Improved full-length killer cell immunoglobulin-like receptor transcript discovery in Mauritian cynomolgus macaques

Abstract

Killer cell immunoglobulin-like receptors (KIRs) modulate disease progression of pathogens including HIV, malaria, and hepatitis C. Cynomolgus and rhesus macaques are widely used as nonhuman primate models to study human pathogens, and so, considerable effort has been put into characterizing their KIR genetics. However, previous studies have relied on cDNA cloning and Sanger sequencing that lack the throughput of current sequencing platforms. In this study, we present a high throughput, full-length allele discovery method utilizing Pacific Biosciences circular consensus sequencing (CCS). We also describe a new approach to Macaque Exome Sequencing (MES) and the development of the Rhexome1.0, an adapted target capture reagent that includes macaque-specific capture probe sets. By using sequence reads generated by whole genome sequencing (WGS) and MES to inform primer design, we were able to increase the sensitivity of KIR allele discovery. We demonstrate this increased sensitivity by defining nine novel alleles within a cohort of Mauritian cynomolgus macaques (MCM), a geographically isolated population with restricted KIR genetics that was thought to be completely characterized. Finally, we describe an approach to genotyping KIRs directly from sequence reads generated using WGS/MES reads. The findings presented here expand our understanding of KIR genetics in MCM by associating new genes with all eight KIR haplotypes and demonstrating the existence of at least one KIR3DS gene associated with every haplotype.

Keywords: Killer cell immunoglobulin-like receptors; Mauritian cynomolgus macaques; PacBio long-amplicon sequencing; Whole Genome and Macaque Exome Sequencing.

Fig. 1. KIR2DL4 genomic alignment

Genomic reads mapped to BX842591 KIR2DL4 exon 1. Each sequence represents a consensus of all reads mapped. Degenerate bases are used to represent sites of heterogeneity. Detected genotypes are represented by two to five animals each. Previous forward primer sequence is shown in the black arrow. Mismatches within the previous forward primer targeted region are boxed. The revised primer binding region is shown by the white arrow.

Fig. 2. PacBio CCS genotypes for representative MCM

The number of identical CCS reads for each transcript is presented within cells. Columns represents CCS reads detected for individuals. Haplotype associations are designated by color. * denotes sequence read counts that are shared between 2 haplotypes. KX8921661 (K2, KIR3DL7) was not detected in CY0116 and CY0568 and therefore isn’t displayed; however, it was detected in K2/K5 animal CY0163. Likewise, KX892687 (K6, KIR3DS) was not detected in CY0568 and therefore isn’t displayed; however, it was detected in K4/K6 animal CY0330.

Fig. 3. WGS/MES genotypes for representative MCM

The number of sequence reads mapped to representative exons are presented within a single cell. Three numbers are given for KIR3DS/KIR3DL alleles, two for KIR2DL4 alleles, and one for KIR1D alleles. The three genes detected in WGS/MES but not in PacBio amplicon sequencing are displayed with “EU” GenBank accession numbers. Haplotype associations are designated by color. Reads displayed in pink specify those that map to genes whose haplotype associations are currently unknown.

Fig. 4. Revised model for Mafa-KIR haplotype genomic organization

Haplotypes are displayed in order of descending frequency within the MCM population. The population frequency for haplotypes is shown to the right of each row and was derived from Bimber et al. (2008) combined with genotypes reported in this study (n = 294). Note that recombinant haplotypes account for 12% of the population but are not displayed. The organization of genes was inferred based on the published rhesus macaque KIR haplotype (Sambrook et al. 2005). Presumed pseudogenes are shown (white box). The number of unique alleles per gene within the population is given below.