Targeted mutations in myostatin by zinc-finger nucleases result in double-muscled phenotype in Meishan pigs

Myostatin (MSTN) is a dominant inhibitor of skeletal muscle development and growth. Mutations in MSTN gene can lead to muscle hypertrophy or double-muscled (DM) phenotype in cattle, sheep, dog and human. However, there has not been reported significant muscle phenotypes in pigs in association with MSTN mutations. Pigs are an important source of meat production, as well as serve as a preferred animal model for the studies of human disease. To study the impacts of MSTN mutations on skeletal muscle growth in pigs, we generated MSTN-mutant Meishan pigs with no marker gene via zinc finger nucleases (ZFN) technology. The MSTN-mutant pigs developed and grew normally, had increased muscle mass with decreased fat accumulation compared with wild type pigs, and homozygote MSTN mutant (MSTN−/−) pigs had apparent DM phenotype, and individual muscle mass increased by 100% over their wild-type controls (MSTN+/+) at eight months of age as a result of myofiber hyperplasia. Interestingly, 20% MSTN-mutant pigs had one extra thoracic vertebra. The MSTN-mutant pigs will not only offer a way of fast genetic improvement of lean meat for local fat-type indigenous pig breeds, but also serve as an important large animal model for biomedical studies of musculoskeletal formation, development and diseases.


Production of MSTN mutant Meishan pigs by ZFN technology. ZFN plasmid pairs PZFN1/
PZFN2 were specifically designed to target exon 2 of porcine MSTN gene (Fig. 1A, Supplementary Fig.  S1). The specific cleavage RSRR site for producing the N-terminal propeptide and the C-terminal mature myostatin is located 51 bp downstream of the target site. Modification of the specific ZFN target sites is expected to result in loss of MSTN function. Following co-transfection with PZFN1 and PZFN2 plasmids to Ms primary fetal fibroblasts, PCR products of mixed cells were cloned to TA vectors and sequenced to determine targeting efficiency and types of mutations. The targeting efficiency of this ZFNs is ~4% when transfected into primary porcine fetal fibroblasts. Single-cell colonies were produced by limiting dilution in drug-free cell culture medium. We screened 1800 cell colonies using sequencing method. Of these 1800 cell colonies, 80 MSTN-mutant cell colonies were identified. 80 MSTN-mutant cell colonies including 78 heterozygous mutation (MSTN +/− ) and 2 homozygous mutation (MSTN −/− ) were identified by DNA sequence analysis (Fig. 1B). Over 90% of these mutants were short fragment deletions or insertions (< 20 bp) (Fig. 1C). Of these 80 MSTN-mutant cell colonies, fourteen were selected as donor cells to produce MSTN-mutant piglets by using SCNT (Fig. 1D). Please note that cell colony #172 which is a MSTN homozygous mutant, did not generate any live offsprings (Supplementary Table S1). A total of 19 mutant piglets were born live from cell colony #105, #110 and #127 (Supplementary Table S1). Among these mutant piglets, 8 piglets from cell colony #105 and 1 piglet from cell colony #110 grew to adulthood. Colony #105 had a 15 bp deletion and a T-G mutation in one allele, and colony #110 had an 11 bp deletion in one allele. DNA sequencing confirmed that these nine piglets have the same mutations in MSTN as their donor cell colonies. The cloned male F0 MSTN +/− pigs from colony #105 were mated with WT pigs to generate F1 female and male MSTN +/− pigs. F1 heterozygous pigs were then bred to generate F2 MSTN −/− , MSTN +/− and MSTN +/+ pigs (Fig. 1E), with a ratio of 1:2:1 for these three genotypes (MSTN −/− : MSTN +/− : MSTN +/+ ), conforming to the Mendel's law of inheritance.
In addition, FokI nuclease, CMV, and KanR domain was amplified by PCR (Supplementary Table S2) to examine whether ZFN plasmid DNA had integrated into the porcine genome. No integration of ZFN plasmid was detected in the DNA from all piglets tested ( Supplementary Fig. S2). Sequence analysis of 11 predicted off-target sites indicated that there were no off-target cleavage events observed (Supplementary  Tables S3 and S4).

Expressions of MSTN mRNA and protein in MSTN-mutant Ms pigs.
