INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Streptococcus pneumoniae continues to be a common cause of serious and life-threatening infections, such as pneumonia, bacteremia, and meningitis, in both adults and children (1). Pneumococci can be classified according to differences in capsular polysaccharide structure. As many as 90 different capsular types can be distinguished by serotyping (10). The distribution of serotypes varies in different populations and different geographic areas, and certain pneumococcal serotypes are known to be more virulent than others (24, 28). Pneumococcal serotype 3 isolates are considered to represent the most virulent serotype. These isolates are often responsible for invasive disease (17, 20), particularly in adults (15, 19). Bacteremia caused by this organism is considered to have the highest mortality rate compared to that caused by other serotypes (15, 20). To date, the frequency of penicillin resistance among serotype 3 isolates has remained low (16).

Serotyping as a tool for epidemiological studies has several disadvantages. S. pneumoniae is a naturally transformable species, and frequent exchange of capsular genes occurs (2-4, 13, 23). In addition, serotyping determines the variation in a single genetic locus, i.e., the cps locus. Therefore, several other typing methods have been developed to assist with the identification of relatedness between strains and their cellular structures. These methods include multilocus enzyme electrophoresis (9), penicillin-binding protein (PBP) profile analysis (21, 22), pneumococcal surface protein A typing (22), and DNA fingerprint methods, such as pulsed-field gel electrophoresis, multilocus sequence typing (MLST) (7), ribotyping, restriction fragment end labeling (RFEL) analysis, BOX PCR fingerprinting, and DNA fingerprinting of the PBP genes (14, 32). RFEL analysis provides a high degree of discriminatory power, and RFEL profiles are reproducible and suitable for computerized comparisons (14). In addition, RFEL analysis provides a DNA fingerprint that represents multiple loci in the pneumococcal genome. This technique is routinely used in our laboratory to generate a data library of pneumococcal DNA fingerprints. In this study, we investigated the molecular epidemiological characteristics of S. pneumoniae strains isolated in a nationwide manner from patients with meningitis in The Netherlands in 1994. The genetic relatedness within pneumococcal serotypes was determined. In addition, we studied the molecular epidemiological characteristics of epidemiologically unrelated serotype 3 pneumococci from four distinct countries. The isolates were characterized by serotyping, RFEL analysis, and PBP genotyping.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial isolates. We studied a collection of S. pneumoniae strains (n = 153) isolated from Dutch patients suffering from meningitis in The Netherlands in 1994. These strains were collected by the National Reference Center for Bacterial Meningitis in a nationwide manner and represent all pneumococcal meningitis isolates collected in a 1-year period. In addition, these strains were penicillin susceptible and were presumed to be epidemiologically unrelated. In addition, 42 penicillin-susceptible invasive serotype 3 pneumococci were isolated from patients in the United States (n = 27), Thailand (n = 9), and Denmark (n = 6). The latter strains were also presumed to be epidemiologically unrelated, since they were isolated from various geographic regions within these countries and at different times ranging from 1960 to 1962 and from 1992 to 1998 (Table 1).

Serotyping. Pneumococci were serotyped on the basis of capsular swelling (Quellung reaction) observed microscopically after suspension in antisera prepared at Statens Seruminstitut, Copenhagen, Denmark (8).

RFEL analysis. Typing of pneumococcal strains by RFEL analysis was performed as described by van Steenbergen et al. (33) and adapted by Hermans et al. (14). Briefly, purified pneumococcal DNA was digested with restriction enzyme EcoRI. The DNA restriction fragments were end labeled at 72°C with [-³²P]dATP by using Taq DNA polymerase (Goldstar; Eurogentec, Seraing, Belgium). The radiolabeled fragments were denatured and separated electrophoretically on a 6% polyacrylamide sequencing gel containing 8 M urea. The gel was transferred to filter paper, vacuum dried (HBI, Saddle Brook, N.Y.), and exposed to ECL Hyperfilms (Amersham, Little Chalfont, Bucks, United Kingdom).

