Edited by Simon Cutting.
Molecular insights into the initiation of sporulation in Gram-positive bacteria: new technologies for an old phenomenon
Version of Record online: 9 JAN 2006
FEMS Microbiology Reviews
Volume 29, Issue 2, pages 281–301, April 2005
How to Cite
Stephenson, K. and Lewis, R. J. (2005), Molecular insights into the initiation of sporulation in Gram-positive bacteria: new technologies for an old phenomenon. FEMS Microbiology Reviews, 29: 281–301. doi: 10.1016/j.fmrre.2004.10.003
- Issue online: 9 JAN 2006
- Version of Record online: 9 JAN 2006
- Received 20 August 2004, Revised 13 October 2004, Accepted 19 October 2004
- Signal transduction;
- Sensor histidine kinase;
- Response regulator;
- Bacillus subtilis
- Top of page
- 5Concluding remarks
The last decade has witnessed extensive, and widespread, changes in scientific technologies that have impacted significantly upon the study of the life sciences. Arguably, the biggest advances in our comprehension of simple and complex biological processes have come as a consequence of obtaining the complete DNA sequence of organisms. It is likely that we will become accustomed to hearing of quantum leaps in the study and understanding of the biology of higher eukaryotes in the coming years, now that (near) complete genome sequences are available for man, mouse and rat. In this review, we will discuss the impact of genome sequence data, and the use of new scientific technologies that have emerged largely as consequence of the availability of this information, on the study of the master regulator of sporulation, Spo0A, in low G + C Gram-positive endospore-forming bacteria.
- Top of page
- 5Concluding remarks
Despite being considerably less complex than sophisticated multi-cellular organisms, bacteria can exploit most, if not all, environments on this planet, indicating that they are amongst the most highly evolved organisms known. The evolutionary processes that have allowed bacterial species to disseminate themselves into a wide range of niches, extreme or otherwise, have been driven by the ability to adapt in specific and defined ways to their environment. A key element of the adaptive responses of any bacterium is the ability to sense and respond to environmental change by altering the pattern of gene expression in a coordinated manner.
No matter which ecological niche is being exploited, the environmental conditions (such as nutrient availability) must be suitable to promote bacterial growth and division. When conditions become disadvantageous for growth, certain bacteria have evolved specific mechanisms to ensure the survival of the species. The most drastic response of members of the genera Bacillus and Clostridium to signals that indicate their immediate environment has deteriorated to a point where survival of vegetative cells is threatened is the production of an endospore (Fig. 1). During the sporulation process vegetative growth is abandoned in favour of an asymmetric round of cell division that yields two compartments of unequal size, a larger mother cell and a smaller cell, the prespore. Although at this stage, the two cells contain identical chromosomes, they are programmed to express quite different subsets of genes that ultimately shape their respective fates. Therefore, sporulation is arguably the simplest example of cellular differentiation (for reviews, see for example [1–4]). The prespore becomes engulfed by the mother cell, which collaborates in the spore's maturation, the final stage of which requires lysis of the mother cell (Fig. 1). The spore will survive extreme environmental conditions, lying dormant and awaiting the restoration of favourable growth conditions. As sporulation has a heavy energetic penalty and ultimately kills the mother cell, the initiation of the process is kept under stringent control by an expanded ‘two-component’ signal transduction system called a phosphorelay (Fig. 2). The phosphorelay integrates multiple stimuli to ensure that the cell sporulates only when all other survival strategies have been exhausted [2,5].
Two-component systems, probably the most prevalent signal transduction system found in bacteria, are minimally comprised of two matched, conserved elements, a sensor protein histidine kinase and a response regulator, which together perceive and respond to signals [6,7]. Response regulators are commonly, but not exclusively, transcriptional regulators. In response to specific signal ligands, the sensor kinase autophosphorylates on a conserved histidine residue in a reaction utilizing ATP, and the phosphoryl moiety is subsequently transferred from the histidine to an aspartic acid residue in the cognate response regulator. The phosphorylation of the response regulator elicits a conformational change within it that activates its latent biochemical function, which is frequently the direct modulation of gene expression by activating or repressing transcription. The modularity of this basic scheme permits elaboration to suit specific signalling pathways, whilst maintaining core structures and functions. Sensor kinases are usually embedded in the cytoplasmic membrane, with an extracellular sensing domain coupled to intracellular histidine phosphotransferase and catalytic domains.
1.1The sporulation phosphorelay
For sporulation initiation, the key event is the accumulation of sufficient quantities of the response regulator Spo0A in its activated, phosphorylated form . Once phosphorylated, Spo0A represses genes expressed during post-exponential growth and activates genes required for sporulation. Phosphorylation of Spo0A is achieved by the multi-component phosphorelay , which is an expanded variant of the two-component signalling module (Fig. 2). The sporulation phosphorelay was first discovered in Bacillus subtilis and has since been the subject of extensive and on-going study in this organism. Multiple sensor kinases feed phosphoryl groups into the phosphorelay. Each sensor kinase recognizes a specific signal and responds by catalyzing the ATP-dependent phosphorylation of a conserved histidine residue. The phosphoryl group is subsequently fed into the phosphorelay by the sensor kinase phosphorylating a single domain response regulator, Spo0F, on a conserved aspartate residue. The phosphoryl group is then passed from the aspartic acid on Spo0F ? P to a histidine side chain on Spo0B, a phosphotransferase. The final step in the relay is the phosphorylation of Spo0A by Spo0B ? P, to activate its transcriptional properties . To maintain control, phosphoryl groups can also be drained from the phosphorelay by the activity of specific response regulator aspartyl-phosphate phosphatases. Spo0F ? P is the substrate for Rap phosphatases , whereas Spo0A ? P is dephosphorylated by Spo0E phosphatases .
The explosion in the amount of data made available from genome sequencing has impacted on a number of pre-existing scientific disciplines, whilst at the same time creating some apparently new branches of science. A recent search of the Internet revealed that almost 90 so-called ‘-omic’ scientific technologies now exist. But how are some of these Omics defined? The goal of genomics is to determine the complete DNA sequence for all the genetic material contained in an organism's complete genome. The definition of proteomics is the qualitative and quantitative comparisons of proteomes [PROTEin complement to a genOME] under different conditions to further unravel biological processes. Structural genomics (or structural proteomics) is the systematic and high-throughput effort to gain a complete structural description, by X-ray crystallography or NMR spectroscopy, of a defined set of molecules, ultimately for an organism's entire proteome. The genome-wide study of mRNA levels is called transcriptomics. In this review, in order to determine the impact of Omics on the study of the key regulator of sporulation, Spo0A, we have limited our discussion to those papers that have been published after November 1997, the date of publication of the genome sequence of B. subtilis.
The genomic sequencing of B. subtilis 168 was undertaken by a European-Japanese consortium, and published in 1997 . At that time, only eight other genomes had been sequenced, the first two, Haemophilus influenzae and Mycoplasma genitalium, were published two years earlier. The B. subtilis genome is a little over four mega basepairs in size and comprises about 4100 genes, approximately half of which have been duplicated and exist as paralogues. The presence of 10 prophages, or remnants thereof, indicates the important role that horizontal gene transfer has played during the evolution of this bacterium. At the time of the original analysis, some 40% of the encoded genes could not be assigned any function based on sequence similarity to other polypeptides held in protein databases. Accordingly, approximately 2800 genes were allocated gene names commencing ‘y’, to indicate that a function had yet to be ascertained. At the time of writing of this paper, the complete sequence of 134 microbial genomes can be accessed at TIGR's Comprehensive Microbial Resource (http://www.tigr.org) . B. subtilis was thus in the vanguard of genome sequencing projects. The sequencing of other members of the same, and related, genera have allowed genomes to be compared, and will be discussed in later sections of this review.
