Abstract

BackgroundPilus components of Streptococcus pneumoniae encoded by rlrA were recently shown to elicit protection in an animal model of infection. Limited data are available on the prevalence of the rlrA operon in pneumococci; therefore, we investigated its distribution and its antigenic variation among disease-causing strains

MethodsThe prevalence of rlrA and its association with serotype and genotype were evaluated in a global panel of 424 pneumococci isolates (including the 26 drug-resistant clones described by the Pneumococcal Molecular Epidemiology Network)

ResultsThe rlrA islet was found in 130 isolates (30.6%) of the defined collection. Sequence alignment of 15 rlrA islets defined the presence of 3 clade types, with an overall homology of 88%–92%. The presence or absence of a pilus-encoding operon correlated with S. pneumoniae genotype (P<.001), as determined by multilocus sequence typing, and not with serotype. Further investigation identified a positive trend of rlrA occurrence among antimicrobial-resistant pneumococci

ConclusionsOn the basis of S. pneumoniae genotype, it is possible to predict the incidence of the rlrA pilus operon in a collection of pneumococcal isolates. This will facilitate the development of a protein vaccine

Streptococcus pneumoniae is a commensal pathogen and often part of the normal flora of the human respiratory tract. The most common pneumococcal infections are otitis media and pneumonia, although S. pneumoniae may less frequently cause invasive diseases such as meningitis and sepsis [1]. Prophylactic vaccination with capsular polysaccharide-based vaccines and antibiotic therapy are 2 effective approaches that decrease the risk of disease. However, selective pressures are affecting S. pneumoniae population dynamics. In this respect, the treatment and prevention of pneumococcal infections has been complicated by the spread of clones that have increase in the incidence of isolates resistant to antibiotics [2–5], and, following the introduction of the PCV7 glycoconjugate vaccine (Prevnar), the spread of nonvaccine serotypes (NVT) [6, 7]. This evidence provides the impetus to develop vaccines based on protein antigens, which are capable of serotype-independent protection [8, 9]

Recently, in a mouse model of infection, S. pneumoniae was shown to express a pilus that has subunits able to elicit protection both by active and passive immunization [10–12]. These structures are encoded by a chromosomal element defined as the rlrA pathogenicity islet [13]. The function of pneumococcal pili is currently an area of investigation. To date, pili are known to be involved in adhesion to lung epithelial cells in vitro, as well as in colonization in a murine model of infection [10, 14]. Sequence analysis of available S. pneumoniae genomes revealed that the serotype 6B strain 670 contains a region of DNA that shares 89% sequence identity with the rlrA islet from TIGR4. However, the locus is absent in strains R6 (a nonencapsulated strain derived from serotype 2 D39), G54 (serotype 19F), and Spain-23F-1 (serotype 23F), suggesting that the rlrA islet is not widespread in pneumococcal isolates

Therefore, we investigated the distribution and sequence variability of the rlrA islet among 424 S. pneumoniae clinical isolates (26 of these were derived from the Pneumococcal Molecular Epidemiology Network [PMEN] multidrug-resistant collection) [15]. We evaluated isolates from distinct geographic regions, which were selected for epidemiological and genetic diversity. The isolates were analyzed for the presence and sequence variability of the rlrA islet, as well as serotype and genetic background as defined by multilocus sequence typing (MLST). Here we show that rlrA-positive pneumococci are found worldwide, and that the operon, defined by 3 variants, is strictly correlated with the genotype of S. pneumoniae rather than the serotype. These findings, along with previous data showing that pili are protective against pneumococcal infection [11], suggest that they may be useful components for a vaccine

Materials and Methods

Strain collectionA total of 424 isolates collected worldwide were analyzed for the presence of rlrA (table 1). This collection consisted of 26 clones from the PMEN collection [15]; 53 invasive pneumococcal isolates from the Centers for Disease Control and Prevention in the United States [16]; 74 clinical isolates from Salvador, Brazil (42 isolates recovered from carriers and 32 from patients with meningitis, respectively); 26 clinical isolates from the Unites States and Finland (15 recovered from carriers, 2 from patients with otitis media, and 9 from patients with invasive diseases); 15 invasive clinical isolates from Ghana [17]; 14 invasive clinical isolates from Bangladesh; 40 clinical isolates from Italy; 12 invasive clinical isolates from Kenya; 76 clinical isolates from Sweden; and 11 laboratory strains from the Novartis collection. Additionally, data about rlrA presence have been extrapolated for 74 isolates from a previously published report [18] and for 3 isolates from sequence data available at the Sanger Institute Web site (http://www.sanger.ac.uk)

