Protein arginine methyltransferase 7–mediated microRNA-221 repression maintains Oct4, Nanog, and Sox2 levels in mouse embryonic stem cells

The stemness maintenance of embryonic stem cells (ESCs) requires pluripotency transcription factors, including Oct4, Nanog, and Sox2. We have previously reported that protein arginine methyltransferase 7 (PRMT7), an epigenetic modifier, is an essential pluripotency factor that maintains the stemness of mouse ESCs, at least in part, by down-regulating the expression of the anti-stemness microRNA (miRNA) miR-24-2. To gain greater insight into the molecular basis underlying PRMT7-mediated maintenance of mouse ESC stemness, we searched for new PRMT7-down-regulated anti-stemness miRNAs. Here, we show that miR-221 gene–encoded miR-221-3p and miR-221-5p are anti-stemness miRNAs whose expression levels in mouse ESCs are directly repressed by PRMT7. Notably, both miR-221-3p and miR-221-5p targeted the 3′ untranslated regions of mRNA transcripts of the major pluripotency factors Oct4, Nanog, and Sox2 to antagonize mouse ESC stemness. Moreover, miR-221-5p silenced also the expression of its own transcriptional repressor PRMT7. Transfection of miR-221-3p and miR-221-5p mimics induced spontaneous differentiation of mouse ESCs. CRISPR-mediated deletion of the miR-221 gene, as well as specific antisense inhibitors of miR-221-3p and miR-221-5p, inhibited the spontaneous differentiation of PRMT7-depleted mouse ESCs. Taken together, these findings reveal that the PRMT7-mediated repression of miR-221-3p and miR-221-5p expression plays a critical role in maintaining mouse ESC stemness. Our results also establish miR-221-3p and miR-221-5p as anti-stemness miRNAs that target Oct4, Nanog, and Sox2 mRNAs in mouse ESCs.

Embryonic stem cells (ESCs) 3 can be derived from inner cell masses of blastocysts (1) and are defined by two main characteristics: 1) long-term self-renewal and 2) the ability to form all three germ layers and differentiate into all kinds of different cell types (pluripotency) (2). These stemness characteristics of ESCs are maintained by pluripotency transcription factors (e.g. Oct4, Nanog, and Sox2) and their regulatory networks (3,4). These factors co-occupy and activate their own genes and other numerous genes important for maintaining ESC pluripotency (e.g. Oct4, Nanog, Sox2, STAT3, and Zic3) while repressing lineage-specific transcription factor genes (e.g. Hox clusters, Pax6, and Meis1) to prevent ESC differentiation (5,6).
In addition to transcription factors, microRNAs (miRNAs) regulate pluripotency (7)(8)(9). miRNAs are small single-stranded RNAs of 21-25 nucleotides that negatively regulate gene expression. The miRNA-mediated targeting of mRNAs induces post-transcriptional repression through Argonaute-2mediated mRNA degradation, translational repression, and mRNA deadenylation (10 -13). miRNA-mediated regulation of stemness often results from changes in miRNA levels between the ESC state and the differentiated state (14,15). For example, miR-27a has been identified as a differentiation-associated miRNA that is induced during ESC differentiation and directly targets the pluripotency factor Foxo1 and signal transducers (gp130 and smad3) to inhibit ESC pluripotency (16).
The expression patterns of miRNAs can be highly cell typespecific and thus are important to regulating cellular differentiation and development (17)(18)(19). It has been shown that the expression of miRNAs is regulated by multiple different mechanisms, including transcriptional control, epigenetic modulation, and post-transcriptional regulation (20,21). Dysregulation of miRNAs is linked to cancer and other diseases (22,23). For instance, the expression of multiple miRNAs (e.g. miR-124, miR-34, miR-9, and miR-200 families) is silenced by DNA hypermethylation in many types of cancer (24,25).
We have reported previously that protein arginine methyltransferase 7 (PRMT7), a transcriptional co-repressor, is essential for maintaining mouse ESC stemness. In the same study, we showed that miR-24-3p and miR-24-2-5p levels are highly up-regulated by PRMT7 knockdown and are increased during mouse ESC differentiation (26). We also characterized miR-24-3p and miR-24-2-5p as anti-stemness miRNAs that can induce mouse ESC differentiation and directly inhibit the expression of the major pluripotency factors Oct4, Nanog, Sox2, Klf4, and c-Myc (26). We further showed that PRMT7-mediated repression of the expression of the miR-24-2 gene encoding miR-24-3p and miR-24-2-5p is required for maintaining mouse ESC stemness.
To better understand how PRMT7 maintains mouse ESC stemness, we sought to identify new anti-stemness miRNAs that are repressed by PRMT7. We thus re-analyzed our previous miRNA expression profile data of control and PRMT7depleted mouse ESCs to determine which miRNAs in mouse ESCs are highly up-regulated by PRMT7 knockdown. We found that miR-221-3p and miR-221-5p act as anti-stemness miRNAs by targeting the 3Ј untranslated regions (3ЈUTRs) of mRNA transcripts of the major pluripotency factors Oct4, Nanog, and Sox2. Our results also revealed that negative regu-lation of miR-221-3p and miR-221-5p expression by PRMT7 is necessary to maintain ESC pluripotency.

