Animals respond to stress by activating specialized protective gene expression programs (de Nadal et al., 2011). While these programs depend on specific signaling and transcriptional activators, they may also be impacted by histone methylation and the production of SAM. For example, we found that C. elegans lacking sams-1 die rapidly on pathogenic bacteria, have low global H3K4me3 and fail to upregulate immune response genes (Ding et al., 2015). In contrast, heat shocked animals survive better without sams-1 (Ding et al., 2018). sams-1(RNAi) animals induced heat shock genes to normal levels and acquired additional changes in the transcriptome, including downregulation of many metabolic genes. However, the H3K4me3 methyltransferase set-16/MLL was essential for survival (Ding et al., 2018), suggesting that methylation was required. We hypothesized that other SAM synthases could play an important role in mediating survival during heat shock (Figure 1A).

In order to test these hypotheses, we first compared expression of each synthase, SAM levels and gene expression after RNAi in adult unstressed animals. ModEncode data Gerstein et al., 2010 from young adult animals shows that in young adult levels, sams-1 is expressed at the highest levels, comparable to the metabolic enzyme GAPDH (gpdh-1) (Figure 1, Figure 1—figure supplement 1A). sams-3 and sams-4 are expressed at lower levels, but comparable to other enzymes of the 1-Carbon cycle such as metr-1, whereas sams-5 is minimally expressed (Figure 1—figure supplement 1A). In order to determine the tissue-specific patterns of the SAM synthases expressed in adult animals, we obtained strains where each protein was tagged with RFP, GFP or mKate, via CRISPR (Figure 1B, Figure 1—figure supplement 1B, C). RFP::sams-1 and GFP::sams-4 animals were also crossed to allow visualize expression of both synthases (Figure 1B). RFP::SAMS-1fluorescence was evident in much of the adult animal, including intestine, hypodermis and cells in the head (Figure 1B, Figure 1—figure supplement 1B), in line with mRNA expression patterns derived from tissue-specific RNA seq (Kaletsky et al., 2018). However, RFP::SAMS-1was not present in the germline, which did express GFP::SAMS-4 and SAMS-3::mKate (Figure 1B, Figure 1—figure supplement 1C). GFP::SAMS-4 and SAMS-3::mKate was also present in intestinal and hypodermal cells (Figure 1B, Figure 1—figure supplement 1C), demonstrating that these tissues, which are major contributors to the stress response (McGhee, 2007) contain each of these SAM synthases. sams-3 and sams-4 are expressed bidirectionally from the same promoter and share 95% sequence identity at the nucleotide level thus RNAi targeting is likely to affect both genes. Indeed SAMS-3::mKate and GFP::SAMS-4 were reduced after either RNAi (Figure 1—figure supplement 1C). Next, we used mass spectrometry to compare SAM levels after sams-3 and sams-4 RNAi and found that like sams-1 (Ding et al., 2015; Walker et al., 2011), reduction in any synthase significantly reduced but did not eliminate SAM (Figure 1—figure supplement 1D).

In order to compare gene expression after RNA of each SAM synthase in basal conditions, we used RNA sequencing (RNAseq). Principal component analysis showed that sams-1(RNAi) and sams-5 formed distinct clusters on the first two principal components; however, sams-3 and sams-4 were overlapping (Figure 1—figure supplement 2A; Supplementary file 1: Tabs A-C). About half of the genes upregulated after sams-4 knockdown also increased in sams-1(RNAi) animals (Figure 1—figure supplement 2B). To determine if genes related to distinct biological processes were present, we compared genes upregulated after sams-1 RNAi (Ding et al., 2018) with those changing in sams-4 RNAi with WormCat (Holdorf et al., 2020), which provides enrichment scores for three category levels (Cat1, Cat2, Cat3) for broad to more specific comparisons. WormCat finds that gene function categories at the Cat1 and Cat 2 level, such as METABOLISM: Lipid (Figure 1—figure supplement 2C) or STRESS RESPONSE: Pathogen (Figure 1—figure supplement 2D–F), are enriched at lower levels and contain different genes in sams-4(RNAi) animals (Supplementary file 1: Tabs D-F). Notably, fat-7 and other lipid synthesis genes that respond to low PC in sams-1 animals are not upregulated after sams-4(RNAi) (Supplementary file 1:Tab:B). These findings strengthen the idea that these SAM synthases could have distinct functions.

In order to determine if other SAM synthases expressed in adult animals contributed to survival in heat shock (Figure 1—figure supplement 3A), we compared the heat shock survival phenotypes of C. elegans with deletions in sams-1, sams-3 and sams-4 to avoid effects of co-targeting by RNAi. sams-1(ok3033) has a deletion covering the majority of the open reading frame and extracts from these animals lack SAMS-1 protein in immunoblots (Ding et al., 2015); therefore, we refer to this allele as sams-1(lof). sams-4(ok3315) animals have a deletion that removes around a third of the open-reading frame. Strikingly, sams-4(ok3315) mutants had the opposite phenotype from sams-1(lof), and died rapidly after heat shock (Figure 1C and D, Supplementary file 2:Tabs B, C). sams-3(2932) harbors a deletion removing most of the ORF, but in contrast to sams-4 and sams-1, is indistinguishable from wild type animals in a heat shock response (Figure 1—figure supplement 3B). Although sams-3 may be co-targeted in RNAi experiments, we will refer solely to sams-4 in our discussion because it has the most direct link to the heat shock phenotypes. Finally, sams-4(RNAi) phenotypes in the heat stress response were not linked to a general failure to thrive, as sams-4(RNAi) animals under basal conditions had modestly enhanced lifespan (Figure 1—figure supplement 3C; Supplementary file 2: Tab A).

