Modeling Yeast Organelle Membranes and How Lipid Diversity

Subscriber access provided by UNIV OF NEBRASKA - LINCOLN

Article

Modelling Yeast Organelle Membranes and How Lipid Diversity influences Bilayer Properties Viviana Monje-Galvan, and Jeffery B. Klauda Biochemistry, Just Accepted Manuscript • Publication Date (Web): 26 Oct 2015 Downloaded from http://pubs.acs.org on October 27, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Biochemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Modelling Yeast Organelle Membranes and How Lipid Diversity influences Bilayer Properties Viviana Monje-Galvan1 and Jeffery B. Klauda1,2* 1

Department of Chemical and Biomolecular Engineering and 2Biophysics Program, University of Maryland, College Park, MD 20742, USA *To whom correspondence should be addressed: [email protected]

1 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 40

ABBREVIATIONS CC=cross-chain, CG=carbon of glycerol, CHOL=choleseterol, DB=bilayer thickness, DC=half the hydrocarbon thickness, DHH=head-to-head distance, EDP=electron density profile, ER=endoplasmic reticulum, ERG=ergosterol, FA=fatty acid, IPC=inositol phosphoceramide, MD=molecular dynamics, NMR=Nuclear Magnetic Resonance, KA=compressibility modulus, Ld=liquid disordered, Lo=liquid ordered, OPM=orientation of proteins in membranes, PA=phosphatidic acid, PBC=periodic boundary conditions, PC=phosphatidylcholine, PE=phosphatidylethanolamine, PME=particle mesh Ewald, PI=phosphatidylinositol, PS=phosphatidylserine, SA=surface area, SANS=small angle neutron scattering, SAXS=small angle x-ray scattering, SCD=deuterium order parameters, TGN=trans-Golgi network, VDW=van der Waals.


Page 3 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

ABSTRACT Membrane lipids are important for the health and proper function of cell membranes. We have improved computational membrane models for specific organelles in yeast Saccharomyces cerevisiae to study the effect of lipid diversity on membrane structure and dynamics. Previous molecular dynamics (MD) simulations were performed by Jo et al. (Biophys J. 97: p50) on yeast membrane models having six lipid types with compositions averaged between the endoplasmic reticulum (ER) and the plasma membrane (PM). We incorporated ergosterol (ERG), phosphatidic

acid

(PA),

phosphatidylcholine

(PC),

phosphatidylethanolamine

(PE),

phosphatidylserine (PS), and phosphatidylinositol (PI) lipids in our models to better describe the unique composition of the PM, ER, and trans-Golgi Network (TGN) bilayers of yeast. Our results describe membrane structure based on order parameters (SCD), electron density profiles (EDPs), and lipid packing. The average surface area per lipid (SA) decreased from 63.8±0.4 Å2 in the ER to 47.1±0.3 Å2 in the PM, while the compressibility modulus (KA) varied in opposite direction. The high SCD values for the PM lipids indicated a more ordered bilayer core, while the corresponding lipids in the ER and TGN models had lower parameters by a factor of at least 0.7. The hydrophobic core thickness (2DC) as estimated from EDPs is the thickest for PM, which is in agreement with estimates of hydrophobic regions of transmembrane proteins from the Orientation of Proteins in Membranes (OPM) database. Our results show the importance of lipid diversity and composition on a bilayer’s structural and mechanical properties, which in turn influences interactions with the proteins and membrane-bound molecules.


Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 40

1. Introduction Saccharomyces cerevisiae, referred to generally here as yeast, is a common reference organism in modern biology because its complete genome sequence is known and its gene homology is high with other eukaryotes. It is the eukaryote most widely studied over the past 20 years because of the ease to manipulate it experimentally.1, 2 We used molecular dynamics (MD) simulations to study the influence of lipid diversity in a membrane to develop more accurate models for specific organelles in yeast. Specific lipid composition is needed in a given membrane to ensure proper function of the cell; the diversity of each membrane determines the proteins that interact with it to allow its proper function.3 Protein distribution in the membrane bilayer is not random, it is in fact linked to the lipid landscape in a bilayer.1, 4 Lipids are free to move laterally in a bilayer and can form dynamic phases or domains at certain sterol concentrations, which in turn modulate the mechanical properties and function of the membrane.5, 6 Cell signaling and sensing, as influenced by lipid composition, have not yet been fully understood.

