Molecular population genetic analysis of aStreptococcus pyogenesbacteriophage-encoded hyaluronidase gene: recombination contributes to allelic variation

Microbial Pathogenesis 1997; 22: 209–217

MICROBIAL PATHOGENESIS

Molecular population genetic analysis of a Streptococcus pyogenes bacteriophageencoded hyaluronidase gene: recombination contributes to allelic variation Ann Marie Marciela b, Vivek Kapura∗ & James M. Mussera b† a

Section of Molecular Pathobiology, Department of Pathology, and bDepartment of Microbiology and Immunology, Baylor College of Medicine, Houston, TX 77030, U.S.A. (Received September 11, 1996; accepted in revised form October 29, 1996)

Many strains of the human pathogenic bacterium Streptococcus pyogenes produce hyaluronidase, an enzyme that degrades hyaluronic acid, a major component of the extracellular matrix. Degradation of hyaluronic acid is thought to aid in host tissue invasion and dissemination of S. pyogenes. The molecular population genetics of the bacteriophage-encoded hyaluronidase gene (hyl) was analysed by sequencing the gene from 13 streptococcal strains representing seven well-differentiated multilocus enzyme electrophoretic types and eight M or T protein serotypes. Substantial levels of allelic polymorphism were identified, and the analysis found strong statistical evidence that recombinational processes have contributed to the generation of molecular variation in this gene. A 111 base pair segment of hyl encoding a collagenous motif, that may bind collagen, was absent in a serotype M14 isolate and 13 serotype M18 multilocus enzyme electrophoretic type 20 strains examined. The analysis provides a molecular population genetics framework for studies examining the role of naturally occurring hyaluronidase variation in host–pathogen interactions.  1997 Academic Press Limited

Key words: Streptococcus pyogenes, group A Streptococcus, hyaluronidase, recombination, genetics, bacteriophage.

Introduction Streptococcus pyogenes (Group A Streptococcus) causes many human diseases including pharyngitis, glomerulonephritis, acute rheumatic fever, ∗ Present affiliation: Department of Veterinary Pathobiology, University of Minnesota, St. Paul, MN 55108, U.S.A. † Author for correspondence. 0882–4010/97/040209+09 $25.00/0/mi960100

and severe invasive infections such as toxicshock-like-syndrome and necrotizing fasciitis. A recent increase in the frequency of occurrence of severe invasive diseases caused by this pathogen has led to renewed interest in several putative virulence factors. Hyaluronidase is a degradative enzyme thought to be produced by most group A streptococci. The enzyme potentially assists bacterial spread through host tissue by its ability to hydrolyse glucosaminic  1997 Academic Press Limited

210

bonds in hyaluronic acid [1], a major component of the extracellular matrix. S. pyogenes produces two distinct hyaluronidase enzymes. One is of bacterial origin, the other is bacteriophage encoded [2]. The bacterial hyaluronidase is serologically identical among isolates assigned to distinct Lancefield streptococcal serogroups [3]. In contrast, phageencoded hyaluronidase is antigenically similar among isolates expressing the same M serotype, but distinct among organisms assigned to different M types [2, 4]. Hynes and Ferretti cloned and sequenced a bacteriophage-encoded hyaluronidase gene (hylP) from a serotype M49 S. pyogenes strain [5]. More recently, Hynes et al. [6] characterized a second bacteriophage-encoded hyaluronidase gene (hylP2) from a serotype T22 strain, and showed it to be 66.5% identical to hylP. The size of hylP2 differed from hylP by 102 bp encoding 34 amino acids. PCR analysis of 115 strains found that 30% and 54% had a PCR product consistent with hylP and hylP2, respectively. PCR products corresponding to both genes were found in 14% of strains. Inasmuch as there is accumulating evidence that horizontal transfer and recombinational processes are contributing to the generation of allelic variation in group A streptococci virulence genes [7–12], we tested the hypothesis that polymorphism in these bacteriophageborne genes may be due, in part, to intragenic recombinational events. Our data demonstrate the existence of mosaic alleles and thereby show that recombination has played a significant role in the diversification of streptococcal bacteriophage-encoded hyaluronidase.