To detect the impact of the MSTN gene mutations on MSTN mRNA expression, we extracted total RNA from longissimus dorsi of MSTN +/+ , MSTN +/− and MSTN −/− pigs, produced from colony #105, at the age of 8 months, then the coding sequence of MSTN mRNA was amplified and sequenced ( Fig. 2A). In the reverse transcriptase-PCR, Scientific RepoRts | 5:14435 | DOi: 10.1038/srep14435 it is clear that MSTN −/− homozygote only have a small PCR fragment compared with the wild-type, which had a big PCR product. In MSTN +/− , there are two PCR products. Both PCR fragments were cloned to the TA vector and sequencing analysis of the two PCR products indicate that the small fragment is 887 bp and the big fragment is 1080 bp. The T-G mutation derived from colony #105 disrupted the normal splicing of RNA, resulting in a 193-nt deletion in exon 2 (Fig. 2B). The altered RNA splicing caused a frameshift and premature termination of translation, resulting in possible production of a peptide of 188aa (Fig. 2C). The transcription level of MSTN was measured via real-time PCR (qRT-PCR). To detect intact MSTN mRNA levels, we designed two sets of primers to amplify MSTN mRNA. The primer pairs used for detecting MSTN-total are located in Exon1 region (1546-1565) and the beginning region of Exon2 (3534-3552) respectively, and the primer pairs used for detecting MSTN-intact are located in the end region of Exon2 (3744-3766) and Exon3 (5903-5922). The detailed information about the primer sequences and MSTN genomic DNA sequences were reported in Supplementary Fig. S1. Since the segment with 193-nt deletion in mRNA is located 3605-3798, the primer set of MSTN-intact would not be able to amplify any fragment with mRNA samples from MSTN-mutant pigs. The results demonstrated that the mRNA sample from wild-type pigs can be successfully amplified by the primer sets of the MSTN-total and MSTN-intact. However, there was no PCR amplification with mRNA samples from MSTN −/− pigs with the primer set of MSTN-intact. The level of intact MSTN mRNA in heterozygous individuals decreased compared with WT pigs (Fig. 2D). The results with the primer set for MSTN-total showed successful amplifications for all pigs. The qPCR results indicated that total MSTN mRNA in MSTN −/− pigs was substantially greater than in MSTN +/+ pigs, the amplified products apparently are from mutant mRNA since the intact mRNA were not detected. The length of amplified DNA fragments is shorter in MSTN −/− pigs than that in wild-type pigs. The transcription level detected by these primers in MSTN −/− pigs suggest that mutant MSTN gene can be transcripted to RNA in MSTN −/− pigs, but can not be successfully translated to functional myostatin protein due to the 193-bp deletion in Exon2. As a result of the defects in MSTN translation, feedback signal may enhance the transcription and accumulate more products of mutant MSTN mRNA. The results from Western blotting showed MSTN precursor (52 kDa) and C-terminal mature myostatin dimer (26 kDa) bands in skeletal muscle extracts from the wild-type pigs and heterozygote (MSTN +/− ), but both protein bands were not detectable in MSTN −/− pigs (Fig. 2E). We also employed ELISA to detect myostatin protein in serum. Serum concentration of mature myostatin is 8 ng/ml and 6 ng/ml in wild-type and MSTN +/− pigs, respectively, and it was not detectable in the serum of MSTN −/− pigs (Fig. 2F). Therefore the results from Western blotting and ELISA confirmed that there was no functional MSTN protein in MSTN −/− Ms pigs.  (Fig. 1E). There were not significant differences in body weight in the early growth stage up to 6 months of age among three genotype pigs. However, the average body weight of male MSTN −/− pigs was heavier than MSTN +/+ pigs after 6 months of age. At the age of 8 months, the average body weight of male MSTN −/− pigs significantly increased by 14.6% compared to MSTN +/+ pigs (Fig. 3A). To confirm if the changes in body weight of MSTN −/− pigs were caused by skeletal muscle mass, we evaluated the carcass of male pigs at 8-month-old. The average carcass weight of MSTN −/− pigs was 47.07 ± 1.40 kg, significantly heavier by 23.77% than MSTN +/+ control pigs (38.03 ± 0.54 kg). There were significant differences (P < 0.01) in carcass weights between wild type (WT) and mutant pigs (MSTN −/− and MSTN +/− ), and MSTN −/− carcass was significantly heavier than MSTN +/− carcass (P < 0.01) (Fig. 3B). Further analysis of muscle, fat, skeleton, and skin tissue indicated that the lean percentage was greater in MSTN −/− pigs by 8.08% and Data are expressed as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, Student's t test.

