![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Biology, Washington University, St. Louis, Missouri 63130
Received 24 July 2007/ Accepted 4 September 2007
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Like tubulin, FtsZ assembly dynamics result from GTP binding and hydrolysis. In vitro GTP binding induces FtsZ assembly into single-stranded polymers. These polymers then interact laterally to form the bundles (19, 23) that are thought to constitute the FtsZ ring in vivo (21, 25). Subsequent GTP hydrolysis destabilizes FtsZ polymers, which become curved and rapidly disassemble (25).
In the cell, FtsZ exists in a precise balance between the membrane-associated single-stranded polymers and bundles that constitute the FtsZ ring and the cytoplasmic monomers and small multimers that constitute the pool of FtsZ subunits available for exchange into the cytokinetic ring (21, 25). Disturbances in FtsZ assembly dynamics result in aberrant division events or potentially lethal filamentation (10, 21, 25). The FtsZ ring itself is a highly dynamic structure. Fluorescence recovery after photobleaching (FRAP) experiments have indicated that the half-life of an individual subunit is between 4 and 20 s (1). Maintaining the precise balance between FtsZ assembly and disassembly is critical for coordinating division with cell growth and to ensure that daughter cells each receive a full genetic complement following cytokinesis (10, 21, 25).
FtsZ levels are constant throughout the cell cycle, and the spatial and temporal control of division is governed primarily at the level of FtsZ assembly (30). To date, a few proteins that modulate FtsZ assembly dynamics in vivo have been identified (6, 10). One of these proteins, EzrA, is an inhibitor that prevents aberrant FtsZ assembly at the poles of exponentially growing cells (8, 17, 19).
EzrA is found throughout the low-G+C-content gram-positive bacteria and exhibits significant structural conservation. EzrA is anchored in the plasma membrane by an N-terminal transmembrane domain. The remainder of the EzrA polypeptide consists of a series of predicted cytoplasmic coiled-coil domains (17). This somewhat unusual topology (N terminus out topology) is shared with ZipA, a cell division protein that is found in the gamma subdivision of the Proteobacteria (9). Null mutations in ezrA lower the concentration of FtsZ required for assembly in vivo, leading to the formation of polar FtsZ rings and septa (17) and stabilizing FtsZ assembly at midcell (19).
A purified thioredoxin-EzrA fusion lacking the EzrA transmembrane anchor (Thio-EzrA) interacts with FtsZ in vitro and inhibits FtsZ assembly in both sedimentation and 90°-angle light-scattering assays (8). Deletion analysis has indicated that EzrA interacts with the C-terminal 16 residues of FtsZ to inhibit assembly in vitro (28). It has recently been reported that addition of the native cytoplasmic portion of EzrA results in a modest (1.6-fold) increase in FtsZ GTP hydrolysis (5), suggesting that EzrA may increase subunit turnover. However, the precise mechanism by which EzrA modulates FtsZ ring formation remains unknown, as an EzrA fusion protein lacking the transmembrane domain effectively inhibits FtsZ assembly in vitro without altering FtsZ's intrinsic GTPase activity (8).
EzrA exhibits two distinct patterns of subcellular localization. In predivisional cells EzrA is uniformly distributed throughout the plasma membrane. Paradoxically, given its ability to inhibit FtsZ assembly, EzrA also concentrates at the division site in an FtsZ-dependent manner (17). EzrA's role at midcell is unclear. EzrA localization to the medial ring may simply be a consequence of residual interaction with FtsZ. Alternatively, EzrA may promote subunit turnover at midcell. Arguing against the latter model, FRAP experiments on FtsZ rings have indicated that there is little change in subunit turnover in the absence of EzrA (1). However, cells lacking EzrA are significantly longer than wild-type cells (4, 14, 17), consistent with altered FtsZ assembly dynamics at midcell that disrupt the coupling of division with cell growth.
