PrkC, a Transmembrane Serine/Threonine Protein Kinase, Regulates Bacterial Chain Length in Bacillus anthracis

Chaining phenotype acts as a virulence factor in Bacillus anthracis. This is the first study that identifies a signal transduction protein with an ability to regulate the chaining phenotype in Bacillus anthracis.

ABSTRACT

Anthrax is a zoonotic disease caused by Bacillus anthracis, a spore-forming pathogen that displays a chaining phenotype. It has been reported that the chaining phenotype acts as a virulence factor in B. anthracis. In this study, we identify a serine/threonine protein kinase of B. anthracis, PrkC, the only kinase localized at the bacterium-host interface, as a determinant of B. anthracis chain length. In vitro, a prkC disruption strain (BAS ΔprkC) grew as shorter chains throughout the bacterial growth cycle. A comparative analysis between the parent strain and the BAS ΔprkC strain indicated that the levels of proteins BslO and Sap, associated with the regulation of bacterial chain length, were upregulated in the BAS ΔprkC strain. BslO is a septal murein hydrolase that catalyzes daughter cell separation, and Sap is an S-layer structural protein required for the septal localization of BslO. PrkC disruption also had a significant effect on bacterial growth, cell wall thickness, and septum formation. Upregulation of ftsZ in the BAS ΔprkC strain was also observed. Altogether, our results indicate that PrkC is required for maintaining optimum growth, cell wall homeostasis, and, most importantly, the chaining phenotype.

IMPORTANCE Chaining phenotype acts as a virulence factor in Bacillus anthracis. This is the first study that identifies a signal transduction protein with an ability to regulate the chaining phenotype in Bacillus anthracis. We show that the disruption of the lone surface-localized serine/threonine protein kinase, PrkC, leads to the shortening of the bacterial chains. We report upregulation of the dechaining proteins in the PrkC disruption strain. Apart from this, we also report, for the first time, that PrkC disruption results in attenuated cell growth, a decrease in cell wall thickness, and aberrant cell septum formation during the logarithmic phase of growth, a growth phase where PrkC is expressed maximally.

INTRODUCTION

Bacteria exhibit diverse shapes and morphologies, the result of long evolutionary processes that select genotypes best suited for bacterial survival. Apart from the variation in shape, bacteria display multicellular structures, such as aggregates, biofilms, and chains/filaments (1, 2). Bacillus anthracis, a Gram-positive spore-forming pathogen of grazing mammals and the etiological agent of anthrax, grows as chains of rod-shaped cells (3–6). In the environment, B. anthracis persists primarily as metabolically inert oblong spores. Germination happens in the presence of an optimal signal within the host (7).

B. anthracis spores can infect humans through three routes, gastrointestinal, inhalational, and cutaneous (7–10). Multiple factors act as virulence determinants (11–13). However, the secreted binary exotoxins (lethal toxin and edema toxin) and the antiphagocytic poly-γ-d-glutamic acid capsule, encoded by the virulence plasmids pXO1 and pXO2, respectively, act as the primary virulence factors (7, 14, 15).

Among other factors that play a role in virulence, the bacterial chaining phenotype has been shown to contribute significantly (4, 5). During initial stages of infection, B. anthracis spores phagocytosed by macrophages germinate and grow in chains before causing cell rupture (16). In mice, the high pathogenicity of systemically inoculated B. anthracis strains making capsules but not the toxins (encapsulated but nontoxinogenic strain) was linked to chain length-dependent blockade of alveolar capillaries, leading to hypoxia, lung tissue injury, and death (4, 5). Of note, the lung is the terminal organ targeted by B. anthracis, irrespective of the route of infection (5, 17, 18). These studies indicate that the chaining phenotype presents a survival advantage to B. anthracis within its host during both early and late stages of infection.

Intrigued by the relevance of this morphotype to the biology of Bacillus species, various groups have tried to identify the mechanisms controlling bacterial chain length in both pathogenic and nonpathogenic strains (19–27). In B. anthracis, one of the determinants of bacterial chain length is the septal peptidoglycan hydrolase, BslO (Bacillus surface layer O). BslO is a Bacillus S-layer-associated protein (BSL) with N-acetylglucosaminidase activity that catalyzes daughter cell separation (22). Restrictive deposition of BslO to the septal region is, in turn, attributed to sequential coverage of the cell wall by the primary S-layer proteins (SLPs), Sap (surface array protein), and EA1 (extractable antigen 1) (21). SLPs and BSLs associate with the pyruvylated secondary cell wall polysaccharides (SCWP) through their conserved S-layer homology domain (SLH) (28, 29). While several enzymes that influence the chaining phenotype through their role in synthesis/modification of SCWPs and, hence, the attachment of SLPs to SCWPs, have been identified in B. anthracis (29, 30), a sensory molecule with the potential to regulate the chaining phenotype, possibly through regulation of one of these factors, remains unknown.

Membrane-localized serine/threonine protein kinases (STPKs) containing extracellular PASTA (penicillin-binding proteins and Ser/Thr kinase-associated) repeats are known to sense external stimulus and relay it to the cellular core (31). In B. anthracis, three STPKs have been characterized, namely, PrkC, PrkD, and PrkG (32–34). Among these, PrkC is the only membrane-associated protein kinase with the peptidoglycan binding PASTA repeats (33, 35). In B. subtilis and B. anthracis, PrkC has been shown to interact with peptidoglycan fragments generated by neighboring growing cells, triggering germination of the spores (36). Apart from a role in germination, B. subtilis PrkC has been implicated in stationary-phase processes, cell wall metabolism, cell division, sporulation, and biofilm formation (35, 37–41).

In this study, we identify B. anthracis PrkC as a determinant of bacterial chain length. We show that the B. anthracis Sterne 34F2 prkC mutant strain (BAS ΔprkC) cannot attain a chaining phenotype throughout the bacterial growth cycle. Both BslO and Sap are found to be upregulated in the BAS ΔprkC strain, which probably creates a condition that favors dechaining. PrkC is also shown to influence bacterial cell division, possibly through the regulation of the cytoskeletal protein, FtsZ. Through this work, we propose that PrkC, a transmembrane kinase with a sensor domain, perceives growth-permissive signals and maintains the levels of the primary proteins involved in dechaining to regulate the chaining phenotype.

RESULTS

prkC disruption results in bacteria with shorter chain length.

