Introduction

Streptococcus pneumoniae is a leading causative pathogen responsible for acute otitis media (AOM) that frequently colonizes the nasopharynx [3, 6]. This pathogen has long been susceptible to penicillin, and AOM caused by pneumococci are easily improved by oral antimicrobial therapy. However, recent dramatic increases of antimicrobial resistance in S. pneumoniae are making the treatments of AOM with oral antibiotics more difficult [5, 8, 9]. There are urgent demands to prevent pneumococcal AOM through vaccinations. Naopharyngeal colonization with causative pathogens is one of the more important risk factors for developing AOM. Consequently, reducing the frequency of nasopharyngeal carriage of S. pneumoniae is an important step towards preventing the development of AOM. However, there is less evidence for preventing AOM by vaccines [19]. A newly developed 10-valent vaccine conjugated with H. influenzae protein D shows more efficacy for reducing nasopharyngeal carriage [13]. Nevertheless, recent reports have shown a decrease in the carriage of vaccine serotypes and a parallel increase in non-vaccine serotypes following vaccination [1, 14, 20].

In order to be able to carry out a comprehensive evaluation of vaccine efficacies, it is first necessary to understand the prevalence of vaccine serotypes as well as the antimicrobial resistances of pneumococci associated with AOM. The determination of serotypes by immunological methods requires the isolation of pneumococci, is time consuming and expensive, and the results are difficult to interpret. In an earlier publication, we reported the successful applications of the multiplex PCR approach for determining pneumococcal serotypes [2]. We also showed that there was a correlation between the gene mutation and the minimal inhibitory concentration (MIC) to beta-lactam antibiotics. However, in that study, the gene mutation was investigated using isolated pneumococci [6]. To date, there has been no study which simultaneously evaluates the pneumococcal serotypes and gene types present in nasopharyngeal secretions (NPS). We hypothesized that the PCR technique could be used for identifying serotyping, and genotyping pneumococci present in NPS, thereby also enabling us to determine whether more than one serotype can co-exist there.

The aims of this study were (1) to simultaneously identify and determine serotypes of S. pneumoniae in the NPS of acute otitis media (AOM) patients, (2) to determine the genotypes of penicillin-binding proteins (PBPs) and macrolide-resistant traits in NPS of AOM patients and, in addition, (3) to examine whether one or more serotypes of S. pneumoniae may co-exist as nasopharyngeal flora in patients.

Materials and methods

Nasopharyngeal secretions

A total of 60 NPS were collected from pediatric patients (12–60 months old) with AOM at the clinics of Otolaryngology – Head and Neck Surgery, Wakayama Medical University Hospital. The NPS samples were collected by suction using a fine, flexible plastic catheter (no 5, French) and syringe. Informed consent was obtained from the parents or guardians of the patients prior to the collection of samples, in accordance with the guidelines of the institutional review board of Wakayama Medical University.

Bacterial culture

A portion of each of the 60 NPS samples was cultured on 5% sheep blood agar plates and chocolate agar plate (Nippon Becton Dickinson, Tokyo, Japan) for 48 h at 37°C in a humidified atmosphere suppled with 5% CO2. S. pneumoniae was identified on the basis of alpha-hemolysis and colony morphology, Gram-stained smear, optochin disk sensitivity and bile solubility. Determinations of the MICs to penicillin G (PCG) and clarithromycin (CAM) were performed using CLSI methods [12], and the serotypes were determined using a standard laboratory method. Briefly, a bacterial suspension was mixed with group-specific or type-specific antisera (The Statens Serum Institute, Copenhagen, Denmark) [15]. The quellung and agglutination were assessed by phase-contrast microscopy.

Preparation of genomic DNA

Total genomic DNA was purified from both NPS and S. pneumoniae isolates. Prior to purification of the genomic DNA from the NPS, the samples were diluted three times with sterilized saline and then centrifuged to remove inhibitory substances. The NPS pellets and a single colony of pneumococci were then digested with a lysis solution [1 M Tris, pH 8.9, 4.5 (v/v) nonident P-40, 4.5 (v/v) Tween 20, 10 mg/ml Proteinase K) for 1 h at 60°C. Following centrifugation, the supernatant was mixed with 3 M sodium acetate and the total genomic DNA was precipitated by ethanol.

