Animal Health Trust, Lanwades Park, Kentford, Newmarket, Suffolk CB8 7UU, UK1
Author for correspondence: Neil Chanter. Tel: +44 1638 751000. Fax: +44 1638 750794. e-mail: neil.chanter{at}aht.org.uk
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Keywords: truncated M-protein, phagocytosis, persistent colonization, guttural pouch colonization, Streptococcus equi
The EMBL accession numbers of the M-protein gene partial sequences of S. equi isolates A1A3 and BE are AJ249868AJ249874, respectively.
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Streptococcus pyogenes, an important human respiratory pathogen, produces a class of surface M-proteins with separate characteristic and conserved regions associated with cell wall anchorage (C-terminal), central tandem repeats and an N-terminal signal sequence (Kehoe et al., 1996 ). Following the signal sequence, some M-proteins have a hypervariable region that can elicit production of opsonic antibody, believed to be important for protective immunity, and that generally does not cross-react with other M-types. One or more separate M-proteins can be produced by individual strains, each with one or more host-protein-binding capabilities. In the absence of opsonic antibody, the opsonogenic M-proteins can increase resistance of cells to phagocytosis and this activity is dependent on the presence of fibrinogen. S. equi produces a surface M-protein (Timoney et al., 1997
) that also increases resistance of cells to phagocytosis in vitro in serum supplemented with equine fibrinogen (Chanter et al., 1994
) and it is highly opsonogenic in vivo, but not protective (Timoney & Eggers, 1985
).
The DNA sequence of the S. equi M-protein gene (Timoney et al., 1997 ) is more than 99% identical to the sequence of a recently detected S. equi gene for a fibrinogen-binding protein (Meehan et al., 1998
). Apart from regions involved in cell wall anchorage, cell wall spanning and the signal sequence, there is little sequence or structural similarity between the S. equi and S. pyogenes M-proteins. The amino acid residue repeat sequences within the central region of the S. equi protein are smaller and generally fewer in number than those present within the S. pyogenes proteins. Incongruity between the variation seen in the M-proteins of S. pyogenes and other markers of genetic variation suggests a good deal of horizontal gene transfer (Whatmore et al., 1994
). Further, the M-proteins of S. pyogenes can vary in size during infections and outbreaks due to intragenic recombination between identical central tandem repeats (Fischetti et al., 1986
), and insertion and point mutations can occur in the hypervariable N terminus (Harbaugh et al., 1993
). In contrast, the M-protein gene of S. equi does not appear to have transferred to the closely related Streptococcus zooepidemicus (Timoney et al., 1997
) and changes have hitherto not been detected in the M-protein, which has been regarded as invariant (Galan & Timoney, 1988
).
Four strangles outbreaks investigated by endoscopy after cases had convalesced were each shown to include persistent carriers in which S. equi most frequently resided in the guttural pouches, often associated with clinically silent inflammation (Newton et al., 1997 , 2000
). Some of these carriers were colonized by S. equi with a truncated M-protein gene. The objective of this study was to characterize the isolates and the nature of the M-protein gene truncation.
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
S. equi isolates.
One hundred and sixty-seven isolates of S. equi from separate cases of strangles were compared with 17 isolates from 17 outwardly healthy carriers. The 167 isolates were composed of 136 from clinical samples submitted to the Animal Health Trusts diagnostic laboratory from 40 veterinary practices in 22 counties of mainland Britain between 1990 and 1998, 14 from Australia, 7 from Ireland and 10 from Canada. The 17 isolates from 17 healthy animals were largely detected by repeated endoscopic examination and sampling of the guttural pouches of cases and contacts (n=150 horses) in 4 outbreaks (Newton et al., 1997 , 2000
). Fifteen of these isolates were from persistent carriers that had been either cases or contacts during the outbreaks. The sixteenth isolate was cultured from an outwardly healthy horse, also with guttural pouch empyema, shortly after it arrived in quarantine. The seventeenth isolate was from the only S. equi sample from a healthy pony with normal guttural pouches that was subsequently S. equi negative. Isolates were stored on Cryobeads (ProLab) at -50 °C after two or three passages on 5% (v/v) horse blood agar. Studies were conducted after 23 further subcultures of the frozen stock on 5% horse blood agar. All 184 isolates failed to ferment trehalose, sorbitol, ribose and lactose but fermented salicin, and all were Lancefield group C (Grant et al., 1993
) and 16S23S RNA gene intergenic spacer type D1, typical of S. equi (Chanter et al., 1997
). All isolates were tested by PCR1 (see below) for the M-protein gene after two further passages of single colonies on 5% horse blood agar. There were considerable variations in colony morphology after 18 h incubation at 37 °C. The colonies of many isolates were white, dry and 1 mm in diameter, whilst others were honey-coloured, glossy and 2 mm in diameter, fitting the typical description for S. equi. The colonies of all but one isolate had zones of haemolysis extending from 2 to 6 mm in diameter and this variation had no apparent relationship with other aspects of colony morphology. The different colony morphology of each of the isolates was stable with further passages.
