Decision letter: SET-9 and SET-26 are H3K4me3 readers and play critical roles in germline development and longevity

Abstract

Article Figures and data Abstract eLife digest Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract C. elegans SET-9 and SET-26 are highly homologous paralogs that share redundant functions in germline development, but SET-26 alone plays a key role in longevity and heat stress response. Whereas SET-26 is broadly expressed, SET-9 is only detectable in the germline, which likely accounts for their different biological roles. SET-9 and SET-26 bind to H3K4me3 with adjacent acetylation marks in vitro and in vivo. In the soma, SET-26 acts through DAF-16 to modulate longevity. In the germline, SET-9 and SET-26 restrict H3K4me3 domains around SET-9 and SET-26 binding sites, and regulate the expression of specific target genes, with critical consequence on germline development. SET-9 and SET-26 are highly conserved and our findings provide new insights into the functions of these H3K4me3 readers in germline development and longevity. https://doi.org/10.7554/eLife.34970.001 eLife digest Cells keep their DNA organized by wrapping it around groups of proteins called histones. These structures not only keep the genetic code tidy, they also affect how and when a cell uses its genes. This is because small chemical groups that are added to histones, such as a methyl group added to the fourth position of histone H3 (known as H3K4me3), affect which proteins can access the surrounding genes. This in turn determines whether those genes are likely to be on or off. Many proteins help to regulate histone modifications, including proteins that add or remove the specific chemical groups. Enzymes that add a methyl group to histone usually contain a region called SET; while proteins containing a structure called a PHD finger can recognize histone modifications and help to amplify the signal to switch a gene on or off. SET-9 and SET-26 are two proteins containing both SET regions and PHD fingers. Found in the worm Caenorhabditis elegans, these proteins are 97% identical. Changes in histone modifications can affect the lifespan of these worms, and the number of offspring they produce. Recent work revealed that loss of SET-9 and SET-26 makes the worms live longer. Now, Wang et al. use gene editing to better understand how these proteins have their effects. Experiments with worms lacking the gene for SET-9 or SET-26 or both revealed that, despite looking almost identical, SET-9 and SET-26 have different roles. Every cell in the worm makes SET-26 protein and getting rid of it increases their lifespan by affecting the activity of a protein called DAF-16. But, only the cells in the reproductive system make SET-9, and both proteins play a role in fertility. A technique called ChIP-seq revealed where each protein attached to the genome. The PHD fingers of SET-9 and SET-26 bound to around half of the possible H3K4me3 modification sites. Not all the possible sites actually had a methyl group attached, and the pattern of binding matched the pattern of modifications. This indicates that the two proteins arrive only once the positions already have their methyl groups. Getting rid of the SET-9 and SET-26 proteins increased the number of H3K4me3 sites with methyl groups attached. This suggests that the role of SET-9 and SET-26 is to stop the spread of H3K4me3 modifications, controlling the use of certain genes. In mammals, the proteins SETD5 and MLL5 likely do the job of SET-9 and SET-26. Understanding how they work in worms could further our understanding of fertility and ageing in humans. https://doi.org/10.7554/eLife.34970.002 Introduction Dynamic regulations of histone methylation status have been linked to many biological processes (Santos and Dean, 2004). Recent studies have revealed that specific histone methyltransferases and demethylases can play key roles in regulating germline functions and/or modulating longevity (Han and Brunet, 2012; Greer and Shi, 2012). In C. elegans, loss of the COMPASS complex, which is critical for methylating histone H3 lysine 4, results in a global decrease in H3K4 trimethylation (H3K4me3) levels, and reduced brood size (Li and Kelly, 2011; Robert et al., 2014) and extended lifespan (Greer et al., 2010) phenotypes. Interestingly, deletion of spr-5 or rbr-2, both of which encode demethylases that erase methyl marks on H3K4, also results in fertility defects (Alvares et al., 2014; Katz et al., 2009) and altered lifespan (Alvares et al., 2014; Greer et al., 2010). These findings suggest that histone methylations need to be precisely controlled to maintain longevity and germline function. It is important to note that the molecular mechanisms whereby these H3K4 modifying enzymes effect reproduction and longevity functions, including the genomic regions that they act on and how the altered H3K4 methylation levels contribute to the biological outcomes in these mutants, are largely unknown. SET (Su(var)3–9, Enhancer of Zeste, Trithorax) domain-containing proteins represent a major group of histone methyltransferases (Dillon et al., 2005). We previously carried out a targeted RNAi screen to identify the SET-domain containing proteins that play a role in longevity in C. elegans. We found that RNAi knockdown of set-9 and set-26, two closely related genes, results in significant lifespan extension (Ni et al., 2012). SET-9 and SET-26 share 96% sequence identity and both proteins contain highly conserved PHD and SET domains. PHD domains are known to bind to specific histone modifications (Shi et al., 2007; 2006), suggesting that SET-9 and SET-26 could be recruited to chromatin via binding to specific histone marks. Interestingly, the SET domain of SET-9 and SET-26 contains mutations in conserved residues thought to be key for methylating activities (Ni et al., 2012), making it unclear whether SET-9 and SET-26 could be active enzymes. Nevertheless, a recent study reported that the SET domain of SET-26 exhibits H3K9me3 activity in vitro (Greer et al., 2014). In this work, we demonstrated that, despite their high sequence identity, SET-26, but not SET-9, plays a key role in heat stress response and longevity. In addition, we revealed a novel redundant function of SET-9 and SET-26 in germline development. We also confirmed that SET-26 is broadly expressed, whereas SET-9 is only expressed in the germline, which likely accounts for their distinct and redundant functions in lifespan and reproduction. Indeed, genetic and transcriptomic analyses supported the notion that SET-26 acts through the FOXO transcription factor DAF-16 in the soma to modulate longevity. Furthermore, we showed that the PHD domains of SET-9 and SET-26 bind to H3K4me3 in vitro and that the genome-wide binding patterns of SET-9 and SET-26 are highly concordant with that of H3K4me3 marking in C. elegans, indicating that SET-9 and SET-26 are recruited to H3K4me3 marked regions in vivo. Although the SET domain of SET-26 was reported to methylate H3K9me3 in vitro (Greer et al., 2014), our results indicated that loss of set-9 and set-26 does not affect global H3K9me3 levels and the genome-wide binding patterns of SET-9 and SET-26 are highly divergent from that of H3K9me3. Instead, we found that loss of set-9 and set-26 results in expansion of H3K4me3 marking surrounding most if not all of the SET-9 and SET-26 binding sites specifically in the germline, and significant RNA expression change of a subset of germline specific genes bound by SET-9 and SET-26. We propose that SET-9 and SET-26 are recruited to the chromatin via binding to H3K4me3, where they function to restrict H3K4me3 spreading and to regulate the expression of specific genes, and together these activities contribute to the proper maintenance of germline development. Results set-26, but not set-9, single mutant exhibits prolonged lifespan and heightened resistance to heat stress RNAi knockdown of the highly similar paralogs set-26 and set-9 were previously shown to significantly extend lifespan in C. elegans (Greer et al., 2010; Ni et al., 2012). However, due to their high sequence similarity, RNAi likely knocks down both set-26 and set-9 in those experiments. We confirmed the lifespan extension phenotype with multiple available set-26 single mutants, but a set-9 single mutant was not available at the time (Ni et al., 2012). To delineate whether set-9, like