Combinatorial transcriptomic and genetic dissection of insulin/IGF‐1 signaling‐regulated longevity in Caenorhabditis elegans

Abstract Classical genetic analysis is invaluable for understanding the genetic interactions underlying specific phenotypes, but requires laborious and subjective experiments to characterize polygenic and quantitative traits. Contrarily, transcriptomic analysis enables the simultaneous and objective identification of multiple genes whose expression changes are associated with specific phenotypes. Here, we conducted transcriptomic analysis of genes crucial for longevity using datasets with daf‐2/insulin/IGF‐1 receptor mutant Caenorhabditis elegans. Our analysis unraveled multiple epistatic relationships at the transcriptomic level, in addition to verifying genetically established interactions. Our combinatorial analysis also revealed transcriptomic changes associated with longevity conferred by daf‐2 mutations. In particular, we demonstrated that the extent of lifespan changes caused by various mutant alleles of the longevity transcription factor daf‐16/FOXO matched their effects on transcriptomic changes in daf‐2 mutants. We identified specific aging‐regulating signaling pathways and subsets of structural and functional RNA elements altered by different genes in daf‐2 mutants. Lastly, we elucidated the functional cooperation between several longevity regulators, based on the combination of transcriptomic and molecular genetic analysis. These data suggest that different biological processes coordinately exert their effects on longevity in biological networks. Together our work demonstrates the utility of transcriptomic dissection analysis for identifying important genetic interactions for physiological processes, including aging and longevity.

analysis of genes crucial for longevity using datasets with daf-2/insulin/IGF-1 receptor mutant Caenorhabditis elegans.Our analysis unraveled multiple epistatic relationships at the transcriptomic level, in addition to verifying genetically established interactions.
Our combinatorial analysis also revealed transcriptomic changes associated with longevity conferred by daf-2 mutations.In particular, we demonstrated that the extent of lifespan changes caused by various mutant alleles of the longevity transcription factor daf-16/FOXO matched their effects on transcriptomic changes in daf-2 mutants.
We identified specific aging-regulating signaling pathways and subsets of structural and functional RNA elements altered by different genes in daf-2 mutants.Lastly, we elucidated the functional cooperation between several longevity regulators, based on the combination of transcriptomic and molecular genetic analysis.These data suggest that different biological processes coordinately exert their effects on longevity in biological networks.Together our work demonstrates the utility of transcriptomic dissection analysis for identifying important genetic interactions for physiological processes, including aging and longevity.

| INTRODUC TI ON
Epistasis is defined as the interaction of genetic changes in two or more genes, which are determined by analyzing phenotypes (Bateson et al., 1909;Phillips, 2008).Epistasis includes hierarchical relationships and non-hierarchical relationships, such as feedback and feedforward loops (Azpeitia et al., 2011).Classical epistasis analysis is useful for understanding the genetic basis of the signaling and molecular interactions underlying particular phenotypes, but it has several limitations.For example, phenotypic epistasis analysis usually requires extensive and subjective experiments and a large number of individuals to detect significant genetic interactions.
Analysis of polygenic and quantitative traits also presents challenges in identifying the number of genes that participate in specific interactions.In addition, classical epistasis analysis has difficulty analyzing non-binary phenotypes, including partial suppression (Avery & Wasserman, 1992).Therefore, effective strategies are required to overcome the limitations of phenotype-based classical epistasis analysis.
Transcriptomic analysis using microarray and/or RNA sequencing (RNA-seq) (Heller, 2002;Wang et al., 2009;Wilhelm & Landry, 2009) has been applied to systematically characterize the interactions of genes crucial for various biological processes such as neuronal activities, hypoxia-inducible factor 1-regulated responses to low oxygen levels, pathogenic avoidance, aging, and immunosenescence (Angeles-Albores et al., 2018;Ham et al., 2022;Kaletsky et al., 2016Kaletsky et al., , 2020;;Lee et al., 2021).Transcriptomic dissection of phenotypes offers several advantages over classical genetic epistasis analysis.For example, transcriptomic analysis is adequate for the identification of many genes whose expression changes influence specific phenotypes (Garber et al., 2011;Wang et al., 2009;Wilhelm & Landry, 2009).Transcriptomic dissection analysis also provides quantitative measurements of genetic interactions in complex phenotypes (Angeles-Albores et al., 2018;Evans et al., 2023).Thus, transcriptomic analysis of phenotypes represents an effective strategy for identifying multiple previously undetected genetic interactions.
