Systematic analysis of small RNAs associated with human mitochondria by deep sequencing: detailed analysis of mitochondrial associated miRNA

Abstract

Mitochondria are one of the central regulators of many cellular processes beyond its well established role in energy metabolism. The inter-organellar crosstalk is critical for the optimal function of mitochondria. Many nuclear encoded proteins and RNA are imported to mitochondria. The translocation of small RNA (sRNA) including miRNA to mitochondria and other sub-cellular organelle is still not clear. We characterized here sRNA including miRNA associated with human mitochondria by cellular fractionation and deep sequencing approach. Mitochondria were purified from HEK293 and HeLa cells for RNA isolation. The sRNA library was generated and sequenced using Illumina system. The analysis showed the presence of unique population of sRNA associated with mitochondria including miRNA. Putative novel miRNAs were characterized from unannotated sRNA sequences. The study showed the association of 428 known, 196 putative novel miRNAs to mitochondria of HEK293 and 327 known, 13 putative novel miRNAs to mitochondria of HeLa cells. The alignment of sRNA to mitochondrial genome was also studied. The targets were analyzed using DAVID to classify them in unique networks using GO and KEGG tools. Analysis of identified targets showed that miRNA associated with mitochondria regulates critical cellular processes like RNA turnover, apoptosis, cell cycle and nucleotide metabolism. The six miRNAs (counts >1000) associated with mitochondria of both HEK293 and HeLa were validated by RT-qPCR. To our knowledge, this is the first systematic study demonstrating the associations of sRNA including miRNA with mitochondria that may regulate site-specific turnover of target mRNA important for mitochondrial related functions.

Figure 1. Analysis of isolated mitochondria.

The mitochondria were isolated and purified from HEK293 and HeLa. (A) The protein contents of whole cell lysate, and purified mitochondria were normalized, resolved on 12.5% SDS-PAGE, transferred to PVDF membrane and probed with NDUFS2 and RPS9 antibody. (B) RNA was isolated from purified mitochondria and total cell. The subsequent cDNA was used for PCR amplification of mitochondrial encoded ND4 and cytosolic/nuclear specific β-actin. M: mitochondrial fraction; C: cellular lysate. (C) RNA was isolated from mitochondria. The nuclear RNA contamination in mitochondrial RNA was assessed by checking relative enrichment of mitochondrial encoded RNA (ND4, CYB) and nuclear encoded mRNA (TRIM4, MITA) by RT-qPCR as described in method section.

Figure 2. Generation and analysis of sRNA sequences from mitochondria.

sRNA library generated from mitochondria from HEK293 and HeLa were sequenced using Illumina Hiseq 2000 platform that generated 19089819 and 17312962 clean sequence respectively. (A) Venn diagram showing distribution of common and specific sRNA total sequence reads amongst the two libraries. (B) Venn diagram showing distribution of common and specific sRNA unique sequence reads amongst the two libraries. (C) Length distribution and frequency percent of sequences in HEK293 and HeLa mitochondrial sRNA libraries.

Figure 3. Frequency of distribution of different classes of RNA associated with mitochondrial sRNA libraries.

The unique sequences obtained from the sRNA libraries were subjected to a series of sequence similarity searches using specific databases (rRNAs, tRNAs, snRNAs, snoRNAs, miRNAs, other non-coding RNAs). The sequences that did not match with any known sequence were considered as unannotated sequences. (A) An overview of sRNA associated with mitochondria of HEK293. (B) An overview of sRNA associated with the mitochondria of HeLa. The unique clean tags of repeat associated sequences were further categorized to determine the diversity of repeat associated RNA. (C) Detailed clustering of repeat-associated RNAs from mitochondria of HEK293. (D) Detailed clustering of repeat-associated RNAs from mitochondria of HeLa.

Figure 4. Analysis of differential association of miRNAs to mitochondria from HEK293 and HeLa.

(A) Scatter Plot depicting the differential association of miRNAs from libraries of HeLa and HEK293. The X and Y axis shows association level of miRNAs with mitochondria from two cell lines. Red points represent miRNA with ratio>2; Blue points represent miRNA with 1/210,000 counts) with mitochondria of HEK293. (D) The frequency of highly associated miRNAs (>10,000 counts) with mitochondria of HeLa.

Figure 5. Analysis of putative novel miRNAs associated with human mitochondria.

The putative novel miRNAs were predicted from unannotated clean reads using Mireap software. (A) The distribution of putative novel miRNAs levels with respect to frequency. Numbers of sequence reads are taken as miRNA levels and the values are represented in the form of range of values in both HEK293 and HeLa mitochondrial sRNA libraries. (B) The percentage of first nucleotide base bias of 18–23 nt putative novel miRNAs from mitochondria-associated sRNA library of HEK293. (C) The percentage of first nucleotide base bias of 18–23 nt putative novel miRNAs from mitochondria-associated sRNA library of HeLa.

Figure 6. Mapping sRNAs to human mitochondrial genome.

The sequences obtained from the mitochondrial sRNA libraries from HEK293 and HeLa were aligned to mitochondrial genome. (A) An overview of mitochondria-associated sRNAs from HEK293 that aligned to mitochondrial genome. (B) An overview of mitochondria-associated sRNAs from HeLa that aligned to mitochondrial genome. (C) The miRNAs and putative novel miRNAs that aligned to mtDNA (determined by SOAP, MapMi, BLASTN and RNAhybrid) were mapped on mtDNA using Dynamo Software tool. The locations of known miRNAs and putative novel miRNAs that aligned to mtDNA are marked in red and blue respectively. If more than 1 miRNA aligned to same position, only 1 miRNA was marked.

Figure 7. Validation of miRNAs and their targets associated with mitochondria.

The enrichment of target mRNA of mitochondria-associated miRNAs with outer mitochondrial membrane were validated. (A) The association of 5S rRNA and negative control (hsa-miR-145) to mitochondria was analyzed. The relative enrichment of 5S rRNA and hsa-miR-145 was determined by qPCR as described in methods. (B) The miRNAs (frequency >1000) which were common in both the libraries were selected for validations by RT-qPCR. RNA was prepared from purified mitochondria and cDNA synthesized. The mean CT values of miRNAs (with CT <30 and distinct melt curve) from the mitochondria of both HEK293 and HeLa are listed. (C) The mitoplast was prepared from HEK293 as described in method section. The purity of mitoplast preparation was assessed by western blotting by probing with antibody against NDUFS2 (mitochondrial inner membrane protein) and cytochrome c (inter membrane space protein). TC: total cell; MP: mitoplast; MT: mitochondria. The targets of 3 miRNAs associated with mitochondria (let-7b: STAT3; hsa-miR-107: MFN2; hsa-miR-320a: XIAP) were determined by Starbase and validated by qPCR. (D) Validation of target mRNA associated with mitoplast as compared to mitochondria from HEK293. (E) Validation of target mRNA associated with RNase A treated mitochondria from HEK293.