Results DNA sequence analysis of hyl from 13 S. pyogenes strains The entire hyl gene sequence was obtained from 13 strains, and nine distinct alleles were identified (Fig. 1; Table 1). The open reading frame for nine of the sequences ranged from 1110 to 1116 bp, but three ET20/M18 organisms, and one ET88/M14 strain, lacked the same 111 nucleotides (bp 167 to 277) encoding 37 amino acids. Compared to the published hylP sequence [5] the number of polymorphic nucleotide sites ranged from 20 sites (1.8%) in two ET14 isolates to 225 sites (20.2%) in an ET88/M14 strain. Not

A. M. Marciel et al.

including the 111 bp region, the polymorphic nucleotide sites would result in a total of 89 amino acid substitutions. In general, strains classified as the same clone by multilocus enzyme electrophoresis lacked hyl nucleotide diversity. For example, the three ET20/M18 strains had the identical hyl sequence, and similarly, an ET14/M4 and one ET14/T4 isolates had the same hyl allele. ET2/M3 strains were unusual in that one of the three strains sequenced in entirety had a single nucleotide change (C→A) located at position 636. This mutation would result in a S212R substitution in the inferred amino acid sequence.

Lack of a collagenous motif in all 13 ET20/M18 strains and 1 ET88/M14 strain Stern and Stern [13] noted that the hyaluronidase gene from S. pyogenes encoded a region of 30 amino acids with 10 repeats of the Gly-X-Y triplet characteristic of the collagen structure. Interestingly, the hyl gene in the three ET20/ M18 strains sequenced in entirety, and the ET88/ M14 organism lacked 111 bp which would encode amino acids spanning a putative collagenlike domain. Ten additional ET20/M18 organisms were then screened for this polymorphism by examining size variation in hyl PCR products using primers flanking the gene. The 10 strains had smaller amplified fragments than those obtained from all other non-ET20/ M18 organisms studied, except an ET88/M14 isolate, a result suggesting absence of the region in these bacteria. To localize the size variation to the region of hyl encoding the collagenous motif, the amplification procedure was repeated using internal primers that flank the 111 bp region. The resulting PCR products were also smaller than those obtained from all other organisms, except the ET88/M14 isolate (data not shown).

hyl variation among ET2/M3 isolates Strain MGAS 274 (ET2/M3) had a single base pair change compared to the other two ET2/ M3 organisms sequenced. To determine if other ET2/M3 organisms had this polymorphism, 36 ET2/M3 strains recovered from diverse localities, diseases, and decades were examined with a restriction fragment polymorphism strategy. The PCR products from all 36 ET2/M3

Recombination in S. pyogenes hyaluronidase gene

211

212

A. M. Marciel et al.

Figure 1. Alignment of the hyl alleles identified in 13 S. pyogenes isolates. A previously published [5] sequence of hylP is included for comparison. Dots represent identity with hylP. Dashed lines indicate the absence of nucleotides. Sequences were aligned by the clustal method with the MegAlign program of DNASTAR.

strains were digested once with MspA1 I, indicating a cytosine at nucleotide site 636. Strain MGAS 274 regrown from a freezer stock and again subjected to hyl PCR amplification and RFLP analysis yielded identical results to those obtained with the original culture of strain MGAS 274.

Evidence for intragenic recombination Visual inspection of a schematic representation of hyl polymorphism (Fig. 2) identified structural

mosaicism, suggesting that intragenic recombination had contributed to the generation of nucleotide variation identified among these strains. For example, apparent recombination is readily illustrated by the hyl sequence from MGAS 1719 and the three ET20/M18 strains (MGAS 156, MGAS 300, MGAS 1622). Sawyer’s sum of the squares of the condensed fragment lengths (SSCF) analysis [14] with synonymous silent sites was then used to test for statistical evidence of recombination. Results from this analysis method strongly supported

Recombination in S. pyogenes hyaluronidase gene

Table 1. Properties of 13 S. pyogenes isolates. a

ET

MGAS no.b

Serotype

1 1 2 2 2 14 14 20 20 20

249 326 259 274 1835 321 318 156 300 1622

M1 M1 M3 M3 M3 M4 T4 M18 M18 M18

43 66 88

1719 1838 1352

T8 M27 M14

c

Disease

TSLS SID SID Invasive TSLS Invasive TSLS Invasive Rheumatic fever Invasive Invasive Scarlet fever

