JOURNAL OF EXPERIMENTAL ZOOLOGY (MOL DEV EVOL) 308B:325-335 (2007)

Development and Evolution of Chordate Cartilage 1 2 1 AMANDA L. RYCHEL • AL'lD BILLIE J. SWALLA ,2 * I Biology Department and Center for Developmental Biology, Uni versity of Washington, Seattle, Washington 98195 2Friday Harbor Laboratories, University of Washington, Friday Harbor, Washington 98250

ABSTRACT Deuterostomes are a monophyletic group of an imals containi ng vertebrates, lancelets, tunicates, hemichordates, ech inoderms, and xenoturbellids. Four out of these six extant groups-vertebrates, lancelets, tunicates, and hemichordates-have pharyngeal gill slits. All groups of deuterostome an imals that have pharyngeal gill slits also have a pharyngeal sk eleton supporting the pharyngeal openings, except tunicates. We previously found that pharyngeal cartilage in hemichordates and cephalochordates contains a fibrillar collagen protein similar to vertebrate type II collagen, but unlike vertebrate cartilage, the in vertebrate deuterostome cartilages ar e acellular. We found SaxE and fibrillar collagen expression in the pharyngeal endodermal cells adjacent to where the cartilages form . These same endodermal epithelial cells also express Pax.l /9 , a marker of pharyngeal endoderm in vertebrates, lancelets, tunicates, and hemichordates. In situ exp eriments with a cephalochordate fibrillar collagen also showed expression in pharyn geal endode rm, as well as the ectoderm and the mesodermal coelomic pou ches lining the gill bars. These results indicate that the pharyngeal en dodermal cells are respon sible for secretion of the cartilage in hemichordates, wherea s in lancelets, all the pharyngeal cells surrounding the gill bars, ectodermal, endodermal, and mesodermal may be responsible for cartilage formation. We propose that endoderm secretion was primarily the ancestral mode of making pharyngeal cartilages in deutero stomes. Later the evolu tionary origin of neural crest allow ed co-option of the gene network for th e secretion of pharyngeal cartilage matrix in the new migratory neural crest cell populations found in vertebrates. J. Exp. Zool. (Mol. Dev. Evo!.) 308B:325- 335, 2007. D 2007 Wiley-Liss, In c.

How to cite this article: Rychel AL, Swalla BJ. 2007. Development and evolution of chordate cartilage. J. Exp. Zool. (Mol. Dev. Evol.) 308B:325-335.

Two separate groups of invertebrate deuteros­ this paper, we use collagen and SaxE expression to tomes, Hemichordata and Cephalochordata (Ian­ show that the gill bar cartilages of hemichordates celets), have cartilages supporting the pharyngeal and lancelets are secreted by the pharyngeal gills (Fig. 1) (Schaeffer, '87; Cameron, 2002; Smith epithelia surrounding them, which is endodermal et al ., 2003 ; Cole and Hall, 2004a; Rychel et a1. , in hemichordates and a combination of endoderm, 2006). Members of the invertebrate chordate ectoderm, and mesoderm in lancelets. group Tunicata do not have a pharyngeal skeleton Cartilage tissu e in vertebrates contains an despite having pharyngeal slits (P ennachetti, '84). extensive and characteristic extracellular matrix Vertebrates also have cartilage supporting their (ECM). Chondrocytes, cells that produce cartilage pharyngeal gills (Fig. 1), and this tissue is secr eted by chondrocytes of neural crest origin (Noden, '78;

Kimmel et a1., 2001 ). A major difference, how ever, Gra nt sponsor: Lerner Gray Grant fro m the American Museu m of is that vertebrate pharyngeal cartilage is a cellular Natural History; Grant spo nsor: NIHDevelopm ental Biology T raining Gr an t; Grant number: T32 HD007183-26A1; Gr ant sponsor: UW NSF tissue while the hemichordate and lancelet phar­ ADVAt\lCE gr ant. yngeal cartilages are acellular (Cole and Hall, 'Correspo ndence to : RJ. Swall a, Box 35 1800, Department of Biology, Univers ity of Was hington, Seat tle, WA 9819 5-1800 . 2004a; Rychel et al. , 2006). Since lancelets and E-mail: bjswallaeeu .washington .edu hemichordates lack chondrocytes in their phar­ Received 15 Septem ber 2006 ; Revised 6 December 2006; Accepted 8 yngeal cartilage, it is not obvious which cells are J anuary 2007 Published online 14 Mar ch 2007 in Wiley InterScience (w ww. responsible for secreting the cartilage matrix. In in terscience.wiley.corn). 001: 10.1002/jez.b.21157. : . ffiWlLEY . InterSciencee © 2007 WILEY-LISS, INC. •• 326 A.L. RYCHEL AND B.J. SWALLA

Ambulacraria Chordata

til til <:J Ll! m til Qj "E <:J 0 -e s: til 0 :::l u +-' s: +-' s .Q ~ u 0 til c u It! .0 'E OJ £ OJ X C a. ~ I :::l OJ OJ I- U >

/'1

(ad~pted from smm; and others,'03) (adapted from David and others,'02) pharyngeal skeleton of collagenous cartilage SoxEregulation of fibrillar collagens in cartilage

Fig. 1. Deuterostome phylogeny. Deuterostomes are composed of two major clades, Ambulacraria (Hemichordata, Echinodermata, and Xenoturbellida) and Chordata (invertebrates Tunicata, Cephalochordata, and Vertebrata). For this study, Paxl/9, fibrillar collagen and SaxE expression was examined in hemichordates and cephalochordates (boxed) that, although distantly related invertebrate deutero stomes, share a similar gill skeleton structure. A whole mount alcian blu e prep of a hemichordate is shown below th e large arrow from Hem ichordata (Smith et aI., 2003). A whole mount alcian blue prep from a zebrafish is shown beneath the large arrow drawn from Vertebrata (David et aI., 2002 ).

