MEN epsilon/beta nuclear-retained non-coding RNAs are up-regulated upon muscle differentiation and are essential components of paraspeckles

Abstract

Studies of the transcriptional output of the human and mouse genomes have revealed that there are many more transcripts produced than can be accounted for by predicted protein-coding genes. Using a custom microarray, we have identified 184 non-coding RNAs that exhibit more than twofold up- or down-regulation upon differentiation of C2C12 myoblasts into myotubes. Here, we focus on the Men epsilon/beta locus, which is up-regulated 3.3-fold during differentiation. Two non-coding RNA isoforms are produced from a single RNA polymerase II promoter, differing in the location of their 3' ends. Men epsilon is a 3.2-kb polyadenylated RNA, whereas Men beta is an approximately 20-kb transcript containing a genomically encoded poly(A)-rich tract at its 3'-end. The 3'-end of Men beta is generated by RNase P cleavage. The Men epsilon/beta transcripts are localized to nuclear paraspeckles and directly interact with NONO. Knockdown of MEN epsilon/beta expression results in the disruption of nuclear paraspeckles. Furthermore, the formation of paraspeckles, after release from transcriptional inhibition by DRB treatment, was suppressed in MEN epsilon/beta-depleted cells. Our findings indicate that the MEN epsilon/beta non-coding RNAs are essential structural/organizational components of paraspeckles.

Figure 1.

Men ε/β ncRNA localization in C2C12 nuclei. (A) RNA FISH revealed that the Men ε/β transcripts are localized in discrete foci in C2C12 myoblast nuclei. (B) In C2C12 myotubes, the foci of Men ε/β ncRNAs are enlarged and present in greater numbers than in C2C12 myoblasts. Scale bars, 10 μm. (C) Both Men ε and Men β isoforms are up-regulated upon C2C12 myoblast differentiation into myotubes as assessed by Q-PCR. Gapdh was used as a normalization control. The data in the histogram are shown as mean and standard deviation values of three independent experiments.

Figure 2.

The Men ε/β locus produces two non-coding RNAs. (A) The Men ε/β ncRNAs are transcribed from a single promoter located on mouse chromosome 19qA. Men ε is highly conserved among mammals and does not contain any repetitive elements. (Arrows) Transcription start site and polyadenylation site. (Black box) Northern and (▲) RNase protection assay (RPA) probe positions. The position of the primer pairs used in the Q-PCR (Fig. 1C) is indicated. The 3′-UTR of a nearby protein coding gene is located ∼6 kb upstream of the transcription start site. (B) Northern blot analysis using 10 μg of total RNA from 8-wk-old C57BL6 mice. (Arrow) A single band of ∼3.2 kb was detected using a probe complementary to the Men ε transcript. Relative levels of Men ε in various tissues are depicted as a histogram. Beta-actin (also known as Actb) was used as a loading control. (C) An RNase protection assay was performed using 10 μg of total RNA from the same mice. Relative levels of Men β in various tissues are depicted as a histogram. Beta-actin was used as a loading control. The full-length beta-actin probe and Men β probe are 304 nt and 221 nt, respectively. (Arrow) The protected MEN β fragment is 201 nt. (Arrowheads) Protected beta-actin fragments, 245-nt major fragment, and two smaller muscle-specific fragments.

Figure 3.

RNase P cleavage generates the 3′ end of Men β. (A) Northern blot analysis using 20 μg of total RNA showed that the human MEN β ortholog is expressed in (left) HeLa cells. The designated oligonucleotide probes were then used to roughly map (right) the 3′ end of MEN β. Beta-actin was used as a loading control. (B) RNase H digestion followed by Northern blot analysis was used to more finely map the 3′ end of MEN β. Oligo 1 is complementary to nucleotides 64,969,484–64,969,533 of human chromosome 11. Oligo 2 is complementary to nucleotides 64,969,492–64,969,541 of human chromosome 11. (C) The in vitro RNase P cleavage site, which corresponds to the 3′ end of the mature MEN β transcript, was mapped by ligation-based RNA cloning procedures. Numbers at the top indicate the position on human chromosome 11. (D) Men β is a substrate for human RNase P. (E) Recombinant His-tagged human RNase Z cleaves Men β in vitro. (F) Northern blot analysis using 25 μg of total RNA from EpH4-EV, C2C12, or mouse liver showed that the Men β tRNA-like small RNA is selectively stabilized in liver. U6 was used as a loading control.

Figure 4.

