Genetic Basis for the New Pneumococcal Serotype, 6C

Abstract

We have recently reported a new pneumococcal serotype (6C), which is closely related to serotype 6A (I. H. Park et al., J. Clin. Microbiol. 45:1225-1233, 2007). To investigate the genetic basis for serotype 6C, we studied the capsule gene loci of 14 6C isolates from three different continents, including one isolated in Alabama 27 years ago. The wciN region of all 6C isolates has a 1,029-bp-long sequence that replaces the 1,222-bp-long sequence of the 6A wciN region. This recombination event has created a new 1,125-bp-long open reading frame which encodes a product that is also homologous to glycosyl transferases. Flanking this introduced gene is 300 bp upstream and 100 bp downstream with only about 90% homology with 6A and which is identical in all 6C isolates. Transfer of the wciN region converts 6A to 6C. Determination of the DNA sequence of the entire capsule gene locus of one 6C isolate showed that the 6C capsule gene locus is almost identical (>98% homologous) to that of 6A except for the wciN region. These findings indicate that the 6C capsule type originated more than 27 years ago by a single recombination event in a 6A locus in which 6A wciN was replaced by a gene of unknown origin.

Streptococcus pneumoniae is a major human pathogen that can cause various diseases, including pneumonia, otitis media, meningitis, and sepsis (9). Pneumococci produce at least 91 different types (serotypes) of the polysaccharide (PS) capsule (11, 19), which shields pneumococci from host phagocytes and is required for virulence (1). Antibodies to the capsular PS can abrogate the shielding effect of the capsule and thereby provide a host with capsule-type-specific protection (6). Consequently, currently available pneumococcal vaccines are designed to induce the production of antibodies to capsular PS.

Among the 91 different serotypes, certain serotypes are clinically important and have been extensively studied. For instance, 6A and 6B serotypes accounted for 4.6% and 10.6% of pneumococcal meningitis or sepsis in young children before the introduction of a pneumococcal conjugate vaccine (24). Extensive biochemical studies since their discovery in 1929 (7) have shown the 6A/6B PS to be isopolymers containing galactose-glucose-rhamnose-ribitol-phosphate, differing only in the rhamnose-ribitol linkage (12, 22). Genetic studies of the capsule loci from many isolates showed that the capsule loci of 6A and 6B are largely identical and that only one nucleotide in wciP (G584A; S195N) is consistently different between the capsule loci of the two serotypes (16). This suggests that the mutation at position 584 is responsible for the difference between 6A and 6B serotypes.

Notwithstanding such intensive studies, we have recently identified two subtypes in pneumococci that were serotyped as “6A” by classical quellung reactions (14, 19). Chemical studies of the capsular PS showed that the major subtype produces capsular PS matching the 6A PS structure described in the literature (galactose-glucose-rhamnose-ribitol-phosphate) and, thus, the major subtype kept the serotype 6A designation (19). In contrast, the repeating unit of the minor subtype was found to be glucose-glucose-rhamnose-ribitol-phosphate, and the minor subtype was named serotype 6C (19). In the present study, we describe the genetic differences between 6A and 6C.

MATERIALS AND METHODS

Bacterial strains and culture.

The pneumococcal strains used in this study are listed in Table 1. In addition to the 6C isolates from Brazil that were reported earlier (14), we identified additional 6C strains by retyping the preexisting pneumococcal isolates archived in our laboratory as the “6A” serotype. One isolate (BGO-2197) was obtained in 1979 from the nasopharynx of a healthy child in Birmingham, AL (S. K. Hollingshead, unpublished). The collection includes 6A isolates used for studies by Robinson (23) and Mavroidi (16). The TIGR4JS4 strain is a noncapsulated variant of the TIGR4 strain (26) and was produced by replacing the type 4 capsule gene locus with a Janus cassette (kanA-rpsL⁺) and backcrossing three times to wild-type TIGR4 (25, 27; Hollingshead, unpublished). TIGR6AX, TIGR6A4, TIGR6C4, and TIGR4JS4 are four variants of TIGR4 expressing, respectively, no capsule, 6A capsule, 6C capsule, or no capsule. These variants were produced as described below.

