Evidence that Autophosphorylation of the Major Sporulation Kinase in Bacillus subtilis Is Able To Occur in trans

ABSTRACT

Entry into sporulation in Bacillus subtilis is governed by a multicomponent phosphorelay, a complex version of a two-component system which includes at least three histidine kinases (KinA to KinC), two phosphotransferases (Spo0F and Spo0B), and a response regulator (Spo0A). Among the three histidine kinases, KinA is known as the major sporulation kinase; it is autophosphorylated with ATP upon starvation and then transfers a phosphoryl group to the downstream components in a His-Asp-His-Asp signaling pathway. Our recent study demonstrated that KinA forms a homotetramer, not a dimer, mediated by the N-terminal domain, as a functional unit. Furthermore, when the N-terminal domain was overexpressed in the starving wild-type strain, sporulation was impaired. We hypothesized that this impairment of sporulation could be explained by the formation of a nonfunctional heterotetramer of KinA, resulting in the reduced level of phosphorylated Spo0A (Spo0A∼P), and thus, autophosphorylation of KinA could occur in trans. To test this hypothesis, we generated a series of B. subtilis strains expressing homo- or heterogeneous KinA protein complexes consisting of various combinations of the phosphoryl-accepting histidine point mutant protein and the catalytic ATP-binding domain point mutant protein. We found that the ATP-binding-deficient protein was phosphorylated when the phosphorylation-deficient protein was present in a 1:1 stoichiometry in the tetramer complex, while each of the mutant homocomplexes was not phosphorylated. These results suggest that ATP initially binds to one protomer within the tetramer complex and then the γ-phosphoryl group is transmitted to another in a trans fashion. We further found that the sporulation defect of each of the mutant proteins is complemented when the proteins are coexpressed in vivo. Taken together, these in vitro and in vivo results reinforce the evidence that KinA autophosphorylation is able to occur in a trans fashion.

IMPORTANCE Autophosphorylation of histidine kinases is known to occur by either the cis (one subunit of kinase phosphorylating itself within the multimer) or the trans (one subunit of the multimer phosphorylates the other subunit) mechanism. The present study provided direct in vivo and in vitro evidence that autophosphorylation of the major sporulation histidine kinase (KinA) is able to occur in trans within the homotetramer complex. While the physiological and mechanistic significance of the trans autophosphorylation reaction remains obscure, understanding the detailed reaction mechanism of the sporulation kinase is the first step toward gaining insight into the molecular mechanisms of the initiation of sporulation, which is believed to be triggered by unknown factors produced under conditions of nutrient depletion.

INTRODUCTION

Bacterial cells are directly exposed to a fluctuating environment. To survive under such conditions, they must sense changes in various environmental factors such as nutrients, temperature, and osmolarity and respond rapidly by changing their gene expression and subsequent cellular processes (1, 2). To transduce such environmental signals to cellular responses, bacterial cells use two-component signal transduction pathways, which involve a sensor histidine kinase and its cognate substrate, a response regulator (2–4). To monitor environmental conditions, the sensor histidine kinase is typically a membrane-bound homodimeric protein with an extracytoplasmic sensory domain linked to a cytoplasmic transmitter domain through transmembrane helices (3). By directly responding to environmental stimuli, the sensor kinase undergoes autophosphorylation on a histidine residue located in the C-terminal cytoplasmic transmitter domain (3). Subsequently, the phosphoryl group at the histidine is transferred to an aspartic acid residue on the response regulator. In most cases, the response regulator protein becomes active only upon phosphorylation, resulting in binding to a target DNA sequence and thereby regulating the expression of downstream genes.

Generally, bacterial two-component systems involve reversible protein phosphorylation and dephosphorylation to regulate gene expression and adapt to changes in the environment (3). For this, the cellular level of the phosphorylated response regulator is strictly controlled through the bifunctional sensor kinase possessing both kinase and phosphatase activities toward the response regulator, as well as with additional auxiliary protein phosphatases (3, 5, 6). Therefore, when cells adapt to the environment, the specific cellular response is diminished and the steady state is restored by changing the level of the active response regulator.

Upon nutrient depletion, entry into sporulation in Bacillus subtilis is governed by a multicomponent phosphorelay, a complex version of the two-component system which consists of the major sporulation kinase KinA, two phosphotransferases (Spo0F and Spo0B), and the master transcriptional regulator Spo0A in a His-Asp-His-Asp signaling pathway (Fig. 1) (4, 7, 8). While many of the bacterial two-component systems control rapid and transient gene expression in response to various and specific stimuli (3), the phosphorelay system is involved in the control of the crucial and highly coordinated cell fate decision, which ultimately leads to the development of spores (2, 4, 9–11). Accumulated evidence indicates that a temporal and spatial increase in the level and activity of the master regulator Spo0A is required for sporulation to proceed properly (12–14). Under nutrient-rich conditions, the DNA-binding protein AbrB represses genes involved in the transition from vegetative growth to stationary phase, and only a basal level of Spo0A is expressed from the relatively weak, constitutive σ^A-controlled promoter Pv (15). When cells enter stationary phase, σ^H RNA polymerase activity increases slightly, leading to the increased expression of Spo0A from a stronger σ^H-controlled promoter, Ps (15). This increased level of Spo0A is sufficient to repress abrB transcription by binding directly to the cis element of the abrB gene. The resulting decrease in the concentration of AbrB leads to the derepression of the σ^H gene (Fig. 1) (4, 9). As a result, the increased σ^H RNA polymerase activity stimulates the expression of genes for KinA, Spo0F, and Spo0A, causing a further increase in Spo0A. In addition, the transcription of genes for Spo0F and Spo0A is positively regulated by Spo0A (16). Through this feedback regulation of the phosphorelay, a gradual increase in the cellular level and activity of Spo0A is achieved (12, 13).

FIG 1.

The sporulation phosphorelay that leads to the phosphorylation of the master regulator Spo0A. The phosphorelay includes both posttranslational (solid lines) and transcriptional (dashed lines) regulatory interactions. Posttranslationally, the kinase KinA transfers a phosphoryl group to the master regulator Spo0A (0A) via the two phosphotransferases Spo0B (0B) and Spo0F (0F). Transcriptionally, 0A∼P controls the expression of kinA, 0F, and 0A both directly and indirectly via AbrB and σ^H forming multiple transcriptional feedback loops. To modulate proper timing of sporulation, the phosphatases RapA and Spo0E inactivate the signaling pathways by dephosphorylating 0F∼P and 0A∼P, respectively.

