Introduction

Aging is a major risk factor for the onset of cancer, cardiovascular disease, dementia, and many other serious illnesses. Studying the aging process is crucial because a more comprehensive understanding can lead to the development of therapeutics that treat multiple diseases simultaneously. The use of model organisms, from yeast to mammals, has helped define genetic and environmental pathways that influence aging¹. Many of these pathways, such as dietary restriction (DR), defined as a decrease in nutrient intake without malnutrition, involve active modification of metabolism that robustly and reproducibly improve health and longevity across species^2–4. Similarly, the response to oxygen deficiency, or hypoxic response, can lead to increased longevity, healthspan, and stress resistance⁵. Interestingly, in C. elegans, both DR and the hypoxic response converge upon a single gene, fmo-2, that is necessary and sufficient to improve health and longevity⁶.

Flavin-containing monooxygenases (FMOs) are a family of enzymes that use oxygen and NADPH to oxygenate nucleophilic substrates. FMOs oxygenate a wide array of xenobiotic substrates and were discovered ∼50 years ago for their role in drug metabolism^7,8.

Consequently, a majority of published data on FMOs relate to this role. Recently, studies have begun to focus on the endogenous role(s) of FMOs. Results show that multiple mammalian FMO proteins are involved in systemic metabolism^9,10. We initially discovered a role for C. elegans fmo-2 in regulating stress resistance and longevity downstream of DR and hypoxia and have continued to identify role(s) for FMO-2 in C. elegans metabolism and longevity¹⁰.

However, the conserved mechanisms for FMO-mediated health benefits are still unclear, as are the roles of individual FMO enzymes. Thus, to best understand how FMOs could be leveraged to improve health, we need to understand the conserved mechanism of these enzymes in longevity and metabolism.

There are five C. elegans FMOs, and three of the five share significant structural overlap with mammalian FMOs. This overlap involves a predicted endoplasmic reticulum (ER) localization and a membrane spanning domain, which are each found in fmo-1, fmo-2, and fmo-4. In comparison to the more well-studied longevity gene fmo-2, fmo-4 in particular shares 88% identity in the catalytic domain amino acids, suggesting that FMO-2 and FMO-4 may bind similar substrates and could plausibly have overlapping endogenous roles. Published data also show that DR induces both fmo-2 and fmo-4 gene expression, implying a role for fmo-4 in longevity regulation⁸. Based on this, we hypothesized that fmo-4 may play a role in C. elegans longevity and that studying this role can help to understand this gene family and its relevance to aging.

FMOs are transmembrane enzymes thought to primarily reside in the ER. While mainly studied for its involvement in protein and lipid synthesis, the ER also plays a key role in metabolism and maintaining homeostasis within the cell. For instance, the ER responds to misfolded proteins by eliciting its unfolded protein response (UPR^ER), thereby removing dysfunctional proteins and restoring homeostasis¹¹. The ER is also responsible for storing and releasing calcium in a coordinated manner with the cytosol and mitochondria¹². The ER establishes calcium homeostasis through processes involving 1) sequestering calcium in the lumen with the calreticulin chaperone^12,13, 2) releasing calcium into the cytosol through the inositol triphosphate receptor (IP₃R)^12,14, and 3) bringing calcium into the ER lumen through the sarcoplasmic endoplasmic reticulum ATPase (SERCA) pump^12,15. Interestingly, the ER’s role in regulating metabolic and calcium homeostasis in the cell has recently been linked to longevity in C. elegans. Knocking out activating transcription factor (atf)-6, one of the three branches of the UPR^ER, modulates calcium signaling from the ER to the mitochondria, resulting in increased mitochondrial turnover and lifespan extension in C. elegans¹⁶.

Given that 1) fmo-4 is induced by a longevity-promoting pathway (DR), 2) fmo-4 is predicted to be an ER transmembrane protein, and 3) the ER’s role in cellular homeostasis is linked to aging, we hypothesized that fmo-4 regulates longevity by modulating ER-related processes. Here, we test this hypothesis by modifying fmo-4 expression and interrogating interactions between fmo-4, stress resistance, longevity, and the ER. Our resulting data show that fmo-4 promotes longevity and paraquat stress resistance downstream of mTOR and DR through calcium homeostasis. We find that not only is fmo-4 a regulator of multiple longevity promoting pathways, but it is also sufficient to extend lifespan and confer paraquat stress resistance when overexpressed either ubiquitously or in the hypodermis. Transcriptomics data and stress assays reveal that calcium signaling is altered downstream of fmo-4 expression, and that fmo-4 interacts with calcium regulation between the ER and mitochondria to promote longevity and paraquat stress resistance. Together, these results establish fmo-4 as a longevity promoting gene and provide evidence as to how fmo-4 extends C. elegans lifespan through calcium-mediated ER to mitochondrial processes.