Table 1

Sources and composition of the strain collection

Table 1

Sources and composition of the strain collection

rlrA islet detection and sequencingTIGR4 [19] and 670 (genomic sequence available at http://www.TIGR.org) rlrA nucleotide sequences were analyzed, and a set of 30 oligonucleotide primers was designed (table 2): 22 matching inside the islet, 2 annealing in 2 conserved genes (SP459-SP470 of TIGR4) flanking the operon, and 6 clade-specific primers. The set of primers was defined to detect the presence and location of the operon within all the isolates in the collection, to amplify and sequence the entire locus, and, for rlrA-positive strains, to determine the clade type (I, II, or III). In brief, the genomic location of the operon was determined by simultaneously assessing 4 polymerase chain reaction (PCR) amplifications: the first using primers (459 for, 470 rev) matching regions flanking the operon (when rlrA was absent, a lower fragment size was detectable); the second with 459 for and 1 rev, matching inside the islet (the fragment was detectable only when the islet was present and inserted in the genomic region between SP459-SP470); the third and fourth amplifying in conserved regions of the operon (amplification detected whenever the islet was present in the genome) (table 2)

Table 2

Nucleotide sequences of the primers used in this study

Table 2

Nucleotide sequences of the primers used in this study

Multilocus sequence typing (MLST)MLST was performed as previously described [20]. In brief, internal fragments of the aroE, gdh, gki, recP, spi, xpt and ddl genes were amplified by PCR directly from the bacteria by use of the primer pairs indicated at http://spneumoniae.mlst.net/misc/info.asp#experimental. Sequences were obtained on both DNA strands by use of an ABI 3730xl DNA Analyzer. Alleles from the MLST Web site (http://spneumoniae.mlst.net) were downloaded for alignment analyses and sequence type (ST) determination. In MLST, an ST is uniquely determined by the allelic profile. New allelic profiles have been submitted to the MLST database for ST assignment

Clonal complex (CC)CCs are groups of STs that share a recent common ancestor. The eBURST algorithm defines clonal complexes by partitioning the MLST data set into groups of single-locus variants (SLVs), i.e., profiles that differ at 1 of the 7 MLST loci [21]. This partitioning associates each ST with a clonal complex and identifies the most likely founder ST, which is defined as being the ST with the greatest number of SLVs within the CC. To explore the relationship between rlrA presence in our data set and CC, we ran eBURST with default settings on the entire MLST database and subsequently assigned each ST in our data set to a CC. In this work, we have named CCs in accordance with the ST number of the founder predicted by eBURST

Statistical analysisAll statistical analyses of association were performed using R (version 2.4.0; CRAN) and the vcd package (version 1.0–3; CRAN) for the analysis of categorical data. Pearson χ2P values and Cramer V coefficients were reported to test the null hypothesis of independence and to measure the strength of the observed correlations [22]

Gene prediction and multiple alignmentThe 7 TIGR4 rlrA islet gene sequences were used to perform Smith-Waterman searches against the 13 rlrA sequences of rlrA-positive PMEN clones to find the genes in each locus. Multiple alignments of the homologous genes were obtained by translating the predicted nucleotide sequences into peptides, aligning, and then back-translating to the original nucleotide sequences with T-Coffee (version 4.70) [23]. The rlrA multiple alignment was reconstructed by concatenating the 7 aligned genes with the corresponding intergenic regions. Phylogenetic trees were reconstructed from nucleotidic multiple alignment by use of MEGA (version 3.1) [24], by use of unweighted pair group method with arithmetic mean and K2P distance correction. Multiple alignments were used with MEGA (version 3.1) [24] to measure the nucleotide and amino acid identity within and between clades I, II, and III

Results

Prevalence of the rlrA operon inS. pneumoniaeTo define the rlrA islet distribution, a panel of 424 isolates was analyzed The collection was obtained by sampling clinical isolates worldwide; it includes 70 serotypes and 200 different STs, which can be grouped into 76 CCs [20]. rlrA islet presence was determined by PCR and confirmed (for the 26 PMEN clones) by Southern blot analysis with a conserved fragment of the srtB gene (data not shown). Out of the entire collection, 130 (30.6%) of the S. pneumoniae isolates were rlrA positive and showed the pilus operon consistently inserted in the same genomic region

Association of the rlrA islet with serotypeTo identify common features among rlrA-positive isolates, we investigated its association with capsular serotype. Figure 1A represents a histogram of isolates stratified by serotype [25]; the dark area corresponds to the fraction of rlrA-positive isolates. Of the most common serotypes (1, 3, 4, 5, 6B, 7F, 9V, 14, 18C, 19A, 19F, and 23F, which accounted for 257 (60.6%) of all pneumococcal isolates), serotypes 1, 3, 5, 7F, and 18C were not associated with the presence of the rlrA operon. However, with the exception of Serotype 9V, most rlrA-positive serotypes included variable fractions of rlrA-negative strains. Therefore, although there was an association between the capsular serotype and the presence of the islet, the strength of this association was minimal (Cramer V=0.68)

