Discussion
CRISPR–Cas9 mediated genome editing is a powerful biological technology that has widespread application. Its application to livestock has been slow [21]. In this study, by applying genome editing without the integration of the transgene, an effective knockout condition was established at the embryo level and live-born edited offspring were produced. The MSTN knockout was used for proof of concept because of its clear double-muscling phenotype.
The MSTN gene consists of three domains (a signal sequence, a pro-peptide, and a mature-region). After transcription, translation, and two cleavage events (pro-peptide convertase by furin and tolloid protease by BMP-1 metalloprotease), the released mature MSTN protein dimer regulates the inhibition of skeletal muscle growth [22]. In more detail, the first cleavage occurs at the 266th position by Furin, followed by cleavage at 76th by BMP-1/Tolloid metalloproteinase, and finally, the released active MSTN protein dimer binds to the receptor (ActRIIB), resulted in inhibition of muscle growth [22-24]. Two representative cattle breeds, Belgian Blue, and Piedmontese, show natural mutations in this gene, 11 bps deletion, and one base mutation on mature MSTN domain, respectively. These cattle breeds phenotypically indicate that mature MSTN domain mutations contribute to muscle growth [5]. To mimic or reproduce those natural mutations using genome editing technologies, in a previous study, ZFNs disrupted exon1 locus (Signal Sequences region), subsequently, the mature MSTN domain was broken [25]. In another study, the mature domain locus via TALEN was directly targeted and mutated in microinjected embryos [19]. In both studies, the phenotype was observed after mutation of the mature locus of MSTN .
In our study it was assumed that disruption of the mature MSTNlocus might occur by applying effective sgRNA on the pro-peptide locus using CRISPR-Cas9. Microinjected embryos were transplanted, and the muscle outgrowing phenotype was observed in one calf (#17). Thus, we thought that gene editing on the pro-peptide locus region worked well and predicted that the sequence of mature domain locus might be mutated by CRISPR-Cas9. However, one interesting finding was observed as a result of sequencing. This was an in-frameshift (-12bps deletion) knockout in the target locus that did not disrupt the mature MSTNdomain region or amino acids of two cleavage regions. In other words, the 266th-, and 76th- amino acids for furin and proteinase were respectively conserved. Thus, the active MSTN protein dimer may be formed and muscle production is suppressed normally, and finally a wild type offspring should be born. Interestingly, the typical phenotyping (muscle outgrowth) was observed in #17 calf and the expression of MSTN mRNA was decreased (Figure 2). Because there have been no reports of phenotyping because of this type of mutation, it is hard to explain why this phenomenon occurred. One possibility is that the 156–160 position can be thought of as another molecular biological function in addition to the previously known two cleavage events. Similarly, in-frameshift mutation of the MSTNpro-peptide in mice showed a muscle gain phenotype [26]. Importantly, the blood test results of the mutated calves were normal (Table 1) and the calves showed no issue in their general health. In the future, we will monitor the growth, including germline transmission, and investigate how this mutation may have affected the function ofMSTN .
Microinjection commonly results in mosaic F0 founder animals that are then screened for the exact knockout/knockin in the F1 generation following subsequent breeding. This technique is very effective in rodent experiments but is not suited to cattle because of their long gestational periods and single pregnancies. A cattle F0 and F1 system would take more than 3 years and require high costs. Consequently, most genome edited cattle are produced using a SCNT approach. However, live, healthy calf offspring are limited when SCNT is employed because of abnormal reprograming during embryogenesis. In our study, microinjection was used to produce live, healthy genome edited calves. Randomly selected blastocysts were analyzed in vitro by sgRNA/Cas9 mRNA and a 81.3 ± 17.2% knockout efficiency rate was found. Embryo transfer was performed and a lower MSTN mutant cow generation rate of 17.6% was found in vivo. It is possible that non-mutated blastocysts were selected during the randomly selected process. In the future, to improve the efficiency of producing mutated offspring a portion of the blastocysts could be biopsied prior to transfer to identify possible mutations [27].
Genotyping analysis showed another interesting result. When mRNA of sgRNA and Cas9 was introduced into cells and embryos (blastocysts) various mutant pattern (-12, -10, -3, -2, -1, +1; Figure 1B) were shown, but only one mutant pattern (-12bps) was observed in genome edited calves. It is difficult to explain why only one pattern is observed in all MSTN mutated calves. One possible theory is that the cells with the other mutated pattern may be embryonic lethal at some time after the point of embryo transfer. Future studies will focus on improving in vitro and in vivo efficiency.
In conclusion, we demonstrated, for the first time, that microinjection of Cas9 mRNA and sgRNA for MSTN into in vitro fertilized embryos can produce live, genome edited, Korean beef calves—including one calf with biallelic mutation. These calves will be served as a model for the future development of CRISPR–Cas9 technology in the agricultural industries.