Ahr-bacterial diet interaction modulates aging and associated pathologies in C. elegans

Abstract

Genetic and environmental factors concurrently shape the aging process. The nematode Caenorhabditis elegans has been instrumental in the past 30 years to identify most factors and pathways nowadays known to modulate aging. The aryl-hydrocarbon-receptor (AhR) is a transcription factor originally discovered as a dioxin-binding protein and involved in the metabolism of different environmental toxicants and xenobiotics. Since its discovery, the variety of pathophysiological processes regulated by the AhR has grown and previously we disclosed it promotes aging phenotypes across species. Here, we followed up on our previous study and show that lack of C. elegans AHR-1 differentially affects health and lifespan in a context-dependent manner. Notably, we identified a critical role for AhR-bacterial diet interaction in animal lifespan, stress resistance, and age-associated pathologies, and narrowed down the dietary factor involved in the observed AhR-dependent features to a bacteria extruded product, likely derived from tryptophan metabolism.

Introduction

Aging affects every human and is accompanied by increased morbidity (e.g., diabetes, cardiovascular diseases, neurodegenerative diseases, and cancer) and risk of death (Kaeberlein, 2016, Lopez-Otin, Blasco et al., 2013). In the past decades, also thanks to the growing number of researchers exploiting simple but powerful model organisms such as the nematode Caenorhabditis elegans (C. elegans), numerous hallmarks of aging as well as genetic and environmental factors regulating it have been identified (Dato, Rose et al., 2017, Lopez-Otin et al., 2013, Tigges, Krutmann et al., 2014).

The aryl hydrocarbon receptor (AhR) is a highly conserved transcription factor of the bHLH PAS family originally discovered as a dioxin-binding protein (Poland, Glover et al., 1976) and involved in the metabolism of different environmental toxicants and xenobiotics. Since its discovery, the variety of pathophysiological processes regulated by the AhR has grown and range from cell death, to immune response and neuronal development (Esser, Lawrence et al., 2018). Contradictory studies also indicated a role for AhR in the aging process (reviewed in (Brinkmann, Ale-Agha et al., 2020)) and more recently, pro-aging functions of AhR (Eckers, Jakob et al., 2016, Williams, Mouchiroud et al., 2014) have been described in an evolutionarily conserved manner from C. elegans to mammals. In mammals, in basal conditions, AhR is bound by HSP90, AIP, and p23, which retain it in the cytoplasm, in a ligand-affine state. Ligand binding of an AhR agonist leads to the dissociation of the AhR binding complex and AhR nuclear translocation (Ashida, Nishiumi et al., 2008, Ikuta, Eguchi et al., 1998). In the nucleus, AhR dimerizes with the AhR nuclear translocator (Arnt), and the AhR-Arnt heterodimer then binds to the xenobiotic responsive elements (XREs) on AhR target genes (Ashida et al., 2008). Similar to its mammalian counterpart, the C. elegans AhR homolog, AHR-1, forms a heterodimer with the C. elegans Arnt homolog AHA-1 and binds to XREs (Powell-Coffman, Bradfield et al., 1998). It is expressed in several types of neurons and plays a role in neurodevelopment and long-chain unsaturated fatty acid synthesis (Aarnio, Storvik et al., 2010, Huang, Powell-Coffman et al., 2004, Qin & Powell-Coffman, 2004, Qin, Zhai et al., 2006, Smith, O’Brien et al., 2013). However, unlike mammalian AhR it does not bind to TCDD or β-naphthoflavone (Powell-Coffman et al., 1998), but no other modulators have been tested so far for their effect on C. elegans AHR-1 activity and downstream effects. Although AhR was originally discovered as a dioxin-binding protein (Poland et al., 1976) in more recent years many other compounds have been identified which modulate its activity and can be mainly divided into four categories: xenobiotics (e.g., 2,3,7,8-tetrachlorodibenzodioxin (dioxin) and benzo(a)pyrene (BaP)), dietary factors (e.g., kaempferol, and curcumin), endogenous modulators (e.g., 6-formylindolo[3,2-b]carbazole (FICZ) and kynurenine) and ligands generated by the microbiota metabolism (e.g., indole-3-acetate and tryptamine) (Amakura, Tsutsumi et al., 2003, Fritsche, Schafer et al., 2007, Jin, Lee et al., 2014, Lee, McPherson et al., 2017, Opitz, Litzenburger et al., 2011, Poland et al., 1976, Rannug, Rannug et al., 1987, Shimizu, Nakatsuru et al., 2000). Interestingly, many of these AhR modulators may impact on aging and age-associated pathologies (Alavez, Vantipalli et al., 2011, Caesar, Jonson et al., 2012, Lim, Chu et al., 2001), but whether these effects are AhR-dependent and whether AhR itself influences the progression of these diseases has been largely unexplored. Here, using C. elegans as a powerful model organism for aging and toxicology studies, we followed up on our previous finding indicating an anti-aging effect of C. elegans ahr-1 depletion (Eckers et al., 2016) and investigated the effect of possible AHR-1 modulators on aging and other age-related features, such as stress response and associated pathologies. Our findings indicate that C. elegans can be exploited to study evolutionarily conserved functions of AhR modulators and their regulated processes and therefore to contribute to the discovery of novel information about AhR in mammals. Most importantly, we identify aging and associated pathologies as novel life traits regulated by the AhR in a diet-dependent manner bringing further complexity to the landscape of AhR-microbiota regulated processes.

Results

C. elegans AHR-1 is differentially involved in stress response

Loss of AHR-1 extends C. elegans’ health- and lifespan (Eckers et al., 2016). Here we assessed the role of ahr-1 in different age-correlated features. Lifespan extension often correlates with increased resistance to different types of stressors (Cypser & Johnson, 2002). We assessed AhR role in stress resistance by exposing wild-type as well as ahr-1 mutants to either heat shock, metabolic stressors (i.e. high concentrations of glucose or the hypoxia mimetic iron chelator 2,2’Bipyridyl, BP) or UVB radiations. In agreement with the literature, the development and fertility of wild-type animals were affected by all insults in a dose-dependent manner (Figure 1) (Rieckher, Bujarrabal et al., 2018, Schiavi, Maglioni et al., 2015). Consistent with our previous data (Eckers et al., 2016) loss of ahr-1 increased heat-stress resistance (Figure 1 A, B), however, surprisingly, it did not confer resistance to any of the newly tested insults (Figure 1 C - H). Actually, ahr-1(ju145) mutants were even more sensitive to UVB radiations since their developmental delay and embryonic lethality were significantly more affected than in wild-type worms (Figure 1 G, H).

Figure 1. Loss of ahr-1 differentially regulates resistance to stressors.

A) Survival in response to heat shock. Curves show the pooled data of 60 worms in 3 independent replicates. Statistical test: Log-Rank test, * significance vs. wild-type. B) Fertility after heat stress. Shown are the number (left panel) and viability (right panel) of eggs laid from gravid adults treated with heat shock for the indicated time. Mean + SEM of pooled data from 18 worms/condition in 3 independent experiments are shown. C - D) Development and fertility in response to the indicated concentration of glucose. Means (+SEM) of 3 independent replicates are shown. N = number of individuals in panel C, 9 individuals were used in panel d. E – F) Development and fertility in response to the indicated concentration of iron chelator (BP). Means (+SEM) of 3 and 4 independent replicates are shown. N = number of individuals in panel E, 9 individuals were used in panel F. G – H) Development and fertility in response to indicated doses of UVB. Fertility was assessed at a dose of 600 J/m². Means (+SEM) of 3 independent replicates are shown. N = number of individuals in panel G, 9 individuals were used in panel H. B – H) Statistical test: 2-way ANOVA with Tukey’s multiple comparisons test, * significance vs. wild-type, # significance vs. control (untreated), p-value < 0.05.

