclude that in theory aswell as in practice, cytoplasm increasesuneventfully and exponentially during thedivision cycle.

3.1.2 Variation in the Rate of CytoplasmIncreaseThe continuous variation in allowablegrowth rates may be understood by parti-tioning the cytoplasm into those moleculesinvolved in cytoplasm synthesis and thatare absolutely necessary for growth (ribo-somes, RNA polymerases, etc.) and thosemolecules that are involved in dispensable

mcb200300028


syntheses (e.g. enzymes involved in aminoacid biosynthesis). In a minimal medium,in which many dispensable enzymes arebeing synthesized, the ribosomes may beconsidered to be involved in the synthesisof ribosomal protein and a large number ofdispensable proteins. Since only a fractionof the ribosomes are making ribosomalprotein at any moment, the time for a dou-bling in the ribosomal content is relativelylong. As one increases the number of nu-trients in the medium, thus relieving thecell of having to make the dispensable en-zymes, a larger fraction of ribosomes areinvolved in ribosomal protein synthesis.Thus, in rich medium the time for ribo-some number to double is shorter thanin poor medium. The variation in growthrates is thus a reflection of the contin-uous ability of the cell to partition theribosomes (or the RNA polymerases) be-tween making ribosomal components ornonribosomal components.

Considering the time required for aribosome to make another ribosome (i.e.using 15 amino acids incorporated ina protein chain in a second), one cancalculate that it would take eight minutesfor one ribosome to synthesize a singleribosome. If a cell were in some waysupplied with all proteins other thanribosomes, one could consider that thefastest rate of cell doubling would be eightminutes, the time for ribosomes to doubletheir number.

3.2DNA Synthesis during the Division Cycle

One of the most important generaliza-tions regarding the regulation of linearmacromolecule synthesis is that the rateof synthesis is regulated at the point of ini-tiation of polymerization. This principle isexhibited most clearly in DNA replication,

where the process of initiation of replica-tion can be clearly dissociated from thecontinued replication of DNA that hasalready initiated replication. If one in-hibits the initiation of replication, then therounds of DNA replication in progress willcontinue until their normal termination.

3.2.1 Pattern of DNA Synthesis during theDivision CycleThe DNA synthesis pattern during thebacterial division cycle (specifically forE. coli and related bacteria) is composedof one or more periods of constant rates ofDNA accumulation (Fig. 3). In cells witha 60-min interdivision time, there is aconstant rate of DNA synthesis for thefirst 40 min, with one bidirectional pair ofreplication forks, followed by a zero rate ofsynthesis during the last 20 min. In cellswith a 30-min interdivision time, there is aconstant rate of DNA synthesis for the first10 min, which falls to a constant rate thatis two-thirds of the initial rate for the last20 min. In cells with a 20-min interdivisiontime, there is a constant rate of DNAsynthesis throughout the division cycle.Other growth rates have similar particularpatterns of DNA synthesis.

The constant rates of DNA synthesiscan be understood in terms of themolecular aspects of DNA synthesis.DNA is synthesized by the movementof a replication point along the parentaldouble helix, leaving two double helicesin its wake. The rate of movement ofa growing point appears to be invariantand independent of its location in thegenome. This means that the rate of DNAsynthesis is proportional to the extantnumber of replication points. The numberof replication points is constant for anyperiod of the division cycle. Thus, the rateof DNA synthesis during any period ofthe division cycle is also constant. The

mcb200300028


3

2

6

1.75

3.5

3

6

1

1

Interdivisiontime [min]

203060

Rat

e of

DN

AS

ynth

esis

Cell age during division cycle

2

10 10 10

Am

ount

of D

NA

Cell size at birth

10 10 10

Fig. 3 DNA synthesis during the division cycle. Three differentpatterns are illustrated for cells growing with 20, 30, and 60 mindoubling times. For each growth rate, the time from initiation totermination of DNA replication is 40 min. The time betweentermination of replication and cell division is 20 min. The proposedchromosome patterns at the start and finish of the division cycle areillustrated above the graphs. In addition to the rate of DNA synthesisduring the division cycle, the pattern of accumulation of DNA duringthe division cycle is presented. Accumulation of DNA is composedof periods of linear synthesis. The rates during these periods areproportional to the existing number of growing points. The graph ofthe rate of synthesis is the differential of the accumulation plot. Arepresentation of the cell size expected for cells at the start of thedivision cycle is presented below each synthetic pattern. The size ofthe newborn cells is in the ratio of 1 : 2 : 4 in the three culturesillustrated here. The number of new initiations at the start of eachdivision cycle is also in the ratio of 1 : 2 : 4.

pattern of constant rates of synthesis isderived from, and is consistent with, ourunderstanding of the biochemistry of DNAsynthesis.

Embedded in the patterns of Fig. 3 arethe two major rules for the determinationof the pattern of replication of DNA duringthe bacterial division cycle. The first ruleis that once DNA replication is initiated

at the origin of replication, the time fora complete round of replication – the C-period – is relatively invariant (40 min inE. coli at 37 ◦C). The second rule is that thetime between termination of replicationand cell division – the D-period – is alsorelatively invariant.

The constant C- and D-periods are mostclearly evident in the 60-min cells in Fig. 3.

mcb200300028


At other growth rates, the patterns ofsynthesis appear more complicated, butthis is due to the possibility of initiationoccurring before completion of previousrounds of replication and the possibilityof initiation of replication occurring inone division cycle and replication ofthe chromosome being completed in asubsequent division cycle.

3.2.2 Regulation of Initiation of DNAReplicationWhy is replication initiated at theseparticular times during the division cycle?How is the initiation process regulated?Rounds of replication are initiated in abacterial cell when the amount of cellmass or something proportional to thecell mass (the ‘‘initiator’’) is present at acertain amount per origin. Initiation ofnew rounds of replication occurs onceper division cycle when the mass (or‘‘initiator’’) per origin reaches a fixed value.In the 60-min cell, initiation occurs atone origin with a cell of unit size (size1.0). In the 30-min cells, there are twoorigins in the newborn cell and the sizeof the newborn cell is twice that of a 60-min cell. And the 20-min cell has fourorigins initiated in the newborn cell andthe cell size is four times the newborn60-min cell. Thus, not only does theDNA per cell increase with growth ratebut the average cell size increases withgrowth rate as well. In the cells shown inFig. 3, the relative sizes of the newborncells, as well as the average cell in theculture, are in the ratio of 1 : 2 : 4. Thisincrease in cell size with growth rate isconsistent with the exponential increasein cell size as a function of the inverse ofthe doubling time.

