A Conserved Serine Residue Is Required for the Phosphatidate Phosphatase Activity but Not the Transcriptional Coactivator Functions of Lipin-1 and Lipin-2^*

Animal Studies

C57BL/6J mice were obtained from the Jackson Laboratory (Bar Harbor, ME). All mice were fed Purina 5001 mouse chow and were maintained on a 12:12-h light:dark cycle. Animal studies were performed under approval of the UCLA and University of Alberta, Edmonton, Institutional Animal Care and Use Committees.

Lipin Expression Constructs and Site-directed Mutagenesis

V5 epitope-tagged lipin-1 and lipin-2 expression plasmids were generated as described previously (6). Site-directed mutagenesis was performed using the QuickChange site-directed mutagenesis kit (Stratagene). Oligonucleotides used are as follows: lipin-1A^S721L (f, TACAAGTTTCTCTATTGTTTAGCACGTGCCATTGGGATGGCGGAC, and r, GTCCGCCATCCCAATGGCACGTGCTAAACAATAGAGAAACTTGTA); lipin-1A^S721D (f, TACAAGTTTCTCTATTGTGACGCACGTGCCATTGGGATGGCGGAC, and r, GTCCGCCATCCCAATGGCACGTGCGTCACAATAGAGAAACTTGTA); lipin-2^S731L (f, TGGCTACAAGTTTCTGTACTGTTTAGCACGTGCCATCGGCATGGC, and r, GCCATGCCGATGGCACGTGCTAAACAGTACAGAAACTTGTAGCCA); and lipin-2^S731D (f, TGGCTACAAGTTTCTGTACTGTGACGCACGTGCCATCGGCATGGC, and r, GCCATGCCGATGGCACGTGCGTCACAGTACAGAAACTTGTAGCCA).

Cell Culture and Transfection

HEK 293 and Hepa 1–6 cells (American Type Culture Collection, Manassas, VA) were propagated in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and antibiotics and incubated at 37 °C in 5% CO₂. Cells were transfected with Effectene (Qiagen). Two days after transfection, the cells were lysed and assayed for PAP or luciferase reporter activity.

PAP Activity

PAP assays were performed as described previously (6). Briefly, HEK 293 cells were homogenized in 0.25 m sucrose containing 2 mm dithiothreitol (to stabilize PAP activity), protease inhibitor mixture (EDTA-free, Roche Diagnostics), phosphatase inhibitor cocktails I and II (Sigma), and 0.15% Tween 20. We assayed PAP activity in a final volume of 0.1 ml containing 100 mm Tris maleate buffer, pH 7.5, 5 mm Mg²⁺, and 0.6 mm PA labeled with [³H]palmitate (about 75 Ci/mol) dispersed in 0.4 mm PC and 1 mm EDTA (22, 23). The final concentration of Tween 20 in the assay was adjusted to 0.05%. Reactions were stopped after incubation at 37 °C with 2.2 ml of chloroform containing 0.08% olive oil as carrier for neutral lipids. Then 0.8 g of basic alumina was added to adsorb the PA and any [³H]palmitate formed by phospholipase A-type activities (23). The tubes were centrifuged, and 1 ml of the chloroform, which contains the [³H]diacylglycerol product, was dried and quantitated by scintillation counting. Protein concentrations (measured by the Bradford method) and the times of incubation (normally 30 min) were adjusted so that <10% of the PA was consumed during the incubation. Total PAP activities were calculated from measurements at three different protein concentrations to ensure the proportionality of the assay. The activities measured in parallel assays containing 8 mm N-ethylmaleimide allowed us to estimate the contribution of lipid phosphate phosphatase activity to the total activity. This former activity was negligible compared with that of the overexpressed PAP activity, and therefore no correction was necessary in these experiments.

PAP activity was normalized to lipin protein expression by quantitative Western blot analysis. Equivalent amounts of protein from each sample were electrophoresed in 3–8% SDS-polyacrylamide gradient gels (Invitrogen) and transferred to nitrocellulose membranes using a semi-dry electroblotter (Bio-Rad). Membranes were blocked with 5% nonfat milk powder in Tris-buffered saline and incubated overnight with mouse monoclonal anti-V5 antibody conjugated to horseradish peroxidase (Invitrogen). Protein bands were detected using ECL-Plus chemiluminescent reagent (Amersham Biosciences). Fluorescent signal was detected with the Typhoon Imaging System and quantified with ImageQuant software.

