The microRNA machinery regulates fasting-induced changes in gene expression and longevity in Caenorhabditis elegans

Intermittent fasting (IF) is a dietary restriction regimen that extends the lifespans of Caenorhabditis elegans and mammals by inducing changes in gene expression. However, how IF induces these changes and promotes longevity remains unclear. One proposed mechanism involves gene regulation by microRNAs (miRNAs), small non-coding RNAs (∼22 nucleotides) that repress gene expression and whose expression can be altered by fasting. To test this proposition, we examined the role of the miRNA machinery in fasting-induced transcriptional changes and longevity in C. elegans. We revealed that fasting up-regulated the expression of the miRNA-induced silencing complex (miRISC) components, including Argonaute and GW182, and the miRNA-processing enzyme DRSH-1 (the ortholog of the Drosophila Drosha enzyme). Our lifespan measurements demonstrated that IF-induced longevity was suppressed by knock-out or knockdown of miRISC components and was completely inhibited by drsh-1 ablation. Remarkably, drsh-1 ablation inhibited the fasting-induced changes in the expression of the target genes of DAF-16, the insulin/IGF-1 signaling effector in C. elegans. Fasting-induced transcriptome alterations were substantially and modestly suppressed in the drsh-1 null mutant and the null mutant of ain-1, a gene encoding GW182, respectively. Moreover, miRNA array analyses revealed that the expression levels of numerous miRNAs changed after 2 days of fasting. These results indicate that components of the miRNA machinery, especially the miRNA-processing enzyme DRSH-1, play an important role in mediating IF-induced longevity via the regulation of fasting-induced changes in gene expression.


Intermittent fasting (IF) is a dietary restriction regimen that extends the lifespans of Caenorhabditis elegans and mammals by inducing changes in gene expression. However, how IF
induces these changes and promotes longevity remains unclear. One proposed mechanism involves gene regulation by microRNAs (miRNAs), small non-coding RNAs (ϳ22 nucleotides) that repress gene expression and whose expression can be altered by fasting. To test this proposition, we examined the role of the miRNA machinery in fasting-induced transcriptional changes and longevity in C. elegans. We revealed that fasting up-regulated the expression of the miRNA-induced silencing complex (miRISC) components, including Argonaute and GW182, and the miRNA-processing enzyme DRSH-1 (the ortholog of the Drosophila Drosha enzyme). Our lifespan measurements demonstrated that IF-induced longevity was suppressed by knock-out or knockdown of miRISC components and was completely inhibited by drsh-1 ablation. Remarkably, drsh-1 ablation inhibited the fasting-induced changes in the expression of the target genes of DAF-16, the insulin/IGF-1 signaling effector in C. elegans. Fasting-induced transcriptome alterations were substantially and modestly suppressed in the drsh-1 null mutant and the null mutant of ain-1, a gene encoding GW182, respectively. Moreover, miRNA array analyses revealed that the expression levels of numerous miRNAs changed after 2 days of fasting. These results indicate that components of the miRNA machinery, especially the miRNA-processing enzyme DRSH-1, play an important role in mediating IF-induced longevity via the regulation of fasting-induced changes in gene expression.
Dietary restriction (DR) 2 increases lifespan and prevents agerelated diseases in many organisms, ranging from yeast to mice (1)(2)(3)(4). Many forms of DR, including caloric restriction (CR), intermittent fasting (IF), and protein restriction (1,2,5), exist. An increasing number of studies indicate that fasting stimuli induce health benefits, such as lifespan extension and the prevention of diabetes and cardiovascular disease, and are considered to be a plausible intervention for slowing the rate of aging in humans (1,(5)(6)(7)(8). Our previous studies have shown that IF significantly extends the lifespan of Caenorhabditis elegans, and IF-induced longevity is mediated by the fasting-induced transcriptional alterations by two transcription factors: DAF-16, the insulin/IGF-1 signaling (IIS) pathway effector (9), and AP-1, the stress-activated MAP kinase JNK pathway effector (10). Thus, the importance of transcriptional changes in IF-induced longevity has been well documented. However, roles of post-transcriptional regulation in fasting-induced signaling remain ambiguous.
MicroRNAs (miRNAs) are a class of small non-coding RNAs that post-transcriptionally regulate gene expression (11)(12)(13). In the miRNA pathway, primary miRNA transcripts are cleaved by the microprocessor complex, which is composed of the ribonuclease (RNase) III enzyme Drosha/DRSH-1 and its cofactor DGCR8 (DiGeorge syndrome critical region gene 8)/PASH-1 (11)(12)(13). The processed products, termed precursor miRNAs, are exported to the cytoplasm, where the precursor miRNA stem-loop is processed by another RNase III, Dicer, thus generating mature miRNAs (11)(12)(13). Mature miRNAs form the miRNA-induced silencing complex (miRISC) with Argonaute protein (ALG-1 or ALG-2) and GW182 protein (AIN-1 or AIN-2), and the miRISC recognizes and represses target gene expression (12,14,15). Multiple roles of miRNAs in animals and plants have been reported in many biological processes, including aging (16 -23). Recently, mir-71 and mir-228 have been shown to regulate CR-induced longevity through transcription factors PHA-4 and SKN-1 (21). In addition, mir-80 null mutants show age-related phenotypes that are similar to those of eat-2 mutants (the model of CR) (22). The expression of Dicer decreases with aging; this decrease is suppressed by CR in mouse adipose tissues and C. elegans (23). The expression levels of several miRNAs are altered after 12 h of fasting in C. elegans at the L4 stage (24). However, the involvement of miRNA in IF-induced longevity remains unaddressed.
In this study, we examined the role of the miRNA machinery in fasting-induced transcriptional changes and longevity in C. elegans. Our results indicated that the miRISC components and the miRNA-processing enzyme Drosha/drsh-1 were upregulated by fasting and that knockdown or knock-out of miRISC components suppressed IF-induced longevity. Remarkably, drsh-1 null mutation completely suppressed IF-induced longevity and inhibited the fasting-induced up-regulation of the target genes of DAF-16, the IIS pathway effector. Also, the majority of the fasting-induced transcriptome alterations were suppressed in drsh-1 null mutants. Additionally, our miRNA array analysis indicated that the expression levels of a number of miRNAs changed after 48 h of fasting starting at day 2 of the adult stage. These results reveal involvement of the miRNA machinery, particularly the miRNA-processing enzyme DRSH-1, in fasting-induced changes in gene expression and IF-induced longevity.

