Gene silencing by ingested dsRNA during larval development does not persist into adulthood.

Silencing in the germline was measured after exposure of gtbp-1::gfp animals to bacteria expressing dsRNA by imaging separate cohorts at indicated stages (a) or day 3 of adulthood (b). a and b, left, Schematics depicting stages and durations of exposure to dsRNA. a and b, right, GFP intensity (a.u.) in gtbp-1::gfp animals quantified in germ cells (larvae) or eggs in utero (adults) (a) or in day 3 adult (a3) animals (b) after exposure to control dsRNA (black) or gfp-dsRNA (red). The numbers of animals scored at each stage (n) are indicated. Asterisks indicate P < 0.05 with Bonferroni correction using Mann-Whitney U test for two-sided comparisons between animals exposed to control or gfp-dsRNA. Also see Extended Data Fig. 1.

Oxidative damage of neurons enhances gene silencing by neuronal dsRNA in the adult germline.

a, Schematic illustrating exposure to blue light of animals expressing a singlet oxygen generator (miniSOG) and gfp-dsRNA in neurons, and subsequent release of dsRNA. Such extracellular dsRNA is expected to enter the cytosol of the germline through the dsRNA importer SID-1 and silence gfp::h2b mRNA from a two-gene operon that expresses mCherry::h2b and gfp::h2b as part of a single pre-mRNA. b-d, Images of single gonad arms in adult animals with the two-gene operon (mex-5p::mCherry::h2b::gfp::h2b) showing fluorescence (black) of mCherry::H2B (magenta outline) or of GFP::H2B (green outline). Punctate autofluorescence from the intestine can also be seen. Numbers of animals assayed (n) and percentages of adult animals with the depicted expression patterns are indicated. Scale bars, 50 μm. b, mCherry::H2B fluorescence is seen throughout the germline (left) and GFP::H2B fluorescence is seen in the oocytes and in the distal gonad (right). c, GFP::H2B fluorescence in sid-1(+) and sid-1(-) animals expressing membrane-localized miniSOG (PH::miniSOG) and gfp-dsRNA driven by a neuronal promoter (rgef-1p) from a multi-copy transgene (Ex, jamEx214) without (left) or with (right) exposure to blue light at 48 hours post L4-stage of parent. d, mCherry::H2B fluorescence in sid-1(+) animals with the transgene Ex. Silencing of mCherry is enhanced in the distal gonad (third row) and sperm (fourth row) after exposing animals to blue light at 48 hours and 54 hours post L4-stage of parent. By region, silencing after exposure to light (right) in the proximal germline (57% = 10 + 18 + 29) > distal germline (47% = 18 + 29) > sperm (29%). Also see Extended Data Fig. 2 and Extended Data Fig. 3.

Transport of dsRNA from parental circulation to progeny occurs through two routes with distinct substrate selectivity.

a, Hermaphrodite animals of indicated genotypes (in red) were injected in the body cavity with 50-bp unc-22-dsRNA synthesized with a 5’-OH (short dsRNA, left bars) or unc-22-dsRNA with a 5’ triphosphate transcribed from a ∼1.1 kb template (mixed dsRNA, right bars). Hermaphrodite self-progeny of injected animals were scored for unc-22 silencing (fr. Unc-22: strong, black; weak, grey). Numbers of injected parents and scored progeny (P0; F1 n) are indicated. Also see Extended Data Fig. 3 and Extended Data Fig. 4. b, Fluorescence images of progeny from animals with a gfp tag of the ubiquitously expressed gene gtbp-1 (gtbp-1::gfp) that were not injected (left), injected with 50-bp gfp-dsRNA (short dsRNA injection, middle), or injected with dsRNA transcribed from a ∼730-bp template (mixed dsRNA injection, right). Complete silencing is not observed in neurons or in the developing vulva; brackets indicate additional regions with dim GFP fluorescence. Numbers of animals assayed (n) and percentages of L4-staged animals with the depicted expression patterns are indicated. Scale bar, 100 μm. Also see Extended Data Fig. 5. c, Requirements for intergenerational transport of extracellular dsRNA. (top left) Differential Interference Contrast (DIC) and fluorescence images of a developing embryo from an animal injected in the body cavity with 50-bp dsRNA of the same sequence as in b and labeled at the 5’ end of the antisense strand with Atto-565. Accumulation within the intestinal lumen (arrowhead), number of embryos imaged (n), and percentage of embryos with depicted pattern of fluorescence are indicated. Scale bar, 20 μm. (top right and bottom) Hermaphrodite animals of the indicated genotypes were injected with short dsRNA (left bars) or mixed dsRNA (right bars) and self-progeny (top right) or cross progeny after mating with wild-type males (bottom) were analyzed as in a. Cases of no observable silencing are indicated with ‘0’. d, Schematic summarizing requirements for transport of dsRNA from parental circulation to developing progeny. See text for details. Asterisks in a and c indicate P < 0.05 with Bonferroni correction using χ2 test.

