The Ser/Thr protein kinase PrkC imprints phenotypic memory in Bacillus anthracis spores by phosphorylating the glycolytic enzyme enolase

Bacillus anthracis is the causative agent of anthrax in humans, bovine, and other animals. B. anthracis pathogenesis requires differentiation of dormant spores into vegetative cells. The spores inherit cellular components as phenotypic memory from the parent cell, and this memory plays a critical role in facilitating the spores' revival. Because metabolism initiates at the beginning of spore germination, here we metabolically reprogrammed B. anthracis cells to understand the role of glycolytic enzymes in this process. We show that increased expression of enolase (Eno) in the sporulating mother cell decreases germination efficiency. Eno is phosphorylated by the conserved Ser/Thr protein kinase PrkC which decreases the catalytic activity of Eno. We found that phosphorylation also regulates Eno expression and localization, thereby controlling the overall spore germination process. Using MS analysis, we identified the sites of phosphorylation in Eno, and substitution(s) of selected phosphorylation sites helped establish the functional correlation between phosphorylation and Eno activity. We propose that PrkC-mediated regulation of Eno may help sporulating B. anthracis cells in adapting to nutrient deprivation. In summary, to the best of our knowledge, our study provides the first evidence that in sporulating B. anthracis, PrkC imprints phenotypic memory that facilitates the germination process.

Deciphering the role of metabolic enzymes that orchestrate morphogenic transition states in bacteria is a fundamental question. Bacillus anthracis is the causative agent of anthrax in humans, bovine, and other animals (1,2). It is known to survive hostile environments by forming spores that remain quiescent and retain minimum metabolic activity (3,4). As a pathogen, the success of B. anthracis depends on the spore's ability to develop into a growing vegetative cell. Under favorable conditions, spore metabolism is triggered to support energy needs and to develop into a fully functional cell (5). The metabolic checkpoints and energy reserves in the spore provide different stimuli at an early growth stage and ensure the completion of the developmental program. Therefore, the transformation of a dormant spore into a vegetative cell is a key step in the pathogenic cycle of B. anthracis. However, the molecular events leading to successful spore dormancy and later to germination remain to be fully elucidated.
During the process of germination, some spores disintegrate their protective structure and grow into vegetative cells (5). Sporulating cells carry a substantial set of macromolecules (including several proteins) through their journey from progenitor cell to spore (6). The role of these proteins remains largely unknown. This carryover of cellular components is termed "phenotypic memory" (7). The efficiency of spore germination has been shown to be determined by this phenotypic memory inherited from the parent cell (7,8). In a recent study, the role of one such protein, alanine dehydrogenase, which controls alanine-induced outgrowth in cellular memory, was described (7). Because the protein cargo remains constant in the spore, these proteins decide the fate of the reviving spore depending on environmental conditions and sensory signaling molecules (9). There have been considerable efforts toward understanding the spore germination process and involvement of signaling mechanisms (10 -12).
The link between metabolism and the spore revival process is not explored well. At the onset of germination, metabolism resumes without requiring new macromolecular synthesis (13). Reports in Bacillus subtilis have highlighted the essentiality and  Table S1. 1 To whom correspondence may be addressed: Dept. of Zoology, University of Delhi, Delhi 110007, India. E-mail: arorag1983@gmail.com. 2 To whom correspondence may be addressed: Dept. of Zoology, University of Delhi, Delhi 110007, India. E-mail: ysinghdu@gmail.com.
interactions of glycolytic enzymes phosphofructokinase (Pfk), 3 phosphoglyceromutase (Pgm), and enolase (Eno) (14). Eno and Pgm play an essential role in both glycolysis and gluconeogenesis where Pgm reversibly converts 3-phosphoglyceric acid (3-PGA) to 2-PGA, and Eno catalyzes the penultimate step of glycolysis by conversion of 2-PGA to phosphoenolpyruvate (PEP), thus deciding the flux of pathway. Bacteria survive harsh conditions efficiently by keeping an alternate source of energy, 3-PGA, which is used during the early events of spore germination (15). A balanced ratio of 3-PGA and 2-PGA is maintained at the time of spore formation by keeping the spore metabolically dormant. Furthermore, the dehydrated, acidic core of spore diminishes the metabolic activity to maintain the 3-PGA reserve (16). Because spores of Bacillus sp. are known to hold significant levels of 3-PGA reserves, we decided to investigate the role of Eno in the spore germination process of B. anthracis. Previous studies have shown that an Eno deletion-mutant strain was defective in sporulation, whereas its inhibition led to a lag in ATP production during germination, thus uncovering a role for Eno in the bacterial life cycle (17,18). Besides this, Eno is a component of the adsorbed anthrax vaccine and has been indicated in helping bacteria evade the innate immune cells by binding to host plasminogen (19,20).
In this study, we address the role of Eno in spore germination. Our results show that Eno acts as an intrinsic memory controller that influences the germination process and is regulated by the Ser/Thr protein kinase PrkC.