Anti-stemness function of miR-221
To determine whether miR-221 expression is directly repressed by PRMT7, we performed quantitative chromatin immunoprecipitation (ChIP) experiments. ChIP results showed that PRMT7 occupied the promoter region in the miR-221 gene in V6.5 mouse ESCs (Fig. 1, C and D). It has been generally accepted that PRMT7 monomethylates arginine residues, such as H4R3, for gene repression (27). We and others have also shown that PRMT7 represses gene expression by indirectly establishing symmetric dimethylation at H4R3 (H4R3me2s) (26 -29). We therefore examined the effect of PRMT7 knockdown on monomethylated H4R3 (H4R3me1) and H4R3me2s levels at the miR-221 promoter. Our results showed that H4R3me1 and H4R3me2s levels at the miR-221 promoter were decreased by PRMT7 depletion (Fig. 1E). Together, these results indicate that PRMT7 represses the expression of miR-221, at least in part, by up-regulating repressive histone marks (e.g. H4R3me1 and H4R3me2s) at the miR-221 promoter in V6.5 mouse ESCs.

Anti-stemness function of miR-221 CRISPR-mediated deletion of the miR-221 gene impedes spontaneous differentiation of PRMT7-depleted mouse ESCs
To confirm that the repression of miR-221-3p and miR-221-5p expression by PRMT7 is required for mouse ESC stemness, we sought to determine whether miR-221 loss impedes the spontaneous differentiation of PRMT7-depleted mouse ESCs. To generate miR-221-null mouse ESCs, we used CRISPR-Cas9, which is a powerful tool for genome editing in living cells (32). Specifically, we used a double nicking strategy involving two guide RNAs and Cas9-D10A nickase (a mutant form of the RNA-guided, double-strand cleaving DNA endonuclease Cas9) because this strategy may introduce DNA doublestrand breaks with significantly increased specificity around the target region (33). For this double nicking strategy, we first cloned two 20-bp-long single guide RNAs (sgRNAs) into a Cas9-D10A nickase expression plasmid that was used to target the mouse miR-221 gene (Fig. 6A). We then transfected these plasmids into V6.5 mouse ESCs and screened mouse ESC colonies to obtain miR-221-null clones. Because V6.5 mouse ESCs were derived from a male mouse embryo and the miR-221 gene is located on chromosome X, one allele of chromosome X needs to be deleted for the generation of miR-221-null mouse ESCs. Our genomic PCR results demonstrated that the miR-221 gene was deleted in two clones (miR-221-KO-4 and miR-221-KO-7) (Fig. 6B). DNA sequencing of genomic PCR products also confirmed the knockout of miR-221 (Fig. 6C). Furthermore, miRNA-specific quantita-

Discussion
In the present study, the anti-stemness functions of miR-221-3p and miR-221-5p are supported by several lines of evidence. We showed that transfection of miR-221-3p or miR-

