MicroRNA miR-128 represses LINE-1 (L1) retrotransposition by down-regulating the nuclear import factor TNPO1

Abstract

Repetitive elements, including LINE-1 (L1), comprise approximately half of the human genome. These elements can potentially destabilize the genome by initiating their own replication and reintegration into new sites (retrotransposition). In somatic cells, transcription of L1 elements is repressed by distinct molecular mechanisms, including DNA methylation and histone modifications, to repress transcription. Under conditions of hypomethylation (e.g. in tumor cells), a window of opportunity for L1 derepression arises, and additional restriction mechanisms become crucial. We recently demonstrated that the microRNA miR-128 represses L1 activity by directly binding to L1 ORF2 RNA. In this study, we tested whether miR-128 can also control L1 activity by repressing cellular proteins important for L1 retrotransposition. We found that miR-128 targets the 3' UTR of nuclear import factor transportin 1 (TNPO1) mRNA. Manipulation of miR-128 and TNPO1 levels demonstrated that induction or depletion of TNPO1 affects L1 retrotransposition and nuclear import of an L1-ribonucleoprotein complex (using L1-encoded ORF1p as a proxy for L1-ribonucleoprotein complexes). Moreover, TNPO1 overexpression partially reversed the repressive effect of miR-128 on L1 retrotransposition. Our study represents the first description of a protein factor involved in nuclear import of the L1 element and demonstrates that miR-128 controls L1 activity in somatic cells through two independent mechanisms: direct binding to L1 RNA and regulation of a cellular factor necessary for L1 nuclear import and retrotransposition.

Keywords: LINE-1 (L1); genomic instability; inhibitor; miR; miR-128; microRNA (miRNA); microRNA mechanism; mobile elements; nuclear import; nuclear transport; restriction, TNPO1, retrotransposition.

Figure 1.

Identification and verification of TNPO1 as a cellular target of miR-128. A, change in colony count of neomycin-resistant foci was used to determine the level of active retrotransposition in HeLa cells stably transduced with lentiviral constructs encoding a control miRs (control 1, 2, and 3), anti-miR-128, or miR-128 transfected with the L1 expression plasmid (wild-type L1, left panel). Colony formation assays were performed as described above using a miR-128–resistant L1 expression plasmid (Mutant) or reverse transcriptase–incompetent L1 expression plasmid (RT dead L1). Data are shown as mean ± S.E. (n = 3 independent biological replicates; *, p < 0.05; **, p < 0.01; ***, p < 0.001). B, schematic of the miR-128 qPCR screen approach. HeLa cells were transiently transfected with miR-128 or control miR mimic, cells were harvested after 72 h, RNA was isolated, and qPCR was performed for predicted miR-128 targets using GAPDH as a housekeeping gene (bottom panel). Thirteen targets were validated as down-regulated in miR-128–treated cells (supplemental Fig. 3), and relative levels of five targets, TAPT1, CASC3, SOX7, Bmi1, and TNPO1 RNA, normalized to B2M are shown as mean ± S.E. (n = 3 independent biological replicates; *, p < 0.05; **, p < 0.01) (top panel). C, relative levels of TNPO1 RNA normalized to B2M in HeLa cells stably transduced or transiently transfected with control miR, anti-miR-128, or miR-128 are shown as mean ± S.E. (n = 3 independent biological replicates; *, p < 0.05; **, p < 0.01). D, HeLa cells were stably transduced with control, anti-miR-128, or miR-128 lentiviral constructs, and Western blot analyses were performed for TNPO1 (top left panel), L1 ORF1p (bottom left panel), α-tubulin, or GAPDH protein. One representative example of three is shown. Quantification of results (n = 3) normalized to tubulin (TNPO1) or GAPDH (L1 ORF1p) are shown (right panels). E, relative levels of TNPO1 RNA normalized to B2M were determined in a teratoma cell line (Tera) stably transduced with control miR, anti-miR-128, or miR-128 and iPSCs (IMR90-1) transiently transfected with control miR, anti-miR-128, or miR-128 mimics (n = 3 independent biological replicates; *, p < 0.05; **, p < 0.01). Throughout the figure, *, p < 0.05; **, p < 0.01 by two-tailed Student's t test. Uncropped versions of blots are shown in supplemental Fig. 5.

