Daily rhythmicity of clock gene transcript levels in fast and slow muscle fibers from Chinese perch (Siniperca chuatsi)

Clock genes are considered to be the molecular core of biological clock in vertebrates and they are directly involved in the regulation of daily rhythms in vertebrate tissues such as skeletal muscles. Fish myotomes are composed of anatomically segregated fast and slow muscle fibers that possess different metabolic and contractile properties. To date, there is no report on the characterization of the circadian clock system components of slow muscles in fish. In the present study, the molecular clock components (clock, arntl1/2, cry1/2/3, cry-dash, npas2, nr1d1/2, per1/2/3, rorα and tim genes) and their daily transcription levels were characterized in slow and fast muscles of Chinese perch (Siniperca chuatsi). Among the 15 clock genes, nrld2 and per3 had no daily rhythmicity in slow muscles, and cry2/3 and tim displayed no daily rhythmicity in fast muscles of the adult fish. In the slow muscles, the highest expression of the most clock paralogs occurred at the dark period except arntl1, nr1d1, nr1d2 and tim. With the exception of nr1d2 and tim, the other clock genes had an acrophase at the light period in fast muscles. The circadian expression of the myogenic regulatory factors (mrf4 and myf5), mstn and pnca showed either a positive or a negative correlation with the transcription pattern of the clock genes in both types of muscles. It was the first report to unravel the molecular clock components of the slow and fast muscles in vertebrates. The expressional pattern differences of the clock genes between the two types of muscle fibers suggest that the clock system may play key roles on muscle type-specific tissue maintenance and function.


Background
Fish skeletal muscles are the most abundant tissue in the body mass and play an important role in the process of certain physiological metabolism [1]. Similar to other peripheral tissues of the body, skeletal muscles have circadian rhythms [2]. These rhythms are regulated by a transcriptional-translational and post-translational feedback network termed as the molecular clock [3]. There are several major components in the molecular clock, including circadian locomotor output cycles kaput (clock), aryl hydrocarbon receptor nuclear translocatorlike protein 1 (arntl1 or bmal1), cryptochrome (cry), and period protein (per) etc. [4]. Two transcriptional activation proteins of the molecular clock, namely clock and arntl1, are the basic-helix-loop-helix (bHLH) transcription factors that form into a heterodimer in the nucleus. Together, they transactivate per and cry gene expression via binding to the E-box elements (CACGTG) at their promoter sequences [5][6][7]. Per and cry then translocate into the cell nucleus, in which they inactivate clock and arntl activity, thereby repressing their own transcription. The clock mechanism plays a pivotal role in myogenesis, gene transcription, and maintenance of muscle metabolism [8,9]. The molecular clock components have been identified in skeletal muscles and showed a circadian rhythms of expression. In addition to these clock genes, more than 2300 other genes have circadian pattern of expression in skeletal muscles. These muscle genes with circadian pattern of expression are believed to be regulated by the major molecular clock genes and the clock-controlled transcription factors [10]. Fish are excellent model species for investigating the regulation of skeletal muscle physiology in vertebrates because it has several structural features making convenient to experimental analysis [11,12]. The slow-contracting red and fast-contracting white muscles are the two main muscle fiber types in fish [13,14]. Particularly, they are localized into physically distinct area of the fish body. The fast muscle fibers are the main component of skeletal muscles that distribute along the spine of the whole body. Their explosive force is strong for fast swimming using energy from glycolysis [15,16]. Slow fibers, on the other hand, contain high contents of mitochondria and their metabolism is completely aerobic [16,17]. These unique features enable slow fibers to maintain sustained swimming and support oxygen respiration [18].
