Characterization of the fibrinogen binding domain of bacteriophage lysin from Streptococcus mitis

The binding of bacteria to human platelets is a likely central mechanism in the pathogenesis of infective endocarditis. Platelet binding by Streptococcus mitis SF100 is mediated in part by a lysin encoded by the lysogenic bacteriophage SM1. In addition to its role in the phage life cycle, lysin mediates the binding of S. mitis to human platelets via its interaction with fibrinogen on the platelet surface. To better define the region of lysin mediating fibrinogen binding, we tested a series of purified lysin truncation variants for their abilities to bind this protein. These studies revealed that the fibrinogen binding domain of lysin is contained within the region spanned by amino acid residues 102 to 198 (lysin(102-198)). This region has no sequence homology to other known fibrinogen binding proteins. Lysin(102-198) bound fibrinogen comparably to full-length lysin and with the same selectivity for the fibrinogen Aα and Bβ chains. Lysin(102-198) also inhibited the binding in vitro of S. mitis to human fibrinogen and platelets. When assessed by platelet aggregometry, the disruption of the lysin gene in SF100 resulted in a significantly longer time to the onset of aggregation of human platelets than that of the parent strain. The preincubation of platelets with purified lysin(102-198) also delayed the onset of aggregation by SF100. These results indicate that the binding of lysin to fibrinogen is mediated by a specific domain of the phage protein and that this interaction is important for both platelet binding and aggregation by S. mitis.     

Prevalence of the amylase-binding protein A gene (abpA) in oral streptococci

Salivary amylase binds specifically to a number of oral streptococcal species. This interaction may play an important role in dental plaque formation. Recently, a 585-bp gene was cloned and sequenced from Streptococcus gordonii Challis encoding a 20.5-kDa amylase-binding protein (AbpA). The goal of this study was to determine if related genes are present in other species of oral streptococci. Biotinylated abpA was used in Southern blot analysis to screen genomic DNA from several strains representing eight species of oral streptococci. This probe hybridized with a 4.0-kb HindIII restriction fragment from all 13 strains of S. gordonii tested. The probe did not appear to bind to any restriction fragments from other species of amylase-binding oral streptococci including Streptococcus mitis (with the exception of 1 of 14 strains), Streptococcus crista (3 strains), Streptococcus anginosus (1 strain), and Streptococcus parasanguinis (1 strain), or to non-amylase-binding oral streptococci including Streptococcus sanguinis (3 strains), Streptococcus oralis (4 strains), and Streptococcus mutans (1 strain). Primers homologous to sequences within the 3' and 5' ends of abpA yielded products of 400 bp following PCR of genomic DNA from the Southern blot-positive strains. Several of these PCR products were cloned and sequenced. The levels of similarity of these cloned products to the abpA of S. gordonii Challis ranged from 91 to 96%. These studies reveal that the abpA gene appears to be specific to S. gordonii and differs from genes encoding amylase-binding proteins from other species of amylase-binding streptococci.     

Streptococcus mitis Phage-Encoded Adhesins Mediate Attachment to α2-8-Linked Sialic Acid Residues on Platelet Membrane Gangliosides

