|Home | About | Journals | Submit | Contact Us | Français|
In this study, we analyzed 118 penicillin-nonsusceptible Streptococcus pneumoniae (PNSP) isolates (MICs, ≥0.12 μg/ml) recovered in Poland in 2003 to 2005 from patients with respiratory tract diseases and invasive infections. Seven different serotypes (14, 9V, 23F, 19F, 6B, 19A, and 6A, in order of descending frequency), seven alleles of the murM gene (murMA, murMB6, and the new murMB12 to -16 alleles), and 31 multilocus sequence types (STs) were observed. The vast majority of the PNSP isolates (90.7%) belonged to the international multiresistant clones, and among these, the Spain9V-ST156 clonal complex was the most prevalent (56 isolates) and was significantly overrepresented in invasive infections. The clone has been evolving rapidly, as demonstrated by the observed number of STs, the diversity in multiple-locus variable-number-tandem-repeat analysis (MLVA) types, and the polymorphism of pbp and pspA genes (coding for penicillin-binding proteins and the pneumococcal surface protein A, respectively). The presence and structure of the rlrA islet (encoding the pneumococcal pilus) were very well conserved. The Spain9V-ST156 clonal complex has been largely responsible for a decreasing susceptibility to penicillin among pneumococci in Poland in recent years, in spite of a relatively moderate antimicrobial use.
Streptococcus pneumoniae (pneumococcus) is a leading cause of community-acquired respiratory tract infections (RTIs) and one of the major agents of invasive bacterial diseases such as sepsis and meningitis (4). Its diminishing susceptibility to antimicrobials, especially to β-lactams (3, 13, 43, 44, 46), is of the highest concern. In a clinical isolate, the loss of β-lactam susceptibility results from DNA acquisition from a penicillin-nonsusceptible S. pneumoniae (PNSP) strain (11, 23, 51) or a viridans group streptococcus (14). Changes in three of the six pbp genes, encoding the penicillin-binding proteins (PBPs), which are active in peptidoglycan synthesis, are the major resistance factors (56). However, other determinants, such as variants of the murM gene, play a role in resistance expression as well (7, 12, 19, 20).
Modern typing methods, mainly multilocus sequence typing (MLST) (1), greatly facilitate tracking the geographic spread of specific S. pneumoniae strains and following the dynamics of microbial populations over time. It was found that relatively few clones, the so-called multiresistant international clones defined by the Pneumococcal Epidemiology Network (PMEN) (34; www.sph.emory.edu/pmen), cause increases in pneumococcal resistance to β-lactams and other drugs (29, 36, 44, 47, 57). The “epidemiologic success” of a particular S. pneumoniae clone may also arise from factors augmenting its fitness. These would include virulence factors which improve the survival of pneumococci in the human host (24), such as the highly polymorphic pneumococcal surface protein A (PspA) (25), involved in the protection of pneumococci against the host immune system (53). Another factor, a pilus, was recently proposed to contribute to the spread of the Spain9V-ST156 clone (45).
In a previous report on Polish PNSP in 1995 to 2002 (44), we described several cases of importation of the international clones and the presence of those that had evolved locally. One of the imported clones, Spain9V-ST156, contributed much to the increase in PNSP frequency in 2002. Ongoing surveillance showed further PNSP growth in the next years, 2003 to 2005, and in this work we reveal this phenomenon and characterize the main PNSP clones.
(Parts of this work were presented at the 6th International Symposium on Pneumococci and Pneumococcal Diseases, 8 to 12 June 2008, Reykjavik, Iceland, and at the Conference Europneumo 2009, 4 to 7 June 2009, Bern, Switzerland.)
