Molt cycle regulation of protein synthesis in skeletal muscle of the blackback land crab, Gecarcinus lateralis, and the differential expression of a myostatin-like factor during atrophy induced by molting or unweighting

Skeletal muscle is a highly adaptable tissue. In mammals, muscle mass is regulated by an array of physiological and pathological conditions, such as nervous activity, hormones, passive stretching, disuse and disease (reviewed by Favier et al., 2008; Tisdale, 2009). In decapod crustaceans, skeletal muscle retains a great deal of plasticity in the adult (reviewed by Mykles, 1997), which is exemplified in a reversible atrophy that occurs under two disparate conditions. One of these is the atrophy of claw muscle as part of a developmental program leading up to ecdysis, or molting (reviewed by Mykles and Skinner, 1982a). A reduction in mass facilitates withdrawal of the large closer muscle in the propodus through the narrow basi-ischial segment at ecdysis (reviewed by Mykles, 1999). Both Ca²⁺-dependent and ubiquitin-dependent proteolytic systems are elevated in atrophic claw muscle (reviewed by Mykles, 1999). Calpains (Calp) degrade myofibrillar proteins and cDNAs encoding three G. lateralis calpains (Gl-CalpB, -M and -T) have been characterized; eyestalk ablation causes a transient increase in Gl-CalpT and ecdysone receptor (Gl-EcR) expression (Kim et al., 2005a). An extensive remodeling of the contractile apparatus is coincident with claw muscle atrophy, but no change in fiber phenotype occurs (Ismail and Mykles, 1992; Mykles and Skinner, 1981; Mykles and Skinner, 1982b). Protein synthesis is elevated during premolt (Skinner, 1965). It is hypothesized that the resulting increase in protein turnover facilitates rearrangement of the myofilament lattice in atrophic muscle (Mykles, 1997). The fully reversible atrophy of the claw muscle during premolt is induced by elevated ecdysteroid titers in the hemolymph and does not occur in leg or thoracic muscles (reviewed by Mykles and Skinner, 1982a; Mykles, 1999). Claw and thoracic muscles express different assemblages of Gl-RXR (retinoid X receptor) isoforms, suggesting that the two muscles differ in response to ecdysteroids (Kim et al., 2005b).

Reversible atrophy of thoracic muscles occurs during intermolt in response to unweighting, which is caused by autotomy, or voluntary loss, of the associated limb in three decapod species (Carcinus maenas Linnaeus 1758, Uca pugilator Bosc 1802 and Procambarus spp.) (Moffett, 1987; Schmiege et al., 1992). Continued atrophy despite coordinated movement of the regenerating limb bud indicates that mass of thoracic muscles is not strictly regulated by activity. Furthermore, atrophy of thoracic muscles in response to limb autotomy does not appear to involve ecdysteroid-dependent signaling, as precocious molting is not induced if less than five limbs are autotomized in G. lateralis (Skinner and Graham, 1972). Ultrastructural examination of atrophic thoracic muscle demonstrates that the reduction in fiber area is associated with disruption of sarcomere organization, loss of organelles, and increased erosion of myofilaments (Schmiege et al., 1992). This stands in contrast to the highly ordered restructuring of muscle fibers in the premolt claw (Ismail and Mykles, 1992; Mykles and Skinner, 1981), and suggests that proteolytic regulation in atrophic thoracic muscle differs from that of the claw muscle.

A myostatin-like protein is expressed in G. lateralis (Gl-Mstn), and may be involved in regulating muscle mass in crustaceans. Myostatin, a member of the transforming growth factor-β (TGF-β) superfamily of cytokines, is a negative regulator of muscle mass in mammals (reviewed by Lee, 2004; Rodgers and Garikipati, 2008), the function of which is conserved in at least one species of lower vertebrate (Medeiros et al., 2009). Expression in non-muscle tissues does, however, suggest a broader regulatory role for myostatin, especially in lower vertebrates (reviewed by Rodgers and Garikipati, 2008). Like other TGF-β family members, myostatin is secreted as a latent autocrine/paracrine factor, and is activated by proteolytic cleavage of the propeptide; the mature peptide binds to activin receptors in the plasma membrane as a homodimer (reviewed by Herpin et al., 2004; Lee, 2004). Activation of these serine/threonine kinase receptors initiates signaling through Smad transcription factors (reviewed by Liu, 2003; Xu, 2006) that initiate a shift in the balance between anabolic and catabolic pathways in skeletal muscle, thereby causing a net loss of muscle protein. In mammals, myostatin stimulates ubiquitin-dependent protein degradation and inhibits target of rapamycin (TOR)-dependent protein synthesis (reviewed by Matsakas and Patel, 2009; Tisdale, 2009). Fiber atrophy is induced by elevated expression of myostatin (Amirouche et al., 2009; Durieux et al., 2007; Reisz-Porszasz et al., 2003; Zimmers et al., 2002). Conversely, fiber hypertrophy is induced by reduced myostatin signaling (Magee et al., 2006; Medeiros et al., 2009; Nakatani et al., 2008; Qiao et al., 2008). The Gl-Mstn cDNA encodes a 497-amino-acid (aa) polypeptide that includes a 110-aa mature peptide domain which has 52% identity and 69% similarity to human myostatin (Covi et al., 2008b). The function of this putative crustacean myostatin is unknown, but its predominant expression in skeletal muscle of the land crab suggests a role in regulating muscle mass.

