Chromosome Dynamics in Bacteria: Triggering Replication at the Opposite Location and Segregation in the Opposite Direction

Bacteria can accomplish surprising levels of organization in the absence of membrane organelles by constructing subcellular asymmetric protein gradients. These gradients are composed of regulators that can either trigger or inhibit cell cycle events from distinct cell poles. In Caulobacter crescentus, the onset of chromosome replication and segregation from the stalked pole are regulated by asymmetric protein gradients. We show that the activators of chromosome replication and segregation are not restricted to the stalked pole and that their organization and directionality can be flipped in orientation. Our results also indicate that the subcellular location of key chromosomal loci play important roles in the establishment of the asymmetric organization of cell cycle regulators.

T riggering the onset of chromosome replication under the right conditions is central to cell survival and proliferation. In bacteria, DnaA is the highly conserved initiator of chromosome replication that opens the origin of replication (ori) by forming a helical right-handed polymeric structure (1)(2)(3)(4)(5). Once DnaA opens the double-stranded chromosome, the replication machinery assembles at ori and initiates chromosome replication bidirectionally. The subcellular location where DnaA initiates chromosome replication is established by the position of ori inside the cell. Depending on the bacterial species, the subcellular location of ori varies significantly. For instance, ori in Caulobac-ter crescentus (referred to hereafter as Caulobacter) is found at one pole (the stalked pole), whereas in Escherichia coli and Bacillus subtilis, ori is found near mid-cell. Within a single species, the subcellular location of ori is strictly retained at the same position in nondividing cells and reestablished soon after chromosome replication and segregation initiate in actively dividing cells.
In nondividing Caulobacter, ori is retained near the stalked pole by the interaction between the anchoring protein PopZ and the ParB partition protein bound to the centromere-like chromosomal parS (referred to hereafter also as the centromere) (6)(7)(8).
DnaA initiates replication at ori near the stalked pole (Fig. 1). Once chromosome replication initiates at ori, the replication fork must pass through the centromere chromosomal region (parS, 8 kb away from ori) for chromosome segregation to initiate (9). One of the two newly replicated ParB-coated centromeres is segregated to the opposite pole (also referred to as the new pole) by direct interactions between ParB and the ATPase ParA (9)(10)(11)(12)(13)(14). It has been proposed that ParA forms a stable gradient with concentrations gradually decreasing from the new pole to the stalked pole, which are critical for establishing segregation directionality (10)(11)(12)15).
In this report, we asked whether the onset of replication and segregation are restricted to the intrinsic localization of ori and the centromere, which in Caulobacter are near the stalked pole. We genetically engineered a Caulobacter strain where movement of ori and the parS centromere can be triggered in the absence of replication initiation (16). Once these chromosomal foci were translocated to the opposite cell pole, we tested for competency of replication initiation and chromosome segregation. Our data demonstrate that DnaA and ParA are able to initiate chromosome replication and segregation irrespective of the subcellular localization of ori and the centromere. The organization of the ParA gradient flips in orientation once the unreplicated centromere is relocalized to the new pole, revealing the robustness and flexibility of the orientation of chromosome replication and segregation.

