Disruption of tRNA biogenesis enhances proteostatic resilience, improves later-life health, and promotes longevity

Abstract

tRNAs are evolutionarily ancient molecular decoders essential for protein translation. In eukaryotes, tRNAs and other short, noncoding RNAs are transcribed by RNA polymerase (Pol) III, an enzyme that promotes ageing in yeast, worms, and flies. Here, we show that a partial reduction in Pol III activity specifically disrupts tRNA levels. This effect is conserved across worms, flies, and mice, where computational models indicate that it impacts mRNA decoding. In all 3 species, reduced Pol III activity increases proteostatic resilience. In worms, it activates the unfolded protein response (UPR) and direct disruption of tRNA metabolism is sufficient to recapitulate this. In flies, decreasing Pol III's transcriptional initiation on tRNA genes by a loss-of-function in the TFIIIC transcription factor robustly extends lifespan, improves proteostatic resilience and recapitulates the broad-spectrum benefits to late-life health seen following partial Pol III inhibition. We provide evidence that a partial reduction in Pol III activity impacts translation, quantitatively or qualitatively, in both worms and flies, indicating a potential mode of action. Our work demonstrates a conserved and previously unappreciated role of tRNAs in animal ageing.

Fig 1. Partial inhibition of Pol III in worms alters tRNA expression specifically.

(A) Volcano plots of differential expression of short RNAs (<300 b) between worms treated with rpc-1 RNAi and vector controls, with different classes of transcripts coloured in different panels. “Other” refers to snoRNAs identified in a previous study (see main text). Horizontal line indicates the p value threshold for 10% FDR. (B) Northern blots of RNA from worms treated with the vector control or rpc-1 RNAi. Actin band from 1 wild-type sample was accidentally lost from the blot. Data underlying this figure can be found in S1 Data. Uncropped images of blots are presented in S1 Fig.

Fig 2. Pol III inhibition and interference in tRNA metabolism both trigger UPR in worms.

(A) Tunicamycin and (B) heat shock survival of worms treated with rpc-1 RNAi or the vector control. Demography and statistical analyses of data presented in A, B and additional repeats are provided in S1 Table. (C) Expression of the hsp-4-4::GFP reporter in rpc-1 RNAi- or vector control-treated worms (n = 11 rpc-1 RNAi/12 control; effect of RNAi p < 10⁻⁴, effect of age p < 10⁻⁴, RNAi by age interaction p < 10⁻⁴, LM). Fluorescence intensities are relative to control on day 2. (D) Representative images. (E) Expression of the hsp-4::GFP reporter (day 5 of adulthood) upon RNAi against genes encoding aminoacyl tRNA synthases, as well as rpc-1 RNAi and tunicamycin treatment (n = 10–25; ANOVA p < 10⁻⁴; p value from Dunnet’s test to vector control is indicated). Boxplots show quantiles with individual data points overlayed. Data underlying this figure can be found in S1 Data.

Fig 3. Partial inhibition of Pol III in flies alters tRNA expression and induces BiP.

(A) Volcano plots of differential expression of short RNAs (<300 b) in midguts of female Polr3D^EY/+ flies relative to wild type, with different classes of transcripts coloured in different panels. “Other” refers to U6 and several snoRNAs identified in a previous study (see main text). Horizontal line indicates the p value threshold for 10% FDR. (B) Northern blots of RNA from dissected guts, representative images. (C) Northern blots, quantification of relative tRNA expression (n = 5, Polr3DEY/+ versus wild type p = 0.035 one-tailed t test on least-squares means within LM; LM summary: effect of genotype p = 0.07, no significant effect of tRNA type or the genotype by type interaction). (D) Relative expression of pre-tRNA^His assessed by qPCR (n = 7–8, p = 1.5 × 10⁻³ t test). (E) TPM estimated in the RNA-Seq data for 3′ processed region of pre-tRNA^His and mature tRNA^His (n = 3, Polr3D^EY/+ versus wild type for pre-tRNA and tRNA separately p < 0.05 post hoc t tests on least-squares means within LM; LM summary: effect of genotype p > 0.05, effect of tRNA sequence p < 10⁻⁴, genotype by sequence interaction p = 2 × 10⁻³). (F) Relative expression of BiP mRNA in whole bodies of female Polr3D^EY/+ flies (n = 10, p < 10⁻⁴ t test). Boxplots show quantiles with individual data points overlayed. Data underlying this figure can be found in S1 Data. Uncropped images of blots are presented in S3 Fig.

Fig 4. TFIIIC activity promotes ageing in the fruit fly.