Figure 1

Distribution of rlrA-positive Streptococcus pneumoniae by serotype (A) and by clonal complex (CC) (B). Histograms show the relative frequency of rlrA-positive (black) and rlrA-negative (gray) isolates in the collection, stratified by serotype or CC. In panel A the column “others” contains serotypes represented in our collection by a single isolate; in panel B “others” clusters together CCs represented by a single isolate and isolates (singletons) that could not be assigned to a CC. Asterisks CCs for which 1 sequence type was tested. Total number of isolates, N=424. C eBURST graphic representation of CC 176, obtained by eBURST analysis on entire MLST S. pneumoniae database. Sequence types analyzed for rlrA presence are represented by larger font

Figure 1

Distribution of rlrA-positive Streptococcus pneumoniae by serotype (A) and by clonal complex (CC) (B). Histograms show the relative frequency of rlrA-positive (black) and rlrA-negative (gray) isolates in the collection, stratified by serotype or CC. In panel A the column “others” contains serotypes represented in our collection by a single isolate; in panel B “others” clusters together CCs represented by a single isolate and isolates (singletons) that could not be assigned to a CC. Asterisks CCs for which 1 sequence type was tested. Total number of isolates, N=424. C eBURST graphic representation of CC 176, obtained by eBURST analysis on entire MLST S. pneumoniae database. Sequence types analyzed for rlrA presence are represented by larger font

Association of the rlrA islet with genotypeThe genetic structure of the S. pneumoniae collection was determined by MLST [26]. As described above, each isolate was assigned to a CC after ST determination. Figure 1B shows a histogram representing the number of isolates analyzed for each CC and illustrates the fraction of rlrA-positive and rlrA-negative S. pneumoniae isolates. Among the CCs most represented in this study (figure 1B), at least 10 different STs were analyzed for the presence of rlrA. Almost all the STs analyzed were homogeneous for rlrA presence, with a few exceptions. In brief, of 18 ST156 (CC162) and 7 ST205 (CC205) isolates analyzed, 1 and 2 isolates, respectively, were negative for the rlrA islet. Figure 1B clearly shows that CCs are composed of either rlrA-positive or rlrA-negative isolates, which indicates that its presence is dependent on the genetic background P<.001; Cramer V=0.96). CC176 represents an exception to the scenario described above (figure 1C). Eight distinct STs were tested (comprising 16 isolates): 5 were rlrA positive and 3 were rlrA negative. By mapping the STs analyzed onto a predictive eBURST diagram (which displays hypothetical patterns of descent), we observed that rlrA-positive and rlrA-negative STs were grouped into 2 defined areas of the CC

rlrA pilus islet sequence variabilityIn the TIGR4 strain, pili are encoded within the rlrA islet, a 12 kb genomic region flanked by a pair of IS1167 elements, a feature characteristic of mobile genetic elements [13, 27]. The islet consists of 7 genes, namely, rlrA which encodes a RofA-like transcriptional regulator, rrgA, rrgB and rrgC, which encode LPXTG cell-wall anchored proteins that constitute the structure of the pilus, and srtB, srtC and srtD, which encode for the 3 sortase enzymes that catalyze the pilus polymerization reaction [10, 12, 28]

The well-characterized PMEN collection was used as a reference collection for initial assessment of the variability of the rlrA islet in a global panel of pneumococcal isolates [15]. By PCR, the rlrA islet was found to be present in 13 (50%) of 26 PMEN clones (table 3). Sequencing of the full-length locus was carried out for the 13 PMEN strains (GenBank accession numbers EF560625–EF560637) and an additional 18 isolates from the collection (data not shown). Multiple alignment of entire nucleotide sequences revealed the presence of 3 main families, herein referred to as clades I, II, and III (figure 2)

Table 3

Presence of the rlrA islet in the Pneumococcal Molecular Epidemiology Network collection

Table 3

Presence of the rlrA islet in the Pneumococcal Molecular Epidemiology Network collection

Figure 2

Genetic variability among rlrA islets in Streptococcus pneumoniae isolates. Distance-based tree obtained from the multiple sequence alignment of 15 rlrA islets (TIGR4, 6B 670, 13 rlrA-positive Pnemococcal Molecular Epidemiology Network clones)

