Histone H3K4me3 modification is a transgenerational epigenetic signal for lipid metabolism in Caenorhabditis elegans

As a major risk factor to human health, obesity presents a massive burden to people and society. Interestingly, the obese status of parents can cause progeny’s lipid accumulation through epigenetic inheritance in multiple species. To date, many questions remain as to how lipid accumulation leads to signals that are transmitted across generations. In this study, we establish a nematode model of C. elegans raised on a high-fat diet (HFD) that leads to measurable lipid accumulation, which can transmit the lipid accumulation signal to their multigenerational progeny. Using this model, we find that transcription factors DAF-16/FOXO and SBP-1/SREBP, nuclear receptors NHR-49 and NHR-80, and delta-9 desaturases (fat-5, fat-6, and fat-7) are required for transgenerational lipid accumulation. Additionally, histone H3K4 trimethylation (H3K4me3) marks lipid metabolism genes and increases their transcription response to multigenerational obesogenic effects. In summary, this study establishes an interaction between a network of lipid metabolic genes and chromatin modifications, which work together to achieve transgenerational epigenetic inheritance of obesogenic effects. Transgenerational inheritance (TEI) mechanisms are to some extent conserved across species, but how TEI mediates lipid accumulation is unknown. Here the authors reveal that a network of lipid metabolic genes and chromatin modifications mediated by transcription factors and H3K4 trimethylation work together to achieve multigenerational obesogenic effects in C. elegans fed with a high-fat diet.

M etabolic diseases are the result of a combination of genetic and environmental effects. Common environmental factors affecting metabolic diseases include nutrition, microbiome, exercise, and individual behaviors [1][2][3][4] . Interestingly, studies in animal models and humans have indicated that parents' experience can not only influence their own health but also affect their descendants, initiating a pass-down of disease risk across generations. For example, the parental experiences before and during pregnancy, are now recognized as an important factor in the incidences of obesity and type II diabetes (T2D) 5 . In modern societies, many people live in a condition of overnutrition that causes a high prevalence of obesity, and thus a variety of metabolic diseases such as T2D, hypertension, and coronary heart disease. Therefore, it is important to study whether and how obesity induces multigenerational inheritance, and to dissect the underlying molecular mechanisms.
The nematode worm Caenorhabditis elegans is a relevant and tractable model for the study of multigenerational obesogenic effects. Nematode genes for lipid metabolism are mostly conserved in mammalian models, as core lipid metabolic pathways in C. elegans have homologs in humans. The known transgenerational epigenetic inheritance (TEI) mechanism is also conserved across species. For example, histone modifications such as methylation of H3K4, H3K9, and H3K27, and inherited small RNAs are involved in TEI among multiple species in different scenarios [6][7][8][9] . Recently, some progress has been made in the understanding of TEI using nematodes as a model. For example, C. elegans learn to avoid pathogens and transmit the memory to their progeny for up to four generations [10][11][12] . An artificial environment such as Bisphenol A (BPA) exposure can cause transgenerational inheritance, with one-generational parental (P0) exposure of BPA leading to a decrease of H3K9me3 and H3K27me3 for five generations 8,13 . The memory of mitochondrial stress was reported to be transgenerationally inherited 14,15 . In our previous study, we found that hormetic heat stress-induced survival advantage, which had been reported in another study 16 , could be passed down to progeny through histone H3K9me3 and DNA N 6 -mA modifications 17 . A recent study reported that H3K9me2 was involved in TEI induced by heat shock 18 .
The question of whether, and if so how, TEI plays a role in lipid metabolism has attracted the attention of researchers in recent years. For example, it was reported that the TEI obesogenic effects of sulfomethoxazole are associated with histone H3K4me3 modification 19 ; another study indicated that a reduction in lipid accumulation induced by benzylisoquinoline can be transmitted to progeny 20 . However, the detailed mechanisms of the underlying multigenerational obesogenic effects remain largely unknown. For example, how are genes related to fat metabolism linked to chromatin modifiers to achieve the lipid metabolismassociated TEI? In this study, we use a high-fat diet (HFD, egg yolk) to feed C. elegans to establish a lipid accumulation model. Using this model, we characterize the critical factors functioning for "implementation" or "transmission" in TEI of lipid accumulation in C. elegans.

