Slow myosin heavy chain 1 is required for slow myofibril and muscle fibre growth but not for myofibril initiation

Slow myosin heavy chain 1 (Smyhc1) is the major sarcomeric myosin driving early contraction by slow skeletal muscle ﬁ bres in zebra ﬁ sh. New mutant alleles lacking a functional smyhc1 gene move poorly, but recover motility as the later-formed fast muscle ﬁ bres of the segmental myotomes mature, and are adult viable. By motility analysis and inhibiting fast muscle contraction pharmacologically, we show that a slow muscle motility defect persists in mutants until about 1 month of age. Breeding onto a genetic background marking slow muscle ﬁ bres with EGFP revealed that mutant slow ﬁ bres undergo terminal differentiation, migration and ﬁ bre formation indistinguishable from wild type but fail to generate large myo ﬁ brils and maintain cellular orientation and attachments. In mutants, initial myo ﬁ brillar structures with 1.67 μ m periodic actin bands fail to mature into the 1.96 μ m sarcomeres observed in wild type, despite the presence of alternative myosin heavy chain molecules. The poorly-contractile mutant slow muscle cells generate numerous cytoplasmic organelles, but fail to grow and bundle myo ﬁ brils or to increase in cytoplasmic volume despite passive movements imposed by fast muscle. The data show that both slow myo ﬁ bril maturation and cellular volume increase depend on the function of a speci ﬁ c myosin isoform and suggest that appropriate force production regulates muscle ﬁ bre growth.


Introduction
Myosin are motor proteins that drive actin-based motility (Szent--Gyorgyi, 2004).There are more than 35 classes of myosin, each consisting of a myosin heavy chain (MyHC) and a number of smaller proteins, each class being defined by its distinct structure and function (Thompson and Langford, 2002;Sweeney and Houdusse, 2010).Class II myosins are found across metazoa, arose early in eukaryotic evolution (Richards and Cavalier-Smith, 2005;Thompson and Langford, 2002) and are composed of a heterohexamer containing two motor domains each made from a MyHC subunit, bound to a pair of non-identical myosin light chains (MyLCs) and linked by a coiled-coil MyHC tail that attaches to cargo.The sarcomeric myosins of striated muscle are a subset of class II MyHCs that form the regular filamentous array of thick myosin filaments interdigitating with parallel arrays of thin actin filaments (Hall et al., 1946;Hanson and Huxley, 1953).These structures are stabilised by M-lines and Z-lines that crosslink myosin thick filaments and actin thin filaments respectively, to form a sarcomere, many of which are linked in series to form a myofibril (Alberts et al., 2014;Howard, 1997).Myosin motor function within the bipolar arrays exerts force on actin thin filaments, shortening the sarcomeres, and thereby the myofibril, bringing about muscle contraction.
Structurally, a sarcomeric class II MyHC molecule has a conserved Nterminal motor head domain, which has actin-activated ATPase activity, followed by a rod region.The rod has a lever arm (containing the two MyLC binding sites), followed by a helical light meromyosin (LMM) region that dimerizes to form a coiled-coil, which then aligns with other heterohexamers to form a thick filament.The cooperative properties of these molecules within the myofibril determine the contractile characteristics of the muscle fibre, in particular its force/velocity relationship and energy efficiency (Irving, 2017).
Mammalian skeletal muscle fibres can be divided into slow-twitch and fast-twitch fibre types.Slow-twitch type I fibres contain slow MyHC (encoded by the MYH7 gene) which gives slow fibres a low maximal shortening velocity.Fast-twitch type II fibres generally contain MyHCs with one of three progressively faster maximal shortening velocities (IIa, IIx and IIb, encoded by MYH2, MYH1 and MYH4, respectively).Coordinated metabolic properties give slow-twitch fibres oxidative metabolism and slow activation and relaxation rates, whereas the faster fibres have glycolytic metabolism and are rapidly activated/ inactivated (Bottinelli, 2001;Schiaffino and Reggiani, 2011;Weiss et al., 2001).A number of murine adult fast MyHC genes have been genetically ablated, but lead to rather mild physiological phenotypes arising after 6 weeks of age that appear to be compensated by upregulation of other fast MyHC genes (Acakpo-Satchivi et al., 1997;Allen and Leinwand, 2001).Deletion of earlier developmental fast MyHC isoforms, on the other hand, can have stronger effects (Agarwal et al., 2020), https://www.mousephenotype.org/data/genes/MGI:1339712.Thus, the importance of individual fast MyHC genes depends on details of their function/s that are not immediately apparent from their known biochemistry.
Mammals also employ slow MyHC in cardiac muscle.MYH7 encodes the so-called β-cardiac slow MyHC that is expressed by ventricular cardiomyocytes.A genetically-linked duplicate gene, MYH6, expresses α-cardiac MyHC in adult atrial cardiomyocytes, conferring upon them distinct contractile properties (Alpert et al., 2002).Homozygous mutation of MYH6 or MYH7 slow MyHCs is lethal in mice due to early heart defects (Gifford et al., 2019;Jones et al., 1996).In contrast, heterozygous missense mutations in either gene are known to cause cardiac disease in humans that can be modelled in mice (Blankenburg et al., 2014;Gifford et al., 2019).To date, analysis of the effect of loss of slow MYHC function specifically in mammalian skeletal muscle has not been performed.
Slow myosins in non-mammalian vertebrates have undergone distinct gene duplication events, but are, nevertheless, found to be expressed primarily in cardiac and slow skeletal muscle (www.zfin.org).In zebrafish, one genomic locus contains at least four tandemly-arrayed smyhc genes that are not expressed in the heart (Elworthy et al., 2008;McGuigan et al., 2004).Elsewhere in the zebrafish genome, various other genes with greatest homology to mammalian MYH6/7 exist (named myh6, myh7, myh7l, myh7ba, and myh7bb) and are expressed in heart and/or skeletal muscle (Berdougo et al., 2003;Shih et al., 2015;Thisse and Thisse, 2004;Wang et al., 2011).Smyhc1 is the primary gene expressed in the earliest slow skeletal muscle of the zebrafish (Bryson-Richardson et al., 2005;Elworthy et al., 2008;Thisse and Thisse, 2004), which is the first functional muscle.Mammals also express slow myosin in their earliest muscle fibres (Kelly and Rubinstein, 1980).
Zebrafish slow muscle fibres derive from adaxial cells that undergo terminal differentiation in the medial somite, prior to migrating to the lateral myotome surface as fast muscle fibres form behind them (Blagden et al., 1997;Devoto et al., 1996;Hinits and Hughes, 2007).In each myotome, approximately 20 mononucleate slow fibres align anteroposteriorly and gradually increase in number as the myotome grows by fibre addition at the dorsal and ventral myotomal extremes (Barresi et al., 2001).Subsequently, slow fibres increase in size, maturing to form a specialised slow muscle with a uniform fibre type at the lateral midline (Rowlerson and Veggetti, 2001).Biomechanical arguments suggest that the positioning of oxidative slow fibres in this location enables continuous low-speed swimming at minimal energetic cost (Videler, 1993).The importance of the Smyhc1 protein for each of these developmental, cell biological and physiological processes is poorly understood.
Here we describe generation and characterisation of new loss of function smyhc1 mutants, which show significant differences in phenotype from some smyhc1 mutant alleles previously published.Mutant fish are viable and fertile but show a slow muscle specific motility defect during early life that recovers in the late larva at the time the smyhc2 gene is normally induced.The embryonic slow muscle fibres in mutants are formed, migrate, align and assemble immature myofibril-like structures containing other MyHC molecules, but fail to build an abundant array of large myofibrils or correctly increase cytoplasmic volume to form a fibre of the normal size and shape.We conclude that specific MyHC genes are required not only for appropriate motility, but also for the correct assembly of muscle fibre structure.