Next, we used immunostaining to compare global levels of H3K4me3 in sams-1 and sams-4 RNAi nuclei during heat shock. In contrast to the reduction in H3K4me3 in basal conditions in sams-1(lof), sams-4(ok 3315) or RNAi animals (Figure 1E–F–), we detected robust levels of H3K4me3 in sams-1(lof) nuclei after heat shock (2 hr at 37 °C) (Figure 1E and H), suggesting that sams-1-independent mechanisms act on H3K4me3 during heat shock. These increases in H3K4me3 did not appear in heat shocked sams-4(RNAi) intestinal nuclei Figure 1F, I, raising the possibility that sams-4 contributed to the effects in sams-1(lof) animals. Next, we wanted to test effects of reducing both sams-1 and sams-4 levels on H3K4me3 during heat shock. Loss of multiple SAM synthases reduces viability in C. elegansTowbin et al., 2012. In order to circumvent this, we used dietary choline to rescue PC synthesis and growth of sams-1(RNAi) or (lof) animals during development (Ding et al., 2015; Walker et al., 2011). sams-1(lof); sams-4(RNAi) animals were raised on choline until the L4 stage, then moved to normal media for 16 hr before heat shock. Immunostaining of sams-1(lof); sams-4(RNAi) intestines showed that sams-4 is required for the H3K4me3 in heat shocked sams-1(lof) nuclei (Figure 1G and J). These results were identical when we used RNAi to reduce sams-1 in sams-4(ok3315) animals (Figure 1—figure supplement 3D). We also asked if sams-4 was necessary for the increased survival of sams-1 animals after heat shock and found that the survival advantage in sams-1(RNAi) was decreased in sams-4(ok3315) animals (Figure 1—figure supplement 3E). These results suggest that H3K4me3 may be remodeled during heat shock with SAM from distinct synthases and that sams-4-dependent methylation is critical for survival. Previously, it was shown that H3K4me3 deposition is independent of sams-4 in embryonic nuclei (Towbin et al., 2012), however, our finding that it is broadly decreased in sams-4(RNAi) intestinal nuclei suggests it may have important roles in H3K4 methylation in adults.

Increases in H3K4me3 have also been shown to occur in budding yeast when blocks in phospholipid synthesis relieve a drain on SAM and increase levels (Ye et al., 2017), which we have confirmed in C. elegans (Ding et al., 2018). In order to determine if SAM levels could explain differences in H3K4me3 in sams-1 and sams-4 animals during heat shock, we used targeted LC/MS to compare SAM, it’s precursor methionine and S-adenosylhomocysteine (SAH), the product after methyl transfer, before and after heat shock. As in our previous assays, SAM decreased significantly after sams-1 or sams-4(RNAi) in basal conditions (Figure 1—figure supplement 3F), whereas SAM levels increased in each population as sams-1 or sams-4 animals were shifted to 37 °C for 2 hr (Figure 1—figure supplement 3F). Levels of methionine and SAH also decreased when comparing control, sams-1 or sams-4(RNAi) animals in basal vs heat-shocked conditions (Figure 1—figure supplement 3G, H), consistent with increased production and utilization of SAM. The increase in SAM in heat-shocked animals is consistent with our data showing the contribution of SAMS-4 to H3K4me3 and survival in heat-shocked sams-1 animals; however, a reduction in demand for SAM if other metabolic processes are reduced after heat shock could also contribute. Finally, levels of SAM in heat-shocked sams-4(RNAi) animals also rise to levels comparable to control animals at basal temperatures; however, H3K4me3 remains low in these conditions.

SAM is necessary for histone methylation; however, histone methylation dynamics are also influenced by methyltransferase (KMT) or demethylase (KDMT) activity (Bannister and Kouzarides, 2011). Therefore, changes in histone methylation dynamics could also impact H3K4me3 patterns during heat shock. H3K4me3 is catalyzed by multiple versions of the COMPASS complex, which each consist of one of several SET domain histone methyltransferases and several shared accessory subunits (Shilatifard, 2012). In mammals, seven methyltransferases in the SET1, MLL or THX groups can methylate H3K4. C. elegans contain single orthologs from two of these groups: set-2/SET1 and set-16/MLL, respectively, with roles in embryonic development (Li and Kelly, 2011; Xiao et al., 2011; Wenzel et al., 2011), lipid accumulation and transgenerational inheritance (Greer et al., 2010; Han et al., 2017). In adult C. elegans, set-2 RNAi results in extensive loss of H3K4me3 in intestinal nuclei and although set-16(RNAi) causes an intermediate reduction in bulk H3K4me3 levels, it has a broader requirement for survival during stress (Ding et al., 2018). Because specificity for H3K4 mono, di or trimethylation has not been verified on a genome-wide scale for KDMTs, we examined multiple members of the H3K4 KDM family.

