Expression and Functional Characterization of Smyd1a in Myofibril Organization of Skeletal Muscles

Background Smyd1, the founding member of the Smyd family including Smyd-1, 2, 3, 4 and 5, is a SET and MYND domain containing protein that plays a key role in myofibril assembly in skeletal and cardiac muscles. Bioinformatic analysis revealed that zebrafish genome contains two highly related smyd1 genes, smyd1a and smyd1b. Although Smyd1b function is well characterized in skeletal and cardiac muscles, the function of Smyd1a is, however, unknown. Methodology/Principal Findings To investigate the function of Smyd1a in muscle development, we isolated smyd1a from zebrafish, and characterized its expression and function during muscle development via gene knockdown and transgenic expression approaches. The results showed that smyd1a was strongly expressed in skeletal muscles of zebrafish embryos. Functional analysis revealed that knockdown of smyd1a alone had no significant effect on myofibril assembly in zebrafish skeletal muscles. However, knockdown of smyd1a and smyd1b together resulted in a complete disruption of myofibril organization in skeletal muscles, a phenotype stronger than knockdown of smyd1a or smyd1b alone. Moreover, ectopic expression of zebrafish smyd1a or mouse Smyd1 transgene could rescue the myofibril defects from the smyd1b knockdown in zebrafish embryos. Conclusion/Significance Collectively, these data indicate that Smyd1a and Smyd1b share similar biological activity in myofibril assembly in zebrafish embryos. However, Smyd1b appears to play a major role in this process.


Introduction
Members of the Smyd family are newly identified proteins that have been implicated in diverse biological functions in embryonic development and cancer [1]. Currently, five smyd genes (Smyd1, 22, 23, 24, and 25) have been identified in vertebrates based on the presence of both SET and MYND domains in their protein sequences [1]. Smyd1, also known as skm-Bop, represents the first identified member of the Smyd family [2,3]. Smyd1 is specifically expressed in skeletal and cardiac muscles and plays a key role in muscle development and embryonic survival in mice and zebrafish [4][5][6]. Targeted disruption of the smyd1 gene resulted in defective cardiac morphogenesis and early embryonic lethality of mouse embryos [4]. Knockdown or mutation of smyd1b gene in zebrafish led to disruption of myofibril organization in skeletal and cardiac muscles in zebrafish embryos [5,6].
The smyd1gene is a direct downstream gene target of myogenic regulatory factors MyoD, Myogenin and Mef2 that control the muscle specific expression of smyd1 in skeletal muscles during embryogenesis and in adult muscle tissues [7][8][9][10]. A recent report showed that smyd1 expression is also regulated by serum response factor (SRF) through direct binding to the promoter region of smyd1 [11]. In addition, smyd1 gene expression can be repressed by Hepatoma-derived growth factor through interaction with a transcriptional co-repressor C-terminal binding protein (CtBP) [12]. Consistent with the idea of being a downstream factor of MyoD and Mef2, loss of Smyd1 function had no effect on myoD and myogenin gene expression and myoblast specification [5]. However, loss of Smyd1 function resulted in defective sarcomere organization in myofibers of skeletal and cardiac muscles, suggesting that Smyd1 is required in the late stage of muscle cell differentiation and myofiber maturation [5,6].
At present, little is known about the mechanism by which Smyd1 functions in myofibrillogenesis. In vitro studies have shown that Smyd1 has a histone methyltransferase (HMTase) activity [5,13], and could function as a transcriptional repressor in a histone deacetylase (HDAC)-dependent manor [4,14]. However, Just and colleagues reported recently that the Smyd1 mutant lacking the HMTase activity was biologically active in myofibril assembly [6], arguing against Smyd1 being a HMTase activity dependent transcriptional regulator. Interestingly, Just and colleagues showed that GST-tagged Smyd1 was capable of pulling down skeletal muscle-specific myosin heavy chain [6]. Consistent with a potential role of Smyd1 outside of the nucleus, a nuclear to cytoplasmic translocation was observed during myoblast differentiation into myotubes [15], and Smyd1 is localized on the M-lines of sarcomeres although the biological significance of the sarcomeric localization is not clear [6,16].
Recent studies demonstrated that zebrafish genome contains two highly related smyd1 genes, smyd1a and smyd1b [17]. Most of the previous studies were focused on smyd1b, very little is known about smyd1a. It is not clear whether Smyd1a plays a similar role as Smyd1b in myofibrillogenesis. To investigate the function of Smyd1a in myofibril assembly, we isolated smyd1a from zebrafish, and characterized its expression and function during muscle development. The results showed that smyd1a was specifically expressed in skeletal muscles of zebrafish embryos. smyd1a expression came several hours later than smyd1b during myogenesis in zebrafish embryos. Functional analysis revealed that knockdown of smyd1a alone had little effect on myofibril assembly in zebrafish skeletal muscles. However, knockdown of smyd1a and smyd1b together resulted in a stronger phenotype in myofibril disorganization. Moreover, the myofibril defects from smyd1b knockdown could be rescued by an ectopic expression of the zebrafish smyd1a or mouse Smyd1 transgene. Together, these data indicate that Smyd1a and Smyd1b share similar biological activity in myofibril assembly although the function of Smyd1b appears to be more critical.

