Amphioxus FGF signaling predicts the acquisition ofvertebrate morphological traitsStephanie Bertranda,1, Alain Camassesa,1,2, Ildiko Somorjaia,3, Mohamed R. Belgacema, Olivier Chabrolb,Marie-Line Escandea, Pierre Pontarottib, and Hector Escrivaa,4

aCentre National de la Recherche Scientifique, Unité Mixte de Recherche 7232, Université Pierre et Marie Curie Paris 06, Observatoire Océanologique, F-66651Banyuls-sur-Mer, France; and bCentre National de la Recherche Scientifique, Unité Mixte de Recherche 6632, Evolutionary Biology and Modelization,University of Marseilles,13331 Marseilles, France

Edited* by Walter J. Gehring, University of Basel, Basel, Switzerland, and approved April 19, 2011 (received for review September 25, 2010)

FGF signaling is one of the few cell–cell signaling pathways con-served among all metazoans. The diversity of FGF gene contentamong different phyla suggests that evolution of FGF signalingmay have participated in generating the current variety of animalforms. Vertebrates possess the greatest number of FGF genes,the functional evolution of which may have been implicated inthe acquisition of vertebrate-specific morphological traits. In thisstudy, we have investigated the roles of the FGF signal duringembryogenesis of the cephalochordate amphioxus, the best proxyfor the chordate ancestor. We first isolate the full FGF gene com-plement and determine the evolutionary relationships betweenamphioxus and vertebrate FGFs via phylogenetic and synteny con-servation analysis. Using pharmacological treatments, we inhibitthe FGF signaling pathway in amphioxus embryos in different timewindows. Our results show that the requirement for FGF signalingduring gastrulation is a conserved character among chordates,whereas this signal is not necessary for neural induction in amphi-oxus, in contrast to what is known in vertebrates. We also showthat FGF signal, acting through the MAPK pathway, is necessaryfor the formation of the most anterior somites in amphioxus,whereas more posterior somite formation is not FGF-dependent.This result leads us to propose that modification of the FGF signalfunction in the anterior paraxial mesoderm in an amphioxus-likevertebrate ancestor might have contributed to the loss of segmen-tation in the preotic paraxial mesoderm of the vertebrate head.

evo-devo | somitogenesis | gene family

Only a few cell–cell signaling molecules are known to playa major role during metazoan embryonic development.

Among them, the FGFs were discovered in the mid-1970s in ver-tebrates. FGFs are small proteins, generally secreted, character-ized by a conserved functional domain and acting through bindingto their transmembrane receptors [FGF receptors (FGFRs)],causing them to homodimerize. This leads to intracellular auto-phosphorylation and further activation of cytoplasmic signalingcascades, such as the MAPK and the PI3K pathways (1).Sequencing of several complete metazoan genomes has shown

that FGF and FGFR genes were already present in the commonancestor of diploblastic and bilaterian animals (2). In vertebrates,at least 22 genes coding for FGFs and 4 coding for their FGFRsare known. In contrast, only 3 genes coding for FGFs and 2 codingfor FGFRs have been described in Drosophila (2). In the uro-chordate Ciona intestinalis, a member of the sister group of ver-tebrates, only 6 FGF genes are present, for which some orthologyrelationships with vertebrate FGFs are still not resolved (3). Thislarge number of FGFs and FGFRs in vertebrates raises thequestion of their implication in the evolution of vertebrate-specificmorphological characteristics.In vertebrates, FGF signaling is involved in different devel-

opmental processes, among which are neural and mesoderminduction in the early embryo, and somitogenesis and limb budformation at later stages (4). To investigate the evolution ofFGF signaling pathway function during embryogenesis at theinvertebrate chordate-to-vertebrate transition, we undertooka study of the embryonic role of this pathway in the cepha-

lochordate amphioxus (Branchiostoma lanceolatum). Amphioxusis an amenable model system to address this question because ofits basal position in the chordate tree. Moreover, its genomic,morphological, and developmental characteristics are probablyhighly similar to those of the chordate ancestor (5).We describe here the content of the amphioxus FGF genes,

as well as their evolutionary relationships with vertebrate FGFs,using phylogeny and deep synteny conservation analyses. De-tailed study of the expression patterns of all amphioxus FGFsindicates that FGF-dependent developmental processes in am-phioxus are as varied as those of vertebrates. Moreover, ourfunctional analysis shows that (i) during early development, FGFsignal in amphioxus controls cellular movements through aMAPK-independent pathway, similar to some vertebrates; (ii)in contrast to what is described in vertebrates, neural and me-soderm induction are not FGF-dependent in amphioxus; and (iii)a MAPK-dependent FGF signal is necessary during gastrulationfor the formation of the rostral but not the posterior somites,where it is instead required for their normal maturation. Overall,our work shows that the role of the FGF signaling pathwayduring neural and mesoderm induction is a specific vertebrateacquisition. In addition, changes in the requirement for the FGFsignal during somitogenesis, particularly at the rostral level, couldexplain the appearance of the “new head” in vertebrates (6).

ResultsAmphioxus FGF Gene Content. BLAST searches on the Bran-chiostoma floridae genome for FGF and FGFR genes resultedin the identification of eight putative genes containing an FGFdomain and one FGFR coding gene. The corresponding full-length cDNAs were cloned in B. lanceolatum. All the predictedFGF proteins contain a highly conserved FGF domain, and fivecontain a putative signal peptide in their N-terminal region(Table S1). The unique FGFR gene was previously described asthe ortholog of the four vertebrate FGFRs (7). We used phylo-genetic reconstructions based on the alignment of the FGFdomains to assign the orthology relationships of the amphioxus

Author contributions: S.B., A.C., I.S., P.P., and H.E. designed research; S.B., A.C., I.S., M.R.B.,O.C., M.-L.E., and H.E. performed research; S.B., I.S., P.P., and H.E. analyzed data; and S.B.,I.S., and H.E. wrote the paper.

The authors declare no conflict of interest.

*This Direct Submission article had a prearranged editor.

Data deposition: The sequences reported in this paper have been deposited in the Gen-Bank database (accession nos. EU606032.1, EU606035.1, EU606036.1, EU606033.1,EU606034.1, EU606038.1, HM854710, EU606037.1, HM854709, HM359129, HM359124,HM359125, and HM359126).1S.B. and A.C. contributed equally to this work.2Present address: Institut de Génétique Moléculaire de Montpellier, Centre National de laRecherche Scientifique, Unité Mixte de Recherche 5535, 34293 Montpellier Cedex 5,France.

3Present address: Departament de Genètica, Facultat de Biologia, Universitat de Barce-lona, E08028 Barcelona, Spain.

4To whom correspondence should be addressed. E-mail: hescriva@obs-banyuls.fr.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1014235108/-/DCSupplemental.

9160–9165 | PNAS | May 31, 2011 | vol. 108 | no. 22 www.pnas.org/cgi/doi/10.1073/pnas.1014235108

FGF genes. No matter the method used (maximum likelihood orBayesian inference), only three FGF genes have a well-sup-ported position in the phylogenetic tree (Fig. S1), namely, FGF1/2, FGF8/17/18, and FGF9/16/20, which are placed at the base ofthe vertebrate FGF1/2, FGF8/17/18, and FGF9/16/20 paralogygroups, respectively. The other five genes were named FGFA,FGFB, FGFC, FGFD, and FGFE. To go further, we looked intosynteny conservation of the FGF and FGFR gene loci betweenamphioxus and vertebrates (Fig. 1). For FGF1/2 (Fig. 1A),FGF8/17/18 (Fig. 1D), and FGFR (Fig. 1F), synteny conserva-tion with vertebrates reinforces their orthology relationships. Wealso found amphioxus/vertebrate synteny conservation forFGFA, FGFB, and FGFC genes (Fig. 1 B, C, and E), indicatingthat they are orthologs of the FGF3/7/10/22, FGF4/5/6, andFGF19/21/23 vertebrate paralogy groups, respectively.

