Intraclonal variations among Streptococcus pneumoniae isolates influence the likelihood of invasive disease in children.

BACKGROUND: Pneumococcal serotypes are represented by a varying number of clonal lineages with different genetic contents, potentially affecting invasiveness. However, genetic variation within the same genetic lineage may be larger than anticipated.
METHODS: A total of 715 invasive and carriage isolates from children in the same region and during the same period were compared using pulsed-field gel electrophoresis (PFGE) and multilocus sequence typing. Bacterial genome sequencing, functional assays, and in vivo virulence mice studies were performed.
RESULTS: Clonal types of the same serotype but also intraclonal variants within clonal complexes (CCs) showed differences in invasive-disease potential. CC138, a common CC, was divided into several PFGE patterns, partly explained by number, location, and type of temperate bacteriophages. Whole-genome sequencing of 4 CC138 isolates representing PFGE clones with different invasive-disease potentials revealed intraclonal sequence variations of the virulence-associated proteins pneumococcal surface protein A (PspA) and pneumococcal choline-binding protein C (PspC). A carrier isolate lacking PcpA exhibited decreased virulence in mice, and there was a differential binding of human factor H, depending on invasiveness.
CONCLUSIONS: Pneumococcal clonal types but also intraclonal variants exhibited different invasive-disease potentials in children. Intraclonal variants, reflecting different prophage contents, showed differences in major surface antigens. This suggests ongoing immune selection, such as that due to PspC-mediated complement resistance through varied human factor H binding, that may affect invasiveness in children.

Streptococcus pneumoniae is a human-adapted commensal pathogen colonizing the nasopharynx in up to 60% of preschool children [1, 2]. Even though invasive disease is an unlikely event of infection, this organism is estimated to cause the death of >800 000 children <5 years of age worldwide annually [3]. Epidemiological studies have revealed that pneumococci expressing particular capsular serotypes dominate in invasive pneumococcal disease among children, and it is against such pediatric serotypes that polysaccharide-conjugate vaccines (PCVs) have been directed. Comparisons of the prevalence of carriage and invasive isolates within the same geographical area and period demonstrate that particular serotypes (ie, those with a high invasive-disease potential) are associated with a higher potential than others to cause invasive disease [4-6].

It has been difficult to demonstrate whether different clonal lineages of pneumococci, independent of capsular type, differ in their ability to cause invasive disease. One reason is that many serotypes, including those with a high invasive-disease potential, such as serotypes 1, 4, and 7F, belong to only 1 or a few dominating genetic lineages [5, 7, 8]. Among serotypes with a low invasive-disease potential, such as serotype 19F, many different genetic lineages are found, but since the number of isolates representing each lineage usually is low the study has to be sufficiently large to demonstrate statistical significance. The same holds true for genetic lineages that may appear with different capsules, such as clonal complex 156 (CC156) [9].

Because of the highly recombinogenic nature of pneumococci, different isolates belonging to the same genetic lineage, as determined by multilocus sequence typing (MLST), may differ from one another in gene content [10, 11]. Therefore, even within the same genetic lineage, there may be specific subclones or variants with a higher capacity than others to cause invasive disease. Deletions, insertions, and DNA rearrangements are expected to create differences in pulsed-field gel electrophoresis (PFGE) patterns among isolates belonging to the same genetic lineage as determined by MLST.

In this study, we examined a large set of nasopharyngeal carriage and invasive pneumococcal isolates recovered from children from the Stockholm area during 1997–2004. We calculated the odds ratio (OR) for causing invasive disease and found that different clonal types but also intraclonal variants may exhibit different invasive-disease potentials. Pneumococcal isolates representing intraclonal variants of the most prevalent clonal lineage with different disease outcomes in children were characterized genetically and functionally.

MATERIALS AND METHODS

Clinical Isolates Studied

Nasopharyngeal carriage pneumococcal isolates (n = 550) and invasive pneumococcal isolates (n = 165) were recovered from children (<18 years of age) during 1997–2004 from the Stockholm area. The mean age was 2.5 years for children with invasive pneumococcal disease and 3.3 years for healthy carriers.

Serotyping

All 715 isolates were serotyped using gel diffusion with 46 serotype/group antisera obtained from Statens Serum Institut in Copenhagen, Denmark [2].

PFGE Analysis

All invasive and carriage isolates from children were subjected to PFGE adapted from the procedure described by Hermans et al [12]. The clones were defined by using the criteria described by Tenover et al [13].

MLST Analysis

A total of 200 isolates—165 invasive isolates from children and 35 representatives of clonal types found with PFGE among the carriage isolates—were subjected to MLST analysis. MLST was performed according the procedure described by Enright et al [10]. The sequences obtained were submitted to the MLST database (available at: http://www.mlst.net) and assigned a sequence type (ST). The isolates were assigned to different clonal complexes (CCs), as defined on the basis of the entire collection of isolates within the MLST database [14], using an algorithm on the MLST Web site from 2008 [15]. In brief, a CC is defined as a group of STs in a population that shares 6 of 7 alleles with at least 1 other ST in the group. Isolates belonging to a CC may be assumed to have a recent common ancestor [15]. For isolates without MLST data, we assigned a CC to those that belong to a PFGE clone for which the MLST is available for at least 1 isolate.