Characterization of Smyd1a in Zebrafish
Sequence analysis revealed that zebrafish genome contains two highly related smyd1 genes (smyd1a and smyd1b) with similar gene structures [17]. smyd1a and smyd1b are believed to be generated by gene duplication. The zebrafish smyd1a is located on chromosome 5, whereas smyd1b is located on chromosome 8 ( Figure S1). Sequence analysis revealed a strong synteny arrangement of zebrafish smyd1a gene and human smyd1 gene with the thronine synthase like 2 gene (THNSL2) and fatty acid binding protein1 gene (FABP1) in zebrafish and mouse genome ( Figure S1). A similar synteny arrangement was found with the zebrafish smyd1b and fatty acid binding protein 1b like gene (fabp1b). However, no synteny arrangement was found with the smyd1b gene and thronine synthase like 2 gene on chromosome 8 in zebrafish. Sequence alignment revealed that zebrafish Smyd1a contains the highly conserved SET and MYND domains involved in protein methylation and proteinprotein interactions, respectively ( Figure S1). Both Smyd1a and Smyd1b share high sequence identity with Smyd1 from other vertebrate species ( Figure S2), although only smyd1 gene has been identified in mice and humans.

Temporal and Spatial Expression of Smyd1a in Zebrafish Embryos
It has been reported previously that both smyd1a and smyd1b are specifically expressed in developing muscles of zebrafish embryos [5,9,17]. To compare their patterns of expression during embryonic development, we carried out a temporal and spatial expression analysis by RT-PCR and whole mount in situ hybridization in zebrafish embryos. Data from the expression analysis showed that smyd1a and smyd1b had different pattern of temporal expression. smyd1a was expressed several hours later than smyd1b in zebrafish embryos. smyd1b expression was first detected at 6 hours post-fertilization (hpf) with a strong expression starting around 9 hpf [5]. In contrast, smyd1a expression could not be detected until 19 hpf (Fig. 1A).
Previous studies have demonstrated that smyd1b encodes two muscle-specific isoforms of mRNA transcripts, designated as smyd1b_tv1 and smyd1b_tv2 [5,16]. smyd1b_tv1 and smyd1b_tv2 are generated by alternative splicing of the 39 bp exon 5. smyd1b_tv1 transcript contains the 39 bp sequence from exon 5, whereas smyd1b_tv2 transcript lacks this 39 bp sequence and thus is 13 aa shorter than smyd1b_tv1 [5,16]. Sequence analysis revealed that the zebrafish smyd1a gene also contained the small exon 5 of 39 bp. To test whether exon 5 could be alternatively spliced in smyd1a, we performed RT-PCR analysis using cDNA from zebrafish larvae of different stages. In contrast to smyd1b where two different isoforms were amplified (Fig. 1B), only the longer isoform of smyd1a was found to be expressed (Fig. 1C), suggesting that exon 5 in smyd1a was not alternatively spliced in zebrafish embryos.
To determine whether smyd1a expression is restricted to muscle cells, we analyzed its spatial pattern of expression by whole mount in situ hybridization. The data showed that similar to smyd1b, smyd1a is specifically expressed in skeletal muscles of zebrafish embryos ( Fig. 1D-I). However, unlike smyd1b, little or no expression was detected for smyd1a transcripts in cardiac muscles (Fig. 1D, G).

