P311 functions in an alternative pathway of lipid accumulation that is induced by retinoic acid

Lipid droplets, also known as lipid bodies, are intracellular inclusions that generally consist of a core of neutral lipids, predominantly triglycerides and steryl esters, surrounded by a monolayer of phospholipids and proteins (Martin and Parton, 2005). Although typically associated with the phenotype and functions of adipocytes and steroidogenic cells, lipid droplets are also found in many other cell types and typically form in response to elevated fatty-acid levels (Pol et al., 2004). The size, composition, and function of these lipid droplets may vary, however, according to cell type or metabolic state (Martin and Parton, 2006). Initially thought of as inert intracellular lipid stores, these cytosolic lipid droplets have recently been shown to be dynamic, motile organelles that participate in a variety of cellular functions. Lipid droplets regulate the storage and turnover of neutral lipids and cholesterol esters that can be used for energy metabolism, membrane synthesis, steroid synthesis and synthesis of lipid mediators. Defects in intracellular lipid homeostasis are found in many diseases, such as obesity, insulin resistance, diabetes, atherosclerosis and lipid-storage diseases. Although functionally linked to vital cellular processes, the biology of the lipid droplet is still poorly understood, including potential differences in the factors that stimulate their formation in alternative cellular contexts.

P311 (also known as chromosome 5 open reading frame 13, C5orf13, D0H4S114) is a gene originally discovered as an abundant transcript at sites of active embryonic and postnatal neurogenesis (Studler et al., 1993) and has been subsequently shown to induce neurite extension in cultured neurons and nerve regeneration in vivo (Fujitani et al., 2004). It encodes an 8 kDa, 68 aa protein with no structural motifs suggestive of its molecular functions. It has also been identified in other cell types, such as glioblastoma cells (Mariani et al., 2001) and chondrocytes (Sironen et al., 2000), and has been shown to induce myofibroblast differentiation in vitro (Pan et al., 2002). We identified P311 as a gene substantially upregulated during lung alveolarization, a stage of distal lung development that generates the respiratory gas exchange surface (Zhao et al., 2006). This stage of lung development is associated with an abundance of lipid-containing fibroblasts (McGowan and Torday, 1997). Interestingly, the known expression pattern and function of P311 in the neuronal system and the distal-lung parallel that of retinoic acid – a powerful hormone well-known for its ability to stimulate regeneration in the nervous system (Maden and Hind, 2003) as well as the distal-lung structure (Maden and Hind, 2003; Maden and Hind, 2004). In this study we sought to further investigate the cellular and molecular functions of P311. We report here that P311 upregulates classes of genes associated with lipid-uptake and synthesis, increases cellular levels of cholesterol and triglycerides, and increases intracellular lipid droplets. Interestingly, we find that P311 specifically mediates lipid-droplet accumulation that is induced by retinoic acid in lung fibroblasts. Our data thus demonstrate a novel function for P311 in lipid accumulation and implicate its functions in retinoic-acid regulation of lipid-droplet biogenesis.

To explore the potential cellular functions of P311, we stably expressed P311 in C3H10T1/2 mouse embryonic fibroblasts under the tetracycline-regulated conditional expression system. Semi-quantitative reverse transcriptase (RT)-PCR confirmed the induction of P311 transcript levels in cells stably expressing P311 upon doxycycline administration (Fig. 1A). In addition, a low basal level of P311 transcripts without exposure to doxycycline was observed (Fig. 1A). These basal levels were sufficient to cause a morphological change in the cells stably expressing P311 when seeded at a higher density (10⁶ cells per 100-mm plate) (Fig. 1B-E). These changes included cell enlargement and the presence of a dark perinuclear ring. To determine any changes in gene expression that might be responsible for this alteration in cell morphology induced by the expression of P311, we examined the global transcript expression profile of over 25,000 genes in C3H10 cells stably expressing P311 seeded at the higher cell density compared with C3H10 control stable cells. Microarray analysis revealed 579 genes that were statistically significantly upregulated at least twofold in cells stably expressing P311. To determine the functional significance of these upregulated genes, we used Ensemble attribute profile clustering data (Semeiks et al., 2006) to functionally classify them. The analysis revealed that primarily genes associated with the synthesis and metabolism of lipid and cholesterol were significantly upregulated when P311 was overexpressed in the cells. Specifically, these genes encode proteins that function in multiple aspects of lipid physiology, including the transport of fatty acids, the synthesis and metabolism of sterols and steroid molecules, and the accumulation of cholesterol esters (Table 1).

