Genes and Immunity (2005) 6, 398–406 & 2005 Nature Publishing Group All rights reserved 1466-4879/05 $30.00 www.nature.com/gene FULL PAPER Evidence for natural selection in the HAVCR1 gene: high degree of amino-acid variability in the mucin domain of human HAVCR1 protein T Nakajima1,2, S Wooding3, Y Satta4, N Jinnai1, S Goto1, I Hayasaka5, N Saitou6, J Guan-jun7, K Tokunaga8, LB Jorde3, M Emi2 and I Inoue1 1Division of Genetic Diagnosis, The Institute of Medical Science, The University of Tokyo, Tokyo, Japan; 2Department of Molecular Biology, Institute of Gerontology, Nippon Medical School, Kawasaki, Japan; 3Department of Human Genetics, University of Utah Health Sciences Center, Salt Lake City, UT, USA; 4Department of Biosystems Science, Graduate University for Advanced Studies, Hayama, Japan; 5Kumamoto Primates Park, Sanwa Kagaku Kenkyusho Co. Ltd, Kumamoto, Japan; 6Division of Population Genetics, National Institute of Genetics, Mishima, Japan; 7Red Cross Blood Center of Harbin, Harbin, China; 8Department of Human Genetics, Graduate School of Medicine, The University of Tokyo, Tokyo, Japan The family of genes encoding T-cell immunoglobulin and mucin-domain containing proteins (Tim), which are cell-surface molecules expressed in CD4 þ T helper cells, has important roles in the immune system. Here, we report three unusual patterns of genetic variation in the human hepatitis A virus cellular receptor 1 gene (HAVCR1) that are similar to patterns observed in major histocompatibility complex loci. First, levels of polymorphism in exon 4 of HAVCR1 were exceptionally high in humans (nucleotide diversity (p) ¼ 45.45 Â 10À4). Second, nonsynonymous substitutions and insertion/deletion variants were more frequent than synonymous substitutions in that exon (10 out of 12 variants). The rate of the mean number of nucleotide substitutions at nonsynonymous sites to synonymous sites at HAVCR1-exon 4 is 41(PA/PS ¼ 1.92 and pA/pS ¼ 2.23). Third, levels of divergence among human, chimp, and gorilla sequences were unusually high in HAVCR1-exon 4 sequences. These features suggest that patterns of variation in HAVCR1 have been shaped by both positive and balancing natural selection in the course of primate evolution. Evidence that the effects of natural selection are largely restricted to the mucin domain of HAVCR1 suggests that this region may be of particular evolutionary and epidemiological interest. Genes and Immunity (2005) 6, 398–406. doi:10.1038/sj.gene.6364215; published online 12 May 2005 Keywords: HAVCR1; Tim gene family; natural selection Introduction ponses and facilitates the development of immuno- logical tolerance.4,5 The TIM gene family, HAVCR1 (human hepatitis A virus HAVCR1, the first TIM family member to be identified, cellular receptor 1), HAVCR2, and TIMD4, encodes a was initially recognized as a receptor for the hepatitis A small group of type I transmembrane glycoproteins virus (HAV)6 and was later implicated in susceptibility to composed of an immunoglobulin-like domain positioned allergic asthma in mice7 and humans.8 This evidence, atop a mucin-like domain. This structure is similar to along with the participation of other TIM genes in the that of several known adhesion proteins, and points to a 1,2 human immune system, suggests that functional varia- possible role of TIM proteins as adhesion molecules. tion in HAVCR1 could be a particularly important The expression of these genes in T helper (Th) cells determinant of the immune response. Yet, little is known suggests that these proteins might have particularly about patterns of functional variation in this gene. important roles in regulating the immune response. For Information about the presence of functional variation example, HAVCR2, which is highly expressed by Th1 in genes can be obtained in a variety of ways. In humans, cells, appears to be involved in the differentiation of CD4 þ Th cells and the activation of macrophages.3 The linkage and association studies are often used to identify HAVCR2 pathway also provides an important mechan- genotype–phenotype correlations; in mice, targeted gene ism for downregulating Th1-dependent immune res- knockouts are often useful. While these approaches have been highly successful, they can be prohibitively ex- pensive. A major goal of this study was to generate new Correspondence: Dr T Nakajima, Department of Molecular Biology, hypotheses about the presence of functional polymorph- Institute of Gerontology, Nippon Medical School, 1-396 Kosugi-cho, Nakahara-ku, Kawasaki 211-8533, Japan. E-mail: [email protected] isms by testing for signatures of natural selection in the Received 18 January 2005; revised 15 March 2005; accepted 22 HAVCR1 gene. As natural selection acts on variable March 2005; published online 12 May 2005 phenotypes, the effects of natural selection will be Evidence for natural