2115RESEARCH ARTICLE

INTRODUCTIONThe formation and growth of skeletal muscles in vertebrates

proceeds via successive waves or phases of myogenesis that occur

during embryonic and foetal development under the influence of

various intrinsic and extrinsic signals (reviewed by Buckingham,

2007; Pownall et al., 2002; Shi and Garry, 2006). Vertebrate muscles

contain a spectrum of fibre types varying between fatigue-resistant

slow twitch fibres with oxidative metabolism and fast twitch fibres

with glycolytic metabolism for brief powerful activity. The myosin

heavy chain (MyHC) isotype is a major determinant of these

contrasting contractile properties (Bottinelli, 2001; Weiss et al.,

2001). In a wide range of vertebrate genomes, numerous MyHC

genes are organised as members of tandem arrays or as individual

genes. Mammals have a tandem array with fast MyHC isoform

genes termed embryonic, IIa, IId, IIb, perinatal and extraocular

(Shrager et al., 2000; Weiss et al., 1999). Mammals also have the

cardiac MyHCα and MyHCβ gene pair (Myh6 and Myh7 – Mouse

Genome Informatics) with MyHCβ providing the slow MyHC

isoform for skeletal as well as cardiac muscle (Lompre et al., 1984;

Mahdavi et al., 1984). MyHC genes elsewhere in the mammalian

genome provide fast isoforms for particular craniofacial muscles

(Korfage et al., 2005a). During development and regeneration,

mammalian muscle fibres express different MyHC genes in a

specific temporal sequence (Agbulut et al., 2003). Once fibres are

formed, neural activity can remodel fibre type to suit the induced

contraction behaviour (Buller et al., 1960; Buller et al., 1969).

In contrast to the single MyHCβ gene in the mammalian genome,

the chick has a tandem array of three slow MyHC genes with

differing developmental expression patterns (Chen et al., 1997; Page

et al., 1992; Sacks et al., 2003). Analysis of chick MyHC isoforms

demonstrated that embryonic muscle fibres are initially patterned as

slow or fast twitch independently of neural activity (Crow and

Stockdale, 1986).

In teleost fish, such as carp and medaka, different MyHC genes

are expressed during different developmental stages and at different

water temperatures (Liang et al., 2007; Nihei et al., 2006; Ono et al.,

2006). The previously described zebrafish smyhc1 gene (Bryson-

Richardson et al., 2005; Nakano et al., 2004; Rauch et al., 2003) is

within a tandem array of five MyHC genes that has been extensively

studied on a genomic sequence basis as a classic example of gene

conversion events (McGuigan et al., 2004).

In the zebrafish embryo, the first muscle fibres to form are of the

slow twitch type (Devoto et al., 1996; van Raamsdonk et al., 1978).

These fibres derive from precursors known as adaxial cells, so called

because they lie adjacent to the axial midline structures that secrete

Hedgehog (Hh) family proteins (Devoto et al., 1996). Adaxial cells

respond to Hh signalling by expressing the Prdm1 transcription

factor, the activity of which is both necessary and sufficient for

adaxial cells to adopt the slow twitch fibre identity (Baxendale et al.,

2004) and express slow MyHC (Bryson-Richardson et al., 2005;

Devoto et al., 1996), and the homeodomain protein Prox1 (Glasgow

and Tomarev, 1998; Roy et al., 2001). Once specified, the cells

undergo a radial migration from their original medial location to

form a superficial monolayer of slow twitch muscle fibres covering

the embryonic myotome and subsequently a lateral wedge-shaped

strip of slow muscle called the lateralis superficialis in juveniles

(Devoto et al., 1996).

Following the initial wave of primary fibre differentiation in the

zebrafish, further slow twitch fibres differentiate in a variety of

locations. The specification of some of these secondary slow

twitch fibres has been shown to be independent of Hh signalling

(Barresi et al., 2001). Here, we have investigated the molecular

identity of these secondary slow fibres and explored the role of

Prdm1 in their specification. We show that Prdm1 activity is

Expression of multiple slow myosin heavy chain genesreveals a diversity of zebrafish slow twitch muscle fibres withdiffering requirements for Hedgehog and Prdm1 activityStone Elworthy, Murray Hargrave, Robert Knight, Katharina Mebus and Philip W. Ingham*

The zebrafish embryo develops a series of anatomically distinct slow twitch muscle fibres that characteristically express genesencoding lineage-specific isoforms of sarcomeric proteins such as MyHC and troponin. We show here that different subsets of theseslow fibres express distinct members of a tandem array of slow MyHC genes. The first slow twitch muscle fibres to differentiate,which are specified by the activity of the transcription factor Prdm1 (also called Ubo or Blimp1) in response to Hedgehog (Hh)signalling, express the smyhc1 gene. Subsequently, secondary slow twitch fibres differentiate in most cases independently of Hhactivity. We find that although some of these later-forming fibres also express smyhc1, others express smyhc2 or smyhc3. We showthat the smyhc1-positive fibres express the ubo (prdm1) gene and adopt fast twitch fibre characteristics in the absence of Prdm1activity, whereas those that do not express smyhc1 can differentiate independently of Prdm1 function. Conversely, some smyhc2-expressing fibres, although independent of Prdm1 function, require Hh activity to form. The adult trunk slow fibres express smyhc2and smyhc3, but lack smyhc1 expression. The different slow fibres in the craniofacial muscles variously express smyhc1, smyhc2 andsmyhc3, and all differentiate independently of Prdm1.

KEY WORDS: Zebrafish, Myotome, Fibre type, Slow myosin heavy chain, Troponin, Prox1, Shh, Blimp1, u-boot, prdm1, Craniofacial

Development 135, 2115-2126 (2008) doi:10.1242/dev.015719

MRC Centre for Developmental and Biomedical Genetics and Department ofBiomedical Science, University of Sheffield, Western Bank, Sheffield S10 2TN, UK.

*Author for correspondence (e-mail: p.w.ingham@sheffield.ac.uk)

Accepted 15 April 2008 DEVELO

PMENT

Development ePress online publication date 14 May 2008

2116

required only by a specific subset of these slow fibres that are

distinguished from other secondary slow fibres by their expression

of the smyhc1 gene.

MATERIALS AND METHODSZebrafish, mutants, genotyping and cyclopamine treatmentZebrafish embryos were kept and staged following standard methods

(Kimmel et al., 1995; Westerfield, 1995). Pigmentation was inhibited by

immersion in 0.2 mM N-phenylthiourea. Zebrafish mutants ubotp39, nrdm805,

smob641 and smohi1640, and the transgenic line Tg(acta1:GFP)zf13 have been

described previously (Barresi et al., 2000; Baxendale et al., 2004; Chen et

al., 2001; Higashijima et al., 1997; Hernandez-Lagunas et al., 2005; Roy et

al., 2001; van Eeden et al., 1996; Varga et al., 2001). All experiments using

Tg(PACprdm1:gfp)i106 were with smob641. For other experiments, either

smob641 or smohi1640 were used. Genotyping for ubotp39 used PCR with

primers tagtggggaacgacccttta+ggcctatttccaagtgtagtgc followed by digestion

with AciI, which cuts at the site of the ubotp39 point mutation but not the wild-

type allele. Cyclopamine treatment followed a standard method with

immersion in 50 μM cyclopamine (Toronto Research Chemicals) from the

four-cell stage or 20-somite stage (Hirsinger et al., 2004; Wolff et al., 2003).

Smyhc cDNART-PCR from 14-somite stage embryos using CODEHOP (Rose et al., 1998)

primer pairs gaatccgaatctgccgaaarggnttycc+cacgcccatgaaggctckdatrttcca

and gaagatggagggagacctgaaygaratgga+ctgctccttcttcagctcctcngccatcat

followed by 5� and 3� Smart RACE (Clontech) gave products used to design

the primer pair cggggacacaggacaactcgaggtaag+gctagattagaaccagtactt -

gaacatggc used to amplify an RT-PCR product spanning the smyhc1(Accession Number EF030714)-coding sequence cloned as psmyhc1-CDS.

This is the same gene as described by Bryson-Richardson et al. (Bryson-

Richardson et al., 2005) and Rauch et al. (Rauch et al., 2003). smyhc2(Accession Number BK006466) is represented by the existing cDNA

IMAGE_7433531, smyhc3 (Accession Number BK006465) is represented

by cDNA IMAGE_7041248. 5� Smart RACE from 120 hpf embryos with

primer tggcgggttctgagggtgacagt gave EU526313 (Accession Number) with

the smyhc2 5�UTR, whereas primer gctcccggtccgacttcctgaga gave

EU218877 (Accession Number) with the smyhc3 5�UTR.

