Precise spatiotemporal regulation of the SIX1 homeoprotein is required to coordinate vital tissue development, including myogenesis. Whereas SIX1 is downregulated in most tissues following embryogenesis, it is re-expressed in numerous cancers, including tumors derived from muscle progenitors. Despite crucial roles in development and disease, the upstream regulation of SIX1 expression has remained elusive. Here, we identify the first direct mechanism for Six1 regulation in embryogenesis, through microRNA30a (miR30a)-mediated repression. In zebrafish somites, we show that miR30a and six1a and six1b (hereafter six1a/b) are expressed in an inverse temporal pattern. Overexpression of miR30a leads to a reduction in six1a/b levels, and results in increased apoptosis and altered somite morphology, which phenocopies six1a/b knockdown. Conversely, miR30a inhibition leads to increased Six1 expression and abnormal somite morphology, revealing a role for endogenous miR30a as a muscle-specific miRNA (myomiR). Importantly, restoration of six1a in miR30a-overexpressing embryos restores proper myogenesis. These data demonstrate a new role for miR30a at a key node in the myogenic regulatory gene network through controlling Six1 expression.
Embryonic skeletal trunk muscle in vertebrates arises from progenitor cells in the paraxial mesoderm that are induced to form the primary myotome. These precursor cells differentiate into both fast- and slow-twitch muscle. Early myogenesis is regulated through expression of basic helix-loop-helix (bHLH)-domain-containing myogenic regulatory factors (MRFs), which are controlled by a complex genetic network that includes Sine oculis homeobox (SIX) and Eyes absent (EYA) transcriptional regulators (Bryson-Richardson and Currie, 2008; Heanue et al., 1999). SIX family members are homeodomain-containing proteins that contribute to the development of muscle, and many other tissues, through promoting the transcription of genes involved in cell proliferation, survival, differentiation, migration and invasion (Christensen et al., 2008). SIX1 and SIX4 directly activate expression of several MRFs in the mouse, including myogenic differentiation 1 (Myod1), myogenic factor 5 (Myf5), myogenin (Myog) and Myf6 (Giordani et al., 2007; Grifone et al., 2005; Hinits et al., 2007; Spitz et al., 1998). In parallel, MRF transcription can be regulated by the homeodomain-containing paired box gene 3 (PAX3) during primary myogenesis (Buckingham and Relaix, 2007), the expression of which is controlled by SIX1 and SIX4, along with their co-transcriptional activators EYA1 and EYA2, in a subset of developing muscles (Grifone et al., 2007; Grifone et al., 2005).
Because the SIX family of transcription factors are upstream activators of the myogenic program, their regulation during muscle development is crucial. During early myogenesis, knockdown of mouse and zebrafish Six1 results in severe muscle hypoplasia and a decrease in fast-twitch fibers (Bessarab et al., 2008; Grifone et al., 2005; Laclef et al., 2003a; Nord et al., 2013), demonstrating a role for Six1 in both muscle progenitor activation and the promotion of fast muscle differentiation. However, overexpression of Six1 in zebrafish also prevents fast twitch fiber formation (Nord et al., 2013), indicating that improper Six1 levels, either too high or low, can negatively affect early muscle differentiation. To date, mechanisms controlling Six1 expression during embryogenesis have not been elucidated in any tissue. Because microRNAs (miRs) can tightly regulate protein levels in a developmental context, we examined potential miR-mediated control of Six1.
MiRs are small, non-coding RNAs known to exert essential spatiotemporal gene regulation in a diverse array of developmental and disease programs, including myogenesis (Chen et al., 2009; Sayed and Abdellatif, 2011; Yekta et al., 2008). MiRs function by base-pairing to a ‘seed’ sequence located in target mRNAs, mediating mRNA degradation or translational repression (Bartel, 2009; Filipowicz et al., 2008). In both mice and zebrafish, recent studies aimed at eliminating the function of an enzyme essential for general miR-processing, Dicer, have demonstrated important roles for miRs in embryonic myogenesis, because the resulting phenotype is decreased muscle mass and abnormal muscle fiber morphology (Mishima et al., 2009; O'Rourke et al., 2007). In addition, members of the miR1 and miR206, and miR133 families, referred to as muscle-specific miRNA (myomiRs) (Goljanek-Whysall et al., 2012; McCarthy, 2008), are known to regulate genes that participate in adult myoblast activation, including Histone Deacetylase 4, DNA Polymerase α and Connexin 43 (Anderson et al., 2006; Chen et al., 2006; Goljanek-Whysall et al., 2012; Kim et al., 2006). However, few miRs have been identified that directly control early myogenic transcriptional regulators. Of the known embryonic MRF transcriptional activators, only Pax3 is reported to be miR-regulated (Gagan et al., 2012).
