Phosphorylation and functional analysis of the sporulation initiation factor Spo0A from Clostridium botulinum


*E-mail; Tel. (+1) 858 784 7905; Fax (+1) 858 784 7966.


The initiation of sporulation in aerobic Bacillus species is regulated by the phosphorelay consisting of several sensor histidine kinases, the Spo0F response regulator, the Spo0B phosphotransferase and the Spo0A transcription factor that upon phosphorylation represses genes for growth and activates the developmental process. Clostridium species lack both Spo0F and Spo0B and the identities of the sensor histidine kinases are unknown. The amino acid sequence of Spo0A is highly conserved in Clostridium botulinum relative to Bacillus subtilis but the cloned C. botulinum Spo0A was unable to complement a spo0A mutant of B. subtilis for sporulation. However, it was able to repress the abrB gene of B. subtilis. Active site mutations in Spo0A still repressed, indicating this activity was independent of phosphorylation. An orphan sensor histidine kinase of C. botulinum appeared to normally phosphorylate C. botulinum Spo0A and expression of this kinase in combination with C. botulinum Spo0A in B. subtilis was lethal, suggesting phosphorylation of C. botulinum Spo0A repressed essential growth genes as a prerequisite to sporulation but could not compensate for this effect by inducing sporulation. A chimera Spo0A consisting of a B. subtilis Spo0A response regulator domain fused to a C. botulinum DNA-binding domain was capable of restoring sporulation to a spo0A mutant of B. subtilis albeit at less than wild-type levels. The data suggest that induction of sporulation requires interactions of both domains of Spo0A with other conserved proteins and despite the high conservation of the amino acid sequence of C. botulinum Spo0A, some of these interactions have been lost.


The initiation of sporulation is one of the decisive moments in the life cycle of sporulating bacteria. In the aerobic Bacilli, a complex system of controls exists that ensures the orderly transition from the dividing vegetative cell to a sporulating cell. Much of this regulation revolves around the Spo0A transcription factor which activates the transcription of crucial genes for the sporulation process in addition to its other functions in the regulation of stationary phase genes. In all, several hundred genes are known to be under the control of Spo0A either directly or indirectly by its control of other regulators such as AbrB (Liu et al., 2003; Molle et al., 2003).

The transcription functions of Spo0A are activated by phosphorylation of the N-terminal response regulator domain of this two domain transcription factor. The C-terminal domain binds to promoter DNA regions containing the Spo0A box ‘TGA/TCGAA’ (Zhao et al., 2002). The signal transduction pathway that leads to phosphorylation of Spo0A is the phosphorelay consisting of several sensor histidine kinases, the single domain response regulator Spo0F and a phosphotransferase Spo0B (Fig. 1) (Burbulys et al., 1991). Signals of unknown composition interact with the sensor histidine kinases activating an ATP-dependent autophosphorylation reaction on a histidine residue. The sensor histidine kinases target the Spo0F protein to which the phosphoryl group is transferred producing an aspartyl-phosphate. This phosphoryl group is transferred to a histidine residue of Spo0B and subsequently to an aspartate residue of Spo0A. This phosphoryl transfer pathway is subject to a variety of cellular and metabolic controls mediated by phosphatases that modulate the phosphorylation level to ensure the initiation of sporulation does not occur at inopportune times (Perego and Hoch, 2002). The genes for the basic pathway and many of its controlling elements are conserved in the genomes of all aerobic sporulating Bacilli suggesting that the pathway is essentially similar in all.

Figure 1.

The phosphorelay in sporulation initiation. Sporulation in B. subtilis is controlled and initiated by the phosphorelay. No Spo0B or Spo0F orthologues were found in C. botulinum, suggesting Spo0A receives its phosphate directly from a sensor histidine kinase. S: sensor domain; HKA: histidine kinase domain; ATP: ATP-binding domain; RR: response regulator domain; DNA: DNA-binding domain.

The situation in anaerobic sporulating bacteria, however, is different. The gene for the Spo0A transcription factor is readily apparent in all of the sequenced genomes of Clostridia and other anaerobic sporeformers due to the extremely high conservation of its amino acid sequence (Brown et al., 1994). However, genes for both Spo0F and Spo0B, if they exist, were not readily found in the genomes of Clostridium acetylbutylicum, Clostridium difficile, Clostridium perfringens, Clostridium thermocellum and Clostridium botulinum (Stephenson and Hoch, 2002; Stragier, 2002). This raises several questions of how the bacteria of the Clostridium species initiate and regulate sporulation because one of the key regulatory sites is Spo0F phosphorylation/dephosphorylation in Bacillus subtilis and probably other Bacillus species. The simple explanation that Spo0A is directly phosphorylated by sensor histidine kinases without the intermediate control points afforded by Spo0F and Spo0B lacks proof that such a reaction by sensor histidine kinases exists (Fig. 1). In order to examine this possibility in more details, a bioinformatic analysis of two-component systems in C. botulinum to identify probable sporulation sensor histidine kinases was carried out. Cloning and functional analyses of genes for C. botulinum Spo0A and sensor histidine kinases were accomplished and their relationship to the phosphorylation of Spo0A was examined.


