Sox9 in Stickleback and Zebraﬁsh

DEVELOPMENTAL DYNAMICS 228:480–489, 2003

PATTERNS & PHENOTYPES

Genome Duplication, Subfunction Partitioning, and Lineage Divergence: Sox9 in Stickleback and Zebraﬁsh

William A. Cresko, Yi-Lin Yan, David A. Baltrus, Angel Amores, Amy Singer, Adriana Rodrı´guez-Marı´, and John H. Postlethwait*

Teleosts are the most species-rich group of vertebrates, and a genome duplication (tetraploidization) event in ray-fin fish appears to have preceded this remarkable explosion of biodiversity. What is the relationship of the ray-fin genome duplication to the teleost radiation? Genome duplication may have facilitated lineage divergence by partitioning different ancestral gene subfunctions among co-orthologs of tetrapod genes in different teleost lineages. To test this hypothesis, we investigated gene expression patterns for Sox9 gene duplicates in stickleback and zebrafish, teleosts whose lineages diverged early in Euteleost evolution. Most expression domains appear to have been partitioned between Sox9a and Sox9b before the divergence of stickleback and zebrafish lineages, but some ancestral expression domains were distributed differentially in each lineage. We conclude that some gene subfunctions, as represented by lineage-specific expression domains, may have assorted differently in separate lineages and that these may have contributed to lineage diversification during teleost evolution. Developmental Dynamics 228:480–489, 2003. © 2003 Wiley-Liss, Inc.

Key words: genome duplication; subfunctionalization; transcription factor; chondrogenesis; sex determination; macroevolution; tetraploidization

Received 28 July 2003; Accepted 4 August 2003

INTRODUCTION chavi et al., 2001). Genetic mapping Meyer and Schartl, 1999; Taylor et studies showed that duplicated ze- al., 2001a, b, 2003; Van de Peer et Half of all vertebrate species are te- brafish genes map in duplicated al., 2002; Amores et al., 2003). Did leost fish (Nelson, 1994), the most chromosome segments co-ortholo- genome duplication play a role in speciose and diverse group of vertebrates (Fig. 1). What evolutionary gous to portions of individual human the teleost radiation, and if so, how mechanisms contributed to this re- chromosomes (Amores et al., 1998; did it spur lineage diversification and markably successful explosion of Postlethwait et al., 1998; Gates et al., morphologic variation? biodiversity? A clue comes from the 1999; Woods et al., 2000). The most To clarify the relationship between observation that chromosomally parsimonious conclusion was that genomic and phenotypic complex- diploid teleosts often have several there had been a genome duplica- ity, we must first understand the pro- paralogous copies of single copy tion in the zebrafish lineage, and cesses by which duplicate genes tetrapod genes (Morizot et al., 1991; subsequent analysis shows that this evolve. Initial models of duplicate Ekker et al., 1992, 1997; Akimenko et genome amplification in the ray-fin gene evolution assumed that each al., 1995; Smith et al., 2000, 2001; Ad- fish lineage occurred before the te- gene performs one or few functions amska et al., 2001; Robinson-Re- leost radiation (Amores et al., 1998; (Ohno, 1970; Nei and Roychoudhury,

Institute of Neuroscience, University of Oregon, Eugene, Oregon Grant sponsor: National Institutes of Health; Grant numbers: R01RR10715; 5 F32 GM020892; Grant sponsor: National Science Foundation; Grant numbers: IBN 0236239; IBN 9728587. *Correspondence to: John H. Postlethwait, Institute of Neuroscience, University of Oregon, Eugene, Oregon, 97403-1254. E-mail: [email protected] DOI 10.1002/dvdy.10424

