A Daphnane Diterpenoid Isolated from Wikstroemia polyantha Induces an Inflammatory Response and Modulates miRNA Activity

MicroRNAs (miRNAs) are endogenously expressed single-stranded ∼21–23 nucleotide RNAs that inhibit gene expression post-transcriptionally by binding imperfectly to elements usually within the 3′untranslated region (3′UTR) of mRNAs. Small interfering RNAs (siRNAs) mediate site-specific cleavage by binding with perfect complementarity to RNA. Here, a cell-based miRNA reporter system was developed to screen for compounds from marine and plant extracts that inhibit miRNA or siRNA activity. The daphnane diterpenoid genkwanine M (GENK) isolated from the plant Wikstroemia polyantha induces an early inflammatory response and can moderately inhibit miR-122 activity in the liver Huh-7 cell line. GENK does not alter miR-122 levels nor does it directly inhibit siRNA activity in an in vitro cleavage assay. Finally, we demonstrate that GENK can inhibit HCV infection in Huh-7 cells. In summary, the development of the cell-based miRNA sensor system should prove useful in identifying compounds that affect miRNA/siRNA activity.

Initially, most miRNAs are transcribed by RNA polymerase II as a long primary miRNA transcript (pri-miRNA) [21] which is capped and polyadenylated [22]. The pri-miRNA, which can contain clusters of miRNAs, is processed by the RNAase III enzyme Drosha [23] and the dsRNA binding protein, DGCR8 [24], into precursor miRNAs (pre-miRNAs) in the nucleus. The pre-miRNA folds into a stem loop structure and is recognized and exported to the cytoplasm by Exportin 5 [25][26][27]. The ,70 nucleotide pre-miRNA is then processed by another RNAse III enzyme, Dicer [28][29][30], and its binding partner, TRBP [31,32], into a mature 21-23 nucleotide dsRNA containing 2 nucleotide 59 overhangs. The dsRNA is delivered to the RNA-inducing silencing complex (RISC), which contains an Argonaute (Ago) protein [31,[33][34][35]. One of the strands of the dsRNA (the guide strand) remains with the RISC complex; selection is determined by the most stable 59 end of the duplex [36,37]. The miRNA-containing RISC complex binds to the target sequence within the 39UTR of the mRNA [38]. If the miRNA pairs with the target sequence with perfect complementarity, the target sequence undergoes endonucleolytic cleavage specifically by Ago2 and the mRNA is subsequently degraded [39]. By contrast, most miRNAs pair with imperfect complementarity leading to translational repression and/or degradation of the mRNA [38]. The mechanism by which miRNAs exert translational repression remains controversial [40]. Several models have been proposed including miRNA-directed inhibition of the cap-binding complex, inhibition of translation elongation, and deadenylation stimulation leading to subsequent degradation of target mRNA. Besides the main miRNA biogenesis pathway described, a subset of miRNAs mature via alternate pathways including Drosha-independent and Dicer-independent pathways and can originate from tRNA precursors and introns [41,42].
The maturation of miRNAs can be regulated at the transcriptional and post-transcriptional levels and are impacted by distinct signaling pathways. Specifically, miRNA biogenesis can be regulated at distinct steps thus altering the rate at which pri-, pre-and mature miRNAs are processed. For instance, Drosha activity can be regulated positively and negatively to affect pri-miRNA maturation. Drosha-mediated processing of the pri-miRNA let-7 can be blocked by lin-28 [43]. Furthermore, estradiol stimulation can inhibit pri-miRNA processing of a subset of miRNAs by inducing estrogen receptor-a interactions with Drosha [44]. By contrast, it has been shown that TGF-b signaling pathway can promote the processing of pri-to pre-miRNA of miR-21 through Smad association with Drosha [45]. KSRP, an RNA binding protein known for its role as a splicing factor, is also a component of Dicer and Drosha complexes and is involved in the biogenesis of a subset of miRNAs [46]. Dicer activity can be regulated by the MAP/ERK kinase pathway via phosphorylation of its binding partner TRBP [47]. Finally, RISC activity can be targeted. Growth factor treatment of cells enhance the stability of Ago2, thus effectively promoting global miRNA and siRNA activities [48].
Chemical biology approaches have provided insights into the signaling pathways that control miRNA activity. Enoxacin, which was discovered via a cell-based high-throughput screen, enhances siRNA-mediated suppression of a target mRNA, and promotes miRNA biogenesis by acting at the TRBP-mediated stage [49,50]. In other studies, a modified azobenzene compound was identified to specifically inhibit miR-21 activity by targeting pri-miRNA transcription [51] and small molecules have been found to modulate the liver-specific miR-122 levels [52]. In addition, Watashi et al. (2010) discovered two compounds, polylysine and trypaflavine, that inhibit Dicer and Argonatute activity, respectively [53]. Such studies can generate valuable tools for elucidating regulatory and signaling pathways of miRNA/siRNA function and may provide lead compounds for treatment of miRNA-associated diseases. In this study, we have used a cell-based high-throughput approach to screen for compounds from marine and plant extracts that inhibit miRNA/siRNA activity. We have identified a compound of the Daphne diterpenoid family -called genkwanine M (GENK) [54]-which affects diverse cellular functions including the inhibition of miRNA activity.

