Adult caudal fin shape is imprinted in the embryonic fin fold

Summary

Appendage shape is formed during development (and re-formed during regeneration) according to spatial and temporal cues that orchestrate local cellular morphogenesis. The caudal fin is the primary appendage used for propulsion in most fish species, and exhibits a range of distinct morphologies adapted for different swimming strategies, however the molecular mechanisms responsible for generating these diverse shapes remain mostly unknown. In zebrafish, caudal fins display a forked shape, with longer supportive bony rays at the periphery and shortest rays at the center. Here, we show that a premature, transient pulse of sonic hedgehog a (shha) overexpression during late embryonic development results in excess proliferation and growth of the central rays, causing the adult caudal fin to grow into a triangular, truncate shape. Both global and regional ectopic shha overexpression are sufficient to alter fin shape, and forked shape may be rescued by subsequent treatment with an antagonist of the canonical Shh pathway. The induced truncate fins show a decreased fin ray number and fail to form the hypural diastema that normally separates the dorsal and ventral fin lobes. While forked fins regenerate their original forked morphology, truncate fins regenerate truncate, suggesting that positional memory of the fin rays can be permanently altered by a transient treatment during embryogenesis. Ray finned fish have evolved a wide spectrum of caudal fin morphologies, ranging from truncate to forked, and the current work offers insights into the developmental mechanisms that may underlie this shape diversity.

Introduction

The development and regeneration of biological shapes requires precise deployment and temporal interpretation of spatial signals (1,2); developmental shifts that alter ultimate shape can profoundly impact the function of an organ, with disordered or adaptive effects (3). The homocercal caudal fin is a major evolutionary innovation of teleosts (ray finned fish), and shows an elegant external skeletal structure that is complex enough to be developmentally informative, yet simple enough that essential aspects of form may be mechanistically disentangled. The shape of the external caudal fin is primarily derived from the difference in length between the outer dorsal and ventral (peripheral) and central fin rays. This overall shape varies considerably across species with different swimming ecologies, and the shape of the fin corresponds to different hydrodynamic tradeoffs (4–6). Across the spectrum of caudal fin diversity, there are two classes of fin shapes. Some teleosts possess triangular–truncate–shapes (e.g. in medaka, trout) with central fin rays as long as or longer than the peripheral rays; this morphology provides a relatively large surface area for efficient acceleration (5–7). Other teleost groups have evolved a forked fin shape (e.g. tuna, carp), with central rays that are relatively shorter than peripheral rays; this forked shape reduces the overall fin surface and is believed to maximize either efficient cruising or stability (4,5,8,9). The zebrafish caudal fin exhibits a distinctly forked shape and, along with the other fins, is intensively studied as a model for skeletal growth regulation and regeneration (e.g. see 10–12).

The zebrafish caudal fin is characterized by mirror-image symmetry of the rays, reflected around the central hypural diastema, a cleft that separates the central-most skeletal elements (13–15). This external symmetry contrasts with the highly asymmetric caudal fin endoskeleton, where most fin rays are supported by hypurals—modified ventral spines (16,17). During development, central caudal fin rays appear first, ossifying in pairs around the hypural diastema and ventral to the notochord, with peripheral rays appearing later in sequence (13,18). The notochord flexes upward as the fin develops, re-orienting the organ from ventral to posterior and ultimately establishing the dorsoventrally-symmetrical organ (13,15,19). Zebrafish fins are highly regenerative, and the caudal fin can regrow to its original size and forked shape within weeks of amputation (20–22).

The outgrowth of the caudal fin is initiated by pulses of cell proliferation at the distal end of the caudal fin fold mesenchyme (23,24). Skeletal precursors differentiate into osteoblasts that secrete the mineralized collagenous matrix that forms the fin rays (25,26). Signaling pathways such as Wnt, Shh, and BMP, among others, regulate the timing of skeletal differentiation, proliferation, and migration throughout the median fin fold as the organ develops and regenerates (12,27–32). Previous studies have identified several factors that act at the organ level to shape and pattern the appendage: Hox factors initially govern fin ray length, number and identity (14,33); ion channels and gap junctions govern fin ray growth by modulating tissue-level bioelectricity (34–37); thyroid hormone and osteoclast activity regulate patterning of the rays and location of ray branches (38,39). Disrupting any of these pathways profoundly disrupts the phenotype of the fin; notably however, the forked shape of the organ remains remarkably consistent even if length or skeletal patterning (or both) are disrupted (38,40). Unlike length and skeletal patterning, the developmental pathways that regulate caudal fin shape remain unresolved.

