The effect of immune selection on the structure of the meningococcal opa protein repertoire

Abstract

The opa genes of the Gram negative bacterium Neisseria meningitidis encode Opacity-associated outer membrane proteins whose role is to promote adhesion to the human host tissue during colonisation and invasion. Each meningococcus contains 3-4 opa loci, each of which may be occupied by one of a large number of alleles. We analysed the Opa repertoire structure in a large, well-characterised collection of asymptomatically carried meningococci. Our data show an association between Opa repertoire and meningococcal lineages similar to that observed previously for meningococci isolated from cases of invasive disease. Furthermore, these Opa repertoires exhibit discrete, non-overlapping structure at a population level, and yet low within-repertoire diversity. These data are consistent with the predictions of a mathematical model of strong immune selection upon a system where identical alleles may occupy different loci.

Figure 1. Schematic of the model setup.

Each isolate has two opa loci, each having two HV regions (HV1 and HV2). Each HV region can express one of two possible variants, a or b for HV1 and x or y for HV2, which may be the same for both opa genes. There is no dose dependence; an isolate with locus 1 expressing ‘ax’ and locus 2 expressing ‘ax’ is considered to be just ‘ax’. This simple 2-locus system can be generalized to more loci without affecting the qualitative model outcome.

Figure 2. Different dynamics generated by the model.

a) All strains coexist (no strain structure NSS), but strains expressing 3 and 4 epitopes are suppressed, γ = 0.65. b) Cyclical dynamics (cyclical strain structure CSS) with successive dominance of discordant sets of strain expressing 2 variants, and suppressed oscillations of strains expressing 3 and 4 alleles , γ = 0.83. c) The pathogen population is dominated by a subset of non-overlapping strains (DSS), each expressing two alleles, all other strains are suppressed, γ = 0.9.

Figure 3. The combination of HV1-HV2 epitopes occurring in all carried isolates from the Czech Republic between March and June of 1993, presented as a heatmap.

HV1 and HV2 epitopes are classified into alleles as described in the Materials and Methods section. The colours indicate the total number of observations of each particular combination. Each opa locus is treated independently, so each isolate can contribute more than one combination. (See Text S1 for an alternative representation of the data).

Figure 4. Observed and expected numbers of Opa loci per repertoire with identical HV1-HV2 combinations.

Only unique repertoires were included, in order to control for prevalent repertoires, and the x-axis shows how many of these loci have identical HV combinations. The ‘random’ distribution was generated using the same number of HV combinations found in the data, and these were randomly distributed among isolates. Note that the deviation from the data would be even greater if the HV combinations were truly randomized, ie. if all combinations had the same probability of occurring in a given repertoire. The observed number of isolates with two or more of the same HV combination departs significantly from the expected number (p<0.0001).