Phosphatidylcholine synthesis for lipid droplet expansion is mediated by localized activation of CTP:phosphocholine cytidylyltransferase

Abstract

Lipid droplets (LDs) are cellular storage organelles for neutral lipids that vary in size and abundance according to cellular needs. Physiological conditions that promote lipid storage rapidly and markedly increase LD volume and surface. How the need for surface phospholipids is sensed and balanced during this process is unknown. Here, we show that phosphatidylcholine (PC) acts as a surfactant to prevent LD coalescence, which otherwise yields large, lipolysis-resistant LDs and triglyceride (TG) accumulation. The need for additional PC to coat the enlarging surface during LD expansion is provided by the Kennedy pathway, which is activated by reversible targeting of the rate-limiting enzyme, CTP:phosphocholine cytidylyltransferase (CCT), to growing LD surfaces. The requirement, targeting, and activation of CCT to growing LDs were similar in cells of Drosophila and mice. Our results reveal a mechanism to maintain PC homeostasis at the expanding LD monolayer through targeted activation of a key PC synthesis enzyme.

Figure 1. CCT Is a Major Phospholipid Synthesis Enzyme Required during LD Expansion

(A) RNAi-mediated depletion of several PC synthesis enzymes reduces PC levels in LDs as analyzed by TLC. (B) RNAi depletion of PE synthetic enzymes does not affect LD morphology in Drosophila S2 cells after loading with 1 mM oleate for 12 hr. Confocal midsections (upper panels) and 3D reconstructions (lower panels) of BODIPY stained cells are shown. (C) RNAi-mediated depletion of several PC synthesis enzymes results in giant LDs. (D) Choline deficiency leads to giant LDs. S2 cells were grown in normal or choline-deficient medium, oleate loaded, and analyzed as in (B). (E and F) PC stabilizes artificial LDs better than other phospholipids. Stability and size of artificial droplets consisting of the indicated lipids were determined by fluorescence microscopy (E) and light scattering (F). Values are mean ±SD of three experiments. Bars, 5 μm.

Figure 2. Among PC Synthesis Enzymes, CCT Uniquely Localizes to LDs

(A) Overview of the reactions of the Kennedy pathway. (B) Transiently expressed mCherry-tagged CCT1 and CCT2, but not CK and CPT, (left panels) localize to LDs in S2 cells loaded with 1 mM oleate for 12 hr. Overlays and a magnification are shown (right two panels). Bar, 5 μm (overview) or 1μm (magnification). (C) Endogenous CCT localizes to LDs. Normalized peptide intensities determined by LC-MS/MS are shown for fractions of a LD purification. CCT, hormone-sensitive lipase (HSL, LD marker), disulfide isomerase (PDI; ER marker), lamin (nuclear marker), cytochrome oxidase subunit Va (CoVa; mitochondrial marker), and aldehyde dehydrogenase (ALDH; cytosol marker) are shown. Bottom panels show western blots for CCT1, KDEL receptor, and Gapdh and a TLC for TG of each fraction normalized to protein.

Figure 3. CCT1 Targets to LDs after an Initial Delay and Relocalizes to the Nucleus when Oleate Is Removed

(A) mCherry-CCT1 (left panels) localizes to LDs (BODIPY, second left panels) after 3 hr of oleate loading and relocalizes to the nucleus when S2 cells are shifted to lipid-free medium. Confocal midsections of each time point are shown, as well as overlays and a magnification of LDs. Scale bar, 5 μm (overview) or 1 μm (magnification). (B) LDs grow during oleate loading. Shown are box plots of LD diameters at different time points of oleate loading (n > 100 for each time point). (C) Number of cells with targeting of CCT1 to LDs increases between 1 and 3 hr of oleate loading and decreases after oleate removal. Values are means ±SD of three experiments. (D) The amount CCT1 on LDs increases in the first 12 hr of oleate loading and decreases after oleate removal. Shown is quantification of mCherry-CCT1 fluorescence intensity outside the nucleus. Values are means ±SD of three experiments. (E) PC/TG ratio of LDs decreases for up to 5 hr of oleate loading but stabilizes afterwards. Values are mean ±SD of three experiments. (F) Cellular TG levels increase steadily in oleate-loaded cells. Values are mean ±SD of three experiments. (G) After an initial delay, cellular PC levels increase linearly over time in oleate-loaded cells. Values are mean ±SD of three experiments.