matrix, differentiate from mesodermal mesench­ not in vertebrate cartilages (Dziadek, '95), yme cells, or are derived from neural crest cells whereas cartilage is characterized by a specific in the vertebrate head (Goldring et al., 2006). sulfated proteoglycan, chondroitin sulfate, in the Developmentally, prechondrogenic mesenchyme ECM (Hall, 2005). cells have a mesodermal origin, and neural crest SoxE (Sox8/9/1O) is the invertebrate ortholog to cells have an ectodermal origin. Most connective three vertebrate Sox gene duplicates: Sox8, Sox9, tissues are known and understood best in verte­ and SoxlO (Bowles et a1., 2000). The expression of brates, but many invertebrates from diverse phyla these genes is often overlapping and there is some, also have extensive connective tissues, including but not total functional redundancy between Sox8, cartilage (Cole and Hall, 2004a,b). Many of the Sox9, and SoxlO (Kellerer et a1., 2006; O'Donnell invertebrate cartilages described by Cole and Hall et al., 2006). In vertebrates-both agnathans and (2004a,b) are cellular, like vertebrate cartilage, gnathostomes-pharyngeal cartilage development but their developmental origins are largely un­ is characterized by the early expression of Sox9 known. However, the acellular cartilages found in in the neural crest cells that migrate into the hemichordates and cephalochordates are peculiar pharyngeal pouches (Yan et al., 2002; McCauley in that they resemble thickened basal lamina and Bronner-Fraser, 2006; Zhang et al., 2006). (Hyman, '59; Rychel et a1. , 2006) . Basal laminae do Sox9 is responsible for directly activating tran­ not typically contain the fibril-forming collagens scription of several genes coding for proteins that found in vertebrate cartilage, instead they contain compose vertebrate cartilage ECM. One of the mesh-forming collagens, like type IV collagen. most abundant proteins specific to vertebrate Also, basal laminae have abundant structural cartilage ECM is type II collagen (Col2al ). Both proteins such as laminin and nidogen that are Sox9 and SoxlO can directly activate Col2al (Bell

J. Exp . Zool. (Mol. Deu. Eual.) DOl 1O.l002/jez.b DEVELOPMENT AND EVOLUTION OF CHORDATE CARTILAGE 327 et al., '97; Bi et al., '99; de Crombrugghe et al., Park, WA, during low tide by digging in the mud 2000; Sekiya et al., 2000; Suzuki et al., 2006) and in locations where small fecal casts are seen. in Xenopus, both Sox8 and Sox9 can directly Saccoglossus kowalevskii was collected at Smith­ activate SoxlO (O'Donnell et al., 2006). Prechon­ sonian Marine Station at Fort Pierce, FL, by drogenic mesenchymal cells derived from neural digging in the sand to the right of the SMS dock. crest as well as differentiated non-hypertrophic Branchiostoma floridae was ordered from Gulf chondrocytes express Sox9, SoxlO, and Col2al in Specimen Marine Laboratories, Inc. in Panacea, the pharyngeal arches (Ng et al., '97; Zhao et al., FL. Branchiostoma virginiae was collected by '97; Suzuki et al., 2006). In vertebrate cartilage dredging sand with the Sunburst at the Smithso­ development, chondrocyte cells are trapped within nian Marine Station in Fort Pierce, FL. the ECM they secrete, creating a cellular cartilage. Animals were fixed overnight in 4% paraforrnal­ Placodes and neural crest-like cells have not been dehyde in 0.5M NaCI, 0.1 M MOPS, pH 7.5 buffer. described in hemichordates and are lacking in The following day they were either transferred lancelets, even though tunicates clearly contain to 50% then 80% ethanol and stored at - 20'C or placodes and may have neural crest-like cells dehydrated through a 30, 50, 80, 100% ethanol (Swalla, 2007 for review). Yet both hemichor­ series and embedded in polyester wax. Sections of dates and lancelets have fibrillar collagen in their 7 urn were made on a Spencer 829 microtome and pharyngeal cartilages as seen by transmission mounted on gelatin-subbed slides. electron microscope (TEM) and immunohisto­ chemistry (Rahr, '82; Pardos and Benito, '88; In situ hybridization Rychel et al., 2006), even though the gill bar cartilages are acellular (Cole and Hall, 2004b; Slides were deparaffinized in 100% ethanol, and Rychel et al., 2006). rehydrated through an alcohol series into PBT. We have presented evidence previously that Next, the slides were digested with 1 ug/rnl gill slits in hemichordates, lancelets, and verte­ proteinase K for 10 min at 37°C. Then the slides brates are homologous (Rychel et al., 2006; Swalla, were transferred to a 2 mg/ml glycine PBT solu­ 2007) and that gill bars in hemichordates and tion, and post fixed in 4% paraformaldehyde in cephalochordates are composed of a protein PBS for 1 hr. Slides were subsequently dipped into similar to vertebrate type II collagen (Rychel triethanolamine and then treated with triethano­ et al., 2006). Now we present new data that lamine with 0.25% anhydrous acetic acid. Hybridi­ show that the pharyngeal endodermal cells in zation with a DIG-labeled RNA probe was carried hemichordates and lancelets express a fibrillar out at 45°C overnight for Paxl/9 and collagen and collagen and the pharyngeal endoderm in 37°C for Sb-SoxE. Slides were washed extensively hemichordates also expresses the invertebrate in 4X SSC, 0.1% Tween-20, 2X SSC, 0.1% Tween­ homolog to vertebrate Sox8/9/1O, SaxE. Lancelets 20, and then transferred to Solution A (4M NaCI, also show collagen expression in the ectodermal 1M Tris, pH 8, 0.5M EDTA, 0.1% Tween-20). pharyngeal epithelia and mesodermal coelom Slides were treated with Lug/ml RNAse for 20 min lining the atrial surface of the gill bars. at 37°C, transferred back into Solution A, then 2X These data indicate that the evolutionary origin SSC, 0.1% Tween-20, and IX SSC, 0.1% Tween-20, of a collagenous pharyngeal cartilage was near followed by PBT. A 0.1% blocking reagent (Roche, the time of deuterostome diversification, not at Indianapolis, Indiana) in PBT was added for the base of the vertebrates. Our results show 30 min, and a 1:2000 dilution ofa DIG-AP antibody pharyngeal cartilage is made using a similar was applied to the slides overnight. Slides were genetic pathway as in vertebrates but rather than then extensively washed in PBT and transferred neural crest cells secreting the matrix, the role into Buffer 3 pH 8 (100 mM Tris pH 8, 100 mM of matrix secretion lies in the endoderm for NaCI, 50 mM MgClz) and then into Buffer 3 pH 9.5 hemichordates, and partially in the endoderm (100mM Tris pH 9.5, 100mM NaCI, 50mM in lancelets. MgClz). An AP detection solution containing Buffer 3 pH 9.5, 200 mM levamisole, and NBT/ MATERIALS AND METHODS BCIP (Roche, Indianapolis, Indiana) was then applied to the slides and the color reaction was Tissue collection and preparation allowed to proceed for approximately 3-5 hrs. The Enteropneust hemichordates Saccoglossus bro­ reaction was stopped in PBS, and then mophenolosus were collected at Bay View State the slides were dehydrated, counterstained with