Mammalian AhR typically regulates the expression of phase-I and phase-II detoxification enzyme genes like cyps (e.g., CYP1A1 or CYP1B1), ugts (e.g., UGT1A1 or UGT1A6) and gsts (e.g., GSTA1 or GSTA2) (Ashida et al., 2008, Xue, Li et al., 2017, Yueh, Huang et al., 2003). Thus, to further characterize animals’ stress response, we quantified the expression of C. elegans transgenic reporters for different detoxification genes in control and upon ahr-1 RNA interference (RNAi). Out of nine tested reporters, we found five genes (cyp-35A2, cyp-35B1, gst-4, cyp-37A1, and ugt-29) differentially expressed by at least ten percent upon ahr-1 RNAi (Figure EV1; Appendix Table S1). Unexpectedly, when crossed into the ahr-1(ju145) mutant background, we observed different effects for cyp-35B1 and gst-4 expression, while the expression of ugt-29 remained reduced in the ahr-1 mutant (Figure EV1 C, D, E), suggesting either different mode of action of ahr-1 silencing and mutation or tissue-specific effects disclosed by the RNAi treatment. In line with the differential effect between ahr-1 RNAi and mutant on cyp-35B1 and gst-4 expression, ahr-1 suppression by RNAi did not have any impact on the lifespan of the wild-type animals (Table 1; Figure EV1 G). Moreover, when ahr-1 RNAi was fed to different C. elegans strains, in which silencing is effective only or primarily in specific tissues (Calixto, Chelur et al., 2010, Espelt, Estevez et al., 2005, Kumsta & Hansen, 2012, Qadota, Inoue et al., 2007, Schmitz, Kinge et al., 2007, Sijen, Fleenor et al., 2001, Tijsterman, Okihara et al., 2002, Yigit, Batista et al., 2006), we observed that in most cases tissue-specific ahr-1 RNAi did neither affect health- or lifespan, which were instead significantly shortened only when RNAi was applied systemically but enhanced in the nervous system (Table 1; Figure EV1 F). Taken together these data reveal that a complex AHR-1 role in C. elegans’ lifespan and response to stress in a tissue- and insult-dependent manner.

Table 1. Survival analysis upon tissue-specific ahr-1 RNAi.

Mean life- and healthspans +/- SEM of pooled data are shown. The L4440 plasmid was used as a negative control. Statistical analysis was performed between ahr-1 and control RNAi using Log-rank test. The number of worms is specified as N, while the number of replicates is specified as n.

AHR-1 differentially regulates lifespan in response to potential modulators of its activity

AhR knockout in basal conditions revealed its role in aging and other physiological processes (Eckers et al., 2016, Mulero-Navarro & Fernandez-Salguero, 2016), yet, in mammals, AhR primarily exerts its functions in response to several different ligands and modulators of its activity, namely exogenous substances such as xenobiotics or dietary compounds, and endogenous products of metabolism such as tryptophan derivatives and microbiota-associated factors (Lee et al., 2017). Thus, we then investigated whether representative AhR modulators of each class affect C. elegans life traits and gene expression through AHR-1. The polyphenol curcumin is a dietary component well known to extend lifespan in C. elegans and can modulate AhR signaling in mammals (Amakura et al., 2003, Liao, Yu et al., 2011, Nishiumi, Yoshida et al., 2007). Consistent with a potential conserved activity through AhR, curcumin reproducibly and significantly extended life- and health-span in an ahr-1-dependent manner (Figure 2 A, B; Figure EV2 A). In human cancer cell lines, curcumin alters the expression of the AhR target genes Cyp1A1 and Cyp1B1 (Choi, Chun et al., 2008, Rinaldi, Morse et al., 2002). However, quantification of 47 C. elegans cyps by quantitative PCR revealed that only cyp-13B1 was significantly upregulated either by ahr-1 depletion or by curcumin in an ahr-1-dependent manner in C. elegans, while other two cyps (i.e. cyp-13A5 and cyp-13A8) were increased in an ahr-1-independent manner (Figure EV2 B - D). In mammals, AhR-mediated induction of cyps mediates some of the toxic effects of the xenobiotic BaP (Shimizu et al., 2000) and in C. elegans BaP induces cyp genes expression (Menzel, Bogaert et al., 2001). We observed that although BaP significantly increased the expression of the cyp-35B1 reporter strain and affected animals’ development in a dose-dependent manner, these effects were largely ahr-1-independent (Figure S3 A, B). Of note, we found that BaP treatment from adulthood, consistent with its toxic effects, significantly shortened animals’ health- and lifespan, with a more pronounced effect on ahr-1 mutants, indicating a protective role of AHR-1 against BaP-curtailed longevity (Figure 2 C, D).

Figure 2. AHR-1 displays evolutionarily conserved functions and affects aging in a diet-dependent manner.

A – B) Curcumin extends lifespan in an ahr-1-dependent manner. Pooled lifespan/healthspan curves of 290 (wt DMSO) and 300 (all other conditions) worms/condition in 5 independent experiments treated wither with DMSO or 100 µM of curcumin are shown. C – D) ahr-1 is more sensitive to xenobiotic stress. Pooled lifespan/healthspan curves of 120 worms/condition in 2 independent replicates treated with either DMSO or 5 µM BaP from adulthood are shown. E – F) ahr-1 is more sensitive to UVB stress. Pooled lifespan/healthspan curves of 300 (wt control, ahr-1 control), 180 (wt UVB) and 178 (ahr-1 UVB) worms/condition in 3 independent replicates either left untreated or treated with 1200 J/m² UVB from adulthood are shown. G – H) AHR-1 affects aging in a diet-dependent manner. Pooled lifespan/healthspan curves of 170 (wt OP50) and 180 (all other conditions) worms/condition in 3 independent replicates grown either on HT115 or OP50 are shown. A – H) Statistical test: Log-Rank test, # significance vs. control/HT115, * significance vs. wt, p-value < 0.05.

As opposed to the exogenous modulators, 6-formylindolo[3,2-b]carbazole (FICZ), a photoproduct of tryptophan produced in response to UVB light, is a well-described AhR endogenous high-affinity ligand (Fritsche et al., 2007, Rannug et al., 1987). In C. elegans UVB radiation shortens lifespan, induces embryonic lethality and germline cells apoptosis in a dose-dependent manner (Lans, Marteijn et al., 2010, Torgovnick, Schiavi et al., 2018). In Fig 1 we showed that ahr-1(ju145) embryos derived from irradiated meiotic cells - those prompted to undergo apoptosis - are more sensitive to UVB-induced embryonic lethality. In line with these results, we found that UVB-induced germline cells apoptosis was significantly more increased in the ahr-1 mutant’s germline compared to wild-type animals (Figure S3 C). UVB irradiation increases Cyp1A1 expression in human keratinocytes in an AHR-dependent manner (Fritsche et al., 2007). Interestingly, similar to the above results with BaP, although UVB tent to increase cyp-35B1 expression in an ahr-1-independent manner (Figure S3 D), it decreased animals’ lifespan and motility with a significantly stronger effect on the ahr-1(ju145) compared to wild-type animals (Figure 3 E, F). These results agree with the notion that loss of AhR sensitizes mammalian cells to UVB-induced apoptosis (Frauenstein, Sydlik et al., 2013) and support an evolutionarily conserved role of AhR in response to UVB.

