Fig 1.
Disrupting Glycolysis-promoting Genes Extends C. elegans Healthspan, Whereas High Glucose and Gluconeogenic Gene Disruptions Shorten Healthspan.
(A) Left-hand graph: swimming rates for WT animals raised on control media or media containing 4% glucose. We recorded the number of head bends (i.e. head “thrashes”) / 30 seconds for individual animals placed in liquid media on day 8 of life. 4% glucose significantly decreases WT locomotory ability (*** P = 0.0005, unpaired t test; averages from three trials, n = 30 animals per condition; error bars represent SEM). Trials with 5 and 13 day old animals also showed impaired older age locomotion in 4% glucose (Supplemental S1A and S1B Fig). Right-hand graph: We recorded pharyngeal pumping of individual wild-type 5 day old animals raised from the egg stage on control media or media containing 4% glucose; 30 second records. 4% glucose significantly impairs pumping rate (*** P = 0.0008, unpaired t test, n = 35 animals per condition, error bars represent SEM; averages from three trials per condition). (B) C. elegans glycolytic and gluconeogenic pathways. Although most enzymes of glycolysis catalyze both forward (green, glycolysis) and reverse (red, gluconeogenesis) reactions, PFK/FBP and PYK/PCK are enzyme pairs that catalyze one-way, opposing reactions: glycolytic enzymes PFK (phosphofructokinase) and PYK (pyruvate kinase) and gluconeogenic enzymes FBP (fructose-1,6-bisphosphatase) and PCK (phosphoenolpyruvate carboxykinase). Glu-6-P = glucose-6-phosphate; Fru-6-P = fructose-6-phosphate; Fru-1,6-BP = fructose-1,6-bisphosphate; DHAP = dihydroxyacetone phosphate; G3P = glyceraldehyde-3-phosphate; PEP = phosphoenolpyruvate. * Gene disruptions studied most extensively in this work (C. elegans ORF designations indicated). We note that, in our hands, disruption of C50F4.2 (pfk-1.2), F25H5.3 (pyk-1), and W05G11.6 (pck-1) did not impact lifespan. Throughout the paper, data on genes that normally promote glycolysis are represented in green; data on genes that normally promote gluconeogenesis are presented in red. (C) Swimming rate profiles for WT animals treated with glycolysis-limiting pfk(RNAi) and pyk-2(RNAi) or gluconeogenesis-limiting fbp-1(RNAi) and pck-2(RNAi). age-1(RNAi) is used as a positive control known to extend old age locomotory ability. Animals treated with glycolysis-limiting pfk(RNAi) and pyk-2(RNAi) swim significantly better than controls on days 10 and 15 of life; (pfk(RNAi) confers a 42% increase in swimming ability on day 10 and a 59% increase on day 15; pyk-2(RNAi) confers a 63% increase on day 10 and an 83% increase on day 15). It may be worth noting that these increases we recorded in this experiment are higher than those seen for the age-1 positive control RNAi (for age-1(RNAi), a 26% and a 46% increase in swimming ability on days 10 and 15, respectively, in this study) so locomotory impact appears relatively substantial, assuming age-1(RNAi) efficacy. On day 15, fbp-1(RNAi) decreases swimming rates by 44%, and pck-2(RNAi) animals moved too slowly to measure (or were dead) (marked as N/A). Data shown are pooled from two similar independent trials, n = 40 animals per condition per trial. Error bars represent SEM. *** P < 0.001; **** P < 0.0001; n.s. = not significant; all statistical analyses by one way ANOVA with Dunnett’s multiple comparisons test; N/A = not available. (D) Survival curves of WT animals treated with pfk(RNAi), pyk-2(RNAi), fbp-1(RNAi), and pck-2(RNAi). Animals treated with RNAi targeting glycolytic genes pfk and pyk-2 show increases in median survival (an 8.85% increase for pfk disruption and a 20.83% increase for pyk-2 disruption) and have significantly increased survival relative to controls (P < 0.0001 for both pfk and pyk-2 disruptions, Log-rank test), although maximal lifespan is not significantly affected (Supplemental S1D Fig). Conversely, animals treated with RNAi directed against gluconeogenic genes fbp-1 and pck-2 have lowered median survival (a 10.94% decrease for fbp-1 disruptions and a 21.88% decrease for pck-2 disruptions) and exhibit lower survival relative to controls (P < 0.0001 for both fbp-1 and pck-2 disruptions, Log-rank). Animals treated with the age-1(RNAi) positive control show a 46.88% increase in median survival and significantly increased survival (P < 0.0001, Log-rank; refer to S1 Table for survival data with age-1(RNAi)). Data shown are pooled from 6 independent trials, n = 60 animals per condition per trial.
Fig 2.
DAF-16/FOXO is Required for Locomotory Healthspan and Lifespan Benefits of Genetic Glycolysis Disruption.
