The fatty acid 2-hydroxylase (FA2H) is essential for synthesis of 2-hydroxylated fatty acids in myelinating and other cells, and deficiency of this enzyme causes a complicated form of hereditary spastic paraplegia also known as fatty acid hydroxylase-associated neurodegeneration. Despite its important role in sphingolipid metabolism, regulation of FA2H and its interaction with other proteins involved in the same or other metabolic pathways is poorly understood. To identify potential interaction partners of the enzyme, quantitative mass spectrometry using stable isotope labeling of cells was combined with formaldehyde cross-linking and proximity biotinylation, respectively. Besides other enzymes involved in sphingolipid synthesis and intermembrane transfer of ceramide, and putative redox partners of FA2H, progesterone receptor membrane component-1 (PGRMC1) and PGRMC2 were identified as putative interaction partners. These two related heme-binding proteins are known to regulate several cytochrome P450 enzymes. Bimolecular fluorescence complementation experiments confirmed the interaction of FA2H with PGRMC1. Moreover, the PGRMC1 inhibitor AG-205 significantly reduced synthesis of hydroxylated ceramide and glucosylceramide in FA2H-expressing cells. This suggests that PGRMC1 may regulate FA2H activity, possibly through its heme chaperone activity.

Introduction

Fatty acid 2-hydroxylase (FA2H; EC 1.14.18.6), an enzyme of the endoplasmic reticulum (ER), is required for synthesis of 2-hydroxylated sphingolipids in the nervous system and other organs (for review, see [1]). Although free fatty acids are substrates for FA2H in vitro [2,3], the crystal structure of the yeast homolog (SCS7) strongly suggests that ceramides are the physiological substrates of the enzyme [4]. Deficiency of FA2H in humans causes a complicated form of hereditary spastic paraplegia, SPG35, also known as fatty acid hydroxylase-associated neurodegeneration (FAHN) [5,6], which can be associated with leukodystrophy and brain iron accumulation [6,7]. Mice deficient in FA2H develop late-onset axon and myelin degeneration [8] and show deficits in learning and memory [9]. How absence of FA2H causes the symptoms observed in FAHN or FA2H-deficient mice, as well as the physiological role of 2-hydroxylated sphingolipids is largely unknown.

Synthesis of 2-hydroxylated sphingolipids requires functional interaction of FA2H with several other enzymes (enzymes for the transfer of electrons to the cytochrome b5 domain of FA2H, as well as the enzymes involved in ceramide synthesis). In the present study, we addressed the question whether these functional interactions are reflected by physical interactions. Binding of enzymes involved in the same metabolic pathways has been shown in several studies. Examples in the area of sphingolipid metabolism include the interaction of glycosyltransferases in the Golgi apparatus [10], nucleotide-sugar transporters and UDP-galactose:ceramide galactosyltransferase (EC 2.4.1.45) in the ER [11], and the interaction of fatty acid elongases (EC 6.2.1.3), 3-ketoacyl-CoA reductases (EC 1.1.1.330) and trans-2,3-enoyl-CoA reductase (EC 1.3.1.38) with ceramide synthase 2 (CerS2; EC 2.3.1.24) [12].

Besides coordination of enzymes in the same metabolic pathway, physical interactions may be also important for the regulation of enzymes and the coordination of different metabolic pathways. For example, the physical interaction between sphingolipids and sterols is well documented and the reciprocal interference of their metabolic pathways was clearly shown in several studies (see ref. [13] and references therein). A subunit of the serine palmitoyltransferase (EC 2.3.1.50) complex, serine palmitoyltransferase long chain-1 (SPTLC1), binds to the cholesterol transporter ABCA1 and inhibits its activity [14]. Functional interactions of FA2H with other (metabolic) pathways are suggested by significant alterations in wax ester synthesis in sebaceous glands in FA2H-deficient mice [15], but also from studies in plants showing interaction of the Arabidopsis thaliana FA2H homolog At-FA2H with the Bax inhibitor [16].

To identify functionally important interaction partners of the mammalian FA2H, we performed a mass spectrometry screen. Because initial experiments provided no evidence for strong and stable interactions of FA2H, we used two complementary approaches that are able to reveal also weak interactions : formaldehyde cross-linking [17] and proximity biotinylation using promiscuous biotin protein ligase BirA* [biotin protein ligase BirA mutant (R118G); EC 6.3.4.15] (BioID; [18]), both in combination with stable isotope labeling with amino acids in cell culture (SILAC) [19]. To confirm the interaction of FA2H with identified proteins, we used bimolecular fluorescence complementation (BiFC) experiments. We could identify progesterone receptor membrane component-1 (PGRMC1) as an FA2H interaction partner. Moreover, the PGRMC1 inhibitor AG-205 reduces synthesis of 2-hydroxylated sphingolipids in FA2H-expressing cells, suggesting that PGRMC1 may be involved in regulating FA2H activity.

Experimental procedures

Plasmids

All oligonucleotides used in this work (Table 1) were obtained from MWG/Eurofins Genomics (Ebersberg, Germany). A mammalian expression vector for FA2H carrying an N-terminal Twin-StrepTag (Twin-Strep-FA2H) was generated using the StarGate combinatorial cloning system (IBA Lifesciences, Göttingen, Germany). For this, FA2H was amplified by PCR (using primers Twin-Strep-FA2HFW and Twin-Strep-FA2HRV; see Table 1), inserted into the entry vector pENTRY-IBA51 and finally shuttled into the acceptor vector pESG-IBA105, following the instructions of the manufacturer. For BioID assays, FA2H was amplified by PCR (using primers BioID-FA2H(KpnI)RV and BioID-FA2H(EcoRI)FW) and subcloned via KpnI and EcoRI restriction sites into the mammalian BioID expression vector pcDNA3.1-mycBioID [18] [pcDNA3.1 mycBioID was a gift from Kyle Roux (Addgene plasmid #35700)] C-terminally of myc-tagged BirA*(R118G). For BiFC assays, cDNAs for FA2H, PGRMC1, acyl-CoA synthetase long-chain family member 3 (ACSL3) and SPTLC1, amplified using oligonucleotides shown in Table 1, were subcloned (using EcoRI and ClaI restriction sites) into vectors adding a C-terminal citrine YFP fragment tag (cYFP1 or YFP2; [20]) to the expressed proteins. The parent vectors, pcDNA3-MCFD2-cYFP1 and pcDNA3-MCFD2-YFP2, were generously provided by Dr Veronika Reiterer (Biotechnologie Institut Thurgau, Switzerland). A control plasmid encoding FLAG-tagged mouse Rimklb has been described previously [21].

Table 1
Oligonucleotides used in the present study
Oligonucleotide Sequence 
BiFC-FA2H_C(ClaI)RV TCATCGATCTGCATCTTCGGGTGGGC 
BiFC-FA2H_C(EcoRI)FW ATGAATTCATGGCCCCCGCTCCGC 
BiFC-ACSL3_C(ClaI)RV TCATCGATTTTTCTTCCATACATTCGCTCAA 
BiFC-ACSL3_C(EcoRI)FW ATGAATTCATGAATAACCACGTGTCTTCAAA 
BiFC-PGRMC1_C(ClaI)RV TCATCGATATCATTTTTCCGGGCACTCTCATC 
BiFC-PGRMC1_C(EcoRI)FW ATGAATTCATGGCTGCCGAGGATGTGGT 
BiFC-SPTLC1_C(ClaI)RV TTATCGATGAGCAGGACGGCCTGGG 
BiFC-SPTLC1_C(EcoRI)FW CTGAATTCATGGCGACCGCCACGGAG 
BioID-FA2H(KpnI)RV TGGGTACCTCACTGCATCTTCGGGTG 
BioID-FA2H(EcoRI)FW ATGAATTCATGGCCCCCGCTCCGC 
Twin-Strep-FA2HFW AGCGGCTCTTCAATGGCCCCCGCTCCGCCC 
Twin-Strep-FA2HRV AGCGGCTCTTCTCCCCTGCATCTTCGGGTGGGC 
Oligonucleotide Sequence 
BiFC-FA2H_C(ClaI)RV TCATCGATCTGCATCTTCGGGTGGGC 
BiFC-FA2H_C(EcoRI)FW ATGAATTCATGGCCCCCGCTCCGC 
BiFC-ACSL3_C(ClaI)RV TCATCGATTTTTCTTCCATACATTCGCTCAA 
BiFC-ACSL3_C(EcoRI)FW ATGAATTCATGAATAACCACGTGTCTTCAAA 
BiFC-PGRMC1_C(ClaI)RV TCATCGATATCATTTTTCCGGGCACTCTCATC 
BiFC-PGRMC1_C(EcoRI)FW ATGAATTCATGGCTGCCGAGGATGTGGT 
BiFC-SPTLC1_C(ClaI)RV TTATCGATGAGCAGGACGGCCTGGG 
BiFC-SPTLC1_C(EcoRI)FW CTGAATTCATGGCGACCGCCACGGAG 
BioID-FA2H(KpnI)RV TGGGTACCTCACTGCATCTTCGGGTG 
BioID-FA2H(EcoRI)FW ATGAATTCATGGCCCCCGCTCCGC 
Twin-Strep-FA2HFW AGCGGCTCTTCAATGGCCCCCGCTCCGCCC 
Twin-Strep-FA2HRV AGCGGCTCTTCTCCCCTGCATCTTCGGGTGGGC 

Cell culture and transfection

HEK-293 and HEK-293T cells were cultivated in Dulbecco's modified Eagle's medium (DMEM) with GlutaMAX supplemented with 10% fetal bovine serum, 100 U/ml penicillin and 100 μg/ml streptomycin (all from Life Technologies, Darmstadt, Germany) at 5% CO2 and 37°C. BHK cells were cultivated in DMEM–GlutaMAX supplemented with 5% fetal bovine serum, 100 U/ml penicillin and 100 µg/ml streptomycin (all from Life Technologies) at 5% CO2 and 37°C.

For SILAC labeling, HEK-293 and HEK-293T cells were cultivated in DMEM for SILAC supplemented with 10% dialyzed fetal bovine serum for SILAC (both from Thermo Scientific, Waltham, MA, U.S.A.), 2 mM glutamine, 100 U/ml penicillin and 100 µg/ml streptomycin (all from Life Technologies). In addition, the media were supplemented with either a light (natural lysine and arginine) or a heavy amino acid mixture (15N213C6-lysine and 15N413C6-arginine) (both from Cambridge Isotope Laboratories, Andover, U.S.A.). Incorporation of isotopes was allowed for at least six cell divisions. The successful label incorporation was assessed by MALDI-TOF/TOF.

HEK-293 and HEK-293T cells were transfected with either the calcium phosphate method [22] or with TurboFect (Thermo Scientific). For calcium phosphate transfections, the cells were seeded 1 day prior to transfection at a density of 3 × 106 per 10 cm dish. The transfection mixture was prepared by first mixing 10 µg of plasmid DNA together with 65 µl of 2 M CaCl2 in a total volume of 500 µl in sterile H2O. Then, this mixture was combined with the same volume of 2× HBS (50 mM HEPES, 280 mM NaCl, 1.5 mM Na2HPO4, pH 7.0). After 30 min incubation at room temperature, the formed precipitates were evenly spread onto the cells. The cells were then incubated at 37°C and in 5% CO2 for 24 or 48 h. TurboFect transfections were performed according to the manufacturer's protocol.

