Effects of decreased muscle activity on developing axial musculature in nic^b107 mutant zebrafish (Danio rerio) Free

Embryonic development is not just a sequence of one gene expression following another. Mechanical loading influences development(Bagatto et al., 2001; Cho, 2004; Hutson et al., 2003; Prendergast, 2002; Vandenburgh et al., 1991). Muscle and bone are tissues that generate and support mechanical loads, which also regulate their growth, maintenance and differentiation(Hoppeler and Fluck, 2002; Huiskes et al., 2000; Pette, 2001). We will consider the effects of an absence of mechanical load by immobility on muscle development in a model system for vertebrate development, the zebrafish(Danio rerio Hamilton).

Increasing muscle activity by forced swimming in fish stimulates red muscle development, enhances muscle enzyme activity, increases blood oxygen carrying capacity, increases mitochondrial density, improves swimming efficiency and increases hypoxia tolerance (Bagatto et al., 2001; Davison, 1989, 1997; De Graaf et al., 1990; Kieffer, 2000; Pelster et al., 2003). Decreased muscle activity transiently downregulates the activity of the muscle enzyme succinate dehydrogenase (De Graaf et al., 1990). Decreased activity by immobility in zebrafish and Xenopus laevis embryos correlated with abnormal muscle fibre distribution and axial musculature architecture in some(Droin and Beauchemin, 1974; Granato et al., 1996; Van Raamsdonk et al., 1977, 1982) but not all cases(Granato et al., 1996). The effects of altered muscle activity on gene expression levels during development have been only rarely investigated.

The bulk of zebrafish axial muscle consists of fast white fibres. Slow red fibres form a superficial layer, and intermediate pink fibres are located in between (Waterman, 1969). The white trunk axial musculature is arranged in a series of myomeres, which are separated by collagenous sheets, the myosepta. Muscle fibres in adult fish run between the myosepta in a pseudo-helical pattern(Alexander, 1969; Ellerby and Altringham, 2001; Gemballa and Vogel, 2002; Johnston et al., 1995; Mos and Van Der Stelt, 1982; Van der Stelt, 1968; Van Leeuwen, 1999). This pseudo-helical arrangement is thought to be an optimisation for muscle work output (Alexander, 1969) and,as such, may be influenced by mechanical loading. At 18 hours post-fertilisation (18 hpf),zebrafish embryos start making their first feeble movements (Westerfield, 1995). At this age, slow muscle fibres in the zebrafish embryo are adjacent and parallel to the notochord, can be stained for heavy myosin chains and are about to migrate laterally through the somite(Devoto et al., 1996). Fast fibres, which already form the bulk of the muscle mass, do not stain for heavy myosin chains until after 23 hpf (Devoto et al., 1996). Helical arrangements of zebrafish muscle fibres are first observed at 4 days after hatching(Van Raamsdonk et al., 1974). This timeline of development and use of the axial musculature also suggest that early use of the axial musculature is crucial for its proper development(Van Raamsdonk et al.,1977).

As the use of the musculature appears crucial for proper development of the axial musculature, lack of use, i.e. immobility, is expected to hamper proper development. We used the nic^b107 mutant to study the effect of immobility on axial muscle development. In this mutant, theα-subunit of the acetylcholine receptor is defective, which blocks assembly of functional acetylcholine receptors on the muscle fibres(Sepich et al., 1998). As a result of this lack of innervation, the muscle fibres fail to contract in vivo. They are mechanically intact, however, as they are able to contract upon electrical stimulation (Westerfield et al., 1990). In the present paper, we report several effects of immobility on the morphogenetic development of axial musculature in early zebrafish larvae by studying the expression levels of a selection of structural as well as regulatory muscle genes and studying muscle structure at different organisational levels.

Embryos from the nic^b107 strain were bought from the Zebrafish International Resource Center (ZIRC) in Oregon, USA (NIH-NCRR grant#RR12546) and raised in our facility. Zebrafish embryos were generated by natural spawnings of heterozygous parents. Homozygous mutants were selected on the basis of immobility and an inability to respond to touch. Embryos were reared at 28.5°C and removed from the egg capsule at 24 hpf using needles.

Embryos were positioned in sedative (1 g l^-1 MS-222 and 1.5 g l^-1 Na₂CO₃. H₂O) on a 1% agarose gel. Lateral photographs were taken using an Olympus DP50 digital camera mounted on a Zeiss Stemi SV11 microscope with AnalySIS software V3.1 (Soft Imaging System GmbH, Münster, Germany). At 24, 48, 72, 96 and 120 hpf,total body length and, at anus level, muscle height, notochord height, somite size in anterior–posterior direction and angle of the somite were measured from photos using AnalySIS software in both nic^b107 embryos and their wild-type siblings. Five to eight animals were used per group per measurement. Statistical differences between wild-type and nic^b107 data were detected using the Mann–Whitney U test and were considered significant when P<0.05.