As the initial event in sporulation, it has been proposed that an unidentified signaling molecule is produced by cells only under starvation conditions and is received by the sensor domain of one or a combination of the kinases, resulting in autophosphorylation of the histidine residue at the C-terminal domain of kinase (4, 10, 17). However, the molecular details are not yet clear because the putative signaling molecule has never been identified. Recently, using a synthetic biology approach in combination with computer modeling, an alternative model has been proposed for entry into sporulation in which the initial event of phosphorelay activation is an increase in the cellular level of KinA, which is constitutively active in a manner independent of an unidentified ligand(s) and, once the protein concentration reaches a threshold level, sporulation can be triggered (13, 18–21). The mechanism(s) controlling the cellular concentration of the kinase is unknown, and thus, this model is still under debate.

While the typical two-component sensor kinases are bifunctional enzymes having both kinase and phosphatase activities (3, 5), KinA shows little or no phosphatase activity on the cognate response regulator phosphorylated Spo0F (Spo0F∼P) (22). Instead, levels of Spo0F∼P and Spo0A∼P are regulated by specific phosphatases, RapA and Spo0E, that remove phosphoryl groups from Spo0F∼P and Spo0A∼P, respectively (6, 23, 24). The rapA gene is subject to control by Spo0A∼P directly, and the spo0E gene is known to be repressed by AbrB (25, 26). Thus, the phosphorelay network is controlled at both the transcriptional and posttranslational levels as a whole, resulting in both temporal and spatial regulation of the activity of Spo0A (Fig. 1) (13, 14, 18, 26, 27). Based on the findings and facts described above, it has been proposed that the complexity of the phosphorelay provides multiple entry points to control the precise concentration of the master regulator Spo0A during the development and progression of sporulation (13, 27).

Nevertheless, a detailed characterization of the kinase autophosphorylation is a first step toward better understanding of not only sporulation but also bacterial signal transduction systems in general. Until now, extensive biochemical and genetic data have provided evidence that the autophosphorylation of the sensor histidine kinases occurs in either a cis (28, 29) or a trans (30–34) fashion, although the physiological and mechanistic significance of these two types of autophosphorylation reactions remains obscure.

The recent results of a computational approach suggest that the autophosphorylation reaction of KinA occurs within the same protomer in the homodimer in a cis fashion (33). In contrast, our recent study demonstrated that KinA forms a homotetramer, not a dimer, mediated by the N-terminal domain, as a functional unit. Furthermore, when the N-terminal domain was overexpressed in the wild-type strain cultured under starvation conditions, sporulation was impaired (20). This impairment of sporulation can be explained by the formation of a nonfunctional heterotetramer of the kinase in a dominant negative effect, resulting in reduced levels of Spo0A∼P (20). These results suggest that autophosphorylation of KinA occurs in trans, but no direct experimental data are available to support this notion.

In this study, using a combination of biochemical and genetic approaches, we obtained direct evidence that the autophosphorylation of KinA is able to occur in a trans manner, in which ATP binds to one protomer and then the γ-phosphoryl group of the bound ATP is transferred to the other protomer within the same homotetramer complex.

MATERIALS AND METHODS

Strains, plasmids, and oligonucleotides.

All B. subtilis strains (see Table S1 in the supplemental material) were derived from the prototrophic strain PY79 (35). Details of the constructions are available upon request. All plasmid constructions were performed in Escherichia coli DH5α using standard methods (36). The E. coli BL21(DE3) pET vector system (Novagen) was used for protein overexpression. The plasmids used in this study are listed in Table S2 in the supplemental material. Oligonucleotides used for plasmid construction are listed in Table S3 in the supplemental material.

Media and culture conditions.

To induce protein synthesis under the control of an isopropyl-β-d-thiogalactopyranoside (IPTG)-inducible hyper-spank promoter (Phy-spank) (37) in the engineered B. subtilis cells, the desired concentration of IPTG was added to Luria-Bertani (LB) cultures during the exponential growth phase (optical density at 600 nm [OD₆₀₀], 0.5) as the rich medium conditions. To induce protein synthesis under normal sporulation conditions, the engineered cells were grown in casein hydrolysate (CH) medium at 37°C. At the mid-exponential phase of growth (OD₆₀₀, 0.5) in CH medium, cells were suspended in Sterlini-Mandelstam (SM) medium (38) supplemented with IPTG at the concentrations indicated below.

Ni-NTA pulldown assays.

Pulldowns were performed by expressing the endogenous full-length KinA_his6 (bait) with the IPTG-induced full-length and the mutant derivatives of green fluorescent protein (GFP)-tagged KinA (prey). Cells were induced for sporulation in SM medium (39) with 0.1 mM IPTG for 3 h before being harvested by centrifugation. For pulldown assays, cells were suspended in lysis buffer (50 mM Tris-HCl [pH 8], 20% [wt/vol] glycerol), lysed by sonication, and clarified by centrifugation. Cells expressing untagged KinA with the IPTG-induced GFP-tagged KinA served as a negative control. Lysates were mixed with 100 μl (50% slurry) of nickel-nitrilotriacetic acid (Ni-NTA)–agarose (Qiagen) equilibrated in lysis buffer and incubated with rocking at 4°C for 1 h. The resin was pelleted by centrifugation, the supernatant removed, and the resin washed five times with 1 ml of lysis buffer containing 5 mM imidazole. After the last wash, 100 μl of elution buffer (50 mM Tris-HCl [pH 8], 20% [wt/vol] glycerol, 200 mM imidazole) was added to the pelleted resin. Total sonicated proteins (2 μg) and an equal volume of eluted pulldown protein complexes were analyzed by SDS-PAGE followed by immunoblotting using anti-GFP antibodies (20).

Immunoblot analysis.

