A median fin derived from the lateral plate mesoderm and the origin of paired fins

The development of paired appendages was a key innovation during evolution and facilitated the aquatic to terrestrial transition of vertebrates. Largely derived from the lateral plate mesoderm (LPM), one hypothesis for the evolution of paired fins invokes derivation from unpaired median fins via a pair of lateral fin folds located between pectoral and pelvic fin territories1. Whilst unpaired and paired fins exhibit similar structural and molecular characteristics, no definitive evidence exists for paired lateral fin folds in larvae or adults of any extant or extinct species. As unpaired fin core components are regarded as exclusively derived from paraxial mesoderm, any transition presumes both co-option of a fin developmental programme to the LPM and bilateral duplication2. Here, we identify that the larval zebrafish unpaired pre-anal fin fold (PAFF) is derived from the LPM and thus may represent a developmental intermediate between median and paired fins. We trace the contribution of LPM to the PAFF in both cyclostomes and gnathostomes, supporting the notion that this is an ancient trait of vertebrates. Finally, we observe that the PAFF can be bifurcated by increasing bone morphogenetic protein signalling, generating LPM-derived paired fin folds. Our work provides evidence that lateral fin folds may have existed as embryonic anlage for elaboration to paired fins.

The development of paired appendages was a key innovation during evolution and facilitated the aquatic to terrestrial transition of vertebrates. Largely derived from the lateral plate mesoderm (LPM), one hypothesis for the evolution of paired fins invokes derivation from unpaired median fins via a pair of lateral fin folds located between pectoral and pelvic fin territories 1 . Whilst unpaired and paired fins exhibit similar structural and molecular characteristics, no definitive evidence exists for paired lateral fin folds in larvae or adults of any extant or extinct species. As unpaired fin core components are regarded as exclusively derived from paraxial mesoderm, any transition presumes both co-option of a fin developmental programme to the LPM and bilateral duplication 2 . Here, we identify that the larval zebrafish unpaired pre-anal fin fold (PAFF) is derived from the LPM and thus may represent a developmental intermediate between median and paired fins. We trace the contribution of LPM to the PAFF in both cyclostomes and gnathostomes, supporting the notion that this is an ancient trait of vertebrates. Finally, we observe that the PAFF can be bifurcated by increasing bone morphogenetic protein signalling, generating LPM-derived paired fin folds. Our work provides evidence that lateral fin folds may have existed as embryonic anlage for elaboration to paired fins.
Two alternate hypotheses have been proposed to explain the evolutionary origin of vertebrate paired appendages (fins and limbs). Derivation from posterior gill arches was posited by Gegenbaur 3 , whilst a number of anatomists later invoked a rival theory, the lateral fin fold hypothesis. This proposed that paired fins derived (either phylogenetically or ontogenetically) from longitudinal bilateral fin folds that were then subdivided 4,5 . Whilst recent molecular studies have provided some evidence in support of each hypothesis 2,6,7 , there remains significant criticism of the lack of substantiation in the fossil record or in embryology 1,8 . Certain stem vertebrates, including anaspid-related fossils, show evidence of lateral fin folds; however, these fin folds mostly consist of soft tissue with only sporadic skeletal elements and are thus poorly preserved. This has led to conflicting interpretations [9][10][11][12][13] . The developmental programme for paired fins has been postulated to have been first assembled in median fins, which appear in the fossil record before the origin of paired fins 2 . A number of studies have traced the cellular origin of median fins in lamprey, catshark, zebrafish and Xenopus 2,14,15 . All median fins, both larval and adult, assessed so far have shown derivation from the paraxial mesoderm (PM), whilst paired fins are known to be derived from the lateral plate mesoderm (LPM). Thus, the median fin programme was most likely transferred to the LPM from the PM 2 , possibly before the formation of hypothesized lateral fin folds. How or when this transition occurred is unclear. As only a subset of median fins in zebrafish has been assayed, we expanded the characterization of the composition and origin of median fins in zebrafish to determine if PM derivation was an invariant characteristic.
As with most surveyed jawed vertebrates (gnathostomes), larval zebrafish possess two median unpaired fin folds. A caudal fin fold (or major lobe) runs continuously from the dorsal midline around the caudal end of the larva and then ventrally to the anus (Fig. 1a). A pre-anal fin fold (PAFF; or minor lobe) runs along the underside of the yolk sac extension, immediately anterior to the anus 16 (Fig. 1a). The median fin folds are resorbed during metamorphosis, and the caudal fin fold is Article replaced by three separate PM-derived median adult fins 14,17 : the caudal fin, the dorsal fin and the anal fin. No adult median fin replaces the PAFF, which is a developmentally transient structure 16 . The mesenchyme of the caudal fin fold can be labelled by the photoconvertible Kaede green-to-red fluorescent protein expressed under control of the tbx16l promoter, confirming that this median fin is derived from the PM 18 . In contrast, we consistently failed to observe any Kaede labelling in the PAFF ( Fig. 1b and Extended Data Fig. 1a). This labelling difference is not due to de novo activity of the tbx16l promoter specifically in the caudal median fin fold, as we photoconverted Kaede in anterior somites above the PAFF at 24 h post-fertilization (hpf) and traced the photoconverted Kaede to the mesenchyme of the dorsal portion of the caudal fin fold; again, we never found labelled cells in the PAFF (n = 11) (Extended Data Fig. 1b-d). The absence of PAFF labelling by a PM-lineage trace is not due to absence of mesenchyme cells, as we observed an abundance of these cells by Nomarski optics at 3 days post-fertilization (dpf) and in the enhancer trap line Et(krt4:EGFP) sqet37 (ET37), which labels all fin mesenchyme cells (Extended Data Fig. 1e,f). The ET37 line further allowed us to visualize the morphology of the mesenchyme cells, which showed a distally polarized stellate shape, indistinguishable from the mesenchyme of all other fins (Fig. 1c-e and Extended Data Fig. 1f-h). PAFF mesenchymal cells expressed known differentiation markers of fin mesenchyme, including fibulin1 (fbln1) and integrin beta 3b (itgb3b), and were weakly positive for lyve1b reporter activity in the lyve1b:DsRed2 transgenic line, which also expresses in the post-anal ventral fin mesenchyme 19 (Fig. 1f-h and Extended Data Fig. 1i). Hence, despite their divergent developmental origin, PAFF mesenchyme has a similar morphology and expression profile to fin mesenchyme of the major median fin lobe.
