Extremely Variable Conservation of γ-Type Small, Acid-Soluble Proteins from Spores of Some Species in the Bacterial Order Bacillales

Abstract

γ-Type small, acid-soluble spore proteins (SASP) are the most abundant proteins in spores of at least some members of the bacterial order Bacillales, yet they remain an enigma from both functional and phylogenetic perspectives. Current work has shown that the γ-type SASP or their coding genes (sspE genes) are present in most spore-forming members of Bacillales, including at least some members of the Paenibacillus genus, although they are apparently absent from Clostridiales species. We have applied a new method of searching for sspE genes, which now appear to also be absent from a clade of Bacillales species that includes Alicyclobacillus acidocaldarius and Bacillus tusciae. In addition, no γ-type SASP were found in A. acidocaldarius spores, although several of the DNA-binding α/β-type SASP were present. These findings have elucidated the phylogenetic origin of the sspE gene, and this may help in determining the precise function of γ-type SASP.

INTRODUCTION

Bacterial spores of species of the Firmicutes phylum contain a number of small, acid-soluble proteins (SASP) that comprise 10 to 15% of the protein in the spore's central region or core (30, 32). The following two types of major SASP have been identified in spores: (i) α/β-type SASP that are products of a multi-ssp gene family and have extremely similar sequences both within and across species and (ii) γ-type SASP that are almost always encoded by a single sspE gene; this is the most abundant protein found in spores of a number of species and comprises 5 to 8% of total spore protein (18–20, 29–32, 39). In contrast to the highly conserved sequences of α/β-type SASP, sequences of γ-type SASP are not well conserved across species, and this has allowed the use of sspE and SASP-γ sequences to distinguish closely related Bacillus strains and species (17).

In Bacillus subtilis, genes encoding both α/β-type and γ-type SASP are transcribed in parallel late in spore development when the various SASP are synthesized, and the transcription of ssp genes is mediated by the RNA polymerase sigma factor, σ^G (22). The SASP are degraded soon after spores complete the germination process and begin outgrowth, and one function of these proteins is to serve as a reservoir of amino acids (aa) for protein synthesis early in outgrowth (12, 30, 32). The latter is an important function, since spores become deficient in a number of amino acid biosynthetic enzymes during spore formation and synthesize these enzymes only during spore outgrowth. In addition to serving as a reservoir of amino acids, the α/β-type SASP have an additional function, as these proteins saturate spore DNA and protect it from many types of damage and are thus very important for long-term spore survival (30–32). However, other than serving as an amino acid reservoir, no additional function has been demonstrated for γ-type SASP (12, 30, 32).

In the current work, we have examined genome sequence information for spore-forming Firmicutes and have confirmed that (i) spore-forming Clostridiales species appear to lack sspE genes; (ii) most but not all spore-forming Bacillales species, including those in the clade encompassing Paenibacillus species, appear to contain a single sspE gene; and (iii) some Bacillales species, including Alicyclobacillus acidocaldarius and Bacillus tusciae, appear to lack an sspE gene. We have also analyzed SASP in spores of several of these species and have found that (i) Paenibacillus polymyxa spores contain a γ-type SASP homolog that is related only distantly to γ-type SASP of Bacillus species, and (ii) A. acidocaldarius spores appear to lack a γ-type SASP but do contain at least two α/β-type SASP. These observations have allowed the determination of the phylogenetic origin of the sspE gene in the order Bacillales, and this information could lead to suggestions for additional functions of γ-type SASP besides that of an amino acid reservoir.

MATERIALS AND METHODS

Preparation of P. polymyxa spores and SASP extraction and analysis.

P. polymyxa (ATCC 842) was obtained from the American Type Culture Collection. Spores of this species, as well as those of B. subtilis PS832, a laboratory derivative of strain 168, were prepared and purified as described previously (14, 23). The purified P. polymyxa spores (5 to 6 mg [dry weight]) were lyophilized and dry ruptured with 10-μm glass beads (100 mg [dry weight]) as the abrasive, with 10 1-min periods of shaking interspersed with 1-min periods of cooling (23). The dry-ruptured powder was extracted twice with 1 ml cold 3% acetic acid, and the two supernatant fluids were pooled and dialyzed at 4°C in Spectra/Por 3 tubing (molecular weight cutoff, 3,500) for ∼18 h against two changes of 1 liter cold 1% acetic acid as described previously (23). Acetic acid extracts of ∼4 mg (dry weight) B. subtilis spores were prepared as described previously (23) and used as a source of markers for α/β- and γ-type SASP in acid gel electrophoresis. The final dialysates were centrifuged, and the supernatant fractions were lyophilized.