Antigens derived from KinA were overexpressed in E. coli using pET vectors containing the strong T7-lac promoter. pET vector-based plasmids pMF400 and pMF402 were designed to express the N terminus of KinA (KinA_N; 266 amino acids [aa]) and the C terminus of KinA (KinA_C; 155 aa), respectively. More details for plasmid constructions are provided in the supplemental material. The resulting N- and C-terminal domains were overexpressed in E. coli and purified using Ni-NTA–agarose (Qiagen). Antibodies against KinA_N and KinA_C were prepared in rabbits by Cocalico Biologicals, Inc. Whole-cell lysates for immunoblot analysis were prepared by sonication. Protein concentration was measured by Bradford method (Pierce). Total proteins were subjected to SDS-PAGE (16% acrylamide) and transferred to a nitrocellulose filter. Immunoblot analysis was done with polyclonal anti-GFP antibodies (a gift from David Rudner) (40) or anti-KinA antibodies. Alkaline phosphatase-coupled secondary antibodies were used for recognizing the primary antibodies and detected using substrate nitro blue tetrazolium–5-bromo-4-chloro-3-indolyl phosphate (NBT-BCIP) (Promega). The intensities of the individual bands were quantified with an image analyzer (FluorChem; Alpha Innotech). σ^A, a constitutively expressed protein, detected by anti-σ^A antibodies was used as a loading control (38).

Sporulation efficiency and β-galactosidase assays.

Assays of sporulation efficiency and β-galactosidase activity were performed as described previously (19).

Protein purification.

All His-tagged protein purifications were carried out at 4°C as described previously (16, 41).

In vitro phosphorylation.

Phosphorylation reactions were performed as described previously (42).

RESULTS

Construction of the phosphoryl-accepting histidine point mutant and the catalytic ATP-binding domain point mutant of KinA.

While the typical two-component histidine kinases act as homodimers (3, 43), our recent studies demonstrated that KinA forms a functional homotetramer mediated by the N-terminal domain (19, 20). Furthermore, the overexpression of the N-terminal domain in the sporulating wild-type strain interferes with sporulation, suggesting that the formation of nonfunctional heterotetramers composed of the overexpressed N terminus and the endogenous KinA results in a reduced level of autophosphorylation in a dominant negative effect and leads to the reduced production of Spo0A∼P through the phosphorelay (20). Taking into account the fact that the gradual increase in the levels of Spo0A∼P over the course of starvation is critical for triggering sporulation, a subtle reduction of the Spo0A∼P levels is sufficient to cause a significant decrease in sporulation (18). These results support a trans autophosphorylation mechanism. To provide direct evidence for this mechanism, we generated a series of B. subtilis strains expressing homogeneous or heterogeneous protein complexes consisting of various combinations of the phosphoryl-accepting histidine point mutant protein and the catalytic ATP-binding domain point mutant protein.

In the previously published studies, neither a single amino acid substitution in the phosphoryl-accepting histidine residue (H405) (22) nor a single mutation of the catalytic ATP-binding (CA) domain in the KinA histidine kinase displayed autophosphorylation activity (e.g., Q508) (33). Based on these results, we hypothesized that if a trans autophosphorylation occurs, it should be possible that the heterogeneous tetramer complex of KinA consisting of the phosphoryl-accepting histidine mutant protein and the catalytic ATP-binding domain mutant protein could complement for the autophosphorylation reaction.

By examining the published data (29, 33), we identified a conserved amino acid residue (N516) in KinA, corresponding to the ATP-binding N380 of the histidine kinase HK853 of Thermotoga maritima (Fig. 2), and replaced it with alanine (N516A). Independently, the phosphoacceptor histidine residue (H405) of KinA (22) was replaced with alanine (H405A). To measure autokinase activities of the KinA and its mutant proteins, each of the genes was placed under the control of an IPTG-inducible promoter (Phy-spank) (37) and the construct was integrated at the amyE locus of the B. subtilis chromosome. To exclude the endogenous pathways which activate Spo0A through the phosphorelay, the kinA and kinB genes were deleted from their original loci on the chromosomal DNA. As controls, the full length, the N terminus, the C terminus, and the deletion of the C-terminal catalytic ATP-binding domain (ΔCA) of KinA were similarly constructed. In vivo autokinase activity of KinA can be measured only indirectly with a reporter for the phosphorylation status of the terminal response regulator Spo0A in the multicomponent phosphorelay system (20). Thus, a lacZ fusion to the Spo0A-directed spoIIG operon promoter (PspoIIG-lacZ), which has been generally used to detect high-threshold Spo0A activity sufficient for entry into sporulation (26), was introduced at the thrC locus of each of the strains as described above. In this engineered system, now known as artificial sporulation initiation (ASI) (18), KinA and its mutant derivatives can be expressed by adding IPTG, and the resulting autokinase activity and sporulation efficiency can be determined under nutrient-rich conditions in which the possible involvement of an unknown starvation signal(s) can be excluded for simplicity.

FIG 2.

Domain architecture of KinA and amino acid sequence alignment of the conserved region in the C-terminal domain of KinA, KinB, KinC, and HK853. The diagram at the top shows three PAS domains, the dimerization and histidine phosphotransfer (DHp) domain, and catalytic ATP-binding (CA) domains of KinA. The conserved four amino acid sequences of KinA (8), KinB (56), KinC (57, 58), and HK853 (44) are aligned. The mutated amino acid residues are bold and underlined. An asterisk indicates an identical or conserved residue, a colon indicates a conserved substitution, and a period indicates a semiconserved substitution.

As shown in Fig. 3, cells expressing either the phosphoacceptor mutant KinA^H405A or the catalytic ATP-binding mutant KinA^N516A displayed no detectable β-galactosidase activities and low sporulation efficiencies (Fig. 3, lanes 4 and 5), similar to those of the nonfunctional mutants (N terminus, ΔCA, and C terminus) (Fig. 3, lanes 2, 3, and 6). In contrast, cells expressing the full-length KinA showed significant β-galactosidase activities and sporulation efficiencies (10⁸ spores/ml of culture), similar to those in the sporulating wild type, as reported previously (Fig. 3, lane 1) (20). We verified that all the proteins were expressed at similar levels by immunoblotting (Fig. 3). These results indicated that the failure to detect the reporter activities is due to the loss-of-function mutation and not due to degradation or a lack of protein expression.

FIG 3.