Identification of two-component systems in C. botulinum

Bioinformatic analyses were undertaken of the genomic sequence of C. botulinum searching for genes coding for sensor histidine kinases and response regulators. Genes coding for proteins showing homologies to either kinases or response regulators were identified by blast analysis using the amino acid sequences of conserved domains of two-component sensor histidine kinases and response regulators. The histidine kinase domain, HisKA (starting 15 amino acids N-terminal to the histidine), and the ATP binding domain of Kinase A of B. subtilis were used as a probe to identify all possible sensor histidine kinases. Response regulator genes were identified using Spo0F of B. subtilis as a probe.

The amino acid sequences of candidates for the histidine kinases and response regulators were examined for the existence of the conserved active site residues. This analysis found 35 kinases and 41 response regulators (Table 1). Those located in immediate proximity on the genome, presumably comprising a single transcription unit, were assumed to form two-component systems leaving the rest as orphans.

Table 1.  Two-component system homologues in C. botulinum.
ClassFamilyHistidine kinaseClosest homologue in B. subtilisaPutative trans-membrane domainsResponse regulatorClosest homologue in B. subtilisaGene organizationb
  • a. 

    e-values are from blast analysis of the C. botulinum genes against B. subtilis proteins at the SUBTILIST website.

  • b. 

    H, histidine kinase gene; R, response regulator gene. Transcription occurs from left to right.

CBO1922ResE(7e- 33)4CBO1921YcbL(1e-51)RH
None CBO0506 Orphan
CBO0509 Orphan
CBO0998 Orphan
CBO1924 Orphan

All of the kinases and response regulators were sorted into classes and families. The classes were determined by the homology around the phosphorylatable histidine in the kinases (Table 2), whereas the family refers to homologies within the DNA binding domain of the response regulators. In addition the amino acid sequences of the kinases and response regulators were used as probes to blast search the B. subtilis genome for homologues (Table 1). Where possible, the putative orthologues were confirmed by looking at homologies among the surrounding genes.

Table 2.  Classification of kinases by the sequence around the phosphorylatable histidine.aThumbnail image of

The histidine kinases could be divided into five families (Table 2): I, II, IIIA, IV and V (Fabret et al., 1999). This was obtained by blast analysis using the histidine kinase domain (starting 15 amino acids N-terminal and ending 45 amino acids C-terminal of the active site histidine residue) of YesM (class I), DegS (class II), PhoR (class IIIA), CitS (class IV) and CheA (class V) from the B. subtilis genome as probes. A differentiation between class IIIA and IIIB as in B. subtilis was not made. Five of the histidine kinases turned out to be orphans.

To further analyse the homologies of the response regulators, they were blast analysed and aligned with response regulators of B. subtilis belonging to different families. The amino acid sequences of LytT (Others-B family), DegU (NarL family), PhoP (OmpR family), CitT (Others-A family) and CheY (CheY family) were used as probes. As the regulator domains are highly conserved throughout the response regulators, just the downstream region belonging to the DNA-binding domain (leaving out the first 120 amino acids) was used to be able to differentiate between distinct families. As CheY is just 120 amino acids long, the whole sequence was used in this case.

Three response regulators form the Others-B family, one fell into the NarL family, 26 into the OmpR family, one into the NtrB family, two into the Others-A family and one into the CheY family whereas four of them were orphans. Without exception the response regulators forming a family could be paired with the kinases forming a class. In addition, each family and class had identical gene order within the class.

CBO1872 was identified as coding for Spo0A with an e-value of −83. No Spo0F or Spo0B orthologues were found. No multidomain kinase-response regulator composite proteins were found.

Cloning and characterization of the C. botulinum spo0A gene

The spo0A gene, CBO1872, was polymerase chain reaction (PCR) amplified from chromosomal DNA and cloned in the plasmid pHT315S which placed the gene under control of the spac promoter. The primer used cloned 20 base pairs upstream of the ATG start codon and included the ribosome binding site but excluded the Spo0A binding site in the promoter region. This plasmid was transformed into B. subtilis strains JH642 (Spo+), JH646 (spo0A) and JH648 (spo0B). In these strains the C. botulinum Spo0A was constitutively expressed from the spac promoter. The transformants were characterized by small colonies that segregated larger colonies suggesting growth inhibition by the C. botulinum Spo0A (data not shown). In order to overcome this problem the spo0A gene was subcloned into plasmid pJM119 which placed the gene under lacI control of the spac promoter giving rise to plasmid pK2; the cloned region was then integrated into the chromosomal amyE gene as a single copy in the wild-type strain JH642 generating strain JH28011 (Fig. 2).

Figure 2.

Clostridial spo0A insertion in the B. subtilis chromosome. Clostridial spo0A (spo0AC.b.) was expressed in single copy following insertion from plasmid pK2 into the non-essential amyE locus by double crossover. Features relevant for expression are shown. The protein is expressed from the IPTG-inducible spac promoter (PSpac) and controlled by the penP-driven lacI product as described (Yansura and Henner, 1984). Position of a single BspDI site used to disrupt the lacZ gene (see Experimental procedures) for compatibility with β-galactosidase assays is shown.