TABLE 1.

List of pneumococcus strains

Strain name	Serotype	Tissue location	Country of origin (yr[s] of isolation)	Source or reference
CHPA37	6C	Nasopharynx	United States (1999-2002)	17
CHPA388	6C	Nasopharynx	United States (1999-2002)	17
BGO-2197	6C	Nasopharynx	United States (1979)	Hollingshead, unpublished
MX-67 CMN	6C	Bronchus	Mexico (1996)	23
ACA-C21	6C	Nasopharynx	Canada (1995)	23
BZ17	6C	CSFa	Brazil (2003)	14
BZ39	6C	CSF	Brazil (2003)	14
BZ86	6C	CSF	Brazil (2003)	14
BZ650	6C	CSF	Brazil (2003)	14
ST260	6C	CSF	Brazil (2003)	This study
KK177	6C	Oropharynx	Korea (2005)	This study
CH66	6C	Nasopharynx	China (1997)	23
CH158	6C	Nasopharynx	China (1997)	23
CH199	6C	Nasopharynx	China (1998)	23
CHPA67	6A	Nasopharynx	United States (1999-2002)	17
CHPA78	6A	Nasopharynx	United States (1999-2002)	17
BZ652	6A	CSF	Brazil (2003)	14
KK58	6A	Oropharynx	Korea (2005)	This study
AAU-33	6A	Blood	United States (1998)	16
TIGR4JS4	Noncapsulated	Derived from TIGR4b	Not applicable	27
TIGR6A4	6A	Derived from TIGR4JS4	Not applicable	This study
TIGR6AX	Noncapsulated	Derived from TIGR6A4	Not applicable	This study
TIGR6C4	6C	Derived from TIGR6A4	Not applicable	This study

^aCSF, cerebrospinal fluid.

^bTIGR4 expresses serotype 4 and was originally isolated from blood (26).

PCR and DNA sequencing.

The PCR mixture had 10 to 30 ng of chromosomal DNA, 1 μl of each primer from a 100-pmol stock, 2 μl of 10 mM deoxynucleoside triphosphate, 5 μl of 10× buffer solution, 0.5 μl (2.5 U) of Taq polymerase (Takara Biomedical, Shiga, Japan), and 39.5 μl of sterile water (Sigma, St. Louis, MO). Chromosomal DNA was isolated with a Wizard genomic DNA purification kit (Promega, Madison, WI), according to the manufacturer's instructions. Thermal cycling conditions were customized for each primer pair. The size of the PCR products was determined by electrophoresis in a 1% to 1.5% agarose gel.

The primers used to amplify the wciN, wciO, and wciP genes were described by Mavroidi et al. (16). Additional primers were designed by us using the DNA sequences of the 6A and 6B capsule gene loci in GenBank (accession numbers CR931638 and CR931639, respectively). The newly designed PCR primers are listed in Table 2.

TABLE 2.

