Regulation of lipid droplet size and phospholipid composition by stearoyl-CoA desaturase

Fatty acid desaturation regulates membrane function and fat storage in animals. To determine the contribution of stearoyl-CoA desaturase (SCD) activity on fat storage and development in the nematode Caenorhabditis elegans, we analyzed the lipid composition and lipid droplet size in the fat-6;fat-7 desaturase mutants independently and in combination with mutants disrupted in conserved lipid metabolic pathways. C. elegans with impaired SCD activity displayed both reduced fat stores and decreased lipid droplet size. Mutants in the daf-2 (insulin-like growth factor receptor), rsks-1 (homolog of p70S6kinase, an effector of the target of rapamycin signaling pathway), and daf-7 (transforming growth factor β) displayed high fat stores, the opposite of the low fat observed in the fat-6;fat-7 desaturase mutants. The metabolic mutants in combination with fat-6;fat-7 displayed low fat stores, with the exception of the daf-2;fat-6;fat-7 triple mutants, which had increased de novo fatty acid synthesis and wild-type levels of fat stores. Notably, SCD activity is required for the formation of large-sized lipid droplets in all mutant backgrounds, as well as for normal ratios of phosphatidylcholine (PC) to phosphatidylethanolamine (PE). These studies reveal previously uncharacterized roles for SCD in the regulation of lipid droplet size and membrane phospholipid composition.


Fatty acid composition and lipid analysis
Fatty acid composition of young adult nematodes was determined by gas chromatography/mass spectrometry (GC/MS) as previously described ( 6,28 ). Separation of the TAG and phospholipid fractions used a two-solvent TLC protocol. Approximately 10,000 young adult stage C. elegans were washed from NGM plates several times in water. Most of the water was removed, and worm pellets were frozen in liquid nitrogen. Lipids were extracted by adding 5 ml of ice-cold chloroform:methanol (1:1) and incubating overnight at Ϫ 20° with occasional shaking. A solution of 0.2M H 3 PO 4 and 1M KCl was added to samples, which resulted in phase separation of the organic and aqueous phase. The organic phase was removed and dried under argon, then resuspended in chloroform. Samples were loaded in triplicate, and TLC plates were developed two thirds of the way up the plate in the fi rst solvent system: chloroform: methanol:water:acetic acid (65:43:3:2.5), dried, and then the second solvent system hexane:diethylether:acetic acid (80:20:2) was developed to the top of the plate. Lipids were visualized under UV light after spraying the plate with 0.005% primuline, and spots corresponding to TAG and the major phospholipids were scraped, spiked with a known standard (15:0), and transesterifi ed for GC/MS analysis to determine the fatty acid composition as well as to determine the relative levels of TAG and phospholipid (PL) fractions. At least three biological replicates were used for TLC analysis. Signifi cance was determined with one-way ANOVA analysis and Tukey's multiple com parison posttest using GraphPad Prism 5 software.
Stable isotope labeling of fatty acids was performed essentially as described ( 29 ). Briefl y, equal amounts of bacteria grown in either LB ( 12 C media) or isogrow (98.5% 13 C-enriched, Sigma) were mixed and plated onto agarose plates. For each sample, approximately 30,000 synchronized L1 nematodes prepared from hypochlorite treatment of gravid adults were added to the plates and grown for 48 h at 20°C (worms reached L4 larval stage). Nematodes were washed off the plates, their lipids were extracted, and fatty acids were analyzed by GC/MS as described ( 29 ). Isotopomers were monitored in a scanning ion mode corresponding to the fatty acid species of in terest: 16:0 was scanned from m/z 270-286 , 18:0 was scanned from m/z 298-316, and 18:1(n-7) was scanned from m/z 296-314.
Quantifi cation of lipid droplet size. For measurement of lipid droplet size, at least ten young adult worms of each genotype stained with postfi xed Nile Red were photographed ( 13 ). For each worm photograph, a 26 × 26 m square was placed arbitrarily over the mid-intestinal region, and within the square, each visible Nile Red-stained droplet was manually traced using the circle tool of Image Pro Plus software, which recorded the diameter of each droplet. For each worm, the average lipid droplet diameter contains a ⌬ -12 fatty acid desaturase (FAT-2) that enables it to synthesize polyunsaturated fatty acids from oleic acid ( 8 ). Polyunsaturated fatty acids (PUFA) are not present in the C. elegans laboratory diet, because they are not synthesized by E. coli . Thus, in C. elegans , SCD is the fi rst desaturase required for the biosynthesis of a wide range of PUFAs ( 9 ).
