Molecular Biology and Evolution

MBE Advance Access originally published online on September 24, 2007
Molecular Biology and Evolution 2007 24(12):2698-2706; doi:10.1093/molbev/msm208

This Article Abstract FREE Full Text (PDF) Supplementary Material All Versions of this Article: 24/12/2698    most recent msm208v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Request Permissions Google Scholar Articles by Mank, J. E. Articles by Ellegren, H. Search for Related Content PubMed PubMed Citation Articles by Mank, J. E. Articles by Ellegren, H. Social Bookmarking          What's this?

© The Author 2007. Published by Oxford University Press on behalf of the Society for Molecular Biology and Evolution. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org

Research Articles

Rapid Evolution of Female-Biased, but Not Male-Biased, Genes Expressed in the Avian Brain

Judith E. Mank, Lina Hultin-Rosenberg, Erik Axelsson and Hans Ellegren

Department of Evolutionary Biology, Evolutionary Biology Centre, Uppsala University, Uppsala, Sweden

E-mail: Hans.Ellegren{at}ebc.uu.se.

	Abstract

TOP Abstract Introduction Materials and Methods Results Discussion Supplementary Material Acknowledgements References

 
The powerful pressures of sexual and natural selection associated with species recognition and reproduction are thought to manifest in a faster rate of evolution in sex-biased genes, an effect that has been documented particularly for male-biased genes expressed in the reproductive tract. However, little is known about the rate of evolution for genes involved in sexually dimorphic behaviors, which often form the neurological basis of intrasexual competition and mate choice. We used microarray data, designed to uncover sex-biased expression patterns in embryonic chicken brain, in conjunction with data on the rate of sequence evolution for >4,000 coding regions aligned between chicken and zebra finch in order to study the role of selection in governing the molecular evolution for sex-biased and unbiased genes. Surprisingly, we found that female-biased genes, defined across a range of cutoff values, show a higher rate of functional evolution than both male-biased and unbiased genes. Autosomal male-biased genes evolve at a similar rate as unbiased genes. Sex-specific genomic properties, such as heterogeneity in genomic distribution and GC content, and codon usage bias for sex-biased classes fail to explain this surprising result, suggesting that selective pressures may be acting differently on the male and female brain.

Key Words: gene expression • sexual selection • sex bias • d_N/d_S • chicken

	Introduction

TOP Abstract Introduction Materials and Methods Results Discussion Supplementary Material Acknowledgements References

 
Many of the numerous behavioral, physiological, and anatomical differences between males and females originate from autosomal genes that are identical in both sexes. In order to produce distinct female and male phenotypes from the same gene complement, many coding regions are expressed at different levels in the sexes, a phenomenon referred to as sex-biased gene expression (Ellegren and Parsch 2007). Sex-biased expression is common, and microarray gene expression analyses have indicated that a large proportion of the transcriptome shows different expression between females and males in several metazoan species (Ranz et al. 2003; Parisi et al. 2004; Marinotti et al. 2006; Yang et al. 2006). Sex-biased genes are likely responsible for a great deal of intersexual variation and are often associated with various aspects of reproduction (Rinn and Snyder 2005); therefore, these genes are frequently subject to the powerful selective pressures of sexual selection, mate choice, species recognition, and sperm competition (Swanson and Vacquier 2002).

Reproductive pressures could theoretically increase the rate of sequence and protein evolution for both male- and female-biased genes, a condition that has been demonstrated primarily for coding regions associated with the reproductive tract (Civetta and Singh 1999; Swanson et al. 2001, 2004; Zhang et al. 2004; Jagadeeshan and Singh 2005). However, sex-biased genes are not limited to the reproductive tract and are present in somatic tissues as well (Ranz et al. 2003; Parisi et al. 2004; Yang et al. 2006), giving rise to behavioral, physiological, and anatomical sexual dimorphism. Studies of sex-biased somatic genes have suggested that primarily male-biased genes undergo faster rates of functional evolution (Meiklejohn et al. 2003; Pröschel et al. 2005); however, these studies have been primarily limited to Drosophila melanogaster, which may not be an appropriate generalized model for sex-biased gene evolution (Metta et al. 2006). Sex-biased genes expression in brain are the probable cause of many, if not most of the behavioral differences between males and females (Dauwalder et al. 2002; Drapeau et al. 2003; Kadener et al. 2006), and sex-biased brain genes are therefore important for various aspects of premating sexual selection, including courtship and breeding behavior, as well as mate choice and species recognition (Andersson 1994). This situation warrants an examination of the relationship between sex-biased gene expression in brain and rate of molecular evolution in organisms in which sexual selection is pronounced. Many avian species show elaborate morphological and behavioral dimorphisms, many of which have been linked to sexual selection, making birds an ideal clade for such an investigation.

It is possible to estimate the rate of sequence evolution for different classes of genes by comparing the ratio of nonsynonymous (or coding) substitutions to synonymous (or silent) substitutions in the alignment of orthologous coding regions for 2 or more species. This ratio corrects for neutral mutation rate variation and gives an indicator of how much functional change a given gene has undergone since those species diverged. We used this approach to compare the divergence of sex-biased and unbiased brain-expressed genes in birds in order uncover the pattern of molecular evolutionary properties of the coding regions that underlie behavioral dimorphisms. We utilized microarray expression data from chicken (Gallus gallus) brain tissue, designed to reveal differences in expression between male and female embryos (Ellegren et al. 2007). We then used the annotated chicken genome sequence information (ICGSC 2004) for roughly 4,000 identified brain-expressed genes, in conjunction with orthologous zebra finch (Taeniopygia guttata) brain expressed sequence tag (EST) data from the ‘Songbird Neurogenomics Initiative’ (Clayton 2004) to estimate the rate of sequence evolution that has occurred since the chicken and zebra finch diverged, roughly 100 MYA (van Tuinen et al. 2000). We also examined other genomic properties, such as chromosome location, Gene Ontology terminology, GC content, and bias in codon usage that have been linked to rate of sequence evolution in birds and other organisms (ICGSC 2004; Meunier and Duret 2004; Zhang et al. 2004; Axelsson et al. 2005; Hambuch and Parsch 2005; Spencer et al. 2006).