In order to determine if KMTs or KDMT dynamics played a role in the change of H3K4me3 during heat shock, we used RNAi to deplete them in sams-1(lof) or sams-4(ok3315) animals and measured survival after heat shock and intestinal H3K4me3 levels. RNAi of set-2/SET1 (Figure 2A) or set-16/MLL (Figure 2B) increased survival in sams-1(lof) animals after heat shock (also Supplementary file 2:Tabs:C, E) and did not limit heat shock-induced H3K4me3 in sams-1(RNAi) nuclei (Figure 2D and E; GH). RNAi of two KDMTs, rbr-2 (Figure 2C) and spr-5 (Figure 2—figure supplement 1A) had opposite effects from the KMTs, moderately reducing survival (Supplementary file 2: Tab F), whereas amx-1 and lsd-1 had no effect (Figure 2—figure supplement 1B, C; Supplementary file 2: Tabs I, J). RNAi of set-2 or set-16 had slight, but statistically significant effects, increasing survival of sams-4(ok3315) animals (Figure 2—figure supplement 1D and E; Supplementary file 2: Tabs G, H). However, survival was still significantly below controls in sams-4(ok3315) with or without the KMT RNAi. Taken together, this suggests that set-2 and set-16 may act redundantly in the deposition of H3K4me3 after heat shock and are important to survival in sams-1(lof) animals. Furthermore, our data illustrate that the context is critical for understanding role of SAM and H3K4me3 in stress; sams-4 and set-16 are generally required for survival after heat shock, but loss of either H3K4 KTM enhances survival in sams-1(lof) animals.

H3K4me3 is a prevalent modification enriched near the transcription start sites (TSSs) of actively expressed genes (Eissenberg and Shilatifard, 2010). Differing global patterns of H3K4me3 in sams-1(RNAi) and sams-4(RNAi) nuclei suggest this histone modification at specific sites could also be distinct. In order to identify loci that might link H3K4me3 to these phenotypes, we used CUT&Tag, (Cleavage Under Targets and Tagmentation, C&T) (Kaya-Okur et al., 2019), to determine genome-wide H3K4me3 levels in Control RNAi, sams-1 and sams-4(RNAi) in basal (15 °C) and after heat shock (37 °C/2 hr) from two biologically independent replicates along with no antibody controls. C&T is uniquely suited to the small sample sizes available from these stressed populations. In this approach, a proteinA-Tn5 transposase fusion protein binds to the target antibody in native chromatin and DNA libraries corresponding to antibody binding sites are generated after transposase activation. After sequencing of libraries, we used the HOMER analysis suite Heinz et al., 2010 to analyze reads mapped to the C. elegans genome and called peaks using ChIPSeqAnno Zhu et al., 2010 for more detailed peak annotation. Bar plots from ChIPSeqAnno annotations and TSS plots generated with HOMER show robust mapping of H3K4me3 to promoter-TSS regions, validating this approach (Figure 3A; Supplementary file 3: Tabs A-F). While promoter-TSS regions were the largest feature in each sample, heat shocked sams-4(RNAi) animals had fewer overall peaks (Figure 3A). Correlation plots also show strong similarity between replicates (Figure 3—figure supplement 1A). Because C&T has not been extensively used in C. elegans, we compared data from basal conditions in our study to three previously published ChIP-Seq data sets (Ho et al., 2014; Pu et al., 2015; Wan et al., 2022). We compared our C&T data from wild type young adult animals grown at 15 °C on control RNAi food (HT115) against ModEncode (L3 animals), glp-1(e2141) mutants from Pu et al., 2018 and wild type adults grown at 20 °C on OP50 bacteria from Wan et al., 2022 by computing a pair-wise Pearson correlation. We found our C&T clustered most closely with the ChiPSeq from wild type animals in Wan et al., along with one of the modEndode replicates (Figure 3—figure supplement 1B) with moderate correlation scores. Both our C&T data and the Wan ChiPseq data correlated poorly with the Pu et al. ChIP seq, which is likely due to the lack of germline nuclei in these animals. The moderate correlation between our data and ChiP seq from Wan et al may be due to differences in growth temperature and bacterial diet. As a part of our quality control, we visually inspected browser tracks around the pcaf-1 gene, which is a long gene and has been used by our labs and others as a positive control for H3K4me3 localization in the five prime regions (Ding et al., 2015; Xiao et al., 2011). H3K4me3 peaks are prominent upstream of the transcript as expected and the no antibody libraries showed few reads (Figure 3—figure supplement 1C).