The regulatory systems responsible for inducing and reversing the atrophy of skeletal muscle in crustaceans are largely unknown. Atrophy of claw muscle during premolt is triggered by ecdysteroids, whereas atrophy of thoracic muscle during intermolt is ecdysteroid-independent and may be triggered by reduced load. The purpose of this study was to examine the role of Gl-Mstn in the atrophy of claw and thoracic muscle. We hypothesized that Gl-Mstn expression would increase in atrophic skeletal muscle, and that an Mstn/Smad signaling pathway would upregulate the expression of ecdysteroid receptor and Gl-CalpT. Expression of elongation factor 2 (Gl-EF-2), a ‘housekeeping’ gene, would not be affected. Molting was induced by eyestalk ablation (ESA) or multiple limb autotomy (MLA) (reviewed by Mykles, 2001; Skinner, 1985b). Gene expression was assessed in claw closer, weighted thoracic and unweighted thoracic muscles over the molting cycle. Transcripts for Gl-Mstn (Covi et al., 2008b), ecdysteroid receptor (Gl-EcR and Gl-RXR) (Kim et al., 2005a; Kim et al., 2005b), Gl-EF-2 (Kim et al., 2004) and Gl-CalpT (Kim et al., 2005a) were quantified by real-time reverse transcription-polymerase chain reaction (qRT-PCR). Protein synthesis was measured by incorporation of [³⁵S]methionine in claw closer muscle cultured in vitro. Protein synthesis and Gl-Mstn mRNA were inversely correlated with hemolymph ecdysteroid levels. By contrast, unweighting increased Gl-Mstn expression in thoracic muscle from intermolt animals. Regulation of muscle atrophy clearly differs between the two muscles. However, analysis of the data suggests that Gl-Mstn regulates protein turnover in molt-induced and disuse muscle atrophies.

Adult male Gecarcinus lateralis (Fréminville 1835) used in [³⁵S]methionine incorporation experiments were obtained from Bermuda and maintained as described previously (Skinner, 1962); molting was induced by MLA. Adult animals used in qRT-PCR experiments were obtained from the Dominican Republic and maintained at 22°C in 75–90% relative humidity. It is relevant to note that G. lateralis from Bermuda, Puerto Rico and the Dominican Republic all experience similar atrophy of claw musculature during premolt. Thus, differences between Bermuda and Dominican Republic populations are assumed to be insignificant. Intermolt animals were kept in communal (7–12 animals per cage) plastic cages lined with aspen bedding wetted with 5 p.p.t. Instant Ocean (Aquarium Systems, Mentor, OH, USA). Limb-autotomized animals were kept in individual quart size (∼1 l) plastic cages lined with sand wetted with 10 p.p.t. Instant Ocean. Cages for limb-autotomized animals were shielded from room lighting with cloth, as constant darkness shortens the premolt period (Bliss and Boyer, 1964). Animals were maintained under a 12 h:12 h light:dark cycle and fed iceberg lettuce, carrots and raisins twice weekly. Animals used in qRT-PCR experiments were kept in captivity for less than 2 months prior to harvesting of tissues.

Precocious molting is induced by ESA and autotomy of at least five walking legs (reviewed by Mykles, 2001; Skinner, 1985b). Regeneration of autotomized limbs occurs in two phases: basal growth, which occurs during intermolt, and proecdysial growth (Hopkins, 1993). Progression through premolt in eyestalk-ablated and multiple limb-autotomized crabs was monitored by measuring the length of the limb regenerate at the position of the third walking leg and calculating the regeneration index [R-index; =length of regenerate×100/carapace width (Bliss, 1956)]. In G. lateralis, the R-index ranges between 0 and 24 (Holland and Skinner, 1976; Yu et al., 2002). Eyestalk-ablated animals missing more than three limbs by the time of dissection were not used. All eight walking limbs were autotomized for the MLA experiments, and the weighted thoracic muscle harvested in this experiment was associated with the two intact chelipeds. ESA was done as described previously (Covi et al., 2008a).

Penicillin and streptomycin were obtained from Sigma (St Louis, MO, USA). All other reagents used in [³⁵S]methionine incorporation experiments were the same as described previously (Mykles, 1990). [³⁵S]methionine (specific activity ∼4.07×10¹³ Bq mmol^–1) was purchased from ICN Biochemicals (Costa Mesa, CA, USA).

Autotomized claws were packed in crushed ice for 3–4 h to induce apolysis (O'Brien et al., 1986). Claw closer muscle was cultured as described previously (Mykles, 1990). Briefly, tissue was equilibrated for 30–60 min in a filter-sterilized supplemented culture medium containing 10 mmol l^–1 Hepes–NaOH (pH 7.4), 316 mmol l^–1 NaCl, 5.4 mmol l^–1 KCl, 8.8 mmol l^–1 CaCl₂, 6.8 mmol l^–1 MgSO₄, 1.0 mmol l^–1 sodium phosphate, 5.6 mmol l^–1d-glucose, 13.6 mmol l^–1 sodium citrate, 1.0 mmol l^–1 pyruvic acid, 4 mmol l^–1 glutamine, 0.2 mmol l^–1l-alanine, 0.6 mmol l^–1l-arginine, 0.34 mmol l^–1l-asparagine, 0.3 mmol l^–1 aspartic acid, 0.1 mmol l^–1l-cystine, 0.3 mmol l^–1l-glutamic acid, 0.3 mmol l^–1 glycine, 0.2 mmol l^–1l-histidine, 0.4 mmol l^–1l-isoleucine, 0.4 mmol l^–1l-leucine, 0.4 mmol l^–1l-lysine, 0.2 mmol l^–1l-phenylalanine, 0.3 mmol l^–1l-proline, 0.3 mmol l^–1l-serine, 0.4 mmol l^–1l-threonine, 0.05 mmol l^–1l-tryptophan, 0.4 mmol l^–1l-valine, penicillin (10 i.u. ml^–1), and streptomycin (10 μg ml^–1). After equilibration, the medium was replaced with fresh medium containing [³⁵S]methionine (3.7×10⁵ Bq ml^–1) and the culture was incubated on an orbital shaker for 12 h. Equilibration of radioactively-labeled amino acid occurs within 1 h (Skinner, 1965) (M. Haire and D.L.M., data not shown).

