Assembly of the cytoskeletal protein FtsZ into a ring-like structure is required for bacterial cell division. Structurally, FtsZ consists of four domains: the globular N-terminal core, a flexible linker, 8–9 conserved residues implicated in interactions with modulatory proteins, and a highly variable set of 4–10 residues at its very C terminus. Largely ignored and distinguished by lack of primary sequence conservation, the linker is presumed to be intrinsically disordered. Here we employ genetics, biochemistry and cytology to dissect the role of the linker in FtsZ function. Data from chimeric FtsZs substituting the native linker with sequences from unrelated FtsZs as well as a helical sequence from human beta-catenin indicate that while variations in the primary sequence are well tolerated, an intrinsically disordered linker is essential for Bacillus subtilis FtsZ assembly. Linker lengths ranging from 25 to 100 residues supported FtsZ assembly, but replacing the B. subtilis FtsZ linker with a 249-residue linker from Agrobacterium tumefaciens FtsZ interfered with cell division. Overall, our results support a model in which the linker acts as a flexible tether allowing FtsZ to associate with the membrane through a conserved C-terminal domain while simultaneously interacting with itself and modulatory proteins in the cytoplasm.
The prokaryotic cytoskeletal protein FtsZ plays a central role in bacterial cell division. In rod-shaped bacteria such as Bacillus subtilis and Escherichia coli, FtsZ assembles into a ring shaped structure (Z ring) at midcell that establishes the location of the future division site and serves as a platform for assembly of the division machinery. As cell division progresses, the Z ring constricts to begin separation of the newly forming daughter cells. The Z ring is required for recruitment of downstream cell division proteins, including those required for cross-wall synthesis, to the nascent division site (for reviews see: Adams and Errington, 2009; Kirkpatrick and Viollier, 2011; Lutkenhaus et al., 2012).
High-resolution microscopy suggests that the Z ring consists primarily of heterogeneously distributed short filaments (∼ 100 nm in length) held together via lateral interactions (Fu et al., 2010; Strauss et al., 2012). FtsZ protofilaments are believed to associate with the membrane through interactions with FtsA and other, less widely conserved, cell division proteins. In addition, several studies have shown that FtsZ filaments targeted to the membrane in in vitro liposomes are able to generate visible distortions where FtsZ is bound and suggest the force being applied is solely through FtsZ (Osawa et al., 2008; 2009). Computational modelling suggests that GTP-hydrolysis-dependent bending of FtsZ filaments generates sufficient force to drive inward cell-wall growth during division (Hsin et al., 2012).
In vitro, FtsZ assembly into single stranded ‘protofilaments’ is stimulated by GTP (Romberg and Levin, 2003; Erickson et al., 2010). FtsZ, a tubulin homologue, binds to GTP as a monomer. Polymerization leads to the formation of active sites for GTP hydrolysis at the interface between FtsZ subunits. FtsZ protofilaments are able to associate laterally to form filament bundles, spirals, tubes, and sheets depending on the buffer conditions. Lateral interaction potential varies between FtsZs from different species and is dependent in part on the charge of the variable residues at the very C-terminus of the protein (Mukherjee and Lutkenhaus, 1999; Gueiros-Filho and Losick, 2002; Popp et al., 2009; Gündoğdu et al., 2011; Buske and Levin, 2012).
The FtsZ monomer is divided into 5 domains: an unstructured N-terminal peptide, a highly conserved globular core, an unstructured C-terminal linker (CTL), a conserved set of ∼ 11 residues referred to here as the C-terminal constant region (CTC, C-terminal tail), and a small, highly variable group of residues at the extreme C-terminus of FtsZ termed the C-terminal variable region (CTV) (Vaughan et al.,2004; Erickson et al., 2010; Buske and Levin, 2012). For simplicity, we treat the N-terminal peptide and core as a single unit. The core contains residues required for GTP binding and hydrolysis as well as residues involved in longitudinal interactions between subunits (Lu et al., 2001; Redick et al., 2005). In vitro, the core has also been shown to be sufficient for filament formation and has also been implicated in protofilament bundling (Wang et al.,1997). The very C-terminus of FtsZ, a 15 residue region in E. coli FtsZ and a 17 residue region in B. subtilis FtsZ that includes both the CTC and the CTV, has been implicated in interactions between FtsZ and a host of modulatory proteins. Such proteins include FtsA, ZipA and MinC in E. coli and EzrA and SepF in B. subtilis (Ma and Margolin, 1999; Mosyak et al., 2000; Yan et al., 2000; Haney et al., 2001; Singh et al., 2007; Shen and Lutkenhaus, 2009; Król et al., 2012; Szwedziak et al., 2012). The CTV also appears to be a key determinant of lateral interaction potential. We have designated the region containing the CTC and CTV (residues 366–382 of B. subtilis FtsZ) the ‘grappling hook peptide’ (GHP) to reflect its role in mediating interactions between modulatory proteins in both the cytoplasm and the plasma membrane.
While recent work has begun to clarify the role of the core, the CTC and the CTV in FtsZ assembly and Z-ring integrity, the function of the unstructured C-terminal linker has remained largely mysterious. The CTL spans the gap between the core and the CTC, suggesting it may be required to maintain a flexible connection between FtsZ protofilaments and modulatory proteins bound to the GHP of FtsZ subunits. Phylogenetic analysis of FtsZs across domains of life showed that there is little sequence conservation of the CTL even among bacteria from the same taxonomic phylum.