Anti-stemness function of miR-221
221-5p mimics into two different ESC lines V6.5 and R1 caused the stemness loss of these ESC lines and that the loss of the miR-221 gene inhibited the spontaneous differentiation of PRMT7-depleted mouse ESCs. We also found that the expression levels of miR-221-3p and miR-221-5p were lower in mouse ESCs than in differentiated somatic cells (3T3 fibroblasts) (data not shown) and were increased during RA-induced differentiation. The results of our reporter assay, in combination with mutagenesis, showed that miR-221 miRNAs can target the 3ЈUTRs of the major pluripotency factors Oct4, Nanog, Sox2, Klf4, and Prmt7, indicating that miR-221 acts as an anti-stem-ness miRNA by targeting the mRNAs of multiple pluripotency genes, including Oct4, Nanog, Sox2, and Prmt7 (Fig. 4). In line with this, LNA-miRNA-mediated inhibition or CRISPR-mediated deletion of miR-221-3p and miR-221-5p restored the expression of Oct4, Nanog, Sox2, and Prmt7 in PRMT7-depleted mouse ESCs (Figs. 5 and 7).
The differentiation of ESCs requires both the down-regulation of pluripotency factors and the up-regulation of lineagespecific markers (35,36). In this respect, our results showed that miR-221 loss in PRMT7-depleted mouse ESCs not only recovers the expression of pluripotency markers (e.g. Oct4,

Anti-stemness function of miR-221
Nanog, and Sox2) but also down-regulates the expression of mesoderm and endoderm markers (Fig. 7). Interestingly, miR-221 is up-regulated in fully differentiated neurons (37) and plays a role in neuron differentiation (38). Therefore, it is likely that miR-221 not only antagonizes the stemness of mouse ESCs but also positively regulates terminal differentiation of certain types of stem cells. Similar to miR-221, there are many other anti-stemness miRNAs that inhibit ESC stemness and facilitate cell differentiation. For example, human miR-145 is an antistemness miRNA that promotes endoderm and ectoderm differentiation by targeting pluripotency factors, such as Oct4, Sox2, and Klf4 (39). In addition, miR-9 promotes the differentiation of neural stem cells while inhibiting their proliferation (40).
It has been shown that the miR-221 gene is transcriptionally regulated by several transcription factors. Estrogen receptor-␣ binds to the miR-221 promoter to repress miR-221 expression (41). The transcription factor FOSL1 activates miR-221 expression through its interaction with the miR-221 promoter in breast cancer cells (42). Fornari et al. (43) showed that p53 can regulate miR-221 levels in hepatocellular carcinoma. In the current study, our results indicate that PRMT7 binds to the miR-221 promoter and increases the repressive epigenetic marks H4R3me1 and H4R3me2s to down-regulate miR-221 expression. Thus, our findings provide a new miR-221 regulatory mechanism in which the expression of miR-221 is epigenetically repressed by the arginine methyltransferase PRMT7 in mouse ESCs.
Our results showed that miR-221-5p targeted its own repressor PRMT7 in addition to the well-known pluripotency factors Oct4, Nanog, and Sox2 (Fig. 4D), whereas PRMT7 directly repressed the miR-221 gene (Fig. 1). These results indicate a mutually antagonistic relationship between miR-221-5p and PRMT7. Interestingly, we have previously reported an additional antagonistic relationship between PRMT7 and miR-24-2 (i.e. miR-24-3p and miR-24-2-5p) (26). In the same study, we have also demonstrated that mouse ESC stemness requires PRMT7-mediated repression of miR-24-2 expression (26). Unexpectedly, our results in the current study showed that the CRISPR-mediated deletion of the miR-221 gene alone was sufficient to block the spontaneous differentiation of PRMT7-depleted mouse ESCs (Fig. 7A), suggesting the possibility that the PRMT7-mediated repression of the miR-221 gene plays a more predominant role in maintaining mouse ESC stemness than does the PRMT7-mediated repression of the miR-24-2 gene. However, miR-24-3p and miR-24-2-5p levels in PRMT7-depleted mouse ESCs were highly decreased by miR-221 loss (Fig.  7E). In addition, increased levels of miR-24-3p or miR-24-2-5p via transfection of miR-24-3p or miR-24-2-5p mimics induced spontaneous differentiation of mouse ESCs, as shown in our previous study (26). Therefore, it is likely that the repressed states of both miR-24-2 and miR-221 genes are critical for maintenance of mouse ESC stemness.
Interestingly, miR-221 is up-regulated in many types of tumors, including breast cancer, prostate cancer, lung cancer, and colorectal cancer (30). Tumor suppressor targets of miR-221 have been identified. For example, miR-221 directly targets the cell cycle inhibitor p27 (Kip1) and positively affects prolif-eration potential in prostate cancer (44). miR-221 also targets the tumor suppressor and cell cycle inhibitor p57 (CDKN1C) in hepatocarcinoma (45). Garofalo et al. (46) showed that miR-221 targets the tumor suppressors PTEN and TIMP3 to enhance the tumorigenicity of non-small cell lung cancer and hepatocellular carcinoma. Overexpression of miR-221 in several cancers is linked to resistance to various cancer therapies, in addition to a growth advantage in cancer cells (30). For instance, the expression of miR-221 is increased in several chemoresistant cancer cells (47,48). For these reasons, miR-221 is considered an oncomiR that plays an important role in cancer development, and the inhibition of miR-221 in combination with other cancer treatments may be relevant to a new therapeutic strategy for cancer treatment (49,50). Distinct from these studies, results reported here showed that miR-221 has an anti-stemness function to enhance the differentiation of mouse ESCs by down-regulating Oct4, Nanog, Sox2, and Prmt7 levels.
In summary, our results showed that miR-221-3p and miR-221-5p target the 3ЈUTRs of the major pluripotency factors Oct4, Nanog, and Sox2 in mouse ESCs, indicating an anti-stemness and pro-differentiation role for miR-221-3p and miR-221-5p. Because our results also uncovered that PRMT7 epigenetically represses the expression of miR-221-3p and miR-221-5p in mouse ESCs and that miR-221-5p silences the expression of Prmt7, it is possible that miR-221-5p and Prmt7 form a negative feedback loop (Fig. 7F). Finally, we provide evidence that the PRMT7-mediated repression of miR-221-3p and miR-221-5p expression is necessary for maintaining mouse ESC stemness.

