Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Cell. Author manuscript; available in PMC 2009 September 19.
Published in final edited form as:
PMCID: PMC2614312

A multimeric pole-organizing protein critical for chromosome attachment, division and protein localization in Caulobacter


Cell polarization is an integral part of many unrelated bacterial processes. How intrinsic cell polarization is achieved is poorly understood. Here, we provide evidence that Caulobacter crescentus uses a multimeric pole-organizing factor (PopZ) that serves as a hub to concurrently achieve several polarizing functions. During chromosome segregation, polar PopZ captures the ParB•ori complex and thereby anchors sister chromosomes at opposite poles. This step is essential for stabilizing bipolar gradients of a cell division inhibitor and setting up division near midcell. PopZ also affects polar stalk morphogenesis and mediates the polar localization of the morphogenetic and cell cycle signaling proteins CckA and DivJ. Polar accumulation of PopZ, which is central to its polarizing activity, can be achieved independently of division and does not appear to be dictated by the pole curvature. Instead, evidence suggests that localization of PopZ largely relies on PopZ multimerization in chromosome-free regions, consistent with a self-organizing mechanism.


Bacterial cells display an intrinsic polarization affecting many aspects of their life (Bardy and Maddock, 2007; Ebersbach and Jacobs-Wagner, 2007). Numerous functionally unrelated proteins localize to the cell poles, affecting a wide variety of processes including chemotaxis, signal transduction, polar morphogenesis and pathogenesis. Cell polarization also plays an essential role during DNA segregation and cell division. Equal partitioning of chromosomes and plasmids relies on sister DNA origins localizing near opposite poles (Ebersbach and Gerdes, 2005; Thanbichler and Shapiro, 2006a) and the proper location of cell division is governed by polar localization of cell division inhibitors (Lutkenhaus, 2007). In recent years, we have learned more about how specific molecules or complexes are individually targeted or retained at a pole. However, it remains ill-defined whether bacteria, like eukaryotic cells, use organizing factors to govern multiple polarization events simultaneously.

We address this fundamental question in the α-proteobacterium Caulobacter crescentus whose life cycle is rich in well-documented polarized events (Lawler and Brun, 2007). Organelles such as stalks, flagella and pili form at specific poles during the cell cycle (Figure 1A). The coupling between polar morphogenesis and the cell cycle is achieved through an intricate regulatory network, including several histidine kinases that exhibit polar localization during the cell cycle (Goley et al., 2007). In C. crescentus, the origin of replication (ori) is located at the 'old' pole (that existed in the previous cell cycle) (Figure 1A). The DNA partitioning protein ParB binds to a parS centromeric sequence nearby ori (Mohl and Gober, 1997). After initiation of DNA replication, one of the duplicated ParB•ori complexes rapidly migrates toward the 'new' pole (created by the most recent division), in a process that requires the MreB cytoskeleton (Gitai et al., 2005; Jensen and Shapiro, 1999). What retains ParB•ori at the pole has remained mysterious. Bipolar localization of ParB•ori is, however, crucial for setting up division (Thanbichler and Shapiro, 2006b). MipZ, an inhibitor of the FtsZ cytokinetic structure, binds to bipolar ParB•ori, favoring FtsZ assembly in the central cell region where MipZ concentration is low (Thanbichler and Shapiro, 2006b). Thus, in C. crescentus, cell polarization is critical for temporal and spatial execution of chromosome segregation, cell division and polar morphogenesis.

Figure 1
A multicopy plasmid screen identifies a factor involved in cell division placement

While cell polarization is clearly crucial for many functions in C. crescentus, we know little about the mechanisms involved in pole recognition and organization. Recently, TipN was identified as a spatial cue that specifies the site of the last division (Huitema et al., 2006; Lam et al., 2006). In the absence of TipN, several new-pole markers exhibit an abnormal frequency of old-pole localization (Huitema et al., 2006; Lam et al., 2006). How these markers localize to the wrong pole is unclear. Moreover, old-pole and bipolar markers appear largely unaffected by TipN. Thus, additional mechanisms must be involved in spatially organizing the cell.

In this study, we identify a self-assembling, multifunctional protein that serves as a pole-organizing center to mediate several polarizing functions important for chromosome attachment, cell division, stalk morphogenesis and protein localization.


A screen for polarization factors involved in cell division

In bacteria, cell polarization plays an important role in cell division site selection (Lutkenhaus, 2007). Consistently, overproduction of the TipN polarity factor causes cell filamentation, minicelling and cell branching in C. crescentus (Lam et al., 2006). We exploited this observation to search for additional polarity factors, and designed a screen for genes that cause a division phenotype when overexpressed from a multicopy plasmid (See supplemental text). From this screen, we isolated the hypothetical gene cc_1319, now renamed popZ (for pole-organizing protein that affects FtsZ; see below). Overexpression of popZ from the xylose-inducible promoter (Pxyl) on a multicopy plasmid caused a division defect (Figure 1B). Time-lapse microscopy in the presence of xylose revealed that PopZ overproduction initially resulted in one or several divisions near the new pole (opposite the stalk), generating minicells (Figure 1C). Over time, the cells stopped dividing except for occasional erratic divisions, generating a heterogeneous population of cell filament sizes. The pattern of FtsZ ring localization (seen as FtsZ-YFP bands) in popZ-overexpressing cells was consistent with this division pattern (Movie 1).

Based on its NCBI annotation, PopZ is a 177-residue protein predicted to be cytoplasmic. While the PopZ sequence does not match any known domain or motif, BLAST analysis showed that PopZ homologs are widely present in α-proteobacterial genomes.