Metabolic labeling and lipid analysis

CHO-K1 cells were maintained in DMEM:Nut Mix F12 (1 : 1) supplemented with 1% FCS and 2 mM l-glutamine. For transfection, cells were seeded in six-well plates at a density of 250 000 cells per well 20–24 h before transfection in medium containing 2.5% FCS. Transfection was done using Lipofectamine LTX (Thermo Scientific) according to the manufacturer's instruction. Six hours after transfection, 10 µM of AG-205 (Sigma; dissolved in DMSO) and/or 20 µM of hemin (Sigma; dissolved in DMSO) or DMSO were added. Cells were lysed 25 h later (31 h after transfection) in lysis buffer [50 mM Tris–HCl (pH 8.0), 100 mM NaCl, 0.5% NP-40, 2 mM EDTA, 1× HALT protease inhibitor mix; Thermo Scientific] for western blot analysis. Alternatively, cells were metabolically labeled for 24 h with [U-14C]-l-serine (37 kBq per 35 mm well), 7 h after transfection (and 1 h after treatment with AG-205 and/or hemin). Lipids were extracted according to ref. [23] and separated by thin layer chromatography (TLC) on silica gel 60 HPTLC plates (Merck, Darmstadt, Germany). To separate non-hydroxy-VLCFA, non-hydroxy-LCFA and 2-hydroxylated glucosylceramide (GlcCer), TLC plates were developed twice in chloroform/methanol/water (144 : 25 : 2.8; v/v/v) [24]. Radioactive signals were visualized using Bioimager screens (Fuji) and quantified using the AIDA software (Elysia-raytest GmbH, Straubenhardt, Germany).

Formaldehyde cross-linking of protein complexes

Formaldehyde cross-linking of proteins was based on the protocol described by Klockenbusch and Kast [25]. Briefly, cells expressing the bait protein of interest were pelleted and washed thrice with 1× PBS to remove all cell culture media remains. Then, the cells were resuspended in 1 ml of 37°C pre-warmed paraformaldehyde (PFA) solution (0.25–1% in 1× PBS) per 1 × 107 cells and incubated at room temperature under constant agitation for 15 min. Subsequently, cells were pelleted at 2000×g for 5 min and the cross-linking stopped by resuspension in 1 ml of ice-cold stopping solution (1.25 M glycine in PBS) per 1 × 107 cells. Pelleting and washing the cells once again with stopping solution completed the procedure. Afterwards, the cell pellets were either stored at −20°C or lysed directly.

In vivo biotinylation of FA2H-interacting proteins

Cells previously transfected with BirA* constructs were grown in a culture medium without biotin for 24 h. To induce biotinylation, 50 µM biotin was added to the cell medium, followed by another 24 h of incubation. Afterwards, the cells were harvested and either stored at −20°C or immediately used for experiments.

Cell lysis

Cell disruptions were carried out at 4°C. The cells were lysed in different buffers depending on the experiment performed: StrepTag buffer [1% TX-100 (v/v), 7.5% glycerol (v/v), 50 mM Tris–HCl, 150 mM NaCl, 2 mM CaCl2, 2 mM MgCl2, pH 7.4] for StrepTag experiments without PFA, RIPA buffer (0.1% SDS, 0.5% sodium deoxycholate, 1% Nonidet-P40, 150 mM NaCl, 50 mM Tris, pH 8.0) for StrepTag experiments with PFA cross-linking and BioID buffer [2% (v/v) Triton X-100, 0.4% (w/v) SDS, 500 mM NaCl, 5 mM EDTA, 1 mM DTT, 50 mM Tris, pH 7.4] for BioID experiments. All buffers were supplemented with HALT protease inhibitor cocktail (Thermo Scientific). The cells were mixed with the appropriate amount of lysis buffer and resuspended by pipetting. Afterwards, cell lysates were incubated for 30 min on an overhead tumbling incubator at 4°C, followed by 30 min centrifugation at 18 000×g and 4°C. The lysates were then transferred to pre-cooled microtubes. For PFA-treated cells, lysed in RIPA, two steps were added after the 30 min incubation: an additional homogenization step (Dounce homogenizator 1 ml, tight-fitting, 25 strokes) followed by another 15 min incubation on a tumbling incubator. Protein concentrations were determined using the Bio-Rad DC Assay (Bio-Rad, Hercules, U.S.A.) following the manufacturer's instructions.

StrepTactin-affinity purification, SDS–PAGE and in-gel digestion

After mixing equal protein amounts of cell lysates from control (pcDNA3) and Twin-Strep-mFA2H (pESG-IBA105-FA2H)-transfected cells, StrepTag-affinity purification was performed at 4°C by overnight incubation with 50 µl (50% slurry) of StrepTactin-Macroprep beads (IBA Lifesciences). Afterwards, beads were washed three times for 5 min with either 1 ml of Triton X-100 or RIPA buffer. Proteins were eluted three times with 50 µl of wash buffer supplemented with 2 mM biotin. Pooled eluates from the StrepTag-affinity purification were precipitated by chloroform/methanol precipitation [26], resuspended in 30 µl of reducing 2× Laemmli buffer, sonicated for 5 min and heated to 95°C for 10 min. Finally, proteins were alkylated by the addition of 56 mM acrylamide and incubation for 30 min at room temperature. Then, proteins were separated by SDS–PAGE and visualized by colloidal Coomassie blue staining [27]. Each lane was cut into 4–6 pieces (depending on the sample complexity) of approximately equal protein amounts and subjected to tryptic in-gel digestion including peptide extraction steps [28].

Neutravidin affinity purification of biotinylated proteins and on-bead digestion

Cell lysates of equal protein amount from BirA* (pcDNA3-1-mycBioID) or BirA*-FA2H (pcDNA3.1-mycBioID-FA2H) expressing cells were separately subjected to Biotin Affinity Purification. This was achieved overnight at 4°C using NeutrAvidin-agarose beads (Thermo Scientific). Briefly, 200 µl of 50% bead slurry was used per cell lysate derived from one confluent 15 cm cell culture plate. After incubation, the beads were washed with four different buffers to reduce unspecific binding, as described previously [18]. First, the beads were washed twice with 2% SDS and once with 0.1% sodium deoxycholate, 1% TX-100, 500 mM NaCl, 1 mM EDTA, 50 mM HEPES, pH 7.5, followed by a wash with 250 mM LiCl, 0.5% NP-40, 0.5% sodium deoxycholate, 1 mM EDTA, 10 mM Tris–HCl (pH 8.1), and finally washed twice with 50 mM Tris and 50 mM NaCl (pH 7.4). Since the NeutrAvidin–biotin bond is too strong for efficient and complete protein elution, and peptides for mass spectrometry analysis were generated by on-bead digestion. Accordingly, the beads were first treated with 10 mM DTT in 100 mM NH4HCO3 and heated to 56°C for 30 min to accomplish protein reduction. Afterwards, the reduced disulfide bridges were alkylated by adding 56 mM acrylamide and incubating at room temperature for 30 min. Following washing three times with 100 mM NH4HCO3, trypsin was added to achieve protein digestion overnight at 37°C. The resulting peptide solutions were recovered and still bead-bound peptides re-extracted twice with 5% ACN and 0.1% FA. Now, the digests from BirA* and BirA*-FA2H were combined at a 1 : 1 ratio (v/v) and analyzed by LC–MS/MS (liquid chromatography coupled to tandem mass spectrometry).

LC–MS/MS peptide analysis

All peptide samples were first desalted using self-prepared C18 StageTips [29]. Afterwards, they were dried in a speed-vac and finally dissolved in 5% FA, 5% ACN and 90% water (v/v/v). Peptides were analyzed by an LC–MS/MS set-up consisting of an EASY-nLC 1000 nano LC (Thermo Scientific) coupled to an Orbitrap Velos hybrid mass spectrometer (Thermo Scientific). Peptide separations were achieved on C18-reversed-phase capillary columns, packed in-house (Magic C18, 5 µm, 100 µm × 150–200 mm, Bruker-Michrom), via linear 30 min (BioID) or 60 min (Twin-Strep pulldowns) gradients (1–35% solvent B) consisting of defined mixtures of solvent A [0.1% FA, 5% DMSO, 94.1% water (v/v/v)] and B [0.1% FA, 5% DMSO, 94.1% ACN (v/v/v)] at a flow rate of 400 nl/min [30]. The mass spectrometer was set to acquire one Orbitrap survey scan (scan range: 400–1200 m/z, resolution: 30 000, AGC target: 1 × 106, maximum fill time: 300 ms) followed by 10 data-dependent product ion scans in the LTQ ion trap (normal scan rate, AGC target: 1 × 104, maximum fill time: 100 ms). 2+, 3+ and 4+ charged ions were selected for data-dependent CID fragmentation (norm. coll. energy: 35%, act. q = 0.25, act. time = 10 ms) with dynamic exclusion enabled (excl. duration: 30 s, parent mass accuracy: 10 ppm). An instrument lock mass of 401.92 m/z (DMSO) was used.

MS data analysis

MS/MS spectra were searched against the human UniProtKB database and a common contaminants database, using the MaxQuant 1.5.3.8 software package with its built-in Andromeda search algorithm (http://MaxQuant.org) [31,32]. The algorithm was set to use trypsin as a proteolytic enzyme (Trypsin/P) allowing up to two missed cleavages and assuming propionamide as a fixed cysteine modification (+71.03711 Da). For the PFA-cross-linking experiment, the variable modifications were: oxidation (M, +15.99491 Da), acetyl (protein N-Term, +42.01056 Da), deamidation (NQ, +0.98401 Da) and cation: Fe[III] (DE, +52.91146 Da). In the case of the BioID experiment data, the same variable modifications were selected, except for cation: Fe[III] (DE, +52.91146 Da). Mass tolerance was set to 20/4.5 ppm (first/main search) for MS and 0.5 Da for MS/MS. A false discovery rate of 1%, determined by searching against a reverse database, was set both for peptides and proteins. Moreover, a minimal Andromeda score of 40 and a delta score of 6 were used for modified peptides. Both the match between run (alignment window: 20 min, match window: 0.5 min) and re-quantify option were enabled. Quantification of SILAC pairs (heavy labels: Arg+10, Lys+8) was performed with the default MaxQuant settings with a minimum ratio count of 2. Perseus 1.5.2.6 (http://www.perseus-framework.org) [33] was used for the statistical evaluation and visualization of the protein results. Protein groups identified by MaxQuant as contaminants or reverse hits first filtered out. Subsequently, the normalized H/L-expression ratios of each protein group, as provided by MaxQuant, were linearized by log2 transformation. Only protein groups with SILAC values in all three experiments were further evaluated, and a one-sample t-test with Benjamini–Hochberg P-value correction (false discovery rate [FDR] = 0.05) was performed to find significantly changed protein groups. Proteins enriched at least three-fold in the FA2H bait sample were selected as putative FA2H interaction partners. Subcellular localization of identified proteins was determined manually using information provided at the UniProt database (http://www.uniprot.org).