Men ε/β transcripts are localized to nuclear paraspeckles. (A) A C2C12 cell line stably expressing EYFP fused to PSPC1 (also known as PSP1α) was established. Immunoblotting showed that the endogenous PSPC1 level was reduced in (lane 2, lower band) C2C12 EYFP-PSPC1 stable cells compared to in (lane 1) wt C2C12 cells. (Lanes 3,4) Lamin B1 served as a loading control in a duplicate blot. (Lanes 5,6) After stripping the anti-Lamin B1 antibody, immunoblotting using an anti-GFP antibody confirmed that the band at ∼100 kDa corresponds to EYFP-PSPC1. (*) Residual Lamin B1 signal. (B) mCherry fused to NONO (also known as p54/nrb) was transiently expressed in C2C12 EYFP-PSPC1 stable cells. The foci of mCherry-NONO are colocalized with EYFP-PSPC1. Scale bar, 10 μm. (C) RNA FISH analysis showed that the Men ε/β transcripts are localized to paraspeckles. A probe that detected both the Men ε and Men β transcripts, as well as a probe that only detects Men β, exhibit the same localization patterns. Scale bar, 10 μm. (D) A mouse monoclonal antibody to NONO, designated 9-99, was generated. Immunoblotting analysis using C2C12 whole-cell lysate detected a single band confirming the specificity of the 9-99 monoclonal antibody. (E) A coimmunoprecipitation assay revealed that the Men ε/β transcripts directly interact with NONO. Immunoblotting using anti-PSPC1 or anti-SFPQ antibodies after co-IP showed that PSPC1 and SFPQ (also known as Psf) directly interact with NONO. Ten percent input was used for immunoblotting. cDNA was generated using random hexamers from the IP fraction. RT-PCR revealed the existence of the Men ε/β transcripts in the NONO protein complex. Both Men ε/β transcripts are ∼20-fold enriched in the same IP fraction, assessed by Q-PCR. Gapdh was used as a normalization control in Q-PCR. The data in the histogram are shown as mean and standard deviation values of three technical replicates.

Figure 5.

The MEN ε/β transcripts are essential for the integrity of nuclear paraspeckles. (A) Antisense oligonucleotides (ASO) were designed to knock down MEN ε/β or MEN β expression in HeLa cells. Three ASOs (ASO 1, 2, and 3) target both MEN ε and β isoforms, while ASO 4 targets only the MEN β transcript. (Arrows) The positions where each ASO targets the MEN ε/β transcripts. (B) Twenty-four hours after transfection of ASOs into HeLa cells stably expressing EYFP-PSPC1 (also known as PSP1α), an ∼70% knockdown of MEN ε/β (ASO 1, 2, or 3) or 50% knockdown of MEN β (ASO 4) was achieved, as assessed by Q-PCR. Beta-actin was used as a normalization control. The data in the histogram are shown as mean and standard deviation values of three independent experiments. (C) RNA FISH was performed 24 h after HeLa EYFP-PSPC1 cells were transfected with a control ASO or ASOs targeting the MEN ε/β or β transcript, to identify cells in which the RNAs were knocked down. In cells transfected with the control ASO, MEN ε/β transcripts are localized to paraspeckles. Cells transfected with ASO 3 or 4 did not show paraspeckles. (Arrows) Residual paraspeckles in a cell where knockdown of the MEN β transcript was not complete. Scale bar, 10 μm. (D) The portion of paraspeckle-positive cells was reduced to ∼20% by ASO 1, 2, or 3, while control ASO did not appear to influence the integrity of paraspeckles. Treatment with ASO 4 resulted in a loss of paraspeckles, although to a lesser extent. The data in the histogram are shown as mean and standard deviation values of three independent experiments. Approximately 100 cells were counted per experiment. (E) Knockdown of the MEN ε/β transcripts did not result in degradation of the PSPC1 protein. Lamin B1 serves as a loading control.

Figure 6.

The reformation of paraspeckles after release from transcriptional inhibition is suppressed in MEN ε/β-depleted cells. (A) Upon DRB treatment for 1 h, EYFP-PSPC1 (also known as PSP1α) relocalized to the periphery of nucleoli. The MEN β transcript lost paraspeckle localization, while the MEN ε transcript relocalized to speckles. Upon removal of DRB and recovery for 2 h, paraspeckles reformed and colocalized with the MEN ε/β transcripts. (B) Cells were treated with ASOs to knock down MEN ε/β expression prior to DRB treatment and recovery. In cells treated with a control ASO, paraspeckles re-formed within 2 h of recovery. In contrast, paraspeckles did not re-form when MEN ε/β (ASO 3) or MEN β alone (ASO 4) was depleted. (Arrows) Residual paraspeckles in a cell where knockdown of the MEN ε/β transcripts was not complete. Scale bar, 10 μm. (C) Q-PCR was used to assess the ASO knockdown efficiency after 6 h of ASO treatment. A 40%–70% knockdown of MEN ε/β (ASO 1, 2, or 3) or ∼75% knockdown of MEN β (ASO 4) was achieved. Beta-actin was used as a normalization control. The data in the histogram are shown as mean and standard deviation values of three independent experiments. (D) The percentage of paraspeckle positive cells was reduced to 20%–30% by ASO 1, 2, or 3, while the control ASO did not influence the integrity of paraspeckles. Treatment with ASO 4 also resulted in a loss of paraspeckles, although to a lesser extent. The data in the histogram are shown as mean and standard deviation values of three independent experiments. Approximately 100 cells were counted per experiment.