Anti-stemness function of miR-221 RNA interference in mouse ESCs
The shRNA plasmids (30 g) were transfected into mouse ESCs (5 ϫ 10 5 cells in 0.8 ml) using the Gene Pulser Xcell electroporation system (500 microfarads and 250 V; Bio-Rad) according to the manufacturer's instructions. We have reported previously that the transfection in this electroporation condition was efficient (26). The cells were plated on a 6-cm dish and treated with puromycin (1.0 g/ml). These cells were cultured for 14 days and then harvested for further analysis.

Quantitative PCR for miRNA and mRNA expression
Total RNAs were isolated using TRIzol RNA isolation reagents (Life Technology). To measure mRNA levels, cDNA was synthesized using the iScript cDNA synthesis kit (Bio-Rad) according to the manufacturer's instructions. Then, quantitative PCR was performed using the CFX384 real-time PCR detection system (Bio-Rad). GAPDH levels were used as the internal control.
For miRNA measurement, qScript microRNA cDNA synthesis kit (Quantabio) was used to synthesize microRNA cDNA. In brief, total RNAs (1 g) were used in a poly(A) polymerase reaction that adds a poly(A) tail to miRNA, and polyadenylated miRNAs were further reverse-transcribed with qScript reverse transcriptase to synthesize miRNA cDNA. Quantitative PCR was performed. PCR data were normalized to sno66 to determine relative miRNA levels.

Chromatin immunoprecipitation assay
A ChIP assay was performed according to a previously described protocol with minor modifications (28,51,52). Mouse ESCs were first fixed with 1% formaldehyde. Cell pellets were then lysed with ChIP lysis buffer and sonicated for 15 min (30 s on and 30 s off for 15 cycles) to shear DNA using Bioruptor (Diagenode). Antibodies were added and incubated overnight at 4°C. Preblocked protein A beads were added and incubated for 1-2 h to capture the antibody-DNA complex. The beads were then washed once with the following buffers: low-salt buffer, high-salt buffer, LiCl buffer, and TE buffer. ChIP DNA was then eluted by the ChIP elution buffer (1% SDS and 0.1 M NaHCO 3 ). The eluate was reverse cross-linked, and ChIP DNA was purified by the phenol/chloroform extraction method.

RA-induced differentiation of mouse ESCs
RA-induced differentiation of mouse ESCs was performed as described previously (26). In brief, mouse ESCs were trypsinized and cultured in a 10-cm Petri dish (Fisher) with ESC media, without adding leukemia inhibitory factor, to generate embryoid body (EB). After 5 days, embryoid bodies were transferred to a gelatin-coated tissue culture dish, and 0.5 M RA was added to induce differentiation.

Statistical analysis
The statistical significance between the two groups was analyzed by Student's t test using Prism software (GraphPad Software, Inc.). Data are presented as the mean Ϯ standard deviation (S.D.) of at least three independent experiments. *, p Ͻ 0.05; **, p Ͻ 0.01; and ***, p Ͻ 0.001 indicate statistically significant changes.