To clarify the role of EzrA at midcell, we employed a combination of deletion analysis and site-directed mutagenesis to identify regions of EzrA that are required for medial localization. Using this approach, we identified a conserved seven-residue patch (designated the "QNR patch" after the three most conserved residues) that is required for EzrA localization to the FtsZ ring. A single substitution in the QNR patch abolishes medial EzrA localization, but it does not alter the ability of EzrA to inhibit FtsZ assembly in vitro and results in only a negligible increase in polar FtsZ assembly. However, despite lacking extra FtsZ rings, ezrA QNR patch mutant cells are longer than their wild-type counterparts and exhibit stabilized FtsZ assembly at midcell, similar to ezrA null mutant cells. Together, our results support a model in which EzrA has two distinct functions in vivo: (i) preventing aberrant FtsZ ring formation by inhibiting de novo FtsZ assembly at cell poles and (ii) maintaining proper FtsZ assembly dynamics within the medial FtsZ ring.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
The ezrA::kan null mutation was created by inserting a 1.4-kb kanamycin cassette between codons 147 and 285 of ezrA. The kan cassette was removed from pGK67 (15) by SmaI/PstI (NEB) digestion and ligated into a 1.1-kb ezrA fragment within a derivative of pCR2.1 (Invitrogen) digested with PstI/Eco47III (NEB). The resulting plasmid was transformed into JH642 for ezrA::kan integration at the native ezrA locus by double-crossover homologous recombination with selection for kanamycin resistance. Proper integration was verified by PCR.
C-terminal EzrA truncations that were fused to green fluorescent protein (GFP) were created in pPL65 (a derivative of pUS19 [2] containing a 
900-bp 3' ezrA fragment fused to gfp). The wild-type ezrA fragment was cut from the vector by EcoRI/XhoI (NEB) digestion and replaced with a relevant ezrA truncation fragment amplified by PCR from the JH642 chromosome. All ezrA-gfp point mutations and the ezrA
501-511-gfp deletion were created using a QuikChange site-directed mutagenesis kit (Stratagene) on pPL65. Tagless mutants lacking GFP were constructed by PCR amplifying the ezrA fragment (residues 355 to 562) with EcoRI/BamHI (NEB) linkers from the pPL65 derivatives and placing them into pUS19. The resulting plasmids were transformed into JH642 with selection for single-crossover recombination events at the native ezrA locus by growth on spectinomycin-containing media. Mutations were confirmed by plasmid sequencing, and integration was confirmed by chromosomal ezrA locus sequencing.
Strains used for the minCD and ftsZ(Ts) suppression assays were constructed by transforming the chromosome from the tagless ezrA mutant strains into the relevant background and selecting for the appropriate antibiotic resistance. minCD and ftsZ(Ts) suppression assays were conducted as previously described (19, 31).
The Thio-EzrATM(R510D)-His6 fusion [Thio-EzrA(R510D)] under PBAD control (7) was constructed by amplification of full-length ezrATM(R510D) from PL1780, followed by TOPO cloning (Invitrogen) as described previously for the wild-type EzrA fusion (Thio-EzrA) (8). After the mutation within the plasmid was verified by sequencing, the plasmid was transformed into "One Shot chemically competent" TOP10 cells (Invitrogen) and then isolated and transformed into a BB101 background (3) for protein induction and purification.
Microscopy. Fluorescence microscopy was performed as described previously (17). An Olympus BX51 microscope with Chroma filters and a Hamamatsu OrcaERG camera were used for image capture. Images were processed using Openlab version 5.0.1 (Improvision) and Adobe Photoshop CS version 8.0 (Adobe Systems). All cell or ring measurements for collected images were obtained with a minimum population of 200 cells/strain.
GFP fusions were visualized in live cells as described previously (16). Briefly, cells grown to mid-exponential phase (optical density at 600 nm, 0.4) in liquid culture were stained with the vital membrane dye FM 4-64 (Invitrogen) at a dilution of 1:2,000 for 1 to 3 min and then placed on 1% agarose in 1x phosphate-buffered saline pads (16) on glass microscope slides. EzrA-GFP localization to midcell was scored as strong, weak, or none based on fluorescent band intensity. Statistical analysis of EzrA-GFP localization between strains was performed using a 
2 test with 2 df and a significance level (
) of 0.001.