In situ hybridisation and immunofluorescenceWhole-mount in situ hybridisation (Thisse and Thisse, 1998) and

cryosection in situ hybridisation (Braissant and Wahli, 1998) were as

previously described. Probes were from plasmids pBlimp1 (Baxendale

et al., 2004) for prdm1, pslowtroponinC-ISH subcloned from

IMAGE_6899234 for slowtroponinC zgc:86932 (Thisse et al., 2001; Xu et

al., 2006), psmyhc1-CDS with bp 1-6128 of EF030714, psmyhc1 (1 to 249

bp) with bp 1-249 of EF030714, psmyhc2 (1 to 324 bp) with bp 1-324 of

EU526313, psmyhc2 (1 to 155 bp) with bp 1-155 of EU526313 and psmyhc3(1 to 129 bp) with bp 1-129 of EU218877. psmyhc1-CDS detects the

smyhc1, smyhc2 and smyhc3 transcripts. psmyhc2 (1 to 324 bp) provides

more sensitive detection of smyhc2 than does psmyhc2 (1 to 155 bp) but

requires high stringency conditions to avoid cross-hybridisation to smyhc1.Chromogenic in situ hybridisation microscopy used a Zeiss Axioplan with

Spot4 digital camera.

Immunofluorescence followed Devoto et al. (Devoto et al., 1996),

Serbedzija et al. (Serbedzija et al., 1998) Hamade et al. (Hamade et al., 2006)

and Hernandez et al. (Hernandez et al., 2005) using: mouse monoclonals

F310 anti-fast myosin light chain (1:50, DSHB), MF20 anti-light

meromyosin (1:500, DSHB), S58 anti-slow MyHC (1:10, DSHB), II-II6B3

anti-collagen 2a (1:500 DSHB), 15B8 anti-β-catenin (1:500 Sigma) and

Bu20a anti-bromodeoxyuridine (1:100, DakoCytomation); and rabbit

polyclonals anti-GFP (1:1000, Torrey Pines Biolabs) and anti-Prox1

(1:5000). Anti-Prox1 was raised against recombinant zebrafish Prox1

purified from E. coli (A. M. Taylor, personal communication). TOTO-3

iodide (Molecular Probes) was used at 200 nM. Imaging used Leica TCS SP,

Olympus Fluoview FV1000 or Zeiss LSM 510 confocal microscopes with

ImageJ software (Abramoff, 2004).

Fluorescent in situ hybridisation used anti-digAP and Fast Red (Roche)

or anti-digPOD (at 1:10,000) and TSA Cyanine3 (Perkin Elmer).

smyhc1:gfp transgenesisA GFP-SV40pA-FRT-Kn-FRT recombineering targeting cassette and red

recombineering system in EL250 cells were used to insert EGFP with an

SV40 polyadenylation site at the smyhc1 ATG start site in BAC ZC227E6

(Lee et al., 2001), which has at least 80 kb of upstream sequence and an

insert size of 170 kb, as determined by CHEF electrophoresis and Southern

blotting. This modified BAC, linearised with PI-SceI, was used to generate

stable transgenic line Tg(BAC smyhc1:gfp)i108 (Higashijima et al., 1997).

Reporter plasmid p9.7kbsmyhc1:gfp-I-SceI was constructed with I-SceI sites

flanking 9.7kb of smyhc1 upstream sequence joined at the smyhc1 ATG start

site to EGFP with an SV40 polyadenylation site. p9.7kbsmyhc1:gfp-I-SceI

was used to generate five stable Tg(9.7kb smyhc1:gfp) transgenic lines.

Lines Tg(9.7kb smyhc1:gfp)i101, i102, i103 and i104 have the same expression

pattern as Tg(BAC smyhc1:gfp)i108. Line Tg(9.7kb smyhc1:gfp)i105 differs in

having additional ectopic expression in the brain. For this work, we used the

two lines that gave the strongest signal: Tg(BAC smyhc1:gfp)i108 and

Tg(9.7kb smyhc1:gfp)i104.

prdm1:gfp transgenesisA GFP-SV40pA-FRT-CmR-FRT recombineering targeting cassette was

used to insert EGFP with an SV40 polyadenylation site at the prdm1 ATG

start site in PAC F2219 (Amemiya and Zon, 1999; Baxendale et al., 2004).

The modified PAC, linearised with NotI, was used to generate stable

transgenic lines Tg(PACprdm1:gfp)i106 and Tg(PACprdm1:gfp)i107

(Higashijima et al., 1997). A further red recombineering step removed all

sequences downstream of the SV40 polyadenylation site. This PAC,

linearised with Not1, was used to generate stable transgenic line

Tg(–60prdm1:gfp)i111. Tg(PACprdm1:gfp)i106 was used in smo mutants.

Tg(PACprdm1:gfp)i106 and Tg(PACprdm1:gfp)i107 partially rescue the ubotp39

phenotype and so Tg(–60prdm1:gfp)i111 was used in ubo (prdm1) mutants.

Tg(–60prdm1:gfp)i111 gives the same expression pattern as

Tg(PACprdm1:gfp)i106 and Tg(PACprdm1:gfp)i107 when transmitted from

males, but gives ubiquitous maternal expression. Tg(–60prdm1:gfp)i111

males were crossed to non-transgenic females to provide the embryos used

for this study.

BrdU analysisBrdU pulse labelling followed Park and Appel (Park and Appel, 2003).

Fixed, BrdU-labelled embryos were exposed to 1.7 N HCl for 60 minutes

prior to a 90 minute 10 μgml-1 proteinase K digestion and standard

immunofluorescence. Specimens were examined over a three-somite wide

view at the level of the anus. The full thickness of the myotome was scanned

at 0.57 μm intervals by confocal microscopy and all the individual images

in the confocal stack were inspected to count BrdU labelled nuclei.

RESULTSThe expression of different slow MyHC genesdefines distinct groups of slow twitch musclesThe slow MyHC specific antibodies F59 and S58 (raised against

chicken myosin) have been widely used as markers for

investigations of muscle fibre type specification in zebrafish

embryos (Devoto et al., 1996; Hernandez et al., 2005). Expression

of a slow MyHC gene, designated smyhc1, has previously been

reported in F59-positive muscle fibres in early embryos,

suggesting that it may encode the epitope recognised by F59 and

S58 in zebrafish (Bryson-Richardson et al., 2005; Nakano et al.,

2004; Rauch et al., 2003). Using in situ hybridisation, we

confirmed that a probe complementary to the coding sequence of

this gene (smyhc1-CDS) detects slow fibres in a diverse set of

muscles (Fig. 1A) (Rauch et al., 2003). To investigate the

expression of this gene further, we generated gfp reporter

constructs; surprisingly, we found that transgenic animals carrying

these constructs expressed GFP in only a subset of the S58-positive

muscles (Fig. 1A). This could imply that the smyhc1 expression

pattern is not fully recapitulated by our reporter genes;

alternatively, it may be that the smyhc1-CDS probe detects

RESEARCH ARTICLE Development 135 (12)

DEVELO

PMENT

transcripts from more than one Smyhc gene. Consistent with this

latter hypothesis, genomic analysis has shown that zebrafish

smyhc1 resides within a tandem array of five genes that have a

remarkable level of DNA sequence identity (Fig. 1C) (McGuigan

et al., 2004). The long stretches of DNA sequence identity between

smyhc1 and the other genes in the tandem array, including a 270

bp stretch of 100% nucleotide identity between smyhc1 and the

adjacent gene in the tandem array cause the CDS probe to cross

hybridise, confounding expression analysis of the individual

genes. We therefore synthesised gene-specific 5�UTR probes to

resolve the expression of these genes in different slow fibres by in

situ hybridisation (Fig. 1; see Fig. S1 and Table S1 in the

supplementary material; Fig. 3).

The smyhc1- (1 to 249 bp) 5�UTR probe confirmed that the

smyhc1:gfp reporter genes precisely recapitulate the expression

pattern of the endogenous gene. The adaxial cells express smyhc1,

as do the primary surface slow fibres and muscle pioneers to

which they give rise (Fig. 1D; see Fig. S1 in the supplementary

material).

From 48 hpf, a series of muscles differentiate with slow twitch

fibres that express smyhc2 but not smyhc1 (Fig. 1D-G, see Fig. S1

and Table S1 in the supplementary material). The third gene in the

tandem array, which we call smyhc3, has very short UTR sequences,

making detection by in situ hybridisation difficult. Nevertheless, we

could observe at the limit of detection smyhc3 expression in some

of the smyhc2-positive trunk muscles (Fig. 1H; see Table S1 in the

supplementary material).

In contrast to the differential expression of the Smyhc genes, slowtroponin C is expressed in all slow MyHC-expressing fibres

throughout the embryo (see Fig. S2 in the supplementary material)

(Thisse et al., 2001). Expression is restricted to S58-positive fibres

except in the pectoral fin muscle that is S58 negative but strongly

expresses slow troponin C (see Fig. S2 in the supplementary

material); the physiological status of this muscle is unclear.