Western blot analyses

Following SDS–PAGE separation, protein samples were transferred onto nitrocellulose membranes using a semi-dry set-up. After transfer, membranes were blocked with skim milk followed by primary and secondary antibody incubations, with the secondary antibodies either linked to HRP or fluorescent dyes. The following primary antibodies were used: rabbit anti-FA2H (dilution 1 : 2000; [15]), rabbit anti-SPTLC1 (1 : 2000; [34], kindly provided by Prof. Thorsten Hornemann, University Hospital Zurich), rabbit anti-PGRMC1 (1 : 1000, Abcam, Cambridge, U.K., ab80941, lot number 887309), rabbit anti-ACSL3 (1 : 1000, Thermo Scientific, PA5-29507, lot number OK17866023) and mouse anti-α-tubulin (1 : 40 000; Sigma–Aldrich, T5168, lot number 103M4773V). Secondary antibodies were: bovine anti-goat Ig-peroxidase conjugate (1 : 5000, product number 805-035-180, Dianova, Hamburg, Germany), goat anti-rabbit Ig-peroxidase conjugate (1 : 5000, product number 111-035-003, Dianova), goat anti-mouse Ig-peroxidase conjugate (1 : 5000, product number 115-035-044, Dianova), donkey anti-goat Ig-DyLight800 (1 : 10 000, product number SA5-10092, Thermo Scientific), goat anti-rabbit Ig-DyLight650 (1 : 10 000, product number SA5-10034, Thermo Scientific), goat anti-rabbit Ig-DyLight800 (1 : 10 000, product number SA5-35571, Thermo Scientific) and goat anti-mouse Ig-DyLight800 (1 : 10 000, product number 35521, Thermo Scientific). Twin-Strep-FA2H was detected using a StrepTactin-HRP conjugate (1 : 5000, product number 2-1502-001, IBA Lifesciences). Bound secondary antibodies were visualized using Pierce™ ECL Western Blotting Substrate (Thermo Scientific) in combination with a CCD camera system (Vilber Lourmat, Eberhardzell, Germany). Quantification of protein bands was achieved using the system standard software package (FUSION-Capt Advance Solo 4).

Bimolecular fluorescence complementation

BHK or CHO-K1 cells were transfected with the respective plasmids combinations and grown for 48 h. Transfection mixtures were composed of 10% cYFP1 plasmid, 10% YFP2 plasmid and 80% empty vector, pcDNA3. For BiFC-competition experiments, plasmids encoding cYFP1- and YFP2-tagged FA2H, PGRMC1 and SPTLC1 (10% each) were co-transfected with plasmids encoding non-tagged FA2H, PGRMC1 or SPTLC1 (80%) or, as controls, plasmids encoding irrelevant proteins (pFLAG-Nat8l or pFLAG-Rimklb). Subsequently, either immunofluorescence microscopy was performed for qualitative analyses or the fluorescence (BiFC signal) was measured with a plate reader for quantitative analyses. For the latter, the cells were first detached using trypsin, washed two times in 1× PBS, resuspended in 200 µl of 1× PBS and transferred to a black 96-well plate. Afterwards, the measurement was performed using one of the following devices: (A) Mithras LB 940 plate reader (Berthold Technologies, Bad Wildbad, Germany) equipped with an excitation filter of 510 and an emission filter of 535 nm. (B) Infinite 200 PRO (TECAN, Crailsheim, Germany) equipped with a monochromator set to an excitation of 505 nm and emission of 540 nm. As a positive control, cells were co-transfected with cYFP1- and YFP2-tagged ER–Golgi intermediate compartment 53 kDa protein (ERGIC53), a lectin that oligomerizes in the ER and is known to produce a strong BiFC signal [35].

Statistics

Data were analyzed using (paired or unpaired) t-test or ANOVA with post hoc Tukey HSD test, as indicated.

Results

Screening for FA2H interactors by quantitative mass spectrometry using formaldehyde cross-linking and proximity biotinylation

To identify putative interaction partners of FA2H, we initially performed co-immunoprecipitation experiments in which Twin-Strep-FA2H (Figure 1A) from transiently transfected HEK-293 cells was immunoprecipitated and co-purified proteins identified by mass spectrometry. By comparing these identifications with the ones obtained in a parallel-performed control purification missing the bait protein, interacting proteins are expected to be enriched in the bait sample. To confidently determine differences in protein abundance between bait and control samples, SILAC was employed for these and all further experiments described below. This approach, however, did not show any significant enrichment of proteins in the bait-expressing cells (data not shown). These results suggest that FA2H interactions are probably weak and transient and interaction partners may get lost during cell lysis and affinity purification. We therefore modified the MS-based strategy and used two different approaches that are able to recognize also weak and transient interactions. In the formaldehyde cross-linking (PFA) method, cells were treated with formaldehyde prior to cell lysis and FA2H precipitation (Figure 1B). The second approach (Figure 1C) depends on proximity-based biotinylation using the constitutively active Escherichia coli biotin ligase BirA mutant BirA* (R118G) [18] fused to the amino-terminus of FA2H (Figure 1A). The addition of biotin to cells expressing this construct leads to the biotinylation of proteins interacting with or present in close proximity to the BirA*-FA2H fusion protein and thus allows their selective enrichment by Biotin Affinity Purification (for review, see ref. [36]).

Mass spectrometry-based approaches used to identify FA2H interaction partners.

Figure 1.
Mass spectrometry-based approaches used to identify FA2H interaction partners.

(A) Schematic structure of the two FA2H bait constructs, Twin-Strep-FA2H and BirA*-FA2H. (B) Experimental workflow for the PFA-assisted SILAC Twin-Strep-FA2H pull-down approach. (C) Experimental workflow for the SILAC BioID approach. FA2H: fatty acid 2-hydroxylase; Twin-Strep: Twin-StrepTag-affinity tag; Cytb5: cytochrome b5 domain; TM: transmembrane domain; myc: myc-tag (epitope of the 9E10 anti-myc antibody); BirA*: constitutively active E. coli biotin ligase mutant; PFA: paraformaldehyde; SILAC: stable isotope labeling of amino acids in cell culture; light: cells labeled with light SILAC medium; heavy: cells labeled with heavy SILAC medium; GeLC–MS/MS: in-gel tryptic digestion followed by liquid chromatographic separation and tandem mass spectrometry; OFFGEL: isoelectric gel-based peptide fractionation system; LC–MS/MS: liquid chromatography coupled to tandem mass spectrometry.

Figure 1.
Mass spectrometry-based approaches used to identify FA2H interaction partners.

(A) Schematic structure of the two FA2H bait constructs, Twin-Strep-FA2H and BirA*-FA2H. (B) Experimental workflow for the PFA-assisted SILAC Twin-Strep-FA2H pull-down approach. (C) Experimental workflow for the SILAC BioID approach. FA2H: fatty acid 2-hydroxylase; Twin-Strep: Twin-StrepTag-affinity tag; Cytb5: cytochrome b5 domain; TM: transmembrane domain; myc: myc-tag (epitope of the 9E10 anti-myc antibody); BirA*: constitutively active E. coli biotin ligase mutant; PFA: paraformaldehyde; SILAC: stable isotope labeling of amino acids in cell culture; light: cells labeled with light SILAC medium; heavy: cells labeled with heavy SILAC medium; GeLC–MS/MS: in-gel tryptic digestion followed by liquid chromatographic separation and tandem mass spectrometry; OFFGEL: isoelectric gel-based peptide fractionation system; LC–MS/MS: liquid chromatography coupled to tandem mass spectrometry.

The treatment of cells expressing Twin-Strep-tagged FA2H with formaldehyde leads to formation of high molecular mass products as demonstrated by western blotting (Figure 2A). For the mass spectrometric screen, ‘SILAC-labeled’ HEK-293T cells were transiently transfected with Twin-Strep-tagged FA2H (‘heavy’ labeled) or empty vector (‘light’ labeled) and treated with 0.5% formaldehyde, and FA2H was isolated from cell lysates using Streptactin beads. Bound proteins were digested with trypsin and peptides were analyzed by LC–MS/MS to identify cross-linked proteins. In total, 676 protein groups were identified in all three experiments (Figure 3B). Using an at least three-fold enrichment as the criterion for putative FA2H interactors (with an FDR of 0.05), 43 candidate proteins were identified (Figure 3B and Table 2). As expected, the majority of identified proteins are ER proteins (with few additional proteins localized to the nuclear membrane, Golgi apparatus, plasma membrane and endosomes) (see Table 2). In the BioID approach, biotinylated proteins from cells transiently expressing FA2H N-terminally fused to the BirA* biotin ligase were compared with cells expressing the cytosolic BirA* ligase (Figure 2B). Using this approach, 9 of 400 identified proteins were reproducibly enriched at least three-fold (FDR = 0.05), all of which localized to the ER or nuclear membrane (Figure 3A and Table 3). Seven of these proteins were also found in the PFA screen (indicated by bold letters in Tables 2 and 3).

Western blot analysis of PFA-cross-linking and proximity biotinylation experiments.

Figure 2.
Western blot analysis of PFA-cross-linking and proximity biotinylation experiments.

(A) Western blot analysis of a PFA-cross-linking experiment. HEK-293T cells were transiently transfected with a plasmid encoding Twin-Strep-FA2H or the empty vector (control). After fixation/protein cross-linking with 0.5% PFA for 20 min, cells were lysed in RIPA buffer and subjected to affinity purification using StrepTactin beads. Aliquots of each fraction obtained from the purification protocol were analyzed by western blotting with anti-StrepTag. The percentage rates indicated the amounts (percent of total) loaded from each fraction. P, 20 000×g pellet of lysate; L, 20 000×g supernatant of lysate; NB, non-bound fraction; W1–W3, wash fractions; E1–E3, eluate fractions; B, proteins remaining bound to the beads after elution. (B) Western blot analysis of a representative proximity biotinylation (BioID) experiment. HEK-293T cells were transiently transfected with a plasmid encoding Twin-Strep-FA2H or the empty vector (mock). pcDNA3.1-mycBioID (BirA*) or pcDNA3.1-mycBioID-FA2H (BirA*-FA2H). After the addition of biotin (50 µM) for 24 h, cells were lysed in StrepTag lysis buffer and examined by western blotting with the streptavidin-peroxidase conjugate.

Figure 2.
Western blot analysis of PFA-cross-linking and proximity biotinylation experiments.

(A) Western blot analysis of a PFA-cross-linking experiment. HEK-293T cells were transiently transfected with a plasmid encoding Twin-Strep-FA2H or the empty vector (control). After fixation/protein cross-linking with 0.5% PFA for 20 min, cells were lysed in RIPA buffer and subjected to affinity purification using StrepTactin beads. Aliquots of each fraction obtained from the purification protocol were analyzed by western blotting with anti-StrepTag. The percentage rates indicated the amounts (percent of total) loaded from each fraction. P, 20 000×g pellet of lysate; L, 20 000×g supernatant of lysate; NB, non-bound fraction; W1–W3, wash fractions; E1–E3, eluate fractions; B, proteins remaining bound to the beads after elution. (B) Western blot analysis of a representative proximity biotinylation (BioID) experiment. HEK-293T cells were transiently transfected with a plasmid encoding Twin-Strep-FA2H or the empty vector (mock). pcDNA3.1-mycBioID (BirA*) or pcDNA3.1-mycBioID-FA2H (BirA*-FA2H). After the addition of biotin (50 µM) for 24 h, cells were lysed in StrepTag lysis buffer and examined by western blotting with the streptavidin-peroxidase conjugate.

MS-based proteomic analyses reveal FA2H-interacting proteins.

Figure 3.
MS-based proteomic analyses reveal FA2H-interacting proteins.

Volcano plots depicting the fold change and P-values (one-sample t-test against 0) of identified protein groups using SILAC BioID (A) and PFA-assisted SILAC Twin-Strep-FA2H pulldown (B). Only data of protein groups identified in all replicates are shown (BioID: 400; PFA: 676). All proteins enriched significantly at least three-fold in the FA2H-expressing samples with an FDR of 0.05 (Benjamini–Hochberg correction for multiple comparisons) are indicated by black triangles. The gene names of all significantly enriched proteins are shown for the BioID screen (A) (see also Table 3). (B) For the PFA screen only, some gene names are given (see Table 2 for the complete list).

Figure 3.
MS-based proteomic analyses reveal FA2H-interacting proteins.