Few functions of fin mesenchyme cells are defined. The zebrafish frilly fin (frf) mutants display ruffling of the caudal larval fin fold due to mutations in bone morphogenetic protein 1a (bmp1a), which is expressed in fin mesenchyme 20 , including in the PAFF (Extended Data Fig. 1j). The overall morphology of the PAFF also displays ruffling in frf mutants (Fig. 1i,j). Further, immunostaining of collagen II revealed ordered parallel collagen II fibres in all median fins of the wild type (WT), while these fibres were disorganized in both the caudal median fin folds and the PAFF of frf mutants (Fig. 1i,j), indicating that Bmp1a in PAFF mesenchyme is also required for maturation of collagen. Collectively, these results indicate that mesenchymal cells of the PAFF show functional overlap with those of the caudal median fin fold yet have a distinct developmental origin.
The transcription factor Hand2 is expressed in several LPM progenitors, including cardiac, pharyngeal, mesothelial and pectoral fin progenitors 21,22 . Upon examining the TgBAC(hand2:EGFP) transgenic line that accurately recapitulates endogenous hand2 gene expression 23 , we observed the expected enhanced green fluorescent protein (eGFP) expression in larval pectoral fins at 2-3 dpf (Fig. 2a,b). Although we observed no TgBAC(hand2:EGFP) expression in the major unpaired larval fin lobe, the PAFF mesenchyme was strongly eGFP positive, suggesting an LPM origin (Fig. 2a,b and Extended Data Fig. 2a,b). Additionally, in situ hybridization at 3 and 5 dpf indicated that the PAFF mesenchyme, but no other median fin fold, expressed hand2 (Fig. 2c,d and Extended Data Fig. 2c,d). If the LPM uniquely contributes to the PAFF, then we might expect that the PAFF, but no other median fin fold, would be affected in an LPM mutant. The zebrafish hand2 mutant hands off (han s6 ) exhibits defects in LPM derivatives, including the heart, pectoral fins and mesothelium 21,22 . Consistent with an LPM origin, the  fin height of PAFFs in han s6 mutant larvae was significantly reduced compared with WT at 3 and 5 dpf, whilst the ventral section of the major median fin lobe was unaffected (Extended Data Fig. 3a,b,g,h,m). We observed similar results with hand2 antisense morpholino (MO)-based knockdown (Extended Data Fig. 3c-f,n). hand2 knockdown in the ET37 line severely reduced the number of eGFP-positive mesenchymal cells in the PAFF at 3 and 5 dpf compared with WT, whilst the mesenchyme of the ventral caudal median fin fold was unaffected (Extended Data Fig. 3i-l,o). Transplantation of hand2 MO; ET37 mutant cells into WT suggested that the loss of Hand2 reduced clone size and altered morphology and migration of mesenchyme cells autonomously (Extended Data Fig. 4). Together, these observations are consistent with an LPM origin of PAFF mesenchyme.
To confirm that this expression of hand2 indicates an LPM origin of the PAFF rather than de novo expression, we permanently labelled the LPM in a mosaic fashion using a Tg(drl:creERT2; hsp70l:Switch) transgenic combination 24 (Fig. 2e). During gastrulation and early somitogenesis, drl:creERT2 is expressed in LPM-primed mesoderm with increasing selectivity to the LPM, suitable for lineage labelling of this mesodermal lineage using 4-OH-tamoxifen (4-OHT) induction and loxP reporters 24 . Consistent with an LPM origin of paired appendages, we first documented that this LPM lineage tracing approach labels cells of the paired fins, in particular the larval pectoral fins and pelvic fins. We found lineage labelling in the endoskeletal disc and mesenchyme of 4.5 dpf pectoral fins and also, the intraray fibroblasts and Zns5-positive osteoblasts of the adult pelvic fins (Extended Data Fig. 5a-d). In addition, we observed LPM lineage labelling of the PAFF mesenchyme at 3 and 5 dpf, whilst the labelling in the caudal fin fold was limited to myeloid cells, which are also LPM derived ( Fig. 2f and Extended Data Fig. 5e-g). These data further support the contribution of the LPM to the PAFF mesenchyme but not to other median fin folds. We next used drl-driven expression of the Dendra2 green-to-red photoconvertible fluorophore to localize the precise LPM origin of PAFF mesenchyme. During early to mid-somitogenesis (8-10-somite stage (ss)), drl:H2B-Dendra2 is expressed in the nuclei of the LPM (Extended Data Fig. 6a,b). Photoconversion of posterior-most LPM at this stage, using ultra-violet laser illumination, resulted in selective Dendra2-red labelling of this field of LPM ( Fig. 2g and Extended Data Fig. 6c,d). These cells were then traced to the migrating PAFF mesenchyme at 40 and 48 hpf (n = 4) ( Fig. 2h and Extended Data Fig. 6e-j,q-t), whilst unconverted control cells did not show any Dendra2-red PAFF labelling (Extended Data Fig. 6k-p,u-x). This confirmed PAFF mesenchyme originates from the LPM, with significant contribution from the posterior-most LPM at early segmentation stages.
To investigate whether an LPM-seeded median PAFF is a zebrafish apomorphy or if it is a more ancestral feature shared with other vertebrates, we examined the larvae of other taxa for the presence of a PAFF and used hand2 in situ hybridization as a proxy marker of an LPM origin of constituent mesenchymal cells. Medaka (Oryzias latipes) larvae have a markedly smaller PAFF that possesses only a sparse number of mesenchymal cells as detected by Nomarski imaging (Fig. 3a,b). These mesenchymal cells have cell extensions projecting distally and were mostly proximal in the fin where the interstitial space is thickest (Fig. 3b). Accordingly, medaka hand2 was expressed proximally in a punctate pattern in PAFFs, consistent with mesenchymal expression, and was not expressed in other median fin folds ( Fig. 3c and Extended Data Fig. 7a-c). To confirm this, we expressed eGFP in early medaka LPM by injection of a -6.35drl:EGFP construct. We saw expression in mesenchymal cells in the PAFF transiently and also, in stable transgenic embryos (eight PAFF mesenchyme cells seen in four stage 36 Tg(-6.35drl:EGFP) transgenic medaka embryos) (Extended Data Fig. 7d,e). In addition, we used 1,1′-dioctadecyl-3,3,3′,3′tetramethylindocarbocyanine perchlorate (DiI) lipophilic dye labelling to label the posterior LPM territory at stage 20 (four-somite stage). Consistent with our positional mapping of PAFF-seeding mesenchyme in zebrafish, posterior LPM injection of DiI consistently labelled mesenchyme in the medaka PAFF (total of 27 cells in nine of nine labelled embryos) (Extended Data Fig. 7f-

