Previous Article | Next Article ![]()
Infection and Immunity, September 2007, p. 4482-4489, Vol. 75, No. 9
0019-9567/07/$08.00+0 doi:10.1128/IAI.00510-07
Copyright © 2007, American Society for Microbiology. All Rights Reserved.

Departments of Pathology,1 Microbiology, University of Alabama at Birmingham, 845 19th Street South, BBRB 614, Birmingham, Alabama 352942
Received 9 April 2007/ Returned for modification 13 May 2007/ Accepted 10 June 2007
|
|
|---|
|
|
|---|
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.
|
|
|---|
|
View this table: [in a new window] |
TABLE 1. List of pneumococcus strains
|
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.
|
View this table: [in a new window] |
TABLE 2. List of PCR primers
|
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 wciN6A gene with wciN6C 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 wciN6C 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.
![]() View larger version (12K): [in a new window] |
FIG. 1. wciN region exchange experiment diagram. In step A, the wchA/wciN6A/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 wciN6C, 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.
|
Nucleotide sequence accession numbers. The nucleotide sequences of the capsule gene locus of pneumococcal serotype 6C and the wciN6C region of several pneumococcal serotype 6C isolates have been submitted to GenBank and assigned accession numbers EF538714 to EF538718.
|
|
|---|
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 wciN6A and wciN6C, respectively.
![]() View larger version (8K): [in a new window] |
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).
|
![]() ![]() View larger version (65K): [in a new window] |
FIG. 3. A. The nucleotide sequence of the wciN6C 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 wciN6C ORF is shown below the nucleotide sequence. Also shown are putative termination sites of wchA and wciN6C as well as putative initiation sites of wciN6C and wciO. wciO has two potential initiation sites. B. DNA sequences of wciN6A and wciN6C 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 wciN6C and potential initiation sites of wciN6C 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 wciN6A. 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 wciN6A locus is replaced with new DNA (1,029 bases; black). The replacement creates a new ORF (named wciN6C) 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).
|
With this gene replacement, wciN6C has a new open reading frame (ORF) that is 1,125 bases long and encodes a peptide with 374 amino acids, which is designated WciN6C (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 wciN6C 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 wciN6C gene has 22% amino acid identity and 44% similarity to the translated sequence of the capH gene of Staphylococcus aureus (15). The wciN6C gene product is a member of the WaaG family (20). Incidentally, the waaG gene product of Escherichia coli K-12 is an
-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 wciN6A region with the wciN6C 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 wciN6A 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 wciN6A gene. The wciN6C region was then inserted into TIGR6AX using cassette 2, which contained the wciN6C 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 wciN6C gene at the expected location when the region was sequenced. This confirmed that the wciN6C gene and the 1-kb regions surrounding wciN6C 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).
![]() View larger version (14K): [in a new window] |
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.
|
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. |
|
|---|
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 wciN6C 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 wciN6A and wciN6C genes are very different. Perhaps the wciN6C 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 wciN6C shows evidence for the two flanking regions that may have participated in the homologous recombination.
The source of wciN6C is not yet known. wciN6C resembles no pneumococcal genes, including the capsule gene locus genes of the 90 different, non-6C serotypes (2). About 100 bases of the wciN6C 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 wciN6C. The protein sequence of WciN6C 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 wciN6C 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 wciN6C region was examined for a 27-year-old isolate and multiple recent isolates from different continents, flanking region sequences to wciN6C 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 wciN6C 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 wciN6C 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.
The work was supported by NIH funding AI-031473 and AI-30021 to M.H.N.
Published ahead of print on 18 June 2007. ![]()
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»