Polymerase Chain Reaction (PCR) for pcpA and Typing of Prophage DNA

PCR was performed as previously described [9]. The presence of pcpA was determined using specific primers from the coding region (Supplementary Table 1). The presence of phage genes and individual phage typing was determined as described by Romero et al [16].

Lysogenic Bacteriophage Induction by Use of Mitomycin C

The protocol was adapted from the report by Romero et al [16]. In brief, S. pneumoniae isolates were grown in C + Y medium at 37°C until the culture reached an optical density (OD) at 620 nm of 0.2. Mitomycin C was added to a final concentration of 100 ng/mL to induce release of temperate bacteriophages. Growth was monitored in microtiter plates (5 replicate wells/isolate) with Bioscreen equipment (Labsystems, Finland). Measurements were made at OD600 every fifth minute for 16 hours.

Sequencing of pspC and pspA

Sequencing of pspC1 was performed for 4 serotype 6B isolates, and sequencing of pspC2 and pspA was performed for 14 and 12 serotype 6B isolates, respectively, as described previously [17, 18].

Whole-Genome Shotgun Sequencing

Whole-genome shotgun sequencing was performed on 4 serotype 6B isolates (BHN237, BHN427, BHN418, and BHN191). Chromosomal DNA was prepared using a Qiagen DNA prep kit (Genomic DNA Buffer Set and Genomic-tip 100/G). The DNA was subjected to sequencing using a Roche Genome Sequencer FLX (GS FLX). Assembly of the 454 data was performed using Newbler v 2.3 [19]. Genes were predicted using GeneMark.hmm prokaryotic, version 2.6p [20]. To align the contigs to the reference genome, MUMmer [21] was used; to visualize the data, Artemis Comparison Tool was used [22]. Results of this whole-genome shotgun project has been deposited at DDBJ/EMBL/GenBank under accession number ASHQ00000000.

Factor H (FH) Binding by Use of Far Western Blot and Enzyme-Linked Immunosorbent Assay (ELISA) Analyses

Strains were grown in C + Y medium to an optical density at 620 nm of 0.6 and harvested by centrifugation. Bacterial pellets were lysed by boiling in 1 × Nupage sodium dodecyl sulfate (SDS) loading solution (Invitrogen). The samples containing 15 µg of proteins from whole-cell lysate of pneumococci (quantified by Bradford Reagent, Sigma-Aldrich) were separated by SDS polyacrylamide gel electrophoresis and transferred to a polyvinylidene fluoride membrane. The membrane was incubated with 1 µg/mL human FH, followed by incubation with goat anti-FH antibody (1:2000 dilution; Calbiochem) and with rabbit anti-goat immunoglobulin G conjugated with horseradish peroxidase (HRP; 1:8000 dilution; Invitrogen). Detection was performed using the ECL prime Western blotting detection reagent (GE Healthcare) and developed using Gel-doc (Biorad).

FH binding was also quantified using whole-cell ELISA. Microtiter plates were coated overnight with 5 µg/mL FH at 4°C. Subsequent steps were conducted at room temperature for 2 hours each. After washes with 0.05% Tween 20 in phosphate-buffered saline (PBS), nonspecific binding sites were blocked with 1% skim milk in PBS and washed again, and 2 × 106 bacteria were added. Bound bacteria were detected using a 6B specific polyclonal anticapsule (Statens Serum Institut, Denmark) diluted 1:1000 in 0.1% BS and anti-rabbit-HRP. SigmaFast OPD (Sigma) was used for detection, and absorbance was measured at 492 nm. Three different experiments were done in triplicate.

Animal Model

C57BL/6 mice were infected intranasally with 5 × 106 colony-forming units (CFU). Blood samples were collected every 24 hours, and the bacterial load in lungs was determined at the time of euthanization. The experiments were approved by the local ethics committee (Stockholms Norra djurförsöksetiska nämnd).

Statistical Analysis

ORs were calculated to estimate the potential of causing invasive disease for different serotypes and clonal types, as follows: OR = [ad]/[bc], where a the number of invasive clone X isolates, b is the number of carriage clone X isolates, c is the number of invasive clone non-X isolates, and d the number of carriage clone non-X isolates. ORs and 95% confidence intervals (CIs) were calculated using the Fisher method implemented in the Epitools package for statistical software R, version 2.13.0 (available at: http://www.r-project.org). An OR of >1, along with a 95% CI that did not include 1, indicated an increased risk for causing invasive disease.

The data from animal experiments were analyzed using GraphPad Prism 4 software. Survival curves were analyzed using the log-rank test, and blood CFU data were analyzed using the Kruskal-Wallis test.