Fig. 2 | The PAFF is an LPM-derived median fin fold. a,b, Confocal images of
Tg(hand2:EGFP) embryos at 2 dpf (a) and 3 dpf (b) showing eGFP labelling of the mesenchyme of the pectoral (green outline and arrowheads) and PAFFs (yellow outline and arrowheads, and magnified in inset) but not the caudal fin fold (cyan outline). c,d, In situ hybridization of hand2 at 3 dpf shows fin expression of hand2 only in the PAFF (yellow arrowheads (c), and higher magnification with Nomarski optics indicates expression in the mesenchyme (d). e, Schematic of the LPM lineage tracing transgenes. f, Lineage tracing of LPM using transgenics depicted in (e) following 4-OHT treatment and heat shock before imaging shows that PAFF mesenchyme is derived from the LPM (yellow arrowhead and magnified in inset). g,h, Ventral (g) and lateral (h) confocal images of the drl:H2B-Dendra2 transgenic line at the 10-somite stage (10 ss) (g) and 48 hpf (h) following ultra-violet laser photoconversion in the region of the LPM outlined in (g). h, Photoconverted PAFF mesenchyme is indicated by yellow arrowheads. Scale bars, 100 µm (a,c,f); 50 µm (a (inset),f (inset),g,h); 200 µm (b); 20 µm (d).

