Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2019 Jun 13;177(7):1814-1826.e15.
doi: 10.1016/j.cell.2019.04.029. Epub 2019 Jun 6.

Neuronal Small RNAs Control Behavior Transgenerationally

Affiliations
Free PMC article

Neuronal Small RNAs Control Behavior Transgenerationally

Rachel Posner et al. Cell. .
Free PMC article

Abstract

It is unknown whether the activity of the nervous system can be inherited. In Caenorhabditis elegans nematodes, parental responses can transmit heritable small RNAs that regulate gene expression transgenerationally. In this study, we show that a neuronal process can impact the next generations. Neurons-specific synthesis of RDE-4-dependent small RNAs regulates germline amplified endogenous small interfering RNAs (siRNAs) and germline gene expression for multiple generations. Further, the production of small RNAs in neurons controls the chemotaxis behavior of the progeny for at least three generations via the germline Argonaute HRDE-1. Among the targets of these small RNAs, we identified the conserved gene saeg-2, which is transgenerationally downregulated in the germline. Silencing of saeg-2 following neuronal small RNA biogenesis is required for chemotaxis under stress. Thus, we propose a small-RNA-based mechanism for communication of neuronal processes transgenerationally.

Keywords: C. elegans; epigenetic inheritance; neuronal small RNAs; non-Mendelian inheritance; small RNA inheritance; transgenerational inheritance.

