An antisteatosis response regulated by oleic acid through lipid droplet-mediated ERAD enhancement

Abstract

Although excessive lipid accumulation is a hallmark of obesity-related pathologies, some lipids are beneficial. Oleic acid (OA), the most abundant monounsaturated fatty acid (FA), promotes health and longevity. Here, we show that OA benefits Caenorhabditis elegans by activating the endoplasmic reticulum (ER)-resident transcription factor SKN-1A (Nrf1/NFE2L1) in a lipid homeostasis response. SKN-1A/Nrf1 is cleared from the ER by the ER-associated degradation (ERAD) machinery and stabilized when proteasome activity is low and canonically maintains proteasome homeostasis. Unexpectedly, OA increases nuclear SKN-1A levels independently of proteasome activity, through lipid droplet-dependent enhancement of ERAD. In turn, SKN-1A reduces steatosis by reshaping the lipid metabolism transcriptome and mediates longevity from OA provided through endogenous accumulation, reduced H3K4 trimethylation, or dietary supplementation. Our findings reveal an unexpected mechanism of FA signal transduction, as well as a lipid homeostasis pathway that provides strategies for opposing steatosis and aging, and may mediate some benefits of the OA-rich Mediterranean diet.

Fig. 1.. OA-dependent SKN-1A activation in GSC(−) C. elegans.

(A) Canonical proteasome recovery pathway mediated by SKN-1A/Nrf1 (18, 20). HRD-1 and the CDC-48/p97 adenosine triphosphatase (ATPase) participate in the ERAD mechanism, which translocates SKN-1A/Nrf1 from the ER into the cytosol (18, 20, 21). (B) A genome-scale RNA interference (RNAi) screen (see table S1 and data S1 for details) identified ERAD and lipid-related genes as regulators of SKN-1 transcriptional activity in GSC(−) animals. We blocked GSC formation genetically, using a temperature-sensitive (ts) mutation in the glp-1 (Notch) gene [glp-1(ts) in the figures; see Materials and Methods] (12). Suppressor or enhancer refers to genes for which knockdown decreased or increased reporter expression, respectively. dsRNA, double-stranded RNA. (C) SKN-1A nuclear accumulation (arrowheads) is increased in GSC(−) animals. Scale bar, 20 μm. Animals were classified (left) and quantified (right). See data S2 for details. (D) SKN-1A nuclear accumulation in GSC(−) animals is dependent on the ERAD retrotranslocon HRD-1 (see data S2 for details). (E) Contribution of SKN-1A to gst-4 expression in GSC(−) animals (see data S2 for details). Scale bar, 200 μm. (F) Abbreviated FA desaturation pathway (Δ-9 SCD, stearoyl–coenzyme A 9 desaturases; Δ-12, delta-12-FA desaturase) (30). (G) OA dependence of SKN-1A activation in GSC(−) animals. Unless otherwise indicated, all analyses were performed at day 1 of adulthood. Numbers above the bars denote the sample size (biological replicates). **P < 0.01 and ***P < 0.001. N.S., not significant (P > 0.05).

Fig. 2.. OA-dependent SKN-1A activity promotes longevity and enhances proteostasis.

(A) Life-span extension from GSC ablation requires SKN-1A (see table S2 for replicates and statistics). (B) Life-span extension from knockdown of mTORC1 pathway components requires SKN-1A. (C and D) Life-span extension from daf-2 knockdown (C) or DR (by food dilution) (D) partially requires SKN-1A. The life span of GSC(−) animals (A) was analyzed at 25°C, while other experiments (B to D) were run at 20°C. IGF, insulin-like growth factor. (E) Gene categories modulated by SKN-1A specifically in GSC(−) animals. RNA sequencing (RNA-seq) was analyzed with the gene annotation tool WormCat (77). We analyzed (E) and fig. S2B separately because the germ line accounts for about two-thirds of all adult nuclei (12), complicating direct comparisons between GSC(+) and GSC(−) transcriptomes (12). (F and G) SKN-1A positively regulates the proteasome subunit gene rpn-6 (F) and many ERAD genes (G), as measured by quantitative reverse transcription polymerase chain reaction (qRT-PCR) and RNA-seq, respectively (shown as means ± SD; N = 3; t test). (H) SKN-1A increases clearance of ubiquitinated proteins largely independently of other SKN-1 isoforms (quantification at the bottom reflect the average of two biological replicates of 3000 animals each). Unless otherwise indicated, survival experiments were initiated at the L4 stage (time 0). GAPDH, glyceraldehyde-3-phosphate dehydrogenase. *P < 0.05, **P < 0.01, and ***P < 0.001.

Fig. 3.. SKN-1A reduces fat accumulation in LDs.

(A) Regulation of lipid metabolism genes in GSC(−) animals by SKN-1A (RNA-seq data are shown; means ± SD; n = 3; t test). (B) SKN-1A reduces TAG accumulation, visualized by fixed ORO staining of 1-day-old animals. Scale bars, 20 μm. (C) SKN-1A mediates skn-1 effects on TAG levels [quantification of fig. S3D; one-way analysis of variance (ANOVA), Tukey post hoc]. (D to F) SKN-1A reduces LD number (E) and size (F), visualized by the LD marker DHS-3::GFP (79). Scale bar, 50 μm. *P < 0.05, **P < 0.01, and ***P < 0.001. N.S., P > 0.05.

