Breed-Dependent Transcriptional Regulation of 5′-Untranslated GR (NR3C1) Exon 1 mRNA Variants in the Liver of Newborn Piglets

Glucocorticoids are vital for life and regulate an array of physiological functions by binding to the ubiquitously expressed glucocorticoid receptor (GR, also known as NR3C1). Previous studies demonstrate striking breed differences in plasma cortisol levels in pigs. However, investigation into the breed-dependent GR transcriptional regulation is hampered by lacking porcine GR promoter information. In this study, we sequenced 5.3 kb upstream of the translation start codon of the porcine GR gene, and identified seven alternative 5′-untranslated exons 1–4, 1–5, 1–6, 1–7, 1–8, 1–9,10 and 1–11. Among all these mRNA variants, exons 1–4 and 1–5, as well as the total GR were expressed significantly (P<0.05) higher in the liver of newborn piglets of Large White (LW) compared with Erhualian, a Chinese indigenous breed. Overall level of CpG methylation in the region flanking exons 1–4 and 1–5 did not show breed difference. However, nuclear content of Sp1, p-CREB and GR in the liver was significantly (P<0.05) higher in LW piglets, associated with enhanced binding of p-CREB, and higher level of histone H3 acetylation in 1–4 and 1–5 promoters. In contrast, GR binding to promoters of exons 1–4 and 1–5 was significantly diminished in LW piglets, implicating the presence of negative GREs. These results indicate that the difference in the hepatic expression of GR transcript variants between two breeds of pigs is determined, at least partly, by the disparity in the binding of transcription factors and the enrichment of histone H3 acetylation to the promoters.


Introduction
Glucocorticoids (GCs) are involved in the regulation of almost all the biological processes ranging from growth and development to immunity and reproduction [1,2]. The biological functions of GCs require the functionality of their intracellular binding protein, glucocorticoid receptor (GR) which is also known as nuclear receptor subfamily 3, group C, member 1 (NR3C1). GR is expressed in almost all the cells studied. The intracellular concentration of GR protein varies greatly among cell types and determines, in large part, the cellular responses to the hormone. Therefore, processes that regulate the expression of GR gene are critical to the maintenance of appropriate functions of glucocorticoids.
The GR gene spans more than 80 kb and is conserved among humans, rats and mice [3,4]. Human GR contains nine untranslated alternative first exons and eight coding exons (exons 2-9) [3]. Since the alternative first exons are each preceded by their own promoter, the tissue-specific usage of these alternative promoters results in different GR mRNA transcripts [4]. The 59-heterogeneity of GR mRNA transcripts and promoter usage represent complex mechanisms for the regulation of GR transcription in different tissues under different conditions. Among the nine untranslated alternative first exons, seven are located in the proximal promoter region approximately 5 kb upstream of exon 2. The proximal GR promoter comprises of a CpG island which shows high sequence homology between rats and humans [5].
A wide variety of transcription factors have been identified to bind to their cis-acting elements or binding sites on the CpG island of GR promoter to regulate GR transcription, among which are specificity protein 1 (Sp1) [6], YinYang 1 (YY1) [7], cAMP response element binding protein (CREB) [8], Nerve growth factor-inducible protein A (NGFI-A) [9] and activator protien-1 (AP-1) [10]. GR, a transcription factor itself, works together with other transcriptional factors to fine-tune its own transcription. The auto-regulation of GR transcription can be stimulatory or inhibitory, depending on the nature of the glucocorticoid response elements (GREs). Several positive GREs or negative GREs (nGREs) have been identified in CpG island promoters of human, rats or mice GR genes [11,12].
Epigenetic mechanisms, such as DNA methylation and histone modification, act in concert with transcription factors to regulate GR transcription. For instance, low maternal care increases the CpG methylation of the NGFI-A binding site in the GR promoter 1-7, which is associated with the inhibition of histone acetylation and NGFI-A binding to GR 1-7 promoter in the hippocampus of offspring rats. These epigenetic alterations subsequently cause lower GR expression in the hippocampus and higher stress responses in the offspring of low caring mothers [13]. Epigenetic programming of GR is also reported in peripheral tissues. Maternal dietary protein restriction during pregnancy leads to decreased methylation of GR 1-10 promoter and increased GR expression in the liver of rat offspring [14].
Our knowledge on GR gene and its transcriptional regulation is mostly based on studies in humans, mice and rats [3]. Pigs serve as excellent model for human metabolic research because they share high similarity to humans in anatomy, physiology, development, metabolism, and omnivorous habits [15]. Previous studies demonstrate striking breed differences in plasma cortisol concentration and GR mRNA expression in pigs [16,17], Chinese indigenous Erhualian (EHL) pigs demonstrate 2-fold higher plasma cortisol concentration than Pietrain pigs [18]. Moreover, GR expression in hippocampus [17], liver and muscle [16] also differs between Chinese and Western pig breeds, which is associated with the breed-specific characteristics in stress responses, hepatic gluconeogenesis, and intramuscular fat deposition, respectively. Accumulating evidence suggests that dysfunction of hepatic GR is involved in the development of metabolic diseases including obesity, diabetes and fatty liver [19,20]. Studies on the regulation of hepatic GR expression in the pig may shed light on the pathogenesis of GR related metabolic disorders in human liver. However, investigation into GR transcriptional regulation in the pig is hampered by lacking genomic information on porcine GR promoters. Therefore, the present study was aimed first, to clone and sequence the porcine GR promoters; second, to compare the hepatic expression of 59-untranslated GR first exon mRNA variants between two pig breeds differing in plasma cortisol concentrations and; third, to investigate the breed-dependant transcriptional regulation of GR in porcine liver.