Functional characterization of the phosphoacceptor mutant and the catalytic ATP-binding mutant. Genes for full-length KinA and each of the mutant forms were placed under the control of the Phy-spank promoter and integrated at the nonessential amyE locus of the kinA and kinB double deletion mutant. To monitor autokinase activity indirectly, the Spo0A-directed PspoIIG-lacZ reporter was introduced at the nonessential thrC locus of each of the strains. Cells were grown in LB medium and IPTG was added to a final concentration of 0.1 mM at mid-exponential phase (OD₆₀₀ = 0.5). Strains used were MF5554 (KinA), MF5522 (N, KinA_N), MF5563 (ΔCA, KinA^ΔCA), MF5564 (H405A, KinA^H405A), MF5566 (N516A, KinA^N516A), and MF5561 (C, KinA_C). A schematic diagram of KinA or its mutant derivatives is shown at the top of each lane of the gel. The position of the amino acid substitution is marked by an asterisk. Cells for immunoblot analysis were collected at hour 3 after IPTG addition. Total proteins (2 μg) were subjected to SDS-PAGE (16% acrylamide) and transferred to a nitrocellulose filter. Immunoblot analysis was done with polyclonal anti-KinA_N and anti-KinA_C antibodies as described in Materials and Methods. σ^A, a constitutively expressed protein detected by anti-σ^A antibodies, was used as a loading control. The amino acid length for each protein was indicated. β-Galactosidase (β-gal) activities were measured at hour 3 after IPTG addition. The means of three samples for β-galactosidase are shown with standard deviations. Numbers of spores (per milliliter of culture) were determined as described in Materials and Methods. The means from three analyses are indicated.

Next, we tested whether the newly constructed mutants of KinA form stable complexes mediated by the N-terminal domain as reported (20). For this, we first constructed a series of strains harboring GFP-tagged KinA and its mutants (KinA_N, KinA_C, KinA^ΔCA, KinA^H405A, and KinA^N516A) under the control of the IPTG-inducible Phy-spank promoter at the amyE locus. Then, each of these constructs was introduced into a strain which contained a hexahistidine-tagged copy of KinA (KinA_his6) at the native locus. We note that the GFP- and His-tagged KinA are functional (12, 20, 21). In this system, KinA_his6 can be expressed from the native σ^H-controlled promoter in the starvation medium (SM medium) via the phosphorelay system (Fig. 1) (39), and each of the GFP-tagged proteins can be simultaneously synthesized in excess by adding IPTG. Accordingly, the resulting engineered cells were cultured in SM medium in the presence of IPTG (0.1 mM) to allow coexpression of the proteins. Cells were collected 3 h postinduction, and the cell lysates were subjected to Ni-NTA–agarose column chromatography. The bound proteins were recovered and separated via SDS-PAGE followed by immunoblotting with anti-GFP antibodies. As shown in Fig. 4 (lanes 3 to 6 and 9 to 14), the complexes were detected when the N-terminal domain was present in the GFP-tagged prey proteins (KinA_N, KinA^ΔCA, KinA^H405A, and KinA^N516A) in the eluted fraction with the His-tagged KinA (bait). In contrast, the GFP fusion of the C-terminal domain of KinA as a prey protein was not detectable in the eluted fraction with the His-tagged KinA (bait) (Fig. 4, lanes 7 and 8). As a negative control, the GFP fusion of the full-length KinA (prey) was not detected in the eluted fraction with the untagged KinA (bait) (Fig. 4, lanes 1 and 2). These results confirmed that KinA^H405A and KinA^N516A form a stable complex mediated by the N-terminal domain. Based on the fact that KinA forms a homotetramer (20), each of these protein complexes is likely composed of a tetramer of each of the wild-type and mutant protomers.

FIG 4.

Detection of the homo- and heterocomplex formations between KinA and its mutant forms by Ni-NTA pulldown assay. B. subtilis cells expressing the full-length KinA_his6 (bait) from the native promoter with the IPTG-inducible full-length or mutant derivatives of GFP-tagged KinA (prey) protein were constructed. Cells were cultured in SM medium in the presence of IPTG (0.1 mM) to induce both the bait and prey expression (39) and collected at 3 h after induction. A Ni-NTA pulldown assay was carried out as described in Materials and Methods. Total proteins (T; 2 μg) and an equal volume of eluted pulldown protein fractions (E) were analyzed by SDS-PAGE followed by immunoblotting using anti-GFP antibodies (top). A schematic diagram of KinA or its mutant derivatives is shown at the bottom of each lane of the gel. The position of the amino acid substitution is marked by an asterisk. Strains used were MF4255 (no bait [minus]; prey, KinA-GFP) (lanes 1 and 2), MF4886 (bait, KinA_his6; prey, KinA-GFP) (lanes 3 and 4), MF4887 (bait, KinA_his6; prey, KinA_N-GFP [N]) (lanes 5 and 6), MF4888 (bait, KinA_his6; prey, KinA_C-GFP [C]) (lanes 7 and 8), MF4894 (bait, KinA_his6; prey, KinA^ΔCA-GFP [ΔCA]) (lanes 9 and 10), MF5690 (bait, KinA_his6; prey, KinA^H405A-GFP [H405A]) (lanes 11 and 12), and MF5692 (bait, KinA_his6; prey, KinA^N516A-GFP [N516A]) (lanes 13 and 14). Plus and minus signs indicate positive and negative results for complex formation between the bait and prey proteins, respectively, as shown in the bottom part.

Overexpression of the kinase activity-deficient KinA protein in starving wild-type cells interferes with Spo0A activity and sporulation.

