Rheb, an activator of target of rapamycin, in the blackback land crab, Gecarcinus lateralis: cloning and effects of molting and unweighting on expression in skeletal muscle

Skeletal muscle is a plastic tissue that responds to a variety of physiological conditions. In mammals, nutrition, steroid hormones, innervation, stretch, exercise and disease alter fiber size (reviewed by Cardozo et al., 2010; Drummond et al., 2009a; Favier et al., 2008; Matsakas and Patel, 2009b; Zhang et al., 2007). Mammalian muscle atrophies in response to aging, denervation, glucocorticoids, unloading, sepsis and cancer cachexia (reviewed by Frost and Lang, 2011; Furlow et al., 2010; Glass, 2010; Schakman et al., 2009; Urso, 2009). In decapod crustaceans, two types of muscle atrophy are documented: (1) a molt-induced claw muscle atrophy in response to molting hormone (ecdysteroids), which allows easier passage of the claw through the small basi-ischial joint when the old exoskeleton is shed at ecdysis (reviewed by Mykles, 1997; Mykles, 1999; Mykles and Skinner, 1982a); and (2) an unweighting atrophy of the thoracic muscles in response to autotomy, or voluntary loss, of the associated limb (Griffis et al., 2001; Moffett, 1987; Schmiege et al., 1992). Unlike claw muscle atrophy, thoracic muscle atrophy can occur during the intermolt period when hemolymph ecdysteroid levels are low (Mykles, 2011). Claw muscle atrophy coincides with an extensive remodeling of the contractile apparatus, resulting from the preferential loss of thin filaments (Ismail and Mykles, 1992; Mykles and Skinner, 1981; Mykles and Skinner, 1982b). It is thought that the remaining thick filaments provide a framework for the incorporation of thin filaments into the sarcomere during the post-molt period (Mykles and Skinner, 1981). An accelerated protein turnover, resulting from an increase in protein synthesis, occurs during premolt and is hypothesized to facilitate the exchange of myofilaments between soluble and myofibrillar protein pools required for remodeling (Mykles, 1997).

Myostatin (Mstn) is an autocrine/paracrine factor that inhibits muscle growth in vertebrates (reviewed by Glass, 2010; McCarthy and Esser, 2010; Rodgers and Garikipati, 2008; Stinckens et al., 2011). It acts by suppressing protein synthesis at the same time as it increases protein degradation, resulting in a reduction of skeletal muscle mass (reviewed by Matsakas and Patel, 2009a; McCarthy and Esser, 2010; Tisdale, 2009; Zhang et al., 2007). Mstn may also regulate muscle growth in invertebrates. cDNAs encoding Mstn-like proteins have been characterized in mollusks and arthropods (Covi et al., 2008; De Santis et al., 2011; Hu et al., 2010; Kim et al., 2004; Kim et al., 2009; Kim et al., 2010; Lo and Frasch, 1999; MacLea et al., 2010). In blackback land crabs (Gecarcinus lateralis), molt-induced atrophy and unweighting atrophy have opposite effects on Gl-Mstn expression in claw closer and thoracic muscles, respectively. Unweighting increases Gl-Mstn mRNA levels 3-fold in thoracic muscles in intermolt animals (Covi et al., 2010). By contrast, Gl-Mstn mRNA levels in claw muscle decrease ∼17-fold as protein synthesis of myofibrillar and soluble proteins increase 11-fold and 13-fold, respectively, during premolt (Covi et al., 2010). Gl-Mstn mRNA level in weighted thoracic muscle decreases only ∼5-fold (Covi et al., 2010). In the American lobster, Homarus americanus, Ha-Mstn expression in claw closer muscles over the course of the molt cycle varies in very similar ways to that observed in G. lateralis, underscoring the conservation of the transcriptional regulatory mechanism among decapod crustaceans (MacLea et al., 2010).

In mammals, Mstn represses protein synthesis by inhibiting the metazoan target of rapamycin (mTOR) (reviewed by Drummond et al., 2009a; Glass, 2010; Matsakas and Patel, 2009a; McCarthy and Esser, 2010; Otto and Patel, 2010). mTOR is a serine/theonine protein kinase that associates with Rheb (Ras homolog enriched in brain), Raptor and GβL to form the mTOR Complex 1 (mTORC1) (reviewed by Proud, 2009; Wullschleger et al., 2006). mTORC1 is a downstream target of the insulin/PI₃ kinase/Akt signaling pathway and stimulates protein synthesis at the translational level by phosphorylating ribosomal S6 kinase (s6k) and 4E-binding protein 1 (Proud, 2009; Wullschleger et al., 2006). Binding of GTP to Rheb activates mTORC1 while hydrolysis of GTP to GDP, which is stimulated by Rheb-GAP [tuberous sclerosis complex 1 and 2 (TSC1/2)], inactivates mTORC1 (Aspuria and Tamanoi, 2004; Wullschleger et al., 2006). Akt inactivates Rheb-GAP, thus maintaining mTORC1 in the activated state by inhibiting GTP hydrolysis. Mstn signaling inhibits mTOR via Akt in mammalian skeletal muscles (Amirouche et al., 2009; Trendelenburg et al., 2009). Mstn downregulation stimulates protein synthesis and Akt and s6k phosphorylation (Lipina et al., 2010; Morissette et al., 2009; Welle et al., 2009; Welle et al., 2011). These data indicate that there is cross-talk between Mstn and mTOR signaling pathways in mammalian skeletal muscles.

Ecdysteroid regulation of protein synthesis in crustacean muscle is primarily at the translational level. Elevated ecdysteroid levels have little effect on actin and myosin heavy chain mRNA levels (El Haj, 1999; Medler et al., 2005; Whiteley and El Haj, 1997). Increased protein synthesis during premolt is correlated with increased ribosomal activity (El Haj et al., 1996; Skinner, 1965; Skinner, 1968). Autoradiographic analysis of soluble and myofibrillar proteins shows that the large increase in protein synthesis in atrophic claw muscle is consistent with enhanced translation of existing mRNA into protein (Covi et al., 2010). Crustacean Mstn, like mammalian Mstn, may suppress translation by inhibiting insulin/mTOR signaling. A potential target is Rheb, a member of the Ras superfamily of GTP-binding proteins (reviewed by Reuther and Der, 2000). Rheb is a crucial upstream activator of mTOR (Aspuria and Tamanoi, 2004; Frost and Lang, 2011; Li et al., 2004; Manning and Cantley, 2003). Overexpression of Rheb stimulates cell growth while knockdown of Rheb expression inhibits protein synthesis and cell growth in insects (Hall et al., 2007; Patel et al., 2003) [see additional references in Aspuria and Tamanoi (Aspuria and Tamanoi, 2004)].