B. anthracis Eno influences spore germination
To decipher the contribution of some key metabolic enzymes in spore germination, we cloned and expressed the glycolytic pathway genes pgk, pgm, and eno that are important for maintaining the 3-PGA reserve. The B. anthracis Sterne (Bas) strains overexpressing Pgk, Pgm, and Eno were analyzed for their sporulation and germination efficiencies. As shown in Fig. 1A, an increase in the expression of either glycolytic protein led to a decrease in sporulation efficiency. For germination efficiency, there was an ϳ75% decrease in Eno-overexpressing spores, whereas spores overexpressing Pgk and Pgm showed only about 20 and 40% decreases, respectively (Fig. 1B). To assess the levels of Eno in the overexpression strain, we generated Enospecific polyclonal antibodies in mice (Fig. S1). Immunoblotting with these antibodies showed that the expression of Eno was increased by ϳ1.5-fold in the recombinant strain as compared with parent strain (Fig. S2).

Eno expression is reduced in B. anthracis spores
Our results suggest that overexpression of Eno causes a decrease in spore germination. Therefore, we decided to check the intrinsic regulation of Eno expression in spores as well as in vegetative cells. Using Eno-specific polyclonal antibodies, we determined the expression of Eno in whole-cell lysates at different stages of the B. anthracis lifecycle. Immunoblotting with anti-Eno antibody detected a specific band at ϳ45 kDa corresponding to the molecular mass of Eno. After quantification of band intensities, differential expression of Eno was observed in several growth stages relative to early log phase where the maximum expression was observed (Fig. 2). The expression of Eno consistently decreased in the later growth stages (log phase, late log phase, and stationary phase) until in a sporeforming stage where only 30% of the protein remained with respect to early log phase. Because spores have lower levels of Eno as compared with vegetative cells and overexpression of Eno leads to reduced fitness of the Bas strain during spore germination, there seems to be a decisive role of Eno in germination.

Eno is phosphorylated in vitro by the B. anthracis Ser/Thr protein kinase PrkC
Signaling mechanisms regulate the transition of B. anthracis from dormancy to vegetative state (21,22). Interestingly, there is a growing body of evidence supporting the notion that PrkC could play an important role in the spore's exit from dormancy (12,23,24). In our previous studies, we found that glycolytic enzymes are subjected to regulation by phosphorylation (25,26). Large-scale phosphoproteome analysis in B. subtilis also were used for spore formation followed by the addition of nutrient-rich medium for germination. Sporulation (A) and germination (B) efficiencies were calculated considering the efficiency of Bas-wt as 100% and plotted using GraphPad Prism. Error bars represent S.D. of three independent experiments. *, p Յ 0.05; **, p Յ 0.01 as determined by two-tailed unpaired Student's t test.

