Sensor kinase KinB and its pathway‐associated key factors sense the signal of nutrition starvation in sporulation of Bacillus subtilis

Abstract Bacillus subtilis responds to environmental stress cues and develops endospores for survival. In the process of endospore formation, sporulation initiation is a vital stage and this stage is governed by autophosphorylation of the sensor histidine kinases. The second major sensor kinase KinB perceives the intracellular changes of GTP and ATP during sporulation. However, determination of the environmental signals as well as its related signaling pathway of KinB requires further elucidation. Our current study found that, contrary to the sporulation failure induced by ΔkinA in the nutrient‐rich 2× SG medium, the sensor kinase KinB sensed the environmental cues in the nutrient‐poor MM medium. Two other membrane proteins, KapB and KbaA, also responded similarly to the same external signal as KinB. Both KapB and KbaA acted upstream of KinB, but they exerted their regulation upon KinB independently. Furthermore, we demonstrated that both the SH3 domain and the α‐helix structure in KapB are required for sensing or transducing the signal of sporulation initiation. Collectively, our work here supplied the direct evidences that KinB and its pathway sense the external signal of nutrient starvation in MM medium, and further analyzes the interrelationship among KinB, KbaA, and KapB.


| INTRODUCTION
Nutrient starvation or other adverse environmental conditions generally prompts Bacillus subtilis to form endospores for survival (Cano & Borucki, 1995). The change from vegetative growth to endospore formation represents a significant shift in life history for a unicellular bacterium to survive in hostile environments (Hall-Stoodley, Costerton, & Stoodley, 2004;Stewart & Costerton, 2001). During this process, regulation of sporulation initiation is believed to be the most critical. The specific environmental cues stimulate autophosphorylation of the sensor kinases and then a phosphoryl group will be transferred through a multicomponent phosphorelay system (Fujita & Losick, 2005). Briefly, the activated sensor histidine kinase first phosphorylates the relay protein Spo0F into Spo0F~P. The phosphoryl group of Spo0F, in turn, is transferred to Spo0B. Subsequently, Spo0B donates the phosphoryl group to the key regulator Spo0A. As the intracellular activation of Spo0A~P reaches a certain threshold, it turns on transcription of those sporulation-related genes (Burbulys, Trach, & Hoch, 1991;Fujita & Losick, 2005). The extracellular or intracellular signal activating the sensor histidine kinases is one of the earliest events for endospore formation.
In B. subtilis, the sensor kinases include at least five members, such as KinA-KinE, that sense different signals of sporulation (Jiang, Shao, Perego, & Hoch, 2000;LeDeaux, Yu, & Grossman, 1995;Piggot & Hilbert, 2004). Among them, KinA, as an intracellular sensor kinase having no protein domains outside the cell, has been suggested to respond to the shift in available ATP pool (Stephenson & Hoch, 2001).
But anyway, KinA and KinB are believed to be the major kinases for initiating sporulation.
KinB is the second major sensor kinase. During sporulation, a concurrent change of GTP decrease and ATP increase upregulates the transcriptions of kinB and kinA, which ultimately leads to the increase in Spo0A~P and the activation of the sigma cascade to produce endospores (Tojo, Hirooka, & Fujita, 2013). Structural analysis of KinB demonstrated that it was a membrane protein and composed of six transmembrane domains, a DHp domain, and a CA domain (Bick et al., 1992). Furthermore, the structure prediction indicated that KinB contained only loop regions and an N-terminal segment as an extracellular sensor rather than an integral domain outside functioning as a sensor of extracellular signals (Parkinson & Kofoid, 1992). This characteristic of KinB implies that it is unlikely to receive extracellular signals directly, and some other membrane or membrane-related proteins should be coupled with KinB (Dartois, Djavakhishvili, & Hoch, 1996;Phillips & Strauch, 2002). Indeed, it has been reported that another gene, kapB, was localized in the same operon as kinB using a single promoter. Their functions seemed to be related since the inactivation of kapB also led to a sporulation defect in the mutant strain MB340 (DkinA96) just as the KinB mutant did, suggesting that KapB plays a regulatory role in the expression of KinB or KapB is essential for the activation of KinB (Dartois, Djavakhishvili, & Hoch, 1997;Trach & Hoch, 1993). KbaA is another integral membrane protein containing six potential membrane-spanning helices. It has been described that KbaA may execute a positive role to trigger the activation of KinB at the onset of sporulation (Dartois et al., 1996).
Although it has been described that the roles of proteins KbaA and KapB are coupled with KinB, most of the experimental evidences in this area were from the strain MB340 (DkinA96) that is a mutant strain with kinA deletion (Dartois et al., 1996(Dartois et al., , 1997Trach & Hoch, 1993

| Bacterial strains, plasmids, and media
The strains of B. subtilis and Escherichia coli as well as the plasmids used in this study are listed in Table 1 Table 2.

| Assay for growth kinetics
Growth rates of different strains, including ΔkinA, ΔkinB, ΔkbaA, ΔkapB, and wild-type strain B. subtilis 168, in LB and MM media were determined.
One ml of overnight culture of each strain was added to 100 ml of fresh LB and MM medium, and the optical density (OD600) was measured every 2 hr for a growth kinetic. The data for each strain was collected from at least three biological replicas to determine their growth rates.

