Genome stabilization by RAD51‐stimulatory compound 1 enhances efficiency of somatic cell nuclear transfer‐mediated reprogramming and full‐term development of cloned mouse embryos

Abstract Objectives The genetic instability and DNA damage arise during transcription factor‐mediated reprogramming of somatic cells, and its efficiency may be reduced due to abnormal chromatin remodelling. The efficiency in somatic cell nuclear transfer (SCNT)‐mediated reprogramming is also very low, and it is caused by development arrest of most reconstituted embryos. Materials and Methods Whether the repair of genetic instability or double‐strand breaks (DSBs) during SCNT reprogramming may play an important role in embryonic development, we observed and analysed the effect of Rad 51, a key modulator of DNA damage response (DDR) in SCNT‐derived embryos. Results Here, we observed that the activity of Rad 51 is lower in SCNT eggs than in conventional IVF and found a significantly lower level of DSBs in SCNT embryos during reprogramming. To address this difference, supplementation with RS‐1, an activator of Rad51, during the activation of SCNT embryos can increase RAD51 expression and DSB foci and thereby increased the efficiency of SCNT reprogramming. Through subsequent single‐cell RNA‐seq analysis, we observed the reactivation of a large number of genes that were not expressed in SCNT‐2‐cell embryos by the upregulation of DDR, which may be related to overcoming the developmental block. Additionally, there may be an independent pathway involving histone demethylase that can reduce reprograming‐resistance regions. Conclusions This technology can contribute to the production of comparable cell sources for regenerative medicine.


| INTRODUC TI ON
Mammalian oocytes endow terminally differentiated somatic cells with totipotency through the reprogramming that occurs somatic cell nuclear transfer (SCNT), which has resulted in the production of genetically identical animals of more than 20 species. 1,2 Moreover, SCNT can be applied to therapeutic cloning through the generation of immunocompatible pluripotent stem cells (PSCs) from patientderived somatic cells, which makes SCNT a promising technology for in vitro disease modelling and cell transplantation. 3 Various efforts have been undertaken to comprehensively recognize the pivotal molecular factors affecting successful generation of mammalian SCNT-cloned embryos, conceptuses and offspring. These factors determine among others (a) the provenance of nuclear donor cells, [4][5][6][7] (b) quality of nuclear recipient oocytes related to their meiotic, epigenomic and cytoplasmic maturity status. [8][9][10] The aforementioned factors also determine (c) epigenetic reprogrammability of donor cell nuclear genome in SCNT-derived oocytes and resultant embryos, 11-15 (d) intergenomic crosstalk between nuclear and mitochondrial compartments in SCNT-derived oocytes and cloned embryos [16][17][18][19][20] and (e) the incidence of apoptosis-or autophagydependent events in the ex vivo-expanded nuclear donor cells and cultured cloned embryos. [21][22][23] Although SCNT presents tremendous potential for practical application in various areas, there are technical limitations that cause poor embryonic development, which has been largely attributed to the abnormal epigenetic status of the transplanted nuclei derived from somatic cells. 24 In fact, the success rate of SCNT is mainly determined by the efficiency with which enucleated oocytes reprogramme the epigenetic identity of donor somatic nuclei before the onset of zygotic genome activation (ZGA), which occurs at the 2-cell stage in mice and the 8-cell stage in humans. [25][26][27] In mice, the treatment of SCNT-derived embryos with histone deacetylase inhibitors such as TSA significantly improves their development and epigenetic status. 28 Furthermore, the application of another histone-modifying enzyme, lysine-specific demethylase 4D (KDM4D), for the reprogramming of somatic nuclei overcomes epigenetic barriers to embryo development at the 2-cell stage through the regulation of the H3K9 methylation of donor nuclei. 25 Additionally, the overexpression of the H3K27me3-specific demethylase KDM6A, but not KDM6B, improves the efficiency of mouse SCNT. 29 Recently, the derivation rate of human SCNT-PSCs has been greatly increased by the injection of KDM4A mRNA to reduce H3K9me3 activity, but this methodology still does not seem to be as efficient as possible. 27 Another mechanism that is highly likely to be associated with SCNT efficiency is genome stability. In another type of reprogramming procedure, the generation of induced pluripotent stem cells (iPSCs) is negatively related to the activation of TP53 by damaged DNA and the introduction of transcription factors (TFs) increases the phosphorylated form of H2AX (γH2AX) foci, which are markers of DNA double-and single-stranded breaks. 30,31 Additionally, it was recently reported that significant genomic DNA breaks occur when the somatic cell nucleus is remodelled through chromosome condensation prior to entry into S-phase. 32 Similarly, γH2AX foci are detected at the time of paternal DNA demethylation, 33 suggesting that genome stability is challenged by DNA breakage during zygotic reprogramming. Given the potential for genetic instability to impede embryonic development after SCNT, it can be postulated that increasing genome stability is another approach for enhancing the reprogramming efficiency of SCNT.
Rad51 homologous 1 (RAD51) is a DNA-binding protein that has pleiotropic functions in maintaining genome stability and a regulating protein to control the DNA damage response (DDR), homologous recombination, DNA replication and repair. 34 In fact, cells with spontaneous DNA damage display high expression of rad51, which is induced by multiple types of genotoxic stress. 35 Furthermore, during the development of porcine parthenotes, inhibiting RAD51 induces apoptosis, reactive oxygen species accumulation, an abnormal mitochondrial distribution, delayed division during development and inhibition of development to the blastocyst stage. 36 In 2017, Chia et al. 32 found that genetic instability marked by frequent chromosome segregation errors and DNA double-strand breaks (DSBs) arose prior to transcriptional activity during human and mouse SCNT reprogramming, and the absence of the repair this damage resulted in delayed DNA replication and severely abnormal mitosis. On the basis of these results, we hypothesized that the repair of genetic instability or DSBs during SCNT reprogramming plays an important role in embryonic development.
In the present study, we found that increased RAD51 activity induced by supplementation with RAD51-stimulatory compound 1 (RS-1), a homology-directed repair (HDR) enhancer, increases genomic remodelling during the early phase of mouse SCNT and subsequently enhances reprogramming efficiency. This approach may contribute to the reprogramming of somatic cells independent of previous epigenetic modifications, and the use of both approaches as a combined protocol will have a synergistic effect in improving the efficiency of reprogramming by SCNT.