The expression pattern of SID-1 varies during development.

a, Schematic depicting insertion of mCherry sequence that lacks piRNA binding sites42,43 (jam195[mCherryΔpi]) into the sid-1 gene using Cas9-mediated genome editing. b and c, Representative images showing fluorescence from SID-1::mCherry (black) in (b) the adult gonad arm, (c, left) developing embryos, (c, middle) L1-stage animals, or (c, right) L4-stage animals with sid-1(jam195[mCherryΔpi]) compared to autofluorescence in wild-type animals of the same stages. Numbers (n) of each stage imaged are indicated (100% of animals exhibited the depicted expression patterns). For animals imaged in b, the distal germline was obstructed by the intestine in 1/10 sid-1(jam195[mCherryΔpi]) and 5/9 wild-type animals. d, Airyscan image of an L1-staged animal assembled by stitching four different Z-stacks after depth-coding and taking maximum projections, illustrating the expression of SID-1::mCherry throughout the animal. Scale bar for adult gonad arms in b and embryos in c, 20 μm; scale bar for larvae in c, 50 μm andin d, 10 μm. Also see Technical comments on “Making a sid-1 translational reporter” in Methods.

Ancestral loss of SID-1 causes transgenerational changes in the mRNA levels of two germline genes that are subject to RNA regulation.

a, Schematic of modifications at the sid-1 gene generated using Cas9-mediated genome editing. Deletion of the entire coding sequence (jam113[deletion]), a nonsense mutation (jam80[nonsense]), and its reversion to wild-type sequence (jam86[revertant]) are depicted. b, Fractions of animals with the indicated genotypes that show silencing in response to unc-22-dsRNA (grey) or bli-1-dsRNA (black). Numbers of animals scored (n), significant differences using two-tailed test with Wilson’s estimates for single proportions (asterisks, P < 0.05 with Bonferroni correction) and 95% CI (error bars) are indicated. c, Principal components explaining the variance between wild type (black), sid-1(jam80[nonsense]) (red), and sid-1(jam86[revertant]) (grey) polyA+ RNA samples. Almost all of the variance between samples is explained by PC 1. d, Volcano plots of changes in the abundance of polyA+ RNA in sid-1(jam80[nonsense]) (top) and sid-1(jam86[revertant]) (bottom) animals compared with wild-type animals (black, q < 0.05; red, both q < 0.05 and change in the same direction in sid-1(jam80[nonsense]) and sid-1(jam113[deletion]); see Extended Data Fig. 7). While sid-1 transcript levels in sid-1(jam86[revertant]) are comparable to that in wild type (grey), sdg-1 (W09B7.2/F07B7.2) and sdg-2 (Y102A5C.36) transcript levels remain elevated in sid-1(jam86[revertant]) (red). e, Levels of spliced sid-1 (top), sdg-1 (middle) and sdg-2 (bottom) transcripts measured using RT-qPCR. The median of three technical replicates is plotted for each of three biological replicates (bar indicates median) assayed before and after 1 year of passaging animals (year 1, dark grey; year 2, light grey). Asterisks indicate P < 0.05 with Bonferroni correction using two-tailed Student’s t-test. f, Heatmap showing changes in the levels of transcripts (total RNA or mRNA) or antisense small RNAs (22G RNA) from sid-1, sdg-1, sdg-2, and tbb-2 (abundant germline transcript for comparison). Fold changes (expressed as LogFC, indicating log2 for (m)RNA, log10 for piRNA binding, and log10 for 22G RNA) were deduced by integrating reports (study) of 21 experiments that identify subsets of genes as being subject to RNA-mediated regulation within the germline (# genes). These prior studies include comparisons of RNA or 22G RNA from wild-type animals with that from mutant animals (e.g., mut-16(-) 22G RNA), biochemical detection of piRNA binding to transcripts (piRNA-bound mRNA), and biochemical detection of 22G RNA binding to an Argonaute (HRDE-1-bound 22G RNA). ‘NS’ indicates cases where changes, if any, were not significant based on the criteria used in the study. A conservative value of 2-fold is assigned to all genes reported as changing >2-fold in ref.58.

