Abstract
The mechanisms that restrict peptidoglycan biosynthesis to the pole during elongation and re-direct peptidoglycan biosynthesis to mid-cell during cell division in polar-growing Alphaproteobacteria are largely unknown. Here, we demonstrate that although two of the three FtsZ homologs localize to mid-cell, exhibit GTPase activity and form co-polymers, only one, FtsZAT, is required for cell division. We find that FtsZAT is required not only for constriction and cell separation, but also for the termination of polar growth and regulation of peptidoglycan synthesis at mid-cell. Depletion of FtsZ in A. tumefaciens causes a striking phenotype: cells are extensively branched and accumulate growth active poles through tip splitting events. When cell division is blocked at a later stage, polar growth is terminated and ectopic growth poles emerge from mid-cell. Overall, this work suggests that A. tumefaciens FtsZ makes distinct contributions to the regulation of polar growth and cell division.
Introduction
The spatial and temporal regulation of cell division is a vital process across bacterial species with implications in the development of antimicrobial therapies [1]. The cell division process must coordinate membrane invagination(s), peptidoglycan (PG) biosynthesis and remodeling, and the physical separation of the two daughter cells, all while maintaining cellular integrity. Furthermore, cell division must be precisely regulated to be orchestrated with other key cell cycle processes including cell elongation, DNA replication, and chromosome segregation to ensure that each daughter cell is of sufficient size and contains a complete genome [2, 3].
To initiate bacterial cell division, the tubulin-like GTPase, FtsZ, polymerizes and forms a discontinuous ring-like structure at the future site of cell division [4-10]. The presence of FtsZ at mid-cell leads to the recruitment of many proteins that function in cell division, collectively called the divisome [11-14]. The divisome includes cell wall biosynthesis proteins, such as the penicillin-binding protein, PBP3, and FtsW, which contribute to PG biosynthesis and remodeling necessary to form new poles in daughter cells [11]. Once the divisome is fully assembled, FtsZ filaments treadmill along the circumference of the mid-cell, driving the Z-ring constriction [9, 10]. The movement of FtsZ filaments is correlated with the movement of enzymes that function in septal PG biogenesis. These finding are consistent with the notion that FtsZ not only recruits enzymes that function in PG biogenesis to mid-cell but also regulates their activities to promote proper cell wall biogenesis [15-17].
In most rod-shaped model organisms used to study cell division, a block in cell division leads to the production of long, smooth filamentous cells. This phenotype suggests that assembly or activation of some divisome components is necessary not only to enable the cells to divide but also to stop cellular elongation. Indeed, in Escherichia coli, FtsZ (along with the Z-ring stabilizing proteins FtsA, ZipA, and ZapA) has been proposed to have an early function in the switch from lateral PG biogenesis to mid-cell PG biosynthesis [18]. Following maturation of the divisome by recruitment of additional PG remodeling enzymes and cell division proteins, PG biosynthesis is coordinated with membrane invagination, enabling cells to constrict and separate [19].
Conversely to e.g. E. coli, polar growing rods in the alphaproteobacterial clade Rhizobiales exhibit branched morphologies when cell division is blocked [20-27]. Examination of the cell morphologies resulting from the block in cell division suggests that different types of branched morphologies arise [28]. Drug treatments that block DNA replication cause an early block in cell division, resulting in a “Y” morphology in which the branches are formed from existing growth poles [25, 26]. In contrast, antibiotics that target PBP3 cause mid-cell bulges and branches with some cells adopting a “T” or “+” morphologies [25, 27]. These observations suggest that polar-like PG synthesis is redirected to mid-cell when cell division is blocked at a later stage. The manifestation of two distinct phenotypes during early and late blocks in cell division suggests that divisome assembly and activation may contribute to termination of polar growth, onset of mid-cell PG biosynthesis, cell constriction, and ultimately cell separation.
In Agrobacterium tumefaciens, homologs of FtsZ and FtsA fused to fluorescent proteins localize at the growth pole during elongation and at mid-cell during division [27, 29, 30]. FtsZ was found to arrive at mid-cell considerably earlier than FtsA [30], indicating that FtsZ may be able to initiate Z-ring formation prior to FtsA recruitment to the divisome. This observation is consistent with the described order of divisome assembly in Caulobacter crescentus [31] and suggests that a distinctive time-dependent role of these proteins in cell division.
Here, we take advantage of the ability to deplete essential proteins in A. tumefaciens [32] to explore the function of cell division proteins FtsZ, FtsA, and FtsW in a polar growing alphaproteobacterium. Although the genome of A. tumefaciens encodes three FtsZ homologs, we find that only one, henceforth referred to as FtsZAT, is essential for cell survival. FtsZAT is required to recruit division proteins to mid-cell and likely regulates the activity of PG biosynthesis enzymes at mid-cell. In the absence of FtsZAT, cells not only fail to divide but are also unable to terminate polar growth. Depletion of either FtsA or FtsW also causes a block in cell division, but unlike FtsZAT depletion, growth at the poles is halted and instead, polar-like PG synthesis is redirected to mid-cell. These observations suggest that only FtsZ is required to terminate polar growth and initiate cell division-specific PG biosynthesis at mid-cell, whereas FtsZ, FtsA, and FtsW are exclusively required for cell division. Together these findings suggest that A. tumefaciens uses sequential regulation of cell division, a theme that is broadly conserved in bacteria.
Results and Discussion
FtsZAT is required for cell division and termination of polar growth
Agrobacterium tumefaciens contains three homologs of Escherichia coli’s FtsZ, Atu_2086, Atu_4673, and Atu_4215 (Figure 1A) [27]. E. coli FtsZ is comprised of three regions: the conserved N-terminal tubulin-like GTPase domain, a C-terminal linker (CTL), and a conserved C-terminal peptide (CTP), which anchors FtsZ to the membrane via interactions with FtsA [33]. Atu_2086 contains each of these domains out of which the GTPase domain and CTP share 52% and 67% identity to their respective domain in E. coli FtsZ, whereas the CTL is extended in length [27]. The gene encoding Atu_2086 is found in a putative operon with genes encoding DdlB, FtsQ, FtsA [34, 35] and is predicted to be essential for cell survival based on saturating transposon mutagenesis [36]. Atu_2086 localizes to mid-cell in wildtype (WT) pre-divisional cells (Figure 1B) [27, 29]; consistent with a role in cell division. Atu_4673 (called FtsZ1; consistent with the genome annotation) contains a complete GTPase domain with 49% identity to tubulin domain of E. coli FtsZ but lacks both the CTL and CTP [27]. Although Atu_4673 is not predicted to be required for cell survival based on saturating transposon mutagenesis [36], it localizes to mid-cell in pre-divisional cells, suggesting a possible role in cell division (Figure 1B). Atu_4215 (termed FtsZ3 in this work) contains a partial GTPase domain with 48% identity to the N-terminal portion of the E. coli FtsZ tubulin domain and lacks both the CTL and CTP [27]. FtsZ3 is not essential for survival of A. tumefaciens based on saturating transposon mutagenesis [36] and exhibits a diffuse localization pattern (Figure 1B). Together, these data suggest that Atu_2086 is the canonical FtsZ protein required for cell division, and this protein will be referred to as FtsZAT throughout this work (although it is annotated as FtsZ2 in the A. tumefaciens C58 genome [34, 35]).
To characterize the function of each FtsZ homolog, we constructed deletions of ftsZ1 and ftsZ3 and a depletion strain of ftsZAT. Since we were unable to construct a deletion of ftsZAT, we used a depletion strategy in which ftsZAT is present as a single copy under the control of an isopropyl β-D-1-thiogalactopyranoside (IPTG) inducible promoter at a neutral site in the chromosome [21, 32]. Using western blot analysis, we have confirmed the depletion of FtsZAT in the absence of IPTG (Figure 1-Figure Supplement 1A).
Deletion of ftsZ1 or ftsZ3 does not impact cell viability (Figure 1C), cell morphology (Figure 1D; Table 1; Figure 1-Figure Supplement 1B), microcolony formation (Figure 1D), constriction rate or position (Table 1) when compared to WT cells. Similarly, when FtsZAT is expressed in the depletion strain (labeled in Figures as +FtsZAT) the cells remain viable (Figure 1C), are similar in size to WT cells (Table 1), properly position constrictions (Table 1), and form microcolonies (Figure 1D). In contrast, depletion of FtsZAT (labeled in Figures as –FtsZAT) causes a marked decrease in cell viability (Figure 1C) and triggers the formation of large cells with complex branched morphologies (Table 1; Figure 1D). To quantify changes in morphology during depletion of FtsZAT, the cell area of at least 100 cells was calculated based on phase contrast images of cells acquired immediately after removal of the inducer (-FtsZAT 0 h), 8 h after removal of the inducer (-FtsZAT 8 h), and 14 h after removal of the inducer (-FtsZAT 14 h) (Table 1, Figure 1-Figure Supplement 1C). Initially, the FtsZAT depleted cells are similar to WT in cell size, but after 8 h of FtsZAT depletion the cell area has nearly doubled (Table 1, Figure 1-Figure Supplement 1C). Within 14 h of FtsZAT depletion, the average cell area has dramatically increased (Table 1, Figure 1-Figure Supplement 1C). Together, these results demonstrate that only the FtsZAT homolog is required for proper cell growth and division.
