Integrator of Stress Responses Calmodulin Binding Transcription Activator 1 (Camta1) Regulates miR-212/miR-132 Expression and Insulin Secretion*

Altered microRNA profiles have been demonstrated in experimental models of type 2 diabetes, including in islets of the diabetic Goto-Kakizaki (GK) rat. Our bioinformatic analysis of conserved sequences in promoters of microRNAs, previously observed to be up-regulated in GK rat islets, revealed putative CGCG-core motifs on the promoter of the miR-212/miR-132 cluster, overexpression of which has been shown to increase insulin secretion. These motifs are possible targets of calmodulin binding transcription activators Camta1 and Camta2 that have been recognized as integrators of stress responses. We also identified putative NKE elements, possible targets of NK2 homeobox proteins like the essential islet transcription factor Nkx2-2. As Camtas can function as co-activators with NK2 proteins in other tissues, we explored the role of Camta1, Camta2, and Nkx2-2 in the regulation of the miR-212/miR-132 cluster and insulin secretion. We demonstrate that exposure of control Wistar or GK rat islets to 16.7 mm glucose increases miR-212/miR-132 expression but significantly less so in the GK rat. In addition, Camta1, Camta2, and Nkx2-2 were down-regulated in GK rat islets, and knockdown of Camta1 reduced miR-212/miR-132 promoter activity and miR-212/miR-132 expression, even under cAMP elevation. Knockdown of Camta1 decreased insulin secretion in INS-1 832/13 cells and Wistar rat islets but increased insulin content. Furthermore, knockdown of Camta1 reduced K+-induced insulin secretion and voltage-dependent Ca2+ currents. We also demonstrate Camta1 and Nkx2-2 protein interaction. These results indicate that Camta1 is required not only for expression of the miR-212/miR-132 cluster but at multiple levels for regulating beta cell insulin content and secretion.

CAMTAs 2 are members of a family of calmodulin binding transcription activators that are highly conserved across a wide range of multicellular eukaryotes ranging from Arabidopsis through Drosophila and Caenorhabditis to humans (1)(2)(3). CAMTAs were first identified in plants where they are involved in cell differentiation in response to environmental cues (4), and a role for CAMTAs as common integrators to stress responses has been suggested (5). In vertebrates, this protein family appears to consist of only two members, CAMTA1 and CAMTA2.
Recent data have drawn our attention to the possibility of the involvement of CAMTAs in type 2 diabetes (T2D), in particular evidence from genome-wide association studies, which have identified a common genetic variant in CAMTA1 as associated with T2D (6,7). Being part of a neuroendocrine tissue, pancreatic islet cells share many functional characteristics with neuronal cells. In this context, it is relevant to point out that several studies have identified involvement of CAMTA1 in neuronal malfunction. A polymorphism in human CAMTA1 has been associated with neuropsychological test performance (8) and a genetic variation in human CAMTA1 predisposes individuals to episodic memory performance (9). In neuroblastoma cells, CAMTA1 primarily regulates genes involved in neuronal function and differentiation and is a candidate tumor suppressor (10).
Previous studies have demonstrated that regulation of CAMTAs involves association with class II histone deacetylases (11) and NK2 family proteins. CAMTA2, for example, is activated by dissociation of HDAC5, the activation of which subsequently promotes transcription of genes implicated in cardiac hypertrophy via interaction with an NK2 homeobox transcription factor Nkx2-5 (11,12). NK2 and related homeobox factors recognize and bind NKE sequence elements with the consensus sequence 5Ј-NAAGTG-3Ј (13). NK2-related proteins are essential regulators of development and differentiation in a variety of organisms and tissues (14 -17). Nkx2-2 belongs to this family of homeobox domain-containing proteins; it is a vital transcription factor involved in the development of the ventral vertebrate central nervous system (16), in neuroendocrine differentiation of islet cells in the pancreas (18), and in direct regulation of the insulin gene (19). It is critically involved in establishing mature functional alpha, beta, and PP cells in islets (18), in the maintenance of mature beta cell identity and function (20,21), and in beta-to-alpha cell reprogramming (22). Additionally, the influence of CAMTAs on neuronal survival has been suggested to be subject to signal-dependent regulation by the MEF2/HDAC5/CaMKII pathway (23), which has recently been targeted as a novel treatment for both type 1 and type 2 diabetes (24).