Intriguingly, while the length of the CTL can range 2–330 residues (Vaughan et al., 2004), it is ∼ 50–100 residues long in the majority of FtsZs sequenced that contain all structural domains. An exception to this rule is FtsZs from the Alpha-proteobacteria, in which linkers are significantly longer, between 119 amino acids (aa) in Rickettsia prowazeckii and 251 aa in Agrobacterium tumefaciens. Among the model organisms for bacterial cell division, the CTL of B. subtilis and E. coli is ∼ 50 residues, while that of Caulobacter crescentus, a member of the Alpha-proteobacteria, is 176. Notably, the CTL is irresolvable on crystal structures of bacterial FtsZ (Leung et al., 2004; Oliva et al., 2007; Haydon et al., 2008; Läppchen et al., 2008; Raymond et al., 2009; Matsui et al., 2012), and therefore has been presumed to be an intrinsically disordered peptide (IDP) (Erickson et al., 2010). Previous work suggests the CTL is flexible with a contour length of 17 nm, and an average end-to-end distance of 5.2 nm for the relaxed peptide (Ohashi et al., 2007).
Here we employ genetics, cell biology, and biochemistry to clarify the role of the FtsZ CTL. Our data indicate the CTL is required for protofilament assembly in vitro and formation of a curved ring in vivo, suggesting a role for the CTL in the geometry of longitudinal interactions between individual subunits. While changes in linker length were generally tolerated in vitro, increases in length greater than 100 residues led to disruptions in the frequency and position of FtsZ assembly in vivo. Changes in the linker's primary sequence had little impact on assembly in vivo or in vitro when retained as an IDP, whereas replacing the linker with an inflexible domain from human beta-catenin was not tolerated. Notably, our findings in B. subtilis closely mirror those from the Erickson laboratory working in E. coli (Gardner et al., 2013), who similarly determined CTL sequence to be unimportant while length being crucial. Taken together our findings have significant implications for our understanding of FtsZ assembly dynamics at the molecular level, as well as the role of various regions of FtsZ in assembly of a productive cytokinetic ring.
The FtsZ linker is required for normal function in B. subtilis
Deletion analysis indicated that the linker region of FtsZ was essential for efficient assembly in vitro and ring formation in vivo. As a first step towards assessing the role of the linker in FtsZ activity, we generated an FtsZ mutant in which the linker was deleted, leaving only the core followed immediately by the CTC and CTV domains (FtsZ ΔCTL50) (Fig. 1). Transmission electron microscopy (TEM) of wild-type FtsZ assembled in our standard FtsZ polymerization buffer [50 mM 2-(N-morpholino)ethanesulfonic acid (MES), pH 6.5, 50 mm KCl, 2.5 mM MgCl2, 1 mM EGTA, 1 mM GTP] indicated the present of filament bundles and sheets (Fig. 2A and B), consistent with our previous findings (Buske and Levin, 2012). In contrast, FtsZ ΔCTL50 appeared impaired for filament formation and filament bundling. By TEM we observed primarily short oligomers (< 25 nm) interspersed among single- or double-stranded filaments of FtsZ ΔCTL50 5 min after initiation of polymerization by GTP (Fig. 2C). Fifteen minutes after addition of GTP, FtsZ ΔCTL50 filaments became more pronounced and longer, suggesting the rate of filament formation was delayed compared with wild type (WT), though many small oligomers could still be observed also suggesting gross filament formation was diminished (Fig. 2D).
FtsZ ΔCTL50 assembly was GTP-dependent, as we observed a modest increase in 90° light scattering signal following the addition of 1 mM GTP (Fig. 2E). Here, as previously, we use light scattering to measure increase inpolymer mass and FtsZ bundling (Mukherjee and Lutkenhaus, 1999; Buske and Levin, 2012), and define assembly as a global term to include head-to-tail polymerization into protofilaments and lateral interactions. Consistent with its limited assembly in TEM micrographs, the FtsZ ΔCTL50 light scattering signal was 37-fold lower than that of wild-type FtsZ in the presence of GTP. Similarly, the critical concentration (Cc) of FtsZ ΔCTL50 was 1.08 μM compared with 0.68 μM for WT, suggesting a reduction in cooperative assembly. We also observed that time needed to reach a peak light scattering value was on the order of minutes whereas wild-type FtsZ was on seconds. The delay in filament formation suggested that head-to-tail assembly of filaments was impaired without the FtsZ linker. The GTPase activity of FtsZ ΔCTL50 was 1.48 GTP FtsZ−1 min−1, significantly lower than the 3.82 GTP FtsZ−1 min−1 measured for WT (Table 1). Notably, CD spectra indicated that FtsZ ΔCTL50 folded normally despite the loss of the linker, eliminating a trivial explanation for its defect in polymerization (Fig. S1).