There is expression of smyhc2 at 72 hpf and smyhc3 at 96 hpf in

a few fibres lateral to the horizontal myoseptum that we have called

the embryonic lateralis superficialis (Fig. 1D,F,H; see Fig. S1 in the

supplementary material). These few fibres may later become

2117RESEARCH ARTICLEPrdm1 and zebrafish slow muscle fibres

Fig. 1. Different Smyhc genes have distinct expression patterns. (A) At 96 hpf, the smyhc1:gfp reporter gene is expressed in a subset of theslow fibres labelled with the S58 anti slow MyHC Ab or the smyhc1-CDS in situ hybridisation (ISH) probe. The iob is a major muscle that does notexpress smyhc1:gfp in its slow fibres. (B) High magnification lateral view of smyhc1-CDS in situ hybridisation in posterior trunk showing transcript isconcentrated at the ends of the superficial slow fibres. (C) A diagram illustrating the tandem array of Smyhc genes as described by McGuigan et al.(McGuigan et al., 2004) together with the gene names used here. We followed the Bryson-Richardson et al. (Bryson-Richardson et al., 2005)renaming of myhcE as smyhc1 and similarly renamed the adjacent genes smyhc2 and smyhc3. (D) Differential expression of Smyhc genes at 96 hpf,as shown by in situ hybridisation using smyhc1 (1 to 249 bp) or smyhc2 (1 to 324 bp) probes. The sca, els, iob, sh, scp and icp somite-derivedmuscles, and dorsal and ventral craniofacial muscles express smyhc2. (E) smyhc1:gfp together with smyhc2 (1 to 324 bp) in situ hybridisation at 96hpf. (F) At 96 hpf, smyhc2 (1 to 324 bp) in situ hybridisation shows colocalisation with slow MyHC S58 antigen but not with smyhc1:gfp in the scamuscle (dorsal view anterior trunk), els or iob muscles (lateral view anterior trunk) or scp or icp muscles (lateral view posterior tail). (G) A deep focusventral view shows smyhc2 (1 to 324 bp) expression in the oesophagus at 96 hpf. (H) The low sensitivity smyhc3 (1 to 129 bp) in situ hybridisationweakly detects expression in the sca and els muscles at 96 hpf, as shown by dorsal and lateral views of the anterior trunk. Scale bars: 25 μm.Abbreviations: dm, head dorsal muscles; els, embryonic lateralis superficialis; icp, infracarinalis posterior; iob, inferior obliquus; oes, oesophagus;sca, supracarinalis anterior; scp, supracarinalis posterior; sh, sternohyoideus; vm, head ventral muscles. Muscle nomenclature is taken from previouswork (Schilling and Kimmel, 1997; Stiassny, 2000; Winterbottom, 1974).

DEVELO

PMENT

2118

incorporated into the lateralis superficialis muscle that forms at this

location in juveniles predominantly from adaxially derived fibres

(Devoto et al., 1996). At 32 dpf, smyhc1 and smyhc2 are co-

expressed in the lateralis superficialis (Fig. 2A). By 42 days post

fertilisation (dpf), the trunk lacks smyhc1 expression (data not

shown), and smyhc2 and smyhc3 are expressed in the lateralis

superficialis muscles (Fig. 2C). GFP is known to persist long after

reporter transcription has ceased (Tallafuss and Bally-Cuif, 2003).

This presumably explains why lateral fibres of the 42 dpf lateralis

superficialis are still marked with smyhc1:gfp (Fig. 2B). The adult

lateralis superficialis lacks smyhc1 expression (data not shown) and

has a complex complementary expression of smyhc2 and smyhc3(Fig. 2D).

From 48 hpf, a series of craniofacial muscles develop from cranial

paraxial mesoderm (Schilling and Kimmel, 1997). Most of these

muscles comprise both slow fibres and fast fibres (Hernandez et al.,

2005). As in the trunk and tail, slow troponin C is expressed in all

S58-positive craniofacial muscles (Fig. 3A,B,E). Most of these slow

fibres express both smyhc1 and smyhc2 (Fig. 3A-D; see Table S1 in

the supplementary material). Strikingly, however, smyhc1

expression is absent from the slow fibres of a disparate subset of the

craniofacial muscles (Fig. 3; see Table S1 in the supplementary

material; Fig. 1D). Furthermore, some muscles contain both clusters

of fibres that co-express smyhc1 with smyhc2, and also clusters that

just express smyhc2 (Fig. 3D).Despite the limitations of the smyhc3 probe, intense smyhc3

expression was detected in the levator arcus palatini muscle that

lacks smyhc1 expression (Fig. 3A). Weaker smyhc3 expression was

detected in certain other craniofacial muscles (Fig. 3A,B).

Using S58 staining, Barresi et al. (Barresi et al., 2001) showed

that slow fibres are added all along the dorsal and ventral margins

of the primary superficial slow fibre layer after 24 hpf. As most of

this region is devoid of smyhc2 and smyhc3 (Fig. 1D; see Fig. S1 in

the supplementary material) it seems probable that these secondary

fibres express smyhc1. To distinguish such fibres from their primary

counterparts, we adopted the BrdU labelling technique used by

Barresi et al. (Barresi et al., 2001). Exposure of smyhc1:gfpembryos to BrdU at 18 hpf, resulted in extensive labelling of fast

muscle and external cell nuclei across all dorsal ventral levels of the

myotome at 48 hpf, in accordance with the recent report of

Stellabotte et al. (Stellabotte et al., 2007) (Fig. 4A,B). By contrast,

only smyhc1:gfp-positive slow fibres at the dorsal and ventral

margins of the layer of superficial slow fibres were BrdU labelled

(Fig. 4B-D; see Movie 1 in the supplementary material). Exposure

to BrdU at 35 hpf similarly led by 96 hpf to BrdU-labelled,

smyhc1:gfp positive, fibres only at the dorsal and ventral margins

(Fig. 4C,D). These findings establish that these fibres do indeed

express smyhc1. We refer to these fibres as ‘secondary superficial

slow fibres’ to distinguish them from the series of smyhc2-

expressing secondary slow fibres.

smyhc1 and smyhc2 expressing secondary fibresdiffer in their requirement for HedgehogsignallingPrevious studies have shown that some secondary slow fibres can

differentiate normally in the absence of Hh signalling activity, in

contrast to their primary counterparts (Barresi et al., 2001). We

investigated the identity of these Hh-independent fibres using the

gene-specific Smyhc probes. Embryos either mutant for the Hh

signal transducer Smoothened (smo) or treated with the Smo

antagonist cyclopamine, have extensive secondary superficial slow

fibres at the dorsal and ventral margins of the posterior trunk and

the tail that express smyhc1 and not smyhc2 (Fig. 5A,C,D).

Expression of smyhc2 in smo mutants is similar to that in wild

type; however, the embryonic lateralis superficialis fibres are

absent, as are the smyhc2-expressing fibres of the tail (Fig. 5D; see

Table S1 in the supplementary material). This lack of the smyhc2-

expressing fibres of the tail is strikingly similar to the loss of third

wave slow muscle reported for cyclopamine-treated Xenopusembryos (Grimaldi et al., 2004). We treated embryos with

cyclopamine from the 20-somite stage to determine whether

zebrafish have a similar requirement for Hh signalling during slow

fibre myogenesis late in embryogenesis. These treated embryos

had a normal pattern of smyhc1:gfp and smyhc2 expression except

for the lack of the tail smyhc2-expressing slow fibres (Fig. 5E).

The posterior tails of these embryos had a narrower dorsal-ventral

extent and lacked any muscle dorsal or ventral of the smyhc1-

expressing layer (Fig. 5F), suggesting that Hh signalling is

required for the formation of this muscle. This contrasts with its

role in specifying the fibre type of adaxial cells but is strikingly

similar to the generation of third wave slow muscle in Xenopus(Grimaldi et al., 2004).

RESEARCH ARTICLE Development 135 (12)

Fig. 2. Expression of smyhc1, smyhc2 and smyhc3 in the juvenileand adult trunk. All panels show trunk transverse cryosections. (A) At32 dpf smyhc1 (1 to 249 bp) in situ hybridisation shows expression inthe lateralis superficialis, while smyhc2 (1 to 324 bp) in situhybridisation shows colocalisation with smyhc1:gfp in the lateralissuperficialis. (B) At 42 dpf, smyhc1:gfp shows colocalisation with slowMyHC S58 antigen in the lateralis superficialis, which is not fast muscleF310 antigen positive. (C) At 42 dpf, smyhc2 (1 to 324 bp) and smyhc3(1 to 129 bp) in situ hybridisation shows expression in the lateralissuperficialis. smyhc1 (1 to 249 bp) in situ hybridisation on adjacentsections failed to detect any expression. (D) Adjacent sections of 22month post-fertilisation adult with smyhc2 (1 to 324 bp) and smyhc3(1 to 129 bp) in situ hybridisation showing expression in subsets of theslow MyHC S58 antigen positive, fast muscle F310 antigen negative,lateralis superficialis. smyhc1 (1 to 249 bp) in situ hybridisation onadjacent sections failed to detect any expression throughout the trunk.Scale bars: 25 μm.