Lipid synthesis and oxidation are regulated by various nutrient and energy sensing pathways. These include the conserved insulin/insulin growth factor-1 (IIS) signaling pathway, the target of rapamycin (TOR) pathway, the transforming growth factor ␤ (TGF ␤ ) pathway, and the AMPactivated protein kinase pathway (AMPK). C. elegans mutants with defi cient IIS or TGF ␤ signaling, such as daf-2 and daf-7 , tend to arrest during development as dauer larvae, and under conditions where the mutants reach adulthood, they are resistant to environmental stresses, have a greatly extended lifespan, and store excess fat compared with wildtype (WT) (10)(11)(12)(13)(14). Similarly, the C. elegans TOR mutant let-363 , along with Raptor mutant daf-15 , arrest in larval stages with excess fat stores (15)(16)(17). Mutants in the gene encoding the TOR complex 2 component Rictor, rict-1 , do not arrest as larvae, but they exhibit slow growth, increased fat stores, and shortened lifespan ( 18,19 ). Similarly, rsks-1 mutants in the downstream ribosomal S6 kinase also exhibit slow growth and altered lifespan ( 20 ). Finally, overexpression of the C. elegans AMPK gene aak-2 increases lifespan ( 21 ), and AAK-2 activity is necessary for proper energy expenditure during dauer larval stages ( 22 ).
Lipid droplets are fat-storing organelles consisting of a hydrophobic core of TAG and cholesterol ester surrounded by a phospholipid monolayer containing various proteins ( 23 ). Proteomic studies of lipid droplets from various organisms, including C. elegans, reveal that lipid droplets are associated with a complex mixture of proteins that are predicted to play roles in lipid synthesis and degradation, membrane traffi cking, and protein degradation ( 24,25 ). C. elegans contains lipid droplets in intestinal, hypodermal, and gonadal tissues ( 9,26 ).
The fat-6;fat-7 strain exhibits opposite fat storage phenotypes compared with the daf-2 (IIS), daf-7 (TGF ␤ ), and previously characterized mutants in TOR signaling. The SCD-defi cient fat-6;fat-7 double mutants have reduced fat stores, whereas the other mutants cause nematodes to store higher amounts of TAG. We constructed triple-mutant strains to determine whether SCD activity is necessary for high fat stores in these strains. We found a striking requirement for endogenous SCD activity for the regulation of lipid droplet size and discovered that SCD activity infl uences the relative ratios of membrane phospholipid species.

Nematode strains and growth conditions
Nematode growth media (NGM) was used to maintain C. elegans with the E. coli (OP50) at 20°C. The WT strain was N2. The strains used in this study were BX106 fat-6(tm331) , BX107 fat-5(tm420) , BX110 fat-5(tm420);fat-6(tm331) , BX156 fat-6(tm331);fat-7(wa36) , to determine the frequency of lipid droplets of various sizes. The individual lipid droplets in WT ranged in diameter from 0.5 m to 3.7 m, with an average size of 1.4 m, while lipid droplets in fat-6;fat-7 were much smaller, with a range from 0.2 m to 1.5 m, and an average diameter of 0.5 m ( Fig. 1B ). In addition, we examined lipid droplet sizes from the single SCD mutant strains fat-5 , fat-6 , and fat-7 single mutants, as well as the fat-5;fat-6 double mutant, and we found that the lipid droplet diameters in these strains do not appreciably differ from WT (supplementary Fig. I-A). Thus, the C. elega ns SCD strain with the most severe lipid droplet-size defect is fat-6;fat-7 , which also has the most severe fatty acid composition defect among the SCD single-and double-mutant strains ( 6,7 ).
Next, we analyzed the fatty acid composition of individual membrane phospholipids in the fat-6;fat-7 double mutants. In WT, PUFA is detected mostly in the phospholipid fraction, in contrast to TAG, which contains only small amounts of PUFA ( 7,29,36 ). When SCD activity is blocked by simultaneous mutations in fat-6 and fat-7 , normal PUFAs are not formed, because the FAT-6 and FAT-7 desaturases are required to synthesize PUFA de novo from acetyl CoA or to synthesize them from fatty acid precursors (primarily palmitic acid) derived from the bacterial diet ( 7,29 ). In the absence of FAT-6 and FAT-7, unusual C18 PUFAs are synthesized via desaturation and elongation of palmitoleic acid (16:1n-7), which is synthesized from palmitic acid by the ⌬ 9 desaturase FAT-5 ( 7,29 ). For this study, we examined the fatty acid composition of TAG faction, as well as the membrane phospholipids phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylserine (PS), and phosphatidylinositol (PI) in the fat-6;fat-7 double mutants. Similar to the distribution of PUFAs in WT lipids, we detected the unusual C18 PUFAs in PI and PS, as well as PC, but they are present to a lesser extent in PE and in the TAG fraction (supplementary Table I). In agreement with our previous studies, fat-6;fat-7 had overall lower levels of TAG than WT ( Fig. 1C , Table 1 ) and a greatly increased composition of saturated fatty acids and decreased composition of polyunsaturated fatty acids ( Fig. 1D ).