Knockdown of smyd1a Expression in Zebrafish Embryos had Little Effect on Muscle Development
To determine whether smyd1a plays a role in muscle development, we knocked down smyd1a expression in zebrafish embryos using two smyd1a-specific antisense morpholino oligos, E8I8-MO and E9I9-MO. The E8I8-MO and E9I9-MO were specifically targeted to the sequences at the exon-8/intron-8, or exon-9/ intron-9 junctions, respectively ( Fig. 2A). RT-PCR analysis confirmed that both E8I8-MO and E9I9-MO could knock down the normal splicing of smyd1a transcripts and resulted in the production of defectively spliced smyd1a mRNA (Fig. 2B). However, the knockdown was not complete with an efficiency of approximately 50% (Fig. 2B). Sequence analyses revealed that E8I8-MO caused a defective splicing at the exon 8 and intron 8 junction, resulting in a 14 bp deletion at the end of exon 8 ( Fig. 2A). This 14 deletion caused a reading frame shift, leading to the production of a mutant protein without the 105 aa C-terminal sequence. Similarly, the E9I9-MO caused a defective splicing at the end of exon 9 with a 58 bp deletion ( Fig. 2A), which resulted in the production of a mutant protein without the last 63 aa Cterminal sequence.
The smyd1a knockdown embryos were examined morphologically for several days following the MO injection. Unlike the smyd1b knockdown embryos that had no skeletal and cardiac muscle contraction, the smyd1a knockdown embryos had normal skeletal and cardiac muscle contraction when observed at 24, 48 and 72 hpf. The E8I8-MO injected embryos appeared morphologically normal (Fig. 2E, F) compared with the control (Fig. 2C, D). Embryos injected with the smyd1a E9I9-MO, however, showed small head and developmental delay, suggesting some off-target toxic effect. The off-target toxic effect could be alleviated by coinjection with the p53-MO (Fig. 2G, H). To confirm that knockdown of Smyd1a did not affect early muscle development, we analyzed the specification of slow and fast muscle precursors and their subsequent migration and differentiation. Compared with control ( Fig. 2I), MyoD expression appeared normal in slow and fast muscle precursors of smyd1a knockdown embryos (Fig. 2I, J). In addition, fiber-type-specific expression of slow or fast myosin heavy chain (MHC) genes also appeared normal (Fig. 2M, N). Slow fibers were clearly localized at the superficial layer of the myotome in smyd1a knockdown embryos (Fig. 2M), suggesting that migration of slow muscle cells was not affected by smyd1a knockdown. To analyze the effects of smyd1a knockdown on sarcomere assembly in skeletal muscles, we first examined the organization of thick filaments in slow myofibers of smyd1a knockdown zebrafish embryos. The results showed that knockdown of smyd1a alone had no detectable effect on the sarcomere organization of the myosin thick filaments ( Fig. 3D-F) in slow fibers of zebrafish embryos at 24, 48 and 72 hpf. The lack of muscle defect from smyd1a knockdown could be due to the inefficient knockdown. To determine whether increasing the efficiency of smyd1a knockdown could affect muscle development, we co-injected the two smyd1a splicing MOs into zebrafish embryos. Compared with the single MO injection, the results showed that co-injection of the two MOs increased the efficiency of smyd1a knockdown (Fig. 2B). However, it did not result in more defects on myofibril assembly in slow muscles ( Fig. 3G-I).
To test the possibility that the lack of muscle defect from smyd1a knockdown was due to the redundant function from smyd1b, we carried out the double knockdown of smyd1a and smyd1b in zebrafish embryos. The data showed that knockdown of smyd1a and smyd1b together resulted in a stronger muscle phenotype compared with smyd1a or smyd1b knockdown alone. As shown in Fig. 3J, knockdown of smyd1b alone resulted in disruption of myosin thick filament organization at 28 hpf. However, the defective thick filament organization was partially recovered in smyd1b knockdown zebrafish embryos at 48 and 72 hpf (Fig. 3K, L). Double knockdown studies showed that the partial recovery from smyd1b knockdown was diminished when smyd1b was knocked down together with smyd1a ( Fig. 3N,O), suggesting that Smyd1a may have a partial function redundancy with Smyd1b in myofibril assembly.
To determine whether this was also true for other sarcomeric structures, we analyzed the a-actin thin filaments and Z-line organization in the single or double knockdown embryos. The results showed that the knockdown of smyd1a and smyd1b together had a stronger effect on thin filament disruption at 72 h pf ( Fig. 4O) compared with the knockdown alone ( Fig. 4I, L). A similar finding was also observed with the Z-line structure (Fig. 5). Collectively, these data indicate that Smyd1a and Smyd1b are required for myofibril organization and assembly in skeletal muscles of zebrafish embryos although Smyd1b plays a dominant role compared with Smyd1a.
In situ expression analysis showed that smyd1a was expressed in fast muscles (Fig. 1). To determine whether knockdown of smyd1a had any effect on fast muscles, sarcomere organization was characterized in fast muscles of smyd1a knockdown embryos. The results showed that knockdown of smyd1a alone had no effect on myofibril assembly in fast muscles (Fig. 6).