Volcano plots depicting the fold change and P-values (one-sample t-test against 0) of identified protein groups using SILAC BioID (A) and PFA-assisted SILAC Twin-Strep-FA2H pulldown (B). Only data of protein groups identified in all replicates are shown (BioID: 400; PFA: 676). All proteins enriched significantly at least three-fold in the FA2H-expressing samples with an FDR of 0.05 (Benjamini–Hochberg correction for multiple comparisons) are indicated by black triangles. The gene names of all significantly enriched proteins are shown for the BioID screen (A) (see also Table 3). (B) For the PFA screen only, some gene names are given (see Table 2 for the complete list).

Table 2
Putative FA2H interaction partners identified by formaldehyde cross-linking
Gene name Protein name Function/metabolic pathway Fold change P-value 
UBB/UBC Polyubiquitin B/C Protein degradation/ERAD 10.3 0.00139 
LRRC59 Leucine-rich repeat-containing protein 59 Nuclear import 9.0 0.00164 
ERLIN2 Erlin-2 Protein degradation/ERAD 8.8 0.00063 
GOLGA2 Golgin subfamily A member 2 Vesicle docking/Golgi structure 8.4 0.00091 
EMC2 ER membrane protein complex subunit 2 Protein folding 82 0.00226 
SRPRB Signal recognition particle receptor subunit β ER protein translocation 7.6 0.00112 
ESYT1 Extended synaptotagmin-1 Plasma membrane-ER contact sites 7.4 0.00173 
CKAP4 Cytoskeleton-associated protein 4 Anchoring ER to microtubles 7.1 0.00015 
SEC22B Vesicle-trafficking protein SEC22B Vesicular transport 7.0 0.00350 
SPTLC1 Serine palmitoyltransferase 1 Sphingolipid synthesis 6.7 0.00082 
SPTLC2 Serine palmitoyltransferase 2 Sphingolipid synthesis 6.6 0.00192 
ERLIN1 Erlin-1 Protein degradation/ERAD 6.1 0.00147 
POR NADPH cytochrome P450 reductase Electron transfer to heme 5.9 0.00232 
RPN1 Dolichyl-diphosphooligosaccharide protein glycosyltransferase subunit N-glycosylation 5.6 0.00018 
CCDC47 Coiled-coil domain-containing protein 47 Protein degradation/ERAD 5.5 0.00134 
MIA3 Melanoma inhibitory activity protein 3 Vesicular trafficking 5.4 0.00106 
ATP2A2 Sarcoplasmic/endoplasmic reticulum calcium ATPase 2 Calcium transport into ER 5.3 0.00059 
VAPA Vesicle-associated membrane protein-associated protein A ER–Golgi contact sites 5.2 0.00108 
VAPB Vesicle-associated membrane protein-associated protein B ER–Golgi contact sites 5.1 0.00268 
SLC27A4 Long-chain fatty acid transport protein 4 Lipid metabolism 5.1 0.00142 
TMED9 Transmembrane emp24 domain-containing protein 9 Vesicular trafficking 4.7 0.00417 
RDH11 Retinol dehydrogenase 11 Lipid metabolism 4.6 0.00345 
TFRC Transferrin receptor protein 1 Iron metabolism 4.6 0.00351 
EPHX1 Epoxide hydrolase 1 Biotransformation 4.5 0.00292 
SSR4 Translocon-associated protein subunit γ ER protein translocation 4.3 0.00293 
RAB6A Ras-related protein Rab-6A Vesicle transport 4.2 0.00139 
NSDHL Sterol-4-α-carboxylate 3-dehydrogenase. decarboxylating Sterol metabolism 4.1 0.00021 
SCFD1 Sec1 family domain-containing protein 1 Trafficking ECM components 4.0 0.00421 
ZW10 Centromere/kinetochore protein zw10 homolog Vesicular trafficking 4.0 0.00437 
ALDH3A2 Fatty aldehyde dehydrogenase Lipid metabolism 3.9 0.00129 
DDOST Dolichyl-diphosphooligosaccharide protein glycosyltransferase subunit N-glycosylation 3.8 0.00036 
DHCR7 7-Dehydrocholesterol reductase Sterol metabolism 3.8 0.00306 
CANX Calnexin Protein folding 3.8 0.00462 
HACD3 Very long-chain (3R)-3-hydroxyacyl-CoA dehydratase 3 Fatty acid elongation 3.5 0.00532 
RAB7A Ras-related protein Rab-7A Vesicle transport 3.4 0.00377 
FAF2 FAS-associated factor 2 Protein degradation/ERAD 3.4 0.00395 
PGRMC1 Membrane-associated progesterone receptor component 1 Heme binding 3.4 0.00455 
EMD Emerin Nuclear structure 3.3 0.00276 
PGRMC2 Membrane-associated progesterone receptor component 2 Heme binding 3.1 0.00166 
APMAP Adipocyte plasma membrane-associated protein Arylesterase 3.1 0.00307 
TMPO Lamina-associated polypeptide 2. isoforms β/γ; Thymopoietin Nuclear lamina 3.1 0.00485 
RPN2 Dolichyl-diphosphooligosaccharide-protein glycosyltransferase subunit N-glycosylation 3.0 0.00283 
TOR1AIP1 Torsin-1A-interacting protein 1 Torsin A activator 3.0 0.00522 
Gene name Protein name Function/metabolic pathway Fold change P-value 
UBB/UBC Polyubiquitin B/C Protein degradation/ERAD 10.3 0.00139 
LRRC59 Leucine-rich repeat-containing protein 59 Nuclear import 9.0 0.00164 
ERLIN2 Erlin-2 Protein degradation/ERAD 8.8 0.00063 
GOLGA2 Golgin subfamily A member 2 Vesicle docking/Golgi structure 8.4 0.00091 
EMC2 ER membrane protein complex subunit 2 Protein folding 82 0.00226 
SRPRB Signal recognition particle receptor subunit β ER protein translocation 7.6 0.00112 
ESYT1 Extended synaptotagmin-1 Plasma membrane-ER contact sites 7.4 0.00173 
CKAP4 Cytoskeleton-associated protein 4 Anchoring ER to microtubles 7.1 0.00015 
SEC22B Vesicle-trafficking protein SEC22B Vesicular transport 7.0 0.00350 
SPTLC1 Serine palmitoyltransferase 1 Sphingolipid synthesis 6.7 0.00082 
SPTLC2 Serine palmitoyltransferase 2 Sphingolipid synthesis 6.6 0.00192 
ERLIN1 Erlin-1 Protein degradation/ERAD 6.1 0.00147 
POR NADPH cytochrome P450 reductase Electron transfer to heme 5.9 0.00232 
RPN1 Dolichyl-diphosphooligosaccharide protein glycosyltransferase subunit N-glycosylation 5.6 0.00018 
CCDC47 Coiled-coil domain-containing protein 47 Protein degradation/ERAD 5.5 0.00134 
MIA3 Melanoma inhibitory activity protein 3 Vesicular trafficking 5.4 0.00106 
ATP2A2 Sarcoplasmic/endoplasmic reticulum calcium ATPase 2 Calcium transport into ER 5.3 0.00059 
VAPA Vesicle-associated membrane protein-associated protein A ER–Golgi contact sites 5.2 0.00108 
VAPB Vesicle-associated membrane protein-associated protein B ER–Golgi contact sites 5.1 0.00268 
SLC27A4 Long-chain fatty acid transport protein 4 Lipid metabolism 5.1 0.00142 
TMED9 Transmembrane emp24 domain-containing protein 9 Vesicular trafficking 4.7 0.00417 
RDH11 Retinol dehydrogenase 11 Lipid metabolism 4.6 0.00345 
TFRC Transferrin receptor protein 1 Iron metabolism 4.6 0.00351 
EPHX1 Epoxide hydrolase 1 Biotransformation 4.5 0.00292 
SSR4 Translocon-associated protein subunit γ ER protein translocation 4.3 0.00293 
RAB6A Ras-related protein Rab-6A Vesicle transport 4.2 0.00139 
NSDHL Sterol-4-α-carboxylate 3-dehydrogenase. decarboxylating Sterol metabolism 4.1 0.00021 
SCFD1 Sec1 family domain-containing protein 1 Trafficking ECM components 4.0 0.00421 
ZW10 Centromere/kinetochore protein zw10 homolog Vesicular trafficking 4.0 0.00437 
ALDH3A2 Fatty aldehyde dehydrogenase Lipid metabolism 3.9 0.00129 
DDOST Dolichyl-diphosphooligosaccharide protein glycosyltransferase subunit N-glycosylation 3.8 0.00036 
DHCR7 7-Dehydrocholesterol reductase Sterol metabolism 3.8 0.00306 
CANX Calnexin Protein folding 3.8 0.00462 
HACD3 Very long-chain (3R)-3-hydroxyacyl-CoA dehydratase 3 Fatty acid elongation 3.5 0.00532 
RAB7A Ras-related protein Rab-7A Vesicle transport 3.4 0.00377 
FAF2 FAS-associated factor 2 Protein degradation/ERAD 3.4 0.00395 
PGRMC1 Membrane-associated progesterone receptor component 1 Heme binding 3.4 0.00455 
EMD Emerin Nuclear structure 3.3 0.00276 
PGRMC2 Membrane-associated progesterone receptor component 2 Heme binding 3.1 0.00166 
APMAP Adipocyte plasma membrane-associated protein Arylesterase 3.1 0.00307 
TMPO Lamina-associated polypeptide 2. isoforms β/γ; Thymopoietin Nuclear lamina 3.1 0.00485 
RPN2 Dolichyl-diphosphooligosaccharide-protein glycosyltransferase subunit N-glycosylation 3.0 0.00283 
TOR1AIP1 Torsin-1A-interacting protein 1 Torsin A activator 3.0 0.00522 

Protein with more than three-fold increase in abundance compared with controls (FDR = 0.05).

Some proteins have additional functions that are not listed here.

Proteins also identified in the BioID screen (see Table 3) are shown in bold letters.

Table 3
Putative FA2H interaction partners identified by BioID
Gene name Protein name Function/metabolic pathway Fold change P-value 
CCDC47 Coiled-coil domain-containing protein 47 ERAD 4.7 0.00224 
PGRMC2 Membrane-associated progesterone receptor component 2 Heme binding 4.4 0.00092 
PTPN1 Tyrosine-protein phosphatase non-receptor type 1 Regulation of ER unfolded protein response 4.1 0.00065 
TOR1AIP1 Torsin-1A-interacting protein 1 Torsin A activator 3.6 0.00028 
SRPRB Signal recognition particle receptor subunit β ER protein translocation 3.6 0.00399 
VAPA Vesicle-associated membrane protein-associated protein A ER–Golgi contact sites 3.4 0.00103 
VAPB Vesicle-associated membrane protein-associated protein B ER–Golgi contact sites 3.3 0.00363 
SMPD4 Sphingomyelin phosphodiesterase 4 Sphingolipid metabolism 3.2 0.00683 
ALDH3A2 Fatty aldehyde dehydrogenase Lipid metabolism 3.1 0.00903 
Gene name Protein name Function/metabolic pathway Fold change P-value 
CCDC47 Coiled-coil domain-containing protein 47 ERAD 4.7 0.00224 
PGRMC2 Membrane-associated progesterone receptor component 2 Heme binding 4.4 0.00092 
PTPN1 Tyrosine-protein phosphatase non-receptor type 1 Regulation of ER unfolded protein response 4.1 0.00065 
TOR1AIP1 Torsin-1A-interacting protein 1 Torsin A activator 3.6 0.00028 
SRPRB Signal recognition particle receptor subunit β ER protein translocation 3.6 0.00399 
VAPA Vesicle-associated membrane protein-associated protein A ER–Golgi contact sites 3.4 0.00103 
VAPB Vesicle-associated membrane protein-associated protein B ER–Golgi contact sites 3.3 0.00363 
SMPD4 Sphingomyelin phosphodiesterase 4 Sphingolipid metabolism 3.2 0.00683 
ALDH3A2 Fatty aldehyde dehydrogenase Lipid metabolism 3.1 0.00903 

Protein with more than three-fold increase in abundance compared with controls (FDR = 0.05).