The purpose of this study was to determine the effects of molt induction and unweighting on the expression of insulin/mTOR signaling components in G. lateralis claw closer and thoracic muscles, as well as to determine if the expression of Gl-Mstn is coupled to the expression of Gl-Rheb or other components of the insulin/mTOR signaling pathway. Our knowledge of the insulin/mTOR pathway in crustaceans and its role in development and growth is fragmentary at best. Insulin-like peptides have been isolated from crustacean tissues (Gallardo et al., 2003; Hatt et al., 1997) and the androgenic gland expresses an insulin-like growth factor that determines male development (Banzai et al., 2011; Manor et al., 2007; Rosen et al., 2010; Sroyraya et al., 2010; Ventura et al., 2009). Insulin receptor tyrosine kinases and phosphotyrosyl phosphatases are present in crustacean tissues (Chuang and Wang, 1993; Chuang and Wang, 1994; Kucharski et al., 1999; Kucharski et al., 2002; Lin et al., 1993). In the present study, we report the characterization of cDNAs encoding Rheb, Akt, mTOR and s6k from G. lateralis and Rheb from the green shore crab (Carcinus maenas) and H. americanus. Quantitative polymerase chain reaction (qPCR) was used to quantify the expression of Gl-Rheb, Gl-mTOR, Gl-Akt and Gl-s6k in claw closer, weighted thoracic and unweighted thoracic muscles. Molting was induced by eyestalk ablation (ESA) and multiple leg autotomy (MLA). cDNA from a previous study (Covi et al., 2010) was used to directly compare the expression of mTOR signaling components with Gl-Mstn expression and hemolymph ecdysteroid level.

Adult male blackback land crabs (Gecarcinus lateralis Fréminville 1835) were shipped to Colorado, USA, from the Dominican Republic and maintained as described (Covi et al., 2010). Molting was induced by either MLA or ESA (for reviews, see Mykles, 2001; Skinner, 1985). Progression of animals through the premolt period for both methods was monitored by measuring the length of a limb regenerate at the position of the third walking leg, called the regeneration (R) index (R=length of regenerate×100/carapace width) (Holland and Skinner, 1976; Yu et al., 2002). For MLA animals, all eight walking legs were autotomized. For ESA animals, one walking leg was autotomized and a basal regenerate was allowed to form prior to ESA. One claw was also autotomized from intact and MLA animals to determine the effects of unweighting; weighted muscles were harvested from the thoracic segment with a claw, and unweighted muscles were harvested from the thoracic segment without a claw from the same animal.

Intermolt adult green shore crab, Carcinus maenas Linnaeus 1758, and American lobster, Homarus americanus H. Milne-Edwards 1837, were obtained from Bodega Marine Laboratory, Bodega Bay, CA, USA (MacLea et al., 2010; McDonald et al., 2011).

Reverse transcriptase-polymerase chain reaction (RT-PCR) and rapid amplification of cDNA ends (RACE) were used to clone the cDNAs encoding Gl-Rheb, Gl-mTOR, Gl-Akt and Gl-s6k. RNA isolation from G. lateralis skeletal muscles and cDNA synthesis were the same as described in Covi et al. (Covi et al., 2010). PCR used 0.5 μl of the first-strand cDNA as a template and forward and reverse primers (10 pmol each; Table 1) synthesized by Integrated DNA Technology (IDT; Coralville, IA, USA). For the cloning of Gl-Rheb, specific primers directed against the Cm-Rheb expressed sequence tag (EST) sequence (GenBank #HM989970) were used (see below). For all other genes degenerate primers were designed using iCODEHOP (Boyce et al., 2009) or by hand using a multiple sequence alignment of various homologous proteins and a codon chart to target areas of high conservation in protein sequences from other arthropod sequences.

Cm-Rheb (GenBank #HM989970) and Ha-Rheb (#HM989972) were cloned from C. maenas or H. americanus cDNA, respectively, on the basis of sequences determined by multiple sequence alignment of several EST clones provided by the Mount Desert Island Biological Laboratory (Towle and Smith, 2006). RNA isolation and cDNA synthesis were the same as described for G. lateralis above. The sequence for each open reading frame (ORF) was verified by RT-PCR using cDNA prepared as described below.

All initial cloning reactions for the above genes were PCR amplified using GoTaq Green master mix (Promega, Madison, WI, USA). After denaturing the cDNA at 96°C for 3 min, 35 cycles of PCR were completed with the following programme: 96°C for 30 s, the lowest annealing temperature of a primer pair (see Table 1) for 30 s, and 72°C for 30 s to 1 min. Final extension was for 7 min at 72°C. Amplified products, verified as single bands by 1% agarose gel electrophoresis, were purified using the GeneJet PCR Cloning kit (Fermentas, Glen Burnie, MD, USA), ligated into the pJET1.2 vector using the CloneJet PCR Cloning kit (Fermentas) and, after insert verification by PCR with vector primers, sequenced using a T7 primer (Davis Sequencing, Davis, CA, USA).

To acquire the 5′ and 3′ sequences for Gl-Rheb, the FirstChoice RLM-RACE kit (Applied Biosystems, Austin, TX, USA) was used according to the manufacturer’s instructions. RACE conditions were as follows: 0.4 μl of RACE template cDNA was used in each reaction, with 8 pmol of each gene specific (Table 1) and kit primer, and other components identical to the initial PCR reactions (see above). After denaturation at 94°C for 3 min, 35 cycles of 94°C for 30 s, the lowest annealing temperature of a primer pair (Table 1) for 30 s, and 72°C for 30 s to 1 min, were completed. Final extension was for 7 min at 72°C. All products were cloned and sequenced as described above. The cloned coding sequences were deposited in GenBank (see Table 2 for accession numbers).