The direct binding of bacteria to human platelets contributes to the pathogenesis of infective endocarditis. Platelet binding by Streptococcus mitis strain SF100 is mediated in part by two bacteriophage-encoded proteins, PblA and PblB. However, the platelet membrane receptor for these adhesins has been unknown. In this study, we demonstrate that these proteins mediate attachment of bacterial cells to sialylated gangliosides on the platelet cell surface. Desialylation of human platelet monolayers reduced adherence of SF100, whereas treatment of the platelets with N- or O-glycanases did not affect platelet binding. Treatment of platelets with sialidases having different linkage specificities showed that removal of α2-8-linked sialic acids resulted in a marked reduction in bacterial binding. Preincubation of SF100 with ganglioside GD3, a glycolipid containing α2-8-linked sialic acids that is present on platelet membranes, blocked subsequent binding of this strain to these cells. In contrast, GD3 had no effect on the residual binding of platelets by strain PS344, an isogenic ΔpblA ΔpblB mutant. Preincubating platelets with specific monoclonal antibodies to ganglioside GD3 also inhibited binding of SF100 to platelets, but again, they had no effect on binding by PS344. When the direct binding of S. mitis strains SF100 and PS344 to immobilized gangliosides was tested, binding of PS344 to GD3 was reduced by 70% compared to the parent strain. These results indicated that platelet binding by SF100 is mediated by the interaction of PblA and PblB with α2-8-linked sialic acids on ganglioside GD3.Among the viridans group streptococci, Streptococcus mitis is a leading cause of infective endocarditis (14). The binding of these microorganisms to human platelets appears to be a central process in the pathogenesis of this disease (20). Platelet binding by S. mitis strain SF100, an endocarditis isolate, was previously shown to be mediated in part by proteins PblA and PblB (3, 4). The genes encoding these proteins reside within the temperate bacteriophage SM1, a member of the Siphoviridae family (23). PblA and PblB are unusual because neither protein has strong similarity to known bacterial adhesins, but instead they resemble structural components of bacteriophages. Disruption of pblA and pblB results in phage particles lacking tails, further indicating that these proteins are involved in phage morphogenesis (4). Expression of the SM1 phage-encoded holin and lysin results in the permeabilization of a subpopulation of SF100 cells, thereby releasing PblA and PblB into the surrounding medium. Following this, both PblA and PblB attach to choline residues within the cell walls of viable bacteria, where they mediate the binding of viable bacteria to platelets. This novel method of adhesin expression on the S. mitis cell surface directly links bacterial platelet binding activity to disease pathogenesis, as deletion of pblA and pblB results in a significant loss of virulence as measured in an animal model of endocarditis (20).The interaction of bacteria with sialylated oligosaccharides on host molecules may be a factor in the colonization of the oral cavity by viridans group streptococci by allowing them to adhere to salivary glycoproteins (27). Sialic acid is commonly found as a terminal sugar on oligosaccharides on host glycoproteins and glycolipids. These oligosaccharides serve as receptors for bacterial adhesion (13, 21, 25, 26), in addition to acting as a hydrolysable source of metabolic sugars (7). For example, the serine-rich repeat glycoprotein GspB of Streptococcus gordonii recognizes α2-3-linked sialic acid oligosaccharides present on host receptors and mediates adhesion of S. gordonii to salivary mucin (27). While GspB may allow S. gordonii to colonize the oral cavity in a commensal fashion, it can also contribute to the virulence of the organism in the setting of infective endocarditis by promoting attachment of bacterial cells to platelets via the platelet glycoprotein GP1b (5, 26, 29).Gangliosides are sialylated glycosphingolipids present in the outer leaflet of the plasma membrane of all mammalian cells. They serve as receptors for a variety of bacteria and bacterial products, such as Haemophilus influenzae, Helicobacter pylori, Neisseria meningitidis, and cholera toxin (1, 6, 11, 15, 19, 28). Two gangliosides (GM3 and GD3) have previously been shown to be present on platelet membranes (10, 18). GM3 is a monosialylated ganglioside that is located in lipid microdomains and is further sialylated to GD3 during an early stage of platelet activation that precedes calcium and granule release (18). GD3 has also been shown to participate in signaling through the FcλRIIa receptor in platelets (10).Although platelet binding by SF100 is mediated predominantly by PblA and PblB, the platelet binding site for this interaction has not been characterized. In this study, we demonstrate that α2-8-linked sialic acid residues on platelet membrane ganglioside GD3 are the principal targets for PblA/PblB-associated binding to human platelets.     

Platelet receptors for the Streptococcus sanguis adhesin and aggregation-associated antigens are distinguished by anti-idiotypical monoclonal antibodies.

Platelets aggregate in response to an adhesin and the platelet aggregation-associated protein (PAAP) expressed on the cell surfaces of certain strains of Streptococcus sanguis. We sought to identify the corresponding PAAP receptor and accessory adhesin binding sites on platelets. Since the adhesion(s) of S. sanguis for platelets has not been characterized, an anti-idiotype (anti-id) murine monoclonal antibody (MAb2) strategy was developed. First, MAb1s that distinguished the adhesin and PAAP antigens on the surface of S. sanguis I 133-79 were selected. Fab fragments of MAb1.2 (immunoglobulin G2b [IgG2b]; 70 pmol) reacted with 5 x 10(7) cells of S. sanguis to completely inhibit the aggregation of human platelets in plasma. Under similar conditions, MAb1.1 (IgG1) inhibited the adhesion of S. sanguis cells to platelets by a maximum of 34%, with a comparatively small effect on platelet aggregation. Together, these two MAb1s inhibited S. sanguis-platelet adhesion by 63%. In Western immunoblots, both MAb1s reacted with S. sanguis 133-79 87- and 150-kDa surface proteins and MAb1.2 also reacted with purified type I collagen. The hybridomas producing MAb1.1 and MAb1.2 were then injected into BALB/c mice. Enlarged spleens were harvested, and a panel of MAb2 hybridomas was prepared. To identify anti-ids against the specific MAb1s, the MAb2 panel was screened by enzyme-linked immunosorbent assay for reaction with rabbit polyclonal IgG antibodies against the 87- and 150-kDa antigens. The reactions between the specific rabbit antibodies and anti-ids were inhibited by the 87- and 150-kDa antigens. When preincubated with platelets, MAb2.1 (counterpart of MAb1.1) inhibited adhesion to platelets maximally by 46% and MAb2.2 (anti-MAb1.2) inhibited adhesion to platelets maximally by 35%. Together, both MAb2s inhibited the adhesion of S. sanguis to platelets by 81%. MAb2.2 also inhibited induction of platelet aggregation. MAb2.2 immunoprecipitated a biotinylated platelet membrane antigen of 170 kDa (unreduced); MAb2.1 precipitated membrane antigens of 175- and 230-kDa (unreduced). Therefore, platelet binding sites and the receptor for the S. sanguis adhesin and PAAP, respectively, are distinguished by the anti-id MAb2s.     