RESULTS AND DISCUSSION

All 715 invasive and carriage isolates were subjected to serotyping and molecular typing to study genetic relatedness, using PFGE and/or MLST (Table 1). The nomenclature for PFGE variants indicates a particular pattern within the dominating serotype, whereas the MLST data allow grouping of isolates into different CCs. The ORs for invasive disease were calculated for each PFGE pattern and are given in Table 1. Isolates belonging to CCs associated with the highest invasive-disease potential (CC306, CC205, and CC191) were represented by isolates belonging to mainly 1 serotype (types 1, 4, and 7F, respectively).

Table 1.

Odds Ratio (OR) of Clonal Types Determined Using Pulsed-Field Gel Electrophoresis (PFGE)

	Isolates, No.
PFGE clone	Total	Invasive	Carriage	Other serotypes	CCa	OR (95% CI)	Comment
SWE14-6b	7	7	0	-	15	∞ (4.94-∞)	All invasive
SWE11A-1	13	3	10	-	62	1.00 (0.17-3.95)
SWE19F-5	8	0	8	-	63	0.00 (0.00-1.95)	All carriage
SWE18C-1	17	2	15	10A,6A, 18B	113	0.44 (0.05-1.91)
SWE18C-3b	6	5	1	6B	113	17.07 (1.89-809.23)
SWE14-2	14	3	11	-	124	0.91 (0.16-3.49)
SWE14-3b	8	5	3	-	124	5.68 (1.09-36.97)
SWE14-4	6	2	4	-	124	1.67 (0.15-11.8)
SWE23F-2	21	4	17	19A, 19F, 35F	124	0.78 (0.19-2.44)
SWE6B-1	61	16	45	6A	138	1.20 (0.62 -2.25)
SWE6B-2	34	9	25	-	138	1.21 (0.49-2.75)
SWE6B-3b	7	6	1	6A	138	20.62 (2.47-949.68)
SWE14-1	23	3	20	9V,19F	156	0.49 (0.09-1.69)
SWE3-1	12	2	10	-	180	0.66 (0.07-3.16)
SWE7F-1b	18	12	6	-	191	7.09 (2.41-23.4)
SWE14-5	5	0	5	19A	199	0.00 (0.00-3.64)	All carriage
SWE19A-1	25	6	19	19F, 15B, 15C	199	1.05 (0.34-2.81)
SWE4-1	6	4	2	6A	205	6.78 (0.96-75.51)
SWE19F-6	5	0	5	-	271	0.00 (0.00-3.64)	All carriage
SWE19F-7	5	0	5	-	271	0.00 (0.00-3.64)	All carriage
SWE1-1b	14	13	1	-	306	46.68 (6.91-1977.78)
SWE19F-4	11	0	11	-	309	0.00 (0.00-1.32)	All carriage
SWE38-1	6	0	6	-	393	0.00 (0.00-2.83)	All carriage
SWE16F-1	16	2	14	-	414	0.47 (0.05-2.08)
SWE19F-2	7	1	6	-	425	0.55 (0.01-4.61)
SWE23F-1	29	4	25	19A	439	0.52 (0.13-1.54)
SWE23F-3	5	0	5	-	439	0.00 (0.00-3.64)	All carriage
SWE6A-4	14	2	12	-	460	0.55 (0.06-2.51)
SWE6A-9	8	0	8	-	460	0.00 (0.00-1.95)	All carriage
SWE6A-8	5	0	5	-	473	0.00 (0.00-3.64)	All carriage
SWE6A-1	22	4	18	-	490	0.73 (0.18-2.27)
SWE6B-4	5	0	5	-	553	0.00 (0.00-3.64)	All carriage
SWE19A-2b	7	5	2	-	NPF	8.53 (1.38-90.45)
SWE19F-1	9	0	9	-	NPF	0.00 (0.00-1.68)	All carriage
SWE19F-9	8	0	8	-	NPF	0.00 (0.00-1.95)	All carriage
SWE35B-1	11	0	11	-	NPF	0.00 (0.00-1.32)	All carriage
SWE35F-1	6	0	6	-	NPF	0.00 (0.00-2.83)	All invasive
SWE6A-2	5	0	5	-	NC	0.00 (0.00-3.64)	All carriage
SWE19F-3	5	0	5	-	NC	0.00 (0.00-3.64)	All carriage
SWE19F-8	5	0	5	-	NC	0.00 (0.00-3.64)	All carriage
NT	5	0	5	-	NC	0.00 (0.00-3.64)	All carriage
Otherc	211	45	166			0.87 (0.57-1.30)
Total	715	165	550

PFGE clones representing ≥5 isolates are shown. A PFGE clone was given its name according to the most prominent serotype in the clone.

Abbreviations: CC, clonal complex; NC, nonconclusive; NPF, no predicted founder.

a CCs are named after the predicted founder as of the 2008 Mars Project in the MLST database.

b Significantly associated with invasive disease.

c PFGE clones with <5 isolates belonging to serotype 6A, 19F, 6B, 23F, 14, 10A, 9 V, 18C, 35F, NT, 3, 21, 11A, 23A, 9N, 15B, 7F, 19A, 33F, 16F, 23B, 4, 8, 31, 38, 11B, 15C, 17F, 22F, and 24F, in decreasing order.