tions that perform a variety of tasks in different tissues and at different developmental times. This view of genes comprising multiple subfunctions has important implications for understanding both the evolution of duplicate gene pairs and the evolution of phenotypic complexity (Force et al., 2003). Instead of the acquisition of novel functions, the partitioning of ancestral subfunctions among descen- dant gene duplicates by the recipro- cal, neutral fixation of degenerative regulatory mutations can contribute to permanent preservation of both copies (Hughes, 1994, 1999; Force et al., 1999; Hughes, 1999; Stoltzfus, 1999). The partitioning process for duplicate gene retention was formalized in the Duplication, Degeneration, Comple- mentation (DDC) model (Force et al., 1999), and the partitioning of ancestral subfunctions has since been dem- onstrated in many duplicated teleost gene pairs (co-orthologs) in single lineages (De Martino et al., 2000; Bingham et al., 2001; Bruce et al., 2001; Chiang et al., 2001; Lister et al., 2001; McClintock et al., 2001, 2002; Altschmied et al., 2002; Yu et al., 2003). A key question for understanding the evolution of gene duplicates and the role of gene duplication in the teleost radiation is whether ancestral subfunctions assorted before or after the divergence of various teleost lineages. Ancestral subfunctions that have not partitioned between duplicates before lineage divergence remain available for subsequent differential partitioning in different lineages, a process that can contribute to reproductive isolation (Lynch and Force, 2000). Sub- functions of paralogous duplicates in the different lineages will be largely identical if partitioning happens soon Fig. 1. A family tree for bony fish (Osteichthys) after (Nelson, 1994). after duplication. If partitioning is slow, however, it may frequently occur independently in different lineages, 1973; Bailey et al., 1978; Takahata and Empirical data, however, show that and as a consequence, duplicate Maruyama, 1979; Li, 1980; Watterson, duplicate pairs are retained at a paralogs will retain different combina- 1983). On the basis of this assumption, higher rate than the classic models of tions of ancestral subfunctions in dif- retention of both copies of dupli- duplicate gene evolution predict (Al- ferent teleost lineages. A central cated genes was hypothesized to be lendorf et al., 1975; Bisbee et al., 1977; question is the following: how often rare and would occur only if one Ferris and Whitt, 1979; Graf and Kobel, does subfunction partitioning occur copy acquired a novel, positively se- 1991; Hughes and Hughes, 1993). This independently in different lineages? lected function (neofunctionalization; may be due to the modular and To date, there is no generalizable an- Walsh, 1995; Sidow, 1996; Cooke et complex nature of many genes, swer, because only a single case of al., 1997; Nadeau and Sankoff, 1997). which comprise numerous subfunc- duplicate gene pairs has been ana- 482 CRESKO ET AL. lyzed in detail in two different teleost drogenesis, studies on Xenopus em- seven positive clones. Restriction en- lineages (Lister et al., 2001; Altschmied bryos indicate that reducing the zyme analysis divided these clones et al., 2002). In this case, subfunction production of Sox9 protein with mor- into two distinct classes, and we partitioning appears to have oc- pholino antisense-oligonucleotides chose two clones from each class to curred before the divergence of ze- inhibits the formation of neural crest, analyze in depth. A combination of brafish, swordtail, and pufferfish lin- the progenitors of craniofacial carti- directed and shotgun sequencing eages at the base of the teleost lage (Spokony et al., 2002). Thus, of the subclones produced two dis- radiation (see Fig. 1). Sox9 is a multifunctional gene, play- tinct genomic contigs 9908 and 6495 To test the hypothesis that dupli- ing important roles in testis determi- base pairs long that each contain a cate genes can be partitioned inde- nation, the formation of neural crest, Sox gene with DNA sequence simi- pendently, we have examined ex- and chondrogenesis. larity to Sox9. Gene prediction anal- pression patterns for two duplicates Zebrafish has two Sox9 genes: ysis using GENSCAN (Burge and Kar- of Sox9 in the threespine stickleback Sox9a and Sox9b (Chiang et al., lin, 1998) showed that each gene Gasterosteus aculeatus and the ze- 2001). Gene phylogenies and ge- comprises three exons and two in- brafish Danio rerio lineages that di- netic mapping show that these are trons and is predicted to encode verged early in the teleost radiation co-orthologs of the tetrapod Sox9 translated products of 464 and 477 (Fig. 1). Stickleback is an emerging gene and likely arose in the ray-fin amino acids that have high se- model system for the study of the genome duplication event (Chiang quence similarity to Sox9. The pre- evolution of development, particu- et al., 2001). Importantly, expression sumptive HMG domain at the N-ter- larly for rapid morphologic changes patterns of these duplicates in ze- minus of the protein is well in bony armor plates and defensive brafish exhibit overlapping subsets of conserved between these pre- spines over short periods of time (Bell the tetrapod expression domains dicted peptides and Sox9 co-or- et al., 1993; Bell and Foster, 1994; Bell (Chiang et al., 2001; Yan et al., 2002) thologs from teleosts (pufferfish, ze- and Orti, 1994; Ahn and Gibson, and together approximate the ex- brafish, rice eel; Bagheri-Fam et al., 1999; Bell, 2001; Peichel et al., 2001; pression pattern of the Sox9 gene in 2001; Chiang et al., 2001; Zhou et al., Gibson, 2002). Thus, these data are mouse, suggesting the partitioning 2002) and tetrapods (chick, mouse not only useful for macro-evolution- of ancestral subfunctions. To deter- and human; Wright et al., 1993; ary studies across divergent lineages mine whether subfunction partition- Wagner et al., 1994; Healy et al., but also lay the groundwork for ex- ing occurred before or after the te- 1999). The C-termini of the predicted amining how microevolutionary leost radiation, we have cloned two proteins, however, are much less change in cartilage and bone reg- Sox9 genes from stickleback, estab- conserved both between the stick- ulatory genes might contribute to lished their phylogenetic relation- leback paralogs and across the the origin of phenotypic variation ships to tetrapod and zebrafish Sox9 other orthologs. among stickleback populations. genes, and compared their expres- A phylogenetic analysis of these Mutations in human, mouse, and sion patterns with those of zebrafish sequences confirmed that stickle- zebrafish have shown that Sox9 is an and tetrapods. We found that the back have at least two different important gene for the regulation of event that produced duplicated te- Sox9 genes (Fig. 2). Both of these se- cartilage formation and the subse- leost Sox9 genes occurred before quences fell squarely inside the Sox9 quent development of cartilage-re- the divergence of stickleback and clade with a very high bootstrap placement bones (Wagner et al., zebrafish lineages, and so did the value (1,000 of 1,000), showing that 1994; Bi et al., 2001; Kist et al., 2002; partitioning of most subfunctions as- they are not orthologs of the closely Yan et al., 2002). Sox9 is a member of sayed by embryonic expression anal- related Sox8 and Sox10 clades. Fur- the Sox protein family of transcrip- ysis. These results have important im- thermore, one stickleback se- tion factors, which contain a SRY-like plications for the generalization of quence clusters within the Sox9a HMG-box that binds and bends DNA experimental results among teleost clade with pufferfish, rice eel, and (Ng et al., 1997; Wegner, 1999). In model systems, for the ways in which zebrafish (Chiang et al., 2001; Zhou mammals, Sox9 is involved in testis duplicated genes evolve, and for the et al., 2002), whereas the other falls determination (Wagner et al., 1994; mechanisms that generated the within the Sox9b clade alongside Vidal et al., 2001) and regulates car- magnificent teleost radiation. pufferfish and rice eel (Bagheri-Fam tilage formation by binding to a cis- et al., 2001; Zhou et al., 2002) with regulatory sequence in COL2A1, the high bootstrap support (728 of 1000). human type II collagen gene (Bell et RESULTS AND DISCUSSION Because zebrafish sox9b fell as an al., 1997; Lefebvre et al., 1997). Sox9 Sox9 Duplication Occurred outgroup to the other teleost Sox9 expression occurs in mesenchymal Before Divergence of genes, it was important to be sure it condensations before and during Stickleback and Zebrafish was not a Sox8, Sox9, or Sox10; there- chondrogenesis, and the expression fore, we identified a zebrafish sox8 pattern mirrors that of COL2A1 in tet- Lineages gene as EST fi23c10 (AW153579) iso- rapods and teleosts (Zhao et al., By using redundant primers de- lated in the Washington University 1997; Healy et al., 1999; Chiang et signed to amplify the first exon of Zebrafish EST Project, and mapped it al., 2001; Yan et al., 2002). In addition Sox9, we screened Fosmid genomic to LG3 in a region with conserved to the critical role played in chon- libraries for stickleback and found synteny with human chromosome Sox9 EXPRESSION IN STICKLEBACK AND ZEBRAFISH 483

