LPS Impairs Phospholipid Synthesis by Triggering β-Transducin Repeat-containing Protein (β-TrCP)-mediated Polyubiquitination and Degradation of the Surfactant Enzyme Acyl-CoA:Lysophosphatidylcholine Acyltransferase I (LPCAT1)^*

Cells and Reagents

Murine lung epithelial (MLE) cells were maintained with HITES medium complemented by 10% FBS in a 37 °C incubator with a supplement of 5% CO₂. Rabbit LPCAT1 polyclonal antibody was generated by Covance (Princeton, NJ). V5 antibody, mammalian expression plasmid pcDNA3.1/HisV5-topo, and Escherichia coli Top10 competent cells were purchased from Invitrogen. Wild type, hyperactive, or kinase-dead GSK-3β plasmids were generous gifts from Dr. John Engelhardt (15). The HA-ubiquitin construct was a generous gift from Dr. Peter M. Snyder. Phosphoserine antibodies were from Cell Signaling (Danvers, MA). The siRNAs and GSK-3, mouse V5 monoclonal, and phosphoGSK-3 antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA). Immobilized protein A/G beads were obtained from Pierce. Proteasome inhibitor MG132 was from Calbiochem. HA antibody, leupeptin, cycloheximide, E. coli lipopolysaccharide (LPS), and phosphatase inhibitor mixtures were from Sigma. Gel extraction kits and QIAprep spin miniprep kits were from Qiagen (Valencia, CA). All materials in highest grades used in the experiments are commercially available.

Plasmid Construction and Expression

Full-length mouse Lpcat1 cDNA was amplified by PCR using forward primer 5′-CACCATGAGGCTGCGGGGCCGCGGGCCG-3′ and reverse primer 5′-GTCCGCTTTCTTACAAGAATTCTTTCTCCCAAAG-3′ and Lpcat1 ORF (Open Biosystems) as a template. The PCR product was cloned into the pcDNA3.1/His/V5 vector, and the accuracy of the cDNA sequence was confirmed by DNA sequencing. Site-directed mutagenesis was performed by using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA) following the directions of manufacturer using pcDNA3.1/hisV5/Lpcat1 plasmid as a template and appropriate primers (16). Individual SCF F-box subunits were cloned as described previously (17). Cells were nucleofected with plasmids as described previously (16). Briefly, 1 million MLE cells in their exponential growth stage were suspended in 100 μl of nucleofection buffer (20 mm of Hepes in PBS buffer) and well mixed with 3 μg of plasmid DNA in an electroporation cuvette. Electroporation was performed with pre-set program T-013 in Nucleofection^TM II system (Amaxa Biosystems, Gaithersburg, MD), and the cells were cultured in 2 ml of complete HITES medium for 18 h. siRNA oligos were also delivered into cells by using nucleofection with the same protocol, and 0.5 μm siRNA oligos was used in each nucleofection (16).

LPCAT1 Activity

Cells were harvested in lysis buffer (250 mm sucrose, 10 mm Tris-HCl, pH 7.4), and equal amounts of microsomal cellular protein or cell lysate were used in the assay. One nmol of 1-palmitoyl-sn-glycero-3-phosphocholine was added for every microliter of 5× assay buffer (final working concentration 65 mm Tris-HCl, pH 7.4, 10 mm MgCl₂, 12.5 mm fatty acid-free BSA, 2 mm EDTA) and sonicated. 35 μl of H₂O and cellular protein (10 μg) was added to 10 μl of sonicated assay buffer containing 5 μl of [1-¹⁴C]acyl-CoA (0.1 μCi, 1.8 nmol) for a total assay volume of 50 μl. Upon addition of [1-¹⁴C]acyl-CoA, samples were incubated at 30 °C for 10 min, after which the reaction was terminated by addition of chloroform/methanol/H₂O (1:2:0.70, v/v). Total cellular lipids from reaction mixtures were extracted by the method of Bligh and Dyer and spotted on LK5D plates, and PtdCho was resolved using TLC and detected using a Bioscan AR-2000 scintillation plate reader (7).

