A Pre-microRNA-149 (miR-149) Genetic Variation Affects miR-149 Maturation and Its Ability to Regulate the Puma Protein in Apoptosis

Background: The role of miRNA SNPs in disease susceptibility remains ill-defined. Results: The polymorphism rs71428439 is associated with the risk for myocardial infarction and affects miR-149 maturation. Puma is a target of miR-149. Conclusion: This polymorphism contributes to the risk of myocardial infarction through the miR-149-Puma axis. Significance: This polymorphism, miR-149, and Puma can be targets for the development of individualized treatment for myocardial infarction. to inhibit Puma expression and the con-sequent apoptosis. Our results warrant future studies to explore this SNP of pre-miR-149 in the pathophysiology of myocardial infarction.

logical and pathological processes, including development, apoptosis, cell proliferation, hematopoiesis, and tumorigenesis (7)(8)(9)(10). A growing number of studies have demonstrated that the aberrant expression of miRNAs was closely related to the etiology, diagnosis, and prognosis of many diseases, including heart diseases (11).
Single nucleotide polymorphisms (SNP) in protein-coding genes have been confirmed to link to many kinds of diseases. Although the role of miRNA SNPs in disease susceptibility remains largely unknown, its importance has been greatly implicated in many cancers such as thyroid cancer (12), renal cell carcinoma (13), and breast cancer (14). It is evident that common SNPs in miRNAs and SNPs within their targets may affect miRNA target expression and functions (12,15). Moreover, SNPs in miRNAs have been demonstrated to affect the miRNA expression (12,16). Recently, it was reported that a naturally occurring miR-499 mutation outside the critical seed sequence modifies mRNA targeting and end-organ function of miR-499 (17). All these studies suggest that the sequence variation caused by SNP or mutation may greatly contribute to the heterogeneity of complex diseases.
This study conducted a systematic search of SNPs within pre-miRNAs that have been verified to express in the heart (11, 18 -20). We found that an A3 G polymorphism (rs71428439) identified in the miR-149 gene is associated with the risk of myocardial infarction. This polymorphism is located in the stem region outside the mature miR-149 sequence and results in a change to the structure of the miR-149 precursor. Intriguingly, we observed that the G-allelic miR-149 precursor displayed a low efficiency to produce mature miR-149 compared with the A-allelic one. In follow-up functional analysis, our results showed that the miR-149 suppressed apoptosis through targeting the pro-apoptotic protein Puma. The A to G variation caused by the SNP in the miR-149 precursor leads to a reduced effect of pre-miR-149 to inhibit Puma expression and the consequent apoptosis. Our results warrant future studies to explore this SNP of pre-miR-149 in the pathophysiology of myocardial infarction.

EXPERIMENTAL PROCEDURES
Study Subjects-All study subjects were Han people enrolled from Hubei Province in China. The diagnosis of myocardial infarction was based on the following criteria: 1) typical chest pain lasting longer than 30 min; 2) characteristic electrocardiographic patterns of myocardial infarction; 3) elevation of cardiac enzymes (creatine kinase and lactate dehydrogenase), and troponin I or T in blood. The control individuals were randomly selected and matched to cases based on age and gender. Hypertension was defined as receiving ongoing medication for hypertension, systolic blood pressure Ն140 mm Hg, or diastolic blood pressure Ն90 mm Hg. Diabetes mellitus was defined as ongoing therapy for diabetes or a fasting plasma glucose level Ն7.0 mM/liter. The investigations complied with the ethical guidelines of the 1975 Declaration of Helsinki and were approved by appropriate local institutional review boards on human subject research. Informed consent was obtained from all participants.
Genotyping-Genomic DNA was extracted from peripheral blood samples. Genotypes were analyzed by polymerase chain reaction (PCR)-based direct DNA sequencing. DNA specimens were amplified by using standard PCR protocols. The PCR products were purified and sequenced in both directions. The sequencing results were analyzed by using DNAStar SEQMAN software. The primers, 5Ј-TCTCATGTCCAGGACCACAA-3Ј and 5Ј-GAGAGGCATGGAGAGGTGAG-3Ј, were used to amplify a 493-bp fragment covering the A3 G polymorphism site (rs71428439) in the miR-149 precursor.
