Identification of Genes of VSH-1, a Prophage-Like Gene Transfer Agent of Brachyspira hyodysenteriae

Abstract

VSH-1 is a mitomycin C-inducible prophage of the anaerobic spirochete Brachyspira hyodysenteriae. Purified VSH-1 virions are noninfectious, contain random 7.5-kb fragments of the bacterial genome, and mediate generalized transduction of B. hyodysenteriae cells. In order to identify and sequence genes of this novel gene transfer agent (GTA), proteins associated either with VSH-1 capsids or with tails were purified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The N-terminal amino acid sequences of 11 proteins were determined. Degenerate PCR primers were designed from the amino acid sequences and used to amplify several VSH-1 genes from B. hyodysenteriae strain B204 DNA. A λ clone library of B. hyodysenteriae B204 DNA was subsequently screened by Southern hybridization methods and used to identify and sequence overlapping DNA inserts containing additional VSH-1 genes. VSH-1 genes spanned 16.3 kb of the B. hyodysenteriae chromosome and were flanked by bacterial genes. VSH-1 identified genes and unidentified, intervening open reading frames were consecutively organized in head (seven genes), tail (seven genes), and lysis (four genes) clusters in the same transcriptional direction. Putative lysis genes encoding endolysin (Lys) and holin proteins were identified from sequence and structural similarities of their translated protein products with GenBank bacteriophage proteins. Recombinant Lys protein hydrolyzed peptidoglycan purified from B. hyodysenteriae cells. The identified VSH-1 genes exceed the DNA capacity of VSH-1 virions and do not encode traditional bacteriophage early functions involved in DNA replication. These genome properties explain the noninfectious nature of VSH-1 virions and further confirm its resemblance to known prophage-like, GTAs of other bacterial species, such as the GTA from Rhodobacter capsulatus. The identification of VSH-1 genes will enable analysis of the regulation of this GTA and should facilitate investigations of VSH-1-like prophages from other Brachyspira species.

Brachyspira hyodysenteriae is an anaerobic spirochete and the etiologic agent of swine dysentery, a severe, mucohemorrhagic intestinal disease afflicting animals in the postweaning period of growth (20). B. hyodysenteriae cells contain a defective prophage, VSH-1, which is induced when mitomycin C is added to cultures of growing cells (27). This prophage has been purified and characterized previously (28). VSH-1 particles resemble λ virions in morphology but, with a head diameter of 45 nm and a tail length of 64 nm, are substantially smaller.

VSH-1 virions contain random 7.5-kb fragments of host genomic DNA (28), making it difficult to identify VSH-1 genes. Purified virions are “noninfectious,” that is, they do not lyse spirochete cells when added to cultures of B. hyodysenteriae or of other Brachyspira species. VSH-1 transfers bacterial genes between B. hyodysenteriae cells (28, 52) and has likely contributed to the recombinant population structure of B. hyodysenteriae (55). This generalized transduction activity of VSH-1 is useful for constructing B. hyodysenteriae mutant strains (35). VSH-1 is the first natural gene transfer mechanism to be described for a spirochete.

The ability to recognize VSH-1-specific genes would be a major advance for investigating intracellular and intercellular activities of VSH-1 and for evaluating its ecological significance to its host bacterium. To overcome the inability to identify distinct VSH-1 genes carried by virions, structural protein genes were identified from the sequences of VSH-1 head and tail proteins. Genes for endolysin and holin proteins, proteins enabling the release of bacteriophage from bacterial cells, were discovered during nucleotide sequence analyses. Recombinant endolysin was shown to degrade peptidoglycan. Based on its genome and biological properties, VSH-1 resembles prophage-like gene transfer agents (GTA) reported for other bacteria (3, 32, 37, 46, 56).

MATERIALS AND METHODS

Separation of VSH-1 whole virions and tailless heads.

VSH-1 particles were harvested from B. hyodysenteriae strain B204 cultures (400 ml) in NT broth 7 to 8 h after mitomycin C treatment (28). Following polyethylene glycol precipitation, VSH-1 particles from five cultures were combined, extracted once with an equal volume of CHCl₃, and harvested by CsCl gradient ultracentrifugation (28). Two bands, one containing intact VSH-1 particles and the other containing tailless heads, were collected from the gradient. The virions and heads were concentrated by ultrafiltration (Microcon-100), resuspended in 200 μl of SM buffer (48), examined by electron microscopy (28), and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE).

