Exportin-5 orthologues are functionally divergent among species

Abstract

Exportin-5, an evolutionarily conserved nuclear export factor belonging to the importin-beta family of proteins, is known to play a role in the nuclear export of small noncoding RNAs such as precursors of microRNA, viral minihelix RNA and a subset of tRNAs in mammalian cells. In this study, we show that the exportin-5 orthologues from different species such as human, fruit fly and yeast exhibit diverged functions. We found that Msn5p, a yeast exportin-5 orthologue, binds double-stranded RNAs and that it prefers a shorter 22 nt, double-stranded RNA to approximately 80 nt pre-miRNA, even though both of these RNAs share a similar terminal structure. Furthermore, we found that Drosophila exportin-5 binds pre-miRNAs and that amongst the exportin-5 orthologues tested, it shows the highest affinity for tRNAs. The knockdown of Drosophila exportin-5 in cultured cells decreased the amounts of tRNA as well as miRNA, whereas the knock down of human exportin-5 in cultured cells affected only miRNA but not tRNA levels. These results indicate that double-stranded RNA binding ability is an inherited functional characteristic of the exportin-5 orthologues and that Drosophila exportin-5 functions as an exporter of tRNAs as well as pre-miRNAs in the fruit fly that lacks the orthologous gene for exportin-t.

Figure 1

In vitro binding of the Exp5 orthologues to miRNA (A) The exportins indicated in each lane were expressed in E.coli and purified as indicated in Materials and Methods. After separation by SDS–PAGE, they were visualized by Coomassie blue staining. The positions of molecular weight markers are indicated in the left in kilo Daltons. (B) Binding of purified recombinant hsExp5 (lanes 1–5), dmExp5 (lanes 6–10) or Msn5p (lanes 11–15) to pre-miRNA was examined by EMSA. In the presence of 2 μM of hsRanQ69L-GTP (lanes 1– 5, 16, 18) or its orthologous mutants from cognate species (dmRanQ69L-GTP, lanes 6–10; yeast gsp1pQ71L-GTP, lanes 11–15), 0.07 (lanes 1, 6 and 11), 0.35 (lanes 2, 7 and 12), 0.7 (lanes 3, 8 and 13), 3.5 (lanes 4, 9 and14) and 7 (lanes 5, 10 and 15) pmol of each exportin was mixed up with radio-labeled pre-miR-30 probe. The migration of the probe was visualized by 6% non-denaturing PAGE followed by autoradiography. The positions of free and bound probe are indicated on the right. Purified Exp1 (lanes 16 and 17) and CAS (lanes 18 and 19) (7 pmol each) were also subjected to binding reactions as described above in the presence (+) (lanes 16 and 18) or absence (−) (lanes 17 and 19) of hsRanQ69L-GTP.

Figure 2

Binding specificities of the Exp5 orthologues. (A) Interaction between radio-labeled pre-miR-30 and hsExp5 (0.07 μM, lanes 1–14), dmExp5 (0.07 μM, lanes 15–28) or Msn5p (0.21 μM, lanes 29–42) in the presence (lanes 2–14, 16–28 and 30–42) or absence (lanes 1, 15 and 29) of hsRanQ69L-GTP or its orthologous mutants from cognate species was tested as in Figure 1. The binding reactions were performed in the presence of 0.5, 1, 2.5 or 5 pmol of the indicated competitors (unlabeled pre-miRNA, lanes 3–6, 17–20, 31–34; unlabeled dsDNA, lanes 7–10, 21–24, 35–38; unlabeled dsRNA, lanes 11–14, 25–28, 39–42). (B) Radio-activities co-migrating with the shifted bands were quantified using a phosphorimager. The values obtained without competitors were arbitrary set at 100 and those obtained with the indicated amounts of each competitor were calculated. Data are represented by means + SD of three independent experiments. (C) The binding of hsExp5 (lanes 1–3) or Msn5p (lanes 4–6) to radio-labeled pre-miR-30 is competed by 2.5 pmol of dsRNAs with the 5′ (lanes 2 and 5) or 3′ (lanes 3 and 6) protruding ends. The structures of dsRNAs are schematically indicated below the panel. In lanes 1 and 4, the binding reactions were performed without the unlabeled competitors. Arrows indicate the positions of free and bound probe.

