Menin Is Required for Optimal Processing of the MicroRNA let-7a*

Background: Menin represses pancreatic beta cell proliferation. Results: Menin promotes processing of let-7a, whose target IRS2 plays an important role in insulin signaling and beta cell proliferation. Conclusion: Menin represses beta cell proliferation partly via regulation of miRNA biogenesis. Significance: Understanding how menin represses beta cell proliferation will aid toward improving therapies targeting endocrine tumors and metabolic diseases including diabetes. Multiple endocrine neoplasia type I (MEN1) is an inherited syndrome that includes susceptibility to pancreatic islet hyperplasia. This syndrome results from mutations in the MEN1 gene, which encodes menin protein. Menin interacts with several transcription factors, including JunD, and inhibits their activities. However, the precise mechanism by which menin suppresses gene expression is not well understood. Here, we show that menin interacts with arsenite-resistant protein 2 (ARS2), a component of the nuclear RNA CAP-binding complex that is crucial for biogenesis of certain miRNAs including let-7a. The levels of primary-let-7a (pri-let-7a) are not affected by menin; however, the levels of mature let-7a are substantially decreased upon Men1 excision. Let-7a targets, including Insr and Irs2, pro-proliferative genes that are crucial for insulin-mediated signaling, are up-regulated in Men1-excised cells. Inhibition of let-7a using anti-miRNA in wild type cells is sufficient to enhance the expression of insulin receptor substrate 2 (IRS2) to levels observed in Men1-excised cells. Depletion of menin does not affect the expression of Drosha and CBP80, but substantially impairs the processing of pri-miRNA to pre-miRNA. Ars2 knockdown decreased let-7a processing in menin-expressing cells but had little impact on let-7a levels in menin-excised cells. As IRS2 is known to mediate insulin signaling and insulin/mitogen-induced cell proliferation, these findings collectively unravel a novel mechanism whereby menin suppresses cell proliferation, at least partly by promoting the processing of certain miRNAs, including let-7a, leading to suppression of Irs2 expression and insulin signaling.

Multiple endocrine neoplasia type I (MEN1) 2 is an inherited syndrome, with development of neoplasia in several endocrine organs including pancreatic islets (1)(2)(3)(4). The gene mutated in this syndrome, MEN1, encodes a nuclear protein of 610 amino acids, menin (5,6). Menin interacts with multiple proteins and is involved in a variety of cellular processes including gene transcription, cell proliferation, apoptosis, and genome stability (7)(8)(9)(10)(11)(12), but the precise mechanisms regarding menin-mediated suppression of cell proliferation remain to be elucidated. The recently solved crystal structure of menin reveals that it contains a deep pocket that binds short MLL1 or JunD peptides in the same manner, but has opposite effects on gene transcription (13), supporting the capacity of menin to act as either a contextdependent transcription activator or repressor.
MicroRNAs (miRNAs) are short (20 -24 nucleotides), endogenous non-protein-coding RNA molecules that negatively regulate gene expression post-transcription. MiRNAs either bind to their target gene mRNA and promote their degradation and/or inhibit protein translation (14,15). The primary transcripts of miRNA genes (pri-miRNA) are processed in the nucleus by the Microprocessor complex consisting of DGCR8 and Drosha to ϳ70-nucleotide hairpin structures called pre-miRNAs (16 -18). Pre-miRNAs are exported out of the nucleus and processed to ϳ22-nucleotide double stranded miRNA duplexes by Dicer (19 -21). The miRNA duplexes are unwound, and the mature miRNA is incorporated into the RNA-induced silencing complex containing the Argonaute protein (22,23). MiRNAs are involved in diverse biological processes during development and disease (24,25) including type 2 diabetes (26 -28). Aberrant expression of miRNAs has also been linked to several types of cancer (29,30).
miRNA-mediated regulation of pancreatic islet cell proliferation and insulin secretion has previously been demonstrated (31)(32)(33)(34). Conditional deletion of Dicer1 early in pancreas development results in dramatic reduction of insulin-producing beta cells (31). Similarly, Dicer1-deficient adult beta cells show dramatic decrease in insulin content, insulin-promoter activity, and insulin mRNA levels resulting in development of diabetes in these animals (32). Furthermore, miR-7a inhibition leads to activation of mTOR signaling and promotes adult beta cell replication in mouse primary islets (33). Analysis of miRNAs showing differential expression between normal pancreas and pancreatic endocrine tumors (PET) identified 87 up-regulated and 8 down-regulated miRNA in PETs (34). Among the down-regulated miRNAs, miR-155 showed lack of detectable expression in PETs upon comparison with normal pancreas (34). Whereas inactivation of the MEN1 gene is among the most frequent genetic events identified in sporadic PETs (35), little is known as to whether menin is involved in regulating miRNAs during the pathogenesis of the MEN1 syndrome.
