Streptococcus suis MsmK: Novel Cell Division Protein Interacting with FtsZ and Maintaining Cell Shape

ABSTRACT Bacteria of different shapes have adopted distinct mechanisms to faithfully coordinate morphogenesis and segregate their chromosomes prior to cell division. Despite recent focuses and advances, the mechanism of cell division in ovococci remains largely unknown. Streptococcus suis, a major zoonotic pathogen that causes problems in human health and in the global swine industry, is an elongated and ellipsoid bacterium that undergoes successive parallel splitting perpendicular to its long axis. Studies on cell cycle processes in this bacterium are limited. Here, we report that MsmK (multiple sugar metabolism protein K), an ATPase that contributes to the transport of multiple carbohydrates, has a novel role as a cell division protein in S. suis. MsmK can display ATPase and GTPase activities, interact with FtsZ via the N terminus of MsmK, and promote the bundling of FtsZ protofilaments in a GTP-dependent manner in vitro. Deletion of the C-terminal region or the Walker A or B motif affects the affinity between MsmK and FtsZ and decreases the ability of MsmK to promote FtsZ protofilament bundling. MsmK can form a complex with FtsZ in vivo, and its absence is not lethal but results in long chains and short, occasionally anuclear daughter cells. Superresolution microscopy revealed that the lack of MsmK in cells leads to normal septal peptidoglycan walls in mother cells but disturbed cell elongation and peripheral peptidoglycan synthesis. In summary, MsmK is a novel cell division protein that maintains cell shape and is involved in the synthesis of the peripheral cell wall. IMPORTANCE Bacterial cell division is a highly ordered process regulated in time and space and is a potential target for the development of antimicrobial drugs. Bacteria of distinct shapes depend on different cell division mechanisms, but the mechanisms used by ovococci remain largely unknown. Here, we focused on the zoonotic pathogen Streptococcus suis and identified a novel cell division protein named MsmK, which acts as an ATPase of the ATP-binding cassette-type carbohydrate transport system. MsmK has GTPase and ATPase activities. In vitro protein assays showed that MsmK interacts with FtsZ and promotes FtsZ protofilament bundling that relies on GTP. Superresolution microscopy revealed that MsmK maintains cell shape and is involved in peripheral peptidoglycan synthesis. Knowledge of the multiple functions of MsmK may broaden our understanding of known cell division processes. Further studies in this area will elucidate how bacteria can faithfully and continually multiply in a constantly changing environment.

IMPORTANCE Bacterial cell division is a highly ordered process regulated in time and space and is a potential target for the development of antimicrobial drugs. Bacteria of distinct shapes depend on different cell division mechanisms, but the mechanisms used by ovococci remain largely unknown. Here, we focused on the zoonotic pathogen Streptococcus suis and identified a novel cell division protein named MsmK, which acts as an ATPase of the ATP-binding cassette-type carbohydrate transport system. MsmK has GTPase and ATPase activities. In vitro protein assays showed that MsmK interacts with FtsZ and promotes FtsZ protofilament bundling that relies on GTP. Superresolution microscopy revealed that MsmK maintains cell shape and is involved in peripheral peptidoglycan synthesis. Knowledge of the multiple functions of MsmK may broaden our understanding of known cell division processes. Further studies in this area will elucidate how bacteria can faithfully and continually multiply in a constantly changing environment. with the most highly conserved divisome component, FtsZ. Conserved-domain analysis showed that the N-terminal region of MsmK contains the characteristic motifs of an ATPase (Fig. 1A). Walker A and B motifs are associated with ATP binding and hydrolysis, respectively (16). For follow-up protein experiments, His-tagged MsmK protein and its derivate proteins, N_MsmK (amino acids [aa] 1 to 234), C_MsmK (aa 235 to 378), A_MsmK, and B_MsmK were expressed and purified (Fig. 1A). A_MsmK and B_MsmK are the protein mutants generated upon deletion of the Walker A (GPSGCGKS; aa 40 to 47) and B (VFLMD; aa 158 to 162) motifs, respectively. The authenticity of the purified proteins was checked using an anti-His monoclonal antibody and Western blotting (Fig. 1B).