RESULTS
Subphysiological levels of DnaA result in translocation of ori away from the stalked pole. In Caulobacter, the centromere is the first chromosomal locus to segregate away from the stalked pole (9). Previous analyses of cells expressing subphysiological levels of DnaA (insufficient to initiate replication) revealed a DnaA-dependent and replication-independent segregation of the centromere (16). Subphysiological levels of DnaA cause the unreplicated centromere to move in a ParA-dependent manner from the stalked pole to the new pole (16). We asked whether subphysiological levels of DnaA could also trigger the movement of ori independently of replication. We tracked the localization of ori by constructing a strain with a fluorescent tag to be inserted near ori using the Yersinia pestis parS(pMT1) chromosomal sequence and its corresponding gene encoding ParB(pMT1) (17). This strain is called PM500 [xylX::cfp-parB(pMT1), parS(pMT1) at nucleotide 1108, dnaA::⍀, vanA::dnaA; details on strain construction found in Materials and Methods]. Using growth curves and CFU, we showed that this strain, PM500 with the parS(pMT1) insertion near ori and expression of cfp-parB(pMT1), exhibits no significant effect on doubling time and viability compared to the wild type (see Fig. S1 in the supplemental material). To reach subphysiological levels of DnaA in the cell, we used the same vanillate promoter that was used to regulate dnaA expression when tracking centromere movement (16). Our results show that subphysiological levels of DnaA trigger the translocation of ori away from the stalked pole independently of chromosome replication ( Fig. 2A; see also Movie S1 in the supplemental material). As the time of DnaA depletion increased, cells displayed unreplicated ori foci progressively moving toward the new pole (Fig. 2B). Within 3 h of DnaA depletion, ϳ75% of cells exhibited ori localized at the new pole. These results resemble the replication-independent movement of the centromere, suggesting that ori and the centromere translocate in a DnaA-dependent manner likely due to their close proximity on the chromosome (16).
DnaA's ability to initiate replication is not restricted to the stalked pole. In Caulobacter, DnaA initiates replication only once per cell cycle and only in stalked cells, which have their ori localized at the stalked pole (Fig. 1). To determine whether DnaA can initiate replication at the opposite cell pole, we analyzed cells that had undergone replication-independent translocation of ori. To track replication initiation, we induced dnaA from a vanillate promoter in cells with ori localized to the opposite cell pole and followed the appearance of two newly replicated ori foci ( Fig. 3A; Movie S2). We quantified the frequencies of replication initiation based on the initial location of ori prior to induction of dnaA expression (Fig. 3B). Within 30 min of vanillate supplementation, ϳ63% of cells with ori localized at the new pole had initiated chromosome  replication as evidenced by two clearly separated ori foci. Notably, chromosome replication initiated slower in the subpopulation of cells that retained ori at the stalked pole: ϳ19% of cells displayed two ori foci after 30 min of vanillate supplementation. Our data revealed that DnaA's activity as a replication initiator is not restricted to the stalked pole.
Centromeres are effectively segregated in the opposite direction. On the basis of our results that DnaA is able to initiate replication from the new pole (Fig. 3), we then asked about chromosome segregation. Can the partitioning system ParABS initiate segregation of the centromere from the new pole toward the stalked pole, which in this case would be in the opposite direction? To test this, we used a Caulobacter strain in which the native parB gene was replaced with the fusion gene encoding cyan fluorescent protein (CFP)-ParB and in which the only copy of dnaA was regulated under the vanillate promoter (PM109) (16). In Caulobacter, the partitioning protein ParB binds directly to the centromere (9,18). Thus, we can track centromere movement by using cells expressing a functional fusion protein CFP-ParB. When the vanillate inducer is removed from the growth medium of strain PM109, cells are exposed to subphysiological levels of DnaA insufficient to initiate chromosome replication but sufficient to trigger movement of the unreplicated centromere (16). Fluorescent imaging of PM109 cells depleted of DnaA (incubated without vanillate for 3 h) revealed the localization of a single CFP-ParB focus with the following distribution: ϳ37% at or near the new pole, ϳ54% at or near the stalked pole, and ϳ10% at/around mid-cell (Fig. S2), consistent with previous analyses (16). One potential explanation for the difference in frequencies of replication-independent parS-ori translocation between strains PM109 and PM500 (ϳ37% versus ϳ75%) may have to do with expression of the Y. pestis parB(pMT1). Although we did not observe defects on doubling times and/or viability from expression of parB(pMT1) in Caulobacter, our translocation frequency data suggest that Y. pestis ParB may influence the activity of Caulobacter's ParA in chromosome segregation. From here on, we use only Caulobacter strains with its native parB fluorescently tagged.
