, Volume 59, Issue 6, pp 517–522

The expanded cattle KIR genes are orthologous to the conserved single-copy KIR3DX1 gene of primates


    • Department of Structural Biology
  • Laurent Abi-Rached
    • Department of Structural Biology
  • John A. Hammond
    • Department of Structural Biology
  • Peter Parham
    • Department of Structural Biology
Brief Communication

DOI: 10.1007/s00251-007-0214-x

Cite this article as:
Guethlein, L.A., Abi-Rached, L., Hammond, J.A. et al. Immunogenetics (2007) 59: 517. doi:10.1007/s00251-007-0214-x


Cattle are the only non-primate species for which expansion of the killer cell immunoglobulin-like receptor (KIR) genes has been reported. We analyzed cattle KIR sequences to determine their relationship to the two divergent lineages of primate KIR: one comprising the KIR3DX1 gene of unknown function, the second comprising all other primate KIR genes, which encode variable major histocompatibility complex class I receptors. Phylogenetics and analysis of repetitive elements shows that cattle KIR subdivide into the same two lineages as primate KIR. Unlike the primates, the lineage of variable and likely functional cattle KIR corresponds to the KIR3DX1 lineage of primate KIR, whereas the variable lineage of primate KIR is represented in cattle by one KIR gene and a related gene fragment.


KIR receptorsCattleEvolutionPhylogeny

The killer cell immunoglobulin-like receptors (KIR) of primates are diverse, polymorphic major histocompatibility complex class I receptors expressed by subpopulations of natural killer (NK) cells and T cells (Uhrberg et al. 2001). The family of genes encoding these variable and rapidly evolving KIR is located in the leukocyte receptor complex (LRC), flanked on one side by the leukocyte immunoglobulin-like receptor (LILR) gene family, and on the other side by FCAR, the gene encoding the immunoglobulin A receptor (FcαRI) of myeloid cells (Wende et al. 1999; Wilson et al. 2000). An additional KIR gene, named KIR3DX1, was recently discovered in a location distinct from the other KIR. It is in the middle of the LILR gene family, between the two duplicated blocks of LILR genes and bounded by the LAIR2 and LILRA2 genes (Sambrook et al. 2006b). KIR3DX1 is conserved as a single-copy gene in four of the five primate species examined with an apparent species-specific duplication in the fifth. From sequencing KIR3DX1 cDNA from single human and chimpanzee individuals, Sambrook et al. 2006b identified limited polymorphism: with a single synonymous substitution between the two human alleles and five non-synonymous substitutions between the two chimpanzee alleles. KIR3DX1 is unique in not having the LTR33A/MLT1D element in intron 3 that is present in all other primate KIR (Martin et al. 2000).

KIR have been reported for several non-primate species. An analysis of horse cDNA provides evidence for a single EqcaKIR3DL gene (Takahashi et al. 2004). Genomic analysis showed that the pig has a single KIR gene (SuscKIR2DL1), placed between LILR and FCAR in the homologous position occupied by the expanded lineage of primate KIR (Sambrook et al. 2006a). SuscKIR2DL1 is of the same lineage as the expanded primate KIR and contains the LTR33A/MLT1D repeat in intron 3. This element also characterizes the rat KIR (Hoelsbrekken et al. 2003), which is present in the homologous location between LILR and FCAR, as well as the two mouse KIR, which are not in the corresponding location due to their transposition to the X chromosome (Hoelsbrekken et al. 2003; Welch et al. 2003). A search of GenBank revealed a single cat KIR gene that is flanked by LILR and FCAR and also contains the intron 3 element. This composite element comprises an LTR33A fragment interrupted by the insertion of a MLT1D element. Its presence in a wide range of mammalian KIR genes suggests that the insertion event occurred early in KIR evolution and that the inserted element is a useful lineage marker.

