Disruption of the creb3l1 gene causes defects in caudal fin regeneration and patterning in zebrafish Danio rerio

Abstract

Background

The gene cAMP‐Responsive Element Binding protein 3‐like‐1 (CREB3L1) has been implicated in bone development in mice, with CREB3L1 knock‐out mice exhibiting fragile bones, and in humans, with CREB3L1 mutations linked to osteogenesis imperfecta. However, the mechanism through which Creb3l1 regulates bone development is not fully understood.

Results

To probe the role of Creb3l1 in organismal physiology, we used CRISPR‐Cas9 genome editing to generate a Danio rerio (zebrafish) model of Creb3l1 deficiency. In contrast to mammalian phenotypes, the Creb3l1 deficient fish do not display abnormalities in osteogenesis, except for a decrease in the bifurcation pattern of caudal fin. Both, skeletal morphology and overall bone density appear normal in the mutant fish. However, the regeneration of caudal fin postamputation is significantly affected, with decreased overall regenerate and mineralized bone area. Moreover, the mutant fish exhibit a severe patterning defect during regeneration, with a significant decrease in bifurcation complexity of the fin rays and distalization of the bifurcation sites. Analysis of genes implicated in bone development showed aberrant patterning of shha and ptch2 in Creb3l1 deficient fish, linking Creb3l1 with Sonic Hedgehog signaling during fin regeneration.

Conclusions

Our results uncover a novel role for Creb3l1 in regulating tissue growth and patterning during regeneration.

Key Findings

• novel patterning factor

• creb3l1 bone regeneration

• proximodistal axis patterning

• creb3l1 regulator of proximodistal patterning

• creb3l1 modulates sonic hedgehog signalling

1. INTRODUCTION

Although bone and the skeleton may seem like relatively static tissues that support our bodies with their rigid mineralized structures, their growth and development during embryogenesis and during regeneration require dynamic expression of many interacting proteins. Together, these proteins facilitate a complex program of processes to regulate fate determination, patterning, and bone matrix deposition (reviewed in References 1, 2, 3). An analogous program of events can be temporarily reactivated during bone repair and regeneration. Numerous factors involved in different aspects of bone development have been identified and interrogated. One such factor is the cAMP Responsive Element‐Binding Protein 3 Like 1 (CREB3L1), also known as Old Astrocyte Specific Induced Substance (OASIS). ⁴

Creb3l1 is a member of the ATF/CREB family of transmembrane transcription factors, which have a unique topology: an N‐terminal cytoplasmic transcription activating (TA) region containing the DNA‐binding basic leucine zipper motif (bZIP) domain, a single transmembrane (TM) domain that tethers CREB3L1 in the membrane of the endoplasmic reticulum (ER), and a small C‐terminal luminal domain (Figure 1A). This topology allows tight regulation of CREB3L1 activity, as it is transcriptionally inert while in the ER, and must be trafficked to the Golgi where it is cleaved by S1P and S2P proteases to release the TA fragment. ⁵ , ⁶ The ER to Golgi transport is initiated in response to ER stress or to a developmental or a physiological stimulus. After cleavage, the released TA fragment is able to enter the nucleus to regulate transcription of target genes, including genes related to bone development and secretory pathway expansion. ⁵ , ⁶ , ⁷ , ⁸ , ⁹ The best‐understood function of CREB3L1 is its transcriptional regulation of COL1A1, a gene that encodes type 1 collagen, the major structural protein of bones. CREB3L1 regulates COL1A1 gene expression by binding to an unfolded protein response element upstream of the COL1A1 gene. ¹⁰ In addition, CREB3L1 has been shown to respond to a variety of stimuli in different cell types by increasing the transcription of components that regulate the efficiency of the secretory pathway. ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ This upregulation of secretory efficacy is especially important in the secretion of the bulky, rigid collagen cargo that makes up bones. ⁶ , ⁸ , ¹⁰ , ¹⁵ , ¹⁶ , ¹⁷ Thus, Creb3l1 involvement in both collagen production and secretion makes CREB3L1 an essential factor in bone development. ⁷ , ¹⁰ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰

FIGURE 1.

Creb3l1 is a transmembrane transcription factor highly conserved between humans and zebrafish. (A) A schematic of Creb3l1 protein showing the N‐terminal transcription activating (TA) region (amino acids 1–374) containing the DNA‐binding bZIP domain (amino acids 291–356), followed by a single transmembrane (TM) domain (amino acids 375–395) and the C‐terminal luminal domain (amino acids 427–519). The sites cleaved by the S1P and S2P proteases are indicated by arrows. Creb3l1 resides in the endoplasmic reticulum and in response to stress is transported to the Golgi (step 1), where it is cleaved by S1P and S2P proteases (step 2). The released TA fragment then enters the nucleus to regulate transcription of target genes (step 3). (B) Alignment of human and zebrafish Creb3l1. The TA domain is indicated by a blue line and the TM domain is marked by an orange line. The DNA‐binding bZIP domain is outlined in red. The sites cleaved by S1P and S2P are boxed in green. The highest level of amino acid conservation is in bZIP, the TM and regions immediately upstream from bZIP. The S1P and S2P cleavage sites also reside within conserved domains. Identical residues are shaded in blue. (C) The genetic loci of human and zebrafish Creb3l1 demonstrate high synteny, sharing 8 genes within the same ~1 megabase span in their respective genomes.

The critical importance of CREB3L1 in bone formation is underscored by human patients with CREB3L1 mutations that exhibit severe osteogenesis imperfecta. ¹² , ¹³ , ¹⁸ Such patients exhibit diminished expression of COL1A1, as well as reduced expression of components of the secretory pathway, including proteins involved in anterograde vesicle transport. ²⁰ The critical role of CREB3L1 in bone development is also evident in Creb3l1 knock‐out (KO) mice, which are severely osteopenic. ¹⁰ The KO mice exhibit lower bone mineral density, spontaneous ankle fractures, and a lower rate of bone formation/bone surface. Importantly, Creb3l1 is not expressed in osteoclasts, ¹⁰ indicating that these defects are a result of poor bone deposition by osteoblasts, and not enhanced bone absorption by osteoclasts. In support, Creb3l1 ^−/− osteoblasts have a lower expression of Col1a1 and overall poorer mineralization capability. ¹⁰

In addition to its essential role in bone formation during development, Creb3l1 is required for bone regeneration. In a mouse femur fracture model, Creb3l1 KO mice show defects in fracture healing, most likely due to reduced synthesis of collagen caused by a deficit in Col1a1 transcription. ¹¹ , ¹⁷ , ²¹ However, the mammalian model of bone regeneration has severe limitations, as mammalian bone has limited regenerative capability. In contrast, Danio rerio zebrafish exhibit high capacity for bone regeneration, and this model has been widely used to discover and clarify key pathways involved in bone development and regeneration. A major benefit of the zebrafish is that they are able to fully regenerate numerous tissues, including complex structures containing bony tissue such as the caudal fin. ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ Bone regeneration is often studied after a partial amputation of the caudal fin, which within 21 days regenerates a tail of exact pre‐amputation size and with pre‐amputation patterning and tissue organization. ²³ , ³⁰ , ³¹ , ³² Further, the thin and translucent tail allows direct visualization of distinct cell types during successive stages of regeneration using cell type‐specific fluorescent‐labeled transgenic fish models or staining of fixed or live tissue. ²² , ²⁸ , ³³ , ³⁴

Caudal fin regenerative outgrowth starts with the formation of a blastema, a mass of undifferentiated, proliferative cells, over each regenerating fin ray. The progenitor cells in the blastema differentiate to reform all the distinct tissues within the fin. Multiple pathways participate in regeneration, with FGF signaling required for blastema formation, Wnt signaling facilitating proliferation of progenitor cells, and Sonic Hedgehog (Shh) signaling playing a major role in controlling differentiation into bony tissues, as well as pattern establishment required for ray bifurcation. ²² , ²⁵ , ³⁵ , ³⁶ , ³⁷ , ³⁸ One of the key events required to produce or reconstitute the normal architecture of the tail during development and regeneration is the establishment of the proximo‐distal axis. The restoration of the axis is necessary to inform the characteristic bifurcation fin patterns that form when a single ray proximal to the amputation plane splits into two smaller rays. ²⁵ , ³⁴ , ³⁷ , ³⁹ , ⁴⁰ The position of the bifurcation point is tightly regulated, and bifurcation is initiated within each ray at distinct lengths from the amputation plane, depending on the position of the ray within the tail. The proximo‐distal placement of the bifurcations in each ray (except the most lateral ones) is critical for overall fin architecture. Defects in the reformation of the proximo‐distal axis are easily scored by evaluating the bifurcation parameters of the regenerating fin. ²⁵ , ³⁴ , ³⁷ , ³⁹ , ⁴⁰ Thus, the zebrafish caudal fin regeneration model allows a minimally invasive analysis of the role of Creb3l1 in caudal fin regeneration and patterning of the regenerating tissues, and may provide insights to its intersections with signaling pathways controlling regeneration and the establishment of the proximo‐distal axis.

