Alternative Splicing of the Porcine Glycogen Synthase Kinase 3β (GSK-3β) Gene with Differential Expression Patterns and Regulatory Functions

Background Glycogen synthase kinase 3 (GSK3α and GSK3β) are serine/threonine kinases involved in numerous cellular processes and diverse diseases including mood disorders, Alzheimer’s disease, diabetes, and cancer. However, in pigs, the information on GSK3 is very limited. Identification and characterization of pig GSK3 are not only important for pig genetic improvement, but also contribute to the understanding and development of porcine models for human disease prevention and treatment. Methodology Five different isoforms of GSK3β were identified in porcine different tissues, in which three isoforms are novel. These isoforms had differential expression patterns in the fetal and adult of the porcine different tissues. The mRNA expression level of GSK3β isoforms was differentially regulated during the course of the insulin treatment, suggesting that different GSK3β isoforms may have different roles in insulin signaling pathway. Moreover, GSK3β5 had a different role on regulating the glycogen synthase activity, phosphorylation and the expression of porcine GYS1 and GYS2 gene compared to other GSK3β isoforms. Conclusions We are the first to report five different isoforms of GSK3β identified from the porcine different tissues. Splice variants of GSK3β exhibit differential activity towards glycogen synthase. These results provide new insight into roles of the GSK3β on regulating glycogen metabolism.


Introduction
The serine/threonine kinase glycogen synthase kinase 3 (GSK3) was first characterized for its role in glycogen metabolism with phosphorylating and inactivating the enzyme glycogen synthase [1,2]. There are two mammalian GSK3 isoforms encoded by distinct genes: GSK3a and GSK3b [3]. The difference in size is due to a glycine-rich extension at the N-terminus of GSK3a. GSK3 performs an important role in several signalling pathways including IGF-1, Wnt and Hedgehog signal transduction, involved in the regulation of cell fate, embryonic development, protein synthesis, glycogen metabolism, mitosis and apoptosis [4]. In the absence of a Wnt signal, active GSK3 is present in a multiprotein complex that targets b-catenin for degradation via ubiquitinmediated degradation [5,6]. In Wnt stimulated cells, GSK3 phosphorylation of b-catenin is prevented, leading to accumulation of b-catenin and subsequent interaction with TCF/LEF transcription factors to trans-activate target genes [7]. IGF-1/ Insulin may stimulate GS via the linear signaling cascade PI3K/ Akt/GSK3b that leads to the phosphorylation of GSK3 and GSK3 kinase with inhibited activity. The inhibition of GSK3 by insulin results from its phosphorylation at Ser21 in GSK3a and Ser9 in GSK3b and that this is catalyzed by protein kinase B [8]. This inactivation leads to a decrease in phosphorylation of GS resulting in its activation and promotes glycogen synthesis [9].
The actions of GSK3 are also often regulated by the phosphorylation state of its substrates including cytoskeletal proteins, transcription factors and membrane receptors [10,11]. Previous studies have shown that GSK3b has two alternatively splice variants (GSK3b1 and GSK3b2) [12,13], and exhibits differential activity towards specific substrates. Recently studies show that GSK3b1 phosphorylates the microtubule-associated protein tau and MAP1B more effective in vitro than GSK3b2 [14][15][16][17], suggesting that different GSK3b isoforms could alter substrate specificity.
So far, most of the studies on GSK3 have been carried out in humans and mouse, few results have been reported on pigs.
Although mouse has been used as human disease models, the pig is considered as an important experimental animal model of human disease, because pigs and humans share similar anatomical, physiological, and pathological characteristics. The knowledge of porcine GSK3, therefore, will also contribute to the understanding and development of porcine models for human disease (eg: Alzheimer's disease and diabetes) prevention and treatment [18,19].
Here we show that porcine GSK3b is multiple alternatively spliced in different tissues. The mRNA expression level of GSK3b isoforms is differentially regulated during the course of the insulin treatment. To further understand different GSK3b isoforms roles in glycogen metabolism, the expression level of porcine glycogen synthase gene and glycogen synthase activity were detected in pig kidney epithelial cells (PK15). The characterization of porcine GSK3b will undoubtedly help in further understanding its roles in mammals and development of porcine models for human disease prevention and treatment.

