microRNA-1 regulates sea urchin skeletogenesis by directly targeting skeletogenic genes and modulating components of signaling pathways

Abstract

microRNAs are evolutionarily conserved non-coding RNAs that direct post-transcriptional regulation of target transcripts. In vertebrates, microRNA-1 (miR-1) is expressed in muscle and has been found to play critical regulatory roles in vertebrate angiogenesis, a process that has been proposed to be analogous to sea urchin skeletogenesis. Results indicate that both miR-1 inhibitor and miR-1 mimic-injected larvae have significantly less F-actin enriched circumpharyngeal muscle fibers and fewer gut contractions. In addition, miR-1 regulates the positioning of skeletogenic primary mesenchyme cells (PMCs) and skeletogenesis of the sea urchin embryo. Interestingly, the gain-of-function of miR-1 leads to more severe PMC patterning and skeletal branching defects than its loss-of-function. The results suggest that miR-1 directly suppresses Ets1/2, Tbr, and VegfR7 of the skeletogenic gene regulatory network, and Nodal, and Wnt1 signaling components. This study identifies potential targets of miR-1 that impacts skeletogenesis and muscle formation and contributes to a deeper understanding of miR-1’s function during development.

Graphical Abstract

Introduction

microRNA-1 (miR-1) is among the most evolutionarily conserved microRNAs (Nguyen and Frasch, 2006). miR-1 is classified as a myomiR because of its enriched expression in the muscle of vertebrates and its important role in myogenesis, angiogenesis, and vascularization (Mansfield et al., 2004; McCarthy, 2011; Sokol and Ambros, 2005; Wienholds et al., 2005; Zhao et al., 2005). In mice, miR-1 loss-of-function and gain-of-function led to aberrant heart morphogenesis, myogenic differentiation, and cell proliferation (Chen et al., 2006; Zhao et al., 2007; Zhao et al., 2005). It was proposed that vertebrate miR-1 finetunes the levels of transcripts encoding proteins that are essential for heart function, rather than suppressing non-muscle genes in other tissues (Mishima et al., 2007). Extensive studies have demonstrated miR-1’s role in blocking cardiomyocyte proliferation and skeletal muscle differentiation in both vertebrates and invertebrates, functioning in a tissue-specific way (Nguyen and Frasch, 2006; Sokol and Ambros, 2005; Zhao et al., 2007; Zhao et al., 2005). In this study, we use the purple sea urchin embryo, a deuterostomic invertebrate, to perform both loss-of-function and gain-of-function studies of miR-1 to further understand its role in developmental gene regulatory networks (GRNs).

The developmental processes of the sea urchin and humans are remarkably similar at the cellular and molecular level (Adonin et al., 2020). Similar to vertebrates, they utilize highly conserved signaling pathways such as the canonical Wnt (cWnt)/β-catenin signaling for anterior-posterior (AP) axis formation (Kiecker and Niehrs, 2001; Kimura-Yoshida et al., 2005; Logan et al., 1999; Wikramanayake et al., 1998) and BMP signaling for specification of the dorsal-ventral (DV) secondary body axis (Dal-Pra et al., 2006; De Robertis, 2006; Duboc et al., 2004; Floc’hlay et al., 2021; Khokha et al., 2005; Lapraz et al., 2009; Xu et al., 2014). Immediately after fertilization, maternal inputs, zygotic transcription, and signaling mechanisms help define distinct GRNs (Logan et al., 1999; Revilla-i-Domingo et al., 2007; Sherwood and McClay, 2001; Sweet et al., 2002; Wikramanayake et al., 1998). By the mesenchyme blastula stage, germ layer specification has already occurred; during gastrulation, the three germ layers are differentiated by cross-regulation among signaling pathways and GRN interactions (Davidson et al., 2002a; Davidson et al., 2002b; Oliveri et al., 2002; Revilla-i-Domingo et al., 2007; Yuh et al., 2002). We have also shown that post-transcriptional regulation mediated by miRNAs can also regulate skeletogenic primary mesenchyme cells (PMCs), immune cells, endodermally derived gut morphology, and neuronal development (Konrad and Song, 2022; Konrad et al., 2023; Sampilo et al., 2021; Sampilo et al., 2018; Song et al., 2011; Stepicheva and Song, 2015).

Although the body plans and structures of deuterostomes are diverse, the sea urchin embryo and vertebrates utilize conserved factors for analogous structures. For example, sea urchin skeletogenesis is thought to be analogous to vertebrate angiogenesis and vascularization since these processes utilize a common set of transcription factors (TFs) (Ets1/2, Erg, Hex, Tel, and FoxO) and signaling pathways (Vegf, Nodal/Bmp, Notch, and Angiopoetin) (Gildor et al., 2021; Morgulis et al., 2019). In response to the Vegf ligand in the ectoderm, the sea urchin VegfR-expressing PMCs initiate differentiation, patterning, and the formation of the skeletal rudiment (Adomako-Ankomah and Ettensohn, 2013; Duloquin et al., 2007; Morgulis et al., 2021). The migrating PMCs form a syncytium, connected by filopodial membranes between cell bodies where biomineralization enzymes form calcite granules (Bradham et al., 2004; Khor and Ettensohn, 2022; Winter et al., 2021). Similarly, the sea urchin larva utilizes similar TFs as vertebrates (FoxA, GataE, Xlox, Cdx) for gut differentiation (Annunziata et al., 2019; Annunziata and Arnone, 2014; Cole et al., 2009). The tripartite gut is compartmentalized with the cardiac, pyloric, and anal sphincters. For the nervous system, the sea urchin embryo uses orthologous neuronal transcriptional factors as those expressed in the vertebrate forebrain (Six3, ZIC2, Achaete-scute, NKX2.1 and FEZ) (Range and Wei, 2016; Wei et al., 2009). Thus, using the sea urchin as a simple and experimentally tractable organism, we can better understand complex molecular mechanisms that occur in vertebrate systems.

In the sea urchin, we take advantage of their well-characterized GRN to examine the function of miR-1 in early development. Previously, we found that miR-1 is one of the most highly expressed miRNAs in the purple sea urchin embryo (Song et al., 2011). The sea urchin has ~50 annotated miRNAs, which is a relatively small number in contrast to the 519 miRNAs in humans (Fromm et al., 2015; Kadri et al., 2011; Song et al., 2011; Wheeler et al., 2009). The sea urchin embryo contains a single miR-1, making it feasible to use this embryo to provide a deeper understanding of miR-1’s function in development. Here we address the regulatory role of miR-1 in mesodermally-derived tissues of the sea urchin embryo, using loss- and-gain of function perturbations. We discovered that miR-1 regulates circumpharyngeal muscle structures, skeletogenesis, and the positioning of PMCs. Using site-directed mutagenesis and reporter constructs, we identified that miR-1 modulates skeletogenesis by directly suppressing components of the PMC developmental GRN (Ets1/2, Tbr, and VegfR7), and Nodal, Notch, and Wnt1 signaling components. Additionally, the gain-of-function of miR-1 resulted in more severe skeletal branching defects and PMC patterning than its loss-of-function.

Materials and Methods

Animals

Adult purple sea urchin, Strongylocentrotus purpuratus (Sp), were obtained from Point Loma Marine Invertebrate Lab (Lakeside, CA) and Marinus Scientific, LLC (Long Beach, CA). Adult males and females were intracoelomically injected with 0.5 M KCl to obtain sperm and eggs. Filtered natural seawater (FSW) (collected from Indian River Inlet; University of Delaware) or artificial seawater (ASW) made from Instant Ocean^© was used for embryo cultures incubated at 12°C.

