Evolution of Developmental Regulation

8/13/2019 Evolution of Developmental Regulation

1/26

Evolution of Developmental Regulation

in the Vertebrate FgfD SubfamilyRICHARD JOVELIN1, YI-LIN YAN2, XINJUN HE2, JULIAN CATCHEN2,3,ANGEL AMORES2, CRISTIAN CANESTRO2, HAYATO YOKOI2, ANDJOHN H. POSTLETHWAIT21Center for Ecology and Evolutionary Biology, University of Oregon, Eugene,Oregon2Institute of Neuroscience, University of Oregon, Eugene, Oregon3Department of Computer and Information Science, University of Oregon,Eugene, Oregon

ABSTRACT Fibroblast growth factors (Fgfs) encode small signaling proteins that help regulateembryo patterning. Fgfs fall into seven families, including FgfD. Nonvertebrate chordates have asingleFgfDgene; mammals have three (Fgf8, Fgf17,and Fgf18); and teleosts have six (fgf8a, fgf8b,

fgf17, fgf18a, fgf18b, and fgf24). What are the evolutionary processes that led to the structuralduplication and functional diversification ofFgfDgenes during vertebrate phylogeny? To study thisquestion, we investigated conserved syntenies, patterns of gene expression, and the distribution ofconserved noncoding elements (CNEs) in FgfD genes of stickleback and zebrafish, and comparedthem with data from cephalochordates, urochordates, and mammals. Genomic analysis suggests that

Fgf8, Fgf17, Fgf18, and Fgf24 arose in two rounds of whole genome duplication at the base of thevertebrate radiation; that fgf8 and fgf18 duplications occurred at the base of the teleost radiation;and that Fgf24is an ohnolog that was lost in the mammalian lineage. Expression analysis suggeststhat ancestral subfunctions partitioned between gene duplicates and points to the evolution of novelexpression domains. Analysis of CNEs, at least some of which are candidate regulatory elements,suggests that ancestral CNEs partitioned between gene duplicates. These results help explain theevolutionary pathways by which the developmentally important family of FgfD molecules arose and

the deduced principles that guided FgfD evolution are likely applicable to the evolution ofdevelopmental regulation in many vertebrate multigene families. J. Exp. Zool. (Mol. Dev. Evol.)312B, 2009. r 2009 Wiley-Liss, Inc.

How to cite this article: Jovelin R, Yan Y-L, He X, Catchen J, Amores A, Canestro C, Yokoi H,Postlethwait JH. 2009. Evolution of developmental regulation in the vertebrate FgfDsubfamily. J. Exp. Zool. (Mol. Dev. Evol.) 312B:[page range].

The development of body form and the differ-entiation of cells depend on a variety of processes,including cell adhesion accomplished by celladhesion proteins (e.g. cadherins, integrins, selec-tins), cell communication performed by extracel-lular signaling proteins (e.g. Notch, Hedgehog,

Wnt, Tgf-b, Fgf), and differential gene expressioncontrolled by transcription factors (e.g. homeobox,bHLH, C2H2-zinc finger, bZIP). Families ofstructurally related genes encode these proteinsand constitute an ancestral toolkit for the con-struction of animal embryos. Recent genomicinvestigations reveal a surprisingly complete gene

set already present in the last common ancestor ofall metazoans (Technau et al., 2005; Guder et al.,2006; Nichols et al., 2006; Kasbauer et al., 2007;Matus et al., 2007), but this ancestral toolkit

Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jez.b.21307

Received 11 February 2009; Revised 4 May 2009; Accepted 27 May2009

Richard Jovelin and Yi-Lin Yan contributed equally to the work.Grant sponsor: The National Center for Research Resources; Grant

number: 5R01RR020833; Grant sponsor: National Institutes ofHealth; Grant number: P01 HD22486.

Correspondence to: John H. Postlethwait, Institute ofNeuroscience, 1254 University of Oregon, Eugene, OR 97403.E-mail: [email protected]

2009 WILEY-LISS, INC.

JOURNAL OF EXPERIMENTAL ZOOLOGY (MOL DEV EVOL) 312B (2009)


2/26

diversified by gene duplication and gene deletionin the great lineages of animal life.

Stem chordates appear to have inherited asubset of toolkit genes present in stem metazoans,and these genes amplified into several members of

each subfamily, many in two rounds of wholegenome duplication (WGD) (R1 and R2) thatappear to have occurred at the base of thevertebrate radiation before the divergence ofchondrichthyans and osteichthyans about 530million years ago (Kumar and Hedges, 98;Kortschak et al., 2001; Ornitz and Itoh, 2001;Lundin et al., 2003; Dehal and Boore, 2005;Garcia-Fernandez, 2005; Bourlat et al., 2006;Delsuc et al., 2006; Guder et al., 2006; Jacob andLum, 2007; Kitisin et al., 2007; Sundstrom et al.,2008; Yu et al., 2008; Lynch and Wagner, 2009).Genes derived from a WGD event are called

ohnologs to emphasize their special characterswith respect to genes duplicated by other mechan-isms, such as tandem gene duplication, unequalcrossing-over, or retrotransposition (Wolfe, 2000;Postlethwait, 2007). An example is the family ofHox clusters, which is initially amplified bytandem duplication into a gene cluster that thenreplicated in vertebrates into four clusters thatappear to be related to each other in the relation-ship ((A,B)(C,D)), as would be expected by the 2Rmodel (Meulemans and Bronner-Fraser, 2007;

Amemiya et al., 2008).

The widespread occurrence of multigene fa-milies and subfamilies raises two major questionsregarding the evolution of developmental mechan-isms: What are the processes by which individualmembers of gene families evolve their specializedroles in development? And, to what extent doesthe origin of individual family members by geneduplication contribute to morphological innova-tion and lineage divergence?

A classic multigene family essential for animaldevelopment, the fibroblast growth factor (Fgf)gene family, encodes small signaling proteinsdeployed in many aspects of embryo patterning(Ornitz and Itoh, 2001). Human and mousegenomes have 22 Fgf genes grouped into sevensubfamilies, each containing two to four members(FgfA (1/2), FgfB (3/7/10/22), FgfC (4/5/6), FgfD(8/17/18), FgfE (9/16/20), FgfF(11/12/13/14), andFgfG (19/21/23)) (Itoh and Ornitz, 2004; Popoviciet al., 2005). The Fgf gene complement of theurochordate Ciona intestinalis, whose lineagediverged from the vertebrate lineage before theR2 events (Oda et al., 2002; Delsuc et al., 2006;

Vienne and Pontarotti, 2006; Wada et al., 2006;

Delsuc et al., 2008), helps elucidate the pattern ofdiversification of mammalian Fgf subfamilies(Satou et al., 2002). Based on strong evidencefrom sequence similarities, two of the sixC. intestinalis Fgf genes are orthologs of FgfD

and FgfF subfamilies, respectively, and moderatesupport, based on conserved protein signaturemotifs and conserved genomic syntenies, suggeststhat three other genes are orthologs of the FgfB,FgfC, and FgfE subfamilies (Satou et al., 2002;Popovici et al., 2005). It is not clear whetherCi-FgfL is orthologous to the FgfA subfamily orwhether it is unique to C. intestinalis.

Consider the origin of theFgfB, FgfC, andFgfGsubfamilies. A representative of FgfG appears tobe missing from Ciona and thus either arose bygene duplication in vertebrates or was secondarilylost in the urochordate lineage. Because FGF3,

FGF4, and FGF19 of the FgfB, C, and Gsubfamilies are contiguous in a cluster in humanchromosome 11q13.3, these three genes likelyarose by tandem duplication. Those tandemduplications must have occurred before R1 be-causeFGF6 and FGF23(FgfCand FgfGsubfami-lies) are nearest neighbors in chromosome12p13.32 (the otherFgfBmember was apparentlylost from this chromosome), and FGF21 andFGF22 (FgfB and FgfG) are linked on humanchromosome 19 (Hsa19) (the FgfC member wasapparently deleted from this chromosome). This

genomic pattern suggests the model that anancient FgfB/C/G gene tandemly duplicated toFgfB and FgfC/G before the divergence ofurochordates and vertebrates and that subse-quently, FgfC/G duplicated to FgfC and FgfGbefore R1; after R2, chromosome rearrangementsand gene loss led to the current human genomicarrangement.

Unexpectedly, the cnidarian Nematostella vec-tensishas at least 13 Fgfgenes, twice as many asCionaand five or six times more thanDrosophilamelanogaster and Caenorhabditis elegans (Matuset al., 2007). Phylogenetic analysis suggests thatnine Nv-Fgfs may be species-specific with no clearrelationship to vertebrate Fgfs, whereas fourNv-Fgfs fall within the FgfD subfamily (Matuset al., 2007).

In mammals, the FgfD subfamily includes Fgf8,Fgf17, and Fgf18 (Ornitz and Itoh, 2001) and inzebrafish, fgf24 (Itoh and Ornitz, 2004; Popoviciet al., 2005). The vertebrate FgfD phylogenysuggests either the hypothesis that Fgf24 waslost in the tetrapod lineage (Jovelin et al., 2007), orthe hypothesis that fgf24 was an innovation in

R. JOVELIN ET AL.2

J. Exp. Zool. (Mol. Dev. Evol.)


3/26

teleosts. FgfD members play critical roles inmesoderm and neuronal induction across largephylogenetic distances. For instanceFgf8, Fgf17,and Fgf18 are expressed at the midbrainhind-brain border (MHB) (Sato et al., 2004). FGF18 is

required for skeletal development (Ohuchi et al.,2000; Liu et al., 2002; Ohbayashi et al., 2002), andFgf8 plays an important role in MHB and limbbud development (Lewandoski et al., 2000; Moonand Capecchi, 2000; Sun et al., 2002; Boulet et al.,2004). Remarkably, FGF signaling in cnidariansmay be important during gastrulation and neuralinduction as well (Matus et al., 2007).

Until recently, the comparison of Fgf genecontent between mammals and zebrafish did notreflect the R3 hypothesis, a WGD occurring in theteleost lineage after the teleosts diverged fromother ray-fin fish (Sidow, 96; Amores et al., 98;

Postlethwait et al., 98; Wittbrodt et al., 98; Meyerand Schartl, 99; Christoffels et al., 2004; Hoegget al., 2004; Naruse et al., 2004; Meyer and Van dePeer, 2005; Crow et al., 2006). Our previousphylogenetic analysis of the vertebrate FgfDsubfamily showed that fgf8 and fgf18 wereduplicated early in the teleost lineage and pro-vided evidence for a diversification compatiblewith the ray-fin fish genome duplication (Jovelinet al., 2007). Duplicates of human fgf6, fgf10,andfgf20 genes have also recently been identified inzebrafish (Itoh and Konishi, 2007).

Because of the central role of FGF signaling indevelopment, mutations in Fgf genes and theirreceptors cause hereditary human diseases (Itoh,2007). An understanding of the precise evolution-ary relationships among Fgf genes and insightsinto the mechanisms of their functional diversifi-cation are required to fully connect teleostresearch to human biology. For instance, inmouse, Fgf8 is required for heart development(Abu-Issa et al., 2002; Frank et al., 2002; Ilaganet al., 2006). In zebrafish, fgf8ais required for theexpression of cardiac genes (Reifers et al., 2000b),butfgf8bis not expressed in the heart; conversely,in stickleback, fgf8bbut not fgf8ais expressed inthe heart (Jovelin et al., 2007). After the ray-finfish genome duplication, independent evolution ofregulatory elements in lineages leading to zebra-fish and stickleback led to the differential parti-tioning offgf8subfunctions in heart developmentbetween paralogs so that orthologous genes in thetwo species have come to express different func-tions (Jovelin et al., 2007).