Results
Lipid accumulation from the HFD exposure can be transmitted across generations. Supplementation of different lipids, including oleic acid (OA) or palmitoleic acid (PA), in food, can significantly increase fat accumulation in C. elegans 21,22 . Similarly, using Oil-Red-O (ORO) staining, we found that the ORO level was obviously elevated when animals were fed with OA, PA, cholesterol, or egg yolk compared with the OP50 medium, and that supplementation with egg yolk showed the most significant lipid accumulation (Fig. 1a). Therefore, in the subsequent assays, we used egg yolk as a HFD to feed C. elegans and established a fat accumulation model. To test whether the HFD-induced lipid accumulation phenomenon can be transmitted to the descendants raised on OP50, we randomly selected one wild-type hermaphrodite worm and obtained a population with the same genetic background through its self-fertilization. Then we fed wild-type worms with HFD and bleached the mothers to obtain F1, F2, and Fn generations (Fig. 1b). We found that the ORO levels were significantly increased, in both HFD-fed parents and their F1 and F2 progeny recovered on normal OP50 (Fig. 1c, d), suggesting that HFD can induce a multigenerational epigenetic inheritance phenotype on descendants from a single exposure of the P0 generation. Considering that the F1 generation may have been exposed to HFD while still inside the mother, in which case the obesogenic effect in the F1 worms would be considered intergenerational inheritance instead of transgenerational inheritance. Therefore, we fed P0 animals with HFD from L1 to L4 larvae, then transferred them to nematode growth medium (NGM) plates with normal OP50 to prevent F2 primordial germ cells from exposure. We found that pre-exposure of P0 worms to HFD until L4 larvae could also induce the lipid accumulation of their naive F1 and F2 progeny (Fig. 1e), suggesting that HFD can induce a TEI obesity phenotype. Notably, animal populations in these assays were synchronized using L1 diapause. One concern is that the synchronization procedure might cause starvation in L1, which would influence the metabolism of animals and affect transgenerational gene expression 23 . Therefore, to ease our concern, we compared the TEI effects of worms with or without synchronization to find that synchronization did not affect TEI in our settings (Fig. 1h).
Different consecutive generations of exposure to the same environmental stress can enhance animals' adaptation to stress for different generations of descendants 15,24,25 . In our heritable obesity effect model, we found that F2 populations exhibited a measurable but minor lipid accumulation phenotype. In order to find if a longer exposure might enhance F2s' lipid accumulation, we exposed multiple consecutive generations of animals to HFD. We found that the exposure of four consecutive generations of animals (P0, F1, F2, and F3), three (F1, F2, and F3), two (F2, and F3), or one (F3), did not affect lipid accumulation of their recovered F4 or F5 descendants ( Fig. 1f-g and Supplementary  Fig. 1f-g). In addition, we observed that lipid accumulation induced by OA or PA can also transmit to their naive progeny ( Supplementary Fig. 1a-c). Altogether, these results demonstrated that exposure of HFD induces a transgenerational obesity effect.
We next wanted to determine how long a parent must be exposed to HFD to transmit lipid accumulation signal to progeny. For this, we fed P0 animals with HFD at the L1 stage for 12 or 24 h, and then transferred them to NGM plates with normal food. The results showed that the ORO levels were significantly increased in HFD-fed parents and their recovered F1 progeny, although to a lesser extent than worms from parents fed with HFD for longer periods, but not F2 progeny ( Supplementary  Fig. 1d, e), suggesting that parents need to be fed HFD for enough long time before they can show a TEI phenotype. Transgenerational information has been shown to be passed down through both male and female germline 17 . To determine the contributions of the sperm and oocytes in transmitting information of lipid accumulation, we conducted reciprocal mattings and analyzed lipid levels of cross progeny (Fig. 1i). Our results showed that both females fed with HFD and males fed with HFD transmitted information of lipid accumulation to the next generation, suggesting a role of both sperm and oocytes in the transgenerational obesity effect.
It is reported that lipid metabolism and chromatin modifications are associated with the regulation of lifespan [26][27][28] . In our settings, we found that animals fed with egg yolk increased lipid accumulation. We wondered whether egg yolk affected the lifespan of C. elegans. Lifespan analyses revealed no difference between worms fed with HFD and OP50, suggesting that egg yolk did not affect the survival of C. elegans ( Supplementary  Fig. 2).
HFD-induced transgenerational inheritance is mediated by nuclear receptors NHR-49, NHR-80, and transcription factors SBP-1 and DAF-16. To dissect the molecular mechanisms underlying HFD-induced TEI, we evaluated the contribution of several nuclear receptors and transcription factors that play essential roles in lipid metabolism in C. elegans. SBP-1, a sterol regulatory element-...

