Final Report Summary - DNATRAFFIC (DNA traffic during bacterial cell division)
During cell proliferation, DNA synthesis, chromosome segregation and cell division must be coordinated to ensure the stable inheritance of the genetic material. In eukaryotes, this is achieved by checkpoint mechanisms that separate these processes in time. No such strict temporal separation exists in bacteria. Instead, bacteria display an integrated cell cycle, in which DNA synthesis, chromosome segregation and cell division are concomitant, their coordination being achieved by their interdependence. For instance, it was observed in Escherichia coli, Bacillus subtilis and Caulobacter crescentus that chromosomes serve as a scaffold for the positioning of cell division regulators, which coordinates the assembly of the bacterial cell division apparatus to the last stages of chromosome segregation and positions it between sister nucleoids. Reciprocally, the initiation of cell division activates a DNA pump, FtsK, which ensures the equal partition of genetic information in conditions that compromise chromosome segregation.
Bacterial chromosomes are generally covalently closed circular DNA molecules. They carry a single origin of bidirectional replication, oriC, which defines two replication arms. Replication terminates in a region opposite of oriC on the circular chromosome map, the terminus (ter). Two apparently opposed cellular arrangements of chromosomes have been observed in model rod-shaped bacteria. Indeed, the unique chromosome of E. coli adopts a transversal organisation, with oriC being positioned at mid-cell and flanked by the replication arms in newborn cells. In contrast, the unique chromosome of C. crescentus adopts a longitudinal organisation, with oriC and ter being positioned at opposite poles and the replication arms extending side by side along the long axis of the cell. Correspondingly, the mechanisms driving the spatial arrangement and the segregation of bacterial chromosomes are not conserved. In addition, they are often redundant. As a result, the overarching principles driving chromosome organisation in bacteria are still debated. To tackle this question, we decided to focus on FtsK-driven segregation because it seems to be a general feature of the bacterial cell cycle and take an evolutionary approach by comparing its action in E. coli and Vibrio cholerae, two closely related species in the phylogenetic tree of bacteria that nevertheless displayed contrasting cellular organisations and cell cycles.
FtsK can be divided in two domains. The N-terminal part of the protein, FtsKN, is an integral membrane domain. In E. coli, it is recruited to the septum soon after the formation of the tubulin-like FtsZ ring and participates in the recruitment of the peptidoglycan synthesis and degradation machinery. Thus, FtsK is an essential component of the cell division machinery. The carboxy-terminal domain of FtsK, FtsKC, assembles into hexamers over double stranded DNA and has been shown to function as a fast DNA translocase. As FtsKN anchors FtsK in the division septum, it results in the mobilization of chromosomal DNA. Thus, FtsK serves as a last safeguard to avoid division over partially segregated chromosomes. This probably explains why DNA translocation by FtsK becomes essential when the activity of the major cellular decatenase, TopoIV, is compromised, when the organization and/or the packaging of the chromosome are altered by the deletion of the condensin-like mukBEF system or by large chromosomal inversions, and when chromosome replication is affected. However, most of our knowledge on E. coli FtsK came from the study of chromosome dimer resolutio. Ondd numbers of crossovers due to homologous recombination between replicating circular sister chromatids lead to the formation of chromosome dimers. In E. coli, it occurs in 15% of the cells at each cell generation under laboratory growth. If left unresolved, chromosome dimers stay trapped in closing septa. Daughter cells eventually separate, but it results in the shearing of the chromosomes and cell death. Chromosome dimers are resolved by the addition of a crossover by two dedicated tyrosine recombinases, XerC and XerD, which target a unique 28-bp sequence od the chromosome, dif. FtsK plays two roles in this process: (i) specific DNA motifs, the KOPS, dictate the orientation of FtsK translocation by orienting its loading on DNA. KOPS are skewed along the two replichores of the E. coli chromosome. They point away from oriC towards ter. As a result, FtsK pumps the origin-proximal regions of a chromosome dimer away from the division. dif is located within the terminus region, precisely in the zone of convergence of the KOPS. As a consequence, FtsK translocation brings the two dif sites carried by a chromosome dimer together at mid-cell. (ii) FtsKC then activates recombination at dif via a direct interaction with XerD.
