Catabolic Signaling Pathways, Atrogenes, and Ubiquitinated Proteins Are Regulated by the Nutritional Status in the Muscle of the Fine Flounder

A description of the intracellular mechanisms that modulate skeletal muscle atrophy in early vertebrates is still lacking. In this context, we used the fine flounder, a unique and intriguing fish model, which exhibits remarkably slow growth due to low production of muscle-derived IGF-I, a key growth factor that has been widely acknowledged to prevent and revert muscle atrophy. Key components of the atrophy system were examined in this species using a detailed time-course of sampling points, including two contrasting nutritional periods. Under basal conditions high amounts of the atrogenes MuRF-1 and Atrogin-1 were observed. During fasting, the activation of the P38/MAPK and Akt/FoxO signaling pathways decreased; whereas, the activation of the IκBα/NFκB pathway increased. These changes in signal transduction activation were concomitant with a strong increase in MuRF-1, Atrogin-1, and protein ubiquitination. During short-term refeeding, the P38/MAPK and Akt/FoxO signaling pathways were strongly activated, whereas the activation of the IκBα/NFκB pathway decreased significantly. The expression of both atrogenes, as well as the ubiquitination of proteins, dropped significantly during the first hour of refeeding, indicating a strong anti-atrophic condition during the onset of refeeding. During long-term refeeding, Akt remained activated at higher than basal levels until the end of refeeding, and Atrogin-1 expression remained significantly lower during this period. This study shows that the components of the atrophy system in skeletal muscle appeared early in the evolution of vertebrates and some mechanisms have been conserved, whereas others have not. These results represent an important achievement for the area of fish muscle physiology, showing an integrative view of the atrophy system in a non-mammalian species and contributing to novel insights on the molecular basis of muscle growth regulation in earlier vertebrates.