Morphology of temperate bacteriophage SfV and characterisation of the DNA packaging and capsid genes: the structural genes evolved from two different phage families

The entire genome of SfV, a temperate serotype-converting bacteriophage of Shigella flexneri, has recently been sequenced (Allison, G.E., Angeles, D., Tran-Dinh, N., Verma, N.K. 2002, J. Bacteriol. 184, 1974-1987). Based on the sequence analysis, we further characterised the SfV virion structure and morphogenesis. Electron microscopy indicated that SfV belongs to the Myoviridae morphology family. Analysis of the proteins encoded by orf1, orf2, and orf3 revealed that they were homologous to small and large terminase subunits, and portal proteins, respectively; the protein encoded by orf5 showed homology to capsid proteins. Western immunoblot of the phage with anti-SfV sera revealed two antigenic proteins, and the N-terminal amino acid sequence of the 32-kDa protein corresponded to amino acids 116 to 125 of the ORF5 protein, suggesting that the capsid may be processed. Functional analysis of orf4 showed that it encodes the phage capsid protease. The proteins encoded by orfs1, 2, 3, 4, and 5 are homologous to similar proteins in the Siphoviridae phage family of both gram-positive and gram-negative origin. The capsid and morphogenesis genes are upstream and adjacent to the genes encoding Myoviridae (Mu-like) tail proteins. The organisation of the structural genes of SfV is therefore unique as the head and tail genes originate from different morphology groups.     

Discrimination of Streptococcus pneumoniae from viridans group streptococci by genomic subtractive hybridization

Two oligonucleotide primer sets for the discrimination of Streptococcus pneumoniae from "pneumococcus-like" oral streptococcal isolates by PCR were developed. Genomic subtractive hybridization was performed to search for differences between Streptococcus pneumoniae strain WU2 and the most closely related oral streptococcus, Streptococcus mitis strain 903. We identified 19 clones that contained S. pneumoniae-specific nucleotide fragments that were absent from the chromosomal DNA of typical laboratory strains of S. mitis and other oral bacteria. Subsequently, oligonucleotide PCR primers for the detection of S. pneumoniae were designed from the sequences of the subtracted DNA fragments, and the specificities of the 19 primer sets were evaluated by PCR using chromosomal DNAs extracted from four S. pneumoniae clinical isolates and from 20 atypical organisms classified as S. mitis or S. oralis, which harbored genes encoding the pneumococcal virulence factors autolysin (lytA) or pneumolysin (ply), as templates. Of the 19 primer sets, two (Spn9802 and Spn9828) did not amplify PCR products from any of the pneumococcus-like streptococcal strains that we examined. The genes containing the Spn9802 and Spn9828 sequences encoded proteins of unknown function that did not correspond to any previously described proteins in other bacteria. These new oligonucleotide primers may be very useful for early and correct diagnosis of S. pneumoniae infections.     