Characterization of the Porcine GSK3a and GSK3b Gene
Analysis of the cDNA sequences of the porcine GSK3 revealed the following: (1) The deduced cDNA of porcine GSK3a consists of 2077 bp that contains an ORF of 1452 bp encoding a protein of 483 residues with a calculated molecular mass of 50.9 kDa and an isoelectric point (pI) of 9.04. It contains a 59-untranslated region of 23 bp (59-UTR) and a 39-untranslated region of 602 bp (39-UTR) with a putative polyadenylation signal AATAAA located at 2052 to 2057 bp. (2) The porcine GSK3b cDNA consists of 1517 bp; computer analysis revealed a 1263-bp ORF flanked by a 130-bp 59-UTR and a 124-bp 39-UTR. The porcine GSK3b gene is predicted to encode a polypeptide of 420 amino acids with a molecular mass of 46.8 kDa and a pI of 8.98. The sequences of porcine GSK3a and GSK3b were deposited in GenBank (GenBank accession no. HM214803 and JN387127).
Similar to their orthologous genes in both human and mouse, porcine GSK3a and GSK3b showed the highest homology toward their kinase domains (98% identity). The main structural differences between GSK3a and GSK3b isoforms lie in the Nand C-terminal regions. There is a glycine-rich extension at the Nterminal of GSK3a. Activities of GSK3 are positively regulated by phosphorylation of tyrosine residues 279 and 216 for a and b isoforms, respectively, and negatively regulated by N-terminal serine phosphorylation (residue 21 and 9 for a and b, respectively) [20].

Identification of Multiple Alternative Transcripts of Porcine GSK3b
Previous studies have reported the identification of two forms of GSK3b in human and mouse brain, which in this article are referred to as GSK3b1 and GSK3b2. During the cloning of porcine GSK3b gene, sequencing of individual clones revealed three novel forms of GSK3b mRNA. The sequences for the cDNAs encoding porcine GSK3b1, 2, 3, 4 and 5 were obtained and are deposited in GenBank as JN387128-JN387132, respectively.
These isoforms (GSK3b1, GSK3b2, GSK3b3, GSK3b4 and GSK3b5) encode predicted proteins with 420, 433, 399, 387 and 400 amino acids, respectively (Fig. 1). Interestingly, GSK3b2 contained an additional 39 nucleotides insertion between exons 8 and 9, resulting in a 13 amino acid insert in the kinase domain. We call this minor exon, exon 8 b. This putative exon is surrounded by typical splice consensus sites. GSK3b3 contained an additional 50 nucleotides insertion between exons10 and 11, named exon 10 b. The translational stop codon of GSK3b3 exists in exon10. GSK3b3 contains the whole kinase domain but lacks 21 amino acids at the N-terminus compared to isoform 1. GSK3b4 lacks exon10 which encodes a region outside the catalytic domain, which is poorly conserved between species and between a and b isoforms. GSK3b5 lacks 60 nucleotides in exon6, resulting in a 20 amino acid extraction in the kinase domain (Fig. 2).

Genomic Structure of the Porcine GSK3b Gene
To get more information about the genomic structure of the GSK3b, we searched the pig nucleotide database by BLASTN, and found two overlapped contigs, which encode the GSK3b cDNAs (Fig. 2). The porcine GSK3b gene spans about 221 kb on chromosome 13. After comparing the genomic sequence with mRNA sequences, we found that there were thirteen exons, which were alternatively spliced to generate multiple GSK3b isoforms (Fig. 2). There were eleven exons for transcript of GSK3b1 and GSK3b5, twelve for GSK3b2 and GSK3b3, and ten for GSK3b4. When porcine GSK3b was compared with the GSK3b of human and mouse, an alignment of intron-exon junctions of GSK3b of these species was done, which showed that splicing sites (gt-ag) in all these introns of GSK3b were conserved in mammals. The GSK3b2 used a region of intron 8 sequence as its exon 8 b, and GSK3b3 used a region of intron 10 sequence as its exon 10 b (Fig. 2). The sequence of GSK3b3, GSK3b4 and GSK3b5 has not been cloned in other species.