Results
Porcine GR gene promoter demonstrated higher homology to human than the rat We first compared the published human (AY436590), rat (AJ271870) and mouse (X66367) GR promoter sequences and identified the highly conserved regions. Using these highly conserved sequences and the porcine GR cDNA sequence (AY779185.1) as probes, we screened the porcine BAC library and hit a positive clone (CH242-105G5). The BAC clone was then priority sequenced by the Sanger Institute upon our request. Within the complete sequence (CU928713) of the clone, we verified, by sequencing the overlapping PCR products amplified from porcine genomic DNA extracted from liver, a 5300 bp 59 flanking sequence of porcine GR exon 2 (Fig. 1). Using primers shown in Table 1, we identified seven untranslated alternative exons (1-4, 1-5, 1-6, 1-7, 1-8, 1-9,10, and 1-11) with RT-PCR in porcine total RNA extracted from various types of tissues.
The 39 boundaries of the seven alternative first exons were determined ( Fig. 2A) and the sequences were aligned with the relevant regions of rat (AJ271870) [21], and human (AY436590) [4] for homology analysis. Compared to the rat sequence, the alternative first exons of the porcine GR showed greater homology with that of the human GR gene (Fig. 2B). Moreover, similar to human GR gene, the exon 1-9,10 of porcine GR is a single exon, while in rats exon 1-9 and exon 1-10 are 2 independent exons separated by an intron [3].
The reported nucleotide sequence of the porcine GR promoter is published in NCBI database with the accession number JN624280.
Body weight, liver index and serum cortisol concentration were significantly different between Large White (LW) and Erhualian (EHL) piglets at birth As shown in Table 1, LW piglets were about two-fold heavier in body weight than EHL piglets at birth (P,0.01). Accordingly, liver weight was also significantly higher in LW piglets (P,0.01), yet the liver index (liver weight relative to body weight) did not differ between breeds. Serum cortisol concentration was two-fold higher in EHL piglets compared with LW piglets at birth (P,0.01). These findings are in agreement with the previously published results of breed comparison studies in pigs [18,22].
Hepatic expression of total GR mRNA and 59untranslated GR Exon 1 mRNA variants were different between LW and EHL piglets at birth We determined the abundance of total GR mRNA and the first alternative exon mRNA variants in the liver of newborn piglets with qRT-PCR using primers shown in Table 2. All the seven GR first exon transcript variants were expressed in the liver of newborn piglets. The abundance of GR exon 1 mRNA variants was estimated relative to that of the total GR mRNA, using 18S as an internal standard.

