Carnitine promotes recovery from oxidative stress and extends lifespan in C. elegans

Carnitine is required for transporting fatty acids into the mitochondria for β-oxidation. Carnitine has been used as an energy supplement but the roles in improving health and delaying aging remain unclear. Here we show in C. elegans that L-carnitine improves recovery from oxidative stress and extends lifespan. L-carnitine promotes recovery from oxidative stress induced by paraquat or juglone and improves mobility and survival in response to H2O2 and human amyloid (Aβ) toxicity. L-carnitine also alleviates the oxidative stress during aging, resulting in moderate but significant lifespan extension, which was dependent on SKN-1 and DAF-16. Long-lived worms with germline loss (glp-1) or reduced insulin receptor activity (daf-2) recover from aging-associated oxidative stress faster than wild-type controls and their long lifespans were not further increased by L-carnitine. A new gene, T08B1.1, aligned to a known carnitine transporter OCTN1 in humans, is required for L-carnitine uptake in C. elegans. T08B1.1 expression is elevated in daf-2 and glp-1 mutants and its knockdown prevents L-carnitine from improving oxidative stress recovery and prolonging lifespan. Together, our study suggests an important role of L-carnitine in oxidative stress recovery that might be important for healthy aging in humans.


INTRODUCTION
Carnitine, a water-soluble quaternary amine (3-hydroxy-4-N, N, N-trimethyl amino butyric acid), is synthesized from the essential amino acids lysine and methionine [1,2]. Cells utilize L-carnitine for transporting long chain fatty acids into mitochondria for β-oxidation [3]. L-carnitine is essential for lipid metabolism in all cell types and highly enriched in skeletal and cardiac muscle [4]. Various forms of carnitine, including L-carnitine, acetyl-L-carnitine, and propionyl-L-carnitine, have long been used as dietary supplements for weight loss and enhancing performance [5]. Impaired carnitine uptake due to genetic mutations in the carnitine-transporters can cause cardiomyopathy, skeletal-muscle weakness, and hypoglycemia [6][7][8]. Certain conditions such as chronic renal failure and antibiotics abuse can also reduce carnitine retention in the body, manifesting symptoms of carnitine deficiency [8,9]. Carnitine has been approved by the Food and Drug Administration (FDA) to treat primary and certain secondary carnitine-deficiency syndromes [10].
Carnitine levels decrease during aging, which may contribute to the decline in mitochondrial function and aging [11]. In rats, dietary supplementation with high doses of acetyl-L-carnitine and alpha-lipoic acid improves mitochondrial function and decreases oxidative stress, resulting in improved performance on memory-requiring AGING tasks [12][13][14]. In humans, several studies suggest that carnitine could improve mental health in older adults with mild cognitive impairment or Alzheimer's disease [15]. Carnitine may also improve other age-associated diseases such as cardiovascular diseases [16][17][18][19], atherosclerosis [20], neurometabolic disorders [21] and diabetic symptoms [22][23][24][25]. In addition, carnitine has also been used in adjuvant therapy to treat conditions such ashemodialysis [26]. Acetyl-L-carnitine can extend chronological lifespan of budding yeast [27,28]. Most of the therapeutic effect of carnitine have been linked to cellular redox balance [29,30]. However, the detailed mechanisms of action at the cellular and molecular levels have not been fully investigated.
The round worms C. elegans has been widely used to study oxidative stress response and aging [31,32]. In response to chemicals that generate reactive oxygen species (ROS) such as paraquat or juglone, C. elegans activates SKN-1 and DAF-16 dependent transcription programs to alleviate oxidative damage, which leads to lifespan extension [33,34]. Activation of the SKN-1 pathway contributes to the extended lifespan of glp-1 mutant lacking germline and daf-2 mutant with impaired insulin receptor [35,36]. The roles of SKN-1 and DAF-16 are conserved in mammals (Nrf2 and FOXO3, respectively), providing useful targets for disease intervention [37]. C. elegans has also been used for study human Aβ toxicity, in which ectopic expression of human Aβ causes paralysis [38]. Very few studies have been carried out in C. elegans regarding carnitine metabolism and function. In one study, L-carnitine supplement in the medium can prevent glucose toxicity on C. elegans survival [39]. In another study, feeding worms with acetylcarnitine reduces age-related neuronal damage and improves learning behavior [40]. Whether carnitine has a role in OSR and aging in C. elegans has not been reported.
In this study, we show that carnitine supplement in the medium facilitates recovery from oxidative stress during aging, promoting resistance to oxidative toxicity induced by H2O2 and human amyloid protein Aβ  aggregates. Carnitine's beneficial effect is dependent on SKN-1, DAF-16 and a potential carnitine transporter T08B1.1. T08B1.1 expression is increased in the longevity mutant glp-1 and required for the long lifespan. Our study is the first of its kind to report the effect of carnitine on aging in C. elegans and reveal novel roles of carnitine in extending lifespan.