Since the microarray data suggest a role for P311 in lipid metabolism, we wished to determine whether lipid synthesis and accumulation are altered in cells stably expressing P311. We added radio-labeled fatty acids to control cells and those stably expressing P311, and assayed for the amount incorporated into cellular triglycerides. We found that there was an approximately threefold higher level of labeled triglycerides in cells that stably expressed P311 relative to control cells (Fig. 2A). Total cellular cholesterol levels were also elevated approximately threefold in cells stably expressing P311 following fatty-acid administration (Fig. 2B). In the process of performing these biochemical assays we observed the visible accumulation of lipid droplets following fatty-acid supplementation in cells that stably expressed P311. To better visualize these lipid droplets, we stained C3H10 cells stably expressing P311 and control cells following the addition of fatty acids with Oil-Red-O. In contrast to control cells, cells that stably expressed P311 exhibited strong Oil-Red-O staining (Fig. 2C,D), indicating a substantially increased ability to synthesize lipid droplets. These lipid droplets exhibited some perinuclear clustering but also localized throughout the cytoplasm. Interestingly, these droplets did not associate with adipophilin or TIP47, marker proteins typically associated with lipid droplets in adipocytes (Fig. 2E-L), and we did not observe changes in the mRNA levels of these markers (data not shown). These data indicate that the cellular expression of P311 leads to an increased ability to synthesize lipids and accumulate lipid droplets.

Since the cellular expression of P311 induced lipid synthesis and accumulation we asked whether P311 has a function in adipogenesis. We assayed for changes in P311 gene expression during the adipogenic induction of NIH3T3-L1 cells, an in vitro model of adipocyte differentiation. We observed a substantial initial decrease followed by a gradual increase in P311 transcript expression (Fig. 3A). To determine whether P311 has a role in this process, lentiviruses expressing short hairpin RNAs (shRNAs) targeting the P311 transcript were used to silence P311 transcript expression (Ventura et al., 2004). Quantification by real-time RT-PCRof P311 expression in NIH3T3-L1 cells infected with control and P311-shRNA-expressing lentiviruses demonstrated a substantial decrease in P311 transcript expression following the infection with of the P311 shRNA (Fig. 3B). Infected cells were subsequently induced for adipogenic differentiation (see Materials and Methods). Accumulation of large lipid droplets in almost all control- and P311-shRNA-lentiviral-infected cells was observed (Fig. 3C-F), and quantification of Nile-Red-stained cells followed by flow cytometry demonstrated no significant decrease in intracellular lipids in P311-shRNA-lentivirus-infected cells relative to control-lentivirus-infected cells (Fig. 3G). These results indicate that P311 gene silencing has no significant effects on adipogenic differentiation and/or lipogenesis in this system. Thus, typical levels of P311 expression are not required for lipid accumulation in this in-vitro model of adipogenesis.

Since P311 did not have a significant role in a conventional model of adipogenesis, yet causes increased cellular lipid uptake and accumulation when expressed in C3H10 cells, we sought to determine the identity and significance of signaling pathways that regulate P311 expression and P311 functions. We examined whether fatty acids regulate P311 mRNA expression, because fatty acids are known to stimulate lipogenesis through the peroxisome proliferators-activated receptor (PPAR) pathway (Green, 1995). In addition, we speculated that retinoic acid also modulates P311 mRNA expression, because there is evidence suggesting that P311 is associated with the retinoic-acid signaling pathway: (1) P311 expression is high in brain regions that have extensive retinoic-acid signaling (Studler et al., 1993), (2) P311 promotes neuronal regeneration – an effect also attributed to retinoic acid (Maden and Hind, 2003) and, (3) retinoic-acid signaling is markedly increased during lung alveolarization (Chytil, 1996), a period when P311 expression is substantially increased (Zhao et al., 2006). We therefore isolated lung fibroblasts from 1-week-old mice, because both lipid accumulation and retinoid signaling are markedly increased in this cell type during the early postnatal period (Chen et al., 1998; Kaplan et al., 1985; Liu et al., 1993; Vaccaro and Brody, 1978) and examined the influence of fatty-acid and retinoic-acid treatment on P311 expression in these cells. Treatment of lung fibroblasts from 1-week-old mice with fatty acids alone did not induce P311 mRNA expression, as assayed by real-time quantitative RT-PCR (Fig. 4A). retinoic acid, however, induced P311 mRNA expression ∼2.7-fold after one week, and the addition of fatty acids did not further affect the level of P311 induction by retinoic acid (Fig. 4A). The induction of P311 expression by retinoic acid occurs by 48 hours and is dose-dependent (Fig. 4B). These data demonstrate that P311 expression is regulated by retinoic acid.

