M Protein Gene Typing of Streptococcus pyogenes by ...

Download PDF
3 downloads 0 Views 1MB Size Report
Comment
A new approach for the typing of Streptococcus pyogenes is described. Oligonucleotide probes of 30 nucleotides in length were derived from currently known ...
JOURNAL OF CLINICAL MICROBIOLOGY, Sept. 1992, p. 2391-2397

Vol. 30, No. 9

0095-1137/92/092391-07$02.00/0 Copyright © 1992, American Society for Microbiology

M Protein Gene Typing of Streptococcus pyogenes by Nonradioactively Labeled Oligonucleotide Probes ACHIM KAUFHOLD,l* ANDREAS PODBIELSKI,' DWIGHT R. JOHNSON,2 EDWARD L. KAPLAN,2 AND RUDOLF LUJTICKEN1 Institute of Medical Microbiology, Technical University (RWTH) Aachen, Pauwelsstrasse 30, D-5100 Aachen, Germany, 1 and Department of Pediatrics and World Health Organization Collaborating Center for Reference and Research on Streptococci, University of Minnesota Medical School, Minneapolis, Minnesota 554552 Received 14 April 1992/Accepted 10 June 1992

A new approach for the typing of Streptococcus pyogenes is described. Oligonucleotide probes of 30 nucleotides in length were derived from currently known sequences of the N-terminal regions of M protein genes (emm genes). The oligonucleotides were labeled with digoxigenin-dUTP and hybridized to dot-blotted genomic DNA from 116 group A streptococcal strains of serotypes M-1, M-2, M-3, M-5, M-6, M-12, M-18, M-19, M-24, and M-49. Hybridization reactions were visualized with a chemiluminescent substrate. In comparison with conventional serological typing of expressed M proteins, the binding of the probes to the corresponding emm genes exhibited 100% sensitivity and specificity. The results emphasize the high degree of type-specific conservation of the N-terminal regions of emm genes from reference strains and epidemiologically unrelated U.S. and European clinical isolates. The existence of two distinct genetic subgroups among eight investigated M-49 strains was unequivocally shown by hybridization assays and further confirmed by nucleotide sequence data obtained from four selected M-49 strains. Because oligonucleotide probes are relatively easy to prepare, easy to handle, and known to give consistent interlaboratory results, the "oligotyping" technique appears to offer potential advantages over conventional serological typing methods.

Streptococcus pyogenes (a Lancefield group A streptococcus) continues to be a significant cause of suppurative and nonsuppurative human infections throughout the world (1). Accurate identification of group A streptococci is an essential tool in studies of the epidemiology, pathogenesis, and therapy of streptococcal infections. Streptococcal serological classification systems are based primarily on the presence of specific M and T protein antigens. The M protein, a fibrillar protein on the cell surface that confers to the streptococcus the ability to resist phagocytosis in the absence of type-specific antibodies, is the most specific marker. Therefore, despite some shortcomings, serological identification of group A streptococci by use of the specific M protein is the well-established "gold standard" for reliable strain identification (21). Characterization of group A streptococci may also be accomplished by the detection of serum opacity factor (OF) production, which is consistently associated with specific M protein-containing streptococcal strains (5, 26-28). Since the first description of the streptococcal M protein over 60 years ago by Rebecca Lancefield (12), many studies have elucidated the function, immunochemistry, genetic structure, and antigenic variations of this bacterial molecule (for a review see reference 4). At present, about 80 serologically different types of this protein are known. After the publication of at least partial nucleotide sequences of several M protein genes (emm genes) from different types, we decided to explore the possibility of applying this information to identifying M serotypes of group A streptococci by the detection of type-specific emm genes. Initial experiments in our laboratory had suggested the usefulness of a molecular approach to detecting specifically the serotype 1 M protein *

gene (18). We have extended these studies, and in the present communication we describe a reproducible and reliable typing system that uses a panel of nonradioactively labeled oligonucleotide probes.

