Base editing in bovine embryos reveals a species-specific role of SOX2 in regulation of pluripotency

The emergence of the first three lineages during development is orchestrated by a network of transcription factors, which are best characterized in mice. However, the role and regulation of these factors are not completely conserved in other mammals, including human and cattle. Here, we establish a gene inactivation system with a robust efficiency by introducing premature codon with cytosine base editors in bovine early embryos. By using this approach, we have determined the functional consequences of three critical lineage-specific genes (SOX2, OCT4 and CDX2) in bovine embryos. In particular, SOX2 knockout results in a failure of the establishment of pluripotency in blastocysts. Indeed, OCT4 level is significantly reduced and NANOG barely detectable. Furthermore, the formation of primitive endoderm is compromised with few SOX17 positive cells. RNA-seq analysis of single blastocysts (day 7.5) reveals dysregulation of 2074 genes, among which 90% are up-regulated in SOX2-null blastocysts. Intriguingly, more than a dozen lineage-specific genes, including OCT4 and NANOG, are down-regulated. Moreover, SOX2 level is sustained in the trophectoderm in absence of CDX2. However, OCT4 knockout does not affect the expression of SOX2. Overall, we propose that SOX2 is indispensable for OCT4 and NANOG expression and CDX2 represses the expression of SOX2 in the trophectoderm in cattle, which are all in sharp contrast with results in mice.


Introduction 45
Mammalian preimplantation development is characterized of the two earliest cell fate 46 decisions. The first cell fate decision gives rise to the inner cell mass (ICM) and the 47 trophectoderm (TE) and subsequently the ICM generates the primitive endoderm (PE) 48 and the epiblast (EPI) during the second cell fate decision. TE and PE will develop 49 into placenta and extra-embryonic cells, respectively, whereas the pluripotent EPI 50 contributes to the embryo proper (1, 2). The mechansims that regulate these events 51 have been mostly obtained from mouse model. Recent gene-expression and functional 52 analyses suggest that these mechanisms in the mouse may differ in other mammals, 53 including human and cattle. Investigation of these mechansims is important for 54 assisted reproductive technology, regenerative medicine as well as understanding 55 early embryonic mortality in human and agricultural animals. 56 The establishment and maintenance of pluripotency is regulated by a variety of 57 transcription factors, including core pluripotency factors, OCT4, SOX2 and NANOG 58 (1, 2). The functional importance and relationship of these core transcription factors 59 have been relatively well-characterized in mouse embryos. Interestingly, unlike the 60 universal expression pattern of OCT4 in morula, SOX2 is specifically restricted into 61 the inside cells of morula that become the ICM and is considered the earliest 62 pluripotency marker in mice (3). The HIPPO pathway plays a critical role in temporal 63 and spatial expression of SOX2 in mouse preimplantation embryos (3, 4). However, 64 SOX2 is not required for the first cell fate decision although SOX2-null mouse with base editor mRNA. Results showed injection of the base editor components did 105 not affect embryonic development to morula stage (Fig S2A and S2B). Using BE3, 106 results indicated successful editing of three, two and single target genes in 25.8%, 107 25.8% and 12.9% embryos, respectively ( Fig S2C and Table 1). For ABE7.10, results 108 indicated successful editing of three, two and single target genes in 23.3%, 56.7% and 109 20.0% embryos, respectively ( Fig S2D and Table 2). Taken together, these data 110 present proof-of-evidence of base editing with high efficiency in bovine embryos.  Fig S3A). Results shows the edited efficiency of sgRNA1 is 25.92% and 118 sgRNA2 reaches 74.07% (Fig S3B and S3C). Overall, premature stop condons were 119 successfully introduced in 77.8% (21 out of 27) embryos, suggesting gene disruption. 120 As a side-by-side experiment, immunostaining analysis confirmed that OCT4 can be 121 efficiently deleted in all blastomeres in these blastocysts (Fig S3D and S3E). Next, we 122 tested if OCT4-null embryos generated here recapitulate the phenotype of OCT4 123 knockout embryos produced via somatic cell nuclear transfer as reported previously 124 (21). Remarkbaly, NANOG is barely detectable in absence of OCT4 at blastocyst 125 stage (Fig S3D and S3F). The developmental potential to form blastocysts is grealty 126 inhibited in OCT4 KO groups (Fig S3G). These results are consistent with the 127 previous study. Thus, we establish a powerful and reliable system to accomplish 128 efficient base editing in bovine embryos, which will facilitate studies of gene 129 functions in bovine embryos.

130
Expression pattern of SOX2 protein in bovine early embryos 131 To functionally characterize SOX2 in bovine embryos, we first determine its 132 expression pattern in detail in bovine embryos. SOX2 was first found in the 8-cell 133 stage and continued to express thereafter (Fig 2A). In contrast to mouse embryos, SOX2 was not detected during oocyte maturation and the early development to the four-cell stage (Fig 2A). It is noteworthy that SOX2 gradually accumulates in the 136 ICM cells along with blastocyst expansion. Specifically, SOX2 was evenly distributed 137 in both TE and ICM in early blastocysts. Then, SOX2 was lost in subsets of TE cells 138 in middle blastocysts, and eventually restricted into ICM in late blastocysts (Fig 2B).