PopZ exhibits polar localization

PopZ fused to a tetracysteine (TC) motif (encoded at the native locus as the only copy) was found at one or both poles in an asynchronous cell population (Figure S1A). Time-lapse experiments with synchronized cells expressing a functional popZ-yfp fusion from the native locus as the only copy of popZ (see supplemental text) revealed that PopZ-YFP localizes at the old pole of swarmer cells, later adopting a bipolar localization (Figure 2A and Figure S1B). Division breaks this symmetry and yields daughter cells with PopZ-YFP at their old pole.

Figure 2
PopZ colocalizes with ParB at the poles

The fact that the chromosomal ori exhibits a similar unipolar localization after division and bipolar localization after duplication and segregation (Jensen and Shapiro, 1999) led us to simultaneously visualize PopZ-YFP and ori using CFP-ParB, which binds to the parS sequence adjacent to ori (Mohl and Gober, 1997; Thanbichler and Shapiro, 2006b). This showed that PopZ-YFP starts accumulating at the new pole to adopt a bipolar localization at about the time when the segregation of CFP-ParB•ori is completed (Figure 2B).

PopZ captures the ParB•ori complex and thereby anchors the chromosomes at the poles

A popZ deletion strain was viable but grew more slowly than wild-type and exhibited defects in cell division placement, generating minicells and elongated cells (Figure 3A). Replication and segregation of ori still occurred in ΔpopZ cells, as evidenced by the presence of two or more CFP-ParB•ori foci (Figure 3B). However, poles were often devoid of CFP-ParB foci. This was not due to filamentation of ΔpopZ cells because filamentous FtsZ-depleted cells showed bipolar localization of CFP-ParB in addition to internal CFP-ParB foci (Figure 3B). Normally, the origins remain at opposite poles after segregation (Jensen and Shapiro, 1999; Viollier et al., 2004), as if attached (Figure 3C). Conversely, in ΔpopZ cells, CFP-ParB foci moved within a constrained region (Movie 2 and Movie 3, Figure 3D). This motion was evident in kymograph analyses of CFP-ParB distribution from time-lapse recordings (Figure 3E). In contrast, polar CFP-ParB foci in FtsZ-depleted cells were held in place while internal CFP-ParB foci moved rapidly within a restricted region (Movie 4, Figure 3F). Since PopZ-YFP localized at both poles in FtsZ-depleted cells (Figure 3G), our observations argued that PopZ docks ParB•ori at the poles.

Figure 3
Deletion of PopZ results in minicelling, cell filamentation and failure of polar chromosome attachment

This notion was supported by the behavior of CFP-ParB in a strain in which a xylose-inducible mYFP-PopZ fusion is the only source of PopZ. After growth without xylose to obtain a PopZ phenotype, the cells were placed on a slide with xylose to induce myfp-popZ expression and were imaged by time-lapse microscopy (Movie 5Movie 6, Figure 3H). At first, there was no detectable mYFP-PopZ, and all CFP-ParB•ori foci displayed motion as expected. Over time, mYFP-PopZ accumulated primarily at the poles while the CFP-ParB•ori foci remained mobile, as the polar localization of PopZ preceded that of CFP-ParB. Eventually, during its motion, a CFP-ParB•ori focus would cross the polar mYFP-PopZ site. From then on, this focus stopped moving and remained colocalized with mYFP-PopZ, suggesting a ‘capture’ mechanism (Movie 5Movie 6, Figure 3H). On few occasions, the newly synthesized mYFP-PopZ also accumulated at an ectopic location along the cell length where it subsequently captured a nearby CFP-ParB•ori complex (Figure 3H, Movie 5), indicating that PopZ can recruit ParB•ori independently of the poles.

Co-immunoprecipitation (Co-IP) experiments using anti-GFP antibody (which recognizes YFP) and cell lysates of CB15N popZ::popZ-yfp revealed that ParB is specifically pulled down with PopZ-YFP (Figure 3I). Similar experiments with purified poly-histidine-tagged ParB and PopZ (Figure S2) showed that His6-PopZ co-immunoprecipitated with His6-ParB (Figure 3J), suggesting a physical interaction between PopZ and ParB. This is supported by the E. coli experiments presented below (Figure 6B).

Figure 6
PopZ localizes to the pole in E. coli

PopZ-mediated docking of ParB•ori to the poles is required for proper temporal and spatial regulation of division

In C. crescentus, cell division and chromosome segregation are connected, as the FtsZ assembly inhibitor MipZ binds to ParB (Thanbichler and Shapiro, 2006b). The instability of ParB•ori in ΔpopZ cells resulted in erratic movement of MipZ-YFP in ΔpopZ cells (Movie 7), which likely decreased the probability of FtsZ assembly at one specific location. Consistent with this, an FtsZ-YFP reporter exhibited rapid movements in ΔpopZ cells and was able to form a stable and functional structure (seen as a band) only on occasion and at aberrant places (Movie 8), reflecting the ΔpopZ division pattern described above (Figure 3A). This argues that polar anchoring of ParB•ori is critical for proper division by stabilizing MipZ gradients at the poles, thereby allowing stable FtsZ assembly near midcell.

All together, our findings indicate that polar PopZ captures the migrating ParB•ori complex during chromosome segregation, which, in turn, affects where and when division occurs.

PopZ self-associates into multimers

Since PopZ overproduction results in division defects (Figure 1B), we examined its distribution under this condition. Overproduction of an N-terminal GFP or C-terminal tetracysteine fusion to PopZ (GFP-PopZ or PopZ-TC) recapitulated the cell division phenotypes observed with untagged PopZ (Figure 1B–C). Visualization of FlAsH-stained PopZ-TC revealed a large region of accumulation from the old pole extending into the cell body (Figure 4A). Often, there was a second, weaker signal at the opposite pole. The PopZ accumulation appeared homogeneous rather than concentrated at the membrane, suggesting cytoplasmic accumulation. This was confirmed by 3D-reconstruction of optical sections after 3D-deconvolution (not shown). The fact that excess PopZ did not disperse but partitioned at the pole suggested that PopZ has a cohesive (i.e., self-associating) property.