(A) Fluorescence intensity levels of wild-type animals expressing a GFP transcriptional reporter for sod-3, which expresses superoxide dismutase and is a direct transcriptional target of DAF-16. We measured in vivo fluorescence using a spectrofluorimeter on day 5 of life. As expected from previous work [74], decreased insulin signaling with age-1(RNAi) results in increased SOD-3::GFP expression. Similarly, knockdown of glycolytic genes via pfk(RNAi) and pyk-2(RNAi) significantly increases SOD-3::GFP fluorescence levels. Disrupting the fbp-1 and pck-2 gluconeogenic genes, on the other hand, does not significantly change GFP fluorescence levels. Data are pooled averages from 5 independent trials, n = 50 animals per condition per trial. Error bars represent SEM. *** P < 0.001; **** P < 0.0001; n.s. = not significant; all by one way ANOVA with Dunnett’s multiple comparisons test. (B) Swimming rates of daf-16(mgDf50) mutants treated with glycolytic and gluconeogenic RNAi. Disrupting glycolysis-promoting genes pfk or pyk-2 in the daf-16 null background eliminates old-age benefits that occur with pfk(RNAi) or pyk-2(RNAi) (refer to Fig 1C), indicating a requirement for daf-16 in the benefits of glycolysis impairment. In contrast, disruption of gluconeogenic genes fbp-1 or pck-2 results in significant decreases in swimming rates in daf-16 animals, similar to wild-type. Data shown are pooled from two independent trials, n = 40 animals per condition per trial. Measurements were taken on day 10 of life. Error bars represent SEM. *** P < 0.001; **** P < 0.0001; n.s. = not significant; all by one way ANOVA with Dunnett’s multiple comparisons test. daf-16 mutants are documented to have lowered locomotory ability vs. WT later in life [91, 92]; we note that, although comparison of scores across experiments is not possible due to variation in baseline values, the swimming rate of daf-16(mgDf50) control animals was slightly lower than WT controls on the same day of life (in Fig 1C; mean of 46.17 head thrashes/30 sec. for daf-16 vs. 52.17 head thrashes/30 sec. for WT on day 10). (C) Survival curves of daf-16(mgDf50) null mutants treated with RNAi against glycolytic genes pfk and pyk-2. The beneficial effects of glycolytic gene disruptions seen in wild-type animals (refer to Fig 1D) are mostly eliminated in the daf-16 mutant background: median survival is not increased in daf-16 animals treated with either RNAi, and the survival curve of daf-16 mutants raised on pyk-2(RNAi) is not significantly (n.s.) different than controls (Log-rank test). pfk(RNAi) slightly increases the survival of daf-16 mutants later in life (as indicated by a right-shifted survival curve in late-life; Log rank calculated P = 0.0128), suggesting possible daf-16-independent mechanisms for lifespan extension in older animals with glycolytic inhibitions. Data are pooled from 5 independent trials, n = 60 animals per condition per trial, details in S1 Table.
Fig 3.
Increased Locomotory Healthspan and Survival with pck-2 Overexpression, as Well as Expression of pck-2 in the Intestine, Require daf-16.
(A) Survival curves of wild-type animals expressing a transcriptional reporter for pck-2 (Ppck-2gfp, used as the control strain) or the intact pck-2 expressed from its own promoter (Ppck-2pck-2::gfp, “pck-2 OE”, used as the pck-2 over-expressor strain) and treated with empty vector control (EV RNAi) or F47B8.10(RNAi). F47B8.10 encodes glucose-6-phosphate translocase (G6PTase), an enzyme that functions in a complex with G6P phosphatase to catalyze the final step of gluconeogenesis in vertebrates [61]. pck-2 overexpression significantly increases survival (P = 0.0020, Log-rank) that depends on G6PTase (survival curves of control animals and pck-2 over-expressors treated with F47B8.10/G6PTase(RNAi) are not significantly different (n.s., Log-rank)). Data shown are from three trials, n = 60 animals per condition. (B) Wild-type animals expressing pck-2::gfp from the native pck-2 promoter (Is[Ppck-2pck-2::gfp]) and treated with an empty vector control RNAi display a GFP signal in cells in the anterior (“A”) and posterior (“P”) intestine (arrows, top left panels; day 7 of life; approximately 35 out of 60 animals displayed this intestinal GFP pattern on day 7) as well as in the pharynx (arrowhead; all animals displayed pharyngeal GFP on day 7). daf-16(RNAi) treatment specifically eliminates the intestinal GFP signal (top right-hand panels; no intestinal GFP was detected in any of the 60 observed animals on day 7). The predicted DAF-16 binding site in the pck-2 promoter is shown on the bottom left in red; green highlighting indicates altered base pairs in the binding site mutant. Disruption of the DAF-16 binding site 1086 