P. polymyxa spores (∼12 mg [dry weight]) were incubated in an alkaline decoating solution containing SDS and urea for 30 min at 65°C to remove the spore coat and outer membrane protein while retaining SASP and spore viability (5). After the decoated spores were washed and dried, an aliquot (∼5 mg [dry weight]) was disrupted and acetic acid extracts were prepared, processed, and dried as described above. Intact P. polymyxa spores (6 mg [dry weight]) were also germinated for 60 min at 37°C in ∼30 ml of 1 mM dodecylamine in 20 mM Tris-HCl buffer (pH 8.4) (28), and phase-contrast microscopy indicated that ≥90% of the spores had germinated. After centrifugation and lyophilization, ∼4 mg (dry weight) of the germinated spores was disrupted and acetic acid extracts were prepared, processed, and dried as described above.

The dried acetic acid extracts from various types of spores were dissolved in 30 μl of 8 M urea plus 15 μl diluent for acid-acrylamide gel electrophoresis, samples were run on acrylamide gel electrophoresis at a low pH, and the gels were stained with Coomassie blue (23). In some experiments, proteins separated by polyacrylamide gel electrophoresis at a low pH were transferred to polyvinylidene difluoride membranes, and the proteins on these membranes were stained with Coomassie blue. Selected stained protein bands were then subjected to automated N-terminal sequence analysis on an Applied Biosystems Procise 494 HT protein sequencer in the Keck Biotechnology Resource Center at the Yale University School of Medicine.

Preparation of A. acidocaldarius spores and SASP extraction and analysis.

A. acidocaldarius strain NRS 1662 was obtained from the American Type Culture Collection, and spores of this strain were prepared on agar plates using a modification of the basal medium for A. acidocaldarius as described previously (10). This medium had final concentrations of 1.5 mM (NH₄)₂SO₄, 1.7 mM CaCl₂, 2 mM MgSO₄, 4.4 mM KH₂PO₄, and 1 g/liter yeast extract, and the mixture of these components was adjusted to pH 3.7 with 10 N H₂SO₄ prior to autoclaving and held at 50°C. Autoclaved glucose, agar, and filter-sterilized MnCl₂ (all also held at 50°C) were added separately to final concentrations of 1 g/liter, 15 g/liter, and 250 μM, respectively, just prior to being poured into plates. The A. acidocaldarius strain was streaked on a plate made as described above except without MnCl₂, and the plate was incubated overnight at 55°C. A loop of this overnight culture was inoculated into 20 ml of the medium as described above but without agar, the culture was grown for 5 to 8 h at 55°C to an optical density at 600 nm of ∼0.7, and 200-μl aliquots were spread on 40 plates containing MnCl₂ as described above. The plates were incubated in bags for ∼72 h at 55°C until maximum sporulation had occurred and cooled to 23°C for a few hours, and the cell/spore mix was scraped from the plates and placed in 4°C deionized water. The spores were purified initially by multiple rounds of sonication, followed by centrifugation and water washing as described previously (23), and the crude spores were suspended in 10 ml of 4°C water. Final removal of cells, cell debris, and germinated spores was by layering 1 ml of the crude spores on each of six 13.2-ml ultracentrifuge tubes with 8.5 ml 50% Nycodenz (Sigma Chemical Company, St. Louis, MO) at 20°C and centrifugation at 13,000 rpm in a Beckman 48 Ti rotor for 20 min at 20°C. Under these conditions, cells, cell debris, and germinated spores remained at the water-Nycodenz interface while the pure spores pelleted. The pelleted spores were washed ∼5 times with 4°C water to remove Nycodenz and finally suspended in ∼5 ml water. This procedure yielded ∼45 mg (dry weight) of A. acidocaldarius spores that were >98% free of cells, cell debris, and germinated spores as observed by microscopy.

Twelve milligrams (dry weight) of the purified A. acidocaldarius spores was dry ruptured and extracted with acetic acid as described above but using 30 min of rupturing and with dialysis for only ∼4 h. The dialysate was lyophilized, the residue was dissolved in 25 μl of 8 M urea plus 12.5 μl acid gel diluent (23), 15-μl aliquots were subjected to acid gel electrophoresis, the proteins on the gel were transferred to a polyvinylidene difluoride membrane, the membrane was stained with Coomassie blue as described above, and proteins in stained bands were sequenced as described above.

RESULTS

Sequences of γ-type and α/β-type SASP.

