|Home | About | Journals | Submit | Contact Us | Français|
The initiation of sporulation in Bacillus species is regulated by the phosphorelay signal transduction pathway, which is activated by several histidine sensor kinases in response to cellular and metabolic signals. Comparison of the protein components of the phosphorelay between Bacillus subtilis and Bacillus anthracis revealed high homology in the phosphorelay orthologs of Spo0F, Spo0B, and Spo0A. The sensor domains of sensor histidine kinases are poorly conserved between species, making ortholog recognition tenuous. Putative sporulation sensor histidine kinases of B. anthracis were identified by homology to the HisKA domain of B. subtilis sporulation sensor histidine kinases, which interacts with Spo0F. Nine possible kinases were uncovered, and their genes were assayed for complementation of kinase mutants of B. subtilis, for ability to drive lacZ expression in B. subtilis and B. anthracis, and for the effect of deletion of each on the sporulation of B. anthracis. Five of the nine sensor histidine kinases were inferred to be capable of inducing sporulation in B. anthracis. Four of the sensor kinases could not be shown to induce sporulation; however, the genes for two of these were frameshifted in all B. anthracis strains and one of these was also frameshifted in the pathogenic pXO1-bearing Bacillus cereus strain G9241. It is proposed that acquisition of plasmid pXO1 and pathogenicity may require a dampening of sporulation regulation by mutational selection of sporulation sensor histidine kinase defects. The sporulation of B. anthracis ex vivo appears to result from any one or a combination of the sporulation sensor histidine kinases remaining.
The initiation of sporulation in Bacillus subtilis and most likely in aerobic Bacillus species in general is controlled by the phosphorelay signal transduction system (4) (Fig. (Fig.1).1). The ultimate goal of the phosphorelay is to activate by phosphorylation the Spo0A transcription factor, which represses certain genes and promotes the transcription of a large number of genes for stationary-phase functions as well as sporulation (16, 23). The signals that initiate the phosphorelay reactions are recognized and interpreted by several sensor histidine kinases (10, 11, 13, 14). While the nature of these signals in Bacillus subtilis and other Bacillus species remains a mystery, the target of their activation is the Spo0F response regulator, to which they transfer a phosphoryl group. The phosphoryl group is subsequently transferred to Spo0A through the Spo0B phosphotransferase. This basic pathway is subject to a surfeit of secondary controls that regulate the transcription of its genes and the flow of phosphoryl groups through its proteins to the Spo0A response regulator transcription factor (25, 26).
The Spo0A transcription factor is known to play a major role in the regulation of expression of the component genes of the toxin synthesized by Bacillus anthracis through its regulation of the transition state regulator AbrB. In a B. anthracis strain mutant for AbrB, the pagA gene coding for the protective antigen component of the anthrax toxin is more highly expressed than in a wild-type strain (29). The pagA gene is also subject to AbrB repression when cloned in B. subtilis (3). Phosphorylation of Spo0A by the phosphorelay results in repression of AbrB, which, in turn, allows anthrax toxin gene expression. AbrB is a transition state regulator that binds to a wide variety of stationary-phase promoters and prevents their expression during active growth (35). It has the unique property of recognizing these promoters without an apparent consensus nucleotide sequence, making it difficult to identify its targets from sequence information (37). While it was clear that the promoter for the pag gene coding for the protective antigen component of the anthrax toxin is AbrB regulated, it has been only recently reported that this effect may be the result of direct repression of the atxA promoter by the AbrB protein (34).
The Spo0A transcription factor/response regulator is very highly conserved in amino acid sequence in all of the Bacillus and Clostridium species (33). In the DNA-binding domain all of the residues known to make contact with the DNA target of Spo0A, the 0A box, are identical, which suggests that the sequence of the 0A box is invariant in sporulating bacteria (39). Amino acid sequence conservation is also apparent in the Spo0F and Spo0B components of the phosphorelay (33). More importantly there is essentially invariant conservation of the amino acids of these proteins that are known to make interactions between them, and it is these invariant residues that underlie the molecular basis for recognition specificity (33, 36, 38). Thus, this amino acid conservation in the interaction surfaces allowed the initial identification of the sensor histidine kinases that may be involved in the regulation of sporulation in B. anthracis. The sensor domains of sporulation sensor histidine kinases are not highly conserved between different species of bacilli (33), and this species variability complicates ortholog identification.
This communication describes nine sensor histidine kinases of B. anthracis that were identified by amino acid conservation surrounding the active-site histidine of sporulation sensor kinase A of B. subtilis. These enzymes were cloned and expressed in B. subtilis and individually deleted in B. anthracis in order to uncover the signal transduction pathway to sporulation in this pathogenic organism.
