Tbx4/5 Gene Duplication and the Origin of Vertebrate Paired Appendages

Tbx4/5 gene duplication and the origin of vertebrate paired appendages

Carolina Minguillona,1, Jeremy J. Gibson-Brownb,2, and Malcolm P. Logana,3

aMedical Research Council-National Institute for Medical Research, The Ridgeway, Mill Hill, London, NW7 1AA, United Kingdom; and bDepartment of Biology, Washington University, 1 Brookings Drive, St. Louis, MO 63130

Edited by Clifford J. Tabin, Harvard Medical School, Boston, MA, and approved November 2, 2009 (received for review September 16, 2009) Paired ﬁns/limbs are one of the most successful vertebrate inno- 12). Both genes encode transcription factors that directly regu- vations, since they are used for numerous fundamental activities, late the expression of Fibroblast growth factor-10 (Fgf10) and including locomotion, feeding, and breeding. Gene duplication establish an FGF signaling loop that drives limb outgrowth (7, 9). events generate new genes with the potential to acquire novel To investigate the relationship betwen duplication of a single, functions, and two rounds of genome duplication took place ancestral, Tbx4/5 locus to give rise to separate Tbx5 and Tbx4 loci, during vertebrate evolution. The cephalochordate amphioxus di- and the acquisition of vertebrate paired appendages, we com- verged from other chordates before these events and is widely pared the functions of the amphioxus Tbx4/5 gene, and its used to deduce the functions of ancestral genes, present in single genomic regulatory landscape, to those of the mouse Tbx4 and copy in amphioxus, compared to the functions of their duplicated Tbx5 gene loci. A priori, one can envisage two alternative vertebrate orthologues. The T-box genes Tbx5 and Tbx4 encode scenarios to explain the origin of paired appendages following two closely related transcription factors that are the earliest factors duplication of the single Tbx4/5 gene. First, before duplication of required to initiate forelimb and hind limb outgrowth, respec- the Tbx4/5 gene, mutations within the coding region endowed tively. Since the genetic components proposed to be responsible the descendant paralogous genes, Tbx5 and Tbx4, with the for acquiring a trait during evolution are likely to be involved in the potential to initiate limb outgrowth; specifically, the proteins formation of that same trait in living organisms, we investigated acquired the ability to activate Fgf10 in the LPM. Second, that whether the duplication of an ancestral, single Tbx4/5 gene to give after the divergence of cephalochordates, mutation(s) in the rise to distinct Tbx4 and Tbx5 genes has been instrumental in the regulatory regions of the ancestral Tbx4/5 locus, led ultimately to acquisition of limbs during vertebrate evolution. We analyzed the expression of the vertebrate Tbx4 and Tbx5 genes in the early whether the amphioxus Tbx4/5 gene is able to initiate limb LPM before limb bud stages, when this tissue is competent to outgrowth, and assayed the amphioxus locus for the presence of initiate limb outgrowth (13, 14). To investigate this further, we limb-forming regulatory regions. We show that AmphiTbx4/5 is undertook two complementary approaches: first, we tested the able to initiate limb outgrowth and, in contrast, that the genomic potential of the amphioxus Tbx4/5 protein to initiate limb locus lacks the regulatory modules required for expression that outgrowth in the forelimbless Tbx5 conditional knockout mouse would result in limb formation. We propose that changes at the (10) and second, we tested the ability of the AmphiTbx4/5 and level of Tbx5 and Tbx4 expression, rather than the generation of mouse Tbx5 and Tbx4 genomic regions to drive gene expression novel protein function, have been necessary for the acquisition of in the early LPM by transient transgenesis experiments in the paired appendages during vertebrate evolution. mouse.