Differential Expression of GSK3b Isoforms in Fetal and Adult Porcine Tissues
To analyze expression patterns of these alternatively spliced GSK3b isoforms, the cDNAs synthesized from different fetal and adult tissues were analyzed by non-quantitative RT-PCR. As shown in Fig. 3A, five isoforms of the GSK3b expression are detectable in all tissues except for skeletal muscle. Porcine GSK3b4 mRNA was undetectable in adult skeletal muscle. In fetal skeletal muscle, the expression of GSK3b2, GSK3b3, GSK3b4 and GSK3b5 was absent. In order to determine whether three novel GSK3b isforms exists in mouse, we designed primer pairs between exon8 and exon10 for PCR, based on mouse GSK3b cDNA sequence information. After sequence analysis, we found that there was only two kinds of alternative splicing events of GSK3b (GSK3b1 and GSK3b2) in mouse different tissues (Fig. 3B). They are identical to the form reported previously [12,13], and other GSK3b isoforms have not been found.
The expression patterns were further confirmed and quantitated using isoform specific primer pairs (Table S1) to selectively amplify each of the GSK3b isoforms by qRT-PCR (Fig. 4). The mRNA of GSK3b1 was expressed in all twenty tissues examined, with the high level in adult liver, ovary, testis and fetal kidney. GSK3b2 and GSK3b3 have similar expression pattern with the high level in adult testis and fetal liver, whereas expressions in other tissues were relatively weak. The GSK3b4 was abundantly expressed in adult liver, ovary, testis and fetal liver but not in adult heart and skeletal muscle. In adult spleen, kidney and testis, the expression patterns of GSK3b5 were different from those in other isoforms. We also found that the GSK3b5 was abundantly expressed in adult spleen and kidney, whereas expressions in testis were relatively weak. In testis, all transcripts of GSK3b mRNA were at the highest level except for GSK3b5. Interestingly, the mRNA abundance of GSK3b1 of adult liver was significantly higher (p,0.01) than that in embryo. However, the expression level of GSK3b2 was significantly lower (p,0.01) in adult liver than that in fetal liver. Porcine GSK3b5 was most abundant in adult kidney, and undetectable in fetal kidney. The tissue distribution of five GSK3b isoforms raises the possibility that these isoforms may play different functions in porcine various tissues.

Immunohistochemistry
Immunostaining performed on cross-sectional liver, spleen and testis cryosections using antibody directed against human GSK3b protein, porcine liver sections showed strong positive staining within the hepatocyte cytoplasm (Fig. 5A). The GSK3b protein was present in the nucleus of Sertoli cells, spermatogonia and spermatocytes in adult porcine testis (Fig. 5C). In the spleen, although porcine GSK3b can be seen in white pulp, GSK3b immunoreactivity was mainly detected in the red pulp (Fig. 5E). In pig kidney epithelial cells, porcine GSK3b was detected in the nuclei and cytoplasm.

Effect of Insulin on Regulation of Different Porcine GSK3b Isoforms
To determine whether insulin affect the expression patterns of the GSK3b transcript variants, PK15 cells was treated for 0, 5, 10, 15, 20 or 30 min with 100 nM insulin. As shown in Fig. 6A, the porcine total GSK3b mRNA was up-regulated from the 0 min to 10 min after induction, reaching its highest expression at 10 min (p,0.01) and then declined. However, insulin did not significantly affect the total GSK3b protein level (Fig. 6B). In addition, the expression level of different GSK3b isoform is differentially regulated during the course of the insulin treatment (Fig. 6C). GSK3b1, GSK3b2 and GSK3b4 expression level was upregulated from the 0 min to 10 min after induction, reaching its highest expression at 10 min (p,0.01) and then declined. Although the expression patterns of GSK3b1, GSK3b2 and GSK3b4 were similar, the mRNA expression level of GSK3b2 and GSK3b4 was much lower than that of GSK3b1. Moreover, porcine GSK3b3 was up-regulated at an earlier time following insulin treatment, increased to a peak at the 5 min (p,0.01), and then was down-regulated from the 5 min to 30 min after induction. No significant differences were detected in transcription of GSK3b5 during the course of the insulin treatment. These results suggest that GSK3b splicing variants may have different roles in insulin signaling pathway or others signaling pathways in pig through their different gene expressions.