Figures

None
Figure S1
Figure S1
High-Copy Expression of RDE-4 in Neurons Regulates a Subset of STGs, Related to Figure 1 (A) A typical image demonstrating the neuronal expression pattern of the rescued RDE-4 (Prgef-1::rde-4::SL2::yfp), as monitored by examination of a trans-spliced YFP fluorescent reporter. Prgef-1::rde-4::SL2::yfp was co-injected with Punc122::GFP expressed in coelomocytes (marked in red). Bar = 20 μm. (B and C) smFISH staining of yfp transcripts (magenta) and DAPI nuclei staining (blue) in one typical worm expressing Prgef-1::rde-4::SL2::yfp. Shown are focal plains focusing on the neuronal ventral chord (B), yellow dashed lines), and the germline (C), white dashed lines). Bar = 20 μm. (D and E) Expression of STGs in rescued Psng-1::rde-4 (D) and Prgef-1::rde-4 (E) worms (y axis) compared to rde-4(ne299) mutants (x axis). Shown are the averaged expression values (log2 of RPM) of STGs (See also Table S2). Each dot represents an STG. Red dots: STGs that display differential expression between groups (analyzed with Deseq2, adjusted p value < 0.1). (F and G) x-fold enrichment and depletion values of upregulated STGs (upper panel) and downregulated STGs (lower panel) following RDE-4 High-Copy rescue in neurons. P values for enrichment were calculated using 10,000 random gene sets identical in size to the tested group (See “STAR Methods” for details). For the clarity of display, complete depletion (linear enrichment = 0) appears with the smallest value in the scale. Enrichments were considered significant if p < 0.05. ns- p > 0.05. ∗∗∗- p < 0.001. ∗∗∗∗- p < 10−4.
Figure 1
Figure 1
Characterization of Small RNA Changes following Rescue of RDE-4 in Neurons (A) Nervous system-specific rescue of RDE-4. A typical image demonstrating the neuronal expression pattern of the rescued RDE-4 (Psng-1::rde-4::SL2::yfp), as monitored by examination of a trans-spliced YFP fluorescent reporter. Bar, 20 μm. (B and C) smFISH staining of yfp transcripts (magenta) and DAPI nuclei staining (blue) in one typical worm expressing the integrated single-copy Psng-1::rde-4::SL2::yfp pan-neuronal rescue transgene. Shown are focal plains focusing on the neuronal ventral chord (B, yellow dashed lines), and the germline (C, white dashed lines). Bar, 20 μm. (D) Expression of STGs in rescued rde-4(n299);Psng-1::rde-4 worms (y axis) compared to rde-4(ne299) mutants (x axis). Shown are the averaged expression values (log2 of RPM) of STGs (see also Table S2). Each dot represents an STG. Red dots, STGs that display differential expression between groups (analyzed with Deseq2, adjusted p value < 0.1). (E) x-fold enrichment and depletion values of upregulated STGs and downregulated STGs following RDE-4 rescue in neurons. We tested the enrichment of the RDE-4-dependent STGs against lists of STGs that are known to require DCR-1 for their biogenesis, to bind the Argonaute ERGO-1, and to depend on somatic mut-16 activity (Welker et al., 2010, Zhang et al., 2011, Vasale et al., 2010). p values for enrichment were calculated using 10,000 random gene sets identical in size to the tested group (see STAR Methods for details). Enrichments were considered significant if p < 0.05. Not significant [ns], p > 0.05; ∗∗∗p < 0.001; ∗∗∗∗p < 10−4. (F) STGs distributions. Shown are STGs normalized read counts (y axis) as function of genomic location (x axis) of small RNAs targeting the genes ser-5 and C46G7.5 STGs (in red) in N2 wild-type worms, rde-4 mutants, and Psng-1::rde-4 rescue worms. Exons appear on a gray background. Blue arrow points to the direction of transcription. See also Figure S1.
Figure 2
Figure 2
Sorting of Neurons Followed by RNA Sequencing Allows the Characterization of Neuronal RDE-4-Dependent Small RNAs (A) Scheme depicting the production of RNA libraries specifically from neurons of C. elegans. Single-cell suspensions were produced out of wild type, rde-4(−), and Psng-1::rde-4 strains that express Prab-3::rfp in neurons (Kaletsky et al., 2016, Stefanakis et al., 2015), followed by immediate fluorescence-activated cell sorting (FACS). Sorted RFP+ neurons were collected for total RNA isolation. (B) Expression levels of NeuroSTGs in rescued Psng-1::rde-4 worms (y axis) compared to rde-4(ne299) mutants (x axis). Shown are the averaged expression values (log2 of rpm) of NeuroSTGs (see also Table S5). Each dot represents a NeuroSTG. 46 NeuroSTGs (red) displayed differential expression between groups (analyzed with Deseq2, adjusted p value < 0.1). (C) x-fold enrichment or depletion values of upregulated NeuroSTGs (left) and downregulated NeuroSTGs (right) following neuronal RDE-4 rescue. See also Figure 1E. For the clarity of display, complete depletion (linear enrichment = 0) appears with the smallest value in the scale. ns, p > 0.05; ∗∗∗∗p < 10−4. (D) Changes in neuronal mRNA levels (y axis) in Psng-1::rde-4 compared to rde-4(−) neurons, plotted against changes in their associated NeuroSTGs (x axis). Each dot represents the values for one gene, and the 46 genes with significant changes in their corresponding NeuroSTGs are shown (analyzed with Deseq2, adjusted p value < 0.1). Nine genes (red) exhibited also differential mRNA expression (analyzed with Deseq2, adjusted p value < 0.1). See also Figure S2 and Table S3.