The induction of P311 expression by retinoic acid in lung fibroblasts and the ability of P311 to increase lipid-droplet accumulation in C3H10 cells led us to hypothesize that retinoic acid stimulates lipid-droplet accumulation in lung fibroblasts. This would be consistent with the presence of active retinoic acid signaling and intracellular lipid accumulation in the lung mesenchyme during alveolarization. We therefore determined whether retinoic acid induces intracellular lipid accumulation in lung fibroblasts isolated from 1-week-old mice. We found that a large portion of freshly isolated lung fibroblasts at this stage contain lipid droplets (data not shown), consistent with previous observations of high amounts of lipid-containing fibroblasts in the lung during alveolarization. However, maintenance of these lung fibroblasts for at least 1 week in culture without fatty-acid supplementation depleted most cells of their lipids (data not shown). Re-supplementation of fatty acids to these cells resulted in the accumulation of intracellular lipids as assayed by Nile-Red staining followed by flow cytometry (Fig. 5), suggesting increased levels of fatty-acid metabolism in lung fibroblasts at the alveolarization stage of lung development (Abumrad et al., 1981; Abumrad et al., 1984; Dutta-Roy, 2000). Treatment of the cells with retinoic acid caused a further accumulation of intracellular lipids (Fig. 5), demonstrating that retinoic acid stimulates lipid accumulation in these lung fibroblasts.

We next determined whether P311 expression is required for retinoic-acid-mediated lipid accumulation in lung fibroblasts. We silenced P311 expression by using a lentiviral-mediated shRNA approach. Quantification of P311 expression by real-time RT-PCR in lung fibroblasts that had been infected with either control- or P311-shRNA-expressing lentiviruses demonstrated a significant knockdown of P311 transcripts in P311-shRNA-lentivirus-infected cells (Fig. 6A). Control- or P311-shRNA-lentivirus-infected cells were then treated with retinoic acid in the presence of fatty acids, and intracellular lipid accumulation was assayed by Nile-Red staining. A higher fatty-acid concentration was used in these experiments because it decreases the time to observe lipid-droplet accumulation induced by retinoic acid. P311-shRNA-lentivirus-infected lung fibroblasts exposed to retinoic acid and fatty acids accumulated substantially less lipid droplets compared with control-lentivirus-infected cells (Fig. 6B). Quantification by flow cytometry indicated an ∼85% decrease in intracellular lipids in P311-shRNA-lentivirus-infected cells relative to control-lentivirus-infected cells (Fig. 6C). The incomplete inhibition of lipid accumulation in cells infected with P311 shRNA lentivirus might be due to the incomplete knockdown of P311 expression. Overall, these data demonstrate that the knockdown of P311 expression has a significant effect on the induction of lipid accumulation by retinoic acid in lung fibroblasts in vitro and suggest that P311 mediates the effects of retinoic acid in this process.

P311 has been previously identified to be differentially regulated in many different cellular contexts, and has been implicated in diverse cellular functions, including cell proliferation, migration and differentiation. However, the molecular function of P311 is still unknown. Here, we identified a novel function for P311, demonstrating its ability to upregulate genes that are associated with lipid synthesis and metabolism, increase cellular levels of triglycerides and cholesterol, and induce lipid-droplet accumulation. Since lipid droplets regulate many cellular processes, this function of P311 might contribute to the pleiotropic effects of P311 expression. Interestingly, P311 is not required for lipid accumulation during adipogenic differentiation but, instead, has an important role in the induction of lipid-droplet accumulation mediated by retinoic acid. Our data thus identify the presence of distinct pathways in the accumulation of lipid droplets and places P311 within the context of cellular functions mediated by retinoic acid.