L-carnitine shortens the length of oxidative stress response (OSR)
It has been shown that L-carnitine has antioxidant activity in vitro [41]. By using paraquat, a well-known ROS generator affecting mitochondrial respiration and dynamics [42,43], we tested if L-carnitine would have antioxidant activity in vivo in C. elegans. We raised the animals on agar plate containing nematode growth (NG) medium supplemented with and without 100 µM of Lcarnitine from L1 stage to L4 stage, then oxidatively stressed the animals with 1mM paraquat for 24 hours on NGM plate. For convenience, we examined the OSR marker gst-4::gfp [44]. We found strong activation of gst-4:gfp by paraquat treatment, indicating a robust oxidative stress condition. L-carnitine supplementation did not significant affect gst-4:gfp induction by paraquat ( Figure 1A, 1B). However, animals treated with Lcarnitine had a better recovery from paraquat, as shown by a faster clearance of the OSR marker gst-4::gfp after transferring to paraquat-free medium ( Figure 1C, 1D). To confirm these results, we also used another ROS generating compound juglone [45]. Similarly, gst-4:: gfp levels decreased faster in animals treated with Lcarnitine compared with non-treated animals ( Figure 1E), suggesting that L-carnitine promotes oxidative stress recovery in C. elegans.

L-carnitine promotes the recovery from oxidative stress
To further confirm the new finding that L-carnitine improves recovery from oxidative stress, we directly measured the ROS levels. Animals raised with and without 100 µM L-carnitine supplement from L1 stage to L4 stage were challenged with 1mM paraquat for 24 hours, then transferred to paraquat-free medium plate. After 12, 24, 36 and 48 hours, animals were stained with Dihydroethidium (DHE), a widely used ROS indicator [35,46]. The results showed that L-carnitine treatment robustly reduced the ROS levels in the intestine after 36 and 48 hours of recovery from paraquat (Figure 2A, 2B).
Next, we tested multiple heath parameters related to oxidative stress. First, we examined if muscle strength could be improved by L-carnitine, by measuring the bending movement. When in liquid, the worms keep bending or thrashing ( Figure 2C), which has been used to indicate the muscle strength and mobility [47]. By treating the animals with paraquat and L-carnitine as mentioned above, we measured the thrashing speed after 48 hours of recovery. The results showed that the impaired mobility by paraquat treatment could be largely rescued by L-carnitine ( Figure 2D). L-carnitine did not obviously improve mobility under normal conditions. Second, we tested if the faster recovery from oxidative stress could result in a better tolerance to oxidative damage by H2O2. Similarly, animals treated with paraquat and L-carnitine were collected 48 hours after recovery from paraquat and incubated with 50mM of H2O2 for 1 hour. We found that paraquat-treated worms AGING Figure 1. L-carnitine shortens the length of oxidative stress response (OSR) induced by paraquat and juglone. (A) L-carnitine did not affect induction of OSR. C. elegans expressing the OSR marker gst-4::gfp were synchronized at L1 larvae stage and raised on NG medium supplemented with or without 10 µM L-carnitine to L4/young adult stage. The ROS generator paraquat was then added to the medium to the final concentration of 1 mM. After 24 hours, animals were imaged with a fluorescent microscope. Representative images were shown. (B) Quantification of images taken from at least 3 independent experiments in (A) by ImageJ and relative expression levels were plotted. Statistical analysis was performed by two-tailed, unpaired student's t-test (ns, not significant. ****, P<0.0001). Error bars indicate the standard deviation of the mean. (C) L-carnitine reduced the gst-4::gfp expression during recovery from oxidative stress. C. elegans worms were prepared and treated as in (A). After 24 hours of paraquat treatment, worms were transferred to fresh plate. OSR marker gst-4::gfp were examined at indicated time. (D) Quantification of gst-4::gfp expression in (C). Images from 3 independent experiments were quantified using ImageJ and normalized to the value at time 0. Statistical analysis was performed by two-tailed, unpaired student's t-test (*, P<0.05. ***, P<0.001). Error bars indicate standard deviation of the mean. (E) L-carnitine reduced gst-4::gfp expression during recovery from juglone treatment. C. elegans worms were prepared and treated as in (C) except that 300 µM juglone (final concentration) was added to the NG medium. Images from 2 independent experiments were quantified using ImageJ and normalized to the average at 12 