Mol Genet Genomics DOI 10.1007/s00438-015-1126-3
ORIGINAL ARTICLE
Comparison of skeletal muscle miRNA and mRNA profiles among three pig breeds Xinhua Hou1 · Yalan Yang1,2 · Shiyun Zhu1 · Chaoju Hua1 · Rong Zhou1 · Yulian Mu1 · Zhonglin Tang1,2 · Kui Li1
Received: 1 July 2015 / Accepted: 28 September 2015 © Springer-Verlag Berlin Heidelberg 2015
Abstract The pig is an important source of animal protein, and is also an ideal model for human disease. There are significant differences in growth rate, muscle mass, and meat quality between different breeds. To understand the molecular mechanisms underlying porcine skeletal muscle phenotypes, we performed mRNA and miRNA profiling of muscle from three different breeds of pig, Landrace (lean-type), Tongcheng (obese-type), and Wuzhishan (mini-type) by Solexa sequencing. Forty-three genes and 106 miRNAs were differentially expressed between Landrace and Tongcheng pigs, 92 genes and 151 miRNAs were differentially expressed between Tongcheng and Wuzhishan pigs, and 145 genes and 156 miRNAs were differential expressed between Landrace and Wuzhishan pigs. Gene ontology analysis suggested that genes differentially expressed between Landrace and Tongcheng pigs were mainly involved in the biological processes of oxidative
stress and muscle organ development. Meanwhile, for Tongcheng vs Wuzhishan and Landrace vs Wuzhishan pigs, the differentially expressed genes were involved in fatty acid metabolism, oxidative stress, muscle contraction, and muscle organ development, processes that are closely related to meat quality. To investigate the molecular mechanisms underlying meat quality diversity based on differentially expressed genes and miRNAs, interaction networks were constructed, according to target prediction results and integration analysis of up-regulated genes with downregulated miRNAs or down-regulated genes with up-regulated miRNAs. Our findings identify candidate genes and miRNAs associated with muscle development and indicate their potential roles in muscle phenotype variance between different pig breeds. These results serve as a foundation for further studies on muscle development and molecular breeding.
Communicated by S. Hohmann.
Keywords miRNAs · Transcriptome · Solexa sequencing · Pig · Skeletal muscles
X. Hou and Y. Yang contributed equally to this study. Electronic supplementary material The online version of this article (doi:10.1007/s00438-015-1126-3) contains supplementary material, which is available to authorized users. * Zhonglin Tang
[email protected];
[email protected] 1
Key Laboratory of Farm Animal Genetic Resources and Germplasm Innovation of Ministry of Agriculture, Institute of Animal Science, Chinese Academy of Agricultural Sciences, Beijing 100193, People’s Republic of China
2
Agricultural Genome Institute at Shenzhen, Chinese Academy of Agricultural Sciences, Shenzhen 518120, People’s Republic of China
Introduction The pig is an important livestock animal that provides a protein source for humans. The pig is also an ideal disease model because it is similar to humans in terms of physiology, anatomy, and genome structure. However, there is significant phenotype variance in traits such as growth rate, muscle mass, and myofiber characteristics between pigs of different genotypes. Under intensive selection for improving growth rate and muscularity, western pig breeds, such as Landrace pigs, typically show rapid growth rates and high lean carcass percentages. Tongcheng pigs, a wellknown indigenous Chinese obese-type pig, exhibit slow
13
growth rate, high fat percentage, and low muscle mass (Fan et al. 2006). Wuzhishan pigs are an indigenous Chinese mini-type pig in terms of size and body weight, which live in tropical areas in southern China (Min et al. 2014). The molecular mechanism of muscle phenotype variance in pigs is unclear. microRNAs (miRNAs) are important in the epigenetic regulation of gene expression. Precursor miRNAs are generated in the nucleus and form a double-stranded RNA structure with a hairpin loop. Once transported into the cytoplasm, they are processed by Dicer to generate miRNA:miRNA* duplexes. Mature miRNAs are about 21–24 nucleotides long and can combine with the Argonaute protein to form a miRNA-Induced Silencing Complex (miRISC). The ‘seed sequence’ of the miRNA, a stretch of seven nucleotides spanning nucleotides 2–8 on the 5′ end of the miRNA, can guide the miRISC to targetgene mRNAs through imperfect base pairing to their 3′ untranslated regions (3′-UTRs), and down-regulate gene expression by post-transcriptional repression or inhibition of translation. Ørom et al. have reported that miRNAs can also recognize the 5′-UTR of target genes and enhance their translation (Ørom et al. 2008). miRNAs can influence the cell cycle and proliferation and, therefore, abnormal expression of miRNAs can induce cancer. They also play important roles in developmental processes, including adipogenesis, osteogenesis, neurogenesis, and myogenesis. Myogenesis is a complex, continuous process including myogenic precursor determination, migration, myoblast proliferation, differentiation, fusion, myotube formation, and maturation of myofibers. miRNAs play a crucial role in skeletal muscle development, as skeletal muscle specific knockout of Dicer in mice leads to skeletal muscle hypoplasia and perinatal lethality (O’Rourke et al. 2007). miR-1, miR-206, and miR-133 are well-known myomiRs specifically expressed in muscle that play important roles in the regulation of muscle cell differentiation and proliferation (Chen et al. 2006). Other miRNAs that are not muscle-specific have also been implicated in the regulation of myogenesis, such as miR-127 (Yang et al. 2014), miR-148 (Zhang et al. 2012), miR-155 (Zhao et al. 2012), and miR21 (Bai et al. 2015). Meat quality is an important consideration in pig breeding. Meat quality is complex and includes traits such as water-holding capacity, color, fat content and composition, pH, juiciness and flavor, and is influenced by factors such as feeding conditions, slaughter methods, nutrition, age, and breed (Rosenvold and Andersen 2003; Davoli and Braglia 2007). Meat quality is closely related to the size of the muscle fibers and the intramuscular fat (IMF) content. The size of muscle fibers is determined during muscle development, and the IMF content correlates with the potency of fatty acid metabolism in skeletal muscles as
