Pneumococcus with the “6E” cps Locus Produces Serotype 6B Capsular Polysaccharide

Abstract

Genetic studies of serogroup 6 isolates of Streptococcus pneumoniae identified putative serotype 6E. Although its capsular polysaccharide structure has not been elucidated, putative serotype 6E is described in an increasing number of studies as a potentially new serotype. We show here that SPEC6B, which is widely used as a target strain for serotype 6B opsonophagocytosis assays, has the genetic features of the putative serotype 6E but produces capsular polysaccharide identical to 6B capsular polysaccharide as determined by one-dimensional (1D) and 2D nuclear magnetic resonance (NMR). Thus, putative serotype 6E is a mere genetic variant of serotype 6B. Also, SPEC6B is appropriate as a target strain for serotype 6B opsonophagocytosis assays. This example illustrates the difficulties of assigning new bacterial serotypes based on genetic findings alone.

INTRODUCTION

Streptococcus pneumoniae is a Gram-positive bacterial species capable of producing more than 90 distinct capsule types (1). Due to the significance of S. pneumoniae as a human pathogen, the capsule diversity of pneumococci has been extensively studied serologically, biochemically, and genetically (1). In 2006, the genetic loci for capsule synthesis (cps loci) were sequenced for all known pneumococcal capsule types (2). Subsequent studies have found that genetic information in the cps locus correlates well with serologic and biochemical data for the capsular polysaccharide (PS). Thus, several well-described PCR protocols have been developed for serotyping purposes (e.g., see reference 3 and http://www.cdc.gov/streplab/downloads/pcr-oligonucleotide-primers.pdf; reviewed in reference 1), and it has become common to use genetic information to determine capsule serotypes.

Genetic studies of serotype 6B cps have revealed two subgroups of cps loci that differ by >5% in the cps locus (4). The major subgroup was called class 1; the minor subgroup was called class 2 and has an ∼300-bp indel element between wciN and wciO as its genetic hallmark. In addition, the class 2 locus contains a 9-nucleotide (nt) in-frame deletion in wze and four open reading frames upstream of wzg (5). Because of the stark genetic difference from class 1, it was suggested that class 2 may be a new serotype, “serotype 6E” (6). Although the authors of this study concluded that “serologic and biochemical characterization” is needed to confirm it as a new serotype (6), the putative serotype 6E is being increasingly treated as a distinct serotype in publications describing its discovery in all parts of the world (5, 7–10) and its association with antibiotic resistance (7, 10). Consequently, a recent study stated that there is an urgent need to determine the structure of capsular polysaccharide from the putative serotype 6E (10).

Recently, several serotype 6B reference strains from the Pneumococcal Molecular Epidemiology Network collection were found to have “serotype 6E” genetic loci (10). Furthermore, SPEC6B, which is used as the serotype 6B reference target strain for the multiplexed opsonophagocytosis assay (MOPA) (11), was found to have the genetic features of “serotype 6E” (unpublished information). As SPEC6B is extensively used in evaluating pneumococcal vaccine immunogenicity (12–14), it became necessary to examine the PS structure made by SPEC6B. Here, we demonstrate that SPEC6B has genetic features of “serotype 6E” but produces PS identical to that of serotype 6B in molecular structure.

MATERIALS AND METHODS

Genetic analysis by PCR.

To detect the presence of indel, PCR was performed using two previously described primer sets (forward primers 5106F [5′-TACCATGCAGGGTGGAATGT] and 5101F [5′-ATTTGGTGTACTTCCTCC] in independent reactions with primer 3101R [5′CCATCCTTCGAGTATTGC]) (8) and genomic DNA from SPEC6B, MNZ2, and DS2212-94. SPEC6B has been described previously (11), DS2212-94 came from the CDC (Atlanta, GA), and MNZ2 came from K. H. Kim (Ewha Womans University, Seoul, South Korea). DS2212-94 has a class 1 cps sequence (unpublished information), and we have previously determined that MNZ2 has a class 2 cps sequence. DS2212-94 and MNZ2 were included as the serotype 6B and “serotype 6E” control strains, respectively.

