Identification of A Novel Link between the Intermediate Filament Organizer IFO-1 and Cholesterol Metabolism in the Caenorhabditis elegans Intestine

The intestine is an organ essential to organismal nutrient absorption, metabolic control, barrier function and immunoprotection. The Caenorhabditis elegans intestine consists of 20 cells harboring a dense intermediate filament network positioned below the apical plasma membrane that forms a junction-anchored sheath around the intestinal lumen. This evolutionarily conserved arrangement provides mechanical and overall stress-protection, and it serves as an important model for deciphering the role of intestinal architecture in metazoan biology. We recently reported that the loss-of-function mutation of the intestinal intermediate filament organizer IFO-1 perturbs this architecture, leading to reduced body size and reproduction. Here, we demonstrate that the IFO-1 mutation dramatically affects cholesterol metabolism. Mutants showed an increased sensitivity to cholesterol depletion, reduced cholesterol uptake, and cholesterol transfer to the gonads, which is also observed in worms completely lacking an intermediate filament network. Accordingly, we found striking similarities to transcriptome and lipidome profiles of a nuclear hormone receptor (NHR)-8 mutant. NHR-8 is homologous to mammalian LXR (liver X receptor) that serves as a sterol sensor and transcriptional regulator of lipid metabolism. Remarkably, increasing exogenous cholesterol partially rescues the developmental retardation in IFO-1 mutants. Our results uncover a novel link of the intestinal intermediate filament cytoskeleton to cholesterol metabolism that contributes to compromised growth and reproduction.


Introduction
The intestinal epithelium delineates the border between the exterior environment and the body interior. It serves a dual function by providing a protective barrier against the harmful contents passing through the intestinal lumen and by facilitating the secure uptake of biomolecules needed for organismal metabolism. Elucidating these seemingly opposing properties and identifying the contributing cellular components are keys to understanding the physiologic function of the gut. The simplicity of the Caenorhabditis elegans intestine provides an ideal in vivo system to study intestinal

Lipids Are Reduced in the Intestine of ifo-1 Mutants
The loss-of-function allele kc2 of the intestinal intermediate filament organizer gene ifo-1 leads to reduced body size and progeny production [6], which may be caused by compromised food uptake in the intestine. Functional and biochemical evidence for this idea, however, is still lacking. Ultrastructural analyses of intestinal cells revealed luminal widening, disordered microvilli, and large intermediate filament polypeptide-containing aggregates at cell-cell junctions [6]. A close look at the cytoplasm of the mutant intestine suggested a reduction in the electron-lucent lipid droplets, while the amount of the darker yolk granules seemed not to be altered ( Figure 1A,B). To quantify differences in lipid droplet composition, worms were stained with the lipophilic dye Nile red (for discussion of specificity, see [15]). A small but significant decrease in Nile red fluorescence was detected in ifo-1(kc2) compared to wild-type control animals ( Figure 1C-E).

ifo-1 Mutants Exhibit Increased Sensitivity to Cholesterol Depletion
Because lipid metabolism is influenced by cholesterol [14], we next asked whether mutants showed an altered cholesterol homeostasis. Embryos were placed on cholesterol-depleted plates, and their development was followed. Wild-type N2 embryos developed into adults and reproduced in the absence of cholesterol. However, they had fewer progeny in the F2 generation, which arrested at the L1/L2 stage (for comparable observations, see [13]). In contrast, ifo-1(kc2) mutant embryos did not develop into adults in the absence of cholesterol but arrested as L1/L2 larvae already in the F1 generation (Figure 2A,B,D).
To find out whether the ifo-1(kc2)-dependent larval arrest phenotype was caused by the re-organization of the intermediate filament-containing endotube, we made use of ifb-2(kc14) knockout animals, which lack an endotube altogether [3,4]. Monitoring the growth of ifb-2(kc14) in the absence of cholesterol surprisingly showed that they produced F2 progeny, which, similar to the wild type, arrested at early larval stages ( Figure 2C,D). Additional analyses revealed that the time of development of ifb-2(kc14) was not different from N2 ( Figure 2E). Furthermore, the brood size was reduced in comparison to N2 (see also [3]), which, however, was independent of the presence or absence of cholesterol ( Figure 2F). These findings suggested that the increased sensitivity of ifo-1(kc2) to cholesterol depletion may be caused by the junctional intermediate filament aggregates. On the other hand, the loss of intermediate filament polypeptides in ifb-2(kc14) did not affect cholesterol-dependency in the performed assay.

