Starvation-induced transgenerational inheritance of small RNAs in C. elegans

Abstract

Evidence from animal studies and human famines suggests that starvation may affect the health of the progeny of famished individuals. However, it is not clear whether starvation affects only immediate offspring or has lasting effects; it is also unclear how such epigenetic information is inherited. Small RNA-induced gene silencing can persist over several generations via transgenerationally inherited small RNA molecules in C. elegans, but all known transgenerational silencing responses are directed against foreign DNA introduced into the organism. We found that starvation-induced developmental arrest, a natural and drastic environmental change, leads to the generation of small RNAs that are inherited through at least three consecutive generations. These small, endogenous, transgenerationally transmitted RNAs target genes with roles in nutrition. We defined genes that are essential for this multigenerational effect. Moreover, we show that the F3 offspring of starved animals show an increased lifespan, corroborating the notion of a transgenerational memory of past conditions.

Figure 1. A Protocol for Analyzing the Starvation-Induced Transgenerational Small RNA Response

(A) Aschematic description of the experimental protocol: secondary (RdRP-amplified) small RNAs were cloned and sequenced from three biological replicates of young adult worms that were either continuously fed or severely starved for 6 daysasL1s and from theirF3 progeny. Following adapters trimming, the reads were mapped to the WS220 version of the C. elegans genome. Reads which mapped antisense to annotated gene were counted and then analyzed by principle component analysis to reveal the similarities/differences between the experimental conditions (see Figure S1 for an additional PCA of reads which were mapped to the entire genome). (B) Visualization in the UCSC browser of the vhl-1 gene. An STG (small rnas targeting a given gene) is defined as the sum of the small RNAs that are antisense to a certain gene (see Table S1 for the fold-change and p value ranking of all the STGs).

Figure 2. Transcriptome Analysis of P0 Adults that Experienced Starvation at the L1 Stage

(A) Overlaps of up/downregulated STGs of the different endo-siRNA species (see Figure S2 for an MA Plot visualization of the differentially expressed STGs of each small RNA species). (B) A functional annotation of the 22G STGs (secondary small RNAs which are 22 nt long and possess a 5′ guanosine) that are differentially expressed following starvation in wild-type animals. Annotation was performed using the GOrilla web-server. (I) Functional enrichment map. (II) A table of the specific GO terms, statistical significances, the number of genes that share each specific GO term, and their relative enrichments. An additional GO analysis, which also takes into account the fold-change ranking of the differentially expressed STGs is presented in Figure S3. (C) Changes in the mRNA levels that correspond to changes in STGs. The x axis represents the two species of small RNAs that were examined: 22Gs (secondary small RNAs) and 26Gs (primary small RNAs which are 26 nt long and possess a 5′ guanosine). The y axis represents the total number of STGs which were upregulated following L1-starvation, in correspondence to the changes in mRNA levels (see Table S3 for fold-change ranking of the genes).

Figure 3. Transgenerational Inheritance of Small RNAs Following L1 Starvation

(A) An example of a gene (hlh-30) targeted by heritable 22Gs. See Table S2 for the fold-change and p value ranking of all the inherited STGs. (B) The total number of inherited STGs of different small RNA species and their overlapping targets. (C) A functional annotation of STGs that are differentially expressed in an inherited manner following L1-starvation in WT animals. (I) GOrilla functional enrichment map. (II) A table of the specific GO terms, statistical significances, the number of genes that share each specific GO term, and their relative enrichments (see Figure S4 for genomic visualization of the nutrient reservoir activity genes).

Figure 4. Transgenerationally Inherited Small RNAs Differentially Regulate Their Cognate mRNAs and Are RDE-4 and HRDE-1-Dependent

(A) Plotting of the log2-fold changes of the inherited STGs, versus the mRNA log2-fold changes of their cognate targets. Shown separately are (I) HRDE-1 (II) CSR-1 targets (Buckley et al., 2012; Claycomb et al., 2009). Somatic genes are represented by blue dots and germline-expressed genes are represented by red dots. Shown is a linear correlation line (R² = 0.907) between the inherited STGs that were previously shown to bind CSR-1, and their germline-expressed mRNA targets. (B) The majority of the inherited STGs are depleted in rde-4 and hrde-1 mutants. Shown are the numbers of the small minority of STGs which are nevertheless inherited in the RNAi mutants and their corresponding mRNA changes.

Figure 5. Animals with a Starvation History Show an Extended Lifespan

A comparison of the lifespan of N2 worms whose great-grandparents had experienced starvation at the first larval stage with animals whose great-grandparents were continuously fed. The x axis shows lifespan in days of adulthood. The y axis shows the fraction of worms alive. N2 derived from starved great-grandparents show extended lifespan in comparison to N2 that were continuously fed over generations (live 37% longer, log-rank, p < 0.0001). Each of the lifespan experiments was performed three times (data from two additional experiments are shown in Figure S6).

Figure 6. A Model for L1 Starvation-Induced Transgenerational Inheritance

Food (bacterial lawn) is shown in dark green. Starvation-induced transcripts, which are transcribed either during or after worms experience L1 starvation, enable the production or depletion of specific RDE-4-dependent small RNAs, which regulate the levels of their mRNA targets (involved in nutrition). The production of the heritable small RNAs might be initiated in the germline, or, if endogenous small RNAs act systemically, it could originate also in somatic tissues. The changes in small RNAs persist in the absence of the trigger (starvation) to the next generations. The germline-expressed argonaute, HRDE-1, carries the heritable 22G small RNAs in the germline.