Phenotypic memory in B. anthracis
indicates phosphorylation of Eno, which is a close homolog of B. anthracis Eno (ϳ80% sequence similarity) (27). Additionally, in our previous study, comparison of phosphoproteomic analyses of B. anthracis WT (Bas-wt) and prkC deletion mutant (Bas⌬prkC) identified phosphorylated isoforms of Eno (28). Therefore, we hypothesized that Eno could be regulated by PrkC-mediated phosphorylation, and this regulation might be important for B. anthracis morphogenesis.
To address this hypothesis, in vitro kinase assays were performed using recombinant PrkC and Eno with [␥-32 P]ATP. As shown in Fig. 3A, Eno was found to be phosphorylated by PrkC. No phosphotransfer was observed when Eno was incubated alone or with the Ser/Thr phosphatase PrpC. Phosphotransfer analysis using in vitro kinase assay and two-dimensional gel electrophoresis indicated phosphorylation of Eno at early time points (Fig. 3, B-D). Thus, these assays confirmed that Eno is a substrate of PrkC. Subsequently, the phosphorylation of Eno was also confirmed in vivo using coexpression with PrkC and PrpC in Escherichia coli. Eno was found to be phosphorylated when coexpressed with PrkC (phosphorylated Eno (Eno-P)), whereas no phosphorylation was observed when coexpressed with PrpC (unphosphorylated Eno (Eno-UP)) as confirmed by Pro-Q Diamond phosphorylation-specific staining (Fig. 3E). Eno-P isolated from cells metabolically labeled with [ 32 P]orthophosphoric acid showed a 32 P-labeled Eno demonstrating PrkC-specific phosphorylation under native conditions in E. coli (Fig. 3F).

Eno is phosphorylated on serine and threonine residues
To identify the residues in Eno that are phosphorylated by PrkC, purified Eno-P and Eno-UP proteins were subjected to MS (Fig. S3). We found nine phosphorylated serine and threonine residues in Eno-P, whereas no phosphorylated residues were identified in Eno-UP. To identify the location of phosphorylation sites in Eno, we generated a homology model of B. anthracis Eno using the crystal structure of B. subtilis Eno (Protein Data Bank (PDB) code 4A3R) (29). Fig. 4A shows the Bas Eno homology model with the phosphorylation sites marked. Three phosphorylated residues (Ser 336 , Thr 363 , and Ser 367 ) were localized in the C-terminal hydrophobic region, whereas the remaining six residues were present at the protein surface (Fig. 4A). Among these sites, Ser 367 was highly conserved among the Eno homologs and was found to be present in the flexible loop responsible for catalysis, as identified by multiple sequence alignment of Eno and its homologs (Fig. S4). Previously, Eno Ser 367 was also identified as a phosphorylated residue in the spore phosphoproteome of B. subtilis (27). Further structural analysis of the three C-terminal residues indicated that Ser 367 could interact with the catalytic site residues Lys 340 and Ile 339 (Fig. 4B), whereas Thr 363 and Ser 336 were present on the parallel and antiparallel strands, stabilizing the structure of the protein (Fig. 4C).
To determine the role of individual phosphorylation sites, we chose to study the three C-terminal residues located in the hydrophobic pocket. These residues were mutated to generate single mutants Eno T363A , Eno S336A , and Eno S367A ; a double mutant, Eno S336A/T363A ; and a triple mutant, Eno S336A/T363A/S367A . Equal amounts of Eno and its mutant derivatives were used in an in vitro kinase assay with PrkC. Quantification of phosphorylation levels indicated a significant decrease in signal for all the single Eno mutants, whereas the double mutant Eno S336A/T363A showed a Ͼ60% loss in phosphorylation levels ( Fig. 4D). In the triple mutant (Eno S336A/T363A/S367A ), there was a significant loss (Ͼ80%) in phosphorylation ( Fig. 4E), showing that Ser 336 , Thr 363 , and Ser 367 were important phosphorylation sites.
The role of these three residues in regulating the enzyme activity was also analyzed. In glycolysis, Eno catalyzes the penultimate step by converting 2-PGA to PEP. We evaluated the phosphorylation-mediated variations on the activity of Eno by using a kit-based colorimetric assay that measures the formation of an intermediate product during this conversion. As shown in Fig. 4F, a considerable loss (40 -50%) in the activity of Eno S336A and Eno S367A mutants was observed, whereas Eno T363A exhibited an increase in activity. Furthermore, the triple mutant showed a ϳ40% decrease in the overall protein activity as compared with Eno-wt (Fig. 4G).
To understand the effect of mutations on the structural integrity of Eno, we performed circular dichroism (CD) spectroscopy using Eno purified from E. coli (Eno-wt), Eno-P (Eno coexpressed with PrkC), Eno-UP (Eno coexpressed with PrpC), and the triple mutant Eno S336A/T363A/S367A . To our surprise, all four samples showed a similar ␣-helix and ␤-sheet composition, suggesting that the mutations as well as phosphorylation  The intensity of phosphorylation of protein bands was calculated using QuantityOne software. Eno phosphorylation after 30 min was taken as 100% (signal saturation), and relative phosphorylation was calculated. The error bars show S.D. of three independent experiments. C and D, time-dependent phosphorylation of Eno by PrkC using cold ATP. Isoforms were resolved by 2D gel electrophoresis, and subsequent blots were probed with anti-Eno antibodies to determine the stoichiometry of Eno phosphorylation at different time points. Multiple species (encircled) were observed after 30 min of the reaction as compared with 0 and 10 min, and their relative intensity was calculated and plotted using GraphPad Prism. E and F, His-Eno purified from E. coli cells expressing Eno with PrkC (Eno-P) or PrpC (Eno-UP) were resolved by SDS-PAGE and stained with Pro-Q Diamond followed by analysis using a Typhoon imager (E). E. coli cells overexpressing Eno-P or Eno-UP were subjected to metabolic labeling using [ 32 P]orthophosphoric acid (F). In both panels, Eno-P was found to be phosphorylated, whereas no phosphorylation was observed on Eno-UP.