GSK3b5 have a Different Role on Regulating the Glycogen Synthase Gene Expression
To evaluate the effective overexpression of different GSK3b isoforms, whole cell lysates were harvested for mRNA extraction. All four constructs were overexpressed roughly 10-fold ( Fig. S1) after 48 h transfection (with the exception of the GSK3b4, the constuction of pcDNA3.1-GSK3b4 was failed). Moreover, cells were transfected with pEGFP-GSK3b1, fluorescence detection revealed that 80-90% of the cells expressed GSK3b1 (Fig. S1). In order to clarify the role of different GSK3b isoform on glycogen synthase in PK-15 cells, the expression level of porcine GYS1 and GYS2 gene encoding glycogen synthase was detected. We found that GSK3b5 had a different role on the expression level of porcine GYS1 and GYS2 gene. Compared to vehicletransfected cells (Vector), overexpression of GSK3b1, GSK3b2 and GSK3b3 significantly reduced GYS1 (p,0.01) and GYS2 (p,0.05) mRNA expression level (Fig. 7A,B,C). However, the mRNA level of GYS1 and GYS2 was not significant change in cells overexpressing GSK3b5. No significant expression differences among cells overexpressing GSK3b1, GSK3b2 and GSK3b3 were observed.

Regulation of Glycogen Synthase Phosphorylation and Activity Following Over-expression of Different GSK3b Isoforms
To define the activity of GSK3b isoforms in vitro, different GSK3b isoforms were significantly overexpressed in PK-15 cells.   Then we assessed the phosphorylation level of GSK3b in vitro. As shown in Fig. 8A, Increased serine 9 phosphorylation and decreased tyrosine 216 phosphorylation were observed in PK-15 cells transfected with GSK3b1, GSK3b2 or GSK3b3, while overexpression of GSK3b5 caused a decrease in serine 9 phosphorylation levels of GSK3b, respectively.
Since glycogen synthase is a direct substrate of GSK3b, we examined its phosphorylation status and activity in cells expressing different GSK3b isoforms in absence or presence of insulin treatment. As shown in Fig. 8B, overexpression of GSK3b5 did not cause phosphorylation of glycogen synthase in PK-15 cells compared the control cells. Conversely overexpression of GSK3b1, 2 or 3 caused more phosphorylation of glycogen synthase (Fig. 8B). Moreover, the treatment with 100 nM insulin significantly reduced the phosphorylation level of glycogen synthase in PK-15 cells transfected with GSK3b1, GSK3b2 or GSK3b3, demonstrating the reduced activity of GSK 3b. These data show that increase in Ser9 phosphorylation of GSK 3b correlates with inhibition of GSK-3 kinase activity.
We further examined the effect of different GSK3b isoform on glycogen synthase activity by assaying the incorporation of radioactive UDP-glucose into glycogen in the presence or absence  of the allosteric activator glucose 6-phosphate (G6P). As shown in Fig. 7D, overexpression of GSK3b1, GSK3b2, GSK3b3, but not GSK3b5, decreased significantly (p,0.01) glycogen synthase enzyme activity in PK-15 cells.

Subcellular Localization of Porcine GSK3b Isoforms
The cellular locations of GSK3b1 and GSK3b5 isoforms were determined by fluorescence and confocal analysis of PK15 cells transiently transfected with pGFP-GSK3b1 and pGFP-GSK3b5 respectively. After labeling nuclei by staining with DAPI, these GFP fusion proteins were all found to localize both in cytoplasm and nuclei, displayed a nucleocytoplasmic distribution (Fig. 9). There is no different cellular localization among GSK3b1, GSK3b2, GSK3b3 and GSK3b5 isoforms (data not shown).

