The TMEM189 gene encodes plasmanylethanolamine desaturase which introduces the characteristic vinyl ether double bond into plasmalogens

Significance Although sequencing of the human genome was completed years ago, we still do not know about the physiological significance of thousands of predicted proteins, particularly of predicted membrane proteins. On the other hand, for approximately 100 human enzymes, no coding gene is known even though their enzymatic reaction has been well characterized. In this work, we assign one of those predicted membrane proteins (transmembrane protein 189; TMEM189) to one of the enzymatic reactions with an uncharacterized gene (plasmanylethanolamine desaturase). This enzyme catalyzes the final step in the biosynthesis of plasmalogens, an abundant class of glycerophospholipids that is depleted in such diseases as Alzheimer’s. Our findings enable interpretation of the previously characterized impaired growth phenotype of Tmem189-deficient mice.

Transfection with expression plasmids and siRNAs HEK293T cells were seeded at a density of 5 x 10 5 in 6 well plates and transfected on the next day with 4 µg of expression plasmid and 6 µl TurboFect (Thermo-Fisher, Vienna Austria) in a total volume of 4 ml. Mammalian expression plasmids (CMV promoter) were from the Mammalian Genome Collection (for the murine proteins DEGS1, DEGS2, FADS1, FADS2, FADS6, SCD1, SCD2, TMEM189, Dharmacon, Lafayette, CO, USA) or from Origene (for the human proteins FADS3, FA2H, Origene, Rockville, MD USA). Human or murine sequences were chosen according to availability of error-free clones suited directly for mammalian expression as well as according to pricing. Untagged clones were used to avoid potential inactivation of the enzymatic activity by the tag, C-terminally 6xmyc-tagged clones were prepared from these to enable monitoring of recombinant protein formation by Western blot. Site directed mutagenesis was performed with the Quikchange Kit (Stratagene, La Jolla, CA, USA). Clones were regrown, purified with an endotoxinfree plasmid preparation protocol with the endo-free Kit (Qiagen) and sequenced (Microsynth). HAP1 cells were seeded at a density of 8 x 10 5 in 6 well plates and transfected with 1 µg DNA and 5 µl TurboFectin (Origene) in a total volume of 2 ml. A431 cells were seeded at a density of 3 x 10 5 and treated with 25 pmol siGenome Smart pools (Dharmacon) and 4 µl ScreenFect siRNA reagent (ScreenFect, Eggenstein-Leopoldshafen, Germany) in a total volume of 1.5 ml.

Real time live confocal microscopy
Murine Tmem189 cDNA was cloned into pEGF-N1 (Clontech, Mountain View, CA, USA), the individual histidine to alanine mutations inserted using the Quikchange Kit (Stratagene) and the plasmids transfected to HEK293T cells which were seeded in a concentration of 5 x 10 4 cells/well in 8-well chambered cover glass slides (Ibidi μ-slide 8 well™ slides, Ibidi, Munich, Germany). Real time confocal imaging was performed 48 hours after transfection with a spinning disk confocal system (UltraVIEW VoX; Perkin Elmer, Waltham, MA) connected to a Zeiss AxioObserver Z1 microscope (Zeiss, Oberkochen, Germany). GFP-constructs were excited with a wavelength of 488 nm. Hoechst 33342 (Thermo Fisher Scientific, Waltham, MA, USA) was excited with 405 nm. ER Tracker Red (Thermo-Fisher) was added 100 nM to the cells and excited with 561 nm. Images were acquired with the Volocity software (Perkin Elmer) using a 40 x water immersion objective with a numerical aperture of 1.2.

Feeding of cells with 1-O-pyrenedecyl-sn-glycerol
To monitor the formation of labeled plasmalogens in intact cells, cells were fed with 5 µM 1-Opyrenedecyl-sn-glycerol (Otava, Vaughan, Ontario, Canada) which was first dissolved 4 mM in ethanol, diluted 100 fold with culture medium and then added to the cells.