MATERIALS AND METHODS Bacterial strains. Several clinical isolates (especially a number of strains of serotype M-1) and reference strains of group A streptococci used in this study have been described previously (18). The sources and numbers of strains investigated in the present study are summarized in Table 1. Altogether, 116 strains of 10 serotypes were included in this study. Forty of these strains of group A streptococci of different M protein serotypes, obtained from the strain collection of the World Health Organization Collaborating Center for Reference and Research on Streptococci, University of Minnesota, Minneapolis, were coded and sent for a blinded evaluation of the probe technology to the Institute of Medical Microbiology, Aachen, Germany. To reduce the possibility of clonal association between strains of a given serotype, we selected strains from a variety of infection types, from widely diverse geographic locations, and from different years. Serological M protein typing was done by the standard immunodiffusion method with hydrochloric acid extracts of bacterial cultures and absorbed type-specific antisera raised in rabbits (21). Nucleic acid techniques. Streptococcal genomic DNA was prepared by the protocol of Huang et al. (9) from 30-ml bacterial cultures in brain heart infusion broth. DNA was blotted in 20-p.g amounts onto Biodyne B membranes (Pall BioSupport, Dreieich, Germany) by conventional methods

(23).

Corresponding author. 2391

2392

J. CLIN. MICROBIOL.

KAUFHOLD ET AL.

TABLE 1. Reference strains and clinical isolates of group A streptococci investigated in the present study No. of strains investigated from each of the following strain collections:

Serological M protein type

1 2 3 5 6 12 18 19 24 49 Total

Aachena

Pragueb

Minneapolisc

31 6 9 1 1 6 1 1 0 1 57

2 2 2 2 2 1 2 2 2 2 19

4 5 5 4 5 5 5 1 1 5 40

a These strains were obtained from the collection of the Institute of Medical Microbiology, Aachen, Germany. The majority of the M-1 strains were listed in a recent publication (18), whereas the strains of the other serotypes were primarily pharyngitis isolates collected from 1975 to 1976. b These strains were reference strains and clinical isolates obtained from the collection of the Center of Epidemiology and Microbiology, National Institute of Public Health, Prague, Czechoslovakia. c These strains were from the collection of the World Health Organization Collaborating Center for Reference and Research on Streptococci, University of Minnesota, Minneapolis, and were isolated primarily from 1988 to 1990 from sick patients in various states throughout the United States (a few isolates originated from Eastern Asia) or were standard reference strains.

Published nucleotide sequence data for emm genes (emm 1[7],emm 2 [19], emm 5 [15], emm 6 [8], emm 12 [20], emm 18 [19], emm 19 [19], emm 24 [16], and emm 49 [6]) and the unpublished N-terminal sequence of emm 3 (kindly provided by Jim Dale and Robert W. Baird, Memphis, Tenn., for our studies) were analyzed with the aid of the PC GENE (IntelliGenetics, Mountain View, Calif.) and OLIGO (National Biosciences, Plymouth, Minn.) computer programs to construct suitable oligonucleotide probes. Sequences (all but 1 of 30 nucleotides in length) that were complementary to the coding strands of parts of the N-terminal regions of the respective emm genes were selected as probes. The sequences of these oligonucleotide probes are shown in Table 2. The oligonucleotides were prepared on a model 200A DNA synthesizer (Beckman Instruments,

Munich, Germany). Further preparation of the oligonucleotides and hybridization assays were done as previously described in detail (18). In brief, the oligonucleotides were desalted by the use of a Sephadex G-50 minicolumn (Pharmacia, Freiburg, Germany) and dissolved in distilled water. Labeling of the oligonucleotides with digoxigenin-dUTP (Boehringer, Mannheim, Germany) was carried out as follows: 10 ,u of a 5 ,uM oligonucleotide solution (450 ng of oligonucleotide for a 30-mer), 2.5 ,u of digoxigenin-dUTP (100 p,M final concentration), 2.5 RI of lOx tailing buffer (1.4 mM potassium cacodylate, 300 mM Tris [pH 7.2], 10 mM CoCl2, 1 mM dithiothreitol), and 20 U of terminal transferase (Pharmacia) were mixed, distilled water was added to 25 pI, and the solution was incubated for 4 h at 37°C. For visualization of hybridization reactions, the chemiluminescent substrate 3-(2'-spiroadamantan)-4-methoxy-4-(3 "-phosphoryloxy)-phenyl-1,2-dioxetan (Boehringer) was used in accordance with the instructions of the manufacturer. The polymerase chain reaction (PCR) for amplification of the emm genes of serotype M-49 strains, agarose gel electrophoresis, and Southern blotting were carried out as previously described (19). As described previously (19), the PCR products of emm genes of four M-49 strains were purified by electroelution from agarose gels and ligated into the polylinker site of plasmid vector pUC18. Escherichia coli DH5ao was used as a host for the plasmid vector. With pUC standard primers, the N termini of the cloned PCR products were sequenced with the aid of an automated DNA sequencer (Applied Biosystems, Weiterstadt, Germany). Given that the nomenclature of emm genes, emm-like genes, and emm-related genes is currently a subject of debate, for reasons of convenience the genes are simply referred to as emm genes in this paper. RESULTS Since the nucleotide sequence corresponding to the N-terminal amino acids of the mature M protein is thought to be type specifically conserved, oligonucleotide probes were derived from this part of the M protein molecule. Computer analysis confirmed that the published N-terminal sequences of the genes of different serotypes showed the least homology to each other and negligible homology to other known