Results

Cefmo-4 is required for DR and mTOR-mediated lifespan extension

C. elegans fmo-4 shares significant similarity to its family member and longevity gene, fmo-2, including 88% conservation in catalytic residues (Supplementary Fig 1), a transmembrane domain (Fig 1A) and predicted subcellular localization in the ER (Supplementary Fig 2)¹⁷. Additionally, fmo-4 is induced by the longevity intervention DR⁸. These structural similarities and DR-mediated induction led us to hypothesize that fmo-4 may play a role in aging. To test this, we first asked whether fmo-4 is required for well-studied longevity pathways, including DR. We utilized fmo-4 (ok294) knockout (KO) animals¹⁸ on five conditions reported to extend lifespan in C. elegans. Since fmo-4 is induced by dietary restriction, we started with fed and DR (sDR¹⁹) conditions. Our results show that loss of fmo-4 has no significant effect on control fed worms but prevents nearly all the lifespan extension seen under DR (Fig 1B). Wild-type (WT) worms on DR experience a ∼35% lifespan extension compared to fed WT worms, but when fmo-4 is knocked out this extension is reduced to ∼10% and this interaction is significant by cox regression (p-value < 4.50e^-6). These data support that fmo-4 is required for the DR longevity pathway (Fig 1B). Having established this role, we continued lifespan analyses of fmo-4 KO worms exposed to RNAi knockdown of the S6-kinase gene rsks-1 (mTOR signaling)²⁰, the von hippel lindau gene vhl-1 (hypoxic signaling)²¹, the insulin receptor daf-2 (insulin-like signaling)²², and the cytochrome c reductase gene cyc-1 (mitochondrial electron transport chain, cytochrome c reductase)²³ (Fig 1C-F). The resulting data show that fmo-4 is fully required for rsks-1-mediated longevity as marked by a complete abrogation of the rsks-1 RNAi lifespan extension when fmo-4 is knocked out (Fig 1C), placing fmo-4 downstream of mTOR. fmo-4 was not required for the hypoxic response (Fig 1D), insulin- like signaling (Fig 1E), or cytochrome c reductase inhibition pathways (Fig 1F) to increase lifespan. It is notable that, unlike fmo-4, fmo-2 is required for vhl-1-mediated longevity⁶. This result, coupled with fmo-4 and fmo-2 each being required for DR, suggests that these two genes in the same family are overlapping but distinct in their requirement. Together, these data suggest that fmo-4 plays a necessary role in C. elegans longevity regulation downstream of at least two pathways: DR and mTOR signaling.

Figure 1.

fmo-4 is required for fmo-2-mediated longevity, stress resistance, and healthspan

Having established that, like fmo-2, fmo-4 is required for DR-mediated longevity, we next hypothesized that fmo-2 and fmo-4 may be acting in the same pathway. To test this, we crossed fmo-2 overexpressing (OE) worms with fmo-4 KO worms to create an fmo-2 OE; fmo-4 KO strain (fmo-2OE;4KO). We validated this strain via qPCR analysis (Supplementary Table 1).

Upon measuring their lifespan compared to fmo-2 OE animals, we find that knocking out fmo-4 completely abrogates the lifespan extension from fmo-2 OE (Fig 2A). This result suggests that fmo-4 is required for fmo-2-mediated longevity and thus may act downstream of fmo-2 to promote longevity in C. elegans. We previously showed that fmo-2 OE is not just sufficient for lifespan extension but is also sufficient to promote resistance to multiple forms of stress, including oxidative stress (paraquat), heat stress, and ER stress (tunicamycin)⁶. Since fmo-4 is required for fmo-2- mediated longevity, we hypothesized that fmo-4 would also be required for fmo-2-mediated stress resistance. Following exposure to 5mM paraquat, 37°C heat stress, or 5µg/mL tunicamycin, we find that the fmo-2OE;4KO worms survive similar to WT worms and are no longer resistant to these stresses (Fig 2B-D). Thus, knocking out fmo-4 blocks fmo-2-mediated broad stress resistance.

Figure 2.

Healthspan is a crucial component to longevity, and we previously found that fmo-2 OE is sufficient to promote healthspan benefits with age in C. elegans⁶. As fmo-4 is required for both the lifespan extension and stress resistance observed with fmo-2 OE, we hypothesized that fmo-4 would also be required for fmo-2-mediated healthspan benefits. To test this, we measured the thrashing rate of fmo-2 OE;4KO worms at day 2 and day 10 of adulthood. We find that knocking out fmo-4 abrogates the healthspan benefits of the fmo-2 OE worms with age (Fig 2E-F). Thus, fmo-4 is required for fmo-2-mediated health and resilience benefits, further supporting its acting downstream of fmo-2 to promote health and longevity.

Overexpression of fmo-4 promotes longevity, healthspan, and paraquat stress resistance

While being necessary for increased healthspan, lifespan, and stress resistance shows that fmo-4 is required for these benefits, it does not test whether it can modulate aging directly.

Therefore, to better understand fmo-4’s role in longevity and health, we next asked whether fmo-4 is sufficient for lifespan extension, stress resistance, and healthspan benefits. We created a ubiquitous fmo-4 OE worm strain with fmo-4 expressed under the eft-3 promoter via multicopy extrachromosomal array followed by random integration. qPCR analysis shows that fmo-4 is overexpressed ∼150 fold over WT in the ubiquitous fmo-4 OE worms (Supplementary Table 1), and their development time is slightly (∼1.5 hours) but statistically significantly delayed compared to WT (Supplementary Fig 3). Interestingly, we find that ubiquitous fmo-4 OE is sufficient to reproducibly extends lifespan by ∼20% (Fig 3A). Epistatic analysis with fmo-2 shows that fmo-4 OE lifespan extension is not additive with fmo-2 OE (Supplementary Fig 4A), as determined by comparing the individual OE strains to the fmo-2 OE;fmo-4 OE (fmo- 2OE;4OE) worm strain (Supplementary Table 1). This further supports that these two genes in the same family act in the same pathway. To test whether fmo-4 requires fmo-2, we overexpressed fmo-4 in the context of fmo-2 knockout (fmo-4 OE;2KO) (Supplementary Table 1). The results show that fmo-4 OE does not require fmo-2 for its longevity benefits (Supplementary Fig 5A). This further validates that fmo-4 likely acts downstream of fmo-2.

Figure 3.