Transcriptional Coactivation Assay

Hepa 1–6 cells were transiently transfected with a PPRE-firefly luciferase reporter plasmid along with phRL-TK-renilla luciferase control vector (Promega), pCMX-PPARα, pCMX-PGC-1α, pCMX-RXR (kindly provided by Dr. Peter Tontonoz), and wild-type or mutant lipin expression plasmids. As a negative control to demonstrate specificity of the response, the PPRE-Luc reporter was replaced with a mutant version in which the PPRE sequences were deleted (ΔPPRE). Two days after transfection, cells were washed in 1× phosphate-buffered saline and lysed in Passive Lysis Buffer (Promega), and luciferase assays were performed using the Dual-Luciferase assay system (Promega). Firefly luciferase activity was normalized to Renilla luciferase activity to account for differences in transfection efficiency. In each experiment, samples were analyzed in quadruplicate.

In Vitro Translocation Assay for Lipins

Measurement of the capacity of wild-type and mutant lipin-2 proteins to translocate to endoplasmic reticulum membranes was based upon our previous work (24). Livers of male Sprague-Dawley rats were perfused with 60 ml of sterile ice-cold phosphate-buffered saline, pH 7.4, and homogenized in 0.25 m sucrose containing 2 mm dithiothreitol and 20 mm HEPES buffer, adjusted to pH 7.4 with KOH. The homogenates were centrifuged at 4 °C for 15,300 × g for 20 min, and the supernatant was then centrifuged for a further 400,000 × g for 40 min to pellet the microsomal fraction. This pellet was resuspended at 37 °C in homogenization buffer containing 6 mg/ml fatty acid-free bovine serum albumin and incubated for 20 min to remove fatty acids and dissociate bound lipin from the membranes. The microsomal membranes were then collected after re-centrifuging at 4 °C as above and resuspended in the same buffer. Recombinant lipins were expressed in HEK 293 cells, which were then homogenized in 0.25 m sucrose containing 2 mm dithiothreitol and 20 mm HEPES buffer, and adjusted to pH 7.4 with KOH and phosphatase inhibitor cocktails 1 and 2 from Sigma. The mitochondria-free supernatant was then isolated after centrifugation for 16,000 × g for 20 min to collect lipin-2 bound to endoplasmic reticulum membranes and in the cytosol.

For the translocation experiments, 350 μg of liver microsomal protein was incubated for 5 min at 37 °C in a final volume of 20 μl of homogenizing buffer containing 6 mg/ml bovine serum albumin and 0–750 μm K⁺ oleate. The extract from the HEK 293 cells containing recombinant lipin (about 250 μg of total protein) was added, and the tubes were then incubated for 10 min at 37 °C to equilibrate the lipins between cytosol and membrane fractions (24). Following this, the samples were cooled on ice for 5 min. The soluble proteins were then collected after centrifugation at 400,000 × g for 40 min at 4 °C. The pellets (membranes) were resuspended in 45 μl of homogenization buffer. The cytosolic fractions were treated with 4 volumes of acetone at −20 °C. The precipitated protein was collected by centrifugation and resolubilized in 45 μl of homogenization buffer. Samples for Western blotting were prepared with 4× NuPAGE LDS SB from Invitrogen and then stored at −80 °C.

PCR-Restriction Fragment Length Polymorphism Assay

Total RNA was isolated from C57BL/6J mouse liver with TRIzol (Invitrogen), and cDNA was synthesized from 1 μg of RNA using Omniscript reverse transcriptase kit (Qiagen, Valencia, CA). The PCR-restriction fragment length polymorphism assay was performed by simultaneous amplification of lipin-2 and lipin-1 using the following primers (f, ACCATCTACCTGTGGAA, and r, AGAAACTTGTAGCCATTCT), followed by restriction enzyme digestion of amplified fragment with either EcoRV or PvuII (New England Biolabs). Products were analyzed by electrophoresis in 2% agarose.