Recently, several studies have been reported on the clock rhythmicity in fish. In zebrafish, the major clock genes showed similar circadian expression patterns in fast muscles compared with the central organs, such as retina and brain [19,20]. It has been shown that insulinlike growth factor binding proteins (igfbp3 and igfbp5b) and myogenic regulatory factor 4(mrf4) were controlled by the core clock genes in zebrafish skeletal muscles [20,21]. In Atlantic cod fast muscles, similar circadian clock system has been identified [2]. Another myogenic regulatory factor myf5 exhibited a significant correlation with the core clock genes at the transcription levels [2]. However there is little information on the circadian clock system components in fish slow muscle. The Chinese perch (Siniperca chuatsi) is one of the most important species in aquaculture in China [22]. Its high nutritional value, high protein content and appealing taste have led to its expanded large-scale aquaculture in china [23]. In the present study, we report the characterization of circadian clock system in both fast and slow muscles of Chinese perch, and the correlation analysis between core clock gene expression and 11 myogenic related genes in the two types of muscles.
The conserved structural and functional domains were characterized based on the predicted protein sequences. The clock protein contains the conserved bHLH, Per-Arnt-Ser (PAS) A and PAS B domains. The amino acid sequences of these domains showed 100%, 98% and 100% similarities to those of Larimichthys crocea, respectively ( Figs. 1 and 2). The per1 has a PAS domain, a G-box binding protein multifunctional mosaic region (MFMR) and a 2/3 C-terminal region of period protein.
The amino acid sequences of these domains showed 99%, 91% and 93% similarities to those of L. crocea, respectively. The cry1 contains two functional domains, a flavin adenine dinucleotide (FAD)-binding domain and a DNA photolyase domain. The amino acid sequences of The predicted domain structure of clock, cry1, per1 and nr1d2 are shown as above and presented below the sequence structure. The split domain of clock proteins is shown in different colors. Both clock and per1 proteins have a PAS domain. cry1 includes a photosensitive domain of DNA photolyase, and nr1d2 has a unique zinc finger structure that connects with the nuclear DNA these two domains showed 98% and 98% similarities to those of L. crocea, respectively. The nr1d2 has two nuclear hormone receptors DNA-binding domains and the amino acid sequences of these domains showed 100% and 98% similarities to those of L. crocea, respectively.
The rhythmicity of clock genes during a daily cycle in fast and slow muscles The expression pattern of the clock genes was determined for daily rhythmicity in fast and slow muscle (Tables 1 and 2, and Figs. 3 and 4). Among the 15 clock genes, nr1d2 and per3 have no daily rhythmicity in slow muscles, while cry3, npas2, and tim have no rhythmicity in fast muscle. Cry2 and cry-dash have no daily rhythmicity in neither slow nor fast muscles. In contrast, arntl1, arntl2, clock, cry1, per1, per2, nr1d1 and rorα displayed the daily rhythmicity in both fast and slow skeletal muscles. In fast muscles, transcriptional activation factors arntl1, arntl2 and clock displayed the daily rhythmicity with an acrophase during the light phase (Fig. 3). There was no much temporal difference between arntl2 and clock expression. In slow skeletal muscles, arntl1, arntl2 and clock still displayed the daily rhythmicity but arntl1 had an acrophase during the dark phase (Fig. 4). The temporal expression of arntl1 and clock showed no apparent difference. Finally, npas2 showed the daily rhythmicity with an acrophase during the light phase only in slow muscles (Fig. 4).
The transcriptional repressors cry2, cry3 and tim were arrhythmic but cry1, per1, per2 and per3 were rhythmic in fast muscles. With the exception of tim, the other clock genes were expressed with the acrophase during the light phase in fast muscles. In slow muscles, the cry2 and per3 were arrhythmic but cry1, cry3, per1, per2 and tim were rhythmic. The tim was highly expressed during the light phase with the acrophase at Zeitgeber time (ZT) = 0.16 h, but the other genes were expressed with the acrophase during the dark phase in slow muscles. The nuclear receptors rorα, nr1d1 and nr1d2 displayed the daily rhythmicity in fast muscles. rorα and nr1d1 had an acrophase during the light phase (ZT = 5.05 h) but nr1d2 had an acrophase during the dark phase (ZT = 20.51 h). In slow muscles, rorα and nr1d1 exhibited a daily rhythmic expression but nr1d2 was arrhythmic. The rorα and nr1d1 genes had an acrophase during the light phase, but nr1d2 exhibited a similar process during the dark phase (ZT = 19.08 h).
