Evolution of the Major Pilus Gene Cluster of Haemophilus influenzae

Haemophilus influenzae is a ubiquitous colonizer of the human respiratory tract and causes diseases ranging from otitis media to meningitis. Many H. influenzae isolates express pili (fimbriae), which mediate adherence to epithelial cells and facilitate colonization. The pilus gene (hif) cluster of H. influenzae type b maps between purE and pepN and resembles a pathogenicity island: it is present in invasive strains, absent from the nonpathogenic Rd strain, and flanked by direct repeats of sequence at the insertion site. To investigate the evolution and role in pathogenesis of the hif cluster, we compared the purE-pepN regions of various H. influenzae laboratory strains and clinical isolates. Unlike Rd, most strains had an insert at this site, which usually was the only chromosomal locus of hif DNA. The inserts are diverse in length and organization: among 20 strains, nine different arrangements were found. Several nontypeable isolates lack hif genes but have two conserved open reading frames (hicA and hicB) upstream of purE; their inferred products are small proteins with no data bank homologs. Other isolates have hif genes but lack hic DNA or have combinations of hif and hic genes. By comparing these arrangements, we have reconstructed a hypothetical ancestral genotype, the extended hif cluster. The hif region of INT1, an invasive nontypeable isolate, resembles the hypothetical ancestor. We propose that a progenitor strain acquired the extended cluster by horizontal transfer and that other variants arose as deletions. The structure of the hif cluster may correlate with colonization site or pathogenicity.