evolution ͉ limb ͉ Tbx5 Results We have used our limb-rescue assay (11) to test the ability of the ince Ohno’s visionary hypothesis concerning the origin of amphioxus Tbx4/5 protein to initiate limb outgrowth. We gen- Svertebrate innovations by genome duplication (1), numerous erated three independent transgenic lines (Prx1-Amphi4/5#1–3) reports have been published supporting this concept (reviewed in which the expression of the full-length AmphiTbx4/5 cDNA in ref. 2). The cephalochordate amphioxus, a limbless extant was driven under the control of the Prx1 promoter (Fig. 1 A and invertebrate relative of the vertebrates (3), has been extensively B) (15) in early, presumptive forelimb, LPM. In these transgenic used to study the ancestral functions of genes, present in single lines the amphioxus gene is ectopically expressed in forelimbs as copy in amphioxus, that have been duplicated in the vertebrate well as in hind limbs (Fig. 2 A–C). Analysis of e17.5-rescued lox/lox lineage. Amphioxus exhibits many basal chordate characteristics, embryos (Tbx5 ; Prx1-Cre; Prx1-Amphi4/5) revealed that including the presence of a dorsal nerve cord, a notochord, and AmphiTbx4/5 is able to initiate and maintain limb outgrowth Љ Љ segmented paraxial mesoderm, but lacks many vertebrate char- (Fig. 2 A –C ) as judged by comparison to the ‘‘no-forelimb’’ Ј acteristics such as migratory neural crest cells, a cranium, or an phenotype of the Tbx5 conditional knock-out (Fig. 3A ). Skeletal endoskeleton (4). We sought to investigate the origin of one of preparations of these embryos showed that all of the skeletal the most successful vertebrate innovations, paired appendages, elements along the entire proximo-distal axis of the limb (scap- which include the pectoral and pelvic fins of fish and their derived homologues, the forelimbs and hind limbs of tetrapods. Author contributions: C.M., J.J.G.-B., and M.P.L. designed research; C.M. and M.P.L. per- The T-box genes Tbx4 and Tbx5 are paralogous genes that formed research; C.M. and M.P.L. analyzed data; and C.M., J.J.G.-B. and M.P.L. wrote the arose by duplication of a single, ancestral Tbx4/5 gene. Extant paper. amphioxus possesses a single Tbx4/5 gene (AmphiTbx4/5) (5) and The authors declare no conﬂict of interest. lacks paired appendages, whereas all jawed vertebrates with two This article is a PNAS Direct Submission. pairs of paired appendages have distinct, postduplication Tbx4 Data deposition: The sequence reported in this paper has been deposited in the GenBank and Tbx5 genes. In vertebrates, Tbx5 is expressed in the lateral database (accession no. EU084005). plate mesoderm (LPM) of the presumptive pectoral fin/forelimb 1Present address: Consejo Superior de Investigaciones Cientiﬁcas, Instituto de Biologia region, whereas Tbx4 is expressed in the pelvic fin/hind limb Molecular de Barcelona. c/ Baldiri i Reixac, 10, 08028 Barcelona, Spain. region (reviewed in ref. 6). Tbx5 and Tbx4 are the earliest factors 2Present address: Institute for Evolutionary Discovery, 909 Hiawatha Drive, Mount Pleasant, required for the initiation of limb outgrowth (7–10), and are MI 48858. sufficient to initiate outgrowth in an otherwise limbless state (11, 3To whom correspondence should be addressed. E-mail: [email protected].

21726–21730 ͉ PNAS ͉ December 22, 2009 ͉ vol. 106 ͉ no. 51 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0910153106 Downloaded by guest on September 30, 2021 Fig. 1. Cloning of the amphioxus Tbx4/5 cDNA. (A) Alignment of the deducted amino acid sequence of the AmphiTbx4/5 protein to the mouse Tbx5 and Tbx4 proteins. The N-terminal domain is underlined in blue, the T-domain is underlined in red, and the C-terminal domain is underlined in green. (B) Schematic representation of the Prx1-driven transgenic lines (15). (C) Schematic representation of the Tbx5 and Tbx4 genomic regions cloned in the BGZA reporter vector (16). e1, exon1; e2, exon2; i1, intron1; i2, intron2.