Cloning

To test potential miR-1 targets, we cloned transcripts containing miR-1 seed site (ACATTCC) downstream of luciferase constructs. To obtain 3’UTR of target genes, PCR primers or Fragment GENE DNA fragments (Genewiz, South Plainfield, NJ) were designed based on sequence information available from the sea urchin genome (echinobase.org) (Arshinoff et al., 2022) (Table S1). Amplified PCR products of Bmp2/4, Ets1/2, Notch, Nodal, and Tbr 3’UTRs were first cloned into ZeroBlunt vector (Thermo Fisher Scientific, Waltham, MA), and then subcloned into the Renilla luciferase (Rluc) reporter construct. Wildtype (WT) constructs of IgTM and Nodal were commercially synthesized. VegfR7 and Wnt1 3’UTRs were previously cloned (Sampilo et al., 2021; Stepicheva and Song, 2015) (Table S1). Mutations were generated within the miR-1 seed sequences using the QuikChange Lightning or QuikChange Multi Site-Directed Mutagenesis Kit (Agilent Technologies, Santa Clara, California) to disrupt miR-1’s binding and regulatory function (Staton and Giraldez, 2011; Stepicheva and Song, 2015). The predicted miR-1 seed sites (canonical site: 5’ ACATTCC 3’) within the 3’UTRs of Ets1/2, Dri, Tbr, VegfR7, Notch, Nodal, Bmp2/4, Wnt1, and IgTM were modified at the third and fifth base pair (Remsburg et al., 2019; Staton and Giraldez, 2011) (Table S1). We have previously demonstrated that truncated miRNA seed sequence differing in one nucleotide at the 5’ end is sufficient in miRNA-mRNA target recognition and function (Sampilo et al., 2021). Nodal and Wnt1 contain truncated miR-1 sites (6 out of 7 bps), and Bmp2/4 and IgTM contain mismatched miR-1 seed sites which differ in one nucleotide within the seed sequence (Table S1). Only two of the three potential miR-1 binding sites within Nodal 3’UTR were mutated due to sequence complexity. Firefly luciferase was used as a normalization control in the dual luciferase assay as previously described (Stepicheva and Song, 2015). Each of the construct sequences were verified by DNA sequencing (Genewiz, South Plainfield, NJ). Luciferase constructs containing Ets1/2, VegfR7, Nodal, and IgTM 3’UTRs were linearized with EcoRI, while Tbr, Bmp2/4, and Wnt1 3’UTRs were linearized with NotI. The luciferase constructs and Firefly luciferase mRNA were in vitro transcribed using the Sp6 mMessage machine kit (Ambion Inc, Austin, Texas). In vitro transcribed mRNAs were purified using Macherey-Nagel Nucleospin^® RNA Clean-up kit (Macherey-Nagel, Bethlehem, PA) prior to injections.

Microinjections

Microinjections were performed as previously described (Cheers and Ettensohn, 2004; Stepicheva and Song, 2014) with modifications. All injection solutions were prepared in a 2.5 μl solution consisting of 20% glycerol and 0.4 μg/μl of 10,000 MW neutral non-fixable Texas Red dextran (Thermo Fisher Scientific, Waltham, MA). Approximately 1–2 picoliter was injected into each newly fertilized egg based on the size of the injection bolus at about one-fifth of the egg diameter. miR-1 miRCURY LNA miRNA Power Inhibitor and miRCURY LNA miRNA mimic were obtained from Qiagen (Germantown, MD). miR-1 inhibitor (Hsa-miR-1-3p, ID# YI04100840; 5’- ACATACTTCTTTACATTCCA −3’) and miR-1 miRCURY LNA miRNA mimic (ID#YM00472818; 5’ UGGAAUGUAAAGAAGUAUGUAU 3’) were used at 10 μM, 30 μM and 40 μM concentrations. miRCURY LNA inhibitors are single-stranded antisense oligonucleotides with high specificity to their target miRNA (Davis et al., 2006; Orom et al., 2006; Roberts et al., 2006). Control embryos were injected with dextran with or without Cel-miR-39-3p LNA mimic (ID#YM00479902; 5’ UCACCGGGUGUAAAUCAGCUUG 3’) (not present in the sea urchin). miR-1 inhibitor and miR-1 mimic were co-injected at a 1:1 molar ratio (40 μM miR-1 inhibitor + 40 μM miR-1 mimic) to test the specificity of miR-1 inhibitor.

Dual-luciferase quantification

The injection solutions for the dual-luciferase assay contained 20% sterile glycerol, 0.4 μg/μl 10,000 MW Texas Red lysine-charged dextran, 100–200 ng/μl Firefly mRNA, and 100 ng/μl Rluc mRNA (Ets1/2, Dri, Tbr, VegfR7, Notch, Nodal, Bmp2/4, Wnt1, and IgTM). 20–50 embryos were collected at the mesenchyme blastula stage (24 hpf). Dual luciferase assays were performed using the Promega Dual-Luciferase Reporter (DLR^™) Assay Systems with the Promega GloMax 20/20 Luminometry System (Promega, Madision, WI) (Sampilo et al., 2021; Sampilo et al., 2018; Stepicheva and Song, 2015). The Rluc values were normalized to the Firefly signal to account for microinjection volume differences. Rluc data with mutated miR-1 seed sites were normalized to the Rluc with wildtype (WT) 3’UTR constructs. P-value was analyzed using Student’s t-test. All error bars represent Standard Error (SEM).

Immunofluorescence

Gastrulae and larvae were fixed in 4% paraformaldehyde (PFA) (20% stock; EMS, Hatfield, PA) in FSW overnight at 4°C. Three PBS-Tween (0.05% Tween-20 in 1X PBS) were performed, followed by 1 h block with 4% sheep serum (MilliporeSigma, St. Louis, MO). 1D5 antibody was used at 1:50 to visualize PMCs (McClay et al., 1983) and diluted in PBS-Tween with 4% sheep serum. Embryos were incubated overnight to 3 days at 4°C and washed 3 times with PBS-Tween, followed by goat-anti mouse secondary antibody (Thermo Fisher Scientific, Waltham, MA) at 1:300 for 1 h at RT. Embryos were then washed 3 times with PBS-Tween and mounted on slides for confocal imaging. For visualization of DNA, embryos were counterstained with Hoechst dye (Lonza, Walkersville, MD), DAPI (Thermo Fisher Scientific, Waltham, MA), or VECTASHIELD^® Antifade Mounting Medium with DAPI (Vector Laboratories, Burlingame, CA).

Whole mount and fluorescent in situ hybridization (WMISH and FISH)

The steps performed for fluorescence RNA in situ hybridization (FISH) are described previously with modifications. The Hsa-miR-1-3p (ID# YD00619868; 5’ UGGAAUGUAAAGAAGUAUGUAU 3’) miRCURY LNA detection probe (Qiagen, Germantown, MD) was used to visualize sea urchin miR-1 (at 0.5 ng/μl). The scramble-miR LNA negative control (ID# YD00699004; 5’ GTGTAACACGTCTATACGCCCA 3’) detection probe was used as a negative control. Probes were incubated with embryos in hybridization buffer at 50°C for 5–7 days as previously described with modifications (Konrad and Song, 2022; Sethi et al., 2014; Stepicheva and Song, 2015).

