An Internationally Spread Clone of Streptococcus pneumoniae Evolves from Low-Level to Higher-Level Penicillin Resistance by Uptake of Penicillin-Binding Protein Gene Fragments from Nonencapsulated Pneumococci

Abstract

Low-level penicillin resistance in an international Streptococcus pneumoniae serotype 19F clone emerging in Switzerland was characterized by mutations in the penicillin-binding protein PBP2x. Some isolates of this clone had evolved to higher resistance levels (penicillin MICs of 0.094 and 1 μg/ml), probably by acquisition of pbp2x fragments from local nonencapsulated pneumococci.

Streptococcus pneumoniae acquires β-lactam resistance by importation of foreign DNA sequences which code for low-affinity penicillin binding proteins (PBPs) (6, 12). Low-affinity PBPs probably arise in commensal oral streptococci by accumulation of point mutations (5, 25). High MICs of penicillin for S. pneumoniae require a combination of altered PBPs PBP1a, -2b, and/or -2x (5, 20, 26). Low-level resistance is mainly based on alterations in PBP2x and can be detected by the oxacillin disk test (4). Whether such low-level resistance provides a selective advantage in vivo is unknown at present. Early surveillance studies suggested a gradual increase of isolates with low-level penicillin resistance (MICs of ≥0.03 and ≤0.06 μg/ml) between the 1980s and 1990s in the United States (3). In this study, we demonstrate that low-level penicillin resistance of an international serotype 19F clone which emerged in Switzerland in 1998 (18, 23) was mediated by alterations in the pbp2x gene and that isolates of this clone evolved to higher MICs and acquired pbp2x gene fragments shared with nonencapsulated pneumococci.

(Some of the data in this study were presented at the 4th International Symposium on Pneumococci and Pneumococcal Diseases, Helsinki, Finland, May 2004.)

Nasopharyngeal isolates (n = 2,111) of Streptococcus pneumoniae collected between 1998 and 2002 in a nationwide Sentinel study were tested against oxacillin (1-μg disk), erythromycin, and cotrimoxazol by the disk diffusion method. Penicillin MICs were determined by the E-test method (AB Biodisk; DMD AG, Arlesheim, Switzerland). Capsular serotypes were determined, and pulsed-field gel electrophoresis (PFGE) was conducted as previously described (18). The present study describes the analysis of all 108 serotype 19F isolates with an oxacillin disk diameter of <20 mm. In addition, the analysis of PBPs included five nonencapsulated pneumococcal isolates reported previously (13). Twenty-two (20.4%) of the 108 serotype 19F isolates were further analyzed by multilocus sequence typing (MLST) (10, 15).

Restriction fragment length polymorphism (RFLP) analysis of pbp2x was done as previously described (24, 28). Fragments of the pbp genes pbp1a, -2b, and -2x were amplified in 100-μl reaction mixtures with 16 μl of 0.2 mM deoxynucleoside triphosphates, 5 μl (10 μl for pbp1a) each of forward and reverse 1 μM primers, 10 μl of 1.5 mM MgCl-DNA polymerase buffer (Roche Molecular Biochemicals, Rotkreuz, Switzerland), 1 μl of 1 mM MgCl, 57.6 μl (47.6 μl for pbp1a) of water, 2 U of Taq DNA polymerase (Roche Molecular Biochemicals), and 200 ng of DNA in 5 μl of Tris-EDTA buffer by using published primers and primers designed for this study (2, 7, 19). Primer pbp2x_down was adapted from the work of Muñoz et al. (19): 5′-GGCGAATTCCAGCACTGATGGAAATAAACATATTA-3′. PCR conditions were the following: initial incubation at 94°C for 5 min, followed by 30 cycles of 95°C for 1 min, 50°C for 2 min, and 70°C for 6 min. PCR products were purified using a QIAGEN PCR purification kit (QIAGEN AG, Basel, Switzerland). Alternatively, pbp1a was amplified using primers pbp1a_f3 (5′-GTAAAACAAGATGAACAAACCAACGA-3′) and pbp1a_b2185 (5′-TTATGGTTGTGCTGGTTGAGGAT-3′) with the following conditions: initial incubation at 94°C for 5 min, followed by 40 cycles of 95°C for 1 min, 54°C for 1 min, and 70°C for 3 min, and a final elongation at 70°C for 10 min. The recipient laboratory strain R6 was transformed with 1 μg of PCR products of pbp 1a, 2b, or 2x (28). Transformed cells (100 μl) were plated on cation-adjusted Mueller-Hinton blood agar containing 0.025 μg of benzylpenicillin (Sigma-Aldrich Chemie GmbH, Steinheim, Germany)/ml. PCR products of pbp2x, pbp2b, and pbp1a were sequenced by Microsynth GmbH (Balgach, Switzerland).