B. subtilis strains used in this study are listed in Table Table1.1. Transformations were carried out as previously described (1). All the plasmids used in B. subtilis transformations are derivatives of pJM115 and carry the kanamycin resistance gene aphA-3; a multiple cloning site containing unique EcoRI, SmaI, and BamHI sites; the promoterless lacZ gene of Escherichia coli; and chromosomal regions for double crossover integration at the amyE locus. Transformants were selected for kanamycin (2 μg/ml) or chloramphenicol (5 μg/ml) resistance and screened for amyE activity on tryptose blood agar base plates containing 0.75% starch. Clones showing a loss of amylase activity were used for further study.
B. anthracis strains used in this study are listed in Table Table2.2. Transformations were carried out as previously described by Koehler et al. (12). The vector pTCVlac (28) was used for lacZ transcriptional fusion constructs. Transformants were selected on LB plus kanamycin (7.5 μg/ml). Strains JHA1001 to JHA1012 were obtained by transforming the Sterne strain 34F2 (pXO1+/pXO2−) with the pTKP plasmids. All cloning to generate knockout mutants was done in the vector pORI-Cm. Transformants were selected on LB plus spectinomycin (150 μg/ml) and screened on LB plus chloramphenicol (7.5 μg/ml) to test for loss of Cmr. Strains JHA1013 to JHA1020 were made by transforming 34F2 with the pORI-Cm gene-knockout derivatives.
E. coli DH5α and TB1 were used for plasmid construction and propagation. LB broth was used for growth of E. coli cultures. Transformants were plated on LB plus kanamycin (30 μg/ml) plus X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside; 40 μg/ml) for all pJM115 and pTCVlac-derived plasmids or LB plus ampicillin (100 μg/ml) for pJM116 derivatives, or on LB with spectinomycin (150 μg/ml) or with chloramphenicol (7.5 μg/ml) for pORI-Cm-derived plasmids. E. coli strains C600 and SCS110 (Stratagene) were used to generate nonmethylated plasmid DNA for B. anthracis transformations.
For all kinase genes of B. anthracis, the corresponding coding sequence and promoter region were amplified by PCR using the high-fidelity enzymes Pwo DNA polymerase (Roche) or Accuzyme DNA polymerase (Bioline). Chromosomal DNA was purified from the B. anthracis Sterne strain 34F2 (pXO1+/pXO2−) using the UltraClean microbial DNA isolation kit (MoBio Labs, Solana Beach, CA). The oligonucleotides used are listed in Table S1 of the supplemental material. The pJAK vectors were constructed by inserting a PCR product containing the full open reading frame and promoter region into the vector pJM115 or pJM116. PCR products that amplified the coding sequences of BA1351, BA3702, BA1478, BA4223, BA5029, and BA2636 were digested with EcoRI and BamHI or BglII while the vector was digested with EcoRI and BamHI. About 400 bp of the region upstream from the putative ribosome binding site was included in each construct. The vector pJAK1356 was constructed by cloning the undigested PCR product into SmaI-digested pJM115. The pJKP and pTKP vectors are transcriptional lacZ fusion constructs of each gene promoter region constructed in pJM115 or pTCVlac, respectively. PCR products of BA3702, BA4223, BA1478, and BA5029 were digested with EcoRI and XmnI to eliminate the 3′ end of the coding region and terminator. They were then cloned into either pJM115 or pTCVlac digested with EcoRI and SmaI to make the pJKP and pTKP constructs, respectively. The PCR product of BA1356 was cut with XmnI and cloned into pJM115 cut with SmaI to make pJKP1356. For the construction of pTKP1356 the PCR product obtained with oligonucleotides BA1356-5′Eco and BA1356-3′BglII was digested to completion with BamHI (which cut within the open reading frame) and partially with EcoRI. A 1.5-kb fragment was recovered and cloned into pTCVlac digested with EcoRI and BamHI. The PCR product of BA2636 was digested with EcoRI/BamHI and cloned into similarly digested pJM115 and pTCVlac to generate pJKP2636 and pTKP2636, respectively. The BA1351 lacZ constructs were generated by cloning the PCR fragment obtained with oligonucleotides BA1351-5′Eco2 and BA1351-3′Bam2 and digested with EcoRI-BamHI into similarly digested pJM115 and pTCVlac.
