In Vivo Effects of Sporulation Kinases on Mutant Spo0A Proteins in Bacillus subtilis

Abstract

The phosphorylated form of the response regulator Spo0A (Spo0A∼P) is required for the initiation of sporulation in Bacillus subtilis. Phosphate is transferred to Spo0A from at least four histidine kinases (KinA, KinB, KinC, and KinD) by a phosphotransfer pathway composed of Spo0F and Spo0B. Several mutations in spo0A allow initiation of sporulation in the absence of spo0F and spo0B, but the mechanisms by which these mutations allow bypass of spo0F and spo0B are not fully understood. We measured the ability of KinA, KinB, and KinC to activate sporulation of five spo0A mutants in the absence of Spo0F and Spo0B. We also determined the effect of Spo0E, a Spo0A∼P-specific phosphatase, on sporulation of strains containing the spo0A mutations. Our results indicate that several of the mutations relax the specificity of Spo0A, allowing Spo0A to obtain phosphate from a broader group of phosphodonors. In the course of these experiments, we observed medium-dependent effects on the sporulation of different mutants. This led us to identify a small molecule, acetoin, that can stimulate sporulation of some spo0A mutants.

In response to starvation and high cell density, Bacillus subtilis can initiate a developmental pathway leading to the formation of dormant endospores (3, 13, 46). The transcription factor Spo0A is a critical regulator of the shift from exponential growth to sporulation. Spo0A is activated by phosphorylation {reviewed in references 3 and 13), and sporulation begins with accumulation of a threshold amount of phosphorylated Spo0A (Spo0A∼P) (5).

Spo0A is a response regulator, but unlike most members of this protein family, Spo0A does not obtain phosphate directly from a histidine kinase. Spo0A phosphorylation is controlled by a phosphotransfer pathway, known as the phosphorelay, composed of Spo0F, Spo0B, and at least four histidine kinases, KinA, KinB, KinC, and KinD (1, 2, 20, 21, 24, 27, 36, 47). The kinases donate phosphate to Spo0F, a response regulator with no output domain (2, 36). The phosphate from Spo0F∼P is transferred to Spo0B and finally from Spo0B∼P to Spo0A (2). spo0B and spo0F null mutants do not sporulate, indicating that phosphorelay-independent phosphorylation of Spo0A is normally insufficient to activate sporulation.

A range of intra- and extracellular signals control sporulation by affecting the flow of phosphate through the phosphorelay. KinA, KinB, KinC, and KinD all have different roles in Spo0A activation and probably respond to distinct stimuli (7, 28, 47). KinA is the most important kinase for sporulation in rich sporulation media (28, 47). KinB is the most important kinase when cells are grown in glucose minimal medium (28). KinC and KinD contribute very little to sporulation under any condition but appear to regulate Spo0A during vegetative growth (20, 21, 24, 27).

Spo0A∼P levels are also regulated by phosphatases. Spo0E is a Spo0A∼P-specific phosphatase (33). Several phosphatases indirectly decrease Spo0A∼P levels by dephosphorylating Spo0F∼P. At least three members of the Rap (response regulator aspartyl-phosphate phosphatase) family of proteins dephosphorylate Spo0F∼P and are regulated by small peptides involved in cell-cell signaling (19, 26, 35, 37, 42).

Several spo0A mutations allow initiation of sporulation in the absence of spo0F and spo0B. These mutations have been isolated from many different screens and are named sof and sur0F (suppressors of spo0F), rvt (revertant), coi (control of initiation), sur0B (suppressor of spo0B), and sad (spo0A dominant) (18, 23, 34, 40, 41, 45). These mutations also bypass many of the regulatory signals that feed into the phosphorelay (15, 17, 18, 34, 40, 41, 45).

It is not fully understood how these spo0A mutations permit sporulation in the absence of spo0F and spo0B. Genetic analysis has shown that the spo0A(N12K) (sof-1) and spo0A(Q90R) (rvtA11) mutations alter Spo0A so that it depends on KinC to support sporulation without Spo0F or Spo0B (24, 27). This has generally been interpreted as evidence that altering residues on Spo0A can permit a better interaction with certain alternate sources of phosphate (4, 45), such as KinC (24, 27). Several lines of evidence suggest that even wild-type Spo0A can obtain a low level of phosphate directly from KinC in the absence of Spo0F (20, 27).