PCR for M-protein gene fragments.
Three to five colonies were resuspended in 500 µl isotonic PBS. Resuspended bacteria or samples of the nasopharynx and guttural pouch were centrifuged at 10000 g for 5 min. The supernatant was discarded and the deposit was resuspended in 1x GeneAmp buffer II (Perkin Elmer; supplied as 10x concentration), 0·5% (v/v) Tween 20 and 100 µg proteinase K ml-1 (Sigma-Aldrich). A 25 µl aliquot of buffer was generally used, but if a large deposit was produced, buffer was added to at least twice the volume of the pellet. The samples were incubated at 55 °C for 30 min and then boiled for 5 min. After centrifugation at 10000 g for 5 min, the supernatant fraction was recovered for immediate PCR. Aliquots of the original samples were stored at -20 °C.
The position of the primers for PCR1 were selected to amplify most of the M-protein gene (Fig. 1) from DNA isolated from cultures. To reduce the possibility of amplification of a related sequence, the reverse primer was chosen from sequences just upstream of the putative anchor sequence to avoid potential homology with the genes of other cell-wall-anchored proteins. Details of the oligonucleotide sequences of the primers are given in Table 1
. Sample DNA (5 µl) was mixed with 2·5 µl 10x GeneAmp PCR buffer (Perkin Elmer), 0·5 µl 10 mM dNTP mix, 1 µl of each primer at 25 pmol µl-1, 0·1 µl (0·5 U) AmpliTaq (Perkin Elmer), 14·9 µl water and 30 µl light mineral oil for PCR (Sigma-Aldrich). The mixture was heated at 95 °C for 2 min, followed by 30 cycles of 94 °C for 1 min, 54 °C for 2 min and 72 °C for 3 min, and then a period of 7 min at 72 °C.
|
|
DNA sequencing.
DNA from the product of PCR1 was sequenced using the Big Dye Terminator Cycle Sequencing kit (Perkin Elmer) after purification with QIAquick columns (Qiagen). The DNA concentrations were estimated from absorbance at 260 nm and diluted in water to 3090 ng DNA µl-1. Diluted DNA (1 µl) was added to 8 µl Big Dye Terminator mix, 2 µl sequencing primer 1 or 2 (Fig. 1; Table 1
) at 1·6 pmol µl-1 and 9 µl water. The mixture was amplified by 25 cycles of 96 °C for 30 s, 50 °C for 15 s and 60 °C for 4 min. The DNA was precipitated from the sequencing reaction by mixing with 2 µl 3 M sodium acetate pH 4·6 and 50 µl 95% (v/v) ethanol for 15 min. After centrifugation at 10000 g for 20 min, the supernatant fraction was discarded, the pellet rinsed in 250 µl 70% ethanol, recentrifuged and air-dried at 37 °C for 10 min. The samples were then sequenced on an ABI Prism 377 apparatus (Perkin Elmer).
Immunoblotting.
