Reduced longevity response to DR in ClkJrk mutant flies.

(A-B) Survival curves of wild-type control flies (iso31) and ClkJrkhomozygous mutant flies in 1%, 5%, 10%, 15%, and 20% sucrose-yeast (SY) diets (total dilution). (C) Mean lifespan plots of iso31 and ClkJrk flies across different concentrations of SY diets. Basic survival parameters from the Kaplan-Meier method for each diet and genotype are in Supplementary File 1. P values represent the probability of main (Gene (G) as a nominal variable and Diet (D) as a continuous variable) and interaction effects (G x D) by the likelihood ratio chi-square test from Cox proportional hazards regression analysis to evaluate ability of tested genes to modify lifespan in the specified range of diets. Separate analyses were performed for normal DR response (from 5SY to 20SY) and malnutrition-like response (from 5SY to 1SY). An independent replication of the experiment is presented in Figure 1-figure supplement 2. See Supplementary File 1 for additional details of the statistical analysis.

Reduced fecundity response to diets in ClkJrk mutant flies.

(A and B) Cumulative and daily average number of eggs produced per fly over 7 days in wild-type control flies (iso31) and ClkJrk homozygous mutant flies on 5%, 10%, 15%, and 20% sucrose-yeast (SY) diets. (C) Cumulative number of eggs produced per fly over 7 days. ** p < 0.01 by t-test for specified pairwise comparisons. All error bars represent SEM.

Effects of DR on circadian transcriptome in the abdominal fat body.

(A) Lifespan extension by DR. Samples for RNA-Seq analysis were collected after ∼5 days under either control or DR diets. (B) Number of rhythmic genes (RAIN, FDR<0.1 and log2 fold change >0.6). (C) Reorganization of the diurnal transcriptome by DR. The heatmap represents the relative expression (Z-score) of rhythmic genes in each group across 48 h at 2 h intervals (2 replicates of 12 time points over 24 h). Genes in the top panels are rhythmic in both control and DR diets (common); those in the middle panels are rhythmic in the control diet (left) but arrhythmic in the DR diet (right); those in the bottom panels are rhythmic in the DR diet (right) but arrhythmic in the control diet (left). (D) DR failed to affect the expression patterns of core clock genes. (E) Effect of DR on overall time-averaged expression of rhythmic genes in each group. TPMs of rhythmic genes in each group were averaged across all time points and normalized to those in the control diet. (F) Increased overall expression in the common rhythmic genes under DR. The heat map represents relative expression (Z-score) of common rhythmic genes across all time points from both control and DR diets. (G) Examples of common rhythmic genes with increased expression under DR.

Identification of the “proteasome module” as a diet-dependent differentially regulated module enriched with oscillating genes.

(A) Identification of a co-expression module enriched with proteasome subunit genes under DR. Gene co-expression networks under the DR diet were built using the WGCNA/r package. A topological overlap matrix (TOM) was then computed to evaluate the neighborhood similarity between genes and to classify network genes into modules using hierarchical clustering and dynamic tree cut (see Methods). (B) Venn Diagram shows the number of modules enriched with rhythmic genes (FDR < 0.05 and odd ratio > 0) and differentially regulated by diet (q < 0.1 and MDC > 1). Five modules were enriched with cycling genes and were deferentially connected between DR and control conditions. (C) Heatmap showing gene expression across two 24-h light/dark cycles in the five common modules. Each row represents a gene, and each column represents a sample ordered by sampling time. Expression levels were standardized across all samples and are presented as Z-scores. (D) Physical interaction map among the genes in the proteasome module in Bayesian network. Physical interaction mapping among the genes in the proteasome module was analyzed using a Bayesian network approach implemented in the bnlearn R package. (E) KEGG pathway enrichment analysis for the five differentially connected cycling modules. Enrichment scores in p-values were corrected using Benjamini-Hochberg approach with a threshold of 0.05.