Cells were prepared for immunofluorescence microscopy by paraformaldehyde and glutaraldehyde treatment, as described previously (16, 17), with lysozyme incubation typically lasting 7 to 10 min. FtsZ was detected using affinity-purified polyclonal rabbit anti-FtsZ sera (18) in combination with donkey anti-rabbit sera conjugated to Cyanine-3 (Jackson Immunoresearch). Cell walls were visualized with wheat germ agglutinin conjugated to fluorescein (Invitrogen). Statistical analysis of FtsZ localization between strains was performed using a 2 test (5 df; = 0.001.
Cell lengths were measured with fixed cells, using a 3- to 4-min lysozyme treatment that better preserved cell wall structure. Septa were defined as bands of fluorescent signal that fully traversed the cell width with intensity equal to or greater than the intensity of lateral wall fluorescence. Cell length data were obtained with Openlab software and exported to Microsoft Excel (version 11.1) for analysis. Statistical analysis of cell lengths between strains was performed using a 2 test (5 df; = 0.001). Length-to-ring ratios were determined as described previously (29, 32).
Protein purification. B. subtilis FtsZ was purified from PL1184, an E. coli ER2566 derivative, as described previously (8, 31). Thio-EzrA and Thio-EzrA(R510D) were purified from E. coli BB101 (3) derivatives (PL1364 and PL1821, respectively) as described previously (8). A size exclusion chromatography step was added to our original protocol to remove EzrA degradation products that frequently caused aggregation and precipitation of the final product.
One-liter cultures of cells containing induced Thio-EzrA or Thio-EzrA(R510D) were pelleted, washed in EzrA induction buffer (50 mM NaPO4 [pH 8.0], 300 mM NaCl), repelleted, and frozen at –80°C for later use. On the day of purification, cell pellets were thawed and resuspended in 30 ml of ice-cold EzrA induction buffer containing 1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride (Sigma). Cells were lysed by two or three passes in a prechilled, 30-ml-capacity French press cell at 1,000 lb/in2. All subsequent steps were performed at 4°C. Lysates were cleared by centrifugation as described previously (8) and loaded in EzrA buffer 350 (50 mM HEPES [pH 7.5], 350 mM NaCl) containing 50 mM imidazole onto two tandem 5-ml Hi-Trap chelating HP columns (GE Healthcare) charged with nickel on an ÁKTA prime low-pressure purification system (GE Healthcare). Bound EzrA was washed with EzrA buffer 350 containing 75 mM imidazole, and bound protein was eluted in one step with EzrA buffer 350 containing 200 mM imidazole. Peak fractions were pooled, concentrated to 1 to 2 ml using Centricon YM-50 filters (Fisher Scientific), and applied to an S300 gel filtration column (GE Healthcare) preequilibrated with EzrA buffer 175 (50 mM HEPES [pH 7.5], 175 mM NaCl, 1 mM EGTA). Both the Thio-EzrA and Thio-EzrA(R510D) proteins eluted in three peaks: a small inactive aggregate peak, an active peak with the apparent molecular weight of a dimer, and an inactive peak with the apparent molecular weight of a monomer. Fractions from the active peak were collected, pooled, and concentrated with YM-50 filters. Glycerol was added to a final concentration of 10%, and aliquots were flash frozen at –80°C.
The FtsZ protein concentration was determined by using an average of bicinchoninic acid (Pierce) and Coomassie Plus (Pierce) assay estimates, while EzrA fusion concentrations were estimated by bicinchoninic acid and Coomassie Plus assays, as well as by using A280 (molar extinction coefficient, 58,656.0 A280/mol as determined by Gene Inspector, version 1.6.3 [Textco]) and a SPECTRAmax Plus spectrophotometer (Molecular Devices).
90°-angle light-scattering assay. Light-scattering assays were conducted essentially as described previously (8, 31) using a DM-45 spectrofluorimeter (Olis). Readings were taken four times per second at 30°C, and a baseline was determined for 1 min before addition of 1 mM GTP to the cuvette. The reaction mixtures contained 5 µM FtsZ diluted in polymerization buffer [50 mM morpholineethanesulfonic acid (MES) (pH 6.5), 2.5 mM Mg(CH3COO–)2, 1 mM EGTA, 50 mM KCH3COO–) and EzrA buffer 175 containing 10% glycerol with or without Thio-EzrA or Thio-EzrA(R510D). Data were collected by SpectralWorks (Olis) and exported into Microsoft Excel for data processing. Baseline corrections were applied in Excel to remove the background signal from unassembled FtsZ and the EzrA fusions.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
551-562-GFP, a comparatively small C-terminal truncation, showed the same localization as full-length EzrA-GFP (Fig. 2A).
|
We next employed site-directed mutagenesis to identify specific residues within the QNR patch that are required for medial EzrA localization. In these experiments, we targeted the three most conserved residues for alanine substitution: Q506, N509, and R510. We also constructed an R510D charge reversal (Fig. 1C).