After incubation, tissues were frozen in liquid N₂ and stored at –80°C. Muscles were homogenized in 6 volumes of buffer A (20 mmol l^–1 Tris–HCl, pH 7.5; 20 mmol l^–1 KCl; 1 mmol l^–1 EDTA; 1 mmol l^–1 dithiothreitol) with a Polytron homogenizer for 30 s at high speed. The homogenates were centrifuged at 16,000 g for 20 min at 4°C. The soluble protein in the supernatant fraction was dialyzed against two 1-l changes of 20 mmol l^–1 ammonium acetate, frozen, lyophilized, and dissolved in 1–2 ml SDS sample buffer (Medler et al., 2007) at 60°C for 10–15 min. The myofibrillar protein in the pellet obtained from centrifugation was washed three times with 40 ml buffer A and extracted with five volumes 0.6 mol l^–1 NaCl and 5 mmol l^–1 sodium phosphate (pH 7.4) at 4°C for 30 min. The soluble and myofibrillar protein solutions were clarified by centrifugation at 16,000 g for 15 min at 4°C.

The filter paper disk assay (Bollum, 1968) was used to quantify the incorporation of radioactivity into protein. Protein concentration was determined by fluorescence emission spectroscopy (Avruch and Wallach, 1971). Samples (100 μg) of soluble and myofibrillar proteins were applied to 1.5–2 cm squares of 3MM paper and air dried. Filters were washed briefly in 10% trichloroacetic acid (TCA) at room temperature, washed in 5% TCA at 90°C for 15 min, rinsed in 95% ethanol at room temperature, and air dried. Filters were placed in vials containing Ecoscint O (National Diagnostics, Atlanta, GA, USA) and the radioactivity was measured with a Beckman LS 7000 liquid scintillation counter. Incorporation was expressed as nmol [³⁵S]Met mg^–1 protein h^–1. For autoradiography, proteins were separated using 10% SDS-PAGE (Medler et al., 2007) and exposed to X-ray film at –80°C.

All reagents were of molecular biology grade. Animals were anesthetized by immersion in crushed ice for 5 min prior to dissection. Claw closer and thoracic muscles were harvested within 10 min, frozen in liquid N₂ and stored at –80°C. Weighted and unweighted thoracic muscle was determined by the presence or absence, respectively, of the associated pereiopod. Sufficient time had elapsed after autotomy for a basal limb regenerate to have formed.

Total RNA was isolated using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. Briefly, total RNA was treated with DNase I (Invitrogen) to remove genomic DNA. After DNase treatment, total RNA was subjected to phenol:chloroform:isoamyl alcohol extraction (25:24:1), followed by precipitation of RNA to remove DNase, salts and digested DNA. Total RNA was precipitated with 1 volume of isopropanol and resuspended in 30 μl of RNA Storage Solution (Ambion, Austin, TX, USA). Nucleotide concentration was determined by absorbance at 260 nm after 1:1 dilution with Tris-EDTA (TE) buffer (Integrated DNA Technology), and 3.75 μg was used for 30 μl cDNA synthesis reaction using SuperScript III reverse transcriptase (Invitrogen) and oligo-dT(22)VN primer (50 μmol l^–1). Resulting cDNA was treated with RNase H (New England Biolabs, Ipswich, MA, USA) to remove complementary RNA.

Quantitative analysis of Gl-Mstn [GenBank EU432218 (Covi et al., 2008b)], Gl-EF-2 [GenBank AY552550 (Kim et al., 2004)], Gl-EcR [GenBank AY642975 (Kim et al., 2005a)], Gl-RXR [GenBank DQ067280 (Kim et al., 2005b)] and Gl-CalpT [GenBank AY639154 (Kim et al., 2005a)] was conducted using LightCycler FastStart DNA Master Plus SYBR Green I reaction mix and a LightCycler 480 thermal cycler (Roche Applied Science, Indianapolis, IN, USA). Reactions consisted of 1 μl first strand cDNA, 5 μl 2× SYBR Green Master Mix, 0.5 μl each of 10 mmol l^–1 forward primer and 10 mmol l^–1 reverse primer, and 3 μl PCR-grade water. PCR conditions included an initial denaturation at 95°C for 5 min, followed by 45 cycles of denaturation at 95°C for 5 s, annealing at gene-specific temperature (Table 1) for 5 s, and extension at 72°C for 20 s. Gl-Mstn primers were designed using the 3′ end of the mature peptide domain (Covi et al., 2008b), and, thus, amplified all splice variants that encoded a full-length mature peptide. Primers for RXR were designed using a sequence common to all known splice variants (Kim et al., 2005b). At the end of each run, an analysis of the PCR product melting temperature was conducted. Transcript concentrations were calculated with LightCycler 480 software (Roche; version 1.2) using standards curves produced by serial dilution of purified PCR product (10 ag μl^–1 to 10 ng μl^–1). The calibrator template used to verify consistency between runs contained cDNA pooled from claw muscle, thoracic muscle, heart, hepatopancreas, eyestalk ganglia, thoracic ganglia, testis and ovary. Ecdysteroid levels in the hemolymph were quantified by radioimmunoassay as described previously (Medler et al., 2005).

Statistical analysis was performed using JMP 5.1.2 software (SAS Institute, Inc., Cary, NC, USA). qRT-PCR data were log transformed prior to analysis to reduce differences in variance from the mean. Variances among log-transformed data were determined to be equivalent using Brown–Forsythe tests (P<0.05). Means for transcript abundance were compared among developmental stages for individual muscle types using analysis of variance (ANOVA). Post-hoc multiple comparisons were made using Tukey–Kramer HSD tests. A paired t-test was used to compare means for transcript abundance between claw and thoracic muscles, or weighted and unweighted thoracic muscles, at each developmental stage. All data not plotted as individual points are presented as mean ± 1 s.e.m. and the level of significance was set at α=0.05 for all statistical analyses.