Deletion of ftsZ1 and ftsZ3 does not change the FtsZAT depletion phenotype
Since the ftsZ1 and ftsZ3 single deletions do not have an obvious impact on cell morphology, growth, or division, we constructed double and triple mutants to determine if there is an increasing effect when removing multiple ftsZ homologs. Double deletion of ftsZ1 and ftsZ3 does not cause a decrease in cell viability (Figure 2A, top panel), cell morphology (Table 1), or microcolony formation (Figure 2A, bottom panel). Furthermore, ΔftsZ1 ΔftsZ3 cells properly place constrictions and have an average constriction rate similar to WT (Table 1). Next, we introduced the ΔftsZ1, ΔftsZ3, and ΔftsZ1 ΔftsZ3 mutations into the ftsZAT depletion strain to determine if loss of multiple ftsZ homologs further aggravated the ftsZAT depletion phenotypes. The combination of the ftsZAT depletion strain with ΔftsZ1, ΔftsZ3, or ΔftsZ1 ΔftsZ3 mutations did not result in a further decrease in cell viability (Figure 2B, top panel) or a worsening of cell morphology (Figure 2B, bottom panel) when compared to FtsZAT depletion alone. Together, these results suggest that the FtsZ1 and FtsZ3 homologs do not have a major impact on cell division under the conditions tested.
ftsZ gene duplications have occurred independently in several alphaproteobacterial lineages and in chloroplasts and some mitochondria [37]. In most of the cases that have been studied, one FtsZ homolog plays a canonical role in cell or organelle division while the other plays a regulatory or specialized role. However, little is known about the roles of multiple ftsZs in certain alphaproteobacteria species. In both Rhizobium meliloti and Magnetospirillum gryphiswaldense, one of the FtsZs (containing a CTL and CTP similar to FtsZAT) is essential and the other (truncated after the GTPase domain similar to FtsZ1) is dispensable [38, 39]. In the case of M. gryphiswaldense, the truncated ftsZ is dispensable for division but important for biomineralization in this magnetotactic species under certain growth conditions. Similarly, it is possible that FtsZ1 or FtsZ3 may have important contributions to cell growth or division of A. tumefaciens in different environments as e.g. in its plant-associated life-style.
FtsZ1 requires FtsZAT to localize to mid-cell and to polymerize in vitro
Since FtsZ1 localizes to mid-cell (Figure 1B), we hypothesized that FtsZ1 may be a nonessential divisome component. To test this, we examined the localization of FtsZ1-sfGFP in both WT and the ftsZAT depletion strain (Figure 3). In WT and FtsZAT induced cells, FtsZ1-sfGFP does not localize in newborn cells but forms FtsZ-like rings at the future site of division in pre-divisional cells (Figure 3A, top and middle panel). This Z-like ring constricts to form a single focus in dividing cells. These observations suggest that FtsZ1 may be a divisome component despite the absence of a cell division phenotype in the ΔftsZ1 strain. To explore the possibility of interactions arising due to the loss of FtsZ1 and FtsZAT, we next visualized FtsZ1-sfGFP localization during the depletion of FtsZAT (Figure 3A, bottom panel). We pre-depleted FtsZAT for 4 h in liquid to avoid cell crowding caused by division events prior to sufficient FtsZAT depletion. Early during the depletion of FtsZAT, FtsZ1-sfGFP localizes in a FtsZ-like ring near mid-cell. However, as the FtsZAT depletion continues, FtsZ1-sfGFP rings and foci progressively fade away, demonstrating that localization of FtsZ1-sfGFP to mid-cell requires the presence of FtsZAT.
Since FtsZ1 is recruited to mid-cell by FtsZAT, we hypothesized that FtsZAT and FtsZ1 may form co-polymers. To first test the ability of FtsZAT and FtsZ1 to independently form polymers, each protein was purified and subjected to polymerization studies. Right angle light scattering assays of wildtype FtsZAT revealed that this protein exhibits a GTP-dependent increase in light scattering at concentrations above 2 µm, consistent with its polymerization (Figure 3B, blue lines). Negative stain transmission electron microscopy (TEM) confirmed that FtsZAT forms gently curved protofilaments in the presence of GTP (Figure 3C, left panel) and it rapidly releases inorganic phosphate suggesting that GTP is hydrolyzed (Figure 3D, blue lines; 4.7 ± 0.2 GTP min-1 FtsZ-1 at 8 µM FtsZAT, n=3). Surprisingly, we did not observe polymerization of wildtype FtsZ1, even at high protein concentrations either in light scattering (Figure 3B, red line), TEM (Figure 3C, center panel), or GTP hydrolysis assays (Figure 3D, red line).
In light of the dependence of FtsZ1 on FtsZAT for mid-cell localization, we next sought to determine if FtsZAT and FtsZ1 can form co-polymers. To conduct these experiments, FtsZ1-L71W and FtsZAT-L72W were purified to enable monitoring of protein polymerization using tryptophan fluorescence. The leucine to tryptophan mutation introduces a tryptophan on the surface of FtsZ that increases in fluorescence when it is buried in the subunit interface upon polymerization [40]. While wildtype FtsZAT (with no tryptophan) does not change in fluorescence on addition of GTP (Figure 3E, solid blue line), FtsZAT-L72W fluorescence increases rapidly after GTP addition reflecting polymerization (Figure 3E, dashed blue line). When wildtype FtsZAT is added to FtsZAT-L72W, bringing the total FtsZ concentration to 8 µM, fluorescence again increases, but then drops back to baseline upon complete consumption of GTP by this high concentration of FtsZ (Figure 3E, dotted blue line). Conversely, on its own or combined with wildtype FtsZ1, FtsZ1-L71W maintains a constant tryptophan fluorescence level before and after addition of GTP, consistent with our conclusion that it does not polymerize on its own (Figure 3E, red lines). Remarkably, tryptophan fluorescence increases when FtsZ1-L71W and FtsZAT are mixed, indicating that the FtsZ1-L71W is incorporated into polymers in the presence of FtsZAT (Figure 3C, purple dashed line). When FtsZAT-L72W is mixed with FtsZ1, fluorescence increases above the level observed for FtsZAT-L72W alone and drops to baseline faster than FtsZAT-L72W on its own, again indicating co-polymerization. Finally, equimolar concentrations of FtsZAT alone or mixtures of FtsZAT and FtsZ1 exhibit similar rates of GTP hydrolysis (Figure 3D) and form qualitatively similar polymers by TEM (Figure 3C, right panel). Together, these observations indicate that FtsZ1 cannot polymerize independently, but that FtsZAT and FtsZ1 form co-polymers with similar structure and GTP hydrolysis rates as FtsZAT polymers.
Though multiple FtsZs are present in a number of bacterial and chloroplast lineages, their co-assembly properties have only begun to be characterized. In contrast to our observations, each of the FtsZs of M. gryphiswaldense was able to independently polymerize in vitro, but they also appeared to directly interact, perhaps reflecting an ability to co-polymerize [39]. Chloroplast FtsZs from Arabadopsis thaliana are also able to co-polymerize and, at least under some conditions, to independently polymerize [41]. Conversely, one of the FtsZs from tobacco chloroplasts cannot polymerize on its own but promotes polymerization of its partner homolog [42]. Finally, the FtsZ pair from the chloroplasts of representative green and red algae co-polymerize into polymers with altered assembly dynamics from either homopolymer [43]. It is likely that in each of these cases, the assembly or co-assembly properties of the duplicated FtsZs have evolved to suit a niche regulatory function. We hypothesize the FtsZ1 from A. tumefaciens has low affinity for itself, but higher affinity for FtsZAT, limiting its homopolymerization but allowing for co-polymerization both in vitro and in cells. Since FtsZ1 cannot polymerize independently, FtsZAT must first polymerize at mid-cell after which FtsZ1 can be recruited by co-polymerization. The biological relevance of these biochemical and cell biological properties awaits further study.