Several criteria dictated our choice of V. cholerae as a new model organism for the study of the overarching principles driving chromosome organisation in bacteria:
(i) V. cholerae is close to E. coli in the phylogenetic tree of bacteria and its genome encodes for most of the replication (dam/SeqA), chromosome segregation (MukB, MatP) and cell division regulators (Min, Noc) of E. coli. In particular, we previously demonstrated that the V. cholerae chromosome dimer resolution system follows the E. coli paradigm.
(ii) However, the genome of V. cholerae is divided between two chromosomes, Chr1 and Chr2. Chr1 derives from the mono-chromosomal ancestor of the Vibrios. Chr2 results from the domestication of a horizontally acquired mega-plasmid. The preliminary information we had on the choreographies of segregation of the two V. cholerae chromosomes at the start of the DNA traffic project suggested that they differed between each other and from the E. coli paradigm.
(iii) Chr1 and Chr2 both harbour a partition system (ParAB/parS and HubP), like the unique chromosomes of B. subtilis or C. crescentus (ParAB and PopZ). Thus, the ancestor of E. coli and V. cholerae was at a crossroad in the evolution.
• Choreographies of segregation of the two V. cholerae chromosomes
We performed the first detailed characterization of the arrangement of the genetic material in a multipartite genome bacterium. We demonstrated that the two chromosomes of V. cholerae, chr1 and chr2, are longitudinally arranged. However, the smaller one only extended over the younger cell half.
Role of the ParAB/parS system of Chr1
• We found that disruption of the Par system of Chr1 released its origin from the pole but preserved its longitudinal arrangement and showed that the addition of an ectopic oriC perturbed this arrangement. These results suggest that the replication program directly contributes to chromosomal organisation.
• Role of the MatP/matS system
The ter of the unique chromosome of most bacteria locates at mid-cell at the time of cell division. In several species, this localization participates in the necessary coordination between chromosome segregation and cell division, notably for the selection of the division site, the licensing of the division machinery assembly and the correct alignment of chromosome dimer resolution sites. However, our fluorescent microscopy observations suggested that although the ter regions of Chr1 and Chr2 replicated at the same time, Chr2 sister termini separated before cell division whereas Chr1 sister termini were maintained together at mid-cell, which raised questions on the management of the two chromosomes during cell division. To confirm this possibility, we simultaneously visualized the location of the dimer resolution locus of each of the two chromosomes. We then studied the impact of the MatP/matS macrodomain organization system on Chr1 and Chr2 ter sister separation. We found that Chr1 terI loci remained in the vicinity of the cell centre in the absence of MatP and a genetic assay specifically designed to monitor the relative frequency of sister chromatid contacts during constriction suggested that they kept colliding together until the very end of cell division. In contrast, even though it was not able to impede the separation of Chr2 ter sisters before septation, the MatP/matS macrodomain organization system restricted their movement within the cell and permitted their frequent interaction during septum constriction. These results suggest that multiple redundant factors, including MatP in the enterobacteriaceae and the Vibrios, ensure that sister copies of the terminus region of bacterial chromosomes remain sufficiently close to mid-cell to be processed by FtsK.
We are currently finishing a study on the comparative roles of MatP in the two V. cholerae chromosomes and the E. coli chromosome. We expect to be able to submit an article at the end of 2017.
• Role of the MukB condensin
Following our work on the partition machineries and the MatP system, we have started to study the comparative roles of MukB in the cellular arrangement and segregation of the two V. cholerae chromosomes and the E. coli chromosome. Part of the results will be published in the incoming MatP article paper.
• Replication termination in V. cholerae
We have explored replication termination in Vibrio cholerae. An article is under preparation on the subject.
