Bacterial cell cycle: Regulatory strategies to increase survival

(1)

Bacterial cell cycle:

Regulatory strategies to increase survival

June 2016

Santiago Caño-Muñiz

University of Groningen

(2)

(3)

1. Introduction ... 1

2. Cell cycle ... 1

2.1 B-period ... 2

2.2 C-period ... 3

2.3 D-period ... 3

3. Regulation of cell cycle ... 4

3.1 Signal driven regulation ... 4

3.2 Noise-based regulation ... 6

3.3 The joint model ... 7

4. Outlook ... 8

5. References ... 8

Change is an inherent property of environ- ments, and thus organisms must adapt to change in order to survive. During the last few decades, the knowledge regard- ing how cells accommodate to changes in nutrient availability within the cell cycle has increased exponentially. Additionally, there has also been described a completely new way of regulating the cell cycle based on stochastic switch independent of exter- nal signals. In this master thesis, the signal mechanism used by E. coli to regulate the cell cycle and the genetic base responsible for the stochastic switch are explained, fol- lowed by a discussion of potential conse- quences of both mechanisms.

1. INTRODUCTION

2. CELL CYCLE

Author affiliation:

Caño-Muñiz, S.

29 May, 2016

Supervisor: Scheffers, D.J.

¹

1 - Department of Molecular Microbiology, University of Groningen, The Netherlands

One property that distinguishes biological processes from most other physical processes is the pro- priety of life to reproduce (Weber). Reproduction entails that cells self-sustain and are able to con- struct self-replicas. Logically, the notion of replication has been proposed as the sine qua non for evolution (Weber), on grounds that selection acts over populations and generations rather than just individ- uals. Prokaryotes represent the simplest systems that fill all these requirements to be considered as living systems. However, beyond the apparent simplicity, we have not yet been able to create artificial self- replicating systems because cell division relies on an intricate regulatory structure. This regulatory structure is necessary because life is a far-from-equilib- rium process, so it requires a constant input of energy and matter from the environment to allow growth.

As such, the regulatory structure connects the cell cycle with the environmental conditions (Wang &

Levin, 2009).

Natural environments tend to oscillate, as what

happens with day-night cycles or tides, but also abrupt changes such as drought, toxins, or antibiotics can disrupt the habitat conditions (Kussell & Leibler, 2005). These natural fluctuations constrain cell division in such a way that they limit the availability of building blocks required for growth (Wang & Levin, 2009). Consequently, natural selection has favoured a set of mechanisms to cope with these natural fluctuations and adjusts the cell cycle in order to com- pensate for changes in the environment (Wang &

Levin, 2009). How bacteria might adjust their machinery is a vast field of research with a plethora of open questions.

Recent work suggests that bacterial cell division depends on multiple signalling pathways, which transmit nutritional and growth information directly to the cell cycle machinery (Wang & Levin, 2009).

Nevertheless, since the last decade, a group of novel mechanisms that regulate cell division independent of signalling, and fundamentally based on noise, has also been reported (Yamaguchi & Inouye, 2011).

These recent mechanisms allow populations to generate diversity, which contributes enormously to bac- terium evolution (Yamaguchi & Inouye, 2011). The present work aims to review both of the following cat- egories, the signal driven or canonical mechanisms, and the noise based, or non-canonical mechanisms, in order to integrate the role of both in the process of cell division in prokaryotes. For that purpose, lit- erature surrounding the model organism Eschericha coli (E. coli) will primarily be reviewed.

During the process of cell division, an existing cell acts as the “template” and must double in size, repli- cate its DNA, and separate the resulting “sister” chromosome molecules, in order to finally separate from the new cell and produce two “identical” daughter cells. All of these events must be coordinated and controlled spatially and temporally as well as with the environment to ensure that each daughter cell is fully functional.

(5)

2.1 B-period

During the B-period, the cell increases in size and mass until it begins a new round of chromosomal replication (DePamphilis & Bell, 2010). However, it is crucial to notice that in rich environments, where generation times are shorter than the chromosome duplication time, the B period is completely omitted and new born cells start directly in the C-period (see section 2.2).

In E. coli, the key event that separates the B-period from the C-period is the formation of a replication fork by the binding of the DnaA protein to the origin of the replication sequence (oriC). DnaA is a monomer (inactive) that, when it polymerizes (active), acts as plat- form for the loading of DnaC, which in turn recruits

Despite that the cell cycle in prokaryotes is not as precisely defined as in eukaryotes, it has traditionally been divided into three phases: The B-period, which covers the time since cell birth until the initiation of chromosomal replication, the C-period, which entails the time required for the duplication of the chromosome, and the D-period, which includes the end of the replication process and the complete separation of the new born cells, also known as cytokinesis (DePamphilis & Bell, 2010).