Fasting induces changes in the expression of miRISC components
We examined the expression levels of miRISC components in worms under fed and fasting conditions. Remarkably, the genes encoding Argonaute proteins, alg-1 and alg-2, and the genes encoding GW182 proteins, ain-1 and ain-2, were up-regulated 2-3-fold after 48 h of fasting (Fig. 1A). Our immunoblot analysis revealed that the ALG-1 protein (GFP::ALG-1) level was significantly up-regulated by fasting (Fig. 1B, left), whereas the ALG-2 protein (ALG-2::HA) level was only slightly up-regulated although not statistically significant (Fig. 1B, right). These results indicate that the expression levels of miRISC components undergo large changes in response to fasting.

miRISC is involved in IF-induced longevity
Because fasting-induced gene expression alterations underlie IF-induced longevity (9, 10), we considered the possibility that fasting might induce gene expression changes and longevity, at least partly through the miRNA machinery. To test this possibility, we measured the lifespan of the null mutants of miRISC components, alg-1(gk214), alg-2(ok304), ain-1(tm3681), and ain-2(tm1863), and found that the IF-induced longevity was significantly suppressed in the null mutants of alg-1, ain-1, and ain-2 (Fig. 2, A and B, Table 1). Because we observed developmental defects in alg-1 null mutants, we knocked down miRISC components (alg-1, alg-2, ain-1, and ain-2) by using RNAi. Because the Argonaute proteins are required for normal development (25,26), we performed alg-1 or alg-2 RNAi after completion of development to prevent developmental defects. In the case of ain-1 and ain-2, we performed RNAi starting from the egg stage. The lifespan measurements revealed that the RNAi of alg-1, alg-2, or ain-1, compared with the control RNAi, caused a significant decrease in the IF-induced longevity (Fig. 3, A and B, Table 1). These results demonstrate that miRISC components are involved in regulation of the IF-induced longevity.