Identification of Chemical Inhibitors of miRNA Activity
To identify compounds that interfere with miRNA and siRNA regulation and/or processing, we screened natural products using a cell-based reporter system that monitors miRNA and siRNA activity. A stably transfected clonal HeLa cell line expressing a reporter RNA encoding the eGFP gene and a sequence that bears perfect complementarity to the let-7a miRNA within the 39UTR (CMV-GFP-let-7) was generated ( Figure 1A). Under basal conditions, endogenously-expressed let-7a binds to the complementary sequence leading to Ago2-mediated cleavage of the reporter RNA resulting in low GFP expression ( Figure 1B). By contrast, transfection of an antisense 29-O-methyl let-7 (AS-let-7) but not a randomized antisense RNA (AS-Ran) resulted in GFP expression, indicating that GFP expression is under the control of let-7a ( Figure 1B). Using this cell-based reporter system, a library of crude marine and plant extracts (,12,000) was screened and four crude extracts were identified that led to an increase in GFP expression ( Figure 1B). Specifically, we searched for extracts that led to a .3 fold increase GFP expression over background which is similar to that observed when cells were transfected with the ASlet-7. Guided by the reporter GFP system, sequential fractionation of one of the extracts led to the isolation of genkwanine M (GENK) (Figure 2A). Unfortunately, the other three extracts did not yield a pure compound. As a result, we focused on the characterization of GENK and its cellular effects.

Isolation of GENK
The plant extract containing GENK was prepared from the leaves and twigs of a 0.5 m tall shrub of Wikstroemia polyantha (family Thymelaeaceae), collected on the centerline of Peninsula Malaysia approximately 50 miles south of the Thai border by E. Soepadmo and M. Suhaimi under contract with the University of Chicago at Illinois. A voucher numbered Q66O4184 was deposited in the Field Museum in Chicago and at the National Herbarium at the Smithsonian. The plant material was transferred to the NCI Open Repository where it was dried and then extracted with CH 2 Cl 2 and MeOH. The crude extract was entered into the Open Repository screening plates as sample number NO44759. Three grams of the crude extract were supplied to UBC by D. Newman of the NCI Open Repository. Detailed procedures on the isolation and purification of GENK are described in the Materials and Methods and in the Supplementary Materials S1. (calcd for C 34 H 38 O 9 Na, 613.2414), requiring 16 sites of unsaturation. The diterpenoid constitution and most of the relative configuration of GENK was determined by detailed analysis of 1D and 2D NMR data (Supplemental Materials S1). However, some of the relative configurations in the region C-1 to C-7 could not be assigned with certainty from the NMR data. Therefore, GENK ( Figure 2A) was converted to its diacetyl derivative ( Figure 2B) (see Supplemental Materials S1 for NMR and MS data), which gave crystals from a 9:1 CCl 4 /hexane solution that were suitable for single crystal x-ray diffraction analysis. An ORTEP diagram that shows the constitution and absolute configuration of the diacetyl derivative is shown in Figure 2B (Supporting Materials S1).The NMR data obtained for GENK showed no evidence for the presence of acetate groups, indicating that the structure of the natural product as shown in Figure 2A, which is identical to that reported for the daphnane diterpenoid genkwanine M [54].