Partial coding sequences of Bmp2/4, Nodal, Not1, Vegf3, and Wnt1 were cloned into ZeroBlunt vector to generate RNA probes (Thermo Fisher Scientific, Waltham, MA) Constructs were linearized using FastDigest^™ (Thermo Fisher Scientific, Waltham, MA) and in vitro transcribed with DIG RNA Labeling Kit (Millipore Sigma, St. Louis, MO) (Table S2). Vegf3 and Wnt1 RNA probes were previously cloned (Sampilo et al., 2021; Stepicheva and Song, 2015). WMISH was conducted according to previous publications (Arenas-Mena et al., 2000; Sampilo et al., 2021). Probes were used at 1 ng/μl and incubated at 50°C for 5–7 days.

Imaging and Phenotyping

Representative images were taken with Zeiss LSM 880 scanning confocal microscope using Zen software or ZEISS Observer Z1 using AxioVision software (Carl Zeiss Microscopy, LLC, White Plains, NY). For videos of gut contractions, live embryos were collected 5 dpf and mounted in FSW onto protamine sulfate (PS)-coated coverslips, creating a positively-charged surface (Stepicheva and Song, 2014). For live behavior examination, control injected, miR-1 inhibitor, and miR-1 mimic-injected embryos were mounted on the same multichambered PS-coated coverslip to avoid variability of environmental conditions and response. To examine the percentage of ingressed PMCs (Fig. S1), we used ZEISS Observer Z1 microscope with the AxioCam305 camera to take Z-stacks of brightfield images of VegfR10 mRNA labeled cells along with DAPI staining to visualize the shape and position of PMCs to characterize their EMT state. To measure dorsoventral connecting rod (DVC) length or PMC migration, ZEISS Observer Z1 microscope was used to take Z-stacks of differential interference contrast (DIC). ZEISS AxioCam105 color camera was used to take in situ images. AxioVision or Zen 3.1 software (Carl Zeiss Microscopy, White Plains, NY) was used to measure the length of DVCs, PMC migration distance, in situ expression domains and to determine the center of gastrulae in vegetal views to measure angles of Vegf3, Wnt1, Nodal, Not1, and Bmp2/4 expression domains, ventral ectodermal (VE), and dorsal ectodermal (DE) domains. N is the total number of embryos examined unless otherwise stated. NS = not significant, *p < 0.05, **p < 0.001, ***p < 0.0001. All error bars represent SEM.

Real-Time, quantitative PCR (qPCR)

To examine the levels of endogenous miR-1 expression within a developing embryo, 200–500 embryos were collected at various developmental stages. Purification of total RNA was done using miRNAeasy Micro Kit (QIAGEN, Germantown, MD). cDNA synthesis of 100 ng total RNA was performed with miRCURY LNA RT Kit (10 μl volume reaction) which adds a 5’ universal tag of a poly(A) tail to mature miRNA templates (QIAGEN, Germantown, MD). cDNA template was diluted 1:10, and miRNA qPCR was performed using miRCURY LNA miRNA PCR Assays (QIAGEN, Germantown, MD) in QuantStudio 6 Real-Time PCR cycler system (Thermo Fisher Scientific, Waltham, MA). Sea urchin miR-200 were used as normalization controls due to its similar expression from the cleavage to the larval stages (Song et al., 2011). Results are shown as fold changes compared to the egg stage using the Ct^−2ΔΔ method as previously described (Konrad and Song, 2022). miRCURY LNA miRNA PCR Primer Mix is against human miR-1 (Hsa-miR-1-3p).

To measure the transcriptional changes of transcripts that encode TFs of the skeletogenic GRN and biomineralization enzymes, we injected zygotes with control, miR-1 inhibitor, and miR-1 mimic. 100 of these blastulae were collected at 24 hpf. Total RNA was extracted by using the Macherey-Nagel Nucleospin^® RNA Clean-up XS kit (Macherey-Nagel, Bethlehem, PA). cDNA was synthesized using iScript cDNA synthesis kit (Bio-Rad, Hercules, CA). qPCR was performed using 2.5 embryo equivalents for each reaction with the Fast SYBR or PowerUp Green PCR Master Mix (Thermo Fisher Scientific, Waltham, MA) in the QuantStudio 6 Real-Time PCR cycler system (Thermo Fisher Scientific, Waltham, MA). Results were normalized to the mRNA expression of ubiquitin and depicted as fold changes compared to control embryos using the Ct^−2ΔΔ method as previously described to analyze the relative changes in gene expression (Stepicheva et al., 2015). Primer sequences were designed using the Primer 3 Program (Rozen and Skaletsky, 2000) and are listed in Table S3. 3–6 biological replicates were conducted. Statistical significance was calculated using two-tailed unpaired Student’s t-tests.

Results

Expression of miR-1 peaks at the gastrula and larval stages with some enriched expression.

To investigate the spatial and temporal expression of miR-1 throughout sea urchin development, we used miRNA real-time quantitative polymerase chain reaction (qPCR) and fluorescent in situ hybridization (FISH). Using qPCR, we observed miR-1 transcript levels decreased 2-fold by 12 hours post fertilization (hpf; early blastula) compared to the egg (Fig. 1A). By 48 hpf (gastrula) and 72 hpf (larvae), miR-1 transcript levels have increased over 10-fold. Using miR-1 FISH, we observed miR-1 to be maternally expressed (Fig. 1B). At the 32-cell stage, the expression of miR-1 is enriched in the perinuclear region. Consistent with miRNA qPCR data, miR-1 expression peaks at the gastrula stage with ubiquitous expression and some enrichment in the mesenchymal cells, gut, and vegetal plate area. At the larval stage, miR-1 expression is still ubiquitous with slight enrichment in the gut and ciliary band (Fig. 1B).

Figure 1. miR-1 is dynamically expressed throughout development.

(A) miR-1 expression is measured by relative miRNA RT-qPCR at various developmental stages. Red dashed lines indicate 2-fold expression difference. Blue circles represent datum points. 3 biological replicates. (B) FISH was used to detect miR-1 at various developmental stages and counterstained with DAPI against DNA (blue). miR-1 is expressed perinuclearly in the 32-cell stage embryo (arrow). miR-1 has increased expression during gastrula stage with enrichment in the mesenchymal cells, gut, and vegetal plate area. miR-1 is enriched in the larval ciliary band and gut. 3 biological replicates. Scale bar = 50 μm.

Perturbation of miR-1 results in circumpharyngeal muscle defects

To examine the function of miR-1 in early development, we injected miR-1 inhibitor for loss-of-function studies or miR-1 mimic for gain-of-function studies. miR-1 miRCURY LNA inhibitor is complementary in sequence with miR-1, so it binds to the endogenous miR-1 to inhibit its function in the embryo. miRCURY LNA mimics are double stranded RNA that are designed to be recognized by the RNA-induced silencing complex (RISC) and consist of three RNA strands, including the specific miRNA and two segmented passenger strands that are rapidly degraded, once the specific miRNA is incorporated into the RISC complex (Owczarzy et al., 2011). Thus, the miR-1 mimic should not bind to the miR-1 LNA inhibitor in co-injection rescue experiments. Since miR-1 is enriched and plays a function in the vertebrate muscle, we examined the effect of miR-1 perturbations on the structure of the circumpharyngeal muscles that surround the larval gut (Fig. 2). As indicated by the miR-1 FISH followed by miR-1 inhibitor injection, miR-1 level in the miR-1 inhibited embryos was greatly reduced compared to the control; however, this is a qualitative assessment as conventional FISH is not quantitative (Fig. 2A). The number of filamentous actin-rich fibers (F-actin) within the muscle fiber ring was significantly decreased and less structured in both miR-1 inhibitor and miR-1 mimic-injected larvae compared to the control. Almost 20% of miR-1 mimic-injected larvae had a complete loss of detectable F-actin (Fig. 2B). The miR-1 inhibitor-induced defects were rescued by co-injection with the miR-1 mimic, indicating that the muscle fiber defects are due to miR-1 perturbation. Interestingly, both miR-1 inhibitor and miR-1 mimic-injected larvae have significantly fewer foregut and hindgut contractions compared to the control, indicating a potential defect in gut functions (Movies 1–3) that may be related to their circumpharyngeal muscle structures.