	Materials and Methods

TOP Abstract Introduction Materials and Methods Results Discussion Supplementary Material Acknowledgements References

 
Expression Data Collection
RNA was collected from the brain tissue of 4 female and 3 male 18-day chicken embryos, and each sample was analyzed individually using the Chicken Affymetrix GeneChip microarray platform. Further details regarding sample preparation and methodology are found in Ellegren et al. (2007) and are also available in the supplementary material online. All preprocessing and statistical analysis of microarray data was performed in R version 2.4.1 using Bioconductor packages release 1.9 (Ihkaka and Gentleman 1996). CEL files were processed with GC Robust Multi-array Average (Wu and Irizarry 2004), a background adjustment method that corrects for the GC content of probes when assessing nonspecific binding, followed by quantile normalization and median polish summarization of probe intensities into probe set intensities. A linear model was then fitted to the log₂ of the expression levels based on all probe sets using the Limma package (Smyth 2004). After preprocessing and linear model fitting, the probe sets were filtered on expression; an expression threshold was set on both average expression level and absent/present calls from the R implementation of the Affymetrix MAS 5.0 algorithm. Only probe sets with average expression over a defined threshold and present in more than half of the samples within each sex were considered as significantly expressed. Average relative fluorescence was calculated among replicates of each sex, and sex-bias estimates were calculated as log₂ (average male expression) – log₂ (average female expression). Data were then averaged across replicate probe sets that corresponded to the same coding region. These data are available at the NCBI Gene Expression Omnibus repository under the accession number GSE8693.

For brain-expressed coding regions, we identified both male-biased and female-biased genes according to a range of fold-change cutoff criteria. Previous studies of reproductive tract genes have used a 2-fold cutoff to define sex-biased genes (Zhang et al. 2004; Zhang and Parsch 2005); however, the testis and the ovary are the most sexually dimorphic of all tissues and would therefore have much higher levels of sex-biased gene expression than other, less dimorphic tissues (such as brain). We therefore used lower thresholds to delineate sex-biased coding regions. Three categories were used based on expression in one sex compared with the other: genes were classified as sex biased if they were upregulated by at least 125%, 133%, or 150% in one sex compared with the other, with a t-test–adjusted P value corresponding to a 10% false discovery rate. Genes that did not meet these fold-change cutoffs were classified as unbiased; genes with sex-biased expression over the defined threshold but with insufficient adjusted P values were removed from the analysis at appropriate fold-change cutoffs. This accounts for the slight variation in total coding regions at different fold-change cutoffs. Results are unchanged if these ambiguous genes are included in the unbiased expression category (supplementary tables S1 and S2, Supplementary Material online); however, sex-biased genes are inherently more variable in expression than unbiased genes, which directly effects significance statistics. This implies that it may not be fair to assume that genes exceeding fold-change cutoffs but lacking significant P values are unbiased.

Alignment and Estimation Gene-Specific Substitutions
Detailed methodology for the estimation of d_N and d_S estimates for the coding regions in this study has been previously described (Mank et al. 2007). Briefly, for all significantly expressed genes, we searched for corresponding zebra finch (T. guttata) EST contig data generated by the Songbird Neurogenomics Initiative, available at http://titan.biotec.uiuc.edu/songbird/ (Clayton 2004). The transcripts, all derived from multiple cDNA libraries made from telencephalon of embryonic, juvenile, and adult birds of both sexes, were aligned with their chicken orthologues, as defined by BlastN search to all known and ab initio predicted protein-coding chicken genes identified in Ensembl (www.ensembl.org/Gallus_gallus/index.html). Orthology was established using the principle of best reciprocal hit, given a minimum e value of 10^–30 for a match. Because e scores are influenced by sequence length, we excluded all alignments less than 100 bp from further analysis in order to reduce the possibility of false homology between short alignments.

DNA sequences were subsequently translated into amino acid sequences and aligned using DIALIGN2 (Morgenstern 1999), and the CODEML PAML package, version 3.15 (Yang 1997) was used to estimate the nonsynonymous (d_N, defined as the number of nonsynonymous substitutions per nonsynonymous site) and synonymous (d_S, the number of synonymous substitutions per synonymous site) divergence for each set of orthologues. The maximum likelihood estimation was performed separately for each alignment, which allows for variation in codon usage, a factor implicated in Drosophila sex-biased gene expression (Hambuch and Parsch 2005). For all analyzed coding regions, we inferred the chromosomal location from the May 2006 Ensembl galGal3 assembly of the chicken genome.

We excluded 41 coding regions from our data set with d_S > 3 as divergence estimates are not reliable for saturated sites. The expression pattern of these 41 genes (3 female biased, 5 male biased, and 33 unbiased) did not significantly differ from the expectation given the overall expression pattern (² = 2.45, P = 0.29, 2 degrees of freedom [df]); therefore, we have no reason to suspect that the removal of these genes causes a bias in our analysis. Additionally, all transcripts with premature stop codons were removed as these either represent pseudogenes or errors introduced by reverse transcription. Finally, we screened all zebra finch sequences for high complexity contamination using the Gallus-derived repeat library in RepeatMasker (Smith et al. 2002; http://www.repeatmasker.org) As the Ensembl gene annotation methodology automatically removes repetitive DNA elements via RepeatMasker (Curwen et al. 2004), it was not necessary to screen the orthologous chicken coding regions for retroviral contamination. No contamination in the zebra finch sequences was identified.

Estimation of Rates of Molecular Evolution for Sex-Biased Genes
For the entire genomic complement of genes in our data set, mean categorical values of d_N and d_S for biased and unbiased categories, as well as the genomic average, were calculated by dividing the sum of the number of substitutions over genes by the sum of the number of sites over genes. This avoids problem of infinitely high d_N/d_S values arising from genes with no synonymous substitutions and has the added advantage that data for individual genes are weighted by alignment length. We assessed the 95% confidence intervals (CIs) for d_N/d_S ratio estimates with bootstrapping (1,000 repetitions). We also compared the various sex-biased categories to unbiased genes for all brain-expressed categories with a permutations test (1,000 repetitions).