Molecular typing by PFGE of the 108 isolates revealed one major cluster comprising 70 (64.8%) isolates and two clusters comprising 5 and 3 isolates, respectively. The remaining 30 isolates each had a unique PFGE banding pattern. Thirteen of the seventy isolates belonging to the major cluster were typed by MLST. Eleven had the international MLST sequence type 177 (ST177) (allelic pattern: aroE, 7; gdh, 14; gki, 4; recP, 12; spi, 1; xpt, 1; ddl, 14 [http://www.mlst.net]), and two were ST179, which differs from ST177 at the gki locus only (allele 40) (18). ST177 is the genotype of the recognized global clone Portugal^19F-21, which has been isolated in Portugal, Spain, and Iceland (22-24). The clonal complex comprising ST177 and ST179 will be referred to as PFGE cluster H from now on (22). For study strains belonging to PFGE cluster H, penicillin MICs were 0.023 to 1.0 μg/ml. Low-level penicillin resistance (MIC of ≥0.03 and ≤0.06 μg/ml) was frequent (36 of 70 isolates; 51.4%), as described earlier (18). Penicillin MICs for a further 33 (47.1%) isolates in PFGE cluster H were between >0.06 and ≤0.125 μg/ml, and for 1 isolate, the MIC was higher at 1 μg/ml (isolate 304.80) (Table 1). The MIC range among the 38 isolates not belonging to PFGE cluster H was 0.012 to 2.0 μg/ml, but only 3 isolates (7.9%) showed low-level penicillin resistance.

TABLE 1.

MLST and PBP analysis of 14 pneumococcal study strains with low- or intermediate-level penicillin resistance belonging to international serotype 19F PFGE clone H

Isolate	MLST sequence type	MICa (μg/ml)	RFLPb			Sequencec
			pbp2x	pbp2b	pbp1a	pbp2x	pbp2b	pbp1a
104.33	177	0.047	A	R6	R6	A	R6	R6
110.36	177	0.047	A	R6	R6	A	R6	R6
109.07	177	0.047	A	−	−	A	−	−
109.34	177	0.047	A	−	−	A	−	−
201.33	177	0.047	A	−	−	A	−	−
301.49	177	0.047	A	−	−	−	−	−
104.38	177	0.064	A	−	−	A	−	−
108.76	177	0.064	A	R6	R6	A	R6	R6
107.81	177	0.064	A	R6	R6	A	R6	R6
209.74	177	0.064	A	−	−	−	−	−
302.41	177	0.064	A	−	−	−	−	−
202.47	177	0.094	B	R6	A	B	R6	A
207.41	179	0.064	A	R6	R6	−	−	−
304.80	179	1.0	C	A	B	C	A	B

^aPenicillin MIC determined by E-test; values represent means for at least three independent measurements.

^bRFLP of pbp1a, -2b, and -2x, R6 indicates an RFLP pattern identical to that of the laboratory strain R6; A, B, and C describe patterns that differ from that of R6; −, no RFLP was obtained.

^cComparison of sequence fragments of pbp1a, -2b, and -2x; R6 indicates close homology to the laboratory strain R6 (>99.10%), as laboratory strain R6; A, B, and C describe sequence patterns with sequence homology to R6 of < 96.5%; −, no sequence was obtained.