Dietary oleic acid promotes PUFA production but does not restore lipid droplet size to WT levels
Our previous work showed that the slow growth and reduced brood size in fat-6;fat-7 worms were partially rescued by dietary oleic acid ( 7 ). To see whether dietary oleic acid can rescue the lipid droplet and fat storage defects of fat-6;fat-7 we grew the worms on plates containing 0.1-0.5 mM sodium oleate. We examined the lipid droplet size in the oleate-supplemented worms and found they had a larger size distribution compared with unsupplemented fat-6;fat-7 worms, indicating that dietary oleate induces large-sized lipid droplets in SCD-defi cient worms, although oleate levels up to 0.5 mM cannot fully rescue lipid droplet size to that of WT ( Fig. 1A, B , supplementary Fig. I-B). GC/MS analysis revealed that fatty acid uptake did not increase signifi cantly above that seen in 0.2 mM oleate; therefore, the 0.2 mM concentration was was calculated. Statistical comparisons (one-way ANOVA and Tukey's multiple comparison test) were performed using GraphPad Prism 5 software.

Fatty acid oxidation assay
Fatty acid oxidation was performed on L4 nematodes essentially as described ( 30 ), except that tritiated palmitic acid (30 Ci/ mmol, Perkin Elmer, Waltham, MA) was used as substrate.

Quantitative RT-PCR analysis
WT, fat-6;fat-7 , and daf-2;fat-6;fat-7 nematodes were synchronized and harvested at L4 stage. RNA and cDNA was prepared as described ( 7 ). Real-time PCR assays were run and monitored with an ABI Prism 7000 Sequence Detection system (Applied Biosystems, Foster City, CA). The real-time PCR was conducted on three treatment groups, with each individual treatment group in triplicate. Threshold values (Ct) for the gene of interest and a housekeeping gene tbb-2 were determined using ABI Prism SDS software version 1.1 (Applied Biosystems). The expression level of the gene of interest was evaluated using the 2 Ϫ ( ⌬ ⌬ Ct) method ( 31 ).

Physiological assays
Growth rate. Eggs were isolated from gravid adults using hypochlorite treatment and then plated onto NGM plates seeded with E. coli strain OP50. The number of adults and total number of nematodes were determined at various time points.
Brood size. For analysis of total progeny produced per worm, 10-20 L4s were transferred individually to fresh NGM plates seeded with E. coli strain OP50. Worms were transferred daily until they did not produce any more progeny. Two days after removal of the adult, the live progeny of each genotype were counted.

SCD activity is necessary for large-sized lipid droplets
In the nematode C. elegans , the lack of SCD activity in the fat6;fat-7 double-mutant strain has profound consequences for the fatty acid composition of membrane phospholipids and TAG storage lipids ( 7 ). In our previous study, we showed faint Nile Red staining in fat-6;fat-7 animals, which increased in intensity when the animals were supplemented with 0.1 mM sodium oleate ( 7 ). However, recent work in many labs has demonstrated that Nile Red fl uorescence in live animals does not correlate with actual TAG stores but instead stains lysosome-related organelles ( 13,(32)(33)(34). In contrast, the intensity of Nile Red, Sudan black, or Oil-Red O staining in nematodes that have been fi xed with isopropanol or paraformaldehyde corresponds well with biochemical measurements of TAG ( 13,33 ) and stains lipid droplets, not lysosome-related organelles ( 35 ). We therefore analyzed the lipid-staining pattern of fi xed, young-adult-stage nematodes (with 0-2 eggs), and we found that the SCD-defi cient fat-6;fat-7 double mutants have fewer and smaller lipids droplets than WT controls ( Fig. 1A ). We used imaging software to measure lipid droplet diameter of cross-sections of the mid-intestinal region

Interactions between SCD and conserved metabolic pathways regulating fat storage and fat oxidation
To examine the role of SCD in the formation of lipid droplets, TAG storage, and growth and development, we generated the triple-mutant strains that allowed us to examine the role of SCD activity in the background of mutations that confer high fat stores or altered fat oxidation.