The pJMR constructs were made as follows. The promoter region of gene BA4223 was PCR amplified using primers Ba4810Eco5′ and BA4223UPBam-Nco and was designated P4223. These oligonucleotides introduced an EcoRI site in the upstream region of the promoter and adjacent NcoI and BamHI sites at the 3′ end with the NcoI site falling immediately upstream of the BamHI site. The ATG of the NcoI restriction site overlapped the translational start site for BA4223. PCR-amplified P4223 was digested with EcoRI/BamHI and cloned into EcoRI/BamHI-digested pJM115, generating plasmid pJMRNco. The coding sequences of BA1351 and BA5029 were PCR amplified using primers that generated a BspHI site overlapping the translational start site at the 5′ end of the coding sequence and a BamHI site at the 3′ end downstream of the terminator. Following a BspHI/BamHI digestion, the products were cloned into NcoI/BamHI-digested pJMRNco to generate pJMR1351 and pJMR5029, respectively. Similarly, BA4223 was amplified with an NcoI site at the ATG start codon at the 5′ end of the coding sequence and a BamHI site provided with oligonucleotide BA4223-3′Bam. Following an NcoI-BamHI digestion, this product was ligated into NcoI/BamHI-digested pJMRNco to generate pJMR4223. For pJMR1356, the vector was digested with NcoI, treated with mung bean nuclease (New England Biolabs), and then digested with BamHI. The coding sequence of BA1356 was then amplified with primers BA1356-5′Nde and BA1356-3′BglII (the latter introduced a BglII site at the 3′ end just downstream of the terminator). The sample was digested with BglII and then cloned into the digested pJMRNco, yielding pJMR1356.
BA2291 was PCR amplified using oligonucleotide primers BA22915′Kpn and BA22913′Hind and cloned in vector pJM116 to give rise to plasmid pJAK2291. The same fragment was also cloned in the multicopy vector pHT315 (2) to give plasmid pHT315-2291. BA2291 was also PCR amplified using oligonucleotides BA2291 5′Kpn and BA2291 3′Hind, and the undigested PCR fragment was cloned into pJM115 digested with SmaI. The orientation of the fragment was determined by EcoRI digestion. Digestion of this plasmid with BamHI and religation resulted in the deletion of 135 bp at the 3′ end of the gene inclusive of a putative transcription terminator. The resulting plasmid thus carried a transcriptional fusion to the E. coli lacZ gene contained in pJM115. This plasmid was named pJKP2291. The fragment carried by pJKP2291 was recovered by EcoRI-BamHI digestion (1,600 bp) and cloned into vector pTCVlac (28) digested with EcoRI-BamHI to obtain a transcriptional lacZ fusion plasmid suitable for studies in B. anthracis. This plasmid was named pTKP2291.
Vector pJM115 is a derivate of the transcriptional lacZ fusion plasmid pDH32 (24) with the kanamycin resistance cassette replacing the chloramphenicol resistant determinant. Vector pJM116 is also a derivative of pDH32 carrying an extended multiple cloning site (EcoRI, PmlI, BstBI, HindIII, PmeI, and BamHI) in place of the EcoRI-SmaI-BamHI multiple cloning site (M. Perego, unpublished).
The spectinomycin cassette plasmid pJM134 was constructed by cloning the spectinomycin resistance gene amplified from plasmid pIC33 (32) as an EcoRV fragment into the EcoRV site of pBluescript (Stratagene, La Jolla, CA) (M. Perego, unpublished). The oligonucleotides used for the PCR amplification reaction, Spc5′EV and Spc3′EV, are reported in Table S1 of the supplemental material.
Sporulation assays were carried out in Schaeffer's sporulation medium (SM) (30) or in LB broth as indicated. Cells were grown at 37°C in 5 ml of liquid culture for the time indicated in the tables. Serial dilutions were plated on SM agar plates before and after treatment with chloroform.
β-Galactosidase assays were carried out as previously described (6). Activity was expressed in Miller units (21). For B. subtilis, 100 ml SM was inoculated with a loopful of overnight culture from an SM agar plate. Five-milliliter samples were taken at 1-h intervals starting 1.5 h postinoculation. Assays with B. anthracis were similarly done in LB broth containing the appropriate antibiotic.
To generate pORI-Cm, the temperature-sensitive vector pCASPER (5) was digested with EcoRV to eliminate the kanamycin resistance cassette, and the resulting 2,496-bp fragment was ligated to a 1,051-bp chloramphenicol resistance cassette. The resistance cassette was obtained from pJM105A (24) by digestion with EcoRI and HindIII, followed by Klenow treatment to generate blunt ends. The orientation of the insert was confirmed by restriction digest with EcoRI and XmnI, and the resulting vector, which possesses the temperature-sensitive RepA-dependent replicon from pVE6002 (17), originally from pWV01 (15), was designated pORI-Cm (Fig. (Fig.22).
The method used for inactivation of the sensor histidine kinases of B. anthracis was based on the plasmid pORI-Cm. Each kinase gene was amplified by PCR from chromosomal DNA of the B. anthracis Sterne strain 34F2 using a 5′ EcoRI oligonucleotide primer and a 3′ BamHI or BglII oligonucleotide primer (see Table S1 in the supplemental material). After restriction nuclease cleavage the amplified DNA was cloned by ligation in pORI-Cm double digested with EcoRI and BamHI. The ligated DNA was transformed into E. coli strain DH5α and incubated at 30°C, and recombinants were selected on LB with 7.5-μg/ml chloramphenicol plates. Using the constructs obtained, the cloned kinase genes were inactivated by inserting the spectinomycin resistance cassette in each of them. This was accomplished, in general, by using a convenient single unique restriction site within the kinase gene as the site of insertion of an EcoRV-digested spectinomycin cassette. After ligation and transformation the recombinants were selected in E. coli as spectinomycin resistant, purified, and tested by restriction mapping and nucleotide sequencing.