Regardless of expression bias, sex-linked genes in birds have been shown to evolve more rapidly (Mank et al. 2007), a likely consequence of the so-called Fast-Z evolution resulting from the hemizygous exposure in the heterogametic sex (Charlesworth 1991). They also undergo higher levels of mutation than autosomal genes (Axelsson et al. 2004). Additionally, the lack of complete dosage compensation in chicken and other birds means that the majority of genes on the Z chromosome are expressed at higher levels in males (ZZ) than in females (ZW) (Ellegren et al. 2007; Itoh et al. 2007). Together, these observations suggest that an increased d_N/d_S ratio for male-biased genes on the Z chromosome may not be due to sex-biased expression, but rather Z linkage. We therefore removed Z-linked genes from the data set and compared the d_N, d_S, and d_N/d_S estimates for different sex-biased expression categories on the autosomes as described above for the full genomic data set. This data set was used for all subsequent analyses unless explicitly noted.

Tissue of Maximal Expression
Many genes are expressed in multiple tissues and organs throughout the body (Waxman and Peck 1998; Ericson et al. 2006). Microarray studies reveal whether a coding region is expressed in a given tissue, but they do not generally reveal the relative expression of that coding region throughout the organism. This means that a brain microarray data set such as the one we employ here is a blend of brain-specific genes, genes that exhibit maximal expression levels in the brain but are expressed at lower levels in other tissues, genes that are expressed in many tissues at similar levels, and genes that are expressed at low levels in the brain but achieve maximal expression in other organs. Although expression in other tissues does not mean that a given coding region is not important in brain function and development, a hodgepodge of expression characteristics can obscure evidence of sequence evolution as genes with maximal expression in tissues other than brain may be responding to selection for extraneurological functions.

We therefore identified genes that are confined to, or primarily expressed in, the central nervous system using data from the UniGene EST assemblage (ftp://ncbi.nlm.nih.gov/repository/unigene/). This database is a repository of EST libraries from a variety of organisms across a wide array of tissues. We selected EST libraries from chicken that represented a single defined tissue or organ, and the resulting data set comprised nonnormalized chicken EST libraries for blood, cartilage, central nervous system, digestive tract, epiphyseal growth plate, eye, genitourinary tract, heart, limb, and the lymphoreticular system. Data for each coding region were standardized to number of transcripts per million (TPM) for each EST library. This metric is a measure of the number of transcripts detected for a given gene per million transcripts analyzed from tissue i and is defined as

N is the number of tissues examined, and TPM_max is the highest expression level detected for a given gene over all tissues examined. We then determined which tissue possessed the highest TPM for each coding region in our study, detailed in supplementary table S3 (Supplementary Material online) and identified those coding regions where expression was highest in, or confined to, the central nervous system. We assigned the resulting 313 coding regions into sex-biased expression categories, using only the 125% cutoffs and adjusted P values in order to maximize sex-biased sample sizes for statistical power and compared d_N, d_S, and d_N/d_S for each category with permutation tests (1,000 repetitions).

Expression Level
We calculated the relationship between expression level and rate of evolution for all the genes in the 125% expression bias class. We assessed the correlation with linear regression between d_N/d_S and the log₂ average expression, based on normalized relative fluorescence of the microarray data, for each sex separately. We also calculated the average the expression level for unbiased, male-biased, and female-biased genes and determined the 95% CIs based on bootstrapping (1,000 repetitions).

Gene Ontology
In order to investigate whether different functional classes of genes could be influencing our results, we tested for overrepresentation of Gene Ontology (GO) terms (Gene Ontology Consortium 2000) using ONTOLOGIZER 2.0 (Robinson et al. 2004). We first tested whether there was an overabundance of GO terms for Z-linked genes using all mapped coding regions as the population. We then tested for overrepresentation of GO terms in both autosomal female- and male-biased expression classes (at the 125% fold-change level) using the full complement of autosomal genes as the population set. For all these analyses, we used term-for-term comparisons and implemented the Bonferroni correction for multiple comparisons.

Heterogeneity Estimation
Chickens have a wide range of chromosome sizes, from the large macrochromosomes (chromosomes 1–5), the intermediate chromosomes (chromosomes 6–10), and the small microchromosomes (chromosomes 11–38) (ICGSC 2004). These chromosomal classes have distinctive properties, with microchromosomes generally being more gene dense, having higher rates of recombination and higher GC content than the other chromosomal classes (ICGSC 2004). These properties have been shown to influence the rate of sequence evolution in birds (ICGSC 2004; Axelsson et al. 2005); therefore, in order to investigate whether genomic distribution influences our observed rates of evolution, we compared the distribution of sex-biased genes to the distribution of all brain-expressed genes for macrochromosomes and microchromosomes with a ² test (1 df).

GC and third position GC (GC3) content has been shown to influence diversity and substitution (Williams and Hurst 2000; Smith et al. 2002). Therefore, we calculated GC and CG3 content for sex-biased classes, based on Ensembl annotation of the chicken genome sequence, and calculated the statistical significance of the difference among sex-bias classes using the 125% cutoff only in order to maximize statistical power through large sex-biased sample size, with permutation testing (1,000 repetitions). We also assessed the role of GC and GC3 on d_N/d_S among 125% sex-biased expression categories using both linear regression and analysis of covariance (ANCOVA) (GC and GC3 as independent variables, 2 df).

Male-biased genes in flies have been shown to exhibit less codon bias than other expression categories (Hambuch and Parsch 2005), which, if such a phenomenon were present in birds as well, could explain the high d_S values in male-biased brain genes. We therefore calculated the effective number of codons (ENC) for each gene (Wright 1990) based on the Ensembl annotation of the chicken genome. ENC can take values from 20, in the case of extreme bias where only one codon is used for each amino acid, to 61, where all alternative codons are equally likely. We then tested ENC among expression categories for statistical difference at the 125% threshold using analysis of variance (ANOVA) (2 df).