MicroRNAs (miRNAs) are small non-coding RNAs involved in the regulation of expression of specific target genes at the mRNA level. There is evidence of their involvement in the regulation of insulin secretion and the development of T2D (25). Altered expression of miRNAs has been demonstrated in several experimental models of T2D, including the pancreatic islets of the diabetic Goto-Kakizaki (GK) rat model (26) where 24 microRNAs were found to be up-regulated and six downregulated when compared with the control Wistar rat (27). However, not much is known regarding the transcriptional regulation of miRNAs. We have previously demonstrated that several of the miRNAs showing altered expression in the GK rat are glucose-dependent (27), and we and others have shown cAMP-dependent regulation of miRNAs (28 -30). Specifically, the miR-212/miR-132 cluster has been suggested to be regulated by cAMP through a PKA-dependent mechanism (30) involving cAMP-response element (CRE)-binding proteins and CRTC1 (31).
In this study, we performed a bioinformatic analysis of the promoter of the miR-212/miR-132 cluster, and we identified putative transcription factor binding sites for CAMTAs and NK2 factors. We investigated the possibility of regulation of miR-212/miR-132 expression by CAMTAs and Nkx2-2 in pancreatic islets and clonal beta cells, and we examined effects of CAMTAs on insulin secretion.

Results
Bioinformatic Analysis of miR-212/miR-132 Promoter-To approach the problem of the abnormal miRNA profiles observed in the diabetic GK rat (27), we hypothesized that this might result from disturbed transcriptional regulation of the miRNA genes. To evaluate this, we performed a bioinformatic analysis of the promoters of several up-or down-regulated miRNAs in GK rat islets reported in the literature (27). Using a custom Perl script, we analyzed the frequency of occurrence of DNA sequence motifs, conserved in human, rat, and mouse. This analysis suggested the presence of putative CM2 sequence motifs (A/C/G)CGCG(C/G/T), also called the CG-1 element, on the promoter of the miR-212/miR-132 cluster. This type of motif has been shown to interact with CAMTAs via its N-terminal CG-1 binding domain (5,(32)(33)(34). We also identified one putative conserved NKE element 5Ј-NAAGTG-3Ј identical to motifs that bind to NK2related factors (13) and an additional NK2 element conserved in rats and humans only (Fig. 1).
mRNA Expression of Camta1, Camta2, and Nkx2-2 Is Decreased in GK Rat Islets-To determine whether Camta1, Camta2, or Nkx2-2 gene expression might be relevant to diabetes in the GK rat, we investigated mRNA expression of these genes in GK rat islets compared with age-matched control Wistar rat islets. The relative expression of mRNA of the transcription factors was evaluated by real time quantitative reverse transcription PCR after conditioning islets for 24 h in 16.7 mM glucose, to mimic the level of blood glucose observed in the diabetic GK rat model, or for 6 days in 5 mM glucose to expose islets to normal physiological glucose levels for an extended period. In both conditions, we observe a significant reduction of Camta1, Camta2, and Nkx2-2 mRNA expression in GK rat islets compared with Wistar rat islets (Fig. 2, A-C). This is an indication that bringing glucose to physiological levels for a sustained period in GK islets does not take Camta1, Camta2, or Nkx2-2 mRNA expression to the levels observed in control Wistar rat islets. For Camta1 and Camta2, we did not observe any difference in mRNA expression between the two conditions in the same animal model. However, mRNA expression of Nkx2-2 was significantly increased at 16.7 mM glucose in Wistar islets, and a borderline increase was also observed in GK islets. The increase was significantly less in GK (only 63% increase) than in Wistar islets (2.4-fold increase).
