Genomics 95 (2010) Contents lists available at ScienceDirect. Genomics. journal homepage:

Transcription

1 Genomics 95 (2010) Contents lists available at ScienceDirect Genomics journal homepage: Evolutionary genomics of the Fox genes: Origin of gene families and the ancestry of gene clusters Sebastian M. Shimeld a,, Bernard Degnan b, Graham N. Luke c a Department of Zoology, University of Oxford, South Parks Road, Oxford OX1 3PS, UK b School of Biological Sciences, University of Queensland, Brisbane QLD 4072, Australia c School of Biological Sciences, University of Reading, Whiteknights, Reading, UK article info abstract Article history: Received 8 July 2009 Accepted 3 August 2009 Available online 10 August 2009 Keywords: Fox gene Gene cluster Gene duplication Gene loss Over the past decade genomic approaches have begun to revolutionise the study of animal diversity. In particular, genome sequencing programmes have spread beyond the traditional model species to encompass an increasing diversity of animals from many different phyla, as well as unicellular eukaryotes that are closely related to the animals. Whole genome sequences allow researchers to establish, with reasonable confidence, the full complement of any particular family of genes in a genome. Comparison of gene complements from appropriate genomes can reveal the evolutionary history of gene families, indicating when both gene diversification and gene loss have occurred. More than that, however, assembled genomes allow the genomic environment in which individual genes are found to be analysed and compared between species. This can reveal how gene diversification occurred. Here, we focus on the Fox genes, drawing from multiple animal genomes to develop an evolutionary framework explaining the timing and mechanism of origin of the diversity of animal Fox genes. Ancient linkages between genes are a prominent feature of the Fox genes, depicting a history of gene clusters, some of which may be relevant to understanding Fox gene function Elsevier Inc. All rights reserved. Introduction: The Fox genes The Fox genes take their name from the Forkhead box, encoding a DNA binding domain first identified via the Drosophila melanogaster fork head gene [1]. The forkhead domain is approximately 100 amino acids in size and is usually present in single copy within a typical Fox protein, with a small number of divergent genes known to encode multiple domains. Fox proteins use the forkhead domain for sequence-specific DNA binding, marking them as transcription factors. Numerous regulatory roles in development and metabolism have been described (reviewed by [2,3,4,5]). Notable members include the FoxO genes with a key role in mediation of insulin signalling, FoxA genes with roles in early endoderm, liver, lung and ventral neural development and FoxJ1 with a key role in cilia formation [2,3]. Exactly when the forkhead domain evolved is unknown. The domain itself is of the winged-helix type and as such bears structural similarity to some bacterial domains that interact with DNA. In eukaryotes forkhead domains have been identified in animal and fungal genomes and in a small number of unicellular eukaryotes that are closely related to animals and fungi. Together these species comprise a eukaryotic lineage called the opisthokonts [6], and Corresponding author. Fax address: sebastian.shimeld@zoo.ox.ac.uk (S.M. Shimeld). forkhead domains have not been detected in other eukaryotes including plants, slime molds and myriad unicellular forms. Fox gene diversity and classification in the Bilateria Forty-six Fox genes have been annotated in the mouse genome and 18 in the D. melanogaster genome, with molecular phylogenetic analyses used to construct an evolution-based classification system [4,7,8,9]. Groups of orthologous genes in the Bilateria (hereafter called families) are designated by a letter: currently named families range from FoxA to FoxS, although FoxR and FoxS are vertebrate-specific [7,10]. Four families (FoxL, J, N and Q) have been further subdivided and an additional family (FoxAB) erected as more invertebrate members were discovered and incorporated into the analyses [7,8,9]. This yields 22 definitive bilaterian Fox families (Fig. 1), one member of each of which would have been present in the genome of the last common ancestor of protostomes and deuterostomes. Note that here we take this ancestor as synonymous with the last common ancestor of the Bilateria; while the later may also include the acoel flatworms and allies which evidence suggests branched before the protostome-deuterostome last common ancestor [11], insufficient genome data precludes their inclusion in this discussion of Fox gene diversity. Although these 22 families appear to represent the primitive condition for the Bilateria, most species appear to have lost one or /$ see front matter 2009 Elsevier Inc. All rights reserved. doi: /j.ygeno