To probe the role of Creb3l1 in bone development and regeneration, we generated a zebrafish mutant that expresses Creb3l1 lacking the majority of its coding sequence, including the DNA‐binding bZIP domain (creb3l1 ^{ΔbZIP/ΔbZIP} ). The mutant Creb3l1 is unable to activate its target genes and the fish are transcriptionally null for creb3l1. Analyses of the mutant fish indicate that their skeletal development is largely normal, in agreement with the phenotype of the Creb3l1 KO mice. ¹⁰ However, while mammals lacking CREB3L1 or expressing mutant CREB3L1 exhibit low bone density and predisposition to bone fracture, creb3l1 ^{ΔbZIP/ΔbZIP} zebrafish show normal bone density and lack spontaneous fractures. Yet, creb3l1 ^{ΔbZIP/ΔbZIP} mutant fish exhibit decreased bifurcation complexity in the caudal fin, suggesting a developmental patterning defect. Moreover, we document that Creb3l1 activity is required for caudal fin regeneration, with creb3l1 ^{ΔbZIP/ΔbZIP} zebrafish displaying a significantly decreased overall re‐growth of the fin. This phenotype is linked to perturbations in cell proliferation due to the disruption in the cell cycle in creb3l1 ^{ΔbZIP/ΔbZIP} regenerating caudal fins. Moreover, the creb3l1 ^{ΔbZIP/ΔbZIP} zebrafish show strong defects in the patterning of the regenerated tissue. Specifically, creb3l1 ^{ΔbZIP/ΔbZIP} animals have significantly reduced number of bifurcations, as well as an increased distalization of the bifurcation point. Mechanistically, this may be caused by a perturbation in the Shh signaling pathway, which we show by in situ hybridization to be spatially affected in the creb3l1 ^{ΔbZIP/ΔbZIP} regenerating fin. Together, our results demonstrate a critical requirement for Creb3l1 function during regenerative bone growth, and identify Creb3l1 as a novel factor regulating the proximo‐distal patterning of the caudal fin during regeneration.

2. RESULTS

2.1. Creb3l1 is a transmembrane transcription factor highly conserved between humans and zebrafish

Human and zebrafish Creb3l1 exhibit 58.9% overall amino acid sequence identity (Figure 1B). The highest sequence conservation is observed in the DNA‐binding bZIP domain, with ~98% sequence identity (indicated in red rectangle). The TM domain (indicated with an overlying orange line) is similarly conserved, with ~90% identity between the human and the fish sequences. The cleavage sites for the S1P and S2P proteases (indicated with green rectangles) are also highly conserved and are present within conserved motifs bracketing the cleavage sites. The C‐terminal lumenal domain of human Creb3l1 shares only 28% identity with its zebrafish homolog. The high level of conservation within the DNA‐binding domain and the TM region suggests a similar mechanism of Creb3l1 transcriptional action within the nucleus, as well as similar regulation of Creb3l1 activity through intramembranous cleavage in humans and in zebrafish.

Beyond the sequence similarity, human CREB3L1 and zebrafish creb3l1 also share high genomic synteny in their respective loci, with an overlap of 8 genes within their respective neighboring megabases containing 18 genes in human and 22 genes in zebrafish (adapted from Ensembl data ⁴¹ ) (Figure 1C). The genomic synteny suggests that this group of genes, including human CREB3L1 and zebrafish creb3l1, may form a gene cluster that undergoes simultaneous transcriptional regulation. ⁴² , ⁴³

2.2. Creb3l1 mutant fish show overall normal skeletal development

The Creb3l1 loss‐of‐function allele (creb3l1 ^{ΔbZIP/ΔbZIP} ) was generated by CRISPR/Cas9 – mediated genome editing using a guide targeting exon 2 (Figure 2A). The creb3l1 ^{ΔbZIP/ΔbZIP} fish have a guide‐induced insertion of seven base pairs (boxed in green in Figure 2A) at 189 bp of the coding mRNA (green arrow). This leads to a mutation at amino acid 65 (change of D to P), followed by a frameshift resulting in the addition of 10 amino acids (in gray) not present in wild‐type Creb3l1 before the TGA stop codon (in red).

FIGURE 2.

creb3l1 mutant zebrafish are normal size and exhibit normal overall skeletal morphology and bone density. (A) Exon 2 of the zebrafish creb3l1 gene was targeted to generate creb3l1 ^{ΔbZIP/ΔbZIP} fish with a +7 bp insertion (in green) that causes a frameshift mutation at amino acid 65 (D to P change). This mutation generates a Creb3l1 fragment encoding only the N‐terminal 65 amino acids of Creb3l1, followed by 10 additional amino acids not present in wild‐type Creb3l1 (in gray). Yellow highlights wild‐type codons while gray highlights mutant codons and amino acids. TGA stop codon is in red. cDNA, consensus coding sequence and protein sequences are shown. (B) creb3l1 mRNA levels were measured by RT‐PCR in 7 dpf creb3l1 ^+/+ and creb3l1 ^{ΔbZIP/ΔbZIP} larvae. Expression of creb3l1 is significantly elevated in the creb3l1 ^{ΔbZIP/ΔbZIP} larvae. n = 3–4, ***p < .001. Each data point represents a pool of ~30 larvae. (C) Larvae from heterozygous creb3l1 ^ΔbZIP/+ incrosses were genotyped at 7 dpf and show the expected Mendelian ratios. n and p values are indicated. (D) The length of age‐matched 6 mpf creb3l1 ^+/+ and creb3l1 ^{ΔbZIP/ΔbZIP} fish was measured from the snout to the posterior‐most point of the tail. There was no significant difference in length between the wild‐type and the creb3l1 mutant fish. n = 27–34. Each data point represents an individual fish. (E) Age‐matched 3 mpf creb3l1 ^+/+ and creb3l1 ^{ΔbZIP/ΔbZIP} fish were analyzed by μCT and representative scans of male and female animals are shown. The creb3l1 ^{ΔbZIP/ΔbZIP} fish exhibit overall normal skeletal morphology. Scale bar = 5 mm. (F) Bone density (mgHA/cm³) of the entire skeleton of male and female creb3l1 ^+/+ and creb3l1 ^{ΔbZIP/ΔbZIP} fish was determined from the μCT scans. n = 9–11. No significant difference was detected. Each data point represents an individual fish.

To test for missense mediated decay of the mutant mRNA, we performed RT‐PCR of creb3l1 mRNA in wild‐type and mutant larvae (primers are listed in Table 1). As shown in Figure 2B, transcription of creb3l1 was upregulated ~1.5‐fold in the creb3l1 ^{ΔbZIP/ΔbZIP} fish, suggesting that the mutant mRNA is not degraded. The mutant mRNA encodes a 75 amino acid peptide that includes only the N‐terminal 64 amino acids of Creb3l1. Irrespective of whether this peptide is translated and stable in the creb3l1 ^{ΔbZIP/ΔbZIP} fish, it lacks the majority of the Creb3l1 sequence including the DNA‐binding bZIP domain, making it inactive as a transcription factor. The upregulation of creb3l1 transcription may reflect a possible mechanism to compensate for the lack of functional Creb3l1 in the mutant fish.

TABLE 1.

Details of primers utilized to perform RT‐PCR, generate in situ probes, or clone from zebrafish cDNA.

Gene	Primers	Use
creb3l1	F: ACAATCCCCCTCCATACC	RT‐PCR
	R: GTTTCCGGGCAGTTCTCT
elfa (housekeeping gene)	F: TACCCTCCTCTTGGTCGC	RT‐PCR
	R: TTGGAACGGTGTGATTGAG
col1a1a	F: CGATGGCTTCCAGTTCGAGT	Generation of in situ probe
	R: TAATACGACTCACTATAGGGCCAGGGGGATTTTACACGCT
creb3l1	F: GATGCGGCCGCCTGTGTTTTGTGTTGGT	Generation of plasmid for in situ probe template
	R: GATCGTCGACGAAATATTCAGCTCCTCC

The offspring of heterozygous creb3l1 ^+/ΔbZIP incrosses were genotyped at 7 days post fertilization (dpf) and show that zebrafish larvae are recovered at relatively normal Mendelian ratios (Figure 2C). This suggests that the lack of functional Creb3l1 is not lethal to larval survival. Moreover, mature fish were obtained from the incrosses (see below), consistent with results from the Creb3l1 KO mice, which also survive to adulthood. ¹⁰

The creb3l1 ^{ΔbZIP/ΔbZIP} fish grow to a size analogous to that of wild‐type creb3l1 ^+/+ fish (Figure 2D), and show no obvious skeletal malformations (Figure 2E; larger images of the micro computed tomography (μCT) scans are included in Supplemental Figure 1). Skeletal development appeared normal in both female and male creb3l1 ^{ΔbZIP/ΔbZIP} fish, indicating lack of sexual dimorphism in the response to creb3l1 mutation.

CREB3L1 has been shown to be required for normal bone density in mammals, as Creb3l1 KO mice have porous fragile bones, a phenotype also observed in human patients with CREB3L1 mutations. ¹² , ¹³ , ¹⁸ , ⁴⁴ , ⁴⁵ , ⁴⁶ In contrast, comparisons of μCT scans of creb3l1 ^+/+ and creb3l1 ^{ΔbZIP/ΔbZIP} fish showed no significant difference in bone density as measured by the amount of hydroxyapatite (HA, the mineral component of zebrafish bone) in milligrams (mg) per cubic centimeter (cm³) (Figure 2F). Thus, the lack of Creb3l1 function is not deleterious to the growth of zebrafish, the development of their skeletons, or the density of their bones.

2.3. Creb3l1 mutant fish show decreased fin ray bifurcations

To provide a baseline for subsequent caudal fin regeneration studies, we performed analyses of intact fins in wild‐type and creb3l1 mutant zebrafish. To visualize the behavior of osteoblasts, we crossed creb3l1 ^+/+ and creb3l1 ^{ΔbZIP/ΔbZIP} fish to a transgenic fish line expressing eGFP under the promoter for the osteoblast‐specific Sp7 transcription factor (Tg(Ola.Sp7:EGFP‐CAAX)pd51; ZIRC line pd51Tg, ²⁷ referred to in this text as Tg(Sp7:eGFP)). We observed robust expression of sp7::eGFP in intact caudal fins of creb3l1 ^{+/+,Tg(Sp7:eGFP)/+} and creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} fish that marks osteoblast distribution within individual rays and showcases the characteristic bifurcations of each ray (Figure 3A).