FGFs Show a Dynamic Expression Pattern During Embryogenesis. Weanalyzed the expression pattern of FGFR and the eight FGFgenes from the eight-cell stage to the larval stage. The FGFRgene is ubiquitously expressed at all the studied stages, except inthe epidermis, with a higher expression level in the mesoderm(Fig. S2). Expression of two of the FGF genes (FGFB andFGFD) could not be detected by whole-mount in situ hybrid-ization, and FGF1/2 shows ubiquitous expression until the larvalstage, except in the epidermis. The remaining five FGFs showa restricted expression pattern (Fig. 2, more detailed expressionpatterns are shown in Figs. S3–S7) (8, 9). Interestingly, onlyFGF8/17/18, FGFA, and FGFE show transient expression in thepresumptive mesoderm. Indeed, FGF8/17/18 is expressed in theposterior dorsal mesendoderm in the gastrula stage embryo (Fig.2 A and B), but this expression fades rapidly and is no longer

visible in the mesoderm at the midneurula stage (Fig. 2D).FGFA expression in the anterior dorsal mesendoderm is onlyobserved at the early midneurula stage (Fig. 2M). FGFE is alsoexpressed in a very restricted mesodermal domain, the first leftsomite, and only very transiently in midlate neurula embryos(Fig. 2Q). Four amphioxus FGFs, including FGF8/17/18, FGF9/16/20, FGFA, and FGFE, are expressed in the central nervoussystem, although the expression is very dynamic. The first threeare detected in the anterior neural plate and cerebral vesicle butat different developmental stages (Fig. 2V). FGF9/16/20 is theearliest neurally expressed gene, with restricted expression in thepresumptive neural plate of the gastrulating embryo (Fig. 2 Gand H). It is thereafter expressed segmentally in the neural tubeuntil the late neurula stage (Fig. 2 J and K); at that time, FGFEexpression starts to be observed in some neurons up until thelarval stage (Fig. 2 R and S). With respect to the endoderm,specific expression in the gut is first observed for FGFC in thelate neurula stage embryo (Fig. 2T). Subsequently, in the larva,FGF9/16/20, FGFA, and FGFE are expressed in the midgutregion, with the first two also being expressed in the anus (Fig. 2 Land P). The pharyngeal endoderm is the embryonic region show-ing the most complex FGF expression combination (Fig. 2V). In-deed, in the larva, all the FGF genes having a restricted expressionpattern are expressed in the pharynx. FGFA and FGF8/17/18 areexpressed around the mouth; FGF8/17/18, FGF9/16/20, andFGFA in the first gill slit; FGF8/17/18 and FGFC in the preoralpit; FGF9/16/20, FGFA, FGFE, and FGFC in the club-shapedgland; and FGFC in the most anterior part of the pharyngealendoderm (Fig. 2 F, L, P, S, and U). The expression in the larvaseems to reflect the expression in different regions of the pha-ryngeal endoderm at the late neurula stage (Fig. 2V).

Fig. 1. Synteny conservation among vertebrate and amphioxus FGF chromosomal regions. For clarity, only synteny conservation with human FGF1/2 (A),FGF3/7/10/22 (B), FGF4/5/6 (C), FGF8/17/18 (D), FGF19/21/23 (E), and FGFR1/2/3/4 (F) paralogy groups is represented. Names of the human genes are accordingto the Ensembl database, and amphioxus names are according to our phylogenetic analyses. Genes for which synteny conservation was found using theCASSIOPE system are marked by a black star. A white star marks amphioxus genes for which phylogenetic signal is insufficient; in this case, the name isaccording to the best reciprocal BLAST hit result.

Bertrand et al. PNAS | May 31, 2011 | vol. 108 | no. 22 | 9161

EVOLU

TION

Early FGF Signaling Controls Gastrulation via a MAPK-IndependentPathway. To understand the role of FGF signaling during gas-trulation, we treated embryos with SU5402, an FGFR inhibitor,and U0126, a MAPK pathway inhibitor. Treatments were per-formed at two different developmental stages, the four- to eight-cell stage and the blastula stage (treatments 1 and 2 in Fig. 3A).When FGF signaling is inhibited with SU5402 at the earlier stage,the embryos are not arrested in their development (Fig. S8) butfail to gastrulate and form a sphere of cells with two distinctregions (Fig. 3C). In these embryos, Otx and Nodal are expressedin one-half of the sphere, possibly corresponding to the pre-sumptive mesendoderm cells that failed to invaginate (Fig. 3 F, F′and G, G′). Brachyury2 is expressed as a ring that suggests thepresumptive blastoporal region (Fig. 3 H, H′), and Neurogenin,a proneural marker, is expressed in patches in one-half of theembryo (Fig. 3 I, I′). Altogether, these results show that meso-dermal and neural induction is not abolished. On the other hand,treatment with U0126, even at toxic concentrations (50 μM),leads to milder developmental defects without abolishing gas-trulation, suggesting that FGF signaling is not acting through theMAPK pathway to control these cell movements (Fig. 3 B, D, andE). Interestingly, when embryos are treated with either SU5402 orU0126 at the blastula stage (treatment 2 in Fig. 3A), corre-sponding to the maternal-zygotic transition, gastrulation occursnormally. These results suggest that gastrulation movements arecontrolled by a maternally inherited FGF signal acting before theblastula stage.

FGF Signaling Controls Anterior Somitogenesis but Not PosteriorElongation of the Embryo. Several treatments were performedusing both SU5402 and U0126, as shown in Fig. 3A. First, weinvestigated the expression pattern of several marker genes inlate gastrula embryos after treatment 2. With either treatment,the proneural gene Neurogenin shows normal expression re-stricted to the neural plate (Fig. 4 A–A′′). Chordin, an axialdorsal mesendoderm marker (presumptive notochord domain),also shows normal expression, although the expression domain isenlarged in treated embryos (Fig. 4 B–B′′). In contrast, the ex-pression of paraxial dorsal mesendoderm marker genes is highlysensitive to both treatments. Indeed, the expression of Bra-chyury2, Delta, Snail, and MRF1 is completely abolished in themost anterior presumptive somitic region after FGF or MAPKpathway inhibition, whereas their expression in other regions of

the embryo is not altered (Fig. 4 C–C′′, D–D′′, E–E′′, and F–F′′).Interestingly, Nodal expression is not abolished in the pre-sumptive anterior somites, suggesting that it is not under thecontrol of the FGF signaling pathway (Fig. 4 G–G′′). We theninvestigated the expression of MRF1 as well as that of MLC(Myosin Light Chain) at later stages. In midlate neurula SU5402-treated embryos, MRF1 starts to be expressed in the paraxialmesoderm but only in the posterior half of the embryo, whereasthe anterior half shows no expression (Fig. 4 H and I). No la-beling is observed for MLC in midlate neurula SU5402-treatedembryos (Fig. 4 H′ and I′). The same result is obtained forMRF1just before the mouth opens; there is an anterior region of theparaxial mesoderm free of MRF1 expression (Fig. 4 J and K).Moreover, although expression of MLC in the notochord ofSU5402-treated embryos persists all along the body axis, its ex-pression in the paraxial mesoderm is only detected more poste-riorly than in control embryos (Fig. 4 J′ and K′). These resultsindicate a total absence of anterior somite formation as con-firmed by embryo sections (Fig. S9). Similar results are observedin U0126-treated embryos (compare Fig. 4 K, K′ with L, L′). Toensure that the difference in sensitivity to FGF signaling in-hibition between the anterior and posterior somites was not at-tributable to degradation of SU5402, we treated the embryos atthe late gastrula stage, before the formation of the first somites(treatment 3 in Fig. 3A). In this case, as shown by the expressionof MRF1 and MLC, both anterior and posterior somites form,although somite and notochord morphology are abnormal (com-pare Fig. 4 J, J′ with M, M′). These results indicate that FGFsignaling, through activation of the MAPK pathway, is specifi-cally required between the blastula and the late gastrula stagesfor the formation of the most anterior somites.To deepen our understanding of the implications of FGF

signaling during somitogenesis, we treated embryos with SU5402at a later stage, just before the somites start to form directly fromthe tailbud by schizocoely and not by enterocoely from thepresomitic mesoderm (treatment 4 in Fig. 3A). As shown by theexpression of MRF1 just before the mouth opens, posteriorsomites still form following treatment (Fig. 5 A, A′). Later on,at the larval stage, somite budding remains normal. However,the larvae show a strong posterior phenotype. Morphology andmaintenance of Brachyury2 expression indicate that maturationof the posterior notochord is altered (Fig. 5 B, B′). Moreover,although the somites formed from the tailbud during treatment

Fig. 2. Amphioxus FGF genes show very dynamic embryonicexpression patterns. Expression of FGF8/17/18 (A–F), FGF9/16/20 (G–L), FGFA (M–P), FGFE (Q–S), and FGFC (T and U) fromgastrula to larva. (V) Schematic representation of the ex-pression of FGF8/17/18, FGF9/16/20, FGFA, FGFE, and FGFCfrom the late gastrula stage to the late neurula stage. Littleoverlap of expression is observed in the pharyngeal region.Only stages for which a restricted expression pattern wasobserved are presented. Anterior is to the left and dorsal isto the top. For detailed descriptions of the expression pat-terns see Figs. S3–S7.