Methylation levels of some CpG sites on porcine GR promoter differ between breeds
Using Methyl Primer Express@ V1.0 (ABI, USA), we predicted a CpG island in the porcine GR gene promoter at positions 25278 to 21265 upstream of the ATG translation start codon in exon 2 (Fig. 4A). All the seven alternative first exons are located in this CpG island. We analyzed the level of methylation on each CpG site located within the 4014 bp CpG island of porcine GR promoter using SEQUENOM's MassARRAY method. Nine pairs of primers covering almost the entire length of the CpG island were used. The raw data from MassARRAY analysis indicating the methylation rates of each CpG units are shown as Table S1. The average percentages of CpG methylation across the whole length of GR promoter were similar in the liver of LW and EHL piglets at birth (Fig. 4B). Neither the intronic nor the exonic regions of each alternative first exons showed significant breed differences in the level of CpG methylation (Data not shown). However, breed differences occurred on some specific CpG sites (Fig. 4B). The methylation rates on CpG 59 (P = 0.05), 70 (P = 0.05), and 76 (P,0.05) were higher in LW than in EHL pigs, whereas the opposite was true for CpG 162 (P,0.05), 260 (P = 0.06), and CpGs 271,274 (P = 0.06). As CpG sites 271,274 are located in close proximity, the MassARRAY method could only report the methylation rate collectively for these 4 adjacent sites. CpG sites 59, 70 and 76 are located in the promoter of exon 1-5, whereas CpGs 260 and 271,274 are located in the promoter of exon 1-11. The breed disparities on the methylation rates of these CpG sites were not in accordance with the differential expression of the corresponding GR mRNA variants in the liver of the two pig breeds [18,22].

Hepatic expression of Sp1, p-CREB and GR in newborn piglets was breed-specific
The cis-acting elements of transcription factors Sp1 [6], CREB (7) and GR [12] are located in human GR promoter 1-4 and 1-5 and involved in the transcriptional regulation of GR gene. To clarify whether these transcriptional factors are expressed in a breed-dependant manner, we analyzed the nuclear abundance of these transcriptional factors in the liver of newborn LW and EHL piglets. Western blot analysis revealed that the nuclear abundance of Sp1, p-CREB and GR in the liver was significantly (P,0.05) higher in LW piglets (Fig. 5).
To clarify whether the predicted cis-acting elements on porcine GR 1-4 and 1-5 promoters are involved in the regulation of the breed-dependent expression of GR exon 1-4 and 1-5 variants, chromatin immunoprecipitation (ChIP) analysis was used to determine the binding of p-CREB, Sp1 and GR to the corresponding regions of the promoters. LW piglets demonstrated significantly (P,0.01) higher p-CREB binding to GR promoter 1-4 ( Fig. 6A), yet GR bindings to both promoter 1-4 and 1-5 were significantly (P,0.05) diminished in the liver of LW piglets (Fig. 6). The higher p-CREB binding to GR promoter 1-4 in LW piglets correlated with higher expression of GR exon 1-4 in this breed, implicating a positive role of p-CREB in the regulation of GR exon 1-4 transcription. On the contrary, LW piglets expressing higher GR exon 1-4 and 1-5 mRNA variants exhibited diminished GR binding to the respective promoters, supporting the prediction of negative GREs in porcine GR 1-4 and 1-5 promoters.
Moreover, ChIP assay revealed significantly higher (P,0.05) level of histone H3 acetylation in promoter 1-4 (P,0.05) and 1-5 (P,0.01) in LW piglets, which is in line with the higher abundances of GR exon 1-4 and 1-5 mRNA variants detected in the liver of LW piglets (Fig. 6). This finding agrees with the    common notion that histone acetylation is usually associated with transcriptional activation [25].