The plasmids obtained carrying the deletion constructs were first amplified in the E. coli dcm dam strain SCS110 (Stratagene) or C600 and subsequently used to electroporate the B. anthracis Sterne strain 34F2 using 0.5 μg of DNA and the transformation protocol described by Koehler et al. (12). Transformants were selected on LB medium with 7.5 μg/ml chloramphenicol, at 30°C. Clones were purified by single colony isolation. At this point it is likely that the majority of the plasmids existed as free-replicating entities. In order to select for integration of the spectinomycin cassette in the chromosomal copy of each gene, clones were grown at 30°C in LB medium containing spectinomycin at 150 μg/ml to an optical density at 600 nm of 0.3 to 0.5 and shifted to 37 or 42°C for several hours. Dilutions of the culture were then plated on both chloramphenicol and spectinomycin plates to verify that there was a general loss of the chloramphenicol resistance-conferring vector from the culture. Colonies from the spectinomycin plate were individually picked onto spectinomycin or chloramphenicol plates to identify those colonies in which the spectinomycin was chromosomally integrated and the plasmid was lost.
Chromosomal DNA from presumptive clones containing an integrated spectinomycin cassette in the gene of interest and free of vector was subjected to several control PCRs to verify the constructions. First, oligonucleotide primers internal to the cat gene were used to ensure that this gene was not present in some other cryptic form. Second, the chromosomal structure of the integrated spectinomycin cassette was confirmed by two PCRs using primers within the spectinomycin cassette and directed outward to chromosomal primers outside the original cloned region. Finally the presence of pXO1 was confirmed using primers that amplified the atxA gene. Only clones that passed all of these tests were deemed inactivated for the gene and subsequently used for sporulation assays.
A bioinformatic approach was taken as the primary identification of genes coding for sensor histidine kinases with the possibility of being functional in the sporulation phosphorelay of B. anthracis. The premise underlying this approach is based on the identification of residues in the interaction surface of Spo0F by alanine scanning (36) and by the crystal structure of the Spo0F::Spo0B complex (38). The cocrystal structure of Spo0F and Spo0B identified and oriented those Spo0F residues contacting the helical hairpin containing the phosphorylated histidine of Spo0B. Subsequent studies showed that these residues were identical in the Spo0F of B. anthracis, indicating that the interaction residues of the helical hairpin carrying the active-site histidine of sensor histidine kinases known by genetic and biochemical analysis to be functional in B. subtilis sporulation must be highly conserved in B. anthracis (33). Thus, this region and surrounding residues (11 residues N terminal and 34 residues C terminal to the active-site phosphorylated histidine), the HisKA domain, from kinase A of B. subtilis were used to probe proteins identified from genomic sequences of B. anthracis strains using the Blastp program on genomic sequences at the National Center for Biotechnology Information.
These Blastp analyses identified nine kinases with significant homology in the HisKA domain to that of kinase A (Table (Table3).3). All of the genes for the nine sensor histidine kinases were “orfans” lacking an adjacent response regulator gene, a property characteristic of the five genes for sporulation sensor histidine kinases in B. subtilis. Most of the identical residues are C terminal to the phosphorylated histidine where the alpha 1 helix of Spo0F would be expected to make contact with the helical hairpin of the kinase (38). While this homology in the residues of the major interaction surface is strong presumptive evidence for functionality, other residues of the Spo0F surface contact other regions of the sensor histidine kinase and specificity for Spo0F cannot be ensured by homology in this region alone. Thus, functionality studies were required to truly identify those sensor histidine kinases with the capability to phosphorylate Spo0F.
An identical search was conducted on the proteins of three strains of the closely related but nonanthrax organism Bacillus cereus, to determine if B. anthracis carried sensor histidine kinases distinct from those of B. cereus. This comparison relating gene numbers of three sequenced B. anthracis strains to those of B. cereus is shown in Table Table4.4. There were no substantial additions or deletions in the sensor histidine kinase genes of these strains compared to B. anthracis based on the HisKA homology. The Ames strain gene designation will be used throughout.