DUE

R1 IHF R5 I1 I2 R2 Fis R3 I3 R4

Fis

DUE

IHF

DUE C/D-period

Early B-period Late B-period

Initiation chromosome duplication +ATP

+Fis

+ATP +IHF -Fis

the helicase DnaB, and will subsequently generate the DNA unwinding region (DUE) (DePamphilis

& Bell, 2010; Mott & Berger, 2007). The key factor that triggers the initiation of the replication is the ratio of active/inactive DnaA (Mott & Berger, 2007;

Donachie & Blakely, 2003). The oriC region contains 5 binding boxes accessible to the active and inactive form and three that are only accessible to the active DnaA (Mott & Berger, 2007). (see Figure 1). In addition, initiation of replication is coordinated with two other transcription factors, Factor for inversion stimulation (Fis) and the Integration Host Factor (IHF) (Mott

& Berger, 2007; DePamphilis & Bell, 2010).

Fis suppresses IHF binding and DnaA binding at R3 (see Figure 1), which consequently causes a delay in replication and apparently links cell growth through delaying the initiation time (DePamphilis & Bell, 2010). Conversely, IHF-dependent stimulation is most likely due to its ability to bend the oriC backbone (see Figure 1), which redistributes the DnaA location and helps in recruiting additional DnaA molecules (Mott

& Berger, 2007). The opposing activity of these two molecules apparently ensures the abrupt transition between phases, increasing both precision of chromosome replication timing and initiation synchrony (DePamphilis & Bell, 2010).

Figure 1: Initiation of the replication organization at oriC. The oriC region contains 5 binding boxes for active/inactive DnaA (black boxes) and three which are only accessible for the active form. During the B and C-periods oriC is practically free. Association of Fis changes the structure of oriC facilitating DnaA (orange elipses) binding to weak sites (Eraly B-period). An increase of ATP plus IHF binding causes a displacement of Fis and more DnaA join the structure resulting in the partial unwinding of the DUE.

(6)

2.2 C-period

The C-period is characterized by the elongation of the DNA replicon until it reaches its terminus region (terC in E. coli) and begins the process of chromosomal segregation. From the oriC, two replication forks run in opposite directions.

At the head of this process are the topoisomer- ase II (also called gyrase) that releases the tension and the helicase that separates the two DNA strings.

Behind this, the DNA pol III polymerizes the new DNA strands. Interestingly, it has been proven in Bacillus subtilis (B. subtilis) and E. coli that doubling the time can still be shorter than the time required for the genome duplication, which is somewhat constant, with an approximate 40-minute duration. The solution to this paradox came in 1968 when Cooper &

Helmstetter presented the multi-fork theory (Cooper

& Helmstetter, 1968). In rapidly growing cells, each chromosome re-initiates a new replication fork on the newly synthetized chromosomes before the first round has terminated, thereby sustaining four to eight sets of replication forks simultaneously (Figure 2A). Moreover, cells must ensure that at least one

Sister chromosomes are moved apart in a multistep process, which involves the active segregation of the newly synthesized origin regions, a condensation- driven partitioning of the bulk of the chromosomes, and finally, the separation of the terminus regions.

To ensure proper sorting of the chromosomes, the cell uses molecular signposts called KOPS. KOPS are short, conserved sequence motifs that become highly overrepresented in the genome region towards terC in the terminus region (Grainge, 2010; Thanbichler, 2010). In those regions, the FtsK protein forms hex- americ complexes that translocate in the direction determined by the polarity of the KOPS elements (Thanbichler, 2010). Finally, when FtsK reaches the resolution point, it interacts with XerD-XerC (see

Figure 2: A) Cooper & Helmstetter model. Slow-growing cells accomplish a single round of replication per division cycle.

During replication each cell has only two copies of oriC (Stars) and one copy of the terC (black box). Rapidly growing cells’ chromosomes re-initiate replication before the previous one has terminated, although only cytokinesis is achieved.

During multifork replication, cells can have four or more copies of the region proximal to oriC and one copy of the region proximal to terC. B) Chromosome resolution. FtsKC interacts with the recombinase XerD (green spheres) and thus induces the first pair of strand exchanges. The recombinase XerC (red spheres) then finalizes the resolution, restoring the two original chromosomes. C) Assembly of the divisome comples. First players appear in solid colour. EzrA tether FtsZ to the membrane in cooperation with FtsA and ZapA. Later, the divisiome components (dashed colored) are recruited. EnvC and AmiC act on the cell wall catalysing the cell separation.

round of replication is finished before cytokinesis, in order for each daughter to receive at least one complete chromosome. Once genome duplication begins, the two copies of oriC migrate towards opposite cell poles, and the terminus region terC moves towards the centre of the cell where cytokinesis will occur (Thanbichler, 2010).

2.3 D-period

Slow Growth

Fast growth

CM CW Divisome

Complex

FtsZ ZapA

FtsA

SepF EzrA XerC FtsK

XerD

A B C

EnvC AmiC

(7)

3. REGULATION OF CELL CYCLE

Figure 2B). This complex catalyses chromosome dimer resolution (Badrinarayanan & Le, 2015).