The sdg-1 gene is prone to stochastic changes in gene expression that can become heritable.

a, Representative images showing fluorescence of SDG-1::mCherry (black) in a wild-type background. While most animals showed symmetric expression in the germline (left), animals with >2-fold difference in fluorescence between both gonad arms (bright anterior, middle and bright posterior, right) were also observed. Punctate fluorescence in the intestine likely represents autofluorescence. Scale bar, 50 μm. b, Quantification of SDG-1::mCherry fluorescence intensity (arbitrary units, a.u.) in adult gonad arms (anterior arm, dark grey; posterior arm, light grey) of sdg-1(jam137[mCherryΔpi]) animals starting in one generation (x) and continuing in successive generations as indicated. Numbers of gonad arms quantified (n) is indicated. Expression in one generation was not significantly different when compared to that in the previous tested generation using Mann-Whitney U test for two-sided comparisons and Bonferroni correction. c, Lineages and estimated relative sdg-1 expression 10 generations after mating wild-type (open circle) males with sdg-1::mCherryΔpi (filled circle) hermaphrodites and vice versa, and isolating sdg-1(+) or sdg-1::mCherry animals from F1 heterozygotes (half-filled circle). Expression of sdg-1 in the F10 generation was measured by RT-qPCR of sdg-1 mRNA purified from pooled wild-type animals of mixed stages or by quantification of SDG-1::mCherry fluorescence in gonad arms of adult sdg-1::mCherryΔpi animals. Relative levels of sdg-1 mRNA and SDG-1::mCherry fluorescence intensity were converted to units of estimated relative sdg-1 expression (see Methods) for comparison. See Extended Data Fig. 9a for raw data. d-f, Fluorescence intensity measurements (quantified as in b) in adult animals with sdg-1::mCherryΔpi (+) and additionally with mutations in genes introduced through genetic crosses (in regulators of dsRNA import rme-2, sid-2 or sid-5, or in regulators of RNA silencing mut-16 or eri-1) or through genome editing (in regulators of dsRNA import sid-1 or sid-3, or in regulators of RNA silencing rde-1 or deps-1). Asterisks indicate P < 0.05 with Bonferroni correction using Mann-Whitney U test for two-sided comparisons between animals with sdg-1::mCherryΔpi (+) and animals with additional mutations. Nonsense mutations (nonsense) or deletions (deletion) introduced through genetic crosses (isolate numbers #1, #2, etc. in d) or genome editing (different alleles in e and f) and numbers of gonad arms (n) quantified for each isolate are indicated. Mutations in genes required for dsRNA import or subsequent silencing resulted in fewer animals with asymmetric fluorescence between gonad arms (a combined proportion of 21/197 for sid-1, sid-3, rde-1 and deps-1 mutants versus 22/84 for wild type, P = 0.0009 using two-tailed test with Wilson’s estimates for single proportions). Animals with at least one gonad arm brighter than the dimmest wild-type gonad arm in a and with asymmetric gonad arms were found in different genotypes (anterior bright: sid-1(-) – 5/122, sid-3(-) – 1/29, rde-1(-) – 2/22, deps-1(-) – 4/24, and posterior bright: sid-1(-) – 6/122, rde-1(-) – 2/22, deps-1(-) – 1/24). g, Fluorescence intensity measurements as in b of animals with sdg-1::mCherryΔpi that show loss of fluorescence when a nonsense mutation is introduced in sid-1 using genome editing ∼30 generations (gen.) later remain changed despite reversion of sid-1 nonsense mutation to wild-type sequence after ∼20 additional generations. Subsequent mutation of deps-1 after another ∼110 generations restored SDG-1::mCherry fluorescence to wild-type levels. Also see Extended Data Fig. 8 and Extended Data Fig. 9.

