MicroRNA miR-199a* Regulates the MET Proto-oncogene and the Downstream Extracellular Signal-regulated Kinase 2 (ERK2)*

MicroRNAs (miRNAs) constitute a class of small noncoding RNAs that play important roles in a variety of biological processes including development, apoptosis, proliferation, and differentiation. Here we show that the expression of miR-199a and miR-199a*(miR-199a/a*), which are processed from the same precursor, is confined to fibroblast cells among cultured cell lines. The fibroblast-specific expression pattern correlated well with methylation patterns: gene loci on chromosome 1 and 19 were fully methylated in all examined cell lines but unmethylated in fibroblasts. Transfection of miR-199a and/or -199a* mimetics into several cancer cell lines caused prominent apoptosis with miR-199a* being more pro-apoptotic. The mechanism underlying apoptosis induced by miR-199a was caspase-dependent, whereas a caspase-independent pathway was involved in apoptosis induced by miR-199a* in A549 cells. By employing microarray and immunoblotting analyses, we identified the MET proto-oncogene as a target of miR-199a*. Studies with a luciferase reporter fused to the 3′-untranslated region of the MET gene demonstrated miR-199a*-mediated down-regulation of luciferase activity through a binding site of miR-199a*. Interestingly, extracellular signal-regulated kinase 2 (ERK2) was also down-regulated by miR-199a*. Coordinated down-regulation of both MET and its downstream effector ERK2 by miR-199a* may be effective in inhibiting not only cell proliferation but also motility and invasive capabilities of tumor cells.

function as regulators, primarily by inhibiting the translation of target mRNA in the cytoplasm (1,2). miRNAs inhibit the expression of target genes by affecting the translation and/or stability of target mRNA through binding to a target site in the 3Ј-UTR of target mRNA (3). More than 500 human miRNAs have been identified, and over 1000 miRNAs are predicted to exist in humans (4,5). Recently, it was reported that miRNAs can regulate the expression of up to 30% of total cellular proteins, thus constituting a vast gene regulatory network (6). miRNAs are transcribed primarily by RNA polymerase II, processed into pre-miRNA by a microprocessor complex comprising Drosha and DGCR8 in the nucleus, exported out of the nucleus by exportin 5, and then processed into single-stranded miRNA by Dicer in the cytoplasm (7)(8)(9)(10)(11)(12). Accumulating evidence indicates that the expression of miRNA is deregulated in many cancers (13,14). miRNA appears to be an important determinant in diagnosis and prognosis of a variety of diseases and may have substantial value as a treatment for incurable diseases such as cancer and viral diseases (15)(16)(17)(18).
Although hundreds of miRNAs have been discovered, relatively little is known about their biological functions. The partially complementary binding of miRNAs to their target sites in the 3Ј-UTR of target genes makes it difficult to predict target genes using a computer program. In the case of cancers, accumulating evidence suggests that miRNAs can function as tumor suppressors or oncogenes. For example, let-7, which is downregulated in lung cancer, targets RAS (19) and HMGA2 (20,21) oncogenes, and miR-15 and miR-16, which are down-regulated in chronic lymphocytic leukemia, control BCL2 at the posttranscriptional level (22). In addition, miR-127 and miR-124a are epigenetically regulated, and down-regulate BCL6 and CYCLIN D KINASE 6, respectively (23,24).
MET proto-oncogene is a transmembrane tyrosine kinase receptor for hepatocyte growth factor (HGF) ligand (25). MET is involved in the control of invasive growth not only during tumorigenesis but also in embryonic development, organ development, inflammatory responses, and wound healing processes (26,27). There is considerable evidence that MET is involved in the initiation and progression of tumors. 1) Transgenic mice overexpressing MET or HGF form metastatic tumors (28); 2) MET is overexpressed or constitutively activated by mutations in many kinds of cancer (29 -32); and 3) down-regulation of MET by RNA interference resulted in the inhibition of cell proliferation and tumor invasion and induced substantial apoptosis (33)(34)(35)(36)(37). Accordingly, MET is considered an important target for anti-cancer therapy, and inhibitors of MET signaling such as ligand antagonists (38 -40), kinase inhibitors (41), and receptor competitors (42,43) have been developed. In addition, the inhibition of MET expression by siRNA technology may be a useful approach.