Septum formation finishes the process of cell division. The septum conforms the structure that will physically separate the two daughter cells. The first step is the proto-ring assembly, promoted by ZapA, ZipA and FtsA, which stabilize bundles of polymerized FtsZ, tethering these bundles to the inner membrane and forming the Z-ring (See figure 2C) (Skoog & Daley, 2012; Thanbichler, 2010). Then, after the chromosome segregation, which likely is signalled through FtsK (Grainge, 2010), a macrostructure known as

“divisome” complex is formed (see figure 2C) (Eraso, et al., 2014; Trip & Scheffers, 2015). The membrane- bound cell division articulates the invagination of the cell wall and membrane to form a division septum.

Septum formation is complete and peptidoglycan hydrolases hydrolyse the completed cross wall, pro- ducing two new-born cells.

Replication is a costly process that requires duplicat- ing all cell components. As such, the metabolic state is a conditioner of cell division; implying information is integrated into the cell cycle through checkpoints before starting the essential and delicate process.

Nevertheless, responsive strategies generate homo- geneous responses, which means when a population “senses” that the conditions are appropriate, all cells will trigger growth behaviour. Thus, microorgan- isms combine responsive mechanisms with stochastic switch between growing and non-growing phenotype in order to introduce variation in the population. In the ensuing text, the basis of both mechanisms and their further implications are discussed.

3.1 Signal driven regulation

Taking the birth of a cell (B-period) as starting point, the initiation of chromosome replication is the entry point for metabolic control of cell division (Wang &

Levin, 2009). DnaA is the central focus for regulation of this initiation process. In E. coli, it has been demonstrated that the affinity of DnaA for 3 of its binding boxes in oriC are dependent on ATP-DnaA binding (Mott & Berger, 2007; McGarry, et al., 2004). This functions as a metabolic indicator for initiating the chromosome duplication (DePamphilis & Bell, 2010).

Additionally, in order to initiate the chromosome duplication, a critical ratio of active/inactive DnaA is required (see section 2.1). Within each cell cycle,

DnaA is diluted and it has to be synthetized de novo, which makes DnaA highly sensitive to amino acid starvation (Donczew, et al., 2014). Moreover, expression of DnaA is controlled by ppGpp (Zyskind & Smith, 1992). This molecule has been called the “stress alar- mone” because it is induced during amino acid starvation (via RelA), and also during carbon limitation (via SpoT) (Ferullo & Lovett, 2008). Accordingly, an increase of ppGpp concentration has been related to a dimmed transcription of the DnaA operon, however it still remains unclear whether this effect is direct or indirect (see Figure 3) (Nazir & Harinarayanan, 2016).

After the formation of the replication fork, cells activate a safety mechanism to prevent the sequen- tial initiation of chromosome replication. The DnaA binding boxes in oriC are occupied by SeqA, which prevents the re-initiation of chromosome duplication (Donczew, et al., 2014). Nevertheless, the com- petition for the binding sites with the methylase Dam overrides sequestration during rich nutrient conditions such that it becomes possible to form the multifork structure (see section 2.2) (Skarstad & Løbner- Olesen, 2003).

C-period of E. coli was originally estimated to be constant independently of conditions (Cooper &

Helmstetter, 1968; Donachie & Blakely, 2003), but when generation times are longer than 1 hour, the C-period can increase by more than twofold (Wang

& Levin, 2009). The prolongation of the C-period is under control of ppGpp. It is not apparent why replication rates are sensitive to the growth rates in E.

coli, it is, however, clear that since elongation is an expensive process that requires large amounts of cellular resources (mainly dNTP), nutrient availability has a substantial impact on the elongation of replication (Figure 3). Ribonucleotide reductase is the enzyme that catalyses the reduction of NDPs to dNTP. This enzyme plays a key role in determining the rate of DNA synthesis during the C-period (Herrick & Sclavi, 2007). Regulation occurs on two levels: through allo- steric control, and by regulation of transcription. The essential regulators are cell cycle gene controllers such as DnaA (Sun & Fuchs, 1992), oxidative stress, oxygen content and stress on the replication machinery (Gon & Beckwith, 2006). This tight regulation reflects that dNTP pool imbalance results in chromosome anomalies and mutation (Gon & Beckwith, 2006).

Furthermore, the regulation at the end of the cell cycle is dominated by the formation of the Z-ring.

(8)

During the D-period, carbon availability is the primary metabolic controller of FtsZ (Figure 3), a major com- ponent of the Z-ring (Daniel & Errington, 2003). Here, UDP-glucose is used in proxy of the metabolic state.

When E. coli is growing in a carbon rich medium, the pool of UDP-glucose increases, binding it to OpgH, upon which a conformational change happens. This change happens such that The OpgH-UDP glucose activated can sequestrate free monomers of FtsZ preventing maturation of the cytokinesis ring, delaying division and thereby increasing cell size (Hill, et al., 2013). Conversely, under conditions in which UDP–glucose levels are low, such as a carbon poor medium, interaction between UDP-glucose and OpgH is reduced such that FtsZ assembly is able to proceed unimpeded, thus reducing cell size (Hill, et al., 2013).

