High rates of recombination in otitis media isolates of non-typeable Haemophilus influenzae

Abstract

Non-typeable (NT) or capsule-deficient, Haemophilus influenzae (Hi) is a common commensal of the upper respiratory tract of humans and can be pathogenic resulting in diseases such as otitis media, sinusitis and pneumonia. The lipopolysaccharide (LPS) of NTHi is a major virulence factor that displays substantial intra-strain and inter-strain variation of its oligosaccharide structures. To investigate the genetic basis of LPS variation we sequenced internal regions of each of seven genes required for the biosynthesis of either the inner or the outer core oligosaccharide structures. These sequences were obtained from 25 representative NTHi isolates from episodes of otitis media. We found abundant evidence of recombination among LPS genes of NTHi, a finding in marked contrast to previous analyses of biosynthetic genes for capsular polysaccharide, a well-documented virulence factor of Hi. We found mosaic sequences, linkage equilibrium between loci and a lack of congruence between gene trees. These high rates were not confined to LPS genes since evidence for similar amounts of recombination was also found in eight housekeeping genes in a subset of the same 25 isolates. These findings provide a population based foundation for a better understanding of the role of NTHi LPS as a virulence factor and its potential as a candidate vaccine.

Fig. 1

LPS structure of Hi strain RM7004 indicating the inner and outer core LPS biosynthetic genes used in this study. Note that the genes lgtC and lic2A add sugars to two different locations in the LPS structure. (Abbreviations–Galactose (Gal); Glucose (Glc); Heptose (Hep); 2-keto-3-deoxyoctulosonic acid (Kdo); Phosphate (PO₄); Phosphoethanolamine (PEA); phosphocholine (ChoP)).

Fig. 2

Species-level Hi ribotype dendrogram based on >400 strains to indicate the genetic diversity of a set of 25 Finnish non-typeable isolates obtained from children with otitis media. The isolates in the vertical column A, indicated by numbers, are the core set of the 16 non-redundant isolates selected on the basis of their ribotype for the analysis of linkage equilibrium and congruence. The isolates in column B are either: (i) genetically identical isolates to those in column A taken from the same patient on the same day (isolate number in parenthesis); or (ii) isolates that have the same ribotype as those in column A. In the case of the latter, the numbers of lipopolysaccharide (LPS) and housekeeping (HK) alleles with identical nucleotide sequences are shown in parentheses. An exceptional difference between ribotype and LPS allele sequence was found between isolates 176 and 477 marked (⋆).

Fig. 3

Polymorphic sites in a fragment of the LPS biosynthesis gene of rfaF of NTHi; the consensus sequence is shown for the uppermost allele and in subsequent alleles only sites which differ from this are identified. The polymorphic sites are numbered above the sequences in vertical format and the isolate numbers indicated on the left-hand side. The mosaic sequence is highlighted in bold type. Group A comprises 13 isolates that share the same sequence as isolate 486 for the first 266 bases. Group B comprises 11 isolates that share the same sequence as isolate 486 from base 267–426.