Cholesterol Uptake and Transport Is Reduced in ifo-1 and ifb-2 Mutants
The above findings suggested that the uptake of cholesterol and/or cholesterol transport to the developing embryos are compromised in the ifo-1(kc2) mutants. To distinguish between them, worms were fed with radioactive cholesterol for measurements of cholesterol uptake and cholesterol transfer to the developing progeny. chup-1(gk245) was used as a control for the first set of experiments because it has been shown to impede cholesterol uptake [16]. The wild-type chup-1 gene encodes an intestinal cholesterol transporter, the precise localization of which, however, has not been resolved to date [16]. Figure 3A illustrates that cholesterol uptake was significantly reduced in chup-1(gk245). An even stronger reduction was detected in ifo-1(kc2). A reduction in cholesterol uptake was also noted in ifb-2(kc14). Yet this reduction was less pronounced than in ifo-1(kc2) but similar to that in chup-1(gk245).
To quantify the transfer of cholesterol to progeny, cholesterol content was determined after radioactive cholesterol uptake in the F1 generation ( Figure 3B). ncr-1(nr2022) and ncr-2(nr2023) double mutants were used in this instance as controls. ncr-1 and ncr-2 have been implicated in intracellular sterol transport [17]. A slight but statistically insignificant reduction was noted for the double mutant in comparison to the wild type. A significantly reduced cholesterol transfer, however, was detected for ifb-2(kc14) and was much more pronounced for ifo-1(kc2).
Taken together, we conclude that ifo-1 and ifb-2 are both needed for cholesterol uptake and transport, albeit to different degrees. The lesser effect of the ifb-2 mutation may explain why overall development is affected in ifo-1(kc2) but not in ifb-2(kc14) (as is the case for chup-1(gk245); [16]).

Transcriptome Profiling Identifies Striking Similarities between ifo-1 and nhr-8 Mutants
Transcriptome analyses had been performed previously using dissected ifo-1(kc2) and wild-type intestines from L4 larvae to identify pathways that may be affected in the mutant [4]. When we compared the altered genes with other published mutants, considerable similarities were noted with the nhr-8(hd117) mutants [14]. nhr-8 encodes a nuclear hormone receptor homologous to mammalian LXR and has been implicated as a sterol-sensor with important functions for nematode cholesterol homeostasis, lipid metabolism, and steroid hormone production [14]. Additionally, 169 of the dysregulated genes in nhr-8(hd117) total lysates that were obtained from L3 larvae grown under low cholesterol at 25 • C were also dysregulated in dissected ifo-1(kc2) intestines that were prepared from L4 larvae maintained at normal cholesterol at 20 • C ( Figure 4A; Table S1). Only five of those genes were oppositely regulated in the different genetic backgrounds. No changes in the mRNA production for ifo-1 in nhr-8(hd117) or, conversely, for nhr-8 in ifo-1(kc2), could be detected.
To find out whether the ifo-1(kc2)-dependent larval arrest phenotype was caused by the reorganization of the intermediate filament-containing endotube, we made use of ifb-2(kc14) knockout animals, which lack an endotube altogether [3,4]. Monitoring the growth of ifb-2(kc14) in the absence of cholesterol surprisingly showed that they produced F2 progeny, which, similar to the wild type, arrested at early larval stages ( Figure 2C,D). Additional analyses revealed that the time of development of ifb-2(kc14) was not different from N2 ( Figure 2E). Furthermore, the brood size was reduced in comparison to N2 (see also [3]), which, however, was independent of the presence or absence of cholesterol ( Figure 2F). These findings suggested that the increased sensitivity of ifo-1(kc2) to cholesterol depletion may be caused by the junctional intermediate filament aggregates. On the other hand, the loss of intermediate filament polypeptides in ifb-2(kc14) did not affect cholesteroldependency in the performed assay.  double mutants were used in this instance as controls. ncr-1 and ncr-2 have been implicated in intracellular sterol transport [17]. A slight but statistically insignificant reduction was noted for the double mutant in comparison to the wild type. A significantly reduced cholesterol transfer, however, was detected for ifb-2(kc14) and was much more pronounced for ifo-1(kc2).
Gene ontology enrichment based on biological process, molecular function, and cellular compartment ( Figure 4B-D) further underscored the significance of the overlapping gene expression profiles in ifo-1(kc2) and nhr-8(hd117). Both mutant loci affected innate immune response, oxidative-reductive processes, and microbial defense, all of which have been implicated in the reduced stress resilience of ifo-1 mutants [4] and appear to be co-regulated. Additionally, altered metabolic processes ranked second in the affected biological function of both mutant backgrounds.