A bacterial two-hybrid assay was conducted. The fusion strains carrying full-length msmK and ftsZ grew well on the screening medium, whereas the negative control and the self-activation control did not grow (Fig. 1C). The finding confirmed the interaction of MsmK and FtsZ. Surface plasmon resonance (SPR) was then used to investigate the specificity of MsmK-FtsZ interaction in vitro. After conditioning and activation of sensor chips, His-FtsZ was quantitatively assembled on the chip surface as the ligand. Injecting full-length MsmK (Fig. 1D) or its N-terminal domain (Fig. 1E) produced a response proportional to the injected amount. No such signal was obtained when the C-terminal domain (Fig. 1F) or buffer instead of proteins was injected. At the same binding constant (K a ), the dissociation constant (K d ) or equilibrium dissociation constant (K D ) of the MsmK-FtsZ complex is lower than that obtained from the N_MsmK-FtsZ complex. This result revealed a high-affinity interaction between MsmK and FtsZ and a weak binding of N_MsmK to FtsZ.
The associations of FtsZ-A_MsmK (Fig. 1G) and FtsZ-B_MsmK (Fig. 1H) were subsequently tested. When two mutant proteins were used as analytes, the binding abilities of FtsZ-A_MsmK and FtsZ-B_MsmK were higher than that of MsmK-FtsZ. According to calculated K D values, the affinity among the four FtsZ-interacting MsmK constructs was in the order A_MsmK . B_MsmK . MsmK . N_MsmK (Fig. 1D to H). This result indicated that the deletion of the Walker A or B motif did not abolish but increased the affinities of MsmK and FtsZ. MsmK and its derivatives were further subjected to far-Western blot analysis to investigate their binding properties to FtsZ. FtsZ bound to MsmK, N_MsmK, A_MsmK, and B_MsmK but not to C_MsmK (Fig. 1I). These observations strongly support the direct interaction of MsmK with FtsZ via its N terminus in vitro.
Dual ATPase and GTPase activity of MsmK. The MsmK protein has ATP binding and ATP hydrolysis activities (15). Here, we found that MsmK could hydrolyze GTP in a concentration-dependent manner ( Fig. 2A). The amount of free GDP released from the substrate GTP increased in the presence of 1, 2, and 4 mM MsmK. Equivalent experiments on N_MsmK presented similar tendencies toward increases in GDP release compared with the full-length MsmK. Equal quantities of C_MsmK or bovine serum albumin did not show hydrolysis behavior. After the addition of ATP to the GTP enzyme reaction mixtures, the GTP hydrolysis efficiency of MsmK was decreased (Fig. 2B) compared with that of the reaction system without ATP. 32 P-labeled GTP and/or 32 P-labeled ATP were simultaneously incubated with MsmK, and the results confirmed that MsmK could display ATPase and GTPase activities (Fig. 2C).
The polymerization characteristics of MsmK and its derivatives in the presence of K 1 , Mg 21 , and ATP, and/or GTP were examined by transmission electron microscopy (TEM). As shown in Fig. 2D, the incubation of MsmK with GTP resulted in the formation of regular short slices that spontaneously gathered together. The combination of MsmK with ATP polymerized into spherical particles with various diameters. MsmK could form the third sea urchin-like form in the presence of GTP and ATP. No distinct structure was found when MsmK was incubated in buffer without GTP or ATP (data not shown). These observations indicated that MsmK can form distinct polymers in an ATPand/or GTP-dependent manner. Moreover, N_MsmK and B_MsmK showed polymerization patterns roughly similar to that of full-length MsmK in the presence of GTP, ATP, or GTP and ATP. Incubating A_MsmK with GTP still produced sea urchin-like polymers, The dotted lines represent the amino acid sequences of the motifs that have been deleted in various protein mutants. (B) Western blot assay for the characterization of the His tag-fused proteins. Purified proteins were probed with anti-His monoclonal antibodies. (C) Bacterial two-hybrid assay of the MsmK-FtsZ interaction. pBT-and pTRG-related plasmids were cotransformed into