Figure 2.

miR-128 represses TNPO1 by binding directly to the 3′ UTR of TNPO1 mRNA. A, schematic of the three predicted miR-128 binding sites in the TNPO1 mRNA (coding DNA sequence (CDS) and 3′ UTR are shown). miR-128 binding site 1 in the TNPO1 3′ UTR is a perfect 8-mer seed site; site 2 (in the 3′ UTR) and site 3 (in the coding region sequence) are both 7-mer seed binding sites. B, relative luciferase levels of HeLa cells transfected with constructs expressing a luciferase gene fused to the WT binding sequence for sites 1, 2, or 3 or the positive control sequence corresponding to the 22-nt perfect match of miR-128 along with transfections of control or miR-128 mimics were determined 48 h post-transfection. Results are shown as mean ± S.E. (n = 3 independent biological replicates. C, schematic of miR-128 binding to WT TNPO1 3′ UTR mRNA or mutant seed site TNPO1 mRNA (top panel). Relative luciferase levels of HeLa cells transfected with the reporter plasmid for WT site 1 or mutated site 1 co-transfected with control miR or miR-128 mimics were determined 48 h post-transfection. Results are shown as mean ± S.E. (n = 3 independent biological replicates; ***, p < 0.001). D, schematic of the Ago immunopurification strategy of miR-128-TNPO1 mRNA complexes (Ago-RIP) (top panel). HeLa cell lines are generated where miR-128 is either stably neutralized (by anti-miR-128) or overexpressed. Relative expression of TNPO1 mRNA normalized to B2M is shown for input samples (bottom left panel); relative fraction of TNPO1 transcript levels associated with Ago complexes is shown for IP samples (bottom right panel). TNPO1 IP fractions normalized to the levels of TNPO1 in input are shown as “corrected” (bottom right panel). Results shown as mean ± S.E. (n = 3 independent biological replicates; *, p < 0.05; ***, p < 0.001). E, relative levels of GAPDH in the same input and IP samples were determined as a negative control. Results are shown as mean ± S.E. (n = 3 independent biological replicates).

Figure 3.

TNPO1 knockdown reduces L1 activity, whereas TNPO1 overexpression enhances L1 retrotransposition. A, relative expression of TNPO1 RNA normalized to B2M in the same samples was determined (right). Results shown as a mean ± S.E. (n = 3 independent biological replicates; *, p < 0.05; ***, p < 0.001). B, de novo retrotransposition was determined by quantification of neomycin-resistant foci of HeLa cells stably transfected with plasmids encoding controls (Control), shTNPO1, FL-Control, or FL-TNPO1 co-transfected with the L1 expression plasmid (Wild-type L1). Data are shown as mean ± S.E. (n = 3 independent biological replicates; *, p < 0.05; **, p < 0.01). C, relative expression of amount of ORF2 normalized to B2M in HeLa cells stably transfected with a shControl (Control), shTNPO1, FL-Control (Control), or full-length TNPO1 overexpression (FL-TNPO1) plasmid (left). D, Western blot analysis of TNPO1 and α-tubulin (protein levels in HeLa cells stably transduced with controls, shTNPO1, or FL-TNPO1 plasmid) (left). One of three representative examples is shown. Quantification of results (n = 3) normalized to α-tubulin is shown (right). Uncropped versions of blots are shown in supplemental Fig. 5.

Figure 4.