Previously, our group had shown that B. anthracis PrkC-mediated phosphorylation plays an essential role in germination and biofilm formation, and some of the components of the PrkC-mediated signaling cascade leading to these processes were identified (GroEL and enolase) (42, 43). Work done by other groups implicated B. subtilis PrkC in later stages of bacterial growth and germination (35, 36, 38). Even though prkC is expressed maximally during the logarithmic phase of in vitro growth, prkC deletion has never been reported to result in any apparent defect in morphology, viability, or growth during this phase in either B. subtilis or B. anthracis (32–35, 37, 38, 44). PrkC is, however, recognized as an infection-specific kinase and is critical for B. anthracis survival in macrophages (33, 34).

While working on the B. anthracis Sterne 34F2 prkC mutant strain (BAS ΔprkC), we observed that logarithmic cultures of the BAS ΔprkC strain allowed to stand at room temperature formed a compact pellet, whereas the parental wild-type strain (BAS WT) did not (Fig. 1A). In a study on a B. anthracis bslO mutant strain, Anderson et al. had shown that compact pellets were formed when bacteria grew as shorter chains, while loose pellets were formed when bacteria exhibited extensive chaining (22). This suggested that the absence of PrkC was leading to the shortening of the bacterial chains. To validate this, exponentially growing BAS WT and BAS ΔprkC strains were visualized under a phase-contrast microscope. As shown in Fig. 1B, disruption of prkC resulted in bacteria growing as shorter chains, and this phenotype was reversed in a prkC-complemented strain (BAS ΔprkC::prkC). Further, to determine if prkC disruption resulted in a defect in cell morphology that causes bulging, shrinking, or changes in cell width or shape, BAS WT, BAS ΔprkC, and BAS ΔprkC::prkC bacterial cells were examined by a scanning electron microscope (SEM). However, as seen in Fig. 1C, no morphological defect was apparent, apart from the shortening of bacterial chains, indicating that the prkC disruption influenced only bacterial chain length.

FIG 1.

prkC disruption results in bacteria with short chain length. (A) Photograph of culture sediments in microcentrifuge tubes after standing incubation (9 h) at room temperature of BAS WT (left) and the BAS ΔprkC strain (right) grown in LB medium. (B) Phase-contrast images of BAS WT, BAS ΔprkC, and BAS ΔprkC::prkC strains in mid-log phase. Cells were grown in LB broth at 37°C, and a 1-ml sample was taken from cultures in mid-log phase. Cells were pelleted and washed with PBS and visualized under the 100×/1.4 oil DIC objective of a Zeiss Axio Imager Z2 upright microscope. Scale bar, 10 μm. (C) Scanning electron microscopy of BAS WT, BAS ΔprkC, and BAS ΔprkC::prkC strains in mid-log phase. Cells were grown in LB broth at 37°C and harvested in mid-log phase. These were then washed with 0.1 M sodium phosphate buffer and fixed with Karnovsky’s fixative, followed by 1% osmium tetroxide. A critical-point drying technique was used for drying the samples, followed by gold coating of 10 nm using an aluminum stub coated with agar sputter. Cells were visualized under a Zeiss Evo LS15. Scale bars, 2 μm; magnification, ×5,000.

Effect of prkC disruption on chaining morphotype during different phases of bacterial growth.

If PrkC is the sensor molecule required for maintaining the chaining phenotype, its absence from the BAS ΔprkC strain would result in shorter chains throughout the bacterial growth cycle. To examine this and to provide a basis for our experiments, we first monitored the growth of the BAS WT strain through the entire growth cycle (Fig. 2A). To determine growth stage-specific changes in chaining phenotype, culture aliquots were taken out at indicated time points and observed under a phase-contrast microscope. As seen in Fig. 2C and 3, BAS WT exhibited extensive chaining until an optical density (OD; A₆₀₀) of ∼3.0, at 4 h, after which a sudden shortening of bacterial chains was observed. The average chain length at 4 h was measured as 115.90 μm (standard deviation [SD], ±46.271 μm; n = 50), while at 5 h it shortened to 63.53 μm (±20.1150 μm; n = 50) (Fig. 3). Interestingly, this time point correlated with the end of the exponential phase and the start of the deceleration phase (Fig. 2A), a stage where bacterial replication rate starts decreasing owing to nutrient deprivation and accumulation of metabolic by-products (45). Next, to determine the effect of prkC disruption on the chaining phenotype, similar growth curve analysis and chain length determinations/measurements were carried out for the BAS ΔprkC strain as described for BAS WT (Fig. 2B and C and 3). As shown in Fig. 2C and 3, the BAS ΔprkC strain grew as shorter chains throughout the growth cycle. Of note, we did observe some chaining in BAS ΔprkC cultures during the lag phase (t = 2 h) (Fig. 2C), which could be due to an insufficient number of the dechaining molecule(s) synthesized at this stage or the time taken for the assembly of these molecules at the desired location. In the presence of PrkC, synthesis of this molecule(s) is probably downregulated to allow bacteria to grow as chains. Introduction of an ectopic copy of prkC (BAS ΔprkC::prkC strain) also reverted the bacterial chaining morphology comparably to that of BAS WT cells. We observed 86% reversion in chain length upon complementation at the 3-h time point (see Fig. S2 in the supplemental material). These results indicate that PrkC senses a growth-permissive signal(s) and regulates the levels of molecules associated with dechaining to maintain the long-chain phenotype, a morphology found during nutrient abundance (3, 22).

FIG 2.

Effect of prkC disruption on chaining morphotype during different phases of bacterial growth. (A) Growth kinetics of BAS WT. BAS WT strain was grown in LB broth at 37°C. For absorbance, the OD (A₆₀₀) was recorded at the indicated time points. Error bars denote standard deviations, n = 3. (B) Growth kinetics of the BAS ΔprkC strain. BAS ΔprkC strain was grown in LB broth at 37°C. For absorbance, the OD (A₆₀₀) was recorded at the indicated time points. Error bars denote standard deviations, n = 3. (C) Phase-contrast images of BAS WT and BAS ΔprkC strains at different phases of the bacterial growth cycle. Cells were grown at 37°C in LB broth, and 1 ml sample was harvested at the time points indicated in panels A and B. Cells were pelleted and washed with PBS and visualized under a 100×/1.4 oil DIC objective of a Zeiss Axio Imager Z2 upright microscope. Scale bars, 10 μm.