The results presented in Table 1 suggest that isolates with different PFGE patterns and/or different CCs differ in their OR for causing invasive disease. Thus, isolates of PFGE SWE14-6 belonging to CC15 (all serotype 14) had a higher invasive-disease potential than serotype 14 isolates of CC124 (PFGE SWE23F-2, SWE14-2, SWE14-3, and SWE14-4) and CC156 (39% of serotype 14; Table 1). For serotype 19F, a multitude of different PFGE types and CCs were found, and the 3 isolates of PFGE pattern SWE19F-10 (CC251; data not shown) were solely from individuals with invasive disease. The other PFGE clones of serotype 19F belonged to different CCs and were almost exclusively found among healthy carriers (Table 1). These findings suggest that the invasive-disease potential for a serotype can be affected by its clonal distribution within that serotype.

Of particular interest was the presence of different PFGE patterns with different invasive-disease potentials belonging to the same genetic lineage (CC) and serotype, as exemplified for CC138 (type 6B), CC124 (type 14), and CC113 (type 18C; Table 1).

Comparative Genomics on CC138 Isolates Reveal Important Differences That May Influence Disease Outcome

CC138 was the most common genetic lineage among children with invasive disease and healthy carriers; this CC was observed in 110 isolates, of which 107 were of serotype 6B. One subclone, SWE6B-3 by PFGE, belonging to CC138 showed a high OR for causing invasive disease, in contrast to 2 other CC138 subclones, SWE6B-1 and SWE6B-2 (Table 1). Four serotype 6B isolates representing these 3 PFGE patterns with different invasive-disease potentials were selected for whole-genome sequencing (Figure 1A and 1B and Tables 1 and 2). The invasive isolate BHN191 represented SWE6B-3, strongly associated with invasive disease, while carriage isolate BHN418 came from the less invasive SWE6B-2. SWE6B-1 was represented by BHN237 from an individual with invasive disease and by BHN427 from a healthy carrier.

Figure 1.

Genomic differences between (A) invasive BHN191 from SWEB-3 and carriage BHN418 from SWEB-2 and (B) between invasive BHN237 and carriage BHN427, both from SWE6B-1.

Figure 2.

Factor H binding to the 4 strains of 6B and clonal complex 138, using Far Western blotting.

Intraclonal Variations in the Presence, Location, and Function of Prophages

The pneumococcal genome contains a large number of accessory regions (ARs), defined as at least 3 consecutive genes only present in a subset of clinical isolates [7]. The 4 CC138 isolates contained roughly the same ARs. The 2 exceptions were AR13, encoding a putative type 1 restriction system, which was missing in BHN237 and BHN427, and AR29, encoding a putative ABC transporter in which about two thirds of the genes were present in BHN237 and BHN427 but missing in the other isolates. Our previous identification of ARs [7] was based on the TIGR4 and R6 genomes, both of which lack prophages. However, pneumococcal prophages are common [16] and represent large segments of integrated DNA that might affect PFGE patterns in otherwise genetically related strains. Analyses of the 4 genomes revealed that 3 of the 4 CC138 isolates carried clustered genes homologous to identified prophage genes of pneumococci and commensal streptococci. Only carrier isolate BHN418 of SWE6B-2 lacked integrated prophage DNA (Figure 1A and 1B; data not shown). Phages of group 3 were present in the 2 invasive isolates BHN191 and BHN237, each located between the genes homologous to the TIGR4 genes SP_0020 and SP_0021. These 2 group 3 phages were very similar in the 2 invasive isolates, with an overall nucleotide sequence identity of 92% (counted over aligned sequences but not including sequence areas between contigs where sequence information is lacking). The sequence identity varied over the phage genome and over areas covering 89% of the phage sequence, and the nucleotide identity was 97%. Phages of group 2 were found in the 2 isolates of SWE6B-2 (BHN237 and BHN427) and were also similar but not identical in the 2 genomes (82% nucleotide sequence identity over alignable areas). BHN191 belonging to SWE6B-3 contained a second prophage that did not correlate with any of the 3 phage groups described by Romero et al [16], but similar sequences can be found in the 2 sequenced isolates SP19-BS75 and JJA (GenBank accession numbers ABAF00000000 and CP000919). In both BHN191 and JJA, the phage is located between TIGR4 genes SP_1908 and SP_1909.

To further investigate the possible role of the prophages in intraclonal variation and invasive-disease potential, we next studied the presence of phages in all serotype 6B isolates by PCR and performed functional assays of lysogeny by phage induction using mitomycin C [16] (Table 2). All isolates belonging to SWE6B-3 (all but 1 were invasive isolates) and 58 of 61 isolates belonging to SWE6B-1 (invasive and carrier isolates) were shown to harbor at least 1 phage type, with phage group 3 being the most prominent, whereas prophages were absent in most (25 of 34) isolates belonging to SWE6B-2 (the majority of which were carrier isolates). Isolates that had a lysogenic phage (Table 2) were also positive by PCR. However, we found lysogenic isolates that did not fully react with any of the 3 phage group–specific PCR primers, suggesting that there are additional phage groups present. Also, some isolates reacted positively with use of phage-specific primers, even though no phage induction was seen with mitomycin C, suggesting the presence of incomplete prophages. For SWE6B-2, we noted that 4 of the 6 isolates harboring inducible phages were invasive isolates, but we found no correlation as such between the presence of phage DNA and either carriage or invasive disease (Table 2).

