Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2011 Sep 24.
Published in final edited form as: Mol Cell. 2010 Sep 24;39(6):975–987. doi: 10.1016/j.molcel.2010.08.027

Imaging-based identification of a critical regulator of FtsZ protofilament curvature in Caulobacter

Erin D Goley 1, Natalie A Dye 1, John N Werner 2, Zemer Gitai 2, Lucy Shapiro 1,*
PMCID: PMC2945607  NIHMSID: NIHMS232221  PMID: 20864042

SUMMARY

FtsZ is an essential bacterial GTPase that polymerizes at midcell, recruits the division machinery, and may generate constrictive forces necessary for cytokinesis. However, many of the mechanistic details underlying these functions are unknown. We sought to identify FtsZ-binding proteins that influence FtsZ function in Caulobacter crescentus. Here, we present a microscopy-based screen through which we discovered two FtsZ-binding proteins, FzlA and FzlC. FzlA is conserved in α-proteobacteria and was found to be functionally critical for cell division in Caulobacter. FzlA altered FtsZ structure both in vivo and in vitro, forming stable higher order structures that were resistant to depolymerization by MipZ, a spatial determinant of FtsZ assembly. Electron microscopy revealed that FzlA organizes FtsZ protofilaments into striking helical bundles. The degree of curvature induced by FzlA depended on the nucleotide bound to FtsZ. Induction of FtsZ curvature by FzlA carries implications for regulating FtsZ function by modulating its superstructure.

HIGHLIGHTS.

  • An imaging screen identifies uncharacterized FtsZ-binding proteins

  • FzlA is critical for Caulobacter cell division and alters FtsZ structure in vivo

  • Nucleotide-dependent helical bundles are formed by FtsZ and FzlA in vitro

  • Induction of FtsZ curvature by FzlA represents a unique mode of FtsZ regulation

INTRODUCTION

In nearly all bacteria, cell division is mediated by the divisome, an assembly of proteins that functions to constrict the cell envelope. The divisome is centered on the tubulin-like GTPase, FtsZ, which polymerizes into a ring (the Z-ring) that serves as a scaffold for recruitment of the rest of the division proteins (Adams and Errington, 2009). In addition to its structural role, in vitro reconstitution experiments suggest that membrane-anchored FtsZ may directly generate forces that constrict the inner membrane (Osawa et al., 2008).

Despite a decades-long appreciation for the central role of FtsZ in cell division, the molecular mechanisms whereby it achieves its function are unclear. FtsZ is highly dynamic in vivo, turning over on the order of seconds (Anderson et al., 2004; Stricker et al., 2002). In vitro it assembles into a variety of structures, from linear protofilaments to bundles, sheets, rings, or tubes depending on the experimental conditions (Bramhill and Thompson, 1994; Erickson et al., 1996; Lowe and Amos, 1999; Mukherjee and Lutkenhaus, 1994; Popp et al., 2009). However, it is not known which structures are physiologically relevant. Modeling based on cell biological and biochemical observations (Erickson, 2009; Osawa et al., 2008) and cryoelectron tomogram imaging of FtsZ in cells (Li et al., 2007) indicate that the Z-ring is probably composed of a series of short, overlapping protofilaments. Several hypotheses for how FtsZ might generate constrictive force have been put forth (Erickson, 2009), including those that invoke lateral interactions between filaments (Ghosh and Sain, 2008; Horger et al., 2008; Lan et al., 2009) and those that rely on bending of filaments (Allard and Cytrynbaum, 2009; Erickson, 1997; Ghosh and Sain, 2008; Li et al., 2007). However, these models remain speculative, and how FtsZ-mediated force generation is implemented and regulated in the cell is a major outstanding question.

Similar to actin and tubulin, each of which has hundreds of interacting proteins, factors that bind directly to FtsZ are central to carrying out its function in vivo. To date only twelve proteins have been demonstrated to interact directly with FtsZ in Escherichia coli and Bacillus subtilis, the best-studied Gram-negative and Gram-positive organisms, respectively (Corbin et al., 2007; Ebersbach et al., 2008; Graumann, 2007). Although key mechanisms of FtsZ function are likely to be broadly conserved, many of these FtsZ regulators are not. Moreover, known FtsZ regulatory proteins display limited functional diversity, falling into three basic categories: those that link FtsZ to the inner membrane, those that stabilize filaments by promoting lateral interactions, and those that destabilize filaments by promoting FtsZ GTPase activity or inhibiting lateral interactions. Notably, no factors have been identified that specifically influence FtsZ filament curvature, which might be critical if filament bending is relevant to the molecular basis of FtsZ function. Based on this lack of functional diversity among known FtsZ regulators, we sought to develop a method to identify additional FtsZ-binding proteins in Caulobacter crescentus (referred to hereafter as Caulobacter), an organism characterized by its asymmetric cell division and cellular polarity. Here we describe an imaging screen that led to identification of FzlA, a functionally critical FtsZ-binding protein. Surprisingly, biochemical studies demonstrate that FzlA induces formation of unique helical bundles of FtsZ, suggesting that FzlA acts as a cellular regulator of filament bending.

RESULTS

An in vivo microscopy-based assay for FtsZ-binding proteins

To screen for undiscovered FtsZ-binding proteins in Caulobacter, we developed an assay for this activity based on a previously reported overexpression phenotype. E. coli FtsZ-G105S is a temperature-sensitive mutant with slower subunit turnover in vivo (Stricker et al., 2002) and reduced GTPase activity in vitro (de Boer et al., 1992; RayChaudhuri and Park, 1992) as compared to wild type. Caulobacter cells overexpressing the analogous ftsZ mutant (ftsZ-G109S, referred to as ftsZ*) adopted a distinct morphology, wherein extended constrictions separated the cell bodies (Fig 1A and (Wang et al., 2001)). FtsZ localized exclusively to these constrictions (Fig 1C and (Wang et al., 2001)) and electron cryotomography of similar cells revealed abundant FtsZ filaments lining the inner membrane of the constrictions (Li et al., 2007).

Figure 1. A microscopy-based assay distinguishes division proteins that bind to FtsZ from those that do not.

Figure 1

Open in a new tab

A) DIC images of Caulobacter bearing a plasmid with PxylX driving expression of ftsZ* grown for 0 h or 1.5 h. B) Localization of indicated fluorescent fusions to Caulobacter cell division proteins in otherwise wild type cells. Scale bar in B = 2 μm and applies to B-E. C and D) Localization of the indicated fluorescent fusions in cells overexpressing ftsZ*. E) Localization of Venus-FtsX or Venus-FtsW in cells overexpressing only ftsZ* (left) or ftsE and ftsZ* (right). F) Localization of the mCherry-MipZ in cells overexpressing ftsZ*. In B-F, fluorescence (red) overlaid on phase contrast (PC) (top) and fluorescence (bottom) are shown. Scale bar = 2 μm.

The presence of stable filaments in the constrictions of ftsZ* overexpressing cells raised the possibility that proteins that bind to FtsZ might be recruited there. To test this, we generated strains bearing inducible fluorescent fusions to Caulobacter divisome proteins integrated at the xylX chromosomal locus. We visualized proteins that have been demonstrated in Caulobacter and/or other organisms to interact directly with FtsZ (FtsA, ZapA, and FtsE) (Corbin et al., 2007; Din et al., 1998; Gueiros-Filho and Losick, 2002; Ma and Margolin, 1999; Wang et al., 1997), as well as those that are recruited to the division site downstream of these FtsZ-binding proteins (FtsX, FtsQ, FtsW, FtsI) (Dajkovic and Lutkenhaus, 2006; Harry et al., 2006). In the presence of xylose inducer in wild type cells, fluorescent fusions to FtsZ, FtsA, FtsE, FtsQ, and FtsI formed a band or focus at the division site (Fig 1B and (Costa et al., 2008; Moll and Thanbichler, 2009)). Identical midcell localization patterns were observed for fluorescent fusions to ZapA, FtsX, and FtsW (data not shown). When visualized in cells overexpressing ftsZ*, however, the fusions fell into two distinct localization patterns. The FtsZ-binding proteins, FtsA, FtsE, and ZapA, were significantly enriched in the constrictions (Fig 1C). In contrast, FtsQ, FtsI (Fig 1D), FtsX, and FtsW (Fig 1E), which do not interact directly with FtsZ, were observed throughout the cell.