Phenotypic memory in B. anthracis
do not affect the secondary structure of the protein and maintain overall conformation (Table 1).
Previous studies on Eno structure from diverse species have shown that Ser 367 is a part of a flexible loop as mentioned earlier (30). This flexible loop is involved in binding to 2-PGA and subsequent conformational change. Additionally, we found this loop to be conserved in Eno from different species (Fig. S4). Therefore, we speculate that this flexibility of the loop and associated conformational change in the absence of substrate might allow the residue to become phosphorylated by the kinase (31).

Eno is phosphorylated in B. anthracis vegetative cells and spores
We next investigated the in vivo phosphorylation status of Eno. We overexpressed Eno with a C-terminal polyhistidine tag in Bas-wt (Bas Eno) and Bas⌬prkC (Bas⌬prkC Eno) Sterne strains. Purified proteins from respective strains were subjected to immunoblotting using anti-pSer and anti-pThr antibodies. Eno-P (coexpressed with PrkC) was used as a positive control, and E. coli Eno (not coexpressed with PrkC) was used as a negative control. pSer-and pThr-specific antibodies recognized Eno purified from Bas-wt, whereas the phosphorylation signal was significantly reduced, although not completely absent, in the Bas⌬prkC strain (Fig. 5, A and B). This result indicates that Eno is predominantly phosphorylated by PrkC in vivo but may also be phosphorylated by other kinases in the absence of PrkC.
Furthermore, we analyzed the stoichiometry of in vivo phosphorylation of Eno by resolving the phosphorylated and unphosphorylated isoforms on 2D gel electrophoresis. Whole-cell protein extracts from Bas-wt and Bas⌬prkC were subjected to electrophoresis followed by immunoblotting with anti-Eno antibody. Four Eno isoforms were observed in Bas⌬prkC cells (Fig. 5C). However, in Bas-wt cells, we identified seven Eno isoforms among which four were present at a pI similar to that in the Bas⌬prkC strain (pI 4.5-5.0), whereas the remaining three isoforms migrated at acidic pI range (toward 4.0; Fig. 5C), thus confirming PrkC-specific phosphorylation of Eno in vegetative cells.
Next, we investigated whether Eno was phosphorylated in spores. We overexpressed and purified Eno from B. anthracis spore lysate (Eno Bas Spore) followed by immunoblotting with anti-pThr antibody. Eno phosphorylation was detected in spores despite its low expression (Fig. 5D). The phosphorylation was also confirmed by using anti-pSer antibodies (Fig. S5).
These results show that Eno is regulated by PrkC-mediated phosphorylation in spores of B. anthracis, confirming its role in spore germination.