Figure S2
Figure S2
Isolation and Sequencing of Neuronal Small RNAs and Neuronal mRNAs from N2 Wild-type Worms, Related to Figure 2 (A) Normalized mRNA levels reported by Kaletsky et al. (2016) (y axis), versus the mRNA levels (log2 of RPM) measured in four samples of N2 worms collected in independent experiments (x axis). Each dot represents a gene. As many points may overlap each other, and to better visualize the distribution of the data, we added a color code reflecting the number of genes in each bin. (B) Shown is a histogram indicating the proportion of STGs (y axis) with different expression levels, displayed by log2 of RPM average (x axis). The vertical red line corresponds to value of 5 RPM (linear scale). We used a cut-off of > 5RPM to create a list of robustly expressed STGs in N2 neurons (See Table S3). Please note the scale in the y axis is changing, with steps of 0.01 in the range of 0 to 0.1, and steps of 0.1 for proportions higher than 0.1. (C) Enriched GO terms for the sub-set of 412 genes targeted by NeuroSTGs with RPM > 5, which are upregulated in isolated neurons in comparison to STGs extracted from the entire animal. As can be seen, 6 of the 7 most enriched GO terms depict a variety of neuronal processes. Analysis done using the GOrilla tool (Eden et al., 2009); all the enriched GO terms with FDR < 0.05 are displayed.
Figure 3
Figure 3
The Germline of Psng-1::rde-4 Worms Is Devoid of Functional RDE-4 (A) Worms with the indicated genotype (y axis) were allowed to lay eggs on plates containing dsRNA-producing bacteria targeting the germline-expressed genes pos-1 and mel-26 or an empty-vector control. Shown are the percentage of hatched eggs per plate (x axis) following exposure to RNAi. Each dot represents one tested plate (biological replicate). Bars represent mean ± SD. Each group was tested in at least three independent experiments. p values were determined by two-way ANOVA with Tukey’s post hoc correction for multiple comparison, ∗∗∗∗p < 10−4; ns, p > 0.05. (B) Multiple sequence alignment of all the sequencing reads aligned to the genomic locations in the vicinity of the insertion defining the rde-4(ne299) allele. We combined all the reads (30) obtained from three independent replicate gonads samples from Psng-1::rde-4 worms. The wild-type rde-4 sequence is shown at the bottom row. We display only reads in which the insertion site is neither in the edge of the read nor included in soft clipping region of the CIGAR string. Shown is the complementary strand of the rde-4 gene, with the insertion position (chr-III: 10,218,186) marked in a red rectangle. See also Figure S3.
Figure S3
Figure S3
Worms Expressing Pan-Neuronal RDE-4 off High-Copy Transgenes Regulate mRNA Targets Transgenerationally, Related to Figures 3, 4, and 5 (A) Worms with the indicated genotype (y axis) were allowed to lay eggs on plates containing dsRNA-producing bacteria targeting the germline-expressed genes pos-1 & mel-26 or an empty-vector control. Shown are the percentage of hatched eggs per plate (x axis) following exposure to RNAi. Each dot represents one tested plate (biological replicate). Each group was tested in at least three independent experiments including n > = 2 biological replicates. P values were determined by Two-way ANOVA with Tukey’s post hoc correction for multiple comparison. ns- p > 0.05. ∗∗∗∗- p < 10−4. Related to Figure 3. (B) Clustering of STGs based on changes in whole-animal samples from worms rescued with the indicated transgene, compared to rde-4(-) mutants. “P0” depicts the data for the rescued lines, and “F1” and “F3” depict the data for their progeny that have lost the High-Copy transgene (See “STAR Methods” section and Table S4). Shown are all STGs displaying significant differential expression in P0 (analyzed with Deseq2, adjusted p value < 0.1). Genes which did not show significant differential expression in F1 or F3 (adjusted p value > = 0.1) are colored in gray. (C) (i) Representative images of smFISH staining against C18D4.6 (upper panel) and C55C3.3 (lower panel) in the posterior gonad of indicated genotypes. The stained worms were synchronized as late L4s. For representation, all images were filtered according to the FISH-quant software (Mueller et al., 2013), projected in the Z axis by maximum intensity and threshold adjusted, identically between conditions. Scale bars = 20μm (ii) Quantification of C18D4.6 and C55C3.3 germline mRNA expression (FISH-quant, see methods) in the indicated genotypes. Each dot represents one quantified worm, and worms were tested on three independent experiments. P values were determined by Kruskal-Wallis test with Dunn’s post hoc correction for multiple comparison and asterisks represent P values in comparison to rde-4(-). ∗∗∗∗- p < 10-4,∗∗- p < 0.01. Bars represent mean ± SD. (D and E) smFISH quantification of worms with the indicated genotype (x axis) to determine the impact of hrde-1 (D) and sid-1 (E) on the germline regulation of C18D4.6 (upper panel) and C55C3.3 (lower panel) by High-Copy Psng-1::rde-4. n.s- p > 0.05. ∗∗∗- p < 0.001. ∗∗∗∗- p < 10−4.