ula, stylopod, zeugopod, and autopod) were present in the mouse regulatory regions required to drive expression of Tbx5 in AmphiTbx4/5-rescued limbs (Fig. 2 AЉ–CЉ). The extent of rescue, forelimb LPM and Tbx4 expression in hind limb LPM. To this from partial (Fig. 2B) to complete (Fig. 2 A and C) is presumably end, we cloned genomic DNA fragments that span regions dependant on the amount of AmphiTbx4/5 protein expressed in upstream of exon3 of the murine Tbx5 and Tbx4 genes into the each independent line. No effect of ectopic amphiTbx4/5 was LacZ reporter vector BGZA (Fig. 1C) (16) for transient trans- observed in the hind limb (Fig. 2 AЉ–CЉ). As additional controls, genesis assays in the mouse. Regulatory elements present Ͻ12 we generated transgenic lines in which the mouse Tbx4, Tbx5, or Kilobase pairs (Kb) upstream of exon3 of the mouse Tbx5 gene

chimeric, domain-swap forms of Tbx4 and Tbx5, were driven by are sufficient to drive expression of the LacZ reporter in the e9.5 EVOLUTION the same Prx1 promoter. All of the rescued embryos for these forelimb LPM (Fig. 4A). Similarly, regulatory sequences Ͻ11 Kb genotypes had forelimbs (Fig. 3) consistent with Tbx4 and Tbx5 upstream of exon3 of mouse Tbx4 are sufficient to drive expres- having common roles in limb initiation, but not having roles in sion of the reporter in the e9.5 hind limb LPM (Fig. 4B). When limb-type specification (11). From these experiments, we con- an equivalent region from the AmphiTbx4/5 locus (Bf4/5–10) was clude that the single AmphiTbx4/5 gene product has the ability used for transgenesis, expression of the reporter in the mouse to initiate limb outgrowth when expressed at the correct time and LPM was never observed (0/12 positive transgenic embryos). To in the appropriate tissue (i.e., the prelimb bud stage LPM). determine whether any LPM-regulatory region was located These data demonstrate that mutations at the level of the coding further away from the amphioxus Tbx4/5 gene, we assayed a sequence cannot account for the limb-forming potential of the bacterial artificial chromosome (BAC; CH302 78M15) contain- mouse and other vertebrate Tbx5 and Tbx4 proteins. ing the entire AmphiTbx4/5 coding region, and 73.3 Kb and 72.3 An alternative scenario is that changes at the level of cis- Kb of flanking region upstream and downstream, respectively, in regulatory regions could account for the acquisition of limbs transient transgenic mice. Whole-mount in situ hybridization on during vertebrate evolution. To address this, we compared the 15 e9.5 transgenic embryos using the full-length amphioxus potential of the amphioxus Tbx4/5 genomic region to drive gene Tbx4/5 as a probe, showed that this BAC did not contain the expression in the LPM, at stages when this territory is competent regulatory regions capable of driving expression in the prelimb to form a limb, to equivalent genomic regions of the murine Tbx5 bud LPM. These experiments indicate that the amphioxus Tbx4/5 and Tbx4 loci. As positive controls, we first sought to locate the genomic region lacks the regulatory elements required to drive