Five hundred sixty-four nonrepetitive pneumococcal isolates were collected in 2003 to 2005 from all over Poland as part of continuous surveillance of community-acquired RTIs and meningitis. The β-lactam susceptibility data for 261 isolates from RTIs from 2003 to 2004 were presented elsewhere (46), and those for the remaining 303 isolates were obtained in this study. Isolates were reidentified by colony morphology, susceptibility to optochin (bioMérieux, Marcy l'Etoile, France), and bile solubility. Susceptibility to penicillin was tested by the microdilution method according to CLSI recommendations (9). In order to keep consistency with earlier reports, the previous breakpoints of 0.12 μg/ml and 2 μg/ml, for intermediate and resistant phenotypes (37), respectively, were used, irrespective of the isolation site (42). Additionally, the epidemiological cutoff of 0.12 μg/ml, defined as the value which separates wild-type S. pneumoniae strains from those with acquired nonsusceptibility to penicillin (http://www.eucast.org; last accessed 18 January 2010), was applied. PNSP isolates identified in this study were also tested against amoxicillin, ceftriaxone, cefepime, and meropenem. Total DNA was purified using a Genomic DNA Prep Plus kit (A&A Biotechnology, Gdynia, Poland). Serotyping was performed by PCR as described by others (32, 40). Serotypes 6A and 6B were identified by amplicon sequencing as proposed by Pai et al. (39), and isolates of serotype 6A were further examined for serotype 6C by PCR (41).
The subgroup of RTI isolates from 2003 was also used in another analysis (28). They were screened for resistance to tetracycline, and the susceptibility, serotyping, and MLST data for three tetracycline- and penicillin-nonsusceptible isolates were reported previously (28).
The MLST analysis was performed according to a standard procedure (16). DNA sequences were analyzed with the SeqMan software in the Lasergene package (DNAStar Inc., Madison, WI). For identification of allele numbers and sequence types (STs), sequence data were compared to the MLST database (www.mlst.net). New alleles and allelic profiles were reported to the database curator for assignment of ST numbers. eBURST analysis (17) was used to group STs into clonal complexes and to construct a population snapshot, applying a definition by which members of a clonal complex share six of seven MLST loci (17). Multiple-locus variable-number-tandem-repeat analysis (MLVA), based on the determination of the number of repeats in 17 polymorphic loci, was performed as described by others (30; http://www.mlva.eu).
Fingerprints of the pbp1a, pbp2b, and pbp2x genes were determined by PCR-restriction fragment length polymorphism (PCR-RFLP) analysis (10, 13, 36) with the HinfI restriction enzyme (Fermentas, Vilnius, Lithuania). The resulting RFLP patterns were compared with those obtained for the reference strain Spain9V-ST156 (ATCC 700671). Selected pbp genes were sequenced as described in a previous study (27). A part of the murM gene was amplified and sequenced as described by Filipe et al. (19).
Genes belonging to the rlrA islet were detected by PCR as proposed previously (45). The PspA families were identified by PCR of the pspA genes with primers LSM12 and SKH63 (family 1) and primers LSM12 and SKH52 (family 2), as reported by Vela Coral et al. (54). The clade-determining region (CDR) of pspA was sequenced using a family-specific primer, SKH63 or SKH52, and the deduced amino acid sequences were clustered with reference sequences for the described PspA clades (25), using the ClustalW algorithm in the MagAlign software of the Lasergene package.
The two-tailed Fisher exact test (FET) was used to evaluate the change in the proportion of penicillin-resistant isolates among all PNSP isolates, the differences in PNSP frequency between invasive and RTI isolates, and association of the Spain9V-ST156 clonal complex with invasive disease. Differences in penicillin MIC distributions between the groups with the murMA and murMB genes were verified by the t test. The diversity index (DI) was calculated as described by Grundmann et al. (22); the adjusted Rand and Wallace coefficients were determined according to methods used in other studies (26, 55), using an online tool available at http://www.comparingpartitions.info/. Sequences of pbp2b, pbp2x, murM, and pspA were compared with those deposited in the GenBank database, using BLAST software (www.ncbi.nlm.nih.gov/blast/Blast.cgi).
New sequences determined in this study have been assigned the following GenBank numbers: EU876858 (partial pbp2x sequence of isolate 226/2005), EU876859 (partial pbp2x sequence of isolate 3671/2005), FJ744155 to FJ744158 (murMB13 to -16 alleles), and EU542617 (partial pspA sequence of isolate 2748-03).