In tetrapod limbs, Shh establishes anteroposterior axis patterning, regulating limb growth and posterior skeletal identities (41–44), making the pathway a strong candidate to regulate caudal fin shape. Although shh is not expressed in the early caudal fin fold primordium (45), the morphogen is produced later as the skeleton develops, initially expressed along the rays and eventually localizing to the growing distal tips (45–47). During both development and regeneration, the Shh pathway promotes ray branching by trafficking pre-osteoblasts distally with migrating basal epidermis (29,47,48). However, despite the involvement of Shh in patterning vertebrate appendages and specifically regulating fin ray growth, the pathway has not previously been shown to contribute to establishing the shape of the caudal fin. Here, we demonstrate that modulating Shh in the embryonic fin fold is capable of producing a novel caudal fin shape, shifting zebrafish from a forked to a truncate caudal fin morphology.

Results

Transient, premature shha overexpression during embryonic development alters shape and pigmentation of the caudal fin

In wild-type (WT) zebrafish caudal fins, the shortest central rays are ~65% the length of the longest peripheral rays, creating a forked shape (Fig. 1A–B, E). To investigate the role of Shh in the development of this shape, we used the hsp70l:shha-EGFP transgenic zebrafish line (49) to drive transient, precocious shha overexpression by heat shock before any of the caudal fin skeleton begins to form: 2 days post-fertilization (dpf); hereafter this treatment is referred to as “shha pulse”. After treatment with a shha pulse, the central rays grew to nearly the same length as the peripheral rays (often >85%), resulting in a truncate fin shape reminiscent of the caudal fins of medaka or killifish (Fig. 1C–E). Notably, while the shape of the fin was changed, the overall size (as measured by the longest peripheral rays) was unchanged (Fig. S1). Fin shape showed no correlation with sex (Fig. S2). Fish treated with a shha pulse exhibited fewer principal rays, varying between 8 and 17 instead of the typical 18 (Fig. 1F), and showed a striking loss of the hypural diastema (Fig. 1D, G).

Fig. 1: Pulse of premature shh during early fin fold development disrupts adult caudal fin shape.

(A-B) Caudal fins of control zebrafish and (C-D) transgenic zebrafish subjected to transient shh overexpression 2 dpf (shh pulse). (B and D) are cleared and stained caudal fins from juvenile zebrafish. Dashed outlines indicate the overall shape of the fins. Arrow indicates the location of the hypural diastema separating the dorsal from ventral lobes in B; asterisk indicates the absence of the diastema in (D). (E) The inheritance of the hsp70l:shha-eGFP transgene and the activation of the promoter by heat shock are both necessary in order to induce the truncate fin phenotype. Significance determined by ANOVA followed by Tukey’s post hoc test. An embryonic shh pulse (F) increases the number and variance of principal fin rays and (G) causes a loss of the hypural diastema. Significance determined using Welch’s two-sample T-tests. (H-J) Treatments with the Smoothened inhibitor BMS-833923 after shh pulse can partially rescue both (I) forked fin shape and (J) horizontal stripes of pigmentation. Significance determined by ANOVA followed by Tukey’s post hoc test. Scale bars, 1 mm.

To determine if the phenotypic effects of a shha pulse were mediated through the canonical Shh signaling pathway (29,48,50), we inhibited the Shh effector Smoothened with the antagonist drug BMS-833923, 24 and 48 hours after the global shha pulse (48,51,52). This inhibition partially rescued the wild-type forked fin shape to shha pulsed fish (Fig. 1H–I), demonstrating that the truncate phenotype is acquired through disruption of canonical Shh signaling.