We performed a genetic screen to identify mutations that would allow Spo0A to obtain phosphate specifically from KinA, KinB, or KinC. Despite many differences between our screen and previous screens, all of the spo0A mutations that we isolated had been identified previously as suppressors of spo0F or spo0B null mutations. To assess the in vivo effects of the spo0A mutations on the interactions between Spo0A and various phosphodonors, we measured the ability of KinA, KinB, and KinC to activate sporulation of each spo0A mutant in the absence of spo0F and spo0B. Our results indicate that many of the mutations in spo0A relax specificity, allowing Spo0A to obtain phosphate from a broader range of phosphodonors. In the course of these experiments we observed medium-dependent effects on the sporulation of different mutants. This led us to identify a small molecule, acetoin, that can stimulate sporulation of strains carrying certain spo0A alleles.

MATERIALS AND METHODS

Media.

Routine growth and maintenance of B. subtilis and Escherichia coli was done in Luria-Bertani medium (32). Cells were grown and allowed to sporulate in two different nutrient sporulation media: DS medium (a nutrient broth) (39) and 2×SG (29) (2×SG contains twice the nutrient broth of DS medium plus 0.1% glucose). Acetoin (Fluka, Buchs, Switzerland) was melted in an 80°C water bath, cooled to room temperature, and mixed with water to make a 500 mM stock. The stock was made fresh for each use and was diluted into DS medium to a final concentration of 20 mM prior to inoculation. Acetoin did not affect the exponential growth of cells but did permit more residual growth during the transition to stationary phase. IPTG (isopropyl-β-d-galactopyranoside) was used at a concentration of 1 mM.

Strains.

The B. subtilis strains used (Table 1) are derivatives of JH642 (38) and contain the trpC2 and phe-1 mutations. Strains with spo0BΔPst are phe⁺. Transformations and strain constructions were performed by standard methods (12). Pspac-kinA (16), kinA::Tn917(mls) and kinA::Tn917(neo) (essentially the neo cassette replacing mls) (16), spo0BΔPst (48), spo0FΔS (23), and acoA::pLGSA700 (14) have been previously described. spo0A-spc was constructed similarly to the previously described spo0A-cat (10), and the spc cassette is approximately 90% linked to spo0A by transformation.

TABLE 1.