Figure 3

tRNA is a potent inhibitor of the interaction between pre-miR-30 and dmExp5. (A) In vitro translated hsExp5, but not dmExp5, bound pre-miR-30. Left panel: Expression of FLAG-tagged hsExp5 (lane 2) and dmExp5 (lane 3) using an in vitro translation system was confirmed by western blot using anti-FLAG antibody. In lane 1 (mock), an in vitro translation reaction programmed with an empty vector was loaded. Right panel: In vitro translated hsExp5 or dmExp5 as indicated in the left panel was mixed with radio-labeled pre-miR-30 in the presence (+, lanes 2, 4 and 6) or absence (−, lanes 3, 5 and 7) of hsRanQ69L-GTP (lanes 2 and 6) or dmRanQ69L-GTP (lane 4). The RNA–protein complex was detected by EMSA. In lane 1, radio-labeled pre-miR-30 alone was loaded. Arrows indicate the positions of bound and free probe. (B) Bacterially expressed recombinant dmExp5 failed to bind pre-miR-30 in the presence of a reticulocyte lysate. Bacterially expressed recombinant hsExp5 (lanes 1 and 2) or dmExp5 (lanes 3 and 4) along with RanQ69L-GTP were mixed with radio-labeled pre-miR-30 in the presence (indicated by +, lanes 1 and 3) or absence (indicated by −, lanes 2 and 4) of a reticulocyte lysate. The RNA-protein complexes were detected by EMSA. Arrows indicate the positions of bound and free probe. A quantification of the shifted bands was done by using a phosphor imager. The values obtained without the competitor were arbitrary set at 1 and those obtained with the competitor were calculated. The data are indicated below each lane. (C) The binding of dmExp5 to pre-miR-30 was inhibited by RNAs extracted from reticulocyte lysate. Bacterially expressed recombinant hsExp5 (lanes 1 and 2) or dmExp5 (lanes 3 and 4) along with RanQ69L-GTP were mixed with radio-labeled pre-miR-30 in the presence (lanes 1 and 3) or absence (lanes 2 and 4) of an RNA fraction isolated from reticulocyte lysate. The RNA–protein complexes were detected by EMSA. Arrows indicate the positions of bound and free probe. A quantification of the shifted bands was done as in (B). The data are indicated below each lane. (D) Co-immunoprecipitation of tRNAs with exportins. Bacterially expressed recombinant exportins indicated in each lane were mixed with a rabbit reticulocyte lysate in the presence (+) or absence (−) of RanQ69L-GTP or its orthologous mutants from cognate species. After immunoprecipitation with anti-pentaHis antibody-bound Protein G Sepharose, co-precipitated RNAs were extracted from the immunepellets and analyzed by denaturing PAGE followed by ethidium bromide staining (lower panel). Background binding was also examined in the absence of exportins (mock). Co-precipitated proteins were also detected by western blot using an anti-pentaHis antibody (upper panel). Ten percent of input was loaded on the left most lane (total). The amounts of the precipitated exportins and tRNAs in the presence of Ran-GTP were determined by densitometric scanning of the blot and the gel. The relative amounts of co-precipitated tRNAs, which are normalized to the amounts of each exportin, are indicated below the panel.

Figure 4

dmExp5 binds tRNA more preferentially than hsExp5. Upper panel: Bacterially expressed dmExp5 (lanes 1–9) or hsExp5 (lanes 10–14) were mixed with radio-labeled pre-miR-30 and RanQ69L-GTP orthologues from cognate species in the absence (lanes 1 and 10) or presence of 5 pmol (lanes 2, 6 and 11), 10 pmol (lanes 3, 7 and 12), 25 pmol (lanes 4, 8 and 13) or 50 pmol (lanes 5, 9 and 14) of indicated competitor RNAs. Protein–RNA complexes were detected by EMSA. Arrows indicate the positions of bound and free probe. Lower panel: Binding efficiency was calculated as in Figure 2B. Data are represented by means + SD of three independent experiments. Note that the concentrations of competitor RNAs were 10 times higher than those in Figure 2.

Figure 5

Knockdown of Exp5 in human and Drosophila cultured cells. (A) Let-7a, but not tRNAs, was depleted upon the knock down of hsExp5. Left panels: Human 293F cells were treated with siRNA against DsRed (DsRedi, which was used as a negative control) or hsExp5 (hsExp5i) for three days. The expression of hsExp5 in the cells was confirmed by western blot using anti-hsExp5 antibody (upper panel). The expression of GAPDH was detected as a control (lower panel). Right panels: Total RNAs were extracted from the siRNA treated cells. Northern blot analysis was performed using specific probes indicated on the right of each panel. The signal intensities were quantified by a phosphorimager and normalized by those of U6 snRNA. The relative expression level of each RNA was indicated below each panel. (B) Both miR-2 and tRNA were depleted upon dmExp5 knockdown. Left panels: Drosophila S2 cells were treated with 74 nM of the indicated dsRNA. Total RNAs were isolated 6 days later. GFP/dsRNA was used as a negative control. Expression of dmExp5 and GAPDH was confirmed by RT–PCR (upper panel). The amount of dmExp5 mRNA but not that of GAPDH was decreased upon dmExp5 depletion (lower panel). Right panels: Northern blot analysis was performed as in (A). Both miR-2 and tRNA levels were decreased by dmExp5 knockdown. (C) The amount of tRNAs was unaffected when miRNA processing factors were depleted. Left panels: S2 cells were treated with the indicated dsRNAs for four days. The knockdown of the expression of each mRNA was confirmed by RT–PCR. Right panels: Northern blot analysis was performed as in (A). Decrease in tRNA-Ser level was observed only when dmExp5 was knocked down.

Figure 6

The repression of miRNA and tRNA levels induced by knockdown of dmExp5 was restored by overexpression of hsExp5. S2/hsExp5, a S2 cell derivative stably expressing hsExp5, was treated with 111 nM of the indicated dsRNAs as described in Figure 6B. S2/Emp, which harbors an empty vector, was used as a negative control. CuSO₄ was added to the culture media 24 h after dsRNA treatment (see Materials and Methods). (A) Depletion of dmExp5 was confirmed by RT–PCR (upper and middle panels). Expression of hsExp5 was confirmed by western blot using anti-FLAG M2 antibody (lower panel). (B) Northern blot analysis was performed as described in Figure 6A. The repression of miRNA and tRNA induced by knockdown of dmExp5 was partially restored by the expression of hsExp5.