- Top of page
- 5Concluding remarks
Initially, the definition of proteomics applied only to the identification of proteins that had been separated by two-dimensional gel electrophoresis, but increasingly the definition of proteomics extends to the study of the properties of a protein such that a full, molecular understanding of its function can be determined. There is thus little to distinguish between biochemistry and proteomics. Pre-genomics, clues to the function of a protein might have been gleaned by genetics and molecular biology, and subsequently tested in vitro. However, in the post-genomic era, even those genes of unknown function can be investigated. If a gene-disruption leads to no clear phenotype, and the amino acid sequence bears no homology to any other protein of known function, how can the cellular role of the gene be determined? Protein sequence evolves more quickly than protein folding. If the three-dimensional structure of a protein has conformational similarity to the structure of an unrelated (by sequence) protein, then clues in its function might be obtained by structural comparison, which can then be tested back in the laboratory.
The derivation of the structures of proteins or domains of no clear function is a methodology that is commonly called ‘structural genomics’. However, the advances in practices that have been driven by the structural genomics community – in effect, an increase in throughput by automation – are now being applied by more biologically targeted ‘structural proteomics’ groups. The goals of these groups/consortia range from deriving all the structures in the cancer proteome, to the particular challenges posed by the structural biology of integral membrane proteins, to those that aim to solve all the protein structures from a single organism. At least 20 large-scale, well-funded structural genomics/proteomics groups and consortia have been established across the globe (http://www.rcsb.org/pdb/strucgen.html). High-throughput techniques have to be applied in order to automate as many aspects of the structural biology as possible, and many thousands of targets are initially selected, as the attrition rate is high. Although B. subtilis and Bacillus anthracis are amongst the model systems under investigation in this way, structural proteomics has yet to have any discernable impact on the study of the initiation of sporulation in B. subtilis and its close relatives. Many of the proteins that regulate the phosphorelay have been selected as targets by the structural genomics/proteomics community (including orthologues of some known structures!), but at the time of writing, only RapC, RapG and RapJ have been crystallized, and their structures have yet to be determined. Thus it remains to be seen whether structural genomics/proteomics will impinge significantly upon the study of the initiation of sporulation. More hypothesis-driven structural and proteomics studies of the phosphorelay have revealed interesting features of the proteins involved. Since the publication of the genome sequence of B. subtilis, structures of Spo0B , the Spo0F:Spo0B complex , N-Spo0A , N-Spo0A ? P , C-Spo0A  and the C-Spo0A:DNA complex  have been derived by protein crystallography, in addition to the previously derived structures of Spo0F by both crystallography  and NMR . In the main, these structures have provided the framework for a number of subsequent in vitro and in vivo studies that have illuminated how the sporulation phosphorelay and other related pathways function.
2.1Phosphorylation of Spo0A
Almost all response regulators have a modular domain organisation comprised of an N-terminal receiver domain with the phosphorylatable aspartate and a C-terminal effector domain that generally mediates DNA binding. The few exceptions are the simple receiver domain only found in Spo0F and CheY, which have specific roles in the sporulation phosphorelay and chemotaxis, respectively. The structures of a variety of response regulator receiver domains have been solved since the first determination of the structure of CheY, more than a decade ago . Globally, these structures have the same doubly wound (β/α)5-fold and are in essence indistinguishable. The early structures were all determined in the non-phosphorylated state, as the lability of the phosphorylated form of response regulators has made structural analysis non-trivial. The molecular basis of Spo0A activation and, by extension, all response regulators, has been provided by the derivation of the crystal structure of the phosphorylated receiver domain of Spo0A (N-Spo0A ? P) from Bacillus stearothermophilus. This was the first structure of an activated response regulator receiver domain to be reported; several others have subsequently been determined [25–32] and much of the activation process described for Spo0A appears to be relevant to the wider response regulator OmpR-type family .
The crystallisation of N-Spo0A ? P was fortuitous as no attempts were made to phosphorylate Spo0A prior to crystallisation. It has been demonstrated since that Spo0A phosphorylation occurs during its preparation, either by intracellular small molecule phosphodonors, or by an Escherichia coli histidine kinase . The relative stability of Spo0A ? P might reflect the time-scale over which Spo0A must remain active during the initiation of sporulation. The overall fold of N-Spo0A was the same as other receiver domains, and the active site architecture was also conserved (Fig. 3). The calcium ion in N-Spo0A occupies a similar position to that of the magnesium in the native crystal structure of CheY.
Functionally significant conformational changes in Spo0A as a consequence of phosphorylation were deduced by comparison to an ensemble of ‘resting’ receiver domain structures, in addition to the domain-swapped form of N-Spo0A. The numbering scheme in the discussion below is that of Spo0A from B. subtilis. Three conserved aspartic acid residues, Asp10, Asp11, and the site of phosphorylation Asp56, which together comprise the ‘aspartyl pocket’ and the invariant Lys108 remain static upon phosphorylation. A near-planar, six-membered chelate ring is formed by the atoms of the phosphorylated aspartate, Oδ1, Cγ and Oδ2, the calcium ion, and one oxygen atom and the phosphorus from the phosphoryl group, as originally proposed in 1952 . The main chain amide nitrogens of Ala87 and Ile58, the Nζ of Lys108 and the adjacent calcium make hydrogen-bond contacts to the phosphoryl group, in addition to that made by the hydroxyl group of the fifth, characteristic conserved amino acid of receiver domains, Thr86 (Fig. 3). The side chain of this residue repositions towards the phosphoryl group, from a position pointing away from the active site. The conserved aromatic residue, Phe105, tracks the movement of the side chain of Thr86, and switches from a solvent-exposed position, to one that is buried (Fig. 3). This activation mechanism is termed the “aromatic switch”, and suggests the importance of the surface of Spo0A defined by the ‘α4–β5–α5’ face in signalling.
The significance of the ‘α4–β5–α5’ face in signal transduction in Spo0A is confirmed by the observation that alanine substitution of Tyr104 and Phe105 confers a sporulation-defective phenotype to B. subtilis. Suppressors of Tyr104Ala could not be found by the experimental regime, perhaps indicating the importance of this residue – invariant amongst Spo0A orthologues, despite being solvent-exposed – to the function of Spo0A. Of the suppressors to Phe105Ala, it is most straightforward to explain the mode of suppression by Thr94Met . To accommodate the bulky, Phe105 side chain during the aromatic switch, α4 shifts away from core of the protein, else Phe105 would clash severely with Thr94. The movement of α4 also causes a reorganisation of the structure of the β4–α4 loop , which might aid stabilisation of the aspartyl phosphate, by shielding it from the attack of water. Methionine is a larger amino acid than threonine, and one can imagine how this bulkier residue could perform the same role as an aromatic, in the case of the alanine-substitution of Phe105, in filling the hydrophobic void behind Thr86 on its re-positioning during phosphorylation.