Double-muscled phenotype of MSTN
11.62%, respectively, than that in MSTN +/− and MSTN +/+ pigs. On the other hand, the percentage of body fat is lower in MSTN −/− pigs by 3.27% and 7.23%, respectively, than that in MSTN +/− and MSTN +/+ pigs. Ratios of skeleton and skin to carcass weights decreased in MSTN −/− pigs compared with MSTN +/− and MSTN +/+ pigs (Fig. 3C), indicating that the increased carcass weights in MSTN −/− pigs came from skeletal muscle mass. To further investigate the effects of targeted MSTN mutation on individual muscles, we dissected longissimus dorsi, triceps, semitendinosus, semimembranosus, and gastrocnemius from 8 month-old male pigs. The average weights of individual muscle in MSTN −/− pigs increased significantly by 59.85% to 101.71% compared with MSTN +/+ pigs. Particularly, the weights of semimembranosus and longissimus dorsi increased by 101.71% and 96.67%, respectively (Fig. 3D, Supplementary Table S5). These dramatic increases in semimembranosus and longissimus dorsi are the major reasons why MSTN −/− pigs have large hips and wide backs, important characteristics of DM beef cattle body structure.
Increased muscle mass of MSTN −/− Ms pigs results from muscle fiber hyperplasia rather than hypertrophy. To determine if the increased muscle mass is due to hyperplasia and/or hypertrophy of muscle fibers, longissimus dorsi and semitendinosus were examined histologically (Fig. 4A). The average size of myofibers in longissimus dorsi from MSTN −/− pigs (1286.01 ± 77.36 μ m 2 ) was substantially smaller than in MSTN +/+ pigs (1975.51 ± 111.20 μ m 2 ). On the other hand, the average size of myofibers in MSTN +/− pigs (2837.31 ± 214.41 μ m 2 ) increased in comparison with MSTN +/+ pigs (Fig. 4A). Distribution of different sizes of myofibers showed that the relative percentage of smaller fiber cells in MSTN −/− pigs is greater than in MSTN +/+ pigs, while the relative percentage of larger fiber cells in MSTN +/− pigs is higher than in MSTN +/+ pigs (Fig. 4A). Corresponding to these observations in cell size distribution, the average density of myofibers in longissimus dorsi from MSTN −/− pigs (475.25 ± 18.13/ mm 2 ) was significantly higher than in MSTN +/+ pigs (396.35 ± 10.51/mm 2 ). However, the average density of myofibers in MSTN +/− pigs is 348.88 ± 15.57/mm 2 , lower than in MSTN +/+ and MSTN −/− pigs (Fig. 4A). The average size of myofibers in semitendinosus from MSTN −/− pigs (1508 ± 93.33 μ m 2 ) was significantly smaller than in MSTN +/+ pigs (1954 ± 34.99 μ m 2 ) and the average density of myofibers in semitendinosus from MSTN −/− pigs (514.8 ± 39.47/mm 2 ) was significantly higher than in MSTN +/+ pigs (412.9 ± 4.36/mm 2 ) (Fig. 4A), although the difference is not at the same magnitude as observed for longissimus dorsi. Additionally, the degree of difference in cell size distribution of myofibers in semitendinosus is also less significant than observed in longissimus dorsi (Fig. 4A). These results are consistent with what were observed for relative percentage of individual muscles. For example, when compared with MSTN +/+ pigs, the average weight of longissimus dorsi in MSTN −/− pigs increased by 96.67% while the average weight of semitendinosus in MSTN −/− pigs increased by 64.52% (Table S5). The relative numbers of myofibers were calculated from the loin eye area and density of myofibers for longissimus dorsi (which has a greater increase in mass and regular cellular shape). At the age of 8 months, the loin eye areas of longissimus dorsi are 33.02 ± 2.72 cm 2 , 24.46 ± 0.35 cm 2 , and 19.66 ± 0.76 cm 2 , respectively, in MSTN +/+ , MSTN +/− and MSTN −/− pigs (Fig. 4B). The calculated relative number of myofibers is 1.57 × 10 6 in MSTN −/− Ms pigs which is significantly greater than that in MSTN +/− (8.54 × 10 5 ) and MSTN +/+ (7.79 × 10 5 ) pigs (Fig. 4B). These results clearly reflect the differences in the number of myofibers among three genotypes of MSTN pigs. Thus, we conclude that the increased muscle mass in MSTN −/− Ms pigs primarily results from hyperplasia rather than hypertrophy of individual muscle fibers. In the meantime, the increased muscle mass of MSTN +/− Ms pigs appears to result from fiber hyperplasia and hypertrophy.  Table 1). Apparently, vertebrae pattern T15L5 is not present in the wild-type pig population. T15L5 with an extra thoracic vertebra is likely the consequence of targeted myostatin mutation in Meishan pigs since approximately 20% MSTN-mutant pigs (19.7% in MSTN +/− , and 20% in MSTN −/− ) ( Table 1). These results imply that myostatin also plays a regulatory role in the development and formation of thoracic vertebrae.