Immunoprecipitation and Immunoblot Analysis

Cell lysates were prepared by brief sonication of harvested cells in Buffer A (150 mm NaCl, 50 mm Tris-HCl, 1 mm EDTA, 2 mm dithiothreitol, 0.025% of sodium azide, 1 mm phenylmethylsulfonyl fluoride, 1% of Triton X-100, and appropriate amounts of protease inhibitor mixture, pH 7.4). For immunoprecipitation, 1 ml of cleared cell lysates (contains 1 mg of protein) were incubated with 5 μg of specified antibody at 4 °C for overnight rotation, and 20 μl of protein-AG-agarose beads were added in and mixed by rotation for another 2 h. Agarose beads were then collected and washed three times by using 1 ml of Buffer A, and the precipitates were separated by SDS-PAGE and subjected to immunoblotting analysis. Immunoblotting was performed as described, and images were obtained with Carestream Kodak in vivo Image F Pro System. Antibodies were diluted as 1:1000 except when specifically indicated.

In Vitro Ubiquitin Conjugation Assay

³⁵S-Labeled substrate LPCAT1 was synthesized using TnT cell-free reticulocyte lysate in vitro transcription/translation system with pcDNA3.1/His/V5/Lpcat1 as template (Promega). SCF-box protein Skp1, Cullin1, Rbx1, and substrate-specific E3 ubiquitin ligase subunit, β-TrCP, were expressed in mammalian HEK293 cells and purified with Ni²⁺ beads. The ubiquitination assay was performed in a total volume of 25 μl containing 50 mm Tris, pH 7.6, 5 mm MgCl₂, 0.6 mm DTT, 2 mm ATP, 1.5 ng/μl E1 (Boston Biochem), 10 ng/μl Ubc5, 10 ng/μl Ubc7, 1 μg/μl ubiquitin (Calbiochem), 1 μm ubiquitin aldehyde, 2 μl of each His-purified recombinant Cullin1, Skp1, Rbx1, and β-TrCP at room temperature for 30 min. Reaction mixtures were then resolved on SDS-PAGE, followed by transferring to a nitrocellulose membrane and autoradiography.

Fluorescence Microscopy

PCR-based strategies were used to create compatible restriction enzyme sites that would allow construction of chimeric cDNAs for microscopy as described previously (28). Lpcat1 and β-TrCP were amplified using PCR prior to cloning into pAmCyan-C1 or pZsYellow1-C1 (Clontech), generating Lpcat1-CFP or β-TrCP-YFP, respectively. For analysis of LPCAT1 and β-TrCP localization, cells were first plated at 0.12 × 10⁶ cells/well in a two-chamber cover glass system and co-transfected with Lpcat1-CFP and β-TrCP-YFP (2 μg of plasmid/chamber) with FuGENE 6 (6 μl). LPCAT1 and β-TrCP were detected at the single cell level using a combination laser-scanning microscope system (Nikon A1; Nikon, Japan). To achieve excitation, the 458- and 514-nm line of an argon ion laser was focused through the ×60 oil differential interference contrast objective lens onto the cell. Emissions of YFP and cyan fluorescent proteins were collected through 535–595- and 470–500-nm barrier filters, respectively. The results were analyzed through Nikon NIS-Elements software.

In Vitro GSK-3β Kinase Phosphorylation Assay

V5-tagged wild type LPCAT1 and S178A, S182A, and S178A/S182A mutated LPCAT1s were expressed in cells and lysed in Buffer A with sonication. The cleared cell lysates were incubated with LPCAT1 antibody overnight and protein A/G-agarose beads for 2 h with rotation at 4 °C. The beads were washed three times with 1 ml of GSK-3β kinase assay buffer (25 mm MOPS, 12.5 mm β-glycerol phosphate, 25 mm MgCl₂, 5 mm EGTA, 2 mm EDTA, 0.25 mm DTT, pH 7.2). In vitro phosphorylation reactions were conducted by combining 40 μl of protein A/G-agarose bead- bonded LPCAT1 and 10 μl of kinase assay buffer containing 1 μCi of [γ-³²P]ATP and 0.1 μg of active GSK-3β (per each reaction) for 30 min at room temperature. Active GSK-3β was denatured at 95 °C for 10 min as the kinase negative control. The reactions were terminated by addition of SDS loading buffer, and the samples were separated by SDS-PAGE, and phosphorylated LPCAT1 was visualized by autoradiography.