	Results

TOP Abstract Introduction Materials and Methods Results Discussion Supplementary Material Acknowledgements References

 
The microarray experiments yielded gene expression data for 8,612 Ensembl-annotated chicken coding regions. We found orthologous sequence data in zebra finch for 4,251 of those genes, though 96 genes were removed due to premature stop codons, 37 had total aligned sequences of >100 bp and were removed, and a further 41 resulted in a d_S > 3 and were removed due to potential problems with saturation. Our data set is therefore comprised of 4,077 chicken–zebra finch orthologous coding regions, corresponding to 2.36 Mb of sequence data. Of these, 176 genes were mapped to the Z chromosome.

For the entire genomic data set, there were more male- than female-biased genes across the range of cutoff values (table 1), though this disappeared after sex-linked coding regions were removed (see below). For all genomic coding regions, female- and male-biased genes showed a significantly higher d_N/d_S ratio than unbiased genes expressed in brain across the range of cutoff values (125%, 133%, and 150% higher expression in either sex, table 1), though only female-biased expression categories showed consistent significant differences across all fold-change cutoffs. In every case, this was due to an increased d_N for sex-biased categories. Bootstrap testing of expression categories indicated that female-biased genes showed both higher d_N/d_S ratios as well as wider 95% CIs than male-biased genes (fig. 1).

View this table: [in this window] [in a new window]   Table 1 Mean Divergence Values for Sex-Biased Brain Genes for All Nuclear-Coding Regions

 

View larger version (12K): [in this window] [in a new window] [Download PowerPoint slide]   FIG. 1.— dN/dS ratios for unbiased-, female-, and male-biased genes for both genomic (A, C, and E) and autosomal(B, D, and F) coding regions. Data from 125% (A and B), 133% (C and D), and 150% (E and F) fold-change cutoff values are shown. The 95% confidence whiskers for sex-biased genesets are based on bootstrap values (1,000 repetitions). Sex-biased values that differ significantly (P < 0.05) from the unbiased gene class in each fold-change category are marked (*).

 
There are sound reasons to focus on the autosomal set of genes as sex-linked genes may evolve unusually fast due to the Fast-Z effect (Mank et al. 2007), an evolutionary phenomenon completely independent of sex-biased expression (Charlesworth 1991). Additionally, recent data suggest that birds lack dosage compensation, therefore, most Z-linked genes are male biased in expression due solely to gene dose (Ellegren et al. 2007; Itoh et al. 2007). These 2 characteristics create the possibility that the Fast-Z effect (analogous to Fast-X, Charlesworth 1991) could be mistakenly taken for rapid evolution of male-biased genes if the Z chromosome is included in the analysis. Consistent with this reasoning, confining further analysis to the 3,901 autosomal coding regions comprising 2.24 Mb significantly altered the picture of molecular evolution for some sex-biased categories. Although female-biased genes maintained a consistently higher d_N/d_S ratio in the autosomal data set (fig. 1 and table 2, permutation test with 1,000 repetitions, P 0.05), male-biased autosomal genes did not have a d_N/d_S ratio that differed statistically from the unbiased average. For example, for genes with a 150% difference in expression level between the sexes, d_N/d_S was roughly 43% higher for female-biased genes (0.1145) than for unbiased (0.08) and male-biased (0.0654) genes. As in the full genomic data set, the high d_N/d_S of female-biased genes was due to significantly higher d_N than unbiased genes. Male-biased genes, though generally exhibiting higher d_S than unbiased genes, lacked significantly higher d_N values.

View this table: [in this window] [in a new window]   Table 2 Mean Divergence Values for Autosomal Sex-Biased Brain Genes

 
Microarray studies such as the one employed here detect gene expression in given tissues, but they do not differentiate between broadly or narrowly expressed genes. Many genes expressed in brain are also expressed in other tissues throughout the organism (detailed in supplementary table S3, Supplementary Material online) and may be responding to selective pressures for extraneurological function. In general, pleiotropic genes are constrained in the evolution of sex-biased expression (Mank et al. 2008). In order to focus on genes potentially involved in cognitive and behavioral processes, we analyzed the abundance of EST transcripts in chicken cDNA libraries from a variety of tissues from the UniGene EST assemblage. The EST data indicated that 313 coding regions from our autosomal data set were either limited to or primarily expressed in the central nervous system. This subset of data exhibited the same substitution profile as the full autosomal data set, with female-biased d_N/d_S (0.123 at 125% fold change) genes nearly 2-fold higher than our estimate for unbiased genes (0.066). Similar to the complete autosomal data set, male-biased genes (0.069) did not have a higher overall d_N/d_S compared with unbiased genes (table 3).

View this table: [in this window] [in a new window]   Table 3 Mean Divergence Values for Sex-Biased Genes Limited to or Primarily Expressed in the Chicken Central Nervous System Based on UniGene EST Data

 
Expression level has been tied to rates of evolution, with highly expressed genes generally showing decreased rates of functional change than genes with lower expression levels (Pal et al. 2001; Subramanian and Kumar 2004; Liao and Zhang 2006). Expression level in our data, averaged separately over male and female samples, was weakly though significantly associated with d_N/d_S for both sexes (r² = 0.0289 for female expression level, r² = 0.0308 for male expression levels; P < 0.001 in both cases), with both relationships showing that more highly expressed genes evolve more slowly. However, although female- and male-biased classes had lower average expression levels than unbiased genes, they did not differ from each other (fig. 2).

View larger version (16K): [in this window] [in a new window] [Download PowerPoint slide]   FIG. 2.— Average expression of expression categories (125% fold-change level). Shown are the average relative expression levels averaged across within-sex samples, for female-, male-, and unbiased gene expression categories. The 95% confidence whiskers are shown, based on bootstrapping (1,000 replicates).

 
Different functional classes of genes can show different rates of evolution in response to stronger or weaker selection pressures; therefore, we tested for overrepresentation of GO terms in our gene categories. After correction for multiple comparisons, there were no significant differences in the occurrence of GO terms between sex-biased gene classes and the full autosomal gene set. In all cases, Bonferroni-adjusted P values exceeded 0.1.