Data are presented as a function of three variables: time, R-index and molt stage. Time was used to describe events following ESA and ecdysis, but did not allow comparisons with premolt animals that were induced to molt by MLA. R-index was used to compare the effects of molt induction by ESA and MLA during premolt. The molt stage classifications developed by Drach (Drach, 1939), and later modified by Skinner on G. lateralis (Skinner, 1962), were used to integrate analyses based on time and R-index.

For anecdysial, or intermolt (stage C₄) animals, [³⁵S]methionine incorporation into soluble protein was 2.7-fold greater than incorporation into myofibrillar proteins (0.167 vs 0.063 nmol [³⁵S]Met mg^–1 protein h^–1; Fig. 1A; P<0.0001). During early premolt (R-index=8 to 13), incorporation began to increase, reaching synthetic rates for myofibrillar and soluble proteins that were 11 and 13 times greater (P<0.0001), respectively, by mid premolt than those in claw muscles from intermolt animals (Fig. 1A). The stoichiometry between the synthesis of soluble and myofibrillar proteins changed during the late D₀ stage of premolt; the synthesis of soluble protein increased at a greater rate than myofibrillar protein (1.4 to 1) during the first half of stage D₀, but the relative increase in synthetic rates for the two protein pools were equivalent thereafter (Fig. 1B). Slopes for linear regressions of myofibrillar and soluble synthetic rates at R<13 and R>13 are significantly different (P<0.05), and the synthesis of soluble and myofibrillar proteins within the same muscle was highly correlated (Fig. 1B). Overall, the synthesis of soluble protein during premolt was about 3.2-fold greater than that of myofibrillar protein (Fig. 1A).

SDS-polyacrylamide gel electrophoresis showed no quantitative or qualitative changes in the protein composition of soluble and myofibrillar preparations during premolt (Fig. 2A,B). Autoradiographs showed increased global synthesis of soluble and myofibrillar proteins (Fig. 2C,D). There were no qualitative differences in the proteins synthesized at any molt stage. Furthermore, the relative synthesis of each of the proteins in the soluble and myofibrillar fractions did not change during the premolt period (Fig. 2C,D).

Titers of hemolymph ecdysteroids increased in response to both ESA and MLA, but were higher in ESA animals at all R-index values sampled (compare Fig. 3A and Fig. 4A). ESA induced a five-fold increase in ecdysteroids over the first 24 h that was followed by a much slower and prolonged increase; by 2 weeks post-ESA, the hemolymph ecdysteroid level was 12-fold higher than that observed in intact animals (Fig. 3A; R=15.5). During the third week following ESA, ecdysteroids increased rapidly again, and were 32-fold greater than levels in intact animals by day 20 (Fig. 3A; R=19.9). The initial increase in hemolymph ecdysteroids observed in response to MLA was more gradual than that observed for ESA. Mean ecdysteroid levels were only 7-fold greater than in intact animals 45±11 days after MLA (Fig. 4A; R=15.1). Hemolymph ecdysteroids increased rapidly in late premolt (R=23.4), reaching a concentration that was 28-fold greater on average than that in intermolt animals. Ecdysteroid levels decreased to <7 ng ml^–1 in postmolt animals and remained low for at least 10 days following ecdysis (Fig. 4A).

ESA had small or only transient effects on Gl-EF-2 expression during the first two weeks post-ESA. The copy number of mRNA transcripts for Gl-EF-2 did not change significantly in claw muscle during the first 14 days following ESA (Fig. 3C; R≤17). However, by 20 days post-ESA (R=20), Gl-EF-2 levels had increased 2.1-fold relative to that in intact animals (Fig. 3C). In thoracic muscle, Gl-EF-2 copy number decreased by 35% during the first 24 h following ESA and returned to intact levels within 2 weeks (Fig. 3C). These levels are similar to previously published values (Kim et al., 2005a). Transcript copy number for Gl-EF-2 was consistently lower in thoracic muscle than in claw muscle at all molt stages (Fig. 3C).

ESA caused a large reduction in Gl-Mstn expression, but the timing and extent of this reduction differed between claw and thoracic muscles. No significant change in the number of copies of Gl-Mstn mRNA occurred in claw muscle during the first 3 days following ESA (Fig. 3D; R≤8). However, the beginning of a downward trend in Gl-Mstn transcript abundance in claw muscle was apparent at day 7 post-ESA (R=10). By day 20 (R=19.9), Gl-Mstn transcript level had decreased by 81% (Fig. 3D). No significant change in Gl-Mstn copy number was apparent until day 20 post-ESA in thoracic muscle (R=19.9), and the magnitude of the decrease in copy number (68%) was smaller than that observed for claw muscle (Fig. 3D). This difference in the response to ESA is supported by statistical comparison of transcript abundance in the two muscle tissues at each molt stage; transcript copy number for Gl-Mstn was significantly lower in thoracic muscle in intact and early D₀ premolt animals (R≤7.2), but was significantly higher in thoracic muscle at R=15 (Fig. 3D).

Effects of ESA on the expression of the subunits of the ecdysteroid receptor (Gl-EcR and Gl-RXR) were dependent on tissue and subunit type. ESA had no significant effect on transcript abundance of Gl-EcR in claw muscle, but induced a steady and significant increase in thoracic muscle (Fig. 3B). By day 20 (R=19.9), Gl-EcR transcript levels had increased by 3.2-fold relative to the level in intact animals (Fig. 3B). It is relevant to note that variances around the mean for days 1 and 3 following ESA were higher relative to those at later time points and for other gene transcripts (Fig. 3). Consequently, for Gl-EcR trends spanning these earlier time points could be obscured by the high degree of variation among individuals during the immediate response to ESA. Copy number for Gl-EcR in thoracic muscle was significantly lower than that of claw muscle at low and high R-index values. Transcript copy number for Gl-RXR did not change significantly in claw muscle during the first 3 days following ESA (Fig. 3E; R≤8). Gl-RXR transcript levels dropped 51% between day 3 and 7 post-ESA (R=7.2 and 10, respectively), and slowly returned to normal levels thereafter (Fig. 3E). A similar trend was apparent in thoracic muscle, although the drop between day 3 and 7 post-ESA was slightly larger (65%) and the subsequent return to normal levels was not observed (Fig. 3E). It is important to note that the qRT-PCR primers amplified a region conserved among all known isoforms. Therefore, the data are the sum of all Gl-RXR variants, which differ in expression between claw and thoracic muscles (Kim et al., 2005b).

