Regulation of lipid saturation without sensing membrane fluidity

Abstract

Cells maintain membrane fluidity by regulating lipid saturation, but the molecular mechanisms of this homeoviscous adaptation remain poorly understood. We have reconstituted the core machinery for regulating lipid saturation in baker's yeast to study its molecular mechanism. By combining molecular dynamics simulations with experiments, we uncover a remarkable sensitivity of the transcriptional regulator Mga2 to the abundance, position, and configuration of double bonds in lipid acyl chains, and provide insights into the molecular rules of membrane adaptation. Our data challenge the prevailing hypothesis that membrane fluidity serves as the measured variable for regulating lipid saturation. Rather, we show that Mga2 senses the molecular lipid-packing density in a defined region of the membrane. Our findings suggest that membrane property sensors have evolved remarkable sensitivities to highly specific aspects of membrane structure and dynamics, thus paving the way toward the development of genetically encoded reporters for such properties in the future.

Fig. 1. The activation of Mga2 is controlled by the ER membrane composition.

a Model of the OLE pathway: the transcription factor Mga2 forms inactive dimers in the ER membrane (Mga2 p120^dimer) with highly dynamic TMHs exploring alternative rotational orientations. Loose lipid packing (left) caused by unsaturated lipids stabilizes conformations with two sensory tryptophan residues (W1042; red) pointing away from the dimer interface toward the lipid environment. Tight lipid packing (right) stabilizes alternative rotational conformations with the sensory tryptophans facing each other in the dimer interface (right). The E3 ubiquitin ligase Rsp5 is required to ubiquitylate (Ub) Mga2, thereby facilitating the proteolytic processing by the proteasome and the release of transcriptionally active Mga2 (p90). b Secondary structure prediction of the juxtamembrane and transmembrane region (residue 951–1062) of Mga2 using Phyre2. Source data are provided as a Source Data file.

Fig. 2. An in vitro sense-and-response system for membrane lipid saturation.

a Schematic representation of the sense-and-response constructs. The fusion proteins are composed of the maltose-binding protein (MBP, blue) and Mga2^950–1062 (green) with the Rsp5-binding site (LPKY), three lysine residues as targets of ubiquitylation (K⁹⁸⁰, K⁹⁸³, and K⁹⁸⁵), a predicted disordered juxtamembrane region, and the C-terminal TMH. An optional N-terminal leucine zipper derived from Gcn4 (gray, Gcn4^249–281) supports dimerization. b Isolation of the zipped sense-and-response construct by affinity purification. 0.1 OD units of the lysate (L), soluble (S), flow-through (FT), and two wash fractions (W_1,2), as well as 1 µg of the eluate were subjected to SDS-PAGE followed by InstantBlue^TM staining. The protein was further purified by preparative SEC (Supplementary Fig. 1a). c One hundred micrograms in 100 µl of the purified sense-and-response constructs either with (+ZIP) or without zipper (−ZIP) were loaded onto a Superdex 200 10/300 Increase column (void volume 8.8 ml). d Schematic representation of the in vitro ubiquitylation assay. Proteoliposomes containing ^ZIP-MBPMga2^950–1062 were mixed with ^8xHisUbiquitin (^HisUb), an ATP-regenerating system, and cytosol prepared from wild-type yeasts to facilitate Mga2 ubiquitylation at 30 °C. e The reaction was performed with the ^ZIP-MBPMga2^950–1062 wild-type (WT) construct, a variant lacking the Rsp5-binding site (∆LPKY), and a variant with arginine residues instead of the lysine residues K⁹⁸⁰, K⁹⁸³, and K⁹⁸⁵ (3KR), thus lacking the target residues of ubiquitylation. These variants were reconstituted in liposomes composed of 100 mol% POPC at protein-to-lipid ratio of 1:5000. After the indicated times, the reactions were stopped using sample buffer and were subjected to SDS-PAGE. For analysis, an immunoblot using anti-MBP antibodies was performed. f Ubiquitylation reactions were performed as in e with the WT sense-and-response construct reconstituted in the indicated lipid environments at a molar protein-to-lipid ratio of 1:5000. Source data are provided as a Source Data file.