Discussion
Previous studies in the human and mouse showed that GSK3b consists two different isoforms, GSK3b1 which distributes in many organs and the minor long form (GSK3b2), which has a 13residue insert in the kinase domain, is present in central nervous system [12,13]. In the present study, we observed multiple alternative splicing events of the GSK3b that occurred in porcine different tissues. One kind of the splicing occurs in the kinase domain, which results in a 13 amino acid insert (GSK3b2) and a 20 amino acid extraction (GSK3b5) in the kinase domain. The other contains the identical serine/threonine kinase domain (GSK3b3 and GSK3b4), but vary in their C-terminal (Fig. 1).
Interestingly, GSK3b5 lacks a 20 amino acid (203-222) in the kinase domain inculding three important phosphorylation site Lys 205, Tyr 216, Tyr 220 (Fig. 1). The crystal structure of GSK-3b reveals that the activation loop (residues 200-226) runs along the surface of the substrate binding groove [21]. Phosphorylation of Tyr 216, located on the activation loop, increases the catalytic activity of GSK3b [22]. Previous studies also show that the activities of GSK3 are regulated negatively by serine phosphorylation, but positively by tyrosine phosphorylation. Activities of GSK3 are positively regulated by phosphorylation of tyrosine residues 279 and 216 for a and b isoforms, respectively, and negatively regulated by N-terminal serine phosphorylation (residue 21 and 9 for a and b, respectively) [23,24]. Our results show that GSK3b5 may lack positively regulated activation and plays different functions in regulating the substrate compared other GSK3b isoforms. GSK3b3 and GSK3b4 do not present Ser 389, which has been proposed as an inhibitory domain (Fig. 1). p38 mitogen-activated protein kinase (MAPK) inactivates GSK3b by direct phosphorylation at Ser 389 site and p38 MAPK-mediated phosphorylation of GSK3b at Ser 389 is sufficient to inhibit GSK3b activity [25]. So, GSK3b3 and GSK3b4 may lack negatively regulated activation and p38 MAPK could not inhibit GSK3b activity through Ser 389. However, the physiological significances of these variants are unclear.
Insulin/IGF-1 stimulate glycogen and protein synthesis mainly mediated via the signaling cascade PI(3)K/Akt/GSK3b that leads to the phosphorylation of GSK3 and GSK3 kinase with inhibited activity [26]. This inactivation leads to a decrease in phosphorylation of GS and eIF2B resulting in its activation and promotes glycogen and protein synthesis [27]. In our previous study, insulin can promote phosphorylation of GSK3b and dephosphorylation of GS in differentiated porcine satellite cells, and insulin can promote the of activity GS [28]. In this study, we observed multiple alternative splicing events of the GSK3b, and the mRNA expression level of GSK3b isoforms is differentially regulated during the course of the insulin treatment, suggesting that different GSK3b isoforms may have different roles in insulin signaling pathway.
We found that GSK3b5 have no effect on the mRNA expression level of porcine GYS1 and GYS2 gene. However, overexpression of GSK3b1, GSK3b2 and GSK3b3 significantly reduced the mRNA expression level of GYS1 and GYS2 (Fig. 7). There are numerous transcription factors that have been proposed to be substrates for GSK-3 including CREB [29], NF-kappaB [30], AP-1 [31], b-catenin, c-jun, heat shock factor 1, p53, and Bax [32]. In our previous study, the porcine GYS1 and GYS2 promoter both contains several binding sites for transcription factors of the NF-kappaB (nuclear factor B), and CREB (cAMP responsive element binding) [28]. In addition, GSK-3a and GSK-3b have been shown to have differentially regulated transactivation in causing cAMP-responsive element and NF-kappaBdependent transactivation [33]. So, we infer that GSK3b may When we examined the effect of different GSK3b isoform on glycogen synthase activity, found that overexpression of GSK3b1, GSK3b2, GSK3b3, but not GSK3b5, decreased significantly (p,0.01) glycogen synthase enzyme activity in PK-15 cells. Pervious studied showed that the inhibition of GSK3 promotes the dephosphorylation and activation of glycogen synthase, contributing to the stimulation of glycogen synthesis [34,35]. GSK3 activity is significantly reduced by phosphorylation of Ser9 in GSK3b and Ser21 in GSK3a. In opposition to the inhibitory modulation of GSK3b that occurs by serine phosphorylation, tyrosine phosphorylation of GSK3b increases the enzyme's activity, and its activity is dependent on tyrosine phosphorylation on Tyr-216 [23,24,36]. In the present study, we detected a novel GSK3b isoform, named GSK3b5, which lacks a 20 amino acid (203-222) in the kinase domain including three important phosphorylation site Lys 205, Tyr 216, Tyr 220. So we infer that GSK3b5 lack the GSK3 kinase positively regulated activity and have no effect on glycogen synthase enzyme activity.
The subcellular distribution of GSK3b regulates its actions by controlling its accessibility to substrates, such as many transcription factors in the nucleus [20]. In this study, we found that GSK3b5 have a different role on the mRNA expression level of porcine GYS1 and GYS2 gene. So we further examined the cellular locations of each GSK3b isoform. These GSK3b-GFP fusion proteins were all found to localize both in cytoplasm and nuclei, displayed a nucleocytoplasmic distribution. Previous study revealed that the nuclear distribution of GSK3b is regulated by the NLS locating in 85-103 residues of N-terminal [37]. The five different GSK3b isoforms both contain the nuclear localization sequence (NLS). Nuclear GSK3b levels decrease, when treated by proliferative growth factors [38]. In addition, nuclear export of GSK3b is partially mediated by FRAT-1, which binds GSK3b in the nucleus followed by export of the complex via the nuclear export sequence (NES) in FRAT-1 [39,40]. So, the more precise localizations of them in various conditions require further investigations to help us deeply understand their different biological functions. Determination of the unique properties, possibly protein partners of GSK3b isoforms will likely provide valuable insight into the special actions of their substrates.