13
Mol Genet Genomics
well as the development of intramuscular adipocytes. Pig breeds have diverse meat quality phenotypes, which might be determined by skeletal muscle development. There are two waves of muscle fiber generation in porcine prenatal stages (Wigmore and Stickland 1983). Primary myofibers are mainly formed at 38–64 days post coitus (dpc) and secondary myofibers at 54–90 dpc. In early embryonic periods, primary fibers in indigenous Chinese pigs are formed earlier than in western pigs and the muscle fiber diameter of western pigs is larger than that of indigenous Chinese pigs in later developmental periods (Zhao et al. 2011, 2015). Diversity in myofiber formation and characteristics, which might be caused by different proliferative abilities of satellite cells, might be a reason for the differences in muscle growth rates between indigenous Chinese pigs and western pigs (Wang et al. 2012). Global gene expression profile analysis is a powerful approach to study differences of muscle phenotypes across breeds. Differentially expressed transcripts relating to muscle growth have been analyzed by Long Serial Analysis of Gene Expression (LongSAGE) at 33, 65, and 90 days post coitus in Tongcheng and Landrace pigs. The results showed that developmental molecular changes are more complicated in Tongcheng pigs than in Landrace pigs, and that genes responsible for increased cellular growth and myoblast survival are expressed at higher levels in Landrace pigs (Tang et al. 2007). Zhao et al. identified differentially expressed genes between Lantang (an indigenous Chinese pig breed located in Guangdong province of China) and Landrace pigs using Solexa sequencing at 10 time points and found a series of differentially expressed genes that might give rise to the differences in myogenesis and adipogenesis between the two breeds (Zhao et al. 2011). The results from twodimensional fluorescence difference gel electrophoresis suggest that longissimus muscles of western lean-type pigs might have a greater oxidative capacity than those of indigenous Chinese obese-type pigs (Li et al. 2013). The transcriptome profiles of skeletal muscle from Tongcheng and Yorkshire pigs across 11 developmental stages have been determined by Solexa sequencing and revealed that muscle fiber formation in Tongcheng pigs is initiated earlier than in Yorkshire pigs (Zhao et al. 2015). miRNAs from skeletal muscle and adipose tissues of Lantang and Landrace pigs have been identified by both Solexa sequencing and microarray analysis and confirmed that miRNAs relevant to muscle development are differentially expressed between these two pig breeds (Li et al. 2012). Although porcine muscle development and meat quality have been investigated by profiling the transcriptome or miRNAome in different pig breeds, an integrated analysis of both has rarely been reported. In this study, Landrace (a leaner western breed), Tongcheng (a typical indigenous Chinese breed), and Wuzhishan (a Chinese miniature
Mol Genet Genomics
breed) pigs were chosen to systematically explore the molecular mechanisms underlying porcine meat quality by the integrated analysis of both mRNA and miRNA expression in longissimus muscles. This study also provides valuable information for porcine genetic breeding.
Materials and methods Tissue collection and RNA extraction All animal studies were conducted according to the guidelines established by HuBei Province, People’s Republic of China, for the Biological Studies Animal Care and Use Committee. Landrace, Tongcheng, and Wuzhishan adult pigs (three pigs for each breed at postnatal day 240) were fasted and the following day humanely slaughtered, and longissimus dorsi muscle samples were collected. All fresh tissue samples were collected, frozen in liquid nitrogen, and stored at −80 °C until RNA extraction. Total RNA was extracted from samples using TRIzol Reagent (Invitrogen, CA, USA). The concentration and quality of RNA were validated using an Agilent 2100 Bioanalyzer (Agilent Technologies, CA, USA) and by agarose gel electrophoresis. RNA sequencing and data analysis An RNA library was constructed for each breed from total RNAs pooled from three individuals. mRNA was isolated from total RNA using oligo(dT) magnetic beads and disrupted into short fragments of about 200 bp. These short fragments and random hexamer primers were used to synthesize first-strand cDNA. Then, second-strand cDNA was synthesized using reaction buffer, dNTPs, RNaseH, and DNA polymerase. Short double-stranded cDNA fragments were purified using a QiaQuick PCR extraction kit (Qiagen, Venlo, Netherlands) and then ligated to sequencing adapters. After purification by agarose gel electrophoresis, the fragments were enriched by polymerase chain reaction (PCR) amplification and sequenced using an Illumina HiSeq™ 2000 (Illumina, CA, USA). The raw sequence data were assessed and tags that contained adaptors, were of low quality, or showed more than 10 % unidentified nucleotides, were removed. Only clean reads were used for subsequent analysis, and the unique reads were used for identifying differentially expressed genes with log2 ratio ≥1 and false discovery rate (FDR) ≤0.001 (Liu et al. 2012; Li et al. 2014).
