Mechanisms of copy number variation and hybrid gene formation in the KIR immune gene complex

Abstract

The fine-scale structure of the majority of copy number variation (CNV) regions remains unknown. The killer immunoglobulin receptor (KIR) gene complex exhibits significant CNV. The evolutionary plasticity of the KIRs and their broad biomedical relevance makes it important to understand how these immune receptors evolve. In this paper, we describe haplotype re-arrangement creating novel loci at the KIR complex. We completely sequenced, after fosmid cloning, two rare contracted haplotypes. Evidence of frequent hybrid KIR genes in samples from many populations suggested that re-arrangements may be frequent and selectively advantageous. We propose mechanisms for formation of novel hybrid KIR genes, facilitated by protrusive non-B DNA structures at transposon recombination sites. The heightened propensity to generate novel hybrid KIR receptors may provide a proactive evolutionary measure, to militate against pathogen evasion or subversion. We propose that CNV in KIR is an evolutionary strategy, which KIR typing for disease association must take into account.

Figure 1.

Segregation of contracted haplotypes in a CEPH family. KIR haplotypes were determined by segregation analysis in all members of a three-generation family. Allele designations correspond to HGNC nomenclature (http://www.genenames.org/genefamily/kir.php).

Figure 2.

Gene copy number determination of KIR2DL4 at the genomic level using quantitative real-time multiplex PCR. The results of duplicate experiments are expressed as the mean relative ratio of KIR2DL4 to a reference gene with an SD. Individuals carrying the j or t haplotype possess one copy of KIR2DL4.

Figure 3.

Gene annotation of the sequenced portions of the j and t haplotypes. Locations of fosmid clone insert sequences are shown below each haplotype.

Figure 4.

Segregation of KIR2DL3/2DP1 and KIR2DL1/2DS1 in the CEPH and the Caucasian family from Northern Ireland, as analysed by gene-specific LR-PCR. Results from the corresponding GAPDH genomic PCR are shown.

Figure 5.

Segregation of contracted haplotypes in a Caucasian family from Northern Ireland. ^lHaplotype sequencing subsequently showed that the t haplotype possesses a novel KIR3DL2 allele. This allele was provisionally labelled *007-like because it only differs from allele *007 by a single non-synonymous nucleotide substitution in exon 5; an adenosine to guanosine corresponding to a substitution of a glutamic acid to a glycine residue in the D2 domain of the predicted translated product.

Figure 6.

Dot matrix analysis of the j and t KIR haplotype sequences. The plot shows the four KIR genes of the j haplotype on the x-axis and the 5 KIR genes of the t haplotype on the y-axis. Regions of similarity are identified as a concentration of dots forming diagonal lines. Minisatellites can be visualized as boxes. The deletion breakpoints on both haplotypes are indicated by dashed lines. The intergenic regions and introns of the KIR loci are well conserved.

Figure 7.

The presumed order of genomic deletions in the formation of the j haplotype. First, a >50 kb deletion fused the KIR2DL1 and KIR2DS1 genes creating the KIR2DL1/S1 hybrid gene. Subsequently, a smaller ∼17 kb deletion fused the KIR2DL3 and KIR2DP1 genes to form the KIR2DL3/2DP1 gene. A representative KIR ‘B’ haplotype is shown below the deletion haplotypes as an example of one potential ancestral haplotype involved in the derivation of haplotype t. The represented haplotype was the most frequent observed in a panel of 85 Caucasoid individuals with a frequency of 0.124 (41). Apart from the ∼17 kb deletion and a single nucleotide, the j and t haplotype sequences are identical, raising the possibility of intra-chromosomal recombination in the formation of the j haplotype, and implicating the t haplotype as the single precursor of the j haplotype.

Figure 8.

The KIR2DL3/2DP1 gene is a product of recombination between KIR2DL3 (CU464054) and KIR2DP1 (CU464061). Nucleotide positions that differ between CU464054 and CU464061 are shown.

Figure 9.

Potential ssDNA secondary structures formed at the KIR2DL3 (left) and KIR2DP1 (right) breakpoint regions ssDNA. These structures can serve as recognition signals to induce strand breaks that cause genomic rearrangements by recombination-repair. The dashed line indicates the breakpoint intervals. The solid grey line depicts the full ∼282 bp Alu sequence. The black circle contains the core 22 nucleotides of the Alu recombination hotspot (5′-TGTAATCCCAGCACTTTGGGAG-3′).

Figure 10.

Proposed mechanism behind the formation of the KIR2DL3/2DP1 gene; deletion of ∼17 kb by homologous unequal intra-chromosomal recombination (Alu-mediated). A DNA break is introduced at the recombination hotspot site of an Alu repeat within a KIR gene. A second break occurs 17 kb away that contains sequence homologous to the first break site. The two homologous sequences serve as a substrate for double-strand break repair, which leads to deletion of the intervening sequences between the break sites.

Figure 11.

The KIR2DL1/2DS1 gene is a product of recombination between KIR2DL1 (CU45907) and KIR2DS1 (AL133414). Nucleotide positions that differ between CU45907 and AL133414 are shown. KIR2DL1/2DS1a (j, t haplotypes, Ukrainian and African American) and KIR2DL1/2DS1b (Han Chinese).

Figure 12.

Potential ssDNA conformation formed from the MLT1 sequence at the KIR2DS1 breakpoint region. The breakpoint interval sequences are circled. Such protrusive non-B DNA structures may catalyse KIR rearrangements.

Figure 13.

Schematic of the 300 bp KIR bi-directional core promoter region of KIR2DL1/S1 (upper) with respect to KIR2DS1 (lower). The positions of known and predicted TFBSs are indicated by boxes. Shaded grey boxes represent core promoter TFBSs. Coloured boxes represent significant TFBS matches corresponding to nucleotide polymorphisms, with their identity (HUGO gene symbol nomenclature) given against the gene in which the TFBS is present. For example, the Sp1 site is present in KIR2DS1 but not in KIR2DL1/2DS1. Transcription initiation sites for forward and reverse promoters of the bidirectional promoter are shown by the rightward and leftward arrows, respectively (Supplementary Material Fig. S7). The vertical lines indicate the positions of polymorphic nucleotides. Numbering indicates the positions of the polymorphic residues relative to the translation initiation codon of the KIR2DL1/S1 gene, where the base A of the ATG codon is denoted nucleotide +1. The nucleotide present at each variable position is shown for both genes. The YY1 site is present neither in KIR2DS1 nor KIR2DL1/S1.

Figure 14.

Hybrid KIR gene organization and predicted protein structures. The coding regions of the exons are represented as grey boxes; their size in base pairs is shown. Pseudoexons are indicated with a dashed line. The way in which the exons code for each protein domain/region is shown. The main structural characteristics of KIR proteins are shown where the domains and regions are represented as boxes of different colours according to the key. The approximate length (amino acids) of each domain or region is shown next to their corresponding box. Predicted KIR protein structures are displayed adjacent to their respective gene. The structural characteristics of two immunoglobulin-like domain KIR proteins are shown. The intact ITIM site of KIR2DL3/2DP1 is shown as a red box. The disrupted ITIM is represented by a dashed red box. The KIR2DL1/S1 is presented with the associated adaptor molecule, DAP12.