Expression of the miR-212/miR-132 Cluster Is Decreased in GK Rat Islets under Glucotoxicity-After a 6-day incubation of islets in medium at 5 mM glucose, the expression of pri-miR-212/132, miR-212-3p, and miR-132-3p was reduced and identical in diabetic GK and control Wistar rats islets (Fig. 2, D-F). However, after incubation of islets at 16.7 mM glucose for 24 h, to mimic the blood glucose levels observed in the diabetic GK rat, expression of the primary miRNA transcript, pri-miR-212/132 (Fig. 2D), containing both miRNAs and expression of the two 3p mature miRNAs miR-212-3p (Fig. 2E) and miR-132-3p (Fig. 2F) increased almost 5-fold in the control Wistar rat islets but significantly less so, only 3-fold, in the GK rat islets. Published data on miR-212-3p and miR-132-3p in freshly picked untreated GK islets show that these miRNAs are highly overexpressed when compared with untreated Wistar islets (27). We now show that under glucotoxic conditions in normal Wistar islets the expression of this cluster is even higher than in GK. These data sug-FIGURE 1. Location of putative CGCG and NKE motifs conserved in human, rat, and mouse and found on the miR-212/miR-132 promoter. Also represented is the position of two conserved TATA boxes and previously reported CRE (36,54). Between positions Ϫ1000 and Ϫ750 bp upstream of the miR-212 hairpin, there are four CGCG motifs, three positioned between two TATA boxes and one just 48 bp downstream of the second TATA box; around Ϫ500 bp, there is one CGCG motif 26 bp downstream of the putative NKE element (AAGTG); just upstream of a pair of known CRE sites, there is one additional NKE element, conserved in human and rat, with two additional CGCG motifs 5 and 17 bp downstream of the CRE sites. The start of each microRNA hairpin also harbors a CGCG motif, and an additional motif lies between the two microRNA hairpins.
gested that when rat islets are exposed to high glucose levels, Camtas and Nkx2-2 may be necessary for full activation of this miRNA cluster.
Knockdown of Camta1 Reduces Glucose and High K ϩ -stimulated Insulin Secretion but Increases Total Insulin-To investigate whether reduced expression of Camtas could affect glucose-stimulated insulin release, insulin secretion was measured in clonal INS-1 832/13 cells after knockdown of Camta1 or Camta2. Knockdown of Camta1 significantly reduced glucosestimulated insulin secretion at 16.7 mM glucose by 57% (Fig.  5A), although a 3-4-fold increase in total insulin content was observed (Fig. 5B) under both basal (2.8 mM glucose) and glucose-stimulated (16.7 mM glucose) insulin secretion. Knockdown of Camta2 did not affect glucose-stimulated insulin secretion (data not shown). In Wistar islets, knockdown of Camta1 also significantly reduced insulin secretion (Fig. 5C). To see whether Camta1 acts at the final stage of the beta-cell stimulus-secretion coupling, i.e. Ca 2ϩ influx and/or exocytosis, we circumvented mitochondrial glucose metabolism and the normal cell depolarization by using a high K ϩ buffer (50 mM) at 2.8 mM glucose (35); this method causes an artificial cell depolarization by eliminating the K ϩ gradient across the membrane. Insulin secretion collected during a 30-min exposure to high K ϩ buffer was significantly reduced in INS-1 832/13 cells after knockdown of Camta1 (Fig. 5D).
Ca 2ϩ Currents Are Reduced by Knockdown of Camta1-To further investigate effects downstream of depolarization, electrophysiological measurements on INS-1 832/13 cells were used to measure differences in exocytosis and voltage-dependent Ca 2ϩ currents. Exocytosis was elicited by a train of 10 depolarizing pulses from Ϫ70 to 0 mV and was measured as changes in cell membrane capacitance. We observed no difference in the capacitance traces between knockdown and control cells (data not shown). Voltage-dependent Ca 2ϩ currents were triggered by membrane depolarizations from a holding potential of Ϫ70 mV to membrane potentials ranging from Ϫ50 to
Knockdown of Camta1 Decreases Expression of Luciferase from the miR-212/miR-132 Promoter under cAMP Stimulation-To investigate the regulation of the miR-212/miR-132 cluster by Camta1, we cloned a 2.6-kb fragment of the miR-212/ miR-132 promoter into a secreted metridia luciferase plasmid pMetLuc2-reporter. After a 6-h incubation in medium containing 16.7 mM glucose or 5 mM glucose with 1 M forskolin (Fsk) and 10 M IBMX, we detected increased expression of secreted metridia luciferase (Fig. 7A) confirming that, as expected, the cloned miR-212/miR-132 promoter fragment was activated by both high glucose and cyclic AMP stimulation (31,36). After a 72-h knockdown of Camta1 in INS-1 832/13 cells using siRNA, we transfected the plasmid construct into the cells and observed that after a 6-h incubation at 5 mM glucose with 1 M forskolin and 10 M IBMX, activation of metridia luciferase expression was significantly reduced (Fig. 7B). No difference in luciferase expression was observed in the same samples after 16 h of incubation at 5 mM glucose (data not shown). After Camta1 knockdown, we also observed a significant down-regulation of pri-miR-212/132 expression by qPCR after a 2-h incubation in Fsk and IBMX (Fig.  7C). However, a 2-h incubation in GLP-1 did not significantly alter pri-miR-212/132 expression (Fig. 7D).