Figure 4
Figure 4
Neuronal RDE-4 Expression Leads to Transgenerational Inheritance of Endogenous Small RNAs and Germline Regulation of Cognate mRNA Targets (A) Expression of GermSTGs in rescued rde-4(n299);Psng-1::rde-4 worms (y axis) compared to rde-4(ne299) mutants (x axis). Shown are the averaged expression values (log2 of RPM) of GermSTGs (see also Table S4). Each dot represents a GermSTG. Red dots, GermSTGs that display differential expression between groups (analyzed with Deseq2, adjusted p value < 0.1). (B) Clustering of STGs based on changes in whole-animal samples from rescued rde-4(−);Psng-1::rde-4 worms compared to rde-4(−) mutants (left), and F3 rde-4(−) progeny of rde-4(−);Psng-1::rde-4(+/−) heterozygote rescue worms compared to rde-4(−) (right). Shown are all STGs displaying significant differential expression in P0 (analyzed with Deseq2, adjusted p value < 0.1) (see also Table S4). Genes that did not show significant differential expression in F3 (adjusted p value ≥0.1) are colored in gray. (C) x-fold enrichment or depletion values of differentially expressed STGs, for small RNAs bound to the germline Argonautes CSR-1 (Claycomb et al., 2009) and HRDE-1 (Buckley et al., 2012). Tissue and generation of the analyzed STG samples are indicated. “P0” denotes samples extracted from rescued rde-4(−);Psng-1::rde-4 worms compared to rde-4(−) mutants. “F3” denotes samples extracted from F3 rde-4(−) progeny of rde-4(−);Psng-1::rde-4(+/−) compared to rde-4(−). Enrichments were considered significant if p < 0.05. ns, p > 0.05; p < 0.05; ∗∗p < 0.01; ∗∗∗∗p < 10−4. (D) Changes in germline mRNA levels (y axis) in Psng-1::rde-4 compared to rde-4(−), plotted against changes in their associated GermSTGs (x axis) (see also Table S4). Each dot represents the values for one gene, and the 124 genes with significant changes in germline mRNAs are shown (analyzed with Deseq2, adjusted p value < 0.1). 40 genes (colored) exhibited also differential STGs expression (analyzed with Deseq2, adjusted p value < 0.1). 18/40 genes (triangles) are regulated by the ZFP-1/DOT-1.1 complex. 7/40 genes (in blue) encode for histone proteins. saeg-2 is marked by a black circle. (E) Changes in germline mRNA levels (y axis) in F3 rde-4(−) progeny of Psng-1::rde-4(+/−) compared to rde-4(−), plotted against changes in their associated GermSTGs (x axis) (see also Table S4). Shown are the 40 genes with differentially expressed mRNA and STGs from (D). Five genes (colored as in D, with their full name indicated) displayed differentially expressed mRNA and STGs also in the F3 generation (analyzed with Deseq2, adjusted p value < 0.1). See also Figure S3 and Table S4.
Figure 5
Figure 5
Neuronal RDE-4-Dependent Small RNAs Regulate Germline Expression of saeg-2 Transgenerationally in a HRDE-1-Dependent Manner (A) STGs read distribution along the saeg-2 gene. Shown are STGs normalized read counts (y axis) against genomic location (x axis) of the small RNAs that target saeg-2 (in red) aligned to the gene locus in N2 wild-type worms, rde-4 mutants, and rde-4(−);Psng-1::rde-4 rescue worms. Exons appear on a gray background. Blue arrow points to the direction of transcription. (B) Representative images of smFISH staining against saeg-2 in worms of the indicated genotype. The stained worms were synchronized as late L4s. For representation, all images were filtered according to the FISH-quant software (Mueller et al., 2013), projected in the z axis by maximum intensity and threshold adjusted, identically between conditions. Scale bar, 20 μm. (C–E) Quantification of saeg-2 germline mRNA expression by smFISH in the indicated genotypes. Levels of saeg-2 mRNA in the germline are transgenerationally downregulated by neuronal RDE-4 (C), in a hrde-1-dependent (D) and sid-1-independent (E) manner. The groups were tested on three separate trials (except for N2s data in C obtained from two trials). Each dot represents one quantified worm. Black bars represent mean ± SD. p values were determined by Kruskal-Wallis test with Dunn’s post hoc correction for multiple comparison. ns, p > 0.05, p < 0.05, ∗∗p < 0.01, ∗∗∗∗p < 10−4. See also Figures S3 and S4 and Table S5.
Figure S4
Figure S4
SID-1 Regulates a Subset of GermSTGs that Depend on Neuronal RDE-4, Related to Figure 5 Shown is a Venn diagram depicting the criteria applied in order to detect sid-1-dependent GermSTGs in high confidence. The number of STGs passing each criteria appears in parenthesis. We kept only STGs that were: (1) upregulated in the germline in Psng-1::rde-4 versus rde-4(-), (2) downregulated in sid-1;Psng-1::rde-4 versus Psng-1::rde-4, (3) not downregulated in sid-1;rde-4(-) versus rde-4(-), (4) not upregulated in sid-1;rde-4(-) versus sid-1;Psng-1::rde-4.