The sequences of spores' single γ-type SASP are not as conserved as those of the α/β-type SASP, and γ-type SASP vary from 49 to 139 residues in length (Fig. 1). All γ-type SASP do, however, contain one, two, or three repeats of an evolutionarily conserved 7-aa sequence, TEFASET. In B. subtilis and Bacillus megaterium SASP-γ, this sequence is the recognition site for cleavage of this protein by a specific protease, termed GPR, that initiates degradation of both α/β-type and γ-type SASP early in spore outgrowth (Fig. 1 and 2, sequences shaded in red) (30, 32). The spacing between the conserved heptapeptide sequences in γ-type SASP that have multiple repeats of this sequence varies considerably (32). In contrast to the α/β-type SASP present in spores of all Firmicutes species that have been studied, γ-type SASP and sspE genes have not been identified in spore-forming species of the order Clostridiales (32).

Fig. 1.

Comparison of amino acid sequences of γ-type SASP from spore-forming Bacillales species. The sequences are given in the one-letter code for the following species: Afl, Anoxybacillus flavithermus; Bam, Bacillus amyloliquefaciens; Ban, Bacillus anthracis; Bat, Bacillus atrophaeus; Bav, Bacillus aminovorans; Bbr, Brevibacillus brevis; Bce, Bacillus cereus; Bcl, Bacillus clausii; Bcy, Bacillus cytotoxis; Bfi, Bacillus firmus; Bha, Bacillus halodurans; Bli, Bacillus licheniformis; Bme, Bacillus megaterium; Bmy, Bacillus mycoides; Bpf, Bacillus pseudofirmus; Bps, Bacillus pseudomycoides; Bpu, Bacillus pumilus; Bsu, Bacillus subtilis; Bth, Bacillus thuringiensis; Bwc, Bacillus weihenstephanensis chromosome; Bwp, B. weihenstephanensis plasmid; GC5, Geobacillus sp. C56-T3; Gka, Geobacillus kaustophilus; Gst, Geobacillus stearothermophilus; Gth, Geobacillus thermodenitrificans; GWC, Geobacillus sp. WCH70; GY1, Geobacillus sp. Y412MC61; Pym, Paenibacillus sp. Y412MC10; Hha, Halobacillus halophilus (originally Sporosarcina halophila); Lsp, Lysinibacillus sphaericus; Oih, Oceanobacillus iheyensis; Pcu, Paenibacillus curdlanolyticus; Pjd, Paenibacillus sp. JDR-2; Pot, Paenibacillus sp. oral taxon 786 strain D14; Ppo, Paenibacillus polymyxa strain E681; Sur, Sporosarcina ureae; and Tth, Thermoactinomyces thalpophilus. For species whose abbreviations are underlined, the amino acid sequences were from the sspE gene cloned from this species (18, 25, 33); all other sequences were from the species' completed genome sequences. The seven-residue sequences shaded in red are the sites for recognition and cleavage by the SASP-specific protease GPR during spore outgrowth, with cleavage between the bold residues (29, 32). The sequence in the N-terminal region of Ppo in purple is the protein sequence determined in this work, and the yellow blocks of sequences in Ppo and Bsu represent two long blocks of repeated sequences in each of these proteins.

Fig. 2.

Comparison of amino acid sequences of α/β-type SASP from A. acidocaldarius, B. subtilis, B. tusciae, G. kaustophilus, and Paenibacillus species. The sequences are shown in the one-letter code, with the asterisk in the Btu3 sequence denoting a stop codon and the dashes introduced to maximize sequence alignments. The sequences are from the following species: Bsu, B. subtilis; Gka, G. kaustophilus; Aac, A. acidocaldarius; Btu, B. tusciae; Pym, Paenibacillus sp. Y412MC10; Pcu, P. curdlanolyticus; Pjd, Paenibacillus sp. JDR-2; Pot, Paenibacillus sp. oral taxon 786 strain D14; and Ppo, P. polymyxa. The sequences above the break are those from species that are more distantly related to those whose sequences are listed below this break. The long regions of sequence shaded in red and yellow are sequences that each form long α-helices that are important structural elements in these proteins' binding to DNA (16). The site of cleavage of these proteins during spore outgrowth by the SASP-specific protease is between the bold residues in the red-shaded regions. The sequence blocks shaded in green seem to be duplications of the GPR cleavage site region, and in all of the proteins with this duplication, the spacing between the red and yellow blocks of sequence is greatly increased. The Ppo1, Ppo2, Aac1, and Aac2 sequences in purple were obtained by protein sequence analysis in this work, although the complete sequences of the P. polymyxa proteins are from strain E681 and the sequences of the A. acidocaldarius proteins are from strain DSM 446.