base pairs upstream of the pck-2 start codon [63] in the pck-2 promoter in Ex[Ppck-2mutpck-2::gfp] (bottom panels) also eliminates the intestinal pck-2::gfp signal (bottom right; no intestinal GFP was detected in any of the 60 observed animals on day 7). Bar, 200 μm. (C) Swimming rates of wild-type animals expressing either a pck-2 transcriptional reporter that lacks any pck-2 coding sequences (Ppck-2gfp, used as the control strain) or the intact pck-2 gene expressed from its own promoter (Ppck-2pck-2::gfp, here indicated as “pck-2 OE”, used as the pck-2 over-expressor strain) treated with an empty vector control RNAi (EV RNAi) or daf-16(RNAi). pck-2 overexpression (OE) significantly increases swimming rates vs. wild-type controls (P < 0.01, one way ANOVA). Swimming fitness is disrupted with daf-16(RNAi) in both WT controls and pck-2 OE animals (P < 0.01, one way ANOVA). Swimming rates of control and pck-2 OE animals treated with daf-16(RNAi) are not significantly different from one another, one way ANOVA. Data shown are from averages from three trials, n = 40 animals per condition. Measurements were taken on day 9 of life. Error bars represent SEM. ** P < 0.01. All one way ANOVA analyses were performed with Dunnett’s multiple comparisons test. (D) Survival curves of wild-type animals expressing a transcriptional reporter for pck-2 (Ppck-2gfp, used as the control strain) or the intact pck-2 expressed from its own promoter (Ppck-2pck-2::gfp, “pck-2 OE”, used as the pck-2 over-expressor strain) and treated with empty vector control (EV RNAi) or daf-16(RNAi). pck-2 overexpression (OE) significantly increases survival (P = 0.0017, Log-rank test). The pck-2 OE survival benefit is abolished with daf-16(RNAi) (control and pck-2 OE animals treated with daf-16(RNAi) have shortened lifespans that are not significantly (n.s.) different from one another, Log-rank test). Data shown are pooled from 2 independent trials, n = 60 animals per condition per trial. **** P < 0.0001. Details in S1 Table.
Fig 4.
Dietary Restriction Induces Gluconeogenic Gene pck-2 Expression via DAF-16, and pck-2 Expression is Needed for DR-associated Healthspan Benefits.
(A) Representative wild-type animals expressing a GFP transcriptional reporter for pck-2 (pck-2 promoter GFP fusion, lacking pck-2 coding sequences) under ad libitum (AL) or dietary restricted (DR) conditions, day 4 of life. DR animals were raised on diluted (7 x 109 cfu/ml) OP50; AL is undiluted (7 x 1010 cfu/ml) bacteria. A = anterior, P = posterior, intestinal GFP fluorescence indicated by arrows. Quantitation of GFP observations is described in (B). Bar, 200 μm. (B) pck-2 transcriptional reporter GFP expression under three distinct food limitation conditions. Left: as in (A), wild-type animals expressing the Ppck-2gfp reporter were raised on plates seeded with undiluted (7 x 1010 cfu/ml) or diluted (7 x 109 cfu/ml) bacteria and the number of animals displaying GFP fluorescence was scored on day 4 of life. Ampicillin was added to plates to inhibit bacteria growth. The number of animals with GFP fluorescence increases with bacterial dilution (P = 0.009, unpaired t test). Data are pooled averages from 3 independent trials, n = 10 animals per condition per trial. Error bar represents SEM. * P < 0.05. Center: metformin is a drug used to treat type-2 diabetes that induces features of dietary restriction-like metabolism in C. elegans [41]. Animals expressing Ppck-2gfp were raised from the egg stage on plates containing either no metformin or 50 mM metformin, and GFP expression was observed on day 4 of life. The number of animals with GFP expression increases on 50 mM metformin plates (P = 0.04, unpaired t test). Data are pooled averages from 3 independent trials, n = 60 animals per condition per trial. Error bar represents SEM. * P < 0.05. Right: GFP fluorescence in eat-2(ad1116) mutants expressing Ppck-2gfp and treated with either an empty vector control RNAi or daf-16(RNAi). Under DR, daf-16(RNAi) reduces the frequency of observing GFP expression on all tested days (data shown are from day 8 of life, P = 0.05, unpaired t test). Wild-type animals carrying Is[Ppck-2gfp] showed no GFP signal on day 8. Data are pooled averages from 3 independent trials, n = 60 animals per condition per trial. Error bar represents SEM. * P < 0.05. (C) Left-hand graph: qRT/PCR-evaluated pck-2 expression levels from early-to-late life in wild-type (black bars) and dietary-restricted (white bars) eat-2(ad1116) mutants. While pck-2 transcript levels are lower in eat-2 mutants (vs. wild-type, P = 0.007, unpaired t test) early in life, pck-2 expression remains nearly constant over time in the eat-2 background (vs. wild-type at 10 days, P = 0.0008, unpaired t