Expression of the miRNA-processing enzyme Drosha/DRSH-1 is enhanced by fasting, and drsh-1 ablation suppresses IF-induced longevity
To further examine the involvement of miRNA machinery in the fasting response, we focused on the miRNA-processing enzyme Drosha/DRSH-1 and its partner Pasha/PASH-1. Our quantitative RT-PCR (qRT-PCR) analysis revealed that the expression levels of drsh-1 and pash-1 were significantly upregulated after fasting, similarly to the miRISC components (Fig. 4A). To test whether DRSH-1 is involved in IF-induced longevity, we used the drsh-1 null mutants. Our lifespan measurements indicated that the null mutation of drsh-1 completely suppressed IF-induced longevity but did not affect the lifespan in ad libitum feeding conditions ( Fig. 4B and Table 1). However, it has previously been shown that a loss-of-function mutation of pash-1 resulted in shortened lifespans under ad libitum conditions in the absence of fluorodeoxyuridine (FUDR) (27). As the difference in the use of FUDR might affect the result, we also measured the lifespan of drsh-1 mutants in the absence of FUDR. The obtained results showed that the lifespan of drsh-1 mutants was shorter than that of wild type in the absence of FUDR ( Fig. 4C and Table 1), consistent with a previous report (27). In any case, our results suggest that the miRNA-processing pathway is activated by fasting and is required for IF-induced longevity.
libitum conditions in the mutant compared with the WT worms (Fig. 5). To further investigate the relationship between DRSH-1 and the IIS pathway, we examined expression of the fasting-repressed DAF-16 target gene ins-7 (9). The results indicated that ins-7 expression was not down-regulated by fasting in drsh-1 null mutants, whereas the expression was substantially decreased by fasting in WT worms (Fig. 5). These results indicate that DRSH-1 is involved in the regulation of the IIS pathway.
We then examined the effect of simultaneous depletion of miRNA synthesis and insulin signaling on the lifespan and expression of the DAF-16 target genes. daf-16 RNAi shortened the lifespan of drsh-1 mutants ( Fig. 6A and Table 1), indicating that DAF-16 regulates the lifespan in a DRSH-1-independent manner under ad libitum conditions. daf-16 RNAi partially suppressed IF-induced longevity in wild type and did not fur-  ther suppress IF-induced longevity in drsh-1 mutants (Fig. 6A), which is consistent with our idea that miRNAs synthesis is involved in regulation of the IIS pathway under IF conditions. drsh-1 depletion did not suppress expression of DAF-16 target genes under fed conditions, when daf-16 RNAi suppressed it (Fig. 6B). daf-16 RNAi partially suppressed fasting-induced changes in expression of DAF-16 target genes, drsh-1 depletion, and simultaneous depletion of daf-16 and drsh-1 completely suppressed them. These results suggest that DRSH-1 and DAF-16 act in parallel pathways in lifespan regulation under ad libitum conditions, and that DRSH-1 is involved in DAF-16 regulation under fasting conditions, which could contribute to IF-induced longevity.

Fasting-induced transcriptome alterations are suppressed by the ablation of miRNA machinery components
These results suggest that DRSH-1 plays an essential role in IF-induced longevity and is involved in regulation of the IIS pathway. It has previously been shown that the IIS pathway also plays a partial but important role in IF-induced longevity (9,10). Thus, we considered the possibility that DRSH-1 might also be involved in other pathways in the fasting response in addition to the IIS pathway. We performed microarray analysis using drsh-1, ain-1, and daf-16 null mutants under both fed conditions and fasting conditions. To validate the involvement of drsh-1 and ain-1 in fasting-induced gene expression changes, we compared the induction rates of all genes in the mutants with the induction rates of all genes in WT worms (Fig. 7). The correlation between WT worms and drsh-1 null mutants (r ϭ 0.601) was low compared with the correlation between WT worms and ain-1 null mutants (r ϭ 0.907) or daf-16 null mutants (r ϭ 0.808), thus suggesting a greater role of DRSH-1 in the fasting response.

Fasting induces substantial changes in the expression of miRNAs
The expression levels of several miRNAs change in response to fasting at the L4 stage in C. elegans (24). To examine the fasting-induced changes in miRNA expression in adult worms, we performed miRNA array experiments. The results indicated that the expression levels of numerous miRNAs underwent substantial changes in response to fasting (Fig. 8, left). The top 10 miRNAs whose expression was up-regulated or down-regulated by more than 1.5-fold after fasting are listed (Fig. 8, right). Previous reports have identified "age-related miRNAs" as miRNAs whose expression levels change during aging (19,20,28). Our analysis indicated that this group of age-related miRNAs (Fig. 8, colored miRNAs, supplemental Tables S1 and S2) is significantly enriched in fasting-induced up-regulated and down-regulated miRNAs (p Ͻ 0.0002, Fisher's exact test).