Generation of smyhc1 mutant alleles
To generate smyhc1 mutants, CRISPR/Cas9 genome editing with distinct gRNAs was used to target smyhc1 in the first and third coding exons.Several mutant alleles were generated, among these mutations smyhc1 kg179 and smyhc1 kg180 were chosen for analysis.Smyhc1 kg179 contains a 3 bp deletion and 1 bp insertion in exon 4 leading to a frameshift at amino acid 134 and an early stop codon at amino acid 148 after a 15 amino acid nonsense tail (Fig. 1A).Smyhc1 kg180 contains a 4 bp deletion in exon 2 leading to a frameshift at amino acid 28 and an early stop codon at amino acid 32 after a 5 amino acid nonsense tail (Fig. 1A).In situ RNA hybridisation for smyhc1 mRNA in smyhc1 mutant and their siblings from smyhc1 kg179/þ and smyhc1 kg180/þ in-crosses reveal nonsense-mediated decay of mutant smyhc1 kg179 and smyhc1 kg180 mRNA at 24 hpf.In a randomlyselected sample from a smyhc1 kg179/þ incross, 4/21 had reduced expression and were shown to be mutant; among 17 'normal' expressors 11/20 were heterozygotes and 6/20 were wild type upon sequence genotyping.From the smyhc1 kg180/þ incross, 6/29 embryos had low expression and were shown to be mutant; among 23 normal expressors 15/23 were heterozygous and 8/23 were wild type upon sequence genotyping.(C) Quantitative RT-PCR for smyhc1, smyhc2 and smyhc3 on pools of ten 2 dpf embryos from a smyhc1 kg179/þ incross that had been sorted into immotile mutant and motile siblings at 1 dpf.Each symbol represents a separate reverse transcription with symbol shapes distinguishing biological replicate lays from separate parents.The dashed line represents the motile sibling level (D).Maximum intensity projections of Slow MyHC-stained 24 hpf embryos showing region centred on somite 17 (S17).Wild-type siblings (left) show slow fibre staining and smyhc1 kg179/kg179 and smyhc1 kg180/kg180.mutant siblings show no slow fibre stain (right).
Whole mount in situ mRNA hybridisation and quantitative RT-PCR showed that nonsense-mediated mRNA decay of smyhc1 mRNA occurs in mutants (Fig. 1B and C).Whereas smyhc1 mRNA was readily detected in slow muscle of sibling embryos at 24 hpf, much less signal was observed in about a quarter of embryos, the genotype of which were subsequently confirmed as smyhc1 kg179 and smyhc1 kg180 homozygous mutants (Fig. 1B).Similarly, immotile embryos contained less smyhc1 mRNA than motile siblings (Fig. 1C, Supplementary Fig. 1).Thus, the mutations cause premature ribosomal termination leading to mRNA decay and therefore little mutant protein will be produced.Indeed, mutants lacked detectable Smyhc1 protein, whereas their siblings expressed in approximately 20 superficial slow fibres (SSFs) and several fibres deeper in the horizontal myoseptum, the slow muscle pioneers (Fig. 1D).We conclude that smyhc1 kg179 and smyhc1 kg180 are likely strong loss of function, and possible null, alleles.

smyhc1 mutants are viable and fertile
Lays from smyhc1 kg179/þ and smyhc1 kg180/þ in-crosses were sorted based on immotility at the 23 somite stage (23ss) and examined under a bright-field microscope at 1 to 5 dpf to identify any morphological and skeletal muscle defects.Mutant larvae were detected at the expected frequency, indicating that null mutations in smyhc1 are not embryonically lethal.In many separate in-cross lays of smyhc1 kg179/þ , one lay from smyhc1 kg180/þ , and one transheterozygote lay, no change in head, somite, tail, yolk sac, fin, pigmentation, or body length was observed (Fig. 2A).However, as previously reported in studies of antisense morphants and distinct smyhc1 mutants (Codina et al., 2010;Li et al., 2020;Whittle et al., 2020;Xu et al., 2012), for both kg179 and kg180 immotile embryos were genotyped as mutant at 24 hpf (Fig. 2B; 21/22 and 12/12 immotile embryos, respectively).By 34 hpf, however, all embryos were motile irrespective of genotype, explaining the ability of mutants to hatch by 3 dpf and swallow air and inflate the swim bladder by 4 dpf (Fig. . 2A and  3).Thus, lack of Smyhc1 in kg179 and kg180 mutants leads to fish that appear immotile at early stages but morphologically normal.
To determine whether smyhc1 mutation affects survival beyond 5 dpf and into adulthood, F3 embryos generated from smyhc1 kg179/þ and smyhc1 kg180/þ in-crosses were reared.One hundred randomly-selected embryos from each cross were monitored for 4 months.Growth of all siblings from crossed fish were divided in tanks of 50 larvae of mixed sex and genotype to ensure competition.At 4 months post-fertilisation (mpf), 82% and 94% survival was observed from smyhc1 kg179/þ and smyhc1 kg180/þ crossed fish, respectively.Genotyping of a randomlyselected subset of 4 mpf adult fish revealed that both lays did not differ from expected Mendelian ratios (Fig. 2C, p > 0.05 Х 2 -tests), although kg180 had fewer heterozygotes than expected (p ¼ 0.016).Mutants remained indistinguishable from siblings during later life (Fig. 2D).Length and weight measurements were taken on mutant fish reared with their siblings at 4 mpf.Fish from the smyhc1 kg180/þ incross were overall smaller in length and weighed less than fish from the smyhc1 kg179/þ incross, possibly reflecting uncontrolled genetic background or, perhaps more likely, environmental rearing conditions.Nevertheless, when comparing between siblings, no significant difference in length or weight was observed between sex-matched wild-type, heterozygote or mutant fish (Fig. 2E).Homozygous smyhc1 kg179/179 and smyhc1 kg180/kg180 mutant males and females were fertile.We conclude that smyhc1 is a nonessential gene for life in an aquarium.