To test the ability of the C. botulinum Spo0A to function in B. subtilis, chromosomal DNA from strain JH28011 was used to transfer the C. botulinum spo0A to strains JH646 and JH648 selecting for kanamycin resistance (strains JH28015 and JH28013 respectively). Control transformants without a cloned insert were also obtained by integration of pJM119 in the amyE gene (strains JH28014 and JH28012). Sporulation was tested after 23 and 49 h of growth in sporulation medium (SM) containing isopropyl-β-D-thiogalacto-pyranoside (IPTG). The wild-type strain JH642 produced normal levels of spores in these conditions but the integrated C. botulinum spo0A showed no ability to complement the sporulation deficiency of either JH646 (spo0A) or JH648 (spo0B) (data not shown).

An early function of phosphorylated Spo0A is to promote the transcription of the spoIIA and spoIIG genes (Kumar et al., 2004; Seredick and Spiegelman, 2004). The ability of the C. botulinum Spo0A to activate the transcription of the spoIIA and spoIIG genes was assayed in strain JH646 carrying either a spoIIA-lacZ or spoIIG-lacZ fusion. The results of these assays showed that this protein was unable to activate the transcription of either gene to an appreciable extent (data not shown). Thus the C. botulinum Spo0A protein appeared non-functional in B. subtilis either inherently or because of instability.

In order to distinguish between these possibilities, experiments were designed to determine if the C. botulinum Spo0A was unstable in B. subtilis. Extracts of the B. subtilis strains described above were prepared and subjected to SDS electrophoresis and Western blot analysis using an antibody prepared against B. subtilis Spo0A. As the sequence of the C-terminal DNA-binding domain of this protein is very highly conserved between the two strains, it was predicted that some of the epitopes should be identical and recognized in C. botulinum Spo0A by the antibodies. The C. botulinum Spo0A was found to be expressed and migrated in the SDS gel at a molecular weight predicted for the intact protein (Fig. 3). Thus it appeared that stability was not a factor in the inability of C. botulinum Spo0A to complement in B. subtilis.

Figure 3.

Clostridium botulinum spo0A is stably expressed in B. subtilis. To address whether clostridium Spo0A is stable in B. subtilis, whole cell protein extracts of the following B. subtilis strains were analysed immunologically with an anti-Spo0A antibody. Lane one contained purified his-tagged B. subtilis Spo0A. Extract in lanes 2 and 3 derived from strain JH28010 (wild-type), lanes 4 and 5 from JH28014 (spo0A) and lanes 6 and 7 from strain JH28015 (JH28014 expressing clostridial spo0A). Protein extract in lanes 2, 4 and 6 was derived from strains grown in the absence of 1 mM IPTG, whereas protein extract in lanes 3, 5 and 7 was derived from strains grown in the presence of 1 mM IPTG. The molecular weight of C. botulinum Spo0A is 30 840 Da and B. subtilis Spo0A is 29 693 Da.

Clostridium botulinum Spo0A has repressor function in B. subtilis

Phosphorylated Spo0A is both a repressor and an activator of genes in B. subtilis. Classically the abrB gene is repressed by Spo0A and this repression is one of the earliest manifestations of low-level Spo0A phosphorylation (Perego et al., 1988). In a spo0A mutant strain the inability to repress abrB results in an extracellular protease-deficient phenotype (Ferrari et al., 1988). The strains described above were tested for protease production on milk plates in the presence or absence of IPTG to induce C. botulinum Spo0A. The results showed a distinct protease halo in both JH646 and JH648 strains bearing the C. botulinum spo0A gene that was dependent on IPTG induction and that was absent from these strains bearing the pJM119 insert negative control (Fig. 4). These results indicated that C. botulinum Spo0A was capable of repressing abrB despite being unable to complement spo0A deficient strains for sporulation. Furthermore, if this repression activity required phosphorylation, it was not dependent on Spo0B.

Figure 4.

Clostridial Spo0A is able to repress abrB independently of phosphorylation.
A. The ability of C. botulinum Spo0A to repress the negative regulator abrB was tested on milk plates. A halo surrounding a colony indicates protease production as a result of abrB repression. B. subtilis strains JH642 (wild-type), JH646 (spo0A) and JH648 (spo0B) harbouring in the amyE locus either an empty pJM119 vector or expressing either wild-type or phosphorylation point mutants of C. botulinum spo0A were spotted as indicated. Plates contained either 0 or 1 mM IPTG as indicated.
B. To demonstrate that C. botulinum spo0A and the four phosphorylation point mutants were expressed to similar levels JH646 harbouring either spo0Ac.b. (lanes 1 and 2) spo0Ac.b.D58A (lanes 3 and 4), spo0Ac.b.D58E (lanes 5 and 6), spo0Ac.b.D58H (lanes 7 and 8) or spo0Ac.b.D58N (lanes 9 and 10) were grown in the absence (odd lanes) or the presence (even lanes) of 1 mM IPTG. C. botulinum Spo0A was visualized immunologically as described for Fig. 3.