To determine the ability of ParABS to trigger segregation of the centromere in the opposite direction (from the new pole toward the stalked pole), we tracked the movement of CFP-ParB in PM109 cells with translocated unreplicated centromeres subsequent to the addition of vanillate (Fig. 4). Two clearly separated centromeres were observed soon after dnaA expression was reestablished, irrespective of the initial localization of centromere ( Fig. 4A; Movie S3). Upon induction of dnaA expression, we observed a similar pattern in the rates of replication initiation when tracking the number of parS centromeres. The appearance of two CFP-ParB foci occurred slightly sooner in cells with a centromere at the new pole compared to cells with a centromere at the stalked pole (Fig. S3). Notably, cells with two CFP-ParB foci were able to segregate their centromeres to the cell poles, irrespective of the initial localization of centromere prior to dnaA induction (Fig. 4A). Quantification of these data revealed that the rate at which centromeres were segregated to the cell poles were significantly faster in cells with centromeres that departed from the mid-cell (Fig. 4B). These results can potentially be explained based on the shorter distance that the replicated centromeres had to travel from mid-cell to reach the cell poles. The rates of centromere segregation that initiated from either the stalked pole or the new pole were not statistically different. These data suggest that the partitioning protein ParA was able to quickly rearrange its gradient in order to segregate centromeres from mid-cell or from the new pole with no significant delays.
Active ParA is required for centromere segregation in the opposite direction. Caulobacter cells expressing ParA variants unable to hydrolyze ATP cannot segregate their centromere to the cell poles (9,10,14). To determine whether centromere segregation observed from the new pole or mid-cell is ParA dependent, we tracked CFP-ParB localization in a merodiploid strain that carries the wild-type allele of parA at the native locus and a dominant-negative mutant parA allele unable to hydrolyze ATP expressed from a xylose-inducible promoter. This dominant-negative allele contains a missense mutation in the ATPase domain (ParA D44A ) of ParA that inhibits chromosome segregation (19). To test segregation in the opposite direction, we first allowed cells to translocate their unreplicated centromeres by growing them in growth media devoid of vanillate. The growth media were then supplemented with xylose and vanillate so that replication initiation was induced in the presence of the dominant-negative ParA D44A . Our data revealed that Ͼ80% of cells expressing wild-type ParA were able to segregate the centromeres to the poles, as evidenced by one CFP-ParB focus at each pole (Fig. 5A). However, cells expressing the dominant-negative ParA D44A after replication initiated at the opposite cell pole failed to segregate their replicated centromeres as evidenced by two CFP-ParB near each other ( Fig. 5A and B). These data strongly suggest that the segregation of the centromere in the opposite direction, from the new pole to the stalked pole, requires an active chromosome segregation machinery.
Relocalization of the centromere locus triggers rearrangement of the ParA gradient. In Caulobacter, ParA forms a visible gradient with concentrations gradually decreasing from the new pole to the stalked pole (10-12, 20) (Fig. 1). Our observation that the centromere could be segregated in the opposite direction suggested that cells with centromeres at the new pole rearrange the gradient of ParA. To determine whether ParA could change the orientation of its gradient, we assessed the localization patterns of ParA using the background of a parA merodiploid strain that contained the wild-type allele of parA at the native locus and a fluorescently tagged parA (ParA-mCherry) under the inducible promoter for xylose (19). We found that the simultaneous overexpression of ParA (native ParA plus ParA-mCherry) and DnaA depletion resulted in chromosome replication initiation in approximately 45% of cells, suggestive of the coregulation of DnaA and ParA observed in B. subtilis (21). However, all cells that retained their unreplicated centromere at the stalked pole displayed the ParA-mCherry gradient that resembles the gradient of wild-type cells (Fig. 6). Notably, cells with unreplicated centromeres that had translocated to the new pole displayed a flipped pattern of ParA-mCherry (high levels at the stalked pole and low levels at the new pole). When the centromere was localized at mid-cell, ParA-mCherry displayed what appears to be two separate gradients with high levels starting from both cell poles (Fig. 6). Thus, our data suggest that the subcellular localization of parS-ori plays a role in dictating the organization of ParA.