Fluorescence microscopy of live cells expressing various EzrA-GFP fusions indicated that single substitutions in the QNR patch disrupted EzrA localization to the FtsZ ring, albeit to different degrees (Fig. 2B and Table 2). The majority of ezrA-gfp cells displayed strong medial EzrA localization (Fig. 2A and Table 2). In contrast, in cells with ezrA(Q506A)-gfp there was a twofold decrease in the population exhibiting strong medial EzrA localization and a corresponding increase in weak localization (Fig. 2B and Table 2). This phenotype was more pronounced in ezrA(N509A)-gfp and ezrA(R510A)-gfp cells, and the majority (60 to 67%) of these cells showed no medial EzrA localization (Fig. 2B and Table 2). Notably, EzrA localization to the FtsZ ring was completely abolished in ezrA(R510D)-gfp cells (Fig. 2B and Table 2).
|
Mutations in EzrA's QNR patch stabilize FtsZ assembly at midcell. Loss of EzrA stabilizes FtsZ assembly at midcell (17, 19), suggesting that EzrA plays an active role at this position. To test this possibility, we examined the effect of mutations in EzrA's QNR patch on the lethality associated with overexpression of the division inhibitor MinCD. During exponential growth the MinCD complex is involved in preventing inappropriate FtsZ ring formation at cell poles (20, 22); however, >12-fold overexpression of MinCD shifts the cellular balance of FtsZ dynamics towards unassembled FtsZ, blocking FtsZ ring formation and causing lethal filamentation in B. subtilis (20, 22). Stabilizing FtsZ assembly at midcell via a mutation in a division inhibitor gene, such as ezrA, permits FtsZ ring formation and division even in the presence of excess MinCD (19, 29, 31).
For these experiments we cloned the four ezrA QNR point mutants (Fig. 1C) under control of the native ezrA promoter and without GFP tags into otherwise wild-type cells. Each ezrA QNR patch mutant was then transformed into a strain encoding an IPTG-inducible minCD overexpression construct, Pspachy-minCD. Viability was tested by plating cells onto medium containing 1 mM IPTG to induce >12-fold overexpression of MinCD. Removal of the GFP tag did not alter EzrA mutant stability or expression, and FtsZ levels were wild type in all strain backgrounds (data not shown).
Suppression of MinCD-induced lethality by mutations in ezrA's QNR patch (Fig. 3A) correlated with the ability of individual QNR patch mutants to localize to the FtsZ ring (Fig. 2B). As expected, viability was reduced 28,000-fold following MinCD overexpression in ezrA+ cells (Fig. 3A). In contrast, the ezrA null mutant counterparts were fully viable in the presence of inducer (Fig. 3A). Of the four QNR patch mutants, those that were severely impaired in medial localization (Fig. 2B and Table 2) fully restored viability to cells in the presence of inducer. On the other hand, the ezrA(Q506A) mutation only weakly suppressed the lethality associated with MinCD overexpression (Fig. 3A), consistent with its intermediate localization phenotype (Fig. 2B and Table 2).
|
3,000-fold reduction in viability (17). Loss-of-function mutations in ezrA or other inhibitors of cell division restore FtsZ ring formation and viability to ftsZ(Ts) cells at the restrictive temperature (17, 31). In confirmation of an active role for EzrA at midcell, loss of medial EzrA localization correlated with suppression of ftsZ(Ts) heat sensitivity. As expected, an ezrA null mutation fully restored viability to ftsZ(Ts) cells at the restrictive temperature (Fig. 3B). The QNR patch mutants that were most severely impaired in medial EzrA localization (Fig. 2B and Table 2) restored full viability to ftsZ(Ts) cells at the restrictive temperature (Fig. 3B). Conversely, EzrA(Q506A), which exhibits an intermediate medial EzrA localization phenotype (Fig. 2B and Table 2), conferred variable levels of suppression that ranged from low levels to nearly ezrA null mutant levels (note the relatively large error bar in Fig. 3B). These results confirm the results of the MinCD overexpression assay and are consistent with a model in which EzrA acts at midcell to destabilize FtsZ assembly at this position.