in length were purified from a polyacrylamide electrophoresis gel, and adaptors were ligated to the 5′ and 3′ ends. After 17 cycles of amplification by reverse-transcription (RT)-PCR, the 90-bp PCR products were isolated from 4 % agarose gels. Then, the purified cDNA fragments were directly sequenced using an Illumina Hiseq 2000 (Illumina). After removing 3′ and 5′ adaptor contamination, sequences with a poly(A) tail, and fragments of less than 18 nt, the clean reads were compared to the NCBI GenBank and Rfam databases to annotate all known rRNA, tRNA, scRNA, snRNA, and snoRNA small RNA sequences. The unannotated sequences were searched against the known miRNA precursors and mature miRNAs in the miRbase database to identify known miRNAs. miRNAs with log2 ≥1 and p ≤ 0.05 were considered to be differentially expressed between the different pig breeds (Hou et al. 2012). Real‑time qPCR cDNA was synthesized with random primers for mRNA and stem-loop primers for miRNA, using the RevertAid™ First Strand cDNA Synthesis Kit (Thermo Scientific, Vilnius, Lithuania) and a method similar to that previously described by Hou et al. (2012). qPCR was performed using SYBR® Premix Ex Taq™ (Takara, Otsu, Japan) on the 7500 FAST Real-Time PCR System (Applied Biosystems, CA, USA) according to the manufacturer’s instructions. GAPDH, ACTB, and HPRT were used as endogenous control genes for mRNA, and U6, Met-tRNA, and 5 s for miRNA. All data were analyzed using the 2−ΔΔCT method. Each qPCR reaction was performed in triplicate, and the data were expressed as mean ± standard error (n = 3). The primers for reverse transcription and qPCR are listed in Additional File 1. Bioinformatics analysis GO enrichment was performed using the DAVID Gene Ontology database (http://david.abcc.ncifcrf.gov/), and interactions between differentially expressed genes were constructed using the STRING database (http://string-db.org/). The target genes of differentially expressed miRNAs were predicted from the miRanda database (http://www.microrna. org/) based on both human and mouse genes, owing to the absence of porcine gene data in the current versions. The differentially expressed genes, both those that existed in the cluster of predicted targets and those that were negatively correlated with given differentially expressed miRNAs, were considered to be potential porcine target genes.
Small RNA sequencing and data analysis
Supporting data information
A small RNA library for each breed was produced by pooling total RNA from three individuals. Fragments 18–30 nt
The raw transcriptome data of Landrace, Tongcheng, and Wuzhishan pigs have been submitted to the NCBI
13
Sequence Read Archive (http://www.ncbi.nlm.nih.gov/ Traces/sra/) under accession No. SRP058340.
Results Overview of the transcriptome and the miRNAome A total of 11,848,194, 11,876,260, and 12,454,394 raw reads were generated from longissimus dorsi muscles of adult Landrace, Tongcheng, and Wuzhishan pigs, respectively, by Illumina HiSeq 2000 sequencing. After removing adaptors, contamination, and low-quality reads, 11,680,303, 11,706,227, and 12,284,240 clean reads, respectively, were obtained (Additional File 2). Approximately 69.95, 69.04 and 70.08 % of clean reads, respectively, mapped to the porcine reference genome (version 10.2) in a unique manner (Additional File 3), while 55.57, 54.16 and 55.79 %, respectively, mapped to porcine reference genes (Additional File 4). Unmapped reads, or reads that mapped multiple times, were removed from further analysis. To decipher the characteristics of miRNAs in skeletal muscle among different pig breeds, total RNA from longissimus dorsi muscle of Landrace, Tongcheng, and Wuzhishan pigs was used to construct small RNA libraries. After quality filtering and trimming of contaminant and adaptor sequences, 7,929,450 (93.39 % of raw reads), 16,136,367 (97.86 % of raw reads), and 10,913,097 (98.80 % of raw reads) clean reads were obtained from the three pig breeds, respectively (Additional File 5). About 45.63, 49.05, and 52.89 % of clean reads, respectively, were unique small RNAs mapping to the reference genome (Additional File 6). After removing small RNAs that matched to rRNAs, tRNAs, snRNAs, snoRNAs, or scRNAs, a total of 1,949, 2,112, and 1,860 unique small RNAs were obtained, representing 214, 222 and 207 mature miRNAs in skeletal muscles of Landrace, Tongcheng, and Wuzhishan pigs, respectively (Additional File 7). Differentially expressed genes and GO analysis To elucidate the causality between phenotype and gene expression, the genes with more than twofold difference in expression level and FDR