Figure 6
Figure 6
Neuronal RDE-4 Controls Behavior Transgenerationally via the Germline Small RNA Machinery Results for experiments testing chemotaxis to benzaldehyde (1:100) at day 1 of adulthood of worms (ethanol was used as control odor). Chemotaxis index = ((# worms at benzaldehyde) − (# worms at ethanol))/((# total worms on plate) − (# worms at origin)). Each dot represents one plate with >200 worms. All groups were tested on at least three independent trials, each including several biological replicates. Black bars represent mean ± SD. For convenience, each biological group was assigned a letter label. p values were determined by Kruskal-Wallis test with Dunn’s post hoc correction for multiple comparison to the rde-4(−) group. ns, p > 0.05; p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 10−4. (A) Chemotaxis experiments on the F3 rde-4(−) progeny of rde-4(+/−) P0 ancestors and their hrde-1 and sid-1 double mutants, together with control strains. (B) Chemotaxis experiments on the F3 rde-4(−) progeny of Psng-1::rde-4(+/−);rde-4(−/−) P0 ancestors and their hrde-1 and sid-1 double mutants, together with control strains. (C) Chemotaxis experiments on the rde-4;saeg-2 double mutants alleles, together with control strains. Three double mutant strains were generated via CRISPR/Cas9. See also Figure S5.
Figure S5
Figure S5
rde-4 Mutants Are Defective in Chemotaxis under a Stressful Temperature but Display Normal Activity in the AWC Sensory Neuron, Related to Figure 6 (A and B) Results for experiments testing chemotaxis at day 1 of adulthood of worms. Chemotaxis index = ((# worms at stimulus)-(# worms at control)) / ((# total worms on plate)-(# worms at origin)). Each dot represents one plate with > 200 worms. All groups were tested on at least three independent trials (n = > 9). (A) Chemotaxis screens were performed on RNAi factor mutants at both 20 (blue dots) and 25 degrees (red dots). Chemotaxis indices (x axis) were tested for the strains (y axis) N2(wild-type), rde-4(ne299), dcr-1(mg375), ergo-1 (tm1860), nrde-3 (gg66), rrf-1(ok589), rrf-3(pk1426), mut-16(pk710) and eri-6(mg379). The odor stimulus used was benzeldahyde (1:100). P values were determined by two-way ANOVA, ∗∗∗∗-p < 10−4. (B) rde-4 mutants are defective in chemotaxis to multiple stimuli at high temperature. N2(wild-type) (black dots) and rde-4 mutants (green dots) raised at 25 degrees for chemotaxis to Benzaldehyde (1:102), Butanone (1:104), Diacetyl (1:103) and NaCl (50mM). P values were determined by Two-way ANOVA with Sidak’s post hoc correction for multiple comparisons. ∗∗∗∗- p < 10−4. (C) A Principal Component Analysis (PCA) projection of 12 samples based on normalized STGs read counts. Each symbol represents one independent replicate. The corresponding genotype and temperature are indicated. The % variances, out of the total original variance in the high-dimensional space, spanned by the first and second Principal Components are indicated on the x- and y- axis, respectively. Related to Figure 6. (D) Activity of sensory neuron AWC was quantified by GCaMP2 fluorescence intensity in a microfluidic device controlling stimulus exposure. N2 (wild-type) (n = 15) and rde-4 (n = 20) mutants were loaded into chips and exposed to the stimulus Isoamyl alcohol (1:104) for 1 minute followed by a switch to buffer (indicated by the red arrow). Each row represents an individual worm. Shown are the fluorescence intensity values normalized from 0 to 1 (Ft-Fmin)/(Fmax-Fmin) across time (seconds). (E) Maximum fluorescence intensity increase (x axis) of the N2 and rde-4 worms (y axis) of AWCneurons in response to odor removal. Mean peak was defined as ΔF (2 s pre-stimuli) – max ΔF (post-stimuli). (F) Time (x axis) it took from the moment of odor removal for AWC neurons in N2 and rde-4 worms (y axis) to reach maximum fluorescence intensity. P values were determined by Mann-Whitney tests. n.s.- p > 0.05.
Figure S6
Figure S6
Worms that Are Homozygous for the hrde-1 Mutant Allele pig4 Are Defective in RNAi Inheritance, Related to STAR Methods Worms heterozygote to the indicated genotype and expressing a germline GFP reporter were exposed to bacteria expressing anti-GFP dsRNA (P0 generation). We used heterozygotes so that the worms could initiate an RNAi response (rde-4 mutants do not initiate RNAi responses). GFP silencing levels were tested in the P0 heterozygotes and their homozygote mutated F2 progeny. These experiments were conducted in three independent replicates (n = > 30). (A) Relative GFP (GFP/empty vector) fluorescence in F2 homozygotes. Each dot represents one worm. P values were determined by one-way ANOVA and Tukey’s post hoc correction for multiple comparison. ∗∗∗∗- p < 10−4; n.s.- p > 0.05. Data shown are means ± SD. (B) Relative GFP fluorescence levels (GFP/empty vector) between worm populations exposed to GFP RNAi at the P0 generation were averaged across trials.