EzrA QNR patch mutants retain the ability to inhibit polar FtsZ ring formation. Aberrant, polar FtsZ ring formation is a hallmark of a loss-of-function mutation in ezrA (17). Polar ring formation is most likely due to an increase in FtsZ stability that overcomes MinCD activity at cell poles (19). The inability of EzrA QNR patch mutants to localize to midcell and the concurrent stabilization of the medial FtsZ ring are therefore consistent with a defect in EzrA's ability to inhibit FtsZ assembly throughout the cell. To test this possibility, we examined FtsZ ring formation in the three ezrA QNR patch mutants [the ezrA(N509A), ezrA(R510A), and ezrA(R510D) mutants] that showed the greatest reduction in medial EzrA-GFP localization (Fig. 2B and Table 2).
Remarkably, the three QNR patch mutants retained the ability to inhibit polar FtsZ ring formation (Fig. 4), despite the loss of medial localization (Fig. 2). Consistent with previous work, the frequency of polar FtsZ ring formation in ezrA null mutant cells was 15-fold higher than that in wild-type cells (Fig. 4B). Cells of two of the three QNR patch mutants [ezrA(N509A) and ezrA(R510A)] displayed FtsZ ring localization patterns that were indistinguishable from that of wild-type cells. ezrA(R510D) mutant cells exhibited a mild (3.9-fold) increase in polar ring formation (Fig. 4B). However, this increase was not statistically significant as determined by 2 analysis ( = 0.001.) Together, these data indicate that the loss of localization to midcell does not eliminate EzrA's ability to inhibit FtsZ assembly at cell poles.
|
Consistent with previous results (8), addition of 1 mM GTP to a reaction mixture containing 5 µM purified FtsZ led to a rapid increase in light scattering, while the presence of 10 µM purified Thio-EzrA decreased FtsZ light scattering by 60% (Fig. 5A). This result was typical for wild-type EzrA purified using the two-step protocol described in Materials and Methods. Strikingly, Thio-EzrA(R510D) inhibited FtsZ assembly to the same degree as its wild-type counterpart (Fig. 5A), even at low concentrations (Fig. 5B). These data indicate that Thio-EzrA(R510D) is wild type with regard to its ability to interact with FtsZ in vitro and are consistent with data indicating that QNR patch mutants interact with FtsZ in vivo to inhibit polar FtsZ ring formation (Fig. 4).
|
To determine whether the loss of EzrA activity at midcell is the proximal cause of the increase in cell length observed in ezrA null mutants, we measured the lengths of QNR patch mutant cells in both nutrient-rich (LB) and nutrient-poor (S750-minimal glucose [13]) media. If EzrA plays an active role at midcell, then ezrA QNR patch mutant cells should be longer than wild-type cells despite the low frequency of polar ring formation. Conversely, if the increased length of ezrA null mutant cells is a consequence of extra FtsZ ring formation at cell poles, then ezrA QNR patch mutants should exhibit a wild-type cell length distribution.
Our data indicate that EzrA localization to midcell is essential for maintaining proper cell size, regardless of polar FtsZ ring formation. Wild-type cells cultured in LB medium had a mean length of 4.0 µm, while ezrA null mutant cells were 75% longer (mean length, 7.0 µm). For the QNR patch mutants, cell length (Fig. 6) correlated with localization pattern (Fig. 2 and Table 2). The cells of the two mutants that displayed an intermediate EzrA localization phenotype [the ezrA(N509A) and ezrA(R510A) mutants] had average lengths that were 23% greater (mean length, 4.9 µm) than the length of wild-type cells. ezrA(R510D) mutant cells, which are completely defective in medial EzrA localization, were 45% longer than wild-type cells (mean length, 5.8 µm). Together, these data show that there was a significant shift in cell length distribution upon disruption of medial EzrA localization (Fig. 6A), even when few polar rings were present.