2 S.M. Shimeld et al. / Genomics 95 (2010) Fig. 1. Fox gene family membership in the bilaterians, three basal animal lineages and the single-celled choanoflagellate M. brevicollis. Fox families from A to Q2 are shown, divided into two major clades [19] (see text for details). Numbers under species names indicate the total number of Fox genes identified in that genome. Black squares show where an orthologue has been described in the respective species. The striped squares indicate possible orthologues where family membership is less robust, and a square bridging two families indicates a gene which may be related to both families. Bilaterians are represented by a single branch and are by definition inferred to have primitively had members of all 22 Fox families, although most individual species have lost one or more families (see Fig. 2). The placozoan T. adhaerens and cnidarian N. vectensis are shown as equally related the bilaterians, reflecting uncertainty over their genuine phylogenetic relationships. Between them they have orthologues of 20 Fox families. The sponge A. queenslandica is depicted as the earliest-branching animal lineage and has members of most clade II families but only four clade I families. Of the eight M. brevicollis genes, convincing members of only three families can be identified. For sources of data and orthologue descriptions, please see supplementary material. more family at some point in their evolutionary history. For example vertebrates appear to lack a FoxAB gene, as do sea squirts, although clear FoxAB genes have been identified in amphioxus and sea urchin genomes [7,8,12]. Similarly placental mammals lack a FoxQ2 gene (though it is found in other vertebrates including monotremes), and D. melanogaster is missing members of the FoxAB, E, H, I, Q1, J2/3 and M families [9]. Missing families could reflect incomplete genome data or, since family membership is based on sequence identity of the forkhead domain, they could represent genes that have diverged in forkhead domain sequence until they are unclassifiable. However few unclassifiable Fox genes have been found in most animal genomes (there are none for example in the well-sequenced mouse and human genomes), suggesting gene loss primarily explains missing families. The evolutionary correlates of such Fox gene losses are currently unknown. So far the only species in which at least one member of all 22 families have been identified is the amphioxus Branchiostoma floridae [7]. Timing the origin of Fox gene families: Fox gene complements in cnidarian, sponge, placozoan and choanoflagellate genomes While there is little controversy regarding the presumed monophyly of the Bilateria, the phylogenetic relationships of the basal metazoan taxa are less firmly established. Amongst the protozoa, the choanoflagellates are the closest living relatives of the metazoa and one species, Monosiga brevicollis [13], has a sequenced genome and acts as a proxy for the single-celled ancestors of the Metazoa. Sponges are historically considered to be the earliest diverging metazoan phylum. They may be paraphyletic, but only a single species has a sequenced genome (Amphimedon queenslandica) [14]. Of the three remaining taxa (ctenophores, cnidarians and placozoans), two include species with sequenced genomes; the cnidarian Nematostella vectensis [15], and the placozoan Trichoplax adhaerens [16]. The relationship between placozoans, cnidarians and bilaterians is disputed [16,17,18]. Here, we consider N. vectensis and T. adhaerens as equally related to the bilaterians, with A. queenslandica as having diverged prior to the radiation of these three lineages (Fig. 1). Surveying the Fox gene complements of N. vectensis and T. adhaerens reveals a striking pattern; both species have clear orthologues of most of the bilaterian families, with 16 out of 22 in T. adhaerens and 19 out of 22 in N. vectensis (Fig. 1) [15,16,19, our unpublished analyses]. Between them, they include 20 out of 22 families, with only FoxH and FoxI absent (Fig. 1). This shows most of the family level diversity of Fox genes seen in the Bilateria evolved very early in animal evolution. Larroux et al. [19] have suggested that the Fox gene families can be divided into two major groups, clade I and clade II. Clade II gene families primitively contained introns in conserved sites in the forkhead box. Clade I gene families did not primitively contain introns in the forkhead box (though introns have evolved in some genes in specific lineages).the clade II condition can be inferred to be primitive for all Fox genes as it is found in the M. brevicollis genes, while the clade I condition is not. With this in mind, the situation in A. queenslandica is different from other animals; this sponge genome includes members of seven out of eight clade II families, but only 4 out of 14 clade I families (Fig. 1) [19]. In M. brevicollis, only three Fox genes can be attributed to families, all in clade II. Simplistically, this suggests eight families (four clade I, four clade II) evolved in the early animal lineage before the divergence of A. queenslandica, and nine (eight clade I, one clade