Cell Cultures, Hydrogen Peroxide Treatment-Primary mouse cardiomyocytes were isolated from hearts of 1-day-old mouse pups, as we described elsewhere (21), with minor modifications. Briefly, non-myocyte contaminants were removed by two rounds of pre-plating for 1.5 h on 100-mm plastic cell culture dishes in a humidified incubator at 37°C with 5% CO 2 . Cardiomyocytes were plated separately into 24-well culture plates or 60-mm culture dishes with serum-containing medium. Following a 24-h incubation period, serum-containing medium was replaced with no serum medium. The treatment with hydrogen peroxide was performed as we described (22).
Construction of miR-149 Expression Vectors-To create the allelic A and G pre-miR-149 expression vectors separately, the 493-bp DNA fragments encompassing the miR-149 precursor sequence and its 5Ј-and 3Ј-flanking regions (199 and 205 bp, respectively) were amplified from human genomic DNA from normal blood donors (determined to have the AA or GG genotype) and cloned into the XhoI and XbaI sites of the vector pcDNA3 (Invitrogen). The sequences of both vectors were confirmed by direct sequencing, and the only difference was in the SNP. The yielded vector with the AA genotype was designated pc3.0-miR-149-A, and the one with the GG genotype was named pc3.0-miR-149-G. pc3.0-miR-423, which contained the miR-423 precursor sequence and its 5Ј-and 3Ј-flanking regions (253 and 117 bp, respectively), and pc-3.0-cel-miR-39, which contained cel-miR-39 precursor and its 5Ј-and 3Ј-flanking regions (85 and 195 bp separately), were constructed as negative controls.
Production of miR-149 Duplex or Inhibitor-Chemically modified antagomir reverse complementary to miR-149 was used to inhibit its expression. The miR-149 antagomir (As-miR-149) sequence was 5Ј-GGGAGUGAAGACACGGAGC-CAGA-3Ј. Chemically modified oligonucleotide 5Ј-CAGUAC-UUUUGUGUAGUACAA-3Ј was used as antagomir negative control (designated as As-NC). All of the bases were 2Ј-OMe modified. miR-149 mimic (designated as miR-149), designed as described elsewhere (23), was employed to enhance its expression. It contains an RNA strand with the sequence identical to the mature miR-149 and an artificial strand that was partially complementary to the mature miR-149 sequence. The miR-149 antagomir and mimic used in this study were both synthesized by GenePharma Co. Ltd. miRNA duplexes (sense, 5Ј-UUCUCC-GAACGUGUCACGUTT-3Ј; antisense, 5Ј-ACGUGACACG-UUCGGAGAATT-3Ј) were used as a negative control (designated as miR-NC). The transfection was performed using Lipofectamine 2000 according to the manufacturer's instruction (Invitrogen).

miR-149 Regulates Apoptosis
Analysis of miR-149 and Puma by Quantitative Reverse Transcription-PCR (qRT-PCR)-qRT-PCR was carried out according to the previously described method (24). miR-149 levels were measured by qRT-PCR using a TaqMan MicroRNA Assays kit according to the manufacturer's instructions in an ABI Prism 7000 sequence detection system (Applied Biosystems). Total RNA was extracted using TRIzol reagent. After DNase I (Takara) treatment, RNA was reverse-transcribed with reverse transcriptase (ReverTra Ace, Toyobo). The results of qRT-PCR were normalized to that of U6. The sequences of U6 primers were forward 5Ј-GCTTCGGCAGCA-CATATACTAA-3Ј and reverse 5Ј-AACGCTTCACGAATTT-GCGT-3Ј. qRT-PCR analysis for Puma was performed using the SYBR Green Real Time PCR Master Mix (Takara) according to the manufacturer's instructions. The data analyzed by qRT-PCR were normalized to that of mouse glyceraldehyde-3phosphate dehydrogenase (GAPDH). The sequences of Puma primers were forward 5Ј-AGGAGGGGGTCTGTGAAGAG-3Ј and reverse 5Ј-CTGGGCACTGGGTTAAGAAG-3Ј. The sequences of mouse GAPDH primers were forward 5Ј-TGTGTC-CGTCGTGGATCTGA-3Ј and reverse 5Ј-CCTGCTTCACCA-CCTTCTTGA-3Ј.