The nine sensor histidine kinases identified by homology in the HisKA domain were characterized by a number of functionality and expression studies: (i) each kinase gene and its promoter region were cloned in the integrative vector pJM115 or pJM116 and tested for complementation of sporulation kinase mutants of B. subtilis; (ii) the promoter region from each kinase gene was tested for its ability to drive the expression of the lacZ gene in B. subtilis and in B. anthracis; (iii) each kinase gene was inactivated in B. anthracis and the mutant strain was tested for sporulation; (iv) comparative analysis of the protein encoded by each kinase gene was extended to all sequenced B. anthracis and B. cereus strains.
Several in vivo approaches were taken to test whether the sensor histidine kinases identified by homology were functional in the sporulation phosphorelay. It was first determined whether the cloned B. anthracis sensor histidine kinase could complement the sporulation deficiency of B. subtilis mutants devoid of kinase A or KinA and KinB, the two major sporulation sensor histidine kinases. This was accomplished by cloning the B. anthracis genes and corresponding upstream region derived from the Sterne strain of B. anthracis in the integration vector pJM115 or pJM116. These plasmids integrate into the amyE gene as a single copy when transformed into B. subtilis, thus placing the cloned gene on the chromosome under control of its own promoter. Functionality studies of the B. anthracis sensor histidine kinases were complicated by the extreme lethality of some recombinant constructs in E. coli. Thus, not all of the genes could be tested for complementation by simply cloning them in this vector. BA2644 resisted all attempts to clone it intact in plasmid pJM115.
B. anthracis sensor histidine kinases were integrated in the amyE gene in the wild-type strain and in strains mutant for kinA or kinA kinB, and sporulation frequencies were measured. In these studies sporulation was assayed in Schaeffer's and in LB medium, in which sporulation is normally repressed in B. subtilis. Several different phenotypes were observed (Table (Table55).
Inactive kinases were further tested by cloning their coding sequence downstream of the promoter for BA4223 in a vector constructed to integrate in the amyE gene. These constructions tested weak complementing genes for functionality when expressed from a B. anthracis promoter known to be active in B. subtilis and eliminated variability from innate promoters of B. anthracis expressing in B. subtilis. These results are also presented in Table Table5.5. The results for each kinase are discussed below.
The activity of the promoters for several of the B. anthracis sensor kinases was assayed by cloning the upstream and 5′ ends of the genes as a lacZ transcriptional fusion in plasmid pJM115, a vector which places the fusion construct in the amyE gene as a single copy in B. subtilis. The same fragments were placed in plasmid pTCVlac and transformed in the 34F2 strain for assay of promoter activity in B. anthracis. The results of these analyses are shown in Fig. Fig.33 and and4.4. The results for each kinase are discussed below.
Genes for each of the sensor histidine kinases were inactivated by insertion of a spectinomycin resistance cassette as described in Materials and Methods. Briefly, the spectinomycin (Spc) cassette was inserted in the open reading frame of each gene while cloned in the temperature-sensitive vector pORI-Cm. These constructions were transformed by electroporation into strain 34F2. Transformants in which the Spc cassette had recombined by double crossover in the chromosomal copy of the sensor histidine kinase gene with elimination of the plasmid were sought. The structure of the recombinant chromosome was verified by several control PCRs.
The effect of insertional inactivation of each gene on sporulation was determined in LB medium and in Schaeffer's sporulation medium. All of the strains showed high levels of sporulation after 48 h of incubation. At 24 h, however, there was a noticeable delay in sporulation in several strains, especially the strain with gene BA2291 inactivated. This effect was overcome by the expression of BA2291 in trans from the gene cloned on the multicopy plasmid pHT315 (2) (data not shown). Experiments to quantitate the level of sporulation in liquid cultures of B. anthracis are complicated by the sporulation of these organisms in long filaments, and free spores are often obtained only after several days of incubation. Thus, a delay in sporulation in strains is a visual qualitative observation. The important point is that deletion of any single kinase does not strongly affect sporulation, unlike the situation in B. subtilis, where a kinA inactivation drops the sporulation frequency to 5% of wild-type levels.
Sensor histidine kinase BA4223 is the most highly active of the B. anthracis kinases in complementation of the sporulation deficiency of both kinA mutant and kinA kinB double mutant strains (Table (Table5).5). In addition it is the only B. anthracis sensor kinase that remains active in complementation in LB medium or in the presence of glucose, in which B. subtilis does not normally sporulate (Table (Table55 and data not shown). The promoter for this gene is the most active of all of the promoters tested in B. subtilis (Fig. (Fig.3)3) and the second most active in B. anthracis (Fig. (Fig.4).4). In both bacteria the promoter is most active during the transition phase and early stationary phase when sporulation is induced.
