MBE Advance Access originally published online on November 28, 2007
Molecular Biology and Evolution 2008 25(2):352-361; doi:10.1093/molbev/msm260
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Published by Oxford University Press 2007.
Research Articles |
Duplication of Accelerated Evolution and Growth Hormone Gene in Passerine Birds
,1
* Department of Vertebrate Zoology, National Museum of Natural History, Smithsonian Institution, 4210 Silver Hill Road, Suitland, MD
Department of Zoology, University of Florida
Behavior, Ecology, Evolution, and Systematics Program, University of Maryland
E-mail: tyuri{at}ufl.edu
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
We report the discovery of a duplication of the growth hormone (GH) gene in a major group of birds, the passerines (Aves: Passeriformes). Phylogenetic analysis of 1.3-kb partial DNA sequences of GH genes for 24 species of passerines and numerous outgroups indicates that the duplication occurred in the ancestral lineage of extant passerines. Both duplicates and their open-reading frames are preserved throughout the passerine clade, and both duplicates are expressed in the zebra finch brain, suggesting that both are likely to be functional. The estimated rates of amino acid evolution are more than 10-fold higher in passerine GH genes than in those of their closest nonpasserine relatives. In addition, although the 84 codons sequenced are generally highly conserved for both passerines and nonpasserines, comparisons of the nonsynonymous/synonymous substitution ratios and the rate of predicted amino acid changes indicate that the 2 gene duplicates are evolving under different selective pressures and may be functionally divergent. The evidence of differential selection, coupled with the preservation of both gene copies in all major lineages since the origin of passerines, suggests that the duplication may be of adaptive significance, with possible implications for the explosive diversification of the passerine clade.
Key Words: relaxation of selection • positive selection • subfunctionalization • Passeriformes
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Growth hormone (GH) is a polypeptide hormone found in all vertebrate lineages (Kawauchi et al. 2002
). It is generally encoded by a single gene (Agellon et al. 1988; Rentier-Delrue et al. 1989), is released from pituitary somatotrophs into the circulation, and exerts actions that promote growth and differentiation at distant target sites (Etherton and Bauman 1998). However, it is less well known that GH is also produced in many extrapituitary sites (Harvey and Hull 1997) where it may participate as a local growth factor or cytokine in the autocrine/paracrine regulation of cellular differentiation during embryonic development (Waters et al. 1999; Sanders and Harvey 2004). In birds, it is also involved in a variety of important secondary functions such as egg production, aging, and reproduction (Aramburo et al. 2000; Ip et al. 2001; Zhao et al. 2004). Because of its important functions, and perhaps the constraints imposed by multiple functions, the evolutionary rate of GH has generally been slow. However, bursts of rapid changes have been noted in some mammals, amphibians, and teleosts (Wallis 1996; Wallis OC and Wallis M 2001).
In mammals, 2 particularly marked episodes of rapid change have occurred, in the Cetartiodactyla (Cetacea plus Artiodactyla, sensu Montgelard et al. 1997; Wallis OC and Wallis M 2001; Maniou et al. 2004) and primates (e.g., Wallis 1981, 1994; Ohta 1993; Liu et al. 2001). Interestingly, duplications of the GH gene have been reported within both of these mammalian groups. Some caprine ruminants appear to have 2 GH-like genes (Wallis et al. 1998), and, in higher primates, a series of duplications have given rise to a cluster of GH-related genes, several of which are expressed in the placenta (Chen et al. 1989; Wallis OC and Wallis M 2002). There are also several cases of duplicated GH genes in amphibians and teleosts, which may be associated with tetraploidy (Devlin 1993; Huang and Brown 2000; McKay et al. 2004).
In the course of our work on Early Bird, a large-scale, collaborative project to determine the interrelationships of all major groups of birds (http://www.fieldmuseum.org/research_collections/zoology/zoo_sites/early_bird), we discovered a duplication of the GH gene in passerines, or perching birds. The passerines are the largest order of birds, comprising more than half of all living avian species, and 2 copies of the GH gene are present throughout the clade. Our analyses suggest that both duplicates have evolved rapidly since the duplication event and are under different selective pressure from the original single-copy GH gene. This is the first case of GH gene duplication reported in birds or in Diapsida (birds and traditional reptiles).
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
DNA Sequence Data Collection
Our sample includes 24 passerine species and 138 outgroup taxa that represent the diversity of extant avian taxa (names and sources in supplementary table S1, Supplementary Material online). Approximately 1.3-kb sequences of the GH gene, including complete intron 2, exon 3, intron 3, and flanking regions of exons 2 and 4 (based on the Chicken Genome: NC006114, International Chicken Genome Sequencing Consortium 2004
) were collected. The target fragment was amplified by a nested, 2-step polymerase chain reaction (PCR), using the first set of primers GH-F874 (5'-CCTTCCCWGCCATGCCCCTTTCCAACC-3') and GH-R3108 (5'-CCGTAGTTCTTCAGCAGGGCSTCCTCG-3'), followed by the second set of primers GH-F897 (5'-TGTTTGCCAACGCTGTGCTGAGG-3') and GH-R1925 (5'-TCCCTTCTTCCAGGTCCTTTART-3'). The resulting PCR products were sequenced using primers GH-F897, GH-R1925, GH-F1391 (5'-GATGTCTCCACAGGAACGYA-3') and GH-1476 (5'-GATTTCTGCTGGGCATCATCCTTCC-3'). In some cases, additional taxon-specific primers were designed to sequence intron regions. Those primer sequences may be obtained from the authors.