The actual nature of the initiator isnot known, but a number of candidatemolecules have been identified. Among

these, the DnaA protein appears as agood candidate for the regulator of DNAinitiation.

Alternative models of initiation havebeen proposed, such as the suddenaccumulation of an inhibitor of initiationthat is diluted out by cell growth. Thismodel does not accommodate the constantsynthesis of cell-cytoplasmic components.In all cases, it is indistinguishable from theinitiator accumulation model, and there isno evidence for such an inhibitor beingsynthesized during the division cycle.

3.2.3 The Specific Origin of ReplicationA specific sequence of DNA in the bacterialgenome is the origin of replication (oriC).The required sequence (for E. coli) is245 bp. Within this sequence are 5 highlyconserved sequence blocks of 15 to 20 bp.These conserved regions are interspersedwith sequences that are random. Regionsof random sequence but of fixed lengthseparate the conserved regions. Thus,the length of separation of conservedsequences is more important than thespecific sequence between the conservedsequences. Four sequences of 9 bp arespecific recognition sites for the bindingof dnaA, the specific initiator protein. Inaddition, there are three 13-bp regionsthat are sites for opening the duplex DNAfor initiation. Finally, there are 14 GATCsequences located in the origin (oriC)region and the neighboring regions. These14 four-bp sequences are expected to occurat a much lower frequency if they were nota necessary part of the E. coli origin.

The ultimate proof that this sequence isthe origin of replication is the productionof minichromosomes, small replicatingcircles of DNA that have the bacterialorigin of replication. An E. coli cell canhave many copies of the minichromosome

mcb200300028


without any apparent impairment of cellgrowth.

3.2.4 Protein Factors Controlling DNAReplicationInitiation of replication of DNA is afunction of numerous proteins. The keyregulatory protein appears to be theproduct of the dnaA gene. The dnaAprotein binds to the 9-bp ‘‘dnaA boxes’’in the oriC region. After binding to theorigin, the protein promotes melting of theDNA (i.e. separation of the DNA strandsin a local region) and helps in loadingadditional proteins such as the dnaBprotein. After this initial opening of theDNA strands, replication fork movementis allowed to proceed bidirectionally fromthe origin.

Initiation of replication is distinct fromthe continued replication of DNA thathas been initiated. Once initiated, DNAreplication can continue even thoughthe accumulation of cell cytoplasm isinhibited.

3.2.5 Chromosome Organization andBidirectional ReplicationThe phenomenon of bidirectional replica-tion has two effects on the organization ofthe chromosome. One effect is the varia-tion in the relative abundance of G residuesversus C residues in one of the strands.While G = C in the double helix, such anequality need not exist in a given strand orin a particular region of a strand. As a gen-eral rule, it is found that in the two halvesof the bidirectionally replicated bacterialgenome, one half of the chromosome (e.g.that from noon past three down to six ona clock representation of the genome) hasG > C and the other half (noon, past nine,down to six) has G < C. If one calculatesthe G − C/G + C for a running windowof nucleotides, and accumulates the value

of this ‘‘GC-skew,’’ there is a valley thatcorresponds to the origin and a peak thatcorresponds to the terminus. As long asone reads the strand in the 5′ to 3′ direc-tion, the location of the peak and the valleyis independent of the strand being used todefine the origin and the terminus.

The bidirectional organization of thegenome also affects the direction of genetranscription. As a general rule (withnumerous exceptions), the direction ofgene transcription is in the same directionas the movement of the replication fork.The rationale for this phenomenon isthat this correspondence of direction ofreplication and transcription reduces the‘‘head-on collision’’ of replication andtranscription. The fact that there are manyexceptions means that the replicationsystem can accommodate running intotranscriptional units without a problem.

3.2.6 DnaA Protein Is Involved inInitiation of DNA ReplicationThe DnaA protein, identified by mutantsaffecting the initiation of DNA replicationand not the progression of rounds ofreplication already in process, appearsto be the ultimate regulatory proteinwith regard to the initiation of DNAsynthesis. The initiation process involvesthe cooperative binding of approximately20 DnaA protein molecules to the originof replication. This complex then leads tothe open complex in which a region ofAT-rich 13-mer repeats denatures so thatsingle-stranded DNA is exposed. Otherproteins such as the DnaB helicase arenow introduced into the forks of the meltedDNA where it expands the region of single-stranded DNA. Other molecules thenjoin the initiation complex with the finalloading of DNA polymerase III. When thisoccurs, the replication of DNA progresses

mcb200300028


around the genome to the opposite end ofthe genome where replication terminates.

3.3Surface Synthesis during the Division Cycle

3.3.1 Peptidoglycan Synthesis during theDivision CycleThe cell surface is made to perfectlyenclose, without excess or deficit, thecytoplasm synthesized by the cell. Cell cy-toplasm increases continuously during thedivision cycle (described above), and there-fore the cell surface is made continuously.How is the bacterial cell surface made andduplicated during the division cycle? Whatis the rate and topography of cell-surfacesynthesis during the division cycle?

Consider an imaginary cell in which thecytoplasm is enclosed in a cell-surface tubethat is open at each end; the cytoplasmremains within the bounds of the tube.The cytoplasm in the newborn cell isthus encased in the cylinder of cellsurface that is made up of membraneand peptidoglycan. As the cytoplasmincreases exponentially (see above), thetube length increases to exactly enclosethe newly synthesized cytoplasm. In thisimaginary case, the cell surface increasesexponentially in the same manner asthe cytoplasm. When the cell cytoplasmdoubles, the tube divides into two new cellsand the cycle repeats. In this imaginarycell, the cytoplasm increases exponentially,the internal volume of the cell increasesexponentially, the surface area increasesexponentially, and the density of the cell,that is, the total weight of the cell per cellvolume, is constant during the divisioncycle.

