Structural requirements for pre-microRNA binding and nuclear export by Exportin 5

Abstract

The biogenesis and function of mature human microRNAs is dependent on the nuclear export of pre-microRNA precursors by Exportin 5 (Exp5). The precursor for the human miR-30 microRNA, which is a 63 nt long RNA hairpin bearing a 2 nt 3' overhang, forms a specific complex with Exp5 and the Ran-GTP cofactor. Here, we have examined the structural requirements for pre-microRNA binding by Exp5. Our data indicate that pre-miR-30 binding requires an RNA stem of >16 bp and is facilitated by a 3' overhang. Although a blunt-ended derivative of the pre-miR-30 stem-loop remained capable of binding Exp5, 5' overhangs were inhibitory. miR-30 variants that had lost the ability to bind Exp5 effectively were not efficiently exported from the nucleus and were also expressed at reduced levels. Furthermore, formation of a pre-microRNA/Exp5/Ran-GTP complex inhibited exonucleolytic digestion of the pre-miRNA in vitro. Together, these data demonstrate that pre-microRNA binding by Exp5 involves recognition of almost all of the RNA hairpin, with the exception of the terminal loop. Moreover, these results argue that Exp5 binding not only mediates pre-microRNA nuclear export but also prevents nuclear pre-microRNA degradation.

Figure 1

Exp5 binding to pre-miR-30 mutants with different stem lengths. Gel-shift assays were performed as described in Materials and Methods. Lane 1, ³²P-labeled WT pre-miR-30 probe in the absence of Exp5/RanQ69L-GTP; lanes 2–17, with Exp5/RanQ69L-GTP; lanes 3–17, 10 or 40 ng of the indicated unlabeled RNAs were used as competitors. Binding efficiencies (%) were calculated as the intensities of the shifted bands, quantified by a PhosphorImager, divided by that of the shifted band in lane 2 (without competitors). The sequences and predicted secondary structures of the WT and mutant pre-miR-30 RNAs are presented below the autoradiograph. Since SP6 RNA polymerase starts transcription with a ‘G’, the corresponding position in pre-miR-30 was changed from ‘U’ to ‘G’, and its base pairing nucleotide was changed from ‘A’ to ‘C’. These changes are indicated by a box. The 2 nt 3′ overhang was also changed to ‘UU’. These changes did not affect Exp5 binding (data not shown).

Figure 2

Exp5 binding to pre-miR-30 mutants with different terminal loops. Lane 1, WT pre-miR-30 probe in the absence of Exp5/RanQ69L-GTP; lanes 2–13, with Exp5/RanQ69L-GTP; lanes 3–13, 10 or 40 ng of the indicated unlabeled RNA competitors were added. Binding efficiencies (%) were calculated as the intensities of the shifted bands divided by that of the shifted band in lane 2. Sequences and predicted secondary structures of the loop deletion mutants are shown below the autoradiograph.

Figure 3

Exp5 binding to pre-miR-30 mutants with different 5′ and 3′ termini. Lane 1, ³²P-labeled WT pre-miR-30 probe without Exp5/RanQ69L-GTP; lanes 2–17, with Exp5/RanQ69L-GTP; lanes 3–17: 10 or 40 ng of the indicated unlabeled RNA competitors were added. Binding efficiencies were calculated as in Figures 1 and 2. The terminal sequences and predicted secondary structures of the competitor RNAs used are listed below the autoradiograph as well in columns 1 and 2 of Table 1.

Figure 4

Expression and function of pre-miR-30 mutants in transfected cells. (A) 293T cells were transfected with 1 μg of pH1-GFP and 1 μg of pSUPER-miR-30 (either WT or one of the mutants) per well in 6-well plates. Two days after transfection, RNAs were isolated from nuclear (N) and cytoplasmic (C) fractions and northern analyses performed. Blots were first probed for miR-30, then stripped and probed for the GFP siRNA. Lane 9: total RNA from untransfected 293T cells (NEG). Intensities of the pre-miRNA bands were quantified by a PhosphorImager, with those of the cytoplasmic and nuclear pre-miRNA expressed from pSUPER-miR-30(WT), both set as 100%. The terminal sequences and predicted secondary structures of the pre-miRNA transcripts are depicted in column 3 of Table 1. Positions of DNA markers are shown at the right side of the autoradiograph. (B) 293T cells in 24-well plates were transfected with 5 ng of the pCMV-Luc-8xmiR-30(P) indicator plasmid, 0.5 ng of the Renilla luciferase internal control plasmid pRL-CMV (Promega) and 20 ng pSUPER or pSUPER-miR-30 (WT or mutants 1 to 7). A dual luciferase assay was performed 2 days after transfection. The firefly luciferase activity (normalized against Renilla luciferase activity) observed upon pSUPER co-transfection was set at 100%. Experiments were performed in triplicate with SD indicated.

Figure 5

Inhibition of ExoT or Dicer cleavage of pre-miR-30 by Exp5 in vitro. (A) This assay measures the ability of ExoT to remove the 3′ overhang present on the pre-miR-30(19) RNA probe (see Figure 2) in the presence or absence of Exp5 and Ran-GTP. ExoT cleavage assays were performed according to Materials and Methods. For lanes 1, 5, 6 and 10, RNAs were directly mixed with 2× loading buffer (98% formamide, 10 mM EDTA, 0.1% bromophenol blue and 0.1% xylene cyanol) at the indicated time points, without going through extraction or precipitation. Lower RNA levels in lanes 2–4, 7–9 resulted from sample loss during extraction and precipitation. Shown on the left-hand side of the autoradiograph are DNA size markers. (B) Dicer was incubated with radiolabeled pre-miR-30 in the absence (lanes 3, 5 and 7) or presence (lanes 4, 6 and 8) of Exp5/RanQ69L-GTP. At the indicated time points, reactions were stopped and RNA isolated. Lane 1, pre-miR-30 without Dicer, no incubation at 37°C; lane 2, pre-miR-30 without Dicer, after incubation at 37°C. Positions of DNA markers are shown on the right-hand side of the autoradiograph. After quantification, the fraction of product formation was calculated as product intensity/(substrate intensity + product intensity) and is presented below the autoradiograph. Squares represent reactions in the absence of Exp5/Ran-GTP, while triangles represent reactions in the presence of Exp5/Ran-GTP.