DEVELO

PMENT

Hedgehog independent smyhc1 expressingsecondary superficial slow fibres require prdm1Using the highly sensitive smyhc1-CDS in situ hybridisation probe,

we have compared the ontogeny of slow twitch muscle fibres in

prdm1 mutant, smo mutant and wild-type embryos (Fig. 6A; we use

‘Smyhc’ to mean transcripts detected with the smyhc1-CDS probe).

Consistent with the report of Barresi et al. (Barresi et al., 2001), we

found that in smo mutant embryos, slow fibres are virtually absent

at 24 hpf, but subsequently appear at the dorsal and ventral margins

of the somites at 36 hpf, increasing in number by 48 hpf. By contrast,

in prdm1 mutants at 24 hpf, we found that many fibres scattered

through the myotome show low levels of Smyhc expression, in line

with previous descriptions of the ubo(prdm1)tp39 mutant phenotype

(Roy et al., 2001). This Smyhc expression was observed not only in

embryos homozygous for the ubo(prdm1)tp39 mis-sense allele but

also in those homozygous for nrd(prdm1)m805, which has a stop

codon at amino acid 154 and is thus a presumptive null allele

(Hernandez-Lagunas et al., 2005) (Fig. 6A). Notably, a few fibres

have a less dramatic reduction in expression and these tend to be

more dorsal and more numerous in posterior somites. By 48 hpf, the

scattered expression of Smyhc throughout the myotome is greatly

diminished, but expression persists in fibres at the dorsal margin.

The location of these latter fibres is similar to that of the secondary

superficial slow fibres, raising the possibility that at least some

secondary superficial slow fibres can form in the absence of prdm1function. Alternatively, these could correspond to adaxially derived,

primary slow fibres that escape the requirement for wild-type levels

of prdm1 function.

2119RESEARCH ARTICLEPrdm1 and zebrafish slow muscle fibres

Fig. 3. Differential expression of Smyhc genes in craniofacial muscles. (A) Lateral and ventral views of the dorsal group of craniofacialmuscles at 96 hpf with slow troponin C (stnnC), smyhc1 (1 to 249 bp), smyhc2 (1 to 324 bp) or smyhc3 (1 to 129 bp) in situ hybridisation (seeFig. 1A,D,E,G for gross location of these muscles). The lap and do lack smyhc1 expression while smyhc3 expression is restricted to the lap.(B) Ventral views showing ventral craniofacial muscles at 96 hpf with slow troponin C (stnnC), smyhc1 (1 to 249 bp), smyhc2 (1 to 324 bp) orsmyhc3 (1 to 129 bp) in situ hybridisation. The ima, sh, iob and posterior tv lack smyhc1 expression. (C) Lateral and ventral views of the dorsalgroup of craniofacial muscles at 96 hpf showing colocalisation of slow MyHC S58 antigen with smyhc2 (1 to 324 bp) in situ hybridisation and,in the ao and lo, with smyhc1:gfp. (D) Ventral view at 96 hpf showing colocalisation of smyhc1:gfp and smyhc2 (1 to 324 bp) in situhybridisation with slow MyHC S58 antigen. In the am, some fibres express smyhc2 and S58 antigen but not smyhc1:gfp (blue arrows). Somefibres express S58 antigen but neither smyhc2 nor smyhc1:gfp (white arrows). The extraocular muscles have zones of S58 antigen-expressingfibres that co-express smyhc1:gfp and zones that do not (pink arrows). (E) Ventral view at 96 hpf showing expression of slow troponin C in situhybridisation (stnnC) in all slow MyHC S58 antigen-expressing fibres. Scale bars: 25 μm. Abbreviations: abh, abductor hyoideus; am, adductormandibulae; ao, adductor operculi; do, dilator operculi; hh, hyohyoidus; ih, interhyoideus; ima, intermandibularis anterior; imp,intermandibularis posterior; io, inferior oblique; iob, inferior obliquus; ir, inferior rectus, lap, levator arcus palatini; lo, levatator operculi; sh,sternohyoideus; tv, transversus ventralis. D

EVELO

PMENT

2120

To distinguish between these possibilities, we eliminated primary

slow muscle fibres from ubo(prdm1)tp39 mutant embryos by treating

them with the Smo antagonist cyclopamine (Hirsinger et al., 2004;

Wolff et al., 2003). Such embryos are essentially devoid of both

primary and secondary superficial slow fibres at 48 hpf, as

indicated by the absence in the tail and posterior trunk of both

Smyhc and slow troponin C expression (Fig. 6B). It follows from

this finding that the persistent dorsal slow fibres seen in

ubo(prdm1)tp39 mutants are adaxially derived primary slow fibres

and that prdm1 function is essential for the differentiation of the

secondary superficial slow fibres that form normally in the absence

of Hh signalling.

By contrast, the smyhc2-expressing fibres are largely unaffected

by the loss of prdm1 activity, though the embryonic lateralis

superficialis muscles are lost in nrd(prdm1)m805 mutants as in smomutants (Fig. 6C; see Table S1 in the supplementary material). This

may be a secondary consequence of the aberrant somite morphology

caused by lack of the horizontal myoseptum.

The craniofacial muscles generally have wild-type levels of

expression of both smyhc1 and smyhc2 in nrd(prdm1)m805

mutants (see Fig. S3B and Table S1 in the supplementary

material; Fig. 6C). The only perturbations to the craniofacial slow

fibres in prdm1 mutants are associated with the major loss of

skeletal elements in the posterior head (see Fig. S3A in the

supplementary material) (see also Wilm and Solnica-Krezel,

2005).

prdm1:gfp expression in primary and secondaryslow fibresWe next investigated whether prdm1 is expressed in the precursors

of the secondary superficial (smyhc1+ve) slow fibres. Low-level

prdm1 expression can be detected by in situ hybridisation at 35 hpf

in cells tentatively identified as newly forming secondary superficial

slow fibres in smo mutants (Fig. 7A). To characterise this expression

further, we took advantage of the stability of GFP transcript and

protein (Tallafuss and Bally-Cuif, 2003) to monitor prdm1 expression

using prdm1:gfp reporter lines. The expression of GFP in these lines

(Fig. 7B) recapitulates the pattern of endogenous prdm1 transcription

in adaxial cells (Baxendale et al., 2004) but persists as the cells

migrate and differentiate into the superficial layer of mononucleated

slow fibres that covers the mytotome at 24 hpf (Fig, 7C). As

expected, in smo mutants carrying the prdm1:gfp transgene, no such

labelled fibres are present at 24 hpf (Fig. 7C). At 48 hpf, however,

prdm1:gfp-positive fibres are present at the dorsal and ventral

margins of the myotome, locations consistent with their being

secondary superficial slow fibres. Simultaneous staining for Smyhc

or slow troponin C expression confirmed that the prdm1:gfp-positive

fibres are indeed secondary superficial slow fibres (Fig. 7D). In

addition, we also found that these fibres accumulate Prox1 in their

nuclei (although at lower levels than their primary counterparts) (Fig.

7E) and lack expression of the fast muscle F310 antigen (Fig. 7F).

Intriguingly, prdm1:gfp is specifically expressed not only in the

smyhc1 expressing slow fibres but also in the slow fibres of muscles

that express smyhc2 and not smyhc1 (Fig. 7G, see also Fig. 1D,F).

Transformation of secondary superficial slowfibres to a fast twitch character in ubo (prdm1)mutantsWe next analysed the fate of primary and secondary superficial slow

fibre precursors in ubo(prdm1)tp39 mutants using the prdm1:gfptransgene as a lineage marker. At 24 hpf, prdm1:gfp colocalises with

the scattered weakly Smyhc-positive cells typical of ubo(prdm1)tp39

mutants, indicating that these are derived from adaxial cells that

have failed to migrate properly owing to the loss of prdm1 function

(Fig. 8A). Although by 48 hpf Smyhc expression is lost from all but

the most dorsal fibres, the scattered prdm1:gfp positive cells persist

(Fig. 8A) and differentiate into fibres that are multinucleate and

F310 positive, characteristics that are typical of fast twitch fibres

(Hamade et al., 2006; Roy et al., 2001) (Fig. 8B,C). Similarly, even

the dorsal fibres in which slow MyHC expression persists are

marked by both prdm1:gfp and the fast muscle F310 antigen (Fig.

8D) and lack expression of Prox1 (Glasgow and Tomarev, 1998)

(Fig. 8E).

At 48 hpf, cylopamine-treated transgenic ubo(prdm1)tp39 mutant

embryos have the normal complement of prdm1:gfp-positive fibres

at the location of the secondary superficial slow fibres in their

genetically wild-type siblings. However, unlike those in the latter,

these fibres do not express slow MyHC and do express the fast fibre-

specific F310 antigen (Fig. 8F).