| Sporulation assays
To determine the efficiencies of sporulation, the bacterial strains were first incubated in a shaker in LB broth at 37°C for 8-10 hr. After  1995). The data for each strain was collected from at least three biological replicas.

| β-Galactosidase assays
To analyze the expression of spoIIG, the promoter region of spoIIG was fused to pDG1728 that contained the reporter gene lacZ. The pspoIIG reporter plasmid was successfully constructed by Shanghai Generay Biotech Co. Ltd. Then the reporter plasmid was transformed into the wild-type strain B. subtilis 168 as well as the other mutants, respectively.
After cultivating in 2× SG or MM medium at 37°C for 24 hr to induce sporulation, the strains containing lacZ fusions were analyzed for β-galactosidase activities as previously reported (Ferrari, Henner, Perego, & Hoch, 1988). Briefly, the activities were assessed with οnitrophenylβd-galactopyranoside as a substrate and were expressed in Miller units. All assays were repeated at least three times for each strain.

| Prediction of protein domain boundaries
Structural prediction of KapB was carried out at http://smart.emblheidelberg.de/ (Schultz, Milpetz, Bork, & Ponting, 1998). Based on the results of online prediction, we constructed two vectors to overexpress the mutant proteins of KapB, including one protein without αhelix (1-240 nt) and the other without SH3 domain (120-387 nt). The amplified gene fragments were digested with SalI and SphI and cloned into the plasmid pDG148 to obtain plasmids pDG148-kapB(1-80) and pDG148-kapB(41-124), respectively. These two recombinant plasmids above, as well as the blank vector pDG148, were all transformed into the ΔkapB mutant.

| Western blotting
The bacterial strains were grown in LB broth at 37°C to an optical density of OD 600 0.6, at which point isopropyl βd-

| Statistical analysis
All data were calculated and expressed as the mean ± standard deviation (SD) before statistical analyses. Statistical comparisons were performed by a one-way analysis of variance (ANOVA) followed by Dunnett's t test.

| KinB is involved in sensing nutrient starvation in MM medium to initiate sporulation
To differentiate the environmental signals that the two major sensor kinases KinA and KinB perceive during sporulation initiation, we compared the frequencies of sporulation of △kinB with the wild-type and △kinA strains after they were each placed in both the nutrient-rich 2× SG medium and the nutrient-poor MM medium. Because the different strains may have the different growth rates and the growth rates can sequentially influence sporulation rates, we analyzed their growth rates before the sporulation frequencies were determined. In growth kinetics of the wild-type strain B. subtilis 168 and ΔkinA and ΔkinB mutants, it was shown that all displayed similar growth rates in LB medium and the cells reached their stationary phase within 12 hr (Figure 1a). However, in the nutrient-poor MM medium, the cell densities of B. subtilis 168, ΔkinA, and ΔkinB were much lower than those in LB medium, and their growth rates were more variable ( Figure 1b). Thus, in our sporulation assay, the tested bacterial strains were first propagated in LB broth at 37°C for 8-10 hr, and after the cultures were washed with 2× SG or MM medium, all cells were then spotted into the same fresh medium to induce sporulation. Our results of sporulation demonstrated that, comparing with the wild-type strain (45.8 ± 1.5% and 70.4 ± 7.5%), the sporulation frequency of the ΔkinA mutant (3.8 ± 0.8% and 13.9 ± 0.7%) dropped significantly at either 24 hr or 36 hr in 2× SG medium (p < .05), but no significant difference was observed between the ΔkinB mutant (46.7 ± 3.6% and 82.9 ± 5.3%) and the wild-type strain at the same time points (p ≥ .05) ( Figure 1c). However, when the frequencies of sporulation of ΔkinA and ΔkinB mutants were determined again in MM medium, the results were opposite: the sporulation frequency in the ΔkinB mutant (1.9 ± 0.0% and 2.0 ± 0.1%) decreased sharply at either 24 hr or 36 hr (p < .05); ΔkinA (41.2 ± 7.3%) retained the similar capability for forming endospore at 36 hr to the wild-type strain (40.2 ± 4.4%) (p ≥ .05) ( Figure 1d). These results suggest that, between the most common sensor kinases KinA and KinB, the latter preferentially responds to the environmental cues of nutrient starvation in MM medium directly or indirectly during sporulation.
To further verify our hypothesis that KinB was more responsible for sensing nutrient starvation in MM medium and to activate the downstream sporulation-related genes, the promoter region of spoIIG, a gene known to be under the direct control of Spo0A (Satola, Baldus, & Moran, 1992), was cloned and fused to the reporter plasmid pDG1728 to drive the expression of β-galactosidase to show the activation of phosphorelay system during sporulation initiation. Altogether, three reporter strains were successfully constructed by transforming the recombinant plasmid into the wild-type strain B. subtilis 168, ΔkinA, and ΔkinB, respectively.
After analyzing their β-galactosidase activities, it was found that compared to the wild-type strain (78.6 ± 9.7 nmol/ml/hr), the β-galactosidase activity in the ΔkinA mutant was reduced significantly (p < .05), about half of the wild-type strain at 24 hr in 2× SG medium (32.4 ± 5.2 nmol/ml/hr).
Meanwhile, qPCR experiment was employed to determine the expressional level of kinA after transferred into MM medium. The result showed that the expression of kinA could be induced to more than twofold after 30 hr (Figure 1g). Collectively, the results above suggest that KinB played a role in perceiving the environmental signal in the MM medium though the other histidine kinase KinA retained its normal expression synchronously.
Furthermore, the results from β-galactosidase assays were consistent with those of sporulation (Figure 2c,d).