In fast muscles, the mRNA transcript levels of clock were positively correlated with the expression pattern of per2 and per3 in fast muscles with a higher correlation index (r > 0.8; Table 3). Arntl2 also exhibited a positive correlation with per2 and per3. Nr1d2, however, showed a moderate negative correlation to arntl1, clock, npas2, nr1d1and rorα. In slow muscles,clock showed a moderate positive correlation to per2 and per3. Arntl2 also displayed a moderate positive correlation with per2 expression. Nr1d2 showed no correlation with other genes (Table 4).

Daily expression of myogenic related genes and the correlation with the clock components in fast and slow muscles
In fast muscles, mbnl1, mrf4, mstn, myf5, myoD, myoG and pcna displayed a daily rhythmic expression (Table 1 and Fig. 5), but only pcna had an acrophase during the dark phase (ZT = 22.91 h). In slow muscles, foxk2, mrf4, mstn, myf5 and pcna exhibited a daily rhythmic expression ( Table 2   Note: Expression levels of clock genes and muscle-related genes are highlighted in bold while they displayed daily rhythmicity. The P value is defined as the noise/signal ratio of the oscillation amplitude. Daily rhythmicity is indicated when P value is less than 0.3 Note: Expression levels of clock genes and muscle-related genes are highlighted in bold while they displayed daily rhythmicity. The P value is defined as the noise/signal ratio of the oscillation amplitude. Daily rhythmicity is indicated when P value is less than 0.3 The mRNA transcript levels of most myogenic related genes had either a positive or a negative correlation with the daily expression of clock genes in fast and slow muscles (Tables 5 and 6). In fast muscles, the transcript levels of pcna displayed a moderate negative correlation with arntl2, cry1 and per1 (−0.8 < r < −0.5). Mbnl1 showed a strong positive correlation with the daily expression of arntl2, cry1, nr1d1, per2, per3 and clock (r > 0.8). MyoG was also positively correlated with arntl2 and per2. In slow muscles, the transcript levels of pcna showed a moderate positive correlation with arntl2, cry1 and per1. myf5, mrf4, pcna and mstn showed a strong positive correlation with cry3 and npas2 in slow muscles. The foxk2 gene also showed a strong correlation with transcriptional activators, such as arm arntl2.

Discussion
In the present study, we identified for the first time 15 clock genes, including four transcriptional activation factors (arntl1, arntl2, clock and npas2), eight transcriptional repressors (cry1, cry2, cry3, cry-dash, per1, per2, per3 and tim), and three orphan nuclear receptors (nr1d1, nr1d2 and rorα) in Chinese perch skeletal muscles. As expected, some of the clock genes exhibited a robust oscillation during the light-dark cycle in the slow or fast skeletal muscle.
The full length cDNA sequences of several key clock genes, such as clock, cry1, per1 and nr1d2, were cloned from Chinese perch. Silico structural analysis of the deduced amino acid sequences indicated that clock and per1 proteins contain the conserved PAS and bHLH domains [24]. These two domains are required for the circadian clock functions and are highly conserved in different species during evolution [25]. The transcriptional repressor cry1 has a typical DNA-photolyase and flavin adenine dinucleotide(FAD)-binding domains that are present in all cryptochrome genes [26,27]. Together with photolyase DNA repair enzymes, FAD-binding domain containing proteins form the cryptochrome/ photolyase complex. This active complex has been used in blue light-induced gene expression to affect biological rhythm [26,28,29]. In nr1d2 protein, two core nuclear domains were also identified from its deduced amino acid sequence. The nr1d2 binds to arntl1 and clock via a unique zinc finger structure domain to form a complex, which binds and blocks the protein complex formation by Circadian Locomotor Cycles [30,31]. The daily rhythmicity was observed in many clock genes in fast muscles of several species, indicating a potential regulatory function in muscle physiology and metabolism ( [2,20,32], and [33]). However, the circadian clock gene expression in slow muscles has not been reported. In the study, we analyzed the arntl1 and arntl2 gene expression in Chinese perch and showed that they exhibited a light-biased expression in fast muscles. Several earlier reports in rainbow trout and mouse suggested that the photosensitivity profile in fast muscles is regulated by their homologous expression in either central or peripheral clocks ( [10,34,35]). However, their dark-biased expression in Chinese perch slow muscle is similar to that observed in Atlantic cod and zebrafish fast muscles. The correlation between arntl1/2 and clock/npas2 in Chinese perch slow muscles indicated that the mechanism underlying the transcriptional activation of the clock system may be similar to those reported from other fish species in slow muscles. Npas2 which shares a high sequence homology with clock protein is able to substitute for clock function in the master brain clock and regulates the circadian rhythmicity in the brain. The antiphase profile between arntl and clock in Chinese perch fast muscles suggested that arntl may correlate with other bHLH-PAS factors in fish muscles.