Northern Blot Analysis-To detect miR-149, the probes were labeled with the nonradioactive digoxigenin using an End Tailing Kit (Roche Applied Science). 15 g of total RNA were loaded onto a precast 15% denaturing TBE-urea polyacrylamide gel and transferred by electroblotting to Nybond Nϩ (GE Healthcare). After UV cross-linking, the membrane was hybridized with digoxigenin-labeled probes. The probe sequence used against miR-149 was 5Ј-GGGAGTGAAGA-CACGGAGCCAGA-3Ј. The hybridizations and washes were performed using an Ultrahyb-Oligo buffer (Ambion) at 42°C according to the manufacturer's instructions. The detections were achieved by the alkaline phosphatase-conjugated anti-digoxigenin antibody (Roche Applied Science) and the CDP-Star reagent (GE Healthcare).
Reporter Constructions and Luciferase Assay-3ЈUTR of Puma was amplified from mouse genomic DNA. The primers for Puma 3ЈUTR amplification were 5Ј-TCCGCCTTCTGA-CACCCT-3Ј and 5Ј-AACCACTGAGCCATTTCT-3Ј. To generate reporter vector bearing miR-149-binding sites, 3ЈUTR of Puma was cloned into the pGL3 vector (Promega) immediately downstream of the stop codon of the luciferase gene. The mutated 3ЈUTR was generated, and mutation (the wild type Puma 3ЈUTR, GCCA; the mutated Puma 3ЈUTR, AGGC) was introduced into the binding site. For luciferase assay, cells in 24-well plates were co-transfected with the plasmid constructs of 200 ng/per well of empty pGL3, pGL3 harboring the wild type 3ЈUTR (pGL3-Puma-WT-3ЈUTR), or the mutated 3ЈUTR (pGL3-Puma-MUT-3ЈUTR) of Puma, along with 60 nmol/liter miR-149 using Lipofectamine 2000 (Invitrogen). pRL-TK vector containing Renilla luciferase cDNA served as the internal control. Mimic control (miR-NC) served as a negative control. 36 h after transfection, cells were lysed, and luciferase activity was measured with the Dual-Luciferase kit (Promega) according to the manufacturer's instructions.
Adenovirus Construction and Infection-Adenoviruses harboring the coding sequence of Puma without 3ЈUTR (Puma-W/ O-3ЈUTR) and Puma with mutated 3ЈUTR (Puma-MUT-3ЈUTR) were constructed using the Adeno-X Expression System (Clontech) according to the manufacturer's instructions. The mutations were introduced to the binding site of miR-149 on Puma 3ЈUTR. CCAG in the wild type Puma 3Ј-UTR were converted to AGCA using QuikChange II XL sitedirected mutagenesis kit (Stratagene). Adenoviral Puma with 3ЈUTR (Puma-W-3ЈUTR) and adenoviral ␤-galactosidase (␤-gal) were used as we described (21). To construct adenoviruses encoding the A-allelic (miR-149-A) and G-allelic pre-miR-149 (miR-149-G), the same sequences cloned into pcDNA3.0 above were amplified and finally cloned into the adenoviral system. Adenoviruses encoding Cel-miR-39 was constructed using the same system. Viruses were amplified in HEK293 cells. Cells were infected with the viruses at the indicated multiplicity of infection (m.o.i.) for 1 h. After washing with PBS, the culture medium was added, and cells were cultured until the indicated time.
Preparation of RNAi Construct of Puma-The RNAi constructs were designed using the siRNA Design Tools from Ambion. Puma RNAi (Si-Puma) sense sequence was 5Ј-CCT-GGAGGGTCATGTACAATCTCTT-3Ј; Si-Puma antisense sequence was 5Ј-AAGAGATTGTACATGACCCTCCAGG-3Ј. The scramble Puma-siRNA (Scr-Puma) sense sequence was 5Ј-TACATCATTGTCGTGCTGCGAGTCA-3Ј; Scr-Puma antisense sequence was 5Ј-TGACTCGCAGCACGACAAT-GATGTA-3Ј. They were cloned into pSilencer adeno 1.0-CMV vector (Ambion) according to the manufacturer's instructions. The specificity of the oligonucleotides was confirmed by comparison with all other sequences in GenBank TM using Nucleotide BLAST. There was no homology to other known mouse DNA sequences.