Substitution rates as well as selective constraint in avian genomes have been shown to relate to several genomic parameters in birds. Specifically, genes on the microchromosomes have generally lower d_N/d_S and higher d_S than macrochromosome genes (ICGSC 2004; Axelsson et al. 2005). Given the d_N/d_S differences among female-biased, male-biased, and unbiased brain genes described above, it is important to test whether these differences can be explained by genomic parameters known to influence substitution patterns. There is no observed heterogeneity in the genomic distribution of sex-biased categories (supplementary table S4, Supplementary Material online) as no chromosome class deviated from random expectation regarding the number of male- or female-biased genes (² tests, P > 0.07 in all cases, 1 df for each test).

GC and third position GC (GC3) content varied by expression category according to permutation testing (tables 4 and 5). In birds, regions of high GC, and corresponding high incidence of mutable CpG sites, have a corresponding elevated d_S level (Webster et al. 2006). Our data are consistent with these as female-biased genes in general have lower GC and GC3 content, which is consistent with their low d_S, than brain-expressed genes. Male-biased GC and GC3 were significantly higher, consistent with their high d_S. The relationship between GC content and d_N/d_S was less clear. There was an overall negative correlation between GC content and d_N/d_S in the different expression categories (for all correlations <0.001), and we detected no significant heterogeneity among best-fit lines between GC and d_N/d_S (ANCOVA F = 0.30, P = 0.74, 2 df). Statistically different d_N/d_S levels among expression classes remained after accounting for variance in GC (ANCOVA F = 5.45, P = 0.004), suggesting that the differences in d_N/d_S we observed for male-biased, female-biased, and unbiased expressed categories are, at least in part, independent of GC content. Similarly, the relationship between GC3 and d_N/d_S did not differ among expression categories (ANCOVA F = 0.10, P = 0.901, 2 df); however, GC3 did not differ after accounting for d_N/d_S differences in expression classes (ANCOVA F = 0.64, P = 0.529, 2 df). Although GC and GC3 can indicate bias in codon usage, we observed no difference in codon bias among female-, male-, and unbiased genes (ANOVA F = 2.264, P = 0.104, 2 df) (data provided by Matthew Webster).

View this table: [in this window] [in a new window]   Table 4 Average GC Content for Sex-Biased Expression Categories

 

View this table: [in this window] [in a new window]   Table 5 Average GC3 for Sex-Biased Expression Categories

 

	Discussion

TOP Abstract Introduction Materials and Methods Results Discussion Supplementary Material Acknowledgements References

 
Sex-biased genes expressed in the avian brain showed a higher d_N/d_S ratio, suggesting that they have a faster functional rate of evolution than unbiased genes. The autosomes and Z chromosome showed somewhat different patterns however, which merit separate discussion. For autosomal genes, only female-biased genes evolve faster than unbiased genes.

The lack of dosage compensation on the chicken Z chromosome (Ellegren et al. 2007; Itoh et al. 2007) results in male-biased expression for most Z-linked coding regions, but this does not mean that selection is acting explicitly on male-biased expression for genes located on the Z, as has been reported in other species (Swanson et al. 2001; Meiklejohn et al. 2003; Parisi et al. 2004; Zhang et al. 2004; Pröschel et al. 2005). Z-linked genes in birds, regardless of expression bias, exhibit a faster rate of functional change than autosomal genes (Mank et al. 2007). This may be either due to the hemizygous exposure of Z-linked genes in the heterogametic sex, which facilitates the fixation of nondominant advantageous mutations, or due to the reduction in effective population size of the Z compared with autosomes, which reduces the effectiveness of selection against weakly deleterious alleles (Charlesworth 1991). Both these phenomena, which are independent of sex-biased expression, could result in elevated d_N/d_S and a spurious correlation between expression bias and rate of evolution in this study. The fact that male-biased autosomal genes do not show the elevated ratios of synonymous to nonsynonymous substitution in our analysis suggests that the increased d_N/d_S ratio we observed on the Z chromosome is more in-line with the Fast-Z evolution than to the effects of male bias per se.

Focusing our analysis on autosomal genes results in a pattern that is somewhat contradictory to previous studies of reproductive genes on other organisms where male-biased categories showed higher rates of functional change (Swanson et al. 2001; Torgerson et al. 2002; Zhang et al. 2004; Panhuis et al. 2006). By contrast, our data indicate that female-biased autosomal categories exhibit the highest d_N/d_S ratios, a pattern for which there are several possible explanations that can be divided into selectionist (either natural or sexual) and neutral nonselectionist categories.

We investigated a series of neutral nonselectionist genomic properties that could theoretically have produced the pattern of evolution we observed here. First, we tested whether heterogeneity in the distribution of sex-biased genes among the different chromosome classes could have produced our observations. Size differences in the avian karyotype result in different rates of recombination and mutation, as well as different GC content, for their associated gene complements (ICGSC 2004; Axelsson et al. 2005). However, as there was no significant bias in expression category distribution, the avian karyotype alone cannot account for the pattern we observed.

Next, we examined whether expression level could account for the elevated rate of evolution for female-biased genes. In general, highly expressed genes evolve more slowly than genes with low expression (Pal et al. 2001; Subramanian and Kumar 2004; Yang et al. 2005; Liao and Zhang 2006). Although sex-biased genes exhibit a lower expression level than unbiased genes (fig. 2), the expression between female- and male-biased genes was too similar (accounting for inversion of sex bias) to explain the difference between female- and male-biased d_N/d_S levels. GC and GC3 levels were significantly higher in male-biased genes (tables 4 and 5) and significantly lower in female-biased genes compared with unbiased gene classes. As GC correlates with recombination rate (Chen et al. 2006; Khelifi et al. 2006; Webster et al. 2006), this may indicate that recombination is more common in male-biased genes. Unfortunately, the lack of a high-density sex-specific recombination map in chicken prevents further testing of this possibility. Finally, despite the difference in GC and GC3, there was no significant difference among expression categories with regards to codon usage; therefore, it is not appropriate to invoke codon bias as a possible explanation for the elevated d_N/d_S levels for female-biased genes.