set-26, also plays a role in lifespan determination, we used CRISPR-cas9 to generate a set-9 mutant (Figure 1A). The set-9 mutant we obtained carries a mutation that causes a premature stop codon and is expected to produce a truncated SET-9 protein lacking the conserved PHD and SET domains (Figure 1A). We tested the lifespan phenotype of this set-9 single mutant along with the set-26 single and set-9 set-26 double mutants. Consistent with previous results, the set-26 single mutant lived longer than wild-type worms (Figure 1B). Surprisingly, although SET-9 and SET-26 proteins share 97% identity in protein sequence, the set-9(rw5) mutation did not alter lifespan in either wild-type or the set-26 mutant background (Figure 1B). Similar to the lifespan phenotype, set-26, but not set-9 single mutant, was more resistant to heat stress compared to wild-type worms (Figure 1C). These results suggested that inactivation of set-26, but not set-9, extends lifespan and improves heat resistance. Figure 1 Download asset Open asset set-26 but not set-9 is important for longevity. (A) Schematic of the set-9(rw5) and set-26(tm2467) mutants. Red star indicates the position of the sgRNA (single guide RNA) targeting the set-9 gene. Premature stop codons caused by deletions of 38 nucleotides in the set-9 gene and 1090 nucleotides in the set-26 gene are depicted as red hexagons. Loss of set-26 gene but not set-9 gene extended lifespan (B), and increased resistance to heat stress (C). Survival curves for N2, set-26(tm2467), set-9(rw5), and set-9(rw5) set-26(tm2467) strains from representative experiments are shown. Quantitative data for all replicates are shown in Supplementary file 1 Table S1. https://doi.org/10.7554/eLife.34970.003 set-9 and set-26 act redundantly to maintain germline function While propagating the set-9 set-26 double mutant, we noticed a possible fertility defect. To more thoroughly assess the roles of SET-9 and SET-26 in reproduction, we assayed the brood size of set-9, set-26 single and double mutants. We found that the progeny number produced by set-9 and set-26 single mutants was slightly smaller compared to that of wild-type worms (Figure 2A). Interestingly, the homozygous set-9 set-26 double mutant derived from heterozygous parents (first generation, i.e. F1) also exhibited a mild brood size defect, and this defect became significantly more severe in the second and later generations (F2 to F6, Figure 2A). Deficiency of several histone modifiers has been previously reported to exhibit a ‘mortal germline’ phenotype. We performed a classical mortal germline assay and found that the set-9 set-26 double mutant indeed displayed a mortal germline phenotype (Figure 2B). Further detailed analyses indicated that the set-9 set-26 double mutant exhibited a high sterile rate and a low brood size through the F2-F6 generations that we assayed (Figure 2A and Figure 2—figure supplement 1A). Figure 2 with 1 supplement see all Download asset Open asset set-9 and set-26 act redundantly to maintain fertility. (A) The set-9(rw5) set26(tm2467) double mutant worms derived from heterozygous parents (F1) displayed a mild fertility defect. The double mutant worms displayed a much more severe fertility defect at later generations (F2–F6). Average brood size of N2, set-26(tm2467), set-9(rw5), and set-9(rw5) set-26(tm2467) strains at the indicated generation were shown (*p<0.05, ***p<0.001). The error bars represent standard errors. n = 9 ~10 for N2, set-9 mutant, set-26 mutant, and F1 set-9 set-26 double mutant worms; n = ~50 for F2-F6 set-9 set-26 double mutants. (B) The set-9(rw5) set26(tm2467) double mutant exhibited a mortal germline phenotype. At each generation, 6 L1s for N2, set-26(tm2467), set-9(rw5) and set-9(rw5) set-26(tm2467) strains were transferred to a new plate. Plates were scored as not fertile when no progeny were found. % of fertile lines indicated percentage of plates that were fertile. n = 6 for N2, set-9 and set-26 mutants; n = 25 for set-9 set-26 double mutants. (C) Maternal contribution of set-9 and set-26 appeared important for alleviating the fertility defect in the double mutant. Average brood size of the set-9(rw5) set26(tm2467) double mutants