FIGURE 3.

creb3l1 mutant zebrafish exhibit perturbations in caudal fin patterning. (A) Tails of age‐matched 6 mpf creb3l1 ^{+/+,Tg(Sp7:eGFP)/+} and creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} fish were imaged by fluorescence and representative images are shown. Orange lines from the caudal peduncle to the tip of third rays mark distance measured as fin length. Blue lines mark the distance of the most proximal bifurcation point from the caudal peduncle of second rays. Scale bar = 1 mm. (B) Images analogous to those in (A) were used to measure the length of second and third ray on both sides of the fin, and the average length/fin is shown. Each data point represents an individual fish. No difference in tail length was detected. n = 15–17. (C) Images analogous to those in (A) were used to measure the distance between the primary proximal bifurcation and the caudal peduncle of the second and third ray on both sides of the fin. The average distance was calculated and is shown. Each data point represents an individual fish. No difference in the proximo‐distal position of the bifurcation point was detected. n = 16–17. (D) Higher magnification of images analogous to those in (A) (the images were “stretched” along the x horizontal axis to more clearly visualize the distinct bifurcations). Rays are numbered starting with the most lateral. Secondary bifurcations of the third ray are indicated by yellow arrows, of the fourth ray by sienna arrows, and of the fifth ray by orange arrows. Secondary bifurcations of the sixth and seventh rays are indicated by blue arrows in wild‐type fish and are absent in mutant fish. (E, F) Images analogous to those in (A) were used to measure the total number of bifurcations per fin (E), as well as the number of primary (F), and secondary bifurcations per fin (G). The complexity of bifurcations is lower in the creb3l1 mutant fins, with fewer secondary bifurcations per fin than in wild‐type creb3l1 fins.

Tail length was measured from the tip of the tail to where the ray met the base of the caudal peduncle (defined as the end of the body wall muscle where the caudal fin ray meets the body of the fish) in adult 6 months post fertilization (mpf) wild‐type and creb3l1 mutant zebrafish (Figure 3A, orange lines). Tail length was not significantly different in creb3l1 ^{+/+,Tg(Sp7:eGFP)/+} and creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} fish (Figure 3B). Similarly, the distance from the base of the caudal peduncle to the point of first bifurcation of second and third rays in both lobes of each fin (Figure 3A, blue lines) is not significantly altered in the creb3l1 ^{ΔbZIP/ΔbZIP} zebrafish (Figure 3C).

In contrast, the complexity of bifurcations is decreased in creb3l1 mutant zebrafish tails (Figure 3D). Quantification of this phenotype, as assayed by the number of total bifurcations per fin, shows a significant decrease in the creb3l1 mutant zebrafish (Figure 3E). This phenotype is the result of a decrease in secondary bifurcations, as the number of primary bifurcations remains the same in wild‐type and mutant fish (Figure 3F). Analysis of secondary bifurcations of each lepidotrichium shows that while in wild‐type zebrafish tails, the third, fourth and fifth rays show secondary bifurcation of both lepidotrichia (Figure 3D, two yellow, sienna and orange arrows), in the creb3l1 mutant zebrafish, the analogous rays show only a single secondary bifurcation on the lateral lepidotrichium of each ray (Figure 3D, one yellow, sienna and orange arrows). Similarly, the sixth and seventh rays in wild‐type fish tails show secondary bifurcation of the lateral lepidotrichia (Figure 3D, blue arrows), but in creb3l1 mutant fish tails, the same rays have no secondary bifurcations. Quantification of this phenotype shows a significant decrease in secondary bifurcation number in the creb3l1 mutant fins (Figure 3G).

2.4. Regeneration of caudal fins is compromised in creb3l1 mutant fish

Although there are no notable changes in skeletal size or bone density during creb3l1 ^{ΔbZIP/ΔbZIP} fish development, the observed defect in bifurcation patterning prompted us to test whether changes in the levels of functional Creb3l1 may affect acute bone regeneration. We used the caudal fin amputation model in which ~50% of the tail fin is removed, causing an injury that is repaired in wild‐type zebrafish within a few weeks. ²³ We first examined the expression of creb3l1 in wild‐type fish at different days post amputation to define the most likely time frame for the possible action of Creb3l1 during regeneration. Expression of creb3l1 has been shown to be upregulated in the regenerating blastema, ⁴⁰ suggesting a possible role in regeneration. As shown in Figure 4A, creb3l1 mRNA was weakly expressed at 1 dpa and was more easily detectable in distinct domains at 3 and 5 dpa, with a more diffuse expression at 7 dpa. The signal in the 3 dpa regenerate was in a single “cap”‐like structure over as yet unbranched fin rays or was present as two independent domains in rays that had already bifurcated (Figure 4A, blue arrowhead). In the 5 and 7 dpa regenerates, the creb3l1 signal was detected in each sister lepidotrichium of each regenerating ray, but appeared more diffuse. Additionally, at the later time points, creb3l1 mRNA signal was detected within the inter‐ray spaces populated by mesenchymal cells (Figure 4A, white arrow).

FIGURE 4.

creb3l1 mutant zebrafish show defects in the regeneration of the caudal fin. (A) Approximately 50% of the caudal fin was amputated from 6 mpf wild‐type fish, and creb3l1 expression was monitored by in situ hybridization at the indicated dpa. creb3l1 signal is predominantly detected at the distal tip of each regenerating ray and appears strongest at 3 dpa (blue arrow). Representative images are shown. n = 3. Amputation planes are indicated with dashed line. Scale bar = 1 mm. (B) Approximately 50% of the caudal fin was amputated from 6 mpf creb3l1 ^+/+ and creb3l1 ^{ΔbZIP/ΔbZIP} fish. Regenerates at 3 dpa were analyzed by in situ hybridization to assess the expression of col1a1a. In wild‐type fins, robust col1a1 signal is detected over each regenerating ray (blue arrows) and immediately below the amputation plane (white arrows). In creb3l1 mutant fins, col1a1 signal is less apparent over the regenerating rays (blue arrows), and is undetectable at the amputation plane (white arrows). n = 3. Scale bar = 1 mm. C‐F) Approximately 50% of the caudal fin was amputated from age‐matched 6 mpf creb3l1 ^{+/+,Tg(Sp7:eGFP)/+} and creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} fish, and regeneration was assessed by fluorescence imaging at the indicated dpa. Amputation planes are indicated with white dashed lines. Scale bar = 1 mm. (C) Representative images at each dpa. (D) To quantify regenerate length, lines were drawn manually from the amputation plane to the tip of the second and third lateral rays on both sides of the regenerate (blue lines), measured and averaged. Ray width was measured for all rays within the regenerate (width of representative rays is marked by orange lines) and averaged. The Sp7⁺‐labeled fluorescent area was determined by thresholding the fluorescence signal in each regenerate (representative image is shown in lower panel). (E) Images analogous to those in (C) were used to measure the lengths of the second and third most lateral rays on both sides of the fin as in (D). Each data point represents an individual fish. The creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP/+)} fish show significantly shorter regenerates at 3, 7, and 9 dpa. n = 13–23. *p ≤ .05; **p ≤ .01; ***p ≤ .001. (F) Images analogous to those in C were used to calculate the area occupied by Sp7⁺ labeled osteoblasts within the regenerate, corrected for the average ray width (total Sp7⁺ area/average ray width) as in (D). Each data point represents an individual fish. The creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP/+)} fish show significantly decreased Sp7⁺ area/ray at all timepoints. n = 13–23. **p ≤ .01; ***p ≤ .001; ****p ≤ .0001.

Because creb3l1 expression in the regenerating tail appeared most robust at 3 dpa, we assessed the expression of a Creb3l1 target gene, the zebrafish col1a1a (a homolog of mammalian COL1A1) by in situ hybridization at 3 dpa. In wild‐type zebrafish, col1a1a was expressed in a cap‐like structure over each regenerating ray (Figure 4B, blue arrows), as well as on the tips of the bony fin rays right below the amputation plane (Figure 4B, white arrows). In contrast, in creb3l1 mutant zebrafish, col1a1a signal appeared fainter over the regenerating rays (Figure 4B, blue arrows), and was almost absent at the amputation plane (Figure 4B, white arrows).

To probe the role of Creb3l1 in caudal fin regeneration, age‐matched 6 mpf creb3l1 ^{+/+,Tg(Sp7:eGFP)/+} and creb3l1 ^{ΔbZIP/ΔbZIP,(Sp7:eGFP)/+} fish were subjected to 50% caudal fin amputation and regeneration was examined at 3, 7, 9, and 14 dpa by fluorescence (Figure 4C). Images analogous to those in Figure 4C were used to determine the length of the regenerated rays at each time point after amputation by measuring the length of Sp7⁺ osteoblast‐containing structures within the second and third rays on each side of the tail manually in FIJI ⁴⁷ (representative 3 dpa image is shown in Figure 4D, upper panel, blue lines). As shown in Figure 4E, creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP/+)} fish exhibit a significant decrease in ray length of the regenerate at 3, 7, and 9 dpa.

