Caenorhabditis elegans Lipin 1 moderates the lifespan‐shortening effects of dietary glucose by maintaining ω‐6 polyunsaturated fatty acids

Abstract Excessive glucose causes various diseases and decreases lifespan by altering metabolic processes, but underlying mechanisms remain incompletely understood. Here, we show that Lipin 1/LPIN‐1, a phosphatidic acid phosphatase and a putative transcriptional coregulator, prevents life‐shortening effects of dietary glucose on Caenorhabditis elegans. We found that depletion of lpin‐1 decreased overall lipid levels, despite increasing the expression of genes that promote fat synthesis and desaturation, and downregulation of lipolysis. We then showed that knockdown of lpin‐1 altered the composition of various fatty acids in the opposite direction of dietary glucose. In particular, the levels of two ω‐6 polyunsaturated fatty acids (PUFAs), linoleic acid and arachidonic acid, were increased by knockdown of lpin‐1 but decreased by glucose feeding. Importantly, these ω‐6 PUFAs attenuated the short lifespan of glucose‐fed lpin‐1‐inhibited animals. Thus, the production of ω‐6 PUFAs is crucial for protecting animals from living very short under glucose‐rich conditions.

Fluorescence imaging of worms. Fluorescence imaging was performed as described previously with modifications (Lee et al., 2015). For measuring the levels and the subcellular localization of GFP::LPIN-1, transgenic worms were fed with control or lpin-1 RNAi bacteria, or OP50 with or without additional 2% glucose from hatching. Synchronized young (day 1) adult worms were then anesthetized with 5 mM levamisole on a 2% agarose pad on a slide glass before imaging.
Lifespan assays. Lifespan assays were performed as described previously with slight modifications (Lee et al., 2015). After dsRNA-expressing bacteria were cultured in liquid LB containing 50 μg/mL ampicillin at 37°C overnight, the bacteria were seeded on NGM plates and incubated at 37°C overnight. One mM IPTG was added on the bacteria-seeded NGM plates for the induction of dsRNA. Briefly, worms that were grown on dsRNA-expressing bacteria-seeded NGM plates from eggs were transferred to 10 μM FUDR-containing plates to inhibit progeny hatching with or without additional 2% glucose at young (day 1) adult stage. For double RNAi treatment, two different types of dsRNA-expressing bacteria were separately cultured using liquid LB media containing 50 μg/ml ampicillin (USB, Santa Clara, CA, USA) at 37°C overnight until the optical density (OD) value at 600 nm reached 0.9. The bacteria were then mixed to 1:1 ratio. The mixed bacteria were seeded on NGM plates and incubated at 37°C overnight. Ten mM IPTG was added on the bacteria-seeded NGM plates for the induction of dsRNA. Dead worms that did not show any movement upon gentle touch with a platinum wire were counted. The lifespan assays for each experiment were performed by at least two researchers independently for reproducibility. Worms that crawled off the plates, burrowed, or displayed ruptured vulvae or internal hatching were classified as censored worms but were included in subsequent statistical analysis. Fatty acid-treated plates were prepared as described previously with modifications (Lee et al., 2015;F. Yang et al., 2006). Specifically, 600 μM of linoleic acids (18:2n-6, Sigma, St Louis, MO, USA), arachidonic acids (20:4n-6, Sigma), or oleic acids (18:1n-9, Sigma, St Louis, MO, USA) dissolved in ethanol were mixed with NGM containing 0.1% St Louis, MO, USA) and 50 μg/mL ampicillin for specific fatty acid feeding assays. For saturated fatty acid (SFA) feeding, 600 μM of myristic acid (14:0, Sigma, St Louis, MO, USA) and palmitic acid (16:0, Sigma, St Louis, MO, USA) dissolved in ethanol were mixed with NGM containing 0.1% NP-40 and 50 μg/mL ampicillin. For control plates for fatty acid feeding assays, NGM was mixed with ethanol, 0.1% NP-40 and 50 μg/mL ampicillin. Statistical analysis was performed by using OASIS2 (https://sbi.postech.ac.kr/oasis2) (Han et al., 2016), which calculates p values using log-rank (Mantel-Cox method) test. RNA seq analysis. RNA was extracted as described previously with minor modifications (Lee et al., 2019). Worms were fed with control or lpin-1 RNAi bacteria on control or 2% glucosecontaining NGM plates from hatching. Synchronized worms were harvested when worms reached young (day 1) adult stage. RNA was extracted using RNAiso plus (Takara Bio Inc., Shiga, Japan). TruSeq (unstranded) mRNA libraries (Illumina, CA, USA) were constructed and paired-end sequencing of Illumina platform was performed by Macrogen (Seoul, South Korea).