Sensor histidine kinase BA2291 is capable of partial complementation in the B. subtilis kinase-deficient strains, but in multicopy in plasmid pHT315 it completely prevented sporulation in the wild-type B. subtilis strain JH642 while exhibiting the typical large transparent Spo0 colony phenotype on solid sporulation medium (Table (Table5).5). The interpretation of this phenotype is that BA2291 kinase acts as a phosphatase on the sporulation phosphorelay and probably at the level of Spo0F. Biochemical characterization of BA2291 activity is under way in our laboratory, and it will be described in an independent report. Transcriptional analyses revealed that the BA2291 promoter is the most active of all sensor histidine kinases in B. anthracis, in which it is mainly active after the transition state and at the early hours of stationary phase (Fig. (Fig.4A).4A). This promoter is also highly active in B. subtilis in approximately the same time frame.
Sensor histidine kinase BA1351 completely prevented sporulation in the wild-type B. subtilis strain when expressed in single copy from the amyE locus under the promoter for BA4223 (Table (Table5).5). The phenotypic interpretation is the same as that for BA2291. The promoter for this kinase is moderately expressed in B. anthracis with a peak of expression at the early hours of stationary phase (Fig. (Fig.4B).4B). In B. subtilis, expression of BA1351 is also moderate, it is maximal early in exponential phase, and it essentially stops by the beginning of stationary phase (Fig. (Fig.3B).3B). This may explain why, when expressed in single copy in the amyE gene from its own promoter, BA1351 did not seem to inhibit sporulation at a quantitatively measurable level. A differential transcription mechanism must regulate expression from this promoter in the two organisms.
Expression of the sensor histidine kinase BA1356 in the B. subtilis kinase-deficient strains resulted in complementation of the sporulation defect albeit at a much lower level than that observed with BA4223 (Table (Table5).5). It is expressed in B. subtilis at moderate levels (Fig. (Fig.3B),3B), but placing the gene under control of the highly expressed promoter of BA4223 did not increase its ability to complement sporulation (Table (Table5).5). The moderate activity of this kinase appears not to be due to its transcription level in B. subtilis. Its transcription in B. anthracis increases throughout growth (Fig. (Fig.4B4B).
Weak complementation by the sensor histidine kinase BA5029 could be enhanced by expression of its gene under the BA4223 promoter (Table (Table5).5). The promoter for BA5029 is very weak in B. subtilis (Fig. (Fig.3C)3C) but is moderately expressed in B. anthracis (Fig. (Fig.4B4B).
The HisKA domain of the sensor histidine kinase BA2636 is very highly homologous to that of kinase A of B. subtilis (Table (Table3),3), yet the proof of its function remains enigmatic. The cloned gene is nonfunctional in complementation (Table (Table5)5) and is not expressed in B. subtilis or B. anthracis (Fig. (Fig.3C3C and and4B).4B). The gene has resisted being cloned under the BA4223 promoter in E. coli, perhaps because of some lethality to this organism.
The ortholog of BA2636 in B. cereus strain ATCC 10987 is gene BCE2662. The protein encoded by BCE2662 is 96 amino acids longer at the N-terminal end than its B. anthracis ortholog. This is also the case for the ortholog in B. cereus ATCC 14579. In addition the B. cereus translation start codon is favorably situated with a ribosome binding site. This frame shift to shorten BA2636 occurs in codon 60 (relative to the B. cereus ortholog) and is conserved in all of the available sequenced B. anthracis strains including Sterne, A2012, Ames, Ames 0581, A1055, Australia 94, CNEVA-9066, and Vollum. This mutational conservation eliminates the trivial possibility of a sequencing error. The BA2636 ortholog found in the B. cereus strain G9241 is also frameshifted but at codon 351 and not at codon 60. Thus, the inability to observe complementation with BA2636 for sporulation may be due to a mutation in this gene in B. anthracis as well as its nonexpression in B. subtilis. Attempts to clone intact BCE2662 in E. coli were unsuccessful. The presumed promoter region for BA2636 was transcriptionally inactive in both B. subtilis and B. anthracis (Fig. (Fig.3C3C and and4C4C).
The sensor histidine kinase BA3702, with high homology to the HisKA domain of kinase A, is essentially nonfunctional in complementation of B. subtilis kinase mutants (Table (Table5),5), yet it is expressed well in B. subtilis and very moderately in B. anthracis (Fig. (Fig.3C3C and and4B).4B). The inability of BA3702 to complement for sporulation is not due to a frameshift in any B. anthracis strain. Strangely all 510 amino acids of this protein are identical in all of the sequenced B. anthracis strains, whereas the B. cereus orthologs have up to 30 missense mutations, only some of which are the same in the various strains analyzed. Three possibilities may be entertained for its lack of complementation: this sensor histidine kinase may not be used to phosphorylate Spo0F, the B. anthracis proteins are uniformly inactive compared to B. cereus, or the proteins are inactive because of lack of signal activation in B. subtilis. Attempts to clone the B. cereus ortholog were not successful.
We were unable to clone the gene under the BA4223 promoter in E. coli to test if overexpression would increase its activity.