PBP genotyping. Genetic polymorphisms of the penicillin resistance genes pbpla, pbp2b, and pbp2x were investigated by restriction fragment length polymorphism analysis. PCR amplification of the PBP-encoding genes was performed with a 50-µl PCR buffer system containing 75 mM Tris-HCl (pH 9.0), 20 mM (NH₄)₂SO₄, 0.01% (wt/vol) Tween 20, 1.5 mM MgCl₂, 0.2 mM each deoxynucleoside triphosphate, 10 pmol of each primer, 0.5 U of Taq DNA polymerase (Goldstar), and 10 ng of purified chromosomal DNA. Cycling was performed with a PTC-100 programmable thermal controller (MJ Research, Watertown, Mass.) and consisted of the following steps: predenaturation at 94°C for 1 min; 30 cycles of 1 min at 94°C, 1 min at 52°C, and 2 min at 72°C; and final extension at 72°C for 3 min. The primers used to amplify the genes pbp1a, pbp2b, and pbp2x were described previously (3, 6, 21). The amplification products (5 µl) were digested with restriction endonuclease HinfI and separated by electrophoresis in 2.5% agarose gels (27). Gels were scanned and printed with a Geldoc 2000 system (Biorad, Veenendaal, The Netherlands). The different PBP genotypes are represented by a three-number code (e.g., 06-14-43), referring to the restriction fragment length polymorphism patterns of the genes pbp1a (pattern 6), pbp2b (pattern 14), and pbp2x (pattern 43), respectively.

Computer-assisted analysis of the DNA banding patterns. The RFEL types were analyzed with the Windows version of Gelcompar software, version 4 (Applied Maths, Kortrijk, Belgium), after imaging of the RFEL autoradiograms with Image Master DTS (Pharmacia Biotech, Uppsala, Sweden). DNA fragments in the molecular size range of 160 to 400 bp were documented. The DNA banding patterns were normalized with pneumococcus-specific bands present in the RFEL banding patterns of all strains. Comparison of the banding patterns was performed by the unweighted pair-group method with arithmetic averages (26) and with the Jaccard similarity coefficient applied to peaks (31). Computer-assisted analysis and the methods and algorithms used in this study were in accordance with the instructions of the manufacturer of Gelcompar. A tolerance of 1.2% in band positions was applied during comparison of the DNA patterns.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Epidemiology of invasive pneumococcal isolates in The Netherlands. The epidemiology of S. pneumoniae strains isolated in a nationwide manner from patients with meningitis in 1994 in The Netherlands was investigated. These strains (n = 153) were all found to be penicillin susceptible and were analyzed by serotyping, PBP typing, and RFEL typing. The results are shown in Fig. 1 and Table 2. The invasive isolates represented 31 serotypes: 1 (n = 3), 3 (n = 8), 4 (n = 3), 5 (n = 2), 6A (n = 7), 6B (n = 15), 7F (n = 7), 8 (n = 4), 9N (n = 4), 9V (n = 7), 10F (n = 2), 10A (n = 4), 11A (n = 2), 14 (n = 12), 15A (n = 1), 15C (n = 2), 16F (n = 2), 18F (n = 1), 18B (n = 2), 18C (n= 12), 19F (n = 18), 19A (n = 2), 22F (n = 1), 23F (n = 16), 23A (n = 1), 23B (n = 2), 24F (n = 3), 32A (n = 1), 33F (n = 5), 34 (n = 1), and 38 (n = 3).

View larger version (49K): [in this window] [in a new window]   FIG. 1.   Genetic relatedness of 153 penicillin-susceptible invasive pneumococcal isolates, based on the RFEL banding patterns of the isolates. The country code (NL, The Netherlands), strain codes, RFEL types, and serotypes are depicted. Codes I to XI refer to genetic clades of pneumococcal strains; genetic clusters are indicated by a grey box in the dendrogram (for definitions, see Materials and Methods).