Figure 3. AHR-1 mutants increase aggregation but extend lifespan in a diet-dependent manner.

A – B) Kaplan Meier curves of polyQ;wt and polyQ;ahr-1 of 180 worms/condition in 3 independent experiments are shown. * p-value < 0.05 vs. polyQ;wt, # p-value < 0.05 vs. HT115, statistical test: Log-rank test. C) Representative fluorescence images of 10-days old polyQ;wt and polyQ;ahr-1 on HT115 and OP50. D) Quantification of aggregates in 10-days old polyQ;wt and polyQ;ahr-1. Mean + 95 % CI of pooled data from 34 (wt HT115), 29 (wt OP50), 26 (ahr-1 HT115), and 35 (ahr-1 OP50) worms in 3 independent replicates is shown. Statistical test: One-way ANOVA with Tukey’s multiple comparisons test, * p-value < 0.05 vs. polyQ;wt, # p-value < 0.05 vs. HT115. E – F) Kaplan Meier curves of a-syn;wt and a-syn;ahr-1 of 120 worms/condition in 2 independent experiments are shown. * p-value < 0.05 vs. a-syn;wt, # p-value < 0.05 vs. HT115, statistical test: Log-rank test. G) Representative fluorescence images of the head muscles of 7-days old a-syn;wt and a-syn;ahr-1 on HT115 and OP50. H) Quantification of aggregates in 7-days old a-syn;wt and a-syn;ahr-1. Mean + 95 % CI of pooled data from 77 (wt HT115), 82 (wt OP50), 88 (ahr-1 HT115), and 92 (ahr-1 OP50) worms in 3 independent replicates is shown. Statistical test: One-way ANOVA with Tukey’s multiple comparisons test, p-value < 0.05 vs. a-syn;wt, # p-value < 0.05 vs. HT115.

Finally, to assess the possible interaction of AhR with microbiota-associated factors we took advantage of two commonly used Escherichia coli strains used as a food source for C. elegans, HT115(DE3) and OP50(xu363). We observed that, consistent with the growing body of evidence indicating a crosstalk between microbiota and AhR in mammals, the activation of a cyp-35B1-GFP reporter was reduced in the absence of ahr-1 when animals were fed HT115 but not OP50 (Figure EV3 F). Strikingly, we found that the beneficial effects on lifespan, motility, pharyngeal pumping, and heat resistance elicited in the ahr-1 mutants fed HT115 are completely abolished when animals are fed OP50 (Figure 3 G, H; Figure EV3 E, G). These differences are neither due to extrinsic bacterial differences nor to the effect of the two different bacteria on the ahr-1 expression (Figure EV3 H) thus likely relying on the modulation of AHR-1-regulated processes.

Results described so far indicate that known AhR modulators differentially affect C. elegans lifespan in an ahr-1-dependent manner but that classical detoxification related genes (cyps) are possibly not the major targets of C. elegans AHR-1. Thus, to gain insight into molecular mechanisms of AhR-regulated stress response and aging we turned to the results of a transcriptomic analysis recently carried out in the lab, which, according to the described role of C. elegans AHR-1 in neuronal determination (Huang et al., 2004, Qin & Powell-Coffman, 2004, Qin et al., 2006, Smith et al., 2013), revealed enrichment in processes linked to neuronal development and differentiation (Brinkmann et al. in preparation). A thorough analysis of the most differentially expressed genes between wild-type and ahr-1(ju145) revealed that many of them have been described in C. elegans to be affected not only during aging but also by dietary compounds (e.g. quercetin, resveratrol, bacteria) known to modulate AhR activity in mammals (Appendix Tables S2, S3). Interestingly, quantitative PCR analysis of some of these genes (atf-2, K04H4.2, egl-46, T20F5.4, ptr-4, dyf-7, clec-209, C01B4.6, C01B4.7, F56A4.3) confirmed their ahr-1 dependency in basal conditions and identified a specific AhR-bacteria diet regulatory effect (Figure EV4). Namely, while UVB or curcumin did not influence the expression of these genes neither in wt nor in the absence of the ahr-1, the reduced expression of the genes in the ahr-1(ju145) mutants fed HT115 was abolished in animal fed and OP50 diet (Figure EV4). Altogether, results shown so far support an evolutionarily conserved protective role for AhR against BaP and UVB and identify a new role for AhR in environmentally regulated aging with dietary bacteria as an important component in ahr-1-signaling-mediated longevity.

Loss of ahr-1 affects age-associated pathologies in a diet-dependent manner

The critical role of AHR-1 in C. elegans neurons (Huang et al., 2004, Qin & Powell-Coffman, 2004, Smith et al., 2013) and motility (Eckers et al., 2016, Williams et al., 2014) prompted us to investigate whether diet-dependent effects also influence age-related neuromuscular pathologies. A growing body of evidence indeed indicates microbiota can affect various aspects of C. elegans’ health-span (Brooks, Liang et al., 2009, Maier, Adilov et al., 2010, Munoz-Lobato, Rodriguez-Palero et al., 2014, Reinke, Hu et al., 2010, Xiao, Chun et al., 2015). In line with previous results, loss of ahr-1 extended life- and health-span of two age-associated disease models, namely animals with muscle expression of aggregation-prone proteins polyQ₄₀ (Morley, Brignull et al., 2002) and α-synuclein (van Ham, Thijssen et al., 2008) (Figure 3). Strikingly, when we looked at the diet-dependent effects, we found that similar to the wild-type strain, also the life- and health-span improvement promoted by ahr-1 deficiency in these disease models were significantly suppressed when animals were fed an OP50 diet instead of HT115 (Figure 3 A, B, E, F). Somewhat to our surprise, despite being longer-lived, ahr-1 mutants fed HT115 displayed a greater number and size of polyQ₄₀ and α-synuclein aggregates, which is nonetheless in line with studies revealing no direct correlation between aggregates and toxic effects (van der Goot, Zhu et al., 2012). Of note, the increase in aggregation was mainly regulated in a diet-independent manner (Figure 3 C, D, G, H), indicating that the ahr-1-diet interaction effect on health-span is independent of its effect on protein aggregation. Very surprisingly, when we investigated the effect of ahr-1 deficiency in another pro-aggregation model - a C. elegans strain expressing a pan-neuronal human Aβ1–42 peptide (Fong, Teo et al., 2016) - we found that contrary to the other models, it reduced animals’ life- and health-span, yet interestingly in a diet-dependent manner (Figure EV5).

Given that temperature may affect aging and associated pathologies depending on the bacterial diet and genetic background (Maier et al., 2010, Miller, Fletcher et al., 2017) we next tested whether the healthy aging phenotype of ahr-1 is also temperature-dependent. Despite the increased resistance to heat shock of ahr-1(ju145) mutants, and opposite to the lifespan at 20 °C, ahr-1 mutants were short-lived on an HT115 diet at 25 °C. Yet, again, this difference in the lifespan of wild-type and ahr-1 was abolished on the OP50 diet, which per se already shortened the lifespan at 25 °C of wild-type animals (Figure 4 A, B). The higher temperature also prevented the beneficial diet-dependent effects of loss of ahr-1 on life- and health-span in the polyQ strain (Figure 4 C, D), while the number of polyQ aggregates increased according to the rise in temperature but still in a diet-independent manner (Figure 4 E, F). Overall our data revealed a new diet-dependent, protein aggregation-independent effect of AhR in modulating age-associated neuromuscular pathologies.