Next, we compared TSS distributions and examined overlap between H3K4me3 peaks in Control RNAi animals in basal and heat shock conditions and found moderate reductions occurred with heat shock (Figure 3B). Around 20–30% of peaks were specific to at either at basal (15 °C) vs. heat shock (37 °C) temperature (Figure 3C), suggesting that H3K4me3 could be remodeled upon heat shock in C. elegans. TSS enrichment of H3K4me3 was sharply reduced in both sams-1 and sams-4 samples at 15 °C; however, this difference was less marked in heat-shocked animals, in line with lower TSS localization in Control animals (Figure 3D and E). While aggregate TSS enrichment for H3K4me3 was similar for sams-1 and sams-4 RNAi animals, this analysis could miss distinct sets of H3K4me3 marked genes in each condition. Indeed, Control, sams-1 and sams-4(RNAi) animals each showed 500–1000 specific peaks in basal conditions, with moderate increases in these numbers after heat shock (Figure 3D and E). As H3K4me3 is a widely occurring modification, we hypothesized that we might better understand potential SAM synthase-specific requirements if we focused on peaks that change in the Control RNAi heat shock response and asked how they are affected by loss of sams-1 or sams-4. We used two different methods for comparing potential SAM synthase requirements for H3K4me3 in the heat shock response. First, we used differential peak calling ChIPPeakAnno Zhu et al., 2010 followed by WormCat category enrichment to determine the classes of genes which might be affected (Figure 3—figure supplement 2A–F; Supplementary file 3; Tabs G-I). Peaks present in both basal and heat shocked conditions were enriched for genes in the METABOLISM category (including Lipid: phospholipid, sphingolipid, sterol and lipid binding, along with mitochondrial genes) as well as in core function categories such as those involved in trafficking, DNA or mRNA functions (Figure 3F, Figure 3—figure supplement 2D–E; Supplementary file 3: Tabs G-I). There was no significant category enrichment specific to 15 °C animals, but after heat shock, Control RNAi animals gain enrichment in peaks at the Category 1 level in PROTEOSOME PROTEOLYSIS (Figure 3F). This enrichment was driven by increases in H3K4me3 at E3: Fbox genes (Figure 3—figure supplement 2A, B; Supplementary file 3:Tab A,B), which could be important for eliminating mis-folded proteins during heat shock. Comparison of peaks differentially present in sams-1 and sams-4 RNAi animals showed that only sams-1(RNAi) exhibited a similar enrichment to Control RNAi in the PROTEOLYSIS PROTEOSOME category (Figure 3G, Figure 3—figure supplement 2C, D), which could help explain the reduced survival of sams-4(RNAi) animals relative to sams-1(RNAi) animals. sams-1 RNAi animals also gained enriched peaks in a wide range of gene categories within METABOLISM, whereas sams-4(RNAi) enriched peaks in these categories were more limited (Figure 3—figure supplement 2C–F). Thus, loss of sams-1 or sams-4 differentially affects H3K4me3 peaks within functional gene classes that also change in the heat shock response.

Next, we hypothesized that H3K4me3 at peaks in Control RNAi animals might reflect multiple differently regulated populations, some which are linked to SAM synthase function and others that are regulated at other levels. In order to test this, we divided peaks in Control animals at 15 °C or 37 °C into those that did not change after SAM synthase RNAi (sams-1 or sams-4 independent peaks) or those that were dependent on sams-1 or sams-4 and examined aggregations around TSS regions. There was little difference between TSS plots of sams-1 or sams-4-independent genes at either temperature (Figure 3H1). However, in basal conditions, Control peaks that depended on sams-1 had more marked TSS localization (Figure 3J), demonstrating that sams-1 and sams-4 dependent peaks have distinct TSS architectures. TSS localization was low in all 37 °C samples, following the general trend of decrease after heat shock (Figure 3K). We next separated Control peaks into those that were generally SAM synthase-dependent and those that were specific to loss of sams-1 or sams-4. Aggregation of these peaks shows that peaks in Control 15 °C samples that were lost only in sams-4 RNAi also had the lowest levels of H3K4me3 in TSS regions, whereas promoters that lost this modification only after sams-1 RNAi had higher levels of H3K4me3 (Figure 3—figure supplement 2G). Control 37 °C samples exhibited a similar pattern, with a lower H3K4me3 level overall consistent with what we have observed in heat shock samples (Figure 3—figure supplement 2H). Thus, genome wide H3K4me3 contain multiple populations with distinct TSS patterns. Peaks that are present even when sams-1 or sams-4 are depleted have the highest levels, whereas sams-1-dependent peaks have moderate H3K4me3, and peaks that are lost after sams-4 RNAi have the lowest levels. Taken together, this shows that individual SAM synthases are linked to distinct sets of H3K4me3 within the genome.

Our C&T and RNA seq assays were performed on whole animals. While sams-1 and sams-4 are co-expressed in the intestine and hypodermis, which are major stress-responsive tissues, the germline nuclei contain only sams-4 (Figure 1B and Figure 1—figure supplement 1B). This aligns with our previous observations that sams-1(RNAi) animals had normal patterns of H3K4me3 in germline nuclei (Ding et al., 2015), whereas RNAi of sams-4 abrogates H3K4me3 staining in germline nuclei (Figure 3—figure supplement 3A). However, embryo production and development appear broadly normal in sams-4 RNAi embryos (not shown). In order to determine how H3K4me3 might align with tissue-specific expression patterns, we aggregated peaks from tissue-specific RNA seq data published by Serizay et al., 2020. Serizay et al. separated nuclei based on tissue specific GFP expression and defined gene sets that were expressed that were ubiquitously, as well as those that were present only in a single tissue. They also performed ATAC seq (Assay for Transposase-Accessible Chromatin using sequencing). Serizay, et al. defined transcripts by expression pattern and defined sets that were specific to (tissue_only), or represented in across multiple tissues (tissue_all). ubiquitious_all and Germline_only genes had the most defined patterns of open chromatin around TSSs (Serizay et al., 2020). We compared our C&T data with Ubiquitious_all, Germline_only and Intestine_only genes and found that we identified peaks for around half of these genes in Control RNAi animals at 15 °C or 37 °C (Figure 3—figure supplement 3B–D). We found the ubiquitous_all and germline_only genes also had strong H3K4me3 peaks that were reduced equally by sams-1 or sams-4 RNAi in both temperature conditions (Figure 3—figure supplement 3D, G; E, H). Intestine_only genes showed lower TSS enrichment but were similarly reduced after sams-1 or sams-4 RNAi (Figure 3—figure supplement 3H, I). These data suggest that differences in germline expression for sams-1 and sams-4 are not sufficient to explain differential effects on H3K4me3 peak populations.