Clumping factor A mediates binding of Staphylococcus aureus to human platelets

The direct binding of bacteria to platelets may be an important virulence mechanism in the pathogenesis of infective endocarditis. We have previously described Staphylococcus aureus strain PS12, a Tn551-derived mutant of strain ISP479, with reduced ability to bind human platelets in vitro. When tested in an animal model of endocarditis, the PS12 strain was less virulent than its parental strain, as measured by bacterial densities in endocardial vegetations and incidence of systemic embolization. We have now characterized the gene disrupted in PS12 and its function in platelet binding. DNA sequencing, Southern blotting, and PCR analysis indicate that PS12 contained two Tn551 insertions within the clumping factor A (ClfA) locus (clfA). The first copy was upstream from the clfA start codon and appeared to have no effect on ClfA production. The second insertion was within the region encoding the serine aspartate repeat of ClfA and resulted in the production of a truncated ClfA protein that was secreted from the cell. A purified, recombinant form of the ClfA A region, encompassing amino acids 40 through 559, significantly reduced the binding of ISP479C to human platelets by 44% (P = 0.0001). Immunoprecipitation of recombinant ClfA that had been incubated with solubilized platelet membranes coprecipitated a 118-kDa platelet membrane protein. This protein does not appear to be glycoprotein IIb. These results indicate that platelet binding by S. aureus is mediated in part by the direct binding of ClfA to a novel 118-kDa platelet membrane receptor.     

Characterization and immunogenicity of pyrogenic mitogens SePE-H and SePE-I of Streptococcus equi.

Two pyrogenic mitogens, SePE-H and SePE-I, were characterized in Streptococcus equi, the cause of equine strangles. SePE-H and SePE-I have molecular masses of 27.5 and 29.5 kDa, respectively, and each is almost identical to its counterpart in Streptococcus pyogenes M1. Both genes are adjacent to a gene encoding a phage muramidase of 49.7 kDa and are located immediately downstream from a phage genomic sequence almost identical to a similar phage sequence in S. pyogenes M1. Strong mitogenic responses were elicited by both proteins from horse peripheral blood mononuclear cells. However, although both were pyrogenic for rabbits, only SePE-I was pyrogenic in ponies. Convalescent sera contained antibody to each mitogen and horses recovered from strangles or immunized with SePE-I were resistant to the pyrogenic effect of SePE-I. The immunogenicity of SePE-I suggests that it should be included in new generation strangles vaccines. In isolates of S. equi sepe-I and sepe-H were consistently present but they were absent from the closely related Streptococcus zooepidemicus, suggesting that phage mediated transfer was an important event in the formation of the clonal, more virulent, S. equi from its putative S. zooepidemicus ancestor.     

Identification of a Porphyromonas gingivalis receptor for the Streptococcus gordonii SspB protein

Colonization of the plaque biofilm by the oral pathogen Porphyromonas gingivalis is favored by the presence of antecedent organisms such as Streptococcus gordonii. Coadhesion between P. gingivalis and S. gordonii can be mediated by the SspB protein of S. gordonii; however, the P. gingivalis cognate receptor for this protein has not been identified. In this study, we identified a surface protein of P. gingivalis that interacts with the SspB protein. Coprecipitation between P. gingivalis outer membrane proteins and purified SspB protein demonstrated that a 100-kDa P. gingivalis protein bound to SspB. The 100-kDa protein also bound to an engineered strain of Enterococcus faecalis that expresses the SspB protein on the cell surface. Monospecific polyclonal antibodies to the 100-kDa protein inhibited the binding between P. gingivalis and S. gordonii in a dose-dependent manner up to 86%. Amino acid sequencing of the 100-kDa protein showed homology to a protein previously identified as the P. gingivalis minor fimbria. The minor fimbrial protein may exist as a complex with a hemagglutinin-like protein since the genes encoding these proteins are adjacent on the chromosome and are cotranscribed. Thus, the P. gingivalis receptor for S. gordonii SspB is a 100-kDa protein that structurally may be a minor fimbria-protein complex and functionally effectuates coadhesion.     

An Eikenella corrodens toxin detected by plaque toxin-neutralizing monoclonal antibodies.