In this study, we investigated how the develop-mental roles ofFgfD subfamily members evolved

as these genes originated from R1 to R3. First, weanalyzed the conservation of syntenies amongchordate FgfD genes to discern evolutionaryrelationships ambiguous in phylogenies (Satouet al., 2002; Itoh and Ornitz, 2004; Popovici

et al., 2005; Jovelin et al., 2007). Second, we usedin situ hybridization in stickleback and zebrafish,whose lineages diverged early in teleost phylo-geny, about 300 million years ago (Hedges, 2002),to examine how functions of duplicate genesevolve in separate lineages. Third, we analyzedthe conservation of noncoding regions in thevicinity of FgfD genes in the context of genephylogeny and gene expression to learn howdivergence in regulatory elements in the face ofdifferential gene loss could affect functionaldivergence of surviving ohnologs.

Results confirmed the presence of anFgf24gene

at the base of gnathostome vertebrates, thepartitioning of functions between Fgf24 andFgf8, and revealed the origin of teleost fgfDparalogs. Expression analyses identified sharedancestral functions and newly evolved functions invarious members of theFgfDgroup, and explora-tion of conserved noncoding elements (CNEs) gavestrong support for a pattern of functional diver-gence through subfunctionalization. The generalpattern of gene duplication and functional diver-gence that this study reveals is likely to apply tomany other vertebrate multiple gene families.

MATERIALS AND METHODS

Conserved synteny

Genome sequences were investigated usingthe Ensembl databases (http://www.ensembl.org/index.html) for zebrafish Zv7, human NCBI 35,mouse NCBI m36, and C. intestinalisversion 2.0.Orthologies were determined by best reciprocalBLAST hit analysis (Wall et al., 2003) automatedin a genome-wide scale in our Synteny Databasegenomic analysis software (Catchen et al., 2009).From the elephant shark genomic database weisolated the following FgfD gene fragments:Fgf8: exon2 AAVX01442442;Fgf17: no fragmentsidentified; Fgf18: exon2 AAVX01015792 and

AAVX01460739, exon3 AAVX01088747, and exon4AAVX01419960;Fgf24: exon2 AAVX01560519 andAAVX01583584, exon3 AAVX01466056 andAAVX01521786, and exon4 AAVX01108229. Theabsence ofFgf17is not significant owing to the low(1.4) covera ge of the shark genome (Venkateshet al., 2007).

FgfDGENES IN EVOLUTION AND DEVELOPMENT 3



4/26


5/26

suggesting that these may be paralogons from theR1 and R2 genome duplication events.

To determine whether the human orthologs ofgenes neighboringC. intestinalis Fgf8/17/18wereclustered on human chromosomes or widelydistributed, we used circle plots. Figure 1B showsthat human orthologs of the neighbors of theC. intestinalis Fgf8/17/18gene cluster around thehuman Fgf8, Fgf17,and Fgf18 genes rather thanbeing splattered over the full extent of eachhuman chromosome. We conclude that syntenieshave tended to be conserved in this region of thegenome, which would be expected if local inver-sions were more frequent than translocationsduring chordate evolution. Furthermore, thisresult is as predicted from the hypothesis thatstem olfactores (urochordates1vertebrates) had achromosome segment that duplicated twice, givingrise to paralogons on Hsa5, 8, and 10 containing

FGF18, FGF17, andFGF8,and another paralogoncontaining parts of Hsa4 or Hsa2p that today lacksan FgfDgene.

The human genome has four FgfD-relatedparalogons

The hypothesis that FGFD genes arose in tworounds of chromosome duplication predicts that adot plot that localizes paralogs of genes surround-ing FGF8 on Hsa10 should identify four para-logons, including parts of Hsa5 and Hsa8, and oneother chromosome segment. To test this predic-tion, we investigated the chromosome locations ofthe paralogs of all genes in the 10 Mb intervalsurroundingFGF8, from 98 to 108 Mb on Hsa10.Figure 1C shows dots for genes from the region ofHsa10 surrounding FGF8 and their paralogs onother chromosomes displayed by red pluses in a

Cin5q

Hsa2

Hsa5

Hsa8

Hsa10

FGF18

FGF17

FGF8

(FGF24)

B

Human Chromosomes

108Mb

Hsa10Paralogs

C

98Mb22212019181716151413121110987654321YX

FGF8

FGF17

FGF18

MVP Hsa16

RIOK3 Hsa18

FOLR1 Hsa11 B3GAT1 Hsa11SCP2 Hsa1

KIAA0368 Hsa9

HD Hsa4FMOD Hsa1; ECM2 Hsa9; DCN Hsa12

3.50 Mb 3.55 Mb 3.60 Mb 3.65 Mb 3.70 Mb 3.75 Mb 3.80 Mb 3.85 Mb 3.90 Mb

TSGA14 Hsa7

POLE4 Hsa2

HMGCS1 Hsa5TMEM20 Hsa10

SLC7A2 Hsa8

FBXW4 Hsa10FGF8 Hsa10; FGF17 Hsa8; FGF18 Hsa5

PPAPDC1B Hsa8; PPAPDC1A Hsa10 FBXW11 Hsa5; BTRC Hsa10

CYP2U1 Hsa4; CYP2E1 Hsa10; CYP2C18 Hsa10

ADD2 Hsa2; ADD1 Hsa4; ADD3 Hsa10FBXW11 Hsa5

MCCC2 Hsa5

RNPEPL1 Hsa2; RNPEP Hsa1

TMEM20 Hsa10

Cin5q

NPM3 Hsa10; NPM2 Hsa8; NPM1 Hsa5

A

Fig. 1. The chromosome region containingFgf8/17/18 in the urochordateCiona intestinalisshares conserved synteny withFGFD-containing regions of the human genome. (A) A portion ofC. intestinalischromosome 5q (rectangle) is enlarged to showthe positions of predicted C. intestinalis genes listed with the symbols for their human orthologs, often several, and theirlocations in the human genome. Black font represents Ciona genes with human orthologs near FGFD genes and gray fontrepresents other locations in the human genome. The black rectangle marks a trio of genes preserved in order and orientationfor 800 million years since the divergence of urochordate and vertebrate lineages. (B) A circle plot of the Fgf8/17/18region of theC. intestinalis genome with arcs linking to human orthologs on chromosomes around the circumference of the circledemonstrating the limited position of the orthologs on each human chromosome. ( C) A dot plot of paralogs for genes in a 10 Mbwindow surroundingFGF8on Hsa10, with the location ofFGF8, FGF17,and FGF18circled. Paralogs to Hsa10 genes are shownon their chromosome directly above the location of the gene on Hsa8. Paralogs lie mainly on Hsa2, 4, 5, 8, and 10.




6/26

column above or below the Hsa10 gene. Resultsshowed that chromosomes 2, 4, 5, and 8 have 13,13, 8, and 9 paralogs of the genes indicated by graycircles on Hsa10, but all other chromosomes havesix or fewer paralogs, with an average of 1.5

paralogs per chromosome. These results supportthe conclusion from the analysis of conservedsyntenies shared between C. intestinalis andhuman genomes that parts of Hsa2p and/or Hsa4are members of theFgfD-containing paralogon.

Teleost FgfD genes reflect R3 and anohnolog gone missing

Having identified paralogous chromosome seg-ments in the human genome that have propertiesexpected for ohnologons arising from the R1and R2 WGD events, we turned to understanding

the relationship of the six teleost FgfD genes tothe one Ciona and three humanFgfDgenes.

Isfgf24 an ohnolog gone missing from thehuman genome?

Recognizing that the primary fate of duplicategenes is nonfunctionalization (pseudogenation) of

one of the two paralogs (Bershtein and Tawfik,2008), we explored the hypothesis that thepredicted fourth R2-derived co-ortholog of theprototypical FgfD gene is Fgf24 and that thisgene was preserved in the teleost lineage but was

lost in the human lineage, as suggested byphylogenetic analysis (Jovelin et al., 2007). Analternative hypothesis to be ruled out is thatfgf24did not exist in the last common ancestor ofteleosts and humans but was a teleost innovation.

The R2 hypothesis makes three predictions.First, the teleost fgf24 gene should be embeddedin a chromosome segment that is conserved withthe human genome; second, the conserved regionshould have paralogs near other human FGFDgenes; and third, this segment should showconserved synteny with theCiona Fgf8/17/18geneneighborhood shown in Figure 1. Here we test

these predictions by examining conserved synte-nies around zebrafish fgf24, starting at the firstgene to its right (Fig. 2).

The gene lying immediately to the right offgf24in the zebrafish and stickleback genomes (Q566X7and ENSGACT00000022099, respectively) encodeproteins that form a well-supported sister clade to

Dre14

Hsa2

74.56Mb74.64Mb

Hsa2

AUP1PCGF1

TLX2LBX2 LOXL3

bM57.1bM22.1

Dre14

aup1

tlx2npm4fgf24

loxl3lbx2

3.12Mb 3.57Mb

lbx2

tpm4fgf24

pcgf1

pcgf1

GacIV

A

B

C GacIV

tlx2

Fig. 2. Conserved syntenies forfgf24. (A) A portion of Hsa2p with orthologs to neighbors of teleostfgf24genes. (B) A portionof zebrafish (Danio rerio) linkage group 14 (Dre14) containing fgf24. (C) A portion of stickleback (Gasterosteus aculeatus)linkage group IV (GacIV) containingfgf24. This evidence is as expected if the human orthologs offgf24and npm4were deletedfrom the human genome. Lines connect orthologs. Positions on chromosomes given in megabases (Mb).

R. JOVELIN ET AL.6



7/26

tetrapod NPM1 and the zebrafish gene npm1; wewill call this gene npm4 for now (Fig. 2B andSuppl. Fig. S1). Thus, npm4 is either an npm1duplicate (e.g. npm1b), which would make fgf24aduplicate of fgf18, or npm4 is a gene, like its

neighbor fgf24, without a human ortholog. Thehuman genome has three NPM genes: NPM1(5q35), which is adjacent to FGF18 (8p21.3);NPM2, which is the nearest neighbor of FGF17,andNPM3, which is the nearest neighbor ofFGF8(10q24) (Figs. 3B, 4C, 5A). Thus, an FgfDgene andan Npm gene are neighbors for all extant humanFgfD genes, as would be expected if this was theancestral condition before R1, and neither npm4nor its neighbor fgf24 have unique humanorthologs.

The next full gene to the right of fgf24(ENSDARG00000011273) is annotated as tlx3a

but is actually an ortholog ofTLX2(Wotton et al.,2008). Human and zebrafish genomes both have

three TLX family genes (Andermann andWeinberg, 2001; Langenau et al., 2002): TLX1 isfive genes distant from FGF8 and tlx1 is threegenes distant fromfgf8a. TLX3is the second genefrom FGF18, whereas tlx3b is 2.8 Mb from fgf18l.