Proteins also identified in the PFA screen (see Table 2) are shown in bold letters.

Proteins that are involved in translation and protein folding in the rough ER are to some extent obvious FA2H interaction partners without a probably specific role for the regulation of FA2H enzyme activity. In addition, the transient overexpression may result in cross-linked interactions due to nonspecific aggregation of unfolded bait protein. On the other hand, it is unlikely that a mis- or unfolded bait protein interacts with a native oligomeric complex. We therefore selected candidate proteins for further evaluation by the following criteria: the putative interaction partners should appear in both experimental settings (BioID and PFA cross-linking) and/or more than one component of known protein complexes should be present. Following this approach, we identified the following protein groups as interesting candidates for further analyses: (A) PGRMC1 has been shown to interact with PGRMC2, NADPH cytochrome P450 oxidoreductase (POR; EC 1.6.2.4) and transferrin receptor-1 (TFRC) [37,38]. All four proteins were enriched in the FA2H bait samples of the PFA-cross-linking screen and PGRMC2 was also identified in the BioID approach (see Tables 2 and 3). PGRMC1 is known as an activator and a possible heme chaperone for cytochrome enzymes. In addition, POR is already known as a functional interaction partner of FA2H, because it functions as an electron donor in an FA2H enzyme assay [2,3]. (B) SPTLC1 and SPTLC2, both of which were found in the PFA screen, are two components of the serine palmitoyltransferase [34] and are thus acting in the same metabolic pathway as FA2H. (C) VAMP-associated protein A and B (VAPA and VAPB; both are found in both screening approaches; see Tables 2 and 3) are known to form dimers and are found at ER–Golgi and ER–endosome contact sides where they interact with several lipid transfer proteins, including the ceramide transfer protein CERT [39]. The interaction of FA2H with VAPA and VAPB could therefore potentially affect ceramide transfer from the ER to the site of sphingomyelin and glucosylceramide synthesis in the Golgi apparatus.

In the present study, we further examined the potential interaction of FA2H with PGRMC1 and SPTLC1, because both interactions could potentially be involved in the regulation of (hydroxy-) sphingolipid synthesis. Analogous to the mass spectrometry screening approaches, Twin-Strep-FA2H expressed in HEK-293 cells was precipitated after PFA cross-linking, and samples were boiled to destroy the cross-links. Western blots of the precipitates were probed with antibodies directed against PGRMC1 and SPTLC1 (Figure 4). These experiments verified the FA2H interactors tested, all of which showed no signal in the control, but a clear signal in the bait precipitate. However, in all three cases, the signal in the bait precipitate was quite low compared with the total lysate (Input), indicating that just a small fraction of each protein was engaged in the interaction with FA2H during PFA cross-linking, suggesting only weak and transient interactions.

Co-immunoprecipitation of PGRMC1 and SPTLC1 with FA2H.

Figure 4.
Co-immunoprecipitation of PGRMC1 and SPTLC1 with FA2H.

Interactions of FA2H with PGRMC1 and SPTLC1 were confirmed by western blotting after affinity pulldown of Twin-Strep-FA2H from control (empty vector) and Twin-Strep-FA2H-transfected cells. To stabilize protein interactions, cells were PFA-cross-linked before lysis. Note that only 1% of the total cell lysate was loaded, whereas eluates were applied completely.

Figure 4.
Co-immunoprecipitation of PGRMC1 and SPTLC1 with FA2H.

Interactions of FA2H with PGRMC1 and SPTLC1 were confirmed by western blotting after affinity pulldown of Twin-Strep-FA2H from control (empty vector) and Twin-Strep-FA2H-transfected cells. To stabilize protein interactions, cells were PFA-cross-linked before lysis. Note that only 1% of the total cell lysate was loaded, whereas eluates were applied completely.

Verification of FA2H interaction with PGRMC1 by BiFC

We further tested the possible interactions of FA2H with PGRMC1 and SPTLC1 in living cells using BiFC assays. FA2H and the putative interaction-partner proteins, all tagged at the C-terminus (which is located in the cytosol in all examined proteins) with complementary fragments (cYFP1 and YFP2) of citrine YFP (cYFP), were expressed in BHK (Figure 5A,C) and CHO-K1 cells (data not shown). ERGIC53, which is known to dimerize and give a strong BiFC signal [35], served as a positive control. Cells co-transfected with FA2H and PGRMC1 or SPTLC1 split-cYFP constructs showed fluorescence complementation (Figure 5A,B). In contrast with ERGIC53, BiFC signals of FA2H/PGRMC1 and FA2H/SPTLC1 split-cYFP fusion partners did not show an ER-like staining pattern, probably because the C-terminal fusions inactivated the C-terminal ER retrieval signals of FA2H and PGRMC1. As a control, a cYFP1 fusion protein of an identified but, according to our criteria, not significantly enriched protein, ACSL3 (EC 6.2.1.3), was co-expressed with FA2H-YFP2. ACSL3-cYFP1 did not functionally complement the FA2H–YFP2 fusion protein (Figure 5A).

Bimolecular fluorescence complementation (BiFC).

Figure 5.
Bimolecular fluorescence complementation (BiFC).

(A) BiFC was observed for pairs FA2H-YFP2/PGRMC1-cYFP1 and FA2H-cYFP1/SPTLC1-YFP2, but not for FA2H-YFP2/ACSL3-cYFP1. Co-expression of sscYFP1-ERGIC53 and ssYFP2-ERGIC53 served as a positive control. Similar results were obtained with BHK and CHO-K1 cells. Data shown here are from BHK cells. (B) Relative fluorescence intensities (normalized to BiFC intensity obtained with sscYFP1-ERGIC53 and ssYFP2-ERGIC53) were determined 48 h after transfection. Shown are the mean ± SD of three independent experiments. (C) Schematic drawing of the membrane topology of the examined split-cYFP fusion constructs.

Figure 5.
Bimolecular fluorescence complementation (BiFC).

(A) BiFC was observed for pairs FA2H-YFP2/PGRMC1-cYFP1 and FA2H-cYFP1/SPTLC1-YFP2, but not for FA2H-YFP2/ACSL3-cYFP1. Co-expression of sscYFP1-ERGIC53 and ssYFP2-ERGIC53 served as a positive control. Similar results were obtained with BHK and CHO-K1 cells. Data shown here are from BHK cells. (B) Relative fluorescence intensities (normalized to BiFC intensity obtained with sscYFP1-ERGIC53 and ssYFP2-ERGIC53) were determined 48 h after transfection. Shown are the mean ± SD of three independent experiments. (C) Schematic drawing of the membrane topology of the examined split-cYFP fusion constructs.

To further support the specificity of FA2H interactions with PGRMC1 and SPTLC1, BiFC-competition assays were performed, where an excess (eight-fold) of untagged FA2H or PGRMC1 was co-expressed with their split-cYFP-tagged counterparts (competition experiments using overexpression of SPTLC1 caused massive cell death, for unknown reasons, and therefore were excluded from the analysis). The competition should lead to a reduction in fluorescence. A complete inhibition of BiFC is, however, unlikely if the interaction is transient, because after split-cYFP dimerization, the formed cYFP is relatively stable [40]. Control cells were co-transfected with plasmids encoding an irrelevant cytosolic protein (Rimklb), because co-transfection with the empty vector led to significantly higher expression of the split-cYFP fusion proteins (data not shown). With this approach, the interactions of FA2H with PGRMC1 was confirmed by the significantly reduced BiFC in the presence of untagged PGRMC1 or FA2H (Figure 6A). Because the PGRMC1 antibody apparently did not recognize the endogenous protein in CHO-K1, it was not possible to determine the degree of overexpression in comparison with endogenous PGRMC1.

BiFC competition confirms the interaction of FA2H with PGRMC1.

Figure 6.
BiFC competition confirms the interaction of FA2H with PGRMC1.

(A) BiFC experiments were performed in CHO-K1 cells by co-transfection of the indicated cYFP1 and YFP2 fusion constructs (see also Figure 4) in the presence of an eight-fold excess of the plasmid encoding the untagged interaction partner (+FA2H; +PGRMC1). Controls were co-transfected with an irrelevant plasmid encoding a non-interacting protein (FLAG-tagged Nat8L or FLAG-tagged Rimklb). Fluorescence intensities were measured 48 h after transfection and data from three or six independent experiments were tested for significant differences using a paired t-test (*P < 0.05; n.s., not significant). Shown are the relative fluorescence intensities (mean ± SD; n = 3–6) obtained after subtracting the background fluorescence of the medium. (B) Western blot analysis (20 µg protein of total cell lysates per lane) of split-cYFP fusion proteins and competitor proteins (FA2H, PGRMC1). α-tubulin served as a loading control. Co-expressing of the competitor proteins resulted in similar or slightly increased concentrations of the split-cYFP fusion proteins compared with controls (co-expressing an irrelevant protein), indicating that reduced fluorescence signals were not the result of reduced protein expression. Because blots were first developed with FA2H antiserum, followed by PGRMC1 antibody, the FA2H signals remained visible in the PGRMC1 staining. Note that endogenous PGRMC1 was not detectable (the human PGRMC1 antibody is directed against a sequence of the C-terminus of PGRMC1 that differs between human and hamster at several positions and may thus not recognize the CHO-K1 enzyme). (C) Western blot analysis (20 µg protein of total cell lysates per lane) of split-cYFP SPTLC1 and FA2H fusion proteins. Co-expressing of the untagged FA2H competitor resulted in similar to slightly decreased concentrations of the split-cYFP fusion proteins compared with controls (co-expressing an irrelevant protein).

Figure 6.
BiFC competition confirms the interaction of FA2H with PGRMC1.

(A) BiFC experiments were performed in CHO-K1 cells by co-transfection of the indicated cYFP1 and YFP2 fusion constructs (see also Figure 4) in the presence of an eight-fold excess of the plasmid encoding the untagged interaction partner (+FA2H; +PGRMC1). Controls were co-transfected with an irrelevant plasmid encoding a non-interacting protein (FLAG-tagged Nat8L or FLAG-tagged Rimklb). Fluorescence intensities were measured 48 h after transfection and data from three or six independent experiments were tested for significant differences using a paired t-test (*P < 0.05; n.s., not significant). Shown are the relative fluorescence intensities (mean ± SD; n = 3–6) obtained after subtracting the background fluorescence of the medium. (B) Western blot analysis (20 µg protein of total cell lysates per lane) of split-cYFP fusion proteins and competitor proteins (FA2H, PGRMC1). α-tubulin served as a loading control. Co-expressing of the competitor proteins resulted in similar or slightly increased concentrations of the split-cYFP fusion proteins compared with controls (co-expressing an irrelevant protein), indicating that reduced fluorescence signals were not the result of reduced protein expression. Because blots were first developed with FA2H antiserum, followed by PGRMC1 antibody, the FA2H signals remained visible in the PGRMC1 staining. Note that endogenous PGRMC1 was not detectable (the human PGRMC1 antibody is directed against a sequence of the C-terminus of PGRMC1 that differs between human and hamster at several positions and may thus not recognize the CHO-K1 enzyme). (C) Western blot analysis (20 µg protein of total cell lysates per lane) of split-cYFP SPTLC1 and FA2H fusion proteins. Co-expressing of the untagged FA2H competitor resulted in similar to slightly decreased concentrations of the split-cYFP fusion proteins compared with controls (co-expressing an irrelevant protein).