Figure 2. Perturbation of miR-1 results in circumpharyngeal muscle morphological defects.

(A) Zygotes injected with miR-1 inhibitor or control (40 μM) were subjected to miR-1 or scrambled FISH RNA probes. Endogenous miR-1 is greatly reduced in miR-1 inhibitor-injected embryos. N=20. 3 biological replicates (B) Circumpharyngeal muscles were labeled with phalloidin to detect F-actin. Z-stack of confocal images were taken to count muscle fiber rings. Compared to control, miR-1 perturbed larvae exhibited less or a complete loss of F-actin, in addition to irregular F-actin morphology (boxed areas). miR-1 inhibitor induced defects were rescued by co-injection of miR-1 mimic (40 μM). Colored asterisks correspond to phenotypes in the bar graph. NS = not significant, *p < 0.05, **p < 0.001, ***p < 0.0001. All error bars represent SEM. 4 biological replicates. Student’s t-test. Scale bar = 50 μm. (C) miR-1 perturbed larvae exhibit significantly less contractions of the foregut and the midgut compared to control. *p < 0.05, **p < 0.001. All error bars represent SEM.

Perturbation of miR-1 results in skeletal defects.

Although the sea urchin larval skeleton is not myogenic, skeleton is mesodermally-derived similar to muscles. Since miR-1 regulates mesodermally-derived muscles (Mansfield et al., 2004; McCarthy, 2011; Sokol and Ambros, 2005; Wienholds et al., 2005; Zhao et al., 2005), we also examined other mesodermally-derived cell types, including the PMCs. We found both miR-1 loss-of-function and gain-of-function resulted in a significant delay of PMC ingression during gastrulation (Fig. S1); however, this delay is transient as larval skeleton is formed (Fig. 3). miR-1 perturbed embryos had an average of three PMCs less than control embryo, not likely to make a significant difference in skeletogenesis (Fig. S1).

Figure 3. Perturbation of miR-1 results in the shortening of the DVCs and abnormal skeletal branching.

(A) miR-1 inhibitor was injected at various concentrations, resulting in decreased DVC length in a dose-dependent manner. Shortened DVCs can be partially rescued by co-injection of miR-1 inhibitor with the miR-1 mimic. Arrows indicate the length of DVCs. NS = not significant, ***p < 0.0001. All error bars represent SEM. (B) miR-1 mimic-injected embryos had a dose-dependent decrease of DVCs. N is the total number of embryos examined. NS = not significant, ***p < 0.0001. All error bars represent SEM. Student’s t-test. Scale bar = 50 μm. (C) miR-1 inhibitor-injected larvae exhibited duplicated Body Rod (BR) branching, lack of BR convergence (LBRC), and underdeveloped (UD) larvae. miR-1 mimic-injected embryos were underdeveloped and exhibited severe defects of duplicated branching (arrows) off the Postoral Rods (PORs), Anterolateral Rods (ALRs), and/or BRs in a dose-dependent manner. 3 biological replicates. Scale = 50 μm.

During gastrula stage, we observed that miR-1 inhibitor and miR-1 mimic-injections resulted in decreased length of DVCs as well as all radii of the tri-radiate in a dose-dependent manner (Fig. 3A,B). miR-1 inhibitor-induced skeletal defects were partially rescued by co-injection of a miR-1 mimic, indicating that this defect is specifically induced by miR-1 inhibition (Fig. 3A). In addition, miR-1 inhibitor-injected larvae were underdeveloped, lacked body rod (BR) convergence, and/or occasionally exhibited duplicate BR branching (Fig. 3C). On the other hand, although miR-1 mimic-injected gastrulae have significantly shortened radii of the triratiate, miR-1 mimic-injected larvae had a dose-dependent severity of abnormal and supernumerary skeletal branching off the postoral rods (POR), anterolateral rods (ALR), and BRs (Fig. 3C). We also observed independent skeletal elements developed perpendicular to the larval BRs. Overall, these results indicate that miR-1 plays a critical role in the initial formation and elongation of the skeletal spicules and that miR-1 mimic-injections induced a more severe larval skeletal defect than miR-1 inhibitor-injections.

Perturbation of miR-1 results in PMC patterning defects.

Since we observed skeletal branching defects (Fig. 3), we examined the effect of miR-1 perturbations on the patterning of PMCs, which are cells that make the skeleton (Ettensohn and McClay, 1986). In miR-1 inhibitor-injected gastrulae, we found that while the patterning of PMCs was not greatly affected, PMCs were clustered posteriorly and had less anterior migration compared to the injected control using PMC specific antibody and VegfR10 RNA probe (Fig. 4Ai–ii). This decreased migration in miR-1 inhibitor-injected embryos was partially rescued by co-injection of miR-1 mimic, indicating that this defect is due to inhibition of miR-1. miR-1 inhibitor-injected larvae had occasional PMCs positioned off a skeletal branch (red arrow in Fig. 4Ai) and an occasional duplicated body rod (Fig. 4C), indicating that miR-1 inhibition is likely to mildly affect biomineralization rather than PMC patterning. In miR-1 mimic-injected gastrulae, several PMCs migrated to the animal pole (referred to as scattered PMCs) (arrows in Fig. 4Bi–ii). Interestingly, miR-1 mimic-injected larvae also had several PMCs migrate off skeletal branches with apparent syncytium to the main skeletal body (white arrow in Fig. 4Cii). The defective migration and patterning of PMCs in miR-1 mimic-injected embryos with various duplicated branching suggest that overexpression of miR-1 disrupts skeletal patterning cues and promote ectopic branching formation (Fig. 4Cii).

Figure 4. Perturbation of miR-1 results in ectopic PMC patterning.

Gastrulae were immunolabeled with 1D5 PMC antibody (McClay et al., 1983) or hybridized with the VegfR10 RNA probe. (Ai) PMCs recognized by 1d5 antibody in miR-1 inhibitor-injected gastrulae exhibit less anterior migration (delineated by red arrows) compared to the control that was able to be partially rescued by co-injection of miR-1 mimic. Student’s t-test. 3 biological replicates. (Aii) VegfR10-expressing cells also exhibit clustering and lack of anterior migration compared to control, similar to immunolabeling with the 1D5 antibody. Occasional PMCs positioned off a skeletal branch (red arrow). (Bi) Control and miR-1 mimic-injected gastrulae were immunolabeled with 1d5 antibody for PMCs. miR-1 mimic-injected gastrulae exhibited scattered PMCs (white arrows) that have migrated near animal pole and clustered PMCs at the posterior end of the embryo with no anterior migration (AM) (red arrow). (Bii) miR-1 mimic-injected embryos have VegfR10-expressing cells at the most anterior end of the embryo compared to control, similar to the 1d5 immunolabeling results. (Ci) Larvae were immunolabeled with 1D5 antibody for PMCs and counterstained with DAPI (blue). Compared to control, miR-1 inhibitor-injected larvae exhibited duplicated branching off the BR (arrow). (Cii) miR-1 mimic-injections resulted in severe PMC patterning defects, correlating to abnormal skeletal branching. Scale = 50 μm.

miR-1 directly targets components of skeletogenic GRN and signaling pathways.