Endpoint PCR was used to qualitatively assess the tissue distribution of Gl-Rheb (no. HM989971), Gl-EF2 (no. AY552550), Cm-Rheb (no. HM989970), Cm-EF2 (no. GU808334), Ha-Rheb (no. HM989972) and Ha-EF2 (no. FJ790217) transcripts. For G. lateralis, endpoint PCR used the same samples used previously to analyze Gl-Mstn and Gl-EF2 tissue expression in Covi et al. (Covi et al., 2008). For C. maenas, RNA was isolated from various tissues as described (Covi et al., 2010). Tissues were harvested from one intermolt adult male, except for Y-organ, eyestalk ganglia and thoracic ganglia, which were pooled from three intermolt males. For H. americanus, the analysis used the same samples from an intermolt female adult lobster from a previous study (MacLea et al., 2010). Reactions contained 1 μl of template cDNA and 5 pmol each of the appropriate expression primers (Table 1) in GoTaq Green master mix (Promega). After denaturation at 94°C for 3 min, 28–35 cycles of 94°C for 30 s, the lowest annealing temperature of a primer pair (Table 1) for 30 s, and 72°C for 30 s, were completed. Final extension was for 7 min at 72°C. Ha-Rheb used 29 and 35 cycles, Ha-EF2 and Cm-EF2 used 28 cycles, Cm-Rheb used 29 cycles, Gl-EF2 used 30 cycles and Gl-Rheb used 35 cycles. After PCR was terminated, products were separated on a 1% agarose gel containing TAE (40 mmol l^–1 Tris acetate and 2 mmol l^–1 EDTA, pH 8.5). The gels were stained with ethidium bromide and visualized with an ultraviolet (UV) light source. All PCR products were sequence verified.

Samples and methods for RNA isolation and cDNA synthesis were the same as described in Covi et al. (Covi et al., 2010). A LightCycler 480 thermal cycler (Roche Applied Science, Indianapolis, IN, USA) was used to quantify the levels of Gl-Rheb (HM989971), Gl-mTOR (HM989973), Gl-Akt (HM989974) and Gl-s6k (HM989975) mRNAs in claw closer and thoracic (weighted and unweighted) muscles. Reactions consisted of 1 μl of first-strand cDNA or standard, 5 μl of 2× SYBR Green I Master mix (Roche Applied Science), 5 pmol each of forward and reverse primers (Table 1) and 3 μl of nuclease-free water. PCR conditions were as follows: an initial denaturation at 95°C for 5 min, followed by 45 cycles of denaturation at 95°C for 10 s, annealing at 62°C for 20 s and extensions at 72°C for 20 s, followed by melting curve analysis of the PCR product. Transcript concentrations were determined by use of the LightCycler 480 software (Roche Applied Science, version 1.5) using a series of double-stranded DNA (dsDNA) gene standards produced by serial dilutions of PCR product for each gene (10 ag μl^–1 to 10 ng μl^–1). The absolute amounts of transcript in copy numbers per μg of total RNA in the cDNA synthesis reaction were calculated based on the standard curve and the calculated molecular weight of dsDNA products. All fold changes in expression indicated in the text refer to changes in the raw expression levels of the genes.

Statistical analysis was performed using JMP 5.1.2, 6.0.0 or 8.0.2 (SAS Institute, Cary, NC, USA). Group variances were analyzed using a Brown–Forsythe test and found to be equal (P<0.05). Means for different developmental stages were compared using analysis of variance (ANOVA). All data not plotted as individual points are represented as means ± 1 s.e.m. and the level of significance was set at α=0.05. Paired t-tests were used for comparisons of the weighted and unweighted thoracic muscle mean comparisons, with the same α-level for significance. Spearman’s rank correlation analysis and linear regression analysis of log-transformed data were performed using Excel 2007 (Microsoft Corporation, Redmond, WA, USA) and JMP (SAS Institute). The slopes of each regression line were compared with y=0 by ANOVA and P-values are reported in the relevant figures and text, along with the Spearman correlation values (ρ) and their corresponding P-values. Multiple sequence alignments were produced with ClustalX version 2.0.12 (Thompson et al., 1997) using deduced amino acid sequences. Grapher 8.2.460 (Golden Software, Golden, CO, USA), Excel and Illustrator 10 (Adobe Systems, San Jose, CA, USA) were used for constructing/annotating graphs and figures. In silico analysis of predicted RNA characteristics was completed using RNAz (http://rna.tbi.univie.ac.at/RNAz) (Gruber et al., 2007), RNA analyzer (http://wb2x01.biozentrum.uni-wuerzburg.de/) (Bengert and Dandekar, 2003) and UTRscan (http://itbtools.ba.itb.cnr.it/utrscan) (Mignone et al., 2005). Estimated protein molecular weights were calculated using the Compute pI/MW tool (http://expasy.ch/tools/pi_tool.html).

BLAST searches identified clones of the Rheb gene among EST libraries for C. maenas (Cm-Rheb) and H. americanus (Ha-Rheb) (Table 2). A BLAST search also identified a Rheb ortholog from a suppression subtractive hybridization cDNA library of the green mud crab, Scylla paramamosain (Sp-Rheb, ACY66406; incorrectly annotated as Ras). The Gl-Rheb cDNA (969 bp) contained the 5′ untranslated region (UTR), complete ORF and an incomplete 3′ UTR, when compared with the 3′ UTRs in the Ha-Rheb and Cm-Rheb cDNA sequences (Fig. 1, Fig. 4, Table 2). All four decapod Rheb cDNAs encoded 182-amino acid proteins with estimated masses of 20.2 kDa for Gl-Rheb and 20.4 kDa for Cm-Rheb, Sp-Rheb and Ha-Rheb.