Comment in

Similar articles

See all similar articles

Cited by 6 articles

See all "Cited by" articles

References

    1. Alcazar R.M., Lin R., Fire A.Z. Transmission dynamics of heritable silencing induced by double-stranded RNA in Caenorhabditis elegans. Genetics. 2008;180:1275–1288. - PMC - PubMed
    2. Alcazar, R.M., Lin, R., and Fire, A.Z. (2008). Transmission dynamics of heritable silencing induced by double-stranded RNA in Caenorhabditis elegans. Genetics 180, 1275-1288. - PMC - PubMed
    1. Anava S., Posner R., Rechavi O. The soft genome. Worm. 2015;3:e989798. - PMC - PubMed
    2. Anava, S., Posner, R., and Rechavi, O. (2015). The soft genome. Worm 3, e989798. - PMC - PubMed
    1. Anders S., Pyl P.T., Huber W. HTSeq--a Python framework to work with high-throughput sequencing data. Bioinformatics. 2015;31:166–169. - PMC - PubMed
    2. Anders, S., Pyl, P.T., and Huber, W. (2015). HTSeq--a Python framework to work with high-throughput sequencing data. Bioinformatics 31, 166-169. - PMC - PubMed
    1. Aoki K., Moriguchi H., Yoshioka T., Okawa K., Tabara H. In vitro analyses of the production and activity of secondary small interfering RNAs in C. elegans. EMBO J. 2007;26:5007–5019. - PMC - PubMed
    2. Aoki, K., Moriguchi, H., Yoshioka, T., Okawa, K., and Tabara, H. (2007). In vitro analyses of the production and activity of secondary small interfering RNAs in C. elegans. EMBO J. 26, 5007-5019. - PMC - PubMed
    1. Ashe A., Sapetschnig A., Weick E.M., Mitchell J., Bagijn M.P., Cording A.C., Doebley A.L., Goldstein L.D., Lehrbach N.J., Le Pen J. piRNAs can trigger a multigenerational epigenetic memory in the germline of C. elegans. Cell. 2012;150:88–99. - PMC - PubMed
    2. Ashe, A., Sapetschnig, A., Weick, E.M., Mitchell, J., Bagijn, M.P., Cording, A.C., Doebley, A.L., Goldstein, L.D., Lehrbach, N.J., Le Pen, J., et al. (2012). piRNAs can trigger a multigenerational epigenetic memory in the germline of C. elegans. Cell 150, 88-99. - PMC - PubMed

Publication types

MeSH terms

LinkOut - more resources

Feedback