II) between the divergence of A. queenslandica and the remaining animal phyla. The conundrum this poses is that homologous genes evolve by duplication and divergence. If this explains the increasing diversity of Fox genes in animal evolution, then A. queenslandica genes should not be neatly attributable to bilaterian families, but should branch basally to the groups of bilaterian families into which their ancestor duplicated and diversified. This general pattern is not confined to Fox genes, and is also seen for homeobox, Sox and other genes [19]. Two hypotheses could explain this discrepancy. First, as suggested for homeobox genes [20], extensive gene loss may characterise these early diverging lineages, such that the common ancestor of the animals had genes orthologous to most bilaterian Fox families. Many of these would have to have been lost by A. queenslandica. This has not

3 258 S.M. Shimeld et al. / Genomics 95 (2010) been disproved, though it is unsatisfying as it just pushes the problem of understanding metazoan gene family diversity a little further back in time. Ctenophore and additional sponge genomes may help test this hypothesis. An alternate hypothesis is that Fox (and other) phylogenetic analyses may be misleading, in that A. queenslandica genes that appear to fall into a bilaterian family are in fact orthologous to more than one family. This could happen if, following duplication, one paralogue were to retain the ancestral sequence while the other diverged rapidly before stabilising and founding a second family. This would require that lineages that diverged during this time and which might reveal an intermediate state have become extinct. This may seem counterintuitive, however it does reflect aspects of the way many transcription factors function and are positioned in gene regulatory networks [21]. The near stasis of DNA binding domains over hundreds of millions of years of evolutionary time presumably results from them being locked in to such networks, via regulation of a battery of target genes via specific binding sites; changing the DNA binding domain would also therefore involve changes in binding specificity at many sites in the genome, with possible pleiotropic consequences. When constraint is relaxed, however, the sequence of the DNA binding domain can change rapidly then perhaps again become stabilised; examples of this have been documented, most notably bicoid in insects, which has evolved from a Hox gene by divergence then stabilisation [22]. Genomic organisation of Fox genes: Clusters past and present Assembled genome sequences allow the study of gene family evolution to be extended beyond molecular phylogenetic analyses. Specifically, the identification of gene clusters can shed light on mechanisms of diversification and functional constraints on evolution. Homeobox genes are the canonical example in this context, with such studies suggesting an ancient origin of many ANTP class genes by tandem duplication [23], and of functional underpinnings for Hox cluster longevity [24]. Recent studies however suggest other gene families may also be informatively analysed in this way, including the Wnt, GATA and Fox genes [25,26,27]. Fig. 2 illustrates linkages between Fox genes in representatives of three bilaterian clades, three basally diverging metazoan lineages and a choanoflagellate. Linkages of genes could arise in two ways; they could reflect recombination uniting previously isolated genes, or they could reflect common ancestry and origin via tandem duplication, a process known to be common in genome evolution. Addressing such alternatives statistically is currently beyond our models of genome organisation and evolution; however, we can consider two factors. First, the closer together two genes are, the more likely the association is due to tandem duplication. Second, linkages found in multiple deeply diverged lineages are likely to be homologous. Based on this, we have extrapolated two potential ancestral Fox gene organisations present before the divergence of the placozoan, cnidarian and bilaterian lineages, one conservative (based on close linkages found Fig. 2. Fox gene linkages in selected animal genomes. Lines indicate linkages, with thick lines indicating an intergenic distance less than 0.1% of the total predicted genome size and a thin line between 0.1% and 1%. Several additional linkages greater than 1% of genome size are present in the better-assembled genomes (particularly human) but are not shown. In some instances multiple genes are shown for a particular family in a species, for example, human FoxC. These represent paralogues deriving from lineage-specific duplication within a family and are only shown when relevant to linkages. At the top of the figure are two models for linkage in the common ancestor of bilaterians, N. vectensis and T. adhaerens. Note gene order is not the same as in Fig. 1, to simplify depiction of linkages. For the data sources on which this figure is based please see the supplementary information.