J. Exp. Zool. (Mol. Dev. Evol.) DOl lO.l002/jez.b 328 A.L. RYCHEL Al"1D B.J. SWALLA

Eosin-Orange G, transferred into xylene, and sequences from the fibrillar collagen triple helical mounted with Permount (Fisher, Houston, Texas). domains were aligned in Clustal X (Jeanmougin et al., '98). Parsimony heuristic searches with Cloning Sb-SoxE with degenerate primers 1,000 random additions were performed in PAUP* 4.0b10 (Swofford, '99) and bootstrapping was done The following degenerate primers were designed with 1,000 pseudoreplicates. using an amino acid alignment of the HMG domain of SoxE genes from several vertebrates and invertebrates. SoxE F: CCNATGAAYGCNT RESULTS TYATG, and SoxE R: TCNGGRTGRTTYTTRTG. Fibrillar collagen expression was examined These primers amplified an approximately 700 bp in two groups of invertebrate deuterostomes for fragment using genomic DNA from S. bromophe­ this study: an enteropneust hemichordate and a nolosus using the following program: 94°C for cephalochordate (Iancelet) (boxes; Fig. 1). Herni­ 4 min, then 35 cycles of 94°C for 30 sec, 42°C for chordates and lancelets were chosen for this study 30 sec, 72°C for 1 min, with a final extension of because they are the only invertebrate deuteros­ 72°C for 10 min. The PCR reaction contained tomes that have an acellular pharyngeal cartilage. 2 mM MgCh, 0.2 mM dNTPs, 0.5 11M each primer, These are considered cartilages because of the fact and 0.75 U of Taq polymerase. Bands were cut that they resist KOH digestion (Smith et al. , out and purified using Sephaglas (Amersham ; 2003), stain blue green using a Milligan's tri­ Piscataway, NJ), and cloned into a pCR II plasmid chrome histology stain, and also stain with a (Invitrogen; Carlsbad, CA). The clone's identity vertebrate type II collagen antibody (Rychel et al., was confirmed with a BLAST search and gene tree 2006). The main component of vertebrate phar­ analysis. The Sb-SoxE sequence has been depos­ yngeal cartilage is fibrillar collagen, so we exam­ ited in Genbank with the accession number ined collagen expression in the pharyngeal EF156377. cartilages of juvenile and adult hemichordates and lancelets. Gill bars are continuously added at Fibrillar collagen clones and gene trees the posterior so it is possible to look at developing A fibrillar collagen clone containing a partial gill bars in adults. portion of the triple helical domain and the Fibrillar collagens are best characterized in complete C terminal domain from S. kowalevskii vertebrates, but so far only a single fibrillar was obtained from John Gerhart and Marc collagen has been found in hemichordates and Kirschner. Two RNA probes were made from this cephalochordates (Fig. 2). Trace lancelet se­ clone: one full length and one with only C terminal quences in GenBank were blasted using the domain sequences. The C terminal probe was conserved C terminal domain of collagen and the made by linearizing the plasmid cDNA with the sequences retrieved were all highly similar to the restriction enzyme Pst I, which cut at a unique site clone we report here, suggesting lancelets have a 100 nucleotides 5' of the C terminal domain. This single fibrillar collagen gene. Expressed sequence probe contains 100 nucleotides 5' of the C terminal tags from several different development stages of domain, 700 nucleotides in the C terminal domain, S . kowalevskii have suggested a single fibrillar and 1,040 nucJeotides in the 3' UTR. Sense and collagen gene in hemichordates, as well (John antisense probes were synthesized using a DIG Gerhart, personal communication). Gene trees RNA labeling kit SP6/T7 (Roche, Indianapolis, built with all collagen genes known from several Indiana). Several clones from a single B. floridae invertebrates, including mosquitos and sea urch­ fibrillar collagen gene were obtained from Georgia ins, show three types of ancestral fibrillar col­ Panopoulou (Panopoulou et al., 2003). The six lagens in animals, the A, B, and C clades clones represented the sequence from a partial (Aouacheria et al., 2004; Rychel et al., 2006; Wada triple helical domain and the complete C terminal et al., 2006). The fibrillar collagens found in domain. Clones from both the triple helical hemichordates and cephalochordates belong to domain and C terminal domain were used for in the A clade of fibrillar collagens (Fig. 2) (Rychel situ hybridization. The parsimony gene tree was et al., 2006). This clade also contains vertebrate constructed as previously reported (Rychel et al., type I (Collal) and type II (Co2al) collagens. 2006) and two new type II collagen sequences from Hemichordates and cephalochordates have at least lamprey (Zhang et al., 2006) were added to the one A clade collagen, whereas echinoderms (sea dataset. Briefly, complete or partial amino acid urchins) have at least three closely related copies,

J. Exp. Zool. (Mol. Deu. Euol.) DOl lO.1002/jez.b DEVELOPMENT AND EVOLUTION OF CHORDATE CARTILAGE 329

99 ,------Bclade

.------Collal t------Colla2 t------Col3al 91 .------Col2a1 lampreyCol2a1a

60 lampreyCol2al b t------Col5a2 L...- Ciona

Urchin2 Urchin5

54 L..- Urchinl

84 L..- Hemichordate

t------Cephalochordate .------Bee 62 .------Abalonel 78 1----__ Abalone2

L...- Polychaete 53 ------Cclade

L..- - Cniderian

Fig. 2. Fibrillar collagen gene tree. Parsimony analysis yielded a tree with three clades (A-C), but only the A clade is shown here in detail with bold branches. Bootstrap values greater than 50 are shown on the tree, and clades with bootstrap values with less than 50% were collapsed. The A clade contains the vertebrate collagens, ColIa l , Colla2, Col3al, Col2al, Col5a2 as well as several invertebrate collagens from members of the Lophotrochozoa (abalone and polychaete) and Ecdysozoa (bee). Type II collagens from humans and lampreys and the fibrillar collagens from hemichordates and cephalochordates are shown in bold. Invertebrate deuLerostome sequences from cephalochordates, hemichordates, and sea urchins group with the vertebrate deuterostome sequences, but not strongly with a particular copy. Multiple copies of sea urchin (2, 5, 1), abalone tL, 2), and lamprey type II collagens (Col2ala and Col2alb) group together, suggesting independent gene duplications in these groups.