In Situ Hybridization

To generate templates for sense and antisense RNA generation, we PCR-amplified a 661-bp fragment of mouse lipin-2 cDNA (oligonucleotide primers f, cttcctaggccaccactcag, and r, tgaaataatctgccccaagg) and subcloned it into pCR 2.1-TOPO (Invitrogen). Plasmids were sequences to select templates for sense and antisense probe generation and linearized by digestion with BamHI. In situ hybridization was performed using the services of Phylogeny, Inc. (Columbus, OH). ³⁵S-UTP-labeled cRNA of transcripts was synthesized by in vitro transcription and hybridized to mouse whole body or tissue sections. Whole body sections were prepared from adult mice fasted for 4–5 h in the morning before sacrifice. Additional tissues were prepared from animals with and without fasting, as indicated. Sections were frozen, cut into 10 μm thick sections, mounted on gelatin-coated slides, and stored at −80 °C. Sections were then hybridized overnight at 55 °C in 50% deionized formamide, 0.3 m NaCl, 20 mm Tris-HCl, pH 7.4, 5 mm EDTA, 10 mm NaH₂PO₄, 10% dextran sulfate, 1× Denhardt's, 50 μg/ml total yeast RNA, and 50,000–80,000 cpm/μl ³⁵S-labeled cRNA probe. The tissues were subjected to stringent washing at 65 °C in 50% formamide, 2× SSC, 10 mm dithiothreitol and washed in phosphate-buffered saline before treatment with 20 μg/ml RNase A at 37 °C for 30 min. Following washes in 2× SSC and 0.1× SSC for 10 min at 37 °C, the slides were dehydrated, exposed to Kodak BioMaxMR x-ray film for 2–5 days, and then dipped in Kodak NTB nuclear track emulsion and exposed in light-tight boxes with desiccant at 4 °C for 10–15 days. Photographic development was carried out in Kodak D-19. Slides were counterstained lightly with hematoxylin and analyzed using both light and dark field optics. Sense control cRNA probes (identical to the mRNAs) always gave background levels of hybridization signal.

qPCR of Lipin Gene Expression in Mouse Tissues

Blood and tissues were isolated from adult male C57BL/6J mice. Red blood cells were isolated from fresh whole blood after centrifugation at 200 × g for 10 min at 4 °C. The spleen was divided into two pieces, and the white blood cell fraction was isolated from one piece by selective lysis of erythrocytes using 2.5 ml of lysis buffer (0.15 m NH₄Cl, 10 mm KHCO₃, 0.1 mm Na₂EDTA), incubating at room temperature for 5 min, and collecting remaining cells (enriched in lymphocytes) by centrifugation at 200 × g for 10 min. Cells were washed in phosphate-buffered saline containing 3% fetal bovine serum, and the final pellet was used for RNA isolation. RNA and cDNA were prepared from blood cells or flash-frozen tissues (liver, spleen, and thymus) as described above. qPCR was performed with SYBR Green as described previously (6, 25). Primers were as follows: lipin-1, GCTCCCGAGAGAAAGTGGTGGA, GGCTTTCCATTCTCGCAGCTCCT; lipin-2, AGTTGACCCCATCACCGTAG, CCCAAAGCATCAGACTTGGT; and lipin-3, TGGAATTGGGATGACAAGGT, CACTGCAAGTACCCCTTGGT.

Statistical Analysis

Experimental groups were compared with appropriate control groups as indicated in the figure legends using the Student's t test and were considered statistically significant for p < 0.05.

Mutating the Majeed Serine Residue Abolishes PAP Enzyme Activity

The Majeed S734L mutation in human lipin-2 alters a serine residue that is located in the C-LIP domain downstream of the enzyme active site and coactivator motifs (Fig. 1A). The affected serine residue is conserved in lipin-1, -2, and -3 in mammals and in orthologous proteins in other species. We investigated the biochemical basis for the Majeed syndrome in individuals carrying an S734L missense mutation in lipin-2 by analyzing PAP activity of recombinant mouse lipin-2 with a leucine substitution at the analogous serine residue, Ser 731. We also produced the corresponding mutation in mouse lipin-1A at Ser 724. Lipin-1 protein occurs in two forms (lipin-1A and lipin-1B) that result from alternative mRNA splicing, both of which have PAP activity (6). The lipin-1A splice variant was utilized in all studies described here. We expressed lipin-1A^S724L and lipin-2^S731L in HEK 293 cells and determined PAP activity in cell lysates. Wild-type lipin-1A and lipin-2 served as positive controls.

Wild-type lipin-1A and lipin-2 had robust PAP activity (63 ± 8.4 and 36 ± 3.1 nmol/mg/min, respectively) (Fig. 1B). As expected, mutation of the PAP motif (D679E) or the coactivator motif (IL693FF) abolished PAP activity, as has been reported previously (5). The serine to leucine substitution found in Majeed patients likewise completely abolished PAP activity in both lipin-1 and lipin-2 backgrounds. Wild-type lipin-2 and lipin-2^S731L proteins were expressed at similar levels, confirming that the lack of PAP activity is not related to reduced protein levels or protein stability (Fig. 1C and data not shown). Mutation of the Ser residue to aspartic acid likewise abolished PAP activity in both lipin-1A and lipin-2, indicating that activity cannot be restored by the presence of a negative charge (Fig. 1B).