Movement defects persist in smyhc1 mutant
We next determined whether the defective movement in smyhc1 mutants persists beyond 24 hpf.Previous studies have shown immotility in zebrafish at 24 hpf in smyhc1 morphants or mutants (Codina et al., 2010;Li et al., 2020;Whittle et al., 2020;Xu et al., 2012), when movements are primarily driven by early-formed slow fibres (Drapeau et al., 2002).Smyhc1 kg179/þ were in-crossed to generate wild type, heterozygous and homozygous mutant embryos.Microscopical observation of anterior somites at or shortly after 24 hpf revealed minor twitching movements in mutants, likely driven by nascent fast fibres, which do not express smyhc1, but begin to assemble striated myofibrils at this stage (Hinits and Hughes, 2007).Chorions were removed at 24 hpf and tail-coiling movement was analysed to separate motile from immotile fish, and their movement assayed over subsequent days (Fig. 3A).At 48 hpf, immotile mutants had regained tail muscle motility and superficially appeared to move similarly to wild-type and heterozygous siblings.Nevertheless, when swimming velocity upon touch stimulation was assayed, homozygous smyhc1 kg179/kg179 mutants showed reduced swimming velocity (136 mm s À1 ) compared to their wild type (p ¼ 0.011) and heterozygous siblings (285 mm s À1 ; p ¼ 0.016) (Fig. 3B).At 5 dpf, smyhc1 kg179/kg179 mutants had reduced swimming velocity (375 mm s À1 ) compared to their wild-type and heterozygous siblings (543 mm s À1 ) (Fig. 3C).From 17 to 30 dpf, no difference between smyhc1 kg179/kg179 and their siblings was apparent (Fig. 3D-F).Thus, loss of Smyhc1 results in reduced swimming capacity in young fish.
To examine motility driven by slow fibres, 2 dpf embryos were treated with 50 μM N-benzyl-p-toluene sulphonamide (BTS), an inhibitor for fast muscle myosin II (Cheung et al., 2002;Li and Arner, 2015) and their swimming velocity was recorded (Fig. 3B).All fish showed strongly reduced swimming velocity after treatment with BTS (Fig. 3B-F).However, at 2, 5, 17 and 20 dpf homozygous smyhc1 kg179/kg179 mutants were more affected than their wild-type and heterozygous siblings, showing either little twitching or no movement (p < 0.000001; Fig. 3B-E).At 30 dpf, BTS reduced swimming velocity to a similar extent as at 20 dpf, but there was no significant difference between homozygous smyhc1 kg179/kg179 mutants and their siblings (Fig. 3F).Thus, slow fibre motility remains compromised in young mutant larvae, before recovering during the fourth week of life.

Defective sarcomere organisation in mutant slow fibres
To understand the basis for motility defects and recovery in smyhc1 mutants, we next investigated sarcomere assembly in specific muscle cell populations (Fig. 4).Using antibody S58, which is known to detect specifically slow MyHC isoforms in a variety of species (Crow and Stockdale, 1986;Devoto et al., 1996), we observed a lack of signal in mutants in regions known to express smyhc1 in wild type fish (Fig. 4A).At 3 dpf mutants entirely lacked slow MyHC in adaxial cell-derived SSFs in trunk and tail whereas, depending upon the position along the rostrocaudal axis, siblings had 10-25 SSFs on the surface of each myotome (Fig. 4A).The same mutant larvae, however, retained S58 staining in cardiac, cranial and the somitically-derived anterior hypaxial muscles that extend over the yolk and into the head forming the sternohyoid muscle (Fig. 4A).Moreover, several small groups of fibres known to express smyhc2 (Elworthy et al., 2008) retained S58 signal in specialised muscle at the dorsal edges of anteriormost somites, the dorsal and ventral edges of the caudalmost somites and in rare cells at the dorsoventral edges and horizontal myoseptum of all somites (Fig. 4A).We conclude that smyhc1 mutation removes all slow MyHC from the adaxially-derived SSFs, but does not affect Smyhc2 protein accumulation.Fig. 2. Zygotic smyhc1 mutation reduces embryo motility but survive to adulthood.A) Representative bright-field images of 1,2,3,4 and 5 dpf larvae from smyhc1 kg179/þ and smyhc1 kg180/þ heterozygote in-crosses.Fish are shown anterior towards the left and dorsal upwards with genotyped heterozygotes and mutants below their respective wild type siblings.Scale bars ¼ 0.5 mm.B) Randomly selected larvae were dechorionated at 24 hpf and examined for presence or absence of tail coiling movement in smyhc1 kg179/þ in-cross (n ¼ 82) and in smyhc1 kg180/þ in-cross (n ¼ 52).The genotype of the fish was determined after the examination by DNA sequencing.Fish derived from several in-crosses of smyhc1 kg179/þ (n ¼ 101) and smyhc1 kg180/þ (n ¼ 100) fish were reared with their siblings and genotyped at 4 months post-fertilization (mpf).C) Adults showed the expected Mendelian ratios at 4 mpf.Fish numbers above each bar.D) Adults at 12 mpf.Scale bars ¼ 1 cm.E) Length and mass at 4 mpf of genotyped siblings showed no significant difference between genotypes.Small symbols indicate individuals, large symbols the mean for each sex and genotype AE S.E.M. Overall length and weight were less in adult fish from the smyhc1 kg180/þ lay compared to adult fish from the smyhc1 kg179/þ lay, a difference that may reflect an uncontrolled environmental or genetic background effect.S1.Note the log 10 scale on the Y-axes to show both control Fish Water and BTS-treated data accurately.Summary Tukey post hoc statistics for separate two way ANOVAs on Fish Water and BTS-treated are shown below.In each case, the upper table shows the overall significance, whereas the lower table shows the Tukey post hoc significance levels for the overall effects of genotype independent of age.Values of within-age Tukey post hoc test comparisons are indicated above graphs (black lines and p-values), none of which had p < 0.05 in Fish Water.To mitigate the effect of changing variance due to the increase in velocity with age and large effect of BTS, individual one way ANOVAs on genotype were performed at each age and Tukey post hoc tests supported the reduced velocity of mutants in Fish Water at 2 and 5 dpf only (green lines and p-values).
To understand the effect of loss of slow MyHC in SSFs we examined other components of the sarcomere (Fig. 4B).Compared to their siblings, the young somites of 24 hpf mutant embryos that entirely lacked slow MyHC showed disruption of thin filament structures, with absence of Zlines marked by α-actinin and a great reduction in F-actin thin filament arrays (Fig. 4B).F-actin present within the slow fibre region of mutants was concentrated at the vertical myosepta along with αÀactinin (Fig. 4B).F-actin also appeared to have irregular periodicity within the mutant myotome, although no αÀactinin periodicity was observed (Fig. 4B, yellow boxes).Thus, Smyhc1 appears to be necessary for efficient assembly of both thick and thin filaments within nascent slow muscle fibres.
Subsequently, however, mutant larvae gained slow MyHC immunoreactivity (Fig. 4C).In smyhc1 kg179 mutants at 72 hpf, S58 positive SSFs began to appear at dorsal and ventral somitic extremes and the horizontal myoseptum (Fig. 4C).At each somite extreme between zero and three nascent S58-reactive fibres formed between 1 and 3 dpf.Dual staining with A4.1025, an antibody that recognizes a conserved epitope in the head of all sarcomeric MyHCs, primarily recognized the abundant slow MyHC in SSFs in siblings, and clearly showed that fast fibres had normal sarcomere assembly within the myotome of smyhc1 kg179 mutants (Fig. 4C).Loss of staining in SSFs was observed with the F59 antibody, which also detects zebrafish Smyhc1 strongly and other MyHCs more weakly (Fig. 5A).Thus, slow MyHC isoforms distinct from Smyhc1 are recognized by the S58 and F59 antibodies and accumulate in specific fibres within the growing myotome.