urchinl, 2, and 5 (Fig. 2). Sea urchins also have both (Ogasawara et al., '99), is expressed in S. kowa­ Band C clade collagen (Aouacheria et al., 2004). The levskii in the pharyngeal region of the young adult sea urchin genome has been completed (Sea Urchin worm (Fig. 3C and D). A section similar to Figure Genome Sequencing Consortium, 2006), whereas 3C and D was hybridized with the hemichordate the hemichordate and cephalochordate genomes fibrillar collagen (Fig. 3F and G) and Sb-SoxE (Fig. have not, so it is possible that more collagens may 31 and J). The gill bar cross sections in Figure 31 be identified through genome analysis. and J are a different shape than those in Figure We examined the expression of fibrillar collagen 3C-G because they were located slightly more in two harrimaniid hemichordates, S. kowalevskii ventral in the animal. Figure 3F and G show a and S. bromophenolosus, by in situ hybridization. positive purple signal for fibrillar collagen RNA in A diagram of a hemichordate is shown in the pharyngeal endoderm in the same cells that Figure 3A, with area sectioned for in situ express Paxl/9, shown in the panel above. Most of hybridization indicated with a red line. A lancelet the RNA is localized in the endodermal epithelium is shown in Figure 3B, also with the area sectioned that is closest to the skeletal elements (Fig. 3D and for in situ hybridization indicated with a red line. G). Results were similar in the closely related Paxl/9, a marker for pharyngeal endoderm species S. bromophenolosus (data not shown). We

J. Exp. Zool. (Mol. Deu. Euol.) DOl lO.l002/jez.b 330 A.L. RYCHEL AND B.J. SWALLA

A Hemichordate B Cephalochordate L-N ~

D E GB

GB

Fig . 3. Fibrillar collagen, SaxE, and Paxl/9 expression. (A) Diagram of a hemichordate worm with the areas sectioned and examined in (C-K) indicated with a red line. (B) Diagram of a cephalochordate with the areas sectioned and examined in L-N indicated with a red line. (C) Lower magnification in situ for Pax 119 collagen around hemichordate (S . kawalevskii) gill bars. (D) Higher magnification of boxed area in (C) showing the purple signal for Paxl/9 RNA within th e cells surrounding the eosin stained gill bar (GBl. (E) Higher magnification of a control section treated with a sense probe for hemichordate Pax1l9 and stained with eosin. There is no purple label in the cells surrounding the gill bars. (F) Lower magnification in situ for Paxl/9 around hemichordate (5. kowalevskii) gill bars. (G) Higher magnification of the boxed area shown in F. The purple signal for fibrillar collagen RNA expression is within the cells surrounding the gill bar, which is counterstained with eosin pink. (H) Higher magnification of a control section treated with a sense prob e for hernichordate fibrillar collagen. There is no purple label in the cells surrounding the gill bars. (1) Lower magnification in situ for SaxE around hemichordate (5 . bromophenolosus) gill bars, (J ) Higher magnification of the boxed area shown in (I). (J ) The purple signal for SaxE is present within the cells around th e gill bar. (K) Higher magnification of a control section treated with a sense probe for SoxE. (K). (L ) Lower magnification in situ for fibr illar collagen around cephalochordate (B . uirginiae) gill bars. (M) Higher magnification of the boxed area in (I) showing the purple signal for fibrillar collagen within the cells surroun ding the eosin sta ined gill bar. (N ) Higher magnification of a control section treated with a sense probe for cephalochordate fibrillar collagen and stained with eosin. There is no purple label in the cells surrounding the gill bars. GB indicates gill bar matrix. Scale bars = 50 urn (C, F, I); 20 um (D- E, G, H, J, K, M, N).

also examined the expression of Sb-SoxE in strong Sb-SoxE expression is seen in oocytes (not S. bromophenolosus, and expression is seen in shown). the pharyngeal endodermal cells, similar to Paxl/9 In the adult lancelets B, floridae and and fibrillar collagen (Fig. 31 and J), Additionally, B. virginiae, fibrillar collagen expression was very

J. Exp. Zool. (Mol. Dev. Evol.) DOl lO.lO02/jez.b DEVELOPMENT AND EVOLUTION OF CHORDATE CARTILAGE 331

strong in the pharyngeal endodermal cells, the gill cartilages of hemichordates and cephalochor­ pharyngeal ectoderm and the mesodermal coelo­ dates are remarkably similar in morphology mic pouches on the atrial side of the gill bars (Schaeffer, '87; Smith et al., 2003 ) and structure (Fig. 3L and M). (Rychel et al., 2006 ). Furthermore, as in verte­ The acellular gill bar cartilage of hemichordates brates the cartilage of hernichordate worms is appears to form as a result of SaxE and collagen resistant to KOH digestion and stains with alcian expression in the endoderm, followed by secretion blue dye, (Fig. 1)(Smit h et al., 2003 ). Fibrillar of collagens into the ECM . The acellular cartilage collagen expression in hemichordate and lancelet of lancelets is also formed by the secretion of pharyngeal cartilage cells suggests that pharyn­ fibrillar collagen by the pharyngeal endoderm geal collagenous cartilages evolved well before the along with the pharyngeal ectoderm and meso­ vertebrate ancestor diverged from more basal derm. Collectively, these results indicate that deuterostomes. the deuterostome ancestor also was likely to We explored the evolutionary origins of phar­ have had a pharyngeal cartilage composed of yngeal skeletons by examining the invertebrate fibrillar collagen that may have been secreted by deuterostomes (lancelets and hemichordates) that endoderm. have an elaborate pharyngeal skeleton (Smith et al., 2003). Both groups live in the sand or DISCUSSION mud and filter water in a burrow (Kardong, 2006 ). In contrast, tunicates live on hard substrates, Until very recently, it was thought that only and have an extremely divergent adult body jawed vertebrates had type II collagen in their plan, but are classified as chordates based on pharyngeal cartilage, since lampreys were under­ larval characteristics (Swalla, 2007 ). Traditionally stood to have head and pharyngeal cartilages cephalochordates have been placed as the inverte­ composed of a novel protein, lamprin (Robson brate chordate sister group to vertebrates because et al., '93; Wright et aI., 2001 ). However, recent of certain aspects of the genome organization tHox studies have shown that lamprey embryos do genes) as well as the presence of somites (Garcia­ express type II collagen in pharyngeal cartilage, Fernandez and Holland, '94; Ikuta et aI., 2004; and that collagen expression is dependent on the Passamaneck and Di Gregorio, 2005) . In recent expression of the up stream transcription factor genome phylogenetic analyses, tunicates were SaxE (McCauley and Bronner-Fraser, 2006; Zhang moved from the most basal member of the et al ., 2006 ). Now, similar overlapping expression chordates to the closest invertebrate relative to patterns have been shown for hemichordates with the vertebrates (Blair and Hedges, 2005; Delsuc a fibrillar collagen and SaxE, indicating that this et al., 2006). Although this hypothesis poses even shallow gene network has been conserved in the stronger evidence for a worm-like deuterostome formation of pharyngeal cartilage since the stem ancestor (Cameron et al., 2000), these results deuterostomes. should be interpreted with caution since mito­ The viscerocranium or pharyngeal region of the chondrial genome analyses place cephalochordates vertebrate head skeleton was hypothesized to have as the sister group to vertebrates (Bourlat et al ., evolved from the pharyngeal gill skeleton of an 2006) . Increased numbers of echinoderm and invertebrate ancestor similar to the lancelet hemichordate sequences were recently added to pharyngeal skeleton (De Beer, '37). This hypoth­ nuclear gene analyses, giving a tree with a well­ esis was later refuted in light of data that supported clade that includes vertebrates, cepha­ indicated that the pharyngeal skeleton of lance­ lochordates, and tunicates, but the relationships lets, hagfish, and lampreys did not contain within are still unresolved (Bourlat et al., 2006 ). collagen structural proteins (Wright et al., 2001 ). Wada et al .(2006) found fibrillar collagen TEM studies and our own antibody studies expression in the floor plate nerve cells of the suggest that cephalochordates as well as hemi­ neural canal in larval cephalochordates, and we chordates have collagenous pharyngeal cartilages found similar expression in lancelet adults. Type (Ra hr, '82; Pardos and Benito, '88; Rychel et al., II collagen (Cal2al) is known to be expressed in 2006 ) and lend stronger support to the idea that a variety of vertebrate embryonic neural tubes the viscerocranium evolved from ancestral phar­ (Kosher and Solursh,'89; Cheah et al. , '91; Yan yngeal cartilages like those found in hemichor­ et al ., '95; Ng et al ., '97), yet its function there has dates and lancelets. Despite their phylogenetic never been explored. The expression of a lancelet distance from each other (Fig. 1), the pharyngeal type A fibrillar collagen in many of the same places