Aliquots (1 ml) of cultures of S. equi grown overnight at 37 °C in ToddHewitt broth were centrifuged at 10000 g for 10 s. The pellets were resuspended in 100 µl Laemmli sample buffer and heated at 100 °C for 5 min before centrifugation at 10000 g for 5 min. The polypeptides in 35 µl of the supernatant fractions were separated by electrophoresis through 12% (w/v) acrylamide and molecular masses were estimated by comparison with those of pre-stained molecular mass standards (Sigma-Aldrich). After electrophoresis, the separated polypeptides were transferred onto nitrocellulose (Towbin et al., 1979 ) by electroblotting at 150 V for 4 h. The nitrocellulose sheet was air-dried before incubating for 1 h at 20 °C in 1% (w/v) dried skimmed milk dissolved in 0·1 % (v/v) Tween 20 in PBS (PBST) before washing in three changes of PBST and shaking for 1 h at 20 °C in mouse antiserum to recombinant M-protein (Meehan et al., 1998
), diluted 1/2000 in PBST. The blot was washed in five changes of PBST before incubation with shaking for 1 h at 20 °C in goat antiserum to mouse immunoglobulin conjugated to horseradish peroxidase (Dako) diluted 1/500 in PBST. After five washes in PBST, the blots were developed using the enhanced ECL Chemiluminescence system (Amersham Pharmacia Biotech) and autoradiography.
Phagocytosis.
Previously described methods for testing the susceptibility of S. equi to phagocytosis (Chanter et al., 1994 ) were used, with the exception that a standard serum from an animal known to have had no contact with strangles (stored at -70 °C) was used in place of specific-pathogen-free foal serum. Neutrophils were purified from a pony (also known to have had no contact with strangles) on different days for serial replicates of the test. As specified in the results, some cultures were grown with 30 µg hyaluronidase (Sigma-Aldrich) ml-1 to remove the hyaluronate capsule. Some of the phagocytosis tests were conducted with 3 mg equine fibrinogen (Sigma-Aldrich) ml-1 added since this has previously been shown to increase resistance to phagocytosis (Chanter et al., 1994
). Since fibrinogen also binds to the full M-protein (Meehan et al., 1998
), it was possible that deficiencies in resistance to phagocytosis by the truncated M-protein variants of S. equi would be more apparent under these conditions. Observations were made of colony morphology of pairs of isolates from the same outbreak, as described above, to see if there were variations in capsulation or haemolytic activity that might explain differences in resistance to phagocytosis. Portions of the bacterial cultures were also spread onto glass slides, dried, stained by the Gram method and examined microscopically for differences in bacterial cell chain length that might affect resistance to phagocytosis. Forty chains were counted from fresh cultures on three separate days for each strain of S. equi.
Statistical methods.
Fishers exact test was used to test the null hypothesis that there was no difference in the proportion of isolates that were variant (producing a truncated M-protein) between S. equi recovered from clinical cases (n=167) and outwardly healthy animals (n=17).
The Wilcoxon test was used to test the null hypothesis that there was no difference between variant and typical S. equi isolates in their resistance to phagocytosis by purified neutrophils in serum under different test conditions. Analyses were conducted between sequential isolates (A1, A2 and A3) from a single outwardly healthy carrier, as well as variant and typical isolates from separate horses in the same outbreaks. The effectiveness of the pairing of laboratory tests on variant and typical isolates in the Wilcoxon test was assessed using the Spearman non-parametric correlation coefficient, rs, and a P value calculated by GraphPad Prism Software; a positive rs and a low P value indicate effective pairing.
The unpaired t-test [after checking for equal variances (F-test)] was used to test the null hypothesis that there was no difference in mean chain length between variant and typical S. equi isolates from different horses, between and within outbreaks. The analyses were conducted using pooled data (total of 120 chains per isolate) comprising observations made on 40 chains on three separate days.
All statistical tests were conducted using GraphPad Prism Software with a significance level set at 5%.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Two earlier isolates were available from the carrier of isolate A3. PCR of A2, taken 2 months earlier, gave a product larger than obtained from A3 but smaller than expected of the full-length M-protein gene. PCR of A1, taken 6 months before A3 when the pony had typical signs of strangles, gave a product consistent with the presence of the full-length M-protein gene.
The PCR test gave a 1·5 kb product as expected of the full gene when applied to 166/167 S. equi isolates from 167 strangles cases in mainland Britain, Australia, Canada and Ireland (99% of active cases). Isolate B, from the remaining strangles case, gave a PCR product of approximately 1·2 kb. Isolation of variant S. equi was statistically significantly less likely from strangles cases than from outwardly healthy yet S. equi positive horses (Fishers exact test, P=0·0002).