PopZ subcellular localization patterns based on parS localization. The swarmer exhibits a single PopZ focus at the pole bearing the flagellum. As the swarmer cell differentiates into a stalked cell and initiates DNA replication, a second PopZ focus is established at the opposite pole (new pole) (6)(7)(8). The directionality of ParA's function in centromere segregation from the stalked pole to the new pole has been proposed to be influenced by the localization of the anchoring protein PopZ (19,22). To determine whether the onset of chromosome replication/segregation from the new cell pole altered PopZ localization dynamics, we tracked the localization of PopZ by using cells expressing a functional fusion protein mCherry-PopZ. In our control experiment with cells grown in the presence of vanillate (the dnaA inducer), mCherry-PopZ exhibited foci at each pole upon the onset of chromosome replication and segregation (6)(7)(8)16). In cells with translocated unreplicated centromeres, mCherry-PopZ localization was dependent on the localization of CFP-ParB bound to the parS centromere (Fig. 7A). About 90% of cells with CFP-ParB at the stalked pole displayed a single mCherry-PopZ focus also localized at the stalked pole. In cells with CFP-ParB localized at the new pole, ϳ90% displayed PopZ-CFP foci located at each pole. Notably, cells with CFP-ParB localized at mid-cell displayed an equal combination of cells with either one mCherry-PopZ focus localized at the stalked pole or mCherry-PopZ foci localized at each pole. Upon the induction of chromosome replication by the addition of the inducer vanillate, cells with bipolar localization of PopZ remained bipolar (Fig. 7B). Regardless of where the centromere was localized when replication initiation was induced by vanillate supplementation, ϳ100% of cells displayed bipolar localization of mCherry-PopZ subsequent to the onset of chromosome replication. Our data suggest that the initiation of chromosome replication and segregation influence the bipolar localization of PopZ.
Effects in viability from initiating chromosome replication/segregation from outside the stalked pole. To determine whether initiating chromosome replication and/or segregation from outside the stalked pole altered the viability of Caulobacter, we analyzed CFU of cells that had undergone ori or centromere translocation away from the stalked pole in the absence of replication (Fig. 8). Cells of strains PM500 Chromosome Replication and Segregation in Caulobacter ® (fluorescent label near ori) and PM109 (fluorescently labeled ParB-centromere) were spotted immediately after DnaA depletion (1 h to 3 h) on plates containing the inducer for dnaA expression (supplemented with vanillate). Our data revealed no significant differences between cells that initiated chromosome replication and segregation from the new pole (or mid-cell) compared to wild-type conditions. Our data suggest that cells can recover relatively quickly after initiating chromosome replication/segregation from outside the stalked pole.

DISCUSSION
By inducing the movement of ori and centromere away from their intrinsic subcellular locations in a replication-independent manner, we have shown that the molecular machinery involved in chromosome replication and segregation are remarkably flexible with respect to their subcellular orientation. We found that the activity of DnaA and ParA are not restricted to a single polar microdomain and can successfully induce chromosome replication and segregation from the opposite polar microdomain composed of distinct client proteins. These results suggest that the chromosomal loci ori and parS play key roles in the localization of cell cycle regulators. Cells that initiate chromosome replication and segregation from outside the stalked pole displayed no detectable viability defects compared to wild-type conditions. Therefore, our data reveal the ability of the cell to rapidly reorganize chromosome orientation along with the set of proteins involved in chromosomal replication and segregation.