Increased longevity is frequently observed in tandem with increased healthspan and resistance to toxic stress²⁴. We assessed the healthspan of the ubiquitous fmo-4 OE worms at days 2 and 10 of adulthood by measuring thrashing rates (movement). We find that ubiquitous fmo-4 OE is sufficient for an improvement (p = 0.011) in healthspan benefits with age (Fig B-C). We next measured the stress resistance of the ubiquitous fmo-4 OE worms to 5mM paraquat (oxidative stress), 37°C heat stress, 5µg/mL of tunicamycin (ER glycosylation stress), and 1mg/mL thapsigargin (ER calcium stress). We included two ER stresses because FMO-4 is predicted to be an ER transmembrane protein. We find that ubiquitous fmo-4 OE is sufficient to convey paraquat stress resistance (Fig 3D) but does not offer any protection from heat (Fig 3E) or tunicamycin stress (Fig 3F). Interestingly, ubiquitous fmo-4 overexpressing animals are sensitive to thapsigargin, as shown by the worms’ inability to develop compared to WT control worms (Fig 3G-I). These results suggest a narrow stress resistance and even some sensitivity for fmo-4 in comparison to other longevity genes. To test whether fmo-4 OE would interact with fmo-2 OE, we combined the strains and measured their stress resistance. We find that the fmo- 2 OE strain is still resistant to paraquat, heat, and tunicamycin stress in the presence of increased fmo-4, suggesting that overexpression of fmo-4 does not affect resistance to these stresses (Supplementary Fig 4B-D). Interestingly, the fmo-2 OE strain is sensitive to thapsigargin stress in the presence of increased fmo-4, suggesting that fmo-4 likely also acts downstream of fmo-2 in this stress pathway (Supplementary Fig 4E-I). Overall, these results suggest that fmo-4 1) is sufficient to improve healthspan, 2) is both necessary and sufficient to promote paraquat resistance, 3) is necessary but not sufficient to improve heat and tunicamycin resistance, and 4) promotes sensitivity to calcium-mediated ER stress from thapsigargin. As expected under the hypothesis that fmo-4 acts downstream of fmo-2, the fmo-4OE;2KO strain phenocopies the fmo-4 OE strain in all stress assays (Supplementary Fig 5B-I). Together, these data suggest that fmo-4 OE is sufficient to promote lifespan extension and paraquat stress resistance as well as a healthspan benefit, while exhibiting sensitivity to ER calcium stress.

Hypodermal overexpression of fmo-4 is sufficient for longevity and paraquat resistance

While using the ubiquitous overexpressing worm strain is helpful to probe into fmo-4’s involvement in longevity, stress resistance, and healthspan, it is also important to note that fmo- 4 is normally localized and expressed primarily in the hypodermis (Fig 4A-C)²⁵. Thus, we asked whether hypodermal-specific fmo-4 overexpression is sufficient for these health benefits. We created a hypodermal-specific fmo-4 OE worm strain expressing fmo-4 under the dpy-7 promoter via multicopy extrachromosomal arrays followed by random integration. qPCR data show that fmo-4 is expressed ∼45 fold over WT in the hypodermal-specific fmo-4 OE worms (Supplementary Table 1), and their developmental time does not differ from WT (Supplementary Fig 3). We measured lifespan, resistance to paraquat, heat, tunicamycin and thapsigargin stress, as well as healthspan of the hypodermal-specific fmo-4 OE worms (fmo-4 OE^Hyp). We find that the fmo-4 OE^Hyp strain exhibits a lifespan extension (Fig. 4D), resistance to paraquat (Fig. 4E) but not heat or tunicamycin stress (Fig. 4F-G), and sensitivity to thapsigargin stress (Fig. 4H-J). However, the fmo-4 OE^Hyp strain did not have a statistically significant effect on healthspan unlike the ubiquitous fmo-4 OE strain (Fig 4K-L). Overall, these results agree with the C. elegans single cell atlas as well as previously published work²⁵ showing that fmo-4 is mostly localized to the hypodermis (Fig 4A) and are consistent with the ubiquitous strain primarily benefiting from expressing additional fmo-4 in the hypodermis. Thus, ubiquitous or hypodermal-specific overexpression of fmo-4 is sufficient for longevity and paraquat stress resistance (Fig 3A,D, Fig 4D-E), does not affect heat or tunicamycin resistance (Fig 3E-F, Fig 4F-G), and results in thapsigargin sensitivity (Fig 3G-I, Fig 4H-J).

Figure 4.

fmo-4 OE transcriptomics reveal a link to calcium regulation

To identify the downstream effects of fmo-4 expression, we analyzed the transcriptome of fmo-4 OE animals and compared them with control animals (Fig 5A). Based on a p-value of < 0.05, we find ∼800 transcripts are upregulated and ∼500 transcripts are downregulated when fmo-4 is overexpressed (Supplemental Data 1). Of the upregulated and downregulated transcripts unique to the fmo-4 OE profile, we noticed that some of the significant pathways include transcription, Wnt signaling, TGFβ signaling, protein processing in the ER, and other subsets of signaling, as determined by the Database for Annotation, Visualization and Integrated Discovery (DAVID) Functional Annotation Bioinformatics Microarray Analysis (Fig 5A). Based on its known role as a xenobiotic enzyme, we were intrigued to find that fmo-4 OE is modulating signaling pathways. We were further intrigued to find that one of the signaling pathways affected by fmo-4 is calcium signaling (Supplementary Table 2). This was interesting considering that fmo-4 OE worms are sensitive to thapsigargin (Fig 3G-I), implicating an interaction between calcium signaling and fmo-4. To further verify this possibility, we used another tool called PANTHER Classification System Protein Class, which identified FMO-4 as a putative transmembrane signal receptor (G-protein coupled receptor), ion transporter, and/or calcium-binding protein (Supplemental Data 2). Together, these results support the possibility that FMO-4 is involved in calcium-related processes.