Lipidomics Reveal Common Features of ifo-1, ifb-2 and nhr-8 Mutants
Based on the observed alterations in gene transcription, we asked whether lipid composition in ifo-1(kc2) and ifb-2(kc14) was affected and how they compared to the lipid composition in nhr-8(hd117). A fatty acid methyl ester (FAME) analysis of fatty acid composition showed similar though not identical changes of saturated and unsaturated fatty acids in ifo-1(kc2) and nhr-8(hd117) ( Figure 5A). A comparison of the percentage of saturated fatty acids (SFAs), monounsaturated fatty acids (MUFAs), polyunsaturated fatty acids (PUFAs), and branched fatty acids showed an increase in SFAs in all mutant backgrounds ( Figure 5B). A significant decrease in MUFAs was observed in ifo-1(kc2) and ifb-2(kc14), and only a minor decrease was observed in nhr-8(hd117). This reduction was in line with the observed downregulation of fat-5 and fat-7 mRNA expression in ifo-(kc2) ( Figure 4E). A comparable downregulation, together with a significant reduction in MUFAs, however, has been reported for nhr-8(hd117) grown at a low cholesterol level [14]. PUFAs were decreased in ifo-1(kc2) and in nhr-8(hd117) but not in ifb-2(kc14). The role of PUFAs for lipid delivery to oocytes was recently shown by coherent anti-Stokes Raman scattering (CARS) microscopy in live C. elegans [20]. A low amount of branched fatty acids was observed but was unchanged in all genetic backgrounds.
Biosynthetic intermediates of the cholesterol-derived steroid hormone dafachronic acid [21] were also quantified. We observed a reduction in 7-dehydrocholesterol that was more pronounced in ifo-1(kc2) than in ifb-2(kc14) but not in ∆ 4 -and ∆ 7 -dafachronic acid ( Figure 5C). The reduced amount of 7-dehydrocholesterol was in line with the previously reported downregulation of the Rieske oxygenase-encoding daf-36 in nhr-8(hd117) [14]. This likely occurs through mechanisms independent of daf-36 transcription, since we did not see any changes in daf-36 mRNA levels in in ifo-1(kc2) or ifb-2(kc14) mutants.
Gene ontology enrichment based on biological process, molecular function, and cellular compartment ( Figure 4B-D) further underscored the significance of the overlapping gene expression profiles in ifo-1(kc2) and nhr-8(hd117). Both mutant loci affected innate immune response, oxidativereductive processes, and microbial defense, all of which have been implicated in the reduced stress resilience of ifo-1 mutants [4] and appear to be co-regulated. Additionally, altered metabolic processes ranked second in the affected biological function of both mutant backgrounds.