Figure 2

Genetic variability among rlrA islets in Streptococcus pneumoniae isolates. Distance-based tree obtained from the multiple sequence alignment of 15 rlrA islets (TIGR4, 6B 670, 13 rlrA-positive Pnemococcal Molecular Epidemiology Network clones)

rlrA sequences within each clade shared at least 98.8% DNA sequence identity, whereas sequences belonging to different clades shared from 88.1% to 92.1% identity (clades I and III being the most similar) (table 4). Data on protein sequence similarity between clades highlights the fact that the transcriptional regulator RlrA was the most conserved (100%), followed by the 3 sortases and the RrgC pilus subunit (98%–99%). Conversely, the RrgA (84%–98%) and RrgB (49%–67%) pilus components are the least conserved, with RrgB existing in 3 different variants, corresponding to the 3 clades of the rlrA operon (table 4). A set of clade-specific primers designed from the most variable regions of rrgB (table 1B) allowed the determination of rlrA clade type by PCR. Of the 130 rlrA-positive isolates, 81 (62.2%) were defined as clade I; 35 (26.8%), clade II; and 14 (10.7%), clade III. Furthermore, all strains belonging to the same CC were homogeneous for rlrA clade type

Table 4

Sequence variability of the rlrA operon

Table 4

Sequence variability of the rlrA operon

Sequence analysis of IS1167 target site duplicationsThe genomic region flanking the operon was analyzed both in rlrA-positive and rlrA-negative isolates. The objective was to investigate whether the presence of the rlrA islet was the result of a single insertion event, which may have evolved into 3 separate clades, or whether different clones acquired the islet independently

In TIGR4, the rlrA islet is flanked by 2 IS1167 mobile genetic elements characteristic of S. pneumoniae These 2 elements, IS1167-1 (SP0460) and IS1167-2 (SP0469), share 61% identity at the nucleotide level. IS1167 has been shown to be stable during laboratory passage and to exhibit a high degree of variability [29]. Moreover, IS1167 sequences are characterized by the presence of imperfect terminal inverted repeats (which are responsible for homologous recombination) and are bracketed by 8-bp direct repeats, generated by the duplication of the IS1167 target sequence

Sequence analysis of this region in 30 isolates revealed the absence of target site duplication on the outer extremities of the operon (including IS elements), as well as on either side of the single IS elements. The lack of direct repeats suggests that the region was not actively inserted by these IS elements as a composite transposon, but could have been acquired by homologous recombination. To date, 2 different insertion sites have been found for the IS1167-1 element (these sites are clade-type independent, but CC related). No differences were found for IS1167-2 insertion sites in rlrA-positive isolates

Sequence analysis of rlrA-negativeS. pneumoniaerlrA-negative pneumococci revealed 3 genetically distinct arrangements (figure 3): (1) a region of 800 bp that codes for a hypothetical protein is present (as seen in R6 and in a majority of rlrA-negative strains), (2) an IS1167-2 bracketed by 8-bp direct repeats is inserted in the genomic context described in the first scenario, and (3) the pilus operon is absent and both IS1167-1 and IS1167-2 are present. The latter situation, found in an ST156 isolate and in 2 ST205 isolates mentioned previously, could represent an event of rlrA islet loss, whereas the second scenario, described in 20 clonally related strains, supports the idea that this locus represents a hot spot region for the insertion of IS1167 family members

Figure 3

Schematic representation of the genomic region in rlrA-negative Streptococcus pneumoniae isolates (A–C) and in rlrA-positive isolates (D). A genomic background (R6-like); B genetic arrangement derived from IS1167 insertion on an R6 genomic background; C genetic arrangement potentially derived from loss of rlrA. White triangles direct repeats formed by target site duplication. Gray boxes open reading frame coding for a hypothetical protein; white boxes open reading frames flanking the region

Figure 3

Schematic representation of the genomic region in rlrA-negative Streptococcus pneumoniae isolates (A–C) and in rlrA-positive isolates (D). A genomic background (R6-like); B genetic arrangement derived from IS1167 insertion on an R6 genomic background; C genetic arrangement potentially derived from loss of rlrA. White triangles direct repeats formed by target site duplication. Gray boxes open reading frame coding for a hypothetical protein; white boxes open reading frames flanking the region

Antimicrobial-resistant clones are more likely to carry the pilus isletGiven the high prevalence of rlrA-positive strains in the antibiotic-resistant PMEN collection (50%), the relationship between antimicrobial resistance and rlrA presence was assessed further. In this analysis, we used the complete S. pneumoniae database downloaded from the MLST web site (http://www.mlst.net) instead of our collection, because antimicrobial resistance data was incomplete for our collection. All pneumococci for which minimum inhibitory concentration (MIC) values were available were classified as susceptible (S) or nonsusceptible (NS) to 3 major antibiotic compounds: penicillin, erythromycin, and tetracycline [30]. STs were subsequently classified as NS if they were composed of at least 50% NS isolates. STs not fulfilling this requirement were classified as S. In most cases, a single ST was composed of either S or NS strains (Cramer V correlation coefficients between ST and antibiotic resistance were 0.94, 0.78, and 0.77 for penicillin, erythromycin and tetracycline, respectively). In contrast to the distribution of the rlrA islet, several CCs were heterogeneous for resistance (the Cramer V correlation coefficients between CC and resistance were 0.81, 0.73, and 0.67 for penicillin, erythromycin and tetracycline, respectively)

