ROD1 is a seedless target gene of hypoxia-induced miR-210

Abstract

Most metazoan microRNA (miRNA) target sites have perfect pairing to the "seed" sequence, a highly conserved region centering on miRNA nucleotides 2-7. Thus, complementarity to this region is a necessary requirement for target prediction algorithms. However, also non-canonical miRNA binding can confer target regulation. Here, we identified a seedless target of miR-210, a master miRNA of the hypoxic response. We analyzed 20 genes that were inversely correlated to miR-210 expression and did not display any complementarity with miR-210 seed sequence. We validated ROD1 (Regulator of Differentiation 1, also named PTBP3, Polypyrimidine Tract Binding protein 3) as a miR-210 seedless transcript enriched in miR-210-containing RNA-induced silencing complexes. ROD1 was not indirectly targeted by a miR-210-induced miRNA. Conversely, we identified a "centered" miR-210 binding site in ROD1 involving 10 consecutive bases in the central portion of miR-210. Reporter assays showed that miR-210 inhibited ROD1 by the direct binding to this sequence, demonstrating that ROD1 is a bona fide seedless target of miR-210. As expected, both ROD1 mRNA and protein were down-modulated upon hypoxia in a miR-210 dependent manner. ROD1 targeting by miR-210 was biologically significant: the rescue of ROD1 inhibition significantly increased hypoxia-induced cell death. These data highlight the importance of ROD1 regulation by miR-210 for cell homeostasis.

Figure 1. Increased or decreased miR-210 levels are associated to higher or lower ROD1 transcript levels in the RISC, respectively.

HEK-293 were co-transfected with expression vectors for either miR-210 or a scramble sequence, and c-myc-Ago2 (A, miR-210 column and B). Alternatively, HEK-293 (4×10³/cm²) were co-transfected with anti-miR-210 or anti-scramble LNA-oligonucleotides and c-myc-Ago2 (A, anti-210 column and C). Then, c-myc antibody was used to immuno-precipitate c-myc-Ago2-containing complexes. A) miRNA levels in RISCs derived from transfected cells were assayed by qPCR. As expected, miR-210 was enriched or deprived in RISCs derived from cells in which miR-210 levels were up- or down-modulated, respectively, whereas other miRNAs did not show any significant modulation. Average values are expressed using a log2 scale. Green and red colors indicate down- or up-regulation, respectively (n = 3; p<0.01). B) ROD1 mRNA levels in RISCs derived from miR-210-enriched cells were assayed by qPCR. Background controls were represented by c-myc-immuno-precipitates derived from cells transfected with miR-210, but not mAgo2 (n = 7; *p<0.001). C) ROD1 mRNA levels in RISCs derived from miR-210-deprived cells were assayed by qPCR. Background controls were represented by c-myc-immuno-precipitates derived from cells transfected with anti-miR-210 LNA-oligonucleotides, but not mAgo2 (n = 3; *p<0.001; #p<0.01).

Figure 2. miR-210 and hypoxia decrease ROD1 expression.

HEK-293 were transfected with plasmids encoding either miR-210 (miR-210) or a scramble sequence (scramble). Then, 24 and 48 hrs later, cells were collected and ROD1 protein and mRNA were assayed (A–C). Alternatively, HEK-293 were infected with lentiviruses expressing anti-scramble or anti-miR-210 sponges and, 24 hrs later, cells were exposed to hypoxia for further 72 hrs. Then cells were collected and ROD1 protein and mRNA were assayed (D-F). A) A representative western blot is shown; TUB: α-tubulin. B) The bar graph represents average ROD1 protein expression levels, measured by densitometric analysis and normalized for α-tubulin levels (n = 4; *p<0.001). C) ROD1 mRNA levels were measured by qPCR (n = 3; *p<0.001). D) A representative western blot is shown; TUB: α-tubulin. E) Expression levels of ROD1 protein were evaluated by densitometric analysis and normalized for α-tubulin protein levels (n-normoxia = 9; n-hypoxia = 5; *p<0.001). F) ROD1 mRNA levels were measured by qPCR (n = 3; *p<0.05; #p<0.005).

Figure 3. miR-210 targets ROD1 directly.

A) miR-210 (green) and ROD1 (red) complementarity. ROD1 transcript variant 6 (NM_001244898) is the longest ROD1 isoform and it was used for base numeration. B) RNA-duplexes prediction between miR-210 (green) and ROD1, EFNA3 and ISCU (red). Their calculated thermodynamic stabilities are indicated (mfe = minimum free energy). The mRNA sequences displayed are the following (gene name, RefSeq, bases, location): ROD1-isoform 6, NM_001244898, 1216–1236, cds; EFNA3, NM_004952,798–804, 3′UTR; ISCU, NM_014301, 111–118, 3′UTR. C) HEK-293 were transfected with firefly luciferase constructs that contain either an intact (pLUC-ROD1-wt) or a mutated miR-210 binding site (pLUC-ROD1-del). pLUC plasmids were co-transfected with a plasmid encoding renilla luciferase along with a plasmid encoding either miR-210 (miR-210) or a scramble sequence (scramble). Firefly luciferase values were normalized according to renilla luciferase activity (n = 4; *p<0.001).

Figure 4. The override of ROD1 down-modulation by hypoxia increased apoptosis.

HEK-293 were transfected with plasmids encoding ROD1 (pCMV-ROD1) or with vector alone (pCMV) and the next day were exposed to 1% hypoxia for the indicated time. A) HEK-293 growth curve. Data are expressed as % of T0 normoxic control. Significant differences between pCMV-ROD1 and pCMV transfected cells in the same experimental condition are indicated (n = 5–11; *p<0.001; #p<0.008). B) Cell death assessed by Trypan blue exclusion assay. Data are expressed as % of dead cells for each experimental point. Differences between hypoxic cells transfected with either pCMV-ROD1 or pCMV are statistically significant (n = 5–11; *p<0.001; #p<0.008). C) Transfected cells were exposed to hypoxia for 48 hrs and apoptosis was measured assessing the apoptotic fragmentation of cytoplasmic DNA (n = 3; *p<0.007; #p<0.03).