Fig. 2. A phylogenetic tree for Sox9-related genes. Numbers are bootstrap values for 1,000 trials. Dre, Danio rerio, zebraﬁsh; Gac, Gasterosteus aculeatus, threespine stickleback; Gga, Gallus gallus, chicken; Hsa, Homo sapiens, human; Mal, Monopterus albus, rice eel; Mmu, Mus musculus, mouse; Tru, Takifugu rubripes, pufferﬁsh. Dre Sox8 AW153579; Dre Sox9a AY090034; Dre Sox9b AAG09815; Dre Sox10 AF402677; Gga Sox9 U12533; Gga Sox10 AF152356; Gga Sox8 AF228664; Hsa Sox8 NP_055402; Hsa Sox9 NP_000337; Hsa Sox10 NP_008872; Mal Sox9a AF378150; Mal Sox9b AF378151; Mmu Sox8 XP_128601; Mmu Sox9 NP_035578; Mmu Sox10 XP_128139; Tru Sox9a AAL32172 (mayfold000587); Tru Sox9b mayfold 000421 (fugu assembly 3 at http://fugu.hgmp.mrc.ac.uk/).

Hsa16p13.3, the location of human (Chiang et al., 2001), the location of guish between evolutionary changes SOX8 (data not shown). This ze- SOX9 (Wagner et al., 1994), we con- that occurred after duplication but brafish sequence clusters strongly clude that Sox9a and Sox9b arose in before lineage divergence and with the other vertebrate Sox8 the hypothesized whole genome those that occurred between lin- genes, which supports the conclu- duplication event in the ray-fin fish eage divergence and the present. sion that zebrafish Sox9b is a Sox9 lineage (Amores et al., 1998; Stickleback embryos are cultured at ortholog, despite its somewhat unex- Postlethwait et al., 1998; Taylor et al., 20°C and are larger and develop pected position in the tree. Impor- 2001a, 2003). Thus, stickleback, ze- more slowly than zebrafish embryos tantly, the branching pattern of the brafish, pufferfish, and rice eel all ap- cultured at 28.5°C, but the embry- Sox9a and Sox9b fish clades with re- pear to have retained both Sox9 onic development of each is similar spect to the tetrapod Sox9 clade, copies formed through the duplica- enough that appropriate compari- while not completely unambiguous, tion of an ancestral fish genome ap- sons can be made between em- shows that the duplication event proximately 300 million years ago. bryos at morphologically similar de- that produced the teleost Sox9a velopmental stages. Stickleback and Sox9b clades occurred before Shared and Divergent embryos at 32 hours postfertilization the divergence of stickleback and (h) show Sox9a expression in cells zebrafish lineages. Because the ze- Expression Patterns of Sox9a around the eye and otic placode brafish Sox9a and Sox9b genes map Because stickleback and zebrafish (Fig. 3A). A zebrafish embryo at ap- in duplicated chromosome seg- Sox9 genes were duplicated before proximately the same stage of de- ments that are co-orthologous to the species lineages diverged, ex- velopment shows strong sox9a ex- the human chromosome Hsa17 pression pattern analysis can distin- pression in the region of the otic 484 CRESKO ET AL.

Fig. 3. Sox9a expression patterns. A,C,E: Sox9a expression in 32 hours postfertilization (h), 70-h, and 92-h stickleback embryos, respec- tively. B,D,F: sox9a expression in 14-h, 24-h, and 48-h zebraﬁsh embryos. a1, mandibular arch; a2, hyoid arch; cb; ceratobranchial arches; e, eye; fb, forebrain; hb, hindbrain; mhb, midbrain–hindbrain border; nc, neural crest; op, otic placode region; pf, pectoral ﬁn; s, somite.