The shha pulse also caused a dramatic shift in pigment pattern: truncate fins developed stripes organized in vertical arches rather than in the stereotypical pattern of horizontal stripes (Fig. 1A, C). The pigment pattern induced by shh pulse is reminiscent of the vertical bars on the fins of certain species with evolved truncate fins, including clownfish (53) and some Corvis wrasses including Corvis flavovittata, which develop horizontal stripes on the body and vertical pigmentation on the rounded fin (D. Parichy, personal communication; (54,55)). This pigment disruption in the shh-pulsed group was partially rescued by treatment with BMS-833923, as well (Fig. 1H, J).

Transient embryonic shha overexpression alters fin shape in a dose-dependent, local manner

To discern the developmental window during which the shha pulse induces truncate fin development, we heat-shocked embryos and larvae at different days post-fertilization. Transgenic embryos heat-shocked at 2 or 3 dpf, developed truncate fins, but not at 4 dpf or later (Fig. 2A). We note that the effect of the heat-shock promoter diminishes at later stages of development, which may contribute to the observed critical window (Fig. S3A). We then quantified the duration of excess shha following the 2-dpf pulse using RT-qPCR. Six hours after heat shock, shha mRNA levels increased approximately 10-fold and returned to control baseline by 24 hours after treatment (3 dpf; Fig. 2B). GFP fluorescence remained detectable for 4 days after induction (6 dpf; Fig S3B), consistent with the stability of the c-terminal Shh-GFP fusion protein that remains in the cytoplasm following cleavage of the Shh protein and secretion of the active n-terminal fragment (49). In WT larvae, the earliest detected Shh signaling activity in the caudal fin fold coincides with fin ray emergence, typically around 8 dpf (13,45,56), two days after transgenic Shh-GFP levels have returned to baseline in the shha-pulsed larvae.

Fig. 2: Transient embryonic shh pulse disrupts caudal fin development in a local and dose-dependent manner.

(A) shh pulse results in truncate fin development when zebrafish are heat shock induced on 2 or 3 dpf. Significance determined by ANOVA followed by Tukey’s post hoc test. (B) Overexpression of shh during hours following heat shock. Significance determined using Welch’s two-sample T-tests, and the correlation between readout and time following heat shock determined by linear-mixed effects model. (C-E) Locally induced shha pulse is sufficient to induce truncate phenotype. (C-C’) Embryo subjected to local posterior heat shock at 2 dpf did not show GFP fluorescence and grew into an adult with a forked fin. (D-D’) Local posterior heat shock induced GFP in transgenic embryo (brackets), which grew into an adult with a truncate fin.Scale bars, 500µm. (E) Local posterior heat shocks in transgenic embryos are capable of inducing truncate fin shape. Inducing local shh pulse in the anterior of the embryo produces no change in fin shape. Significance determined by ANOVA followed by Tukey’s post hoc test. (F) Fish sorted by relative brightness of GFP expression 1 day after whole-body HS (4 dpf) show different caudal fin shapes as adults. Shown below the graph are representative images of individuals in each brightness category. Significance determined by ANOVA followed by Tukey’s post hoc test. Scale bar, 1 mm. (G) Quantified copy number of GFP transgene amplified from genomic DNA correlates with caudal fin shape. Significance between mean Rq and caudal fin shape is determined by linear-mixed effects model.

As a global shha pulse induced a truncate caudal fin phenotype, we predicted that shha overexpression solely at the posterior end of the tail would similarly produce a truncate fin. We locally activated the hsp70l:shha-gfp transgene using local heat shock (57) at two anteroposterior locations along the body axis. As predicted, only localized Shh activation at the posterior end of the tail (adjacent to the region where fin rays will develop (13)) was sufficient to induce truncate fin development (Fig. 2C–E).

We asked if the severity of the truncate phenotype correlated with the amount of shha transgene activation. Indeed, we found that GFP brightness following heat shock (Fig. 2F) as well as the quantity of genomic gfp (quantified by qPCR; Fig. 2G) each predicted the severity of the truncate phenotype. These relationships suggest that greater abundance of genomic shha-gfp acts in a dose-dependent manner to induce the aberrant fin shape.