• Xer recombination activation mechanism
The Xer machinery belongs to the family of tyrosine recombinases. It is very similar to Cre, the resolvase of phage P1, whose action mechanism has been thoroughly characterised at the atomic resolution. However, several features of the Xer machinery differentiate it from Cre and most other tyrosine recombinases. First, it is by default inactive and requires a direct interaction with FtsK to add a crossover at dif. Secondly, it generally consists of two tyrosine recombinases, XerC and XerD, each of them catalysing the exchange of a specific pair of strands during recombination. By default, XerD recombinases are inactive and XerC recombinases promote the formation of a low level of HJs. However, they are rapidly eliminated by reverse reactions. This futile recombination cycle is broken during chromosome dimer resolution by the action of a cell division protein: FtsK. A direct interaction between its extreme C-terminal domain, FtsKγ, and XerD serves to trigger the exchange of a first pair of strands by XerD-catalysis between dif sites engaged in a recombination synapse. The resulting HJ is converted into product by XerC-catalysis. As the DNAtraffic project relied on the role played by FtsK in the activation of Xer recombination to monitor its activity, we needed to gain gained a thorough understanding of the Xer recombination process itself. To this end, we exploited the ability of Integrative Mobile Elements to exploit the Xer machinery of their host, the IMEXs. In particular, we studied two V. cholerae lysogenic phages, CTXφ and TLCφ. Our results on TLCφ were particularly interesting for the DNAtraffic project since we demonstrated that the integration and the excision of TLCφ followed the same pathway of recombination as dif-recombination (first strand exchange catalysed by XerD, second strand-exchange by XerC), but that it was independent of FtsK. We identified a Xer activation factor in the genome of TLCφ, XafT. XafT is a small cytoplasmic protein with no sequence or structural similarities to FtsK. It contains a domain of unknown function that is encoded in many other IMEX. We have been able to reconstitute a full Xer recombination reaction in vitro using purified XafT (unpublished results). We are now taking advantage of the simple limited component Xer reaction promoted by XafT to characterise in vitro the structural changes that occur in the recombination synapse upon activation.
• Coordination between chromosome segregation and cell division
During the course of our work on V. cholerae FtsK, we discovered that it located a cell pole in newborn cells and only relocated to mid-cell at a late stage of the cell cycle. This choreography was strikingly different form the choreography of E. coli cell division proteins. As part of DNAtraffic, it was therefore crucial to understand the processes that led to the late recruitment of FtsK at division site in V. cholerae. To this end, we exploited the fluorescence video-microscopy developed in the course of the project to analyse the choreography of the cell division proteins in V. cholerae.
The position and timing of assembly of the divisome is coordinated with chromosome segregation. In E. coli, this is achieved by the Min and Nucleoid Occlusion (NO) systems, which negatively regulate FtsZ polymerization. Of these, Min is considered the most significant. In its absence, Z-rings assemble at places other than mid-cell leading to division events near the cell poles and the formation of anucleate mini-cells24. NO is mainly effected by SlmA, a protein that binds to specific DNA motifs that are repeated on the E. coli genome, the SlmA Binding Sites (SBSs). The role of SlmA becomes apparent only when problems arise during DNA replication or segregation, or in the absence of Min.. The genome of V. cholerae encodes for putative homologues of the large majority of E. coli cell division proteins, including FtsZ, FtsA, ZapA, FtsK, FtsI, FtsN, Min and SlmA. We showed that SlmA is the main regulator of cell division in V. cholerae and the role of Min is only apparent when chromosome organization is altered. We further showed that the distribution of SBSs on Chr1 and Chr2 confined FtsZ and FtsK to the new cell pole of newborn cells and seemed to delay Z-ring assembly to a very late stage of the cell cycle, after most of Chr1 and Chr2 has been replicated and segregated. Finally, we investigated the dynamics of formation of the V. cholerae divisome using fluorescence microscopy, temperature sensitive mutants and a chemical inhibitor of FtsI. Our results demonstrated that FtsZ polymerisation presides over the recruitment of the other cell division components, which occurs in two distinct steps. They further suggested that early pre-divisional Z-rings form between 40 and 50% of the cell cycle. Pre-divisional Z-rings evolve into mature divisome at about 80% of the cell cycle when late cell division proteins such as FtsK are recruited.
• DNA metabolic pathways leading to a requirement for FtsK
Microscopic observations of the cellular arrangement of pairs of chromosome loci under slow growth conditions recently suggested that FtsK translocation served to release the MatP-mediated cohesion and/or cell division apparatus-interaction of ter sisters in a KOPS-oriented manner, placing it at the centre of the coordination between the E. coli replication/segregation and cell division cycles.