DNA sequencing of the SPEC6B cps locus.

The SPEC6B cps locus was amplified in fragments and sequenced. A portion of the DS2212-94 cps locus (containing wze-wciP) was also sequenced. Sequences were compared to reference sequences and assembled: for SPEC6B, GenBank accession number AF246897 was used; for DS2212-94, GenBank accession number CR931639 was used. Divergences from the reference sequence were confirmed by sequencing of an independent PCR.

Polysaccharide purification.

For each indicated strain, 1-liter cultures were prepared in chemically defined medium (JRH Biosciences, Lenexa, KS) supplemented with choline chloride (1 g/liter), sodium bicarbonate (2.5 g/liter), and cysteine HCl (0.73 g/liter). After overnight incubation at 37°C, lysates were prepared by adding deoxycholate (0.05% final concentration), adjusting the pH to ∼7 and incubating the solution for ∼1 h at room temperature (RT) with stirring. After centrifugation (15,344 × g for 20 min) to remove debris, contaminants were precipitated from the supernatants by sequential incubations (48 h at 4°C) with 30% and 50% ethanol, with centrifugation (15,344 × g for 20 min) after each incubation to remove debris. PS was precipitated from the supernatants by incubation for 48 h at 4°C in 80% ethanol. The precipitate was collected by centrifugation (15,344 × g), suspended in 100 ml of water, and transferred to dialysis tubing (3,500-molecular-weight cutoff). After dialysis against water, a final dialysis was performed against 5 mM Tris, pH 7.4.

The dialyzed product was applied to a DEAE Sepharose (GE Healthcare, Uppsala, Sweden) column (∼60 ml). After a wash with >3 volumes of 5 mM Tris, pH 7.4, elution was performed using a NaCl gradient ranging from 0 M to 1 M over 48 fractions (∼5 ml/fraction).

PS was detected in the fractions using an inhibition enzyme-linked immunosorbent assay (ELISA) in which plates were coated with 6B PS (ATCC, Manassas, VA) and washed, and inhibitors (fractions or 6B PS from ATCC as a standard) were added. After addition of a monoclonal antibody against 6B (Hyp6BM8 [15]), plates were incubated at RT for 30 to 60 min. Plates were washed, and an anti-mouse Ig-alkaline phosphatase (Sigma-Aldrich, St. Louis, MO) was added. After incubation at RT for 30 to 60 min, plates were washed and a phosphatase substrate (Sigma-Aldrich) was added. After incubation at RT for 30 to 60 min, the optical density at 405 nm was determined. The concentration of the PS in the fractions was estimated using purified 6B PS (ATCC) as a standard.

Based on PS concentration and absorbance at 260 and 280 nm, fraction 21 of both preparations was chosen for further analysis. After dialysis against water, each product was lyophilized and suspended in D₂O for nuclear magnetic resonance (NMR) analysis.

NMR spectroscopy.

NMR data were collected on a Bruker Avance II (700 MHz; ¹H) spectrometer equipped with a cryogenic triple-resonance probe. NMR experiments were executed as described previously (16). Briefly, ∼5 mg of lyophilized PS was dissolved in 0.5 ml of D₂O. One-dimensional (1D) ¹H and two-dimensional (2D) ¹H-¹³C heteronuclear multiple quantum coherence (HMQC) data were obtained at 45°C, processed with NMRPIPE (17), and analyzed with NMRVIEW (18). Purified 6A PS (Statens Serum Institute, Denmark) and 6B PS (ATCC) were included as controls.

Nucleotide sequence accession numbers.

Sequences of the SPEC6B cps locus and partial DS2212-94 cps locus were deposited in GenBank under accession numbers KT907353 and KU168827, respectively.