Embryos were fixed in 4% paraformaldehyde (PFA) in PBS overnight at 4°C, then stored in 1% PFA in PBS, postfixed in 10% PFA in PBS and embedded in 15% gelatin in PBS and fixed overnight at 4°C in 4% PFA in PBS. Transverse sections of 100 μm thickness in the area just behind the anus were cut on a vibratome 1500 (Vibratome, St Louis, MO, USA) and stained overnight at 4°C with propidium iodide (1 μg ml^-1) in PBS. They were incubated in 25% (1 h), 50% (1 h), 75% (overnight) and 90% glycerol in PBS (3 h) and then embedded in 90% glycerol in PBS immediately prior to examination.

A 15 μm-thick Z-stack of 1 μm-thick consecutive optical sections was created from the 100 μm-thick transverse section using a laser scanning microscope (ZEISS LSM-510). The Z-stack was exported as individual TIFF files to AnalySIS software and calibrated. Each fibre that was present as a complete cross section in 15 consecutive sections (i.e. over a distance of 15 μm from anterior to posterior) was individually and manually tracked. This implies that fibres close to myosepta were not digitised, due to tapering. For each cross section, the centre of area (CA) in coordinates of the Z-stack was computed with AnalySIS software. The fibre orientations in X_zY_zZ_z coordinates of the Z-stack were computed in Matlab 6.5 (The Mathworks, Inc., Gouda, The Netherlands) from the CA in the first section with that of the nearest CA in the second section. As a control of the validity, two such computations were made per embryo, using different sections. For final analyses, optical sections that were 5 μm apart were analysed. At this distance, individual fibres can be easily identified and tracked, a straight vector is a relatively accurate description of the local fibre orientation and the error in the computed orientation was determined to be less than 5°.

In general, the computed orientation of the fibres in the X_zY_zZ_z coordinates of the Z-stack is not a fair representation of the fibres in a fish-bound X_fY_fZ_f coordinate system, because the sections in the X_zY_z plane are not exactly parallel to the X_fY_f plane (transverse plane) of the fish (Fig. 1). Based on the left–right symmetry of the fish, the computed fibre orientations were rotated over three perpendicular axes to obtain a visually left–right symmetrical vector field. The result is a series of vectors describing the elevation and azimuth of each individual fibre in a fish-bound X_fY_fZ_f coordinate system. Elevation is the angle of the fibre with the horizontal plane (or X_fZ_f plane). Azimuth is the angle of the projection of the fibre on the horizontal plane with the sagittal plane(Y_fZ_f). For visualisation purposes,these vectors were projected on the first section. The distance δbetween sections is 5 μm. From δ and the length of the projections in X_f and Y_f directions, the size of the azimuth and elevation of each vector can be calculated. Azimuth =atan(L_x/δ), and elevation =atan[L_y/√(L_x² +δ ²)], where L_x is the length of the projection in the X_f direction and L_yis the length of the projection in the Y_f direction.

Total RNA was isolated using the RNeasy kit (Qiagen, Venlo, The Netherlands) from five single wild-type and five nic^b107embryos at each of the time points 24, 48, 72, 96 and 120 hpf. Concentrations were determined spectrophotometrically, and integrity was ensured by analysis on a 1.5% agarose gel. RNA was stored at –80°C until further use. First-strand cDNA was constructed from total RNA. For each sample, a non-template control (NT) was included. Total RNA (max. 1 μg) was concentrated in a vacuum dryer. 1 μl 10× DNase I reaction buffer, 1μl DNase I (Invitrogen, Breda, The Netherlands) and milliQ water (Elga LabWater, Lane End, UK) to a final volume of 10 μl were added to the RNA pellet and incubated for 15 min at room temperature. DNase I was inactivated by adding 1 μl 25 mmol l^-1 EDTA and incubating at 65°C for 10 min. To each sample, a mastermix containing 300 ng random hexamers, 1 μl 10 mmol dNTP mix, 4 μl 5× first strand buffer, 2 μl 0.1 mol l^-1 DTT and 10 U RNase inhibitor (Invitrogen) were added and incubated for 10 min at room temperature and an additional 2 min at 37°C. To each positive sample, 1 μl Superscript RNase H free reverse transcriptase (Invitrogen) was added. To each NT, 1 μl milliQ water was added. Reactions were incubated for 50 min at 37°C. All samples were filled up with demineralised water, such that 1 μl of water was added per nanogram of RNA used, and stored at –20°C until further use.

