Cis-regulatory control of mammalian Trps1 gene expression

Abstract

Background:

TRPS1 serves as the causative gene for tricho-rhino phalangeal syndrome, known for its craniofacial and skeletal abnormalities. The Trps1 gene encodes a protein that represses Wnt signaling through strong interactions with Wnt signaling inhibitors. The identification of genomic cis-acting regulatory sequences governing Trps1 expression is crucial for understanding its role in embryogenesis. Nevertheless, to date, no investigations have been conducted concerning these aspects of Trps1.

Results:

To identify deeply conserved noncoding elements (CNEs) within the Trps1 locus, we employed a comparative genomics approach, utilizing slowly evolving fish such as coelacanth and spotted gar. These analyses resulted in the identification of eight CNEs in the intronic region of the Trps1 gene. Functional characterization of these CNEs in zebrafish revealed their regulatory potential in various tissues, including pectoral fins, heart, and pharyngeal arches. RNA in-situ hybridization experiments revealed concordance between the reporter expression pattern induced by the identified set of CNEs and the spatial expression pattern of the trps1 gene in zebrafish. Comparative in vivo data from zebrafish and mice for CNE7/hs919 revealed conserved functions of these enhancers. Each of these eight CNEs was further investigated in cell line-based reporter assays, revealing their repressive potential. Taken together, in vivo and in vitro assays suggest a context-dependent dual functionality for the identified set of Trps1-associated CNE enhancers.

Conclusion:

This functionally characterized set of CNE-enhancers will contribute to a more comprehensive understanding of the developmental roles of Trps1 and can aid in the identification of non-coding DNA variants associated with human diseases.

Graphical Abstract

Introduction

The TRPS1 gene, located on human chromosome 8q24, encodes a zinc finger transcription factor spanning approximately 260.5 kb with seven exons (Momeni et al., 2000). It is the causative gene for human disease phenotypes, tricho-rhino phalangeal syndrome type I and III (TRPSI: OMIM190350 and TRPSIII: OMIM 190351), both of which are autosomal dominant disorders and characterized by craniofacial and skeletal malformations (Giedion, Burdea, Fruchter, Meloni, & Trosc, 1973). Craniofacial abnormalities in TRPSI patients include sparse scalp hairs, distinctive pear-shaped nose tip, a long flat philtrum, and protruding ears with a thin vermillion border (Momeni et al., 2000). Skeletal abnormalities in individuals associated with TRPS1 include cone-shaped epiphyses at phalanges, hip malformations in most cases, and occasionally short stature (Momeni et al., 2000). TRPSIII phenotypes differ from TRPSI, exhibiting severe brachydactyly and retarded growth (Ludecke et al., 2001).

TRPS1 belongs to the GATA family of transcription factors (Malik et al., 2001). Unlike other family members (GATA1 to GATA-6), that possess two C4-type zincfingers (ZFs) and function as transcriptional activators (Molkentin, 2000), TRPS1 exhibits a distinctive structure.TRPS1 is equipped with nine ZFs, with only one of them being a GATA C4-type, and it functions as a transcriptional repressor (Kaiser et al., 2003; Malik et al., 2001; Suemoto et al., 2007). These 9 ZFs spans from the N terminus to C terminus of TRPS1, with the first six lacking significant similarities in protein databases, the seventh resembling the GATA domain, and the remaining two at the C-terminus being similar to the ZFs of the IKAROS family of transcription factors (Chang et al., 2000; Momeni et al., 2000). Numerous missense mutations have been reported in the GATA domain of TRPSI patients (Ludecke et al., 2001; Smaili et al., 2017). Deletion of this GATA motif in a mouse model, is sufficient to induce known TRPS phenotypes (Malik, Von Stechow, Bronson, & Shivdasani, 2002). For proper DNA binding, an intact GATA ZF is crucial, while for the repressive activity, TRPS1 depends on two IKAROS-resembling C-terminal ZF domains (Brown et al., 1997). These C-terminal ZF domains,resembling IKAROS, mediate protein-protein interactions, facilitating the binding of TRPS1 to various repressor proteins such as CtBP, Brg1 and Sin3 (Koipally & Georgopoulos, 2000; Koipally, Renold, Kim, & Georgopoulos, 1999; J. H. Wang et al., 1996). A truncated Trps1 protein with 119 residues missing in its C-terminal is unable to repress GATA-4 mediated transcription, whereas a fusion protein containing GATA-4 and C119 is capable of such repression (Brown et al., 1997).

TRPS1 primarily functions as a transcriptional repressor, playing a crucial role in chondrocyte regulation and the development of the perichondrium (Napierala et al., 2008). Notably, TRPS1 also acts as a transcription activator, participating in the transcriptional activation of Wnt inhibitors such as Wif1, Apcdd1, and Dkk4 through binding to their promoter regions (Fantauzzo & Christiano, 2012). This represents a distinct role for Trps1, involving transcriptional activation in contrast to its well-established function as a transcriptional repressor (Sun, Gui, Shimokado, & Muragaki, 2013). At the genomic level, Trps1 has been extensively investigated for its active involvement in transcriptional repression through direct association with recognized corepressors and inhibitors (Witwicki et al., 2018). In vitro assays conducted by Malik et al., 2001 demonstrated that Trps1 acts as a transcriptional repressor, with this activity is determined by its DNA binding GATA domain and IKAROS like ZF motifs (Malik et al., 2001). Recently reported data has elucidated the context-dependent regulation of Trps1 during the differentiation of normal mammary epithelial cells and the development of breast cancer (Cornelissen et al., 2020).