Our standard PCR amplifications were performed using DNA Engine Tetrad Thermal Cyclers (MJ Research, now Bio-Rad, Hercules, CA) as follows: (1) The first reaction using the primer pair GH-F874 and GH-R3108 was performed in 12.5 µl final volume containing 10–20 ng genomic DNA, 0.25 µM of each primer, 0.2 mM dNTPs, 1.25 µl Ex Taq buffer, and 0.5 U of Takara Ex Taq (Takara Bio, Madison, WI), using a "touchdown" cycling program with 10 cycles of denaturation at 94 °C for 30 s, annealing at 70 °C 
61 °C (1 °C decrease per cycle) for 30 s, and extension at 72 °C for 2–3 min, followed by 30 cycles of denaturation at 94 °C for 30 s, annealing at 60 °C for 30 s, and extension at 72 °C for 2–3 min. (2) The second reaction using the primer pair GH-F897 and GH-R1925 was performed in 50 µl final volume containing 1 µl of the diluted PCR product (1 to 1/100 in dilution) from the first reaction, 0.25 µM of each primer, 0.2 mM dNTPs, 5 µl PCR buffer (standard 10x buffer, GeneChoice, Frederick, MD, or Biolase NH4 reaction buffer, Bioline, Taunton, MA), 1.5 mM MgCl2 (only with Biolase reaction buffer), and 1.25 U of Taq (GeneChoice or Bioline), using a cycling program of 27 cycles of denaturation at 94 °C for 30 s, annealing at 60 °C for 30 s, and extension at 72 °C for 90 s. Sequencing reactions were performed with ABI BigDye Terminator v3.1 Cycle Sequencing Kits, and the resulting products were analyzed on ABI 3100 or 3130xl Genetic Analyzers. DNA sequences used in this study were deposited in GenBank under accession numbers EF521416
[GenBank]
–EF521598
[GenBank]
.
Phylogenetic Inference of GH Gene Tree
Alignment of all GH gene sequences was performed using ClustalX followed by manual adjustment. The aligned sequences were analyzed phylogenetically to reconstruct a gene tree and to calculate bootstrap support of its nodes using GARLI v0.951 (Zwickl 2006; http://www.bio.utexas.edu/faculty/antisense/garli/Garli.html), which performs heuristic phylogenetic searches under the general time reversible (GTR) model of nucleotide substitution. It uses a genetic algorithm approach to simultaneously find the topology, branch lengths, and model parameters that maximize the log likelihood score (Lewis 1998). For all our analyses, the default settings of GARLI were used (with base frequencies, 4-category 
-distributed rate heterogeneity, and a proportion of invariant sites estimated). The resulting tree topology was confirmed by Bayesian analysis using MrBayes v3.1 (Huelsenbeck and Ronquist 2001; Ronquist and Huelsenbeck 2003) and the same model structure as GARLI, but with model parameters estimated separately for 4 partitions of the data: introns and codon positions 1, 2, and 3 of exons. Two sets of 4 Markov chains were run for 10 million generations sampling every 100 generations. The convergence of the 2 sets of analyses was confirmed by the correlation between posterior probabilities for the 2 analyses and the potential scale reduction factor (Gelman and Rubin 1992) approaching 1 for each parameter. The first 2000 samples were discarded as burn-in. The position of zebra finch (Taeniopygia guttata) GH genes on the tree was estimated by adding two 254-bp exon sequences from the Songbird EST project (http://titan.biotec.uiuc.edu/cgi-bin/ESTWebsite/estima_start?seqSet=songbird) to the data set and repeating the procedures above.
Evolutionary Rate Comparisons
Relative rate tests (RRTs) were performed to compare evolutionary rate between passerines and nonpasserines. To assess whether the observed rate acceleration in passerines is specific to GH genes, we also analyzed 3 other genes as controls (ALDOB: Aldolase B, fructose-bisphosphate; CRYAA: Crystallin, Alpha A; and RHO: Rhodopsin; Kimball RT, Braun EL, unpublished data). These 3 genes were selected because they include codon sequences equal to or longer than our GH codon sequences as well as intron regions, and their predicted amino acid sequences include nonautapomorphic variation in ingroup taxa.
Ten passerines and 10 nonpasserines were used as ingroup taxa for the RRTs (supplementary table S1, Supplementary Material online and fig. 1). The ingroup taxa were selected because they include complete data for all the genes examined and represent the diversity of the passerines and their closest relatives. All of the ingroup taxa belong to the smallest well-supported clade that includes both passerines and nonpasserines (Hackett S, Kimball RT, Reddy S, Bowie RCK, Braun EL, Braun MJ, Chojnowski JL, Cox WA, Han K-L, Harshman J, Huddleston CJ, Marks BD, Miglia KJ, Moore WS, Sheldon FH, Steadman DW, Witt CC, Yuri T, unpublished data). To avoid possible anomalies associated with a particular outgroup, we used 3 outgroup taxa, Eudocimus albus (white ibis), Caprimulgus longirostris (band-winged nightjar), and Aramus guarauna (limpkin). These taxa come from widely separated lineages (fig. 1) representing the rest of Neoaves (all extant birds except Galliformes, Anseriformes, and Paleognathae).