derived from four different crosses were shown (***p<0.001, n.s. no significant). n = 11 ~ 12 for assessing the brood size of the set-9(rw5) set26(tm2467) homozygous progeny from heterozygous male(P0) X heterozygous hermaphrodite(P0) and homozygous male(F1) X heterozygous hermaphrodite(P0); n = 30 ~34 for progeny from homozygous male (F1) X heterozygous hermaphrodite(P0) and heterozygous male(P0) X homozygous hermaphrodite(F1). (D) The set-9(rw5) set26(tm2467) double mutant worms that remained fertile nevertheless exhibited reduced number of mitotic germ cells. Whole worms or dissected gonads of fertile set-9(rw5) set26(tm2467) mutants were stained by DAPI and the mitotic cells were counted. D2 adults were scored. n = 18 ~ 27, ***p<0.001. Analyses of sterile set-9(rw5) set26(tm2467) double mutant worms are shown in Figure 2—figure supplement 1. Quantitative data are shown in Supplementary file 1 Table S2. https://doi.org/10.7554/eLife.34970.004 Since we noted a large difference between the brood size of the set-9 set-26 double mutant in the F1 and F2 generations and suspected a possible maternal influence, we therefore performed a series of crosses to test this possibility. We found that set-9 set-26 double mutants derived from homozygous set-9 set-26 hermaphrodites crossed with heterozygous fathers exhibited a significantly more severe brood size defect compared to those from heterozygous hermaphrodites crossed with homozygous mutant fathers (Figure 2C). In other words, heterozygous mothers, but not heterozygous fathers, helped to maintain better germline function in the progeny. These data supported the notion of a maternal contribution in germline maintenance in the set-9 set-26 double mutant worms. We next used DAPI staining to monitor the germ cells of the set-9 set-26 double mutant at the F3 and F4 generations. For the double mutant worms that became sterile, we observed variable germline phenotypes, including a very small mitotic region with no differentiated cells, and a small mitotic region with sperms only or a largely normal mitotic region with oocytes only (Figure 2—figure supplement 1B), suggesting problems with both the germline stem cells and their subsequent differentiation. For the double mutant worms that remained fertile, we observed germlines with a smaller but stable number of mitotic cells (Figure 2D). The results together indicated that SET-9 and SET-26 act redundantly to maintain normal germline function and they may regulate both the proliferation and differentiation of the germline stem cells. SET-26 is broadly expressed but SET-9 is only detectable in the germline Given the high degree of sequence identity between SET-9 and SET-26, and given their differential roles in lifespan and heat resistance, we wondered whether these two proteins could be expressed in different tissues. In an attempt to resolve the expression patterns of SET-9 and SET-26, we previously used RT-PCR at precise temperatures, as well as an antibody that recognized both SET-9 and SET-26, in wild-type, set-26 single, and germlineless mutant worms, and deduced that SET-26 is likely broadly expressed and SET-9 is likely expressed in the germline (Ni et al., 2012). To unambiguously determine the expression patterns of the SET-9 and SET-26 proteins, we used CRISPR-cas9 to knock-in a GFP tag at the C-terminus of the endogenous set-9 and set-26 loci and monitored their expression patterns. Consistent with our previous report (Ni et al., 2012), we found that GFP-tagged SET-9 was only detected in germline cells of C. elegans (Figure 3A). In contrast, the GFP-tagged SET-26 was broadly expressed in both the somatic and germline cells (Figure 3B). As expected, expression of these two proteins was restricted to the nucleus, which is consistent with their possible roles in chromatin regulation. The ubiquitous expression of SET-26, but not SET-9, likely explains why SET-26 alone has a role in lifespan and heat resistance. Figure 3 with 2 supplements see all Download asset Open asset SET-26 is broadly expressed and SET-9 is only detectable in the germline. (A, B) Fluorescent micrographs of worms carrying gfp knock-in at the C-terminus of set-9 or set-26 gene (set-9::gfp