Camta1 Interacts with Nkx2-2 Differentially at Low and High Glucose and with Respect to Nuclear and Cytoplasmic Localization-Using immunofluorescence confocal microscopy and nuclear co-localization analysis along with a CAMTA1 antibody that recognizes exon 9 (NBP1-93620, Novus Biologicals), located between the N-terminal CG-1 domain and the downstream TIG domain, we detected Camta1 in both the nucleus and the cytoplasm in INS-1 832/13 cells (Fig. 8A). We verified the specificity of the antibody by knockdown of Camta1 with siRNA Silencer Select assay s179516, which targets exons 2 and 3 within the N-terminal CG-1 binding domain. This resulted in decreased Camta1 in both the nucleus and the cytoplasm with an overall reduction of 50% (Fig. 8B). To investigate and validate protein interaction between Camta1 and Nkx2-2 in a physiologically relevant context, we performed a proximity ligation assay, which detects protein interactions at less than 40 nm with each fluorescent spot representing a single interaction event. INS-1 832/13 cells were cultured for 48 h in 2.8, 5, or 16.7 mM glucose and subsequently fixed and stained for detection of Camta1 and Nkx2-2 interaction using the Duolink proximity ligation assay. We detected fluorescent signals (Fig. 8, C and D) at all three glucose concentrations with a significant overall decrease observed both at 2.8 and 16.7 mM glucose compared with 5 mM glucose (Fig. 8D). In addition, we observed a significantly lower number of interaction signals in the nucleus at both 2.8 and 16.7 mM glucose, whereas at 5 mM glucose we saw an identical number of interactions in the nucleus and the cytoplasm (Fig. 8D).

Discussion
In T2D altered miRNA expression profiles have been observed in human as well as in animal models such as in diabetic GK rat islets, yet the transcriptional regulation of miRNAs has remained largely unexplored. In this study, we present evidence for the involvement of the transcriptional co-activator Camta1 in the regulation of the miR-212/miR-132 cluster, which in turn affects insulin secretion (31). We also show that CAMTAs are expressed in both rodent and human pancreatic islets and that in rat clonal beta cells Camta1 interacts with Nkx2-2 and influences insulin production and secretion via pathways other than miRNA regulation.
Diabetic GK rats present with elevated blood glucose levels ranging from 13 to 22 mM (27). Incubation of GK or Wistar rat islets at 16.7 mM glucose for 24 h, to mimic the effects of high blood glucose observed in the GK rat, resulted in down-regulated mRNA expression of Camta1, Camta2, and Nkx2-2 in GK rat islets compared with Wistar islets. GK and Wistar islets from the same animals were also incubated for 6 days at 5 mM glucose to see how the expression of these genes in GK rat islets would compare with that of Wistar islets under normal glucose conditions. However, we still observed lower levels of Camta1, Camta2, and Nkx2-2 (Fig. 2) in GK islets under physiological (5 mM) glucose. These data indicate that the down-regulation of these genes may be constitutive in GK rat islets or at least that the down-regulation observed has a relative permanence because it cannot be reversed by incubation of islets at physiological 5 mM glucose for 6 days.