PCR-based genotyping

Seven sets of oligonucleotide primers were used to amplify pbp genes (pbp1a, pbp2x, pbp2b), macrolide-resistant genes (mefA and ermB), the pneumococcal common autolysin gene (lytA) and the pneumolysin gene (ply) (Table 1) [16–18]. Each primer mixture contained multiplex PCR5, multiplex PCR6, multiplex PCR7 and PCR8 primers for lytA and pbp1a, pbp2x and pbp2b, mefE and ermB, and ply, respectively. The PCR reaction mixtures consisted of 1 μl of bacterial lysate or purified DNA from NPs, 8 μl 25 mM of dNTP mixture, 2.5 U of Tth DNA polymerase (Takara Biomedicals, Kyoto, Japan), 10 μl 10× PCR buffer, pH 8.3 and 60 ng of primer mixtures (multiplex PCR5, multiplex PCR6, multiplex PCR7 and PCR8) in 100 μl of solution. The reaction mixture was then subjected to 30 cycles (20 s at 94°C, 20 s at 55°C and 15 s at 72°C) of amplification in the programmable thermal cycler (Gene Amp PCR System 9700; Perkin-Elmer, Norwalk, Conn.). Amplified DNA fragments were analyzed using 3% agarose gel electrophoresis (Fig. 1).

Table 1 Primers used in this study
Full size table
Fig. 1
figure 1

PCR-based genotypes of seven clinical isolates of Streptococcus pneumoniae. MW Molecular-weight marker (100 bp).Strain:1 No mutation in pbp genes,2 mutation in pbp2b,3 mutation in pbp2x,4 mutations in pbp2b and pbp2x,5 mutations in pbp1a and pbp2b,6 mutations in pbp1a and pbp2x,7 mutations in pbp1a, pbp2x and pbp2b,8 ermB-positive,9 ermB- and mefE-positive,10 mefE-positive

Full size image

PCR-based serotyping

The oligonucleotide primers were used to amplify capsular serotypes 1, 3, 4, 14, 19A, 19F, 23F and serogroups 6, 18,19 and 23 for multiplex PCR-based serotyping (Table 1) [4, 10]. The PCR reaction mixtures consisted of 1 μl of bacterial lysate or purified DNA, 0.5 μl 10 mM of dNTP mixture, 0.5 μl Taq DNA polymerase, 2.5 μl 10× PCR buffer, 4 μl 25 mM MgCl2, 5.0 μl Q-solution (Qiagen, Hilden, Germany), 1.25 μl 1 M KCl and 0.25 μl of each primer sets (multiplex PCR1, multiplex PCR2, multiplex PCR3, multiplex PCR4) in 25 μl of solution. For serotyping, the reaction mixture was subjected to denaturation at 94°C for 10 min, followed by 32 cycles of amplification consisting of denaturation at 94°C for 30 s, annealing at 61°C for 30 s and extension at 72°C for 1 min and a further extension at 72°C for 5 min for serotyping. Strains of serotype 4 (ATCC BAA-334) and 19F (ATCC 49619) obtained from the American Type Culture Collection (ATCC, Manassas, Va.) were used for quality control in every reaction (Fig. 2).

Fig. 2
figure 2

A multiplex of PCR-based serotypes.MM Molecular-weight marker.MP1 Serogroup 6, serogroup 19, serogroup 18 and serogroup 23,MP2 serotype 1 and serotype 19F,MP3 serotype 3, serotype 14 and serotype 23F,MP4 serotype 4 and serotype 19A

Full size image

Results

Identification of S. pneumoniae in NPS

S. pneumoniae were identified in 30 (50.0%) and 36 (60.0%) of the 60 NPS samples tested by conventional culture and PCR analysis, respectively. Not only were culturable S. pnuemoniae identifiable by conventional bacterial culture methodology and PCR analysis, but even samples containing pneumococci that were unidentifiable by culture techniques showed amplifications of both lytA and ply with PCR analysis.