Discussion
To our knowledge, this is the first report on the cloning and analysis of the proximal promoter sequence of porcine GR gene. Our results agree with the previous notion that GR promoter is highly conserved among mammalian species studied, including human [4], rat [21], mice [26], etc. It is worth noting that pig is more close to human than the rodents in terms of the sequence homology and structure similarity of the GR promoter, implying that pig can serve as a better model, compared to the rodents, for human metabolic diseases related to the dysregulation of GR expression. Although the nomenclature used to name the alternative first exons differs between human and pig, the seven alternative first exons identified in pig GR promoter match well with the corresponding counterparts reported in human GR promoter [27]. Nevertheless, we noticed minor differences in human GR exon 1G and 1C, in comparison with porcine exon 1-8 and 1-10 respectively. Human exon 1G was predicted from the sequence homology analysis with the rat exon 1-8 [4], but has not been experimentally validated. The pig GR exon 1-8 identified in this study demonstrated high similarity (89% homology) with the predicted human GR exon 1G, further supporting the existence of exon 1G in human GR gene. Human GR exon 1C is reported to have three variants, 1C1, 1C2 and 1C3 [28], but we failed to identify similar variants for the corresponding exon 1-9,10 in pig GR promoter. No information is available on the functional significance of the exon 1C variants in human GR promoter.
Numerous studies have demonstrated disparity in GR expression between different breeds or strains of rats [29], mice [30], chickens [31,32] and pigs [16,17,33,34]. Majority of the studies focused on GR expression in the brain, while some concern GR expression in peripheral tissues such as liver [31], muscle [16], adipose tissue [33] and placenta [34]. Here we found that total GR mRNA expression was significantly lower in the liver of newborn EHL piglets, which was associated with higher serum cortisol level. Such negative correlation between hepatic GR expression and blood corticosterone level agrees with the findings in Italian chicken breeds [31], yet contradicts to the previous reports on porcine muscle [18] and adipose tissue [33], where Chinese indigenous pigs, EHL or Meishan, demonstrated higher GR expression at later ages. Therefore, the breed-specific GR expression in the pig appears to be dependent on the age and/ or tissue-type.
Studies on human and rodents indicated that the 59-heterogeneity of GR mRNA transcripts are tissue-specific and regulated by complex mechanisms under different conditions [21,28]. We report here, for the first time, the breed-specific expression of GR exon 1 mRNA variants in porcine liver. Among seven untranslated GR first exon transcripts, only exon 1-4 and 1-5 mRNA variants expressed differently between breeds, in a same fashion as total GR mRNA. A similar phenomenon was observed in rat hippocampus concerning GR exon 1-7. Maternal care resulted in up-regulation of hippocampal expression of GR exon 1-7 transcript, which was accompanied by increases of total GR mRNA and protein [13]. In another study, adult rats treated with antidepressant drug Fluoxetine expressed higher GR exon 1-7 together with total GR mRNA in hippocampus, while the constitutional GR mRNA variant, exon 1-10, remained unchanged [35]. However, exon 1-4 and 1-5 mRNA variants, which expressed in much lower abundances compared to the constitutive exon 1-9/10, are unlikely to be the main and only contributors for the breed differences in total GR expression. It is noted that, apart from the proximal exon 1 mRNA variants detected in this study, there might be also distal exon 1 variants and/or unidentified variants expressed in the liver. The percentages of these variants are not determined and it is unknown whether some of them are expressed in a breed dependant manner.