The presence of the rlrA islet was then projected onto the MLST data set (figure 4), by classifying as rlrA positive those STs in our collection that were composed of at least 50% rlrA-positive isolates. The percentage of rlrA-positive STs among strains susceptible to penicillin, erythromycin, or tetracycline was 21%, 23%, and 32%, respectively, whereas the percentage of rlrA-positive STs among clones not susceptible to penicillin, erythromycin, or tetracycline was 51%, 53%, and 41%, respectively. The Pearson χ2 statistic showed that for penicillin and erythromycin, the presence of the rlrA islet was significantly associated with resistance (P<.001 and P<.001, respectively). A less significant association was found for tetracycline (P<.37), probably due to the lower number of strains for which MIC data were available

Figure 4

Distribution of the rlrA islet in antimicrobial-resistant S. pneumoniae sequence types (STs). A Histogram showing the relative frequency of rlrA+ (black) and rlrA (gray) STs. Strains within the multilocus sequence typing database collection for which minimum inhibitory concentration (MIC) values were provided were analyzed and stratified by ST and antibiotic susceptibility class (susceptible [S] and nonsusceptible [NS]). NS strains were defined by MIC values of >0.12 μg/mL, >0.5 μg/mL, and >4 μg/mL for penicillin, erythromycin, and tetracycline, respectively. Those STs in which >50% of the strains were NS were defined as NS. The total number of STs analyzed was 151 for penicillin, 137 for erythromycin, and 100 for tetracycline. The fraction of rlrA-positive STs within S and NS groups (95% confidence interval [CI], based on the binomial distribution of probability) were as follows: the fraction of rlrA-positive STs among strains susceptible to penicillin, erythromycin, or tetracycline was 0.21 (95% CI, 0.14–0.30), 0.23 (95% CI, 0.15–0.32), and 0.32 (0.21–0.45), respectively, whereas the fraction of rlrA-positive STs among clones not susceptible to penicillin, erythromycin, or tetracycline was 0.51 (95% CI, 0.36–0.67), 0.53 (0.36–0.70), and 0.41 (0.25–0.58), respectively

Figure 4

Distribution of the rlrA islet in antimicrobial-resistant S. pneumoniae sequence types (STs). A Histogram showing the relative frequency of rlrA+ (black) and rlrA (gray) STs. Strains within the multilocus sequence typing database collection for which minimum inhibitory concentration (MIC) values were provided were analyzed and stratified by ST and antibiotic susceptibility class (susceptible [S] and nonsusceptible [NS]). NS strains were defined by MIC values of >0.12 μg/mL, >0.5 μg/mL, and >4 μg/mL for penicillin, erythromycin, and tetracycline, respectively. Those STs in which >50% of the strains were NS were defined as NS. The total number of STs analyzed was 151 for penicillin, 137 for erythromycin, and 100 for tetracycline. The fraction of rlrA-positive STs within S and NS groups (95% confidence interval [CI], based on the binomial distribution of probability) were as follows: the fraction of rlrA-positive STs among strains susceptible to penicillin, erythromycin, or tetracycline was 0.21 (95% CI, 0.14–0.30), 0.23 (95% CI, 0.15–0.32), and 0.32 (0.21–0.45), respectively, whereas the fraction of rlrA-positive STs among clones not susceptible to penicillin, erythromycin, or tetracycline was 0.51 (95% CI, 0.36–0.67), 0.53 (0.36–0.70), and 0.41 (0.25–0.58), respectively

Discussion

The polysaccharide conjugate vaccine for S. pneumoniae has decreased the incidence of invasive disease caused by vaccine types, although the increased incidence disease due to nonvaccine serotypes makes difficult the estimate of polysaccharide conjugate vaccine coverage [6, 7]. Moreover, the epidemiology of the pneumococci is complicated by the diversity of circulating genotypes often disguised by multiple capsular types. The introduction of a protein-based vaccine could overcome limitations due to serotype replacement and serotype-dependent coverage. A tool to rapidly estimate the distribution and variability of protein antigens in S. pneumoniae could facilitate vaccine development

In this study, a collection of isolates that represented both the capsular and genetic diversity of S. pneumoniae was used to determine the prevalence and antigenic variability of the rlrA pilus. In this article, we report 3 major findings. First, we demonstrate that the rlrA operon previously identified in laboratory strains is found in multiple isolates from diverse parts of the world. In fact, 30% of isolates in our pneumococcal collection contained this genetic islet, confirming a recent report about strains isolated from the Native American population [31]. This suggests that pili could become one of several components in a protein-based vaccine for S. pneumoniae