| Mice
Eight-to 10-week-old female B6D2F1 mice (Orient-Bio Inc) were used both for the collection of recipient oocytes and as SCNT donors. Eight-to 10-week-old female ICR mice were used as the foster mothers in embryo transfer. To induce pseudopregnancy, these mice were mated with vasectomized male mice of the same strain.
The protocols for the use of animals in these studies were ap-

| Preparation of oocytes and nuclear donor cells
Both oocytes and cumulus cells were prepared by the superovulation of 8-to 10-week-old B6D2F1 female mice. Superovulation was induced by sequential injection 5 IU of pregnant mare serum gonadotropin (Sigma-Aldrich), followed by 5 IU of human chorionic gonadotropin (hCG, Sigma-Aldrich) with an interval of 48 hours.
Cumulus-oocyte complexes (COCs) were collected from oviducts in M2 medium (Sigma-Aldrich) at 14 hours after hCG injection and were treated with M2 containing 0.1% bovine testicular hyaluronidase (Sigma-Aldrich) to obtain dissociated cumulus cells and oocytes.
The cumulus-free oocytes were then cultured in potassium simplex optimized medium (KSOM; Millipore) at 37°C under 5% CO 2 until further use. Dispersed cumulus cells were dissociated by hyaluronidase treatment, diluted in M2 medium and collected. The pellet was then resuspended in a small volume of polyvinylpyrrolidone (PVP) in M2 medium in a manipulator chamber.

| Preparation of Kdm4a mRNA
In vitro transcription was performed as described previously. 27 In brief, full-length mouse Kdm4a/Jhdm3a cDNA was cloned into a pcDNA3.1 plasmid containing poly(A)83 at the 3′ end of the cloning site by using an In-Fusion Kit (#638909, Clonetech). Messenger RNA was synthesized from the linearized template plasmids by in vitro transcription using a mMESSAGE mMACHINE T7 Ultra Kit (#AM1345, Life Technologies). The synthesized mRNA was dissolved in nuclease-free water. The concentration of mRNA was measured using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies); aliquots of mRNA were stored at −80°C until use.