However, a real rod-shaped cell doeshave ends, and therefore the pattern of cell-surface synthesis during the division cycle

is not simply exponential. If the cell sur-face were synthesized exponentially, thecell volume cannot increase exponentially.Since the cytoplasm increases exponen-tially, there would have to be a change incell density. (In theory, it may be thatmajor changes in cell shape could al-low a pure exponential increase in cellmass and cell surface, but such changesare clearly not observed during the cell-division cycle.) Cell density, however, isconstant during the division cycle. A pro-posal for cell-surface synthesis that allowsan exponential increase in cell volume,and therefore a constant cell density, ispresented in Fig. 4. Before invagination,the cell grows only in the cylindrical side-wall. After invagination, the cell growsin the pole area and the sidewall. Anyvolume required by the increase in cellcytoplasm that is not accommodated bypole growth is accommodated by an in-crease in sidewall. The cell is considered apressure vessel, and when the pressure inthe cell increases, there is a correspondingincrease in cell surface area. The pole ispreferentially synthesized and any resid-ual pressure due to new cell cytoplasmthat is not accommodated by pole growthis relieved by an increase in cylindricalsidewall area.

The resulting pattern of synthesis isapproximately exponential. The formuladescribing surface synthesis during thedivision cycle is a complex one, includingterms for the shape of the newborn cell,the cell age at which invagination starts,the pattern of pole synthesis, and the ageof the cell. One consequence of this modelof surface synthesis is that at no timeis surface synthesis exponential, since therate of surface synthesis is not proportionalto the amount of surface present over anyperiod of time.

mcb200300028


Fig. 4 Cell-surface synthesis during thedivision cycle. Before invagination, thecell grows only by cylinder extension.Each cell is drawn to scale, with thevolume of the cells increasingexponentially during the division cycle.The shaded regions of the cell indicatethe amount and location of wall growth(whether in pole or sidewall) each tenthof a division cycle. The width of theshaded area is drawn to scale.Cell-surface growth actually occursthroughout the sidewall and not in anarrow contiguous zone. The variablelocations of wall growth in this figurehave no specific meaning, but are meantto show synthesis occurring all over thesidewall during the division cycle. Beforeinvagination, the ratio of the surfaceincrease to volume increase is constant.When pole synthesis starts, at age 0.5 inthis example, there is an increase in theratio of the surface increase to thevolume increase. Any volume notaccommodated by pole growth isaccommodated by cylinder growth. Atthe start of pole growth, there is areduction in the rate of surface growthin the cylinder. As the pole continues togrow, there is a decrease in the volumeaccommodated by the pole and anincrease in the rate of growth in thesidewall.

0.0

0.1

0.2

0.3

0.4

0.7

0.8

0.9

1.0

0.5

0.6

Cell age

As with cytoplasm and DNA synthesis,we can derive the cell-cycle pattern ofsurface synthesis from our understandingof the molecular aspects of peptidoglycansynthesis. The peptidoglycan sacculus ofthe cell is composed of glycan strandsencompassing the cell, perpendicular tothe long axis, as shown in Fig. 5. Thesestrands are cross-linked by peptide chains.Because the cross-linking in the load-bearing layer is not complete, one canhave new strands in place, below the tautload-bearing layer, prior to the cutting ofthe bonds linking old strands. As shown inFig. 5, the load-bearing layer is stretchedby the turgor pressure in the cell. An

infinitesimal increase in cytoplasm leadsto an infinitesimal increase in the turgorpressure of the cell. This increase in turgorpressure places a stress on all of thepeptidoglycan bonds, thus reducing theenergy of activation for bond hydrolysis.The result is that there is an increase inthe cutting of stressed bonds between theglycan chains. When a series of cuts ismade, allowing the insertion of a singlenew strand into the load-bearing layer,there is an infinitesimal increase in cellvolume. This increase in volume leads toa reduction of the stress on the remainingbonds. By a continuous repetitive seriesof cytoplasm increases, surface stresses,

mcb200300028


g adaa

a g d a agd

a

a g d a

Fig. 5 Three-dimensionalrepresentation of peptidoglycanstructure. This is an idealizedrepresentation of peptidoglycanstructure as seen from the outer surface.The thick bars represent glycan chainsat the outside of the peptidoglycan layer.The thinner bars represent glycan chainsbelow the outer layer. The stretchedchains of circles represent amino acidscross-linking the glycan chains. Thechains below the stretched surface ofthe cell rise to the outer layer when thetaut layers of the peptidoglycan arehydrolyzed. The figure shows across-sectional view through the glycanchains illustrating the taut outer layerand the more loosely inserted innermaterial. The letters indicate the aminoacids composing the cross-links;a = alanine, d = diaminopimelic acid,and g = glutamic acid, with italicizedletters present in the D-configuration.

enzymatic cuts, and volume increases,one gets an increased cell volume thatprecisely accommodates the increase incell cytoplasm.

An inside-to-outside mode of peptido-glycan growth, similar to that for Bacillussubtilis, has been proposed for E. coli. Thisproposal is based on the observed recy-cling of murein, the calculated amount ofpeptidoglycan per cell, and the existenceof trimeric and tetrameric fragments thatare consistent with a multilayered pepti-doglycan structure. The insertion of newpeptidoglycan strands in an unstressedconfiguration prior to their movement intothe load-bearing layer of the peptidogly-can can explain all of these observations.The recycling of peptidoglycan may bea strain-specific result, as there is nomajor release or recycling of peptido-glycan in Salmonella typhimurium or inE. coli B/r. At this time, the inside-to-outside mode of surface growth cannotbe excluded.

Although this discussion of the rate ofsurface synthesis during the division cyclehas dealt primarily with peptidoglycan,it applies equally to membranes andother surface-associated elements. Thecell membrane grows in response to theincrease in peptidoglycan surface andcoats the peptidoglycan without stretchingor buckling. The area of the membraneincreases in the same way in whichbacterial peptidoglycan increases.