FIG 3.

Quantitative analysis of chain length variation in prkC disruption strain. Shown is a scatter dot plot denoting BAS WT and BAS ΔprkC strain chain length measurement during different phases of bacterial growth. Phase-contrast images of BAS WT and BAS ΔprkC strains were used for measurement of the bacterial chain length using ImageJ software. Some data points in chain length quantitation (BAS WT at 2 h, 3 h, and 4 h) represent the maximum observable chain length obtained in the phase-contrast images. The vertical and horizontal black lines in the data set denote SD and means, respectively, for both strains at the indicated time points, n = 50. These values are indicated in a separate table. Statistical significance of chain length distribution in BAS WT and BAS ΔprkC strains was analyzed using two-way ANOVA and denoted in the graph in the form of asterisks indicating significant results. ****, P < 0.0001; ns, not significant.

PrkC regulates the expression of Sap, EA1, and BslO.

In B. anthracis, Sap and then EA1 sequentially form monomeric paracrystalline bidimensional surface S-layers during exponential and stationary growth phases, respectively (21, 46). The saturating presence of Sap and EA1 on the cell wall confines the S-layer-associated protein BslO (with a similar SLH domain) to the septal region (21). Disruption of sap has been shown to cause chain length elongation mainly because, in the absence of Sap, BslO is no longer restricted to the septal region and, hence, is incapable of carrying out murein hydrolysis effectively (21). To understand whether PrkC maintains the chaining phenotype through modulating the levels of Sap, BslO, and EA1, their expression levels were determined at the indicated time points in the BAS WT and BAS ΔprkC strains, using GroEL as an internal control (Fig. 2A and B; see also Fig. S3 in the supplemental material). As shown in Fig. 4A, prkC disruption resulted in the upregulation of Sap at most of the time points for which the samples were collected. In BAS WT, a sharp increase in expression was observed toward the end of the exponential growth phase (OD [A₆₀₀], ∼3; time, 4 h) (Fig. 2A and 4A). Interestingly, the initial time points when Sap expression was low were also the time points where long chains were observed (2 h and 3 h) (Fig. 2C, 3, and 4A). However, in the BAS ΔprkC strain, Sap levels were found to be higher than those of the BAS WT strain even at the initial time points (2 h and 3 h) (Fig. 4A and S5). This probably is the reason for the dechaining observed in the BAS ΔprkC strain from the beginning of the growth cycle. These results are in agreement with the previous report, where the absence of Sap was shown to result in a long-chain phenotype (21).

FIG 4.

PrkC regulates the expression of Sap, BslO, and EA1. (A) Differential expression of Sap protein in BAS WT and BAS ΔprkC strains. Equal amounts of protein at different time points (2 h, 3 h, 4 h, 5 h, 6 h, 7 h, 8 h, 9 h, 10 h, 14 h, 18 h, 22 h, 26 h, and 30 h) were loaded onto an SDS-PAGE gel, transferred onto a nitrocellulose membrane, and probed by anti-Sap antibody (1:50,000) raised in rabbit. The same blot was then stripped and probed by anti-GroEL antibody (1:50,000) raised in mice. Densitometry analysis was done using ImageLab software, and the Sap/GroEL ratio was used to plot a graph showing differential expression of Sap protein in the BAS WT and BAS ΔprkC strains in a growth phase-dependent manner. Error bars denote standard deviations, n = 3. Representative images are from one of the three independent experiments. (B) Differential expression of BslO protein in BAS WT and BAS ΔprkC strains. Equal amounts of protein at different time points (2 h, 3 h, 4 h, 5 h, 6 h, 7 h, 8 h, 9 h, 10 h, 14 h, 18 h, 22 h, 26 h, and 30 h) were loaded onto an SDS-PAGE gel, transferred onto a nitrocellulose membrane, and probed by anti-BslO antibody (1:10,000) raised in mice. The same blot was then stripped and probed by anti-GroEL antibody (1:50,000) raised in mice. Densitometry analysis was done using ImageLab software, and the BslO/GroEL ratio was used to plot a graph showing differential expression of BslO protein in the BAS WT and BAS ΔprkC strains in a growth-dependent manner. Error bars denote standard deviations, n = 3. Representative images are from one of the three independent experiments. (C) Differential expression of EA1 protein in BAS WT and BAS ΔprkC strains. Equal amount of protein at different time points (2 h, 3 h, 4 h, 5 h, 6 h, 7 h, 8 h, 9 h, 10 h, 14 h, 18 h, 22 h, 26 h, and 30 h) was loaded onto an SDS-PAGE gel, transferred onto a nitrocellulose membrane, and probed by anti-EA1 antibody (1:50,000) raised in rabbit. The same blot was then stripped and probed by anti-GroEL antibody (1:50,000) raised in mice. Densitometry analysis was done using ImageLab software, and the EA1/GroEL ratio was used to plot a graph showing differential expression of EA1 protein in the BAS WT and BAS ΔprkC strains in a growth-dependent manner. Error bars denote standard deviations, n = 3. Representative images are from one of the three independent experiments.

Next, we wanted to determine the levels of BslO in the BAS WT and BAS ΔprkC strains. Experiments were conducted in a manner similar to that described above. Interestingly, we observed the stable expression of BslO throughout the growth cycle in BAS WT (Fig. 4B). Per our understanding, this is the first report where the levels of BslO have been determined at various stages of bacterial growth. Previous studies have conclusively established the role of BslO in dechaining and have identified it as the primary murein hydrolase driving the dechaining process (21, 22). As observed for Sap, prkC disruption resulted in the upregulation of BslO at most of the time points for which the samples were collected (Fig. 4B and S5). The results obtained for Sap and BslO indicated that an increase in the levels of Sap and BslO in the absence of PrkC creates a condition that is most suitable for dechaining. Increased Sap would restrict BslO to the septal region, which would carry out dechaining, and an increase in the levels of BslO would further add to this effect. Apart from this, to confirm if the higher expression of BslO alone can trigger dechaining, we introduced an ectopic copy of bslO in plasmid pYS5 via electroporation in BAS WT (BAS WT::bslO strain). We observed a striking decrease in the bacterial chain length, suggesting that BslO level in the cell directly correlates to chain length phenotype (Fig. S4).