***
binding protein, is a crucial transcription factor governing fat metabolism 21 . NHR-49, a functional homolog of mammalian peroxisome proliferator-activated receptor alpha, and NHR-80, a homolog of mammalian hepatocyte nuclear factor 4, are both important nuclear hormone receptors involved in the control of fat consumption and fatty acid composition in C. elegans [29][30][31][32] . Forkhead transcriptional factor DAF-16/FOXO, a central downstream effector of the insulin/insulin-like growth factor signaling pathway, is a critical and well-conserved metabolic regulator 33,34 . Our results demonstrated that the elevated fat level induced by HFD in P0 was not affected in a sbp-1 loss-of-function (lof) mutant ( Fig. 2a), an nhr-80 lof mutant ( Fig. 2b) or an nhr-49 lof mutant (Fig. 2c). However, lipid accumulation was abrogated in F1 or F2 descendants of these mutants. By contrast, the elevated lipid phenotype induced by HFD was abolished in P0 parents as well as in their recovery F1 and F2 progeny in the daf-16 lof mutant (Fig. 2d). These results indicated that SBP-1, NHR-49 and NHR-80 work in the F1 generation to mediate TEI of fat accumulation induced by HFD; whereas DAF-16 not only acts in the F1 generation, but also functions for the execution of the lipid metabolism response in the P0 generation. To characterize the relationship of DAF-16, NHR-49, NHR-80 and SBP-1 in F1 generation, we performed epistasis analyses. We found that overexpressing DAF-16 in nhr-49, nhr-80 and sbp-1 mutant could not suppress the phenotype of nhr-49, nhr-80 and sbp-1 lof mutant ( Supplementary Fig. 3a-c). Here, we applied sbp-1 RNAi in the background of overexpressing DAF-16 to analyze lipid level as we failed to obtain daf-16 (OE); sbp-1 (lof) double mutant by numerous matings. These results demonstrated that daf-16, sbp-1, nhr-49 and nhr-80 function in parallel pathway during the stress of lipid accumulation in F1 generation. As daf-16 is responsible for HFD-induced lipid accumulation in P0 generation, to confirm the role of daf-16 in the recovered F1, we used RNAi to knock down daf-16 exclusively in F1 derived from P0 fed with HFD. We found that specific knockdown of daf-16 in F1 also abrogated the elevated fat level (Fig. 2e, f). Moreover, we also performed similar analyses using nhr-49, nhr-80 and sbp-1 RNAi. We found that silencing nhr-49, nhr-80 and sbp-1 in F1 progeny also abolished their lipid accumulation ( Supplementary  Fig. 4a-e), suggesting that daf-16, nhr-49, nhr-80 and sbp-1 act in the progeny to increase lipid level.
As a key regulator, DAF-16 responds to environmental stress by activating a series of target genes 34 . We found that the mRNA level of the target genes of DAF-16 (sod-3 and dod-3) was significantly upregulated in P0 fed with HFD or their naive F1 progeny ( Supplementary Fig. 5c). The expression of SOD-3::GFP was consistent with the mRNA level of sod-3 ( Supplementary  Fig. 5a, b). Altogether, these results indicated that DAF-16 functioned in response to alter lipid metabolism induced by HFD in P0 and their recovered progeny.
To more definitively dissect whether the role of daf-16, nhr-49, nhr-80 and sbp-1 is to implement regulation of lipid level (an "executor"), to transmit the heritable memories (a "transmitter"), or both, we used RNAi to silence daf-16, nhr-49, nhr-80 and sbp-1 exclusively in the P0 generation, and then analyzed the fat level of F1 generation raised on OP50 (Fig. 2g). As reported in our recent study 17 , if a gene is a transmitter, the silencing of that gene in the P0 generation will prevent the transmission of transgenerational memory, which will result in the elimination of the lipid accumulation of the F1 generation. Alternatively, if a gene only functions as an executor, the silencing of that gene in the P0 generation should only influence the lipid level of P0 generation, so that we would still detect the elevation of fat level in F1 generation. Altogether, the results showed that silencing of nhr-49 and nhr-80 did not abrogate the lipid accumulation of F1 progeny (Fig. 2i, j), suggesting that nhr-49 and nhr-80 are purely executors but not transmitters; the silencing of daf-16 or sbp-1 in P0 generation caused the loss of lipid accumulation of F1 (Fig. 2h, k), indicating their roles as transmitters.
Taken together, our results demonstrated that daf-16, nhr-49, nhr-80 and sbp-1 were required for HFD-induced TEI of lipid accumulation. Among them, nhr-49 and nhr-80 functioned solely as executors; sbp-1 was responsible for transmitting the heritable memories, though we could not rule out its role as an executor; for daf-16, it not only functions as an executor to regulate lipid accumulation but also as a transmitter to pass down heritable memory to progeny.