Ectopic Expression of Zebrafish smyd1a or Mouse Smyd1 Transgene Could Rescue the Myofibril Defects in smyd1b Knockdown Zebrafish Embryos
It has been suggested that actinopterygian fish, such as zebrafish, had gone through one more round of whole genome duplication compared with mammals during evolution [18]. The two Smyd1 genes in zebrafish were likely generated by the gene duplication event in teleosts because mouse and human genomes contain only one Smyd1 gene. Sequence analysis revealed that zebrafish Smyd1a and Smyd1b share high sequence similarity with the mouse and human Smyd1 (Fig. S2). However, functional studies revealed that knockdown of smyd1a or smyd1b resulted in strong differences in muscle phenotype. Knockdown of smyd1b significantly disrupted the myofibril organization, whereas knockdown of smyd1a alone had little effect. The different severity of the knockdown phenotypes could be due to structural changes of smyd1a protein that made it less active compared with smyd1b. Alternatively, the different knockdown phenotype could be caused by their distinct patterns of temporal expression because smyd1a was expressed several hours later than smyd1b in zebrafish embryos.
To clarify these questions, we tested whether ectopic expression of smyd1a early in zebrafish embryos could rescue the smyd1b knockdown phenotypes. A rescue experiment was performed in the smyd1b knockdown zebrafish embryos using the smyd1a transgene (Smyd1-zfsmyd1a myc ) directed by the smyd1b promoter. The Smyd1-zfsmyd1a myc transgene expressing a myc-tagged zebrafish Smyd1a was co-injected with smyd1b ATG-MO into zebrafish embryos. The smyd1b ATG-MO could specifically knock down the expression of smyd1b, however it had no effect on the expression of the zebrafish smyd1a myc transgene because the transgene did not contain the smyd1b ATG-MO target sequence. Myofibril organization and ectopic Smyd1a myc expression was analyzed in the coinjected zebrafish embryos by double staining with anti-MyHC (F59) and an anti-myc antibodies. The data showed a clear rescue of thick filament organization in the co-injected embryos (Fig. 7A). The rescue appeared in a mosaic pattern, consistent with the pattern of gene expression through DNA injection. Double staining revealed a perfect match between the rescued myofibers and the expression of the myc-tagged zebrafish Smyd1a (Fig. 7C, E). Collectively, these data indicate that Smyd1a could replace Smyd1b function if expressed early in muscle cells, arguing that the different severity of smyd1a and smyd1b knockdown phenotypes was likely due to distinct pattern of temporal expression.
Sequence analysis revealed that Smyd1a and Smyd1b share similar sequence identity with mouse and human Smyd1 (Fig. S2). To test whether Smyd1 function is conserved during evolution, we performed a rescue experiment in the smyd1b knockdown zebrafish embryos using the mouse smyd1 transgene directed by the zebrafish smyd1b promoter. The mouse Smyd1 transgene (smyd1-mSmyd1 myc ) expressing a myc-tagged mouse Smyd1 was co-injected with the deletion of 14. Defective splicing using a cryptic splicing site in exon 9 in E9I9-MO injected embryos resulted in a deletion of 58 bp. B. RT-PCR showing the defective splicing induced by the E8I8-MO or E9I9-MO or both MOs. Compared with the PCR results from the wild type (wt) control embryos where a single band was generated, two bands were observed in the E8I8-MO injected embryos. Sequence analysis revealed that the smaller band contained a 14 bp deletion at the end of exon 8. The upper band represented the heteroduplex formed by the normal and defectively spliced products. Similarly, three bands were detected in E9I9-MO injected embryos. Sequence analysis revealed that the smaller band contained a 58 bp deletion at the end of exon 9. The middle band was the normal spliced product, whereas the upper band represented the heteroduplex formed by the normal and defectively spliced products. Two major bands were detected in E8I8-MO and E9I9-MO co-injected embryos. Sequence analysis revealed that the smaller band resulted from defective splicing. smyd1b ATG-MO into zebrafish embryos. Double staining revealed that expression of the myc-tagged mouse Smyd1 could rescue the myofibril defects from the smyd1b knockdown in a cell autonomous manner (Fig. 7B, D, F). Collectively, these data indicate that Smyd1function in myofibril organization is likely conserved during evolution.