We reasoned that if the KinA homotetramer autophosphorylates in a trans manner, as partially demonstrated previously (20), the mutant KinA could exert a dominant negative effect to interfere with the normal function of wild-type KinA when these two protomers form a heterotetramer. To test this, we constructed a series of strains expressing the GFP-tagged wild-type and mutant forms of KinA under the control of the IPTG-inducible Phy-spank promoter from the amyE locus in the strain expressing the untagged wild-type KinA from the native promoter. To monitor KinA activity, we introduced the Spo0A-directed PspoIIG-lacZ reporter into these constructs. To exclude the possible involvement of alternative sporulation kinase KinB, the kinB gene was deleted in all strains used for this assay (see Table S1 in the supplemental material). Cells of the engineered strains were induced to sporulate in SM medium (39) in the presence of IPTG (0.1 mM) to overproduce each of the GFP-tagged proteins as described above. As shown in Fig. 5A, we confirmed that all proteins tagged with GFP were detectable. When the GFP-tagged nonfunctional mutant proteins (KinA^H405A KinA^N516A) were overexpressed in the cells simultaneously expressing the native, untagged, wild-type KinA, β-galactosidase activities and sporulation efficiencies were significantly reduced compared with those in the noninduced controls (Fig. 5B and C, columns 13 to 16). When wild-type KinA was induced by adding IPTG in this system, β-galactosidase activities were significantly higher than those in the other constructs due to overexpression of the functional KinA (Fig. 5B and C, columns 5 and 6) (21). The control experiments using the other nonfunctional truncated proteins (KinA_N, KinA_C, and KinA^ΔCA) showed results similar to those obtained with the point mutants (Fig. 5B and C, columns 7 to 12). As additional controls, we used two strains, one expressing the GFP-tagged KinA and the other expressing the untagged KinA, both of which were under the control of the native promoter without the IPTG-inducible construct (Fig. 5, lanes 1 to 4). In these control strains, sporulation efficiencies were normal (∼10⁸ spores/ml), and no significant changes in the reporter activities and sporulation efficiencies were observed in either the presence or absence of the inducer (Fig. 5B and C). To more quantitatively assess the overexpression levels of the wild-type and the mutant forms of KinA, total proteins prepared from cells expressing KinA-GFP from the native promoter and both KinA-GFP and KinA_N-GFP from the Phy-spank promoter in the presence of IPTG (the same samples as in lanes 4, 6, and 8 in Fig. 5A) were serially diluted and processed for immunoblot analysis (Fig. 5D). By measuring the band intensities, the overexpression levels from the Phy-spank promoter were estimated to be at least 7-fold higher than those expressed from the native promoter. It is reported that autophosphorylation activity of wild-type KinA is independent of enzyme concentration, suggesting that the functional subunit assembly is independent of the concentration of the individual protomers (19). These results indicated that the overexpression of nonfunctional KinA from the Phy-spank promoter in the cells coexpressing wild-type KinA from the native promoter leads to the predominant formation of a nonfunctional heterocomplex with these two proteins, causing a dominant negative effect on Spo0A activity and sporulation. We note that the inhibitory effects caused by the overexpression of the C-terminal domain of KinA (KinA_C) is due not to the formation of a nonfunctional heterocomplex (Fig. 4, lanes 7 and 8) but rather to the reverse phosphotransfer reaction in the phosphorelay as demonstrated previously (20). Taken together, these observations are consistent with the idea that a dominant negative effect is achieved when the mutant KinA complexes with its wild-type counterpart to form a nonfunctional tetramer. We further note that in the above-described overexpression system, the cellular concentration of KinA expressed from the native promoter is constant regardless of IPTG addition (Fig. 5, lanes 3 and 4), indicating that overexpression of the mutant form interferes with the kinase activity but not with the expression level of the wild-type KinA per cell.

FIG 5.

Overexpression of a nonfunctional KinA in the wild-type background causes a dominant negative effect on Spo0A activity and sporulation. (A) Each of the GFP-tagged wild-type and mutant forms of KinA was expressed under the control of the Phy-spank promoter in the B. subtilis strain harboring a kinB deletion (ΔkinB) and the PspoIIG-lacZ reporter (lanes 5 to 16). As controls, two strains, one expressing the GFP-tagged KinA (lanes 3 and 4) and the other expressing the untagged KinA (lanes 1 and 2), both of which were under the control of the native promoter without the IPTG-inducible construct, were used. Cells were cultured in SM medium in the presence (+) and absence (−) of IPTG (0.1 mM). Total proteins (2 μg) prepared from cells collected at 3 h after induction in the presence and absence of IPTG were subjected to SDS-PAGE (16% acrylamide) and processed for immunoblot analysis using anti-GFP antibodies as described in Materials and Methods. σ^A was used as a loading control. (B) β-Galactosidase activities were measured at 3 h after induction in the presence and absence of IPTG. The mean values from three samples are shown with standard deviations. (C) Numbers of spores (per milliliter of culture) were determined as described in Materials and Methods. The means from three analyses are indicated. Sample numbers in panels B and C correspond to those in pane A. (D) Total proteins prepared from the same samples as in lanes 4, 6, and 8 in panel A (KinA-GFP from the native promoter [Nat] and both KinA-GFP and KinA_N-GFP from the Phy-spank promoter [Phy]) were serially diluted (1, 0.5, and 0.25 μg) and processed for immunoblot analysis using anti-GFP antibodies. Strains used were MF4840 (native KinA expressed from its own promoter, lanes and columns 1 and 2), MF6785 (GFP-tagged KinA expressed from the native kinA promoter, lanes and columns 3 and 4), MF4844 (GFP-tagged KinA expressed from the Phy-spank promoter and native KinA expressed from its own promoter, lanes and columns 5 and 6), MF4845 (N, GFP-tagged N-terminal domain of KinA expressed from the Phy-spank promoter and native KinA expressed from its own promoter, lanes and columns 7 and 8), MF4846 (C, GFP-tagged C-terminal domain of KinA expressed from the Phy-spank promoter and native KinA expressed from its own promoter, lanes and columns 9 and 10), MF4852 (ΔCA, GFP-tagged kinA^ΔCA expressed from the Phy-spank promoter and native KinA expressed from its own promoter, lanes and columns 11 and 12), MF5757 (H405A, GFP-tagged KinA^H405A expressed from the Phy-spank promoter and native KinA expressed from its own promoter, lanes and columns 13 and 14), and MF5759 (N516A, GFP-tagged KinA^N516A expressed from the Phy-spank promoter and native KinA expressed from its own promoter, lanes and columns 15 and 16).

Cells expressing both wild-type KinA and the autokinase-deficient mutant protomers in a 1:1 stoichiometric ratio exhibit a reduced Spo0A activity compared with those expressing only KinA.