AMPK mutants have WT lipid composition and do not suppress low fat in SCD-defi cient worms.
Mutations in the C. elegans aak-2 gene, which encodes the ␣ subunit of AMPK, affect the regulation of fat oxidation, especially under conditions of starvation ( 22 ). Because experiments in C. elegans indicate that the low fat stores in SCD mutants are associated with increased expression of fat oxidation genes ( 7 ) and because experiments in mice indicate that SCD inhibition leads to activation of AMPK ( 37 ), we tested whether the low fat stores in the fat-6;fat-7 double mutants could be suppressed by a mutation in aak-2 . We fi rst examined the lipid droplet size and measured the lipid composition of aak-2 mutants. We found that, under growth conditions providing ample food, the aak-2 mutants were indistinguishable from WT with respect to lipid droplet size, lipid composition, and growth rate ( Fig. 2 A-E , Table 2 , chosen for detailed lipid analysis. Analysis of TAG and phospholipid fractions indicated that 0.2 mM dietary oleate also signifi cantly increased the TAG levels of the fat-6;fat-7 strain ( Fig. 1C , Table 1 ). The 0.2 mM concentration of oleate resulted in an incorporation of oleic acid and its downstream products [linoleic acid (18:2n-6), ␥ -linolenic acid (18:3n-6), dihommogamma-linolenic acid (20:3n-6), and eicosapentaenoic acid (20:5n-3)] such that they made up 15-17% of total fatty acids in the supplemented fat-6;fat-7 worms (supplementary Table I). Interestingly, the fatty acid composition of the supplemented animals still showed high levels of stearic acid (18:0), similar to unsupplemented fat-6;fat-7 double mutants, and signifi cant amounts of unusual C18 PUFAs (6%, compared with 12% in unsupplemented double mutants and undetectable in WT animals). Thus, dietary oleate promoted PUFA synthesis in both WT and fat-6;fat-7 , and it altered the fatty acid composition profi le of fat-6;fat-7 animals but did not restore it to WT ( Fig. 1D , supplementary Table I). The failure of dietary oleate to completely rescue lipid droplet size in fat-6;fat-7 worms indicates a requirement for endogenous synthesis of unsaturated fatty acids for optimal lipid droplet size and TAG stores.  Values are the average (SEM) from three independent lipid extractions of young adult stage C. elegans . Those not sharing a common letter within the same column differ ( P < 0.05) is infl uenced by the TOR pathway, and mutants in conserved members of TOR complexes 1 and 2 show alterations in fat storage ( 15,16,18,19 ). In mammals, a downstream target of TOR signaling is the p70 ribosomal S6 kinase, which in C. elegans promotes germline proliferation and is required for the proper regulation of lifespan ( 20,38 ). We examined lipid droplet size and fatty acid composition in the TOR complex 2 component rict-1 mutants as well as in rsks-1 mutants. We found that, in both strains, the formation of large lipid droplets depends on FAT-6 and FAT-7 ( Fig. 2A, B , supplementary Fig. II-B). In contrast, the requirements for RICT-1 and RSKS-1 for producing optimal growth rate and brood size act in parallel to SCD, because the triple mutants have a more severe growth and brood size reduction than the fat-6;fat-7 strain ( Fig. 2E, F , supplementary Fig. II-C).
Because fat composition had not been previously studied in the rsks-1 mutant, we examined the lipid composition in this mutant strain, as well as in the rsks-1;fat-6;fat-7 triple-mutant strain. We discovered that rsks-1 mutants, similar to other mutants in the TOR pathway, have increased TAG content compared with WT ( Fig. 2C , supplementary supplementary Table II). In addition, the triple-mutant strain aak-2;fat-6;fat-7 was indistinguishable from fat-6;fat-7 with respect to lipid droplet size, lipid composition, and growth rate ( Fig. 2A-E , Table 2 , supplementary Table II). We detected a smaller brood size in the aak-2 mutant compared with WT and a severely small brood size in the aak-2;fat-6;fat-7 triple mutant, suggesting a role for AAK-2 in optimal fecundity ( Fig. 2F ). The severe reduction in live progeny in the triple mutant is consistent with additive developmental defects in the aak-2 and fat-6;fat-7 strains.