Primer Express software (Applied Biosystems, Nieuwerkerk a/d IJssel, The Netherlands) was used to design primers for use in real-time quantitative PCR(RQ-PCR). Primers are given in Table 1 in supplementary material. Primers were tested for specificity and efficiency of template amplification by sequencing the PCR product and using a known template dilution series in RQ-PCR,respectively. For RQ-PCR, 5 μl cDNA and forward and reverse primers (1.5μl of 5 μmol l^-1 each) were added to 12.5 μl Quantitect SYBR Green PCR Mastermix (Qiagen, Venlo, The Netherlands) and filled up with 4.5 μl demineralised water to a final volume of 25 μl. RQ-PCR (15 min at 95°C, 40 cycles of 15 s at 94°C, 30 s at 60°C, 30 s at 72°C)and melt analysis (60–90°C in 1° steps) was carried out on the Rotor-Gene 200 Real Time cycler (Corbett Research, Mortlake, New South Wales,Australia). Data were analysed using the Pfaffl method to include primer efficiencies (Pfaffl, 2001). This method requires the use of a designated `housekeeping gene', a gene whose expression is not affected by the treatment, decreased muscle activity in this case. Two such housekeeping genes were used to normalise the expression between nic^b107 and wild-type embryos: ribosomal protein 40S and β-actin. Results were similar with either gene, and the results for 40S are shown. Immobility might influence the process of muscle development but it might also influence the final muscle structure. We therefore chose to test several groups of genes: structural components of sarcomeres (myosins, titin, troponin C), muscle activity genes [myoglobin,phosphofructokinase for muscle (PFK-m), mitochondrial diaphorase (NADHd) and succinate dehydrogenase complex subunit A (SDHa)] and muscle growth factors(insulin-like growth factor and its receptors, myostatin/growth and differentiation factor 8, myogenin). Two collagens were tested as they represent two different tissue types that may be indirectly influenced by muscle activity. Collagen type 1 α 2 is found mainly in skin in embryonic stages and later in bone. Collagen type 2 α 1 is found around the notochord in early development and later in cartilage. To determine whether differences were statistically significant, analysis of variance(ANOVA) contrasts were determined in SPSS v. 12.0.1 (SPSS Inc., Chicago, IL,USA) between nic^b107 and wild-type gene expression at the same age, while taking into account whether or not variances were homogeneous. Differences were considered significant when P<0.05.

Embryos were fixed in 4% PFA in PBS overnight at 4°C, washed in PBS,incubated in 5% sucrose for 30 min, embedded in 1.5% agarose in 5% sucrose. After overnight storage at 4°C in 25% sucrose, the agarose block was incubated for 3 h in fresh 25% sucrose before being flash frozen in liquid nitrogen. 10 μm-thick sections were cut on a cryostat and collected on polysine slides (Menzel-Gläser, Braunschweig, Germany). The sections were washed twice for 5 min in PBS, incubated for 10 min with blocking solution[10% normal calf serum (NCS) in PBS], for 60 min with 1:10 F59 supernatant (a generous gift of Dr Frank E. Stockdale; Crow and Stockdale, 1986),which stains slow muscle cells in zebrafish(Devoto et al., 1996), in 3%NCS in PBS, washed three times for 5 min in PBS, incubated with 1:50 FITC-labelled goat anti-mouse antibody (DAKO, Hevelee, BE) in 3% NCS in PBS,washed three times for 5 min in PBS and mounted in Vectashield with propidium iodide (Vector Laboratories, Burlingame, CA, USA). Photographs were taken using an Olympus DP50 digital camera mounted on a Nikon MicroPhot microscope and AnalySIS software.

Embryos were fixed for at least one day in 4% PFA in PBS at 4°C,postfixed for 60 min in a 0.1 mol l^-1 cacodylate buffer (pH 7.2)containing 1% osmium (OsO₄), 2% glutaraldehyde and 1%K₂Cr₂O₇ on ice, rinsed twice with milliQ water and stored in 70% ethanol until embedding in epon.

Early in development, the nic^b107 embryos are indistinguishable from their wild-type siblings, apart from their immobility and non-inflated swimbladder. If the embryos are left in the egg capsule they will not hatch and, when removed after 48 hpf, they will remain curved laterally as a result of being curled up in one position in the egg capsule. For this reason, all embryos were taken out of the egg capsule at 24 hpf. From 72 hpf onwards, ∼50% of the mutant embryos show progressive dorsal curvature of the tail. Posterior to the anus, the curvature is evident(Fig. 2). All morphological measurements, except total length, were taken at anus level and are explained in Fig. 3A. Total length of the embryos increases from ∼2.5 mm to ∼4.0 mm(Fig. 3B). No significant differences between wild-type and nic^b107 embryos were detected. Muscle height increases from ∼150 μm to 260 μm in wild-type larvae and increases from 150 μm to 230 μm in nic^b107 embryos (Fig. 3C). The difference in height is significant from 48 hpf onwards. Notochord height increases from 40 to 70 μm in wild-type and to 65 μm in nic^b107 larvae (Fig. 3D). At 96 and 120 hpf, the difference in height is statistically significant (P<0.01). Somite size increases from ∼50 μm to∼100 μm in wild-type larvae, but in the nic^b107 the size increases slightly faster and, at 120 hpf, a significant(P<0.01) difference in somite size can be observed(Fig. 3E). The angle of the anal somite remains constant at ∼90° in both wild-type and nic^b107 embryos (Fig. 3E).