TABLE 2. Oligonucleotide probes corresponding to N-terminal sequences of M protein genes (emm genes) Nucleotide positions of the probe sequenceb

Nucleotide sequence of the probea

Probe

Reference or source from which the emm probe sequence was derived

emml emm2 emm3

5' TTC TAT AAC TTC CCT AGG ATT ACC ATC ACC 3' 5' TGC TTC TTT TTT GAC AGG GAC AGG GTT CTT 3' 5' CAT GTC TAG GAA ACT CTC CAT TAA CAC TCC 3'

1183-1154 180-151

emm5 emm6 emml2 emml8 emml9 emm24 emm49H emm49K emm49HKl emm49HK2

5' 5' 5' 5' 5' 5' 5' 5' 5' 5'

GGC 3' CCT 3' ATG 3'

198-169 611-582 1435-1406

TAA 3' CAC 3'

184-155 180-151 327-298 193-164 193-185, 172-164

19 Jim Dale and Robert W. Baird 15 8 20 19 19 16 6 6

223-194 253-224

6 6

CGG GTC ATT TAT TGC TTT GTC CGG ACG TTG TTT TTC COGT CTT TAT TGT ATC TTC TGG CGT TTC TTG TAC TTT ATT CTC CGC AAC ATT CTC CGC AGC TTT TTC TCT TCT TGT AAG ATC GGC

TGT GTT TGC CTG ATG TTC CTC CTC TGC GAT

ACC CCT AGT CAC TTC TAC CGT CCC GAC TAA ATC ACT CTG TAG CTC GAG CCT AGT ATA ACG CAG AGT ATC TGT AAC TTT AGC CTC

CTG 3'

AAC 3'

3' AAC AAC GCT AGA CAC GTT 3' TTG GTC GTA TAG CTC 3'

a The sequence is complementary to the coding strand of the corresponding emm gene sequence. The positions correspond to those in the originally published sequence.

b

7

VOL. 30, 1992

M PROTEIN GENE TYPING OF S. PYOGENES 2

3

4

5

2

6

A A

*

*

*

0

3

4

5

6

7

8

9

10

2393 11

12

S

B

C B D

C

FIG. 1. Genomic DNA from several strains of serotypes M-1, M-2, M-3, and M-12 was dot blotted onto a nylon membrane and hybridized with the digoxigenin-labeled emml2 probe. Except for the Lancefield reference strains, the strain designations are those of the Institute of Medical Microbiology, Aachen, Germany. Dots 1 to 6 in row A contained the following M-12 strains: Lancefield reference strain T12/126, Di 1122, Di 1161, Di 1215, Di 1217, and M 8785, respectively. Dots 1, 2, and 6 in row B contained M-3 strains Di 960, Di 652, and Di 944, respectively. Dots 3, 4, and 5 in row B contained M-1 strains Di 714, Di 741, and Di 876, respectively. Dots 1 to 6 in row C contained the following M-2 strains: Lancefield reference strain T2/44/RB4, Di 1083, Di 1093, Di 1180, Di 1198, and Di 1286, respectively. Note the specific identification of the emm genes of all serotype M-12 strains.

streptococcal genes. Since it has been demonstrated that at least 10 p,g of genomic DNA must be dot blotted for the unambiguous detection of a hybridization reaction (18), we decided to use 20 ,ug of target DNA in every assay. In that study (18), it was also shown that the hybridization and washing steps for a probe of 30 nucleotides in length can be carried out at 42°C without a loss of specificity. To screen for multiple point mutations in the type-specific segments of the emm genes, we performed these reaction steps in duplicate at 45 and 50°C. Thus, the stringency obtained in the procedures would not allow more than two or three mismatches, depending on the probe species. For visualization of hybridization signals, X-ray films (Cronex 4; Du Pont de Nemours, Bad Homburg, Germany) were exposed to the blotted membranes. All blotted membranes were hybridized sequentially with the different probes. After each reaction, the probe was reliably removed by boiling of the blotted membrane in 0.1 x SSC (lx SSC is 0.15 M NaCl plus 0.015 M sodium citrate)-O.lx sodium dodecyl sulfate (SDS) for 2 min. Even after nine stripping procedures, no substantial loss of the hybridization signal was observed. The labeled oligonucleotide probes could be stored at -20°C for at least 6 months without a loss of quality. The new probe technique was initially used to evaluate 76 strains for which serological M protein typing results were already known (i.e., strains obtained from Prague as well as reference strains and clinical isolates obtained from Aachen; Table 1). All of the isolates of serotypes M-2, M-3, M-5, M-6, M-12, M-18, M-19, and M-24, as well as both M-1 strains obtained from Prague and 16 of 31 M-1 strains obtained from Aachen were tested with all described probes and identified with 100% sensitivity and specificity (Fig. 1). The remaining 15 strains of serotype M-1 from the Aachen strain collection