9162 | www.pnas.org/cgi/doi/10.1073/pnas.1014235108 Bertrand et al.

develop muscle fibers, these are not correctly aligned (Fig. 5 C, C′).Our transmission electron microscopy (TEM) analysis shows thatthere are no major differences between control and treatedembryos, except maybe some intercellular spaces that appearbetween somitic cells and the epidermis (Fig. 5 D, D′).

DiscussionEvolution of the FGF Gene Family in Chordates. An explosion of theFGF family diversity occurred at the base of the vertebrate lin-eage. Indeed, at least 22 FGF genes are known in vertebrates,whereas only 6 genes were found in C. intestinalis (3). Phyloge-netic studies have thus far failed to decipher the orthology rela-tionships between all urochordate and vertebrate FGFs, such thatthe evolutionary history of the FGF family in chordates remainsunresolved. In this study, using both phylogenetic reconstructionsand precise synteny conservation analysis, we show that all ver-tebrate FGF paralogy groups have an amphioxus pro-ortholog,with the exception of FGF homologous factors (FHF) (FGF11/12/13/14). Interestingly, FGFE, one of the two amphioxus FGFs forwhich we find no clear orthology, lacks a putative signal peptide,as do all the vertebrate FHFs (Table S1). Moreover, FGFEgene expression at the late neurula stage, in a subpopulation ofneurons in the neural tube (Fig. 2R), resembles that of zebrafishand Xenopus FGF13 (10, 11). These arguments suggest thatFGFE may be the amphioxus ortholog of the FHF group. Alto-gether, these data indicate that the chordate ancestor alreadypossessed one FGF ortholog of each vertebrate paralogy group.This implies that the high number of FGF genes in vertebrates isthe result of high gene retention after the two rounds of whole-genome duplication that occurred at the base of this lineage(12). The origin of the amphioxus FGFD is still unknown, andwe currently have no clue as to whether it appeared specificallyin the cephalochordate lineage or if an ortholog was present inthe chordate ancestor and was then lost before the urochordate/vertebrate divergence.Our study shows that amphioxus FGF genes have very dy-

namic and diverse expression patterns, even though some over-lap is observed (Fig. 2V). These results clearly suggest that, as invertebrates, FGF signaling is implicated in many developmentalprocesses in cephalochordates (4). It is nevertheless difficult to

compare the detailed expression of amphioxus FGFs with theirvertebrate orthologs. On the one hand, many data are availableshowing that FGFs sometimes have different expression patternsin different vertebrate species; on the other hand, we often onlyhave expression data for a single vertebrate species. Neverthe-less, some useful comparisons can be made. Indeed, FGF8/17/18is the earliest gene showing a restricted expression pattern, whichwas first observed in the presumptive mesoderm. In zebrafish as inXenopus, FGF8 is one of the FGFs showing the earliest restrictedexpression, which is also observed in the future mesoderm (10,13), suggesting a conserved role of at least this paralogue betweenamphioxus and vertebrates. Another noticeable conserved ex-pression domain is the one described above for FGFE in somespecific neurons (10, 11). In contrast, FGF9/16/20 is characterizedby early expression, first restricted to the whole neural plate inamphioxus, whereas no comparable expression has been describedfor any of the vertebrate FGFs belonging to the FGF9/16/20paralogy group (10), suggesting nonconserved functions betweenamphioxus and vertebrates. If we compare our data at the level ofthe whole FGF family with what is known in vertebrates, itappears that expression of FGF ligands in the central nervoussystem and the pharyngeal region is a conserved feature amongchordates. However, vertebrate FGFs are also expressed in manystructures that are absent in amphioxus, such as neural crest cellderivatives, placodes, and brain-specific regions, suggesting func-tional acquisitions during evolution. The most striking differencebetween amphioxus and vertebrates is the absence of specific FGFexpression in amphioxus somites and notochord, except the earlyFGF8/17/18 expression in the presumptive mesoderm at the gas-trula stage. This might be associated with the differences in therequirement for the FGF signal during somitogenesis that wefunctionally demonstrate in this study.

FGF Function During Chordate Gastrulation. In vertebrates, FGFs areknown to be critical for normal gastrulation movements (14–16).Here, we show that inhibition of the FGF signaling pathway duringearly development in amphioxus leads to gastrulation failure withabsence of blastoderm invagination, as previously suggested by Yuand Zhang (17). Although cell movements during amphioxusgastrulation differ considerably from those seen in vertebrates, therole of the FGF pathway in shaping the gastrula during these early

Fig. 3. Inhibition of FGF and MAPK signaling pathways after fertilizationproduces different phenotypes. (A) Scheme of the pharmacological treatmentsperformed in this study with SU5402 and U0126. Black arrows represent thetimewindow for each continuous treatment, and red arrows represent the timepoints at which embryos where fixed. (B–E) Side views of embryos fixed at theearly midneurula stage after treatment 1. Expression of Otx (F and F′), Nodal(G and G′), Brachyury2 (H and H′), and Neurogenin (I and I′) in early mid-neurula embryos after treatment 1. In the side views, anterior is to the leftand dorsal is to the top.

Fig. 4. FGF and MAPK signaling inhibition induces loss of the most anteriorsomites. Expression patterns by whole-mount in situ hybridization of Neu-rogenin (A–A′′), Chordin (B–B′′), Brachyury2 (C–C′′), Delta (D–D′′), Snail (E–E′′),MRF1 (F–F′′ and H–M), Nodal (G–G′′), and MLC (H′–M′) after treatments withSU5402 (50 μM) or with U0126 (25 μM). Embryos after treatment 2 werefixed at the late gastrula stage (A–G′′), at the midneurula stage (H, H′ and I, I′),or at the premouth stage (J–L and J′–L′). Embryos after treatment 3 werefixed at the premouth stage (M, M′). A–A′′, E–E′′, and G–G′′ are blastoporeviews. B–B′′; C–C′′; D–D′′; F–F′′; H, H′; I, I′; and J′–M′ are dorsal views. J–M arelateral views. The most anterior limit of MRF1 and MLC is marked by a blackarrow. Anterior is to the left in dorsal and lateral views, and dorsal is to thetop in side and blastopore views.

Bertrand et al. PNAS | May 31, 2011 | vol. 108 | no. 22 | 9163

EVOLU

TION

events is an ancestral characteristic of chordates. In sea urchin,FGF signaling is also important for normal invagination duringgastrulation and for migration of the primary mesenchyme cells.This implies that FGF may be implicated in early embryonic cel-lular movements in all deuterosomes (18).Interestingly, SU5402 and U0126 treatments at the four to

eight-cell stage in amphioxus do not produce the same pheno-type. In effect, U0126-treated embryos do gastrulate. Hence, theintracellular signal activated by FGFs during early developmentalstages in amphioxus, and which is necessary for normal cellmovements, is not the MAPK pathway. It is interesting to notethat the same conclusions were drawn in Xenopus even thoughthe time at which this MAPK-independent FGF signal controlsmorphogenesis is different (19, 20).In vertebrates, FGF signaling during early development is not

only important for normal gastrulation movements but for twoessential processes, mesoderm and neural induction. On FGFsignal inhibition, vertebrate embryos fail to form mesoderm asdemonstrated by the absence of expression of early mesodermalmarker genes like Brachyury (4). The “default model” for neuralinduction, proposed more than 10 y ago, postulated that the ec-toderm will acquire a neural fate in the absence of any signal andthat bone morphogenic protein (BMP) acts as the master signalfor epidermal fate induction in vertebrates (21). However, it hasbeen shown that this model is incomplete, because FGF is alsorequired for neural induction (21). In this study, we have shownthat early inhibition of FGF signaling in amphioxus abrogatesnormal gastrulation but that the spherical embryo still containscells with presumptive neural, mesodermal, and endodermalfates, as shown by marker gene expression. This suggests that the

role of the FGF signal in mesoderm and neural induction wasacquired in the chordate lineage after cephalochordate di-vergence. These results lead us to propose that the “defaultmodel” could be applied to amphioxus, supported by the fact thatactivation of the BMP signaling pathway in amphioxus is suffi-cient to abolish neural induction (22, 23).