The previously reported endogenous expression pattern analysis of the Trps1 gene in mice conducted through in situ methods unveiled a highly complex and dynamic expression profile during early embryonic development (Kunath, Ludecke, & Vortkamp, 2002). Specifically, Trps1 mRNA is detectable in mouse embryos before embryonic stage E7.5 (embryonic day), with peak levels observed at E11.5 (Malik et al., 2001). Between E12.5 to E14.5, Trps1 expression intensity is notably high in facial regions, pharyngeal arches, limb joints, maxilla, mandible, snout, developing phalanges, and hair follicles (Kunath et al., 2002; Malik et al., 2001; Suemoto et al., 2007). Moreover, various organs including the gut, developing lungs, kidneys and mesonephric ducts, exhibit Trps1 expression in mouse (Gai et al., 2009; Kunath et al., 2002). Consistent with this broad and dynamic expression pattern, newborn Trps1−/− mice display reduced hair follicles and various abnormalities, including short bones in limbs, defects in facial and thoracic regions, as well as tissues in the lungs and kidneys (Gai et al., 2009; Malik et al., 2002; Michikami et al., 2012; Suemoto et al., 2007). The multifaceted functions and expression dynamics of the mammalian Trps1 gene underscore a highly orchestrated control of its endogenous expression during early embryonic development.

Genes involved in vertebrate development are often evolutionarily deeply conserved. Consequently, there has been a significant interest over the past couple of decades in identifying conserved non-coding elements (CNEs) in and around mammalian developmental genes, implicating them as transcriptional enhancers in gene regulation (Chatterjee & Ahituv, 2017; Kioussis, Vanin, deLange, Flavell, & Grosveld, 1983). Genetic variations in these CNE-enhancers have frequently been associated with mammalian/human disease phenotypes (Jeong et al., 2008; Lettice et al., 2003; Ragvin et al., 2010). In this study, we employed a comparative genomics approach to predict Trps1-associated transcriptional enhancers by comparing genomes of mammals with slowly evolving fish such as coelacanth and spotted gar (Figure 1). These comparative alignments led to the discovery of eight anciently conserved CNEs in the intronic regions of the Trps1 gene. These intragenic CNEs underwent functional testing in zebrafish transgenesis and cell lines-based reporter assays. They functioned as transcriptional enhancers in the zebrafish transgenic assay, driving expression in the pharyngeal arches, heart, and pectoral fin. Cell lines-based reporter assays revealed their cellular context dependent repressive role in transcription, suggesting their dual functionality.

Figure 1: Genomic conservation map of the TRPS1 locus across multiple species.

The Vista alignment graphical output generated by the multi-lagan tool using the human TRPS1 gene sequence as the baseline. The black arrow at thetop indicates the direction of transcription for Trps1. The light orange highlighted regions represent CNEs selected for functional testing in the zebrafish transgenic assay, with a cutoff criterion of 50 bp width and 50 % conservation. Exons are highlighted in light blue. The Y-axis represents the percent conservation of genomic sequences, while the X-axis displays the length of conserved sequences. Ex: exon, CNE: conserved noncoding element, K: Kilobases.

Materials and Methods

Search for cis-acting regulatory elements in TRPS1 gene

The human TRPS1 gene sequence along with a 100 kb flanking region, and its orthologs were retrieved from the Ensemble Genome Browser (https://www.ensemble.org) utilizing the GRCh38 assembly for human, GRCm38 for mouse, ROS Cfam 1.0 for dog, Loxafr3.0 for elephant, LatCha1 for coelacanth, GRCz11 for zebrafish, and LepOcu1 for spotted gar (Cunningham et al., 2022) (Supporting Information: Figure S1). A multiple species sequence comparison was conducted employing the MLAGAN global alignment tool (Brudno et al., 2003). The human TRPS1 gene sequence served as a baseline and was annotated by using exon/intron information available on the Ensemble Genome Browser (Martin et al., 2022). The MLAGAN alignments were then visualized using VISTA tools (Mayor et al., 2000). The alignment was performed with a 50 bp window and a cutoff score of 50% identity (Figure 1; Supporting Information: Figure S1).

Putative TFBSs discovery and motifs search

To identify putative transcription factor binding sites (TFBSs), the orthologous sequences of the identified CNEs were submitted to MEME tool (Bailey & Gribskov, 1998). MEME tool is capable of predicting un-gapped sequence motifs in query sequences (Bailey, Williams, Misleh, & Li, 2006). The search for TF binding motifs had a minimum length criterion set at 6–12 base pairs (bp) (Supporting Information: FigureS2). The Identified sequence motifs were further analyzed using STAMP tool to predict binding sites for transcription factors (TFs) against the TRANSFAC database (v.11.3) (Matys et al., 2003). The predicted TFs were subsequently examined for their endogenous expression using the MGI (Mouse Genome Informatics) database (Supporting Information: Table S1).