Caudal fin shape is established by regional differences in cell proliferation and growth rates

To examine the skeletal basis of the shha pulse-induced fin abnormalities, we tracked fin ray ossification and hypural chondrogenesis throughout larval development (Fig. 3A–B). In control larvae, hypurals appear from anterior to posterior, and fin rays appear sequentially in pairs around the hypural diastema from central to peripheral (see (13,15)). In shh-pulsed fish, the hypural complex was malformed and lacked a diastema as soon as hypurals appeared (Fig. 3A’–B’), while fin ray growth was delayed (Fig. 3C).

Fig. 3: Divergent fin shape is accompanied by disruptions in skeletal growth and presaged by differences in regional cell proliferation.

Development of the caudal skeleton in (A) control and (B) shh pulsed larvae. sp7 reporter-expressing osteoblasts shown in yellow; sox10 reporter-expressing chondrocytes shown in magenta. Arrow indicates the location of the hypural diastema separating the dorsal from ventral lobes in A; asterisk indicates the absence of the diastema in B in both the ossified rays and the cartilaginous hypural complex (*). Bar, 500 uM. A’ and B’ show higher magnification images of boxed areas. (C) In these early stages of fin development (~6.0 SL), the length of the longest rays is less in shh-pulsed larvae than in control siblings. Significance determined using Welch’s two-sample T-tests. (D) control (E) and shh pulsed sibling caudal fin growth from 14 to 36 dpf. Dashed lines indicate the distal edge and overall shape of the fins. Scale bars, 500 µm. (F) The emergence of WT forked fin shape (gray lines) is the result of a lower growth rate in central rays (solid lines) relative to peripheral rays (dashed lines). Following embryonic shh pulse (green lines), central rays exhibit increased growth rates throughout development while peripheral rays retain a WT growth trajectory. (G-H) Dual Fucci reporter showing non-proliferating cells in red and cells in G2, S or M phase in cyan in the dorsal lobe of caudal fin folds of (G) control siblings and (H) larvae that experienced shh pulse. Bar, 200µm. (I) In control developing forked fins, proliferation is relatively lower in central regions, while shh pulse causes increased proliferation in central regions of the developing truncate fin Significance is determined by ANOVA followed by Tukey’s post hoc test. (J) Across the entire organ, proliferation becomes more uniform following shh pulse (closer to 1.0) compared to WT. Significance determined by Welch two-sample T-test. Difference between central / peripheral proliferation comparing WT to shh pulse is still significant when outlier in shh pulse group is removed.

We asked whether the truncate phenotype induced by shha pulse involved a change in the growth rate between central and peripheral rays. In control caudal fins, central rays grow slower than peripheral rays, causing the forked shape to become progressively pronounced as the fish grow (20) (Fig. 3D, Fig. S1D). In shha-pulsed caudal fins, peripheral rays grow at indistinguishable rates from control siblings. Strikingly however, shha-pulsed individuals showed 35% faster central ray growth throughout juvenile development compared to their control siblings (Fig. 3E–F).

We tested whether these changes in linear growth rates of rays following a shha pulse would correspond to altered rates of regional cell proliferation. We used the Dual z-Fucci cell-proliferation reporter transgenic line (55) to quantify proliferating cells in peripheral and central regions of control fins starting as early as ~5.9mm SL, when peripheral and central regions show similar cell population sizes (see Fig. S4A). Differences in regional cell proliferation (previously described in (24,59)) emerged during this developmental stage, with central regions of fins proliferating slower than peripheral regions (Fig. 3G, I). As the fish grows, increased peripheral cell proliferation rates lead to similarly larger cell population sizes in peripheral regions compared to center (Fig. S4B). The shha pulse led to increased proliferation specifically in the central region of fins (Fig. 3H–J) leading to larger central cell populations compared to control siblings (Fig. S4A–B). Peripheral proliferation levels did not change relative to controls (Fig. 3I). These results suggest that an embryonic shha pulse changes proliferation fates, evident at later larval stages during skeletogenesis, subsequently altering fin ray growth rates, changing the ultimate shape of the caudal fin.