Even more extensive differences between creb3l1 ^{+/+,Tg(Sp7:eGFP)/+} and creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} fish were observed when the area containing differentiated osteoblasts within the regenerate was measured. Fluorescent images analogous to those in Figure 4C were thresholded, and the area of the resulting binary image (representative 3 dpa image is shown Figure 4D, lower panel) was measured to define the total Sp7⁺ area per regenerate. This value was divided by the average ray width of all rays proximal to the amputation plane (width of two rays is indicated with orange lines in Figure 4D) to obtain Sp7⁺ area per ray. As shown in Figure 4F, creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} fish exhibit a significantly decreased Sp7⁺ osteoblast area within the regenerate at all timepoints after amputation.

To assess whether the regeneration defects observed in 6 mpf fish were age‐dependent, as well as determine whether the lack of functional Creb3l1 might affect bone mineralization during regeneration, we performed 50% tail amputation on 3, 6, and 8 mpf creb3l1 ^+/+ and creb3l1 ^{ΔbZIP/ΔbZIP} fish and at 9 dpa stained them with the calcein fluorophore. Regenerates were then examined via bright‐field imaging (Figure 5A, upper panels) to assess the size of the regenerate and by fluorescence (Figure 5A, lower panels) to measure the extend of new bone formation. The area of regenerates in bright‐field images was measured as the area enclosed by the amputation plane (Figure 5A, 3 mpf, upper panel, dashed white line) and the regenerate outline traced by hand (Figure 5A, 3 mpf, upper panel, red line). So measured area represents the total regenerated area (REGEN). To correct for any differences in the size of the fins prior to amputation, the REGEN area was normalized to the total stump width (STU) (Figure 5A, 3 mpf, upper panel, black line). As shown in Figure 5B, the REGEN/STU area of creb3l1 ^{ΔbZIP/ΔbZIP} 6 mpf fish is lower than that of wild‐type fish, consistent with the data shown above in Figure 4F. Moreover, a significant decrease in the size of the regenerate is also apparent in 3 mpf and 8 mpf creb3l1 ^{ΔbZIP/ΔbZIP} fish (Figure 5B). Thus, the regeneration defect appears not to be age‐dependent.

FIGURE 5.

creb3l1 mutant zebrafish show defects in the ossification of the regenerating caudal fin. (A) Approximately 50% of the caudal fin was amputated from 3 mpf, 6 mpf, and 8 mpf creb3l1 ^+/+ and creb3l1 ^{ΔbZIP/ΔbZIP} fish. At 9 dpa, the tail regenerates were visualized in bright‐field (upper panels) and after staining with calcein fluorophore to visualize mineralized tissue (green, lower panels). Representative images are shown. A representative bright‐field image shows manually traced regenerate outline (red line) and the stump width (black line). A representative fluorescent image shows manually traced width of rays (orange lines mark two rays) within the fin. Amputation planes are indicated with dashed line. Scale bar = 1 mm. (B) Bright‐field images analogous to those in (A) were used to calculate the regenerate area enclosed by the red and black lines (shown in A) and corrected for the stump width (REGEN/STU). Each data point represents an individual fish. The creb3l1 mutant fish show significantly smaller regenerates than the wild‐type fish. n = 6–19. **p ≤ .01; ***p ≤ .001. (C) Representative threshold of calcein‐stained regenerate used to analyze mineralized area. (D) Fluorescence images analogous to those in (A) were thresholded and used to calculate the area of mineralized bone corrected for the average width of the rays (RMA/RAY). Each data point represents an individual fish. The creb3l1 mutant fish show significantly less mineralization than wild‐type fish. n = 6–19. *p ≤ .05; ****p ≤ .0001.

To assess whether lack of Creb3l1 affects bone mineralization during regeneration, calcein‐stained regenerates analogous to those in Figure 5A were imaged, the fluorescent signal was thresholded and the area of the resulting binary image (representative image shown in Figure 5C) was then measured to define the real mineralization area (RMA). To correct for the size of the fins prior to amputation, total RMA was normalized to the average ray width (RAY) within the tail (representative rays are marked by orange lines in Figure 5A, 3 mpf, fluorescence panel). As shown in Figure 5D, mineralized area was not significantly different in 3 mpf creb3l1 ^{ΔbZIP/ΔbZIP} fish relative to creb3l1 ^+/+ fish. However, mineralized area was significantly lower in the lepidotrichia of 6 mpf and 8 mpf creb3l1 ^{ΔbZIP/ΔbZIP} fish relative to creb3l1 ^+/+ fish.

Taken together, these results indicate that the lack of functional Creb3l1 causes a decrease in the overall size of the regenerate in 3, 6, and 8 mpf fish, suggesting that this phenotype is not age‐dependent. In contrast, differences in mineralization of the regenerated bone are only apparent as creb3l1 ^{ΔbZIP/ΔbZIP} fish age, with no difference seen in 3 mpf animals, but significantly decreased mineralization in 6 and 8 mpf fish.

2.5. Decreased regeneration in creb3l1 mutant fish correlates with defects in cell cycle

The observed decrease in the area of the regenerate (visualized by both, bright‐field microscopy and the presence of sp7:eGFP‐containing osteoblasts) prompted us to examine whether increased apoptosis and/or decreased cell proliferation may account for the decrease in size. creb3l1 ^{+/+,(Sp7:eGFP)/+} and age‐matched 6 mpf creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} fish were subjected to 50% caudal fin amputation and at 3 dpa processed for immunofluorescence with antibodies against the cleaved form of caspase‐3, a marker of apoptotic cells, ⁴⁸ or against phosphohistone‐H3 (PH3), a marker of mitotic cells. ⁴⁹ The anti‐PH3 antibody used recognizes the Ser10 phosphosite on PH3. As shown in Figure 6A, and quantified in Figure 6B, no significant difference in the number of cells containing cleaved caspase was seen in the creb3l1 mutant fish relative to wild‐type fish. Similarly, the number of proliferative cells marked by PH3 in the creb3l1 mutant fish regenerates was not significantly different than in wild‐type fish regenerates (Figure 6C and quantified in Figure 6D). This finding seemed surprising since the decrease in the regenerate area in creb3l1 mutant fish suggested either increased apoptosis (that should be reflected by increased cleaved caspase staining) or decreased proliferation (that should be detected by decreased PH3 staining).

FIGURE 6.

creb3l1 mutant zebrafish show mitotic defects during regeneration. (A–G) Approximately 50% of the caudal fin was amputated from age‐matched 6 mpf creb3l1 ^{+/+,Tg(Sp7:eGFP)/+} and creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} fish. At 3 dpa, the regenerating tails were processed with antibodies to cleaved caspase‐3 or PH3, followed by secondary antibodies conjugated to Alexa‐594 (red) fluorophore, and then imaged by bright‐field (not shown) and fluorescence (representative images of regenerates stained for cleaved caspase (A) or PH3 (C)). Amputation planes are indicated with dashed line. Scale bar = 1 mm. (B) Bright‐field and fluorescent images analogous to those in (A) were used to measure total regenerate area (as in Figure 5A) and to count the number of cells positive for cleaved caspase‐3 within the entire regenerate. The results are presented as the number of stained cells per whole regenerate area. Each data point represents an individual fish. n = 6–10. No significant difference was seen between the wild‐type and creb3l1 mutant fish. (D) Bright‐field and fluorescent images analogous to those in C were used to measure total regenerate area (as in Figure 5A) and to count the number of cells positive for PH3 in the entire regenerate. The results are presented as the number of stained cells per whole regenerate area. Each data point represents an individual fish. n = 6–10. No significant difference was seen between the wild‐type and creb3l1 mutant fish. (E) Regenerates stained for PH3 were imaged at higher magnification to identify metaphase and anaphase cells (white arrows). Representative images are shown. Metaphase cells were defined as cells with an elongated, non‐circular nuclear signal (the metaphase plate). Anaphase cells were defined as cells with chromosomes appearing to be pulled to opposite poles of the cell. Scale bar = 20 μm. (F, G) Fluorescent images of regenerates stained for PH3 were used to count cells in metaphase (F) and anaphase (G) and are presented as a ratio to total PH3⁺ cells counted within the imaged regenerate. creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} fish have significantly fewer cells in metaphase and anaphase. n = 6–8. *p ≤ .05, **p ≤ .01. (H) Diagram of proposed mechanism to reconcile similar PH3 staining with decreased regenerate size in creb3l1 ^{ΔbZIP/ΔbZIP} fish relative to creb3l1 ^+/+ fish. In wild‐type fish, PH3⁺ cells complete mitosis to generate two daughter cells, and rapidly loose the PH3 signal in late anaphase. In contrast, in creb3l1 mutant fish, PH3⁺ cells remain arrested at lateG2/early mitosis M, and continue to be stained with PH3, but do not complete cell division. Such cell cycle arrest would lead to overall smaller regenerates in creb3l1 mutant fish.

Since we observed neither, we explored whether the PH3 staining might be spurious in that the PH3 stained cells could be arrested during cell cycle. Serine 10 (Ser10) of Histone H3 is phosphorylated during mitotic chromatin condensation in late G2 and M phase of the cell cycle, but quickly dissipates after late anaphase. ⁴⁹ , ⁵⁰ Thus, robust PH3 staining would be visible in cells that are stalled in G2/M, and in mitotic cells during prophase, metaphase and early anaphase. To determine whether the PH3‐positive cells in creb3l1 mutant regenerates may exhibit cell cycle arrest, we examined the number of PH3⁺ cells in metaphase or anaphase within the regenerates of creb3l1 ^{+/+,Tg(Sp7:eGFP)/+} and creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} fish. We defined cells in metaphase as those having PH3‐stained DNA condensed in an equatorial position, and cells in anaphase as those that had two clearly separated PH3‐stained DNA structures (representative images are shown in Figure 6E). Analysis of analogous images detected a significant decrease in the number of metaphase (Figure 6F) and anaphase (Figure 6G) cells in the creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} regenerate. A possible explanation for these results is that the PH3‐positive cells in wild‐type regenerates undergo mitosis to generate two daughter cells (Figure 6H). In contrast, many of the PH3‐positive cells in creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} regenerates could be at a stage of a cell cycle prior to metaphase (Figure 6H). Thus, it appears that the lack of functional Creb3l1 causes a mitotic arrest, thereby providing a likely explanation for the smaller size of regenerates in the creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} fish.