| Mouse SCNT procedure and RS-1 treatment
All MII stage oocytes with distinct first polar bodies were enucleated in M2 medium containing 5 µg/mL cytochalasin B. For nuclear transfer, cumulus cells were injected into enucleated oocytes in M2 medium using a PIEZO-driven micromanipulator (Primetech). After nuclear transfer, the reconstructed oocytes were incubated in KSOM medium for 1-2 hours before activation. Activation was performed by incubation in M16 (Millipore) medium containing 10 mmol/L SrCl 2 , 2 mmol/L EGTA, and 5 µg/mL cytochalasin B for 6 hours, and then, the oocytes were cultured in KSOM in a humidified atmosphere of 5% CO 2 at 37°C. During activation, the treated groups were supplemented with RAD51-stimulatory compound 1 (RS-1; R9782, Sigma-Aldrich), which was continued for 22 hours after activation, until the 2-cell stage. In preliminary experiment, embryonic development was  Figure 1A,B). The embryonic development of cloned embryos was assessed for 5 days (120 hours) after activation.

| In vitro fertilization and embryo culture
Sperm masses were collected from the cauda epididymis and placed in drops of incubated HTF (Millipore) medium covered with mineral oil. The spermatozoa were capacitated for 1 hour under 5% CO 2 at 37°C before being used for insemination. As an experimental control, MII oocytes containing COCs were collected from B6D2F1 female mice that had undergone superovulation and were inseminated with preincubated spermatozoa and maintained under 5% CO 2 at 37°C.
The final concentration of spermatozoa in the insemination medium was ~150 spermatozoa/µL. Approximately 5-6 hours after insemination, zygotes were transferred to KSOM for further cultivation.

| Embryo transfer experiment
Somatic cell nuclear transfer embryos cultured to the 2-cell stage under RS-1 treatment or non-treatment conditions were transferred into the oviduct of pseudopregnant female ICR mice that had been mated with a vasectomized male the night before transfer (0.5 days postcoitum [dpc]). Caesarean section was carried out at day 19.5 dpc and the surviving pups were fostered by lactating ICR females.

| Array based comparative genomic hybridization
The ESCs and PSCs derived from experimental groups (IVF, SCNT and SCNT-RS-1) were washed in PBS prior to loading into a PCR tube containing lysis buffer. Genomic DNA was isolated by using a DNA multisample kit (Thermo Fisher Scientific) and stored at −80°C until being used in experiments. Whole-genome amplification was performed using the GenomePlex Whole-Genome Amplification (WGA) kit (Sigma-Aldrich) to achieve the representative amplification of genomic DNA. Array based comparative genomic hybridization (aCGH) was performed using a SureScan unrestricted high-definition CGH microarray (Agilent Technologies). After the scanning the microarray slides, the images were analysed using the Feature Extraction software provided by Agilent as both a standalone programme and an integral component of CytoGenomics software (Agilent).

| Immunostaining
One-cell and 2-cell cloned embryos were washed in PBS containing 0.1% polyvinyl alcohol (PVA; Sigma-Aldrich) and then fixed with 4% The percentage of embryos with RAD51 foci at the 1-cell (left) and 2-cell stages (right) cloned embryos in the IVF, SCNT, and SCNT-RS-1 groups ( * P < .05). This experiment was repeated more than three times and the number of embryos was more than 10 (per each group). E, DNA breaks with the average comet tail lengths in G1-phase (5 h) and S-phase (8 h) 1-cell stage embryos from the IVF, SCNT, and SCNT-RS-1 groups determined through the comet assay. This parameter essentially represents the product of the percentage of the total DNA in the tail and the distance between the centres of mass of the head and tail regions ( ** P < .01). F, Representative photographs of the comet assay results at G1-phase and S-phase in the IVF, SCNT, and SCNT-RS-1 groups diluted in PBS-0.1% BSA buffer. The same conditions were maintained between different samples.