Minguillon et al. PNAS ͉ December 22, 2009 ͉ vol. 106 ͉ no. 51 ͉ 21727 Downloaded by guest on September 30, 2021 Fig. 2. The single amphioxus Tbx4/5 gene rescues forelimb outgrowth. (A–C) Whole-mount in situ hybridization to e10.5 Prx1-Amphi4/5 transgenic embryos for the amphioxus Tbx4/5 cDNA (all are dorsal views with anterior to the top). (AЈ–CЈ) Rescued e17.5 embryos (all are lateral views of the right side). (AЉ–CЉ) skeletal preparations (33) showing the presence of skeletal elements in the Tbx4/5-rescued right forelimbs. au, autopod; ﬂ, forelimb; hl, hind limb; sc, scapula; st, stylopod; z, zeugopod. Fig. 3. Mouse Tbx4, Tbx5, or chimeric domain-swap forms of Tbx4 and Tbx5 rescue forelimb outgrowth. (A) Whole-mount in situ hybridization for Fgf10 gene expression in early, vertebrate LPM, the domain in which to e9.5 Tbx5 conditional knock-out embryo showing expression in hind limbs Tbx5 and Tbx4 are required to execute their limb-forming and lack of expression in the forelimb region (arrow). (B–G) Whole-mount in function. situ hybridization to e10.5 Prx1-transgenic embryos. Probes used are the Finally, we analyzed the expression of AmphiTbx4/5 during C-terminal domain of mTbx5 for B, E, and F, and the C-terminal domain of mTbx4 for C, D, and G.(AЈ–GЈ) e17.5 embryos showing rescue of forelimb amphioxus development (Fig. 4C and ref. 17). Expression of this Ј Ј gene was not observed in embryonic stages. Expression was first outgrowth in all cases (B –G ) in comparison to the absence of a forelimb in the Tbx5 conditional knock-out (AЈ). (AЉ–GЉ) Skeletal preparations showing the detected in late larval (56-h) stages in a posterior, ventral presence of all forelimb elements in all of the rescued embryos (BЉ–GЉ)in mesoderm domain (arrow in Fig. 4C). comparison to the lack of bones in the Tbx5 conditional knock-out (AЉ). A schematic representation of the transgenic construct used for the rescue, as Discussion well as the genotypes of the rescued embryos are indicated. The orientations It is widely accepted that two rounds of whole genome dupli- shown are the same as in Fig. 2. au, autopod; ﬂ, forelimb; hl, hind limb; sc, cation took place early during vertebrate evolution (18). This scapula; st, stylopod; z, zeugopod. genetic material can endow the organism with novel traits, such as in the case of vertebrates, the so-called vertebrate innovations, including two sets of paired appendages. The cephalochordate appendages and possesses a single Tbx4/5 gene (AmphiTbx4/5), amphioxus diverged from other chordates before these genome whereas all vertebrates with paired appendages (from sharks to duplications took place and may therefore be used to deduce the tetrapods) express Tbx5 in the rostral pair and Tbx4 in the caudal functions of ancestral genes, present in single copy in amphioxus, pair (19, 20). We show here that the single AmphiTbx4/5 is able as compared to the functions of their duplicated vertebrate to initiate limb outgrowth when expressed in the correct tem- orthologues. The T-box genes Tbx5 and Tbx4 are required to poral and spatial domains in mice. We propose that changes at initiate forelimb and hind limb outgrowth, respectively. In the the level of the regulation of Tbx5 and Tbx4 expression, rather than absence of Tbx5, forelimbs do not form, and hind limbs fail to the generation of novel protein function, was necessary for the develop in the absence of Tbx4 (9, 10). Amphioxus lacks paired acquisition of paired appendages during vertebrate evolution.

21728 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0910153106 Minguillon et al. Downloaded by guest on September 30, 2021 homologous to vertebrate LPM (2, 27). Hence, a second scenario is that the ancestral Tbx4/5 gene was already expressed in ventral LPM-like tissue but at a later time point during development than its derived vertebrate counterparts, when it is not competent to initiate outgrowth. This suggests that before the duplication of this gene in the vertebrate lineage, but after the divergence of cephalochordates, acquisition of an early enhancer led to the activation of these genes at stages, and in regions, of the LPM that are competent to form a limb. Functional studies on amphioxus embryos have been severely hampered by the short length of the breeding season and, consequently, the limited timeframe during which experimental manipulations can be performed. Recently, however, progress has been made in the induction of spawning in captive animals, which should greatly facilitate the use of amphioxus as an experimental system (28). It will be fascinating to determine whether the ventral mesoderm of amphioxus larvae has any outgrowth potential, and, if so, to investigate the temporal window of this competence, similar to previous experiments, which show that an exogenous source of FGF applied to the chick flank can induce ectopic limb formation (13, 14).

