Chromosomal analysis of group B streptococcal clinical strains; bac gene-positive strains are genetically homogenous

Transcription

1 FEMS Microbiology Letters 208 (2001) 93^98 Chromosomal analysis of group B streptococcal clinical strains; bac gene-positive strains are genetically homogenous Abstract b A. Dmitriev a, Y.Y. Hu b, A.D. Shen b, A. Suvorov a, Y.H. Yang b; a Institute of Experimental Medicine, Academy of Medical Sciences, acad. Pavlov str. 12, Saint Petersburg , Russia Beijing Children's Hospital A liated to Capital University of Medical Sciences, Nan Lishi Road, Beijing , PR China Received 1 April 2001; received in revised form 7 December 2001; accepted 14 December 2001 First published online 1 February 2002 A collection of 45 epidemiologically unrelated Streptococcus agalactiae strains (group B Streptococcus, GBS), belonging to different serotypes, isolated from pregnant women in China and Russia was studied. Strains were characterized by pulsed-field gel electrophoresis (PFGE) employing hybridization with nine genes potentially involved in virulence. Molecular sizes of GBS genomes varied from 2030 to 2290 kb. Location of the genes under study bac, bca, glna, scpb, cyl, hylb, lmb, scaa and cfb on the GBS genomes was found to be conserved irrelevant to the serotype. Potential virulence genes scpb, hylb, lmb were located on a 91-kb SmaI fragment that is equal to 4.5% of total genome. Ribotyping of the strains under study revealed three different HindIII, nine EcoRI and 12 PvuII ribotypes among 45 strains. A strong correlation between the PvuII ribotype and the presence of the bac gene was observed, with 21 of 22 bac-positive strains belonging to the same PvuII ribotype P1. PFGE patterns of bac-positive strains were also similar. The possibility of close genetic relatedness of all bac-positive strains is discussed. ß 2001 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords: Gene bac; Ribotyping; Streptococcus agalactiae genome 1. Introduction Streptococcus agalactiae (group B streptococcus, GBS) is an important human pathogen that causes signi cant pathology during pregnancy and in neonates [1]. In the United States, for example, a rate of 1.8 cases of S. agalactiae sepsis per 1000 live birth has been reported [2]. Multiple factors, such as the polysaccharide capsule and a number of surface and extracellular proteins, such as BAC (L-antigen), BCA (K-antigen), SCPB (C5a peptidase), LMB (laminin binding protein), are known to contribute to the virulence of GBS [3,4]. At present, numerous molecular epidemiological and immunological studies of GBS have employed restriction fragment-length polymorphism (RFLP) analysis of chromosomal DNA as well as serotyping, ribotyping and multilocus enzyme genotype analysis [5^11]. However, none of these investigations has revealed any signi cant correlation between the speci c genotype of a strain and its serotype, and most of the molecular epidemiology studies on GBS have been limited in tools of investigation used for analysis. Recently, employing hybridization analysis of several gene probes to GBS chromosomes, we demonstrated that GBS strains carrying bac gene encoding IgA binding protein di er from GBS strains without this gene [12,13]. To further characterize the genetic organization of clinical GBS strains of di erent serotypes, we analyzed 45 GBS strains by pulsed- eld gel electrophoresis (PFGE), DNA^DNA hybridization with gene probes and ribotyping to look at the relationship between PFGE and ribotype pro les and the presence of di erent genes essential for virulence. 2. Materials and methods 2.1. Bacterial strains * Corresponding author. Tel.: +86 (10) ; Fax: +86 (10) address: yangyonghong@btamail.net.cn (Y.H. Yang). A total of 45 randomly chosen, unrelated GBS strains were analyzed including 31 strains isolated from pregnant / 01 / $22.00 ß 2001 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S (02)