A second isoform of the ␣ subunit of AMP kinase is encoded by the aak-1 gene. To ensure that AMPK activity was effi ciently knocked down, we treated the WT, aak-2 , and aak-2;fat-6;fat-7 strains with aak-1 RNAi. We found no change in lipid droplet size in any of the strains with aak-1 RNAi treatment (supplementary Fig. II-A). With respect to the regulation of fat storage, our results show no role for AMPK in modulating the fat stores in the SCD mutants.
S6K regulates fatty acid composition and fat stores. TOR is a conserved serine/threonine kinase that regulates growth in response to nutritional signals. Fat content in C. elegans droplets in in daf-2 mutants ranged 0.5-7.5 m, with an average size of 2.0 m ( Fig. 2A, B ), and, as previously reported ( 13,42 ), the daf-2 mutant strain contain high levels of TAG and slightly reduced levels of PUFAs ( Fig. 2C, D , Table 2, supplementary Table II ). The daf-2;fat-6;fat-7 strain had lipid droplets that ranged 0.3-3.0 m, with an average lipid droplet size of 1.1 m, which was closer to the size of the fat-6;fat-7 strain than was daf-2 and which was smaller than WT ( Fig. 2A, B ). This indicates that SCD activity is necessary to produce the large-sized lipid droplets (>3 m) that are found in WT and daf-2 mutants. However, qualitative staining with Nile Red indicated that the lipid droplets in daf-2 , fat-6;fat-7 worms were more abundant than those in the fat-6;fat-7 strain, and lipid analysis revealed that the percentage of fatty acids found in the TAG fraction was signifi cantly higher in young adult daf-2 , fat-6;fat-7 worms than in fat-6;fat-7 , similar to the levels measured in similar-aged WT nematodes ( Fig. 2C , Table 2 ). This indicates that, whereas FAT-6 and FAT-7 SCD activity is required to produce large-sized lipid droplets, it is only partially required for increased fat stores in the daf-2 mutant background. While the daf-2;fat-6;fat-7 strain contained increased PUFA in all of the lipid classes compared with the fat-6;fat-7 strain ( Fig. 2D , supplementary Table II), the increased fat stores and increased PUFAs did not confer any suppression of the slow growth or reduced brood size of the fat-6;fat-7 double mutants; in fact, the daf-2;fat-6; fat-7 triple mutants grew at a slower rate than the fat-6;fat-7 double mutants ( Fig. 2E, F ). Taken together, these fi ndings reveal an interaction between IGF signaling and SCD activity in the regulation of fat stores.

Increased fat stores in daf-2;fat-6;fat-7 mutants are not a consequence of reduced fat oxidation but are associated with increased de novo fatty acid synthesis and increased FAT-5 activity
In mice and nematodes, SCD defi ciency leads to increased expression of ␤ -oxidation genes, which may lead to increased fat oxidation and reduced fat stores ( 2, 7 ). Therefore, we hypothesized that the increased fat stores in the daf-2;fat-6;fat-7 mutants may be a result of decreased expression of ␤ -oxidation genes in the daf-2 mutant background. We used real-time quantitative RT-PCR to examine expression of mitochondrial and peroxisomal Table II). However, in combination with SCD defi ciency, TAG accumulation was similar to the fat-6;fat-7 double mutant, indicating that SCD activity is required for the accumulation of high fat stores and large lipid droplets in the rsks-1 mutants ( Fig. 2A-D , Table 2 , supplementary Table II). Interestingly, while rict-1 mutants have a nearly WT fatty acid composition (supplementary Fig II-D), all of the lipids in the rsks-1 strain showed an increase in C20 omega-6 PUFAs (20:3n-6 and 20:4n-6) and a decrease in the C20 omega-3 fatty acids (20:4n-3 and 20:5n-3) (supplementary Fig. II-E, supplementary Table II). We found a fatty acid composition defect similar to rsks-1 in RNAi treatment of more than 30 ribosomal proteins that we screened [C20 fatty acid composition of rps-9(RNAi) is shown in supplementary Fig. II-F]. This indicates that for fatty acid composition, rsks-1 mutants, not rict-1 , phenocopy the altered omega-6/omega-3 fatty acid composition changes that result from the disruption of protein translation.
Overlapping developmental requirements of DAF-7 and SCD. The C. elegans TGF ␤ mutant daf-7 , similar to daf-2 mutants, tends to arrest as dauer larvae ( 39 ). However, when grown under conditions that allow the worms to develop to adulthood, the daf-7 mutants store excess fat ( 13,40 ). Our attempts to construct the daf-7;fat-6;fat-7 triplemutant strain were unsuccessful. We observed that the daf-7;fat-6 double mutants showed increased incidence of dauer formation at 20°C compared with the daf-7 strain (data not shown), and after numerous attempts to obtain the triple mutant, only one worm was confi rmed by PCR to possess the daf-7;fat-6;fat-7 genotype, and this nematode did not produce live progeny. Therefore, we conclude that the additive developmental defects of TGF ␤ and SCD defi ciency led to lethality of the daf-7;fat-6;fat-7 triple mutants.