Figure 4. At 25 °C, ahr-1(ju145) is short-lived and loses its protection against the toxicity of polyQ₄₀ aggregates.

A – B) Kaplan Meier curves of wild-type and ahr-1(ju145) at 25 °C. Pooled data of 180 worms/condition in 3 independent replicates are shown. * p-value < 0.05 vs. wt, # p-value < 0.05 vs. HT115, statistical test: Log-rank test. C – D) Kaplan Meier curves of polyQ;wt and polyQ;ahr-1. Worms were grown at 25 °C from day 3 (indicated by arrowhead). Pooled data of 150 worms/condition in 3 independent replicates are shown. * p-value < 0.05 vs. wt, # p-value < 0.05 vs. HT115, statistical test: Log-rank test. E) Representative fluorescence images of 10-days old polyQ;wt and polyQ;ahr-1 grown at 25 °C from day 3. F) Quantification of aggregates in 10-days old polyQ;wt and polyQ;ahr-1 grown at 25 °C from day 3. Mean + 95 % CI of pooled data from 21 (wt HT115), 19 (wt OP50), 20 (ahr-1 HT115) and 19 (ahr-1 OP50) worms/condition in 2 independent replicates are shown. * p-value < 0.05 vs. wt, statistical test: One-way ANOVA with Tukey’s multiple comparisons test.

Bacterial tryptophan metabolism mediates the beneficial effect of the ahr-1 mutants

In search of the potential bacterial factor responsible for the diet-dependent effects, we first asked whether metabolically active bacteria are required for the observed differences. To this end, we compared the heat-stress resistance of animals fed alive bacteria with that of animals fed bacteria either killed before seeding on plates or killed on the plates two days after seeding (thus allowing metabolites secretion). Very interestingly, killing bacteria before seeding completely abolished the differences between wild-type and ahr-1 fed HT115 (Figure 5 A – B), indicating that a factor produced by metabolically active HT115 bacteria may influence the AHR-1-mediated effects. In support of this possibility, the increased resistance to stress of ahr-1 animals observed on living HT115 bacteria persisted when bacteria were killed after growing for two days on the feeding plates (Figure 5 C). In line with the heat shock experiments, killing the bacteria before seeding also completely suppressed the increased lifespan of the ahr-1 mutants on HT115 with no major effects on animals fed OP50 (Figure 5 D). These data point towards HT115 secreted metabolites playing a role in ahr-1-mediated life- and health-span, likely via gut ingestion or neuronal sensing (Maier et al., 2010). In mammals, tryptophan (Trp) and its metabolites (such as indole) modulate AhR activity (Jin et al., 2014, Manzella, Singhal et al., 2018, Miller, 1997) and in C. elegans Trp was shown to abolish some of the different phenotypes observed between alive HT115 and OP50 (Gracida & Eckmann, 2013). Of note, supplementation of L-tryptophan completely abolished the heat-stress resistance of ahr-1(ju145) on HT115 - and partially suppressed the difference also between wild-type and ahr-1(ju145) on OP50 (Figure 5 E, F). Mass spectrometric analysis of the two bacteria supernatants revealed that the subtracted spectrum between HT115 and OP50 showed three prominent peaks (Figure 5 G) with m/z 361 likely being an arginine-tryptophan dipeptide, thus supporting a role for Trp in mediating the differential effects observed between HT115 and OP50. Overall, these data point towards a potential role in Trp metabolism in modulating ahr-1-bacteria-regulated age-associated phenotypes.

Figure 5. Metabolically active bacteria are required for the differences in lifespan between wild-type and ahr-1 on HT115.

A – C) Survival upon heat stress in 7-days old wild-type and ahr-1(ju145) feeding on either HT115 or OP50. Pooled data of 60 worms/condition in 3 independent experiments are shown. * p-value < 0.05 vs. wt, # p-value < 0.05 vs. HT115, statistical test: Log-rank test. A) Living bacteria were used as a food source. B) Bacteria were killed by UVB irradiation before seeding to the NGM. C) Bacteria had grown on the NGM for 2 days before being killed by UVB irradiation. D) Kaplan Meier curves of wild-type and ahr-1(ju145) on UVB-killed bacteria. Pooled data of 120 (wt HT115, ahr-1 HT115, ahr-1 OP50) and 110 (wt OP50) worms in 2 independent experiments are shown. * p-value < 0.05 vs. wt, # p-value < 0.05 vs. HT115, statistical test: Log-rank test. E – F) Heat stress survival after tryptophan supplementation with indicated concentrations of tryptophan. Survival curves of 7-days old worms feeding on HT115 supplemented with tryptophan are shown. The curves show pooled data from or 40 worms/condition in 2 independent replicates (1.25 mg trp) or 60 worms/condition in 3 independent experiments (2.5 mg trp). * p-value < 0.05 vs. wt, # p-value < 0.05 vs. HT115, statistical test: Log-rank test. G) Positive ESI MS analysis. The HT115 BPC after the subtraction of the OP50 BPC is shown. Masses in peak 2 are shown as inset.

Discussion

C. elegans AHR-1, similar to mammalian AhR, forms a heterodimer with the C. elegans Arnt homolog AHA-1 and binds to xenobiotic responsive elements (XREs), but unlike its mammalian counterpart, it does not bind the classical AhR activators such as TCDD or β-naphthoflavone (Powell-Coffman et al., 1998). However, no other potential AhR ligands or ligand-independent modulators have ever been tested in C. elegans, thus precluding to fully exploit the versatility of this model organism to unravel conserved or novel biological function for this environmentally relevant transcription factor. Only recently, conserved functions of the C. elegans AHR-1 in health and lifespan (Eckers et al., 2016, Williams et al., 2014) have been identified.

In this new study, we revealed a more general and complex role for this transcription factor in stress response and aging in basal conditions and in response to known modulators of its activity (i.e. curcumin, UVB and BaP). The differential effects on lifespan and gene expression we observed upon ahr-1 genetic-or RNAi-mediated suppression interestingly reveal the importance of fine-tuning its activity in a dose and/or tissue-dependent manner, which very nicely recapitulate the variety of phenotypic features described in mammals upon dosage effects or tissue-specific depletion (Brinkmann et al., 2020). Moreover, to our surprise, we found that besides increased resistance to heat-shock, lack of ahr-1 did not confer resistance to any other investigated insult, such as metabolic stress or radiation, and in fact sensitize animals to UVB. We specifically decided to focus our attention on the influence on stress response and aging of different classes of mammalian AhR modulators: environmental (xenobiotic) and dietary factors as well as by microbiota and endogenous factors produced by the organism. These mainly induce AhR’s transcriptional activity in a ligand-dependent manner, but ligand-independent, as well as antagonistic functions, have been also suggested (Ashida et al., 2008). Our data showed that C. elegans ahr-1(ju145) mutants are more sensitive to the lifespan shortening effects of UVB and BaP, two classical activators of mammalian AhR, thus further supporting a role for ahr-1 in the aging process. Of note, ahr-1 mutants also conferred sensitivity to UVB-reduced fertility, which we hypothesized to reflect increased sensitivity to germ cells apoptosis in irradiated animals as also found in human keratinocytes and mice (Frauenstein et al., 2013, Pollet, Shaik et al., 2018). Indeed, loss of AHR-1 function increased apoptosis in basal condition and upon radiation indicating a conserved anti-apoptotic function of the AhR in response to UVB. Similar to the detrimental effect of UVB, exposure to BaP in mammals decreases the lifespan and causes a variety of cancers as well as neurotoxicity (Chepelev, Moffat et al., 2015, Sakakibara, Nakagawa et al., 2005), and loss of AhR prevents BaP- and UVB-induced carcinogenicity in mice (Pollet et al., 2018, Shimizu et al., 2000). In this study, we showed for the first time that BaP has a conserved detrimental effect on lifespan, which is significantly worsened in the C. elegans ahr-1 mutants. Along with the UVB data, these findings support a protective role of AHR-1 in response to classical mammalian activators. However, surprisingly, both stressors induced cyp-35B1 expression in an ahr-1-independent manner thus pointing to a non-canonical mode of activation of AHR-1 in C. elegans. Interestingly, BaP and the UVB-generated ligand FICZ are big and planar molecules which, based on preliminary in silico analysis (Brinkmann et al. in preparation) likely do not fit into the ligand binding pocket of the C. elegans AHR-1. Thus, one can speculate that AHR-1, like its mammalian homolog, is activated by reactive oxygen species (ROS) (Kubli, Bassi et al., 2019), which can be produced either by UVB (Heck, Vetrano et al., 2003) or BaP (Cui, Chen et al., 2019, Wu, Huang et al., 2015).