Sequential nasopharyngeal or guttural pouch samples taken over several months were available from only two of the carriers (A3 and C). PCR amplification of these samples gave homogeneous products of sizes corresponding to the genes detected in the isolates. This suggested that only one type of S. equi, whether variant or typical, was present in the samples. Variant S. equi with an M-protein gene of comparable size to that of A3 were isolated from an empyemic guttural pouch on numerous occasions over a 2·5 month treatment period before infection and inflammation were eventually eliminated. The carrier of isolate C harboured variant S. equi for at least 2 years.
DNA sequence analysis of the truncated M-protein genes
DNA sequencing of the PCR products from the variant S. equi (Fig. 1) showed that the M-protein genes contained deletions of 363441 bp between the end of the signal sequence and the start of the central repeat region (Figs 1
and 2
). The deletions were in-frame and retained the signal sequence. The deletion sites in the M-protein genes from separate variants on different premises were slightly different, suggesting that they had emerged independently. The translated DNA sequence suggested that between 0 and 6 amino acid residues of the N terminus of the mature surface-expressed M-protein were retained in the variant S. equi from the healthy horses (A3, C, E, D). In contrast, the translated M-protein gene from the only variant that came from a strangles case (B) retained 26 amino acid residues of the N terminus. The region of the M-protein gene of isolate A1 encompassing, and just downstream of, the region of the deletions in isolates A2 and A3, had 8 bp variations from the published sequence (Timoney et al., 1997
), translating into four amino acid changes. The sequence for isolate A2 had a further 3 bp changes, two bases downstream of the deletion site. The M-protein gene sequence from A1 varied from the sequence published by Timoney et al. (1997)
and resulted in three amino acid residue changes (G63R, R143S and T181Q) that were also present in the sequence published by Meehan et al. (1998)
. Examination of the full-length M-protein gene over the putative deletion sites did not reveal any inverted repeat DNA sequences. A search of the DNA databases did not reveal any homology between the DNA sequence encompassing the deletions or flanking regions and any insertion elements or transposons.
|
Expression of a truncated M-protein from variant S. equi detected by antibody to recombinant M-protein
Several polypeptides in extracts of the variant S. equi separated by SDS-PAGE and transferred to nitrocellulose reacted with antiserum to the M-protein (Fig. 3a). No reaction was detected using pre-immune serum. Extracts of cultures of typical S. equi also contained several polypeptides reacting with the antiserum but their molecular masses were larger than for the variant S. equi. The slowest migrating and strongest reacting polypeptide in the typical S. equi was estimated to be approximately 17 kDa larger than the slowest and strongest reacting polypeptide in the variant S. equi. The calculated differences between the molecular masses of the truncated and full-length M-protein gene products were from 13·1 to 16·6 kDa. Polypeptides reacting with the antiserum in isolate B, with a larger M-protein gene than the other variants, had slightly higher molecular masses than the reactive polypeptides in the other variant S. equi (A3, C, D and E). The three serial isolates A1A3 contained polypeptides of incrementally smaller sizes reacting with the antiserum, proportional to the sizes of their M-protein genes (Fig. 3b
).
|
|
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The emergence of carriers of variant S. equi lacking the N terminus of the M-protein and the relative lack of such variants from active strangles cases may be an indication of the roles of different domains of M-protein in persistence and disease. One hypothesis regarding the mechanism of the emergence of the variants is that colonization for many months stimulates sufficient immunity to force an escape mutation. M-protein is a potent opsonogenic determinant and a large proportion of antibody responses are directed to the N terminus (Timoney et al., 1999 ). The occurrence of M-protein size variations in the same strain of S. pyogenes in a single patient and individual outbreaks (Fischetti et al., 1986
) and the appearance of point and insertion mutations in single M-types (Harbaugh et al., 1993
) could be due to the selective pressure of host immunity (Fischetti, 1991
). However, an effective assessment of the role of immunity in selecting an escape mutation in S. equi would require serial serum samples from animals that become carriers from before infection until after the emergence of variant S. equi. This would require large numbers of samples to be taken from many animals because it is not possible at the moment to predict which horses will become carriers and in which of these variants of S. equi will emerge. The mechanism of M-protein gene truncation is as unclear as the driving force behind the emergence of the variants. The deletion sites and the flanking DNA did not contain nucleotide sequences indicative of inverted repeats or potential insertion sequence or transposon activity. One feature of the process is the creation of nucleotide base variations in M-protein gene sequence in the regions immediately flanking the deletions, some of which result in changes in amino acid residues. Nonetheless, the identification of three serial isolates from a single carrier with different-sized M-protein genes suggests that the full extent of M-protein gene deletion in at least some variant S. equi might occur incrementally. Preliminary application of PFGE to RFLPs, as part of an ongoing, major molecular typing study, has identified eight types and isolates A1, A2 and A3 all fall into one of these (unpublished observations).