SDS-PAGE and sequencing of VSH-1 proteins.

VSH-1 whole-virion proteins (80 μg of total protein) were separated by SDS-PAGE. The acrylamide concentrations of the resolving gels ranged from 12 to 15% depending on the molecular masses of the proteins targeted for separation. Proteins within gels were stained with Coomassie blue and electrotransferred to polyvinylidene difluoride membranes. Selected protein bands were cut from the membranes, and their N-terminal amino acid sequences were determined by Edman degradation and high-pressure liquid chromatography analysis at the Iowa State University Protein Facility (Ames, Iowa). Reagents, techniques, and equipment for electrophoresis, electroblotting, and sequencing of proteins have been described previously (28, 41, 51, 54).

Throughout this article, VSH-1 virion (structural) proteins are designated Hvp (hyodysenteriae viral protein), followed by their molecular mass as estimated from electrophoretic migration. For example, Hvp38 is the VSH-1 major capsid protein with an apparent molecular mass of 38 kDa and hvp38 is the gene encoding Hvp38.

VSH-1 genome sequencing.

A 1.1-kb region of VSH-1 DNA was amplified from B. hyodysenteriae B204 DNA by using degenerate PCR primers. The forward primer, 5′-AAAAT(T/A)AC(T/A)GAAAAAAA(T/C)AT was designed from the N-terminal sequence of Hvp19, and the reverse primer, 5′-TGAAT(T/A)CC(T/A)GCTTT(T/A)AT(T/A)AT, was based on the N-terminal sequence of Hvp13. Based on the 1.1-kb sequence of the amplicon, PCR primers were designed and used to amplify flanking DNA by the inverse-PCR method (44). B. hyodysenteriae B204 DNA was digested with AseI (New England Biolabs). DNA fragments were circularized by ligation, and samples (0.2 to 1.5 μg) were used as PCR templates, yielding a 3.6-kb PCR amplicon containing genes for the Hvp19, Hvp13, and Hvp38 proteins. The remaining VSH-1 genes and B. hyodysenteriae flanking genes were determined by a “chromosome walking” strategy. DNA probes complementary to sequenced DNA were used to screen by filter hybridization (48) a clone library of sheared B. hyodysenteriae B204 DNA prepared in λ ZAPII (Stratagene). The DNA from purified plaques was amplified in PCRs by using the hybridization probe as one primer and a second primer targeting the T3 or T7 sequence flanking the cloning site of the λ ZAP II vector. Purified PCR amplification products were sequenced (18) at the ISU Nucleic Acid Facility, Ames, IA. Each nucleotide base was determined at least twice from sequences of both DNA strands. Additionally, VSH-1 gene order and linkages were confirmed by PCR amplification, from B. hyodysenteriae DNA, of intergenic regions of both VSH-1 and B. hyodysenteriae genes.

PCR conditions.

For most amplifications of B. hyodysenteriae B204 DNA, AmpliTaq Gold DNA polymerase and GeneAmp reagents (Perkin-Elmer, Applied Biosystems) were used according to the manufacturer's recommendations, except that 0.2 μM PCR primers (0.4 μM for degenerate primers) and 2.5 mM MgCl₂ were used. Annealing temperatures varied between 45 and 55°C, depending on the melting temperature of the primer. Reaction mixes (100 μl) containing 100 to 200 ng of template DNA were incubated through 36 cycles in an UNO-Thermoblock thermal cycler (Biometra).

Cloning the VSH-1 endolysin gene (lys).