DnaA's activity as replication initiator and cell asymmetry. In the absence of membrane-bound organelles, bacteria rely on proteins organized in gradients to establish cellular polarity and perform asymmetric functions. One example of such organization is the phosphorylated regulator CtrA (CtrAϳP), which binds ori and inhibits DnaA from initiating replication at one cell pole (23)(24)(25). In predivisional cells, a phospho-signaling relay at the cell poles has been proposed to generate an asymmetric concentration gradient of CtrAϳP with the highest levels at the new pole that gradually decrease toward the stalked pole (26). Alterations to this proposed asymmetric concentration gradient of CtrAϳP eliminate the asymmetric regulation of DnaA, resulting in cells initiating replication from the new pole (26). Our data revealed that DnaA can also trigger replication initiation from the new pole in cells with altered subcellular location of ori (Fig. 3). Our results can be explained by the undetected levels of CtrA in cells depleted of DnaA (27,28). This is because DnaA is a transcriptional regulator of gcrA, and GcrA is a transcriptional regulator of ctrA (27)(28)(29). Consequently, the expression of ctrA is indirectly dependent on the levels of DnaA. Thus, cells with ori localized at the new pole with depleted levels of DnaA are likely to have no CtrAϳP or a minimal CtrAϳP gradient that is insufficient to inhibit replication initiation from either pole.
Regulation of periodicity of DnaA activity. Most of the regulators of DnaA activity that have been identified thus far are negative regulators that prevent the overinitiation of chromosome replication. However, positive regulators that trigger DnaA to initiate replication with such efficient periodicity remain limited. This periodicity of DnaA activity is maintained even in E. coli cells that were artificially designed to have two spatially separated ori foci. In those cells, DnaA productively initiated replication synchronously from both ori foci (30). In E. coli and Helicobacter pylori, a recruiter of DnaA has been characterized that promotes the assembly of polymeric DnaA at ori (31)(32)(33). Notably, constitutive expression of dnaA in Caulobacter has been shown to Chromosome Replication and Segregation in Caulobacter ® have no effect on the periodicity of DnaA activity, suggesting that dnaA transcriptional regulation is not the principal modulator of DnaA periodicity (34). In Caulobacter, CtrA regulates the spatial activity of DnaA so that chromosome replication initiates only in the stalked cells (24,25,34). However, CtrA is not involved in the periodicity of DnaA's activity (34). Thus, the molecular mechanism that triggers DnaA to initiate chromosome replication with such precise periodicity remains unclear. A hypothetical scenario is that some type of regulator that facilitates this periodicity process is found within the microenvironment at ori's location at the time of replication initiation. Our data suggest that there is no regulator/modulator of DnaA that is fixed at the stalked pole's microdomain. We cannot however exclude the possibility that a potential modulator does exist and that this modulator could migrate along with ori because either it binds directly to ori or it is recruited by DnaA. Biochemical characterization of DnaA's activity at ori is required to identify the mechanism that DnaA uses to regulate its temporal activity with such remarkable accuracy.
Localization of the centromere dictates the orientation of ParA's activity. The partitioning protein ParA is another example of a bacterial protein organized in a gradient to establish cellular polarity. In Caulobacter, ParA forms a gradient with concentrations gradually decreasing from the new pole to the stalked pole (10)(11)(12)20). Interestingly, this stable gradient of ParA is established well before chromosome replication and segregation are initiated (10)(11)(12). Thus, the question remains as to what activates this asymmetric organization of ParA. Our data suggest that the organization of the ParA gradient can be reconstructed in the opposite orientation by rearranging the location of the parS-ParB complex (Fig. 6). Cells with parS-ParB at the new pole displayed higher levels of ParA at the stalked pole than at the new pole.