SCD defi ciency leads to low PC:PE ratio
Our extensive lipid analysis allowed us to compare the ratios of the two major membrane phospholipids, PC and PE, in all of the strains. The ratio of PC to PE is important for proper membrane function. We found that the ratio of PC:PE among independently grown batches of young adult WT nematodes ranged 1.41-1.65, with an average ratio of 1.51 ( Fig. 4A , Table 1 ). The ratio of PC:PE was somewhat higher than WT in the aak-1 , rsks-1 , and daf-2 mutant strains, ranging 1.57-2.26, with an average of 1.78 for aak-2 , 1.96 for rsks-1 , and 1.69 for daf-2 mutants. In contrast, the ratio of PC:PE was less variable and signifi cantly lower in the fat-6;fat-7 mutant strain, ranging 1.11-1.17, with an average of 1.13. Strikingly, in combination with aak-2 , rsks-1 , and daf-2 , the ratio remained low, ranging 0.99-1.25 in the triple-mutant strains, with an average of 1.08 for aak-2;fat-6;fat-7 , 1.04 for rsks-1;fat-6;fat-7 , and 1.13 for daf-2;fat-6;fat-7 ( Fig. 4A , Table 2 ). This fi nding indicates that the low PC:PE ratio correlates with the fatty acid composition of the animals, with either the high content of saturated fatty acids or the low content of polyunsaturated fatty acids infl uencing the relative amounts of PC and PE in membranes.
Reduced amounts of PC are associated with large lipid droplets in Drosophila S2 cells ( 45 ) and fatty livers in mice ( 46 ). Therefore, we were surprised that the small lipid droplets in fat-6;fat-7 mutants were associated with reduced amounts of PC relative to PE. In C. elegans , depletion of sams-1 leads to diminished PC ( 47 ). The sams-1 gene encodes S-adenosylmethionine synthase, which is required to transfer methyl groups onto PE to form PC, one pathway of PC synthesis. Because the sams-1 mutants have large lipid droplets ( 47 ), consistent with reduced PC content, we used RNAi to deplete sams-1 in the fat-6;fat-7 double mutant ( Fig. 4B ). The large lipid droplets induced by depletion of sams-1 did not form in the SCD-defi cient fat-6; fat-7 worms, indicating that SCD is essential for lipid droplet expansion, even when PC levels are reduced. Taken together, these studies reveal an important role for SCD activity in the regulation of lipid droplet size, independent of fat accumulation and membrane phospholipid ratios.

DISCUSSION
Construction of the fat-6;fat-7 triple-mutant strains enabled us to ascertain whether various mutations that, on their own, lead to high fat stores in C. elegans , are able to overcome the low fat stores of the SCD-defective fat-6;fat-7 mutants. We found that development and fat storage is affected in unique ways. The combination of TGF ␤ with SCD in the daf-7;fat-6;fat-7 strain is lethal, whereas the combination of reduced insulin signaling and SCD in the daf-2;fat-6;fat-7 triple-mutant strain leads to the ability to store WT levels of fat. In other mutant strains, in which SCD defi ciency is combined with mutations in aak-2 , sams-1 , and rsks-1 , growth and fat storage resemble the SCD mutants, revealing an essential role for SCD in effi cient ␤ -oxidation genes as well as fatty acid-binding protein genes in WT, fat-6;fat-7 , and daf-2;fat-6;fat-7 worms. We found that, compared with WT, 12 of 29 genes tested showed increased expression in fat-6;fat-7 worms. However, the expression remained high in the daf-2;fat-6;fat-7 worms, indicating that reduced expression of genes encoding ␤ -oxidation machinery is unlikely to be the mechanism for high fat content in daf-2;fat-6;fat-7 worms (supplementary Table III). Furthermore, a direct assay of fatty acid oxidation activity ( 30 ) demonstrated that both the fat-6;fat-7 and daf-2;fat-6;fat-7 strains showed higher rates of oxidation of palmitic acid (16:0) than did WT ( Fig. 3A ). We conclude that high TAG accumulation in daf-2 ; fat-6;fat-7 compared with WT and fat-6;fat-7 is not a result of decreased fatty acid oxidation.