FtsZAT depletion results in tip splitting events
Once we identified FtsZAT as the primary homolog involved in cell division we next analyzed the growth phenotype during FtsZAT depletion more carefully. Compared to FtsZAT induced cells (Figure 4A, top), observation of cells during FtsZAT depletion by time-lapse microscope reveals remarkable changes in cell morphology (Figure 4A, bottom; Movie 1). Early during the depletion of FtsZ, an ectopic pole forms near mid-cell. We hypothesize that this occurs due to the ability of the remaining FtsZ to identify the mid-cell and recruit PG biosynthesis machinery to that site. Both the original growth pole and the ectopic pole are growth-active, resulting in the presence of multiple growth poles. These growth poles are unable to terminate cell elongation and ultimately most growth active poles are split, leading to the accumulation of many growth active poles (Figure 4A, bottom; Movie 1) and the rapid increase in cell area until the cell lyses.
The branched morphology observed during FtsZAT depletion is in stark contrast to FtsZ depletion observed in other organisms. In species like E. coli and B. subtilis, which utilize laterally localized peptidoglycan biosynthesis during elongation, depletion of FtsZ results in long, smooth filamentous cells. We hypothesize that the branching morphology of the A. tumefaciens FtsZAT depletion strain can be attributed to polar elongation. During the block in cell division, the growth pole continues to grow and presumably recruits additional peptidoglycan biosynthesis proteins. This could lead to an over-accumulation of elongasome proteins causing the pole to split into two poles. A similar branching pattern has been characterized during typical growth of Streptomyces coelicolor [44]. In this polar growing bacterium, the established elongasome splits, leaving a small portion of the elongasome behind as growth continues. With time, the subpolar elongasome accumulates in size and eventually forms a new growth pole. Although the polar growth molecular mechanisms are not conserved between A. tumefaciens and S. coelicolor, the fundamental principle of tip splitting as a consequence of polar growth appears to be shared.
PopZ-YFP accumulates at growth poles in the absence of FtsZAT
In WT A. tumefaciens, deletion of popZ has been shown to cause ectopic poles and cells devoid of DNA, demonstrating a role in coordinating cell division with chromosome segregation [45, 46]. We hypothesize that PopZ-dependent coordination of cell division likely involves FtsZ. In WT, PopZ-YFP localizes to the growing pole during elongation and is recruited to mid-cell just prior to cell separation (Figure 4B, top panel) [45-47]. When FtsZAT is expressed in the ftsZAT depletion strain, PopZ-YFP has a similar localization pattern as in WT cells (Figure 4B, middle panel). When FtsZAT is depleted, PopZ-YFP stays at the growth poles and as tip splitting events lead to the production of new growth poles, PopZ-YFP appears to be split and retained at all growth active poles (Figure 4B, bottom panel). These observations indicate that FtsZ is required for mobilizing PopZ from the growth pole to mid-cell. Remarkably, both FtsZ and FtsA are mislocalized in the absence of PopZ, leading to the establishment of asymmetric constrictions sites and a broad range of cell lengths [45]. Together, these data suggest that the presence of both PopZ and FtsZ are important for proper positioning and functioning of the divisome.
In addition to its function in maintaining proper cell division, A. tumefaciens PopZ is also required for chromosome segregation and tethers the centromere of at least one chromosome to the growth pole [46]. Thus, we examined the DNA content of cells depleted of FtsZAT. In both WT cells and in conditions where ftsZAT is induced in the ftsZAT depletion strain, DNA labeled with Sytox orange is diffuse throughout most cells (Figure 4C, top two panels). In late divisional cells, true separation of nucleoids is observed indicating successful completion of chromosome segregation (Figure 4C, marked with an asterisk in the top two panels). In cells depleted of FtsZAT for both 8 and 14 h, DNA is diffuse throughout the elongated branches (Figure 4C, bottom two panels). The absence of distinct nucleoids may suggest that final stages of chromosome segregation are coordinated with cell separation as has been described for other bacteria including E. coli and C. crescentus [48].
To look more carefully at genomic content, we visualized YFP-ParB1, which serves as a proxy to track centromere partitioning in A. tumefaciens [46], in WT and ftsZAT depletion cells. In both WT cells and cells expressing FtsZAT in the ftsZAT depletion strain, a single YFP-ParB1 focus is present at the old pole in new cells generated by a recent cell division event (Figure 4D, top and middle panels). As the cells elongate, a second focus appears and translocates across the longitudinal axis to the growing pole (Figure 4D, top and middle panels). After 4 h of FtsZATpre-depletion, YFP-ParB1 foci can be seen at both poles, but when the cell fails to divide, a third focus of YFP-ParB1 appears and translocates along the longitudinal axis of the cell before taking a rapid turn toward a new ectopic pole formed from near mid-cell (Figure 4D, bottom panel). Next, we quantified the number of YFP-ParB1 foci relative to cell area (Figure 4E). In WT and FtsZAT expressing cells in the ftsZAT depletion strain, small cells have only a single focus of YFP-ParB1. This is followed by a transition period in which elongating cells accumulate a second focus of YFP-ParB1. Finally, the largest, pre-divisional cells have two YFP-ParB1 foci (Figure 4E). Cells depleted of FtsZAT for 8 h accumulate YFP-ParB1 foci as they increase in area (Figure 4E). Cells with an area larger than 3 μm2 all have at least 2 YFP-ParB1 foci, suggesting that chromosome replication is not blocked during FtsZ depletion. Furthermore, in larger cells additional YPF-ParB1 foci accumulate. These data suggest that cell division is not strictly required for the initiation of DNA replication in A. tumefaciens, although completion of chromosome segregation may be coordinated with cell division.
Loss of septal PG synthesis results in altered total PG composition
Since polar growth appears to continue in the absence of FtsZ (Figure 4A, bottom panel), we used fluorescent-D-amino acids (FDAAs), to probe sites enriched in peptidoglycan synthesis [49] during depletion of FtsZAT. In WT cells, FDAAs localize at a single pole in elongating cells and at mid-cell in pre-divisional cells (Figure 5A) [49]. As FtsZAT is depleted, FDAAs are targeted strictly to the poles, confirming that polar peptidoglycan synthesis is responsible for the observed increase in cell biomass after 8 h and 14 h of depletion (Figure 5A).
Since cells depleted of FtsZAT fail to terminate polar growth and do not produce septal peptidoglycan, we hypothesized that the peptidoglycan composition may reveal chemical signatures of peptidoglycan derived from polar growth. Thus, we characterized the peptidoglycan composition of both WT cells and the ftsZAT depletion strain in both the presence and absence of IPTG using ultra-performance liquid chromatography (UPLC) [50]. The major muropeptides found in WT A. tumefaciens PG and their quantification are shown in (Figure 5-Figure Supplement 1) and include monomeric (M), dimeric (D), and trimeric (T) muropeptides. The muropeptide composition and abundance is similar between WT cells, WT cells grown in the presence of IPTG, and the ftsZAT depletion strain grown in the presence of IPTG such that FtsZAT is expressed (Figure 5-Figure Supplement 1). These findings suggest that there are no major changes in PG composition due to IPTG and that the presence of IPTG leads to complementation in the ftsZAT depletion strain. In contrast, when the ftsZAT depletion strain is grown in the absence of IPTG for 14 h, marked changes in muropeptide composition are observed (Figure 5B-D). While the overall abundance of monomeric, dimeric, and trimeric muropeptides are not dramatically impacted (Figure 5C), the abundance of specific muropeptides is modified. When FtsZAT is depleted, there is a significant increase in monomeric disaccharide tripeptide (M3) and a decrease in the abundance of the monomeric disaccharide tetrapeptide M4 (Figure 5B). This observation is consistent with the possibility that the absence of FtsZAT leads to an increase in LD-carboxypeptidase activity, which would remove the terminal peptide from M4, leading to both a reduction in the levels of M4 and an increase in the abundance in M3. Following FtsZAT depletion, the overall degree of muropeptide crosslinking decreases (Figure 5D). In particular, there is a marked decrease in DD-crosslinkages, which are formed by the DD-transpeptidase activity associated with penicillin-binding proteins (PBPs). The dominant dimeric muropeptide formed in the presence of FtsZAT is D44, which contains a DD-crosslink; in contrast, the dominant dimer formed in the absence of FtsZAT is D33, which contains an LD-crosslink (Figure 5B). These data suggest that the activity of LD-transpeptidases is increased and the activity of PBP-mediated DD-transpeptidases is decreased during FtsZAT depletion. The increased pool of M3 may provide additional acceptor substrate for LD-transpeptidases to increase the production of D33 relative to D44. In addition, increased LD-carboxypeptidase activity could contribute to increase further the levels of D33 using D34 as a substrate.