Discussion
In this study, we demonstrate that miRNA machinery, particularly the miRNA-processing enzyme Drosha/DRSH-1, is involved in fasting-induced changes in gene expression and IFinduced longevity in C. elegans. Our analysis revealed that miRISC components (alg-1, alg-2, ain-1, and ain-2) and the miRNA-processing enzyme drsh-1 are up-regulated by fasting, thus suggesting that the miRNA machinery is activated in response to fasting. The expression of miRNA machinery proteins (Argonaute, Dicer, and Drosha) in mouse adipose tissues has been reported to decrease with aging, and these decreases are suppressed by CR (23). The age-dependent decrease of Dicer in C. elegans is also suppressed by CR (23). This previous report has indicated that activity of the miRNA machinery is altered during aging and is regulated by food availability in mice and C. elegans. However, the involvement of the miRNA machinery in DR-induced longevity has remained unaddressed. In the present study, our results indicated that IF-induced longevity is suppressed by knock-out or knockdown of miRISC components and is completely inhibited by drsh-1 null mutations. Because miRISC and DRSH-1 are required for miRNA synthesis and function (11)(12)(13)(14)(15), miRNAs appear to play an important role in IF-induced longevity. Our analyses indicated that fasting-induced transcriptome alterations are significantly and modestly suppressed by the abrasion of drsh-1 and ain-1, respectively. This result correlates with the complete and partial suppression of IF-induced longevity observed after the ablation of drsh-1 and ain-1, respectively, and suggests that the fasting-induced gene expression alteration underlies the IF-induced longevity. Collectively, our results suggest that miRNAs play an important role in the IF-induced longevity by mediating the fasting-induced gene expression alterations. The expression of many miRNAs changes in response to fasting, which is consistent with our idea that the miRNA machinery is activated in response to fasting. Interestingly, we found that age-related miRNAs whose expression levels change during aging were significantly enriched in the fasting-induced miRNAs. These results suggest that the fasting-induced miRNAs could be involved in the regulation of longevity under IF conditions. We also noted that the fasting-induced miRNAs, which were reported in the previous study (24) result from the difference in the stage (L4 stage in the previous study versus day 2 adult in our study) and the duration of fasting stimulus (12 h in the previous study versus 48 h in our study). Two days of fasting decrease the expression of mir-80, which is shown to be an anti-longevity miRNA (22), and increases the expression of mir-34, a pro-longevity miRNA (29). These fasting-induced changes in miRNAs expression may contribute to IF-induced longevity. The IIS pathway plays an important role in IF-induced longevity (9, 10). Our qRT-PCR measurements indicated that fasting-induced changes in expression of the DAF-16 target genes are completely suppressed by drsh-1 null mutation, thus suggesting that DRSH-1 may regulate IF-induced longevity at least partly through the IIS pathway. Our microarray analyses showed that not only DAF-16 target genes but also the majority of the fasting-induced genes are regulated by DRSH-1, thus suggesting that other pathways are also regulated by the miRNA machinery. The pathways that are related to miRNA machinery in fasting conditions remain to be determined. Because the miRNA machinery is conserved between nematodes and mammals (30 -32), our findings may provide a new approach to the prevention of age-related diseases in humans.
Additional studies are needed to better understand the mechanisms of miRNA-mediated IF-induced longevity.

MicroRNA microarray analysis
Synchronized WT eggs were obtained by the bleaching method (33). Worms from these synchronous eggs were raised in normal conditions, and young adult animals were transferred to NGM plates that contained 200 g/ml of FUDR. The day on which the animals were transferred to FUDR-containing NGM plates was defined as t ϭ 0 day. On day 2, the animals were transferred to FUDR-containing NGM plates that were seeded with or without UV-killed OP50. The animals were collected at day 2 in fed conditions and at day 4 in fed or fasting conditions, and total RNA was extracted with TRIzol reagent (Invitrogen) from frozen animals. The miRNA content in total RNA was analyzed using an Agilent 2100 Bioanalyzer. Then, 1 g of total RNA was labeled with a FlashTag TM Biotin HSR RNA Labeling Kit (Affymetrix, Santa Clara, CA) for Affymetrix GeneChip miRNA arrays (Affymetrix) according to the manufacturer's recommendations. A simple colorimetric enzyme-linked oligosorbent assay was used to confirm successful biotin labeling. After labeling, the samples were hybridized on Affymetrix GeneChip miRNA arrays, washed, stained, and scanned according to the manufacturer's instructions (Affymetrix). The array data were normalized by global normalization with Robust Multichip Average and Detection Above Background by using the Expression Console software (Affymetrix).

