Functionally important glycosyltransferase gain and loss during catarrhine primate emergence

Abstract

A glycosyltransferase, alpha1,3galactosyltransferase, catalyzes the terminal step in biosynthesis of Galalpha1,3Galbeta1-4GlcNAc-R (alphaGal), an oligosaccharide cell surface epitope. This epitope or antigenically similar epitopes are widely distributed among the different forms of life. Although abundant in most mammals, alphaGal is not normally found in catarrhine primates (Old World monkeys and apes, including humans), all of which produce anti-alphaGal antibodies from infancy onward. Natural selection favoring enhanced resistance to alphaGal-positive pathogens has been the primary reason offered to account for the loss of alphaGal in catarrhines. Here, we question the primacy of this immune defense hypothesis with results that elucidate the evolutionary history of GGTA1 gene and pseudogene loci. One such locus, GGTA1P, a processed (intronless) pseudogene (PPG), is present in platyrrhines, i.e., New World monkeys, and catarrhines but not in prosimians. PPG arose in an early ancestor of anthropoids (catarrhines and platyrrhines), and GGTA1 itself became an unprocessed pseudogene in the late catarrhine stem lineage. Strong purifying selection, denoted by low nonsynonymous substitutions per nonsynonymous site/synonymous substitutions per synonymous site values, preserved GGTA1 in noncatarrhine mammals, indicating that the functional gene product is subjected to considerable physiological constraint. Thus, we propose that a pattern of alternative and/or more beneficial glycosyltransferase activity had to first evolve in the stem catarrhines before GGTA1 inactivation could occur. Enhanced defense against alphaGal-positive pathogens could then have accelerated the replacement of alphaGal-positive catarrhines by alphaGal-negative catarrhines. However, we emphasize that positively selected regulatory changes in sugar chain metabolism might well have contributed in a major way to catarrhine origins.

Fig. 1.

PAML-estimated ω (i.e., dN/dS) and (N*dN, S*dS) values under the M1 “free ratio” model, which allows the ω value to vary on each branch. Estimated number of nonsynonymous and synonymous sites under this model are 827.4 and 318.6, respectively. Branches leading to unprocessed (UPG) and processed (PPG) loci are depicted in black. Branches leading to GGTA1 genes are depicted in color, with green indicating those branches with dN/dS values <0.5 (i.e., purifying selection) and red indicating those branches with dN/dS values >0.5 (i.e., relaxed purifying selection). Depicted in gray is the branch on which the UPG locus originated. Tree topology also reflects the optimal maximum likelihood and Bayesian phylogenetic results (−lnL = 7125.71 and 7140.76, respectively); maximum parsimony bootstrap results differed only in showing a trichotomy among the active New World monkey genes.

Fig. 2.

Phylogeny of the GGTA1 locus. Shown in black are the lineages in which GGTA1 remained active. Shown in red are the lineages in which GGTA1 was inactivated. Blue and red arrows indicate, respectively, the origin of the processed pseudogene (PPG) and the origin of the unprocessed pseudogene (UPG). The latter event corresponds to the time of GGTA1 inactivation that occurs in the stem-catarrhine lineage and precedes the last common ancestor of living catarrhines by only a few million years. The vertical line at the left indicates time (millions of years ago).