Competition for XPO5 binding between Dicer mRNA, pre-miRNA and viral RNA regulates human Dicer levels

Abstract

MicroRNAs (miRNAs) are a class of small, noncoding RNAs that function by regulating gene expression post-transcriptionally. Alterations in miRNA expression can strongly influence cellular physiology. Here we demonstrated cross-regulation between two components of the RNA interference (RNAi) machinery in human cells. Inhibition of exportin-5, the karyopherin responsible for pre-miRNA export, downregulated expression of Dicer, the RNase III required for pre-miRNA maturation. This effect was post-transcriptional and resulted from an increased nuclear localization of Dicer mRNA. In vitro assays and cellular RNA immunoprecipitation experiments showed that exportin-5 interacted directly with Dicer mRNA. Titration of exportin-5 by overexpression of either pre-miRNA or the adenoviral VA1 RNA resulted in loss of Dicer mRNA-exportin-5 interaction and reduction of Dicer level. This saturation also occurred during adenoviral infection and enhanced viral replication. Our study reveals an important cross-regulatory mechanism between pre-miRNA or viral small RNAs and Dicer through exportin-5.

Figure 1

Regulation of Dicer protein level by XPO5. (a) Expression of XPO5, Dicer, Drosha and DDX6 in HeLa cells transfected with control scrambled (Scr) or specific siRNA, as determined by western blotting. (b) Expression of Dicer, Myc-XPO5 and tubulin in HeLa cells transfected with either Myc-XPO5 or empty expression vectors, as determined by western blotting. (c) Dicer mRNA levels of samples described in a, as analyzed by qRT-PCR after total RNA extraction. Error bars, s.d.

Figure 2

XPO5 knockdown results in accumulation of Dicer mRNA in the nucleus. Quantification of Dicer mRNA, U6 small nuclear RNA and GAPDH mRNA by qRT-PCR in nuclear and cytoplasmic fractionated RNA from HeLa cells transfected with XPO5-specific siRNA or control scrambled siRNA. Cytoplasmic RNAs were normalized to GAPDH mRNA and nuclear RNAs to U6 snRNA. Results are expressed in arbitrary units and are representative of three independent experiments. Error bars, s.d.

Figure 3

Dicer mRNA specifically interacts with XPO5 in vivo and in vitro. (a) HeLa cell extracts were subjected to immunoprecipitation using IgG (control) or antibodies to CRM1, XPO5 or TAP–p15 in the presence of RanQ69L-GTP. Fractions of the unbound (FT) material and immunoprecipitates (IPs) were analyzed by western blotting using specific antibodies (Supplementary Fig. 2), and the rest of the IPs were used for RNA extraction. Purified RNA were reverse transcribed and subjected to qRT-PCR using specific primers for U3 snoRNA, pre-miR-16, GAPDH mRNA and Dicer mRNA. (b) Electrophoretic mobility shift assay using a radiolabeled Dicer mRNA probe (see Online Methods). Complex was formed in the presence of increasing amounts of recombinant XPO5 and RanQ69L-GTP (lanes 1–7). Complexes were subjected to treatment with RNase and analyzed on nondenaturing 5% acrylamide/TBE gels. Increasing amounts of unlabeled pre-miR-30 (10, 50, 250 ng) were used as specific competitor. Both binding assays were carried out in the presence of RanQ69L-GTP to increase specificity (Supplementary Fig. 3). (c) Radiolabeled pre-miR-30 was incubated with 200 ng of XPO5 in the presence of RanQ69L-GTP as described in Online Methods. Complexes were formed in the absence (−) or in the presence of increasing amounts (0.5, 1 and 2 μg) of the competitor RNAs as indicated. Note that this EMSA is done in the absence of RNase treatment. Complexes were analyzed on nondenaturing 5% acrylamide/TBE gels.

Figure 4

Overexpression of pre-miRNA or adenoviral VA1 RNA affects Dicer protein levels in cells. (a) HeLa cells were transfected with XPO5- or Dicer-specific siRNA, empty vector, or increasing amounts of a vector expressing pre-miR-30. Dicer and tubulin expression were analyzed by western blot (left), CRM1, XPO5 or TAP–p15 were immunoprecipitated from cell extracts and associated RNAs were extracted. Recovered RNAs were subjected to qRT-PCR to quantify U3 snoRNA, U6 snRNA, pre-miR-16 and GAPDH mRNA, as controls (Supplementary Fig. 4), as well as Dicer mRNA. (b) As in a except that cells were transfected with either empty vector (pVV2) or increasing amounts of plasmid expressing VA1 RNA (pVA1).

Figure 5

XPO5 inhibition enhances adenovirus replication. (a,b) HeLa cells were infected with wild-type (Ad5) or Ad720 mutant (dlsub720) adenovirus. Infected cells were harvested at different times after infection. A fraction of each cell extract was analyzed for viral capsid, Dicer, XPO5 and tubulin by western blotting (a) and the remainder was subjected to RNA purification. Dicer and GAPDH mRNAs (b, left), U6 snRNA and VA1 RNA (b, right) were quantified by qRT-PCR. (c,d) Top panels show knockdown efficiency of XPO5 (c) and Dicer (d) in HeLa cells. siRNA-transfected HeLa cells were infected with Ad5 or Ad720 (0.1 particle per cell). Cells were harvested every 12 h up to 48 h after infection (c) or just at 48 h after infection (d). Bottom panels show levels of adenoviral DNA as quantified by qPCR using primers amplifying the DBP viral gene present in both viruses. Results are expressed after normalization with respect to GAPDH. Error bars, s.d.