2.2DNA recognition by Spo0A
Activated Spo0A can stimulate transcription from promoters that are either σA- or σH-dependent [38,39]. Derivation of the crystal structure of the effector domain of Spo0A (C-Spo0A) indicated that the overall fold of C-Spo0A is unique , not necessarily surprising since the amino acid sequence of Spo0A is only found in Spo0A orthologues. The structure comprises six α-helices, αA–αF, two of which (αC–αD) form a classical helix–turn–helix (HTH) DNA-binding motif. The region of Spo0A in the immediate vicinity of αE has been shown to be important in transcription activation from σA-dependent promoters [40–43], and this part of the structure forms a flexible section in close proximity to the recognition helix of the HTH. It is not yet clear whether the same region, or that defined by the extreme C-terminus of Spo0A, is required for σH-dependent transcription activation . The rest of the structure of C-Spo0A is a reasonably compact five-helical bundle. Patterns of highly conserved amino acids are clustered to one contiguous surface spanning the ‘recognition’ helix of the HTH DNA-binding motif and including flanking sequences from the αA–αB loop and αB, in addition to the αE–αF loop and αF (Fig. 4).
The molecular surface of C-Spo0A reveals that a convex protrusion is formed by the loop between helices αA and αB, the base of which is formed by residues His162 and Leu174 (Fig. 4). Mutation of these two amino acids (to arginine and phenylalanine, respectively) suppresses the sporulation-deficient phenotype imparted by the spo0A9V mutation, Ala257Val . Ala257 is situated on the opposite face of C-Spo0A, in a concave depression formed by residues from the αE–αF loop, αF and the αB–αC loop. Although His162 and Leu174 are not in intramolecular contact with Ala257, the surface complementarity of these two regions were demonstrated in the structure determination of the DNA-bound form of C-Spo0A . Here, two molecules of C-Spo0A form a tandemly repeated dimer, with the dimer interface being formed by the conserved residues in the αA–αB loop in one molecule, and αF in the other (Fig. 4). Hence sequence conservation in C-Spo0A has been maintained for functional reasons, either sequence-specific DNA recognition, or dimerisation (Fig. 4). Although His162/Leu174 and Ala257 are not in direct intermolecular contact in the C-Spo0A tandem dimer, it is apparent how amino acid substitutions in these positions will disrupt the local structure, and explain the suppressive effects. For instance, replacement of the buried Ala257 with valine will alter the disposition of αF with respect to the protein core, and mutation of His162 to the longer side chain of arginine may produce compensatory interactions with, for example, the main chain carbonyl of Phe236, or the side chain of the solvent-exposed carboxylate of Asp258.
DNA-binding is mediated by a series of conserved residues from the recognition helix, αD, and its surroundings. Perhaps surprisingly, only Glu213, Arg214 and Arg217 make direct contact to the DNA bases, to impart sequence-specificity on the interaction, and just three of the seven bases (positions 2, 4 and 5) in the canonical ‘0A’ box, 5′-TGTCGAA-3′, are contacted directly by these three amino acids. Mutation in Spo0A at Glu213 and Arg214 abolishes DNA-binding and transcription activation and repression [46,47]. This structure validates the findings that mutation of ‘0A’ boxes at various promoters at positions 4 and 5 that increase Spo0A activity restore the consensus sequence whereas mutations away from the consensus at positions 4 and 5 reduce Spo0A activity . All other contacts to the DNA are either via water molecules, or to the phosphate backbone. Interestingly, of the two molecules in the C-Spo0A dimer, only one forms contacts to a canonical ‘0A’ box; the other molecule binds to the linker in the synthetic oligonucleotide between ‘0A’ boxes. There is little to distinguish between sequence-specific and non-specific DNA binding by C-Spo0A. Intriguingly, tandemly arranged ‘0A’ boxes in the promoters of genes/operons that are regulated by Spo0A are rare, suggesting that although dimerisation (or multimerisation) of Spo0A ? P may be essential for the activation of Spo0A-dependent genes, binding of a Spo0A ? P dimer may only require one true ‘0A’ box. Formation of dimers/multimers in those promoters that do contain more than one upstream binding site may actually affect promoter strength.
2.3Maintenance of fidelity in two-component signalling
The histidine phosphotransferase, Spo0B, catalyses the reversible transfer of phosphoryl groups from Spo0F to Spo0A, a pre-requisite for information flow through the phosphorelay and the initiation of sporulation. The structure of Spo0B reveals a dimeric organisation (Fig. 5(a)), and each protomer comprises two domains . One domain mediates dimerisation (Fig. 5(a)), by forming a four-helical bundle that closely resembles the dimerisation/histidine phosphotransferase domains of other sensor kinases [48–50]. Each of the two substrate histidines are located on solvent-accessible surfaces on opposite sides of the helical bundle. The other domain resembles the catalytic domain of histidine kinases , but crucially lacks the five signature motifs that are required for ATP binding in these kinases (Fig. 5b). Hence, it is likely that the sporulation phosphorelay evolved partly through the duplication of genes for a histidine kinase/response regulator pair. The structure of Spo0F has the same fold as that of a canonical response regulator receiver domain, and maintains the conserved lysine and tri-aspartate cluster that are necessary for phosphorylation [22,23].
The mechanism of phosphoryl transfer between histidine phosphotransferases and response regulators, and how signalling fidelity is maintained in B. subtilis, has been revealed by the crystal structure of the complex between Spo0F and Spo0B . The phosphorylatable residues, His30 of Spo0B and Asp54 of Spo0F, are ideally aligned in the complex, poised for phosphoryl transfer to occur and little structural rearrangement occurs in either protein on complex formation. It is mostly the helical bundle of Spo0B that interacts with Spo0F, providing hydrophobic and hydrophilic contacts, stability and specificity (Fig. 6). Residues in Spo0F from the loops surrounding the aspartyl pocket and from α1 complete the molecular interface. Eight out of 10 of the hydrophilic residues in the Spo0B-binding region of Spo0F are conserved in Spo0A, and of the 22 Spo0F amino acids in van der Waals contact with Spo0B, 15 are maintained in Spo0A, which is consistent with Spo0B acting to shuttle phosphoryl groups between Spo0F and Spo0A.
The identity of some residues in the Spo0F:Spo0B interface had previously been determined by alanine-scanning mutagenesis of 80 residues of Spo0F, and subsequent mutant screening by sporulation efficiencies in vivo and phosphotransfer kinetics in vitro . Here, alanine mutation of residues Ile15, Leu18, Lys56, Ile57, Thr82, Tyr84, Glu86, Leu87, Phe106 and Ile108 produced a strong sporulation defective phenotype. The structure of the Spo0B:Spo0F complex revealed a remarkable correlation between the two methodologies; all but two of the amino acids predicted to be placed in the Spo0B-binding region of Spo0F were subsequently found to be in the Spo0B:Spo0F interface (Fig. 6). Of the two ‘outliers’, mutation to alanine of Ile57 appears to destabilize the conformation of the phosphorylatable aspartate (Asp54) and Glu86Ala is affected in the binding of Spo0F to KinA .
Structures of the Spo0B:Spo0F complex , and activated forms of Spo0F  and N-Spo0A  also reveal how dissociation of the response regulator from the complex with Spo0B might occur after phosphorylation. While the aromatic residue of the ‘aromatic switch’ (His101 of Spo0F) is not involved in intermolecular protein–protein contacts, the β3–α3 and β4–α4 loops of Spo0F make several hydrophobic contacts to Spo0B. Both these loops are significantly re-modelled on phosphorylation of Spo0A, and activation of Spo0F by the phosphoryl group analogue, beryllium fluoride. The β3–α3 loop becomes more closely associated with the response regulator, and is in essence pulled away from the Spo0B helical bundle. In contrast, the β4–α4 loop is rearranged towards Spo0B such that the side chain of Ala87 of Spo0A, which is conserved as a small residue in order to permit close approach of phosphotransfer partner proteins, sterically clashes with the imidazole ring of His30 of Spo0B. Together, these movements are likely to dissociate the newly phosphorylated response regulator from Spo0B.