GENK Effects on miRNA Activity
To confirm the effects of GENK on let-7a, CMV-GFP-let-7 reporter RNAs were monitored by Northern blot analysis. Treating CMV-GFP-let-7-expressing HeLa cells with GENK at 1 and 10 mg/mL (1.7 mM and 17 mM) increased the levels of GFP RNA by approximately two fold, similar to that observed when cells were transfected with AS-let-7 but not AS-RAN ( Figure 3A). To determine whether this effect is specific to let-7 in HeLa cells, GENK was tested in Huh-7 hepatocarcinoma cells stably transfected with the CMV-GFP-miR-122 reporter. miR-122 is highly expressed in Huh-7 cells, representing ,70% of all miRNAs expressed in this cell line [55,56]. Similar to that observed in HeLa cells, GENK treatment resulted in an increase in GFP RNA with an EC 50 of 8.3 mg/mL or 14 mM ( Figure 3B). In addition, a minor product in the extracts, which was identified as the new daphnane diterpenoid genkwanine P ( Figure 2C, Supplemental Materials S1) was determined to be 70% as active as GENK ( Figure 3C) in the stably expressing CMV-GFP-miR-122 Huh-7 cell line. Treatment of HeLa and Huh-7 cells with 10 mM GENK for up to 24 or 48 hours did not significantly affect cell proliferation or viability as measured by the MTT assay (Supplementary Figure S1). Due to the consistent effects of GENK observed in the stably-transfected Huh-7 cells, we focused on this cell line for the remainder of the study.
To determine if GENK specifically inhibits siRNA activity, we tested if GENK can alter GFP mRNA levels in Huh-7 cells stably expressing a reporter RNA lacking the miR-122 binding site  (CMV-GFP) ( Fig 1A). Unexpectedly, GENK treatment resulted in an increase in GFP RNA levels in this cell line ( Figure 3E, 3F), indicating that GENK may act at the level of transcription and not through miR-122. However, closer examination of the kinetics of the increase in reporter RNA levels during GENK treatment showed that GENK has a specific impact on the reporter mRNA containing the miR-122 binding site ( Figure 3D, 3E). GENK treatment of CMV-GFP-miR-122 cells led to an increase in GFP RNA levels that peaked at 2 hours after treatment and then decreased over the next 6-8 hours ( Figure 3D). By contrast, GENK treatment of CMV-GFP expressing cells resulted in a gradual increase in reporter RNA levels that continued to increase 8-10 hours after treatment ( Figure 3E). Plotting the reporter mRNA levels (normalized to GAPDH mRNA levels) showed that GENK treatment stimulated the increase of CMV-GFP-miR-122 mRNA levels more quickly than that of the CMV-GFP mRNA levels, suggesting that miR-122 activity is temporally inhibited during GENK treatment ( Figure 3F). To explore this further, we constructed expression plasmids containing different promoters. We selected the SV40 promoter. We constructed an SV40-Renilla and an SV40-Renilla-miR-122 expression plasmid (Supplementary Figure S2A). After 6 hours of GENK treatment in Huh-7 cells, the SV40 promoter constructs were induced 1.3 fold and 1.8 fold in the absence and presence of the miR-122 site, respectively, which is a similar induction observed with the CMV-RL construct (Supplementary Figure S2B, S2C). Thus, GENK treatment induced expression from both CMV-and SV40-driven reporter constructs. To separate the promoter and miRNA effects by GENK, we constructed an expression plasmid containing the eukaryotic elongation factor 1A (eEF1A) promoter ( Figure 1A), which is predicted to direct transcription constitutively [57]. We transiently transfected either eEF1A-Renilla-miR-122 (eEF1a-RL-miR-122) or eEF1A-Renilla (eEF1a-RL) ( Figure 1A) expression plasmids into Huh-7 cells and monitored reporter RNA levels after GENK treatment. In eEF1A-RL transfected Huh-7 cells, reporter RNA levels remained constant for 8 hours of 10 mM GENK treatment, demonstrating that unlike that observed with the CMV-and SV40-promoter driven reporters, GENK had minimal effects on the transcription from the eEF1A promoter ( Figure 3G, 3H). For the eEF1a-RL-miR-122 construct, GENK reproducibly increased reporter RNA levels to a minor extent (1.2 fold) after 4 to 8 hours of treatment ( Figure 3G, 3H). Thus, GENK can both stimulate CMV and SV40-promoter-dependent transcription and inhibit miR-122 activity.