The deduced amino acid sequences of the four crustacean Rheb proteins were highly conserved, showing high identity and similarity to each other and to Rheb protein sequences in insects and human (Fig. 2). The decapod Rheb sequences were aligned with other arthropod sequences (human body louse and red flour beetle), as well as the human Rheb amino acid sequence. A different paralog within the Ras gene family (human K-ras) was included as an ‘outgroup’ for comparison purposes. Sequence identity was particularly high within the ‘G box’ motifs (G1–G5) in all the Rheb proteins. The C-terminal lipid modification site was also conserved in the arthropod Rhebs. There was less identity/similarity between the Rheb and K-ras sequences, particularly in the G1 and G2 motifs (Fig. 2). The segregation of Rheb and Ras into two distinct gene families was confirmed by phylogenetic analysis (Fig. 3). Bootstrapping values indicated a high degree of relatedness among the four decapod Rheb sequences and among the Ras-like sequences from human and insects. Bootstrapping values were lower between the crustacean and insect Rheb sequences. The multiple sequence alignment between translated crustacean, insect and human Rheb protein sequences (Fig. 2) showed significant sequence divergence outside of the G box motifs, in particular the regions between the G2 and G3 boxes and the regions flanking the G5 box.

Multiple alignment of the nucleotide sequences identified highly conserved regions in the 3′ UTRs of the four decapod Rheb cDNAs (Fig. 4). Putative functional elements in the 3′ UTR were identified using RNA analysis software (see Materials and methods). A poly A signal sequence (AATAAA) was identified in Cm-Rheb and Ha-Rheb, indicating that these cDNAs contained a complete 3′ UTR. The nucleotide sequences of the Cm-Rheb and Ha-Rheb were 47% identical, with the greatest amount of identity in the first ∼275 nucleotides of the 3′ UTR. A K-box sequence (TGTGAT) (Lai et al., 1998) was located six nucleotides after the stop codon in all four decapod Rheb cDNAs (Fig. 4). Among the other arthropod species for which complete Rheb sequences are available (e.g. copepod, sand fly, silkworm and mosquito), however, no K-box sequence was found in the 3′ UTR. Among the mammalian sequences in GenBank, the K-box is generally absent. Only human (NM005614) and cattle (NM001031764) Rheb sequences contain a K-box, which overlaps with the stop codon. A ∼100-nucleotide region (nucleotides #161 to #264 in the Cm-Rheb 3′ UTR) was predicted to form a stem-looped secondary structure consistent with mRNAs (see Materials and methods).

Endpoint RT-PCR showed that Rheb and EF2 were ubiquitously expressed (Fig. 5). Cm-Rheb and Cm-EF2 transcripts were present in skeletal muscle, gill, hindgut, hepatopancreas, heart, midgut, testis, thoracic ganglion and Y-organ (Fig. 5A). Gl-Rheb and Gl-EF2 transcripts were present in skeletal muscle, hindgut, heart, eyestalk ganglion, thoracic ganglion and Y-organ (Fig. 5B). Ha-Rheb and Ha-EF2 were present in antennal gland, skeletal muscle, hindgut, hepatopancreas, heart, midgut, ovary and thoracic ganglion (Fig. 5C).

cDNAs encoding additional components of the mTOR signaling pathway were cloned from G. lateralis tissues using RT-PCR. Partial sequences for Gl-Akt, Gl-mTOR and Gl-s6k were obtained (Table 2). The deduced sequences shared high degrees of identity and similarity to the human ortholog of each cDNA at the protein level (60–74% identity/77–84% similarity; Table 2), with even higher identity and similarity to orthologs in insects (data not shown). The Gl-mTOR partial cDNA encoded the kinase domain, which is located near the 3′ end of the ORF in the human mTOR (reviewed by Abraham and Wiederrecht, 1996). Gl-Akt encoded the kinase and pleckstrin domains, and Gl-s6k encoded both kinase domains and the substrate-binding region (Table 2).

qPCR was used to analyze levels of land crab mTOR signaling components in claw and thoracic muscles as animals progressed through the molt cycle. The same cDNA samples from Covi et al. (Covi et al., 2010) were used in the analysis. ‘Weighted’ thoracic muscle served as an internal control, as this muscle does not atrophy during premolt. The data were graphed as a function of three variables: time, R-index and molt stage. Events post-ESA or post-ecdysis were measured in time after the treatment or event. The R-index was used to compare the two methods (ESA and MLA) of molt induction. Finally, a molt stage classification, as applied to G. lateralis by Skinner (Skinner, 1962), was used to integrate the data from time and R-index. Molt induction increased hemolymph ecdysteroid titers; the data from Covi et al. (Covi et al., 2010) are reproduced in Fig. 6E and Fig. 7E to compare gene expression with hemolymph ecdysteroid titer.

The expression of Gl-Akt, Gl-Rheb, Gl-mTOR and Gl-s6k each responded differently to molt induction by ESA (Fig. 6A–D). The expression of a ‘housekeeping’ gene, Gl-EF2, is for the most part expressed at constant levels, except at 20 days post-ESA, when levels increase 2.1-fold in claw muscle but not in thoracic muscle (Covi et al., 2010). Gl-Akt showed a similar pattern of expression in both muscle types (Fig. 6A). Gl-Akt mRNA level increased 1.8-fold in claw muscle and 3.5-fold in thoracic muscle by 7 days post-ESA (R=10.0). From these peak levels, Gl-Akt decreased 2.6-fold in claw muscle and 2.4-fold in thoracic muscle by 20 days post-ESA (R=19.9).

ESA had a differential effect on the expression of Gl-Rheb in claw and thoracic muscles (Fig. 6B). A lower level of Gl-Rheb in thoracic muscle from intact animals contributed to an immediate 3.4-fold increase 1 day post-ESA (R=6.6). However, Gl-Rheb expression in thoracic muscle was relatively constant during premolt. By contrast, Gl-Rheb mRNA in claw muscle increased 3.9-fold over the level in intact animals at 7 days post-ESA (R=10.0) and remained elevated at 14 days (R=15.5) and 20 days (R=19.9) post-ESA. Gl-Rheb transcript levels in claw muscle were significantly higher than those in thoracic muscle at every molt stage, except 1 and 3 days post-ESA (Fig. 6B, asterisks).

ESA increased Gl-mTOR mRNA levels but the timing of the peaks differed between the two muscle types (Fig. 6C). In claw muscle, a maximum level occurred at 7 days (R=10.0) post-ESA, which was 2.8-fold higher than the level in intact animals; mRNA levels remained elevated at 14 days and 20 days post-ESA (R=15.5 and 19.9, respectively). In thoracic muscle, a maximum occurred at 14 days post-ESA, which was 2.4-fold higher than intact animals. However, unlike claw muscle, thoracic muscle Gl-mTOR mRNA levels decreased 2.5-fold after the 14 days post-ESA peak, returning to a level comparable to that from the intact animals. At two points in the molt cycle (R=10.0 and 19.9), Gl-mTOR mRNA levels were significantly lower in thoracic muscle than in claw muscle (Fig. 6C, asterisks).