Table 1. FtsZ GTPase turnover rates and critical concentrations
GTP FtsZ−1 min−1
FtsZ Polymers at pH 6.5, 50 mM KCl
Filament bundles and rings
In vivo, FtsZ ΔCTL50 was unable to support either FtsZ assembly or division. Taking advantage of a strain that allowed for depletion of wild-type FtsZ, we expressed FtsZ ΔCTL50 as the sole copy of FtsZ in B. subtilis cells under the control of an IPTG – inducible promoter at the amyE locus (see Experimental procedures). Expression FtsZ ΔCTL50 in the absence of wild-type FtsZ led to the rapid demise of cells: cell density as measured by absorbance at 600 nm (A600) declined ∼ 2-fold 4 h after reaching a peak density of 0.4 following induction of FtsZ ΔCTL50 and depletion of wild-type FtsZ (Fig. 3A). It should be noted we were only able to observe growth when cells were grown in inducer at a starting A600 twice that of all other constructs used in the study, suggesting that expression of FtsZ ΔCTL50 had an immediate and deleterious impact on cell viability.
Immunofluorescence microscopy of FtsZ ΔCTL50 cells 2.5 h post induction in the absence of wild-type FtsZ supported the idea that the linker was required for proper FtsZ assembly and the formation of an intact cytokinetic ring. Cells were filamentous (> 80 μm, the length of our microscopy field) and FtsZ was mostly in punctate dots (88% of cells) throughout the cytoplasm in contrast to the Z rings of cells expressing wild-type FtsZ (Fig. 2F, arrow). Strikingly, in 30% of cells FtsZ ΔCTL50 formed long and relatively straight filaments that extended longitudinally through the cell similar to what we observed in vitro and suggested defects in the geometry of FtsZ ΔCTL50 assembly. Coexpression of FtsZ ΔCTL50 with wild-type FtsZ resulted in a dominant lethal phenotype and loss of cell viability when assessed by dilution plating assay (Fig. S2).
Together, these findings suggested the linker played a critical role in FtsZ filament assembly counter to previous reports indicating that E. coli FtsZ mutants lacking the entire C-terminal region (CTR) readily assemble in vitro (Wang et al., 1997; Yu and Margolin, 1997). In light of this discrepancy, we examined the ability of B. subtilis and E. coli FtsZ core constructs lacking the entire CTR (B. subtilis residues 316–382) to assemble in vitro under our standard conditions (Fig. S4). Somewhat surprisingly, both the B. subtilis and E. coli core domains were defective in both protofilament formation and lateral interactions, counter to the earlier studies (Wang et al., 1997; Yu and Margolin, 1997; Singh et al., 2007). A potential explanation for this discrepancy may be buffer conditions: while earlier studies employed 10 mM MgCl2 or CaCl2, MgCl2 was at 2.5 mM in our standard conditions. Divalent cations have been shown to promote the formation of stabilizing lateral interactions between protofilaments most likely through charge shielding of the core (Mukherjee and Lutkenhaus, 1999).
FtsZ function is independent of its linker sequence
We next determined if the linear sequence of the FtsZ linker was critical for function. The linker exhibited little or no conservation between even closely related bacterial species, suggesting that its presence and potentially its flexible nature, but not its precise sequence, were critical for function. To explore this possibility we generated an FtsZ mutant in which the sequence of linker was randomly scrambled (FtsZ CTLScr), keeping the same aa as wild-type B. subtilis FtsZ but changing their order. In addition we swapped the 50-residue B. subtilis linker with the equivalent 50 residues from the E. coli linker to yield a linker of 54 residues (CTLE) and with the first 46 aa of the 251 residue long linker from A. tumefaciens FtsZ to yield a linker of 50 residues (CTLA50) (Fig. 1).
Our data suggested that linker sequence had a small impact on FtsZ assembly in vitro. FtsZ CTLScr showed no evidence of compromised protofilament assembly and readily formed single stranded protofilaments and small bundles 3–5 filaments thick as seem by TEM (Fig. 4B). In agreement with TEM data, the 90° light scattering assays of FtsZ CTLScr indicated robust, GTP-dependent assembly to ∼ 60% that of FtsZ WT (Fig. 4E). This level of assembly was consistent with the smaller-than-wild-type bundles we observed by EM but significantly (95–98%) higher than that of FtsZ mutants that were incapable of forming lateral interactions (Buske and Levin, 2012). Evidence of bundling was also seen by a lowered GTPase rate of 2.68 GTP FtsZ−1 min−1. The Cc of CTLScr was 0.88 μM, similar to the wild-type value of 0.68 μM (Table 1).
When FtsZ CTLScr was expressed as the sole copy in B. subtilis, cells were able to divide normally and Z rings localized to midcell as WT (Fig. 4F). Here, and in all cells types that displayed the ability to form Z rings, we employed measurement of the length-to-ring (L/R) ratio as proxy for cell size. As some cell types became filamentous, using interseptal distance underestimated the average length of cells in a given population due to an inability to distinguish between partial and completed septa in these longer cells. FtsZ CTLScr-expressing cells had an average 5.48 μm/Z ring compared with 5.82 μm/Z ring for WT (Fig. 3B).