Introduction
The maintenance of skeletal muscle mass is a complex and controlled process that is largely influenced by the nutritional and physiological state of the animal. This dynamic process is regulated by a balance between protein synthesis and protein degradation; however, when rates of protein degradation exceed rates of protein synthesis, then muscle mass is lost, leading to atrophy of this tissue [1]. Muscle atrophy occurs as a consequence of denervation, injury, joint immobilization, glucocorticoid treatment, sepsis, cancer, and aging [1]. Food deprivation and undernourishment are also two main conditions that promote muscle atrophy; thus, highlighting that nutritional status has a major role in skeletal muscle mass regulation [2,3]. The major route that increases overall rates of protein degradation during muscle atrophy is the ubiquitin-proteasome pathway [4,5]. Polyubiquitination of proteins is a multiple-step process that requires ATP and the participation of three components in the formation of the ubiquitin-protein complexes, the ubiquitinactivating enzyme (E1), a ubiquitin-conjugating enzyme (E2), and a ubiquitin-ligase (E3) [6,7,8], in order to covalently attach multiple ubiquitin molecules to the protein substrate [6]. Subsequently, these tagged proteins are recognized and degraded by the 26S proteasome, resulting in short peptides [8]. Particularly, the ubiquitin-ligases are a family of key enzymes responsible for transferring an activated ubiquitin molecule to a targeted protein, subsequently marking the protein for proteasomal degradation [8]. Indeed, an increase in the capacity for protein degradation via the proteasome is dependent on an increase in ubiquitin-ligase expression [9]. Numerous ubiquitin-ligases have been identified; however, differential expression screening studies, originally planned to detect high-fidelity markers of muscle atrophy, led to the discovery of two genes that encode ubiquitinligases, MuRF-1 (Muscle Ring Finger protein-1) and Atrogin-1 (also called Muscle Atrophy F-box (MAFbx) [10], which have been shown to be upregulated in several models of skeletal muscle atrophy, validating them as reliable markers of atrophy [11,12,13].
The transcriptional regulation of these ubiquitin-ligases, also named atrogenes, during catabolic-atrophic processes in skeletal muscle has been linked to the activation of different signaling pathways, such as the mitogen-activated protein kinases (MAPKs), specifically the P38; the protein kinase B (Akt)/forkhead box O (FoxO); and the inhibitor of kappa, alpha (IkBa)/nuclear factor kappa B (NFkB) [13]. The Akt/FoxO signal transduction is the only signaling pathway able to regulate the expression of both atrogenes, upregulating their transcription when Akt activation decreases and reducing the phosphorylation FoxO1/3 transcription factors, promoting their nuclear translocation [9]. On the other hand, the other two signaling pathways independently stimulate the expression of these atrogenes, showing functional separation of the expression of MuRF-1 and Atrogin-1 [13]. The activation of the IKKb/IkBa/NFkB signaling pathway promotes the upregulation of MuRF-1, causing severe muscle wasting and atrophy, a phenomenon that is reverted when this signaling pathway is blocked [14]. In general terms, NFkB in an inactivated state is located in the cytosol of muscle cells and is associated with the inhibitory protein IkBa. Upon stimulation, IkBa is phosphorylated by the IkB kinase (IKK), ubiquinated, and degraded, thus dissociating the complex formed by IkBa and NFkB and allowing the translocation into the nucleus of NFkB, ultimately promoting the transcription of MuRF-1 [13,14]. Conversely, Atrogin-1 upregulation is promoted by the P38/MAPK signaling pathway [15], through FoxO4 [16], and the Akt/FoxO1/3 signaling pathway [9,17].
Knowledge concerning the molecular mechanisms that modulate muscle atrophy in fish has been scarce and limited, mainly focusing on the cloning and evaluation of expression patterns of both atrogenes [18][19][20][21][22]. A few studies have gone further in the understanding of muscle atrophy in teleosts by using transcriptomic and proteomic approaches [23][24][25]. However, assessment of the intracellular pathways modulating muscle catabolism, as well as the main components of the atrophy system, has not been performed (e.g. P38/MAPK, Akt/FoxO, and IkBa/NFkB signaling pathways, MuRF-1 and Atrogin-1, ubiquitin proteasomedependent proteolysis). Hence, the study of this system represents important progress for the area of fish muscle physiology, especially due to the unique features displayed by this tissue in this group of vertebrates, which include muscle atrophy as part of a natural seasonal cycle, due to changes in food availability, and indeterminate growth by hyperplasia and hypertrophy [26][27][28][29].
The aim of this study was to assess the atrophy system in the skeletal muscle of a teleost fish using a detailed time-course of sampling points including two contrasting nutritional periods: an extended catabolic period of fasting, leading to muscle atrophy, followed by an anabolic period of refeeding, leading to compensatory muscle hypertrophy. We used the fine flounder (Paralichthys adspersus) as a model, which is a unique and intriguing flatfish species that exhibits natural growth deficiency [30,31]. The responsible mechanism for this impairment in growth has been associated to an inherent growth hormone (GH) resistance in skeletal muscle. Thus, this fish represents a unique model of low basal production of muscle-derived insulin-like growth factor-I (IGF-I) [31], which is a key molecule that induces muscle hypertrophy via the Akt/mTOR pathway, concomitant with preventing atrophy [10,13,32], by downregulating the expression of MuRF-1 and Atrogin-1 [9,33,34] acting through the PI3K/Akt/ FoxO pathway [9,33].

Ethics Statement
The study adhered to animal welfare procedures and was approved by the bioethical committees of the Universidad Andres Bello and the National Commission for Scientific and Technological Research of the Chilean government.

Fish Husbandry, Experimental Design, and Sampling
Three-year old, sexually immature juvenile fine flounder (Paralichthys adspersus), with an average weight of 300610 g, were maintained under natural temperature (12uC63) and photoperiod (13 hrs light: 11 hrs dark) conditions corresponding to the geographic location (33u139S; 71u389W) of the Centro de Investigacion Marina de Quintay (CIMARQ) in the southern hemisphere during the spring season of 2009 (October-December). Fish were hand fed once daily with 9-mm commercial pellet containing 45% protein, 22% lipids, 16% carbohydrates, 1% crude fiber, 7% ashes, and 10% humidity (Skretting, Puerto Montt). Fish were randomly divided into two circular (ø 3.85 m) fiberglass tanks (150 fish per tank). The tanks were 75 cm in height with a water column of 30 cm, and a water turn-over of 15.4 L min 21 (one water clearance per hour).
The experimental design consisted in fish acclimatized for two weeks under satiety feeding conditions. At the start of the experiment (week 0), food was withheld from fish for three weeks, inducing a nutritional catabolic state of fasting. Then, fish were subjected to a four week satiety refeeding period, which returned the fish to an anabolic state. Body weight and length were measured to show the effects of nutritionally-induced catabolic/ anabolic states on growth performance [31].
Samples were obtained weekly over the trial (i.e. fasting and refeeding) with the purpose of studying long-term changes in the atrophy system of skeletal muscle. Also, in order to study shortterm changes of this system during the refeeding period, fish were sampled at 2, 4, and 24 hours after the first meal supply following the three weeks of fasting. For each sampling point, three individuals were sampled (n = 3). Sampling was performed under anesthesia (3-aminobenzoic acid ethyl ester, 100 mg/l 21 ), and white-fast myotomal muscle was subsequently collected. Tissue was frozen immediately in liquid nitrogen and stored at 280uC until processing.