Contrarily, the contribution of daf-16b to longevity is marginal (Kwon et al., 2010;Lee et al., 2001;Lin et al., 2001).Compared with extensive classical genetic analysis, transcriptomic analysis of aging and lifespan phenotypes in C. elegans remains largely unexplored.
In this study, we aimed at comprehensive dissection of genetic interactions within the IIS pathway by analyzing the transcriptome of double and triple mutants in the daf-2 mutant background.We verified established genetic interactions and sought to unveil unexpected epistatic relationships among downstream components in the IIS pathway.We revealed that the extent of lifespan changes caused by various daf-16 mutant alleles correlated with those of transcriptomic changes in daf-2 mutants.By analyzing RNA elements affected by different genetic interventions, we found that several genetic alterations affected specific subsets of structural and functional RNA elements in daf-2 mutants.Finally, we identified and experimentally validated several unexpected genetic interactions between longevity regulators, which act coordinately to contribute to longevity in daf-2 mutants.Overall, our study demonstrates that transcriptomic dissection analysis is useful for identifying genetic interactions for physiological processes and complements phenotypebased classical genetic approaches.

| Preprocessing of RNA-seq data for transcriptomic analysis
Alignment and quantification of RNA-seq data were performed by adopting parameters described in the guidelines of the Encyclopedia of DNA Elements (ENCODE) long RNA-Seq processing pipeline (ht tps:// w w w. encod eproj ec t .org / pipel ines/ ENCPL 0 02LPE/ ).

| Transcriptomic epistasis
The epistasis analysis of RNA-seq data is composed of WT animals (control), single mutants (A and B), and double mutants (AB).

| Global comparisons of gene expression levels
Confounding factors between different datasets were adjusted by upper quartile normalization followed by the Remove Unwanted Variation from RNA-Seq Data (RUVSeq) (v.1.22.0)(Risso et al., 2014).
In particular, the Remove Unwanted Variation Using Control Genes (RUVg) was used and genes with nominal p > 0.1 were defined as in silico empirical negative controls.After the removal, the counts were converted to regularized log (rlog) values (Love et al., 2014) for multiple dimensional scaling using two dimensions with relative Euclidian distances and a pairwise comparison matrix using Pearson's correlation coefficients.Overlapping targets of longevity regulators were compared by using UpSet plots (Lex et al., 2014).

| Gene set enrichment analysis
Biological terms enriched in genes of interest were identified by using the WormCatalog (WormCat) (Higgins et al., 2022;Holdorf et al., 2020).Global expression changes of the WormCat terms and target genes of certain transcription factors caused by the genetic inhibition of different components were represented as normalized enrichment scores (NES) by using GSEA (v.3.0)(Subramanian et al., 2005).Biomarkers of individual cell types defined by single cell RNA-seq data analysis of wild-type and daf-2 mutants (Preston et al., 2019) were used for the single cell level transcriptome analysis.Cumulative fractions with expression levels of genes of interest were also used for the global comparisons.R (v.4.1.0,http:// www.r-proje ct.org) was used for all the data plotting and statistical tests unless stated otherwise.
Reads that span junctions and reads within exons were counted.

| Reconstruction of networks using a consensus approach
From log 2 transformed fold changes of genes compared to controls, intensities of links between different genetic components were calculated by using the COnSensus Interaction Network InFErence Service (cosifer) (Manica et al., 2021) with following parameters: -no-standardize -samples_on_rows -index 0 -combiner summa.

| Lifespan assays
Lifespan assays were performed at 20°C on NGM plates seeded with OP50 for experiments with mutants as described previously with minor modifications (Jung et al., 2021).RNAi-mediated lifespan assays were performed at 20°C on NGM plates containing 1 mM of isopropyl β-D-1-thiogalactopyranoside (IPTG; Gold Biotechnology, St. Louis, MO, USA) and 100 μg/mL ampicillin (Thermo Fisher Scientific, Waltham, MA, USA) seeded with HT115 bacteria expressing double-stranded RNA targeting pfd-6.HT115 bacteria were cultured in Luria broth (LB) containing 100 μg/mL ampicillin overnight at 37°C.One hundred μL of bacteria was seeded on plates and incubated overnight at 37°C.One mM of IPTG was added and incubated at room temperature for 24 hours before use (Ham et al., 2022;Park et al., 2022).Briefly, synchronized prefertile young adult animals were transferred to plates containing 5 μM 5-fluoro-2'-deoxyuridine (FUDR; Sigma-Aldrich, MO, USA) to prevent progeny from hatching.All lifespan assays were conducted by at least two independent researchers, and a minimum of four plates were used for each condition, except for Figure 6d,e.Animals that ruptured, displayed internal hatching, or crawled off the plates were censored but included in the lifespan analysis as censored animals.