hours. Statistical analysis was performed by twotailed, unpaired student's t-test (ns, not significant. *, P<0.05. **, P<0.001). Error bars indicate standard deviation of the mean. clearance of ROS. N2 wild-type C. elegans were synchronized at L1 larvae stage and raised on NG medium supplemented with or without 10 µM L-carnitine to L4/young adult stage. The ROS generator paraquat was added to the medium to the final concentration of 1mM. After 24 hours, animals were transferred to new paraquat-free plate with and without L-carnitine for recovery. After recovery for 12, 24, 48 hours, worms were stained with ROS dye dihydroethidium (DHE) and imaged with fluorescent microscope. Representative images at 48-hour recovery were shown. (B) Quantification of DHE signal from at least 3 independent experiments in (A) by ImageJ and relative expression levels were plotted. Statistical analysis was performed by two-tailed, unpaired student's t-test (ns, not significant. *, P<0.05. ***, P<0.001). Error bars indicate standard deviation of the mean. (C) Thrashing assay for C. elegans. Worms were pick from agar plate to 1 mL M9 buffer in 24-well plate and examined under dissecting microscope. Worms were moving left and right rapidly and the movement from one side to the other side then back to the original position was counted as 1 thrash. (D) L-carnitine mitigated the toxicity of paraquat on mobility. Worms were treated with paraquat and L-carnitine as in (A) and recovered for 48 hours. Trashing speed (thrash/min) under different treatments were measured for at least 3-independent experiments with 10 animals/experiment. Statistical analysis was performed by two-tailed, paired student's t-test (**, P<0.01. ***, P<0.001). Error bars indicate standard deviation of the mean. (E) L-carnitine rescued the H2O2 hypersensitivity of paraquat-treated worms. Worms were treated with paraquat and L-carnitine as in (A) and recovered for 48 hours. Worms were then incubated in 50mM H2O2 for 1 hour. Data were pooled from 2 independent experiments and survival rates under different treatment were compared. Statistical analysis was performed by two-tailed, paired student's t-test (*, P<0.05. **, P<0.01). Error bars indicate standard deviation of the mean. (F) L-carnitine mitigated the toxicity of paraquat on embryonic survival. Worms were treated with paraquat and L-carnitine as in (A) and recovered for 48 hours.10 worms were transferred to a new agar plate to allow egg laying for 2 hours and total number of eggs were counted. Hatching were AGING examined after 24 hours. Experiments were conducted 2 times with 5 replicates each time. Data were normalized to control group for comparison. Statistical analysis was performed by two-tailed, unpaired student's t-test (**, P<0.01. ***, P<0.001). Error bars indicate standard deviation of the mean. (G) L-carnitine mitigated paraquat-induced amyloid protein aggregation. Worms expressing human amyloid protein fragment Aβ  in body wall muscle (CL2006) were treated with paraquat and L-carnitine as in (A) and recovered for 48 hours. Worms were then homogenized in high-salt RAB buffer. Soluble and insoluble fraction were analyzed by western blot using anti-Aβ antibody. Total lysate was analyzed by western using anti-actin antibody. Shown was representative results of 2 independent experiments. (H) L-carnitine mitigated Aβ(1-42)-induced paralysis. Animals were raised at 25° C to young adult stage (day-0) and examined every day thereafter for paralysis. Data pooled from 2 independent experiments (n>120) were plotted. Log-rank test was performed (P<0.01). Error bar indicated the standard deviation of the mean.

AGING
were sensitive to H2O2 toxicity, which was largely rescued by L-carnitine ( Figure 2E). Third, we examined the survival rate of progenies. Consistently, egg hatching was reduced by paraquat, but such reduction was significantly mitigated by L-carnitine ( Figure 2F). Fourth, we examined if human amyloid protein aggregation could be mitigated by supplementing Lcarnitine in the medium. Amyloid-beta (Aβ1-42) has been known to aggregate upon oxidative stress or during aging [48,49]. By separating the soluble and insoluble fractions of the whole worm lysate and western blotting, we found that L-carnitine reduced the paraquat-induced Aβ1-42 aggregation ( Figure 2G). Interestingly, despite the lack of effect of L-carnitine on Aβ1-42 aggregates under normal culturing conditions ( Figure 2G), the animals treated with L-carnitine slightly but significantly decreased amyloid-induced paralysis ( Figure 2H). Together, these functional assays confirmed that Lcarnitine reduced oxidative stress damage in C. elegans.