Previous studies have demonstrated that daf-2 mutants display increased de novo fatty acid synthesis compared with WT ( 29 ). We used stable isotope labeling and GC/ MS analysis to determine whether the daf-2;fat-6;fat-7 triple mutant had increased de novo synthesis compared with the fat-6;fat-7 double mutant. We found an increased abundance of isotopomers associated with de novo fatty acid synthesis in the daf-2;fat-6;fat-7 triple-mutant strain compared with fat-6;fat-7 and WT ( Fig. 3B ). Interestingly, the fat-6;fat-7 double mutant showed increased de novo synthesis of stearic acid (18:0), although the relative synthesis of palmitic acid (16:0) and vaccenic acid (18:1n-7) was not different than WT ( Fig. 3C ). In contrast, the daf-2; fat-6;fat-7 showed increased de novo synthesis compared with WT and fat-6;fat-7 in all of the fatty acids analyzed ( Fig. 3C ). Therefore, mutation in the daf-2 gene promotes de novo fat synthesis in SCD-defi cient worms as well as in WT, and this increased synthesis correlates with increased TAG levels in strains containing the daf-2 mutation.
The third isoform of ⌬ 9 desaturase in C. elegans , FAT-5, is a palmitoyl-CoA desaturase that only is only active on palmitic acid (16:0) and does not desaturate stearic acid (18:0) ( 6,43 ). The fat-5 gene is upregulated in fat-6;fat-7 double mutants ( 7 ) as well as in daf-2 mutants ( 44 ). To estimate FAT-5 activity, we compared the abundance of the product (16:1) to the precursor (16:0) to estimate the FAT-5 activity. We found that the ratio of palmitoleic acid (16:1) to palmitic acid (16:0) was increased in the daf-2 and fat-6;fat-7 background compared with WT, and strikingly, the 16:1 to 16:0 ratio was greatest in the triple-mutant strain ( Fig. 3D, E ). The ratio of 16:1 to 16:0 was highest in the daf-2;fat-6;fat-7 strain in both fat storage lipids (TAG, Fig. 3D ) and membrane phospholipids (PC, Fig. 3E ). In addition, every lipid class contained higher amounts of PUFAs in the daf-2;fat-6;fat-7 mutants compared with the fat-6;fat-7 strain. Because strains containing mutations in both fat-6 and fat-7 form unusual PUFAs through the FAT-5 pathway ( 7 ), the increased FAT-5 activity is consistent with increased PUFAs in PC as well as other lipid classes in daf-2;fat-6;fat-7 compared with fat-6;fat-7 ( Fig. 2D , supplementary Table II). This fi nding suggests that increased FAT-5 activity may facilitate increased TAG synthesis in the daf-2;fat-6;fat-7 triple mutants.  Table II). Although upstream components of the TOR signaling pathway, such as daf-15 /RAPTOR, rict-1 /RICTOR, and let-363 /TOR, also have increased fat stores ( 15,19 ), we did not identify the alteration in the membrane omega-3:omega-6 ratios in rict-1 mutants, although we detected similar fatty acid composition changes in ribosomal protein RNAi knockdown worms (supplementary Fig. II). Therefore, the rsks-1 mutants and knockdowns in proteins required for translation have an altered fatty acid composition that is not observed in rict-1 mutants.
In spite of the decreased lipid droplet size, the abundance of lipid droplets, as well as the total TAG levels in the worms, was increased in the daf-2 mutant background. This was not due to a decrease of ␤ -oxidation conferred by the daf-2 mutation, because we found that fatty acid oxidation rates were increased in both the fat-6;fat-7 fat storage as well as in ensuring proper growth and development. Notably, the FAT-6 and FAT-7 SCDs regulate the size of lipid droplets and ratios of cellular phospholipids in every strain examined. Fig. 5 depicts a model summarizing the pathways examined in these studies. A link between SCD activity and lipid droplet size was previously observed in cell lines cultured from patients with Berardinelli-Seip congenital lipodystrophy ( 48 ). This study showed that patients carrying mutations in the Seipin gene had increased proportions of saturated fatty acid in their lipids, indicating decreased SCD activity and decreased size and abundance of lipid droplets.