The present study demonstrates that MET proto-oncogene is negatively regulated by miR-199a*. The expression of miR-199a/a* was silent in proliferating cells except fibroblasts and in accordance with its silent expression, the miR-199a/a* locus was heavily methylated in non-expressing cell lines. When introduced into tumor cells, miR-199a/a* induced pronounced apoptosis, suggesting that miR-199a/a* is a putative tumor suppressor. Intriguingly, mitogen-activated protein kinase (MAPK) ERK2, which is one of the downstream effectors of MET, was also down-regulated by miR-199a*. Thus, promoting apoptosis through the inhibition of the MET signaling pathway by miR-199a/a* may hold great promise as a potential therapy for a variety of primary and metastatic tumors.
Quantitative Real-time PCR Analysis of MicroRNAs-Total RNAs were isolated from cultured cells using the mirVana miRNA isolation kit (Ambion, Austin, TX). Total RNAs from human tissues were obtained from Ambion. For quantitative analysis of miRNAs, two-step TaqMan real-time PCR analysis was performed using primers and probes obtained from Applied Biosystems (Foster City, CA). Briefly, cDNA was made from total RNA in 15-l reactions using murine leukemia virus reverse transcriptase and specific primers for each miRNA contained in the TaqMan MicroRNA Reverse Transcription kit (Applied Biosystems). The reverse transcriptase reaction was performed by sequentially incubating at 16°C for 30 min, 42°C for 30 min, and 85°C for 5 min. Each PCR mixture (20 l) contained 1.3 l of reverse transcriptase product, 10 l of Taq-Man 2ϫ Universal PCR Master Mix, and 1 l of the appropriate TaqMan MicroRNA Assay (20ϫ) containing primers and probes for the miRNA of interest. The mixture was initially incubated at 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 60 s. PCR were performed in triplicate using a DNA Engine Opticon system (MJ Research Inc.). All RNA samples were normalized relative to human 18 S rRNA.
DNA Methylation Analysis Using Combined Bisulfite Restriction Analysis (COBRA) and Bisulfite Sequencing-Genomic DNA was extracted from cultured cells using the DNA Wizard Genomic DNA Purification Kit (Promega). DNA from human normal and tumor tissues were obtained from Biochain (Hayward, CA). Prior to treatment with bisulfite, DNA was digested with NcoI (New England Biolabs, Ipswich, MA). The digested DNA was purified with phenol/chloroform extraction, precipitated with ethanol, and resuspended in Tris-EDTA buffer. A bisulfite reaction was performed with 0.8 g of NcoI-digested DNA using EZ DNA Methylation Kit (Zymo Research, Orange, CA). After bisulfite reaction and purification, 250 ng of bisulfite-converted DNA was used as a template for each PCR analysis. PCR was carried out in a 20-l reaction mixture containing specific primers (10 pmol each) using AccuPower PCR premix (Bioneer, Korea). The amplification cycle was 94°C for 40 s, 57°C for 50 s, and 72°C for 60 s for 35 cycles. Primers used to amplify the DNA fragment (19-F1) containing the miR-199a/a* genomic locus on chromosome 19 were 5Ј-GGTGGTGGAA-AATGATATTTATTTG-3Ј and 5Ј-AAATTTCCTAAAAAC-CCAAAACTTT-3Ј. Primers used to amplify the DNA fragment (19-F2) upstream of the miR-199a/a* locus on chromosome 19 were 5Ј-TTTTGGGTTTTTAGGAAATTTT-AAAG-3Ј and 5Ј-AATCACAAACCATTCCAACTAATAC-3Ј. Primers used to amplify the DNA fragment (1-F2) upstream of the miR-199a/a* locus on chromosome 1 were 5Ј-TGAAT-AGGTAGTTTGAATATTGGGG-3Ј and 5Ј-CATATATAAA-CTCTCCAACCCAACC-3Ј. Primers used to amplify the DNA fragment (1-F3) upstream of the miR-199a/a* locus on chromosome 1 were 5Ј-GGGTTGGGTTGGAGAGTTTAT-ATAT-3Ј and 5Ј-ACTTTTCCATACTAAAACCCACTTC-3Ј. A DNA fragment (1-F1) containing the miR-199a/a* locus on chromosome 1 was amplified using nested PCR. First, DNA fragment was amplified using a forward primer (5Ј-TGGAAA-TAGTTTATTTTGTTTTTAG-3Ј) and a reverse primer (5Ј-CAACCCTTAAATATATCTCAATCAAC-3Ј). The product (1 l) of the initial PCR was then used as a template for a second PCR to amplify DNA fragment 1-F1, using a forward primer (5Ј-TGGAAATAGTTTATTTTGTTTTTAG-3Ј) and a nested reverse primer (5Ј-AAAAAACTTCTAAAAATCCTACTCC-3Ј). After PCR, reaction mixtures were purified using the DNA Clean and Concentrator kit from Zymo Research. The purified PCR fragments were cloned into pGEM-T easy vector (Promega), and individual clones were sequenced. For methylation analysis by COBRA (45), PCR fragment (19-F2) was further amplified using nested primers: a forward primer, 5Ј-ATTTT-AAAGAGTGGGGGAGG-3Ј and a reverse primer, 5Ј-CACA-TCTAAAACTATTTACA-3Ј. After nested PCR, 3 l of reaction mixtures, without purification, were digested with HhaI (New England Biolabs) for 3 h at 37°C. The restriction products were electrophoresed on a 3% agarose gel and visualized by staining with ethidium bromide.