List of PCR primers

Primer direction and name	Primer site of accession no. CR931638	Descriptiona	Sequence	Source or reference
Forward primers
5101	6949-6966	In wciN, for INDEL detection	5′-ATTTGGTGTACTTCCTCC	16
5103	8146-8168	In wciO, for sequencing the 6C capsule gene	5′-AAACATGACATCAATTACA	This study
5106	5897-5916	In wchA, for wciN detection	5′-TACCATGCAGGGTGGAATGT	This study
5108	8350-8370	In wciP, for wciP allele detection	5′-ATGGTGAGAGATATTTGTCAc	16
5112	Not applicable	In kanA-rpsL+ with XbaI site	5′-CTAGTCTAGAGTTTGATTTTTAATGG	This study
5113	4870-4894	In wze, for fragment C	5′-GGGAAAAATAAAAAATAGGTCGGG	This study
5118	7613-7636	In wciO with BamHI site	5′-CGCGGATCCAGAAAAACTATGTCGCCTGCTAAa	This study
5120	1-30	In dexB, for fragment A	5′-TGTCCAATGAAGAGCAAGACTTGACAGTAg	27
5122	2187-2206	In wzg, for fragment B	5′-TTCGTCCATTCACACCTTAG	This study
5123	8775-8794	In wciP, for fragment D	5′-TGCCTATATCTGGGGGTGTA	This study
5124	11274-11293	In wzx, for fragment E	5′-AATGATTTGGGCGGATGTTT	This study
5125	13864-13883	In rmlC, for fragment F	5′-AGTGATTGATGCGAGTAAGG	This study
5140	9531-9551	In wzy, for wzy allele detection	5′-CCTAAAGTGGAGGGAATTTCG	16
5141	11459-11478	In wzx, for wzx allele detection	5′-TTCGAATGGGAATTCAATGG	16
Reverse primers
3101	7888-7905	In wciO, for INDEL and wciN detection	5′-CCATCCTTCGAGTATTGC	16
3103	9468-9487	In wzy, for Janus cassette and fragment C	5′-AACCCCTAACAATATCAAAT	This study
3107	9226-9245	In wciP, for wciP allele detection	5′-AGCATGATGGTATATAAGCC	16
3112	Not applicable	In kanA-rpsL+ with BamHI site	5′-CGCGGATCCGGGCCCCTTTCCTTATGCTTTTGG	This study
3113	6203-6224	In wchA with XbaI site	5′-CTAGTCTAGAAATAAAATTTCAATATCTTTCCAG	This study
3121	3676-3660	In wzd, for fragment A	5′-GATTGCGATTCACTACG	This study
3122	5380-5361	In wchA, for fragment B	5′-AACTCCCCAACAACCTCATT	This study
3123	12978-12959	In rmlA, for fragment D	5′-AAAATCAAGGCAACGCTATC	This study
3124	14618-14600	In rmlB, for fragment E	5′-ACGGAGAGCTTGGGTTGTA	This study
3126	17611-17584	In aliA, for fragment F	5′-CAATAATGTCACGCCCGCAAGGGCAAGT	27
3143	10135-10115	In wzy, for wzy allele detection	5′-CCTCCCATATAACGAGTGATG	16
3144	12068-12049	In wzx, for wzx allele detection	5′-GCGAGCCAAATCGGTAAGTA	16

^aFragments A through F refer to the corresponding fragments of the serotype 6C capsule gene locus used for capsule gene locus sequencing.

The PCR products were purified with a Wizard PCR clean-up system (Promega), and the DNA sequences of the PCR products were determined by the UAB genomics core facility. DNA sequences were analyzed with Lasergene v. 5.1 software (DNASTAR, Madison, WI) and the Basic Local Alignment Search Tool (BLAST; (http://www.ncbi.nlm.nih.gov/BLAST/). Potential operons and genes, promoters, and transcription terminators in the capsule gene locus were identified using fgenesB, BPROM, and FindTerm (Softberry Inc.), which are available at the website www.molquest.com. The sequences from the capsule gene locus were compared with the sequences previously reported (16).

Production of TIGR4 variants with 6A and 6C capsule gene loci.