Figure 5.

To begin exploring this interaction, we determined how changes in intracellular calcium levels impact fmo-4 expression and lifespan. We first manipulated calcium levels by supplementing the acetylcholine agonist, carbachol. Carbachol activates acetylcholine receptors, increasing overall intracellular calcium levels²⁶. We measured fmo-4 gene expression upon exposure to carbachol using an fmo-4p::mCherry transcriptional reporter strain.

Interestingly, we find that supplementation of 300µM carbachol induces fmo-4 promoter expression fluorescence nearly 2-fold over a water control (Fig 5B-C). This was similar to the level of induction of fmo-4 in fasted (DR-like) conditions (Fig 5B-C). We postulate that fmo-4 is induced by carbachol because increased intracellular calcium activates fmo-4 gene expression. Since carbachol supplementation induces fmo-4, we hypothesized that carbachol may affect lifespan and would interact with fmo-4. We measured lifespan of worms supplemented with 50µM carbachol and find that while WT lifespan is significantly extended by carbachol, this extension is not additive with fmo-4 OE (Fig 5D), and requires fmo-4, as the fmo-4 KO is shorter-lived when exposed to carbachol (Fig 5E).

To test whether reducing calcium has an opposing effect to increasing it, we utilized EthyleneDiamineTetraAcetic acid (EDTA) to deplete calcium levels²⁷. First, we placed fmo-4p::mCherry reporter worms on plates containing 10mM EDTA and measured fluorescence. Surprisingly, we find that EDTA also significantly induces fmo-4 expression (Fig 5B-C). We then assessed the survival of WT, fmo-4 OE and KO worms on plates supplemented with 50µM EDTA. We find that EDTA significantly extends WT lifespan without further extending fmo-4 OE lifespan (Fig 5F), while shortening the lifespan of fmo-4 KO animals (Fig 5G). Overall, these data suggest that fmo-4 is highly sensitive to changing calcium levels in either direction, and that both increasing and depleting calcium 1) induces fmo-4, 2) extends WT lifespan, 3) is not additive with fmo-4 OE lifespan, and 4) is deleterious to fmo-4 KO animals. Thus, these results support an interaction between fmo-4 and calcium signaling.

fmo-4 genetically interacts with calcium signaling to promote longevity and paraquat resistance

Having established an interaction between fmo-4 and calcium perturbations, we next asked if fmo-4 interacts with genes involved in calcium signaling. These genes include calreticulin (crt-1), inositol triphosphate receptor (IP₃R, itr-1), and the mitochondrial calcium uniporter (mcu-1)¹⁶. Calreticulin is responsible for binding and sequestering calcium in the ER lumen^12,13. It was previously shown that FMO protein extracted from rabbit lung can form a complex with calreticulin, suggesting a physical interaction²⁸. When testing for genetic interactions between crt-1 and fmo-4, we find that crt-1 RNAi extends WT lifespan, as previously reported¹⁶, and that this lifespan extension is not additive with fmo-4 OE (Fig 6A). Treating fmo- 4 KO worms with crt-1 RNAi can significantly extend lifespan (Supplementary Fig 6A), but this result was inconsistent (Supplemental Data 3). Overall, this suggests that fmo-4 and crt-1 are acting in the same genetic pathway. IP₃R, or itr-1 in C. elegans, is located in the ER membrane and is critical for exporting calcium from the ER lumen into the cytosol^12,14. Previous results suggest that loss of itr-1 decreases WT lifespan¹⁶, which our results confirm (Fig 6B).

Figure 6.

Importantly, we also find that itr-1 RNAi decreases lifespan of fmo-4 KO worms similarly (Supplementary Fig 6B) and fmo-4 OE (Fig 6B) worms to a greater extent than WT, negating the relative longevity of fmo-4 OE worms compared to WT. Thus, itr-1 is required to increase lifespan downstream of fmo-4. MCU (mcu-1) is located in the inner mitochondrial membrane and allows cytosolic calcium into the inner mitochondrial matrix^12,15. We also find that mcu-1 is required downstream of fmo-4 OE, as mcu-1 RNAi abrogates the fmo-4 OE lifespan (Fig 6C). The fmo-4 KO lifespan is slightly decreased on mcu-1 RNAi (Supplementary Fig 6C). These data support that fmo-4 extends lifespan through its interactions with the calcium signaling genes, crt-1, itr-1, and mcu-1.

In addition to lifespan assessment, we were curious whether these genes also interact with fmo-4 expression to modify resistance to paraquat. We exposed WT, fmo-4 OE, and fmo-4 KO worms to 5mM paraquat plates with RNAi for crt-1, itr-1 or mcu-1 and assessed survival over time. Interestingly, we find that crt-1 RNAi promotes resistance in WT worms and this effect is not additive in fmo-4 OE worms (Fig 6D). Further, crt-1 RNAi extends the lifespan of fmo-4 KO worms (Supplementary Fig 6D), suggesting that crt-1 and fmo-4 act in the same pathway and that crt-1 may act downstream of fmo-4 to promote paraquat stress resistance. Neither itr-1 RNAi nor mcu-1 RNAi confers resistance to paraquat, and fmo-4 OE resistance is lost when exposed to these RNAi (Fig 6E-F). The fmo-4 KO worms show a decrease in paraquat survival when treated with itr-1 and mcu-1 RNAi (Supplementary Fig 6E-F). These data support a model where fmo-4 OE converges onto a shared pathway mediating paraquat stress resistance and longevity regulation. Together, we conclude that fmo-4 interacts with ER to mitochondrial calcium signaling through crt-1, itr-1, and mcu-1, to promote longevity and paraquat stress resistance.