We focused our investigation on miRs conserved across species, and identified miR30a as a potential regulator of zebrafish Six1 and of myogenesis. Previously, miR185 has been shown to regulate SIX1 expression in a kidney cancer xenograft model (Imam et al., 2010); however, this miR is not present in zebrafish, nor is it reported to be expressed embryonically. The miR30 family includes five members (a–e) that share the same eight-nucleotide seed sequence and are conserved from zebrafish to humans. During Xenopus embryogenesis, the miR30 family regulates pronephros development through targeting the transcription factor Xlim1/Lhx1, and miR30a specifically is implicated in hepatobiliary duct formation in zebrafish (Agrawal et al., 2009; Hand et al., 2009). In situ analyses in Xenopus embryos also revealed expression of miR30 family members in the somites (Agrawal et al., 2009), which is consistent with a role in myogenic regulation. Here, utilizing molecular and embryological techniques in the zebrafish, we demonstrate that miR30a function is crucial for primary myogenesis. Furthermore, we demonstrate in vivo that the contribution of miR30a to muscle development is through direct modulation of Six1 protein expression.
miR30a and six1a/b expression inversely correlate during primary myogenesis
Because precise modulation of Six1 protein levels is essential for proper myogenesis, we hypothesized that Six1 could be regulated through a miR-mediated mechanism. Owing to a whole-genome duplication event early in the teleost lineage, there are two SIX1 orthologs in zebrafish, six1a and six1b (that latter of which was formerly known as six1a; hereafter, when referring to both isoforms, we use six1a/b). Several publicly available miR:mRNA target prediction algorithms (TargetScan, RegRNA, Zebrafish microRNA Targets) were utilized to identify species-conserved miRs predicted to target both Six1 transcripts. We focused on miR30a, as the 3′ untranslated region (UTR) of six1a contains three predicted miR30a seed target sequences, and the coding sequence of six1b contains one predicted target site (Fig. 1A). In addition, both the mouse and human SIX1 3′UTRs contain two predicted miR30a seed target sequences.
To determine whether miR30a is a potential regulator of Six1, in situ hybridization was performed to examine early embryonic expression patterns for six1a, six1b and miR30a. The paraxial mesoderm is divided along the anterior–posterior axis of the embryo into segments called somites, which will give rise to skeletal muscle. At 24 hours post-fertilization (hpf), mature miR30a expression was absent in the somites, but was observed at 48–72 hpf (Fig. 1B–D). Tranverse sectioning confirmed that miR30a was expressed in the somites specifically (Fig. 1K), whereas real-time PCR analysis revealed a significant increase in miR30a from 24 to 72 hpf (Fig. 1L). In contrast, six1a/b expression was strongest in the somites at 24 hpf, with transcript levels decreasing dramatically by 48 hpf, and remaining absent at 72 hpf (Fig. 1E–J). This temporally reciprocal expression pattern in the somites between miR30a and predicted targets six1a/b is consistent with a role for miR30a in downregulating expression of six1a/b. Supporting this relationship, expression of six1b is additionally observed in non-myogenic lateral line neuromasts in the trunk at 48–72 hpf, where miR30a is not expressed (Fig. 1I,J).
