The Haemophilus cryptic genospecies Cha adhesin has at least two variants that differ in host cell binding, bacterial aggregation, and biofilm formation properties

Abstract

The Haemophilus cryptic genospecies (HCG) causes genital tract infections in pregnant and postpartum women and respiratory infections in neonates. The major surface adhesin in HCG is called Cha, which mediates bacterial adherence to cultured human epithelial cells. In this study, we report that there are two antigenically distinct variants of Cha, dubbed Cha1 and Cha2. These variants are encoded by the same genetic locus in diverse strains and have nearly identical N-terminal export and C-terminal surface anchoring domains but significantly different internal adhesive domains. Based on the comparison of derivatives of a laboratory strain of Haemophilus influenzae expressing either surface-associated Cha1 or surface-associated Cha2, Cha1 mediates a higher level of adherence to cultured human epithelial cells and Cha2 mediates a higher level of adherence to abiotic surfaces. We hypothesize that variation in the Cha1 and Cha2 internal region results in changes in binding specificity or binding affinity and may be associated with adaptation to different host environments during colonization and disease.

FIG 1

HCG strain 1610 expresses a variant Cha adhesin called Cha2. (A) Whole bacteria from the indicated HCG strains were spotted onto a nitrocellulose membrane and examined by immunoblot assay for surface expression of Cha1. The indicated strains were also tested for adherence to Chang respiratory epithelial cells. Adherence levels below 5% of the inoculum were scored as negative. Strains 420 and 427 contain untranscribed cha-1 alleles. Strain 2446 contains untranscribed cha-2. Strain 1595 cha-1::Tn has a transposon insertion in the cha-1 allele, does not produce Cha1, and is nonadherent. Strain 2452 contains a cha-1 allele containing ∼95 tandem repeat domains, rendering it nonadherent. (B) Domain comparison of Cha in HCG 1595 (Cha1) and the variant Cha in HCG strain 1610 (Cha2). Domain annotations were determined by daTAA algorithm analysis (http://toolkit.tuebingen.mpg.de/dataa). The dashed line indicates domains in Cha1 that are not present in Cha2. The bracket indicates the region of greatest diversity between Cha1 and Cha2, and the shaded block indicates the region of greatest homology between Cha1 and Cha2. Amino acid numbers are indicated above and below protein depictions.

FIG 2

Cha2 is an adhesin that binds to respiratory and cervical epithelial cells but does not mediate aggregation. (A) HCG 1595 clone A9 (produces Cha1 with 0 repeats), wild-type HCG1610 (produces Cha2 with ∼60 repeats), and 1610cha-2::Tn (produces no Cha2) were inoculated onto monolayers of Chang conjunctival, Detroit-562 pharyngeal, HeLa cervical, and HEC-1-B endometrial epithelial cells. (B and C) H. influenzae Rd derivatives expressing nearly full-length Cha1 and Cha2 passenger domains were used to inoculate Chang respiratory epithelial cells (B) or HeLa cervical epithelial cells (C). In all panels, adherence is expressed as the percentage of the result for the inoculum. Error bars indicate standard deviations. Data shown are representative of three replicate assays. *, P ≤ 0.05.

FIG 3

Binding domain in passenger domain is longer in Cha2 than in Cha1. (A) Cha2 domains expressed in H. influenzae Rd fused with a FLAG epitope and the Cha membrane anchor are depicted, and numbers represent amino acid positions. (B) FLAG-tagged Cha1 and Cha2 partial passenger domains expressed on the surface of H. influenzae Rd were tested for adherence to Chang cells. Error bars represent standard deviations of the means. Data shown are representative of three replicate experiments. *, P ≤ 0.02.

FIG 4

A high number of tandem repeats do not interfere with Cha1- and Cha2-mediated adherence to plastic. (A) The indicated H. influenzae Rd derivatives expressing Cha1 or Cha2 constructs were resuspended in BHI broth to an optical density at 600 nm (OD₆₀₀) of 0.8 to 1. Bacterial aggregation and settling were assessed by measuring absorbance every 30 min for 3.5 h. (B to E) The indicated HCG strains (B and E) and H. influenzae Rd strains (C and D) were incubated in 96-well plates for 48 to 72 h. Wells were washed repeatedly, stained with crystal violet, and washed again, and the remaining stain was solubilized in dilute acetic acid. Plates were then read in a spectrophotometer, reflecting the fact that absorbance is positively correlated with the number of bacteria adherent to each well. Error bars represent standard deviations of absorbance from quadruplicate wells. Each graph is representative of three replicate experiments. *, P ≤ 0.04.