These results led us to ask whether rays growing in different regions of the fin showed different levels of Shh pathway activation. Ptch2 is a receptor of Shh and also serves as a readout of canonical Shh activity (60), and expression of ptch2 was elevated 12–24 hours after shha pulse treatment (Fig. S5A), before rays developed. Both shha and ptch2 are concentrated at the distal tips of growing rays (28,44), with Ptch2 fluorescence (60) enriched in peripheral rays compared to central rays (Fig. S5B, D) (29). In contrast, shha-pulsed individuals showed comparatively elevated ptch2 reporter activity in the central regions (Fig. S5C–D).

Embryonic shha overexpression alters tissue memory and acts independently from mechanisms that regulate fin length and ray pattern

In control fins, the forked shape, length, and skeletal patterning of the caudal fin are rebuilt during regeneration (Fig. 4A, C). We asked if the shha pulse permanently altered memory of fin shape: indeed, amputated truncate fins restored their original truncate shape (Fig. 4B–C). The embryonic shha treatment induced not just a developmental change in positional information, but also a shift in positional memory consistent through adulthood.

Fig. 4: Altered memory of the adult caudal fin.

(A-B) Control forked caudal fins (A) restore a forked shape 30 days after amputation. (B) Truncate fins restore a truncate shape after amputation. Bar, 1 mm. (C) Quantification showing the fin shape of each individual before and after regeneration. Significance between factors determined via a linear mixed-effects model.

Previous research characterized longfin and shortfin mutants, which show marked differences in fin size relative to SL (34,35,57); hypothyroid fish show a proximalized fin ray patterning with delayed branching (38). In these altered phenotypic contexts, the overall caudal fin shape remains forked (see Fig. S6A, C, E). We asked if introducing an embryonic shha pulse in these altered phenotypic contexts could induce a truncate phenotype. Indeed, lengthened (lof), shortened (sof), or proximalized (conditionally hypothyroid) backgrounds each produced truncate fins following the embryonic shha pulse (Fig. S6B, D, F), suggesting that length, patterning, and shape are each regulated by independent signaling pathways that may be decoupled to produce a vast range of phenotypic fin diversity.

Discussion

We have demonstrated that the caudal fin shape of zebrafish is imprinted in the embryonic fin fold tissue, and that it may be re-patterned by excess shh. Shh is produced by the zone of polarizing activity (ZPA) in the posterior regions of growing limb buds, and confers posterior identity in tetrapod appendages (43). In tetrapods, loss of function to the Shh pathway prevents formation of distal limb structures (62,63), while ectopic Shh expression at the anterior border of the limb bud induces mirror-image digit duplications with posterior identities (42,43,64), with greater amounts or longer duration of Shh activation inducing supernumerary posterior digits (65,66). The posteriorizing role of Shh predates the fin-to-limb transition, a Shh-producing ZPA is both present in the pectoral and pelvic (paired) fins as well as dorsal and anal (median) fins of chondrichthyans (67,68) and bony fishes (69–71). Impairment of the Shh pathway blocks paired fin formation entirely (69,72). However, in contrast to the other median fins, no tissue with ZPA-like activity has been identified in the early caudal fin fold (44,55). Here, we have shown that an embryonic shha pulse in the caudal fin fold–5–6 days before shh is normally expressed in these tissues (45,56)–is sufficient to re-pattern shape of the organ by specifically elongating the central rays, without an obvious posteriorization effect. In contrast to phenotypes induced by excess Shh in paired appendages, the shh pulse in zebrafish decreases caudal fin ray number and appears to “peripheralize” the rays, with central rays now growing to lengths that match those of the peripheral rays.

Neither shha nor shhb are expressed in the posterior tail mesenchyme during zebrafish embryogenesis, and early caudal fin morphogenesis does not appear to require activation of the Shh pathway (45,56). Blocking the Shh pathway specifically during juvenile development inhibits growth of the peripheral-most non-branching rays (29), but the pathway does not appear to be required for early patterning. Indeed, although loss of fin shh expression blocks formation of the paired fins (above), the caudal fin (as well as the anal fin) appears to be unaffected (69). Likewise, we found that we found that early treatment with a Smo inhibitor did not affect development of the forked caudal fin shape (see Fig. 1I). As previously suggested for the anal fin (70), it appears that the caudal fin utilizes regulatory signals that do not require the Shh pathway. Nonetheless, these tissues remain sensitive to premature expression of shha, which we show is sufficient to repattern the growing organ. Indeed, although shh is not expressed in the early fin fold primordium, downstream Shh effectors (including gli3, smo and ptch2) are expressed in these tissues (45), and these canonical effectors likely allow precocious shha to activate the downstream signaling cascade and reshape the fin.

