Inter and transgenerational impact of H3K4 methylation in neuronal homeostasis

Abstract

Epigenetic marks and associated traits can be transmitted for one or more generations, phenomena known respectively as inter- or transgenerational epigenetic inheritance. It remains unknown if genetically and conditionally induced aberrant epigenetic states can influence the development of the nervous system across generations. Here, we show, using Caenorhabditis elegans as a model system, that alteration of H3K4me3 levels in the parental generation, caused by genetic manipulation or changes in parental conditions, has, respectively, trans- and intergenerational effects on H3K4 methylome, transcriptome, and nervous system development. Thus, our study reveals the relevance of H3K4me3 transmission and maintenance in preventing long-lasting deleterious effects in nervous system homeostasis.

Figure 1.. Transgenerational effects of H3K4me3 modifiers on PVQ development.

(A) Top: ventral view of transgenic animals expressing oyIs14 fluorescent reporter visualizing PVQ neurons in WT (top) and in set-2(zr1208) mutant (bottom). In WT, PVQ cell bodies are positioned posteriorly and their axons project anteriorly on each side of the ventral midline. In set-2 mutants, the axon of left PVQ fails to respect the midline and defasciculates to the right ventral nerve cord (designated by white arrow) to later returns. Posterior is on the right. Scalebar 50 μm. Bottom: schematic depiction of WT and mutant PVQs. (B) Schematic depiction of the outcrossing strategy employed to follow wild-type descendants (WT-des) and mutant descendants. P₀ is the crossed parental generation. F₁ is the first generation where cross-progeny is selected, and hermaphrodites are left to self-fertilize. F₂ is the second generation from which independent lines are established. F_N is the third and subsequent generations. (C) PVQ defects at indicated generations of independent mutant descendants lines from WT males crossed with set-2(zr1208) hermaphrodites. The number of lines with defects increases over generations, from two lines in the third generation (F3) to all 10 lines in the sixth generation (F6). (D) PVQ defects in F3 WT-des independent lines, from WT males crossed with set-2(zr1208) hermaphrodites. (E) PVQ defects in F3 WT-des independent lines, from WT males crossed with set-2(2012) hermaphrodites. (F) PVQ defects in F3 WT-des independent lines, from WT males crossed with ash-2(tm1905) hermaphrodites. (G) PVQ defects in F3 WT-ctrl independent lines, from WT males crossed with WT hermaphrodites. Zero lines are significantly different compared with WT. (H) PVQ defects in F3 WT-des independent lines coming from WT males crossed with unc-6(ev400) hermaphrodites. Zero lines are significantly different compared with WT. (I) PVQ defects in F3 WT-des independent lines, from WT males crossed with set-2(zr1208) hrde-1(tm1200) hermaphrodites. In C-I, each grey bar represents the scoring from a single independent line coming from 1-d-old hermaphrodites (n = 50–52). Phenotypic threshold (red stippled line) is defined as the lowest PVQ penetrance required for an independent line (n = 50–52) to be significant different (P < 0.05) from the WT PVQ penetrance defect (6%, n = 200) using chi-square method or Fisher’s exact test.

Figure S1.. H3K4me3 levels in set-2 mutants and PVQ defects in H3K4me3 modifiers at different generations.

(A) Western blots of H3K4me1/2/3 in WT and set-2(zr1208) mutants (left) and in WT and set-2(zr2012) mutants (right). H3 was used as loading control. Western blots were carried out at least three times and numbers underneath represent mean levels ± SEM. (B) PVQ defects in WT-des lines at the fourth (F4) generation. Seven WT-des-independent lines from WT males crossed with set-2(zr1208) hermaphrodites with PVQ defects at the third generation were analysed. (C) PVQ defects in WT-des lines at third (F3) generation. 38 independent lines from set-2(zr1208) males crossed with WT hermaphrodites. Two lines are significantly different compared with WT. (D) PVQ defects in the WT-des fourth (F4) generation. Six independent lines from WT males crossed with set-2(zr2012) hermaphrodites with PVQ defects in the third generation were analysed. Two lines are significantly different compared with WT. (E) PVQ defects in the WT-des fourth generation. Five independent lines from WT males crossed with ash-2(tm1905) hermaphrodites with PVQ defects in the third (F3) generation were analysed. Three lines are significantly different compared with WT. (F) PVQ defects in WT and hrde-1(tm1200) mutants. n ≥ 150 pr. condition. n.s, not statistically significant compared with WT. Data are expressed as mean ± SEM. Each black dot represents an independent experiment. Statistical significance was assessed using t test comparing with WT-ctrl. (G) Summary of the PVQ defects in experiments presented in Figs 1D–I and S1C. Each black dot represents an independent experiment (n = 50–52). Statistical significance and P-value were assessed using t test. In (B, C, D, E), every grey bar represents the scoring from a single independent line coming from 1-d-old hermaphrodite (n = 50–52). Phenotypic threshold (red stippled line) is defined as the lowest PVQ penetrance required for an independent line (n = 50–52) to be significantly different (P < 0.05) from the WT PVQ penetrance (6%, n = 200) using chi-square method or Fisher’s exact test.