We propose that the localization of the centromere complex directs the arrangement of the ParA gradient. This model is consistent with how ParA partitioning systems segregate low-copy-number plasmids to maintain inheritance. Using in vitro reconstitution assays, the ParB-parS complexes of these plasmids were shown to chase and rearrange the ParA gradient (35)(36)(37). For chromosome segregation, the specific component of the centromere complex that triggers the organization of ParA remains to be determined. There are several proteins that bind parS and/or interact with parS-ParB that could serve as potential regulators of ParA's subcellular organization. One possibility is that ParB itself triggers ParA's organization by inducing ParA to hydrolyze ATP (9,10,14), consistent with what has been observed with partitioning of plasmids (35)(36)(37). Another possibility is MipZ, the inhibitor of FtsZ polymerization, which colocalizes with the parS-ParB complex (38). The replication initiator DnaA is another candidate because it also binds parS and has been proposed to be involved in the onset of centromere segregation (16). Last, the anchoring protein PopZ, which interacts with ParB, has been proposed to be involved in ParA activity and directionality (6-8, 19, 22). Our data suggest that the bipolar localization of PopZ is elicited primarily by centromere segregation, as suggested previously (39). About 90% of cells displayed PopZ bipolarly localized in cells with an unreplicated centromere translocated to the new pole (Fig. 7). However, only upon the onset of chromosome replication did we observe 100% of cells with bipolarly localized PopZ, suggesting that replication initiation may also influence PopZ localization. It remains to be determined whether PopZ plays a role in the organization of ParA.
We have demonstrated that ParA can successfully segregate the parS centromere from the new pole to the stalked pole, which is the reverse direction to that observed in wild-type cells. We propose that once a stable gradient of ParA is formed in cells with translocated unreplicated centromeres, the ParA-DNA interaction relay previously shown to provide the force necessary for centromere segregation (20) can initiate and segregate one centromere in the reverse direction.
Robustness of cells to reorganize. Establishment of cellular polarity is required for asymmetric cell division. Notably, the signaling factors involved in establishing polarity in C. crescentus are conserved among bacteria from diverse environmental niches (22,(40)(41)(42)(43)(44), like Brucella abortus (causative agent for brucellosis in mammals) (45,46), Sinorhizobium meliloti (plant symbiont) (47), and Agrobacterium tumefaciens (plant pathogen) (48). However, little is known about how the gradients of these signaling polarity factors are formed or how they function in bacteria with diverse life styles. In this work, we asked what happens to the ability of cells to grow when the organization of the two highly conserved chromosomal loci (ori and parS centromere) are flipped in orientation. We showed that the regulators (DnaA and ParA) can easily adapt to the new locations of these sites and proceed with their activities, and in the case of ParA proceed to orient segregation in the reverse direction. Remarkably, cells were able to recover the "forced" rearrangement of these chromosomal loci and continue to grow with no measurable delays. Our results revealed the robustness and flexibility that cells have to rearrange their signaling polarity factors.

MATERIALS AND METHODS
Bacterial strains and growth conditions. Lists of strains, plasmids, and primers used in this work are provided in Tables S1 and S2 in the supplemental material. Plasmids constructed in this study were created by cloning PCR products amplified using wild-type CB15N (NA1000) or Yersinia pestis KIM5 pMT1 genomic DNA into pNPTS138, pXCHYC-2, or pXCFPN-2 vectors (49). The constructs were transformed into E. coli DH5␣ cells and grown at 37°C in Luria-Bertani (LB) medium. All primers used for cloning are listed in Table S2. Plasmid carrying cfp-parB(pMT1) was done by the parB(pMT1) gene sequence isolation with KpnI and NheI restriction from PM396 (LS5269) and ligation to the equally treated xylose-inducible integrating plasmid pXCFPN-2 (Kan r ) (9). The parA gene was cloned into integrating pXCHYC-2 (Kan r ) plasmid under a xylose-inducible promoter to express mCherry-tagged C-terminal protein fusions (49). The Gibson cloning method (50) was used to construct the plasmids used to delete or insert a gene into the Caulobacter genome.