Modulation of AHR-1 activity by ROS could also explain the potential inhibitory effect of curcumin on AHR-1. Indeed, polyphenols are dietary AhR modulators (Xue et al., 2017) with conserved beneficial effects on health and aging from invertebrates to humans (Monroy, Lithgow et al., 2013) often ascribed to their antioxidant properties (Liao et al., 2011). Here we found that curcumin promotes life- and health-span depending on ahr-1 but, similar to the other modulators, it did not modulate the expression of classical ahr-1-dependent genes such as cyps. Interestingly, curcumin and AHR-1 loss-of-function lead to non-additive effects on lifespan extension compared to untreated wild-type animals, suggesting that curcumin might actually act as an AHR-1 inhibitor, as also reported in mammals (Nishiumi et al., 2007). Whether the effects of curcumin are mediated through direct binding to AHR-1 or indirect modulation of its activity through other mechanisms (e.g. ROS reduction) remain to be clarified but our preliminary in silico analysis suggests that it is unlikely through a direct binding of curcumin to the AHR-1 ligand binding domain. Accordingly, while direct ligand-binding has not been verified for many of the polyphenolic AhR modulators, ligand-independent effects on AhR signaling, like binding to HSP90 (Palermo, Westlake et al., 2005) or inhibition of AhR phosphorylation by the protein kinase C (Das, Ramani et al., 2016, Nishiumi et al., 2007), have been instead described.

Most notably, we identified for the first time a critical role for AHR-1-bacterial diet interaction in regulating aging and associated phenotypes. Specifically, ahr-1 mutants displayed an extended health- and life-span on HT115 but not on the OP50 bacterial diet. Loss of ahr-1 also ameliorated health- and life-span in models of age-associated pathologies with aggregation-prone proteins expressed in the muscles in a diet-dependent manner. Yet, surprisingly, the beneficial effect did not correlate with the amount and size of the aggregates, indicating either that aggregation in this context actually has a beneficial effect, or that loss of ahr-1 promotes health-span independently from mechanisms regulating proteotoxic aggregation. Regardless, we hypothesized that bacteria-associated factors from HT115 or OP50 might mediate the diet-dependent changes in health-span and narrowed them down to secreted bacterial metabolites. Although the exact metabolite(s) responsible for the diet-dependent effects still remain to be identified, our data point to the potential involvement of Trp metabolism. This again supports evolutionarily conserved functions of AHR-1, as its activity is known to be influenced by Trp and its metabolites in mammals (Jin et al., 2014, Manzella et al., 2018, Miller, 1997). Interestingly, Trp and its metabolites have been already shown to differentially modulate C. elegans’ health- and life-span. Trp supplementation increases heat-stress resistance and lifespan of C. elegans (Edwards, Canfield et al., 2015) and reduces the proteotoxicity in neurodegenerative disease models (van der Goot et al., 2012). Instead, different Trp metabolites display opposite effects: Trp degradation through the Kynurenine pathway increases proteotoxicity in neurodegenerative disease models (van der Goot et al., 2012), while the tryptophan metabolite indole, from commensal bacteria, increases the lifespan of C. elegans through AHR-1 (Sonowal, Swimm et al., 2017). The primary site of action of the bacterial metabolite (e.g. intestine, sensory neurons) as well as the affected AhR-dependent molecular process involved in modulation of the aging process (e.g. immune response, redox reactions), are very challenging and interesting points which remain to be addressed.

In summary, we demonstrated that C. elegans ahr-1 displays evolutionarily conserved functions, such as its protective activity against xenobiotic and UVB, but opposite roles in the aging process in response to environmental or dietary interventions. Most importantly, we identified a direct link between ahr-1-regulated processes and bacterial diet as a key determinant of aging and associated pathologies, most likely trough tryptophan metabolism. Overall our findings support a central role for AhR in the aging process in a context-dependent manner, thus expanding the already vast panel of activities played by the AhR in different pathophysiological conditions and establishing C. elegans as a powerful model organism to unravel new AhR-regulated processes in response to conserved modulators of its activity.

Material and Methods

C. elegans strains and cultivation

C. elegans strains used in this study are listed in Appendix Table S4. We created the following strains for this study by crossing CZ2485 with different transgenic strains to obtain: NV33a: ahr-1(ju145); cyp-35B1p::GFP + gcy-7p::GFP, NV35a: ahr-1(ju145); (pAF15)gst-4p::GFP::NLS, NV38b: ahr-1(ju145); unc-54p::Q40::YFP, NV42a: ahr-1(ju145); unc-54p::alpha-synuclein::YFP, NV43a: ahr-1(ju145); hsp-6p::GFP, NV47a: ahr-1(ju145); ugt-29p::GFP. For maintenance, worms were kept synchronized by egg lay at 20 °C on Nematode Growth Media (NGM) plates and fed with E. coli OP50 according to methods described in (Stiernagle, 2006). For the experiments, worms were synchronized on plates supplemented with E. coli HT115(L4440) or OP50(L4440) according to the condition of interest. Unless stated otherwise, plates were supplemented with 1 mM IPTG when using HT115(L4440) or OP50(L4440).

Crossing CZ2485 with transgenic strains

CZ2485 males were generated by treating hermaphrodite L4 larvae with heat stress (4 h at 30 °C). 10 adult CZ2485 males were then left with one hermaphrodite L3/L4 larvae for mating (P0 generation). Single worms of the F1 and F2 generations of this cross were checked for the ju145 allele by PCR (ahr-1F CGGAAAGTTGATGTCTCTAC, ahr-1R TGCTGACTAGACGATATACC) followed by a restriction with AlwI (New England Biolabs, R0513S), which cuts the PCR fragment at the position of the ju145 point mutation. Single worm PCR was performed according to standard protocols (He, 2011).

Gene silencing by RNA-mediated interference (RNAi)

Gene silencing was achieved through feeding bacteria expressing plasmids with dsRNA against specific genes (Timmons & Fire, 1998). For RNAi, either E. coli HT115(DE3) or E. coli OP50(xu363) (Xiao et al., 2015) were used. RNAi feeding was applied continuously from birth to death.

E. coli strains and growth

Bacteria were grown in LB medium at 37 °C overnight. When using E. coli carrying vectors the LB medium was supplemented with 0.01 % of ampicillin and 0.0005 % of tetracycline. E. coli HT115(L4440), HT115(ahr-1), and OP50 were obtained from the Ahringer C. elegans RNAi library (Kamath & Ahringer, 2003). E. coli OP50(xu363) was a gift from Shawn Xu.