(A) Schematic of promoter types and TFs used by Pol III to initiate transcription. (B) Lifespans of GTF3C5^KG/+ and wild-type female flies (n = 134/4 and 124/7 dead/censored flies respectively, p = 2.3 × 10⁻¹³ log-rank test). (C) Changes in median lifespan relative to wild-type observed for the indicated genotypes in several experimental trials (see S4 Fig for details; p values from log-rank test are indicated). (D) Tunicamycin resistance of GTF3C5^KG/+ and wild-type female flies (n = 165/0 and 158/0 dead/censored flies, respectively, p = 1.3 × 10⁻⁸ log-rank test). (E) Relative expression of pre-tRNA^Leu (n = 5–8, p = 0.04 t test). (F) Lifespans of TIGS>GTF3C3^RNAi in the presence or absence of the RU486 inducer (n = 144/2 and 147/2 dead/censored flies respectively, p = 2.2 × 10⁻⁴ log-rank test). Data underlying this figure can be found in S1 Data.

Fig 5. Pol III and TFIIIC limit fly health in late life.

(A) Cumulative proportion of partial and full smurfs (flies with impaired gut barrier function) in cohorts of Polr3D^EY/+, GTF3C5^KG/+ and wild-type females (n = 52–788, effect of genotype: mutants versus wild-type p = 0.11 and GTF3C5^KG/+ versus Polr3D^EY/+ p = 0.88, age p < 2 × 10⁻¹⁶, age by genotype interaction: mutants versus wild-type p = 9.8 × 10⁻⁴ and GTF3C5^KG/+ versus Polr3D^EY/+ p = 0.42, ordinal logistic regression). Mean and standard error of the mean shown. (B) Time spent in the central zone of the arena during exploratory walking by Polr3D^EY/+ and wild-type females (n = 14–16, effect of genotype p = 0.01, age p = 2.7 × 10⁻³, age by genotype interaction p = 9 × 10⁻³, LM). (C) Hight climbed by Polr3D^EY/+ and wild-type females in a climbing (negative geotaxis) assay (n = 99–126, effect of genotype p = 4.7 × 10⁻⁶, age p < 2 × 10⁻¹⁶, age by genotype interaction p = 2.7 × 10⁻⁶, LM). (D) Hight climbed by Polr3D^EY/+, GTF3C5^KG/+ and wild-type females in a climbing assay (n = 95–108, effect of genotype: mutants versus wild-type p = 0.09 and GTF3C5^KG/+ versus Polr3D^EY/+ p = 1.7 × 10⁻³, age p < 2 × 10⁻¹⁶, age by genotype interaction: mutants versus wild-type p = 0.046 and GTF3C5^KG/+ versus Polr3D^EY/+ p = 0.32; LM). Boxplots show quantiles with individual data points overlayed. Data underlying this figure can be found in S1 Data.

Fig 6. Partial inhibition of Pol III alters tRNA expression in mice.

Volcano plots of differential expression of short RNAs (<300 b) in the superior duodenum between Polr3B^-/+ and wild-type mice, with different classes of transcripts coloured in different panels. (A) Female. (B) Male. “Other” refers to 7SK, RNase P, MRP, U6, Vault, and Y RNA; horizontal line indicates the p value threshold for 10% FDR. (C) Thapsigargin LD50 of primary fibroblasts in culture, isolated from female or male, wild-type or Polr3B^-/+ mice (n = 5–7, effect of genotype p = 6 × 10⁻³, effect of sex or interaction p < 0.05, LM). Boxplots show quantiles with individual data points overlayed. Data underlying this figure can be found in S1 Data.

Fig 7. Predicted and observed changes in translation upon partial inhibition of Pol III across worms, flies, and mice.

Computationally predicted fold changes in codon decoding times resulting from changes in the tRNA pool in animals with a reduced function of Pol III compared to controls, based on the RNA-Seq data. (A) C. elegans. (B) Fruit fly. (C) Female mouse. Note differences in Y axes between panels. (D) Translation rates measured by puromycin incorporation in worms, example blot and quantification (n = 3, p = 0.033 paired t test). (E) Translation rates measured in female fly guts, example blot and quantification (n = 5, p = 0.64 paired t test). Boxplots show quantiles with individual data points overlayed. (F) Volcano plot of differential translation efficiency in whole bodies of female Polr3D^EY/+ flies relative to wild type, estimated from Ribo-/RNA-Seq analysis. Position of Clbn is indicated. Data underlying this figure can be found in S1 Data. Uncropped images of blots are presented in S10 Fig.