To reveal the regulatory molecular mechanism of miR-1, we bioinformatically searched for potential miR-1 seed sites within transcripts that encode regulators of the PMC GRN and key components of signaling pathways (Fig. 5A). We examined the direct suppression of miR-1 of Ets1/2, Dri, Tbr, VegfR7 of the PMC GRN (Kurokawa et al., 1999; Oliveri et al., 2002; Rottinger et al., 2004; Stepicheva and Song, 2015), Nodal, Bmp/4, and Wnt1 of the signaling pathways that may impact PMCs (Duboc et al., 2010; Duboc et al., 2004; Sampilo et al., 2021; Saudemont et al., 2010), and IgTM which has been found to play an important role in skeletal morphogenesis by regulating the number of initial branches that arise within each skeletal primordium (Ettensohn and Dey, 2017). We cloned the 3’UTRs of these genes downstream of Renilla luciferase (Rluc) and compared luciferase levels of Renilla luciferase reporter mRNAs containing wild type or mutated miR-1 binding sites with control Firefly luciferase flanked by Xenopus β-globin UTRs (Stepicheva and Song, 2015). The Rluc readout was normalized to the Firefly luciferase to account for injection differences. miRNA’s binding to the 3’UTR of a target gene would silence its translation, whereas mutating the miRNA binding sites would abolish miRNA binding to 3’UTR of a target transcript, leading to increased translation of luciferase. We determined that miR-1 directly suppresses luciferase reporters bearing the 3’UTRs of Ets1/2, Tbr, VegfR7, Notch, Nodal, and Wnt1. miR-1 may have weak miRNA-mRNA binding affinity with Dri (Fig. 5B).

Figure 5. miR-1 targets multiple components skeletogenic GRN and signaling pathways.

(A) Schematic of simplified PMC GRN is shown. PMCs are specified by the Wnt/β-catenin signaling pathway which activates the transcriptional repressor Pmar1, leading to the activation of skeletogenic TFs Ets1/2, Tbr, Alx1, Dri and Erg, as well as other endodermally and mesodermally-derived TFs (not shown). In the ectoderm (ECT), Wnt/β-catenin signaling activates Eve which in turn activates Wnt1. Nodal in the ventral ectoderm activates BMP signaling to restrict Wnt1 to the posterior-ventral side while Wnt1 prevents Nodal expression in the posterior region. Nodal activates Not1, and Not1 represses Vegf3 in the ventral ectoderm, thus restricting Vegf3 expression to the two lateral ectodermal domains. Validated miR-1 targets are highlighted in orange. (B) Schematics of the luciferase constructs are shown. 3’UTR (indicated by position +1) of select transcripts were cloned downstream of Renilla luciferase vector. The predicted reverse complement of the miR-1 seed sequence (ACATTCC) was altered at nucleotides 2 and 5 using site-directed mutagenesis. Dual luciferase assays with constructs with wild type or mutated miR-1 sites were used to validate miR-1 targets. miR-1 directly suppresses Ets1/2, Tbr, VegfR7, Notch, Nodal and Wnt1. NS=not significant. All error bars represent SEM. Student’s t-test.

miR-1 mimic-injections result in ectopic expression domains of several factors.

To identify the molecular mechanism of how miR-1 regulates skeletogenesis, we examined the spatial expression of key factors involved in PMC patterning, including Vegf3, Wnt1, Nodal, Not1, and Bmp2/4 in gastrulae (Fig. 6). In miR-1 inhibitor-injected gastrulae, the lateral and vegetal expression domains of Vegf3 were significantly decreased compared to control (Fig. 6A). Concurrently, the vegetal expression domains of Nodal and Not1 were expanded, without expression domain change of Bmp2/4 (Fig. 6Bi, iii). In miR-1 mimic-injected gastrulae, the lateral expression of Vegf3 had no expression domain change; however, its vegetal expression domain was significantly expanded (Fig. 6Aii, iii). In contrast to miR-1 inhibitor-injected embryos, Nodal, Not1, and Bmp2/4 expression domains were decreased in the vegetal expression domains of the miR-1 mimic-injected embryos, while the expression domain of Wnt1 was increased compared to control (Fig. 6Bii). In general, miR-1 loss- and gain-of-function resulted in reciprocal changes in Vegf3, Nodal, and Not expression domains.

Figure 6. Perturbation of miR-1 results in ectopic expression domains.

The expression domains of Vegf3, Wnt1, Nodal, Not1 and Bmp2/4 were assessed. (Ai) The lateral expression domain of Vegf3 (red dashed lines) was decreased in miR-1 inhibited gastrulae compared to the control, whereas the lateral expression domain of Vegf3 in miR-1 mimic-injected embryos did not change compared to the injected control. (Aii) miR-1 inhibited gastrulae have decreased ventral Vegf3 expression domain compared to the control. The vegetal expression domain of Vegf3 of miR-1 mimic-injected embryos was significantly expanded compared to the injected control embryos. (Aiii) The graph indicates the ventral expression domain of Vegf3 was measured in injected control, miR-1 inhibitor, and miR-1 mimic-injected gastrulae. (Bi) In miR-1 inhibitor-injected gastrulae, vegetal expression domains of Nodal and Not1 were expanded, while the expression domains of Bmp2/4 and Wnt1 were not altered. (Bii) In miR-1 mimic-injected embryos, Nodal, Not1 and Bmp2/4 vegetal expression domains decreased, while the expression domains of Wnt1 and Vegf3 expanded. (Biii) The graph indicates the expression domains of Wnt1, Nodal, Not1, and Bmp2/4 of injected control, miR-1 inhibitor and miR-1 mimic-injected embryos. N is the number of domains measured. Student’s t-test. 2–3 biological replicates. Scale = 50 μm. SEM is graphed.

miR-1 regulates the expression of key skeletogenic transcripts.

To identify the underlying molecular mechanism that led to PMC patterning and skeletal branching defects, we examined the relative expression levels of transcripts that encode PMC specification and patterning (Ets1/2, Dri, Tbr, Snail, Nodal, Bmp2/4, Not1, Vegf3, Wnt1, FgfA and CDC42) (Adomako-Ankomah and Ettensohn, 2013; Duboc et al., 2010; Duloquin et al., 2007; Rottinger et al., 2004; Rottinger et al., 2008; Sepúlveda-Ramírez et al., 2018), biomineralization enzymes (P19, p58A, SM30, SM50) (Adomako-Ankomah and Ettensohn, 2011; Cheers and Ettensohn, 2005; Livingston et al., 2006), PMC-specific cell surface protein Msp130 (Leaf et al., 1987), PMC adhesion protein KirrelL (Ettensohn and Dey, 2017), and markers of dorsal PMCs (Tbx2/3 and GataC) (Duboc et al., 2010) (Fig. 7). In general, miR-1 inhibited blastulae have increased expression for genes involved in PMC GRN (Ets1/2, Dri, Tbr, Snail) and Notch, whereas miR-1 mimic-injected blastulae have decreased expression for these same genes. Specifically, miR-1 inhibited blastulae have significantly increased Tbr mRNA compared to the injected control; significantly decreased GataC and Tbx2/3 mRNAs expressed in dorsal PMCs; significantly decreased p19, SM50, IgTM, Kirrel, and FgfA mRNAs. In contrast, miR-1 mimic-injected blastulae have significantly decreased p58A mRNA, significantly increased Wnt1 mRNA, and significantly decreased Bmp2/4, Not1, Notch, and Cdc42 mRNAs (Fig. 7). Both miR-1 inhibitor and miR-1 mimic-injected embryos have no change or decreased expression of genes involved in biomineralization, suggesting that miR-1 indirectly regulates biomineralization transcripts to impact skeletogenesis.