RESULTS

In order to determine if the SPEC6B cps locus contained the indel element, PCRs were performed using previously described primers (8). We performed separate reactions rather than the previously reported combined reaction, and the amplicon sizes predicted by the serotype 6B reference sequence (GenBank accession no. CR931639) were 2,009 bp and 957 bp rather than the single 1,800-bp product reported by Baek et al. for serotypes 6C and 6D (8). Figure 1A depicts the binding site for each of the primers as well as the predicted size of the products of each primer set. The PCR products were separated and visualized on an agarose gel. As shown in Fig. 1B, the PCR product obtained with each primer set was consistent in size between SPEC6B and MNZ2, and both were slightly larger than the product for DS2212-94. This indicates that SPEC6B likely contains the indel in its cps. A previously reported PCR scheme (8) was utilized to detect the additional nucleotides upstream of wzg associated with the “6E” cps locus (∼2.7 kbp); the product of SPEC6B and a “6E” control were of similar size, larger than the DS2212-94 product, and consistent with the reported “6E” size (data not shown).

FIG 1.

Electrophoresis of PCR products. (A) Genetic arrangement, primer numbers (from reference 8), and the predicted product sizes for each of the two PCRs that were performed to detect the indel element between wciN and wciO (8). The expected results for a serotype 6B (class 1) cps locus and a “serotype 6E” (class 2) cps locus are shown. (B) Agarose gel of the PCR products for each of the three strains. DS2212-94 and MNZ2 were included as controls.

To unambiguously show the presence of the indel and other genetic features of “6E” (e.g., 9-nucleotide [nt] deletion in wze), the entire cps locus of SPEC6B was sequenced and deposited in GenBank (accession number KT907353). When the SPEC6B cps locus was compared to a known “serotype 6E” cps locus (a class 2 serotype 6B locus, GenBank accession number AF246897 [4]), the same 9-nt in-frame deletion in wze was found to be present in both strains, and the sequences from wciN to wciO (which includes the indel sequence) were identical except for a same-sense mutation in the 3′ end of wciN. In contrast, nucleotide identity between SPEC6B and a serotype 6B reference sequence (GenBank accession no. CR931639) was only 92.9% over their shared regions, similar to other observations (7). For comparison, the cps locus of DS2212-94 was partially sequenced and deposited in GenBank (accession number KU168827). In contrast to SPEC6B, DS2212-94 sequence did not exhibit the wze deletion or contain the indel sequence and had characteristic features of a class 1 6B cps locus.

In order to determine the structure of the capsular PS, ion-exchange chromatography was utilized to purify capsular PS from SPEC6B and DS2212-94. The ion-exchange elution chromatographs are shown in Fig. 2 for PS from SPEC6B (upper panel) and DS2212-94 (lower panel). The PS concentrations in fractions 13 to 36 (identified in pilot experiments as the PS-containing fractions) were determined and are indicated with the solid black lines, using the y axis on the left side of the graph. The absorbances at 260 nm and 280 nm were also determined for fractions 13 to 30 (covering the PS peak) and are indicated by the dashed and dotted lines, respectively, using the y axis on the right side of the graph. The solid gray line in each graph represents the NaCl gradient used for elution, ranging from 0 M (at fraction 1) to 1.0 M (at fraction 50). Based on the PS concentration and absorbances at 260 nm and 280 nm, fraction 21 was selected for NMR analysis for both PSs.

FIG 2.

Ion-exchange elution chromatographs. For SPEC6B (upper panel) and DS2212-94 (lower panel), bound materials were eluted from the DEAE ion-exchange resin using a NaCl gradient (ranging from 0 M to 1 M, gray line). For the indicated fractions, the PS concentration (solid black line), absorbance at 260 nm (dashed line), and absorbance at 280 nm (dotted line) were determined.

The ¹H NMR spectra for SPEC6B PS, DS2212-94 PS, 6B PS, and 6A PS are shown in Fig. 3. The spectra of PSs from SPEC6B and DS2212-94 are identical to that of the 6B PS, indicating that these PSs have the same chemical structure but are significantly different from 6A PS. The PS from SPEC6B and DS2212-94 were further analyzed by 2D NMR. Figure 4 shows the overlay of the ¹H-¹³C HMQC spectra of PS from SPEC6B (red) and DS2212-94 (black). Again, their spectra are identical, indicating that the PSs have the same chemical structure.