Replacing the native linker with heterologous sequences had modest impacts on FtsZ assembly, particularly lateral interactions between single stranded protofilaments. Both FtsZ CTLE and FtsZ CTLA50 readily assembled in the presence of GTP (Fig. 4C and D); however, the structures appeared to be primarily single stranded by TEM. Similarly, the light scattering signals for CTLE and CTLA50 were ∼ 62-fold and 37-fold lower than FtsZ WT, respectively, consistent with a reduction in lateral interactions and the formation of higher order structures. Finally, the GTPase activity of CTLE and CTLA50 was also elevated relative to wild-type FtsZ, 5.04 GTP FtsZ−1 min−1 and 5.51 GTP FtsZ−1 min−1 for CTLE and CTLA50 (Table 1), a phenotype that is also consistent with a reduction in lateral interaction potential (Buske and Levin, 2012). The apparent Cc was near WT for both chimeras.
In vivo, FtsZ CTLE and FtsZ CTLA50 all appeared to support normal division, despite their reduction in lateral interaction potential in vitro. Expression of FtsZ CTLE and FtsZ CTLA50 in the absence of wild-type FtsZ fully supported normal division and FtsZ localization (Fig. 4F) with near wild-type L/R ratios (Fig. 3B). Cells expressing these FtsZ chimeras displayed growth rates near that of WT (Fig. 3A). Some irregular cell wall staining seen in these mutants we attributed to lysozyme treatment for immunofluorescence microscopy (IFM), as we observed similar staining in wild-type cells (Fig. S3A). The ability of these mutants to support normal division despite reduced lateral interaction potential was counter to our previous findings that ability to form lateral FtsZ interactions in vitro affected cell division in vivo with a CTV mutant (Buske and Levin, 2012). This finding will be addressed in the Discussion.
Linker length is an important determinant for normal FtsZ function
To determine the role of linker length in FtsZ function we generated FtsZ chimeras of different length and examined the impact of these changes on FtsZ assembly in vitro and in vivo.
Reducing the length of the native FtsZ linker by deleting the last 25 residues (FtsZ ΔCTL25), significantly altered FtsZ assembly in vitro. Purified FtsZ ΔCTL25 formed short, irregular protofilaments with little evidence of lateral interactions (Fig. 5B). Consistent with reduced assembly, FtsZ ΔCTL25 assembly peaked at a level ∼ 79-fold lower than wild-type FtsZ in 90° light scattering assays (Fig. 5E and F). FtsZ ΔCTL25 hydrolysed GTP at a rate of 4.66 GTP FtsZ−1 min−1, higher than wild-type FtsZ and FtsZ CTLScr, presumably due to a reduction in stabilizing lateral interactions. Interestingly, the Cc for FtsZ ΔCTL25 was 0.30 μM, about half that for WT (Table 1).
Surprisingly, despite significantly altering FtsZ assembly dynamics in vitro, reducing the length of the linker by 50% had little effect on FtsZ ring formation or division in vivo. Expression of FtsZ ΔCTL25 in B. subtilis fully supported Z-ring assembly and cell division. Cells were able to grow at a rate near that of WT (Fig. 3A), and FtsZ ΔCTL25 assembled into apparently normal FtsZ rings that were appropriately localized to midcell (Fig. 5G). The L/R ratio of FtsZ ΔCTL25-expressing cells was also near WT at 5.62 μm/Z ring (Fig. 3B).
In a complementary set of experiments we generated FtsZ chimeras with longer linkers by swapping the native linker sequence with either the entire linker (FtsZ CTLA249) from A. tumefaciens (FtsZ CTLA249) or the first 96 residues (FtsZ CTLA100) to make a total linker length of 100 residues (FtsZ CTLA100) (Fig. 1). Both FtsZ CTLA249 and FtsZ CTLA100 assemble into single-stranded protofilaments in the presence of GTP as observed by TEM (Fig. 5C and D). However, there were few, if any, bundles apparent in any field of view, suggesting a reduction in lateral interactions. Consistent with this view FtsZ CTLA249 and FtsZ CTLA100 exhibited ∼ 43 and ∼ 33-fold decreases in signal relative to wild-type FtsZ in 90° light scattering assays (Fig. 5E and F). The apparent critical concentrations for both chimeras were near WT; FtsZ CTLA249 had a Cc of 0.80 μM and FtsZ CTLA100 a Cc of 0.78 μM. GTPase activities were 4.40 and 5.09 GTP FtsZ−1 min−1 for FtsZ CTLA249 and FtsZ CLTA100 respectively (Table 1).
Despite the near wild-type in vitro assembly dynamics of the two chimeras, only FtsZ CTLA100 was able to support division in vivo. Cells expressing FtsZ CTLA100 were essentially WT with regard to growth rate, FtsZ ring assembly, and localization with a wild-type L/R ratio of 5.44 μm/Z ring (Fig. 3). In contrast, cells expressing FtsZ CTLA249 as their only copy of FtsZ were unable to divide and formed extremely long filaments with what appeared to be partial invaginations of the cell wall. Cell growth was significantly reduced as A600 reached a peak value of ∼ 0.5 4 h post-induction (Fig. 3A). Cells expressing FtsZ CTLA249 were also not viable in a plating assay, though coexpression with wild-type FtsZ did restore some growth (Fig. S2B). Although FtsZ rings were present in FtsZ CTLA249 cells, their localization appeared somewhat irregular, with closely spaced rings frequently followed by longer lengths of cells with no rings at all (Fig. 5G). The L/R of FtsZ CTLA249 cells was 4.66 μm/Z ring, which was not statistically significantly different from cells expressing wild-type FtsZ (Fig. 3B).