The BA1478 gene was found to be inactive in complementation for sporulation and was expressed at very low levels in the B. subtilis kinA strain or the kinA kinB strain (Fig. (Fig.3C3C and data not shown); the β-galactosidase levels were slightly higher than the ones obtained with the empty vector pJM115 (Fig. (Fig.3C).3C). The promoter fragment for gene BA1478, however, did drive β-galactosidase expression in B. anthracis (Fig. (Fig.4B).4B). This differential expression may be due to the multicopy nature of the construct in B. anthracis, or perhaps additional factors missing in B. subtilis are required for its transcription.
A comparative analysis of gene BA1478 revealed another frameshift situation. In B. cereus strains G9241, ATCC 10987, and ATCC 14579 the orthologs to BA1478 are 423 amino acids in length. In the A2012, Ames, Ames 0581, and CNEVA-9066 strains a mutation occurs at codon 330 (B. cereus numbering), resulting in a frameshift inactivation. However, the frameshift does not exist in B. anthracis strains Australia 94, Vollum, and A1055. The Sterne strain, from which the gene was cloned in this study, carries the frameshift, and this may explain why it is unable to restore sporulation in complementation experiments.
BA2644 remains untested in complementation due to some lethality problem when attempts were made to clone the gene from either B. anthracis or B. cereus ATCC 10982 in pJM115. Furthermore, the position of its promoter is unclear and it may be cotranscribed with the gene upstream of it. However, the putative promoter region for the upstream gene also proved impossible to clone intact. This gene is intact in most B. cereus strains except ATCC 14579 and in B. anthracis A2012, where it is frameshifted.
This study of the sensor histidine kinases of B. anthracis with potential to induce sporulation revealed nine such proteins with homology to the histidine kinase (HisKA) region of B. subtilis kinase A. Four of the nine possible sporulation sensor histidine kinases were found not to be active in sporulation. The genes for two of these, BA2636 and BA1478, were promoter inactive or very poorly active, respectively, in B. subtilis and were frameshifted in B. anthracis. One gene, BA3702, has a promoter that is functional in both B. subtilis and B. anthracis but does not complement sporulation in B. subtilis. The BA2644 gene could not be tested in these conditions.
The BA4223 sensor histidine kinase is the most highly active of the B. anthracis kinases in complementation of the sporulation deficiency of both kinA mutant and kinA kinB double mutant strains. In addition it is the only B. anthracis sensor kinase that remains active in complementation in LB medium or in the presence of glucose, in which B. subtilis does not normally sporulate (Table (Table5).5). The promoter for this gene is the most active of all of the promoters tested in B. subtilis and the second most active in B. anthracis. In both bacteria the promoter is most active at the transition phase and early stationary phase, when sporulation is induced. However, inactivation of this sensor histidine kinase in the Sterne strain 34F2 did not significantly alter the sporulation of this strain. Thus, there must be mechanisms for keeping this highly expressed kinase under control in B. anthracis, especially in vivo (i.e., in the infected host).
Two of the sensor histidine kinases, BA1356 and BA5029, showed weak complementation of sporulation in B. subtilis. Expression of the BA1356 sensor histidine kinase in the B. subtilis kinase-deficient strains results in complementation of the sporulation defect, albeit at a much lower level than that observed with BA4223. The gene is expressed in B. subtilis at moderate levels, but placing the gene under control of the highly expressed promoter of BA4223 did not increase its ability to complement sporulation. The moderate complementation activity of this kinase appears not to be due to a low transcription level in B. subtilis.
Weak complementation by the BA5029 sensor histidine kinase was enhanced by expression of its gene under the BA4223 promoter. The promoter for BA5029 is very weak in B. subtilis but is moderately expressed in B. anthracis. The poor complementation by both of these kinases may result from a lack of activating signals in B. subtilis.
Two sensor histidine kinases, BA1351 and BA2291, were inferred to be involved in sporulation by their strong inhibition of sporulation when cloned in B. subtilis. BA1351 may be an ortholog of the B. subtilis kinase KinD. The BA2291 kinase has no ortholog in B. subtilis and is found in all B. anthracis and B. cereus strains. There are several possible explanations for the inhibitory activity of these kinases on sporulation. They may act as phosphatases on phosphorylated Spo0F or Spo0A, or they may sequester one or both of these proteins, preventing access by B. subtilis sporulation kinases. While many two-component sensor histidine kinases have phosphatase activity as a facet of their control of response regulator phosphorylation, the sporulation kinase KinA of B. subtilis does not possess this activity, and it can be argued that phosphatase activity is not advantageous for a phosphorelay in which several kinases can phosphorylate a common response regulator. Perhaps the B. subtilis environment lacks signals that normally activate these kinases or they exist in complexes with other proteins in B. anthracis, which keeps their phosphatase activities in check. Deletion of the BA2291 gene in B. anthracis has the most effect of any deletion on the ability of B. anthracis to sporulate, suggesting that it may play a role in sporulation.