nhr-8(RNAi) Mimics the Effect of Cholesterol Depletion in ifo-1(kc2) and ifb-2(kc14)
To examine the genetic interactions between nhr-8 and ifo-1/ifb-2, nhr-8 activity was downregulated by RNAi in ifo-1(kc2) and ifb-2(kc14) ( Figure 6A,A'), with normal levels of cholesterol in the medium. We observed that after 63 h, nhr-8(RNAi) enhanced the weak ifb-2(kc14) phenotype but did not further enhance the already strong ifo-1(kc2) phenotype. Some worms were recovered from larval arrest after 72 h, presumably because animals were able to transport sufficient cholesterol through other pathways to grow with time or because of a subpar RNAi effect. The observations are consistent with the idea that nhr-8 works in overlapping pathways with ifo-1 and ifb-2.
We next asked whether these phenotypes could be modulated by cholesterol deprivation ( Figure 6B,B'). As previously seen, nhr-8(hd117) animals halted development in early larval stages when grown in the absence of cholesterol [14]. Cholesterol depletion had no further effect on ifo-1(kc2) but enhanced ifb-2(kc14), similar to nhr-8(RNAi) above. Notably, ifo-1(kc2) animals displayed L1/L2 arrest with and without cholesterol at 63 h ( Figure 6B). At the later time point of 72 h, however, ifo-1(kc2) partly resumed development-more in the presence than in the absence of cholesterol ( Figure 6B'). Residual traces of cholesterol in the used bacteria suspension may explain why some ifo-1(kc2) animals managed to escape larval arrest.
ifb-2(kc14) animals-which, in contrast, did not show any developmental defects under control conditions-displayed a comparable amount of larval arrest under both nhr-8(RNAi) and cholesterol deprivation, indicating a more subtle yet significant influence on nhr-8-regulated cholesterol handling that was not observed in the experiments shown in Figure 2E, possibly due to the use of normal growth medium (NGM) plates, instead of agarose plates, that were used for the experiments depicted in Figure 6. The obvious difference between ifb-2(kc14) and ifo-1(kc2) was most likely due to the more profound morphological changes and the presence of cytoplasmic and junctional intermediate filament polypeptide-containing aggregates in the ifo-1(kc2) background.
It should be noted that the comparable amount of larval arrest between nhr-8(hd117) -cholesterol and ifo-1(kc2) + cholesterol supports the hypothesis of reduced nuclear hormone receptor (NHR)-8 activity in ifo-1(kc2). Our findings strongly suggest a regulatory role of the IFO-1-controlled intermediate filament network upstream of NHR-8.

