mRNA isoform diversity can obscure detection of miRNA-mediated control of translation

Abstract

Reporter-based studies support inhibition of translation at the level of initiation as a substantial component of the miRNA mechanism, yet recent global analyses have suggested that they predominantly act through decreasing target mRNA stability. Cells commonly coexpress several processing isoforms of an mRNA, which may also differ in their regulatory untranslated regions (UTR). In particular, cancer cells are known to express high levels of short 3' UTR isoforms that evade miRNA-mediated regulation, whereas longer 3' UTRs predominate in nontransformed cells. To test whether mRNA isoform diversity can obscure detection of miRNA-mediated control at the level of translation, we assayed the responses of 11 endogenous let-7 targets to inactivation of this miRNA in HeLa cells, an intensively studied model system. We show that translational regulation in many cases appears to be modest when measuring the composite polysome profile of all extant isoforms of a given mRNA by density ultracentrifugation. In contrast, we saw clear effects at the level of translation initiation for multiple examples when selectively profiling mRNA isoforms carrying the 5' or 3' untranslated regions that were actually permissive to let-7 action, or when let-7 and a second targeting miRNA were jointly manipulated. Altogether, these results highlight a caveat to the mechanistic interpretation of data from global miRNA target analyses in transformed cells. Importantly, they reaffirm the importance of translational control as part of the miRNA mechanism in animal cells.

FIGURE 1.

Manipulation of let-7 activity by anti-miR transfection. HeLa cell lysates were prepared 9 h after transfection with R-luc-3xB reporter and anti-miR against let-7 or control anti-miR, followed by density gradient ultracentrifugation and RNA analysis. (A) Schematic of the Renilla luciferase R-luc-3xB reporter. (B) Representative absorbance trace at 254 nm is shown at the top to locate positions of polysomes, ribosomes (80S), and subunits (60S, 40S). Gel analysis of gradient fraction RNA integrity is shown below. (C) Gradient distribution of R-luc-3xB mRNA after anti-let-7 (gray squares) or control anti-miR transfection (black circles) was measured by qRT–PCR. R-luc-3xB mRNA abundance in each fraction is expressed as a percentage of the total R-luc-3xB mRNA in the gradient (see Supplemental Fig. S1 for normalization strategy). Dotted line represents border between polysomal and subpolysomal complexes. Data are an average of three experiments, and bars represent standard error.

FIGURE 2.

mRNA isoform diversity complicates measurement of translational control by let-7. HeLa cells were transfected with anti-miRs, harvested 6–17 h later, and lysates fractionated by gradient centrifugation as in Figure 1. (A) 3′ UTR configurations of the 11 experimentally verified let-7 target mRNAs under study. Alternative polyadenylation sites (blue triangles) were identified from Refseq, EST, and the PACdb (Brockman et al. 2005) databases (see Supplemental Materials and Methods). Let-7 target sites (purple lines) were as predicted by Targetscan (Friedman et al. 2009) and other sources (MYC, Kong et al. 2008; GPR56, Selbach et al. 2008; HMGA2, Lee and Dutta 2007; IGF2BP1, Mayr and Bartel 2009). Also shown are binding sites for miRNAs expressed at the level of ≥10% that are seen for let-7 in HeLa cells (determined by small RNA-Seq; yellow lines). Binding sites for these miRNA were derived from published studies and Targetscan predictions (details in Supplemental Table S2). (B) Gradient distribution of let-7 target mRNAs after transfection anti-let-7 (gray squares) or control anti-miR (black circles). All qRT–PCR measurements used primers designed against coding regions. Profiles are the averages of four to nine independent experiments (bars represent standard error). An exception is KRAS, which is an average of three independent experiments (6–9 h after anti-miR transfection).

FIGURE 3.

Combined inhibition of let-7 and miR-34a reveals greater translational regulation of MYC mRNA. (A) Schematic of binding sites for miR-34a and let-7 in the MYC 3′ UTR (Kong et al. 2008; Kim et al. 2009a). Red sequence denotes miRNA seed regions, green denotes UTR. (B) Gradient distribution of MYC mRNA by coding-region-focused qRT–PCR, 17 h after transfection with anti-miRs against let-7 (red squares), miR-34a (orange triangles), both (purple diamonds), or a control (blue circles). Data shown are representative of two independent experiments.

FIGURE 4.

Analysis of long 3′ UTR isoforms reveals stronger regulation of translation initiation by let-7. (A) Schematic of IGF2BP1, TMEM2, and HMGA2 3′ UTRs showing let-7 binding sites (lines), alternative polyadenylation sites (triangles), and long 3′ UTR isoforms PCR amplicons (open black box). (B) Gradient distribution of let-7 target mRNAs IGF2BP1, TMEM2, and HMGA2 after transfection anti-let-7 (gray squares) or control anti-miR (black circles). Profiles are averages of four to nine independent experiments (bars represent standard error) and were generated either with coding regions primers as in Figure 2 (left) or with primers selective to the longest 3′ UTR isoform of each mRNA (right). (C) Stabilization of long 3′ UTR isoforms by let-7 17 h after anti-miR transfection. This panel shows the fold stabilization of HMGA2, IGF2BP1, and TMEM2 measured using coding-region primers or primers in the 3′ UTR. Bars represent standard error of three to five independent experiments. (D) Poly(A) tail lengths of two 3′ UTR isoforms of IGF2BP1 mRNA (cleavage sites at 2416 and 8769, see Fig. 2A) were measured by LM–PAT assay 6, 9, and 12 h after anti-miR transfection. GAPDH mRNA is shown as an invariant control; TVN is a marker for ∼12 nt of poly(A) tail.

FIGURE 5.

Analysis of GPR56 5′ UTR isoforms reveals regulation of translation initiation by let-7. (A) Schematic of alternate 5′ UTRs of GPR56 mRNA with positions of putative upstream open reading frames (uORFs) boxed in light gray and isoform-specific amplicons in open gray boxes. (B) Gradient distribution of GPR56 mRNA after transfection with anti-let-7 (gray squares) or control anti-miR (black circles). Profiles are averages of four independent experiments (bars represent standard error), either with coding regions primers as in Figure 2 (left) or primers specific to GPR56 5′ UTR isoform 1 (middle) and isoform 2 (right), respectively.