The Subcellular Localization of Zebrafish Smyd1a and Mouse Smyd1 in Myofibers
Previous studies by us and others have shown that Smyd1b is localized on the M-line of sarcomeres in both skeletal and cardiac muscles [6,16]. This was especially evident for the longer isoform Smyd1b_tv1 [16]. It is not clear whether zebrafish Smyd1a and mouse Smyd1 could also localize on M-lines of sarcomere in skeletal muscles. To clarify this question, we examined the subcellular localization of myc-tagged zebrafish Smyd1a and mouse Smyd1 in skeletal muscles of zebrafish embryos. The results showed that, similar to Smyd1b, mouse Smyd1 was localized on the sarcomere of skeletal muscles (Fig. 7D). However, unlike Smyd1b, zebrafish Smyd1a showed no clear sarcomeric localization in myofibers (Fig. 7C).
To enhance the sensitivity of the detection, we generated a transgene, pTol2-Smyd1a -EGFP, expressing a GFP-tagged zebrafish Smyd1a fusion protein. The pTol2-Smyd1a-EGFP transgene was injected into zebrafish embryos. Smyd1a-EGFP expression and subcellular localization were characterized in the injected zebrafish embryos by direct observation. The results showed that Smyd1b-EGFP had a clear sarcomeric localization (Fig. 8B, E). In contrast, Smyd1a-EGFP was not localized on the sarcomeres (Fig. 8A, D). A diffuse cytosolic localization was detected for Smyd1a-EGFP in myofibers of zebarfish embryos (Fig. 8A, B), similar to EGFP control (Fig. 8C, F). Together, these data indicate that Smyd1a and Smyd1b have distinct subcellular localization, indicating that the sarcomeric localization of Smyd1b might not be directly linked with the Smyd1 function in sarcomere organization because Smyd1a could rescue the myofibril defects from Smyd1b knockdown although Smyd1a was not localized on the sarcomere. This is consistent with previous findings that either one of the Smyd1b isoforms, Smyd1b_tv1 or Smyd1b_tv2, from alternative splicing was able to rescue the smyd1b knockdown phenotype although their proteins showed distinct subcellular localization [5,16].

Discussion
In this study, we have characterized the muscle phenotype from knockdown of smyd1a or smyd1b alone, or together in zebrafish embryos. We demonstrated that in contrast to smyd1b which is absolutely required for myofibril organization in skeletal muscles, knockdown of smyd1a alone had very little effect on myofibril organization. However, knockdown of smyd1a and smyd1b together resulted in a stronger muscle phenotype compared with knockdown of smyd1a or smyd1b alone. We further demonstrated that the muscle defects from smyd1b knockdown could be rescued by the ectopic expression of the zebrafish smyd1a or mouse Smyd1, suggesting that Smyd1 function is likely to be conserved during evolution. Finally, we showed that similar to Smyd1b, mouse Smyd1 was localized on sarcomeres in skeletal myofibers. In contrast, zebrafish Smyd1a did not show any sarcomeric localization although it could rescue the Smyd1b knockdown defects.