One must make distinctions among cellcomponents with regard to their locationin the cell, rather than with regard totheir chemical properties. Cell proteinscan be divided into two categories – thoseassociated with the cytoplasm and thoseassociated with the surface. Proteins asso-ciated with the membranes are includedin the surface category of synthesis duringthe division cycle. If a bacterial cell had ahistone-like protein associated with DNA,there would be a third category of protein-synthetic patterns, proteins synthesized

mcb200300028


during the division cycle with constantrates of synthesis. Proteins are synthesizedduring the division cycle with a patternthat is presumably consistent with theirfinal location or category. Proteins are nota monolithic group nor are they dividedinto a large number of groups. The three-category system proposed here allows us toconceptualize the possible patterns of pro-tein synthesis observed during the divisioncycle.

3.3.2 The Invagination ProcessThe division of the bacterial cell into twodaughter cells occurs by the formation oftwo new poles in the middle of the dividingcell. As there are very few DNA-less cellsproduced during balanced growth, thedivision process must be coordinated withthe replication process so that divisiondoes not occur until two genomes arepresent. The invagination process thatforms two poles in the middle of the cellinvolves a complex process whereby thereis an inward growth of the peptidoglycanand the associated membrane of the cellsurface.

One of the earliest events related to theinvagination process is the localization ofthe FtsZ protein to the future site of poleformation. The FtsZ protein is localizedto the division site in a pattern called theFtsZ ring. The FtsZ protein accumulates atthe mid-cell site where division will occurbefore any visible invagination. Duringthe invagination process, the FtsZ proteinis located at the leading edge of theinvagination process. Thus, the FtsZ ringcontinues to define a ring of decreasingsize as invagination proceeds.

Other proteins necessary for invagina-tion and division are apparently localizedat the center of the cell because of theirassociation with FtsZ.

3.3.3 FtsZ Protein Is Involved in SeptumFormation and Cell DivisionRecent results indicate that the tubulin-like FtsZ protein plays a central role incytokinesis or cell division as a majorcomponent of a contractile cytoskeleton.An indication that the FtsZ protein is thekey initiator of septation is that mutantsof this protein that form filaments showno indication of any constriction. Otherseptation mutants presumed to be laterin the pathway of septum formation doshow some indication of septation. TheFtsZ protein resembles the eukaryotictubulin molecule and this similarity hasled to speculations that the FtsZ proteinhas cytoskeletal properties that allow itto form the new pole in the middle ofthe cell.

3.3.4 Localization of the Division SiteThe actual mechanism by which the divi-sion site is localized at mid-cell position isunknown. One model of the division-sitelocalization proposes that the site existsprior to any observation of FtsZ localiza-tion. An alternative model proposes thatthe division site is determined in someway by the reaction of the cell surface tothe presence of DNA; the absence of DNAin the center of the cell would thus lead toinvagination-site localization.

An understanding of site location isstrongly influenced by the existence ofminicells. If invagination occurs, upon oc-casion, near a pole rather than in the centerof a cell, a very small DNA-less ‘‘mini-cell’’ is produced that is essentially twopoles without an intervening sidewall. Thegenetics of minicell formation is compli-cated, but the main idea is that the productsof the min genes have a negative functionand normally inhibit abnormal invagina-tion sites. Thus, inactivation of any of anumber of min genes leads to minicell

mcb200300028


formation owing to the inappropriate loca-tion of invagination sites.

4Plasmid Replication during the DivisionCycle

4.1High-copy Plasmids

If a cell has a large number (20–100)of plasmids per newborn cell, then onecan consider that these plasmids replicateduring the division cycle in proportionto those present (i.e. exponentially) andthen assort themselves randomly intothe newborn cells. Because the plasmidnumber is high, there is little chance ofone of the daughter cells ending up withno plasmids and with all of the plasmidssegregating to the other daughter cell. Thispattern appears to be the case for naturallyoccurring, high-copy number plasmids.

One possible exception is the artificialplasmid made by cloning the origin (oricC)region of the bacterial chromosome. Thisappears to have a random assortmentmechanism, but it is synthesized at aprecise time during the division cycle thatcoincides with the normal time at whichthe chromosome initiates replication.

4.2Low-copy plasmids

Low-copy plasmids are usually larger thanhigh-copy plasmids and are present onthe order of 1 to 2 per cell. Examplesof such plasmids are the F-factor andthe P1 plasmid. These replicate in aprecise way at a particular time duringthe division cycle. The rules of theirreplication are similar to that regulatingchromosome replication, with initiation of

plasmid replication occurring when a fixedamount of cell mass is present per plasmidorigin.

4.3Minichromosome Replication during theDivision Cycle

Minichromosomes replicate at the sametime as normal chromosome initiation.The behavior of minichromosomes isparadoxical in that in a cell with numerousadditional origins of replication therewould be a competition for the ‘‘initiator’’of replication. Such a competition wouldlead to abnormal cell sizes or somealteration in cell growth pattern. No suchalteration is observed.

5Segregation of Cell Components at Division

Producing the right amount of materialso that the cell can divide is not enoughfor a successful division cycle. At division,the material in the mother cell must alsobe properly apportioned to the newborncells. In a cell that divides to produce twoequivalent daughter cells, the segregationproblem is to ensure that both cells havethe same amount of each portion ofthe cell material. Let us see how thethree categories of material in the mothercell are segregated properly to the newdaughter cells.