Another metabolic mechanism controlling cell division is the concentration of pyruvate through PdhR.

In the case of a low glycolytic flux, indicated by lower levels of pyruvate, PdhR represses mraZ expression (Göhler, et al., 2011), which subsequently regulates the dcw gene cluster that controls cell wall formation during cytokinesis (Eraso, et al., 2014).

Figure 3: Schematic representation of the cell cycle regulation. Red lines symbolize repression and green lines represent positive induction. DnaA represses its own expression (negative feedback loop) and also competes with SeqA for the binding boxes at oriC. ATP activates DnaA and induces replication initiation. Nutrient limitation inhibits both initiation and replication indirectly and via ppGpp. FtsZ polymerization is a key regulator of cytokinesis. This is affected directly by indole (causes membrane depolarization) and indirectly by pyruvate and nutrient availability. N: nitrogen, C: carbon, aa:

amino acids.

On top of these interactions, the formation of the Z-ring also receives information about the stress state of the cell from a variety of channels. First, indole is a quorum-sensing signalling molecule that disrupts the membrane potential of E. coli. By reducing the elec- trochemical potential across the cytoplasmic membrane, indole deactivates MinCD oscillation and prevents formation of the FtsZ ring that is a prerequisite for division (Chimerel, et al., 2012; Strahl & Hamoen, 2010). This mechanism is triggered during the transition to a stationary phase controlled by the catabo- lite repressor protein and also by the alkaline sensor TorP. Secondly, DNA damage can also arrest the cytokinesis. SulA is a cell division inhibitor that is produced as part of the SOS response in E. coli (Adams

& Errington, 2009), and rapidly stalls cell division by both preventing the assembly of nascent Z rings and facilitating the disassembly of existing Z rings. Once the DNA lesions are repaired, SulA levels rapidly decline (Adams & Errington, 2009).

B-period

C-period

D-period Cell incrase in mass and size

0-25 min

Chromosome duplication 40 min Cytokinesis:

Cell split 1 min

DnaA OriC

ATP ppGpp

aa starvation C,N starvation OriC Tritiation SeqA

+

FtsZ Pyruvate

Chromosome Elongation aa starvation

C,N starvation UDP-Gluc.

OpgH

PdhD

MraZ

Quorum Sensing Indole

(9)

3.2 Noise-based regulation

Diversity represents a key factor for average fitness of populations (Kussell, et al., 2005; Kussell & Leibler, 2005). If cell cycle regulation would rely only on a signal-driven mechanism, a typical bacterial population will exhibit a high degree of homogeneity, which imposes a handicap in changing environments (Kussell & Leibler, 2005). Therefore, along with signal-driven mechanisms, natural selection must favour certain mechanisms regulating cell division based on intrinsic factors that increase cell diversity. It has been proposed that the regulatory architecture of Toxin- Antitoxin (TA) modules rely on cell noise, allowing populations to generate diversity independent from the environment (Koh & Dunlop, 2012; Feng, 2014).

As such, TA modules are good candidates for a source of population variance.

TA modules are operons composed of two genes:

the toxin gene encodes a toxic protein that causes cell arrest, whereas the antitoxin can be either RNA or protein, which counteracts the cognate toxin (Yamaguchi & Inouye, 2011). These TA modules have been reported in almost all free-living organisms and they are highly redundant in pathogen organisms (Pandey, 2005); for instance, it has been revealed that E. coli has 37 putative TA modules (8 of them well characterized) (Pandey, 2005). TA modules are affected by environemntal conditions (especially stress indicators), however their loose regulation mechanisms causes that low noise fluctuation leads to a great impact in phenotype. The resulting cell-to-cell differences allow for the emergence of phenotypic subpopulations (Koh & Dunlop, 2012);

Feng, 2014), which consequently increase population fitness (Kussell & Leibler, 2005). The TA modules are divided into 5 families, but in E. coli, only four of these families have been reported (Table 1) (Yamaguchi &

Inouye, 2011).

TA type I are defined by a protein toxin and a small non-coding RNA that blocks toxin translation (see figure XA). There are three described modules in E.

coli. tisBA is the most representative case. TisB is a protein that penetrates the cytoplasmic membrane and consequently dissipates the membrane potential. As a result, cells are neither able to assemble the Z-ring nor sustain ATP synthesis (Dörr, et al., 2010).

The antitoxin TisA is a phantom gene and only the fragment istR-1 is transcribed to RNA. This sequence inhibits TisB translation, although the IstR-1 binding

site is distant from the TisB ribosome binding site.

Induction of the SOS response causes increased transcription of tisB (Unoson & Wagner, 2008) and deple- tion of the IstR-1 pool, thus leading to accumulation of the tisB mRNA, translation of the TisB peptide, and slowed growth. Interestingly, tisBA module is only active during exponential growth. HokB-SokB is another TA type described in E. coli. HokB encodes a polypeptide that induction experiments have shown to be toxic to E. coli host cells, resulting in cell growth arrest, morphological changes, and rapid cell killing.