The FtsZ linker must be flexible and unstructured
To determine if linker flexibility and disorder was essential for FtsZ function, we replaced the linker from B. subtilis FtsZ with residues 398–455 from the human beta-catenin protein (FtsZ CTLH) (Fig. 1). This region of beta-catenin was selected both for its rigid nature – it encompasses ∼ 4 of the 12 alpha-helical Armadillo repeats from the protein – and for its neutral charge, to eliminate potentially interfering effects from electrostatic interactions (Xing et al., 2008).
Our data suggested that linker flexibility was critical for proper FtsZ function both in vitro and in vivo. We observed only small oligomers of FtsZ CTLH by TEM and there was only a negligible increase in signal following the addition of GTP to a 90° light scattering assay (∼ 136 times lower than FtsZ WT) (Fig. 6B and C). Although it was unable to assemble to a significant degree, FtsZ CTLH hydrolysed GTP at a rate of 0.71 GTP FtsZ−1 min−1. The apparent Cc for FtsZ CTLH was 0.70 μM (Table 1). Together, these findings suggested that FtsZ CTLH was still able to bind GTP but was significantly impaired for hydrolysis. We could not attribute assembly defects toward any large changes in FtsZ subunit secondary structure as CD spectra of FtsZ CTLH was nearly identical to that of WT (Fig. S1).
Consistent with its impaired assembly in vitro, FtsZ CTLH was unable to support division in the absence of wild-type FtsZ (Fig. 6D, bottom panel). Cells expressing FtsZ CTLH in the absence of wild-type FtsZ were extremely filamentous and unable to divide. Immunofluorescence microscopy indicated that instead of forming a coherent ring, FtsZ CTLH localized in small punctae that were distributed randomly throughout the filaments (Fig. 6D, bottom panel). This punctate localization pattern was consistent with the inability of FtsZ CTLH to assemble in vitro (Fig. 6B). We also observed some cross-walls in cells that we attributed to remnants of previous cell divisions before wild-type FtsZ was depleted, as we saw similar cross-walls in cells in which no FtsZ was expressed (Fig. S3B). Expression of FtsZ CTLH also had a deleterious effect on growth as the growth curve reveals cells only reached a peak A600 of ∼ 0.2 3 h after induction followed by a slow 1.7-fold decrease in cell density over 4 h (Fig. 3A). Coexpression with wild-type FtsZ was also dominant lethal as indicated by dilution plating (Fig. S2B).
Previous studies of FtsZ have largely ignored its C-terminal linker, in large part because this region is irresolvable in crystal structures. Our data suggest the presence of a flexible and disordered C-terminal linker is critical for protofilament assembly and the architecture of the cytokinetic ring. At the same time, the specific sequence composition and/or length of the CTL appear to be less critical, with FtsZ assembly and Z-ring formation being largely WT in B. subtilis expressing FtsZ variants with heterologous CTLs up to 100 residues in lengths. Below we discuss the implications of these data in the context of our current understanding of FtsZ assembly and the ultrastructure of the FtsZ ring.
An intrinsically disordered linker is essential for FtsZ assembly and Z-ring formation
Somewhat unexpectedly, our data support a role for the flexible linker not only in vivo, but also in longitudinal interactions between FtsZ subunits in vitro. FtsZ ΔCTL50 exhibits a long delay in protofilament assembly (Fig. 2D and E) and an elevated critical concentration (Table 1), suggesting the linker is needed for efficient incorporation of subunits into a filament. Its delayed assembly dynamics provide a potential explanation for the long straight filaments formed by FtsZ ΔCTL50 in vivo (Fig. 2F), as do aberrant interactions with modulatory proteins via its C-terminal GHP. At the same time, we cannot discount possibility that in the absence of the linker, the GHP itself sterically interferes with subunit interactions. A role for the unstructured linker in co-ordinating longitudinal interactions between FtsZ subunits assembly is supported by data indicating that a chimeric FtsZ in which the linker has been replaced with a helical repeat (FtsZ CTLH) is unable to assemble in vitro or in vivo (Fig. 6).
Conservation of an intrinsically disordered peptide
The inability of the helical repeat to functionally replace the FtsZ linker suggests that the ability of the linker to behave as an IDP is a key determinant of linker functionality. Regardless of primary sequence, our FtsZ CTLScr, FtsZ CTLE, and FtsZ CTLA50 chimeras support cell division (Fig. 4), and all maintain disorder when run through multiple secondary structure predictors (Rost et al., 2004; Ishida and Kinoshita, 2008). Importantly, the lack of sequence conservation and the ability of highly divergent linkers to substitute for one another in vivo strongly argue against the linker as a site for interaction with modulatory proteins.
Although disorder is a key feature of the functional FtsZ linker, there appear to be some constraints on its size in the cell as a large increase in linker length has a negative impact on B. subtilis FtsZ assembly in vivo. While FtsZ CTLA249 assembles into ring-like structures in vivo (Fig. 5G), Z-ring localization was abnormal (4.66 μm length-to-ring ratio) and cells were unable to constrict. Specifically, why FtsZ CTLA249 fails to support division in B. subtilis, despite its relatively normal in vitro assembly dynamics, is not clear. One possibility is the large linker interferes with interactions between FtsZ and modulatory proteins. Alternatively, the longer linker may provide too much freedom to FtsZ subunits, disrupting the ultrastructure of the FtsZ ring. It remains to be seen why such long linkers are present in FtsZs from the Alpha-proteobacteria, where exceptionally long (119–330 aa) are the norm and where different sets of FtsZ-interacting proteins exists.