Table 2.

Characteristics of All Isolates Belonging to Serotype 6B

BHN	Serotype	PFGE	ST	CCa	Invasive or Carriage	Phage Group(s)b	Mitomycin C–Induced Lysis	pcpAc	pspAd	pspC1e	pspC2e
BHN 191f	6B	SWE6B-3	138	138	I	3	+	+	Fam2, clade 3	PspC6.9	PspC9.4
BHN 328	6B	SWE6B-3	138	138	I	3	+	+	Fam2	…	…
BHN 387	6A	SWE6B-3	138	138	I	1, 3	+	+	Fam2	…	…
BHN 238	6B	SWE6B-3	138	138	I	3	+	+	Fam2, clade 3	…	PspC9.4
BHN 249	6B	SWE6B-3	138	138	I	3	+	+	Fam2	…	…
BHN 250	6B	SWE6B-3	138	138	I	3	+	+	Fam2, clade 3	…	PspC9.4
BHN 460	6B	SWE6B-3	138	138	C	Unclassified	+	+	Fam2, clade 3	…	…
BHN 273	6B	SWE6B-2	176	138	I	Unclassified	+	−	Fam1	…	…
BHN 212	6B	SWE6B-2	138	138	I	3	−	+	Fam2, clade 3	…	PspC9.4
BHN 259	6B	SWE6B-2	138	138	I	2	+	+	Fam2	…	…
BHN 305	6B	SWE6B-2	138	138	I	−	−	+	Fam2	…	PspC9.4
BHN 310	6B	SWE6B-2	138	138	I	3	+	+	Fam2	…	…
BHN 217	6B	SWE6B-2	138	138	I	Unclassified	+	+	Fam2	…	…
BHN 379	6B	SWE6B-2	138	138	I	3	−	+	Fam2	…	…
BHN 381	6B	SWE6B-2	138	138	I	−	−	+	Fam2	…	…
BHN 386	6B	SWE6B-2	138	138	I	3	−	+	Fam2	…	…
BHN 461	6B	SWE6B-2	138	138	C	Remnant	−	+	Fam2	…	…
BHN 462	6B	SWE6B-2	138	138	C	−	−	+	Fam2	…	…
BHN 463	6B	SWE6B-2	138	138	C	−	−	+	Fam2	…	…
BHN 418f	6B	SWE6B-2	138	138	C	−	−	+	Fam1, clade 1	PspC6.9	PspC9.4
BHN 464	6B	SWE6B-2	138	138	C	−	−	+	Fam2	…	…
BHN 465	6B	SWE6B-2	138	138	C	−	−	+	Fam2	…	…
BHN 466	6B	SWE6B-2	138	138	C	−	−	+	Fam2	…	…
BHN 467	6B	SWE6B-2	138	138	C	Remnant	−	−	Fam2	…	…
BHN 468	6B	SWE6B-2	138	138	C	Remnant	−	−	Fam2	…	…
BHN 469	6B	SWE6B-2	138	138	C	−	−	+	Fam2	…	…
BHN 470	6B	SWE6B-2	138	138	C	−	−	+	Fam2	…	…
BHN 471	6B	SWE6B-2	138	138	C	−	−	−	Fam2	…	…
BHN 510	6B	SWE6B-2	138	138	C	−	−	+	Fam2, clade 3	…	PspC9.4
BHN 50	6B	SWE6B-2	176	138	C	−	−	+	Fam2	…	…
BHN 472	6B	SWE6B-2	176	138	C	Remnant	−	+	Fam1	…	…
BHN 473	6B	SWE6B-2	176	138	C	Remnant	−	+	Fam2	…	…
BHN 543	6B	SWE6B-2	138	138	C	−	−	+	Fam1	…	…
BHN 544	6B	SWE6B-2	138	138	C	−	−	+	Fam2	…	…
BHN 545	6B	SWE6B-2	138	138	C	Remnant	−	+	Fam2	…	…
BHN 546	6B	SWE6B-2	138	138	C	3	+	+	Fam1	…	…
BHN 547	6B	SWE6B-2	138	138	C	−	−	+	Fam2	…	…
BHN 548	6B	SWE6B-2	138	138	C	−	−	+	Fam2	…	…
BHN 549	6B	SWE6B-2	138	138	C	−	−	+	Fam2	…	…
BHN 550	6B	SWE6B-2	138	138	C	Unclassified	+	+	Fam2	…	…
BHN 474	6B	SWE6B-2	176	138	C	Remnant	−	+	Fam2	…	…
BHN 237f	6B	SWE6B-1	176	138	I	2, 3	+	+	Fam1, clade2	PspC6.1	PspC9.1
BHN 314	6B	SWE6B-1	176	138	I	1, 3	+	+	Fam1	…	…
BHN 382	6B	SWE6B-1	176	138	I	3	−	+	…	…	…
BHN 200	6B	SWE6B-1	176	138	I	−	−	−	Fam2, clade 5	…	PspC9.1
BHN 322	6B	SWE6B-1	176	138	I	Unclassified	+	+	Fam1	…	…
BHN 264	6B	SWE6B-1	176	138	I	3	+	+	Fam1	…	…
BHN 269	6B	SWE6B-1	176	138	I	3	+	+	Fam1	…	…
BHN 287	6B	SWE6B-1	176	138	I	1	+	+	Fam1	…	…
BHN 307	6B	SWE6B-1	176	138	I	1	+	+	Fam1	…	…
BHN 296	6B	SWE6B-1	176	138	I	3	+	+	Fam1	…	PspC9.1
BHN 333	6B	SWE6B-1	176	138	I	1, 3	+	+	Fam1	…	…
BHN 248	6B	SWE6B-1	176	138	I	1, 3	+	+	Fam1	…	…
BHN 253	6B	SWE6B-1	176	138	I	1, 3	+	+	Fam1	…	…
BHN 300	6B	SWE6B-1	176	138	I	3	+	+	Fam1	…	…
BHN 266	6B	SWE6B-1	171	138	I	3	−	+	Fam1	…	…
BHN 271	6B	SWE6B-1	138	138	I	Unclassified	+	+	Fam1	…	…
BHN 427f	6B	SWE6B-1	176	138	C	2	+	−	Fam1, clade2	PspC6.1	PspC9.1
BHN 475	6B	SWE6B-1	176	138	C	1	+	+	Fam1	…	…
BHN 476	6B	SWE6B-1	176	138	C	1	+	+	Fam1	…	…
BHN 477	6B	SWE6B-1	176	138	C	1	+	+	Fam1	…	…
BHN 525	6A	SWE6B-1	176	138	C	1	+	+	Fam1	…	…
BHN 478	6B	SWE6B-1	176	138	C	1, 2, 3	+	+	Fam1	…	…
BHN 479	6B	SWE6B-1	176	138	C	1, 3	+	+	Fam1	…	…
BHN 480	6B	SWE6B-1	176	138	C	Unclassified	+	−	Fam1	…	PspC9.1
BHN 481	6B	SWE6B-1	176	138	C	2	+	−	Fam1	…	…