Smyhc1 is required for SSF stability, cytoplasmic organisation and growth
To investigate the fate of SSFs that lack slow MyHCs, we next bred the mutants onto the Tg(smyhc1:EGFP) i104 stable transgenic line that labels SSFs due to EGFP expression from the smyhc1 ATG start codon (Elworthy et al., 2008).Mutants lacked MyHC in almost all slow fibres, but expressed in fast fibres (Fig. 5A).Transheterozygotes of the two smyhc1 mutant alleles contained slow muscle pioneer fibres at the horizontal myoseptum and normal numbers of migrated SSFs located on the lateral surface of the myotome throughout the body axis (Fig. 5A and B).The SSFs in mutants lacked slow MyHC, but underlying obliquely-oriented fast fibres showed weak immunoreactivity with F59 (but not S58), as previously reported (Fig. 5A; (Devoto et al., 1996)).At 1-2 dpf, SSFs in mutants were initially correctly orientated parallel to the anteroposterior body axis and attached to the vertical myosepta at both anterior and posterior somite boundaries (Fig. 5A lower, B).The volume of slow muscle was unaffected at early stages (Fig. 5D).As mutant larvae matured, EGFP-marked SSFs persisted on the myotome surface until 7 dpf but, by 8 dpf, some had lost contact with one or both vertical myosepta, becoming orientated parallel to underlying superficial fast fibres, and showed a greatly disorganized cell structure and reduced fibre volume (Fig. 5B-D).
The internal structure of SSF cytoplasm was dramatically altered in smyhc1 mutants.SSFs in wild type and heterozygote siblings had three main sarcoplasmic structures revealed by the distribution of EGFP signal (Fig. 5C).First, an ordered myofibrillar array filling most of the fibre volume.Sarcomeres consisted of narrow bright I-band regions and wider and dimmer A-band domains with a markedly brighter central M-line.
Sarcomere length in our live preparations was 1.95 AE 0.03 μm (mean AE SD, n ¼ 10 sarcomere lengths measured in each of 10 fibres at 8 dpf).Second, the sarcoplasm contained a variety of vacuoles that excluded EGFP, presumably membranous organelles.Numerous smoothly rounded or oblate vacuoles were located near fibre ends and around the nucleus.Less frequent elongated vacuoles appeared to extend anteroposteriorly between myofibrils, some being up to 10 μm long.Most SSFs had numerous vacuoles in clusters on their lateral surface, some with complex (caption on next column) heads), the mutant showed prominent nuclei, a thin disorganized cytoplasm filled with vacuoles between which ran immature and poorly-aligned material with closer striations (open arrowheads).Some fibre regions in sibling were devoid of myofibrils (arrowhead) and contained prominent and extensive vacuoles (Inset yellow box is a slice 5 μm more superficial to the dashed yellow box region displaying the abundant vacuolar structures at the myotome surface that are less apparent deeper within the SSF layer.D) Quantification of gradually reduced EGFP cell volume of SSFs in smyhc1 kg179 ;Tg(smyhc1:EGFP) mutants compared to their siblings at the same age.Mean (AE SEM when N allowed, calculated by to include propagation of error in both sibling and mutant measures) of the proportion for scanned segment of at least three somites centred on somite 18.On each column, fraction is number of mutant/sibling fish analysed with identical scan and post-processing parameters.Thus, at 1-2 dpf four embryos had SSFs of similar volume to those in five of their non-mutant siblings whereas, at 6-8 dpf, nine mutant larvae had SSFs about half the size of those in 11 non-mutant siblings.P-values above columns represent t-test on measured volumes of mutants versus siblings at each age.E,F).Confocal stacks (E) and slices (F) of 2 h-old somites at 23ss showing the emergence of small MyHCcontaining myofibrils in mutant SSFs (arrowheads).Crosshairs in F indicate slice planes.Asterisks mark the initiation of deep fast MyHC accumulation in more anterior somites.Bars 50 μm (A,B,C upper, E) and 10 μm (C lower, F). multilobular forms (Fig. 5C inset).Narrow cytoplasmic bridges extending between vacuoles frequently aligned with Z-or M-lines.Third, some SSFs appeared to have extended regions containing rather uniform EGFP signal lacking myofibrils but containing vacuoles (Fig. 5C).SSFs in mutants were thin and contained numerous vacuoles that were smaller than those in siblings, even when the SSF was correctly aligned and anchored to both myosepta (Fig. 5C).Moreover, mutant SSFs lacked myofibrillar structures with the characteristic sarcomeric banding.Instead, some regions of cytoplasm had more uniform EGFP signal that exclude vacuoles and appeared to constitute fibrillar structures.The fibrillar EGFP signal was interrupted by narrow dark striations (Fig. 5C), with a striation repeat length of 1.69 AE 0.07 μm, 13% shorter than the sarcomeres in siblings, and more variable.Mutants did have rare small SSFs at the horizontal myoseptum and dorsal and ventral extremes of the myotome that showed a more normal sarcomeric pattern (Fig. 5C), correlating with the location of residual fibres expressing another slow MyHC isoform (Fig. 4A).Nuclear morphology and positioning in mutants was not distinguishable from that in siblings, in which the nuclei were routinely localized at the lateral surface of the fibre adjacent to dermomyotomal cells and the periderm (Fig. 5C).Quantification of total SSF and MP fibre volume across a series of adjacent somites centred on somite 18 revealed that, on average, mutant fibres had only grown to around half the volume of slow fibres in siblings (Fig. 5D).In summary, without Smyhc1, SSFs have reduced fibre volume, altered cytoplasmic vacuolation, lack mature myofibrils and fail to maintain anchorage to vertical myosepta.The underlying fast fibres, unlike the SSFs, are not orientated parallel to the anteroposterior body axis (Fig. 4C and 5A).However, SSFs lacking slow MyHC still aligned anteroposteriorly, indicating that they did not convert to a fast myogenic programme.Nevertheless, in young somites, signs of myofibrils with sarcomeric organisation were apparent in the GFP distribution in mutant SSFs (Fig. 5A, open arrowheads).Shortly after mutant SSFs complete their migration to the lateral myotome surface, sarcomeric MyHC was detected in a single myofibril-like structure in many SSFs using the broad-spectrum MyHC A4.1025 antibody (Fig. 5E  and F).A similar result was obtained with a second broad-spectrum MyHC antibody (MF20; data not shown).Mutant incross lays were examined for actin structure at 24 hpf with phalloidin-Alexa488.In the absence of Smyhc1 protein, we observed that actin filament organisation was severely defective (Fig. 4A).In wild type siblings, F-actin was organised into sarcomeric thin filament units arrayed at regular intervals along the slow muscle fibre length into myofibrils.In mutant embryo slow fibres, by contrast, overall F-actin signal was reduced and disrupted thin filament organisation was observed.Nevertheless, there were a few fibres in some regions with organised F-actin filaments (Fig. 4B).We conclude that an additional MyHC is expressed in SSFs and can initiate myofibril formation.Nevertheless, without Smyhc1, myofibril growth is severely restricted.

