Composition dependence of cholesterol flip-flop rates in physiological mixtures
Introduction
Cholesterol in the human body is taken up from the diet or synthesized in the liver, trafficked to cells by lipoprotein particles, and stored in cellular membranes, especially the plasma membrane, (Lange et al., 1989) endosomes, and the trans Golgi (Van Meer et al., 2008; Ikonen, 2008) Within the cell, the concentration of cholesterol in different membranes varies widely and is tightly controlled. For example, cholesterol varies from only a few mole percent (<5 %) in the endoplasmic reticulum (ER) to 30–40 mole percent in the plasma membrane (PM) — an observation made all the more remarkable by the fact that the two membranes are physically connected at the PM-ER junction (Mesmin et al., 2013). Because of its abundance and critical role in regulating membrane physical properties, cholesterol trafficking and homeostasis are critical to a wide range of cellular and physiological processes.
Cholesterol homeostasis requires interleaflet flip-flop, since cholesterol is either synthesized exclusively on the inner leaflet of the ER or delivered to the lumenal leaflet in endosomes after digestion of cholesterol esters from lipoprotein particles (Ikonen, 2008). Following the delivery of cholesterol to the plasma membrane by either mechanism, the ratio of cholesterol in each leaflet will be equilibrated by flip-flop, on a timescale set by the typical time for a spontaneous flip-flop event (Lange et al., 1981). If the lipid composition in each leaflet were identical, then the abundance of cholesterol (Allender et al., 2019; Ayuyan and Cohen, 2018) in each leaflet would be equal, but this need not be the case. It is well established that the lipid composition of the plasma membrane is asymmetric and maintained by active processes, (Lorent et al., 2020) which may well lead to an asymmetric distribution of cholesterol between the leaflets, even when cholesterol flip-flop is fast enough for tyhe two populations to be in equilibrium. Cho and coworkers, for example, recently reported a 12-fold higher concentration of cholesterol in the outer leaflet as compared to the inner leaflet (Liu et al., 2017; Steck and Lange, 2018).
The rate of cholesterol and lipid flip flop between leaflets of biological membranes is therefore important both to understanding cholesterol trafficking processes, and for establishing bounds on the rate at which energy must be consumed to maintain asymmetry. In this context, it is noteworthy that the range reported for the rate of cholesterol flip-flop varies over six orders of magnitude, from μsec−1 to sec−1 (Bennett et al., 2009; Bennett and Tieleman, 2012; Bruckner et al., 2009; Filipe et al., 2018; Garg et al., 2011; Gu et al., 2019; Javanainen and Martinez-Seara, 2019; Jo et al., 2010; Steck et al., 2002). This range includes both experimental measurements and various types of simulations.
An extensive prior literature reports cholesterol flip-flop rates and translocation free energies via a variety of simulation techniques including unbiased all-atom simulation, (Filipe et al., 2018; Gu et al., 2019; Javanainen and Martinez-Seara, 2019) umbrella sampling methods, (Bennett et al., 2009; Jo et al., 2010) and replica molecular dynamics, (Choubey et al., 2013) obtaining a range of flip-flop rates that depend strongly (over 3–4 orders of magnitude) on composition — highly saturated, ordered membranes reduce flip-flop rates relative to unsaturated, disordered membranes. However, all previous publications have simulated very simple membranes, usually comprised of at most a few lipid types most often with two identical acyl chains, representative of neither the complexity nor the typical lipid structure of mammalian lipidomes.
In this work, cholesterol flip-flop rates from 1.6(103) to 1.9(104) sec−1 are observed by unbiased, all-atom molecular dynamics simulation. The simulated compositions are based on lipidomic data for synaptic vesicles (Takamori et al., 2006) and an asymmetric lipidome (inner and outer leaflet obtained separately) of the human red blood cell (Lorent et al., 2020). These compositions contain between 4 and 11 different lipids, carefully selected to capture the headgroup, chain, and backbone chemistry of natural lipidomes. The high melting temperature lipids in these lipidomes are almost entirely sphingolipids, and the lower melting temperature lipids have a variety of hydrophobic chains at the sn-1 and sn-2 position. The present work therefore adds important new insights into cholesterol dynamics across physiologically relevant variations in composition, across leaflets, and among different membranes.