To investigate the role of the wciN gene in 6C capsule expression, we first inserted the 6A capsule locus into the genetic background of TIGR4 and then replaced the wciN_6A gene with wciN_6C using two different DNA cassettes (labeled cassette 1 and cassette 2 in Fig. 1) as described below. Cassette 1 has three parts: the target DNA and two flanking DNAs. The target DNA contains the kanamycin resistance (kanA) and streptomycin sensitivity (rpsL⁺) genes of the Janus cassette (25). The two flanking DNAs were obtained from either wchA or wciO-P genes from AAU-33 (a 6A strain) by PCR using the primer pairs described in Fig. 1 and Table 2. The three DNA fragments in cassette 1 were then linked together by digestion with the appropriate restriction enzyme, followed by ligation with T4 DNA ligase (New England BioLabs, Beverly, MA). The ligation product was then amplified by PCR using primers 5113 and 3102. Cassette 2, which was used to replace the antibiotic selection genes with the wciN_6C gene, was prepared by PCR of CHPA388 (a 6C strain) DNA using primers 5113 and 3102. The PCR products were purified with the Wizard PCR clean-up system (Promega). After confirming the DNA sequences, the PCR products were used for transformation.

FIG. 1.

wciN region exchange experiment diagram. In step A, the wchA/wciN_6A/wciO-P region of TIGR6A4 was replaced with cassette 1. Cassette 1 has three parts (central core and two flanking regions), and each part is about 1 kb long. The central core has antibiotic susceptibility genes kanA and rpsL⁺. The two flanking regions were made with wchA and wciO-P regions from the AAU-33 strain. In step B, cassette 1 in TIGR6AX is replaced with cassette 2. Cassette 2 has wciN_6C, wchA, and wciO-P regions from a 6C strain (CHPA388). TIGR6C4 shows the final product that is obtained after cassette 2 is inserted. XbaI and BamHI sites in the PCR primers, which were introduced to simplify genetic manipulations, are shown.

The wciN replacement was performed as follows. TIGR4JS4, which has a Janus cassette (25) in place of TIGR4's original capsule gene locus (27), was transformed with genes from a 6A strain, AAU-33. A streptomycin-resistant but kanamycin-sensitive isolate was then selected, and DNA from it was used to backcross into TIGR4JS4. The backcross was repeated three times, each time testing for expression of 6A capsular PS. A final backcrossed transformant expressing 6A capsular PS was selected and named TIGR6A4. To remove the wciN_6A gene from TIGR6A4, it was transformed with cassette 1, and a kanamycin-resistant transformant was obtained. After the resulting transformant was backcrossed three times on the TIGR6A4 strain background, the final capsule-deficient kanamycin-resistant transformant was selected and named TIGR6AX. To insert the wciN_6C gene into TIGR6AX, it was then transformed with cassette 2, and a kanamycin-sensitive but streptomycin-resistant transformant was obtained. Backcrossing once more for three times on the TIGR6AX background resulted in a final 6C capsular PS transformant that was named TIGR6C.

Nucleotide sequence accession numbers.

The nucleotide sequences of the capsule gene locus of pneumococcal serotype 6C and the wciN_6C region of several pneumococcal serotype 6C isolates have been submitted to GenBank and assigned accession numbers EF538714 to EF538718.

RESULTS

Identification of additional 6C strains among our “6A” collections.

To obtain a representative collection of 6C serotypes from various locations, we retested our preexisting collection of “6A” strains with our monoclonal antibodies (19) and identified nine additional 6C isolates from five countries on three different continents (Table 1). These isolates were obtained from spinal fluid, blood, and the nasopharynx samples, indicating that 6C can cause invasive pneumococcal infections as well as be isolated in asymptomatic carriages. One strain (BGO-2197) was isolated in 1979 in Birmingham, AL. This finding shows that the 6C serotype has been in existence for at least 27 years and is now found throughout the world.

The capsule gene loci of 6A and 6C differ in the region between the wchA and wciO genes.

Because galactose of 6A PS is replaced with glucose in 6C PS (19), we hypothesized that the galactosyl transferase gene in the 6A serotype is replaced with a new glycosyl transferase gene in 6C. Since the wciN gene encodes galactosyl transferase, we used PCR to compare the sizes of wciN genes of 6A and 6C isolates. The wciN PCR products of all 6C isolates were about 1.8 kb long, whereas the wciN PCR products of all 6A isolates were about 2 kb long (Fig. 2). To distinguish between the two wciN genes from the 6A and 6C serotypes, we have named them wciN_6A and wciN_6C, respectively.