|
We used the package HYPHY v1.00b (Kosakovsky Pond et al. 2005
; http://www.hyphy.org) to calculate the unconstrained maximum likelihood estimates (MLEs) of the 3-taxon trees (one passerine, one close relative of passerines, and one outgroup) and the MLEs of the 3-taxon trees with ingroup branch lengths constrained to be equal. RRTs were performed separately for amino acid, codon, and intron sequences, and nonsynonymous and synonymous rates in codons were tested independently. MLEs were obtained under the following models: for amino acids, the model using WAG matrix (Whelan and Goldman 2001) with among-site rate variation ( and I); for codons, the model of Goldman and Yang (1994) with nucleotide frequency tabulated separately for 1st, 2nd, and 3rd positions to estimate equilibrium codon frequencies; and for introns, the GTR model with among-site rate variation ( and I). A likelihood ratio test (LRT) was performed to determine whether the alternative hypothesis of unconstrained rate variation was significantly better than the null hypothesis that rates along 2 given branches are equal. The method of false discovery rate (FDR) of Benjamini and Yekutieli (2001) and Bonferroni procedure of Hochberg (1988) were used for multiple test correction using p.adjust function of the statistical package R (R Development Core Team 2007; http://www.R-project.org). The FDR is the expected proportion of true null hypotheses rejected out of the total number of null hypotheses rejected and is a less stringent correction method than Bonferroni corrections (Bonferroni 1936; Miller 1981), which are known to be highly conservative. The method of Benjamini and Yekutieli (2001) controls FDR consistently in multiple tests under dependency.
Analyses of Selection and Functional Divergence
The ratio of nonsynonymous to synonymous substitutions (dN/dS or 
) was used to estimate selective pressure at the protein level for the 2 passerine GH gene duplicates. Values of significantly greater than 1 indicate positive selection, whereas values significantly smaller than 1 indicate purifying selection. We performed an LRT that compares 2 models of selection, the null model M7 and the alternative hypothesis M8, using the program codeml in the PAML package v3.15 (Yang 1997). M7 assumes a β distribution of , in which codon sites are classified into 10 rate categories, each corresponding to a distinctive value within an interval 0 < 
1. The alternative model, M8, is constructed by adding an 11th rate category reflecting positive selection ( > 1) to M7. A rejection of M7 by LRT indicates that the coding region includes sites subject to positive selection. This program utilizes the codon-based evolutionary model of Goldman and Yang (1994) and explicitly takes into account the evolutionary relationships among the sequences. For these analyses, we used the GH gene tree estimated from the GARLI analysis described above. When positive selection was detected, the amino acid residues likely to be under positive selection were identified as those with high site-specific posterior probability of greater than 1 using naive empirical Bayes (Nielsen and Yang 1998; Yang et al. 2000) and Bayes empirical Bayes inferences (Yang et al. 2005).
We also estimated functional divergence between the 2 passerine GH gene duplicates by calculating the coefficient of functional divergence (
, a measure of replacement rate correlation over amino acid residues between gene duplicates) using the program DIVERGE v1.04 (Gu 1999). The program performs an LRT to test whether is significantly greater than zero, which would indicate that the replacement rate of the amino acid sequences differ significantly between the duplicates and thus suggest their functional divergence since the duplication event. When functional divergence was detected, the amino acid residues likely to be involved in functional divergence were identified using site-specific posterior probabilities of rate differences higher than baseline difference, with a cutoff value of 0.67 (Wang and Gu 2001).
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
We amplified via PCR fragments of about 1.3 kb from the GH genes of 162 avian taxa, including most major living lineages. In contrast to nonpasserines, the PCR products from many passerine species contained 2 strong bands in the 1.0- to 1.8-kb size range when visualized after electrophoresis. When the 2 bands from a passerine species were gel isolated and sequenced, they were found to contain nonidentical GH gene-like DNA sequences, differing at 19–25% of nucleotide sequence sites (p-distances for exons: 0.08–0.16, introns: 0.22–0.29). We designate the shorter copy "S" (sequence length: 1.0–1.3 kb) and the longer copy "L" (sequence length: 1.2–1.7 kb). In some cases, the 2 bands were not cleanly separable on agarose gels, so we cloned and sequenced 2–8 clones of the PCR products, allowing us to identify S and L copies by sequence similarity. In such cases, the sequence variation among clones within either the S or L class was low (0.0–2.3%) and attributable to either allelic differences or polymerase error. The lengths of homologous exons are the same for all of our passerine sequences; thus, any difference in the sequence length between S and L copies stems from length variation in introns.
Of 24 taxa chosen to represent the diversity of passerines (supplementary table S1, Supplementary Material online), we were able to recover both S and L copies of the GH gene for 21 taxa including Acanthisitta, the earliest branching taxon among extant passerines (Sibley and Ahlquist 1990; Barker et al. 2004). The 3 passerine taxa for which we have only one sequence are Climacteris (L), Grallaria (S), and Malurus (S). The PCR products of these taxa were cloned, and 4–8 clones of each product were sequenced to confirm that they were derived from a single copy of GH gene. Although we cannot exclude the possibility of gene loss, we believe that the negative results are due to PCR failure and these taxa probably have 2 copies of the GH gene for 2 reasons. First, because the 3 taxa are not close relatives (fig. 1) and both S and L copies are missing, a single loss cannot account for the missing sequences; at least 3 independent losses would be required. Second, typical PCR failure rates were around 5–10% of taxa tested for the "universal primers" used to amplify more than 20 nuclear genes in the Early Bird project. Thus, 3 PCR failures out of 48 attempts on passerine GH genes (6.3%) are not unusual. Only partial sequences were recovered for the L copy of Acanthisitta and Thamnophilus (approximately 50% and 80% of total length recovered, respectively). These partial sequences were used only for phylogenetic analyses.