Fig. 4. Sox9b expression patterns. A,C,E: Sox9a expression in 32 hours postfertilization (h), 70-h, and 92-h stickleback embryos. B,D,F: sox9a expression in 14-h, 24-h, and 48-h zebrafish embryos. a1, mandibular arch; a2, hyoid arch; e, eye; fb, forebrain; hb, hindbrain; n. notochord; nc, neural crest; op, otic placode region; pcp, prechordal plate; pf, pectoral fin; s, somite; t, tectum. placode and in the developing in this stage only after prolonged ing the pharyngeal arches, perhaps somites (Fig. 3B). Zebrafish Sox9a ex- staining (data not shown). derived from the Sox9a-positive cells pression is weak around the eye, By 70 h, stickleback embryos con- in the otic region, stain intensely, and reciprocally, stickleback Sox9a tinue to have staining in the eye re- and expression has begun in the expression is weak in the somites, gion, as well as in the forebrain (Fig. pectoral fin bud. Expression is partic- both expression domains appearing 3C). At this time, crest cells populat- ularly strong in the first (mandibular) Sox9 EXPRESSION IN STICKLEBACK AND ZEBRAFISH 485 and second (hyoid) arches and retina of the eye and in the forebrain genes mitfa and mitfb, which pro- weaker in the ceratobranchial and tectum (Fig. 4C). In the hind- duce these alternative isoforms ex- arches. Weak expression of Sox9a brain, Sox9b is expressed in both te- pressed in general in similar patterns appears in the tail somites in 70-h leosts in a striped pattern as for to their tetrapod orthologs (Lister et embryos (Fig. 3C). Most of these ex- Sox9a in zebrafish (Fig. 3F). Sox9b is al., 2001; Altschmied et al., 2002). pression domains are similar in ze- expressed in the pharyngeal arches Thus, for the only two cases studied, brafish embryos of approximately more strongly in the ceratobranchi- the general result is that most ances- the same stage of development, als than in the mandibular and hyoid tral gene subfunctions appear to with strong expression in forebrain, arches in both teleost embryos. At have partitioned before the teleost crest cells populating the arches, tail this stage, Sox9b is expressed weakly radiation. Clearly more cases need somites (Fig. 3D), and pectoral fin in the somites of stickleback but not to be examined in detail before a (data not shown). Two major differ- zebrafish embryos, and, like ze- strong generalization can be made. ences occur in the midbrain–hind- brafish, Sox9b is expressed in the For some ancestral expression do- brain border and the somites, where crest at the end of the tail (Fig. mains, however, each co-ortholog is zebrafish has significantly stronger 4C,D). expressed in a species-specific man- expression than stickleback. By 92 h, expression of Sox9b in the ner. Examples include the strong ex- In 92-h stickleback embryos, Sox9a stickleback brain is mainly in the pe- pression of Sox9a in the zebrafish continues to be expressed in the ripheral parts of the tectum and in midbrain–hindbrain border and pharyngeal arches, and in the outer cells lining the ventricles (Fig. 4E), as somites, but undetected or weak ex- mesenchyme of the pectoral fin, in zebrafish (Fig. 4F). In the fin bud, pression of Sox9a in these domains in more strongly in the outer portion of Sox9b transcripts accumulate in the stickleback, and reciprocally, the the mesenchyme than the central central core of the mesenchyme in strong expression of Sox9b around core, as it is in zebrafish (Fig. 3E,F). In both stickleback and zebrafish. In the eye in stickleback but the weak zebrafish, Sox9a is expressed in a general, the expression patterns of or absent expression of sox9b in the striped pattern in the hindbrain, likely Sox9b are more similar between the same region of zebrafish embryos. In in glial cells as it is in mouse (Pom- two teleosts than are the expression general, for the Sox9 gene pair, the polo and Harley, 2001). These results patterns of Sox9a. species-specific features appear to show that the patterns of Sox9a ex- be mainly quantitative rather than pression in stickleback and zebrafish Conclusions and Future qualitative. Such unshared domains are largely the same, but differ in have the features predicted for reg- detail, with the stickleback gene Directions ulatory subfunctions that had not stronger in the eye region, and the Stickleback and zebrafish embryos partitioned at the time of the stickle- zebrafish gene stronger in the have largely similar embryonic ex- back/zebrafish lineage divergence somites and midbrain–hindbrain pression patterns for Sox9a and but then evolved differently in the border. Expression domains that are Sox9b, and these combined Sox9a two lineages. taxon-specific involve evolutionary and Sox9b domains are generally A third class of expression patterns change in developmental regula- possessed by tetrapod outgroups as appears to be shared by Sox9a and tion after the divergence of stickle- a function of a single Sox9 ortholog, Sox9b in both teleost embryos. Ex- back and zebrafish lineages. suggesting that they are ancestral. pression in the hindbrain and por- Most partitioned expression differ- tions of the pectoral fin mesen- Shared and Divergent ences between the duplicate genes chyme appear to be examples. If are common to both teleost species, higher resolution cell-by-cell analysis Expression Patterns of Sox9b such as strong expression in the shows that individual cells are tran- Like Sox9a, Sox9b is expressed in mandibular and hyoid arches for scribing both duplicates at the same generally similar patterns in stickle- Sox9a and strong expression in the time, then these would be examples back and zebrafish. At 32 h, stickle- trunk crest for Sox9b. These cases dis- of ancestral regulatory elements back embryos express Sox9b around play the behavior expected if an- that remain unpartitioned to the the eye and otic vesicle, and in the cestral regulatory elements driving present. However, the possibility re- neural crest in the head and trunk these expression domains recipro- mains that, even though at a gross (Fig. 4A). This stickleback expression cally partitioned between the dupli- level of examinations the domains of pattern is largely the same as the cate genes in the time interval be- expression appear to be overlap- pattern for the orthologous gene in tween the gene duplication event ping, a more detailed analysis might zebrafish (Fig. 4B), with the main dif- and the divergence of stickleback reveal fine scale partitioning across ference being that crest expression and zebrafish lineages. This is also cells within structures such as the fin in the trunk is comparatively stronger the case for Mitf gene duplicates in mesenchyme. In-depth studies of in zebrafish. teleosts. In tetrapods, there is a