the reporter strain, and positive cotransformants were selected. The liquid culture of each cotransformant was spotted onto selective (13AT1Str; medium with 3amino-1,2,4-triazole [3-AT] and streptomycin) and nonselective (23AT-Str; medium without 3-AT and streptomycin) plates. CK1, a cotransformant containing pBT-LGF2 and pTRG-Gal as a positive control; CK2, a cotransformant containing pBT and pTRG as a negative control. A cotransformant containing pTRG-msmK and pBT was the self-activation control. (D to H) SPR assays of the interaction between FtsZ and MsmK (D), N_MsmK (E), C_MsmK (F), A_MsmK (G), and B_MsmK (H). His-tagged FtsZ was immobilized onto sensor chips. Buffer without protein was used as a negative control in all SPR assays. (D) Full-length MsmK was diluted into a series of concentration gradients and was used as the analyte. K a , 1.00Â 10 3 1/Ms; K d , 1.00Â 10 25 1/s; K D , 1.00 Â 10 28 M. "1/Ms" and "1/s" are units of the rate constants. (E) The N terminus of MsmK was diluted into three concentration gradients and was used as the analyte. K a , 1.00Â 10 3  but the ability of MsmK to form spherical polymers with ATP was severely disrupted by the deletion of the Walker A motif.
MsmK promotes FtsZ protofilament bundling in vitro. TEM was performed to investigate the morphological changes of FtsZ during protofilament formation and bundling. FtsZ assembled into an extensive network of long and thin polymers in the presence of GTP but not ATP (Fig. 3A). As shown in Fig. 3B, incubating FtsZ and MsmK with GTP resulted in the formation of thick filaments with diameters of approximately 300 nm. We observed that the scattered FtsZ protofilaments were connected by a central point, which likely corresponds to the polymerization mode of the two proteins. The mixture of N_MsmK and FtsZ with GTP could also induce protofilament bundling. Interestingly, a large number of clusters with attached FtsZ protofilaments were observed. C_MsmK showed no effect on FtsZ polymerization (see Fig. S2 in the supplemental material). However, the addition of A_MsmK or B_MsmK under the same conditions reduced the FtsZ bundling, especially in the case of A_MsmK. A large number of protein masses were attached to FtsZ protofilaments when FtsZ was incubated with A_MsmK in the presence of GTP. Also, the addition of extra ATP in mixtures negatively affected the formation of the above structures.
Ultracentrifugation was used to sediment FtsZ protofilaments and any bound proteins (Fig. 4A). The addition of increasing MsmK amounts continuously increased FtsZ sedimentation, indicating that FtsZ polymerization is positively promoted by MsmK. Incorporating ATP marginally reduced the amount of FtsZ in pellets compared with that in the mixtures containing MsmK, FtsZ, and GTP. Sediments with low efficiency of FtsZ polymer formation were observed in the presence of N_MsmK. In contrast, A_MsmK and B_MsmK inhibited FtsZ polymerization. These results confirmed TEM findings. In addition, the MsmK constructs recovered in the pellet fraction were qualitatively analyzed (Fig. 4B). Under the same conditions, the cosedimentation proportion of MsmK constructs was in the order MsmK . A_MsmK . MsmK with ATP .
Formation of the MsmK-FtsZ complex in vivo. Wild-type (WT) S. suis strain SC19 (serotype 2) was isolated from a diseased pig during an epidemic outbreak in 2005 in Sichuan, China (17). The erythromycin marker mutant (DmsmK strain) was previously prepared (15). The E. coli-S. suis shuttle vector pSET2 carrying the msmK expression cassette under the control of its native promoter was transformed into the DmsmK strain to construct the strain CDmsmK for functional complementation (15). During the growth experiment, the DmsmK strain showed a growth trend similar to that of the WT strain, although the values of the optical density at 600 nm (OD 600 ) slightly decreased at the late logarithmic phase (Fig. 5A). The protein expression level of FtsZ in the mutant was not affected, as revealed by Western blot analysis (Fig. 5B).