The muscle fibre directions at anus level for 48, 72 and 96 hpf are given in Fig. 4. The projection of the fibres is taken from anterior to posterior, with the anterior end of each fibre as the centre of a circle. The nic^b107 larvae(Fig. 4D–F) show a topology very similar to the wild-type larvae(Fig. 4A–C). Overall, the fibre orientations in wild-type and nic^b107 are similar,and helices can be discerned as early as 48 hpf. The nic^b107 fibre orientation patterns are somewhat more irregular and dorso-ventrally flattened at 96 hpf, with fibres displaying smaller elevation angles than in wild-type embryos, especially laterally(Fig. 4F). The lack of data near the myosepta and the different locations of the myosepta in different sections prevent an absolute comparison of the elevation and azimuth data between wild-type and nic^b107 embryos or to follow changes over time. For the fibres on the sections shown here, the mean absolute values for the elevation and azimuth angles are both ∼8°.

The fast and slow muscle fibre distribution can be visualised using antibodies. The distribution of fibre types in wild-type embryos(Fig. 5A) is indistinguishable from that in nic^b107 embryos(Fig. 5B). Slow fibres are present in a single cell layer in the periphery of the muscle mass, and fast fibres form the inner bulk of muscle.

Sarcomere architecture results in a repeated banding pattern in striated muscle, a result of stacking of sarcomeres on top of one another in a single myofibril (Fig. 6A,B). In wild-type embryos, myofibrils are positioned in parallel to one another over some distance. The sarcomeres in adjacent myofibrils are juxtaposed regularly and the banding pattern stretches out over multiple myofibrils. In the nic^b107 embryos, the myofibrils are arranged in parallel over shorter distances than in the wild type. This can be inferred from the tapering of the myofibrils (Fig. 6C,D). In addition, sarcomeres are present, but the stacking of sarcomeres is less regular than in the wild-type embryos. The sarcomere bands are not strictly juxtaposed, but a shift in banding between myofibrils in nic^b107 embryos is present(Fig. 6C,D).

Collagen type 1 α 2 expression is lower in nic^b107 than in wild-type embryos(Fig. 7A) and collagen type 2α 1 fluctuates around similar levels as in wild-type larvae, being slightly lower at 48 hpf and slightly higher at 96 hpf(Fig. 7B). Fast muscle specific myosin heavy chain 2 was expressed less in nic^b107 embryos than in wild-type embryos (Fig. 7D). Slow muscle specific myosin heavy chain 5 expression was similar in both groups, only at 48 hpf was a decrease seen(Fig. 7E). Two more fast muscle specific myosins were tested, light chain 2, whose expression is similar in both nic^b107 and wild-type embryos(Fig. 7C), and myosin light chain 3 (Fig. 7F), which showed a pattern similar to slow muscle specific heavy chain 5. Titin expression is elevated in the nic^b107 embryos at all stages examined(Fig. 7G). Both types of troponin C show lower expression in the nic^b107 embryos(Fig. 7H,I). Insulin-like growth factor (IGF) expression is not altered in nic^b107embryos (Fig. 7J). The expression of IGF receptor a (IGF-Ra) is elevated in the nic^b107 embryos only at 24 hpf(Fig. 7K), while the expression of IGF-Rb is elevated in the nic^b107 embryos at all stages examined (Fig. 7L). Expression of the muscle growth inhibitor gdf8 is elevated 100-fold in nic^b107 embryos at 24 hpf, is comparable with expression in wild type at 48 hpf and 72 hpf and is lower in nic^b107embryos at 96 hpf (Fig. 7M). Expression of the muscle growth stimulant myogenin is elevated in nic^b107 embryos at 24 and 96 hpf but is comparable with wild-type expression at the intermediate time points(Fig. 7N). Expression of muscle oxygen carrier myoglobin (Fig. 7O) is only slightly elevated at 96 hpf. Genes that represent tissue activity, such as muscle phosphofructokinase(Fig. 7P), which is involved in glycolysis, and succinate dehydrogenase complex, subunit A/flavoprotein(Fig. 7Q) and cytochrome b-5 reductase/NADHd (Fig. 7R), which are involved in oxidative phosphorylation, did not show significantly altered expression in the nic^b107embryos.