Discussion
Our report identifies IFO-1 and the intermediate filament system as regulators of cholesterol and lipid metabolism. It adds to observations concerning mammalian vimentin, which is typically produced in non-epithelial mesenchymal cells and has been linked to lipid metabolism presumably by altering intracellular trafficking and organelle positioning [24][25][26]. In the present study, we observed an increased sensitivity to cholesterol depletion that was caused by reduced cholesterol uptake and transfer. Given the many open questions concerning intestinal cholesterol metabolism in vertebrates (review in [27,28]), and the limited observations in C. elegans [13,14,29], multiple scenarios can be envisioned (Figure 8). As a first step, cholesterol must be taken up from the intestinal lumen into the intestinal cell. In contrast to previous belief, it is not a simple, energy-independent diffusion process; instead, it is one that involves protein components such as Niemann-Pick Type C1-like protein 1 (NPC1L1), which is known to mediate cholesterol absorption [30]. Transport protein candidates in C. elegans are ChUP-1 [16] and NCR-1/NCR-2, which are related to NPC1L1 [17]. may the P-type ATPase TAT-2, whose mRNA is increased in ifo-1(kc2) (this study) and which has been implicated in sterol metabolism, may also be relevant [19]. Though the likely apical plasma membrane localization of these polypeptides remains to be shown, it is of interest to note that intermediate filaments contribute to the apical targeting of membrane proteins [31][32][33][34]. A simple hypothesis is that the loss of the apical intermediate filament network in ifb-2(kc14) and ifo-1(kc2) [3,6] leads to the inefficient localization of the above candidate transporters and/or other hitherto unidentified transporters. The observation that cholesterol transfer is less affected in ncr-1(nr2022); ncr-2(nr2023) in comparison to ifo-1(kc2) shows that the latter must involve other regulators of cholesterol metabolism. Similarly, the less pronounced inhibition of cholesterol uptake in chup-1(gk245) versus ifo-1(kc2) argues for the involvement of additional transporters. Furthermore, the more pronounced phenotype on cholesterol uptake of ifo-1(kc2) versus ifb-2(kc14) may be attributed to the accumulation of extensive intermediate filament-containing junctional aggregates that are uniquely seen in ifo-1(kc2) or to loss of IFO-1 functions that are independent of the intermediate filament system.  The overlap in altered gene transcription of ifo-1(kc2), which is predominantly localized in the apical cytoplasm [6], and nhr-8(hd117), which localizes to both the cytoplasm and the nucleus [14], suggests that NHR-8 is a downstream target of pathways regulated by IFO-1. A consequence of this linkage are shared changes in lipid metabolism. Of interest is the reduction of PUFAs, which have been implicated in the delivery of lipids to the gonads [20]. The desaturases FAT-5 and FAT-7 may As a next step, cholesterol is transported through the cell. The scheme in Figure 8 highlights two aspects that are relevant to the present findings, i.e., the activation of the sterol sensor NHR-8 and additional mechanisms, that together lead to the formation of yolk granules that are released into the pseudocoelom for the nourishment of oocytes and developing embryos. We postulate that the reduced cholesterol uptake in ifo-1(kc2) and ifb-2(kc14) leads to the reduced activation of NHR-8. Several lines of evidence support this hypothesis.
(I) The overlap of transcriptional dysregulation between nhr-8(hd117) and ifo-1(kc2) (Figure 4) is quite remarkable considering that the compared microarray data stemmed from independent sources with different experimental setups. The analyses were done at different developmental stages (L3 versus L4), in different tissues (complete animals versus dissected intestines), and in different growth conditions (low versus normal cholesterol). This may account for the comparatively large number of differentially expressed, non-overlapping genes in the different genetic backgrounds (396 versus 895 genes). We cannot exclude that this also applies to some of the jointly regulated genes but would like to emphasize the striking similarities in concordant up-and down-regulated gene expression (164 of 169 genes).
(II) ifo-1(kc2) animals display striking phenotypic similarities with nhr-8(hd117) animals under low cholesterol conditions, notably a reduced uptake of cholesterol, an impaired supplementation of embryos with maternal cholesterol, and the accumulation of saturated fatty acids accompanied by a loss of mono-and poly-unsaturated fatty acids (Figures 3 and 5; [14]).
(III) The knockdown of nhr-8 by RNAi has the same consequence as cholesterol deprivation in ifo-1(kc2) animals, leading to early larval rest in the F1 generation ( Figure 6).
The overlap in altered gene transcription of ifo-1(kc2), which is predominantly localized in the apical cytoplasm [6], and nhr-8(hd117), which localizes to both the cytoplasm and the nucleus [14], suggests that NHR-8 is a downstream target of pathways regulated by IFO-1. A consequence of this linkage are shared changes in lipid metabolism. Of interest is the reduction of PUFAs, which have been implicated in the delivery of lipids to the gonads [20]. The desaturases FAT-5 and FAT-7 may play a central role, since the transcription of the corresponding genes is considerably reduced in ifo-1(kc2) and nhr-8(hd117). The resulting reduction in MUFAs is predicted to decrease production of PUFAs, which is catalyzed by FAT-1-4, as well as ELO-1 and ELO-2 [35]. Additionally of interest in this context is the reported reduction in overall intestinal lipid observed in fat-6/fat-7 double mutants [18], which is in line with the reduced Nile red staining described in this report. Additionally, the knock down of the C. elegans sterol-regulatory-element-binding protein (SREBP) homolog SBP-1 also leads to the impaired expression of fat-5 and fat-7, which is accompanied by growth defects, reduced fat content, and the accumulation of saturated fatty acids [36]. It remains to be shown whether low cholesterol availability in ifo-1(kc2)/ifb-2(kc14) affects SBP-1 activation.
The transcriptional dysregulation of the ceramide glucosyltransferases CGT-1 and CGT-2 is also shared by ifo-1(kc2) and nhr-8(hd117). The much more pronounced downregulation of cgt-2 in ifo-1(kc2) versus nhr-8(hd117) ( Figure 4E) may explain the resulting decrease in d17iso-GlcCer in ifo-1(kc2) but not in nhr-8(hd117) ( Figure 5D). The impaired synthesis of glycosphingolipids could be responsible for the growth defects in ifo-1(kc2) animals in accordance with previous reports, which have shown that the removal of CGT activity induces growth arrest [37,38]. In contrast to previous reports, which have ascribed important functions to CGT-1 and CGT-3, the present study additionally suggests the importance of CGT-2.
The upregulation of bre-3 mRNA in ifo-1(kc2) and nhr-8(hd117) implicates another mechanism whereby lipid metabolism may be affected. The depletion of BRE-3 has been shown to increase intracellular cholesterol mobilization by increasing phosphoethanolamine glucosylceramides (PEGCs) [22]. Conceivably, the observed upregulation of bre-3 mRNA might be responsible for the impaired mobilization of the intracellular cholesterol pool.
The more severe phenotype of ifo-1(kc2) than that of nhr-8(hd117) and, conversely, the additive effect of nhr-8(RNAi) in ifo-1(kc2) suggest that NHR-8 is certainly not the only effector of IFO-1 and that, on the other hand, IFO-1-independent functions of NHR-8 exist. A major conundrum in this context is the downregulation of the C. elegans apolipoprotein genes vit-1 and vit-2 in nhr-8(hd117), which is not observed in ifo-1(kc2) animals. Additionally, the transcription of the daf-36 mRNA is reduced in nhr-8(hd117) but not in ifo-1(kc2). daf-36 encodes the Rieske oxygenase DAF-36, which catalyzes the reaction of cholesterol to 7-dehydrocholesterol. However, levels of 7-dehydrocholesterol are reduced not only in nhr-8(hd117) but also, and even more pronounced, in ifo-1(kc2). A possible explanation for this paradox is the reduced availability of the DAF-36 substrate cholesterol in ifo-1(kc2) (Figure 3). It is unclear, however, why the levels of ∆ 4 -and ∆ 7 -dafachronic acids, which are derived from 7-dehydrocholesterol, are unchanged in ifo-1(kc2), since nhr-8 mutants show lower levels of these compounds under cholesterol depletion, which has been linked to dauer formation. It is possible that feedback mechanisms for the production of dafachronic acid can compensate [39,40].
Aside from their similarities, NHR-8 and IFO-1/IFB-2 also show differences. NHR-8-independent targets of IFO-1/IFB-2 may include components of the cytoplasmic lipid transport system. The mammalian vimentin intermediate filament cytoskeleton has been shown to interact with endosomal/lysosomal sorting machinery, thereby affecting the lysosomal exit of LDL (low-density lipoprotein)-derived cholesterol [41,42]. Moreover, the involvement of the NPC proteins in this process [43][44][45] raises the question of whether the localization or function of the C. elegans homologs NCR-1 and NCR-2 are affected in the ifo-1/ifb-2 mutant backgrounds.
Taken together, we conclude that the disruption of the apical intermediate filament network impairs the uptake of cholesterol, thus resulting in the dysregulation of cholesterol-dependent cellular processes and physiology that are, in part, mediated through NHR-8. A striking example of the physiological consequences of impaired cholesterol uptake was recently reported for mutants of chup-1, which was shown to be involved in the cholesterol-and NHR-8-dependent innate immune response to microbes [46]. The clarification of cholesterol-dependent molecular mechanisms connecting organismal growth, reproduction, and stress response with cytoskeletal organization, sterol sensing, and trafficking remains a great challenge. The cholesterol-dependence of C. elegans gut physiology offers an experimentally tractable scenario to address this.