This leaves sexual and natural selection as viable alternatives to explain the observed pattern, and we can offer several possible alternatives. One possibility is that the force of sexual selection is stronger on female-specific behaviors, something that could potentially result in a faster rate of evolution for female-biased brain genes. However, this seems unlikely as the rigors of female choice and male–male competition, common selective pressures in birds that involve sex-specific behaviors (Gould and Gould 1997), would exert a powerful evolutionary force on male breeding behaviors and therefore male-biased brain genes. Sexual selection could still be acting on male-biased brain genes that are expressed at a different time in the life cycle. Little is known about the temporal aspects of sex-biased expression differences through an organism's life span, nor how this would affect our results. We collected tissue for microarray analysis at day 18 of development, well after circulating sex hormones initiate somatic sexual differentiation in chickens (Bruggeman et al. 2002). However, it is possible that the genes involved in male-specific reproductive behaviors, which would presumably be subject to sexual selection, are simply expressed at a different life stage than that which forms the data set for this analysis.

Additionally, we do not yet have sufficient gene expression data for zebra finch to determine the extent to which sex-biased gene expression is conserved between these study species or among birds in general. Gene expression in general in the brain shows a high degree of conservation (Strand et al. 2007). Additionally, the experimental designs of both this microarray study and the source for the zebra finch orthologs (Clayton 2004) ensures that the genes employed here are expressed in the brains of both chicken and zebra finch. However, sex-biased expression for some genes has been shown to alter relatively rapidly among species in response to changing selective regimes (Ranz et al. 2003; Metta et al. 2006). We used data here on sex-biased expression differences in chicken, but the pattern of sex bias may be somewhat different for some subset of our genes in zebra finch.

Natural selection, as opposed to sexual selection, may be acting differently on the male and female brain to produce the d_N/d_S patterns described here. For instance, natural selection could be acting on female-specific behaviors involved in brood care or other aspects of reproduction. Additionally, a high d_N/d_S is not necessarily a consequence of adaptive evolution but could also reflect relaxed selective constraint. Male-biased and unbiased genes may, for some reason, be subject to stronger purifying selection, which would result in lower d_N/d_S values. A similar discussion is made for brain genes of humans and chimpanzee, where elevated rates of sequence evolution could either be due to adaptive evolution or relaxed constraint (Dorus et al. 2004; Khaitovich et al. 2005; Shi et al. 2006; Wang et al. 2006).

In summary, this study represents an attempt to look at the rate of molecular evolution of neurological sex-biased genes that are likely contributors to sexually dimorphic behaviors. Taken as a whole, our results suggest that sexually dimorphic behaviors result from strong evolutionary pressures acting on the avian brain, primarily in females. Future work, involving both gene expression and avian sequence data, will be helpful in further clarifying sex-specific evolutionary processes acting on avian neurology.

	Supplementary Material

TOP Abstract Introduction Materials and Methods Results Discussion Supplementary Material Acknowledgements References

 
Supplementary tables S1–S4 are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).

	Acknowledgements

TOP Abstract Introduction Materials and Methods Results Discussion Supplementary Material Acknowledgements References

 
Support for this work was provided by the Wenner-Gren Foundation and the Swedish Research Council. We thank Matthew Webster and Martin Zwahlen for help with data analysis and 2 anonymous reviewers for helpful comments.

	Footnotes

 
David Irwin, Associate Editor

	References

TOP Abstract Introduction Materials and Methods Results Discussion Supplementary Material Acknowledgements References

 

Andersson M. Sexual selection (1994) Princeton (NJ): Princeton University Press.

Axelsson E, Smith NGC, Sundström H, Berlin S, Ellegren H. Male-biased mutation rate and divergence in autosomal, Z-linked and W-linked introns of chicken and turkey. Mol Biol Evol (2004) 21:1538–1547.[Abstract/Free Full Text]

Axelsson E, Webster MT, Smith NGC, Burt DW, Ellegren H. Comparison of the chicken and turkey genomes reveals a higher rate of nucleotide divergence on microchromosomes than macrochromosomes. Genome Res (2005) 15:120–125.[Abstract/Free Full Text]

Bruggeman V, van As P, Decuypere E. Developmental endocrinology of the reproductive axis in the chicken embryo. Comp Biochem Physiol A (2002) 131:839–846.[CrossRef][Medline]

Charlesworth B. The evolution of sex chromosomes. Science (1991) 251:1030–1033.[Abstract/Free Full Text]

Chen JF, Lu F, Chen SS, Tao SH. Significant positive correlation between the recombination rate and GC content in the human pseudoautosomal region. Genome (2006) 49:413–419.[Medline]

Civetta A, Singh RS. Broad-sense sexual selection, sex gene pool evolution, and speciation. Genome (1999) 42:1033–1041.[Medline]

Clayton DF. Songbird genomics: methods, mechanisms, opportunities, and pitfalls. Ann N Y Acad Sci (2004) 1016:45–60.[CrossRef][Web of Science][Medline]

Curwen V, Eyras E, Andrews TD, Clarke L, Mongin E, Searle SMJ, Clamp M. The Ensembl automatic gene annotation system. Genome Res (2004) 14:942–950.[Abstract/Free Full Text]

Dauwalder B, Tsujimoto S, Moss J, Mattox W. The Drosophila takeout gene is regulated by the somatic sex-determination pathway and affects male courtship behavior. Genes Dev (2002) 16:2879–2892.[Abstract/Free Full Text]

Dorus S, Vallender EJ, Evans PD, Anderson JR, Gilbert SL, Mahowald M, Wyckoff GJ, Malcom C, Lahn BT. Accelerated evolution of nervous system genes in the origin of Homo sapiens. Cell (2004) 119:1027–1040.[CrossRef][Web of Science][Medline]

Drapeau MD, Radovic A, Wittkopp PJ, Long AD. A gene necessary for normal male courtship, yellow, acts downstream of fruitless in the Drosophila melanogaster larval brain. J Neurobiol (2003) 55:53–72.[CrossRef][Web of Science][Medline]