|
3.2 µm. Although shorter than cells cultured in LB medium, ezrA null cells cultured under these conditions were still 59% longer than wild-type cells (mean length, 5.0 µm). Like mutants cultured in LB medium, ezrA QNR patch mutants grown in minimal medium also showed a statistically significant increase in cell length compared to the wild-type strain. Cells of the ezrA(N509A) and ezrA(R510A) mutants were both 6% longer than wild-type cells (mean length, 3.6 µm), and cells of the more severe ezrA(R510D) mutant were 31% longer (mean length, 4.2 µm). The intermediate length of the ezrA(R510D) mutant cells in LB and minimal glucose media (Fig. 6) suggests that this mutant retained some residual activity at midcell, despite showing no medial EzrA localization by fluorescence microscopy (Fig. 2). The frequency of minicell formation in ezrA null mutants and ezrA QNR patch mutants was less than 4% in both minimal glucose and LB media (data not shown); thus, polar septation events that occur at the expense of binary fission are unlikely to be the explanation for increases in cell length.
As further evidence of a disturbance in cell size homeostasis, the ezrA QNR patch mutants exhibited an increase in the ratio of cell length to FtsZ rings (L/R ratio). Wild-type cells exhibited an L/R ratio of 6.5 µm/ring under a range of growth conditions (29). Conversely, the L/R ratio of ezrA null mutant cells increased by 31% (8.5 µm/ring) after growth in LB medium, despite the presence of extra FtsZ rings. Two of the QNR patch mutants, the ezrA(N509A) and ezrA(R510A) mutants, each had a 7% increase in the L/R ratio (6.9 µm/ring), and the ezrA(R510D) mutant had a 19% increase in the L/R ratio (7.7 µm/ring). The widths of wild-type and ezrA mutant cells were statistically indistinguishable (data not shown).
Importantly, the percentage of ezrA mutant cells that contained FtsZ rings was identical to the percentage of wild-type cells that contained FtsZ rings (Fig. 4), and the mass doubling times were also identical (data not shown). These data indicate that the timing of FtsZ ring formation relative to cell birth and the duration of the Z-period were not perturbed. If loss of medial EzrA activity affected either of these two parameters, we would have expected to see changes in FtsZ ring frequency in an ezrA null mutant cell population compared to the wild-type cell population. This is counter to previous reports suggesting that FtsZ ring constriction is delayed in ezrA null mutants during slow growth (14). The difference between our results and those of Kawai and Ogasawara (14) may reflect differences in strain background, media, and/or temperature (e.g., our measurements were obtained with cells cultured in liquid medium at 37°C, whereas Kawai and Ogasawara used slow-growing cells cultured on solid medium at 24°C [14]). Together, these data indicate that the increased length of ezrA mutant cells is due to loss of EzrA activity at midcell rather than to dilution of division components in the presence of extra FtsZ rings at cell poles.
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The isolation of ezrA mutants that are specifically defective for medial localization allowed us to clarify the role of EzrA at the cytokinetic ring. Notably, our data indicate that EzrA contributes to FtsZ assembly dynamics within the medial ring, thereby helping to coordinate division with cell growth and ensuring that cells are the proper length upon cytokinesis. QNR patch mutants were resistant to overexpression of the MinCD division inhibitor (Fig. 3) and exhibited increases in cell length (Fig. 6), consistent with stabilization of FtsZ assembly at midcell. The increases in cell length were unlikely to be due to a reduction in the cytoplasmic pool of division components, as ezrA QNR patch mutant cells exhibited a pattern of FtsZ ring formation that was statistically indistinguishable from that of wild-type cells. Moreover, we saw a similar increase in ezrA null mutant cell length in minimal medium (Fig. 6), conditions that do not support the formation of extra FtsZ rings (17). It is possible that ezrA QNR patch mutants possess a higher percentage of large FtsZ polymers even in the absence of extra FtsZ rings, a situation that might effectively sequester components of the division apparatus away from midcell. However, this explanation seems unlikely given the significant stabilization of FtsZ assembly at midcell in the ezrA QNR patch mutant backgrounds.