Prior to this observation, FtsK translocation was only considered as a safeguard against the formation of chromosome dimers. Chromosome dimers are generated by homologous recombination events between chromatid sisters during or after replication. They physically impede the segregation of genetic information at cell division, which generates a substrate for FtsK translocation. They are resolved by the addition of a crossover at dif by a dedicated pair of chromosomally encoded tyrosine recombinases, XerC and XerD. Xer recombination-deficient Growth competition and dif-cassette excision experiments in E. coli indicated that dif only functioned within the ter region, at the zone of convergence of the KOPS motifs. Finally, density label and dif-cassette excision assays showed that recombination at dif took place at a late stage of cell division, after the initiation of septum constriction. The roles played by FtsK in chromosome dimer resolution explains the spatial restriction of the activity of dif on the chromosome to the KOPS convergence zone while the temporal restriction of Xer recombination at chromosomal dif sites suggests that the action of FtsKC is delayed compared to its recruitment at the septum. Correspondingly, ectopic production of FtsKC was sufficient to activate dif recombination outside of the KOPS convergence zone, independently of cell division and of recA. Taken together, these results suggested that FtsK normally only acted on chromosome dimers. The mild phenotype of FtsK translocation deficient mutants and their suppression by the inactivation of recA corroborated this hypothesis, in apparent contradiction with the general role of FtsK in the release of MatP-mediated cohesion and/or cell division apparatus-interaction of ter sisters.
We showed that in V. cholerae, as in E, coli, dif-cassette excision was restricted to the KOPS convergence zone within chr1 and chr2 ter regions and only took place after the initiation of septum constriction. However, dif-cassette excision on both V. cholerae chromosomes was independent of recA. In addition, it was too high to be solely explained by the estimated rate of chromosome dimer formation at each generation. Using a combination of careful dif-cassette excision assays and newly available fluorescence microscopy techniques, we demonstrated that the E. coli recA-dependency and V. cholerae recA-independency of dif-cassette excision are not determined by differences in the dimer resolution machineries of the two bacteria but by differences in the timing of segregation of their chromosomes: whatever the growth conditions, V. cholerae chr1 ter sister copies remain together at mid-cell until the onset of constriction, which increases the chances for FtsK to activate recombination at dif independently of recA. Likewise, we showed that in slow growth conditions, E. coli ter sister copies separate after the onset of constriction and dif-recombination is independent of recA. In contrast, our results suggested that MatP did not prevent E. coli ter sister copies from separating away from each other and from mid-cell before constriction in fast growth conditions. Finally, we showed that separation of ter sisters was independent of FtsK, which explained why recombination at dif becomes dependent on the formation of chromosome dimers by homologous recombination.
Following this work, we have collaborated with the teams Bénédicte Michel within the institute and D. Leach (UK) to explore DNA traffic problems resulting from the inactivation of the RecBCD recombination repair enzyme of E. coli cells. The results have been submitted for publication in PLoS Genetics.
Bacterial chromosomes are generally covalently closed circular DNA molecules. They carry a single origin of bidirectional replication, oriC, which defines two replication arms. Replication terminates in a region opposite of oriC on the circular chromosome map, the terminus (ter). Two apparently opposed cellular arrangements of chromosomes have been observed in model rod-shaped bacteria. Indeed, the unique chromosome of E. coli adopts a transversal organisation, with oriC being positioned at mid-cell and flanked by the replication arms in newborn cells. In contrast, the unique chromosome of C. crescentus adopts a longitudinal organisation, with oriC and ter being positioned at opposite poles and the replication arms extending side by side along the long axis of the cell. Correspondingly, the mechanisms driving the spatial arrangement and the segregation of bacterial chromosomes are not conserved. In addition, they are often redundant. As a result, the overarching principles driving chromosome organisation in bacteria are still debated. To tackle this question, we decided to focus on FtsK-driven segregation because it seems to be a general feature of the bacterial cell cycle and take an evolutionary approach by comparing its action in E. coli and Vibrio cholerae, two closely related species in the phylogenetic tree of bacteria that nevertheless displayed contrasting cellular organisations and cell cycles.