To further examine the inhibitory effect of the expression of the KinA mutant form in a more quantitative manner, we constructed a strain coexpressing wild-type KinA and the autokinase-deficient protein independently from two different loci (amyE and thrC) under the control of the same IPTG-inducible promoter. In this system, each of the proteins is supposed to be expressed at approximately equal levels to each other in a cell in the presence of IPTG (20). As controls, we used three strains, each expressing either the wild-type protein, the phosphorylation deficient protein (KinA^H405A), or the ATP-binding-deficient protein (KinA^N516A) under the control of the IPTG-inducible promoter. We hypothesized that if autophosphorylation occurs in a trans fashion, the activity of KinA per cell would decrease when the protomers of the wild-type and the autokinase-deficient mutant (KinA^H405A or KinA^N516A) are expressed in a 1:1 stoichiometric ratio, leading to the formation of a set of inactive heterocomplexes. In other words, KinA specific activity would not change if autophosphorylation occurs in an intramolecular (cis) fashion because the cellular concentration of the wild-type protomer is constant regardless of the expression of the mutant form. When either the phosphorylation-deficient protein (KinA^H405A) or the ATP-binding-deficient protein (KinA^N516A) was expressed together with wild-type KinA at an IPTG concentration sufficient to induce sporulation under nutrient-rich conditions (10 μM IPTG in LB), β-galactosidase activities derived from PspoIIG-lacZ for Spo0A activity were significantly reduced compared with those in the single expression system of the wild-type KinA (Fig. 6). As a result, sporulation efficiencies in the strains exhibiting the reduced Spo0A activities dropped to below 5% of the wild-type level (Fig. 6). These inhibitory effects were modest compared with those in the overexpression of the mutant forms with the native KinA (Fig. 5). These results suggest that the nonfunctional KinA protein interferes with the autokinase activity of the wild-type KinA protein to a certain degree in a concentration-dependent manner and that autophosphorylation is inhibited when the nonfunctional KinA protomer is hybridized with the wild-type KinA protomer to create a heterotetramer. Based on these results, we speculate that most of the heterocomplexes formed in a cell might not be fully capable of transfering a phosphoryl group from one protomer to the other in a trans fashion.

FIG 6.

Cells expressing both wild-type KinA and the autokinase-deficient mutant protomers in a 1:1 stoichiometric ratio exhibit a reduced Spo0A activity compared with those expressing only KinA. Strains coexpressing wild-type KinA and the autokinase-deficient form of KinA, both under the control of the IPTG-inducible Phy-spank promoter, were constructed and examined the Spo0A activities and sporulation efficiencies. Strains used were MF5154 (KinA wild type, single expression [wt]), MF5564 (KinA^H405A, single expression [H405A]), MF5566 (KinA^N516A, single expression [N516A]), MF6897 (wt and KinA^H405A, coexpression [wt H405A]), and MF6898 (wt and KinA^N516A, coexpression [wt N516A]). Cells of each strain grown in LB medium to the mid-exponential phase (OD₆₀₀, 0.5) were induced for the production of the proteins by adding IPTG (10 μM). β-Galactosidase activities were measured at 3 h after induction. The means from three samples are shown with standard deviations. Numbers of spores (per milliliter of culture) were determined as described in Materials and Methods. The means from three analyses are indicated.

In vitro trans complementation of the autophosphorylation activity of the ATP-binding deficient mutant protein KinA^N516A by the phosphorylation-deficient mutant protein KinA^H405A.

Taking advantage of the fact that KinA forms a stable tetramer mediated by the N-terminal PAS-domain (Fig. 4) (20), we constructed a series of strains expressing both KinA^N516A and KinA^H405A under the control of the Phy-spank promoter in a stoichiometric ratio (1:1) in the same strain; we purified the resulting heterotetramers and examined their activities in vitro. As controls, homotetramer complexes of KinA, KinA^N516A, and KinA^H405A were processed in the same way.

For this, each of the His-tagged KinA^N516A and the GFP-tagged KinA^H405A constructs was placed under the control of the Phy-spank promoter and integrated at the nonessential amyE and thrC loci on the chromosomal DNA as a single copy, respectively. Because each gene for the mutant proteins is placed under the control of the same promoter but at different locations on the chromosomal DNA in the same cell, we anticipated that two different protomers would be expressed in a stoichiometric ratio (1:1). We also constructed a strain coexpressing GFP-tagged KinA and His-tagged KinA^H405A both under the control of the Phy-spank promoter. As further controls, three strains, each expressing either His-tagged KinA, KinA^N516A, or KinA^H405A, were constructed in the same way.

Protein expression was induced by adding IPTG to the engineered B. subtilis cell cultures to a final concentration of 0.1 mM when cell growth reached the exponential phase in LB medium, followed by purification using Ni-NTA–agarose column chromatography. Each of the purified proteins was separated by SDS-PAGE and processed for immunoblotting with anti-KinA antibodies to detect KinA and its mutant forms.