test). Expression level for all results are represented as fold change with respect to the wild type day 3 level. * P < 0.05; ** P < 0.005. P values were calculated based on three technical repeats per strain per day tested, with approximately 10,000 animals per sample. Right-hand graph: qRT/PCR-evaluated G6P translocase (G6PTase) F47B8.10 expression levels in wild-type and dietary-restricted eat-2(ad1116) mutants on day 3 of life. The gluconeogenic-specific enzyme glucose-6-phosphatase is significantly up-regulated in the eat-2 background (P < 0.0005, unpaired t test). P values were calculated based on three technical repeats per strain, with approximately 10,000 animals per sample. (D) Swimming rates of eat-2(ad1116) mutants disrupted in glycolytic or gluconeogenic activity. Disrupting glycolytic activity with pfk(RNAi) results in small but significant increases in swimming rates early in life (day 5; P < 0.01, one way ANOVA, a 12% increase) and no later change; glycolytic gene pyk-2(RNAi) did not significantly affect swimming ability on any tested day. Disrupting gluconeogenic gene pck-2 significantly reduces swimming rates on day 15 (P < 0.0001 by one way ANOVA, a 58% reduction). Data are pooled from 3 independent trials, n = 40 animals per condition per trial. Error bars represent SEM. ** P <0.001; *** P < 0.0001. All one way ANOVA analyses performed with Dunnett’s multiple comparisons test. (E) Survival curves of long-lived eat-2(ad1116) mutants treated with RNAi against pck-2. pck-2(RNAi) decreases median survival by 13.79%, and also results in decreased survival (survival difference P < 0.0001, Log-rank). Data are pooled from 5 independent trials, n = 60 animals per condition per trial. (F) Survival curves of wild-type animals raised under ad libitum (AL) or dietary restriction (DR) conditions, with and without pck-2(RNAi). Median survival of animals expressing the empty vector control was higher under DR conditions as compared to well-fed controls (an 18.62% increase, see S1 Table for details) (survival differences P = 0.0039, Log-rank test). Disrupting gluconeogenic gene pck-2 abolished these beneficial DR effects: animals raised on diluted gluconeogenic RNAi bacteria had survival curves that were not significantly different from those of well-fed animals (n.s., Log-rank). Data are pooled from 2 independent trials, n = 60 animals per condition per trial. ** P < 0.005; n.s. = not significant. (G) Excitation wavelengths corresponding to peak age pigment fluorescence intensities in wild-type and eat-2(ad1116) dietary restricted mutants raised on empty vector control RNAi, and in wild-type animals treated with RNAi against glycolytic gene pfk. Dietary restriction lowers age pigment peak excitation wavelengths, and this shift to lower excitation wavelengths (the ExMax shift) can be used as a biomarker for DR [64]. Like eat-2 mutants, pfk(RNAi) animals show ExMax shifts to lower excitation wavelengths on day 5 of life (the average ExMax value for wild-type animals with pfk disruptions is 334.0 nm vs. 344.0 nm for wild-type controls; ExMax values for eat-2(ad1116) animals is 330.0 nm; P < 0.0001 for eat-2(ad1116) + vector control and P < 0.001 for WT + pfk(RNAi), one way ANOVA). Data are averages from 2 independent trials, n = 30 animals per condition per trial. Error bars represent SEM. ** P < 0.001; *** P < 0.001. All one way ANOVA analyses performed with Dunnett’s multiple comparisons test. (H) Wild-type animals expressing Is[Ppck-2pck-2::gfp] and treated with a vector control or with RNAi against glycolytic gene pfk. On day 7 of life, pfk(RNAi) significantly increases GFP fluorescence in Ppck-2pck-2::gfp-expressing animals, as quantitated in a spectrofluorimeter (P = 0.0383, unpaired t test). Data are averages from 3 independent trials, n = 30 animals per condition per trial. Error bar represents SEM. * P < 0.05. (I) Summary model for interactions of glycolysis, gluconeogenesis, and healthy aging. Under ad libitum (AL) conditions, insulin-like signaling promotes glucose transport and processing via glycolysis. Glycolytic activity, in turn, suppresses the healthspan-promoting transcription factor DAF-16 to limit healthy aging. Under dietary restriction (DR) or RNAi disruption of pfk or pyk-2, glucose levels and glycolytic activity decrease, while DAF-16 is activated to increase transcription of pro-healthspan targets including gluconeogenic gene pck-2. Elevated gluconeogenic activity (dependent on DAF-16-promoted PCK-2 expression) acts in a feed-forward capacity to further enhance the gluoconeogenic pathway and may inhibit glycolysis/insulin signaling to promote/maintain DAF-16 signaling and healthy metabolism (inhibition represented by a dashed line).