One of the features of two-component signalling is the relative abundance of these protein pairs even within one organism, yet cross-talk between closely related non-cognate partners is avoided inside the cell. Sequence analysis of the major Spo0B binding determinants of Spo0F, α1, and loops β4–α4 and β5–α5, reveals that signal transduction fidelity is maintained by under-stated variations in sequences of response regulators . For instance, the chemical nature of amino acids Pro105, Phe106 and Ile108 in loop β5–α5 and Ile15, Leu18 in α1 and Met81 in loop β4–α4 are maintained in all B. subtilis response regulator sub-families except the NarL group. In contrast, the other residues from the β4–α4 loop, Tyr84, Gly85, Glu86 and Leu87, are non-conserved, and this particular region of the response regulator presumably imparts kinase specificity on the binding reactions. What is striking though, is that the same surface of Spo0A is used for the recognition of Spo0B and also the Spo0E phosphatase . Furthermore, the same surface of Spo0F is involved in binding to the KinA sensor histidine kinase [54,55], Spo0B  and the RapB phosphatase . Although Spo0B is most likely a vestigial histidine kinase, Spo0E and the Rap families of phosphatases are completely unrelated to each other, and to Spo0B or sensor kinases. A complete understanding of these complex and subtle molecular recognition patterns awaits further studies of different interacting systems.
2.4Activation of Spo0A
Whilst an activation mechanism for Spo0A (and other response regulators) seems to have been revealed by structural analysis of the isolated receiver domain, an explanation of the transcriptional regulatory function of Spo0A has not emerged [57,58]. For instance, deletion of Asp75 in B. subtilis Spo0A renders Spo0A constitutively active , yet mutation of Asp75 to serine confers a sporulation-deficient phenotype on B. subtilis, and in vitro transcription stimulation was drastically reduced . It is therefore possible that the α3–β4 loop of the regulatory domain of Spo0A contacts and inhibits the attached effector domain until the inhibition is relieved by phosphorylation some 20 Å away. However, in the absence of structures of intact Spo0A, both in the resting and active conformations, other (proteomics) techniques must be used to answer this question.
One of the ways that energy barriers are overcome in catalysis is by increasing the local concentration of the reactants. In molecular recognition events, such as a transcription factor binding to target DNA, the local concentration of the protein can be increased by multimerisation. The DNA-bound form of C-Spo0A in the crystal describes a dimer . Whilst the domain-swapped structure of N-Spo0A also describes a dimer, this particular structure is a non-reversible low pH-induced artifact with little physiological relevance . Nonetheless, the evidence that dimerisation of Spo0A is an important feature of its activation is beginning to become overwhelming [21,36,61]. For instance, it is possible to purify monomers and low pH-induced dimers of a non-phosphorylatable mutant (D56Q) form of Spo0A . The locked, domain-swapped dimers of Spo0A (D56Q) were as effective in stimulating transcription from the spoIIG promoter as phosphorylated wildtype Spo0A, whereas the monomeric Spo0A (D56Q) counterparts were as ineffective in transcription stimulation as non-phosphorylated wildtype Spo0A . Furthermore, solution and sedimentation studies have revealed that non-phosphorylated Spo0A is a monomer, which becomes a dimer upon phosphorylation [36,61]. In the absence of DNA, the predominant dimerisation-determinants are restricted to the N-terminal domain of Spo0A . The sporulation-deficient phenotypes of Spo0A mutants Tyr104Ala and Phe105Ala are caused by an inability of these proteins to form dimers upon phosphorylation . Suppressor mutations that map to a proposed dimer interface (Fig. 7), based on the surface used by FixJ to dimerise , and by CheY to bind to FliM peptides , restore dimerisation . The residues in this surface, from the β4–α4 loop (88–90), consecutive turns of α4 (91 and 92, 94 and 95, 98) and β5 (104–107) are without exception either totally conserved, or are conserved within the genus Bacillus, and conserved within the genus Clostridium, but homologous between the two. For instance, Lys95 in the bacilli has a glutamate at the equivalent position in the clostridia. Similarly, in the bacilli Phe88, Glu91, Val98, Ile106 and Leu107, are replaced in the clostridia by valine, aspartate, isoleucine, valine and valine, respectively.
The conservation of these residues implies that this contiguous region of Spo0A has functional importance (Fig. 7). Indeed, some of the Spo0A sof mutants, which are affected in their interactions with sensor kinases and RNA polymerase map to similar regions . The most obvious explanation for the observation that B. subtilis cells expressing Spo0A ? P monomers have a sporulation-deficient phenotype, and cells with Spo0A ? P dimers are sporulation-proficient is that dimerisation is a key link between phosphorylation and DNA binding and transcriptional regulation. The N-terminal domain of Spo0A harbours the most significant dimerisation determinants [36,61], and the ability of this domain to impair Spo0A-directed transcription, and thus sporulation  could be a consequence of the formation of transcriptionally inactive, heterologous Spo0A:N-Spo0A dimers. However, it is not yet established whether the ‘α4–β5–β5’ face is used by Spo0A solely in the formation of new, intermolecular interactions (dimerisation), or also in the disruption of pre-existing intramolecular interactions (domain–domain rearrangements) on activation. The molecular details of the signal transduction pathway in Spo0A remain elusive.
- Top of page
- 5Concluding remarks
In addition to the genome sequence of B. subtilis 168 , which is archived and annotated at the SubtiList web server (http://genolist.pasteur.fr/SubtiList/) the genome sequences of Bacillus cereus ATCC 14579 , B. anthracis Ames , Bacillus halodurans C-125 , Oceanobacillus iheyensis HTE831 , Clostridium acetobutylicum ATCC 824 , Clostridium tetani E88  and Clostridium perfringens 13  have been completed. Other genome sequences, including eight strains of B. anthracis, one of B. stearothermophilus and also one of Clostridium difficile are still in progress. As these bacteria reside in diverse natural habitats, comparative genomics will be a powerful tool in deducing evolutionary links and establishing the genetic roots to, for example, halophilicity. In the following sections, we will discuss how comparative genomics has refined our understanding of the initiation of sporulation in anaerobes and pathogens, which are less tractable model systems than B. subtilis.
3.1Signal sensing and the initiation of sporulation
The sensor kinases are the most upstream elements of the phosphorelay and function as the transducers of signal-ligand binding. Signals regulating the activity of sensor kinases are thought to be small molecule effectors, the chemical nature of which may be a reflection of the deterioration of a variety of extracellular or intracellular parameters with the potential to affect cell viability. The sensors have a modular organisation and comprise minimally of two regions; a non-conserved N-terminal signal input region and a C-terminal catalytic region with consensus ATP binding and histidine phosphotransferase domains. In B. subtilis at least five sensor kinases, KinA and KinB , KinC , KinD  and KinE  have the capacity to activate the phosphorelay, although under normal laboratory conditions KinA, and to a lesser extent KinB, are the most efficient activators of the pathway.