FIG. 2.

Electrophoretic patterns of the PCR products of the wciN region of 6A and 6C isolates. The primers used were 5106 and 3101, which are located in wchA and wciO, respectively. M, DNA ladder. Standard markers with 2,000 and 1,650 bp are indicated on the left. Lanes 1 to 13 contain PCR products of 6C isolates CHPA37 (lane 1), CHPA388 (lane 2), BG2197 (lane 3), BZ17 (lane 4), BZ39 (lane 5), BZ86 (lane 6), BZ650 (lane 7), KK177 (lane 8), CH66 (lane 9), CH158 (lane 10), CH199 (lane 11), MX-67 (lane 12), and ACA-C21 (lane 13). Lanes 14 to 18 contain PCR products of 6A isolates CHPA67 (lane 14), CHPA78 (lane 15), BZ652 (lane 16), KK58 (lane 17), and AAU-33 (lane 18).

To further investigate wciN_6C, we determined the DNA sequences of the capsule gene region, including the wchA, wciN_6C, and wciO genes from five 6C strains (BZ17, BZ86, CHPA388, KK177, and ST-260). Since their sequences were almost identical, the actual DNA sequence is shown for only CHPA388 (Fig. 3A), and the sequences of the other isolates were deposited in GenBank (accession numbers EF538715 to EF538718). The sequence of the wciN_6C gene from CHPA388 was then compared with the 6A sequence of the corresponding region available in GenBank (accession no. CR931638). A summary of the comparison is shown in Fig. 3B.

FIG. 3.

A. The nucleotide sequence of the wciN_6C ORF is shown along with the nucleotide sequences of the 3′ end of wchA and the 5′ end of wciO. The derived amino acid sequence of the wciN_6C ORF is shown below the nucleotide sequence. Also shown are putative termination sites of wchA and wciN_6C as well as putative initiation sites of wciN_6C and wciO. wciO has two potential initiation sites. B. DNA sequences of wciN_6A and wciN_6C regions (shown in boldface) of a 6A strain (GenBank accession no. CR931638) and a 6C strain (CHPA388). The sequence of the nonhomologous mid-region of wciN (about 900 to 1,110 bases) is not shown. Sites of PCR primers (5106, 3101, 5118, and 3113) are shown. Also shown are potential termination sites of wchA and wciN_6C and potential initiation sites of wciN_6C and wciO. C. Genetic map of the capsule loci surrounding wciN of 6A and 6C isolates. The map shows wchA (hatched), wciN (horizontal bars or black), wciO (checkered), and wciP (wavy lines) genes. The 6A locus has two unexpressed DNA fragments (indicated by arrows) upstream (95 bases long) or downstream (312 bases long) of wciN_6A. An alternative initiation site for wciO is located 32 bases upstream of the initiation site shown (position 2721 for 6A). For 6C isolates, the native DNA (1,222 bases; indicated by horizontal bars) in the wciN_6A locus is replaced with new DNA (1,029 bases; black). The replacement creates a new ORF (named wciN_6C) that has 1,125 bases. Nucleotide position 1 in this figure corresponds to nucleotide position 4902 of the 6A capsule genome sequence (GenBank accession no. CR931638).