In order to better understand the genetic history of sspE, we aimed to position Firmicutes phylogeny with respect to the presence or absence of the sspE gene. Computational analysis consisted of two steps: (i) exhaustive, controlled searching of completed Firmicutes species genomes for sspE orthologs and (ii) construction of a Firmicutes species phylogenetic tree based on 16S rRNA sequences. These efforts should phylogenetically position the sspE gene to divergence points in Firmicutes evolution, thus suggesting critical evolutionary events that led to the emergence of SASP-γ as a protein that facilitates spore outgrowth and might play some additional but unknown role. Exhaustive detection of a homologous gene across multiple genomes is computationally intensive (11). Although tools such as BLAST (2) exist for finding similar protein sequences, the fact that proteins evolve at different rates and by different mechanisms mandates that exhaustive searches for orthologs utilize iteration against different thresholds and alignment methods. We thus constructed an as-yet-unpublished software tool for performing a “multidimensional” sequence search which allows for analysis of all “best hit” proteins across multiple genomes with respect to not just one but a set of several target proteins of interest (the source code can be acquired by contacting the authors directly; the database structure was obtained from reference 35). This tool was applied to Firmicutes proteomes to reveal sspE genes.

The sequences of γ-type SASP were updated and extracted to a relational database containing over 3 million NCBI proteins from completed proteomes and placed in a smaller data mart containing proteins in the 58 fully curated genomes of Firmicutes species (27). The proteome sizes in these 58 species ranged from 2,080 proteins (Ammonifex degensii) to 6,238 proteins (Paenibacillus sp. Y412MC10 [originally classified as a Geobacillus species]). The protein sequences for major sporulation proteins Spo0A, Spo0E, Spo0F, SpoIIE, SpoIIGA, SpoIVB, SpoVAA, SpoVFA, SpoVFB, CwlJ, and CotE from different species were identified and visualized as histograms alongside the best hits when stratified against SASP-γ proteins from Bacillus clausii, B. subtilis, and Lysinibacillus sphaericus, each of which has SASP-γ proteins with rather different amino acid sequences (Fig. 1). Non-trivial best hits (those with a pairwise sequence similarity of 0.3 on a scale from 0 to 1) were investigated for all genomes in order to validate a list of known sspE-containing organisms, and for the significant SASP-γ sequence feature, the TEFASET motif was investigated (32), as it appears to be well conserved and thus likely plays a critical functional role for the SASP-γ protein family. The final list obtained represented a complete set (with respect to NCBI-curated Firmicutes species genomes) of sspE-containing organisms and can be utilized for phylogenetic analysis. Clear sspE homologs were found in all Bacillales species (Fig. 1) except A. acidocaldarius, B. tusciae, P. polymyxa, Paenibacillus sp. JDR-2, Paenibacillus sp. Y412MC10 and Bacillus selenitireducens. However, B. selenitireducens likely does not sporulate and lacks genes for a number of important spore proteins, including α/β-type SASP (6, 24; data not shown). No sspE genes were identified in spore-forming Clostridiales species either, as was found previously by the analysis of acid-soluble spore proteins or completed genome sequences (7, 8, 32).

Comparison of amino acid sequences of SASP-γ obtained from completed genomes of spore-forming Firmicutes species, as well as those identified by targeted gene cloning (Fig. 1) (18, 25, 34), showed that γ-type SASP exhibit some conserved sequences but are most notable for the following characteristics: (i) a significant enrichment of Gln and Asn (8- and 6-fold enriched with respect to normal frequencies in the proteomes of Paenibacillus sp. JDR-2 and Alicyclobacillus acidocaldarius [∼0.036 and 0.105, respectively]); (ii) very low levels of hydrophobic residues, with substantially lower hydrophobic residue abundances in γ-type SASP, although these residues tend to be less naturally abundant in general (e.g., levels of L, I, and P are, respectively, 3, 2.5, and 1.8 standard deviations below proteomic backgrounds); and (iii) 1 to 3 repeats of the relatively well-conserved TEFASET GPR cleavage site, although the spacing between these last conserved sequences varies considerably (Fig. 1). Only a single sspE gene was identified in the completed genomes searched, with the exception of Bacillus weihenstephanensis, which contained sspE genes encoding almost identical proteins on both the chromosome and a plasmid (Fig. 1).