Western Blotting
Total protein were extracted and prepared according to Fuentes et al., [31,35]. Protein concentration was determined by a PierceH BCA Protein Assay Kit (Thermo Scientific, IL, USA).
Western blot assays were performed according to Fuentes et al., [31,35]. Briefly, proteins were transferred to polyvinylidene diflouride membranes (PVDF) (Millipore, Bedford, MA, USA) and blocked for 1 h at room temperature in 2% ECL Advance TM blocking agent (GE Healthcare, Buckinghamshire, UK) dissolved in tris-buffered saline (TBS 16). Primary antibody incubations (phosphorylated (p) P38, Akt, FoxO, IkBa and total Ubiquitin) were performed at 4uC overnight. All antibodies were commercial antibodies purchased from Cell Signaling (Beverly, MA). Details of the antibodies are found in Table 1. Membranes were visualized by a high sensitivity enhanced chemiluminescence kit, the ECL Advance TM Western Blotting Detection Kit (GE Healthcare, Buckinghamshire, UK), according to the manufacturer's instructions. Subsequently, membranes were stripped and blotted for total P38, Akt, FoxO, and IkBa antibodies respectively (Table 1), therefore obtaining the ratio between the phosphorylated protein with the total protein. For ubiquitinated protein assessment, they were normalized respectively by comparison with Coomassie blue (Cb) staining of total protein. The obtained films were then scanned, and densitometric analysis of the bands was performed with the Image J program (National Institute of Health, USA). Western blots were carried out on three individual samples (N = 3 per sampling time-point), showing a representative blot film. All graphs for long-term changes during fasting and refeeding are expressed as a fold change over the basal levels found at the beginning of the trial (week 0), whereas for short-term refeeding, graphs are expressed as a fold change over levels found at the end of fasting (week 3).
In order to validate and confirm that antibodies developed against mammalian epitopes cross-react with samples of skeletal muscle of the fine flounder, comparative western blots were performed using rat myosatellite cell primary cultures as a control, indicating that all antibodies cross-react with their orthologs in the flounder (Fig. 1). Akt and pAkt antibodies were previously validated by Fuentes et al., [35].

RNA Extraction and cDNA Synthesis
Total RNA was extracted from skeletal muscle using the RNeasy Mini Kit (Qiagen, Austin, TX, USA) following manufacturer's recommendations. RNA was quantified using NanoDrop technology with the Epoch Multi-Volume Spectrophotometer System (BioTek, Winooski, VT, USA). Assessment of RNA quality was performed by electrophoresis on a 1.2% formaldehyde agarose gel containing ethidium bromide. Only RNAs with an A260/280 ratio between 1.9 and 2.1 were used for cDNA synthesis. Residual genomic DNA was removed using the genomic DNA wipeout buffer included in the QuantitectH reverse transcription kit (Qiagen, Austin, TX, USA). Subsequently, 800 ng of RNA were reverse transcribed into cDNA for 30 min at 42uC using the manufacturer's recommendations.