RESEARCH ARTICLE Development 135 (12)

Fig. 4. BrdU labelling shows that wild type embryos havesmyhc1-expressing secondary superficial slow fibres. All panelsshow lateral views of posterior trunk, level with the anus. (A) At 48 hpf,embryos exposed to BrdU at 18 hpf show extensive BrdU co-labellingwith acta1:gfp in nuclei that have the characteristic fast muscleelongated and diagonal morphology. (B) At 48 hpf, embryos exposedto BrdU at 18 hpf have BrdU-labelled nuclei superficial to thesmyhc1:gfp-labelled surface slow fibres in the presumptive external celllayer (nuclei superficial to the surface slow fibres; these nuclei have acharacteristic large flattened disk morphology) and immediately medialto the surface slow fibres, but not in the main body of the surface slowfibre layer (see also Movie 1). (C) BrdU and smyhc1:gfp co-labelling atdorsal somite margin at 48 hpf after BrdU exposure at 18 hpf, and atventral somite margin 96 hpf after BrdU exposure at 35 hpf. (D) Scatterplots showing number of BrdU and smyhc1:gfp co-labelled nuclei in thedorsal or ventral myotome within a three-somite view at the level of theanus. Each dot represents the count in one embryo. The lines show themean. Data are shown for exposure to BrdU at 18 hpf or 35 hpf, andwith analysis at 48 hpf or 96 hpf. The BrdU labelled smyhc:gfp nucleiwere always the smyhc1:gfp nuclei closest to the dorsal or ventralmargin of the myotome. (E) At 48 hpf, embryos exposed to BrdU at 35hpf have extensive labelling in cell types other than the surface slowfibres. Scale bars: 25 μm. D

EVELO

PMENT

DISCUSSIONThe zebrafish embryo develops a series of distinct muscle types

with slow twitch fibres. We show here that these different slow

fibres can be distinguished by the expression of the smyhc1,

smyhc2 and smyhc3 genes that form a tandem array in the

genome. During the juvenile period, there is an apparent transition

from smyhc1 to smyhc2 and smyhc3 expression in the lateralis

superficialis. This is reminiscent of the developmental transitions

in MyHC isoform expression in amniotes. The chick also has

three tandemly arrayed slow MyHC genes that are expressed

with differing developmental control, timing and subcellular

transcript localisation (Sacks et al., 2003). We do not know

whether this arrangement of slow MyHC genes is evolutionarily

ancient or arose independently in each lineage. A molecular

phylogenic assessment of this issue is hampered by gene

conversion events within the zebrafish Smyhc tandem array

(McGuigan et al., 2004).

The expression of slow troponin C indicates that pan slow lineage

expression can be driven by elements associated with a single

transcription unit. This raises the issue as to what evolutionary

constraints have produced the Smyhc tandem array genomic structure.

McGuigan et al. have previously described the extraordinary level of

DNA sequence conservation between the genes in the Smyhc tandem

array (McGuigan et al., 2004). Although frequent and recent gene

conversion events have acted to maintain this sequence conservation,

it is conceivable that the remaining amino acid differences between

the different gene products may have functional significance. The

regions of most amino acid sequence divergence correspond to the 25-

50 junction and 50-20 junction. These regions show low sequence

conservation across the myosin classes (Cope et al., 1996), although

there is evidence that their sequence can influence the enzymatic

activity of myosins (reviewed by Murphy and Spudich, 2000).

Establishing whether these differences confer functional changes will

require biochemical and physiological analyses.

2121RESEARCH ARTICLEPrdm1 and zebrafish slow muscle fibres

Fig. 5. Hedgehog dependent and independent smyhc1 and smyhc2 expression. All panels show lateral views. (A) smyhc1 (1 to 249 bp) insitu hybridisation at 48 hpf showing expression in secondary superficial slow fibres at the dorsal and ventral edges of the somites of a smomutant. (B) smyhc2 (1 to 155 bp) in situ hybridisation at 48 hpf showing wild-type levels of expression in the iob of smo mutant (see Fig. S1 inthe supplementary material for comparison with wild type). (C) Untreated and cyclopamine treated 55 hpf embryos showing Hh-independentsmyhc1:gfp in the secondary superficial slow fibres and Hh-independent smyhc2 (1 to 324 bp) in situ hybridisation signal in the sca and iob.(D) 96 hpf wild-type sib and smo mutant. The smo mutant has slow MyHC S58 antigen in iob, sca and secondary superficial slow fibres; smyhc2(1 to 324 bp) in situ hybridisation signal in iob and sca; but lacks scp and icp smyhc2 (1 to 324 bp) in situ hybridisation signal. (E) Untreatedembryo and embryo exposed to cyclopamine from the 20-somite stage. This late cyclopamine exposure does not eliminate the smyhc1:gfpsurface slow fibres but does eliminate smyhc2 (1 to 324 bp) in situ hybridisation signal in the scp and icp, as well as causing a dorsoventralnarrowing of the posterior tail. Fifteen out of 15 embryos with this treatment showed this phenotype. (F) High-magnification views of posteriortail in untreated embryo and embryo exposed to cyclopamine from the 20-somite stage. In untreated embryos, smyhc2 (1 to 324 bp) in situhybridisation signal colocalises with the general muscle MF20 antigen in the scp and icp dorsal and ventral to the smyhc:gfp surface slow fibrelayer. The treated embryos lack scp and icp smyhc2 (1 to 324 bp) in situ hybridisation signal and general muscle MF20 antigen dorsal or ventral tothe smyhc:gfp surface slow fibre layer. Scale bars: 25 μm. Abbreviations: icp, infracarinalis posterior; iob, inferior obliquus; sca, supracarinalisanterior; scp, supracarinalis posterior.

DEVELO

PMENT

2122

The evolution of the tandem array genomic structure may have

been driven by transcriptional control constraints. The fact that our

9.7 kb promoter smyhc1:gfp reporter gene recapitulates the

expression of the endogenous gene, indicates that, for smyhc1expression, at least, the adjoining parts of the tandem array are not

required in cis for transcriptional control. The tandem arrangement

of mammalian MyHC genes is constrained by bidirectional

expression of MyHC genes and antisense transcripts required for

transcriptional control of adjacent MyHC genes (Haddad et al.,

2003; Pandorf et al., 2006). We do not know whether such a

mechanism exists in zebrafish.

The mammalian craniofacial muscles express an especially

diverse selection of MyHC genes. This is thought to reflect intricate

functional requirements in these muscles (Korfage et al., 2005a;

Korfage et al., 2005b; Noden and Francis-West, 2006). The

complex expression patterns of Smyhc genes in the zebrafish

embryonic craniofacial muscles may have an analogous functional

significance.

There are two additional predicted genes [called myhA and

myhcB by McGuigan et al. (McGuigan et al., 2004)] in the same

tandem gene array as the three Smyhc genes we describe here.

Although myhA and myhB are not represented in current EST

databases, we isolated 5�RACE products for two alternative

5�UTRs of myhA using RNA from 120 hpf embryos. We were

unable to detect any expression in embryos by in situ

hybridisation. Possibly, the predicted genes are expressed

principally at later stages of the life cycle or under different growth

conditions.

Prdm1 acts as a transcriptional repressor to drive the

differentiation of adaxial cells into primary slow twitch fibres in

response to Hh signalling (Baxendale et al., 2004; von Hofsten et al.,

2008). Our analysis shows that prdm1 is also required for secondary

superficial slow fibres to adopt a slow twitch character

independently of Hh signalling. How expression of prdm1 is

activated in the progenitors of these secondary superficial slow

fibres remains to be determined.

RESEARCH ARTICLE Development 135 (12)

Fig. 6. Hedgehog-independent secondary superficial slow fibres require Prdm1, whereas smyhc2 expression does not. (A) Lateral viewsof smyhc1-CDS in situ hybridisation on 24 hpf, 36 hpf and 48 hpf smo mutants, ubo (prdm1)tp39 mutants, nrd (prdm1)m805 and wild-type (wt)siblings; and transverse sections of 24 hpf ubo (prdm1)tp39 mutants and wild-type siblings. The near absence of Smyhc-expressing primary slowfibres in smo mutants at 24 hpf allows secondary slow fibres to be identified unambiguously at later stages. At 24 hpf, ubo (prdm1)tp39 or nrd(prdm1)m805 mutants have abundant, misplaced, weakly Smyhc-expressing fibres. Thus, it is unclear whether Smyhc-expressing fibres at later stagesin prdm1 mutants are simply the remnants of those present at 24 hpf. (B) When primary slow fibre specification is prevented by cyclopaminetreatment, 48 hpf genotyped ubo (prdm1)tp39 mutants lack almost all smyhc1-CDS in situ hybridisation and slow troponin C (stnnC) in situhybridisation in secondary superficial slow fibres, whereas genotyped wild-type siblings have robust expression. (C) smyhc2 (1 to 155 bp) in situhybridisation showing 96 hpf nrd (prdm1)m805 mutants have near wild-type levels of expression in the sca, iob, scp, icp and craniofacial muscles, butlack smyhc2-expressing els fibres. High magnification dorsal views show sca and lateral views show iob and scp/icp. Scale bars: 25 μm.Abbreviations: els, embryonic lateralis superficialis; icp, infracarinalis posterior; iob, inferior obliquus; sca, supracarinalis anterior; scp, supracarinalisposterior.