Further, we determined the levels of EA1, another structural S-layer protein, that shows its presence as the culture approaches the stationary phase (21, 46). As seen in Fig. 4C, EA1 levels in the BAS ΔprkC strain were downregulated at most of the time points for which the samples were collected. Since Sap acts as a transcriptional repressor of the extractable antigen gene encoding EA1 (21, 46), this decrease can be due to the increased levels of Sap in the BAS ΔprkC strain (Fig. 4A). Additionally, introduction of an ectopic copy of prkC in the mutant strain (BAS ΔprkC::prkC strain) complemented the expression level of Sap and BslO comparably to the BAS WT (Fig. S6).

Altogether, these results indicate that during bacterial growth, the presence of PrkC helps in the maintenance of an optimum level of BslO, Sap, and EA1 to maintain the chaining phenotype.

prkC disruption results in decreased cell wall width, cell septum thickness, and increased multiseptum formation.

Throughout the course of these experiments, we observed that the BAS ΔprkC growth curve was not superimposable on that of BAS WT (Fig. 2A and B). To validate our observation, we carried out a comparative growth curve analysis with BAS WT and BAS ΔprkC strains until extended stationary phase. Interestingly, as shown in Fig. 5A, the BAS ΔprkC strain showed an attenuated replication rate throughout the bacterial growth cycle. Disruption of chaining alone could not have caused the observed defect in growth. To identify the reason(s) for the observed defect, we carried out microscopic analysis at the ultrastructural level, and both BAS WT and BAS ΔprkC strains were subjected to transmission electron microscopy. Interestingly, at the ultrastructural level, an apparent decrease in the cell wall width and septal thickness was observed in the prkC disruption strain (mid-log phase) (Fig. 5B and C and S7). This overall decrease in cell wall thickness both at septal and nonseptal regions made BAS ΔprkC cells more sensitive to cell wall-targeting antimicrobials like ceftazidime (30 μg) and cefepime (30 μg) than the BAS WT. In contrast, there was no significant difference between BAS WT and BAS ΔprkC when the cells were treated with the protein synthesis inhibitor, erythromycin (Fig. 6).

FIG 5.

prkC disruption results in decreased cell wall width and septum thickness. (A) Growth kinetics of BAS WT and BAS ΔprkC strains. Bacterial strains were grown in LB broth at 37°C for an extended period of 65 h. For absorbance, the OD (A₆₀₀) was recorded at the indicated time points. Error bars denote standard deviations, n = 3. The inset shows the expanded growth profile of BAS WT and BAS ΔprkC strains up to 6 h. (B) Transmission electron micrographs representing the ultrastructural details of difference in cell wall thickness and septum thickness of BAS WT and BAS ΔprkC strains. Cells were harvested at mid-log phase, and primary fixation was done using Karnovsky’s fixative. Secondary fixation was done using 1% osmium tetroxide, and the samples were embedded into araldite resin mixture (TAAB). Scale bars, 100 nm. (C) Bar graph representing the difference in cell wall thickness of BAS WT and BAS ΔprkC strains. Transverse sections of around 100 mid-log-phase cells of each strain were used to calculate the cell wall thickness and plotted. Statistical significance of the data set was analyzed using two-tailed Student's t test and denoted in the graph in the form of asterisks indicating significant results. ****, P < 0.0001.

FIG 6.

Sensitivity of the ΔprkC strain to cell wall-targeting antibiotics. (A) Representative LB agar plate images for antimicrobial susceptibility profile using BAS WT, BAS ΔprkC, and BAS ΔprkC::prkC strains with 30 μg ceftazidime (CAZ) (column I), 30 μg cefepime (FEP) (column II), and 15 μg erythromycin (E) (column III). (B) Bar graph denoting inhibition zone diameter for BAS WT, BAS ΔprkC, and BAS ΔprkC::prkC strains. Antimicrobial susceptibility disks were placed on LB agar plates containing cefepime (30 μg), ceftazidime (30 μg), and erythromycin (15 μg). After 24 h of incubation at 37°C, the inhibition zone was measured and plotted using GraphPad software (Prism 9) and analyzed by t test for statistical significance (*, P < 0.05).

Additionally, we also observed an increase in multiseptum formation in the prkC disruption strain during later stages of bacterial growth through both confocal and transmission electron microscopy (Fig. 7A to C). PASTA domain-containing kinases from other bacterial species have been shown to play a role in the regulation of cell division machinery and cell wall homeostasis (31). We surmise that similar signaling mechanisms are operational in B. anthracis as well.

FIG 7.

prkC disruption results in increased multiseptum formation. (A) Representative transmission electron microscopy images of BAS WT and BAS ΔprkC stationary-phase cells showing multiseptum formation. Cells were harvested at stationary phase, and primary fixation was done using Karnovsky’s fixative. Secondary fixation was done using 1% osmium tetroxide, and the samples were embedded into araldite resin mixture (TAAB). Scale bar, 100 nm. (B) Staining of live bacterial cells with FM4-64 membrane stain. BAS WT and BAS ΔprkC strains grown up to exponential phase were diluted to an initial OD (A₆₀₀) of 0.035, and 1 μl was spread on an agarose pad. The pads were incubated at 37°C. Cell membrane was stained with FM4-64 (final concentration of 1 μg/ml), and images were captured by a Leica SP8 confocal microscope at 3 h (upper) and 12 h (lower). Arrows indicate the presence of multiple septa in the images. Scale bars, 10 μm. (C) Graph indicating ratio distribution of multiseptum formation with respect to the total number of cells in BAS WT and BAS ΔprkC strains at 3 h and 12 h. Around 1,500 cells were considered for the calculation of each bar in the graph. (D) Comparative gene expression analysis of ftsZ in the BAS ΔprkC strain compared to the BAS WT strain during mid-log and stationary phases. The data were normalized to the expression of rpoB from each sample. Error bars represent standard deviations from three biological and three technical replicates. Statistical significance of ftsZ gene expression in BAS WT and BAS ΔprkC strains at both time points was analyzed using two-way ANOVA and denoted in the graph in the form of asterisks indicating significant results. ***, P < 0.001; ns, not significant.