Transfection-One day before transfection, cells were seeded into 6-well tissue culture plates at a density of 5-8 ϫ 10 4 cells per well for RNA preparation at 48 -72 h post-transfection. Transfection of A549 and HeLa cells was accomplished using Lipofectamine RNAiMAX transfection reagent (Invitrogen) as described previously (46). Briefly, cells were washed once with Opti-MEM (Invitrogen) and 500 l of Opti-MEM was added to each well. For each transfection, 100 nM miRNA duplexes (miRIDIAN miRNA mimetics; Dharmacon) and siRNAs in 250 l of Opti-MEM were mixed with 250 l of Opti-MEM containing 3.5 l of RNAiMAX and incubated for 20 min at room temperature. The mixture was then added to cells in the 6-well plate, giving a transfection volume of 1 ml. After 6 h incubation with the transfection solution, the Opti-MEM medium containing the complexes was replaced with 2 ml of standard growth media and cultured at 37°C. Cells were harvested at different time points for microarray and other experiments.
Assessment of Apoptosis-Cells were seeded into 24-well plates at a density of 1.8 ϫ 10 4 cells per well and transfected after 24 h incubation. Three days after transfection, cell morphology was examined by light microscopy. For Annexin V staining, transfected cells were stained with Annexin V conjugated with Alexa 568 (Roche Applied Science) following the manufacturer's protocol. For the experiments with the caspase inhibitor, cells were pretreated with 100 M Z-VAD-fmk (Calbiochem) for 1 h before transfection.
RNA Isolation and Microarray Experiments-Total RNA was isolated from transfected cells by using the RNeasy mini kit (Qiagen, Hilden, Germany). Amplified and biotinylated cRNA was generated from 500 ng of total RNA using the Illumina TotalPrep RNA Amplification Kit (Ambion). After purification with the RNeasy kit, 700 ng of labeled cRNA was hybridized to the Illumina HumanRef-8 BeadChip following the manufacturer's instructions (Illumina, Inc., San Diego, CA). After hybridization with strepavidin-Cy3 and washing, the arrays were scanned with an Illumina BeadArray Reader and data analysis was carried out using the Illumina BeadStudio program. To validate microarray data, quantitative reverse transcriptase-PCR was performed using gene-specific primers as described previously (46).
Western Blot Analysis-One day before transfection, cells were seeded into 10-cm culture dishes at a density of 5-8 ϫ 10 5 cells per dish for protein preparation at 48 or 72 h post-transfection. Whole cell lysates were prepared using CelLytic-M cell lysis reagent (Sigma) following the manufacturer's instruction. Total protein extracts (30 g) were electrophoresed on 4 -20% precast protein gel (Pierce) and transferred to Protran Nitrocellulose membrane (S&S, Dassel, Germany). The blots were incubated first with a primary antibody and then, with horseradish peroxidase-conjugated secondary antibody (Pierce). Luciferase Reporter Assay-A luciferase reporter carrying the 3Ј-UTR of MET was constructed as follows. An 1850-bp fragment containing the majority of the 3Ј-UTR fragment of the MET gene, including the predicted target site for miR-199a*, was amplified by PCR from a full-length MET cDNA clone (NM_000245, obtained from RZPD, Imagenes GmbH, Berlin, Germany) using F1 and R1 primers that create XhoI and NotI sites, respectively (see supplemental Table S1 for primer sequences). The XhoI-NotI-digested product was cloned into the 3Ј-UTR of the luciferase gene in the psiCHECK-2 vector (Promega). To delete the predicted miR-199a* target site from the 3Ј-UTR fragment of MET, DNA fragments containing 1035 bp upstream and 795 bp downstream of the target site were separately amplified using F1/R2 and F2/R1 primers, respectively. After digestion with XbaI, ligation of upstream and downstream fragments, and digestion with XhoI and NotI, the 3Ј-UTR fragment with a deletion in the miR-199a* binding site was cloned into XhoI/NotI-digested psiCHECK-2. The cells were cotransfected in 6-well plates using Lipofectamine 2000 (Invitrogen) with 400 ng of the 3Ј-UTR-luciferase report vector and 10 nM miRNA mimetics or negative control mimetic (Dharmacon). Forty-eight h after transfection, firefly and Renilla luciferase activities were measured consecutively by using dual-luciferase assays (Promega) according to the manufacturer's protocol.