The A. tumefaciens genome contains 14 LD-transpeptidases, 7 of which are specific to Rhizobiales. The Rhizobiales-specific LD-transpeptidase encoded by Atu_0845 (referred to here as LDTP0845) has been shown to localize to the growing pole in WT cells and has been hypothesized to contribute to polar growth [30]. This localization pattern was confirmed in both WT and FtsZAT induced cells (Figure 5E, top and middle panels). We find that LDTP0845 localizes at growth poles during depletion of FtsZAT (Figure 5E, bottom). This observation suggests that this LDTP0845 may contribute to changes in PG composition during FtsZAT depletion and supports a potential role for LD-transpeptidases in polar growth during elongation. The localization and function of putative periplasmic LD-carboxypeptidases in A. tumefaciens remain to be explored. Overall, these findings suggest that LD-carboxypeptidase and LD-transpeptidase activities are increased during FtsZAT depletion, indicating that these classes of enzymes may contribute to polar growth of A. tumefaciens.
The C-terminal Conserved Peptide (CTP) of FtsZAT is required for proper termination of polar growth
To better understand the mechanism by which FtsZAT terminates polar growth, we constructed truncated proteins to analyze the function of the C-terminal conserved peptide (CTP) and the C-terminal linker (CTL) (Figure 1A). The CTP is a highly conserved domain which binds to proteins such as FtsA, that tether FtsZ to the membrane [37, 51, 52]. The CTL is an intrinsically disordered region of variable length found in FtsZ proteins, which functions in the regulation of PG biosynthesis and protofilament assembly [17, 53-55]. To probe the function of the FtsZAT CTP and CTL domains, we expressed FtsZATΔCTP and FtsZATΔCTL in both WT and FtsZAT depletion backgrounds.
In order to execute these experiments, we constructed a vector with an alternative “inducible” promoter system, which is compatible with the chromosomal IPTG depletion system. We modified pSRKKm [56] by replacing lacIq with the gene encoding the cumate responsive repressor CymR [57, 58] and replacing the lacO operator sites with cuO operator sites (Figure 6-Figure Supplement 1A). This approach allows the same promoter to drive expression of both chromosomal full-length ftsZAT using IPTG and plasmid-encoded ftsZ variants using cumate. For simplicity, henceforth we referred to IPTG induction as mediated by Plac and cumate induction as mediated by Pcym. Expression of sfGFP from Pcym requires the presence of cumate (Figure 6-Figure Supplement 1B) and is comparable to expression of sfGFP from Plac (Figure 6-Figure Supplement 1C). Although higher concentrations of cumate inhibit growth of WT A. tumefaciens, 0.01 mM cumate does not impair growth of WT cells (Figure 6-Figure Supplement 1D; left) and is sufficient to complement growth of the ftsZAT depletion strain in the absence of IPTG (Figure 6-Figure Supplement 1D; right).
In the ftsZAT depletion strain, we introduced 4 vectors: an empty vector with Pcym (pEmpty), Pcym- ftsZAT (pFtsZAT), Pcym-ftsZATΔCTP (pFtsZΔCTP) or Pcym-ftsZATΔCTL (pFtsZΔCTL). When full-length ftsZAT is expressed from the chromosome, the viability of cells is not impacted by the presence of the Pcym vectors (Figure 6A, top left panel). In the absence of induction of ftsZ from the chromosome, the presence of the uninduced Pcym vectors, including pFtsZAT, is not sufficient to rescue viability of the FtsZ-depleted cells (Figure 6A, top right panel); however, viability is significantly restored by expression of plasmid-encoded FtsZAT in the presence of cumate (Figure 6A, bottom left panel). Expression of plasmid-encoded FtsZATΔCTP partially rescues the depletion of FtsZAT (Figure 6A, bottom left panel). In contrast, expression of plasmid-encoded FtsZATΔCTL does not rescue the depletion of FtsZAT (Figure 6A, bottom left panel) and when both chromosomal full-length FtsZAT and FtsZATΔCTL are expressed, viability is impaired, suggesting that FtsZATΔCTL may have a dominant negative phenotype (Figure 6A, bottom right panel).
Next, we observed cell morphology of the ftsZAT depletion strain carrying each of the four vectors under conditions where the chromosomal FtsZAT is depleted and the plasmid-encoded FtsZ variants are expressed for 6 or 14 h (Figure 6B). The presence of pEmpty does not impact the FtsZAT depletion phenotype: branched cells with multiple growth active poles are observed (Figure 6B). Plasmid-encoded FtsZAT rescues the chromosomal FtsZAT depletion, resulting in the production of typical rod-shaped cells with PG biosynthesis occurring at a single pole or mid-cell (Figure 6B, middle left). The partial rescue of FtsZAT depletion in viability by expression of FtsZATΔCTP was matched by a less severe defect in cell morphology (Figure 6B, middle right). Although cells are branched, they are much shorter and have fewer branches than FtsZAT depletion. FDAA labeling reveals that the expression of FtsZATΔCTP enables mid-cell labeling (Figure 6B, middle left), suggesting that PG is synthesized at mid-cell and that some cells may undergo division. Indeed, time-lapse microscopy of the FtsZAT depletion strain expressing only FtsZΔCTP reveals that the cells are capable of cell division events (Figure 6C, top panel). Remarkably, the sites of cell constriction and cell division are often asymmetric, giving rise to a cell population with a broad length distribution. Furthermore, polar growth is not terminated efficiently and both polar elongation and tip splitting events are evident. Together, these observations suggest that the FtsZ CTP contributes to proper termination of polar growth and divisome assembly. Expression of plasmid-encoded FtsZATΔCTL in the absence of chromosome-encoded FtsZAT gives rise to a distinct cell morphology (Figure 6B, far right panel). After 6 hours of FtsZATΔCTL expression, some cells contain mid-cell bulges. Remarkably, in these cells, FDAA labeling reveals that PG biosynthesis is occurring in the bulges and not at either pole. After 14 h, most cells have mid-cell swelling and multiple ectopic poles. Time-lapse microscopy reveals that polar growth is terminated and growth appears to be directed to mid-cell (Figure 6C, bottom panel, 320 min). When cell division fails, ectopic growth poles emerge from the mid-cell bulges (Figure 6C, bottom panel 520 min). The ectopic poles elongate, polar growth is terminated, and new ectopic growth poles are formed near the initial bulge site (Figure 6C, bottom panel). These observations suggest that the CTL of FtsZAT is required for proper cell division but is not required for the termination of polar growth.
The CTL of FtsZAT is required for proper PG composition
The mid-cell bulges observed during FtsZΔCTL expression are reminiscent of those observed when FtsZCCΔCTL is expressed in C. crescentus [17]. In C. crescentus, the CTL was shown to be required for robust PG biosynthesis [17]. We therefore hypothesized that the altered PG composition observed during depletion of FtsZAT could be due to absence of the CTL. To test this hypothesis, we introduced plasmids containing no FtsZ (empty vector control, pEmpty), full-length FtsZAT (pFtsZAT), or FtsZATΔCTL (pFtsZATΔCTL) into the ftsZ depletion strain. Each strain was grown under conditions in which expression of FtsZAT from the chromosomal copy is depleted and expression of the FtsZ variant (if present) from the plasmid is induced. PG was isolated from these strains following induction/depletion and analyzed. Induction of full-length FtsZAT from the plasmid yields lower levels of monomeric muropeptides compared to other strains, especially M3, and increased levels of dimeric and trimeric muropeptides, including D44 and T444 (Figure 7A-B). Overall the expression of full-length FtsZAT leads to an increased level of muropeptides with DD-crosslinks (Figure 7C). These observations indicate that expression of plasmid-encoded full-length FtsZAT compensates for the loss of FtsZAT from the chromosome. In contrast, the expression of FtsZATΔCTL did not compensate for the loss of full-length FtsZ as the PG composition is more similar to the PG profile of FtsZ-depleted cells (Figure 7A-C). This observation suggests that the CTL of FtsZAT likely function in the regulation of proper PG biosynthesis at mid-cell.
FtsZ CTL regulates protofilament assembly
Work in C. crescentus has shown that the FtsZCCCTL directly regulates protofilament structure and dynamics [53]. To determine if the CTL of FtsZAT similarly regulates its assembly, we purified FtsZATΔCTL and a control FtsZAT+CTL protein containing the same restriction sites at the junctions with the GTPase domain and CTP as the ΔCTL construct, but with the CTL in place. FtsZAT+CTL formed mostly single, gently curved protofilaments when visualized by TEM under all conditions tested (Figure 7D), similar to those observed for wildtype FtsZAT (Figure 3C). In contrast, under high salt conditions we observed extended bundles of FtsZATΔCTL (Figure 7D). Furthermore, we saw a decreased rate of GTP hydrolysis by FtsZATΔCTL under conditions that promote bundling (Figure 7D; 3.3 ± 0.2 GTP min-1 FtsZ-1 for FtsZAT+CTL and 2.1 ± 0.1 GTP min-1 FtsZ-1 for FtsZATΔCTL with 300 mM KCl, n =3). Together, these results suggest an important role for the CTL in limiting lateral interactions between protofilaments and promoting polymer turnover. These results in A. tumefaciens are consistent with effects of the CTL on polymer bundling reported in C. crescentus [17, 53]and E. coli [59]. Moreover, in light of our observations that FtsZATΔCTL does not restore proper PG chemistry to FtsZAT-depleted cells (Figure 7A,B), these data are in line with the growing body of evidence linking FtsZ dynamics and polymer superstructure to the regulation of PG biosynthesis.