GENK does not Affect miRNA Levels
Given that GENK inhibited miRNA activity, we explored the possibility that it might affect miRNA levels. Total RNA from cells treated with GENK was isolated and loaded on a high percentage acrylamide gel to resolve miRNAs [58]. Northern blotting showed that mature miR-122 levels in Huh-7 cells ( Figure 4A) and mature let-7a levels in HeLa cells ( Figure 4B) were not affected during GENK treatment, suggesting that GENK may affect miRNA activity directly rather than the transcription or stability of miRNAs. To investigate this, we tested whether GENK can inhibit miRNA-dependent cleavage in an in vitro assay. Towards this, an in vitro transcribed reporter RNA, containing a perfectly complementary site for miR-451 or a miR-451 site in the reverse complement direction, was incubated in rabbit reticulocyte lysates (RRL) (Supplementary Figure S3A). miR-451 is highly expressed in RRL [59]. Specific cleavage of the reporter RNA (388 nucleotides) directed by miR-451 resulted in a 322 nucleotide RNA fragment (Supplementary Figure S3A). Incubation of the 59-  Figure S3B). Several conditions were tested in which GENK was preincubated with the extract prior to addition of the reporter RNA and GENK was incubated at different concentrations and for longer times, none of which altered the cleavage rate (data not shown). In summary, these data suggests that GENK treatment does not inhibit the maturation of miRNA or affect miRNA or siRNA activity as monitored by an in vitro cleavage assay.

GENK Induces an Inflammatory Response
The observation that GENK treatment stimulated CMV transcription, but not eEF1A-driven transcription suggests GENK may stimulate transcription of a subset of genes. To explore this further, we profiled the transcriptome of Huh-7 cells treated with GENK for 1 and 4 hours using microarray analysis. GENK induced the expression of 0.07% (32 mRNAs) and 0.17% (94 mRNAs) of total mRNAs by 2-fold or more at 1 and 4 hours GENK treatment respectively ( Figure 5A, 5B) (pvalue ,0.01, BH-adjusted). Based on the two time points, several trends in gene expression are observed. One, a significant subset of mRNAs (23%) are up-regulated at both 1 and 4 hours GENK treatment (Supplementary Figure S4). Second, a small number of mRNAs were significantly downregulated in GENK-treated cells (13 mRNAs at 4 hours) (Supplementary Figure S4). Third, the majority of mRNAs did not change significantly during GENK treatment (,2 fold). Supplementary Tables S1 and S2 display the top mRNAs that are up or down-regulated by more than 2 fold under GENK treatment for 1 and 4 hours, respectively. Supplementary Table S3 displays the top mRNAs that are up-regulated more than 2 fold at both 1 and 4 hours GENK treatment. The majority of genes include chemokines/cytokines, proteins involved in signal transduction pathways, and transcription factors or modulators. The genes induced after 1 hour GENK treatment closely matched the chemokines and cytokines induced by TNFa stimulation in 3T3 fibroblasts in a study published by Hao and Baltimore (2009) (Supplementary Table S1 and S2) [60]. Furthermore, classic early immediate genes described in that study such as c-Fos, c-Jun, Irf1, Cscl2 and Ier3 [60] were induced during GENK treatment. Thus, the results indicate that GENK treatment induces an early inflammatory response.
To confirm these results, we monitored Fos and Jun mRNA levels during GENK treatment. Unlike GAPDH mRNA levels, Fos and Jun ( Figure 5D, 5E) mRNA levels were transiently upregulated within the first few hours of GENK treatment before decreasing at later time points, in agreement with previous reports and the microarray analysis, and consistent with the idea that GENK treatment induces the early inflammatory response [60]. The transient induction of Fos and Jun ( Figure 5D, 5E) mRNA levels is similar to the trend observed with the induction of the reporter CMV-GFP-miR-122 mRNA levels ( Figure 3D, 3F).