sin- apparently overlapping domains In 70-h stickleback embryos, the gle Mitf gene with transcripts that should be completed before con- expression pattern of Sox9b is more are expressed from different promo- clusions are drawn about whether complex and extensive than Sox9a tors and 5Ј exons (Yasumoto et al., domains remain unpartitioned in in the cranial region. At this stage, 1998; Udono et al., 2000). In teleosts, these tissues. Sox9b transcripts accumulate in the in contrast, there are duplicated The results with Sox9 and Mitf 486 CRESKO ET AL. genes suggest that most gene sub- shared across lineages could pro- CopyControl Fosmid Library Produc- functions may have assorted be- vide the unity of type, whereas later tion Kit from Epicentre (catalog no. tween duplicated genes before the lineage-specific, largely quantitative CCFOS110). Each library was made teleost explosion. But did the parti- partitioning, could provide the ge- from a single individual, both col- tioning happen rapidly after dupli- netic fodder that allows lineages to lected from the wild in Alaska. One cation or did it take a long period of acquire different distributions of traits individual, collected from Rabbit time relative to the duplication appropriate for their conditions of Slough, was from an anadromous event and the present? An answer existence, and thus their divergence. population exhibiting extensive depends upon knowing the date of Because genome duplication pro- bony lateral plate and pelvic armor. the genome duplication and the vides a genome’s worth of gene du- The other individual, collected from dates at which teleost lineages di- plicates, even within a largely shared Bear Paw Lake, was from a popula- verged. Unfortunately, the dating of set of partitioning that happened tion of stickleback whose members both events is in question. Molecular soon after duplication, it would still of- have lost almost all bony armor. clock analysis has suggested that fer enormous opportunities for subse- These fish were transported to the the ray-fin fish genome duplication quent independent subfunction parti- University of Oregon stickleback fa- occurred more than 300 million tioning and multiple dimensions along cility and reared for 2 months before years ago (MYA) (Taylor et al., which lineages could diverge. being killed. Stickleback clone in- 2001a) and that the teleost radiation The above considerations are serts average approximately 50 kbp occurred approximately 140 MYA based on a very small data set. in length, and the arrayed libraries (Hedges and Kumar, 2002), but pa- What is now needed is developmen- provide approximately 12ϫ cover- leontologic evidence suggests that tal genetic analysis of a large num- age of the stickleback genome, as- the radiation may have occurred ber of gene duplicates in a variety of suming 600 Mbp in the haploid ge- 200 MYA or more (Santini and Tyler, different teleosts, including models nome (Vinogradov, 1998). 1999). Genomic analysis is needed for microevolution, such as stickle- We screened the library for Sox9 for basally diverging teleosts such as back and swordtails (Walter and Ka- genes by using polymerase chain re- eels, and basally diverging ray-fin zianis, 2001) and models amenable action (PCR) primers designed to fish, such as bowfin and bicher, to to mutagenesis such as zebrafish conserved regions of the first exons further resolve the question. At the and medaka (Grunwald and Streis- of the Fugu rubripes Sox9a and extreme, however, the times be- inger, 1992; Loosli et al., 2000) to un- Sox9b genes. The forward primer was tween genome duplication and the derstand the relative frequency of 5Ј-TGAATCTCCTCGACCCTTACC-3Ј, teleost radiation, and the radiation conserved vs. independent parti- and the reverse primer was 5Ј-TG- to the present, were about equal— tioning, and what role independent CAGCCTGAGCCCACAC-3Ј. Seven approximately 150 million years. partitioning plays in divergence. On independent positive clones were However, we hypothesize that the a practical note, establishing this fre- identified. We sheared and sub- teleost radiation may have oc- quency is very important because it cloned four fosmid clones that were curred sooner after the genome du- will provide a sense of how often we both positive for Sox9, but which had plication than previously thought, should expect model systems that two distinct restriction enzyme diges- with perhaps only 50–100 million have experienced a genome dupli- tion patterns. Shearing was per- years intervening between the two. cation event, such as pufferfish, formed by using a nebulizer and the In either timing scenario, however, medaka, and zebrafish, to unambig- ends of the subclones were repaired most subfunctions appear to have uously provide insight into the func- by using a mixture of T4 polymerase partitioned before the teleost radiation of genes in other organisms, par- and Klenow fragment. The blunted tion, according to currently avail- ticularly humans. At a more subclones were then ligated into the able results. This timing would sup- fundamental level, the hypothesis PCR4-BLUNT cloning vector (Invitro- port the idea that subfunction that subfunction partitioning played gen TOPO Shotgun Subcloning Kit, partitioning after duplication may a role in the teleost radiation makes catalog nos. K7000-01, K7010-01, be a relatively rapid process. the testable prediction that lineages K7050-01, and K7060-01). The aver- The partitioning of ancestral sub- with the most independent partition- age size of inserted DNA was 2–5 kb. functions among duplicated genes ing should be the most phenotypi- Subclones were arrayed into 96-well may contribute to phenotypic evo- cally diverse. Ruling out this hypoth- plates as single clones and then lution. Darwin said that it is generally esis would focus attention on other screened by means of PCR with the acknowledged that all organic be- causes for the most spectacular ver- same first exon primers described ings have been formed on two great tebrate example of biodiversity. above. Positive clones were se- laws - Unity of Type and Conditions quenced from both ends of the vec- of Existence (Darwin, 1859). The re- tor. Additionally, 16 randomly cho- sults of subfunction partitioning in te- EXPERIMENTAL PROCEDURES sen subclones from each Fosmid leost Sox9 co-orthologs suggests the Isolation of Stickleback Sox9a were sequenced. In all, this method hypothesis that subfunction parti- provided full-length genomic se- tioning may contribute mechanisti- and Sox9b quence for Sox9a (9908 bp) and cally to Darwin’s generalization. The We constructed two stickleback Fos- Sox9b (6495 bp), including 5Ј and 3Ј early subfunction partitioning that is mid genomic