Figure 4
PopZ multimerizes and during overproduction, it forms a large polar accumulation that excludes DNA and ribosomes

Time-lapse recordings of GFP-PopZ during its overproduction revealed that the biased old-pole accumulation is the result of inheritance (Figure 4B). At first, GFP-PopZ progressively accumulated at both poles. Division created new poles, and the daughter cell that inherited the older pole had more GFP-PopZ than its sibling because of its initial head start. Minicelling, which typically occurred at the new pole, enhanced this asymmetry. Repetition of this pattern in subsequent division cycles resulted in expansion of the GFP-PopZ region at the older poles.

Freely diffusible GFP was evenly dispersed in cells overproducing untagged PopZ, indicating that the PopZ-rich region contains cytoplasmic content (Figure 4C). Even though PopZ was untagged and therefore could not be visualized directly, we were able to recognize the PopZ-rich region in DIC images because of a small but reproducible shading difference between the region of PopZ accumulation and the rest of the cell (Figure 4C).

Intrigued by the accumulation of PopZ and the optical effect it generates, we examined popZ-overexpressing cells by electron cryotomography. Tomograms of wild-type C. crescentus cells have shown that the entire cytoplasm is packed with electron-dense ribosomes (Briegel et al., 2006). In striking contrast, in popZ-overexpressing cells, the PopZ-rich region was largely depleted of ribosomes (Figure 4D). This provides an explanation for the shading difference, as a ribosome-rich region should be optically different than a ribosome-depleted region. Chromosomal DNA was also excluded from the PopZ-rich region as visualized by fluorescence microscopy using DAPI and FlAsH to stain DNA and the PopZ-TC region, respectively (Figure 4E).

One possible interpretation of these data is that PopZ multimerizes into a matrix that excludes DNA and ribosomes (the latter possibly because they are attached to DNA through transcription and translation coupling) but not small components (such as free GFP). Consistent with PopZ self-association, purified His6-PopZ, which has a predicted molecular weight (MW) of 21 kDa (but runs as a 35–36 kDa protein by SDS–PAGE; Figure S2), migrates as a single band between the 480 and 720-kDa markers on a native gel (Figure 4F). To determine the MW of the His6-PopZ multimer independently of its shape (Folta-Stogniew, 2006), His6-PopZ was subjected to HPLC size exclusion chromatography (SEC) coupled with UV, on-line laser light scattering (LS) and refractive index (RI) detectors (SEC-UV/LS/RI) (Figure 4G). Measurements indicated that His6-PopZ forms ordered structures between 120 and 170 kDa. The slow migration of His6-PopZ in the native gel and its early elution position from the SEC column (ahead of the 475 kDa standard) indicate that His6-PopZ self-assembles into a non-spherical oligomer.

Formation of large PopZ structures affects the localization of ori, ParB and MipZ

The cumulative profile of PopZ accumulation suggests that the multimeric state of PopZ increases with overproduction, creating a large matrix that expands from the pole into the cell body. One and sometimes two distinct ori were consistently found at the border of the expanding PopZ regions (red arrows, Figure 4H; data not shown), as visualized by the lacO array/LacI-CFP system (Viollier et al., 2004). Thus, PopZ matrix expansion pushes the anchored DNA origin farther from the pole. CFP-ParB was found to 'fill' the entire PopZ-rich region while also forming a focus at its border (Figure 4I) where a DNA origin resides (Figure 4H). Weak CFP-ParB foci were also occasionally observed along the cell length, likely marking segregated ori (Figure 4I). The retention of CFP-ParB in the DNA-free PopZ-rich region is consistent with the observed affinity between ParB and PopZ. MipZ-YFP was largely found at the border of the PopZ-rich region and occasionally formed one or two weaker foci along the cell length (Figure 4J). Since MipZ affects FtsZ assembly, sequestration of MipZ at the PopZ matrix border may cause the division defect observed in the PopZ overproduction strain.

PopZ achieves multiple polarizing functions

PopZ plays a role in chromosome capture and cell division by mediating the stable localization of ParB•ori, and thereby MipZ, at the poles. This is reminiscent of the function of the Bacillus subtilis DivIVA protein, a multimeric polar protein that anchors the chromosomes at the poles during sporulation through a (direct or indirect) interaction with origin-associated RacA (Ben-Yehuda et al., 2003; Stahlberg et al., 2004; Thomaides et al., 2001; Wu and Errington, 2003). During vegetative growth, DivIVA plays a distinct role by maintaining the cell division inhibitor complex MinCD at the poles (Marston and Errington, 1999). Despite their similar functional properties, PopZ and DivIVA share no sequence similarity. Furthermore, DivIVA has a largely coiled-coil structure (Edwards et al., 2000) whereas PopZ has no predicted coiled-coil motifs. Another notable difference is that the two polarizing functions of DivIVA do not overlap in time (normal growth vs sporulation) and are thought to be competitive (Ben-Yehuda et al., 2003) whereas PopZ achieves multiple polarizing tasks simultaneously as we show below. In addition to chromosome capture and division, PopZ is involved in polar stalk formation since stalks were undetectable when ΔpopZ cells were analysed by SEM (Figure S3). Moreover, the polar localization of the histidine kinase DivJ, an old-pole marker involved in polar morphogenesis and cytokinesis sensing (Matroule et al., 2004; Sommer and Newton, 1991; Wheeler and Shapiro, 1999), was disrupted in ΔpopZ cells (Figure 5A). In some cells, DivJ-YFP showed some accumulation at a pole (in addition to the patchy distribution elsewhere), but this polar accumulation was unstable and transient (Movie 9). The cell cycle signaling protein CckA, which exhibits unipolar and bipolar localization during the cell cycle (Jacobs et al., 1999), was also dramatically affected by the loss of PopZ function (Figure 5A). These defects were probably not caused by the cell elongation phenotype of some ΔpopZ cells as both DivJ and CckA retain a polar localization in elongated FtsZ-depleted cells (Biondi et al., 2006; Matroule et al., 2004).