TNFa Inhibits miR-122 Activity
Given that GENK treatment activated the inflammatory response and also inhibited miR-122 and let-7 miRNA, we investigated whether a well-established inducer of the inflammatory response, TNFa, has an effect on miRNA inhibition. Huh-7 cells were transiently transfected with eEF1a-RL and eEF1a-RL-122 and then treated with 10 ng/mL TNFa. Under this treatment, the reporter RNA increased to the same degree and had similar kinetics as that observed with cells treated with GENK ( Figure 6). As shown previously, TNFa treatment resulted in an increase in Fos mRNA [60]. Thus, these results suggest that GENK and TNFa activate a similar response, which includes a moderate inhibition of miRNA activity and induction of CMV-driven transcription ( Figure 6).

Discussion
MiRNAs play important roles in cellular homeostasis, development, and disease. Despite the discovery of the main factors involved in miRNA biogenesis, the regulation and signaling pathways that impact miRNA maturation and function are poorly understood. In this study, we have used a miRNA-sensitive reporter cell line to identify a small molecule from a collection of plant and marine extracts that induces the early inflammatory response and moderately inhibits miRNA function. Using reporter constructs, we demonstrated that GENK inhibits miRNAmediated repression in a time-dependent manner, thus suggesting that tight control of miRNA activity is involved in the early inflammatory response.
Gene profiling revealed that GENK treatment of Huh-7 cells induced an early inflammatory response as several signature markers such as Fos and Jun mRNA levels are increased early during GENK treatment of cells (Supplementary Table S1 and S2, Figure 5). The expression profiles of GENK-treated cells are remarkably similar to those of TNFa treated cells. It has been shown that the early inflammatory response mediated by cytokines such as TNFa is regulated by transcription factors such as NF-kB, AP-1 and interferon-regulatory factors [60]. The CMV and SV40 promoters contain binding sites for the transcription factors NF-kb and AP-1 [61,62]. Thus, the activation of the CMV and SV40driven reporter constructs by GENK treatment in our studies is likely due to the induction of an inflammatory response via NF-kb and AP-1 signaling.
Both TNFa and GENK treatments lead to a rapid increase of Fos and Jun mRNAs followed by rapid degradation (Figure 5) [60]. This early burst of expression is mediated by several mechanisms including transcriptional induction and increased mRNA stability through AU-rich elements (AREs) in the 39UTR. Early immediate genes such as Fos and Jun contain 39UTR AREs which destabilize mRNAs under basal conditions. In response to external signals such as ionizing radiation or immune signaling, ARE-mediated decay is inhibited for a short time to allow stabilization of and translation of ARE-containing mRNAs [60]. Prolonged or chronic misregulation of this posttranscriptional response can lead to inflammatory disorders and disease. For example, targeted deletion of the ARE in TNFa in mice leads to chronic inflammatory arthritis and Crohn's-like inflammatory bowel disease [63]. Interestingly, a previous study reported that the miRNA pathway may play a role in regulating the stability of  AREs and that TNFa treated cells can repress siRNA-mediated RNA interference [8,64]. In our study, we have shown that miRNA-dependent repression can be inhibited by GENK treatment, albeit to moderate levels. The level of repression was reproducible and consistently ,1.2 fold using reporter constructs under the control of the constitutive eEF1A promoter ( Figure 3G, 3H). Although the repression was moderate, there was a significant impact on the kinetics of expression of reporter constructs driven by the CMV promoter. In the miRNA-regulated reporter CMV-GFP-miR-122 construct, the reporter RNA levels peaked at 2 hours before decreasing ( Figure 3D, 3F), which is a pattern similar to that observed with other classic early response genes such as Fos and Jun during an inflammatory response ( Figure 5D) [60]. By contrast, the CMV-GFP reporter RNA lacking a miRNA binding site showed a gradual increase in expression over time ( Figure 3E, 3F). This strongly indicates that miR-122-mediated repression of the CMV-driven reporter construct is inhibited during GENK treatment to allow for the rapid increase in reporter mRNA levels. These results support the idea that the combination of miRNAmediated repression, albeit minor, and transcriptional induction during GENK treatment contribute to the tight regulation of genes during the inflammatory response, possibly those mRNAs containing AREs.
GENK does not alter mature let-7a or miR-122 levels, suggesting that the inhibition of miRNA activity is likely at the level of RISC activity (Figure 4). GENK does not affect siRNAmediated cleavage in an in vitro cleavage assay. How GENK inhibits miRNA/siRNA activities in vivo remains to be investigated. Given the pleiotropic effects of GENK, it seems likely that inhibition of miRNA activity is a result of the induction of the inflammatory response.
GENK inhibits HCV virus infection in Huh-7 cells (Figure 7), which is likely due to activation of an inflammatory response and to a minor extent inhibition of miR-122 activity by GENK. Previously, it has been reported that HCV replication is resistant to TNFa [65]. Thus, despite similarities of the effects tested in this study by GENK and TNFa, GENK may be inhibiting HCV virus infection specifically and distinct of the effects observed by TNFa.
Chemical biology approaches provide an alternative to genetic screens in identifying signaling pathways that affect major biological processes. The use of reporter constructs that sense miRNA or siRNA activity has been extremely useful for identifying chemical inhibitors of this pathway [49][50][51][52]. Although the CMV and SV40 promoters are strong transcriptional promoters and have been used widely for expression studies, these promoters are highly inducible especially during signaling events that activate an inflammatory response [57,61,62]. The use of a constitutive promoter such as the eIF1A promoter used in this study should be useful for future screenings of chemical inhibitors of miRNA function.