J. Exp. Zool. (Mol. Dev. Evol.) DOl 1O.l002/jez.b 332 A.L. RYCHEL AND B.J. SWALLA

as the vertebrate type II collagen (Cal2al) lends water vascular system independently of hemichor­ strength to the speculation that the ancestral dates (Swalla, 2007) . The tunicate innovation was clade A fibrillar collagens are functionally similar the evolution of a cellulose tunic that serves to to vertebrate type II collagen. A recent study protect the animal and its gill pharynx (Zeng and mapped expression patterns in vertebrate bone, Swalla, 2005). Tunicates also filter feed, and like cartilage, and notochord on a deuterostome fibril­ some crinoids, are often attached to a hard lar collagen gene tree, and concluded that the substrate, whereas others are planktonic (Zeng recruitment of fibrillar collagen genes to each of and Swalla, 2005) . An alternative hypothesis to these tissues occurred independently in the three homology between cephalochordate and hemichor­ major clades (A, B, and C) (Wada et al., 2006) . date pharyngeal cartilages is that they are in­ An alternative explanation that we favor is that dependently derived. We feel this hypothesis is an ancestral SaxE gene regulated fibrillar collagen unlikely since the type of gill skeleton found in expression (A, B, and C) and that the regulatory hemichordates and cephalochordates is similar to module was selectively retained in cartilage the pharyngeal skeleton preserved in stem group tissues in the deuterostome ancestor (Fig. 1). vertebrates like the fossil Yunnanozoon (Chen The two deuterostomes that have independently et al., '95). Like hemichordates and cephalochor­ lost their pharyngeal cartilages (Swalla, 2006), dates, Yunnanozoon and similar stem vertebrates echinoderms and tunicates, have fibrillar col­ were thought to be marine benthic filter feeders, lagens that may have evolved new expression yet they had an enlarged head region that included patterns. In echinoderms, there are independent eyes and a brain (Chen et al., '95). duplications of A clade fibrillar collagens. These We are investigating evolutionary relationships and other copies of fibrillar collagens in the B and between the skeletal elements found in hemichor­ C clades (Aouacheria et al., 2004) may have been dates and lancelets and pharyngeal cartilages in co-opted into a role in biomineralization of the vertebrates. Mayor et al. ('99) considered lancelets calcium carbonate endoskeleton (Livingston et al ., as completely lacking head skeleton and crafted 2006). It is possible that the echinoderm collagens a hypothesis that describes how cranial neural coevolved with other genes involved in biominer­ crest cells developing from the ectoderm gained alization of the echinoderm calcium carbonate the potential to make cartilage. In their model, endoskeleton, such as spicule matrix proteins mesoderm cells first had the genetic program for (Benson et al., '87; George et al., '91; Katoh-Fukui cartilage production, and this program was trans­ et al ., '91; Harkey et al., '95 ; Killian and Wilt, '96; ferred to the ectoderm-this is a transfer of Lee et al., '99; Illies et al ., 2002). chondrogenic potential from mesenchyme to Hemichordate and cephalochordate pharyngeal epithelia. Our results suggest the opposite, that cartilage is acellular and appears to be wholly the endodermal epithelial cells transferred the or partially secreted from endodermally derived ability to make cartilage to mesenchymal neural epithelia. This type of ECM formation is rare in crest cells and head mesoderm. vertebrates, but is comparable with how acellular In the context of the pharyngeal skeleton and bone forms in the vertebral column of teleost fish the transition from invertebrate chordates to (Ekanayake and Hall, '87, '88; Fleming et al., vertebrates, we speculate that the task of making 2004) . This sort of acellular bone found in modern cartilage and the gene networks required for teleosts is secondarily derived, but acellular bone cartilage formation were transferred from the preceded cellular bone in vertebrate gnathostome pharyngeal epithelium (mostly endoderm) to the evolution (Donoghue et al. , 2006), so acellular neural crest cells. The acellular manner of making matrix secretion was likely to be common among pharyngeal cartilage in hemichordates and cepha­ stem group jawed vertebrates. lochordates is likely the ancestral mode of The deuterostome ancestor has been hypothe­ constructing pharyngeal cartilage in the deutero­ sized to be a filter feeding benthic worm with gill starnes. Next, in vertebrates, neural crest cells slits and an acellular cartilaginous pharyngeal invaded this matrix and may have been influenced skeleton (Rychel et al., 2006). This hypothesis by it. Finally, endoderm evolved reduced suggests that there have been evolutionary losses matrix secretion, but still continued to pattern in the deuterostomes. Tunicates have lost a gill neural crest migration (Fig. 4) (Graham et al ., skeleton and echinoderms and xenoturbellids have 2005) . We will continue to test and refine these lost both gill bars and gill slits. Echinoderms evolutionary hypotheses experimentally with evolved a calcium carbonate endoskeleton and a extant animals.