Expression of miR-199a/a* Is Confined to Fibroblasts among
Proliferating Cell Lines-We recently constructed a library of small RNAs from synovial tissue isolated from rheumatoid arthritis (RA) patients. Sequencing of 2000 clones revealed that miR-199a/a* is one of several highly expressed miRNAs in RA miR-199a* Regulates Met Proto-oncogene synovial tissues. This finding led us to examine miR-199a/a* expression in a variety of tissues and cell lines using quantitative real-time PCR methods.
As shown in Fig. 1, miR-199a/a* was well expressed in synovial fibroblasts and CCD normal skin fibroblasts. Interestingly, the expression of miR-199a/a* was almost completely silenced in all other examined tumor and normal cell lines. Because even the precursor forms of miR-199a/a* were not detected by Northern blot analysis (supplemental Fig. S1), it is likely that the expression is silenced at the transcriptional level. In the tested tissues, the expression of miR-199a/a* was relatively high in breast, colon, and testis and relatively low in thymus, kidney, and liver. Expression was very low in brain, suggesting possible negative effects on brain function.
miR-199a/a* Loci on Chromosome 1 and 19 Are Fully Methylated in Proliferating Cell Lines Except Fibroblasts-Because the expression of miR-199a/a* is almost completely silenced in proliferating normal and cancer cell lines except fibroblasts, and DNA methylation is one of the fundamental mechanisms involved in gene silencing (47,48), we determined the methylation patterns of the miR-199a/a* gene using bisulfite sequenc-ing and COBRA (45). The miR-199a/a* gene exists on both chromosomes 1 and 19 as a consequence of gene duplication: one locus is located in intron 15 of DYNAMIN 2 gene on chromosome 19 and the other in intron 14 of DYNAMIN 3 on chromosome 1.
As shown in Fig. 2, CpG density is high in the miR-199a/a* locus on chromosome 19 and a predicted CpG island (CpG island searcher) is present between ϳ130 and 540 bp upstream of the mature miR-199a sequence (another program, the CpG plot program at EBI predicted a 134-bp CpG-dense region between 269 and 402 bp upstream of mature miR-199a). In contrast, the miR-199a/a* locus on chromosome 1 is relatively CpG poor. Sequencing analysis of 20 clones obtained from the genomic DNA of A549 and HeLa cells following bisulfite conversion and PCR amplification revealed that 94.7 and 94.6% of analyzed CpG sites around the miR-199a/a* locus on chromosome 19 were methylated in A549 and HeLa cells, respectively (Fig. 2). On chromosome 1, a similar level of DNA methylation was detected: 94.6 and 97.6% in A549 and HeLa cells, respectively. In contrast, the same CpG sites were unmethylated in synovial fibroblasts and CCD-986sk skin fibroblast cells. Therefore, there is a good correlation between DNA hypomethylation and miR-199a/a* gene expression.
Next, we extended our methylation analysis to additional normal and cancer cell lines using bisulfite PCR followed by COBRA. The COBRA technique allows the quantification of the methylation level in CpG sites within CpG islands by assessing the availability of CpG-containing restriction sites for enzymatic cleavage following bisulfite treatment. In this case, GCGC sites (HhaI restriction site) were assessed. GCGC sites can only be cleaved by HhaI if the internal CpG is methylated, which protects the cytosine from conversion to uracil during bisulfite treatment. The ratio of restricted (methylated) to unrestricted (unmethylated) PCR products is a quantitative measure of methylation (see top map in Fig. 2 for the location of 5 CpG sites analyzed by COBRA). As shown in Fig. 3A, the analyzed CpG sites on chromosome 19 were almost completely methylated in all examined normal and tumor cell lines (except fibroblasts) as indicated by complete digestion of the PCR fragment by HhaI. Due to lack of enzyme sites suitable for COBRA analysis, the methylation patterns of 5 CpG sites upstream of the miR-199a/a* locus on chromosome 1 were examined by bisulfite sequencing. As shown in Fig. 3E, the locus on chromosome 1 was heavily methylated in all examined cell lines except K562 erythroleukemia cells. The hypomethylation of the miR-199a/a* locus on chromosome 1 appears to be specific to K562 cells because the same locus was heavily methylated in Jurkat (T cell lymphoma) and Raji (B lymphocytes from Burkitt's lymphoma) cells as well as peripheral blood leukocytes (supplemental Fig. S2). These results suggest a critical role of DNA methylation in suppressing the expression of the miR-199a/a* gene in most proliferating cells except fibroblast cells.
