Complete Sequence of the cap Locus of Haemophilus influenzae Serotype b and Nonencapsulated b Capsule-Negative Variants

Hib cap locus of Hi 1007.

Although there is some nucleotide sequence from the Hib cap locus (accession no. X54987 , X78559 , and S62752 ), the sequence for the entire partially duplicated Hib cap locus, from end to end, particularly region III and the end junctions within the chromosome, has not been previously described. Hi 1007 contains a complete copy of region I with four genes, bexDCBA, that are 99% identical to the previously described Hib region I genes (Fig. 1B) (16) and a second, truncated copy of region I with bexDCB and bexA with the deletion previously described (19). IS1016-V5 (accession no. X58177 ), a truncated copy of IS1016, was found at the end adjacent to the partially deleted bexA. IS1016-V5 is one of six IS1016 variants (V1 to V6) previously described by Kroll et al. (17). Table 2 shows a comparison of Hi 1007 proteins from region I with other capsular polysaccharide (CPS) biosynthesis proteins.

Immediately adjacent to both copies of region I is region II, containing four serotype b-associated open reading frames (ORFs) previously identified by van Eldere et al. (34) (Fig. 1B). The nucleotide sequences for region II of Hib strain RM135 (accession no. X78559 ) and H. influenzae type a (Hia) strain Hia RM107 (accession no. Z37516 ) have been reported. Only the protein products from Hib RM135 orf1 and Hia acs1 have been assigned a function, that of bifunctional ribulose 5-phosphate reductase/CDP-ribitol pyrophoshorylase (9, 34). The predicted protein encoded by Hi 1007 region II orf1 was 99% similar to the previously described Hib region II orf1 protein and 96% similar to Acs1 (Table 2). The predicted protein encoded by Hi 1007 region II orf2 is 67% similar to Acs2 from Hia. We therefore suggest that Hib region II orf1 and orf2 be named bcs1 and bcs2, for b capsule specific, based on the nomenclature proposed for the serotype a capsule-specific genes acs1 and acs2 (9).

Hi 1007 region II orf3 is 100% identical to the previously sequenced Hib region II orf3 but showed no nucleotide similarity to acs3, from region II of Hia, and their products showed no amino acid similarity. There was low-level similarity between the predicted protein of Hi 1007 orf3 and a probable sugar transferase from Campylobacter jejuni and a protein of unknown function from the cap locus of Streptococcus pneumoniae type 19B (Table 2). Hi 1007 region II orf4 is 100% identical to the previously sequenced Hib region II orf4 but showed no nucleotide similarity to acs4, from region II of Hia, and their products showed no amino acid similarity. There was low-level similarity between the predicted protein of Hi 1007 orf4 and Cps14K, a CPS synthesis protein from Streptococcus pneumoniae type 14 (Table 2). We propose that capsule-specific region II ORFs orf3 and orf4 be named bcs3 and bcs4.

Although region III of Hib is common to all serotypes, little is known about it. No sequence data from any H. influenzae serotype had been published for region III. We have identified two ORFs in region III of Hi 1007. Table 2 shows the strong similarity of both genes to cap genes involved in postpolymerization of capsules from E. coli, N. meningitidis, A. pleuropneumoniae, P. multocida, and M. haemolytica. The function of these genes in H. influenzae has yet to be determined. It has been proposed that LipA and LipB from N. meningitidis are involved in phospholipid substitution, necessary for translocation of the capsule (10). The strong consensus among the various genes and proteins involved in CPS biosynthesis suggests a shared function. In many cases the genes and proteins from region III, as well as those from region I, have been shown to be functionally interchangeable with other region III and region I genes, respectively, by complementation studies (10, 24). Furthermore, evidence strongly suggests that a phospholipid moiety is covalently associated with the Hib CPS (22). We propose naming orf1 and orf2 from region III hcsA and hcsB for Haemophilus capsule synthesis.

Hib⁻ GA834 and Hi 373 cap loci.

The two Hib⁻ strains identified did not contain a central 9-kb EcoRI fragment, suggesting that there was only one copy of regions I, II, and III and that the region I present was the site of the previously described IS1016-bexA deletion (19). We confirmed this finding and sequenced a single truncated copy of region I and only one copy of region II and III in both strains as described above. The sequence and arrangement of region I from GA834 and Hi 373 were found to be >99% identical to the those of the truncated copy of region I in Hi 1007. Furthermore, regions II and III from GA834 and Hi 373 were nearly identical to regions II and III from Hi 1007 with one exception. Hi 373 contained a 24-bp direct repeat found at the 3′ end of orf4 (now bcs4), an in-frame insertion that results in a slightly larger protein.

Location of the cap locus within the H. influenzae chromosome.

The cap locus from division I H. influenzae strains is flanked by direct repeats of IS1016 (17). However, it has not been previously determined whether there is a common location for the cap locus within the chromosome. The nonencapsulated H. influenzae Rd strain, for which the first completed H. influenzae genomic sequence is available, contains copies of IS1016 at three different locations within its chromosome. HI1018, a complete, putative copy of IS1016, is found in section 97 at chromosomal coordinates 1082514 to 1081941 on the negative strand of the Rd genome (8) (accession no. NC_000907 ). Two incomplete copies of IS1016, HI1329 and HI1577 (chromosomal coordinates 1407406 to 1407095, negative strand, and 1646128 to 1646346, respectively), are found in sections 127 and 143 of the Rd genome (8). Rd, a division I H. influenzae strain, was formerly an encapsulated serotype d strain. Therefore, if Rd lost its capsule in a recombination event between the IS1016 repeats flanking the ends of the cap locus, the predicted result would be a complete copy of IS1016 remaining, as found in section 97 (HI1018; 1082514 to 1081941 on the negative strand of the complete Rd genome). In addition, Herbert et al. suggest that the site of insertion of the cap locus in an Rd b⁺ transformant was at HI1018 (11).