The absence of an obvious sspE gene from the completed genomes of A. acidocaldarius, B. tusciae, Paenibacillus sp. Y412MC10, Paenibacillus sp. JDR-2, and P. polymyxa was surprising, but examination of available draft genomic sequence information for Paenibacillus curdlanolyticus and Paenibacillus sp. oral taxon 786 strain D14 also revealed no obvious sspE genes, although these species as well as A. acidocaldarius, B. tusciae, Paenibacillus sp. Y412MC10, Paenibacillus. sp. JDR-2, and P. polymyxa did contain multiple genes encoding α/β-type SASP (Fig. 2). The amino acid sequences of α/β-type SASP from Bacillales species compared in previous work are extremely well conserved (30, 32). In particular, these proteins exhibit two 18- to 19-aa regions of a highly conserved sequence that are commonly separated by a 3-aa spacer (Fig. 2, red and yellow highlighted regions and in B. subtilis and Geobacillus kaustophilus sequences) (30, 32). These two conserved regions each form long α-helices that interact with the minor groove of DNA and are also a major element of the SASP-SASP dimer interface when the protein is bound to DNA (16). Notably, the spacing between these two conserved regions is 3 aa in almost all of the α/β-type SASP sequences examined to date (32). However, in all but two of the available α/β-type SASP sequences from Paenibacillus species, A. acidocaldarius, and B. tusciae, there are ≥5 aa between these two highly conserved structural elements (Fig. 2 and data not shown), as has also been observed in almost all α/β-type SASP from Clostridiales species (32). In addition, even in the two highly conserved regions in the α/β-type SASP from A. acidocaldarius, B. tusciae, and Paenibacillus species there are amino acids present that are not found at these locations in α/β-type SASP from other Bacillales species (Fig. 2). These last differences are consistent with the phylogenetically distant relationship between almost all Bacillales species and Paenibacillus species, B. tusciae, and A. acidocaldarius (3, 33; see below). It was notable that the amino acid sequences of the four α/β-type SASP from Paenibacillus sp. Y412MC10 are rather different from those of most Bacillales species, including those of G. kaustophilus (Fig. 2) and a number of other Geobacillus species (33; data not shown). These data and the presence of an obvious sspE gene in all completed genomes of Geobacillus species except Paenibacillus sp. Y412MC10 are consistent with the recent assignment of Paenibacillus sp. Y412MC10 as a Paenibacillus species, as it had originally been classified as a Geobacillus species (see below).

Analysis of SASP in P. polymyxa spores and possible sspE genes in Paenibacillus species.

Although the absence of obvious sspE genes from Paenibacillus species, B. tusciae, and A. acidocaldarius was unexpected, the divergence of these organisms from most other Bacillales species was an ancient event (3, 33). Consequently, the fact that Bacillales species that appeared to lack an sspE gene were related only distantly to sspE-containing Bacillales species suggested the possibility that an sspE gene might indeed be present in these distantly related organisms but has diverged sufficiently to preclude recognition by normal sequence comparison programs. Consequently, we examined spores of A. acidocaldarius and P. polymyxa for a protein that might be an ortholog of SASP-γ.

The acetic acid extract from purified P. polymyxa spores produced a large number of bands on polyacrylamide gel electrophoresis at a low pH, with most of these bands migrating faster than B. subtilis SASP (Fig. 3, lane 1, B. subtilis SASP migration positions denoted by arrows). An obvious question is whether all of the prominent bands seen in the P. polymyxa spore extracts were really SASP. To help answer this question, an aliquot of the acetic acid extract from decoated P. polymyxa spores was also subjected to polyacrylamide gel electrophoresis at a low pH, and this revealed that decoating did not reduce the intensities of bands 1, 2, and 3 but greatly reduced the intensities of all other bands (data not shown). This suggested that many of the latter bands were due to spore coat proteins. To obtain further evidence that one or more proteins in bands 1, 2, and 3 were indeed SASP, an aliquot of germinated P. polymyxa spores was subjected to polyacrylamide gel electrophoresis at a low pH (Fig. 3, lane 2). Bands 1, 2, and 3 were almost completely absent from the germinated spore extract, while the intensities of most other bands were not notably affected. These data are consistent with the proteins in bands 1, 2, and 3 being SASP. Automated N-terminal amino acid sequence analysis of proteins transferred to polyvinylidene difluoride membranes, following polyacrylamide gel electrophoresis at a low pH, produced N-terminal sequences of AQGNNGNS and SRRRNNLQV for bands 1 and 3, respectively. A search of the completed P. polymyxa E681 genome indicated that the proteins in these bands were α/β-type SASP, and these were designated Ppo1 and Ppo2, respectively (Fig. 2).

Fig. 3.