Analysis of meat quality in wild-type and MSTN mutant Ms pigs.
To investigate the effect of loss-of-function mutation in MSTN on meat quality, we measured several parameters such as pH, color, drip loss, cooking holding percentage (CHP), and tenderness. Results indicated that there are no obvious differences noted in meat quality parameters between MSTN mutant pigs and WT pigs except that the "a" value of meat color and shear force in MSTN −/− pigs decreased compared to MSTN +/+ pigs (Table 2). Additionally, the main nutrition components (including protein, fat, moisture and amino acid) Scientific RepoRts | 5:14435 | DOi: 10.1038/srep14435 of longissimus dorsi from 8-month male Ms pigs indicated that there were no differences in total proteins and amino acids between three genotypes of pigs. However, the total fat in MSTN −/− pigs decreased significantly compared to MSTN +/+ pigs (Supplementary Table S6), which is consistent with relative percentage of body fat in carcass (Fig. 3C). These results further confirmed that loss of myostatin function via targeted myostatin mutation have negative effects on body fat accumulation, including fat deposition in skeletal muscles.  Table S7) and biochemical parameters among all three genotypes, except that the levels of serum creatinine (CR), glucose (GLU) increased while the triglyceride (TG) decreased in MSTN −/− Ms pigs compared with MSTN +/+ pigs (Supplementary Table S8). Serum creatinine is 125.60 ± 14.17 μ mol/L in MSTN +/+ pigs, 117.90 ± 11.79 μ mol/L in MSTN +/− pigs and 209.50 ± 12.50 μ mol/L in MSTN −/− pigs. The higher CR level in MSTN −/− Ms pigs than wild-type pigs is certainly due to the increased muscle mass. The increased serum creatinine level has been reported in DM cattle 33 . There has been reports that serum glucose level in MSTN mutant mice is lower than WT mice 23 . But it is interesting to note that serum glucose level is greater in MSTN −/− Ms pigs than that in MSTN +/+ pigs at age of 8 months old. Further studies are required to investigate this phenomenon in pigs. The decreased serum TG level in MSTN −/− Ms pigs is consistent with their lower percentage of body fat than wild-type Ms pigs.