Importantly, the chromosomal partitioning protein ObgE triggers the toxic activity of HokB-SokB, especially in the presence of ppGpp (Verstraeten, 2015).

TA Type II are the most abundant in E. coli and they are defined as a protein toxin and a labile protein antitoxin that blocks toxin activity as well as represses TA operon expression (see figure XB). Within this family, the system mazEF (similar to chpBK and hicAB) is the best-characterized TA module (Kolodkin-Gal, 2007).

MazF functions as endoribonuclease that, when

Family Module Toxin Target

Type I HokSok Cell Membrane

TisBA Cell Membrane

SymRE mRNA

Type II MazEF mRNA

ChpBI-ChpBX mRNA

HicAB mRNA

MqsRA mRNA

RelBE Ribosome

DinJ-YafQ Ribosome

HigAB Ribosome

Type IV CbtAB FtsZ

Type V GhoTS Cell Membrane

Table 1: Main TA modules in E. coli.

(10)

induced, cleaves almost all cellular mRNAs, thereby inhibiting cell growth (Yamaguchi & Inouye, 2011).

The antitoxin MazE binds MazF, blocking toxin activity and also blocking the expression of the operon;

in addition, MazE is very labile due to its continu- ous degradation by proteases (Kolodkin-Gal, 2007;

Christensen & Gerdes, 2003). This regulatory architecture creates a system very sensitive to nutrient influx (Kolodkin-Gal, 2007). In addition, it has been reported that mazEF responds to ppGpp and sigma stress factors, which therefore triggers this system under various stress, such as antibiotic stress, heat shock, DNA damage and other kinds of stresses which are shown to elicit mazEF cell arrest (Hazan, et al., 2004; Hazan & Engelberg-Kulka, 2004). Another example of endoribonuclease is mqsRA. The toxin MqsR cleaves RNA. MqsR is the toxin that cleaves to CGU sites (Yamaguchi, et al., 2009), and the antitoxin MqsA binds MqsR and thereby inhibits its activity and also represses the expression of the operon (Yamaguchi & Inouye, 2011). Interestingly, this TA module responds to the addition of the quorum-sensing autoinducer, AI-2 (Barrios, 2006), which makes this TA module dependent on population context.

For example, it has been demonstrated that MqsR is induced during biofilm formation (Barrios, 2006).

MqsR may inactivate all E. coli mRNAs except for 14 genes that are essential for biofilm formation (Yamaguchi & Inouye, 2011).

Alternatively, TA modules can block protein synthesis by occupying the A-site of ribosomes, leading to the cleavage of mRNAs preferentially between the second and third nucleotides of the termination codon protease (Christensen & Gerdes, 2003). This results in the inhibition of translation elongation, and hence the blockade in protein synthesis causes cell arrest. (Christensen & Gerdes, 2003). Here, relBE is the best example of this family (RelBE, DinJ-YafQ and HigAB). Remarkably, the action mechanism of the antitoxin is the same as that of the mazEF family, a labile antitoxin binds to the toxin blocking its activity and repressing TA expression. Furthermore, RelBE is another biological sensor for amino acid starvation.

When cells are starved of amino acids, Lon protease degrades RelB; RelB degradation frees RelE and de-represses transcription of relBE (Christensen &

Gerdes, 2003).

Furthermore, in E. coli, the only one TA family that has been described, which directly affects the cell division machinery, is TA type IV. The protein pair CbtA and CbtB directly interact with FtsZ and MreB.

3.3 The joint model

The uncertainty of nature constantly challenges bacteria, generating stress. However, even the simplest of organisms, such as bacteria, are capable of processing information in a highly sophisticated manner, adapting to varying environments and evolv- ing new functions. As detailed in this review (see section 3.1), the model organism E. coli uses a set of inner parameters such as UDP-glucose (Hill, et al., 2013) or ATP (DePamphilis & Bell, 2010) to infer the external environment’s state. Due to the fragility of the replication process, cells try to maximize information, so check points can be seen as a solution to it.

For instance, the use of DnaA as an estimator of the cellular mass ensures that cell size will remain constant along generations (Chien, et al., 2012; Anders, et al., 1989). This mechanism provides high fidelity information by silencing the noisy metabolic oscilla- tions, and only allowing the formation of a replication fork once a threshold is surpassed (DePamphilis

& Bell, 2010). This is at the expense of a very intricate regulation network around DnaA and a lag-time required to incorporate the signals from ppGpp, ATP and so on. Therefore, under certain circumstances, a “blind bet-hedging” can be a better strategy than

“always knowing” (Kussell & Leibler, 2005).

The toxin CbtB inhibits GTP-dependent polymerization of FtsZ and ATP-dependent polymerization of MreB (Masuda, et al., 2012), and its counterpart CbtA stimulates these polymerizations. The other two members of the family, Ykfl-YafW, and YpjF-YfjZ have not been characterized (Yamaguchi & Inouye, 2011).