Figure 2.. H3K4me3 and transcriptome changes in descendants.

Wild-type descendant (WT-des) and mutant descendant (MUT-des) lines originates from crossing WT males with set-2(zr1208) hermaphrodites. (A) Principal component analysis of H3K4me3 ChIP-seq data from WT-des and MUT-des lines at F3 and F5 generations. (B) Box plot of median H3K4 signal (all peaks) based on H3K4me3 ChIP-seq data from F3 and F5 generations in WT-des and MUT-des lines. Boxes are 25^th to 75^th percentile, whiskers represent min and max. (C) H3K4 profile plots based on H3K4me3 ChIP-seq data from independent samples from WT-des and MUT-des lines at F3 and F5 generations. (D) Number of H3K4me3 peaks in WT-des and MUT-des lines at F3 and F5 generation. All peaks: union of peaks found in all conditions. Common peaks: overlapping peaks in all conditions. (E) Principal component analysis of mRNA expression the L4 stage in WT-des, MUT-des, and WT-ctrl at F3 and F5 generations. Every dot is an independent line. WT-ctrl originates from WT males crossed with WT hermaphrodites. (F) Heatmap based on RNA-seq data showing up- and down-regulated genes at the L4 stage in F3 and F5 descendants. The genes shown are differentially regulated between WT-ctrl (F3 and F5) and MUT-des (F3 and F5). Gene expression is shown as a relative Z-score across samples.

Figure S2.. ChIP H3K4me3 analyses.

(A) H3K4me3 ChIP-seq heatmap generated from three F3 and F5 WT-des and MUT-des lines. The signal is plotted over the peak center (peak center ±1 kb). (B) H3K4me3 tracks in WT-des F3 and F5 compared with Jänes et al (2018). L4 stage. (C) Box plot of median H3K4 signal at F3–F5 common peaks, based on ChIP-seq data from WT-des lines at F3 and F5 generations. Boxes are the 25^th to 75^th percentiles, whiskers represent min and max. (D) Number of peaks based on H3K4me3 ChIP-seq data in WT-des lines at F3 and F5 generations. The number of peaks in common between F3 and F5 is also shown. (E) Number of peaks identified by H3K4me3 ChIP-seq in Jänes et al (2018) (L4 stage), and in WT-des lines at F3 and F5 generations. Common peaks between the groups are also reported. (F) Spearman correlation heatmap and hierarchical clustering of H3K4me3 ChIP-seq signal from Jänes et al (2018) (L4 stage), WT-des lines at F3 and F5 generations at common peaks between the three groups.

Figure 3.. Parental conditions can affect neuronal developmental, H3K4me3 level, and transcriptome in F1 descendants.

(A) Schematic depiction of different conditions used. (B) PVQ defects in starved animals. WT animals were starved for 24 h at the L4 stage. n ≥ 150 pr. condition. n.s. not statistically significant comparing with progeny from non-starved animals. (C) PVQ defects in WT L4 animals exposed to heat (25°C) at indicate generations, and after moving the animals (F3 generation) back to low temperature (20°C). n ≥ 210 pr. condition. *P < 0.05, **P < 0.01, ****P < 0.0001, and n.s. not statistically significant, when comparing with WT grown at 20°C or between columns, as indicated by black lines. (D) PVQ defects in progeny in ageing WT animals (1, 3- and 4-d old adults). n ≥ 190 pr. condition. **P < 0.01, ****P < 0.0001, comparing with progeny from 1-d old hermaphrodites or between columns, as indicated by black lines. In (B, C, D), data are expressed as mean ± SEM. Each black dot represents an independent experiment. (B, C, D) Statistical significance was assessed using t test (B) or one-way ANOVA Tukey’s multiple comparison (C, D). (E) Representative Western blot of H3K4me3 levels (L4 stage). Left: animals grown at 20°C and shifted at 25°C for two generations. Middle: 1-day-old adults and 4-day-old adults (4DOA). Right: progeny (F1) from 1DOA and 4DOA. H3 is used as loading control. Western blots were carried out at least three times. Numbers represents mean levels ± SEM. (F) Principal component analysis of mRNA expression under different conditions. Each dot represents an independent biological replicate in the RNA-seq analysis. Control (black) represents the progeny from 1-d-old animals grown at 20°C. Maternal age (green) represents progeny from 4DOA at 20°C. F1-25°C (red) represents the progeny from 1DOA exposed to 25°C for one generation. F1-back (yellow) represents the progeny from 1DOA exposed to 25°C for two generations and moved back to 20°C for one generation. (G) Differentially expressed (DE) genes in animals (L4 stage) exposed to different conditions. F1-4DOA, DE genes identified comparing the L4 progeny from 1DOA and from 4DOA. F1-25°C, DE genes identified comparing L4 progeny from animals kept at 20°C and 25°C for one generation. F1-back, DE genes identified comparing L4 progeny from animals kept at 20°C with animals kept at 20°C for one generation after being exposed to 25°C for two generations. Dots represents DE genes (log₂FC > ± 1, Padj < 0.05). Red bar is the average log₂FC. Numbers represent numbers of genes up- and down-regulated under indicated conditions.