pSET2 carrying the msmK expression cassette with a coding sequence for a His tag at the C terminus was constructed and electroporated into the DmsmK strain, which produced the strain CDmsmK-His, to study in vivo complexes containing MsmK and FtsZ. Exponentially growing cells of CDmsmK-His were cross-linked using formaldehyde and then subjected to immunoprecipitation experiments on nickel-nitrilotriacetic acid (Ni-NTA) resins. The input mixed bacterial proteins and the final eluted protein fractions were detected by Western blotting analysis by using antibodies to His tag, MsmK, and FtsZ. Control mock pulldowns were run on cross-linked cells of CDmsmK expressing MsmK instead of the His-tagged MsmK and were used as the background control. The prey protein, endogenous FtsZ of S. suis, was attached to the bait protein (Fig. 5C), showing that MsmK could form complexes with FtsZ in vivo.
The subcellular localization of MsmK and FtsZ in S. suis cells was examined using immunofluorescence microscopy. The CDmsmK-GFP strain, which expressed MsmK with green fluorescent protein (GFP) at the C terminus (Fig. 5D), was probed using the FtsZspecific antibody and the Cy3-conjugated anti-mouse IgG. Cells at different division stages (18) were labeled (designated stages 1 to 3). The superresolution images obtained are arranged sequentially in Fig. 5E. Three-dimensional microscopy confirmed that MsmK-GFP colocalized with the ring of FtsZ throughout the cell cycle, revealing that MsmK is a FtsZ-interacting partner and localizes to the site of division.
Absence of MsmK results in the formation of chains and short cells. Gram-staining test showed that MsmK absence lengthens the bacterial chains (19). In vitro and in vivo protein assays indicated that MsmK is linked to cell division. Scanning electron microscopy (SEM) (Fig. 6A) and TEM (Fig. 6B) were accordingly performed to examine the morphology of the S. suis strains. SEM confirmed the Gram-staining findings. More chains were present in the field of view for the mutant cells than the WT cells (Fig. 6A). Analysis of TEM images showed that chains (at least three well-structured cells) were Electron micrographs revealed that WT cells appeared as characteristic elongated ellipsoids, and the bacteria lacking MsmK formed shorter cells than the WT cells (Fig. 6B). Cell longitudinal length (long axis) and cell width were measured using the automated Microbe-Tracker software (20). The distribution of distinct cell length classes differed between the WT and DmsmK cells (Fig. 7A), and their cell width classifications did not substantially vary (Fig. 7B). The WT cells were concentrated in a cell length interval of 0.  (Fig. 7C). The widths of DmsmK cells did not differ from those of WT cells (Fig. 7D). Therefore, the length-to-width ratio in mutants (1.35 6 0.28) was significantly lower than that in WT cells (1.58 6 0.21) (Fig. 7E). Calculated data for the three S. suis strains revealed that the cell wall in DmsmK cells (25.14 6 1.65 nm; n = 352) was also significantly thinner than that in WT cells (23.70 6 1.95 nm; n = 370) (Fig. 7F). No difference in capsule thickness was observed among the three strains (Fig. 7G).
Although MsmK deletion resulted in the formation of chains and nearly spherical cells, the DmsmK cells still displayed symmetrical cell division sites and equal-sized daughter cells either in chains or in diplococcal morphology (Fig. 6B). Division occurs in successive parallel planes that are perpendicular to their cell long axis, and small round cells continuously dividing with clearly visible division invagination rings were frequently observed (Fig. 6B). Consequently, the absence of MsmK did not disturb proper newborn septum placement and septal cell wall construction, thereby maintaining the capability of mutant cells to effectively divide. msmK complementation restores the mutant phenotype. The recombinant plasmid with the msmK expression cassette under the control of its native promoter was successfully expressed in the mutants (Fig. 5B), and CDmsmK cells showed phenotype characteristics similar to those of the WT cells (Fig. 6B). Analysis and comparison of cell sizes revealed that the lengths of CDmsmK cells ranged 0.8 to 1.3mm (n = 380) with an average value of 0.98 6 0.15mm (Fig. 7C). On average, the cell wall thickness of CDmsmK cells was 24.96 6 0.15 nm (n = 380). Therefore, msmK complementation fully restored the mutant morphology and size, and the above defective phenotypes of the DmsmK cells were caused by msmK deletion. In summary, lack of MsmK leads to cell rounding, suggesting that this protein is involved in the maintenance of S. suis cell shape.