H3K4me3 is found at the promoters of many actively transcribed genes, but it is not necessarily required for gene expression (Bannister and Kouzarides, 2011). However, studying chromatin modification in stress responses may reveal additional regulatory effects (Weiner et al., 2012). We previously found using ChIP-PCR in the context of the stress response in C. elegans that H3K4me3 increased at promoters of genes that responded to bacterial stress in a sams-1-dependent manner (Ding et al., 2015). However, during the stress response, H3K4me3 did not change at multiple non-stress responsive genes, suggesting that stress-responsive loci might be more sensitive to SAM levels (Ding et al., 2015). In order to identify genes that changed in SAM-deficient animals, we performed RNA seq, then compared genes induced by heat shock in control and sams-1(RNAi) (Ding et al., 2018) with genes induced in sams-4(RNAi) animals (Supplementary file 4: A-C). Upregulated genes for control and sams-1(RNAi) animals appeared closely grouped in principal component analysis, with sams-4(RNAi) upregulated genes and all downregulated gene sets forming distinct groups (Figure 4—figure supplement 1A). We previously noted that while sams-1(RNAi) animals could not mount the full transcriptional response to bacterial stress, most genes activated by heat increased similarly to controls (Ding et al., 2018). sams-4(RNAi) animals, in contrast, activate less than 25% of the genes induced by heat in control animals (Figure 4A). sams-1(RNAi) and sams-4(RNAi) animals also induce more that 600 genes in response to heat that are SAM-synthase-specific and which do not increase in control animals (Figure 4A). WormCat pathway analysis shows that sams-4(RNAi) animals lack the robust enrichment in STRESS RESPONSE (Cat1) and STRESS RESPONSE: Heat (Cat2) evidenced in both Control and sams-1(RNAi) samples (Figure 4B; Supplementary file 4: D-F). In addition, enrichment of the CHAPERONE, PROTEOLYSIS PROTEOSOME categories occurring in sams-1(RNAi) animals does not occur after sams-4(RNAi), reflecting lack of induction of these genes which could be important for proteostasis in the heat shock response (Figure 4C). Thus, reduction in sams-1 or sams-4 results in distinct gene expression programs in both basal conditions (Figure 1—figure supplement 1A and B) and during the heat stress response (Figure 4A–C). This differentiation of gene expression programs clearly shows that sams-1 and sams-4 have distinct functional roles.

Gene expression changes occurring after sams-1 or sams-4 depletion could result from direct effects on H3K4me3 or other potential methylation targets, or from indirect effects. Evaluating the impact H3K4me3 on gene expression is also complex, as this modification is generally associated but not necessary for expression of actively transcribed genes Bannister and Kouzarides, 2011. In our analysis of H3K4me3 peaks during the heat stress response, we found evidence of multiple peak populations that depend on or occur independently of sams-1 or sams-4 (Figure 3H–K, Figure 3—figure supplement 3A-F). We reasoned, therefore, that it was also critical to determine H3K4me3 levels at sams-1- or sams-4-dependent genes in the heat shock response.

First, we examined H3K4me3 peak levels at genes with increased in Control RNAi, sams-1(RNAi) or sams-4(RNAi) during heat shock. We found that genes dependent on sams-1 or sams-4 in the heat shock response were marked by lower overall H3K4me3 levels at the TSSs (Figure 4D). However, this analysis included large numbers of upregulated genes in sams-1 or sams-4 outside of the wildtype heat stress response. Therefore, we next focused on genes normally upregulated during heat shock and divided them according to SAM synthase dependence. Strikingly, isolating the sams-1-dependent genes revealed a strong peak 5’ to the TSS, which was not evident in the larger subset of Control or sams-4(RNAi)-dependent upregulated genes (Figure 4E and F). Among the genes with robust peaks in heat-shocked sams-1(RNAi) animals were two F-box proteins, fbxa-59 and T27F6.8, which were robustly expressed in sams-1 but not sams-4 animals (Figure 4C–). Downregulation of T27F6.8 did not affect the survival of the animals after heat shock (Figure 4—figure supplement 1B) while survival of animals fed fbxa-59 RNAi was modestly affected (Figure 4—figure supplement 1C). Survival in heat shock may be multi-genic and rely on pathway responses rather than single genes. However, our data reveals genes upregulated in the heat shock response may have different H3K4me3 levels depending on requirements for sams-1 or sams-4. In addition, our results suggest that roles for H3K4me3 may become clearer when genome-wide methylation populations are separated into biologically responsive categories.