TLX2 is in Hsa2p13.1-p12 without a nearbyFGFD-family gene, but tlx2 (NP_705937) is thesecond gene from fgf24, which favors the modelthat fgf24 is an ohnolog gone missing from thehuman genome.

The next gene to the right, ENSDARG-00000063176 from Ensembl release 45, is notannotated in Zv7 despite the continued presenceof its nucleotide sequence in the position indicatedin Figure 2B. ENSDARG00000063176 has asa best human BLAST hit KAZALD1 locatedat 10q24.31 seven genes from FGF8, b u t i n a phylogenetic tree, it forms, with its ortholog in

the pufferfish Tetraodon nigroviridis, an out-group to the tetrapod1zebrafish KAZALD1 clade

0.52Mb0.62Mb

Dre14

fgf18a

169.68Mb 171.58Mb

Hsa5

3XLT1PINCK FGF18

FBXW11STK10

16.71Mb

Dre10

ENSDARG00000062737

zgc:63728

zgc:101531npm1

fgf18b

tlx3b

NPM1

fbxw11

GacIVA

B

C

D

bM04.01bM07.9

fgf18LOC567566

777600001COL1hdcp

fbxw11etf1pcdh1

pcdh1pcdh1

pcdh1pcdh1

pcdh1

LOC567566etf1 pcdh1

pcdh1

Dre10

Hsa5

Dre14

GacIVE.

PCDH1FGF18

18.04Mb

Fig. 3. Conserved syntenies for Fgf18. (A) A portion of stickleback chromosome GacIV containingfgf18. (B) A portion ofzebrafish chromosome Dre10 containingfgf18b. (C) A portion of human chromosome Hsa5 containingFGF18. (D) A portion ofDre14 containing fgf18a. (E) Locations of each region shown related to the whole chromosome. Black genes have orthologs inthese sections, gray genes do not.




8/26

(Suppl. Fig. S2). This phylogenetic groupingsuggests that, like npm4 and fgf24, EN-SDARG00000063176 is a paralog without a humanortholog and should be called kazald2 (Wottonet al., 2008).

The two neighbors to the immediate left ofzebrafish fgf24 (lbx2 and pcgf1, Fig. 2B) haveorthologs adjacent on Hsa2p (Fig. 2A). In human,LBX2, PCGF1, andTLX2are adjacent, but in thezebrafish genome fgf24 and npm4 lie betweenpcgf1andtlx2(Fig. 2A and B). Moving to the rightof TLX2, two of the next four genes in human,AUP1andLOXL3,have orthologs also to the rightof fgf24 in zebrafish (Fig. 2A and B). Thus, thezebrafish and human share most of the genes in

this region of Hsa2p/Dre14, but orthologs offgf24andnpm4are specifically excised from the humangenome. Stickleback linkage group IV (GacIV),which has five genes in a row orthologous to thezebrafish segment in Dre14 just discussed, and anortholog ofaup1syntenic on GacIV, but about halfa chromosome away (Fig. 2C). These resultssuggest that the five genes (lbx2, pcgf1, fgf24,npm4, tlx2) were neighbors before the divergenceof zebrafish and stickleback lineages and thatinversions in the stickleback lineage removedaup1from its former neighbors.

The most parsimonious explanation of these datais that fgf24and npm4 were adjacent in the FgfDparalogon in the last common ancestor of zebrafish

zgc:136955

29.31Mb

Dre1

poll

lbx1bzgc:153412

bM82.4bM27.3

GacVI

TLX1C6orf57

OGFRL1

ENSGACG00000003828fgf8a

2TAG3B4WXBF

102.95Mb 105.13Mb

Hsa10

LOC100007451

NP_001082838.1nt5c2

cyp17a1 si:ch211-103o12.1

fgf8afbxw4

zgc:114169

A2ATX7ldb4

C10orf26AS3MT

CNNM2

INA

LBX1FGF8

FBXW4LDB1 CYP17A1 NT5C2

PCGF6

RP11-529I10.4

POLL

Dre13

NP_739571.1si:busm1-265n4.4

ogflr11

slc2a5 21.33Mbzgc:91925zgc:103568A2ATX5

HPS6

fgf8b

zgc:111946

zgc:136933 gefiltin

zgc:15365129.69Mb

bM49.01bM27.01

GacIX

ENSGACG00000018247NP_056263.1

LBX1CENPK

INAfgf8bPOLL

C10orf26AS3MT

cenpk

NPM3

A

B

C

D

E

20.255Mb

Fig. 4. Conserved syntenies for Fgf8. (A) A portion of stickleback chromosome GacVI containingfgf8a. (B) A portion ofzebrafish chromosome Dre13 containingfgf8b. (C) A portion of human chromosome Hsa10 containingFGF8. (D) A portion ofDre1 containingfgf8b. (E) A portion of GacIX containingfgf8b. Black genes have orthologs in these sections, gray genes do not.

R. JOVELIN ET AL.8



9/26

and human, and that they were lost in the humanlineage. Note that the orientation of TLX2/tlx2relative toLBX2/lbx2and PCGF1/pcgf1is oppositein human and zebrafish, suggesting that a localinversion stirred the sequences, either in thehuman or zebrafish lineage. The human FGF18region can serve as an outgroup to order the vector

of evolutionary change: TLX3, NPM1, and FGF18are all in the same orientation, suggesting that thedeletion of the tetrapodFgf24and Npm4ohnologswas associated with a small local inversion.

When did the duplication event that producedFgf24occur? BLAST searches of the trace files ofthe genome sequencing project of the elephantshark Callorhinchus milli revealed that chon-drichthyes have at least an Fgf8 gene (exon 2

AAVX01442442) and both an Fgf18 gene (exon 3AAVX01088747) and an Fgf24 gene (exon 3AAVX01521786), as shown by exclusive aminoacid positions in an alignment of exon-3 of FgfDgenes from various vertebrates (Suppl. Fig. S3A).Because the shark genome, which diverged fromthe bony vertebrates before the divergence ofteleost and tetrapod lineages, has an Fgf24 gene,this gene must have already existed in the lastcommon ancestor of all jawed vertebrates. Theinclusion of shark sequences forFgf18 and Fgf24in a phylogenetic tree of exon 3 results in thetopology ((Fgf8,Fgf17)(Fgf18,Fgf24)) as expectedfrom their origin in the R1 and R2 genomeduplications (Suppl. Fig. S3B).

Altogether, the best interpretation of thesefindings is that the vertebrate FgfD subfamilyoriginated from a single gene present in theancestor of vertebrates and urochordates by tworounds of genome duplication. These R1 and R2events lead to four Fgfgenes in the last commonancestor of teleosts and tetrapods but the ortholog

of fgf24 was subsequently lost in the tetrapodlineage.

fgf18aandfgf18b are duplicates from theteleost genome duplication

Phylogenetic analysis shows that zebrafish hastwo genes (fgf18a (NM_001013264) and fgf18b(also called fgf18l; NM_001012379)) in a cladewith FGF18 as the most closely related humangene. Zebrafish Fgf18b is sister-taxon to a teleostFgf18a clade (Itoh and Konishi, 2007; Jovelinet al., 2007; Kikuta et al., 2007) as would beexpected from gene duplication after the diver-gence of ray-fin and lobe-fin fish (Itoh and Konishi,2007; Jovelin et al., 2007; Kikuta et al., 2007).Four other sequenced teleost genomes have anortholog of fgf18a but lack an annotated fgf18band BLAST searches failed to identify an orthologof fgf18b in genome databases for stickleback,medaka, Tetraodon, or fugu, all of which arepercomorph fish. This situation could have ariseniffgf18aand fgf18barose by the R3 duplication instem teleosts and fgf18bwas lost in the lineage of

21.93Mb 25.40Mb

Hsa8

bM31.74bM46.54

Dre8

FGF17CHMP7

ENTPD4

KCTD9

zgc:92668

zgc:113115lgi3

chmp7

LGI3

nr5a1a

zgc:123194

LOC571805

LOC100004958cyr61

fgf17b

0.30Mb 0.78Mb

fgf17ENSGACG00000003433

ENSGACG00000003446

ENSGACG00000003456

1A5RN17430000000GCAGSNE

GacXIII

NPM2

A

B

C

Fig. 5. Conserved syntenies forFgf17. (A) A portion of Hsa8 containingFGF17. (B) A portion of Dre8 containingfgf17.(C) A portion of GacXIII containingfgf17. Black genes have orthologs in these sections, gray genes do not.




10/26

the Percomorpha but was retained in the lineageof the Ostariophysi (including zebrafish). Analternative hypothesis is that fgf18b arose bytandem duplication in the zebrafish lineage afterit diverged from the percomorphs.

Conserved syntenies can help distinguish the R3and tandem duplication hypotheses for the originof fgf18 paralogs. The fgf18a gene of zebrafish ison Dre14 near fgf24 along with fbxw11, theortholog of which is separated from FGF18 by asingle hypothetical gene (Fig. 3C and D), thussupporting the orthology of zebrafish fgf18aand human FGF18. Note that the C. intestinalisgenome has two orthologs of FBXW11, onelocated 10 genes distant and one 13 genes awayfrom Fgf8/17/18 (NP_001027626 and ENSCINT-00000018903) (Fig. 1A); this is another example ofan ancientFgfD-region synteny conserved for 800

million years. A cluster of pcdh1 genes flankingfgf18 (not shown) are, as a group, co-orthologousto PCDH1 protocadherin genes in Hsa5q31, abit distant but syntenic to FGF18 in Hsa5q35(Fig. 3CE). These gene clusters were apparentlytandemly duplicated at least partly independentlyin the zebrafish and human lineages. Together,these conserved syntenies support orthology offgf18aand FGF18.