and set-26::gfp). GFP-fused SET-26 was detected in all cells and GFP-fused SET-9 was only detected in the germline. Star indicates head, arrow indicates germline in the images. The signal outside of the germline detected in the set-9::gfp strain represented autofluorescence (marked by hashtag), which appeared yellow under the microscope. (C, D) Germline-specific knockdown of set-9 and set-26 was not sufficient to extend lifespan. RNAi Knockdown of set-9 and set-26 or wdr-5.1 extended lifespan in N2 worms (C). RNAi knockdown of wdr-5.1, but not set-9 and set-26, extended lifespan in the rrf-1(pk1417) mutant worms. Quantitative data are shown in Supplementary file 1 Table S3. https://doi.org/10.7554/eLife.34970.006 We noted that the knock-in worms expressing GFP-tagged SET-26 lived slightly longer than wild-type (but significantly shorter than the set-26 mutant) (Figure 3—figure supplement 1A) and had a slight heat resistance phenotype (Figure 3—figure supplement 1B), and the knock-in worms expressing both SET-9::GFP and SET-26::GFP had a slightly lower brood size compared to wild-type worms, but a significantly larger brood size than the set-9 set-26 double mutant worms (Figure 3—figure supplement 1C). The data together suggested that the GFP-tags somewhat compromise the functions of SET-9 and SET-26, but the tagged proteins remain largely functional. We next wondered whether the germline or somatic expression of SET-26 is important for lifespan modulation. We previously showed that RNAi knockdown of set-9/–26 (RNAi targets the two genes due to high sequence identity) in glp-1(e2141) germlineless mutant worms extended lifespan to a similar degree as in wild-type worms (Ni et al., 2012), suggesting that somatic set-26 is important for lifespan modulation. To further test this possibility, we used the rrf-1 mutant, in which RNAi is efficient in the germline but not somatic cells (Sijen et al., 2001), to assess whether knockdown of set-26 (and set-9) in the germline alone can extend lifespan. As a control for tissue-specific RNAi, we monitored SET-26::GFP expression in wild-type or rrf-1 mutant worms treated with set-9/–26 RNAi. As expected, set-9/–26 RNAi greatly reduced SET-26::GFP expression in most tissues except neurons in wild-type worms (Figure 3—figure supplement 1D), whereas set-9/–26 RNAi treatment specifically knocked down SET-26::GFP expression in the germline in rrf-1 mutant worms (Figure 3—figure supplement 1D). We next assessed the lifespan of wild-type or rrf-1 mutant worms treated with set-9/–26 RNAi. We included wdr-5.1 RNAi as a positive control as wdr-5.1 is known to act in the germline to modulate lifespan (Greer et al., 2010). As expected, RNAi knockdown of wdr-5.1 extended lifespan in both wild-type and rrf-1 mutant worms. In contrast, set-9/–26 RNAi knockdown extended lifespan in wild-type but not in the rrf-1 mutant background (Figure 3C and D), indicating that inactivation of set-26 (and set-9) in the germline is not sufficient for lifespan modulation. These results corroborated with our previous findings, and indicated that SET-26 likely acts in the somatic cells to modulate longevity and heat stress response, but SET-9 and SET-26 act redundantly in the germline to maintain reproductive function. Transcriptional profiling revealed candidate longevity and germline function genes regulated by SET-9 and SET-26 To gain insights into the molecular changes that may contribute to the somatic SET-26 effect on lifespan, we investigated the transcriptional profiles of the long-lived germlineless glp-1; set-26 double mutant. We isolated total RNA from glp-1; set-26 double and glp-1 single mutant worms, and performed RNA sequencing after removing ribosomal RNAs (ribo-minus RNA-seq). We next used edgeR, an RNA-seq analysis tool in the R package (Robinson et al., 2010), to identify the genes that showed statistically significant expression change in the glp-1; set-26 double mutant compared to glp-1 mutant (Figure 4A). We identified 887 up-regulated and 946 down-regulated genes in response to set-26 loss in the soma (Figure 4A), and gene ontology (GO) analyses indicated that these genes were over-represented in multiple functional