In order to ensure that observed protease expression results were due to Spo0AC.B. acting directly on the abrB promoter, β-galactosidase expression from an abrB-lacZ transcription fusion was assayed in a spo0B strain in the presence or absence of Spo0AC.B.(Fig. 5). Spo0AC.B. repressed abrB-lacZ expression as expected for a direct interaction with the abrB promoter.

Figure 5.

Clostridial Spo0A represses abrB in B. anthracis. Shown are (A) growth in OD units and (B) β-galactosidase activity in Miller Units of B. subtilis spo0B strains containing a chromosomal transcriptional abrB-lacZ fusion and that either did or did not express clostridial spo0A. Strains were JH28062 [spo0B136 abrB::lacZ amyE::(aph3A, lacZΔ)] (black circles) and JH28063 [spo0B136 abrB::lacZ amyE::(spo0AC.b., aph3A, lacZΔI)] (white squares).

Assuming the repression of abrB required phosphorylation of the C. botulinum Spo0A, a number of mutant B. subtilis sensor histidine kinase strains were tested as possible sources of this phosphorylation without success. Acetyl-phosphate is known to be able to phosphorylate some response regulators directly (Lukat et al., 1992). The only source of biosynthetic acetyl-phosphate is from conversion of acetyl-CoA by phosphotransacetylase (Presecan-Siedel et al., 1999). A deletion mutant of this enzyme was transformed in the strains described above and these strains continued to produce protease upon IPTG induction (data not shown). While these studies were not comprehensive, the notion that the abrB repression activity might not require phosphorylation was entertained.

Generation of mutations in the phosphorylated aspartate of Spo0A was deemed a more direct way to test the requirement for phosphorylation in abrB repression. This aspartate, D58, was replaced individually with alanine, histidine, asparagine and glutamate. The mutant spo0A genes were cloned in plasmid pJM119 and the insert was integrated in strains JH642, JH648 and JH646. All strains expressed the C. botulinum Spo0A in response to IPTG (Fig. 4B). These strains were then tested for protease production on milk plates (Fig. 4A). Any and all of the D58 mutants were still capable of producing protease when induced with IPTG. Thus the simplest conclusion is that C. botulinum Spo0A can act as a repressor of abrB in the absence of phosphorylation.

Phosphorylation of Spo0A in C. botulinum

The problem of identifying the sensor histidine kinase(s) responsible for phosphorylation of Spo0A in C. botulinum was approached by attempting to clone the sensor histidine kinase genes of this organism into B. subtilis. There were five orphan sensor histidine kinase genes in C. botulinum that might be directly involved in phosphorylation of Spo0A (Table 1). Four of these (CBO0340, CBO1120, CBO2762 and CBO0336) were amplified from chromosomal DNA and cloned into pHT315S. The only gene successfully cloned was CBO1120; the others did not give complete inserts probably because of lethality in Escherichia coli. Regardless, the pHT315S::CBO1120 plasmid was transformed into JH646 (spo0A) with the pJM119 vector carrying C. botulinum spo0A gene integrated in the amy E region (strain JH28015). The strains with the four non-phosphorylatable mutant C. botulinum spo0A genes described above and the control strain carrying pJM119 lacking spo0A were also recipients of the plasmid. Each strain was tested for growth and sporulation with and without IPTG. The results showed that the combination of the CBO1120 sensor histidine kinase and induced C. botulinum Spo0A was lethal to B. subtilis(Fig. 6). However, CBO1120 alone, in combination with the non-phosphorylatable mutant Spo0A or with uninduced Spo0A, did not show this phenotype. Neither the spo0A colony phenotype nor the sporulation deficiency of these strains was reversed in any case. Thus the conclusion was drawn that CBO1120 is likely to be a sensor histidine kinase for sporulation in C. botulinum that directly phosphorylates Spo0A. Moreover the lethality may be a result of enhanced repression by Spo0A upon phosphorylation.

Figure 6.

Clostridial Spo0A but not Spo0A phosphorylation point mutants are lethal to B. subtilis when coexpressed with the histidine kinase CBO1120. B. subtilis strains expressing the clostridial kinase CBO1120 from the multicopy plasmid pHT315S and harbouring in the amyE locus either empty vector pJM119 (streak 1), or C. botulinum spo0A (2), spo0AD58A (3), spo0AD58E (4), spo0AD58H (5), spo0AD58N (6) were grown in the absence (left plate) or presence (right plate) of 1 mM IPTG.

