MIGA2 binds and transfers lipids between membranes. (A) Left: Native mass spectra of hMIGA2_long-6xhis purified from bacteria, showing exclusive presence of the dimeric protein. Each charge state is further accompanied by a series of up to four lipid bound peaks. Right: Expansion of the 20 + charge state of native MS for hMIGA2_long-6xhis dimer, showing up to four glycerophospholipids bound, with the average mass of 752 ± 5 Da. (B) Untargeted lipidomics of 3xFLAG-hMIGA2_long, purified from mammalian (Expi293) cells. At left: Distribution of lipid classes bound to hMIGA2_long. At right: Distribution of glycerophospholipids bound to hMIGA2_long as compared to their abundance in cells (Lees et al., 2017). (C) In the end-point transfer assay, hMIGA2 constructs are tethered between “heavy” acceptor liposomes (containing 0.75 M sucrose; lipid composition: 65% DOPC, 30% PE, and 5% PI[4,5]P₂) and light donor liposomes containing a single species of fluorescent lipid (containing no sucrose; lipid composition: 63% DOPC, 30% PE, 2% NBD- or Rh-lipid, 5% DGS-NTA). After lipid transfer (after 30 min), proteinase K was added to digest proteins, the light donor and heavy acceptor liposomes were separated by centrifugation, and the fluorescence increase of the heavy acceptor liposomes monitored. Transferred lipid was quantitated based on liposome standards incorporating 2, 1, 0.5, 0.25, or 0% of the appropriate fluorescent lipid. Both hMIGA2_long and hMIGA2_C transport NBD-PE_acyl, -PS_acyl, -PA_acyl, and NBD-PC_acyl; NBD-PS_head is transferred to a lesser extent, and Rh-PE_head is not transferred, indicating that modification of the lipid headgroup can interfere with transport by MIGA2. In a similar experiment, we monitored transfer of unmodified PS using a PS-specific protein probe (C2-domain of lactadherin [Moser von Filseck et al., 2015]) to assess PS transfer. Donor liposomes initially contained 5% PS. Donor and acceptor liposomes were separated by addition of EDTA and imidazole rather than protease, so as not to destroy the PS-probe. Supernatant (containing donor liposomes) and pellet (acceptor liposomes) fractions, in the presence and absence of MIGA2, were analyzed by SDS-PAGE to quantitate the PS-probe associated with each fraction. Transfer efficiency of acyl chain modified NBD-PS, as monitored by fluorescence, is comparable to that of natural PS, indicating that the acyl chain modification does not interfere with transfer by MIGA2. (D) In the FRET-based transfer assay, donor and acceptor liposomes (compositions indicated) were tethered together in the presence or absence MIGA2 linked to the donor liposomes. The donor liposomes initially contain Rh-PE_head and NBD-lipids (-PS_acyl, -PC_acyl, -PA_acyl, -PE_acyl), where FRET between the Rh and NBD initially reduces NBD fluorescence. As lipids are transferred from donor to acceptor liposomes, the Rh- and NBD-labeled lipids are diluted, resulting in reduced FRET and an increase in NBD fluorescence. hMIGA2_long can transport NBD-PS_acyl, NBD-PA_acyl, NBD-PE_acyl, and NBD-PC_acyl. A donor only control shows that the fluorescence increase was due to transfer of lipids between liposomes rather than solely lipid extraction by hMIGA2. After the transfer reaction was completed, dithionite was added to rule out the possibility of fusion between donor and acceptor liposomes, which would also result in a fluorescence increase (Fig. S2 C). Each experiment was performed in triplicate. SDs are shown. NBD-PS_head transfer is less efficient (Fig. S2 D). (E) A similar FRET experiment was carried out to assess transfer from artificial LDs (rather than donor liposomes) to acceptor liposomes and shows that MIGA2 transfers NBD-PE_acyl and NBD-PC_acyl between the artificial LDs and liposomes. Positive stain transmission electron microscopy was used to assess the quality of the LD preparation (Fig. S2 E) as in Wang et al. (2016).