FtsA is required for cell division but not termination of polar growth
FtsA is an actin-like protein that associates with the membrane through an amphipathic helix and binds the FtsZ CTP to anchor FtsZ polymers to the membrane [51, 60]. In C. crescentus, recruitment of FtsA to mid-cell occurs well after the establishment of the FtsZ-ring and is dependent on the presence of FtsZ [13, 61]. In A. tumefaciens, FtsA-sfGFP is retained at the growth pole prior to appearing at mid-cell just before cell division [27, 30]. Here, we confirm that FtsA-sfGFP is observed as a focus at the growth pole until transitioning to a ring-like structure at mid-cell (Figure 8A, top panelIn fact, at some timepoints, both a polar focus and a mid-cell ring of FtsA are observed. Eventually, the polar focus disappears as the FtsA-sfGFP ring becomes more intense just prior to cell division. During the depletion of FtsZAT, a focus of FtsA-sfGFP can be found at the growing pole, and at a newly formed ectopic pole near mid-cell (Figure 8A, bottom panel). FtsA-sfGFP remains associated with each growth pole, and as the poles undergo tip splitting events, each focus of FtsA-sfGFP is also split, resulting in the presence of FtsA-sfGFP in each of the 4 growth-active poles. These observations suggest that FtsZAT is required not only for proper mid-cell localization of FtsA to mid-cell prior to cell division but also contributes to release of FtsA-sfGFP from the growth pole.
Since FtsA tethers FtsZ to the membrane and enables divisome assembly [37, 51, 52] in E. coli, we expected that the depletion of FtsA would phenocopy the depletion of FtsZ. Although a saturating transposon mutagenesis screen indicated that ftsA is not essential for A. tumefaciens cell survival [36], we were unable to construct a ΔftsA mutant. Thus, we constructed a depletion strain in which expression of ftsA is controlled by Plac. Under conditions where FtsA is present in the ftsA depletion strain, cells maintain proper rod-shaped morphology, polar growth, and cell division occurs from constrictions formed near mid-cell (Figure 8B-C, top panels). In contrast, when FtsA is depleted, cells exhibit a marked change in morphology (Figure 8B, bottom panel; Movie 2). During the depletion of FtsA, rod-shaped cells initially elongate from a growth pole (Figure 8B, bottom panel, 0 min). Polar growth is terminated and growth is re-initiated from near mid-cell, typically resulting in the formation of two ectopic poles perpendicular to the original longitudinal axis of the cell (Figure 8B, bottom panel, 170 min). Cells depleted of FtsA continue multipolar growth (Figure 8B, bottom panel, 360 min), terminate growth from both poles and reinitiate growth from near mid-cell resulting in the formation of a new pair of ectopic growth poles (Figure 8B, bottom panel, 510 min). This pattern of multipolar growth, polar growth termination, and new branch formation is continued until cells eventually bulge at the mid-cell and lyse. Overall these observations indicate that the phenotypes caused by FtsZ and FtsA depletion are distinct from one another and suggest that only FtsZ is required for proper termination of polar growth.
To confirm that polar growth occurs and is terminated during FtsA depletion, cells were labeled with FDAAs (Figure 8C, bottom panel). Indeed, FDAAs label the tips of two poles, which are emerging from near mid-cell consistent with the re-initiation of polar growth. To further confirm that polar growth is terminated during FtsA depletion, we observed the localization of PopZ-YFP (Figure 8D, top panel). PopZ marks the growth poles [47] and becomes trapped at growth poles during depletion of FtsZ (Figure 4B). During FtsA depletion, PopZ-YFP is initially present at the growth pole (Figure 8D, top panel, 0 min). Next, PopZ-YFP disappears from the growth poles and reappears near mid-cell (Figure 8D, top panel, 80 min) indicating that polar growth is terminated. Throughout the FtsA depletion, PopZ-YFP continues to disappear from growth poles and reappears at the tips of newly emerging growth poles. Overall, these observations clearly indicate that FtsA is not necessary for termination of polar growth; however, FtsA has an essential function at a later stage of cell division since the cells fail to divide and are prone to lysis.
The ability of cells to target growth to near mid-cell during FtsA depletion suggests that FtsZ-rings may form, enabling the termination of polar growth. Indeed, FtsZAT-sfGFP-rings form near mid-cell early during FtsA depletion (Figure 8D, bottom panel). FtsZAT-sfGFP is briefly retained at new growth poles before reappearing to mark the site where a new growth pole will emerge. These observations are consistent with the finding the FtsA is retained at the growth pole longer than FtsZ [27, 62], and suggest that FtsA arrives at mid-cell after Z-ring assembly and the initiation of FtsZ-dependent cell wall biogenesis. The results observed here in A. tumefaciens are consistent with the observation that FtsA arrives to mid-cell after FtsZ and the onset of mid-cell cell wall biogenesis in C. crescentus [13, 61]. In both A. tumefaciens and C. crescentus, the late arrival of FtsA to the divisome suggests that other proteins contribute to proper tethering of FtsZ to the membrane. In C. crescentus, the FtsZ-binding protein, FzlC, functions as a membrane anchor early during the establishment of the divisome [31, 63]. A homolog of FzlC is readily found in the A. tumefaciens genome (Atu2824) and may contribute to the ability of FtsZ-rings to form in the absence of FtsA.
Depletion of the downstream divisome component FtsW phenocopies depletion of FtsA
Having observed a distinct effect on cell morphology in the absence of ftsA, we wondered if the phenotype observed during ftsA depletion could be recapitulated in the absence of another late-arriving divisome protein. To test this hypothesis, we constructed a depletion strain of FtsW, which is recruited to mid-cell after FtsA in both E. coli and C. crescentus divisome assembly models [11, 13]. Depletion of FtsW results in a phenotype which is strikingly similar to the depletion of FtsA (Figure 9). When FtsW is induced normal growth is observed (Figure 9A, top panel). During FtsW depletion, polar growth is terminated, resulting in the establishment of growth-active poles from near mid-cell (Figure 9A, bottom panel; Movie 3). Multiple rounds of termination of polar growth followed by reinitiation of growth from near mid-cell occur until the mid-cell bulges and the cells ultimately lyse (Figure 9A, bottom panel). Labeling of growth active poles with FDAAs (Figure 9B) or by tracking PopZ-YFP localization (Figure 9C, top panel) confirmed that new branches which emerge from mid-cell are formed by polar growth.
Finally, we confirmed that FtsZ-rings form during the depletion of FtsW and the presence of an FtsZAT-sfGFP-ring typically marks the site where an ectopic growth pole will form (Figure 9C, bottom panel). Together, these observations suggest that FtsZ-rings are formed in the absence of FtsW, enabling the initiation of cell wall biogenesis. Given that FtsW drives septal PG biosynthesis, [64] these findings indicate that the cell wall biogenesis that occurs during depletion of FtsA or FtsW may require the elongation machinery, which typically functions in polar growth. Since the elongation machinery for A. tumefaciens remains to be identified, it is possible that there is considerable overlap between the machineries that contribute to polar and septal PG biosynthesis.
Concluding Remarks
While many questions remain unanswered about the regulation of cell wall biogenesis in A. tumefaciens, our work sheds light on the transition from polar growth to mid-cell growth. We find that FtsZAT, FtsA, and FtsW are required for constriction and cell separation, but FtsZAT is also required to terminate polar growth and initiate mid-cell peptidoglycan synthesis. How might the formation of an FtsZAT-ring at mid-cell cause the termination of polar growth? We find that PopZ, and LDTP0845 become trapped at the growth poles during FtsZ depletion (Figure 5). It is possible that one or more of these proteins contributes to both polar peptidoglycan biosynthesis and mid-cell peptidoglycan synthesis and that the FtsZ-dependent targeting of these proteins (and likely others) to mid-cell triggers the termination of polar growth. While the mid-cell localization of PopZ is dependent on the presence of FtsZAT (Figure 4), FtsZAT-ring stability and placement are impacted by the absence of PopZ [45]. Furthermore, deletion of popZ impairs termination of polar growth and results in cell division defects [45, 46, 65]. The apparent co-dependence of FtsZ and PopZ for localization may suggest that these proteins function together during the early stages of cell division, particularly the termination of polar growth and onset of mid-cell PG biosynthesis.