FGF Control of Somitogenesis. Amphioxus somites form in an an-terior-to-posterior sequence, with somitogenesis classically di-vided into two phases. The most anterior early-arising somitesform as bilateral pairs by means of enterocoelic evagination ofthe wall of the archenteron. Later on, the remaining posteriorsomites form from the tailbud by schizocoely, alternately on theleft and right sides of the embryo (Fig. 5E). Beyond this onto-logical division, expression of several genes clearly functionallydivides anterior from posterior somites. Engrailed, OligA, andPbx (24) are specific gene markers of the anterior enterocoelicsomites, and Axin, Lcx, and Paraxis specifically label the tailbudduring posterior somitogenesis (24). However, several datasuggest an additional division of the amphioxus somites, partic-ularly within the anterior enterocoelic ones. Ontogenetically, themost anterior of these enterocoelic somites form simultaneously,whereas the most posterior form sequentially. Moreover, geneexpression also differentiates these two somitic regions. Indeed,Mox is never expressed in the most anterior simultaneouslyformed somites, suggesting a functional difference from those inthe posterior (25). Our data show that inhibition of FGF orMAPK pathways at the blastula stage induces a complete loss ofthe most anterior enterocoelic somites, whereas formation of allthe most posterior somites is independent of these two pathways.Therefore, we clearly establish the presence of three differentsomitic populations in amphioxus: (i) the most anterior entero-coelic, FGF-sensitive,Mox-negative, andEngrailed-positive somites;(ii) the posterior enterocoelic, FGF-insensitive,Mox- andEngrailed-positive somites; and (iii) the posterior schizocoelic, FGF-insensitive,Engrailed-negative, and Mox-, Axin-, Lcx-, and Paraxis-positivesomites (Fig. 5E). This implies that regulation of the formationof tailbud-derived somites by a posterior FGF signal, controllingthe position of the “wavefront” as proposed in the “clock andwavefront” model (26), would have been acquired specifically invertebrates, although we cannot prove that it is not a secondaryloss in the cephalochordate lineage.

FGF and the Vertebrate Head. Since it was proposed by Goethe200 y ago, the question of whether the vertebrate head is seg-mented or not has been debated (27). Some anatomical studiessupported the idea that vertebrates were primitively segmentedfrom one end to the other, with anterior paraxial mesodermclaimed to form somite-like units or “somitomeres” in the head(28). A segmented vertebrate head upheld the argument forevolution by simple elaboration of the anterior part of an am-phioxus-like ancestral animal. However, the bulk of morpho-logical comparative data among different vertebrate species (29),as well as the absence of any periodical pseudosegmental struc-tures in lampreys (30) and of comparable expression patterns ofgenes known to function in somitogenesis, present major obsta-cles to this hypothesis and instead support the absence of ce-phalic somitomeres in vertebrates.Accepting that the vertebrate head is unsegmented, two major

hypotheses may explain how the vertebrate body plan was ach-ieved from a completely segmented hypothetical amphioxus-likeancestor. First, as Gans and Northcutt (6) proposed, a “newhead” appeared in vertebrates, mostly derived from the neuralcrest and epidermal placodes. This “new head” was added to therostral part of the body. The second hypothesis is that the headwould have been acquired by losing the epithelial segmentationof the most anterior paraxial mesoderm. This implies that theanterior end of amphioxus would be homologous to the anteriorend of the vertebrate head.Of these two hypotheses, the second is the more plausible

because different expression patterns and anatomical studies

Fig. 5. Posterior somite budding is not dependent on FGF signaling. Expres-sion of MRF1 at the late neurula stage in control embryos (A) and SU5402-treated embryos (treatment 4) (A′). (Scale bars = 50 μm.) Expression of Bra-chyury2 in control embryos (B) and in embryos fixed at the larva stage aftertreatment 4 with SU5402 (B′). In the control embryos, Brachyury2 is expressed inthe most anterior notochord and in the tailbud, whereas in the treated em-bryos, Brachyury2 is still expressed in the posterior notochord anterior to thetailbud. (Scale bars = 50 μm.) (C and C′) F-Actin staining of larval posteriorstriatedmuscle fibers in control embryos and in embryos after treatment 4 withSU5402. (Scale bars = 50 μm.) TEM pictures of transverse sections at the larvastage of the posterior region of control embryos (D) and of embryos aftertreatment 4 with SU5402 (D′). Arrows indicate the space between the somiticcells and the epidermis.(Scale bars = 10 μm.) (E) Schematic representation of thedifferent somitic regions in amphioxus. Dorsal view; anterior is to the left.

9164 | www.pnas.org/cgi/doi/10.1073/pnas.1014235108 Bertrand et al.

support the homology between the rostral parts of amphioxusand vertebrate embryos (31). In fact, the evolution of the ver-tebrate head was probably possible not only through the ap-pearance of the neural crest and placodes but also via loss ofmesoderm segmentation in the anterior region, thereby relaxingthe developmental constraints imposed by the somites. Ourresults provide a mechanistic explanation as to how this evolu-tion might have occurred; namely, through loss of FGF functionin the development of the paraxial mesoderm during early gas-trulation in a hypothetical vertebrate ancestor. Indeed, simpleinhibition of the FGF signal in amphioxus specifically inducesloss of the most anterior somites, creating an anterior region withunsegmented paraxial mesoderm, as in the vertebrate head. Insupport of this, data obtained in Xenopus suggest that inhibitionof the FGF signal in the anterior region of vertebrates is nec-essary for head formation. Indeed, Shisa, which antagonizes FGFfunction by inducing FGFR retention in the endoplasmic re-ticulum, is expressed in the anterior part of the vertebrate gas-trula, and its knock-down in Xenopus leads to an anteriortruncation of the head (32). It is tempting to propose that FGFsignal inhibition in the anterior region of vertebrates could havebeen partly achieved by the recruitment of Shisa.

ConclusionsIn this work, we present data suggesting that both loss and gainof FGF function have been instrumental in the evolution of thevertebrate body plan. Modification of the FGF requirement forsomitogenesis during gastrulation could have been a major eventin the appearance of the vertebrate head. Likewise, gain of FGFfunction during posterior somitogenesis probably contributed to

the plasticity evident in the posterior elongation process of ver-tebrates.

Materials and MethodsSynteny Conservation Analysis. Artificial contigs for each FGF human paralogygroup and for the FGFR group were constructed by assembling the chro-mosomic regions containing the human FGF or FGFR genes and 50 genesupstream and downstream. Conserved regions in vertebrates and amphioxuswere found using CASSIOPE (33) (SI Materials and Methods).

Embryo Methods. Gametes were obtained as previously described (34), andembryos were fixed and processed for whole-mount in situ hybridization asdescribed (35). The accession numbers of the sequences used for probesynthesis are provided in SI Materials and Methods.

SU5402 and U0126 (Calbiochem), dissolved in DMSO at 10−2 M, wereadded to cultures of embryos (SI Materials and Methods and Fig. S10). Themaximum final DMSO concentration was 0.5%. Control embryos were raisedwith 0.5% DMSO in sea water.

For F-Actin staining, embryos were washed in PBS + 0.1% Tween 20 afterfixation and then in PBS + 0.1% Triton X100. They were subsequently in-cubatedwith Texas RedX-Phalloidin (Invitrogen) dissolved at 5 U/mL in PBS for1 h. After washes in PBS, embryos were transferred to Mowiol (Calbiochem)for confocal imaging.

TEM was performed according to Lacalli and West (36) with some minormodifications: Postfixation was performed with 1% OsO4, and aqueousuranyl acetate was avoided before embedding in Epon resin.

ACKNOWLEDGMENTS. We thank Laurent Kodjabachian for providing theSU5402 and Laeticia Bariat for technical assistance. The laboratory of H.E. issupported by the Agence Nationale de la Recherche BLAN 1716 01. S.B. holdsa fellowship from the Association pour la Recherche sur le Cancer.

1. Eswarakumar VP, Lax I, Schlessinger J (2005) Cellular signaling by fibroblast growth

factor receptors. Cytokine Growth Factor Rev 16:139–149.2. Itoh N, Ornitz DM (2004) Evolution of the Fgf and Fgfr gene families. Trends Genet 20:

563–569.3. Satou Y, Imai KS, Satoh N (2002) Fgf genes in the basal chordate Ciona intestinalis.

Dev Genes Evol 212:432–438.4. Thisse B, Thisse C (2005) Functions and regulations of fibroblast growth factor

signaling during embryonic development. Dev Biol 287:390–402.5. Schubert M, Escriva H, Xavier-Neto J, Laudet V (2006) Amphioxus and tunicates as

evolutionary model systems. Trends Ecol Evol 21:269–277.6. Gans C, Northcutt RG (1983) Neural crest and the origin of vertebrates: A new head.

Science 220:268–273.7. D’Aniello S, et al. (2008) Gene expansion and retention leads to a diverse tyrosine

kinase superfamily in amphioxus. Mol Biol Evol 25:1841–1854.8. Meulemans D, Bronner-Fraser M (2007) Insights from amphioxus into the evolution of

vertebrate cartilage. PLoS ONE 2:e787.9. Yu JK, Meulemans D, McKeown SJ, Bronner-Fraser M (2008) Insights from the

amphioxus genome on the origin of vertebrate neural crest. Genome Res 18:

1127–1132.10. Lea R, Papalopulu N, Amaya E, Dorey K (2009) Temporal and spatial expression of FGF

ligands and receptors during Xenopus development. Dev Dyn 238:1467–1479.11. Thisse B, Thisse C (2004) Fast release clones: A high throughput expression analysis.