In paired limbs and fins, Shh signaling functions in concert with 5’ Hox factors to establish anteroposterior patterning (73,74). In tetrapod limbs, continuous Shh expression is required for maintenance and later expression of HoxA and D cluster genes (73), and Shh inhibits the repressor form of Gli3, activating 5’ HoxD expression (41,75). Recent zebrafish mutant analyses demonstrate that while hoxA and D cluster mutations impair pectoral, pelvic, dorsal and anal fin development, the caudal fin remains unaffected (76). Instead, the caudal fin is regulated by HoxB and C cluster genes, specifically hoxb13a and hoxc13a, which are expressed in a region-specific manner at the posterior tail at 2–4 dpf, before the emergence of any skeletal elements (14). While double mutants for both hoxb13a and hoxc13a fail entirely to form a caudal fin, single knock out of either hoxb13a or hoxc13a reduces fin ray numbers, abolishes the hypural diastema, and alters fin shape by shortening the peripheral rays (14). The notable phenotypic similarities between single hoxb13a or hoxc13a loss-of-function phenotypes and the truncate fin phenotype presented here suggests that the shha pulse may act by disrupting hox13 gene expression or regulation to repattern fin shape during development.

Desvignes et al. (13) proposed that a central organizing center establishes the hypural diastema and defines the axis of symmetry for the growing fin. According to this model, the organizing center splits the ventroposterior mesenchyme in the central endoskeleton into two plates of connective tissue, forming the diastema between the central hypurals (hypurals 2 and 3). This is followed by progressive, paired emergence of fin rays from central to peripheral; the diastema is theorized to inhibit the growth potential of the earliest developing rays at the center of the organ (13). The early pulse of shha may disrupt formation or activity of the diastema organizing center; in the absence of a central organizing signal, elongating rays may default to peripheral identity and growth. We found that shha pulse applied at 4 dpf or later was ineffective at producing a truncate fin (see Fig. 2A) even though these later heat shocks were capable of inducing transgene expression (see Fig. S3B). This suggests that the positional information that establishes relative ray length and caudal fin shape is imprinted before 4 dpf, well before shh is normally expressed in these tissues (45,56). Alternatively, we note it is possible that decreasing efficiency of the heat shock promoter with age may not produce sufficient shha to induce the positional shift at later stages.

Our data suggest that shha is capable—directly or indirectly—of inducing peripheral characteristics in centrally located rays, in terms of their skeletal growth and increased underlying proliferation (see Fig 3). In developing forked fins, peripheral regions proliferate at a higher rate than central regions (24), likely supporting the accelerated growth of these rays (see Fig. 3G, I). The embryonic shha pulse induced increased proliferation rates in the now rapidly-growing central region of the emerging truncate fin (see Fig. 3H–J). The Shh pathway can directly regulate proliferation rates: in parallel to its role patterning posterior limbs, Shh signaling promotes proliferation by regulating cell-cycle G1–S progression in the distal limb mesenchyme (77–79). The mesenchymal cells proliferate at higher rates in the outer regions of the distal limb bud relative to the center as the organ grows during embryonic development (80). Reduction of Shh signaling after early patterning decreases limb mesenchymal cell proliferation, leading to time-dependent progressive digit loss (78). Shh signaling affects growth by upregulating cyclin/kinase pair Ccnd1 and Cdk6 transcription through the inhibition of the repressor form of Gli3 (79). This mechanism is evolutionary conserved, as gli3 mutants in medaka increase ccnd1 and cdk6 transcription, which have conserved Gli3-regulated promoters across vertebrate genomes (70). Our data presented here demonstrates that relative modulation of regional proliferation rates is associated with different adult fin shapes, providing evidence that intrinsic positional characteristics inform local growth rates across the growing fin to progressively sculpt the shape of the organ.