2 94 A. Dmitriev et al. / FEMS Microbiology Letters 208 (2001) 93^98 women in Beijing (PR China) in 1996^1998 and 14 strains isolated from pregnant women in Saint Petersburg (Russia) in 1989^1995. These strains were identi ed with group B-speci c antiserum and were found to be in the following serotypes; Ia (13 strains), Ib (5 strains), II (15 strains), III (10 strains), V (1 strain). One strain was nontypeable. Bacteria were grown either in Todd^Hewitt broth (Difco, Detroit, MI, USA) or on 1.5% horse blood agar at 37³C overnight DNA techniques Most of the molecular genetic procedures were carried out according to previously described protocols [14]. Plugs containing genomic DNA for PFGE were prepared as described [7]. Plugs were subjected to electrophoresis in 1% agarose gel in TBE bu er with a contour-clamped homogenous electric eld (CHEF DRIII; Bio-Rad Laboratories, Hercules, CA, USA) over 20 h. Pulse times varied from 1 to 120 s at 200 V depending on the sizes of DNA fragments analyzed. The pulse times equal to 5, 30, 40 and 50 s were optimum for DNA fragments separation in the regions 55^65 kb, 305^365 kb, 440^495 kb and 500^540 kb, respectively. DNA labeling with digoxigenin, hybridization analysis and immunological detection were accomplished with the Enzo1 DNA Labeling and Detection kit (Roche, Germany). Ribotyping of the strains was done using rrs gene as a 16S rrna probe (Roche, Germany). DNA fragments of the di erent GBS genes were ampli- ed by polymerase chain reaction (PCR) at annealing temperature 50³C. The following primers were used for ampli cation of the genes: 5P-GACCTTAGCCCCAAC-AG- ACTTAGC-3P and 5P-CCTCTGATAATTGCATTCCA- CGTG-3P (glna gene), 5P-ATGGGATTTGGGATAAC- TAAGCTAG-3P and 5P-AGCGTGTATTCCAGATTT- CCTTAT-3P (cfb), 5P-CAGGTACTTGGAACAGCATT- AATCG-3P and 5P-TGAGACATGGTGGAAAGCA- GAA-3P (hylb), 5P-TTATCATCCAGCGCCTCCTAG-3P and 5P-GTGGTGATAACTGACTTCTTGGGA-3P (lmb), 5P-ACGGCTTGTCCATAGTAGTGTTTG-3P and 5P- AACGACACTGCCATCAGCAC-3P (cyl), 5P-ACAATG- GAAGGCTCTACTGTTC-3P and 5P-ACCT-GGTGT- TTGACCTGAACTA-3P (scpb), 5P-TGTAAAGGACGA- TAGTGTGAAGAC-3P and 5P-CATTTGTGATTCC- CTTTTGC-3P (bac), 5P-CAGAGTACAGGAAGGG- CTAGTC-3P and 5P-TTCTTCCGTCCACTTAGGATC- 3P (bca), 5P-ACGGTATCAACCTTGAAACTGG-3P and 5P-TCAGTGTTGATTTCCCAGATGTA-3P (scaa). PCR products to be used as DNA probes were isolated from agarose gels using the Golden Beads PCR Products Puri- cation kit (Sangon Ltd., Shanghai, PR China) Computer analysis The sequences of the genes were accessed through the Entrez database ( The sequences of the primers were constructed using the computer program OLIGO. After electrophoresis, the sizes of DNA fragments were estimated by the computer program SEQAID using a 100-bp ladder, V-phage concatemers and Saccharomyces cerevisiae chromosomes (Bio- Rad Laboratories) as DNA molecular size standards. 3. Results and discussion 3.1. Heterogeneity of PFGE restriction patterns and genome sizes of di erent strains Comparative analysis of PFGE restriction pro les of GBS strains is usually based on the size variation of large SmaI fragments [7,11,15,16]. RFLP of the relatively small SmaI fragments is usually not analyzed due to the di culties of separation and comparison of the fragments. In the present study, digestion of chromosomal DNA of 45 GBS strains with restriction endonuclease SmaI resulted in 9^14 SmaI fragments, ranging in size from 20 to 1120 kb. Analysis of the mobilities of the relatively small SmaI fragments demonstrated a high degree of size heterogeneity and allowed di erentiation of the strains. When DNA fragments larger than 215 kb in size were used for comparative analysis, 23 di erent PFGE patterns were discovered. However, when the fragments in the region between 42 and 100 kb in size were compared, we found an even higher degree of heterogeneity of PFGE patterns, increasing the number of PFGE