Delta-9 desaturases mediated the heritable memories. In C. elegans, delta-9 desaturases are well-characterized targets downstream of highly conserved transcription factors SBP-1, NHR-49, NHR-80 and DAF-16 ( Fig. 3a) 23,31,35,36 . Because SBP-1, NHR-49, NHR-80 and DAF-16 have been detected to mediate TEI of lipid accumulation in our settings, we wondered whether delta-9 desaturases participated in the TEI of lipid accumulation. We used the delta-9 desaturases-related mutants including fat-5, fat-6, and fat-7 lof mutants, and fat-5; fat-6 and fat-5; fat-7 double mutants to test our hypothesis. Our results showed that the fat accumulation phenotype remained largely unaffected in the P0 generation of these mutants; however, these mutations abrogated the elevated fat level in F1 and F2 progeny ( Fig. 3b-f). These results revealed that delta-9 desaturases function in TEI induced by HFD. To more definitively specify the role of delta-9 desaturases in HFD-induced TEI, we conducted RNAi experiments similar to the above-mentioned treatments (daf-16, nhr-49, nhr-80 and sbp-1). We found that silencing fat-5 in F1 abolished the lipid accumulation (Fig. 3g), and silencing fat-5, fat-6 or fat-7 in P0 generation did not affect the phenotype of fat accumulation in their recovered F1 progeny ( Fig. 3h and Supplementary  Fig. 4f-h), suggesting that delta-9 desaturases solely execute regulation of lipid metabolism in HFD-induced TEI, but not transmit heritable memory. a Quantification of ORO staining of wild-type (N2) treated with different lipid foods (n ≥ 28 worms per condition). PA palmitoleic acid, OA oleic acid. b Experimental scheme. c, d Transgenerational inheritance of lipid accumulation induced by HFD (egg yolk) in wild-type C. elegans (n ≥ 28 per condition), scale bar = 100 μm. e Transgenerational inheritance of lipid accumulation induced by HFD in wildtype C. elegans. Worms were fed with HFD from L1 to L4 larvae before they were transferred to normal food until day 1 adulthood. Embryos collected by bleaching to obtain F1 animals that were fed with OP50. F2-Fn generations were obtained using the same protocol (n ≥ 34 per condition). f, g Quantification of ORO staining of F4 or F5 wild-type (N2) from parents exposed to HFD four (P0, F1, F2 and F3), three (F1, F2 and F3), two (F2 and F3) and one (F3) generation of animals (n ≥ 28 per condition). h Quantification of ORO staining of F1 or F2 from parents fed with HFD and without synchronizationcaused starvation (n ≥ 30 per condition). i Lipid accumulation information induced by HFD could be transmitted through both male and female germlines. Quantification of ORO staining of recovered F1 progeny of fathers fed with HFD or mothers fed with HFD (n ≥ 30). For a and d-i, graph data are presented as mean ± SD (arb. units: arbitrary units), statistical analyses were performed by unpaired two-tailed Student's t-test and ANOVA analysis; ***p < 0.001 and *p < 0.05. Source data are provided as a Source Data file. ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-28469-4 Memory of HFD-induced lipid metabolism was conferred by histone H3K4me3 modification. Reportedly, a set of chromatinmodifying enzymes, including deacetylases and H3K4me3 methyltransferase, influence lipid metabolism 27,37 . To understand whether methyltransferase mediated HFD-induced TEI of obesogenic effect, we performed a targeted mutant screening by selecting genes involved in histone methylation. We found that loss of wdr-5.1, a H3K4me3 complex component, but not other H3K4me3 modifiers (i.e., set-2 or ash-2), abolished the lipid accumulation in progeny ( Fig. 4a and Supplementary Fig. 6a, b). Furthermore, wdr-5.1 did not execute in responding to lipid metabolism induced by HFD, because ORO signaling of the wdr-5.1 mutant remained mostly increased in the P0 parents (Fig. 4a). Moreover, we detected that the mRNA level of wdr-5.  Fig. 6c, d). Furthermore, we observed that H3K4me3 methylation was significantly increased in the P0 generation fed with HFD ( Fig. 4b) and in their recovered progeny ( Fig. 4b-d and Supplementary Fig. 6e, f). Moreover, upregulation of H3K4me3 methylation induced by feeding with HFD was abolished in a wdr-5.1 mutant (Supplementary Fig. 6i). These results suggested that histone H3K4me3 modification, and only histone H3K4me3 modification conferred by wdr-5.1, plays a specific role in transmitting the heritable memory induced by HFD.
To more clearly characterize the role of wdr-5.1 in HFDinduced TEI, we also conducted two sets of RNAi experiments similar to the above-mentioned treatments (daf-16, nhr-49, nhr-80 and sbp-1). We found that silencing wdr-5.1 in F1 abrogated the lipid accumulation ( Supplementary Fig. 7a), and silencing wdr-5.1 in P0 generation also abolished the phenotype of fat accumulation in their recovered F1 progeny ( Supplementary  Fig. 7b), suggesting the role of wdr-5.1 as a transmitter instead of an executor. Collectively, these results suggested that histone H3K4me3 modification plays a specific role in transmitting the heritable memory induced by HFD.
To test whether lipid accumulation induced by HFD is associated with heritable changes of H3K4me3 at specific loci, we conducted RNA sequencing (RNA-seq) and chromatin immunoprecipitation and sequencing (ChIP-seq) to compare changes in the enrichment of H3K4me3 peaks. The results exhibited that when animals fed with HFD, the transcription level and H3K4me3 level of many genes, especially lipid metabolism-associated genes (i.e., daf-16, sbp-1 and gmd-2), was upregulated ( Supplementary Fig. 7c-i). We used ChIP-qPCR and found that H3K4me3 occupancy was significantly increased at the promoter of these genes related to metabolism in the P0, F1 and F2 generations (Fig. 4f). The mRNA level of these genes was also significantly upregulated in the P0, F1 and F2 generations by analysis of RT-qPCR (Fig. 4g).