Fig. 3. FRET reveals membrane-dependent conformational changes in the sense-and-response construct.

a Representation of constructs. The Atto488 dye is linked to K983C at a position, which is ubiquitylated by Rsp5 in vivo. The Atto590 dye was linked to K969C in the Rsp5-binding site. b Fluorescence emission spectra reveal FRET in detergent solution. Each construct (2 µM) were used to record fluorescence emission spectra (ex: 488 nm, em: 500–700 nm) of the donor (K983^{D only}), acceptor (K969^{A only}), and the combined (K983^D + K969^A) FRET pair. c Fluorescence emission spectra were recorded for serial dilutions K983^D + K969^A in detergent solution as in b. The spectra were normalized to the maximal intensity at the donor emission. d Zipped donor (2 µM) and acceptor (2 µM) pairs were mixed and incubated in detergent solution for 10 min, to allow for protomer exchange and equilibration. This sample was titrated with an unlabeled competitor either with (+ZIP) or without zipper domain (−ZIP). Emission spectra were recorded as in b. The relative FRET efficiency was determined from the acceptor-to-donor intensity ratio and plotted as mean ± SD from two independent experiments. e Emission spectra indicate energy transfer within the membrane-reconstituted, dimeric sense-and-response construct. The donor construct was premixed either with an unlabeled (K983^{D only}) or a labeled acceptor construct (K983^D + K969^A) prior to the reconstitution in POPC liposomes at a protein-to-lipid ratio of 1:8000. Fluorescence emission spectra (em: 500–700 nm) upon donor excitation (ex: 488 nm; solid line) and acceptor excitation (ex: 590 nm; dotted line) are plotted. f Donor (K983^D) and acceptor (K969^A) were mixed equimolarly and were incubated in detergent solution prior to a reconstitution in indicated lipid environments. Emission spectra were recorded as in e and were normalized to the maximal acceptor emission after direct acceptor excitation (ex: 590 nm). g The relative FRET efficiency was derived from the fluorescence spectra in f and plotted as mean ± SD of at least three independent reconstitutions (n_DOPC = 4; n_{1:1. (POPC:DOPC)} = 6; n_{3:1(POPC:DOPC)} = 3; n_POPC = 6). A two-tailed, unpaired t-test was performed to test for statistical significance (*p < 0.05, **p < 0.01, ***p < 0.001). Source data are provided as a Source Data file.

Fig. 4. The conformation and activity of the sense-and-response construct does not correlate with membrane viscosity.