Reduced longevity response to DR by prosβ3 and rpn7 knockdown in the abdominal fat body.

(A–B) Survival curves of flies with prosβ3 and rpn7 knockdown (+RU) in adult abdominal fat body (S106-GeneSwitch (GS) driver) and their controls (−RU) on control and DR diets. Survival curves were pooled from 3–4 independent trials (see also Figure 5-figure supplement 1, Supplementary File 1 for survival curves and detailed statistical analyses of independent trials). P < 0.0001 and p = 0.0169 for prosβ3 and rpn7, respectively, represent gene × diet interaction effects from Cox proportional hazards analysis.

Model for how clock and diet impact proteasome expression to regulate lifespan.

Acute DR increases the number of diurnally rhythmic genes, including proteasome genes, without significantly affecting core clock genes in the abdominal fat body. The diet-dependent gain of oscillation in proteasome genes enhances proteostasis and contributes to DR-mediated lifespan extension. In contrast, prolonged DR may further modulate the circadian clock itself, potentially increasing the rhythmicity of clock-output genes.

Reduced DR effect in ClkJrk mutant flies compared to iso31 wild-type and per01 mutant flies.

(A-C) Survival curves of iso31, ClkJrk and per01homozygous mutant flies on control (15% sucrose-yeast; 15SY) or DR (5% sucrose-east; 5SY) diets. DR extended lifespan in iso31, per01, and ClkJrk by 29.8%, 24.8%, and 11.3%, respectively. n = 117–200 flies (Supplementary File 1). Statistical significance of the diet x genotype interaction for ability to extend lifespan by DR between genotypes was analyzed by Cox proportional hazards regression analysis. p < 0.0001 for iso31 vs ClkJrk and ClkJrkvs per01, and p = 0.0635 for iso31 vs per01. See Supplementary File 1 for additional details of the statistical analysis.

Reduced longevity response to DR in ClkJrk mutant flies.

Independent replication of the experiment presented in Figure 1. (A and B) Survival curves of wild-type control flies (iso31) and ClkJrk homozygous mutant flies on 1%, 5%, 10%, 15%, and 20% sucrose-yeast (SY) diets. n = 198–208 flies (Supplementary File 1). (C) Mean lifespan plots of iso31 and ClkJrk flies across different concentrations of SY diets. P values represent the probability of main (Gene (G) as a nominal variable and Diet (D) as a continuous variable) and interaction effects (G x D) by the likelihood ratio chi-square test from Cox proportional hazards regression analysis to evaluate ability of tested genes to modify lifespan in the specified range of diets. Separate analyses were performed for normal DR response (from 5SY to 20SY) and malnutrition-like response (from 5SY to 1SY). See Supplementary File 1 for additional details of the statistical analysis.

Food Consumption in wild-type (iso31) and ClkJrk flies.

Young mated females (∼2 days old) were maintained on a designated diet for 5 days and then transferred to the same diet containing 1% blue dye (FD&C Blue No. 1) for 48 hours. The amount of dye in both internal tissues and excreta was quantified using a spectrophotometer at 630 nm and calculated using a standard curve. n = 5 replicates (3 flies per replicate). Two-way ANOVA with Sidak’s multiple comparisons: Diet: p < 0.0001, Genotype: p = 0.0112, Diet × Genotype interaction: p = 0.0476. A significant difference between WT and ClkJrk flies was observed only in the 1SY malnutrition diet, but not in the DR or control diets.

Reduced longevity response to yeast (Brewer’s yeast; whole cell lysates) restriction in ClkJrk mutant flies.