FtsZ is a cytoskeletal protein of cell division machinery that localizes at mid-cell and forms the initial Z ring. It also serves as the scaffold for further assembly of cell division machinery (47). STPKs from other bacterial systems have been shown to phosphorylate and regulate the activity of FtsZ (48, 49). Our initial results suggest that ftsZ is constantly upregulated in the prkC disruption strain (Fig. 7D). This probably is the reason for an increase in multiseptum formation observed in the prkC disruption strain, possibly due to the mislocalization of FtsZ. Further experiments are under way to delineate the PrkC-mediated signaling cascades, the disruption of which results in the observed defects in cell division, cell wall homeostasis, and multiseptum formation.

DISCUSSION

Chaining phenotype acts as a virulence factor in several bacterial pathogens (1, 4, 50–55). In Legionella pneumophila, chaining morphology helps the pathogen evade phagosomal killing by interfering with phagosomal morphogenesis (52). In Streptococcus pneumoniae, bacteria growing as long chains display increased attachment and adherence to epithelial cell surfaces, possibly via multivalent binding sites (51). Bacillus cereus, a close relative of B. anthracis and a cause of foodborne and opportunistic infections in humans, also displays a chaining phenotype and has been shown to attach to the invertebrate gut through long filaments (54, 55). In B. anthracis, in a mouse model, chain length-dependent physical sequestration of an encapsulated nontoxinogenic strain in lung capillaries is thought to result in hypoxia and associated lung tissue injury, leading to host death (4, 5).

In this study on the B. anthracis Sterne strain, we identify a transmembrane serine/threonine protein kinase, PrkC, with an extracellular sensory PASTA domain as a determinant of the chaining phenotype. Interestingly, PrkC homologs are found in all the above-mentioned chain-forming pathogens (see Fig. S1 in the supplemental material). Although PASTA kinases do not necessarily carry out similar signaling processes across various bacterial species (56, 57), it would be worthwhile to explore whether PASTA kinases of these pathogens also form the primary messenger molecule for maintaining the chaining phenotype, as reported for B. anthracis in this study. Notably, PASTA motifs are unique to bacteria, and their absence from eukaryotes makes them an attractive drug target (56).

In this study, we show that in the prkC disruption strain, the S-layer protein Sap and septal N-acetylglucosaminidase, BslO, are upregulated. On the contrary, stationary-phase S-layer protein EA1 shows downregulation (Fig. 4). S-layers are found in many bacterial species, where they form a cell cover and play important roles (58, 59). In B. anthracis, the S-layer is made up of two primary structural proteins, Sap and EA1, and several S-layer-associated proteins, called BSLs, that carry out diverse roles (60). Previous studies have shown that Sap levels rise until the onset of the stationary phase, and the EA1 amount is minimal during the logarithmic phase (46). As Sap levels go down, EA1 is upregulated and replaces Sap as the primary constituent of the S-layer in the stationary phase. Both Sap and EA1 act as the transcriptional repressors of the eag gene encoding EA1 (46). Our results also show a gradual decline in the Sap levels from late log phase onwards (OD [A₆₀₀] of ∼5.0 to 6.0; time, 6 to 8 h) (Fig. 2A and 4A). EA1 levels, as reported earlier, were minimal during lag phase and early exponential phase but increased gradually until the last point of measurement (Fig. 4C). Interestingly, BslO levels remained constant throughout the growth cycle in BAS WT, which implies that the stage/growth phase-dependent dechaining in the wild-type strain is primarily dependent on its localization, which in turn is controlled by the levels of Sap on the cell surface. Upregulation of both Sap and BslO in the BAS ΔprkC strain would create a condition that would favor dechaining from the initial stages of bacterial growth. Additionally, we observed reversion of the BAS WT chaining phenotype (Fig. 1B and C and Fig. S2) and the Sap and BslO expression levels (Fig. S5) on the introduction of an ectopic copy of prkC under its native promoter in the BAS ΔprkC strain. Altogether, our results suggest that PrkC keeps a check on the levels of Sap, BslO, and EA1 during optimum growth conditions, thereby maintaining the chaining phenotype.

B. anthracis PrkC is a key messenger molecule that plays a central role in various cellular processes, including infection in macrophages, biofilm formation, and germination (33, 34, 36, 42). We report, for the first time, that prkC disruption results in the attenuation of growth rate during the logarithmic phase of growth (Fig. 5A). Apart from this, morphological defects, such as decreases in cell wall width and multiseptum formation, were also observed (Fig. 5B and C and 7A to C). The decrease in cell wall thickness and shorter bacterial chains in the ΔprkC strain could be the reasons for the observed higher sensitivity of the ΔprkC strain to cell wall-targeting antibiotics (Fig. 6). In a study on PknB, a membrane-localized PASTA kinase from Mycobacterium tuberculosis, depletion or overexpression of the kinase was shown to have a significant effect on bacterial morphology, leading to cell death (61). Therefore, PrkC, like some other PASTA domain-containing kinases, seems to be important for the maintenance of cell morphology during exponential growth.

In conclusion, through this study, we show that PrkC, the transmembrane kinase of B. anthracis with a sensor PASTA domain, regulates cell growth, morphology, and, most importantly, the chaining phenotype. Since the disruption strain of PrkC shows decreased virulence in a mouse model of pulmonary anthrax (33), it will be relevant to see if the observed effect is due to a difference in the chaining phenotype, as shown in this report. If so, therapeutic intervention against PrkC could help control bacterial chain size and, hence, lung tissue injury and its pathophysiological consequences.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

Escherichia coli DH5α (Invitrogen) and SCS110 (Stratagene) strains were used for cloning, and the BL21(DE3) (Invitrogen) strain was used for expression of recombinant proteins. The final antibiotic concentrations used were 100 μg/ml ampicillin, 25 μg/ml kanamycin, and 150 μg/ml spectinomycin. LB broth (Difco) with appropriate antibiotic was used to grow bacterial cultures at 37°C with proper aeration (1:5 headspace) and constant shaking at 200 rpm. B. anthracis Sterne strain 34F2 (BAS WT) was obtained from Colorado Serum Company, and the prkC knockout strain (BAS ΔprkC) was a gift from Jonathan Dworkin, Department of Microbiology, Columbia University. Descriptions of plasmids and strains used in this study are provided in Tables 1 and 2, respectively. The sequences of primers used in this study are provided in Table 3.

TABLE 1.