Discussion
High lean meat yield combined with low body fat has long been one of the ultimate goals in the breeding strategy of livestock industry 34 . In this study, we successfully generated MSTN loss-of-function mutant pigs by specifically targeting the exon 2 site of porcine MSTN gene with ZFN technology and somatic cell nucleus transfer. The percentage of lean meat yield of MSTN −/− pigs at 8 months of age is 67%, approximately 12% greater than the corresponding MSTN +/+ pigs. Although DM cattle such as Belgian Blue beef cattle have 20% more muscle mass on average, less bone, lower fat, the breed indeed has several disadvantages, in particular, the reproduction issues due to unusually heavy and bulky offspring and reduced reproduction tract 15 . On the other hand, MSTN-mutant pigs generated in this study are as healthy as normal littermate control pigs. They develop and grow normally where they are raised and fed  with the same normal diets as control pigs. So far, these MSTN-mutant pigs also have the same normal fertility as WT pigs without any abnormal pregnancy and other reproduction problems. Thus, these MSTN-mutant pigs will have an obvious advantage for the livestock meat industry. In addition, Meishan pigs are considered a famous Chinese pig breed for super-prolificacy because of its large litter size 30 . Like most of local Chinese breeds, they grow slowly and deposit huge amounts of body fat mass with low feed efficiency 35 . The dramatic increase in skeletal muscle mass in MSTN-mutant Meishan pigs demonstrate a way of fast genetic improvement for local fat-type indigenous pig breeds using gene targeting technology. Although the animals were generated by genetic manipulations via somatic cell nuclear transfer, the MSTN-mutant pigs do not have any foreign DNA and associated protein as a result of DNA sequence deletion by ZFN. To date, there is not any genetically modified (GM) livestock for food production 36 . The safety of the GM food is an important concern for consumers if the pig is used for commercial pig productions. It is notable that ZFN-edited MSTN-mutant pigs generated in our current study are very similar to those naturally occurring loss-of-function MSTN mutations observed in DM cattle. The pork from MSTN-mutant pigs will be as safe as the MSTN-mutant beef such as Belgian Blue, an important and popular beef cattle breed in the commercial beef production. Therefore, our ZFN-edited MSTN-mutant pigs should be safe to enter into commercial food supply following the review and approval process by government regulatory agencies. The increase in muscle size in MSTN +/− and MSTN −/− mutant mice compared to wild type mice is due to fiber hyperplasia and hypertrophy of skeletal muscle fibers 6 . In our study, both hypertrophy and hyperplasia of skeletal muscle fibers at 8 months of age were observed in MSTN +/− Ms pigs, but only fiber hyperplasia was observed in MSTN −/− Ms pigs, which is consistent with the observation in DM cattle 33,37 . MSTN is an essential regulator of proliferation and differentiation of muscle cells during muscle development, which is detected from early myogenetic stage during embryonic myotome formation to adult skeletal muscle developmental stage 1,[37][38][39][40] . Studies on muscle development demonstrated that muscle fiber number is primarily determined before birth, and the diameter of myofibers expands after birth [41][42][43] . The different patterns of myofibers observed in MSTN +/− and MSTN −/− are likely related to time of myofiber formation and postnatal muscle hypertrophy in MSTN-mutant pigs. Loss of MSTN functions can lead to an increase in the number of myofibers during embryonic phase and then an increase in diameter of myofibers after birth 38 . We believe that the number of myofibers increases significantly during embryonic phase in MSTN −/− pigs, the significantly increased number of fibers may limit the enlargement or hypertrophy of all the myofiber after birth in the MSTN −/− pigs with complete loss of myostatin function. In MSTN +/− pigs, the function of MSTN gene is only partially lost, or with one copy of functional MSTN gene. The number of myofibers increases to some degree during embryonic phase, the magnitude is much lower than in MSTN −/− pigs. During postnatal development, the diameter of myofibers in MSTN +/− pigs can increase to some degree after birth compared with MSTN +/+ pigs, there resulting in muscle phenotype of fiber hyperplasia and hypertrophy.
Interestingly, the results from vertebrae counting showed an additional thoracic vertebra (T15L5, Fig. 5) in about 20% of MSTN-mutant pigs, which was not seen in the wild-type littermate control Ms pigs. To our knowledge, this phenomena was not observed or have not been studied in either MSTN knockout mice or other MSTN-mutant animals. Although the number vertebrae in most mammals are fixed at 19, swine appears variable due to genetic selection for meat production 44,45 . Wild boars have 19 vertebrae while genetically selected improved breeds have vertebrae numbers from 20 to 23 46 47 . GDF11 −/− mice contain more thoracic and lumbar vertebrae compared to GDF11 +/+ mice. However, loss of both alleles of MSTN in GDF11 −/− mice resulted in a much more severe phenotype. Over-expression of transgene GDF-11 propeptide cDNA in bone tissues have resulted in transformation of the seventh cervical vertebra into a thoracic vertebra, and promote bone formation, and density in mice 48,49 . We also performed sequence analysis for the MSTN-mutant pigs (Supplementary Table S10), and did not observe any mutations in GDF11 gene of MSTN-mutant pigs, implying that the T15L5 supernumerary thoracic vertebrae is the effect of targeted MSTN mutation not the GDF11 on skeletal development. Recent studies have showed the positive effects of myostatin inhibition on bone densities in MSTN-knockout mice through regulation of stem/progenitor cell potency 50 . Also, administration of soluble myostatin decoy receptor (ActRIIB-Fc) increased bone mass and bone volume fraction (BV/TV) significantly in the distal femur and lumbar vertebrae 51 . Clearly, there are animal experimental evidences that support positive effects of inhibition of myostatin function on skeletal formation and maintenances. Further investigations are required on molecular mechanisms of loss of myostatin function on skeletal formation and development, in particular the mechanism of MSTN and GDF11 redundancy in musculoskeletal formation and development. Regardless of the mechanism, there is a positive relationship between the number of thoracic vertebrae and pork yield. The increased number of thoracic vertebra has an economic benefit for pork production.
