Growth Conditions Regulate the Requirements for Caulobacter Chromosome Segregation

Abstract

Growth environments are important metabolic and developmental regulators. Here we demonstrate a growth environment-dependent effect on Caulobacter chromosome segregation of a small-molecule inhibitor of the MreB bacterial actin cytoskeleton. Our results also implicate ParAB as important segregation determinants, suggesting that multiple distinct mechanisms can mediate Caulobacter chromosome segregation and that their relative contributions can be environmentally regulated.

One of the most fundamental biological processes is the faithful replication and segregation of genetic material during cell division. The MreB actin homolog and the ParAB system have emerged as two candidates that could mediate bacterial chromosome segregation (19). However, different studies have reported differences in the necessity for chromosome segregation of both MreB and ParAB in various model systems. For example, early studies that MreB plays an important role in Escherichia coli, Bacillus subtilis, Caulobacter crescentus, and Vibrio cholerae chromosome segregation (9, 13, 14, 17, 18) have been recently contradicted by reports that MreB is not essential for either E. coli or B. subtilis chromosome segregation (2, 7, 12). Similarly, while ParA and ParB plasmidic homologs have been clearly shown to mediate plasmid segregation (1, 4), the specific functions of chromosomally encoded ParAB remain unclear, and studies from different species report differences in the extent of their roles in chromosome positioning and/or translocation (3, 5, 6, 10, 16, 21, 23). Here we address a potential origin of these differences by examining the influence of growth conditions on chromosome segregation in Caulobacter crescentus and by using a rapid and specific chemical antagonist of MreB, A22 [S-(3,4-dichlorobenzyl)isothiourea]. A22 was originally identified as a compound that increased the frequency of anucleate cells in E. coli (11) and was subsequently found to specifically perturb MreB in Caulobacter, E. coli, and V. cholerae (9, 13, 18). Caulobacter represents an ideal model system for studying bacterial chromosome segregation, since replication occurs once and only once per cell cycle and the ready ability to isolate swarmer cells that contain a single chromosome enables chromosome segregation to be followed in synchronized cell populations.

Descriptions of the strains used in this study are provided in Table 1. Detailed descriptions of the materials and methods for the experiments described here can be found in the supplemental material.

TABLE 1.

C. crescentus strains

Strain	Relevant characteristic(s)	Reference
MT174	egfp-parB	20
MT148	pxyl-parB eyfp-mipZ	20
LS3814	pxyl-gfp-mreB	8
LS3827(MO1)	(LacO)n::CC0006; pxyl-lacI-cfp tetR-yfp	22

A22 has different effects on chromosome dynamics when cells are grown in liquid or on solid medium.

Before replication initiates in the Caulobacter cell cycle, the single origin of replication (ori) is localized to a cell pole. Soon after ori replication, one of the two ori's is rapidly translocated to the opposite cell pole, such that ori localization provides a powerful assay for Caulobacter chromosome segregation. We followed ori localization using the strain MT174 (20), which contains as its only copy of ParB a functional green fluorescent protein (GFP)-ParB fusion that rescues the parB viability and division defects. ParB specifically binds to ori-proximal DNA sequences, such that this strain serves as a faithful reporter for ori localization (16, 20). To examine the effects of growth conditions on Caulobacter chromosome positioning, GFP-ParB cells were synchronized by density centrifugation and then grown in either liquid medium or on 1% solid agarose pads in the presence of either 10 μg/ml A22 or the equivalent volume of the methanol diluent. This concentration of A22 is above the MIC for A22 for growth in either liquid or solid medium (see below for details) and is sufficient to delocalize MreB (9). The percentage of cells with bipolar GFP-ParB foci was scored at regular intervals under each of these four growth conditions. Since imaging the presence and location of GFP-ParB foci must be performed on solid pads, the cells grown in liquid were transferred to pads and rapidly imaged. In these experiments the cells' exposures to the pads (2 to 3 min) were significantly shorter than the time required for completing ori translocation (∼6 to 8 min) (22).

Cells grown on agarose pads rapidly translocated the newly replicated ori without A22 treatment but showed a severe block in ori translocation in the presence of A22 (Fig. 1a and c). Some cells also exhibited ori's that did move away from the initial pole but failed to find the opposite pole, consistent with a defect in ori targeting (Fig. 1c). Surprisingly, however, A22 treatment of cells grown in liquid media resulted in a delay, but not a block, of ori translocation (Fig. 1b and c). The above experiments were repeated with a strain, LS3827, harboring a lacO array near the ori. Induced LacI-CFP (cyan fluorescent protein) was used to visualize the lacO-labeled ori. This strain showed ori dynamics similar to MT174 under liquid and solid growth conditions, both with and without A22 treatment (see Fig. S1 in the supplemental material). These results demonstrate that A22 blocks origin dynamics of cells grown under solid conditions but only mildly delays origin dynamics of cells grown in liquid medium.