List of plasmids used in this study

Name	Description	Resistance marker(s)	Reference or source
pYS5	Used for complementation in B. anthracis; Ampr in E. coli; Kanr in B. anthracis	Kanamycin, ampicillin	68
pYS5-prkC*	Expressing PrkC under own promoter in B. anthracis	Ampicillin, spectinomycin	This study
pPROEX-HTc	E. coli expression vector with N-terminal His6 tag	Ampicillin	Invitrogen
pET28a	E. coli expression vector with N- and C-terminal His6 tags	Kanamycin	Invitrogen
pPROEX-HTc-sap	Expression of His6-Sap in E. coli	Ampicillin	This study
pPROEX-HTc-eag	Expression of His6-EA1 in E. coli	Ampicillin	This study
pPROEX-HTc-groEL	Expression of His6-GroEL in E. coli	Ampicillin	This study
pET28a-bslO	Expression of His6-BslO in E. coli	Kanamycin	This study
pYS5-bslO	Expression of BslO under PA constitutive promoter	Kanamycin	This study

TABLE 2.

Bacterial strains used in this study

Name	Genotype	Resistance marker(s)	Reference or source
E. coli
DH5α	F− endA1 glnV44 thi-1 recA1 relA1 gyrA96 deoR nupG purB20 ϕ80dlacZΔM15 Δ(lacZYA-argF)U169 hsdR17(rK− mK+) λ−	Invitrogen
BL21(DE3)	B strain; F− ompT gal dcm lon hsdSB(rB− mB−) λ(DE3 [lacI lacUV5-T7p07 ind1 sam7 nin5]) [malB+]K-12 (λS)	Invitrogen
SCS110	endA mutant derivative of the JM110 strain; rpsL (Strr) thr leu endA thi-1 lacY galK galT ara tonA tsx dam dcm supE44 Δ(lac-proAB) [F′ traD36 proAB lacIqZΔM15]	Stratagene
B. anthracis Sterne 34F2
BAS WT	pXO1+, pXO2−	NIAID, NIH
BAS ΔprkC	ΔprkC	Kanamycin	42
BAS ΔprkC::prkC	ΔprkC::prkC	Kanamycin, spectinomycin	This study
BAS WT::bslO	Sterne::bslO	Kanamycin	This study

TABLE 3.

List of primers used in this study

Category and no.	Primer name	Primer sequence 5′→ 3′a
Real-time PCR
P1	BAS 3757_ftsZ RT FP	TGCCTCTAACATTGGCGTGT
P2	BAS 3757_ftsZ RT RP	CAAGCGGCATCTGGTATTGC
P3	BAS0102 _rpoB_RT_FP	AACTTGCGCACATGGTTGAC
P4	BAS0102 _rpoB_RT_RP	CTGTCCACCGAACTGAGCTT
Protein purification
P5	BAS0841 Fp (BamHI)	GGGGGATCCCCATGGCAAAGACTAACTC
P6	BAS0841 Rp (XhoI)	GGGCTCGAGATTATTTTGTTGCAGGTTTTGC
P7	BAS0842 Fp (BamHI)	CCCGGATCCCCATGGCAAAGACTAACTCTTAC
P8	BAS0842 Rp (XhoI)	GGGCTCGAGTTATAGATTTGGGTTATTAAG
P9	BAS0253_groEL_BamHI_FP	AATCCAAGGGGGTGGATCCTTATGGCAAAAG
P10	BAS0253_groEL_XhoI_RP	TTAGGGCAAACTCGAGTTACATCATTCCGCCC
P11	BAS1683 FP (EcoRI)	CCCGAATTCATGAAAAAAGTTATTTCTAATGTG
P12	BAS1683 RP (XhoI)	CCCCTCGAGTTGTATTTTTAAGTTCTTCTTCAATGTCC
Complement strain generation
P13	prkC SpeI Fp	CCACTAGTCGTGCTGATTGGAAAACGCTTAAATG
P14	prkC BamHI Rp	CCGGATCCTTATTGTGTTGGATATGGTACTTCTTTG
P15	prkC Promoter KpnI Fp	CCGGTACCATTGTCGGTCGTGGTACAGAAACTG
P16	prkC promoter SpeI Rp	CCACTAGTATGGCTCGTCCTCTTTCTTTTTC
bslO overexpression strain (BAS-WT::bslO) generation
P17	bslO O/E SpeI Fp	GCGGACTAGTCATGAAAAAAGTTATTTCTAATGTG
P18	bslO O/E BamHI Rp	CCCGGATCCTTATATTTTTAAGTTCTTCTTCAATGTCC

^aUnderlined nucleotides signify restriction sites.

Generation of prkC complement and bslO overexpression strains.

The prkC gene and its promoter sequence were amplified using primers P13 to P16. These genes were subsequently cloned in the shuttle vector pYS5 (62), and the positive clone was transformed in SCS110 cells prior to electroporation in the BAS ΔprkC strain using a Bio-Rad Gene Pulser Xcell (2.5 kV, 400Ω, 25 μF, using 0.2-cm Bio-Rad Gene Pulser cuvette). The complemented strain obtained was named the BAS ΔprkC::prkC strain. A similar protocol was used for the generation of the bslO overexpression strain (BAS WT::bslO) under the PA promoter in the pYS5 shuttle vector (using primers P17 and P18).

Cloning, gene expression, and protein purification.

Genes for sap, eag, bslO, and groEL were amplified using BAS genomic DNA as the template and sequence-specific primers (P5 and P6, sap; P7 and P8, eag; P9 and P10, groEL; P11 and P12, bslO). The amplified products obtained for sap, eag, and groEL were cloned in the pPROEX-HTc vector (Invitrogen), while the bslO gene was cloned in the pET28a vector (Invitrogen). The resulting plasmids encode His₆-tagged fusion proteins. Plasmids were transformed into E. coli BL21(DE3), and proteins were purified using affinity chromatography as described previously (63). Briefly, overnight-grown cultures were diluted in LB broth (1:50) with appropriate antibiotic, ampicillin (100 μg/ml) or kanamycin (25 μg/ml), and grown at 37°C and 200 rpm. Cultures were induced with 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) at an OD (A₆₀₀) of 0.5 to 0.8 and incubated overnight at 16°C. After this, cells were pelleted, resuspended in sonication buffer (50 mM Tris-HCl [pH 8.5], 5 mM β-mercaptoethanol, 1 mM phenylmethylsulfonyl fluoride [PMSF], 1× protease inhibitor cocktail [Roche Applied Science], and 300 mM NaCl), and sonicated (9 cycles of 20% amplitude, 10 s on and 30 s off). The recombinant proteins were purified by affinity purification using a Ni-nitrilotriacetic acid (NTA) column, and the final elution was done using 200 mM imidazole. Protein estimation was done using a Pierce bicinchoninic acid (BCA) protein assay kit (Thermo Fisher Scientific).