Finally, recently a new category of TA modules, named Type V, has been reported (Wang, et al., 2012). What characterizes this family is that the antitoxin blocks toxin activity by cleaving toxin mRNA (Wang, et al., 2012). The only reported member is the operon GhoT. GhoT is predicted to be a small, hydrophobic polypeptide that causes cell lysis when overproduced (Wang, et al., 2012). Similar to the SOS-induced inner membrane toxin TisB (Unoson &

Wagner, 2008), GhoT damages the cytosolic membrane, which reduces membrane potential and subsequently ATP production (Wang, et al., 2012), and when it is produced for long periods can also lead to cell death (Cheng, et al., 2014). Interestingly, this operon’s expression is stimulated by MqsR (Cheng, et al., 2014; Wang, et al., 2012), which might imply a role during biofilm formation (Wang, et al., 2012).

(11)

4. OUTLOOK

5. REFERENCES

Currently, the process that allows cells to divide is well described. In addition, the last few years have been especially prolific in giving us the first mechanism that connects metabolism to the division machinery. However, we are far from understanding even the most fundamental behaviour of cells and organisms from the systemic point of view. Recently, the discovery of bet-hedging and non-growing cells has posed a challenge to the way in which we per- ceive cell division. They break within the field of cell division regulation. Strikingly, these modules are not based on external signals, but in internal cellular noise. The evolutionary meaning of this is a matter

Adams, D. W. & Errington, J., 2009. Bacterial cell division:

assembly, maintenance and disassembly of the Z ring. Nature Reviews Microbiology.

Anders, L.-O.y otros, 1989. The DnaA protein determines the initiation mass of Escherichia coli K-12. Cell, 57(5), pp.

881-889.

Arends, R. S. J. & Weiss, D. S., 2004. Inhibiting cell division in Escherichia coli has little if any effect on gene expression.

Journal of Bacteriology, 186(3), pp. 880-884.

Badrinarayanan, A. & Le, T. B. K., 2015. Bacterial Chromosome Organization and Segregation. Annual Review of Cell and Developmental Biology , Nov, Volumen 31, pp. 171-199 .

Barrios, A. Z. R. H. Y. a. Y. L., 2006. Autoinducer 2 controls biofilm formation in Escherichia coli through a novel motility quorum-sensing regulator (MqsR, B3022).. J. of Bacteriology, 188(1), pp. 305-316.

Buts, L. y otros, 2005. Toxin–antitoxin modules as bacterial metabolic stress managers. Trends in biochemical science, Dec, 30(12), pp. 672-679.

Cheng, H. Y., Soo, V. W. C. & Islam, S., 2014. Toxin GhoT of the GhoT/GhoS toxin/antitoxin system damages the cell membrane to reduce adenosine triphosphate and to reduce growth under stress. Environmental Microbiology, January, 16(6), pp.

1741-1754.

Chien, A. C., Hill, N. S. & Levin, P. A., 2012. Cell size control in bacteria. Current biology, February, 25(3), pp. 385-391.

Chimerel, C. y otros, 2012. Indole prevents Escherichia coli cell division by modulating membrane potential.. Biochim.

Biophys. Acta, 1818(7), pp. 1590-4.

Christensen, S. K. & Gerdes, K., 2003. RelE toxins from bacteria and Archaea cleave mRNAs on translating ribosomes, which are rescued by tmRNA. Molecular microbiology, June, 48(5), pp. 1389-1400.

Cooper, S. & Helmstetter, C. E., 1968. Chromosome replication and the division cycle of Escherichia coliBr. Journal of Molecular Biology, 31(3), pp. 519-540.

When environmental fluctuations are irregular, the increase of complexity in the sensing machinery to provide reliable information becomes too expensive, thus natural selection will punish the complexity of such organisms (Donaldson-Matasci & Bergstrom, 2013; Kussell & Leibler, 2005). Moreover, the use of non-optimal sensors can read white noise giving contradictory signals indicating a constant change of phenotype (Perkins & Swain, 2009). In these cases, a blind bet-hedging strategy gives a better perfor- mance (Kussell & Leibler, 2005). Bet-hedging arose as a product of stochastic switching between dif- ferent phenotypes (Rivoire & Leibler, 2014). The TA modules might not make sense from the single cell point of view since they cause a decrease in growth (Buts, et al., 2005), but on the population level, these cells increase diversity where non-growing cells act as an “insurance policy” against unpredictable risk (Rivoire & Leibler, 2014). Consequently, this causes an increase of inclusive fitness (Kussell, et al., 2005;

Gefen & Balaban, 2009).