Sequence comparison revealed clear differences in wciN_6A and wciN_6C genes. The 6C serotype has 1,029 bp in place of 1,222 bp in 6A (Fig. 3B and C). The two wciN genes are completely different, with a sequence homology of only about 50%. The decreased homology begins immediately after the termination of the wchA gene (position 1368) and ends 130 bases upstream of the beginning of wciO (6A-2631 and 6C-2398) (Fig. 3B and C). The different regions are indicated in Fig. 3B. When the DNA sequences flanking the replaced gene were compared between 6A and 6C, significantly more DNA polymorphisms were found in the two flanking regions than in the regions outside of the two flanking regions. For instance, the 300 bases upstream of the replaced gene have 25 different nucleotides, but the 150 bases located immediately upstream from the 300 bases have only one different base (P < 0.001 by Fisher's exact test) (Fig. 3B). Similarly, in the 3′ direction, 20 bases differ in the proximal 110 bases (positions 2398 to 2508 for CHPA388), but only 1 base differs in the next 300 bases (P < 0.001 by Fisher's exact test) (Fig. 3B). These findings are not unique to this particular 6A sequence (accession no. CR931638), because similar results were obtained with the sequence of seven different 6A strains: AAU-33, D020-1B, HS3050, CHPA78, KK65, ST19, and ST558. These findings suggest that the two flanking regions were also introduced with wciN_6C into 6A to create 6C. The fact that all 6C isolates have the identical flanking region sequences is strong evidence that the major genetic replacement generating 6C took place only once and that the capsule gene locus present in all the 6C isolates must have originated from that initial replacement event.

With this gene replacement, wciN_6C has a new open reading frame (ORF) that is 1,125 bases long and encodes a peptide with 374 amino acids, which is designated WciN_6C (Fig. 3A). The termination codon of the new ORF is between the two potential start codons for wciO, which are located at positions 2497 and 2528 of the sequence of CHPA388 (Fig. 3A). When the sequence of the wciN_6C gene was compared with the sequences in the database, 110 bases (from 1627 to 1736 in 6C) of 6C demonstrated 81% homology to the 90 bases of the exopolysaccharide synthesis gene of Streptococcus thermophilus strain CNRZ1066 (3). Also, the translated sequence of the wciN_6C gene has 22% amino acid identity and 44% similarity to the translated sequence of the capH gene of Staphylococcus aureus (15). The wciN_6C gene product is a member of the WaaG family (20). Incidentally, the waaG gene product of Escherichia coli K-12 is an α-1,3-glucosyltransferase involved in lipopolysaccharide synthesis (10).

The wciN gene region is responsible for conversion from the 6A to 6C serotype.

Although the above studies showed that the major difference is in the wciN region, it is possible that some other genetic differences outside of the capsule gene locus could be involved in the 6C expression. To show that only the wciN region is involved, we determined whether the replacement of the wciN_6A region with the wciN_6C region could convert the 6A serotype to the 6C serotype using the “Janus” cassette (27). As shown in Fig. 1, we first produced TIGR6A by replacing the capsule locus of TIGR4 with the 6A capsule gene locus from strain AAU-33. We then removed the wciN_6A gene from TIGR6A by transforming it with cassette 1. The resulting strain, named TIGR6AX, was nonencapsulated and was found, via PCR, to have lost the wciN_6A gene. The wciN_6C region was then inserted into TIGR6AX using cassette 2, which contained the wciN_6C and wciO genes from CHPA388 as well as parts of wchA and wciP genes. The 3′ flanking regions of CHPA388 and AAU-33 had only minimal differences: they differed by two amino acids between their wciO gene products and by one amino acid between the parts of the wciP gene products involved in the genetic exchange. After the insertion of cassette 2, the resulting strain was named TIGR6C, and it was found to express serotype 6C. In addition, TIGR6C was found to have the wciN_6C gene at the expected location when the region was sequenced. This confirmed that the wciN_6C gene and the 1-kb regions surrounding wciN_6C are sufficient for the serotype conversion. Also, although 6A and 6C have few amino acid differences in the wciO and wciP gene products, the wciN gene is most likely responsible for the serotype conversion from 6A to 6C.

The sequences of the capsule gene loci of the 6A and 6C serotypes differ only slightly in regions other than the wciN gene.