We further investigated whether the transcription factors (sbp-1 and daf-16) and nuclear receptors (nhr-49 and nhr-80), which have been identified to mediate HFD-induced TEI, also regulated the levels of histone H3K4me3 modification. We found that all mutations (daf-16, nhr-49 and nhr-80) except sbp-1 had no effect on H3K4me3 modification in animals fed with HFD when compared with controls ( Fig. 4e and Supplementary Fig. 8a-d). In addition, the elevated transcription level of wdr-5.1 induced by HFD was abrogated in the sbp-1 mutant (Supplementary Fig. 6g). These findings suggested that sbp-1 might mediate HFD-induced memory transmission of lipid accumulation through upregulating wdr-5.1 expression to increase the H3K4me3 level.
Characterization of tissue-specific functions of daf-16 and sbp-1 in HFD-induced TEI. To detect tissues in which daf-16 and sbp-1 function in response to lipid metabolism and in regulating histone modifications, we conducted tissue-specific gene knockdown assays by using strains that can process RNAi efficiently only in specific tissues, such as germline, muscle, or neuron (Fig. 5a). Firstly, we performed the tissue-specific RNAi of daf-16 to detect the ORO staining. We found that muscle-specific RNAi of daf-16 in the P0 generation led to the suppression of lipid accumulation (Fig. 5g). However, other tissue-specific RNAi, including in the intestine, neuron or germline, could not compromise the elevated fat level (Fig. 5h-k). It is interesting that DAF-16 plays a role in the muscle to regulate HFD-induced lipid accumulation which mainly occurs in the intestine. We wondered how DAF-16 senses the signal. Could it be the components of the insulin signaling pathway upstream of DAF-16 that play a role to transmit signals from the intestine to the muscle? To determine this hypothesis, we performed the lipid metabolism-associated germline or intestine-specific RNAi of daf-2 (an insulin receptorlike gene) to detect the ORO staining. Our results exhibited that both germline and intestine-specific RNAi of daf-2 did not abrogate lipid accumulation induced by HFD ( Supplementary  Fig. 9d-f), suggesting that DAF-16 senses other signals instead of the signals transmitted by daf-2 to regulate lipid metabolism.
It is well known that DAF-16 executes a response to environmental stress by translocating to the nucleus 39 . Indeed, we found that DAF-16 displayed obvious aggregation in the nucleus, especially in muscle, in the P0 generation fed with HFD and in their recovery F1 progeny (Supplementary Fig. 9a). Moreover, we found that protein levels of DAF-16 were obviously elevated in P0 worms fed with HFD and their recovered F1 progeny ( Supplementary Fig. 9b, c).
We then performed tissue-specific knockdown of sbp-1 to assess the change of histone H3K4me3 induced by HFD. We found that intestine-specific, muscle-specific, germline-specific, or intestine-and germline-specific RNAi of sbp-1 abolishes the elevated H3K4me3 modification levels (Fig. 5b-f). These results indicated that sbp-1 functions in all detected tissues to regulate HFD-induced H3K4me3 modification. Similar to daf-16, we also conducted the tissue-specific RNAi of sbp-1 to detect the ORO staining. Consistent with the results from sbp-1 lof mutant, knockdown of sbp-1 in different tissues could not abrogate lipid accumulation induced by HFD ( Supplementary Fig. 10a-e). Collectively, tissue-specific RNAi studies demonstrated that daf-16 and sbp-1 function in one or more tissues, suggesting that signals from different tissues might coordinately regulate TEI of lipid accumulation induced by HFD. Fig. 2 NHR-49, NHR-80, SBP-1, and DAF-16 contribute to transgenerational inheritance of lipid accumulation. Transgenerational inheritance test of lipid accumulation induced by HFD in sbp-1 mutant (n ≥ 27 per condition) (a), nhr-80 mutant (n ≥ 29 per condition) (b), nhr-49 mutant (n ≥ 29 per condition) (c), and daf-16 mutant (n ≥ 29 per condition) (d). For a-d, graph data are presented as mean ± SD (arb. units: arbitrary units); unpaired two-tailed Student's t-test; ***p < 0.001, ns: not significant. The same WT control was shown in a-d (for a-d, performing ORO staining analyses at the same time and using the same WT control). e, f P0 worms were fed with or without HFD and without any RNAi exposure, and then F1 progeny were exposed to RNAi with daf-16 to test the requirement of daf-16 in F1s. Quantification of ORO staining of F1 worms (mean ± SD; n ≥ 30 per condition; unpaired two-tailed Student's t-test; ***p < 0.001, and ns: not significant). g-k DAF-16 and SBP-1, but not NHR-49 or NHR-80, contribute to transmit the transgenerational information of lipid accumulation. g Experimental scheme. h-k P0 worms were fed with or without HFD, and subjected to sbp-1 RNAi (n ≥ 31 per condition) (h), nhr-80 RNAi (n ≥ 30 per condition) (i), nhr-49 RNAi (n ≥ 30 per condition) (j) and daf-16 RNAi (n ≥ 27 per condition) (k), then F1 generation raised at OP50. Quantification of ORO staining of F1 worms (mean ± SD; unpaired two-tailed Student's t-test; ***p < 0.001, **p < 0.01, and ns: not significant). For a-d and f, scale bar = 150 μm. Source data are provided as a Source Data file.