Phylogenetic analysis of full-length cDNA sequences shows the division of KIR into two lineages (Fig. 1a). One lineage (the 3DX lineage) contains the primate KIR3DX1 genes and all of the cattle KIR except BotaKIR2DL1. The second lineage (the 3DL lineage) contains all the other KIR. We estimated that the divergence of the two lineages (Fig. 1b) occurred 135.5 ± 10.5 million years ago (mya; 95% confidence interval of 116.7 to 158.3 mya). This places the duplication event that formed the ancestors of the two lineages to a time predating the radiation of placental mammals (95.3–113 mya), after the genesis of the monotremes (162.5–191.1 mya), and coincident with the emergence of the marsupials (124.3–138.8 mya; Benton and Donoghue 2007). The divergence time analysis was performed on the exon 4 sequences (encoding the D1 domain), as this exon provided the best resolved tree.
Fig. 1

KIR divide into two lineages because of an ancient duplication that preceded the diversification of placental mammals. a Phylogenetic analysis shows two lineages of KIR. The KIR3DX1 lineage comprises the KIR3DX1 genes of primates and all cattle KIR except BotaKIR2DL1; the KIR3DL lineage comprises all other KIR, including BotaKIR2DL1. Shown is a neighbor-joining (NJ) bootstrap analysis of full-length coding region KIR sequences (1,000 replicates, pair-wise deletion comparisons, Tamura–Nei model) that was performed using Mega 3.1 (Kumar et al. 2004). The tree was rooted at the midpoint and the support for each node (expressed as a percentage) is shown when > 50%. Performing the analysis after removal of exon 5 sequences (shown to be homogenized in the cattle sequences) does not alter the tree topology but does increase the bootstrap support at the deeper nodes. b NJ phylogenetic tree of exon 3 (encoding the D0 domain) and exon 4 (encoding the D1 domain) used to estimate the divergence time for the 3DL/3DX duplication. Phylogenetic analysis was performed as for (a). The D0 group (outgroup) was collapsed for simplification. The root of the ingroup tree, which corresponds to the duplication between 3DL and 3DX, is marked by a filled rectangle; the observed divergence time (± SD and with 95% confidence interval in brackets) is also indicated. KIR D0 and D1 sequences were aligned using MAFFT (Katoh et al. 2002) and corrected manually. Divergence time analysis was performed using the Bayesian relaxed molecular clock approach with the MULTIDISTRIBUTE program package (Thorne et al. 1998; Thorne and Kishino 2002), as previously described (Abi-Rached and Parham 2005). The root of the ingroup tree was set to 213 ± 59 mya to cover the range 95–331 mya. The lower end corresponds to the minimum divergence time for the human and cattle split (Benton and Donoghue 2007) as both species possess 3DX and 3DL; the upper end corresponds to the maximum divergence time for the mammal–bird split (Benton and Donoghue 2007), as the emergence of the mammalian LRC gene families occurred after this split (Nikolaidis et al. 2005). Eight internal calibration points were also used (marked with filled circles); minimum and maximum divergence times were setup according to fossil data (Benton and Donoghue 2007). The datasets used for the analyses are available upon request. The elephant KIR used in the analysis were obtained from the Ensembl assembly, scaffolds 7872 (elephant (1)) and 162321 (elephant (2))

Outside of the primates, the only species reported to have an expanded KIR gene family is cattle, for which four functional KIR genes and one pseudogene were described from analysis of cDNA (McQueen et al. 2002; Storset et al. 2003). In addition, McQueen et al. looked for the presence of the intron 3 MLT1D/LTR33A repeat in the bovine KIR by polymerase chain reaction amplification of genomic DNA followed by cloning and sequencing of the introns 3. They described two clones that contained the MLT1D/LTR33A element but were unable to assign them to a locus. To assess the relationships of the cattle KIR to the two lineages of primate KIR, we searched the database from the cattle genome project (Bovine genome build 2.1 at for further KIR sequences. We identified 22 KIR-containing contigs: Nine contained one or more full-length KIR genes along with flanking sequences, 13 contained partial KIR gene sequences.