The VSH-1 endolysin gene (lys) was amplified from B. hyodysenteriae B204 genomic DNA (50 ng) by using a high-fidelity proofreading polymerase, PfuTurbo (Stratagene, La Jolla, CA). RF2F (5′-CACCAATCAAGGAGTTTAATAATATGAT) and RF1R (5′-ACCTTGTAATATTTTAAGAATAAT) were used as primers. The RF2F primer contained a 5′ sequence (underlined) complementary to the plasmid vector, a putative Shine-Dalgarno sequence (bold), and the lys start codon (ATG in italics). Amplification conditions consisted of 95°C for 2 min, followed by 35 cycles of 95°C for 30 s, 48°C for 30 s, and 72°C for 1 min and a final 10-min extension at 72°C. The lys gene amplicon was ligated into the plasmid expression vector pBAD102/D-TOPO (Invitrogen) and used to transform Escherichia coli One Shot Top10 chemically competent cells (Invitrogen). A recombinant strain, designated BAD/lys, was purified by colony subculture. The pBAD/lys plasmid in this strain was confirmed by sequence analysis to carry an authentic copy of the lys gene. E. coli Top10 cells were also transformed with plasmid pBAD102/D/lacZ, which carries a gene for β-galactosidase. E. coli strain BAD/lacZ was isolated and used as a positive control for protein expression experiments and as a negative control in peptidoglycan degradation assays.

Identifying VSH-1 endolysin activity.

Cells of E. coli strain BAD/lys and strain BAD/lacZ were cultured at 37°C in 1 liter of LB broth containing ampicillin (50 μg/ml). Expression of His-tagged Lys and LacZ proteins was induced by adding l-arabinose (0.003%, wt/vol, final concentration). Cells were disrupted in a French press and centrifuged at 48,000 × g for 2 h at 4°C. Fusion proteins were purified from the supernates by affinity column chromatography (His-Select Nickel; Sigma). The purity of recombinant Lys was assessed by SDS-PAGE with Coomassie blue staining and by Western immunoblot analysis targeting the His tag, following instructions of the manufacturer (Invitrogen).

E. coli cells from a 1-liter culture were treated with CHCl₃ to remove their outer membranes (5). The chloroform was removed, and the cells were washed twice with 50 mM Tris-HCl buffer (pH 7.0), resuspended at an optical density at 600 nm of 1.0 in Tris-HCl buffer with 0.1% Triton X-100, and held on ice. Purified recombinant Lys protein was added (0.2 μg/ml suspension), and lysis of the CHCl₃-treated bacteria was monitored as a reduction in suspension turbidity (optical density at 600 nm).

To assess the muralytic ability of recombinant Lys, peptidoglycan was purified from B. hyodysenteriae B204 cells cultured in a 12-liter fermentor (29, 51). Approximately 70 mg of peptidoglycan was obtained from 25 g (wet weight) of spirochete cells. Peptidoglycan was also obtained from E. coli cells (2). Samples (0.5 mg) of B. hyodysenteriae peptidoglycan, E. coli peptidoglycan, and crab shell chitin (Sigma) were incubated with VSH-1 endolysin (20 μg/ml) at 37°C in 50 μl of 50 mM sodium phosphate buffer (pH 6.5). After incubation, the reaction mixtures were centrifuged at 20,000 × g, and the supernatants were lyophilized and dissolved in 20 μl of 1% K₂B₇O₄. Reduced N-acetylamino sugars were detected at 585 nm by using the Morgan-Elson reaction (19). Mutanolysin (Sigma), with acetylmuramidase activity, and Serratia marcescens chitinase (Sigma) were used in parallel control assays.

Computer analysis of VSH-1 DNA sequences.

Sequence data were assembled and analyzed by using DNASIS (v. 7.11; Hitachi of America) and Vector NTI suite 8 (InforMax). PCR primers were designed by using Oligo version 6.0 (Molecular Biology Insights). Gene sequences were compared to GenBank sequences via BLASTX (NCBI website, National Library of Medicine). The BLASTP and PSI-BLAST programs were used to compare proteins with GenBank sequences and to identify proteins from their conserved domains (1, 36). VSH-1 holin was analyzed for intramembrane domains by TMpred (26) and SOSUI (25), accessed through EMBnet (http://www.ch.embnet.org/index.html) and the Tokyo University of Agriculture and Technology (http://sosui.proteome.bio.tuat.ac.jp/sosuiframe0.html), respectively.

Nucleotide sequence accession number.

The sequences of the VSH-1 genes and flanking B. hyodysenteriae genes have been deposited in GenBank under accession number AY971355 (1 April 2005, release date).

RESULTS

Identification and sequencing of VSH-1 virion proteins.