atf-6 KD regulates fmo-4-mediated longevity and paraquat resistance

Our data establish that fmo-4 interacts with ER and mitochondrial calcium signaling to promote lifespan extension and resistance to paraquat. A previous study linked many of these components to a major regulator of UPR^ER, activating transcription factor-6 (atf-6)¹⁶. When atf-6 is lost, lifespan is extended through changes in calcium signaling between the ER and mitochondria requiring crt-1, itr-1 and mcu-1¹⁶. Based on our transcriptomics data and fmo-4’s interactions with calcium signaling genes, we hypothesized that fmo-4 could interact with atf-6 to promote longevity and paraquat resistance. To test whether atf-6 is regulating fmo-4, we utilized our fmo-4p::mCherry transcriptional reporter strain. We measured fmo-4 gene expression after RNAi knockdown of atf-6 and find that fmo-4p::mCherry worms on atf-6 RNAi show a consistent ∼2-fold increase in fluorescence (Fig 7A-B), similar to what we observe from calcium perturbations (Fig 5B-C). Since atf-6 is one of three branches of UPR^ER, we were curious to see if fmo-4 interacts specifically with atf-6 or also with the other two branches, ire-1/xbp-1 and pek- 1/atf-4¹¹. We treated the fmo-4p::mCherry reporter worms with ire-1, xbp-1, pek-1, or atf-4 RNAi and measured fluorescence. We find that knocking down the components of the IRE-1 or PEK-1 branches does not induce fmo-4 gene expression (Supplementary Fig 7A-B). Thus, only knockdown of the ATF-6 branch of UPR^ER induces fmo-4 expression.

Figure 7.

We hypothesized that atf-6 limits fmo-4 expression and thus fmo-4 acts downstream of atf-6 knockdown to modulate lifespan extension. To test this, we assessed survival and find that while atf-6 RNAi extends the lifespan of WT worms, as reported¹⁶, this effect is abrogated when fmo-4 is knocked out (Fig. 7C). This result suggests that fmo-4 is indeed acting downstream of atf-6 to promote longevity. To determine if atf-6 is also involved in fmo-4-mediated paraquat resistance, we exposed WT and fmo-4 KO worms to 5mM paraquat plates with atf-6 RNAi. We find that atf-6 RNAi promotes stress resistance to paraquat, but this effect is lost when fmo-4 is knocked out, supporting the hypothesis that atf-6 is involved in fmo-4-mediated paraquat resistance (Fig 7D). Together, these data suggest that fmo-4 modulates lifespan and paraquat stress resistance downstream of the reduction in atf-6, ultimately regulating ER to mitochondria calcium signaling (Fig 7E).

Discussion

Together, our data present a model where fmo-4 acts downstream of DR and mTOR to positively affect healthspan and longevity (Fig 1B-C). fmo-4 overexpression is sufficient to provide these benefits (Fig 3) and does not require the DR-mediating family member, fmo-2 (Fig 2). Our results also suggest that fmo-4 gene expression is upregulated upon changes in intracellular calcium, and that fmo-4 interacts with calcium signaling through key genes like crt- 1, itr-1, and mcu-1, to promote longevity and paraquat stress resistance (Fig 5-6). The relationship between fmo-4 and calcium is further illustrated by fmo-4 OE’s susceptibility to thapsigargin, a calcium-mediated ER stress. Furthermore, we find that knocking down the UPR^ER transcription factor and calcium regulator, atf-6, induces fmo-4 gene expression and requires fmo-4 to promote longevity (Fig 7). Collectively, our data suggest that fmo-4 promotes longevity and paraquat stress resistance by regulating calcium homeostasis between the ER and mitochondria (Fig 7E).

Previous studies from our lab have elucidated how C. elegans fmo-2 promotes longevity through endogenous metabolism¹⁰. Based on the conservation within the fmo gene family and that multiple Fmos are induced in long-lived worms and mammals, it was reasonable to hypothesize that Fmos could play overlapping roles in longevity regulation. Our data here support the broader hypothesis that Fmos have a conserved role in aging, but interestingly, that the mechanisms by which they modulate aging are separable. Importantly, we find that fmo-4 is required for multiple longevity-promoting pathways, including DR and mTOR, but is not required for lifespan extension through the hypoxic response, insulin-like signaling, or cytochrome c reductase pathways. We also find that fmo-4 does not require fmo-2, but that fmo-2 does require fmo-4. This is interesting because not only do these data tell us that fmo-4 is important in the context of longevity, but also that it acts distinctly from fmo-2.

After creating ubiquitous and hypodermal-specific fmo-4 overexpressing strains, we find that fmo-4 is sufficient for longevity and paraquat stress resistance but not heat, tunicamycin, or thapsigargin stress. These results further differentiate fmo-4 from fmo-2 and also provide insight into fmo-4’s potential roles in the cell. For instance, fmo-4 OE worms’ resistance to paraquat could suggest a role in responding to or controlling reactive oxygen species (ROS) production in the mitochondria. Additionally, fmo-4 OE worms’ sensitivity to thapsigargin likely points towards an involvement in ER calcium regulation. These are both interesting plausible roles for fmo-4, and may go hand in hand, as calcium levels are known to impact ROS levels²⁹, and changes in ROS levels are known to impact ER calcium release²⁹. Based on this and the data with calcium perturbations, we speculate that FMO-4 acts as a calcium sensor, and when calcium levels are too high or too low, FMO-4 responds by regulating downstream calcium signaling proteins to restore homeostasis in the cell. As FMO-4 is a predicted ER transmembrane protein, we hypothesize that FMO-4 is sensing changes in calcium levels in the ER and/or cytosol.