PAP Deficiency of Lipin-2^S731L and Lipin-2^S731D Does Not Result from Impaired Membrane Association

Studies performed several years ago, prior to the molecular identification of the lipin proteins, demonstrated that PAP enzymes reside in the cytosolic compartment of the cell. They translocate reversibly to endoplasmic reticulum membranes to catalyze the PAP reaction when cells are stimulated with unsaturated rather than saturated fatty acids (24, 26). Although lipin-2 is known to associate largely with the endoplasmic reticulum when expressed in HeLa cells (10), fatty acid-induced translocation has never been directly demonstrated for lipin-2. Therefore, we investigated whether wild-type lipin-2 exhibits oleate-induced membrane translocation and whether membrane association is impaired by mutation of the Majeed serine residue. Combined membrane and cytosolic fractions from HEK 293 cells expressing similar levels of recombinant wild-type lipin-2 or lipin-2^S731L were used for these experiments rather than cytosolic fractions alone so as not to exclude forms of lipin-2 that might preferentially bind to membranes. These forms of lipin-2 were added to rat liver microsomes, which had been depleted of bound lipins by washing with albumin, in the presence of increasing oleate concentrations (0, 375, and 750 μm). This incubation allowed lipin-2 to redistribute between membrane and cytosolic fractions. The relative amounts of lipin-2 in the cytosol and membrane fractions that were re-isolated were determined by Western blot analysis. Cytosolic- and membrane-associated lipin-2 levels were normalized to levels of glyceraldehyde-3-phosphate dehydrogenase and calnexin, respectively, which do not alter their localization in the presence of oleate.

Wild-type lipin-2 was present at similar levels in cytosolic and membrane fractions in the absence of oleate, and it translocated to the membrane fraction with increasing levels of oleate (Fig. 1, C, left panel, and D). Lipin-2^S731L showed no reduction in membrane association and in multiple trials exhibited a trend toward increased membrane localization compared with wild-type lipin-2, particularly in the absence of oleate (Fig. 1, C, right panels, D). However, there was no statistically significant difference in membrane localization between wild-type and mutant lipin-2 (p = 0.09, Student's t test). Likewise, the lipin-2^S731D protein exhibited normal membrane localization (Fig. 1E), suggesting that the amino acid residue at this position is not a determinant of membrane association. Significantly, wild-type as well as the S731L and S731D mutant forms of lipin-2 resolved into two bands (see Fig. 1C). The cytosolic lipin-2 consisted primarily of the slower migrating band, whereas the membrane-bound forms had a high proportion of the faster migrating band. This result is compatible with that for lipin-1 where a faster migrating species (hypophosphorylated lipin-1) associated preferentially with membranes (7, 10).

Thus, the defective action of lipin-2 in the Majeed syndrome cannot be attributed to a failure of the protein to associate with endoplasmic reticulum membranes. It remains possible, however, that the mutant lipin-2 protein assumes an abnormal configuration in its membrane interaction that could affect catalysis.

Lipin-2 Amplifies PGC-1α-mediated Transcriptional Coactivation, and the Majeed Mutation Does Not Impair Coactivator Activity

Lipin-1 functions as a coactivator in the PGC-1α/PPARα pathway, leading to transcriptional coactivation of PPARα target genes in hepatocytes (5). Because the PPAR-binding motif present in the C-LIP domain of lipin-1 is conserved in lipin-2, we tested the ability of lipin-2 to coactivate transcription through a PPRE. Transcriptional coactivation assays were performed using a PPRE-luciferase (PPRE-Luc) reporter gene in Hepa 1–6 hepatoma cells. Cells were transfected with the reporter construct and PPARγ and retinoid X receptor-α expression vectors in the presence and absence of PGC-1α and wild-type or mutant lipin expression constructs. Similar to previous results for lipin-1B (5), lipin-1A activated PPRE-luciferase expression ∼2-fold (Fig. 2, compare bars for lipin-1A to PPRE-Luc) and amplified the effect of the PGC-1α coactivator (compare activity of PGC-1α + lipin-1A with that of PGC-1α alone). This effect required a functional PPRE, as deletion of this regulatory element (ΔPPRE-Luc) prevented activation. Lipin-2 had nearly identical coactivator activity as lipin-1, showing modest activation of the reporter on its own (compare lipin-2 with PPRE-luc) and amplifying of the effect of PGC-1α (compare PGC-1 α + lipin-2 with PGC-1 α). Notably, the lipin-2^S731L mutant showed identical coactivator activity to wild-type lipin-1A and lipin-2. Thus, the serine that is required for PAP activity of both lipin-1 and lipin-2 is not required for coactivator function.