Discussion
During the last several decades, the number of obese people has been growing rapidly in many countries, causing a huge economic burden to society 40 . However, for metabolic diseases such as obesity and T2D, besides genetic and environmental factors, family history has increasingly been considered to be an important factor. In fact, genetic background only contributes~5-10% of the overall risk for developing T2D 5 . Recently, considerable efforts have been done to elucidate the TEI-associated obesity induced by overnutrition. For example, studies in Drosophila have shown that obese male flies fed with a high-sugar diet can transgenerationally transmit lipid accumulation information through epigenetic markers (histone methylation modification) to their offspring and induce obesogenic phenotype 41 . Other studies  have shown that obese mice fed with HFD can transmit lipid accumulation information to their offspring and increase adiposity, which is conferred by changes in sperm miRNA and reduction in total global germ cell DNA methylation 42,43 . However, many mechanisms remain uncovered. In this work, we use C. elegans as an obesogenic model by feeding them HFD (egg yolk). Our results demonstrated that an obesity effect in parents in C. elegans could be induced by feeding egg yolk, and could be transmitted to their naive progeny that had never been exposed to HFD. In addition, our study revealed that TEI of lipid accumulation was mediated by histone H3K4me3 modification. Consistent with previous reports 34, 35 , we found that DAF-16 is an essential regulator in response to change in lipid metabolism induced by HFD. Furthermore, we observed that sbp-1, daf-16, nhr-49, nhr-80, and delta-9 fatty acid desaturase genes (fat-5, fat-6, and fat-7) were responsible for TEI of obesogenic effects. Among them, nhr-49 and nhr-80 contributed to regulating lipid metabolism; sbp-1 transmitted transgenerational memory; while daf-16 acted in both regulation of lipid metabolism and transmission of obesogenic information. Strikingly, sbp-1 itself also contributed to regulating the level of histone H3K4me3 modification. Consistent with our observations, a previous study has discovered that in C. elegans, a complex transcriptional network consisting of SBP-1, MDT-15, DAF-16, NHR-49, and NHR-80 regulates the activity of the delta-9 fatty acid desaturase genes (fat-5, fat-6, and fat-7). In turn, these factors affect the modification of H3K4me3, ultimately affecting the lifespan and lipid metabolism of the nematodes 27 .
Taken together, these results suggest that, upon animals fed with HFD, the stress of lipid accumulation in the parental generation induces the activity of lipid metabolic transcription factors sbp-1, daf-16, nhr-49, and nhr-80. At the same time, sbp-1 regulates histone H3K4me3 modification and establishes the epigenetic marks in descendants. In turn, the H3K4me3 marks in the progeny promote the recruitment of lipid metabolism-related genes (i.e., sbp-1 and daf-16) and facilitate their activation; simultaneously, the activation of daf-16 and sbp-1 might recruit nhr-49 or nhr-80 to form complex, and then daf-16, sbp-1, nhr-49 and nhr-80 synergistically induce the expression of lipid metabolism-related genes (i.e., fat-5, fat-6 and fat-7) to respond lipid metabolism and ultimately reset the metabolic processes, thereby completing the TEI of obesity effect (Fig. 6).
In this study, egg yolk was fed as HFD to induce an obesity phenotype in C. elegans. We found that, except for daf-16, loss of function of other genes, including nhr-49, nhr-80, and sbp-1, which are recognized as central regulators of lipid metabolism, could not completely block lipid accumulation induced by HFD in the P0 generation. A possible explanation is that the composition of the egg yolk is complex so that these genes only partially mediate the egg yolk-induced lipid accumulation. Because our work focused on the transgenerational inheritance of lipid accumulation, we have not yet fully elucidated the underlying details of molecular mechanisms that regulate lipid metabolism changes induced by egg yolk. This will be a subject of future work. Because egg yolk is used as HFD to induce the phenotype of lipid aggregation, it is worth noting that the executors for lipid accumulation in the P0 and F1/F2 are different since they are fed with a different diet. Therefore, in our present study, for P0 generation, only daf-16, but not nhr-49, nhr-80, sbp-1, fat-5, fat-6 and fat-7 was necessary as an executor to respond to lipid metabolism induced by egg yolk; while, for F1 generation, all of them (including daf-16, nhr-49, nhr-80, sbp-1, fat-5, fat-6 and fat-7) are required as executors to respond to lipid accumulation.