Figure 5
PopZ affects the localization of CckA, DivJ and TipN

The severe loss of polar localization for both DivJ and CckA suggested that PopZ is required for the polar recruitment and/or maintenance of these proteins. Consistent with this idea, both CckA and DivJ co-immunoprecipitated with PopZ-YFP (Figure 5B). In addition, membrane-bound CckA and DivJ were found at the membrane periphery of the PopZ-rich region of PopZ-overproducing cells (Figure 5C). In contrast, a GFP fusion to the CheA chemotaxis protein formed a tight focus at the polar tip of the PopZ-rich region (Figure S4), indicating that the localization pattern of CckA and DivJ is a result of a (direct or indirect) interaction with PopZ, rather than the consequence of an expansion of the polar region.

Thus, PopZ mediates several distinct polarizing functions at the same time, suggesting that PopZ forms a multifunctional platform that organizes the poles. PopZ also affects the localization pattern of the TipN polarity factor. In wild-type cells, TipN localizes at the new pole until the end of the cell cycle when it delocalizes from the pole to accumulate at the site of division, producing progeny with TipN at their new poles (Huitema et al., 2006; Lam et al., 2006). ΔpopZ cells, on the other hand, often failed to release TipN from the new pole, but still formed a focus at the division site (Figure 5D). This resulted in a high frequency (87%) of cells with bipolar TipN localization, which is not observed in filamentous FtsZ-depleted cells (Huitema et al., 2006; Lam et al., 2006). Unipolar components that are affected by TipN (such as the FliG flagellar protein, the PleC histidine kinase, CheA, and the CpaE pilus assembly protein) all remained polarly localized in ΔpopZ cells but displayed varying degree of abnormal bipolar localization (Table S1), which may reflect the effect of PopZ on TipN localization. Thus, PopZ also affects several aspects of cellular asymmetry. Combining ΔpopZ and ΔtipN mutations was synthetically lethal, and depletion of TipN in a ΔpopZ background yielded a severe cell filamentation defect (Figure 5E), indicating that cell polarity is required for cell division and viability in general.

PopZ polarly localizes in E. coli where it recruits ParB

A key aspect of PopZ function is its polar accumulation. Remarkably, PopZ-TC retained this ability even when it was artificially produced in the vastly divergent, γ-proteobacterium E. coli MC1000, which lacks PopZ homologs (Figure 6A). The localization was mostly unipolar in minimal M9 medium whereas the percentage of cells with bipolar localization increased in rich LB medium (not shown).

Since evidence suggests that ParB and PopZ interact without the need of intermediate factors (Figure 3J), these proteins should interact even when artificially produced in E. coli, which lacks homologs of both. When popZ-tc expression was not induced in E. coli, CFP-ParB had an expected diffuse distribution in the cytoplasm given the absence of a parS binding sequence (Figure 6B). However, when PopZ-TC synthesis was induced with arabinose, CFP-ParB was recruited to the PopZ-TC pole (Figure 6B). CFP-MipZ, used here as a control, retained a dispersed cytoplasmic distribution in E. coli cells even when PopZ-TC was polarly present (Figure 6C). These data further substantiate the notion of a physical association between PopZ and ParB and provide evidence that PopZ is functional in the heterologous E. coli system.

Polar accumulation of PopZ involves PopZ multimerization and DNA occlusion

Pole recognition in bacteria is an unresolved question of fundamental importance. Typically, a polar ‘anchor’ protein is invoked. Our findings strongly argue that PopZ plays such a role for several polar components. But how does PopZ accumulate at the poles? The fact that PopZ ‘recognizes’ E. coli poles despite the evolutionary distance between E. coli and C. crescentus indicates a commonality in some property of bacterial poles. Such a universal property is consistent with the previous finding that B. subtilis DivIVA localizes to the poles (and septum) of E. coli (Edwards et al., 2000). Since poles originate from division, PopZ may recognize a remnant of division. To test this, we looked at PopZ in TipN-overproducing cells, which produce ectopic poles independently of division events (Lam et al., 2006). GFP-PopZ localized at the ectopic poles (Figure 6D), indicating that cell division is not essential for PopZ localization. A similar conclusion was reached for DivIVA (Hamoen and Errington, 2003; Harry and Lewis, 2003). Another popular hypothesis is that polar proteins are retained at the poles through direct or indirect binding to the inert peptidoglycan cell wall. We might then expect digestion of the peptidoglycan to result in dispersion of the protein. However, PopZ still formed tight foci in E. coli protoplasts obtained by lysozyme treatment (Figure 6E).

PopZ localizes near the cell periphery, suggesting an affinity for the membrane. Since cell poles exhibit a higher degree of membrane curvature than the lateral sides, another attractive hypothesis is that polar accumulation of proteins might rely on geometric requirements, perhaps through interactions with high-curvature phospholipids that cluster at the poles (Howard, 2004; Huang et al., 2006). To explore the idea that PopZ clustering is dependent on pole curvature, we performed a series of experiments using spherical cells obtained by treating E. coli with mecillinam or A22, which inhibits the function of rod shape determinants PBP2 or MreB, respectively (Gitai et al., 2005; Iwai et al., 2002; Tamaki et al., 1980). Only the A22 experiments will be presented as we obtained similar results with both drugs. To test whether the formation of YFP-PopZ foci is independent of geometric constraints, cells were pre-treated with A22 until a spheroid shape was achieved. These spheroid cells were then placed on a slide containing arabinose and A22 to initiate yfp-popZ expression while maintaining a spherical shape. Time-lapse microscopy (Figure 6F) revealed that YFP-PopZ was first uniformly distributed, then formed small, mobile foci that appeared and dissolved rapidly near the cell periphery. But ultimately all foci coalesced into a single focus that displayed little mobility and that grew in intensity. Note that the contrast of the entire time-lapse sequence was greatly enhanced to permit visualization of weak YFP-PopZ signals in early time points. A similar sequence of events was observed when untreated, rod-shaped E. coli cells were used (Figure 6G, Movie 10). Collectively, these experiments suggest that self-association of PopZ plays a major role in protein clustering and that specific membrane curvature is not a key factor in the clustering process.