Isolation and Cloning of Atrogenes
Once high quality cDNA was obtained from muscle as previously described, MuRF-1 and Atrogin-1 were isolated and deposited in the GeneBank, GeneBank Accession Numbers are indicated in Table 2. Primers used for obtaining the sequences of atrogenes were designed by multiple alignments of sequences from fish species using ClustalW (Table 2). PCR was performed using 1mL cDNA template, 5 mL of PCR buffer 10X, 200 mM of each dNTP, 500 nM of each primers, 0.3 mL of Taq DNA polymerase (12 U/mL) (Promega, Madison, WI, USA), and RNAse-free water to a final volume of 50 ml. Thermal cycling conditions were the following: initial activation of 10 min at 95uC, followed by 40 cycles of 30 s at 95uC; 30 s at 55uC (MuRF-1) or 57uC (Atrogin-1); and 30 s at 72uC with a final extension of 10 min at 72uC.
Cloning procedure was carried out using the TOPO TA CloningH system (Invitrogen, Carlsbad, CA, USA) using manufacturer's recommendations. In short, PCR products were ligated into the T/A pCR4-TOPO vector (Invitrogen, Carlsbad, CA, USA), and subsequently, One ShotH TOP10 competent E. coli (Invitrogen, Carlsbad, CA, USA) were transformed with the vector. Individual colonies were cultured, and plasmids were isolated, purified using the QIAGENH Plasmid Purification (QIAGEN), and subsequently sequenced.
All the procedures were carried out according to the protocol outlined by Bower et al., [36] with minor modifications. All qPCR assays were carried out to complying with the Minimum Information for Publication of Quantitative Real-Time PCR experiments MIQE guidelines [37].
Total RNA extraction and cDNA synthesis from skeletal muscle of all sampling points were performed as described above. Quantitative PCR (qPCR) was performed using the Stratagene MX3005P qPCR system (Stratagene). Each qPCR reaction mixture contained 7.5 mL Brilliant II SYBR green master mix (Agilent Technologies, SC, USA), 6 mL cDNA (40fold dilution), 250 mM of each primer, 5 mM ROX, and RNAse-free water to a final volume of 15 mL. Amplifications were performed in triplicate with the following thermal cycling conditions: 95uC for 10 min, followed by 35 cycles of 30 s at 95uC, 30 s at 54uC (MuRF-1) or 56uC (Atrogin-1), and 30 s at 72uC. In order to confirm the presence of a single PCR product, dissociation curve analysis of the PCR products was performed. Products were also evaluated by electrophoresis on a 1.5% agarose gel to confirm that a single product was amplified. With the purpose of estimating the efficiency of the assays, 2-fold dilutions series were created. Efficiency values were estimated from the slope of the curve following the equation: efficiency E = 10(21/slope) 21 (Table 2). Control reactions included a no template control (NTC) and a control without reverse transcriptase (-RT). An interplate calibrator (IPC) was used in all runs in order to correct plate to plate variation. Moreover, in order to corroborate the results, qPCR experiments were independently performed two times. qPCRs were carried out on three individual samples (N = 3 per sampling time-point). For long-term observations (i.e. fasting and refeeding), graphs are expressed as a fold change over the basal levels found at the beginning of the trial (week 0). For short-term observations (i.e. short-term refeeding), graphs are expressed as a fold change over levels found at the end of fasting (week 3). For atrogenes' mRNA contents, graphs are represented as arbitrary units. These genes were obtained by using the geNorm program, which obtained the normalization factor and subsequent relative expression levels of both atrogenes [38].

Heat Map Summary of Hierarchical Clustering of Atrophy System Data
In order to establish relationships among all the components of the atrophy system, a heat map summary and hierarchical clustering analysis were performed using Permutmatrix [39]. All data points obtained from the activation of the signaling pathways, relative expression profiles of both atrogenes, and ubiquitinated proteins throughout the time-course events in different feeding statuses were incorporated into the Permutmatrix software. Clustering and seriation were based on Pearson's correlation coefficient of z-score normalized abundance values (scaled from 0 to 1). McQuittys method was used as a hierarchical clustering. The Permutmatrix program was also used to illustrate signaling pathway activation, relative expression profiles, and the amount of ubiquitinated proteins throughout the time-course events in different feeding statuses [39].

Statistical Analyses
Statistical analyses used for studying differences in gene expression were based on an advanced linear model. This model was the general linear model (GLM), and was followed by Tukey's analyses as post-test. Also, correlations between different parameters (i.e. molecules of the atrophy system) were assessed through multiple linear regressions, obtaining the coefficient of determination (R 2 ) and the P-value. All statistical analyses were performed using the STATISTICA 7 software (Tulsa, OK, USA).