Previous work demonstrated that miR-212 and miR-132 are up-regulated in freshly isolated GK rat islets compared with Wistar islets (27) and that expression of these miRNAs is reduced with increasing glucose concentration in GK islets after 24 h. Here, we confirm and strengthen our previous observation by demonstrating that the up-regulation of both the mature miR-212-3p and miR-132-3p and the pri-miR-212/132 (Fig. 2, D-F) in GK islets is significantly diminished after incubation at 16.7 mM glucose for 24 h when compared with the up-regulation observed in control Wistar islets under the same conditions. In contrast, no change was observed in expression of this miRNA cluster after a 6-day incubation at physiological (5 mM) glucose, which suggests that, in contrast to the transcription factors above, expression of these miRNAs is normalized by normoglycemia. Knockdown of Camta1 in INS-1 832/13 cells, followed by transfection of a luciferase construct driven by the miR-212/miR-132 promoter, produced no difference in luciferase expression after a 16-h incubation at 5 mM glucose. However, the observation that cAMP-dependent expression of luciferase from the miR-212/miR-132 promoter was significantly reduced after knockdown of Camta1 (Fig. 7, B  and C), and the close proximity of the putative CG-1-binding sites to known cAMP regulatory elements on the miR-212/ miR-132 promoter (Fig. 1), suggests that Camta1 may be required for cAMP-dependent transcriptional regulation of the miR-212/miR-132 cluster. However, after knockdown of Camta1, a 2-h incubation of cells with the incretin GLP-1, which also elevates cAMP, did not alter pri-miR-212/132 expression (Fig. 7D). This may be due to differing temporal distribution and amplitude of intracellular cAMP signals selectively regulating downstream events (37). These results are particularly relevant in combination with our recently published observation that the miR-212/miR-132 cluster is regulated by cAMP through de-phosphorylation of CRTC1 and phosphorylation of CREB/ATF1 (31). It would be of interest to further investigate whether Camta1 and CRTC1 act together to regulate the miR-212/miR-132 cluster.  In INS-1 832/13 cells, Camta1 and Camta2 were down-regulated by glucotoxicity (Fig. 3, A and B). However, in the Wistar rat, we observed no significant change in Camta1 or Camta2 under normoglycemia conditions compared with a 24-h glucotoxic condition (Fig. 2, A and B), while in human islets both CAMTA1 and CAMTA2 mRNAs were up-regulated under glucotoxic conditions (Fig. 6, A and B). A recent study revealed differential expression of CAMTA1 in alpha cells compared with beta cells of both human adult and fetal origin, with a 3.7and 1.8-fold higher expression of CAMTA1 mRNA in fetal and adult alpha cells, respectively (38). In contrast, human islets have been shown to have a greater proportion of alpha cells (39,40), although in rat islets beta cells are vastly predominant. Although we have not had the opportunity to explore expression of CAMTA1 in response to glucose in alpha and beta cells separately, this may explain the differences in response to glucose observed in INS-1 832/13 clonal beta cells, Wistar rat islets, and human islets, as glucose may induce opposite responses of CAMTA1 expression in alpha and beta cells.
After knockdown of Camta1 in INS-1 832/13 cells and Wistar islets, we observed a significant decrease of glucosestimulated insulin secretion (Fig. 5, A and C) but dramatically increased total insulin content (Fig. 5B). The decrease in mRNA expression of the rat insulin genes Ins1 and Ins2, both downstream targets of Nkx2-2 ( Fig. 9, C and D), is likely to be a feedback mechanism to down-regulate insulin production in the face of excessive levels of insulin present in the cell.
High K ϩ buffer stimulation of beta cells allows evaluation of ATP-independent mechanisms of insulin secretion with preferential activation of fast moving insulin granules at the cell periphery, which are more critically dependent on localized Ca 2ϩ pools (41,42). In contrast, glucose has a stimulatory effect on both fast-and slow-moving insulin granules whether near or distanced from the plasma membrane. In the context of high K ϩ -induced secretion being dependent on Ca 2ϩ ion pools localized at the plasma membrane, the observed reduced Ca 2ϩ current through plasma membrane Ca 2ϩ channels after Camta1 knockdown is in agreement with the reduced insulin secretion upon high K ϩ stimulation. This result suggests that Camta1 may be involved either directly in the regulation of voltage-dependent Ca 2ϩ currents through its calcium-calmodulin binding domain (3) or through regulation of genes coding for the voltage-dependent Ca 2ϩ channels themselves or associated regulatory factors.