TNPO1 knockdown reduces nuclear import of L1 (ORF1p), whereas induced expression of TNPO1 enhances nuclear import of L1 (ORF1p). A, localization of L1 ORF1p-HA was determined in HeLa cells stably expressing controls, shTNPO1, or FL-TNPO1 and then co-transfected with full-length WT L1 and ORF1p-HA or control vector. Representative orthogonal views of z-stack images are shown. Quantification of L1 ORF1p-HA localization to the nucleus is shown (represented as a percentage of L1 ORF1p in the nucleus/all L1 ORF1p in the image; arrows indicate examples of ORF1p nuclear staining in the single-channel images). Results are shown as the mean percentage of L1 ORF1p in the nucleus ± S.E. (n = 50 technical replicates of 3 independent biological replicates; ****, p < 0.0001). TAF15, a verified TNPO1 cargo, was used as a positive control (supplemental Fig. 4D). B, subcellular fractionation analysis was performed on TNPO1-modulated HeLa cells that were co-transfected with full-length WT L1 and ORF1p-HA or control vector. Western blot analysis of L1-ORF1p-HA, Lamin A/C, or α-tubulin protein levels in nuclear (N) or cytoplasmic (C) fractions of HeLa cell protein-containing lysates stably expressing controls, shTNPO1 or FL-TNPO1 (one representative of three) is shown. Quantification of results (n = 3) normalized to Lamin A/C (nuclear), or α-tubulin (cytoplasmic) is shown. **, p < 0.01. C, HeLa cells were transfected with ORF1-HA expression plasmid, HA was immunoprecipitated, and co-immunoprecipitated TNPO1 was determined by blotting for native TNPO1. One representative example of two is shown. D, localization of L1 ORF1p-HA was determined in HeLa cells stably transduced with miR control (control), anti-miR-128, or miR-128 and then transfected with the ORF1p-HA expression plasmid. Representative orthogonal views of z-stack images are shown. Quantification of L1 ORF1p-HA localization to the nucleus (represented as a percentage of L1 ORF1p in the nucleus/all L1 ORF1p in the image; arrows indicate examples of ORF1p nuclear staining) is shown as the mean percentage of L1 ORF1p in the nucleus ± S.E. (n = 50 technical replicates of 3 independent biological replicates; *, p < 0.05; **, p < 0.01; ****, p < 0.0001). Uncropped versions of the blots are shown in supplemental Fig. 5.

Figure 5.

TNPO1 partly rescues miR-128-induced repression of L1 retrotransposition and genomic integration. A, de novo retrotransposition was determined by the change in colony count of neomycin-resistant foci of HeLa cells stably transduced with control miR (control) or miR-128 transfected with FL-control, FL-TNPO1, or FL-TNPO1mut (miR-128–resistant) and co-transfected with the miR-128 mutant L1 expression plasmid. Results are shown as mean ± S.E. (n = 3 independent biological replicates; *, p < 0.05; **, p < 0.01; supplemental Fig. S6A). B, localization of L1 ORF1p-HA was determined in HeLa cells stably transduced with control miR (control) or miR-128, then transfected with FL-control, FL-TNPO1, or FL-TNPO1mut, and co-transfected with the miR-128 mutant L1 expression plasmid. Representative orthogonal views of z-stack images are shown. Quantification of L1 ORF1p-HA localization to the nucleus is shown as the mean percentage of L1 ORF1p in the nucleus (represented as a percentage of L1 ORF1p in the nucleus/all L1 ORF1p in the image; arrows indicate examples of ORF1p nuclear staining in the single-channel images) ± S.E. (n = 50 technical replicates of 3 independent biological replicates; *, p < 0.5; ***, p < 0.001; ****, p < 0.0001; supplemental Fig. S6B). C, new retrotransposition events were determined by change in colony count of neomycin-resistant foci in HeLa cells stably transduced with control miR (control) or anti-miR-128 transfected with shControl or shTNPO1 and co-transfected with the WT L1 expression plasmid. Results are shown as mean ± S.E. (n = 3 independent biological replicates; *, p < 0.05; **, p < 0.01). D, localization of L1 ORF1p-HA was determined in HeLa cells stably transduced with control miR (control) or anti-miR-128, then transfected with shControl or shTNPO1, and co-transfected with the WT L1 expression plasmid. Representative orthogonal views of z-stack images are shown. Quantification of L1 ORF1p-HA localization to the nucleus is shown as the mean percentage of L1 ORF1p in the nucleus (arrows indicate examples of ORF1p nuclear staining in the single-channel images) ± S.E. (n = 50 technical replicates of 3 independent biological replicates; *, p < 0.05; **, p < 0.01). E, relative expression of the ORF2 amount normalized to B2M in HeLa cells stably transduced with control miR (control) or anti-miR-128, transfected with shControl or shTNPO1, and co-transfected with the wild-type L1 plasmid. Results are shown as mean ± S.E. (n = 3 independent biological replicates; *, p < 0.05; **, p < 0.01).

Figure 6.

miR-128 regulates L1 retrotransposition by a dual mechanism. Shown is a schematic of miR-128–induced repression of L1 retrotransposition and genomic integration. miR-128 inhibits L1 activity by directly targeting L1 RNA as well as indirectly by repressing the levels of the cellular co-factor TNPO1, on which L1 is dependent for nuclear import and replication.