ZFIN Direct Data Submission, Available at http://zfin.org.12. Dehal P, Boore JL (2005) Two rounds of whole genome duplication in the ancestral

vertebrate. PLoS Biol 3:e314.13. Thisse B, et al. (2001) Expression of the zebrafish genome during embryogenesis. ZFIN

Direct Data Submission, Available at http://zfin.org.14. Amaya E, Musci TJ, Kirschner MW (1991) Expression of a dominant negative mutant

of the FGF receptor disrupts mesoderm formation in Xenopus embryos. Cell 66:

257–270.15. Ciruna BG, Schwartz L, Harpal K, Yamaguchi TP, Rossant J (1997) Chimeric analysis

of fibroblast growth factor receptor-1 (Fgfr1) function: A role for FGFR1 in

morphogenetic movement through the primitive streak. Development 124:2829–

2841.16. Sun X, Meyers EN, Lewandoski M, Martin GR (1999) Targeted disruption of Fgf8

causes failure of cell migration in the gastrulating mouse embryo. Genes Dev 13:

1834–1846.17. Yu D, Zhang PJ (2006) Mechanisms of amphioxus gastrulating cell movements: Forces

that drive the most primitive form of chordate gastrulation. Orig Life Evol Biosph 36:

337.18. Röttinger E, et al. (2008) FGF signals guide migration of mesenchymal cells, control

skeletal morphogenesis [corrected] and regulate gastrulation during sea urchin

development. Development 135:353–365.

19. Nutt SL, Dingwell KS, Holt CE, Amaya E (2001) Xenopus Sprouty2 inhibits FGF-mediated gastrulation movements but does not affect mesoderm induction andpatterning. Genes Dev 15:1152–1166.

20. Sivak JM, Petersen LF, Amaya E (2005) FGF signal interpretation is directed by Sproutyand Spred proteins during mesoderm formation. Dev Cell 8:689–701.

21. Stern CD (2006) Neural induction: 10 years on since the ‘default model’. Curr Opin CellBiol 18:692–697.

22. Yu JK, et al. (2007) Axial patterning in cephalochordates and the evolution of theorganizer. Nature 445:613–617.

23. Onai T, Yu JK, Blitz IL, Cho KW, Holland LZ (2010) Opposing Nodal/Vg1 and BMPsignals mediate axial patterning in embryos of the basal chordate amphioxus. DevBiol 344:377–389.

24. Beaster-Jones L, et al. (2008) Expression of somite segmentation genes in amphioxus:A clock without a wavefront? Dev Genes Evol 218:599–611.

25. Minguillón C, Garcia-Fernàndez J (2002) The single amphioxus Mox gene: Insights intothe functional evolution of Mox genes, somites, and the asymmetry of amphioxussomitogenesis. Dev Biol 246:455–465.

26. Pourquié O (2003) Vertebrate somitogenesis: A novel paradigm for animalsegmentation? Int J Dev Biol 47:597–603.

27. de Beer GR (1937) The Development of the Vertebrate Skull (Oxford Univ Press,London).

28. Meier S (1981) Development of the chick embryo mesoblast: Morphogenesis of theprechordal plate and cranial segments. Dev Biol 83:49–61.

29. Freund R, Dörfler D, Popp W, Wachtler F (1996) The metameric pattern of the headmesoderm—Does it exist? Anat Embryol (Berl) 193:73–80.

30. Kuratani S, Horigome N, Hirano S (1999) Developmental morphology of the headmesoderm and reevaluation of segmental theories of the vertebrate head: Evi-dence from embryos of an agnathan vertebrate, Lampetra japonica.. Dev Biol 210:381–400.

31. Holland LZ (2009) Chordate roots of the vertebrate nervous system: Expanding themolecular toolkit. Nat Rev Neurosci 10:736–746.

32. Yamamoto A, Nagano T, Takehara S, Hibi M, Aizawa S (2005) Shisa promotes headformation through the inhibition of receptor protein maturation for the caudalizingfactors, Wnt and FGF. Cell 120:223–235.

33. Rascol VL, et al. (2009) CASSIOPE: An expert system for conserved regions searches.BMC Bioinformatics 10:284.

34. Fuentes M, et al. (2007) Insights into spawning behavior and development of theEuropean amphioxus (Branchiostoma lanceolatum). J Exp Zoolog B Mol Dev Evol 308:484–493.

35. Somorjai I, Bertrand S, Camasses A, Haguenauer A, Escriva H (2008) Evidence for stasisand not genetic piracy in developmental expression patterns of Branchiostomalanceolatum and Branchiostoma floridae, two amphioxus species that have evolvedindependently over the course of 200 Myr. Dev Genes Evol 218:703–713.

36. Lacalli TC, West JE (1986) Ciliary band formation in the doliolaria larva of Florometra.I. The development of normal epithelial pattern. J Embryol Exp Morphol 96:303–323.

Bertrand et al. PNAS | May 31, 2011 | vol. 108 | no. 22 | 9165

EVOLU

TION

Supporting InformationBertrand et al. 10.1073/pnas.1014235108SI Materials and MethodsSynteny Conservation Analysis. Conserved syntenic regions in thegenomes of Monodelphis domestica, Canis familiaris, Gallusgallus, Xenopus tropicalis, and Tetraodon nigroviridis, which areavailable in the Ensembl database (http://www.ensembl.org/index.html), and of Branchiostoma floridae, which is available atthe US Department of Energy Joint Genome Institute (JGI)Web site (http://genome.jgi-psf.org/Brafl1/Brafl1.home.html),were obtained with CASSIOPE (9). When no statistically sig-nificant conserved regions containing an FGF gene were foundin amphioxus, we searched its genome for orthologs of the genesthat were found as part of the conserved regions in at least twovertebrates. Phylogenetic analyses to verify the orthology of eachgene within the conserved regions were performed using RaxMLversion 7.0.4 with the WAG + G + I model, 100 bootstrapreplicates, and the rapid bootstrapping algorithm (10). All theaccession numbers for genes found in the conserved syntenicregions are given in Table S2.

Immunostaining. Embryos were fixed in 4% (wt/vol) paraformal-dehyde as for in situ hybridization but were thereafter kept inPBS + 0.1% Tween 20 at 4° C until use. Three 10-minute washesin PBS + 0.1% Triton X100 were performed, followed by a 1-hincubation in PBS + 1% Triton X100. Embryos were blocked inPBS + 0.1% Triton X100 + 5% (vol/vol) sheep serum + 0.2%BSA for several hours. They were then incubated in blockingsolution with antibodies diluted at a ratio of 1/250 [monoclonalantiacetylated-tubulin produced in mouse (T7451; Sigma) andantiphospho-histone H3 (Ser10) (06-570; Millipore)] overnight at4° C. They were subsequently washed six times for 1 h in PBS +

0.1% Triton X100 and blocked again for several hours. They werethen incubated overnight at 4° C in the blocking solution con-taining secondary antibodies coupled to FITC or Texas Red at aratio of 1/250. Embryos were then washed three times for 10 minin PBS + 0.1% Tween 20 and mounted in glycerol with 2.5%DABCO (Sigma) for photographs.

Pharmacological Treatments. SU5402 and U0126 were dissolved at10−2 M in DMSO. A range of concentrations, ranging from 10to 250 μM, was tested. At 10 μM, there is no effect of SU5402treatment, whereas at >100 μM, development is completely ar-rested. The highest concentration at which we observe a specificeffect is 50 μM, and at this dose, we observe a complete loss ofthe expression of two orthologs of vertebrate FGF signalingtarget genes: Dusp6/7/9 and ER81/Erm/Pea3 (Fig. S8). Wetherefore subsequently performed our experiments using 50 μMSU5402. For U0126, the effect on embryogenesis is specific be-tween 10 and 25 μM; therefore, we used 25 μM except whenspecified otherwise in the text.

Sections. Sections ranging from 1 to 1.5 μM in thickness wereperformed after embedding in Epon resin and subsequentlystained using Ponceau Red or Richardson Blue.

Accession Numbers. Accession numbers of the sequences used forprobe synthesis are as follows: FGF1/2 (EU606032.1), FGF8/17/18(EU606035.1), FGF9/16/20 (EU606036.1), FGFA (EU606033.1),FGFB (EU606034.1), FGFC (EU606038.1), FGFD (HM854710),FGFE (EU606037.1), FGFR (HM854709), Snail (HM359129), andDelta (HM359124).