B. subtilis strains used

Strain	Genotype
BB1012	spo0FΔS spo0BΔPst kinA::Tn917(mls) ΔkinB::phl ΔkinC::kan spo0A(Q90R)-spc Δspo0E::tet Δpta::cat
JH642	Wild type (trpC2 pheA1)
JQ33	spo0FΔS spo0BΔPst ΔkinC::kan kinA::Tn917(neo) kinB::pJQ8 (Pspac-kinB-cat)
JQ38	spo0FΔS spo0BΔPst ΔkinC::kan kinA::Tn917(neo) kinB::pJQ8 (Pspac-kinB-cat) spo0A+-spc
JQ51	spo0FΔS spo0BΔPst kinA::Tn917(neo) ΔkinB::phl kinC::pJQ7 (Pspac-kinC-cat)
JQ53	spo0FΔS spo0BΔPst ΔkinC::kan ΔkinB::phl kinA::(Pspac-kinA-cat) spo0A+-spc
JQ61	spo0FΔS spo0BΔPst ΔkinC::kan ΔkinB::phl kinA::(Pspac-kinA-cat)
JQ62	spo0FΔS spo0BΔPst kinA::Tn917(neo) ΔkinB::phl kinC::pJQ7 (Pspac-kinC-cat) spo0A+-spc
JQ63	spo0FΔS spo0BΔPst kinA::Tn917(neo) ΔkinB::phl kinC::pJQ7 (Pspac-kinC-cat) spo0A+-spc Δspo0E::tet
JQ64	spo0FΔS spo0BΔPst ΔkinC::kan kinA::Tn917(neo) kinB::pJQ8 (Pspac-kinB-cat) spo0A+-spc Δspo0E::tet
JQ65	spo0FΔS spo0BΔPst ΔkinC::kan ΔkinB::phl kinA::(Pspac-kinA-cat) spo0A+-spcΔspo0E::tet
JQ82	spo0FΔS spo0BΔPst kinA::Tn917(neo) ΔkinB::phl kinC::pJQ7 (Pspac-kinC-cat) spo0A(Q90R)-spc
JQ95	spo0FΔS spo0BΔPst kinA::Tn917(neo) ΔkinB::phl kinC::pJQ7 (Pspac-kinC-cat) spo0A(Q90R)-spc Δspo0E::tet
JQ97	spo0FΔS spo0BΔPst kinA::Tn917(mls) ΔkinB::phl ΔkinC::kan
JQ134	spo0FΔS spo0BΔPst kinA::Tn917(mls) ΔkinB::phl ΔkinC::kan spo0A+-spc Δspo0E::tet
JQ137	spo0FΔS spo0BΔPst kinA::Tn917(mls) ΔkinB::phl ΔkinC::kan spo0A(Q90R)-spc Δspo0E::tet
JQ214	spo0FΔS spo0BΔPst kinA::Tn917(mls) ΔkinB::phl ΔkinC::kan spo0A(N12K)-spc
JQ219	spo0FΔS spo0BΔPst ΔkinC::kan ΔkinB::phl kinA::(Pspac-kinA-cat) spo0A(N12K)-spc
JQ220	spo0FΔS spo0BΔPst ΔkinC::kan kinA::Tn917(neo) kinB::pJQ8 (Pspac-kinB-cat) spo0A(N12K)-spc
JQ221	spo0FΔS spo0BΔPst kinA::Tn917(neo) ΔkinB::phl kinC::pJQ7 (Pspac-kinC-cat) spo0A(N12K)-spc
JQ225	spo0FΔS spo0BΔPst kinA::Tn917(mls) ΔkinB::phl ΔkinC::kan spo0A(N12K)-spc Δspo0E::tet
JQ227	spo0FΔS spo0BΔPst ΔkinC::kan ΔkinB::phl kinA::(Pspac-kinA-cat) spo0A(D92Y)-spc
JQ235	spo0FΔS spo0BΔPst ΔkinC::kan ΔkinB::phl kinA::(Pspac-kinA-cat) spo0A(E14A)-spc Δspo0E::tet
JQ236	spo0FΔS spo0BΔPst ΔkinC::kan ΔkinB::phl kinA::(Pspac-kinA-cat) spo0A(P60S)-spc
JQ263	spo0FΔS spo0BΔPst kinA::Tn917(neo) ΔkinB::phl kinC::pJQ7 (Pspac-kinC-cat) spo0A(N12K)-spc Δspo0E::tet
JQ363	spo0FΔS spo0BΔPst kinA::Tn917(mls) ΔkinB::phl ΔkinC::kan spo0A(E14A)-spc Δspo0E::tet
JQ367	spo0FΔS spo0BΔPst kinA::Tn917(mls) ΔkinB::phl ΔkinC::kan spo0A(E14A)-spc
JQ370	spo0FΔS spo0BΔPst kinA::Tn917(neo) ΔkinB::phl kinC::pJQ7 (Pspac-kinC-cat) spo0A(E14A)-spc Δspo0E::tet
JQ372	spo0FΔS spo0BΔPst ΔkinC::kan kinA::Tn917(neo) kinB::pJQ8 (Pspac-kinB-cat) spo0A(E14A)-spc Δspo0E::tet
JQ387	spo0FΔS spo0BΔPst kinA::Tn917(mls) ΔkinB::phl ΔkinC::kan spo0A(D92Y)-spc
JQ388	spo0FΔS spo0BΔPst kinA::Tn917(mls) ΔkinB::phl ΔkinC::kan spo0A(D92Y)-spc Δspo0E::tet
JQ395	spo0FΔS spo0BΔPst kinA::Tn917(mls) ΔkinB::phl ΔkinC::kan spo0A(P60S)-spc
JQ396	spo0FΔS spo0BΔPst kinA::Tn917(mls) ΔkinB::phl ΔkinC::kan spo0A(P60S)-spc Δspo0E::tet
JQ397	spo0FΔS spo0BΔPst kinA::Tn917(neo) ΔkinB::phl kinC::pJQ7 (Pspac-kinC-cat) spo0A(E14A)-spc
JQ532	spo0FΔS spo0BΔPst kinA::Tn917(mls) ΔkinB::phl ΔkinC::kan spo0A(Q90R)-spc Δspo0E::tet acoA::pLGSA700 (cat)
JQ536	spo0FΔS spo0BΔPst ΔkinC::kan ΔkinB::phl kinA::(Pspac-kinA-cat) spo0A(Q90R)-spc
JQ537	spo0FΔS spo0BΔPst ΔkinC::kan kinA::Tn917(neo) kinB::pJQ8 (Pspac-kinB-cat) spo0A(Q90R)-spc
JQ538	spo0FΔS spo0BΔPst ΔkinC::kan kinA::Tn917(neo) kinB::pJQ8 (Pspac-kinB-cat) spo0A(P60S)-spc
JQ539	spo0FΔS spo0BΔPst kinA::Tn917(neo) ΔkinB::phl kinC::pJQ7 (Pspac-kinC-cat) spo0A(P60S)-spc
JQ540	spo0FΔS spo0BΔPst ΔkinC::kan kinA::Tn917(neo) kinB::pJQ8 (Pspac-kinB-cat) spo0A(D92Y)-spc
JQ541	spo0FΔS spo0BΔPst kinA::Tn917(neo) ΔkinB::phl kinC::pJQ7 (Pspac-kinC-cat) spo0A(D92Y)-spc

pJQ4, carrying the ΔkinB::phl deletion-insertion, was constructed as follows. pNY113 (28), carrying the kinB-kapB operon, was digested with SacII and ClaI to remove the kinase domain of kinB. The phl cassette was cut from pUC18-ble-1 (50) with HindIII and EcoRI. Both the digested pNY113 and the phl cassette were blunted and ligated to give pJQ4.