a Diffusion coefficients of the fluorescent Atto488-DPPE lipid in giant unilaminar vesicles with indicated compositions were determined by confocal point FCS. Data are represented as mean ± SD (n_DOPC = 172; n_{(1:1)DOPC:POPC} = 81; n_POPC = 153; n_40%POPE = 100) and subjected to a Kolmogorov–Smirnov test (*p < 0.05, **p < 0.01, ***p < 0.001). b The lipid packing of the indicated lipid compositions were determined by C-Laurdan spectroscopy and expressed as generalized polarization (GP). Data are shown as mean ± SD (n_DOPC = 6; n_{(1:1)DOPC:POPC} = 10; n_POPC = 9; n_40%PE = 6) and analyzed using a two-tailed, unpaired t-test (*p < 0.05, **p < 0.01, ***p < 0.001). c cwEPR spectra were recorded at −115 °C for a fusion protein composed of MBP and the TMH of Mga2 (^MBPMga2^1032–1062) labeled at position W1042C after reconstitution at a molar protein:lipid ratio of 1:500 in indicated liposomes. The semi-quantitative proximity index I_Lf/I_Mf indicating the interspin distance was derived from the cwEPR spectra as previously published. Higher values indicate a lower average interspin distance. Data are plotted as mean ± SD (n_0%PE = 6; n_20%PE = 3; n_40%PE = 4) and are analyzed by a two-tailed, unpaired t-test (*p < 0.05; ns not significant). d Relative FRET efficiencies calculated from fluorescence emission spectra (ex: 488 nm, em: 500–700 nm) of the (K983^D + K969^A) FRET pair reconstituted in liposomes composed of 50 mol% DOPC, 10 mol% POPC, and 40 mol% POPE. The FRET efficiencies for the other lipid compositions are the same as in Fig. 3g and shown for comparison. Data are plotted as mean ± SD (n_{(1:1)DOPC,POPC} = 6; n_POPC = 6; n_40%PE = 7) and analyzed using a two-tailed, unpaired t-test (**p < 0.01, ***p < 0.001). e In vitro ubiquitylation of the zipped sense-and-response construct (^ZIP-MBPMga2^950–1062) reconstituted in liposomes of the indicated lipid compositions at a molar protein-to-lipid ratio of 1:8000. After the reaction was stopped, the samples were subjected to SDS-PAGE and analyzed by immunoblotting using anti-MBP antibodies. f Densiometric quantification of ubiquitylation from immunoblots as in e was performed using Image Studio^TM Lite. Data are plotted as mean ± SD (n_DOPC = 20; n_{(1:1)DOPC,POPC} = 12; n_POPC = 18; n_40%PE = 14) and analyzed using a two-tailed, unpaired t-test (*p < 0.05, **p < 0.01, ***p < 0.001). Source data are provided as a Source Data file.

Fig. 5. The double bond in unsaturated lipid acyl chains affects the configuration and activity of the sense-and-response construct.

a Diffusion coefficients of the fluorescent lipid Atto488-DPPE in GUVs with indicated compositions were determined by confocal point FCS. Data for DOPC (∆9-cis) are the same as in Fig. 4a. Data are plotted as mean ± SD (n_(∆9-cis) = 172; n_(∆6-cis) = 162; n_(∆9-trans) = 163) and analyzed using Kolmogorov–Smirnov tests (***p < 0.001). b Lipid packing in liposomes was determined by C-Laurdan spectroscopy. Data for DOPC (∆9-cis) are the same as in Fig. 4b. Data are plotted as mean ± SD (n_(∆9-cis) = 6, n_(∆6-cis) = 6; n_(∆9-trans) = 5) and analyzed via unpaired two-tailed, Student’s t-tests (***p < 0.001). c Intensity-normalized cwEPR spectra recorded at −115 °C for ^MBPMga2^1032–1062 labeled at position W1042C after reconstitution at a molar protein:lipid ratio of 1:500 in liposomes with indicated composition. d Semi-quantitative proximity index I_Lf/I_Mf derived from cwEPR spectra. High values indicate low average interspin distances. Data are plotted as mean ± SD (n_(∆9-cis) = 5; n_(∆6-cis) = 6; n_(∆9-trans) = 4) and analyzed using two-tailed, unpaired t-tests (***p < 0.001). e Fluorescence emission spectra of the membrane-reconstituted FRET pair (K983^D + K969^A) (ex: 488 nm, em: 500–700 nm) are plotted after normalization to the maximal emission upon acceptor excitation (ex: 590 nm). Data for DOPC (∆9-cis PC) are the same as in Fig. 3f. f Relative FRET efficiencies were calculated as in e, plotted as mean ± SD (n_(∆9-cis) = 4; n_(∆6-cis) = 4; n_(∆9-trans) = 6), and analyzed by two-tailed, unpaired t-tests (*p < 0.05; **p < 0.005). Data for DOPC (∆9-cis PC) are the same as in Fig. 3g. g In vitro ubiquitylation of the zipped sense-and-response construct (^ZIP-MBPMga2^950–1062) reconstituted in indicated liposomes at a molar protein-to-lipid ratio of 1:8000 and at 30 °C. After stopping the reaction, the samples were subjected to SDS-PAGE and analyzed by immunoblotting using anti-MBP antibodies. h Densiometric quantification of ubiquitylation at the indicated time points from immunoblots as in g). Data are plotted as mean ± SD (n_(∆9-cis) = 20; n_(∆6-cis) = 9; n_(∆9-trans) = 9. Data for DOPC (∆9-cis PC) are the same as in Fig. 4f. Source data are provided as a Source Data file.