six1a and six1b are both required for proper primary myogenesis
Previously, six1b was characterized as a transcription factor important for fast muscle myogenesis in zebrafish (Bessarab et al., 2004; Bessarab et al., 2008; Nord et al., 2013), whereas six1a had not been extensively studied in muscle development. However, six1a was recently shown to promote proliferation of fast muscle progenitors (Nord et al., 2013). Because six1a has three binding sites for miR30a in its in 3′UTR, we first examined the broader contribution of six1a to early myogenesis. Injection of a translation-blocking morpholino (MO) directed against six1a (13 ng, specificity confirmed in supplementary material Fig. S1A–D) into one- or two-cell embryos, led to perturbed somite morphology, indicative of abnormal muscle formation (Fig. 2A). Immunofluorescence staining for Myosin Heavy Chain (MyHC) and fast muscle marker F310 indicated perturbed muscle fiber alignment in the six1a morphants at 24 hpf (Fig. 2B–D). Immunofluorescence staining revealed areas where muscle fibers crossover each other (Fig. 2C, red arrows), and that the overall cross-sectional diameter was decreased (Fig. 2D, inset), which is consistent with what has been previously reported using a splice-blocking six1a MO (Nord et al., 2013). In addition, loss of the ability to refract polarized light (birefringence) demonstrated that six1a knockdown led to irregular muscle fiber arrangement (Fig. 2E).
Given that these phenotypes are similar to those observed upon six1b knockdown (5 ng six1b MO) (Bessarab et al., 2008), we investigated the combined contribution of the paralogs to muscle development. Using lower concentrations of MO that individually do not affect muscle morphology (6.5 ng sixla MO, 2.5 ng six1b MO, supplementary material Fig. S2A), dual knockdown of six1a/b leads to markedly enhanced alterations in gross morphology, including a V-to-U shape change of somites and a curving of tails (Fig. 2F; supplementary material Fig. S2A). Additionally, we found that the six1a/b paralogs cooperate to promote survival during muscle development, which had not previously been shown. Importantly, knockdown of each paralog alone resulted in increased apoptosis, as marked by the presence of cleaved caspase-3, with the greatest levels observed with six1a/b dual inhibition (Fig. 2G,H). These results complement a recent study that reported a decrease in muscle area and inhibition of proliferation in six1a/b morphants, utilizing a different six1a MO (Nord et al., 2013), and together indicate a crucial role for Six1 in proper zebrafish muscle development. Finally, we performed in situ analyses for muscle markers in our six1a/b double knockdown, as well as each single knockdown, at an earlier timepoint (9 somites, ∼12 hpf) when six1b has been reported to delay expression of the MRF myogenin (myog) and the muscle differentiation marker tropomyosin-α (tpma). We found that the strongest delayed expression of both markers occurred with six1a/b double knockdown (Fig. 2I; supplementary material Fig. S2B). These data demonstrate that zebrafish six1a and six1b cooperate to mediate primary myogenesis.
miR30a overexpression phenocopies six1a/b knockdown
To assess the function of miR30a in vivo during primary myogenesis, we injected 22-nucleotide RNA duplexes representing the processed mature form of the miR into one- or two-cell embryos and examined multiple markers of myogenic differentiation. Similar to six1a/b knockdown, miR30a overexpression resulted in V-to-U shape changes in somite morphology, as visualized using phalloidin staining for actin, which is indicative of overall muscle fiber morphology (Fig. 3A). In addition, disarrayed muscle fiber arrangement was assessed by immunofluorescence staining for MyHC and fast-twitch muscle fibers specifically (by using antibody F310), and a loss of birefringence, which assays for muscle fiber alignment (Fig. 3B–E). Phenocopying six1a/b loss, muscle fibers crossed over each other (Fig. 3B, white arrows) and had a decreased fast muscle cross-sectional diameter (Fig. 3C, inset) when miR30a levels are elevated. Interestingly, miR30a overexpression also appeared to decrease the intensity of F310 staining, as observed when the fluorescent gain was set equally (Fig. 3C, inset). Quantification of the number of somitic cells positive for cleaved caspase-3 revealed that, similar to what is observed with six1a/b loss, there was an increase in apoptosis upon miR30a overexpression (Fig. 3F). Importantly, the loss of birefringence and elevation in cell death were specific to miR30a, as injection of a control miR duplex did not change these properties significantly from control non-injected embryos (quantified in Fig. 3E,F). Thus, miR30a overexpression results in the same phenotypes as six1a/b knockdown, suggesting that miR30a can regulate muscle development through its ability to target six1.