Lifespan assay
We performed an intermittent fasting lifespan assay as indicated in Ref. 9. Synchronized WT eggs were obtained by the bleaching method (33). Worms from these synchronous eggs were raised in normal conditions, and young adult animals were transferred to NGM plates containing 200 g/ml of FUDR. The day on which the animals were transferred to FUDR-containing NGM plates was defined as t ϭ 0 day. On day 2, the animals were divided into an ad libitum group and an IF group. The animals in the ad libitum group were fed with UV-killed OP50 throughout their lifespan. The animals in the IF group were alternately placed on a plate with (2-3 days) or without (2-3 days) UV-killed OP50 as food. We scored the death events every 2-3 days. The animals were considered to be dead when they failed to respond to touch with a picker.
The animals were subjected to RNAi for 3 days from egg stage to young adult stage in ain-1 and ain-2 knockdown experiments and for 3 days from the L4 larva stage to day 2 of adulthood in alg-1 and alg-2 knockdown experiments. We define the term of "fasting" as a single duration of fasting stimulus (not repeated), and "IF (intermittent fasting)" as repeated fasting intervals in this study.

Immunoblot analysis
An immunoblot analysis was performed according to standard protocols. The pellet was lysed with SDS-sample buffer and subjected to an immunoblot analysis using primary antibodies (anti-GFP antibody (TAKARA) and anti-Histone H3 antibody (Abcam)) and secondary antibodies (anti-rabbit IgG antibody (GE Healthcare) and anti-mouse IgG antibody (GE Healthcare)).

Microarray analysis
Total RNA was isolated using TRIzol reagent (Invitrogen) in WT worms, drsh-1(ok369), ain-1(tm3681), and daf-16(mu86) mutants under fed and fasting conditions. cDNA synthesis from the total RNA was performed using a GeneChip 3Ј IVT PLUS Reagent Kit according to the manufacturer's protocol. RNA degradation and cRNA elongation were verified with an Agilent 2100 Bioanalyzer. The fragmented cRNA was hybridized using a GeneChip C. elegans Genome Array (Affymetrix) at 45°C for 16 h and a Hybridization Oven 640 (Affymetrix), and then was washed, stained in a GeneChip Fluidics Station 450 (Affymetrix), and scanned using an Affymetrix GeneChip Scanner. The scanned chip images were analyzed with Affymetrix GeneChip Command Console version 2.0 (AGCC) and processed using default settings. The Affymetrix outputs (CEL files) were imported into the GeneSpring GX 11.0.2 (Agi- lent Technologies) microarray analysis software for the presentation of the expression profiles. A hierarchical clustering analysis was performed with the squared Euclidean distance as the distance metric and average linkage as the cluster method by using GeneSpring GX. GO analyses were performed using GeneSpring GX.

Accession numbers
The microarray data are available in the Gene Expression Omnibus (GEO) database (www.ncbi.nlm.nih.gov) under the accession numbers GSE89609 and GSE89614.

Statistical analysis
Data are presented as mean Ϯ S.D. Statistical analysis was carried out using Prism software (GraphPad). For statistical analysis of 2 groups, unpaired Student's t test was used; for comparison of 3 or more groups, analysis of variance (ANOVA) followed by Tukey's test was applied.
Author contributions-A. K. and M. U. conceived the study and designed and performed the experiments. A. K., M. U., and E. N. wrote the manuscript. T. I. conducted the immunoblot analysis. E. N. supervised the project. All authors discussed the results and commented on the manuscript. . The red dots represent the genes whose expression changes were more than 1.5 or less than 0.67 in response to fasting. The expression levels of day 4 adult worms were normalized to the expression levels of day 2 adult worms. The solid lines indicate FC ϭ 1.5 and ϭ0.67. The tables list the top 10 miRNAs whose expression level was changed after 48 h of fasting. The up-regulated (middle) and down-regulated (right) genes are shown. The colored miRNAs represent age-related miRNAs (refer to text). p Ͻ 0.0002, Fisher's exact test.