Polyacrylamide gel electrophoresis at low pH of acetic acid extracts from spores of P. polymyxa ATCC 842 (lanes 1 and 2) and A. acidocaldarius NRS 1662 (lane 3). P. polymyxa spores were isolated and purified, and ∼5 mg (dry weight) was disrupted before or after germination; the dry powder was extracted, dialyzed and lyophilized; aliquots were run via polyacrylamide gel electrophoresis at a low pH; and the gel was stained as described in Materials and Methods. The samples run in lanes 1 and 2 are from dormant P. polymyxa spores (lane 1) and germinated P. polymyxa spores (lane 2). Bands labeled 1 and 3 in lane 1 are the Ppo1 and Ppo2 α/β-type SASP, respectively (Fig. 2), while band 2 is the product of an sspE-like gene (Fig. 1), as determined by amino-terminal sequence analysis of these protein bands as described in the text. Bands labeled a to h are ones that were largely or completely removed by decoating treatment, while bands 1 to 3 were not (data not shown). (Lane 3) Dormant A. acidocaldarius spores (12 mg [dry weight]) were ruptured and extracted, an aliquot from ∼4 mg spores was run via polyacrylamide gel electrophoresis at a low pH, proteins on the gel were transferred to a polyvinylidene difluoride membrane, and the membrane was stained as described in Materials and Methods. Bands labeled 1 and 2 in lane 3 are the A. acidocaldarius α/β-type SASP Aac1 and Aac2, respectively, as described in the text. Lanes 1 and 2 are from the same gel, while lane 3 is from a separate gel. The labeled horizontal arrows adjacent to lanes 1 and 3 denote the migration positions of B. subtilis SASP-α and -γ, which were determined by running an aliquot of an acetic acid extract of B. subtilis spores in lanes that are not shown but were adjacent to lanes 1 and 3.

The N-terminal amino acid sequence of the protein in band 2 was PNQGGSXN, and this sequence matched that of a protein, termed Ppo, encoded in the P. polymyxa E681 genome (Fig. 1). This protein is clearly not an α/β-type SASP as it lacks the two large blocks of conserved sequence found in these proteins. However, Ppo has a number of similarities with γ-type SASP. In particular, Ppo (i) contains 13% Gln, (ii) has two repeats of a 7-aa sequence with high similarity to the TEFASET motif, where GPR cleaves the B. subtilis γ-type SASP (Fig. 1, sequences shaded in red), and (iii) has extended sequence repeats (Fig. 1, yellow blocks in the Ppo and Bsu sequences) (30). In addition, the ppo translational start codon was preceded by a strong Gram-positive bacterial ribosome binding site (RBS) and had appropriately spaced −10 and −35 sequences preceding the RBS that were extremely similar to those in promoters of genes encoding highly expressed Bacillus sporulation proteins, including SASP that are recognized by RNA polymerase containing σ^G (Fig. 4) (22). Similar RBS and appropriately separated −10 and −35 sequences are also upstream of the coding sequences of the genes encoding the P. polymyxa α/β-type SASP Ppo1 and Ppo2 (Fig. 4). In addition, following their translation stop codons, the ppo, ppo1, and ppo2 genes each had an inverted repeat sequence, followed by a T-rich region that is likely a rho-independent transcription terminator, as found in all genes encoding major SASP (30).

Fig. 4.

Putative upstream and downstream regulatory regions for genes encoding α/β-type and γ-type SASP from various species. The genes from the various species are those in Fig. 1 and 2, and the names of the species are as follows: Aac, A. acidocaldarius; Btu, B. tusciae; Pym, Paenibacillus sp. Y412MC10; Pcu, P. curdlanolyticus; Pjd, Paenibacillus sp. JDR-2; Pot, Paenibacillus sp. oral taxon 786 strain D14; and Ppo, P. polymyxa. The upstream and downstream sequences of the genes are from the NCBI Microbial Genomes database. The optimal −10 and −35 promoter sequences and their spacing for strong σ^G promoters are listed above the upstream sequences and are from highly expressed σ^G-dependent genes of four Bacillus species, including those encoding α/β-type and γ-type SASP (22). Bold nucleotides in the −10 and −35 sequences are >90% conserved in these sequences, and nucleotides that are not bold are 50 to 70% conserved. Note that while the Bacillus genes are almost exclusively recognized by σ^G, the recognition sequence for σ^F overlaps that of σ^G to a significant extent (36), so the sequences shown upstream of genes in other species that do or may encode SASP might be recognized by σ^F rather than σ^G. A perfect RBS sequence for Bacillus mRNAs is also listed above the upstream sequences. For each gene, in the upstream sequences the putative −10 and −35 σ^G promoter sequences are highlighted in purple, the RBS in yellow, and the translation start codon in red; for the downstream sequences the translation stop codon is highlighted in green and an inverted repeat followed by a T-rich region is in cyan.