Results
The Atrophy Signaling Pathways (P38/MAPK, Akt/FoxO, and IkBa/NFkB) in the Skeletal Muscle of the Fine Flounder During Nutritionally-induced Catabolic and Anabolic Periods The P38/MAPK, Akt/FoxO, and IkBa/NFkB signal transduction pathways were assessed during fasting and refeeding in the teleost species the fine flounder in order to study the activation dynamics of the main signal transductions associated with muscle atrophy. The P38/MAPK activation pathway displayed a steady decrease during fasting, becoming statistically significant after two weeks ( Fig. 2A). During long-term refeeding, this signaling pathway returned to basal levels of activation, except during the first week of refeeding where P38 significantly increased ( Fig. 2A). During short-term refeeding after the three weeks of fasting, P38/ MAPK signaling pathway activation increased steadily and showed significant differences at 24 hours (Fig. 2B). Akt/FoxO signaling pathway activation also decreased during fasting and became significantly diminished after three weeks (Fig. 3A, 3C). This pathway returned to basal levels during longterm refeeding; however, it was Akt, not FOXO, activation that increased significantly at the end of long-term refeeding (Fig. 3A,  3C). During short-term refeeding after the three weeks of fasting, Akt/FoxO signaling pathway activation increased steadily, showing significant differences at 4 and 24 hours (Fig. 3B, 3D).
On the other hand, activation of the IkBa/NFkB signaling pathway rapidly increased during fasting, with significant differences found in the activation of IkBa from the first until the last week of fasting (Fig. 4A). IkBa activation returned to basal levels during long-term refeeding (Fig. 4A). Short-term refeeding observations showed that IkBa activation decreased rapidly, showing significant differences as early as two hours after refeeding (Fig. 4B). A significant decrease from the first until the third week of fasting was observed when assessing total IkBa protein contents, indicating degradation of IkBa (Fig. 4C). IkBa protein contents returned to basal levels during long-term refeeding. During shortterm refeeding total IkBa protein contents increased, reflecting a decrease in the degradation of this protein (Fig. 4D).

Expression of Atrogenes in Skeletal Muscle during Nutritionally-induced Catabolic and Anabolic Periods
In order to study the dynamic of the two main atrogenes involved in muscle atrophy, which are direct downstream targets of the signaling pathways assessed above, the expression of MuRF-1 and Atrogin-1 was studied. During fasting, MuRF-1 expression in muscle increased rapidly, with almost four-fold higher mRNA levels than basal levels from the first to the third week (Fig. 5A). During long-term refeeding, MuRF-1 expression was restored to basal levels (Fig. 5A). During short-term refeeding, the most drastic changes were observed. In particular, MuRF-1 decreased abruptly during this period, with a difference of more than one thousand-fold lower mRNA levels after 4 to 24 hours of refeeding than at the immediate end of fasting (09 hours) (Fig. 5B).
Like MuRF-1, Atrogin-1 also increased during fasting; nevertheless, changes in the expression of this atrogene were more radical and displayed a steady increase. More than one hundred-fold higher mRNA levels of Atrogin-1, as compared to basal conditions, were found at the end of fasting (Fig. 5C). During long-term refeeding, Atrogin-1 decreased to lower than basal levels in all sampling points (Fig. 5C). During short-term refeeding, Atrogin-1 also showed a drastic decrease, displaying more than twenty-fold lower mRNA levels during all sampling points than at the immediate end of fasting (09 hours) (Fig. 5D).
By assessing specific levels of expression for both atrogenes, it is possible to observe that they were differentially expressed at basal levels, with higher levels of mRNA contents for MuRF-1 than for Atrogin-1 (MuRF-1 was expressed five-fold more than Atrogin-1) (Fig. 5E). During fasting, both MuRF-1 and Atrogin-1 mRNA levels increased significantly, reaching similar levels (Fig. 5E, 5F). During long-term refeeding, MuRF-1 was more expressed than Atrogin-1, and a similar trend to basal levels was found; however, the difference between the atrogenes at the end of refeeding was ninety six-fold (Fig. 5E). During short-term refeeding, both atrogenes dropped significantly; however, MuRF-1 decreased even further than Atrogin-1 mRNA levels, with a seventy three-fold lower expression of MuRF-1 than of Atrogin-1 (Fig. 5F).