Article
American paddlefish (Polyodon spathula), also forms a transient larval PAFF 8 ; we identified the expression of P. spathula hand2 in PAFF mesenchymal cells between stages 36 and 39, but hand2 expression was notably absent from the caudal median fin fold (Fig. 3d,e and Extended Data Fig. 7j-l). PAFF hand2 mesenchyme expression persisted until PAFF regression (around stages 45 and 46). From stage 39, we observed hand2 expression in the core of the nascent pelvic fins (Extended Data Fig. 7k,l). We conclude that a larval PAFF with an LPM mesenchyme is ancestral for actinopterygians. To establish if an LPM contribution to median fins occurred before the origin of paired fins, we examined the larvae of a jawless vertebrate, the sea lamprey (Petromyzon marinus).
Lampreys also possess a small transient larval PAFF 25 , within which strong, specific expression of the lamprey hand2 orthologue, HandA, was detected (Fig. 3f,g), although there was evidence of fainter HandA  expression in the dorsal anterior fin. Finally, both the embryos of the sarcopterygian lungfish 26 and amphibian tadpoles (e.g., Xenopus laevis 27 , Xenopus tropicalis and axolotls 15 ) also possess a small transient PAFF (Fig. 3h). Fluorescence in situ hybridization demonstrated individual hand2-positive mesenchymal cells invading this fin in X. tropicalis but not the major median fin lobe at stage 42 ( Fig. 3h-j).
We confirmed this LPM contribution by injecting -6.35drl:EGFP into X. laevis embryos. X. laevis has a larger PAFF than X. tropicalis, and we observed extensive transient eGFP labelling in the PAFF compared with the caudal fin fold (Extended Data Fig. 7m-o). Thus, we show an LPM origin for PAFF mesenchyme in the larvae of representatives of major extant vertebrate linages: cyclostomes, actinopterygians and sarcopterygians. Given the LPM origin of the median PAFF, we considered that it may have importance in the transition from PM-derived median fins to LPM-derived paired fins. We tested if this LPM median fin fold could be duplicated into paired LPM-derived lateral fin folds. Zebrafish larvae mutant for the bone morphogenetic protein (BMP) antagonist chordin develop ventral caudal fin fold duplications, although whether the PAFF is duplicated in these mutants has not been reported 28 . Injection of a low dose of chrd MO into zebrafish recapitulated various ventralized phenotypes, which included individuals with duplicated or multiple PAFFs (Fig. 4a,b,f and Extended Data Fig. 8a,b). We visualized these PAFF phenotypes by chrd MO injection into the ET37 transgenic line, which demonstrated that each fin fold of duplicated PAFFs contained mesenchyme (Fig. 4c,d). Lineage tracing using the Tg(drl:creERT2; hsp70l:Switch) transgenic combination injected with chrd MO demonstrated that the mesenchyme cells of the duplicated PAFFs are indeed derived from the LPM (Fig. 4e). By generating chrd morphants in the Tg(hand2:EGFP) line and then imaging by light-sheet microscopy, we generated three-dimensional reconstructions documenting duplicated PAFFs along the yolk extension, all of which harboured eGFP-positive mesenchymal cells, further demonstrating that the mesenchyme of these duplicate fin folds was LPM derived (Fig. 4g-i,  Extended Data Fig. 8c-j and Supplementary Video 1). Of note is the generation of multiple parallel fin folds in individual embryos following chrd MO injection (Fig. 4f). These observations indicate that bifurcation of the LPM-seeded PAFF readily arises from modulating ventral BMP signalling, such as by reducing the dose of the dorsal inhibitor Chordin.
To determine if the generation of multiple PAFFs could have evolved spontaneously as part of natural variation, we examined the twin-tail goldfish strain, Ranchu, which has been shown to display bifurcated fin folds including PAFFs due to a loss-of-function mutation in a chordin paralogue (chdA E127X ) [29][30][31] . The bifurcated PAFFs and caudal fin folds of larval Ranchu appeared similar to those of zebrafish larvae injected with low doses of chrd MO (Fig. 4j and Extended Data Fig. 8k,l). These PAFFs were all populated by mesenchymal cells (Fig. 4k). Using hand2 in situ hybridization as a proxy of LPM lineage to infer the origin of the Ranchu PAFF mesenchyme, we observed that the core of both duplicated PAFFs expressed hand2, whilst the caudal fin fold did not (Fig. 4l-n and Extended Data Fig. 8m,n). This suggests that, as in the zebrafish PAFF, the bifurcated Ranchu PAFFs are also LPM derived. Notably, in addition to those displaying simple PAFF duplication, some individuals had three or more parallel PAFFs ( Fig. 4j and Extended Data Fig. 8n). Thus, we conclude that LPM-derived paired fin folds that can be readily generated in zebrafish larvae have also arisen spontaneously in twin-tail goldfish; thus, they represent a viable morphological innovation. Notably, duplicated LPM-derived PAFFs were situated in ventrolateral locations between the future pectoral and pelvic fin domains, overlying the proposed competence stripes of appendage formation in gnathostomes 32 .
Combining several models and techniques, we here have identified a larval median fin with a mesenchyme core derived from LPM and propose it as an intermediate between PM-derived median fins and LPM-derived paired fins. While apparently conserved from cyclostomes to amphibia, most extant chondrichthyan embryos do not appear to have a discernible PAFF. We used micro-computed tomography (microCT) to examine prehatching stages of the elasmobranch Epaulette shark Hemiscyllium ocellatum, which did not show the presence of a clear pre-anal fin (Extended Data Fig. 9a-i). Although this pattern appears to support a loss of the PAFF somewhere in the chondrichthyan lineage, we cannot rule out the possibility that a PAFF arose independently in cyclostomes and Osteichthyes. It is interesting to note that embryos and adults of the frilled shark, Chlamydoselachus, do show bifurcated tropeic folds in the ventral midline (Extended Data   Fig. 9j-m), structures previously invoked as supporting the fin fold hypothesis 33 .