Figure S3.. PVQ defects and histone PTM levels at different maternal ages, generations, and in daf-2/daf-16 mutants.

(A) PVQ defects in F1 and F2 progenies from 4-day-old adult (4DOA) WT hermaphrodites. n ≥ 190 pr. condition. ****P < 0.0001, n.s. not statistically significant comparing with the progeny from 1-d-old hermaphrodites. (B) PVQ defects in the progenies from WT, daf-2(e1370), and daf-16(mu86) mutants from 1-day-old (1DOA), 3-day-old (3DOA), and 4-day-old (4DOA) adults. n ≥ 200 pr. condition. **P < 0.01, ****P < 0.0001, n.s. not statistically significant comparing with progeny from 1DOA hermaphrodites of the same genotype or between columns, as indicated by black lines. In (A, B), data are expressed as mean ± SEM. Each black dot represents independent experiments. Statistical significance was assessed using one-way ANOVA Tukey’s multiple comparison. (C) Quantification of histone PTM levels from Western blots of 4DOA compared with 1DOA. H3 was used as loading control. Western blots were carried out at least twice. Black dots represent independent experiments and bars represents mean levels ± SEM. (D) Quantification of histone PTM levels from Western blots of the progeny from 4DOA compared with the progeny from 1DOA. H3 was used as loading control. Western blots were carried out at least three times. Black dots represent independent experiments and bars represent mean levels ± SEM.

Figure 4.. Alterations related to parental conditions are SET-2 dependent.

(A) PVQ defects in the progeny from 1-day-old adults (1DOA) and from 4-day-old adults (4DOA) WT, set-2(zr1208) and set-16(zr1804). n ≥ 150 pr. condition. *P < 0.05, ***P < 0.001, n.s. not statistically significant, comparing with the progeny from 1DOA same genotype or between columns, as indicated by black lines. (B) PVQ defects in WT, set-2(zr1208) and set-16(zr1804) animals grown at 20°C and at 25°C for three generations. n ≥ 150 pr. condition. ***P < 0.001, ****P < 0.0001, n.s. not statistically significant comparing 25°C with 20°C same phenotype or between columns, as indicated by black lines. (C) Left. Differentially expressed genes (log₂FC > ± 1, Padj > 0.05) based on RNA-seq, identified comparing the L4 progeny from 1DOA and 4DOA WT animals. log₂FC of the same genes in the L4 progeny from 1DOA and 4DOA set-2(zr1208) mutants (set-2) and in the L4 progeny from 4DOA hermaphrodites crossed with young males compared with progeny from 1DOA hermaphrodites crossed with young males (wt crossed). Right: differentially regulated genes (log₂FC > ± 1, Padj > 0.05) based on RNA-seq, identified comparing the L4 progeny from WT grown at 20°C and WT kept for one generation at 25°C (WT). log₂FC of the same genes in progeny of set-2(zr1208) mutants grown at 20°C and set-2(zr1208) kept one generation at 25°C (set-2). (D) PVQ defects in self-progeny from 4DOA WT hermaphrodites and in cross-progeny from hermaphrodites crossed with males, in different age combinations. n ≥ 190 pr. condition. **P < 0.01, ***P < 0.001, n.s. not statistically significant, comparing with self-progeny from 4DOA hermaphrodites. (E) PVQ defects in cross-progeny from 4DOA hermaphrodites crossed with L4 males either WT, set-2(zr1208) or set-16(zr1804). n ≥ 150 pr. condition. **P < 0.01, n.s. not statistically significant, comparing with self-progeny from 4DOA hermaphrodites. (F) PVQ defects in cross-progeny from hermaphrodites kept at 25°C for two generations crossed with males kept at 20°C until adulthood (crossing done at 25°C) either WT, set-2(zr1208) or set-16(zr1804). n ≥ 150 pr. condition. **P < 0.01, n.s. not statistically significant, comparing with the self-progeny from 4DOA hermaphrodites. In (A, B, D, E, F), data are expressed as mean ± SEM. Each black dot represents an independent experiment. Statistical significance was assessed using one-way ANOVA Tukey’s multiple comparison. (G) Schematic model for H3K4me3 contribution in neuronal homeostasis across generations.