Among the 3 period genes expressed in Chinese perch slow muscles, their dark-biased expression was in agreement with that reported from other fish species, such as goldfish, European sea bass and zebrafish [20,28,36,37]. On the other hand, their light-biased expression in Chinese perch fast muscle is also consistent with the central and peripheral clocks in Senegalese sole [24]. The three period genes with the daily rhythmic expression also displayed a positive correlation with the cry1 gene, the only cryptochrome gene with the daily rhythmicity in Chinese perch fast muscle. Of the period and cry genes with the daily rhythmic expression in Chinese perch slow muscles, per1/2 showed the positive correlation with both cry1 and cry3. These results suggested that per may interact with cry to control the transcriptional activation and function in the circadian feedback loop.
Nr1d1, nr1d2 and rorα are members of nuclear receptor family, which are involved in stabilizing the circadian  [20,[38][39][40]. Nr1d1 and nr1d2 were identified as the constitutive transcriptional repressors of arntl1, whereas rorα is the arntl1 transcriptional activator [2,41]. In addition, nr1d1 was considered to be interwoven into the core clock mechanism via downregulating the clock expression. As reported in mammals, Npas2 expression was repressed by these nuclear receptors [40,42]. In this study, the three nuclear receptors displayed a daily rhythmic expression in Chinese perch fast muscles and nr1d2 exhibited a negative correlation with the expression of rorα gene. This is reflected in their tight regulation with the circadian mechanism. Nr1d2 also showed a negative correlation with clock and npas2. Therefore, it is possible that nrld2 may function by repressing npas2 and clock expression in fast muscles. However, nr1d2 had no daily rhythmic expression in Chinese perch slow muscles and nr1d1 showed a direct relationship with the rorα expression. It could suggest that Chinese perch slow muscles may have a different circadian mechanism in maintaining the stabilization loop.
In this study, the transcription levels of eleven genes related to myogenesis during the daily cycle were investigated. The data revealed that seven genes had a daily rhythmic expression in fast muscles and five in slow muscles. Myogenic regulatory factors(MRFs) (such as myoD, myf5, mrf4 and myoG), belong to the same class of helix-loop-helix transcription factors that play distinct and overlapping roles in regulating muscle development and growth [43,44]. It has been reported that the circadian regulation of myoD expression by lock/arntl was crucial for the skeletal muscle phenotype and function in mouse [32]. Our study confirmed that myoD in Chinese perch fast muscles exhibited a typical daily rhythmicity. Based on our observation, it is possible that the  circadian regulation of Chinese perch myoD may function in a similar way as reported in mouse. In contrast, we have not obtained any direct evidence for circadian expression of myoD in Chinese perch slow muscles. Therefore, it is possible that the differentially expressed MRFs may result from the lineage-specific differences by clock genes. However, our work demonstrated that one of the MRFs, mrf4, in Chinese perch slow muscles exhibited a rhythmic expression pattern similar to that described for myoD in fast muscles, suggesting that mrf4 may have a potential function in the maintenance of muscle phenotype and function.