To sum up, cell cycle regulation is a complex machine that provides an optimal solution for predict- able problems such as starvation. Nevertheless, this network also has its limitations, and it cannot provide completely reliable information. Therefore, despite stochastic switch of phenotype displacing optimal fitness, when cells are dealing with high uncertainty, a risk diversification strategy is more robust (Kussell, et al., 2005; Kussell & Leibler, 2005). This can be reflected in the fact that pathogenic organisms present high redundancy in TA modules, while commensal microbes living inside a constant environment present low numbers of TA, with wild organisms lying somewhere between these two extremes (Pandey, 2005).

of debate, where bet-hedging as a risk diversification strategy is gaining supporters, but other theories could also resolve some specific cases where “multi- cellular” behaviour is required.

These steps provide a foundation for unravelling the bacterial cell cycle as a whole. Together with advances in systems biology, that combine theoreti- cal modelling with empirical proving, it will be possible to understand the cell cycle, which would encap- sulate many areas of biology. For instance, the clinical relevance of cell arrest in persistent cells is undeni- able. There are some preliminary results that indicate it is possible to use metabolic signals to trigger persistent awakening. Another approach is to force cell cycle arrest in microbial factories to boost the yield of bioproducts. However, one of the most exciting dis- coveries will be the synthetic creation of self-dividing structures, which would lead to a full understanding of cell division and a revolution in the bioeconomy.

(12)

Dörr, T., Vulić, M. & Lewis, K., 2010. Ciprofloxacin causes persister formation by inducing the TisB toxin in Escherichia coli. PLoS Biology.

Daniel, R. A. & Errington, J., 2003. Control of cell morpho- genesis in bacteria: two distinct ways to make a rod-shaped cell.

Cell, June, 113(6), p. 767–776.

DePamphilis, M. & Bell, S., 2010. Genome Duplication. New York: Garland Science.

Donachie, W. D. & Blakely, G. W., 2003. Coupling the initiation of chromosome replication to cell size in Escherichia coli..

Curr. Opin. Microbiol., 6(2), pp. 146-50.

Donaldson-Matasci, ,. M. C. & Bergstrom, C. T., 2013. When unreliable cues are good enough. The American Naturalist, 182(3).

Donczew, R., Zakrzewska-Czerwińska, J. & Zawilak-Pawlik, A., 2014. Beyond DnaA: the role of DNA topology and DNA meth- ylation in bacterial replication initiation. Journal of Molecular Biology, 426(12), pp. 2269-2282.

Eraso, J. M. y otros, 2014. The highly conserved MraZ protein is a transcriptional regulator in Escherichia coli. Journal of Bacteriology, 196(11), pp. 2053-2066.

Feng, J. K. D. B.-J. E. a. L. H. 2., 2014. Growth feedback as a basis for persister bistability.. Proc. Nat. Acad. of Sci., 111(1), p. 544–549.

Ferullo, D. J. & Lovett, S. T., 2008. The stringent response and cell cycle arrest in Escherichia coli. {PLoS} Genet.

Göhler, A.-K.y otros, 2011. More than just a metabolic regulator-elucidation and validation of new targets of PdhR in Escherichia coli. BMC systems Biology.

Gefen, O. & Balaban, N. Q., 2009. The importance of being persistent: heterogeneity of bacterial populations under antibiotic stress.. {FEMS} Microbiol. Rev., 33(4), pp. 704-17.

Gon, S. & Beckwith, J., 2006. Ribonucleotide reductases:

influence of environment on synthesis and activity.. Antioxid.

Redox Signal., 8(5-6), pp. 773-80.

Grainge, I., 2010. FtsK--a bacterial cell division checkpoint?.

Mol. Microbiol., Dec, 78(5), p. 1055–7..

Hazan, R. & Engelberg-Kulka, H., 2004. Escherichia coli mazEF-mediated cell death as a defense mechanism that inhibits the spread of phage P1. Molecular Genetics and Genomics.

Hazan, R., Sat, B. & Engelberg-Kulka, H., 2004. Escherichia coli mazEF-mediated cell death is triggered by various stressful conditions. Journal of bacteriology.

Herrick, J. & Sclavi, B., 2007. Ribonucleotide reductase and the regulation of DNA replication: an old story and an ancient heritage.. Mol. Microbiol., 63(1), pp. 22-34.

Hill, N. S., Buske, P. J., Shi, Y. & Levin, P., 2013. A Moonlighting Enzyme Links Escherichia coli Cell Size with Central Metabolism.

PLoS Genetics, 9(7), p. e1003663.

Hutchison, C. A. y otros, 2016. Design and synthesis of a minimal bacterial genome.. Science, 351(6280), p. aad6253.

Hu, Z. & Lutkenhaus, J., 2001. Topological regulation of cell division in E. coli: spatiotemporal oscillation of MinD requires stimulation of its ATPase by MinE and phospholipid. Molecular cell, 7(6), pp. 1337-1343.

Koh, R. S. & Dunlop, M. J., 2012. Modeling suggests that gene circuit architecture controls phenotypic variability in a bacterial persistence network. {BMC} systems biology.

Kolodkin-Gal, I. H. R. G. A. a. C. S., 2007. A linear pentapep- tide is a quorum-sensing factor required for mazEF-mediated cell death in Escherichia coli. Science, (2007)..