Unlike classical two-component systems, the genes encoding the sporulation phosphorelay sensor kinases are not located next the cognate response regulator on the chromosome, and are therefore known as orphans . The strict conservation of the active site of sporulation sensor kinases and the orphan nature of their respective genes means that potential sporulation-specific sensor kinases can be identified in sequenced genomes with relative ease. In a previous study, these criteria were used to identify five sensor kinases of B. anthracis, B. halodurans and B. stearothermophilus with the capacity to phosphorylate Spo0F . At that time, the genome sequence of B. anthracis was unfinished and that of B. cereus was not available; we have now expanded this analysis to take into account the complete sequences of these organisms (all genome sequences were analysed via the TIGR website, http://www.tigr.org). Instead of the five potential phosphorelay sensor kinases identified previously, analysis of the now complete genome sequence of B. anthracis reveals a total of eight putative sporulation sensor kinases (Table 1). Somewhat surprisingly, we found that B. cereus has 11 sensor kinases that could potentially use Spo0F as their substrate and activate the phosphorelay (Table 1). This suggests that the range of signals used to initiate the developmental programme in B. cereus may be considerably more extensive than in the other bacilli. It is unlikely that traditional experimental genetic screens would have revealed this.
|Organism||Active site residuesa||Orthologues|
The analysis of the complete genomes of B. anthracis and B. cereus revealed the presence of orthologues of B. subtilis KinB and KinD in these organisms (Table 1). Furthermore, six of the B. anthracis sensor kinases are also found in B. cereus (Table 1) and the respective orthologues share greater than 80% amino acid identity. The presence of orthologous sensor kinases might be expected, given the very close evolutionary relationship of B. anthracis and B. cereus, and may also be a reflection of the fact that a number of the signals required to activate the sporulation pathways of these organisms are shared. The chemical nature of these common signals, or those sensed by the sensor kinases unique to each organism, is a mystery and is impossible to predict solely on the basis of amino acid sequence or domain organisation.
Each species has to sense a different range of signals to activate the phosphorelay pathway that are specific for the particular environmental niches that the different organisms have evolved to exploit. Consequently, there is generally little conservation in the signal input regions of different sporulation sensor kinases, either within or between species. In contrast, the catalytic regions are very well conserved in all cases to facilitate autophosphorylation and phosphoryl group transfer to Spo0F. Furthermore, despite their inherent diversity the signal input regions of sporulation sensor kinases frequently contain sub-domains with recognised sensory functions: a common feature is the presence of PAS domains . PAS domains are evolutionary conserved sensory modules that sense changes in, for instance, energy, redox or oxygen, and frequently contain small molecule signal ligands, such as FMN .
Signal sensing mediated by PAS domains is critical for the proper functioning of the sporulation phosphorelay in B. subtilis[76,77] and most likely all Bacillus species . Of the 14 PAS domains encoded by the B. subtilis genome, seven are involved in phosphorelay signalling and are distributed throughout three sporulation sensor kinases. For example, the major sporulation sensor kinase in B. subtilis, KinA, has three PAS domains located within the N-terminal signal input region. The most N-terminal of which, termed PAS-A, binds ATP with relatively high affinity and catalyses the exchange of the γ-phosphoryl group between ATP and nucleoside diphosphates in a mechanism that is thought to regulate the activity of the catalytic domain . Although there is an obvious link between the cellular energy pool and the initiation of sporulation in response to nutrient deprivation, the dissociation constant of the interaction of PAS-A with ATP (?20 μM) is well below the cellular concentration of ATP (?2 mM) indicating that the ATP concentration is not the parameter sensed directly by KinA. In every Bacillus species analysed to date, at least one of the putative sporulation sensor kinases contains a PAS domain, and frequently multiple PAS domains are present in the same protein. The reason for this apparent redundancy is unknown; perhaps each PAS domain senses a different cellular parameter.
3.2Conservation of the interaction surfaces of phosphorelay proteins in different spore-forming bacteria
Sequence data, combined with structural information gained largely from the analysis of the B. subtilis system, allows the initiation of sporulation via the phosphorelay to be evaluated in different organisms that have not been genetically analysed, or those for which the experimental tools do not exist. The genomes of all sequenced Bacillus species harbour genes encoding Spo0F, Spo0B and Spo0A orthologues that are readily identifiable on the basis of amino acid similarity and by comparison to the surrounding genes . Spo0F amino acids that make up the interaction surface with sensor kinases and Spo0B are highly conserved in orthologues in different Bacillus species, as are the Spo0B amino acids that make up the interaction surface with Spo0F and Spo0A. In particular, the amino acids immediately C-terminal to the ‘active site’ histidine of Spo0B are known to be most important for the interaction with the response regulator receiver domains of Spo0F and Spo0A , and those amino acids found on the molecular surface surrounding the phosphorylatable aspartate of Spo0F are absolutely conserved in sequence from different Bacillus species. This conservation implies that the region of the sensor kinase that is in contact with Spo0F must be highly maintained across these species in order to facilitate phosphoryl group transfer. These surfaces have resisted evolutionary change as the different species have diverged. The contacts and interactions of the proteins are highly conserved to ensure the proteins come together in the proper orientation to create the physico-chemical environment for phosphoryl group transfer between histidine and aspartate residues of the signalling partners. This ensures that the flow of information through the pathway, i.e., the relay of the phosphoryl moiety, is maintained.
The interaction of Spo0A with the ‘0A’ box, in or near the promoters of Spo0A-regulated genes is mediated by the αD recognition helix in the C-terminal effector domain [20,21]. Structural analysis of the DNA complexed structure of the isolated effector domain of B. subtilis Spo0A shows that αD lies in the major groove of DNA, perpendicular to the helical axis of the DNA . This structure reveals that nine amino acid residues are involved in DNA recognition and binding, three of which (Glu213, Arg214 and Arg217) make direct contacts with the bases of the ‘0A’ box . As these residues are strictly conserved in Spo0A from Bacillus and Clostridium species, it can be concluded that DNA recognition and binding by Spo0A occurs by the same mechanism in all spore-forming species. The presence of consensus ‘0A’ boxes upstream of putative Spo0A-regulated genes in all species analysed supports this. Therefore, as well as the protein–protein interactions, the protein–DNA interactions have also been conserved in the phosphorelays of different spore-forming organisms.
3.3Control of phosphoryl group transfer: phosphatases and anti-kinases
To facilitate accurate signal transduction, and retain control of the system, a key element in all two-component and phosphorelay signalling pathways is the reversibility of phosphoryl group transfer. In a classical two-component system, this may be achieved by an autophosphatase activity of the sensor kinase catalytic domain that can remove the phosphoryl group from the phosphorylated form of the cognate response regulator. However, in the phosphorelay, the activity of sensor kinases provides positive input into the pathway in the form of phosphoryl groups in response to signals promoting sporulation and no autophosphatase activity has been detected for any of these sensor kinases. Instead, the activities of specific aspartyl-phosphate phosphatases remove phosphoryl groups from the active site aspartates of phosphorelay response regulators. These phosphatases provide a mechanism for signals antithetical to sporulation to impact on the pathway. Consequently, the phosphorelay functions as a signal integration circuit in which the opposing activities of sensor kinases and aspartyl phosphatases influence the phosphorylation status of Spo0A.