No. of polymorphic sites 77 77 72 73 72 20 20 109d 109d 109d 69 59 114d

a

ET, Electrophoretic type. Musser group A Streptococcus collection number. c SID, severe invasive disease; TSLS, toxic-shock-like-syndrome. d Excludes 111 nucleotides (bp 167–bp 277) comprising a collagenous motif. b

the contention that recombinational processes have contributed to hyl allelic variation (P= <0.0001).

213

Sliding window analysis A sliding window analysis of the number of synonymous changes per synonymous site (PS) and coding changes per non-synonymous site (PN), with a window of 30 codons, identified several regions with substantial variance in the ratio of PS to PN (Fig. 3). Interestingly, the region containing codons 58 to 83 encoding much of the collagenous motif is characterized by a virtual absence of either non-synonymous or synonymous substitutions.

Discussion This analysis identified substantial allelic variation and mosaic organization in the hyaluronidase gene among distinct clones of group A Streptococcus. Nucleotide variation was dispersed throughout the gene, and regions of apparent recombination were found by visual inspection and statistical analysis. Recombination in genes encoding virulence factors has been identified in many bacteria including Salmonella enterica and Escherichia coli [15–17], Neisseria meningitidis [18], and Haemophilus influenzae [19].

Figure 2. Schematic representation of hyl polymorphism. The schematic illustrates the mosaic organization of the hyl alleles that is suggestive of recombination. The entire gene is represented and the vertical lines indicate the position of nucleotide variation compared to a consensus sequence. The black bar region in strain MGAS 1622 (ET20/M18) and GAS 1352 (ET88/M14) illustrates the location of a 111 bp sequence encoding a collagenous motif found in other strains, but not in the ET20/M18 and ET88/M14 organisms. For simplicity, data from only one strain of each clone is shown.

214

A. M. Marciel et al.

Proportion of site differences

0.6 0.5 0.4 0.3 0.2 0.1

0

30

60

90

120

150

180 210 Codon

240

270

300

330

360

Figure 3. Sliding window analysis of hyl alleles. The analysis was done using a window size of 30 to compare numbers of synonymous changes per synonymous site (PS; —Φ—) and coding changes per nonsynonymous site (PN; —•—). A plot of the proportion of site differences vs codon position shows considerable variance in the ratio of synonymous to non-synonymous changes distributed across hyl. The collagenous motif (codons 58–83) is unusual because it is characterized by restricted nucleotide and amino acid variation.

The considerable nucleotide variation in bacteriophage-encoded hyl is in agreement with the observations reported by Benchetrit, Wannamaker and Gray [4] who found that immunological heterogeneity exists among phageassociated hyaluronidase made by different M types. Segments of variability interspersed among homologous regions may account for the hyaluronidase serologic variability observed among heterologous M type strains. It is noteworthy that a region thought to encode a collagenous motif was absent in the hyl gene in an ET88/M14 and all 13 ET20/M18 organisms. Because single stranded collagen can bind to the triple helices of collagen [20], it is possible that the collagenous motif in hyaluronidase assists attachment of the enzyme to host connective tissue. This process would effectively concentrate hyaluronidase activity at a local site. Although the exact importance of this region has not been demonstrated, its distribution among streptococci may have functional consequences. Sliding window analysis of the hyl gene identified substantial variance in the PS/PN ratio across the gene. For most regions, PS was greater than PN, as expected in the absence of significant selective pressure. It is noteworthy that one prominent exception is the region (approximately codons 58 to 83) encoding most of the collagenous motif. This region is characterized by a virtual absence of either synonymous or non-synonymous substitutions. The