Properties of a B. subtilis–C. botulinum Spo0A chimera

The results at this point indicated that the C. botulinum Spo0A was likely to be non-phosphorylated in B. subtilis and that introducing a compatible sensor histidine kinase from C. botulinum resulted in phosphorylation but the Spo0A protein was now a lethal factor rather than a complementing protein. Certainly some new properties had been acquired by phosphorylation but not enough to be quantitatively similar to the B. subtilis orthologue. The most highly conserved portion of the C. botulinum Spo0A is the C-terminal DNA-binding domain, suggesting that the less conserved N-terminal domain may be non-functional in B. subtilis even if phosphorylated (Fig. 8). In order to test this notion, a chimera was constructed. The N-terminal domain and the linker between the two domains from B. subtilis were fused to the C-terminal domain of C. botulinum Spo0A and cloned in plasmid pHT315S. This plasmid was transformed into strain JH646MS (spo0A, abrB) and the transformants were assayed on SM plates for sporulation. The colony phenotype of the transformants was converted from a spo0A transparent phenotype to a partially opaque colony resembling a strain capable of sporulation (Fig. 7). After 72 h the colonies contained about 1% spores whereas the control colonies with a cloned intact B. subtilis spo0A gave about 20% spores. The chimera colonies were treated with chloroform to kill non-spores and plated on SM. The colonies that grew from these spores were identical to the strain from which they were derived and PCR analysis of their DNA showed no rearrangements or loss of cloned Spo0A (data not shown). The chimera results suggest that the amino acid differences in both the N-terminal and the C-terminal domain found in C. botulinum Spo0A are of consequence to sporulation in B. subtilis as only a partial suppression was observed.

Figure 8.

The C-terminal DNA-binding domain of B. subtilis and C. botulinum Spo0A is highly homologous. Shown is an alignment of B. subtilis and C. botulinum Spo0A amino acid sequences. The secondary structure of the C-terminal DNA binding domain is displayed. Residues that interact with the DNA are in bold and highlighted (Zhao et al., 2002). The arrow depicts the position at which a chimera of the two proteins was generated.

Figure 7.

Suppression of the Spo0A colony phenotype by B. subtilis/C. botulinum Spo0A chimera. Strains were grown on SM medium and antibiotics.
A. Strain 646MS (spo0A, abrB15)/pHT315S.
B. Strain 646MS (spo0A, abrB15)/pHT315S-spo0A (pBsu0A).
C. Strain 646MS (spo0A, abrB15)/pHT315S-spo0A chimera (pB/C0A).


The original intent of this research was to identify histidine sensor kinases that serve to phosphorylate Spo0A directly and regulate the onset of sporulation. Several factors impeded the rapid attainment of this goal. The most serious was the surprising finding that in spite of the high conservation of amino acid sequence in the DNA-binding domain of Spo0A and the virtual identity of the residues known to make contact with the ‘0A’ box in promoters (Zhao et al., 2002), the C. botulinum Spo0A was unable to complement the sporulation deficiency of a spo0A mutant of B. subtilis. Expression experiments showed that this was most likely due to an inherent cross species defect and not due to lack of expression or destruction in B. subtilis. The Spo0A chimera produced with the N-terminal response regulator domain of B. subtilis and the DNA-binding domain of C. botulinum was able to complement sporulation in a spo0A mutant of B. subtilis albeit not at wild-type levels. Thus it appears that the infrequent amino acid changes in the C. botulinum DNA-binding domain render this domain less effectively in B. subtilis. The high conservation of amino acid sequence of Spo0A in Bacillus and Clostridium species was noticed some time ago (Brown et al., 1994). The most probable reason for this conservation is that Spo0A must interact with other proteins whose amino acid sequences are conserved. It is known to interact with RNA polymerases containing either Sigma A or Sigma H (Buckner and Moran, 1998). The αE region of Spo0A in which mutations affect Sigma A-dependent transcription (Buckner and Moran, 1998; Hatt and Youngman, 1998) shows significant amino acid variability in the two species (Fig. 8), which may explain the reduced ability of Spo0AC.B. to activate sporulation gene transcription in B. subtilis. In addition Spo0A may contact (Buckner et al., 1998) as well the septation regulating proteins (McLeod and Spiegelman, 2005). It may have other interactions that are unknown at present and the sum of these may serve to constrain randomness in its sequence.

The N-terminal domain of Spo0A is more prone to evolve its amino acid sequence (Fig. 8). It cannot be easily determined whether the observed changes in this domain in C. botulinum would render it non-functional in B. subtilis because it is unknown if there is some level of phosphorylation occurring in B. subtilis. It remains possible that if phosphorylated it could function. The chimera results were interpreted to indicate that the C. botulinum N-terminal domain is not phosphorylated in B. subtilis.

The C. botulinum Spo0A has the ability to repress the abrB gene as evidenced by protease production. and this repression still occurs when the phosphorylated aspartate is mutated to a non-phosphorylatable amino acid. The simple explanation for this observation is that repression only requires Spo0A binding to the affected promoter to block progression of RNA polymerase and the amino acid residues known to bind to ‘0A’ boxes on DNA are unchanged in C. botulinum. Why then doesn’t B. subtilis Spo0A do the same thing in strains that do not allow it to be phosphorylated such as a spo0B mutant? The answer may lie in the relative cellular amount of Spo0A in the two instances. In the C. botulinum Spo0A case the protein is induced from the spac promoter and may be present in much larger quantities than in normal B. subtilis where it has only been tested when expressed from its natural promoter. In the latter case phosphorylation may increase its DNA-binding activity substantially.