When polyethylene glycol-precipitated VSH-1 particles from mitomycin C-treated B. hyodysenteriae cultures were purified by CsCl density gradient ultracentrifugation, two bands, approximately 1 cm apart, appeared in the gradient. By electron microscopy, the lower, prominent band contained VSH-1 whole virions (Fig. 1A), whereas the upper, faint band contained tailless VSH-1 heads (Fig. 1B). Most of the tailless heads appeared somewhat electron dense and were similar in diameter to whole-virion heads, suggesting they are mature capsids separated from their tails.

FIG. 1.

Transmission electron micrographs of (A) VSH-1 virions and (B) VSH-1 heads. Preparations were negatively stained with 2% (wt/vol) phosphotungstic acid (pH 7.0). Bar = 0.1 μm.

Proteins of intact virions were separated by electrophoresis and extracted from gels by electroblotting, and their N-terminal amino acid sequences were determined. Unambiguous sequences of 22 to 25 amino acids were obtained for 11 proteins, nine of which are labeled in Fig. 2. Although not apparent in Fig. 2, two weakly staining proteins, Hvp24 and Hvp28, were detected under different electrophoresis and staining conditions.

FIG. 2.

SDS-PAGE of proteins from VSH-1 whole virions (WV) and heads (H). Molecular mass standards (kDa) are indicated at the right of the figure, and protein designations are indicated on the left. Each lane contained 5 μg of protein.

VSH-1 Hvp45, -38, -24, -22, -19, and -13 were present in both whole virions (Fig. 2, lane WV) and heads (Fig. 2, lane H) and were presumptive head proteins. By contrast, Hvp101, -60, -53, -32, and -28 either were not detected among the head proteins or were reduced in staining intensity. For this reason, they were considered potential tail proteins. The sequences of these 11 proteins were used to identify VSH-1 genes.

Identification and sequencing of VSH-1 head and tail genes.

It was assumed that VSH-1 genes would be contiguous within the bacterial host chromosome, similar to other prophages. Consequently, N-terminal amino acid sequences of VSH-1 proteins were used to design both forward and reverse degenerate PCR primers for amplification of adjacent VSH-1 genes. Of different primer combinations tested, a forward primer targeting the 5′ end of the hvp19 gene and a reverse primer targeting the 5′ end of hvp13 yielded a 1.1-kb product. The amplicon contained the predicted coding sequences for both Hvp19 and Hvp22 proteins (hvp19/22 gene), as discussed below.

A “genome walking” strategy using inverse PCR and overlapping B. hyodysenteriae chromosomal DNA fragments cloned in λ ZAPII enabled DNA sequences upstream and downstream of the 1.1-kb region to be determined. In total, a 29.1-kb segment of the B. hyodysenteriae genome, encompassing VSH-1 genes and flanking B. hyodysenteriae B204 genes, was sequenced and analyzed (Fig. 3). In subsequent experiments, the arrangement and orientation of VSH-1 genes and B. hyodysenteriae genes, depicted in Fig. 3, were confirmed by PCR amplification of intergenic regions from B. hyodysenteriae genomic DNA.

FIG. 3.

Map of genes identified in these studies. VSH-1 genes (filled arrows), B. hyodysenteriae genes (striped arrows), and ORF greater than 240 bp with no GenBank homolog (open arrows) are oriented according to their direction of transcription.

Genes hvp45, hvp19, hvp23, hvp13, hvp38, and hvp24, encoding head-associated (capsid) proteins, were arranged in a cluster (Fig. 3). Two unidentified open reading frames (ORF), orfA and orfB, were contained within this cluster. Tail-associated genes hvp53, hvp32, hvp101, and hvp28, along with orfC, orfD, and orfE, formed a second cluster. A gene encoding the Hvp60 protein was not located within the sequenced chromosomal region (Table 1), suggesting hvp60 is located elsewhere in the genome. Alternatively, the 60-kDa protein could be a B. hyodysenteriae protein. All of the identified VSH-1 structural genes, and the lysis genes described below, were in the same transcriptional orientation.

TABLE 1.