kinC deletion-insertions were constructed using the pLK24 (27) backbone. pLK24, containing kinC, was digested with PstI and BglII to remove the kinase domain of kinC and ligated with either a kan cassette to make pJQ2 (ΔkinC::kan) or a phl cassette to make pJQ3 (ΔkinC::phl). The kan cassette was obtained by PstI-BglII digestion of pDG792 (11) and the phl cassette was obtained by HindIII-EcoRI digestion of pUC18-ble-1.

pJQ7 (Pspac-kinC) and pJQ8 (Pspac-kinB) were constructed by amplifying a portion of the genes (+3 to +357 for kinC; −1 to +306 for kinB [+1 indicating the transcription start site]) and cloning into pDH88 (49), to place the gene fragments downstream of the LacI-repressible IPTG-inducible promoter Pspac. These plasmids were inserted by single crossover into the chromosome, disrupting the native locus and placing the intact kinase gene under the control of Pspac.

Sporulation assays.

Cells were grown in DS or 2×SG medium at 37°C. Spores were assayed approximately 20 to 25 h after the end of exponential growth. Viable cells and spores were counted by plating before and after a heat treatment (80°C for 20 min) as previously described (27). Percent sporulation is calculated as 100 times the number of spores per milliliter divided by the number of viable cells per milliliter. Data presented are from representative experiments. Similar results were obtained in at least two and usually three or more independent experiments.

Mutagenesis.

Six different strains, JQ33, JQ51, JQ61, JQ63, JQ64, and JQ65 (Table 1; see also Results), were mutagenized. Each contained loss-of-function mutations in spo0F, spo0B, and two of the three kinase genes (kinA, kinB, and kinC). The third kinase gene was fused to the LacI-repressible IPTG-inducible promoter Pspac. Three strains were spo0E⁺ and three were spo0E⁻. For mutagenesis, strains were grown to mid-exponential phase in Luria-Bertani medium, pelleted by brief centrifugation, and resuspended in 10 mM (NH₄)₂SO₄–5 mM K₂PO₄–50 mM MOPS (morpholinepropanesulfonic acid)-KOH, pH 7. Ethyl methanesulfonate (Sigma) was added (1%), and cells were incubated for 30 min at 37°C. Mutagenized cells were collected by centrifugation, washed in DS medium, inoculated into DS medium with 1 mM IPTG, and grown until 20 h after the end of exponential growth, at which point cells capable of sporulating should have done so. Cells were killed by heating at 80°C for 20 min, and the surviving spores were plated and allowed to grow on plates containing DS medium plus 1 mM IPTG. The unmutagenized parent strains sporulate extremely poorly. Mutations were back-crossed to determine linkage to spo0A (spc linked) or Pspac (cat linked).

Pspac(hy) mutation.

Mutations that were not linked to spo0A were tightly linked to the chloramphenicol resistance gene linked to Pspac-kinC. The integrated plasmid (containing Pspac-kinC) and nearby portions of kinC were rescued from the mutant and its parent by digesting total DNA from the mutant strains (using SphI), ligating, and transforming into E. coli, selecting for Amp^r. Plasmid isolates were sequenced, and the mutation was identified as a G-to-T change at the −1 position of Pspac relative to the transcription start site (mapped in reference 49). This mutation, called Pspac(hy), raised the level of expression 10- to 20-fold over that from Pspac without affecting the fold induction by IPTG. As previously shown, overexpression of kinC permits sporulation in the absence of the phosphorelay (27). Pspac(hy) is a useful promoter for overexpression of a variety of genes of interest. Plasmids have been constructed for gene expression from Pspac(hy) either as a single crossover (pJQ43) or as a double crossover at the amyE locus (pPL82). Another mutation (A to G at +1) in Pspac that increases expression in the absence of Lac repressor has been described (22).

RESULTS AND DISCUSSION

Phosphorelay bypass mutations are identical to previously isolated mutations.