miR30a targets the Six1 homeoprotein in vivo
Given that the phenotypes displayed by miR30a overexpression are consistent with miR30a-mediated inhibition of Six1, we asked whether miR30a directly regulates Six1 levels in vivo. At 24 hpf, miR30a overexpression did not appear to substantially decrease the level of six1a/b mRNA in the somites, as measured by in situ hybridization (Fig. 4A,B). As miR function does not always occur through target mRNA degradation, and instead often occurs through a block in translation (Guo et al., 2010), we next assessed Six1 protein levels in response to miR30a overexpression. Immunofluorescence staining indicates that Six protein expression decreased specifically in the somites at 24 hpf as a result of miR30a overexpression (Fig. 4C). The antibody used to detect Six1 by immunofluorescence also detects other SIX family members (Qamar et al., 2012), and thus, although it is an indicator of decreased levels of Six1, the immunofluorescence result might mask the extent of Six1 regulation owing to its ability to recognize SIX family members not targeted by miR30a. To show that Six1 expression was specifically regulated, western blot analysis was performed with an antibody that allows for specific detection of Six1 (Qamar et al., 2012). Importantly, miR30a injection almost completely abolished Six1 protein levels at 24 hpf, suggesting that miR30a works primarily through inhibiting Six1 translation (Fig. 4D). Six1 inhibition is specific to miR30a, as a control RNA duplex that does not target the Six1 3′UTR does not change Six1 protein expression compared to non-injected controls (Fig. 4D).
To determine whether downregulation of Six1 is directly mediated by miR30a, we generated a reporter construct that consists of the six1a 3′UTR containing the three seed target sites for miR30a downstream of the coding sequence for GFP. When mRNA transcribed from this plasmid (GFP–3′UTR) is injected during the one-cell stage, green fluorescence could be detected in 90.6% of embryos at 24 hpf (Fig. 4E,F). This expression was inhibited when miR30a is overexpressed, as only 25.9% of the embryos were GFP positive, indicating that miR30a can directly target the 3′UTR of six1a (Fig. 4E,F). To determine whether binding of miR30a to the six1a seed sequence is required for repression of GFP expression, site-directed mutagenesis of the fifth and seventh nucleotides of the seed target sequence for all three miR30a sites in the six1a 3′UTR was performed. Injection of GFP–six1a-mutated-3′UTR mRNA (GFP–3Xmut) resulted in 88.8% GFP-positive embryos, as was also the case in the presence of miR30a overexpression (88.0% GFP-positive embryos) (Fig. 4E,F). In addition, the control miR did not affect GFP expression when co-injected with GFP–3′UTR mRNA; and neither miR30a nor the control miR downregulated fluorescent protein expression when co-injected with GFP mRNA lacking the six1a 3′UTR (data not shown). We further attempted to determine whether miR30a regulates six1b expression, in addition to six1a expression, using this reporter system. However, mRNA containing the six1b coding sequence did not express GFP, even in the absence of miR30a. Although we were unable to demonstrate direct regulation of six1b by miR30a, it should be noted that the Six1 antibody is predicted to recognize both six1a and six1b, and indeed, MO experiments performed in our laboratory suggest that it does detect both paralogs (supplementary material Fig. S3). Because miR30a overexpression almost completely eliminated the Six1 signal as observed by immunoblotting (Fig. 4D), our data suggest that both six1a/b are targeted by miR30a, although six1b might be targeted through direct or indirect mechanisms. Overall, these experiments demonstrate that miR30a downregulates Six1 protein expression in zebrafish embryos, and that the six1a 3′UTR is directly targeted by miR30a.
miR30a knockdown demonstrates an endogenous role for this miR in myogenesis
The above-described studies demonstrate that, when overexpressed, miR30a can modulate Six1 protein levels and alter myogenesis. However, they do not address the role of endogenous miR30a in zebrafish myogenesis. To investigate this question, we performed MO-mediated knockdown of mature miR30a (MO controls are shown in supplementary material Fig. S4), and assessed somite and myogenic phenotypes at timepoints when miR30a is normally expressed. The loss of miR30a led to abnormal somite shape at 48 and 72 hpf, similar to that observed with Six1 overexpression achieved by injecting six1a mRNA (Fig. 5A,B; Fig. 6A). Furthermore, proper muscle fiber morphology was disrupted upon miR30a knockdown at 48 hpf as observed with MyHC staining (Fig. 5C) and loss of birefringence (Fig. 5D), again phenocopying Six1 overexpression. Importantly, injection of increasing concentrations of miR30a MO resulted in a dose-dependent upregulation of Six1 protein as observed by western blot analysis (Fig. 5E,F), indicating that miR30a represses Six1 at 48 hpf in wild-type embryos. These data reveal that miR30a functions in vivo to regulate proper somite morphology, likely through its regulation of Six1 levels.