Genes and proteins of VSH-1 and B. hyodysenteriae identified in these studies^a

Gene designation	Protein mass (kDa)b		Protein identity	Identification basisc	Closest protein match (GenBank accession no.)
	Estimated	Predicted
VSH-1
hvp45	45	47.9	Head	Protein-gene match	None
hvp19/22	19	36.5	Head	Protein-gene match	None
hvp19/22	22	36.5	Head	Protein-gene match	None
hvp13	13	12.7	Head	Protein-gene match	None
hvp38	38	39.4	Head	Protein-gene match	None
orfA		11.6			None
orfB		15.1			None
hvp24	24	20.9	Head	Protein-gene match	None
hvp53	53	52.7	Tail	Protein-gene match	None
orfC		23.6			None
orfD		10.7			None
orfE		14.1			None
hvp32	32	31.1	Tail	Protein-gene match	None
hvp101	101	97.2	Tail	Protein-gene match	None
hvp28	28	21.5	Tail	Protein-gene match	None
lys		23.2	Endolysin	Enzyme activity	Putative endolysin of Salmonella phage epsilon15 (AAO06088); E = 5 × 10−29
orfF		33.9			None
hol		9.6	Holin	Predicted properties	None
orfG		14.2			None
B. hyodysenteriae
trep		107.6	Unknown		T. denticola hyp. proteind (AAS12502); E = 6 × 10−30
glt		41.8	Amino acid transport	Conserved domains	F. nucleatum hyp. Ser/Thr-Na+ symporter (AAL95344); E = 8 × 10−93
oxd		37.3	Fe-S oxidoreductase	Conserved domains	C. tetani hyp. Fe-S oxidoreductase (AAO34830); E = 3 × 10−38
mcpB		69.2	Methyl-accepting sensory protein	Conserved domains	B. hyodysenteriae hyp. McpA protein (AAP58978); E = 4 × 10−94
mcpC		65.5	Methyl-accepting sensory protein	Conserved domains	B. hyodysenteriae hyp. McpA protein (AAP58978); E = 3 × 10−89

^aA gene encoding Hvp60 (N′-M_K_MPYHFLRNKIYKLPPAPYINF ...) was not among the identified genes.

^bEstimated masses are based on gel electrophoretic migration (Fig. 2). Predicted masses are based on translated protein sequences.

^c“Protein-gene match” signifies that 22 N-terminal amino acids of the purified protein were identical to those of the protein predicted from the gene sequence.

^dhyp., hypothetical.

The staining intensities of Hvp38 and Hvp53 relative to those of other proteins (Fig. 2) suggest that Hvp38 and Hvp53 are, respectively, major head and tail proteins of VSH-1 virions. Two VSH-1 proteins, Hvp19 and Hvp22, appeared as products of posttranslational modification (that is, proteolytic cleavage) of a primary gene product. Hvp19 was encoded by the 5′ end of hvp19/22, and Hvp22 (N-terminal sequence, MQKLKNVLEKLISEEKEIEKQAR) was encoded by the 3′ end (bp 468 to 951). One other protein, Hvp28, displayed the electrophoretic mobility of a protein with a mass greater than that predicted (21.5 kDa) from the hvp28 sequence. The difference may signify posttranslational modification of the hvp28 gene product, although there is no direct evidence for this modification.

Identification of VSH-1 lysis genes.

None of the translated VSH-1 head and tail proteins had significant sequence similarity with GenBank sequences. Neither did any putative proteins encoded by orfA to orfF interspersed among these structural protein genes. Downstream of the hvp28 tail gene (Fig. 3), however, two ORF, lys and hol, were identified as genes likely to encode proteins important for the escape of VSH-1 virions from B. hyodysenteriae cells.

The gene lys was predicted to encode a 22.9-kDa protein (197 amino acids) with conserved domains typical of glycoside hydrolase/chitinase enzymes (CD00325; PSI-BLAST, NCBI Entrez Database). Lys shares 38% sequence identity over 191 amino acids with a putative endolysin, a peptidoglycan-degrading enzyme, of enterobacterial phage epsilon15. As described below, the recombinant Lys protein was found to degrade peptidoglycan.

The gene hol was predicted to encode a protein with characteristics of bacteriophage class II holin proteins (57). It is small (9.6 kDa, 85 amino acids) and has a hydrophilic C-terminal end and two hydropathic, predictably transmembrane, domains (amino acids 15 to 37 and 49 to 71).