Phosphorylation decreases catalytic activity and Mg 2؉ cofactor affinity of Eno
After establishing that Eno is a substrate of PrkC, we investigated the effect of phosphorylation on Eno activity. In the activity assay, Eno-P was found to be ϳ60% less active than Eno-UP, suggesting that phosphorylation causes a decrease in the activity of Eno (Fig. 6A).
Interestingly, spores contain high concentrations of divalent cations such as Ca 2ϩ , Mg 2ϩ , and Mn 2ϩ to keep themselves dehydrated, thus maintaining a low metabolic profile (32, 33). As Eno is a metalloenzyme requiring Mg 2ϩ as a cofactor (34), we investigated whether phosphorylation had an impact on Eno-Mg 2ϩ interaction. Eno-P and Eno-UP (coexpressed with PrkC and PrpC, respectively) were used to study the phosphorylation-mediated regulation of the Eno-Mg 2ϩ interaction. A fluorimetry-based method was used to measure variations in the intrinsic tryptophan fluorescence of Eno upon addition of MgCl 2 (35). We observed that Eno-Mg 2ϩ complex formation was correlated with an enhancement of tryptophan fluorescence intensity. Our data showed that fluorescence of Eno-UP with Mg 2ϩ was higher than that of Eno-P, indicating a negative effect of phosphorylation on metal binding (Fig. 6B). Different concentrations of Mg 2ϩ were titrated with Eno-UP or Eno-P, and the binding constant (K a ) was calculated. For Eno-UP, the K a was higher (0.254 Ϯ 0.01 M) than the K a for Eno-P (0.181 Ϯ 0.03 M). Hence, these results show that Eno phosphorylation reduces Mg 2ϩ binding affinity, which is in agreement with the decreased activity of phosphorylated Eno.

PrkC is involved in the regulation of Eno expression in spores
We further analyzed the effect of PrkC on Eno expression in spores by comparing the protein levels of Eno in spores of Bas-wt and Bas⌬prkC Sterne strains (Fig. 7A) as well as during exponential growth (Fig. 7B). Our results showed 2-3-fold higher protein levels of Eno in Bas⌬prkC spores compared with Bas-wt spores, whereas the expression was similar in both strains during the exponential phase (Figs. 7 and S6). Thus, Eno protein levels might be regulated by PrkC. These results could in part explain the compromised virulence phenotype of the PrkC-null mutant as observed in earlier studies (36). However, it remains to be determined whether this PrkC-dependent regulation is direct or indirect. It could be hypothesized that PrkC phosphorylates a yet to be discovered transcription factor that regulates the expression of Eno. Such regulation could occur  (lower panel). F and G, relative activity of Eno mutants with respect to the native Eno (100%). Native Eno is named as WT, and mutants are named with respective mutation sites. All experiments were performed thrice, and error bars represent S.E. of three independent values. *, p Յ 0.05; **, p Յ 0.01; ***, p Յ 0.001 as determined by two-tailed unpaired Student's t test. Phenotypic memory in B. anthracis upon a signal received by PrkC at the onset of sporulation to trigger Eno down-regulation.

Eno phosphorylation status does not regulate its secretion but may affect its cellular localization
Surface-exposed signaling proteins like PrkC might be involved in maintaining the cellular dynamics during morphogenesis by modifying the activities and localization of metabolic enzymes to fulfill cellular requirements. Therefore, we examined whether the phosphorylation status of Eno could be linked to its cellular localization. To determine the localization of Eno within the vegetative cells, we probed the native protein in Bas-wt cells in exponential phase by anti-Eno antibodies using immuno-EM. Eno was found to be localized at the cell membrane and cytoplasm (Figs. 8, A and B, and S7). Subsequently, we compared the localization of Eno in Bas⌬prkC and prkCcomplemented strains at the exponential stage and noticed variability in Eno localization patterns among all three strains. Eno was found to be predominantly localized on the cell membrane in the Bas⌬prkC strain as compared with the Bas-wt and prkCcomplementedstrains (Fig.8B),indicatingthattheunphosphorylated isoform preferentially localizes at the membrane in vegetative cells.
In B. subtilis, Eno is known to be secreted via the nonclassical secretion pathway through a membrane-embedded hydropho-bic domain (37,38). We checked whether Eno secretion could also be impacted upon phosphorylation. So, culture supernatants and cell lysates of the Bas-wt and Bas⌬prkC cells grown to the late exponential phase were subjected to immunoblotting with anti-Eno antibodies. As a control, we used a very wellknown secretory protein from B. anthracis, the protective antigen (PA), which also forms part of the anthrax vaccine. The supernatant fraction was probed with the polyclonal antibody to PA to confirm that the supernatant contained the secreted proteins. Immunoblotting showed that B. anthracis Eno is secreted from the cell, but secretion was independent of its phosphorylation status (Fig. 8C). Thus, we conclude that PrkC affected the localization of Eno but not its secretion.

