As shown in Fig. 2C, C/EBPα activated the reporter gene through the E2 fragment. Next, we performed mutational analyses on predicted HNF binding sites. Because the multiple putative C/EBPα target sites were arranged in a tandem array, we did not perform mutation analysis on these sites. As shown in Fig. 2D, mutagenesis of certain conserved PD 332991 sites (F4A-3, F3B-1, and F1A-3) abolished the effects of HNFs on the reporter, but mutations of all nonconserved sites made no difference. Remarkably, the F4A-3 site was a crucial site, because its mutation completely abolished the miR-122 promoter function. The F3B-1 and F1A-3 sites overlapped (Table 1), and mutations in these sites eliminated the effects

of both HNF1α and HNF3β. These data

demonstrate that the HNFs could directly bind to the miR-122 promoter. This conclusion was further confirmed by way of chromatin immunoprecipitation assay. As shown in Fig. 2E, the three HNFs directly bind to the miR-122 promoter in Huh7 cells. To test whether the LETFs could up-regulate miR-122 expression in HCC cells, we performed overexpression studies. As shown in Fig. 2F, cells transfected with LETF-expressing vectors display an obvious up-regulation of miR-122, I-BET-762 cost especially for C/EBPα. Moreover, this finding is consistently observed in the three cell lines used. Together, these results show that C/EBPα, HNF1α, HNF3β, and HNF4α are involved in the transcriptional regulation of miR-122, which also suggests that miR-122 functions as an effector of these LETFs during liver development. Cellular proliferation and differentiation are the two most important processes for organ development.23 Numerous studies have established the pivotal roles of LETFs in the regulation of both processes during liver development.17-19 To search for the functional targets of miR-122, we primarily focused on candidate target genes with the potential to suppress differentiation and/or promote proliferation, which are contrary to the roles of LETFs. Eleven

candidate targets were arbitrarily selected from the results predicted by Targetscan 4.2 for further confirmation (Table 2). In addition, CCNG1 and BCL2L2, two known targets of miR-122, were 上海皓元 employed as positive controls.16, 24 The 3′-UTR segments of each target were synthesized and subcloned downstream of the Renilla luciferase in the psiCHECK-2 dual luciferase reporter vector (Fig. 3A), and reporter assays were performed as indicated. Surprisingly, as shown in Fig. 3B, 11 reporters were significantly repressed by miR-122 to different degrees (30%-70% reduction), including the two known targets. MSN (moesin) and serum response factor (SRF) were not significant in this group. These data indicate that most candidate genes could be directly repressed by miR-122. To further confirm this hypothesis, we performed mutational analyses on each predicted site.