We reasoned that cell division proteins that do not bind to FtsZ might fail to localize due to disrupted stoichiometry of normal protein-protein interactions in the divisome. Evidence from E. coli and B. subtilis suggests that a complex set of interactions is required to recruit these downstream divisome proteins (Harry et al., 2006). In strains used for our assay, ftsZ* is highly overexpressed from a high copy number plasmid upon induction with xylose. The gene encoding the target protein is mildly overexpressed, as it is present in two copies on the chromosome: one is untagged at the native locus and the other encodes the xylose-inducible fluorescent fusion at the xylX locus. Together this leads to a stoichiometric imbalance between FtsZ*, the fusion protein, and the factors required for localization of the fusion protein. We tested this idea by comparing the localization of Venus-FtsX in cells overexpressing ftsZ* alone to its localization in cells overexpressing both ftsE and ftsZ*. FtsX forms a heterodimer with FtsE (de Leeuw et al., 1999), but has not been shown to bind to FtsZ directly. When only FtsZ* was overproduced, Venus-FtsX was dispersed throughout the cell (Fig 1E). However, in cells overproducing both FtsE and FtsZ*, Venus-FtsX was robustly localized in the constrictions. In contrast, Venus-FtsW, which has not been demonstrated to bind to either FtsE or FtsZ, was diffuse under both conditions (Fig 1E). Thus, proteins that do not bind FtsZ are only efficiently recruited to the constrictions when the factors downstream of FtsZ that are required for their localization are present in sufficient quantities. MipZ, a negative regulator of FtsZ assembly in Caulobacter, localized to the poles in wild type cells (Thanbichler and Shapiro, 2006) and continued to exhibit this localization pattern upon overexpression of ftsZ* even though it binds to FtsZ directly (Fig 1F). Therefore, FtsZ-binding proteins that do not co-localize with FtsZ at the division site may not be recruited to the constrictions in this assay. We conclude that observing the localization of xylose-induced fluorescent fusions in cells overexpressing ftsZ* distinguishes division site-localized proteins that interact with FtsZ directly from those that do not.

Screen for FtsZ-binding proteins

Having established ftsZ* overexpression as a method for identifying proteins that bind to FtsZ, we used it to discover additional proteins with this activity. We did this by interrogating a library of strains bearing xylose-inducible mCherry fusions to 289 different Caulobacter ORFs (Werner et al., 2009). The strains in this collection were selected from a library containing mCherry fusions to 2786 Caulobacter ORFs (74% of those annotated in the genome) as those exhibiting distinct, non-diffuse localization, and are hereafter referred to as the localized library. We introduced the ftsZ* overexpression construct into each strain in the localized library by conjugation (Werner et al., 2009) and cultured the resulting strains in 96-well format. We then imaged each strain after inducing expression of both ftsZ* and the fluorescent fusion in question and manually classified the localization pattern of each fusion protein.

Proteins that localized to midcell in wild type cells (Werner et al., 2009) fell into several localization patterns in FtsZ*-overproducing cells. Some were diffuse, similar to the localization observed for divisome proteins that do not bind FtsZ (Fig 1D), but many others localized to foci in the cell bodies or the constrictions, reflecting “midcell” localization distinct from the divisome (Fig S1A). Notably, however, we identified four fusion proteins that were targeted to the constriction site, as anticipated for FtsZ-binding proteins (class I proteins) (Fig 2A). Surprisingly, a second set of five fusion proteins was specifically excluded from the site of FtsZ localization (class II proteins) (Fig 2B).

Figure 2. FzlA and FzlC are uncharacterized FtsZ binding proteins.

Figure 2

Open in a new tab

A) Localization of mCherry fusions to class I candidate proteins in cells overexpressing ftsZ*. Fluorescence (red) overlaid on PC (top) and fluorescence (bottom) are shown. Scale bar = 2 μm. B) Localization of mCherry fusions to class II proteins in cells overexpressing ftsZ*. See also Fig. S1. C) Co-localization of mCherry fusions to FzlA and FzlC (red) with FtsZ-CFP (green) in wild type cells. In the merged image, fluorescence is overlaid on phase contrast. Scale bar = 2 μm. D) Localization of mCherry fusions to FzlA and FzlC in cells depleted of FtsZ (-FtsZ) and after FtsZ depletion followed by repletion of FtsZ for 1 h (FtsZ repletion). Fluorescence and PC images are shown. Scale bar = 2 μm. E) Coomassie-stained SDS PAGE of supernatant (right) and pellet (left) from co-sedimentation reactions containing 3 μM of the candidate protein ± 3 μM FtsZ.

Although we set out to identify FtsZ-binding proteins in our screen, the class II proteins stood out, as they were clearly excluded from the extended constrictions in FtsZ* overproducing cells. These include three subunits of the F1F0 ATP synthase (AtpA/CC3449, AtpB/CC0368, and AtpD/CC3447), and two uncharacterized proteins that we named FzeA (CC1334, putative electrotransfer ubiquinone oxidoreductase) and FzeB (CC3441, putative lipase/esterase) for FtsZ-excluded proteins. Each of these proteins is predicted to be an integral membrane protein or to associate with such a protein, except FzeB, which is predicted to be periplasmic (Gardy et al., 2005). To determine if exclusion of these factors from the site of FtsZ assembly occurs under normal conditions, we imaged the class II mCherry fusion proteins in wild type cells also expressing ftsZ-cfp. In each case, we detected a dip in fluorescence intensity of the class II protein fusion at the site of peak FtsZ-CFP intensity, indicating that they are absent from the division site (Fig S1B). A control inner membrane protein, encoded by CC2207, was not excluded from the site of FtsZ assembly in wild type cells or in cells overexpressing ftsZ*, indicating that exclusion is specific to a subset of proteins associated with the envelope (Fig S1B,C). We hypothesize that spatially restricted lipid domains lead to accumulation of these factors away from the division site (Matsumoto et al., 2006). Interestingly, ATP synthase was reported to be absent from the septum in B. subtilis, indicating that this is a broadly conserved phenomenon (Meredith et al., 2008). We focused the remainder of our study on the class I candidate FtsZ-binding proteins from our screen.

Candidate FtsZ-binding proteins are well conserved and co-localize with FtsZ in wild type cells

Two of the class I candidates, MurG and MreB, were previously shown to co-localize with and to be functionally related to FtsZ. MurG, a glycosyltransferase, is essential for cell wall synthesis and localizes to the division site in an FtsZ-dependent fashion (Aaron et al., 2007). MreB, an actin homolog, forms a band at midcell in stalked cells and is important for cell shape maintenance and cell polarity (Figge et al., 2004; Gitai et al., 2004). The remaining class I hits are uncharacterized proteins that we named FzlA (encoded by CC3639) and FzlC (CC0100) for FtsZ-localized protein. FzlA is a putative glutathione S-transferase (GST) family member, and FzlC is a hypothetical protein with limited sequence similarity to heparinase II/III family proteins. Both are predicted to be cytoplasmic (Gardy et al., 2005) and BLAST searching indicates that they are widely conserved in α-proteobacteria. FzlA is absent from only a few isolated species, while FzlC is absent from the genus Rickettsia, members of which have a reduced genome.

To obtain further evidence that these factors might bind to FtsZ, we assessed their localization in wild type cells. Fluorescent fusions of both FzlA and FzlC were observed in foci that overlapped with FtsZ-CFP at the incipient division site (Fig 2C), consistent with the possibility that they bind directly to this protein. If this is true, they should require FtsZ for localization. To test this, we imaged fluorescent fusions to each protein in cells in which expression of the only functional copy of ftsZ was under the control of the PxylX promoter, allowing depletion of FtsZ in the absence of xylose (Wang et al., 2001). In the case of FzlA, the mCherry fusion was diffuse in filamentous cells depleted of FtsZ by growth without xylose, but formed midcell foci upon repletion of FtsZ for one hour (Fig 2D), indicating a dependence on FtsZ for division site localization. mCherry-FzlC was also dependent upon FtsZ for midcell localization, but it was observed in a patchy and/or polar localization pattern in the absence of FtsZ (Fig 2D).

