MicroRNAs with a nucleolar location

Abstract

There is increasing evidence that noncoding RNAs play a functional role in the nucleus. We previously reported that the microRNA (miRNA), miR-206, is concentrated in the nucleolus of rat myoblasts, as well as in the cytoplasm as expected. Here we have extended this finding. We show by cell/nuclear fractionation followed by microarray analysis that a number of miRNAs can be detected within the nucleolus of rat myoblasts, some of which are significantly concentrated there. Pronounced nucleolar localization is a specific phenomenon since other miRNAs are present at only very low levels in the nucleolus and occur at much higher levels in the nucleoplasm and/or the cytoplasm. We have further characterized a subset of these miRNAs using RT-qPCR and in situ hybridization, and the results suggest that some miRNAs are present in the nucleolus in precursor form while others are present as mature species. Furthermore, we have found that these miRNAs are clustered in specific sites within the nucleolus that correspond to the classical granular component. One of these miRNAs is completely homologous to a portion of a snoRNA, suggesting that it may be processed from it. In contrast, the other nucleolar-concentrated miRNAs do not show homology with any annotated rat snoRNAs and thus appear to be present in the nucleolus for other reasons, such as modification/processing, or to play roles in the late stages of ribosome biosynthesis or in nonribosomal functions that have recently been ascribed to the granular component of the nucleolus.

FIGURE 1.

Purity of nucleolar fraction. (A) Normalized log signal in nucleolar, nucleoplasmic, and cytoplasmic fractions from L6 rat myoblasts to human U6 and snoRNA probes (provided as controls on Exiqon miRCURY chips). Signal (Hy3) was normalized using total RNA (Hy5), which was an equimolar mixture of nucleoplasmic, nucleolar, and cytoplasmic RNA. The sequence of human and rat U6 are identical at 104/106 nt. Human snoRNAs have varying homology with rat snoRNAs, but both homologous and nonhomologous human snoRNAs are shown to demonstrate specificity and purity of the rat cell fractions used in this study. Cross-species homology is as follows: hsa_SNORD6, 4-nt difference; hsa_SNORD4A, expression not verified in rat; hsa_SNORD3@, 3-nt difference; hsa_SNORD2, 4-nt difference; hsa_SNORD15A, 18-nt difference; hsa_SNORD14B, not identified in rat; hsa_SNORD13, 19-nt difference; and hsa_SNORD12, 9-nt difference. (B) RT-qPCR of rat Y1 RNA and rat snoRNA E2 showing relative quantities (RQ) in nucleolar, nucleoplasmic, and cytoplasmic fractions from L6 rat myoblasts.

FIGURE 2.

Microarray analysis of nucleolar, nucleoplasmic, and cytoplasmic subcellular fractions from L6 rat myoblasts. miCURY LNA miChips were used for analysis. (A) Heat map showing relative levels of five significantly nucleolus-concentrated miRNAs. NO indicates nucleolus; NU, nucleus; and CY, cytoplasm. (B) Normalized Hy3/Hy5 ratios of selected miRNAs showing relative signal detected in each fraction. Hy3 indicates Hy3-labeled RNA from single fraction; Hy5, Hy5-labeled pooled RNA used as normalization control, containing equal amounts of nucleolar, nucleoplasmic, and cytoplasmic RNA.

FIGURE 3.

miRNA LNA in situ hybridization in L6 myoblasts. Both fluorescence (top) and phase (bottom) images of typical cells are shown for each miRNA tested, with the name of the miRNA listed below the pair. Images are scaled to allow visualization of intranuclear localization patterns and should not be used to compare relative quantities between microRNAs. Exposure times were 3 sec for miR-125, miR-199, let-7a, miR-206, and miR-21; 500 msec for miR-351 and miR-664; 200 msec for miR-1 and miR-A; and 100 msec for miR-494. Bar, 3 μm.

FIGURE 4.

Detection of mature miRNAs in the nucleolus. TaqMan RT-qPCR microRNA assays were employed to characterize and quantify miRNAs identified in the nucleolus-enriched fraction of L6 myoblasts. In this method, a hairpin primer selectively hybridizes to a particular mature miRNA target and primes reverse transcription of only that mature miRNA and not its precursor or primary transcript within an RNA fraction. The RT product is then amplified using miRNA-specific PCR primers, and the number of copies is measured by hybridization of a specific fluorescent probe as real time PCR proceeds (Chen et al. 2005; Schmittgen et al. 2008). Relative copy numbers were normalized between two experiments using parallel RT reactions that were spiked with a control RNA, IPC (see Materials and Methods). The graph illustrates the levels of mature miRNAs relative to one another (in arbitrary units expressed as ΔC_t in log base 2) in nucleolus-enriched fractions. The y intercept for the x axis has been set at the value for Y1 (a prevalent cytoplasmic small RNA), which represents background. Also shown are results for snoRNA E2, a small RNA that is highly concentrated in the nucleolus. Error bars, SE between the two experiments.

FIGURE 5.

Subnucleolar mapping of miRNAs. L6 myoblasts were grown to ∼70% confluence; fixed and subjected to immunostaining for fibrillarin, a fiduciary marker for the nucleolar DFC; followed by detection of miRNAs by in situ hybridization. Optical stacks were then captured, deconvolved, and color-merged. Images show magnified nucleoli cropped from a central plane of deconvolved stacks. (A) miR-351; (E) miR-494; (I) miR-664. (B,F,J) Fibrillarin. (C,G,K) Color merge of two images in row. The white line across the nucleolus indicates the position of the densitometric line scans shown in D, H, and L, respectively. Line scans indicate the relative intensities of fibrillarin (green) and the miRNA of interest (red). Scale bars, 0.5 μm.