Multilocus Sequence Typing System for Group B Streptococcus

Strain collection.The study collection consisted of 152 isolates of GBS from North America, New Zealand, Thailand, Singapore, Israel, Japan, and the United Kingdom. In addition, two well-characterized strains were included, the NCTC-8541 strain (isolated from a vaginal carrier; Public Health Laboratory, United Kingdom) and the NEM316 strain (ATCC 12403, isolated from a case of fatal neonatal sepsis, country of origin unknown), whose genome has been fully sequenced (11). The collection was globally diverse and included strains from asymptomatic carriers as well as human infections. Most capsular serotypes of GBS were represented, although serotypes IV and VII, which are rarely associated with disease in humans, were not included. Table 1 further describes the strain collection.

Identification of GBS.Isolates were grown on blood agar and identified as group B streptococcus by the following criteria (21): β-hemolysis on a Columbia agar plate containing 5% horse blood (two strains of the 152 analyzed were nonhemolytic), Gram staining showing gram-positive cocci in pairs or short chains, negative reaction with catalase reagent, and Lancefield grouping with type B antisera (Oxoid, Basingstoke, United Kingdom).

Characterization of strains.Capsular serotyping was carried out on all strains in the laboratory of John Bohnsack. Capsular serotyping was repeated in a second laboratory with different method for 25 (16.5%) randomly selected isolates (Streptococcal Reference Laboratory, Central Public Health Laboratory, Colindale, United Kingdom [n = 20], and Clinical Microbiology Laboratory, KK Women's and Children's Hospital, Singapore [26] [n = 5]). Forty strains had previously been characterized by RDP (4).

DNA extraction.The DNeasy kit (Qiagen GmbH) was used to extract DNA, and the gram-positive bacterial protocol was followed. A single colony of each strain was streaked across a Columbia agar plate containing 5% horse blood. Several colonies were picked off into phosphate-buffered saline and centrifuged at 5,500 × g. The cell pellet was resuspended in 180 μl of enzymatic lysis buffer containing lysozyme (20 mg/ml) and incubated for 30 min at 37°C. Then 25 μl of proteinase K (10 mg/ml) was added, and incubation was continued at 70°C for 30 min. The DNA in the clear viscous lysates was precipitated with 95% (vol/vol) ethanol and added to DNeasy minicolumns. Ethanol (70%, vol/vol)-based buffers AW1 and AW2 were added sequentially to the columns and centrifuged at 5,500 × g. The supernatants were discarded, and the DNA was resuspended in sterile water and stored at −20°C.

Choice of loci for MLST.Ten candidate loci, encoding enzymes involved in intermediary metabolism, were identified by searching the genome sequence of the GBS strain NEM316 (11) with homologous sequences from other bacteria. Suitable genes were then chosen on the basis of chromosomal location and sequence diversity observed in pilot studies with a restricted set of GBS strains. Three genes were excluded, two because they failed to distinguish between GBS strains and one which was not reliably amplified. The following seven loci were selected for the MLST scheme (Table 2): alcohol dehydrogenase gbs0054 (adhP), phenylalanyl tRNA synthetase (pheS), amino acid transporter gbs0538 (atr), glutamine synthetase (glnA), serine dehydratase gbs2105 (sdhA), glucose kinase gbs0518 (glcK), and transketolase gbs2105 (tkt). The chromosomal locations of these housekeeping loci (Table 2) suggested that it was unlikely for any of them to be coinherited in the same recombination event, as the minimum distance between two loci was 20 kb.

Amplification and nucleotide sequence determination.PCR products were amplified with oligonucleotide primer pairs designed from the NEM316 GBS genome sequence (11). A range of primers were tested, with those shown in Table 3 providing reliable amplification from a diverse range of GBS isolates. Each 50-μl amplification reaction mixture comprised 10 ng of GBS chromosomal DNA, 100 pmol of each PCR primer (MWG Biotech, Ebersberg, Germany), 1× PCR buffer with 1.5 mM MgCl₂ (Qiagen GmbH), 0.5 U of Taq DNA polymerase (Qiagen GmbH), and 1.6 mM deoxynucleoside triphosphate mix (ABgene, Epsom, United Kingdom). The reaction conditions were denaturation at 94°C for 1 min, primer annealing at 55°C for 45 s, and extension at 72°C for 1 min for 30 cycles.

The amplification products were purified by precipitation with 20% polyethylene glycol and 2.5 M NaCl (8), and their nucleotide sequences were determined at least once on each DNA strand with internal nested primers (Table 3) and ABI Prism BigDye Terminators version 3.0 reaction mix (Applied Biosystems, Foster City, Calif.) in accordance with the manufacturer's instructions. Unincorporated dye terminators were removed by precipitation of the termination products with sodium acetate (3 M, pH 5.2) and 95% (vol/vol) ethanol, and the reaction products were separated and detected with an ABI Prism 3700 DNA analyzer (Applied Biosystems). Sequences were assembled from the resultant chromatograms with the Staden suite of computer programs and edited to resolve any ambiguities (24).