patterns up to 32 of the 45 GBS strains (Fig. 1). Strains A43 and 87/29, and D2 as well as and B1 were isolated in di erent geographical regions at di erent dates and were not distinguished by PFGE. However, analysis of the strains employing ribotyping demonstrated several di erences (Table 1), suggesting that combined analysis of the well-separated PFGE fragments and ribosomal types is desirable and perhaps necessary for accurate epidemiological studies. Estimated molecular size of the entire GBS genome of these strains varied from 2030 kb (strain A12) to 2290 kb (strains 96-8 and A211). The average genome size (2173 kb) was almost identical to the genome sizes of two GBS strains discovered in our previous investigations [12]. The maximum di erence in the genome sizes of GBS strains was 260 kb. However, some of the strain speci c restriction patterns (D4 and D60, and B1, etc.) were di erent by only one fragment with di erence in size equal C Fig. 1. Schematic representation of SmaI physical genome maps of GBS strains with location of virulence genes. The numbers above the maps indicate the sizes of SmaI restriction fragments in kb.

3 A. Dmitriev et al. / FEMS Microbiology Letters 208 (2001) 93^98 95

4 96 A. Dmitriev et al. / FEMS Microbiology Letters 208 (2001) 93^98 Table 1 Ribosomal types of group B streptococcal strains Ribosomal types bac gene presence bca gene presence Number of the strains Serological types Origin PvuII HindIII EcoRI P1 H1 E II China P1 H1 E NT China P1 H1 E Ia China P1 H1 E , 94/95, 29/95 Ib Russia P1 H1 E , 6, 10, 43, 89, 137, 254, 60 II China P1 H1 E , 87/29 II Russia P1 H1 E , 71, 82 II China P1 H1 E Ia China P1 H1 E Ib Russia P5 H1 E Ib China P2 H3 E5 ^ ^ 97-3, III China P3 H1 E1 ^ ^ Ia China P3 H3 E3 ^ ^ 15 Ia China P4 H1 E7 ^ ^ Ia China P4 H1 E7 ^ ^ 278 Ia Russia P4 H1 E7 ^ + 97, 123, 407 Ia Russia P4 H3 E1 ^ ^ III China P4 H3 E4 ^ ^ 2 III China P6 H3 E1 ^ ^ 80 II Russia P7 H1 E2 ^ Ia China P7 H1 E7 ^ ^ 211 III China P8 H1 E7 ^ ^ 96-8 III China P8 H2 E1 ^ ^ 1, , III China P9 H1 E2 ^ ^ 73 V China P10 H1 E1 ^ ^ 99 III China P10 H1 E1 ^ ^ Bax III Russia P11 H1 E9 ^ ^ 171 Ia Russia P12 H1 E1 ^ Ia Russia Total NT: nontypeable. to 5^25 kb (Fig. 1). It is possible that this size heterogeneity re ects acquisition or excision of di erent mobile genetic elements such as bacteriophages, IS-elements or transposons Genome mapping of GBS strains The preliminary genetic maps of two GBS strains were previously established with location of several virulence and `house keeping' genes on the maps [12]. In this study, the relative location of nine GBS genes potentially important for virulence phenotype expression was tested for 45 strains. The major goal of these experiments was to investigate the comparative genome organization of signi cant number of strains and analyze the possibility of genetic linkages of virulence genes on GBS genomes as it was found in virulence islands of other bacteria [17]. All of the strains were found to possess the genes glna (glutamine synthetase), scpb (C5a peptidase), cyl (L-hemolytic activity), hylb (hyaluronidase), lmb, scaa (aggregation factor) and cfb (CAMP protein). The sizes of the fragments ampli ed by PCR were in accordance with the previously reported DNA sequences. The genes bca (Kantigen) and bac (L-antigen with IgA binding capacity), however, were found only in a subset of the strains: 22 of 45 strains were bac and 27 of 45 strains were bca. Twenty two strains possessed both bac and bca genes, and ve strains possessed only the bca gene (Table 1). The genes under study were located on SmaI restriction fragments by hybridization analysis. The glna and cfb genes as well as the scpb, hylb and lmb genes were located on the same fragments in all the strains. The smallest fragment containing all three genes (scpb, hylb, lmb) was equal to 91 kb (Fig. 1). Comparison of strain speci c