Article
PAFFs are mostly transient larval structures that generally lack a mineralized skeleton, which may explain their poor documentation in the fossil record. Although a true PAFF does not persist into adulthood in most species, it is seen in hagfish adults, where unlike the caudal median fin, it does not possess cartilage rays 34 . In the fossil record, a pre-anal fin is seen in some stem vertebrate fossils, including Haikouella and Haikouichthys, while the pre-anal fin of Kerreralepis shows well-developed plates 35,36 . Zebrafish larval fin mesenchyme is known to persist and contribute to the lepidotrichia of adult fins 18 , and LPM-derived dermal fibroblasts form cartilage during axolotl limb regeneration 37 . These data suggest the potential of the PAFF mesenchyme to generate skeletal outcomes.
Spontaneous duplication in twin-tail goldfish species indicates that paired LPM-derived PAFFs could have readily arisen during evolution. In Ranchu, we frequently observed multiple PAFFs, permitting retention of a median PAFF and divergence of duplicates into paired fins. Although there is no evidence of lateral fin folds present in extant vertebrates, adult specimens of the hagfish species Neomyxine biniplicata possess anterior lateral paired folds that terminate close to the pre-anal fin 38 , resembling partial duplications of the pre-anal fin, although their relevance to paired fin evolution is disputed 39 . In the fossil record, there is evidence that both mineralized and unmineralized lateral paired fin folds were common in anaspids, although homology of anapsid and gnathostome paired fins is highly contentious, as is their relevance to paired fin evolution and the lateral fin fold hypothesis 1,11 . Pharyngolepis, Cowielepis and Euphanerops display pre-anal ventrolateral paired ribbon structures or triangular fins with radials, while in Jamoytius, the ribbons lacked mineralized structures 12,40,41 . The galeaspid, Tujiaaspis, has been recently described bearing skeletal ventrolateral fins composed of skeletal units 42 . Subsequent regionalization of these adult paired elongate fins through anterior restriction to the pectoral fin region is proposed and exemplified by Pharyngolepis and Rhyncholepis 11,43 .
Our work gives weight to a model for paired fin evolution through co-option of a larval median fin programme to the LPM followed by fin duplication and subsequent regionalization to pectoral and pelvic fins. We present one possible model for how the PAFF may have led to the generation of LPM-derived paired fins (Fig. 5). The PAFF may have originated as a small ectoderm-only fin fold (as seen in amphioxus larvae 44 ) with an LPM contribution evolving subsequently upon alterations in LPM topology, such as persistence of a somatopleure and/or a lateral mesodermal divide 45,46 . Similar LPM tissue contexts may have then led to PAFF elongation and following duplication, paired fin regionalization. As the PAFF possesses characteristics of both unpaired and paired fins, the PAFF may represent a novel evolutionary module or at least demonstrate components of the developmental mechanisms that contributed to the emergence of paired appendages.