Second, we show that presence of the rlrA islet segregates according to genotype, rather than phenotype. Moreover, the islet was probably acquired prior to the formation of the clonal complexes and steadily maintained during clonal diversification, even in CCs that show evidence of a complex evolutionary history. Following these observations, the occurrence of rlrA-negative STs in CC 176, the founder of which is rlrA positive, could be explained as an instance of pilus operon loss. This event can be inferred to have taken place between ST 176 and ST 361 (figure 1C), the rlrA-negative genotype then having been inherited by the descendant STs. Alternative explanations could be acquisition of the rlrA pilus islet, subsequent to the formation of distinct subclones of the CC; or artificial grouping by eBURST of isolates from 2 independent clonal expansions, one of which originated from an rlrA-positive strain and the other from an rlrA negative strain, although testing of the eBURST algorithm suggests its robustness under the evolutionary processes [32]

Third, the genetic variability of the rlrA pilus islet can be organized into 3 clades, as demonstrated by sequence analysis of the operons from genetically diverse pneumococci. Single protein alignments highlight RrgA and RrgC as the most promising components for a serotype-independent vaccine, while the variability of RrgB makes this protein less attractive. Additionally, the rlrA clade type within a CC was determined to be identical, strongly suggesting clonal inheritance of this islet; this was in contrast with serotype, which frequently varies within a CC. These findings clearly reveal that the association between the rlrA islet and serotype depends on the genetic link between serotype and genotype. In fact, there is a significant correlation between serotype and the presence of the operon only for those serotypes that correspond to a restricted number of CCs, such as serotype 9V (CC 162) and serotype 3 (CC 180)

Furthermore, sequence analysis of the genomic region in rlrA-positive and rlrA-negative isolates suggests that the islet could have been acquired by homologous recombination during multiple events mediated by IS1167 insertions. This latter observation raises questions regarding the potential genetic source of the rlrA genetic element in other Streptococcus species. To date, 25 S. mitis isolates have been analyzed and tested negative for the rlrA islet by Southern blot analysis and PCR (data not shown). Moreover, the probability that alternative elements could code for a pilus in S. pneumoniae strains lacking the rlrA islet, as is the case for S. agalactiae and S. pyogenes [33, 34], will be verified by sequencing new genomes

In our study we also noted a correlation between the presence of the rlrA islet and antibiotic resistance, suggesting its clinical relevance with respect to the development and spread of pneumococcal disease. The reasons for the apparent association between the rlrA pilus operon and antibiotic resistance are not clear. It could be that lineages that contain both the rlrA islet and specific resistance genes are more commonly subject to recombination and more likely to have taken up both elements [35]. Other scenarios may be suggested in which pili aid adhesion during the colonization of a nasopharynx devoid of other bacterial flora as a result of antibiotic treatment. Elucidation of this relationship is beyond the scope of the current work. We note however, that this association is another attractive feature of pilus proteins as a vaccine component, as they could provide protection against drug-resistant clones

This analysis provides the potential to predict the actual distribution of rlrA among a collection of isolates with defined genotypes, even given the current limits dictated by epidemiological surveillance. The efficacy of the pilus as a protective antigen furthermore supports a global effort to understand its epidemiology in the design of a protein-based vaccine

Acknowledgments

We would like to acknowledge those people who provided invaluable S. pneumoniae clinical isolates for this study: Active Bacterial Core surveillance ([ABCs] a collaboration between the Centers for Disease Control and Prevention [CDC], state health departments, and universities) http://www.cdc.gov/ncidod/dbmd/abcs/index.htm) and Bernard Beall from the CDC for supplying genotyped and serotyped ABCs strains; Annalisa Pantosti (ISS Italy); Albert Ko and Mitermayer G. Reis (Oswaldo Cruz Foundation, Brazil); Gerd Plushke (Swiss Tropical Institute); Samir Saha (Dhaka Shishu Hospital, Bangladesh); and Anthony Scott (Kenyan Medical Research Center, Kenya). We would like to thank Mogen Kilian (University of Aarhus, Denmark) for providing us with S. mitis genomic DNA. We would also like to thank Julian Parkhill from the Sanger Institute for rlrA sequence information and helpful discussions on mobile genetic elements, as well as the Novartis Sequencing Facility, Beatrice Rogolino, Silvia Guidotti, and Monica Giraldi for technical assistance