PFGE patterns and hybridization patterns resulted in the construction of the genetic maps of 45 GBS strains (Fig. 1). The order of some fragments (smaller than 100 kb), which did not hybridize to the gene probes used in the study, could not be established with certainty. However, the analysis of the genetic maps of the GBS strains suggests that the entire genome organization of group B streptococci is highly conserved (Fig. 1) as it was revealed for group A streptococci [18] Restriction polymorphism of rdna patterns Several recent publications have been devoted to the ribotyping of GBS strains in order to nd a correlation between ribotypes and serological types; however, no

5 A. Dmitriev et al. / FEMS Microbiology Letters 208 (2001) 93^98 97 Fig. 2. A: Schematic representation of PvuII ribosomal types of GBS strains. Tracks 1^12 correspond to ribotypes P1^P12; the numbers on the right represent the number of kb pairs of V-HindIII fragments. B: Schematic representation of HindIII ribosomal types of GBS strains. Tracks 1^3 correspond to ribotypes H1^H3; the numbers on the right represent the number of kb pairs of V-HindIII fragments. C: Schematic representation of EcoRI ribosomal types of GBS strains. Tracks 1^9 correspond to ribotypes E1^E9; the numbers on the right represent the number of kb pairs of V-HindIII fragments. correlation between the HindIII or EcoRI ribotypes and the serological types or genotypes of the strains tested. Twenty-eight of 45 strains (62.2%) were characterized by either EcoRI ribotype E1 or ribotype E2 and 36 of 45 strains (80%) belonged to the HindIII ribotype H1 (Table 1). Among PvuII ribosomal types, ribotype P1 was observed in 21 of the 22 strains containing the gene bac with only one bac strain belonging to ribotype P5 (Table 1). The 23 bac 3 strains of di erent serotypes were divided into 10 di erent ribotypes, P2^4, P6^P12. These results demonstrate that ribotyping employing the PvuII enzyme was the most valuable for di erentiation of GBS strains, while HindIII and EcoRI enzymes were giving the same patterns for the large number of genetically unrelated strains. The analysis of the HindIII, EcoRI and PvuII ribotypes, taken together, allowed the detection of 21 combined ribotypes. Three of them, P1H1E1, P1H1E2 and P4H1E7, were speci c for 21 of 45 strains (Table 1). It is possible that these three genetic lines are typical for GBS strains isolated in Russia and China Genetic features of bac gene-positive GBS strains The gene bac was discovered in 22 of 45 GBS strains tested by PCR and DNA^DNA hybridization (Fig. 1). These strains belonged to di erent serotypes (Table 1). The occurrence of the gene bac was compared with ribosomal types of the GBS strains. We found that 21 of 22 strains with the bac gene isolated from di erent countries belonged to the same PvuII ribotype P1 (Table 1). PFGE analysis also showed a strong correlation between the PFGE pattern and the presence of bac gene in the genome. Not one of bac gene-negative strains was characterized by PFGE restriction pro le speci c for bac gene-positive strains (Fig. 3). To our knowledge none of the bac gene positive strains had any epidemiological relationship. However, due to the high degree of the similarity in genetic maps, genotypes, ribotypes and serotypes, these compelling correlation of such kind has been demonstrated [5,6,10]. In our study, three di erent HindIII, nine EcoRI and 12 PvuII rdna patterns were revealed among 45 randomly chosen GBS strains isolated in Saint Petersburg and Beijing (Fig. 2). We could not nd a strong Fig. 3. PFGE restriction patterns of GBS strains. Pulse time ^ 30 s, run time ^ 20 h, voltage ^ 200 V. Lanes 1^5 show the bac gene-negative strains 96-8, , , , B15; lanes 6^11 show the bac genepositive strains A12, D82, B71, A254, A6, A89; lane 12 shows molecular mass marker (Saccharomyces cerevisiae chromosomes, kb).