We note that our data reveal genetic interactions between fmo-4 and other longevity pathways like DR, mTOR, FMO-2, and ATF-6. However, it is important to also understand on the protein level how FMO-4 is interacting with these pathways and how FMO-4 is regulating downstream calcium signaling components to promote longevity and paraquat stress resistance. For instance, it was previously shown that rabbit FMO protein forms a complex with calreticulin²⁸. It is possible that an increase in FMO-4 leads to more binding between FMO-4 and calreticulin, which ultimately prevents calreticulin from binding ER calcium. This would then allow for proper calcium flux from the ER to the mitochondria via the IP₃R and MCU, respectively. Our future work will look into these protein interactions so that we can further tease apart the FMO-4-mediated longevity pathway.

While our transcriptomics analysis and calcium manipulations suggest fmo-4 involvement in calcium regulation, there are other measures that can better assess an interaction between fmo-4 and calcium. Carbachol and EDTA supplementation are not perfect assessments of altered calcium levels, as carbachol does not act in every tissue and EDTA may be depleting more than just calcium from the cell. Thus, future studies should measure calcium levels and calcium flux in fmo-4 OE and KO worms using GCaMP expressing worms to determine the direct link between fmo-4 expression and calcium in the cell. Additionally, to determine if FMO-4 is acting as a calcium sensor or interacting with one to regulate calcium homeostasis, development of FMO-4 antibodies for immunoprecipitation assays would be useful³². Together, it will be important to establish the biochemical parameters of the FMO-4 calcium interaction.

As calcium signaling occurs between the ER and the mitochondria, our results could have interesting implications in ER-mitochondrial metabolism. We would expect that FMO-4 regulates ER-mitochondria metabolism because 1) FMO-4 is a predicted ER transmembrane protein and blocking the expression of another ER transmembrane protein, ATF-6, induces the expression of fmo-4, 2) fmo-4 OE worms are highly sensitive to ER calcium stress but resistant to ROS induced stress, 3) lifespan extension driven by fmo-4 depends on mitochondrial calcium import (mcu-1), and 4) Fmo gene expression is induced by mitochondrial inhibitors³³.

Additionally, changes in mitochondrial calcium import affect various mitochondrial measures like respiration, dynamics (i.e., fission and fusion), membrane potential, and TCA cycle activity³⁴.

Future studies will delve into how fmo-4 perturbations impact each of these measures and if they act in the fmo-4-mediated longevity pathway. Based on our data, we speculate that fmo-4 OE may increase mitochondrial calcium by regulating mcu-1 activity, and that fmo-4 OE worms may in turn be regulating ROS production in the mitochondria. We hypothesize that changes in fmo-4 expression will alter mitochondrial respiration, dynamics, membrane potential, and TCA cycle activity. Taken together, future experiments will shed light on the role that FMOs play in calcium regulation and mitochondrial metabolism and will help to further tease apart the FMO- mediated longevity pathway.

Methods and Materials

Strains and Growth Conditions

Standard C. elegans cultivation procedures were used as previously described⁶. Briefly, all worm strains were maintained on solid nematode growth media (NGM) using E. coli OP50 throughout life except where double stranded (ds) RNAi (E. coli HT115) were used. Worms were transferred using a platinum wire. All worm strains were kept at 20°C. RNAi used is listed in Supplementary Table 3. Worm strains are listed in Supplementary Table 4. Genotyping primers are listed in Supplementary Table 5.

Development Assays

Animals were synchronized by placing 10 gravid adult worms on NGM plates seeded with E. coli OP50 to lay eggs for 1 hour at 20°C. The gravid adult worms were then removed, and the eggs were allowed to hatch and develop at 20°C until larval stage 2 (L2). At this point the L2 worms were moved to individual 35mm NGM plates seeded with E. coli OP50, one worm per plate. This was done for each worm strain tested and ten total 35mm plates per worm strain were prepared. The worms were then followed through development and watched hourly during young adulthood to score the time of the first egg lay, as previously described⁶. Development time was reported in hours since egg lay. Three replicate experiments were performed.

Stress Resistance Assays

Paraquat stress assay

Paraquat (Methyl viologen dichloride hydrate, 856177, Sigma-Aldrich) was used to induce oxidative stress. Worms were synchronized from eggs on either NGM plates seeded with E. coli OP50 or RNAi plates seeded with HT115 strain expressing dsRNAi for a particular gene. At L4 stage, 30 worms were transferred to either NGM plates or RNAi-FUdR (40690016, Bioworld) plates containing 5mM paraquat dissolved in water. Three plates per strain per condition were prepared, for a total of 90 worms per condition. As previously described, worms were then scored every other day and considered dead when they did not move in response to prodding under a dissection microscope⁶. Worms that crawled off the plate were not considered, but ruptured worms were noted and considered. Three replicate experiments were performed. Log-rank test was used to derive p-value for paraquat stress resistance survival assays using p < 0.05 cut-off threshold compared to EV or wild-type controls. Cox regression was also used to assess interactions between genotype and condition for paraquat stress resistance survival assays using p < 0.01 cut-off threshold compared to controls.

Heat stress assay

Worms were synchronized from eggs on NGM plates seeded with E. coli OP50. Once the worms reached day 1 of adulthood, 25 worms of each strain were transferred to 35mm NGM plates seeded with E. coli OP50¹⁶. Four plates per strain per condition were prepared, for a total of 100 worms per condition. The plates were placed in a single layer on a tray and incubated at 37°C for 3 hours. After 3 hours, the plates were removed and placed at 20°C for 24 hours to give the worms time to recover. After 24 hours, the number of dead and alive worms was counted. % Alive was calculated as (# alive / # total)×100 and graphed in Graphpad Prism using unpaired two-tailed t tests with Welch’s correction as well as one-way ANOVA to determine significance. Three replicate experiments were performed.