Previous studies suggest that the methylation status of some specific CpG sites in the CpG island of GR promoter can significantly affect transcription of GR mRNA variants, such as exon 1-7 transcript in hippocampus [36], and exon 1-10 transcript in liver [14]. Maternal protein restriction during pregnancy changed the level of methylation in PPAR-a gene promoter in the liver of juvenile rat offspring [37]. Based on the consideration that EHL pigs were traditionally raised under low protein diet, we speculated that the methylation level of GR promoter of EHL may differ from LW piglets. However, we failed to detect breed differences in the overall level of CpG methylation for either the whole length of GR promoter or the proximal promoters for each of alternative first exons. It is possible that methylation changes on some CpG sites in GR promoter would be adequate to fine-tune GR transcription. Indeed, some CpG sites were found to be differently methylated between breeds, among with are CpG70 and 76 containing predicted binding sites for transcription factors, Sp1 and GR.
Transcriptional control within the CpG island of human GR promoter has been well characterized [11]. Here we provide the first evidence that Sp1, CREB and GR could translocate to the nucleus and bind to porcine GR exon 1-4 and 1-5 promoters. Moreover, we found breed disparities not only in nuclear content of GR, Sp1 and CREB, but also in the binding of CREB to GR exon 1-4 promoter and GR to both exon 1-4 and 1-5 promoters. Sp1 and CREB are reported to positively regulate gene transcription, while GR can regulate gene transcription either positively or negatively [38]. Enhanced binding of p-CREB to promoter 1-4 in LW piglets was associated with higher hepatic GR exon 1-4 transcript, which is in agreement with the previous reports [12]. However, binding of GR to promoter 1-4 and 1-5 was negatively correlated with the abundance of GR exon 1-4 and 1-5 transcripts, which supports the prediction of nGREs in porcine GR promoters. This negative feedback regulation of GR on the transcription of itself may explain, to some extent, why high serum cortisol level in EHL piglets was associated with lower hepatic GR expression.
Histone acetylation causes relaxation of chromatin structure and increases the recruitment of transcription factors and RNA polymerase II to initiate transcription [39], thus is often associated with gene activation. Indeed, higher expression of GR exon 1-4 and 1-5 mRNA variants detected in LW piglets was accompanied by increased enrichment of acetylated histone H3 on their promoters. Recent epigenetic studies suggest that DNA methylation, histone modifications, and the binding of transcription factors on a specific locus are closely related and work in concert to regulate gene transcription [40]. Further analytical studies are required to decipher the interactions among these regulatory elements and to evaluate their contributions to the breed-specific expression of 59-untranslated GR first Exon mRNA variants in the liver of newborn piglets.
Taken together, our results indicate that the difference in the hepatic expression of GR exon 1-4 and 1-5 transcript variants between two breeds of pigs is determined, at least partly, by the disparity in the binding of CREB and GR, as well as the enrichment of histone H3 acetylation to the GR exon 1-4 and 1-5 promoters. Methylation of some CpG sites may also contribute to the transcriptional regulation of GR in the liver of newborn piglets. The genomic information on the promoter of porcine GR gene may provide essential basis for further investigations into GR gene regulation in the pig for both agricultural and human medical research.