FIG. 1.

Effect of A22 treatment on ori segregation under solid and liquid growth conditions. (a and b) Percentages of cells (MT174) with bipolar GFP-ParB foci at different times after synchronization. Cells were grown with (red) or without (blue) 10 μg/ml A22 on M2G agarose pads (a) or in liquid M2G medium (b). Error bars indicate standard deviations (n > 100). (c) Representative images of the overlay of phase and GFP micrographs are shown for fields from the 0-min (left), 60-min (center), and 120-min (right) time points from the experiments in panels a and b. The arrow points to a cell with a defect in ori translocation. Bar, 1 μm.

Differences in A22-dependent chromosome dynamics are not due to differential MreB perturbations.

To examine whether the differential effects of A22 on chromosome positioning during solid and liquid growth are due to differential effects of A22 on MreB, we determined whether other A22-induced phenotypes, such as growth, cell shape, and MreB delocalization, are also more severe during solid as opposed to liquid growth. The growth of Caulobacter cells was more sensitive to A22 when cells were grown in liquid media than on solid media, with MICs of 1.25 μg/ml and 2.5 μg/ml, respectively. Since GFP-MreB is only partially delocalized by 2.5 μg/ml of A22 but completely delocalized by 10 μg/ml of A22 (data not shown), we used 10 μg/ml of A22 for the remainder of our experiments. Similar to the MIC results, cells grown overnight in the presence of A22 under liquid or solid conditions exhibited more pronounced cell shape defects in the liquid medium (Fig. 2c and d). Finally, as detailed in the supplemental materials, we found that A22 had a similar effect on the localization of a GFP-MreB fusion under either liquid or solid growth conditions. Thus, of all A22-induced phenotypes, chromosome positioning is the only one that is specifically exacerbated by growth on solid media, and most other phenotypes are more pronounced during liquid growth. Consequently, the differences in A22's effects on ori positioning under the two growth conditions are not due to differential effects on MreB, but rather reflect differential requirements for ori translocation. In addition, the inverse correlations of A22's effects on cell shape and chromosome positioning indicate that the chromosome positioning defects observed during growth on solid pads are not a secondary consequence of cell shape perturbations.

FIG. 2.

Effects of A22 on Caulobacter morphology and MreB localization in liquid versus solid medium. A strain expressing GFP-MreB was exposed to 10 μg/ml A22 in liquid M2G medium and on M2G agarose pads. Phase (top) and GFP (bottom) images were taken of cells grown in liquid medium (a and c) or on pads (b and d) with (c and d) or without (a and b) A22 treatment for 3 min, 90 min, or overnight (O/N). Bar, 1 μm.

A22 treatment and ParB depletion have a synthetic effect on chromosome segregation in liquid.

The presence of two different mechanisms for chromosome segregation could explain the lack of effect of A22 treatment observed under liquid culture conditions. One likely candidate for such a second mechanism is the Par system, which has been implicated in plasmid and chromosome dynamics in a number of species (10, 16, 23). In Caulobacter, the ParA and ParB proteins are essential due to their role in cell division (20). Overexpression of ParA or ParB results in a mild segregation defect (16), while overexpression of a dominant-negative ParA point mutant strongly inhibits the completion of chromosome segregation (21). To study ori dynamics upon Par system perturbation, we used a strain, MT148, that allows for conditional xylose-dependent expression of parB while imaging ori localization with a yellow fluorescent protein (YFP) fusion to another protein that colocalizes with the ori, MipZ (20).

In order to study the role of the Par system in chromosome segregation, cells were depleted of ParB for 5 h before the experiment, which resulted in a ∼3-fold reduction in ParB levels (see Fig. S2 in the supplemental material) (15, 20). Consistent with our previous results, MT148 cells grown in liquid without ParB depletion rapidly translocated their origins in the absence of A22, and A22 treatment delayed but did not block ori translocation (Fig. 3a and b). Cells depleted of ParB in the absence of A22 showed a delay in origin translocation similar to that of A22 treatment alone (Fig. 3a and b). This delayed ori translocation could reflect either the incomplete depletion of ParB in these cells or the presence of a second segregation mechanism that can partially compensate for the reduced ParB. When ParB depletion was combined with A22 treatment, cells showed a dramatically more severe translocation defect than either ParB depletion or A22 treatment alone (Fig. 3a and b). This synthetic interaction between ParB depletion and A22 treatment raises the possibility that these two perturbations target different aspects of the process of chromosome segregation.