MsmK absence affects peripheral PG synthesis. Peripheral cell wall synthesis occurs between the future equator and newborn septum of dividing cells (18). Blocking this process through the deletion of peripheral growth components, such as PBP2b (18) and MreCD (21) in S. pneumoniae, produces similar-looking chains of nearly spherical cells. The decreased length-to-width ratio in DmsmK cells suggests that MsmK might be involved in peripheral cell wall synthesis. Here, the morphology of nascent biosynthesized PG was examined by using TAMRA-D-lysine (TDL; a fluorescent carboxytetramethylrhodamine derivative of D-alanine) (22). S. suis strains were simultaneously labeled with TDL, membrane dyes, and 49,6-diamidino-2-phenylindole (DAPI) and subjected to structured illumination microscopy to achieve colocalization. Superresolution micrographs of the WT cells displayed a typical pattern, with cell wall synthesis labeled at the midcell and equators of future daughter cells (Fig. 8A). As expected for the disruption of peripheral PG synthesis, the septal PG walls of 30.84% (70/227) of the mutant cells were constructed in mother cells without cell elongation (Fig. 8A), thereby generating two nascent equatorial rings that were extremely close to the parental division septum. Approximately 16.74% (38/227) of the mutant cells lacked nucleoids (Fig. 8A), and minicells (cell length of approximately 0.31mm) with no nucleoid were observed. Moreover, part of the residual structure of PG did not fully disintegrate, resulting in incomplete cell separation.
S. suis strains were grown in the presence of HCC-amino-D-alanine (HADA), a fluorescent hydroxycoumarin derivative of D-alanine (22), for one or two generations to label the cell contour and further validate the effect of the absence of MsmK on peripheral PG synthesis. Compared with WT and CDmsmK cells, part of the mutant cells displayed visualized septal walls and poles but an invisible peripheral cell wall (Fig. 8B). The missing blue contours in the DmsmK cells represent the unlabeled older cell wall, in which new PGs were not inserted into the corresponding area at least in one or two generations. Therefore, MsmK is a component of the elongation system of S. suis.
MsmK is not the phosphorylation target of STKs. Eukaryotic-type serine/threonine kinases (STKs) are control elements in ovococci that coordinate several key processes over the cell cycle (5). The S. suis genome encodes a single STK with autophosphorylation and substrate phosphorylation activities and has several cell division-related targets, including FtsZ, FtsA, GpsB, DivIVA, and MapZ (23). After incubation with StkP and [g-32 P]ATP in vitro, no phosphorylation signal was observed on MsmK or its fragments (Fig. S3), thereby indicating that MsmK could not be phosphorylated by STK.

DISCUSSION
Faithful cell proliferation relies on the strict coordination of related cell division processes (3). To date, more than 30 proteins are proven or likely to function in the divisome and/or the elongation machineries in pneumococcal growth and division (3). Here, we identified a new cell division protein, MsmK, an ATPase that provides energy for multiple-carbohydrate transport in S. suis (15). The absence of MsmK in cells is not lethal but results in more chains and shorter daughter cells, some of which are anuclear, compared with the WT strain. These abnormal cells still possess symmetrical morphology postsplitting, indicating that the lack of MsmK does not affect the division site selection, whereas the absence of DivIVA (24), LocZ (20), or EzrA (25) causes cells to divide asymmetrically in ovococci. In addition, the absence of MsmK does not influence septal wall synthesis and cell division, except for the decreased separation degree between daughter cells.
The two-state model of PG synthesis in ovococci indicates the possibilities of ovococcus-to-rod or ovococcus-to-coccus transitions when septal and peripheral PG syntheses are separately inhibited (1). This hypothesis has been tested and approved by functional observations in several species, such as S. pneumoniae, Lactococcus lactis, Streptococcus agalactiae, Streptococcus bovis, and Enterococcus hirae (1). Conversely, related proteins could also be assigned to septal or peripheral PG synthesis based on cell morphology defects caused by deletion (18). Proteins with different functions, such as MreCD, RodA, RodZ, PBP1a, and PBP2b (4), are assigned to the elongation machinery of ovococci, and their deletions result in the same round phenotype. With regard to the production of chains and spherical cells caused by the absence of MsmK, we speculated that MsmK is involved in peripheral but not in septal wall synthesis. Subsequent electron and superresolution micrographs supported this hypothesis.