Ethics Statement
All research involving animals were conducted according to the regulation (No. 5 proclaim of the Standing Committee of Hubei People's Congress) approved by the Standing Committee of Hubei People's Congress, P. R. China. Sample collection was approved by the ethics committee of Huazhong Agricultural University (No. 30700571 for this study).

RNA Isolation and cDNA Synthesis
Adult Meishan pigs (females, four-month-old, n = 3; males, fourmonth-old, n = 3) and pig embryos [65day post conception (dpc), n = 3] were all obtained from Jingpin Pig Station of Huazhong Agricultural University (Wuhan, China). Porcine adult heart, liver, spleen, lung, kidney, stomach, longissimus dorsi muscle, subcutaneous adipose tissue, ovary, uterus, testis, brain, and embryo heart, liver, spleen, lung, kidney, stomach, longissimus dorsi muscle and brain were freshly collected and then immediately frozen in liquid nitrogen pending RNA extraction. Male mouse (kunmingbai, twoweek-old, n = 3) were purchased from a supplier (Chengdu Center for Disease Prevention and Control, Sichuan, China). Mouse heart, liver, kidney, testis, brain and skeletal muscle were also collected. The RNA was extracted using Trizol reagent (Invitrogen) according to the manufacturer's protocol, treated with RNase-free DNase I (Takara, Japan) to remove contaminating genomic DNA and stored at 280uC. The first strand cDNAs was synthesized using M-MLV reverse transcriptase (Promega, Madison, WI, USA) as described in protocol. The corresponding cDNA was stored at -20uC.

In Silico Cloning of Porcine GSK3a and GSK3b Gene
Human cDNA sequences of GSK3a and GSK3b (GenBank accession no. NM_019884.2 and NM_002093.3) were compared to all sequences available in the pig EST databases using the BLAST algorithm (http://www.ncbi.nlm.nih.gov/blast). We selected the porcine ESTs that shared more than 85% sequence identity to the corresponding human cDNA to assemble the porcine GSK3a and GSK3b gene using the DNA Star program (Madison, WI, USA). Two primer pairs (CDS-AF, CDS-AR for GSK3a and CDS-BF, CDS-BR for GSK3b, Table S1) in both 59 and 39 untranslated region covering the entire coding sequences were designed to PCR amplify coding regions of porcine GSK3a and GSK3b gene.

Detection of Splice Forms of GSK3b
For identification of splice forms of porcine GSK3b, RT-PCR was carried out using porcine adult liver, testis, brain cDNA and porcine fetal liver, testis, brain cDNA using the primer pairs CDS-BF and CDS-BR. The PCR products were separated by 2.0% agarose gel electrophoresis, the all PCR bands were cut out of the agarose gel and purified using a Gel Extraction Kit (Sangon, Shanghai, China). The purified products were then sub-cloned into the pMD-18T vector (Takara, Japan). Plasmids randomly isolated from individual colonies from porcine adult liver, testis, brain cDNA and fetal liver, testis, brain cDNA. The sequences were analyzed by Beijing AuGCT Biotechnology Company.