5.1Segregation of Cytoplasm

The cytoplasmic components of the cellappear to segregate randomly at division.There does not appear to be any compart-mentalization or hindrance to segregationof cytoplasm (Fig. 6).

mcb200300028


1

0.50.5

0.25 0.25

Cytoplasm

DNA

0.5 0.5

∼0.30∼.20pq

Peptidoglycan

Fig. 6 Segregation of cytoplasm, surface, and DNA at division.The cytoplasm of the cell is segregated randomly at division.Each daughter cell shares equally in the parental cytoplasm(ribosomes, soluble proteins, ions, etc.), which can beconsidered randomly distributed throughout the dividing cell.The DNA is segregated nonrandomly, but each daughter cell getsa complete and equal complement of DNA from the dividing cell.The nonrandom segregation is at the level of a single strand.Since each DNA duplex is made up of an ‘‘older’’ strand and anew complement made on the older strand, and each cell iscomposed of an ‘‘older’’ pole and a pole just made at theprevious division, one can see whether strands of one age gopreferentially toward one or the other age pole. Experimentalanalysis indicates that the DNA strands segregate as though theolder DNA strand goes preferentially toward the pole that it wentto at the previous division. This may be phrased as the newerstrand goes toward the newer pole. The degree of nonrandomsegregation is not perfect, but is stochastic, and is approximately60/40 rather than 50/50. The surface of the cell segregates to thedaughter cells at division such that the poles are conserved andthe sidewall segregates with no apparent conservation of sidewallpeptidoglycan. It is not clear whether insertion of new wallmaterial over the side is perfectly uniform or whether there issome nonuniform pattern of insertion.

5.2Segregation of DNA

DNA segregates so that each daughtercell receives a complete genome fromthe dividing mother cell (Fig. 6). As the

DNA nucleoid is relatively compact withina dividing cell, if the nucleoids wereapportioned completely randomly, onemight expect that at some divisions oneof the daughter cells would not receive anyDNA and the other daughter cell would

mcb200300028


receive two genomes. That this is a rareevent supports the notion that segregationof the DNA is a regulated and deterministicprocess that leads to very few nonnucleatedcells in a culture.

In addition to the segregation of individ-ual nucleoids, the separate strands of theDNA have rules that describe their segre-gation and which are compatible with thedeterministic equipartition of DNA to thetwo daughter cells. As all rod-shaped cellsare asymmetric in that one of the polesis the new pole (produced at the last di-vision) and one pole is necessarily older,one can see how older DNA strands andnewer DNA strands (the newer strandshaving just been made on an older, pre-existing strand) segregate with regard topole age. The general rule is that the olderstrands segregate slightly preferentially tothe older poles and the younger strandssegregate preferentially to the newer oryounger poles. This phenomenon is con-sistent with other data that indicate someattachment of the DNA to the cell surface.

5.3Segregation of the Cell Wall

The peptidoglycan is a large macro-molecule that grows by intercalation ofnew material between old material. Sinceall the peptidoglycan is covalently linkedin one macromolecule, there is no abil-ity for the peptidoglycan components tomove with respect to each other. The seg-regation of cell wall material is thereforedetermined by the location of the cell wallmaterial. Once in place, the wall segre-gates into the newborn cells so that the oldpoles go to each cell, and the sidewall issegregated to the newborn cells (Fig. 6).

6Are There Events during the Division Cycle?

There are only four discrete events thatoccur during the division cycle of agrowing rod-shaped bacterial cell. Theyare (1) the initiation of DNA replication,(2) the termination of DNA replication,(3) the initiation of pole formation, and(4) the completion of pole formation.Other aspects of cell growth are toocontinuous to be considered cell-cycle-specific events. For example, the additionof new nucleotides at position 43,567 onthe chain of DNA may be considereda unique event, and one that occurs ata particular time in the division cycle,but because of cell-cycle variability, thetime of this synthetic occurrence cannever be precisely obtained. The time ofinsertion of a particular nucleotide pairhas no meaning for the cell cycle; onlythe initiation and termination of DNAreplication are important. With regardto the division cycle, the DNA is anamorphous material with no biochemicalspecificity along the strand.

Similar considerations apply to the cellsurface and the cytoplasm. No aspect ofpeptidoglycan strand insertion is uniquein time during the division cycle. A newstrand is inserted at random sites in re-sponse to the stretching of the cell surface.This may occur in one cell at position0.376 of the cell length; at the same time,in another cell, the new strand may beinserted at position 0.549. For an analysisof the cell cycle, we should consider poleand sidewall growth as uneventful exten-sions of the cell surface. Regarding thecytoplasm, one may consider the increasein ribosome content from 37,411 to 37,412an event; but such an individual event isnot measurable and is without meaning inthe cell cycle. Thus, cytoplasm synthesis,

mcb200300028


DNA synthesis, and cell-surface synthesisare uneventful during the division cycle.

In contrast to our understanding of theprinciples of cytoplasm, cell-surface, andDNA synthesis, the biochemical mecha-nisms of the four events occurring duringthe division cycle – the initiation and ter-mination of DNA synthesis and the initia-tion and cessation of pole formation – arestill unknown. While a great deal is knownabout the biochemistry of initiation ofDNA replication, the mechanism by whichthis initiation is regulated is less well un-derstood. In addition, very little is knownof the biochemical principles involved ineither the termination of DNA replicationor the start and cessation of pole forma-tion. A cell-surface structure, the periseptalannulus, has been proposed as a possiblefirst step in the formation of a new pole.If the periseptal annulus is the start ofpole formation, the important question iswhether there are definable steps betweenthe formation of the periseptal annulusand the start of invagination.

7Do Checkpoints Control the Bacterial CellCycle?

In eukaryotic cell-cycle studies, the conceptof ‘‘checkpoints’’ looms large. Checkpointsare the postulated functions of a cell thatdetermine that one function is properlycompleted before a subsequent functionoccurs. Thus, in eukaryotic cells it is pos-tulated that a checkpoint function deter-mines that DNA has replicated completelyand properly before allowing mitosis tooccur. The checkpoint is thus a valuablefunction that prevents abnormal functionsfrom occurring.

Studies of this type of phenomenonin bacteria can provide a corrective lens

for understanding or more rigorouslydefining checkpoints. In bacteria, if DNAsynthesis is inhibited, cells with only onegenome (i.e. cells that have not completedreplication of a total chromosome) do notdivide. Even when DNA is inhibited ordamaged, a cell with two genomes willdivide, while a cell with one genome willnot. Thus, in bacteria, DNA damage per sedoes not activate a checkpoint to preventdivision. It is believed that the presenceof a single genome in the center of a cellprevents division by its central presence,not the activation of a function that checksfor completion of DNA replication.