The Unequal Role of Smyd1a and Smyd1b in Myofibril Assembly
Gene duplication plays a vital role in evolution [19]. Sequence analysis revealed that zebrafish genome contains two smyd1 genes, smyd1a and smyd1b that were likely generated by the whole genome duplication in actinopterygian fish during evolution [17,18,20]. smyd1a and smyd1b are paralogous genes that are dispersed on different chromosomes in zebrafish. Sequence analysis revealed that zebrafsh smyd1a and mouse Smyd1 showed a highly conserved synteny arrangement with the surrounding THNSL2 and FABP1 genes. The synteny arrangement with fabp1was maintained for smyd1b in zebrafish. However, this synteny was not found between the zebrafish smyd1b and thnsl2 genes, suggesting that a genomic rearrangement or deletion might occur in this region after the whole genome duplication.
It has been suggested that after whole genome duplication, the duplicated genes could evolve in several directions without the selective pressure. Some of the duplicated genes could gain new functions, while many others turned into nonfunctional pseudogenes [19,20]. We showed in this study that both smyd1a and smyd1b are functional genes after the duplication. However, they do not play an equal role in myofibril organization. Smyd1b appears to be more critical than Smyd1a because knockdown of smyd1b gave a stronger muscle phenotype compared with smyd1a knockdown. Our data indicate that functional difference between Smyd1a and Smyd1b could be caused by their different patterns of temporal expression, different activity due to changes of protein structure, or different levels of gene expression. We showed that early ectopic expression of smyd1a by using the smyd1b promoter could rescue the myofibril defects from smyd1b knockdown, suggesting that Smyd1a could function as Smyd1b if expressed in a similar temporal and spatial pattern in zebrafish embryos. At present, it is not clear whether Smyd1a gained new functions during evolution. Knockdown of smyd1a alone failed to reveal any significant phenotype in zebrafish embryos, unless it was knocked down together with smyd1b. Together, these studies suggest that Smyd1a is involved in myofibril assembly, but with a minor role in the process.

The Smyd1 Function in Myofibril Organization is Conserved during Evolution
Our studies demonstrated that Smyd1 function in myofibril organization is likely to be conserved during evolution because ectopic expression of mouse Smyd1 could rescue the muscle defects from smyd1b knockdown in zebrafish embryos. Unlike zebrafish, there is only one Smyd1 gene in mouse and human genomes. Consistent with the idea that zebrafish smyd1b is the ortholog of Smyd1 in mice, it has been reported that knockout of Smyd1 in mice resulted in early embryonic lethality from cardiomyogenesis defects [4]. A cardiac muscle defect was also observed in Smyd1b knockdown or mutant zebrafish embryos [5,6]. In addition to their functional similarity, the zebrafish Smyd1b and mouse Smyd1 share other common features in gene expression and protein localization. First, both zebrafish smyd1b and mouse smyd1 express two alternatively spliced mRNA transcripts that differ by 39 bp encoded by exon 5. Second, both mouse Smyd1 and zebrafsh Smyd1b are localized on the M-lines of sarcomeres in skeletal and cardiac muscles [6], which is in contrast to the zebrafish Smyd1a that showed no sarcomeric localization.