Identification of B. hyodysenteriae genes.

Flanking the VSH-1 gene clusters were genes homologous to bacterial genes (Fig. 3). Predicted proteins encoded by mcpB and mcpC genes have conserved domains of bacterial methyl-accepting sensory proteins (smart00283; E values of 6 × 10⁻⁴⁶ and 3 × 10⁻⁴¹, respectively). The McpB and McpC proteins share 50% sequence identity, and they have 34 to 35% identity with the putative McpA protein of B. hyodysenteriae.

Upstream of VSH-1 genes, bacterial genes were oriented in the opposite transcriptional direction (Fig. 3). The predicted Glt protein had conserved domains of monovalent cation/dicarboxylate symporters (COG1301; E value of 7 × 10⁻⁵⁹) and high overall sequence similarity with putative amino acid transporters from various bacterial genera (closest match, 52% identity over 386 amino acids with the serine/threonine-Na⁺ symporter of Fusobacterium nucleatum). The translated product of the oxd gene had conserved domains predicted for Fe-S oxidoreductase enzymes (COG1242; E value of 7 × 10⁻⁵⁹) and high overall sequence similarity with oxidoreductases from various bacteria (closest match, 33% identity over 315 amino acids with a predicted Fe-S oxidoreductase of Clostridium tetani). The Trep protein shared sequence similarity with conserved putative proteins with unknown function from Treponema spp. (closest match, 25% identity over 663 amino acids with a hypothetical protein of Treponema denticola; NP 972591; E value of 6 × 10⁻³⁰).

Peptidoglycan-degrading activity of VSH-1 Lys.

Recombinant Lys was purified by affinity chromatography from the soluble fraction of E. coli BAD/lys cells that had been treated with l-arabinose to induce Lys production. Approximately 1 mg of the recombinant protein was routinely purified from a 1-liter culture.

Neither cell lysis nor loss of viability was observed when Lys was added (100 μg/ml, final concentration) to E. coli or to B. hyodysenteriae cell suspensions, indicating that the outer membranes of these bacteria are a barrier to the Lys protein. By contrast, E. coli cells treated with CHCl₃ to remove their outer membranes were rapidly disrupted by purified Lys (Fig. 4). In separate experiments (data not shown), EDTA treatment of E. coli cells, which produces holes in the outer membrane without killing the bacteria (39, 40), made the cells sensitive to lysis by both VSH-1 Lys and egg white lysozyme.

FIG. 4.

Lysis of CHCl₃-treated E. coli cells by VSH-1 Lys protein. E. coli cells treated with CHCl₃ to remove their outer membranes were incubated at 25°C with 0.2 μg/ml recombinant Lys (•), 0.2 μg/ml β-galactosidase (▴), or no enzyme (▪).

Purified Lys protein hydrolyzed B. hyodysenteriae peptidoglycan, yielding N-acetylamino sugars detected by the Morgan-Elson reaction (Fig. 5). Similar activity was observed when B. hyodysenteriae peptidoglycan was digested with mutanolysin (Fig. 5). In separate assays, Lys did not hydrolyze crab chitin. These results indicated that recombinant VSH-1 Lys is a muralytic enzyme that cleaves the polysaccharide chains of B. hyodysenteriae peptidoglycan (19).

FIG. 5.

Peptidoglycan hydrolysis by VSH-1 Lys protein. Purified B. hyodysenteriae peptidoglycan (0.5 mg) was incubated at 37°C with 1 μg of recombinant Lys (•), 1 μg mutanolysin (▴), or no enzyme (▪). Amino sugars (reducing sugars) released from peptidoglycan were detected at 585 nm by the Morgan-Elson reaction.

DISCUSSION

The findings of these investigations indicate that VSH-1 has features of commonly known bacteriophages but clearly is a different type of prophage. VSH-1 genes, hvp45 through orfG (Fig. 3), span a 16.3-kb region of the B. hyodysenteriae B204 genome. Identified genes and predicted ORF are oriented in the same transcriptional direction and organized in head (seven genes), tail (seven genes), and lysis (four genes) modules. The clustering of similar-function genes and the serial head-tail-lysis arrangement of the clusters are typical of late operons of various temperate prophages (4, 7, 11, 21, 32). VSH-1 genes have an average G + C content of 28 mol% (range of 20 to 33 mol%).

