A potential link between transgene silencing and poly(A) tails

Abstract

Argonaute proteins function in gene silencing induced by double-stranded RNA (dsRNA) in various organisms. In Drosophila, the Argonaute proteins AGO1 and AGO2 have been implicated in post-transcriptional gene-silencing (PTGS)/RNA interference (RNAi). In this study, we found that AGO1 and AGO2 depletion caused the accumulation of multicopied enhanced green fluorescence protein (EGFP) transgene transcripts in Drosophila S2 cells. Depletion of AGO1, the essential factor for miRNA biogenesis, led to an increased transcriptional rate of the transgenes. In contrast, depletion of AGO2, the essential factor for siRNA-directed RNAi, resulted in EGFP mRNA stabilization with concomitant shortening of the EGFP mRNA poly(A) tail. Our findings suggest that AGO1 and AGO2 mediate multicopied transgene silencing by different mechanisms. Intriguingly, Dicer2 depletion phenocopies AGO2 depletion, with an increase in EGFP protein levels and shortening of the EGFP mRNA poly(A) tail. The possibility that AGO2 and Dicer2 involve, at least in part, poly(A) length maintenance of transgene mRNA suggests a potentially important link between transgene silencing and poly(A) tails.

FIGURE 1.

Depletion of members of the Argonaute family of proteins by RNAi causes an increase in expression of multicopied EGFP transgenes. (A) S2-EGFP cells were transfected with double-stranded RNAs (dsRNAs) corresponding to the indicated cDNAs (AGO1 and AGO2). Four days later, cells were harvested and the expression of EGFP protein analyzed. Values shown represent the expression of EGFP protein in AGO1 or AGO2 dsRNA-transfected cells relative to EGFP protein expression from cells transfected with control dsRNA. Four independent experiments were performed. Expression levels of ribosomal protein P1 (RpP1) were also checked as a control. (B) Northern blots show that depletion of AGO1 or AGO2 causes the accumulation of EGFP transcripts. Values shown represent the expression of EGFP mRNA in AGO1 or AGO2 dsRNA-transfected cells relative to EGFP mRNA expression from cells transfected with control dsRNA. Three independent experiments were performed. EGFP mRNAs are ~100 nt shorter, specifically when AGO2 is depleted. The levels of the RpP1 transcripts are also shown as a control, indicating that both amount and length of RpP1 mRNA are not altered upon depletion of AGO1 or AGO2. (C) Northern blots show that the levels of transiently expressed EGFP mRNA do not significantly change upon depletion of AGO1 or AGO2. The levels of RpP1 mRNA are shown as a control. (D) Northern blots show the nuclear and cytoplasmic distribution of EGFP mRNA. EGFP mRNA levels in the cytoplasmic fractions relative to those in the nuclear fractions are 1.52- and 1.53-fold in AGO1- and AGO2-depleted cells, respectively. Total RNAs (5 μg) prepared either from the nuclear or the cytoplasmic fraction were applied per lane. 5S rRNA is shown as an internal control. Western blots (bottom) show that lamin is only observed in the cytoplasmic fraction and cytochrome c only in the nuclear fraction, indicating that both fractions were well separated from each other. (E) AGO2 depletion leads to poly(A) shortening of the EGFP transcripts. The length of the EGFP mRNA poly(A) was compared under conditions with or without AGO2 expression. Northern blots after treatment of total RNAs with RNaseH and a specific oligo DNA hybridizing to EGFP 3′ UTR sequence show that the poly(A) tails are shorter by an ~100 nt, but only in AGO2-depleted cells (left). The length of tubulin mRNA poly(A) tail was not altered upon AGO2 depletion (right).

FIGURE 2.

AGO1 and AGO2 are involved in silencing of EGFP transgenes in two ways. (A) AGO2 functions in PTGS of the transgenes. S2-EGFP cells were first treated with dsRNAs corresponding to AGO1 or AGO2 cDNAs. Four days later, transfected cells were treated with Actin-omycinD. At the times indicated, total RNAs in each case were isolated and EGFP transcripts analyzed by Northern blot analyses. Ribosomal RNAs (rRNA) stained with ethidium bromide are shown as a loading control. A rapid decrease of the EGFP mRNA was observed in AGO1-depleted cells, as in the control. In contrast, the rate of decrease was much slower in AGO2-depleted cells, indicating that the EGFP transcripts are stabilized upon AGO2 depletion. Quantitation of the RNA bands by a PhosphorImager from three independent experiments is shown below. Tubulin mRNA levels were not altered upon AGO2 depletion in contrast to EGFP mRNA levels. (B) Nuclear run-on shows that an increase in transcription of the EGFP transgenes was observed only upon AGO1 depletion. Four days after dsRNA treatment, cells were harvested and the nuclei isolated. Transcription was then performed in nuclei in the presence of [³²P]UTP. The data shown below are the average ± standard deviation for three trials. No increase in transcription of tubulin was observed when AGO1 was silenced by RNAi. Tubulin mRNA levels are shown as an internal control. (C) Detection of EGFP antisense transcripts in S2-EGFP cells. EGFP antisense transcripts were not detected in S2-EGFP cells by Northern blotting (data not shown). RT–PCR analysis, however, reveals that EGFP antisense transcripts are indeed present in S2-EGFP cells, suggesting that EGFP dsRNAs may exist in S2-EGFP cells. RT–PCR was also performed for tubulin as an internal control. (D) Heterochromatin formation at the EGFP promoter region. Dot-blot hybridization was performed on DNA fragments isolated from the immunoprecipitated complexes with anti-HP1 antibody (right). The probe used contains the sequence corresponding to the EGFP promoter. (n.i.) nonimmune IgG shows the background level of this experiment. Western blot using anti-HP1 (left) shows that anti-HP1, but not n.i., specifically immunoprecipitates HP1 protein. (h.c. and l.c.) Protein bands corresponding to IgG heavy and light chains, respectively.

FIGURE 3.

Dicer2 depletion phenocopies AGO2 depletion. (A) Western blots show that depletion of Dicer2, but not Dicer1, causes the accumulation of EGFP protein. S2-EGFP cells were transfected with dsRNAs corresponding to the indicated cDNAs. Three days later, cells were harvested and the expression of EGFP protein analyzed by Western blotting. Values shown represent the expression of EGFP protein in Dicer1 or Dicer2 dsRNA-transfected cells relative to EGFP protein expression from cells transfected with control dsRNA. Three independent experiments were performed. (B) Dicer2 depletion phenocopies AGO2 depletion, with an increase in EGFP mRNA levels and poly(A) shortening. Of note, is that an increase in EGFP mRNA levels was observed in both Dicer1-depleted and Dicer2-depleted cells. Tubulin mRNA levels are also shown as an internal control. (C) Depletion of CAF1, a component of the mRNA deadenylation complex, together with depletion of AGO2 or Dicer2 caused the restoration of the poly(A) length of the EGFP transcripts.