Genomics conservation and interactions analysis

Predicted enhancers are frequently associated with a specific target gene body through a comparative syntenic investigation of the enhancer containing locus and an analysis of the endogenous expression patterns of neighboring genes (Ali et al., 2016; Parveen et al., 2013). In this study, the human locus (GRCh38: Chromosome 8: 115,408,496–115,809,673) containing predicted TRSP1 CNE-enhancers underwent syntenic comparison with other vertebrate species (Mouse and Zebrafish) using the region comparison option available at the Ensemble Genome Browser (https://asia.enemble.org) (Martin et al., 2022). Based on this analysis, a schematic was created to illustrate the conserved gene neighborhood spanning approximately 3 Mb on either side of the TRPS1 gene (Figure 2A). The conserved syntenic association between CNE-enhancer regions and one or more neighboring genes might indicate a functional association (Parveen et al., 2013). Furthermore, the reported endogenous expression patterns (MGI: RNA in-situ hybridization) of the neighboring genes (Trps1, Eif3h and Csmd3) were examined to precisely assign the target gene body to the predicted CNE-enhancers (Supporting Information: Table S2). However, synteny comparison and inspection of the endogenous expression patterns of bracketing genes failed to assign the precise target gene body to the identified set of CNE-enhancers (Figure 2A;Supporting Information: Table S2). To address this challenge, we utilized the 3D genome browser (http://3dgenome.fsm.northwestern.edu/) and the Juicebox web app (https://www.aidenlab.org/juicebox/), which provide tools for visualizing Hi-C data from multiple tissues and cell lines available in their database or provided locally (Robinson et al., 2018; Y. Wang et al., 2018). Hi-C data from H1-ESC (Dixon2015-raw) were employed to identify spatial interactions by visualizing chromatin topology at the TRPS1 locus (Figure 2B). Chromatin interactions and physical contacts in the region containing the TRPS1 promoter and predicted CNE-enhancers were observed from topologically associated domains (TADs) structure at the TRPS1 locus (Figure 2B; Supporting Information: Figure S3 & S4). The heatmap in Figure 2 and Figure S3 illustrating chromatin interactions at the TRPS1 locus, was generated by the 3D genome browser (Robinson et al., 2018). Additionally, the heatmaps in Figure S4 were generated by the Juicebox web app using data from GM12878 cells (ENCFF053VBX), K562 cells (ENCFF013TGD), zebrafish embryos (GSM4724554),and mESC (4DNFI4OUMWZ8). The TRPS1 gene schematics, showing exons and predicted CNEs, are aligned according to the scale of their coordinates (Figure 2B).

Figure 2: Synteny map of the TRPS1 locus and physical interaction of the TRPS1 promoter with conserved CNEs.

A: Comparative synteny among human, mouse, and zebrafish Trps1-containing loci demonstrates deep conservation of several genes, including TRPS1, CSMD3, EIF3H, UTP23, RAD21, AARD, SLC30A8, and MED30, extending down to zebrafish. The identified set of Trps1 intronic CNEs cannot be associated with a single target gene body through syntenic analysis. Syntenically conserved genes are color coded, while non-conserved genes are depicted in black. Loci are approximately drawn to scale. B: The chromatin contact (Hi-C) heatmap at 25-kb resolution of 1.5-Mb region on chromosome 8 spanning the TRPS1 gene (promotor and predicted CNEs) in human embryonic stem cells (H1-ESC) (data from Jin et al., 2013). H3K27Ac marks (indicative of regulatory elements) from ENCODE and vertebrate conserved regions from PhyloP60Way tracks are displayed below the heatmap. The TRPS1 embedding region exhibits high color intensity, indicating increased chromatin contacts within the region. In panel B, the TRPS1 promoter is represented by a green rectangle labeled with the letter P, exons are depicted as purple boxes labeled with letter E, and CNEs are represented by red boxes. kb: kilobases, P: promoter, E: exon, CNEs: conserved noncoding elements.

Zebrafish transgenesis assay

The human versions o fthe identified set of TRPS1-associated CNEs were amplified from human genomic DNA using the primers provided in the Table S3. PCR products of the correct size were purified using a Gel purification kit (Thermo scientific) and cloned into the pCR8/GW/TOPO vector through direct ligation, following the manufacturer’s instructions (Invitrogen, Life Technologies). To confirm the orientation of our inserts, both flanking regions and inserts were sequenced via Sanger sequencing. Subsequently, the CNEs inserts were transferred to the destination vector (pGW-cfos-EGFP) through gateway cloning. LR recombination, utilizing attL and attR sites, was facilitated by the LR clonase enzyme, combining the entry construct (100ng/uL) and destination vector (100ng/uL). The resulting destination constructs were verified through restriction digestion and Sanger sequencing.

Adult healthy zebrafish were maintained and bred following standardized protocols. Freshly spawned eggs were collected from mating wild type zebrafish. To synthesize mRNA encoding the transposase enzyme, the pCS-TP vector was linearized using NotI enzyme digestion for subsequent in vitro transcription with the SP6 mMessagem Machine kit (Ambion). The freshly synthesized mRNA was precipitated using Lithium chloride and washed with 100 % ethanol. For microinjection into zebrafish embryos, the protocol provided by Fisher et al., 2006 was followed (Fisher et al., 2006). The injection solution, consisting of 1 uL of the reporter vector (pGW-CNE-cfos-EGFP, 125 ng/uL), 1 uL transposase RNA (175 ng/uL), 0.5 uL phenol red (0.5%), and 2.5 uL nuclease free-water, was calibrated to approximately 2 nL and injected into the animal pole of zebrafish embryos at the one cell stage. After injection, embryos were maintained in E3 media (1X) at 28.5°C.,Phenylthiourea (PTU) at 0.003% was added to the media after 10 hours to prevent pigmentation.

The embryos were raised for 5 days and manually dechorionated using sharp forceps. GFP expression in transgenic embryos was screened at various stages, including 24 hpf, 48 hpf, 72 hpf, 96 hpf and 120 hpf. Tricaine was used to anesthetize embryos for microscopy under an IX71 inverted fluorescence microscope (Olympus, Japan). Live images of zebrafish transgenic embryos were captured using the DP72 camera with monochrome software. For an element to be considered positive, it had to exhibit consistent expression in 20% or more of the transgenic embryos.

Luciferase reporter assay

For the luciferase reporter assay, CNEs were amplified by PCR and cloned into the pGL3 vector, which contains the TK (thymidine kinase) minimum promoter (pGL3-tk-mini). The clones were verified by Sanger sequencing. NIH3T3 and C3H10T1/2 cells were cultured in DMEM supplemented with 10% FBS and 10U/mL penicillin/streptomycin. In each well of a 48-well plate, 3 × 10⁴ cells were plated. Following the manufacturer’s (Promega) instructions, Fugene 6 was utilized to transfect cells with 100 ng of luciferase reporter constructs and 5 ng of pRL-TK. Cells were harvested 40–44 hours post transfection, and luciferase activities for all wells were measured using the Promega dual reporter assays system. Experiments were conducted with n=4, and the results were presented as averages along with standard deviation. The data presented in Figure 6 represent a typical outcome from more than three independent experiments.