Altogether, 118 PNSP isolates (MICs, ≥0.12 μg/ml) were identified among all 564 S. pneumoniae isolates from the study period, showing significant differences in frequency between 2003 and 2005. Namely, this frequency increased from 15.5% in 2003 and 14.4% in 2004 to 30.2% in 2005. PNSP isolates were isolated in 46 centers located in 32 towns and were derived mainly from sputum samples (49 isolates [41.5%]), bronchoalveolar lavage samples (24 isolates [20.3%]), cerebrospinal fluid (15 isolates [12.7%]), the blood of patients with pneumonia (11 isolates [9.3%]), and the blood of patients with meningitis (8 isolates [6.8%]). The remaining isolates were from pleural fluid, eye, ear, throat, nose, and a wound (one to three isolates per specimen). Thirty-seven invasive isolates accounted for 31.4% of the PNSP isolates. PNSP isolates were more frequent among RTI isolates (23.2%) than among invasive isolates (17.3%); however, the difference was not statistically significant (FET; P = 0.14).
The MIC of penicillin was ≥2 μg/ml (94 isolates [79.7%]) for the majority of PNSP isolates; the MICs for the remaining 24 isolates (20.3%) ranged from 0.12 to 1 μg/ml. According to the recent CLSI breakpoints for penicillin in parenteral use (9), which consider nonmeningeal isolates with MICs of 4 and ≥8 μg/ml to be intermediate and resistant, respectively, 24 RTI isolates (5.3% of nonmeningeal isolates) were intermediate and 3 isolates (0.67%) were resistant to penicillin. With the breakpoint of 0.12 μg/ml for meningeal isolates, all such PNSP isolates were resistant, constituting 18.0% of all meningeal isolates from the study period. The distribution of MICs for other β-lactams with respect to meningeal and nonmeningeal isolates is shown in Fig. Fig.1.1. While amoxicillin, ceftriaxone, and cefepime retained relatively good activity against nonmeningeal isolates, most meningeal isolates showed nonsusceptibility to the last two compounds. The majority of nonmeningeal PNSP isolates were nonsusceptible to meropenem.
The PNSP group contained isolates of seven different serotypes: serotypes 14 (40 isolates [33.9% of the PNSP isolates), 9V (24 isolates [20.3%]), 23F (23 isolates [19.5%]), 19F (14 isolates [11.9%]), 6B (12 isolates [10.2%]), 19A (4 isolates [3.4%]), and 6A (1 isolate). Fifteen isolates (11.5% of the PNSP isolates) were obtained from children below 2 years of age. Serotypes in this group included serotypes 14 (8 isolates), 6B, 19F, 23F (2 isolates each), and 9V (1 isolate), thus corresponding to complete coverage by the 7-valent conjugated vaccine (PCV7).
Thirty-one different allelic profiles were observed in the PNSP group (Table (Table1),1), among which four were new (STs 2336, 2337, 2338, and 2955). Twenty STs with 82 isolates were grouped into five clonal complexes, whereas 11 STs with 36 isolates were singletons (Fig. (Fig.2).2). All of the clonal complexes and three singletons (23 STs with 107 isolates [90.7% of all PNSP isolates]) belonged to the pandemic PMEN clones. The biggest complex, associated with the Spain9V-ST156 clone, included 8 STs and 56 isolates (47.5% of the PNSP isolates). Other international clones or their variants included Spain23F-ST81, Spain6B-ST90, England14-ST9, Taiwan19F-ST236, Taiwan23F-ST242, Poland23F-ST173, and Poland6B-ST315 (Table (Table1).1). For the whole group, the MLST-based DI of the PNSP isolates was equal to 89.38% (confidence interval [CI], 86.02% to 92.74%). DIs for the RTI isolates and invasive isolates differed, although not significantly, and were equal to 91.33% (CI, 88.17% to 94.48%) and 83.33% (CI, 74.72% to 91.95%), respectively.