The gradual increase of isolates with low-level resistance in early surveillance studies (3) suggested that low-level penicillin resistance may mediate a selective advantage. This hypothesis is further supported by the epidemic emergence of PFGE cluster H in the West region of Switzerland in the winter season 1998-1999. Its prevalence (among all pneumococcal isolates) was 2.68% in 1998 and 8.16% in 1999 (18). Thereafter it declined to 1.86% in 2001 and 3.84% in 2002 (chi-square test for trend since 1999; P = 0.04). With the decrease in the West region, cluster H appeared gradually in the East region with a prevalence of 0% in 1998, 2.62% in 1999, 3.09% in 2001, and 5.66% in 2002 (chi-square test for trend, P = 0.006). As described earlier, a higher level of antibiotic usage in the West region may have promoted the epidemic emergence of PFGE cluster H (18).

Dowson et al. (4) showed that various alterations in PBP2x can induce low-level penicillin resistance. In this study, pbp1a, -2b, and -2x were analyzed for a selection of isolates from PFGE cluster H using RFLP and sequencing. Perfect homology (100%) of pbp2x to a previously reported strain of PFGE cluster H (24) was found for 12 of 14 isolates for which penicillin MICs were 0.047 to 0.064 μg/ml (Table 1). Strain 202.47 (penicillin MIC of 0.094 μg/ml) and strain 304.80 (penicillin MIC of 1.0 μg/ml) showed different sequence patterns for pbp2x (Tables 1 and 2). The Thr₃₃₈Ala and Leu₅₄₆Val shift in pbp2x of strain 304.80 has been associated with an elevated MIC of penicillin (Table 2) (1, 20). The pbp2b sequences of six isolates in PFGE cluster H (104.33, 110.36, 107.81, 108.76, and 207.41 [penicillin MICs, 0.047 to 0.064 μg/ml] and isolate 202.47 [penicillin MIC of 0.094]) were 99.93% homologous to the corresponding sequence of R6 (Table 1). Isolate 304.80 (penicillin MIC of 1.0 μg/ml) exhibited a Thr₄₄₅Ala mutation in SSN motif II and the mutations Glu₄₇₅Gly and Thr₄₈₈Ala (Table 2) (5, 20). The pbp1a sequence of five isolates in PFGE cluster H (104.33, 110.36, 107.81, 108.76, and 207.41 [penicillin MICs, 0.047 to 0.064 μg/ml]) showed ≥99.14% homology to the corresponding sequence in R6 (Table 1). Strain 202.47 (penicillin MIC of 0.094) and strain 304.80 (penicillin MIC of 1.0 μg/ml) exhibited the mutations NTGY (574 to 577) (26, 27), and in strain 304.80 also a Thr₃₇₁Ala shift was present (26) (Table 2).

TABLE 2.

Sequence comparison of PBP1a, -2x, and -2b from three pneumococcal strains with low- or intermediate-level penicillin resistance

Strain	Gene	Comparison with strain R6					Comparison with wild-type strainsa
		Nucleotidesb		Change in motifc			Nucleotidesb		Straind; serotype (reference)
		Total (bp)	% Homology	I	II	III	Total (bp)	% Homology
107.81	pbp2x	1994	94.93				1,979	100	AF446240; 19F (24)
107.81	pbp2b	1408	99.93
107.81	pbp1a	2093	99.62				2,091	99.76	X67872; 13 (17)
202.47	pbp2x	1980	92.42	T338S			1,969	97.46	AF446237; 23F (24)
							1,981	99.03	205.09; nontypeable
202.47	pbp1a	2129	96.67				2,205	98.63	AF210752; 23F (11)
202.47	pbp2b	1407	99.93
304.80	pbp2x	1986	85.65	T338A		L546V	1,986	98.70	AJ238580
							1,987	98.68	AF210759; 23F (11)
							1,957	99.80	106.44; nontypeable (13)
304.80	pbp1a	2121	88.31	T371A			2,120	100	AF210749; 23F (11)
304.80	pbp2b	1408	92.33		T445A		1,408	96.24	AJ243053; 23F (9)

^aDescribes the sequence with the closest homology by either BLAST analysis using GenBank or comparison with strains of this study.

^bTotal, length of the fragment analyzed; % homology represents homology with R6.