In B. subtilis, two groups of protein phosphatases specific for phosphorelay response regulators have been identified, the Rap and Spo0E families [10,11]. It has been demonstrated that three members of the Rap family (RapA, RapB and RapE), and all members of the Spo0E family (Spo0E, YisI and YnzD) specifically dephosphorylate Spo0F ? P and Spo0A ? P, respectively, limiting the flow of phosphoryl groups through the pathway. Although both Rap and Spo0E phosphatases have the same biochemical activity, perform similar cellular functions and interact with protein domains exhibiting essentially the same fold inside the cell, they show no conservation at the amino acid level.
Like the main components of the sporulation phosphorelay, genes encoding proteins homologous to the Rap and Spo0E phosphatases can also be identified in sequenced Bacillus genomes. This is probably not surprising given the strict conservation of Spo0F and Spo0A in all Bacillus species. Although putative sporulation-associated aspartyl-phosphate phosphatases can be identified on the basis of sequence homology and the presence of characteristic amino acid motifs, direct orthologues of the major B. subtilis sporulation-associated aspartyl phosphatases, RapA and Spo0E, are not obvious in any of the bacilli. Since aspartyl phosphatase gene expression and enzymatic activity are regulated by physiological conditions contra to sporulation, the absence of direct RapA and Spo0E orthologues may reflect differences in the negative signals that impact on the pathway in the other bacilli. This situation is reminiscent of the diversity observed in signal sensing by phosphorelay sensor kinases of different organisms described above. With the exception of C. tetani, neither Rap nor Spo0E phosphatases can be found in Clostridial genomes on the basis of homology to the B. subtilis proteins (see below).
Further control over the flow of phosphoryl groups through the phosphorelay is attained by the action of the anti-kinases such as KipI , and the unrelated Sda . Sda acts on the phosphorelay by inhibiting autophosphorylation of KinA, and more specifically by forming a barrier to intramolecular communication and phosphotransfer between the catalytic and phosphotransfer domains of KinA . The residues of Sda that bind to the phosphotransfer domain of KinA are conserved in Sda orthologues, which are only found in the bacilli, and cluster onto one face of the helical hairpin that describes the simple structure of Sda . In contrast, the molecular basis of KinA–KipI interactions, and the molecular mechanism of KipI antagonism by KipA are completely uncharacterised. These two proteins may function to regulate other two-component histidine kinases and are found in the genomes of many bacteria, including the asporogenous Gram-negative E. coli, but, interestingly, not in the clostridia.
3.4Insights into the initiation of sporulation in anaerobic spore-forming bacteria
Spo0A is known to regulate sporulation in C. acetobutylicum and C. perfringens and most likely all Clostridium species. The cellular events that lead to the activation of Spo0A in the genus Clostridium differ from those that occur in Bacillus. For instance, obvious orthologues of Spo0F cannot be identified in sequenced Clostridium genomes on the basis of amino acid sequence or chromosomal location. It therefore seems to be possible that the signal transduction pathway regulating entry into the developmental programme in clostridia has not expanded by evolution into a phosphorelay. Spo0A might thus be the direct phosphorylation recipient for one or more sensor kinases in the clostridia. As multi-component phosphorelays are generally regarded as signal integration circuits that allow many signals to be processed and interpreted, the absence of a phosphorelay suggests that the initiation of sporulation in anaerobes is under less stringent control than in Bacillus species. However, the possibility that other, as yet unidentified proteins function to transfer phosphoryl groups from sensor kinases to Spo0A cannot be ruled out. Consequently, it is difficult to identify sensor kinases with the potential to initiate sporulation on the basis of sensor kinase active site homologies alone. Although a number of orphan sensor kinase genes can be identified in the genome sequences of Clostridium species, without experimental evidence to show phosphoryl group transfer to Spo0A, it is impossible to deduce which of the sensor kinases are involved in the initiation of sporulation in the clostridia.
When the sequences of Spo0A from different Bacillus and Clostridium species are compared, the greatest variation exists in the receiver domain. A number of residues can be identified in the receiver domain where an amino acid is strictly conserved across the bacilli, but is replaced by different residues in the clostridia. These amino acids cluster to regions of Spo0A that are known to mediate the interaction with histidine-containing phosphotransferase domains in other response regulators. It would seem highly likely that these amino acids promote interactions with the phosphoryl group donors that are specific to Clostridium species. Whether these donors are sensor kinases, or discrete phosphotransferase proteins analogous by function, if not by sequence, to Spo0B remains to be determined experimentally.
The gene encoding the third component of the phosphorelay in the bacilli, spo0B, is located immediately upstream from the gene that codes for an essential G protein known as Obg. Although Spo0B orthologues in the bacilli generally exhibit less overall amino acid identity than the other elements of the pathway, the major interaction surface, the α1-helix, and the phosphorylatable histidine are well conserved . Spo0B orthologues cannot be identified in any of the genomes of Clostridium species on the basis of amino acid identity to the Bacillus proteins. However, in C. tetani there is a gene immediately upstream of a putative obg gene that encodes a hypothetical protein with a region that exhibits similarity to α1 of B. subtilis Spo0B. This protein (designated NT02CT2179 by TIGR) contains a putative phosphorylatable histidine, and six of the 12 residues of B. subtilis Spo0B that are known to make contact with Spo0A and Spo0F are identical. In the genomes of C. perfringens and C. acetobutylicum, equivalents to the obg gene are present in the same chromosomal location, but a gene corresponding to NT02CT2179 is missing. Whether NT02CT2179 is a functional phosphotransferase in C. tetani, or merely a non-functional evolutionary vestige is open to speculation.
Rap phosphatase homologues cannot be identified in the genomes of any of the Clostridium species analysed, consistent with the absence of their cognate protein substrate, Spo0F. Similarly, despite the presence of highly conserved Spo0A proteins, Spo0E phosphatases do not appear to be present in C. acetobutylicum or C. perfringens; this is consistent with the theory that the control of sporulation initiation is less complex in these organisms . Somewhat surprisingly, two Spo0E-like proteins can be identified in C. tetani. These putative phosphatases, designated NT02CT1455 and NT02CT1370, are more homologous to B. subtilis Spo0E than to YisI or YnzD, exhibit 63% amino acid identity with each other and carry the signature sequence motif characteristic of the Spo0E family of phosphatases. However, the genes surrounding either of these putative phosphatase genes do not correspond to those adjacent to spo0E, yisI or ynzD on the B. subtilis chromosome. The functional significance of dual Spo0E-like phosphatases in C. tetani and their absence from other Clostridium species is intriguing. Perhaps it is more than coincidence that C. tetani also possesses a putative phosphotransferase protein (NT02CT2179) with the potential to act as a phosphoryl group donor for Spo0A?
3.5Involvement of the phosphorelay in the pathogenic lifestyle of spore-forming bacteria
Virulence and pathogenicity are the properties of some bacteria that encounter suitable host environments. For a pathogenic species to persist and to survive at specific host sites, the bacteria must be able to sense and respond to signals emanating from their immediate environment. One way that bacteria accomplish this is via two-component and phosphorelay signal transduction systems. These signalling pathways are integral elements of the virulence responses of most, if not all, bacterial pathogens and are recognised targets for the development of novel antimicrobial drugs [83–87].