Over half of the isolates (67 isolates [56.8% of the PNSP isolates]) harbored the wild-type murMA allele (Table (Table1).1). This allele was characteristic of Spain23F-ST81, Poland6B-ST315, and Taiwan19F-ST236 isolates, the majority of England14-ST9 isolates (STs 15, 1492, and 1815), and almost all sporadic clones. murMA was also present in a part of the Spain9V-ST156 clonal complex, including STs 557 and 1569 and all but one isolate of ST156 (Table (Table1).1). Six different murMB alleles were found in the remaining 51 isolates, including murMB6, known previously (19), and five new alleles (with proposed designations murMB12 to -16). The murMB12 allele, characteristic of Spain6B-ST90 and the sporadic clone ST319, shared 100% identity with that of S. pneumoniae strain 02J1095 (GenBank accession no. DQ056830). The newly observed murMB13 allele occurred in the Taiwan23F-ST424 complex, murMB14 occurred in Poland23F-ST173, and murMB15 and murMB16 occurred among the Spain9V-ST156 isolates (Table (Table1).1). MICs of penicillin were significantly higher for isolates that carried murMB alleles than for isolates with murMA alleles (t test; P < 0.0001), and isolates with high-level resistance (MIC = 8 μg/ml) occurred only in the first group (Table (Table11).
The results of the detailed analysis of the Spain9V-ST156 isolates are described below and shown in Table Table11.
Among 56 isolates in this group (Table (Table1),1), 26 isolates were invasive (46.4%), constituting 70.3% of all invasive PNSP isolates and being significantly overrepresented among these (FET; P = 0.0013). Within the complex, isolates of serotype 14 predominated (59.0%), followed by the original serotype 9V (39.3%) and another capsular variant, 19F (a single isolate of ST557). Serotype 14 was characteristic for all of the isolates of STs 143, 144, 790, 1569, and 2336 and for 25% of the ST156 isolates, while serotype 9V was associated with the remaining ST156 isolates and a single ST2306 isolate. The MLST-based DI for this clonal complex equaled 60.45% (CI, 52.78% to 68.13%).
MLVA typing yielded 31 different profiles, MLVA types (MTs) 684 to 714 (profiles are available at the MLVA database [http://www.mlva.eu]), none of which had been reported previously. Only the Spneu15, Spneu26, and Spneu41 loci did not show any polymorphism in the number of repeats, while for the most polymorphic locus, Spneu19, seven variants were observed. The majority of the isolates (n = 52) repeatedly did not yield any product for the Spneu36 locus within the trzA gene. The eBURST analysis grouped the isolates into four clonal complexes and five singletons (Fig. (Fig.3).3). While no MT was associated with more than a single ST, two main STs, 143 and 156, contained several MTs each. Except for a small complex composed of MTs 700 and 703, the MLVA complexes were associated with 2 or 3 different STs. The DI for MLVA was equal to 90.39% (CI, 84.82% to 95.96%), confirming the much higher resolution of this method than that of MLST. Owing to this fact, the two methods correlated poorly (adjusted Rand coefficient, 0.2796); however, the Wallace coefficient for MLVA versus MLST was equal to 1.0, i.e., a given MT predicted an ST with 100% accuracy.
All of the isolates showed the same PCR-RFLP pattern for the pbp1a gene as that produced by the reference Spain9V-ST156 strain ATCC 700671 (Table (Table1).1). Almost all isolates also shared the pbp2b pattern of the ATCC strain (pbp2b pattern 1), except for a single isolate of ST2306 (pbp2b pattern 2). Similarly, this isolate (pbp2x pattern 4), together with three isolates (pbp2x pattern 3) and one isolate (pbp2x pattern 2) of ST143, were the only ones possessing specific PCR-RFLP patterns for pbp2x. Interestingly, the amoxicillin MIC (8 μg/ml) was even higher than the penicillin MIC (4 μg/ml) for the ST2306 isolate. Its pbp2b pattern was identical to that of S. pneumoniae strain S172, from China (GenBank accession no. DQ438992), while its pbp2x pattern was the same as that for isolates 685.2FR99 (GenBank accession no. EF501684) and 595.4IT95 (GenBank accession no. EF501632), from France and Italy, respectively (51). The pbp2b sequence was 87.2% and 91.5% identical to the corresponding pbp2b parts in the penicillin-susceptible reference strain R6 and the Spain9V-ST156 ATCC strain, respectively, while the deduced PBP2B amino acid sequence was 91.4% and 93.5% identical