The bacterial experience thus suggeststhat a distinction should be made betweenthe existence of a function external tothe tested function (i.e. some activity thatchecks on whether DNA replication hasbeen completed) and the actual functionitself acting as a brake on further activity.The failure to complete DNA replication isthe inhibitor of division, and there is nochecking by some ‘‘checkpoint’’ activity.

A similar analysis may be applied to theubiquitous postulation of restriction pointsin eukaryotic cells. Restriction points arepoints or events postulated to exist in theG1-phase of the division cycle. Passage ofa cell through the restriction point is arequirement for entering the S-phase andeventual division. The restriction point wasdefined by the effect of different starvationregimens on DNA replication and theresumption of DNA replication followingrelease from starvation.

It is interesting that a bacterial restrictionpoint was postulated to exist before theproposal of the eukaryotic restrictionpoint, although the term restriction pointwas not used. Subsequent analysis ofthe bacterial restriction point indicatedthat the proposed control point was anexperimental artifact related to leakage of

mcb200300028


synthesis under starvation conditions. Theeukaryotic restriction point is subject tothe same critique.

8Temporal Variation of the Bacterial DivisionCycle

Division cycles in a culture are not pre-cisely timed. There is a great deal ofvariability in division cycles, with a culturehaving a 60-min doubling time proba-bly having cells with interdivision timesbetween 40 and 100 min. The variabilityof the division cycle can be visualizedas due to the variability of the separatecomponents that make up the divisioncycle. Thus, for two identical newborncells one can have the subsequent divi-sion occurring at different times becauseof (1) variability in mass synthesis in thetwo cells leading to variation in the timeof initiation of DNA replication, (2) varia-tion in the rate of DNA replication in thetwo cells, (3) variation in the time betweentermination and division, (4) and variationin the equality of division (which leads tovariation in the next division cycle).

9Alternative Bacterial Division Cycles

9.1Caulobacter crescentus Division Cycle

At first glance, the division cycle of thestalked bacterium Caulobacter crescentusappears quite different from that ofE. coli. At division, two different-sizedand different looking cells are produced(division is not in the middle of themother cell and one cell is stalked andthe other is flagellated), the DNA patternsduring the division cycle differ in each

of the daughter cells, and there arecell-cycle-specific syntheses of flagellum-related materials. The most importantdifference is that the poles of the daughtercells are different in that one cell hasa flagellum and the other has a stalk.Because of the production of two differentappearing cells, the growth and division ofCaulobacter has been described as a modelof cell differentiation.

The division cycle of Caulobacter canbe shown to correspond to that of thestandard E. coli model by considering thatthe different DNA patterns are due tothe different sizes of the two daughtercells arising by division not occurring inthe middle of the dividing mother cell,and that the cell-cycle-specific pattern ofexpression of flagellum-related genes isrelated to the completion of poles after theact of cell division. Flagellum synthesis isthus considered a part of surface synthesis.The cycle-specific pattern observed in cellsis related more to the window of time inwhich poles are made rather than synthesisat a particular time during the divisioncycle (Fig. 7).

9.2Bacillus subtilis Division Cycle

The division cycle of the gram-positivebacterium Bacillus subtilis is relativelysimilar to that of E. coli. The majordifference between the organisms lies inthe fact that the B. subtilis cell wall is muchthicker (giving rise to its gram-positivecharacter), and this requires an inside-to-outside pattern of peptidoglycan growth.The cell wall is made adjacent to thecytoplasmic membrane, and this wall layermoves out as subsequent layers are madebetween it and the cytoplasmic membrane.This leads to the multilayered B. subtilispeptidoglycan.

mcb200300028


C D

E′E′′

C DC DB B

C DC D B C

0

1a2a

1b 2b

X

Y

Fig. 7 The Caulobacter cell cycle. Caulobactercells divide unequally, producing a smallerswarmer cell that is motile and a larger stalkedcell that is generally attached to a surface by thestalk. As drawn here, the development ofCaulobacter can be followed by concentrating onthe two new poles formed by the first division.The development of the poles in both cells is asuccession from bald, to flagellated, to stalked.Because the swarmer cell is smaller thedevelopment is slightly delayed, suggesting asize control for both the initiation of DNA

replication and the completion of poledevelopment. As drawn here, thecell-cycle-specific synthesis of flagella protein(flagellin) is a result of pole developmentfollowing cell separation. The ultimate end ofpole development is a stalk, after which nofurther changes in the pole occur. Thus, withregard to the general model of cytoplasmsynthesis proposed here, the Caulobactercytoplasm (excluding surface) does not exhibitcell-cycle-specific patterns of gene expression.

9.3Streptococcal Growth during the DivisionCycle

The growth of spherical bacteria such asStreptococci reveals a different pattern ofsurface growth. There is no apparent in-tercalation of new peptidoglycan materialinto the preexisting older peptidoglycan.Rather, the old peptidoglycan is conservedand new peptidoglycan grows at the cen-ter of the cell to make a completely newcell pole.

10Mass Increase Is the Driving Force of theDivision Cycle

The driving force of the division cycle iscytoplasm synthesis. Some part of the en-ergy used by the cell to make cytoplasmdrives the biosynthesis of the cell surfaceby causing stresses along the cell surface

and this stress leads to the breaking ofpeptidoglycan bonds. The insertion of newpeptidoglycan then leads to the increase incell size. Cytoplasm increase also regulatesDNA synthesis. Cell size at birth is greaterat faster growth rates (Fig. 3) because theinitiation of new rounds of DNA replica-tion occurs when the cell has a unit (orcritical) amount of cytoplasm per origin ofDNA. Cell size at initiation is proportionalto the number of replication points in thecell, and that is why cells are larger at fastergrowth rates; at faster growth rates thereare more origins per cell at initiation. Thecell ‘‘titrates’’ the amount of cytoplasm orsome specific molecule that is a constantfraction of the cytoplasm. Thus, DNA syn-thesis is initiated at all available originswhen the amount of cytoplasm per originreaches a particular value.