TUNEL Assay-Cells were fixed in 4% paraformaldehyde in PBS and permeabilized with 0.1% Triton X-100 in 0.1% sodium citrate. An in situ apoptotic cell death detection kit (fluorescein, Roche Applied Science) based on TUNEL assay was used as per the manufacturer's instructions to detect apoptotic cells. Negative controls were included in each case by omitting TUNEL enzyme terminal deoxynucleotidyltransferase reaction mixture and incubating the cells with the label solution. PBS containing 5 g/ml 4Ј,6Ј-diamidino-2-phenylindole (DAPI; Vector Laboratories) was prepared to stain nuclei. Sections were examined with a Zeiss LSM510 META microscope. The percentage of apoptotic nuclei was calculated. 100 -150 cells were counted in 20 -30 random fields. For apoptosis analysis by TUNEL assay in heart sections, the procedure was the same except that cardiomyocytes were stained with the ␣-actinin antibody (A7811, Sigma). An investigator blind to the treatment quantified 20 random fields of samples.
Caspase-3 Activity Assay-Caspase-3 activity was measured using an Apo-ONE homogeneous caspase-3/7 assay kit (Promega) according to the manufacturer's protocol. Briefly, Apo-ONE caspase-3/7 reagent was added, and the mixtures were incubated at room temperature for up to 6 h. The level of fluorescence was measured using a Synergy 4 Hybrid Microplate Reader (BioTek Instruments) with excitation/emission at 499/ 521 nm.
Mitochondrial Staining and Immunofluorescence-Cells were plated onto the coverslips coated with 0.01% poly-L-lysine. After treatment they were stained for 20 min with 0.02 mol/ liter MitoTracker Red CMXRos (Molecular Probes). Immunofluorescence was performed as we described previously (25). The samples were imaged using a laser scanning confocal microscope (Zeiss LSM 510 META).
Animal Studies-We obtained C57BL/6 mice from the Institute of Laboratory Animal Science of the Chinese Academy of Medical Sciences (Beijing, China). Myocardial ischemic model was established by ligating left anterior descending of the coronary artery to cause myocardial ischemia as we described previously (24). Briefly, mice were anesthetized and placed on an HX-300S animal ventilator. Body temperature was maintained at 37°C on a heating pad. The beating heart was accessed via a small left anterior thoracotomy. After removing the pericardium, a descending branch of the left anterior descending coronary artery was ligated with a nylon suture. Ligation was confirmed by the whitening of a region of the left ventricle. Adenoviral miR-149-A, miR-149-G, or ␤-gal was injected immediately after left anterior descending ligation into the myocardium bordering the infarct zone at a dose of 1 ϫ 10 11 viral genome particles per animal using an insulin syringe with a small gauge needle, respectively. The chest was closed, and the animals were moved back to cages after the occurrence of spontaneous breathing. Cardiac function of these groups of animals was evaluated by echocardiographic analysis 14 days after the surgery. Trichrome stain of the heart section was performed by employing the trichrome stain (MASSON) kit HT-15 (Sigma).
Statistical Analysis-Pearson's 2 test was used to evaluate the differences in the distribution of genotypes between cases and controls. Hardy-Weinberg equilibrium was assessed using a goodness-of-fit 2 test. Association between the miR-149 polymorphism and the risk of myocardial infarction was analyzed by multivariate unconditional logistic regression, adjusted for gender, age, hypertension, and diabetes status. All analyses were performed using SPSS software (version 13.0, SPSS). All measurement data are shown as means Ϯ S.D. or means Ϯ S.E. of at least three independent experiments. Statistical analyses of the differences between groups were conducted by one-way analysis of variance followed by the LSD post hoc test for multiple comparisons. Paired data were evaluated by Student's t test. p Ͻ 0.05 was considered statistically significant.