lack of nucleotide sequence variation in this region suggests relatively recent “capture” of this segment by the hyl gene. Under this hypothesis, this region has not yet had sufficient time to accumulate the levels of expected variation characteristic of other hyl segments. Consistent with this idea of recent capture, Hynes et al. [6] amplified DNA fragments corresponding to the size of hylP with the collagenous motif in only 30% of isolates studied. Recently the sequence for hylP2, a bacteriophage hyaluronidase gene from a S. pyogenes serotype T22 strain was published [6]. It was reported that 102 nucleotides (encoding 34 amino acids) present in hylP were absent from hylP2. As shown in Figs 1 and 2, the number of amino acids missing in the region of the collagenous motif differ in length between hylP2 and all four of our strains analysed which also lack the collagenous motif. This observation suggests either that additional sequence variation has occurred since the divergence of the two genes from a common ancestor, or that two distinct lineages have independently converged by a mechanism involving loss of a region of approximately 100 bp. Although we currently lack sufficient information to distinguish between the two possibilities, we believe convergence-by-loss is less likely. The importance of hyaluronidase as a virulence factor has been demonstrated in higher organisms, such as the Ancylostoma helminth larvae [21], and insects and snakes [22], where

Recombination in S. pyogenes hyaluronidase gene

the enzyme is used to degrade hyaluronic acid in the connective tissue of the host and aid in the spread of itself or other substances. The important role of hyaluronidase has also been shown in Streptococcus pneumoniae where the level of hyaluronidase activity is apparently one of the more important factors mediating bacterial penetration of the blood–brain barrier during meningitis episodes [23]. Quinn and Liao showed that 76% of patients infected with Group A Streptococcus and 100% of patients with active rheumatic fever had an anti-hyaluronidase antibody titer above normal [24]. By analogy with the findings from S. pneumoniae [23], the level of hyaluronidase expression by group A streptococci could be an important variable in mediating disease type and severity. However, more investigation will be required to test this hypothesis. The hyl alleles characterized in this study show strong evidence for the involvement of genetic diversification mechanisms other than simple random point mutations. The analysis showed that clonally unrelated organisms, as indexed by multilocus enzyme electrophoresis, had regions of high sequence diversity interspersed with regions of restricted nucleotide variation. This apparent mosaic organization of hyl adds to substantial accumulating evidence [12] that horizontal transfer and intragenic recombination is a critical mechanism employed by S. pyogenes to enhance the level of genetic diversity in nature. Gene transfer mediated by bacteriophages is known to be an important factor in generating bacterial variation [25]. The hyl in our study is bacteriophage-encoded, and therefore it is reasonable to hypothesize that transduction has been involved in mediating hyl diversification. Generation of new alleles by recombination is an effective mechanism to create proteins with considerable variation in structure and/or function. The recurring theme that horizontal gene transfer is an important mechanism in generating genome diversity and allelic variants of virulence genes [12] suggests that the molecular mechanisms of recombination in S. Pyogenes warrant additional study.

Materials and methods Bacterial strains The entire hyl gene was sequenced from 13 S. pyogenes strains recovered from patients with

215

diverse clinical syndromes and widespread localities (Table 1). These bacteria represent seven well-differentiated multilocus enzyme genotypes and eight M or T protein serotypes. Bacteria were grown overnight on brain–heart infusion agar at 37°C in 5% CO2. Chromosomal DNA was extracted and used for gene amplification and sequence analysis. Thirty-six additional ET2/M3 organisms were studied. These bacteria include nine strains recovered from scarlet fever patients in East Germany between 1969 and 1990, seven strains isolated from acute rheumatic fever patients, and 20 strains from patients with severe invasive disease or toxic-shock-like-syndrome in 10 U.S.A. states. Ten additional ET20/M18 organisms from the United States were also used, including nine strains isolated from acute rheumatic fever patients, and one organism recovered from a patient with severe invasive disease.