Plasmids
CMV-GFP-let-7 and CMV-eGFP-miR-122 were constructed by inserting a perfectly complementary let-7 or miR-122 binding site within an MfeI restriction enzyme site between the eGFP and SV40 poly A signal in pEGFP-N1 vector (Clontech). eEF1A-RL and eEF1A-RL-miR-122 were constructed by swapping the CMV promoter of CMV-GFP-let-7 or CMV-GFP-miR-122 with the eEF1A promoter from pTy-eGFP (kindly provided by Ivan Sadowski, University of British Columbia) using the restriction site BglII and HindIII. eEF1a-RL-122 was constructed by PCR with primers containing an Nhe1 site with the initial ORF of RL and a primer containing the end of the reverse complement ORF of RL, miR-122 binding site and XbaI site. This fragment was subsequently swapped into the eEF1a-RL, replacing the RL with NheI and XbaI restriction sites. SV40-RL (provided by Martin Bushell, University of Notingham) was used as a template to construct SV40-RL-miR122 and CMV-RL. Specifically, SV40-RL-miR122 was constructed by introducing miR122 into the Xba1 site. And CMV-RL was made by swapping in SV40 promoter with full length CMV promoter (from pCDNA3, Invitrogen) using BglII and HindIII restriction sites.

Screen for miRNA Inhibitors
HeLa cells stably expressing EGFP-let-7 by CMV promoter were seeded in PerkinElmer View 96-well plates at 8,000 cells per well in 100 mL DMEM at 37uC. 24 hours after seeding, 340 nL of marine and plant extracts were added to each well using a Biorobotics Biogrid II equipped with a 0.7 mm diameter 96-pin tool. Plates were incubated for 8 hours at 37uC, the medium was aspirated and cells were fixed with 3% paraformaldye in PBS for 15 min at room temperature. After two washes with PBS, the cells were incubated with 500 ng/mL Hoechst 33342 for 15 min at room temperature. Plates were read in a Cellomics TM Arrayscan VTI automated fluorescence imager. Cells were photographed with a 106 objective in the Hoechst and FITC (XF-93 filter) channels. 10 images per well were collected. The target activation analysis algorithm was used to identify cells via Hoechst stain of the nuclei (ObjectAreaCh1 = 5) in the first channel and apply a cytoplasmic mask in the second FITC channel around the nucleus. This allows us to determine the average cytoplasmic GFP intensity of each cell. The average GFP intensity of all the cells in each well was computed at fixed 200 pixel intensity units. Guided by the ASlet-7 transfection experiments, we determined that a 3 fold increase in GFP expression in cells over that observed in untreated cells is the minimal threshold to identify a positive candidate extract. Wells that showed less than 4,000 cells were discarded as toxic extracts. 12,000 plant and marine extracts from the NCI open repository were screened using this protocol.