miR-149 Polymorphism rs71428439 Is Associated with the
Risk for MI-We analyzed the samples of myocardial infarction and the controls, and their characteristics were summarized in Table 1, including gender, age, hypertension, and diabetes status. The observed genotype distribution of the miR-149 rs71428439 polymorphism in both cases and control groups was in agreement with Hardy-Weinberg equilibrium (p ϭ 0.312 and p ϭ 0.659, respectively), indicating no population stratification within the cohort. We observed that the genotype distribution of this polymorphism in myocardial infarction cases was statistically significantly different from that in control subjects ( 2 ϭ 12.639 and p ϭ 0.002). The GG genotype in the patients was remarkably more frequent than that in the controls. A similar trend was also observed in the frequency of the AG genotype ( Table 2). The association between the genotype and the risk of myocardial infarction was further analyzed using multivariate unconditional logistic regression with adjustment for gender, age, hypertension, and diabetes status. We observed a statistically significant association of the GG phenotype with the increased risk for myocardial infarction (before adjustment, odds ratio (OR) ϭ 2.340, 95% confidence interval (CI) ϭ 1.448 -3.783, p ϭ 0.001; after adjustment, OR ϭ 2.323, 95% CI ϭ    Table 2. Moreover, the ORs for myocardial infarction risk were higher in the recessive model (after adjustment, GG versus AAϩAG, OR ϭ 1.922, and 95% CI ϭ 1.235-2.993) than that in the dominant model (after adjustment, GG ϩAG versus AA, OR ϭ 1.635, and 95% CI ϭ 1.165-2.293), indicating a recessive role of the G allele for the prognostication of myocardial infarction risk.
rs71428439 Polymorphism Affects the Maturation of miR-149-To detect the consequence of SNP rs71428439 bringing to miR-149, we used the mfold web server to predict the minimum free-energy secondary structure of pre-miR-149-A and pre-miR-149-G and found that the G allele causes a difference to the structure of pre-miR-149 from the A allele (Fig. 1A). Sequence-mediated differences in processing miRNAs have

miR-149 Regulates Apoptosis
been reported (12,27). We thus explored whether the SNP could influence the processing of mature miR-149. To this end, HEK293 cells were transfected with pcDNA3.0-pre-miR-149-A (pc3-miR-149-A) or pcDNA3.0-pre-miR-149-G (pc3-miR-149-G), along with a plasmid expressing Caenorhabditis elegans miR-39 (cel-miR-39, which shares no homologous sequence in human and mouse) as a control of transfection and expression efficiency. Real time PCR was used to detect the effect of different alleles on the expression levels of mature miR-149. As shown in Fig. 1B, the expression levels of mature miR-149 from the G allele were less than that from the A allele.
To further confirm these results, Northern blot was used to detect the expression levels of miR-149 between the different alleles. We utilized the cells transfected with pcDNA3.0-pre-miR-423 plasmids as a control to exclude the influence of foreign sequence on miR-149 levels. The amount of miR-149 was also observed to be diverse between the different alleles. The expression levels of mature miR-149 from the G allele were lower than that from the A allele (Fig. 1C). To test whether the polymorphism would also affect miR-149 maturation in cardiomyocytes, we infected cells with adenoviral pre-miR-149-A (miR-149-A) or pre-miR-149-G (miR-149-G) and observed that pre-miR-149-G still produced less mature miR-149 than pre-miR-149-A in cardiomyocytes (Fig. 1D). These in vitro results prompted us to examine the association between rs71428439 genotypes and mature miR-149 expression levels in human samples. There were significantly lower levels of mature miR-149 in human peripheral blood mononuclear cells from people with at least one G allele (AG or GG genotype) in genome than from those with the A allele (AA genotype) (Fig.  1E). These data suggest that the G to A substitution in pre-miR-149 attenuates the processing of mature miR-149.
miR-149 Inhibits Apoptosis-To understand the function of miR-149, we used chemically synthesized 2Ј-OMe-oligonucleotides (AS-miR-149), the sequence of which is complementary to mature miR-149, to knock down endogenous miR-149. The scramble oligonucleotides were used as a negative control (designated as AS-NC). AS-miR-149 could efficiently reduce endogenous miR-149 levels ( Fig. 2A). At first, we assessed the effects of H 2 O 2 on endogenous miR-149 expression. The cells were exposed to a dose range of H 2 O 2 at different time points. miR-149 levels were reduced in a time-and dose-dependent manner (Fig. 2, B and C). TUNEL analysis revealed that knockdown of miR-149 caused the cells to be more susceptible to apoptosis upon hydrogen peroxide treatment (Fig. 2D). In addition, miR-149 knockdown promoted caspase 3 activation induced by H 2 O 2 (Fig. 2E).