E

FIG. 2. Genomic DNA from 40 strains of group A streptococci of different serotypes, obtained from Minneapolis (Table 1), and from several other isolates was dot blotted onto a nylon membrane and hybridized with the digoxigenin-labeled emm6 probe. Distinct and specific signals were obtained with emm genes from the five M-6 strains.

were only tested with the M-1-specific probe, which identified all these strains unambiguously. The results obtained for serotype M-49 are described below. Following these investigations, 40 strains of different serotypes, obtained from Minneapolis (Table 1), were examined in a one-sided blinded experimental design, i.e., without disclosure of the M serotypes of these strains. Again, all of the probes were hybridized sequentially with dot-blotted genomic DNA from these strains, and all group A streptococcal strains of the above-mentioned serotypes were identified, with concordant results for the serological method and the probe technology (Fig. 2). In comparison with those of strains of all other investigated M serotypes, the N-terminal regions of strains of serotype M-49 were found to be more heterogeneous. When dot-blotted M-49 genomic DNA was hybridized to a probe (designated emm49H) whose sequence corresponded to nucleotide positions 164 to 193 (i.e., spanning the codons of amino acids 5 to 14 of the mature protein) of the sequence published by Haanes and Cleary (6), only three of eight investigated M-49 strains were identified. The deduced amino acid sequence of the M-49 protein of Haanes and Cleary (6) differs by a four-amino-acid residue insertion relative to pepM49, the sequence of which was published by Khandke et al. (11). Probe emm49H spans the gene segment of 12 nucleotides encoding the four-amino-acid insertion. Therefore, another probe of 18 nucleotides, directed against the same region but without the 12-nucleotide insertion sequence, was constructed. This probe (designated emm49K) specifically identified the remaining five M-49 strains. The strains previously detected by probe emm49H, when retested with probe emm49K, showed no or very weak hybridization signals, depending on the stringency of the hybridization assay. However, when probes (designated emm49HK1 and emm49HK2; for details, see Table 2) directed to sequences located further downstream of the sequences used for probes emm49H and emm49K were applied, all available M-49 strains were specifically recognized. The results of the dot blot assays were further confirmed as follows. The emm genes of the eight M-49 strains were amplified by the PCR with universal M primers as described recently (19). Subsequently, Southern blots of the amplified

2394

J. CLIN. MICROBIOL.

KAUFHOLD ET AL. 1 2 34 5 6 7 8 9

as 11 1 2 3 4 5 6 7 8 9 10 11

a

A

a

B

l"

5

6 7 8 9 10 11

ar

*..

1 2 3 4

&S

C

FIG. 3. Southern blotting following agarose gel electrophoresis of the emm 49 genes of eight serotype M-49 strains after amplification by PCR and sequential hybridization with three different digoxigenin-labeled probes. Therefore, lanes 1 to 11 in each of the blots in panels A, B, and C represent the same samples. Lanes: 1, negative control, i.e., containing all components of the PCR assay without added target DNA; 2, control for specificity of the probe containing the amplified emm gene of an M-1 strain; 6, HindIII-digested bacteriophage lambda DNA, serving as a molecular size marker. The other lanes represent the amplified emm 49 genes of the following strains: 8314/1945 (lane 3), 49-49/123 (lane 4), B737/137/1 (lane 5), 90-397 (lane 7), 90-306 (lane 8), 88-299 (lane 9), 89-288 (lane 10), and 90-053 (lane 11). Blot A was hybridized with the emm49H probe, blot B was hybridized with the emm49K probe, and blot C was hybridized with both the emm49HK1 and the emm49HK2 probes (with identical results). Further interpretation of the hybridization results is given in the text.