Overall, our results are consistent with a general model, which is highly conserved in bacteria, in which the establishment of a FtsZ-ring leads to the recruitment of many other cell division proteins to mid-cell [11], though many mechanistic questions remain. How is FtsZAT targeted to mid-cell? A variety of mechanisms that contribute to the proper placement of FtsZ at mid-cell have been described (for review see [66, 67]). The most well studied mechanisms of FtsZ positioning include negative regulation by the Min system and nucleoid occlusion. While genes encoding components of the Min system are readily identifiable in the A. tumefaciens genome, the deletion of minCDE has a minimal impact on placement of constriction sites and cell division efficiency [68]. Furthermore, FtsZAT-GFP rings form over DNA prior to nucleoid separation in A. tumefaciens. These observations indicate that additional regulatory mechanisms must contribute to proper division site selection in A. tumefaciens. Following the appearance of FtsZ at mid-cell, how is the FtsZAT-ring stabilized? In E. coli, the FtsZ-ring is stabilized by interactions with FtsA and ZipA, which tether FtsZ filaments to the membrane [51, 52, 69, 70]. In A. tumefaciens, FtsZAT appears at mid-cell well before FtsA [30] and we observe that FtsZAT rings form even when FtsA is depleted (Figure 7C, bottom panel). Furthermore, the position of FtsZAT-GFP rings marks the site of ectopic pole formation. These observations suggest the FtsZAT is stabilized, at least early during cell division by other proteins. While there are no obvious ZipA homologs encoded in the A. tumefaciens genome, a homolog of FzlC, which functions to stabilize FtsZ in C. crescentus [31, 63], is encoded in the genome.
The observation that FtsZ is necessary for the initiation of mid-cell PG biosynthesis suggests that FtsZ is necessary for recruitment of PG biosynthesis enzymes to mid-cell. Septal PG biosynthesis is likely mediated by FtsW, a putative PG glycosyltransferase [71-73], and PBP3 (FtsI), a PG DD-transpeptidase [74]. In A. tumefaciens, depletion of FtsW does not cause a complete block of PG synthesis at mid-cell (Figure 8). This observation suggests that mid-cell PG biosynthesis is mediated by other cell wall biogenesis enzymes while the activity of FtsW contributes to later stages of cell division, consistent with the inability of cells to form constrictions and separate in the absence of FtsW. These observations may indicate that the initial PG biosynthesis at mid-cell comprises the final stage of cell elongation, consistent with descriptions of FtsZ-dependent mid-cell elongation in C. crescentus [16]. The observation that growth-active, ectopic poles emerge from near mid-cell during FtsW depletion (Figure 8B) provides evidence in support of this possibility. Thus, FtsZ-dependent PG biosynthesis may contribute to both elongation and cell division in A. tumefaciens. For a polar growing bacterium, it is tempting to speculate that the retention of PG biosynthesis machinery dedicated to elongation at the site of cell division may prime the newly formed poles to become growth active following cell separation.
Materials and Methods
Bacterial strains and culture conditions
All bacterial strains and plasmids used are listed in Table S4.1. A. tumefaciens strains were grown in ATGN minimal medium with .5% glucose [75] at 28°C. E. coli strains were grown in Luria-Bertani medium at 37°C. When indicated, kanamycin (KM) was used at 300 µg/ml for A. tumefaciens, 50 µg/ml for E. coli DH5α, and 25 µg/ml for E. coli S17-1 λ pir. Gentamicin was used when indicated at 200 µg/ml for A. tumefaciens and 20 µg/ml for E. coli DH5α. IPTG was added at a concentration of 1 mM when indicated. Cumate was added at a concentration of 0.1 mM when indicated.
Construction of expression plasmids and strains
All strains and plasmids used are listed in Table S4.1, while primers used are listed in Table S4.2. For amplification of target genes, primer names indicate the primer orientation and added restriction sites. To construct expression vectors containing ftsZAT-sfgfp, ftsZ1-sfgfp, ftsZ3-sfgfp, and ldtp0845-sfgfp the respective coding sequence was amplified from purified C58 genomic DNA using primers indicated in Table S4.2. The amplicons were digested overnight and ligated into cut pSRKKM-Plac-sfgfp using NEB T4 DNA ligase at 4°C overnight. The newly formed sfgfp fusion of each gene was excised from the plasmid by overnight digestion with NdeI and NheI. Fragments containing ftsZAT-sfgfp, ftsZ1-sfgfp, ftsZ3-sfgfp, and ldtp0845-sfgfp were then ligated into cut pRV-MCS2 to give constitutive expression vectors containing the fusions. To construct the popZ-yfp expression vector, popZ along with the upstream promoter sequence were amplified from purified C58 genomic DNA, digested and ligated into pMR10.
To construct pSRKKM-Pcym, a synthesized gBlock from IDT Integrated DNA Technologies was made containing the regulatory elements of the cumate system similar to previously described plasmid constructs [76, 77]. The Pcym promoter region is annotated in Table S4.2. The sequence encoding the cumate repressor was codon optimized for A. tumefaciens and placed under the control of the constitutive kanamycin promoter from pSRKKm-Plac-sfgfp. The synthesized gBlock was digested overnight with EcoRI and NdeI. The resulting fragment was then ligated into cut pSRKKm-Plac-sfgfp thereby replacing the original lac promoter and repressor with the cumate repressor and cumate regulated promoter.
Next, yfp-parB was excised from pSRKKM-Plac-yfp-parB [46] and ligated into pSRKKM-Pcym to create an expression vector compatible with the depletion strains. To create expression vectors for ftsZAT, ftsZATΔCTP, and ftsZATΔCTL the respective target gene was amplified utilizing indicated primers, digested overnight with NdeI and BamHI and ligated into pSRKKM-Pcym.
All expression vectors were verified by sequencing. All vectors were introduced into A. tumefaciens strains utilizing standard electroporation protocols [78] with the addition of IPTG in the media when introducing plasmids into in depletion backgrounds.
Construction of deletion/depletion plasmids and strains
Vectors for gene deletion by allelic exchange were constructed using recommended methods for A. tumefaciens [78]. Briefly, 500 bp fragments upstream and downstream of the target gene were amplified using primer pairs P1/P2 and P3/P4 respectively. Amplicons were spliced together by SOEing using primer pair P1/P4. The amplicon was digested and ligated into pNTPS139. The deletion plasmids were introduced into A. tumefaciens by mating using an E. coli S17 conjugation strain to create KM resistant, sucrose sensitive primary exconjugants. Primary exconjugants were grown overnight in media with no selection. Secondary recombinants were screened by patching for sucrose resistance and KM sensitivity. Colony PCR with primers P5/P6 for the respective gene target was used to confirm deletion. PCR products from P5/P6 primer sets were sequenced to further confirm deletions.
For depletion strains, target genes (ftsZAT, ftsA, and ftsW) were amplified, digested and ligated into either pUC18-mini-Tn7T-GM-Plac or pUC18-mini-Tn7T-GM-Plac. The mini-Tn7 vectors, along with the pTNS3 helper plasmid, were introduced into C58ΔtetRA::a-attTn7 as described previously [32]. Transformants were selected for gentamicin resistance and insertion of the target gene into the a-att site was verified by colony PCR using the tet forward and Tn7R109 primer. PCR products were sequenced to confirm insertion of the correct gene. Next, the target gene was deleted from the native locus as described above in the presence of 1 mM IPTG to drive expression of the target gene from the engineered site.