Generation of polyclonal antibodies against GroEL and BslO in mice and Sap and EA1 in rabbits.

For the generation of polyclonal antibodies, 30 μg of purified protein (GroEL and BslO) was used for injecting into three BALB/c mice, and 500 μg of purified protein (Sap and EA1) was used for injecting into three rabbits for each protein. Antigens were emulsified in complete Freund’s adjuvant (Sigma-Aldrich) at a 1:1 ratio before subcutaneous injection in the animals. The production of antibody was stimulated at an interval of 21 days, followed by a booster injection of 15 μg protein emulsified in incomplete Freund’s adjuvant for mice and three booster injections of 250 μg protein emulsified in incomplete Freund’s adjuvant for rabbits. Animals were bled to collect serum 14 days after the final injection, and the antibody titer was calculated by enzyme-linked immunosorbent assay (ELISA) and used accordingly for further experiments.

Growth kinetics.

BAS WT and BAS ΔprkC strains were grown overnight in LB broth at 37°C, 200 rpm. These overnight-grown cultures were taken as the inoculum for growth kinetics experiments. The secondary cultures were initiated at a starting OD (A₆₀₀) of 0.001 in triplicate in LB broth at 37°C, 200 rpm (New Brunswick Innova 42 incubator shaker). Following this, the OD (A₆₀₀) was monitored until around 64 h at the indicated intervals (Fig. 5A).

For lysate preparation and microscopy, the secondary cultures for both strains were initiated similarly in triplicates from overnight-grown cultures at a starting OD (A₆₀₀) of 0.001. Culture samples were collected at different time points (2 h, 3 h, 4 h, 5 h, 6 h, 7 h, 8 h, 9 h, 10 h, 14 h, 18 h, 22 h, 26 h, and 30 h) for analysis, and the corresponding OD (A₆₀₀) was plotted (Fig. 2A and B).

Phase-contrast microscopy.

For phase-contrast microscopy, 1-ml culture samples from growing cells of BAS WT and BAS ΔprkC strains were collected at different time points as mentioned above. The cells were pelleted and washed thrice with phosphate buffer (pH 7.4) and resuspended in 100 μl buffer. The cells were then observed under a 100×/1.4 oil differential interference contrast (DIC) objective of a Zeiss Axio Imager Z2 upright microscope. Images were captured using an AxioCam 506 color camera attached to the microscope and processed in ZEN 2 Pro software.

Quantitative immunoblot analysis.

For immunoblot analysis, 5- to 10-ml bacillus culture samples were collected at different time points (2 h, 3 h, 4 h, 5 h, 6 h, 7 h, 8 h, 9 h, 10 h, 14 h, 18 h, 22 h, 26 h, and 30 h) from BAS WT and BAS ΔprkC cultures. The cell pellet obtained was washed with phosphate-buffered saline (PBS) and resuspended in 1 ml lysis buffer (200 mM Tris-HCl [pH 7.5], 1 mM EDTA, 150 mM NaCl, 1 mM PMSF, 5% glycerol, and 1× protease inhibitor cocktail [Roche Applied Science]). After this, sonication of the bacterial pellets resuspended in the buffer was done (9 cycles of 20% amplitude, 10 s on and 30 s off). Protein estimation was done using Pierce BCA protein assay kit (Thermo Fisher Scientific). Five-microgram protein lysate samples from each time point were prepared using SDS sample buffer containing 250 mM Tris-HCl (pH 6.8), 30% (vol/vol) glycerol, 10% SDS, 10 mM dithiothreitol, and 0.05% (wt/vol) bromophenol blue. Samples were heated for 5 min at 95°C prior to loading on 12% polyacrylamide gels, followed by transfer onto an NC membrane (Millipore). Bovine serum albumin (3%) in PBS with 0.05% Tween 20 (PBST) was used for blocking the membranes overnight at 4°C. This was followed by washing with PBST (3 washes of 5 min each). Membranes were then probed with antibodies specific to Sap protein (1:50,000), BslO protein (1:10,000), or EA1 protein (1:50,000) for 1 h, followed by washes with PBST (5 washes of 5 min each). After this, anti-rabbit IgG secondary antibody (for Sap and EA1) or anti-mouse IgG secondary antibody (for BslO) conjugated with horseradish peroxide (1:10,000; Cell Signaling Technology) was used and the blots were incubated for another 60 min, followed by 3 PBST washings of 10 min each. Finally, SuperSignal West Pico Plus chemiluminescent substrate (Thermo Fisher Scientific) was used to detect the signal, and it was visualized and quantified with the luminescent image analyzer (Amersham Imager 600 or ImageLab6.0.1). The blots were then stripped using stripping buffer and probed similarly using anti-GroEL (1:50,000) and anti-mouse IgG secondary antibody conjugated with horseradish peroxide (1:10,000; Cell Signaling Technology) for normalizing the loading pattern.

RNA extraction and quantitative real-time PCR.

BAS WT and BAS ΔprkC strains were grown to mid-log and stationary phases in triplicates for RNA extraction by following the hot lysis method as described previously (64–66), with a few modifications. Cells were harvested at 6,000 × g for 15 min, and the cell pellet was washed once with PBS, resuspended in 500 μl TRIzol (Invitrogen), and frozen at −80°C until ready for further processing. The frozen samples were thawed on ice, and RNA was extracted by the hot lysis method. Briefly, the samples were mixed with 400 μl of buffer (50 mM Tris [pH 8.0], 1% SDS, and 1 mM EDTA) and 400 μl of zirconia beads treated with diethyl pyrocarbonate water. This suspension was incubated at 65°C for 15 min with rigorous intermittent vortexing after every 5 min. The suspension was cooled in ice and mixed well with 100 μl chloroform/ml of TRIzol. Separation of the aqueous phase containing RNA was done by centrifugation at 9,500 × g for 15 min at 4°C. RNA was precipitated from the aqueous phase by adding LiCl₂ (0.5 M) and 3× ice-cold isopropanol followed by 2 h of incubation at −80°C. The RNA pellet obtained by centrifugation at 16,000 × g for 20 min (4°C) was washed using 70% ethanol (Merck) and resuspended in nuclease-free water after air drying. RNA sample was then treated with DNase (Ambion) to remove any residual DNA contamination (according to the manufacturer’s protocol). An RNeasy minikit (Qiagen) was used to obtain pure RNA by using the manufacturer’s protocol. cDNA was prepared using 1 μg of RNA with a first-strand cDNA synthesis kit (Thermo Fisher) according to the protocol provided by the manufacturer. To analyze the expression of the ftsZ gene, 2 μl of cDNA (diluted 10 times) was used for each time point (mid-log and stationary phases) along with gene-specific primer and SYBR green master mix (Roche) in a 10-μl reaction mixture according to the manufacturer's protocol. Reactions were run in triplicates along with a no template control in a LightCycler 480 Instrument II (Roche). The rpoB gene encoding DNA-directed RNA polymerase subunit beta was used as a housekeeping control (67). All of the primers used were sequence specific with a PCR product of 120 bp.

