Nuclear processing and export of microRNAs in Arabidopsis

Abstract

In mammalian cells, the nuclear export receptor, Exportin 5 (Exp5), exports pre-microRNAs (pre-miRNAs) as well as tRNAs into the cytoplasm. In this study, we examined the function of HASTY (HST), the Arabidopsis ortholog of Exp5, in the biogenesis of miRNAs and tRNAs. In contrast to mammals, we found that miRNAs exist as single-stranded 20- to 21-nt molecules in the nucleus in Arabidopsis. This observation is consistent with previous studies indicating that proteins involved in miRNA biogenesis are located in the nucleus in Arabidopsis. Although miRNAs exist in the nucleus, a majority accumulate in the cytoplasm. Interestingly, loss-of-function mutations in HST reduced the accumulation of most miRNAs but had no effect on the accumulation of tRNAs and endogenous small interfering RNAs, or on transgene silencing. In contrast, a mutation in PAUSED (PSD), the Arabidopsis ortholog of the tRNA export receptor, Exportin-t, interfered with the processing of tRNA-Tyr but did not affect the accumulation or nuclear export of miRNAs. These results demonstrate that HST and PSD do not share RNA cargos in nuclear export and strongly suggest that there are multiple nuclear export pathways for these small RNAs in Arabidopsis.

Fig. 1.

The intracellular distribution of miR156 and mir156a^*. (A) Diagram of the 35S::miR156a construct. (B) Northern blots of total RNA isolated from nuclear and cytoplasmic fractions of 16-day-old wild-type (Col) and transgenic (TG) plants expressing miR156a under the regulation of the CaMV 35S promoter. Blots were hybridized with oligonucleotide probes complementary to the functional (miR156) or nonfunctional (miR156a^*) strands of miR156a. The RNA loaded on these blots represents ≈8% (50 μg) of the cytoplasmic yield and 30% (10 μg) of the nuclear yield. U6 small nuclear RNA and tRNA-Met were used as markers of nuclear and cytoplasmic RNA, respectively.

Fig. 2.

hst decreases the accumulation of many, but not all, miRNAs. (A) Northern blot of total RNA from Col, hst-3, hst-1, hst-6, and psd-13 rosette leaves hybridized sequentially with probes complementary to the indicated miRNAs. (B) Blots of total RNA from immature rosette leaves, fully expanded rosette leaves, and flower buds hybridized sequentially with probes complementary to the indicated miRNAs. The intensity of the hybridization signal relative to Col is indicated and was calculated after normalization to U6.

Fig. 3.

hst-1, but not psd-13, reduces the accumulation of miRNAs in both the nucleus and cytoplasm. Blots of RNA from nuclear and cytoplasmic fractions of 3-week-old rosettes (A) and flowers (B) were hybridized sequentially with probes complementary to the indicated miRNAs and the siRNA, ASRP255. Fifty micrograms of RNA was loaded was loaded in the case of cytoplasmic samples; this represented ≈10% of the cytoplasmic yield. The amount of nuclear RNA loaded on these blots was as follows: A,3 μg/15%; B Left,10 μg/20%; B Right,20 μg/40%. The intensity of the hybridization signal relative to Col is indicated and was calculated after normalization to 5S rRNA.

Fig. 4.

miRNA targets accumulate in hst-1. Northern blot of total RNA from 3-week-old rosettes hybridized sequentially with probes to the indicated mRNAs. The cognate miRNA is indicated in parentheses. Actin was used as a loading control.

Fig. 5.

The effect of hst and psd on tRNA accumulation. Blot of cytoplasmic and nuclear RNA from Col, hst-6, psd-13, and hst-6 psd-13 floral buds hybridized with probes complementary to tRNA-Tyr and tRNA-Met. The In-tRNA-Tyr probe hybridizes both to unspliced and spliced forms of this tRNA. The intensity of the hybridization signal relative to Col is indicated and was calculated after normalization to 5S rRNA. The arrow indicates the band used for quantitation.

Fig. 6.

HST is not required for the production/stability of endogenous siRNAs or for transgene silencing. (A) Blots of low-molecular-weight RNA hybridized with probes to the indicated siRNAs. (B) The expression of the L1 transgene in mature rosettes of Col, hst-6, and sgs2-1 (rdr6) plants. Col and hst-6 have similar low levels of GUS expression, whereas GUS expression is elevated in sgs2-1. (C) The expression of 35S::GUS in the absence (Left) and presence (Right) of a 35S::hairpin GUS. hst-6 does not prevent silencing of 35S::GUS.