Ellegren H, Hultin-Rosenberg L, Brunström B, Dencker L, Kultima K, Scholtz B. Faced with inequality: chicken does not have a general dosage compensation of sex-linked genes. BMC Biol (2007) 5:40.[CrossRef][Medline]

Ellegren H, Parsch J. The evolution of sex-biased genes and sex-biased gene expression. Nat Rev Genet (2007) 8:689–698.[CrossRef][Web of Science][Medline]

Ericson E, Pylvanainen I, Fernandez-Ricaud L, Nerman O, Warringer J, Blomberg A. Genetic pleiotropy in Saccharomyces cerevisiae quantified by high-resolution phenotypic profiling. Mol Genet Genomics (2006) 275:605–614.[CrossRef][Web of Science][Medline]

Gene Ontology Consortium. Gene ontology: a tool for the unification of biology. Nat Genet (2000) 25:25–29.[CrossRef][Web of Science][Medline]

Gould JL, Gould CG. Sexual selection: mate choice and courtship in nature (1997) New York: Scientific American Library.

Hambuch TM, Parsch J. Patterns of synonymous codon usage in Drosophila melanogaster genes with sex-biased expression. Genetics (2005) 170:1691–1700.[Abstract/Free Full Text]

[ICGSC] 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]

Ihkaka R, Gentleman R. R: a language for data analysis and graphics. J Comput Graph Stat (1996) 5:299–314.[CrossRef]

Itoh Y, Melamed E, Yang X. 13 co-authors. Dosage compensation is less effective in birds than in mammals. J Biol (2007) 6:2.[CrossRef][Medline]

Jagadeeshan S, Singh RS. Rapidly evolving genes of Drosophila: differing levels of selective pressure in testis, ovary, and head tissues between sibling species. Mol Biol Evol (2005) 22:1793–1801.[Abstract/Free Full Text]

Kadener S, Villella A, Kula E, Palm K, Pyza E, Botas J, Hall JC, Rosbash M. Neurotoxic protein expression reveals connections between the circadian clock and mating behavior in Drosophila. Proc Natl Acad Sci USA (2006) 103:13537–13542.[Abstract/Free Full Text]

Khaitovich P, Hellmann I, Enard W, Nowick K, Leinweber M, Franz H, Weiss G, Lachmann M, Paabo S. Parallel patterns of evolution in the genomes and transcriptomes of humans and chimpanzees. Science (2005) 309:1850–1854.[Abstract/Free Full Text]

Khelifi A, Meunier J, Duret L, Mouchiroud D. GC content evolution of the human and mouse genomes: insights from the study of processed pseudogenes in regions of different recombination rates. J Mol Evol (2006) 62:745–752.[CrossRef][Web of Science][Medline]

Liao BY, Zhang JZ. Low rates of expression profile divergence in highly expressed genes and tissue-specific genes during mammalian evolution. Mol Biol Evol (2006) 23:1119–1128.[Abstract/Free Full Text]

Mank JE, Axelsson E, Ellegren H. Fast-X on the Z: rapid evolution of sex-linked genes in birds. Genome Res (2007) 17:618–624.[Abstract/Free Full Text]

Mank JE, Hultin-Rosenberg L, Zwahlen M, Ellegren H. Pleiotropic constraint hampers the resolution of sexual antagonism in vertebrate gene expression. Am Nat (2008) Forthcoming.

Marinotti O, Calvo E, Nguyen QK, Dissanayake S, Ribeiro JMC, James AA. Genome-wide analysis of gene expression in adult Anopheles gambiae. Insect Mol Biol (2006) 15:1–12.[CrossRef][Web of Science][Medline]

Meiklejohn CD, Parsch J, Ranz JM, Hartl DL. Rapid evolution of male-biased gene expression in Drosophila. Proc Natl Acad Sci USA (2003) 100:9894–9899.[Abstract/Free Full Text]

Metta M, Gudavalli R, Gibert JM, Schlotterer C. No accelerated rate of protein evolution in male-biased Drosophila pseudoobscura genes. Genetics (2006) 174:411–420.[Abstract/Free Full Text]

Meunier J, Duret L. Recombination drives the evolution of GC-content in the human genome. Mol Biol Evol (2004) 21:984–990.[Abstract/Free Full Text]

Morgenstern B. DIALIGN 2: improvements of the segment-to-segment approach to multiple sequence alignment. Bioinformatics (1999) 15:211–218.[Abstract/Free Full Text]

Pal C, Papp B, Hurst LD. Highly expressed genes in yeast evolve slowly. Genetics (2001) 158:927–931.[Free Full Text]

Panhuis TM, Clark NL, Swanson WJ. Rapid evolution of reproductive proteins in abalone and Drosophila. Phil Trans R Soc B (2006) 361:261–268.[CrossRef][Medline]

Parisi M, Nuttall R, Edwards P. (12 co-authors). A survey of ovary-, testis-, and soma-biased gene expression in Drosophila melanogaster adults. Genome Biol (2004) 5:R40.[CrossRef][Medline]

Pröschel M, Zhang Z, Parsch J. Widespread adaptive evolution of Drosophila genes with sex-biased expression. Genetics (2005) 174:893–900.[Web of Science]

Ranz JM, Castillo-Davis CI, Meiklejohn CD, Hartl DL. Sex-dependent gene expression and evolution of the Drosophila transcriptome. Science (2003) 300:1742–1745.[Abstract/Free Full Text]

Rinn JL, Snyder M. Sexual dimorphism in mammalian gene expression. Trends Genet (2005) 21:298–305.[CrossRef][Web of Science][Medline]

Robinson PN, Wollstein A, Böhme U, Beattie B. Ontologizing gene-expression microarray data: characterizing clusters with Gene Ontology. Bioinformatics (2004) 20:979–981.[Abstract/Free Full Text]

Shi P, Bakewell MA, Zhang J. Did brain-specific genes evolve faster in humans than in chimpanzees? Trends Genet (2006) 22:608–613.[CrossRef][Web of Science][Medline]

Smith NGC, Webster MT, Ellegren H. Deterministic mutation rate variation in the human genome. Genome Res (2002) 12:1350–1356.[Abstract/Free Full Text]

Smyth GK. Linear models and empirical Bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol (2004) 3:3.