The idea that EzrA destabilizes FtsZ assembly at midcell appears to conflict with FRAP data indicating that FtsZ turnover rates are identical in wild-type and ezrA null mutant cells (1). However, FRAP also shows that the half-life of an FtsZ subunit in the cytokinetic ring varies significantly from 4 to 20 s, even in wild-type cells (1). Thus, it is possible that the loss of ezrA has a modest effect on subunit turnover that is undetectable by FRAP, yet sufficient to alter the stability of the medial FtsZ ring. Alternatively, the loss of EzrA may stabilize the FtsZ ring in a manner that does not significantly decrease subunit turnover itself (e.g., boosting nucleation potential at midcell or increasing the number of polymers in the ring itself).
Our data indicating that the frequency of FtsZ ring formation and the duration of the Z-period are not perturbed in ezrA mutant cells (Fig. 4) argue against a model in which the increase in cell length is due to changes in the timing of cytokinesis and/or in the persistence of the FtsZ ring, as previously suggested (14). Instead, we favor a model in which the loss of EzrA activity at midcell increases the stability of the medial FtsZ ring, rendering it resistant to the factors responsible for coordinating division with cell size. This situation would be analogous to the increase in FtsZ polymer stability that overcomes the activity of the MinCD complex at the poles of ezrA null mutant cells (19).
Despite being defective in medial localization (Fig. 2), EzrA QNR patch mutants are similar to wild-type EzrA with regard to the ability to inhibit aberrant FtsZ assembly at cell poles (Fig. 4). The data indicate that QNR patch mutants have not lost the ability to interact with FtsZ, consistent with our in vitro data (Fig. 5). It remains formally possible that in the absence of medial EzrA localization there is an increase in EzrA concentrations at the cell poles that compensates for a decreased EzrA-FtsZ interaction, thus allowing inhibition of polar FtsZ assembly to persist. However, EzrA(R510D), which is completely defective in medial localization, inhibits FtsZ assembly in vitro to wild-type levels (Fig. 5), suggesting that the EzrA-FtsZ interaction remains robust.
Based on these results, we propose that EzrA has at least two separate regions for interaction with FtsZ. The first region, located in the N-terminal portion of EzrA, is required for the interaction between EzrA and the free FtsZ monomers and multimers that constitute the cytoplasmic pool of FtsZ. The second region, contained in the QNR patch, promotes interaction with the single-stranded polymers and bundles that constitute the FtsZ ring but is dispensable for inhibition of de novo FtsZ assembly. Alternatively, the QNR patch may be required for interaction between EzrA and another component of the cytokinetic ring, thereby bringing EzrA into close proximity with FtsZ at this position. A fusion of the QNR patch to the C terminus of GFP was unable to localize to the FtsZ ring on its own, indicating that, while necessary, this fusion is not sufficient for medial localization (D. P. Haeusser, unpublished data).
Together, our results support a model in which EzrA has two distinct functions in vivo: one in which it prevents aberrant FtsZ assembly at cell poles, as previously characterized (8, 17, 19), and one in which it destabilizes FtsZ assembly at midcell, enhancing the dynamic nature of the medial FtsZ ring and ensuring that division is coupled to cell growth. Future work will address whether the QNR patch of EzrA is required for interaction with preformed FtsZ polymers or for interaction with other division proteins. Our results highlight the need for a more comprehensive understanding of the molecular nature of the FtsZ interaction with its modulating proteins, particularly those that appear to interact with FtsZ at more than one subcellular location.
|
ACKNOWLEDGMENTS |
|---|
This work was supported in part by Public Health Service grant GM64671 from the NIH and by National Science Foundation CAREER award MCB-0448186 to P.A.L. A.Z.B. was funded in part by a Ruth L. Kirschstein National Research Service Award from the NIH (grant F32-GM077828). A.C.G. was funded in part by a summer undergraduate research fellowship from the Howard Hughes Medical Institute.
|
FOOTNOTES |
|---|
Published ahead of print on 14 September 2007.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
B levels and activity in Bacillus subtilis. J. Bacteriol. 175:2347-2356.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Appl. Environ. Microbiol. | Infect. Immun. | Eukaryot. Cell |
|---|---|---|
| Mol. Cell. Biol. | J. Virol. | Microbiol. Mol. Biol. Rev. |
| ALL ASM JOURNALS |