Conclusions
Development of spore to a vegetative cell requires a shift in metabolism. During spore formation, bacteria down-regulate their metabolism by halting transcriptional and translational machinery and storing energy reserves. At the beginning of germination, a spore requires substantial metabolic supplements. PrkC is known to sense germination cues from the environment and helps in regulating translation and metabolism during the spore germination. One of the primary questions in cellular development is how the metabolic machinery switches from dormancy during sporulation to exponential growth dur-

Phenotypic memory in B. anthracis
ing germination. Phenotypic memory, which is a carryover of cellular material from the mother cell to the spore, is key to germination. A critical step in this developmental program is the regulation of glycolytic enzymes that store energy reserves and balance metabolic needs when germination begins. We studied the expression of three glycolytic enzymes and identified Eno as a metabolic switch that holds the key to germination and contributes to phenotypic memory. Our results further identified PrkC to augment the germination process by maintaining Eno quantity, localization, and its enzymatic activity (Fig. 9).
In conclusion, our results provide evidence of Eno regulation by phosphorylation and its involvement in the process of germination of B. anthracis spores. Our study elucidates the phosphorylation of Eno and identifies its regulation by PrkC. The role of PrkC has been previously discussed in spore germination (12). Our results show that during sporulation PrkC initiates imprinting phenotypic memory by modifying the metabolic protein Eno. This imprinted memory helps B. anthracis to survive the nutritional shift and helps in spore germination.

Bacterial strains and growth conditions
E. coli strains DH5␣ (Novagen) and BL21-DE3 (Agilent) were used for cloning and expression of recombinant proteins, respectively. E. coli cells were grown and maintained with constant shaking (200 rpm) at 37°C in LB broth supplemented with 100 g/ml ampicillin or 25 g/ml chloramphenicol for the coexpression system. Bas-wt, PrkC deletion strain (Bas⌬prkC) and prkC-complemented strains (Bas⌬prkC complement) (12) were grown in LB medium supplemented with antibiotics as required. LB agar was used as the solid medium for culturing both E. coli and B. anthracis in the presence of selective antibiotics.

Cloning and mutagenesis of B. anthracis genes
Cloning and mutagenesis were performed according to the standard molecular biology procedures described earlier using B. anthracis Sterne strain genomic DNA (39,40). The coding sequences of eno (Bas4985), pgm (Bas4986), and pgk (Bas4988) from B. anthracis were amplified by PCR using primers containing SpeI and BamHI restriction sites with six-histidine repeat sequences in the reverse primer (Table S1). The resulting PCR product was cloned into E. coli/B. anthracis shuttle vector pYS5. Clones were then confirmed with restriction digestion and DNA sequencing (SciGenome). The confirmed plasmid was electroporated in Bas-wt or Bas⌬prkC strain using BTX Electro Cell Manipulator 600. For cloning and expression in E. coli, eno was amplified by PCR using primers containing BamHI and XhoI restriction sites. The resulting PCR product was cloned in pPro-Ex-HTc. To generate the site-specific mutants of Eno, site-directed mutagenesis was performed using the QuikChange XL Site-Directed Mutagenesis kit (Agilent).

Expression and purification of recombinant B. anthracis proteins
The recombinant His 6 -tagged fusion proteins were overexpressed and purified from E. coli and B. anthracis as described previously (28). E. coli BL21-DE3 strains harboring plasmid pACYC-PrkC or pACYC-PrpC were cotransformed with Bas4985 (eno)-containing plasmid (pProEx-HTc) to overexpress and purify Eno-P or Eno-UP, respectively. The phosphorylation status was checked by staining with Pro-Q Diamond phospho-specific stain (Molecular Probes, Life Technologies) according to the manufacturer's instructions. The protein amounts were checked by SYPRO Ruby protein gel stain (Molecular Probes, Life Technologies) and Coomassie Brilliant Blue stain.

