CRISPR/Cas9‐mediated gene knockout for DNA methyltransferase Dnmt3a in CHO cells displays enhanced transgenic expression and long‐term stability

Abstract CHO cells are the preferred host for the production of complex pharmaceutical proteins in the biopharmaceutical industry, and genome engineering of CHO cells would benefit product yield and stability. Here, we demonstrated the efficacy of a Dnmt3a‐deficient CHO cell line created by CRISPR/Cas9 genome editing technology through gene disruptions in Dnmt3a, which encode the proteins involved in DNA methyltransferases. The transgenes, which were driven by the 2 commonly used CMV and EF1α promoters, were evaluated for their expression level and stability. The methylation levels of CpG sites in the promoter regions and the global DNA were compared in the transfected cells. The Dnmt3a‐deficent CHO cell line based on Dnmt3a KO displayed an enhanced long‐term stability of transgene expression under the control of the CMV promoter in transfected cells in over 60 passages. Under the CMV promoter, the Dnmt3a‐deficent cell line with a high transgene expression displayed a low methylation rate in the promoter region and global DNA. Under the EF1α promoter, the Dnmt3a‐deficient and normal cell lines with low transgene expression exhibited high DNA methylation rates. These findings provide insight into cell line modification and design for improved recombinant protein production in CHO and other mammalian cells.

attachment regions, 10,11 ubiquitous chromatin opening elements [12][13][14] and insulators. 15 Transgene silencing is also associated with the methylated cytosine on the CpG sites of promoters in recombinant protein-producing CHO cells. 16 Yang et al 17 found an increase in the methylation of the human cytomegalovirus major immediate-early (hCMV-MIE) promoter, which controls the monoclonal antibody expression in unstable cell lines with low productivity. Moreover, they discovered that the lost productivity of specific mAb can be partially restored by treating these cells with 5-Aza-2-deoxycytidine (which is a DNA methylation inhibitor). These findings suggested that epigenetic factors, such as DNA methylation, possibly contribute to the loss of productivity and play important roles in regulating gene expression. 18 Therefore, elucidating the effects of DNA methylation and its mechanism on transgene expression would benefit the development of recombinant CHO cell lines that produce both high and stable production.
Previous studies on the effect of DNA methylation on the transgene expression in CHO cells have mainly focused on promoter modification, such as mutating the cytosine within promoters, 19 using CpG-free 20 or synthetic promoters 21 and inserting core CpG island elements into promoters. 22 Nonetheless, although a promoter without CpG dinucleotides could mitigate early gene silencing, it still cannot improve the long-term expression stability in transfected CHO cells. 20 DNA methylation is catalysed by DNA methyltransferases, including those which establish methylation (Dnmt3a and Dnmt3b) 23 and maintain methylation (Dnmt1). 18 In particular, Dnmt3a and Dnmt3b mediate epigenetic silencing through histone modification and subsequently DNA methylation. 24 As inhibitors, quinolone analogues are potent against human Dnmt3a, which can induce the re-expression of a reporter gene controlled by a methylated CMV promoter in leukaemia KG-1 cells. 25 Thus, we proposed that modifying the CHO cell line through the gene knockout of DNA methyltransferase would improve the expression stability of permanently transfected CHO cells during long-term cultivation. To validate this hypothesis, we knocked out the Dnmt3a gene through CRISPR/Cas9 genome editing technology to establish a Dnmt3adeficient CHO cell line and investigated the transgene expression level and stability in the deficient cells.

| sgRNA target design for Dnmt3a KO and plasmid construction
A DNA fragment 3a3 + 4 of Dnmt3a (NW_003613640.1) containing exon1 was amplified by polymerase chain reaction (PCR) from the genomic DNA of CHO-K1 cells using primers (Table 1). Two pairs of sgRNAs used for targeting the exon1 of Dnmt3a (Table 1) were designed and synthesized (TaKaRa, Dalian, China). If the first nucleotide was not G, we added an additional G nucleotide at the 5 0 end, because the U6 promoter prefers a G nucleotide for transcriptional initiation. The sense and antisense single-stranded oligos (100 mmol/L) were annealed in NEBuffer4 (New England Biolabs, Ipswich, MA, USA) by incubating the oligos mix at 95°C for 5 minutes. The annealed sgRNA oligos were cloned into pX330 vectors (Addgene, #42230) by conducting Bbs I digestion and ligation on the hybridized oligos, and the resulting CRISPR vectors were confirmed by sequencing and were referred to as pX330-3a1 and pX330-3a2, respectively. The plasmid was purified for cell transfection using an EndoFree Maxi Plasmid Kit (Tiangen, Beijing, China).   Briefly, the cells were seeded into 6-well plates at 1 9 10 5 cells/mL and allowed to grow for 2, 4, 6, 10, 24 and 48 hours at 37°C under 5% CO 2 prior to cell counting using a haemocytometer measurement tool and trypan blue staining assay. The t D value was determined by the formula t D = (t 2 Àt 1 ) 9 (log(2)/log(C 2 /C 1 )), where C 2 and C 1 are the cell concentrations at times t 2 and t 1 , respectively, under a constant growth rate.

| Long-term stability analysis for stably transfected cells
Stably transfected cell polycolonies were selected by adding Geneticin

| Statistical analysis
The data were expressed as means AE standard deviation. All statistical analyses were performed on the SPSS 18.0 (SPSS, Inc., Chicago, IL, USA). One-way ANOVA was used to compare multiple groups.
Paired-sample Student's t test was used for statistical analysis when only 2 groups were tested. P < .05 was considered statistically significant. All experiments were performed at least thrice, and all samples were tested in triplicate.

| Analysis of cells characteristics
The detection of cell proliferation and apoptosis indicated that Dnmt3a KO did not alter the cell morphology and the growth status    Dnmt3a-deficient cells transfected with CMV or EF1a had higher expression levels than those of the control normal CHO-K1 cells.

| Significant improvements by Dnmt3a KO in long-term expression stability
To

| DNA methylation
To determine whether DNA methylation affected the long-term trans- Rates of DNA methylation in the analysed 25 CpG sites of EF1a at the start and at high passage (P50) were low and had no differences in the transfected 3a-30 and normal CHO cell lines ( Figure 6B).

| DISCUSSION
Previous reports have shown that the DNA hypermethylation in the promoter region can cause instability in stable recombinant CHO cell lines. 16,17 Previous report showed that Dnmt3a knockout CHO cell lines were constructed using the CRISPR-Cas9 system without homologous regions. 28  The hCMV-MIE provides high levels of transgene expression; however, its loss of productivity has been reported with prolonged cultivation time. 37,38 hCMV-MIE is prone to transgene silencing, which is associated with DNA methylation. 16,17,37,39 Moreover, recombinant CHO cell lines with EF1a exhibited loss of productivity during long-term culture. 26,40 Therefore, to monitor the effects of are critical sites that prevent or inhibit transcription. 41 Compared with CMV, EF1a displayed decreased rates of DNA methylation, which presented no significant differences between the 3a-30 and normal control CHO cells at the start and at 50 passages, suggesting that the density of methylation in a given region determines whether or not transcription would be prevented. 42 The possible reason is that the effect of CpG methylation on gene expression is context dependent 43 ; that is, the degree of CpG methylation change, the density or the genomic location of CpG, methyl-CpG "readers" and TF binding are

CONFLI CT OF INTERESTS
The authors declare that they have no competing interests.