Here, we show that menin interacts with arsenite-resistance protein 2 (ARS2), a component of the nuclear cap-binding complex, and facilitates processing of primary-miRNA (pri-miRNA) including pri-let-7a and pri-miR-155, reducing the levels of Insr and Irs2, both let-7a targets. Moreover, anti-miRmediated knockdown of let-7a results in increased levels of insulin receptor substrate 2 (IRS2). These findings unravel the role of menin in regulating the biogenesis of miRNAs via interaction with ARS2 and expression of a key component of the insulin-signaling pathway, IRS2.

EXPERIMENTAL PROCEDURES
Plasmids and Cell Culture-Retroviral plasmids expressing FLAG-tagged menin have been described previously (36). Lentiviral constructs expressing Ars2 were obtained from Open Biosystems. Lentiviral packaging plasmids, pMD2G and pAX2G, were purchased from Addgene. MEFs and HEK293 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% FBS and 1% Pen/Strep.
RNA Extraction and Quantitative Real-time PCR-Total RNA was extracted from cultured cells with TRIzol and an RNeasy extraction kit from Qiagen. One g of RNA was transcribed into cDNA using Superscript III RT from Invitrogen, and real-time PCR was performed on the 7500 Fast Real Time PCR machine from Applied Biosystems. Either Gapdh or Hprt1 mRNA levels were used as endogenous controls, and analysis was done using the relative quantification method according to instructions from ABI.
Isolation of miRNA and cDNA Synthesis-miRNA was isolated using the mirVana miRNA isolation kit from Invitrogen. Briefly, cells were lysed in lysis/binding solution, and RNA was extracted with acid-phenol:chloroform. One-third volume of ethanol was added to the aqueous phase and passed through a filter cartridge. Another two-third volume of ethanol was added to the filtrate, mixed, and passed through a second filter cartridge. Finally, the miRNA was eluted with 100 l of elution solution at 95°C. cDNA was synthesized using the NCode miRNA First-Strand cDNA Synthesis kit from Invitrogen. Briefly, a poly(A) tail was added to the miRNA samples using poly(A) polymerase, followed by incubation at 37°C for 15 min. First strand cDNA was synthesized from the polyadenylated miRNA by reverse transcription using the supplied Universal RT Primer Superscript III RT/RNaseOUT enzyme mix and incubated at 50°C for 50 min.
Primary miRNA Processing Assay-To generate pri-let-7a, PCR was performed using primers flanking the pre-let-7a hairpin of the mouse locus. Forward and reverse primers added ssRNA flanks as well as a 5Ј-XhoI and 3Ј-XbaI site. The resulting PCR products were digested and cloned into pcDNA3.1. Pri-let-7a template for in vitro transcription was generated from the pri-miRNA-pcDNA3 construct described above. PCR was employed to add on a 5Ј T7 promoter (5Ј-TCG TAA TAC GAC TCA CTA TAG GGA TAT CCA TCA CAC TGG CGG CC-3Ј) and (5Ј-GCT GAT CAG CGA GCT CTA GC-3Ј). Radiolabeled pri-let-7a was generated using the T7 High Yield RNA Synthesis kit (New England Biolabs) in the presence of [ 32 P]UTP according to the manufacturer's instructions. Cell lysates were prepared from MEFs in radioimmuneprecipitation assay buffer followed by sonication and centrifugation. A reaction mixture containing 15 l of cell lysate, 6.4 mM MgCl 2 , and radiolabeled pri-let-7a transcript was incubated at 37°C for 90 min. RNA was extracted using TRIzol and resolved on a 6% urea-polyacrylamide gel, and the bands were detected by autoradiography.
Statistical Analyses-Statistical analysis was performed using GraphPad Prism (v 5.0, GraphPad Software). The data are presented as the mean Ϯ S.D. A two-tailed Student's t test was used for measuring statistical differences.