6 98 A. Dmitriev et al. / FEMS Microbiology Letters 208 (2001) 93^98 bac gene-positive strains probably are genetically related. This conclusion corresponds to the recommendation for interpreting PFGE patterns [19]. These results are in agreement with our previous data demonstrating that GBS strains with the bac gene share common characteristics and di er from GBS strains without this gene [12,13]. The acquisition of the bac gene in the past by some ancestral GBS recipient strain may have provided a new biological advantage to this strain. This could result in the rapid and e ective spreading of the resulting clone within the human population. However, the origin of the bac gene in GBS is still unknown. Recently, it was reported that the BAC protein of GBS contains domains homologous to eukaryotic immunoglobulin superfamily domains and it was proposed that these domains were acquired horizontally by prokaryotes from eukaryotes after their divergence [20]. In the near future it will be interesting to investigate the mechanisms of the bac gene transfer in GBS. In conclusion, this study presents new data about the GBS genome and the chromosomal DNA size heterogeneity of di erent strains. It was found out that both ribotyping and PFGE analyses taken together provide an additional and valuable source of information for epidemiology. The homogeneity of all bac gene-positive strains from Russia and PR China leads us to the conclusion that all these strains belong to the same genetic subpopulation. Acknowledgements We are thankful to Prof. A.A. Totolian (Institute of Experimental Medicine, Saint Petersburg, Russia) and Prof. J.J. Ferretti (University of Oklahoma, Oklahoma City, OK, USA) for the comments which helped to publish this paper in the present form. This work was supported by funding from the Ministry of Science and Technology, PR China, by the Grants and from the Russian Foundation for Basic Research and Public Health Service Grants AI19304 and TW00188 from the National Institutes of Health. References [1] Gray, B.M. and Dillon, H.C. (1985) GBS infections in mothers and their infants. Antibiot. Chemother. 35, 225^236. [2] Regan, J.A., Klebano, M.A., Nugent, R.P., Eschenbach, D.A., Blackwelder, W.C., Lou, Y., Gibbs, R.S., Rettig, P.J., Martin, D.H. and Edelman, R. (1996) Colonization with group B streptococci in pregnancy adverse outcome. VIP Study Group. Am. J. Obstet. Gynecol. 174, 1354^1360. [3] Jones, A.L., Knoll, K.M. and Rubens, C.E. (2000) Identi cation of Streptococcus agalactiae virulence genes in the neonatal rat sepsis model using signature-tagged mutagenesis. Mol. Microbiol. 37, 1444^ [4] Nizet, V., Ferrieri, P. and Rubens, C. (2000) Molecular pathogenesis of group B streptococcal disease in newborns. In: Streptococcal Infections, Clinical Aspects, Microbiology and Molecular Pathogenesis (Stevens, D.L. and Kaplan, E.L., Eds.), pp. 180^221. Oxford University Press, New York. [5] Blumberg, H.M., Stephens, D.S., Licitra, C., Piggot, N., Facklam, R., Swaminathan, B. and Wachsmuth, I.K. (1992) Molecular epidemiology of group B streptococcal infections: use of restriction endonuclease analysis of chromosomal DNA and DNA restriction fragment length polymorphism of ribosomal RNA genes (ribotyping). J. Infect. Dis. 166, 574^579. [6] Chattelier, S., Huet, H., Kenzi, S., Rosenau, A., Geslin, P. and Quentin, R. (1996) Genetic diversity of rrna operons of unrelated Streptococcus agalactiae strains isolated from cerebrospinal uid of neonates su ering from meningitis. J. Clin. Microbiol. 