Construction of plasmids for protein expression and purification
To construct pET21a FtsZAT, ftsZAT was amplified from C58 genomic DNA with FtsZAT For NdeI and FtsZAT Rev EcoRI, digested with NdeI and EcoRI, and ligated into similarly digested pET21a. To construct pTB146 FtsZ1, ftsZ1 was amplified from C58 genomic DNA with FtsZ1 For SapI and FtsZ1 Rev BamHI, digested with SapI and BamHI, and ligated into similarly digested pTB146. Ligation products were transformed into NEB Turbo (New England Biolabs) and selected for ampicillin resistance. Insertions were verified by colony PCR and Sanger sequencing. Primers FtsZAT-L72W and FtsZ1-L71W were used to mutagenize pET21a FtsZAT and pTB146 FtsZ1, respectively, using the Quikchange Multi Lightning Mutagenesis Kit (Agilent) and following the manufacturer’s protocol to generate pET21a FtsZAT-L72W and pTB146 FtsZ1-L71W. Mutations in the targeted sites were verified by Sanger sequencing.
pET21c FtsZAT+CTL and pET21c FtsZATΔCTL were constructed in several steps. First, the GTPase domain of ftsZAT was amplified from C58 genomic DNA using FtsZAT For NdeI and FtsZAT GTPase Rev KpnI SacI, split into two aliquots, and digested with NdeI and KpnI or NdeI and SacI. The CTL region of ftsZAT was amplified from C58 genomic DNA using FtsZAT CTL For KpnI and FtsZAT CTL Rev SacI and digested with KpnI and SacI. For FtsZAT+CTL, the GTPase domain amplicon (digested with NdeI and KpnI) was combined with the CTL amplicon (digested with KpnI and SacI) and together they were ligated into pXCFPN-1 [79] digested with NdeI and SacI. For FtsZATΔCTL, the GTPase domain amplicon (digested with NdeI and SacI) was ligated into pXCFPN-1 digested with NdeI and SacI. Each was transformed into NEB Turbo, selected for spectinomycin resistance, and confirmed by colony PCR and Sanger sequencing. Next, the CTP was added to each of the above constructs by annealing oligos FtsZAT CTP + and FtsZAT CTP – (engineered with overhangs compatible with SacI and NheI ligation) and ligating into the above constructs digested with SacI and NheI. Each was transformed into NEB Turbo and confirmed as above to generate pX1 FtsZAT+CTL and pX1 FtsZATΔCTL. Finally, FtsZAT+CTL and FtsZATΔCTL were subcloned into pET21c by digestion of pX1 FtsZAT+CTL and pX1 FtsZATΔCTL with NdeI and NheI and ligating into similarly digested pET21c. Each was transformed into NEB Turbo, selected for ampicillin resistance, and confirmed by colony PCR and Sanger sequencing.
DIC and phase contrast microscopy
Exponentially growing cells (OD600 = ∼0.6) were spotted on 1% agarose ATGN pads as previously described [80]. Microscopy was performed with an inverted Nikon Eclipse TiE with a QImaging Rolera em-c2 1K EMCCD camera and Nikon Elements Imaging Software. For time-lapse microscopy, images were collected every ten minutes, unless otherwise stated.
Fluorescence microscopy
Plasmid encoded FtsZAT-sfGFP, FtsZ1-sfGFP, FtsZ3-sfGFP, and LDTP0845-sfGFP fusions were expressed from the Pvan promoter, which provides constitutive low levels of expression (Figure 6-Figure Supplement 1C). Plasmid encoded FtsA-sfGFP and PopZ-YFP fusions were expressed from the native promoters. Expression of plasmid encoded YFP-ParB was induced by the presence of 0.1 mM cumate for 2 hours (h). Cells containing plasmids with fluorescent protein fusions were grown to exponential phase before imaging on agarose pads.
To visualize DNA, 1 ml of exponentially growing cells was treated with 1 µl of Sytox Orange for 5 minutes. Cells were collected by centrifugation and washed with PBS 2 times followed by a final resuspension in PBS. Cells were then imaged on agarose pads.
To visualize sites of active peptidoglycan synthesis 1 ml of exponentially growing cells was labeled with the fluorescent D-amino acid (FDAA), HCC amino-D-alanine (HADA), as previously described [49, 80].
Cell viability and growth curve assays
For cell viability spot assays, exponentially growing cultures were diluted to OD600 = 0.1 and serially diluted in ATGN. 3 µl of each dilution was spotted onto ATGN and incubated at 28°C for 3 days before imaging. When appropriate ATGN plates contained KM 300 µg/ml, IPTG 1mM, and cumate 0.1 mM as indicated in figure legends. For growth curve analysis, exponentially growing cultures were diluted to OD600 = .05 in 200 µl of ATGN in 96-well plates. Plates were shaken for 1 minute before OD600 readings, which were taken every 10 minutes.
Cell morphology and constriction rate analysis
Exponentially growing cells were imaged using phase contrast microscopy as described above. Cell length, area, and constrictions were detected using MicrobeJ software [81].
To calculate constriction rates, cells with detectable constrictions were tracked using time-lapse microscopy. The width of the cell constriction was measured at an initial time-point and the measurement was repeated after 10 minutes. The difference in constriction width was divided by the 10-minute time interval to give a constriction rate.
Western blot analysis
For western blot analysis of FtsZ depletion, the ftsZ depletion strain was grown in 40 ml ATGN with 1 mM IPTG to exponential phase. 2 ml of culture was collected prior to depletion (time 0) by centrifugation at 10,000 x g for 3 minutes. The remaining culture was collected by centrifugation at 3500 x g for 10 minutes, and supernatants were discarded. Cells were washed in sterile water and pelleted again. To deplete FtsZ, the pellet was resuspended in fresh ATGN without IPTG and grown under standard culturing conditions. 2-ml samples were collected by centrifugation after 30, 45, 60, 120, and 240 minutes of depletion. OD600 was taken for each sample prior to centrifugation so that samples could be normalized to an OD600 equivalent to 0.68. The cell pellets were incubated with 100 μl of a master mix containing 1 ml of BugBuster protein extraction reagent (Novagen) and supplemented with 1 EDTA-free protease inhibitor cocktail (Sigma), 10 μl of lysonase (Novagen), 2,500 U/ml DNase I (Thermo Scientific), and 1 mM dithiothreitol (DTT) (Thermo Scientific) for 25 minutes with shaking at room temperature to lyse the cell pellets. The whole-cell lysates were clarified by centrifugation at 10,000 rpm for 15 min. A final concentration of 1 X Laemmli buffer was added to the cleared cell lysates. Samples were boiled at 100°C for 5 min prior to loading on a 4-15% Mini-PROTEAN TGX Precast Gel (Bio-Rad). The separated proteins were electroblotted onto polyvinylidene difluoride (PVDF) membranes (Bio-Rad) and blocked overnight in 5% nonfat dry milk powder solubilized in 1% TBST (Tris-buffered saline [TBS], 1% Tween 20). The blocked PVDF membranes were probed with Escherichia coli anti-FtsZ (1:3000) monoclonal antibody (gift from Joe Lutkenhaus) for 1.5 h in 5% milk-TBST, followed by incubation with anti-rabbit (1:5000) HRP (Pierce 31460) secondary antibody for 1 h in 5% milk-TBST. The secondary antibody was detected using the ECL Plus HRP substrate (Thermo Scientific Pierce).
For comparison of expression from Pvan, Plac, and Pcym promoters, strains were grown in 2 ml ATGN with 200 ug/mL KM to exponential phase. Plac and Pcym were induced with 1 mM IPTG and 50 μM cumate, respectively for 4 h. Cell pellets were lysed as described above and clarified whole-cell lysates were boiled with 1 X Laemmli buffer for 5 min prior to loading on 4-15% Mini-PROTEAN TGX Precast Gel (Bio-Rad). The separated proteins were electroblotted onto PVDF membranes (Bio-Rad), blocked as described above, and probed with anti-GFP (1:3,000) monoclonal antibody (Thermo Scientific Pierce) for 1 h in 5% milk-TBST, followed by incubation with a donkey anti-mouse (1:300) horseradish peroxidase-conjugated secondary antibody (Thermo Scientific Pierce) for 1 h in 5% milk-TBST. The secondary antibody was detected using the ECL Plus HRP substrate (Thermo Scientific Pierce).
Protein expression and purification
FtsZAT, FtsZAT-L72W, FtsZAT+CTL (FtsZAT with restriction sites flanking the CTL), and FtsZATΔCTL (FtsZAT with the CTL deleted, containing the same restriction sites at the GTPase-CTC junction as in FtsZAT+CTL) were expressed and purified in untagged form. Each was produced from a pET21 expression vector (pEG1555 – FtsZAT, pEG1556 - FtsZAT-L72W, pEG1444 - FtsZAT+CTL, pEG1445 - FtsZATΔCTL) in Escherichia coli Rosetta(DE3)pLysS induced at 37°C for 4 h with 0.5 mM IPTG after OD reached 0.8 to 1.0 OD at 600 nm. Cells were harvested by centrifugation at 6000 x g and resuspended in 30 mL FtsZ QA buffer (50 mM Tris-HCl pH 8, 50 mM KCl, 0.1 mM EDTA, 10% glycerol) per liter of culture. Resuspensions were snap frozen in liquid nitrogen and stored at −80°C until purification. To purify, resuspensions were thawed quickly and cells were lysed by incubation with 1 mg/mL lysozyme, 2.5 mM MgCl2, DNAse I, 2 mM PMSF, and a cOmplete mini EDTA-free protease inhibitor tablet (Roche) for 45 min to 1 h at room temperature followed by sonication. Lysates were cleared by centrifugation at 15000 x g for 30 min at 4°C and filtered through a 0.45 µm filter before anion exchange chromatography (HiTrap Q HP 5 mL, GE Life Sciences). Protein was eluted with a linear KCl gradient (FtsZ QA buffer with 50 to 500 mM KCl) and fractions containing FtsZ were verified by SDS-PAGE, pooled, and subjected to ammonium sulfate precipitation. Precipitates (at 17-20% ammonium sulfate saturation depending on the variant) were verified by SDS-PAGE, resuspended in HEK50G (50 mM HEPES-KOH pH 7.2, 0.1 mM EDTA, 50 mM KCL, 10% glycerol, 1 mM β-mercaptoethanol), and further purified by gel filtration (Superdex 200 10/300 GL, GE Life Sciences). Peak fractions were pooled, snap frozen in liquid nitrogen, and stored at −80°C.