Online content
Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/s41586-023-06100-w. Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Animal husbandry and experiments
Adult zebrafish and medaka were maintained under standard conditions, and embryos were obtained through natural crosses, kept in E3 medium at 28.5 °C and staged according to Iwamatsu 47 and Kimmel et al. 48 . Embryos of twin-tail goldfish (Ranchu strain) were purchased from a local breeder in Singapore, cultured in E3 medium (with methylene blue) at room temperature and staged as per Li et al. 49 . The experimental procedures and protocols were approved by the institutional animal care and usage committees (IACUCs) at the Institute of Molecular and Cell Biology, Agency for Science, Technology and Research, Singapore (IACUC number 140924); Nanyang Technological University (IACUC number A18002); University of Zurich and the veterinary office of the Canton of Zurich; and the University of Colorado School of Medicine Anschutz Medical Campus (protocol number 979). Paddlefish embryos (P. spathula) were obtained and staged as previously described 8 , and experiment protocols and animal care were approved by James Madison University (IACUC number 20-1601). Lamprey embryo work was performed based on published protocols and staged per Tahara 50 , with experimental protocols approved by the IACUC of the California Institute of Technology (IACUC number 1436). Experiments involving X. tropicalis and X. laevis were approved by the local ethics committee at the University of Manchester and the Home Office (number PFDA14F2D). X. laevis embryos were injected with 100 pg of the -6.35drl:EGFP construct 24 together with 400 pg of in vitro-transcribed membrane-mCherry messenger RNA in each blastomere at the two-cell stage. Medaka embryos were obtained from the National University of Singapore under IACUC numbers BR19-0120 and BR22-1497. Embryos of Epaulette shark were obtained from brood stock at Monash University as approved by the Monash Animal Ethics Committee under license (number 30347) and staged per Ballard et al. 51 .