DEVELO

PMENT

Intriguingly, the specific expression of prdm1:gfp in slow

fibres extends beyond the smyhc1-positive fibres and includes

the smyhc2-positive fibres for which we could discern no

prdm1 requirement. We consider it likely that this prdm1:gfpexpression reflects the expression of the endogenous gene, as

we have observed it in three independent prdm1:gfp reporter

gene lines. It remains possible that prdm1 acts redundantly with

some other gene in these fibres or has a subtle role we failed to

discern.

Although specification of slow fibre type shows diversity in its

requirement for Hh signalling and prdm1 activity, it is possible that

there is a common basis for slow fibre type determination at a more

downstream step. The specific expression of slow troponin Cacross all slow fibres could result from a shared transcriptional

program specifying slow fibre type that is common to different

slow muscles. Alternatively, the slow troponin C expression may

be determined by a composite of independent transcriptional

control elements for each of the different types of slow fibre. The

complex transcriptional control of mouse Myf5 in myogenic cells

provides an example of numerous independent control elements

acting in distinct muscle types (Carvajal et al., 2001; Hadchouel et

al., 2003).

The origin of the progenitor cells for the various secondary slow

fibres and their relationship to the recently identified, Pax7-

positive, secondary fast muscle progenitors is currently unclear

(Hammond et al., 2007; Hollway et al., 2007; Stellabotte et al.,

2007). An additional issue is the mechanism for myogenic

induction versus progenitor maintenance for secondary slow fibres.

Hh signalling induces the myogenesis of primary slow fibres and

also of lateral fast fibres in the primary myotome (Feng et al., 2006;

Hammond et al., 2007; Lewis et al., 1999). Our data suggest that

Hh signalling is also required for myogenesis of the smyhc2-

expressing slow fibres of the posterior tail. This shows a striking

similarity to third wave slow fibre myogenesis in Xenopus,

suggesting that it is an evolutionarily ancient mechanism (Grimaldi

et al., 2004).

It is unknown how various slow fibres are specified to express

particular Smyhc genes. This question possibly relates to the more

general issue of muscle identity determination. Muscle identity may

be influenced by intrinsic cues, as well as signalling from other cell

types. In zebrafish, anterior trunk somites have an intrinsic axial

identity that enables them to produce appendicular muscle (Haines

et al., 2004). Earlier in development, paraxial mesoderm for

different axial levels is specifically determined by different

2123RESEARCH ARTICLEPrdm1 and zebrafish slow muscle fibres

Fig. 7. The prdm1:gfp transgenicreporter marks primary andsecondary slow fibres. (A) Weakprdm1 in situ hybridisation inpresumptive secondary superficial slowfibre cells (arrows) at 35 hpf in smomutants. (B) prdm1:gfp fluorescence inadaxial cells at the six-somite stage isalmost eliminated in smo mutants. (C) At24 hpf, Prox1-expressing primary slowfibres are marked with prdm1:gfp andare absent in smo mutants. (D) Lateralviews of posterior trunk of 48 hpf smomutants showing prdm1:gfp colocalisedwith smyhc1-CDS in situ hybridisationand slow troponin C (stnnC) in situhybridisation in secondary superficialslow fibres at the dorsal and ventraledges of the somites. (E) In smo mutants,prdm1:gfp colocalises with Prox1 antigenin secondary fibres (arrows) in somite 13at 36 hpf. Note the intense Prox1staining in isolated cells that are externalto the myotome (arrowheads). (F) At 48hpf, in smo mutant posterior trunk, slowMyHC S58 antigen and prdm1:gfp marksecondary superficial slow fibres that donot colocalise with fast muscle F310antigen. (G) Dorsolateral view of the ioband dorsal view of the sca at 96 hpfshowing colocalisation of prdm1:gfp withslow MyHC S58 antigen but not with fastmuscle F310 antigen (see Fig. 1 for grosslocation of these muscles). Scale bars: 25μm. Abbreviations: iob, inferior obliquus;sca, supracarinalis anterior.

DEVELO

PMENT

2124

combinations of Nodal, Bmp and Fgf signals (Szeto and Kimelman,

2006). Numerous Hox genes are expressed in particular regions of

the paraxial mesoderm (Prince et al., 1998). Possibly, certain

paraxial mesoderm cells are predisposed to form certain types of

slow fibre by such intrinsic cues. Intrinsic cues may also influence

whether muscle has a fast twitch or slow twitch character (Nikovits,

Jr et al., 2001).

In the head in particular, the different Smyhc genes reveal an

unprecedented detail of molecular diversity in the slow fibres. A

single Engrailed-expressing condensation of first arch mesoderm

splits to form the adjacent levator arcus palatini and dilator operculi

(Hatta et al., 1990), and yet the levator arcus palatini expresses high

levels of smyhc3 while the dilator operculi has barely detectable

levels. The differential expression of Smyhc genes between different

fibres within craniofacial muscles such as the adductor mandibulae

is equally remarkable.

The different Smyhc genes described here will provide valuable

late differentiation markers for experimental studies required to

understand the specification of muscle identity.

We thank L. Gleadall, M. Green, F. Browne and S. Surfleet for expert zebrafishhusbandry; J. Sanderson for imaging advice; K. Ohymama for cryosection insitu hybridisation advice; K. Berry, C. Moore, A. Burguiere, A. Taylor and C.Davison for generating and characterising reagents; anonymous reviewers forhelpful suggestions; H. Roehl, S. Baxendale and F. van Eeden for helpfuldiscussions; N. Copeland for the EL250 strain; and K. Artinger for nrdm805

zebrafish. This research project was funded by a UK Medical Research Council(MRC) Programme Grant (G0100151) and by the EU MYORES Network ofExcellence. The CDBG Zebrafish Aquaria was supported by an MRC CentreDevelopment Grant (G0400100) to P.W.I. Imaging facilities were funded by theMRC, Wellcome Trust and Yorkshire Cancer Research.

Supplementary materialSupplementary material for this article is available athttp://dev.biologists.org/cgi/content/full/135/12/2115/DC1

RESEARCH ARTICLE Development 135 (12)

Fig. 8. The prdm1:gfp transgenic reporter provides a lineage tracer to follow the fate of prdm1-dependent muscle precursors in ubo(prdm1) mutants. (A) Transverse sections show colocalisation prdm1:gfp with smyhc1-CDS in situ hybridisation in ubo (prdm1)tp39 mutants andwild-type siblings at 24 hpf. At 48 hpf, the prdm1:gfp marked muscle fibres are still present in ubo (prdm1)tp39 mutants, although Smyhc expressionis reduced. Cell boundaries are marked with anti-β-catenin 15B8. (B) Lateral views showing that muscle fibres with prdm1:gfp form a superficiallayer of horizontal, mononucleate fibres in 48 hpf wild-type embryos but are among the fast fibres in ubo (prdm1)tp39 mutants and like fast fibresare frequently multinucleate (TOTO stain) and orientated diagonally. (C) Transverse sections at 48 hpf show colocalisation of F310 fast muscleantigen with prdm1:gfp in ubo (prdm1)tp39 mutants but not in wild-type siblings. (D) In posterior trunk at 48 hpf, in ubo (prdm1)tp39 mutants, thedorsal fibres with persistent slow MyHC S58 antigen are marked with prdm1:gfp and fast muscle F310 antigen. (E) In posterior trunk at 36 hpf, inubo (prdm1)tp39 mutants, even the most dorsally positioned fibres marked with prdm1:gfp lack Prox1 (arrows). Note the intense Prox1 staining inisolated cells external to the myotome (arrowheads). (F) In posterior trunk of cyclopamine-treated embryos, prdm1:gfp marked secondary fibrescolocalise with slow MyHC S58 antigen but not fast muscle F310 antigen in genotyped wild-type siblings and vice versa in genotyped ubo(prdm1)tp39 mutants. Scale bars: 25 μm.

DEVELO

PMENT

ReferencesAbramoff, M. D., Magelhaes, P. J. and Ram, S. J. (2004). Image Processing with

ImageJ. Biophotonics Int. 11, 36-42.Agbulut, O., Noirez, P., Beaumont, F. and Butler-Browne, G. (2003). Myosin

heavy chain isoforms in postnatal muscle development of mice. Biol. Cell 95,399-406.

Amemiya, C. T. and Zon, L. I. (1999). Generation of a zebrafish P1 artificialchromosome library. Genomics 58, 211-213.

Barresi, M. J., Stickney, H. L. and Devoto, S. H. (2000). The zebrafish slow-muscle-omitted gene product is required for Hedgehog signal transduction andthe development of slow muscle identity. Development 127, 2189-2199.