Bacterial plaque from the gingival region of teeth contains cytotoxic agents which lyse undifferentiated human HL60 cells. A small panel of monoclonal antibodies (MAbs) was found to abrogate much of this activity and to detect antigens in certain strains of Streptococcus mitis and Eikenella corrodens. The aim of this study was to determine whether these bacterial antigens might be involved in HL60 cells cytolysis. Saline extracts were obtained by homogenizing washed, stationary-phase cells in 65 mM NaCl with a tight-fitting Potter-Elvehjem homogenizer. The extracts of E. corrodens were toxic to HL60 cells, whereas similar extracts of S. mitis were nontoxic. Adding plaque toxin-neutralizing MAb 3hE5 blocked the toxic effect of E. corrodens extract S. mitis extracts contained a single, strongly reactive antigen of 140 kDa (s140K antigen) detected on Western blots (immunoblots) by three MAbs from the panel. Rabbit antibodies raised to this antigen excised from the gel (anti-s140K serum) detected larger antigens in addition to s140K. E. corrodens extracts contained a number of antigens detected by the MAbs. Immunoglobulin G (IgG) was purified from anti-s140K serum by passage through DE52 cellulose. A 100-fold excess (by weight) of the purified IgG to E. corrodens protein specifically cross-precipitated an 80-kDa antigen plus a nonantigenic 16-kDa protein, presumably attached noncovalently. The remaining supernatant fraction had no toxic activity. A similar ratio of control IgG (from nonimmunized rabbits) did not precipitate these proteins, and the supernatant fraction had the same activity as the extract not treated with IgG. The proteins of 80 and 16 kDa were also detected in the anti-s140K immunoprecipitate by rabbit IgG antibodies to E. corrodens whole cells. The 80-kDa antigen, alone or complexed with the 16-kDa protein, may be involved in mediating the toxic activity in E. corrodens and plaque extracts.     

Genome sequence of the temperate bacteriophage PH10 from <Emphasis Type="Italic">Streptococcus oralis</Emphasis>

Exponential growing cultures of Streptococcus oralis strain OMZ 1038, isolated from human supragingival dental plaque, were found to release a bacteriophage (designated PH10) upon treatment with mitomycin C. The complete genome sequence of phage PH10 was determined. The genome was 31276 bp in size and contained 54 open reading frames. The module encoding structural proteins was highly similar to that of Streptococcus pneumoniae prophage PhiSpn_3. The most abundant phage structural protein was encoded by ORF35 and was likely processed by proteolytic cleavage. The putative endolysin from PH10, which contained a muramidase domain and a choline-binding domain, was purified and shown to have lytic activity with S. oralis, S. pneumoniae and Streptococcus mitis, but not with other streptococcal species.     

The Variable Region of Pneumococcal Pathogenicity Island 1 Is Responsible for Unusually High Virulence of a Serotype 1 Isolate

Streptococcus pneumoniae is the leading infectious cause of death in children in the world. However, the mechanisms that drive the progression from asymptomatic colonization to disease are poorly understood. Two virulence-associated genomic accessory regions (ARs) were deleted in a highly virulent serotype 1 clinical isolate (strain 4496) and examined for their contribution to pathogenesis. Deletion of a prophage encoding a platelet-binding protein (PblB) resulted in reduced adherence, biofilm formation, reduced initial infection within the lungs, and a reduction in the number of circulating platelets in infected mice. However, the region''s overall contribution to the survival of mice was not significant. In contrast, deletion of the variable region of pneumococcal pathogenicity island 1 (vPPI1) was also responsible for a reduction in adherence and biofilm formation but also reduced survival and invasion of the pleural cavity, blood, and lungs. While the 4496ΔPPI1 strain induced higher expression of the genes encoding interleukin-10 (IL-10) and CD11b in the lungs of challenged mice than the wild-type strain, very few other genes exhibited altered expression. Moreover, while the level of IL-10 protein was increased in the lungs of 4496ΔPPI1 mutant-infected mice compared to strain 4496-infected mice, the levels of gamma interferon (IFN-γ), CXCL10, CCL2, and CCL4 were not different in the two groups. However, the 4496ΔPPI1 mutant was found to be more susceptible than the wild type to phagocytic killing by a macrophage-like cell line. Therefore, our data suggest that vPPI1 may be a major contributing factor to the heightened virulence of certain serotype 1 strains, possibly by influencing resistance to phagocytic killing.     

Repeats in an extracellular protein of weakly pathogenic strains of Streptococcus suis type 2 are absent in pathogenic strains.

Streptococcus suis type 2 strains that are pathogenic for pigs produce a 110-kDa extracellular protein factor (EF). Nonpathogenic and weakly pathogenic strains do not produce EF or produce a protein (EF*) that is immunologically related to EF. To study the pathogenesis of S. suis type 2 in pigs and to develop tools and methods for the control of S. suis type 2 infections, we cloned and characterized the genes encoding EF and various EF* proteins. Analysis of the deduced amino acid sequences showed that the first 833 amino acids at the N terminus of the EF and EF* proteins were nearly identical. The proteins differed, however, at their C termini. Unlike the 110-kDa EF protein, the EF* proteins contained several repeated units of 76 amino acids. The number and arrangement of the repeats in the EF* proteins varied. The data suggest that the gene encoding EF could have evolved from an epf* gene by a specific deletion event. The lack of repeated amino acid units in the EF protein may be related to virulence.     