Scanning electron microscopy.

BAS WT, BAS ΔprkC, and BAS ΔprkC::prkC strains were grown in LB broth at 37°C, harvested at mid-log phase, and processed (68). Briefly, the bacterial culture was harvested at 12,000 × g at 4°C, and the pellet obtained was washed thrice using 0.1 M sodium phosphate buffer (pH 7.4). Karnovsky’s fixative (2.5% glutaraldehyde [TAAB] plus 2% paraformaldehyde [Sigma] in 0.1 M sodium phosphate buffer, pH 7.4) was used to fix the bacterial samples overnight at 4°C. Fixed cells were again washed using sodium phosphate buffer, and this step was repeated thrice to remove any residual fixative from the pellet. After this, the pellets were again fixed using 1% osmium tetroxide for 20 min at 4°C. Sequential dehydration was then done for 30-min steps using a range of ethanol (Merck) concentrations (30%, 50%, 70%, 80% [2×], 90%, and 100% [3×]) at 4°C. A critical-point drying technique was used for drying the samples followed by gold coating of 10 nm using an aluminum stub coated with agar sputter. Images were captured using a Zeiss scanning electron microscope EVO LS15 at 20 kV. Comprehensive imaging, processing, and analysis were performed with Smart SEM software (69).

Transmission electron microscopy.

BAS WT and BAS ΔprkC strains were grown in LB broth at 37°C, harvested at mid-log and stationary phases, and processed. Briefly, the bacterial culture was harvested at 12,000 × g at 4°C, and the pellet obtained was washed thrice using 0.1 M sodium phosphate buffer (pH 7.4). Karnovsky’s fixative containing 2.5% glutaraldehyde (TAAB) and 2% paraformaldehyde (Sigma) was made in 0.1 M sodium phosphate buffer, pH 7.4, and was used for primary fixation of the cells overnight at 4°C. Fixed cells were again washed using sodium phosphate buffer, and this step was repeated thrice to remove any residual fixative from the pellet. After this, the pellets were again fixed using 1% osmium tetroxide for 20 min at 4°C. Sequential dehydration was then done for 30-min steps using a range of acetone (Merck) concentrations (30%, 50%, 70%, 80% [2×], 90%, and 100% [3×]) at 4°C. For the clearing process and removal of dehydrating agent, absolute xylene (Merck) was used, and the samples were subjected to bullet preparation using an araldite resin mixture (TAAB). Following this, infiltration was done by raising the concentration of the embedding medium and lowering the concentration of clearing agents gradually. The final bullets were prepared by curing at 55°C for 24 h and for 48 h at 65°C. Sections were obtained using a Leica UC6 ultracut to make the grids, which were then observed in an FEI Tecnai G² Spirit at 200 KV (69).

Confocal microscopy to analyze multiseptum formation.

FM4-64 labeling was used to visualize multiseptum formation in BAS WT and BAS ΔprkC strains using fluorescence. LB-agarose pads were prepared using AB gene frame (17 by 54 mm; Fischer Scientific) on frosted glass slides (Corning micro slide, frosted; 75 by 25 mm). To prepare the agarose pad, 3% low-melting-point agarose (Sigma) was poured on these slides and left for solidification until further use. One-microliter samples of exponentially growing cultures of BAS WT and BAS ΔprkC strains diluted to an OD (A₆₀₀) of 0.035 were spread evenly on the agarose pads along with 1 μg/ml FM4-64 dye for staining the cell membrane. Images were captured using a Leica TCS SP8 confocal laser scanning microscope at 3 h and 12 h using a 63× oil immersion objective (70).

Antibiotic-sensitive assay.

Two-hundred-microliter samples of BAS WT, BAS ΔprkC, and BAS ΔprkC::prkC cultures were plated on an LB agar plate at a starting OD (A₆₀₀) of 0.5. Antimicrobial susceptibility disks (Thermo Scientific) were then placed carefully onto the agar plates and kept for incubation at 37°C overnight. Growth inhibition diameter was measured for each disc and was plotted using GraphPad software (Prism 9).

Phylogenetic analysis.

The amino acid sequences of PrkC protein in different pathogens, Bacillus anthracis Sterne, Bacillus cereus, Legionella pneumophila, and Streptococcus pneumoniae, were procured from NCBI. These sequences were aligned using the T-coffee tool (71). These protein sequences were then used for the generation of a phylogenetic tree by neighbor-joining analysis conducted by MEGA X (72–74). The representation of branch lengths is in units of evolutionary distances computed by the Poisson correction method (75).

Statistical analysis.

GraphPad software (Prism 9) was used for all the statistical analyses. The statistical tests are indicated in the figure legends, and the corresponding two-tailed t test or analysis of variance (ANOVA) P values are reported in the graphs wherever required (P < 0.05, P < 0.01, P < 0.001, and P < 0.0001 were considered significant results). Values indicated in the graphs represent means ± SD, where n = 3 for both strains at each time point, unless specified otherwise in the figure legend. Error bars are indicative of SD (n = 3). All experiments were done in biological triplicates to ensure the reproducibility of the obtained data. Real-time experiments (Fig. 7D) were done in biological and technical triplicates, while growth kinetics (Fig. 2A and B and 5A), Western blot analysis (Fig. 4), and antibiotic sensitivity assay (Fig. 6) were done using biological triplicates. Separate flasks of the same strains are considered biological replicates, while technical replicates refer to multiple readings of the same sample.