Figure 6: Expression pattern of zebrafish trps1 transcripts.

Whole-mount RNA in situ hybridization was conducted in zebrafish embryos at various developmental stages, spanning from 1 hpf to day 5. (A) 1 hpf lateral view. (B) 3 hpf lateral view. (C) 12 hpf lateral view. (D) 24 hpf dorso-lateral view. (E) 48 hpf dorsal view. (F) 48 hpf lateral view. (G) 60 hpf dorsal view. (H) 60 hpf lateral view. (I) 96 hpf dorsal view. (J) 120 hpf dorsal view. At 1 hpf trps1 expression is observed in the animal pole. By 48 hpf, trps1 expression intensifies in the head and tail regions, and at 96 hpf it becomes pronounced in the fins. The black arrow heads indicate the expression domains. hpf: hours post fertilization, h: head,pf:pectoral fins, t: tail, pa: pharyngeal arches, fb: forebrain, mb: midbrain

Whole mount in situ hybridization

To elucidate the endogenous expression pattern of trps1 during zebrafish embryogenesis, whole mount RNA in situ hybridization (WISH) experiments were conducted following the standard protocol outlined by Thisse and Thisse (Thisse & Thisse, 2008). Briefly, forward (5’-GCAAGACGATCTCTCCAGT-3’) and reverse primers (5’-GTCATACTGTTGCCATAGTG-3’) were designed for the amplification of the zebrafish trps1 region, including flanking parts of the final exon and 3’-UTR. To synthesize the antisense probe against trps1 mRNA, the promoter sequence of SP6 RNA polymerase was added to the reverse primer. The amplified PCR product was used to synthesize the riboprobe via in vitro transcription using the SP6 polymerase kit (mMessagem Machine from Ambion). Zebrafish embryos at various stages (24 hpf, 48 hpf, 60 hpf, 96 hpf and 120hpf) were fixed overnight in 4 % paraformaldehyde at 4°C. Embryos at 1 hpf, 3 hpf and 12 hpf were dechorionated post-fixation. For stages beyond 12 hpf, embryos permeabilized by proteinase K digestion. Zebrafish embryos were then hybridized with 50 ng of the trps1 antisense riboprobe in 200 ul of hybridization mix. Subsequently, embryos were washed and stirred in blocking buffer for 1 hour at room temperature. After blocking, embryos were incubated at 4°C under gentle agitation in fresh blocking buffer containing a 6000-fold dilution of Anti-Digoxygenin-Ap Fab fragments (Roche). The embryos were then washed with PBT (PBS+0.01% tween 20). Finally, color development was achieved by labelling embryos with NBT and BCIP in an alkaline tris buffer. Embryos were periodically observed under a microscope to track color development, and images of stained embryos were captured using a Discovery V8 SteREO microscope.

Results

Comparative sequence analysis of the human TRPS1 locus

To identify cis-regulatory elements controlling the spatiotemporal expression of the mammalian Trps1 gene, the human TRPS1 genomic sequence (approximately 260.5 kb) was compared with its orthologous counterparts from mouse, dog, elephant, zebrafish and slowly evolving fish genomes, i.e. coelacanth and spotted gar (Amemiya et al., 2013; Braasch et al., 2016; Howe et al., 2013; Lindblad-Toh et al., 2005; Mouse Genome Sequencing et al., 2002). The criterion for this analysis was set to aminimum of 50% identity over a sequence length of 50 bp (Figure 1). This analysis identified eight human-spotted gar conserved non-coding elements (CNEs) within the first and fifth intron of the human TRPS1 gene. These CNEs are designated as CNE1, CNE2, CNE3, CNE4, CNE5, CNE6 CNE7 and CNE8 (Figure 1; Supporting Information:Figure S1 &Table S4). Among these CNEs, CNE7 overlaps for the most part of its sequence with a previously characterized limb-specific enhancer, hs919 (Visel, Minovitsky, Dubchak, & Pennacchio, 2007). Transgenic mice based functional data for hs919 are available at the Vista Enhancer Browser (https://enhacer.lbl.gov). Each of the identified human-fish TRPS1 CNEs was investigated for the presence of conserved TF binding sites (TFBSs) using MEME Suite and STAMP software (Supporting Information: FigureS2). These in silico analyses revealed binding sites for TFs that are known to be co-expressed with Trps1, such as E2F-1, Ets, Pou3f2, Tbx5,Oct-1, Myogenin (Further details are given in Table S1) (Baldarelli et al., 2021).