miR30a regulates cell death and somite morphology through its ability to target Six1
To confirm that the abnormal muscle phenotypes associated with miR30a overexpression are due to its ability to target Six1, we re-introduced a six1a mRNA lacking its 3′UTR, and therefore any miR30a seed target sites, into embryos overexpressing miR30a, which rescued Six1 protein expression (Fig. 6A). Restoring six1a expression in the background of miR30a overexpression led to a rescue of normal V-shaped somites (Fig. 6B) and normal muscle fiber morphology, as detected in MyHC-stained embryos (Fig. 6C), and to restored birefringence (Fig. 6D). Furthermore, rescuing six1a expression in miR30a-overexpressing embryos decreased cell death, as observed by assessing for cleaved caspase-3, to the levels observed in non-injected controls in 50% of dual-injected embryos (Fig. 6E). Equal segregation of both miR30a and six1a mRNA to all cells in the embryos following injection is difficult to achieve, which likely explains why 50%, rather than a larger percentage, of dual-injected embryos displayed rescued phenotypes. These experiments demonstrate that miR30a-mediated regulation of Six1 expression is crucial for primary myogenesis in zebrafish.
In this manuscript, we demonstrate that the homeoprotein Six1, an early activator of primary myogenesis, is directly downregulated by miR30a in zebrafish. This provides the first description of a mechanism by which Six1 levels are controlled in embryogenesis, as well as the first demonstration that miR30a plays an important </emph>role in embryonic muscle development and is thus a newly identified myomiR. In addition to demonstrating the reciprocal nature of endogenous six1a/b transcripts and miR30a expression, we performed overexpression and knockdown experiments to show that miR30a modulates six1a/b protein levels in vivo, and that, in the case of six1a, miR30a directly binds to the 3′UTR of the mRNA to regulate its levels. Most importantly, using rescue experiments, we demonstrated that regulation of Six1 by miR30a is crucial for early somitogenesis, identifying miR30a as an important new myomiR controlling a key regulatory node in muscle development.
Transcriptional control of myogenesis is initiated with SIX1 and SIX4, and EYA1 and EYA2, activating expression of the MRF genes, either directly or indirectly through PAX3 and/or PAX7 (Bentzinger et al., 2012). Until now, the only genes in this myogenic network shown to be miR targets were PAX3 and/or PAX7. In mouse embryos, miR27b overexpression decreases Pax3 expression, disrupts muscle progenitor cell migration and promotes premature muscle differentiation (Crist et al., 2009). Furthermore, dual inhibition of both miR1 and miR206 in chick embryos results in Pax3 upregulation and a delay in myogenic differentiation (Goljanek-Whysall et al., 2011). After embryogenesis, PAX7 is regulated by miR206, miR1 and miR486 specifically in postnatal muscle progenitors (satellite cells) (Chen et al., 2010; Dey et al., 2011). These data demonstrate an important role for myomiRs in muscle development, but do not address whether myomiRs control other key transcription factors in the myogenic regulatory network. Thus, our demonstration of miR30a-directed regulation of Six1 is the first evidence for miR-mediated control of crucial upstream regulators of Pax3 during primary myogenesis.
Our studies also reveal that both zebrafish paralogs of the transcription factor SIX1, six1a and six1b, function during early myogenesis to direct somitic cell survival and allow proper organization of fast-twitch muscle fibers. Given that knockdown of either paralog alone leads to similar phenotypes, the two paralogs are likely to have partially redundant functions during myogenesis. Of interest, neither six1a nor six1b can completely compensate for the other in the single knockdown experiments, suggesting there are unique roles that can be ascribed to each paralog. Another interpretation of these data is that a certain level of total Six1 protein is necessary for myogenesis, and knockdown of either paralog alone reduces total levels below sufficiency. Given that the double knockdown of six1a/b results in the most dramatic dysregulation of muscle development, we suggest that these two paralogs work together during early myogenesis.