The CBO1120 sensor histidine kinase was identified as being capable of phosphorylating Spo0A and may serve this purpose in vivo. The sequence of the HisKA domain of CBO1120 which makes productive interaction with the response regulator and determines its specificity (Zapf et al., 2000) is very highly related to three other orphan sensor histidine kinases: CBO0336, CBO0340 and CBO2762, suggesting that they might also be capable of initiating sporulation (Table 2). Unfortunately they could not be cloned easily to test this notion. Multiple sensor histidine kinases for sporulation are the rule in both B. subtilis and Bacillus anthracis where they may allow sporulation under a variety of environmental conditions (Brunsing et al., 2005). Comparison of the N-terminal ‘signal recognition’ domain of each of the putative sporulation sensor histidine kinases with translated genomes of the other Clostridia and related anaerobes revealed possible orthologues only in C. perfringens. This observation is consistent with the results of comparing these domains between Bacillus strains and suggests that the signals that drive sporulation may be unique to each species and reflect the environment in which each species resides.

The combination of expression of both the C. botulinum Spo0A and the sensor histidine kinase CBO1120 in the same cell prevents growth. Phosphorylation appears to augment an inherent inhibitory activity resulting from C. botulinum Spo0A expression in B. subtilis. Non-phosphorylatable Spo0A mutants are unaffected by the sensor histidine kinase. The basis for this inhibition is not known but the evidence suggests that C. botulinum Spo0A is an effective repressor of genes known to be repressed by this transcription factor (Liu et al., 2003; Molle et al., 2003) especially if phosphorylated but is unable to activate sporulation genes. It is tempting to suggest that Spo0A as the master regulator of development has a repression activity on some cell division and growth genes as a normal prelude to activating sporulation and this repression is being observed here. In B. subtilis the tight control on phosphorylation of B. subtilis Spo0A by the phosphatases (Perego and Hoch, 2002) of the phosphorelay may prevent this from occurring except as a transient state preceding activation of sporulation. The regulatory events controlling Spo0A activity at the crucial decision point between cell division and sporulation may be revealed by determining the target(s) of C. botulinum Spo0A inhibitory activity.

Experimental procedures

Bioinformatic analysis

DNA sequence and predicted protein sequences of C. botulinum were obtained from Homologues of histidine kinases and response regulators in C. botulinum were identified by performing a blast search on the genome of C. botulinum with the amino acid sequence of the histidine kinase domain (starting 15 amino acids upstream of the phosphorylatable histidine) and the ATP-binding domain of B. subtilis KinA and the sequence of Spo0F of B. subtilis respectively. The numbers of putative transmembrane domains were determined using the TMpred program.

Bacterial strains and growth conditions

All B. subtilis strains used in this study were derivatives of the parental strain JH642 and are shown in Table 3. A spectinomycin cassette mutation in the B. subtilis phosphotransacetylase gene, pta, was a kind gift of E. Presecan-Siedel, Institut Pasteur. Strains were grown in Schaeffer SM (Schaeffer et al., 1965) or Spizizen minimal medium and transformed based on the method of Anagnostopoulos and Spizizen (1961). Transformants were selected on SM plates. Antibiotics were used at the following concentrations: erythromycin 5 µg ml−1 or lincomycin 25 µg ml−1, chloramphenicol 5 µg ml−1, kanamycin 2 µg ml−1, spectinomycin 100 µg ml−1. Integration at the amyE locus of pJM119 derived plasmids was confirmed by screening for amylase deficiency on TBAB plates supplemented with 1% starch. When appropriate, IPTG was used at a concentration of 1 mM for induction.

Table 3. Bacillus subtilis strains used in this study.
StrainaRelevant genotypeDescription
  • a. 

    All strains are derivative of the parental strain JH642, therefore they carry the trpC2, phe-1 auxothrophic markers.

  • b. 

    Wu et al. (1992).

  • c. 

    Strain construction: Strains JH28055 and JH28056 were transformed with JH12556 selecting for CmR and thereby introducing a chromosomal transcriptional abrB-lacZ fusion.