The orthologs of several genes located nearzebrafish fgf18b on zebrafish LG10 are adjacentto or near FGF18 on Hsa5q (Fig. 3B and C),

consistent with the orthology offgf18band FGF18as shown in the phylogeny. The nearest neighborto fgf18b (Fig. 3B) is an ortholog of FBXW11,which is separated from FGF18 by a single smallhypothetical gene (Fig. 3C). Like fgf18a andFGF18neighborhoods, the fgf18bregion of zebra-fish and stickleback both have a series of tandemlyduplicated pcdh genes. The region surroundingfgf18 in stickleback has at least five neighborswith orthologies to genes in the region surround-ing fgf18b in zebrafish, a result expected if thetwo regions are orthologous (Fig. 3A and B). Thisis in conflict with the phylogeny, which shows thatFgf18 in stickleback (ENSGACT00000023719) ismore closely related to Fgf18a than it is to Fgf18b(Jovelin et al., 2007). If the conserved synteniescorrectly suggest orthologies, then the zebrafishFgf18b sequence is evolving more rapidly thanthe Fgf18a sequence. Alternatively, gene lossesmay account for the conflict. If orthologs neigh-boring stickleback fgf18 were lost near fgf18abut not near fgf18b, then the reciprocal bestblast method for identifying orthologies couldfind only genes near fgf18b and call them as

orthologs given the loss of these genes fromthe neighborhood of fgf18a. Both explanations,however, are consistent with the origin offgf18a and fgf18b in the R3 event in teleostevolution.

fgf8duplicates arose in R3

Zebrafish has one uncontested ortholog ofFGF8(Brand et al., 96; Reifers et al., 98), and a secondgene (fgf17a) that was initially assigned as a co-ortholog of tetrapod Fgf17 (Reifers et al., 2000a;Cao et al., 2004), but that was later shown to be afgf8 duplicate and is now called fgf8b (Itoh andKonishi, 2007; Jovelin et al., 2007; Kikuta et al.,2007). A detailed analysis of conserved synteniesin human, zebrafish, and stickleback supports this

conclusion. Orthologs of six genes surroundinghuman FGF8 in a 2.16 Mb region on chromosomeHsa10q (Fig. 4C) are located in a 0.4 Mb region onzebrafish chromosome Dre1 containingfgf8b(Fig.4B). The orthologous portion of stickleback link-age group GacIX containsfgf8band eight genes inthe corresponding region of the human genome(Fig. 4A). The orthologs of a somewhat differentset of human genes flanking FGF8 lie nearfgf8a in zebrafish Dre13 and stickleback LG VI(Fig. 4D and E). We conclude that zebrafishfgf8 (NM_131281) and the so-called fgf17a(NM_182856) are co-orthologs of human FGF8

and support the conclusion that they should benamed fgf8a and fgf8b (Itoh and Konishi, 2007;

Jovelin et al., 2007; Kikuta et al., 2007). Figure 4shows that in addition to FGF8 co-orthologs,zebrafish and stickleback have co-orthologs ofLBX1 (Q5CZV7 and si:busm1-265n4.1) andC10orf26 (zgc:111946 and Q1ECU4) in thesegenomic regions, indicating that that theseduplicated loci also originated in the R3 genomeduplication.

Teleosts have a single copy offgf17Sequenced teleost genomes have a single ortho-

log of FGF17. Several genes in the immediateneighborhood the gene originally called fgf17bonzebrafish Dre8 (Cao et al., 2004) are orthologs ofgenes that are near FGF17 in Hsa8p21 (Fig. 5Aand B). These two segments are largely conservedbetween zebrafish and stickleback (Fig. 5B and C).

We conclude that fgf17b should be called simplyfgf17(NM_182856) and it is the zebrafish orthologofFGF17.

R. JOVELIN ET AL.10



11/26

Expression patterns of FgfD family genes

With this understanding of the evolutionaryrelationships of teleost and tetrapod FgfD genes,we wondered how paralog functions evolved as

judged by gene expression patterns.

Mid-segmentation stages

fgf8 paralogs: At mid-segmentation stages(40 hpf stickleback and 16 hpf zebrafish) (Kimmelet al., 95), the expression of fgf8a and fgf8b aresimilar in zebrafish and stickleback (Reifers et al.,98, 2000a; Draper et al., 2003; Jovelin et al.,2007), with strong expression of both genes inboth species in MBH and somites (Fig. 6AH). In

addition, fgf8a is expressed in the dorsal dience-phalon and the tailbud in both species. In theheart, stickleback and zebrafish express differentfgf8paralogs (Jovelin et al., 2007).

fgf17: At mid-segmentation in stickleback and

zebrafish,fgf17is expressed weakly in the centralnervous system (CNS), and more strongly andspecifically around the tail bud and in thesegmental plate and segmented somites (Fig. 6IL)(Cao et al., 2004; Hamade et al., 2006). The majordifference between sfgf17 (stickleback fgf17)and zfgf17 (zebrafish fgf17) is that the zebrafishgene is expressed in the dorsal diencephalon(Fig. 6K). For both species, fgf8a and fgf8bare both expressed in the MHB, but fgf17 is not,

Gac 40 hpf Dre 16 hpfA B C D

E F G H

I J K L

M N O P

Q R

S T U V

a8fgfa8fgf

b8fgfb8fgf

71fgf71fgf

a81fgf81fgf

fgf18b

42fgf42fgf

mhb mhb

dd dd

r r

ss

tbtb

bhmbhm

mhb

mhb

mhb

s s

ss

s

s

tb tb

umum

ov

ovov

dd

dddd

rr

r

rr

r r

h h

Fig. 6. Expression of stickleback and zebrafishFgfDgenes at midsegmentation stages. A, B, E, F, I, J, M, N, S, T, 40hpfstickleback embryos. C, D, G, H, K, L, O, P, Q, R, U, V, 16 hpf zebrafish embryos. AD, fgf8a. EH, fgf8b. IL, fgf17. MP,fgf18a. Q, R,fgf18b. SV,fgf24. dd, dorsal diencephalon; h, heart; mhb, midbrainhindbrain boundary; ov, otic vesicle; r, retina;s, somite; tb, tail bud; um, unsegmented mesenchyme.




12/26

and all three genes are expressed in the somites.In the tailbud,fgf8ain both species is expressed inthe tailbud itself; fgf8b is not expressed in thetailbud region; and fgf17 is expressed in thesegmental mesoderm that surrounds the tailbud

in a pattern complementary to that offgf8a. Bothzfgf17 and zfgf8a are expressed in the dorsaldiencephalon.

fgf18: The expression of zfgf18a and zfgf18bhave not been previously described. At mid-segmentation, the single copy of fgf18 in stickle-back is expressed weakly in the MHB and in theheart primordium, but not in the somites (Fig. 6Mand N). In zebrafish, zfgf18b, but not zfgf18a, isexpressed in the MHB like zfgf8a, and recipro-cally, zfgf18a, but not zfgf18b,is expressed in theheart like zfgf8b (Fig. 6OR). Although sfgf18 isexpressed only weakly if at all in the somites

(Fig. 6N), zfgf18aand zfgf18bare both expressedin the somites, but zfgf18a is expressed broadly insomites like zfgf8b(Fig. 6H and P), but zfgf18bisexpressed in just a part of each somite like zfgf8a(Fig. 6D and R).

fgf24: In mid-segmentation stage sticklebackembryos, sfgf24 is expressed in the dorsal dience-phalon, MBH, and tailbud, but not in the somites,like sfgf18 (Fig. 6S and T). In zebrafish as instickleback, fgf24 is expressed in the dorsaldiencephalon like fgf8a (Draper et al., 2003). Incontrast to sfgf24, zfgf24 is not expressed in the

MHB (Fig. 6U and V). In both stickleback andzebrafish, fgf24 is expressed in the unsegmentedpresomitic mesoderm.

Long pec stage

fgf8:At the long pec stage (120 hpf sticklebackand 48 hpf zebrafish) (Kimmel et al., 95), theexpression patterns of both fgf8 genes in bothspecies occupy several of the earlier domains, butin addition, in both species, new domainsappear forfgf8a(but not fgf8b) in the telencepha-lon, olfactory epithelium, pharyngeal arches,(Meulemans and Bronner-Fraser, 2007), andapical epidermal ridge (AER) of the pectoral finbud (Fig. 7AD).

fgf17: At the long pec stage, stickleback sfgf17isexpressed in the retina (Fig. 7E) in a pattern thatoverlaps in time and space expression of sfgf8ainthe eye, (Fig. 7A) and is stronger than sfgf8bexpression in the eye at the same stage (Fig. 7C).In addition, sfgf17and zfgf17are expressed in theotic vesicle (Fig. 7E and F). sfgf17was only weaklyexpressed in the pharyngeal arches compared with

zebrafish (Fig. 7E and F). No expression of sfgf17was apparent in the MHB or in the pectoral finbud. In zebrafish, zfgf17 is also expressed in theeye, but broadly in several layers. In mouse, theexpression patterns of Fgf17 and Fgf8 largely

overlap, butFgf17is expressed in a broader regionof the frontal cortex and cerebellum of the brainand is required for dorsal frontal cortex andcerebellar development (Xu et al., 2000; Cholfinand Rubenstein, 2007; Dominguez and Rakic,2008). The regenerating adult zebrafish heart alsoexpressesfgf17(Lepilina et al., 2006).

fgf18: To the expression domains present at mid-segmentation stages (Fig. 6MR), at the long pecstage, stickleback sfgf18and zebrafish zfgf18aaddexpression in the retina, olfactory epithelium, oticvesicle, and pharyngeal arches 37. In contrast,

Gac 5 dpf Dre 2 dpf

A B

C D

E F

G H

I

J K

fgf8a fgf 8a

fgf8b fgf 8b

fgf17 fgf17

fgf18a fgf 18a

fgf18b

fgf24 fgf24

dd

mhb

pa

ov

rt

aeraer

mhb

dd

r

ov

pat

os

mhbmhb

ovov

ov ov

pa

pa

rr

pa

pa pa

ovov

rr

rh

mhb

ov

pa

rn

mhb

ov ov

nmhb

r

n

pa

mhb

naeraer

nn

Fig. 7. Expression of stickleback and zebrafishFgfDgenesat long pec stage. A, C, E, G, J, 5 dpf stickleback embryos. B,D,F,H,I,K, 2 dpf zebrafish embryos. A, B,fgf8a. C, D, fgf8b.E, F, fgf17. G, H, fgf18a. I, fgf18b. J, K, fgf24. aer, apicalectodermal ridge of pectoral fin; dd, dorsal diencephalon; h,heart; mhb, midbrain-hindbrain boundary; n, nose; ov, oticvesicle; os, optic stalk; pa, pharyngeal arches; r, retina;t, telencephalon.

R. JOVELIN ET AL.12



13/26

zebrafish zfgf18b only adds expression in theretina and pharyngeal arches (Fig. 7GI).

fgf24:At the long pec stage, stickleback sfgf24isexpressed in the MHB, olfactory epithelium, mostcells of the otic vesicle, endoderm of the phar-

yngeal pouches, and apical epidermal ridge of thepectoral fin (Fig. 7J). Expression of sfgf24 in theeye and MHB is broader than the expression ofsfgf8a and sfgf18 in these domains (Fig. 7AG).Here we extend previous descriptions of zfgf24expression (Draper et al., 2003; Fischer et al.,2003); zfgf24is expressed in a small group of cellsin the dorsal MHB, in the olfactory epithelium,endoderm of the pharyngeal pouches, and the oticvesicle. In the AER of the pectoral fin in bothstickleback and zebrafish,fgf24is expressed broad-

er and stronger than fgf8a, with the fgf8 domainentirely within thefgf24domain (Fig. 7A, B, J, K).

Eyes and ears

Several FgfD genes are expressed in the eyesand ears, but often in different domains. Theganglion cell layer of the eye expresses sfgf8a,zfgf8a,sfgf17,and sfgf24 (Fig. 8A, C, I, S)(Pickeret al., 2005). In the retina, zfgf24 is expressed inisolated cells scattered in various layers (Fig. 8S,U). Except for the retina and MHB, zfgf24 andsfgf24 are expressed similarly.

The developing ear expresses all FgfD genes atlong pec stages in both stickleback and zebrafish.Some genes are expressed generally throughout

A B C D

E F G H

I J K L

M N O P

Q R

S T U V

fgf18fgf18

fgf8a fgf8a

fgf8b fgf8b

fgf17 fgf17

fgf18l fgf18l

fgf24 fgf24

Gac 120 hpf Dre 48 hpf

ac

am

am

cpac

ac

pc

cp

lc

am

am

am

am

ac

am

pc

ac pc

cp

cp

cp

cp

pcam

cp

pp

ac

pc

gl gl

gl

vp

il

Fig. 8. Expression of stickleback and zebrafishFgfDgenes in the ear. A,B,E,F,I,J,M,N,S,T, 5 dpf stickleback embryos.C, D,G, H, K, L, O, P,Q, R,U,V, 2 dpf zebrafish embryos. AD, fgf8a. EH,fgf8b. IL,fgf17. MP, fgf18a. Q, R, fgf18b. SV,fgf24. ac, anterior crista; am, anterior macula; cp, canal projection; fp, fusion plate; gl, gangloin layer; il, inner nuclear layer; lc,lateral crista; pc, posterior crista; pp, pharyngeal pouches; vp, ventral patch.