Phylogenetic analyses using both GARLI and MrBayes yielded consistent relationships among passerine GH gene sequences. The estimated GH gene tree (fig. 1) has the following features: (1) all passerine sequences are clustered in a single clade, indicating that they are monophyletic, (2) there are 2 sister gene clades within passerines, corresponding to S and L gene copies, and (3) the topologies of the 2 gene clades are generally consistent with each other. Based on this analysis, we concluded that the GH gene was duplicated in the ancestral lineage of extant passerine birds.
We believe that both passerine GH gene paralogs (the 2 copies of GH gene in passerines are called paralogs hereafter) are functional for the following reasons: (1) both paralogs are preserved throughout the passerine clade, (2) the available exon sequences of all the paralogs in passerines contain conserved open-reading frames totaling 84 codons in length (83 or 84 codons in nonpasserines), (3) the predicted amino acid sequences are generally highly conserved, and (4) sequences corresponding to both S and L paralogs for zebra finch (T. guttata) are present in the Songbird EST project database, a database for gene sequences expressed in the zebra finch brain. Our phylogenetic analysis placed one zebra finch sequence in each of the S and L paralog clades, clustered with Old World finches as expected based on current taxonomy (fig. 1).
To examine the evolutionary rate of passerine GH gene paralogs, we first mapped predicted amino acid replacements on the GH gene tree (fig. 1). The resulting phylogram suggests an elevated rate of amino acid replacement in both paralogs (fig. 2). Then, we performed a series of pairwise RRTs to compare the rate of GH gene evolution in passerines and nonpasserines that are closely related to passerines. GH gene sequences from 10 passerines were compared with those from 10 nonpasserines, and virtually all (199 of 200) comparisons at the amino acid level revealed that both passerine GH gene paralogs have evolved more rapidly than any of the GH genes from nonpasserines (table 1). The average results of the RRTs suggest that paralog S evolved 
34-fold faster than nonpasserine GH gene homologs and that paralog L evolved 11-fold faster than nonpasserine homologs. The majority of these rate differences were significant when tested individually, although many comparisons lose significance after multiple test correction (table 1). This probably reflects the limited power associated with RRTs of the short amino acid sequences available. However, it is striking that there was no case in which a nonpasserine rate significantly exceeded the passerine rate.
|
|
The nucleotide sequences of passerine GH gene introns have also generally evolved faster than those of nonpasserines, although only by about 2-fold (table 1). The majority of these rate differences retain significance, even after multiple test correction, probably due to the longer intron sequences available for comparison. This suggests that there is a global acceleration of molecular evolution in passerines, which could be a genome-wide phenomenon possibly due to their small body size, high metabolic rate, and/or short generation time (Martin and Palumbi 1993
). This acceleration in passerines has been noted in a number of studies (e.g., van Tuinen et al. 2000). To determine whether the observed amino acid rate acceleration for GH genes simply reflects the general rate acceleration in passerine genomes rather than a GH gene-specific phenomenon, we performed a series of parallel RRTs for 3 additional nuclear genes (ALDOB, CRYAA, and RHO) as a control (table 1). Like the passerine GH gene paralogs, all 3 of these additional genes showed similar rate increases (1.5- to 1.8-fold) for intron nucleotide sequences. All 3 additional genes also show increases in amino acid rates. However, none of the amino acid rate increases were greater than 3-fold, and none retained significance after multiple test correction. Similar patterns were observed when exon evolutionary rates were compared at the codon level (supplementary table S2, Supplementary Material online). The rate of synonymous codon change in passerines was about 2-fold faster than in nonpasserine close relatives in both GH gene paralogs and the 3 control genes. In sharp contrast, the rate of nonsynonymous codon change was more than 10-fold faster for many of the passerine GH gene paralogs but only 2-fold faster for passerines in the 3 control genes. Thus, an approximately 2-fold greater rate of nucleotide sequence evolution appears to characterize many passerine nuclear genes and possibly represent a genome-wide effect, whereas the higher rate of amino acid sequence evolution appears to be GH gene specific. The RRT results were qualitatively unchanged when the outgroup used for table 1 (Eudocimus) was replaced with either of 2 other outgroups (Caprimulgus and Aramus), both of which are distantly related to Eudocimus. There is a quantitative difference when different outgroups are used; the rate acceleration for paralog L actually appears to be greater than that for paralog S when Caprimulgus is used as outgroup, whereas the apparent acceleration is greater for paralog S than for paralog L with either Eudocimus or Aramus used as outgroup (supplementary table S2, Supplementary Material online).