VSH-1 genes identified by matching amino acid and nucleotide sequences include genes encoding capsid proteins Hvp45, -38, -24, -22, -19, and -13 and tail proteins Hvp101, -53, -32, and -28 (Table 1). Based on their staining intensities (Fig. 2), Hvp38 and Hvp53 are predominant VSH-1 structural proteins and are likely to be, respectively, major capsid and tail proteins. Capsid proteins Hvp19 and Hvp22 appear as derivatives of a truncated primary gene product involved in VSH-1 capsid assembly. Posttranslational modifications of bacteriophage gene products, notably proteolytic cleavage of capsid precursor proteins, commonly occur during virion assembly (22, 38, 43).

Significant similarities between VSH-1 structural proteins and those of other bacteriophages were not detected during analyses of GenBank sequences by us or during targeted searches of phage-specific databases (S. Casjens and R. W. Hendrix, personal communications). Thus, proteins conserved among double-stranded DNA tailed phages, such as terminase, head portal protein, tail tape measure protein, and tail fiber protein (7, 8, 23, 24), could not be identified. A tail assembly protein may be encoded by orfE, based on a potential translational frameshift site (TTTTTTG) within the gene (R. W. Hendrix, personal communication) and the previous finding of such sites in tail assembly genes of double-stranded DNA bacteriophages of phylogenetically diverse bacteria, including a prophage of the spirochete Borrelia burgdorferi (58).

An inability to recognize VSH-1 capsid and tail homologs of other phage proteins is not surprising. Characterized phage proteins in computer databases are still few, relative to the diversity of bacteriophages in nature, and are biased towards certain bacterial taxonomic groups (7, 17). It is also possible that additional VSH-1 genes are located elsewhere in the B. hyodysenteriae genome. A gene encoding Hvp60 (Fig. 2; Table 1) was not found among the VSH-1 genes. While the gene could encode a VSH-1 protein, alternatively, it might encode a B. hyodysenteriae protein associated with VSH-1 particles. The sequence and chromosome location of this gene will likely be identified when the current B. hyodysenteriae genome sequencing project is completed (T. La and D. Hampson, unpublished data).

Consistent with the lysis-mediated release of VSH-1 virions (27), the VSH-1 lys gene encodes an endolysin that was identified from its conserved glycoside hydrolase domains, ability to lyse CHCl₃-treated E. coli cells (Fig. 4), and peptidoglycan-degrading activity (Fig. 5). The VSH-1 hol gene likely encodes a holin protein. The hol gene is strategically positioned near lys. The Hol protein has predicted biochemical features, notably two transmembrane domains, typical of phage class II holins (57). Holins enable endolysins to gain access to peptidoglycan and are important in timing the release of virions (57). Direct evidence for its holin identity would come from demonstrations that the Hol protein forms oligomers, enters membranes, and complements heterologous endolysin in a λ holin-deficient mutant (9, 57).

What is VSH-1? Although VSH-1 resembles bacterial viruses of the Siphoviridae family of tailed phages (28), both biological and genomic properties distinguish this agent from traditional bacteriophages. VSH-1 preparations are noninfectious and a single virion is incapable of self-propagation, based on indirect and direct evidence as follows. VSH-1 identified structural and lysis genes comprise 16.3 kb of DNA. These late-function genes generally represent about 50 to 60% of the genome of self-replicating prophages, such as λ (4). Thus, an entire complement of VSH-1 early- and late-function genes would predictably span 30 kb of DNA. A VSH-1 virion which packages 7.5-kb DNA fragments (28) would need to carry multiple DNA fragments for self-propagation. Potential early genes were not among the identified VSH-1 genes (Fig. 3). Finally, we have not detected increases in VSH-1 gene copy number, a sign of phage gene replication, during mitomycin C induction of VSH-1 (unpublished data). These considerations support the classification of VSH-1 as a GTA-like defective prophage (7). Our use of the term “defective” is intended to indicate that VSH-1 functions in a different manner from traditional prophages and to distinguish it from those phages. VSH-1 appears quite effective in its relationship with B. hyodysenteriae.