Transformation of E. coli OP50(xu363)

For this study, we created OP50(L4440) by PEG transformation of OP50(xu363) according to methods described previously with adjustments (Chung & Miller, 1988). 100 µl of OP50(xu363) were transformed with 100 pg of plasmids isolated from the HT115(DE3) strains (Ahringer Library). The QIAprep Spin Miniprep Kit (Qiagen, 27104) was used for plasmid isolation.

Heat stress survival

The resistance to heat stress was tested with 20 animals/condition per experiment at 35 °C on 3 cm plates wrapped with parafilm in an incubator (Intrafors HT Multitron). When using animals older than 3 days, animals were transferred to fresh NGM plates daily during the fertile phase. The number of dead animals was scored hourly by gently touching the worms with a platinum wire and analysis was performed as described for the lifespan assay.

Reproduction on heat stress, glucose, and BP

Animals were grown on control plates with alive bacteria until day 3 and then transferred to treatment or control plates for 24 hours. After a 24-hour treatment, 3 animals of each condition were transferred to fresh control plates for 4 hours to lay eggs and the number of eggs laid was counted. The number of progenies hatched from these eggs was counted 2 days afterward.

Development

The development on the specific compound was explored by counting the number of eggs, which developed to gravid adults after 72, 96, and 120 hours as well as the number of worms that arrested their development and the number of eggs, which did not hatch.

Glucose treatment

D-Glucose (Merck, 8342) was dissolved in H₂O and supplemented to the NGM after autoclaving and UVB killed E. coli HT115(L4440) were fed as a food source.

2,2′-dipyridyl (BP) treatment

The iron chelator 2,2′-dipyridyl (Carl Roth, 4153) was dissolved in H₂O and supplemented to the NGM after autoclaving.

UVB irradiation

Worms were exposed to ultraviolet radiation on bacteria-free NGM plates, using a Waldmann UV 236 B (UV6) lamp with an emission maximum of 320 nm. Irradiation times were 9 seconds, 27 seconds, and 53 seconds for 100 J/m², 300 J/m² and 600 J/m², respectively with a distance of 18 cm between lamp and plate.

Reproduction after UVB treatment

3 days old worms were treated with UVB and the effect of irradiation on germ-cells in different stages (oogenic stage, meiotic stage, and mitotic stage) was investigated, as described in (Kim, Yang et al., 2005). Briefly, the number and viability of eggs laid between 1 – 8 hours (oogenic stage), 8 – 24 hours (meiotic stage) and 24 – 32 hours (mitotic stage) after the irradiation were analyzed.

Lifespan

The lifespan analysis was started from a synchronized population of worms, which was transferred to fresh NGM plates daily during the fertile period. After the fertile phase, the animals were transferred every alternate day. Dead, alive and censored animals were scored during the transferring process. Animals were counted as dead when they did show neither movement, nor response to a manual stimulus with a platinum wire, nor pharyngeal pumping activity. Animals with internal hatching (bags), an exploded vulva or which died desiccated on the wall were censored. The number of dead and censored animals was used for survival analysis in OASIS (Yang, Nam et al., 2011) or OASIS 2 (Han, Lee et al., 2016). For the calculation of the mean lifespan and the survival curve in OASIS and OASIS 2, the Kaplan Meier estimator was used, and the p-values were calculated using the log-rank test between pooled populations of animals.

Movement/Healthspan

The movement was set as a parameter for healthy aging, and the phase of active movement is referred to as healthspan. It was assessed in the populations used for the lifespan assay. Animals, which were either crawling spontaneously or after a manual stimulus, were considered as moving while dead animals or animals without crawling behavior were considered as not moving. Statistical analysis was done as described for lifespan.

Curcumin treatment

Curcumin (Sigma Aldrich, C7727) was dissolved in DMSO (Carl Roth, 4720) in a concentration of 100 mM and supplied to the NGM after autoclaving. The final concentration of curcumin in the media was 100 µM (0.1 % DMSO). Control plates contained 0.1 % DMSO. Worms were treated continuously starting from eggs.

Benzo(a)pyrene (BaP) treatment

Benzo(a)pyrene (Sigma Aldrich, B1760) was dissolved in DMSO (Carl Roth, 4720) in concentrations 1000 times higher than the desired concentration in the NGM. After autoclaving the NGM, BaP or DMSO were added to the media. For development assays, worms were treated from eggs, while they were treated from the first day of adulthood for lifespan and healthspan assays.

Quantification of polyQ aggregates

PolyQ₄₀ aggregates were visualized by fluorescence microscopy (100x magnification) in worms anesthetized with 15 mM sodium azide (Sigma, S2002). The number and the size of the aggregates were quantified in Fiji (Schindelin, Arganda-Carreras et al., 2012). To assess the number of aggregates, images were stitched using the Fiji pairwise stitching plugin (Preibisch, Saalfeld et al., 2009) to create whole worms. The average size of the aggregates was instead measured in the non-stitched images. The number and the size of aggregates were counted using the plugin “Analyze Particles”.

Quantification of α-synuclein aggregates

α-synuclein aggregates in the head muscles of 7-days old worms were visualized by fluorescence microscopy (400x magnification) in worms anesthetized with 15 mM sodium azide (Sigma, S2002). Pictures were segmented using Ilastik (version 1.3.0) (available on https://www.ilastik.org/) (Sommer, Strähle et al., 2011). The segmented pictures were used to analyze the number and size of the aggregates in Fiji (Schindelin et al., 2012) using the plugin “Analyze Particles”.

Assessment of age-associated features at 25 °C

Lifespan, movement, and polyQ aggregation analysis were performed as described above. PolyQ-expressing worms were kept at 20 °C until reaching the L4 stage and were afterward kept at 25 °C for the rest of their lifespan.

Killing bacteria before seeding

When seeding killed E. coli onto the NGM, the bacteria were pelleted (10 min at 4000 rpm), the supernatant was removed, and the pellet was suspended in S-basal to a final concentration of OD₅₉₅ = 3.6. Then the bacteria suspension was irradiated with a UVB lamp (Waldmann UV 236 B) for 1 hour to kill the bacteria. The suspension of the dead bacteria was again pelleted and re-suspended in fresh S-basal. Killed bacteria were seeded to NGM plates in a concentration of OD₅₉₅ = 3.6 and let dry at room temperature overnight.

Killing bacteria on plates

Bacteria (OD₅₉₅ = 0.9) were seeded to the NGM plates and let grow for 2 days at room temperature. The bacterial lawn was then killed by exposure to UVB light (Waldmann UV 236 B) for 45 min.

Tryptophan supplementation

The concentration of tryptophan and the supplementation procedure was taken from (Gracida & Eckmann, 2013) with minor changes. L-tryptophan (Carl Roth 4858.2) was dissolved in water at a concentration of 12.5 mg/ml and incubated shaking at 30°C for 45 min. Afterward, the solution was filtered (pore size: 0.22 µm, Carl Roth P666.1). For 7 ml of NGM 200 µl of 12.5 mg/ml tryptophan was spotted on the bacterial lawn of an NGM plate, which was kept at room temperature for overnight. After the supplementation, the dish was kept at room temperature for another day.