Primers based on the region of DNA just outside of IS1016 from section 97 in Rd (Table 1) were generated and used in combination with region I (bexB) or region III (endregIII) primers to amplify DNA at the ends of the cap locus from Hi 1007, GA834, and Hi 373, as well as American Type Culture Collection (Manassas, Va.) strains ATCC 9006 (9), ATCC 9007 (14), ATCC 9008, ATCC 8142, and ATCC 700222 (33) (serotypes a, c, d, e, and f, respectively). All PCRs were performed by using an annealing temperature 5°C less than the melting temperature of the primer as calculated by the manufacturer (Sigma-Genosys, The Woodlands, Tex.).

PCR amplification of the end junctions adjacent to region I [primers bexB and Rd 97(8388)] gave a 621-bp product for Hi 1007, GA834, and Hi 373, consistent with a partial deletion of IS1016-bexA at the junction in section 97 (Fig. 2A). When the same primers were used with ATCC 9007 and ATCC 9008 (serotypes c and d), a product of 1.8 kb was seen (Fig. 2A). The increase of 1.2 kb indicates the presence of complete copies of IS1016 and bexA at the left junction in these strains. Each PCR product was sequenced, confirming the presence of the truncated IS1016-V5 and partially deleted bexA in Hi 1007, GA834, and Hi 373 and a complete IS1016 (IS1016-V2, accession no. X58174 ) and a complete bexA in the serotype c and d strains at Rd section 97 bp 8230. PCR analysis of the end adjacent to region III [primers endregIII and Rd97(7400)] demonstrated a 906-bp product for Hi 1007, GA834, Hi 373, and the serotype c and d strains (Fig. 2B). These products were sequenced, and region III was found to contain a complete copy of IS1016-V2 adjacent to the sequence found at Rd section 97 bp 7521. We conclude that the cap loci from serotypes b, c, and d (all division I H. influenzae) appear to be located within the same chromosomal site between bp 7520 and 8230 of Rd in section 97 (Fig. 1). The only variation noted was the IS1016-bexA deletion at the region I junction in duplicated Hib strains and Hib⁻ variants compared with a complete IS1016 and bexA forming the region I junction in serotypes c and d.

• Open in new tab
• Download powerpoint

FIG. 2.

PCR analysis of the end junctions of the H. influenzaecap locus. Lanes 1, H. influenzae serotype a strain ATCC 9006; lanes 2, serotype b strain Hi 1007; lanes 3, Hib⁻ variant Hi 373; lanes 4, Hib⁻ variant GA834; lanes 5, serotype c strain ATCC 9007; lanes 6, serotype d strain ATCC 9008; lanes 7, serotype e strain ATCC 8142; lanes 8, serotype f strain ATCC 700222; lane 9, 1-kb standard marker (Promega Corp., Madison, Wis.). (A) Samples were amplified with primer set bexB-Rd97(8388) (Table 1). The truncated end junctions of Hib (lane 2) and the Hib⁻ variants (lanes 3 and 4) yielded a product of 621 bp due to the 1.2-kb deletion within the IS1016-bexA region. Strains ATCC 9007 (lane 5) and ATCC 9008 (lane 6), serotypes c and d, respectively, yielded a product of 1.8 kb, consistent with an intact IS1016 and bexA. (B) Samples were amplified with primer set endregIII-Rd97(7400) (Table 1). The end junctions adjacent to region III for Hi 1007, Hi 373, GA834, ATCC 9007, and ATCC 9008 are the same size (906 bp) (lanes 2 to 6). No product is seen for serotype a (lane 1), e (lane 7), or f (lane 8).

No PCR product was amplified with either primer pair for serotypes a, e, and f (Fig. 2), suggesting the possibility of an alternative location for the capsule genes in some H. influenzae strains. Capsule serotype f strains and some serotype a strains are classified as division II, distinguished by the lack of an IS1016 element associated with the cap locus (17). Division II cap loci may therefore not be linked to the IS1016 site located in section 97 of Rd. Although serotype e strains are division I, the failure to amplify cap-specific material from the serotype e strain tested suggests that it may also contain capsule genes in a different location. However, given the greater genetic distance between serotype e and other capsule serotypes within division I (27), failure to amplify a PCR product may reflect sequence variation rather than a different cap location. Further studies will be necessary to document cap locus sequence variation among non-b serotypes and to identify alternative chromosomal sites for capsule genes.

The possession of a polysaccharide capsule greatly enhances the ability of H. influenzae to cause invasive disease (26). The Hib conjugate vaccine has resulted in the near elimination of invasive serotype b disease (1). However, the success of the Hib conjugate vaccines should not bring about complacency. The occurrence of invasive Haemophilus disease due to other capsular serotypes as well as nontypeable H. influenzae during the postvaccine era (2, 29, 33, 35) suggests that greater understanding of non-b capsules and events leading to acquisition and/or loss of capsule genes may be relevant in the future. The IS1016-bexA deletion has been demonstrated in some serotype a strains, and it potentially stabilizes encapsulation and enhances virulence in these strains (2, 20). The natural transformability of H. influenzae and the evidence that exchange of cap genes has occurred between strains in the past (20) suggest that further investigation of the cap genes from all six capsular serotypes is warranted.

The nucleotide sequences reported in this paper have been deposited in GenBank under the accession numbers AF549213 (Hi 1007), AF549210 (GA834), and AF549212 (Hi 373).