Animals

Lenti-Lpcat1 viral and control constructs were generated by the virus core at the University of Iowa. C57BL/6J mice were administered intratracheal lenti-empty or lenti-Lpcat1 (10⁸ plaque-forming units/mouse) for 96 h prior to intratracheal inoculation with LPS (5 mg/kg) for 24 h. Animals were then mechanically ventilated, and lung resistance, compliance, and elastance (lung stiffness) were measured, and pressure-volume loops were generated. Each group contained 5–9 mice. Lung tissues from mice were homogenized and sonicated in cell lysis buffer followed by immunoblotting analysis. All procedures were executed in accordance with approved protocols through the University of Pittsburgh Institutional Animal Care and Use Committee.

Statistical Analysis

Statistical analysis was carried out by two-way analysis of variance. Data are presented as mean ± S.D.

LPCAT1 Is Degraded by Ubiquitin-Proteasome

In the process of investigating regulation of proteolytically sensitive enzymes involved in PtdCho biosynthesis (7), we tested the stability of the LPCAT1 protein. MLE cells in their exponential growth phase were treated with 10 μg/ml cycloheximide, and LPCAT1 mass was analyzed over time (Fig. 1A). The results indicated that endogenous LPCAT1 t½ is ∼4.5 h (Fig. 1B). These results were consistent with t½ of overexpressed LPCAT1 using similar methods (Fig. 1C). Two major routes for elimination of proteins include the proteasome and/or lysosome. In cells treated with the proteasomal inhibitor MG132 (20 μm), LPCAT1 levels were remarkably stabilized with accumulation by 24 h (Fig. 1, D and E). In contrast, in cells exposed to a lysosomal inhibitor, leupeptin, LPCAT1 mass remained unchanged overall (Fig. 1F). We next assessed if LPCAT1 is modified by ubiquitination as a sorting signal for proteasomal targeting. In the presence of ubiquitin aldehyde (which inhibits deubiquitination activity) and MG132, we detected a robust polyubiquitinated LPCAT1 signal in cell lysates (Fig. 1G). Furthermore, cells were co-transfected with V5-tagged LPCAT1 and HA-tagged ubiquitin plasmids, and lysates were subjected to HA-tagged immunoprecipitation and probed with V5-antibody. Here, V5-LPCAT1 was also polyubiquitinated (Fig. 1H). These results indicate that both endogenous and overexpressed LPCAT1 are degraded by the ubiquitin proteasome machinery but not within the lysosome in epithelia.

SCF^β-TrCP Triggers LPCAT1 Ubiquitination

Data base analysis identified that LPCAT1 harbors several candidate phosphorylation sites, and the enzyme was effectively phosphorylated in vitro (see below). As SCF E3 ligases often target phosphoenzymes for ubiquitination, we performed an unbiased screen analyzing several candidate SCF F-box components on their ability to degrade LPCAT1. Of the candidate F-box proteins expressed in epithelia, only β-TrCP reproducibly produced LPCAT1 degradation (Fig. 2A). Co-expression of β-TrCP with LPCAT1 in cells led to a marked reduction in immunoreactive basal levels of the enzyme (Fig. 2B). β-TrCP expression also reduced LPCAT1 activity by ∼90% under these conditions (Fig. 2C). E3 ubiquitin ligases specifically bind to their targets to provide a niche for the E2 ubiquitin conjugation enzyme to catalyze the covalent ligation between ubiquitin and the acceptor lysine residues of the targeted protein. Consistent with this mechanism, co-immunoprecipitation studies demonstrated that endogenous LPCAT1 binds to β-TrCP (Fig. 2D). We tested in vitro ubiquitinating activity of β-TrCP by incubating LPCAT1 with Skp1, Cullin1, E2 conjugation enzyme-binding protein Rbx1, and ubiquitin. As shown in Fig. 2E (left panel), each SCF subunit in the ubiquitin conjugation reaction was sufficiently expressed. Here, ubiquitin was conjugated into LPCAT1 substrate with the appearance of multiple slower migrating bands suggestive of polyubiquitinated products. Levels of LPCAT1 were markedly reduced in cells exposed to cycloheximide. However, when cells were exposed to β-TrCP siRNA compared with scrambled RNA, enzyme levels recovered significantly (Fig. 2F). These results demonstrate that β-TrCP is an essential E3 SCF component that regulates stability and enzymatic behavior of LPCAT1.