4 S.M. Shimeld et al. / Genomics 95 (2010) in multiple lineages) and one relaxed (which includes more distant linkages found in fewer lineages; Fig. 2). The conservative model predicts two ancient linkages (D + E and L1 + C + F + Q1). The relaxed model suggests the majority of Fox families were primitively linked. This is probably an overprediction with some linkages inferred as ancestral in fact deriving from lineage-specific recombination or genome misassembly. For example, it seems unlikely that the clade I genes evolved via tandem duplication of a clade II gene given their different intron exon organisation suggests duplication by an alternative route (for example retrotransposition) [19]. Additional assembled genomes and improved assemblies of currently sequenced genomes will help resolve this. However we consider it unlikely that more data will undermine our current conservative model, and hence we predict that as additional genomes are sequenced they will only act to increase our conservative inference of ancient gene linkages. Gene cluster longevity; maintenance versus dispersal in different lineages The FoxD FoxE genes In T. adhaerens and N. vectensis, as well as in the deuterostomes B. floridae and humans, FoxD and FoxE genes show evidence of tight linkage. Distances are in the region of 22 kb between FOXE3 and FOXD2 in humans, 129 kb between FoxEa and FoxD in B. floridae, and 45 kb and 30 kb between FoxD and FoxE in T. adhaerens and N. vectensis, respectively. Whether a pair of genes justifies the name cluster is a semantic point, however we can infer the two evolved as distinct genes by tandem duplication at least before the separation of cnidarian, placozoan and bilaterian lineages and have remained in close proximity in several lineages since this time. This is confirmed by the degree of sequence similarity between FoxE and FoxD genes (they are sister groups in some phylogenetic analyses) [7]. Notably in A. queenslandica there is a single FoxD/E gene which is hard to ascribe to one family or the other, with different analyses giving different results [19] (SMS unpublished analyses). This suggests the duplication occurred after the divergence of this sponge lineage. Such tight linkage over extensive evolutionary time in multiple lineages might reflect a selective pressure to maintain the genes together, though what this pressure might be is unknown. No FoxE genes have been found in ecdysozoans to date, suggesting loss in this lineage, and in many species (Capitella sp. I, Lottia gigantea, Ciona intestinalis; see Fig. 2) the genes are not linked [8, 9, 12]. This suggests loss of linkage has occurred several times in evolution. The FoxL1, FoxC, FoxF and FoxQ1 genes Members of these four gene families are tightly linked in many animal genomes, with cluster sizes of around 250 kb in deuterostomes, although as for the FoxD FoxE gene pair, break-up has occurred multiple times (for example D. melanogaster, C. intestinalis, Strongylocentrotus purpuratus) [8,9,12]. Studies of FoxL1, FoxC, FoxF and FoxQ1 genes in dogfish, amphioxus, molluscs and annelids have suggested that their expression in different mesodermal tissues might require a degree of co-regulation, providing a selective force to maintain the genes as a cluster, at least in some lineages [25,28] (Shimeld et al., unpublished data). However the identification and clustered organisation