Tunicamycin stress assay

E. coli OP50 was cultured at 37°C shaking overnight and then centrifuged and concentrated to 100x. 100µL of bacteria was added to 9.9mL of S-media to create a 1x mixture. 100µL of this bacteria/S-media mixture was added to the wells of a Falcon non tissue culture treated clear flat bottom 96 well plate (Falcon, 351172). Approximately 20 eggs per worm strain were added to the wells of the 96 well plate as previously described⁶. Three replicates per worm strain were added to the 96-well plate, for a total of 60 eggs per worm strain. Then either 0, 1, 2, or 5µg/mL of tunicamycin (Sigma, T7765) diluted in dimethyl sulfoxide (DMSO) (Sigma, D2650) was added to the wells. Total volume per well was 100µL. The plate was incubated covered at room temperature on an orbital shaker for 72 hours. After 72 hours, each well containing worms was rinsed with 100µL M9 and then added to individual 35mm NGM plates seeded with E. coli OP50 to sit covered at room temperature overnight. The following day, plates were assessed for live adults. % Alive Adults was calculated by determining how many of the three technical replicates of a given strain showed at least one live adult and then graphed in Microsoft Excel. Three replicate experiments were performed.

Thapsigargin stress assay

To assess development of worms during chronic ER calcium stress, NGM plates seeded with E. coli OP50 were spotted with 25µL 1mg/mL thapsigargin in DMSO (Sigma) or DMSO only directly on to the E. coli OP50 lawns and allowed to dry overnight. Then, 30 synchronized L1’s per worm strain were added directly to the spotted lawns¹⁶. Forty-eight hours later, worms were picked off these plates, added to a 2% agarose pad on a glass slide, anesthetized in 0.5M sodium azide (Sigma), and imaged at 6.3x magnification (brightfield) with the LASx software and Leica scope. Three replicates were performed. Each worm was recorded in ImageJ. Data were analyzed in GraphPad Prism using unpaired two-tailed t tests with Welch’s correction.

Thrashing Assays

Animals were synchronized by placing 10 gravid adult worms on NGM plates seeded with E. coli OP50 and allowing them to lay eggs for 2 hours at 20°C. The gravid adult worms were then removed, and the eggs were allowed to hatch and develop at 20°C until day 1 adulthood.

Worms were placed in a drop of M9 solution, as previously described⁶. The individual body bends were counted at maximum rate for 30 seconds. Thrashing was assayed on day 2 and day 10 of adulthood. The animals that were not used for the day 2 assay were transferred to fresh fed FUdR plates 4 times until they were ready to be assayed. Three replicates were performed. Data were analyzed in GraphPad Prism using unpaired two-tailed t tests with Welch’s correction as well as one-way ANOVA.

Lifespan Assays

Gravid adults were placed on NGM plates containing 1mM β-D-isothiogalactopyranoside (IPTG), 25μg/ml carbenicillin, and the corresponding RNA interference (RNAi) clone from the Vidal or Ahringer RNAi library. 200µL of HT115 bacteria expressing double stranded (ds) RNA of either the control empty vector (EV) or the RNAi of interest at optical density (OD) of 3.0 and concentration of 3X was added to each plate. After 3 hours, the adults were removed, and the eggs were allowed to develop at 20°C until they reached late L4/young adult stage. From here, 70 worms were placed on each RNAi plate and transferred to fresh RNAi + FUDR plates on day 1, day 2, day 4, and day 6 of adulthood. A minimum of two plates per strain per condition were used per replicate experiment. Experimental animals were scored every 2-3 days and considered dead when they did not move in response to prodding under a dissection microscope. Worms that crawled off the plate were not considered, but ruptured worms were considered as previously described⁶. A similar method was used for non-RNAi lifespan experiments, except the plates did not contain IPTG and worms were fed E. coli OP50. 200µL of E. coli OP50 at OD of 3.0 and concentration of 3X was added to each plate. A similar method was also used for carbachol (ThermoFisher, L06674.14) and ethylenediaminetetraacetic acid (EDTA) (ThermoFisher, AM9260G) supplementation lifespan experiments, except 50µM carbachol or 50µM EDTA was added to the NGM plates without IPTG, and worms were fed E. coli OP50. 200µL of E. coli OP50 at OD of 3.0 and concentration of 3X was added to each plate. Optimal concentrations of carbachol and EDTA (50µM) were determined by assessing survivability of worms exposed to a range of concentrations. Log-rank test was used to derive p- value for lifespan assays using p < 0.05 cut-off threshold compared to EV or wild-type controls. Cox regression was also used to assess interactions between genotype and condition for lifespan assays using p < 0.01 cut-off threshold compared to controls.

Dietary Restriction (sDR) Lifespan Assay

Gravid adults were placed on NGM plates seeded with 200µL E. coli OP50. After 3 hours, the adults were removed, and the eggs were allowed to develop at 20°C until they reached late L4/young adult stage. From here, 70 worms were transferred to fed FUdR plates seeded with 200µL of E. coli OP50 at OD of 3.0 and concentrated 3X on days 1 and 2 of adulthood. On day 3 of adulthood, DR conditions were transferred to plates with 10⁹ seeded lawns and transferred every other day four times while the corresponding controls were transferred equally to fed ad- lib plates. This form of DR is termed solid DR (sDR)¹⁹. 80µL of 10mM palmitic acid (Sigma- Aldrich) dissolved in 100% EtOH was added to the outer rim of the plate to prevent fleeing. A minimum of two plates per strain per condition were used per replicate experiment.