Reverse Transcription-PCR
Reverse transcription-PCR (non-quantification) was used to amplify individual isoforms of GSK3b from the cDNAs of different adult and fetal tissues of Meishan pigs. PCR cycling conditions were as follows: 95uC initial denaturation for 4 min, 35 cycles of 95uC denaturation for 40 s, 60uC annealing for 40 s, and 72uC extension for 20 s. A final extension was performed at 72uC for 7 min. The PCR fragments were purified and directly sequenced to confirm the correct amplification of the individual isoforms. For the specific amplification of each GSK3b cDNA isoform, five new primer pairs were used (Table S1); GSK3B-V2F/2R in exon8 and exon8 b, GSK3B-V3F/3R in exon10 b and exon11, GSK3B-V4F/4R located at exon8 and the junction of exon9 and exon11, and GSK3B-V5F/5R located at the junction of exon6, exon7 and exon9. Primer pairs GSK3B-V1F/1R located at the junction of exon8, exon9 and the junction of exon10, exon11 to amplify the GSK3b1 and GSK3b5 (Fig. 1). Beta-actin was used as an endogenous reference gene.

Quantitative Real Time RT-PCR Analysis
Real-time RT-PCR was used to quantify the expression level of porcine GSK3b isoforms in different adult and fetal tissues using ABI 7300 real-time PCR thermal cycle instrument (ABI, USA), according to the supplied protocol. Each real-time PCR (in 25 mL) reaction contained 12.5 mL SYBRH Green Real time PCR Master Mixture (contains ROX Dye. Toyobo, Jap), 0.25 mM primers and 1 mL normalized template cDNA. The cycling conditions consisted of an initial, single cycle for 3 min at 95uC followed by 40 cycles of cycling consisting of 20 s at 94uC, 20 s at 60uC, 15 s at 72uC, and final extension for 5 min. The primers were the same as those mentioned in the RT-PCR. The specificity of PCR products were confirmed by melting curve analysis. Pooled cDNA from a subset of the liver samples examined in this study was used to generate the standard curves. In this assay, the efficiency of Beta-actin, GSK3b1, GSK3b2, GSK3b3, GSK3b4 and GSK3b5 gene primers were 96.5%, 97.5% 97.1%, 96.2%, 98.2%and 97.7%, respectively. The amplification efficiencies of control and target genes are approximately equal. Gene expression levels were quantified relative to the expression of Beta-actin using Gene Expression Macro software (ABI, USA) by employing an optimized comparative Ct (2-DDCt) value method. All PCR amplifications were performed in triplicate for each RNA sample.

Cryosection and Immunohistochemistry
Immunohistochemical examination was undertaken on liver, slpeen and testis samples from three male Meishan pigs. Tissues were embedded in OCT medium (Tissue Tek, Miles, Elkhart, IN, USA) and frozen at 220uC, and serial 5-mm sections were cut with a cryostat (Leica, Bensheim, Germany). Slides were incubated using anti-GSK3b (1:200) at 4uC overnight. Secondary antibody was horseradish peroxidase-conjugated goat anti-rabbit IgG (1:2000) and was incubated for 30 min. The SABC and DAB visualization methods were used according to the manufacturer's instructions (Boster Company, China). The sections were counterstained with hematoxylin for 10 min. Negative controls were done by omitting the primary antibody or by using an irrelevant primary antibody of the same isotype. Stained muscle crosssections were viewed with an Olympus BX-50F light microscope (Olympus Optical, Tokyo, Japan).

Time-course Effect of Insulin in PK-15 Cells
PK15 cells were plated in 6-well plates at a concentration of 2.5610 5 cell/well (2 mL/well) and cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented 10% (v/v) fetal bovine serum under humidified air containing 5% CO2 at 37uC. To evaluate the time-dependent changes in the levels of each GSK3b isoform, when 70%-80% confluence was observed, PK-15 cells were serum starved for 6 h, then 100 nM bovine insulin (Sigma, Cat. Nr: I0516) was added for 0, 5, 10, 15, 20 or 30 min. Total RNA and protein was extracted for determining the expression of different GSK3b isoform by qRT-PCR and western blotting, respectively.