Isolation of GENK
To Isolate GENK, three grams of the crude extract were partitioned between EtOAc (3650 mL) and H 2 O (100 mL). The combined EtOAc extract was evaporated to dryness to give 2.1 g of dark green oil; 1.00 g of this was chromatographed on a 10 g normal phase Waters Sep-PakPak, employing a step gradient from 95:5 n-Hexanes/EtOAc to EtOAc, and from 90:10 EtOAc/ MeOH to MeOH, to give fractions A-E. Fraction C (38 mg) eluting with 60:40 n-Hexanes/EtOAc, was chromatographed on a SephadexH LH-20 column using 1:1 MeOH/CH 2 Cl 2 as an eluent to give fractions C-(A-G). Pure GENK (Figure 2A) (4.9 mg) and a minor product, genkwanine P ( Figure 2C) (2.1 mg) were obtained aswhite amorphous solids from fraction C-E (15 mg) via C-18 reversed-phase HPLC using 8:2 MeOH/H 2 O as an eluent over 70 min. (flow rate 2 mL/min). A more detailed protocol can be found in the Supplementary Materials S1.

Northern Blot Analysis
Total RNA from cells was purified using TRIzol (Invitrogen). 5 mg of each sample was loaded onto a 1% denaturing agarose gel (16 MOPS and 20% formaldehyde v/v). RNA was transferred and crosslinked to Zeta-probe membrane (Bio-Rad, UV Stratalinker 1800 TM by Strategene). Transfer of RNA was confirmed by methylene-blue staining. The membrane was pre-hybridized with 5 mL hybridization buffer (0.5 M NaPO 4 , 10 mM EDTA, 7% SDS w/v) at 65uC for at least 30 min. [a-32 P] dATP-labeled DNA probes were generated using the Radprimer kit (Invitrogen) and subsequently purified using the QIAquick Nucleotide Removal Kit (Qiagen). For miRNA detection, 20 mg of RNA was loaded onto a 15% denaturing acrylamide gel and subsequently transferred to a nitrocellulose membrane by a Trans-Blot SD Semi-Dry Transfer Cell TM (BioRad) for 1 hr at 12 V. After transfer, RNAs were UVcrosslinked to the membrane (UV Stratalinker 1800 TM by Strategene). 59 end-labeled LNA antisense miR-122 or let-7 oligos (Exiqon) were used for Northern blot analysis. Probes were heated at 95uC before adding to the pre-hybridized membranes. After incubating overnight at 65uC, the membranes were washed three times with 15 mL of Wash Buffer (0.1% SDS w/v, 0.16 SSC) at 50uC. The membranes were analyzed by phosphorimager analysis (Typhoon, GE Healthcare).

In vitro siRNA Cleavage Assay in Rabbit Reticulocyte Lysate
The miR-451 sensor was constructed by PCR amplification using primers that amplified the RL gene with the miR-451 sequence in the sense or antisense direction. The PCR product was subsequently cloned into the pCRII-TOPO vector (Invitrogen). The reporter miR-451 RNA was in vitro transcribed by a T7 RNA polymerase reaction. The reporter RNA was dephosphorylated and then 59 end-labeled with [c-32 P]-ATP by T4 polynucleotide kinase (Fermentas). The RNAs were purified using an RNeasy Kit (Qiagen). Radiolabel RNAs (500,000 cpm) were incubated in Rabbit Reticulocyte Lysates (Promega) pretreated with or without 10 mM GENK and aliquots of the reaction were taken at different times. RNA was extracted from the samples by phenol/chloroform extraction and ethanol precipitation. Equal cpm of RNA were loaded on a 12% acrylamide gel, dried, and analyzed by phosphorimager analysis.