Another common feature among bacterial poles is the absence of DNA, which suggested to us that PopZ complex formation may be facilitated in DNA-free regions. In agreement with this idea, induction of YFP-PopZ synthesis in filamentous E. coli cells blocked for cell division by cephalexin generated multiple YFP-PopZ foci exclusively in chromosome-free regions (based on DAPI staining), including between nucleoids (Figure 6H). A DNA occlusion mechanism would be particularly appropriate for C. crescentus, where the DNA fills virtually the entire cell except for the polar tips (Viollier et al., 2004). Consistent with a DNA occlusion mechanism, PopZ-YFP foci formed largely outside the DAPI-stained DNA in regions corresponding to the tip of the poles (Figure 6I).

If DNA spatially restricts PopZ multimerization to the poles in C. crescentus, then PopZ should be able to form foci in any chromosome-free regions, as we observed in filamentous E. coli (Figure 6H). In C. crescentus, a block in division results in filamentous cells with no obvious separation of nucleoids (Ward and Newton, 1997). However, simultaneous inhibition of the cell division protein FtsA and the topoisomerase IV subunit ParE (using a ftsAts parEts double mutant at the restrictive temperature) creates filamentous cells with large DNA-free regions (Ward and Newton, 1997). Under these conditions, YFP-PopZ formed ectopic foci exclusively in the DNA-free regions (Figure 6J). This was also observed when yfp-popZ was expressed under the native popZ promoter (Figure S5).

Collectively, the data strongly argue that the polar localization of PopZ is achieved through PopZ multimerization in chromosome-free regions near the membrane.


We show here that PopZ achieves multiple polarizing functions in the cell. One is to bind ParB•ori and anchor the chromosomes at the poles. This plays a critical role in the spatio-temporal regulation of cell division by stabilizing MipZ gradients at opposite poles, favoring assembly of FtsZ near midcell (Figure 7A). PopZ also mediates polar localization of DivJ and CckA while maintaining its chromosome binding function. PopZ is also involved in other polarized events because without PopZ, stalk morphogenesis is impaired and the fidelity of asymmetric localization of TipN (and TipN-controlled components) is compromised. The molecular basis for these additional polarity defects is less clear.

Figure 7
Proposed models for PopZ function and localization

Polar accumulation of PopZ, which is clearly central to its multifaceted polarizing activity, appears to rely on a possible self-organizing mechanism. We propose that after nucleation of PopZ at the membrane, self-association of PopZ into a large multimer or matrix is favored in chromosome-free regions, as shown in both E. coli and C. crescentus. The DNA polymer may exert its inhibitory activity by steric hindrance or through a general repulsion mechanism. During normal growth of C. crescentus, localization of PopZ at the old pole is inherited from the previous division cycle and provides a primer for further accumulation. We suggest that the presence of a small DNA-free region at the opposite pole results in a progressive accumulation of PopZ at this location (Figure 7B). The PopZ multimer at the new pole captures the segregated ParB•ori complex. Continued synthesis and self-association of PopZ results in expansion of the PopZ matrix. Co-visualization of PopZ-YFP and CFP-ParB•ori (both present at native levels) revealed that the ParB•ori signal is often found overlapping with the pole-distal tip of the PopZ signal (Figure 6K). This is because ParB is at the interface between the growing PopZ matrix and the polymeric DNA. This effect is amplified under popZ-overexpressing conditions when a massive extension of the PopZ matrix progressively pushes ParB, the attached DNA origin and the rest of the chromosome away from the pole (Figure 4H–I). This is likely because PopZ and the chromosome are both polymeric, and polymers cannot mix well for steric reasons.

During the normal cell cycle, the PopZ matrix also binds, directly or indirectly, CckA and DivJ (and possibly other molecules) (Figure 7B). The localization of ParB•ori at the poles closely parallels that of PopZ during the cell cycle, consistent with a simple relationship. This is, however, not true for DivJ and CckA. While PopZ is present at the poles where and when CckA and DivJ localize, CckA and DivJ, which have distinct cell cycle patterns of localization (Jacobs et al., 1999; Wheeler and Shapiro, 1999), are not always at the poles where and when PopZ is present. This partial overlap indicates that PopZ does not dictate the timing of CckA and DivJ localization. Other factors must achieve this role by modulating the association of PopZ with CckA and DivJ during the cell cycle.

PopZ homologs are widely found among α-proteobacteria, feeding into the notion that cell polarization is an important cellular attribute in α-proteobacteria (Hallez et al., 2004), which form a large class of organisms with extensive medical and agricultural relevance. Perhaps more importantly, this study provides mechanistic insights into how bacteria can organize their poles. Many unlinked bacterial processes involve the polar localization of various proteins. The identification of PopZ suggests that bacteria can use multifunctional proteins that serve as "hubs" to achieve and perhaps coordinate multiple polarizing activities. Such organizing centers may correspond to a primitive version of the centralized polarity system of eukaryotic cells. Furthermore, our study suggests a mechanism of spontaneous self-organization for pole recognition, which relies on protein multimerization, membrane interaction and DNA occlusion. This proposed mechanism might represent an important principle of bacterial cellular organization. It will be interesting to test whether the multimeric DivIVA protein, the lattice-forming chemoreceptors or other proteins follow a similar principle of polar localization.