hybridized sequentially to the described emm49 probes. The hybridization results (Fig. 3) were in complete agreement with the results obtained by the dot blot assays. For clarity, the findings of the hybridization experiments for detection of the emm 49 gene by different technical approaches and probes are summarized in Table 3. Finally, the N termini of emm genes from four selected M-49 strains (designated 49-49/123, 88-299, 90-053, and 8314/ 1945) were sequenced, and segments of 230 nucleotides (beginning with the translation start codon for the signal peptide) are shown in Fig. 4. The nucleotide sequence of the gene of reference strain 49-49/123 was identical to the sequence published by Haanes and Cleary (6), and in the sequence of clinical isolate 88-299, a difference of only one nucleotide from the corresponding sequence of Haanes and Cleary was found. Of note, the other two sequences (emm genes of clinical isolates 90-053 and 8314/1945) showed exactly the 12-nucleotide deletion described by Khandke et al. (11). Thus, the sequencing data were in total agreement with the results predicted by our hybridization assays.

genes were

DISCUSSION For the past five decades, streptococci have been grouped and identified serologically by methods based on the studies conducted by Rebecca Lancefield (13). Group A streptococci are commonly further classified on the basis of the antigenic specificities of their OF, T, and M antigens. These antigenic markers have been routinely relied upon for streptococcal epidemiological studies. Phage (24) and bacteriocin (25) typing have been applied to a few strains, but these methods have not found widespread use. Characterization of group A streptococcal strains on the basis of isoenzyme patterns (17) is the most recent development of typing at the level of expressed proteins. One of the initial molecular approaches for studying the epidemiology of group A streptococci made use of the visual comparison of DNA fingerprints produced by restriction enzyme digests of genomic DNA. With the exception of some variations among strains of the same serotype, DNA fingerprints reflected M antigen specificity (3). However, at present this technique is somewhat limited by technical problems in evaluating larger numbers of strains. Shortcomings are encountered with comprehensive serological M protein typing as well. Many group A streptococcal isolates are currently not M protein typeable because of either a loss of M antigen expression under cultivation or a lack of availability of appropriate M antisera. M antisera are

laborious and expensive to prepare because of the poor immunogenicity of some M proteins and the necessity for the often difficult removal of cross-reactivity by absorption with organisms of selected heterologous serotypes. These problems have restricted comprehensive M protein typing to a few reference laboratories in the world. Even the insufficient standardization of antisera, especially for rarely occurring M antigens, may occasionally lead to contradictory typing results between reference laboratories (unpublished observation). To overcome these limitations, we developed an "oligotyping" system using oligonucleotide probes corresponding to the N-terminal sequences of several emm genes. Currently available amino acid and nucleotide sequence data revealed that the signal peptides and the C termini of M protein molecules are highly conserved, whereas the N terminus of a mature protein is specific for a given M serotype, making this region a suitable candidate to serve as a target sequence for a type-specific hybridization assay. Even though the published partial nucleotide sequences of emm 2, emm 18, and emm 19 are described only as sequences of emm-like genes (19), the probes derived from these sequences worked as well as the probes derived from other, confirmed emm genes. TABLE 3. Results obtained by hybridization of DNA from eight M-49 strains with four different probes (for details, see Table 2) corresponding to the published sequence of the N-terminal region of the emm 49 gene (6)a Strain (source)

Reference strains B737/137/1 (Aachen; Lancefield strain) 49-49/123 (Prague)

Clinical isolates 88-299 (Minneapolis) 90-053 (Minneapolis) 90-397 (Minneapolis) 89-288 (Minneapolis) 90-306 (Minneapolis) 8314/1945 (Prague)

Hybridization' with the following probe: emm49H emm49K emm49HKl emm49HK2

+

-

+

+

+

-

+

+

+

-

-

+ + + + +

+ + +

+ + +

+ + +

+ + +

a Identical results were obtained with both dot blot assays and Southern blotting of the PCR-amplified emm 49 gene. b +, strongly positive hybridization signal; -, no or a very weak (depending on the hybridization stringency) hybridization signal.