In addition to contributing to early muscle development, six1b (formerly six1a) participates in both inner ear and pituitary development in zebrafish embryos (Bricaud and Collazo, 2011; Pogoda and Hammerschmidt, 2009), with analogous functions for six1a currently not reported. In the inner ear, six1b has divergent roles, as it promotes hair cell fate but inhibits neuronal fate (Bricaud and Collazo, 2011). Additional tissues in which SIX1 plays an important embryonic role have been identified in several species, including several embryonic placodes and the kidney, amongst others (Brugmann et al., 2004; Grifone et al., 2005; Ikeda et al., 2007; Laclef et al., 2003b; Sato et al., 2010). Similar to what is observed in myogenesis, many of these tissues require SIX1 early during development to promote cell survival, proliferation and migration, but SIX1 protein expression soon decreases later in embryogenesis and is mostly absent in mature tissues (Christensen et al., 2008). Our studies identify a regulatory mechanism for Six1 for the first time in any developmental context, and it would thus be of interest to determine whether this same mechanism of regulation is at play in other tissues in which Six1 levels are tightly controlled.
Although SIX1 expression in most adult tissues is rare, upregulation of this transcription factor has been detected in over ten different cancers, including tumors that arise from skeletal muscle tissue, called rhabdomyosarcoma (RMS) (Christensen et al., 2008; Patrick et al., 2013; Yu et al., 2004). Interestingly, the expression of miR30a decreases with the progression of lymphoma, leukemia, lung, ovarian, breast and colon cancer (Cheng et al., 2012; González-Gugel et al., 2013; Guan et al., 2012; Lee et al., 2012; Liu et al., 2013; Ma et al., 2012; Võsa et al., 2013), all cancers in which SIX1 overexpression has been detected (Behbakht et al., 2007; Ford et al., 1998; Mimae et al., 2012; Ono et al., 2012; Wang et al., 2011). Importantly, increased SIX1 expression enhances progression of RMS and several additional tumor types by re-activating many of the same signaling networks that SIX1 turns on developmentally to promote cell survival, proliferation and migration (Christensen et al., 2008; Khan et al., 1999; Yu et al., 2006; Yu et al., 2004).
An additional embryonic program co-opted by tumor cells, referred to as an epithelial-to-mesenchymal transition (EMT) (Kalluri and Weinberg, 2009), is known to be hijacked by tumor cells (Micalizzi et al., 2010). In the embryo, SIX1 and SIX4 promote muscle precursors to undergo an EMT as they delaminate from the dermomoytome and invade into the developing limb (Alvares et al., 2003; Bladt et al., 1995; Grifone et al., 2005). In breast cancer and colon cancer, SIX1 induces an EMT and thereby promotes progression of these malignancies (McCoy et al., 2009; Micalizzi et al., 2009; Ono et al., 2012). Of note, miR30a has been implicated as a tumor suppressor in a variety of cancers, primarily through downregulation of EMT-driving proteins (Cheng et al., 2012; Kumarswamy et al., 2012). Given our demonstration that miR30a downregulates Six1 developmentally, and existing data that the two molecules have a reciprocal expression pattern in numerous of the same tumor types, it is tempting to speculate that miR30a loss might contribute to high Six1 expression in multiple types of cancer.
In summary, we have identified the first developmental regulator of the Six1 homeoprotein, miR30a, which we show is a myomiR that controls muscle development by directly regulating Six1. We anticipate that our results will have implications that extend beyond muscle development to other developing tissues in which SIX1 plays an essential role, as well as perhaps to cancers such as RMS, in which SIX1 is inappropriately expressed.
MATERIALS AND METHODS
Zebrafish care and use
Experiments involving zebrafish were approved by the University of Colorado Anschutz Medical Campus IACUC committee and complied with all relevant animal welfare laws, guidelines and policies. Wild-type TAB embryos cared for using established protocols (Westerfield, 1993) were staged by morphology (Kimmel et al., 1995) and age in hours post fertilization (hpf).