| Comet assay
DNA damage in the 1-cell embryos generated from IVF or SCNT was assessed at different time (5 and 8 hours after insemination or SCNT) by DNA comet length using a neutral version of the comet assay, 37,38 as detailed in the manufacturer's instructions of the CometAssay ® Kit (Trevigen, Inc).
To maximize cell lysis efficiency, the zona pellucida of the embryos was removed. After the embryos were suspended in 1% low-meltingpoint agarose (LMA) at 37°C, the LMA-embryo mixture was pipetted onto slide provided with the CometAssay ® Kit. The slides were incubated at 4°C in the dark to achieve complete adherence and then im-

| Quantitative reverse transcriptionpolymerase chain reaction (RT-PCR)
Total RNA of each group was extracted from twenty 1-, 2-or 4cell stage embryos using a TRIzol reagent (Invitrogen). cDNA was synthesized from total extracted RNA using the LeGene Premium

RNA sequencing
For the RNA-Seq analysis, the DriveMap TM library preparation method (Selecta) was used. For cDNA synthesis, each embryo was lysed in 1× TCL lysis buffer, deposited in separate wells of a TurboCapture 96 mRNA plate (Qiagen) and incubated at room temperature for 1 hour After washing the plate three times with cold TCW washing buffer (Qiagen), 10 µL of RT reaction master mix solution was added. The plate was then subjected to incubation at 50°C for 40 minutes, followed by RT inactivation at 95°C for 5 minutes.

| Analysis of RNA-seq data
RNA-Seq data were analysed by using either the ROSALIND tool P-values were calculated with right-sided hypergeometric tests.
Benjamin-Hochberg adjustment was used for the correction of multiple tests. KEGG pathways with P-values <.05 were considered statistically significant.

| Statistical analysis and reproducibility
All data presented with error bars came from consist of at least with Duncan's test using SAS software, while implantation and ESC-derivation rates were analysed with the chi-square test. P < .05 was regarded as statistically significant.

| RS-1 recovered the developmental insufficiency of SCNT-derived mouse embryos
To confirm the insufficiency of embryonic development from SCNT-derived embryos, the rates of blastocyst formation and RAD51 activity (DDR ) were compared in SCNT-and IVF-derived embryos. As shown in Figure S1, unlike IVF embryos, the majority of  Figure 1A,B). However, there was no significant difference in the total cell number and inner cell mass number between blastocysts in the SCNT and SCNT plus RS-1-treated (SCNT-RS-1) groups ( Figure S2).
Based on a previous report, 32 G1-phase, S-phase and G2-phase 1-cell embryos were harvested at 5, 8 and 12 hours post-activation, respectively. To determine the differences in chromatin structure during reprogramming, the numbers of γH2AX spots (a marker of DSB) and RAD51 spots were analysed in IVF-derived and SCNTderived 1-cell embryos (at G2-phase). Interestingly, the number of γH2AX spots in SCNT embryos was much lower than that in IVF embryos, and the RS-1 treatment of SCNT embryos (SCNT-RS-1 group) increased γH2AX spots compared to non-treated SCNT embryos (SCNT group) ( Figure 1C). In most cases, the number of RAD51 spots showed similar patterns to those of γH2AX and Rad51 spots overlapped with γH2AX spots. In addition, the number of SCNT embryos with RAD51 spots was lower than that of IVF embryos, and this number was greatly increased in the SCNT-RS-1 group ( Figure 1D).
To compare the genomic DNA instability of the IVF, SCNT and SCNT-RS-1 group during reprogramming, the frequency of DNA fragmentation was analysed by applying the comet assay. The lengths of the DNA fragments (comet tail lengths) observed at G1-phase in the 1-cell embryo were very similar among the IVF, SCNT and SCNT-RS-1 groups. However, the comet tail length at S-phase was clearly distinguishable between IVF embryos and SCNT embryos. SCNT embryos at S-phase clearly showed notably longer comet tails than IVF embryos, indicating that the genomic DNA of SCNT embryos was severely damaged in S-phase of the 1-cell stage. In contrast, the comet tail lengths of the SCNT-RS-1 group were similar or even shorter than those of the IVF samples ( Figure 1E,F, Figure S3). These observations suggest that RS-1 may have resolved the DNA damage stress caused in SCNT embryos by progression through S-phase of the 1-cell stage.