34, 2741^2747. [7] Elliot, J.A., Farmer, K.D. and Facklam, R.R. (1998) Sudden increase in isolation of group B streptococci, serotype V, is not due to emergence of a new pulsed- eld gel electrophoresis type. J. Clin. Microbiol. 36, 2115^2116. [8] Ellis, S., Kotiw, M. and Garland, S.M. (1996) Restriction endonuclease analysis of group B streptococcal isolates from two distinct geographical regions. J. Hosp. Infect. 33, 279^287. [9] Hauge, M., Jespersgaard, C., Poulsen, K. and Kilian, M. (1996) Population structure of Streptococcus agalactiae reveals an association between speci c evolutionary lineages and putative virulence factors but not disease. Infect. Immun. 64, 919^925. [10] Huet, H., Martin, C., Geslin, P., Grimont, F. and Quentin, R. (1993) Ribotyping of Streptococcus agalactiae strains isolated from vaginas of asymptomatic women. Res. Microbiol. 144, 457^465. [11] Rolland, K., Marois, C., Siquier, V., Cattier, B. and Quentin, R. (1999) Genetic features of Streptococcus agalactiae strains causing severe neonatal infections, as revealed by pulsed- eld gel electrophoresis and hylb gene analysis. J. Clin. Microbiol. 37, 1892^1898. [12] Dmitriev, A., Suvorov, A. and Totolian, A. (1998) Physical and genetic maps of Streptococcus agalactiae, serotypes II and III; rrna operon organization. FEMS Microbiol. Lett. 167, 33^39. [13] Suvorov, A., Dmitriev, A., Ustinovich, I., Schalen, C. and Totolian, A. (1997) Molecular analysis of clinical group B streptococcal strains by use of K and L gene probes. FEMS Immunol. Med. Microbiol. 17, 149^154. [14] Maniatis, T., Fritsch, E.F. and Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. [15] Ferrieri, P., Cho, D.S., Livdahl, C., Rubens, C.E. and Flores, A.E. (1997) DNA restriction pro les of nontypable group B streptococcal clinical isolates. Adv. Exp. Med. Biol. 41, 343^346. [16] Granlund, M., Oberg, L., Sellin, M. and Norgren, M. (1998) Identi- cation of a novel insertion element, IS 1548, in group B streptococci, predominantly in strains causing endocarditis. J. Infect. Dis. 177, 967^976. [17] Hacker, J. (2000) Pathogenicity islands and the evolution of microbes. Annu. Rev. Microbiol. 54, 641^679. [18] Suvorov, A.N., Ferretti, J.J. (2000) Comparative analysis of chromosomal organization of group A streptococcal strains. In: Proceedings of XIV Lance eld International Symposium on Streptococci and Streptococcal Diseases, pp. 381^384 [19] Tenover, F.C., Arbeit, R.D., Goering, R.V., Mickelsen, P.A., Murray, B.E., Persins, D.H. and Swaminsthan, B. (1995) Interpreting chromosomal DNA restriction patterns produced by pulsed- eld gel electrophoresis; criteria for bacterial strain typing. J. Clin. Microbiol. 33, 2233^2239. [20] Bateman, A., Eddy, S.R. and Chothia, C. (1996) Members of the immunoglobulin superfamily in bacteria. Protein Sci. 5, 1939^1941.