Angiogenin-cleaved tRNA halves interact with cytochrome c, protecting cells from apoptosis during osmotic stress

Abstract

Adaptation to changes in extracellular tonicity is essential for cell survival. However, severe or chronic hyperosmotic stress induces apoptosis, which involves cytochrome c (Cyt c) release from mitochondria and subsequent apoptosome formation. Here, we show that angiogenin-induced accumulation of tRNA halves (or tiRNAs) is accompanied by increased survival in hyperosmotically stressed mouse embryonic fibroblasts. Treatment of cells with angiogenin inhibits stress-induced formation of the apoptosome and increases the interaction of small RNAs with released Cyt c in a ribonucleoprotein (Cyt c-RNP) complex. Next-generation sequencing of RNA isolated from the Cyt c-RNP complex reveals that 20 tiRNAs are highly enriched in the Cyt c-RNP complex. Preferred components of this complex are 5' and 3' tiRNAs of specific isodecoders within a family of isoacceptors. We also demonstrate that Cyt c binds tiRNAs in vitro, and the pool of Cyt c-interacting RNAs binds tighter than individual tiRNAs. Finally, we show that angiogenin treatment of primary cortical neurons exposed to hyperosmotic stress also decreases apoptosis. Our findings reveal a connection between angiogenin-generated tiRNAs and cell survival in response to hyperosmotic stress and suggest a novel cellular complex involving Cyt c and tiRNAs that inhibits apoptosome formation and activity.

FIG 1

ANG inhibits caspase 3 activity during early hyperosmotic stress. MEFs were incubated in hyperosmotic medium alone or in the presence of ANG (0.5 μg/ml) for the indicated times. (A and B) Western blot analysis of proteins from total cell extracts for the indicated proteins. (C) Caspase 3 activity in cell extracts from three biological replicates for the indicated treatments. (D) Total RNA 5′ end labeled with [γ-³²P]ATP was analyzed on 10% urea-polyacrylamide gels. The migrations of tRNAs and tiRNAs are indicated with brackets. (E) Cell viability assays from three biological replicates for the indicated treatments using a Muse cell count and viability kit. (F) Western blot analysis of cell extracts from MEFs treated with ANG-targeted shRNA (shANG) or an shRNA control (shCtrl) for the indicated proteins and the indicated times. The significance of differences among means in panels C and E were evaluated using the Student t test (*, P < 0.05; **, P < 0.01). Error bars represent standard errors.

FIG 2

ANG inhibits apoptosis by decreasing Apaf-1 oligomerization and apoptosome activity. (A) MEFs were treated with normal medium (Control), hyperosmotic medium, or hyperosmotic medium in the presence of ANG (0.5 μg/ml) for 2 h. Total cell lysates were fractionated on a Superose 6 gel filtration column. Proteins isolated from the fractions were analyzed by Western blotting for the indicated proteins. The positions of molecular weight standards that were used for calibration of the column are marked at the bottom (arrows). (B) Caspase 9 activity in cell extracts from three biological replicates of MEFs treated with ANG for the indicated times. Error bars represent standard errors of three independent replicates (**, P < 0.01). (C) Western blot analysis of cell extracts from Apaf-1^+/+ and Apaf-1^−/− MEFs incubated in hyperosmotic medium for the indicated proteins and the indicated times. P, phospho; M.W., molecular weight (in thousands).

FIG 3

Cyt c forms complexes with ANG-induced tiRNAs during hyperosmotic stress. MEFs were incubated in hyperosmotic medium alone or in the presence of ANG (0.5 μg/ml) for the indicated times. (A) Total cell lysates were immunoprecipitated with a Cyt c antibody. The inputs, supernatants (Sup), and immunoprecipitates (Cyt c-IP) were analyzed by Western blotting for the indicated proteins. (B) RNA was isolated from the Cyt c-IPs, 5′ [γ-³²P]ATP end labeled, and analyzed as described in the legend of Fig. 1D. (C) Mitochondrial (Mito) and cytosolic (Cyto) extracts were analyzed by Western blotting for the indicated proteins. (D) RNA isolated from the mitochondrial lysates was 5′ [γ-³²P]ATP end labeled and analyzed as described in the legend of Fig. 1D. The migrations of tRNAs and tiRNAs are indicated with brackets.

