Haplotype-Based Analysis of KIR-Gene Profiles in a South European Population-Distribution of Standard and Variant Haplotypes, and Identification of Novel Recombinant Structures

Abstract

Inhibitory Killer-cell Immunoglobulin-like Receptors (KIR) specific for HLA class I molecules enable human natural killer cells to monitor altered antigen presentation in pathogen-infected and tumor cells. KIR genes display extensive copy-number variation and allelic polymorphism. They organize in a series of variable arrangements, designated KIR haplotypes, which derive from duplications of ancestral genes and sequence diversification through point mutation and unequal crossing-over events. Genomic studies have established the organization of multiple KIR haplotypes-many of them are fixed in most human populations, whereas variants of those have less certain distributions. Whilst KIR-gene diversity of many populations and ethnicities has been explored superficially (frequencies of individual genes and presence/absence profiles), less abundant are in-depth analyses of how such diversity emerges from KIR-haplotype structures. We characterize here the genetic diversity of KIR in a sample of 414 Spanish individuals. Using a parsimonious approach, we manage to explain all 38 observed KIR-gene profiles by homo- or heterozygous combinations of six fixed centromeric and telomeric motifs; of six variant gene arrangements characterized previously by us and others; and of two novel haplotypes never detected before in Caucasoids. Associated to the latter haplotypes, we also identified the novel transcribed KIR2DL5B^*0020202 allele, and a chimeric KIR2DS2/KIR2DL3 gene (designated KIR2DL3^*033) that challenges current criteria for classification and nomenclature of KIR genes and haplotypes.

Keywords: KIR; NK cells; copy-number variation; genes; haplotypes; polymorphism.

Figure 1

Centromeric and telomeric KIR-gene haplotypes commonly observed in Caucasoids, and their linkage disequilibrium in Spanish individuals. Positive and negative relative linkage disequilibrium values are represented with red and gray arrows, respectively. Dotted lines indicate non-statistically significant LD, whilst thickness of solid lines indicates the level of statistical significance (p < 0.05/0.01/0.0001). Conserved “framework” genes are represented as solid boxes. Genes and intergenic spaces are not depicted to scale.

Figure 2

Flowchart for haplotype estimation from KIR-gene profiles.

Figure 3

KIR-gene profiles observed in a sample of Spanish individuals. (A) Gene presence or absence is represented by solid gray and empty boxes; allelic forms are indicated for KIR3DP1. Carrier frequency is given on the right side of each genotype. Gene order reflects, approximately, that seen in the KIR complex, with genes forming A-haplotypes in the middle, flanked on both sides by genes characteristic of B-haplotypes. For KIR 2DL5 and 2DS3, genes represented by two paralogues, a diagonal line indicates that the gene is present in a genotype, but most likely in the alternative location. The ID column shows the number by which the genotype is registered in www.allelefrequencies.net (40); since this database does not distinguish between structural/positional variants, a same ID can correspond to several profiles in the table. Two profiles (7 and 81) include each one individual bearing a concealed variant haplotype, as explained in the text. In the lower part of the panel, which compiles gene profiles not explained by conventional haplotypes, thick lines highlight distinctive traits, including missing genes normally associated with ones present in a given genotype. (B) Distribution of the major groups of KIR genotypes. (C) Cumulative frequencies of KIR profiles.

Figure 4

Distribution of centromeric and telomeric gene profiles and haplotypes. (A) Centromeric and telomeric profiles corresponding to combinations of conventional haplotypes are represented as in Figure 2. Profiles derived from variant and novel arrangements are not represent, therefore frequencies do not sum up 100%. (B) Distribution of partial KIR haplotypes. (C) Frequencies of complete haplotypes and phasing ambiguities.

Figure 5

Variant and novel haplotypes detected in Spanish individuals. Colors are used to highlight gene deletions and hybrid structures derived from recombination of two different genes or haplotypes, but they lack a specific meaning. Duplicated genes are represented in parallel.

Figure 6

Characterization of KIR2DL3*033. (A) Gene and protein structure, including homology to other KIR, and an example of the PCR-SSP test to identify the novel allele (lane 1). IPC stands for internal positive control. (B) Flow cytometry plots of NK cells (CD3⁻CD56⁺) from a donor expressing KIR2DL3*033, and from others with common KIR 2DL2/2DS2/2DL3 genotypes.

Figure 7

Characterization of KIR2DL5B*0020106 and *0020202. (A) Gene structure of KIR2DL5B*0020202, including homology to other KIR2DL5 alleles. (B) Comparison of KIR2DL5 alleles carrying nearly identical coding sequences but highly divergent promoter regions; vertical lines represent polymorphisms distinguishing those alleles. Gene transcription or silencing is indicated on the right side. (C) RT-PCR assay showing KIR2DL5B*0020106 and *0020202 transcription, in contrast with their common, silent homologue KIR2DL5B*0020101. Presence or absence of an intact RUNX binding site in the proximal promoter is indicated for each allele. (D) A KIR2DL5 product is undetectable on the surface of NK cells transcribing KIR2DL5B*0020106 and *0020202, as we reported previously for KIR2DL5A*005, which encodes an identical mature polypeptide.