Our laboratory has previously identified P311 as a gene that potentially regulates alveolarization, because the expression of P311 is markedly increased in the mouse lung during alveolar development and is decreased in mice with impaired distal-lung development (Zhao et al., 2006). In this context, it is significant that P311 induces cellular lipid-droplet accumulation and that it mediates this effect of retinoic acid in lung fibroblasts, because during alveolarization there is both a high abundance of lipid-containing interstitial cells (termed lipofibroblasts) (McGowan, 2002; McGowan and Torday, 1997) as well as substantial retinoic-acid signaling in the lungs (Chen et al., 1998; Kaplan et al., 1985; Liu et al., 1993; Vaccaro and Brody, 1978). The function of P311 in alveolar development might be, thus, to trigger retinoic-acid-dependent induction of lipofibroblast formation. However, P311 has also been reported to direct the transition of fibroblasts to myofibroblasts (Pan et al., 2002), the latter of which are smooth-muscle α-actin-containing cells located at sites that require an active contractile function, including at the tip of the alveolar septa (Vaccaro and Brody, 1978). Our cellular expression of P311 under the tetracycline-conditional system does not seem to induce a myofibroblast phenotype, because we found no change in smooth-muscle α-actin expression in our cells that stably expressed P311 (our unpublished observations). This difference in cellular response is possibly due to differences in P311 expression levels, because tetracycline-induced conditional expression is relatively weaker compared with that of cytomegalovirus-promoter-induced expression used in a previous study (Pan et al., 2002). In addition, our observation of lipid accumulation in P311-stable cells is under conditions of higher cell confluency, where there might be differences in cell state and signaling.

It is also intriguing to consider the potential relationship between the lipid-containing lipofibroblasts and the smooth-muscle α-actin-containing myofibroblasts. These two cell types are located in close proximity during the development of lung alveoli – myofibroblasts at the tips of newly formed septa and lipofibroblasts in close proximity to the base of the septa (Vaccaro and Brody, 1978). Lipofibroblasts also share characteristics with the myofibroblasts, including the synthesis of extracellular matrix structural components, such as collagen and elastin (McGowan and Torday, 1997). In the liver, hepatic stellate cells are known to exist in a quiescent lipid-containing state until their development to myofibroblasts (Sato et al., 2003). Thus, lipid-containing cells stably expressing P311 might represent an intermediate cell-type in the pathway leading to myofibroblast development.

Retinoic acid is a potent hormone that is essential for the morphogenesis and function of many tissues. It has pleiotropic cellular effects, including the regulation of cell proliferation, differentiation and survival. Previous data indicate that retinoic acid can also significantly affect lipid accumulation, depending on tissue and cell-type. Retinoic acid increases lipid accumulation in human macrophages (Inazawa et al., 2003), mouse Ob1771 preadipocytes (Safonova et al., 1994), and murine embryonic stem cells (Dani et al., 1997). However, retinoic acid inhibits lipid accumulation in NIH3T3-L1 preadipocytes (Schwarz et al., 1997). We found that retinoic acid significantly increases the accumulation of lipids in lung fibroblasts. In lung, retinoic acid is important for alveolar development, because defects in retinoic-acid signaling are associated with defects in distal-lung formation (McGowan et al., 2000; Snyder et al., 2005; Yang et al., 2003). Retinoic acid can also stimulate alveolar regeneration. Retinoic-acid treatment of adult rodents that have emphysema, induced by dexamethasone or elastase, causes the formation of new alveoli (Hind and Maden, 2004; Massaro and Massaro, 1997; McGowan, 2002). The cellular and molecular bases of retinoic acid functions during alveolarization are unknown. Our data implicate the regulation of intracellular lipid-droplet accumulation as a potential function of retinoic acid signaling during lung development and regeneration, and demonstrate that P311 is an important component of this process. Both retinoic acid and P311 have also been shown to induce neurite outgrowths and nerve regeneration (Fujitani et al., 2004; Maden and Hind, 2003). It will be of substantial interest to determine whether the regulation of cellular lipid stores is also an aspect of retinoic acid and P311 functions in the neuronal system.