References

1
Ortqvist
A
Hedlund
J
Kalin
M
Streptococcus pneumoniae: epidemiology, risk factors, and clinical feature
Semin Respir Crit Care Med
 , 
2005
, vol. 
26
 (pg. 
563
-
74
)
2
Hanage
WP
Huang
SS
Lipsitch
M
, et al.  . 
Diversity and antibiotic resistance among nonvaccine serotypes of Streptococcus pneumoniae carriage isolates in the post-heptavalent conjugate vaccine er
J Infect Dis
 , 
2007
, vol. 
195
 (pg. 
347
-
52
)
3
Sogstad
MK
Hoiby
EA
Caugant
DA
Molecular characterization of non-penicillin–susceptible Streptococcus pneumoniae in Norwa
J Clin Microbiol
 , 
2006
, vol. 
44
 (pg. 
3225
-
30
)
4
Farrell
DJ
Klugman
KP
Pichichero
M
Increased antimicrobial resistance among nonvaccine serotypes of Streptococcus pneumoniae in the pediatric population after the introduction of 7-valent pneumococcal vaccine in the United State
Pediatr Infect Dis J
 , 
2007
, vol. 
26
 (pg. 
123
-
8
)
5
Sjostrom
K
Blomberg
C
Fernebro
J
, et al.  . 
Clonal success of piliated penicillin nonsusceptible pneumococc
Proc Natl Acad Sci U S A
 , 
2007
, vol. 
104
 (pg. 
12907
-
12
)
6
Streptococcus pneumoniae serogroups 15 and 33: an increasing cause of pneumococcal infections in children in the United States after the introduction of the pneumococcal 7-valent conjugate vaccin
Pediatr Infect Dis J
 , 
2006
, vol. 
25
 (pg. 
301
-
5
)
7
Pai
R
Moore
MR
Pilishvili
T
Gertz
RE
Whitney
CG
Beall
B
Postvaccine genetic structure of Streptococcus pneumoniae serotype 19A from children in the United State
J Infect Dis
 , 
2005
, vol. 
192
 (pg. 
1988
-
95
)
8
Barocchi
MA
Censini
S
Rappuoli
R
Vaccines in the era of genomics: the pneumococcal challeng
Vaccine
 , 
2007
, vol. 
25
 (pg. 
2963
-
73
)
9
Ogunniyi
AD
Grabowicz
M
Briles
DE
Cook
J
Paton
JC
Development of a vaccine against invasive pneumococcal disease based on combinations of virulence proteins of Streptococcus pneumonia
Infect Immun
 , 
2007
, vol. 
75
 (pg. 
350
-
7
)
10
Barocchi
MA
Ries
J
Zogaj
X
, et al.  . 
A pneumococcal pilus influences virulence and host inflammatory response
Proc Natl Acad Sci U S A
 , 
2006
, vol. 
103
 (pg. 
2857
-
62
)
11
Gianfaldoni
C
Censini
S
Hilleringmann
M
, et al.  . 
Streptococcus pneumoniae pilus subunits protect mice against lethal challeng
Infect Immun
 , 
2007
, vol. 
75
 (pg. 
1059
-
62
)
12
LeMieux
J
Hava
DL
Basset
A
Camilli
A
RrgA and RrgB are components of a multisubunit pilus encoded by the Streptococcus pneumoniae rlrA pathogenicity isle
Infect Immun
 , 
2006
, vol. 
74
 (pg. 
2453
-
6
)
13
Hemsley
C
Joyce
E
Hava
DL
Kawale
A
Camilli
A
MgrA, an orthologue of Mga, acts as a transcriptional repressor of the genes within the rlrA pathogenicity islet in Streptococcus pneumonia
J Bacteriol
 , 
2003
, vol. 
185
 (pg. 
6640
-
7
)
14
Nelson
AL
Ries
J
Bagnoli
F
, et al.  . 
RgrA is a pilus-associated adhesin in Streptococcus pneumonia
Mol Microbiol
 , 
2007
, vol. 
66
 (pg. 
329
-
40
)
15
McGee
L
McDougal
L
Zhou
J
, et al.  . 
Nomenclature of major antimicrobial-resistant clones of Streptococcus pneumoniae defined by the pneumococcal molecular epidemiology networ
J Clin Microbiol
 , 
2001
, vol. 
39
 (pg. 
2565
-
71
)
16
et al.  . 
Pre- and postvaccination clonal compositions of invasive pneumococcal serotypes for isolates collected in the United States in 1999, 2001, and 200
J Clin Microbiol
 , 
2006
, vol. 
44
 (pg. 
999
-
1017
)
17
Leimkugel
J
Adams Forgor
A
, et al.  . 
An outbreak of serotype 1 Streptococcus pneumoniae meningitis in northern Ghana with features that are characteristic of Neisseria meningitidis meningitis epidemic
J Infect Dis
 , 
2005
, vol. 
192
 (pg. 
192
-
9
)
18
Paterson