Altered tissue loading as a regulating factor in embryonic development has been the focus of relatively few studies. Some examples are endurance training(Bagatto et al., 2001) and immobilization by agar embedding (Van Raamsdonk et al., 1982) of zebrafish embryos and laser ablation to investigate local tissue forces in dorsal closure Drosophila embryos(Hutson et al., 2003). Altered tissue loading in adult tissues has received more attention because of its relevance for medicine, sports, aquaculture and space flight. The techniques used to alter tissue load are extremely diverse but can be separated into two categories based on their effect: decreased muscle activity (DMA) and increased muscle activity (IMA). Examples of DMA are reduced electrical activity through spinal cord isolation and the use of neurotoxins, spinal cord injury or transection, unloading (mainly of the hindlimb), immobilisation of muscle in a shortened position, space flight, mechanical ventilation and bed rest (Carlson et al., 1999; Gayan-Ramirez and Decramer,2002; Lalani et al.,2000; Talmadge,2000; Wehling et al.,2000). Examples of IMA are endurance (swim) training and chronic low-frequency stimulation of muscles (CLFS)(De Graaf et al., 1990; Kraus and Pette, 1997; Pette, 1998; Siu et al., 2004; Walker et al., 2004).

IMA and DMA have opposing effects on muscle, as is shown in Fig. 8. IMA promotes muscle growth amongst others through IGF (Adams,1998) and myogenin signalling(Hasty et al., 1993; Nabeshima et al., 1993; Rescan, 2001), and DMA inhibits muscle growth amongst others by stimulating gdf8 signalling(Lee, 2004; McPherron et al., 1997; Xu et al., 2000, 2003). Gdf8 signalling acts in part through downregulation of myogenin and IGF expression(Amali et al., 2004). In addition to promoting growth, a high relative expression of myogenin over MyoD promotes a shift towards a slow muscle phenotype(Talmadge, 2000). Overcrowding stress acts contradictively as it inhibits gdf8 expression in zebrafish but it also represses muscle growth (not shown in Fig. 8). This was suggested to result from a general depression of muscle protein synthesis that does not spare myostatin (Vianello et al.,2003).

The aim of the present study was to investigate the effects of DMA on the early development of the axial musculature in zebrafish. It was hypothesised that DMA would hamper proper muscle development. As a model for DMA, the nic^b107 mutant was used. Up to 18 hpf, wild-type and nic^b107 mutant embryos are indistinguishable by eye. During the first 18 h of development, a great many processes involving specification and differentiation of somite derivatives have already taken place. In the early hours after fertilisation, involution and convergent extension have created axial and paraxial mesoderm. The axial mesoderm induces local differentiation of the paraxial mesoderm, and slow muscle precursors have been specified. The paraxial mesoderm is segmented into somites from 10.5 hpf onwards, and, by 18 hpf, 18 somites have been formed(Kimmel et al., 1995; Stickney et al., 2000). The wild-type embryos now start making their first feeble coiling movements(Westerfield, 1995) by contraction of the muscle pioneer cells through cholinergic signalling(Melancon et al., 1997). By 21 hpf, they become sensitive to touch and respond with vigorous coiling, and, at 1 day post-fertilisation, fast and slow muscle are both activated in coiling behaviour (Buss and Drapeau,2002). Muscle action is conferred via cholinergic signalling through the release of acetylcholine (ACh), which, after binding to its receptor on the muscle cell, results in an action potential. After subsequent release of calcium in the muscle cell, the sarcomeres, the smallest functional units of muscle, contract. Sarcomeres are composed of thick(myosins) and thin (actin, tropomyosin and the troponin complex) filaments,titin and several other proteins (Marieb,2004). In the nic^b107 mutant, cholinergic signalling is defective because the acetylcholine receptor is not functional(Sepich et al., 1994, 1998; Westerfield et al., 1990), no action potentials are generated(Westerfield et al., 1990) and thus no calcium is released. Consequently, the nic^b107mutant embryos do not perform any movement, not even when stimulated (Sepich et al., 1994, 1998; Westerfield et al., 1990). The lack of signalling and ensuing lack of muscle contraction possibly result in DMA when tissues are still developing. This is expected to result in increased gdf8 expression, decreased myogenin and IGF signalling, a decrease in muscle growth and a slow-to-fast transition of muscle fibres, according to literature for adult muscle (Fig. 8).