FzlA and FzlC bind to FtsZ in vitro

The preceding results are consistent with the possibility that FzlA and FzlC bind to FtsZ. To test this directly, we turned to an in vitro method for assessing protein-protein interactions using purified components. In the presence of GTP and Mg2+, recombinant Caulobacter FtsZ (CcFtsZ) polymerizes into filaments that can be recovered in the pellet after high-speed centrifugation (Fig 2E and (Thanbichler and Shapiro, 2006)). We asked if FzlA or FzlC were able to co-sediment with FtsZ filaments under similar conditions, as this would indicate a direct interaction with polymeric FtsZ. Indeed, we found that purified recombinant His6-FzlA and His6-FzlC were each enriched in the pellet and depleted from the supernatant after incubation with FtsZ under polymerizing conditions (Fig 2E). Co-pelleting of these proteins was specific, as they were found in the supernatant when FtsZ, GTP, or MgCl2 was omitted. Moreover, the GST control did not co-sediment with FtsZ under any conditions tested. These data demonstrate that FzlA and FzlC bind directly to FtsZ filaments, the form of FtsZ found at the division site.

FzlA is essential for efficient cell division

The fact that FzlA and FzlC interact directly with FtsZ suggested that they might function during cell division. To address their cellular roles, we attempted to generate in-frame deletions of the genes encoding FzlA and FzlC. ΔfzlC cells grew with normal rates and morphology, indicating that this gene is dispensable under laboratory conditions (Fig S2A)). However, we were unable to delete fzlA unless a complementing copy of the gene was expressed, indicating that it is essential. We therefore generated a depletion strain (EG312) in which expression of the only functional copy of fzlA is under the control of the xylose-inducible PxylX promoter (Fig 3A). By 15 hours of growth without inducer in PYE, FzlA protein was no longer detectable (Fig 3B) and cells were elongated (Fig 3C). At longer depletion times, cells grew into smooth filaments, implicating FzlA in the earliest stages of envelope invagination during division. When growth in PYE was monitored by an increase in optical density (OD), cells depleted of FzlA stopped growing after approximately 24 hours (Fig S2B). Surprisingly, the FzlA depletion strain was able to grow without inducer in minimal M2G media. However, depletion of FzlA caused a severe reduction in doubling time from ~6 hours in the presence of xylose to >15 hours in the presence glucose (Fig S2B). Additionally, ~20-fold overproduction of FzlA caused mild filamentation (Fig 4B) and inhibited growth of Caulobacter (Fig S2C, D). Because FzlA was found to be critical for cell division, we focused the remainder of this study on FzlA and its interaction with FtsZ.

Figure 3. FzlA is critical for cell division and is cell cycle regulated.

Figure 3

Open in a new tab

A) Genomic context of fzlA (WT) and construction of a xylose-dependent FzlA depletion strain (EG312). B) Immunoblots of cell lysates from wild type (WT) cells or cells from strain EG312 grown in the absence of xylose for the indicated times (hours). Equivalent total OD units of lysates were probed with antibodies recognizing the indicated proteins. C) Phase contrast images of cells of strain EG312 grown in the absence of xylose (FzlA depleted) for the indicated times. Scale bar = 2 μm. D) PC and mCherry-FzlA fluorescence images showing FzlA localization over the cell cycle. Synchronized swarmer cells were placed on M2G-agarose pads and imaged every 30 min. E) Normalized abundance of ftsZ and fzlA transcript levels over the cell cycle. F) Immunoblots using antibodies against the indicated proteins of cell lysates from synchronized wild type cells grown in M2G (top) or PYE (bottom). See also Fig. S2.

Figure 4. FzlA affects FtsZ structure in vivo and promotes formation of higher order FtsZ structures in vitro.

Figure 4

Open in a new tab

A) Localization of FtsZ-CFP in cells from FzlA depletion strain EG312 grown in xylose (fzlA induced) or glucose (FzlA depleted) for 24 h. Arrows indicate morphologically normal FtsZ rings. B) Localization of FtsZ-CFP in Caulobacter bearing a plasmid with PxylX driving overexpression of fzlA. Cells were grown for 23 h in glucose (uninduced) or xylose (FzlA overproduced). Arrows indicate FtsZ-CFP foci associated with the sides of the cell. Hatched arrows indicate multiple mislocalized FtsZ-CFP foci. In A-B, fluorescence (red) overlaid on phase contrast (PC) (left) and fluorescence (right) are shown. Scale bars = 2 μm. C) (Left) Coomassiestained SDS PAGE of pellet fractions after high speed centrifugation of FtsZ polymerization reactions containing the indicated concentrations of FzlA and 2 μM FtsZ. (Right) Bar graph depicting relative amount of FtsZ in the pellet at each FzlA concentration. Graph is normalized relative to FtsZ in the pellet with 0 μM FzlA and mean and SEM are shown (n = 3). D) Coomassie-stained SDS PAGE of pellet fractions after low speed centrifugation of FtsZ polymerization reactions containing 2 μM FtsZ ± 4 μM FzlA. E) Right angle light scattering over time of FtsZ in the presence of the indicated concentrations of FzlA. Arrow indicates addition of FzlA or buffer control and hatched arrow indicates addition GTP. F) Right angle light scattering reactions as in E, but containing 0.1 mM of the indicated nucleotide. 4 μM FzlA is included where indicated. Inset graph is an expanded view of the bottom of the graph at left. G) Inorganic phosphate (Pi) concentration in solution over time in the presence of 2 μM FtsZ, 2 mM GTP, and the indicated concentrations of FzlA. GTPase rates (Pi released per FtsZ molecule per min) for each reaction are indicated. See also Fig. S3.

The localization and abundance of FzlA are cell cycle regulated

Having established that FzlA is essential for efficient cell division, we asked at what point in the cell cycle it might exert its effects. The localization of mCherry-FzlA was followed over the course of the cell cycle in synchronized cells in which the native copy of fzlA was replaced by mCherry-fzlA. mCherry-FzlA was diffuse in newborn swarmer cells but appeared in a band at midcell early in the cell cycle (Fig 3D), concomitant with midcell assembly of FtsZ (data not shown). It remained at the division site until just before completion of cell division, when it became diffuse (t = 120 min, Fig 3D). Although mCherry-FzlA localized to the new cell pole in swarmer cells that expressed vanillate-induced ftsZ-cfp in addition to untagged native ftsZ (Fig 2C, S2E), it was diffuse in swarmer cells that were not expressing additional ftsZ. Therefore, in the presence of native levels of FtsZ, FzlA disappears from the division site before the completion of division and is not recruited to the focus of FtsZ at the new pole. This suggests the existence of a regulatory mechanism to displace FzlA from the division site late in the cell cycle and prevent it from associating with FtsZ at the cell pole.

FzlA was also found to be cell cycle regulated at the transcript and protein levels. Analysis of Affymetrix data from synchronously growing cells (McGrath et al., 2007) indicated that fzlA mRNA levels peaked in pre-divisional cells, similar to those encoding other cell division proteins such as ftsA (Sackett et al., 1998) (Fig 3E). FzlA protein levels were highest in swarmer and early stalked cells when grown in minimal M2G media, and remained relatively constant over the cell cycle when cells were grown in PYE media (Fig 3F). We quantified the number of molecules of FzlA and FtsZ at 0 min and 70 min post-synchrony in cells grown in PYE (swarmer and late pre-divisional cells, respectively). FzlA was present at 262 ± 42 (mean ± SEM) molecules/cell in swarmers (~1.1 μM), and 404 ± 74 molecules/cell in pre-divisional cells (~1.7 μM), whereas FtsZ was present at 193 ± 79 molecules/cell in swarmers (~0.8 μM), and 2509 ± 688 molecules/cell in pre-divisional cells (~10.6 μM). These results indicate that, while FzlA is present at the division site from the time of Z-ring assembly until just before division occurs, it is at its highest levels relative to FtsZ early in the cell cycle, prior to and during assembly of the Z-ring.

FzlA affects FtsZ structure in vivo

Since FzlA directly interacts with FtsZ filaments, it might alter the structure or dynamics of FtsZ during division. To investigate this possibility, the localization of FtsZ-CFP was analyzed in cells depleted of or overproducing FzlA. In wild type Caulobacter, fluorescently tagged FtsZ localizes to a band at midcell that progressively constricts until it appears as a focus in deeply constricted cells (Thanbichler and Shapiro, 2006). Z-rings still formed in filamentous cells depleted of FzlA (Fig 4A) and although some faint dispersed foci of FtsZ-CFP were observed, many Z-rings appeared similar to those found in stalked cells prior to constriction. Similar rings were observed by immunofluorescence using α-FtsZ antibodies (Fig S3A). Moreover, the cell division proteins FtsB and FtsL were still recruited to midcell foci in filamentous cells depleted of FzlA (Fig S3B, C). Thus, although FzlA is critical for invagination of the envelope, it is not absolutely required for FtsZ to form a ring structure or recruit other divisome components.