GK
Mitchell
TJ
The role of Streptococcus pneumoniae sortase A in colonisation and pathogenesi
Microbes Infect
 , 
2006
, vol. 
8
 (pg. 
145
-
53
)
19
Tettelin
H
Nelson
KE
Paulsen
IT
, et al.  . 
Complete genome sequence of a virulent isolate of Streptococcus pneumonia
Science
 , 
2001
, vol. 
293
 (pg. 
498
-
506
)
20
Enright
MC
Spratt
BG
A multilocus sequence typing scheme for Streptococcus pneumoniae: identification of clones associated with serious invasive diseas
Microbiology
 , 
1998
, vol. 
144
 (pg. 
3049
-
60
)
21
Feil
EJ
Li
BC
Aanensen
DM
Hanage
WP
Spratt
BG
eBURST: inferring patterns of evolutionary descent among clusters of related bacterial genotypes from multilocus sequence typing dat
J Bacteriol
 , 
2004
, vol. 
186
 (pg. 
1518
-
30
)
22
Press
WH
Teukolsky
SA
Vetterling
WT
Flannery
BP
Numerical recipes in FORTRAN: the art of scientific computing. 2nd ed
New York: Cambridge University Press
 , 
1992
23
Notredame
C
Higgins
DG
Heringa
J
T-Coffee: a novel method for fast and accurate multiple sequence alignmen
J Mol Biol
 , 
2000
, vol. 
302
 (pg. 
205
-
17
)
24
Kumar
S
Tamura
K
Nei
M
MEGA3: integrated software for molecular evolutionary genetics analysis and sequence alignmen
Brief Bioinform
 , 
2004
, vol. 
5
 (pg. 
150
-
63
)
25
Hausdorff
WP
Feikin
DR
Klugman
KP
Epidemiological differences among pneumococcal serotype
Lancet Infect Dis
 , 
2005
, vol. 
5
 (pg. 
83
-
93
)
26
Maiden
MC
Bygraves
JA
Feil
E
, et al.  . 
Multilocus sequence typing: a portable approach to the identification of clones within populations of pathogenic microorganism
Proc Natl Acad Sci U S A
 , 
1998
, vol. 
95
 (pg. 
3140
-
5
)
27
Hava
DL
Hemsley
CJ
Camilli
A
Transcriptional regulation in the Streptococcus pneumoniae rlrA pathogenicity islet by Rlr
J Bacteriol
 , 
2003
, vol. 
185
 (pg. 
413
-
21
)
28
Hava
DL
Camilli
A
Large-scale identification of serotype 4 Streptococcus pneumoniae virulence factor
Mol Microbiol
 , 
2002
, vol. 
45
 (pg. 
1389
-
406
)
29
Zhou
L
Hui
FM
Morrison
DA
Characterization of IS1167, a new insertion sequence in Streptococcus pneumonia
Plasmid
 , 
1995
, vol. 
33
 (pg. 
127
-
38
)
30
Clinical and Laboratory Standards
Institute
Performance standards for antimicrobial susceptibility testing: 16th informational supplement, document M100-S16
Wayne, PA: Clinical and Laboratory Standards Institute;
 , 
2006
31
Basset
A
Trzcinski
K
Hermos
C
, et al.  . 
Association of the pneumococcal pilus with certain capsular serotypes but not with increased virulenc
J Clin Microbiol
 , 
2007
, vol. 
45
 (pg. 
1684
-
9
)
32
Turner
KM
Hanage
WP
Fraser
C
Connor
TR
Spratt
BG
Assessing the reliability of eBURST using simulated populations with known ancestr
BMC Microbiol
 , 
2007
, vol. 
7
 pg. 
30
 
33
Lauer
P
Rinaudo
CD
Soriani
M
, et al.  . 
Genome analysis reveals pili in Group B Streptococcu
Science
 , 
2005
, vol. 
309
 pg. 
105
 
34
Mora
M
Bensi
G
Capo
S
, et al.  . 
Group A Streptococcus produce pilus-like structures containing protective antigens and Lancefield T antigen
Proc Natl Acad Sci U S A
 , 
2005
, vol. 
102
 (pg. 
15641
-
6
)
35
Hsieh
YC
Wang
JT
Lee
WS
, et al.  . 
Serotype competence and penicillin resistance in Streptococcus pneumonia
Emerg Infect Dis
 , 
2006
, vol. 
12
 (pg. 
1709
-
14
)
Presented in part: “Novel Opportunities to Develop Vaccines to Control Antibiotic Resistant Bacteria: From the Trials Back to the Laboratory” (REBAVAC) Workshop, March 2007, Siena, Italy (abstract A7)
Potential conflicts of interest: M.M., C.D., A.M., V.M., S.C., A.C., R.R., and M.A.B. are employed by Novartis Vaccines. All other authors report no potential conflicts