Given the reduced insulin secretion observed after knockdown of Camta1 (Fig. 5, A, C, and D) and the reduced Ca 2ϩ current (Fig. 5, E and F), we expected that the depolarizationinduced capacitance increase would have diminished, reflecting reduced exocytosis, in the electrophysiological measurements. However, capacitance traces only tell us that we have electrical contact between a vesicle and the plasma membrane and not how much insulin is actually released. The method used to measure capacitance cannot distinguish between full granule fusion and "kiss-and-run" (or transient granule fusion). The fact that we do not see any difference in depolarizationinduced capacitance increase in the context of decreased glucose-induced insulin secretion could be due to increased kissand-run transient granule fusion without release of insulin (43). This interpretation could explain the increased insulin content observed (Fig. 5B) and indicates that Camta1 may have a direct effect on granule fusion, which warrants further investigation.
Knockdown of Camta1 also decreased expression of Nkx2-2, a transcription factor responsible for establishment of beta cell phenotype and for maintenance of beta cell identity. Knockdown of Nkx2-2 is known to result in reduced insulin secretion and beta cell function (18,20,44). In combination with the putative regulatory elements for CAMTAs and NK2 factors identified on the miR-212/miR-132 promoter, these results led us to hypothesize that CAMTA1 and Nkx2-2 might interact to regulate gene expression. CAMTA2, for example, has been implicated in cardiac hypertrophy though interaction with Nkx2-5, an NK2 homeobox transcription factor (11,12). Indeed, by a proximity ligation assay in rat INS-1 832/13 clonal beta cells, we observed interaction between Camta1 and Nkx2-2. The finding of significantly less interaction in the nucleus, at both 2.8 and 16.7 mM glucose, as opposed to at 5 mM glucose (Fig. 8D), suggests that such differential nuclear localization of the interactions occurs as a consequence of abnormal glucose levels, supporting a role for Camta1 in response to physiological stresses. Increase in intracellular Ca 2ϩ is intricately linked to the mechanisms that modulate insulin secretion from beta cells (45). In another context, expression of CAMTA1 has been shown to be up-regulated in cells exposed to increased intracellular Ca 2ϩ , and this has been associated with novel expression of cardiac transcription factors (46). However, transcription factors can often function as both transcriptional repressors and activators, depending on temporal or cell-specific settings, as is, in fact, the case for Nkx2-2 (19 -21, 47, 48).
The knowledge that, in humans, CAMTA1 expression is higher in alpha cells that in beta cells (49) suggests that precisely adequate levels of CAMTA1 may be required to maintain both alpha and beta cell identity in islets along with appropriate expression levels of Nkx2-2. In Fig. 10, we present a summary overview of our results in a rat beta cell. Reduced expression of the glucose transporter Slc2a2 (Glut2) and the islet transcription factor MafA may result from the reduced levels of Nkx2-2 itself after knockdown of Camta1 (Fig. 9). Reduced Slc2a2 expression would contribute to reduced uptake of glucose and thereby to reduced insulin secretion, whereas the reduced expression of MafA, a unique beta cell-specific activator (50) implicated in both insulin biosynthesis and secretion (51), will undoubtedly contribute to the reduced insulin secretion. Our data therefore support the notion that correct levels of Camta1 are crucial at multiple levels, from the detection of glucose, to insulin production and secretion, and beta cell identity. Finally, the normal levels of miR-212 and miR-132 expression obtained in GK rat islets after a 6-day incubation at 5 mM glucose suggest that the requirement for adequate levels of Camta1 may only become evident when blood glucose levels stray beyond normal levels, as is the case in diabetes. This suggests a fundamental role for Camta1 in response to physiological stress. Specifically, our results show that Camta1 is a novel candidate for the regulation of expression of the miR-212/miR-132 cluster. We also show that Camta1 is likely to function alongside Nkx2-2 and cAMP signaling to modulate insulin secretion in the mature islet.
In conclusion, our results provide the first evidence that, in addition to playing a part in the regulation of the miR-212/miR-132 cluster, Camta1 is involved in insulin production and secretion and the pathology of the diabetic GK rat model, suggesting it may play a crucial role in the development of type 2 diabetes.

Experimental Procedures
Materials-All chemicals were purchased from Sigma unless otherwise stated.