libraries by using the untranslated regions (UTRs), introns, Sox9 EXPRESSION IN STICKLEBACK AND ZEBRAFISH 487 and perhaps some regulatory ele- most of the first exon. The Sox9b mitf genes associated with differential ments. Each Sox9-positive contig was probe is 1,800 bp in length, covering degeneration of alternative exons in sequenced to an average of 6ϫ cov- most of exons 1 and 3, and all of fish. Genetics 161:259–267. Altschul SF, Madden TL, Schaffer AA, erage, equally distributed in either di- exon 1 and introns 1 and 2. Embryos Zhang J, Zhang Z, Miller W, Lipman DJ. rection (GenBank accession nos. were fixed in 4% paraformaldehyde 1997. Gapped BLAST and PSI-BLAST: a AY351914 and AY351915). The gene at 20°C for 1 week, after which time new generation of protein database structure of the Sox9 gene in each they were dechorionated by hand search programs. Nucleic Acids Res 25: 3389–3402. contig was predicted by using the and processed for in situ hybridiza- Amores A, Force A, Yan Y-L, Joly L, GENSCAN (Burge and Karlin, 1998) tion as described (Chiang et al., Amemiya C, Fritz A, Ho RK, Langeland Web server (http://genes.mit.edu/ 2001; Yan et al., 2002). Stickleback J, Prince V, Wang Y-L, Westerfield M, GENSCAN.html). embryos were staged based on Ekker M, Postlethwait JH. 1998. Ze- morphologic criteria in analogy with brafish hox clusters and vertebrate genome evolution. Science 282:1711– Phylogenetic Analysis the zebrafish staging series (Kimmel 1714. et al., 1995). Amores A, Suzuki T, Yan Y, Pomeroy J, Protein sequence for zebrafish, rice Singer A, Amemiya C, Postlethwait J. eel, mouse, human, and chicken or- 2003. Developmental roles of pufferfish thologs of Sox8, Sox9, and Sox10 were ACKNOWLEDGMENTS Hox clusters and genome evolution in We thank for support NSF grants IBN ray-fin fish. Genome Research (in obtained from Entrez (http://www. press). ncbi.nlm.nih.gov:80/entrez/query. 0236239 for stickleback, NSF grant IBN Bagheri-Fam S, Ferraz C, Daemaille J, fcgi?dbϭNucleotide). The zebra- 9728587 for sex gene research, NIH Scherer G, Pfeifer D. 2001. Compara- fish Sox9a and Sox9b sequences grant R01RR10715 for zebrafish, and tive genomics of the SOX9 region in were used to BLAST (Altschul et al., NIH grant 5 F32 GM020892 for post- human and Fugu rubripes: conservation of short regulatory sequence ele- 1997) against the Fugu database doctoral training support (W.A.C.). ments within large intergenic regions. (http://ensembl.fugu-sg.org/Fugu_ NSF IGERT grant DGE 9972830 sup- Genomics 78:73–82. rubripes/blastview), and we ob- ported summer undergraduate re- Bailey GS, Poulter RTM, Stockwell PA. tained one single strong alignment searchers Diana Bradley, Rebecca 1978. Gene duplication in tetraploid for each paralog. The gene struc- Loda, Melanie Robinson, and Mark fish: model for gene silencing at un- linked duplicated loci. Proc Natl Acad ture for each was determined by Rothgary, who helped screen libraries. Sci U S A 75:5575–5579. analyzing the genomic sequence We also thank Mike Bell, Susan Foster, Bell MA. 2001. Ontogeny and evolution of using the GENSCAN Web server, as John Baker, and Jeff Walker and their lateral plate number in low morph was done for the stickleback genes. students for developing the system of threespine stickleback fish, Gasteros- teus aculeatus. J Morphol 248:205. The published Sox9 protein sequences Alaskan stickleback populations into Bell MA, Foster SA. 1994. The evolutionary and the inferred pufferfish and stickle- the great resource for studies in micro- biology of the threespine stickleback. back protein sequences were ana- evolution and development that it is Oxford: Oxford University Press. 571 p. lyzed by using ClustalX (version 1.82; today. Bell MA, Orti G. 1994. Pelvic reduction in Thompson et al., 1994) to establish the threespine stickleback from Cook Inlet lakes: geographical distribution and in- phylogenetic relationship of these se- REFERENCES trapopulation variation. Copeia:314– quences. Node confidence was as- 325. certained by bootstrapping the data Adamska M, Wolff A, Kreusler M, Witt- Bell MA, Orti G, Walker JA, Koenings JP. set 1,000 times (Felsenstein, 1985), and brodt J, Braun T, Bober E. 2001. Five 1993. Evolution of pelvic reduction in the resulting tree was visualized by Nkx5 genes show differential expres- threespine stickleback fish: a test of competing hypotheses. Evolution 47: using NJPlot (Perrie`re and Gouy, sion patterns in anlagen of sensory or- gans in medaka: insight into the evolu- 906–914. 1996; http://pbil.univ-lyon1.fr/software/ tion of the gene family. Dev Genes Evol Bell DM, Leung KK, Wheatley SC, Ng L-J, njplot.html). 211:338–349. Zhou S, Ling KW, Sham MH, Koopman P, Ahn D, Gibson G. 1999. Expression pat- Tam PPL, Cheah KSE. 1997. SOX9 di- terns of threespine stickleback Hox rectly regulates the type-II collagen Detecting Gene Expression in genes and insights into the evolution of gene. Nat Genet 16:174–178. Stickleback Embryos by In Situ the vertebrate body axis. Dev Genes Bi W, Huang W, Whitworth DJ, Deng JM, Evol 209:482–494. Zhang Z, Behringer RR, de Crombrug- Hybridization Akimenko M-A, Johnson SL, Westerfield ghe B. 2001. Haploinsufficiency of Sox9 M, Ekker M. 1995. Differential induction results in defective cartilage primordia Linearized genomic clones of each of four msx homeobox genes during fin and premature skeletal mineralization. gene were transcribed in vitro to development and regeneration in ze- Proc Natl Acad SciUSA98:6698–6703. make riboprobes using digoxigenin- brafish. Development 121:347–357. Bingham S, Nasevicius A, Ekker SC, Chan- labeled UTP. The Sox9a probe was Allendorf FW, Utter FM, May BP. 1975. drasekhar A. 2001. Sonic hedgehog generated from a subclone in a Gene duplication within the family Sal- and tiggy-winkle hedgehog coopera- monidae: II. Detection and determina- tively induce zebrafish branchiomotor PCR4-Shotgun/Not1-linearized tem- tion of the genetic control of duplicate neurons. Genesis 30:170–174. plate using T3 RNA polymerase, loci through inheritance studies and Bisbee CA, Baker MA, Wilson AC, Iran- whereas the Sox9b probe was syn- the examination of populations. In: dokht HA, Fischberg M. 1977. Albumin thesized from a PCR4-TOPO/Not1-lin- Markert CL, editor. Isozymes. New York: phylogeny for clawed frogs (Xenopus). Academic Press. p 415–432. Science 195:785–787. earized plasmid also using T3 RNA Altschmied J, Delfgaauw J, Wilde B, Dus- Bruce AE, Oates AC, Prince VE, Ho RK. polymerase. The Sox9a probe is 650 chl J, Bouneau L, Volff JN, Schartl M. 2001. Additional hox clusters in the ze- bp in length, covering the 5ЈUTR and 2002. Subfunctionalization of duplicate brafish: divergent expression patterns 488 CRESKO ET AL.