^cChanges in the amino acid motifs that form the active penicillin-binding site cavities of the PBPs in S. pneumoniae.

^dGenBank accession number or strain number from this study, as applicable.

Increased MICs of penicillin were obtained upon transformation of R6 with PCR fragments of pbp2x from isolates 107.81 and 202.47 (transformation efficiencies, 0.066 and 0.07%) but not by transformation with PCR fragments of pbp1a and -2b of isolate 107.81 and pbp2b of isolate 202.47. Second-step transformation of R6_{Tpbp2x_202.47} with pbp1a of isolate 202.47 increased the MIC of penicillin from 0.064 to 0.094 μg/ml and the MIC of cefaclor from 1.0 to 2.0 μg/ml, which corresponds to the MICs for the donor strain. Since the products of pbp1a, -2b, and -2x of isolate 304.80 contained amino acid changes already known to confer β-lactam resistance, no transformation experiments were performed.

PBP2x alterations mediating low-level penicillin resistance are probably the most common first step in the development of penicillin resistance in pneumococcus (5, 8, 20). Low-level resistance may play a role in the evolution of higher-level resistance. It may select for compensatory mutations which are essential for efficient growth in the presence of multiple altered PBPs (21). This study supports this hypothesis. For two strains (202.47 and 304.80) belonging to PFGE cluster H, higher MICs of penicillin were achieved by uptake of new PBP gene fragments. To our knowledge this is the first in vivo documentation of a stepwise evolution of isolates with low-level resistance to higher resistance for a pneumococcal clone. Interestingly, the pbp2x sequences of isolates 202.47 and 304.80 showed higher homology to the sequences of nonencapsulated isolates (205.09 and 106.44, respectively) than to the R6 sequence (Tables 1 and 2; Fig. 1). For both pbp2x sequences shared by the 19F and nonencapsulated strains, BLAST analysis yielded no other matches with similarly high homology (>99%) in GenBank. In contrast, all the other PBP sequences found in the PFGE type H clone had close BLAST matches with >99% sequence homology. This suggests that nonencapsulated pneumococci may serve as donors of resistance genes to encapsulated S. pneumoniae (29). There is ample evidence that commensal streptococci are the source of mutated PBPs (5, 25). In fact, BLAST analysis revealed high homology (98.53%) between the distal fragments of the transpeptidase region of isolate 304.80, those of the nonencapsulated isolate 106.44, and The Institute for Genomic Research's Streptococcus mitis genome sequence (www.tigr.org). Therefore, S. mitis was probably involved at some point during the generation of this mosaic structure. The nonencapsulated strains may have gathered DNA pieces from several sources, such as commensal streptococci, before subsequent transmission to the 19F strains. Nonencapsulated pneumococci may thus act as a shuttle vector of new resistance alleles for encapsulated pneumococci. Whatmore et al. already speculated that atypical or nonencapsulated pneumococci may act as a source for new resistance-mediating alleles (29). This hypothesis is further supported by the findings of Majewski et al. (16), who showed that the efficiency of transformation increases with decreasing phylogenetic difference between donor and recipient strain. Nonencapsulated pneumococci are, therefore, ideal candidates for the transfer of streptococcal resistance genes to encapsulated pneumococci (13, 29).

FIG. 1.

The mosaicism of pbp2x sequence fragments (base pairs 514 through 2494, numbering according to the work of Laible et al. [14]) for three study strains of serotype 19F (107.81, 202.47, and 304.80) and two nonencapsulated strains (205.09 and 106.44) with low or intermediate penicillin resistance, a previously reported 19F strain (GenBank accession number AF446240) (24), and the S. mitis genome sequence (The Institute for Genomic Research [www.tigr.org]) compared to laboratory strain R6. STMK, SSN, and KSG, conserved motifs that form the active penicillin-binding site cavities of the PBPs in S. pneumoniae, are indicated as vertical lines in the R6 sequence. Shaded boxes refer to sequence blocks with >10% nucleotide divergence from R6. Identical shading within sequence blocks indicate close homology. The number above each sequence block indicates the percent nucleotide divergence from the corresponding sequence in R6.