14/26

the ear, but others are only expressed in specific

regions. The anterior sensory patch shows strongexpression of sfgf8bin stickleback and its paralogzfgf8ain zebrafish, whereas expressing zfgf8bonlyweakly (Fig. 8D, F, H). In addition, the anteriorsensory patch expresses zfgf18b strongly andzfgf17 and zfgf18a only weakly (Fig. 8L, P, R).The anterior and posterior cristae express zfgf24(Fig. 8V).

In both stickleback and zebrafish during theearly pharyngula period, fgf18 and fgf18a areexpressed in specific spinal cord neurons (Fig. 9Aand B). None of the other FgfD genes are

expressed in spinal cord neurons at this stage inzebrafish, and none of the FgfD genes in mouseare expressed in spinal cord neurons at thisdevelopmental stage, suggesting that this expres-sion domain is either an innovation in stemteleosts, or that it was an ancestral function lostin the mammalian lineage.

Of the threeFgfDgenes in mouse, only Fgf18isexpressed in the pancreas (Dichmann et al., 2003).In stickleback and zebrafish long pec stage, sfgf18,zfgf18a, sfgf24, and zfgf24 are also expressed inthe pancreas (Fig. 9CF), suggesting that apancreas regulatory element was present beforethe gene duplication event that produced fgf18and fgf24.

Analysis of conserved noncoding elements

Expression patterns of theFgfDgene family justdescribed show many examples of conservationacross species and across paralogs. These con-served patterns are likely to be controlled byregulatory elements, the functions of which havebeen conserved for hundreds of millions of years.

Because some regulatory elements conserved infunction are also conserved in sequence, wesearched teleost and humanFgfDgenomic regionsfor CNEs.

CNE sharing amongFgfD paralogs withinspecies

Human:To discover ancient CNEs derived fromthe R1 and R2 genome duplication events, weconducted a Vista plot analysis of human FGFDgenes. The analysis allowed several conclusions(Fig. 10A). First, CNEs shared betweenFGF8andFGF18were more frequent and longer than thoseshared by FGF8 and FGF17. Second, CNEsembedded in the genes flanking FGF8 (FBXW4and NPM3) are conserved with those flankingFGF18, which is located adjacent to NPM1 and

near FBXW11. NPM2 is adjacent to FGF17(Fig. 5A), but CNEs embedded in NPM3 (nearFGF8) andNPM1(nearFgf18) are apparently notconserved with NPM2 (near FGF17). Third, aCNE (labeled a) shared by FGF8 and FGF18 isfound within intron four of FBXW4this issignificant because an enhancer trap elementinserted in a similar position in zebrafish confersan fgf8 expression pattern (Kikuta et al., 2007).Finally, three highly conserved CNEs located 3 toFGF8 (labeled d, e, and f) are shared by all threehuman FGFD genes. We conclude that at least

three noncoding regions in the genomic neighbor-hood of FGFD genes have been highly conservedin sequence for at least 500 million years (Kumarand Hedges, 98; Peterson et al., 2004) since theR1 and R2 genome duplication events. Our resultsraise the hypothesis that these conserved se-quences are regulatory elements that drive theexpression ofFGFD genes in domains common toall three genes.

Stickleback: To test if CNEs nearFgfDgenes arealso conserved in teleosts, we constructed Vistaplots for stickleback. Fig. 10B shows conservationfor the exons offgf8a/fgf8band scl2a5a/scl2a5batgreater than threshold, but no CNEs shared by thetwo paralogs. This result is a bit surprising giventhe similarities of expression patterns between thetwo sfgf8ohnologs. In contrast, several CNEs wereapparent in the comparison of sfgf8a and sfgf17,although most exons failed to rise above threshold(Minimum Y value on the VISTA plot550%,Minimum conservation identity570%, Minimumlength for a CNE5100).

Three of the stickleback CNEs (a, b, e) were inintrons offgfDneighbor genes, and two were found

Gac Dre

fgf18

fgf18

fgf18a

fgf18a

fgf24 fgf24

A B

CD

E F

2.5 dpf 1dpf

5 dpf

5 dpf

2 dpf

2 dpf

spn

spn

pan pan

panpan

Fig. 9. Expression of stickleback and zebrafishFgfDgenesin the trunk. A, C, E, stickleback embryos. B,D, F, zebrafishembryos. A, B, fgf18 and fgf18a. C, D, fgf18 and fgf18a. E, F,fgf24. pan, pancreas; spn, spinal cord neurons.

R. JOVELIN ET AL.14



15/26

in intergenic region flanking the fgfD gene. Notethat CNE a in Fig. 10B is in the same intron ofstickleback fbxw4 as CNE a of human FBXW4although it is shared between the fgf8aand fgf17regions for stickleback and FGF8 and FGF18regions for human. Neither the exons nor anyCNEs were detected that were shared betweensfgf8aand either sfgf18or sfgf24, suggesting rapidevolution, translocations, or problems with genomeassembly. The sparse pattern of CNEs in stickle-back fgfD genes contrasts with the rather broadCNEs found for human FGFD genes.

Zebrafish: For zebrafish (Fig. 10C), a strong CNE(labeled d in Fig. 10B and C and a in Fig. 10A)was detected in the same intron of fbxw4 asdetected in stickleback and in human, and it wasshared by all zebrafishfgfDregions except fgf18b.

An enhancer trap insertion nearby (CLGY667)confers reporter expression in telencephalon, opticstalk, midbrainhindbrain boundary, somites,heart, olfactory pits, and tail bud in a patternsimilar to that offgf8 genes (Kikuta et al., 2007).This CNE, conserved since before R1 and main-tained byfgf8a, fgf8b, fgf17, fgf18a,andfgf24, may

Fig. 10. Distribution of CNEs amongFgfDparalogs in human (A), stickleback (B) and zebrafish (C). The higher density ofCNEs among human paralogs is probably the result of the R3 genome duplication in teleosts with subsequent relaxed selectionand partitioning of regulatory elements (Force et al., 1999). The reference sequence is indicated at the top. The position of exonsis indicated along the toop. For each conservation plot, the top line represents 100% conservation with the reference sequence,and the bottom line indicates 50% conservation. Gray arrows show gene orientation. Lower case letters indicate peaksmentioned in the text.




16/26

play a role in defining expression domains sharedby vertebrateFgfDgenes.

Despite their similarity in expression patterns(Fig. 8AH, 9AD), fgf8ashares fewer CNEs withfgf8b than it does with fgf17, fgf18a, and fgf24

(Fig. 10C). Using fgf8a as base, one CNE (d) ispresent near five zebrafishfgfDgenes, three (b, n,and r) are present near three fgfD genes, eight(f, i, k, l, m, q, w, and y) are near two fgfDgenes,and 14 (a, c, e, g, h, j, o, p, s, t, u, v, x, and z) areshared byfgf8aand only one or anotherfgfDgene.fgf8ashares most CNEs (17) with fgf24and aboutan equal number withfgf17andfgf18a(11 and 10,respectively).

Enhancer trap CLG508 is located in betweenfgf8a and fbxw4 and confers expression in theapical ectodermal ridge of the pectoral fin bud(Kikuta et al., 2007), but Vista detected no CNE

shared by fgf8a and other fgfD genes in thisintergenic region despite the expression of bothfgf8 and fgf24 in the AER. Enhancer trapCLGY1030 is inserted near CNE u in zebrafishand matches the expression pattern of fgf8aonlyin the tail bud (Kikuta et al., 2007).

CNE sharing amongFgfD orthologs

CNEs displayed when comparing FgfD geneswithin a species show elements conserved fromthe R2 event, whereas CNEs revealed by compar-

ing a single FgfD gene among species identifieselements conserved since species divergence.fgf17: Few CNEs appeared in comparisons of the

stickleback fgf17gene to its orthologs in humanand several teleosts, although exons were readilyapparent (Fig. 11). At least six CNEs were

apparent when comparing stickleback and zebra-fishfgf17genes, although the flanking genes werenot conserved (Fig. 11). Given that stickleback andpufferfish are more closely related than stickle-back and zebrafish, it was surprising to find more

sfgf17 CNEs shared with zebrafish than puffer-fish. The lack of CNEs near fgf17 in the twopufferfish species could be due in part to incom-plete sequencing in this genomic region: fgf17in fugu (Takifugu rubripes) lies in a smallscaffold without predicted neighbors, and numer-ous sequencing gaps surround fgf17 in thegreen spotted pufferfish T. nigroviridis. Amongthe species examined, only stickleback andT. nigroviridis share a neighboring gene(ENSGACG00000003433) (Fig. 11). Close inspec-tion of chromosome segments carrying fgf17 inzebrafish and stickleback shows that global syn-

teny is conserved but gene orders are not,indicating local genomic micro-rearrangements(Fig. 5). Comparison with human reveals rearran-gements over still larger genomic distances(Fig. 5). Chromosome breaks nearfgf17may havecarried CNEs along with fgf17neighbors disrupt-ing synteny of coding and noncoding sequences inthese regions.

fgf18: BLAST searches identified a single fgf18gene in stickleback, medaka and both pufferfishand the topology of the fgf18 clade suggests thatfgf18bmay have been lost in the lineage leading to

these species (Jovelin et al., 2007). In contrast tothe four percomorph species, zebrafish has twofgf18 genes. A number of sfgf18 CNEs are foundassociated only with zebrafish fgf18a(b, c. d, e, g,h, i, j, l, n, o, q), a few only with zfgf18b(a, f ), andseveral with both (k, m, r) (Fig. 12). This pattern

Fig. 11. Distribution of CNEs amongFgf17orthologs. The low number of CNEs is likely owing to micro-rearrangementsaround the Fgf17locus disrupting synteny of coding and noncoding sequences.

R. JOVELIN ET AL.16



17/26

of reciprocal distribution of CNEs is what wouldbe expected if the single stickleback fgf18 generepresented much of the ancestral distribution ofCNEs and they had become distributed as pre-dicted by subfunctionalization in the zebrafish

lineage after gene duplication.The high density of CNEs surrounding fgf18suggests that fgf18b may have been lost shortlyafter the separation of the lineages leading tozebrafish and stickleback and that CNEs may havebeen maintained by selection in lineages in which

only one fgf18gene subsisted. However, in zebra-fish, which retained both fgf18 duplicates, mostCNEs found aroundfgf18of species retaining justone fgf18 gene have been partitioned betweenfgf18aand fgf18b(Fig. 12).

fgf24:The high density of CNEs in proximity tofgf24 orthologs in all tested species (Fig. 13) isconsistent with our inference that one fgf24duplicate was lost before the divergence of thelineages leading to zebrafish and stickleback(Jovelin et al., 2007). Two particularly strong

Fig. 12. Most of the CNEs conserved among stickleback, medaka and pufferfish fgf18 are partitioned between fgf18a andfgf18bin zebrafish, consistent with the DDC model (Force et al., 99). The high density of CNEs in teleost species having a singlefgf18 gene suggests that the second fgf18 duplicate was lost shortly after the R3 duplication. The region lacking CNEs intetraodon is owing to a gap in the sequence alignment.