Our results reinforce previous work implicating C-terminal charge as an important determinant of lateral interactions between FtsZ protofilaments. While CTLE and CTLA50 have little impact on protofilament assembly and Z-ring formation, both carry net positive charges well below WT (wild-type FtsZ linker net charge +4; CTLE and CTLA50 net charge of +1.5 and −0.3 respectively) and display reduced lateral interaction potential (Fig. 4C–E). The FtsZ CTLScr chimera, which retains the same charge composition as wild-type FtsZ, readily forms bundled structures in vitro, although its light scattering signal is only half that of WT (Fig. 4E), suggesting that the position of charge within the linker may also contribute to lateral interactions. These findings are consistent with data indicating that CTV charge, as well as buffer pH and salt concentration, have a strong impact on FtsZ's lateral interaction potential, most likely by shielding negatively charged core domains on different protofilaments from one another (Pacheco-Gómez et al., 2011; Buske and Levin, 2012).
At the same time, the finding that FtsZ CTLE and CTLA50 are essentially WT for localization and division, despite the absence of lateral interactions in vitro, is somewhat surprising based on our previous work suggesting that loss of lateral interaction potential, via changes in CTV charge, significantly undermines the integrity of the FtsZ ring (Buske and Levin, 2012). One explanation for the more severe phenotype of the CTV mutant (Bs FtsZ CTVE) is that the CTV, like the CTC, is required for interaction between FtsZ and a subset of modulatory proteins in vivo, and that loss of these interactions negatively impacts the stability of the FtsZ ring.
Alternatively, in vivo, CTV charge may be a more important determinant of lateral interaction potential than linker charge, either through a better ability to access adjacent protofilaments or its proximity to the CTC and associated modulatory proteins. In support of the former, a recent crystal structure showed that the CTV was bound to FtsA from Thermotoga maritima (Szwedziak et al., 2012). Similarly, it was also recently reported that mutations in CTV residues 377, 378, 380 and 381 significantly reduce interactions between FtsZ and the bundling protein SepF in B. subtilis (Król et al., 2012). Monitoring cell division in cells expressing the different FtsZ CTL and CTV mutants in the absence of these modulatory proteins will help clarify such questions.
The FtsZ linker is a flexible tether
We propose that the primary role of the linker is to allow FtsZ to be anchored in the membrane via interactions between its GHP and modulatory proteins. At the same time, the linker permits FtsZ filaments to maintain a sufficient distance from the membrane for protofilaments to form the curved structure necessary for assembly of the Z ring and division machinery needed for cytokinesis to proceed along the short axis of the cell, similar to what has been proposed by Erickson et al. (2010) (Fig. 7A). Based on the near-linear localization pattern of the FtsZ ΔCTL50 construct, linkerless FtsZ filaments instead remain in close proximity to the membrane unable to respond to its curvature in a manner that allows Z ring formation (Fig. 7B). A structural role for the linker as a flexible tether is consistent with our finding that relatively large changes in linker length are tolerated both for FtsZ assembly in vitro and Z-ring formation in vivo. FtsZ with linkers ranging in lengths from 25 residues up to 100 residues appear to be WT for both assembly and cell division. The Erickson lab has reported a similar finding for E. coli FtsZ (Gardner et al., 2013).
A flexible linker may also be essential to allow FtsZ to adopt a conformation that allows it to position the CTV in a way that promotes lateral filament interactions, likely with the core of adjacent filaments and similar to what we previously proposed (Buske and Levin, 2012) (Fig. 7A). In the case of native B. subtilis FtsZ, this can be accomplished in the absence of modulatory proteins due to its proclivity toward forming FtsZ bundles. In FtsZs where filaments do not bundle, the interaction between the GHP and modulatory proteins that promote bundling can be positioned near neighbouring filaments through the linker.
While the precise nature of the mechanisms responsible for initiating FtsZ assembly at the nascent division site, modulating the architecture of the FtsZ ring, and driving constriction of the ring at the beginning of cytokinesis remain to be determined, our findings together with those of the Erickson lab (Gardner et al., 2013), reinforce the need to consider the role of the linker in all aspects of FtsZ dynamics.
All B. subtilis strains are derivatives of the strain JH642 (Perego et al., 1988). Cloning and genetic manipulation were performed using standard techniques (Harwood et al., 1990; Sambrook and Russell, 2001). All cloning was done using the E. coli strain AG1111 derivative PL930 (Wang and Lutkenhaus, 1993). PL930 contains the low copy plasmid pBS58 expressing E. coli ftsQAZ, which facilitates cloning of B. subtilis FtsZ. Vent DNA polymerase was used for PCRs (New England Biolabs). Cells were grown in Luria–Bertani (LB) medium at 37°C unless otherwise noted. Antibiotics were used at the following concentrations: ampicillin = 100 μg ml−1, spectinomycin = 100 μg ml−1, chloramphenicol = 100 μg ml−1. Strains and plasmids used in this study are listed in Table 2. Primers used are listed in Table S1.