In contrast with PGRMC1, however, BiFC of FA2H-cYFP1 and SPTLC1-YFP2 was not reduced in the presence of untagged FA2H (Figure 6A). Though this observation argues against a specific interaction of FA2H and SPTLC1, it is possible that the apparent weak and transient interaction together with the mislocalization of the proteins accounts for this result (as noted above, the addition of the cYFP fragment to the C-terminus destroys the ER retrieval signal of FA2H). Western blot analyses confirmed comparable expression levels of the split-YFP fusion proteins (or showed even increased levels in case of FA2H-YFP2 and PGRMC1-cYFP1) after co-transfection with the competitor encoding plasmid compared with control plasmids, indicating that reduced BiFC signals in the presence of untagged PGRMC1 is not caused by reduced expression of the split-cYFP fusion proteins (Figure 6B,C).

Synthesis of 2-hydroxylated sphingolipids is reduced in the presence of the PRGMC1 inhibitor AG-205

Having established the physical interaction of FA2H with PGRMC1, we further examined a possible functional role of this interaction using the PGRMC1 inhibitor AG-205 [41]. For these experiments, we used CHO-K1 cells, because these cells lack galactosylceramide and display a less complex sphingolipid pattern, facilitating the discrimination of hydroxylated and non-hydroxylated sphingolipid species by TLC. CHO-K1 cells were transiently transfected with FA2H (transfection mixtures contained 20% FA2H-expressing plasmid and 80% empty vector) and treated with 10 µM AG-205 for 25 h (in the presence of 2.5% FCS; under these conditions, cell viability was not significantly affected). Cells were metabolically labeled for 24 h with [14C]-serine during incubation with the inhibitor and lipids synthesized were analyzed by TLC (Figure 7A). In addition, cells treated under identical conditions, but not metabolically labeled, were analyzed for FA2H expression by western blotting (Figure 7D). AG-205 caused a significant reduction in the relative amount of HFA (hydroxylated fatty acid)-GlcCer (Figure 7B) and HFA-ceramide (Figure 7C). The addition of hemin reversed the effect of AG-205 and also led to a significant increase in HFA-sphingolipids in the absence of the inhibitor when compared with untreated FA2H-expressing cells (Figure 7C,D). This increase predominantly results from an enhanced synthesis of hydroxylated long-chain fatty acid (LC-HFA) sphingolipids (Figure 7A). Western blot analysis showed that expression of FA2H was strongly reduced when co-expressed with PGRMC1 (Figure 7D). Interestingly, FA2H protein levels in these cells were also strongly reduced compared with AG-205-treated FA2H-expressing cells (and this was also the case at earlier time points; data not shown), though HFA-sphingolipid levels in the latter were significantly lower (Figure 7B,C). This is in agreement with a lower specific activity of FA2H in the presence of AG-205. We concluded that PGRMC1 can stimulate FA2H activity, possibly through its heme chaperone activity [37,4244].

PGRMC1 inhibitor AG-205 reduces synthesis of 2-hydroxylated sphingolipids.

Figure 7.
PGRMC1 inhibitor AG-205 reduces synthesis of 2-hydroxylated sphingolipids.

(A–D) CHO-K1 cells were co-transfected with an FA2H-encoding plasmid and empty vector or FA2H and PGRMC1 expression plasmids. Six hours after transfection, 10 µM AG-205 and/or 20 µM hemin (or an equal volume of the solvent DMSO) were added and one additional hour later cells were metabolically labeled with [14C]-serine for 24 h. (A) Lipids were extracted and separated by TLC followed by visualizing radioactivity using Bioimager screens. Synthesized non-hydroxylated fatty acid containing (NFA) and hydroxylated fatty acid containing (HFA) ceramides and glucosylceramides were quantified by densitometry and the relative amount of HFA-glucosylceramide (B) and HFA-ceramide (C) was calculated. NFA- and HFA-sphingolipids migrate as double bands, which are known to represent very long-chain fatty acid (VLC) and long-chain fatty acid (LC) containing lipids, respectively ([24]). Sphingolipids in CHO-K1 cells mainly contain fatty acids with chain lengths of C24/C22 (VLC) and C16 (LC) [57]. (D) CHO-K1 cells transfected and treated under identical conditions, but not metabolically labeled, were analyzed by western blotting using antibodies against FA2H and PGRMC1. α-Tubulin was used as a loading control (because blots were not stripped before reprobing, a weak signal below the tubulin band is visible in PGRMC1-overexpressing cells, which corresponds to the PGRMC1 dimer band). (E–H) CHO-K1 cells were co-transfected with an FA2H-encoding plasmid and a plasmid encoding an irrelevant protein (Rimklb) or FA2H and PGRMC1 expression plasmids. Six hours after transfection, cells were treated with 20 µM hemin (or an equal volume of the solvent DMSO) and 1 h later cells were metabolically labeled with [14C]-serine for 24 h. Lipids were extracted, separated by TLC and quantified as in (A). The relative amount of HFA-glucosylceramide (F) and HFA-ceramide (G) was calculated. (H) CHO-K1 cells transfected and treated under identical conditions, but not metabolically labeled, were analyzed by western blotting using antibodies against FA2H and PGRMC1. Gels with identical loading were run in parallel and resulting blots were stained separately with antibodies against FA2H and PGRMC1; therefore, two different α-tubulin loading controls are shown. Because blots were not stripped before reprobing with the tubulin antibody, a weak signal below the tubulin band is visible in PGRMC1-transfected cells, which corresponds to the PGRMC1 dimer band (cropped in the PGRMC1 blots). Data shown are the mean + SD of 3–5 independent experiments [n = 5 for the first four groups in (B) and (C); n = 3 for all others]. Diamonds indicate significant differences (P < 0.02) to all groups not labeled with a diamond, and asterisks indicate significant differences (P < 0.02) between two groups, as indicated (one way-ANOVA with post hoc Tukey HSD test).

Figure 7.
PGRMC1 inhibitor AG-205 reduces synthesis of 2-hydroxylated sphingolipids.

(A–D) CHO-K1 cells were co-transfected with an FA2H-encoding plasmid and empty vector or FA2H and PGRMC1 expression plasmids. Six hours after transfection, 10 µM AG-205 and/or 20 µM hemin (or an equal volume of the solvent DMSO) were added and one additional hour later cells were metabolically labeled with [14C]-serine for 24 h. (A) Lipids were extracted and separated by TLC followed by visualizing radioactivity using Bioimager screens. Synthesized non-hydroxylated fatty acid containing (NFA) and hydroxylated fatty acid containing (HFA) ceramides and glucosylceramides were quantified by densitometry and the relative amount of HFA-glucosylceramide (B) and HFA-ceramide (C) was calculated. NFA- and HFA-sphingolipids migrate as double bands, which are known to represent very long-chain fatty acid (VLC) and long-chain fatty acid (LC) containing lipids, respectively ([24]). Sphingolipids in CHO-K1 cells mainly contain fatty acids with chain lengths of C24/C22 (VLC) and C16 (LC) [57]. (D) CHO-K1 cells transfected and treated under identical conditions, but not metabolically labeled, were analyzed by western blotting using antibodies against FA2H and PGRMC1. α-Tubulin was used as a loading control (because blots were not stripped before reprobing, a weak signal below the tubulin band is visible in PGRMC1-overexpressing cells, which corresponds to the PGRMC1 dimer band). (E–H) CHO-K1 cells were co-transfected with an FA2H-encoding plasmid and a plasmid encoding an irrelevant protein (Rimklb) or FA2H and PGRMC1 expression plasmids. Six hours after transfection, cells were treated with 20 µM hemin (or an equal volume of the solvent DMSO) and 1 h later cells were metabolically labeled with [14C]-serine for 24 h. Lipids were extracted, separated by TLC and quantified as in (A). The relative amount of HFA-glucosylceramide (F) and HFA-ceramide (G) was calculated. (H) CHO-K1 cells transfected and treated under identical conditions, but not metabolically labeled, were analyzed by western blotting using antibodies against FA2H and PGRMC1. Gels with identical loading were run in parallel and resulting blots were stained separately with antibodies against FA2H and PGRMC1; therefore, two different α-tubulin loading controls are shown. Because blots were not stripped before reprobing with the tubulin antibody, a weak signal below the tubulin band is visible in PGRMC1-transfected cells, which corresponds to the PGRMC1 dimer band (cropped in the PGRMC1 blots). Data shown are the mean + SD of 3–5 independent experiments [n = 5 for the first four groups in (B) and (C); n = 3 for all others]. Diamonds indicate significant differences (P < 0.02) to all groups not labeled with a diamond, and asterisks indicate significant differences (P < 0.02) between two groups, as indicated (one way-ANOVA with post hoc Tukey HSD test).

Because co-transfection of a second transcribed plasmid can potentially reduce expression of the first gene [45], it is possible that the low FA2H expression level in the PGRMC1-co-transfected cells was not caused by a specific effect of PGRMC1. The experiments were therefore repeated using co-transfection of a plasmid encoding an irrelevant (cytosolic) protein (Rimklb; as in the BiFC experiments, shown in Figure 6) instead of the empty vector (Figure 7E). Although differences in the FA2H expression level were lower, the amount of FA2H was still clearly reduced in PGRMC1-co-expressing cells, compared with cells co-expressing the control protein (Figure 7H). This suggests a possible destabilizing effect of PGRMC1 on FA2H, though further experiments are required to clarify this. The addition of hemin did no longer increase the relative level of HFA-sphingolipids significantly, but rather led to a slight decrease in the relative amount of HFA-sphingolipids (Figure 7F,G), indicating that heme supply is not limiting FA2H activity at a lower expression level. In total, the levels of HFA-sphingolipids were similar in cells overexpressing PGRMC1 or the control protein.

Discussion

In the present study, we used two different methods, proximity biotinylation and formaldehyde cross-linking, both in combination with SILAC, to identify putative interaction partners of FA2H. Because of the restriction for successful proximity biotinylation, which depends on the accessibility of lysine residues in a favored orientation toward the biotin ligase, the relatively low number of proteins significantly enriched in the BioID screen compared with the PFA screen is not unexpected. Nevertheless, the high overlap of positive hits from the BioID and the PFA-cross-linking screens illustrates the feasibility of the latter approach. Obviously, all putative candidates have to be confirmed by independent methods, as done here for PGRMC1. BiFC as well as the use of the PGRMC1 inhibitor AG-205 provided strong evidence for the physical and functional interaction of this protein with FA2H.

PGRMC1 is the best-characterized member of a family of four related membrane-associated progesterone receptors that include PGRMC2, neudesin and neuferricin (for review, see ref. [46]), all containing a cytochrome b5 domain [47]. PGRMC1 binds heme and progesterone [42,44,48,49] and forms homodimers, whereby dimerization depends on heme binding [50]. Structural analysis of PGRMC1 showed that heme-dependent dimerization of PGRMC1 depends on the direct interaction of the bound heme groups [50]. PGRMC1 positively regulates several cytochrome P450 enzymes that are involved in cholesterol synthesis [43,51]. In addition, PGRMC1 was found to bind a sterol regulatory element-binding protein cleavage activating protein and insulin-induced gene (Insig) [52]. The functional interaction of PGRMC1 with enzymes involved in cholesterol metabolism and its regulation, on the one hand, and FA2H, on the other, could potentially co-ordinate both metabolic pathways. Several examples for functional interdependence of sphingolipids and sterols are known (for review, see ref. [13]). Moreover, a mutual interference between sphingolipid hydroxylation and sterol metabolism is suggested by experiments in Drosophila, showing that cholesterol depletion leads to up-regulation of FA2H expression and elevated levels of 2-hydroxylated sphingolipids [53].