FIG. 3.

Synthetic effects of A22 treatment and ParB depletion on ori segregation. (a) Percentages of cells (MT148) with bipolar MipZ-YFP foci at different times after synchrony for cells that were untreated (blue), A22 treated (black), depleted of ParB for 5 h (red), or both A22 treated and ParB depleted (green). Error bars indicate standard deviations (n > 100). (b) Representative images of the overlay of phase and YFP micrographs for fields from the 0-min (left) and 45-min (right) time points from the experiment in panel a. Bar, 1 μm.

The combined effect of A22 treatment and ParB depletion is on ori translocation and not replication.

The previous experiment demonstrated that the combination of A22 treatment and ParB depletion causes a severe defect in Caulobacter ori dynamics. However, our assay follows origins in situ, such that our observation that these cells generally retain a single ori focus cannot resolve a defect in chromosome segregation from a defect in chromosome replication. To distinguish these possibilities, we developed a method for using microscopy and flow cytometry to assay ori localization and DNA content, respectively, in the same cell population (see the supplemental material for details). The microscopy and flow profile for ParB-depleted cells treated with A22 for 3 h after synchronization demonstrated that while these cells primarily contained single polar MipZ-YFP foci, they were overwhelmingly in the fully replicated (2N) state (see Fig. S3 in the supplemental material). These results suggest that there is no block in chromosomal replication as a result of combined depletion of ParB and A22 treatment and that the synthetic defect in ori dynamics results from a defect in ori segregation.

The ability to examine both the DNA content and ori positioning of the same cell population also enabled us to determine whether the A22-induced delay in the ori positioning of cells grown in liquid media (Fig. 1) reflects delayed replication or segregation. We found that these cells exhibited a delay in the initiation of chromosome replication which correlated with their delayed ori positioning (see Fig. S4 in the supplemental material). Since flow cytometry must be performed in liquid, we have not been able to directly assess DNA replication in cells grown on solid pads. Together, our data suggest that when cells are grown in liquid medium, A22 treatment can disrupt ori translocation without blocking DNA replication, but only when ParB is also simultaneously perturbed.

Does MreB function in chromosome segregation?

The potential role of MreB in mediating bacterial chromosome segregation has recently emerged as a controversial topic. Early reports suggested that MreB plays an important role in segregating the chromosomes of E. coli, B. subtilis, Caulobacter, and V. cholerae (9, 13, 14, 17, 18). However, more recent studies in both E. coli and B. subtilis have failed to support the role of MreB in chromosome segregation (2, 7, 12). Our discovery that A22 treatment alone does not significantly affect Caulobacter ori translocation when cells are grown in liquid medium is consistent with the latter reports that MreB is not absolutely necessary for segregation in other species. At the same time, our results indicating that the combination of A22 treatment and ParB depletion strongly perturbs ori translocation without blocking replication raise the possibility that the existence of multiple segregation mechanisms mask the role of MreB in mediating segregation in different contexts. This hypothesis is further supported by the fact that mreB appears to be the specific target of A22 in Caulobacter and other bacterial species (9, 13, 18) and that partially A22-resistant mreB point mutants shift the dose-response curve for A22's effect on ori translocation (data not shown). However, it is still formally possible that the A22 effects we have observed are a consequence of the influence of A22 on another cellular target and that MreB does not play a significant role in Caulobacter chromosome dynamics.

Multiple pathways can mediate chromosome segregation.

The combined effects of A22 treatment and ParB depletion on the ori dynamics of Caulobacter cells grown in liquid medium suggest that multiple pathways can collaborate in a partially redundant fashion to mediate chromosome segregation. In light of the essential nature of chromosome segregation, it is perhaps not surprising that multiple mechanisms have evolved as safeguards. Nevertheless, it is surprising that this partial redundancy is only observed under specific growth conditions. It is possible that under stressed conditions neither pathway is sufficient, rendering each pathway necessary. Since different methods for visualizing chromosomal loci in vivo with different DNA-binding proteins might put different amounts of strain on the system, this scenario could reconcile conflicting reports about the effects of A22 on E. coli chromosome segregation (e.g., references 12 and 13). The relevant difference between liquid and solid growth conditions remains unclear. Preliminary experiments suggest that the impact of A22 on chromosome segregation is not affected by differences in oxygen levels or viscosity. In the future, it will prove interesting to determine what physiological parameters regulate chromosome segregation, how they are sensed, and how they exert their effect.