FIG 3.

1D NMR spectra. Shown are ¹H NMR spectra of capsular polysaccharide purified from SPEC6B, DS2212-94, serotype 6B, and serotype 6A. Serotype 6A (Statens Serum Institute [SSI]) and 6B (ATCC) are purified 6A and 6B PSs from SSI and the ATCC, respectively. Note the major difference in the anomeric region (∼5.0 to 5.2 ppm) between serotypes 6B and 6A.

FIG 4.

2D NMR spectra. Shown is an overlay of ¹H-¹³C HMQC spectra of PS from SPEC6B (red) and DS2212-94 (black). Each peak has been assigned as described previously.

DISCUSSION

Recent studies have led to an urgent need for the determination of the “serotype 6E” capsular PS structure (10). Here we demonstrate that strains DS2212-94 and SPEC6B have genetic hallmarks for serotypes 6B and “6E,” respectively, yet their capsular PSs have identical 2D NMR spectral patterns. Thus, serotypes “6E” and 6B produce PSs with identical chemical structures and are expected to have identical serologic properties. Indeed, no serologic differences between serotypes 6B and “6E” have been detected (10).

Our finding further demonstrates limitations in using genetic data to identify a new pneumococcal “serotype.” Multiple studies have shown that different strains of the same serotype can have significant genetic differences in their cps loci. For instance, many serotype 11A strains isolated from the United States and Europe have large differences in their cps DNA sequences, resulting in misidentification of some 11A isolates as 11D in the past (19–21). Also, serotype 19A cps loci can be significantly different among different isolates (22). Here, we report that class 1 and class 2 serotype 6B isolates produce the same capsular PS despite striking differences in their cps loci.

Conversely, small genetic changes can alter the functioning of enzymes involved in capsular PS production, resulting in changes in PS structure. For instance, it was shown that serotype 6A and 6B differ genetically by three nucleotides out of 17 kb of DNA in the serogroup 6 cps locus (4). Similarly, serotypes 11A and 11D can differ by one nucleotide (21). Recently, we identified single nucleotide differences in the serotype 6A or 6B cps locus leading to the production of hybrid serotypes, which we named 6F, 6G, and 6H (16, 23). Thus, the magnitude of genetic differences is not a good predictor of a new serotype.

We believe that the limitations associated with the use of genetic information to define a new serotype should not be limited to pneumococci but should be extended to other bacterial species. For example, a Neisseria meningitidis strain expressing capsule type Y can become a capsule type W-135 strain following a single amino acid change (24). Also, a single amino acid change in a sialyltransferase, Cst-II, can alter Campylobacter jejuni lipooligosaccharide (LOS) structure (25). C. jejuni LOS can mimic ganglioside structure, and antibodies to C. jejuni LOS can induce Guillain-Barré syndrome (25). Interestingly, the small change in Cst-II is associated with Guillain-Barré syndromes with different neurological manifestations (25).

Taken together, the data show that putative “serotype 6E” is identical to serotype 6B in serologic properties (10) and in chemical structure (this study). Thus, it should be described as a genetic variant of serotype 6B and expressed as serotype 6B with class 2 cps locus. Also, SPEC6B can continue to be used as a target strain for serotype 6B, and the use of “serotype 6E” should be discontinued. The results of this study clearly demonstrate the potential pitfalls when a genetic variant is assumed to be a new serotype. A recent review (1) proposes that a new serotype should have unique serologic properties, a stable genetic difference, and a distinct biochemical structure.

It is worth noting that serotype 6A isolates containing a class 2 cps locus have also been identified (4, 5, 7), albeit at a lower frequency (7); thus, one should expect that any serogroup 6 member can exhibit the class 2 cps locus. Nevertheless, it is possible that while the capsules produced by class 2 loci are identical to their class 1 counterparts, the class 2 genetic elements may modulate the regulation or amount of capsular PS produced. More importantly, observations of the class 2 cps locus may be useful in studies of capsule origins or epidemiological studies such as those investigating antibiotic resistance.