Table 2. Bacterial strains and plasmids used in this work
The pPJ15 plasmid was made by Quick Change II site-directed mutagenesis (Agilent Technologies) of pPJ1 plasmid to delete B. subtilis FtsZ aa 316–365. All other FtsZ chimeras were constructed as follows. Novel restriction sites were introduced into the B. subtilis FtsZ sequence by site-directed mutagenesis (BamHI after aa 315 and XmaI before aa 366) to yield the pPJ19 plasmid. This resulted in the aa pairs GS and PG at the beginning and end of the FtsZ linker. Gene sequences were PCR amplified, restriction digested, and ligated into the pPJ19 plasmid between the BamHI and XmaI sites. FtsZ ΔCTL25 was made by PCR amplification of B. subtilis FtsZ aa 1–338 and ligating into pPJ19. Including the XmaI site, this resulted in a linker length of 25 aa. FtsZ CTLScr was created by assigning each linker codon a number and running it through a random number generator (http://www.random.org) to yield the resulting sequence: KQSTNTPQEKQPSIKVRRQQEVKPSLQKQENTVHIRRHSNAPTSDPPEEQ. A synthetic oligonucleotide encoding this sequence was made (Integrated DNA Technologies) and PCR amplified to insert into pPJ19. Including the restriction sites, the total linker length was 54 aa. The E. coli linker sequence (aa 318–366) was amplified from genomic DNA from strain MG1655 DNA and ligated into pPJ19. Including the restriction sites, the total linker length was 52 aa. FtsZ CTLH was made through the creation of a gBlocks™ Gene Fragments synthetic double-stranded oligonucleotide encoding human beta-catenin residues 398–455 (Integrated DNA Technologies). The oligonucleotide was then restriction digested and inserted into pPJ19. Agrobacterium tumefaciens FtsZ1 linker sequences were amplified from genomic DNA from strain A136 (gift of W. Margolin). FtsZ CTLA249 contained aa 322–566 from A. tumefaciens FtsZ1 in addition to extra residues from the novel restriction sites resulting in a total linker of 249 residues. FtsZ CTLA100 contained aa 322–417 from A. tumefaciens FtsZ1 plus restriction sites yielding a linker total length of 100 aa. FtsZ CTLA50 contained aa 322–367 from A. tumefaciens FtsZ1 along with restriction sites producing a linker of 50 residues.
pPJ27-35 were constructed by ligating restriction-digested full-length FtsZ chimera insert into the digested pDR67 plasmid with forward primers containing the B. subtilis ribosome binding site. All DNA sequences were confirmed by Sanger sequencing.
Bacillus subtilis strains PL3400-07 were constructed as previously described (Buske and Levin, 2012). Briefly, these strains allow depletion of wild-type FtsZ and expression of B. subtilis FtsZ variants. They were constructed by first transforming pPJ27-35 plasmids into wild-type JH642 cells. This allowed the insertion of a copy of FtsZ chimeras with the B. subtilis ribosome binding site upstream of the start codon at the amyE locus under control of the isopropyl IPTG-inducible Pspac promoter. Chromosomal DNA from the resulting strains was then transformed into competent PL2084 cells (JH642 thrC::Pxyl-ftsZBs, ftsZBs::spc, xylA::tet), which have their only copy of native FtsZ expressed from a xylose-inducible promoter at the thrC locus. This strain permits depletion of wild-type FtsZ and is xylose-dependent for normal growth (Weart and Levin, 2003).
Growth conditions were as follows. Bacillus subtilis strains PL3171 (FtsZ WT) and PL3400-3407 (FtsZ variants) were first grown from single colony overnight in LB containing spectinomycin and xylose to a final concentration of 0.5%. The next day, cells were diluted 1:100 in fresh LB medium supplemented with 0.5% xylose and grown to an A600 of 0.4–0.5, at which point 1 ml of cell culture was harvested by centrifugation. Cells were then washed twice with a fresh 1 ml of LB medium and then diluted 1:100 into a fresh 20 ml of LB medium plus isopropyl IPTG to a final concentration of 0.1 mM to induce expression of FtsZs. IPTG concentration was determined by titration in PL3171 to match FtsZ expression to that of wild-type JH642 cells. A600 measures were taken every 30 min to monitor cell growth, and after 2.5 h post-induction with IPTG cells were harvested and prepared for immunofluorescence.
Immunofluorescence was performed as described previously (Buske and Levin, 2012) using 2.5% paraformaldehyde and 0.04% glutaraldehyde. An Olympus BX51 microscope with Chroma filters and a Hamamatsu OrcaERG camera were used for image capture. Images were processed using Openlab version 5.2.2 (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 per strain. Bacillus subtilis FtsZ was detected using affinity-purified polyclonal rabbit anti-FtsZ serum (Levin and Losick, 1996) in combination with goat anti-rabbit serum conjugated to Alexa488 (Life Technologies). Cell walls were visualized with wheatgerm agglutinin conjugated to tetramethylrhodamine (Invitrogen). Nucleoids were stained with DAPI.
Determination of cell length/FtsZ ring ratio
The cell length/FtsZ ring (L/R) ratio was calculated as the sum total length of a population of cells divided by the number of FtsZ rings in that population as described previously (Weart et al., 2007).