Lipid droplets are increasingly recognized as important intracellular organelles that regulate diverse cellular functions (Martin and Parton, 2005). Studies of lipid-droplet biogenesis and function have traditionally focused on adipocytes and hepatocytes, which regulate lipid storage in response to whole-body lipid homeostasis (Badman and Flier, 2007; Yan et al., 2007). However, almost all cell types can form lipid droplets, which are essential for the regulation of cellular lipid homeostasis and function. Lipid droplets are the main storage sites of retinyl esters, the precursors of vitamin A, in liver hepatic stellate cells (Wake and Sato, 1993). They are sites of arachidonic acid metabolism – the pathway of eicosanoids and prostanoids synthesis – in leucocytes (Wan et al., 2007) and macrophages (Dvorak et al., 1983), and sources of cholesterol used in steroid-hormone synthesis in steroidogenic cells in the adrenal cortex (Toth et al., 1997), testes and ovaries (Murphy, 2001). In Drosophila embryos lipid droplets are used as depots to store proteins, including histones (Cermelli et al., 2006). Although the regulation of lipid-droplet formation in adipocytes and hepatocytes has been extensively studied, their regulation in other cell types has not been well characterized. Our data identify differences between the regulation of lipid accumulation in adipocytes and that in fibroblasts. Whereas P311 promotes lipid-droplet accumulation in fibroblasts, the silencing of P311 gene expression did not significantly delay or alter the formation of lipid bodies in the accumulation of lipids in NIH3T3-L1 preadipocytes, a well-established model for adipogenesis (Green and Meuth, 1974), suggesting that P311 is not required for this process. This, importantly, implicates mechanistic differences in the accumulation of lipid droplets that depend on stimulus and cell type. Our study identifies for the first time P311 as part of this mechanistic difference, showing that it has a distinct role in lipid-droplet accumulation mediated by retinoic acid.

The C3H10T1/2 fibroblast (American Type Culture Collection, Manassas, Virginia, USA) and PT67 retroviral packaging cell lines (BD Clontech, Mountain View, CA) were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 4.5 g/l glucose with 10% fetal bovine serum (FBS), 100 U/ml penicillin G and 100 μg/ml streptomycin. For routine sub-culture, control and stable cells were seeded at 2.5×10⁵ cells/100 mm. The pREV-TRE:mP311-myc plasmid was constructed as follows. The cDNA of the coding region of mouse P311 was conjugated with a Myc tag through PCR amplification with modified primers (sense 5′-TTTGGATCCGCCGCCACCATGGTTTACTACCCAGAA-3′; antisense 5′-AAAATCGATTTACAGATCCTCTTCTGAGATGAGTTTTTGTTCAAAAGGGTGGAGGTA-3′), digested with BamHI and ClaI, and cloned into the pREV-TRE retroviral tetracycline-responsive element vector (BD Clontech, Mountain View, CA). The recombinant protein produced by this construct is the full-length murine P311 with a Myc tag at the C terminus.

The stable C3H10T1/2 cell lines containing the Tet-On transactivator and TRE:mP311-myc transgenes were generated by either sequential transfection or double infection. To obtain stable cell pools generated by transfection, the cells were transfected first with the pREV-Tet-On plasmid (BD Clontech, Mountain View, CA). Following selection, single stable Tet-On clones were transiently transfected with the TRE-luciferase plasmid (BD Clontech) and induced with 1 μg/ml doxycycline (Sigma, St Louis, MO) to identify those clones with the highest induced:uninduced ratio of luciferase activity. The clone with highest ratio was subsequently transfected with either pREV-TRE or pREV-TRE:mP311-myc, and selection was performed under 200 μg/ml G418 and 200 μg/ml hygromycin (Invitrogen, Carlsbad, CA) for 2 weeks to select for C3H10T1/2 stably expressing both pREV-Tet-On and either pREV-TRE (control cells) or pREV-TRE:mP311-myc (cells stably expressing P311). To obtain stable cell pools generated by double infection, Tet-On and TRE-mP311-myc plasmids were each transfected into the PT67 retroviral packaging cell line. Retroviral-containing medium was harvested 2 days later. Polybrene (Sigma) was added to the medium at a final concentration of 4 μg/ml and filtered through a 0.45 μm cellulose acetate syringe filter (Fisher, Hampton, NH). Retroviral supernatants containing Tet-On and TRE-mP311-myc were used to co-infect C3H10T1/2 cells at 70% confluency. After 2 days cells were passaged at 1:10 dilution and treated with 200 μg/ml G418 and 200 μg/ml hygromycin for 1 month to select for C3H10T1/2 stably expressing both pREV-Tet-On and pREV-TRE:mP311-myc.