Transcriptional responses to heat shock largely focus on rapidly induced genes that provide protection from changes in proteostasis (Morimoto, 2006; Mahat et al., 2016). However, downregulated genes could also play important roles. For example, the WormCat category of TRANSMEMBRANE TRANSPORT (TM TRANS) is enriched in genes downregulated after heat shock in C. elegans (Figure 5A and B). Previously we observed that heat shocked animals depended on sams-1 for normal expression of nearly 2000 genes, falling within WormCat Categories of METABOLISM, TRANSCRIPTION FACTOR (TF), SIGNALLING, and STRESS RESPONSE (Ding et al., 2018; Figure 5A and B). Interestingly, the metabolic genes dependent on sams-1 include those in lipid metabolism, whereas the TF enrichment was centered around nuclear hormone receptors (NHRs) (Figure 5C and D), which regulate many metabolic and stress responsive genes in C. elegans (Arda et al., 2010). However, neither the shared TM TRANSPORT nor the sams-1 specific categories depend on sams-4 (Figure 5B and C). Thus, as in genes upregulated during the heat shock response, genes downregulated in the heat shock response also have differential requirements for sams-1 and sams-4.

Next, we examined H3K4me3 levels around TSSs of genes that lost expression during heat shock in Control, sams-1 or sams-4(RNAi) animals. Genes decreasing in Control animals had a slight reduction of H3K4me3 peaks when comparing15°C and 37 °C samples, consistent with global levels after heat shock (Figure 5D). RNAi of sams-1 or sams-4 also broadly reduced H3K4me3 TSS enrichment at downregulated genes (Figure 5D–F). However, there were minimal differences before or after heat shock, suggesting expression patterns affecting survival could be established before induction of the stress.

H3K4me3 has been reported to act as a bookmarking modification, therefore we hypothesized that some loci could be affected before heat shock, with expression changing afterward. Therefore, we more closely examined genes with sams-1-dependent H3K4me3 at 15 °C that lost expression during heat shock. Those genes were highly enriched for METABOLISM: Lipid: beta oxidation and NHR transcription factors (Figure 6A). We noted they included multiple members of a regulatory circuit that control expression of a beta-oxidation-like pathway that degrades toxic fatty acids identified by the Walhout lab (Bulcha et al., 2019), including nhr-68, nhr-114 and beta-oxidation genes acdh-1, hach-1 ech-6,–8, and –9 (Figure 6B and C). Indeed, nhr-68, the initiating TF in this regulatory circuit, shows lower levels of H3K4me3 at its promoter in basal conditions, compared to Control or sams-4 RNAi animals (Figure 6D). The H3K4me3 peak overlaps with another gene, pms-2, whose expression does not change after heat shock or upon SAM synthase RNAi (Supplementary file 4: Tabs A-C). In order to test if H3K4me3-dependent regulation of nhr-68 was important for survival during heat shock, we made use of a construct expressing nhr-68 under the intestine-specific ges-1 promoter (Bulcha et al., 2019), where H3K4me3 peaks do not change after RNAi of sams-1 or sams-4 (Figure 6—figure supplement 1A). Expression of nhr-68 under this heterologous promoter had a moderate, but significant effect on survival (Figure 6E). Thus, downregulation of nhr-68 in sams-1 animals after heat shock could be part of a program enhancing survival (Figure 6—figure supplement 1B). Taken together, our results suggest differences in H3K4me3 patterns in sams-1 and sams-4 animals before heat shock may also influence gene expression patterns during the stress response. This demonstrates that sams-1 and sams-4 are required for distinct sets of genes in the heat stress response and contribute to different H3K4me3 patterns.

The genetic tools used in our study provide a method to reduce SAM from a specific enzymatic source. However, the roles for SAM in the cell are broad and can affect methylation of multiple targets. While our metabolomics assays show that SAM increases in heat shocked sams-1(RNAi) animals, we have not demonstrated that this SAM is derived from sams-4. In addition, survival benefits after heat shock occur across broad cellular functions including proteostasis and other methylation marks such as H3K9me3 (Das et al., 2021). Thus, there may be multiple additional methylation-dependent mechanisms that influence survival of sams-1 or sams-4 animals during heat shock. In addition, we measured gene expression and H3K4me3 at 2 hr post heat shock, whereas the survival assay occurs over multiple days. Thus, there may be changes in gene expression or histone modifications occurring at later times that also affect survival.

N2(Caenorhabditis Genetics Center); sams-1(lof)(ok3033); sams-3(ok2932) IV, sams-4((ok3315) IV, Caenorhabditis Genetics Center), tagRFP::SAMS-1 (WAL500, this study); GFP::SAM-4(WALK501, this study); SAMS-1::RFP;GFP::SAMS-4(WAL502, this study), SAMS-3::mKate (WAL305). Pges-1::NHR-68::GFP (VL1296) (Bulcha et al., 2019). CRISPR tagging for WAL500 and WAL501 were done by the UMASS Medical School transgenic core, confirmed by PCR for genotype and outcrossed three times to wild type animals. Next, each strain was crossed to the respective deletion allele to create WAL503 (RFP::sams-1(ker5); sams-1(ok3033)) and WAL504(GFP::sams-4(ker6); sams-4(ok3315)). sams-3::mKate(nu3139) (COP2476) was constructed using CRISPR by In Vivo biosystems then outcrossed three times (WAL305).