Then we tested whether ectopic expression of miR-149 can influence apoptosis. miRNA duplex was used for miRNA overexpression. Principally, the miR-149 duplex (miR-149) contained an RNA strand with a sequence identical to the mature miR-149 and an artificial strand that was partially complementary to the mature miR-149 sequence. miR-NC, used as a negative control, was an miRNA duplex with a scrambled sequence that was nonhomologous to any mouse genome sequence. miR-149 levels were observed to increase by the administration of miR-149 duplex but not miR-NC (Fig. 2F). miR-149 could sig-nificantly attenuate caspase 3 activation and apoptosis induced by H 2 O 2 (Fig. 2, G and H). These data suggest that miR-149 is able to inhibit apoptosis.
To know the biological consequences of impaired processing of miR-149, we tested whether the A to G variation caused by SNP could influence apoptosis. To know whether the consequence brought by SNP can be disturbed by infection or expression efficiency, the infection/expression efficiency was monitored through co-infection with Cel-miR-39, and the levels of mature miR-149 were detected (Fig. 2I). Adenoviruses harboring pre-miR-149 with the A allele elicited a stronger effect on the inhibition of caspase 3 activation and apoptosis than the one with the G allele (Fig. 2, J and K). Thus, it appears that there is a functional effect of the SNP.
Puma Is a Target of miR-149-miRNAs execute their functions through inhibiting translation or promoting degradation of their target mRNAs. To work out the molecular mechanisms by which miR-149 regulates apoptosis, we searched the potential targets of miR-149 using the program of targetscan. Puma has a conservative miR-149-binding site in its 3ЈUTR (Fig. 3A). We observed that the mRNA and protein levels of Puma were increased responding to H 2 O 2 treatment (Fig. 3, B and C). The administration of AS-miR-149 resulted in an elevated Puma protein levels (Fig. 3D). Enforced expression of miR-149 attenuated the increase in Puma levels upon H 2 O 2 treatment (Fig.  3E). These data suggest that miR-149 participates in regulating Puma expression.
To learn whether Puma is a direct target of miR-149, we created luciferase constructs of Puma with a wild type 3ЈUTR (pGL3-WT-Puma-3ЈUTR) and with a mutated 3ЈUTR (pGL3-MUT-Puma-3ЈUTR). Luciferase reporter assays revealed that miR-149 induced a decrease in the luciferase activity of pGL3-WT-Puma-3ЈUTR but not pGL3-MUT-Puma-3ЈUTR (Fig. 3F). H 2 O 2 increased the activity of luciferase constructs containing the Puma wild type 3ЈUTR (Puma-WT-3ЈUTR), and this increase was attenuated by miR-149. However, this effect could not be observed in Puma mutated 3ЈUTR (Puma-MUT-3ЈUTR) (Fig. 3G). Moreover, pc3-miR-149-A caused many more reductions in the luciferase activity of pGL3-WT-Puma-3ЈUTR than pc3-miR-149-G (Fig. 3H). In addition, we infected the cells with adenoviruses containing Puma with wild type 3ЈUTR (Puma-W-3ЈUTR), with mutated 3ЈUTR (Puma-MUT-3ЈUTR), or without 3ЈUTR (Puma-W/O-3ЈUTR) (Fig. 3I). Immunoblot showed that miR-149 could subdue the expression of Puma with 3ЈUTR but not Puma without 3ЈUTR or Puma with mutated 3ЈUTR (Fig. 3J). Next, we infected the cells with adenoviral Puma-W-3ЈUTR along with adenoviral miR-149-A or miR-149-G. The infection/expression efficiency and the levels of miR-149 were detected (Fig. 3K). miR-149-A repressed Puma expression at a greater degree than miR-149-G (Fig. 3L). These data indicate that miR-149 directly regulates Puma expression, and the SNP rs71428439 is able to influence the effect of pre-miR-149 on the expression levels of Puma.
Puma Regulates Mitochondrial Network-To explore the underlying mechanism by which Puma initiates the apoptotic program, we produced Puma siRNA (Si-Puma) and its scrambled form (Scramble). Si-Puma but not its scrambled form could abrogate H 2 O 2 -induced Puma expression (Fig. 4A).

miR-149 Regulates Apoptosis
Enforced expression of Puma could initiate mitochondrial fission, whereas knockdown of Puma attenuated mitochondrial fission induced by H 2 O 2 treatment (Fig. 4B). Puma overexpression promoted caspase 3 activation, and the abrogation of Puma impaired caspase 3 activation induced by H 2 O 2 treatment (Fig. 4C). Concomitantly, overexpression of Puma induced apoptosis, whereas knockdown of Puma abolished apoptosis caused by H 2 O 2 stimulation (Fig. 4D). These results

miR-149 Regulates Apoptosis
indicate a critical role for Puma in regulating mitochondrial network and apoptosis responding to H 2 O 2 induction.