1. Letunic I, Doerks T, Bork P (2009) SMART 6: Recent updates and new developments.Nucleic Acids Res 37(Database issue):D229–D232.

2. Schultz J, Milpetz F, Bork P, Ponting CP (1998) SMART, a simple modular architectureresearch tool: Identification of signaling domains. Proc Natl Acad Sci USA 95:5857–5864.

3. Thompson JD, Gibson TJ, Higgins DG (2002) Multiple sequence alignment usingClustalW and ClustalX. Curr Protoc Bioinformatics, Chapter 2:Unit 2.3.

4. Galtier N, Gouy M, Gautier C (1996) SEAVIEW and PHYLO_WIN: Two graphic tools forsequence alignment and molecular phylogeny. Comput Appl Biosci 12:543–548.

5. Huelsenbeck JP, Ronquist F (2001) MRBAYES: Bayesian inference of phylogenetictrees. Bioinformatics 17:754–755.

6. Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phylogenetic inference undermixed models. Bioinformatics 19:1572–1574.

7. Yu JK, Meulemans D, McKeown SJ, Bronner-Fraser M (2008) Insights from theamphioxus genome on the origin of vertebrate neural crest. Genome Res 18:1127–1132.

8. Meulemans D, Bronner-Fraser M (2007) Insights from amphioxus into the evolution ofvertebrate cartilage. PLoS ONE 2:e787.

9. Rascol VL, et al. (2009) CASSIOPE: An expert system for conserved regions searches.BMC Bioinformatics 10:284.

10. Stamatakis A, Hoover P, Rougemont J (2008) A rapid bootstrap algorithm for theRAxML Web servers. Syst Biol 57:758–771.

Bertrand et al. www.pnas.org/cgi/content/short/1014235108 1 of 11

Fig. S1. Phylogenetic analysis of the FGF family. Only FGF domains, as predicted by the online SMART software (http://smart.embl-heidelberg.de/) (1, 2), wereused for the alignment. Sequences were aligned automatically using ClustalX (3), with manual correction in Seaview (4). Bayesian inference trees were inferredusing MrBayes 3.1.2 (5, 6), with the WAG + G + I model. Two independent runs of 1 million generations each, sampled every 100 generations with two chains,were performed. A 50 majority rule consensus tree was calculated using a 250,000-generation burn-in.

Bertrand et al. www.pnas.org/cgi/content/short/1014235108 2 of 11

Fig. S2. FGFR expression pattern. The embryos were underincubated in the staining solution to detect the embryonic territories that express FGFR at a higherlevel. A probe for the TK domain was synthesized (accession no. HM854709). (A and B) Four-cell and blastula stage embryos showing no or very low expressionof FGFR. (C) Gastrula stage embryo showing a higher level of expression in the anterior mesendoderm. (D and E) Lateral and dorsal views of a late gastrula/early neurula stage embryo. At this stage, expression is higher in the paraxial mesoderm. (F and G) Lateral and dorsal views of a midlate neurula stage embryoshowing a higher FGFR expression level in the mesoderm, particularly in the most anterior and posterior somites. (H) Late neurula before the mouth opensshowing a higher expression level in the notochord, the posterior somites, and the anterior pharyngeal endoderm. (I) Dorsal view of the posterior part of thespecimen shown in H. (J) Larva showing a higher expression level in the notochord and the anterior pharyngeal endoderm. In the gut, the iliocolonic region isless strongly labeled than the other regions. (K) Enlargement of the anterior part of the specimen shown in J. (L) Enlargement at the level of the iliocolonicregion of the specimen shown in J. Side views are shown except when specified. Anterior is to the left, and dorsal is to the top.

Fig. S3. FGF8/17/18 expression pattern. (A–C) Lateral, blastopore, and dorsal views of a gastrula stage embryo showing expression in the dorsal posteriormesendoderm. (D and E) Lateral and dorsal views of a late gastrula/early neurula stage embryo showing expression in the dorsal mesendoderm, with a higherlevel in the posterior part as described by Yu et al. (7). (F and G) Lateral and dorsal views of an early midneurula stage embryo showing a very high level ofFGF8/17/18 expression in the anterior neural plate and a low level of FGF8/17/18 expression in the posterior mesoderm. (H) Lateral view of a midlate neurulastage embryo with expression in the anterior epidermis and in a ventral and a lateral spot in the pharyngeal endoderm as described by Meulemans andBronner-Fraser (8). (I) Late neurula before the mouth opens showing expression in the anterior epidermis and in two regions of the pharyngeal endodermcorresponding to the mouth and first gill slit anlagen. (J) Ventral view of the anterior part of the specimen shown in I. (K) Larva showing expression of FGF8/17/18 in the cerebral vesicle, around the mouth, in the first gill slit, and in the posterior wall of the tailbud. (L) Enlargement of the specimen shown in K at the levelof the first gill slit. (M) Enlargement of the specimen shown in K at the level of the mouth. Side views are shown except when specified. Anterior is to the left,and dorsal is to the top.

Bertrand et al. www.pnas.org/cgi/content/short/1014235108 3 of 11

Fig. S4. FGF9/16/20 expression pattern. (A) Gastrula stage embryo with a higher level of FGF9/16/20 expression in the posterior dorsal ectoderm. (B and C)Lateral and blastopore views of a late gastrula stage embryo showing expression in the neural plate. (D and E) Lateral and dorsal views of an early midneurulastage embryo with labeling in the neural plate and the pharyngeal endoderm. (F–I) Views of a midlate neurula embryo. (F) Labeling is visible in the neural tubeand in the pharyngeal endoderm. (G) Enlargement of the posterior part of the embryo shown in F. Dorsal view (I) and enlargement of the posterior part (H). (J)Late neurula embryo before the mouth opens showing expression in the pharynx and neural tube. (K) Ventral view of the anterior part of the specimen shownin J. (L) Larva showing FGF9/16/20 expression in the club-shaped gland, the first gill slit, the midgut, and the anus. Enlargement at the level of the tail (M) andat the level of the pharynx (N) of the larva shown in L. Side views are shown except when specified. Anterior is to the left, and dorsal is to the top.

Fig. S5. FGFA expression pattern. (A and B) Lateral and dorsal views of an early midneurula stage embryo with expression in the anterior neural plate and inthe pharyngeal endoderm. (C) Midlate neurula stage embryo showing FGFA expression in the cerebral vesicle and the ventral pharyngeal endoderm. (D and E)Enlarged lateral and ventral views of the specimen shown in C. (F) Late neurula before the mouth opens with labeling in the cerebral vesicle and in thepharyngeal endoderm. (G) Ventral view of the pharyngeal region of the specimen shown in F. (H) Larva showing FGFA expression around the mouth, in theendostyle, in the club-shaped gland, in the first gill slit, and in the anus. (I) Enlargement of the pharyngeal region focused on the mouth. (J) Enlargement ofthe pharyngeal region focusing on the right part. (K) Enlargement of the posterior part of the larva shown in H. Side views are shown except when specified.Anterior is to the left, and dorsal is to the top.

Bertrand et al. www.pnas.org/cgi/content/short/1014235108 4 of 11

Fig. S6. FGFE expression pattern. (A) Midneurula stage embryo with restricted labeling in the first left somite. (B) Dorsal view of the anterior region of thespecimen in A. (C) Section of the specimen shown in A at the level of the restricted labeling. (D–G) Late neurula stage embryos before the mouth opens.Dynamic expression is visible in some specific neurons. (H) Larva showing expression in the neural tube, the gut, and the club-shaped gland. (I) Enlargement ofthe anterior region of the larva shown in H. Side views are shown except when specified. Anterior is to the left, and dorsal is to the top.

Fig. S7. FGFC expression pattern. (A) Late neurula stage embryo showing a higher expression level of FGFC in the anterior pharynx, in the midgut, and in thetailbud. (B) Enlargement of the anterior region of the specimen shown in A. (C) Larva with high FGFC expression level in the club-shaped gland, in the anteriormost part of the pharynx, in the preoral pit, and in part of the endostyle. (D and E) Enlargement of the anterior part of the specimen shown in C focused on themouth and on the right side, respectively. Side views are shown except when specified. Anterior is to the left, and dorsal is to the top.