ESA had small effects on the expression of Gl-s6k in both muscle types (Fig. 6D). In claw muscle, there were no changes in Gl-s6k transcript levels, except for a 1.9-fold increase over intact levels at 7 days post-ESA. In thoracic muscle, Gl-s6k expression in intact animals was lower than that in claw muscle but was expressed at similar levels to claw muscle in premolt animals, except at 7 days post-ESA. Thoracic muscle Gl-s6k expression was significantly higher in ESA animals than in intact animals. There was little change in the expression of Gl-s6k in thoracic muscle during premolt, except for a small increase and subsequent decrease 10–20 days post-ESA.

The expression of Gl-Akt, Gl-Rheb, Gl-mTOR and Gl-s6k each responded differently to molt induction by MLA (Fig. 7A–D). Animals entered premolt, as measured by the R-index, and successfully molted. Thus, we were able to compare gene expression in muscles from premolt and post-molt animals. Gl-EF2 mRNA levels showed no significant changes in the same samples (Covi et al., 2010). Gl-Akt was expressed at constant levels in thoracic muscle, except at mid-premolt (R=15.0), where a significant 1.9-fold increase over intact animals was observed (Fig. 7A). By contrast, Gl-Akt was expressed in claw muscle at the same levels in intact (intermolt), early premolt (R=10.0) and mid-premolt (R=15.0) stages and then decreased 2.8-fold by 10 days post-molt (Fig. 7A). Gl-Akt mRNA levels in thoracic muscle were significantly lower than those in claw muscle from intact and premolt animals (Fig. 7A, asterisks); the means were not significantly different at 2 days and 10 days post-molt.

MLA had differential effects on Gl-Rheb expression in claw and thoracic muscles. In claw muscle, Gl-Rheb mRNA levels increased during premolt, reaching a level 3.4-fold higher in late premolt animals (R=23.5) than that in intermolt (intact) animals (Fig. 7B). Following ecdysis, Gl-Rheb mRNA levels decreased and returned to intermolt levels by 10 days post-molt. By contrast, thoracic muscle showed a small increase in Gl-Rheb expression during premolt; Gl-Rheb mRNA levels in late premolt animals (R=23.5) were 1.7-fold higher than the level in intact animals (Fig. 7B). Gl-Rheb mRNA levels in thoracic muscles from post-molt animals were ∼2-fold higher than intact animals. Gl-Rheb expression in thoracic muscle was significantly lower than claw muscle at all molt stages (Fig. 7B, asterisks).

MLA had transient effects on the expression of Gl-mTOR in both muscle types. In claw muscle, Gl-mTOR mRNA levels were constant during premolt and then decreased 3.1-fold from intact levels by 10 days post-molt (Fig. 7C). In thoracic muscle, expression of Gl-mTOR decreased 2-fold in mid-premolt animals compared with intact animals (Fig. 7C; R=15.0). However, thoracic muscle levels returned to the initial level by late premolt (R=23.5) and decreased 1.7-fold in post-molt animals. There was no significant difference in the Gl-mTOR mRNA levels between claw and thoracic muscles at each molt stage, except at 10 days post-molt, in which the level in the claw was significantly lower than that in the thoracic muscle (Fig. 7C, asterisk).

The general pattern of Gl-s6k expression was similar between claw and thoracic muscles over the molt cycle. MLA had little effect on Gl-s6k expression in both muscle types, although there were significant decreases of 1.7-fold and 1.6-fold in Gl-s6k mRNA levels in claw and thoracic muscles, respectively, between 2 and 10 days post-molt (Fig. 7D). In thoracic muscle, a transient 1.4-fold decrease in Gl-s6k expression in early premolt (R=10.0) animals was reversed at later premolt stages (Fig. 7D). Gl-s6k mRNA levels in thoracic muscle were significantly lower than those in claw muscle at all premolt and post-molt stages (Fig. 7D, asterisks).

The levels of Gl-Akt, Gl-mTOR and Gl-s6k mRNAs in both muscle types were not correlated with hemolymph ecdysteroid titers in animals induced to molt by either ESA or MLA (data not shown). Gl-Rheb mRNA level in claw muscle was correlated significantly with hemolymph ecdysteroids in both MLA and ESA animals (Fig. 8A,B, solid lines). There was no significant correlation between Gl-Rheb mRNA and ecdysteroid levels in the corresponding weighted thoracic muscles (Fig. 8A,B, dashed lines).

Gl-Rheb transcript level was quantified using the same cDNA samples as those used for quantifying the Gl-Mstn transcript (Covi et al., 2010). Given the negative correlation of Gl-Mstn expression with ecdysteroids reported in Covi et al. (Covi et al., 2010) and the positive correlation of Gl-Rheb expression with ecdysteroid levels in this study (see above), we expected a negative correlation between Gl-Rheb and Gl-Mstn mRNA levels, indicating a linkage between Gl-Mstn and Gl-Rheb expression. However, a negative correlation was not consistently observed (Fig. 9). In MLA animals, the Gl-Mstn and Gl-Rheb transcript copy numbers in claw muscle were not significantly correlated (Fig. 9A). In ESA animals, however, there was a weak negative correlation between Gl-Mstn and Gl-Rheb mRNA levels that was statistically significant (Fig. 9B). By contrast, Gl-Rheb mRNA level was positively correlated with Gl-Mstn mRNA level in thoracic muscle from ESA and MLA animals (Fig. 9A,B). Both positive correlations in thoracic muscle were statistically significant.