Puma Is Required for miR-149 to Control Mitochondrial Network and Apoptosis-We examined the relationship between Puma and miR-149 in regulating the mitochondrial network and apoptosis. As shown in Fig. 5A, the inhibition of endogenous miR-149 could elevate caspase 3 activation induced by H 2 O 2 . When Puma was knocked down, inhibition of miR-149 failed to promote caspase 3 activation induced by H 2 O 2 . Intriguingly, miR-149 inhibition could promote mitochondrial fission and apoptosis, and this effect was abolished upon knockdown of Puma (Fig. 5B), suggesting that Puma is a target of miR-149 in regulating mitochondrial network and apoptosis.
To test whether Puma is a direct or indirect target of miR-149 in eliciting its effect, we employed the constructs of adenoviral Puma-W-3ЈUTR and Puma-W/O-3ЈUTR. miR-149 significantly reduced caspase 3 activity induced by Puma-W-3ЈUTR but not Puma-W/O-3ЈUTR (Fig. 5C). Concomitantly, miR-149 could significantly decrease mitochondrial fission and apoptosis induced by Puma-W-3ЈUTR but failed to attenuate mitochondrial fission and apoptosis caused by Puma-W/O-3ЈUTR (Fig. 5D). To test whether the SNP could elicit effects on apoptosis, we infected cells with adenoviral miR-149-A or miR-149-G. It was observed that compared with miR-149-A, miR-149-G caused a lesser reduction of caspase 3 activation (Fig. 5E), mitochondrial fission, and apoptosis (Fig. 5F) induced by enforced expression of Puma with wild type 3ЈUTR. These data suggest that Puma is a direct target of miR-149 in regulating the mitochondrial network and that the allelic G impairs the ability of pre-miR-149 to repress the mitochondrial network and apoptosis.
Inverse Expression of miR-149 and Puma in Ischemia-Finally, we attempted to know the expression levels of miR-149 and Puma under the pathological conditions. The pathophysiology of myocardial infarction involves a reduced blood supply to the myocardium. Ischemia can induce cardiomyocyte apoptosis (28). We tested their levels in the ischemic conditions. miR-21 has been verified to be highly expressed in the heart (20). By comparing with miR-21 expression levels in the heart, miR-149 was observed to possess relatively high expression levels in the heart under physiological conditions (Fig. 6A). However, miR-149 showed a low level in the area at risk in the ischemic heart (Fig. 6B). In contrast, Puma increased under ischemic conditions (Fig. 6B). Taken together, these results suggest that miR-149 and Puma show an inverse expression level in ischemia.
A and G Allele of rs71428439 Polymorphism Affects Myocardial Infarction in Mouse Models-We assessed whether miR-149 can influence myocardial infarction and cardiac function. Puma was increased markedly after the MI surgery, which was much more attenuated by miR-149-A than miR-149-G (Fig.  7A). The cardiac functions were more significantly preserved in the mice receiving miR-149-A (Fig. 7B). The infarct size was much reduced in mice receiving miR-149-A (Fig. 7C). Consistently, mice receiving miR-149-A displayed a lower degree of myocyte apoptosis (Fig. 7D). The results in the mouse model motivated us to further compare the clinical features of the MI patients based on the miR-149 SNP genotype. It was observed that fractional shortening in patients carrying the G allele in the genome was significantly lower than that of patients carrying the A allele (Fig. 7E). Together, the results indicate that the SNP rs71428439 is able to influence the effect of pre-miR-149 on myocardial infarction.

DISCUSSION
SNPs in miRNA genes, including pri-miRNAs, pre-miRNAs, and mature miRNAs, may influence the processing and/or target selection of miRNAs. In this study, we first showed that SNP rs71428439 located in miR-149 precursor can affect the expression of miR-149. In particular, it is associated with cardiac infarction. In exploring the underlying molecular mechanism, we found that Puma is a target of miR-149. The reduced expression of miR-149 leads to the up-regulation of Puma in response to apoptotic stimulation. miR-149 regulates mitochondrial fission and apoptosis through Puma. Our results reveal a novel model of SNP that can affect mitochondrial dynamics.