Microarray Analysis
Total RNA extracted from Huh-7 cells pretreated with DMSO or 10 mM GENK for 1 and 4 hours and performed in triplicate. Illumina Direct Hybridization assays were performed by the Vancouver Prostate Centre Laboratory for Advanced Genome Analysis, Vancouver, Canada. Total RNA quality was assessed with the Agilent 2100 bioanalyzer prior to microarray analysis. Samples with a RIN value of greater than or equal to 8.0 were deemed to be acceptable for microarray analysis. An input of 200 ng of total RNA was used to generate biotinlabeled cRNA following the IlluminaH TotalPrep RNA Amplification Kit (Ambion, Inc.). Samples were hybridized on Illumina HumanHT-12 v3 BeadChips following the Illumina Whole-Genome Gene Expression Direct Hybridization Assay Guide (11286331). BeadChips were imaged and quantified with the Illumina iScan scanner and data was processed with Illumina BeadStudio v3.3.8. Data processing included averaging signal intensities for each unique Beadtype. The microarray data has been deposited to GEO with accession number: (will be submitted).
Differential expression analysis was carried out between samples treated with drug for 1 hr and 4 hours relative to untreated control. log 2 normalized expression values were used for statistical analysis using the limma package in R. The p-values were adjusted for multiple correction using the Benjamini-Hochberg algorithm. A stringent criteria consisting of fold change of 2 (up/downregulation) and an adjusted p-alue of 0.01 was used to determine significantly deregulated mRNAs. Clustering of differentially expressed genes from both treatment groups was carried out using complete linkage hierarchical clustering with Euclidean distances with the gplots package in R. Pathway analysis was performed using the Ingenuity Pathway database. A p-value of 0.05 (-log(p) of 1.30) was used to determine significantly enriched pathways.

Detection of HCV-infected Cells
A cDNA copy of HCV JFH-1 (a gift from Dr. Takaji Wakita, National Institute of Infectious Diseases, Tokyo, Japan) was in vitro transcribed to genomic RNA which was then used to produce infectious virus stocks as previously described [67,68]. Filtered HCV JFH-1 stocks were used to infect Huh-7 cells pretreated with 10 mM GENK for 24 hours. At 2 days post-infection, cells were fixed with 4% v/v formaldehyde in PBS, washed with PBS and then permeabilized and blocked with PBS containing 3% BSA, 0.3% Triton X and 10% FBS (Blocking buffer) for 1 hr. Cells were then probed with mouse anti-core C7-50 primary antibody (1:500, Abcam) in Binding Buffer (Blocking buffer without FBS) at 4uC overnight then washed 3 times with PBS. Cells were probed with diluted donkey anti-mouse Alexa-Fluor-568 conjugated secondary antibody (1:1000) and 10 mg/ml Hoecsht 33258 in binding Buffer for 1 hr at room temperature. After three washes with PBS, cells were left in 50 ml PBS and the percentage of HCV infected cells was analyzed using the Cellomics TM Arrayscan VTI automated fluorescence imager. Figure S1 MTT assay of HeLa and Huh-7 cells treated with 10 mM GENK for the indicated times. Viability of HeLa and Huh-7 cells (100 mL volume) was monitored by the MTT colorimetric assay (the addition of final 1.25 mg/mL MTT for 2 hours). Subsequently, cells were lysed with extraction buffer (20% w/v SDS, 50% DMF, 2.5% acetic acid, and 2.5% 1M HCL) overnight. Reduction of MTT was detected at 570 nm. The miR-451 sensor RNA was in vitro transcribed as shown and 59end labeled with P 32 using a T4 kinase assay. The sensor RNA was incubated in rabbit reticulocyte lysate (RRL) (Promega) pretreated with DMSO or 10 mM GENK for the indicated times. Endogenous miR-451 in RRL binds to the miR-451 site leading to cleavage of the in vitro transcribed RNA to generate two fragments, 322 and 66 nucleotides in length. Only the 322 nucleotide fragment will be detected by a phosphorimager. The RNAs were extracted and separated on a 12% acrylamide gel, which was dried and detected by a phosphorimager analysis. (EPS) Figure S4 Clustering of differentially expressed mRNAs from 1 and 4 hours treatment groups. Heatmap showing mRNAs that are differentially expressed at 1 and 4 hours GENK post treatment. Each mRNA needs to be significantly deregulated in at least one treatment group to be included. Colors indicate absolute fold changes according to the specified scale. Red denotes upregulation relative to control while blue indicates down-regulation. Clustering was carried out using the complete linkage clustering algorithm. Whereas a small number of mRNAs are downregulated at 4 hours after treatment, the majority are up-regulated.