Protein glycosylation is essential for all eukaryotes, from disease-causing protists such as malaria and trypanosomes, to yeast and mammals. Secretory proteins are almost invariably N -glycosylated, O-and C -mannosylated, and/or GPI-anchored as they enter the lumen of the endoplasmic reticulum (ER).All ER protein glycosylation reactions occur in the lumen and often involve lumenal mannosylation and glucosylation steps in which mannose and glucose residues are sourced from the glycolipids mannosyl-and glucosyl-phosphoryl dolichol (MPD and GPD, respectively) [1,2]. Paradoxically, these two lipids are synthesized on the cytoplasmic face of the ER and must therefore be flipped across the ER membrane to provide a source of lumenal mannose and glucose. As the spontaneous rate of MPD and GPD flipping is extremely low, specific transporters are needed to facilitate the transbilayer movement of MPD and GPD across the ER membrane at a physiological rate. MPD and GPD transport activities have been demonstrated and characterized in ER microsomes, as well as in vesicles reconstituted with ER membrane proteins [3][4][5][6]. The transport proteins have been shown to be highly structure specific, discriminating between isomers of their lipid substrates, and facilitating lipid movement bidirectionally in an ATP-independent manner (the last point defines them as scramblases, whereas they were previously known as ATP-independent flippases). Although most of the enzymes and co-factors of ER protein glycosylation are known, the molecular identities of the critical dolichol glycolipid scramblases remain a mystery.Unlike MPD scramblase that is required for all ER protein glycosylation reactions, GPD scramblase is needed exclusively for the synthesis of the glucosylated N -glycan precursor Glc 3 Man 9 GlcNAc 2 -PP-dolichol (G3M9-DLO). Even though non-glucosylated N -glycan precursors are substrates for the protein N -glycosylation machinery, the presence of the tri-glucosyl cap -and hence GPD scramblase activity -is critically important for glycosylation efficiency in many eukaryotes, including yeast and humans [7]. Two points are noteworthy. (i) Not all organisms have glucose in their N -glycan precursor [8]. (ii) While the synthesis of glucosylated N -glycan precursors is not essential for the viability of yeast [9,10], yeast cells that are deficient in G3M9-DLO synthesis display numerous phenotypes including under-glycosylation of proteins, abnormal cell shape and altered susceptibility to a variety of chemicals. In humans, optimization of oligosaccharide transfer efficiency provided by the glucosyl cap is critical as evinced by severe human diseases.Taking advantage of the fact that not all N -glycosylation-competent organisms have glucose in their N -glycan precursor, we implemented a bioinformatics approach for assignment of protein function, namely phylogenetic profiling. Using this procedure, we identified a number of polytopic ER membrane proteins as GPD scramblase candidates in yeast. *