The ratio of nonsynonymous to synonymous substitutions (dN/dS or ) was used to estimate selective pressure at the protein level for the S and L paralogs. LRTs indicated positive selection on paralog S (2
l = 7.8, P < 0.006) but not on paralog L (2l = 0.0, P > 0.995). Naive empirical Bayes inference identified amino acid residues 50 and 58 (numbers correspond to the 84 residues predicted by our sequences) under positive selection ( > 1) in paralog S, with posterior probabilities greater than 0.99 (fig. 3a). However, Bayes empirical Bayes inference only weakly indicated positive selection on these 2 residues, with posterior probabilities of 0.77 and 0.80, respectively. In a simulation study, Wong et al. (2004) showed that the false-positive rates of naive empirical Bayes inference for M8 (with positive selection) versus M7 (without positive selection) model comparisons using the cutoff of 0.95 were below 5%. Because Bayes empirical Bayes inference is conservative, particularly for small data sets (Yang et al. 2005), we consider that these results provide good indications of positive selection for the 2 amino acid residues.
|
We also estimated functional divergence between the 2 passerine GH gene paralogs by calculating the coefficient of functional divergence (
) using amino acid replacement rates (Gu 1999). An LRT indicated that was significantly greater than zero (MLE of = 0.53 ± 0.24, 2l = 5.0, P < 0.03), suggesting functional divergence between the protein products of the 2 GH gene duplicates. The site-specific posterior analyses identified 6 amino acid residues that are likely to be responsible for this divergence (fig. 3b). All these residues show relatively marked differences in selective pressure (i.e., mean posterior ) between the S and L paralogs (fig. 3a). Two of these residues, 50 and 58, were the same residues identified as being under positive selection in paralog S by the aforementioned analysis of . Parsimony estimation of codon changes at sites 50 and 58, each of which has the same ancestral condition in all birds and in passerines specifically, indicates that 5 nonsynonymous changes occurred at each site in the clade of paralog S (fig. 4). In contrast, no nonsynonymous change occurred at either site of paralog L, and only 2 and 4 nonsynonymous changes occurred at the codon sites 50 and 58, respectively, in all 138 nonpasserine taxa. At codon site 50 in particular, 3 of the 5 nonsynonymous codon changes require multiple nucleotide substitutions. In addition, the changes are more concentrated in the oscine (songbird) clade. Although random mutations are expected to produce mainly nonsynonymous changes in these codons, the elevated rate of nonsynonymous changes in particular clades is not consistent with the pattern of random changes expected from the relaxation of purifying selection alone.
|
Based on the human GH protein structure (de Vos et al. 1992
), sites 50 and 58 are located on helix 2 and lie close to a randomly coiled region of the polypeptide between helices 3 and 4 (Supplementary fig. S1, Supplementary Material online). The estimated ancestral states of amino acids at sites 50 and 58 in passerines are methionine and valine, respectively, both of which are hydrophobic residues, and the observed replacement residues are also mainly hydrophobic (fig. 4). Hydrophobic amino acids in 
-helices are often thought to be of functional (mainly structural) significance (Dill 1990; O'Neil and Degrado 1990); therefore, changes in these amino acids may contribute to changes in the function of this polypeptide hormone. |
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Our results demonstrate that the GH gene was duplicated in a common ancestor of all extant passerine birds and that both paralogs have been maintained in most or all passerine lineages. Both paralogs are expressed in zebra finch brain, and both are likely to be functional based on maintenance of open-reading frames and generally conservative amino acid evolution. Comparative analyses indicate that both passerine paralogs have evolved more rapidly at the nucleotide and amino acid levels than the GH genes of nonpasserine relatives. Although the roughly 2-fold faster rate of synonymous codon or intron evolution may be a general phenomenon in passerines, related to their small body size, high metabolic rate, and/or short generation time (Martin and Palumbi 1993
), the 10-fold or greater rate of amino acid evolution is likely to be specific to the passerine GH gene paralogs. The paralogs appear to have functionally diverged, but only one of the paralogs shows evidence for positive selection.
Buggiotti and Primmer (2006) pointed out that, of the 6 avian taxa they studied, the most divergent GH amino acid sequence was that of a passerine bird, European pied flycatcher (Ficedula hypoleuca), which differed from the other avian GH polypeptides by 18–27 amino acids, whereas divergence among the other 5 taxa ranged from 2 to 22 amino acids. This level of amino acid sequence divergence is comparable to that found between the green sea turtle (Chelonia mydas) and avian GH polypeptides (23–29 amino acid divergence). Their report on the apparently accelerated rate of GH amino acid evolution in pied flycatcher is consistent with our finding, although they included only GH gene paralog S for the single passerine examined.
Because newly duplicated genes are functionally redundant, selective constraints on the duplicated genes are likely to become relaxed, allowing some mutational variation to be sustained. This variation, in turn, may allow molecular evolution to proceed more rapidly than in single-copy homologs. These duplicated genes are expected to have 1 of 3 possible fates (Ohno 1970; Lynch and Conery 2000; Zhang 2003; Hurles 2004, Sassi et al. 2007): (1) one of the duplicates becomes a pseudogene due to degenerative mutations (nonfunctionalization), (2) one of the duplicates gains a new function due to a new, advantageous mutation (neofunctionalization), and (3) the original functions of the single-copy gene may be partitioned between the duplicates (subfunctionalization). The observed patterns of evolution in the passerine GH genes are unlikely to reflect nonfunctionalization in which only one copy is expected to exhibit an increased rate of evolution, with a value of approaching (but not exceeding) unity (Sassi et al. 2007). Mutations that interrupt the reading frame are also expected after some time, and neither prediction has been met for passerine GH genes.