The function of MsmK in peripheral PG wall synthesis remains unclear. MsmK directly interacts with FtsZ in vitro and colocalizes with FtsZ in vivo. FtsZ has been reported to coordinate and organize not only septal PG synthesis but also peripheral PG synthesis in ovococci, similarly to the role of MreB in preseptal elongation in rods (1). Therefore, MsmK may affect peripheral PG synthesis by interacting with FtsZ and/ or regulating FtsZ polymerization.
The mechanisms underlying the role of MsmK in the peripheral growth may be inspired by the conserved FtsEX complex. The FtsE/X complex is shown to be involved in activating cell wall hydrolases that impact the PG remodeling during cellular elongation in a variety of Gram-positive species (1). The membrane-bound FtsX interacts with the cytoplasmic FtsE ATPase subunit, structurally resembling an ABC transporter (26). The extracellular protein PcsB is the only essential PG autolysin (27), and the FtsEX complex is required for the PcsB function. The PG remodeling activity by PcsB is coordinated with cell division through its interaction with the FtsEX complex, which in turn cooperates with FtsZ and other division proteins through FtsE in S. pneumoniae (26). Notably, S. suis possesses the coding sequences of FtsE (SSUSC84_1271) and FtsX (SSUSC84_1270). Moreover, the chain phenotype of the mutant suggests that the PcsB function is affected in the absence of MsmK. As the ATPase subunit in S. suis (15) and S. pneumoniae (28), MsmK energizes multiple ABC transporters. Hence, MsmK may play the same bridge role as FtsE by linking FtsZ and specific functional proteins involved in the elongation system. The relationship between MsmK, FtsZ, FtsE/FtsX, PcsB, and other components involved in ABC transporter systems containing MsmK needs further study.
A number of proteins, such as CvaB of E. coli (29), SpoIVA of B. subtilis (30), Rep of beak and feather disease virus (31), and NS4B of classical swine fever virus (32), act as nucleoside triphosphatases (NTPases) that hydrolyze ATP and GTP with various levels of activity. Their ATPase and GTPase activities are associated with cell signaling, spore developing, viral replication, or membrane fusion. FeoB, a large cytoplasmic-membrane-associated protein of the ferrous iron transporter system, exhibits GTPase and ATPase activities in Vibrio cholerae and Helicobacter pylori but is solely a GTPase in E. coli (33). Here, we found that MsmK can display ATPase and GTPase activities. TEM and ultracentrifugation experiments showed that MsmK promotes FtsZ protofilament bundling in the GTP-dependent manner, and the addition of ATP decreases this promoting function. Considering that MsmK correspondingly forms distinct polymers in the presence of GTP or ATP, we speculate that the polymerized structure of MsmK relying on its GTPase activity is critical for its function in FtsZ polymerization. In this case, the dual enzyme activities of MsmK are responsible for its different functions. In addition, a series of proteins, such as ZapC (34), ZapA (35), and MciZ (36), promotes polymer synthesis by inhibiting the GTPase activity of FtsZ. Thus, MsmK might also competitively combine with GTP and disturb the GTPase activity of FtsZ, thereby promoting FtsZ polymerization.
MsmK interacts with FtsZ via its N terminus. However, the N-terminal fragment of MsmK showed lower-affinity interaction with FtsZ than full-length MsmK. The N-terminal fragment also showed low efficiency in promoting FtsZ polymerization. These results illustrated that the C-terminal region is important for the functions of MsmK. Walker A and B motifs contribute to the ATPase and GTPase activities of an NTPase, and mutations in the A motif have a relatively remarkable effect on the NTPase activity (31,32). This phenomenon was also observed through the TEM analysis of A_MsmK and B_MsmK. Although the two protein mutants showed strong affinity for FtsZ, they inhibited FtsZ polymerization in vitro. That may be because they were unable to form appropriate structures due to the impaired GTPase activity, thereby severely affecting the formation and bundling of FtsZ protofilaments.