Discussion
The data presented lead to three new insights into myogenesis.First, we resolve some contradictory findings derived from six smyhc1 putative null mutant alleles.Second, we find that myosin accumulation is required in order for muscle fibres to grow, thereby illuminating a novel and likely important aspect of muscle fibre growth control.Third, we provide evidence that in zebrafish, as in amniotes, an embryonic fast myosin initiates myofibril formation in nascent slow muscle fibres.

Allelic differences in smyhc1 mutants
Combining our two new mutant alleles with those previously reported (Li et al., 2020;Whittle et al., 2020), predicted protein-truncating mutations have been made at four positions within Smyhc1 (Fig. 1A).Our early truncations in the first or third coding exons of smyhc1 show nonsense-mediated mRNA decay, as do the alleles characterised by Li et al. (2020), which also have mutations in the third coding exon.These alleles are therefore likely to be functional null.Indeed, our findings are in most respects congruent with the conclusions of Li et al. (2020).Namely, we observe complete loss of slow MyHC in SSFs at early stages, with slow fibre-specific sarcomere and motility defects and recovery in the 20-30 dpf period as other smyhc genes commence expression and generate Smyhc proteins in SSFs.All four smyhc1 mutants lack detectable slow MyHC protein in fibres that normally express smyhc1 mRNA, suggesting that, as we show for smyhc2 and smyhc3, despite the presence of the genetically-linked and highly homologous (93-96% amino acid identity) smyhc2-4 genes, no compensatory upregulation occurs.The similar homozygous phenotypes observed in all early truncation alleles also suggests that, although the guide RNA employed by Li et al. (2020) is perfectly matched not only to smyhc1 but also to smyhc2-4 and the mutagenic status of these linked genes was not reported, mutations in smyhc2 and smyhc3, at least, may not be present in the mb16-18 alleles.Our guide RNAs were designed to sequences specific to smyhc1.
One important difference between our findings and those previously published (Li et al., 2020;Whittle et al., 2020) is the lack of significant larval death in our mutants.Although the reason is unclear, we suggest the better survival of our mutants does not arise from allele-specific effects but rather from genetic background or differences in husbandry in the nursery.One such difference is that our fish are fed on rotifers, whereas the larvae carrying mb16-18 alleles were fed on paramecium (Li et al., 2020).Feeding and survival of smhyc1 stl583 homozygotes was not described, but the allele did worsen survival on the background of a semi-dominant truncating allele (Whittle et al., 2020).As Li et al. (2020) showed feeding was variably reduced in their smhyc1 mb17 homozygotes, the effect of defective motility on efficiency of food intake could differ with prey motility, size and other parameters.
In contrast to the similarities between the kg179, kg180 and mb16-18 alleles, the stl583 allele, which is predicted to truncate Smyhc1 towards the end of the myosin S1 head domain, has a number of reported differences.Perhaps most striking is the complete paralysis of smyhc1 stl583 homozygotes until 48 hpf (Whittle et al., 2020).Whereas reduced motility at or before 24 hpf allowed selection of embryos mutant for the earlier truncations (Fig. 2B and (Li et al., 2020)), once fast muscle had matured we observed movements in homozygotes for each of our alleles such that they were not readily distinguishable from siblings.Nevertheless, if assayed quantitatively, the early truncated alleles had motility defects that persisted beyond the first week (Fig. 3 and (Li et al., 2020).This reduction in swimming velocity was recovering by 17 dpf (Fig. 3D), an age at which Li et al. (2020) reported increased death of mutant larvae.Indeed, because we found that a continued motility defect became readily apparent by eye when fast myosins were inhibited pharmacologically (Fig. 3B-E), we attribute the effective movement of late embryos and young larvae to fast muscle contraction.One intriguing possibility is that the smyhc1 stl583 allele is not null but expresses a partially-functional myosin head fragment that interferes with motility up to 48h hpf.Although smyhc1 stl583 homozygotes lack of F59 immunoreactivity, it may be significant that the F59 epitope is in an unmapped region of the myosin S1 head (Crow and Stockdale, 1986) and could therefore be absent in a truncated protein.Use of a head antibody that maps upstream of the stl583 mutation, such as A4.1025 (Dan-Goor et al., 1990), might reveal the presence of truncated Smyhc1 protein fragments that could have a hypermorphic or neomorphic effect.Such an effect may also explain the variable spinal curves in smyhc1 stl583 adults, a phenotype we did not observe and was not reported by Li et al. (2020) in their surviving adults.The possibility of gain of function effects in the smhyc1 stl583 allele may reduce the concern expressed by Whittle et al. (2020) over the use of non-specific antisense oligonucleotide approaches to remove toxic gain-of-function MyHCs in human conditions (but see further discussion below).