In addition, it is still unclear how sbp-1 regulates the level of histone H3K4me3 modification in this work. Epigenetic mechanisms include chromatin structure and modification, DNA methylation and noncoding RNAs. In this study, we identified histone H3K4me3 modification as a contributing factor to the TEI of obesity effect; however, the function, if any, of DNA methylation and noncoding RNAs remains ambiguous. Therefore, more work is needed to further disentangle the molecular mechanisms of TEI related to obesity, including how and when the epigenetic memory is established, as well as how lipid metabolism genes interplay with genes involved in epigenetic inheritance. Our research provides evidence that the lifestyle of parents plays an important role in the metabolic status of offspring in nematodes, and it is tempting to speculate that these findings may be relevant to mammalian systems.
Mating protocol. Mating experiments were conducted with a GFP + strain MT18143 to distinguish the cross progeny from the self-fertilized progeny. Briefly, synchronized L1 male (GFP + ) or female (wild-type N2) worms were fed with HFD. Afterwards, worms (a ratio of males and females of 3:1) were picked onto OP50 seeded NGM plates to mate for 12 h at 20°C. Two days after mating, GFP + F1 young adult worms were picked and collected to perform ORO staining.
RNA interference. RNAi was essentially conducted using the standard feeding protocol. The E. coli HT115 transformed with vectors expressing the corresponding dsRNA and empty vector were derived from the Ahringer library (Source Bioscience, Nottingham, UK). All vectors were confirmed by sequencing. RNAi bacteria were cultured at 37°C in LB with 100 μg mL -1 ampicillin. Freshly prepared bacteria were spotted on NGM plates with 1 mM isopropyl-B-D-thiogalactoside before use. All RNAi treatments began from synchronized L1 larvae. Fig. 3 Transgenerational inheritance of lipid accumulation depends on fat-5, fat-6 and fat-7. a Schematic presentation of the transcriptional network that regulates delta-9 desaturase gene. b-f Transgenerational inheritance test of lipid accumulation induced by HFD in fat-5 mutant (n ≥ 28 per condition) (b), fat-6 mutant (n ≥ 28 per condition) (c), fat-7 mutant (n ≥ 28 per condition) (d), fat-5; fat-6 double mutant (n ≥ 28 per condition) (e), and fat-5; fat-7 double mutant (n ≥ 30 per condition) (f) (mean ± SD; unpaired two-tailed Student's t-test; ***p < 0.001, and ns: not significant). The same WT control was shown in a-f. g P0 worms were fed with or without HFD, and F1 progeny were exposed to fat-5 RNAi to determine the requirement of fat-5 in the F1 generation. h P0 animals were fed with or without HFD and subjected to fat-5 RNAi, and then lipid levels of F1 progeny were analyzed. For g and h (right panels), graph data are presented as mean ± SD, n ≥ 31 per condition, statistical analyses were performed by unpaired two-tailed Student's t-test; ***p < 0.001, **p < 0.01, and ns: not significant, Source data are provided as a Source Data file.
Fatty acid supplementation. All fatty acids and cholesterol used in this study were purchased from Sigma-Aldrich (Munich, Germany). For the fatty acid supplementation, a final concentration of 4 mM OA, 4 mM PA and 50 mg/mL cholesterol were used per experiment. OP50 were mixed with corresponding fatty acid and then seeded onto plates at room temperature for 6 h before the addition of worms. For egg yolk supplementation, the cooked random egg yolk was ground and mixed with OP50 (volume ratio, 1:5) to provide the C. elegans a food supplement. The mixture of egg yolk and bacteria is freshly made before each experiment.
Conditions for inducing lipid accumulation. To induce lipid accumulation, 500 synchronized L1 larvae were raised on NGM plates with or without HFD. To obtain F1 generation worms, gravid day 1 adults treated with or without HFD were bleached and hatched overnight in 10 mL M9 buffer, and subsequently grown on NGM plates without HFD. This process was repeated to obtain the Fn generations. For worms fed with HFD until L4 larvae, synchronized L1 larvae were grown on NGM plates with HFD until L4 larvae, then transferred them to NGM plates with normal food, Fn generations animals were obtained using the similar protocol mentioned above.

wdr-5.1 (ok1417)
Oil-Red-O staining and quantification. For ORO staining and quantification, all strains were cultured as described under conditions to accumulate lipid to obtain P0, F1 and F2 animals. Then animals were fixed in fixing solution (1× MRWB buffer (160 mM KCl, 40 mM NaCl, 14 mM Na 2 EGTA, 1 mM spermidine-HCl, 0.4 mM spermine, 30 mM Na-PIPES, 0.2% β-mercaptoethanol) containing 1% paraformaldehyde) by gently rocking for 1 h at room temperature. Fixed worms were washed with 1×PBS, resuspended in 60% isopropanol and incubated for 15 min at room temperature. Worms were then incubated and stained using saturated ORO staining solution dissolved in 60% isopropanol overnight with rocking. Afterwards, worms were washed twice with 1× PBS 0.01% Triton X-100 to move excessive dye. Then animals were mounted to agar plates for imaging at ×200 magnification with a Nikon Ti2-U microscope. The same exposure setting was used across all conditions within each experiment. The ORO intensity (arbitrary unit, arb. units) was measured using ImageJ processing software as previously described after background was removed and greyscale converted 45 . Statistical significance was analyzed using unpaired two-tailed Student's t-test by GraphPad Prism (v 8.0.1). At least 30 worms were used per experiment, and the experiment was repeated at least three times.