Genetic relatedness within serotypes in The Netherlands. The genetic relatedness of strains within the nine most predominant serotypes present in the collection was investigated. All strains of serotype 7F (n = 7) belonged to clade IX, and all strains of serotype 9V (n = 7) belonged to clade VII. Strains of serotype 3 (n = 8) belonged to two distinct genetic clades, I and VIII. Strains of serotype 14 (n = 12) represented three distinct genetic clades, III, X, and XI. Strains of serotypes 6B, 18C, and 23F were genetically more heterogeneous. However, most strains of serotypes 6B, 18C, and 23F belonged to one clade. Eight of the 15 serotype 6B strains belonged to clade V, 9 of the 12 serotype 18C strains belonged to clade III, and 7 of the 16 serotype 23F strains belonged to clade IV. Strains with serotypes 6A and 19F displayed the most heterogeneity in this collection of S. pneumoniae strains, as 7 serotype 6A strains were represented by 4 genetic clades and 18 serotype 19F strains were represented by 11 genetic clades (Fig. 1).

Genetic relatedness within serotype 3 isolates of distinct geographic origins. We investigated the molecular epidemiology of serotype 3 strains from The Netherlands (n = 8) and three additional countries: the United States (n = 27), Thailand (n = 9), and Denmark (n = 6). These 50 epidemiologically unrelated serotype 3 strains were characterized by RFEL analysis. Four distinct RFEL clades and seven RFEL types that were less than 85% related to other serotype 3 strains were observed among these strains (Fig. 2). The most predominant RFEL clade, I, represented 29 serotype 3 strains (58%). This RFEL clade was represented by 22 isolates from the United States, 2 isolates from Denmark, and 5 isolates from The Netherlands. RFEL cluster VIII was represented by eight strains (16%)two American, three Danish, and three Dutch strains. RFEL clade XII was represented by four Thai isolates. In addition, two Thai isolates formed a Thai-specific clade. Thus, 43 strains shared RFEL types with at least one other strain (86%). Seven serotype 3 strains with RFEL types 296, 295, 165, 105, 289, 76, and 242 did not match the four genetic clades, and six of them did not match any of the 153 Dutch invasive strains representing 116 RFEL types and 31 serotypes. In contrast, the serotype 3 strain with RFEL type 105 was genetically related (90.9%) to a serotype 19F strain representing RFEL type 352. 

View larger version (40K): [in this window] [in a new window]   FIG. 2.   Genetic relatedness of 50 penicillin-susceptible pneumococcal serotype 3 isolates, based on the RFEL banding patterns of the isolates. RFEL types are depicted. Codes I, VIII, XII, and XXVII refer to genetic clades of pneumococcal strains; genetic clusters are indicated by a grey box in the dendrogram (for definitions, see Materials and Methods). Bars represent the number of isolates per RFEL type.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Few studies have documented genotype analyses of penicillin-susceptible strains (12, 29) and of serotype-specific strains (9, 18). We investigated the epidemiological characteristics of 153 penicillin-susceptible S. pneumoniae strains isolated from patients with meningitis in The Netherlands in 1994. The isolates represented 31 serotypes. The most predominant serotypes were 19F, 23F, 6B, 18C, 14, 3, 6A, 7F, and 9V. Various investigators have reported the occurrence of horizontal transfer of capsular genes (2, 11-13). In Dutch penicillin-susceptible isolates, horizontal transfer of capsular genes has occurred frequently. A high frequency of capsular exchange has been reported in molecular epidemiological studies of penicillin-resistant isolates from many countries (12, 13). This is the first study suggesting the frequent occurrence of horizontal transfer of capsular genes among penicillin-susceptible isolates.