Muscle fibre growth depends on myosin
The mechanisms coordinating the many cell biological processes required for cell growth are poorly understood.For example, in the century since D'Arcy Thompson formulated the problem in print (Thompson, 1917), how cells match their plasma membrane surface area to the volume of their cytoplasm so as to control cell shape, and how cells generate the appropriate quantity of cytoskeletal elements to support the size and shape of the cell remain mysteries.Skeletal muscle fibres may provide an instructive example of cell volume control because a) their form is constrained to an approximate cylinder by their contractile function, b) their cytoplasm is essentially filled with a specialised actomyosin cytoskeleton and c) they are one of the few cell types that can both greatly increase and decrease in size during adult life.Despite these propitious characteristics for insight into the cell size problem, muscle has been hard to analyse because most fibres are multinucleate.Mutation of murine MYH4, encoding MyHC IIB, the major myosin in mouse limb muscles, led to reduced muscle size, but this was accompanied by reduced fibre number, compensatory hypertrophy and an unquantified change in nucleation (Allen et al., 2001).By analysing the early steps in growth of the unusual mononucleate slow fibres of the zebrafish, we reveal the dependence of several other aspects of cell growth on the accumulation of Smyhc1, the major MyHC of these fibres.Smyhc1 mRNA accumulation precisely parallels the initiation of slow muscle fibre terminal differentiation (Hinits and Hughes, 2007;Hinits et al., 2007).Slow muscle fibres assemble myofibrils prior to their migration (Jana Koth and Simon M. Hughes, unpublished) yet, despite lack of their major MyHC, SSFs migrate and orientate on the surface of the myotome in smyhc1 mutants in wild type numbers (Fig. 1 and (Li et al., 2020)).No compensatory accumulation of other MyHC isoforms was detected and myofibrils were greatly reduced.Despite attaining normal fibre length, the cross-sectional area, and thus volume, of SSFs fails to increase in the absence of normal myofibril assembly in smyhc1 mutants (Fig. 5).By 8 dpf, some SSFs become mis-orientated in smyhc1 mutants.Nevertheless, we and others have observed recovery of SSF morphology, function and presumably cell size in smyhc1 mutants once Smyhc2 and Smyhc3 become expressed (Li et al., 2020;Whittle et al., 2020).We hypothesise, therefore, that muscle fibre growth control can be likened to the situation with shopping and bags; the quantity of 'purchased' cytoskeleton dictates the extent of expansion of the sarcolemmal 'bags'.How MyHC content or myofibril assembly is assessed and membrane trafficking and/or osmotic balance thereby controlled is of great interest.

3.
3. An 'embryonic' fast MyHC in fish slow muscle fibres SSFs were present but lacked slow MyHC in smyhc1 mutants.Nevertheless, muscle fibres in three somitic regions continued to show slow MyHC immunoreactivity in mutants.S58 and F59 immunoreactivity were presented in thin muscle fibres at the dorsal and ventral somitic extremes, the horizontal myoseptum and the posterior end of the tail (Fig. 4) These slow MyHC-expressing somitic fibres match in number, timing and smyhc2-3 expression the non-adaxially-derived slow fibres described previously (Barresi et al., 2001;Elworthy et al., 2008).In addition, other slow MyHC-immunoreactive cells persisted in mutants in heart and specialised cranial and somite-derived muscles in which other slow MyHC genes, such as myh6, myh7 and myh7l, are expressed (Fig. 4A) (Berdougo et al., 2003;Elworthy et al., 2008;Shih et al., 2015;Thisse and Thisse, 2004;Wang et al., 2011).As up-regulation of alternative slow MyHC genes was not observed in smyhc1 mutants (Li et al., 2020), we conclude that the residual striated myofibril-like organisation in SSFs does not require slow MyHC protein.
By using other anti-sarcomeric MyHC antibodies with broad species and isoform cross-reactivity (namely, A4.1025 and MF20), however, we observed striated MyHC immunoreactivity in the SSFs of smyhc1 mutants.This persistent MyHC was located in myofibril-like structures that also contained F-actin arrays and were compact enough in some regions to exclude the cytoplasmic EGFP that marked the SSFs, making the myofibril-like structures visible in live larvae.Such structures had reduced sarcomere length, as reported in the residual actin and α-actinin-containing structures in other smyhc1 mutants analysed (Li et al., 2020;Whittle et al., 2020).It is possible that the gene responsible for this MyHC accumulation in SSFs lacking Smyhc1 is an myhz1, which may also be expressed in adaxial SSF-precursors from the initiation of their terminal differentiation (Hinits and Hughes, 2007;Hinits et al., 2007).Thus, the ability of SSFs in smyhc1 mutants to persist and recover function once alternative slow MyHCs start to express around ~20 dpf may reflect the resumption of a developmental progression prevented by the lack of Smyhc1.We hypothesise that, in the presence of Smyhc1, this progression would allow seamless maturation of sarcomere length and expansion of myofibril size in the early embryo SSFs.
There are some similarities between the defective SSFs in smyhc1 mutants and those arising after Mef2 knockdown (Hinits and Hughes, 2007).Reduced Mef2 activity leads to failure to incorporate Smyhc1 into myofibrils, leaving a 'dumbbell' morphology fibre with a single thin myofibril and a cloud of unincorporated Smyhc1 at each end (Hinits and Hughes, 2007).In the absence of Smyhc1, a similar nascent myofibril structure perdures.As Mef2 activity is not required for smyhc1 expression in SSFs, it seems that both Smyhc1 and at least one other Mef2-dependent protein are required for growth of myofibrils.An obvious possibility is the embryonically-expressed fast MyHC present in smyhc1 mutants.
The presence of a distinct MyHC isoform in nascent fibres in zebrafish is reminiscent of the embryonic MyHC, expressed from the MYH3 gene, in nascent mammalian fibres irrespective of the eventual mature fibre type (Condon et al., 1990;Karsch-Mizrachi et al., 1989;Schiaffino et al., 2015;Whalen et al., 1981).Knockout of murine Myh3 leads to reduced viability, muscle weight and fibre cross-sectional area and a relative increase in slow myogenesis, accompanied by knock-on effects on muscle stem cells (Agarwal et al., 2020).Viable Myh3 mutants are compensated by up-regulation of other fast MyHC genes (Agarwal et al., 2020).Our finding of reduced fibre size when Smyhc1 is missing followed by compensation when additional Smyhc proteins accumulate suggests that loss of either of two MyHC isoforms in nascent myofibres may have similar effects.In this light, it is interesting that point mutations in the fast-class MYH3 gene cause several arthrogryposes (Toydemir et al., 2006).As similar mutations in the slow-class smyhc1 gene cause defects in zebrafish (Whittle et al., 2020), it will be interesting to determine whether mutation of the remaining MyHC in SSFs has similar effects.