Plasmid Construction and Transfection
To construct porcine different GSK3b isoform recombinant plasmids for expression in mammalian cells, the open reading frames coding for different porcine GSK3b isoforms were amplified from cDNAs of porcine liver using primer pair (GSK3B-GFPF, GSK3B-GFPR, Table1). The PCR products were first cloned into a pGEMT easy vector (Promega), then digested with KpnI and XhoI enzymes for sub-cloning into the pcDNA3.1(+) and pEGFP-N1 vector to generate the pcDNA3.1-GSK3b and pEGFP-GSK3b plasmids. Correct orientation and reading frame were confirmed by sequencing.
PK15 cells were cultured as mentioned above. Transient transfection was performed using lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions, and then cultured in media an additional 48 h before mRNA extraction or assay of glucose uptake. For detecting the cellular locations of each GSK3b isoform, At 24 h after transfection, cells were washed three times with phosphate-buffered saline (PBS), and then fixed in prewarmed growth medium containing 4% formaldehyde for 15 min at 37uC. After the final washing steps and incubation with 10 mM DAPI for 10 min, the slides were mounted and sealed, and analyzed by confocal microscopy (TCS-SP2). Leica confocal software (Leica IM500) was used to generate images of individual fluorescent markers as well as overlay pictures to demonstrate the relative distribution of the fusion protein.

Western Blotting
Total protein was extracted in preparation buffer [7 M urea, 2 M thiourea, 4% CHAPS, 1% DTT, 2% IPG Buffer pH3-10, 10 ml proteinase inhibitor cocktail (BBI, Kitchener, Canada)]. Then incubated for 30 min at room temperature with occasional vortex, and centrifuged at 20,000 g for 15 min at 4. The supernatant was collected and stored at -80 until analysis. Protein concentrations were determined using the Bradford protein assay. 20 mg of total protein extract from PK-15 cells was separated by 10% SDS-PAGE and subsequently electro transferred to PVDF membrane, and incubated with antibodies to phospho-Ser9-GSK3b (Santa Cruz, CA) and phospho-Tyr216-GSK3b (Santa Cruz, CA) at a 1:400 dilution; GSK3b (Cell signaling, Danvers, MA) and phospho-Ser641-GS (Cell signaling, Danvers, MA) at a 1:1,000 dilution at 4C overnight. Immunoblots were developed using horseradish peroxidase-conjugated goat anti-rabbit IgG (Santa Cruz, CA) at a 1:5,000 dilution, followed by detection with enhanced chemiluminescence.

Glycogen Synthase Assay
PK-15 cells were transfected with different pcDNA3.1-GSK3b plasmids for 24 h then scraped into 600 ml of homogenization buffer (10 mM Tris-HCl (pH 7.8), 150 mM NaF, 15 mM EDTA, 60 mM sucrose, 10 mg/ml leupeptin, and 50 mM sucrose, 1 mM PMSF). Samples were centrifuged at 4uC and 12000 g for 10 min and the supernatant was saved for GS assay. GS activity was assayed by measuring the incorporation of glucose from UDP-[U-14 C] glucose into glycogen [28]. Briefly, assay buffer (50 mM Tris-HCI, pH 7.8, 20 mM EDTA, 25 mM KF, 1% glycogen, 200 UDP-[U-14C] glucose (4.5 mCi/umol) was added to 45 mL cell lysate in the presence and absence of 20 mM glucose-6-phosphate. After a 30 min incubation at 37uC the reaction was stopped by spotting the reaction mix onto Whatman No.5 filter paper and washed three times in 66% (v/v) ethanol for 20 min. Filters were washed in acetone for 5 min and dried before being counted in a liquid scintillation counter. Glycogen synthase activity was expressed as a ratio of activity in the absence divided by that in the presence of its allosteric activator, glucose-6-phosphate. Figure S1 The effect of GSK3b isoforms overexpression. PK-15 cells were transfected with pcDNA3.1-GSK3b1, GSK3b2, GSK3b3, GSK3b5, respectively. Cells were harvested for mRNA extraction after 48 hours transfection. (A) The effective overexpression of different GSK3b isoforms, was identified by Semiquantitative RT-PCR and Beta-actin was used as a control. (B) PK-15 cells were transfected with pEGFP-GSK3b1. Cells were fixed and analyzed by immunofluorescence after 48 hours transfection. (TIF)