A genome-wide RNAi screen using fat-5p::fat-5::gfp. A genome-wide RNAi screen using a liquid culture system was performed as described previously with modifications (Lehner, Tischler, & Fraser, 2006). Each of 19,213 RNAi bacteria from commercially available C. elegans RNAi library (Source BioScience, Nottingham, UK) was cultured in 200 μL liquid LB media with 50 μg/mL ampicillin using 96 well plates overnight at 37°C. RNAi bacteria were treated with 4 mM IPTG to induce dsRNA at 37°C for 1 hr, and the bacteria were spun down by centrifugation (870 g for 10 min) and the supernatant was then discarded. Pellet was resuspended by adding liquid nematode growth media (NGM) with 4 mM IPTG and 50 μg/mL ampicillin.
Worms were synchronized by bleaching gravid adult worms cultured on OP50-seeded high growth (HG) plates with bleach (mix 5 N NaOH and household bleach to 1:2 ratio) solution (Stiernagle, 2006), and the eggs were allowed to hatch to become L1 larvae in 15 mL tube with M9 buffer for synchronization at 20°C. The synchronized L1 larvae in 15 μL M9 buffer drops were counted to calculate the number of worms per volume. Approximately 20 L1 worms were then transferred into each well of RNAi bacteria-containing 96 well plates by using a micropipette. After three days of culture, day 1 adult worms in each well were scored for GFP intensity by four researchers, with a semi-quantitative range from -3 (dimmest) to +3 (brightest).
Control RNAi (L4440) and gfp RNAi were used as control scores 0 and -3, respectively. Cut-off scores were arbitrarily set to include 1% of the total RNAi clones that were examined; RNAi clones that displayed GFP intensity scores greater and less than the arbitrary cut-off scores +2.75 and -2.50 respectively were selected as enhancer and suppressor candidate RNAi clones from the primary screen. The liquid-based screen was subsequently repeated six times using a sub-library that included the candidate RNAi clones from the primary screen.
Confirmation of the genome-wide RNAi screen using fat-5p::fat-5::gfp on solid media. For confirmation of the whole-genome RNAi screen using fat-5p::fat-5::gfp on solid NGM plates, 56 RNAi clones that substantially and reproducibly altered the intensity of GFP fluorescence from the liquid-based RNAi screen were chosen; cut-off values from 6 repeats were arbitrarily set to -1.1 and +0.7 for suppressor and enhancer RNAi clones, respectively. RNAi clones that oppositely (≥ +2 for suppressors and ≤ -2 for enhancers) altered GFP intensity at least one of 6 sets were excluded. The 56 RNAi clones were cultured in LB media with 50 μg/mL ampicillin overnight at 37°C. Subsequently, 100 µL of RNAi bacteria was seeded and grown on solid NGM plates containing 50 μg/mL ampicillin overnight at 37°C. One mM IPTG was added on the RNAi bacteria-seeded NGM plates for the induction of dsRNA for 24 hrs at room temperature.
fat-5p::fat-5::gfp transgenic worms at gravid adult stages were transferred onto the dsRNAexpressing bacteria seeded-plates to allow to lay eggs. The worms were anesthetized with 100 mM sodium azide and subsequently used for imaging at young (day 1) adult stage.