of three of these genes (FoxC, F and Q1) in T. adhaerens argues against such a simple relationship. T. adhaerens has a very simple body plan, with a few distinct cell types organised without clear tissues and no obvious mesoderm. Unless it has degenerated from a more complex form, then we would have to presume that a co-regulated cluster was involved in some other aspect of the biology of the common ancestor and was perhaps coopted during mesoderm evolution. Analysis of Fox gene function in T. adhaerens might help resolve this. Conclusions Comparison of Fox gene complements across animal genomes shows that the common ancestor of bilaterians, cnidarians and placozoans had an extensive repertoire of Fox genes families, with little change in the sequence of the Fox domain having occurred for most genes in most lineages since this time. The single sponge and choanoflagellate genomes available to date suggests that some of these families evolved after the divergence of sponges and some before it, although we note this interpretation is based on one phylogenetic tree out of several that have been proposed and could be subject to change. Ancient clusters of Fox genes provide further insights, both into the mechanism and timing of gene duplication in ancestors, and into putative selective pressure to retain linkages in their descendants. Acknowledgments We are greatly indebted to the sequencing centres responsible for the sequencing and assembly of the genomes we have examined and to the funding agencies that have supported these sequencing programmes. G.N.L. and S.M.S. acknowledge the support of the Biotechnology and Biological Sciences Research Council [G19873/2]. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi: /j.ygeno References [1] D. Weigel, H. Jackle, The fork head domain: a novel DNA binding motif of eukaryotic transcription factors? Cell 63 (1990) [2] G. Tuteja, K.H. Kaestner, SnapShot: forkhead transcription factors I, Cell 130 (2007) [3] G. Tuteja, K.H. Kaestner, Forkhead transcription factors II, Cell 131 (2007) 192. [4] K.H. Kaestner, W. Knochel, D.E. Martinez, Unified nomenclature for the winged helix/forkhead transcription factors, Genes Dev. 14 (2000) [5] P. Carlsson, M. Mahlapuu, Forkhead transcription factors: key players in development and metabolism, Dev. Biol. 250 (2002) [6] S.L. Baldauf, A.J. Roger, I. Wenk-Siefert, W.F. Doolittle, A kingdom-level phylogeny of eukaryotes based on combined protein data, Science 290 (2000) [7] J.K. Yu, F. Mazet, Y.T. Chen, S.W. Huang, K.C. Jung, S.M. Shimeld, The Fox genes of Branchiostoma floridae, Dev. Genes Evol. 218 (2008) [8] Q. Tu, C.T. Brown, E.H. Davidson, P. Oliveri, Sea urchin Forkhead gene family: phylogeny and embryonic expression, Dev. Biol. 300 (2006) [9] F. Mazet, J.K. Yu, D.A. Liberles, L.Z. Holland, S.M. Shimeld, Phylogenetic relationships of the Fox (Forkhead) gene family in the Bilateria, Gene 316 (2003) [10] K.R. Wotton, S.M. Shimeld, Comparative genomics of vertebrate fox cluster loci, BMC Genomics 7 (2006) [11] J. Baguna, P. Martinez, J. Paps, M. Riutort, Back in time: a new systematic proposal for the Bilateria, Philos. Trans. R. Soc. Lond. B Biol. Sci. 363 (2008) [12] K. Yagi, Y. Satou, F. Mazet, S.M. Shimeld, B. Degnan, D. Rokhsar, M. Levine, Y. Kohara, N. Satoh, A genomewide survey of developmentally relevant genes in Ciona intestinalis III. Genes for Fox, ETS, nuclear