Fluctuating and Geographically Specific Selection Characterize Rapid Evolution of the Human KIR Region.

Danillo G Augusto,Ravi Dandekar,Jill A Hollenbach,Paul J Norman

doi:10.3389/fimmu.2019.00989

Abstract

The killer-cell immunoglobulin-like receptor (KIR) region comprises a fast-evolving family of genes that encode receptors for natural killer (NK) cells and have crucial role in host defense. Evolution of KIR was examined in the context of the human genome. Gene-content diversity and single nucleotide polymorphisms (SNP) in the KIR genes and flanking regions were compared to >660,000 genome-wide SNPs in over 800 individuals from 52 populations of the human genome diversity panel (HGDP). KIR allelic diversity was further examined using next generation sequencing in a subset of 56 individuals. We identified the SNP rs587560 located in KIR3DL3 as a marker of KIR2DL2 and KIR2DL3 and, consequently, Cen A and Cen B haplotypes. We also show that combinations of two KIR2DL4 SNPs (rs35656676 and rs592645) distinguish KIR3DL1 from KIR3DS1 and also define the major KIR3DL1 high- and low-expressing alleles lineages. Comparing the diversity of the SNPs within the KIR region to remainder of the genome, we observed a high diversity for the centromeric KIR region consistent with balancing selection (p < 0.01); in contrast, centromeric KIR diversity is significantly reduced in East Asian populations (p < 0.01), indicating purifying selection. By analyzing SNP haplotypes in a region spanning ~500 kb that includes the KIR cluster, we observed evidence of strong positive selection in Africa for high-expressing KIR3DL1 alleles, favored over the low-expressing alleles (p < 0.01). In sharp contrast, the strong positive selection (p < 0.01) that we also observed in the telomeric KIR region in Oceanic populations tracked with a high frequency of KIR3DS1. In addition, we demonstrated that worldwide frequency of high-expression KIR3DL1 alleles was correlated with virus with virus (r = 0.64, p < 10−6) and protozoa (r = 0.69, p < 10−6) loads, which points to selection globally on KIR3DL1 high-expressing alleles attributable to pathogen exposure.

Highlights

Due to their pivotal role in the immune response, much attention has been given recently to variation in the highly polymorphic killer cell immunoglobulin-like receptors (KIR), expressed on the surface of natural killer (NK) [1] cells and a subset of T cells [2]
To understand how diversity within the killer-cell immunoglobulin-like receptor (KIR) region compares to genome-wide diversity within populations, we examined genotypic data from over 660,000
By leveraging publicly available data for of 650,000 single nucleotide polymorphisms (SNP) in one of the most well-characterized sample collections of global populations, we present the first analysis of diversity and patterns of selection in the KIR region in comparison to diversity across the human genome

Summary

Introduction

Due to their pivotal role in the immune response, much attention has been given recently to variation in the highly polymorphic killer cell immunoglobulin-like receptors (KIR), expressed on the surface of natural killer (NK) [1] cells and a subset of T cells [2]. The KIR gene family co-evolves with the genes that encode the human leukocyte antigen (HLA) class I molecules, the ligands for most KIR molecules [3,4,5]. KIR transduce inhibitory and/or activating signals that regulate NK cell activation, and specific KIR and HLA combinations have been associated with numerous diseases, including autoimmunity, cancer and infection [6,7,8,9,10]. The unusual structural polymorphism of the KIR region, yielding variable presence or absence for most of the KIR genes (and numerous observed gene-content haplotypes) [15] combined with pronounced allelic variation at each locus- and their demonstrated importance for human survival [16, 17] make them intriguing targets for disease association and evolutionary studies. KIR genes share substantial sequence similarity with one another, which together with their structural polymorphism, impose technical barriers to their study, at allelic level [22]

Methods

Results

Conclusion