Statistical analysis of the lifespan data was conducted using online application of the survival analysis 2 (OASIS2, http:// sbi.poste ch.ac.kr/ oasis2) (Han et al., 2016).p values were calculated using the log-rank (Mantel-Cox) test.
Collectively, these results suggest that transcriptomic epistasis analysis is useful for confirming established genetic interactions and suggesting unexpected relationships among genes.
Three functionally null or strong loss-of-function alleles, mgDf50, mgDf47, and mu86, delete parts of the a, b, and f isoforms (Figure 3a),  the genetic inhibition of which largely suppresses the longevity of daf-2 mutants (Chen et al., 2015;Heimbucher et al., 2015;Kumar et al., 2015;Lin et al., 2018;Riedel et al., 2013).Similarly, mg54, a nonsense mutation that disrupts daf-16a and daf-16f (Figure 3a), suppresses the longevity of daf-2 mutants (Chen et al., 2015).In contrast, tm5030 and tm5032, which delete parts of the exon specific to daf-16a (Figure 3a), partially suppress the longevity of daf-2 mutants, whereas tm6659, which deletes daf-16f (Figure 3a), has a small effect on the longevity of daf-2 mutants (Chen et al., 2015).We sought to relate transcriptomic changes caused by these seven daf-16 mutant alleles to effects on the longevity of daf-2 mutants (Figure 3b).mgDf50 elicited the greatest impact on the transcriptome, and the changes were distinct from those caused by the other daf-16 alleles (Figures 2 and 3b).This is likely caused by shorter read length and older version of sequencing machines used for generating mgDf50 data than data with the other daf-16 mutant alleles (Appendix S1).We therefore excluded mgDf50 as an outlier from our subsequent comparisons among daf-16 mutant alleles.The effects of the other daf-16 mutant alleles on the transcriptome were positioned along the trajectory formed by daf-16(mgDf47); daf-2 and daf-2 mutations as follows: mgDf47 > mu86 ≈ mg54 > tm5030 ≈ tm5032 > tm6659 > control (Figure 3b).We also identified strong correlations (absolute r = 0.89) between the extents of transcriptomic changes and those of lifespan changes caused by the analyzed six daf-16 mutant alleles in daf-2 mutants (Tables S1 and S2); mgDf50 had no corresponding lifespan data (Kumar et al., 2015).Thus, we used mgDf47 as the basis for further analyses of daf-16 mutant alleles because it had the greatest effect on the transcriptome among the six mutant alleles.We determined the overall magnitude of gene expression changes caused by each daf-16 mutant allele, which was upregulated or downregulated in daf-2 mutants compared to that in mgDf47 mutants (Figure 3c; Figure S3).We found that the extent of gene expression changes caused by mu86 and mg54 was similar to that caused by mgDf47 (Figure 3c; Figure S3c-e).These data are consistent with the strong allelic nature of these mutations and the marginal contribution of daf-16b to gene expression changes (Kwon et al., 2010).We found that the overall extent of changes caused by tm5030 and tm5032 was relatively small compared to that caused by mgDf47 (Figure 3c; Figure S3f,g).tm6659 had the smallest impact on the expression of target genes affected by mgDf47 (Figure 3c; Figure S3h).
Together, these results indicate that the extents of lifespan changes caused by these daf-16 mutant alleles correlate with the levels of transcriptomic changes in daf-2 mutants.