Cell harvest and lipid extraction
Cells were collected by centrifugation, washed with phosphate-buffered saline, the phosphatebuffered saline aspirated, the pellets shock-frozen in liquid nitrogen and stored at -80°C until analyzed for PEDS activity, lipid content, total protein amounts and amounts of 6xmyc-tagged protein or ß-actin by Western blot. Lipids were extracted twice with 500 µl chloroform/methanol (2/1 v/v) and the combined organic phases dried. The dried lipid extract was taken up in 100 µl acetonitrile/ethanol (1/1 v/v) and stored at -20°C until analyzed.

Measurement of PEDS enzymatic activity
PEDS activity assays were performed with cellular homogenates using the fluorescent 1-Opyrenedecyl-sn-glycero-3-phosphoethanolamine as substrate which was purified from lipid extracts of 1-O-pyrenedecyl-sn-glycerol treated RAW.12 cells as described(1). Briefly, extracts were treated with HCl to cleave residual alk-1'-enyl compounds, purified by a first reversed phase HPLC step, then the 2-acyl residue cleaved off by NaOH treatment, and the lyso compound isolated from a second reversed phase HPLC step, the solvent removed by evaporation, stored at -20°C and aliquots reconstituted with methanol to a concentration of 100 µM. Cellular pellets were homogenized in 200 µl buffer containing 0.1 M Tris and 0.25 M sucrose, pH 7.6 (HCl) using glass beads and four cycles of 30 s and 20 Hz in a Retsch MM400 homogenizer (Retsch, Haan, Germany). To ensure measurements in the linear range of the assay, protein was then determined by a Bradford assay (Biorad, Vienna, Austria) and the homogenates diluted with 3.5 mg/ml bovine serum albumin in phosphate buffered saline to yield cellular protein of 0.5 mg/ml (for expression plasmid testing) or 1 mg/ml (for si-RNA and mouse tissue testing). 7.5 µl of the diluted homogenates were mixed with 5 µl assay mix to start the reaction. The assay mix contained the following components (2.5 fold of the final assay concentrations); 0.25 M Tris.HCl, pH 7.2, 0.25 mg/ml catalase (C1345, Sigma), 2.5 mM NADPH, 5 mM EDTA, 5 µM 1-O-pyrenedecyl-sn-glycero-3phosphoethanolamine (added from a 100 µM stock in methanol). After 30 min at 37°C, the reaction was stopped by adding 37.5 µl HCl/acetonitrile (105 parts 2 M aqueous HCl and 895 parts acetonitrile v/v), incubated for 30 min at 37°C, centrifuged for 5 min at 20000 g and 4°C, and 10 µl injected to an Agilent 1200 HPLC system (Agilent, Vienna, Austria). A Zorbax Eclipse XDB C8 column (4.6mm x 50mm, 3.5 µm particle size, Agilent) thermostatted to 25°C was eluted with a flow rate of 1 ml/min with 10 mM potassium phosphate buffer, pH 6.0, containing 79% (v/v) methanol for 3 min followed by a linear gradient to 100% methanol at 10 min. Methanol (100%) was held until 15 min, changed with a linear gradient to starting buffer until 15.5 min, which was held to 17 min. Pyrene labeled compounds were detected by fluorescence (340 nm excitation, 405 nm emission) and quantified by comparing peak areas in relation to external standards of pyrenedecanoic acid (Sigma). Retention time of the product pyrenedecanal was calibrated with external synthetic standard (Ramidus, Lund, Sweden). Pyrenedecanal was not used for quantification of the amounts by area comparison since it was less stable as pyrenedecanoic acid. Parallel incubations were stopped with acetonitrile containing the same amount of aqueous acetic acid instead of HCl to control for free pyrenedecanal formed in the incubations. Pyrenedecanal in these incubations was always below the detection limit, indicating all pyrenedecanal observed in the incubation mixtures originated from alk-1'-enyl compounds.