Nile Red Staining
The lipophilic dye Nile red, which becomes fluorescent in lipid-rich environments [47], was obtained from Fluka (Thermo Fisher Scientific Inc., Waltham, MA, USA). It was diluted to 0.5 ng/mL in water and was added to the bacterial lawn on agar plates followed by incubation overnight in the dark at room temperature. Plates were inoculated with L4 larvae and kept for 22 h at 20 • C. Worms were then transferred to an agarose pad and immobilized in 4 µL 1% levamisole (Sigma-Aldrich, St. Louis, MO, USA). Nile red fluorescence was assessed by microscopic imaging in a Zeiss Apotome at 10x magnification (Carl Zeiss AG, Oberkochen, Germany). Fluorescence was then quantified with the help of Fiji to manually define regions of interest and determine the grey values within.

Electron Microscopy
Electron microscopy was done on high-pressure frozen worms as described previously [3].

Cholesterol Depletion Assay
For cholesterol deprivation experiments, 10 cm agarose plates (1.75% (w/v) agarose, 52 mM NaCl, 25 mM KH 2 PO 4 /K 2 HPO 4 [pH 6.0], 1 mM MgSO 4 , and 0.5 mM CaCl 2 ) were prepared containing either 5 mg/L cholesterol in ethanol (control condition) or an equal amount of pure ethanol. The plates were inoculated with an overnight culture of OP50 E. coli, which was washed two times with the M9 buffer and re-suspended in 1/5 of its volume. Synchronized L1 larvae, which had hatched overnight on NGM plates without food and cholesterol, were transferred onto the test plates and incubated at 20 • C. To prevent overcrowding, a chunk of agar containing F2 larvae (if any) was transferred onto a fresh plate at day 3. Representative pictures were taken at days 4 and 6 using an SMZ1500 stereoscopic zoom microscope equipped with an µEye USB camera and a Plan Apo WD70 1× objective (Nikon Corporation, Tokyo, Japan).
For phenotypic analysis, 6 cm standard NGM plates with and without cholesterol were used. A single embryo of bleached animals was placed on each plate (see above) and incubated at 20 • C. The developmental stage was monitored daily, and gravid adults were transferred to a fresh plate each day to count the total offspring from each animal.
For the measurement of larval arrest, 1 mL of agarose (1.75% (w/v) agarose, 52 mM NaCl) with and without cholesterol, was poured into 24-well plates and inoculated with an overnight culture of OP50 E. coli. The experiment was started by placing several embryos of bleached animals into the wells (see above). Early larvae (L1/L2) and total offspring were counted after 63 and 72 h at 20 • C.