We performed a genetic screen to identify mutations that would allow Spo0A to accept phosphate from KinA, KinB, or KinC in the absence of the phosphotransfer proteins, Spo0F and Spo0B. We constructed strains lacking spo0F, spo0B, and two of the three kinases. The remaining kinase was expressed from the LacI-repressible IPTG-inducible promoter, Pspac. The Pspac kinase expression constructs were all functional, since each could complement a null mutation of the corresponding kinase (data not shown). The strains expressing one kinase from Pspac and lacking the other two kinases, and spo0F and spo0B, were unable to sporulate. These strains (JQ33, JQ51, and JQ61) were mutagenized with ethyl methanesulfonate and then grown in DS medium supplemented with IPTG. Spores were selected by a heat treatment that killed nonsporulated cells. This procedure strongly selected for mutations that permitted sporulation in the absence of spo0F and spo0B. To identify weaker alleles, the screen was also performed in strains (JQ63, JQ64, and JQ65) null for spo0E, the gene coding for a phosphatase that can dephosphorylate Spo0A∼P. This did not change the types of mutations isolated.

All of the mutations characterized were linked to spo0A, except for one which was in Pspac (see Materials and Methods). The spo0A gene from mutant cells was amplified by PCR, cloned, and sequenced to identify the mutations. All of the isolates carried mutations that were identical to those previously isolated as sof, coi, or rvt mutations (Table 2). Although we independently obtained some mutations 10 or more times, other previously identified mutations were never isolated. This selectivity has been observed previously and may depend upon the mutagen used (45). Nonetheless, there is overlap in the alleles isolated from different screens. This overlap suggests that there are very few single mutations that allow Spo0A to stimulate sporulation in the absence of spo0F and/or spo0B. We examined the ability of the five different spo0A alleles to support sporulation in the presence of KinA, KinB, or KinC in vivo.

TABLE 2.

spo0A mutations used

Allele names	Amino acid change	Reference(s)
spo0AN12K; sof1; sof103; sur0F1	Asparagine 12 to lysine	4, 23, 41, 45
spo0AE14A; sof3; sof115	Glutamate 14 to alanine	4, 45
spo0AP60S; sof4; sof118; coi-1	Proline 60 to serine	4, 34, 45
spo0AQ90R; rvtA11	Glutamine 90 to arginine	40
spo0AD92Y; sof5; sof114	Aspartate 92 to tyrosine	4, 45

Cells with spo0A mutations have various sensitivities to the sporulation kinases.

The abilities of kinA, kinB, and kinC, expressed from Pspac, to stimulate sporulation of the various spo0A mutants was assessed in a Δspo0F Δspo0B background (Table 3). Wild-type Spo0A was not sufficiently activated by any of the kinases to allow sporulation. spo0A(N12K) and spo0A(Q90R) could be activated only by KinC, as has been described before when the kinases were expressed from their native promoters (24, 27). Sporulation of the spo0A(N12K) and spo0A(Q90R) strains was lower than reported previously because we used DS medium instead of 2×SG as the sporulation medium. This medium effect is discussed below. spo0A(E14A) could also be activated by KinC (Table 3).

TABLE 3.

Effects of kinA, kinB, and kinC on sporulation of spo0A mutants

spo0A allelea	% Sporulationb
	Pspac-kinA ΔkinBC	Pspac-kinB ΔkinAC	Pspac-kinC ΔkinAB	ΔkinABC
spo0A+	<1.3 × 10−6	<2.9 × 10−6	1.0 × 10−6	<1.9 × 10−6
spo0A(N12K)	3.6 × 10−6	<5.0 × 10−6	1.9 × 10−3	<5.0 × 10−6
spo0A(E14A)	ND	ND	1.5 × 10−3	9.4 × 10−5
spo0A(P60S)	3.1	2.2	4.4	4.7
spo0A(Q90R)	1.6 × 10−6	1.8 × 10−5	1.4 × 10−2	ND
spo0A(D92Y)	0.10	0.20	0.36	0.13

^aAll strains are Δspo0F Δspo0B in addition to the indicated spo0A and kinase alleles. 

^bStrains were grown and allowed to sporulate in DS medium with 1 mM IPTG to induce expression from Pspac. Wild-type (JH642) cultures produced 50 to 80% spores. ND, not determined. 

spo0A(P60S) and spo0A(D92Y) permitted a moderate level of sporulation in the absence of the three kinases and the phosphorelay (Table 3). There was no further stimulation of sporulation in the presence of any of the three kinases. These data indicate that the spo0A(P60S) and spo0A(D92Y) mutations do not improve the ability of Spo0A to accept phosphate directly from KinA, KinB, or KinC, consistent with in vitro results (4). The spo0A(P60S) and spo0A(D92Y) gene products might obtain phosphate from another source. Alternatively, in the absence of the phosphorelay, they probably stimulate sporulation independently of phosphorylation (see below).

Cells with spo0A mutations exhibit differing sensitivities to the phosphatase Spo0E.