Although the association between the SNPs in protein-coding genes and the risk of myocardial infarction has been extensively investigated, there are few reports showing that myocardial infarction is in association with the SNPs in miRNA genes. This study demonstrates that the rare allele of SNP rs71428439 located in miR-149 precursor is closely associated with the risk of myocardial infarction. We observed that the individuals with GG genotype of miR-149 gene are at a higher risk for myocardial infarction compared with the individuals with the AA genotype. Intriguingly, we found that an A to G variation in the miR-149 precursor caused by SNP rs71428439 affects the production of mature miR-149. The G-allelic pre-miR-149 exhibits a lower expression level of mature miR-149 compared with the A-allelic one. This effect may have stemmed from the minimal free-energy secondary structure change of the precursor caused by the A to G variation. A growing body of evidence demonstrates that the integrity of the precursor RNA stem is required for mature miRNA production. For example, the introduction of artificial mutations on the miR-30 precursor are detrimental to mature miR-30 processing (29). It has been reported that the polymorphism rs2910164 located in the stem region of miR-146a precursor influences the production of mature miR-146a (12,30). The G to U variation in the miR-125a precursor results in a reduced production of mature miR-125a (16).
We investigated the biological function of miR-149 and found that miR-149 can conquer mitochondrial related apoptosis through targeting the pro-apoptotic protein Puma. The suppression of miR-149 renders cardiomyocytes more sensitive to undergo apoptosis, although the enforced expression of miR-149 reduces apoptosis. The miRNAs execute their functions by inhibiting their target gene expression (31). This work for the  . Inverse expression of miR-149 and Puma in ischemic myocardial tissue. A, endogenous miR-149 is highly expressed in mouse heart. miR-149 level was detected in different mouse organs by RT-PCR. The data were normalized to U6. miR-149 levels were compared with miR-21 levels in mouse heart (n ϭ 6). B, expression of miR-149 and Puma are inversely related during myocardial ischemia. The areas at risk were collected at the indicated ischemic time. miR-149 levels were analyzed by RT-PCR (n ϭ 6), *, p Ͻ 0.05 versus control. Puma was analyzed by immunoblot.

miR-149 Regulates Apoptosis
first time shows that miR-149 modulates Puma expression in cardiomyocytes. Puma belongs to the Bcl-2 homology 3-only subfamily of Bcl-2-related pro-apoptotic proteins. Recently, it was reported that Puma is required for cardiac cell death upon ischemia/reperfusion in mouse hearts (32). Puma null mice are significantly resistant to ischemia-reperfusion injury. The infarct sizes in the Puma-null hearts are reduced by 50% compared with the wild type or heterozygous mice. Also, Puma is verified to participate in mediating cardiomyocyte apoptosis induced by endoplasmic reticulum stress (33). In this study, we observed that the expression of Puma is negatively regulated by miR-149.
Mitochondria constantly undergo fusion and fission, two necessary processes for the maintenance of organelle fidelity (34). However, the abnormal mitochondrial fission is involved in the initiation of apoptosis (35)(36)(37). We have previously shown that miRNAs can regulate mitochondrial fission and apoptosis in cardiomyocytes by targeting mitochondrial fission proteins (24). In this study, we have found that miR-149 prevents mitochondrial fission in cardiomyocytes, and it regulates mitochondrial fission by targeting Puma. Puma is exclusively located to mitochondria. It is necessary to find out the molecular mechanism by which Puma regulates mitochondrial fission in a future study.
A recent report has summarized the myocardial infarctionrelated miRNAs, and miR-149 is one of them (38), but the underlying mechanism remains largely unknown. This study offered evidence to support the point that miR-149 may play a vital role in controlling apoptosis during myocardial infarction. Furthermore, this work revealed that the A-and G-allelic miR-149 precursors display an obvious difference in their abilities to regulate apoptosis. A sudden insufficient blood supply to the myocardium is a major reason of myocardial infarction, resulting in a loss of cardiomyocytes. Cellular apoptosis is one typical pathological feature of myocardial infarction. Our data suggest that the SNP rs71428439 executes functional influence on apoptosis by conferring different levels of miR-149, which causes the different levels of Puma.
miR-149 has been reported to be down-regulated both in mice heart tissue and human heart tissue of MI (39). In this study, we explored the role of miR-149-Puma axis in myocardial infarction in the mouse model. The employment of human