Based on the CLSI criteria, the pneumococcal isolates consisted of 12 penicillin-sensitive S. pneumoniae (PSSP; 40.0%), 11 penicillin-intermediate S. pneumoniae (PISP) (36.7%) and seven penicillin-resistant S. pneumoniae PRSP; 23.3%). They were also classified into six (20.0%), one (3.33%) and 12 (76.67%) strains that showed sensitivity, intermediate resistance and resistance to CAM, respectively. Based on the serology, the isolates were classified into serotypes 3 (1; 3.3%), 14 (3; 10.0%), 19F (5; 16.7%), 23F (6; 20.0%), 6B (10; 33.3%), 6A (1; 3.3%), 15 (1; 3.3%), 9V (1; 3.3%) and non-11 (2; 6.7%) (Table 2). All serotype 23F strains and three (60.0%) 19F strains were resistant to both PCG and CAM.

Table 2 Results of the multiplex PCR and conventional culture analyses in terms of detecting, genotyping and serotyping strains of S. pneumoniae in pneumococcal-positive nasopharyngeal secretions
Full size table

PCR-based genotypes

Among 36 NPS samples, 19 (52.8%) samples contained strains with mutations in the three PBP genes (Table 2), eight (22.2%) contained either strains with mutations in pbp1a and pbp2x (four samples; 11.1%) or in pbp2x and pbp2b (four samples; 11.1%), eight (22.2%) contained either strains with a mutation in pbp1a (one sample; 2.8%), in pbp2b (one sample; 2.8%) and in pbp2x (six samples; 16.7%) and only one (2.8%) contained the strain without mutations in the three PBP genes.

Susceptibilities to PCG of strains with mutations in three PBP genes, in two types of PBP genes, in one type of PBP gene and without mutations in the PBP genes were 0.12–2, ≤0.06–1, ≤0.06 and ≤0.06 μg/ml, respectively. Twenty-six (72.2%) samples contained strains with macrolide-resistant genes: 15 (44.4%) samples with strains possessing mefE, 13 (19.4%) strains possessing ermB and three (8.3%) strains possessing both macrolide-resistant genes, respectively. Ten (27.8%) samples contained strains having neither type of macrolide-resistant gene. Susceptibilities to CAM of strains with the mefE gene, ermB gene and both were 1–4, >64 and >64 μg/ml, respectively.

PCR-based serotypes

The multiplex PCR was able to determine the serotypes/serogroups of 36 NPS samples; in contrast, conventional methodology determined the serotype/serogroups in only 30 samples (Table 2); these were serotypes 3 (1; 2.8%), 14 (3; 8.3%), 19F (7; 19.4%), 23F (10; 27.8%) and 6 (11; 30.6%) as well as seven ‘others’ (19.4%). With the exception of the typeable pneumococci, all of the serotypes identified by multiplex PCR showed a similar quellung reaction with antisera. Four untypeable samples by multiplex PCR were classified into serogroups 15 (one), 9V (one) and non-11 (two) types by serological determination. In this study we compared the ratio of accurate determination of serotypes by both procedures. The correct determination of the serotypes when both procedures showed same results was determining. If one procedure failed to identify serotypes determined by the other procedure, the cases were defined as false determinations. The multiplex PCR determined the pneumococcal serotypes in 35 (97.2%) of the 36 NPS samples, while a conventional serological method determined the serotypes in 29 (80.5%) of the 36 NPS samples (p < 0.05). Two NPS samples were culture-negative but contained pneumococcal genomes. One sample (no. 8) showed serogroup 23 by multiplex PCR, but showed serotype 6B by serological determination. Two different serotypes were determined concurrently in two NPS by multiplex PCR. The positive and negative predictive value of PCR based on the serotyping of NPS compared with the serological method was 96.3% (26/27 samples) and 100%, respectively.

The multiplex PCR was able to determine the serotypes/serogroups of four NPS samples in which pneumococci were not identified by a conventional bacterial culture and in which it had not been possible to determine pneumococcal serotypes/serogroups.