Phylogenetic analyses were performed on exons 1 + 2 jointly, exons 3, 4, 5, and 6 individually, and exons 7–9 jointly to assess the relationship of the cattle KIR genes to each other and to the previously described cattle KIR cDNA sequences (McQueen et al. 2002; Storset et al. 2003). For all the phylogenetic trees, except that for exon 5, two deeply diverging lineages were observed; the first lineage includes the previously defined BotaKIR2DL1, the second lineage includes the four other previously defined cattle KIR: BotaKIR2DS1, BotaKIR3DL1, BotaKIR3DS1, and BotaKIR3DL1p (Fig. 2).
Fig. 2

Phylogenetic analysis groups cattle KIR into two divergent lineages. NJ analysis and tree display were performed as described in the legend to Fig. 1. Cattle KIR sequences obtained from the genome database are designated by their GenBank contig identifier. A c following the identifier indicates that the reverse complement of the contig was used. For contigs with multiple KIR genes, the genes are numbered in order of their appearance in the contig, gene fragments interrupted by a gap in the contig are indicated by a number followed by A. Previously defined BotaKIR are indicated by white-lettering on black boxes. a Analysis of exon 3 shows the division of the cattle KIR into two divergent lineages, BotaKIR2DL-related KIR (dark gray box) and all other BotaKIR. The later divide into four sub-lineages (light gray boxes). b Phylogenetic analysis of exons 7–9 show the divergence between BotaKIR2DL1 and all other BotaKIR. c Phylogenetic analysis of exons 7–9 of BotaKIR, excluding BotaKIR2DL1-related KIR, reveals two sub-lineages corresponding to KIR encoding short (upper group) and long (lower group) cytoplasmic tails

Figure 2a shows the results of the analysis for exon 3. In addition to showing the divergence of the two main KIR lineages, the tree defines four sub-lineages within the second lineage of cattle KIR. These relationships were similarly observed in the tree of exon 4 sequences (data not shown). The analyses of exons 1 + 2 and exon 6 also revealed the two main cattle KIR lineages but not the separation into sub-lineages, likely due to the small size of these exons. Analysis of exons 7–9, encoding the transmembrane and cytoplasmic tail of the KIR, also showed the lineages corresponding to the BotaKIR2DL1-related sequences vs all others (Fig. 2b). When the BotaKIR2DL1-related KIR were excluded from the analysis of exons 7–9, the resulting tree of the other lineage of cattle KIR showed two sub-lineages corresponding to the KIR having short and long cytoplasmic tails (Fig. 2c).

Phylogenetic trees of exon 5 of the BotaKIR, encoding the D2 domain, lacked the divergent lineages characteristic of the other exons. Comparing the BotaKIR, exon 5 sequences with exon 5 sequences of other mammalian KIR shows that the cattle sequences form a monophyletic group with very strong support. Disrupting this group, to cluster BotaKIR2DL1 with the other 3DL-lineage sequences and the remaining cattle KIR with the 3DX1-lineage sequences results in a tree topology with a markedly reduced parsimony score (Templeton test, α = 0.001). This supports the hypothesis that all of the cattle KIR exon 5 sequences share a unique common ancestor, likely due to a recombination that replaced the exon 5 sequence of one lineage with that of the other. Further analysis, including sequences derived from other species, will be necessary to resolve the direction of the recombination.