A pulse of shha in the embryonic fin can permanently alter the memory of fin shape, since the adult truncate fin phenotype was restored following amputation without additional exogenous Shh (Fig. 4). While modulating bioelectricity or thyroid hormone availability can shift the morphology of a regenerating fin, neither treatment is capable of altering tissue memory: fins revert to a WT morphology if the treatment (calcineurin inhibition or thyroid inhibition) is removed or rescued (30,38). Memory of fin size can be altered by inhibiting overall proliferation: temporarily inactivating an accessory subunit of the DNA polymerase alpha (pola2) during regeneration can permanently alter the memory of fin size (40). Notably, the forked shape is not altered by this treatment (40). Our work therefore identifies an early developmental window of caudal fin development during which positional information is imprinted that will inform both the development and the memory of organ shape.

The evolution of the externally-symmetrical homocercal caudal fin in teleosts allowed for the external skeleton to take on distinct dorsoventral functionalization (19,81–83). This morphological functionalization may have supported the diversification of caudal fin shapes across teleosts. The wide spectrum of fin shape diversity can be categorized into truncate shapes with a flat or rounded edge and forked shapes with a concave edge; truncate and forked fins each offer biomechanical advantages and tradeoffs as propulsive and stabilizing organs (5,6,19,84). Here, we have identified an experimental manipulation capable of changing the zebrafish caudal fin from a forked to a truncate shape (see Fig. 1, 3F, S1C). In addition to the noticeably altered external shape, induced zebrafish truncate fins consistently lacked a hypural diastema (see Fig. 1D, G). The hypural diastema is considered a teleostean novelty (15) (although it was independently acquired in gars (19)). The diastema has been convergently lost at least once in nearly every lineage of crown teleosts, including cusk, swamp and true eels and in derived groups of bony tongues, catfishes, cods, flatfishes and killifishes (15,19,85–87). Notably, many of the clades that have lost the hypural diastema also show a truncate or rounded caudal fin shape, however the evolutionary relationship between hypural diastema and the shape of the fin has not been explored in detail. Our work identifies early activation of the Shh pathway or its downstream targets as capable of reshaping the external fin and abolishing the diastema, without altering length or proximodistal patterning of the organ (see Fig S6), suggesting developmental mechanisms that may underlie natural teleost fin diversity.

Methods

Resource Availability

Lead Contact:

Further information and requests for resources or reagents, including fish lines, should be directed to and will be fulfilled by the lead contact, Dr. Sarah McMenamin (mcmenams@bc.edu).

Materials Availability:

This study did not generate any new reagents or animal strains

Data and Code Availability

• All data reported in this study will be shared by the lead contact upon request

• This paper does not report original code

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Experimental Model and Subject Details

Experimental Animals

Zebrafish were reared under standard conditions at 28°C with a 14:10 light:dark cycle. Fish were fed marine rotifers, Artemia, Adult Zebrafish Diet (Zeigler, Gardners PA, USA) and Gemma Micro (Skretting, Stavanger, NOR). Individuals that experienced a shha pulse during embryogenesis (below) experienced a slight growth delay during larval development, although caught up in size by early juvenile stages (see Fig. S1). Because of the early growth delay, we took care to size-match treated and control individuals and standard length (SL) are reported throughout. Note that prior to development of the hypural complex, notochord length was measured, and is referred to as SL per (18). For developmental serial imaging, siblings positive and negative for the hsp70l:shha-eGFP transgene were reared in individual containers so individuals could be identified. Fish line to induce sonic hedgehog a overexpression was Tg(hsp70l:shha-EGFP) (49). Other lines used were Tg(sp7:GFP)b1212 (88) to visualize osteoblasts, Tg(p7.2sox10:mRFP) (89) for chondrocytes, TgBAC(ptch2:Kaede) (60), and Dual z-Fucci (58,90). Mutants used were longfin^dt2/kcnh2a (35,91), and shortfin^dj7e2/cnx43 (34).