Meat quality analysis results indicated that there were no obvious differences between MSTN mutant pigs and WT pigs except that the "a" value of meat color and shear force in MSTN −/− pigs compared to MSTN +/+ pigs, which is consistent with what was observed in DM cattle 42 . The "a" value of meat color is an indication of meat redness which reflects the level of oxymyoglobin. It has been known that the level of oxymyoglobin is higher in type I than in type IIB myofibers 52 . We speculate that a decrease in the "a" value in MSTN −/− pigs indicates higher frequency of type IIB myofibers in MSTN −/− pigs. We further analyzed Myh4 protein (a specific gene of type IIB myofibers) in longissimus dorsi from MSTN mutant pigs. The results showed that the expression level of Myh4 is significantly higher in MSTN −/− pigs than in MSTN +/+ pigs ( Supplementary Fig. S3), suggesting that MSTN −/− pigs have more type IIB myofibers, which is consistent with the change of "a" value in MSTN −/− pigs. The results of meat biochemical composition analysis indicated that there was no difference in total proteins and amino acids between three genotypes of pigs except the total fat. The decrease in muscle fat is consistent with decreased% fat in carcass. The value of shear force is an indication of meat tenderness, the smaller the shear force, more tenderer the meat. The decrease in shear force in MSTN −/− pigs indicates the meat tenderness increases, which may be related to thinner myofibers in MSTN −/− pigs. Most studies on DM cattle indicate that their meat tenderness is improved in comparison with normal (non-DM) beef cattle.
In conclusion, we have successfully generated MSTN-mutant Meishan pigs using ZFN technology in combination with somatic cell nucleus transfer. These MSTN-mutant pigs developed and grew normally, had increased growth performance after 6 months of age with dramatic skeletal muscle mass, producing a significant amount of carcass lean tissue and less body fat. These unique quality characters are particularly more apparent in MSTN −/− pigs, showing great similarity in body structure to DM cattle with much higher muscle mass than MSTN-mutant heterozygotes and wild-type pigs. Moreover, one very interesting observation is that about 20% of MSTN-mutant pigs contain one extra thoracic vertebrae. This result provides us a new insight to better understand MSTN's function in both skeletal and muscle formation and development in the future studies. The new MSTN-mutant pigs on Meishan pig genetic background generated in this study not only demonstrated a way of fast genetic improvement of lean-meat yield for local fat-type pig breeds, but also could serve as a great large animal model for biomedical studies in musculoskeletal development. Transfection, screening and identification of MSTN mutant cell line. All cell culture medium and reagents were purchased from Invitrogen. Primary porcine fetal fibroblasts were established from fetus collected from a 35 day pregnant Meishan sow and cultured in DMEM supplemented with 10% FBS, 10 mM MEM non-essential amino acids, 8 mM L-glutamine, 1 × sodium pyruvate, 1 × HT supplement, 2 ng/mL bFGF, and 1% Pen/Strep. For ZFN transfections, cells were transfected with the MSTN-ZFNs using Nucleofector TM (AMAXA) according to the manufacturer's instruction, with program T-016 being selected. After 48 h, cells were diluted and re-plated in 10 cm 2 dishes (150 cells/dish on average). Positive colonies containing ZFN-mediated modification were screened and identified by DNA sequencing. Individual cell clones were isolated 11 or 12 days after culture, and then were expanded, sequence analyzed and cryopreserved after a total of 22-24 days in culture. Cells were incubated with 5% CO 2 at 37 °C.

Methods
Generation of MSTN mutant pigs by somatic cell nucleus transfer. Piglets were generated using somatic cell nucleus transfer with modified protocol named handmade cloning method (HMC) as described by Du et al. 54 . Fourteen different MSTN-mutant cell lines including one that contains a MSTN homozygous mutant, #172, were used as donor nuclei to generate the reconstructed embryos. The blastocysts at day 5 to day 7 with clearly visible inner cell mass produced from the HMC were surgically transferred to landrace sows on 5 or 6 days of standing estrus. Pregnancies were diagnosed by ultrasonography on day 28 after the surgery and confirmed every two weeks.