Association of identified CNEs with the target gene locus

Evolutionary conserved enhancers are recognized for maintaining synteny with their target genes across distantly related vertebrates (Parveen et al., 2013). Previous studies have demonstrated that enhancers can influence target gene promoters over long genomic distances and may even reside within introns of other genes (Lettice et al., 2003). Furthermore, the expression of a typical vertebrate developmental gene is controlled by multiple enhancers, and in some cases, a single enhancer can upregulate multiple genes (Parveen et al., 2013; Pennacchio, Bickmore, Dean, Nobrega, & Bejerano, 2013). Consequently, pinpointing precise target gene bodies for enhancers has consistently posed a challenging task (Visel, Rubin, & Pennacchio, 2009). To associate predicted CNEs with their target genes, we employed a combination of comparative synteny analysis, investigation of endogenous expression patterns, and analysis of Hi-C (chromosome conformation capture-based technology) datasets (Figure 2A and B; Supporting Information: Table S2, Figure S3 and S4). CNEs carrying the TRPS1 gene, along with flanking intervals (2 Mb upstream and 3 Mb downstream), were compared with orthologous loci in mouse and zebrafish (Figure 2A). This comparative analysis revealed the conserved syntenic association of Trps1 and its flanking genes Eif3h and Csmd3 in all three animals analyzed (Figure 2A). MGI based endogenous expression patterns investigation revealed the co-expression ofTrps1, Eif3h and Csmd3 in heart, limbs, and pharyngeal arches (Supporting Information: Table S2). These comparative analyses suggest any of these three genes as potential targets for the identified set of Trps1 intronic CNEs (Figure2A). To further evaluate the physical regulatory connection between identified CNEs and their precise target promoter, we used Hi-C data for human embryonic stem cells (H1-ESC), K562 cells and GM12878 cells from the 3D genome browser (http://3dgenome.fsm.northwestern.edu/view.php) (Figure 2B and Figure S3 and S4A, B)(Y. Wang et al., 2018).These cell lines are reported to express TRPS1 transcription factor (Joung et al., 2023; Lombardi et al., 2022). The Hi-C data for H1-ESC and k562 cells revealed that the TRPS1 gene is embedded in a domain with enhanced interactions compared to surrounding chromatin regions (Figure 2B; Supporting Information: Figure S4B). This domain is further positioned in a larger TAD with several long-range interactions (Supporting Information: Figure S3 & Figure S4B). In GM12878 cells, the TRPS1 locus lies separately in a large TAD with no other protein coding genes (Figure S4A). Additionally, we also investigated the topology of the Trps1 locus in zebrafish embryos and mouse embryonic stem cells (mESC) to further elucidate its spatial contacts in different contexts across species (Zebrafish: GSM4724554 and mESC: 4DNFI4OUMWZ8). The Hi-C data for zebrafish embryos shows that the TRPS1 gene body has some long-range physical contacts in addition to interactions with its own promotor (Supporting Information: Figure S4C). On the other hand, mESC Hi-C data revealed that the region encompassing the TRPS1 gene is embedded in an isolated TAD with higher within TAD interactions (Supporting Information: Figure S4D). The flanking genes EIF3H and CSMD3 appear to lie outside of this TAD structure (Supporting Information: Figure S4D). Considering the numerous Hi-C data points collected from diverse species and cell types, we can infer that the conserved non-coding elements (CNEs) identified within the intronic interval of the Trps1 gene are linked to the regulation of its promoter.

Transgenic zebrafish based in vivo analysis of TRPS1 associated CNEs

The cis-regulatory potential of the identified human-spotted gar CNEs associated with Trps1 was assessed through the tol2 transposon assay, as developed by Fisher et al., 2006 for zebrafish transgenesis (Fisher et al., 2006). All eight CNEs (Supporting Information: Table S4) were cloned in an expression vector containing the tol2 transposon from medaka, the c-fos mouse minimal promoter, and EGFP (enhanced green fluorescent protein). To boost transposon activity, the recombinant reporter constructs containing CNEs were injected along with transposase mRNA into zebrafish embryos at the one cell stage (Ali et al., 2021). The injected embryos were monitored for the reporter gene expression at various developmental time points up to 5 days post-fertilization (Figure 3A) (Minhas et al., 2015).

Figure 3: Functional assessment of CNEs in TRPS1-related domains.

(A) The top panel outlines the strategy for evaluating the functions of human CNEs in zebrafish. (B). Six TRPS1-associated CNEs exhibit redundant enhancer activity in zebrafish pectoral fins, marked by white dotted lines, at various developmental stages. CNE1 at 48 hpf, CNE3 and CNE4, at 96 hpf,demonstrating enhancer expression in pectoral fins. CNE5, CNE6 and CNE7 at 96 hpf, 120 hpf and 48 hpf respectively also display reporter gene expression in pectoral fins. Rectangular boxes marked by red color cut lines containing zoomed in images of fins (marked by white color cut lines) are added to each image. (C) CNE1, CNE2, CNE5 and CNE8 at 48 hpf, and CNE7 at 72 hpf, induce robust reporter gene activity in the heart of zebrafish embryos. (D) CNE2, CNE3, CNE4, CNE5 and CNE6 induce strong reporter activity in pharyngeal arches (branchial arches) of zebrafish embryos from 48 hpf onwards. Images are merged bright-field and fluorescent views. hpf: hours post fertilization, pf: pectoral fins, ht: heart, PA: pharyngealarches

Among the identified set of eight human-fish conserved CNEs, six elements (CNE1, CNE3, CNE4, CNE5, CNE6 and CNE7) were found to upregulate reporter expression in the pectoral fin of zebrafish at various stages of development (Figure3B). For instance, both CNE1-or CNE3-induced pectoral fin-specific expression was detected at 48 hpf (44% and 46% of transgenic embryos, respectively) (Figure 3B; Supporting Information: Table S5 and S6). CNE4-or CNE5-induced reporter expression in the pectoral fin was detected at 96 hpf (50% and 40% of transgenic embryos, respectively) (Figure 3B; Supporting Information: Table S5 and S6), while CNE6-and CNE7-triggered reporter expression in the pectoral fin was detected at 72hpf and 120 hpf, respectively (each in 60 % of transgenic embryos) (Figure 3B; Supporting Information: Table S5 and S6). In the context of fin-specific regulation, comparative in vivo analysis of CNE7/hs919 in mice and zebrafish transgenesis revealed the conserved role of this element in limb/fin development (Figure 4A and B).

Figure 4: Conservation of TRPS1 CNE7 enhancer function across vertebrates.

(A) Blast and conservation map derived from UCSC vertebrate conservation tracks for hs919/CNE7, including 100bp upstream and 500bp downstream flanking intervals. (B)Trps1 intronic CNE7/hs919 exhibits conserved expression in fins/limbs of zebrafish/mouse. CNE7 induced reporter gene expression in zebrafish pectoral fins (at 48 hpf and 72 hpf), while hs919 (from vista enhancer browser) induced reporter expression was detected in mouse limb buds (E11.5). pf: pectoral fin, n: number of embryos with enhancer activity

In addition to the pectoral fin, redundant regulation of reporter gene expression was also detected in the heart and pharyngeal arch tissues (Figure 3C and D). For instance, CNE2, CNE3, CNE4 or CNE6-induced reporter expression in pharyngeal arches was detected at 48 hpf (in 65%, 52%, 45% and 60% of transgenic embryos, respectively) (Figure 3D;Supporting Information: Table S5 and S6), whereas CNE5-induced reporter gene expression in pharyngeal arches at 72 hpf (in 54% of transgenic embryos) (Figure 3D;Supporting Information: Table S5 and S6).