Spo0E negatively regulates sporulation by directly dephosphorylating Spo0A∼P (33). We tested the sensitivity of the different spo0A mutations to Spo0E in vivo. If spo0E has an effect on sporulation, then we can infer that mutant Spo0A is phosphorylated in vivo. If spo0E has no effect, then either the mutant Spo0A is not effectively phosphorylated in vivo or it is resistant to dephosphorylation by Spo0E.

We found that wild-type spo0A and the spo0A mutants (N12K, Q90R, and E14A) that were responsive to one or more sporulation kinases (Table 3) were also sensitive to Spo0E (Table 4). Pspac-kinC was used to drive Spo0A phosphorylation in the absence of kinA, kinB, spo0F, and spo0B. Wild-type spo0A did not allow significant sporulation in the presence of spo0E⁺, but in the Δspo0E background it allowed a low level of sporulation (Table 4). Even in the absence of spo0E, wild-type Spo0A could not obtain sufficient phosphate from KinA or KinB to stimulate sporulation (data not shown). These results indicate that wild-type Spo0A can obtain a low level of phosphate from KinC and that Spo0E normally acts to remove this phosphate.

TABLE 4.

Effects of spo0E on sporulation of spo0A mutants

Strainsa (spo0E+/Δspo0E)	spo0A allele	Other genotype	% Sporulationb
			spo0E+	Δspo0E
JQ62/JQ63	spo0A+	Pspac-kinC	1.0 × 10−6	3.0 × 10−4
JQ221/JQ263	spo0A(N12K)	Pspac-kinC	2.0 × 10−3	44
JQ397/JQ370	spo0A(E14A)	Pspac-kinC	1.5 × 10−3c	64
JQ367/JQ363	spo0A(E14A)	ΔkinC	9.4 × 10−5c	0.78
JQ395/JQ396	spo0A(P60S)	ΔkinC	2.4	1.6
JQ82/JQ95	spo0A(Q90R)	Pspac-kinC	7.6 × 10−2	46
JQ387/JQ388	spo0A(D92Y)	ΔkinC	0.33	0.24

^aAll strains are ΔkinA ΔkinB Δspo0F Δspo0B. Strains carrying spo0A(P60S) and spo0A(D92Y) were unaffected by kinC, and therefore the assay was done in a ΔkinC background. Pspac-kinC was induced with 1 mM IPTG added when the culture was inoculated. 

^bStrains were grown in DS medium. Wild-type (JH642) cultures produced 50 to 80% spores. 

^cData are from the experiment shown in Table 3. 

The absence of spo0E caused increased sporulation of spo0A(N12K), spo0A(E14A), and spo0A(Q90R) strains (Table 4), indicating that these altered forms of the Spo0A protein need to be phosphorylated to have significant activity in vivo and that they are substrates for the Spo0E phosphatase in vivo. These in vivo findings are consistent with results indicating that the phosphorylated forms of Spo0A(N12K) and Spo0A(E14A) are dephosphorylated by Spo0E in vitro as efficiently as wild-type Spo0A (4). Even though these mutant proteins are substrates for Spo0E, the spo0A(N12K) and spo0A(E14A) alleles are known to suppress the sporulation defect caused by the hyperactive spo0E mutation, spo0E11 (34, 41, 45). Taken together, these results indicate that these Spo0A mutants can receive phosphate from sources other than Spo0B∼P, overwhelming the hyperactive phosphatase produced in a spo0E11 strain.

spo0A(E14A) can moderately activate sporulation in the absence of KinA, KinB, or KinC, but this residual sporulation was virtually abolished in the presence of spo0E⁺ (Table 4). Therefore, the residual sporulation was due to phosphorylation of Spo0A(E14A) by another, unidentified phosphodonor. Given that spo0A(E14A) exhibits relaxed specificity for phosphodonors in the absence of spo0E, we asked if sporulation of a spo0A(E14A) Δspo0E strain could be stimulated by KinA or KinB and found that KinA significantly stimulated sporulation (Table 5).

TABLE 5.

Effects of kinA, kinB, and kinC on sporulation of the spo0A(E14A) Δspo0E mutant

Straina	Other genotypeb	% Sporulationc
JQ363	ΔkinABC	0.33
JQ235	Pspac-kinA ΔkinBC	35
JQ372	Pspac-kinB ΔkinAC	0.09
JQ370	Pspac-kinC ΔkinAB	47

^aAll strains are spo0A(E14A) Δspo0F Δspo0B Δspo0E. 

^bExpression of kinases from the Pspac promoter was induced with 1 mM IPTG added when the culture was inoculated. 