The minimal consensus recognition sequence for β-TrCP binding within targets is SXXXS. Sequence alignment with lysophospholipid acyltransferases uncovered that this motif is well conserved within LPCAT1 (Fig. 3A). To test whether this motif is an authentic β-TrCP-binding site, we generated constructs where we substituted Ser¹⁷⁸ and/or Ser¹⁸² with Ala in the motif. MLE cells were nucleofected with wild type pcDNA3.1/Lpcat1 or serine-mutated LPCAT1 containing plasmids and treated with MG132 for 16 h. Robust accumulation of LPCAT1 in the presence of MG132 was observed in wild type (WT) LPCAT1 plasmid nucleofected cells but not in cells expressing S178A, S182A, or S178A/S182A mutant constructs (Fig. 3B). These results were confirmed in cells treated over time with the proteasomal inhibitor where again WT LPCAT1 accumulation was detected, but the LPCAT1 variants showed limited, if any, accumulation (Fig. 3C). In studies where V5-LPCAT1 was immunoprecipitated and products immunoblotted with β-TrCP antibody, β-TrCP bound wild type LPCAT1, but not the Ser LPCAT1 mutants (Fig. 3D). These results identify that Ser¹⁷⁸ and Ser¹⁸² are indispensable for β-TrCP docking and subsequent ubiquitination within LPCAT1. Finally, cells transfected with LPCAT1-CFP or β-TrCP-YFP were visualized by confocal microscopy demonstrating co-localization of proteins within the cytoplasm (Fig. 3E).

GSK-3β Phosphorylates LPCAT1

We hypothesized that serine residues in the β-TrCP-docking motif within LPCAT1 might include a GSK-3β-phosphorylated signature for E3 ligase recognition. To verify this hypothesis, we silenced the GSK-3β gene and assessed LPCAT1 degradation. GSK-3β was knocked down by siRNA, and levels were reduced further in the presence of cycloheximide (Fig. 4A). As before, cycloheximide effectively reduced LPCAT1 levels, an effect reversed using GSK-3β siRNA but not control RNA oligos (Fig. 4A). We next co-overexpressed wild type, hyperactive, or kinase-dead GSK-3β plasmids with either WT LPCAT1 or Ser-mutated LPCAT1 plasmids in cells. Expression of kinase-dead GSK-3β or wild type GSK-3β produced no effect or only minimally reduced LPCAT1 levels, respectively; expression of hyperactive GSK-3β significantly reduced LPCAT1 protein (Fig. 4B). Levels of the LPCAT1 Ser mutants were not affected by co-expression of any of the three GSK-3β plasmids indicating that these phosphorylation residues (S178A and S182A) are relevant for LPCAT1 destabilization. As a more direct approach, in vitro phosphorylation assays were conducted using immunoprecipitated LPCAT1 and Ser mutants as substrates using ³²P-labeled ATP (Fig. 4C). WT LPCAT1 was phosphorylated by active but not denatured GSK-3β; phosphorylation of LPCAT1 variants harboring point mutations (S178A or S182A) was reduced, and phosphorylation of the doubly mutated protein (S178A/S182A) was drastically decreased. To assess LPCAT1 phosphorylation in cells, we overexpressed V5-tagged WT and mutated LPCAT1 plasmids in cells and silenced GSK-3β prior to V5 immunoprecipitation and by probing with phosphoserine antibody (Fig. 4D). Serine residues within WT LPCAT1 were robustly phosphorylated; however, knockdown of GSK-3β completely nullified its phosphorylation. In preparations isolated from cells expressing LPCAT1 S178A or S182A plasmids, detectable phosphorylation was reduced. Signal intensity of the S178A mutant was less than the S182A variant, and only a trace phosphorylation signal was detected in the LPCAT1 Ser double mutant (Fig. 4D). Interestingly, when we probed V5 antibody immunoprecipitates with GSK-3β, the kinase directly associated with wild type LPCAT1, but this interaction was again reduced greater in the S178A mutant versus the S182A variant (Fig. 4D, 2nd row). Taken together, these observations suggest that specific molecular sites appear to be phosphorylated by GSK-3β, which in turn signals binding interactions between the acyltransferase and the E3 ligase subunit to regulate LPCAT1 stability in cells.