| Transcriptomic analysis indicates that daf-16/FOXO mutant alleles exert generally common effects on biological processes in daf-2 mutants
To assess the effects of multiple daf-16 mutant alleles on the transcriptome of daf-2 mutants, we next conducted gene set enrichment analysis (GSEA) (Subramanian et al., 2005) based on WormCat, an annotation database of C. elegans genome-scale data (Higgins et al., 2022;Holdorf et al., 2020), at intermediate and low levels (categories 2 and 3) (Figure 3d; Appendix S5).We commonly detected upregulation of the term "Stress response, Oxidative" in all comparisons.The terms "Stress response, Heavy metal" and "Metabolism, Insulin" were upregulated in daf-2 mutants compared to the majority of daf-16; daf-2 mutants except for the animals that contained tm6659, the weakest daf-16 mutant allele.These data suggest that the daf-16f isoform elicits upregulation of the oxidative stress response but not the response to heavy metal stress or insulin metabolism.
Three functionally null alleles, mgDf50, mgDf47, and mu86, tended to downregulate the term "Extracellular material, Collagen", consistent with the pro-longevity roles of collagens (Ewald et al., 2015), whose upregulation may require daf-16a, daf-16b, and daf-16f.We also detected common downregulation of the terms "Metabolism, Amino acid" and "Ribosome, Subunit" in daf-2 mutants compared with the majority of the analyzed daf-16; daf-2 mutants except for the animals carrying the strongest allele mgDf50.These results are consistent with reports indicating that reduced translation and changes in amino acid levels contribute to reduced IIS-mediated longevity (Depuydt et al., 2013;Edwards et al., 2015).In addition, "Stress response, Pathogen" and "Stress response, C-type lectin" were downregulated in daf-2 mutants compared with the majority of daf-16; daf-2 mutants.These data are consistent with the essential but complex role of immunity in the longevity conferred by upregulation of DAF-16/FOXO in daf-2 mutants (Lee et al., 2021;Park et al., 2021;Podshivalova et al., 2017;Wu et al., 2019).
We additionally conducted GSEA using single cell RNA-seq data obtained with wild-type and daf-2 mutants (Preston et al., 2019), to analyze the effects of multiple daf-16 mutant alleles on the transcriptome of daf-2 mutants at a single cell level (Figure 3e).We commonly detected the enrichment of the cell type "001_Sperm" in almost all the comparisons."013_Cholinergic DA motor neurons" was enriched in the dataset with tm6659, but depleted in that with the tm5030 and tm5032 alleles, suggesting the specific roles of daf-16f in these neurons for contributing to physiological changes in daf-2 mutants.Gene expression in various other cell and tissue types, including FLP neurons, the intestine, hypodermis, pharynx, muscle, seam cells, and germline, was enriched in daf-2 mutants compared with the majority of daf-16; daf-2 mutants except for the mutant animals that contained the weakest tm6659 mutant allele.These data suggest that daf-16a alters gene expression in more diverse cell types in daf-2 mutants than daf-16b or daf-16f does.Overall, our transcriptomic analysis revealed the differential effects of individual daf-16 mutant alleles on the transcriptome of daf-2 mutants at various cell types, which also correlate with phenotypic consequences, including lifespan changes.Consistent with our analysis shown in Figure 2, we found that the expression of DAF-16-induced genes was higher in daf-2 mutants than in hel-1; daf-2, daf-2; daf-18(yh1), daf-2; daf-18(nr2037), daf-2; math-33, and daf-2; swsn-1 mutants (Figure 4a).In addition, the expression of DAF-16-induced genes was higher in daf-2 mutants than in pfd-6; daf-2 mutants.These data are consistent with established relationships between the five factors and DAF-16 (Gil et al., 1999;Heimbucher et al., 2015;Mihaylova et al., 1999;Ogg & Ruvkun, 1998;Riedel et al., 2013;Seo et al., 2015;Son et al., 2018).Together, HEL-1, DAF-18, MATH-33, SWSN-1, PFD-6, and DAF-16 appear to cooperate for common gene induction.