JH642ParentalLaboratory Stock
JH646spo0A12Laboratory Stock
JH648spo0B136Laboratory Stock
JH646MSspo0A12, abrB15Laboratory Stock
JH28011amyE::(spo0AC.b., aph3A)JH28011→JH642
JH28012spo0B136, amyE::(aph3A)pJM119→JH648
JH28013spo0B136, amyE::(spo0AC.b., aph3A)JH28011→JH648
JH28014spo0A12, amyE::(aph3A)pJM119→JH646
JH28015spo0A12, amyE::(spo0AC.b., aph3A)pK2→JH646
JH28051amyE::(aph3A, lacZΔ)pK3→JH642
JH28052amyE::(spo0AC.b., aph3A, lacZΔ)pK4→JH642
JH28053spo0A12, amyE::(aph3A, lacZΔ)pK3→JH646
JH28054spo0A12,amyE::(spo0AC.b., aph3A, lacZΔ)pK4→JH646
JH28055spo0B136, amyE::(aph3A, lacZΔ)pK3→JH648
JH28056spo0B136, amyE::(spo0AC.b., aph3A, lacZΔ)pK4→JH648
JH28057spo0A12, amyE::(aph3A, lacZΔ)/pPP155b→JH28053
JH28058spo0A12,amyE::(spo0AC.b., aph3A, lacZΔ)/pPP155b→JH28054
JH28059spo0B136, amyE::(aph3A, lacZΔ)/pPP155b→JH28055
JH28060spo0B136, amyE::(spo0AC.b., aph3A, lacZΔ)/pPP155b→JH28056
JH28062spo0B136, amyE::(aph3A, lacZΔ) abrB::lacZJH12556→JH28055c
JH28063spo0B136, amyE::(spo0AC.b., lacZΔ) abrB::lacZJH12556→JH28056c
JH28064amyE::(spo0AC.b.D58A, aph3A)pK5→JH642
JH28065amyE::(spo0AC.b.D58H, aph3A)pK6→JH642
JH28066amyE::(spo0AC.b.D58E, aph3A)pK7→JH642
JH28067amyE::(spo0AC.b.D58N, aph3A)pK8→JH642
JH28068spo0A12, amyE::(spo0AC.b.D58A, aph3A)pK5→JH646
JH28069spo0A12, amyE::(spo0AC.b.D58H, aph3A)pK6→JH646
JH28070spo0A12, amyE::(spo0AC.b.D58E, aph3A)pK7→JH646
JH28071spo0A12, amyE::(spo0AC.b.D58N, aph3A)pK8→JH646
JH28072spo0B136, amyE::(spo0AC.b.D58A, aph3A)pK5→JH648
JH28073spo0B136, amyE::(spo0AC.b.D58H, aph3A)pK6→JH648
JH28074spo0B136, amyE::(spo0AC.b.D58E, aph3A)pK7→JH648
JH28075spo0B136, amyE::(spo0AC.b.D58N, aph3A)pK8→JH648
JH28077amyE::(spo0AC.b., aph3A)/mlsRpK25→JH28011
JH28078spo0A12, amyE::(aph3A)/mlsRpK25→JH28014
JH28079spo0A12, amyE::(spo0AC.b., aph3A)/mlsRpK25→JH28015
JH28080spo0B136, amyE::(aph3A)/mlsRpK25→JH28012
JH28081spo0B136, amyE::(spo0AC.b., aph3A)/mlsRpK25→JH28013
JH28082spo0B136, amyE::(spo0AC.b.D58A, aph3A)/mlsRpK25→JH28072
JH28083spo0B136, amyE::(spo0AC.b.D58H, aph3A)/mlsRpK25→JH28073
JH28084spo0B136, amyE::(spo0AC.b.D58E, aph3A)/mlsRpK25→JH28074
JH28085spo0B136, amyE::(spo0AC.b.D58N, aph3A)/mlsRpK25→JH28075
JH28091spo0A12, abrB15/mlspBsu0A→JH646MS

Escherichia coli DH5α and TG1 were used for plasmid constructions. Strains were grown in Luria–Bertani (LB) medium. Antibiotics were used at the following concentrations: ampicillin 100 µg ml−1, kanamycin 30 µg ml−1.

Plasmid constructions

All relevant plasmids used in this study are listed in Table 4. Plasmid pHT315S was obtained by cloning the spac promoter from pMUTIN4 (Vagner et al., 1998) in the EcoRI and SacI sites of pHT315 (Arantes and Lereclus, 1991). The clostridial spo0A gene was amplified by PCR from chromosomal DNA of C. botulinum ATCC 3502 using the following oligonucleotides:

Table 4.  Plasmids used in this study.
pHT315E. coli/B. subtilis shuttle vector AmpR ErmR (Arantes and Lereclus, 1991)
pHT315SpHT315 with Pspac promoter from pMUTIN4 (Vagner et al., 1998; V. Dartois and J.A. Hoch, unpublished)
pJM119AmpR KanRamyE integration vector with IPTG-inducible Pspac (M. Perego, unpublished)
pKS(-) AmpRorif1(-) oripUC PlaclacZ (Stratagene, La Jolla CA)
pK1pHT315S-spo0AC.b. (SacI-BamHI)
pK2pJM119-spo0AC.b. (SmaI-BamHI)
pK3pJM119 lacZ
pK4pK3-spo0AC.b. (SmaI-BamHI)
pK5pJM119-spo0A(D58A)C.b. (SmaI-BamHI)
pK6pJM119-spo0A(D58H)C.b. (SmaI-BamHI)
pK7pJM119-spo0A(D58E)C.b. (SmaI-BamHI)
pK8pJM119-spo0A(D58N)C.b. (SmaI-BamHI)
pK20pKS(-)-spoAC.b. (SacI-BamHI)
pK21pKS(-)-spo0A(D58A)C.b. (SacI-BamHI)
pK22pKS(-)-spo0A(D58H)C.b. (SacI-BamHI)
pK23pKS(-)-spo0A(D58E)C.b. (SacI-BamHI)
pK24pKS(-)-spo0A(D58N)C.b. (SacI-BamHI)
pK25pHT315S-CBO1120 (SmaI-BamHI)
pB/C0ApHT315S-spo0A (B. subtilis–C. botulinum chimera)



pK1 and pK2 were constructed by cloning the clostridial spo0A gene into SacI and BamHI sites of pHT315S and SmaI and BamHI sites of pJM119 respectively.