BHN 482	6B	SWE6B-1	176	138	C	2	+	−	Fam1	…	…
BHN 483	6B	SWE6B-1	176	138	C	1	+	+	Fam1	…	…
BHN 484	6B	SWE6B-1	176	138	C	1, 3	+	−	Fam1	…	…
BHN 485	6B	SWE6B-1	176	138	C	1, 3	+	−	Fam1	…	PspC9.1
BHN 486	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 487	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 488	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 489	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 490	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 491	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	PspC9.1
BHN 492	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 493	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 494	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 495	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 496	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 497	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 498	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 499	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 500	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 501	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 502	6B	SWE6B-1	176	138	C	Remnant	−	+	Fam1	…	…
BHN 503	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 504	6B	SWE6B-1	53	62	C	3	+	+	Fam1	…	…
BHN 51	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 505	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 526	6B	SWE6B-1	176	138	C	2	+	−	Fam1	…	…
BHN 528	6B	SWE6B-1	176	138	C	Unclassified	+	+	Fam1	…	…
BHN 529	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 530	6B	SWE6B-1	176	138	C	−	+	+	Fam1	…	…
BHN 531	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 532	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 533	6B	SWE6B-1	176	138	C	2	+	+	Fam2, clade 3	…	…
BHN 534	6B	SWE6B-1	176	138	C	3	+	+	…	…	…
BHN 535	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 536	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 506	6B	SWE6B-1	176	138	C	3	+	+	Fam1	…	…
BHN 553	6B	SWE6B-4	553	553	C	Unclassified	+	+	Fam2	…	…
BHN 554	6B	SWE6B-4	553	553	C	Unclassified	+	+	Fam2	…	…
BHN 555	6B	SWE6B-4	553	553	C	2	+	+	Fam2	…	…
BHN 556	6B	SWE6B-4	553	553	C	2	+	+	Fam2	…	…
BHN 557	6B	SWE6B-4	553	553	C	1, 3	+	+	Fam2	…	…
BHN 559	6B	SWE6B-5	553	553	C	1, 3	+	+	Fam2	…	…
BHN 560	6B	SWE6B-5	553	553	C	1	+	+	Fam2	…	…
BHN 320	6B	SWE6B-6	385	90	I	Unclassified	+	+	Fam1	…	…
BHN 304	6B	SWE6B-7	315	315	I	Unclassified	+	+	Fam2	…	…
BHN 302	6B	SWE6B-7	315	315	I	−	−	+	Fam2	…	…
BHN 272	6B	SWE6B-8	2659	90	I	Unclassified	+	+	Fam1	…	…
BHN 255	6B	SWE6B-8	146	90	I	Unclassified	+	+	Fam1	…	…
BHN 306	6B	SWE6B-9	2660	2660	I	Unclassified	+	+	Fam1	…	…
BHN 278	6B	SWE6B-10	90	90	I	1, 2	+	+	Fam1	…	…
BHN 197	6B	SWE6B-11	138	138	I	2	+	+	Fam2	…	…
BHN 507	6B	SWE6B-12	176	138	C	2, 3	+	+	Fam1	…	…
BHN 514	6B	SWE6B-13	176	138	C	Unclassified	+	+	Fam1	…	…
BHN 515	6B	SWE6B-14	8790	Not present in any group	C	−	−	+	Fam1	…	…
BHN 516	6B	SWE6B-15	2156	156	C	Unclassified	+	+	Fam1	…	…
BHN 517	6B	SWE6B-16	8789	Not present in any group	C	−	−	−	Fam1	…	…
BHN 518	6B	SWE6B-17	8791	Not present in any group	C	2	+	+	Fam1	…	…
BHN 519	6B	SWE6B-18	553	553	C	2	+	+	Fam2	…	…
BHN 520	6B	SWE6B-19	NC	NC	C	Unclassified	+	−	…	…	…
BHN 521	6B	SWE6B-20	710	1121	C	Unclassified	+	+	Fam2	…	…
BHN 522	6B	SWE6B-21	2936	Singleton	C	−	−	+	Fam1	…	…
BHN 523	6B	SWE6B-22	176	138	C	Unclassified	+	+	Fam1	…	…
BHN 509	6B	SWE6B-23	176	138	C	2, 3	−	+	Fam1	…	…
BHN 508	6B	18C-3	138	138	C	3	+	+	Fam2	…	…