FtsK can be divided in two domains. The N-terminal part of the protein, FtsKN, is an integral membrane domain. In E. coli, it is recruited to the septum soon after the formation of the tubulin-like FtsZ ring and participates in the recruitment of the peptidoglycan synthesis and degradation machinery. Thus, FtsK is an essential component of the cell division machinery. The carboxy-terminal domain of FtsK, FtsKC, assembles into hexamers over double stranded DNA and has been shown to function as a fast DNA translocase. As FtsKN anchors FtsK in the division septum, it results in the mobilization of chromosomal DNA. Thus, FtsK serves as a last safeguard to avoid division over partially segregated chromosomes. This probably explains why DNA translocation by FtsK becomes essential when the activity of the major cellular decatenase, TopoIV, is compromised, when the organization and/or the packaging of the chromosome are altered by the deletion of the condensin-like mukBEF system or by large chromosomal inversions, and when chromosome replication is affected. However, most of our knowledge on E. coli FtsK came from the study of chromosome dimer resolutio. Ondd numbers of crossovers due to homologous recombination between replicating circular sister chromatids lead to the formation of chromosome dimers. In E. coli, it occurs in 15% of the cells at each cell generation under laboratory growth. If left unresolved, chromosome dimers stay trapped in closing septa. Daughter cells eventually separate, but it results in the shearing of the chromosomes and cell death. Chromosome dimers are resolved by the addition of a crossover by two dedicated tyrosine recombinases, XerC and XerD, which target a unique 28-bp sequence od the chromosome, dif. FtsK plays two roles in this process: (i) specific DNA motifs, the KOPS, dictate the orientation of FtsK translocation by orienting its loading on DNA. KOPS are skewed along the two replichores of the E. coli chromosome. They point away from oriC towards ter. As a result, FtsK pumps the origin-proximal regions of a chromosome dimer away from the division. dif is located within the terminus region, precisely in the zone of convergence of the KOPS. As a consequence, FtsK translocation brings the two dif sites carried by a chromosome dimer together at mid-cell. (ii) FtsKC then activates recombination at dif via a direct interaction with XerD.
Several criteria dictated our choice of V. cholerae as a new model organism for the study of the overarching principles driving chromosome organisation in bacteria:
(i) V. cholerae is close to E. coli in the phylogenetic tree of bacteria and its genome encodes for most of the replication (dam/SeqA), chromosome segregation (MukB, MatP) and cell division regulators (Min, Noc) of E. coli. In particular, we previously demonstrated that the V. cholerae chromosome dimer resolution system follows the E. coli paradigm.
(ii) However, the genome of V. cholerae is divided between two chromosomes, Chr1 and Chr2. Chr1 derives from the mono-chromosomal ancestor of the Vibrios. Chr2 results from the domestication of a horizontally acquired mega-plasmid. The preliminary information we had on the choreographies of segregation of the two V. cholerae chromosomes at the start of the DNA traffic project suggested that they differed between each other and from the E. coli paradigm.
(iii) Chr1 and Chr2 both harbour a partition system (ParAB/parS and HubP), like the unique chromosomes of B. subtilis or C. crescentus (ParAB and PopZ). Thus, the ancestor of E. coli and V. cholerae was at a crossroad in the evolution.
• Choreographies of segregation of the two V. cholerae chromosomes
We performed the first detailed characterization of the arrangement of the genetic material in a multipartite genome bacterium. We demonstrated that the two chromosomes of V. cholerae, chr1 and chr2, are longitudinally arranged. However, the smaller one only extended over the younger cell half.
Role of the ParAB/parS system of Chr1
• We found that disruption of the Par system of Chr1 released its origin from the pole but preserved its longitudinal arrangement and showed that the addition of an ectopic oriC perturbed this arrangement. These results suggest that the replication program directly contributes to chromosomal organisation.
• Role of the MatP/matS system
The ter of the unique chromosome of most bacteria locates at mid-cell at the time of cell division. In several species, this localization participates in the necessary coordination between chromosome segregation and cell division, notably for the selection of the division site, the licensing of the division machinery assembly and the correct alignment of chromosome dimer resolution sites. However, our fluorescent microscopy observations suggested that although the ter regions of Chr1 and Chr2 replicated at the same time, Chr2 sister termini separated before cell division whereas Chr1 sister termini were maintained together at mid-cell, which raised questions on the management of the two chromosomes during cell division. To confirm this possibility, we simultaneously visualized the location of the dimer resolution locus of each of the two chromosomes. We then studied the impact of the MatP/matS macrodomain organization system on Chr1 and Chr2 ter sister separation. We found that Chr1 terI loci remained in the vicinity of the cell centre in the absence of MatP and a genetic assay specifically designed to monitor the relative frequency of sister chromatid contacts during constriction suggested that they kept colliding together until the very end of cell division. In contrast, even though it was not able to impede the separation of Chr2 ter sisters before septation, the MatP/matS macrodomain organization system restricted their movement within the cell and permitted their frequent interaction during septum constriction. These results suggest that multiple redundant factors, including MatP in the enterobacteriaceae and the Vibrios, ensure that sister copies of the terminus region of bacterial chromosomes remain sufficiently close to mid-cell to be processed by FtsK.