To examine the functional significance of DNA methylation in transcriptional repression of the miR-199a/a* gene, A549 cells were treated with the DNA methyltransferase inhibitor, 5-aza-2Ј-deoxycytidine (5-azaC), and/or the histone deacetylase inhibitor, trichostatin A, and the level of miR-199a/a* transcripts was measured by quantitative real-time PCR. As shown  Expression profiles of miR-199a and -199a*. A, DNase-treated RNA (0.1 g) from normal and tumor cell lines was reverse transcribed and amplified with miRNA-specific primers and TaqMan probes (Applied Biosystems). Relative expression was calculated using the comparative cycle threshold method and miRNA abundance was normalized relative to human 18S rRNA. Results are expressed as the percentage of miRNA levels relative to that in NS1, which was set at 100%. NS1, normal synoviocyte isolated from individual 1; NS2, normal synoviocyte isolated from individual 2; RS1, RA synoviocyte isolated from patient 1; RS2, RA synoviocyte isolated from patient 2; CCD, CCD-986sk. The expression of miR-199a and -199a* was almost completely silenced in other cancer cell lines including MCF7, K562, KB, JEG3, JAR, SiHa, and SK-OV-3 (data not shown). B, miR-199a and -199a* expression in human tissues. RS, RA synovial tissue isolated from 6 patients; THY, thymus; LIV, liver; KID, kidney; BRN, brain; TES, testis; SPL, spleen; BRT, breast; COL, colon. miRNA levels in RS was set at 100%. Filled bar, miR-199a; open bar, miR-199a*. JUNE 27, 2008 • VOLUME 283 • NUMBER 26 in supplemental Fig. S3, the miR-199a/a* genes were significantly reactivated by treatment with 5-azaC alone or 5-azaC followed by trichostatin A. These results indicate that DNA methylation is one mechanism that transcriptionally silences miR-199a/a* genes in A549 cells.

miR-199a* Regulates Met Proto-oncogene
In adult tissues, the miR-199a/a* loci on chromosomes 1 and 19 were heavily methylated in most tissues with breast, colon, and testes being less methylated (Fig. 3, B and F). Consistent with a low level of expression in brain, the miR-199a/a* gene was heavily methylated in that tissue. In fetal tissues, the miR-199a/a* gene was heavily methylated in brain, and less methylated in lung (Fig. 3C). Notably, the miR-199a/a* gene was hypomethylated in fetal skin, which is in good agreement with the lack of methylation in fibroblast cells (supplemental Fig.  S4). In breast tumor tissues, the miR-199a/a* gene was slightly more methylated in 4 of 5 examined tissues compared with matched normal breast tissues (Fig. 3D).
miR-199a/a* Induce Apoptosis in A549 Cells-The silencing of miR-199a/a* in most proliferating cell lines prompted us to investigate whether miR-199a/a* functions as a tumor suppressor. As shown in Fig. 4A, the transient introduction of miR-199a and/or -199a* induced apoptosis in A549 cells compared with cells transfected with negative control mimetic, as assessed by cell morphology and Annexin V staining. miR-199a* caused more pronounced apoptosis than miR-199a in several cancer cell lines such as A549, PC3, KB, and MCF7 cells ( Fig. 4A and supplemental Fig. S5) and its pro-apoptotic activity was concentration dependent (data not shown).