Strains 3171, 3400, 3404, and 3405 were used for dilution plating. Cells were grown overnight in LB medium containing ampicillin at 30°C, and the following morning, cultures were first grown from single colony overnight in LB containing spectinomycin and xylose to a final concentration of 0.5%. The next day, cells were diluted 1:100 in fresh LB medium supplemented with 0.5% xylose and grown to an A600 of 0.4–0.5, at which point 1 ml of cell culture was harvested by centrifugation. Cells were then washed twice with a fresh 1 ml of LB medium upon which serial dilutions from 10−1 to 10−8 were made in LB medium. Ten microlitres of each dilution was then plated in series using a multichannel pipette onto pre-warmed LB agar containing spectinomycin and 0.1 mM isopropyl IPTG with or without 0.5% xylose. Liquid cultures were allowed to dry on the plates at room temperature for upwards of 6 h, and then plates were incubated at 37°C overnight. The plates were then imaged, and relative growth was qualitatively assessed.
FtsZ variants were cloned into the pET-21b(+) expression vector through E. coli strain AG1111. The resulting plasmids were miniprepped and freshly transformed into C41(DE3) cells (Miroux and Walker, 1996) and consequently used for expression of protein. No frozen stocks were used. Briefly, 1 l of LB medium was inoculated 1:100 with overnight culture from a single colony. Cells were grown at 37°C until A600 ∼ 0.6, and then cells were induced with isopropyl IPTG to a final concentration of 1 mM. Cells were grown for an additional 4 h at 37°C, and then cells were harvested by centrifugation, and cell pellets were stored at −80°C. Purification of FtsZs were performed as described previously (Buske and Levin, 2012) with the following modifications. FtsZ CTLE, CTLH CTLA249, CTLA100, and CTLA50 were all precipitated with one ammonium sulphate cut at 30%. The FtsZ was further purified on a MT-20 column manually packed with UNOsphere™ Q beads (Bio-Rad) with a linear gradient of 50–500 mM KCl in 50 mM Tris, pH 8.5, 1 mM EDTA, 10% sucrose. Peak fractions were analysed by SDS-PAGE, pooled together, and dialysed overnight twice, first in 1 l of FtsZ dialysis buffer, pH 6.5 (50 mM MES, pH 6.5, 50 mM KCl, 2.5 mM MgCl2,1 mM EGTA, 10% sucrose), then in 1 l of FtsZ dialysis buffer, pH 7.5 (50 mM HEPES, pH 7.5, 50 mM KCl, 2.5 mM MgCl2, 1 mM EGTA, 10% sucrose). Protein preparations were concentrated in dialysis tubing using PEG 12 000, aliquoted, flash frozen on liquid N2, and stored at −80°C. Prior to use in an assay to remove any insoluble polmyers, FtsZ aliquots were thawed on ice, then spun at 80 000 r.p.m. in a TLA-100 rotor (Beckman) for 10 min at 4°C and the supernatant used. FtsZ was then quantified using Coomassie Plus™ Protein Assay Reagent (Fisher) with porcine tubulin (Cytoskeleton) as a standard.
90° angle light scattering assay
Light scattering experiments were performed essentially as described earlier (Buske and Levin, 2012) using a DM-45 spectrofluorimeter (Olis). Assembly reactions contained 5 μM FtsZ in assembly buffer (50 mM MES, pH 6.5, 50 mM KCl, 2.5 mM MgCl2, 1 mM EGTA, 1 mM GTP).
Electron microscopy was performed essentially as described (Buske and Levin, 2012). 3 μM FtsZ was used because it was found to be the best concentration in which to visualize single FtsZ filaments. FtsZ was assembled with GTP as for light scattering, and samples were visualized using a JEOL 1200EX transmission electron microscope. FtsZ filament lengths were measured using ImageJ, and data were compiled in Microsoft Excel and Kaleidagraph.
GTPase assay and critical concentration
GTPase activity was measured using the continuous, regenerative coupled GTPase assay of Ingerman and Nunnari (2005). Assays were conducted in buffer conditions identical to those used for light scattering. Each reaction included 1–8 μM FtsZ, 1 mM GTP, 1 mM phosphoenolpyruvate, 250 μM NADH, 80 units ml−1 lactose dehydrogenase, and 80 units ml−1 pyruvate kinase. A linear decline of absorbance at 340 nm for NADH was observed at 30°C for 5 min in a 96 well plate (path length calculated as 0.425 cm) using a SPECTRAmax Plus spectrophotometer (Molecular Devices). The raw data of absorbance per minute were then converted to activity using the extinction coefficient for NADH at 340 nm of 6220 M−1 cm−1 and the path length. The raw data were then exported to Microsoft Excel for analysis. The critical concentration was determined as the X-intercept of a plot of the GTP consumed per minute versus FtsZ concentration as has been done previously (Chen et al., 2012).
We thank Harold Erickson for sharing data and helpful discussions on the manuscript. We thank Bill Margolin for the generous gift of A. tumefaciens strain A136. We also thank Wandy Beatty for assistance with electron microscopy and all members of the Levin lab for their helpful discussions and feedback. This work was supported by a National Institutes of Health Public Health Service Grant GM64671 to PAL.