Surprisingly, the complete genomes of several other Paenibacillus species had no strong Ppo homologs (Fig. 1). However, genes encoding proteins with some similarity to γ-type SASP were discovered in Paenibacillus species with either completed or draft genomes available by searching for encoded proteins ≤120 aa that matched at least 6 of the 7 amino acids in the TEFASET GPR cleavage motif. To accommodate sequence divergence, we searched via the consensus motif of [AQT]EF[AGS][AST][EQ][FT]. Examination of these potential γ-type SASP revealed at least one ortholog in a number of Paenibacillus species for which the coding gene had other features of genes encoding γ-type SASP, including (i) the presence of a strong RBS; (ii) the RBS being preceded by sequences with excellent homology to −10 and −35 promoter sequences recognized by RNA polymerase with σ^G, although these genes could potentially be recognized by the other forespore-specific sigma factor σ^F, as the promoter sequences of σ^G and σ^F-dependent genes overlap to some extent (36); (iii) appropriate spacing between the putative −10 and −35 σ^G recognition elements; and (iv) coding sequences followed by likely rho-independent transcription terminators (Fig. 4). In addition, the putative γ-type SASP from Paenibacillus sp. oral taxon 786 strain D14, Paenibacillus curdlanolyticus, Paenibacillus sp. JDR-2, and Paenibacillus sp. YM412MC10 had 9, 12, 14, and 14% Gln, respectively, while the Ppo protein had 13% Gln (Fig. 2). Overall, these data are consistent with the proteins encoded by these genes being γ-type SASP, although this has by no means been proven. In addition, the absence of any close homolog of Ppo encoded by other completely sequenced Paenibacillus genomes suggests that this group of organisms is extremely diverse, perhaps more so than many other clades of Bacillales species.

Analysis of SASP and ssp genes in A. acidocaldarius spores.

In contrast to the P. polymyxa spore extract, the A. acidocaldarius spore extract produced only a single major band (band 1) on acid gel electrophoresis, with this band running slightly faster than B. subtilis SASP-γ, as well as at least one other minor band (band 2) (Fig. 3, lane 3, migration position of B. subtilis SASP-γ denoted by the arrow). No other band in the A. acidocaldarius extract had ≥5% of the intensity of the major band.

Automated sequence analysis of bands 1 and 2 from the A. acidocaldarius spore extract resulted in the sequences ANNSGSNR and ANQN[SG]SNR, which were perfect matches to the N-terminal sequence of A. acidocaldarius α/β-type SASP Aac1 and Aac2, respectively, with the N-terminal methionine residues removed (Fig. 2). The region upstream of the aac1 and aac2 genes contained good matches to likely transcription and translation signals found in other genes encoding major SASP, and both genes had a likely transcription terminator downstream of the translation stop codon (Fig. 4). The identification of only α/β-type SASP in acid extracts of A. acidocaldarius spores suggests that if spores of this species contain a γ-type SASP, it is expressed only poorly, is not acid soluble, or is extremely labile to digestion by an acidic protease. However, it seems more likely that this species does not contain an sspE gene. Indeed, while a search of the completed A. acidocaldarius and closely related B. tusciae genomes did reveal potential genes encoding proteins that had near matches to the 7-aa GPR cleavage site motif, candidate genes lacked many other features of genes encoding γ-type SASP in either Bacillus species or P. polymyxa, such as a strong RBS and putative σ^G promoter sequences (data not shown). Consequently, it appears most likely that A. acidocaldarius and B. tusciae lack an sspE gene.