^cStrains were grown in DS medium. Wild-type (JH642) cultures produced 50 to 80% spores. 

In contrast to mutants that were affected by one or another of the kinases and spo0E, sporulation of the spo0A(P60S) and spo0A(D92Y) strains was not significantly affected by spo0E (Table 4). This indicates that either these mutant spo0A gene products are not significantly phosphorylated in vivo in the absence of the phosphorelay or that they are resistant to dephosphorylation by Spo0E. Previous work showed that Spo0A(D92Y) is efficiently dephosphorylated by Spo0E in vitro (4). Therefore, we favor the interpretation that Spo0A(D92Y) and probably Spo0A(P60S) function independently of phosphorylation. They could do this by stabilizing interactions between Spo0A and RNA polymerase, as has been observed in vitro (4). Though these two mutant proteins appear to function in the absence of phosphorylation, they clearly function better when phosphorylated. Phosphorylation in vitro stimulates their ability to activate transcription (4). In vivo, these mutant proteins allow higher levels of sporulation in the presence of the phosphorelay than in its absence, indicating that activity in vivo is also stimulated by phosphorylation (45). Together, the in vivo and in vitro results are most consistent with the interpretation that, in vivo, these mutant forms of Spo0A do not obtain phosphate from sources other than the phosphorelay.

These experiments divide the spo0A alleles into two groups. The first group of spo0A alleles, spo0A(N12K), spo0A(E14A), and spo0A(Q90R), code for proteins that can support sporulation in the absence of spo0F and spo0B by accepting phosphate directly from KinC. As seen in the absence of spo0E, spo0A(E14A) can also weakly accept phosphate from KinA and at least one unidentified phosphodonor. Sporulation of these spo0A mutant strains is highly sensitive to the phosphatase Spo0E. Consistent with these mutants having a relaxed specificity for alternative phosphodonors, in vitro data indicate that Spo0A(N12K) and Spo0A(E14A) can be phosphorylated directly by KinA more efficiently than wild-type Spo0A (4, 20). The second group of spo0A alleles, spo0A(P60S) and spo0A(D92Y), appear to support a modest level of sporulation in the absence of the phosphorelay without being phosphorylated.

Phosphorylation-dependent spo0A mutants are differentially stimulated by various carbon sources.

In the course of these experiments, we observed that cells carrying spo0A(N12K), spo0A(Q90R), and spo0A(E14A) all sporulated substantially better in 2×SG medium than in DS medium (data not shown). The difference in sporulation was caused by the 0.1% glucose in 2×SG. A spo0A(Q90R) strain sporulated as well in DS medium plus 0.1% glucose as in 2×SG but did not sporulate well in DS medium or DS medium with twice the nutrient broth (data not shown). Similar results were seen with spo0A(N12K) and spo0A(E14A). We found that other carbon sources (fructose, glycerol, and ribose) that are metabolized to pyruvate through part of the glycolytic pathway also stimulated sporulation. Pyruvate itself had little effect on sporulation, perhaps because it is poorly taken up (44). Malate, which is metabolized through the citric acid cycle, did not affect sporulation. It is important to note that high levels of some sugars (>0.5%) inhibit sporulation. The activating effect described here occurs with low levels (0.1%) of sugars.

A metabolic by-product, acetoin, differentially activates Spo0A mutants.

We reasoned that various carbon sources might affect sporulation by stimulating production of a metabolic by-product. B. subtilis growing on DS medium plus 0.1% glucose secretes primarily pyruvate, acetate, acetoin, isovalerate, and isobutyrate into the medium (44). Upon entering stationary phase, the citric acid cycle is activated and acetate and acetoin are rapidly consumed (43, 44).

Acetoin substantially affected sporulation of some spo0A mutants (Table 6). Acetate and pyruvate had little effect on sporulation, and isovalerate and isobutyrate had no effect (data not shown). When supplemented with exogenous acetoin, B. subtilis cultures exhibit increased residual growth after entry into stationary phase (in fact, the growth on acetoin might more properly be called biphasic) (14). Sporulation of spo0A⁺, spo0A(P60S), and spo0A(D92Y) strains was slightly inhibited by acetoin, perhaps due to increased growth. In contrast, sporulation of spo0A(N12K), spo0A(E14A), and spo0A(Q90R) mutants was stimulated by the addition of acetoin to the medium (Table 6). The fact that acetoin affected some spo0A mutants differently from others indicates that it might be contributing to activation of Spo0A. Any effect on downstream sporulation events or transcription and translation of spo0A should affect all the spo0A mutants similarly. Acetoin probably stimulates sporulation by increasing phosphorylation of Spo0A. Acetoin stimulated sporulation of all the spo0A mutants with relaxed specificity for KinC [spo0A(N12K), spo0A(E14A), and spo0A(Q90R)] and did not affect the mutants with normal specificity for the upstream kinases [spo0A(P60S) and spo0A(D92Y)]. Additionally, the acetoin effect on spo0A(Q90R) was decreased in the presence of the phosphatase, Spo0E (data not shown), further supporting the notion that phosphorylation of the mutant Spo0A proteins was affected.