FtsZ1 and FtsZ1-L71W were produced as His6-SUMO fusions and cleaved to yield untagged, scarless proteins. Each was produced from a pTB146 expression vector (pEG1535 - FtsZ1, pEG1542 - FtsZ1-L71W) in E. coli Rosetta (DE3)pLysS as described above. Cells were harvested by centrifugation as above, resuspended in HK300G (50 mM HEPES-KOH pH7.2, 300 mM KCl, 10% glycerol) with 20 mM imidazole, snap frozen in liquid nitrogen, and stored at −80°C until purification. To purify, resuspensions were thawed quickly and cells were lysed by incubation with 1 mg/mL lysozyme, 2.5 mM MgCl2, and DNAse I for 45 min at room temperature followed by sonication. Lysate was cleared and filtered as described above. Protein was isolated by Ni2+ affinity chromatography (HisTrap FF 1 mL, GE Life Sciences) and eluted in HK300G with 300 mM imidazole. Fractions containing His6-SUMO fusions were verified by SDS-PAGE, combined with Ulp1 Sumo protease at a 1:100 (protease:FtsZ) molar ratio, and cleaved by incubation at 30°C for 3.5 h. Cleaved FtsZ1 or FtsZ1L71W was purified away from His6-SUMO by gel filtration (Superdex 200 10/300 GL, GE Life Sciences) in HEK50G. Peak fractions were pooled, snap frozen in liquid nitrogen, and stored at −80°C.
Polymerization kinetics assays
A Fluoromax-3 spectrofluorometer (Jobin Yvon, Inc) was used to monitor FtsZ polymerization by right-angle light scattering and tryptophan fluorescence. FtsZ1 and/or FtsZAT (wild-type or L71W/L72W mutants, as indicated in figures and text) was polymerized in HEK50 (50 mM HEPES-KOH pH 7.2, 50 mM KCl, 0.1 mM EDTA) with 2.5 mM MgCl2. 2 mM GTP was used to induce polymerization for light scattering and 50 µM GTP was used to induce polymerization for tryptophan fluorescence (GTP is fluorescent at the wavelengths used, so low concentrations must be used). GTP was added after baseline light scatter or fluorescence was established. For light scattering, samples were excited at 350 nm and scatter was detected at 350 nm with slits set to 2 nm. For tryptophan fluorescence, samples were excited at 295 nm and emission was detected at 344 nm, with 2 nm slits.
GTPase assay
FtsZ1 and/or FtsZAT was polymerized in HEK50 with 2.5 mM MgCl2 and 2 mM GTP. FtsZAT+CTL or FtsZATΔCTL was polymerized in HEK50 or HEK300 (same as HEK50 but with 300 mM KCl) as indicated, with 10 mM MgCl2 and 2 mM GTP. GTP was added at time 0. Reaction was stopped at 5, 10, 15, 20, and 30 minutes with quench buffer (50 mM HEPES-KOH pH 7.2, 21.3 mM EDTA, 50 mM KCl). Inorganic phosphate in solution (liberated by GTP hydrolysis) over time was measured using SensoLyte MG Phosphate Assay Kit Colorimetric (AnaSpec, Inc, Fremont, California).
Negative stain transmission electron microscopy (TEM)
FtsZ1 and/or FtsZAT were polymerized in HEK50 with 2.5 mM MgCl2 and 2 mM GTP. 4 µM FtsZAT+CTL or FtsZATΔCTL were polymerized in HEK50 or HEK300 as indicated with 10 mM MgCl2 and 2 mM GTP. After a 15-minute incubation at room temperature, samples were applied to carbon-coated glow-discharged grids with 0.75% uranyl formate staining as previously described [17, 82]. TEM samples were imaged using a Philips/FEI BioTwin CM120 TEM equipped with an AMT XR80 8 megapixel CCD camera (AMT Imaging, USA).
Peptidoglycan composition analysis
Six cultures of WT and ftsZ depletion cells were grown in 10 ml of ATGN with IPTG to exponential phase. The 10 ml cell cultures were added to 40 ml of fresh media. The 50 ml cultures were grown to exponential phase and pelleted by centrifugation at 4000 x g for 10 minutes. Cell pellets were washed three times with ATGN by centrifugation and resuspension to remove IPTG. After the final wash 3 cell pellets were resuspended in 50 ml ATGN and the remaining 3 pellets were resuspended in 50 ml ATGN with 1 mM IPTG. Each culture was grown for 14 h. The optical densities of the cells were monitored to ensure the optical density of the cultures never went above OD600 = 0.7 to avoid changes to peptidoglycan content due to stationary phase. If necessary, fresh medium was added to dilute the cultures to maintain exponential growth. After 14 h of growth, 50 ml of the exponential cultures were collected and pelleted by centrifugation at 4000 x g for 20 minutes. Cell pellets were resuspended in 1mL of ATGN and 2 mL of 6% SDS and stirred with magnets while boiling for 4 h. After 4 h, samples were removed from heat but continued to stir overnight. Samples were then shipped to Dr. Felipe Cava’s laboratory for purification and analysis.
Upon arrival, cells were boiled and simultaneously stirred by magnets for 2 h. After 2 h, boiling was stopped and samples were stirred overnight. Peptidoglycan was pelleted by centrifugation for 13 min at 60000 rpm (TLA100.3 Beckman rotor, Optima Max-TL ultracentrifuge; Beckman), and the pellets were washed 3 to 4 times by repeated cycles of centrifugation and resuspension in water. The pellet from the final wash was resuspended in 50 µl of 50 mM sodium phosphate buffer, pH 4.9, and digested overnight with 100 µg/ml of muramidase at 37°C. Muramidase digestion was stopped by boiling for 4 min. Coagulated protein was removed by centrifugation for 15 min at 15000 rpm in a desktop microcentrifuge. The muropeptides were mixed with 15 µl 0.5 M sodium borate and subjected to reduction of muramic acid residues into muramitol by sodium borohydride (10 mg/ml final concentration, 20 min at room temperature) treatment. Samples were adjusted to pH 3 to 4 with orthophosphoric acid and filtered (0.2-µm filters).
Muropeptides were analyzed on a Waters UPLC system equipped with an ACQUITY UPLC BEH C18 Column, 130 Å, 1.7 µm, 2.1 mm × 150 mm (Waters) and a dual wavelength absorbance detector. Elution of muropeptides was detected at 204 nm. Muropeptides were separated at 45°C using a linear gradient from buffer A [formic acid 0.1% (v/v)] to buffer B [formic acid 0.1% (v/v), acetonitrile 20% (v/v)] in a 12 min run with a 0.250 ml/min flow. Peptidoglycan compositional analysis on triplicate samples was completed on two separate occasions.
FUNDING INFORMATION
PB and MH were supported by the National Science Foundation, IOS1557806. This work was funded in part by the National Institutes of Health through R01GM108640 (EDG) and T32GM007445 (training support of PJL). FC and AA receive funding support from Laboratory for Molecular Infection Medicine Sweden, Knut and Alice Wallenberg Foundation, Kempe and the Swedish Research Council. AA is supported by a MIMS/VR PhD position.
Movie 1. Growth and morphological changes during FtsZAT depletion. Cells were washed to remove inducer and grown in liquid ATGN for 4 hours before spotting on a ATGN pad. Images were acquired every ten minutes and movie is played at 10 frames per second for a total of 145 frames.
Movie 2. Growth and morphological changes during FtsA depletion. Cells were washed to remove inducer and grown in liquid ATGN for 2 hours before spotting on a ATGN pad. Images were acquired every ten minutes and movie is played at 10 frames per second for a total of 97 frames.
Movie 3. Growth and morphological changes during FtsW depletion. Cells were washed to remove inducer and grown in liquid ATGN for 4 hours before spotting on a ATGN pad. Images were acquired every ten minutes and movie is played at 10 frames per second for a total of 85 frames.
ACKNOWLEDGEMENTS
We thank members of the Brown lab for helpful discussions and critical reading of this manuscript.
References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.
- 6.
- 7.
- 8.
- 9.↵
- 10.↵
- 11.↵
- 12.
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.
- 23.
- 24.
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