Amplification and DNA sequence analysis of hyl Oligonucleotide primers based on a published hyl sequence [5] were used for PCR amplification of chromosomal DNA prepared from the 13 strains. The primers used were: primer 1 (forward), GGTTGACGGTAATAATGCCAC; primer 2 (reverse), CGACCATCTGCAAAAGATAAGTC; primer 3 (forward), CCATTAAGAGTCCAATTTAAGC; primer 4 (reverse), CTGATTCTGAGCATTTTACC; primer 5 (forward), AGAAACTAATAGTAAAATCAC. Primers 1 and 2 are located outside of the hyaluronidase gene at the 5′ end and 3′ end, respectively. Primers 3, 4, and 5 are located internal to primers 1 and 2. PCR amplification was done with Taq DNA polymerase, deoxynucleotides, and 10× buffer (Perkin Elmer Cetus, Norwalk, CT). Amplification conditions were the same for all primer pairs used: 30 cycles of denaturing at 94°C for 1 min, annealing at 55°C for 2 min, and 2.5 min of extension at 72°C. Automated DNA sequencing was performed with an Applied Biosystems model 373A instrument. Taq DyeDeoxy terminator cycle sequencing kits (Applied Biosystems, Inc., Foster City, CA) were used for sequencing reactions. PCR conditions for sequencing were 25 cycles of 30 s of denaturation at 96°C, annealing for 15 s at 50°C, and extension for 4 min at 60°C.

216

Analysis of DNA sequence data DNA sequences were analysed electronically with DNASTAR (Madison, WI) computer programs, including MegAlign, EditSeq, MapDraw, SeqMan and Align. Alignment of all sequences was obtained with the clustal method and the consensus generated by including all sequences. A statistical test for intragenic recombination [14] was performed with the VTDIST computer program written by S. Sawyer, Department of Mathematics, Washington University. Synonymous and non-synonymous values for sliding window analysis were generated with the program NAGV2 (written by T. Ota, Institute of Molecular Evolutionary Genetics, Pennsylvania State University, PN).

Restriction fragment length polymorphism analysis of hyl from ET2/M3 strains Restriction fragment length polymorphism (RFLP) analysis was used to identify allelic variation located at nucleotide position 636 in the hyl gene of ET2/M3 strains. Restriction enzyme MspA1 I recognizes the sequence CMG′CKG that occurs once in hyl. A nucleotide change C→A at position 636 results in loss of the MspA1 I cleavage site. PCR products of the hyl gene from 36 ET2/M3 strains were obtained using primers 1 and 2 and then screened using this method.

PCR analysis of hyl in ET20/M18 strains Absence of a 111 bp region in hyl in 10 ET20/ M18 strains was assessed by PCR analysis. Chromosomal DNA from these strains was amplified by PCR with primers 1 and 2 which flank hyl. The size of the resulting PCR products was compared by gel electrophoresis to amplified fragments derived from strains sequenced for hyl, including organisms known to contain or lack the 111 bp region of interest (nucleotides 167 to 277). To localize the missing region, internal primers 3 and 4 were also used in PCR reactions, and the size of the resulting amplified gene fragments was compared with products from control strains.

Acknowledgments We thank L.-L. Li and X. Pan for technical assistance. This study was supported by Public

A. M. Marciel et al.

Health Service Grant AI-33119. J.M.M. is an Established Investigator of the American Heart Association.