The sporulation phosphorelay controls the initiation of development and a variety of other post-exponential phase phenomena. In B. subtilis, activated Spo0A influences the expression of more than 500 genes , 121 of which are regulated directly, the remainder indirectly . Consequently, Spo0A has a profound influence on global patterns of gene expression. The sporulation phosphorelay is a complex signal transduction pathway with pleiotropic cellular effects. Therefore, it is not surprising that pathogenic spore-forming bacteria have adapted the phosphorelay to regulate not only sporulation but also the virulence properties that allow them to exploit, and persist, within the environment provided by their chosen host organism. For example, in B. anthracis Spo0A influences toxin synthesis by regulating the expression of the gene encoding the AbrB transition state regulator, a key repressor of toxin gene expression . Similarly, in B. cereus and the insect pathogen Bacillus thuringiensis, Spo0A influences virulence via an effect on the expression of plcR, a gene encoding a positive regulator of the virulence responses of these organisms. Furthermore, Spo0A positively regulates enterotoxin production and sporulation in C. perfringens and one ‘0A’ box can be identified in the promoter region of the cpe enterotoxin toxin gene , providing strong evidence for a direct role in toxin synthesis activation.
The phosphorelay integrates signals emanating from the extracellular environment with those from the cell cycle and metabolism to influence the activation state of Spo0A and the overall pattern of gene expression. The mechanism of Spo0A-mediated gene regulation appears to be the same in all spore-forming bacteria. As bacterial spores are generally recalcitrant to chemical insults, including antimicrobial compounds commonly used in the clinic, the cellular events that control the production of a spore are of obvious relevance to the development of novel therapies to treat or prevent infections caused by spore-forming pathogenic organisms. Furthermore, the identification of the signal ligands for sporulation phosphorelay sensor kinases of any organism has remained elusive, despite intensive investigation. The identity of these signals may hold the key to unravel the complex processes that occur within the host during infection by pathogenic spore-forming bacteria.
- Top of page
- 5Concluding remarks
In recent years, it has become possible to monitor genome-wide changes in transcription with the use of DNA microarray technology. Here, labelled cDNA probes are generated from total cellular RNA, and hybridized against immobilized PCR products corresponding to almost all the ORFs in an entire genome. By using specific mutants or by varying growth conditions, the factors that affect the activity of a transcription factor and the identity of those genes that comprise its regulon can be investigated. This powerful methodology has received wide-spread attention from the Bacillus community, and the regulons for alternative sigma factors [93–96], competence regulators  and two-component regulators have been determined , in addition to that of Spo0A . These studies are often backed up with extensive bioinformatics analysis of, for instance, promoter regions for the presence of consensus binding sites of transcription factors . The details of these computational techniques are beyond the scope of this article. One of the major difficulties in this type of transcriptomic analysis is distinguishing the differences between direct and indirect effects on transcript levels. For instance, Spo0A itself influences the transcription of other regulatory proteins, such as AbrB and SinI, and the sporulation alternative sigma factors σE and σF (reviewed in ). In order to establish which genes are subject to direct control by Spo0A, microarray analyses of Spo0A have been performed alongside those of σF, and more recently chromatin immunoprecipitation (ChIP) of Spo0A has been used in parallel with DNA microarrays , a procedure that has become known as the ‘ChIP-on-chip’ analysis.
4.1Transcriptional profiling in the bacilli
Fawcett et al.  used microarray technology to investigate the influence of Spo0A on global transcription profiles before and following the initiation of sporulation in B. subtilis. In total, the transcript levels of 586 genes were altered by at least threefold between a wildtype and a spo0A deletion mutant, 266 and 283 of which were 5- and 10-fold affected, respectively. These genes could be divided into classes, those that were stimulated (283 genes, class I) or repressed (242 genes, class II) by Spo0A. Within class I, sporulation genes such as spo0F, spoIIA and spoIIG were identified, and were previously known to be under the direct control of Spo0A, as were examples of genes that are indirectly dependent upon Spo0A (e.g., kinA). Genes also found in class I were those expressed during stationary phase that encode proteins with known scavenging functions, such as secreted proteases, peptide antibiotics and those involved in amino acid metabolism. Many of these genes are probably regulated indirectly by Spo0A; for example, some of the protease genes are known to be negatively regulated by the sentinel of the transition state, AbrB, the expression of which is tightly repressed by Spo0A ? P. A group of genes were also identified that are involved in oligopeptide transport, peptide pheromones and the peptide pheromone-regulated Rap phosphatase family. However, the majority of genes identified in class I were of unknown function, indicating that a significant number of the factors governing the initiation of sporulation are not yet completely understood. In class II, genes were identified that were previously known to be directly repressed by Spo0A ? P, such as abrB, and also those that were indirectly repressed via the influence of Spo0A on the expression of other regulators, such as AbrB.
The pleiotropic influence that Spo0A wields in B. subtilis is illustrated by the finding that chemotaxis, motility and autolysin genes were all overexpressed in the spo0A mutant, all of which are σD-dependent. σD transcription is complex as it is positively regulated by SinR, which is negatively regulated by SinI, the transcription of which is Spo0A ? P-dependent! This is typical of the multi-layered control circuits used to regulate post-exponential phase phenomena and sporulation. However, despite the profound effect of Spo0A on global gene expression, it is not clear which of the genes identified above are truly under the control of Spo0A, and which are under the control of other regulators, that are themselves affected by Spo0A. Indeed, the transcriptional profiles of spo0A and sigH mutants are quite similar , but computational analyses of the DNA binding sites of these regulators in promoter regions of the genome are dissimilar . Accordingly, the Spo0A regulon has been re-examined by the ChIP-on-chip’ procedure . Here, from a culture of B. subtilis that harbours an IPTG-inducible constitutively active mutant form of Spo0A (Sad67) , cellular proteins are reversibly crosslinked to DNA, and sheared DNA:Spo0A complexes were isolated by immunoprecipitation with anti-Spo0A antibodies. The cross-links are removed, and the DNA fragments are amplified by PCR and fluorescently labelled, before annealing to DNA microarrays of the B. subtilis genome to reveal the identity of the genes to which Spo0A was bound. The Spo0A-regulated transcriptional units identified in this procedure were validated by transcriptional profile analysis as described above, resulting in the description of a regulon that includes 30 monocistronic genes, and 24 polycistronic operons of 91 genes, corresponding to a total of 121 genes (ca. 3% of the genome) that are directly controlled by Spo0A. Spo0A-binding at the promoters of these genes had only been confirmed for some of these transcriptional units. In the other 60 uncharacterised genes, DNA-binding of C-Spo0A could be demonstrated in upstream regions in 51 of them. Of these newly identified genes that are under the positive control of Spo0A are the sporulation-killing factor and sporulation-delaying protein operons, that have recently been recognized as responsible for regulating the cannibalistic nature of sporulating B. subtilis cells .
4.2Transcriptional profiling in the clostridia
The sporogenous, anaerobic, solventogenic clostridia are challenging organisms to study. Unlike the situation for B. subtilis, only a small number of genes have been functionally studied, the number of available mutants is small and genetic manipulation by chromosomal integration is a serious challenge. B. subtilis was used as the prototype for studies of the clostridia, until the completed genome sequence of C. acetobutylicum re-affirmed that the clostridia are only distantly related to the bacilli [68,102]. The transcriptional profile of approximately one quarter (just over one thousand ORFs) of the genome of wildtype C. acetobutylicum was compared to an asporogenous strain  that harbours an inactivated spo0A gene . A total of 211 genes were identified as being under the control of Spo0A. Of those that are negatively regulated by Spo0A are 20 genes involved in chemotaxis and motility, and one of three genes encoding AbrB paralogues. The repression by Spo0A ? P of the chemotaxis and motility genes in C. acetobutylicum is analogous to the situation in B. subtilis. The transcription profiling also revealed that the transcription of the critical cell division genes, ftsA and ftsZ, are under the control of Sp0A ? P. These genes have essential roles in cell division and sporulation in B. subtilis[104,105], and possibly all eubacteria. Included in the 59 genes that are positively regulated by Spo0A ? P, either directly or indirectly, are sporulation genes that are analogous to the B. subtilis counterparts (e.g., the spoIIA locus), but also the major solventogenic locus (sol) and a number of sugar metabolizing enzymes that are expressed during stationary phase that are unique to C. acetobutylicum. Many of the genes belonging to the Spo0A regulon in the clostridia that were identified by microarray analysis of a spo0A mutant  were subsequently confirmed by the analysis of the effects on transcription profiles of the IPTG-inducible overexpression of Spo0A .