Figure 4. CCT1 Shuttles between Nucleus and Cytosol before Oleate Loading but Stably Associates with LDs during Their Expansion

(A and B) mCherry-CCT1 shuttles between nucleus and cytosol as seen in images (A) and quantitation of three FLIP experiments (B; means ±SD are shown) in which a spot in the cytosol was repeatedly bleached. (C and D) CCT1 is mobile in the nucleus. A FLIP experiment analogous to (A) and (B) is shown, but a spot in the nucleus was bleached repeatedly. (E) mCherry-CCT1 fluorescence (upper panels) of a bleached LD (shown in lower panels) does not recover. Box indicates the photobleached LD, which is shown in higher magnification. Prebleach (left), immediately after bleach (middle), and postbleach (right) images of the experiment are shown. (F) Normalized fluorescence intensity of mCherry-CCT1 on a bleached LD as in (C) over time. Values are means ±SD of three experiments. (G) mCherry-CCT1 (upper panels) fluorescence remains stable on LD (BODIPY labeled, middle panels) in a bleached cell. Lower panels show overlays of the two channels. At time 0 min, all signal except on the indicated LD was photobleached. Prebleach (left), immediately after bleach (middle), and postbleach (right) images of the FRAP experiment are shown. (H) Normalized fluorescence intensity of mCherry-CCT1 of experiments as in (G) over time. Values are mean ±SD of three experiments. Bars, 5 μm.

Figure 5. CCT Activation by LD Binding Is Essential for Normal LDs

(A) Expression of different mCherry-CCT1 constructs (left panel) reveals the requirement of the helical region for LD (BODIPY stained; second left panels) binding. Overlays (merge) of the two channels and a magnification of LDs are shown. Bar, 5 μm (overview) or 1 μm (magnification). (B) Endogenous CCT activity increases when cells are loaded with oleate and decreases in lipid-free medium. CCT activity normalized to enzyme amount before oleate loading, after 12 hr of oleate loading, and after 20 hr in lipid-free medium subsequent to 24 hr oleate loading is shown. Values are mean ±SD of three experiments. (C) CCT is active on LDs. Specific activity of cytosol, membrane, and LD fractions of S2 cells loaded with oleate for indicated times is shown. Values are means ±SD of three experiments. (D) CCT activity from extracts inversely correlates with PC content of added artificial droplets generated with the indicated phospholipid compositions. Values are means ±SD of three experiments. (E) LD binding is essential for CCT1 normal LDs. Different mCherry-CCT1 constructs (left panels) were expressed in cells depleted for endogenous CCT1 by RNAi against its 3′UTR, and the ability of the mutants to rescue the LD phenotype (LDs stained with BODIPY; second left panels)was tested. Merged images and magnifications of LDs are shown. Bar, 5 μm (overview) or 1 μm (magnification).

Figure 6. CCT Is Important for LD Homeostasis In Vivo and in Mammalian Cells

(A) Flies with CCT1 knockdown in larval fat body have giant LDs there. LDs in fat body of CgG4/+ and CgG4/CCT^RNAi flies were stained with BODIPY (middle). Expression of the UAS-Gal driver was monitored by mCherry levels (left). Merged images are shown on the right. Bar, 10 μm. (B) Flies with CCT1 knockdown in larval fat body have increased TG content. Lipids from third-instar larval from Cg /+, +/CCT^RNAi, and CgG4/CCT^RNAi flies were quantified by TLC. (C) Flies with CCT1 knockdown in larval fat body are more resistant to starvation. Fifty percent survival rates of CgG4/+, +/CCT^RNAi, and CgG4/CCT^RNAi flies are indicated. Values are mean ±SD of three experiments. (D) CCTα-GFP (left) targets to LDs (Nile red; middle) and membranes in oleate-loaded Raw267.4 macrophages. Merged images (second right panels) and magnifications (right panels) are shown. Bar, 5 μm (overview) or 1 μm (magnification). (E) CCTα moves from the soluble fraction to membranes and LDs during oleate loading of BMDM. The indicated fractions were probed for CCT α and the marker proteins ADRP, calnexin, and Gapdh. (F and G) CCTα is activated by oleate loading in BMDMs and Raw267.4 macrophages. Values are mean ±SD of three experiments analogous to Figure 5B.(H) CCTα knockdown in HTCN-treated BMDM from Cctα^F/F mice results in giant LDs. Confocal images of BODIPY-stained untreated BMDM of Cctα^F/F mice or of HTNC-treated R26R-EYFP are shown as controls. Bar, 5 μm.

Figure 7. Model of CCT Regulation during LD Expansion

(A–D) See the Discussion for details.