Section snippets
Methods
The four membrane systems were built using packmol (Martínez et al., 2009) to assemble the numbers of lipids indicated in Table 1. Each system was hydrated with at least 50 TIP3P waters per lipid and the salt concentration set to 150 mM NaCl. Lipids and ions were modeled with the CHARMM36 parameters for the lipids (Klauda et al., 2010; Venable et al., 2014; Leonard et al., 2018) and ions with NBFixes applied as described in (Venable et al., 2013). A recent analysis of small angle scattering
Results
Fig. 1 shows the number of cholesterols in each leaflet of all four systems over the course of the simulations, with the location of each cholesterol defined by whether its hydroxyl is above or below the midplane. Cholesterol movement along the bilayer normal appears as discrete steps, with each step having a mirror image in the other leaflet. (These are not flip-flop events, as discussed below.) It is clear that cholesterol is much more dynamic in the less ordered membranes (the synaptic
Discussion and conclusions
Cholesterol flip-flop rates were obtained by unbiased all-atom simulations of physiologically complex lipid mixtures. The compositions of each system were based on lipidomics of representative mammalian membranes: the inner and outer leaflet of the human red blood cell, and the overall lipidome of a synaptic vesicle (with no distinction between inner and outer leaflet, so this system effectively represents a scrambled synaptic vesicle membrane.) An essential difference to previous work on
Acknowledgements
Funding for IL was provided by the NIH/National Institute of General Medical Sciences (GM114282, GM124072, GM120351, GM134949), the Volkswagen Foundation (grant 93091), and the Human Frontiers Science Program (RGP0059/2019). EL and SB were supported by NIH/National Institute of General Medical Sciences GM120351. This work used the Extreme Science and Engineering Discovery Environment (XSEDE), which is supported by US National Science Foundation grant number ACI-1548562. Anton 2 computer time
References (39)
- et al.
Cholesterol-dependent bending energy is important in cholesterol distribution of the plasma membrane
Biophys. J.
(2019) - et al.
The chemical potential of plasma membrane cholesterol: implications for cell biology
Biophys. J.
(2018) - et al.
Molecular simulation of rapid translocation of cholesterol, diacylglycerol, and ceramide in model raft and nonraft membranes
J. Lipid Res.
(2012) - et al.
Flip-flop-induced relaxation of bending energy: implications for membrane remodeling
Biophys. J.
(2009) - et al.
Cholesterol translocation in a phospholipid membrane
Biophys. J.
(2013) - et al.
Noninvasive neutron scattering measurements reveal slower cholesterol transport in model lipid membranes
Biophys. J.
(2011) - et al.
Spontaneous curvature, differential stress, and bending modulus of asymmetric lipid membranes
Biophys. J.
(2020) - et al.
The rate of transmembrane movement of cholesterol in the human erythrocyte
J. Biol. Chem.
(1981) - et al.
Plasma membranes contain half the phospholipid and 90% of the cholesterol and sphingomyelin in cultured human fibroblasts
J. Biol. Chem.
(1989) - et al.
XA four-step cycle driven by PI(4)P hydrolysis directs sterol/PI(4)P exchange by the ER-Golgi Tether OSBP
Cell
(2013)
Numerical integration of the cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes
J. Comp. Phys.
Hexagonal substructure and hydrogen bonding in liquid-ordered phases containing palmitoyl sphingomyelin
Biophys. J.
Probing red cell membrane cholesterol movement with cyclodextrin
Biophys. J.
Molecular anatomy of a trafficking organelle
Cell
CHARMM all-atom additive force field for sphingomyelin: elucidation of hydrogen bonding and of positive curvature
Biophys. J.
Molecular view of cholesterol flip-flop and chemical potential in different membrane environments
J. Am. Chem. Soc.
Particle mesh Ewald: an {N}\cdot \log({N}) method for Ewald sums in large systems
J. Chem. Phys.
Molecular structure of sphingomyelin in fluid phase bilayers determined by the joint analysis of small-angle neutron and X-ray scattering data
J. Phys. Chem. B
Constant pressure molecular dynamics simulation: the Langevin piston method
J. Chem. Phys.
Cited by (7)
Deciphering lipid transfer between and within membranes with time-resolved small-angle neutron scattering
2021, Current Topics in MembranesCitation Excerpt :Even in studies of model membrane systems, where lipid composition is controlled, the reported transfer rates have been inconsistent. For example, the reported half-life for cholesterol's transmembrane flipping varies by five to six orders of magnitude, ranging from several hours (Brasaemle, Robertson, & Attie, 1988; Poznansky & Lange, 1978; Rodrigueza, Wheeler, Klimuk, Kitson, & Hope, 1995) to a few minutes or seconds (Backer & Dawidowicz, 1981; John, Kubelt, Muller, Wustner, & Herrmann, 2002; Leventis & Silvius, 2001; Schroeder et al., 1996; Steck, Ye, & Lange, 2002), and to even a few milliseconds (Baral, Levental, & Lyman, 2020; Bruckner, Mansy, Ricardo, Mahadevan, & Szostak, 2009) to tens of nanoseconds (Bennett, MacCallum, Hinner, Marrink, & Tieleman, 2009; Gu, Baoukina, & Tieleman, 2019). Noninvasive approaches like time-resolved small angle neutron scattering (TR-SANS) or sum-frequency generation vibrational spectroscopy (SFGVP) have shown that the movement of lipids is extremely sensitive to slight chemical structure differences, finding that the transfer rates of unaltered lipid molecules are dramatically different from their chemically tagged counterparts (for example, due to a fluorescent label) (Garg, Porcar, Woodka, Butler, & Perez-Salas, 2011; Liu & Conboy, 2005).
Overlay databank unlocks data-driven analyses of biomolecules for all
2024, Nature CommunicationsLiPyphilic: A Python Toolkit for the Analysis of Lipid Membrane Simulations
2021, Journal of Chemical Theory and Computation