In situ hybridization
In situ hybridization on zebrafish, medaka and goldfish embryos was conducted as described 55 . Probes used were fbln1 (ref. 56), itgb3b 57 , bmp1a 20 , zebrafish hand2 (ref. 22) and medaka hand2 (ref. 58). Goldfish hand2 probe was generated by polymerase chain reaction using goldfish complementary DNA and the primer sequences (5′-3′) ACGTTTTATGGGGAGACAACC and TAATACGACTCACTATAGGGTCTT CCTTGGCGTCTGTCTT, where the T7 RNA polymerase site is underlined. The transverse section of whole-mount in situ hybridization was achieved via cryosections. Briefly, stained embryos were embedded in agar sucrose and cryoprotected in 30% sucrose. Samples were embedded in OCT medium (Tissue-Tek) and cut into 16-to 20-µm sections in a Leica cryostat (CM3050S). Paddlefish in situ hybridization was performed as previously stated, and the paddlefish hand2 probe was used 59 . Whole-mount chromogenic in situ hybridizations and fluorescent in situ hybridizations and sections were performed as previously described 60,61 on X. tropicalis larvae using a hand2 antisense probe from the expressed sequence tag (EST) clone (TTba012k13) identified from the X. tropicalis full-length EST library 62 . Lamprey in situ hybridization was accomplished using HandA as a probe according to the published protocol 63 .

Transplantation
The hand2 cell autonomy experiments were performed by transplanting marginal cells at gastrula stages from ET37 hosts to the margin of unlabelled WT hosts as previously described 64 . Donors were injected with 50 pg H2B-mCherry messenger RNA either with or without 500 µM hand2 MO. Chimeras were then raised, and PAFF mesenchyme was imaged at 3 dpf.

DiI injection of medaka
CellTracker CM-DiI Dye (C7000; Invitrogen) was microinjected using filamented glass capillaries at a concentration of 1 µg µl −1 into the LPM, just posterior to Kupffer's vesicle at the approximately four-somites stage using Eppendorf FemtoJet.

Immunostaining
Antibody staining was executed as previously described 20

Microscopy
Confocal images were taken on an LSM800, LSM880 or LSM980 Zeiss microscope. For high magnification of bright-field or Nomarski images, a Zeiss AxioImager M2 was used. A Zeiss AxioZoom V16 was utilized to capture low magnification of bright-field images. Light-sheet fluorescence microscopy was undertaken using a Zeiss Lightsheet Z.1, and sample preparation and experimental procedures were performed based on the manufacturer's manual. Microscopy images (fluorescence and bright field) were analysed by ZEN Blue v.3.6 (Zeiss), Fiji (ImageJ v.1.52p) and IMARIS v.9.9.1 (Oxford Instruments).

High-resolution X-ray computed tomography
Shark samples previously fixed in 4% paraformaldehyde (PFA) and dehydrated for freezer storage were stained with 1% I 2 dissolved in ethanol according to Metscher 65 . X-ray computed tomography scans were done using a Zeiss Xradia 520 Versa system. Images were generated using Avizo 3D software (ThermoFisher Scientific).