The majority of gene duplications appear to be preserved by subfunctionalization (Lynch and Force 2000), a process that may begin with differences in gene expression reflecting small changes in regulatory regions of the duplicated genes (Force et al. 1999). As many genes perform a multiplicity of subtly distinct functions, selective pressures may have resulted in a compromise between optimal sequences for each role. Once the functions of the duplicates begin to diverge, amino acid changes related to functional specialization of each duplicate are likely to be adaptive, and both duplicates will evolve rapidly until subfunctionalization is complete (Hughes 1994). Therefore, subfunctionalization can explain the rapid amino acid evolution often reported in both gene duplicates after a gene duplication event (Wallis 1996). Because subfunctionalization is more common than neofunctionalization and consistent with the evidence for accelerated amino acid evolution in both passerine GH gene paralogs, we believe that it is the more likely explanation for the preservation of both paralogs. However, we cannot rigorously exclude neofunctionalization as an alternative explanation.
The explosive radiation of passerines has intrigued many avian biologists and systematists for more than a century (e.g., Müller 1878; Ames 1971; Raikow 1982; Edwards et al. 1991; Nee et al. 1992; Barker et al. 2004). However, there are only a few obvious "key innovations" recognized in this group, and some systematists have questioned whether Passeriformes includes an arbitrarily large number of species (e.g., Raikow 1986; Raikow and Bledsoe 2000). Could the GH gene duplication reported here have played a significant role in the passerine radiation? The maintenance of 2 GH gene copies since some time before the separation of New Zealand wrens (Acanthisittidae) from other passerines, between 55 and 100 million years ago (Boles 1995, Ericson et al. 2002, Barker et al. 2004, Pereira and Baker 2006), indicates that the second copy must be functional. Because of the importance of GH to development and the accelerated development observed in passerines relative to many other groups of birds (Ricklefs 1979; Ricklefs and Starck 1998), we speculate that this duplication may be of adaptive significance. Future work on the functions of duplicated GH genes in passerines may yield insight into the evolutionary success of this most speciose group of birds.
|
Supplementary Material |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Supplementary tables S1 and S2 and figure S1 are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).
|
Acknowledgements |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
For tissue samples, we thank the following institutions (in alphabetical order) and collectors (supplementary table S1, Supplementary Material online): American Museum of Natural History, Australian National Wildlife Collection, Burke Museum of Natural History and Culture (University of Washington), Field Museum of Natural History, University of Kansas Natural History Museum & Biodiversity Center, Louisiana State University Museum of Natural Science, Marjorie Barrick Museum (University of Nevada, Las Vegas), Museum of Southwestern Biology (University of New Mexico), Museum of Vertebrate Zoology (University of California, Berkeley), Museum Victoria, National Museum of Natural History, San Francisco Zoological Garden, and Zoological Museum University of Copenhagen. We are grateful to two anonymous reviewers for helpful comments and to our collaborators in the Early Bird project (in alphabetical order), Rauri Bowie, Jena Chojnowski, Shannon Hackett, Kin-Lan Han, John Harshman, Chris Huddleston, Ben Marks, Kathy Miglia, Bill Moore, Sushma Reddy, Fred Sheldon, Dave Steadman, and Chris Witt for providing unpublished trees and valuable insights and comments. This work was supported by the National Science Foundation "Assembling the Tree of Life" program (DEB-0228617, DEB-0228675, DEB-0228682, and DEB-0228688).
|
Footnotes |
|---|
1 Present address: Department of Zoology, University of Florida Gainesville, FL.
Scott Edwards, Associate Editor
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Agellon LB, Davies SL, Lin C-M, Chen TT, Powers DA. Rainbow trout has two genes for growth hormone. Mol Reprod Dev (1988) 1:11–17.[CrossRef][Medline]
Ames PL. The morphology of the syrinx in passerine birds. Bull Peabody Mus Nat Hist (1971) 37:1–94.
Aramburo C, Luna M, Carranza M, Reyes M, Martinez-Coria H, Scanes CG. Growth hormone size variants: changes in the pituitary during development of the chicken. Proc Soc Exp Biol Med (2000) 223:67–74.
Barker FK, Cibois A, Schikler P, Feinstein J, Cracraft J. Phylogeny and diversification of the largest avian radiation. Proc Natl Acad Sci USA (2004) 101:11040–11045.
Benjamini Y, Yekutieli D. The control of the false discovery rate in multiple testing under dependency. Ann Stat (2001) 29:1165–1188.[CrossRef]
Boles WE. The world's oldest songbird (Aves: Passeriformes). Nature (1995) 374:21–22.
Bonferroni CE. Teoria statistica delle classi e calcolo delle probabilità. Pubbl R Ist Super Sci Econ Commer Fir (1936) 8:3–62.
Buggiotti L, Primmer CR. Molecular evolution of the avian growth hormone gene and comparison with its mammalian counterpart. J Evol Biol (2006) 19:844–854.[CrossRef][Web of Science][Medline]
Chen EY, Liao YC, Smith DH, Barrera-Saldana HA, Gelinas RE, Seeburg PH. The human growth hormone locus: nucleotide sequence, biology, and evolution. Genomics (1989) 4:479–497.[CrossRef][Web of Science][Medline]
de Vos AM, Ultsch M, Kossiakoff AA. Human growth hormone and extracellular domain of its receptor: crystal structure of the complex. Science (1992) 255:306–312.
Devlin RH. Sequence of sockeye salmon type 1 and type 2 growth hormone genes and the relationship of rainbow trout with Atlantic and Pacific salmon. Can J Fish Aquat Sci (1993) 50:1738–1748.[CrossRef]
Dill KA. Dominant forces in protein folding. Biochemistry (1990) 29:7133–7155.[CrossRef][Web of Science][Medline]
Edwards SV, Arctander P, Wilson AC. Mitochondrial resolution of a deep branch in the genealogical tree for perching birds. Proc R Soc Lond B Biol Sci (1991) 243:99–107.[Medline]
Ericson PGP, Christidis L, Cooper A, Irestedt M, Jackson J, Johansson US, Norman JA. A Gondwanan origin of passerine birds supported by DNA sequences of the endemic New Zealand wrens. Proc R Soc Lond B Biol Sci (2002) 269:235–241.