Fig. 6. The activity of Mga2 is tuned by mutations in the sensory TMH.

a Dose-dependent rescue of UFA auxotrophy by linoleic acid (18:2). ∆SPT23∆MGA2 strains carrying CEN-based plasmids to produce ^MycMga2 variants with the indicated residues at position 1042 were cultivated for 16 h at 30 °C in SCD-Ura medium supplemented with indicated concentrations of linoleic acid in 0.8% tergitol. The density of the culture was determined at 600 nm (OD₆₀₀) and plotted against the concentration of linoleic acid. Cells carrying an empty vector served as control (gray). Plotted is the mean ± SEM (n = 8). b Rescue of UFA auxotrophy of ∆SPT23∆MGA2 by Mga2 variants. Cells producing mutant Mga2 as in a were cultivated for 24 h in the absence of supplemented UFAs in SCD-Ura medium. Cell density was determined as in a and was plotted against residue surface area of residues installed at position 1042. Plotted is the mean ± SEM of five independent experiments. The dotted line indicates the OD measured for an empty vector control. c Immunoblot analysis of the Mga2 processing efficiency. Wild-type cells (BY4741) producing the indicated ^MycMga2 variants at position 1042 were cultivated in full medium (yeast extract–peptone–dextrose; YPD) to the mid-exponential phase. Cell lysates were subjected to SDS-PAGE and analyzed via immunoblotting using anti-Myc antibodies to detect the unprocessed (p120) and the processed, active form (p90) of Mga2. An immunoblot using anti-Pgk1 antibodies served as loading control. d Densiometric quantification of the ratio of p90:p120 in immunoblots as in c. Signal intensites were quantified using Fiji and plotted as mean ± SD (n_W = 6, n_Y = 5, n_F = 5, n_Q = 5, n_L = 6, n_A = 6). A two-tailed, unpaired t-test was performed to test for statistical significance (*p < 0.05, **p < 0.01, ***p < 0.001). Source data are provided as a Source Data file.

Fig. 7. The local number density of lipid atoms in a specific region of the membrane correlates with Mga2 activation.

a Snapshot from an all-atom MD simulation of a protein-free DOPC bilayer (left) and the derived local number density of lipid atoms (right). A schematic representation of the dimeric TMH of Mga2 in the left panel is provided to indicate the relative position of the sensory tryptophan W1042 (red sticks and dashed line). The position of double bonds in the lipid acyl chains is given by green spheres, whereas a schematic representation of the Mga2-TMH is given to guide the eye. The number density of lipid atoms in cubic boxes with a side of 1 Å was calculated and plotted (right panel) for different depths in the bilayer (distance for bilayer center) and different lateral positions (lateral dimension). Highest local number densities are indicated in yellow and are observed in the region of the lipid headgroups. Lowest densities are indicated in dark blue and observed in the center of the lipid bilayer. For details see the Supplementary Materials. b Difference maps of the local density were determined by subtracting the local number density of lipid atoms of DOPC from the one of the indicated bilayer systems. Increased packing densities relative to the DOPC bilayer are indicated in yellow, whereas decreased densities are indicated in blue. The region probed by the sensory tryptophan is indicated by two dashed lines. Changes in lipid saturation and the lipid headgroup region have qualitatively distinct impact on the distribution of local packing densities. Source data are provided as a Source Data file.