Experimental animals were scored every 2-3 days and considered dead when they did not move in response to prodding under a dissection microscope. Worms that crawled off the plate were not considered, but ruptured worms were considered as previously described⁶.

Transcriptomic Analysis

Approximately 600 day 1 adult worms per biological replicate were washed with M9 buffer three times, frozen in liquid nitrogen, and then stored at −80°C. RNA was extracted by adding 500mL of Trizol reagent to the frozen worm pellets, followed by three freeze-thaw cycles with liquid nitrogen and water bath at 42°C. Then, 500µL of ethanol was added to the samples, and RNA was isolated using the Direct-Zol Miniprep Plus Kit (Cat#R2072). Purified RNA was sent to Novogene (Novogene Corporation Inc) for sequencing on the Illumina HWI-ST1276 instrument. Messenger RNA was purified from total RNA by using poly-T oligo-attached magnetic beads.

After fragmentation, first strand cDNA was synthesized using random hexamer primers, followed by second strand cDNA. The library was constructed, which entailed end repair, A- tailing, adapter ligation, size selection, amplification, and purification, and then was sequenced on an Illumina device using paired-end sequencing. Gene ontology analysis was done using the National Institutes of Health DAVID Bioinformatics tool.

Gene Expression Assays

RNA was isolated from day 1 adult worms (approximately 600 worms per strain) following three rounds of freeze-thaw in liquid nitrogen using Invitrogen’s Trizol extraction method, similar to the method described above. 1µg of isolated and purified RNA was reverse transcribed to cDNA using SuperScript™ II Reverse Transcriptase (18064071, Invitrogen,). Gene expression levels were measured using 600ng of cDNA and SYBR^TM Green PCR Mastermix (A25742, Applied Biosystems) with primers at 10μM concentration. mRNA levels were normalized using Y45F10D.4 as a reference gene³⁵. List of qPCR primers used are in Supplementary Table 6.

fmo-4 induction on RNAi

Gravid fmo-4p::mCherry transcriptional reporter adult animals were placed on NGM plates containing 1mM β-D-isothiogalactopyranoside (IPTG), 25 μg/ml carbenicillin, and the corresponding RNAi clone from the Vidal or Ahringer RNAi library. After 3 hours, the adults were removed, and the eggs were allowed to develop at 20°C until they reached late L4/young adult stage. Then 40-50 worms per plate were transferred to similar plates that contain FUdR for overnight. The following day, ∼20 worms per condition were picked off these plates, added to a 2% agarose pad on a glass slide, anesthetized in 0.5M sodium azide (Sigma), and imaged at 6.3x magnification with the LASx software and Leica scope using the mCherry fluorescence channel³⁶. Three replicates were performed. Fluorescent intensity in the mCherry channel was measured in ImageJ. Brightness of images within a dataset was increased to the same level. Data were analyzed in GraphPad Prism using unpaired two-tailed t tests with Welch’s correction.

fmo-4 induction on Carbachol

Gravid fmo-4p::mCherry transcriptional reporter adult animals were placed on NGM plates seeded with E. coli OP50. After 3 hours, the adults were removed, and the eggs were allowed to develop at 20°C until they reached late L4/young adult stage. Then 30 worms were transferred to NGM plates seeded with E. coli OP50 also containing either 300µM carbachol (ThermoFisher, L06674.14) or water. An optimal concentration of carbachol was determined by first assessing survivability of worms exposed to a range of concentrations and then by assessing fluorescence of the fmo-4p::mCherry reporter worms exposed to a range of concentrations. The worms were incubated for 24 hours at 20°C. After 24 hours, ∼20 worms per condition were then picked off these plates and added to unseeded NGM plates, anesthetized in 0.5M sodium azide (Sigma), and imaged at 6.3x magnification with the LASx software and Leica scope using the mCherry fluorescence channel³⁶. Three replicates were performed. Each worm was measured for fluorescence in ImageJ. Data were analyzed in GraphPad Prism using t tests.

fmo-4 Induction Assay on EDTA

Gravid fmo-4p::mCherry transcriptional reporter adult animals were placed on NGM plates seeded with E. coli OP50. After 3 hours, the adults were removed, and the eggs were allowed to develop at 20°C until they reached late L4/young adult stage. Then 30 worms were transferred to NGM plates seeded with E. coli OP50 topically spotted with 150µL of 0.5M ethylenediaminetetraacetic acid (EDTA) (ThermoFisher, AM9260G) or 150µL of water. An optimal concentration of EDTA was determined by first assessing survivability of worms exposed to a range of concentrations and then by assessing the fluorescence of the fmo- 4p::mCherry reporter worms exposed to a range of concentrations. The worms were incubated for 24 hours at 20°C. After 24 hours, ∼20 worms per condition were then picked off these plates and added to unseeded NGM plates, anesthetized in 0.5M sodium azide (Sigma), and imaged at 6.3x magnification with the LASx software and Leica scope using the mCherry fluorescence channel³⁶. Three replicates were performed. Each worm was recorded in ImageJ. Data were analyzed in GraphPad Prism using t tests.

Statistical Analyses

Log-rank test was used to derive p-value for lifespan and stress resistance survival assays using p < 0.05 cut-off threshold compared to EV or wild-type controls. Cox regression was also used to assess interactions between genotype and condition for lifespans and stress resistance survival assays using p < 0.01 cut-off threshold compared to controls. Supplemental Data 3 provide the results of the Log-rank test and Cox regression analysis, which were run in Rstudio.

Acknowledgements

This work was supported by grants from NIH. AMT was supported by NIH T32AG000114 and the University of Michigan Rackham Research Grant. SFL was supported by R01AG075061 and the Glenn Foundation for Medical Research.