Etherton TD, Bauman DE. Biology of somatotropin in growth and lactation of domestic animals. Physiol Rev (1998) 78:745–761.
Force A, Lynch M, Pickett FB, Amores A, Yan YL, Postlethwait J. Preservation of duplicate genes by complementary, degenerative mutations. Genetics (1999) 151:1531–1545.
Gelman A, Rubin DB. Inference from iterative simulation using multiple sequences. Stat Sci (1992) 7:457–511.[CrossRef]
Goldman N, Yang Z. A codon-based model of nucleotide substitution for protein-coding DNA sequences. Mol Biol Evol (1994) 11:725–736.[Abstract]
Gu X. Statistical methods for testing functional divergence after gene duplication. Mol Biol Evol (1999) 16:1664–1674.[Abstract]
Harvey S, Hull KL. Growth hormone: a paracrine growth factor? Endocrine (1997) 7:267–279.[Web of Science][Medline]
Hochberg Y. A sharper Bonferroni procedure for multiple tests of significance. Biometrika (1988) 75:800–802.
Huang HC, Brown DD. Overexpression of Xenopus laevis growth hormone stimulates growth of tadpoles and frogs. Proc Natl Acad Sci USA (2000) 97:190–194.
Huelsenbeck JP, Ronquist F. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics (2001) 17:754–755.
Hughes AL. The evolution of functionally novel proteins after gene duplication. Proc Biol Sci (1994) 256:119–124.
Hurles M. Gene duplication: the genomic trade in spare parts. PLoS Biol (2004) 2:900–904.[Web of Science]
International Chicken Genome Sequencing Consortium. Sequence and comparative analysis of the chicken genome provide unique perspectives on vertebrate evolution. Nature (2004) 432:695–716.[CrossRef][Medline]
Ip CYS, Zhang X, Leung FC. Genomic growth hormone gene polymorphisms in native Chinese chickens. Exp Biol Med (2001) 226:458–462.
Kawauchi H, Suzuki K, Yamazaki T, Moriyama S, Nozaki M, Yamaguchi K, Takahashi A, Youson J, Sower SA. Identification of growth hormone in the sea lamprey, an extant representative of a group of the most ancient vertebrates. Endocrinology (2002) 143:4916–4921.
Kosakovsky Pond SL, Frost SDW, Muse SV. HyPhy: hypothesis testing using phylogenies. Bioinformatics (2005) 21:676–679.
Lewis PO. A genetic algorithm for maximum-likelihood phylogeny inference using nucleotide sequence data. Mol Biol Evol (1998) 15:277–283.[Abstract]
Liu JC, Makova KD, Adkins RM, Gibson S, Li WH. Episodic evolution of growth hormone in primates and emergence of the species specificity of human growth hormone receptor. Mol Biol Evol (2001) 18:945–953.
Lynch M, Conery JS. The evolutionary fate and consequences of duplicate genes. Science (2000) 290:1151–1155.
Lynch M, Force A. The probability of duplicate gene preservation by subfunctionalization. Genetics (2000) 154:459–473.
Maniou Z, Wallis OC, Wallis M. Episodic molecular evolution of pituitary growth hormone in Cetartiodactyla. J Mol Evol (2004) 58:743–753.[CrossRef][Web of Science][Medline]
Martin AP, Palumbi SR. Body size, metabolic rate, generation time, and the molecular clock. Proc Natl Acad Sci USA (1993) 90:4087–4091.
McKay SJ, Trautner J, Smith MJ, Koop BF, Devlin RH. Evolution of duplicated growth hormone genes in autotetraploid salmonid fishes. Genome (2004) 47:714–723.[Medline]
Miller RG. Simultaneous statistical inference (1981) 2nd ed. New York: Springer-Verlag.
Montgelard C, Catzeflis FM, Douzery E. Phylogenetic relationships of artiodactyls and cetaceans as deduced from the comparison of cytochrome b and 12S rRNA mitochondrial sequences. Mol Biol Evol (1997) 14:550–559.[Abstract]
Müller JP. On certain variations in the vocal organs of the Passeres that have hitherto escaped notice (1878) London: Macmillan.
Nee S, Mooers AØ, Harvey PH. Tempo and mode of evolution revealed from molecular phylogenies. Proc Natl Acad Sci USA (1992) 89:8322–8326.
Nielsen R, Yang Z. Likelihood models for detecting positively selected amino acid sites and applications to the HIV-1 envelope gene. Genetics (1998) 148:929–936.
Ohno S. Evolution by gene duplication (1970) New York: Springer.
Ohta T. Pattern of nucleotide substitutions in growth hormone-prolactin gene family: a paradigm for evolution by gene duplication. Genetics (1993) 134:1271–1276.[Abstract]
O'Neil KT, Degrado WF. A thermodynamic scale for the helix-forming tendencies of the commonly occurring amino acids. Science (1990) 250:646–651.
Pereira SL, Baker AJ. A mitogenomic timescale for birds detects variable phylogenetic rates of molecular evolution and refutes the standard molecular clock. Mol Biol Evol (2006) 23:1731–1740.