Total RNA was extracted from four biological samples of C3H10T1/2 cells stably expressing Tet-On/TRE-REV and C3H10T1/2 cells stably expressing Tet-On/TRE mP311-myc by using the RNeasy mini kit (Qiagen, Valencia, CA). After verifying the quality of the RNA samples on an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA), the samples were amplified once using T7 RNA polymerase from the MessageAmpII aRNA Kit (Ambion, Austin, TX). Cy3 or Cy5 was incorporated into the amplified RNA products using amino allyl-modified dUTPs. Fluorescently labeled amplified RNAs were hybridized to DNA microarrays using SlideHyb Glass Array Hybridization Buffer (Ambion, Austin, TX). Arrays were produced by the UCSF Sandler Microarray facility using the Mouse Exonic Evidence Based Oligonucleotide (MEEBO) set, a 70-mer oligonucleotide gene set that contains over 25,000 genes, including alternatively spliced exons. After hybridization, arrays were washed and scanned using an Axon GenePix 4000B scanner (Molecular Devices Corporation, Sunnyvale, CA) and images were processed using GenePix 5.0 software. The `print-tip loess' normalization and single-channel quantile normalization was used to correct for within-array dye and spatial effects in order to facilitate comparison between arrays. The functions in the library marrayNorm of the R/Bioconductor package were used to perform these normalizations. After normalization, a log ratio (C3H10 Tet-On mP311-myc/C3H10 Tet-On TRE-REV) was determined for each probe on the array. Statistical analyses were performed using the functions in the library limma of the R/Bioconductor package to obtain a B (log posterior odds ratio) value, and those genes with a B value greater than one and with a twofold or greater change in gene expression in cells stably expressing P311 were further assessed by Ensemble attribute profile clustering (Semeiks et al., 2006).

To measure total cellular cholesterol, cells were treated with fatty-acid complexes consisting of 1 mM of Na-oleate and 1% BSA. The next day, cellular lipids were extracted with 2 ml of hexane:isopropanol (3:2 vol/vol) per well of a 6-well plate. The lipid extract was collected into 12×75 mm glass tubes, the wells rinsed with an additional 1 ml of hexane:isopropanol (3:2), and the combined extracts dried under nitrogen and dissolved in 200 μl of isopropanol containing 10% Triton X-100. The levels of free cholesterol were quantified with the Cholesterol Quantitation Kit (BioVision, Mountain View, CA) according to the manufacturer's instructions. To measure triglyceride synthesis and accumulation, cells were cultured in 6-well plates until confluency and supplemented for 24 hours with medium containing 1 mM of Na-oleate, 1% BSA, and 2 μCi of [¹⁴C]oleic acid as a tracer. The next day cells were placed on ice and washed three times with ice-cold PBS 0.2% BSA, two times for 10 minutes with ice-cold PBS 0.2% BSA, and two times with ice-cold PBS. A recovery standard of 0.02 μCi of [³H]triolein was added into each well and lipids were extracted with 2 ml of hexane:isopropanol (3:2 vol:vol) per well of a 6-well plate as described for cholesterol measurements. Lipid extracts were separated by thin-layer chromatography (TLC), the triglyceride band isolated, and radioactivity quantified by scintillation counting. All lipid measurements were normalized to lipid recovery and protein concentration. To determine protein concentration, proteins were extracted from cells remaining on the well after lipid extraction by incubation in 1 ml of 0.1 M NaOH for 10-15 minutes at room temperature, and subsequently quantified using the Bradford method (Bio-Rad, Hercules, CA). Experiments were performed in triplicate and statistical significance determined by the Student's t-test.

Control and P311 stable C3H10 cells were supplemented with 1 mM Na oleate 1% BSA for 24 hours. Cells were then washed twice with PBS and fixed in 10% formalin in PBS at 37°C for 1.5 hours. The cells were then washed three times with water and 2-3 ml of dye solution was added. The dye solution was prepared by dissolving 4.2 g of Oil-Red-O (Sigma) in 1200 ml of absolute isopropanol and left overnight without stirring at room temperature. The solution was vacuum-filtered through a Buchner funnel, diluted with 900 ml of double-distilled water, left overnight at 4°C without stirring and then filtered twice. After incubation of the cells in the dye solution at 37°C for 3 hours, excess dye was removed and the dish left to dry in the incubator for 10 minutes. The red-stained lipid droplets were analyzed under light-microscopy using a Nikon TE100 microscope.

For infection, NIH3T3-L1 cells (ATCC, Manassas, VA) were incubated with recombinant lentivirus in the presence of 4 μg/ml polybrene and subjected to centrifugation at room temperature for 90 minutes at 1000 rpm in a tabletop centrifuge. After 2-3 days at 37°C infected cells were identified by the expression of the GFP co-marker signal. NIH3T3-L1 cells were sub-cultured at 3.3×10³ cells/cm² and grown to high confluency (no visible gaps between cells) in DMEM supplemented with 10% calf serum. Two days post confluency, cells were stimulated with induction medium (DMEM supplemented with 10% FBS, 0.5 mM IBMX, 1 μg/ml insulin, 1 μM dexamethasone) (Sigma). After 2 days the induction medium was changed to insulin-containing medium (DMEM supplemented with 10% FBS, 1 μg/ml insulin). After 2 days, this medium was changed to DMEM supplemented with 10% FBS, which was then changed every 2 days. Full differentiation was achieved at day 8. For lentivirus infection, cells were incubated with recombinant virus in the presence of 4 μg/ml polybrene and subjected to centrifugation at room temperature for 90 minutes at 1000 rpm in a tabletop centrifuge. After 2-3 days, infected cells were subjected to induction of adipogenic differentiation as described above.