Photoconversion
To track cell lineage, Kaede and Dendra-2 photoconversion was conducted as previously described 18 . In brief, 24 hpf Kaede or Dendra-2 green-expressing embryos were mounted in 3% methyl cellulose or 0.5% low-melting point agarose and imaged by confocal microscopy. A selected region of interest was photoconverted to express red fluorescence using 405-nm ultra-violet laser illumination. Embryos were imaged immediately after photoconversion by sequential acquisition using 488-and 561-nm-wavelength channels to confirm successful conversion. Then, the same embryos were re-imaged at 40, 48 or 72 hpf.

Tamoxifen treatment
To induce Cre activity in Tg(drl:creERT2, hsp70l:Switch) transgenic embryos, 4-OHT was added in E3 (final concentration of 10 µM) at the 12-somite stage. Then, 4-OHT was washed off at 24 hpf, and embryos were raised until the desired stage. For specific embryonic and larval time points, embryos were heat shocked for 1 h at 37 °C to initiate the expression of eGFP (2-3 h before fixation). Subsequently, embryos were sampled, fixed and stained with 4′,6-diamidino-2-phenylindole (DAPI), and confocal imaging was undertaken.
Adult zebrafish were heat shocked for 16 h at 37 °C, anesthetized with Tricaine and euthanized in an ice-water bath. Successfully switched individuals were selected based on eGFP-expressing heart (major derivate organ of the drl-expressing LPM), and the paired pectoral and pelvic fins and anal fin were dissected and fixed overnight with 4% PFA.

Statistical analysis and reproducibility
Statistical analysis was performed using Prism v.9 (GraphPad). A two-tailed Mann-Whitney test was used when two conditions were compared. No statistical methods were used to predetermine sample size. All experiments were performed at least three times on different weeks with different biological samples. The experiments were not randomized, and investigators were not blinded. Consistent labelling within each batch of embryos was confirmed, and representative samples were used for imaging.

Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data Availability
The authors confirm that all relevant data are provided in this paper and in its Extended Data files. The data for measurements of fin size and cell number in Extended Data  Fig. 8 | Reduced Chordin leads to paired duplication of PAFF. a-b, Low (a) and high (b) power Nomarski images of duplicated PAFFs (yellow arrowheads) in 5 dpf larvae injected with chrd morpholino. c-j, Confocal images of PAFFs from 8 dpf Tg(hand2:EGFP) either uninjected (c,e) or injected with chrd morpholino (d,f,g-j). Confocal projections (c,d,i,j) and total surface renderings (e,f,g,h) highlight duplicated PAFFs (blue and yellow, f) compared to single PAFF in uninjected larvae (grey, e). Imaging of left (g) and right (h) duplicated PAFFs of chrd morphants indicated eGFP positive mesenchymal cells populated both fin folds (yellow arrowheads, i,j). k-l, Lateral (k) and ventral (l) Nomarski images of the duplicated PAFFs of the Ranchu goldfish strain at 6 dpf. Duplicated PAFFs and ventral caudal fin folds indicated by yellow and cyan arrowheads, respectively. m-n, Lateral (m) and transverse (n) views of 7 dpf Ranchu larvae stained by in situ hybridisation for hand2. Expression in individual mesenchymal cells of the PAFF indicated by yellow arrowheads (m). Occurrence of three PAFFs with core hand2 expression indicated by yellow arrowheads (n). Scale Bars: 100 µm (a,e,g), 50 µm (b), 500 µm (k), 200 µm (l), 20 µm (m,n).

Corresponding author(s): Tom J. Carney
Last updated by author(s): Apr 9, 2023 Reporting Summary Nature Portfolio wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency in reporting. For further information on Nature Portfolio policies, see our Editorial Policies and the Editorial Policy Checklist.

Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.

n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly The statistical test(s) used AND whether they are one-or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section.
A description of all covariates tested A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals) For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors and reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Portfolio guidelines for submitting code & software for further information.

Data
Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: -Accession codes, unique identifiers, or web links for publicly available datasets -A description of any restrictions on data availability -For clinical datasets or third party data, please ensure that the statement adheres to our policy The authors confirm that all relevant data are provided in the results section of this paper, in its Extended Data files and in the Source Data file. The data for