In cells overproducing FzlA, FtsZ-CFP rarely localized to morphologically normal rings at midcell (Fig 4B). Instead, tight foci of FtsZ were observed that were associated with one side of the cell (Fig 4B, arrows). Cells frequently contained more than one bright FtsZ-CFP focus and these were not always located near midcell (Fig 4B, hatched arrows). Such aberrant structures were never observed in cells producing FzlA at wild type levels (Fig 1B, 4A, B). We quantified this effect by categorizing cells as having a midcell band of FtsZ, a midcell focus of FtsZ at a constriction, a polar focus of FtsZ, or non-midcell foci of FtsZ or other aberrant FtsZ localization. In the presence of glucose (no induction of fzlA), 52.5% of cells had a midcell band of FtsZ, 32.1% had a focus of FtsZ at a constriction, 14.8% had a polar focus, and 2.0% had non-midcell foci or otherwise aberrant FtsZ localization (n=196 cells). In the presence of xylose (FzlA overproduction), however, only 6.6% of cells had a midcell band of FtsZ, whereas 48.0% of cells had aberrant FtsZ localization. The number of cells with a midcell focus at a constriction (30.0%) or a polar focus (15.5%) was similar to what was observed with normal levels of FzlA (n = 257 cells). These results suggest that FzlA can affect FtsZ structure in vivo by binding to filaments and either altering their structure directly or, when overproduced, by interfering with other interactions required for formation of a normal Z-ring.

FzlA promotes formation of stable higher order FtsZ structures in vitro

Our findings thus far implicate FzlA in direct regulation of FtsZ structure and/or dynamics during cell division. We examined this possibility more directly using purified components in in vitro assays that allow us to assess FtsZ polymerization in the presence and absence of FzlA. First, we introduced FzlA into FtsZ pelleting reactions at increasing concentrations and found that as FzlA concentration increased, the proportion of FtsZ recovered in the pellet increased (Fig 4C). Inclusion of FzlA at a 1:2 molar ratio to FtsZ resulted in doubling the amount of FtsZ recovered in the pellet, and at 2:1 FzlA:FtsZ a ~4-fold increase in pelleted FtsZ was observed. The amount of FzlA that cosedimented with FtsZ also increased as FzlA concentration increased, although a small fraction of FzlA was found in the pellet in the absence of FtsZ at the highest concentration (Fig S3D). This indicates that FzlA promotes polymerization of FtsZ or alters FtsZ structures such that they are more readily recovered upon centrifugation. Evidence supporting this second possibility came from low speed pelleting of FtsZ in the absence and presence of FzlA (Fig 4D). On its own, very little FtsZ (7%) was found in the pellet after centrifugation at 16000 × g. In the presence of FzlA, however, 47% of FtsZ was recovered in the pellet, suggesting that FzlA drives formation of larger scale FtsZ structures.

We further explored the effect of FzlA on FtsZ polymers using right angle light scattering. As shown previously for FtsZ from other bacteria, we were able to monitor polymerization of CcFtsZ over time by measuring an increase in light scattering. Polymerization was rapid and was observed in the presence of GTP or the poorly hydrolysable analog GMP-CPP, but not in the presence of GDP (Fig S3E). Moreover, steady state polymer mass increased at increasing FtsZ concentrations (Fig S3F). Having established that light scattering is effective for assessing CcFtsZ polymerization, we used it to analyze the effect of FzlA on polymerization dynamics. We allowed FtsZ to reach a baseline light scattering value, then added FzlA (or buffer control) and again allowed light scattering to stabilize, then added GTP to induce polymerization (we also reversed the order of addition of GTP and FzlA with similar results, data not shown). Consistent with our pelleting results, when FzlA was present in polymerization reactions, a dose-dependent increase in light scattering was observed (Fig 4E). Under the conditions used, 4 μM FzlA (with 2 μM FtsZ) caused a >10-fold increase in light scattering at steady state. It is notable that, even prior to addition of GTP, FzlA induced a significant increase in light scattering in the presence of FtsZ (Fig 4E), indicating the formation of a complex of FtsZ and FzlA prior to polymerization. Although FzlA on its own induced a moderate amount of light scattering, this was not GTP dependent and was >10 fold lower than the light scattering observed at the same FzlA concentration in the presence of FtsZ (Fig S3G, H). Moreover, size exclusion chromatography of FzlA alone indicated that it is monomeric (Fig S3I). Collectively, these findings demonstrate that FzlA binds to FtsZ and induces formation of higher order structures.

The FtsZ binding proteins ZapA and SepF have been shown to bundle FtsZ protofilaments and, in so doing, stabilize filaments and decrease the GTPase rate of FtsZ (Gueiros-Filho and Losick, 2002; Singh et al., 2008). To address whether this is true of FzlA, polymerization and depolymerization of FtsZ was monitored in the presence of limiting concentrations of GTP in the presence and absence of FzlA using light scattering. Without FzlA, FtsZ polymerized and polymer mass remained relatively constant until the available pool of GTP was exhausted, at which point FtsZ depolymerized and light scattering returned to baseline (Fig 4F, inset). In the presence of FzlA, however, a significant drop in light scattering was not observed over the time scale tested, similar to the case of FtsZ alone in the presence of low concentrations of GMP-CPP (Fig 4F), suggesting that the FzlA/FtsZ structures are highly stable. We directly tested whether FzlA affects the GTPase activity of FtsZ using a malachite green assay to monitor inorganic phosphate (Pi) release over time. In the absence of FzlA, we measured a GTPase rate of 4.7 Pi released per minute per FtsZ molecule (Fig 4G), similar to the published value of 5.2 per min (Thanbichler and Shapiro, 2006). As predicted from the light scattering results, FzlA potently inhibited this activity in a dose-dependent manner, maximally reducing kcat nearly 4-fold to 1.3 per min (Fig 4G).

FzlA can compete with MipZ in vitro and in vivo

In vivo, Caulobacter FtsZ is directed to assemble at midcell early in the cell cycle by the MipZ ATPase. MipZ localizes in a gradient with its highest concentration at the cell poles and directly inhibits FtsZ polymerization by promoting its GTPase activity. This restricts FtsZ assembly to the site of lowest MipZ concentration, roughly midcell (Fig 5A and (Thanbichler and Shapiro, 2006)). Since FzlA stabilizes FtsZ in vitro, we asked whether FzlA might compete with the inhibitory activity of MipZ. To this end, we monitored pelleting of 2μM FtsZ with increasing concentrations of FzlA in the presence or absence of 2 μM MipZ. The presence of MipZ resulted in a ~45% reduction in the amount of FtsZ in the pellet (Fig 5B), as reported previously (Thanbichler and Shapiro, 2006). However, FzlA was able to cause a substantial increase (up to 4-fold) in the amount of FtsZ polymer in the pellet in both the absence and presence of MipZ (Fig 5B). Even when MipZ was present in 4-fold molar excess over FzlA (at 2 μM and 0.5 μM, respectively), FzlA restored the amount of FtsZ recovered in the pellet to the level observed in the absence of MipZ. Thus, FzlA can stabilize FtsZ polymers even in the presence of a molar excess of MipZ.

Figure 5. FzlA stabilizes FtsZ in vitro and in vivo.

Figure 5

Open in a new tab

A) Cartoon depiction of the localization of MipZ (yellow) and FtsZ (cyan) localization in swarmer cells (0 min post-synchrony) and stalked cells (25 min post-synchrony in PYE). B) (Bottom) Coomassiestained SDS PAGE of pellet fractions after high speed centrifugation of FtsZ polymerization reactions containing the indicated concentrations of FzlA and 2 μM FtsZ ± 2 μM MipZ. (Top) Bar graph depicting relative amount of FtsZ in the pellet at each FzlA concentration. Graph is normalized relative to FtsZ in the pellet with 0 μM FzlA and 0 μM MipZ, and mean and SEM are shown (n = 3). C) Bar graph depicting percentages of cells of EG496 and EG499 grown for 24 h with xylose with the indicated localization of FtsZ and MipZ at 25 min post-synchrony. Mean and SEM are shown. D) Representative images of cells from (B). Arrows indicate polar foci of FtsZ-CFP that overlap with MipZ-YFP. Bar = 2 μm.