Abbreviations: CC, clonal complex; NC, nonconclusive; PCR, polymerase chain reaction; PFGE, pulsed-field gel electrophoresis; ST, sequence type; –, negative; +, positive.

a CCs are named after the predicted founder as of the 2008 Mars Project in the MLST database.

b Defined by a positive PCR product of primer pairs designed by Romero et al [16]. The term “unclassified” denotes a lysogenic strain but no detectable PCR product, suggesting the presents of a novel prophage not belonging to any of the 3 groups previously described by Romero et al [16]. The term “remnant” denotes expression of genes without lyse upon induction of mitomycin C.

c Determined by PCR, using specific primers.

d Determined by PCR, using specific primers, and clades by sequencing.

e Determined by Sanger sequencing. For pspC1, PspC6.9 was from G386, and PspC6.1 was from G31. For pspC2, PspC9.4 was from G386, and PspC9.1 was from G31.

f Selected for whole-genome sequencing.

Intraclonal Allelic Variations of Known Virulence Surface Proteins

The comparative sequence analyses revealed differences among the 4 CC138 isolates in genes encoding the important virulence-associated surface proteins pneumococcal surface protein A (PspA), pneumococcal choline-binding protein A (PcpA), and pneumococcal surface protein C (PspC). These surface proteins are known to represent dominating antigens to which many individuals without prior invasive disease have antibodies [23].

pspA is a mosaic gene that has evolved through extensive recombination and is an important vaccine candidate [18, 24]. PspA is required for pneumococcal virulence and contributes to colonization in mice, and it has been shown to interfere with the fixation of complement C3 [25, 26] and to bind human lactoferrin [27]. PspA can be divided into 3 families and 6 clades, in which the clade-defining region is determined by residues 192–290. It has previously been reported that a single clone usually consists of strains from 1 specific clade. However, in this study of CC138 isolates, we observed extensive intraclonal pspA variation (Table 2). Thus, BHN191 of SWE6B-3 possessed a pspA allele similar to TIGR4 (pspA family 2, clade 3), whereas the other 3 isolates carried a pspA allele similar to R6 (pspA family 1) but still with sequence differences placing them in different clades (Table 2).

All CC138 6B isolates were subsequently tested for the pspA family by use of family-specific primers, and the complete pspA gene was sequenced from a selected number of isolates to determine the clade (Table 2 and Supplementary Figures 1 and 2). All isolates from SWE6B-3 belonged to pspA family 2, and those sequenced were of clade 3. In SWE6B-2, only a minority of isolates possessed a pspA allele of family 1 (clade 1), whereas the majority were of pspA family 2 (clade 3). Finally, for SWE6B-1, most isolates carried pspA of family 1 (clade 2), but 2 were of family 2 (clades 2 and 5). Thus, for the largest single clone among children (CC138) in Stockholm, we identified extensive pspA variation. However, no specific pspA family and clade correlated to either carriage or disease.