Discussion

In this era of antimicrobial resistant S. pneumoniae and vaccine development against this pathogen, it is important to be able to evaluate the characteristics of the pneumococci colonizing the nasopharynx [3, 5, 6]. However, the current serological determinations of capsular serotypes are difficult to interpret in terms of the quellung reactions of pneumococcal cells with anti-polysaccharide sera [2]. Consequently, attempts have made to determine the accurate pneumococcal characteristics by molecular biological procedures. Earlier studies in our laboratory determined that the application of PCR-based serotyping and genotyping to pneumococcal isolates facilitates the identification of pneumococcal isolates [2] and that the multiplex PCR serotyping approach is able to determine particular types of capsular polysaccharides accurately. PCR-based genotyping also revealed that the frequencies of mutations in pbp and the expressions of mefA and ermB are closely related with susceptibilities to β-lactams and macrolides [6]. However, most of the studies applied PCR to pneumococcal isolates – and not directly to clinical samples [4, 10]. In the present study, we applied the multiplex PCR technique directly to NPS specimens and were able to identify and determine the characteristics of 11 pneumococcal serotypes and serogroups simultaneously. Those identified consisted of most of the serotypes used for current pneumococcal vaccines. We also applied PCR-based genotyping to the pbp genes and macrolide resistance genes.

Serogroups 6, 19F or 23F were the predominant capsular types in this study. Approximately 52.8% of the NPS samples contained pneumococci having mutations in the three pbp genes and could be classified as PRSP according to the CLSI criteria. Approximately 72.2% of the NPS samples contained pneumococci expressing either type of macrolide-resistant trait and were resistant to CAM. Most of these strains with mutations in the three pbp genes belonged to serogroups 6, 19F or 23F. The multiple PCR analysis was able to determine the pneumococci and the pneumococcal serotypes and serogroups isolated from the NPS samples more accurately than conventional culture and immunological methodology. In six samples, pneumococcal genomes were identified by the multiplex PCR, while a conventional culture method failed to identify pneumococci. In two cases, strains of serotypes 19F and 23F were simultaneously identified in the NPS, although only one strain of serotype 19F was identified by a conventional culture method. Recent reports on vaccine efficacies against nasopharyngeal colonization with pneumococci have shown that the pneumococci in the nasopharynx can change from vaccine serotypes to non-vaccine serotypes following vaccination [1, 14, 20]. The underlying factors determining these changes are still unclear, although one hypothesis is a replacement of pneumococcal strains in the nasopharynx. Nasopharygeal pneumococcal flora consist of several different strains, with minor pneumococcal populations present in the nasopharynx. Current conventional serological methodology requires the isolation of the pneumococcal strains. Huebner et al. suggested that if the less common serotype represents only 5% of the total pneumococcal population, 59 colonies from each specimen would need to be serotyped to have a 95% probability of picking the second pneumococcal type [7]. In addition, non-PCR methods can be prone to mis-interpretation. In the present study, a strain was determined to be serogroup 23 by the multiplex PCR, while based on the serology, the strain was serotype 6B. It is somewhat difficult to interpret the quellung changes and the serological determination sometimes failed to determine the accurate prevalence of pneumococci in the nasopharynx.

In this study, we have analyzed a limited number of serotypes by the multiplex PCR techniques because of the unavailability of the sequences of other pneumococcal serotypes. However, multiplex PCR-based serotyping can provide an accurate and rapid distribution of pneumococcal serotypes, including minor populations, among the S. pneumoniae populations of the nasopharynx. The multiplex PCR method does not need expertise to interpret the results and can be used to run many samples at one time. Moreover, multiplex PCR-based serotyping and genotyping can replace conventional microbiological methods that are used to identify and determine S. pneumoniae, pneumococcal capsular serotype, penicillin susceptibility and macrolide resistance traits. A follow-up study involving the quantitative evaluation by real-time PCR is necessary, and continuous monitoring of pneumococcal serotypes is essential since it has been shown that the incidence of types responsible for AOM can change over time [11]. We have shown that the direct application of PCR to the NPS enables a feasible analysis of minority strains in a pneumococcal population and confirmed this by studies on the carriage of multiple pneumococcal capsular types.