To examine the molecular mechanism(s) by which miR-199a and -199a* induce apoptosis in cancer cells, Western blot analysis was performed on samples isolated from A549 cells transfected with miR-199a and/or -199a* mimetics at 72 h after transfection. First, the activation of caspase enzymes via a proteolytic cascade was examined by measuring cleavage of caspases and their substrate, poly(ADP-ribose) polymerase (PARP). Cleavage of PARP by caspases is recognized as a hallmark of apoptosis (49). As can be seen in Fig. 4B, PARP was cleaved when cells were transfected with miR-199a but not when they were transfected with miR-199a*. Consistent with PARP cleavage, all examined caspases including initiator (CASPASE 9 and CASPASE 8) and effector (CASPASE 3 and CASPASE 7) were cleaved in cells transfected with miR-199a. These results indicate that the apoptosis pathway induced by miR-199a is caspase-dependent, whereas miR-199a* causes a caspase-independent cell death pathway in A549 cells. To further clarify the mechanisms underlying miR-199a-or -199a*-mediated apoptosis, we next examined the relative levels of pro-apoptotic and anti-apoptotic proteins. As depicted in Fig. 4B, the level of the pro-apoptotic BAX protein was slightly increased in cells transfected with miR-199a, whereas levels of BID and BAD, other pro-apoptotic proteins, were increased when cells were transfected with miR-199a*. The apoptosis inducing factor, which plays an important role in caspase-independent apoptosis, was decreased in cells transfected with miR-199a*. Notably, the amount of anti-apoptotic BCL-X L protein was significantly reduced in cells transfected with miR-199a, suggesting that decrease in the level of BCL-X L may be responsible for the miR-199a-induced apoptosis. Although treatment with Z-VAD-fmk, a pan-caspase inhibitor, inhibited miR-199amediated apoptosis in A549 cells, the same treatment did not block miR-199a*-mediated apoptosis, as expected, given the lack of caspase activation in cells treated with miR-199a* (data not shown).
MET Is a Target of miR-199a*-To identify gene targets regulated by miR-199a/a*, we used a microarray approach to characterize changes in mRNA levels after transfection with miR-199a and -199a* mimetics. Although miRNAs regulate gene expression mostly at the translational level, it is becoming clear that miRNAs can also negatively regulate gene expression through effects on mRNA degradation (3, 50 -52). After transfection with miR-199a and -199a* mimetics, we identified hundreds of genes with altered mRNA expression profiles (73 and 24 genes were up-regulated, and 318 and 183 genes were down-

miR-199a* Regulates Met Proto-oncogene
regulated in cells transfected with miR-199a* and -199a mimetics, respectively). By comparing these genes with miR-199a/a* gene targets predicted by a web-based computer program at the Sanger Institute (miRanda algorithm), we narrowed down the number of putative target genes to 38 genes for miR-199a* and 23 genes for miR-199a (see supplemental Table S2 for a list of putative target genes). Subsequent target validation by Western blot analysis revealed that the MET proto-oncogene is downregulated in A549 and HeLa cells transfected with miR-199a* (Fig. 5A).
To demonstrate that MET is a target of miR-199a*, effects of miR-199a/a* mimetics on expression of the luciferase gene were determined using a MET 3Ј-UTR-luciferase reporter construct. The reporter plasmid bearing the 3Ј-UTR of the MET gene was cotransfected with miR-199a*, miR-199a, or negative control mimetics and luciferase activity was measured using a dual luciferase assay at 48 h after transfection. When cotransfected with miR-199a*, luciferase activity was decreased to 19% of the level in cells transfected with a negative control (Fig. 5B). More importantly, luciferase activity was not reduced when the predicted binding site for miR-199a* was deleted from the 3Ј-UTR. Very similar results were obtained using a shorter MET 3Ј-UTR fragment (data not shown). Taken together, immunoblotting and reporter assay data indicate that the MET proto-oncogene is a target of miR-199a*.
Although the expression of miR-199a/a* was silent in many proliferating cells, miR-199a/a* was significantly expressed in fibroblasts. To determine whether inhibition of miR-199a* in fibroblasts can lead to an increase in the level of MET protein, normal synovial fibroblasts were treated with anti-miR-199a* inhibitor for 24 h and the MET protein was detected using Western blot analysis. Consistent with the role of miR-199a* in down-regulating MET, treatment of fibroblasts with the anti-miR-199a* inhibitor resulted in up-regulation of MET, indicating a relief in miR-199a*-mediated translational inhibition by anti-miR-199a* inhibitor (supplemental Fig. S6).
ERK2 Is Down-regulated by miR-199a*-To examine the effects of MET down-regulation on signaling pathways regulated by MET, we performed Western analysis to determine changes in the amount and phosphorylation of AKT and ERK1/2, which are two major effector molecules downstream of MET. As depicted in Fig. 6A, the Akt level was slightly decreased in cells transfected with miR-199a. Interestingly, the level of ERK2 was significantly decreased in cells transfected with miR-199a*, but ERK1 levels were unaffected, and levels of MAPKs SAPK/JNK and p38 were also unaffected.