To determine if FzlA can stabilize FtsZ polymers in vivo, we again looked at the relationship between FzlA, FtsZ, and MipZ. MipZ and FtsZ localize to opposite poles in swarmer cells, and upon bipolarization of MipZ, FtsZ is rapidly displaced from the cell pole by the inhibitory action of MipZ and re-assembles at midcell (Thanbichler and Shapiro, 2006). If FzlA acts to stabilize FtsZ polymers in vivo, we expect to observe delayed displacement of FtsZ from the cell pole in the presence of excess FzlA. To test this, we created a strain (EG496) bearing mipZ-yfp at the native locus and vanillate-inducible ftsZ-cfp at the vanA locus that allows monitoring of FtsZ and MipZ localization in the same cells. We generated a second strain (EG499) that is isogenic to EG496 but carries a xylose-inducible fzlA overexpression construct. We then grew both EG496 and EG499 in the presence of xylose for 24 h, isolated swarmer cells, and allowed them to grow synchronously in PYE plus xylose. As expected, MipZ and FtsZ were at opposite poles in swarmer cells (time 0) of both strains (data not shown) and at 25 minutes post-synchrony, MipZ-YFP was bipolar in the majority of cells from both strains (80% (n=1148 cells in 2 independent experiments) and 75% (n=972 cells in 2 experiments) for EG496 and EG499, respectively). At 25 minutes post-synchrony, we categorized cells as exhibiting polar FtsZ, midcell FtsZ, or both midcell and polar FtsZ, and quantified the number of cells in which FtsZ and MipZ co-localized (Fig 5C, D). In cells with wild type FzlA levels (EG496), 19.5% of cells exhibited polar FtsZ, 69.2% had midcell FtsZ, and 11.2% had polar and midcell FtsZ localization, with 12.1% of cells showing overlap between FtsZ-CFP and MipZ-YFP (n=1093 cells, 2 independent experiments). When fzlA was overexpressed, however, we observed substantial increases in the percentages of cells with polar FtsZ (33.3%), cells with midcell and polar FtsZ (34.1%), and cells exhibiting co-localization of MipZ and FtsZ (41.1%) (n=952). Thus, in the presence of excess FzlA, FtsZ is less dynamic in vivo and is not as readily displaced from the cell pole upon arrival of MipZ.

FzlA induces the formation of curved FtsZ bundles

Stabilization of FtsZ polymers is often coupled to bundling of protofilaments (Gueiros-Filho and Losick, 2002; Mukherjee and Lutkenhaus, 1999; Singh et al., 2008; Yu and Margolin, 1997). The products of FtsZ polymerization reactions in the presence and absence of FzlA were therefore examined by negative stain electron microscopy (EM) to determine if FzlA promotes FtsZ bundling. As shown previously (Thanbichler and Shapiro, 2006), FtsZ alone formed straight and gently curved single protofilaments in the presence of GTP (Fig 6A, left). In the presence of FzlA, however, striking arcs and helical bundles were observed in addition to protofilaments morphologically similar to those formed by FtsZ alone (Fig 6A, right). The helical bundles were right-handed and exhibited remarkably consistent geometry, with a diameter of 43.8 ± 2.7 nm, a helical pitch of 31.6 ± 1.9 nm, and a strand width of 13.3 ± 0.9 nm. The helical structures required the presence of both FtsZ and FzlA, as FzlA on its own did not form any observable higher order structures (Fig S4A, S5A). Moreover, the His6 tag on FzlA was not responsible for its activity, as FzlA with the tag cleaved off induced formation of identical helices (Fig S4B). Although normal filaments appeared to be fairly abundant, it is important to note that negative stain EM is not quantitative, and we often observed large clusters of helical bundles (Fig S4C). Examination of bundled structures at high magnification revealed that they are composed of three filaments (Fig 6A). We propose that the outside two filaments are FtsZ protofilaments and the third, inner filament is composed of FzlA.

Figure 6. Helical bundles of FtsZ form in the presence of FzlA.

Figure 6

Open in a new tab

A) Negative stain EM of 2 μM FtsZ under polymerizing conditions in the absence (left) or presence (right) of 4 μM FzlA at the indicated magnifications. Scale bars = 100 nm. B) Negative stain EM of 2 μM FtsZ with 4 μM FzlA applied to grids at the indicated times after addition of GTP. Scale bars = 100 nm. C) Negative stain EM of 2 μM FtsZ with the indicated concentrations of FzlA. Scale bars = 50 nm. See also Fig. S4.

In order to learn how FzlA promotes formation of curved FtsZ bundles, we analyzed FzlA/FtsZ structures by EM at increasing time intervals and varying concentrations of FzlA. Time-dependent analysis revealed the presence of short arcs at 30 sec, arcs and short helices at 3 min, and progressively longer helices at 8 and 16 min after induction of polymerization (Fig 6B). We did not observe any bundling of filaments in the background. The formation of bundled helices was also concentration dependent. Helices were most abundant at 2:1 molar ratio of FzlA to FtsZ, but were also readily formed at 1:1 FzlA to FtsZ. Formation of helices was inhibited at higher (5:1 or 10:1) FzlA concentrations (Fig 6C). This may be explained if at very high concentrations all FzlA binding sites on FtsZ filaments are occupied, preventing interaction of FzlA on one filament with a second filament.

Formation of FzlA/FtsZ helical bundles is nucleotide dependent

FtsZ has been reported to form highly curved filaments in the presence of GDP and polycations (Lu et al., 2000). To determine whether the FzlA-induced curvature of FtsZ is similarly regulated by nucleotide, we analyzed polymerization of FtsZ with FzlA in the presence of GTP, GDP, and GMP-CPP. Both GTP and GMP-CPP supported the ability of FzlA to induce high speed pelleting of FtsZ (Fig 7A) and to induce an increase in FtsZ light scattering (Fig 7B). Pelleting of FtsZ was only mildly increased by FzlA in the absence of nucleotide or in the presence of GDP (Fig 7A), and light scattering did not increase upon addition of GDP in the presence of FzlA and FtsZ beyond the initial FzlA-induced increase prior to addition of nucleotide (Fig 7B). To determine the nature of the structures formed in the presence of each nucleotide, we examined them by negative stain EM. FtsZ formed morphologically similar filaments in the presence of GTP and GMP-CPP, but no filaments were observed in the presence of GDP (Fig 7C, top). Upon addition of FzlA, all three nucleotides promoted formation of curved FtsZ structures (Fig 7C, bottom). These required the presence of FtsZ, since FzlA did not form any regular structures on its own regardless of the nucleotide present (Fig S5). Interestingly, the structures formed in the presence of FtsZ and FzlA with different nucleotides were morphologically distinct from one another. Whereas helical bundles formed in the presence of GTP, poorly stained arcs formed in the presence of GDP and large arcs and rings formed in the presence of GMP-CPP (Fig 7C, bottom). The diameters of the three structures were distinct, with FzlA/FtsZ-GDP arcs being the smallest (33.7 ± 4.3 nm), FzlA/FtsZ-GTP helices being intermediate in size (42.8 ± 2.8 nm), and FzlA/FtsZ-GMP-CPP arcs and rings being the largest (74.8 ± 8.6 nm) (Fig 7C). Extended helices were not observed in the presence of either GDP or GMP-CPP. Thus, the degree of FtsZ filament curvature induced by FzlA is nucleotide-dependent. Additionally, MgCl2 was required for the formation of helical bundles, although FzlA stabilized curved FtsZ arcs and circles in the absence of MgCl2 and the presence of GTP (Fig S5B). Collectively these and the previous data allow us to propose a speculative model wherein FzlA nucleates formation of paired helical bundles of FtsZ that elongate over time (Fig 7D). Although size exclusion chromatography indicates that FzlA is a monomer on its own (Fig S3I), our light scattering data suggests that binding of FzlA to FtsZ may induce multimerization prior to polymerization (Fig 4E, S3G,H). GTP hydrolysis is linked to degree of curvature and filament elongation, but the timing of hydrolysis and/or Pi release relative to formation of helical bundles is unknown.

Figure 7. The effect of FzlA on FtsZ polymerization is nucleotide dependent.

Figure 7

Open in a new tab

A) Coomassie-stained SDS PAGE of pellet fractions after high-speed centrifugation of FtsZ polymerization reactions containing 2 μM FtsZ, ± 4 μM FzlA and 0.2 mM of the indicated nucleotide. B) Right angle light scattering of FtsZ polymerization reactions ± 4 μM FzlA and the indicated nucleotides (GTP and GDP were used at 2 mM and GMP-CPP was used at 0.2 mM). C) Negative stain EM of FtsZ polymerization reactions containing 2 μM FtsZ, ± 4 μM FzlA and 0.2 mM of the indicated nucleotide. Mean diameter ± standard deviation is indicated for each. Scale bars = 50 nm. See also Fig. S5. D) Speculative model for formation of FzlA-FtsZ helical bundles in the presence of GTP. FzlA binds to monomers or short oligomers of FtsZ and promotes elongation of paired, curved FtsZ protofilaments with the indicated dimensions.