to those in the respective strains. The PBP2B sequence of this isolate contained the same significant T446A change in the 443SSNT446 motif as the parental Spain9V-ST156 reference strain and, additionally, 10 changes in the 538-582 region (N538D, D561E, 565QLQPT569 to 565AIDTK569, M571I, D578E, and S582A) and 10 changes in the 592-641 region (A592S, G597P, N606D, L609T, A619G, D625G, Q628E, T630N, S640T, and D641E). The pbp2x gene of the ST2306 isolate was 83.0% and 97.0% identical to its homologs in the reference strains R6 and Spain9V-ST156, respectively; at the amino acid sequence level, the identity was 87.5% and 98.2%, respectively. The two aforementioned specific pbp2x variants of several ST143 isolates correlated with resistance to ceftriaxone (MICs, 4 to 8 μg/ml) and had new sequences. The sequences shared 87.4% (GenBank accession no. EU876858) and 84.6% (GenBank accession no. EU876859) identity with pbp2x of the R6 strain and 94.9% and 96.2% identity with the reference Spain9V-ST156 strain, and they were 95.0% identical to each other. The amino acid sequences were 89.6% and 88.3% identical to PBP2X of R6 and 96.5% and 97.6% identical to PBP2X of the Spain9V-ST156 ATCC strain. All of the PBP2X variants of the Spain9V-ST156 isolates had the T338A change in the 337STMK340 motif and an L546V change adjacent to the 547KSG549 motif, as well as I371T, R384G, and N605T changes, which are important for resistance (8). Additionally, the new PBP2X variants of the ceftriaxone-resistant isolates contained the substitution M339F in the 337STMK340 motif.
Three different alleles of the murM gene were found among the isolates of the Spain9V-ST156 complex (Table (Table1).1). Isolates of STs 557 and 1569 and all but one isolate of ST156 possessed the murMA allele, while all isolates of STs 143, 144, 790, and 2336 and a single ST156 isolate carried the new allele murMB15. The single isolate of ST2306 had the new allele murMB16.
Sequencing of the CDR in pspA revealed that the majority of the isolates harbored the same variant as the Canadian isolate AC122 (GenBank accession no. AF071818) (25), which is characteristic of family 2, clade 3. The single isolate of ST2306 possessed the pspA variant originally identified in strain DBL5, of unknown origin (GenBank accession no. AF071810), representing family 1 and clade 2 (25). Finally, a single isolate of ST2336 had a new variant of pspA, belonging to family 2, clade 4. The rlrA, rrgA, rrgB, rrgC, srtB, and srtD genes were detected in all but one isolate, indicating the presence of the intact pilus islet. In the case of the single isolate of ST156 and serotype 14, no product was obtained with primers SP0462F and SP0462R (45), specific for the 5′ part of the rrgA gene. Moreover, PCR with primers SP0461R and SP0462RX (45), annealing upstream of rlrA and in the 3′ part of rrgA, respectively, was also negative, suggesting an insertion in the 5′ part of rrgA.
In our previous study of 131 PNSP isolates obtained from 1998 to 2002, we observed an increase in frequency of penicillin nonsusceptibility (MIC, ≥0.12 μg/ml) among pneumococci in Poland, from 8.1% in 1998 to 20.3% in 2002 (44). This trend persisted in the following years, reaching a PNSP prevalence rate of 30.2% in 2005. Concomitantly, the fraction of resistant strains (MICs, ≥2 μg/ml) among PNSP isolates grew significantly, from 67.9% in 1998 to 2002 to 79.7% (FET; P = 0.04) during 2003 to 2005. We had previously observed a correlation between β-lactam consumption and the increase in the PNSP level (44), similar to the case for several European countries for which the consumption data from 1998 to 2004 were compared to resistance rates for 2004 to 2005 (21, 43). However, recently in Poland the consumption of β-lactams showed a decreasing trend, from 13.44 to 9.74 defined daily doses (DDD)/1,000 inhabitants/day in 2001 and 2004, respectively (www.esac.ua.ac.be). The consumption of macrolides, which are other selectors of PNSP due to frequent coresistance to these two classes of antimicrobials (21), showed a moderate increase, from 1.54 DDD/1,000 inhabitants/day in 1998 to 2.52 DDD/1,000 inhabitants/day in 2004 (www.esac.ua.ac.be); the latter value does not differ much from the average European consumption (2.10 DDD/1,000 inhabitants/day in 2004). Therefore, other factors have probably been driving the observed increase in PNSP in the country.