Mass Spectrometry

Liquid NGM was prepared like solid NGM but without agar to prevent solidification. NGM was poured into petri dishes (7 ml NGM/ 6 cm petri dish) and seeded with bacteria (200 µl, OD₅₉₅ = 0.9) or LB-medium (control). The medium was incubated for two days at room temperature to allow the bacteria to grow. Then, the bacteria were removed by centrifugation (4500 rpm, 10 min) and the medium was filtered (22 µm filter, Carl Roth P666.1) before using it for mass spectrometry analysis. Electrospray ionization mass spectrometry (ESI-MS) was used. For the LC-MS measurements, the liquid NGM samples were diluted 1:20 with methanol before the injection of 10 μl sample volumes. Triterpenes were separated on a Dionex HPG 3200 HPLC system (Thermo Scientific) equipped with a 150 × 2.1 mm, 2.7 μm, C18-CSH column (Waters) with a binary gradient system. Mobile phase A consisted of water + 0.1 % formic acid (FA), and mobile phase B consisted of methanol + 0.1 % FA. The mobile phase gradient was as follows: Starting conditions were 5 % mobile phase B, increased to 95 % B within 10 min, the plateau was held for 4 min, and the system was returned to starting conditions within 1 min and held for another 4.5 min. The flow rate was 0.5 mL/min. The MS and MS/MS analysis were performed with a quadrupole-time-of-flight instrument (maXis 4G, Bruker Daltonics, Bremen, Germany) equipped with an ESI source. The device was operated in positive-ion and negative-ion mode and the operating conditions were as follows: dry gas (nitrogen): 8.0 L/min, dry heater: 220 °C, nebulizer pressure: 1.8 bar, capillary voltage: 4500 V. Data analysis were performed using the software data analysis 4.2 and Metabolite detect 2.1 (Bruker daltonics, Bremen, Germany).

Quantification of transgene expression

The expression of fluorescently-tagged genes was investigated in adult worms (first day of adulthood) by using fluorescence microscopy (ZEISS Imager M2 with an Axiocam MRm camera). For this, worms expressing fluorescently tagged genes were paralyzed with 15 mM NaN₃ and pictures were taken with identical exposure times and settings. The pictures were then analyzed in either Image J, Fiji (Schindelin et al., 2012) or cellprofiler. Depending on the distribution of the expression, the fluorescence was either measured in a defined area of the region of interest or the whole worm. Statistical analysis was performed with the pooled data.

Measurement of cyp-35B1p::GFP intensity in response to BaP

3-days old worms were treated with BaP or UVB for 18 hours. The relative intensity of cyp-35B1p::GFP was then visualized by fluorescence microscopy (100x magnification) and analyzed using Fiji (Schindelin et al., 2012). The integrated density was chosen as a parameter for the expression. The quantified intensities were normalized to the mean of untreated wild-type in each experiment. Statistical analysis was performed with the pooled data.

UVB-induced Apoptosis

To investigate UVB-induced apoptosis, L4 larvae were treated with 600 J/m² UVB and the apoptotic corpses were counted 24 h post-irradiation in the gonad loop region. The apoptotic corpses were identified based on their shape.

Fertility

The number of eggs and progeny of animals were investigated during the main fertile period. Two days after synchronization single L4 larvae were transferred to NGM plates and from then transferred to fresh NGM plates every 24 hours until the 8^th day after hatching. The number of eggs laid during this period was counted. Two days later the progeny hatched on each day were counted.

Pharyngeal pumping using the NemaMetrix ScreenChip System

To measure the pharyngeal pumping rate with the NemaMetrix ScreenChip, worms were washed off the plates with S-basal and collected in a reaction tube. Then, the worms were washed twice with S-basal and twice with 10 mM serotonin (Sigma Aldrich, 14927) and incubated in 10 mM serotonin for 30 min. Worms were loaded on the ScreenChip SC40 with a syringe (0.01 ml – 1 ml). The EPG of single worms was recorded for a duration of approx. 2 minutes. Only worms which showed pumping activity were recorded, while those with no pumping activity were not considered. The following NemAcquire-2.1 and NemAnalysis-0.2 software were used for analysis (https://nemametrix.com/products/software/).

Assessment of mRNA expression by RTqPCR

Samples from 3 independent replicates with approximately 1000 3-days old worms per condition were collected and RNA was extracted. After washing and elution steps the RNA content was quantified by spectrophotometry, and 1 - 2 µg of RNA was used for the cDNA synthesis (Omniscript RT Kit (Qiagen, 205111). Primers were designed using NCBI Primer BLAST (https://www.ncbi.nlm.nih.gov/tools/primer-blast/) (Ye, Coulouris et al., 2012). Each primer pair was designed to span an exon-exon junction. Primer pairs and their features are listed in Appendix Tables S5 – S6. For the Real-time qPCR, the cDNA was diluted 1:20 in 10 mM TRIS (pH 8.0). For the reaction, the qPCR Green Core kit (Jena Biosciences, PCR-333L) or the GoTaq® qPCR kit (Promega, A6001) was used. The samples were run in a MyiQ2 cycler (BioRad), and the expression of each sample was measured in duplicate on the same multi-well plate. The expression was calculated relative to the reference genes act-1 and cdc-42 using the iQ5 software. All data collected were enabled for gene study according to the BioRad user instructions, and the expression was calculated using the normalized expression (ddC_T). The efficiency of each primer pair reaction was added for correct quantification of the normalized expression. The efficiency was assessed with 1:20, 1:100, 1:500, and 1:2500 dilutions of the cDNA. From normalized expression values, the fold-change compared to wild-type was calculated for each replicate.

Statistical analysis

If not stated, statistical analysis was performed in GraphPad Prism 6. For life-/healthspan assays, statistical analysis was done using OASIS (Han et al., 2016, Yang et al., 2011). Statistical analysis of the microarray data was performed in R.

Author Contributions

Conceptualization: N.V.

Formal analysis: V.B., A.Sc., N.V.

Funding acquisition: V.B., N.V.

Investigation: V.B., D.P., A.Sh.

Supervision: N.V.

Resources: R.M., N.V.

Visualization: V.B.

Writing -original draft-: V.B., N.V.

Writing -review and editing-: R.M., A.Sc.

Conflict of interest

The authors declare no competing interests.

Expanded view figure legends

Figure EV1. AHR-1 regulates genes involved in xenobiotic stress response.

A) Quantification of cyp-35A2p::GFP expression in worms treated with control or ahr-1 RNAi. Mean + 95 % CI of pooled data from 3 independent experiments with 43 worms are shown. B) Quantification of cyp-37A1p::GFP expression in worms treated with control or ahr-1 RNAi. Mean + 95 % CI of pooled data from 2 independent experiments with 38 (con RNAi) and 26 (ahr-1 RNAi) worms are shown. C) Quantification of cyp-35B1p::GFP expression in worms treated with control or ahr-1 RNAi (left panel) or worms with either wildtype ahr-1 or the ju145 allele (right panel). Mean + 95 % CI of pooled data from 60 (con RNAi), 40 (ahr-1 RNAi), 52 (wt), and 63 (ahr-1) worms in 3 independent experiments are shown. D) Quantification of ugt-29p::GFP expression in ahr-1-depleted worms. Mean + 95 % CI of pooled data from 3 independent experiments are shown. Number of worms 44 (con RNAi), 36 (ahr-1 RNAi) [left panel], 34 (wt), 40 (ahr-1) [right panel]. E) Quantification of gst-4p::GFP expression in worms treated with control or ahr-1 RNAi (left panel) or worms with either wildtype ahr-1 or the ju145 allele (right panel). Mean + 95 % CI of pooled data from 27 (con RNAi), 29 (ahr-1 RNAi), 27 (wt), and 19 (ahr-1) worms in 2 independent experiments are shown. F) ahr-1 depletion via RNAi does not affect the lifespan of N2. Kaplan Meier curves of control (L4440) or ahr-1 RNAi treated worms. Pooled data of 280 (N2 L4440) and 264 (N2 ahr-1) worms/condition in 5 independent replicates are shown. * p-value < 0.05 vs. wt, statistical test: Log-rank test. G) ahr-1 depletion via RNAi shortens the lifespan in a strain with enhanced RNAi in the nervous system (TU3311). Kaplan Meier curves of control (L4440) or ahr-1 RNAi treated worms. Pooled data of 180 worms/condition in 3 independent replicates are shown. * p-value < 0.05 vs. wt, statistical test: Log-rank test.