Materials and Methods Embryos and Transgenic Lines. Mouse embryos were staged according to Kaufman (29). Noon on the day a vaginal plug was observed was taken to be e0.5 days of development. Transgenic lines were generated by the Procedural Services Section, NIMR. C57BL/10XCBA/Ca F1 hybrid founders were crossed onto an MF1 background. For pronuclear injection, plasmid DNA was injected at a concentration of 2 ng/mL, while BAC DNA was at 1 ng/␮L. To generate rescued embryos (Tbx5lox/lox; Prx1-Cre; Prx1-gene-of-interest) Tbx5lox/ϩ; Prx1- Cre; Prx1-gene-of-interest mice were crossed to homozygote Tbx5lox/lox mice. The conditional Tbx5 allele, the Prx1-Cre transgene and the whole-mount in situ protocol have been described elsewhere (11). Amphioxus embryos were Fig. 4. Lateral plate mesoderm expression driven by the murine Tbx5 and collected as described in ref. 30. Tbx4 promoter region and expression of AmphiTbx4/5.(A and B)b- galactosidase activity staining of an m5–10 (A) and m4–10 (B) e9.5 transgenic Cloning of Full-Length cDNAs and Chimeras. The full-length sequence of the embryo. (C) Fifty-six-hour amphioxus larva showing expression of Amphi- AmphiTbx4/5 gene was deduced from the sequence of the CH302 78M15 BAC Tbx4/5. Mouse embryos are shown in lateral views of the right side and the (17). We used RT-PCR on a 56-h embryonic cDNA sample to obtain the putative amphioxus larva of the left side. (D) Proposed evolutionary scenario leading to full-length coding region. This PCR product was cloned into pSlax to tag it with the acquisition of paired appendages during vertebrate evolution. The an- the flag epitope. For the murine Tbx4 and Tbx5, as well as the chimeric cestral, single Tbx4/5 gene is represented by a purple box and the postdupli- constructs, we used plasmids containing Tbx4 and Tbx5 cDNAs as templates (a cative Tbx5 and Tbx4 genes are indicated by a red and a blue box, respectively. gift from Benoit Bruneau, Gladstone Institute, San Francisco, CA). All these The orange circle represents the ancestral heart enhancer, whereas the green PCR products were also cloned in pSlax to add the flag epitope 3Ј to the ORFs. circle shows the acquired LPM enhancer. A schematic representation of the We used Advantage HF2 polymerase (Clontech) to perform all of the PCR chordate ancestor and a modern fish-like creature are drawn showing expres- reactions, and all constructs were sequenced. Primer sequences and PCR sion in the heart territory (orange) and in the fin territory (green). fl, forelimb conditions are available on request. To produce transgenic constructs, the LPM; H, heart; hl, hind limb LPM; LPM, lateral plate mesoderm. flag-tagged cDNAs were isolated from pSlax and cloned into the Prx1 promoter construct as described in ref. 11. The schematic representation of the murine constructs in Fig. 3 is as follows: full-length Tbx5 (B) is represented with We also show that in extant amphioxus AmphiTbx4/5 is a blue box, and full-length Tbx4 (C) is shown as a green box. The abbreviations expressed in a posterior, ventral mesoderm domain (Fig. 4C)of for the chimeric constructs are as follows: 5N4T4C (D) is composed of the the 56-h larva. Expression in this domain is consistent with the N-terminal domain of Tbx5 and the T- and C-terminal domains of Tbx4. two following evolutionary scenarios that implicate changes at 4N4T5C (E) contains the N- and T-domains of Tbx4 and the C-terminal domain EVOLUTION the level of regulatory elements. The ventral mesoderm is where of Tbx5. 4N5T5C (F) contains the N-terminal domain of Tbx4 and the T- and the rudimentary amphioxus heart tube forms, and the am- C-terminal domains of Tbx5. 5N5T4C (G) contains the N- and T-domain of Tbx5 phioxus orthologues of many other genes implicated in heart and the C-terminal domain of Tbx4. See Fig. 1 for the subdivision of murine Tbx5 and Tbx4 in these domains. development in vertebrates, including Nk2-tinman and BMP2/4 (21, 22), are expressed in this domain. Therefore, the first Cloning of the Regulatory Regions. Genomic regions m5–10, m4–10, and scenario is that the Tbx4/5 gene may have played an ancestral role Bf4/5–10 were cloned using long-range PCR on BACs containing the mouse in heart development and this function has been retained by the Tbx5 (RP-150D8), mouse Tbx4 (RP23–48A17), or amphioxus Tbx4/5 (CH302– vertebrate Tbx5 and Tbx4 genes, since both are expressed in the 78M15) genes, respectively. These regions were subcloned into the BGZA heart anlagen and are required for heart development (23–26). reporter vector (16) using standard techniques. All constructs were verified by This implies that, during evolution of the Tbx4/5 subfamily, sequencing. mutations occurred at the level of their cis-regulatory regions that account for the acquisition of a expression domain in LPM, Whole-Mount in Situ Hybridization. Whole-mount in situ hybridization of mouse embryos was performed as described in ref. 31 and ref. 32 for am- where Tbx5 and Tbx4 can execute their limb-forming function. A phioxus embryos. Hybridized mouse embryos were photographed using a schematic representation of this evolutionary scenario at the Leica digital camera attached to a Leica dissecting scope. Amphioxus embryos Tbx4/5 locus is presented in Fig. 4D. Alternatively, a few reports were photographed with a Q-Imaging camera attached to a Zeiss Axiophot have suggested that cephalochordate ventral mesoderm could be microscope under DIC/Nomarski optics.