Menin Interacts with ARS2, a Component of the Nuclear RNA Cap-binding Complex Important for miRNA Biogenesis-
Menin acts as a scaffold protein coordinating the function of different proteins (38), but the precise mechanisms by which it represses gene transcription and cell proliferation are not fully understood. We sought to identify novel menin-interacting partners to provide further insight into menin-mediated regulation of cell proliferation. To this end, ectopically expressed menin was purified by affinity chromatography using FLAG M2 beads, and the menin-interacting proteins were excised for characterization by mass spectroscopy. Analysis of the purified proteins showed that ARS2 was among the major eluted proteins (Fig. 1A). ARS2 is a component of the nuclear RNA capbinding complex that stabilizes the primary miRNA (pri-miRNA) transcript for processing by the Microprocessor complex consisting of Drosha and DGCR8, and thus plays an essential role in miRNA-mediated silencing (37,39). To confirm ARS2 as a menin-interacting partner, we performed coimmunoprecipitation in cell lysates from HEK293 cells ectopically expressing menin. Immunoprecipitation with an anti-menin antibody effectively pulled down ARS2 (Fig. 1B), indicating ARS2 as a novel menin-interacting partner.
Menin Affects the Levels of Mature let-7a and miR-155-A previous report showed that depletion of ARS2 results in reduced levels of several miRNAs including miR-21, let-7, and miR-155 (37). To determine whether menin plays a role in the biogenesis of certain miRNAs, we compared mature let-7a and miR-155 levels in Men1 l/l ;CreER MEFs, a cell line in which Men1 can be effectively excised upon treatment with 4-hydroxytamoxifen ( Fig. 2A). qRT-PCR analysis showed that mature let-7a and miR-155 expression levels were substantially reduced upon Men1 excision (Fig. 2, B and C). Furthermore, ectopic expression of menin in BON cells, a human neuroendocrine carcinoid cell line that expresses low levels of endogenous menin (40), resulted in increased levels of let-7a (Fig. 2D), consistent with the idea that menin facilitates the processing of let-7a. To determine whether menin affects miRNA levels at the transcriptional level, we examined the levels of pri-miRNA in both WT and Men1-excised cells. The levels of pri-let-7a (Fig. 3A) and pri-miR-155 (Fig. 3B) in Men1-excised cells were similar to levels in control WT cells. These findings indicate that menin promotes the biogenesis of certain miRNAs, including let-7a and miR-155, post-transcriptionally at the level of miRNA processing.
Menin Enhances the Processing of pri-let-7a to pre-let-7a-Because menin does not affect the transcript level of pri-let-7a (Fig. 3A), and ARS2 has been shown to be involved in the processing of pri-miRNA to pre-miRNA, we determined whether menin affects the processing of pri-let-7a to pre-let-7a biochemically. We performed a miRNA-processing assay using 32 P-radiolabeled pri-let-7a as the substrate and cell lysates from either menin-expressing or menin-null cells. Lysates from menin-null cells contained diminished processing activity compared with lysates from control WT cells, which generated a cleavage product of ϳ94 nucleotides (Fig. 4). This result indicates that menin plays a role in the biogenesis of let-7a by enhancing the processing of pri-let-7a to pre-let-7a.
To determine whether the decrease in the processing of prilet-7a in Men1-excised cells results from decreased levels of components of the pri-miRNA processing machinery, we examined the levels of Drosha and CBP80. Our results show that the protein levels of Drosha, an RNase III enzyme that executes the initial step of miRNA processing in the nucleus along with DGCR8 (17), and CBP80, a nuclear cap-binding protein (37), are comparable in both menin-excised and wild type cells (Fig. 5). This indicates that the role of menin in the processing of pri-let-7a does not occur via regulation of the expression levels of the Microprocessor complex, but rather by enhancing the stability of the pri-let-7a transcript.

Functional Interaction between Menin and ARS2 Is Crucial
for Processing of pri-let-7a-To elucidate the causal relationship between the interaction of menin and ARS2 in miRNA processing, we knocked down Ars2 expression in both WT and menin-null cells. A significant reduction in the levels of ARS2 protein was observed in both cell types using three independent clones of shRNA targeting Ars2 (Fig. 6A). qRT-PCR analysis indicated a 40% decrease in let-7a levels in WT cells for all three shRNAs examined (Fig. 6B). On the contrary, Ars2 knockdown did not affect let-7a levels in Men1-excised cells (Fig. 6B). These findings indicate that the impact of ARS2 on miRNA processing depends on the presence of menin, strongly suggesting that menin and ARS2 interact functionally to control processing of let-7a.