L-carnitine decreases oxidative stress during aging and increased lifespan in C. elegans
Since ROS accumulates during aging, L-carnitine could also improve age-related oxidative damage. To test this, C. elegans were cultured from L1 stage on NG medium supplemented with or without 100 µM L-carnitine and gst-4::gfp expression was examined at L4, day-2, day-4 and day-6 of adulthoods. Interestingly, gst-4::gfp was induced after reaching adulthood but gradually declined thereafter, consistent with a previous report [35]. Similar to that of paraquat treatment, aging-induced gst-4::gfp expression was decreased by L-carnitine during the recovery stage but not the induction stage ( Figure 3A, 3B). Despite the decline in gst-4::gfp expression after day-4, DHE-stained ROS continued to accumulate from day-2 to day-10. Importantly, worms raised on Lcarnitine appeared to have less ROS on day-6 and day-10 of adulthood, suggesting a better recovery from oxidative stress during aging ( Figure 3C, 3D).
We also tested if aging was delayed in C. elegans by supplementing L-carnitine in the medium. Animals were cultured under 20° C on NG medium containing various concentrations of L-carnitine throughout life. 50 µM 5'-Fluoro-2-deoxyuridine (FUDR) was added to inhibit progeny growth. The results showed that L-carnitine extended lifespan of C. elegans at 100, 200 and 500 but not 50 µM ( Figure 3E and Supplementary Figure 1A). At 200 and 500 µM, L-carnitine causes slight developmental delay in some worms (Supplementary Figure 1B). We therefore chose 100 µM for all experiments in this study. SKN-1 and DAF-16 are transcription factors known to activate oxidative stress defensing programs and extend lifespan in C. elegans [50]. Both RNAi knockdown prevented L-carnitine from extending lifespan, suggesting that SKN-1 and DAF-16 are both needed for L-carnitine to extend lifespan in C. elegans ( Figure 3F, 3G).

The long-lived mutants daf-2 and glp-1 recover better from oxidative stress than wild-type controls
The requirement for DAF-16 and SKN-1 for lifespan extension by L-carnitine prompted us to test the role of L-carnitine in long-lived mutants glp-1 and daf-2. These animals, similar to carnitine-treated animals, require both DAF-16 and SKN-1 for their long lifespan [36,51]. First, we asked if glp-1(e2144) and daf-2(e2370) mutants would recover from oxidative stress better than wild-type controls. As reported earlier [35], gst-4::gfp expression was much higher in glp-1 mutant reaching adulthood ( Figure 4A). To evaluate the recovery, we normalized gst-4::gfp expression levels at day-2 adulthood and examined the speed of decline in GFP intensity. Indeed, glp-1(e2144) showed a faster decline in gst-4::gfp expression than controls ( Figure 4A, 4B). Similar results were obtained for daf-2(e2370) mutants ( Figure 4B). We then asked if such improved recovery from OSR could lead to enhanced clearance of endogenous ROS levels. Indeed, the experiments showed that both glp-1(e2144) and daf-2(e2370) mutant worms accumulated ROS slower than wild-type control ( Figure 4C), suggesting that the speed to recover from OSR plays an important role in aging.
Second, we tested if L-carnitine could further increase the lifespan of glp-1 and daf-2 mutants. In several biological repeats, we found that L-carnitine could no longer increase lifespan of glp-1(e2144) and daf-2(e2370) mutant ( Figure 4D, 4E). These results suggest that the activation of DAF-16 and SKN-1 in glp-1 and Wild-type C. elegans were synchronized at L1 larvae stage and raised on NG medium supplemented with or without 10 µM L-carnitine. 50 µM FUDR was added to prevent reproduction. Dead and viable worms were counted every 2 or 3 days starting from day-10 of adulthood. Experiments were performed for 2 times (n>120) and survival were analyzed by log-rank test (Supplementary Table 1). (F) L-carnitine did not extend lifespan of C. elegans with skn-1 knockdown. Experiments were performed similar to (E) except RNAi bacteria was used. Data were pooled from 2 experiments (n>120) and analyzed by log-rank test (Supplementary Table 1). (G) L-carnitine did not extend lifespan of C. elegans with daf-16 knockdown. Experiments were performed similar to (E) except RNAi bacteria was used. Data were pooled from 2 experiments (n>120) and analyzed with log-rank test (Supplementary Table 1).
AGING daf-2 mutants dominate the effect of L-carnitine, making these mutants no longer sensitive to L-carnitine.