139
Quantitative results showed SOX2 level in TE is gradually diminished relative to the 140 one in ICM when the blastocyst is expanding (Fig 2C and 2D). Altogether, these data 141 indicate that SOX2 displays a different expression pattern in bovine embryos, which 142 suggests that there are differences in the regulatory mechanism of pluripotency 143 between species.

144
Effects of SOX2 KO on the bovine embryo development 145 We then sought to explore the functional role of SOX2 by disrupting its expression 146 using BE3 (Fig 3A and 3B). Genotyping results show sgRNA2 and 3 are more 147 efficient than sgRNA1 in editing SOX2 (Fig 3C and 3D). Overall, premature stop 148 condon was successfully installed at SOX2 in 87.1% (101 out of 116) bovine 149 blastocysts when these three sgRNAs were injected. Immunostaining results further 150 confirmed that SOX2 signal was drastically diminished in these corresponding 151 blastocysts and only 9 (out of 116) embryos display mosacism ( Fig 3E).

152
In vitro culture of embryos revealed no significant difference in the capability to 153 become blastocysts between SOX2 KO and WT groups ( Fig 3F). Interestingly, the 154 total cell number per blastocyst was significantly reduced at both E7.5 and E8.5 ( Fig   155   3G and 3H). However, the TE cell number (CDX2 positive) was not obviously 156 changed while the number of ICM cells (CDX2 negative) was rather decreased showed that WT and SOX2 KO blastocysts formed two distinguished clusters (Fig   165   S4C). There are a total of 2074 differentially expressed genes (DEGs, Fold changes 166 (FC) >2 or <0.5, P adjusted<0.05), among which 88.53% were remarkably 167 upregulated in SOX2 KO groups (Fig 4B and 4C).  Fig 4D). Surprisingly,

174
NANOG and OCT4 were both sharply downregulated in SOX2 KO bovine blastocysts.

175
In sum, these data suggest SOX2 plays a critical role in maintaining correct gene 176 expression of the pluripotency network.

SOX2 is indispensable for NANOG and OCT4 expression in the ICM of bovine
178 blastocysts 179 We then hypothesized that SOX2 is required for OCT4 and NANOG expression in    195 We next asked if CDX2 is involved in the gradual disappearance of SOX2 in the TE.

196
Both genotyping and immunostaining results indicated CDX2 is completely knocked 197 out in 82.9% embryos (58 out 70; Fig S6A-D). No difference was found on the 198 developmental potential to arrive blastocyst stage in CDX2 KO groups (Fig S6E). 199 Immunostaining analysis revealed that SOX2 signal sustained in the TE of bovine late 200 blastocysts in CDX2 KO groups (Fig 6A-C). To further test the specificity of the role 201 of CDX2 in regulating SOX2 expression during the bovine embryonic development, 202 we microinjected base editor components into one blastomere at 2-cell stage (Fig 6D).   Here, our studies reported that BE3 and ABE7.10 enable us to achieve gene editing 225 with an efficiency above 79% in bovine embryos. Importantly, we found no obvious 226 off-target editing at potential site, indicating the specific effects we documented in the 227 present study. To maximize the editing efficieny, we microinjected 2 or 3 sgRNA 228 together and found the target gene can be deleted completely in all blastomeres in 229 around 80% embryos with only less than 10% embryos exhibit mosaicism. We believe 230 this approach is a powerful tool to dissect gene function and produce genome-edited OCT4. However, we speculate SOX2 also directly regulate the expression of NANOG 245 because OCT4 is not completely lost when SOX2 is deleted (Fig 6G).

258
In conclusion, we demonstrated that the base editing system could be applied to  (Table S1). The DNA sequences were synthesized by 293 Sangon Co., LTD (Shanghai). Then, sgRNA DNA oligos were annealed and cloned 294 into a PX458 vector containing BpiI restriction sites with T7 promoter.

295
In vitro transcription 296 BE3 and ABE7.10 plasmids were purchased from Addgene (#73021 and #102919).   Table S3. 316 Single embryo was subject to whole-genome amplification by using REPLI-g Mini 317 Kits (QIAGEN, Cat. No. 150023). The target sites and 6 potential off-target sites 318 (Table S4) that predicted by an online software were amplified using PCR primers 319 with barcode sequence (Table S5). All amplicons were purified and subject to targeted 320 deep sequencing.  Table S6.  We thank all members of the K. Zhang laboratories for their helpful discussions. This            were detected). E and F. CDX2 KO has no effect on the rate of blastocyst formation 580 (Ten replicates of 20-25 embryos per group). Scale bar = 100 μm.