Spencer CCA, Deloukas P, Hunt S, Mullikin J, Myers S, Silverman B, Donnelly P, Bentley D, McVean G. The influence of recombination on human genetic diversity. PLoS Genet (2006) 2:1375–1385.[Web of Science]

Strand AD, Aragaki AK, Baquet ZC, Hodges A, Cunningham P, Holmans P, Jones KR, Jones L, Kooperberg C, Olson JM. Conservation of regional gene expression in mouse and human brain. PLoS Genet (2007) 3:e59.[CrossRef][Medline]

Subramanian S, Kumar S. Gene expression intensity shapes evolutionary rates of the proteins encoded by the vertebrate genome. Genetics (2004) 168:373–381.[Abstract/Free Full Text]

Swanson WJ, Clark AG, Waldrip-Dail HM, Wolfner MF, Aquadro CF. Evolutionary EST analysis identifies rapidly evolving male reproductive proteins in Drosophila. Proc Natl Acad Sci USA (2001) 98:7375–7379.[Abstract/Free Full Text]

Swanson WJ, Vacquier VD. The rapid evolution of reproductive proteins. Nat Rev Genet (2002) 3:137–144.[Web of Science][Medline]

Swanson WJ, Wong A, Wolfner MF, Aquadro CF. Evolutionary expressed sequence tag analysis of Drosophila female reproductive tracts identifies genes subjected to positive selection. Genetics (2004) 168:1457–1465.[Abstract/Free Full Text]

Torgerson DG, Kulathinal RJ, Singh RS. Mammalian sperm proteins are rapidly evolving: evidence of positive selection in functionally diverse genes. Mol Biol Evol (2002) 19:1973–1980.[Abstract/Free Full Text]

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.[Abstract/Free Full Text]

Wang HY, Chien HC, Osada N, Hashimoto K, Sugano S, Gojobori T, Chou CK, Tsai SF, Wu CI, Shen CK. Rate of evolution in brain-expressed genes in humans and other primates. PLoS Biol (2006) 5:e1.

Waxman D, Peck JR. Pleiotropy and the preservation of perfection. Science (1998) 279:1210–1213.[Abstract/Free Full Text]

Webster MT, Axelsson E, Ellegren H. Strong regional biases in nucleotide substitution in the chicken genome. Mol Biol Evol (2006) 23:1203–1216.[Abstract/Free Full Text]

Williams EJB, Hurst LD. The proteins of linked genes evolve at similar rates. Nature (2000) 407:900–903.[CrossRef][Medline]

Wright F. The effective number of codons used in a gene. Gene (1990) 87:23–29.[CrossRef][Web of Science][Medline]

Wu Z, Irizarry RA. Stochastic models inspired by hybridization theory for short oligonucleotide arrays. J Comput Biol (2004) 12:882–893.[CrossRef][Web of Science]

Yang Z. PAML: a program package for phylogenetic analysis by maximum likelihood. Comput Appl Biosci (1997) 13:555–556.[Free Full Text]

Yang X, Schadt EE, Wang S, Wang H, Arnold AP, Ingram-Drake L, Drake TA, Lusis AJ. Tissue-specific expression and regulation of sexually dimorphic genes in mice. Genome Res (2006) 16:995–1004.[Abstract/Free Full Text]

Yang J, Su AI, Li WH. Gene expression evolves faster in narrowly than in broadly expressed mammalian genes. Mol Biol Evol (2005) 22:2113–2118.[Abstract/Free Full Text]

Zhang Z, Hambuch TM, Parsch J. Molecular evolution of sex-biased genes in Drosophila. Mol Biol Evol (2004) 21:2130–2139.[Abstract/Free Full Text]

Zhang Z, Parsch J. Positive correlation between evolutionary rate and recombination rate in Drosophila genes with male-biased expression. Mol Biol Evol (2005) 22:1945–1947.[Abstract/Free Full Text]

Accepted for publication September 12, 2007.

CiteULike    Connotea    Del.icio.us    What's this?

This article has been cited by other articles:

				Z.-F. Jiang and C. A. Machado Evolution of Sex-Dependent Gene Expression in Three Recently Diverged Species of Drosophila Genetics, November 1, 2009; 183(3): 1175 - 1185. [Abstract] [Full Text] [PDF]

				J. E. Mank and H. Ellegren Are sex-biased genes more dispensable? Biol Lett, June 23, 2009; 5(3): 409 - 412. [Abstract] [Full Text] [PDF]

				N. Derome, B. Bougas, S. M. Rogers, A. R. Whiteley, A. Labbe, J. Laroche, and L. Bernatchez Pervasive Sex-Linked Effects on Transcription Regulation As Revealed by Expression Quantitative Trait Loci Mapping in Lake Whitefish Species Pairs (Coregonus sp., Salmonidae) Genetics, August 1, 2008; 179(4): 1903 - 1917. [Abstract] [Full Text] [PDF]

This Article Abstract FREE Full Text (PDF) Supplementary Material All Versions of this Article: 24/12/2698    most recent msm208v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Request Permissions Google Scholar Articles by Mank, J. E. Articles by Ellegren, H. Search for Related Content PubMed PubMed Citation Articles by Mank, J. E. Articles by Ellegren, H. Social Bookmarking          What's this?

Online ISSN 1537-1719 - Print ISSN 0737-4038

Copyright © 2010 Society for Molecular Biology and Evolution

Oxford Journals Oxford University Press

• Site Map
• Privacy Policy
• Frequently Asked Questions

Other Oxford University Press sites: Oxford University Press Oxford Journals China Oxford Journals Japan American National Biography Booksellers' Information Service Children's Fiction and Poetry Children's Reference Corporate & Special Sales Dictionaries Dictionary of National Biography Digital Reference English Language Teaching Higher Education Textbooks Humanities International Education Unit Law Medicine Music Online Products Oxford English Dictionary Reference Rights and Permissions Science School Books Social Sciences Very Short Introductions World's Classics