| Transcriptomic analysis of gene manipulation with respect to four representative longevity transcription factors in the IIS pathway
We also demonstrated that the expression of SKN-1-induced genes was higher in daf-2 mutants than in hel-1; daf-2 and pfd-6; daf-2 mutants (Figure 4a), suggesting the cooperativity of HEL-1, PFD-6, and SKN-1 in gene induction.Our current analysis recapitulated a previous report (Park et al., 2021) revealing the downregulation of SKN-1-induced genes in daf-2 mutants compared with daf-2; daf-18(nr2037) mutants but not with daf-2; daf-18(yh1) mutants (Figure 4a).In addition, many genes that were induced by HLH-30 in daf-2 mutants were suppressed by daf-18(yh1), daf-18(nr2037), and swsn-1 mutations (Figure 4a), and therefore HLH-30 appears to cooperate with DAF-18 and SWSN-1 in gene induction.This is consistent with previous reports showing the positive relationships between DAF-18 and DAF-16 (Gil et al., 1999;Mihaylova et al., 1999;Ogg & Ruvkun, 1998;Park et al., 2021) and between HLH-30 and DAF-16 (Lin et al., 2018).We found that the increased expression of HSF-1-induced genes in daf-2 mutants was enriched in genes whose expression was upregulated by PFD-6 (Figure 4a), which acts downstream of HSF-1 (Son et al., 2018).Our analysis indicated that the induction of genes by HSF-1/HSF1 was upregulated by SPR-3/SPR-4, and HIS-71/HIS-72 as well (Figure 4a), consistent with the role of HSF-1 in chromatin dynamics (Labbadia & Morimoto, 2015).We also identified unexpected relationships among the four representative longevity-associated transcription factors.We found that the expression of DAF-16-and SKN-1-induced genes was reduced in daf-2 mutants compared to that in daf-2; hlh-30 mutants (Figure 4a), raising the possibility of the distinct role of HLH-30 opposed to DAF-16 and SKN-1 in transcriptional regulation.Our analysis also raises the possibility that HSF-1 is mutually antagonistic to DAF-16 and HLH-30 in overall gene induction (Figure 4a).Collectively, these four established longevity transcription factors appear to work with specific subsets of factors in the IIS pathway to mediate the effects of daf-2 mutations on physiological processes, including aging.

| WormCat analysis of the transcriptomes affected by the inhibition of longevity genes in IIS
To assess the effects of the genetic interventions on the transcriptome of daf-2 mutants, we further compared the expression changes of genes associated with WormCat (Higgins et al., 2022;Holdorf et al., 2020) terms at intermediate and low levels (categories 2 and 3) caused by different genetic interventions (Figure 4b; Appendix S6).
Together these data suggest that the majority of different genetic interventions affect certain signaling pathways to suppress longevity conferred by daf-2 mutations.

| Analysis of intron-derived and non-coding RNAs
In our previous report regarding systematic transcriptome analysis for identifying aging biomarkers using wild-type and daf-2 mutants, q-values were obtained by calculating the false discovery rate corresponding to each NES.Representative terms were selected based on minimum q-values (q = 0.05) among different conditions.The terms were further selected if at least four conditions showed absolute NES greater than 1.5.Terms were sorted based on the number of upregulated terms.(e) NES of transcriptomic changes for biomarkers elicited by various daf-16 mutant alleles in indicated cells (Preston et al., 2019).NES range from −3 (blue) to +3 (red).The size of black circles correlates with −log 10 (q-value).References for datasets with the same mutant alleles were indicated as follows: Riedel mgDf47 (Riedel et al., 2013), Lin mgDf47 (Lin et al., 2018), Chen mu86 (Chen et al., 2015), and Heimbucher mu86 (Heimbucher et al., 2015).
we revealed that intron-and intergenic region-derived transcripts, noncoding RNAs, and the usage of distal 3′ splice sites in mRNA transcripts are upregulated during aging (Ham et al., 2022).We therefore determined whether these biomarkers of aging were affected by the genetic inhibition of each of the 11 genetic factors, which decreases the longevity of daf-2 mutants.However, none of the genetic inhibitions significantly affected intron-derived transcripts (Figure 5a).In contrast, intergenic region-derived transcript levels were increased in daf-2 mutants compared with daf-2 his-72; his-71 mutants but were decreased in daf-2 mutants compared with smg-2; daf-2 mutants (Figure 5b).These data suggest that HIS-71 and HIS-72 regulate the transcription of non-annotated genes.In addition, SMG-2 may contribute to longevity conferred by daf-2 mutations by suppressing the expression of intergenic regionderived transcripts.Additionally, the levels of long intergenic noncoding RNAs (lincRNAs) were lower in daf-2 mutants than in smg-2; daf-2 and spr-4; daf-2; spr-3 mutants (Figure 5c).These data raise the possibility that SMG-2 and REST contribute to longevity in daf-2 mutants by suppressing the induction of lincRNAs.