The cell cytoplasm increases as fast asit can, given the external nutrient con-ditions. The synthesis of DNA and cell

mcb200300028


surface cannot outpace or lag behind cy-toplasm synthesis. DNA synthesis cannotgo faster because it is waiting to initi-ate new rounds of DNA replication inresponse to cytoplasm synthesis nor canit go slower because the cytoplasm wouldincrease without initiating new rounds ofDNA replication. Cell surface is made tojust enclose the newly synthesized cyto-plasm. The regulation of surface and DNAsynthesis by cytoplasm thus explains whythere is no dissociation of these synthesesduring the division cycle and during thegrowth of a culture. Furthermore, the fail-ure to complete DNA replication (so that acell has only one genome) prevents cell di-vision; the cell thus ensures that there willbe enough material (genome, cytoplasm,and cell surface) to allow the productionand survival of two new daughter cells.

11The Bacterial Growth Law during theDivision Cycle

It has been speculated that there isa ‘‘bacterial growth law’’ that can be

discovered by sensitive methods of anal-ysis. Does the cell grow linearly, bilinearly,exponentially, or perhaps follow someother, yet undiscovered, growth law? Ifthere were a general law of cell growth,independent of the biosynthetic patternsof the three categories that comprise thecell, then the individual categories ofbiosynthesis would have to accommodatethemselves to this overall growth law.

There is no simple mathematical growthlaw regulating or describing bacterialgrowth during the division cycle. Bacterialgrowth during the division cycle is thesimple weighted sum of the biosyntheticprocesses that are described by the threecategories proposed here. As presentedin Table 1, it can be seen that thecomponents of the cytoplasm of thecell are 80% of the total cell weight.This means that the growth of the cellis approximately exponential. A slightdeviation from exponential growth is dueto the contribution of cell surface and DNAsynthesis. Cell growth, then, is the resultof a large number of individual reactions,regulated by local conditions, and notconforming to any overriding growth law.

Tab. 1 Composition of Escherichia coli, a typical gram-negative, rod-shaped bacterium.

Cell component Molecules per cell Percent dry weight

CytoplasmProtein 2,350,000 55.0RNA 255,480 20.5Polyamines 6,700,000 1.1Glycogen 4,300 3.5Metabolites, ions, cofactors 2.5

SurfacePeptidoglycan 1 2.5Lipid 22,000,000 9.1Lipopolysaccharide 1,430,000 2.4

GenomeDNA 2.1 3.1

mcb200300028


If one wished to state the growth law,the growth law is simply the sum of theindividual synthetic patterns of the threecategories of the cell material.

12Experimental Analysis of the BacterialDivision Cycle

How can one study the pattern of synthesisof cell components during the divisioncycle? At first thought, it appears that onecould study the division cycle by preparinga synchronized bacterial culture – thatis, a culture composed of cells all ofthe same age, for example, newborncells – and taking measurements of thecell components at different times duringthe division cycle. It turns out that thismethod has rarely, if ever, led to anyconcrete results. This is presumably dueto the introduction of perturbations andartifacts when cells are synchronized, asno naturally synchronized populations areavailable.

The methods that have given results arenonperturbing, nonsynchrony methods.One is the membrane-elution method,or ‘‘baby machine,’’ which analyzes thedivision cycle of unperturbed cells. In thismethod, a label is added to a growingculture, the labeled cells are filtered onto amembrane, the cells bind as though boundat one end or pole, and when medium ispumped through the inverted membrane,newborn cells are eluted from the boundcells. Since the newborn cells arise fromthe bound cells as a function of the originalcell age, the first newborn cells arise fromthe oldest cells at the time of labeling.With time, the newborn cells arise fromcells labeled at younger and youngercell ages. By measuring the amount oflabel per cell during elution, one can

determine the pattern of synthesis duringthe division cycle. This has been donesuccessfully for DNA, plasmids, protein,RNA, peptidoglycan, and cell membrane.

Another method for cell-cycle analysisis flow cytometry, where the amount ofmaterial per cell in a cell population isdetermined for individual cells by a com-plex system of liquid handling and laser-excitation engineering. The quantitativedistribution of cell material can give the rel-ative rate of synthesis during the divisioncycle. This method has been successfulfor DNA replication during the divisioncycle and has confirmed the initial re-sults obtained with the membrane-elutionmethod. It is not clear whether the methodis sensitive enough for bacteria to enablethe discovery of a particular pattern of DNAreplication.

13Genetic Analysis of the Bacterial Cell Cycle

Complementing the physiological and bio-chemical analyses of the bacterial divisioncycle has been a large amount of work onisolating and analysing mutants that affectthe division cycle. Because a successful di-vision requires that the cell make all ofits required material, a mutation affect-ing the synthesis of any required materialis, in some trivial sense, a mutation af-fecting the division cycle. Thus, a mutantunable to make a particular amino acid,such as leucine, that is required for growthwould stop progressing through the divi-sion cycle. However, this would be a trivialrelationship to the division cycle becausethe leucine is required throughout the di-vision cycle and is not related to a specificcontrol mechanism.

More interesting are those mutants thataffect the major events of the division cycle

mcb200300028


such as the start of DNA synthesis andthe start of septation. A large number ofmutations affecting these processes havebeen isolated, and their study has revealedthat a number of enzymes and proteinsare required for cell-surface synthesis andDNA synthesis.

14Die Einheit der Zellbiologie

Are the interrelationships of cell-cycleregulation, proposed here for E. coli, appli-cable to other organisms? Other regulatorymechanisms have been proposed in othercell types. For example, a circular regula-tory system was proposed for Caulobacterand a stochastic mechanism as well asa rate-controlled model for animal cells.Is the mechanism of cell-cycle regula-tion subject to historical accident? Or isthere a unity of mechanisms regulatingthe division cycle of all types of cells? Iconjecture that the E. coli model of cell-cycle regulation is generally applicable toother organisms, and that the optimal de-sign of the cell cycle is exhibited by theE. coli cell cycle.