IRS2 Levels Are Enhanced upon Men1 Excision-Using prediction algorithms, Insr and Irs2 were previously identified as let-7 targets based on sequence complementarity between the miRNA and the 3Ј-UTR of target genes (41). Inhibition of let-7a using anti-miRs prevented down-regulation of IRS2 expression in the liver of mice on a high fat diet, confirming IRS2 as a let-7a target (41). Importantly, IRS2 plays a central role in peripheral insulin signaling and pancreatic beta cell proliferation, and disruption of IRS2 causes type 2 diabetes in mice (42). We have shown previously that excision of the Men1 gene ameliorated pre-existing glucose intolerance and increased both glucosestimulated insulin release and circulating insulin levels in mouse models for diabetes (43). We therefore examined whether a decrease in let-7a levels upon Men1 excision leads to increased IRS2 expression. Our results show that upon Men1 excision, Irs2 expression is increased ϳ2-fold at the mRNA level (Fig. 7A), with substantial increase at the protein level (Fig.  7B). This is consistent with the role of miRNAs in regulating gene expression by degradation of target gene mRNA and/or inhibiting translation (14,15). Furthermore, Men1 excision increased levels of insulin receptor, Insr, another known target of let-7a that phosphorylates and activates IRS2 (41) (data not shown). In agreement with this finding, ectopic expression of menin in mouse ␤HC9 cells, insulin-producing cells derived from pancreatic islets with beta cell hyperplasia, decreases Irs2 mRNA and protein levels (Fig. 7C). Additionally, decreased levels of IRS2 protein levels were observed in BON cells ectopically expressing menin (Fig. 7D). These findings suggest that the elevated level of mature let-7a in menin-expressing cells targets the Irs2 transcript for degradation, and excision of Men1 subsequently relieves this degradation, resulting in enhanced expression of IRS2.
Anti-miR-induced Knockdown of let-7a Restores the Levels of IRS2-To determine whether menin-mediated biogenesis of let-7a plays a key role in menin-induced repression of Irs2 expression, we examined the effect of let-7a inhibition on Irs2 levels. To this end, cells were transfected with either let-7a or control anti-miR, and the levels of IRS2 were quantitated 5 days after transfection. A significant increase in the protein levels of IRS2 was observed in menin-expressing MEFs transfected with let-7a anti-miR compared with control anti-miR-treated cells (Fig. 8, second panel, lane 2 versus 1). As positive controls, the protein level of RAS, a known target of the let-7 microRNA family (44), was also up-regulated in menin-expressing MEFs treated with let-7a anti-miRNA compared with control anti-miR-treated cells, as expected (Fig. 8, third panel, lane 2 versus  1). These results clearly demonstrate that the decreased levels of IRS2 in menin-expressing MEFs compared with Men1-excised cells can be, at least partly, attributed to the presence of elevated levels of let-7a. To examine whether the impact of the let-7a anti-miR is menin/let-7a-specific, we treated menin-null cells with let-7a anti-miR. Our results indicate that the level of IRS2 in Men1-excised cells was not affected by let-7a anti-miR treatment (Fig. 8, lane 4 versus 3), possibly due, in part, to the  Menin is required for processing of let-7a. Radiolabeled pri-let-7a was incubated with cell extracts from WT and menin-null MEFs, and the RNA was resolved on a 6% urea-polyacrylamide gel followed by autoradiography. The substrate, pri-let-7a, and product, pre-let-7a, are as indicated. nt, nucleotides.  relatively low levels of let-7a present in the menin-null cells, as expected. Similarly, the protein levels of RAS were unchanged after let-7a anti-miR treatment in Men1-excised cells compared with control anti-miR-treated cells. Collectively, these results indicate that the difference in IRS2 protein levels in menin-null and menin-expressing cells is largely attributed to menin-regulated production of mature let-7a, which degrades Irs2 mRNA and inhibits protein translation.