The expression of myogenin in the nic^b107 embryos is higher than in wild-type embryos, which is not in line with the effects of DMA in Fig. 8. A study addressing the effects of innervation and blocking neuromuscular transmission in developing and adult rats reported that developmental innervation, although IMA-like, downregulates myogenin expression. Preventing developmental innervation, although it is DMA-like, prevented this downregulation(Witzemann and Sakmann, 1991). Blocking neuromuscular transmission (DMA-like) using toxins that leave the neuromuscular junction itself intact, and thus the possibility of communication, increased myogenin expression(Witzemann and Sakmann, 1991). In our study, developmental innervation takes place; only the neuromuscular transmission is defective. The observed increase in myogenin could thus result from a lack of downregulation. Along a similar line of reasoning, the increased IGF signalling, which also promotes muscle growth(Adams, 1998), may perhaps be viewed. IGF signalling is increased through increases in IGF-Rb expression. IGF-1 and IGF-Ra expression is not affected, which is consistent with effects of altered muscle activity by space flight(Lalani et al., 2000) and resistance training (Walker et al.,2004) but not mechanical ventilation, which decreased IGF-1 expression (Gayan-Ramirez and Decramer,2002). We also observed a transient increase in gdf8 expression,which is in line with Fig. 8,followed by a decrease by 96 hpf. A (transient) increase fits with mammalian literature (Carlson et al.,1999; Lalani et al.,2000). Since increases in opposing growth factor signalling are detected, the question is whether growth of muscle is promoted or inhibited. Titin expression, which is found in red and white muscle alike, is elevated and suggests that growth is promoted.

High relative expression of myogenin has been implied in a fast-to-slow transition in muscle fibres (Talmadge,2000). The observed downregulation of fast muscle specific myosin heavy chain 2 and troponin C expression is consistent with this idea. The downregulation of slow troponin C is not. Troponin C is involved in binding calcium that is released upon cholinergic signalling. In the absence of this calcium release in the nic^b107 mutants, there is no functional demand for troponin C in either muscle type. If a direct relationship between calcium release and troponin C expression exists, this can explain the downregulation of both fast and slow troponin C in the nic^b107 mutant. This would also suggest a downregulation of both types of troponin C in the relaxed mutant, which also lacks calcium signalling, despite successfully generating an electrical signal(Ono et al., 2001).

In zebrafish, increased activity at early life stages increased the mitochondrial content of red and intermediate muscle, suggesting an increased demand for oxidative metabolism (Pelster et al., 2003). Decreased activity resulted in a transient decrease in the activity of SDHa, an enzyme of the oxidative phosphorylation pathway(De Graaf et al., 1990). The nic^b107 mutant does not require energy for movement, only for basal cell metabolism, and therefore the energy-generating pathways may be less active in nic^b107 than in wild-type embryos. We assayed the expression of genes that are involved in cellular energy metabolism. Expression of myoglobin, which acts as a muscle oxygen carrier,was not changed. Glycolysis is the first step to generate energy, and muscle-specific phosphofructokinase is involved in this process. No differences in expression level were found between wild-type and nic^b107 mutant embryos. End products of glycolysis are oxidised in mitochondria in the oxidative phosphorylation pathway, a.o. by the enzymes SDHa and NADHd. These genes are also not altered in expression. The transient decrease in SDHa enzyme activity in adult zebrafish(De Graaf et al., 1990) may have been the result of (post)-translational control of activity or may occur only later in life. Together, these observations suggest that energy metabolism in the nic^b107 mutant is not grossly affected by the lack of muscle activity. Overall, it appears that the composition of developing muscle is altered at the molecular level, in line with DMA effects on adult muscle. Contrary to literature for adult animals, however, muscle growth and energy metabolism are not impeded by DMA during development(Fig. 8).

In addition to affecting muscle, muscle activity may also affect the tissues peripheral (skin) and central (notochord) to muscle by exerting force. Major force-resistant molecules in these tissues are collagens. DMA was expected to result in lower collagen expression in these tissues since less force is applied. The skin collagen fibrils consist mainly of collagen type 1(Dubois et al., 2002; Le Guellec et al., 2004). In fact, the bulk of expression of collagen type 1 α 2 is concentrated in the skin, with minor expression in other organs(Dubois et al., 2002). The expression of this gene is downregulated in nic^b107embryos, suggesting the skin to be less stiff in nic^b107than in wild-type embryos as a result of DMA. Consequently, the skin in the nic^b107 embryos may be less resistant to deformation. In the nic^b107 embryos, in the absence of muscle fibre activity, deformation can still be generated by growth. For a direct effect of muscle fibre activity, see below.

The notochord is a fluid-filled column that stiffens the body axis because it is osmotically pressurised (Adams et al., 1990). Collagen type 2 α 1 (col2α1) is a major component of the notochordal sheet (Adams et al., 1990) during the first 24 h of development in the zebrafish (Yan et al., 1995)and has been suggested to counteract the internal pressure(Aszódi et al., 1998). The expression of col2α1 is not different between nic^b107 and wild-type embryos at 24 hpf, which suggests unchanged stiffness between wild-type and nic^b107embryos.