The Sarcomeric Localization of Smyd1
The biological significance of the sarcomeric localization is not clear. It has been reported that Smyd1b could bind to myosin and showed a transient localization with thick filaments during myofibrillogenesis in zebrafish embryos [6]. However, after the completion of sarcomere organization, the Smyd1b is translocated to the M-line. Interesting, Etard and colleagues have shown that myosin chaperones Unc45b and Hsp90a could shuttle between the A band and the Z line in response to stress or damage to the myofiber [21]. It has been suggested that the sarcomeric localization of myosin chaperones and their translocation within different parts of the muscle cells could be involved in the response of muscle cells to mount efficient physiological responses to muscle stress, load requirements, and/or stretch [21]. It remains to be determined whether the sarcomeric localization of Smyd1b is involved in sarcomere remodeling and response to muscle stress.
We have shown previously that the sarcomeric localization of Smyd1b requires Phe223 and Serine 225 [16]. Substitution of Phe223 or Ser225 with alanine significantly diminished the sarcomeric localization of Smyd1b [16]. Interestingly, the two highly conserved residues Phe223 and Ser225 are present in zebrafish Smyd1a, arguing that in addition to the Phe223 and Se225 residues, other residues or motifs are likely to be involved in the sarcomeric localization. Sequence analysis revealed that Smyd1a differs from Smyd1b significantly at the C-terminal sequence ( Figure S1). Compared with Smyd1a, Smyd1b has seven extra amino acids at the C-terminus including three highly conserved residues (xxLFxxK) that are present in all Smyd1 proteins that have been identified in vertebrates. It remains to be determined whether these residues may contribute to the sarcomeric localization of Smyd1b.

Ethics Statement
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Institutional Animal Care and Use Committee of the University of Maryland (Permit Number:0610009).

Fish Maintenance and Use
Zebrafish were maintained at 28.5 C in 10 gallon aquarium supplied with fresh water and air at a photoperiod of 14 hours of light and 10 hours of dark at the Zebrafish Facility at Aquaculture Research Center (ARC) in Columbus Center (Baltimore). Pairs of adult male and female zebrafish were put in 1 liter tanks for natural spawning after lights were turned on in the facility. Embryos were collected and raised at 28.5 C. To ease pain and facilitate animal handling, fish embryos over one day old were anaesthetized in 0.6 mM Tricaine that has been buffered to neutral pH around 7. The anaesthetized embryos were sacrificed quickly by keeping on ice and used directly for RNA extraction or fixation for immunostaining. All procedures were carried out in compliance with the guidelines stipulated by the Institutional Animal Care and Use Committee of the University of Maryland.

Isolation of Zebrafish smyd1a cDNA
The full-length coding sequence of zebrafish smyd1a (NM_205540) was cloned by RT-PCR using total RNA extracted from zebrafish embryos at 24-72 hpf. The PCR was carried with the Advantage PCR kit using the following primers: zfsmyd1a-P1 (59 -AGCATGACCGTGGAGAAGACGGAC-39) and zfsmyd1a-P2 (59 -GTGTTCATGCTTTGATCTGCACTT-39) derived from DNA sequences near the start codon and stop codon,

Analysis of smyd1a and Smd1b Expression in Zebrafish Embryos by RT-PCR
To analyze the temporal pattern of smyd1a expression during development, total RNA was extracted from zebrafish embryos at 20 min, 3h, 6h, 9h, 12h, 16h, 1d, 2d, 3d, 4d and 5d post fertilization. smyd1a expression was determined by RT-PCR using zfsmyd1a-P1 and zfsmyd1a-P2 primers. To characterize the expression of tv1 and tv2 isoforms from alternative splicing of exon 5, RT-PCR was carried out using two new sets of primers smyd1a-tv-P1/P2 or smyd1b-tv-P1/P2, respectively. PCR using these primers produced small PCR fragments covering the alternatively spliced exon 5.

Synthesis of Morpholino Antisense Oligos
Morpholino antisense oligos were synthesized by Gene Tools. The smyd1b ATG-MO was based on the target sequence surrounding the ATG start codon. The two smyd1a splicing blockers (E8I8-MO, E9I9-MO) were based on the sequence at the exon-8/intron-8 or exon-9/intron-9 junction, respectively. The standard control MO from Gene Tools was used as control.  Morpholino antisense oligos were dissolved in 1x Danieau buffer [22] to a final concentration of 0.5 mM. Approximately 1-2 nl (5-10 ng) of MO was injected into each embryo at 1 or 2 cell stages. For co-injection with transgenes, equal volumes of MO (1 mM) and DNA construct (100 mg/ml) was mixed and 1-2 nl was microinjected into each embryo at 1 or 2 cell stages. All the microinjection experiments and subsequent analyses were carried out at least three times with 100-150 embryos per sample in each experiment.