Phenotypic memory in B. anthracis Sporulation and germination efficiency
Spores were prepared as described in the previous study (41). Different dilutions of heat-treated and nontreated spores were plated on LB agar, and cfu were counted. The sporulation efficiency was calculated as cfu per ml (heat-treated)/cfu per ml (nontreated) and compared with respect to Bas-wt (taken as 100%) (42). For germination efficiency, spores were diluted to an OD 600 of 1 and heat-treated at 70°C for 30 min (43). The heat-treated spores were serially diluted in deionized water, plated on LB agar (without antibiotics), and incubated at 37°C overnight. cfu were counted, and Bas-wt spore cfu were taken as 100%. Statistical analysis was performed using parametric t test.

Eno activity assay
The activity of His 6 -tagged Eno-UP and Eno-P (1 g) was measured using an Eno colorimetric activity assay kit according to the manufacturer's protocol (Biovision). Eno catalyzes the conversion of 2-PGA to PEP. The intermediate product formed reacts with a peroxide substrate to generate color (OD 570 ) at 25°C proportional to the Eno activity. A standard curve was generated with different dilutions of H 2 O 2 , and Eno activity was calculated using the following formula. Where B is the amount (nmol) of H 2 O 2 generated between T initial and T final reaction: Reaction time ϭ T final Ϫ T initial (min). V is the sample volume (ml) added to the well.

Mass spectrometry analysis
Samples were resolved by SDS-PAGE and trypsinized to prepare peptide mixtures for mass spectrometric analysis (45). Peptides were separated and measured by LC-electrospray ionization-MS using the Easy-nLCII HPLC system (Thermo Fisher Scientific) coupled directly to an LTQ Orbitrap Velos TM mass spectrometer (Thermo Fisher Scientific). Proteins were identified by searching all MS/MS spectra against a forwardreverse database that was composed of all protein sequences of B. anthracis Sterne and common contaminants using Sorcerer TM -SEQUEST (version v.27, rev.11, Thermo Fisher Scientific) in conjunction with Scaffold (version 3 00 06, Proteome Software Inc., Portland, OR).  Fig. S7. C, lysates and the secretory fraction from Bas-wt and Bas⌬prkC were subjected to immunoblotting using anti-Eno antibody (upper panel) and normalized by reprobing with anti-PA (for normalization). ***, p Յ 0.001; ns, not significant as determined by two-tailed unpaired Student's t test.

Immuno-EM
Bas-wt, kinase-deletion mutant (Bas⌬prkC), and the complemented strain (Bas⌬prkC complement) were grown at 37°C to mid-log phase and harvested. The cells were fixed in 2% paraformaldehyde and 0.05% glutaraldehyde dissolved in 0.1 M sodium phosphate buffer (PB), pH 7.4, for 2 h and then washed three times with PB. The cells were then resuspended in 2% agar and harvested again. The cell pellets were immersed in 30% sucrose (w/v) overnight at 4°C and subjected to immunolabeling as described previously (46) using a 1:10 dilution of anti-Eno antibodies and preimmune serum as a control. Ultrathin sections (70 nm thick) were cut on an RMC ultramicrotome, stained with 1% uranyl acetate, and imaged in a Tecnai G2 20 twin (FEI) transmission electron microscope. Cell-surface enolase expression was normalized with cytosolic expression using Fiji-ImageJ and plotted using GraphPad Prism.

Magnesium binding assay
The Eno-UP and Eno-P forms of Eno were used in interaction studies with MgCl 2 . The reaction was initiated by the addition of the each of the proteins (1 M) to fluorescence assay buffer, buffer F (1 mM PEP (Sigma), 0.1 M HEPES, pH 7, and 7.7 mM KCl) with MgCl 2 (0.05 mM). A buffer containing PEP, HEPES, and KCl served as a control. The emission spectra were recorded from 310 to 430 nm after excitation at 280 nm (Fluoromax-3 spectrofluorimeter, Jobin Yvon Horiba) with an integration time of 1s. The association of MgCl 2 with Eno was measured by mixing increasing concentrations of MgCl 2 (0.05-10 mM) with a 1 M concentration of Eno (Eno-P and Eno-UP) in buffer F at 25°C. The experiment was performed by monitoring the fluorescence change over time (35).