After a transient decrease, ESA caused a modest increase in Gl-CalpT expression in claw and thoracic muscles; this was followed by a decrease at later premolt stages. Transcript abundance for Gl-CalpTdecreased by 40–50% during the first 24 h following ESA in both muscles, but this decrease was only statistically significant in thoracic muscle (Fig. 3F). Copy number for Gl-CalpT returned to the levels in intact animals in both tissues by day 3 post-ESA (R=7.2). Transcript levels in claw muscle increased by 2.7-fold between days 1 and 14 post-ESA (R=6.6 and 15.5, respectively) whereas levels in thoracic muscle increased by 1.7-fold (Fig. 3F). Transcript abundance decreased between day 14 and day 20 (R=15.5 and 19.9, respectively) post-ESA in both muscles, but this drop was only statistically significant in thoracic muscle (Fig. 3F).

The expression of Gl-EcR and Gl-RXR was examined following MLA. Levels of mRNA were similar between claw and thoracic muscles at all molt stages except 10 days postmolt for both Gl-EcR and Gl-RXR (Fig. 4B,E). Gl-EcR transcript levels in both muscles increased by two- to threefold during premolt, decreased by 95% between late premolt and 2 days postmolt, and remained low for at least 10 days after ecdysis (Fig. 4B). No significant change in Gl-RXR mRNA was observed in either muscle over a molting cycle (Fig. 4E). However, levels of both Gl-EcR and Gl-RXR transcripts were significantly higher in thoracic muscle from 10-day postmolt animals (Fig. 4B,E).

In response to MLA, the abundance of Gl-EF-2 transcripts decreased by 42% in claw muscle and 40% in thoracic muscle during the early D₀ stage of premolt, but this change was only statistically significant for the thoracic muscle (Fig. 4C; R=10). Gl-EF-2 transcript abundance in both muscles returned to intact levels by the late D₀ stage of premolt (Fig. 4C; R=15). Gl-EF-2 levels were generally lower in claw muscle than thoracic muscle (Fig. 4C).

Transcript abundance for Gl-Mstn decreased in both claw and thoracic muscle following MLA, but this decrease was more pronounced in claw muscle. This is illustrated by a direct comparison between claw and thoracic muscles at specific molt stages; Gl-Mstn mRNA levels for claw muscle were significantly greater than that of thoracic muscle during the early D₀ stage of premolt (R=10), but significantly less than thoracic muscle during the late D₁ stage of premolt (R=23.5) and 10 days postmolt (Fig. 4D). Gl-Mstn copy number in the high R-index group was 94% lower than that of intact animals for claw muscle and 82% lower for thoracic muscle (Fig. 4D; intact vs R=22.3). Gl-Mstn transcript levels increased over tenfold in claw and fivefold in thoracic muscle between late premolt (R=22.3) and 2 days postmolt groups (Fig. 4D). A large decrease in Gl-Mstn transcript abundance occurred between 2 and 10 days postmolt for both muscles; a drop of 84% occurred in claw and a drop of 57% occurred in thoracic muscle (Fig. 4D).

Changes in the abundance of Gl-CalpT transcripts differed between claw and thoracic muscles in response to MLA. The transcript copy number for Gl-CalpT transcript did not change significantly in thoracic muscle during premolt (Fig. 4F). In claw muscle, Gl-CalpT mRNA increased significantly during premolt; copy number at late premolt (R=23.5) was 3.2-fold greater than that at early premolt (R=10; Fig. 4F). This is supported by a comparison of claw and thoracic muscle at specific R-index values; transcript levels for claw muscle were significantly less than that of thoracic muscle during early premolt and significantly greater during late premolt (Fig. 4F; R=10 vs R=23.5). However, because transcript abundance decreased by 60% in claw muscle during early premolt, the late premolt value was not significantly higher than that of the intact animals.

Abundance of Gl-Mstn mRNA was generally correlated with circulating ecdysteroids, but this relationship was much stronger in MLA animals than in ESA animals. Gl-Mstn transcript abundance was high at low ecdysteroid levels, and decreased significantly as ecdysteroid titers increased during premolt (Fig. 5). Gl-Mstn mRNA in claw muscle was negatively correlated with hemolymph ecdysteroids in animals induced to molt by MLA (Fig. 5A inset). A similar relationship was evident in ESA animals, but variability was much higher at ecdysteroid levels between 50 and 200 pg μl^–1 (Fig. 5B) and levels of Gl-Mstn mRNA were weakly correlated with hemolymph ecdysteroids (Fig. 5B inset). If intact animals were excluded from the analysis, the correlation coefficient for a linear fit to the log vs log data for ESA claw muscle increased to 0.720, whereas the fit for thoracic muscle did not change appreciably (data not shown). The negative correlation between hemolymph ecdysteroids and Gl-Mstn mRNA was weaker for thoracic muscle than claw muscle in both MLA and ESA experiments (Fig. 5A,B insets). Expression of Gl-EcR, Gl-RXR. Gl-CalpT, and Gl-EF-2 was not correlated with ecdysteroid concentration (data not shown).