To disrupt the lacZ gene for compatibility with β-galactosidase assays, the plasmids pJM119 and pK2 were digested with BspDI and subjected to a Klenow-fill-in reaction and subsequent religation creating a frameshift mutation within the lacZ gene and resulting in plasmid pK3 and pK4 respectively.

The plasmid pK25 was derived by cloning the gene CBO1120, coding for a clostridial kinase, into the SmaI and BamHI sites of pHT315S using oligonucleotides: 5′-TTAT TGAATTCTAGATGTTATAACTGAAAAGG-3′ and 5′-AATTAG GATCCTTTTGAAACGCCCTCTTTATC-3′.

To construct the chimerical spo0A gene DNA fragments corresponding to the B. subtilis Spo0A N-terminal regulatory domain (N-spo0Ab for amino acids 1–158) and the C. botulinum C-terminal transcription activation domain (C-spo0Ac for amino acids 165–273) were PCR amplified using the following pairs of oligonucleotides: 5′-CTGAGCTCTA CATTTGGGGAGGAAGA-3′ and 5′-TTCATGGATGATATCT GTGATGCTCGC-3′ or: 5′-ACAGATATCATACATGAAATAGG CGTG-3′ and 5′-CTAACGGATCCATCAACTAAGCGATT-3′.

Additionally full-length spo0A of B. subtilis was also amplified with the following oligonucleotide pair: 5′-CTGAGCTC TACATTTGGGGAGGAAGA-3′ and 5′-CGGGATCCTGTTTAA GAAGCCTTATGCTC-3′.

The PCR products were cloned into pCR® 4-TOPO vector (Invitrogen) according to the Manufacturer's protocol. SacI/EcoRV fragment DNA containing N-spo0Ab and EcoRV/BamHI fragment DNA containing C-spo0Ac from the correct TOPO clones were purified and cloned into pHT315S vector via three way ligation. SacI and BamHI fragment DNA containing B. subtilis spo0A was also cloned into pHT315S. These constructs were designated as pB/C0A and pBsu0A respectively. The constructs were sequenced to confirm the authenticity of the spo0A fusion and spo0A gene.

Site-directed mutagenesis

The clostridial spo0A gene was cloned into the SacI and BamHI site of pKS(–) (Stratagene) generating plasmid pK20 and site-directed mutagenesis was carried out using the BIO-RAD Muta-Gene Phagemid In Vitro Mutagenesis system based on the method described by Kunkel (1985) essentially as described by the supplier.

The following oligonucleotides were used for the in vitro mutagenesis:





For the final transformation E. coli DH5α cells were used. The mutations were confirmed by sequencing, resulting in plasmids pK21-pK24.

The mutated spo0A genes were subcloned into the SacI and BamHI sites of pJM119 and the resulting plasmids pK5-pK8 were transformed into the appropriate B. subtilis strains.

Sporulation efficiency

The efficiency of sporulation was tested by inoculating a single colony in 5 ml of Schaeffer SM supplemented with the appropriate antibiotics; cultures were grown for 23 h and 49 h, respectively, at 37°C. Serial dilutions were plated before and after treatment with CHCl3 in order to obtain the viable cell count and the spore count, as described.

β-Galactosidase assays

Cells were grown in SM at 37°C in the presence of the appropriate antibiotics. Samples were taken at indicated times. β-Galactosidase activity was determined as previously described (Ferrari et al., 1985) and the activity is reported in Miller Units (Miller, 1972).

Western blot

Strains were grown in LB medium for 6 h, collected by centrifugation, washed twice with protoplast buffer (25 mM potassium phosphate pH 7, 10 mM magnesium chloride, 0.1 mM EDTA, 20% sucrose, and 30 mM sodium lactate) supplemented with 250 µg ml−1 chloramphenicol and resuspended in that buffer at an A525nm = 1.0. One millilitre of cell suspension was treated with 4 mg lysozyme for 30 min at 37°C and the resulting protoplasts were collected by centrifugation. Protoplasts were resuspended and boiled in 100 µl of 1× SDS-loading buffer and 10 µl were separated on a 12% SDS-PAGE gel and transferred to a polyvinylidene difluoride (PVDF) membrane. Detection of Spo0A was carried out in principal as described in the ECL Plus kit by Amersham. The primary rabbit-anti-Spo0AB. subtilis antibody was used at a dilution of 1:10 000 following three pre-absorption treatments with an acetone protein powder of strain JH28014 in order to reduce unspecific binding.

Protease production

Protease production phenotypes were scored on TBAB plates supplemented with 1% skim milk. Formation of a halo around spotted colonies following 14 h incubations at 37°C confirms the production of proteases.


This research was supported in part by Grants GM19416 and AI55860 from the National Institute of General Medical Sciences and the National Institute of Allergy and Infectious Diseases, National Institutes of Health, United States Public Health Service. Oligonucleotide synthesis and sequencing services were supported in part by the Stein Beneficial Trust. The authors are indebted to Dr Marta Perego for invaluable scientific and technical advice, to Dr Eric Johnson, University of Wisconsin, for providing C. botulinum DNA and to Dr Julian Parkhill, The Sanger Institute, for allowing access to the C. botulinum genome sequence. Dr Veronique Dartois is acknowledged for the construction of plasmid pHT315S.