By inspecting the band intensities of the GFP- and His-tagged KinA mutants (Fig. 7A, lanes 4 and 5), we concluded that the ratio between two different protomers in the purified protein fraction is approximately 1:1. These results indicate that the GFP-tagged protomer stably forms a complex with the coexpressed His-tagged protein, and the resulting complex, which is supposed to be a tetramer (20), can be purified with the His-tagged protomer. In total, we purified (i) KinA_his6 (Fig. 7A, lane 1), (ii) KinA^H405A_his6 (Fig. 7A, lane 2), (iii) KinA^N516A_his6 (Fig. 7A, lane 3), (iv) KinA-GFP and KinA^H405A_his6 heterocomplex (Fig. 7A, lane 4), and (v) KinA^H405A-GFP and KinA^N516A_his6 heterocomplex (Fig. 7A, lane 5). For in vitro autophosphorylation, each of the purified proteins was incubated with [γ-³²P]ATP for 20 min at 30°C, and the reaction mixtures were analyzed via SDS-PAGE followed by autoradiography. In Fig. 7B, lane 5, the ³²P-labeled band was detected in a reaction with the purified heterocomplex composed of KinA^H405A-GFP and KinA^N516A-his6 in the presence of [γ-³²P]ATP. As negative controls, neither the purified KinA^H405A_his6 nor KinA^N516A_his6 was phosphorylated (Fig. 7B, lanes 2 and 3). In a positive-control reaction, the purified KinA_his6 was efficiently labeled in the presence of [γ-³²P]ATP (Fig. 7B, lane 1). In another control reaction with the purified heterocomplex composed of KinA-GFP and KinA^H405A_his6, we detected only a low-mobility band corresponding to KinA-GFP but not KinA^H405A_his6 (Fig. 7B, lane 4). Based on these results, we concluded that in the reaction with the purified heterocomplex composed of KinA^H405A-GFP and KinA^N516A_his6, the labeled band corresponds to KinA^N516A_his6 and KinA^H405A-GFP is not labeled with ³²P (Fig. 7B, lane 5). These results suggest that ATP binds to the ATP-binding domain of KinA^H405A in one protomer and the γ-phosphoryl group is transferred from ATP to H405 of KinA^N516A in another protomer within the tetramer in a trans manner. Furthermore, we added the purified Spo0F, an immediate downstream target for KinA, to detect the phosphotransfer reaction from the autophosphorylated KinA heterocomplex to Spo0F. As shown in Fig. 7C, lanes 1, 4, and 5, the labeled Spo0F was detected only in the reactions with the autophosphorylated kinases. These results indicated that when the kinase is autophosphorylated, the phosphotransfer reaction occurs efficiently. We note that with two different protomers, one GFP tagged and the other His tagged, there are five possible stoichiometric combinations of the complexes (Fig. 7D). Under our experimental conditions, the GFP-tagged protein can be purified only when the complex is formed with the His-tagged one. On this basis, among five different heterocomplexes, four complexes whose one protomer is His tagged can be obtained in the purified complex fraction (Fig. 7D). In the 2:2 heterocomplex composed of KinA^H405A-GFP and KinA^N516A_his6 (Fig. 7D), only two histidine residues (H405) in the KinA^N516A_his6 protomer are available for phosphorylation, while four sites are possibly phosphorylated in the wild-type KinA tetramer. In the other three hetero- and homocomplexes (Fig. 7D), less or no activity could be detected because the available number of the H405 residue or the ATP-binding site is only one or none. Therefore, each of the heterocomplexes might not be fully active. This might be the reason for the reduced level of phosphorylation in the reactions with the purified heterocomplex proteins (Fig. 7B, compare lanes 4 and 5 with lane 1), as well as the reduced sporulation efficiencies in the cells expressing the heterocomplex compared with those in the sporulating wild-type strain, as indicated in the following section (Table 1). Taken together, the evidence indicates that ATP initially binds to one or more protomers within the tetramer complex and then the γ-phosphoryl group is transmitted to another in a trans fashion.

FIG 7.

In vitro trans autophosphorylation of the ATP-binding-deficient mutant protein KinA^N516A by the phosphorylation-deficient mutant protein KinA^H405A. (A) Equal volumes of eluted protein fractions from Ni-NTA–agarose were subjected to SDS-PAGE (16%) followed by immunoblotting using anti-KinA antibodies. (B) In vitro autophosphorylation reactions were performed with the same samples as in panel A. (C) In vitro phosphotransfer reactions from the kinase to Spo0F were performed with the same samples as in panel A. The detailed procedures can be found in Materials and Methods. The gray and open arrowheads indicate the position of the GFP- and His-tagged proteins, respectively. (D) Possible tetramers formed by two different protomers expressed in the same cell (lane 4, KinA-GFP and KinA^H405A_his6; lane 5, KinA^H405A-GFP and KinA^N516A_his6) are listed. The molecular masses of each of the protomers and their complexes are indicated, in kilodaltons (size column). ND, not detectable under these experimental conditions. Strains used were MF5268 (KinA_his6, lane 1), MF5527 (KinA^H405A_his6, lane 2), MF5529 (KinA^N516A_his6, lane 3), MF5531 (KinA-GFP and KinA^H405A_his6, lane 4), and MF5756 (KinA^H405A-GFP and KinA^N516A_his6, lane 5).

TABLE 1.

Functional complementation assay results for KinA and derivatives

Protein(s)	No. of viable cells/ml	No. of spores/ml	Efficiencya
KinA	8.8 × 108	2.7 × 108	0.3
KinAH405A	3.6 × 108	<10	NDb
KinAN516A	4.9 × 108	<10	ND
KinAH405A + KinAN516A	7.7 × 108	3.1 × 103	4.0 × 10−6

^aSporulation efficiency was determined in 16-h culture as CFU per milliliter after heat treatment by incubation at 80°C for 10 min, compared with CFU of the pre-heat treatment sample.

^bND, not determined.

In vivo complementation of the autophosphorylation reaction between the ATP-binding-deficient mutant protein KinA^N516A and the phosphorylation-deficient mutant protein KinA^H405A.

To verify the in vitro results (Fig. 7), we performed in vivo complementation assays. We hypothesized that a strain expressing both KinA^H405A and KinA^N516A can complement the kinase activity and restore sporulation. To perform this, we constructed a series of strains that express different combinations of the wild-type and the mutant proteins of KinA under the control of the Phy-spank promoter (Table 1; see also Table S1 in the supplemental material). We note that as reported previously (18, 20), sporulation was triggered efficiently when wild-type KinA protein was expressed to the levels of the sporulating wild-type cells (lacking the IPTG-inducible construct but with the native promoter) using the IPTG-inducible expression system under nutrient-rich conditions. Using this system, we confirmed that neither of the single mutants underwent sporulation when expressed by adding IPTG (10 μM) (no spores were detected in 1 ml of culture), while the wild-type KinA control showed the normal sporulation efficiency (10⁸ spores were detected in 1 ml of culture), similar to the result obtained with the wild-type B. subtilis strain under normal sporulation conditions in SM medium (Table 1) (18, 20). Next, we introduced both the IPTG-inducible KinA^H405A and KinA^N516A constructs at the amyE and thrC loci, respectively, into the same strain and examined sporulation efficiency. When IPTG (10 μM) was added to express KinA^H405A and KinA^N516A simultaneously, fewer spores were detected than in the strain expressing wild-type KinA. However, these numbers (10³ in 1 ml of culture) were significantly higher than those in the strain expressing either KinA^H405A or KinA^N516A (Table 1). We verified that the engineered cells expressed similar levels of each of the mutant and wild-type proteins when IPTG was added as shown in Fig. 7A. As indicated in the above section, there are five possible stoichiometric combinations with two different protomers (Fig. 7D). All of them have fewer available sites for phosphorylation and ATP-binding than the wild-type KinA tetramer, resulting in the reduced autophosphorylation and therefore reduced levels of Spo0A∼P. Under such conditions, it is expected that an insufficient threshold level of Spo0A∼P does not fully trigger entry into sporulation (Table 1) (18, 26). Taken together, all these results indicate that the heterotetramer composed of KinA^H405A and KinA^N516A is, at least, active, although the levels were lower than those of the wild-type KinA homotetramer. Therefore, these in vivo results reinforce the evidence that KinA autophosphorylation is able to occur in a trans fashion.