Conclusion
In this study, we assayed the possible correlation of a functional clock system in Chinese perch slow and fast skeletal muscles. We demonstrated that 10 clock genes and 7 genes related to myogenesis exhibited the daily rhythmicity in fast muscles of Chinese perch. The 11 clock genes and 5 genes related to myogenesis have the daily rhythmicity in slow muscles (Fig. 7). The circadian expression of mrf4, myf5, mstn, and pcna may either positively or negatively regulate the transcription of the clock genes in both types of muscles. It is plausible that muscle type-specific maintenance and function is regulated by the core clock genes. This is based on the evidence of daily rhythmicity and apparent correlation of gene expression of clock genes and genes related to myogenesis. Taken together, our data provide new information on the rhythmic expression of clock genes and a better understanding of the circadian clocks in fish muscle phenotype maintaining and function.

Daily rhythm experiment
Adult Chinese perch (body weight 450 ± 10 g) were stocked in 250 m 3 tanks. Fifteen testing fishes were kept in each tank and a total of six tanks were used in the experiment. To liminate the fish disturbance during sampling, the fish were fed at the same time each day. During the experiment, water temperature was kept at 25 ± 0.8°C, dissolved oxygen at 85 ± 2% and the light intensity of the water surface at 0.84 W*m −2 (200 lx). The testing fishes were acclimated to the above described conditions for 2 weeks during a daily lightdark cycle before sample collection. Six individuals were randomly collected from the each tank in every three hours until 24 h (Zeitgeber time: ZT0, 3, 6, 9, 12, 15, 18, 21 and 24). Sampling of the light treatment at the different time points was basically followed as described by Lazada [2]. Briefly, samples at ZT0 were collected when the light reached to a maximum intensity, the samples at ZT24 were collected when it transited to the light phase; The ZT12 samples were collected at approximately Note: Only correlations with r > +0.5 or r < −0.5 and including at least one gene with significant daily rhythmicity are shown in this table.
The following values were set to define the degree of correlation: data are moderately correlated if 0.5 < r < 0.79 and there is a strong correlation when r ≥ 0.80 which are highlighted in bold 20 min later, while samples of ZT0-9 and ZT15-24 were collected at the time between the light and dark. The fast muscle was sampled from the dorsal muscle tissue, and slow muscle was collected under the body skin [16]. All of the collected muscle samples were washed with cold and sterilized 1 × PBS to remove contaminating blood and then immediately stored in liquid nitrogen at −80°C for total RNA extraction.

Molecular cloning of Chinese perch clock genes
To amplify the cDNA fragments of the clock genes, the primers were designed based on the assembled EST sequences of the Chinese perch muscle database (accession nos: SRX1738860) or the clock gene sequences from several-related fish species (Table 1). The full length cDNAs of the four clock genes were amplified with the SMART RACE cDNA Amplification Kit Table 6 Correlation of expression levels of rhythmic clock and muscle-related genes in the slow muscle Note: The following values were set to define the degree of correlation: data are considered to be moderately correlated if 0.5 < r < 0.79 or −0.79 < r < −0.5 and there is a strong correlation when r ≥ 0.80. And moderate correlation is marked in bold and strong correlation in red color Table 5 Correlation of expression levels of rhythmic clock and muscle-related genes in the fast muscle Note: The following values were set to define the degree of correlation: data are considered to be moderately correlated if 0.5 < r < 0.79 or −0.79 < r < −0.5 and there is a strong correlation when r ≥ 0.80. And moderate correlation is marked in bold and strong correlation in red color according to the manufacture's instruction (Clontech, Palo Alto, CA, USA). Specific nested PCR primers were designed based on the cloned partial sequences (Table 1). For 5′ RACE amplification, the protocols were conducted as follows: 35 cycles of 94°C for 30 s, 60°C for 30 s and 72°C for 1 min. For 3′-RACE, two amplifications were conducted at the same parameters as 5′ RACE the amplification described above.

Bioinformatic analysis of the clock genes
The cDNA fragment identity was confirmed with BLAST (http://blast.ncbi.nlm.nih.gov) and the deduced amino acid sequences were obtained using the ExPASy Proteomics Server (http://www.expasy.ch). The complete amino acid sequences of the clock, cry1, per1 and nr1d2 domains were analyzed through UniProt (http://www.uniprot.org) and the ExPASy proteomics Server (http://www.expasy.ch).