| Improved embryonic development of SCNTderived mouse eggs may be related to changes in gene expression induced by RS-1 treatment
Next, we employed a single-cell RNA-Seq approach to examine the changes in gene expression induced by RS-1 treatment. To that end, at least 10 2-cell embryos from the SCNT, SCNT-RS-1 or IVF groups were individually subjected to RNA-Seq analysis. Our analyses revealed that RS-1 treatment changed the gross gene expression pattern in SCNT-2 cells to a pattern close to that in IVF 2 cells (Figure 2A). In a non-supervised hierarchical analysis, five major gene clusters were identified and subjected to GO analysis.
As shown in Figure 2B However, as indicated in cluster 4, RS-1 by itself is insufficient to activate entire transcriptional programme for early embryonic development. Next, to examine biological functions altered by RS-1, KEGG pathway analysis was performed. As shown in Figure 2C, the spliceosome, RNA transport and mRNA surveillance pathways were markedly changed by RS-1. Oocyte meiosis-and cell cyclerelated genes were also restored by RS-1. The spliceosome and cell cycle pathways were included in the genes of cluster 2 as well ( Figure 2D). Finally, given that epigenetic regulators (ERs)/TFs regulate ZGA, 43 Figure 3A,B, a higher derivation rate of mESCs was obtained in the SCNT-RS-1 group than in the RS-1 non-treated (SCNT) and IVF groups (46 ± 3.0% vs 17 ± 0.3% and 24 ± 15.9%, respectively, P < .05). In addition, treatment of RS-1 improved the derivation efficiency on the IVF embryos and it may also have a beneficial effect on the development and ESC-derivation on IVF embryos (44 ± 6.0% vs 28 ± 5.8%, P <.05) ( Figure S4).
To analyse the long-term safety of RS-1 treatment during embryonic development, we performed array comparative genomic hybridization (CGH) to identify genetic variations in ESCs derived from the IVF, SCNT and SCNT-RS-1 groups. As shown in Figure S4, in ESCs from both the SCNT and SCNT-RS-1 groups, the gain and loss of copy number variation (CNV) during reprogramming and prolonged periods in culture were not different compared to those in the IVF group ( Figure S5). In particular, these data may show that there was no harmful effect of RS-1 treatment on genetic stability because there was no difference between the RS-1-treated and non-treated groups. In addition, to compare the embryonic quality of cloned embryos according to RS-1 treatment, we analysed the delivery rate after embryo transfer. The number of offspring in the SCNT-RS-1 group was significantly increased compared to that in the SCNT group (5.2 ± 1.3% vs 0.4 ± 0.4%, P < .05; Figure 3C,D). All pups were normal morphology, healthy and well-grown into adult. First, we performed another RNA sequencing analysis using 2cell embryos (50 embryos/sample) from the SCNT and SCNT-RS-1 groups, and the data were analysed together with previous RNA-seq results from the SCNT and SCNT-Kdm4a injected (SCNT-Kdm4a) groups. 27 As shown in the Figure 4A (Table S1). In addition, in another round of experiments, the rate of 2-cell block in the SCNT-RS-1 and SCNT-Kdm4a groups were significantly lower than that in the SCNT group (7.8 ± 0.4% and 10.4 ± 3.3% vs 43.2 ± 3.7%, P < .05). None of the reconstituted eggs showed a 2-cell block in the SCNT-RS-1+Kdm 4a group (P < .01). Additionally, embryonic development up to the blastocyst stage was greatly increased in the SCNT-RS-1-Kdm 4a group ( Figure 4C,D). This experiment was repeated three time (≥20 SCNT embryos/each group/trial). These results may suggest that the two treatments have different modes of action and exert synergistic effects to overcome the developmental block of mouse SCNT eggs. were increased in the treated SCNT eggs and γH2AX spots were also increased ( Figure 1A-D). The γH2AX foci that form at sites of DNA damage serve to recruit repair proteins and it has been suggested that this leads to recombination and conformational changes in chromatin. [47][48][49] These results may suggest that RS-1 upregulates RAD51 activity, which greatly increases the genomic . This experiment was repeated more than 3 times and the number of embryos was more than 20 (per each group) modification of nuclear donors by inducing DSBs and HDRs when exposed to oocyte cytoplasm. Similarly, IVF embryos exhibit a greater number of DSBs by the 1-cell stage than parthenogenetic and SCNT embryos. 50 Additionally, high HR-related gene expression in iPSCs has been achieved through reprogramming. 51,52 Therefore, these results may suggest that successful reprogramming during fertilization and the dedifferentiation of somatic cells in oocyte cytoplasm require a proper DSBs in the genome and its repair system.