We are currently finishing a study on the comparative roles of MatP in the two V. cholerae chromosomes and the E. coli chromosome. We expect to be able to submit an article at the end of 2017.
• Role of the MukB condensin
Following our work on the partition machineries and the MatP system, we have started to study the comparative roles of MukB in the cellular arrangement and segregation of the two V. cholerae chromosomes and the E. coli chromosome. Part of the results will be published in the incoming MatP article paper.
• Replication termination in V. cholerae
We have explored replication termination in Vibrio cholerae. An article is under preparation on the subject.
• Xer recombination activation mechanism
The Xer machinery belongs to the family of tyrosine recombinases. It is very similar to Cre, the resolvase of phage P1, whose action mechanism has been thoroughly characterised at the atomic resolution. However, several features of the Xer machinery differentiate it from Cre and most other tyrosine recombinases. First, it is by default inactive and requires a direct interaction with FtsK to add a crossover at dif. Secondly, it generally consists of two tyrosine recombinases, XerC and XerD, each of them catalysing the exchange of a specific pair of strands during recombination. By default, XerD recombinases are inactive and XerC recombinases promote the formation of a low level of HJs. However, they are rapidly eliminated by reverse reactions. This futile recombination cycle is broken during chromosome dimer resolution by the action of a cell division protein: FtsK. A direct interaction between its extreme C-terminal domain, FtsKγ, and XerD serves to trigger the exchange of a first pair of strands by XerD-catalysis between dif sites engaged in a recombination synapse. The resulting HJ is converted into product by XerC-catalysis. As the DNAtraffic project relied on the role played by FtsK in the activation of Xer recombination to monitor its activity, we needed to gain gained a thorough understanding of the Xer recombination process itself. To this end, we exploited the ability of Integrative Mobile Elements to exploit the Xer machinery of their host, the IMEXs. In particular, we studied two V. cholerae lysogenic phages, CTXφ and TLCφ. Our results on TLCφ were particularly interesting for the DNAtraffic project since we demonstrated that the integration and the excision of TLCφ followed the same pathway of recombination as dif-recombination (first strand exchange catalysed by XerD, second strand-exchange by XerC), but that it was independent of FtsK. We identified a Xer activation factor in the genome of TLCφ, XafT. XafT is a small cytoplasmic protein with no sequence or structural similarities to FtsK. It contains a domain of unknown function that is encoded in many other IMEX. We have been able to reconstitute a full Xer recombination reaction in vitro using purified XafT (unpublished results). We are now taking advantage of the simple limited component Xer reaction promoted by XafT to characterise in vitro the structural changes that occur in the recombination synapse upon activation.
• Coordination between chromosome segregation and cell division
During the course of our work on V. cholerae FtsK, we discovered that it located a cell pole in newborn cells and only relocated to mid-cell at a late stage of the cell cycle. This choreography was strikingly different form the choreography of E. coli cell division proteins. As part of DNAtraffic, it was therefore crucial to understand the processes that led to the late recruitment of FtsK at division site in V. cholerae. To this end, we exploited the fluorescence video-microscopy developed in the course of the project to analyse the choreography of the cell division proteins in V. cholerae.