While modifications in the pbp1a, pbp2b, and pbp2x genes play a crucial role in the loss of penicillin susceptibility in clinical S. pneumoniae strains, replacement of the wild-type murMA allele by one of the murMB variants appears to reduce the biological cost associated with the production of altered PBPs and is usually observed in isolates with elevated MICs of penicillin and other β-lactams (7, 12, 50). An association of murMA replacement by several murMB variants with the increased level of penicillin nonsusceptibility was also found in this study. The fact that the majority of murMB alleles observed here were new indicates that the variability of this gene has been explored only partially so far. Moreover, some of these murMB alleles (murMB12, -15, and -16) were identified in representatives of the Spain6B-ST90 and Spain9V-ST156 clonal complexes, instead of the murMB5, murMB6, and murMB10 alleles described previously for these complexes (7, 12). Since the new alleles differed by several nucleotides from the original ones (data not shown), they most likely arose through recombination. Apparent selective pressure on this gene further supports the hypothesis of the importance of modified murM for pneumococci with reduced susceptibility to β-lactams.
Similar to our previous results (44), MLST of the newer PNSP isolates showed the vast predominance of PMEN clones. While some of them, like Poland23F-ST173 and Poland6B-ST315, decreased with time, the Spain9V-ST156 complex continued its spread, starting in Poland in the early 2000s. For the first time, isolates of Taiwan19F-ST236 and penicillin-intermediate and -resistant variants of England14-ST9 were observed. The latter clone originally showed full susceptibility to penicillin (34). Spain23F-ST81, Spain6B-ST90, and Taiwan23F-ST242 have persisted in Poland since the second half of the 1990s (44). However, Taiwan23F-ST242 now showed elevated penicillin MICs (4 to 8 μg/ml), similar to other recent observations (50). Some of the previously sporadic PNSP strains (44) were found again, including ST276, ST319, and ST1027 strains. The penicillin-intermediate Sweden15A-ST63 clone, which was present earlier, was not recovered this time. Thus, constant surveillance for 10 years allowed us to follow the dynamics of the PNSP subpopulation in Poland and revealed the phenomena of clone replacement, introduction, and long-term persistence.
The role of international clones in the spread of antimicrobial resistance in S. pneumoniae is well established (29), and among such clones, Spain9V-ST156 has already achieved a global distribution (29, 45). This clone most likely emerged in Spain in the 1980s (18, 57), having evolved from a penicillin-susceptible strain of ST162 (45). In Poland, the Spain9V-ST156 complex disseminated during a short period (from 22.0% of PNSP isolates in 1998 to 2002 to 47.5% of these isolates in 2003 to 2005) and thus greatly contributed to the elevated PNSP prevalence, especially among invasive isolates. Therefore, we decided to analyze this clone in more detail, aiming at the description of its possible diversification in terms of presumably neutral markers (the MLST and MLVA loci) as well as those that are under selective pressure (capsular type, presence of pili, and the polymorphic genes pbp, murM, and pspA).
This study provides the first detailed report on congruence between the MLST and MLVA results for an S. pneumoniae clonal complex. In the case of Spain9V-ST156, MLVA was much more discriminatory and showed several novel types. The number of loci used in MLVA is larger (15) than that for MLST (7); the latter method, however, is able to index a single nucleotide change within analyzed genes. The analysis of other clones will show whether the rapid evolution of the repeat numbers within the MLVA loci is characteristic of the clone or typical of these loci. In general, MLVA showed a good congruence with MLST, which indicates its potential utility for fast identification of pneumococcal clones.