The E1a subunit of pyruvate dehydrogenase is required for the metabolism of pyruvate at the final stage of glycolysis and positively regulates Z-ring assembly by colocalizing with the chromosome in a pyruvate-dependent manner (37). Gluconate 5-dehydrogenase, which catalyzes the reversible reduction of 5-ketogluconate to D-gluconate, is localized to the cell division site and prevents ectopic Z-ring formation during cell division in S. suis (14). These findings clarified that nutritional conditions coordinate with bacterial division to ensure the survival of newly formed cells (38). We propose that MsmK is a nonessential component of cell division. MsmK is not a phosphorylated substrate of STK and thus is not regulated by STK. As a component of the carbohydrate ABC transporter, MsmK expression is naturally regulated by the nutrient level (39), suggesting that MsmK may be involved in coordinating bacterial division with nutrient availability. Whether MsmK serves as a linker between nutrition and cell division needs further study. Finally, MsmK is a nonunique protein that exists in numerous bacteria and is conserved independently of phylogenetic morphogenesis, thereby implying that its participation in cell division may be a widespread and fundamental physiological process in bacteria. Further developments in this area will ultimately help us understand, at a systems biology level, how bacteria can faithfully and continually multiply in a constantly changing environment.

MATERIALS AND METHODS
Bacterial strains, plasmids, and growth conditions. Bacterial strains and plasmids are listed in Table S1. The WT strain (SC19, S. suis serotype 2) was isolated from a sick pig during an epidemic outbreak of S. suis infections in Sichuan Province, China (17). The msmK gene of SC19 was deleted via a double-crossover method through homologous recombination to generate a DmsmK strain (15). For the construction of complemented strain CDmsmK, the recombinant vector P msmK -msmK was transformed into DmsmK cells (15). The construction of other S. suis strains is detailed in Text S1. All S. suis strains  Table S2 in the supplemental material. Cloning and genetic manipulations were conducted using standard techniques. Recombinant plasmids and strains, protein techniques, general immunoblotting assays, bacterial two-hybrid and SPR analyses, and GTPase, pelleting, and kinase assays were performed as described in Text S1.
Coimmunoprecipitation. In vivo interaction assays were performed as described previously (40) with few modifications. Cultures of the CDmsmK and CDmsmK-His strains were grown exponentially in TSB to an OD 600 of ;0.6 to 0.8. Cells were collected, washed, and resuspended with 1Â phosphate-buffered saline (PBS) buffer at 4°C. A 10% paraformaldehyde solution was added for cross-linking to a final concentration of 0.1% (vol/vol). Mixtures were incubated at 37°C for 1 h. Exactly 1.0 M glycine was added and incubated at 25°C for 10 min to abolish cross-linking reactions. The cells were then collected, washed, and resuspended with lysis buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 5 mg/ml lysozyme, 1% Triton X-100 (vol/vol), and 1% phenylmethylsulfonyl fluoride (vol/vol). After sonication, bacterial debris was removed through centrifugation at 14,000 Â g for 10 min at 4°C. The supernatant of each strain was incubated with Ni-NTA columns (GE Healthcare, USA) at 4°C for 2 h in binding buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, and 25 mM imidazole. The beads were washed four times with 80 mM imidazole for 5 min each. His-tagged protein was eluted by incubating with 250 mM imidazole for 1.5 h at 4°C. The original supernatant (input) and the final elution of each strain were separated by polyacrylamide gel electrophoresis and transferred to a polyvinylidene fluoride membrane (Invitrogen). The membrane was then probed with mouse antihistidase antibody (1:4,000; Abcam, China), in-house-prepared anti-MsmK serum (1:1,000), or in-house-prepared anti-FtsZ serum (1:1,000). This experiment was performed in triplicate.