Nile red staining. Nile red staining was performed as described previously with modification (Ashrafi et al., 2003;Pino, Webster, Carr, & Soukas, 2013). Nile red powder (Sigma, St Louis, MO, USA) was dissolved in acetone to 0.5 mg/mL (stock solution) and stored at -20°C. Gravid adults were allowed to lay eggs overnight on control or lpin-1 RNAi bacteria-seeded NGM plates containing 50 μg/mL ampicillin with or without 2% glucose. Synchronized worms at young adult stage (day 1) were harvested and washed once using M9 buffer with 0.01% triton X-100 (Daejung, Siheung, South Korea). Worms were fixed using 500 μL of 40% isopropanol for 3 min at room temperature. Worms were collected by spinning down briefly and the supernatant was discarded. Fixed worms were stained using 500 μL of 3 μg/mL Nile red solution (3 μL stock solution to 500 μL of 40% isopropanol) for 2 hrs at room temperature in the dark. By adding 500 μL of M9 buffer with 0.01% Triton X-100, worms were destained for 30 min. GFP emission filter was used to detect Nile red signals. Images of worms that were placed on a 2% agarose pad on a slide glass were captured by using a camera (AxioCam HRc, Zeiss Corporation, Oberkochen, Germany) attached to a Zeiss Axioscope A.1 microscope (Zeiss Corporation, Oberkochen, Germany). For quantification, ImageJ (https://imagej.nih.gov/ij/) (Schneider, Rasband, & Eliceiri, 2012)  and placed on a shaker at room temperature for two days. During the two-day shaking, precipitated aggregates from the 0.3% Oil red O were filtered with 0.45 μm minisart filter two more times. Prior to staining experiments, the 0.3% Oil red O was freshly prepared by filtering with 0.45 μm minisart filter once more. Worms were fed with control RNAi and lpin-1 RNAi bacteria on control or 2% glucose-containing NGM plates with 50 μg/mL ampicillin from hatching. Worms at young (day 1) adult stage were harvested with M9 buffer in microtubes and were fixed with 60% isopropanol for two min. Worms were stained with the 0.3% Oil red O in an airtight plastic box with wet paper at 25℃ over 18 hrs. After staining, worms were washed once and de-stained with M9 buffer containing 0.01% Triton X-100 (Daejung, Siheung, South Korea). The samples were stored at 4°C until microphotographs were captured. The differential interference contrast (DIC) images of worms that were placed on a 2% agarose pad on a slide glass were captured by using a camera (AxioCam HRc, Zeiss Corporation, Oberkochen, Germany) attached to a Zeiss Axioscope A.1 microscope (Zeiss Corporation, Oberkochen, Germany). The Oil red O intensity of worms was quantified by using ImageJ (https://imagej.nih.gov/ij/) (Schneider et al., 2012). The backgrounds of the images were subtracted and the images were converted to 8-bit grayscale images. By using circle tools in ImageJ (https://imagej.nih.gov/ij/) (Schneider et al., 2012), the areas that had intensities over 142 (arbitrary threshold) in first two anterior intestinal cells were measured for detecting the Oil red O signals.
Body size measurement assays. Gravid wild-type worms were allowed to lay eggs on 50 μg/mL ampicillin-containing NGM plates seeded with control or lpin-1 RNAi bacteria under control or 2% glucose-diet conditions. Progeny were anesthetized with 5 mM levamisole at young (day 1) adult stage. Anesthetized worms were placed on a 2% agarose pad and bright field images were captured by using a camera (AxioCam HRc, Zeiss Corporation, Oberkochen, Germany) attached to a Zeiss Axioscope A.1 microscope (Zeiss Corporation, Oberkochen, Germany). The area of worms was quantified by using ImageJ (https://imagej.nih.gov/ij/) (Schneider et al., 2012).
Quantitative RT-PCR. Quantitative RT-PCR was performed as described previously with slight modifications (Lee et al., 2015). Worms were fed with control RNAi or lpin-1 RNAi bacteria on control or 2% glucose-containing diets for entire life. Synchronized worms were harvested at young (day 1) adult stage and washed at least twice with M9 buffer. RNA was extracted using RNAiso plus (Takara, Japan) and phase lock gel (VWR, PA, USA). For reverse transcription, random primers (6 mers, Cosmogenetech, South Korea) were used and reverse transcription was performed using ImProm-II™ Reverse Transcriptase kit (Promega, Madison, WI, USA).
Quantitative real time PCR was performed by using StepOne and StepOnePlus Real-Time PCR systems (Applied Biosystems, Foster City, CA, USA) and using SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA, USA). Relative quantity of the mRNA was calculated by employing comparative Ct methods described in the manufacturer's manual. ama-1, an RNA polymerase II large subunit, tba-1, tubulin α, and pmp-3, ATP binding cassette subfamily, was used as endogenous reference genes for normalization (Hoogewijs, Houthoofd, Matthijssens, Vandesompele, & Vanfleteren, 2008;Zhang, Chen, Smith, Zhang, & Pan, 2012). The levels of mRNA were normalized by multiple controls (ama-1, tba-1 and pmp-3). Sequences of primers that were used in this study are as follows.  Lee et al., 2015). Worms treated with lpin-1 RNAi with or without additional 2% glucose were synchronized by using a bleach method (Stiernagle, 2006). Young (day 1) adult worms were collected using ddH2O. Approximately 400 µL of wet worm pellet was frozen in liquid nitrogen and stored at -80°C until use. The worms were sonicated on ice (30 amplitude, 10 sec on, 2 sec off, 6 cycles) and 10 µL of total worm lysate was used for total protein quantification. Total protein amount was measured by using bicinchoninic acid (BCA) protein assay kit (Thermo Scientific, MA, USA). Three hundred µL of worm lysate was then used for the GC/MS analysis. Five mL ice-cold chloroform/methanol (1:1 ratio) was added to a glass tube. The worm lysates mixed with chloroform/methanol were vortexed and incubated for approximately 2 hrs at room temperature. The samples were then treated with 2 mL Hajra's solution (0.2 M H3PO4, 1 M KCl) and mixed thoroughly by shaking.