The six1a (NM_001009904) 3′ untranslated region (UTR) was amplified by RT-PCR from genomic DNA using primers from Eurofins MWG Operon (Huntsville, AL, USA) (forward, 5′-TCTGCAAGGCACCATGAACAATCC-3′; reverse 5′-TCCACTTTGGTGTTTAGTTGACG-3′). To create the GFP–3′UTR construct, the six1a 3′UTR was cloned downstream of GFP in a pCS2+ vector, a kind gift from Abby Olena (Patton Laboratory, Vanderbilt University, Nashville, TN), as previously described (Li et al., 2011). To create the GFP–3Xmut plasmid, each of the three miR30a seed target sites in the six1a 3′UTR in the GFP–3′UTR construct were mutated by site-directed mutagenesis, which was performed utilizing the Stratagene (Santa Clara, CA, USA) Quikchange protocol and primers: M1F, 5′-CTTCTGACGCAGAGGACGTCTGCATCAAGAAAAAAG-3′; M1R, 5′-CTTTTTTCTTGATGCAGACGTCCTCTGCGTCAGAAG-3′; M2F, 5′-GGTTCTTTGTTTTTATTCAGTCTGCATATATAATATG-3′; M2R, 5′-CATATTATATATGCAGACTGAATAAAAACAAAGAACC-3′; M3F, 5′-CAGTTTACATATATAATATGTCTGCATTACAATTGTAG-3′; M3R, 5′-CTACAATTGTAATGCAGACATATTATATATGTAAACTG-3′. To make an expression vector for six1a mRNA lacking the 3′UTR, the six1a coding sequence (CDS) was amplified by RT-PCR from 24 hpf cDNA using primers supplied by Eurofins (forward, 5′-TCGAGGATCCGCCGCCACCATGTCAATCTTGCCCTCGTT-3′; reverse, 5′-CGGCGAATTCCTACGATCCTAAATCCACAAGGC-3′), and cloned into pCS2+. To make capped mRNA, all plasmids were linearized with NotI and transcribed with SP6 RNA polymerase using the mMessage mMachine kit from Ambion (Grand Island, NY, USA). All cloning products were verified by DNA sequencing.
Zebrafish embryos were injected at the one- or two-cell stage with morpholinos (MOs), miRs or mRNA. Gene Tools (Philomath, OR, USA) supplied all MOs, which were injected in the following amounts: 6.5 ng or 13 ng six1a MO (5′-CAAGATTGACATGGCTCCCCTATGC-3′), 2.5 ng or 5 ng six1b MO (5′-TCTCCTCTGGATGCTACGAAGGAAG-3′), 15 ng miR30a MO (5′-ACTTCCAGTCGGGAATGTTTACAAC-3′), or 15 ng of the standard control MO. Integrated DNA Technology (Coralville, IA, USA) supplied single-strand RNA oligonucleotides: miR30a sense, 5′-UGUAAACAUUCCCGACUGGAAG-3′; miR30a antisense, 5′-CUUCCAGUCGGGAAUGUUUACA-3′; control miR-219 sense, 5′-GGAGUUGUGGAUGGACAUCACG-3′; and control miR-219 antisense, 5′-CGUGAUGUCCAUCCACAACUCC-3′. Annealing was performed by mixing equal amounts of sense and antisense oligonucleotides, heating to 83°C for 2 minutes, gradually cooling to room temperature and storing at −80°C. miR duplexes were injected at 0.5, 1 and 3 ng. GFP reporter mRNAs were injected at 100 pg mRNA with or without 2.5 ng miR30a. For overexpression studies, 100 pg of six1a mRNA was injected. For rescue experiments, 2.5 ng miR30a with or without 50 pg six1a mRNA was injected. All experiments were performed in triplicate. In all figure legends, n = the number of embryos represented by each image/the total number of embryos analyzed. Both the standard control MO and control miR were injected at the same concentration as the corresponding experimental reagent, and consistently did not measurably affect any of the phenotypes described throughout this text when compared to non-injected controls.
In situ hybridization
Whole-mount in situ hybridization (ISH) was performed as previously described (Johnson et al., 2011). The six1a and six1b riboprobe plasmids (Bessarab et al., 2008) were provided by Vladimir Korzh (Institute of Medical and Cellular Biology, A*STAR, Proteos, Singapore). The tpma riboprobe plasmid was a kind gift from Sharon Amacher (Ohio State University, Columbus, OH). The myog riboprobe plasmid was provided by Wolfgang Driever, (University of Freiburg, Germany). A Locked Nucleic Acid (LNA) probe for dre-miR30a (5′-CTTCCAGTCGGGAATGTTTACA-3′) labeled with digoxigenin at the 5′ end was synthesized by Exiqon (Woburn, MA, USA). Following ISH, whole embryos were embedded in 5% sucrose/1.5% agar, incubated in 30% sucrose for 24 hours at 4°C, flash-frozen in liquid nitrogen, and cryosectioned at 18 µm.