2.6. creb3l1 mutant fish show aberrant patterning of the regenerating caudal fin

As shown in Figure 3, we detected a decrease in secondary bifurcations within each bony ray in intact tails of creb3l1 ^{ΔbZIP/ΔbZIP} fish. To determine whether a patterning defect also occurs during fin regeneration, we measured the number of ray bifurcations in regenerates in age‐matched 6 mpf creb3l1 ^{+/+,Tg(Sp7:eGFP)/+} and creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} fish by fluorescence imaging at 7, 9, and 14 dpa. As shown in Figure 7A (representative images at 7 dpa) and quantified in Figure 7B, significantly fewer total bifurcations were present in the creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} fish regenerates relative to wild‐type fish regenerates at all examined time points after amputation. The decrease was not due to a difference in primary bifurcations, as analogous number of primary bifurcation sites was present in wild‐type and mutant fish (Figure 7C). In contrast, there was a significant difference in the number of secondary bifurcations: while wild‐type fish showed secondary bifurcations on a number of regenerated rays (Figure 7A, white arrows), the creb3l1 ^{ΔbZIP/ΔbZIP} mutant fish lacked secondary bifurcations (Figure 7A, gray arrows mark lepidotrichia without secondary bifurcations). The decrease in secondary bifurcation was significant at every tested point after amputation, as quantified in Figure 7D.

FIGURE 7.

creb3l1 mutant zebrafish exhibit bifurcation defects during regeneration of the caudal fin. (A) Approximately 50% of the caudal fin was amputated from 6 mpf creb3l1 ^{+/+,Tg(Sp7:eGFP)/+} and creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} fish. At different dpa, the tails were imaged directly by fluorescence. Representative images of 7 dpa regenerates are shown. Amputation planes are marked with dashed white lines. The distance between the amputation plane and the primary bifurcation point in the second and third ray is marked with white lines. In wild‐type fish, secondary bifurcations are indicated by white arrows. In mutant fish, lack of secondary bifurcation is marked with gray arrows. Scale bar = 1 mm. (B) Images analogous to those in (A) were used to quantitate the number of total bifurcations in a 7, 9, and 14 dpa fin regenerate. Each data point represents an individual fish. n = 10–18. **p ≤ .01; ***p ≤ .001. (C) Images analogous to those in (A) were used to quantitate the number of primary bifurcations in a 7, 9, and 14 dpa fin regenerate. Each data point represents an individual fish. n = 10–18. *p ≤ .05. (D) Images analogous to those in (A) were used to quantitate the number of secondary bifurcations in a 7, 9, and 14 dpa fin regenerate. Each data point represents an individual fish. n = 10–18. *p ≤ .05; **p ≤ .01; ***p ≤ .001. (E) Images analogous to those in (A) were used to quantitate the distance from the amputation plane to first bifurcation in the second and third ray in the regenerate. Each data point represents an individual fish. Significant increase in distance is observed in creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} fish at all dpa. n = 12–20. *p ≤ .05. (F) Approximately 50% of the caudal fin was amputated from 6 mpf creb3l1 ^+/+ and creb3l1 ^{ΔbZIP/ΔbZIP} fish. At 9 dpa, the tails were stained with calcein fluorophore and imaged by fluorescence. Representative images of regenerates are shown. Amputation planes are indicated with dashed white lines. The distance between the amputation plane and the primary bifurcation point in the second and third ray is marked with white lines. In wild‐type fish, secondary bifurcations are indicated by white arrowheads. In mutant fish, lack of secondary bifurcation is marked with gray arrows. Scale bar = 1 mm. (G) Images analogous to those in F were used to quantitate the total number of bifurcations in the regenerating fin. Each data point represents an individual fish. n = 14. ***p ≤ .001. (H) Images analogous to those in F were used quantitate the distance from the amputation plane to first bifurcation in the second and third ray in the regenerate. Significant increase in distance is observed in creb3l1 ^{ΔbZIP/ΔbZIP} mutant fish. n = 14. **p ≤ .01.

Images analogous to those in Figure 7A were used to examined the relative position of the most proximal bifurcation point by measuring the distance between the amputation plane and the first bifurcation within the second and third ray of each regenerate of creb3l1 ^{+/+,Tg(Sp7:eGFP)/+} and creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} fish at 7, 9, and 14 dpa. Lines from the amputation plane to the bifurcation point were drawn manually on second and third ray (Figure 7A, white lines) and their length measured. As shown in Figure 7E, a significant increase in the distance between the amputation plane and the first bifurcation was observed in the creb3l1 ^{ΔbZIP/ΔbZIP,Tg(Sp7:eGFP)/+} mutant fish at all points post amputation.

We also measured the number of bifurcations and the position of bifurcations in age‐matched 6 mpf wild‐type and creb3l1 ^{ΔbZIP/ΔbZIP} mutant fish regenerates stained with calcein 9 dpa. As shown in Figure 7F and quantified in Figure 7G, significantly fewer bifurcations were present in the creb3l1 ^{ΔbZIP/ΔbZIP} mutant regenerates relative to wild‐type regenerates. The decrease was predominantly in secondary bifurcations, with wild‐type fish showing secondary bifurcations on a number of rays (Figure 7F, white arrows), while the creb3l1 ^{ΔbZIP/ΔbZIP} mutant fish had no secondary bifurcations (Figure 7F, gray arrows). Moreover, the position of the most proximal bifurcation point in second and third ray relative to the amputation point (Figure 7F, white lines) was increased in the creb3l1 ^{ΔbZIP/ΔbZIP} mutant regenerates relative to wild‐type regenerates (Figure 7H).

Thus, using the sp7:eGFP to visualize osteoblasts, or calcein staining to visualize newly formed mineralized tissue, we observed a significant decrease in the number of secondary bifurcations, as well as the distalization of the proximal branching point in the creb3l1 mutant fish.

2.7. creb3l1 mutant fish show aberrant expression of shha and ptch2

Ray bifurcation defects are often indicative of aberrant establishment of the proximo‐distal axis, an important patterning event in limb development. ³⁹ , ⁴⁰ Our results document significant bifurcation changes in fish lacking functional Creb3l1, and suggest that Creb3l1 participates in the establishment of the proximo‐distal axis during fin regeneration. A key pathway controlling fin regeneration is the Shh pathway, which functions in the basal epidermis to regulate bifurcation of the regenerating fin. ²² Current models postulate that Hedgehog signaling promotes the physical coupling of preosteoblasts to distally migrating basal epithelial cells, and that the tightly regulated expression of shha and its downstream target ptch2 leads to the characteristic branching configurations. ²² , ²⁵

Thus, we examined the expression pattern of shha and ptch2 in the regenerates of age‐matched 6 mpf creb3l1 ^+/+ and creb3l1 ^{ΔbZIP/ΔbZIP} fish at 3 and 7 dpa. The expression of shha in 3 dpa regenerating fins of wild‐type zebrafish is restricted to the distal tip of the regenerating ray, with the shha signal present in two distinct domains capping a branching ray (Figure 8A, black arrows), in agreement with previous reports. ³⁰ In contrast, in the creb3l1 ^{ΔbZIP/ΔbZIP} 3 dpa regenerate, shha expression was detected in a single diffuse domain atop an amputated ray (Figure 8A, blue arrow). This defect in shha patterning in creb3l1 ^{ΔbZIP/ΔbZIP} fish persisted at 7 dpa: while in the wild‐type creb3l1 ^+/+ regenerate, the shha signal was clearly separated into two concentrated foci, each on top of a bifurcated lobe of the regenerating lepidotrichia (Figure 8B, black arrows), in the creb3l1 ^{ΔbZIP/ΔbZIP} regenerate, the shha signal was less tightly segregated within two points, and some rays still contained only a single shha domain on top of as yet unbranched ray (Figure 8B, blue arrow).

FIGURE 8.

creb3l1 mutant fish exhibit aberrant patterning of shha and ptch2. (A, B) Approximately 50% of the caudal fin was amputated from 6 mpf creb3l1 ^+/+ and creb3l1 ^{ΔbZIP/ΔbZIP} fish. Regenerates at 3 dpa (A) and 7 dpa (B) were analyzed by in situ hybridization to assess the expression patterns of shha and ptch2. Amputation planes are indicated with dashed white line. Scale bar = 1 mm. (A) The shha signal in wild‐type creb3l1 ^+/+ fish exhibits the characteristic two‐point staining (black arrows). In contrast, in creb3l1 mutant fish, shha signal is largely in a single domain over the regenerating ray (blue arrow). The ptch2 signal in wild‐type creb3l1 ^+/+ regenerate is segregated into defined caps over regenerating rays with unstained areas between the caps (white arrow). In contrast, patch2 expression is more diffuse in creb3l1 ^{ΔbZIP/ΔbZIP} regenerate (white arrow). (B) In wild‐type creb3l1 fish, the shha signal appears in two clearly separated foci over each bifurcated ray (black arrows). In contrast, shha signal is more diffuse in creb3l1 mutant regenerates (blue arrow). The ptch2 signal in wild‐type creb3l1 ^+/+ fish segregates into two domains over each bifurcated ray (white arrows). In contrast, in creb3l1 ^{ΔbZIP/ΔbZIP} mutant regenerate, ptch2 signal predominantly remains within a single diffuse domain (white arrows). n = 3–7.