Because MET is down-regulated by miR-199a*, it may be that the decrease in signaling from MET may result in the decrease in the level of ERK2. To test this possibility, we transfected 3 siRNAs targeting MET mRNA into A549 cells and assayed their effect on ERK2 levels. As shown in Fig. 6C, MET protein was down-regulated by each of the 3 siRNAs, but ERK2 protein was unaffected. These data suggest that the down-regulation of MET is not a cause of the decrease in ERK2 protein.

miR-199a* Regulates Met Proto-oncogene
As expected from the reduced level of ERK2, the amount of phosphorylated ERK2 was also reduced in cells transfected with miR-199a* (Fig. 6B). Thus, miR-199a* controls MET signaling at multiple points including affecting MET receptor levels and affecting levels of the downstream effector ERK2.

DISCUSSION
In this study, we show that MET proto-oncogene is a target of miR-199a*. Several lines of evidence indicate that miR-199a* is a putative tumor suppressor. 1) The expression of miR-199a/a* is silenced in all proliferating cell lines tested except fibroblasts. 2) Introduction of miR-199a/a* caused apoptosis in cancer cells. 3) miR-199a* down-regulates MET proto-oncogene and also down-regulates ERK2, an effector downstream of MET.
In support of our finding that miR-199a/a* is a putative tumor suppressor, it was recently reported that the expression of miR-199a/a* was significantly reduced in ovarian cancer and hepatocellular carcinoma when compared with normal tissues (53,54). Another interesting role of miR-199a* was recently reported. The miR-199a* homolog in mouse, mmu-miR-199a*, regulates COX-2 expression in mouse uterus during implantation (55).
miR-199 genes are located within the intron of DYNAMIN genes that are large GTPases involved in vesicle formation during receptor-mediated endocytosis (56). There are three DYNAMIN genes in mammals and each human DYNAMIN gene contains an miR-199 gene (57). In human, more than 25% of miRNAs are located in introns of coding genes and in many cases, miRNAs are coexpressed with their host transcripts (58). However, the miR-199a/a* gene is located in the opposite direction to the DYNAMIN genes suggesting it is transcribed from its own promoter. The miR-199a/a* gene on chromosome 1 is relatively CpG poor and not associated with a CpG island. Interestingly, the miR-214 gene is located ϳ5.7 kb downstream of miR-199a/a* gene in the same direction of transcription, suggesting cotranscription by the same promoter. This hypothesis is supported by a similar pattern of expression of miR-199a/a* and miR-214 across a variety of tissues and cell lines (59,60). Analysis of transcripts after knockdown of Drosha by siRNA or transfection of a large genomic DNA fragment containing upstream of the miR-199a/a* gene may be necessary for promoter analysis.
MET tyrosine kinase receptor is frequently amplified or mutated in a variety of solid tumors and the inhibition of its amplified signaling leads to the inhibition of cell proliferation and metastasis. Hence, it is an attractive target for anti-cancer therapy and more than 5 MET inhibitors including monoclonal antibodies and small molecule inhibitors are currently being tested in clinical trials. Recently, MET amplification has been implicated in drug resistance to the epidermal growth factor receptor kinase inhibitors gefitinib and erlotinib in lung cancers (61). ERBB3 (HER3)-dependent activation of phosphatidylinositol 3-kinase, which is triggered by amplified MET receptor, appeared to confer resistance to gefitinib and concurrent inhi- . miR-199a and -199a* induce caspase-dependent and caspaseindependent apoptosis, respectively. A, A549 cells were transfected with mimetics at 100 nM concentration. At 72 h post-transfection, apoptotic cell morphology was observed under a light microscope. For Annexin V staining, transfected cells were stained with Annexin V conjugated with Alexa 568 (Roche Applied Science) and Annexin V-stained cells were observed by using an epifluorescence microscope (Nikon, Tokyo, Japan). 