PGRMC1 forms heterodimers with PGRMC2 [54] and forms a complex with the TFRC and ferrochelatase, suggesting a role in regulating heme metabolism [37]. Piel et al. [37] showed that purified PGRMC1 can transfer heme to cytochrome b5. Thus, PGRMC1 seems to act as a heme chaperone for cytochrome b5 and other cytochrome enzymes [37,51]. We propose that FA2H activation depends at least, in part, on PGRMC1, because the PGRMC1 inhibitor AG-205 significantly reduced synthesis of hydroxylated sphingolipids, which could be prevented by the addition of hemin. This would be in line with a heme chaperone activity of PGRMC1 toward FA2H. Interestingly, hemin increased the synthesis of HFA-sphingolipids in cells highly overexpressing FA2H also in the absence of AG-205, which suggests that full activation of the enzyme under these conditions is limited by heme supply. The increase was mainly caused by a strong increase in hydroxylated LC-sphingolipids, suggesting that FA2H has a preference for very long-chain (VLC)-ceramides. An alternative explanation would be a preferential interaction of FA2H with ceramide synthases synthesizing VLC-ceramides (ceramide synthases CerS2, CerS3 and CerS4; for review, see [55,56]), though we did not find any ceramide synthase as a candidate interaction partner in our screens. A less efficient hydroxylation of LC (C16)-ceramide compared with VLC (C24/C22)-ceramide is also in line with previous mass spectrometric analysis of FA2H-expressing CHO-K1 cells [57].

Further evaluation of the putative FA2H interactors identified in our screenings may also be important for the understanding of FAHN/SPG35 and hereditary spastic paraplegia in general. In some cases, in vitro enzyme activity assays of FA2H harboring missense mutations found in FAHN patients showed only slight or even no reduction in enzymatic activity [6,7]. Although this could be the result of the overexpression of the enzyme in the in vitro assay performed, or decreased protein stability (as suggested by Kruer et al. [7]), it may be possible that the interaction of FA2H with its binding partners is affected by disease-causing mutations, potentially playing a role in pathogenesis. Along this line it is interesting that ERLIN1, which (together with its known interaction partner ERLIN2) appeared in the PFA-cross-linking screen as a putative FA2H interactor and was also identified as an FA2H interaction partner in a yeast two-hybrid screen [58], is known to be mutated in the hereditary spastic paraplegia SPG62 [59]. Whether impaired interactions of different SPG proteins could be relevant for the diseases will require further studies.

Abbreviations

     
  • ABCA1

    ATP-binding cassette sub-family A member 1

  •  
  • ACN

    acetonitrile

  •  
  • AGC

    automatic gain control

  •  
  • BiFC

    bimolecular fluorescence complementation

  •  
  • BioID

    proximity-dependent biotin identification

  •  
  • BirA*

    biotin protein ligase BirA mutant (R118G)

  •  
  • CID

    collision-induced dissociation

  •  
  • DMEM

    Dulbecco's modified Eagle's medium

  •  
  • ER

    endoplasmic reticulum

  •  
  • FA

    formic acid

  •  
  • FA2H

    fatty acid 2-hydroxylase

  •  
  • FAHN

    fatty acid hydroxylase-associated neurodegeneration

  •  
  • FCS

    fetal calf serum

  •  
  • FDR

    false discovery rate

  •  
  • GlcCer

    glucosylceramide

  •  
  • HFA

    hydroxylated fatty acid

  •  
  • HRP

    horseradish peroxidase

  •  
  • LC

    long chain

  •  
  • LCFA

    long chain fatty acid

  •  
  • LC–MS/MS

    liquid chromatography coupled to tandem mass spectrometry

  •  
  • MS

    mass spectrometry

  •  
  • NFA

    non-hydroxylated fatty acid

  •  
  • PFA

    paraformaldehyde

  •  
  • PGRMC1/2

    progesterone receptor membrane component 1/2

  •  
  • POR

    NADPH cytochrome P450 oxidoreductase

  •  
  • SILAC

    stable isotope labeling with amino acids in cell culture

  •  
  • SPTLC1/2

    serine palmitoyltransferase long chain-1/2

  •  
  • TFRC

    transferrin receptor-1

  •  
  • TLC

    thin layer chromatography

  •  
  • VAMP

    vesicle-associated membrane protein

  •  
  • VAPA/VAPB

    VAMP-associated protein A/B

  •  
  • VLC

    very long chain

  •  
  • VLCFA

    very long chain fatty acid

Author Contribution

M.E. conceived the study, performed experiments, analyzed data and wrote the manuscript. R.H. performed experiments, analyzed data and wrote the manuscript. D.W. performed experiments and analyzed data. V.G. analyzed data. All the authors read and approved the final manuscript.

Funding

This work was supported by grants of the Deutsche Forschungsgemeinschaft through SFB645 of the University of Bonn, project B5 (to M.E.) and Z4 (to V.G.).

Acknowledgments

We are very grateful to Prof. T. Hornemann for the kind gift of SPTLC antibodies. Furthermore, we also thank Dr V. Reiterer for the generous gift of plasmids used for the generation of BiFC constructs.

Competing Interests

The Authors declare that there are no competing interests associated with the manuscript.