VOL. 30, 1992

M PROTEIN GENE TYPING OF S. PYOGENES 49-49/123

ATGGCTAGAA AAGATACGAA TAAACAGTAT TCGCTTAGAA AATTAAAAAC

88-299 90-053 8314/1945

49-49/123 88-299

2395

50

----------

AGGTACAGCA TCCGTAGCGG TCGCTGTGGC TGTTTTAGGA GCAGGCTTTG

100

49-49/123 88-299 90-053 8314/1945

CAAACCAAAC AGAAGTTAAG GCTGCGAAA AAAAAGTTGA GGCTAAAGTT

150

49-49/123 88-299

GAGGTTGCGG AGAATAACGT GTCTAGCGTT GCAAGAAGAG AAAAAGAGCT ---------- ------G--- ----------_________

90-053

******

90-053

8314/1945

8314/1945

------

_-----_--

T----- -----* * * ---- T----- ----------*** ----

-_____---

__

49-49/123

ATACGACCAA ATCGCCGATC TTACAGATAA

88-299 90-053

--

-------

----------

8314/1945

----------

----------

-------

200

230

----

----------

FIG. 4. N-terminal nucleotide sequences of the amplified emm 49 genes from four different M-49 strains. The sequences start with the translation start codon. A dash at a given position in the sequences indicates identity with the corresponding sequence of reference strain 49-49/123. The beginning of the sequence for the mature protein is underlined. The 12-nucleotide deletion in strains 90-053 and 8314/1945 is marked with asterisks.

In comparison with the serological M protein typing results, the typing results achieved with the newly designed oligonucleotide probes exhibited 100% sensitivity and specificity for the identification of serotypes M-1, M-2, M-3, M-5, M-6, M-12, M-18, M-19, and M-24. This was true even when a potential investigator's bias was prevented by running one set of assays in a singly blinded fashion. Thus, the results confirm that the segments of diverse emm genes that encode the N terminus of a mature protein not only are significantly different from each other but also, with regard to the stringency of the hybridization reactions, are more than 93 to 96% homologous within the entity of one M serotype. This high degree of homology among strains of the above-mentioned M serotypes could be found in strains originating from three different continents (Europe, North America, and Asia) and three different decades. This finding could be important with regard to theories of how the different M serotypes have been generated. On the basis of our results, a gradual change in the type-specific region of an emm gene because of point mutations resulting in an encoded M protein not being detected by antibodies directed against the original M protein appears to be less probable. The investigation of the strains of serotype M-49 revealed an interesting finding that underscored the specificity of the probe technology. Probes emm49HK1 and emm49HK2 are directed to codons 15 to 24 and codons 25 to 34, respectively, of the sequence coding for the mature M-49 protein of strain CS101 (or B737/137/1, as the strain is designated in the laboratory of the late Rebecca Lancefield, The Rockefeller University, New York, N.Y.), published by Haanes and Cleary (6). The deduced amino acid sequences of these segments are identical to the corresponding amino acid sequences of the M-49 protein derived from strain B915 (11). These regions of the M-49 protein seem to be conserved within this M serotype, because all M-49 strains tested by us were specifically detected by the emm49HK1 and emm49HK2 probes (Table 3 and Fig. 3C). The emm49H probe, directed against codons 5 to 14 in the sequence of strain CS101, annealed to three of eight investigated M-49

strains. In comparison with that of strain B915, this sequence contains a four-amino-acid insertion comprising codons 8 to 11 and, therefore, 12 of the 30 nucleotides of probe emm49H. Another probe, designated emm49K, was specific for M-49 strains lacking this insertion sequence. Thus, probes emm49H and emm49K were designed to allow the identification of the eight strains of serotype M-49 in a mutually exclusive form (Table 3 and Fig. 3A and B), clearly indicating that the M-49 strains segregated into (at least) two distinct genetic subgroups. The specificity of the probe technology was further confirmed by sequencing of the N termini of emm genes from four selected M-49 strains (Fig. 4). These results were in complete agreement with the results obtained by other authors (6, 11), and the natural occurrence of this subgroup phenomenon among M-49 strains is supported for the first time by the investigation of several strains. It is tempting to speculate that a sequence diversity similar to that in the antigenically exposed part of the M-49 protein might be found in the corresponding parts of the M proteins of other OF-positive serotypes as well. This speculative question can be answered as soon as more sequence data for emm genes of OF-positive serotypes are available. Compared with the gold standard, serological identification of M serotypes, the probe technology clearly has some advantages that merit further investigations. For the 10 serotypes tested, the oligonucleotide probes are as sensitive as serological M protein typing. Moreover, on the basis of theoretical grounds as well as the experimental results, they are even more specific than polyclonal antibodies. With regard to the M-49 subgroup phenomenon, excellent specificity can be achieved when probes that anneal to an appropriate subsegment of the gene are selected. Thus, the discriminatory power of this approach may be useful in epidemiological studies to differentiate M-49 strains (and perhaps also strains of other M serotypes) that would otherwise be indistinguishable. The probe technology has some technical advantages as