We next analyzed the changes of alternative splicing caused by genetic inhibition of each of the 11 factors in daf-2 mutants.Among them we found that his-72; his-71 mutations elicited the greatest impact on alternative splicing.Specifically, his-72; his-71 increased the levels of skipped exons and the usage of distal 3′ splice sites in daf-2 mutants while decreasing those of retained introns (Figure 5d,f,h).
These data suggest that HIS-71/HIS-72 inhibits the usage of distal 3′ splice sites to promote longevity in daf-2 mutants.Different from the effects of his-72; his-71 mutations, pfd-6, smg-2, and spr-4; spr-3 mutations decreased the levels of skipped exons (Figure 5d).In addition, pfd-6 mutations decreased the usage of proximal 5′ splice sites and increased the levels of retained introns (Figure 5e,h).smg-2 mutations decreased the usage of distal 3′ splice sites (Figure 5f).spr-4; spr-3 mutations also decreased the usage of distal 3′ splice sites but increased the levels of retained introns (Figure 5f).These data imply The effects of four representative longevity-promoting transcription factors that act downstream of daf-2 mutants on transcriptomic changes.(a) NES of transcriptomic changes of target genes of representative longevity transcription factors acting in the IIS: DAF16/FOXO (Riedel et al., 2013), SKN-1/NRF (Ewald et al., 2015), HLH-30 (Lin et al., 2018), and HSF-1/HSF1 (Lee et al., 2021).q-values were obtained by calculating the false discovery rate corresponding to each NES.Datasets were sorted based on hierarchical clustering.(b) NES of transcriptomic changes in indicated WormCat terms at intermediate levels by the mutations analyzed in this study.q-values were obtained by calculating the false discovery rate corresponding to each NES.Representative terms were selected based on minimum q-values (q = 0.05) among different conditions.The terms were further selected when at least five conditions showed absolute NES greater than 1.5 or at least six conditions showed absolute NES greater than 1.25.Terms were sorted based on hierarchical clustering.NES range from −3 (blue) to +3 (red).The size of black circles correlates with −log 10 (q-value).3.8 | Several longevity regulators that were identified with our transcriptome analysis act together to promote the longevity of daf-2 mutants We sought to identify novel genetic interactions between the pairs of longevity regulators acting downstream of DAF-2/insulin/IGF-1 receptor.We first functionally analyzed the relationship between hlh-30/TFEB and smg-2/UPF1, as mutations in these genes induced transcriptomic changes with the highest similarity among previously uncharacterized interactions in the daf-2 mutant background (interaction intensity = 0.96) (Figure 6a; Figure S4; Appendices S7 and S8).
In addition, we analyzed the interaction between hlh-30/TFEB and pfd-6/PFDN6 because pfd-6 have a higher similarity with hlh-30 (interaction intensity = 0.89) than daf-16 (interaction intensity = 0.19); DAF-16 activity as a transcription factor was enhanced by PFD-6 (Son et al., 2018) (Figure 6a; Figure S5; Appendices S7 and S9).We demonstrated that hlh-30; smg-2 double mutations had slightly stronger suppressive effects on the longevity of daf-2 mutants than the single mutations (Figure 6b).These data suggest that SMG-2/ UPF1 and HLH-30/TFEB regulate common targets, but the extent of regulation by each factor is partially additive.Importantly, we found that the long lifespan conferred by hlh-30 overexpression (Lapierre et al., 2013;Lin et al., 2018) was suppressed by smg-2 mutations (Figure 6c).Conversely, longevity conferred by the overexpression of SMG-1 (Son et al., 2017), which phosphorylates and activates SMG-2/UPF1 to enhance nonsense-mediated mRNA decay (NMD) (Kim & Maquat, 2019;Kwon et al., 2023;Son & Lee, 2017), was suppressed by hlh-30 mutations (Figure 6d).Therefore, hlh-30 and smg-2 act on substantially overlapping transcriptomic changes and promote longevity in concert.We also demonstrated that hlh-30 mutations did not further shorten the lifespan of daf-2 mutants treated with pfd-6 RNAi (Figure 6e).Moreover, we found that the long lifespan conferred by hlh-30 overexpression (Lapierre et al., 2013;Lin et al., 2018) was suppressed by pfd-6 mutation (Figure 6f).Thus, hlh-30 and pfd-6 also appear to promote longevity together by regulating common target genes.These data raise the possibility of unexpected F I G U R E 6 SMG-2/UPF1, HLH-30/ TFEB, and PFD-6/PFDN6 act coordinately to promote longevity caused by daf-2 mutations.