Method Details

Imaging:

Zebrafish were anesthetized with tricaine (MS-222, ∼0.02% w/v in system water). Anesthetized or cleared and stained (92) individuals were imaged on an Olympus SZX16 stereoscope using an Olympus DP74 camera, an Olympus IX83 inverted microscope using a Hamamatsu ORCA Flash 4.0 camera, a Leica Thunder Imager Model Organism using a sCMOS monochrome camera, or a Zeiss AxioImager Z2 using a Hamamatsu Flash4.0 V3 sCMOS camera. Identical exposure times and settings were used to compare experimental treatments and capture repeated images of fins. Images were correspondingly adjusted for contrast, brightness and color balance using FIJI (93), and compiled using BioRender.

Sonic hedgehog overexpression:

Tg(hsp70l:shha-EGFP) crosses at 48–54 hours post fertilization (hpf) were treated with 37° heat shock for 15 minutes. 16–18 hours after treatment, individuals were screened for GFP expression as in (49). Sibling larvae that were treated with heat shock but were negative for GFP were kept as negative WT controls.

Localized shha overexpression:

Localized induction of the heat shock promoter was performed as previously described (57). Local heat shocks were induced at 48–54hpf for 15 minutes; local GFP expression was confirmed ~16–18 hours after the treatment.

Amputations:

Adult caudal fin regeneration experiments were performed on adult zebrafish 27–33 mm SL. Caudal fins were amputated from anesthetized fish under a stereoscope at the 5th ray segment using a razor blade and given 30 days to regenerate.

Drug Treatments:

To rescue the shha overexpression phenotype by Shh pathway inhibition, larvae were treated either with the Smoothened inhibitor BMS-833923 (10mM stock in 100% DMSO, 0.5 µM working solution in fish water) or the vehicle control DMSO (0.5 µM in fish water). A clutch of Tg(hsp70l:shha-eGFP) was treated with heat shock and sorted as above, and transgenic (GFP+) and non-transgenic (GFP-) siblings were treated with either BMS or the vehicle control for 4 hours starting 16–18 hours after the heat shock. The treatment with BMS or vehicle was repeated a second time 24 hours after the first treatment. After washout, fish were reared to adulthood under standard conditions.

To induce hypothyroidism in regenerating fish, fins were amputated as above, and allowed to regenerate in 1.0mM MPI cocktail (1.0mM MMI + 0.1mM KClO₄ + 0.01mM iopanoic acid, diluted in fish water) (38,53), for 21 days with drug changes every 1–2 days.

RT-qPCR:

Larvae were reared and heat shocked as stated above. At 3dpf, larvae were sorted into positive and negative cohorts based on fluorescence, placed into Thermo Fisher’s RNAlater^™ Stabilization Solution (Cat. #: AM7021), and subsequent collection from both groups proceeded until 4dpf. RNA was extracted from these samples using Zymo Research Quick-RNA^™ Microprep Kit (Cat. #: R1050) and cDNA libraries synthetized using Thermo Fisher SuperScript^™ IV Reverse Transcriptase (Cat. #: 18090010). Thermo Fisher PowerUp^™ SYBR^™ Green Master Mix was used for qPCR (Cat. #: A25741); three technical replicates and three biological replicates were run on Thermo Fisher QuantStudio^™ 3 Real-Time PCR System (Cat. #: A28567). Results were analyzed using ThermoFisher Connect Platform. The primer sets used can be found in Table S1.

Proliferation quantification:

Proliferation was measured in different regions of the growing fins using the Dual z-Fucci transgenic line (58). Four rays were measured from each fin: the second dorsal and second ventral rays (the peripheral rays) and the two center-most rays of each lobe (the central rays). The proliferation rate was measured along a line drawn through each ray and was calculated as the number of cyan-expressing cells divided by the total number of fluorescent cells (cyan plus red). Regional proliferation was calculated as the average proliferation rate of the two peripheral rays and the average proliferation of the two central rays.

Statistical analysis:

Analyses were performed in RStudio. Data were analyzed with either Welch two-sample t-test, ANOVA followed by Tukey’s honest significant differences (using 95% family-wise confidence level), Fligner-Killeen test, or a linear mixed-effects model. In graphs showing pairwise comparisons, significance is indicated as follows: * p < 0.05, ** p < 0.01, *** p < 0.001.