Zebrafish lines and maintenance
All lines used were reared at King's College London on a 14/10 h light/dark cycle at 28.5 C with adults kept at 26.5 C, with staging and husbandry as described in (Westerfield, 2007).AB stocks were maintained by breeding at least ten pairs at each generation.Embryos/larvae were reared at 28.5 C in the dark, except for periods outside the incubator until 6 dpf when rotifer feeding commenced.Smyhc1 kg179 and smyhc1 kg180 mutant alleles on AB background were genotyped by High Resolution Melt Analysis (HRM), followed by sequencing using primers indicated (Table 1).Briefly, HRM primers amplified DNA fragments of 107 bp, 103 bp, 115 bp and 113 bp and sequencing primers amplified DNA fragments of 603 bp, 599 bp, 280 bp and 278 bp smyhc1 þ/þ , smyhc1 kg179 , smyhc1 þ/þ and smyhc1 kg180 alleles, respectively.Tg(smyh-c1:EGFP) i104 was back-crossed onto smyhc1 kg179 .All experiments were performed on zebrafish derived from F3 or later generations, in accordance with licences held under the UK Animals (Scientific Procedures) Act 1986 and later modifications and conforming to all relevant guidelines and regulations.
HRM-selected AB wild-type fish were DNA sequenced over the target loci to avoid polymorphisms, crossed and the resulting embryos injected with 1 nl containing 40 pg (gRNA2) or 80 pg gRNA1, 300 pg spCas9 protein with 0.03% rhodamine dextran to select injected embryos.Ten 48 hpf larvae were analysed by HRM to verify mutagenesis, their F0 siblings grown to adulthood and outcross F1 progeny analysed for transmission by HRM.Mutant loci of F1 heterozygotes were sequenced to identify F0s transmitting mutations of interest, F1 siblings grown to adulthood and F1 heterozygotes identified by HRM and sequencing of fin-clip DNA.Subsequent generations were bred by outcross to wild-type AB selected as non-polymorphic at the target locus.

RNA extraction and RT-qPCR
Protocol was previously described (Ganassi et al., 2018;Kelu et al., 2020).Pools of five (5 dpf) or 10 (1, 2 dpf) embryos were snap-frozen in liquid nitrogen with minimal liquid.Samples were sonicated on ice using 350 μL RNeasy Lysis Buffer (Buffer RLT) from RNeasy Mini Kit (Qiagen), lysates treated with RQ1 RNase-free DNase (Promega) and column-purified using RNeasy Mini Kit (Qiagen) according to manufacturers' instructions.Purified total RNA (500 ng) was subsequently reverse transcribed using High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems) with both random and oligo-dT primers (Invitrogen) and RNase inhibitor (NEB).For RT-qPCR, technical triplicates were performed on 5 ng of cDNA using Takyon Low ROX SYBR 2X MasterMix blue dTTP (Eurogentec) using ViiA™ 7 Real-Time PCR System (Thermo Fisher Scientific) in a 10 μL reaction using primers listed (Table 3).Specific primers were designed to target unique regions in the 5 0 -UTR of smyhc1-3 transcripts using Primer-BLAST (Ye et al., 2012) (Supplementary Fig. 1A).An additional primer set targeting 3 0 -UTR of smyhc1 was previously described (Li et al., 2020) (Supplementary Fig. 1), but with a modification in the reverse primer to account for a SNP.eef1a1l1 was used as a housekeeping gene as its expression is high and stable during zebrafish development (McCurley and Callard, 2008).ΔCt was calculated by subtracting the triplicate mean Ct value of eef1a1l1 from that of the target gene.ΔΔCt of each target gene was then calculated by subtracting the mean ΔCt of target and housekeeping samples, and relative gene expression calculated as 2 ÀΔΔCt (Livak and Schmittgen, 2001), followed by normalisation to the mean of 1 dpf motile control samples.

Adult fish analysis
Siblings (4 mpf) from single heterozygote in-crosses were anaesthetised with tricaine (Sigma Aldrich), sexed, blotted dry and weighed on an Ohaus YA102 balance, standard length measured against a ruler and fin-clipped for sequence genotyping.Weights and lengths were compared by unbalanced two way (genotype and sex) ANOVA with Tukey post hoc tests (GraphPad PRISM 8).

Swimming velocity test
Chorions were removed from 1 dpf siblings from single heterozygote in-crosses.Individual embryos were analysed for presence or absence of movement in response to touch stimulus.At 2 dpf and later, embryos/ larvae from 3 to 4 lays were stimulated by touch using a needle and video recorded using Leica MZ16 with Olympus DP70 camera and DP Controller at 30.53 frames.sÀ1 before and after treatment with 50 μM BTS for 10 min.Fish velocity measured using Tracker (https://physlets.org/tracker/).Embryos/larvae were retrospectively genotyped.Statistics were analysed in RStudio 1.4.1103separately for control Fish Water and BTS-treated fish with factors Age, Genotype, Lay and Age*Genotype interaction.Velocity measurements were log-transformed to ensure equal variance (homoscedasticity) based on diagnostic plots of unexplained variance.The Akaike information criterion was used to select the most informative ANOVA, which was two way ANOVA on Age and Genotype with interaction.No effect of Lay was detected.Graphpad Prism was used for one way ANOVA.