Lipin-2 Gene Expression in Liver, Digestive Tract, Red Blood Cells, and Lymphoid Cells

We previously demonstrated using qPCR that the three lipin genes have distinct, but overlapping, expression patterns across a panel of several mouse and human tissues (6). To obtain a more complete view of lipin-2 tissue expression throughout the body, including tissues that are affected in Majeed syndrome such as bone and skin, we used in situ hybridization with antisense lipin-2 RNA on whole mouse sections. Hybridization was observed at extremely high levels in the liver, as well as in specific regions of the gastrointestinal tract, kidney, and brain (Fig. 3). As a negative control, hybridization was performed with lipin-2 sense strand probe and showed minimal background hybridization, primarily in skin and retina (Fig. 3C).

In the liver, lipin-2 showed a uniform distribution pattern (Fig. 4) with labeling observed within hepatocytes (hepatocyte nuclei are visible as darkly stained areas in Fig. 4, C–F). Fasting for 14 h increased labeling by 2-fold (compare Fig. 5, C and D). It has been shown that lipin-1 is expressed in liver, where it is also induced by fasting (5). Using qPCR, we determined that fasting induces hepatic lipin-1 by 25-fold and lipin-2 by 5-fold (Fig. 4G). Thus far, it has been unclear whether lipin-1 or lipin-2 is expressed at higher absolute levels in the liver. To address this, we designed an assay to PCR-amplify lipin-1 and lipin-2 cDNA simultaneously with the same primer set and then to distinguish the products by restriction enzyme digestion (Fig. 4H). The PCR primers amplified a 170-bp fragment from both lipin-1 and lipin-2 cDNA. Subsequent digestion of the amplified product with PvuII cut the lipin-2 fragment into 97- and 73-bp fragments but did not alter lipin-1. In contrast, digestion with EcoRV cut the lipin-1 fragment to produce 125- and 45-bp fragments but did not cut lipin-2 (Fig. 4H). Both restriction reactions showed that lipin-2 mRNA levels in liver are substantially higher than lipin-1 levels, suggesting that lipin-2 may significantly contribute to PAP activity and lipin-associated coactivator activity in this tissue.

In addition to liver, in situ hybridization revealed significant levels of lipin-2 expression in the gastrointestinal tract. Expression was detected in squamous keratinized epithelium of the tongue and the esophagus, the epithelial cell layer of the intestinal mucosa, and in epithelial cell villi of the duodenum, ileum, and jejunum (Fig. 5). Lipin-2 expression in kidney was detected in the cortex, medulla, and papillae (results not shown). In brain, lipin-2 was expressed in the cortex, hippocampus, thalamus, and hypothalamus (Fig. 3B and results not shown). Lipin-2 was also expressed in the submandibular and submaxillary salivary glands, as well as the thymus and lung (Fig. 3B). Notably, we did not detect lipin-2 expression using in situ hybridization in bone or skin, tissues that show pathologic changes in Majeed syndrome (18, 19). The lack of lipin-2 expression in skin was confirmed using the extremely sensitive method of real time PCR; in contrast, lipin-1 mRNA could be detected at low levels in skin (Fig. 6A). We detected low levels of lipin-2 expression in bone and bone marrow, at 10–20% the levels in liver (Fig. 6A).

We also assessed whether the anemia and inflammatory conditions observed in Majeed syndrome may reflect a role for lipin-2 in red blood cells and lymphoid cells. Using qPCR, we detected mouse lipin-2 expression in circulating red blood cells, in whole spleen and the white blood cell fraction of spleen, and in thymus, all at ∼10% of the levels observed in liver, where lipin-2 is expressed at highest levels (Fig. 6B). Although direct comparisons of gene expression among all three lipin genes are not possible using different PCR primer sets, based on the determination that lipin-2 is expressed at much higher levels in liver than lipin-1 or lipin-3 (Fig. 4H) (6), lipin-2 appears to be the prominent lipin mRNA expressed in red blood cells and lymphoid tissues (Fig. 6).