Construction of Zebrafish smyd1b-zfsmyd1a myc Transgene
To generate the smyd1b-zfsmyd1a myc transgene for muscle specific expression of myc-tagged zebrafish Smyd1a, the coding sequence of the zebrafish smyd1a was amplified with pfu DNA polymerase using the pGEM-T-zfsmyd1a plasmid as a template. The PCR was carried out using zfsmyd1a-P1 and zfsmyd1a-39-myc primers. The zfsmyd1a-P1 primer (59-AGCATGACCGTGGA-GAAGACGGAC-39) contained the ATG start codon. The zfsmyd1a-39-myc primer (59-CTAATTCAGGTCCTCTTCAGA-GATGAGCTTCTGCTCTGCTTTGATCTGCACTTGTTC-39) contained a myc-tag coding sequence and a stop codon. The PCR product containing the full-length coding sequence of a myctagged zebrafish smyd1a was phosphorylated with T4 kinase and directly cloned in the SmaI sites of the muscle-specific smyd1b-gfp vector [9]. The resulting plasmid was named smyd1b-zfsmyd1a myc .

Construction of smyd1b-mSmyd1 myc construct
Mouse smyd1a cDNA was cloned by RT-PCR using total RNAs from adult mouse skeletal muscles. The PCR was carried with the Advantage PCR kit using the mouse mSmyd1-ATG (BH1) and mSmyd1-39-myc primers. The PCR product was cloned into pGEM-T vector. The resulting plasmid was named pGEM-T-msmyd1a myc .
To generate the smyd1b-mSmyd1 myc transgene for muscle specific expression of myc-tagged mouse Smyd1, the coding sequence of a myc-tagged mouse Smyd1 was amplified from the pGEM-T-mSmyd1 myc plasmid by pfu DNA polymerase. The PCR was carried using the following primers: mouse Smyd1-ATG (BH1) and mouse Smyd1-39-myc. The PCR product containing the coding sequence of mouse Smyd1 and the myc-tag at the Cterminus was phosphorylated with T4 kinase and directly cloned into the SmaI sites of the muscle-specific smyd1b vector [9]. The resulting plasmid was named smyd1b-mSmyd1 myc .

Construction of Tol2-zfsmyd1a-EGFP Construct
The pTol2-smyd1a-EGFP construct was generated by cloning the zebrafish smyd1a coding sequence in frame upstream of the EGFP coding sequence in the T2A200R150G vector [23]. Briefly, the zebrafish smyd1a coding sequence without the stop codon were generated by PCR using pfu DNA polymerase (Stratagene). The PCR was carried out using the smyd1b-zfsmyd1a myc plasmid as a template together with the zfsmyd1a-F (59-gctAGATCTatgaccgtggagaagacggaccc-39) and zfsmyd1a-R (59-gctAGATCTgctttgatctgcacttgttctcc-39) primers. A BglII site was introduced at the 59 and 39 ends of the smyd1a coding sequence via the respective PCR primers. The PCR products were digested with BglII and then cloned into the compatible BamHI site of the T2A200R150G vector. The DNA sequence at the smyd1a and EGFP junction was confirmed by sequencing.

Whole Mount in Situ Hybridization and Immunostaining
Whole mount in situ hybridization was carried out using digoxigenin-labeled antisense probes as previously described [24][25][26]. The pGEM-T-zfsmyd1a plasmid was digested with NcoI and transcribed with Sp6 RNA polymerase to synthesize digoxigeninlabled antisense RNA probes. Immunostaining was carried out using whole mount zebrafish embryos (1-3 dpf) as previously described [5,27]. Briefly, zebarfish embryos were fixed in 4% paraformaldehyde (in PBS) for 1 hour at room temperature. The fixed embryos were washed for15 minutes 3 times in PBST (0.5% Tween-20 in 1x PBS). Embryos of 48 and 72 hpf were digested with 1 mg/ml collagenase for 45 and 75 minutes, respectively. In situ hybridization for MyoD, slow and fast myosin heavy chains were carried out using the respective digoxigenin-labeled probes as descried previously [28].