(a) A network of various longevity factors reconstructed by using a consensus approach based on correlation coefficients of transcriptome data.The widths and darkness of lines that connect datasets correlate with the intensities of the links from 0.8 (gray) to 1 (black).through deubiquitination (Heimbucher et al., 2015), our finding implies that MATH-33 downregulates DAF-2.The quality-control ubiquitin ligase CHIP triggers the turnover of DAF-2 for affecting IIS and consequently lifespan (Tawo et al., 2017).The stability of CHIP is affected by other ubiquitin ligases and deubiquitylation enzymes (Hohfeld & Hoppe, 2018).Thus, these studies and our current work raise the possibility that MATH-33 contributes to the stability of CHIP that downregulates DAF-2.In addition, our results suggested that HIS-71/HIS-72, SPR-3/SPR-4, and SMG-2 act with DAF-2 in the same pathway and that HEL-1 upregulates DAF-2 via feedforward signaling.These data support the requirement of multifaceted and comprehensive analysis to properly understand longevity regulators in the IIS pathway.
Additionally, we noted that only 12% of genes that were downregulated by spr-4; spr-3 mutations in daf-2 mutants at Day 10 of adulthood overlapped with class I DAF-16 target genes, which are defined as genes upregulated in daf-2 mutants in a DAF-16dependent manner (Murphy et al., 2003).We noted that the roles of PFD-6 is entangled with different regulators.PFD-6 appears to work with SKN-1 and HSF-1 as well as DAF-16.In fact, 66% of genes that were downregulated by pfd-6 mutations did not overlap with those downregulated by daf-16 mutations in daf-2 mutants (Son et al., 2018).Our data also raise another possibility that pfd-6 and hlh-30 act on substantially overlapping transcriptomic changes and promote longevity in concert.Overall, these data suggest that the established proteins crucial for the longevity of daf-2 mutants play distinct roles in the regulation of DAF-16.
We analyzed RNA-seq data obtained from animals with genetic inhibition of factors crucial for longevity attributable to daf-2 mutations in C. elegans.However, transcriptomic changes caused by the genetic inhibition of these factors are possibly associated with altered responses to heat shock and oxidative stress and/or dauer formation independently of longevity.In addition, dissecting the difference between chronological and physiological ages at the transcriptomic level is required to understand the mechanisms underlying each of these physiological processes (Ham et al., 2022).Furthermore, because we analyzed RNA-seq data obtained from whole bodies, we should be cautious about potential caveats, including tissue heterogeneity, signal dilution, and biological noise.Notably, daf-2 mutations contribute to longevity by affecting particular longevity regulators in a tissue-specific manner (Kaletsky et al., 2016;Libina et al., 2003;Roy et al., 2022;Zhao et al., 2021).Tissue-specific analysis remains essential for a nuanced interpretation of the molecular mechanisms underlying longevity conferred by daf-2 mutations.Importantly, transcriptomic analysis reveals correlations in gene expression patterns, but does not directly prove causation.Therefore, identified relationships between genes in a pathway should be further validated using other experimental techniques, such as genetic manipulations and functional physiologic assays.To identify transcriptomic changes specifically associated with longevity and their causality, future research aiming at integrative analysis using other omics data obtained using ChIP-seq, CLIP-seq, single cell RNA-seq (Hwang, 2023;Kim, 2023;Kim & Lee, 2023;Ryu et al., 2023), and mass spectrometry will be necessary.

F
Transcriptomic analysis of seven selected daf-16 mutant alleles in daf-2 mutant backgrounds.(a) Schematic diagram of the daf-16 mutant alleles along daf-16 gene (black) and daf-16 transcript isoforms (gray).Colors of the alleles indicate the extents of their impacts on phenotypic and transcriptomic changes: red > pink > orange > yellow.(b) A multidimensional scaling plot showing relative Euclidian distances among the seven daf-16 mutant alleles that we analyzed in daf-2 mutant backgrounds at transcriptome levels.(c) Cumulative fraction of the genes in an ascending order of the extent of gene expression changes conferred by the daf-16 mutant alleles in daf-2 mutant backgrounds.daf-16 isoforms that are affected by mutant alleles are indicated.(d) Normalized enrichment scores (NES) of transcriptomic changes in indicated WormCat terms at intermediate levels caused by various daf-16 mutant alleles.