Insulin Secretion and Total Insulin-To measure insulin secretion from the INS-1 832/13 cell line, cells were first cul-tured to confluence in 11.1 mM glucose culture medium in 24-well plates. An overnight incubation in 5 mM glucose medium was performed before the secretion assay. Secretion assay buffer composition (in mM) was as follows: 114 NaCl, 4.7 KCl (68.7 NaCl for 50 KCl secretion), 1.2 KH 2 PO 4 , 1.16 MgSO 4 , 20 HEPES, 25.5 NaHCO 3 , 2.5 CaCl 2 , and 0.2% bovine serum albumin, pH 7.2. Secretion assays were performed in 4 -6 technical replicates. The cells were pre-incubated in 2.8 mM glucose for 30 min followed by 1 h or 30 min conditioning in 2.8 mM glucose (basal secretion), 1 h in 16.7 mM glucose (glucose-stimulated insulin secretion), or 30 min in 2.8 mM glucose with 50 mM KCl secretion buffer (high K ϩ -stimulated insulin secretion). The secretion buffer was collected from each well, and secreted insulin was determined by radioimmunoassay (Coat-A-Count TKIN2, Siemens) or ELISA (High Range Rat Insulin ELISA, catalog no. 10-1145-01, Mercodia) according to the manufacturers' protocol. Insulin secretion data were normalized to total protein collected from each well using RIPA buffer as described below; total insulin was determined from total protein by ELISA (Mercodia).
Islet Preparations-Rat islets were extracted by collagenase digestion from pancreases of 9 -10-week-old male diabetic GK rats (Stockholm colony (53), bred in-house) and age-matched male control Wistar rats (Taconic). Tail blood glucose levels were measured in both GK and Wistar rats before sacrificing the animals in a CO 2 chamber; tail blood glucose ranges were 12-22 mmol/liter (GK) and 4 -7 mmol/liter (Wistar). Human islets were obtained through the Human Tissue Lab, Lund University. Islets were hand-picked in Hanks' buffer and incubated in RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum, 100 units/ml penicillin, 100 g/ml streptomycin, 200 mM glutamine, and the indicated glucose concentration.
RNA Extracts and Relative Gene Expression-All RNA was extracted using miRNeasy kit (Qiagen, catalog no. 217004) with on-column DNase (Qiagen, catalog no. 79254) treatment according to the manufacturer's protocols. RNA concentration was determined using Nanodrop 1000 (Thermo Fisher Scientific). Real time quantitative reverse transcription-PCR was performed using TaqMan assay reagents (Life Technologies, Inc.); cDNA for gene expression was prepared using random primers and High Capacity reverse transcription kit (catalog no. 4368814) on 100 ng of sample RNA; cDNA for miRNA expression was prepared using pooled TaqMan miRNA assay stem loop primers on 10 ng of sample RNA. Relative gene expression was determined using TaqMan Universal PCR Master Mix I (Thermo Fisher) for protein-coding genes and primary miRNA transcript; TaqMan Universal PCR Master Mix II (Thermo Fisher) was used for miRNAs along with TaqMan stem loop primers. Amplification and Ct values for three technical replicates of each sample were obtained using ViiA TM 7 real time PCR system and software (Life Technologies, Inc.). Relative gene expression was calculated against expression of two reference genes. The following TaqMan primers were used for rat genes: Camta1 (Rn01051596_m1); Camta2 (Rn01460457_m1); Nkx2-2 (Rn04244749_m1); Ins1 (Rn02121433_ g1); Ins2 (Rn01774648_g1); Slc2a2 (Rn00563565_m1); MafA (Rn00845206_s1); and pri-miR-212/132 (Rn03465057_pri), a In addition, the molecular interaction between Camta1 and Nkx2-2 in the nucleus is reduced when glucose levels in the medium are either above or below normoglycemia (indicated by ϮG m ). Abbreviations used are as follows: G-6-P, glucose 6-phosphate; VDCC, voltage-dependent calcium channel; K ATP channel, ATP-sensitive potassium channel. 69-bp amplicon for primary miRNA transcript containing both miR-212 and miR-132 hairpins and against two reference genes Hprt1 (Rn01527840_m1) and Ppia (Rn00690933_m1). The following TaqMan primers were used for human genes: CAMTA1 (Hs00391998_m1), CAMTA2 (Hs00610789_m1), and NKX2-2 (Hs00159616_m1) against two reference genes HPRT1 (Hs02800695_m1) and B2M (Hs00984230_m1). To detect expression of the mature miRNAs, only the more abundant 3p miRNA sequences were assessed. For rats, the following miRNA TaqMan primers were used: miR-212 (ID 002551) and miR-132 (ID 000457) against two reference small RNAs, U6 snRNA (ID 001973) and U87 (ID 001712). For humans, the following miRNA TaqMan primers were used: miR-212 (ID 000515) and miR-132 (ID 000457) against two reference small RNAs, RNU44 snRNA (ID 001094) and RNU48 (ID 001006).