From 72 hpf onwards, about half of the mutant embryos show curvature of the body with their tails curved up. Abnormal body curvature is a common feature in mutants in the diverse mutagenesis screens. These mutants usually have additional defects. Abnormal body curvature is especially common in mutants with defects in notochord (Stemple et al.,1996) and midline structures(Brand et al., 1996),suggesting that these tissues normally stabilise a straight body shape or induce tissues that do so. Searching the 2966 entries in the zfin database(Sprague et al., 2001) for body curvature (tail up or down and bent body) results in 209 hits. Only one of these is immobile (Brand et al.,1996). Vice versa, out of 145 hits for `nonmotile' or`reduced motility', only 10, representing three loci, show changes in body curvature (Granato et al.,1996). This suggests that there is not a one-to-one relationship between body curvature and mobility, which is also indicated by the fact that the body curvature phenotype is not fully penetrant in the nic^b107 mutants.

Somite size, measured ventrally in the anal somite, is larger in nic^b107 embryos than in wild-type embryos at 120 hpf, a possible result of lower gdf8 expression in nic^b107embryos (Amali et al., 2004). Total length measured along the notochord did not differ. This incongruity correlates with the presence of progressive dorsad curvature of the tail in half of the nic^b107 embryos(Fig. 2B). Despite this total body curvature, the somite angle did not differ significantly between groups. Other authors found larger somite angles and smaller somite sizes at 72 hpf for immobile zebrafish embryos (Van Raamsdonk et al., 1977). They looked at spontaneously immobile embryos and also induced immobility by removal of the brain or making a midbody lesion. Both types of immobile embryos in these experiments were morphologically different from the untreated embryos, e.g. they had shorter,less developed tails (Van Raamsdonk et al., 1977). These results are therefore difficult to compare with ours, in which development was not altered by surgical procedures. Morphometric measurements are not part of the large genetic screens and therefore this type of data is scarce. Experiments with spontaneously immobile tadpole larvae show body curvature with the tail pointing dorsally and larger somite size (Droin and Beauchemin,1974), similar to our findings. Somite angle became smaller in these embryos, whereas total length increased(Droin and Beauchemin, 1974),which contrasts with our results. There appears to be a correlation between the lengthening of the embryonic body and the angle of the somites. Shorter bodies correlate with larger angles (Van Raamsdonk et al., 1977), longer bodies correlate with smaller somite angles (Droin and Beauchemin,1974) and, in our experiments, neither of the two changed. The larger somite size in nic^b107 embryos can be a result of activity of the muscle fibres. Muscle fibres shorten when active and pull adjacent myosepta together, effectively decreasing somite size. The inactive muscle fibres in the nic^b107 embryos will not pull adjacent myosepta together, resulting in larger somite sizes.

No indications for notochord differences between wild-type and nic^b107 embryos were found at the molecular level (see above), but in wild-type embryos extra pressure may be generated by the muscle fibre contractions that exert a compressive stress on the notochord in a longitudinal direction. This activity of the muscle fibres acts to raise internal pressure, which in turn results in greater mechanical stiffness of the notochord (Adams et al.,1990). A more circular transverse circumference is then adopted automatically in the pressurised notochord, as is seen in wild-type embryos. In addition, in a simulation of muscle fibre activity in adult fish, Van Leeuwen (1999) proposed that the external shape of a fish body could be affected by skin stiffness. In simulations with reduced skin stiffness, muscle tissue moved inwards close to the mid-horizontal plane and outwards in the dorsal and ventral regions upon activation of the muscle fibres, leading to two concavities near the mid-horizontal plane (Van Leeuwen,1999). The embryonic skin consists of only two layers of cells with few collagen fibrils (Le Guellec et al., 2004) and can therefore be expected to be weak. This holds for both the wild-type and the nic^b107 embryos, but muscle fibre activity occurs only in the wild-type embryos. Indeed, the predicted concavities are found in the wild-type embryos, especially at 48 and 72 hpf,but are less prominent in nic^b107 embryos, despite the fact that they have a potentially weaker skin, suggesting a direct effect of DMA on cross-sectional body shape.