199a, miR-199a mimetic; 199a*, miR-199a* mimetic; aϩa*, miR-199a and -199a* mimetics; NC, negative control mimetic. For Western blot analysis (B), 5.0 ϫ 10 5 A549 cells in 10-cm dishes were treated with mimetics for 72 h and then, the total protein (30 g) was subjected to Western blot analysis. a, miR-199a mimetic; a*, miR-199a* mimetic; aϩa*, miR-199a and -199a* mimetics; NC, negative control mimetic. Casp9, CASPASE 9; C-Casp9, cleaved CASPASE 9; C-Casp7, cleaved CASPASE 7; C-Casp3, cleaved CASPASE 3; Casp8, CASPASE 8; AIF, apoptosis inducing factor. F, full-length protein; C, cleaved fragment. FIGURE 5. MET is down-regulated by miR-199a*. A, A549 and HeLa cells were transfected with mimetics at 100 nM concentration. At 48 h post-transfection, total cell lysates were prepared and 30 g of total protein was subjected to Western analysis with specific antibodies for MET and ␤-ACTIN. NC, negative control; a, miR-199a; a*, miR-199a*; aϩa*, miR-199a and miR-199a* mimetics. B, a luciferase reporter plasmid (400 ng) carrying a wild-type or mutant 3Ј-UTR of MET was cotransfected with mimetics (10 nM). At 48 h after transfection, luciferase activity was measured using dual-luciferase assays (Promega). Results are representative of three individual experiments. Normalized Renilla luciferase activity in cells transfected with negative control was set at 100%. Luciferase activity of 3Ј-UTR-luciferase construct harboring wild type 3Ј-UTR of MET or mutant 3Ј-UTR of MET in which a binding site for miR-199a* is deleted are shown as filled and open bars, respectively.

miR-199a* Regulates Met Proto-oncogene
bition of both epidermal growth factor receptor and MET suppressed the growth of resistant cell lines. Similar coactivation of multiple receptor tyrosine kinases including MET was recently reported in brain tumor glioblastoma multiforme, suggesting the importance of combination therapy to treat tumors resistant to single agents (62). In this regard, it is important to note that miR-199a* down-regulated both MET receptor and the downstream signaling component ERK2. Because down-regulation of MET by siRNA did not result in the decrease in ERK2, it is likely that miR-199a* down-regulates ERK2 directly by targeting ERK2 mRNA or indirectly by targeting another protein whose reduction in turn affects the level of ERK2 protein. Because siRNA targeting MET or ERK2 alone inhibits cell growth (34 -36, 63, 64), down-regulation of both MET and ERK2 by miR-199a* may be more effective in inhibiting tumor cell proliferation and metastasis.
Although caspase-mediated apoptosis is a major cell death pathway, it is now evident that apoptosis can occur independently of caspase activation (65). Upon induction of mitochondrial membrane permeabilization, which is controlled by the BCL-2 family, several small molecules including cytochrome c, Smac/DIABLO, Omi/HrtA2, apoptosis-inducing factor, and endonuclease G are released from the mitochondrial intermembrane space. Our finding that the level of anti-apoptotic factor BCL-X L was significantly decreased in cells transfected with miR-199a suggests the induction of mitochondrial membrane permeability caused by reduced BCL-X L and subsequent release of pro-apoptotic molecules from mitochondria. In fact, anti-apoptotic factors such as BCL-2 and BCL-X L are overexpressed in a variety of human tumors and down-regulation by siRNA inhibited cell growth and induced apoptosis (66 -69). In A549 cells, down-regulation of MET alone may not be a primary cause of apoptosis because down-regulation of MET by short hairpin RNA did not inhibit proliferation in those cells in a previous study (35) and miR-199a* was more effective in inducing apoptosis than siRNAs targeting MET in our study, suggesting that targeting of multiple signaling molecules by miRNA may be more effective and more broadly applicable to a variety of tumor cells.
Strategies targeting the MET signaling pathways to develop anti-cancer drugs include: 1) antagonism of the ligand/receptor interaction; 2) inhibition of MET kinase activity; and 3) blockade of the intracellular signaling pathway. Alternatively, MET expression can be inhibited by antisense or siRNA. Our findings that miR-199a/a* is a putative tumor suppressor that concurrently down-regulates MET receptor and the downstream effector ERK2 suggest a new and potentially more effective strategy to treat primary and metastatic tumors especially when the MET signaling pathway is activated by gene amplification or mutation. In this regard, it is noteworthy that transfection with miR-199a and/or -199a* mimetics significantly inhibited tumor growth in a KB xenograft model (supplemental Fig. S7). Targeted disruption of the MET signaling pathway both upstream at the MET receptor and downstream at ERK2 by miR-199a* (and perhaps at additional unidentified targets) may prove the most effective type of treatment to prevent tumor growth and metastasis in a wide spectrum of tumor types.