Ancient genome duplications did not structure the human Hox-bearing chromosomes

Abstract

The fact that there are four homeobox (Hox) clusters in most vertebrates but only one in invertebrates is often cited as evidence for the hypothesis that two rounds of genome duplication by polyploidization occurred early in vertebrate history. In addition, it has been observed in humans and other mammals that numerous gene families include paralogs on two or more of the four Hox-bearing chromosomes (the chromosomes bearing the Hox clusters; i.e., human chromosomes 2, 7, 12, and 17), and the existence of these paralogs has been taken as evidence that these genes were duplicated along with the Hox clusters by polyploidization. We tested this hypothesis by phylogenetic analysis of 42 gene families including members on two or more of the human Hox-bearing chromosomes. In 32 of these families there was evidence against the hypothesis that gene duplication occurred simultaneously with duplication of the Hox clusters. Phylogenies of 14 families supported the occurrence of one or more gene duplications before the origin of vertebrates, and of 15 gene duplication times estimated for gene families evolving in a clock-like manner, only six were dated to the same time period early in vertebrate history during which the Hox clusters duplicated. Furthermore, of gene families duplicated around the same time as the Hox clusters, the majority showed topologies inconsistent with their having duplicated simultaneously with the Hox clusters. The results thus indicate that ancient events of genome duplication, if they occurred at all, did not play an important role in structuring the mammalian Hox-bearing chromosomes.

Figure 1

Summary of phylogenetic analyses of families including members on at least two of human Hox-bearing chromosomes (2, 7, 12, and 17), which indicated that one or more gene duplications occurred before the origin of vertebrates or after the mammalian radiation. For each family, the timing of duplications is indicated relative to a phylogeny of organisms, and the chromosomal locations of the corresponding human genes are indicated; e.g., the notation “CDK (12) (7,12,17)” indicates that a member of the CDK family on human chromosome 12 duplicated from the ancestor of other family members on human chromosomes 7, 12, and 17 before the divergence of fungi from animals. The bootstrap confidence level (BCL) for the internal branches of the phylogenetic trees are color-coded. The time window within which the vertebrate Hox clusters are believed to have duplicated is also indicated. Names of gene families are abbreviated as in Table 1.

Figure 2

Phylogenetic tree of the cyclin-dependent kinase (CDK) family, constructed by the neighbor-joining method, on the basis of the proportion of amino acid difference (p). Numbers on the branches are the bootstrap confidence levels; only values ≥50% are shown. For genes on the human Hox-bearing chromosomes, chromosomal locations are indicated by circled numbers. Species included in the phylogeny are the following: Vertebrata: human (Homo sapiens), mouse (Mus musculus), chicken (Gallus gallus), clawed frog (Xenopus laevis), goldfish (Carassius auratus); Insecta: Drosophila melanogaster; Nematoda: Caenorhabditis elegans; Plantae: Arabidopsis thaliana, rice (Oryza sativa); Fungi: yeast (Saccharomyces cerevisiae), Schizosaccharomyces pombe, Kluyveromyces lactis.

Figure 3

Estimated duplication times (±S.E.) of pairs of paralogous genes on human Hox-bearing chromosomes, plotted on a timeline illustrating divergence times of major vertebrate taxa (Kumar and Hedges 1998). Only families for which the hypothesis of a molecular clock could not be rejected were included. The time window within which the vertebrate Hox clusters are believed to have duplicated is indicated by shading. Names of gene families are abbreviated as in Table 1.

Figure 4

Schematic phylogenies of families having members on all four Hox-bearing chromosomes that duplicated early in vertebrate history: (a) Hox clusters (Zhang and Nei 1996); (b) collagen genes (Bailey et al. 1997); (c) ERBB family; (d) phylogeny consistent with the phylogenies of most three-member families (Fig. 5). Numbers on the branches are the bootstrap confidence levels; only values ≥50% are shown. Chromosomal locations are indicated by circled numbers.

Figure 5

Summary of rooted phylogenies of families of genes including members on three of the four human Hox-bearing chromosomes. In each case, the bootstrap confidence level (BCL) of the internal branch is indicated in parentheses. Names of gene families are abbreviated as in Table 1. (INTA-PS1) Vertebrate integrin α chains in the subfamily related to Drosophila integrin α-PS1; (INTA-PS2) vertebrate integrin α chains in the subfamily related to Drosophila integrin α-PS2 (Hughes 2001). For NOS phylogeny, see Hughes (1998b).