Ubiquitination of Total Protein in Skeletal Muscle during Nutritionally-induced Catabolic and Anabolic Periods
During fasting, ubiquitinated proteins increased steadily and were significant at the end of fasting (three-fold higher levels than 0 week) (Fig. 6A). During long-term refeeding, ubiquitinated proteins returned to basal levels (Fig. 6A). On the other hand, short-term refeeding triggered a rapid decrease in ubiquitinated proteins, finding significantly lower levels as early as two hours post-refeeding, and which were maintained until the first 24 hours  (Fig. 6B). The amount of free ubiquitin did not change significantly during the trial (Inserts in Fig. 6A and 6B).

Integration of Atrophy System Components in Skeletal Muscle During Nutritionally-induced Catabolic and Anabolic Periods
Hierarchical clustering of the atrophy system throughout the trial showed two clades. The first clade showed a close relation and co-variation of Akt and FoxO activation, which was subsequently clustered with P38 activation. The second clade clustered ubiquitinated proteins with MuRF-1 expression and, subsequently, Atrogin-1 expression. This clade was clustered with IkBa activation (Fig. 7A).

Discussion
The balance between muscle loss and growth, which determines skeletal muscle size, is dynamically controlled by specific signaling pathways that will either trigger an increase in protein degradation to diminish muscle mass (atrophy) or an increase in protein synthesis to stimulate muscle fiber growth (hypertrophy). The present study simulates this by withholding food from fine flounder for an extended period (3 weeks), inducing a catabolic period of muscle atrophy that is followed by a sustained nutritionally favorable period (4 weeks), which then induces an anabolic stage of compensatory muscle hypertrophy. unpublished data]. During fasting, MuRF-1 expression rapidly increases while Atrogin-1 displays a steadier and more dramatic increase, and both reach similar levels at the end of this period. During refeeding, mRNA levels of both atrogenes drop rapidly and radically. However, MuRF-1 shows an intense transcriptional regulation during only short-term refeeding, whereas Atrogin-1 shows this during both long-and short-term refeeding. These results strongly suggest that different transcriptional controls are influencing these atrogenes.
Merging previous information on the expression of both atrogenes with the activation information of different signaling pathways and of the ubiquitination of proteins, an important relationship and temporal synchronicity among some components of the atrophy system is observed; however, not for all of them.
P38/MAPK activation decreases during fasting and increases during refeeding. The activation of this pathway does not show a similar trend and synchronicity with any of the atrogenes. In mammals, the P38/MAPK pathway was initially described as controlling cellular responses to stressors including pro-inflammatory cytokines, lipopolysaccharides (LPS), and UV light [40]. Particularly, this was corroborated in skeletal muscle by showing that the tumor necrosis factor-alpha (TNF-a), a pro-inflammatory cytokine, stimulates the expression of Atrogin-1 via the P38/MAPK pathway [15]. Other more recent and detailed approaches have also supported this notion [41,42]. On the other hand, studies have shown a positive role of the P38/MAPK pathway during myogenesis through the promotion of differentiation [40,43] and inhibition of proliferation [43]. This dual role played by this signaling pathway is initiated and promoted by IGF-I [43]. Interestingly, P38 activation shows a similar pattern as that in both endocrine (circulatory/systemic, liver-derived) and autocrine/ paracrine (local, muscle-produced) IGF-I of the fine flounder . Fasting and refeeding effects on IkBa/NFkB signaling pathway activation in skeletal muscle. pIkBa/IkBa ratios during longterm fasting and refeeding (A) and short-term refeeding (B). IkBa degradation during long-term fasting and refeeding (C) and short-term refeeding (D). White, black, and grey bars represent periods of feeding, fasting, and refeeding, respectively. A probability level of P,0.05 (lower case letters) and P,0.01 (upper case letters) was used to indicate statistical significances. Results are expressed as means6SEM (n = 3). Different letters indicate significant differences among sampling points of each group, respectively. Abbreviations: 09 = zero hour of short-term refeeding corresponding to the end of fasting period (week 3). doi:10.1371/journal.pone.0044256.g004 The Atrophy System in Fish Skeletal Muscle PLOS ONE | www.plosone.org Figure 5. Transcriptional regulation of MuRF-1 and Atrogin-1 in the skeletal muscle during long-term fasting and refeeding and short-term refeeding. MuRF-1 relative expression during long-term fasting and refeeding (A) and short-term refeeding (B). Atrogin-1 relative expression during long-term fasting and refeeding (C) and short-term