A difference also was observed in the expression pattern of patch2. In 3 dpa wild‐type regenerate, the ptch2 signal was restricted in a cap‐like structure centered over each regenerating ray, with minimal expression between the rays (Figure 8A, white arrow marks the boundary between concentrated foci of patch2 expression). In contrast, in the creb3l1 ^{ΔbZIP/ΔbZIP} 3 dpa regenerates, the patch2 signal was less organized, more punctate, and appeared to be diffusely present within the spaces between the rays (Figure 8A, white arrow). The defect in patch2 patterning in the creb3l1 ^{ΔbZIP/ΔbZIP} regenerates also was observed at 7 dpa: while in wild‐type fish, two clearly separate ptch2 “caps” decorated each bifurcated lobe of the regenerating lepidotrichia (Figure 8B, white arrows), in the creb3l1 ^{ΔbZIP/ΔbZIP} regenerate, the ptch2 signal appeared within a single larger “cap” positioned over the regenerating ray (Figure 8B, white arrow).

Together, these results indicate that a lack of functional Creb3l1 correlates with a perturbation in the Shh signaling pathway by either directly controlling events that regulate shha positioning or by regulating bifurcation events that subsequently may lead to irregularities in shha patterning as a downstream event.

3. DISCUSSION

Herein, we characterize a novel zebrafish model in which to assess the role of the Creb3l1 transcription factor in organismal development and physiology. Our creb3l1 mutant fish express Creb3l1 lacking the majority of its coding region and missing the essential DNA‐binding bZIP domain, thus preventing the Creb3l1 fragment from binding to DNA and acting as a transcription factor. Utilizing this new creb3l1 ^{ΔbZIP/ΔbZIP} zebrafish model, we aimed to understand the role of Creb3l1 in bone development and during the regeneration of the caudal fin as a paradigm for complex tissue regeneration.

Our analyses show that relative to wild‐type animals, the creb3l1 ^{ΔbZIP/ΔbZIP} fish do not exhibit significant changes in growth, skeletal morphology and bone density, but show a lower complexity of bifurcation in the caudal fin rays. Ray bifurcation patterns are characteristic in different fish species. ⁵¹ While in adult wild‐type zebrafish, we observed the characteristic secondary branching of the rays, the number of secondary bifurcations was significantly decreased in the creb3l1 ^{ΔbZIP/ΔbZIP} zebrafish.

The relatively mild phenotypes in zebrafish contrast with the serious defects in mammalian bone development caused by the lack of CREB3L1 or CREB3L1 mutations. Creb3l1 KO mice are significantly smaller than their littermates and despite having normal skeletal development, exhibit poor bone mineralization and frequent fractures. ¹¹ , ¹⁷ Humans with CREB3L1 mutations have severe osteogenesis imperfecta and even perinatal lethality. ¹² , ¹³ , ¹⁸ , ²⁰ , ⁴⁴ These differences highlight some of the limitations of the zebrafish model, as uterine development of humans and mice places more strain on the developing bones, as well as the force placed on bones when humans and mice begin to ambulate. Furthermore, zebrafish possess significantly more intramembranous bone compared to the endochondral bones affected by lack of functional Creb3l1 in humans and mice. Moreover, creb3l1 homologs creb3l2 and creb3l3l are broadly expressed in zebrafish, ⁵² , ⁵³ and their expression overlaps with creb3l1 expression in a cell cluster containing osteogenic markers, suggesting a possible compensatory mechanism during bone development. ⁵³ , ⁵⁴ Zebrafish creb3l2 (an ortholog of human CREB3L2 which regulates chondrogenesis ⁵⁵ ) and creb3l3l (an ortholog of human CREB3L3 enriched in the liver and regulating triglyceride metabolism therein ⁵⁶ , ⁵⁷ ) are broadly expressed in the regenerating fin, ⁵² possibly providing a compensatory mechanism during Creb3l1 deficiency. The occurrence of partial functional redundancy between Creb3l1 homologs is evident from experiments on Creb3l1 and Creb3l2 in endometrial stromal cells, with phenotypes becoming most severe when both CREB3L1 and CREB3L2 are knocked down. ⁵⁸ Thus, expression of creb3l2 and/or creb3l3l may provide a compensatory mechanism during the development of creb3l1 ^{ΔbZIP/ΔbZIP} fish.

We used our creb3l1 ^{ΔbZIP/ΔbZIP} fish as a model to explore the role of Creb3l1 in the regeneration of complex bony tissues. Utilizing the sp7:eGFP‐CAAX transgene (pd51Tg) ²⁷ to generate wild‐type creb3l1 ^+/+ and creb3l1 ^{ΔbZIP/ΔbZIP} fish expressing eGFP in osteoblasts, and the robust system of the regenerating caudal fin, we observed a significantly decreased postamputation regeneration in the creb3l1 mutant fish. The creb3l1 ^{ΔbZIP/ΔbZIP} fish tail regenerates were shorter in length and had a significantly reduced total area. Moreover, by utilizing calcein staining of mineralized tissue, we uncovered decreased mineralized area of bone in the creb3l1 ^{ΔbZIP/ΔbZIP} fish. This is consistent with a delay in the healing of bone fractures in Creb3l1 KO mice, which exhibit poorer mineralization of the bony callus. ¹⁷ Thus, our data suggest that the need for Creb3l1 function during bone regeneration is at least partially conserved between zebrafish and mammals.

The decreased regenerate size in the creb3l1 ^{ΔbZIP/ΔbZIP} fish prompted us to compare the levels of apoptosis and proliferation in wild‐type and creb3l1 mutant fish. Surprisingly, there were no significant changes in the number of cells with cleaved caspase 3 (a marker for cells undergoing apoptosis) or stained for PH3 (a marker for dividing cells) in wild‐type and creb3l1 ^{ΔbZIP/ΔbZIP} fish, implying analogous levels of cell apoptosis and division in wild‐type and mutant fish. Such results appeared contradictory to the smaller regenerates in the creb3l1 ^{ΔbZIP/ΔbZIP} fish, and suggested a possible mitotic arrest. Indeed, quantification of the number of PH3‐positive cells that have entered metaphase and anaphase showed that creb3l1 ^{ΔbZIP/ΔbZIP} fish regenerates have significantly fewer mitotic cells. PH3 associates with condensed chromatin in late G2 phase, and remains bound during mitosis from prophase to late anaphase, but rapidly dissociates from chromatin in late anaphase/early telophase. ⁴⁹ , ⁵⁰ Thus, a possible explanation for the difference in regenerate size in wild‐type and the creb3l1 ^{ΔbZIP/ΔbZIP} fish is that in wild‐type fish, cells divide continuously, with the PH3 pattern reflecting a snapshot in time, while in the creb3l1 ^{ΔbZIP/ΔbZIP} fish, many cells do not progress through mitosis, with the PH3 pattern reflecting an accumulation of mitotically arrested cells. Such a mitotic deficiency could be the result of several factors, including ER stress shown previously to cause cell cycle arrest at the G2/M phase. ⁵⁹ Creb3l1 has been reported as a transducer of ER stress, ⁵ , ⁸ , ¹⁰ , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ⁶⁰ , ⁶¹ and it is likely that cells lacking functional Creb3l1 exhibit increased ER stress, leading to a possible increase in G2/M‐arrested cells. Such a scenario would result in normal numbers of PH3⁺ cells within wild‐type and mutant regenerates, but nevertheless would cause reduced size of the mutant regenerate. Why this effect of CREB3L1 deficiency is only visible during tail regeneration and does not affect fish growth during development remains to be explored.

Examination of the ray patterning of the regenerating fins indicates a significant defect in bifurcations in the creb3l1 ^{ΔbZIP/ΔbZIP} fish at every examined point after amputation. This phenotype is due to a significant decrease in secondary bifurcations, while primary bifurcations appear analogous in wild‐type and mutant fish. Moreover, primary bifurcations in the creb3l1 ^{ΔbZIP/ΔbZIP} fish are significantly more distal to the amputation plane than in wild‐type fish. Such distalization phenotype is highly indicative of defects in the development of the proximo‐distal axis, a highly conserved process required for zebrafish and human limb development. ⁶² It is possible that once the caudal fin is fully regenerated, the number and location of the bifurcations relative to the amputation planes may reach parameters similar to the intact fin of creb3l1 ^{ΔbZIP/ΔbZIP} fish, but this remains to be determined.

The establishment of the proximo‐distal axis in the regenerating caudal fin is regulated by several pathways, including FGF, Wnt, Retinoic Acid, and Shh. ⁴⁰ Of particular interest to this study was Shh, due to its role in caudal fin bifurcation. ²⁵ , ³⁰ Additionally, recent findings in mice demonstrate that Shh signaling is selectively activated during the regeneration of severely damaged bone, while it is dispensable during healing of small fractures. ⁶³ This is especially relevant when considering the complex environment of caudal fin regeneration. When assayed via in situ hybridization, we found a change in the distribution of shha expression in the creb3l1 ^{ΔbZIP/ΔbZIP} regenerating fins: while shha expression in wild‐type regenerates is detected in distinct twin puncta at the distal tip, shha expression in the creb3l1 ^{ΔbZIP/ΔbZIP} regenerate exhibits a more diffuse staining, with the two puncta connecting in some fin rays to create a single point of signaling. Similarly, ptch2, an effector of Shh signaling, is expressed in wild‐type regenerate as distinct domains over the bifurcating ray, but in creb3l1 ^{ΔbZIP/ΔbZIP} regenerate, ptch2 is localized in a single domain at the tip of the regenerating ray. These results suggest that functional Creb3l1 is required to establish a restricted pattern of shha expression and signaling during morphogenic branching.