DISCUSSION

Here, we have described a microscopy-based assay for identification of Caulobacter FtsZ-binding proteins. We used this tool in combination with a Caulobacter localized protein library to discover two proteins with FtsZ-binding activity, as well as to identify a set of proteins specifically excluded from the division site. These findings validate our screening approach and demonstrate the utility of secondary screening of the Caulobacter localized library to identify proteins with a desired activity. Our results establish a framework to investigate the effects of these factors on FtsZ structure and function, as well as to dissect the pathways downstream to other divisome proteins that ultimately function in concert with FtsZ during cellular growth and division.

Of the hits from our screen, FzlA stood out owing to its functional importance for cell division. FzlA is highly conserved among α-proteobacteria and is one of nineteen members of the GST protein family in Caulobacter. Although canonical GSTs function as detoxifying enzymes, other proteins in this family have diverse activities including functioning as transcription factors and ion channels (Oakley, 2005). FzlA does not bind to glutathione (data not shown) and has apparently adapted this protein fold to a cytoskeletal purpose.

The ability of FzlA to promote the formation of highly curved FtsZ filaments implicates FzlA as a regulator of FtsZ filament bending. A number of FtsZ-binding proteins promote stabilization and/or lateral association of FtsZ protofilaments, and FzlA shares some common features with these factors. Like ZapA and SepF (Gueiros-Filho and Losick, 2002; Singh et al., 2008), FzlA promotes formation of bundled FtsZ structures, reduces the GTPase activity of FtsZ, and stabilizes filaments (Fig 4,6). However, the structures formed in the presence of FzlA are dramatically different from those formed with ZapA, SepF, or ZipA. Whereas the latter promote formation of straight, parallel bundles of variable width (Gueiros-Filho and Losick, 2002; Hale et al., 2000; RayChaudhuri, 1999; Singh et al., 2008), FzlA induces assembly of helical bundles with highly regular dimensions (Fig 6). This is not an anomaly of bundling of CcFtsZ, as ZapA from Caulobacter does not promote formation of helical bundles of CcFtsZ (Y.-C. Yeh, N.A.D., H. McAdams and L.S., unpublished). The only published example wherein an FtsZ-binding protein promotes a stable curved FtsZ conformation is that of B. subtilis ZapA in the absence of nucleotide, where it induces formation of ~25 nm mini-rings of FtsZ (Gueiros-Filho and Losick, 2002). However, it is unclear how physiologically relevant such activity might be, since filaments are unlikely to be free of nucleotide in the cell. Thus, FzlA represents a type of FtsZ-modulating protein that is biochemically and functionally distinct from those characterized previously.

The ability of cytoskeletal regulators to organize polymers into distinct superstructures is central to specifying the function of the polymer at a given time and place in the cell. Prior to the present study, observed higher order FtsZ filament structures induced by binding partners were limited to straight parallel bundles. However, the physiological relevance of the curvature-inducing activity of FzlA remains an open question. Given its relative abundance early in the cell cycle and its ability to stabilize FtsZ against the depolymerizing activity of MipZ (Fig 3,5), FzlA may be important for early establishment of the Z-ring. Alternatively, or perhaps in addition to a general stabilizing role, the specific form FtsZ adopts when bound to FzlA could dictate the function of FtsZ. In Caulobacter, the Z-ring initially promotes cell elongation via midcell-localized peptidoglycan sysnthesis before switching to invagination of the cell envelope at the nascent division site (Aaron et al., 2007). Since the FzlA:FtsZ ratio decreases around the time that FtsZ switches from promoting elongation to promoting division, it is possible that changing the fraction of FtsZ in helical FzlA-mediated bundles is relevant to regulating the elongation-to-division switch in FtsZ function.

FtsZ filament curvature has been linked mechanistically to force generation by FtsZ during cytokinesis, thus it is tempting to speculate that FzlA-mediated FtsZ curvature may be relevant to this process. The filament-bending force generation mechanism was originally posited to be regulated by GTP hydrolysis based on the early observation that FtsZ formed straight filaments in the presence of GTP, but highly curved filaments in the presence of GDP (Lu et al., 2000). However, it was shown more recently that FtsZ does not require GTP hydrolysis to deform membranes in vitro (Osawa et al., 2009). Filament bending by an FtsZ-binding protein like FzlA provides an attractive alternative mechanism for regulating when and where FtsZ will assume a curved conformation. This could occur, for example, via post-translational modification of FzlA or by cell cycle variation in the ratios of regulatory proteins to FtsZ and to each other. We have shown that the FzlA to FtsZ ratio changes dramatically as the cell cycle progresses (Fig 3F), as do levels of MipZ (Thanbichler and Shapiro, 2006) and FtsA (Martin et al., 2004). The stoichiometric balance between proteins with competing or complementary activities would therefore be expected to shift through the cell cycle leading to variations in FtsZ structure, dynamics, and therefore, function.

The characterization of the full complement of FtsZ regulatory proteins in a given organism is critical to developing a detailed mechanistic model for FtsZ function in the cell. In this study, we have made progress toward this goal by identifying two FtsZ-binding proteins in Caulobacter and discovering the unique curvature-inducing activity associated with FzlA. Future work will address the structural basis of this activity, as well as the question of how FzlA-induced FtsZ curvature relates to the in vivo function of FtsZ. Ultimately such approaches will enable understanding the complex interplay between the growing number of factors that manipulate FtsZ dynamics and structure in the cell, and how this is relayed downstream to the full divisome during cell division.

EXPERIMENTAL PROCEDURES

Bacterial strains, plasmids, and growth conditions

Detailed descriptions of bacterial strains and growth conditions are included in the Supplemental Information. Strains and plasmids are listed in Supplemental Table S1.

Library construction

pEG106 was introduced into the localized library by high throughput conjugation and strains were isolated and cultured as described previously (Werner et al., 2009).

Fluorescence microscopy

Cells were imaged during log phase of growth after immobilization on 1% agarose pads. High throughput screening was performed as described previously (Werner et al., 2009). DIC imaging of ftsZ*-expressing cells was done on a Nikon Eclipse E800 microscope equipped with a Nikon Plan Apo 100×/1.40 DIC objective, and a 5 MHz MicroMAX cooled CCD camera (Princeton Instruments). All other imaging was done on a Leica DM 6000 B microscope using a HCX PL APO 100×/1.40 Oil PH3 CS objective, Hamamatsu EM-CCD C9100 camera, and KAMS software (Christen et al., 2010). Images were processed with Metamorph 4.5 (Universal Imaging Group) and/or Adobe Photoshop.

Protein purification

FtsZ, His6-FzlA, His6-FzlC, MipZ-His6, and GST were overproduced in E.coli. Untagged FtsZ was purified essentially as described (Thanbichler and Shapiro, 2006) by ammonium sulfate precipitation, anion exchange chromatography, and gel filtration. His6-FzlA, His6-FzlC, and MipZ-His6, were purified by Ni-affinity chromatography using Ni-NTA agarose (QIAGEN). His6-FzlC and His6-MipZ were further purified by cation exchange chromatography and gel filtration, respectively. GST was purified by glutathione affinity chromatography (Glutathione Sepharose 4B, GE Healthcare). All proteins were frozen in liquid nitrogen and stored at -80°C.

Antibody production and purification

Purified His6-FzlA was used to immunize a rabbit for antibody production (Josman, LLC). To affinity purify anti-FzlA antibodies, His6-FzlA was coupled to Affi-gel 10 resin (Bio-Rad) following the manufacturer’s protocol. Affinity purification of anti-FzlA antibodies was performed using standard protocols and antibodies were eluted with 0.2 M glycine [pH 2.5] and 150 mM NaCl, dialyzed into Tris-buffered saline, and diluted 1:1 in glycerol before storage at -20°C. Anti-FzlA antibodies recognized a major band in Caulobacter lysates that runs at ~ 30 kDa by SDS-PAGE.