Figure EV2. Cyp-13B1 expression is increased by curcumin in an ahr-1-dependent manner.

A) Fertility over time in response to 100 µM of curcumin. Mean + SEM of 6 worms/condition in 2 independent experiments are shown. * p-value < 0.05 vs. wt, # p-value < 0.05 vs. DMSO, statistical test: 2-way ANOVA with Tukey’s multiple comparisons test. B) Expression of cyp genes in DMSO-treated ahr-1 relative to DMSO-treated wild-type. C) Expression of cyp genes in curcumin-treated wild-type relative to DMSO-treated wild-type. D) Expression of cyp genes in curcumin-treated ahr-1 mutants relative to DMSO-treated ahr-1 mutants. Mean + SEM is shown in panels B - D. * p-value < 0.05 vs. wt, # p-value < 0.05 vs. DMSO, statistical test: Two-way ANOVA with Tukey’s multiple comparisons test.

Figure EV3. Aging is influenced by the bacterial diet in the ahr-1 mutant.

A) Development of wild-type and ahr-1 on indicated doses of BaP. Pooled data of 3 independent experiments are shown, the number of individuals is presented as N. Statistical test: 2-Way ANOVA with Tukey’s multiple comparisons test, no statistical differences were observed. B) cyp-35B1p::GFP induction in response to 5 µM of BaP. Mean + 95 % CI of pooled data of 58 (wt DMSO), 51 (wt BaP), 53 (ahr-1 DMSO), and 65 (ahr-1 BaP) in 3 independent experiments are shown. *p-value < 0.05 vs. wt, # p-value < 0.05 vs. DMSO, statistical test: One-way ANOVA with Tukey’s multiple comparisons test. C) UVB irradiation and ahr-1 loss of function induce apoptosis. Mean + 95 % CI of pooled data of 39 (wt con, ahr-1 UVB) and 38 (wt UVB, ahr-1 con) worms in 3 independent experiments are shown. Statistical test: 2-Way ANOVA with Tukey’s multiple comparisons test, * p-value < 0.05 vs. wt, # p-value < 0.05 vs. con. D) cyp-35B1p::GFP induction in response to UVB irradiation. Mean + 95 % CI of pooled data of 34 (wt con), 37 (wt UVB), 28 (ahr-1 con), and 35 (ahr-1 UVB) worms in 2 independent experiments are shown. * p-value < 0.05 vs. wt, # p-value < 0.05 vs. control, statistical test: One-way ANOVA with Tukey’s multiple comparisons test. E) Heat shock resistance of wild-type and ahr-1 feeding on HT115 or OP50. Mean survival times + SEM of 3 independent experiments are shown, the number of individuals is presented as N. Statistical test: Two-way ANOVA with Tukey’s multiple comparisons test. * p-value < 0.05 vs. wt, # p-value < 0.05 vs. HT115, F) cyp-35B1p::GFP expression of worms feeding on HT115 or OP50. Mean + 95 % CI of pooled data of 37 (wt HT115), 27 (wt OP50), 44 (ahr-1 HT115), and 19 (ahr-1 OP50) in 2 independent experiments are shown. *p-value < 0.05 vs. wt, # p-value < 0.05 vs. HT115, statistical test: One-way ANOVA with Tukey’s multiple comparisons test. G) Pumping frequencies of pooled data from 3 independent experiments are shown as mean + 95 % CI, the number of individuals is presented as N. * p-value < 0.05 vs. wt, # p-value < 0.05 vs. HT115, § p-value < 0.05 vs. day 3, statistical test: Two-way ANOVA with Tukey’s multiple comparisons test. H) ahr-1 mRNA expression in wild-type worms feeding on HT115 or OP50. Pooled data of 3 independent replicates are shown. No statistical significance was observed with the t-test.

Figure EV4. Gene expression changes in response to Ahr modulators.

Expression of the strongest up- and down-regulated genes in ahr-1 vs. wild-type in response to A) 100 µM of curcumin, B) 1200 J/m² UVB, and C) bacteria. Pooled data from 3 independent replicates are shown. Statistical test: 2-Way ANOVA with Tukey’s multiple comparisons test, * significance vs. wt, # significance vs. control.

Figure EV5. AHR-1 does not influence aging in an Alzheimer’s disease model.

A – B) Kaplan Meier curves of Abeta;wt and Abeta;ahr-1 grown on HT115. Pooled data of 240 (wt) and 236 (ahr-1) worms/condition in 4 independent replicates are shown. * p-value < 0.05 vs. wt, statistical test: Log-rank test. C – D) Kaplan Meier curves of Abeta;wt and Abeta;ahr-1 grown on OP50. Pooled data of 120 worms/condition in 2 independent replicates are shown. Statistical evaluation with log-rank test did not display significant differences.

Acknowledgments

Most nematode strains utilized in this work were provided by the Caenorhabditis Genetics Center, funded by the NIH Office of Research Infrastructure Programs (P40 OD010440). Other nematodes or bacteria strains were kindly provided by David Depomerai (University of Nottingham), Sheila Nathan (University of Malaysia), and Shawn Xu (University of Michigan). We thank Anne Hemmers, Bo Scherer, and Lisa Tschage for technical help, Sabine Metzger (University of Cologne) for the performance of mass spectrometry, and Rene Deenen (University of Düsseldorf) for the performance of the microarray. We further thank Wormbase and the GENiE network funded by the European Cooperation in Science and Technology (COST Action BM1408). This work was financially supported by funding to N.V. from the German Research Foundation (DFG VE366/8-1). V.B. was supported by a Ph.D. scholarship by the Jürgen Manchot Foundation.

Appendix

Appendix Tables

Table S1. Relative expression of transgenes in control or ahr-1 RNAi-treated reporter strains.

The relative expression was measured in worms on their first day of adulthood and normalized to control RNAi treated worms in each replicate. N shows the number of worms and n the number of experiments, Statistical test: 2-tailed unpaired t-test.

Table S2. List of the most robust over-expressed genes in ahr-1 vs. wildtype.

Human orthologs were extracted using the BioMart tool (https://parasite.wormbase.org/biomart). Modulators of these genes with potential relevance for AHR-1 activity were selected from WormBase. For log fold change values (logFC) a base of 2 is used.

Table S3. List of the most robust under-expressed genes in ahr-1 vs. wildtype.

Human orthologs were extracted using the BioMart tool (https://parasite.wormbase.org/biomart). Modulators of these genes with potential relevance for AHR-1 activity were selected from WormBase. For log fold change values (logFC) a base of 2 is used.

Table S4. C. elegans strains used in this study.

Strains are sorted alphabetically.

Table S5. Primer pairs used in this study to validate genes differentially expressed between ahr-1(ju145) and wild-type (Figure S4).

Table S6. Primer pairs used for the assessment of cyp gene expression in DMSO and curcumin-treated wild-type and ahr-1(ju145) (Figure S2).