C. elegans (N2) were cultured using standard laboratory conditions on E. coli OP50 or HT115 expressing appropriate RNAi. RNAi expression was induced using 6 mM IPTG. Adults were bleached onto RNAi plates and allowed to develop to the L4 to young adult transition before stresses were applied. For heat stress applications, animals were raised at 15 °C from hatching then at the L4/young adult transition replicate plates were placed at 15 °C or 37 °C for 2 hr. After each stress, animals were washed off the plates with S-basal, then pellets frozen at –80 °C. RNA was prepared as in Ding et al., 2015. For survival assays, ~10–15 adult N2 animals were bleached on 60 mm RNAi plates. The eggs were allowed to hatch and grow to young adults at 15 °C. Twenty-five to 30 young adults were then moved to 35 mm plates in triplicate (75–90 animals per RNAi treatment) and subjected to heat shock at 37 °C for 2 hr. Animals were kept at 20 °C for the remainder of the assay. Dead animals were identified by gentle prodding, were counted and removed each day. Animals that died of bagging or from desiccation on the side of the plate were not counted. Three independent non blinded biological replicates were carried out and Kaplan-Meir curves were generated with GraphPad Prism v8.0. For lifespan experiments, the N2 adults were bleached on 60 mm RNAi plates. The eggs were allowed to hatch and grow to young adults at 20 °C. Twenty-five to 30 young adults were then moved to 35 mm plates in triplicate (75–90 animals per RNAi treatment). Adults were moved to fresh plates every day and dead animals were identified by gentle prodding and removed each day. Three independent non blinded biological replicates were carried out and Kaplan-Meir curves were generated with GraphPad Prism v8.0.

RNA for deep sequencing was purified by Qiagen RNAeasy. Duplicate samples were sent for library construction and sequencing at BGI (China). Raw sequencing reads were processed using an in-house RNA-Seq data processing software Dolphin at University of Massachusetts Medical School (Yukselen et al., 2020). The raw read pairs were first aligned to C. elegans reference genome with ws245 annotation. The RSEM method was used to quantify the expression levels of genes and Deseq was used to produce differentially expressed gene sets with more than a twofold difference in gene expression, with replicates being within 0.05 in a Students T test and a False Discovery Rate (FDR) under 0.01. Statistics were calculated with DeBrowser (Kucukural et al., 2019). Venn Diagrams were constructed by BioVenn (Hulsen et al., 2008). WormCat analysis was performed using the website https://www.wormcat.com/ (Holdorf et al., 2020; Higgins et al., 2022) and the whole genome annotation version 2 (v2) and indicated gene sets. PCA was conducted by using prcomp in R and graphed with ggplot in R studio.

For H3K4me3 (Cell Signaling Technology, catalogue number C42D8) staining, dissected intestines were incubated in 2% paraformaldehyde, freeze cracked, then treated with –20°C ethanol before washing in PBS, 1% Tween-20, and 0.1% BSA. Images were taken on a Leica SPE II at identical gain settings within experimental sets. Quantitation was derived for pixel intensity over nuclear area for at least seven dissected intestines, with at least three nuclei per intestine. Three biological repeats were carried out for every experiment.

C. elegans (N2) gravid adults (~15–20) were bleached onto 60mm RNAi plates, eggs were allowed to hatch and grow to young adults at 15°C. For heat stress application, replicate plates were placed at either 15°C or 37°C for 2 hr. At the end of the heat stress, worms were collected in S-Basal, and pellets were frozen at –80°C. Four independent biological replicates were collected. To prepare the samples for LC-MS, the pellet was thawed on ice and washed with 0.9% NaCl. Washed worms were then transferred to 2 mL FASTPREP tubes (MP Biomedicals) containing 1.4 mm ceramic beads (Qiagen). The samples were then resuspended in 1 mL 80% methanol (LC-MS grade) and homogenized using a bead beater (6.5 m/s; 20 s). The samples were cooled on ice between cycles. The homogenized samples were then vortexed at 4 °C for 10 min and centrifuged at 21,000 RPM for 10 min at 4 °C. The supernatant was removed at dried under vacuum. The pellet was resuspended in ice cold RIPA buffer and vortexed at 4 °C for 10 min and centrifuged at 21,000 RPM at 4 °C for 10 min. The supernatant was removed and used for protein quantification using Pierce Protein BCA assay kit (ThermoFisher). The protein quantification was then used to resuspend the pellet for an equal input of 0.5 μg/ml of protein per sample.

Metabolomics was conducted on a QExactive Plus bench top orbitrap mass spectrometer equipped with an Ion Max source and a HESI II probe, which was coupled to a Vanquish Horizon HPLC system (Thermo Fisher Scientific, San Jose, CA). External mass calibration was performed using the standard calibration mixture every 7 days. Dried extracts were reconstituted in enough water to achieve a final concentration of 0.5 μg/ml protein per sample. Two μL of this resuspended sample were injected onto a SeQuant ZIC-pHILIC 150x2.1 mm analytical column equipped with a 2.1x20 mm guard column (both 5 mm particle size; Millipore Sigma). Buffer A was 20 mM ammonium carbonate, 0.1% ammonium hydroxide; Buffer B was acetonitrile. The autosampler tray was held at 4°C. The chromatographic gradient was run at a flow rate of 0.150 mL/min as follows: 0–20 min: linear gradient from 80% to 20% B; 20–20.5 min: linear gradient form 20% to 80% B; 20.5–28 min: hold at 80% B. The mass spectrometer was operated in full-scan, polarity-switching mode, with the spray voltage set to 4.0 kV, the heated capillary held at 320°C, and the HESI probe held at 350°C. The sheath gas flow was set to 10 units, the auxiliary gas flow was set to 1 units, and the sweep gas flow was set to 1 unit. MS data acquisition was performed in a range of m/z=70–1000,, with the resolution set at 70,000, the AGC target at 1x10⁶, and the maximum injection time at 20 ms. An additional scan (m/z 220–700) in negative mode only was included to enhance detection of nucleotides. Relative quantitation of polar metabolites was performed TraceFinder 5.1 (Thermo Fisher Scientific) using a 5 ppm mass tolerance and referencing an in-house library of chemical standards. Statistical analysis was carried out for the data using GraphPad Prism (v8.0).