Wild-type zebrafish embryos start to make movements around 18 hpf, and this is also the time that slow muscle fibres migrate laterally through the somite towards the muscle periphery (Devoto et al., 1996). Agar immobilisation of zebrafish embryos resulted in a mosaic pattern of slow and fast fibres in the bulk of fast muscle(Van Raamsdonk et al., 1982),which may be the result of altered migration of slow muscle fibres or phenotypic switching of fast muscle fibres. In any case, it implies altered ratios of slow and fast fibres, i.e. of slow and fast myosins. The expression of a fast-twitch myosin heavy chain is indeed affected(Fig. 7), which is in line with this idea. We therefore stained slow fibres in zebrafish to check for possible mosaic formation of slow fibres in the bulk of fast fibres in situ. The muscle fibres in nic^b107 embryos are normally segregated as slow and fast fibres and no mosaic is found(Fig. 5). Apparently, genetic immobilisation is different from agar immobilisation. In 5% agar, growth defects become apparent; 0.5% agar allows for growth but also for body bending(Van Raamsdonk et al., 1979). Firing of neurones is not inhibited by embedding. Indeed, in 1% agar, the embryos are able to contract their muscles and they will make shrugging movements (T.v.d.M., unpublished), even when the action does not result in body deformation and embryos appear immobile(Van Raamsdonk et al., 1979). It is not implausible that embedded embryos will try to escape the agar confinement or try to refresh the oxygen-poor stagnant agar boundary layer. In their efforts to do so, they may even show IMA. IMA leads to a fast-to-slow transition in muscle fibres (Fig. 8) and could explain the mosaic formation. The difference therefore lies in the activity of muscles, i.e. whether the immobility is intrinsic (in the nic^b107 mutant) or extrinsic (in agarose embedding). Our results showed that in the nic^b107 mutant,fast and slow muscle fibres segregate correctly. Muscle fibre type is determined largely by sarcomere components(Fig. 8), indicating that these are assimilated correctly.

Another feature of muscle development is the development of a pseudo-helical arrangement of white muscle fibres. This pseudo-helical arrangement in adult fish is thought to allow equal amounts of work by each individual fibre of a single myomere and can thus be viewed as an optimisation of the fibre architecture to activity(Alexander, 1969). The development of this pattern occurs only after the embryos make their first movements. We found helical arrangements as early as 48 hpf, which is four days earlier than previously reported (Van Raamsdonk et al., 1974). As both wild-type and nic^b107 embryos arranged their muscle fibres in helical patterns, it appears that the mechanical loading of the early parallel fibres is not necessary for the development of a helical pattern. Loading may be necessary for further development of muscle fibre architecture, as helices in nic^b107 mutants appear less organised and the change in body shape concomitantly affects the shape of the helix. As a result, at 96 hpf, the nic^b107 lateral muscle fibres have a smaller elevation angle. This can be seen in Fig. 4C,vsFig. 4F: the line segments indicating fibre direction have a greater vertical component in lateral wild-type fibres than in lateral nic^b107 mutant fibres. The larger elevation angles in the lateral wild-type fibres are necessary for equalising the strains over the fibres in a single myotome, which forms the basis of the optimised fibre architecture (Alexander, 1969;J.L.v.L., unpublished). In conclusion, the pseudo-helical arrangement of muscle fibres develops in the case of DMA but is affected by it.

Within muscle fibres, stacks of sarcomeres are arranged in myofibrils. DMA in the nic^b107 mutant, results in apparently normal sarcomeres, but the stacking of sarcomeres on top of one another and the myofibril organisation are less regular(Fig. 6). A similar phenotype is observed in the twister mutant. The IMA in this mutant results from increased synaptic decay times, due to a dominant mutation in the αsubunit of the acetylcholine receptor. Contractile filaments appear disorganised and myofibril organisation is severely impaired(Lefebvre et al., 2004). This shows that basal components, such as the sarcomeres, develop more or less normally, but the integration of multiple basal components into a higher-level architecture goes astray when muscle activity deviates from the normal.

Despite the lack of muscle fibre use, slow and fast muscle fibres develop,the intricate sarcomere architecture is built and the necessary genes are transcribed. It appears therefore that the information that is needed to develop and differentiate the `basal components' of muscle does not depend on external signalling from muscle activity. The DMA that results from a lack of cholinergic signalling during development distinguishes nic^b107 embryos from wild-type embryos, however, as body shape, gene expression levels, sarcomere stacking and myofibril arrangement are altered. In conclusion, DMA affects morphological as well as genetic parameters during early development, influencing differentiation of basal components.

When muscle activity is viewed as an, as yet, unknown pathway that affects muscle development and differentiation, it can be interfered with at several steps that result in DMA or IMA. The nic^b107 mutant, a model for DMA, has defective electrical signalling, and downstream muscle activities are affected as well. These include the release of calcium and force production by muscle contraction. In the relaxed mutant,electrical signalling is intact, but a defect in dihydropyridine receptor-mediated release of internal calcium renders them motionless(Ono et al., 2001). Comparing these two mutants will provide information on the importance of the electrical signal in the muscle activity pathway. The twister mutant provides a model for IMA. It has a dominant mutation in the α subunit of the acetylcholine receptor that increases synaptic decay times and causes uncoordinated movements (Lefebvre et al.,2004). It has already been shown to have profound effects on muscle development, but data on the levels of expression of genes that appear affected morphologically are currently not available(Lefebvre et al., 2004).

The authors would like to thank Anja Taverne-Thiele for excellent technical assistance at EM work, De Haar Vissen for taking care of the fish and two anonymous referees for useful suggestions on the manuscript. Fish from the nic^b107 strain were supplied by ZIRC under NIH-NCRR grant number RR12546.