Measurement of labeled alkyl and alk-1'-enyl lipids
Pyrene labeled alkyl and alk-1'-enyl lipids were quantified by reversed phase HPLC with fluorescence detection with the same reversed phase HPLC system as described above for PEDS activity determinations, but with a slightly modified elution protocol. The column was eluted with 10 mM potassium phosphate buffer, pH 6.0, containing 81.25% (v/v) methanol for 5 min followed by a linear gradient to 100% methanol at 10 min. Methanol (100%) was held until 21 min, changed with a linear gradient to starting buffer until 21.5 min, which was held to 23 min. Pyrene labeled compounds were detected by fluorescence (340 nm excitation, 405 nm emission) and quantified by comparing peak areas in relation to external standards of pyrenedecanoic acid (Sigma). To determine the amount of pyrene-labeled alk-1-'enyl lipids, lipid extracts were treated with HCl which cleaves the alk-1'-enyl ether bond to the respective aldehyde. Controls were treated with acetic acid which leads to the same aldehyde derivatives as HCl but leaves the alk-1'-enyl ether bond intact. These control incubations would result in the detection of free pyrenedecanal in the cell extracts which was, however, never observed. In a variation of our described protocol (1) we treated the extracts with HCl or acetic acid in methanol rather than in acetonitrile to monitor the liberation of pyrenedecanal from pyrene-labeled alk-1'-enyl lipids. Under these conditions, pyrenedecanal was converted to pyrenedecanal dimethylacetal which was better resolved as the free pyrenedecanal from the peak of 1-O-pyrenedecyl-sn-glycerol used to feed the cells. 10 µl lipid extract was mixed with 40 µl HCl/methanol (70 parts 2 M aqueous HCl/930 parts methanol v/v) or with 40 µl acetic acid/methanol (70 parts 2 M aqueous acetic acid/930 parts methanol v/v) and incubated for 30 min at 37°C to ensure complete cleavage and derivatization. The retention times of the pyrenedecanal derivatives as well as their formation under the indicated conditions were evaluated using synthetic pyrenedecanal (Ramidus AB, Lund, Sweden).

Measurement of total unlabeled plasmalogen
This was performed as described (1). Briefly, 10 µl lipid extracts were derivatized with 40 µl 0.45 mg/ml dansylhydrazine (Sigma) in acetonitrile in presence of either HCl (105 parts 2 M aqueous HCl/895 parts acetonitrile v/v) or acetic acid (105 parts 2 M aqueous acetic acid/895 parts acetonitrile v/v). The resulting aldehyde derivatives were separated with the same chromatographic method as described above for the determination of PEDS enzymatic activity except for the detection by fluorescence at excitation 340 nm, emission 525 nm. The method was calibrated with external standards 1-(1Z-octadecenyl)-2-oleoyl-sn-glycero-3-phosphoethanolamine, Avanti Polar Lipids, Alabaster, AL, USA) and hexadecanal which had been synthesized as described in (4). The amount of free aldehydes measured by incubations with acetic acid was subtracted from the amount of plasmalogens determined by incubations with HCl. The amount of free aldehydes was, however, typically below 1% of the measured amount of plasmalogens.