Crossing the barrier: evolution and spread of a major class of mosaic pbp2x in Streptococcus pneumoniae, S. mitis and S. oralis

A paradigm for Streptococcus interspecies gene transfer is represented by the mosaic pbp genes encoding the target enzymes for beta-lactam antibiotics, the penicillin-binding proteins, in Streptococcus pneumoniae. We investigated a collection of oral streptococci from three continents by comprehensive multi-locus sequence typing analysis in order to trace the origin of a mosaic block belonging to a dominant family of mosaic pbp2x implicated in penicillin resistance of S. pneumoniae. One widespread family of mosaic pbp2x occurred in all three distinct clusters of S. pneumoniae, Streptococcus mitis and Streptococcus oralis, documenting independent inter- and intraspecies recombination events. Moreover, potential ancestor genes of this mosaic block could be identified in two penicillin-susceptible S. mitis strains from South Africa and Spain, facilitating the identification of pbp2x mutations relevant for resistance development.     

Identification and Molecular Analysis of Rough-Colony-Specific Outer Membrane Proteins of Actinobacillus actinomycetemcomitans

Actinobacillus actinomycetemcomitans, a gram-negative bacterium isolated from the human mouth, has been implicated in the pathogenesis of early-onset periodontitis. Primary isolates cultured from subgingival plaque exhibit an adherent, rough colony phenotype which spontaneously converts to a nonadherent, smooth phenotype upon in vitro subculture. The rough colony variant produces abundant fimbriae and autoaggregates, while the smooth colony variant is planktonic and produces scant fimbriae. To begin to understand the significance of colony variation in biofilm formation by A. actinomycetemcomitans, outer membrane protein profiles of four isogenic rough and smooth colony variants were compared by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Two proteins with relative molecular masses of 43 and 20 kDa were expressed by the rough colony variants exclusively. Expression of these proteins was not found to be dependent on growth phase, oxygen tension, or type of complex medium. N-terminal amino acid sequences of these proteins obtained by Edman degradation were compared with sequences from the University of Oklahoma A. actinomycetemcomitans genome database. Two contiguous open reading frames (ORFs) encoding proteins having sequence homology with these proteins were identified. The 43-kDa protein (RcpA [rough colony protein A]) was similar to precursor protein D of the general secretion pathway of gram-negative bacilli, while the 20-kDa protein (RcpB [rough colony protein B]) appeared to be unique. The genes encoding these proteins have been cloned from A. actinomycetemcomitans 283 and sequenced. A BLASTX (gapped BLAST) search of the surrounding ORFs revealed homology with other fimbria-related proteins. These data suggest that the genes encoding the 43-kDa (rcpA) and 20-kDa (rcpB) proteins may be functionally related to each other and to genes that may encode fimbria-associated proteins.     

Isolation of extracytoplasmic proteins from Serpulina hyodysenteriae B204 and molecular cloning of the flaB1 gene encoding a 38-kilodalton flagellar protein.

Extracytoplasmic proteins were released from Serpulina (Treponema) hyodysenteriae (strain B204) by treatment of whole cells with a nonionic detergent (Tween 20). Centrifugation of the Tween 20-released proteins at 100,000 x g sedimented 10 major extracytoplasmic proteins with approximate molecular masses of 44, 43.5, 42, 39, 38, 34, 33.5, 33, 31, and 29 kDa. Treatment of the sedimented fraction with 6 M urea solubilized all of the proteins except the 39-kDa protein. Peptide sequences were obtained for the purified 42-, 39-, 38-, 34-, 31-, and 29-kDa proteins. The peptide sequences of the 42-, 38-, and 31-kDa proteins indicate that they likely are components of the periplasmic flagella. The amino-terminal peptide sequence of the 38-kDa protein was used to design an oligonucleotide probe and to clone an S. hyodysenteriae DNA fragment containing the gene encoding this protein. The predicted 290-amino-acid protein sequence derived from the cloned gene was highly homologous to those of several other bacterial flagellar proteins and is preceded by consensus sigma D nucleotide sequences found upstream of other flagellar genes. On the basis of its similarity to the FlaB proteins of other spirochetes, we propose to designate the cloned S. hyodysenteriae gene flaB1 and its encoded protein FlaB1. Vaccination of pigs with FlaB1 or its recombinant counterpart did not protect them from an experimental challenge.