Animals
To minimize the interference of postnatal factors such as nutrition (milk yield and milk composition) and maternal behavior, we chose to use newborn piglets in the present study. The newborn piglets were obtained from two neighboring pig breeding farms in Changzhou, Jiangsu Province, China. Six newborn male piglets from three litters (2 from each litter) of purebred EHL and LW sows, respectively, were sacrificed within 5 minutes after birth by exsanguinations. Body weight and liver weight were recorded. The blood was collected from the precaval vein and the serum was gathered and kept at 220uC. Liver samples were harvested within 10 min after slaughter and immediately frozen in liquid nitrogen, and then kept at 280uC until subsequent analysis.

Radioimmunoassay for serum hormone levels
Cortisol was measured with commercial RIA kits purchased from Beijing North Institute of Biotechnology, Beijing, China. The detection limit for cortisol was 2 ng/mL and the intra-and interassay coefficients of variation were 10% and 15%, respectively. Samples were analyzed in duplicate within one assay to avoid inter-assay variations.

RNA extraction and GR mRNA exon 1 variant quantification
Total RNA was extracted from liver using TRNzol Total RNA Kit (Tiangen Biotech Co. Ltd, Beijing, China), according to the recommendations of the manufacturer. Two micrograms of total RNA were reverse-transcribed with M-MLV reverse transcriptase (Promega, Madison, WI, USA) following the standard protocol. Using transcript-specific primer pairs (Table 1), we measured the expression of these alternatively spliced transcripts and the total GR mRNA in liver of LW and EHL piglets. Real-time PCR was performed on a Mx3000P (Stratagene, La Jolla, CA, USA) with SYBR Premix Ex Taq TM (TaKaRa, Dalian, China) to quantitate mRNA of total GR and its 59 first exon variants. The primers were designed to ensure that same PCR condition (denaturing at 95uC for 2 min, followed by 40 cycles of 95uC for 20 s, 64uC for 20 s, 72uC for 20 s, and then 72uC for 10 min) can be used to amplify all the GR mRNA variants with a similar efficiency. To confirm the amplification specificity, the PCR products from each primer pair were subjected to a melting curve analysis and sequencing.

Identification of 59 and 39 boundary locations of the alternative first exons of pig GR gene
The 39 boundaries of the alternative first exons were determined by RT-PCR followed by sequencing. The upstream primers, which were designed according to the sequence homology of pig with human and rat, were located in various first exons, while the downstream primers were located in coding region of exon 2. Seven pairs of primers were used to amplify the alternative first exons with cDNAs, then the PCR products were detected in gel and the expected bands were cut out and cloned into the pMDH18-T Vector (TaKaRa, Dalian, China), after that the clones were sequenced. All the alternative first exons were spliced to the same nocleotide sequence located in exon 2, thus the 39 end of the first exons were determined.
The 59 boundaries were identified by bioinformatic prediction referring to the 59 boundary sequences of the first exons in human and rat GR. All the spliced introns obeyed the GT/AG rule.

Sequenom's Mass ARRAY for quantitative methylation analysis of porcine GR promoter
Genomic DNA was treated with sodium bisulfite using EZ-96 DNA methylation kit (Zymo Research). Sequenom's Mass ARRAY platform (CapitalBio, Beijing, China) was used to perform quantitative methylation analysis. The system utilizes MALDI-TOF mass spectrometry in combination with RNA basespecific cleavage (MassCLEAVE). A detectable pattern is then analyzed for methylation status. PCR primers were designed by using Methprimer (www.urogene.org/methprimer/). For each reverse primer, an additional T7 promoter tag for in vitro transcription was added. In addition, a 10-mer tag was added on each forward primer to adjust for melting-temperature differences between the forward and reverse primers. The sequences of the primers used are not shown here and will be available upon request.
Computer program based analysis of GR promoter 1-4 and 1-5.