LC-MS/MS of glycerophosphocholines and glycerophosphoethanolamines
Analysis was performed similarly to (5) in a modified and extended version allowing quantification and fragmentational elucidation of glycerophosphocholines and glycerophosphoethanolamines. Briefly, the harvested cell pellets (approximately 1 million cells) were stored at -80°C prior to Folch lipid extraction (6), at the beginning of which 5 µM PC (28:0) and PE (28:0) were added as internal standards. Cells were broken in 200 µl deionized water with the help of glass beads (Sigma) and a 10 µl aliquot of the aqueous sample was removed prior to lipid extraction for protein content measurement by a Bradford assay (Biorad, Vienna, Austria) with bovine serum albumin as standard. Samples were then shaken, sonicated on ice, centrifuged at 20000 x g and 4°C, and the lower organic residue transferred to a new glass vial and extraction was repeated without sonication. For LC-MS/MS measurements the extracts were dissolved in HPLC starting solvent (54% A, 46% B (v/v), adjusted by protein content and 10 µl separated by reversed-phase HPLC using a Poroshell 120 EC-C8 2.7 μm 2.1 x 100 mm column (Agilent). Solvent A consisted of 10 mM ammonium formate, 0.2% formic acid in isopropanol/acetonitrile (9/1 v/v). Solvent B consisted of 10 mM ammonium formate, 0.2% (v/v) formic acid in acetonitrile/water 6/4 (v/v). At a flow rate of 0.4 ml/min, the column was first eluted for 2 min with 54% A, 46% B (v/v) followed by a linear gradient to 28% A, 72% B (v/v) at 22 min. After rinsing with 100% B and 0.6 ml/min from 23 to 28 min, the column was re-equilibrated for another 2 min at starting conditions before the next injection. A Dionex Ultimate 3000 HPLC (Thermo Fisher Scientific) coupled to a Velos Pro Dual-Pressure Linear Ion Trap Mass Spectrometer (Thermo Fisher Scientific) was employed. A dilution series of (28:0) and (36:2) PC and PE standards was measured at the beginning and end of the batch. 5 µl of each sample were pooled as quality control sample measured before and after the samples. The samples were measured consecutively in randomized order. Data analysis was performed in Mzmine2 (7), Version 2.40 applying an in-house built custom database paired with manual correction and verification on baseline corrected data. Quantification was performed by external calibration using our in-house R data analysis workflow. For data analysis class-wise normalized profiles were used for mean and SD calculation and plotting. Only a handful of different 1-O-alkyl (plasmanyl) lipids and 1-O-alk-1'enyl (plasmenyl) lipids are commercially available, therefore their behavior (relative to class specific standards and other phospholipids) was extrapolated. Briefly, we found for isobaric 1-O-alkyl (plasmanyl) lipids and 1-O-alk-1'-enyl (plasmenyl) lipids (for PE as well as PC) that the 1-O-alkyl (plasmanyl) lipids have a 45-60 s shorter retention time on our reversed-phase HPLC system as compared to their isobaric plasmenyl (1-Oalk-1'enyl) counterparts. This was matched by the respective MS/MS fragmentation behavior. With our ESI ion source and CID fragmentation, plasmanyl as well as plasmenyl lipids fragment at the sn2 fatty acyl residue, yielding a mass for the sn2-acyl residue and the complementary headgroup-1-O-alkyl/alk-1'-enyl residue. Together with the precursor mass and retention time, this constrains the possible molecular structure down to the fatty acyl level, but does for example not include information about the double bond position on the sn2 residue. See Supplementary Figure 5 for an example for retention times and the fragmentation spectra for selected plasmanyl-and plasmenyl lipid species.

Harvest of mouse tissues for analysis of PEDS activity and plasmalogen content
Animal breeding was approved by the Austrian Ministry of Education, Science, and Culture (BMBWF-66.011/0100-V/3b/2019). Tmem189tm1a(KOMP)Wtsi mice were obtained from the Welcome Sanger Institute (Hinxton, Cambridge, UK) and were maintained on C57bl/6N genetic background. The transgene (tg) was a knockout-first allele leading to an inactive truncated Tmem189 protein by artificially splicing a galactosidase reporter downstream of exon 2 (8). Mice were housed in individual ventilated cages with nesting material, in a 12 h/12 h light/dark cycle with standard chow and water ad libitum. Genotyping was performed as recommended by the supplier, with primers Tmem189_35635F (GCGTGTCCTGCTGAGACTTG) and CAS_R1_Term (TCGTGGTATCGTTATGCGCC) for the transgenic allele and Tmem189_35635F and Tmem189-35635R (CATCCCACCTATCCCACCTG) for wildtype allele. Mice were weighed weekly from 3-8 weeks of age. For tissue harvest 8 week old female and male homozygous Tmem189 deficient mice and their heterozygous and wild type littermates were sacrificed by cervical dislocation. Tissues were snap frozen in liquid nitrogen and stored at -80°C until further analysis.

Statistical calculations
Statistical calculations were performed with PRISM (version 8) GraphPad software, San Diego, CA. USA. Transgene (tg) loci contain an additional cassette downstream exon 2 with a splice acceptor site which causes the formation of a truncated version of the Tmem189 protein fused to a galactosidase protein thereby inactivating the enzymatic activity (8). Lipid extracts of tissues obtained at 8 weeks of age were treated with HCl or acetic acid in presence of dansylhydrazine thus creating fluorescent derivatives from plasmalogens plus free aldehydes (in the presence of HCl), or free aldehydes only (in presence of acetic acid (1)). Free aldehydes were typically less than 1% as compared to plasmalogens and subtracted from the amounts formed in presence of HCl. Results for the analysis of tissues from three male and three female mice (mean ± SEM) are shown. wt, wild type. Blue, male animals, magenta, female animals.