Figure 7. miR-1 regulates key skeletogenic transcripts.

The expression levels of key transcripts encoding factors important for PMC development were assessed by qPCR at the mesenchyme blastula stage. The dashed red line indicates 2-fold differences. 3–5 biological replicates were conducted. Each replicate contains 100 embryos. Student’s t-test was used. *p≤0.05.

Discussion

miR-1 is a myomiR that has been found to regulate muscle development (Mansfield et al., 2004; McCarthy, 2011; Sokol and Ambros, 2005; Wienholds et al., 2005; Zhao et al., 2005). In this study, we found miR-1 to regulate circumpharyngeal muscles, the proper formation of the larval skeleton, and the patterning of mesodermally-derived PMCs. Our results indicate that miR-1 regulates skeletogenesis, likely via its suppression of components of the PMC GRN (VegfR7, Ets1/2 and Tbr) and Nodal and Wnt1 signaling components. In addition, miR-1 indirectly regulates biomineralization transcripts and Vegf3 to potentially impact spicule formation and PMC patterning, respectively.

miR-1 expression is enriched in multiple tissues of the gastrulae and larvae (Fig. 1). The expression pattern of miR-1 may reveal its regulatory mechanism. In vertebrates, miR-1 is enriched in muscle lineages and may function as a developmental switch when it is specifically expressed in cardiac and skeletal tissues. In invertebrates, miR-1’s expression is broader with more variable functions, including regulation in immunity by suppression of clathrin during phagocytosis in shrimp (Liu et al., 2014), neuromuscular junction in nematodes (Simon et al., 2008), midgut regeneration in fruit fly (Takemura et al., 2021), and sex-determination in oriental fruit fly (Peng et al., 2020). Our results indicate that the sea urchin miR-1 has a broader expression with more diverse functions than vertebrates, including in circumpharyngeal muscle structures and skeletal components.

We observed that miR-1 perturbed larvae have decreased number of F-actin muscle fiber rings (Fig. 2B). The defective gut contractions in miR-1 perturbed larvae may be due to problems with the muscle fibers and/or neural coordination (Fig. 2C). Previous work has shown Delta/Notch signaling to play critical roles in vertebrate myogenesis (Conboy and Rando, 2002; Kopan et al., 1994; Schuster-Gossler et al., 2007), and knockdown of Delta or introducing a dominant negative version of Notch that lacks the Notch intracellular domain in the sea urchin embryos resulted in decreased muscle fibers (Sherwood and McClay, 1999; Sweet et al., 2002). Using site-directed mutagenesis and dual luciferase assays, we determined that miR-1 directly suppresses reporter bearing the 3’UTR of Notch. Even though miR-1 inhibition did not significantly alter transcript level of Notch, miR-1 mimic injection resulted in a significant decrease of Notch mRNA, suggesting that miR-1’s regulation of Notch may be in part via inducing transcript degradation (Fig. 7). Since Notch receptor functions with Delta ligand, changes in Notch mRNA alone is not likely to affect the signaling and thus not likely to explain the muscle phenotypes induced by miR-1 perturbations. In addition, although perturbations of the sonic hedgehog pathway in the sea urchin embryo led to disorganized circumesophageal muscle causing an inability to swallow (Walton et al., 2009), we did not identify potential miR-1 binding sites within Sonic Hedgehog, Smoothened, or Patched transcripts.

Another mesodermally-derived tissue that is regulated by miR-1 is the larval skeleton. Our results indicate that all radii of the triradiate at late gastrula stage are significantly shortened in embryos injected with either miR-1 inhibitor or miR-1 mimic (Fig. 3A,B). 40% of these miR-1 mimic-injected gastrulae embryos have PMCs that migrated all the way to the anterior region of the gastrulae (Figs. 3B, 4B), indicating that these mispatterned PMCs do not seem to contribute to skeletal initiation. The shortened tri-radiates phenotype in the miR-1 perturbed gastrulae seems to be a transient effect, since both miR-1 inhibitor and mimic-injected larvae have elongated skeletal structures and ectopic branching (Fig. 3C).

Of note is that the penetrance of miR-1 mimic seems to be better than the miR-1 inhibitor (Fig. 3). This difference in penetrance of the inhibitor and miR-1 mimic is potentially due to the difference in how they are synthesized, designed and prepared by the manufacturer. miRCURY LNA inhibitor is a RNA complementary to miR-1; however, how many LNA residues are designed into the LNA inhibitor sequence is proprietary. On the other hand, miRCURY LNA mimics are double stranded RNA that are designed to be recognized by the RNA-induced silencing complex (RISC) and consist of the miRNA itself with two passenger strands that are rapidly degraded once the specific miRNA is incorporated into the RISC complex (Owczarzy et al., 2011; Qiagen, 2018). Due to these differences, the penetrance of the miR-1 mimic seems to be better than the miR-1 inhibitor.

In general, results suggest that miR-1 inhibition leads to mild biomineralization defects and miR-1 overexpression leads to perturbation of PMC patterning cues (Figs. 3, 4). Results indicate that p19 and SM50 are significantly decreased in miR-1 inhibited blastulae and p58A is significantly decreased in miR-1 mimic-injected blastulae (Fig. 7). The decreased biomineralization transcripts in miR-1 inhibited and overexpressed blastulae may explain the shortened skeletal length in these gastrulae (Figs. 3 and 7). The result is seemingly contradictory in that miR-1 mimic-injected gastrulae have significantly shortened tri-radiates but have multiple duplicated branching in the larval stage (Fig. 3C). However, we do not know the level of these biomineralization transcripts in gastrula and larval stages. In addition, the change in miR-1’s expression pattern throughout early development suggests that miR-1’s regulation of biomineralization genes may change throughout development (Fig. 1B). There may also be unknown factors that miR-1 suppresses that is involved in providing negative signals to skeletal branching, as miR-1 mimic induces supernumerary branching (Fig. 3C).

To understand how miR-1 may be regulating PMC development, we used site-directed mutagenesis and dual luciferase assays to identify miR-1 target transcripts (Fig. 5). We found miR-1 to directly suppress reporters containing 3’UTRs of Ets1/2, Tbr, and VegfR7 of the PMC GRN, and Notch, Nodal, and Wnt1 of signaling pathways (Fig. 5). We analyzed the spatial expression of factors that may affect PMC patterning (Fig. 6), as well as the level of transcripts that encode proteins that affect skeletogenesis (Fig. 7). Since PMCs are mainly responsible for the formation of the larval skeleton (Ettensohn and McClay, 1986), we examined the patterning of PMCs (Fig. 4). It was striking that PMCs are mispatterned in miR-1 mimic-injected embryos (Fig. 4B). Some of these PMCs in miR-1 mimic-injected larvae appear to be mispatterned to areas where we observed spurious skeletal branches in the larvae (Figs. 4C), indicating a disruption of patterning cues.

For both miR-1 inhibitor and miR-1 mimic-injected blastulae, the percentage of PMCs undergoing EMT is significantly and consistently less than the control (Fig. S1). Since we observe that at a later timepoint PMCs ingress into the blastocoel, this EMT delay is transient. However, even though the delay of PMC ingression is transient (Fig. S1), this delay could potentially affect PMC patterning by disrupting the time and distance-sensitive interaction between VegfR10-expressing PMCs and the Vegf3 ligand expressed in the ectoderm, as Vegf3 expression becomes restricted to the Veg1 ectoderm by 30 hpf (early gastrula) (Li et al., 2014).