To determine if the 6A and 6C capsule gene loci differ only in the wciN region, we determined the sequence of the entire capsule locus of a 6C isolate, CHPA388, by PCR amplifying the entire capsule gene locus between dexB and aliA in six overlapping DNA fragments (fragments A to E) using primers shown in Table 2. Figure 4 shows the genetic map of the sequence of the capsule gene locus. The 6C capsule gene locus is about 17 kb long and contained 14 ORFs (Fig. 4) involved in capsular PS synthesis. The 6C ORFs are all on the same strand and correspond exactly to those found in the 6A capsule locus with the exception of the wciN region. wciP of 6C has G at position 584, as wciP of 6A does. The ORFs of 6C begin with cpsA at the 5′ end and terminate with rmlD at the 3′ end. Both 6A and 6C capsule loci have four potential transcription start sites and a putative rho-independent transcription terminator site. Also, there are “transposase-like” sequences (Fig. 4) at both ends of the 6C capsule gene locus, as are commonly found for many pneumococcal capsule cassette (2). The nucleotide sequence of the entire locus has been deposited in GenBank (accession number EF538714).

FIG. 4.

Capsule locus of a 6A strain (GenBank accession no. CR931638) and a 6C strain (CHPA388). All ORFs involved in capsule synthesis are shown as horizontal arrows, and their direction indicates the transcriptional orientation. For both 6A and 6C loci, the putative transcription initiation sites (bent arrows) and putative termination sites (vertical lines with a solid circle) were identified using fgenesB, BPROM, and FindTerm (Softberry Inc.), available at www.molquest.com. “Transposase” sequences (black boxes, labeled “tnp”) are found at either end of the capsule gene locus.

When the 6C sequence was compared with the capsule locus of a 6A strain (GenBank accession no. CR931638), we found that, except for the wciN region described above, the capsule gene locus of 6C was very homologous (∼98%) to that of 6A. Also, homology was significantly low (about 78%) for about 60 bp in the middle of the cpsA ORF, and the “transposase-like” sequences found at either end of the capsule gene loci were different between the 6A and 6C capsule gene loci. The 6C capsule gene locus did not have the INDEL that is present upstream of wciO in some 6A or 6B capsule loci (16). Despite these differences, the most prominent differences between the 6A and 6C capsule loci were found in the wciN region.

DISCUSSION

Following our initial discovery of 6C among “6A” isolates using monoclonal antibodies (14, 19), we now show the genetic basis for the new serotype, 6C. The capsule gene locus of 6C is very similar to the 6A locus except for wciN: 6A strains have wciN_6A but 6C strains have wciN_6C, which is dramatically different in its size (about 200 bases shorter) as well as its sequence (only about 50% DNA homology) compared to wciN_6A. The size differences have been confirmed with additional 6C isolates by PCR (unpublished information). Chemical studies of purified 6C PS found that repeating unit is identical to 6A PS except for glucose in place of the galactose (19). Consistent with this, WciN_6A belongs to glycosyl transferase family 8, which includes many galactosyl transferases (4), but WciN_6C belongs to glycosyl transferase family 1 and resembles RfaG (WwaG), which is a glucosyl transferase involved in E. coli lipopolysaccharide synthesis (8). Furthermore, a 6A strain can be converted to a 6C strain by replacing the wciN_6A region with the wciN_6C region. Also involved in this genetic exchange are wciO and parts of the wciP and wchA genes. When the amino acid sequences of these coincidentally involved genes were compared between 6A and 6C strains, differences were minimal: two amino acids for wciO, one amino acid for the 5′ end of wciP, and nine amino acids for wchA. These findings strongly support that wciN_6A encodes a galactosyl transferase and wciN_6C encodes a glucosyl transferase. Thus, we conclude that the wciN_6C gene is the basis for 6C serotype.

