Phylogeny of primate major histocompatibility complex DRB alleles

Abstract

ABBREVIATIONS ARS antigen recognition site In contrast to alleles at most other loci, alleles at the major histocompatibility complex (MHC) loci often dif- fer by more than one nucleotide substitution. It is there- fore legitimate to look into the lines of descent of the individual MHC alleles and to construct their phyloge- nies. One of the largest collections of MHC alleles thus far assembled is that of the class II DRB loci in pri- mates, specifically in the great apes. Marsh and Bodmer [1] recently compiled sequences of 42 human (HLA- DRB) alleles; we have obtained the sequences of four gorilla (l~lhcGogo-DRB, M. Kasahara and J. Klein, un- published data), three chimpanzee (MhcPatr.DRB, ref. 2), and two orangutan (MhcPopy-DRB, Ch. Sch6nbach, Z.-F. Zhu, and J. Klein, unpublished data.) ~ To deter- mine the relationships among these 51 sequences, we analyzed them using two methods of phylogenetic tree construction: the neighbor-joining method of Saitou and Nei [4] and the maximal parsimony method of Felsenstein [5]. Since the former produced more realis- tic trees than the latter, we shall limit our discussion to the tree obtained by the application of this method to the exon 2 (first domain) DRB sequences, using the l The designations of nonhuman primate MHCs accord with the recent nomenclatorial agreement; see ref. 3. Prom the Dopartment of Microbiology~ and Immunology, University of Miami School of Medicine, Miami, Florida (J.K.; M.K.), and the Max- Planck-lnstitut f~r Bidogie, Abteilung lmmungenetik, Tabingen, Ger- many (J.K.; J.G.; C.$.). Add~ss reprint requests to Professor Jan Klein, Max-Planck-lnstitut f~r Biologie, Abteilung lmmungenetik, D-7400 T~'bingen, Germany. Accepted November 16,1990. 28 0198-8859/91/$3.50 MHC major hismcompatibility complex mouse H-2Eb b sequence as a root (Fig. 1). The restric- tion to exon 2 sequences was dictated by two consider- ations: first, it is the only exon for which sequences of all the 51 alleles are available; and second, when comparing alleles, exons are better evolutionary units than whole genes because recombination may produce hybrid genes that could complicate the interpretation of allelic phylogenies. A remarkable feature of the tree in Fig. 1 is the inter- mingling of sequences from the four primate species, Homo sapiens, Pan troglodytes, Gorilla gorilla, and Pongo pygmeus. This intermingling provides further support for the notion that was first put forward as a hypotheti- cal possibility [6] but has since been supported by a growing body of observations [2,7-11], namely, that the history of many of the bIHC alleles extends beyond the species' boundaries. It can safely be concluded from the tree that all its main branches were established be- fore the divergence of chimpanzee, gorilla, and human lineages from a common ancestral lineage more than 5 million years (my) ago and that at least some of them existed as separate lineages before the split of orangutan ancestors from the gorilla-chimpanzee-human ancestors more than 10 my ago. If we accept that Homo sapiens emerged between 300,000 and 500,000 years ago [12], we must conclude that all alleles differing by more thau one substitution outside the antigen recognition site (ARS) were most likely established before this event [13]. All these alleles, therefore, predate speciation. A phylogenetic tree of the sort shown in Fig. 1 is constructed to reflect the succession of substitutions at Human Immunology 31, 28-35 (1991) © American $ociet,/for Histocompatibillty and lmmunogenetics, 1991