Limb autotomy induced specific changes in gene expression in the corresponding thoracic muscle of intermolt animals. Unweighting had little or no effect on expression of Gl-EF-2, Gl-EcR and Gl-RXR in thoracic muscle (Fig. 6). However, transcript levels of both Gl-Mstn and Gl-CalpT were significantly altered in response to unweighting; Gl-Mstn mRNA increased threefold, whereas Gl-CalpT mRNA decreased by more than 40% (Fig. 6). Abundance of Gl-Mstn mRNA in unweighted thoracic muscle paralleled Gl-MstnmRNA abundance in weighted thoracic muscle during premolt and postmolt (Fig. 7).

The primary goal of the present study was to assess the role of a myostatin-like protein, Gl-Mstn, in regulating the atrophy of claw closer muscle during premolt and thoracic muscle after limb autotomy. The data show that (1) protein synthesis in claw muscle increases in two distinct stages during premolt; (2) transcript level for Gl-Mstn decreases dramatically in skeletal muscles during both premolt and postmolt; (3) transcript level for Gl-Mstn in atrophic claw muscle is negatively correlated with hemolymph ecdysteroids during premolt; (4) transcript level for Gl-Mstn in intermolt animals is significantly elevated in unweighted thoracic muscles; and (5) transcript levels for Gl-CalpT, Gl-EF-2, Gl-EcR, and Gl-RXR in skeletal muscle are not correlated with Gl-Mstn transcript abundance or hemolymph ecdysteroids. Together these data provide evidence supporting the hypothesis that Gl-Mstn mediates reversible atrophy of skeletal muscle in response to elevated ecdysteroids and unweighting, but through two distinct mechanisms.

Skeletal muscle mass is determined by the balance between protein synthetic and degradation rates. In general, muscle atrophy results from the net loss of protein caused by increased protein degradation, decreased protein synthesis, or a combination of the two (reviewed by Sandri, 2008; Tisdale, 2009). The large increase in protein synthesis that occurs in the claw during molt-induced atrophy is an exception to this rule, as protein degradation must increase to an even greater extent to effect the reduction in mass that occurs during the premolt period. This seems counter-productive, especially in consideration of the increased energetic costs associated with upregulated anabolic processes. In mammals, increased protein turnover occurs during remodeling of skeletal muscle undergoing fiber transformation and cardiac muscle undergoing rosiglitazone-induced hypertrophy (Festuccia et al., 2009; Termin and Pette, 1992). We suggest that increased protein turnover is necessary for the extensive remodeling of the sarcomere that is observed as myofibrils are reduced in size (Mykles, 1997). In the isomorphic claws of land crab, G. lateralis, and the major claw of fiddler crab, Uca pugnax, there is a preferential loss of about 11 thin myofilaments for every thick myofilament, which results in a decrease in the thin:thick myofilament ratio from ∼9:1 to ∼6:1; this corresponds to a 31% to 64% decrease in the ratio of actin to myosin heavy chain (Ismail and Mykles, 1992; Mykles and Skinner, 1981; Mykles and Skinner, 1982b). The loss of thin myofilaments increases the thick myofilament packing density between 51% and 72% (Ismail and Mykles, 1992; Mykles and Skinner, 1981). It is generally thought that exchange of myofibrillar proteins between sarcomeric and cytoplasmic pools occurs at the periphery of the sarcomere (Belcastro et al., 1991; Russell et al., 1992) and is the rate-limiting step in the incorporation of newly-synthesized protein into the contractile apparatus (Feng et al., 2009). Thus, a high protein turnover rate would facilitate rearrangement of the myofilaments within the myofibril.

Synthesis of soluble and myofibrillar proteins increased during proecdysis (premolt), confirming results from a previous study on G. lateralis by Skinner (Skinner, 1965). Skinner observed a 6.5-fold increase in the incorporation of [1-¹⁴C]leucine into total protein from skeletal muscle during early to mid premolt (pooled samples from stages D₀ to D₂) that was reversed in late premolt or early postmolt. The present study expands on this work by quantifying protein synthesis in the soluble and myofibrillar protein fractions, and provides a more detailed examination of protein synthesis during stages D₀ to D₂. The increase in protein synthesis that occurs in atrophic claw muscle peaks during mid to late premolt. During premolt, de novo synthesis of myofibrillar and soluble protein in the claw increases approximately 11- and 13-fold, respectively (Fig. 1A; ranging from R=20 to 23). Peak values indicate that the rate of protein synthesis may increase as much as 20-fold in both protein pools by late D₁ (Fig. 1A). Data presented in Fig. 1 and by Skinner (Skinner, 1965) indicate that protein synthesis reaches a maximum in claw muscle between the late D₁ and D₃ stages of premolt, and begins to decrease prior to ecdysis. This coincides with a peak in hemolymph ecdysteroids at stage D₂ (Skinner, 1985b). The decrease in protein synthesis during late premolt may be regulated by the large drop in circulating ecdysteroids at stage D₃ (Skinner, 1985b). Muscle atrophy, and associated remodeling of the contractile apparatus, is completed by this stage (Skinner, 1985b).

The increase in protein synthesis that occurs in atrophic claw muscle appears to involve two distinct phases of regulation. The relationship between protein synthesis in soluble and myofibrillar protein pools changes significantly in late D₀ of premolt (Fig. 1B). The relative increase in protein synthetic rate for the two protein pools is nearly equivalent in animals with an R-index of more than 13, but preferential incorporation of [³⁵S]methionine into soluble protein occurs in animals with an R-index less than 13 (1.4:1, soluble:myofibrillar ratio; Fig. 1B). Although there is a great deal of evidence indicating that proteolytic systems are upregulated in atrophic claw muscle (reviewed by Goll et al., 2003; Mykles, 1998) (see also Kramerova et al., 2005; Murphy et al., 2007), it is as yet unclear what proteins are preferentially synthesized during early premolt. Whatever the case, it is clear that the preferential incorporation into soluble protein early in premolt must involve many proteins, as there are no obvious changes in the pool of radioactively-labeled proteins during premolt (Fig. 2C).