The glass tubes were centrifuged (870 g, 10 min) to separate lipid-containing chloroform from aqueous phase liquids. By using glass Pasteur pipettes, lipid-containing chloroform layer was separated and transferred into new glass tubes. Three mL chloroform was added to the remaining aqueous phase solution and centrifuged to further extract residual lipids. The combined lipidcontaining chloroform layers were evaporated to 120 µL, and 20 µL of the solution was subsequently used for analyzing total lipid extract.
Step I. We initially identified 116 enhancer and 118 suppressor RNAi clones that respectively increased and decreased the fluorescence levels of the fat-5p::fat-5::gfp in a liquid culture system. Step II. After performing the fluorescence measurement experiments 6 more times, we narrowed down the candidates to 18 enhancer and 38 suppressor RNAi clones that reproducibly altered the fat-5p::fat-5::gfp levels (Tables S3 and S4).
Step III. Among the 18 enhancer and the 33 suppressor RNAi clones that did not show developmental defects or L1 arrest on solid media, seven and thirteen RNAi clones significantly increased and decreased the expression of fat-5p::fat-5::gfp, respectively (Tables S3   and S4: written in bold). Cut-off scores for each step are described in the Experimental
Supporting Tables   Table S1. Summary of lifespan screen using far-3p::gfp enhancer RNAi clones. control RNAi-treated animals on control diets or on glucose-rich diets.

Gene ID Gene name Description
Lifespan assays were performed by at least two independent researchers.
Glucose specificity (%) was calculated by subtracting lifespan changes (%) on control diets from lifespan changes (%) on glucose diets.
The gene list was sorted by ascending order of glucose specificity (%).
*cyp-42A1 was not initially included in enhancer RNAi clones (Lee et al., 2015) but instead Y80D3A.6 RNAi was included. Because Y80D3A.6 is predicted as a dead gene (based on WormBase (https://wormbase.org/)), we tested which gene was targeted by Y80D3A.6 RNAi by using BLAST and subsequently found that a portion of cyp-42A1 was targeted. Therefore, we substituted Y80D3A.6 with cyp-42A1. Lifespan data from the same experimental sets were indicated by bold lines and biological repeats were separated by bold dashed lines and stating trial numbers.
All p values were calculated by using log-rank test.
Percent change (%) and p values indicate comparison of each condition with control RNAi on control diets within the same experimental dataset.
ctrl RNAi glc : percent change (%) and p values were calculated by comparing with control RNAi on glucose diets within the same experimental dataset.
* indicates experimental set of pod-2 RNAi, which was used as a positive control that specifically decreased lifespan under glucose-rich conditions (Lee et al., 2015).

T08G11.5 unc-29
Neuronal acetylcholine receptor subunit alpha-4 0.7 Enhancer RNAi clones targeting nineteen genes were sorted in descending order by average scores of six replicates from the liquid culture system ( Figure S4A, step II).
Seven enhancer RNAi clones that increased fat-5p::fat-5::gfp expression on solid media were highlighted in bold ( Figure S4A, step III). Suppressor RNAi clones targeting thirty eight genes were listed in ascending order by average scores of 6 replicates from the liquid culture system ( Figure S4A, Step II).
Thirteen suppressor RNAi clones that decreased fat-5p::fat-5::gfp expression on solid media were highlighted in bold ( Figure S4A, Step III). Gray boxes indicate that RNAi bacteria targeting multiple genes, which were confirmed through sequencing.  Lifespan data from the same experimental sets were indicated by bold lines and biological repeats were separated by bold dashed lines and stating trial numbers.
All p values were calculated by using log-rank test.  Lifespan data from the same experimental sets were indicated by bold lines and biological repeats were separated by bold dashed lines and stating trial numbers.
All p values were calculated by using log-rank test.
Percent change (%) and p values were calculated by comparing each condition with control RNAi on control diets within the same experimental dataset.
lpin-1 RNAi glc : percent change (%) and p values were calculated by comparing with N2 or odr-1p::rfp treated with lpin-1 RNAi on glucose diets within the same experimental dataset.