DISCUSSION

In this study, we established both in vivo and in vitro systems to assess the mechanism of the autophosphorylation reaction of the major sporulation kinase KinA. Using in vivo systems in which pairs of the wild-type and the mutant forms of KinA are coexpressed in B. subtilis, dominant negative effects on sporulation are observed only when the heterotetramer is formed between the mutant and the wild-type proteins. Furthermore, the heterocomplex composed of two different mutant forms of KinA, one defective in phosphorylation with the alanine substitution at histidine 405 (H405A) and the other defective in ATP binding with the alanine substitution at asparagine 516 (N516A), was purified and shown to complement each other's defect of the autophosphorylation activity. Finally, sporulation is partially restored when these two mutant forms are coexpressed in the same cell, while expression of each of the single mutant proteins shows no effect on the restoration of sporulation. All these observations support the notion that autophosphorylation of KinA is able to occur in trans.

One of the most popular models for describing the mechanism of autokinase activation in response to environmental cues is that ligand binding induces a conformational change in the histidine kinase leading to the interaction between the γ-phosphoryl group of ATP bound at the CA domain and the phosphoryl-accepting His at the dimerization/histidine phosphorylation (DHp) domain to catalyze autophosphorylation, similar to the eukaryotic receptor kinase system. A crystallographic study suggests that the binding of a ligand to the sensor domain influences the spatial relationship of the catalytic domain to control autophosphorylation in HK853 of T. maritime (44). However, the active state for the autophosphorylation reaction has never been captured (29, 44).

To overcome these experimental difficulties, computational approaches in combination with structural information have been used to predict the important amino acid residues involved in the specificity and regulation of sensor histidine kinases. For example, one such study suggests that length and handedness of the α-helix bundle loops in the DHp domain are important for determining whether histidine kinases autophosphorylate in cis or in trans (45). Another such study suggests that a certain signal binding to the sensor domain causes helical unwinding in the C-terminal DHp and CA domains, resulting in the transition from the inactive to the active conformation (46). However, actual experimental data are required to directly support these assumptions.

While many of the sensor histidine kinases, including HK853, typically act as transmembrane receptors (3, 47), KinA is a cytosolic enzyme composed of three tandem PAS domains (PAS-ABC) at the N terminus and the catalytic DHp and CA domains at the C terminus (4, 8, 48). Thus, the means by which the cytosolic kinase responds to starvation conditions in the surrounding environment has been a long-standing question. Although no direct supporting evidence has been provided, many researchers are in favor of a receptor kinase model in which the N-terminal “sensor” domain receives a hypothetical ligand produced under starvation conditions, inducing a conformational change to produce the active form of KinA (4, 9, 10, 33). A recent report suggests that KinA is involved in biofilm formation in response to impaired respiration (49). Their model proposes that an oxidized form of NAD (NAD⁺) directly binds to the PAS-A domain of KinA and thereby inhibits the autokinase activity, though it is not clear how the binding of NAD⁺ to KinA influences its activity (49). Other reports indicate that the PAS-A domain is dispensable for its activity (20, 21, 50). To further clarify and extend these findings, we have recently shown that when the hypothetical sensor domain of KinA is replaced with an unrelated protein segment containing two PAS domains derived from E. coli protein YdaM, the resulting chimeric protein (YdaM_N-KinA_C) triggers a massive entry into sporulation in a manner dependent on its concentration regardless of culture conditions (19). Furthermore, we have demonstrated that the chimeric protein forms a tetramer, mediated by the N-terminal domain of YdaM, as was found for KinA, and the autophosphorylation activity of the chimeric kinase is independent of enzyme concentration, similar to the case with wild-type KinA, suggesting that the functional subunit assembly is not cooperative (19). These results suggest that the N-terminal domain of KinA does not necessarily recognize a hypothetical sporulation signal in order to activate the kinase. Instead, we have proposed that the N-terminal domain plays a role in maintenance of the tetramer structure, which is essential for autokinase activity that resides in the C-terminal catalytic domain (19–21). In support of this, both in vivo (20) and in vitro (22) studies indicate that the C-terminal catalytic domain of KinA by itself shows very little, if any, autokinase activity and instead accepts a phosphoryl group from Spo0F∼P in a reverse phosphotransfer reaction. Consistent with these results, the complex formation only with the C-terminal catalytic domain is not detectable in vivo (20). We note that the previous structural studies of KinA use the nonfunctional C-terminal domain of KinA, and thus, the predicted structure might not represent a fully functional form (51–53). Therefore, although the C-terminal DHp domain appears to be important and sufficient for the specific dimer assembly in many other histidine kinases (47, 54), this concept might not be applicable to KinA. Furthermore, we have shown, using a series of unique artificial genetic systems called artificial sporulation initiation (ASI) in combination with conventional genetic and biochemical approaches, that (i) KinA is constitutively active regardless of culture conditions (18), (ii) the concentration of KinA increases gradually over the course of starvation (55), and (iii) a threshold level of KinA is primarily important to trigger entry into sporulation (18). These results, together with the data showing that the PAS-A domain is dispensable for the autokinase activity (20, 21, 50), suggest that the activity of KinA is not directly regulated in response to an unknown sporulation signal(s). Instead, it appears that the cellular concentration of the constitutively active kinase is primarily important for cell fate decision, although the mechanism(s) to control the KinA protein level under starvation conditions is unclear. Nevertheless, the hypothetical starvation signal has not been identified, and thus, the long-standing model in which KinA acts as a receptor kinase is not necessarily proven to be true.

In sum, for further detailed characterization of the major sporulation kinase, we provide direct experimental evidence that autophosphorylation of KinA is able to occur in a trans fashion, although the biological significance of this is unclear. Our in vitro and in vivo experimental systems can provide real data to understand the reaction mechanism and kinetics of the sporulation kinase. While the receptor kinase model for the ligand-induced conformational change may be applicable to many other transmembrane histidine kinases, our results and insights described above suggest that KinA might be an exception. Therefore, current modeling approaches based on the nonfunctional C terminus crystal structure might be insufficient to predict the real features of the KinA activation mechanism.