The position and timing of assembly of the divisome is coordinated with chromosome segregation. In E. coli, this is achieved by the Min and Nucleoid Occlusion (NO) systems, which negatively regulate FtsZ polymerization. Of these, Min is considered the most significant. In its absence, Z-rings assemble at places other than mid-cell leading to division events near the cell poles and the formation of anucleate mini-cells24. NO is mainly effected by SlmA, a protein that binds to specific DNA motifs that are repeated on the E. coli genome, the SlmA Binding Sites (SBSs). The role of SlmA becomes apparent only when problems arise during DNA replication or segregation, or in the absence of Min.. The genome of V. cholerae encodes for putative homologues of the large majority of E. coli cell division proteins, including FtsZ, FtsA, ZapA, FtsK, FtsI, FtsN, Min and SlmA. We showed that SlmA is the main regulator of cell division in V. cholerae and the role of Min is only apparent when chromosome organization is altered. We further showed that the distribution of SBSs on Chr1 and Chr2 confined FtsZ and FtsK to the new cell pole of newborn cells and seemed to delay Z-ring assembly to a very late stage of the cell cycle, after most of Chr1 and Chr2 has been replicated and segregated. Finally, we investigated the dynamics of formation of the V. cholerae divisome using fluorescence microscopy, temperature sensitive mutants and a chemical inhibitor of FtsI. Our results demonstrated that FtsZ polymerisation presides over the recruitment of the other cell division components, which occurs in two distinct steps. They further suggested that early pre-divisional Z-rings form between 40 and 50% of the cell cycle. Pre-divisional Z-rings evolve into mature divisome at about 80% of the cell cycle when late cell division proteins such as FtsK are recruited.
• DNA metabolic pathways leading to a requirement for FtsK
Microscopic observations of the cellular arrangement of pairs of chromosome loci under slow growth conditions recently suggested that FtsK translocation served to release the MatP-mediated cohesion and/or cell division apparatus-interaction of ter sisters in a KOPS-oriented manner, placing it at the centre of the coordination between the E. coli replication/segregation and cell division cycles.
Prior to this observation, FtsK translocation was only considered as a safeguard against the formation of chromosome dimers. Chromosome dimers are generated by homologous recombination events between chromatid sisters during or after replication. They physically impede the segregation of genetic information at cell division, which generates a substrate for FtsK translocation. They are resolved by the addition of a crossover at dif by a dedicated pair of chromosomally encoded tyrosine recombinases, XerC and XerD. Xer recombination-deficient Growth competition and dif-cassette excision experiments in E. coli indicated that dif only functioned within the ter region, at the zone of convergence of the KOPS motifs. Finally, density label and dif-cassette excision assays showed that recombination at dif took place at a late stage of cell division, after the initiation of septum constriction. The roles played by FtsK in chromosome dimer resolution explains the spatial restriction of the activity of dif on the chromosome to the KOPS convergence zone while the temporal restriction of Xer recombination at chromosomal dif sites suggests that the action of FtsKC is delayed compared to its recruitment at the septum. Correspondingly, ectopic production of FtsKC was sufficient to activate dif recombination outside of the KOPS convergence zone, independently of cell division and of recA. Taken together, these results suggested that FtsK normally only acted on chromosome dimers. The mild phenotype of FtsK translocation deficient mutants and their suppression by the inactivation of recA corroborated this hypothesis, in apparent contradiction with the general role of FtsK in the release of MatP-mediated cohesion and/or cell division apparatus-interaction of ter sisters.
We showed that in V. cholerae, as in E, coli, dif-cassette excision was restricted to the KOPS convergence zone within chr1 and chr2 ter regions and only took place after the initiation of septum constriction. However, dif-cassette excision on both V. cholerae chromosomes was independent of recA. In addition, it was too high to be solely explained by the estimated rate of chromosome dimer formation at each generation. Using a combination of careful dif-cassette excision assays and newly available fluorescence microscopy techniques, we demonstrated that the E. coli recA-dependency and V. cholerae recA-independency of dif-cassette excision are not determined by differences in the dimer resolution machineries of the two bacteria but by differences in the timing of segregation of their chromosomes: whatever the growth conditions, V. cholerae chr1 ter sister copies remain together at mid-cell until the onset of constriction, which increases the chances for FtsK to activate recombination at dif independently of recA. Likewise, we showed that in slow growth conditions, E. coli ter sister copies separate after the onset of constriction and dif-recombination is independent of recA. In contrast, our results suggested that MatP did not prevent E. coli ter sister copies from separating away from each other and from mid-cell before constriction in fast growth conditions. Finally, we showed that separation of ter sisters was independent of FtsK, which explained why recombination at dif becomes dependent on the formation of chromosome dimers by homologous recombination.
Following this work, we have collaborated with the teams Bénédicte Michel within the institute and D. Leach (UK) to explore DNA traffic problems resulting from the inactivation of the RecBCD recombination repair enzyme of E. coli cells. The results have been submitted for publication in PLoS Genetics.