In miR-1 inhibited gastrula, Vegf3, Nodal, and Not1 had significant expression domain changes that were reciprocal to that of miR-1 mimic-injected embryos. The loss-of-function of Vegf inhibits skeleton formation (Duloquin et al., 2007). Thus, the decreased expression domain of Vegf in miR-1 inhibited embryos is consistent with overall reduced biomineralization in these gastrulae (Figs. 6A, 7). On the other hand, miR-1 mimic-injected embryos have expanded Vegf that may contribute to the duplicated skeletal element branching observed later in those larvae. Interestingly, Vegf3 transcript level at blastula stage was not altered in both miR-1 inhibited or overexpressed blastulae (Fig. 7), suggesting that miR-1 may regulate factors that restrict the expression domain of Vegf3. In zebrafish, miR-1 has been found to negatively regulate angiogenesis during development by repressing VegfAa (Stahlhut et al., 2012). We bioinformatically identified potential miR-1’s binding site within the sea urchin Vegf3 transcript. Thus, if miR-1 directly regulated Vegf3, it would be likely affecting its protein level, since its transcript was not significantly altered upon miR-1 perturbation (Fig. 7).

We also observed that the expression domain of Nodal is significantly increased in miR-1 inhibited gastrulae and significantly decreased in miR-1-mimic injected gastrulae, further suggesting that miR-1 regulates Nodal (Fig. 6B). Although the level of Nodal transcripts was not greatly altered in miR-1 mimic-injected blastulae assessed with qPCR, we observed a significant >3-fold decrease of Nodal’s known target, Not1 (Fig. 7) (Li et al., 2012; Materna et al., 2013). This result suggests that miR-1 mimic-injected embryos may have decreased Nodal protein, resulting in decreased Not1 transcripts (Fig. 7). This also suggests that miR-1 regulates Nodal at the post-transcriptional level. Prior studies have shown that Nodal activates the expression of itself, Not1, Bmp2/4, and Vegf3 in the ventral ectoderm of the early blastula (Li et al., 2012). Also, Not1 knockdown led to expanded Vegf3 expression in the ventral ectoderm (Li et al., 2012). Therefore, although our skeletal defects do not phenocopy Nodal’s loss-of-function which resulted in broad perturbance of the skeleton, the level of miR-1 modulation of Nodal may indirectly impact the Vegf3 expression domain in the ventral ectoderm (Duboc et al., 2010; Duboc et al., 2004; Layous et al., 2021; Li et al., 2012; Saudemont et al., 2010).

We also observed an expanded expression domain of Wnt1 in miR-1 mimic-injected embryos. Wnt1 can activate its own transcription and is part of highly cross-regulated positive feedback circuitry (Cui et al., 2014; Sampilo et al., 2021). miR-1 mimic-injected embryos resulted in significant increased levels of Wnt1 mRNA, suggesting that Wnt1 transcript is altered in response to other factors regulated by miR-1. miR-1 mimic-injected embryos resulted in significantly decreased Bmp2/4 (Fig. 7). Since miR-1 does not regulate Bmp2/4, this regulation is likely to be indirect (Fig. 5). The significant expression domain changes of Vegf3, Nodal, and Not1 in miR-1 inhibited embryos do not result in a patterning change of PMCs (Figs. 4A, 6). The changes we observed in the spatial expressions of Vegf3, Nodal, Not1, Bmp2/4, and Wnt1 in miR-1 mimic-injected embryos may all contribute to PMC mispatterning (Figs. 4A, 6). Since we did not conduct a time course study to observe changes in gene expression domains, we cannot rule out if these gene expression domain changes are reflective of a developmental delay. However, all control embryos were also injected, so these expression domain changes are not likely due to injections.

Ets1/2, Tbr, and Alx1 are all key regulators of skeletogenesis (Ettensohn et al., 2003; Fuchikami et al., 2002; Oliveri et al., 2008). Of these, we found miR-1 to suppress reporters containing 3’UTRs of Tbr and Ets1/2 (Fig. 5). Tbr plays an essential role in specification of the skeletogenic mesoderm and formation of the larval skeleton where Tbr knockdown resulted in a complete loss of skeleton (Oliveri et al., 2002; Oliveri et al., 2008). Tbr has also been found to be important for PMC EMT, basement membrane remodeling, and apical constriction of PMCs (Saunders and McClay, 2014). We found miR-1 inhibited blastulae have significantly increased Tbr mRNA compared to the control (Fig. 7). The impact of Tbr overexpression on skeletogenesis in not clear. miR-1’s suppression of Tbr may potentially contribute to the initial significant delay of PMC ingression (Fig. S1). miR-1 perturbation may also affect Tbr levels to impact the level of msp130, which is positively activated by Tbr, and encodes one of the biomineralization enzymes (Cary et al., 2017). miR-1 perturbation did not result in significant changes of Ets1/2 mRNA levels, indicating that miR-1 could regulate Ets1/2 post-transcriptionally and indirectly by regulating factors that control its function. Ets1, similar to Alx1, provides positive inputs into a large fraction of PMC effector genes (Rafiq et al., 2014). For example, miR-1 perturbation may also affect Ets1 to impact the level of SM50, which is activated by Ets1 and encodes one of the biomineralization enzymes (Kurokawa et al., 1999). In general, the effect of inhibition of miR-1 during early blastula/blastula stage, when miR-1 is normally expressed at low level (Fig. 1), on the skeletogenic program may be less consequential than overexpression of miR-1. This may explain the stronger miR-1 overexpression induced skeletal and PMC patterning defects compared to miR-1 inhibition (Figs. 3, 4).

Interestingly, miR-1 may repress a negative regulator of Kirrel, since we observed occasional mispatterned PMCs and shorted tri-radiates in miR-1 inhibited gastrulae (Fig. 4A) (Ettensohn and Dey, 2017). Similarly, miR-1 may repress a negative regulator of IgTM, since miR-1 mimic-injected larvae had ectopic skeletal branching, reminiscent of IgTM knockdown gastrulae with multiple branching coming out of the tri-radiate rudiment (Fig. 3C, 4B) (Ettensohn and Dey, 2017). Since we did not find miR-1 to directly regulate IgTM (Fig. 6B), miR-1’s regulation of IgTM is likely to be indirect. miR-1 inhibition resulted in significant decrease of Fgfa transcript (Fig. 7), indicating that miR-1 likely suppresses a negative regulator of Fgfa. Previously it was found that a knockdown of Fgfa resulted in PMC mispatterning and a loss of skeleton (Rottinger et al., 2008). This result suggests that decreased Fgfa in miR-1 inhibited gastrulae may contribute to the initial shortened tri-radiates but did not impact PMC patterning (Figs. 3A, 4A). Of note is that miRNAs functions by repressing translation of its targets and/or recruiting deadenylase complexes to degrade its target transcript (Lee et al., 1993; Lim et al., 2005; Wightman et al., 1993). Thus, miR-1 may regulate these transcripts at the level of post-transcriptional control and not impacting their transcript levels. Without assaying for their protein levels, we cannot determine how miR-1 regulates these transcripts.