The expression of mTOR components was quantified in the same cDNA samples from Covi et al. (Covi et al., 2010). In intact (intermolt) animals, unweighting increased Gl-Rheb mRNA level 2.2-fold and Gl-s6k mRNA level 1.3-fold; there was no effect of unweighting on the expression of Gl-Akt and Gl-mTOR (Fig. 10A). Gl-Rheb was expressed at significantly higher levels in unweighted thoracic muscle than weighted thoracic muscle expression over the molt cycle (Fig. 10B, asterisks). Gl-Rheb expression in unweighted thoracic muscle remained elevated during premolt (Fig. 10B). There were small increases of 1.3-fold and 1.4-fold over intermolt levels at 2 and 10 days post-molt, respectively. The expression of Gl-Akt, Gl-mTOR and Gl-s6k in unweighted thoracic muscle during the premolt and post-molt periods paralleled expression in weighted thoracic muscle (data not shown).

The insulin/mTOR signaling pathway is highly conserved in metazoans and functions as a nutrient sensor that is crucial for growth and development in insects and other invertebrates (reviewed by Hietakangas and Cohen, 2009; Soulard et al., 2009; Teleman, 2010). In order to determine its function in crustaceans, we cloned six mTOR pathway components representing four different genes from three crustacean species (Table 2). cDNAs encoding Rheb were cloned from G. lateralis, C. maenas and H. americanus, and deduced amino acid sequences had high identity with orthologs in other species (Figs 2 and 3; Table 2). Rheb was expressed in all tissues examined (Fig. 5), which is consistent with its ubiquitous expression in other organisms (Aspuria and Tamanoi, 2004). Partial cDNAs encoding Akt, mTOR and s6k were cloned from G. lateralis and showed high levels of sequence identity when compared with human orthologs (Table 2). This level of relatedness was even higher when compared by BLASTX against top hits among insect sequences in the database (68–81% identity and 79–89% similarity).

The decapod crustacean Rheb contained the five highly conserved ‘G boxes’ for GTP binding and GTPase activity and the C-terminal lipid modification site (Figs 1 and 2; CAAX, where C=Cys, A=aliphatic residue, X=any residue but usually Ala, Cys, Met, Gln or Ser) (reviewed in Aspuria and Tamanoi, 2004; Wennerberg et al., 2005), which is required for proper targeting to membranes and the downstream effects of Rheb (Castro et al., 2003). The 15-amino acid effector domain, which includes the G2 box (Fig. 2), is important for Rheb activation of mTOR (Aspuria and Tamanoi, 2004; Ma et al., 2008). The effector domain also mediates binding of Rheb to FKBP38, which is thought to inhibit mTOR activity under nutrient-poor conditions (Dunlop et al., 2009; Ma et al., 2008). However, there is contradictory evidence pertaining to the role of FKBP38 in modulating Rheb activity (Wang et al., 2008).

Sequence alignment of the 3′ UTR identified several highly conserved regions in the crustacean Rheb cDNAs (Fig. 4). Of particular interest was the potential regulation of Rheb by the K-box motif (TGTGAT). The K-box was first identified in Drosophila melanogaster, where it functions in Notch signaling (Lai et al., 1998). Although this motif is absent from many arthropod Rheb genes (e.g. those in copepods, sand fly, silkworm and several mosquitoes) and many mammalian Rheb genes, except human (NM_005614) and cattle (NM_001031764), the complete identity within the decapod Rhebs suggests that the K-box is an important regulatory element in this group of organisms. Other studies have shown that the K-box is regulated by microRNAs in Drosophila and mouse (Lai et al., 2005; Wu et al., 2008). This motif is also important for development in the silkworm (Zeng et al., 2009) and is present in the 3′ UTR of cyclin B in the Chinese mitten crab, Eriocheir sinensis (Fang and Qiu, 2009). Given the successful application of RNAi technology in crustaceans (reviewed in Shekhar and Lu, 2009) and the presence of the microRNA machinery in shrimp (Su et al., 2008), it is likely that post-transcriptional regulation of crustacean Rheb is mediated by microRNAs that may target the K-box motif.

Two molt induction methods were used to determine the effects of ecdysteroids on the expression of mTOR signaling components. Both methods increase hemolymph ecdysteroid titers but over different temporal scales (Skinner, 1985). ESA increased ecdysteroid level by 1 day post-ESA; animals entered premolt within a few days and reached late premolt (Stage D₂) 20 days post-ESA (Fig. 6E). Animals molt ∼25 days post-ESA (Skinner and Graham, 1972). However, ESA animals experience a 100% mortality at ecdysis, indicating that an acute elevation of ecdysteroid does not mimic a natural molt (Covi et al., 2010; Skinner and Graham, 1972). By contrast, animals tolerate MLA well and all successfully molt (Covi et al., 2010; Skinner and Graham, 1972). Animals enter premolt several weeks after MLA, which allows time for the basal limb regenerates to differentiate (Mykles, 2001). Animals molt 3–4 weeks after entering premolt (Skinner and Graham, 1972). Thus, ESA animals reach ecdysis in about half the time as MLA animals (Skinner and Graham, 1972). In this study, hemolymph ecdysteroid levels increased more slowly in MLA animals than in ESA animals, when comparing animals at the same premolt stage. At R≈10 and ≈15, the ecdysteroid titers were higher in ESA animals than in MLA animals; by late premolt, the titers were the same (compare Fig. 6E and Fig. 7E).

The only consistent effect of molt induction on mTOR pathway components was an increase in Gl-Rheb expression in claw muscle. ESA increased Gl-Rheb mRNA level 3.9-fold, which was sustained through the premolt period (Fig. 6B). MLA increased Gl-Rheb mRNA level during premolt, reaching a maximum in late premolt that was 3.4-fold greater than that in intact animals (Fig. 7B). By contrast, molt induction had little or no effect on Gl-Rheb expression in thoracic muscle (Fig. 6B and Fig. 7B). Thoracic muscle served as an internal control, as it is exposed to the same hormonal environment but does not atrophy (Griffis et al., 2001). Qualitative and quantitative differences in the expression of ecdysteroid receptor (EcR/RXR) isoforms may contribute to the differences in response of the claw and thoracic muscles to ecdysteroids (Kim et al., 2005). Moreover, there is a significant correlation between Gl-Rheb mRNA level in the claw muscle and the level of ecdysteroid in the hemolymph but not in the thoracic muscle (Fig. 8). These data strongly suggest that Gl-Rheb expression is upregulated by ecdysteroids in claw muscle.