2396

KAUFHOLD ET AL.

well. Nonradioactively labeled probes can be handled easily in the diagnostic laboratory without special permits and precautions. They can be stored easily for extended periods of time without a loss of quality. Probes are reliably and more cost effectively synthesized in a shorter time than is required for the production and absorption of antisera. It is of pivotal importance that probes can be prepared without problems of standardization in different laboratories. On the other hand, probes could be applied to the standardization of M antigen typing antisera. Also, with respect to hands-on time for nucleic acid preparation and hybridization (18), the described oligotyping method already compares favorably with the conventional serological technique. Further technical improvements are currently under investigation in our laboratories. During the preparation of this manuscript, an alternative method for the isolation of genomic DNA (2) was found to be equally suitable and has reduced the time for preparing DNA from 12 isolates from approximately 6 h to 2 h. The panel of probes for typing of an individual group A streptococcal strain can be reduced by prescreening this strain either by conventional OF detection techniques (10) or by the use of probes that are able to discriminate OF-positive and OF-negative strains (19). Thereafter, testing with only a specific set of OF-positive or OF-negative subpanels of type-specific probes is required. Finally, the introduction of techniques such as the "reverse dot blot" method (22) may significantly shorten the time required for the oligotyping procedure. In this method, a membrane that contains an array of immobilized typing probes is hybridized once with a PCR-amplified emm gene (already labeled by the incorporation of biotinylated primers during the amplification cycles), thus avoiding time-consuming sequential hybridization procedures. In conclusion, the available panel of evaluated probes will allow the reliable identification of many M serotypes that are presently known to be of major clinical significance for group A streptococcal diseases. With sequence data for more emm genes (shortly before this study was completed, the nucleotide sequence of the M-57 protein gene was published [14]), comprehensive M protein gene typing by DNA probes may be possible. When nucleotide sequences of emm genes of group A streptococcal strains that are not typeable by serological means become available, the genetic method may provide a useful approach for the specific identification of these strains as well. We suggest that serological M protein typing and oligotyping should not be regarded as competing methods but rather as different methods that may lead to distinct but complementary classification schemes. The oligotyping method has been shown to enhance serotype resolution, and it may increase group A streptococcal typeability in the future. Such improvements could dramatically benefit epidemiological studies essential for the development of improved public health control methods for this important human pathogen. ACKNOWLEDGMENTS We acknowledge the excellent technical assistance of Jutta Palmen. We are grateful to Jim Dale and Robert W. Baird for providing a portion of the unpublished nucleotide sequence of the emm 3 gene as well as to Paula Kuzemenska, Renata Vesela, and Jiri Havlicek for supplying streptococcal strains used in this study.

REFERENCES 1. Bisno, A. 1991. Group A streptococcal infections and acute rheumatic fever. N. Engl. J. Med. 325:783-793.