Analysis of representative non-coding RNAs and splicing of mRNAs affected by genetic inhibitions that suppress the longevity of daf-2 mutants.(a-c) Overall changes in the levels of intron-derived transcripts (a), intergenic region-derived transcripts (b), and long intergenic noncoding RNA (lincRNA) (c) by mutations in the longevity-promoting genetic factors analyzed in this study.p value of two-tailed Welch's t test is shown on top of each panel, *p < 0.05, **p < 0.01, ***p < 0.001.Genes whose mutations elicited significant changes are indicated in bold in this figure.(d-h) The number of splicing events, including skipped exons (d), proximal 5′ splice sites (e), distal 3′ splice sites (f), mutually exclusive exons (usage of distal exons) (g), and retained introns (h), which were upregulated or downregulated by the mutations analyzed in this study.p values of two-tailed Fisher's exact test for counts are shown on top of each panel, *p < 0.05, **p < 0.01, ***p < 0.001.
HIS-71/HIS-72, PFD-6, SMG-2, and SPR-3/SPR-4 in alternative splicing.Overall, our data suggest that different genetic interventions affect specific subsets of structural and functional elements of RNAs to mediate the effects of daf-2 mutations on aging.
longevity regulators to delay aging and to confer longevity.4 | DISCUSS ION Transcriptomic analysis is useful for identifying novel and multiple genetic interactions that elicit phenotypic consequences.In the current work, we analyzed RNA-seq data featuring the genetic inhibition of various aging-regulating factors that contribute to longevity conferred by daf-2 mutations in C. elegans.Our transcriptomic analysis verified established genetic interactions and unraveled the possibilities of unexpected epistatic relationships of downstream components in the IIS pathway.Our results also indicate that the extent of lifespan changes caused by various daf-16 mutant alleles correlates with the extent of transcriptomic changes in daf-2 mutants.In addition, different genetic interventions appear to affect specific subsets of structural and functional RNA elements in daf-2 mutants.Lastly, we identified genetic interactions between longevity regulators, hlh-30 and smg-2, and pfd-6 and hlh-30, which cause common transcriptomic changes for mediating longevity conferred by daf-2 mutations in concert.Overall, our combinatorial transcriptomic analysis proved useful for determining novel and multiple genetic interactions and for overcoming the limitations of phenotype-based classical genetic approaches.Our unbiased analysis revealed the unexpected actions regarding longevity regulators in daf-2 mutants.Different from a previous study suggesting the role of MATH-33 in the stability of DAF-16 addition to the mRNA analysis, our work demonstrated that different genetic interventions affected specific subsets of intronderived or non-coding RNAs in daf-2 mutants.However, because of the technical limitations of publicly available RNA-seq data, we were not able to analyze non-coding RNAs in depth or to examine whether genetic factors whose depletion decreases longevity caused by daf-2 mutations contribute to universal changes in non-coding RNA levels.Future research is required to overcome these technical constraints by employing advanced sequencing technologies or generating targeted datasets.
approaches have the potential to facilitate transcriptomic analyses of other genetic interventions that contribute to longevity, such as reduced mTOR signaling and mild inhibition of mitochondrial respiration in C. elegans and other model organisms.Moreover, our work can serve as a guideline for the future application of analyses of transcriptomic changes to provide insights into aging and age-associated diseases in humans.AUTH O R CO NTR I B UTI O N S Seokjin Ham: Conceptualization; data curation; formal analysis; investigation; visualization; methodology; writing-original draft.Sieun S. Kim: Formal analysis; validation; methodology; writingoriginal draft.Sangsoon Park: Formal analysis; validation.Hyunwoo C. Kwon: Formal analysis; validation.Seokjun G. Ha: Formal analysis; validation.Yunkyu Bae: Data curation; writing-original draft.Gee-Yoon Lee: Investigation.Seung-Jae V. Lee: Conceptualization; supervision; funding acquisition; investigation; writing-original draft; project administration.