belie equivalent activities of duplicate Healy C, Uwanogho D, Sharpe PT. 1999. divergence after genome duplications hoxB5 genes. Evol Dev 3:127–144. Regulation and role of Sox9 in cartilage early in vertebrate evolution. Genetics Burge CB, Karlin S. 1998. Finding the formation. Dev Dyn 215:69–78. 147:1259–1266. genes in genomic DNA. Curr Opin Hedges SB, Kumar S. 2002. Genomics. Nei M, Roychoudhury AK. 1973. Probabil- Struct Biol 8:346–354. Vertebrate genomes compared. Sci- ity of fixation of nonfunctional genes at Chiang E, Pai C-I, Wyatt M, Yan Y-L, ence 297:1283–1285. duplicate loci. Am Nat 107:362–372. Postlethwait J, Chung B-C. 2001. Two Hughes AL. 1994. The evolution of func- Nelson JS. 1994. Fishes of the world. New sox9 genes on duplicated zebrafish tionally novel proteins after gene dupli- York: Wiley-Interscience. chromosomes: expression of similar cation. Proc R Soc Lond B 256:119–124. Ng L-J, Wheatley S, Muscat GEO, Con- transcription activators in distinct sites. Hughes AL. 1999. Adaptive evolution of way-Campbell J, Bowles J, Wright E, Dev Biol 229:149–163. genes and genomes. New York: Oxford Bell DM, Tam PPL, Cheah KSE, Koop- Cooke JMA, Nowak MA, Boerlijst M, May- University Press. man P. 1997. SOX9 binds DNA, acti- nard Smith J. 1997. Evolutionary origins Hughes MK, Hughes AL. 1993. Evolution of vates transcription, and coexpresses and maintenance of redundant gene duplicate genes in a tetraploid animal, with type II collagen during chondro- expression during metazoan develop- Xenopus laevis. Mol Biol Evol 10:1360– genesis in the mouse . Dev Biol 183:108– ment. Trends Genet 13:360–364. 1369. 121. Darwin C. 1859. The origin of species. Lon- Kimmel CB, Ballard WW, Kimmel SR, Ull- Ohno S. 1970. Evolution by gene duplica- don: John Murray. mann B, Schilling TF. 1995. Stages of tion. New York: Springer-Verlag. De Martino S, Yan Y-L, Jowett T, Postleth- embryonic development of the ze- Peichel CL, Nereng KS, Ohgi KA, Cole BL, wait JH, Varga ZM, Ashworth A, Austin brafish. Dev Dyn 203:253–310. Colosimo PF, Buerkle CA, Schluter D, CA. 2000. Expression of sox11 gene du- Kist R, Schrewe H, Balling R, Scherer G. Kingsley DM. 2001. The genetic archi- plicates in zebrafish suggests the recip- 2002. Conditional inactivation of Sox9: tecture of divergence between rocal loss of ancestral gene expression a mouse model for campomelic dys- threespine stickleback species. Nature patterns in development. Dev Dyn 217: plasia. Genesis 32:121–123. 414:901–905. 279–292. Lefebvre V, Huang W, Harley VR, Good- Perrie` re G, Gouy M. 1996. WWW-query: Ekker M, Wegner J, Akimenko M-A, fellow PN, de Crombrugghe B. 1997. an on-line retrieval system for biologi- Westerfield M. 1992. Coordinate em- SOX9 is a potent activator of the chon- cal sequence banks. Biochimie 78:364– bryonic expression of three zebrafish drocyte-specific enhancer of the pro 369. engrailed genes. Development 116: alpha1(II) collagen gene. Mol Cell Biol Pompolo S, Harley VR. 2001. Localisation 1001–1010. 17:2336–2346. of the SRY-related HMG box protein, Ekker M, Akimenko M, Allende M, Smith R, Li W-H. 1980. Rate of gene silencing at SOX9, in rodent brain. Brain Res 906:143– Drouin G, Langille R, Weinberg E, duplicate loci: a theoretical study and 148. Westerfield M. 1997. Relationships interpretation of data from tetraploid Postlethwait J, Yan Y, Gates M, Horne S, among msx gene structure and func- fishes. Genetics 95:237–258. Amores A, Brownlie A, Donovan A, tion in zebrafish and other vertebrates. Lister J, Close J, Raible D. 2001. Duplicate Egan E, Force A, Gong Z, Goutel C, Fritz Mol Biol Evol 14:1008–1022. mitf genes in zebrafish: complemen- A, Kelsh R, Knapik E, Liao E, Paw B, Ran- Felsenstein J. 1985. Confidence limits on tary expression and conservation of som D, Singer A, Thomson M, Abduljab- phylogenies: an approach using the melanogenic potential. Dev Biol 237: bar T, Yelick P, Beier D, Joly J, Larham- bootstrap. Evolution 39:783–791. 333–344. mar D, Talbot W. 1998. Vertebrate Ferris SD, Whitt GS. 1979. Evolution of the Loosli F, Koster RW, Carl M, Kuhnlein R, genome evolution and the zebrafish differential regulation of duplicate Henrich T, Mucke M, Krone A, Wittbrodt gene map. Nat Genet 18:345–349. genes after polyploidization. J Mol Evol J. 2000. A genetic screen for mutations Robinson-Rechavi M, Marchand O, Es- 12:267–317. affecting embryonic development in criva H, Bardet PL, Zelus D, Hughes S, Force A, Lynch M, Pickett FB, Amores A, medaka fish (Oryzias latipes). Mech Laudet V. 2001. Euteleost fish genomes Yan Y-L, Postlethwait J. 1999. Preserva- Dev 97:133–139. are characterized by expansion of tion of duplicate genes by comple- Lynch M, Force A. 2000. The origin of in- gene families. Genome Res 11:781– mentary, degenerative mutations. Ge- terspecific genomic incompatibility via 788. netics 151:1531–1545. gene duplication. Am Nat 156:590– Santini F, Tyler JC. 1999. A new phyloge- Force AG, Cresko WA, Pickett FB. 2003. 605. netic hypothesis for the order Tetra- Informational accretion, gene duplica- McClintock JM, Carlson R, Mann DM, odontiformes (Teleostei, Pisces), with tion, and the mechanisms of genetic Prince VE. 2001. Consequences of Hox placement of the most fossil basal lin- module parcellation. In: Schlosser G, gene duplication in the vertebrates: an eages. Am Zool 39:10A. Wagner G, editors. Modularity in devel- investigation of the zebrafish Hox para- Sidow A. 1996. Gen(om)e duplications in opment and evolution. Chicago: Uni- loguegroup1genes.Development128: the evolution of early vertebrates. Curr versity of Chicago Press. 2471–2484. Opin Genet Dev 6:715–722. Gates MA, Kim L, Egan ES, Cardozo T, McClintock JM, Kheirbek MA, Prince VE. Smith S, Metcalfe JA, Elgar G. 2000. Iden- Sirotkin HI, Dougan ST, Lashkari D, Aba- 2002. Knockdown of duplicated ze- tification and analysis of two snail gyan R, Schier AF, Talbot WS. 1999. A brafish hoxb1 genes reveals distinct genes in the pufferfish (Fugu rubripes) genetic linkage map for zebrafish: roles in hindbrain patterning and a and mapping of human SNA to 20q. comparative analysis and localization novel mechanism of duplicate gene Gene 247:119–128. of genes and expressed sequences. retention. Development 129:2339– Smith S, Metcalfe JA, Elgar G. 2001. Char- Genome Res 9:334–347. 2354. acterization of two topoisomerase 1 Gibson G. 2002. A genetic attack on the Meyer A, Schartl M. 1999. Gene and ge- genes in the pufferfish (Fugu rubripes). defense complex. Bioessays 24:487– nome duplications in vertebrates: the Gene 265:195–204. 489. one-to-four (-to-eight in fish) rule and Spokony R, Aoki Y, Saint-Germain N, Graf J-D, Kobel HR. 1991. Xenopus laevis: the evolution of novel gene functions. Magner-Fink E, Saint-Jeannet J. 2002. practical uses in cell and molecular bi- Curr Opin Cell Biol 11:699–704. The transcription factor Sox9 is required ology: genetics. In: Kay BK, Peng HB, Morizot DC, Slaugenhaupt SA, Kallman for cranial neural crest development in editors. Methods in cell biology. New KD, Chakravarti A. 1991. Genetic link- Xenopus. Development 129:421–432. York: Academic Press. p 19–34. age map of fishes of the genus Xi- Stoltzfus A. 1999. On the possibility of con- Grunwald DJ, Streisinger G. 1992. Induc- phophorus (Teleostei: Poeciliidae). Ge- structive neutral evolution. J Mol Evol tion of recessive lethal and specific lo- netics 127:399–410. 49:169–181. cus mutations in the zebrafish with ethyl Nadeau JH, Sankoff D. 1997. Compara- Takahata N, Maruyama T. 1979. Polymor- nitrosourea. Genet Res 59:103–116. ble rates of gene loss and functional phism and loss of duplicate gene ex- Sox9 EXPRESSION IN STICKLEBACK AND ZEBRAFISH 489