The preferential upregulation of Gl-Rheb in claw muscle during premolt indicates that Rheb plays a role in regulating protein turnover in molt-induced claw muscle atrophy. As a key regulator of mTOR, Rheb stimulates protein synthesis in mammals and invertebrates (Dunlop and Tee, 2009; McCarthy and Esser, 2010; Soulard et al., 2009; Teleman, 2010). It is likely that increased Gl-Rheb expression is responsible, at least in part, for the increase in global protein synthesis in atrophic claw muscle (Covi et al., 2010). Increased Gl-Rheb transcription would increase the pool of Gl-Rheb protein capable of activating mTOR. This is what occurs when Rheb is overexpressed in mammalian and insect cells (Aspuria and Tamanoi, 2004; Dunlop and Tee, 2009; Hall et al., 2007; Ma et al., 2008). For example, Rheb overexpression increases translational initiation and protein synthesis in rat fibroblasts (Kubica et al., 2008). The increase in Gl-Rheb mRNA (Fig. 7B) parallels the increase in soluble and myofibrillar protein synthesis in animals induced to molt by MLA (Covi et al., 2010). However, the increases in Gl-Rheb mRNA and protein synthesis are not the same magnitudes (∼4-fold vs 11-fold and 13-fold, respectively), indicating that mTOR-mediated protein synthesis is not regulated solely by Gl-Rheb transcription. Gl-Rheb is probably regulated post-transcriptionally as well. Rheb-GAP is a heterodimer of TSC1 and TSC2 subunits that inactivates Rheb by stimulating the hydrolysis of GTP to GDP. Rheb-GAP in turn is inhibited when phosphorylated by Akt (Drummond et al., 2009a; McCarthy and Esser, 2010; Teleman, 2010). Thus, activation of Akt maintains more Rheb in the activated state by inhibiting Rheb-GAP. Overexpression of a constitutively active Akt increases TSC2 phosphorylation and mTOR activity in mouse myotubes (Ohanna et al., 2005). TSC2 phosphorylation is decreased in atrophic mammalian muscle, which is consistent with the reductions in mTOR activity and protein synthesis under these conditions (Amirouche et al., 2009). Overexpression of TSC1 causes atrophy in mouse muscle by inhibiting mTOR (Wan et al., 2006). In human skeletal muscle, increased Rheb expression coincides with increased protein synthesis after resistance exercise, which is inhibited by rapamycin (Drummond et al., 2009b; Drummond et al., 2009c). Taken together, Rheb, an integral component of the mTORC1, is an important regulator of protein synthesis in mammalian and crustacean muscles. It is likely that increases in the levels of Gl-Rheb protein contribute to the large increase in global protein synthesis in premolt claw muscle (Covi et al., 2010; Skinner, 1965).

Transcription of other components of the mTOR signaling pathway in claw and thoracic muscles are not regulated strictly by ecdysteroid titer. Any changes in the mRNA levels of Gl-Akt, Gl-s6k and Gl-mTOR during premolt were transient (Fig. 6A,C,D and Fig. 7A,C,D), and there was no correlation of the mRNA levels with hemolymph ecdysteroid level (data not shown). It is likely that Gl-Akt, Gl-s6k and Gl-mTOR are regulated by phosphorylation. In mammals, various conditions that alter muscle mass can affect Akt, mTOR, 4E-binding protein 1 and s6k phosphorylation with little effect on total protein levels. For example, fasting or Mstn upregulation reduces phospho-Akt, phospho-TSC2, phospho-s6k and phospho-4E-binding protein 1 in mammalian muscle (Amirouche et al., 2009; Trendelenburg et al., 2009; Welle et al., 2007). Conversely, stretching and Mstn downregulation increase phospho-Akt, phospho-s6k and phospho-4E-binding protein 1 (Agata et al., 2009; Welle et al., 2009). This does not mean that molting does not influence Gl-Akt, Gl-mTOR and Gl-s6k expression. There were significant changes in mRNA levels as a function of molt stage and/or muscle type. However, as there were no consistent differences between claw and thoracic muscle with respect to hemolymph ecdysteroid titer, we conclude that transcriptional regulation of Gl-Akt, Gl-mTOR and Gl-s6k does not contribute to the increase in protein synthesis during molt-induced claw muscle atrophy.

Ecdysteroids regulate protein metabolism in the claw muscle (reviewed by Mykles, 1997; Mykles, 1999; Mykles and Skinner, 1982a). The expression of both Gl-Rheb and Gl-Mstn in claw muscle is sensitive to ecdysteroid, as indicated by the correlation between mRNA levels and hemolymph ecdysteroid titer. Gl-Mstn is strongly downregulated by ecdysteroid (Covi et al., 2010) whereas Gl-Rheb was strongly upregulated (Fig. 8). As Mstn signaling regulates gene expression (Kollias and McDermott, 2008), ecdysteroids may increase Gl-Rheb expression by suppressing Gl-Mstn expression, resulting in the stimulation of mTOR-mediated protein synthesis as discussed above. In order to determine whether expression of the two genes is coupled, Gl-Rheb mRNA level was quantified in the same cDNA preparations used to quantify Gl-Mstn mRNA level (Covi et al., 2010). Interestingly, the relationship between Gl-Rheb and Gl-Mstn transcript levels differed in the two muscles. In thoracic muscle, Gl-Rheb log copy number was positively correlated with Gl-Mstn log copy number in both MLA and ESA animals (Fig. 9). In claw muscle, Gl-Mstn and Gl-Rheb mRNA levels were not correlated in MLA animals but were negatively correlated in ESA animals (Fig. 9). These data indicate that the regulation of Gl-Rheb by Mstn is complex and is not solely dependent on Gl-Mstn mRNA level. It is possible that Gl-Rheb expression is regulated by the level and/or activity of the Gl-Mstn mature protein. Proteolytic processing, secretion, proteolytic activation and inhibitory proteins (e.g. follistatin) can affect the activity of Mstn and other transforming growth factor-β (TGFβ) factors (reviewed by Breitbart et al., 2011; Moustakas and Heldin, 2009). Moreover, TGFβ signaling is downregulated by inhibitory Smad, protein phosphatases and ubiquitin-dependent degradation of TGFβ receptors and Smads (reviewed by Liu and Feng, 2010).