M-protein is strongly immunogenic and opsonogenic; it is exposed on the bacterial cell surface and increases resistance to phagocytosis by neutrophils in the presence of fibrinogen in vitro (Chanter et al., 1994 ), and it binds to fibrinogen (Meehan et al., 1998
). The region of M-protein responsible for fibrinogen binding by recombinant M-protein is the part missing from the truncated M-protein of variant S. equi (Meehan et al., 1999
). The greater occurrence of variant S. equi among carriers compared with strangles cases suggests that truncation of the M-protein could be linked to a reduction in virulence. On the other hand, a truncated M-protein may contribute significantly to persistence of S. equi in empyemic guttural pouches since the variants retained the ability to express approximately 80% of the M-protein, as indicated by the detection of truncated M-protein using antisera to recombinant M-protein, under selective pressure to lose 20% of its N terminus.
The reaction of antiserum to recombinant M-protein with several polypeptides in each isolate may be indicative of post-translational processing. The reactive polypeptides in each of the isolates with the full-length M-protein had very similar, if not identical, sizes. The sizes of the different reactive polypeptides in the variants were also of similar molecular masses (Fig. 3). Even those of isolate B, with a larger M-protein gene, gave a similar pattern of polypeptides to the other variants although they were consistently of slightly higher molecular mass. This pattern suggests specific post-translational cleavage sites within the M-proteins. Several polypeptides of different molecular masses with identical N-terminal amino acid sequences have been detected in single M-protein extracts, suggesting specific C-terminal truncation sites (Timoney & Mukhtar, 1993
) and this may explain the presence of the polypeptides described here. S. equi produces at least two easily detectable proteases (Collin et al., 1999
) and these could be involved in the cleavage of M-protein.
The variant S. equi from carriers had reduced resistance to phagocytosis, perhaps partly caused by the deletion in their M-protein genes, yet paradoxically these isolates persisted in their hosts in the presence of purulent inflammation. Full resistance to phagocytosis, however, is likely to be the sum of the activities of a number of S. equi products, including the capsule. Consequently, the reduced resistance to phagocytosis of the variant S. equi may only be partial and insufficient to compromise the ability to persist in the guttural pouches.
The extent of the deletion at the N terminus of the mature surface-exposed M-protein might particularly influence virulence. One variant was from a strangles case and this had 20 more amino acids of the N terminus than were found in the M-protein of variants from outwardly healthy carriers. However, the truncation of M-protein may only partly explain the apparent lack of virulence of variant S. equi from healthy horses. One variant lacked haemolytic activity and it is possible that the other variants have additional alterations to one or more putative virulence determinants. Nonetheless, the naturally occurring variations in M-protein in S. equi from carriers indicate potentially revealing genetic locations for the generation of in-frame allelic replacement mutants for use in experimental infections to characterize the importance of different domains of M-protein for virulence. The naturally emerging variant S. equi may be important for elevating herd immunity and could be used for live vaccination provided the M-protein truncation causes attenuation without compromising the ability to colonize and immunize. Better understanding of the potential immunological and epidemiological significance of these variants will come from detailed investigations of more strangles outbreaks by endoscopy and M-protein PCR, and by experimental tests of virulence and immunity from variant challenge.
![]() |
ACKNOWLEDGEMENTS |
---|
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Chanter, N., Collin, N., Holmes, N., Binns, M. & Mumford, J. (1997). Characterization of the Lancefield group C streptococcus 16S23S RNA gene intergenic spacer and its potential for identification and sub-specific typing.Epidemiol Infect 118, 125-135.[Medline]
Collin, N. C., Chanter, N. & Mumford, J. A. (1999). Detection and characterisation of Streptococcus equi and Streptococcus zooepidemicus proteases to identify potential protective immunogens. In Equine Infectious Diseases VIII: Proceedings of the Eighth International Conference, Dubai, United Arab Emirates, 23rd26th March 1998, pp. 535-537. Edited by U. Wernery, J. F. Wade, J. A. Mumford & O. R. Kaaden. Newmarket: R & W Publications.