A potential carnitine transporter T08B1.1 promotes oxidative stress recovery and increases lifespan in C. elegans
Next, we investigated the mechanisms by taking advantage of the long-lived glp-1 and daf-2 mutants. We focused on genes that are upregulated by both mutants.
One such genes, T08B1.1, encoding a potential cation transporter, showed similarity to the carnitine transporter gene OCTN1 in humans ( Figure 5A). Multiple sequence alignment of T08B1.1 with OCTN1 proteins in human, chicken and Drosophila showed 3 conserved protein domains (Supplementary Figure 2), despite moderate similarity at the primary sequence. We confirmed that T08B1.1 mRNA was upregulated in glp-1 and daf-2 mutants by using qPCR ( Figure 5B). We tested if T08B1.1 protein could mediate carnitine transport by RNAi knockdown. Animals were cultured in the presence of 10 µM L-carnitine with or without T08B1.1 RNAi bacteria. By using the L-carnitine Assay Kit from Abcam, we showed that glp-1 and daf-2 mutants have higher carnitine content than WT controls after feeding L-carnitine for 48 hours from L1 stage ( Figure 5C). Lcarnitine levels were all reduced after T08B1.1 knockdown, suggesting T08B1.1 is a potential carnitine transporter.
We continued to ask if T08B1.1 would be required for recovery from oxidative stress during normal aging. To this aim, we fed age-matched L1 C. elegans expressing gst-4:gfp with T08B1.1 RNAi bacteria on NGM plate in the absence or presence of 10 µM L-carnitine and checked GFP levels on day-4 of adulthood. Consistent with previous results, L-carnitine significantly reduced Error bars indicate standard deviation of the mean. (C) daf-2 worms recovered from oxidative stress faster than wild-type controls. Agematched worms expressing gst-4::gfp were raised at 20° C. gst-4::gfp expression was examined at indicated time points. Two independent experiments were performed. Data were normalized to the average of day-2 adulthood and analyzed by two-tailed, paired student's t-test (*, P<0.05. ***, P<0.001). Error bars indicate standard deviation of the mean. (D) L-carnitine did not further increase lifespan of daf-2 C. elegans. Age-matched wild-type or daf-2 worms were raised at 20° C throughout life with and without 10 µM L-carnitine supplement. Dead and live worms were counted every 2 or 3 days starting from day-10 of adulthood. Data from 2 experiments were pooled and analyzed by log-rank test (Supplementary Table 2). (E) L-carnitine did not further increase lifespan of glp-1 worms. Age-matched wild-type or glp-1 worms were raised at 25° C from L1 to L4 stage and then maintained at 20° C throughout life. 10 µM L-carnitine supplement was added starting from L1 stage. Dead and live worms were counted every 2 or 3 days starting from day-10 of adulthood. Data from 2 experiments were pooled and analyzed by log-rank test (Supplementary Table 2).