The heads and yolk were removed from 25 embryos/group, and total RNA was extracted from the remaining tails with the miRNeasy RNA isolation kit from Qiagen (Germantown, MD, USA) following the manufacturer's protocol. cDNA was reverse-transcribed from 1 mg RNA using the miScript II RT kit from Qiagen. Real-time PCR was performed using the SsoFast EvaGreen Supermix from BioRad (Hercules, CA, USA), and the following primers: miR30a forward, 5′-TGTAAACATTCCCGACTGGAAG-3′; U6 forward, 5′-CGCAAGGATGACA-3′; and universal reverse, 5′-GAATCGAGCACCAGTTACCC-3′ on a BioRad CFX96.
All immunofluorescent analyses were performed as previously described (Johnson et al., 2011). Antibodies used were against: MyHC A4.1025 and F310 (1∶20, Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, IA, USA), cleaved caspase-3 (1∶500, BD Pharmingen, San Jose, CA, USA), Six1 (1∶50, Sigma, St. Louis, MO, USA). Secondary antibodies were Alexa-Fluor-488-conjugated anti-rabbit-IgG and Alexa-Fluor-594-conjugated anti-mouse-IgG (1∶750, Invitrogen, Grand Island, NY, USA). Rhodamine–phalloidin from Invitrogen was also used. Quantification of cleaved caspase-3 was performed by counting positive cells in somites 10–15, with somites visualized by phalloidin or MyHC staining.
At total of 50 whole embryos/group were dechorionated, deyolked using Ginzburg Fish Ringers solution, rinsed in deyolking wash buffer and snap-frozen using liquid nitrogen as previously described (Westerfield, 1993). Lysis was performed in 50 µl RIPA buffer containing protease inhibitors on ice for 1 hour, followed by dounce homogenization, and centrifugation at 13,000 g for 15 minutes. The supernatant was then run on a 10% SDS-PAGE gel, transferred onto a PVDF membrane and probed for Six1 at 1∶1000 (Behbakht et al., 2007) and GAPDH at 1∶5000 (Sigma) as primary antibodies, and horseradish-peroxidase-conjugated anti-rabbit-IgG at 1∶10,000 (Sigma) as the secondary antibody.
Live embryos and embryos processed for ISH were imaged on a Leica (Wetzlar, Germany) M165 FC dissecting microscope. For differential interference contrast (DIC), polarized light and most immunofluorescence studies, embryos were imaged on an Olympus (Center Valley, PA, USA) BX51WI compound microscope with Olympus UPlan FL N objectives at 10× magnification, 0.3 aperture or 20× magnification, 0.5 aperture. Images were captured using a QI Click camera (QImaging, British Columbia, Canada) and QCapture software (QImaging). Muscle fiber immunofluorescence images were captured on a Leica TCS SP5 laser scanning confocal microscope utilizing an HCX PL APO 63× objective with 0.6 aperture, with LAS software (Leica). All embryos were mounted in 80% glycerol and imaged at room temperature.
Prism 5 software (GraphPad, La Jolla, CA, USA) was used for one-way analysis of variance (ANOVA) with Bonferroni's post-hoc test.
We would like to thank Morgan Singleton for help with zebrafish care and Jason Williams (both Department of Craniofacial Biology, University of Colorado, Aurora, USA) for cryosectioning.
J.H.O., K.B.A. and H.L.F. contributed to experimental conception and design, data interpretation and manuscript preparation. J.H.O. and L.H.L. executed all experiments.
This work was supported by a National Cancer Institute NRSA postdoctoral fellowship [grant number F32CA174169 to J.H.O.]; by a grant from the National Cancer Institute [grant number R01CA095277 to H.L.F.]; and by a Golfers Against Cancer grant to H.L.F. and K.B.A. Deposited in PMC for release after 12 months.
The authors declare no competing interests.