Allele and sequence type assignment.For each locus, every different sequence was assigned a distinct allele number in order of identification; these were internal fragments of the gene which contained an exact number of codons. Any change in the nucleotide sequence, whether or not the amino acid sequence was altered, was defined as a new allele. Each isolate was therefore designated by a seven-integer number, constituting its allelic profile. Isolates with the same allelic profile were assigned to the same sequence type (ST), which were numbered in the order of their identification (ST-1, ST-2, etc.). The data have been deposited in a database accessible on the Internet at http://sagalactiae.mlst.net.

Computational analyses.Determination of the number of polymorphic nucleotide sites, calculation of d_n/d_s, where d_n is nonsynonymous substitutions and d_s is synonymous substitutions, and construction of dendrograms with the unweighted pair group method with arithmetic mean (UPGMA) were performed with START (http://www.mlst.net) (14). STs were grouped into lineages or clonal complexes with BURST (START version 1.05, http://www.mlst.net [14]). The members of a BURST lineage were defined as groups of two or more independent isolates where each isolate had identical alleles at six or more loci with at least one other member of the group.

Variation at the seven MLST loci.The sequences of the seven chosen loci were determined for the 152 strains, and allelic profiles were assigned. The alleles defined for the MLST scheme were based on sequence lengths of between 459 (glcK) and 519 bp (sdhA). Between four (glcK) and 11 alleles (adhP) were present at each locus. The average number of alleles at each locus was 6.4, providing the potential to distinguish 4.4 × 10⁵ different genotypes. The proportion of variable nucleotide sites present in the selected housekeeping genes ranged from 1.2% (glnA) to 2.5% (sdhA) (Table 2 and Fig. 1). The proportions of nucleotide alterations that changed the amino acid sequence (nonsynonymous substitutions, d_n) and the proportions of silent changes (synonymous substitutions, d_s) were calculated for each gene. With these data, the d_n/d_s ratios were calculated for all seven loci and were all <1 (Table 2).

• Open in new tab
• Download powerpoint

FIG. 1.

Polymorphic nucleotide sites in GBS MLST genes. Only the variable sites are shown. The nucleotide at each site is shown for allele 1; only those that differ from the nucleotide in allele 1 are shown for the other alleles. Nucleotide sites are numbered in vertical format.

Relatedness of GBS isolates.The 152 isolates were resolved into 29 STs, 14 of which were identified only once (Table 4). One hundred and one isolates (66.5% of the data set) were represented by one of four STs, ST-1, ST-17, ST-19, and ST-23. The most common ST (ST-17) was identified 44 times in the data set, followed by ST-1 (21 isolates), ST-19 (20 isolates), and ST-23 (16 isolates). ST-3 was not identified in this data set but had been identified in a pilot study. UPGMA was used to construct a dendrogram from the matrix of pairwise allelic differences between the 29 STs of all 152 isolates (Fig. 2). BURST grouped the isolates into seven lineages (Fig. 2), which approximated well with the clusters of STs obtained by UPGMA.

• Open in new tab
• Download powerpoint

FIG. 2.

UPGMA dendrogram showing genetic relationships between the 29 STs. The allelic profile of each ST is shown in parentheses. The four most common STs are indicated by an asterisk.

Relationship between ST, capsular serotype, and restriction digest patterns.Capsular serotype was known for all 152 strains (Table 5), and there was complete correlation between capsular serotyping results among three laboratories. Five strains proved to be nontypeable. Serotype III was most common (78 strains, 44 of which belonged to ST-17), followed by serotypes Ia (19 strains), Ib (17 strains), V (15 strains), II (8 strains), VI (6 strains), and VIII (4 strains). Serotypes IV and VII were not represented in the data set. Capsular serotype was generally not restricted to specific STs, and four STs contained isolates with different capsular serotypes.

RDP typing results were known for 40 of the isolates within the data set (Table 6). RDP type correlated closely with ST; isolates of the same RDP type were identical by MLST or differed at only a single locus.

Relationship between lineage, host, and country of origin.The data set presented in this study was not specifically designed to investigate the relationship between isolates and country or host of origin. However, the following observations can be made. ST-1 and ST-19 were significantly associated with carriage (chi-squared test, Yates corrected, P = 0.004 and P = 0.008, respectively), and several different capsular serotypes were represented in these STs. The 44 strains within ST-17 were all serotype III. ST-17 was significantly associated with invasive neonatal disease (chi-squared test, Yates corrected, P = 0.0000001). ST-23 contained serotype Ia strains (11 of 16, 68.6%) from carriage and invasive disease. ST-1, ST-19, ST-17, and ST-23 each contained strains isolated from Australasia, Europe, Asia, and North America, suggesting global dispersal.