DISCUSSION
In this study, we found that L-carnitine could facilitate the recovery from oxidative stress, but did not affect the induction of OSR. L-carnitine delayed ROS accumulation during aging, reduced amyloid protein aggregation and prolonged lifespan. We also found that T08B1.1 gene, encoding a potential carnitine transporter, is required for efficient carnitine uptake and lifespan extension in worms treated with L-carnitine and germline mutant (glp-1) mutant. Together, by using the genetically tractable model, we have gained significant insights into the functions of L-carnitine on aging.
The effect of L-carnitine on oxidative stress recovery is unprecedented. L-carnitine is derived from lysine and methionine and such special structure could directly scavenge free radicals [1,2]. Previous studies have shown Age-synchronized L1 worms were raised on NG medium supplemented with10 µM L-carnitine until day-1 adulthood. Worms were homogenized and relative L-carnitine content measured with L-carnitine Assay kit. Data from 2 experiments were analyzed with two-tailed, paired student's t-test (ns, not significant, *P<0.05, **P<0.01). Error bars indicate standard deviation of the mean. (D) RNAi knocking down of T08B1.1 prevented L-carnitine from decreasing oxidative marker gst-4::gfp. Age-synchronized worms expressing gst-4::gfp were raised on NG medium supplemented with10 µM L-carnitine until day-4 adulthood. Worms were imaged for GFP levels. Shown are representative images from day-4 adulthood. (E) Quantification of day-2 and day-4 data from experiments shown in (D). Ten images from 2 experiments were quantified by ImageJ and analyzed with two-tailed, unpaired student's t-test (***, P<0.001). Error bars indicate standard deviation of the mean. (F) RNAi knocking down of T08B1.1 decreased glp-1 lifespan but not that of wild-type or daf-2 mutant. glp-1 worms were raised at 25° C from L1 to L4 stage and then changed to 20° C. WT and daf-2 worms were kept at 20° C. Dead and live worms were counted every 2 or 3 days starting from day-10 of adulthood. Data from 2 experiments (n>120 for each curve) were analyzed by log-rank test (Supplementary Table 3). (G) RNAi knocking down of T08B1.1 prevented L-carnitine from extending lifespan. Age-matched N2 wild-type worms were kept at 20° C throughout life. Dead and live worms were counted every 2 or 3 days starting from day-10 of adulthood. Data from 2 experiments (n>120 for each curve) were analyzed by log-rank test (Supplementary Table 4).
AGING that L-carnitine functions to reduce oxidative stress in mice, rat and humans [21,52]. Among many other proposed mechanisms such as PI3K, AKT, ERK pathways [53,54], the antioxidant effect of L-carnitine has been attributed to its role in activating Nrf2dependent transcription [55][56][57]. However, we show that carnitine does not activate Nrf2 homolog SKN-1 in C. elegans. Rather, L-carnitine serves to facilitate the recovery from oxidative stress by mitigating SKN-1mediated stress response. Therefore, L-carnitine likely functions through a feedback mechanism to control SKN-1 and oxidative stress. These results contrast with previous thoughts, opening new windows for further investigations into the underlying mechanisms. In addition, there are many antioxidative proteins other than SKN-1 and DAF-16 that could play more important roles in mediating L-carnitine's effect on lifespan, which could be readily revealed by a genome-wide RNAi screen for suppressors of either markers (for example gst-4::gfp) or lifespan.
Our study has shown a direct role of L-carnitine in lifespan extension. L-carnitine has been shown to improve multiple aging phenotypes, including high blood lipid, high blood glucose, neurodegeneration and weak cardiac function [21,[58][59][60]. But its role in lifespan remains poorly studied. Only one study has been reported showing that acetyl-L-carnitine can extend the chronological lifespan in the budding yeast [27,28]. By taking advantage of the genetically tractable model organism C. elegans, we show that L-carnitine was able to extend lifespan in worms, which was mediated through the potential carnitine transporter T08B1.1. Confirming the role of L-carnitine in lifespan regulation, we show that T08B1.1 is elevated in long-lived daf-2 and glp-1 mutants and required for glp-1 mutant to live long. T08B1.1 knockdown does not decrease lifespan of daf-2 mutant, likely due to the fact that the daf-2(1370) mutant relies less on SKN-1-mediated oxidative stress response as shown before [36]. Alternatively, the constitutive activation of SKN-1 in daf-2 mutant could dominate the L-carnitine signaling, making the mutant insensitive to T08B1.1 RNAi knockdown. Our study has established C. elegans as a model for mechanistic study of L-carnitine's effect on delaying aging, which could boost the discovery of new strategies against aging and related disease.

Drug treatment
L-carnitine was prepared in 1 mM stock solution and added to the NGM plate to the final concentration of 100 µM at least 12 hours before use. 10 day-1 adult worms were transferred to the plate for egg laying for 2 hours. Age-matched eggs were then allowed to hatch. For paraquat treatment, 10 mM paraquat stock solution was added to the NGM plate to the final concentration of 1 mM at least 12 hours before use. Worms at L4/young adult stage were transferred to paraquat plates for 24 hours and transferred back to paraquat-free NGM plate for recovery. Juglone was added similarly as paraquat except that 300 µM final concentration was used.