Raikow RJ. Monophyly of the Passeriformes: test of a phylogenetic hypothesis. Auk (1982) 99:431–455.[Web of Science]
Raikow RJ. Why are there so many kinds of passerine birds? Syst Zool (1986) 35:255–259.
Raikow RJ, Bledsoe AH. Phylogeny and evolution of the passerine birds. Bioscience (2000) 50:487–499.[CrossRef][Web of Science]
R Development Core Team. R: a language and environment for statistical computing (2007) Vienna (Austria): R Foundation for Statistical Computing.
Rentier-Delrue F, Swennen D, Mercier L, Lion M, Benrubi O, Martial JA. Molecular cloning and characterization of two forms of trout growth hormone cDNA: expression and secretion of tGH-II by Escherichia coli. DNA (NY) (1989) 8:109–117.
Ricklefs RE. Patterns of growth in birds. V. A comparative study of development in the starling, common tern, and Japanese quail. Auk (1979) 96:10–30.[Web of Science]
Ricklefs RE, Starck JM. Embryonic growth and development. In: Avian growth and development—Starck JM, Ricklefs RE, eds. (1998) New York: Oxford University Press. 31–58.
Ronquist F, Huelsenbeck JP. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics (2003) 19:1572–1574.
Sanders EJ, Harvey S. Growth hormone as an early embryonic growth and differentiation factor. Anat Embryol (2004) 209:1–9.[CrossRef][Medline]
Sassi SO, Braun EL, Benner SA. The evolution of seminal ribonuclease: pseudogene reactivation or multiple gene inactivation events? Mol Biol Evol (2007) 24:1012–1024.
Sibley CG, Ahlquist JE. Phylogeny and classification of birds (1990) New Haven (CT): Yale University Press.
van Tuinen M, Sibley CG, Hedges SB. The early history of modern birds inferred from DNA sequences of nuclear and mitochondrial ribosomal genes. Mol Biol Evol (2000) 17:451–457.
Wallis M. The molecular evolution of pituitary growth hormone, prolactin and placental lactogen: a protein family showing variable rates of evolution. J Mol Evol (1981) 17:10–17.[CrossRef][Web of Science]
Wallis M. Variable evolutionary rates in the molecular evolution of mammalian growth hormones. J Mol Evol (1994) 38:619–627.[Web of Science][Medline]
Wallis M. The molecular evolution of vertebrate growth hormones: a pattern of near-stasis interrupted by sustained bursts of rapid change. J Mol Evol (1996) 43:93–100.[CrossRef][Web of Science][Medline]
Wallis M, Lioupis A, Wallis OC. Duplicate growth hormone genes in sheep and goat—commentary. J Mol Endocrinol (1998) 21:1–5.[CrossRef][Web of Science][Medline]
Wallis OC, Wallis M. Molecular evolution of growth hormone (GH) in Cetartiodactyla: cloning and characterization of the gene encoding GH from a primitive ruminant, the chevrotain (Tragulus javanicus). Gen Comp Endocrinol (2001) 123:62–72.[CrossRef][Web of Science][Medline]
Wallis OC, Wallis M. Characterisation of the GH gene cluster in a new-world monkey, the marmoset (Callithrix jacchus). J Mol Endocrinol (2002) 29:89–97.[Abstract]
Wang YF, Gu X. Functional divergence in the caspase gene family and altered functional constraints: statistical analysis and prediction. Genetics (2001) 158:1311–1320.
Waters MJ, Shang CA, Behncken SN, Tam SP, Li H, Shen B, Lobie PE. Growth hormone as a cytokine. Clin Exp Pharmacol Physiol (1999) 26:760–764.[CrossRef][Web of Science][Medline]
Whelan S, Goldman N. A general empirical model of protein evolution derived from multiple protein families using a maximum-likelihood approach. Mol Biol Evol (2001) 18:691–699.
Wong WSW, Yang Z, Goldman N, Nielsen R. Accuracy and power of statistical methods for detecting adaptive evolution in protein coding sequences and for identifying positively selected sites. Genetics (2004) 168:1041–1051.
Yang Z. PAML: a program package for phylogenetic analysis by maximum likelihood. Comput Appl Biosci (1997) 13:555–556.
Yang Z, Nielsen R, Goldman N, Pedersen AM. Codon substitution models for heterogeneous selection pressure at amino acid sites. Genetics (2000) 155:431–449.
Yang Z, Wong WSW, Nielsen R. Bayes empirical Bayes inference of amino acid sites under positive selection. Mol Biol Evol (2005) 22:1107–1118.
Zhang JZ. Evolution by gene duplication: an update. Trends Ecol Evol (2003) 18:292–298.[CrossRef]
Zhao RQ, Muehlbauer E, Decuypere E, Grossmann R. Effect of genotype-nutrition interaction on growth and somatotropic gene expression in the chicken. Gen Comp Endocrinol (2004) 136:2–11.[CrossRef][Web of Science][Medline]
Zwickl DJ. Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. Ph.D. dissertation, (2006) University of Texas at Austin.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. Harshman, E. L. Braun, M. J. Braun, C. J. Huddleston, R. C. K. Bowie, J. L. Chojnowski, S. J. Hackett, K.-L. Han, R. T. Kimball, B. D. Marks, et al. Phylogenomic evidence for multiple losses of flight in ratite birds PNAS, September 9, 2008; 105(36): 13462 - 13467. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||