Cholesterol Uptake and Transfer Assays
Cholesterol feeding experiments were carried out at 20 • C on 10 cm agarose plates (1.75% (w/v) and 52 mM NaCl), inoculated with 800 µL of an overnight culture of OP50 E. coli, supplemented with 0.5 µCi of [4-14 C]-cholesterol (Perkin-Elmer, Inc., Waltham, MA, USA). Synchronized L1 larvae, which had hatched overnight on NGM-plates without food and cholesterol, were placed on the feeding plates and allowed to grow to adulthood at 20 • C. The fed animals were washed two times with an M9 buffer, and groups of 20 animals each were either transferred to normal NGM plates or collected for lysis. The animals on the NGM plates were allowed to lay eggs for 3 days at 15 • C, after which they were removed and collected for lysis. After 2 days at 20 • C, the F1 generation was flushed off the plates, washed two times with an M9 buffer, and subjected to lysis. Lysis was carried out for 1 h at 65 • C in a single-worm lysis buffer, containing 0.5 µg/µL proteinase K. Afterwards, all samples were mixed with 5 mL of scintillation liquid Irgasafe plus (Perkin-Elmer, Inc., Waltham, MA, USA), and radioactivity was measured in a Hidex 600 SL liquid scintillation counter (Hidex Oy, Turku, Finland).

Cholesterol Rescue Experiment
Cholesterol rescue experiments were carried out on standard NGM-plates supplemented with either standard or 5× cholesterol. Worms were grown from synchronized embryos at 18 • C, and the developmental state was monitored daily.