J. Exp. Zool. (Mol. Dev. Eval.) DOl 1O.l002/jez.b DEVELOPMENT AND EVOLUTION OF CHORDATE CARTILAGE 333

A for their help in collecting Saccoglossus kowalevs­ kii. We would also like to thank John Gerhart, 0 ~ _-.~-~--- Marc Kirschner, and Georgia Panopoulou for - generously providing fibrillar collagen clones used in this study. A.L.R. was supported by NRSA Grant Number T32 HD007183-26A1 from NIH­ ~ 1 NICHD, and by the American Museum of Natural

B History Lerner-Grey Marine Research Fund. This research was partially funded by the UW 0 NSF ADVANCE program. This is contribution ------number 681 of the Smithsonian Marine Station at Fort Pierce, FL.

./ LITERATURE CITED

c Aouacheria A, Cluzel C, Lethias C, Gouy M, Garrone R, Exposito JY. 2004. Invertebrate data predict an early 0 emergence of vertebrate fibrillar collagen clades and an - - - - anti-incest model. J Bioi Chern 279:47711-47719. Bell DM, Leung KK, Wheatley SC, Ng LJ, Zhou S, Ling KW, I ~ ~ Sham MH, Koopman P, Tam PP, Cheah KS. 1997. SOX9 J ~ directly regulates the type-Ll collagen gene. Nat Genet 16: ./ • n • 174-178. Benson S, Sucov H, Stephens L, Davidson E, Wilt F. 1987. A lineage-specific gene encoding a major matrix protein of the sea urchin embryo spicule. I. Authentication of the cloned gene and its developmental expression. Dev Bioi 120: 499-506. Bi W, Deng JM, Zhang Z, Behringer RR, de Crombrugghe B. 1999. Sox9 is required for cartilage formation. Nat Genet 22:85-89. Blair JE, Hedges SB. 2005 . Molecular phylogeny and diver­ Fig. 4. Evolutionary model for the evolution of pharyngeal gence times of deuterostome animals. Mol Bioi Evol 22: cartilage in deuterostomes. (A) In hemichordates, the phyar­ 2275-2284. yngeal endoderm (yellow) is the tissue layer responsible for Bourlat SJ, Juliusdottir T, Lowe CJ, Freeman R, Aronowicz J, secretion of the acelluar cartilage (green) in between the gill Kirschner M, Lander ES, Thorndyke M, Nakano H, Kohn slits. (B) The situation is similar in lancelets as in A, except AB, Heyland A, Moroz LL, Copley RR, Telford MJ. 2006. the pharyngeal epithelium has an ectodermal component on Deuterostome phylogeny reveals monophyletic chordates the atrial side and a mesodermal component in the gill bar and the new phylum XenoturbeJlida. Nature 444:85-88. coelom that is involved in matrix section in addition to the Bowles J, Schepers G, Koopman P. 2000. Phylogeny of the pharyngeal endoderm. (C) The hypothetical intermediate is SOX family of developmental transcription factors based on an animal that also has endodermal (yellow), ectodermal sequence and structural Indicators. Dev Bioi 227:239-255. (blue), and mesodermal (red) components to matrix secretion Cameron CB. 2002. Particle retention and flow in the pharynx like lancelets, except this animal also has migratory neural of the enteropneust worm Harrimania planletophilus: The crest cells (blue circles) invading the pharynx that become filter-feeding pharynx may have evolved before the chor­ influenced by the extracellular matrix (green). (D) The dates. Bioi Bull 202:192-200. current vertebrate situation of pharyngeal cartilage forma­ Cameron CB, Garey JR, Swalla BJ. 2000. Evolution of the tion, where the migratory neural crest cells (blue circles) chordate body plan: new insights from phylogenetic ana­ themselves secrete the extracellular matrix that forms the lyses of deuterostome phyla. Proc Nat! Acad Sci USA 97: cartilages. The mesoderm (red) in the pharyngeal pouch does 4469-4474. not participate in pharyngeal cartilage formation, but does Cheah KS, Lau ET, Au PK, Tam PP. 1991. Expression of the form musculature of the pharynx. mouse alpha HII) collagen gene is not restricted to cartilage during development. Development 111:945-953. Chen JY, Dzik J, Edgecombe GD , Ramskold L, Zhou GQ. 1995. A possible Early Cambrian chordate. Nature 377:720-722. ACKNOWLEDGMENTS Cole AG, Hall BK. 2004a. Cartilage is a metazoan tissue; integrating data from nonvertebrate sources. Acta Zoologica Thanks to Hugh Reichardt and Ashleigh 85:69-80. Cole AG, Hall BK. 2004b. The nature and significance of Symthe for collecting Branchiostoma virginiae at invertebrate cartilages revisited: distribution and histology the Smithsonian Marine Station at Fort Pierce, of cartilage and cartilage-like tissues within the Metazoa. FL. Woody Lee and Gisele Kawauchi are thanked Zoology 107:261-273.