ROS detection
To detect ROS levels by using microscope, animals were stained with fluorescent dye Dihydroethidium (DHE). Specifically, worms raised on agar plates were washed with M9 buffer and then incubated in M9 buffer containing 3 μM DHE for 30 min. Worms were washed extensively in M9 buffer before the ROS signal was imaged with fluorescence microscope. Fluorescence intensity was quantified by using ImageJ software.

Western blot of protein aggregates
Equal number of Aβ-expressing worms (CL2006) at L1 stage were raised to young adult stage and treated as indicated. Worms were washed from plates with ice-cold M9 buffer to remove bacteria. Worms were then sonicated in 1X RIPA Buffer (20 mM Tris-HCl pH 7.5. 150 mM NaCl, 1 mM Na2EDTA. 1 mM EGTA. 1% NP-40) then centrifuged at 18,400Xg for 20 min at 4° C. The supernatant is soluble fraction and the pellet is insoluble fraction. The insoluble pellet was washed 3 times with RIPA buffer then resuspended in 75 ml denaturing buffer (8 M Urea, 2% SDS, 50 mM DTT, 50 mM Tris pH 8). Protein samples from soluble and insoluble factions were boiled in SDS loading buffer and separated by SDS-PAGE and total protein were transferred to PVDF AGING membrane for western blot analysis with anti-Aβ antibody (Abcam, ab10148) and β-actin (Abcam, ab14128).

Aβ-induced paralysis assay
CL2006 worms expressing human Aβ  in body-wall muscles were allowed to lay eggs to new plates for 2 hours. Age-matched progenies were allowed to grow on NGM medium plates at 25° C. 50 µM FUDR (5-Fluoro-2′deoxyuridine) was added at L4/young adult stage to inhibit reproduction. Worms were gradually paralyzed during adulthood. Worms that failed to move forward or backward when touched by a platinum wire were scored as "paralyzed". Worms were scored for paralysis at the indicated time points starting from day-1 of adulthood.

RNAi knockdown
RNAi knockdown was carried out by feeding worms with bacteria expressing double strain RNA (dsRNA) on NGM agar plate. All RNAi bacteria were from the Ahringer RNAi library [62]. To prepare the RNAi agar plate, RNAi bacteria were cultured at 37° C overnight and sub-cultured to early log phase at OD600=0.4. 100 ul of bacteria were added directly on NGM plate agar plate containing 50ug/ml Ampicillin and 1mM Isopropyl β-D-1-thiogalactopyranoside (IPTG) to induce dsRNA expression. For RNAi knockdown, 10 day-1 adult worms were transferred to the RNAi plate for egg laying for 2 hours. Age-matched eggs were then allowed to hatch and grow on RNAi bacteria plate. 50 µM FUDR (5-Fluoro-2′-deoxyuridine) was added at L4/young adult stage to inhibit reproduction.

Lifespan assay
For lifespan assay in C. elegans, age-matched L1 worms were raised on either OP-50 or RNAi bacteria plate until L4/young adult stage, then 50 µM FUDR (5-Fluoro-2′deoxyuridine) was added to inhibit reproduction. Worms may be transferred to new plates as required. The number of live and dead worms was recorded every 2 or 3 days starting from day 10. Exploded and bagged worms and worms with protruding vulva were censored. Death was defined as lack of any visible movement for 5 seconds after touching the tail and head with a platinum wire.

Carnitine content measurement
Worms were grown on NGM medium plate containing 100 µM L-carnitine from hatching. On the day of experiment, worms were washed extensively with M9 buffer to remove bacteria and L-carnitine. Worms were then homogenized by sonication. Worm lysate were centrifuged at 18,400Xg for 20 min at 4° C. L-carnitine levels in the lysate were detected with L-carnitine Assay Kit (Abcam) by following manufacturer's protocol. Protein concentration was measured by using BCA protein Assay Kit (Abcam) and served as input control.

Supplementary Figures
Supplementary Figure 1. (A) L-carnitine extends lifespan at various concentrations. Wild-type C. elegans were synchronized at L1 larvae stage and raised on NG medium supplemented with or without 10 µM L-carnitine. 50 µM FUDR was added to prevent reproduction. Dead and viable worms were counted every 2 or 3 days starting from day-10 of adulthood. Data were analyzed by log-rank test (Supplementary Table 5). (B) 200 and 500 µM L-carnitine cause developmental delay in some worms. Synchronized L1 worms in (A) were checked for development stage at L3, L4 and adult. Percentage of L3 and L4 in the whole population were shown in blue and orange, respectively. Figure 5G.