The HisKA domain of the BA2636 sensor histidine kinase is very highly homologous to that of kinase A of B. subtilis, yet the proof of its sporulation function remains enigmatic. The cloned gene is nonfunctional in complementation and is not expressed in B. subtilis or B. anthracis. The gene has resisted being cloned under the BA4223 promoter in E. coli perhaps because of some lethality to E. coli. However, the reading frame of this sensor kinase in all B. anthracis strains is N terminally truncated compared to this same protein in B. cereus strains. The exception to this is B. cereus strain G9241, where the gene is inactivated by a frameshift at amino acid 351.
The common thread between B. cereus G9241 and the B. anthracis strains is the presence of plasmid pXO1 and the acquisition of pathogenicity (9), which raises the possibility that inactivation of BA2636 is required for successful pathogenesis by these bacteria perhaps to stabilize the replication or expression of some genes from pXO1. There is precedent for sensor histidine kinase/response regulator control of plasmid stability. The DpiA response regulator of E. coli binds at sites of replication initiation of plasmid pSC101 and influences DNA replication and plasmid inheritance, suggesting a mechanism by which environmental stimuli can regulate chromosomal and plasmid dynamics (20). While it is believed that the target of BA2636 activity is Spo0F, which does not bind DNA, it is possible that this histidine kinase is capable of phosphorylating a response regulator with that ability.
There remains the observation that while B. anthracis grows well in the body it does not sporulate there (22). It seems possible that efficient sporulation may be in diametric opposition to successful pathogenesis. Macrophages have the ability to take up and destroy spores, but encapsulated vegetative cells are refractory to this defense mechanism. Any mechanism that would shift the in vivo bacteria from sporulation to vegetative growth would select for increased pathogenicity. Destruction of the ability to sporulate entirely by, for example, inactivating Spo0A would not be advantageous since this would lead to the accumulation of the AbrB transition state regulator and the repression of anthrax toxin synthesis (29) and would put the cells at a serious disadvantage for survival ex vivo. Thus, it seems that the best course would be to dampen sporulation initiation in the body while still retaining the capacity to sporulate in other environments. Pathogenesis-induced prevention of sporulation by sensor histidine kinase inactivation might explain the BA2636 inactivation, the common but not universal inactivation of kinase BA1478 in B. anthracis strains, and the seeming inactivity of sensor histidine kinase BA3702 in sporulation despite its strong homology to the HisKA domain of kinase A.
The relationship between the regulatory processes emanating from the pXO1 plasmid and the regulation dynamics of the cell in growth and sporulation is clearly very complex. Inactivation of the plasmid-encoded AtxA regulator not only prevents toxin gene expression but also increases sporulation and makes the mutant grow poorly on minimal medium (7, 8). There is a complex interaction between the AtxA, PlcR, and PagR regulators on both plasmid and chromosomal genes in addition to effects on sporulation (18, 19). These relationships are now only superficially understood, but clearly the acquisition of plasmid pXO1 has serious regulatory ramifications (19). In addition to the inactivation of the PlcR regulator in all B. anthracis strains examined (31), the inactivation of several probable sporulation sensor histidine kinases in these strains compared to B. cereus suggests that regulatory interactions in these signal transduction systems are substantially altered by pXO1.
The original intent of these studies was to define the sensor histidine kinases involved in sporulation in B. anthracis and to understand this developmental process in a human pathogen. In B. subtilis the activation of sporulation by phosphorylated Spo0A is a two-level process. Low-level phosphorylation of this transcription factor is carried out by one or more of three kinases, KinC, KinD, and KinE. The effect of these kinases on cellular regulation is to cause the repression of the transition state regulator AbrB and thus release inhibition of the transcription of a large number of genes, which in B. anthracis include the virulence toxin genes. High-level phosphorylation and sporulation require the activity of KinA and KinB. In B. anthracis at least two of the kinases closely related to KinA and KinB, BA2636 and BA1478, are not active and inactivation of any of the other kinases does not result in serious sporulation deficiency. At least in laboratory culture, sporulation in B. anthracis appears to result from the activity of several sensor histidine kinases, for which the deletion of any one activity has little discernible effect.
This research was supported in part by grants AI055860 and AI052289 from the National Institute of Allergy and Infectious Diseases and grants GM19416 and GM55594 from the National Institute of General Medical Sciences, National Institutes of Health, United States Public Health Service. Oligonucleotide synthesis and DNA sequencing costs were underwritten in part by the Stein Beneficial Trust.
†Supplemental material for this article may be found at http://jb.asm.org/.
‡Paper no. 17150-MEM from The Scripps Research Institute.