Quantitative RT-PCR. Briefly, about 4000 worms were used to extract total RNA with RNAiso Plus (Takara). Then RNA was reverse transcribed into cDNA using a cDNA Reverse Transcription Kit (ABclonal) as per the manufacturer's instructions. Real-time qPCR was conducted in triplicate for each gene using a CFX96 Real-Time PCR system (Bio-Rad) with SYBR Green select master mix (ABclonal). The quantification was computed by 2 -ΔΔCt method after being normalized to gene cdc-42. In order to rule out the possibility of cdc-42 expression being affected by HFD, we analyzed the level of cdc-42 using other reliable reference genes when animals fed with HFD compared with normal OP50. The primers are summarized in Supplementary Table 1. RNA-seq. RNA-seq libraries were prepared and sequenced by Novogene Corporation (Beijing, China). Briefly, young adult synchronized worms were harvested with cold PBS buffer. Total RNA was extracted using RNAiso Plus (Takara). RNA integrity was assessed using the RNA Nano 6000 Assay Kit of the Bioanalyzer 2100 system (Agilent Technologies, CA, USA). A total amount of 1 μg RNA per sample was used as library and sequence. The sequence was performed by cBot Cluster Generation System using TruSeq PE Cluster Kit v3-cBot-HS (Illumia) according to the manufacturer's instructions. Differential expression analysis was performed using the DESeq2 R package (1.20.0). Genes with a greater than twofold change and false discovery rate <0.01 were defined as differentially expressed genes (DEG). Functional annotation of DEG was analyzed through various databases, including Gene Ontology (GO), Kyoto Encyclopedia of Gene and Genomes (KEGG) and EuKaryotic Ortologous Groups. HFD-induced DEG were assigned functional categories using the Database for Annotation, Visualization and Integrated Discovery.
H3K4me3 ChIP-seq. ChIP-seq libraries were prepared and sequenced by Novogene Corporation (Beijing, China). Briefly, 3 g of C. elegans were harvested and washed with cold PBS buffer, then crosslinked with 1% formaldehyde in M9 buffer. After lysing the sample to obtain chromatins, chromatins were sonicated to get soluble sheared chromatin (average DNA length of 200-500 bp). The soluble chromatin was used for immunoprecipitation by H3K4me3 antibody (5 μL per sample) (Abcam ab8580) and 5% of it was saved as input. DNA purity was checked using the NanoPhotometer® spectrophotometer (IMPLEN, CA, USA). The purified DNA was used for ChIP-seq library preparation and sequence on Illumina platform (Illumina, CA, USA). Library quality was assessed on the Agilent Bioanalyzer 2100 system. Raw data (raw reads) of fastq format were firstly processed using fastp (v 0.19.11) software 47 . Clean data (clean reads) were obtained by removing reads containing adapter and ploy-N, as well as low-quality reads. Index of the reference genome was built using BWA (v 0.7.12) and clean reads were aligned to the reference genome using BWA mem (v 0.7.12). Afterwards, regions of IP enrichment over background were identified by the MACS2 (v 2.1.0) peak calling software (q-value threshold of 0.05 was used for all data sets). Peak-related genes were confirmed by ChIPseeker (v 3.14). And then GO enrichment analysis was performed by the GOseq R package. KEGG analysis was conducted by KOBAS software. The heatmaps, reads distribution of diffpeak, and all peaks were obtained using Deeptool (v 3.2.1).
Fluorescence microscopic imaging. For quantification of SOD-3, synchronized L1 larvae of CF1553 (SOD-3p::GFP) were treated as described above to induce accumulation of lipid. Young adult worms were mounted to agar plates for imaging using a Nikon Ti2-U fluorescence microscope with ×20 air objectives. To exhibit the fluorescence intensity, six to seven randomly selected worms were put together to take images. For quantification purpose, each worm was imaged individually. The GFP fluorescence intensity was quantified by mean fluorescence intensity (total intensity/area) using ImageJ software. At least 30 animals were used per experiment. Statistical analyses were performed using unpaired two-tailed Student's t-test.
Lifespan assay. Lifespan assays were performed at 20°C, based on standard protocols, as previously described 17 . In brief,~100 young adults fed with HFD from L1 stage were transferred to NGM plates with 10 μM 5-fluoro-2'-deoxyuridine (Sigma) seeded with heat-inactivated OP50 to perform survival analyses. This day was defined as day 1. Death events were scored daily. The experiments were repeated at least twice. The mean, SEM, p and lifespan values were processed using the Kaplan-Meir survival analysis in SPSS software (v 26).