selection in the HAVCR1 gene T Nakajima et al 399 strongest on those variants that confer such phenotypes. cells implies that Tim proteins are likely to participate in Thus, evidence of natural selection in a genomic region immune regulation in the human. suggests that functionally important variants are present in that region. By examining patterns of nucleotide Nucleotide polymorphisms in human HAVCR1, diversity both within humans and between humans, HAVCR2, and TIMD4 genes chimpanzees, and gorillas, we hoped to identify those Patterns of DNA sequence variation were evaluated in sites most likely to confer phenotypic variation. We all exons of HAVCR1, HAVCR2, and TIMD4 in 281 found that signatures of balancing natural selection are human samples (94 African-American, 92 Caucasian, present in HAVCR1; however, these signatures are and 95 Japanese). A total of 28 sequence variants were localized to a region encoding an extracellular mucin identified: 17 in HAVCR1, three in HAVCR2, and eight in domain that is likely involved in the recognition of TIMD4 (Figure 2 and Table 1). Many of the nucleotide molecules outside the cell. We hypothesize that these variants detected in coding regions were observed in variants confer phenotypic variation and are thus good exon 4 of HAVCR1 (12 sequence variations), a region candidates for use in future linkage and association encoding the N-terminal half of the mucin domain. Six of analyses. the 12 variants observed in exon 4 of HAVCR1 were nonsynonymous nucleotide substitutions. Four were in- frame insertion/deletion polymorphisms (one 18 bp/6aa deletion and three 3 bp/1aa deletions). The abundance of Results nonsynonymous nucleotide substitutions and insertion/ Expression of Tim-family genes in human CD4 þ Th deletion polymorphisms within one exon suggested an cells evolutionary history of natural selection, as in the case of CD4 þ Th cells were isolated from human peripheral antigen recognition sites in major histocompatibility blood using monoclonal antibodies to CD4 coupled to complex (MHC)-HLA genes.10 magnetic beads (MACS, Miltenyi Biotechnology, Au- To determine whether natural selection has occurred at burn, CA, USA), and stimulated in vitro under Th1- or HAVCR1-exon 4, we determined the complete sequences Th2-differentiating conditions by plate-bound anti-CD3e of all HAVCR1 exons and B1.5 kb of sequence in its along with IL-12 þ anti-IL-4 or by anti-CD28 with IL-4 þ promoter region among 624 individuals from 10 different anti-IFN-g þ anti-IL-12.9 As shown in Figure 1, human populations (73 native African, 94 African-American, HAVCR1, HAVCR2, and TIMD4 were highly expressed 49 Palestinian, 46 Druze, 46 Ashkenazi, 92 Caucasian, 51 in Th1-differentiating conditions 5 days after stimulation. Indian, 95 Japanese, 23 Korean, and 55 Northern Chinese HAVCR1 in the mouse is specifically expressed by (Heilongjian)). TA cloning was performed using TOPOTM differentiating and differentiated murine Th2 cells,7 and TA cloning kits (Invitrogen, Carlsbad, CA, USA) to the expression of Tim-family genes in human CD4 þ Th determine the gametic phase for individuals who were heterozygous in HAVCR1-exon 4. The level of nucleotide diversity (p) for this locus (45.45 Â 10À4 in all human samples) was unusually high, about three times greater a IFN- (ng/ml) than the value for noncoding promoter regions À4 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 (15.41 Â10 in all human samples; Table 2). The patterns of nucleotide substitutions at nonsynon- ymous sites and/or synonymous sites in HAVCR1-exon Th1 4 were evaluated. The mean number of nucleotide substitutions at nonsynonymous sites (PA) and synony- mous sites (PS) were evaluated based on 11 HAVCR1- Th2 exon 4 sequences identified in human population samples. The value of PA at HAVCR1-exon 4 (0.0085) is 0 50 100 150 200 250 300 350 400 much higher than the value for other coding regions of HAVCR1 (0.0011) and the ratio of PA to PS at HAVCR1- IL-5 (pg/ml) exon 4 is 41(PA/PS ¼ 1.92). A similar pattern was found in the mean number of b T-bet nucleotide substitutions at nonsynonymous sites (pA) GATA-3 and the mean number of nucleotide substitutions at 11,12 IFN- synonymous sites (pS) in human population samples. À4 As shown in Table 2, pA (54.09 Â 10 in all human IL-5 À4 samples) is much higher than pS (24.31 Â10 in all HAVCR1 human samples) and the ratio of pA to pS at HAVCR1- HAVCR2 exon 4 is 41(pA/pS ¼ 2.23). This pA/pS value is much higher than the value based on 13 582 bp of coding TIMD4 sequences from 106 genes (2.23 vs 0.64), described by GAPDH Cargill et al.13 Th1 Th2 Haplotype network of HAVCR1-exon 4 in human Figure 1 Expression of HAVCR1 in Th1-differentiated human populations CD4 þ Th cells. (a) Production of IFN-g and IL-5 in CD4 þ Th cells stimulated under Th1- or Th2-differentiating conditions in vitro.