The majority of isolates of the Spain9V-ST156 complex had the same PCR-RFLP profile for the pbp1a, pbp2b, and pbp2x genes, which was in common with that of the reference strain. Generally, higher penicillin MICs were characteristic of Polish isolates of the complex than those of isolates from areas with a low PNSP prevalence, such as Norway or Scotland (47, 48). This might suggest the importance of specific murMB alleles found in some isolates of this clone, the possible presence of mutations in the pbp genes that were undetectable by PCR-RFLP analysis, or the potential role of other factors, e.g., regulatory systems such as the CiaRH system (33), that may vary among particular strains. Increased MICs of other β-lactams (amoxicillin and ceftriaxone) observed for a few isolates paralleled divergence in their pbp patterns. Sequencing of these genes revealed that resistance to cephalosporins in ST143 isolates was most likely associated with acquisition of a pbp2x gene with an M339F change (8), in addition to the T338A change in the 337STMK340 motif which is typical for members of family 1 of mosaic PBP2X (56). Resistance to amoxicillin in the ST2306 isolate was probably mediated by 10 additional changes in the 592-641 region surrounding the 615KTG617 drug-binding site in PBP2B (7, 15, 31), including the crucial A619G substitution (31). Other studies (7) indicated an association between the single N538D mutation in the 538-582 region in PBP2B and the presence of a murMB allele in a strain. The strain analyzed here harbored the new murMB16 allele and the N538D mutation; however, nine other changes were present in the 538-582 region that had previously been described as typical for strains with murMA (7).
Among the Spain9V-ST156 isolates, we found several isolates of serotype 14 and a single isolate of serotype 19F, apart from the basic serotype 9V. Frequent serotype changes have been observed for this complex, including serotypes 3, 4, 6B, 9A, 11, 14, 18C, 19A, 19F, and 23F (51). The serotype 14 variant which now contributes significantly to the complex evolved more than once (2, 10). In contrast to 9V, this serotype is preferentially associated with invasive infections (6) and may be one of the reasons for the overrepresentation of the clone among invasive isolates in the study. Another possible factor is the presence of the pilus, whose genes are well conserved among the isolates. It has been shown that a piliated clinical isolate of Spain9V-ST156 was more virulent and evoked an elevated immune response in comparison to its nonpiliated derivative in a murine model (5). The presence of the pilus is believed to be an important factor behind the recent epidemic success of this clone (45). The pilus, by improving the adherence of bacteria to human epithelial cells, may facilitate nasopharyngeal colonization as well (5), which would ease DNA acquisition and thus promote evolution of the clone. It is difficult to speculate on the adaptive value of the PspA variant associated with the clone, but it was relatively well conserved in the studied group, and the predominant variant was the same as that observed by others (25). In general, the likely neutral markers, such as the MLST and MLVA loci, evolved faster than the resistance and virulence determinants, suggesting that the particular variants of the latter markers might confer a selective advantage for Spain9V-ST156.
Implementation of the PCV7 vaccine, licensed in the United States in 2000 and in the European Union in 2001, has led to a reduction of penicillin resistance in local pneumococcal populations (35, 52). A decrease in vaccine serotypes, such as 6B, 9V, 14, and 23F, has been observed among PNSP isolates, followed by an increase in their genetic variability and in the expansion of nonvaccine serotypes, such as 19A and 35B (42). In Poland, conjugated antipneumococcal vaccines are reimbursed only for children from certain risk groups, and the PCV7 uptake in the early 2000s was minimal, not exerting any pressure on the S. pneumoniae population. This is consistent with the observed high clonality of Polish PNSP isolates, as well as the low prevalence of serotype 19A and the absence of serotype 35B. It is known that some well-adapted clones spread efficiently, even if antimicrobial pressure is decreasing, low, or nonexistent (38, 49), and Spain9V-ST156 is a good example of this. Considering this, as well as the observed evolution of circulating strains toward high penicillin resistance and resistance to other β-lactams, further surveillance is indispensable.
This study was partially financed by a grant from the Polish Committee for Scientific Research (N401 178 32/3572) and by the EU grant HEALTH-F3-2009-223111. We acknowledge the use of the pneumococcal MLST database, which is located at the Imperial College, London, United Kingdom, and is funded by the Wellcome Trust.
We thank Cara Horowitz for a critical reading of the manuscript.
Published ahead of print on 1 March 2010.