Fig. 13. The high density of CNEs among teleostfgf24 orthologs suggests that the secondfgf24 duplicate was lost shortlyafter the R3 duplication before the partitioning offgf24 subfunctions. Medakafgf24 is located on a short scaffold, hence theapparent lack of conservation further away from fgf24.




18/26

CNEs (labeled a and d) lie 30 and 50 respectively tofgf24and two others (b and c) are within introns ofthe fgf24 gene. We checked to see if Fgfd-relatedCNEs could be detected in the location predictedfor the missing human FGF24, but found none.

fgf8: In stickleback and zebrafish, fgf8 dupli-cates exhibit common and distinct expressionpatterns, and comparisons betweenfgf8orthologsshow that partitioning of most ancestral functionsoccurred before the divergence of these species(Jovelin et al., 2007). Thus, we should expectregulatory elements to be largely shared betweenfgf8 orthologs in stickleback and zebrafish ratherthan being species-specific.

For comparisons based on stickleback fgf8a, atleast four CNEs were shared with the humanFGF8 region (Fig. 14A). These included elementsin introns offbxw4(CNEs a, b, and d in Fig. 14A)

and an element in an intron of fgf8a itself(element h). Among the five stickleback fgfDgenes, the fgf8 comparisons were the only onesthat revealed any CNEs shared with human. All ofthe elements sfgf8ashared with human were also

shared with zebrafish. In addition, zebrafishfgf8ashared with stickleback fgf8aseveral elements inthe intergenic spaces 30 and 50 to fgf8a. As notedfor other fgfD genes, stickleback fgf8a had moreand longer CNEs shared with percomorphs thanwith zebrafish. With few exceptions (see elementd, Fig. 14A), the percomorphs had very similarCNE patterns.

The comparison of the stickleback fgf8b regionto that of human revealed a CNE 30 to thesticklebackfgf8bgene (element b, Fig. 14B) sharedwith other teleosts except T. nigroviridis. A CNEin an intron of fgf8b (element c in Fig. 14B) was

Fig. 14. Distribution of CNEs among Fgf8 orthologs. CNEs are partitioned in unique sets between teleost fgf8a (A) andfgf8b(B), reflecting the partitioning of expression domains between fgf8duplicates (this work and (Jovelin et al., 2007)). Fewerfgf8bCNEs are shared with human FGF8.

R. JOVELIN ET AL.18



19/26

shared with zebrafish and fugu. A large number ofCNEs located 50 tofgf8b(e.g. element g) were heldin common by several teleosts that were not seenin the comparison with zebrafish.

Therefore, the overall pattern of conservation of

noncoding sequences in the proximity of fgf8correlates with the evolutionary history of fgf8duplicates and their subsequent functional diver-sification: fgf8a and fgf8b orthologs have uniquesets of CNEs, some of which are shared withhuman as it would be expected if subfunctionshave been partitioned among paralogs.

DISCUSSION

Vertebrate FgfD genes originated asohnologs in R1 and R2

To infer the evolutionary history of expressiondomains and conserved noncoding sequences, it isessential to understand the phylogenetic relation-ships of the four FgfDgenes (Suppl. Fig. S4). Thephylogenetic tree of these genes shows a history of(((Fgf8, Fgf17)Fgf18)Fgf24) (Jovelin et al., 2007).Several other data sets, however, support theconclusion that the four FgfD genes arose in theR1 and R2 WGD events. Evidence supporting thisconclusion comes from four independent sources:(1) The chromosome region surrounding theCiona FgfD gene contains genes with humanorthologs mainly on human chromosome seg-

ments containing FGF8, FGF17, and FGF18, aswell as part of Hsa2p. (2) An automated search forparalogous chromosome segments around FGF8revealed chromosome segments surroundingFGF17, FGF18, and a portion of Hsa2p. (3) Theelephant shark has a copy ofFgf24, demonstratingthe origin of this gene before the divergence ofteleost and tetrapod genomes. (4) The immediategenomic neighborhood of the teleost fgf24gene ispreserved in the human genome with the ratherprecise excision of an ortholog to fgf24 and itsnearest neighbor npm4. These data support the

conclusion that R1 and R2 produced four ohnologsrelated as ((Fgf8, Fgf17)(Fgf18, Fgf24)) and thatthe tetropod Fgf24 gene went missing fromtetrapod genomes.

Comparative analysis of expressionpatterns

After the R2 genome duplication event, the fourFgfD genes would have been preserved either bysubfunctionalization (the duplicate preservationby reciprocal partitioning of ancestral gene sub-

functions) or neofunctionalization (the origin ofnovel domains) (Force et al., 99; Hughes, 99;Stoltzfus, 99; Postlethwait et al., 2004; Chainet al., 2008; Conant and Wolfe, 2008), andsubsequent to preservation, additional expression

domains or other functions could have been lostor invented. To analyze these events, weinvestigated expression patterns ofFgfDgenes intwo teleosts and compared them with publisheddata available on their orthologs in mouse(Suppl. Fig. S4).

The single FgfDgene present in stem olfactores(urochordates1vertebrates) was an ancestorof Fgf8/17/18 in the ascidian urochordateC. intestinalisand its expression pattern serves asan outgroup to the vertebrate FgfD genes. InCiona, Fgf8/17/18 is expressed in the embryonicCNS coincident with En and Pax2/5/8 (Ikuta and

Saiga, 2007). Correspondingly, the vertebrateorthologs of these three Ciona genes (Fgf8, Fgf17,Fgf18, En1, En2, Pax2, Pax5, and Pax8) are co-expressed in the vertebrate CNS at the midbrain-hindbrain boundary (MHB) (Joyner, 96). Althoughurochordates may lack an MHB organizer secon-darily (Canestro et al., 2005; Ikuta and Saiga,2007), this gene set appears to constitute an ancientregulatory cassette in chordate CNS development.The expression of all four vertebrateFgfDohnologsin the MHB region in at least one vertebratespecies (Suppl. Fig. S4) suggests that this was an

ancient expression domain existing before thedivergence of urochordates and vertebrates andthat this expression domain was retained by allfour genes after the R1 and R2 WGD events at thebase of the vertebrate radiation.

Besides its expression in the CNS, Fgf8/17/18 isalso expressed in C. intestinalis in mesodermalcells lateral and anterior to the CNS expressiondomain, as well as cells at the tip of the tail (Imaiet al., 2002; Ikuta and Saiga, 2007). The relation-ship of these two urochordate domains to similardomains in the vertebrate somitic and unsegmen-ted mesodermal expression and tail bud domain isuncertain. In the cephalochordate amphioxus,Fgf8/17/18 is expressed transiently in the rostralCNS and in the pharyngeal endoderm, whichare likely conserved domains with vertebrates(Muelemans and Bonner-Fraser, 2007).

A substantial number of morphological innova-tions or elaborations occurred after the divergenceof urochordates and vertebrates but before thedivergence of bony fishes into ray-fin and lobe-finclades. These include the elaboration of neural crestand placodes, the invention of bone, and the origin




20/26

of paired appendages (Gans and Northcutt, 83;Bassham and Postlethwait, 2005; Jeffery, 2006;Sauka-Spengler and Bronner-Fraser, 2008). Alsooccurring at about this time period were two roundsof WGD, R1 and R2 (Kortschak et al., 2001; Ornitz

and Itoh, 2001; Lundin et al., 2003; Dehal andBoore, 2005; Garcia-Fernandez, 2005; Bourlat et al.,2006; Delsuc et al., 2006; Guder et al., 2006; Jacoband Lum, 2007; Kitisin et al., 2007; Sundstromet al., 2008; Yu et al., 2008) that we show here arehighly likely to have produced the four FgfDohnologs (Fgf8, Fgf17, Fgf18, and Fgf24). Anyexpression domain that appears in the orthologs ofthree or four of these four genes in at least one of thethree species stickleback, zebrafish, and mouse, ishighly likely to have been a site of expression in thelast common ancestor of all bony fishes.

In addition to the MHB and somite/unsegmen-

ted mesoderm, at least one of the three investi-gated vertebrates (stickleback, zebrafish, mouse)expresses eachFgfDgene in the retina, pharyngealarches, somites, and otic vesicle (Suppl. Fig. S4). Weconclude that these expression domains wereprobably present in the four orthologous FgfDgenes in the last common ancestor of ray-fin andlobe-fin fish and were differentially lost fromvarious genes in different lineages. Because thesedomains were in common in all four FgfD genes,we infer that they were also in common in theunduplicatedFgfD gene in the lineage leading to

vertebrates before R1 and R2. An alternativehypothesis is that these shared expression do-mains evolved independently but convergentlymultiple times, a possibility that seems unlikelybased on parsimony. If the hypothesis is true thatexpression domains shared by all fourFgfDgeneswere present in the last pre-R1-duplication FgfDgene, then that vertebrate ancestor would havehad expression domains appropriate for eyes, ears,pharyngeal arches, somites, and MHB. Availabledata forFgfDexpression patterns in cephalochor-dates and urochordates do not contradict thisexplanation (Imai et al., 2002; Ikuta and Saiga,2007); Muelemans and Bonner-Fraser, 2007).

Expression domains in the dorsal diencephalonand the apical epidermal ridge (AER) of thepectoral fin/limb bud are shared by Fgf8 andFgf24 (Figs. 6, 7). Furthermore, the examinationof mutant phenotypes shows that Fgf8is requiredfor paired pectoral appendage development intetrapods, which lack an ortholog of fgf24, andreciprocally, fgf24 is required for paired pectoralappendage development in zebrafish but fgf8a isnot (Reifers et al., 98; Lewandoski et al., 2000;

Draper et al., 2003) (fgf8b is not expressed in theAER of fish pectoral appendages (Jovelin et al.,2007)). If the R1, R2 evolutionary history of thefour FgfD genes is correct, then the pre-duplica-tion FgfD gene should have had an expression

domain corresponding to the dorsal diencephalonand AER domains in modern vertebrates. Accord-ing to this hypothesis, the last ancestor before thevertebrate genome duplications would thus havehad an expression domain for the AER, eventhough, paradoxically, it is unlikely that thatorganism had paired appendages (Coates, 94).Thus, the ancestral function of this expressiondomain would have been co-opted into appendagedevelopment after paired appendages began toevolve in early vertebrates, and it may have beenused in the development of unpaired midline fins(Freitas et al., 2006). Alternatively, one or both of

these expression domains could have evolvedconvergently as neofunctionalizations by the twogenes Fgf8 and Fgf24 well after the R1 and R2events during the evolution of paired fins.

Some additional expression domains are displayedby just two of the four FgfD genes (Suppl. Fig. 4).Expression in tailbud is shared byFgf8 andFgf17andin unsegmented mesoderm and pancreas by Fgf18and Fgf24. This distribution might be predicted bythe (Fgf8, Fgf17)(Fgf18, Fgf24) lineage hypothesis ifthese domains had been a part of the repertoire of theFgf8/17andFgf18/24genes after the R1 WGD event.