In the second decade of this century,Kluyver and his colleagues explicitly pro-posed the principle of the Unity of Bio-chemistry, or as originally proposed, DieEinheit der Biochemie (the oneness ofbiochemistry). Although the original pro-posal dealt primarily with the nature ofoxidation–reduction reactions, the basicprinciple has evolved to encompass evenmore unifying principles. The nature ofthe gene, the genetic code, the mechanismof protein synthesis, the use of enzymesfor biochemical changes, the patterns ofenzyme regulation, and all of the fun-damental aspects of biochemistry appear

similar throughout the different organ-isms on earth. While there are exceptionsand unique aspects to many organisms,the essential principle that there is a uni-fied biochemistry has been supported.

Here we must now add, with regardto the cell cycle, the Unity of CellBiology – Die Einheit der Zellbiologie – thatis analogous to the Unity of Biochemistry.Phenomena suggesting the existence ofother modes of cell-cycle regulation inanimal cells have been shown to beconsistent with the model for the bacterialdivision cycle proposed here. For example,the constancy of C- and D-periods ismimicked by the constancy of theiranalogues in animal cells, the S andG2 periods. The variation in cell sizewith growth rate and the rules regardingsize determination also appear to holdfor all organisms. The rules of the cell-division cycle – that is, the logic anddesign principles of the division cycle – arethe same for Escherichia and escargot,Salmonella and salmon. Although thereare many detailed differences betweeneukaryotic and prokaryotic division cycles,at the deepest level the principle of theUnity of Cell Biology proposes that thedivision cycle is ultimately regulated bythe continuous accumulation of somemolecule that is titrated against a fixedamount of another cell component. Fromthis viewpoint, cell-cycle-specific eventsidentified in eukaryotic cells are theresult or symptom of a deeper regulatoryprinciple that is produced in a cell-cycle-independent manner. The question forfuture analysis is whether there is somedeeper rule that ensures this commonpattern in all cells. I conjecture that this isindeed the case and hope that a search forthe underlying meaning of size control inall cells is now the new goal of cell-cycleresearch.

mcb200300028


Bibliography

Books and Reviews

Bramhill, D. (1997) Bacterial cell division, Annu.Rev. Cell. Dev. Biol. 13, 395–424.

Cooper, S. (1984) The Continuum Model asa Unified Description of the Division Cycleof Eukaryotes and Prokaryotes, in: Nurse, P.,Streiblova, E. (Eds.) The Microbial Cell Cycle,•CRC Press, pp. 7–18.

Q1Cooper, S. (1991) Synthesis of the cell surface

during the division cycle of rod-shaped, gram-negative bacteria, Microbiol. Rev. 55, 649–674.

Cooper, S. (1991) Bacterial Growth and Division:Biochemistry and Regulation of Prokaryotic andEukaryotic Division Cycles, Academic Press,San Diego.

Cooper, S. (1996) Segregation of CellSurface Structures, in: Neidhardt, F.C.,Ingraham, J.L., Low, K.B., Magasanik, B.,Schaechter, M., Umbarger, H.E. (Eds.) Es-cherichia coli and Salmonella: Cellular andMolecular Biology, 2nd edition, AmericanSociety for Microbiology, Washington, DC,pp. 1652–1661.

D’Ari, R. (2001) Cycle-regulated genes and cellcycle regulation, BioEssays 23, 563–565.

Draper, G.S., Gober, J.W. (2002) Bacterial chro-mosome segregation, Annu. Rev. Microbiol. 56,567–597.

Gordon, G.S., Wright, A. (2000) DNA segrega-tion in bacteria, Annu. Rev. Microbiol. 54,681–708.

Helmstetter, C.E. (1996) Timing of SyntheticActivities in the Cell Cycle, in: Neid-hardt, F.C., Ingraham, J.L., Low, K.B., Ma-gasanik, B., Schaechter, M., Umbarger, H.E.

(Eds.) Escherichia coli and Salmonella: Cellularand Molecular Biology, 2nd edition, AmericanSociety for Microbiology, Washington, DC,pp. 1627–1639.

Helmstetter, C.E. (1996) Timing of SyntheticActivities in the Cell Cycle, in: Nei-dhardt, F.C., Curtiss III, R., Ingraham, J.L.,Low, K.B., Magasanik, B., Reznikoff, W.S., Ri-ley, M., Schaechter, M., Umbarger, H.E. (Eds.)Escherichia coli and Salmonella: Cellular andMolecular Biology, ASM Press, Washington,DC, pp. 1627–1639.

Levin, P.A., Grossman, A.D. (1998) Cell cycle:the bacterial approach to coordination, Curr.Biol. 8, R28–R31.

Lutkenhaus, J., Addinall, S.G. (1997) Bacterialcell division and the Z ring, Annu. Rev.Biochem. 66, 93–116.

Lutkenhaus, J., Mukherjee, A. (1996) CellDivision, in: Neidhardt, F.C., Ingraham, J.L.,Low, K.B., Magasanik, B., Schaechter, M.,Umbarger, H.E. (Eds.) Escherichia coli andSalmonella: Cellular and Molecular Biology, 2ndedition, American Society for Microbiology,Washington, DC, pp. 1615–1626.

Messer, W., Weigel, C. (1996) Initiation of Chro-mosome Replication, in: Neidhardt, F.C., In-graham, J.L., Low, K.B., Magasanik, B., Sch-aechter, M., Umbarger, H.E. (Eds.) Escherichiacoli and Salmonella: Cellular and MolecularBiology, 2nd edition, American Society for Mi-crobiology, Washington, DC, pp. 1579–1601.

Newton, A., Ohta, N. (1990) Regulation of the celldivision cycle and differentiation in bacteria,Annu. Rev. Microbiol. 44, 689–719.

Zyskind, J.W., Smith D.W. (1992) DNA replica-tion, the bacterial cell cycle, and cell growth,Cell, 69, 5–8.

mcb200300028

QUERIES TO BE ANSWERED BY AUTHOR (SEE MARGINAL MARKS Q..)

IMPORTANT NOTE: Please mark your corrections and answers to these queries directlyonto the proof at the relevant place. Do NOT mark your corrections on this query sheet.

Q1. Please provide the place of publication for this reference.

bacterial growth and division - u-m personal world wide web server

Documents