Five of the CNEs were found to redundantly activate reporter gene expression in heart tissues of developing zebrafish embryos at 48 hpf (CNE1, CNE2 CNE5 and CNE8) and at 72 hpf (CNE7) (Figure 3C;Supporting Information: Table S5 and S6).

Luciferase reporter assay of TRPS1 associated CNEs

Transgenic zebrafish-based analysis uncovered the in vivo regulatory potential for all conserved non-coding elements (CNEs) within the Trps1 gene, derived from human-spotted gar comparison (Figure 3B, C, and D). To further examine their gene regulatory potential, we employed an in vitro luciferase-based reporter assay (Abbasi et al., 2007; Anwar et al., 2015). For this purpose, CNEs were cloned into a modified pGL3 vector upstream of luciferase gene, which contained a TK minimum promoter (Figure 5A). These CNEs-carrying pGL3 constructs were used for the transfection of NIH3T3 and C3H10T1/2 cell lines, both known to respond to Wnt signaling pathway components, including TRPS1 (S. Chen, McLean, Carter, & Leask, 2007; Fischer, Boland, & Tuan, 2002).

Figure 5: Luciferase reporter activities of TRPS1-associated CNEs in NIH3T3 and C3H10T1/2 cell lines.

(A) Schematic drawings illustrating the core region of luciferase constructs. TRPS1-associated CNEs were cloned into a modified pGL3 luciferase vector containing the TK minimum promoter (pGL3-tk-mini). (B, B’) Luciferase activities of reporter constructs with TRPS1-associated CNEs in NIH3T3 cell line, with pGL3-tk-mini without CNE used as a control. (C, C’) Luciferase activities of reporter constructs with TRPS1-associated CNEs in C3H10T1/2 cell lines, with pGL3-tk-mini without CNE used as a control. Luciferase reporter activities are shown as mean +/− standard deviation.

We first tested activities of seven CNEs located in the intron 5 sequence (CNE1-CNE7). In NIH3T3 cells, we observed weak repression of reporter activities for all TRPS1-associated CNEs as compared to the control pGL3-TK mini vector (having no cloned/inserted CNE) (Figure 5B). When using C3H10T1/2 cell lines, strong repression of reporter activity was observed for all the seven TRPS1-associated CNEs (CNE1-CNE7) (Figure 5C). In contrast, CNE8, located in intron 1, showed different regulatory activities. In NIH3T3 cells, CNE8 showed almost no activities on luciferase reporter expression (Figure 5B’). In C3H10T1/2 cells, CNE8 slightly upregulated reporter activity (Figure 5C’). To explore the reason for differential reporter activities in both cell lines, we transfected both cell lines with Top-tk-luciferase vector alone and co-transfected with a plasmid that activates Wnt/ß-catenin signaling, such as constitutively active (CA) ß-catenin or Wnt3a. The Top-tk-luciferase reporter is known to be activated by the LEF/TCF-ß-catenin complex, and its expression reports Wnt/ß-catenin signaling activities (Yamaguchi, Nagatoishi, Tsumoto, & Furukawa, 2020). In the C3H10T1/2 cell line, Wnt3a efficiently activated the reporter compared to the NIH3T3 cells (Supporting Information: Figure S5A & B). This demonstrates that C3H10T1/2 cells can respond to Wnt signaling more efficiently than NIH3T3 cells.

Using constitutively active ß-catenin, Top-tk luciferase was activated both in C3H10T1/2 and NIH3T3 cells. This shows a direct correlation with the Wnt/ß-catenin pathway, as most cells show a response to the Wnt signaling pathway with an increase in the level of ß-catenin (Goentoro & Kirschner, 2009). In both cell lines Top-tk luciferase show differential stimulation in response to Wnt3a and ß-catenin (Supporting Information: Figure S5A & B). Each cell type is known to respond to a specific Wnt component based on the set of receptors it expresses (van Amerongen, Mikels, & Nusse, 2008). Therefore, it can be inferred that both cell lines have one or more differential set of components that can bind ligands or receptors of the Wnt cascade differently. In Supplementary figure S5 the upregulation of reporter activity by Wnt3a in C3H10T1/2 cells and by ß-catenin in NIH3T3 cells hints at the responsiveness of both cell lines in the context of Wnt signaling.

Expression profile of trps1 gene in developing zebrafish embryos

To the best of our knowledge, the endogenous expression of trps1 in zebrafish embryos has not been previously reported (Bradford et al., 2022). Therefore, we aimed to assess the in situ hybridization-based endogenous expression of the trps1 gene in zebrafish embryos at various developmental stages, namely 1 hpf (hours post-fertilization), 3 hpf, 12 hpf, 24 hpf, 48 hpf, 60 hpf, 96 hpf and 120 hpf (Figure 6). The results unveiled widespread endogenous expression of trps1 during embryonic stages such as 1 hpf, 3 hpf and 12 hpf (Figure 6A, B and C), while during later stages, trps1 mRNA expression was predominantly detected in the embryonic head and tail regions (Figure 6E, F, G, H, I and J). For instance, at 1 hpf, 3 hpf, and 12 hpf (5 somite stage), ubiquitous trps1 expression was observed in the blastomere and anterior-posterior ends of the embryo (Figure 6A, B and C). By 48hpf (long-pec stage), the expression was mostly confined to the head and tail region (Figure 6E). Strong expression in the heart was also observed at the long-pec stage and pec-fin stage (Figure 6F and H). At pec-fin stage (60hpf), several tissues in the head region (eyes, midbrain), the heart and the pectoral fin exhibited trps1 expression (Figure 6G and H). Trps1 expression in the tail region was observed until 60 hpf (Figure 6G). At 96 hpf, the trps1 transcript was detected ubiquitously in the skull region of the zebrafish embryo (Figure 6I). The expression in pectoral fins, which commenced at 48hpf, intensified at 96 hpf (Figure 6I). Additionally, at 120 hpf, certain areas of the head, including forebrain, midbrain and otic vesicles, also displayed strong trps1 expression (Figure 6J).