Experimental Procedures

Strains, plasmids and growth conditions

Strains and plasmids are listed in Table S2. Their construction is described in the supplemental text. Transformations, conjugations and phage transductions were carried out as described (Ely, 1991). C. crescentus strains were grown to log-phase at 30°C in PYE or M2G minimal medium (Ely, 1991). When needed, cell populations were synchronized as described (Evinger and Agabian, 1977). E. coli strains used for microscopy experiments were grown to log-phase at 30°C or 37°C in M9-minimal medium supplemented with 0.2% glycerol and 0.1% casamino acids unless otherwise stated. When required, gene expression was induced by adding 0.03–0.3% xylose, 0.5 mM vanillic acid or 0.02% arabinose unless otherwise stated. When needed, A22 (50µM) or cephalexin (10 µg/ml) were used. E. coli protoplasts were generated as described (Dai et al., 2005).

Light microscopy

Microscopy was performed by using a Nikon E1000 microscope and a Hamamatsu Orca-ER LCD camera, or a Nikon E80i microscope and an Andor iXonEM+ camera. Images were taken and processed with Metamorph 6.1r0 software. Samples were placed on agarose-padded slides containing medium and xylose, vanillic acid, arabinose or IPTG when required. DNA and TC-tagged proteins were visualized by using DAPI and FlAsH or ReAsH (Invitrogen), respectively.


Cells were harvested, washed twice in M2G, and resuspended in fixation solution for 30 min at RT (5% glutaraldehyde, 4% formaldehyde in 0.08M sodium phosphate buffer pH 7.2). Fixed cells were washed in PBS and mounted onto poly-L-lysine coated cover-slips. Cells were dehydrated through an ascending series of ethanol baths ending in 100% ethanol, and were critical point dried and gold-coated. Samples were examined by using a FEI XL-30 ESEM FEG microscope (acceleration voltage: 10.0 kV; spot size: 3; working distance: 7.5 mm or 10 mm).

Electron cryotomography

Data were collected on a FEI Polara™ (FEI Company, Hillsboro, OR, USA), 300 kV FEG transmission electron microscope with a Gatan energy filter configured with a slit width of 20 eV, on the lens-coupled 4k by 4k Ultracam CCD (Gatan). Single axis tilt-series were recorded from −60° to 60° with an increment of 1° at 10 or 12 µm underfocus, using UCSF-Tomo (Zheng et al., 2007), with a pixel size of 0.96 nm and a cumulative dose of up to 200 e/A2. Three-dimensional reconstructions were calculated using IMOD (Mastronarde, 1997). To visualize ribosome distribution, the template-matching program Molmatch (Bohm et al., 2000; Ortiz et al., 2006) was run using the crystal structures of the E. coli ribosome (PDB ID 2aw7, 2awb) as a template. The cross-correlation peaks, selected by a combination of an automatic peak search and manual interpretation, were visualized using the Amira software package (Mercury Computer Systems).

In vivo Co-IP

Cells were harvested, washed with buffer (20 mM HEPES pH 7.5, 100 mM NaCl, 0.05% TX100 containing antiprotease mix from Roche) and resuspended in IP1 buffer (20 mM Hepes pH 7.5, 100 mM NaCl, 20% glycerol, 0.05% TX100 plus antiprotease mix). After lysozyme treatment (0.2 mg/ml) and sonication, lysates were treated with DNaseI (30 U, Roche) and MgCl2 (10 mM) for 10 min on ice before centrifugation. Cleared lysate was mixed with Protein A-agarose beads (Roche) that had been previously incubated with α-GFP antibody (JL-8, Clontech) in IP1 buffer for 30 min at 4°C with gentle shaking followed by 2 washes with IP1 buffer. After 3 h of incubation (overnight for ParB IP), the beads were washed 6 times with IP1 buffer. The samples were subjected to SDS-PAGE and western blotting by using standard protocols. For protein detection, the following antibodies and concentrations were used: α-GFP, 1:10,000 for 1h; α-CckA:, 1:5,000 for 2h; α-DivJ, 1:10,000 for 2h; α-ParB, 1:4,000 for 1h; α-crescentin, 1:10,000 for 1h.

In vitro Co-IP

Protein A-agarose beads were incubated with α-ParB antibody in IP2 buffer (IP1 buffer without glycerol) at 4°C with gentle shaking. After 1h, purified His6-ParB (1 µg) was added and the samples were left shaking overnight. The beads were washed twice with IP2 buffer. BSA (0.25%) was added in IP2 buffer for 1h at RT to block nonspecific binding. Finally, His6-PopZ (1 µg) was added. After 5h of incubation at 4°C, the beads were washed 5 times with IP2 buffer. After SDS-PAGE, proteins were detected by immunoblotting using α-Tetra-His antibody (Qiagen) 1:1,500 for 1h.

Protein purification

IPTG (0.1 mM) was added to LB cultures of BL21/pET28a-PopZ or BL21/pET28a-ParB at OD600 = 0.6 for 3h at 37°C. Cells were harvested, resuspended in wash buffer (50 mM phosphate buffer pH 7.0 (ParB) or pH 8.0 (PopZ), 300 mM NaCl, antiprotease mix) containing DNaseI (30U) and MgCl2 (10 mM), and lysed by using a French press. His6-ParB and His6-PopZ were purified from the supernatant and pellet, respectively, by using Talon metal affinity resin (Clontech) and the standard batch/column protocol provided by the manufacturer. All steps were performed in either 50 mM phosphate buffer pH 8.0, 300 mM NaCl, 6M urea (His6-PopZ) or 50 mM phosphate buffer pH 7.0, 300 mM NaCl (His6-ParB). For washes, imidazole (20 mM for His6-PopZ and 15 mM for His6-ParB) was added whereas 200 mM imidazole was used for elution. Fractions were dialyzed against 5 mM Tris-HCl pH 8.5 (His6-PopZ) or 20 mM HEPES (pH 7.6), 0.1 mM EDTA, 12.5 mM MgCl2, 10% glycerol (His6-ParB).