Fig. 4.. SKN-1A activation depends upon phospholipid membrane homeostasis.

(A) Diagram of ER membrane lipid homeostasis and LD formation mechanisms, highlighting those focused on here: (i) PC biosynthesis enzymes [see (B)]; (ii) phospholipid flipping by TAT-4, ortholog of the transbilayer amphipath transporter (TAT), a P4 ATPase transmembrane flippase (34, 37); and (iii) the ER transmembrane LD biogenesis factor FITM-2 (7, 41) along with the LD membrane–bound protein PLIN-1 (30, 79) (see “OA-induced SKN-1A activation requires membrane homeostasis maintenance and LD functions” and table S3 for more details). (B) In C. elegans, PC biosynthesis requires the PMT-1/2 methyltransferases and methyl groups generated by the S-adenosyl methionine synthase (SAMS-1) (30). (C) SKN-1A activation in GSC(−) animals requires PC synthesis. Choline supplementation restores SKN-1A nuclear localization. sams-1(1:5) refers to a 1-in-5 dilution with EV used to avoid developmental delay (see Materials and Methods for details). (D) SKN-1A activation in GSC(−) animals depends upon TAT-4 and its co-chaperone CHAT-1 (chaperonin of TAT-1; orthologous to transmembrane protein 30) (see data S2 for details) (34, 37). Numbers above the bars denote the sample size (biological replicates). ***P < 0.001.

Fig. 5.. Dependence of OA-mediated SKN-1A activity on LDs.

(A) Suppressors of SKN-1A target (gst-4) activation in GSC(−) animals identified in a screen of LD-related genes (see table S2 and data S1 for complete list and details). (B) Overlap between gst-4 suppressors and an LD proteome from human hepatocarcinoma cells (see data S3 for details) (81). (C) Genes required for SKN-1A activation alter LD morphology, as visualized by DHS-3::GFP. Scale bar, 50 μm. (D) Knockdown of the LD genes fitm-2 and plin-1 blocks SKN-1A nuclear accumulation in GSC(−) animals (see data S2 for details). (E and F) Life-span extension from GSC ablation requires the LD biogenesis/function proteins FITM-2 and PLIN-1 (see table S2 for details). (G) SKN-1A activation requires fitm-2 and plin-1 independently of the CDP-DAG PC production pathway, revealed by choline rescue. Numbers above the graph bars denote the sample size (biological replicates). ***P < 0.001. N.S., P > 0.05.

Fig. 6.. H3K4me3 deficiency and OA extend life span by enhancing SKN-1A activity through LDs and ERAD.

(A) Knockdown of H3K4me3 methyltransferases activates SKN-1A (see data S2 for details). (B) ash-2 knockdown fails to extend life span in skn-1a–null animals (see table S2 for details). (C) SKN-1A is activated by OA feeding dependent upon ERAD (hrd-1), phospholipid flipping (tat-4), PC synthesis (pmt-1), and LD function (plin-1) (see data S2 for details). (D) OA-rich oils, but not SFA-rich coconut oil (see fig. S6E for FA composition), activate SKN-1A in C. elegans (see data S2 for details). (E) Dietary oils increase gst-4 expression in a PC-dependent manner (see data S2 for details). Scale bar, 200 μm. (F) Life-span extension by OA requires SKN-1A (see table S2 for details). (G) Life-span extension by OA is blocked by knockdown of the PC-synthesizing enzyme pmt-1 (see table S2 for details). (H and I) The enhanced longevity of OA-supplemented animals is blocked by impairing LD formation or function (see table S2 for details). Numbers above the bars denote the sample size (biological replicates). *P < 0.05, **P < 0.01, and ***P < 0.001. N.S., P > 0.05.

Fig. 7.. OA and LD homeostasis regulate ERAD efficiency.

(A and B) Impairment of ERAD by interference with PC biosynthesis (A), tat-4, fitm-2, or plin-1 (B); quantification reflects the average of three biological replicates (3000 animals each). (C) Knockdown of the H3K4me3 COMPASS component ash-2 enhances ERAD function (quantification at the bottom reflects the average of three biological replicates of 500 animals each). (D) OA supplementation enhances ERAD function (quantification at the bottom reflects the average of three biological replicates of 3000 animals each). (E) OA-induced enhancement of ERAD (indicated by reduced levels of the mutated CPL-1 substrate) is prevented by interference with the ERAD translocation channel (hrd-1), phospholipid flipping (tat-4), or LDs (fitm-2). ERAD is impaired by interference with OA (fat-6/7) or PC synthesis (pmt-1), with robust rescue by OA feeding seen only with fat-6/7 knockdown (data are presented as means ± SEM; n = 4 of 3000 animals each; one-way ANOVA).*P < 0.05, **P < 0.01, and ***P < 0.001

Fig. 8.. The SKN-1A/Nrf1 lipid homeostasis response.

Model and summary diagram (inset) of the findings shown here. OA activates SKN-1A through a previously unknown mechanism that we term the SKN-1A/Nrf1 lipid homeostasis response. In this pathway, OA is sensed through LD formation and effects on ER membrane homeostasis. These changes enhance ERAD efficiency, thereby increasing levels of SKN-1A, which, in turn, remodels lipid metabolism gene expression and reduces fat storage in LDs. OA acts through this pathway to extend C. elegans life span.