J. CLIN. MICROBIOL. 2. Bollet, C., M. J. Gevaudan, X. de Laballerie, C. Zandotti, and P. de Micco. 1991. A simple method for the isolation of chromosomal DNA from gram positive or acid-fast bacteria. Nucleic Acids Res. 8:1955. 3. Cleary, P. P., E. L. Kaplan, C. Livdahl, and S. Skjold. 1988. DNA fingerprints of Streptococcus pyogenes are M type specific. J. Infect. Dis. 158:1317-1323. 4. Fischetti, V. A. 1989. Streptococcal M protein: molecular design and biological behavior. Clin. Microbiol. Rev. 2:285-314. 5. Gooder, H. 1961. Association of a serum opacity reaction with serological type in Streptococcus pyogenes. J. Cien. Microbiol. 25:347-352. 6. Haanes, E. J., and P. P. Cleary. 1989. Identification of a divergent M protein gene and an M protein-related gene family in Streptococcus pyogenes serotype 49. J. Bacteriol. 171:63976408. 7. Haanes-Fritz, E., W. Kraus, V. Burdett, J. B. Dale, E. H. Beachey, and P. Cleary. 1988. Comparison of the leader sequences of four group A streptococcal M protein genes. Nucleic Acids Res. 16:4667-4677. 8. Hollingshead, S. K., V. A. Fischetti, and J. R. Scott. 1986. Complete nucleotide sequence of type 6 M protein of the group A streptococcus. J. Biol. Chem. 261:1677-1686. 9. Huang, T. T., H. Malke, and J. J. Ferretti. 1989. Heterogeneity of the streptokinase gene in group A streptococci. Infect. Immun. 57:502-506. 10. Johnson, D. R., and L. Kaplan. 1988. Microtechnique for serum opacity factor characterization of group A streptococci adaptable to the use of human sera. J. Clin. Microbiol. 2:2025-2030. 11. Khandke, K. M., T. Fairwell, A. S. Acharya, B. L. Trus, and B. N. Manjula. 1988. Complete amino acid sequence of streptococcal pepM49 protein, a nephritis-associated serotype: conserved conformational design' among sequentially distinct M protein serotypes. J. Biol. Chem. 263:5075-5082. 12. Lancefield, R. C. 1928. The antigenic complex of Streptococcus hemolyticus. I. Demonstration of a type-specific substance in extracts of Streptococcus hemolyticus. J. Exp. Med. 47:91-103. 13. Lancefield, R. C. 1933. A serological differentiation of human and other groups of hemolytic streptococci. J. Exp. Med. 57:571-595. 14. Manjula, B. N., K. M. Khandke, T. Fairwell, W. A. Relf, and K. S. Sriprakash. 1991. Heptad motifs within the distal subdomain of the coiled-coil rod region of M protein from rheumatic fever and nephritis associated serotypes of group A streptococci are distinct from each other: nucleotide sequence of the M57 gene and relation of the deduced amino acid sequence to other M proteins. J. Protein Chem. 10:369-384. 15. Miller, L., L. Gray, E. Beachey, and M. Kehoe. 1988. Antigenic variation among group A streptococcal M proteins. J. Biol. Chem. 263:5668-5673. 16. Mouw, A. R., E. H. Beachey, and V. Burdett. 1988. Molecular evolution of streptococcal M protein: cloning and nucleotide sequence of the type 24 M protein gene and relation to other genes of Streptococcus pyogenes. J. Bacteriol. 170:676-684. 17. Musser, J. M., A. R. Hauser, M. H. Kim, P. M. Schlievert, K. Nelson, and R. K. Selander. 1991. Streptococcus pyogenes causing toxic-shock-like syndrome and other invasive diseases: clonal diversity and pyrogenic exotoxin expression. Proc. Natl. Acad. Sci. USA 88:2668-2672. 18. Podbielski, A., 0. Kuhnemund, and R. Luitticken. 1990. Identification of group A type 1 streptococcal M protein gene by a non-radioactive oligonucleotide detection method. Med. Microbiol. Immunol. 179:255-262. 19. Podbielski, A., B. Meizer, and R. Lutticken. 1991. Application of the polymerase chain reaction to study the M protein(-like) gene family in beta-hemolytic streptococci. Med. Microbiol. Immunol. 180:213-227. 20. Robbins, J. C., J. G. Spanier, S. J. Jones, W. J. Simpson, and P. P. Cleary. 1987. Streptococcus pyogenes type 12 M protein gene regulation by upstream sequences. J. Bacteriol. 169:56335640. 21. Rotta, J., R. M. Krause, R. C. Lancefield, W. Everly, and H. Lackland. 1971. New approaches for the laboratory recognition

VOL. 30, 1992 of M types of group A streptococci. J. Exp. Med. 134:1298-1315. 22. Saiki, R. K., P. S. Walsh, C. H. Levenson, and H. A. Erlich. 1989. Genetic analysis of amplified DNA with immobilized sequence-specific oligonucleotide probes. Proc. Natl. Acad. Sci. USA 86:6230-6234. 23. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 24. Skjold, S. A., and L. W. Wannamaker. 1976. Method for phage typing group A type 49 streptococci. J. Clin. Microbiol. 4:232238. 25. Tagg, J. R., and L. V. Bannister. 1979. "Fingerprinting" betahaemolytic streptococci by their production of and sensitivity to bacteriocine-like inhibitors. J. Med. Microbiol. 12:397-411.

M PROTEIN GENE TYPING OF S. PYOGENES

2397

26. Top, F. H., and L. W. Wannamaker. 1968. The serum opacity factor of Streptococcus pyogenes: frequency of production of streptococcal lipoproteinase by strains of different serological types and the relationship to M protein production. J. Hyg. 66:49-58. 27. Top, F. H., Jr., and L. W. Wannamaker. 1968. The serum opacity reaction of Streptococcus pyogenes: the demonstration of multiple, strain-specific lipoproteinase antigens. J. Exp. Med. 127:1013-1034. 28. Widdowson, J. P., W. R. Maxted, and D. L. Grant. 1970. The production of opacity in serum by group A streptococci and its relationship with the presence of M antigen. J. Gen. Microbiol. 61:343-353.