Plasmid Construct and Luciferase Assay-A 2.6-kb fragment of the promoter region of the miRNA cluster miR-212 and miR-132 was cloned from rat DNA with forward and reverse primers 5Ј-ATCTCGAGGAGAACCTGCTTCACATCATAGG-3Ј and 5Ј-ATGAATTCACTCCTGATCCCATCAGTTCACCA-3Ј, respectively. The fragment was inserted into the promoter region of secreted Metridia Luciferase plasmid pMetLuc2-Reporter (Clontech) using GeneArt seamless cloning and assembly kit (Invitrogen, Life Technologies, Inc.). FuGENE HD transfection reagent (Promega) was used for transfection of the plasmid construct into cells grown in 24-well plates. Secreted Metridia Luciferase expression was evaluated in cell culture medium with Ready-To-Glow TM secreted luciferase reporter assay (Clontech) according to the manufacturer's protocol. Tecan Infinite 200 instrument with Magellan software was used to measure chemiluminescent data in relative light units.
Activation of miR-212/miR-132 Promoter by cAMP-To evaluate robust activation of known cyclic AMP-response elements (CRE) on the promoter of miR-212/miR-132, we stimulated cells with 1 M forskolin to enhance cAMP production, in combination with 10 M IBMX, which inhibits cAMP degradation. The incretin GLP-1 at 100 mM was used to evaluate a more physiological but less sustained cytoplasmic cAMP elevation.
Immunostaining and Laser Scanning Confocal Microscopy-For analysis of immunostaining by confocal microscopy, cells were cultured in 8-well chamber slides (Nunc Lab-Tek II-CC2 TM chamber slide, Sigma) before being fixed with 3% paraformaldehyde, permeabilized with 0.1% Triton X-100, and blocked with 5% normal donkey serum. Immunostaining was performed with rabbit anti-CAMTA1 as primary antibody (NBP1-93620, Novus Biologicals) and donkey DyLight488 anti-rabbit (Jackson 711-485-152) as secondary antibody. Nuclear staining was obtained by incubation in Hoechst (catalog no. 34580, Molecular Probes). Imaging was performed under ϫ60 magnification with oil immersion on a confocal laser scanning microscope (Zeiss LSM-510) equipped with argon 488 nm and 2-photon lasers. Single or Z-stack images were acquired and analyzed for nuclear co-localization using ZEN 2009 software (Zeiss, Germany).
The experiments were conducted on single INS-1 832/13 cells 72-96 h after transfection without visualization of the transfected cells. All experiments were performed at 32-34°C using the standard whole-cell configuration of the patch clamp technique. Changes in membrane capacitance and whole-cell currents were recorded using an EPC-10 patch clamp amplifier and the software PatchMaster (Heka Elektronik, Germany; version 2x90.1). Exocytosis was elicited by a train of 10 depolarizing pulses from Ϫ70 to 0 mV applied as 1 Hz. Voltage-dependent Ca 2ϩ currents were investigated using an IV protocol where the cell membrane was depolarized from Ϫ70 mV to voltages ranging from Ϫ50 to ϩ 40 mV for 100 ms. Charge (Q) was measured as the AUC 12 ms after onset of the pulse to exclude the rapidly inactivating Na ϩ component.
Author Contributions-I. G. M. conceived the project, conducted most experiments, analyzed and interpreted the results, and wrote the paper. H. A. M. conducted experiments, analyzed data, and revised the manuscript critically. A. W. advised on protocols and design, obtained and analyzed electrophysiological data, and revised the manuscript critically. M. O. M. conceived the project and revised the manuscript critically. L. E. conceived the project, advised on protocols, interpreted the results, and wrote the paper.