Genotyping of MSTN mutants in Cloned piglets and offsprings. Genomic DNA was extracted from fetuses and newborn piglets and then was subjected to PCR (Supplementary Table S9, Supplementary  Fig. S4). Following sequencing, the nucleotide sequence were analyzed via BioEdit to determine the genotype. "BioEdit" is a software for sequence editing and analysis. For heterozygote, PCR product of the targeted region of the MSTN gene was subcloned into the pGEM-TEasy Vector system (Promega) per Scientific RepoRts | 5:14435 | DOi: 10.1038/srep14435 manufacturer's protocol, and twenty colonies for each sequence were picked and sequenced using the T7prom primer.

RNA isolation and real-time RT-PCR.
Total RNA was extracted using the Trizol reagent (Invitrogen), and cDNA was synthesized using the RevertAid TM First Strand cDNA Synthesis Kit (Fermentas) with 500 ng total RNA. The coding sequence of MSTN was amplified and sequenced. Each real-time PCR contained SYBR Select Master Mix (Life Technologies), 0.2 μ M forward and reverse primers, 2 μ L template cDNA, and dH 2 O up to a final volume of 20 μ L. Reactions were performed on a 7500 FAST Real-Time PCR System (Applied Biosystems). Cycling conditions consisted of an initial, single cycle of 2 min at 95 °C followed by 40 cycles of 15 s at 95 °C and 1 min at 60 °C. Porcine GAPDH gene was used as an internal control to normalize the RT-PCR efficiency and to quantify the expression of the genes in WT pig, heterozygote-and homozygote MSTN mutant Ms pigs. All data were analyzed by the 2 −ΔΔCT method using 7500 System SDS Software V1.4.0. RT-PCR primer for MSTN and real-time RT-PCR were shown in Supplementary Table S10.
Western Blot and ELISA. Total proteins were extracted from longissimus dorsi using Thermo Scientific T-PER tissue protein extraction reagent (#78510). Protein concentration was measured with Micro BCA Protein Assay Kit from Thermo Scientific (#23235). Equal amount of total protein from MSTN +/+ , MSTN +/− and MSTN −/− samples were loaded and separated by 12% SDS-PAGE. Following transfer of protein from gel to PVDF membrane, Western blot was performed using standard method. The primary antibodies against MSTN are from Abcam (ab55106) diluted at 1:1000. And the second antibodies are from CST (#7076) diluted at 1:2000. Color development was performed with SuperSignal West Pico Chemiluminescent Substrate from Thermo Scientific (#34080). Western blot for GAPDH (MBL, M171-7) was performed as an internal standard. The concentration of MSTN protein in serum was determined using an ELISA kit for porcine myostatin, with the antibody recognizing the C-terminal domain of MSTN protein (Bioss).
Animal euthanasia, carcass dissection, and sample collection. Experimental pigs were euthanized, blood samples were collected. Hairs, heads, hoofs and internal organs were removed. Body weight, length, backfat thickness, and loin eye areas of carcass were measured, and the number of thoracic and lumbar vertebra were counted manually. Skin, bone, muscles and fat were then dissected from left carcass and their individual weights were determined. Total protein, total fat, moisture, amino acid, and essential amino acid were analyzed for longissimus dorsi. Tissue samples used for molecular detection were collected from right carcass and snap-frozen in liquid nitrogen.
Histochemistry and analysis of myofiber size and number. Muscle samples for histochemistry were taken at the same location. After each tissue was fixed and slides prepared, haematoxylin and eosin staining was performed by a standard method. Slides were viewed by Olympus DP72 image system. For each sample, five view fields (areas) were randomly selected and then analyzed using IPP 6.0 software. The cross-sectional areas of 400-500 myofibers were measured, and average area of each myofiber, total number of myofibers and myofiber density were determined. Statistical analysis. Statistical comparisons of body weight, carcass character, myofibers and blood parameters among different genotypes of pigs were performed by the Student t test and p < 0.05 was considered as statistically significant. Statistical analyses were carried out using SAS release 8.1 (SAS Institute Inc., Cary, NC).