Barresi, M. J., D’Angelo, J. A., Hernandez, L. P. and Devoto, S. H. (2001).Distinct mechanisms regulate slow-muscle development. Curr. Biol. 11, 1432-1438.

Baxendale, S., Davison, C., Muxworthy, C., Wolff, C., Ingham, P. W. and Roy,S. (2004). The B-cell maturation factor Blimp-1 specifies vertebrate slow-twitchmuscle fiber identity in response to Hedgehog signaling. Nat. Genet. 36, 88-93.

Bottinelli, R. (2001). Functional heterogeneity of mammalian single muscle fibres:do myosin isoforms tell the whole story? Pflugers Arch. 443, 6-17.

Braissant, O. and Wahli, W. (1998). A simplified in situ hybridization protocolusing non-radioactively labeled probes to detect abundant and rare mRNAs ontissue sections. Biochemica 1, 10-16.

Bryson-Richardson, R. J., Daggett, D. F., Cortes, F., Neyt, C., Keenan, D. G.and Currie, P. D. (2005). Myosin heavy chain expression in zebrafish and slowmuscle composition. Dev. Dyn. 233, 1018-1022.

Buckingham, M. (2007). Skeletal muscle progenitor cells and the role of Paxgenes. C. R. Biol. 330, 530-533.

Buller, A. J., Eccles, J. C. and Eccles, R. M. (1960). Interactions betweenmotoneurones and muscles in respect of the characteristic speeds of theirresponses. J. Physiol. 150, 417-439.

Buller, A. J., Mommaerts, W. F. and Seraydarian, K. (1969). Enzymic propertiesof myosin in fast and slow twitch muscles of the cat following cross-innervation.J. Physiol. 205, 581-597.

Carvajal, J. J., Cox, D., Summerbell, D. and Rigby, P. W. (2001). A BACtransgenic analysis of the Mrf4/Myf5 locus reveals interdigitated elements thatcontrol activation and maintenance of gene expression during muscledevelopment. Development 128, 1857-1868.

Chen, Q., Moore, L. A., Wick, M. and Bandman, E. (1997). Identification of agenomic locus containing three slow myosin heavy chain genes in the chicken.Biochim. Biophys. Acta 1353, 148-156.

Chen, W., Burgess, S. and Hopkins, N. (2001). Analysis of the zebrafishsmoothened mutant reveals conserved and divergent functions of hedgehogactivity. Development 128, 2385-2396.

Cope, M. J., Whisstock, J., Rayment, I. and Kendrick-Jones, J. (1996).Conservation within the myosin motor domain: implications for structure andfunction. Structure 4, 969-987.

Crow, M. T. and Stockdale, F. E. (1986). Myosin expression and specializationamong the earliest muscle fibers of the developing avian limb. Dev. Biol. 113,238-254.

Devoto, S. H., Melancon, E., Eisen, J. S. and Westerfield, M. (1996).Identification of separate slow and fast muscle precursor cells in vivo, prior tosomite formation. Development 122, 3371-3380.

Feng, X., Adiarte, E. G. and Devoto, S. H. (2006). Hedgehog acts directly on thezebrafish dermomyotome to promote myogenic differentiation. Dev. Biol. 300,736-746.

Glasgow, E. and Tomarev, S. I. (1998). Restricted expression of the homeoboxgene prox 1 in developing zebrafish. Mech. Dev. 76, 175-178.

Grimaldi, A., Tettamanti, G., Martin, B. L., Gaffield, W., Pownall, M. E. andHughes, S. M. (2004). Hedgehog regulation of superficial slow muscle fibres inXenopus and the evolution of tetrapod trunk myogenesis. Development 131,3249-3262.

Hadchouel, J., Carvajal, J. J., Daubas, P., Bajard, L., Chang, T., Rocancourt, D.,Cox, D., Summerbell, D., Tajbakhsh, S., Rigby, P. W. et al. (2003). Analysis ofa key regulatory region upstream of the Myf5 gene reveals multiple phases ofmyogenesis, orchestrated at each site by a combination of elements dispersedthroughout the locus. Development 130, 3415-3426.

Haddad, F., Bodell, P. W., Qin, A. X., Giger, J. M. and Baldwin, K. M. (2003).Role of antisense RNA in coordinating cardiac myosin heavy chain geneswitching. J. Biol. Chem. 278, 37132-37138.

Haines, L., Neyt, C., Gautier, P., Keenan, D. G., Bryson-Richardson, R. J.,Hollway, G. E., Cole, N. J. and Currie, P. D. (2004). Met and Hgf signalingcontrols hypaxial muscle and lateral line development in the zebrafish.Development 131, 4857-4869.

Hamade, A., Deries, M., Begemann, G., Bally-Cuif, L., Genet, C., Sabatier, F.,Bonnieu, A. and Cousin, X. (2006). Retinoic acid activates myogenesis in vivothrough Fgf8 signalling. Dev. Biol. 289, 127-140.

Hammond, C. L., Hinits, Y., Osborn, D. P., Minchin, J. E., Tettamanti, G. andHughes, S. M. (2007). Signals and myogenic regulatory factors restrict pax3and pax7 expression to dermomyotome-like tissue in zebrafish. Dev. Biol. 302,504-521.

Hatta, K., Schilling, T. F., BreMiller, R. A. and Kimmel, C. B. (1990).Specification of jaw muscle identity in zebrafish: correlation with engrailed-homeoprotein expression. Science 250, 802-805.

Hernandez, L. P., Patterson, S. E. and Devoto, S. H. (2005). The developmentof muscle fiber type identity in zebrafish cranial muscles. Anat. Embryol. 209,323-334.

Hernandez-Lagunas, L., Choi, I. F., Kaji, T., Simpson, P., Hershey, C., Zhou, Y.,Zon, L., Mercola, M. and Artinger, K. B. (2005). Zebrafish narrowmindeddisrupts the transcription factor prdm1 and is required for neural crest andsensory neuron specification. Dev. Biol. 278, 347-357.

Higashijima, S., Okamoto, H., Ueno, N., Hotta, Y. and Eguchi, G. (1997).High-frequency generation of transgenic zebrafish which reliably express GFP inwhole muscles or the whole body by using promoters of zebrafish origin. Dev.Biol. 192, 289-299.

Hirsinger, E., Stellabotte, F., Devoto, S. H. and Westerfield, M. (2004).Hedgehog signaling is required for commitment but not initial induction of slowmuscle precursors. Dev. Biol. 275, 143-157.

Hollway, G. E., Bryson-Richardson, R. J., Berger, S., Cole, N. J., Hall, T. E. andCurrie, P. D. (2007). Whole-somite rotation generates muscle progenitor cellcompartments in the developing zebrafish embryo. Dev. Cell 12, 207-219.

Kimmel, C. B., Ballard, W. W., Kimmel, S. R., Ullmann, B. and Schilling, T. F.(1995). Stages of embryonic development of the Zebrafish. Dev. Dyn. 203, 253-310.

Korfage, J. A., Koolstra, J. H., Langenbach, G. E. and van Eijden, T. M.(2005a). Fiber-type composition of the human jaw muscles-(part 1) origin andfunctional significance of fiber-type diversity. J. Dent. Res. 84, 774-783.

Korfage, J. A., Koolstra, J. H., Langenbach, G. E. and van Eijden, T. M.(2005b). Fiber-type composition of the human jaw muscles-(part 2) role ofhybrid fibers and factors responsible for inter-individual variation. J. Dent. Res.84, 784-793.

Lee, E. C., Yu, D., Martinez de Velasco, J., Tessarollo, L., Swing, D. A., Court,D. L., Jenkins, N. A. and Copeland, N. G. (2001). A highly efficient Escherichiacoli-based chromosome engineering system adapted for recombinogenictargeting and subcloning of BAC DNA. Genomics 73, 56-65.

Lewis, K. E., Currie, P. D., Roy, S., Schauerte, H., Haffter, P. and Ingham, P. W.(1999). Control of muscle cell-type specification in the zebrafish embryo byHedgehog signalling. Dev. Biol. 216, 469-480.

Liang, C. S., Kobiyama, A., Shimizu, A., Sasaki, T., Asakawa, S., Shimizu, N.and Watabe, S. (2007). Fast skeletal muscle myosin heavy chain gene cluster ofmedaka Oryzias latipes enrolled in temperature adaptation. Physiol. Genomics29, 201-214.

Lompre, A. M., Nadal-Ginard, B. and Mahdavi, V. (1984). Expression of thecardiac ventricular alpha- and beta-myosin heavy chain genes is developmentallyand hormonally regulated. J. Biol. Chem. 259, 6437-6446.

Mahdavi, V., Chambers, A. P. and Nadal-Ginard, B. (1984). Cardiac alpha- andbeta-myosin heavy chain genes are organized in tandem. Proc. Natl. Acad. Sci.USA 81, 2626-2630.