Construction of indicator strain PM500 with fluorescently labeled origin of replication. To track the cellular localization of ori, we engineered a fluorescent tag to be inserted near ori using the Y. pestis parS(pMT1) chromosomal sequence and its corresponding gene encoding ParB(pMT1) (17). To insert the parS(pMT1) site approximately 1 kb far away from the ori, the cloned parS(pMT1) sequence from PM395 (LS5270) and around 600 bp of CCNA0001 C-terminal and CCNA0002 N-terminal sequences were assembled into the pNPTS138 plasmid (9). The parS-ParB(pMT1) system from Yersinia has been previously used in Caulobacter and shown not to interfere with the activity of the native Caulobacter ParABS partitioning system (9). To control the expression of dnaA, we first engineered an additional copy of dnaA to replace the vanA gene, resulting in the expression of dnaA regulated by the VanA promoter (vanA::dnaA) (49). Using this merodiploid strain, the native gene encoding DnaA was deleted, leaving no scars on the genome. The final indicator strain PM500 has the genotype xylX::cfp-parB(pMT1) parS(pMT1) at nucleotide 1108, dnaA::⍀, vanA::dnaA. We refer here to the parS(pMT1) localized near ori simply as ori.
Growth assays. Overnight cultures grown from Caulobacter frozen stocks in M2G liquid medium were diluted to an optical density at 600 nm (OD 600 ) of 0.2 (2 ml) in 13-mm glass tubes. Cultures were incubated at 28°C, and the optical density at 600 nm was monitored every hour to monitor the growth rates of bacteria.
Synchronization. A culture of Caulobacter in M2G (15 ml) was inoculated with a saturated overnight M2G culture and grown to an OD 600 of ϳ0.3. The medium was supplemented with vanillate (250 M) and antibiotics as noted. Cells were pelleted using centrifugation at 6,000 rpm for 10 min at 4°C. The cell pellet was resuspended in about 800 l of 1ϫ M2 salts and mixed well with Percoll (900 l; Sigma-Aldrich) to generate a density gradient. Swarmer cells (bottom layer) were separated out from the stalked/predivisional cells (top layer) by centrifuging at 11,000 rpm for 20 min at 4°C. Collected swarmer cells were washed twice with cold 1ϫ M2 salts by spinning at 8,000 rpm for 3 min at 4°C and resuspended in M2G medium to the appropriate OD 600 . When cells were not synchronized, the cultures grown to an OD 600 of ϳ0.3 were pelleted and washed with 1ϫ M2 salts three times.
Fluorescence microscopy. The cells (1 to 3 l) were spotted on agar pads (1% agarose in M2G) and imaged using phase-contrast and fluorescence microscopy in Zeiss Axio Observer 2.1 inverted microscope, set up with a Plan-Apochromat 100ϫ/1.40 Oil Ph3 M27 (WD ϭ 0.17 mm) objective, AxioCam 506 mono camera and ZEN lite software. Agar pads supplemented with vanillate (250 M) were used in time-lapse assays when needed. Images were analyzed using Fiji software (51), and localization of fluorescent foci was counted using the Cell Counter plugin.
CFU assay. The cultures were serially diluted (10-fold) by mixing 10 l of culture with 90 l of PYE medium in a sterile 96-well plate. Five microliters of each sample was spotted onto PYE agar (1.5%) plates supplemented with vanillate (250 M) if needed. CFU counts were obtained from the plates incubated at 28°C for 2 days.

ACKNOWLEDGMENTS
We thank Lucy Shapiro, Edward de Koning, Carol Potenza, and Lee Uranga for critically reading the manuscript and Ericka Flores for assistance in the construction of plasmids pDNA218/219. This work was funded by NIH SC2GM121180. A.B.M. was supported in part by CONACYT.
A.B.M., I.P.M., and P.E.M. designed the study and wrote the manuscript. A.B.M. and I.P.M. acquired and analyzed data.
We declare that we have no competing interests in relation to the work described in this article.