refeeding (D). Relative mRNA content comparison between MuRF-1 and Atrogin-1 during long-term fasting and refeeding (E) and short-term refeeding (F). White, black and grey bars represent periods of feeding, fasting and refeeding, respectively. Red and blue bars represent Atrogin-1 and MuRF-1 expression respectively. A probability level of P,0.05 (lower case letters) and P,0.01 (by upper case letters) was used to indicate statistical significances. Results are expressed as means6SEM (n = 3). Different letters indicate [31,35]. In other fish species, there are no reports about P38/ MAPK activation in muscle; thus, the present study is the first approach assessing P38 in a teleost species. A complex biology of P38, concomitant with a lack of information of this kinase in fish skeletal muscle, makes the interpretation of the data difficult. More studies are required in order to understand the role of this signal transduction in fish muscle. Meanwhile, our data suggests that P38 is not involved in the Atrogin-1-induced muscle atrophy as is observed in mammals.
The Akt/FoxO signaling pathway shows an opposite, inverse trend in comparison with the expression of both atrogenes. These results are in line with previous reports in both vertebrate models. In mammals, Akt activation, in addition to stimulating skeletal muscle hypertrophy, can significantly inhibit the induction of atrophy signaling [9,33]. Genetic activation of Akt blocks atrophyassociated increases in MuRF-1 and Atrogin-1 transcription [33], and the mechanism by which Akt inhibits the expression of both atrogenes involves the FoxO transcription factor's family and the upregulation of MuRF-1 and Atrogin-1 [9,17,33]. In fish, a single in vitro study has assessed the Akt/FoxO signaling pathway and the expression of both atrogenes, showing similar results to those in mammals [20].
The IkBa/NFkB signaling pathway is rapidly activated, concomitant with IkBa degradation during fasting, whereas during refeeding an opposite phenomenon is observed. This kinetic is highly correlated and synchronized with MuRF-1 significant differences among sampling points of each group, respectively. Abbreviations: 09 = zero hour of short-term refeeding corresponding to the end of fasting period (week 3). doi:10.1371/journal.pone.0044256.g005 Figure 6. Ubiquitination of proteins in the skeletal muscle during long-term fasting and refeeding and short-term refeeding. Ubiquitinated protein during long-term fasting and refeeding (A) and short-term refeeding (B). Inserts in A and B show the amount of free ubiquitin during the trial. Percentage of ubiquitinated proteins during long-term fasting and refeeding (C) and short-term refeeding (D). White, black and grey bars represent periods of feeding, fasting and refeeding, respectively. A probability level of P,0.05 (lower case letters) and P,0.01 (by upper case letters) was used to indicate statistical significances. Results are expressed as means6SEM (n = 3). Different letters indicate significant differences among sampling points of each group, respectively. Abbreviations: Cb = Coomassie blue staining; 09 = zero hour of short-term refeeding corresponding to the end of fasting period (week 3). doi:10.1371/journal.pone.0044256.g006 expression (R 2 = 0.82, P,0.0001), but not with Atrogin-1 (R 2 = 0.52, P.0.05). In mammals, a linear signaling pathway conformed by IKKb/IkBa/NFkB is sufficient to induce atrophy by stimulating the expression of MuRF-1 but not of Atrogin-1 [14]; which is in agreement with our results. In skeletal muscle of fish, this study constitutes the first approach linking the activation of IkBa/NFkB with the expression of MuRF-1.
Altogether, the activation of the IkBa/NFkB signaling pathway and upregulation of MuRF-1, concomitant with the inactivation of the Akt/FoxO pathway and the upregulation of Atrogin-1, are closely related and synchronized with an increase in ubiquitinated proteins, a previous step to protein degradation and muscle atrophy. The involvement of the ubiquitin-proteasome pathway in skeletal muscle atrophy has been well established [2,3,5]. Several target muscle proteins have been identified as being ubiquitinated by either MuRF-1 [44] or Atrogin-1 [44,45,46]. Particularly important was the discovery that the myosin heavy chain (MYH), one of the most abundant structural proteins in muscle, is a substrate of MuRF-1, thus identifying the mechanism by which MYH is depleted under atrophy conditions [47]. Interestingly, MuRF-1 is highly expressed in muscle of the fine flounder in basal conditions and increases during fasting. The fine flounder presents a natural growth deficiency, which is reflected in limited muscle growth [31]. It could be possible to hypothesize that a high abundance of MuRF-1 in skeletal muscle of the fine flounder might be triggering the depletion of the structural protein MYH, thus affecting muscle mass. We have previously shown that MYH expression radically deceases during fasting [