Autotomy-induced atrophy of the associated thoracic muscle segment also affects expression of mTOR signaling components. The unweighted thoracic muscle remains atrophied until a new limb is regenerated and ‘weighting’ is restored, along with full limb functionality (Moffett, 1987). This type of atrophy is akin to unloading atrophy in mammals (Jackman and Kandarian, 2004; Zhang et al., 2007). Unlike molt-induced atrophy, unweighting atrophy is not regulated by ecdysteroids, as this atrophy occurs during intermolt when ecdysteroid levels are low (Fig. 6E and Fig. 7E). Gl-Mstn mRNA level in unweighted thoracic muscle is 3-fold higher than in weighted muscle in intact animals but there is no difference in Gl-Mstn expression between weighted and unweighted muscles in premolt animals (Covi et al., 2010). By contrast, Gl-Rheb mRNA levels increased 2.2-fold in unweighted thoracic muscle in intact animals and remained elevated during premolt and post-molt (Fig. 10B). Contrary to expectations, reduced expression of Gl-Rheb or the other mTOR components (Fig. 10A) was not associated with higher levels of Gl-Mstn (Covi et al., 2010). In fact, Gl-Rheb and Gl-Mstn mRNA levels were positively correlated, emphasizing the differences between molt-induced atrophy and unweighting atrophy. The upregulation of Gl-Rheb and Gl-s6k suggests that unweighting stimulates global protein synthesis in thoracic muscle. This is an area for future study.

In summary, we have shown that Gl-Rheb is transcriptionally regulated by molting and unweighting in claw closer and thoracic muscles, respectively. Of the mTOR signaling genes, only Gl-Rheb mRNA levels were consistently increased in atrophic muscle. However, there were significant differences in how Gl-Rheb was regulated in the two muscles. Gl-Rheb expression in claw muscle appears to be stimulated by ecdysteroids, as Gl-Rheb mRNA level was correlated with hemolymph ecdysteroid level. This is consistent with our hypothesis that ecdysteroids stimulate mTOR-mediated protein synthesis. Gl-Rheb expression in thoracic muscle is not regulated by ecdysteroids, as there was no correlation between Gl-Rheb mRNA and ecdysteroid levels. The signal transduction pathway mediating the increased expression of Gl-Rheb in thoracic muscle in response to unweighting remains to be established. If Gl-Mstn plays a role in both types of atrophy, any unifying mechanism must reconcile the differences between Gl-Mstn and Gl-Rheb expression in the two muscles. In claw muscle, Gl-Mstn is downregulated when Gl-Rheb is upregulated whereas, in unweighted thoracic muscle, both Gl-Mstn and Gl-Rheb are upregulated. One possibility is that Gl-Mstn is a positive regulator of Gl-Rheb in thoracic muscle but is a neutral or negative regulator of Gl-Rheb in claw muscle. In shrimp (Penaeus monodon), RNAi knockdown of Pm-Mstn expression in abdominal muscle slows growth over a 45-day period (De Santis et al., 2011), suggesting that Mstn can stimulate muscle growth. Alternatively, regulation in thoracic muscle may instead be primarily post-transcriptional. The expression of Gl-Mstn and Gl-Rheb in both muscle types was not tightly coupled, indicating that transcriptional regulation of Gl-Rheb is complex and may differ between the claw and thoracic muscles in the relative contributions of ecdysteroid and Mstn regulation. In claw muscle, ecdysteroids may upregulate Gl-Rheb directly and Gl-Mstn may regulate other genes that inhibit the insulin/mTOR signaling pathway. For example, overexpression of Rheb-GAP inhibits mTOR activation and reduces fiber cross-sectional area in mammalian skeletal muscle (Wan et al., 2006). A further complication is that Mstn and its signaling through Smad transcription factors is regulated at multiple levels (e.g. Mstn processing, secretion and activation and modulation of Activin receptor and Smads) (Breitbart et al., 2011; Moustakas and Heldin, 2009) that could be affected by ecdysteroids. As transcription of Gl-Rheb in thoracic muscle does not involve ecdysteroids, Mstn/Smad or perhaps other signaling pathways have a greater role in regulating Gl-Rheb expression in unweighting atrophy. Now that key components of the insulin/mTOR and Mstn/Smad signaling pathways have been identified, future studies can be directed at gaining a mechanistic understanding of the interactions between ecdysteroids and the signaling pathways that control protein metabolism in crustacean skeletal muscle.

 
• ESA

  eyestalk ablation

 
• EST

  expressed sequence tag

 
• GAP

  GTPase-activating protein

 
• Gl

  Gecarcinus lateralis

 
• MLA

  multiple leg autotomy

 
• Mstn

  myostatin

 
• mTOR

  metazoan target of rapamycin

 
• mTORC1

  metazoan target of rapamycin Complex 1

 
• ORF

  open reading frame

 
• qPCR

  quantitative polymerase chain reaction

 
• RACE

  rapid amplification of cDNA ends

 
• Rheb

  Ras homolog enriched in brain

 
• R-index

  regeneration index

 
• RT-PCR

  reverse transcriptase-polymerase chain reaction

 
• s6k

  ribosomal S6 kinase

 
• TGFβ

  transforming growth factor-β

 
• TOR

  target of rapamycin

 
• TSC

  tuberous sclerosis complex

The authors thank Kira Marshall for assistance with tissue cDNA preparation for G. lateralis and C. maenas; Young Ji Kim and Sun Ok Lee for assistance with cloning the Gl-Rheb UTR sequences; Sharon A. Chang for measuring ecdysteroids by radioimmunoassay; Hector C. Horta for collecting G. lateralis; Staci Amburgey, Lisa Axtman, Peter Exner, Maddie Fountain, Jen Gunderson, Stefan LaBere, Kira Marshall, Katie Regelson, Ashanti Robinson and Sere Williams for animal care; and Christine M. Smith and David W. Towle, Mount Desert Island Biological Laboratory, for providing EST clones of lobster and green crab Rheb. We recognize the assistance of Ricardo Colon Alvarez (Executive Director, Consejo Dominicano de Pesca y Acuicultura, Dominican Republic) and his staff for expediting approval of permits, identifying suitable sites and collecting G. lateralis.