Fischetti, V. A. (1991). Streptococcal M protein.Sci Am 264, 58-65.[Medline]
Fischetti, V. A., Jarymowycz, M., Jones, K. F. & Scott, J. R. (1986). Streptococcal M protein size mutants occur at high frequency within a single strain.J Exp Med 164, 971-980.[Abstract]
Galan, J. E. & Timoney, J. F. (1988). Immunologic and genetic comparison of Streptococcus equi isolates from the United States and Europe.J Clin Microbiol 26, 1142-1146.[Medline]
Grant, S. T., Efstratiou, A. & Chanter, N. (1993). Laboratory diagnosis of strangles and the isolation of atypical Streptococcus equi.Vet Rec 133, 215-216.[Medline]
Harbaugh, M. P., Podbielski, A., Hugl, S. & Cleary, P. P. (1993). Nucleotide substitutions and small-scale insertion produce size and antigenic variation in group A streptococcal M1 protein.Mol Microbiol 8, 981-991.[Medline]
Kehoe, M. A., Kapur, V., Whatmore, A. M. & Musser, J. M. (1996). Horizontal gene transfer among group A streptococci: implications for pathogenesis and epidemiology.Trends Microbiol 4, 436-443.[Medline]
Meehan, M., Nowlan, P. & Owen, P. (1998). Affinity purification and characterization of a fibrinogen-binding protein complex which protects mice against lethal challenge with Streptococcus equi subsp. equi.Microbiology 144, 993-1003.[Abstract]
Meehan, M., Nowlan, P. & Owen, P. (1999). Localisation of the ligand binding domain in the fibrinogen-binding protein complex which protects mice against lethal challenge with Streptococcus equi subsp. equi. Abstracts of the 142nd Ordinary Meeting of The Society for General Microbiology, p. 62.
Newton, J. R., Wood, J. L. N., Dunn, K. A., DeBrauwere, M. N. & Chanter, N. (1997). Naturally occurring persistent and asymptomatic infection of the guttural pouches of horses with Streptococcus equi.Vet Rec 140, 84-90.[Medline]
Newton, J. R., Verheyen, K., Talbot, N. C., Timoney, J. F., Lakhani, K., Wood, J. L. N. & Chanter, N. (2000). Control of strangles outbreaks by isolation of guttural pouch carriers identified using PCR and culture of Streptococcus equi. Equine Vet J (in press).
Timoney, J. F. & Eggers, D. (1985). Serum bactericidal responses to Streptococcus equi of horses following infection or vaccination.Equine Vet J 17, 306-310.[Medline]
Timoney, J. F. & Mukhtar, M. M. (1993). The protective M proteins of the equine group C streptococci.Vet Microbiol 37, 389-395.[Medline]
Timoney, J. F., Artiushin, S. C. & Boschwitz, J. S. (1997). Comparison of the sequences and functions of Streptococcus equi M-like proteins SeM and SzPSe.Infect Immun 65, 3600-3605.[Abstract]
Timoney, J. F., Artiushin, S. & Wang, J. (1999). Regions of the protective M-like protein of Streptococcus equi recognised by serum and mucosal antibodies of convalescent horses. In Equine Infectious Diseases VIII: Proceedings of the Eighth International Conference, Dubai, United Arab Emirates, 23rd26th March 1998, pp. 88-94. Edited by U. Wernery, J. F. Wade, J. A. Mumford & O. R. Kaaden. Newmarket: R & W Publications.
Towbin, H., Staehelin, T. & Gordon, J. (1979). Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications.Proc Natl Acad Sci USA 76, 4350-4354.[Abstract]
Whatmore, A. M., Kapur, V., Sullivan, D. J., Musser, J. M. & Kehoe, M. A. (1994). Non-congruent relationships between variation in emm gene sequences and the population genetic structure of group A streptococci.Mol Microbiol 14, 619-631.[Medline]
Received 20 December 1999;
revised 3 February 2000;
accepted 13 March 2000.