It is interesting to note that miR-1 loss- and gain-of-function lead to similar phenotypes, such as muscle fiber defects (Fig. 2), EMT delay (Fig. S1), and skeletal defects (Fig. 3). We do not know the exact molecular regulatory mechanism of miR-1. However, we propose that such a regulatory mechanism needs to consider the expression of miR-1 (Fig. 1) and miR-1’s regulation of multiple targets that impact the same protein or pathway. For example, to explain why miR-1 loss- and gain-of-function lead to similar skeletal defects in the sea urchin, we propose that miR-1 regulates Ets1/2 and an unidentified negative regulator of Ets1/2. The expression and regulation of Ets1 is complex. The Ets1 mRNA and protein are maternally present and zygotically expressed during late cleavage stage; its expression is restricted to the skeletogenic lineage until late mesenchyme blastula stage (Kurokawa et al., 1999; Rizzo et al., 2006; Yajima et al., 2010). The function of Ets1 requires phosphorylation ERK for PMC specification (Fernandez-Serra et al., 2004; Rottinger et al., 2004). We propose that miR-1 post-transcriptionally regulates Ets1/2, as well as a negative regulator that impacts the function of Ets1/2, such as its phosphatase. The sea urchin genome contains five annotated serine/threonine phosphatases, all of which contain potential miR-1 binding sites. miR-1 is expressed at relatively low levels during early blastula/blastula stages, when Ets1/2 is zygotically expressed and becomes localized to the skeletogenic lineage (Kurokawa et al., 1999; Rizzo et al., 2006; Yajima et al., 2010). Since Ets1/2 is regulated by phosphorylation and becomes functional during the blastula stage, the impact of miR-1 inhibition would result in increased translated Ets1/2 and this unknown negative regulator of Ets1/2. Since the normal expression of miR-1 at the early blastula stage is low, inhibition of miR-1 would not be expected to have a dramatic effect on Ets1/2 and its negative regulator. In this case, phosphorylation of Ets1/2 would result in overall increased functional Ets1, leading to some enhanced expression of PMC effector genes and skeletogenesis. In the case of miR-1 overexpression during the early blastula/blastula stages, when miR-1 is usually at low levels, translation of Ets1/2 and its negative regulator would be decreased. However, the translated Ets1 would be mostly phosphorylated and functional, leading to enhanced PMC effector gene expression and skeletogenesis. Thus, in this proposed mechanism, through miR-1’s regulation of Ets1/2 and its phosphatase, the loss- and gain-of-function of miR-1 would result in similar phenotypes.

An alternative regulatory mechanism could be that miR-1 has variable level of suppression on multiple targets that encode factors that influence muscle, PMC EMT, and skeletogenesis. For example, in the case of skeletogenesis, miR-1’s suppression of skeletogenic promoting factors may be less than the level of miR-1’s suppression of skeletogenic repressive factors during early blastula/blastula stages. Thus, the net effect would be that miR-1 inhibition is less impactful during a time when its normal expression is low, so that the overall skeletogenic program would not be greatly affected. Upon miR-1 overexpression, the suppression of skeletogenic repressive factors would be greater than the suppression of skeletogenic promoting factors. Thus, during early blastula/blastula stages, excess miR-1 would greatly suppress the skeletogenic repressive factors more than the skeletogenic promoting factors.

We propose that in cases when miR-1 targets a TF or regulator of the skeletogenic program, we would observe reciprocal responses. However, if miR-1 directly targets a TF or regulator of the skeletogenic program as well as a negative regulator of this target during the early blastula/blastula stages, then we would observe similar defects upon either miR-1 inhibition or overexpression. These proposed mechanisms are speculative and will need to be tested in future studies.

Previously, we identified that miR-31 in the sea urchin regulates skeletogenesis by directly suppressing Eve and Wnt1 (Sampilo et al., 2021). Depletion of miR-31 resulted in expanded vegetal spatial expression of Vegf3 (Sampilo et al., 2021; Stepicheva and Song, 2015), similar to miR-1 mimic-injections. Here we identified miR-1 to likely directly suppresses genes Ets1/2, Tbr, VegfR7 and Wnt1, of which Ets1/2 and Tbr are downstream of miR-31 targets Pmar1 and Eve (Sampilo et al., 2021; Stepicheva and Song, 2015). Thus, both miR-31 and miR-1 target critical components within the PMC GRN and co-regulate skeletogenesis.

Conclusions

Overall, we identified miR-1 to be broadly expressed with diverse functions in the sea urchin embryo. miR-1 regulates not only mesodermally-derived gut muscle structures, but also mediate skeletal development. This study identifies novel functions of miR-1, by identifying its likely direct targets and revealing miR-1 to regulate various transcription factors of the PMC GRN and signaling components to regulate skeletogenesis of the developing embryo.

Supplementary Material

1. Figure S1. Perturbation of miR-1 induced a delay in PMC ingression.

Control and miR-1 perturbed embryos were collected at various times during PMC ingression. Embryos were subjected to RNA in situ hybridization with VegfR-10 probe followed by counterstaining with DAPI for DNA. VegfR10-positive PMCs were counted from a series of Z-stack images. PMCs undergoing epithelial to mesenchymal transition (EMT) appear as ‘bottle-shaped’, whereas cells that have completed EMT have a ‘round cell shape’ (Anstrom, 1992; Ettensohn, 1999; Fink and McClay, 1985; Katow and Solursh, 1981). The percentage of “ingressed, round” PMCs from each embryo were calculated from the total of VegfR10-expressing PMCs (Sampilo et al., 2021). (A) Using the cell shape as a criterion for EMT, we found miR-1 loss-of-function resulted in a significant delay of PMC ingression during gastrulation. Compared to control, miR-1 inhibited blastulae had fewer PMCs. 40 μM of control and miR-1 inhibitor were injected. (B) We found miR-1 gain-of-function resulted in a significant delay of PMC ingression during gastrulation. Compared to control, miR-1 mimic-injected blastulae had fewer PMCs. 30 μM of mimic control and miR-1 mimic were injected. 3 biological replicates. Cochran-Mantel-Haenszel test was used to assess the percentage of ingressed PMCs. NS = not significant, *p < 0.05, ***p < 0.0001. All error bars represent SEM. Student’s t-test was used to assess the number of PMCs. Scale bar = 50 μm.

2. Movie 1. Gut contractions in control embryos.

Live embryos were collected 5 dpf and recorded. Average of contractions was analyzed within 4 min. Scale bar = 20 μm.

3. Movie 2. Gut contractions in miR-1 inhibitor-injected embryos.

Live embryos were collected 5 dpf and recorded. Average of contractions was analyzed within 4 min. Scale bar = 20 μm.

4. Movie 3. Gut contractions in miR-1 mimic-injected embryos.

Live embryos were collected 5 dpf and recorded. Average of contractions was analyzed within 4 min. Scale bar = 20 μm.

Summary statement:

This study identifies miR-1 to regulate circumpharyngeal muscle structure, skeletal branching, and skeletal cell patterning. Using site directed mutagenesis and reporter constructs, we identified that miR-1 directly suppresses Ets1/2, Tbr, and VegfR7 of the skeletogenic gene regulatory network, and Nodal, Notch, and Wnt1 signaling components.

Highlights.

• Sea urchin miR-1 has broad expression and function in the developing embryo.

• miR-1 regulates gut muscle structures.

• miR-1 regulates skeletogenic cell patterning and skeletal spicules.

• miR-1 is likely to directly suppress skeletogenic gene regulatory components.

Abbreviations:

ALR

anterolateral rod

BR

body rod

BE-DVM

border ectoderm-dorsal ventral margin

cWnt

canonical Wnt

DE

dorsal ectoderm

Dpf

days post fertilization

DVC

dorsoventral connecting rod

EMT

epithelial to mesenchymal transition

GRN

gene regulatory network

Hpf

hours post fertilization

PMCs

primary mesenchyme cells

POR

postoral rod

TF

transcription factor

VE

ventral ectoderm

Vegf

vascular endothelial growth factor