McGuigan, K., Phillips, P. C. and Postlethwait, J. H. (2004). Evolution ofsarcomeric myosin heavy chain genes: evidence from fish. Mol. Biol. Evol. 21,1042-1056.

Murphy, C. T. and Spudich, J. A. (2000). Variable surface loops and myosinactivity: accessories to a motor. J. Muscle Res. Cell Motil. 21, 139-151.

Nakano, Y., Kim, R., Kawakami, A., Roy, S., Schier, A. F. and Ingham, P. W.(2004). Inactivation of dispatched 1 by the chameleon mutation disruptsHedgehog signalling in the zebrafish embryo. Dev. Biol. 269, 381-392.

Nihei, Y., Kobiyama, A., Ikeda, D., Ono, Y., Ohara, S., Cole, N. J., Johnston, I.A. and Watabe, S. (2006). Molecular cloning and mRNA expression analysis ofcarp embryonic, slow and cardiac myosin heavy chain isoforms. J. Exp. Biol. 209,188-198.

Nikovits, W., Jr, Cann, G. M., Huang, R., Christ, B. and Stockdale, F. E. (2001).Patterning of fast and slow fibers within embryonic muscles is establishedindependently of signals from the surrounding mesenchyme. Development 128,2537-2544.

Noden, D. M. and Francis-West, P. (2006). The differentiation andmorphogenesis of craniofacial muscles. Dev. Dyn. 235, 1194-1218.

Ono, Y., Liang, C., Ikeda, D. and Watabe, S. (2006). cDNA cloning of myosinheavy chain genes from medaka Oryzias latipes embryos and larvae and theirexpression patterns during development. Dev. Dyn. 235, 3092-3101.

Page, S., Miller, J. B., DiMario, J. X., Hager, E. J., Moser, A. and Stockdale, F.E. (1992). Developmentally regulated expression of three slow isoforms ofmyosin heavy chain: diversity among the first fibers to form in avian muscle. Dev.Biol. 154, 118-128.

Pandorf, C. E., Haddad, F., Roy, R. R., Qin, A. X., Edgerton, V. R. and Baldwin,K. M. (2006). Dynamics of myosin heavy chain gene regulation in slow skeletalmuscle: role of natural antisense RNA. J. Biol. Chem. 281, 38330-38342.

Park, H. C. and Appel, B. (2003). Delta-Notch signaling regulates oligodendrocytespecification. Development 130, 3747-3755.

Pownall, M. E., Gustafsson, M. K. and Emerson, C. P., Jr (2002). Myogenicregulatory factors and the specification of muscle progenitors in vertebrateembryos. Annu. Rev. Cell Dev. Biol. 18, 747-783.

2125RESEARCH ARTICLEPrdm1 and zebrafish slow muscle fibres

DEVELO

PMENT

2126

Prince, V. E., Joly, L., Ekker, M. and Ho, R. K. (1998). Zebrafish hox genes:genomic organization and modified colinear expression patterns in the trunk.Development 125, 407-420.

Rauch, G. J., Lyons, D. A., Middendorf, I., Friedlander, B., Arana, N., Reyes, T.and Talbot, W. S. (2003). Submission and curation of gene expression data.ZFIN Direct Data Submission, http://zfin.org.

Rose, T. M., Schultz, E. R., Henikoff, J. G., Pietrokovski, S., McCallum, C. M.and Henikoff, S. (1998). Consensus-degenerate hybrid oligonucleotide primersfor amplification of distantly related sequences. Nucleic Acids Res. 26, 1628-1635.

Roy, S., Wolff, C. and Ingham, P. W. (2001). The u-boot mutation identifies aHedgehog-regulated myogenic switch for fiber-type diversification in thezebrafish embryo. Genes Dev. 15, 1563-1576.

Sacks, L. D., Cann, G. M., Nikovits, W., Jr, Conlon, S., Espinoza, N. R. andStockdale, F. E. (2003). Regulation of myosin expression during myotomeformation. Development 130, 3391-3402.

Schilling, T. F. and Kimmel, C. B. (1997). Musculoskeletal patterning in thepharyngeal segments of the zebrafish embryo. Development 124, 2945-2960.

Serbedzija, G. N., Chen, J. N. and Fishman, M. C. (1998). Regulation in theheart field of zebrafish. Development 125, 1095-1101.

Shi, X. and Garry, D. J. (2006). Muscle stem cells in development, regeneration,and disease. Genes Dev. 20, 1692-1708.

Shrager, J. B., Desjardins, P. R., Burkman, J. M., Konig, S. K., Stewart, S. K.,Su, L., Shah, M. C., Bricklin, E., Tewari, M., Hoffman, R. et al. (2000).Human skeletal myosin heavy chain genes are tightly linked in the orderembryonic-IIa-IId/x-ILb-perinatal-extraocular. J. Muscle Res. Cell Motil. 21, 345-355.

Stellabotte, F., Dobbs-McAuliffe, B., Fernandez, D. A., Feng, X. and Devoto,S. H. (2007). Dynamic somite cell rearrangements lead to distinct waves ofmyotome growth. Development 134, 1253-1257.

Stiassny, M. (2000). Muscular system. In The Laboratory Fish (ed. G. K.Ostrander), pp. 109-128. New York: Academic Press.

Szeto, D. P. and Kimelman, D. (2006). The regulation of mesodermal progenitorcell commitment to somitogenesis subdivides the zebrafish body musculatureinto distinct domains. Genes Dev. 20, 1923-1932.

Tallafuss, A. and Bally-Cuif, L. (2003). Tracing of her5 progeny in zebrafishtransgenics reveals the dynamics of midbrain-hindbrain neurogenesis andmaintenance. Development 130, 4307-4323.

Thisse, B., Pflumio, S., Fürthauer, M., Loppin, B., Heyer, V., Degrave, A.,Woehl, R., Lux, A., Steffan, T., Charbonnier, X. Q. and Thisse, C. (2001).

Expression of the zebrafish genome during embryogenesis. ZFIN Direct DataSubmission, http://zfin.org.

Thisse, C. and Thisse, B. (1998). High resolution whole-mount in situhybridization. Zebrafish Sci. Monit. 5.

van Eeden, F. J., Granato, M., Schach, U., Brand, M., Furutani-Seiki, M.,Haffter, P., Hammerschmidt, M., Heisenberg, C. P., Jiang, Y. J., Kane, D. A.et al. (1996). Mutations affecting somite formation and patterning in thezebrafish, Danio rerio. Development 123, 153-164.

van Raamsdonk, W., Pool, C. W. and te Kronnie, G. (1978). Differentiation ofmuscle fiber types in the teleost Brachydanio rerio. Anat. Embryol. 153, 137-155.

Varga, Z. M., Amores, A., Lewis, K. E., Yan, Y. L., Postlethwait, J. H., Eisen, J.S. and Westerfield, M. (2001). Zebrafish smoothened functions in ventralneural tube specification and axon tract formation. Development 128, 3497-3509.

von Hofsten, J., Elworthy, S., Gilchrist, M. J., Smith, J. C., Wardle, F. C. andIngham, P. W. (2008). Prdm1 and Sox6 mediated transcriptional repressionspecifies muscle fibre type in the zebrafish embryo. EMBO Rep. (in press).

Weiss, A., McDonough, D., Wertman, B., Acakpo-Satchivi, L., Montgomery,K., Kucherlapati, R., Leinwand, L. and Krauter, K. (1999). Organization ofhuman and mouse skeletal myosin heavy chain gene clusters is highlyconserved. Proc. Natl. Acad. Sci. USA 96, 2958-2963.

Weiss, S., Rossi, R., Pellegrino, M. A., Bottinelli, R. and Geeves, M. A.(2001). Differing ADP release rates from myosin heavy chain isoforms definethe shortening velocity of skeletal muscle fibers. J. Biol. Chem. 276, 45902-45908.

Westerfield, M. (1995). The Zebrafish Book (2nd edn). Oregon: University ofOregon Press.

Wilm, T. P. and Solnica-Krezel, L. (2005). Essential roles of a zebrafishprdm1/blimp1 homolog in embryo patterning and organogenesis. Development132, 393-404.

Winterbottom, R. (1974). A descriptive synonymy of the striated muscles of theTeleostei. Proc. Acad. Nat. Sci. Philadelphia 125, 225-317.

Wolff, C., Roy, S. and Ingham, P. W. (2003). Multiple muscle cell identitiesinduced by distinct levels and timing of Hedgehog activity in the zebrafishembryo. Curr. Biol. 13, 1169-1181.

Xu, J., Srinivas, B. P., Tay, S. Y., Mak, A., Yu, X., Lee, S. G., Yang, H.,Govindarajan, K. R., Leong, B., Bourque, G. et al. (2006). Genomewideexpression profiling in the zebrafish embryo identifies target genes regulated byHedgehog signaling during vertebrate development. Genetics 174, 735-752.

RESEARCH ARTICLE Development 135 (12)

DEVELO

PMENT