Bertrand et al. www.pnas.org/cgi/content/short/1014235108 5 of 11

Fig. S8. Embryos treated with SU5402 at the four- to eight-cell stage are not arrested in their development. Immunostaining experiments were performed forWT morula/blastula stage embryos and for embryos treated with SU5402 at the four- to eight-cell stage and fixed at the early midneurula stage, as well as forthe corresponding control embryos. In the morula/blastula embryos (A), all the cells are dividing, as shown by antiphospho-histone H3 immunostaining (A′), butthere are still no cilia, as demonstrated by the absence of labeling after antiacetylated-tubulin immunostaining (A′′). In the treated embryos (C), only somepatches of cells are labeled using the antiphospho-histone H3 antibody (C′), as in the control embryos (B, B′). Enlarged images of labeled nuclei are shownat the lower right in A′–C′ (boxes). The antiacetylated-tubulin staining shows that in treated embryos, cilia are well developed (C′′), as in control embryos (B′′)(SI Materials and Methods).

Fig. S9. Somite morphology after treatments with SU5402. (A, A′) Sections of embryos after in situ hybridization using a probe for MLC, stained with Ri-chardson Blue and Ponceau Red. In late neurula stage control embryos, the somites are clearly visible and express MLC in the pharyngeal (B and C, sections atthe level of a in A) and posterior (D and E, sections at the level of b in A) regions. In the SU5402-treated (treatment 2; Fig. 3A) late neurula stage embryos, nosomites are formed in the pharyngeal region (B′ and C′, sections at the level of a′ in A′), whereas in the posterior region, somite morphology and MLC ex-pression are normal (D′ and E′, sections at the level of b′ in A′). (F, F′ and G, G′) differential optic contrast optics images of the anterior part of late neurula stageembryos, dorsal views. The anterior somite cavities are clearly visible in the control embryos (F and G) as well as in SU5402-treated embryos (treatment 3; Fig.3A) (F′ and G′). (G, G′) Somite cavities are encircled in yellow and notochord in blue.

Bertrand et al. www.pnas.org/cgi/content/short/1014235108 6 of 11

Fig. S10. SU5402 treatment at 50 μM abolishes expression of Dusp6/7/9 and ER81/Erm/Pea3. Embryos were treated from blastula to gastrula or to earlyneurula stage with 50 μM of SU5402. In situ hybridization experiments were performed using probes for Dusp6/7/9 (HM359125) and ER81/Erm/Pea3(HM359126) on treated and control embryos. (A–D) At the gastrula and early neurula stages, Dusp6/7/9 and ER81/Erm/Pea3 are expressed in the dorsalmesendoderm and ectoderm in control embryos. (A′–D′) In SU5402-treated embryos, expression is totally absent at both stages even after several days ofincubation in the staining solution. Dusp6/7/9 and ER81/Erm/Pea3 are the orthologs of well-known FGF pathway target genes in vertebrates. This experimentleads us to suggest, first, that both genes are FGF signaling target genes and, second, that our SU5402 treatment at 50 μM is able to inhibit the pathwaycompletely.

Table S1. Characteristics of eight predicted amphioxus FGF proteins

Name Protein domain Exon–intron

Amphioxus FGF1/2 FGF 35–163, 1.86e-44 1–205 // 206–310 // 311–501Amphioxus FGF8/17/18 Signal peptide 1–31/FGF 55–181, 1.39e-25/low complexity 200–217 1–41 // 42–80 // 81–262 // 263–369 // 370–654Amphioxus FGF9/16/20 Low complexity 4–19/FGF 48 179, 5.35e-68 1–244 // 245–346 // 347–591Amphioxus FGFA Signal peptide 1–16/FGF 56–189, 2.90e-55/low complexity 206–215 1–265 // 266–369 // 370–651Amphioxus FGFB Signal peptide 1–24/FGF 55–185, 4.75e-51 1–98 // 99–262 // 263–564Amphioxus FGFC Signal peptide 1–20/FGF 60–193, 1.70e-38/low complexity 210–232 1–279 // 280–385 // 386–699Amphioxus FGFD Signal peptide 1–26/FGF 51–199, 6.54e-31/low complexity 202–213 1–258 // 259–393 // 394–687Amphioxus FGFE FGF 61–192, 9.71e-57 1–283 // 284–386 // 387–582

Protein domains of the eight amphioxus FGFs, predicted by SMART (http://smart.embl-heidelberg.de/), are indicated (including the e value for the FGFdomain). The exon-intron structure of the coding sequence of the eight amphioxus FGFs is also indicated.

Bertrand et al. www.pnas.org/cgi/content/short/1014235108 7 of 11

Table

S2.

Accessionnumber

ofthegen

esin

theco

nserved

syntenic

regions

Vertebrates

paralogues,

abbreviations

Vertebrates

paralogues,

nam

esHomosapiens

Monodelphis

domestica

Can

isfamiliaris

Gallusgallus

Xen

opus

tropicalis

Tetrao

don

nigroviridis

Proortholog

nam

eBranch

iostoma

floridae

FGF1

/2BBS7

Bardet–Biedl

syndrome7

ENSG

0000

013

8686

ENSM

ODG00

0000

1892

7EN

SCAFG

0000

000

4128

ENSG

ALG

0000

0011

880

ENST

NIG00

0000

1304

9BBS7

1135

23

GRXCR1

Glutaredoxin,

cysteine-rich

1EN

SG00

0002

152

03EN

SMODG00

0000

2482

0EN

SGALG

0000

0023

032

ENST

NIG00

0000

1573

6GRXCR1/2

6370

6

GRXCR2

Glutaredoxin,

cysteine-rich

2EN

SG00

0002

049

28EN

SMODG00

0000

2374

0EN

SCAFG

0000

002

3537

ENSG

ALG

0000

0023

774

ENST

NIG00

0000

1821

2KIAA01

41KIAA01

41EN

SG00

0000

817

91EN

SMODG00

0000

1045

6EN

SCAFG

0000

000

6236

TBLA

STN

ENST

NIG00

0000

1420

6KIAA01

4111

7215

SPATA

5Sp

ermatogen

esis-

associated

5EN

SG00

0001

453

75EN

SMODG00

0000

1889

5EN

SCAFG

0000

000

3988

ENSG

ALG

0000

0011

833

ENSX

ETG00

0000

0157

5SP

ATA

563

613,

1170

82SP

RY1

Sprouty

homolog

1EN

SG00

0001

640

56EN

SMODG00

0000

1889

0EN

SCAFG

0000

000

3979

ENSX

ETG00

0000

0157

3EN

STNIG00

0000

1306

4SP

RY

1171

02

SPRY2

Sprouty

homolog

2SP

RY3

Sprouty

homolog

3SP

RY4

Sprouty

homolog

4EN

SG00

0001

876

78EN

SMODG00

0000

2375

6EN

SCAFG

0000

000

6250

ENSG

ALG

0000

0007

336

ENSX

ETG00

0000

1956

8EN

STNIG00

0000

1820

6FG

F3/7/10/22

C19

orf29

Unch

aracterize

dprotein

C19

orf29

(ren

alcarcinoma

antigen

NY-REN

-24

)(cactin)

