An intronic microRNA silences genes that are functionally antagonistic to its host gene

Abstract

MicroRNAs (miRNAs) are short noncoding RNAs that down-regulate gene expression by silencing specific target mRNAs. While many miRNAs are transcribed from their own genes, nearly half map within introns of 'host' genes, the significance of which remains unclear. We report that transcriptional activation of apoptosis-associated tyrosine kinase (AATK), essential for neuronal differentiation, also generates miR-338 from an AATK gene intron that silences a family of mRNAs whose protein products are negative regulators of neuronal differentiation. We conclude that an intronic miRNA, transcribed together with the host gene mRNA, may serve the interest of its host gene by silencing a cohort of genes that are functionally antagonistic to the host gene itself.

Figure 1.

Neuronal differentiation parallels AATK mRNA and miR-338 levels. (A) Photomicrograph of M17 and SH-SY5Y cells after 4 days of RA treatment or no treatment. (B) Northern measurement of mRNAs and miRNAs and immunoblot of the proteins in two different cells after the indicated days of RA treatment. In the lowest panels, the M17 cells were transfected with AATK siRNA (Table 1) 18 h before the addition of RA.

Figure 2.

Putative miR-338 response sites and translational status of target mRNAs. (A) Multiple predicted MRE-miR-338 hybrids for the indicated target genes are depicted in the pattern shown above, i.e. for each hybrid, the top sequence is MRE and bottom is miR-338. The seed sequence of miR-338 is underlined. For each target, the number on the left is the position of this nucleotide in the NCBI entry (whose accession numbers are provided in the Results section). (B) Validity of ribosomal fractionation. The ribosome-free fraction from the top and the ribosomal/polyribosomal fraction from the bottom of the gradient were analyzed for the ribosomal protein L13a in western blot (left) and for the presence of rRNA by agarose gel electrophoresis (right). Note that similar fractions were analyzed in (C) as well. Day 0 and 4 mean no RA treatment and 4 days of RA treatment, respectively. (C) Polysomal association of target mRNAs. SH-SY5Y cell were treated as shown on left (i.e. RA treatment with or without transfection with antagomirs against miR-9, -128, -138). Before (day 0) and after 4 days (day 4) of RA treatment, the polysomal and free RNAs were quantified by qRT–PCR and representative samples analyzed on 2% agarose–EtBr gels and photographed under UV light (left panels). Immunoblot (western) of total extract of the corresponding cells shows the different protein levels (right panels). ND, not done.

Figure 3.

Induced endogenous intronic miR-338 promotes neurite growth. Standard SH-SY5Y monolayer was treated with RA and/or transfected with siRNA or antagomirs in the following combination, and at 4 days thereafter, neurite lengths were measured and then cell extracts were made to determine AATK and control profilin protein levels by immunoblot. 1, no treatment; 2, RA alone; 3, RA + anti-AATK siRNA (50 nM); 4, RA + anti-Luc siRNA (100 nM); 5, RA + antagomir-338 (50 nM); 6, RA + antagomir-338 (100 nM); 7, RA + anti-AATK siRNA (50 nM) + antagomir-338 (50 nM); 8, RA + control antagomir-9 (100 nM); 9, RA + control antagomir-128 (100 nM); and 10, RA + control antagomir-138 (100 nM).

Figure 4.

Dox-inducible coexpression of recombinant AATK and its intronic miR-338 is sufficient to promote optimal neurite growth. Note that there is no RA treatment in this panel of experiments, and therefore, the endogenous AATK gene is silent. (A) Measurement of neurite growth. SH-SY5Y cells-containing stable AATK transgene with or without intron-8 were induced with Dox where indicated for 4 days (+) or 6 days (++), or uninduced (−). Synthetic miR-338 mimic was transfected where indicated (+50 nM; ++100 nM). Control miR-139 (100 nM) was transfected in lane c. Transfection of siRNA (50 nM) against MAP1A, NOVA1 or UBE2Q1 was done in lanes M, N and U, respectively. Neurite length was measured and standard error bars from three experiments are shown. AATK and control profilin protein levels in the cell extracts were determined by immunoblot. In the miR-338 panel, intron-8-generated miR-338 was detected as described in Materials and methods section. ND, not done. (B) Representative photomicrographs of cells corresponding to data 1, 2, 4 and 6 of panel A.

Figure 5.

Overexpression of recombinant UBE2Q1 protein impedes neurite growth in SH-SY5Y cells. Recombinant UBE2Q1 transgene with (lanes C and 1–3) or without (lanes 4–8) functional MREs was induced with Dox (+) or not induced (−). Cells were treated with RA (+) (to activate endogenous AATK and miR-338 synthesis) or left untreated (−). Lane 8 expressed the C351A mutant recombinant UBE2Q1 protein. Transfection with exogenous miR-338 mimic (m) and anti-UBE2Q1 siRNA (s) were performed where indicated. In the three immunoblot (western) panels, cell lysates were probed with (from top to bottom) anti-UBE2Q1, anti-His and antiprofilin antibodies. Note that the UBE2Q1 antibody reacts with both the His-tagged recombinant and the endogenous protein, whereas the His antibody reacts with the recombinant only, confirming the identity of the two bands. To achieve a substantial separation between the His-tagged and endogenous protein bands (that differ by ∼2000 Mr), the SDS–PAGE was performed in a 30-cm tall gel apparatus (normally used for DNA sequencing runs).

Figure 6.

Model for silencing of antagonistic genes by intronic miRNA. A pathway (e.g. differentiation) is activated by the product of the gene (e.g. AATK) that also hosts the miRNA (e.g. miR-338) that silences antagonistic genes, providing optimal stimulation of the pathway. Note that many variations of the theme are possible. For example, the mechanism would also apply to examples of repression if the first gene were a repressor of the pathway; the miRNA would then silence a family of antagonistic activators of the pathway to promote optimal repression.