ChIP assay for the analysis of transcription factors binding and histone 3 acetylation on GR promoters
Antibodies against GR (sc-1004, Santa Cruz, CA, USA), Sp1 (sc-14027, Santa Cruz, CA, USA), p-CREB (BS4053, Bioworld Technology, MN, USA), and acetyl-Histone H3 (06-599, Millipore, MA, USA) were used for chromatin immunoprecipitation assay following previously described procedure [42]. Briefly, approximately 300 mg frozen liver tissue was ground in liquid nitrogen and fixed by addition of formaldehyde to a final concentration of 1% for 10 min at room temperature. The reaction was stopped by adding glycine to a final concentration of 125 mM while keeping rotating for 5 min at room temperature. The pellet was resuspended in cell lysis buffer (10 mM Hepes, pH 7.9, 1.5 mM MgCl2, 10 mM KCl, 0.2% NP-40, 0.2 mM sodium orthovanadate, 0.15 mM spermine, 0.5 mM spermidine, 1 mM EDTA, 5% sucrose, 1 mM dithiothreitol, and protease inhibitors), Dounce homogenized, and centrifuged. The pellet was washed once with ice cold PBS and resuspended in 3 ml SDS lysis buffer (50 mM Tris-HCl, 10 mM EDTA, 1% SDS). The crude chromatin released from the nuclei was sonicated (Sonic & Materials Inc., USA) at 50% amplitude for 3 min (10 s on, 10 s off) to produce 500-1000 bp fragments. Sonicated chromatin was quantified on the basis of DNA content at A260. The chromatin equivalent of 40 mg DNA based on the absorption at A260 was used in each immunoprecipitation. One-fifth of chromatin used for each immunoprecipitation was saved and designated as input. Three micrograms of respective antibody or normal IgG (control) were added to the chromatin and incubated overnight at 4uC. The immune complexes were collected using protein A/G plus beads (SC-2003, Santa Cruz, CA, USA), followed by extensive washing with low salt, high salt and LiCl buffers, and then twice with TE buffer (pH 8.0). Then the DNA-protein-antibody complexes were eluted and the DNA-protein formaldehyde cross-links were reversed by incubation at 65uC for 6 h. Proteins were digested by proteinase K at 45uC for 2 h. DNA fragments were purified by phenol/chloroform extraction and ethanol precipitation in the presence of 2 ml 10 mg/ml glycogen. DNA pellet was resuspended in 30 ml of H2O. Three micro liters of ChIPed DNA were used as the PCR template. The primers used for PCR amplification were shown in Table 2.
The amount of precipitated DNA was calculated relative to the input, The results are represented as a percentage of the input DNA (% input), which is calculated as 2 2gCt [normalized ChIP] 6100, where gCt [normalized ChIP] = Ct [ChIP] -(Ct [Input] -Log 2 (Input Dilution Factor)) [43].
Western blot analysis for measuring the nuclear content of transcription factors Nuclear protein extracts was prepared from liver samples as previously described [44]. Briefly, about 200 mg frozen liver tissue was ground in liquid nitrogen and rresuspended in 2 ml of low-salt buffer containing 20 mM HEPES pH 7.9, 10 mM KCl, 0.1 mM NaVO 4 , 1 mM EDTA, 1 mM EGTA, 0.2% NP-40, 10% glycerol, complete protease inhibitor (Roche Applied Science), and incubated for 10 min on ice. Then the mixture was homogenized with glass Dounce homogenizer, centrifuged for 1 min at 12 000 rpm at 4uC. The remaining pellet was dissolved with 200 ml high-salt buffer containing 420 mM NaCl, 20 mM HEPES pH 7.9, 10 mM KCl, 0.1 mM NaVO 4 , 1 mM EDTA, 1mM EGTA, 20% glycerol, and complete protease inhibitors. Samples were kept on ice for 30 min, then centrifuged for 10 minutes at 12 000 rpm at 4uC. Supernatants were taken as nuclear fractions. Protein concentration was measured with a BCA assay kit (Pierce, Rockford, IL) according to the manufacturer's instructions.
Antibodies against GR, Sp1 and p-CREB used for Western blot analysis were the same as used for ChIP assay. Images were captured using VersaDoc 4000MP system (Bio-Rad), and the band densities were calculated using Quantity-One software (Bio-Rad). Histone H1 (BS1656, Bioworld Technology, MN, USA) was used as a loading control. The nuclear contents of the transcription factors are presented as the fold change relative to LW.

Statistical analysis
All data are presented as mean 6 SEM (n = 6), and were analyzed using Independent-Samples T Test with SPSS 13.0 for Windows (SPSS, Inc., Chicago, IL). The mRNA abundances of the alternative first exons are expressed as the fold change relative to that of the exon 1-4 in LW piglets. Nuclear protein abundances are expressed as the fold-change relative to that in LW. * and ** indicate significant difference between breeds at P,0.05 and P,0.01, respectively.