References 1 Linker A, Meyer K, Hoffman P. The production of unsaturated uronides by bacterial hyaluronidases. J Biol Chem 1956; 219: 13–25. 2 Kjems E. Studies on streptococcal bacteriophages. 5. Serological investigation of phages isolated from 91 strains of group A haemolytic streptococci. Acta Pathol Microbiol Scand 1960; 49: 205–12. 3 Wenner HA, Gibson DM, Jaques R. Specificities of hyaluronidases formed by several groups of streptococci. Proc Soc Exp Biol Med 1951; 76:585–7. 4 Benchetrit LC, Wannamaker LW, Gray ED. Immunological properties of hyaluronidases associated with temperate bacteriophages of group A streptococci. J Exp Med 1979; 149: 73–83. 5 Hynes WL, Ferretti JJ. Sequence analysis and expression in Escherichia coli of the hyaluronidase gene of Streptococcus pyogenes bacteriophage H4489A. Infect Immun 1989; 57: 533–9. 6 Hynes WL, Hancock L, Ferretti JJ. Analysis of a second bacteriophage hyaluronidase gene from Streptococcus pyogenes: evidence for a third hyaluronidase involved in extracellular enzymatic activity. Infect Immun 1995; 63: 3015–20. 7 Kapur V, Kanjilal S, Hamrick MR, Li L-L, Whittam TS, Sawyer SA, Musser JM. Molecular population genetic analysis of the streptokinase gene of Streptococcus pyogenes: mosaic alleles generated by recombination. Mol Microbiol 1995; 16: 509–19. 8 Kapur V, Nelson K, Schlievert PM, Selander RK, Musser JM. Molecular population genetic evidence of horizontal spread of two alleles of the pyrogenic exotoxin C gene (speC) among pathogenic clones of Streptococcus pyogenes. Infect Immun 1992; 60: 3513–7. 9 Nelson K, Schlievert PM, Selander RK, Musser JM. Characterization and clonal distribution of four alleles of the speA gene encoding pyrogenic exotoxin A (scarlet fever toxin) in Streptococcus pyogenes. J Exp Med 1991; 174: 1721–4. 10 Whatmore AM, Kapur V, Sullivan DJ, Musser JM, Kehoe MA. Non-congruent relationships between variation in emm gene sequences and the population genetic structure of group A streptococci. Mol Microbiol 1994; 14:619–31. 11 Whatmore AM, Kehoe MA. Horizontal gene transfer in the evolution of group A streptococcal emm-like genes: gene mosaics and variation in Vir regulons. Mol Microbiol 1994; 11: 363–74. 12 Kehoe MA, Kapur V, Whatmore A, Musser JM. Horizontal gene transfer among group A streptococci: implications for pathogenesis and epidemiology. Trends Microbiol 1996; 4: 436–43. 13 Stern M, Stern R. A collagenous sequence in a prokaryotic hyaluronidase [letter]. Mol Biol Evol 1992; 9: 1179–80. 14 Sawyer S. Statistical tests for detecting gene conversion. Mol Biol Evol 1989; 6: 526–38.

Recombination in S. pyogenes hyaluronidase gene 15 Li J, Nelson K, McWhorter AC, Whittam TS, Selander RK. Recombinational basis of serovar diversity in Salmonella enterica. Proc Natl Acad Sci USA 1994; 91: 2552–6. 16 Reeves PR. Evolution of Salmonella O antigen variation by interspecific gene transfer on a large scale. Trends Genet 1993; 9: 17–22. 17 Liu D, Reeves PR. Presence of different O antigen forms in three isolates of one clone of Escherichia coli. Genetics 1994; 138: 7–10. 18 Feavers IM, Heath AB, Bygraves JA, Maiden MCJ. Role of horizontal gene exchange in the antigenic variation of the class I outer membrane protein of Neisseria meningitidis. Mol Microbiol 1992; 6: 489–95. 19 Kroll JS, Moxon ER. Capsulation in distantly related strains of Haemophilus influenzae type b: genetic drift and gene transfer at the capsulation locus. J Bacteriol 1990; 172: 1374–9. 20 Ramachandran GN, Ramakrishnan C. Molecular structure. In: Ramachandran GN, Reddi AH, eds. Biochemistry of Collagen. New York: Plenum, 1976:45–84. 21 Hotez PJ, Narasimhan S, Haggerty J, Milstone L,

217

22 23

24

25

Bhopale V, Schad GA, Richards FF. Hyaluronidase from infective Ancylostoma hookworm larvae and its possible function as a virulence factor in tissue invasion and in cutaneous larvae migrans. Infect Immun 1992; 60: 1018–23. Duran-Reynals F. Tissue permeability and the spreading factors in infection: a contribution to the host-parasite problem. Bact Rev 1942; 6: 197–252. Kostyukova NN, Volkova MO, Ivanova VV, Kvetnaya AS. A study of pathogenic factors of Streptococcus pneumoniae strains causing meningitis. FEMS Immunol Med Microbiol 1995; 10: 133–8. Quinn RW, Liao SJ. A comparative study of antihyaluronidase, antistreptolysin “O,” antistreptokinase, and streptococcal agglutination titers in patients with rheumatic fever, acute hemolytic streptococcal infections, rheumatoid arthritis and non-rheumatoid forms of arthritis. J Clin Invest 1950; 29: 1156–66. Cheetham BF, Katz ME. A role for bacteriophages in the evolution and transfer of bacterial virulence determinants. Mol Microbiol 1995; 18: 201–8.