Lys²²¹ Is a Bona Fide Site for LPCAT1 Ubiquitination

To identify putative ubiquitin acceptor sites within LPCAT1, we substituted several candidate Lys residues with Arg within the functional domains that were predicted to be cytosolically exposed and did not reside within the N-terminal transmembrane region of the enzyme. These constructs were expressed in cells, and LPCAT1 accumulation was measured in the presence of proteasomal blockade. Of several mutants tested, only a K221R LPCAT1 mutant was not observed to be polyubiquitinated in the presence of MG132. As a representative immunoblot, Fig. 5A shows that the K221R mutant, unlike WT LPCAT1, K191R, or K199R mutants, remained relatively unmodified in the presence of MG132 (Fig. 5A). These observations were confirmed in kinetic studies where by 8–12 h all the constructs exhibited increasing intensities of native and upper bands with the exception of the K221R mutant (Fig. 5B). Densitometric analysis of immunoblots is shown in Fig. 5C revealing a ∼2-fold increase in levels of WT LPCAT1, K191R, or K199R mutants versus the K221R mutant. Finally, we co-expressed WT LPCAT1 and Lys mutant plasmids together with HA-tagged ubiquitin (Fig. 5D). Ubiquitinated proteins were immunoprecipitated with an HA antibody, and immunoblotting was performed on the precipitates using a V5 antibody. As showed in Fig. 5D, wild type and K199R mutated LPCAT1 were ubiquitinated in a similar manner, but the K221R mutant LPCAT1 was not ubiquitinated. Thus, Lys²²¹ is a critical residue for site-specific LPCAT1 polyubiquitination.

LPS Induces LPCAT1 Degradation by a β-TrCP-GSK-3β Mechanism

As surfactant phospholipid biosynthesis is reduced during Gram-negative sepsis (18), we hypothesized that LPCAT1 life span would be reduced after exposure to lipopolysaccharide. Indeed, LPS reduced immunoreactive LPCAT1 levels in a concentration-dependent (5–20 μg/ml, see Fig. 6A) and time-dependent (18–24 h, Fig. 6B) manner, an effect that was blocked with MG132 treatment. When cells were exposed to LPS or cycloheximide and then silenced using β-TrCP siRNA, LPCAT1 levels were restored as compared with cells exposed to LPS or protein synthesis inhibitor alone (Fig. 6C). LPS activates GSK-3β (19). GSK-3β knockdown by siRNA also restored immunoreactive LPCAT1 levels after cycloheximide or LPS treatment (Fig. 6D). These data suggest that LPCAT1 is pathophysiologically regulated and implicate a β-TrCP-GSK-3β pathway as a potentially important effector mechanism in LPS-induced LPCAT1 degradation.

To extend the above observations to a sepsis model in vivo, we assessed the major airway-secreted product of LPCAT1, surfactant PtdCho, and measured surfactant lung biophysical properties (Fig. 7). Mice were given diluent, infected with a lentiviral vector alone, or lentivirus encoding LPCAT1 for 96 h prior exposure to LPS for 24 h. In mice given LPS, immunoblotting revealed that LPCAT1 was significantly decreased in lungs compared with untreated mice as revealed by densitometry (p = 0.02, n = 12) (Fig. 7A, upper row). Importantly, LPCAT1 gene transfer significantly increased immunoreactive LPCAT1 content and partially restored these levels in mice given LPS (Fig. 7A, lower row). LPS significantly impaired respiratory mechanics as it increased airway resistance and elastance (lung stiffness) (Fig. 7, B and D), reduced compliance (Fig. 7C), and markedly reduced surfactant phospholipid levels (Fig. 7F). LPCAT1 lentiviral gene transfer partially reversed these adverse effects on lung mechanics in mice (Fig. 7, B–E) that was due, in part, to increased surfactant levels (Fig. 7F). Thus, LPS destabilizes LPCAT1 in vivo leading to reduced surfactant lipid and impaired lung functional properties that can be lessened with gene transfer of the phospholipid remodeling enzyme.