The clock gene sequences of other teleosts were obtained from the NCBI databases and the protein accession numbers are listed in Additional file 2.

Gene expression analysis
Total RNAs were isolated from Chinese perch muscles using the TRIzol R Reagent (Invitrogen, USA Primers for the qRT-PCR assays were designed using the software Primer 5.0 ( Table 7). The reverse transcribed cDNAs from skeletal muscles were used as templates for qRT-PCR assays with SYBR Green PCR reaction kit (Stratagene, Shanghai, China). The qRT-PCR amplification reaction was carried out using the Stratagene Mx3005 system (Stratagene, CA, USA). A total volume of 25 μL reaction was used for the qRT-PCR assays, containing 1 μL cDNA templates, 12.5 μL SYBR Green mix, and 1 μmol/L each of forward and reverse primers. The reaction protocol was used as the standard cycling of the qPCR. Each product identity was verified by dideoxy-mediated chain termination sequencing at Sangon Biotechnology Inc. (Shanghai, China). The relative expression ratio (R) of target mRNA was calculated by R = 2 -ΔΔCt [45,46], where Ct is the cycle threshold. The basic equation employed was ΔΔCt = (Ct target gene − Ct housekeeping gene ) experiment − (Ct target gene − Ct housekeeping gene ) control . The treanscriptional levels of selected muscle-related genes in Chinese perch fast and slow muscles during a daily cycle were also quantitatively assayed using the same qRT-PCR protocol.
The stability of transcription of reference genes was assayed with GeNorm system and total six reference genes were analyzed including Glyceraldehyde-3phosphate dehydrogenase (GAPDH), β-actin, 18 S rRNA gene, hypoxanthine phosphoribosyltransferase 1-like (hprt1), epinephelus coioides ribosomal protein S29 (rps29) and ribosomal protein L13 (rpl13) [47][48][49]. GeNorm analysis revealed that rpl13 was the most stable control gene in skeletal muscles (geNorm stability value M = 0.28). The geometric average of these genes were measured by absolute quantification within all samples Fig. 7 Molecular components of the clock system identified in fast and slow skeletal muscles of Chinese Perch and myogenic genes with daily rhythmic expression. The green and red oval represent genes involved in the peripheral clock components in Chinese Perch skeletal muscles, which is related with the muscle fiber. They comprise members of the transcriptional activator arm (in red: arntl1, arntl2, clock and npas2), transcriptional repressor arm (in green: cry1, cry2, cry3, cry-dash, per1, per2, per3 and tim) and nuclei (in nucleus: nr1d1, nr1d2 and rorα). The daily rhythmicity of mbnl1, mrf4, mstn, myf5, myoD, myoG and pcna play a crucial role in fast muscle specification, and foxk2, mrf4, mstn, myf5 and pcna play an important role in terminal slow muscle differentiation

Statistical analysis
The transcriptional expression levels of the clock and muscle-related genes at each time points were analyzed with the Sigma plot and then were calculated by one-way ANOVA procedures using SPSS software. To compare the difference between the control and experimental groups, Duncan's multiple range tests were used for the analysis. The differences were considered to be statistically significant when the P value was less than 0.05. Data are shown as means ± SE (n = 6). The daily rhythmicity in relation to the expression of the clock and muscle-related genes was assayed with Matlab 7.0 followed as described by earlier studies [2,37]. To perform a COSINOR analysis, the formula ƒ(t) = M + Acos(t/pi/12 -φ) was used, where ƒ(t) stands for the gene expression level at a given time, mesor (M) stands for the mean value, A stands for is the oscillation amplitude, t is the time in hours and φ stands for the acrophase. The P value was defined by the noise/signal of the amplitude (SE(A)/A) and if P Value <0.3, the expression levels could considered to display daily rhythmicity [2,37]. Correlation between the clock and muscle-related gene expression was assayed with Pearson's correlation test (r).