J. Exp. Zool. (Mol. Dev. Euol.) DOl 1O.lO02/jez.b 334 A.L. RYCHEL AND B.J.SWALLA

David NB, Saint-Etienne L, Tsang M, Sch illing TF, Rosa FM. Benson SC, Sucor HM, Davidson EH . 1991. The corrected 2002 . Requirement for endoderm and FGF3 in ventral head structure of the SM50 spicule matrix protein of Strongylo­ skel eton formation. Development 129:4457-4468. centrotus purpuratus . Dev BioI 145:201-202. De Beer GR. 1937. The developm ent of the ver tebrate skull. Keller er S, Schreiner S, Sto lt CC, Scho lz S, Bosl MR, Wegner Oxford: Oxford University Press. 408p . M. 2006. Replacement of the Soxl 0 transcription factor by de Crombrugghe B, Lefebvr e V, Behringer RR, Bi W, Sox8 reveals incomplete func tional equivalence. Develop­ Murakami S, Hu ang W. 2000. Transcriptional mech anisms ment 133:2875-2886. of chondrocyte differentiation. Matrix BioI 19:389-394. Killian CE, Wilt FH. 1996. Characterization of the proteins Delsu c F, Brinkmann H, Chourrout D, P hilippe H. 2006 . comprising the integral matrix of Strongylocentrotus Tunicates an d not cephalochordates are the closest living purpuratus embryonic spicules. J BioI Chern 271: re latives of vertebrates. Nature 439:965-968. 9150-9159. Donoghue PCJ, Sansom IJ, Downs JP. 2006 . Early evolution Kimmel CB, Miller CT, Moens CB. 2001. Specification and of vertebrate skeletal tissues and cellu lar interactions, an d morphogenesis of the zebrafish larval head skeleto n. Dev the canalization of skeletal development. J Exp Zool B: Mol BioI 233:239-257. Dev Evol 306B :278-294. Kosher RA, Solursh M. 1989. Widespread distribution of type Dziadek M. 1995. Role of laminin-nidogen comp lexes in II collagen during embryonic chick deve lopment. Dev BioI basement membrane formation during em bryonic develop­ 131:558-566. ment.Experientia 51:901-913. Lee YH, Br itte n RJ, Davidson EH. 1999. SM37, a skeletogenic Ekanayake S, Hall BK. 1987. Th e development of acellularity gene of the sea urchi n emb ryo linked to the SM50 gen e. Dev, of the vertebral bon e of the Japanese medaka, Oryzias Growth Differ 41:303-312. latipes (Teleostei; Cyprinidontidae). J Morphol 193:253-261. Livingston BT, Killian CE, Wilt F, Cameron A, Landrum MJ, Ekanayake S, Hall BK. 1988. Ultrastructure of the osteogen­ Ermolaeva 0 , Sapojnikov V, Maglott DR, Buchanan AM, esis of acellular vertebral bone in the Japanese medaka, Ettensohn CA. 2006 . A genome -wide analysis of biominer­ Oryzias latipes (teleostei, cyprinidontidae). Am J Anatomy alization-related proteins in the sea urchin Strongylocen­ 182:241-249. trotus purpuratus. Dev Bioi 300:335-348. Fleming A, Keynes R, Tannahill D. 2004 . A central role for the Mayor R, Young R, Vargas A. 1999. Deve lopment of neural notochord in vertebral patterning. Development 131: cre st in Xenopus. Curr Top Dev Bioi 43:85-113. 873-880. McCauley DW, Bronner -Fraser M. 2006 . Import ance of SoxE Garcia-Fern andez J , Holland PWH. 1994. Arch etypal organi­ in neural crest development and the evolution of th e zation of the amphioxus Hox gene clus ter. Nature 370: ph ary nx. Nature 441:750-752. 563- 566. Ng LJ, Wheatley S, Muscat GE, Conway-Campbell J , Bowles J , George NC, Killian CE, Wilt FH. 1991. Characterization and Wright E, Bell DM, Tam PP, Cheah KS, Koopman P. 1997. expression of a gene encoding a 30.6-kDa Strongylocentrotus SOX9 binds DNA, activates transcription, and coexpresses purpuratus spicule matrix protein. Dev BioI 147:334-342. wit h type II collagen during chondrogenesis in the mouse. Goldring MB, Tsuchimochi K, Ijir i K. 2006. The control of Dev BioI 183:108-121. chondrogenesis . J Cellu lar Biochem 97:33-44. Noden DM. 1978. The control of avian cephalic neural crest Graham A, Okabe M, Quinlan R. 2005 . The role of the cytodifferentiation: 1. Skeletal an d connective tissues.Dev endoderm in the development and evolution of the phar­ BioI 67:296-312. yngeal arches. J Anatomy 207:479-487. O'Donnell M, Hong CoS, Huang X, Delnicki RJ , Saint-Jeannet Hall BK 2005 . Cartilage. Bones and cartilage: developmental J -P . 2006 . Functional analysis of Sox8 during neural crest and evolutionary skeletal biology. San Diego: Elsevier development in Xenop us. Developm ent 133:3817-3826. Academic Press. p 33-47. Ogasawar a M, Wada H, Pe ters H, Sa toh N. 1999. Develop­ Harkey MA, Klueg K, Sheppar d P , Raff RA. 1995 . Stru cture, mental expression of Pa:x l /9 genes in urochordate an d expression, and extracellular targeting of PM27, a skeletal hemichordate gills: insight into function an d evolution of protein associated specifically wit h growth of the sea urchin the pharyngeal epithe lium. Development 126:2539-2550. larval spicule. Dev Bioi 168:549- 566. Panopoulou G, Hennig S, Groth D, Krause A, Poustka AJ, Hyman LH. 1959. The enterocoelous coelomates-phylum Herwig R, Vingron M, Lehrach H. 2003. New evidence for Hemichordata. The invertebrates: smaller coelomate genome-wide duplications at the origin of vertebrates usi ng groups. New York : McGraw-Hill. p 72-207. an amphioxus gen e set and completed animal gen omes . Iku ta '1', Yoshida N, Satoh N, Saiga H. 2004 . Ciona intestinalis Genome Res 13:1056-1066. Hox gene cluster: its dispersed structure and residual Pardos F, Benito J . 1988. Blood vessels and related structures colinea r expression in development. Proc Natl Acad Sci 101: in the gill bars of Glossobalanus minutus (Enteropneusta). 15118-15123. Acta Zoologica 69:87-94. IUies MR, Pee ler MT, Dechtiaru k AM, Ettensohn CA. 2002 . Passam aneck YJ, Di Gregorio A. 2005 . Ciona intestinalis: Iden tificat ion and developme ntal expression of new biomi­ Chordate developmen t ma de sim ple. Dev Dyn 233:1-19. neralization protein s in the sea urchi n Strongylocentrotus Pen nachetti CA. 1984. Fu nctional morphology of the bran­ pu rpuratus. Dev Genes Evol 212:419-431. chial basket of Asc idia paratropa (Tunicata , Ascidiacea). Jeanmougin F, Thompson JD, Gouy M, Higgins DG, Gibson Zoomorphology 104:216-222. TJ. 1998. Multiple sequence alignment with Clustal X. Rahr H. 1982. Ultrastructure of gill bars of Branchiostoma Trends in Biochem Sci 23:403-405. lan ceolatum with special reference to gill skeleton and blood Kardong KY. 2006 . Vertebrates: comparative anatomy, func­ vesse ls (Cephalochordata). Zoomorphology 99:167-180. tion, evolu tion. New York: McGraw-Hill . p 47-83. Robso n P , Wright GM, Sitarz E, Maiti A, Rawat M, Youson Kato h-Fukui Y, Noce T, Ueda T, Fujiwara Y, Hashimoto N, JH, Keeley FW. 1993. Characterization of lamprin, an Higashinakagawa '1', Killian CE , Livingston BT , Wilt FH , unusual matrix protein from lamprey cartilage. Implications