Enzyme-linked immunosorbent assay (ELISA)
An indirect ELISA was performed as described earlier with some modifications (47). Briefly, His 6 -tagged Eno (100 ng/well) was dissolved in a coating buffer (carbonate-bicarbonate buffer, pH 9.6) and adsorbed on the surface of a 96-well ELISA plate (Maxisorb, Nunc) for 16 h at 4°C. The test serum was serially diluted (1% in phosphate-buffered saline (PBS)) and added to each well to be kept for 1 h at 37°C. Preimmune serum was used as a control. The experiment was done in triplicates, and the antibody titer was expressed as the reciprocal of the end-point dilution. In Bas-wt cells, there is a basal-level expression of Eno in vegetative cells that is carried over in the spores. The expression and activity are kept low by PrkC through phosphorylation so that the spores can germinate effectively. Middle panel, Eno overexpression strain. The expression is raised to 1.5-fold in this strain with subsequent activation in the spore leading to spore germination defect. Bottom panel, PrkC-deficient strain. Membrane-localized unphosphorylated Eno is carried over in spores with increased expression and activity leading to spore germination defect.

Phenotypic memory in B. anthracis Antibody generation and Western blotting
B. anthracis Eno (Bas4985) expressed in E. coli with a His 6 tag was purified with up to 98% homogeneity (as per SDS-PAGE analysis). The affinity-purified protein was dialyzed against PBS and confirmed to be endotoxin-free using a kitbased assay (Pierce). The protein was then used to immunize a group of Rabbit and BALB/c mice (n ϭ 3) first in combination with Freund's complete adjuvant and subsequently Freund's incomplete adjuvant. After three booster doses (2 weeks apart), the animals were bled, and the isolated serum was analyzed using an indirect ELISA, which showed an efficient titer against Eno. The specificity of the Western blots was determined by probing B. anthracis cell lysates with different dilutions of Eno serum. The antibodies recognize a ϳ45-kDa molecular-mass protein corresponding to Eno. Eno purified from E. coli was taken as a positive control, and GSH S-transferase was used as a negative control. Anti-PA antibody was used from the previous studies (41). For estimating protein size, prestained protein markers were used (Bio-Rad, catalog numbers 26616, 26619, and 26634).

Quantification and statistical analysis
For the radioactivity-based experiments, the autoradiograms were quantified using QuantityOne software (Bio-Rad), and the corresponding Coomassie Brilliant Blue-stained gels were quantified using Fiji-ImageJ. The normalized values were plotted using GraphPad Prism. The Western blots were quantified using Fiji-ImageJ software, and the respective values normalized by the controls were plotted using GraphPad Prism. For time-dependent 2D gel electrophoresis, the amount of phosphorylation at a given time point was calculated using the following formula.
Intensity of phosphorylation (at a given time point) ϭ (Intensity of phosphoisoforms/Intensity of the total input protein)ϫ100 (Eq. 2) For statistical significance, a parametric unpaired t test was performed with Welch's correction.

Circular dichroism spectroscopy
CD spectra of Eno and its mutants (Eno-wt, Eno-UP, Eno-P, Eno S336A/T363A/S367A ) were recorded using a Jasco J-815 CD spectropolarimeter. The spectrum was recorded from 190 to 250 nm at 25°C using a cell of 0.1-cm path length. Experiments were repeated thrice, and molar ellipticity values with respect to wavelength were analyzed using the K2D3 web server to estimate the protein secondary structure (48).

Animal ethics approval
The animal experiments were performed according to the Institutional Animal Ethics Committee (IAEC) of Defense Research and Development Establishment (registration number 37/GO/c/1999/CPCSEA dated April 13, 2011). The animals were maintained according to the approved guidelines of the Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA), Government of India. The study was also approved by the Institutional Biosafety Committee of Defense Research and Development Establishment (DRDO), Ministry of Defense, Government of India (protocol number IBSC/12/BT/AKG/22). For generation of antibodies, animals (mice and rabbits) were used after approval of the Animal Ethics Committee of CSIR-Institute of Genomics and Integrative Biology.