groups (Figure 4—figure supplement 1A), with ‘collagen’ stood out as the most highly enriched GO term. We noted that collagens, as well as some of the other genes with expression change, have been implicated to be important for lifespan in C. elegans (Ewald et al., 2015). It would be interesting to test how altered expression of collagens, and other genes identified in our RNA-seq data, might contribute to the extended lifespan of the set-26 mutant in the future. Figure 4 with 1 supplement see all Download asset Open asset DAF-16-dependent somatic SET-26 regulated genes are enriched for lifespan determinant genes. (A) Venn diagrams show the overlap between up-regulated genes in glp-1(e2141); set-26(tm2467) (comparing with glp-1(e2141)) and down-regulated genes in daf-16(mgDf47); glp-1(e2141); set-26(tm2467) (comparing with glp-1(e2141); set-26(tm2467)); and the overlap between down-regulated genes in glp-1(e2141); set-26(tm2467) (comparing with glp-1(e2141)) and up-regulated genes in daf-16(mgDf47); glp-1(e2141); set-26(tm2467) (comparing with glp-1(e2141); set-26(tm2467)). (B) GO term analysis of DAF-16-dependent somatic SET-26 regulated genes. Gene lists can be found in Supplementary file 2. https://doi.org/10.7554/eLife.34970.009 Since we previously showed that somatic set-26 largely acts through daf-16, which encodes the Forkhead box O (FOXO) transcription factor, to modulate lifespan (Ni et al., 2012), we sought to further identify the transcriptional changes in response to somatic set-26 loss that are also dependent on daf-16. Using similar RNA-seq experiments, we investigated the transcriptional profiles of the germlineless daf-16; glp-1; set-26 triple and glp-1; set-26 double mutants (Figure 4A). We identified 164 genes that were up-regulated, and 131 genes that were down-regulated in the daf-16; glp-1; set-26 triple mutant (Figure 4A). By comparing these gene lists with the gene lists discussed above for the germlineless glp-1; set-26 double mutant vs. glp-1, we deduced the somatic genes whose expression become significantly up-regulated or down-regulated when set-26 is deleted, but those expression changes were reverted when daf-16 was simultaneously lost (down-regulated or up-regulated in the daf-16; glp-1; set-26 triple mutant, respectively) (Figure 4A). We termed these DAF-16-depednent somatic SET-26 regulated genes. Interestingly, GO term analyses revealed that the functional group ‘determination of adult lifespan’ was highly enriched in these DAF-16-dependent somatic SET-26 regulated genes (Figure 4B). Therefore, the transcriptomic analysis corroborated the genetic analysis, and supported a model that DAF-16-mediated gene regulation likely contributes to the lifespan phenotype of the set-26 mutant. We additionally investigated the transcriptional profiles of the long-lived fertile set-26 single mutant and revealed that 869 genes showed significant expression change in the set-26 mutant compared with wild-type worms. As expected, there was a significant and substantial overlap between the genes that exhibited expression change in response to whole-body loss of set-26 and the somatic SET-26 regulated genes discussed above (Figure 4—figure supplement 1B). Interestingly, the analysis using germlineless worms revealed far greater number of genes with expression change compared to that using reproductive worms. This could be due to technical variations between experiments, but might also suggest that some genes exhibit selective expression changes only in somatic cells, and those expression changes could be masked when germ cells were included in the analysis. To gain insights into the molecular changes that may underlie the germline phenotypes, we next compared the transcriptional profiles of the set-9 single, set-26 single, and F1 set-9 set-26 double mutants, all of which were fertile. We identified 162, 334, 1888 genes that were up-regulated, and 545, 534, 1644 genes that were down-regulated in the set-9, set-26, and F1 set-9 set-26 mutants respectively (Figure 5A). Interestingly, although there was significant overlap among the three gene sets, a substantial number of genes appeared to only show expression changes in the F1 set-9 set-26 