SID-1 modifies RNA regulation within the germline, potentially through sdg-1 and other sid-1-dependent genes.

a, (left) Schematic of assay for sensitive detection of pos-1 silencing by ingested dsRNA. (right) Numbers of developed progeny (> 3rd larval stage) laid by wild-type animals, animals with a deletion (Δ) in sdg-1 (jam232, jam241, jam242) or animals with overexpression (sdg-1::mCherryΔpi) of sdg-1 exposed to pos-1 dsRNA (red) or control dsRNA (black) for 16 hours are plotted. Asterisks indicate P < 0.05 using Mann-Whitney U test for two-sided comparisons with Bonferroni correction. b, Cross progeny males that inherited the mex-5p::mCherry::h2b::gfp::h2b transgene (T)43 (also used in Fig. 2) from maternal (left) or paternal (right) parents, both of wild-type, sid-1(-), or sdg-1(-) background, were scored for expression of mCherry and GFP (bright, dim, off). Wild-type data for top set (n = 77 and n = 33) are replotted from ref.43 for comparison. Dashed line separates independent experiments. Asterisk indicates P < 0.05 with Bonferroni correction using χ2 test; n.s. indicates not significant. c, Representative AiryScan images of the distal germline (left; scale bar, 10 µm) or single germline nuclei (right; scale bar, 2 µm) showing SDG-1::mCherry alone (top) or with GFP::ZNFX-1 (bottom, merge and single channel images). The number of animals imaged (n) and the percentage that show enrichment of SDG-1::mCherry in perinuclear foci are indicated. Sites of SDG-1::mCherry enrichment coincide with GFP::ZNFX-1 localization. Boxes in left mark the nuclei shown in right. d, Representative images showing entry of SDG-1::mCherry into the nucleus in -1 oocytes (left) and upon pronuclear fusion in early embryos during the time course indicated (right). Numbers of germlines and embryos imaged are indicated. Scale bars, 20 µm. Also see Extended Data Movies 1-4. e, Representative image of the hermaphrodite germline in animals with a translational (left) or transcriptional (right) reporter of sdg-1. Scale bars, 20 µm. Apparent extracellular punctae of SDG-1::mCherry and mCherry surrounding the proximal germline requires further study, but could be non-specific because similar localization is observed in animals with other promoters driving mCherry expression, but not GFP expression, in the germline (data not shown). The numbers of animals with the depicted fluorescence pattern are indicated. f and g, Response of the transcriptional sdg-1 reporter (sdg-1p::mCherryΔpi[sdg-1(Δ)]::sdg-1 3’ UTR) to the addition of unc-22-dsRNA (f) or loss of rde-4 (g). Quantification and asterisk are as in Fig. 6. h, Models for dsRNA import into the germline (top) and subsequent RNA-mediated regulation of sdg-1 (bottom). See text for details.