The defective prophages of R. capsulatus and Methanococcus voltae resemble small bacteriophages, package short (4.4- to 4.5-kb) random segments of their host bacterial DNAs, and perform generalized transduction (3, 16, 32, 37, 56). The characteristics of VSH-1 and the well-studied GTA from R. capsulatus are compared in Table 2. R. capsulatus GTA gene expression is coordinately regulated with bacterial motility genes through a bacterial two-component signal transduction system (31). Neither environmental nor intracellular factors controlling VSH-1 gene expression are known, although mitomycin C induction implicates the involvement of RecA during an SOS response (10). The availability of VSH-1 gene sequences will facilitate measurements of VSH-1 induction through quantitative mRNA transcript analysis. VSH-1 does not resemble bacteriophages of other spirochete genera (13), specifically the characterized, infectious bacteriophages of Leptospira biflexa (47) and φBB-1, the 32-kb circular-plasmid phage of B. burgdorferi (14, 15). A possible GTA-like prophage has been reported for B. burgdorferi (13) and was predicted from an analysis of recombination in that species (12). Nevertheless, we have not found VSH-1 protein homologs among the sequenced genomes of B. burgdorferi, Treponema pallidum, T. denticola, and Leptospira interrogans serovar icterohaemorrhagiae.

TABLE 2.

Characteristics of Brachyspira hyodysenteriae VSH-1 and Rhodobacter capsulatus GTA prophages^a

Comparison of:
B. hyodysenteriae VSH-1	R. capsulatus GTA
Virions carry random 7.5-kb fragments of host genome (28)	Virions carry random 4.5-kb fragments of host genome (49, 59)
Noninfectious; generalized transduction ability (28, 35, 52)	Noninfectious; generalized transduction ability (37, 50, 56)
18 genes/ORF (16.3 kb DNA) for head, tail, lysis functions (this study)	19 ORF (15 kb DNA) for putative head and tail functions (33, 34)
Released by host cell lysis; endolysin and putative holin genes identified (this study)	Escape mechanism unknown (37); lysis genes not identified (33)
Spontaneously produced by growing cells (52)	Maximum production by bacteria in stationary growth phase (31, 33, 50)
Induced by mitomycin C (27, 28)	Not induced by mitomycin C (37, 49, 59)
Prophage induction presumably regulated by RecA-associated SOS response	Prophage induction regulated by two-component signal transduction system (31)
VSH-1-like agents distributed widely among Brachyspira species (6, 53)	GTA-like agents distributed widely among R. capsulatus strains (37, 56); GTA genes detected in other species of Rhodobacter and other α-proteobacteria (34)
Specific integration site in B. hyodysenteriae B78T and Brachyspira pilosicoli P43/6/78T chromosomes (53, 60)	Specific integration site in R. capsulatus chromosome (33)
GenBank accession no. AY971355	GenBank accession no. AF181080

^aNumbers in parentheses correspond to reference numbers.

Several observations suggest that VSH-1-like GTA have been and continue to be involved in the lifestyles of the genus Brachyspira, a distinct branch in spirochete evolution (45). B. hyodysenteriae population structure has been shaped by a high level of genetic recombination (55). As the only gene transfer mechanism described for Brachyspira, VSH-1 has likely contributed to this genetic diversity (55). VSH-1 genes have an average GC content of 28 mol%, comparable to the 26 mol% of Brachyspira DNA (30, 42), an indication that VSH-1 genes have long been present in Brachyspira species. VSH-1 prophages are widely distributed among Brachyspira species. VSH-1-like genes and inducible virions have been detected in 27 strains of six different Brachyspira species (6, 53). Unfortunately nothing is known about these prophages, including their ability to transduce host genes. Hopefully, the information provided in these investigations will stimulate further study of these VSH-1-like agents.

ADDENDUM IN PROOF

The genome sequence of prophage LE1 from the spirochete Leptospira biflexa was recently described by Bourhy et al. (J. Bacteriol. 187:3931-3940, 2005). We found no significant similarities among VSH-1 predicted protein sequences and those of phage LE1.