Cells were fixed for 6 minutes in 4% paraformaldehyde followed by permeabilization in 0.1% Triton X-100 for 6 minutes. Blocking was performed in PBS+5% bovine serum albumin (BSA). Cells were incubated using anti-adipophilin (Research Diagnostics Inc., Concord, MA) or anti-Tip47 antibody (a kind gift from Perry Bickel, University of Texas Health Science Center, Houston, TX) for 1 hour in PBS+5% BSA, washed, and were then incubated for 1 hour with either Alexa-Fluor-594-conjuated goat anti-mouse antibody (Molecular Probes, Carlsabad, CA) or FITC-conjugated donkey anti-rabbit antibody (Southern Biotech, Birmingham, AL), washed, and mounted in Vectashield mounting medium (Vector Laboratories, Burlingame, CA).

Lung fibroblasts were isolated from lungs of 1-week-old C57 mice as previously described (Bruce and Honaker, 1998). Briefly, following euthanasia lungs were promptly dissected, rinsed in Ca²⁺- and Mg²⁺-free Hank's balanced salt solution (HBSS), minced into 1-mm³ to 2-mm³ pieces, and placed into a solution of 0.3 mg/ml type IV collagenase and 0.5 mg/ml trypsin in HBSS (in 37°C water bath placed on a shaker) for 60 minutes. Minced tissue was passed through a 25-ml pipette at 20-minute intervals to dissociate the cells. Cells in suspension were removed from the lung homogenate and added to an equivalent volume of cold complete medium consisting of a 1:1 (vol/vol) of DMEM/Ham's F12, 10% FBS, and glutamine and plated into a 100-mm cell-culture dish. For experiments involving the fatty-acid exposure and retinoic-acid treatment of lung fibroblasts, cells were seeded at 3.3×10³ cells/cm². The following day cells were treated with fatty-acid complexes consisting of 1 mM Na-oleate and 1% BSA and retinoic acid (Sigma) at the concentrations indicated and for the indicated times. Experiments were performed in triplicate and statistical significance was determined by the Student's t-test.

2 μg total RNA was treated with DNase I (Invitrogen) and reverse transcribed using the SuperScript First-Strand Synthesis System for RT-PCR (Invitrogen, Carlsbad, CA). PCR was performed in the ABI PRISM 7000 Sequence Detection System (Applied Biosystems, Foster City, CA). Primers and probes were designed using the Primer Express program from the manufacturer. The sequences for the primers/probes used are as follows: P311: probe, 5′-AGACTGGCGCTGCCTCGCTGAC-3′, forward primer, 5′-GAAGTGAACCGAAAGAAGATGAG-3′, reverse primer, 5′-GAATTCACGGCTGCCTGG-3′; GAPDH: probe, 5′-CCGCCTGGAGAAACCTGCCAAGTATG-3′, forward primer, 5′-TGT GTCCGTCGTGGATCTGA-3′, reverse primer, 5′-CCTGCTTCACCACCTTCTTGAT-3′. The 5′ and 3′ modification of all the probes are FAM and BHQ, respectively. Thermal cycling conditions were 50°C for 2 minutes, 95°C for 10 minutes, followed by 40 repetitive cycles of 95°C for 15 seconds and 60°C for 1 minute. Following PCR, quantification of mRNA levels was done by determining the number of cycles to threshold (C_T) of fluorescence detection within the geometric region of the semi-log plot of fluorescence detection. Relative gene expression was determined using the comparative C_T method as described previously (Hettinger et al., 2001). Briefly, the ΔC_T value was obtained by subtracting the GAPDH C_T value from the P311 C_T value in each sample. The ΔΔC_T was obtained by subtracting the mean ΔC_T of the control samples from the ΔC_T of each experimental sample. The fold change in gene expression was determined by calculating 2-^ΔΔCT and is expressed as fold change or as percentile change by multiplying by 100. Experiments were performed in triplicate and statistical significance determined by the Student's t-test.