As mentioned above, the tailbud expression domainmay be related to the expression domain at the tip ofthe tail in Ciona and inherited from stem olfactores(Imai et al., 2002; Ikuta and Saiga, 2007). In addition,both Fgf8 and Fgf18 share expression in the heartand Fgf17is expressed in zebrafish heart regenera-tion, consistent with either inheritance from theFgfDgene of stem olfactores followed by loss in theFgf24clade or, less parsimoniously, independent acquisitionby convergent evolution.

Several expression domains appear only on singleclades in the tree. These include segmental plate onlyin teleost fgf17genes, telencephalon only in Fgf8ofteleosts and tetrapods, olfactory epithelium only inteleostfgf8genes, and spinal neurons only in fgf18aof teleosts. The most parsimonious explanation ofthese domains is that they represent neofunctiona-lization events in the corresponding clades.

Correlation of CNEs and expressiondomains

Because many CNEs are shared by fgf8a andother zebrafish FgfD genes (Fig. 10C), we can

R. JOVELIN ET AL.20



21/26

attempt to correlate individual CNEs with expres-sion domains in embryonic zebrafish that aresummarized in Supplementary Fig. 4. The firstconclusion is that none of the CNEs identified wasunique for a single tissue. CNE q is shared only by

fgf8a, fgf18a, and fgf24, and only these genesexpress in the olfactory organ, making this acandidate regulatory element for contributing toolfactory expression, although genes possessingCNE q are expressed in tissues in addition to theolfactory epithelium. The dorsal diencephalonexpresses only fgf8a, fgf17, and fgf24, and thesegenes share CNEs labeled b, d, i, k, l, n, and w.(Fig. 10C); of these shared elements, i, k, l, and ware absent from paralogs that do not showexpression in the dorsal diencephalon, so theseare candidate CNEs for positive regulatory ele-ments contributing to expression in the dorsal

diencephalon of zebrafish embryos. Elements e, g,h, p, t, x, and z are found in both fgf8aand fgf24and in none of the other fgfDgenes, and only thesetwo genes are expressed in the AER at 2 dpf,making these elements candidates for controllingexpression in the apical ectodermal ridge of thepectoral fin bud. For the retina at 16 hpf, onlyfgf18blacks expression, and element d is the onlyelement possessed by all zebrafish fgfD genesexceptfgf18b, and so element d is a candidate forthe early expression in the retina. Other factorsmay turn on fgf18b in the retina at pharyngula

stages. Genes expressed in the heart and thespinal cord neurons have in common the same setof elements (b, c, d, m, n, q, r, s, u, and v). Theunsegmented mesenchyme, segmental plate, andpancreas each have unique CNE combinations notshared with other tissues. Because vertebrates canconserve regulatory function without conservingsequence similarity (Fisher et al., 2006), thesignificance of these conserved noncoding regionsmust be tested by experiment to draw firmconclusions.

CNEs among paralogs reflect theevolutionary history of the FgfD subfamily

Gene fates after duplication include the loss ofone of the two paralogs by nonfunctionalization,or the divergence of gene function by the acquisi-tion of new function(s) (neofunctionalization)(Ohno, 1970) or the partitioning of ancestralfunctions (subfunctionalization) (Force et al., 99;Hughes, 99; He and Zhang, 2005). Subfunctionpartitioning occurs by the complementary fixationof otherwise deleterious mutations in ancestral

control elements of gene duplicates. This modelessentially follows neutral evolution, andtherefore one might expect that subfunctionpartitioning should increase with the time ofcoexistence between duplicates within a lineage.

Consequently, the DuplicationDegenerationComplementation (DDC) model predicts that moreancestral functions should have been partitionedbetween paralogs that have coexisted over longperiods of evolutionary time than in cases whereduplicates have existed for a shorter time period orwhere duplicates rapidly became reduced to singlecopy after the duplication event.

Examples of subfunction partitioning can befound for a variety of gene duplicates (e.g. Listeret al., 2001; Cresko et al., 2003; Liu et al., 2005).The FgfD subfamily in teleosts provides never-theless a rare opportunity to test predictions of the

DDC model because its members share an originthrough the R1, R2, and R3 genome duplicationevents, and because since R3, different FgfDsubfamily members have spent different amountsof time as duplicate copies in various lineages. Thefgf24 and fgf17 genes apparently became singlecopy before the divergence of zebrafish andstickleback lineages; the fgf18a and fgf18b para-logs co-existed from R3 to the present in zebrafish,but became single copy in the stickleback lineageafter it diverged from the zebrafish lineage; andthe fgf8aand fgf8bparalogs have co-existed from

R3 to the present in several lineages. In addition,CNEs identified using phylogenetic footprinting inthe proximity of theFgf8locus have been shown tohave regulatory activity in mouse (Beermannet al., 2006) and in zebrafish (Inoue et al., 2006).Moreover, regulatory modules of fgf8 functionextend to intronic sequences of fgf8 neighbors(Kikuta et al., 2007). It is therefore likely thatCNEs identified around other Fgf genes have afunctional role as well.

The large number of long CNEs detected amongthe three human FGFD genes contrasts with thefewer shorter CNEs found in zebrafish andstickleback. What could be responsible for thisdifference? It is not mere antiquity because thelast common ancestor of the zebrafish fgf8 andfgf18agenes was in fact the same gene as the lastcommon ancestor of the humanFGF8and FGF18genes. Thus, either the human lineage preservedmore broad CNEs than teleost lineages, perhapsamplifying them locally by tandem duplication,or the zebrafish lineage lost more CNEs afterthe divergence of human and zebrafish lineages.The pattern of just a few, short, but nevertheless




22/26

well-conserved CNEs within stickleback and zeb-rafish genomes may relate to the additional WGDin the teleost lineage not suffered by the tetrapodlineage (R3). Post-R3 fgfD paralogs may haveexperienced greater relaxation of selective con-

straints than post-R2 FgfD genes. After bothevents, we would expect the partitioning ofexpression domains by fixation of complementarydeleterious mutations in regulatory elements(Force et al., 99; Lynch a nd Force, 2000).

After R3, one fgf18 duplicate was lost after thedivergence of the lineages leading to zebrafish andstickleback resulting in one fgf18gene in stickle-back, medaka, and pufferfish whereas zebrafishretained twofgf18duplicates. Analysis of CNEs inthe proximity of the sticklebackfgf18 locus showsextensive conservation between stickleback andother percomorph fish (Fig. 12), suggesting that

the integrity of regulatory elements in thisgenomic region is important for fgf18 functionand has been maintained by selection in specieshaving a single fgf18 gene. In contrast, fewerCNEs are shared between zebrafish and thepercomorphs, and most CNEs have been parti-tioned between fgf18a and fgf18b in zebrafish.This pattern of shared and partitioned CNEsmakes two predictions. First, the amount andlevel of conservation among CNEs surroundingfgf18 in percomorphs suggest that little subfunc-tion partitioning may have occurred before the

secondfgf18duplicate was lost in this lineage andthusfgf18in these species may have retained mostof the ancestral pre-R3 fgf18 functions. Second,the pattern of shared CNEs between fgf18a andfgf18b in zebrafish, where both duplicates havebeen maintained over longer evolutionary time,suggests the partitioning of ancestral pre-R3fgf18functions between paralogs. Supporting thesepredictions, stickleback fgf18 is expressed in theheart field precursor and in the MHB at mid-segmentation, whereas in zebrafish, these expres-sion domains are partitioned between fgf18a andfgf18b, respectively (Fig. 6). At the long pec stage,expression of stickleback fgf18 in the olfactoryepithelium and in the MHB is partitioned betweenzebrafish fgf18a and fgf18b (Fig. 7). Similarly, atthe early pharyngula stage, expression in thespinal chord neurons partitioned to zfgf18a(Suppl. Fig. S4). Specifically, one or more ofelements a, f, and r in Fig. 12 arecandidates for expression of fgf18b in the MHB,and elements be, gj, and oq are candi-dates for expression of fgf18a in the nose, spinalneurons, and pancreas, assuming all regulatory

elements act positively, although, of course somemay act to diminish expression in certain tissues.

Sequenced teleost genomes have a single fgf24and fgf17 gene, indicating that one duplicate ofeach locus was lost after R3, before the divergence

of the lineages leading to zebrafish and other fishspecies. The extensive number and length of CNEsaround all fgf24 orthologs also supports thishypothesis and suggests that regulatory controlelements of fgf24 have been maintained byselection (Fig. 13). By contrast, the lower numberof CNEs surrounding Fgf17 genes may be me-chanistically related to the multiple local chromo-some rearrangements disrupting syntenies of bothCNEs and Fgf17 neighboring genes betweenteleosts and human and within teleosts (Figs. 5,11). CNEs detached from their target genes bychromosomal rearrangements are expected to be

lost by genetic drift (Kikuta et al., 2007). Forinstance, it is noteworthy thatFgf17expression isthe least conserved between mouse and teleostsamong FgfD members (Suppl. Fig. S4). Theexpression patterns of fgf17 orthologs in stickle-back and zebrafish, however, are well conserved,suggesting that important ancestral regulatoryelements may have been retained in both lineages,and CNEs ae in Figure 11 are candidates forsuch elements.

Because one fgf24 duplicate was lost before thedivergence of zebrafish and percomorph lineages

and one fgf18 duplicate was lost after theselineages diverged, fgf24 duplicates originatingfrom R3 have coexisted over shorter evolutionarytime than fgf18 duplicates originating from thesame WGD before one copy of each gene was lost.Thus, we would expect the functions of fgf24orthologs to be more conserved than those offgf18orthologs. Indeed, not only does zfgf24 sharemore and longer CNEs with its orthologs thanzfgf18a and zfgf18b share with their orthologs(Figs. 12 and 13) but zfgf24and sfgf24expressiondomains are remarkably conserved with onlyminor differences in the retina and MHB(Figs. 6SV, 7JK, 8SV), whereaas fgf18 ortho-logs show expression differences in somites, retinaand ear (Figs. 6N, P, R, 8MR).

In summary, this comparison of CNEs andexpression domains verifies predictions made bythe molecular mechanisms of subfunction parti-tioning and the DDC model (Force et al., 99) forall teleostfgfDmembers for which synteny has notbeen disrupted by local chromosomal rearrange-ments. Additionally, these results identify a clearrelationship, predicted by the subfunctionalization

R. JOVELIN ET AL.22



23/26

model, between the length of time that paralogscoexist and their distribution of both CNEs andexpression domains. Importantly, this work iden-tifies a number of CNEs that are candidates forregulatory elements for specific expression do-

mains in development. The testing of thesedomains for developmental functions is anongoing project.

ACKNOWLEDGMENTS

We thank Dong Liu for discussions on earmorphology, Amanda Rapp, Tim Mason and theUniversity of Oregon Zebrafish Facility for provid-ing animals and excellent fish care, and PohKheng Loi and Amber Selix of the HistologyFacility for sectioning. This work was supported

by the National Center for Research Resources(5R01RR020833) and National Institutes ofHealth (P01 HD22486); the contents of this studyare solely the responsibility of the authors and donot necessarily represent the official views ofNCRR or NIH. RJ is an Associate of the IGERTTraining Program for Development, Evolution,and Genomics, NSF DGE-9972830.