receptors and NFkappaB, Dev. Genes Evol. (2003) [13] N. King, M.J. Westbrook, S.L. Young, A. Kuo, M. Abedin, J. Chapman, S. Fairclough, U. Hellsten, Y. Isogai, I. Letunic, M. Marr, D. Pincus, N. Putnam, A. Rokas, K.J. Wright, R. Zuzow, W. Dirks, M. Good, D. Goodstein, D. Lemons, W. Li, J.B. Lyons, A. Morris, S. Nichols, D.J. Richter, A. Salamov, J.G. Sequencing, P. Bork, W.A. Lim, G. Manning, W.T. Miller, W. McGinnis, H. Shapiro, R. Tjian, I.V. Grigoriev, D. Rokhsar, The genome of the choanoflagellate Monosiga brevicollis and the origin of metazoans, Nature 451 (2008) [14] S.P. Leys, D.S. Rohksar, B.M. Degnan, Sponges, Curr. Biol. 15 (2005) R114 R115. [15] N.H. Putnam, M. Srivastava, U. Hellsten, B. Dirks, J. Chapman, A. Salamov, A. Terry, H. Shapiro, E. Lindquist, V.V. Kapitonov, J. Jurka, G. Genikhovich, I.V. Grigoriev, S.M. Lucas, R.E. Steele, J.R. Finnerty, U. Technau, M.Q. Martindale, D.S. Rokhsar, Sea anemone genome reveals ancestral eumetazoan gene repertoire and genomic organization, Science 317 (2007) [16] M. Srivastava, E. Begovic, J. Chapman, N.H. Putnam, U. Hellsten, T. Kawashima, A. Kuo, T. Mitros, A. Salamov, M.L. Carpenter, A.Y. Signorovitch, M.A. Moreno, K. Kamm, J. Grimwood, J. Schmutz, H. Shapiro, I.V. Grigoriev, L.W. Buss, B. Schierwater, S.L. Dellaporta, D.S. Rokhsar, The Trichoplax genome and the nature of placozoans, Nature 454 (2008)

5 260 S.M. Shimeld et al. / Genomics 95 (2010) [17] C.W. Dunn, A. Hejnol, D.Q. Matus, K. Pang, W.E. Browne, S.A. Smith, E. Seaver, G.W. Rouse, M. Obst, G.D. Edgecombe, M.V. Sorensen, S.H. Haddock, A. Schmidt-Rhaesa, A. Okusu, R.M. Kristensen, W.C. Wheeler, M.Q. Martindale, G. Giribet, Broad phylogenomic sampling improves resolution of the animal tree of life, Nature 452 (2008) [18] S.L. Dellaporta, A. Xu, S. Sagasser, W. Jakob, M.A. Moreno, L.W. Buss, B. Schierwater, Mitochondrial genome of Trichoplax adhaerens supports placozoa as the basal lower metazoan phylum, Proc. Natl. Acad. Sci. U. S. A. 103 (2006) [19] C. Larroux, G.N. Luke, P. Koopman, D.S. Rokhsar, S.M. Shimeld, B.M. Degnan, Genesis and expansion of metazoan transcription factor gene classes, Mol. Biol. Evol. 25 (2008) [20] K.J. Peterson, E.A. Sperling, Poriferan ANTP genes: primitively simple or secondarily reduced? Evol. Dev. 9 (2007) [21] E.H. Davidson, M.S. Levine, Properties of developmental gene regulatory networks, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) [22] M. Stauber, H. Jackle, U. Schmidt-Ott, The anterior determinant bicoid of Drosophila is a derived Hox class 3 gene, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) [23] S.L. Pollard, P.W. Holland, Evidence for 14 homeobox gene clusters in human genome ancestry, Curr. Biol. 10 (2000) [24] D.E. Ferrier, C. Minguillon, Evolution of the Hox/ParaHox gene clusters, Int. J. Dev. Biol. 47 (2003) [25] F. Mazet, C.T. Amemiya, S.M. Shimeld, An ancient Fox gene cluster in bilaterian animals, Curr. Biol. 16 (2006) R314 R316. [26] R. Nusse, An ancient cluster of Wnt paralogues, Trends Genet. 17 (2001) 443. [27] C. He, H. Cheng, R. Zhou, GATA family of transcription factors of vertebrates: phylogenetics and chromosomal synteny, J. Biosci. 32 (2007) [28] K.R. Wotton, F. Mazet, S.M. Shimeld, Expression of FoxC, FoxF, FoxL1, and FoxQ1 genes in the dogfish Scyliorhinus canicula defines ancient and derived roles for Fox genes in vertebrate development, Dev. Dyn. 237 (2008)