| KapB and KbaA function upstream and regulate KinB independently
To explore the regulatory relationships among KbaA, KapB, and F I G U R E 1 KinB is a more important sensor than KinA in response to the external cues of nutrient starvation in MM medium. (a, b) The growth curves of wild-type, ΔkinA, ΔkinB, ΔkapB, and ΔkbaA in Luria-Bertani (LB) and MM media. All the tested mutants and the wild-type strain showed similar growths in LB media, while they had different growths in MM media. (c, d) Comparing the sporulation frequencies of ΔkinA, ΔkinB, and the wild-type Bacillus subtilis 168 in nutrition-rich 2× SG medium and nutrition-poor MM medium, respectively. Contrary to the sporulation failure due to ΔkinA mutation in the nutrient-rich 2× SG medium, disruption of kinB led to a serious sporulation defect only in MM medium. (e, f) β-Galactosidase activities of ΔkinA, ΔkinB, and the wild-type B. subtilis 168 in nutrition-rich 2× SG medium and nutritionpoor MM medium after growing for 36 hr, respectively. Compared to the ΔkinA mutant and the wild-type strain, ΔkinB had significantly decreased β-galactosidase activity in MM medium. (g) The experiment of qPCR determined the relative expressional level of kinA in MM medium. not significant p ≥ .05, *p < .05 Since our current data indicated that both KabA and KapB functioned upstream of KinB, we further investigated if KapB and KbaA modulated the sensor kinase KinB via the same or independent pathway(s). After constructing the double mutant strain of ΔkbaAΔkapB and analyzing its sporulation frequency and β-galactosidase activity, we found that the double mutant ΔkbaAΔkapB represented much lower sporulation frequency and β-galactosidase activity than either ΔkbaA or ΔkapB (p < .05) (Figure 4a,b). Furthermore, it was also noticed that the sporulation frequency and β-galactosidase activity of the ΔkbaAΔkapB double mutant were similar to those of the ΔkinB mutant. Based on those experimental evidences, it was reasonable to speculate that KapB and KbaA should regulate KinB independently, but not in the same pathway.

| Analysis of the protein domains of KapB required in KinB-dependent sporulation
The protein structure prediction of KapB revealed that it had two main domains, an SH3 domain (3-40 amino acids) and an α-helix domain  (c, d) In the ΔkinB, ΔkapB, ΔkinBΔkapB, pDG148-kinBΔkinBΔkapB, and pDG148-kapBΔkinBΔkapB mutants, the sporulation frequencies and β-galactosidase activities were analyzed. Similarly, only the KinB-complemented strain could restore the majority of sporulation frequencies and β-galactosidase activities of the double mutant strain ΔkinBΔkapB. n.s. p ≥ .05, *p < .05 using the wild-type strain Bacillus subtilis 168 as control, **p < .05 using the double mutant strains ΔkinBΔkbaA or ΔkinBΔkapB as control research group revealed that the kinB gene was expressed at a higher rate than kinA during exponential growth and reached a maximum 1.5 hr before kinA transcription. Correspondingly, the absence of kinB delayed the transcription of spoIIG for 1 hr but its ultimate expressional level was not changed significantly (Dartois et al., 1996). At the same time, another study from the same laboratory also showed that in the absence of kinA and kinB, the phosphorylation mediated by KinC and KinD could also happen at the exponential phase of growth.
Thus, it was concluded that all the kinases were expressed at the same stage and the differential activities observed in growth and sporulation might result from differential activation by the signal ligands unique to each kinase (Jiang et al., 2000).
Since sporulation can be initiated via both the external and in- In the background of ΔkapB, the intact KapB or the truncated KapB without either SH3 domain or α-helix was complemented and expressed and then the sporulation frequencies were determined respectively. Only the complement by intact KapB could rescue the sporulation defect of ΔkapB. n.s. p ≥ .05, *p < .05 caused more serious defects than the single gene deletion. Although it is reasonable to hypothesize that KbaA and KapB might have their impacts on sensing the signal of nutrition starvation, those two proteins might only be involved in maintaining KinB as previously reported (Dartois et al., 1996), because the stability and the subcellular localization of KinB had not been observed when KapB was abnormal.
Collectively, our data here support that KinB plays a more important role during sporulation initiation triggered by the nutrient starvation in MM medium than KinA does, which is consistent with the hypothesis that there are different signals unique to each histidine kinase activating their own sensors and the shared downstream phosphorelay. We further analyzed the interrelationship among