LITERATURE CITED

Abu-Issa R, Smyth G, Smoak I, Yamamura K, Meyers EN.

2002. Fgf8 is required for pharyngeal arch and cardiovas-cular development in the mouse. Development 129:46134625.

Amemiya CT, Prohaska SJ, Hill-Force A, Cook A,

Wasserscheid J, Ferrier DE, Pascual-Anaya J, Garcia-

Fernandez J, Dewar K, Stadler PF. 2008. The amphioxusHox cluster: characterization, comparative genomics, andevolution. J Exp Zool B Mol Dev Evol 310:465477.

Amores A, Force A, Yan YL, Joly L, Amemiya C, Fritz A,

Ho RK, Langeland J, Prince V, Wang YL, Westerfield M,Ekker M, Postlethwait JH. 1998. Zebrafish hox clusters andvertebrate genome evolution. Science 282:17111714.

Andermann P, Weinberg ES. 2001. Expression of zTlxA,a Hox11-like gene, in early differentiating embryonic

neurons and cranial sensory ganglia of the zebrafish

embryo. Dev Dyn 222:595610.Bassham S, Postlethwait JH. 2005. The evolutionary history

of placodes: a molecular genetic investigation of the

larvacean urochordate Oikopleura dioica. Development132:42594272.

Beermann F, Kaloulis K, Hofmann D, Murisier F, Bucher P,Trumpp A. 2006. Identification of evolutionarily conserved

regulatory elements in the mouse Fgf8 locus. Genesis44:16.

Bershtein S, Tawfik DS. 2008. Ohnos model revisited:measuring the frequency of potentially adaptive mutations

under various mutational drifts. Mol Biol Evol 25:23112318.

Blair JE, Hedges SB. 2005. Molecular phylogeny and diver-gence times of deuterostome animals. Mol Biol Evol22:22752284.

Boulet AM, Moon AM, Arenkiel BR, Capecchi MR. 2004. Theroles ofFgf4and Fgf8in limb bud initiation and outgrowth.Dev Biol 273:361372.

Bourlat SJ, Juliusdottir T, Lowe CJ, Freeman R, Aronowicz J,Kirschner M, Lander ES, Thorndyke M, Nakano H,Kohn AB, Heyland A, Moroz LL, Copley RR, Telford MJ.2006. Deuterostome phylogeny reveals monophyletic chor-dates and the new phylum Xenoturbellida. Nature 444:8588.

Brand M, Heisenberg CP, Jiang YJ, Beuchle D, Lun K, Furutani-Seiki M, Granato M, Haffter P, Hammerschmidt M, Kane DA,Kelsh RN, Mullins MC, Odenthal J, van Eeden FJ, Nusslein-

Volhard C. 1996. Mutations in zebrafish genes affecting theformation of the boundary between midbrain and hindbrain.Development 123:179190.

Brudno M, Do CB, Cooper GM, Kim MF, Davydov E,Green ED, Sidow A, Batzoglou S. 2003. LAGAN andMulti-LAGAN: efficient tools for large-scale multiple align-

ment of genomic DNA. Genome Res 13:721731.Canestro C, Bassham S, Postlethwait J. 2005. Development of

the central nervous system in the larvacean Oikopleuradioica and the evolution of the chordate brain. Dev Biol285:298315.

Cao Y, Zhao J, Sun Z, Zhao Z, Postlethwait J, Meng A. 2004.fgf17b, a novel member of Fgf family, helps patterningzebrafish embryos. Dev Biol 271:130143.

Catchen JM, Conery JS, Postlethwait JH. 2009. Automatedidentification of conserved synteny after whole genomeduplication. Genome Res. [Epub ahead of print].

Chain FJ, Ilieva D, Evans BJ. 2008. Duplicate gene evolutionand expression in the wake of vertebrate allopolyploidiza-tion. BMC Evol Biol 8:43.

Cholfin JA, Rubenstein JL. 2007. Patterning of frontal

cortex subdivisions by Fgf17. Proc Natl Acad Sci USA 104:76527657.

Christoffels A, Koh EG, Chia JM, Brenner S, Aparicio S,Venkatesh B. 2004.Fugugenome analysis provides evidencefor a whole-genome duplication early during the evolution ofray-finned fishes. Mol Biol Evol 21:11461151.

Coates MI. 1994. The origin of vertebrate limbs. Dev Suppl169180.

Conant GC, Wolfe KH. 2008. Turning a hobby into a job: howduplicated genes find new functions. Nat Rev Genet9:938950.

Cresko WA, Yan YL, Baltrus DA, Amores A, Singer A,Rodriguez-Mari A, Postlethwait JH. 2003. Genome duplica-tion, subfunction partitioning, and lineage divergence: Sox9in stickleback and zebrafish. Dev Dyn 228:480489.

Crow KD, Stadler PF, Lynch VJ, Amemiya C, Wagner GP.2006. The fish-specific Hox cluster duplication is coin-cident with the origin of teleosts. Mol Biol Evol 23:121136.

Dehal P, Boore JL. 2005. Two rounds of whole genomeduplication in the ancestral vertebrate. PLoS Biol 3:e314.

Delsuc F, Brinkmann H, Chourrout D, Philippe H. 2006.Tunicates and not cephalochordates are the closest livingrelatives of vertebrates. Nature 439:965968.

Delsuc F, Tsagkogeorga G, Lartillot N, Philippe H. 2008.Additional molecular support for the new chordate phylo-geny. Genesis 46:592604.

Dichmann DS, Miller CP, Jensen J, Scott Heller R, Serup P.2003. Expression and misexpression of members of the FGF




24/26

and TGFbeta families of growth factors in the developingmouse pancreas. Dev Dyn 226:663674.

Dominguez MH, Rakic P. 2008. Neuroanatomy of the FGFsystem. J Comp Neurol 509:141143.

Draper BW, Stock DW, Kimmel CB. 2003. Zebrafish fgf24functions with fgf8 to promote posterior mesodermal

development. Development 130:46394654.Fischer S, Draper BW, Neumann CJ. 2003. The zebrafishfgf24 mutant identifies an additional level of Fgf signalinginvolved in vertebrate forelimb initiation. Development130:35153524.

Fisher S, Grice EA, Vinton RM, Bessling SL, McCallion AS.2006. Conservation of RET regulatory function from humanto zebrafish without sequence similarity. Science 312:276279.

Force A, Lynch M, Pickett FB, Amores A, Yan YL,Postlethwait J. 1999. Preservation of duplicate genes bycomplementary, degenerative mutations. Genetics 151:15311545.

Frank DU, Fotheringham LK, Brewer JA, Muglia LJ,Tristani-Firouzi M, Capecchi MR, Moon AM. 2002. An

Fgf8 mouse mutant phenocopies human 22q11 deletionsyndrome. Development 129:45914603.

Frazer KA, Pachter L, Poliakov A, Rubin EM, Dubchak I.2004. VISTA: computational tools for comparative geno-mics. Nucleic Acids Res 32:W273W279.

Freitas R, Zhang G, Cohn MJ. 2006. Evidence that mechan-isms of fin development evolved in the midline of earlyvertebrates. Nature 442:10331037.

Gans C, Northcutt RG. 1983. Neural crest and the origin ofvertebrates: a new head. Science 220:268274.

Garcia-Fernandez J. 2005. The genesis and evolution ofhomeobox gene clusters. Nat Rev Genet 6:881892.

Guder C, Philipp I, Lengfeld T, Watanabe H, Hobmayer B,Holstein TW. 2006. The Wnt code: cnidarians signal theway. Oncogene 25:74507460.

Hamade A, Deries M, Begemann G, Bally-Cuif L, Genet C,Sabatier F, Bonnieu A, Cousin X. 2006. Retinoic acidactivates myogenesis in vivo through Fgf8 signalling. DevBiol 289:127140.

He X, Zhang J. 2005. Rapid subfunctionalization accompaniedby prolonged and substantial neofunctionalization in dupli-cate gene evolution. Genetics 169:11571164.

Hedges SB. 2002. The origin and evolution of model organ-isms. Nat Rev Genet 3:838849.

Hoegg S, Brinkmann H, Taylor JS, Meyer A. 2004. Phyloge-netic timing of the fish-specific genome duplication corre-lates with the diversification of teleost fish. J Mol Evol59:190203.

Hubbard TJ, Aken BL, Beal K, Ballester B, Caccamo M,Chen Y, Clarke L, Coates G, Cunningham F, Cutts T,

Down T, Dyer SC, Fitzgerald S, Fernandez-Banet J, Graf S,Haider S, Hammond M, Herrero J, Holland R, Howe K,

Johnson N, Kahari A, Keefe D, Kokocinski F, Kulesha E,Lawson D, Longden I, Melsopp C, Megy K, Meidl P,Ouverdin B, Parker A, Prlic A, Rice S, Rios D,Schuster M, Sealy I, Severin J, Slater G, Smedley D,Spudich G, Trevanion S, Vilella A, Vogel J, White S,

Wood M, Cox T, Curwen V, Durbin R, Fernandez-Suarez XM, Flicek P, Kasprzyk A, Proctor G, Searle S,Smith J, Ureta-Vidal A, Birney E. 2007. Ensembl 2007.Nucleic Acids Res 35:D610D617.

Hughes AL. 1999. Adaptive evolution of genes and genomes.New York: Oxford University Press.

Ikuta T, Saiga H. 2007. Dynamic change in the expressionof developmental genes in the ascidian centralnervous system: revisit to the tripartite model and theorigin of the midbrain-hindbrain boundary region. Dev Biol312:631643.

Ilagan R, Abu-Issa R, Brown D, Yang YP, Jiao K, Schwartz RJ,

Klingensmith J, Meyers EN. 2006. Fgf8 is required foranterior heart field development. Development 133:24352445.

Imai KS, Satoh N, Satou Y. 2002. Region specific geneexpressions in the central nervous system of the ascidianembryo. Mech Dev 119:S275S277.

Inoue F, Nagayoshi S, Ota S, Islam ME, Tonou-Fujimori N,Odaira Y, Kawakami K, Yamasu K. 2006. Genomic organi-zation, alternative splicing, and multiple regulatory regionsof the zebrafish fgf8 gene. Dev Growth Differ 48:447462.

Itoh N. 2007. The Fgf families in humans, mice, and zebrafish:their evolutional processes and roles in development,metabolism, and disease. Biol Pharm Bull 30:18191825.

Itoh N, Konishi M. 2007. The zebrafish fgffamily. Zebrafish4:179186.

Itoh N, Ornitz DM. 2004. Evolution of the Fgf and Fgfr genefamilies. Trends Genet 20:563569.

Jacob L, Lum L. 2007. Hedgehog signaling pathway. Sci STKE2007:cm6.

Jeffery WR. 2006. Ascidian neural crest-like cells:phylogenetic distribution, relationship to larval complexity,and pigment cell fate. J Exp Zoolog B Mol Dev Evol306:470480.

Jovelin R, He X, Amores A, Yan YL, Shi R, Qin B, Roe B,Cresko WA, Postlethwait JH. 2007. Duplication and diver-gence offgf8functio

Evolution of Developmental Regulation

Documents

Transcript of Evolution of Developmental Regulation