References

References
1
Hama
,
H.
(
2010
)
Fatty acid 2-Hydroxylation in mammalian sphingolipid biology
.
Biochim. Biophys. Acta, Mol. Cell Biol. Lipids
1801
,
405
414
2
Alderson
,
N.L.
,
Rembiesa
,
B.M.
,
Walla
,
M.D.
,
Bielawska
,
A.
,
Bielawski
,
J.
and
Hama
,
H.
(
2004
)
The human FA2H gene encodes a fatty acid 2-hydroxylase
.
J. Biol. Chem.
279
,
48562
48568
3
Alderson
,
N.L.
,
Walla
,
M.D.
and
Hama
,
H.
(
2005
)
A novel method for the measurement of in vitro fatty acid 2-hydroxylase activity by gas chromatography-mass spectrometry
.
J. Lipid Res.
46
,
1569
1575
4
Zhu
,
G.
,
Koszelak-Rosenblum
,
M.
,
Connelly
,
S.M.
,
Dumont
,
M.E.
and
Malkowski
,
M.G.
(
2015
)
The crystal structure of an integral membrane fatty acid α-hydroxylase
.
J. Biol. Chem.
290
,
29820
29833
5
Edvardson
,
S.
,
Hama
,
H.
,
Shaag
,
A.
,
Gomori
,
J.M.
,
Berger
,
I.
,
Soffer
,
D.
et al. 
(
2008
)
Mutations in the fatty acid 2-hydroxylase gene are associated with leukodystrophy with spastic paraparesis and dystonia
.
Am. J. Hum. Genet.
83
,
643
648
6
Dick
,
K.J.
,
Eckhardt
,
M.
,
Paisán-Ruiz
,
C.
,
Alshehhi
,
A.A.
,
Proukakis
,
C.
,
Sibtain
,
N.A.
et al. 
(
2010
)
Mutation of FA2H underlies a complicated form of hereditary spastic paraplegia (SPG35)
.
Hum. Mutat.
31
,
E1251
E1260
7
Kruer
,
M.C.
,
Paisán-Ruiz
,
C.
,
Boddaert
,
N.
,
Yoon
,
M.Y.
,
Hama
,
H.
,
Gregory
,
A.
et al. 
(
2010
)
Defective FA2H leads to a novel form of neurodegeneration with brain iron accumulation (NBIA)
.
Ann. Neurol.
68
,
611
618
8
Zöller
,
I.
,
Meixner
,
M.
,
Hartmann
,
D.
,
Büssow
,
H.
,
Meyer
,
R.
,
Gieselmann
,
V.
et al. 
(
2008
)
Absence of 2-hydroxylated sphingolipids is compatible with normal neural development but causes late-onset axon and myelin sheath degeneration
.
J. Neurosci.
28
,
9741
9754
9
Potter
,
K.A.
,
Kern
,
M.J.
,
Fullbright
,
G.
,
Bielawski
,
J.
,
Scherer
,
S.S.
,
Yum
,
S.W.
et al. 
(
2011
)
Central nervous system dysfunction in a mouse model of Fa2h deficiency
.
Glia
59
,
1009
1021
10
Giraudo
,
C.G.
,
Daniotti
,
J.L.
and
Maccioni
,
H.J.F.
(
2001
)
Physical and functional association of glycolipid N-acetyl-galactosaminyl and galactosyl transferases in the Golgi apparatus
.
Proc. Natl Acad. Sci. U.S.A.
98
,
1625
1630
11
Sprong
,
H.
,
Degroote
,
S.
,
Nilsson
,
T.
,
Kawakita
,
M.
,
Ishida
,
N.
,
van der Sluijs
,
P.
et al. 
(
2003
)
Association of the Golgi UDP-galactose transporter with UDP-galactose:ceramide galactosyltransferase allows UDP-galactose import in the endoplasmic reticulum
.
Mol. Biol. Cell
14
,
3482
3493
12
Ohno
,
Y.
,
Suto
,
S.
,
Yamanaka
,
M.
,
Mizutani
,
Y.
,
Mitsutake
,
S.
,
Igarashi
,
Y.
et al. 
(
2010
)
ELOVL1 production of C24 acyl-CoAs is linked to C24 sphingolipid synthesis
.
Proc. Natl Acad. Sci. U.S.A.
107
,
18439
18444
13
Breslow
,
D.K.
and
Weissman
,
J.S.
(
2010
)
Membranes in balance: mechanisms of sphingolipid homeostasis
.
Mol. Cell
40
,
267
279
14
Tamehiro
,
N.
,
Zhou
,
S.
,
Okuhira
,
K.
,
Benita
,
Y.
,
Brown
,
C.E.
,
Zhuang
,
D.Z.
et al. 
(
2008
)
SPTLC1 binds ABCA1 to negatively regulate trafficking and cholesterol efflux activity of the transporter
.
Biochemistry
47
,
6138
6147
15
Maier
,
H.
,
Meixner
,
M.
,
Hartmann
,
D.
,
Sandhoff
,
R.
,
Wang-Eckhardt
,
L.
,
Zöller
,
I.
et al. 
(
2011
)
Normal fur development and sebum production depends on fatty acid 2-hydroxylase expression in sebaceous glands
.
J. Biol. Chem.
286
,
25922
25934
16
Nagano
,
M.
,
Ihara-Ohori
,
Y.
,
Imai
,
H.
,
Inada
,
N.
,
Fujimoto
,
M.
,
Tsutsumi
,
N.
et al. 
(
2009
)
Functional association of cell death suppressor, Arabidopsis Bax inhibitor-1, with fatty acid 2-hydroxylation through cytochrome b5
.
Plant J.
58
,
122
134
17
Vasilescu
,
J.
,
Guo
,
X.
and
Kast
,
J.
(
2004
)
Identification of protein-protein interactions using in vivo cross-linking and mass spectrometry
.
Proteomics
4
,
3845
3854
18
Roux
,
K.J.
,
Kim
,
D.I.
,
Raida
,
M.
and
Burke
,
B.
(
2012
)
A promiscuous biotin ligase fusion protein identifies proximal and interacting proteins in mammalian cells
.
J. Cell Biol.
196
,
801
810
19
Ong
,
S.-E.
,
Blagoev
,
B.
,
Kratchmarova
,
I.
,
Kristensen
,
D.B.
,
Steen
,
H.
,
Pandey
,
A.
et al. 
(
2002
)
Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurate approach to expression proteomics
.
Mol. Cell. Proteomics
1
,
376
386
20
Nyfeler
,
B.
,
Reiterer
,
V.
,
Wendeler
,
M.W.
,
Stefan
,
E.
,
Zhang
,
B.
,
Michnick
,
S.W.
et al. 
(
2008
)
Identification of ERGIC-53 as an intracellular transport receptor of α1-antitrypsin
.
J. Cell Biol.
180
,
705
712
21
Becker
,
I.
,
Lodder
,
J.
,
Gieselmann
,
V.
and
Eckhardt
,
M.
(
2010
)
Molecular characterization of N-acetylaspartylglutamate synthetase
.
J. Biol. Chem.
285
,
29156
29164
22
Graham
,
F.L.
and
van der Eb
,
A.J.
(
1973
)
A new technique for the assay of infectivity of human adenovirus 5 DNA
.
Virology
52
,
456
467
23
Bligh
,
E.G.
and
Dyer
,
W.J.
(
1959
)
A rapid method of total lipid extraction and purification
.
Can. J. Biochem. Physiol.
37
,
911
917
24
Morell
,
P.
and
Radin
,
N.S.
(
1969
)
Synthesis of cerebroside by brain from uridine diphosphate galactose and ceramide containing hydroxy fatty acid
.
Biochemistry
8
,
506
512
25
Klockenbusch
,
C.
and
Kast
,
J.
(
2010
)
Optimization of formaldehyde cross-linking for protein interaction analysis of non-tagged integrin 1
.
J. Biomed. Biotechnol.
2010
,
927585
26
Wessel
,
D.
and
Flügge
,
U.I.
(
1984
)
A method for the quantitative recovery of protein in dilute solution in the presence of detergents and lipids
.
Anal. Biochem.
138
,
141
143
27
Kang
,
D.-H.
,
Gho
,
Y.-S.
,
Suh
,
M.-K.
, and
Kang
,
C.-H.
(
2002
)
Highly sensitive and fast protein detection with Coomassie Brilliant Blue in sodium dodecyl sulfate-polyacrylamide gel electrophoresis
.
Bull. Korean Chem. Soc.
23
,
1511
1512
28
Winter
,
D.
and
Steen
,
H.
(
2011
)
Optimization of cell lysis and protein digestion protocols for the analysis of HeLa S3 cells by LC-MS/MS
.
Proteomics
11
,
4726
4730
29
Rappsilber
,
J.
,
Mann
,
M.
and
Ishihama
,
Y.
(
2007
)
Protocol for micro-purification, enrichment, pre-fractionation and storage of peptides for proteomics using StageTips
.
Nat. Protoc.
2
,
1896
1906
30
Hahne
,
H.
,
Pachl
,
F.
,
Ruprecht
,
B.
,
Maier
,
S.K.
,
Klaeger
,
S.
,
Helm
,
D.
et al. 
(
2013
)
DMSO enhances electrospray response, boosting sensitivity of proteomic experiments
.
Nat. Methods
10
,
989
991
31
Cox
,
J.
and
Mann
,
M.
(
2008
)
Maxquant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteome-wide protein quantification
.
Nat. Biotechnol.
26
,
1367
1372
32
Cox
,
J.
,
Neuhauser
,
N.
,
Michalski
,
A.
,
Scheltema
,
R.A.
,
Olsen
,
J.V.
and
Mann
,
M.
(
2011
)
Andromeda: a peptide search engine integrated into the MaxQuant environment
.
J. Proteome Res.
10
,
1794
1805
33
Cox
,
J.
and
Mann
,
M.
(
2012
)
1D and 2D annotation enrichment: a statistical method integrating quantitative proteomics with complementary high-throughput data
.
BMC Bioinformatics
13
(
suppl 16
),
S12
34
Hornemann
,
T.
,
Richard
,
S.
,
Rütti
,
M.F.
,
Wei
,
Y.
and
von Eckardstein
,
A.
(
2006
)
Cloning and initial characterization of a new subunit for mammalian serine-palmitoyltransferase
.
J. Biol. Chem.
281
,
37275
37281
35
Nyfeler
,
B.
,
Michnick
,
S.W.
and
Hauri
,
H.-P.
(
2005
)
Capturing protein interactions in the secretory pathway of living cells
.
Proc. Natl Acad. Sci. U.S.A.
102
,
6350
6355
36
Varnaite˙
,
R.
and
MacNeill
,
S.A.
(
2016
)
Meet the neighbors: mapping local protein interactomes by proximity-dependent labeling with BioID
.
Proteomics
16
,
2503
2518
37
Piel
, III,
R.B.
,
Shiferaw
,
M.T.
,
Vashisht
,
A.A.
,
Marcero
,
J.R.
,
Praissman
,
J.L.
,
Phillips
,
J.D.
et al. 
(
2016
)
A novel role for progesterone receptor membrane component 1 (PGRMC1): a partner and regulator of ferrochelatase
.
Biochemistry
55
,
5204
5217
38
Szczesna-Skorupa
,
E.
and
Kemper
,
B.
(
2011
)
Progesterone receptor membrane component 1 inhibits the activity of drug-metabolizing cytochromes P450 and binds to cytochrome P450 reductase
.
Mol. Pharmacol.
79
,
340
350
39
Kawano
,
M.
,
Kumagai
,
K.
,
Nishijima
,
M.
and
Hanada
,
K.
(
2006
)
Efficient trafficking of ceramide from the endoplasmic reticulum to the Golgi apparatus requires a VAMP-associated protein-interacting FFAT motif of CERT
.
J. Biol. Chem.
281
,
30279
30288
40
Kerppola
,
T.K.
(
2006
)
Visualization of molecular interactions by fluorescence complementation
.
Nat. Rev. Mol. Cell Biol.
7
,
449
456
41
Ahmed
,
I.S.
,
Rohe
,
H.J.
,
Twist
,
K.E.
,
Mattingly
,
M.N.
and
Craven
,
R.J.
(
2010
)
Progesterone receptor membrane component 1 (Pgrmc1): a heme-1 domain protein that promotes tumorigenesis and is inhibited by a small molecule
.
J. Pharmacol. Exp. Ther.
333
,
564
573
42
Ghosh
,
K.
,
Thompson
,
A.M.
,
Goldbeck
,
R.A.
,
Shi
,
X.
,
Whitman
,
S.
,
Oh
,
E.
et al. 
(
2005
)
Spectroscopic and biochemical characterization of heme binding to yeast Dap1p and mouse PGRMC1p
.
Biochemistry
44
,
16729
16736
43
Mallory
,
J.C.
,
Crudden
,
G.
,
Johnson
,
B.L.
,
Mo
,
C.
,
Pierson
,
C.A.
,
Bard
,
M.
et al. 
(
2005
)
Dap1p, a heme-binding protein that regulates the cytochrome P450 protein Erg11p/Cyp51p in Saccharomyces cerevisiae
.
Mol. Cell. Biol.
25
,
1669
1679
44
Kaluka
,
D.
,
Batabyal
,
D.
,
Chiang
,
B.-Y.
,
Poulos
,
T.L.
and
Yeh
,
S.-R.
(
2015
)
Spectroscopic and mutagenesis studies of human PGRMC1
.
Biochemistry
54
,
1638
1647
45
Lauret
,
E.
and
Baserga
,
R.
(
1988
)
Inhibition of gene expression at the translational level by cotransfection with competitor plasmids
.
DNA
7
,
151
156
46
Ryu
,
C.S.
,
Klein
,
K.
and
Zanger
,
U.M.
(
2017
)
Membrane associated progesterone receptors: promiscuous proteins with pleiotropic functions — focus on interactions with cytochromes P450
.
Front. Pharmacol.
8
,
159
47
Mifsud
,
W.
and
Bateman
,
A.
(
2002
)
Membrane-bound progesterone receptors contain a cytochrome b5-like ligand-binding domain
.
Genome Biol.
3
,
research0068.1
48
Min
,
L.
,
Strushkevich
,
N.V.
,
Harnastai
,
I.N.
,
Iwamoto
,
H.
,
Gilep
,
A.A.
,
Takemori
,
H.
et al. 
(
2005
)
Molecular identification of adrenal inner zone antigen as a heme-binding protein
.
FEBS J.
272
,
5832
5843
49
Crudden
,
G.
,
Chitti
,
R.E.
and
Craven
,
R.J.
(
2006
)
Hpr6 (heme-1 domain protein) regulates the susceptibility of cancer cells to chemotherapeutic drugs
.
J. Pharmacol. Exp. Ther.
316
,
448
455
50
Kabe
,
Y.
,
Nakane
,
T.
,
Koike
,
I.
,
Yamamoto
,
T.
,
Sugiura
,
Y.
,
Harada
,
E.
et al. 
(
2016
)
Haem-dependent dimerization of PGRMC1/Sigma-2 receptor facilitates cancer proliferation and chemoresistance
.
Nat. Commun.
7
,
11030
51
Hughes
,
A.L.
,
Powell
,
D.W.
,
Bard
,
M.
,
Eckstein
,
J.
,
Barbuch
,
R.
,
Link
,
A.J.
et al. 
(
2007
)
Dap1/PGRMC1 binds and regulates cytochrome P450 enzymes
.
Cell Metab.
5
,
143
149
52
Suchanek
,
M.
,
Radzikowska
,
A.
and
Thiele
,
C.
(
2005
)
Photo-leucine and photo-methionine allow identification of protein-protein interactions in living cells
.
Nat. Methods
2
,
261
268
53
Carvalho
,
M.
,
Schwudke
,
D.
,
Sampaio
,
J.L.
,
Palm
,
W.
,
Riezman
,
I.
,
Dey
,
G.
et al. 
(
2010
)
Survival strategies of a sterol auxotroph
.
Development
137
,
3675
3685
54
Peluso
,
J.J.
,
Griffin
,
D.
,
Liu
,
X.
and
Horne
,
M.
(
2014
)
Progesterone receptor membrane component-1 (PGRMC1) and PGRMC-2 interact to suppress entry into the cell cycle in spontaneously immortalized rat granulosa cells
.
Biol. Reprod.
91
,
104
55
Mullen
,
T.D.
,
Hannun
,
Y.A.
and
Obeid
,
L.M.
(
2012
)
Ceramide synthases at the centre of sphingolipid metabolism and biology
.
Biochem. J.
441
,
789
802
56
Tidhar
,
R.
and
Futerman
,
A.H.
(
2013
)
The complexity of sphingolipid biosynthesis in the endoplasmic reticulum
.
Biochim. Biophys. Acta, Mol. Cell Res.
1833
,
2511
2518
57
Eckhardt
,
M.
,
Yaghootfam
,
A.
,
Fewou
,
S.N.
,
Zöller
,
I.
and
Gieselmann
,
V.
(
2005
)
A mammalian fatty acid hydroxylase responsible for the formation of α-hydroxylated galactosylceramide in myelin
.
Biochem. J.
388
,
245
254
58
Rolland
,
T.
,
Taşan
,
M.
,
Charloteaux
,
B.
,
Pevzner
,
S.J.
,
Zhong
,
Q.
,
Sahni
,
N.
et al. 
(
2014
)
A proteome-scale map of the human interactome network
.
Cell
159
,
1212
1226
59
Novarino
,
G.
,
Fenstermaker
,
A.G.
,
Zaki
,
M.S.
,
Hofree
,
M.
,
Silhavy
,
J.L.
,
Heiberg
,
A.D.
et al. 
(
2014
)
Exome sequencing links corticospinal motor neuron disease to common neurodegenerative disorders
.
Science
343
,
506
511