The exact mechanism through which Creb3l1 may influence Shh signaling remains to be determined. Within the developing intact caudal fin, as well as in the regenerate, Creb3l1 may affect the localization of shha signaling via a direct mechanism. It is also possible that Creb3l1 regulates positioning of branchpoints by intersecting with the Retinoic Acid (RA) pathway, as alterations in retinoic acid homeostasis also shift bifurcation distally. ⁶⁴ Creb3l1 may also directly or indirectly affect the localization and/or function of tartrate‐resistant acid phosphatase‐positive (TRAP⁺) osteolytic tubules (OLTs) shown to regulate branchpoint positioning during zebrafish fin regeneration. ⁶⁵ OLTs have been implicated in bone resorption required to inhibit the osteoblast‐dependent “stitching” fusion of regenerating rays, and inhibition of their osteolytic activity causes distalization of bifurcation points. The bifurcation defects observed in our creb3l1 ^{ΔbZIP/ΔbZIP} fish could be explained by Creb3l1 being required to correctly localize OLTs at the branching sites and/or promoting their bone resorbing activity. Such an effect could be indirect, and Creb3l1 could regulate the expression in osteoblasts of a signaling molecule that controls osteoclast activity, or Creb3l1‐mediated TSH homeostasis (see below) could control osteoclastogenesis.

Alternatively, Creb3l1 may impact the localization of bifurcations by affecting the thyroid hormone (TH) signaling pathway shown to be a regulator of branchpoint positioning in the regenerating tail. Decrease in TH levels causes marked reduction in the number of bifurcations and a significant distalization of branchpoints, without notable effects on the overall size of the fish, a phenotype similar to our creb3l1 ^{ΔbZIP/ΔbZIP} fish. ³⁹ Notably, in the absence of TH, shha domains fail to separate into distinct points, ³⁹ a phenotype observed in our creb3l1 ^{ΔbZIP/ΔbZIP} fish during regeneration. How could Creb3l1 control the TH pathway? TH formation by thyroid follicular cells requires the Sodium/Iodide Symporter (NIS) that mediates active iodide accumulation required for the synthesis of all iodide‐containing THs. ⁶⁶ Importantly, NIS expression has been shown to be modulated by CREB3L1 in rat thyroid follicular cells, with overexpression and inhibition of CREB3L1 inducing an increase and decrease in NIS protein, respectively. ⁶⁷ This raises the possibility that the lack of functional Creb3l1 within the zebrafish thyrocytes decreases the production of NIS, which causes a decrease in NIS‐mediated iodide uptake and decreased TH production. The reduced TH levels then cause defects in shha pathway signaling, leading to defects in bifurcation during development and regeneration. As such, we suggest that the alterations in bifurcation in creb3l1 ^{ΔbZIP/ΔbZIP} fish could be attributed to remote defects in TH production that nevertheless affect signaling locally within the caudal fin.

Together, our findings demonstrate that Creb3l1 plays an important role in the development and regeneration of the zebrafish caudal fin, by showcasing a novel functions for this protein in tissue patterning. Specifically, we show that Creb3l1 participates in re‐establishing the proximo‐distal axis during fin regeneration by impacting the shha signaling pathway. Recent findings on the essential role of Shh's in the healing of severe bone injury in mammals ⁶³ suggest that the intersection of shha signaling with Creb3l1 may have implications for therapeutic approaches to fracture healing in mammals. This is a novel finding, and may provide a foundation for future research on complex tissue regeneration. Our results significantly expand the current understanding of Creb3l1 function(s) by uncovering a previously unknown role for this protein in tissue patterning and as a potential modulation of Shh signaling.

3.1. Experimental procedures

3.1.1. Zebrafish husbandry

Zebrafish (D. rerio) were maintained in accordance with the University of Alabama at Birmingham's Institutional Animal Care and Use Committee, according to the Zebrafish Research Facility and the vertebrate animal welfare protocol (APN 21787). Adult fish were housed on an Aquaneering circulating water system, on a 14:10 light:dark cycle at 28°C. Fish were fed Gemma Micro 300 twice daily with one live Artemia feeding. Fish were bred using the standard system water and two‐piece tank system to prevent oophagy. Eggs were harvested, sorted for healthy eggs, and raised in E3 ⁶⁸ until placed on the system or utilized for experiments. Fish from ages 3–6 months were used and age‐matched for experiments.

3.1.2. Generation of creb3l1 mutant fish and SP7‐GFP transgenes

Alt‐R crRNA target sites were designed with Integrated DNA Technologies Alt‐R CRISPR HDR Design Tool (https://www.idtdna.com/pages/tools/alt‐r‐crispr‐hdr‐design‐tool). Alt‐R CRISPR‐Cas9 crRNA, tracrRNA (Integrated DNA Technologies) and Alt‐R S.p. Cas9 Nuclease V3 (Integrated DNA Technologies) was prepared following manufacturer instructions. Then, 3 μM sgRNA (Guide sequence: GGCGGAGAGGAACCCGTTGC) were obtained through diluting 100 μM crRNA and 100 μM tracrRNA into Nuclease‐Free Duplex Buffer (Integrated DNA Technologies), heating at 98°C for 5 min, then cooling to room temperature. Total of 0.5 μl Cas9 protein was diluted with Cas9 working buffer (20 mM HEPES; 150 mM KCl, pH 7.5) to yield a working concentration of 0.5 μg/μl. The diluted Cas9 protein working solution was mixed 1:1 with 3 μM sgRNA solution and then incubated at 37°C for 10 min to obtain RNP complex. RNP complex was freshly prepared and left on ice until microinjection. Microinjection was performed by injecting ~1 nl of RNP complex into yolk of 1‐cell stage wild‐type (AB) embryos.

F0 adults generated from this injection were crossed to AB fish to generate heterozygous offspring. Heterozygous offspring were then sequenced via Sanger Sequencing (primers: F: 5′‐TTTGTGGTCTCTCTCCAGCA‐3′, R: 5′‐GGATGGATGGCAGGAAAGTA‐3′) to determine sequences of CRISPR indels. Sequenced heterozygous fish were crossed to generate homozygous offspring. These offspring were sequenced once more to verify homozygosity for the mutant allele.

For analysis of fin rays in vivo, creb3l1 mutant fish were outcrossed to AB Ola.Sp7:EGFP‐CAAX ²⁷ fish, henceforth referred to as Sp7:eGFP, or creb3l1 ^{Tg(Sp7:eGFP/+)} fish. Resultant heterozygotes were then incrossed with homozygous creb3l1 mutants to generate fish homozygous for their respective mutations, and heterozygous for the Sp7:eGFP transgene. Fish were screened at 3–5 dpf for fluorescence in the craniofacial bones and raised according to standard husbandry protocols.

3.1.3. Genotyping by high resolution melt analysis in zebrafish

To isolate genomic DNA, adult fish were anesthetized in 250 mg/L Tricaine until opercular motion slowed. Tail clippings were collected from fish and incubated at 98°C for 20 min in 40 μl 25 mM NaOH in a 96‐well plate; then neutralized with 40 μl of 40 mM Tris–HCl. For genotyping of whole embryos or larvae, the incubation time at 98°C was shortened to 10 min. PCR reactions contained 1 μl of LC Green Plus Melting Dye (Biofire Defense), 1 μl of 10x enzyme buffer, 0.2 μl of dNTP Mixture (10 mM each), 0.3 μl of 15 mM MgCl₂, 0.3 μl of each primer (F: 5′‐GATCATCTGTTGGCGGAGAG‐3′, R: 5′‐TTGGAGGGTCCAGATCCATA‐3′) (10 μM), 1 μl of genomic DNA, 0.05 μl of Taq DNA Polymerase (Genscript), and DNA/RNAse free water up to 10 μl. The PCR reaction protocol was 98°C for 30 s, then 45 cycles of 98°C for 10 s, 59°C for 20 s, and 72°C for 15 s, followed by 95°C for 30 s and then rapid cooling to 4°C. Following PCR, melting curves were generated and analyzed using the LightScanner instrument (Idaho Technology) over a 65–95°C range.

3.1.4. RT‐PCR

Here, 7 dpf larvae were anesthetized with 250 mg/L tricaine and collected in a 1.5 ml Eppendorf tube (~30 fish/pool). Excess embryo media was removed and 350 μl of Trizol was added. Pooled larvae were homogenized by pestle in Trizol and purified using a Direct‐zol RNA Miniprep Plus Kit from Zymo Research, including gDNA removal. The concentration of the RNA was measured using a NanoDrop ND‐1000. RNA with a 260/280 of ~2.0 was used for further analyses. Up to 2 μg of RNA was utilized to synthesize cDNA in the High‐Capacity cDNA Reverse Transcription Kit (Applied Biosystems) with RNaseOUT Recombinant Ribonuclease Inhibitor (Invitrogen). Prepared cDNA was used for RT‐PCR, using a SYBR Green Power UP (Applied Biosystems) master mix (all primers are listed in Table 1). A 10 μl final volume reaction was prepared, with 1 μl of cDNA. This was run using the standard cycling time/temperatures for SYBR Green Power UP with 0.8 mM Primer concentration (all primers are listed in Table 1). Fold change was calculated via the 2^−ΔΔCt method. ⁶⁹ ΔCt values were utilized for statistics and were compared via t test or Mann–Whitney test, as described below.