Immunoblotting and quantitation of protein levels

Cells were harvested in log phase of growth and lysed in SDS-PAGE loading buffer by boiling for 5 minutes. SDS-PAGE and transfer of protein to PVDF membrane was performed using standard procedures. Samples were probed with α-FzlA (1:10000) or α-FtsZ (1:10000) (Mohl et al., 2001) primary antibodies. For quantitation, films were scanned and band intensity was determined using Kodak MI software.

FtsZ activity assays

Initial co-sedimentation assays (Fig 2E) were performed with 3 μM FtsZ in polymerization buffer (50 mM HEPES pH 7.2, 50 mM KCl, 0.1 mM EDTA) with 10 mM MgCl2, 2 mM GTP and 0.05% Triton X-100. All subsequent pelleting assays were performed with 2 μM FtsZ in polymerization buffer with 2.5 mM MgCl2 and 2 mM GTP (unless indicated otherwise). All reactions were incubated at room temperature for 15 min prior to centrifugation at high (254000 × g) or low (16000 × g) speed. Light scattering was monitored using a Fluorolog-3 spectrofluorometer (Jobin Yvon, Inc.) Reactions contained 2 μM FtsZ (unless indicated otherwise) in polymerization buffer with 2.5 mM MgCl2. His6-FzlA (4 μM unless indicated otherwise) or FzlA buffer was added, and then GTP (2 mM unless indicated otherwise) or other nucleotides were added.

GTPase activity of FtsZ was assayed using the SensoLyte MG Phosphate Assay Kit (AnaSpec) following the manufacturer’s protocol. GraphPad Prism Software was used to fit curves and determine GTPase rates.

Electron microscopy

Polymerization reactions were set up as for pelleting assays using 2 μM FtsZ, 4 μM FzlA, 2.5 mM MgCl2, and 2 mM GTP and incubated for ~15 min (except where concentrations and/or times are indicated otherwise). Subsequently, samples were applied to glow-discharged carbon-coated grids, stained with 2% uranyl acetate, blotted, and air dried. Images were taken at 80kV on a JEOL TEM1230 transmission electron microscope equipped with a Gatan 967 slow-scan, cooled CCD camera. EM measurements were performed manually using ImageJ.

Detailed descriptions of growth conditions, protein purification, quantitative immunoblotting, and FtsZ activity assays are included in the Supplemental Information.

Supplementary Material

01

Acknowledgments

We are grateful to members of the Shapiro and McAdams laboratories for helpful discussions. We thank Matt Welch for use of the Fluorolog-3 spectrofluorometer, Matt Footer for assistance with size exclusion chromatography, Yi-Chun Yeh and Esteban Toro for strains, Michael Fero for developing KAMS imaging software, and Angela Zippilli and Eric Chen for technical assistance. EDG is a Helen Hay Whitney Foundation postdoctoral fellow. JNW is supported by a postdoctoral fellowship, Grant Number 1F32AI073043-01A1, from the National Institute of Allergy and Infectious Diseases. This work was supported in part by the Office of Science (BER) DOE grant No. DEFG02-05ER64136 (ZG and LS) and by NIH grant R01 GM 32506 (LS).

Footnotes

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.