DISCUSSION
Here, we identify the role of menin in the biogenesis of certain miRNAs including let-7a and miR-155. We show that menin interacts with ARS2, a protein previously known to contribute toward stability and delivery of capped pri-miRNA transcripts to the Microprocessor complex containing Drosha and DGCR8 (37). Excision of Men1 results in decreased levels of several miRNAs including let-7a and miR-155 (Fig. 2). Analysis of the transcript levels indicated that the levels of pri-miRNA were not altered by menin (Fig. 3). Additionally, radiolabeled processing assay indicates that menin plays a role in the processing of pri-miRNA to pre-miRNA (Fig. 4). The recently solved crystal structure of menin (13) suggests that it acts as a scaffold protein and regulates gene expression by interacting with distinct partners including transcription factors JunD (7,13), NF-B (45), PPAR␥ (46), and various histone modification enzymes including MLL (8,13,47,48), PRMT5 (12), EZH2 (49) and HDACs (50). Furthermore, it has previously been shown that menin binds to the promoter of miR-26-a and induces its expression, resulting in decreased SMAD1 levels during osteoblastic differentiation of human adipose tissue-derived stem cells (51). Ars2 knockdown resulted in significant decrease of let-7a levels in menin-expressing cells whereas it had little effect in menin-null cells. This indicates that the functional interaction between menin and ARS2, in addition to the miRNA processing machinery, is essential for proper processing of the let-7a miRNA (Fig. 6). Additionally, this is the first report detailing a post-transcriptional role for menin in regulating gene expression and cell proliferation via regulation of miRNA processing.
The role of miRNAs in regulating pancreatic beta cell proliferation and insulin secretion has been demonstrated extensively (27,28,(31)(32)(33). We show here that menin plays a role in the processing of let-7a, a member of the let-7 family of miRNAs that regulates glucose metabolism in multiple organs (41). Furthermore, global knockdown of let-7a in mice prevents and treats impaired glucose tolerance in diet-induced obese mice, partly by increasing the levels of IRS2 and INSR (41). Similarly, we observed decreased levels of let-7a in Men1-excised cells with significant increase in IRS2 expression (Fig. 7), and inhibition of let-7a using specific anti-miR resulted in derepression of IRS2 (Fig. 8) suggesting that the increased levels of Irs2 upon Men1 excision can be attributed, in part, to reduced let-7a levels. It has previously been reported that IRS2 is markedly overexpressed in rat insulinoma cells compared with rat primary islet ␤ cells (52). Furthermore, the importance of IRS2 in insulin signaling has been demonstrated using transgenic knock-out mouse models where Irs2 Ϫ/Ϫ mice show a marked decrease in beta cell mass and develop type 2 diabetes (53). It is thus conceivable that elevated serum insulin and decreased blood glucose levels in Men1-null transgenic mice (54) can be attributed partly to decreased let-7a with subsequent increase in IRS2 and increased pancreatic beta cell mass. As a corollary, it can also be conceived that disruption of the menin-let-7a-IRS2 axis contributes, at least in part, to tumorigenesis caused by menin mutations in MEN1 syndrome. Moreover, it is likely that menin-mediated processing of miRNAs such as let-7a also contributes partly to the suppression of endocrine tumors.
A recent report described insulin-mediated down-regulation of menin and localization in the cytoplasm in a time-dependent manner via the human insulin receptor (55). It is conceivable that activation of the insulin signaling pathway by insulin leads to reduction in menin levels particularly in the nucleus, consequently resulting in decreased processing of pri-let-7a and increased Irs2 expression. However, in the cell lines used for our  . Anti-miR-induced repression of let-7a expression results in increased levels of IRS2. Men1 l/l and Men1 ⌬/⌬ MEFs were harvested 5 days after transfection with either control (Ϫ) or let-7a (ϩ) anti-miR, and the protein levels of IRS2 and RAS were determined by Western blotting. Immunoblotting for actin is included as a loading control. study, we did not observe any decrease in global menin levels or increased localization in the cytoplasm upon treatment with insulin for either 15 min or 24 h (data not shown). However, increased p-AKT levels were detected in cells treated with insulin (data not shown). It is possible that the effect of insulin on menin expression was not observed in our cells because of the lack of ectopic expression of human insulin receptor.
In addition to let-7a, we observed that menin plays a role in the biogenesis of miR-155 with reduced levels upon Men1 excision (Fig. 2C). A previous screen for miRNAs aberrantly regulated in pancreatic neuroendocrine tumors identified significant down-regulation of miR-155 expression in pancreatic endocrine tumors compared with normal pancreas (34). As Ͼ40% of pancreatic endocrine tumors harbor somatic mutations in the Men1 gene (56), it is possible that the decreased levels of miR-155 observed in pancreatic endocrine tumors can be partly attributed to the dysregulation of miR-155 biogenesis upon Men1 mutation.
In conclusion, we have uncovered a novel mechanism by which menin regulates gene expression post-transcriptionally via regulating biogenesis of miRNAs including let-7a and miR-155, and provide a rationale for possible use of anti-miR-based therapy for alleviating symptoms associated with MEN1 syndrome and metabolic diseases including diabetes.