Based on the phylogenetic analyses, the repetitive elements present in the introns, and the position in the contig, the cattle KIR gene sequences were sorted into 13 groups (Table 1). Four of the five cattle KIR defined previously correspond to groups 1, 3, 4, and 5, and the fifth is likely represented by either group 6 or 7, both of which are partial sequences. The large number of groups is consistent with the number of hybridizing bands observed in Southern blot analysis of cattle KIR (McQueen et al. 2002). The BotaKIR2DL1 gene corresponds to group 1 and is divided between two nonoverlapping contigs. Group 2 represents other KIR of the BotaKIR2DL1 lineage. Intron 3 of these genes contains the 5′ portion of the LTR33A/MLT1D element that is characteristic of all primate KIR except KIR3DX1 (Martin et al. 2000; Sambrook et al. 2005). Truncation of this element in cattle KIR is the result of a deletion beginning 300 bp downstream of the end of exon 3 and continuing through exon 4 to leave only the final 77 bp of the exon. This is consistent with the finding from the cDNA analyses that the BotaKIR2DL1 transcript lacked sequences corresponding to the D1 domain (McQueen et al. 2002; Storset et al. 2003).
Table 1

Thirteen gene groups identified in the cattle KIR sequences

Genes were grouped by similarity in the phylogenetic analysis as well as similarity in intron sequences. The first column shows the number assigned to each group. The second column indicates the identity of the gene when known. The last column indicates the source of the sequence with the contig name (NW_#) as assigned by GenBank, followed by a ‘c’ when the reverse complement of the contig was used. For contigs with multiple genes, the genes are numbered in order of their appearance in the contig, gene fragments interrupted by a gap in the contig are indicated by a number followed by ‘A’. The central nine columns indicate the presence or absence of exonic sequences. A white box indicates absence of a sequence, a shaded box its presence. Shading indicates lineage relationship: BotaKIR2DL1-related are dark gray, KIR3DX1-related are black, and BotaKIR2/3DS-related are light gray. An X indicates the presence of a mutation predicted to disrupt the reading frame, and a P indicates the presence of a partial exon. The group designated as U is related to both groups 9 and 11 but could not be assigned to either.

Groups 3–13 in Table 1 correspond to the 3DX1 lineage cattle KIR genes. As with the primate KIR3DX1, the introns 3 of these genes do not contain the LTR33A/MLT1D element. Thus, the variable lineage of functional KIR in cattle is the result of diversification of the 3DX1 lineage. This contrasts with the situation in primates, where the other KIR lineage expanded and KIR3DX1 remained a conserved single-copy gene.

Analysis of the gene order in those contigs containing multiple KIR genes provided preliminary mapping information. Contig NW_943617 contains the 3′ portion of BotaKIR3DL1 linked to the 5′ portion of BotaKIR2DL1 with an intergenic interval of approximately 2 kb, a distance similar to that observed in the primate KIR region (Wende et al. 1999; Wilson et al. 2000). BotaKIR3DL1p is linked to the BotaKIR2DL1-related gene fragment with a similar 2-kb intergenic interval in three separate contigs (NW_938467, NW_938583, and NW_942117). Two additional linkages were observed that had contiguous sequence for the intergenic region, group 12 to group 5 (BotaKIR3DS1) in NW_937716 and group 7 to group 9 in NW_941846. The intergenic interval of these latter two pairs was approximately 7 kb, much longer than that between BotaKIR3DL1 and BotaKIR2DL1.

Recombination appears to have changed the gene order in the cattle and primate LRCs. The linkage of BotaKIR3DL1 to BotaKIR2DL1 is different from the map positions observed for the two divergent lineages in primates, where KIR3DX1 is separated from the 3DL-lineage KIR containing the LTR33A/MLT1D element by genes of the LILR family (Sambrook et al. 2006b), a distance of approximately 180 kb. NW_001003629 provides evidence that cattle LILR are situated on the other side of the KIR genes compared to primates and other species.

So far, cattle have stood out as the only non-primate species in which the KIR genes have expanded through gene duplication and diversification. In this paper, we have shown that that cattle KIR comprise two lineages. The lineage that is related to the highly diversified primate KIR has limited diversity. In contrast, the highly diversified cattle KIR are related to primate KIR3DX1, a conserved single-copy gene for which the function is unknown.


The research described in this paper was supported by NIH grant AI024258 to P.P.

Copyright information

© Springer-Verlag 2007