RFEL analysis revealed that 52% of the strains belonged to genetic clusters. The amount of genetic clustering was substantially lower among the penicillin-susceptible isolates than among the penicillin-nonsusceptible isolates in other studies (2, 5, 11-13, 25). A comparison of the penicillin-susceptible invasive isolates studied here with 193 penicillin-nonsusceptible strains representing 193 distinct RFEL types in the international data library and representing 16 countries revealed no overlap (12, 13).

The PBP genotypes 02-02-03, 02-02-71, and 02-02-02 were found most frequently. This observation corresponds with the PBP typing results for penicillin-susceptible pediatric carriage isolates in the U.S. population (30). Interestingly, four additional PBP genotypes (02-02-14, 02-02-15, 02-02-16, and 02-02-05) were identified for serotypes 8, 5, 32A, and 19F, respectively. The serotype specificity of the latter PBP genotypes suggests a divergence of the PBP genotypes before the origin of the capsular types 8, 5, 32A, and 19F.

The genetic relatedness within the specific pneumococcal serotypes was highly variable. RFEL genotypes of serotype 6A and 19F strains displayed high levels of heterogeneity; i.e., strains of these serotypes represented many RFEL types that belonged to many genetic clusters and genetic clades. In contrast, the RFEL genotypes of serotype 7F, 9V, 14, and 3 strains were found to be genetically related. Interestingly and consistent with our observations, Canadian penicillin-susceptible isolates of serotypes 3 and 7F were also more genetically related than isolates of other serotypes (18). Moreover, invasive penicillin-susceptible serotype 3 isolates from the United Kingdom also tended to be more closely related to each other than to isolates of other serotypes (9).

We focused on the molecular epidemiological characteristics of epidemiologically unrelated serotype 3 pneumococci and extended our serotype 3 collection with isolates from the United States, Thailand, and Denmark. RFEL analysis demonstrated that serotype 3 strains isolated in these countries displayed a strong degree of genetic relatedness: the vast majority of the strains represented two distinct RFEL clades. Furthermore, both genetic clades harbored isolates from three countries: the United States, Denmark, and The Netherlands. These observations indicate that two serotype 3 clones have been disseminated internationally. In addition, six Thai serotype 3 isolates belonged to two RFEL clades (clades XII and XXVII). The data suggest strong genetic homogeneity within the serotype 3 pneumococci and support the observations for Canada and the United Kingdom (9, 18). Interestingly, the Canadian serotype 3 strains displayed two distinct genotypes, and the majority of the epidemiologically nonrelated serotype 3 strains from the United Kingdom displayed two genotypes. Moreover, MLST analysis of serotype 3 strains isolated in six countries identified two major genetic clusters (M. C. Enright and B. G. Spratt, http://mlst.zoo.ox.ac.uk). Since the strains have been characterized by distinct typing methods, i.e., pulsed-field gel electrophoresis, multilocus enzyme electrophoresis, MLST, and RFEL analysis, and since there is no overlap in the characterized strains, the genetic relatedness between the latter serotype 3 strains and the strains characterized in this study is currently unknown. The remaining six serotype 3 RFEL types each occurred once in our collection. Our observations suggest that these latter strains have been derived from horizontal transfer of the capsular type 3 gene locus to other pneumococcal genotypes.

PBP genotyping of the serotype 3 strains demonstrated limited variation in the pbp1a, pbp2b, and pbp2x genes. All serotype 3 strains from the United States, Denmark, and The Netherlands displayed PBP genotype 02-02-71. However, variation was demonstrated in the Thai serotype 3 isolates. PBP type 09-02-71 was represented by a single Thai isolate. This PBP type was also specific for the penicillin-susceptible phenotype, as there was no overlap with penicillin-nonsusceptible isolates from 16 countries (13).

In conclusion, pneumococcal strains belonging to serotype 3 display limited genetic heterogeneity despite the lack of epidemiological relatedness. We hypothesize that this serotype has recently evolved or has remained unchanged for a prolonged period. The few serotype 3 isolates not belonging to the main clusters are presumably derived from horizontal transfer of capsular genes.