Plasmalogen biosynthesis (oxygen-dependent alkyl-ether desaturation route and anaerobic diacyl-lipid reduction route)

Formation of plasmalogens: glycerolipids that carry a 1-(1Z-alkenyl) (vinyl) ether at sn-1, usually with an acyl chain at sn-2 and a phospho or glyco head group at sn-3. The class is defined by the vinyl-ether bond alone, so it spans hundreds of molecular species that differ in head group, sn-1 alkenyl chain and sn-2 acyl chain. The vinyl ether is made by two routes that share no enzymes and no intermediates. In animals and other eukaryotes the oxygen- dependent route builds an ether lipid from scratch. Peroxisomal GNPAT acylates dihydroxyacetone phosphate, and AGPS exchanges that acyl chain for a fatty alcohol supplied by FAR1, forming the ether bond. PexRAP then reduces the 1-alkyl-DHAP, and ER enzymes acylate, dephosphorylate and add the ethanolamine head group (SELENOI) to give plasmanyl-PE. Finally the PEDS1 (TMEM189) desaturase introduces the vinyl-ether double bond with O2 and cytochrome b5. PEDS1 and its myxobacterial ortholog CarF belong to one PANTHER subfamily, and the eukaryotic pathway is thought to have come from myxobacteria by horizontal gene transfer. In obligate and facultative anaerobic bacteria the vinyl ether is made without an alkyl-ether intermediate and without oxygen. A plasmalogen synthase of the benzoyl-CoA reductase / 2-hydroxyacyl-CoA dehydratase family reduces the sn-1 acyl ester of a complete diacyl phospholipid directly to the alk-1'-enyl ether. This enzyme is either one polypeptide (PlsA) or two (PlsA with PlsR). Because the bacterial enzyme acts on whichever diacyl lipids the cell makes, it gives plasmenyl PE, PG and other head groups. Outside the module are the supply of diacyl phospholipids and of CDP-ethanolamine, plasmenylcholine formation and acyl-chain remodelling (for example EC 2.3.1.25 lysoplasmalogen acyltransferases), and plasmalogen turnover.

Modelling choices. (1) The two routes share no step, substrate or intermediate, so they are whole-pathway alternatives in one variant set. The eukaryotic route is included by reference to ether_lipid_plasmalogen_biosynthesis, not copied. (2) GO:0008611 is the most specific process term available. (3) The anaerobic step has no GO MF term, Rhea reaction or EC number, so its annotons carry a described function without a term, as for BIO1 in the biotin module. A GO new-term request and a relabelling of GO:0047159 ("plasmalogen synthase activity", an EC 2.3.1.25 lysoplasmalogen acyltransferase) have been drafted. (4) PANTHER PTHR32329:SF4 is named after 2-hydroxyacyl-CoA dehydratase activators, but its experimentally characterized members are plasmalogen synthases. The broader PTHR32329 family mixes these with genuine dehydratase activators (HgdC, LcdC, HadI, FldI) and benzoyl-CoA reductase subunits. (5) Myxobacteria use the CarF desaturase. Their upstream ether-bond enzymes are not modelled, so the eukaryotic variant is grounded in eukaryotic proteins apart from the desaturase step. (6) Naming: "PlsA" is the Act1-Act2-Red1 subunit in C. perfringens (PMID:33350306) but the whole single-chain enzyme in E. faecalis and Lactococcus. The Selenomonas ruminantium enzyme is called SrPlsAR (PMID:39725447) and the Megasphaera enzyme MeHAD (PMID:37348645). Labels here follow each organism's paper. (7) SrPlsAR is not a representative member because the cached abstract gives no locus tag or accession to ground it.