Our extensive lipid analysis revealed previously unknown lipid composition defects in the S6kinase-defi cient rsks-1 strain. We found that the overall TAG levels were high in this strain, consistent with larger lipid droplets observed by Nile Red staining of fi xed worms. We also identifi ed increased omega-6 and decreased omega-3 the ER-lipid droplet interface during lipid droplet expansion ( 54 ). Although we neither observed FAT-6::GFP or FAT-7::GFP on the lipid droplet surface nor were these gene products identifi ed in a proteomic analysis of C. elegans lipid droplets ( 25 ), we found that endogenous SCD activity is necessary for effi cient lipid storage in C. elegans , and we found that dietary oleic acid did not restore large-sized lipid droplets, consistent with the metabolic channeling hypothesis. A recent study provides evidence that lipogenic enzymes, specifi cally GPAT4, relocalize from the ER to a subset of expanding lipid droplets ( 55 ). It is possible that proper membrane composition may be required for effi cient relocalization of TAG synthesis enzymes to expanding lipid droplets.
In mice, Drosophila , and C. elegans , disruption of PC synthesis leads to increased TAG and large-sized lipid droplets (45)(46)(47)56 ). Upon lipid loading of cells, PC is synthesized and its presence prevents the coalescence of lipid droplets ( 45 ). Our studies show that all of the SCD-defi cient strains have a decreased amount of PC relative to PE. This data is in agreement with a recent metabolic study of C. elegans SCD mutants which revealed a reduction in some phosphocholine derivatives in SCD mutants compared with WT ( 57 ). The small-sized lipid droplets in the fat-6;fat-7 double mutants seem contradictory to the reduced PC, since reducing PC synthesis leads to large lipid droplets. Even so, we show that SCD activity is required for the large-sized and daf-2;fat-6;fat-7 strains ( Fig. 3 ). Instead, increased de novo synthesis correlates with the increased fat stores in strains containing the daf-2 mutation ( 29 ). Our fi ndings indicate that the induction of the FAT-5 palmitoyl-CoA desaturase facilitates the increase in de novo fat synthesis.
Mice lacking SCD1 are lean and resistant to diet-induced obesity ( 2,49 ). They have reduced fatty acid and TAG synthesis in response to high dietary carbohydrates ( 50 ), and they also have increased fat oxidation rates in various tissues ( 49 ). Cell culture studies of lipid droplet formation rely on oleic acid in the culture media to induce the expansion of lipid droplets ( 45 ). However, our studies in an intact organism reveal that dietary oleic acid does not fully compensate for the endogenous synthesis of monounsaturated fatty acids. This is consistent with mouse studies that revealed that diets containing up to 5% oleate did not rescue the low fat stores in SCD1 mutants ( 50 ) and with human studies that showed that diets high in oleic acid, such as a Mediterranean diet, provide a protective role against obesity, whereas diets high in saturated fats and simple carbohydrates, which induce endogenous SCD activity, lead to excess fat stores ( 51 ).
Interestingly, the SCD1 was shown to colocalize with diacylglycerol acyltransferase (DGAT)2 in the mitochondrialassociated membrane subcompartment of the endoplasmic reticulum (ER) ( 52 ). DGAT enzymes are required for the formation of lipid droplets ( 53 ). The colocalization of SCD1 and DGAT2 suggests that metabolic channeling of endogenously synthesized monounsaturated fatty acids and DGAT2 optimizes TAG synthesis. Work in C. elegans demonstrated that DGAT2 and ACS-22 form a complex at  Mutations in the S6 kinase rsks-1 lead to slow growth, increased fat stores, and altered fatty acid composition. Similarly, mutations genes encoding DAF-7 /TGF ␤ and DAF-2/IIS receptor lead to increased fat stores. Large-sized lipid droplets and high PC:PE ratios in all mutants requires SCD activity, although in daf-2;fat-6;fat-7 mutants, overall fat stores are improved due to increased activity of the palmitoyl-CoA desaturase  droplets in the PC-defi cient sams-1 mutants. Therefore, even though PC levels are reduced in fat-6;fat-7 mutants, the levels are apparently adequate to prevent the coalescence of lipid droplets. Instead, the lipid droplet size appears to be driven by the ability of SCD to synthesize unsaturated fatty acids. It is tempting to speculate that the alteration in the PC:PE ratio represents a compensation for the increased presence of saturated fatty acids and decreased abundance of PUFAs in the SCD-defi cient strains. For example, PC molecules containing unsaturated fatty acid are typically cylinder shaped, whereas PE molecules, with a smaller headgroup, are typically cone shaped ( 58 ). Because the chain length and the relative degree of unsaturation infl uences the shape of the fatty acids, which in turn infl uences the phospholipid shape, we predict that the shape of the PC molecules may be altered in fat-6;fat-7 double mutants. The decrease in PC together with the increase in PE may provide a phospholipid composition that improves membrane function in the context of increased saturated fatty acids in the fat-6;fat-7 strains.