Minguillon et al. PNAS ͉ December 22, 2009 ͉ vol. 106 ͉ no. 51 ͉ 21729 Downloaded by guest on September 30, 2021 ACKNOWLEDGMENTS. We thank David E. K. Ferrier and Peter Osborne for bandry, and Jo Del Buono, Ania Kucharska, Amy Horton, and Ed Oates for help and material for the amphioxus in situs. We are indebted to Sophie technical help and exchange of constructs. This work was supported by the Wood, Procedural Services, National Institute for Medical Research (NIMR) for European Molecular Biology Organization and Medical Research Council pronuclear injections, staff from Biological Services, NIMR for mouse hus- (C.M.).

1. Ohno S (1970) Evolution by Gene Duplication (Springer, Heidelberg, Germany). 18. Dehal P, Boore JL (2005) Two rounds of whole genome duplication in the ancestral 2. Shimeld SM, Holland PW (2000) Vertebrate innovations. Proc Natl Acad Sci USA vertebrate. PLoS Biol 3:e314. 97:4449–4452. 19. Gibson-Brown JJ, et al. (1996) Evidence of a role for T-box genes in the evolution of limb 3. Delsuc F, Brinkmann H, Chourrout D, Philippe H (2006) Tunicates and not cephalochor- morphogenesis and the specification of forelimb/hindlimb identity. Mech Dev 56:93– dates are the closest living relatives of vertebrates. Nature 439:965–968. 101. 4. Holland LZ, Laudet V, Schubert M (2004) The chordate amphioxus: An emerging model 20. Tanaka M, et al. (2002) Fin development in a cartilaginous fish and the origin of organism for developmental biology. Cell Mol Life Sci 61:2290–2308. vertebrate limbs. Nature 416:527–531. 5. Ruvinsky I, Silver LM, Gibson-Brown JJ (2000) Phylogenetic analysis of T-Box genes 21. Holland ND, Venkatesh TV, Holland LZ, Jacobs DK, Bodmer R (2003) AmphiNk2-tin, an demonstrates the importance of amphioxus for understanding evolution of the ver- amphioxus homeobox gene expressed in myocardial progenitors: Insights into evolu- tebrate genome. Genetics 156:1249–1257. tion of the vertebrate heart. Dev Biol 255:128–137. 6. Logan M (2003) Finger or toe: The molecular basis of limb identity. Development 22. Panopoulou GD, Clark MD, Holland LZ, Lehrach H, Holland ND (1998) AmphiBMP2/4, 130:6401–6410. an amphioxus bone morphogenetic protein closely related to Drosophila decapen- 7. Agarwal P, et al. (2003) Tbx5 is essential for forelimb bud initiation following pattern- taplegic and vertebrate BMP2 and BMP4: Insights into evolution of dorsoventral axis ing of the limb field in the mouse embryo. Development 130:623–633. specification. Dev Dyn 213:130–139. 8. Ahn DG, Kourakis MJ, Rohde LA, Silver LM, Ho RK (2002) T-box gene tbx5 is essential 23. Bruneau BG, et al. (1999) Chamber-specific cardiac expression of Tbx5 and heart defects for formation of the pectoral limb bud. Nature 417:754–758. in Holt-Oram syndrome. Dev Biol 211:100–108. 