Kussell, E., Kishony, R., Balaban, N. Q. & Leibler, S., 2005.

Bacterial persistence a model of survival in changing environments. Genetics.

Kussell, E. & Leibler, S., 2005. Phenotypic diversity, population growth, and information in fluctuating environments..

Science, 309(5743), pp. 2075-8.

Masuda, H., Tan, Q., Awano, N. & Wu, K. P., 2012. YeeU enhances the bundling of cytoskeletal polymers of MreB and FtsZ, antagonizing the CbtA (YeeV) toxicity in Escherichia coli.

Mol Microbiol. , 84(5), pp. 979-89.

McGarry, K. C., Ryan, V. T., Grimwade, J. E. & Leonard, A.

C., 2004. Two discriminatory binding sites in the Escherichia coli replication origin are required for DNA strand opening by initia- tor DnaA-ATP.. Proc. Natl. Acad. Sci. {U.S.A.}, 101(9), pp. 2811-6.

Mott, M. L. & Berger, J. M., 2007. DNA replication initiation: mechanisms and regulation in bacteria. Nature Reviews Microbiology.

Nazir, A. & Harinarayanan, R., 2016. (p)ppGpp and the bacterial cell cycle.. J. Biosci., 41(2), pp. 277-82.

Pandey, D. a. G. K. 2., 2005. Toxin–antitoxin loci are highly abundant in free-living but lost from host-associated prokaryotes.. Nucleic acids research..

Perkins, T. J. & Swain, P. S., 2009. Strategies for cellular deci- sion-making. Molecular systems biology.

Rivoire, O. & Leibler, S., 2014. A model for the generation and transmission of variations in evolution. Proc. Natl. Acad.

Sci., 111(19), p. E1940–E1949.

Romberg, L. & Levin, P., 2003. Assembly dynamics of the bacterial cell division protein FtsZ: poised at the edge of stabil- ity. Annual Review of Microbiology, 57(1), pp. 125-154.

Skarstad, K. & Løbner-Olesen, ,. A., 2003. Stable co-exis- tence of separate replicons in Escherichia coli is dependent on once-per-cell-cycle initiation. The {EMBO} journal.

Skoog, K. & Daley, D. O., 2012. The Escherichia coli cell division protein ZipA forms homodimers prior to association with FtsZ. Biochemistry.

Strahl, H. & Hamoen, L. W., 2010. Membrane potential is important for bacterial cell division. E1940–E1949, 107(27), p.

12281–12286.

Sun, L. & Fuchs, J. A., 1992. Escherichia coli ribonucleotide reductase expression is cell cycle regulated.. Mol. Biol. Cell, 3(10), pp. 1095-105.

Thanbichler, M., 2010. Synchronization of chromosome dynamics and cell division in bacteria. Cold Spring Harb Perspect Biol. , 2(1), p. a000331.

Trip, E. N. & Scheffers, D. J., 2015. A 1 MDa protein complex containing critical components of the Escherichia coli divisome.

Scientific reports.

Unoson, C. & Wagner, E. G. H., 2008. A small SOS-induced toxin is targeted against the inner membrane in Escherichia coli.

Molecular microbiology.

Verstraeten, N., Knapen, W. J., Kint, C. I., Liebens, V., Van den Bergh, B., Dewachter, L., Michiels, J. E., Fu, Q., David, C. C., Fierro, A. C., Marchal, K., Beirlant, J., Versées, W., Hofkens, J., Jansen, M. and Fauvart, M. (2015) ‘Obg and membrane Depolarization are part of a microbial bet-hedging strategy that leads to antibiotic tolerance’, Molecular Cell, 59(1), pp. 9–21.

Wang, J. D. & Levin, P. A., 2009. Metabolism, cell growth and the bacterial cell cycle. Nature Reviews Microbiology.

Wang, X., Lord, D. M., Cheng, H. Y. & Osbourne, D. O., 2012.

A new type V toxin-antitoxin system where mRNA for toxin GhoT is cleaved by antitoxin GhoS. Nat Chem Biol., 8(19), pp. 855-861.

Weber, B., . Life. En: E. N. Zalta, ed. The Stanford Encyclopedia of Philosophy. s.l.:The Stanford Encyclopedia of Philosophy (Spring 2015 Edition).

Yamaguchi, Y. & Inouye, M., 2011. Regulation of growth and death in Escherichia coli by toxin–antitoxin systems. Nat

(13)

Rev Microbiol, 9(11), pp. 779-790.

Yamaguchi, Y., Park, J. H. & Inouye, M., 2009. MqsR, a crucial regulator for quorum sensing and biofilm formation, is a GCU-specific mRNA interferase in Escherichia coli. Journal of Biological Chemistry.

Zyskind, J. W. & Smith, D. W., 1992. DNA replication, the bacterial cell cycle, and cell growth.. Cell, 69(1), pp. 5-8.

(14)

Bacterial cell cycle: Regulatory strategies to increase survival