Phylogenetic relationships and the evolution of regulatory gene sequences in the parrotfishes

Abstract

Regulatory genes control the expression of other genes and are key components of developmental processes such as segmentation and embryonic construction of the skull in vertebrates. Here we examine the variability and evolution of three vertebrate regulatory genes, addressing issues of their utility for phylogenetics and comparing the rates of genetic change seen in regulatory loci to the rates seen in other genes in the parrotfishes. The parrotfishes are a diverse group of colorful fishes from coral reefs and seagrasses worldwide and have been placed phylogenetically within the family Labridae. We tested phylogenetic hypotheses among the parrotfishes, with a focus on the genera Chlorurus and Scarus, by analyzing eight gene fragments for 42 parrotfishes and eight outgroup species. We sequenced mitochondrial 12s rRNA (967 bp), 16s rRNA (577 bp), and cytochrome b (477 bp). From the nuclear genome, we sequenced part of the protein-coding genes rag2 (715 bp), tmo4c4 (485 bp), and the developmental regulatory genes otx1 (672 bp), bmp4 (488bp), and dlx2 (522 bp). Bayesian, likelihood, and parsimony analyses of the resulting 4903 bp of DNA sequence produced similar topologies that confirm the monophyly of the scarines and provide a phylogeny at the species level for portions of the genera Scarus and Chlorurus. Four major clades of Scarus were recovered, with three distributed in the Indo-Pacific and one containing Caribbean/Atlantic taxa. Molecular rates suggest a Miocene origin of the parrotfishes (22 mya) and a recent divergence of species within Scarus and Chlorurus, within the past 5 million years. Developmentally important genes made a significant contribution to phylogenetic structure, and rates of genetic evolution were high in bmp4, similar to other coding nuclear genes, but low in otx1 and the dlx2 exons. Synonymous and non-synonymous substitution patterns in developmental regulatory genes support the hypothesis of stabilizing selection during the history of these genes, with several phylogenetic regions of accelerated non-synonymous change detected in the phylogeny.

Figure 1

Phylogenetic hypothesis for the parrotfishes calculated using Bayesian inference. Maximum posterior probability topology resulting from 4 runs of a MCMC analysis each with 6 chains running for 10 million generations. Posterior probabilities are shown for each node (computed from majority rule consensus of post burn-in trees), and the black dots indicate support at the 90% level or greater from a 1000 repetition parsimony bootstrap analysis. The root of the seagrass and reef clades and four clades of the genus Scarus are identified. Photos are by John E. Randall (via Fishbase.org) with permission, identified by initials of genus and species.

Figure 2

A comparison of phylogenetic signal in the data partitioned by regulatory (otx1, dlx2, bmp4) and non-regulatory (12s, 16s, cytb, tmo4c4, rag2) genes. Topologies illiustrated are the majority-rule consensus tree of the 95% credible set of trees from 3 runs of a MCMC analysis each with 5 chains running for 10 million generations. Black dots (•) indicate strongly supported nodes at which both partitioned trees agree with each other and with the combined data tree. Strongly supported nodes that agree with the combined data tree but are not strongly supported in the other partition tree (+) are less frequent, and strongly supported nodes that disagree with the combined data tree (*) are rare.

Figure 3

Phylogram of rates of molecular evolution in the (A) 12s and (B) cytochrome b data partitions illustrated on the Bayesian consensus tree of all concatenated sequences. Numbers on selected branches indicate that the branchlength is N times the length shown.

Figure 4

Rates of molecular evolution in the (A) rag2 and (B) bmp4 data partitions illustrated on the Bayesian consensus tree of all concatenated sequences. Numbers on selected branches indicate that the branchlength is N times the length shown.

Figure 5

Rates of molecular evolution in the (A) dlx2 exon and (B) dlx2 intron data partitions illustrated on the Bayesian consensus tree of all concatenated sequences. Numbers on selected branches indicate that the branchlength is N times the length shown.

Figure 6

Synonyomus (A) and non-synonymous (B) substitution rate trees for cytochrome b, calculated using the Bayesian consensus topology and the Muse-Gaut (1994) codon substitution model implemented in HyPhy (Kosokovsky Pond et al., 2005).

Figure 7

Synonyomus (A) and non-synonymous (B) substitution rate trees for rag2, calculated using the Bayesian consensus topology and the Muse-Gaut (1994) codon substitution model.

Figure 8

Synonyomus (A) and non-synonymous (B) substitution rate trees for bmp4, calculated using the Bayesian consensus topology and the Muse-Gaut (1994) codon substitution model.

Figure 9

Chronogram illustrating the ages of origin and diversification of parrotfish phylogenetic groups, calculated using the Bayesian inference topology and the Bayesian posterior probability algorithms for mean and variance of nodal ages implemented in CodonRates 1.0 (Seo et al., 2004) assuming a root age of the labrid outgroups of 55 million years ago.