9. Naiche LA, Papaioannou VE (2003) Loss of Tbx4 blocks hindlimb development and 24. Garrity DM, Childs S, Fishman MC (2002) The heartstrings mutation in zebrafish causes affects vascularization and fusion of the allantois. Development 130:2681–2693. heart/fin Tbx5 deficiency syndrome. Development 129:4635–4645. 10. Rallis C, et al. (2003) Tbx5 is required for forelimb bud formation and continued 25. Koshiba-Takeuchi K, et al. (2006) Cooperative and antagonistic interactions between outgrowth. Development 130:2741–2751. Sall4 and Tbx5 pattern the mouse limb and heart. Nat Genet 38:175–183. 11. Minguillon C, Del Buono J, Logan MP (2005) Tbx5 and Tbx4 are not sufficient to 26. Krause A, et al. (2004) Tbx5 and Tbx4 transcription factors interact with a new chicken determine limb-specific morphologies but have common roles in initiating limb out- PDZ-LIM protein in limb and heart development. Dev Biol 273:106–120. growth. Dev Cell 8:75–84. 27. Kozmik Z, et al. (2001) Characterization of amphioxus AmphiVent, an evolutionarily 12. Takeuchi JK, et al. (2003) Tbx5 and Tbx4 trigger limb initiation through activation of conserved marker for chordate ventral mesoderm. Genesis 29:172–179. the Wnt/Fgf signaling cascade. Development 130:2729–2739. 28. Fuentes M, et al. (2007) Insights into spawning behavior and development of the 13. Cohn MJ, Izpisua-Belmonte JC, Abud H, Heath JK, Tickle C (1995) Fibroblast growth european amphioxus (Branchiostoma lanceolatum). J Exp Zoolog B Mol Dev Evol factors induce additional limb development from the flank of chick embryos. Cell 308:484–493. 80:739–746. 29. Kaufman MH (2001) The Atlas of Mouse Development (Academic, Cambridge, UK), 14. Ohuchi H, et al. (1995) An additional limb can be induced from the flank of the chick 2nd Ed. embryo by FGF4. Biochem Biophys Res Commun 209:809–816. 30. Holland NaH, LZ (1993) Essential Developmental Biology: A Practical Approach (IRL 15. Martin JF, Olson EN (2000) Identification of a prx1 limb enhancer. Genesis 26:225–229. Press, Oxford, UK). 16. Summerbell D, et al. (2000) The expression of Myf5 in the developing mouse embryo 31. Riddle RD, Johnson RL, Laufer E, Tabin C (1993) Sonic hedgehog mediates the polar- is controlled by discrete and dispersed enhancers specific for particular populations of izing activity of the ZPA. Cell 75:1401–1416. skeletal muscle precursors. Development 127:3745–3757. 32. Holland L, Holland PWH, Holland ND (1996) Molecular Zoology: Advances, Strategies 17. Horton AC, et al. (2008) Conservation of linkage and evolution of developmental and Protocols (Wiley, New York). function within the Tbx2/3/4/5 subfamily of T-box genes: Implications for the origin of 33. Hogan B, Beddington R, Constantini F, Lacy E (1994) Manipulating the Mouse Embryo vertebrate limbs. Dev Genes Evol 218:613–628. (Cold Spring Harbor Press, Cold Spring Harbor, NY) 2nd Ed.

21730 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0910153106 Minguillon et al. Downloaded by guest on September 30, 2021