Capsule production involves many enzymes, and an alteration in the repeating unit of the capsular PS must still be recognized by other enzymes, such as the flippase and polymerase, which are involved in transporting and linking the altered repeating units. It is unlikely, however, that new genes other than wciN_6C are involved in producing 6C PS compared to 6A. We found that homologous recombination between the two wciN regions is sufficient to alter 6A serotype to 6C. Also, when the changes in the repeating unit are small, the other enzymes can accommodate the new repeating unit without any changes. 6A and 6B serotypes have different rhamnosyl transferases and produce PS with different rhamnose-ribitol linkages, but other genes can accommodate both repeating units. Similarly, glucose and galactose have very small chemical differences—a different orientation of the hydroxyl group of the fourth carbon molecule—and this minor chemical difference is often accommodated. For instance, 9L PS has a galactose molecule and 9N PS has a glucose molecule. Their capsule loci resemble each other except for one gene, wcjA, which encodes a galactosyl transferase for 9L and a glucosyl transferase for 9N (2).

In the case of the 9L and 9N serotypes, their wcjAs are very similar and must have originated from a common precursor gene. In contrast, the wciN_6A and wciN_6C genes are very different. Perhaps the wciN_6C gene is from an organism other than pneumococci and was inserted into the 6A capsule gene locus. Horizontal gene transfers between S. pneumoniae and another bacterial species have been demonstrated with antibiotic resistance genes (5, 18) and have been suggested to occur for capsule gene loci because they have low G+C content (2). The “flanking regions” are known to be critical for homologous recombination in pneumococci (21). Indeed, an examination of wciN_6C shows evidence for the two flanking regions that may have participated in the homologous recombination.

The source of wciN_6C is not yet known. wciN_6C resembles no pneumococcal genes, including the capsule gene locus genes of the 90 different, non-6C serotypes (2). About 100 bases of the wciN_6C gene are similar (81% homology) to the epsG gene of S. thermophilus, a gene involved in the synthesis of exopolysaccharide by S. thermophilus. Although the organization of exopolysaccharide synthesis gene loci is similar to pneumococcal capsule gene loci, the homology is short, suggesting that S. thermophilus is not the direct source for wciN_6C. The protein sequence of WciN_6C resembles the waaG (rfaG) gene product of the E. coli K-12 strain, and some pneumococcal genes apparently did come from gram-negative organisms, such as Haemophilus (13). Thus, it is possible that the wciN_6C gene could have come from a gram-negative species as well. Nevertheless, other oral streptococci, such as Streptococcus salivarius, Streptococcus mitis, and Streptococcus oralis, are the leading candidates, since they coexist in the oral cavity with pneumococci and many antibiotic resistance genes came from S. oralis.

When the wciN_6C region was examined for a 27-year-old isolate and multiple recent isolates from different continents, flanking region sequences to wciN_6C were found to be identical. Also, the 6C serotype has only one (or two) capsule gene locus profile (data not shown) (14), whereas the 6A and 6B serotypes have diverse capsule gene locus profiles (16). These findings suggest that a single bacterium producing 6A PS captured wciN_6C and that all the 6C isolates found throughout the world have the capsule gene locus that originated in this 6C founder bacterium. The lack of diversity in the 6C capsule gene loci suggests that the capture of wciN_6C might have occurred recently compared to the history of 6A or 6B capsule gene loci. This singular origin of the 6C capsule gene locus may make 6C a good model for studying bacterial genetic evolution in response to the host immune system. Also, pneumococcal vaccines may not protect us against 6C as well as against 6A and the vaccination may increase the prevalence of 6C. If this happens, it would make 6C an even more interesting model for studying bacterial genetic evolution.

Because of the medical importance, pneumococcal serotypes have been extensively investigated using serological and biochemical as well as genetic tools. Thus, it has been commonly believed that additional pneumococcal serotypes, if found, would be discovered among nontypeable pneumococci. We now show that even a well-characterized serotype such as serotype “6A” harbors a distinct serotype with a relatively large genetic change that has eluded our intensive investigations. Thus, we should be aware that new serotypes can exist among the serotypeable pneumococci, as well as among the nontypeable.