Gene Ontology Analysis
Transcriptomic data for nhr-8(hd117) were previously published in [14], and transcriptomic data for ifo-1(kc2) were published in [4]. Both lists of regulated genes were subjected to gene ontology analysis using DAVID (database for annotation, visualization and integrated discovery) v6.8 using default settings [49].
Lipid extraction from C. elegans: Worms were collected at young adult stage, and homogenates were obtained by beads beating for 30 min at 4 • C using a Qiagen TissueLyser (Qiagen, Venlo, The Netherlands); 0.1% (w/v) of butylated hydroxytoluene (BHT) was added to prevent autoxidation reactions. Protein concentrations were measured using the BCA assay kit (Thermo Fisher Scientific Inc., Waltham, MA, USA) prior to lipid analyses. A volume of homogenate corresponding to 600 µg of protein was used for ceramide analysis, a volume of homogenate corresponding to 200 µg of protein was used for fatty acid analysis (FAME), and a volume of homogenate corresponding to 200 µg of protein underwent solid phase extraction (SPE) for steroid enrichment (see corresponding sections).
Fatty acid analysis (FAME): GC-MS analyses were carried out on a GC/MS (Triple Quadrupole -GC 7890A + 7000 QQQ; Agilent Technologies, Santa Clara, CA, USA) interfaced with a robotic auto sampler PAL 1 system (PAL LHXxt). One microliter of the derivatized worm sample was injected into the liner (4 mm splitless, single taper liner with glass wool) by the pulsed-split mode at 200 • C. Analytes were separated using a DB Ultra Inert column (15,250, and 0.25 mm; Agilent Technologies, Santa Clara, CA, USA). The initial oven temperature was set to 180 • C and held for 1 min, and then the temperature was increased by 30 • C/min up to 300 • C and held at that temperature for 1 min. At the end, the temperature was increased up to 325 • C by 30 • C/min. The temperature of the transfer line was set to 325 • C, and helium was used as the carrier gas at a flow rate of 1.4 mL/min in a constant flow mode. Ions were generated using an electron impact ion source (70 eV and 300 • C) and analyzed by a triple quadrupole. C13:0 fatty acid was used as internal standard. Data were analyzed using MassHunter Workstation Software, Qualitative Analysis, Version B.06.00 (Agilent Technologies, Santa Clara, CA, USA). The relative response was calculated by dividing the peak area of the analyte to the internal standard peak area and further normalized to protein concentration.
Ceramide targeted analysis: Ceramides were extracted by adding a 2:1 (v/v) mixture of methanol and chloroform and incubated for 1 h at 4 • C. Samples were then centrifuged for 5 min at 3000 rcf at 4 • C, and the liquid phase was transferred and dried in a speed vac. Samples were reconstituted with 60:40 (v/v) methanol:water, and 5 µL were injected into the LC-MS system. The identification of ceramides and glucosylated ceramides was performed on a Q Exactive Mass Spectrometer (Thermo Fisher Scientific Inc., Waltham, MA, USA) coupled with a binary pump system (Vanquish, Thermo Fisher Scientific Inc., Waltham, MA, USA). Lipid species were separated using an RP column (Xbridge BEH C8 2.5 µm, 2.1 × 100 mm) using solvents A (60:40 v/v acetonitrile: water with 10 mmol/L ammonium formate) and B (90:10 v/v isopropanol:water with 10 mmol/L ammonium formate), as previously reported [50]. The gradient started from 32% B then ramped to 45% B in 4 min. In 1 min, the gradient was ramped till 52% B, and in 3 min, it was ramped till 58% B. Then, in 3 min, it increased up to 66% B before being ramped in 3 min till 70% B. In the following 4 min, it went till 75% B, and then it ramped in 3 min to 97% B and held for 4 min; finally, in 1 min, it went back to 32% B and held for 4 min. The total time was 30 min, and the column was heated at 45 • C using a flow rate of 100 µL/min. The LC system was flushed in between runs with 75/25 (v/v) isopropanol water with 0.1% formic acid.
Ceramides and glycosylated ceramides were detected in negative ion mode using an H-ESI ion source with the following parameters: sheath gas of 10, auxiliary gas of 2, spray voltage of 3 kV, and ion transfer tube temperature of 250 • C. Data were acquired using a targeted-selected ion monitoring chromatogram mode (t-SIM) with a resolution of 70.000, AGC 5 e 4 , a maximum injection time of 200 ms, and an isolation of 1.0 m/z. C4 Ceramide (d18:1/4:0) and C12 GluCer d18:1/12:0 were used as internal standards. Data were analyzed using Xcalibur version 4.0 and Trace Finder version 4.1 (Thermo Fisher Scientific Inc., Waltham, MA, USA). The relative response for each steroid species was calculated by dividing the peak area of the analyte to the internal standard peak area, and the result was further normalized to protein concentration.
Steroid analysis: HLB prime cartridges (Oasis, Waters, Eschborn, Germany) were pre conditioned with 1 mL of methanol and 1 mL of water with 1% (v/v) formic acid before adding the sample. The homogenate was directly applied to the column, and phospholipids were removed by adding 1 mL of hexane. Steroids were eluted using 500 µL of ethyl acetate, as previously described [51]. The organic fraction was dried in a speed vac. All samples were reconstituted in 20 µL of 50% (v/v) methanol/water, and 5 µL were injected into the LC-MS/MS system. LC-MS analysis was carried out on a triple quadrupole mass spectrometer (QqQMS) (TSQ Altis; Thermo Fisher Scientific Inc., Waltham, MA, USA). Prior to the measurements, gradient optimization was performed using authentic standards. Steroids were separated with a reverse phase column (XSelect HSS T3 2.5 µm 2.1 × 100; Waters, Eschborn, Germany) using a binary pump system (Vanquish, Thermo Fisher Scientific Inc., Waltham, MA, USA) with solvent A as water with 0.1% (v/v) formic acid and eluent B as acetonitrile with 0.1% (v/v) formic acid; 2 µL of the reconstituted sample were injected. The flow rate was 0.1 mL/min for the first 5 min and then increased to 0.150 mL/min thereafter. The column temperature was set at 30 • C. The ESI ionization parameters were as follows: 3.5 kV, 25 a.u. sheath gas, 5 a.u. auxiliary, and 350 • C transfer ion capillary. All the spectra were acquired in positive ion mode.
For the internal standard, 7 ketocholesterol (d7) was used. Data were analyzed using Xcalibur version 4.0 and Trace Finder version 4.1. The relative response for each steroid species was calculated by dividing the peak area of the analyte to the internal standard peak area, and the result was further normalized to protein concentration.