J. Exp . Zoo!' (Mol. Dev. Evo !.) DOl 10.100 2/jez.b DEVELOPMENT AND EVOLUTION OF CHORDATE CARTILAGE 335

for evolution, structure, and assembly of elastin and other Gill R, Hamilton C, Hernandez J, Hines S, Hume J, Jackson fibrillar proteins. J Bioi Chern 268:1440-1447. L, Jolivet A, Kovar C, Lee S, Lewis L, Miner G, Morgan M, Rychel AL, Smith SE, Shirnamoto HT, Swalla BJ. 2006. Nazareth LV, Okwuonu G, Parker D, Pu L-L, Thorn R, Evolution and development of the chordates: collagen and Wright R. 2006. The genome of the sea urchin Strongylo­ pharyngeal cartilage. Mol Bioi Evol 23:541-549. centrotus purpuratus. Science 314:941-952. Schaeffer B. 1987. Deuterostome monophyly and phylogeny. Sekiya I, Tsuji K, Koopman P, Watanabe H, Yamada Y, Evol Bioi 21:179-235. Shinomiya K, Nifuji A, Noda M. 2000. SOX9 enhances Sea Urchin Genome Sequencing Consortium: Sodergren E, aggrecan gene promoter/enhancer activity and is up-regu­ Weinstock GM, Davidson EH, Cameron RA, Gibbs RA, lated by retinoic acid in a cartilage-derived cell line, TC6. J Angerer RC, Angerer LM, Arnone MI, Burgess DR, Burke Biol Chern 275:10738-10744. RD, Coffman JA, Dean M, Elphick MR, Ettensohn CA, Foltz Smith SE, Douglas R, da Silva KB, Swalla BJ. 2003. KR, Hamdoun A, Hynes RO, Klein WH, MarzluffW, McClay Morphological and molecular identification of Saccoglossus DR, Morris RL, Mushegian A, Rast JP, Smith LC, species (Hernichordata: Harrimaniidae) in the Pacific Thorndyke MC, Vacquier YD, Wessel GM, Wray G, Zhang Northwest. Can J Zool 81:133-141. L, Elsik CG, Ermolaeva 0, Hlavina W, Hofmann G, Kitts P, Suzuki T, Sakai D, Osumi N, Wada H, Wakamatsu Y. 2006. Landrum MJ, Mackey AJ, Maglott D, Panopoulou G, Sox genes regulate type 2 collagen expression in avian Poustka AJ, Pruitt K, Sapojnikov V, Song X, Souvorov A, neural crest cells. Dev Growth Differ 48:477-486. Solovyev V, Wei Z, Whittaker CA, Worley K, Durbin KJ, Swalla BJ. 2006. Building divergent body plans with similar Shen Y, Fedrigo 0, Garfield D, Haygood R, Primus A, Satija genetic pathways. Heredity 97:235-243. R, Severson T, Gonzalez-Garay ML, Jackson AR, Milosavl­ Swalla BJ. 2007. New insights into vertebrate origins. In: jevic A, Tong M, Killian CE, Livingston BT, Wilt FH, Adams Moody S, editor. Principles of developmental genetics. San N, Belle R, Carbonneau S, Cheung R, Cormier P, Cosson B, Diego: Elsevier Science/Academic Press. Croce J, Fernandez-Guerra A, Geneviere A-M, Goel M, Swofford DL. 1999. PAUP*: phylogenetic analysis using Kelkar H, Morales J, Mulner-Lorillon 0, Robertson AJ, parsimony (*and other methods). Version 4. Sunderland, Goldstone JV, Cole B, Epel D, Gold B, Hahn ME, Howard­ Mass: Sinaeur Associates. Ashby M, Scally M, Stegeman JJ, Allgood EL, Cool J, Wad a H, Okuyama M, Satoh N, Zhang S. 2006. Molecular Judkins KM, McCafferty SS, Musante AM, Obar RA, evolution of fibrillar collagen in chordates, with implications Rawson AP, Rossetti BJ, Gibbons IR, Hoffman MP, Leone for the evolution of vertebrate skeletons and chordate A, Istrail S, Materna SC, Samanta MP, Stole V, Tongprasit phylogeny. Evol Dev 8:370-377. W, Tu Q, Bergeron K-F, Brandhorst BP, Whittle J, Berney Wright GM, Keeley FW, Robson P. 2001. The unusual K, Bottjer DJ, Calestani C, Peterson K, Chow E, Yuan QA, cartilaginous tissues of jawless craniates, cephalochordates Elhaik E, Graur D, Reese JT, Bosdet I, Heesun S, Marra and invertebrates. Cell Tissue Res 304:165-174. MA, Schein J, Anderson MR, Brockton V, Buckley KM' Yan YL, Hatta K, Riggleman B, Postlethwait JH. 1995. Cohen AH, Fugmann SD, Hibino T, Loza-Coll M, Majeske Expression of a type II collagen gene in the zebrafish AJ, Messier C, Nair SV, Pancer Z, Terwilliger DP, Agca C, embryonic axis. Dev Dyn 203:363-376. Arboleda E, Chen N, Churcher AM, Hallbook F, Humphrey Yan YL, Miller CT, Nissen RM, Singer A, Liu D, Kirn A, GW, Idris MM, Kiyama T, Liang S, Mellott D, Mu X, Murray Draper B, Willoughby J, Morcos PA, Amsterdam A, Chung G, Olinski RP, Raible F, Rowe M, Taylor JS, Tessmar-Raible BC, Westerfield M, Hamer P, Hopkins N, Kimmel C, K, Wang D, Wilson KH, Yaguchi S, Gaasterland T, Galindo Postlethwait JH. 2002. A zebrafish sox9 gene required for BE, Gunaratne HJ, Juliano C, Kinukawa M, Moy CW, Neill cartilage morphogenesis. Development 129:5065-5079. AT, Nomura M, Raisch M, Reade A, Roux MM, Song JL, Su Zeng LY, Swalla BJ. 2005. Molecular phylogeny of the Y-H, Townley IK, Voronina E, Wong JL, Amore G, Branno protochordates: chordate evolution. Can J Zool-Revue M, Brown ER, Cavalieri V, Duboc V, Duloquin L, Flytzanis Canadienne De Zoologie 83:24-33. C, Gache C, Lapraz F, Lepage T, Locascio A, Martinez P, Zhang G, Miyamoto MM, Cohn MJ. 2006. Lamprey type II Matassi G, Matranga V, Range R, Rizzo F, Rottinger E, collagen and Sox9 reveal an ancient origin of the Beane W, Bradham C, Byrum C, Glenn T, Hussain S, vertebrate collagenous skeleton. Proc Natl Acad Sci 103: Manning FG, Miranda E, Thomason R, Walton K, Wikra­ 3180-3185. manayke A, Wu S-Y, Xu R, Brown CT, Chen L, Gray RF , Lee Zhao Q, Eberspaecher H, Lefebvre V, De Crombrugghe B. PY, Nam J, Oliveri P, Smith J, Muzny D, Bell S, Chacko J, 1997. Parallel expression of Sox9 and Col2al in cells Cree A, Curry S, Davis C, Dinh H, Dugan-Rocha S, Fowler J, undergoing chondrogenesis. Dev Dyn 209:377-386.

J. Exp. Zool. (Mol. Dev. Evol.) DOl lO.1002/jez.b ~ ----~===~. ------'