Electron microscopy. All purified proteins were precentrifuged for 10 min at 16,000 Â g at 4°C. Proteins (0.6 mM each) were incubated with 2 mM GTP and/or 1 mM ATP for 10 min at 37°C in buffer P (Text S1) to observe the structure formation of MsmK and its derivatives. For observing FtsZ polymerization in vitro, mixtures of 0.6 mM FtsZ and/or His-tagged MsmK constructs (0.6 mM each) were incubated at 37°C for 2 min in buffer P to analyze FtsZ polymerization in vitro. Afterward, 2 mM GTP and/or 1 mM ATP was added into the mixtures, which were continuously incubated for 10 min at 37°C. Mixtures were then withdrawn and applied to glow-discharged carbon-coated grids. The grids were air-dried and visualized using an H-7650 TEM (Hitachi, Japan).
The cells in TSB were grown to an OD 600 of 0.5 and fixed overnight with 2.5% glutaraldehyde at 4°C to observe the bacterial morphology under TEM. For SEM, the cell cultures were grown in TSB to an OD 600 of 0.5 and spotted onto poly-L-lysine-coated coverslips, followed by washing with PBS buffer. The cells were fixed overnight with 2.5% glutaraldehyde at 4°C. Subsequent dehydration steps were performed using ethanol as previously described (19). The dried samples were covered with a 10-nm-thick layer of gold and observed with a JSM-6390LV SEM (JEOL, Japan).
Fluorescence microscopy. The immunofluorescence microscopy of cytoplasmic proteins was performed as previously described (14), with a few modifications. The CDmsmK-GFP strain expressing the MsmK-GFP fusion protein was cultivated in TSB at 37°C until an OD 600 of 0.6 was achieved. After washing three times with precooled PBS, cultures were incubated with 100% formaldehyde for 20 min at 220°C and with 80% formaldehyde for 1 h at 28°C. Cells were washed once with 80% methanol and permeabilized with 0.3% Triton X-100 in PBS for 30 min at 28°C. After blocking with 5% bovine serum albumin for 30 min at 37°C, cells were sequentially probed with the mouse anti-FtsZ serum (1:100) and the goat antimouse IgG (heavy plus light chain [H1L])-Cy3 conjugate (1:200; Proteintech Group. Inc., USA) for 1 h at 37°C. Cells were transferred onto microscope slides and covered by 1% agarose gel. Superresolution images were obtained using a Nikon structured illumination microscope (N-SIM; Nikon Instruments, Inc.) equipped with an electron-multiplying charge-coupled device camera (iXon DU-897; Andor) and a total internal reflection fluorescence objective (100Â; numerical aperture [NA], 1.49; CFI Apo TIRF; Nikon Instruments, Inc.). Images were captured with multiple phases and angles of illumination pattern before direct reconstruction by using the Nikon NIS-Elements AR software.
TDL and HADA (22) were synthesized by Wuxi AppTec Co., Ltd. TDL or HADA was dissolved in dimethyl sulfoxide (up to 0.2 M) to probe peptidoglycan synthesis. Alexa Fluor 647 (AF 647; Thermo Scientific) dissolved in anhydrous acetonitrile (up to 10 mg/ml) was prepared for cell membrane staining. For the colocalization of PG cell walls, cell membranes, and nucleoids, cultures of WT, DmsmK, and CDmsmK strains were grown to an OD 600 of 0.5 and then centrifuged at 9,000 Â g for 5 min. After washing with PBS, the cells were resuspended in 1 ml of PBS containing 5 ml of TDL and 5 ml of AF 647. The mixtures were incubated for 1 h at 37°C in the dark. The cells were then washed with PBS and stained with 1.5 mg/ml DAPI for 5 min in the dark at room temperature. Following the same incubation and pretreatment steps, the cells were incubated in 1 ml of PBS containing 5 ml of HADA for 30 min at 37°C in the dark to observe the cell wall contours.
Image analysis. Cell sizes were measured with the automated MicrobeTracker software (20) by using TEM images from three independent experiments. GTPase assays and pelleting assays were quantitatively analyzed using ImageJ2x software. Statistical analyses were performed via unpaired Student's t tests in GraphPad Prism 6.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. TEXT S1, DOC file, 0.1 MB. FIG S1, TIF