Discussion

Comparative genomics across vertebrate species has identified numerous anciently conserved genomic intervals in the human (Abbasi et al., 2007; Antonellis et al., 2008; Ritter et al., 2010; Shin et al., 2005). While some of these conserved genomic regions encompass protein-coding exonic regions or non-coding RNAs, a substantial portion does not give rise to functional transcripts and is commonly referred to as conserved non-coding elements (CNEs) (Dermitzakis, Reymond, & Antonarakis, 2005; Turner & Cox, 2014; Vavouri, Walter, Gilks, Lehner, & Elgar, 2007; Woolfe et al., 2005). In recent decades, genetic and functional screening of CNEs has unveiled their roles as enhancers or cis-acting regulatory elements (Ali et al., 2021; Pennacchio et al., 2006; Shin et al., 2005; Venkatesh et al., 2006). Notably, genomic variations in CNEs have been linked to mammalian disease phenotypes (Lettice et al., 2003; Turner & Cox, 2014).

The transcription factor Trps1 plays a crucial role in the development of craniofacial structures and limb joints (Malik et al., 2001; Suemoto et al., 2007). In humans, point mutations in TRPS1 are widely associated with skeletal and craniofacial malformations (Momeni et al., 2000). The phenotypes of Trps1-deficient mice closely resemble those of human TRPS patients, indicating the conserved roles of this gene during mammalian embryonic development (Malik et al., 2002). Consistent with recognized roles of Trps1 in mammalian disease phenotypes, multiple studies in mice have reported Trps1 expression in the facial region, pharyngeal arches, limb joints, and hair follicles (Malik et al., 2001; Suemoto et al., 2007). Despite substantial knowledge about the disease etiology of Trps1 and its role in embryonic development, the cis-regulatory control for its endogenous expression remains to be elucidated. Identifying cis-acting elements that interact with the Trps1 promoter could enhance our understanding of the transcriptional regulatory mechanisms that ensure Trps1 availability in cells responsive to Wnt signalling (Dean, Larson, & Sartorelli, 2021; Furlong & Levine, 2018; Long, Prescott, & Wysocka, 2016).

In the present study, we conducted a comparative analysis across multiple vertebrate species to identify anciently conserved CNEs associated with human TRPS1 gene (Figure 1; Supporting Information: Table S4) (Amemiya et al., 2013; Braasch et al., 2016; Venkatesh et al., 2006). Our comparative sequence analysis successfully identified eight human-spotted gar conserved non-coding elements within intronic regions (intron 1 and intron 5) of human TRPS1 gene, designated as CNE1–8), (Figure 1). A BLAST-based similarity search did not detect paralogous copies of identified CNEs in the human genome (Supporting Information: Table S1 and Table S4). In silico analysis of these CNEs revealed the presence of binding sites for multiple distinct TFs known for their established roles in embryonic development and co-expression with Trps1 in various tissues and organs (Supporting Information: Table S1 and S4). The TRPS1-associated cis-regulatory potential of these CNEs was further assessed through zebrafish transgenic assay, in vitro assays using cell lines, and whole mount RNA in situ hybridization (Figure 3,5 and 6).

In the present study, we employed the Whole Mount RNA In Situ hybridization (WISH) assay to provide first-ever view of trps1 endogenous expression in developing zebrafish embryos (Bradford et al., 2022). The resulting WISH images offera visual representation of the mRNA expression pattern (Figure 6). Consistent with previously reported endogenous expression of Trps1 in mice (Malik et al., 2001), we observed a highly dynamic and widespread expression of trps1, starting in the early stages of zebrafish development (1 hpf) and extending up to 120 hpf, encompassing various organs and tissues (Figure 6A–J). Until the long pec stage (48 hpf), trps1 expression was detected in the anterior and posterior portions of the zebrafish embryo, along with the gut region, and most tissues in the head, including eyes, branchial arches, the intestine, and the caudal fin of the zebrafish embryo (Figure 6A–E). In line with the known endogenous aspects of mouse Trps1 expression, in zebrafish at 60 hpf and beyond, trps1 expression was observed in the anterior-portion of zebrafish embryos, mainly confined to the craniofacial region (includingpharyngeal arches) and pectoral fins (Figure 6G–J) (Malik et al., 2001; Malik et al., 2002; Momeni et al., 2000).