ENSG

0000

010

5298

ENSM

ODG00

0000

0683

7EN

SCAFG

0000

001

9208

C19

orf29

1217

14

LMNA

Lamin

A/C

ENSG

ALG

0000

0006

083

Nuclea

rLa

min

1217

09

LMNB1

Lamin

B1

ENSG

0000

011

3368

ENSG

ALG

0000

0006

083

LMNB2

Lamin

B2

ENSG

0000

017

6619

ENSM

ODG00

0000

0482

4EN

SCAFG

0000

001

9389

ENSG

ALG

0000

0000

470

PIAS1

Protein

inhibitor

ofactiva

ted

STAT,

1

ENSG

0000

003

3800

ENSC

AFG

0000

001

7463

ENSG

ALG

0000

0007

970

ENST

NIG00

0000

0954

9PIAS1

/2/3/4

1941

47,27

9826

PIAS2

Protein

inhibitor

ofactiva

ted

STAT,

2

ENSG

ALG

0000

0001

843

PIAS3

Protein

inhibitor

ofactiva

ted

STAT,

3PIAS4

Protein

inhibitor

ofactiva

ted

STAT,

4

ENSG

0000

010

5229

ENSM

ODG00

0000

0076

5EN

SCAFG

0000

001

9154

ENSG

ALG

0000

0001

836

ENST

NIG00

0000

0324

3

Bertrand et al. www.pnas.org/cgi/content/short/1014235108 8 of 11

Table

S2.Cont. Vertebrates

paralogues,

abbreviations

Vertebrates

paralogues,

nam

esHomosapiens

Monodelphis

domestica

Can

isfamiliaris

Gallusgallus

Xen

opus

tropicalis

Tetrao

don

nigroviridis

Proortholog

nam

eBranch

iostoma

floridae

FGF4

/5/6

C12

orf4

Unch

aracterize

dprotein

C12

orf4

ENSG

0000

004

7621

ENSM

ODG00

0000

1829

9EN

SCAFG

0000

001

5354

ENSG

ALG

0000

0017

289

ENSX

ETG00

0000

0596

6EN

STNIG00

0000

0881

3C12

orf4

1266

42,28

3171

CD9

CD9molecu

leEN

SG00

0000

102

78EN

SMODG00

0000

1805

3EN

SCAFG

0000

001

5172

ENSG

ALG

0000

0017

274

ENST

NIG00

0000

1922

3,EN

STNIG00

0000

1268

5

CD9/81

/TSP

AN2

9659

8

CD81

CD81

molecu

leEN

SG00

0001

106

51EN

SCAFG

0000

001

0121

ENSG

ALG

0000

0006

546

ENST

NIG00

0000

1367

8TS

PAN2

Tetraspan

in2

FoxM

1Fo

rkhea

dboxM1

ENSG

0000

011

1206

ENSM

ODG00

0000

1834

9EN

SCAFG

0000

001

5642

ENSG

ALG

0000

001

3420

ENST

NIG00

0000

1139

4Fo

xM23

2011

TSPA

N4

Tetraspan

in4

ENSG

0000

021

4063

ENSC

AFG

0000

002

5574

ENSG

ALG

0000

0006

837

TSPA

N4/9/CD53

2319

99

TSPA

N9

Tetraspan

in9

ENSG

0000

001

1105

ENSM

ODG00

0000

1833

5EN

SCAFG

0000

001

5466

ENSG

ALG

0000

0014

346

ENST

NIG00

0000

1925

8,EN

STNIG00

0000

1267

0CD53

CD53

molecu

leFG

F8/17/18

CCNJ

Cyclin

JEN

SG00

0001

074

43EN

SCAFG

0000

000

8454

ENSG

ALG

0000

0006

955

ENST

NIG00

0000

0402

3CCNJ/JL

9307

5

CCNJL

Cyclin

J-lik

eEN

SG00

0001

350

83EN

SMODG00

0000

0798

7EN

SGALG

0000

0001

455

ENST

NIG00

0000

0465

1

GPA

MGlycerol-3-

phosphate

acyltran

sferase,

mitoch

ondrial

ENSG

0000

011

9927

ENSM

ODG00

0000

1032

6EN

SCAFG

0000

001

0868

ENSG

ALG

0000

0008

795

GPA

M12

6875

,12

6887

GPA

T2Glycerol-3-

phosphate

acyltran

sferase2,

mitoch

ondrial

ENSM

ODG00

0000

0808

8

HPS

EHep

aran

ase

HPS

E1/2

9310

00HPS

E2Hep

aran

ase2

ENSG

0000

017

2987

ENSM

ODG00

0000

0360

2EN

SCAFG

0000

000

9386

ENSG

ALG

0000

0007

508

PEBP4

Phosphatidyletha

nolamine-binding

protein

4

ENSG

0000

013

4020

ENSM

ODG00

0000

0923

5EN

SCAFG

0000

000

9270

PEBP4

1268

94

RHOBTB

1Rho-related

BTB

domain

containing1

ENSG

0000

007

2422

ENSM

ODG00

0000

1709

7EN

SCAFG

0000

001

2932

ENSG

ALG

0000

0003

073

RHOBTB

1/2

9308

8

RHOBTB

2Rho-related

BTB

domain

containing2

ENSG

0000

000

8853

ENSM

ODG00

0000

0922

9EN

SCAFG

0000

000

9245

Bertrand et al. www.pnas.org/cgi/content/short/1014235108 9 of 11

Table

S2.Cont. Vertebrates

paralogues,

abbreviations

Vertebrates

paralogues,

nam

esHomosapiens

Monodelphis

domestica

Can

isfamiliaris

Gallusgallus

Xen

opus

tropicalis

Tetrao

don

nigroviridis

Proortholog

nam

eBranch

iostoma

floridae

FGF1

9/21

/23

BANF1

Barrier

toau

tointegration

factor1

ENSG

0000

017

5334

ENSM

ODG00

0000

2490

7EN

SCAFG

0000

001

3083

BANF1

/220

2557

BANF2

Barrier

toau

tointegration

factor2

BRMS1

Breastcancer

metastasis

suppressor1

ENSG

0000

017

4744

ENSC

AFG

0000

001

2784

BRMS1

/1L

2604

40

BRMS1

LBreastcancer

metastasis-

suppressor1-lik

e

ENSG

ALG

0000

0010

093

DED

DDea

theffector

domainco

ntaining

DED

D1/2

2026

45

DED

D2

Dea

theffector

domain

containing2

ENSG

0000

016

0570

ENSC

AFG

0000

000

4892

FKBP2

FK50

6binding

protein

2,13

kDa

ENSG

0000

017

3486

ENSC

AFG

0000

001

4572

FKBP2

8334

1,12

3439

KAT5

K(lysine)

acetyltran

sferase

5

ENSG

0000

017

2977

ENSC

AFG

0000

001

3299

KAT5

2604

67

POU2F

1PO

Uclass2

homeo

box1

ENSG

ALG

0000

0015

446

POU2F

1/2/3

1179

08

POU2F

2PO

Uclass2

homeo

box2

ENSG

0000

002

8277

ENSC

AFG

0000

000

4911

POU2F

3PO

Uclass2

homeo

box3

ENSG

0000

013

7709

QPC

TGlutaminyl-

pep

tide

cyclotran

sferase

QPC

T/L

2758

83

QPC

TLGlutaminyl-

pep

tide

cyclotran

sferase-

like

ENSG

0000

001

1478

ENSC

AFG

0000

000

4398

SYMPK

Symplekin

ENSG

0000

012

5755

ENSC

AFG

0000

000

4376

,EN

SCAFG

0000

002

3839

SYMPK

2022

68

Bertrand et al. www.pnas.org/cgi/content/short/1014235108 10 of 11

Table

S2.Cont. Vertebrates

paralogues,

abbreviations

Vertebrates

paralogues,

nam

esHomosapiens

Monodelphis

domestica

Can

isfamiliaris

Gallusgallus

Xen

opus

tropicalis

Tetrao

don

nigroviridis

Proortholog

nam

eBranch

iostoma

floridae

FGFR

1/2/3/4

KIAA02

32KIAA02

32EN

SG00

0001

708

71EN

SMODG00

0000

0290

6EN

SCAFG

0000

001

4342

ENSG

ALG

0000

0015

538

KIAA02

3284

842

PSTK

Phosphoseryl-

tRNA

kinase

ENSG

0000

017

9988

ENSM

ODG00

0000

0780

8EN

SCAFG

0000

001

2641

ENSG

ALG

0000

0009

610

PSTK

8487

7

RGS1

0Reg

ulatorofG

protein

signaling

10

ENSG

0000

014

8908

ENSM

ODG00

000

0088

76EN

SCAFG

0000

002

3428

ENSG

ALG

0000

0009

422

RGS1

0/12

/14

8486

9

RGS1

2Reg

ulatorofG

protein

signaling

12

ENSG

0000

015

9788

ENSM

ODG00

0000

0350

2EN

SCAFG

0000

001

4641

,EN

SCAFG

0000

002

3428

ENSG

ALG

0000

0015

626

RGS1

4Reg

ulatorofG

protein

signaling

14

ENSG

0000

016

9220

ENSM

ODG00

0000

0441

3EN

SCAFG

0000

001

6425

TMED

4Tran

smem

brane

emp24

protein

tran

sport

domain

containing4

TMED

4/9/11

1240

27

TMED

9Tran

smem

brane

emp24

protein

tran

sport

domain

containing9

ENSG

0000

018

4840

ENSM

ODG00

0000

0433

7

TMED

11Tran

smem

brane

emp24

protein

tran

sport

domain

containinggen

e

ENSM

ODG00

0000

0439

9EN

SGALG

0000

0015

722

Theab

breviationsan

dnam

es,as

wellas

theaccessionnumbers,forHomosapiens,Monodelphisdomestica,Can

isfamiliaris,Gallusgallus,Xen

opustropicalis,an

dTe

trao

donnigroviridisareacco

rdingto

the

Ensembld

atab

ase(http://www.ensembl.o

rg).Th

eaccessionnumbersforBranch

iostomafloridae

gen

esareacco

rdingto

theJG

Iamphioxu

sgen

omev1

.0(http://gen

ome.jgi-psf.org/Brafl1/Brafl1.home.htm

l).T

he

accessionnumbersaregiven

only

ifthegen

esarein

theco

nserved

syntenic

regionan

dthey

areitalicized

ifthey

areat

more

than

10kb

from

theFG

ForFG

FRgen

esofea

chregion.

Bertrand et al. www.pnas.org/cgi/content/short/1014235108 11 of 11