pression: a theoretical study with appli- amino acid level: a case study based of the zebrafish genome. Genome Res cation to tetraploid fish. Proc Natl on anciently duplicated zebrafish 10:1903–1914. Acad SciUSA76:4521–4525. genes. Gene 295:205–211. Wright EM, Snopek B, Koopman P. 1993. Taylor JS, Van de Peer Y, Braasch I, Meyer Vidal VPI, Chaboissier MC, de Rooij DG, Seven new members of the Sox gene A. 2001a. Comparative genomics pro- Schedl A. 2001. Sox9 induces testis de- family expressed during mouse devel- vides evidence for an ancient genome velopment in XX transgenic mice. Nat opment. Nucleic Acids Res 21:744. duplication event in fish. Philos Trans R Genet 28:216–217. Yan YL, Miller CT, Nissen R, Singer A, Liu D, Soc Lond B 356:1661–1679. Vinogradov A. 1998. Genome size and Kirn A, Draper B, Willoughby J, Morcos Taylor JS, Van de Peer Y, Meyer A. 2001b. GC-percent in vertebrates as deter- PA, Amsterdam A, Chung BC, Wester- Revisiting recent challenges to the an- mined by flow cytometry: the triangu- field M, Haffter P, Hopkins N, Kimmel C, cient fish-specific genome duplication lar relationship. Cytometry 31:100–109. Postlethwait JH. 2002. A zebrafish sox9 hypothesis. Curr Biol 11:R1005–R1008. Wagner T, Wirth J, Meyer J, Zabel B, Held gene required for cartilage morpho- Taylor J, Braasch I, Frickey T, Meyer A, M, Zimmer J, Pasantes J, Bricarelli FD, genesis. Development 129:5065–5079. Van De Peer Y. 2003. Genome duplica- Keutel J, Hustert E. 1994. Autosomal sex Yasumoto K, Amae S, Udono T, Fuse N, tion, a trait shared by 22,000 species of reversal and campomelic dysplasia Takeda K, Shibahara S. 1998. A big ray-finned fish. Genome Res 13:382– are caused by mutations in and gene linked to small eyes encodes mul- 390. around the SRY-related gene SOX9. tiple Mitf isoforms: many promoters Thompson JD, Higgins DG, Gibson TJ. Cell 79:111–1120. make light work. Pigment Cell Res 11: 1994. CLUSTAL W: improving the sensi- Walsh JB. 1995. How often do duplicated 329–336. tivity of progressive and multiple se- genes evolve new functions? Genetics Yu WP, Brenner S, Venkatesh B. 2003. Du- quence alignment through sequence 110:345–364. plication, degeneration and subfunc- weighting, positions-specific gap pen- Walter RB, Kazianis S. 2001. Xiphophorus tionalization of the nested synapsin- alties, and weight matrix choice. Nu- interspecies hybrids as genetic models Timp genes in Fugu. Trends Genet 19: cleic Acids Res 22:4673–4680. of induced neoplasia. ILAR J 42:299– 180–183. Udono T, Yasumoto K, Takeda K, Amae S, 321. Zhao Q, Eberspaecher H, Lefebvre V, de Watanabe K, Saito H, Fuse N, Tachi- Watterson GA. 1983. On the time for Crombrugghe B. 1997. Parallel expres- bana M, Takahashi K, Tamai M, Shiba- gene silencing at duplicate loci. Ge- sion of Sox9 and Col2a1 in cells under- hara S. 2000. Structural organization of netics 105:745–766. going chondrogenesis. Dev Dyn 209: the human microphthalmia-associ- Wegner M. 1999. From head to toes: the 377–386. ated transcription factor gene con- multiple facets of Sox proteins. Nucleic Zhou R, Cheng H, Zhang Q, Guo Y, Coo- taining four alternative promoters. Bio- Acids Res 27:1409–1420. per R, Tiersch T. 2002. SRY-related chim Biophys Acta 1491:205–219. Woods IG, Kelly PD, Chu F, Ngo-Hazelett genes in the genome of the rice field Van de Peer Y, Frickey T, Taylor J, Meyer P, Yan Y-L, Huang H, Postlethwait JH, eel (Monopterus albus). Genet Sel Evol A. 2002. Dealing with saturation at the Talbot WS. 2000. A comparative map 34:129–137.