double mutant (Figures 5A, 1430 down-regulated, 1781 up-regulated), suggesting a redundant role of SET-9 and SET-26 in regulating gene expression. Since the F1 set-9 set-26 mutant had a mild brood-size phenotype, and gave rise to progeny that exhibited severe defects in germline development (Figure 2), we speculated that many of these SET-9 and SET-26 co-regulated genes could be important for germline function. Indeed, GO analyses revealed that the genes that showed expression change in response to the simultaneous loss of set-9 and set-26 were over-represented for a number of different functional groups, including genes with annotated functions in sperm development and function (Figure 5B–C). We further compared these SET-9 and SET-26 co-regulated genes with genes previously determined to be germline-, oocyte-, and sperm-specific (Reinke et al., 2004). Interestingly, we found a significant over-representation of germline-specific genes among the genes that exhibited up-regulated expression in the F1 set-9 set-26 double mutant, but not the genes that exhibited down-regulated expression (Figure 5D). Both sperm- and oocyte-specific genes (383 and 252, respectively) were among these germline-specific genes that were up-regulated in the F1 set-9 set-26 double mutant. It is possible that up-regulated expression of these germline-specific genes contribute to the reproductive defects of the set-9 set-26 double mutant (Greer et al., 2014; Katz et al., 2009; Kerr et al., 2014). Figure 5 with 1 supplement see all Download asset Open asset Transcriptional profiles of set-9(rw5), set-26(tm2467), and F1 set-9(rw5) set-26(tm2467) mutants. (A) Venn diagrams show the overlap among set-9(rw5), set-26(tm2467), and set-9(rw5) set-26(tm2467) down-regulated (left) and up-regulated (right) gene sets. Hashtag indicates genes that only show expression change in the F1 set-9(rw5) set-26(tm2467) double mutant. GO term analysis of up-regulated (B) and down-regulated (C) genes that only show expression change in the F1 set-9(rw5) set-26(tm2467) double mutant. (D) Venn diagram shows the overlaps between genes that only show expression change in the F1 set-9(rw5) set-26(tm2467) double mutant identified in our RNA-seq data with the previously reported germline-specific gene lists (Reinke et al., 2004). Gene lists can be found in Supplementary file 2. https://doi.org/10.7554/eLife.34970.011 Considering the maternal effect of SET-9 and SET-26 on fertility (Figure 2C), we also profiled the transcriptome of the F3 set-9 set-26 double mutant, which exhibited greatly compromised fertility (Figure 2A), and compared that with the transcriptional profile of the F1 set-9 set-26 double mutant discussed above. We noted that the germline of the F3 set-9 set-26 was morphologically quite different from wild-type and the F1 set-9 set-26 double mutant (Figure 2—figure supplement 1). Interestingly, we found that the genes with significant expression change in the F1 set-9 set-26 and F3 set-9 set-26 double mutants not only substantially overlapped, but they were also enriched for similar functional groups based on GO term analyses (Figure 5—figure supplement 1A, B and C). The fold change of gene expression, compared to wild-type, in the F1 set-9 set-26 and the F3 set-9 set-26 double mutants also positively correlated (Figure 5—figure supplement 1D and E). These results together suggested that the transcriptional profiles of the F1 and the F3 set-9 set-26 double mutants are highly correlative despite that the F3 set-9 set-26 double mutant has a more severely defective germline. We note that the GO term ‘development and reproduction’ was unique for the F3 set-9 set-26 (Figure 5—figure supplement 1B), which may reflect the more severe germline defects in these mutant worms. Loss of SET-9 and SET-26 does not affect the global levels of H3K9me3 We next investigated the possible normal functions of SET-9 and SET-26, which could inform how their inactivations lead to the gene expression changes and biological phenotypes discussed above. The SET domain of SET-26 was recently reported to show H3K9me3 methylation activity in vitro (Greer et al., 2014). This was a somewhat surprising result, as the SET d