| RS-1 overcomes the developmental block of mouse SCNT eggs through a mechanism independent of epigenetic regulators
It is well established that reprogrammed human PSCs contain a number of protein-coding mutations, which may be preexisting mutation or arise during reprogramming process. 53,54 To test the safety of RS-1 treatment on SCNT reprogramming, we analysed chromosome instability using the comet assay in SCNT eggs and array CGH in SCNT-PSCs and then compared the results with those for IVF eggs and IVF-ESCs. In 1-cell eggs, the comet tail lengths of the three groups (IVF, SCNT and SCNT-RS-1) were not different in G1phase, but they were greatly increased in the S-phase of SCNT eggs compared to those of IVF and SCNT-RS-1 eggs ( Figure 1E). These data may suggest that a large amount of genetic instability exists due to DNA damage that occur during SCNT reprogramming, which is a reason for the developmental arrest observed in SCNT eggs. DNA damage can be greatly reduced by treatment with RS-1 ( Figure 1E and Figure S2). However, we did not find much differences in CNVs in ESCs and PSCs among all three groups ( Figure S3). This result is very similar to previous findings indicating that CNVs and InDels in human SCNT-PSCs and IVF-ESCs show no statistically significant differences. 55 In addition, it may prove that the increase in RAD51 activity induced by RS-1 treatment during SCNT reprogramming has no effect on genetic stability after embryonic development and stem cell derivation.
RS-1-treated SCNT eggs have been shown to overcome developmental arrest and present good embryonic development, similar to that of IVF-derived embryos. Following RNA-seq analysis using a single embryo, RS-1-treated SCNT-2-cell embryos showed the reactivation of a large number of genes that were not expressed in SCNT-2-cell embryos, and their expression patterns were also much more similar to those of IVF embryos (Figure 2A). KEGG analysis suggested that upregulation of RNA and protein processing-, cell cycle-and epigenetic regulation-related genes in clusters 3 and 4 ( Figure 2B). In addition, RS-1-treated SCNT embryos showed a higher derivation rate of PSCs than non-treated SCNT embryos in vitro as well as greater production of viable offspring in vivo (Figure 3). Therefore, we suggest that treatment with RS-1 during SCNT reprogramming may upregulate the expression of previously repressed genes by recovering RAD51 activity for HDR-related DNA repair.
In our previous report, we found insufficient gene expression (referred to as reprograming-resistant regions, RRRs) due to the presence of epigenetic barriers such as histone H3 lysine 9 trimethylation (H3K9me3) in SCNT embryos, and we found that the reduction of barrier activity through the ectopic expression of the histone demethylase KDM4A could fully or partially activate some genes of RRRs and greatly improve SCNT embryo development. 27 In the present study, we performed comparative studies of embryos derived from two systems, RS-1-mediated and Kdm4amediated SCNT, to understand their mode of action. In the first study, the gene expression of 2-cell embryos from the SCNT-RS-1 and non-treated SCNT control groups was analysed by RNA sequencing, and the results were compared with those from SCNT-Kdm4a-injected (SCNT-Kdm4a) and non-injected SCNT control groups. 27  In the present study, we observed that transcriptional activity has been altered during SCNT reprogramming and it also may be caused by a lower activity of DNA repair system, resulted in developmental block on further embryonic development. And, we determined the effect of upregulation of RAD51 by treatment with RS-1 on the role of DSB repair during SCNT reprogramming and the improvement of embryonic development by recovery of gene expression. Therefore, we suggest that RS-1 treatment would be an efficient, safe and simple protocol for improving the efficacy of SCNT technology.

CO N FLI C T O F I NTE R E S T
The authors indicate no potential conflicts of interest.

AUTH O R CO NTR I B UTI O N S
DRL and KSP conceived the project. ARL, SHS, KSP and DRL designed the experiments. ARL performed mouse SCNT experiments.