References

  1. Aaron M, Charbon G, Lam H, Schwarz H, Vollmer W, Jacobs-Wagner C. The tubulin homologue FtsZ contributes to cell elongation by guiding cell wallprecursor synthesis in Caulobacter crescentus. Mol Microbiol. 2007;64:938–952. doi: 10.1111/j.1365-2958.2007.05720.x. [DOI] [PubMed] [Google Scholar]
  2. Adams DW, Errington J. Bacterial cell division: assembly, maintenance and disassembly of the Z ring. Nat Rev Microbiol. 2009;7:642–653. doi: 10.1038/nrmicro2198. [DOI] [PubMed] [Google Scholar]
  3. Allard JF, Cytrynbaum EN. Force generation by a dynamic Z-ring in Escherichia coli cell division. Proc Natl Acad Sci U S A. 2009;106:145–150. doi: 10.1073/pnas.0808657106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Anderson DE, Gueiros-Filho FJ, Erickson HP. Assembly dynamics of FtsZ rings in Bacillus subtilis and Escherichia coli and effects of FtsZ-regulating proteins. J Bacteriol. 2004;186:5775–5781. doi: 10.1128/JB.186.17.5775-5781.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bramhill D, Thompson CM. GTP-dependent polymerization of Escherichia coli FtsZ protein to form tubules. Proc Natl Acad Sci U S A. 1994;91:5813–5817. doi: 10.1073/pnas.91.13.5813. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Christen B, Fero MJ, Hillson NJ, Bowman G, Hong SH, Shapiro L, McAdams HH. High-throughput identification of protein localization dependency networks. Proc Natl Acad Sci U S A. 2010;107:4681–4686. doi: 10.1073/pnas.1000846107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Corbin BD, Wang Y, Beuria TK, Margolin W. Interaction between cell division proteins FtsE and FtsZ. J Bacteriol. 2007;189:3026–3035. doi: 10.1128/JB.01581-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Costa T, Priyadarshini R, Jacobs-Wagner C. Localization of PBP3 in Caulobacter crescentus is highly dynamic and largely relies on its functional transpeptidase domain. Mol Microbiol. 2008;70:634–651. doi: 10.1111/j.1365-2958.2008.06432.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Dajkovic A, Lutkenhaus J. Z ring as executor of bacterial cell division. J Mol Microbiol Biotechnol. 2006;11:140–151. doi: 10.1159/000094050. [DOI] [PubMed] [Google Scholar]
  10. de Boer P, Crossley R, Rothfield L. The essential bacterial cell-division protein FtsZ is a GTPase. Nature. 1992;359:254–256. doi: 10.1038/359254a0. [DOI] [PubMed] [Google Scholar]
  11. de Leeuw E, Graham B, Phillips GJ, ten Hagen-Jongman CM, Oudega B, Luirink J. Molecular characterization of Escherichia coli FtsE and FtsX. Mol Microbiol. 1999;31:983–993. doi: 10.1046/j.1365-2958.1999.01245.x. [DOI] [PubMed] [Google Scholar]
  12. Din N, Quardokus EM, Sackett MJ, Brun YV. Dominant C-terminal deletions of FtsZ that affect its ability to localize in Caulobacter and its interaction with FtsA. Mol Microbiol. 1998;27:1051–1063. doi: 10.1046/j.1365-2958.1998.00752.x. [DOI] [PubMed] [Google Scholar]
  13. Ebersbach G, Galli E, Moller-Jensen J, Lowe J, Gerdes K. Novel coiled-coil cell division factor ZapB stimulates Z ring assembly and cell division. Mol Microbiol. 2008;68:720–735. doi: 10.1111/j.1365-2958.2008.06190.x. [DOI] [PubMed] [Google Scholar]
  14. Erickson HP. FtsZ, a tubulin homologue in prokaryote cell division. Trends Cell Biol. 1997;7:362–367. doi: 10.1016/S0962-8924(97)01108-2. [DOI] [PubMed] [Google Scholar]
  15. Erickson HP. Modeling the physics of FtsZ assembly and force generation. Proc Natl Acad Sci U S A. 2009;106:9238–9243. doi: 10.1073/pnas.0902258106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Erickson HP, Taylor DW, Taylor KA, Bramhill D. Bacterial cell division protein FtsZ assembles into protofilament sheets and minirings, structural homologs of tubulin polymers. Proc Natl Acad Sci U S A. 1996;93:519–523. doi: 10.1073/pnas.93.1.519. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Figge RM, Divakaruni AV, Gober JW. MreB, the cell shape-determining bacterial actin homologue, co-ordinates cell wall morphogenesis in Caulobacter crescentus. Mol Microbiol. 2004;51:1321–1332. doi: 10.1111/j.1365-2958.2003.03936.x. [DOI] [PubMed] [Google Scholar]
  18. Gardy JL, Laird MR, Chen F, Rey S, Walsh CJ, Ester M, Brinkman FS. PSORTb v.2.0: expanded prediction of bacterial protein subcellular localization and insights gained from comparative proteome analysis. Bioinformatics. 2005;21:617–623. doi: 10.1093/bioinformatics/bti057. [DOI] [PubMed] [Google Scholar]
  19. Ghosh B, Sain A. Origin of contractile force during cell division of bacteria. Phys Rev Lett. 2008;101:178101. doi: 10.1103/PhysRevLett.101.178101. [DOI] [PubMed] [Google Scholar]
  20. Gitai Z, Dye N, Shapiro L. An actin-like gene can determine cell polarity in bacteria. Proc Natl Acad Sci U S A. 2004;101:8643–8648. doi: 10.1073/pnas.0402638101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Graumann PL. Cytoskeletal elements in bacteria. Annu Rev Microbiol. 2007;61:589–618. doi: 10.1146/annurev.micro.61.080706.093236. [DOI] [PubMed] [Google Scholar]
  22. Gueiros-Filho FJ, Losick R. A widely conserved bacterial cell division protein that promotes assembly of the tubulin-like protein FtsZ. Genes Dev. 2002;16:2544–2556. doi: 10.1101/gad.1014102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hale CA, Rhee AC, de Boer PA. ZipA-induced bundling of FtsZ polymers mediated by an interaction between C-terminal domains. J Bacteriol. 2000;182:5153–5166. doi: 10.1128/jb.182.18.5153-5166.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Harry E, Monahan L, Thompson L. Bacterial cell division: the mechanism and its precison. Int Rev Cytol. 2006;253:27–94. doi: 10.1016/S0074-7696(06)53002-5. [DOI] [PubMed] [Google Scholar]
  25. Horger I, Velasco E, Mingorance J, Rivas G, Tarazona P, Velez M. Langevin computer simulations of bacterial protein filaments and the force-generating mechanism during cell division. Phys Rev E Stat Nonlin Soft Matter Phys. 2008;77:011902. doi: 10.1103/PhysRevE.77.011902. [DOI] [PubMed] [Google Scholar]
  26. Lan G, Daniels BR, Dobrowsky TM, Wirtz D, Sun SX. Condensation of FtsZ filaments can drive bacterial cell division. Proc Natl Acad Sci U S A. 2009;106:121–126. doi: 10.1073/pnas.0807963106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Li Z, Trimble MJ, Brun YV, Jensen GJ. The structure of FtsZ filaments in vivo suggests a force-generating role in cell division. EMBO J. 2007;26:4694–4708. doi: 10.1038/sj.emboj.7601895. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Lowe J, Amos LA. Tubulin-like protofilaments in Ca2+-induced FtsZ sheets. EMBO J. 1999;18:2364–2371. doi: 10.1093/emboj/18.9.2364. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Lu C, Reedy M, Erickson HP. Straight and curved conformations of FtsZ are regulated by GTP hydrolysis. J Bacteriol. 2000;182:164–170. doi: 10.1128/jb.182.1.164-170.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Ma X, Margolin W. Genetic and functional analyses of the conserved C-terminal core domain of Escherichia coli FtsZ. J Bacteriol. 1999;181:7531–7544. doi: 10.1128/jb.181.24.7531-7544.1999. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Martin ME, Trimble MJ, Brun YV. Cell cycle-dependent abundance, stability and localization of FtsA and FtsQ in Caulobacter crescentus. Mol Microbiol. 2004;54:60–74. doi: 10.1111/j.1365-2958.2004.04251.x. [DOI] [PubMed] [Google Scholar]
  32. Matsumoto K, Kusaka J, Nishibori A, Hara H. Lipid domains in bacterial membranes. Mol Microbiol. 2006;61:1110–1117. doi: 10.1111/j.1365-2958.2006.05317.x. [DOI] [PubMed] [Google Scholar]
  33. McGrath PT, Lee H, Zhang L, Iniesta AA, Hottes AK, Tan MH, Hillson NJ, Hu P, Shapiro L, McAdams HH. High-throughput identification of transcription start sites, conserved promoter motifs and predicted regulons. Nat Biotechnol. 2007;25:584–592. doi: 10.1038/nbt1294. [DOI] [PubMed] [Google Scholar]
  34. Meredith DH, Plank M, Lewis PJ. Different patterns of integral membrane protein localization during cell division in Bacillus subtilis. Microbiology. 2008;154:64–71. doi: 10.1099/mic.0.2007/013268-0. [DOI] [PubMed] [Google Scholar]
  35. Mohl DA, Easter J, Jr, Gober JW. The chromosome partitioning protein, ParB, is required for cytokinesis in Caulobacter crescentus. Mol Microbiol. 2001;42:741–755. doi: 10.1046/j.1365-2958.2001.02643.x. [DOI] [PubMed] [Google Scholar]
  36. Moll A, Thanbichler M. FtsN-like proteins are conserved components of the cell division machinery in proteobacteria. Mol Microbiol. 2009;72:1037–1053. doi: 10.1111/j.1365-2958.2009.06706.x. [DOI] [PubMed] [Google Scholar]
  37. Mukherjee A, Lutkenhaus J. Guanine nucleotide-dependent assembly of FtsZ into filaments. J Bacteriol. 1994;176:2754–2758. doi: 10.1128/jb.176.9.2754-2758.1994. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Mukherjee A, Lutkenhaus J. Analysis of FtsZ assembly by light scattering and determination of the role of divalent metal cations. J Bacteriol. 1999;181:823–832. doi: 10.1128/jb.181.3.823-832.1999. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Oakley AJ. Glutathione transferases: new functions. Curr Opin Struct Biol. 2005;15:716–723. doi: 10.1016/j.sbi.2005.10.005. [DOI] [PubMed] [Google Scholar]
  40. Osawa M, Anderson DE, Erickson HP. Reconstitution of contractile FtsZ rings in liposomes. Science. 2008;320:792–794. doi: 10.1126/science.1154520. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Osawa M, Anderson DE, Erickson HP. Curved FtsZ protofilaments generate bending forces on liposome membranes. EMBO J. 2009;28:3476–3484. doi: 10.1038/emboj.2009.277. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Popp D, Iwasa M, Narita A, Erickson HP, Maeda Y. FtsZ condensates: an in vitro electron microscopy study. Biopolymers. 2009;91:340–350. doi: 10.1002/bip.21136. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. RayChaudhuri D. ZipA is a MAP-Tau homolog and is essential for structural integrity of the cytokinetic FtsZ ring during bacterial cell division. EMBO J. 1999;18:2372–2383. doi: 10.1093/emboj/18.9.2372. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. RayChaudhuri D, Park JT. Escherichia coli cell-division gene ftsZ encodes a novel GTP-binding protein. Nature. 1992;359:251–254. doi: 10.1038/359251a0. [DOI] [PubMed] [Google Scholar]
  45. Sackett MJ, Kelly AJ, Brun YV. Ordered expression of ftsQA and ftsZ during the Caulobacter crescentus cell cycle. Mol Microbiol. 1998;28:421–434. doi: 10.1046/j.1365-2958.1998.00753.x. [DOI] [PubMed] [Google Scholar]
  46. Singh JK, Makde RD, Kumar V, Panda D. SepF increases the assembly and bundling of FtsZ polymers and stabilizes FtsZ protofilaments by binding along its length. J Biol Chem. 2008;283:31116–31124. doi: 10.1074/jbc.M805910200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Stricker J, Maddox P, Salmon ED, Erickson HP. Rapid assembly dynamics of the Escherichia coli FtsZ-ring demonstrated by fluorescence recovery after photobleaching. Proc Natl Acad Sci U S A. 2002;99:3171–3175. doi: 10.1073/pnas.052595099. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Thanbichler M, Shapiro L. MipZ, a spatial regulator coordinating chromosome segregation with cell division in Caulobacter. Cell. 2006;126:147–162. doi: 10.1016/j.cell.2006.05.038. [DOI] [PubMed] [Google Scholar]
  49. Wang X, Huang J, Mukherjee A, Cao C, Lutkenhaus J. Analysis of the interaction of FtsZ with itself, GTP, and FtsA. J Bacteriol. 1997;179:5551–5559. doi: 10.1128/jb.179.17.5551-5559.1997. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Wang Y, Jones BD, Brun YV. A set of ftsZ mutants blocked at different stages of cell division in Caulobacter. Mol Microbiol. 2001;40:347–360. doi: 10.1046/j.1365-2958.2001.02395.x. [DOI] [PubMed] [Google Scholar]
  51. Werner JN, Chen EY, Guberman JM, Zippilli AR, Irgon JJ, Gitai Z. Quantitative genome-scale analysis of protein localization in an asymmetric bacterium. Proc Natl Acad Sci U S A. 2009;106:7858–7863. doi: 10.1073/pnas.0901781106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Yu XC, Margolin W. Ca2+-mediated GTP-dependent dynamic assembly of bacterial cell division protein FtsZ into asters and polymer networks in vitro. EMBO J. 1997;16:5455–5463. doi: 10.1093/emboj/16.17.5455. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01
NIHMS232221-supplement-01.doc (6.4MB, doc)

RESOURCES