Increased capacity for protein synthesis during premolt involves an increase in ribosomal RNA content. Levels of rRNA increase by 40% in epidermis, digestive gland and skeletal muscle in G. lateralis during premolt (Skinner, 1968). Since copy number was expressed as a function of total RNA, seemingly steady-state levels of an mRNA may actually reflect a nearly 40% increase in abundance during premolt. Given that only one out of five of the transcripts examined decreased during the period when rRNA content in muscle was increasing, it is reasonable to suggest that either a global increase in transcription or a decrease in RNA degradation occurs in skeletal muscle during premolt. Decreased incorporation of [³H]uridine into rRNA in epithelial tissue during premolt suggests that stability of ribosomes increases in this tissue during premolt (Skinner, 1966).

Ecdysteroids may stimulate protein degradation by activating calpains that degrade myofibrillar proteins (for reviews, see Mykles, 1998; Mykles, 1999), but this activation does not appear to involve transcriptional regulation. Levels of Gl-CalpM and Gl-CalpB mRNA do not change after ESA (Kim et al., 2005a), and, although Gl-CalpT mRNA levels increase threefold in atrophic claw muscle during early premolt, transcript levels initially decrease by 40% to 60% after molt induction (Fig. 3F, Fig. 4F). Thus, there is no significant difference between intermolt and premolt (late D₀ to D₂) levels of Gl-CalpT mRNA. These data differ somewhat from that of Kim et al. (Kim et al., 2005a), who reported a transient increase in Gl-CalpT mRNA in claw 1 day after ESA. The reason for this discrepancy is not clear. However, given the modest changes in Gl-CalpT mRNA levels observed during molt-induced atrophy of the claw (Fig. 3F, Fig. 4F), and the significant decrease in Gl-CalpT mRNA observed during atrophy of thoracic muscle induced by limb autotomy (Fig. 6), we conclude that Gl-CalpT activity is not regulated at the transcriptional level in atrophic crustacean muscle.

Ecdysteroids appear to control Gl-Mstn expression in skeletal muscle over much of the molting cycle. Gl-Mstn mRNA levels decrease by at least 68% in thoracic muscle and 81% in claw muscle during premolt (Fig. 3D, Fig. 4D). Gl-Mstn transcript abundance in claw muscle is also negatively correlated with ecdysteroid levels over the same period (Fig. 5). The relatively low correlation between hemolymph ecdysteroids and Gl-Mstn expression in thoracic muscle may result, in part, from greater suppression at low ecdysteroid levels indicative of intermolt animals; Gl-Mstn expression in intermolt animals is 50–60% greater in claw muscle than in thoracic muscle. At 2 days postmolt, when ecdysteroid titers are low, Gl-Mstn mRNA levels increase in thoracic and claw muscle by 6- and 10-fold, respectively (Fig. 4D). In view of these data, it seems reasonable to suggest that Gl-Mstn is negatively regulated by ecdysteroid during premolt and early postmolt. However, the regulation of Gl-Mstn expression also appears to involve ecdysteroid-independent signals. Despite the presence of unchanging and extremely low ecdysteroid levels following ecdysis, Gl-Mstn mRNA in thoracic and claw muscles decreases by 57% and 84%, respectively, between 2 days and 10 days postmolt (Fig. 4D). A direct link between ecdysteroid signaling and expression of Gl-Mstn remains to be established.

Autotomy-induced muscle atrophy differs from molt-induced atrophy both in its regulatory origin and transcriptional profile. In response to the loss of a limb, the thoracic muscle atrophies rapidly and more than half of the mass is lost; it remains in an atrophied state until a new limb is regenerated and becomes fully functional after molting (Moffett, 1987). However, unlike atrophy of the claw during premolt, atrophy of thoracic muscle in response to limb autotomy does not appear to be regulated by ecdysteroids, as atrophy occurs during intermolt when ecdysteroid levels are low (Fig. 3A, Fig. 4A). The intact innervation and muscle structure (Moffett, 1987) suggests that the mechanism may be a response to physical unweighting. What is clear is that the regulation of Gl-Mstn expression differs between molt-induced atrophy and autotomy-induced atrophy of the thoracic muscle. Although Gl-Mstn mRNA decreases as much as 82% in weighted thoracic muscle and 94% in claw muscle during premolt, levels increase by 300% in unweighted thoracic muscles (Fig. 6). Atrophy in unweighted thoracic muscle may be very similar to some forms of disuse atrophy observed in mammals (reviewed by Favier et al., 2008). An interesting question is whether elevated levels of Gl-Mstn transcript in unweighted thoracic muscle are correlated with a decrease in global protein synthesis.

The regulation of Mstn expression in crustacean muscles induced by ESA and MLA differs from that in mammalian muscles induced to atrophy by various treatments or conditions. In mammals, Mstn mRNA levels are elevated in muscle atrophies induced by diseases (Carneiro et al., 2008; Costelli et al., 2008), denervation (Liu et al., 2007; Shao et al., 2007), unloading (Allen et al., 2009; Reardon et al., 2001; Wehling et al., 2000), glucocorticoid (Ma et al., 2003) and aging (Leger et al., 2008). Glucocorticoids stimulate Mstn transcription via response elements in the promoter region (Ma et al., 2001) (for reviews, see Schakman et al., 2008; Tisdale, 2009). Glucocorticoids also stimulate protein degradation and inhibit protein synthesis in mammalian skeletal muscles via Mstn signaling (for reviews, see Schakman et al., 2008; Tisdale, 2009). Moreover, atrophy induced by dexamethasone is prevented by deletion of the Mstn gene in knockout mice (Gilson et al., 2007). By contrast, Mstn expression in crustacean muscles is inversely correlated with ecdysteroid levels (Fig. 5), and Gl-Mstn mRNA is reduced during molt-induced atrophy (Fig. 3D, Fig. 4D).