Another sequence difference among the 4 completely sequenced CC138 strains was the complete absence of pcpA in the carrier isolates BHN427 of SWE6B-1 but the presence of this virulence-associated gene in the other 3 isolates (Table 2). In BHN427 pcpA was replaced by 2 open reading frames with no similarities to pcpA (data not shown). The presence of pcpA was further analyzed by PCR in all CC138 isolates (Table 2). The pcpA gene was absent in 10 of 72 tested isolates, of which 8 were from healthy carriers and 2 from individuals with invasive disease. PcpA is a potential vaccine candidate, but unlike PspA expression of this protein is not required for colonization in mice [28].

Each of the 4 CC138 isolates possessed 2 different pspC genes, pspC1, encoding a choline binding version of this highly variable protein, and pspC2, encoding a cell-wall-anchored LPXTG protein. Of the 4 completely sequenced CC138 isolates, BHN191 and BHN418 were of pspC1 variant PspC6.9. [29]. Nevertheless, the 2 predicted amino acid sequences differed at a number of locations (Supplementary Figure 1; the 2 sequences can be aligned with 80.9% sequence identity, with 4.0% gaps). The 2 pspC1 genes of BHN237 and BHN427 (both of SWE 6B-1) were, however, identical and of PspC6.1 [29]. All 4 pspC1 sequences contain 2 copies of an RNYPT motif known to bind human FH [29].

The 4 pspC2 sequences were all very similar but not identical. BHN191 and BHN418 both had a pspC2 allele of PspC9.4, and the sequences were identical except for the length of the proline-rich repeat region. BHN237 and BHN427, on the other hand, were of PspC9.1 and were 99.1% identical to each other, with all differences in the repeat region. When additional CC138 isolates were sequenced for pspC2 (Table 2), all SWE6B-3 and SWE6B-2 isolates had a PspC9.4 allele, and the SWE6B-1 isolates were all PspC9.1. The major difference again involved the number of repeats in the proline-rich region.

Intraclonal Variations Within PspC1 and PspC2 and Correlation to Human FH Binding

One human-specific virulence property in pneumococci is the ability to sequester human alternative pathway inhibitor FH, leading to evasion of complement-mediated opsonization. In pneumococci, only PspC1 and PspC2 (Hic) have been shown to bind FH [30, 31]. Since we could identify sequence variations in the different pspC1 and pspC2 genes of the 4 6B CC138 strains, we examined FH binding by use of Far Western blotting (Figure 2). FH was shown to bind 1 major protein in strainT4, corresponding in size (95 kDa) to PspC1 purified from this strain (data not shown). All 4 6B strains were shown to preferentially bind FH to a similarly sized protein as in strain T4, suggesting that it represents PspC1. Interestingly, we found that FH bound more intensively to this protein in invasive strain BHN191, compared with the other 3 strains. Since none of the 4 CC138 strains were possible to transform, we were unable to create mutants in the pspC1 or pspC2 genes. FH binding was further quantified by whole-cell ELISA, using strain T4 as the reference strain, in which pneumococci were attached to a microtiter plate and the binding of FH was analyzed. The 4 6B strains were found to sequester more FH than strain T4 (difference, 2–4-fold), with BHN191 exhibiting the highest degree of binding (data not shown), suggesting that enhanced complement resistance could be one explanation why this strain exhibited a higher invasiveness in children than the other isolates. All 4 strains bound FH with maintained cofactor activity, as representative C3b cleavage products were found when the strains were pretreated with FH and incubated with C3b and factor I (data not shown).

Differences in Virulence Within CC138 in an Experimental Mouse Model May Correlate With the Presence of Virulence Protein PcpA

The 4 CC138 isolates were next tested for virulence in an intranasal murine model. The carrier isolate BHN427 was considerably less virulent than the corresponding invasive isolate BHN237 (Figure 3A–C). It is likely that this difference in mice virulence reflects the lack of PcpA, a known mouse virulence determinant, in the carrier isolate. However, we found no difference in mouse virulence between invasive BHN191 of SWE6B-3 with high invasive-disease potential, compared with carrier BHN418 of SWE 6B-2 and invasive BHN237 of SWE6B-1, suggesting that the high invasive-disease potential of SWE6B-3 in children may represent 1 or more virulence attributes only operating in the human setting, such as sequestration of human FH.

Figure 3.

Survival of mice (A) and the number of colony-forming units (CFU) in the bloodstream at 24 hours (B) after intranasal challenge with the 4 strains of serotype 6B and clonal complex 138 representing clonal types with pulsed-field gel electrophoresis patterns indicating different invasive-disease potentials. *P < .05.

Concluding Remarks

We found that efficient horizontal gene transfer events generate intraclonal variants in dominating lineages that differ in invasive-disease potential in children and in virulence in mice. Also, we observed major differences in phage DNA content and in the presence or sequence of surface antigens. These surface antigens are known to be important virulence factors, and some bind human FH, which has been shown to increase complement resistance and to promote invasion directly [29, 32, 33]. We found differential FH binding among intraclonal variants, potentially affecting invasiveness. We hypothesize that a constantly ongoing selection for immune escape variants of pneumococci occurs during carriage that result in variants that may affect the invasive-disease potential in children by simultaneously affecting the disease-promoting properties of individual proteins.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.