Skeletal abnormalities observed in human TRPSI patients encompass limb defects,including cone-shaped epiphysis in phalanges and brachydactyly (Karaca et al., 2019; Momeni et al., 2000). In the context of mouse development, Trps1 mRNA localize to mesenchymal cells, expressing in various tissue domains, notably limb joints and phalanges (Malik et al., 2001). Northern blot and mRNA in situ hybridization analyses have elucidated heightened levels of Trps1 mRNA in the prospective phalanges of developing mouse limbs (Malik et al., 2001). In line with this, we demonstrated robust expression of trps1 in zebrafish pectoral fins from 48hpf through 96 hpf, followed by subsequent decrease in expression (Figure 6F–J). Correspondingly, zebrafish transgenic analysis uncovered the pectoral fin-specific regulatory potential for multiple independent CNEs (CNE1, CNE3, CNE4, CNE5, CNE6 and CNE7) across various time points spanning 48 hpf to 120 hpf (Figure 3B; Supporting Information: Table S5 and S6). The concurrent expression of multiple elements (CNE1, CNE3, CNE4, CNE5, CNE6 and CNE7) within the target domain, the pectoral fin, suggest complex roles for these cis-acting enhancers in limb/fin morphogenesis. Analogously, key developmental regulators such as Shh, Gli3, Gli2, Gdf5, and Shox2 have been reported to be regulated by multiple independent enhancers for limb/fin-specific expression (Abbasi et al., 2010; Ali et al., 2021; H. Chen et al., 2016; Letelier et al., 2018; Minhas et al., 2015; Mouri et al., 2018; Osterwalder et al., 2018; Vokes, Ji, Wong, & McMahon, 2008). Among these fin related enhancers, CNE7 corresponds to hs919 from the Vista Enhancer Browser (https://enhancer.lbl.gov), which is a limb-specific enhancer exhibiting robust activities, driving the expression of a reporter gene in mouse limb buds at E11.5 (Visel et al., 2007). Consistent with transgenic mice-based data for hs919, this genomic interval (CNE7) exhibited reporter activities in the zebrafish pectoral fin at 48 hpf and 72 hpf (Figure 3B, 4B).The Conserved regulatory activity of CNE7 and hs919 in the zebrafish fin and mouse limbs, respectively, suggests that Trsp1-associated CNE-enhancers are not only conserved in their sequence, but also in their tissue-specificity across diverse vertebrate species (Figure 3B and 4B)(Visel et al., 2007).

Significant expression of Trps1 has been documented in the developing mouse heart tissues at E11.5 (Yokoyama et al., 2009). Specifically, Trps1 expression spans various domains within the heart, encompassing the atrium, pulmonary trunk, aortic sinus, tricuspid, and mitral valves from E11.5 to E16.5 (Nomir et al., 2016). In line with the previously reported heart-specific endogenous expression of mouse Trps1, we observed that CNE1, CNE2, CNE5 and CNE7 were capable ofup-regulting reporter expression in the heart of developing zebrafish embryos from 48 hpf onward (Figure 3C; Supporting Information: Table S5 and S6).

For the proper development of craniofacial skeleton, coordinated interactions among multiple tissues and signaling pathways in the pharyngeal arches are essential (Birkholz, Olesnicky Killian, George, & Artinger, 2009). The development of pharyngeal arches involves contributions from all three germ layers, and its patterning is regulated by multiple signaling pathways, including the Fibroblast Growth factors, Hedgehog, Bone Morphogentic proteins, and Wnt pathways (Clouthier & Schilling, 2004; Crump, Maves, Lawson, Weinstein, & Kimmel, 2004; Eberhart, Swartz, Crump, & Kimmel, 2006; Graham & Smith, 2001; Schilling & Kimmel, 1997). Strong expression of Trps1 has been observed in the pharyngeal arches of mouse embryos at E10.5 and E11.5 (Yokoyama et al., 2009). In the present study, in situ hybridization-based experiments revealed trps1 expression in the pharyngeal arches of zebrafish embryos at 48 hpf and 60 hpf (Figure 6F and H). Corresponding with the endogenous expression data of Trps1 in mice and zebrafish, five identified elements (CNE2, CNE3, CNE4, CNE5 and CNE6) demonstrated the ability to upregulate reporter gene expression in the developing pharyngeal arches of zebrafish embryos at 48 hpf (CNE5 at 72 hpf) (Figure D; Supporting Information: Table S5 and S6). The presence of these multiple enhancers suggests a potential combinatorial regulation of trps1 activity during craniofacial morphogenesis (Schilling & Kimmel, 1997).

In the developing mouse embryo, Trps1 functions to suppress Wnt signaling by activating multiple Wnt inhibitors (Fantauzzo & Christiano, 2012). Additionally, Trps1 exerts transcriptional repression on Sox9 and Stat3 by directly binding to their promotors during mouse embryonic development (Fantauzzo, Kurban, Levy, & Christiano, 2012; Suemoto et al., 2007). Consistent with the documented transcriptional repression roles of Trps1, the identified CNEs (CNE1, CNE2, CNE3, CNE4, CNE5, CNE6 and CNE7) that activated reporter gene expression in various developmental domains of zebrafish, acted to repress transcription in in vitro assays employing NIH3T3 and C3H10T1/2cell lines (Figure 5B and C). Interestingly, contrary to all other identified CNEs, only CNE8 exhibited transcriptional activity both in zebrafish transgenesis assay, and luciferase assays in C3H10T1/2 cells (Figure 5C). Taken together, these in vitro and in vivo based data highlights that the identified Trps1-associated CNEs may function in a context-dependent manner within cells, potentially through interactions with a differential set of trans-acting biomolecules(Abbasi et al., 2007; Anwar et al., 2015; Hersh & Carroll, 2005).

Conclusion

In summary, our findings offer initial glimpse into the endogenous facets of trps1 expression in developing zebrafish embryos and unveil a highly intricate cis-regulatory repertoire governing the spatial and temporal expression of this pivotal developmental gene. Our functional data underscore that the identified set of Trps1-associated CNE-enhancers operate in a combinatorial and context-dependent manner, exerting regulatory control over various facets of vertebrate embryogenesis, including craniofacial and limb/fin development. Comparative in vivo analysis across mouse and zebrafish demonstrate the conserved roles of these enhancers in mammals and fish. Consequently, these identified cis-regulatory modules emerge as promising candidates for discerning disease-causing non-coding mutations within the human genome.

Highlights.

Trps1 intronic region contains eight conserved cis regulatory elements in vertebrates. In vivo, these elements exhibit enhancer activity in various domains, including pectoral fins, pharyngeal arches and the heart. In cell lines-based in vitro assays, these elements act as repressors with cellular context-dependent activity.