However, a true comparison between the Spo0A regulons of B. subtilis and C. acetobutylicum might be problematic given the different chemical and genetic conditions used during the growth conditions prior to transcriptome analysis. Nonetheless, completed genome sequences combined with global transcriptional analysis facilitates detailed investigation of the presence and role of orthologues of the members of the Spo0A regulon in low G + C Gram-positive bacteria, such as B. subtilis and its relatives . Sixty-one percent of the transcriptional units that are activated by Spo0A in B. subtilis have orthologues in the four bacilli with completed genome sequences, whereas 79% of the repressed transcriptional units are maintained. In the three clostridia, C. acetobutylicum, C. perfringens and C. tetani, these values are 27 and 53, respectively. In the closely related non-endospore forming genus Listeria, which does not encode Spo0A, 22% and 70% of the orthologues corresponding to activated or repressed genes, respectively, are conserved. In all cases, genes activated by Spo0A are less well conserved than those that are repressed. The Spo0A-repressed genes include essential genes that are expressed only during vegetative growth and are repressed on the entry into sporulation, and thus perhaps the degree of conservation between the genera is not surprising. Of the genes that are activated by Spo0A, these are more likely to be involved directly in sporulation, and these genes are poorly conserved in the non-sporulating Listeria.
- Top of page
- 5Concluding remarks
In summary, the years after the sequencing of the genome of B. subtilis have been fruitful, with new insights obtained by both established and the new methodologies. One of the most striking findings in recent years has been the observation that B. subtilis cells about to commit to sporulation secrete a ‘killing factor’ that blocks other cells from sporulating, and causes cell death . The nutrients that are released from the lysed cells allow the sporulating cells to feed and continue to grow rather than to complete morphogenesis. It would thus appear from this study that sporulation is truly the final act of the desperate cell, and that any opportunity to delay sporulation, even at the expense of other B. subtilis cells in the immediate vicinity, is taken.
Prospects for new discoveries still exist. For instance, there has been no published systematic study of the changes in protein synthesis on the onset of sporulation by two-dimensional gel electrophoresis and mass spectroscopic proteome analysis. This approach has one advantage over the more commonly used transcript profiling because those proteins whose cellular levels do not fluctuate, but whose activities do (by covalent post-translational modification, e.g., phosphorylation), can still be identified. The chemical signals that indicate to the spore-forming bacterium that sufficient changes in its environment have occurred to require sporulation as a means of survival remain unknown. Furthermore, the so far underdeveloped structural analysis of sporulation sensor kinases has the potential to provide significant insights into the molecular basis of recognition, substrate specificity and the evolution of the interaction surfaces of signalling proteins. Finally, the molecular basis of Spo0A activation remains uncertain. The qualitative evidence supporting dimerisation of Spo0A upon phosphorylation as a prerequisite to activation is compelling, but a more rigorous thermodynamic analysis is required before dimerisation can be fully accepted as a physiologically relevant model.
In contrast, the structural details of key components of the phosphorelay, the interactions they make with macromolecular binding partners and the conformational changes that accompany phosphorylation in response regulator receiver domains have all been elucidated, and have had significant impact beyond the study of sporulation, in the wider two-component signal transduction field. The structural studies have also allowed a molecular analysis of the effects of a number of mutations in phosphorelay components that confer sporulation-deficient phenotypes on B. subtilis and confirmed that key residues in binding or phosphotransfer reactions are conserved in sporogenous bacteria. It is expected that the gallery of structures from the sporulation phosphorelay will soon be complemented by the determination of structures of Rap phosphatase family members, which are predicted – but not experimentally proven – to contain tetratricopeptide repeat motifs for use in molecular recognition events .
The genome sequencing of different spore-forming bacteria has allowed direct comparisons of the physiology and evolution of bacteria that in some cases are difficult or hazardous to culture in the laboratory. Indeed, the sequence of spo0A alone was sufficient to determine that the obligate plant nematode parasite Pasteuria penetrans is also sporogenous . The identity of genes directly and indirectly under the control of Spo0A in B. subtilis, and C. acetobutylicum, have been determined and compared, confirming that only some elements are consistent, such as the Spo0A-directed repression of the transition state regulator AbrB. However, it is clear that the signal transduction events that lead to the activation of Spo0A in Clostridium species differs markedly from the phosphorelay process in the bacilli. Insights such as these would not be possible without the knowledge of genome sequences and the detailed structural analysis of phosphorelay proteins from Bacillus species.
The conservation of the sporulation phosphorelay in the bacilli but not in the clostridia might at first glance appear somewhat surprising. However, each organism's genome has been shaped by the need to adapt to changes in its normal ecological habitat. The anaerobic clostridia occupy very different niches to the aerobic bacilli. Consequently, although the surfaces of interacting partners in the phosphorelay of Bacillis species are conserved, the sensing domains of the sensor kinases are highly variable. This variability no doubt reflects the range of signals that have to be integrated by the signalling systems that control sporulation initiation in the different organisms. Despite the high level of conservation of the Spo0A proteins, the genes that are regulated by this response regulator are not conserved in the bacilli and clostridia. Spo0A is not the only example of the variability of the regulation of important genes between, or even within, closely related genera. If we have thus discovered anything in the post-genomic world, it is that an understanding of the evolution of complex signalling pathways, at both the genetic and molecular levels, is now within our grasp.
- Top of page
- 5Concluding remarks
Research in R.J.L.'s lab is funded by the Wellcome Trust, through a Research Career Development Fellowship, the BBSRC and Newcastle University. K.S.'s research is funded by the Royal Society and Leeds University.
- Top of page
- 5Concluding remarks
- Kunst, F., 150 other authors. (1997) The complete genome sequence of the Gram-positive bacterium Bacillus subtilis. Nature 390, 249–256
- Gardino, A.K., Volkman, B.F., Cho, H.S., Lee, S.Y., Wemmer, D.E., Kern, D. The NMR solution structure of -activated Spo0F reveals the conformational switch in a phosphorelay system. J. Mol. Biol. >331 2003. 245–254
- Hatt, J.K., Youngman, P. Spo0A mutants of Bacillus subtilis with sigma factor-specific defects in transcription activation. J. Bacteriol. >180 1998. 3584–3591 78–83..
- 2004) Disruption of the gene (spo0A) encoding sporulation transcription factor blocks endospore formation and enterotoxin production in enterotoxigenic Clostridium perfringens type A. FMR Microbiol. Lett. 233, 233–240., , , (
- Piggot, P.J., Losick, R. (2002) Sporulation genes and intercompartmental regulation. In: Bacillus subtilis and Its Closest Relatives: From Genes to Cells (Sonenshein, A.L., Hoch, J.A., Losick, R., Eds.), pp.483–517 American Society for Microbiology Press, Washington, DC.