3.1.5. Micro computed tomography

Age‐matched fish were fixed in 4% PFA (paraformaldehyde) in phosphate buffered saline (PBS, pH 7.4) with 0.5% Triton X‐100 overnight at 4°C. Fixed zebrafish were scanned using the Scanco μCT40 desktop cone‐beam micro‐CT scanner (Scanco Medical AG, Brüttisellen, Switzerland) using μCT Tomography v6.4‐2 (Scanco Medical AG, Brüttisellen, Switzerland). Scans were automatically reconstructed into 2‐D slices and slices were analyzed using the μCT Evaluation Program (v.6.5‐2, Scanco Medical). The zebrafish were placed in a 36 mm diameter scanning holder and scanned at the following settings: 18 μm voxel size, 70 kVp, 114 μA, 1000 projections/180° with an integration time of 200 ms. The region of interest was drawn around the fish and the analysis was performed on all slices that any part of the fish appeared in. Bone was thresholded at 140.3 mgHA/cm³ and the 3‐D analysis performed. Data were obtained on bone volume and density. 3D images were obtained from the 3D evaluation software (μCT Ray v.4.2, Scanco Medical).

3.1.6. Measurement of uninjured fish

The length of fish at 6 months post fertilization (mpf) was measured with a ruler and recorded. Tails of both wild‐type and mutant fish carrying the sp7:eGFP transgene were imaged on a Nikon AZ100 dissection microscope with fluorescence filters to visualize gross morphology of the caudal fin. Photos were analyzed in FIJI (ImageJ 1.54f) ⁷⁰ to measure the indicated parameters.

3.1.7. Caudal fin amputation model of regeneration and visualization of mineralized regenerate

Age‐matched fish were anesthetized in a 250 mg/L tricaine until opercular motion was slowed and fish were no longer responsive to touch. Fish underwent 50% tailfin amputation via transection with a scalpel blade. Tails were photographed after amputation. Fish were then recovered in system water and allowed to regenerate for the indicated times. At the indicated day post amputation (dpa), the fish were immersed in a 0.2% calcein solution (pH 7.5) ³³ , ⁷¹ , ⁷² for 20 min, briefly rinsed, and then de‐stained in fresh system water for 20 min. The stained fish were imaged on a Nikon AZ100 dissection microscope in both bright‐field and fluorescence filters to visualize gross morphology and mineralization of the regenerate. Photos were analyzed in FIJI (ImageJ 1.54f) ⁷⁰ to measure the indicated parameters.

The size and patterning of the mineralized regenerate was evaluated by a protocol adapted from Cardeira et al. ²⁶ Briefly, utilizing FIJI (ImageJ 1.53f51), ⁷⁰ the total area of the regenerate (REG) of the amputated tail was measured using the FIJI Measure function, then normalized to the width of the stump at the amputation plane (STU). To measure the RMA, the polygon tool was utilized to outline the fluorescently labeled bones with the distal tip of each lepidotrichia and the base of the regenerated area marking the boundary of the estimated mineralized area. Using this selected region, the threshold tool was utilized to isolate pixels of fluorescently stained bone based on individual image intensity histograms (16 bit image binning). The total area of these particles was then analyzed via the Analyze Particles function in FIJI. In addition, the distance from the amputation plane to bifurcation in both, calcein stained and transgenic fish was measured from the amputation plane to the base of the bifurcation on the second and third most lateral fin rays on the tail, using the line segment tool. Rays with bifurcation points proximal to the amputation plane were excluded from the analysis.

3.1.8. Visualization of Sp7⁺ cell populations during regeneration

For dynamic analyses of caudal fin regeneration, wild‐type and CREB3L1 mutant fish expressing Sp7:eGFP underwent amputation as described above. Fish were returned to system water, and imaged using a Nikon AZ100 dissection microscope at 3, 7, 9, and 14 dpa, under tricaine anesthetic. The images were evaluated for the length of the regenerate and the area of Sp7⁺ cells, using the same methods used to quantify and normalize RMA (see above). In addition, the bifurcation depth in both calcein stained and transgenic fish was measured as described above.

3.1.9. Detection of apoptotic and proliferating cells

Fins were amputated as described above. Then, 3 dpa fins were collected and fixed in 2% PFA in PBS (pH 7.4) overnight at 4°C. Fixed tails were processed as described. ⁵¹ Briefly, fixed fins were washed in PBS, blocked in 5% goat serum in PBS with 0.3% Triton‐X (PBSTX) for 1 h, and stained overnight with primary either anti‐Cleaved Caspase 3 (BD Biosciences, diluted 1:500) or anti‐phospho‐histone H3 (Ser10) (Cell Signaling Technology, diluted 1:1000). Tails were washed in PBSTX twice (once for 5 min, and once for 1 h), and stained with goat‐anti‐rabbit IgG‐Alexa Fluor Plus 594 (Thermo Scientific, diluted 1:500) in blocking solution for 2 h at room temperature. Stained tails were washed as above and mounted with glass coverslips on glass slides in 90% glycerol with 0.5% propyl gallate. Representative images were captured with a Nikon A1R‐HD25 Confocal microscope, with a 10× objective and processed with FIJI. To quantify the number of stained cells, tails were imaged on the Nikon Ti2 widefield microscope at 10× or 20× objective, and the images analyzed via FIJI. ⁷⁰

3.1.10. Probe preparation for in situ hybridization

Probes were generated as described by in vitro transcription with digoxigenin‐labeled UTP (Roche) of PCR products engineered to have a T7 site (primers in Table 1) or linearized plasmids (details in Table 2).

TABLE 2.

Details of plasmid preparation to generate in situ probes.

Gene	Linearizing enzyme (Promega)	Polymerase
creb3l1	NotI	T3 (New England Biolabs)
msxb	SalI	T7 (Fisher Scientific)
ptch2	XbaI	T3
shha	HindIII	T7

3.1.11. In situ hybridization

Then, 3 and 7 dpa fins were amputated via scalpel and fixed overnight at 4°C in 4% PFA in PBS (pH 7.4). The tails were treated as previously described, ⁷³ with slight modifications. Briefly, fins were washed in PBS, and dehydrated in 100% methanol for at least 2 h at −20°C. Dehydrated fins were rehydrated stepwise with 5 min washes in ethanol (75% EtOH/25% PBST, 50% EtOH/50% PBST, 25% EtOH/75% PBST), and washed in PBS with 0.1% Tween 20 (PBST) before 1 h of prehybridization and subsequent overnight hybridization with 0.5 mg/ml digoxigenin‐conjugated probes in hybridization buffer (HB) containing 50% Formamide, 5xSSC saline sodium citrate, 1 mg/ml Torula RNA (Sigma‐Aldrich Cat.#R6625), 100 μg/ml heparin, 1x Denhart's, 0.1% Tween 20, 5 mM EDTA ⁷⁴ in a 65°C water bath. Hybridized tails were washed for 10 min at 65°C in stepwise HB/2xSSC solutions (75% HB/25% 2xSSC, 50% HB/50% 2xSSC, 25% HB/75% SSC) followed by a series of 0.2xSSC/PBS 5 min washes with (75% 0.2xSSC/25% PBST, 50% 0.2xSSC/50% PBST, 25% 0.2xSSC/75% PBST). Washed tails were blocked in 2 mg/ml BSA in PBST for 1 h at RT, and then incubated with tail pre‐adsorbed anti‐DIG antibody fragments (Roche, Sigma‐Aldrich, diluted 1:2000) for 2 h at RT. Tails were then washed 3x with a staining solution (100 mM Tris–HCl [pH 9.5], 50 mM MgCl₂, 100 mM NaCl, 0.1% Tween 20, 1 mM levamisole) and incubated in BM Purple (Roche, Sigma‐Aldrich) until developed (overnight at RT or 4°C to reduce background). Stained tails were washed in multiple PBST washes and fixed with 4% PFA overnight at 4°C. and washed again multiple times. Prepared tails were mounted in a 12 well culture plate in PBST and imaged using a Leica M205 FCA dissection microscope.

3.2. Statistical analyses

All experiments were performed in duplicate or greater. Data were collected and analyzed with GraphPad Prism—Version 10. Mendelian ratio data (Figure 2B) was analyzed using a chi‐square analysis of observed vs. expected values. All other data comparing creb3l1 wild‐type versus mutant fish were first analyzed for normality of the data and descriptive statistics. Based on these analyses, the appropriate analyses of non‐paired data were chosen; for normal data, unpaired t tests were utilized, with or without Welch's correction for differences in standard deviation. For data that were not normally distributed, the nonparametric Mann–Whitney test was utilized. Significance was determined as a p < .05 threshold. (*p ≤ .05; **p ≤ .01; ***p ≤ .001, ****p ≤ .0001).

FUNDING INFORMATION

This work was supported by awards from the National Institute of General Medical Sciences (R01GM122802 to ES) and the National Institute of Dental and Craniofacial Research (5T90DE022736‐09 to PEV).

Supporting information

SUPPLEMENTARY FIGURE 1. creb3l1 mutant zebrafish have normal overall skeletal morphology. Age‐matched 3 mpf creb3l1 ^+/+ and creb3l1 ^{ΔbZIP/ΔbZIP} fish were analyzed by μCT and representative scans of male and female animals are shown. The creb3l1 ^{ΔbZIP/ΔbZIP} fish exhibit overall normal skeletal morphology.

VanWinkle PE, Lee E, Wynn B, et al. Disruption of the creb3l1 gene causes defects in caudal fin regeneration and patterning in zebrafish Danio rerio . Developmental Dynamics. 2024;253(12):1106‐1129. doi: 10.1002/dvdy.726