Samples (pre-filtered through a 0.22 µm filter) were applied on a Superose 6, 10/30, HR SEC column (GE Healthcare) connected to HPLC Alliance 2965 (Waters Corp.) and equipped with an autosampler. Elution from SEC was monitored by a photodiode array (PDA) UV/VIS detector (996 PDA, Waters Corp.), differential refractometer (OPTI-Lab, or OPTI-rEx Wyatt Corp.), and static, multi-angle laser light scattering detector (DAWN-EOS, Wyatt Corp.). The SEC-UV/LS/RI system was equilibrated in 100 mM HEPES, pH 7.4, 150 mM NaCl, 1 mM DTT buffer at a flow rate of 0.3 ml/min. The Millennium software (Waters Corp.) controlled the HPLC operation and data collection from the multi-wavelength UV/VIS detector, while the ASTRA software (Wyatt Corp.) collected data from the refractive index detector, the light scattering detectors, and recorded the UV trace at 280 nm sent from the PDA detector. Data collection and analyses were carried out at the Keck Foundation Biotechnology Resource Laboratory, Yale University.

Supplementary Material













We thank M. Mooseker, E. Dufresne, T. Emonet and J. Wolenski for valuable discussions; K. Gerdes, J. Gober, A. Newton, N. Ohta, L. Shapiro, M. Thanbichler, and P. Viollier for supplying strains or antibody; P. Angelastro, A. Jackson, H. Lam and W. Schofield for construction of strains; H. J. Ding and D. Rosenman for computational help; M. Mooseker, Z. Jiang and G. Charbon for assistance with EM; E. Folta-Stogniew and the Keck Foundation Biotechnology Resource Laboratory at Yale for the biophysical analyses, and the Jacobs-Wagner laboratory and T. Emonet for critical reading of the manuscript. G.E. was supported by postdoctoral fellowships from the Villum Kann Rasmussen Foundation and the Danish Natural Science Research Council. This work was funded in part by National Institutes of Health (GM065835 to C.J.-W. and AI067548 to G.J.J.), gifts to Caltech from the Gordon and Betty Moore Foundation and Agouron Institute (to G.J.J.) and the Pew Scholars Program in the Biological Sciences sponsored by the Pew Charitable trust (to C.J.-W.).