Although Mstn expression in mammalian and crustacean muscles responds differently to steroid hormones, the data presented here support the hypothesis that Mstn inhibits protein synthesis in both groups of organisms through the target of rapamycin complex 1 (TORC1). TORC1, a component of the insulin/insulin-like growth factor pathway controlling growth in eukaryotic cells and is a serine/threonine protein kinase; it regulates protein synthesis at the translational level through the phosphorylation of S6 kinase and 4EF-binding protein-1 (for reviews, see Proud, 2007; Wullschleger et al., 2006). This same pathway mediates nutrient-dependent growth in insects (for a review, see Mirth and Riddiford, 2007). There is growing recognition that TORC1 is central to the effects of growth factors, nutrients, contraction and aging on mammalian muscle mass (reviewed by Drummond et al., 2009; Miyazaki and Esser, 2009; Sandri, 2008). Recent studies indicate that Mstn is a negative regulator of this pathway in mammalian muscle. Mstn inhibits signaling via TORC1 in rat skeletal muscle (Amirouche et al., 2009) and protein synthesis in C₂C₁₂ muscle cells (Taylor et al., 2001). Furthermore, Mstn reduces phosphorylation of Akt (McFarlane et al., 2006), which could lead to inactivation of TORC1 (Proud, 2007). Conversely, Mstn inhibition stimulates protein synthesis via the TORC1 pathway in mouse skeletal muscle (Welle et al., 2009). In decapod crustaceans, elevated ecdysteroids have only a small effect on gene expression of myofibrillar protein, suggesting that transcriptional regulation has a relatively minor contribution to changes in protein levels occurring over the molting cycle (El Haj, 1999; Medler et al., 2005; Whiteley and El Haj, 1997; Whiteley et al., 1992). However, increased protein synthesis is correlated with increased ribosomal RNA and activity in muscles from premolt animals (El Haj et al., 1996; Skinner, 1968). The downregulation of Gl-Mstn is associated with an increase in global synthesis of both soluble and myofibrillar proteins (Figs 1, 2). Taken together, these data provide evidence to suggest that Gl-Mstn, either directly or indirectly, suppresses protein synthesis. Further studies are needed to establish whether Gl-Mstn inhibits translation via TORC1.

For future studies using molt induction, it is relevant to note that an examination of ecdysteroidogenesis and gene expression in skeletal muscle demonstrates that ESA and MLA stimulate similar, but not identical, responses. Both ESA and MLA cause precocious molting, but the two differ in the timing of hormonal cues and premolt processes. ESA triggers an acute response by the Y-organ, resulting in elevated hemolymph ecdysteroid levels within 1 day of ESA and a shortened premolt period (Skinner and Graham, 1972). By contrast, intermolt animals induced by MLA enter premolt 2–3 weeks after the procedure (Holland and Skinner, 1976), and ecdysteroids increase more gradually. Expression of Gl-EcR follows a similar pattern to that reported by Kim et al. (Kim et al., 2005a) during the first 3 days after ESA, wherein a transient increase is observed in claw muscle and a transient decrease is observed in thoracic muscle. Expression of Gl-EcR converges somewhat between these tissues over a period of 2 weeks. By contrast, a consistent increase in Gl-EcR mRNA is observed in both claw and thoracic muscle during premolt induced by MLA. In addition, levels of Gl-RXR mRNA remain unchanged throughout the molt cycle induced by MLA, but decrease transiently during stage D₀ in ESA animals. These data confirm that skeletal muscle is responsive to changes in ecdysteroids over a molting cycle, and that the rate of change in ecdysteroid titers affects the expression of responsive genes, including those for ecdysteroid receptor subunits.

In summary, Gl-Mstn is differentially regulated in two distinct atrophic tissues: premolt claw muscle and unweighted thoracic muscle. The reduction in Gl-Mstn transcript in premolt claw and thoracic muscle is correlated with increasing ecdysteroid levels. These data provide support for the hypothesis that ecdysteroids repress Gl-Mstn expression during premolt when bound to the EcR–RXR heterodimer. This is contrary to the prevailing view, which holds that heterodimeric hormone receptors, in the absence of hormone, interact with co-repressors to prevent transcription; ligand binding induces a conformational change that causes dissociation of the corepressors (reviewed by Gurevich et al., 2007). The most parsimonious explanation for these data is that the binding of ecdysteroid allows the receptor to interact with a co-repressor and block transcription of Gl-Mstn. To our knowledge gene repression by ecdysteroid binding to nuclear receptors has not been reported in arthropods. Ecdysteroid and ESA downregulate gene expression in crayfish digestive gland (Shechter et al., 2007), but it is not known whether the effects are direct or indirect. However, a number of co-repressors that require ligand binding to nuclear receptors have been reported in mammals (reviewed by Gurevich et al., 2007). Interestingly, low ecdysteroid levels in intermolt animals permit the upregulation of Gl-Mstn in unweighted thoracic muscle, but this upregulation is overridden by elevated ecdysteroids during premolt (Fig. 7). We conclude that ecdysteroids repress Gl-Mstn expression in both claw and thoracic muscles. The data presented here support the hypothesis that Gl-Mstn functions as an inhibitor of global protein synthesis in crustacean skeletal muscle.

 
• Calp

  calpain (Ca²⁺-dependent protease)

 
• EcR

  ecdysone receptor

 
• EF-2

  eukaryotic elongation factor 2

 
• ESA

  eyestalk ablation

 
• Gl

  Gecarcinus lateralis

 
• MLA

  multiple limb autotomy

 
• Mstn

  myostatin

 
• qRT-PCR

  real-time reverse transcription-polymerase chain reaction

 
• R-index

  regeneration index

 
• RXR

  retinoid X receptor