Integration of the Atrophy System in Fish Skeletal Muscle
In the present study, it has been shown that the components of the atrophy system in skeletal muscle have appeared early in the evolution of vertebrates, finding some to be evolutionarily conservative (i.e. Akt/FoxO and IkBa atrogenes expression; and ubiquitination of proteins) and others as non-conservative (i.e. P38) mechanisms in comparison with higher vertebrates. Moreover, in order to better elucidate the regulation of muscle mass in this fish species, we have placed the present results into a global framework of growth in skeletal muscle by compiling the present information with our previous reports [31,35,48].
In basal conditions, the fine flounder presents low musclederived IGF-I [31]. We have previously shown that IGF-I is able to activate the MAPK/ERK and the PI3K/Akt in muscle in vivo and contribute to somatic growth in this species [35]. In this context, we suggest that the basal impairment in local muscleproduced IGF-I might be responsible for the excessive expression of MuRF-1 and high expression of Atrogin-1; thus, atrophying muscle and subsequently affecting somatic growth (Fig. 8A). The reason why MuRF-1 is more expressed than Atrogin-1 is unknown. However, considering that we did not find evidence for a P38related expression of Atrogin-1, this might imply that expression of this atrogene could be controlled only by the Akt/FoxO signaling pathway. On the other hand, we did find evidence for a relationship between the IkBa/NFkB pathway and MuRF-1. This atrogene is also inversely related with Akt/FoxO, and thus a plausible hypothesis is that MuRF-1 is being controlled by more extracellular stimuli and signaling pathways than Atrogin-1 (Fig. 8A).
During fasting, local IGF-I levels decrease even further [31]. Thus, we suggest that these low levels of IGF-I may be inactivating the Akt/FoxO signaling pathway and upregulating both atrogenes, therefore enhancing protein ubiquitination. Concomitantly, the activation of the IkBa/NFkB pathway would also be increasing the transcription of MuRF-1, thus contributing even more to protein ubiquitination. We speculate that altogether these intracellular responses lead to muscle atrophy, diminishing somatic growth (body weight, condition factor, specific growth rate) [31,35] (Fig. 8B).
During refeeding a rapid switch in the atrophy system is observed, particularly during short-term refeeding when local muscle-derived IGF-I increases radically [31]. We suggest that an increase in IGF-I levels may be reactivating the Akt/FoxO pathway, drastically diminishing the expression of both atrogenes and the ubiquitination of proteins (Fig. 8C). Likewise, a rapid inactivation of the IkBa/NFkB signaling pathway is observed during this period, which might also be contributing to the downregulation of MuRF-1 (Fig. 8C). Thus, during the first stages of refeeding, strong anti-atrophy effects are observed which match with our previous observation that positive-anabolic signals exceed negative-catabolic signals [31,35,48] (Fig. 8C). Hence, it is possible to hypothesize that this imbalance between positive and negative signals is lay the foundation for promotion of a strong catch-up growth (body weight, condition factor, specific growth rate) during the first stages of refeeding in this species [31,35]. Subsequently, during long-term refeeding, positive and negative signals start to become balanced, with the majority of the components returning to basal levels, as with a low production of muscle-derived IGF-I [31]. Nevertheless, during this period, we still observe that positive regulators of growth are more activated (i.e. Akt (present study), circulating IGF-I, and IGFBP-4, 5 [35,48] while negative regulators of growth are still diminished (i.e. Atrogin-1 (present study), and IGFBP-3, 2 [48]). Therefore, we suggest that altogether, the dynamic between these molecules would be promoting an attenuated growth, which, as a consequence, may be triggering the full-compensation of growth (body weight, condition factor, specific growth rate) in this species, as we have shown previously [28,29] (Fig. 8C).