FIG 4

Next-generation sequencing of RNAs immunoprecipitated with Cyt c shows enrichments for some tiRNAs. (A) Experimental scheme of the RNA-seq experiment for the detection of tiRNAs accumulated in Cyt c-IPs versus tiRNAs in total cell lysates during hyperosmotic stress in the presence of ANG. The two experimental tracks, control (Ctrl-SET, green) and Cyt c-IP (IP-SET, yellow), were performed in triplicates, and samples enriched in small RNAs were sent to deep sequencing after a PAGE gel purification. (B) Correlation plots between two replicates of deep-sequencing data mapping on tRNA genes: control and Cyt c-IP enriched. Each dot represents a tRNA gene; red dots represent mitochondrial tRNA genes. (C) Enrichment plots for selected tRNA genes. Each plot shows the deep-sequencing reads mapped along the tRNA gene for both the control (gray line) and Cyt c-IP-enriched (black line) data sets. Values on y axes are in reads per kilobase per million mapped (rpkM), allowing a direct comparison of the data. Enriched regions are shown with vertical red bars. (D) Cumulative fractions of various small tRNA species as a function of enrichment. (E) Distribution of Cyt c-IP enrichment values for tRNA genes. For clarity, only genes with an enrichment value greater than 4 are shown. Enrichment plots for tRNAs in panel E are shown in Fig. 5. aa, amino acid.

FIG 5

Enrichment plots for selected tRNA genes. Each plot shows the deep-sequencing reads mapped along the tRNA gene for both the control (gray line) and Cyt c-IP-enriched (black line) data sets. Values on y axes are in reads per kilobase per million mapped (rpkM), allowing a direct comparison of the data. Enriched regions are shown with vertical red bars and correspond to IP peaks. The enrichment of a given tRNA is the maximum enrichment value among all enriched peaks. The enrichment value at a peak is the log₂ ratio of the height of the IP curve to the height of the control curve evaluated at the position of the peak of the IP curve.

FIG 6

Cyt c binds tiRNAs in vitro. (A) Cyt c binding was measured by gel mobility shift assays. Representative gel for the tiRNA*^pool is shown. The tiRNA*^pool is the same material that was sent for next-generation sequencing in the IP-SET (Fig. 4A). The asterisk indicates that tiRNAs are not the only components of the tiRNA pool. (B) Binding isothermes for tiRNA*^pool, tiRNA^Arg, and tiRNA^Gly. Error bars represent standard errors of three independent experiments. (C) Apparent affinities (K_1/2) were calculated for the full-length tRNA^Met, the tiRNA*^pool, and the indicated tiRNA species. Error bars represent standard errors of three independent experiments. (D) Electrophoretic mobility shift assay of RNA binding reaction mixtures containing tiRNA^Arg (1 μM) and the indicated concentrations of Cyt c. (E) Electrophoretic mobility shift assay of RNA-binding reaction mixtures containing Cyt c (100 μM), tiRNA^Arg (1 μM), and the indicated concentrations of total yeast tRNA. (F) Electrophoretic mobility shift assay of RNA-binding reaction mixtures containing Cyt c (100 μM), total yeast tRNA (1 μM), and the indicated concentrations of tiRNA^Arg. Empty triangles denote free RNA of either yeast tRNA or tiRNA^Arg. Solid triangles indicate bound complex from Cyt c and RNA.

FIG 7

ANG protects primary neurons from hyperosmotic stress-induced apoptosis. Primary cortical neurons from C57BL/6J pups, cultured for 7 days and incubated for 2 h in either hyperosmotic medium alone or in the presence of ANG, were used for TUNEL staining (representative image) (A), Western blotting of total cell lysates for the indicated proteins (B), and cell viability measured by the Muse cell count instrument (C). Error bars represent standard errors of three independent replicates (*, P < 0.05). (D) Schematic representation of ANG-induced tiRNA-mediated regulation of apoptosome formation during hyperosmotic stress. It is shown that ANG (exogenous or activated by stress) generates tiRNAs by cleavage of tRNAs. Cyt c released from mitochondria forms complexes with the tiRNAs within a not fully characterized RNA-protein complex (RNP, ribonucleoprotein complex). Sequestration of Cyt c in the tiRNA-containing RNPs either limits its availability for association with Apaf-1, thus decreasing apoptosome formation, or inhibits apoptosome activity and induction of apoptosis.