TABLE 6.

Acetoin affects sporulation of spo0A mutants

Straina	spo0A allele	% sporulation
		DS medium	DS medium + 20 mM acetoin
JQ134	Wild type	3.3 × 10−6	2.9 × 10−6
JQ225	spo0A(N12K)	1.6 × 10−5	0.5
JQ363	spo0A(E14A)	0.33	11
JQ396	spo0A(P60S)	1.3	0.25
JQ137	spo0A(Q90R)	2.4 × 10−5	1.5
JQ388	spo0A(D92Y)	0.37	0.048

^aAll strains are ΔkinA ΔkinB ΔkinC Δspo0B Δspo0F Δspo0E in addition to the indicated spo0A allele. 

Acetoin must be metabolized to affect sporulation.

Acetoin metabolism was required for acetoin to stimulate sporulation. A set of proteins encoded by the aco operon metabolizes acetoin to acetyl coenzyme A (acetyl-CoA) and acetaldehyde (14). Acetoin did not stimulate sporulation of a strain containing the acoA and spo0A(Q90R) mutations (Table 7). Strains containing a disruption of acoA, the first gene in the aco operon, no longer showed growth stimulation by acetoin (data not shown). The acoA mutation is probably polar on the entire operon (14).

TABLE 7.

Acetoin metabolism affects sporulation

Straina	Relevant genotype	% sporulation in DS mediumb
		Alone	+ 0.1% glucose	+ 20 mM acetoin
JQ137	spo0A(Q90R)	4.9 × 10−5	4.1 × 10−3	3.2
JQ532	spo0A(Q90R) acoA:: pLGSA700	7.5 × 10−5	3.1 × 10−3	6.5 × 10−5

^aBoth strains are ΔkinA ΔkinB ΔkinC Δspo0F Δspo0B Δspo0E. 

^bStrains were grown in the indicated medium. Spores were assayed approximately 20 h after the end of exponential growth. 

It is unclear how the metabolism of acetoin stimulates sporulation. One possibility is that cells metabolizing acetoin have increased levels of acetyl-phosphate, which is known to donate phosphate to some response regulators (8, 30, 31). Our data do not support this model, however. Both acetyl-CoA and acetaldehyde are readily converted into acetyl-phosphate in E. coli and probably in B. subtilis as well (6, 9). The conversion of acetyl-CoA to acetyl phosphate requires the product of the pta gene. Thus, deletion of pta might attenuate the acetoin effect if the effect depends on the conversion of acetoin to acetyl phosphate. However the effect of acetoin on the sporulation of JQ137 (Tables 6 and 7) is unaffected by deleting pta (strain BB1012) (data not shown).

Altered forms of Spo0A have decreased discrimination among phosphate donors.

Our results indicate that the spo0A(N12K), spo0A(E14A), and spo0A(Q90R) mutations cause the production of altered forms of Spo0A that can activate sporulation in response to phosphorylation from a broader range of possible phosphodonors than wild-type Spo0A. All of the mutants and wild-type Spo0A can support strong sporulation when activated by Spo0B∼P (the phosphorelay) (27, 34, 45). In addition, KinC allows weak sporulation of a spo0A⁺ strain in the absence of the phosphatase Spo0E. spo0A(N12K)-, spo0A(E14A)-, and spo0A(Q90R)-containing strains sporulate well in the presence of KinC and are also stimulated by acetoin. These data suggest that Spo0A may have evolved to be generally stringent so as to prevent all but the most efficient phosphodonor, Spo0B∼P, from stimulating sporulation. Mutations such as spo0A(N12K), spo0A(Q90R), and spo0A(E14A) may alter Spo0A so as to relax the general stringency of Spo0A activation. They may allow a more efficient interaction between Spo0A and various phosphodonors, or they may increase the stability of Spo0A∼P. Increased stability would allow lower levels of phosphorylation from alternative phosphodonors to activate sporulation despite the poor efficiency of the interaction. These mechanisms are not mutually exclusive.

Inappropriate activation of Spo0A impairs cell growth and can be lethal (18, 45). B. subtilis contains genes for 34 putative response regulators and 36 putative histidine kinases (7, 25). Residues that help reject interactions with nonspecific kinases may be important in preventing inappropriate activation of Spo0A and other response regulators.