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


  • Bardy SL, Maddock JR. Polar explorations Recent insights into the polarity of bacterial proteins. Curr Opin Microbiol. 2007;10:617–623. [PubMed]
  • Ben-Yehuda S, Rudner DZ, Losick R. RacA, a bacterial protein that anchors chromosomes to the cell poles. Science. 2003;299:532–536. [PubMed]
  • Biondi EG, Reisinger SJ, Skerker JM, Arif M, Perchuk BS, Ryan KR, Laub MT. Regulation of the bacterial cell cycle by an integrated genetic circuit. Nature. 2006;444:899–904. [PubMed]
  • Bohm J, Frangakis AS, Hegerl R, Nickell S, Typke D, Baumeister W. Toward detecting and identifying macromolecules in a cellular context: template matching applied to electron tomograms. Proc Natl Acad Sci U S A. 2000;97:14245–14250. [PubMed]
  • Briegel A, Dias DP, Li Z, Jensen RB, Frangakis AS, Jensen GJ. Multiple large filament bundles observed in Caulobacter crescentus by electron cryotomography. Mol Microbiol. 2006;62:5–14. [PubMed]
  • Dai M, Ziesman S, Ratcliffe T, Gill RT, Copley SD. Visualization of protoplast fusion and quantitation of recombination in fused protoplasts of auxotrophic strains of Escherichia coli. Metab Eng. 2005;7:45–52. [PubMed]
  • Ebersbach G, Gerdes K. Plasmid segregation mechanisms. Annu Rev Genet. 2005;39:453–479. [PubMed]
  • Ebersbach G, Jacobs-Wagner C. Exploration into the spatial and temporal mechanisms of bacterial polarity. Trends Microbiol. 2007;15:101–108. [PubMed]
  • Edwards DH, Thomaides HB, Errington J. Promiscuous targeting of Bacillus subtilis cell division protein DivIVA to division sites in Escherichia coli and fission yeast. EMBO J. 2000;19:2719–2727. [PubMed]
  • Ely B. Genetics of Caulobacter crescentus. Methods Enzymol. 1991;204:372–384. [PubMed]
  • Evinger M, Agabian N. Envelope-associated nucleoid from Caulobacter crescentus stalked and swarmer cells. J Bacteriol. 1977;132:294–301. [PMC free article] [PubMed]
  • Folta-Stogniew E. Oligomeric states of proteins determined by size-exclusion chromatography coupled with light scattering, absorbance, and refractive index detectors. Methods Mol Biol. 2006;328:97–112. [PubMed]
  • Gitai Z, Dye NA, Reisenauer A, Wachi M, Shapiro L. MreB actin-mediated segregation of a specific region of a bacterial chromosome. Cell. 2005;120:329–341. [PubMed]
  • Goley ED, Iniesta AA, Shapiro L. Cell cycle regulation in Caulobacter: location, location, location. J Cell Sci. 2007;120:3501–3507. [PubMed]
  • Hallez R, Bellefontaine AF, Letesson JJ, De Bolle X. Morphological and functional asymmetry in alpha-proteobacteria. Trends Microbiol. 2004;12:361–365. [PubMed]
  • Hamoen LW, Errington J. Polar targeting of DivIVA in Bacillus subtilis is not directly dependent on FtsZ or PBP 2B. J Bacteriol. 2003;185:693–697. [PMC free article] [PubMed]
  • Harry EJ, Lewis PJ. Early targeting of Min proteins to the cell poles in germinated spores of Bacillus subtilis: evidence for division apparatus-independent recruitment of Min proteins to the division site. Mol Microbiol. 2003;47:37–48. [PubMed]
  • Howard M. A mechanism for polar protein localization in bacteria. J Mol Biol. 2004;335:655–663. [PubMed]
  • Huang KC, Mukhopadhyay R, Wingreen NS. A curvature-mediated mechanism for localization of lipids to bacterial poles. PLoS Comput Biol. 2006;2:e151. [PubMed]
  • Huitema E, Pritchard S, Matteson D, Radhakrishnan SK, Viollier PH. Bacterial birth scar proteins mark future flagellum assembly site. Cell. 2006;124:1025–1037. [PubMed]
  • Iwai N, Nagai K, Wachi M. Novel S-benzylisothiourea compound that induces spherical cells in Escherichia coli probably by acting on a rod-shape-determining protein(s) other than penicillin-binding protein 2. Biosci Biotechnol Biochem. 2002;66:2658–2662. [PubMed]
  • Jacobs C, Domian IJ, Maddock JR, Shapiro L. Cell cycle-dependent polar localization of an essential bacterial histidine kinase that controls DNA replication and cell division. Cell. 1999;97:111–120. [PubMed]
  • Jensen RB, Shapiro L. The Caulobacter crescentus smc gene is required for cell cycle progression and chromosome segregation. Proc Natl Acad Sci U S A. 1999;96:10661–10666. [PubMed]
  • Lam H, Schofield WB, Jacobs-Wagner C. A landmark protein essential for establishing and perpetuating the polarity of a bacterial cell. Cell. 2006;124:1011–1023. [PubMed]
  • Lawler ML, Brun YV. Advantages and mechanisms of polarity and cell shape determination in Caulobacter crescentus. Curr Opin Microbiol. 2007;10:630–637. [PMC free article] [PubMed]
  • Lutkenhaus J. Assembly dynamics of the bacterial MinCDE system and spatial regulation of the Z ring. Annu Rev Biochem. 2007;76:539–562. [PubMed]
  • Marston AL, Errington J. Selection of the midcell division site in Bacillus subtilis through MinD-dependent polar localization and activation of MinC. Mol Microbiol. 1999;33:84–96. [PubMed]
  • Mastronarde DN. Dual-axis tomography: an approach with alignment methods that preserve resolution. J Struct Biol. 1997;120:343–352. [PubMed]
  • Matroule JY, Lam H, Burnette DT, Jacobs-Wagner C. Cytokinesis monitoring during development; rapid pole-to-pole shuttling of a signaling protein by localized kinase and phosphatase in Caulobacter. Cell. 2004;118:579–590. [PubMed]
  • Mohl DA, Gober JW. Cell cycle-dependent polar localization of chromosome partitioning proteins in Caulobacter crescentus. Cell. 1997;88:675–684. [PubMed]
  • Ortiz JO, Forster F, Kurner J, Linaroudis AA, Baumeister W. Mapping 70S ribosomes in intact cells by cryoelectron tomography and pattern recognition. J Struct Biol. 2006;156:334–341. [PubMed]
  • Sommer JM, Newton A. Pseudoreversion analysis indicates a direct role of cell division genes in polar morphogenesis and differentiation in Caulobacter crescentus. Genetics. 1991;129:623–630. [PubMed]
  • Stahlberg H, Kutejova E, Muchova K, Gregorini M, Lustig A, Muller SA, Olivieri V, Engel A, Wilkinson AJ, Barak I. Oligomeric structure of the Bacillus subtilis cell division protein DivIVA determined by transmission electron microscopy. Mol Microbiol. 2004;52:1281–1290. [PubMed]
  • Tamaki S, Matsuzawa H, Matsuhashi M. Cluster of mrdA and mrdB genes responsible for the rod shape and mecillinam sensitivity of Escherichia coli. J Bacteriol. 1980;141:52–57. [PMC free article] [PubMed]
  • Thanbichler M, Shapiro L. Chromosome organization and segregation in bacteria. J Struct Biol. 2006a;156:292–303. [PubMed]
  • Thanbichler M, Shapiro L. MipZ, a spatial regulator coordinating chromosome segregation with cell division in Caulobacter. Cell. 2006b;126:147–162. [PubMed]
  • Thomaides HB, Freeman M, El Karoui M, Errington J. Division site selection protein DivIVA of Bacillus subtilis has a second distinct function in chromosome segregation during sporulation. Genes Dev. 2001;15:1662–1673. [PubMed]
  • Viollier PH, Thanbichler M, McGrath PT, West L, Meewan M, McAdams HH, Shapiro L. Rapid and sequential movement of individual chromosomal loci to specific subcellular locations during bacterial DNA replication. Proc Natl Acad Sci U S A. 2004;101:9257–9262. [PubMed]
  • Ward D, Newton A. Requirement of topoisomerase IV parC and parE genes for cell cycle progression and developmental regulation in Caulobacter crescentus. Mol Microbiol. 1997;26:897–910. [PubMed]
  • Wheeler RT, Shapiro L. Differential localization of two histidine kinases controlling bacterial cell differentiation. Mol Cell. 1999;4:683–694. [PubMed]
  • Wu LJ, Errington J. RacA and the Soj-Spo0J system combine to effect polar chromosome segregation in sporulating Bacillus subtilis. Mol Microbiol. 2003;49:1463–1475. [PubMed]
  • Zheng SQ, Keszthelyi B, Branlund E, Lyle JM, Braunfeld MB, Sedat JW, Agard DA. UCSF tomography: an integrated software suite for real-time electron microscopic tomographic data collection, alignment, and reconstruction. J Struct Biol. 2007;157:138–147. [PubMed]