C. elegans (N2) were cultured using standard laboratory conditions on E. coli OP50. Adults were bleached onto RNAi plates and allowed to develop to the L4 to young adult transition before heat stress was applied. For heat stress applications, animals were raised at 15 °C from hatching then at the L4/young adult transition replicate plates were placed at 15 °C or 37 °C for 2 hr. At the end of the heat stress, animals were washed off the plates with S-basal, then pellets frozen at –80 °C. Worm pellets were washed with S-Basal to remove bacteria, then resuspended in 750 μL of chilled Nuclei Purification Buffer (50 mM HEPES pH = 7.5, 40 mM NaCl, 90 mM KCl, 2 mM EDTA, 0.5 mM EGTA, 0.2 mM DTT, 0.5 mM PMSF, 0.5 mM spermidine, 0.1% tween 20, and cOmplete proteinase inhibitor cocktail (Roche)). The suspension was then transferred to Potter-Elvehjem Tissue Grinder (3 mL). The worms were ground with 2–3 cycles consisting of ~45–50 strokes of the grinder. The samples were chilled on ice for ~5 min between consecutive cycles. The lysates were passed through 100 micron filter (X3) followed by 40 micron (X3) (Pluriselect). The lysates were then centrifuged at 4500 RPM for 10 min at 4 °C. The pellets were resuspended gently in wash buffer (1 M HEPES pH 7.5, 5 M NaCl, 2 M spermidine). Concanavalin bead slurry (10 μL/sample) was added gently to the samples and allowed to incubate at room temperature for 15 min in an end-over-end rotator. The sample tubes were then transferred to a magnetic stand and liquid was gently removed. The nuclei were gently resuspended in 50 μL of chilled antibody buffer (8 μL 0.5 M EDTA, 6.7 µL 30% BSA in 2 mL Dig-wash buffer (400 μL 5% digitonin with 40 mL Wash buffer)). 1 uL anti-H3K4me3 antibody (Cell Signaling Technology, catalogue number C42D8) was added to the suspension and allowed to bind overnight at 4 °C on a nutator shaker. Samples without any antibody added were used as controls to correct for background reads and further processed per the CUT&Tag protocol Kaya-Okur et al., 2019 to generate sequencing libraries. The libraries were amplified by mixing 21 μL of DNA with 2 μL each of (10 μM) barcoded i5 and i7 primers, using a different combination for each sample. 25  μL NEBNext HiFi 2×PCR Master mix (NEB) was added, and PCR was performed using the following cycling conditions: 72  °C for 5  min (gap filling); 98  °C for 30  s; 17 cycles of 98  °C for 10   s and 63  °C for 30 s; final extension at 72  °C for 1  min and hold at 4  °C. 1.1×volume of Ampure XP beads (Beckman Coulter) was incubated with libraries for 10 min at room temperature to clean up the PCR mix. Bead bound DNA was purified by washing twice with 80% ethanol and eluting in 20  μL 10  mM Tris pH 8.0. Size distribution of the libraries was determined by Fragment analyzer and concentration by the KAPA Library Quantification Kit before sequencing to determine the H3K4me3 landscape in basal and heat stress condition in worms fed on control, sams-1 or sams-4 RNAi. Sequencing of the prepared libraries was carried out on Illumina NextSeq 500.

Paired end reads from each sample were aligned to the C. elegans genome (ce11 with ws245 annotations) using Bowtie2 (Langmead et al., 2009) with the parameters -N 1 and -X 2000. Duplicate reads were removed using Picard (RRID:SCR_006525) and the reads with low quality scores (MAPQ <10) were removed. HOMER software suite was used to process the remaining mapped reads (Heinz et al., 2010). The ‘makeUCSCfile’ command was used for generating genome browser tracks. Data was normalized to library size. the ‘findPeaks <tag directory> -style histone -o auto’ command was used for calling H3K4me3 peaks and the ‘annotatePeaks’ command was used for making aggregation plots. Differential peak calling was accomplished using (Zhu et al., 2010) the command “. We used the findOverlapsOfPeaks command in ChipSeqAnno³⁷ with a max gap of 1000 basepairs to determine peak overlap. TSS plots were generated using HOMER (Heinz et al., 2010) and Venn Diagrams were constructed by BioVenn (Hulsen et al., 2008).

Correlation matrices were generated with deeptools version 3.5.1 (Ramírez et al., 2016). Multibamsummary was used to compare bam files from each sample, using default values except --binSize 2000. This data was visualized using plotCorrelation with --removeOutliers and the Pearson method. Previously published datasets were used to compare H3K4me3 Cut and Tag versus previously published data sets. Young adults fed a normal diet were used from Wan et al., 2022. Day 2 glp-1 adults were chosen from Pu et al., 2015. modENCODE ChIP-seq data drew from L3 animals (Ho et al., 2014).