DISCUSSION

The work in this communication indicates that although sspE genes are found in most spore-forming Bacillales species, they are most likely absent from A. acidocaldarius and B. tusciae, two species that are evolutionarily divergent from other spore-forming Bacillales (3, 4, 9, 15, 33, 38; see below). Phylogenetic reconstruction of bacterial evolution can be more straightforward than searches for protein orthologs, and this has been done for Firmicutes species in the past (15). We have reconstructed a Firmicutes phylogenetic tree that includes all known sspE-containing species, along with a number of other Firmicutes, including those of Bacillales species that lack an sspE gene. In order to generate this phylogenetic tree using the most up-to-date information from public databases, we merged 16S rRNA sequences from various prokaryotic genome resources, including the NCBI, RDP, and GreenGenes databases. In particular, the RDP database was used to provide a broad survey of various classes of 16S rRNA sequences to increase the accuracy of our predictions. All reconstructions were performed using the BOSQUE program (26), and critical sequences which appeared to be volatile during alignment were cross-validated via external BLAST searches of EBI prokaryotic 16S rRNA sequences, as well as by comparison with previous reconstructions. Inclusion of sequences of a number of Clostridium species was shown to be critical for training the reconstruction to accurately cluster Thermoactinomyces species in a manner that was consistent with external BLAST searches of current data, as well as previous reconstructions by other groups. The sequence alignment was done using Muscle 3.6, and the tree was derived using AIC with four categories and PhyML evolution by HKY (1, 13). The final tree (Fig. 5) includes the spore-forming Bacillales species with completed genomes, as well as those shown previously to contain an sspE gene, even though these species' genomes have not been sequenced (18, 25, 34), as well as members of other Firmicutes genera and species. Importantly, members of one such genus, Thermoactinomyces, are highly dissimilar to other Firmicutes and have very few closely matching neighbors. A BLAST search of 16S rRNA sequences from Thermoactinomyces species yielded low scores, indicating that the genus Thermoactinomyces may be a phylogenetically coherent group of organisms, as has been suggested in previous analyses (37, 38).

Fig. 5.

Phylogenetic tree for Firmicutes species. The tree was constructed using 16S rRNA sequences as described in the text. Organism names in green contain an sspE gene, organism names in red do not contain an sspE gene, and for organism names in black the completed genome sequence is not available and analysis of the presence or absence of the sspE gene has not been carried out. The red asterisk adjacent to B. selenitireducens indicates that this species almost certainly does not sporulate, as described in the text. The large green asterisk in the region between the ancestor of the Paenibacillus genus and the ancestor of the clade containing Alicyclobacillus species indicates the period in Firmicutes evolution when the sspE gene appeared. The numbers adjacent to interior branch points in the tree are bootstrap values.

Knowledge of the evolutionary positions of species that do and do not contain sspE genes (Fig. 5) now allows us to pinpoint the common ancestor that first acquired the sspE genes as between the ancestor of Paenibacillus species and that of Alicyclobacillus spp./B. tusciae, since P. polymyxa and Brevibacillus brevis contain an sspE gene, Thermoactinomyces thalpophilus contains an sspE gene, and its spores contain a γ-type SASP (Fig. 1) (19, 34). This analysis also predicts that Thermobacillus spp. will contain an sspE-like gene, while Pasteuria species as well as Bacillus schlegelii will lack an sspE gene. The determination of at least the Pasteuria penetrans genome sequence is in progress (21), so a definitive test of this prediction may soon be forthcoming. We note, however, that this overall interpretation assumes that an sspE gene did not emerge and was subsequently lost within these taxa. We also do not know how many other potentially informative ancestral taxa are not available for analysis due to Bacillales undersampling, and in addition, many Bacillales culture collection accessions are misidentified (17).

In addition to the absence of SASP-γ and an sspE gene from A. acidocaldarius and its closely related species, the absence of sspE genes has previously been noted in spore-forming Clostridiales species (7, 8, 32). Clearly a γ-type SASP is not essential for spore formation, spore stability, or spore resistance, although SASP-γ does provide an amino acid reserve that can be used in spore outgrowth (12, 30). However, this must not be an essential function, unlike the essential role in spore DNA protection for the α/β-type SASP present in all spore-forming Firmicutes. α/β-Type SASP degradation can also supply much amino acid for protein synthesis early in spore outgrowth (12, 29). Presumably, the additional gain in amino acid storage capacity in dormant spores that could be provided by a γ-type SASP does not provide spores with a significant evolutionary advantage, or this is compensated for in other ways. Indeed, at least under laboratory conditions, it is very difficult to demonstrate a major phenotypic effect of loss of SASP-γ from B. subtilis spores that contain normal levels of α/β-type SASP (12).

The likely absence of a γ-type SASP from A. acidocaldarius spores and the apparent absence of an sspE gene from B. tusciae as well suggest that spores of members of the clade containing these organisms do not contain a γ-type SASP. Perhaps knowing more details about the properties of spores of members of this clade in comparison with spores that do contain γ-type SASP may suggest possible additional functions for this extremely abundant spore protein, other than simply being an amino acid reservoir. In this regard it is perhaps noteworthy that A. acidocaldarius is an aerobe. Thus, γ-type SASP seem most likely not to play any significant role in spores' long-term tolerance to oxygen, while this might have been suggested as a possibility had the sspE gene appeared only in the transition between the anaerobic Clostridiales and the generally aerobic Bacillales species.