Fig. 1 .
Fig. 1.Genome editing generates likely null alleles of zebrafish smyhc1.(A) Schematic of smyhc1 gene, mRNA sequences and proteins showing nature of smyhc1 kg179 and smyhc1 kg180 mutant alleles.Underline indicates the guide RNA sequence, red and blue bases were deleted and inserted, respectively.Most of the S1 (cyan) and all the S2 (magenta) and LMM domains (purple) of the protein are lost in the mutants, ablating most highly-conserved regions (black).Location of other smyhc1 mb16-18 and smyhc1 stl583 truncation alleles are marked (green).(B)Insitu RNA hybridisation for smyhc1 mRNA in smyhc1 mutant and their siblings from smyhc1 kg179/þ and smyhc1 kg180/þ in-crosses reveal nonsense-mediated decay of mutant smyhc1 kg179 and smyhc1 kg180 mRNA at 24 hpf.In a randomlyselected sample from a smyhc1 kg179/þ incross, 4/21 had reduced expression and were shown to be mutant; among 17 'normal' expressors 11/20 were heterozygotes and 6/20 were wild type upon sequence genotyping.From the smyhc1 kg180/þ incross, 6/29 embryos had low expression and were shown to be mutant; among 23 normal expressors 15/23 were heterozygous and 8/23 were wild type upon sequence genotyping.(C) Quantitative RT-PCR for smyhc1, smyhc2 and smyhc3 on pools of ten 2 dpf embryos from a smyhc1 kg179/þ incross that had been sorted into immotile mutant and motile siblings at 1 dpf.Each symbol represents a separate reverse transcription with symbol shapes distinguishing biological replicate lays from separate parents.The dashed line represents the motile sibling level (D).Maximum intensity projections of Slow MyHC-stained 24 hpf embryos showing region centred on somite 17 (S17).Wild-type siblings (left) show slow fibre staining and smyhc1 kg179/kg179 and smyhc1 kg180/kg180.mutant siblings show no slow fibre stain (right).

Fig. 3 .
Fig.3.Loss of Smyhc1 reduces swimming velocity.Zebrafish larvae from smyhc1 kg179/þ in-crosses in fish water were assayed for swimming activity upon touch stimulation with a pipette before and then after transient treatment with 50 μM fast myosin inhibitor BTS.A) Schematic describing workflow.B-F) Quantification of velocity (mean AE SD, N ¼ indicated on bars) on fish of indicated ages were obtained using at least three separate lays from heterozygous smyhc1 kg179/þ in-crosses.Full dataset is in TableS1.Note the log 10 scale on the Y-axes to show both control Fish Water and BTS-treated data accurately.Summary Tukey post hoc statistics for separate two way ANOVAs on Fish Water and BTS-treated are shown below.In each case, the upper table shows the overall significance, whereas the lower table shows the Tukey post hoc significance levels for the overall effects of genotype independent of age.Values of within-age Tukey post hoc test comparisons are indicated above graphs (black lines and p-values), none of which had p < 0.05 in Fish Water.To mitigate the effect of changing variance due to the increase in velocity with age and large effect of BTS, individual one way ANOVAs on genotype were performed at each age and Tukey post hoc tests supported the reduced velocity of mutants in Fish Water at 2 and 5 dpf only (green lines and p-values).

(
Fig. 4. Defective sarcomere organisation in smyhc1 kg179/kg179 mutants.Sibling zebrafish from smyhc1 kg179/þ in-crosses mounted dorsal up, anterior to left, shown in wholemount (A, upper panel) or maximal intentisity projection confocal stacks.A) Slow MyHC immunofluorescent detection with antibody S58 at 3 dpf shows signal in superficial slow fibres (SSFs) of heterozygote sibling (n ¼ 8/8 sibs) but its absence in mutant (asterisks), despite the presence in mutant of signal in cranial, cardiac (arrows), and specific somite-derived muscle (arrowheads; n ¼ 6/6).White boxes are shown enlarged beneath to highlight S58 signal in thin nascent SSFs at dorsal and ventral somitic extremes and muscle pioneers at the horizontal myoseptum in mutant (arrowheads).B) Immunodetection of slow MyHC with antibody F59 and α-actinin to mark Z-disks and phalloidin staining to reveal filamentous actin in somite 17/18 region at 24 hpf.Note the absence of organised Z-lines and thin filament arrays in the mutant but the persistence of signal at the verticle myoseptum (filled arrowheads).Boxed areas magnified beneath show repeated F-actin structures in mutant (open arrowheads) that appear to lack α-actinin.C) Immunodetection of slow MyHC with antibody S58 and all sarcomeric MyHC with A4.1025 reveals lack MyHC in mutant slow fibres but presence in underlying fast fibres in somite 16-19 region at 3 dpf.Note the poor binding of A4.1025 to fast fibres in the wild type, presumably due to adsorption to slow fibres.Abbreviations: ahmanterior hypaxial muscle, sca -supracranialis anterior, scp -supracranialis posterior.Bars ¼ 100 μm.

Fig. 5 .
Fig. 5. Survival of SSFs without Smyhc1.Sibling zebrafish from smyhc1 kg179/ þ ;Tg(smyhc1:EGFP) i104 male crossed to a female smyhc1 kg180/kg180 or smyhc1 kg179/þ ;Tg(smyhc1:EGFP) i104 mounted dorsal up, anterior to left, shown in live wholemount or confocal maximal intensity projection stacks or slices.A) Upper panels: 23ss embryos stained for sarcomeric MyHC revealing absence of signal in slow fibres marked by GFP in mutant.Note the absence of MyHC in muscle pioneers (arrowhead) and slow fibres in tail regions lacking fast MyHC (bracket).Middle panels: Short stacks of somite 19/20 stained for MyHC with F59 showing the presence of a layer of SSFs marked by EGFP (arrows) in both mutant and sibling at 34 hpf.Note the coincidence of strong slow MyHC and EGFP in SSFs (cyan arrowheads) only in the sibling.Dark areas reflect regions of the stack containing underlying fast fibres.The short stack in the mutant was selected to show that MyHC is also weakly detected by F59 in obliquelyorientated fast fibres (white arrowheads).EGFP signal in SSFs of mutant reveals a striated sarcomere-like pattern despite the absence of MyHC (open arrowheads).Lower panels: Magnified full stacks of the same embryos show EGFP in dorsal myotomes of somites 18-20 revealing normal SSF orientation, nuclear positioning and striation (open arrowheads) in mutant.B) Live smyhc1 kg179 ;Tg(smyhc1:EGFP) mutant and sibling embryos/larvae.Upper panels: Lateral views, with dorsal view below, showing that all SSFs have migrated to the lateral myotome in a 2 dpf mutant.Lower panels: Live 8 dpf larval somites 17-19 showing migrated mutant SSFs detached from the vertical myosepta (arrowheads), compared with a rare fibre defect in the SSF layer in sibling (asterisk).Insets show reduced volume of mutant SSFs in transverse optical sections at the dashed lines.C) Single confocal slices of EGFP in SSF layers with blue boxes magnified below.Whereas the sibling had cytoplasm largely filled with myofibrils containing a regular 1.96 μm sarcomeric array (open arrow- H.-T.A.Hau et al.Developmental Biology 499 (2023)  47-58