(A and B) Survival curves of wild-type control flies (iso31) and ClkJrkhomozygous mutant flies on 1%, 5%, 10%, 15%, and 20% yeast-restricted (Y) diets with a fixed sucrose concentration of 5%. n = 175–181 flies (Supplementary File 1). (C) Mean lifespan plots of iso31 and ClkJrkflies across different concentrations of Y diets. P values represent the probability of main (Gene (G) as a nominal variable and Diet (D) as a continuous variable) and interaction effects (G x D) by the likelihood ratio chi-square test from Cox proportional hazards regression analysis to evaluate ability of tested genes to modify lifespan in the specified range of diets. Separate analyses were performed for normal DR response (from 5Y to 20Y) and malnutrition-like response (from 5Y to 1Y). See Supplementary File 1 for additional details of the statistical analysis.

Reduced longevity response to yeast (yeast extract) restriction in ClkJrkmutant flies.

(A and B) Survival curves of wild-type control flies (iso31) and ClkJrkhomozygous mutant flies in varying concentrations of yeast extract with a fixed sucrose concentration of 5%. (C) Mean lifespan plots of iso31 and ClkJrk(Jrk) flies across different concentrations of yeast extract diets. Mal (malnutrition): 0.01%, DR: 0.5%, AL (ad libitum): 5% yeast extract. Upper and lower panels are from independent trials 1 and 2. See Supplementary File 1 for additional details of the statistical analysis.

Reduced fecundity response to diets in ClkJrk mutant flies.

Independent replication of the experiment presented in Figure 2. (A and B) Cumulative and daily average number of eggs produced per fly over 7 days in wild-type control flies (iso31) and ClkJrk homozygous mutant flies on 5% (n = 17, iso31; n = 18, ClkJrk), 10% (n = 19, iso31; n = 19, ClkJrk), 15% (n = 19, iso31; n = 13, ClkJrk), and 20% (n = 16, iso31; n = 18, ClkJrk) sucrose-yeast (SY) diets. (C) Cumulative number of eggs produced per fly over 7 days. * p < 0.05, ** p < 0.01 by t-test for specified pair-wise comparisons.

Clock is required for diurnal rhythmic gene expression in the abdominal fat body.

(A) Diurnal expression patterns of core clock genes timeless (tim) and period (per) in WT and ClkJrk flies. (B) Diurnal rhythmic expression patterns (red = high, blue = low) of common cyclers in WT flies under control and DR diets, and their expression pattern in ClkJrk mutants under control diet. The heatmap displays relative expressions over 48 hours (2-hour intervals for WT; 4-hour intervals for ClkJrk).

Effect of proteasome suppression in the abdominal fat body on lifespan and DR.

Proteasome subunits were individually knocked down in the adult abdominal fat body with the S106-Gene Switch (GS) driver. Mated females were aged on either control (15% sucrose-yeast) or DR (5% sucrose-yeast) diets and provided fresh food every 2∼3 days (-RU: 1% EtOH, +RU: 200 uM RU486 in 1% EtOH). n = 104–185. See Supplementary File 1 for additional details of the statistical analysis.

Effect of prosβ3 and rpn7 knockdown in the abdominal fat body on lifespan and DR.

Proteasome subunits (prosβ3 and rpn7) were individually knocked down in the adult abdominal fat body using the S106-GS driver. Mated females were aged on either control (15% sucrose-yeast) or DR (5% sucrose-yeast) diets and provided fresh food every 2∼3 days (-RU: 1% EtOH, +RU: 200 uM RU486 in 1% EtOH). For comparison among independent trials, the graphs for prosβ3 and rpn7 from Figure 5-figure supplement 1 are presented again as trial 1. n = 104–294 in independent trials. See Supplementary File 1 for additional details of the statistical analysis.

Effect of prosβ3 and rpn7 knockdown in the gut on lifespan and DR.

Proteasome subunits (prosβ3 and rpn7) were individually knocked down in the adult gut using the TI-GS driver. Mated females were aged on either control (15% sucrose-yeast) or DR (5% sucrose-yeast) diets and provided fresh food every 2∼3 days (-RU: 1% EtOH, +RU: 200 uM RU486 in 1% EtOH). n = 223–253. See Supplementary File 1 for additional details of the statistical analysis.