The lentiviral construct pSicoR and lentiviral packaging constructs VSV-g, REV, and MDL were provided by Tyler Jacks (MIT, Cambridge, MA) and Michael McManus (UCSF, CA). The 5′ phosphorylated, PAGE-purified oligonucleotids used for shRNA constructs to targeting P311 were designed using the pSicoOligomaker program (http://web.mit.edu/ccr/labs/jacks/protocols/pSico.html). The shRNA construct was designed with the following pairs of oligos: forward: 5′-TGAATTCACCTCTCCAGCTATTCAAGAGATAGCTGGAGAGGTGAATTCTTTTTTC-3′, reverse: 5′-TCGAGAAAAAAGAATTCACCTCTCCAGCTATCTCTTGAATAGCTGGAGAGGTGAATTCA-3′. The specificity of oligonucleotids was verified using BLAST. Oligonucleotids were resuspended to a final concentration of 100 μM and annealed in a reaction containing 23 μl ddH₂O, 1 μl sense oligonucleotides, 1 μl antisense oligonucleotid, and 25 μl of 2× annealing buffer (200 mM K-acetate, 60 mM HEPES-KOH pH 7.4, and 4 mM Mg-acetate) and incubated for 4 minutes at 95°C, then at 10 minutes at 70°C, and then slowly cooled down to 4°C. To ligate the annealed oligonucleotids to the pSicoR vector, 1 μl of a 1:10 dilution of the annealed oligonucleotids was added to 50-100 ng of HpaI-XhoI-digested pSicoR in a 10 μl reaction mixture and incubated at ambient temperature for 3 hours. Transformation of 2 μl of the ligation mixture yielded positive clones that were sequence verified. To produce lentiviral particles, the 293T packaging cell line (ATCC) was plated at 2×10⁶/100 mm and transfected with the plasmids encoding components of the lentiviral system (VSV-g, MDL, REV, SicoR). After 48-72 hours, cells were scraped and collected. Cells were disrupted by freeze-thaw three times, and cell debris was removed by centrifugation at 3000 rpm for 10 minutes in a tabletop centrifuge. The supernatant was filtered through a 0.45-μm-pore nitrocellulose membrane. Virus was aliquoted and frozen at –80°C until use. For infection, cells were incubated with recombinant virus in the presence of 4 μg/ml polybrene and subjected to centrifugation at room temperature for 90 minutes at 1000 rpm in a tabletop centrifuge. After 2-3 days at 37°C, infected cells were identified by the expression of the GFP co-marker signal.

Lung fibroblast cells were detached with 0.05% trypsin 0.02% EDTA for 5 minutes, resuspended in complete medium and centrifuged at 1000 rpm for 10 minutes in a tabletop centrifuge. Cells were then resuspended in PBS, filtered through a nylon mesh and stained with Nile Red (Sigma) at a final concentration of 0.1 μg/ml. For flow cytometry, cells were subsequently washed three times, resuspended in PBS, and analyzed for Nile-Red fluorescence with a Becton Dickinson FACSort flow cytometer. Nile-Red fluorescence was detected through a 530/30-nm bandpass filter on FL2. To quantify the flow cytometry data we determined the mean fluorescent intensities (MFI) of each of the samples. The MFI was determined using the FCSPress analysis program (FCSPress, Cambridge, UK). For dual-color flow cytometric analysis, cells were washed three times, resuspended in PBS, and analyzed using a Becton Dickinson FACSort flow cytometer. non-infected and non-stained cells were used as baseline controls for GFP and Nile-Red signals and to adjust FL1 (GFP) and FL2 (Nile Red) PMT voltages. Separate control samples with GFP fluorescence or Nile-Red staining were used for instrument setup and compensation was performed to ensure one fluorochrome did not interfere with the other. All fluorescence data were collected on an arbitrary four-decade log scale and analyzed using FCSPress. FL1 versus FL2 cytograms were used to determine infected (shRNA-containing) cells with lipid droplets (FL1+/FL2+) and without lipid droplets (FL1+/FL2-).

We thank Andrea Barczak and Rebecca Barbeau of the Sandler Microarray Core for their assistance with microarray hybridizations and data analysis. This study was supported by grants to T.H.V. from the National Institute of Health (HL069925, HL075680 and HL073823), from the March of Dimes Research Foundation, and from the Sandler Family Supporting Foundation. J.K.L. was supported by an NIH Ruth L. Kirschstein NRSA Individual Fellowship.