1, 2

The motivation to develop organelle-specific models is to use them in

further studies of lipid-lipid and lipid-protein interactions. The goal of this work is to study the effect of lipid composition on the structural, mechanical, and dynamical properties of yeast membranes. Our models contain characteristic lipid compositions for the endoplasmic reticulum (ER), trans-Golgi Network (TGN), and plasma membrane (PM), organelles involved in the secretory pathway of the cell. The ER is a dynamic network of sheets and tubules that expands from the nucleus to the plasma membrane. It is the starting point of the secretory pathway and the organelle where most proteins and lipids are synthesized.5 The TGN is a sorting station for lipids and proteins in the secretory pathway before they are delivered to their final location. This


Page 5 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

membrane resists many of fluxes coming to and from the ER and PM. In addition, the Golgi apparatus was shown to be a major producer of sphingolipids.7-9 Finally, the PM is the outermost membrane of the cell, and also the most rigid bilayer. Sterol composition in this membrane is critical to ensure its proper rigidity and allow cell growth.6 Sterols are by far the predominant non-polar lipids in biological membranes. Ergosterol (ERG), homologue of cholesterol (CHOL) in mammalian cells, is the major sterol representative in yeast.6, 10 For this work, we have developed models for three membranes in yeast, i.e., ER, PM and TGN. This falls in line with our recent publications of all-atom model membranes of bacteria11-14 and a preliminary model for yeast.15. Since we cannot simulate very large membranes, our models are simpler than those recently published using the MARTINI coarse-grained force field.16, 17 Although the PM and TGN membranes are known to have an asymmetric distribution of lipids between leaflets, our models assume equal composition in each leaflet to provide an baseline comparison to future studies of asymmetric membrane models. The compositions of these membranes are based off experimental measures for each organelle.8, 18-20 Specifically, we developed lipid membranes consisting of lipids shown in Figure S1. None of the lipids in our models are fully saturated, they have at least one monounsaturated tail and the chains length varies between 16 or 18 carbons. We selected five different lipid headgroups, including anionic groups such as phosphatidic acid (PA), phosphatidylserine (PS), and phosphatidylinositol (PI); and zwitterioninc groups like phosphatidylcholine (PC) and phosphatidylethanolamine (PE). 2. Methods Cellular processes can take place over short timescales; MD simulations facilitate their study and can be helpful in the selection of experimental targets. Seven membrane models were developed to study the impact of lipid composition in the membrane properties of the ER, TGN,


Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 40

and PM in yeast. The first model, AVG-Yeast, is a replicate of a simple model developed previously based on an averaged lipid composition between the ER and the PM,15 but with an updated lipid force field. More complex models built specifically to model the PM, ER, and TGN were based on experimental data (see also Table S1)

8, 18-20

and available force field

parameters. Each model was constructed using the Membrane Builder platform of CHARMMGUI (www.charmm-gui.org),21-24 an automated graphical user interface to build heterogeneous systems choosing from a library of over 200 lipids. The main simulation box for each model contained 300 lipids evenly distributed in two leaflets, neutralizing ions, and was fully hydrated (40+ waters per lipid) using the TIP3P model for water

25

. Each model was equilibrated on

CHARMM26 using the typical six-step CHARMM-GUI protocol for 225 ps.15, 27 The first two steps run using NVT dynamics, and the remaining four using the NPT ensemble gradually decreasing restraint force constants.21-24 We used CHARMM (equilibration)26 and NAMD (production)28 software packages to run three replicates of each model with the CHARMM36 lipid force field (C36),29 which includes the most updated parameters for PI lipids30 and sterols31 and reproduces more accurately experimental properties similar to those of interest in this work. The NPT ensemble was used during all production runs to complete trajectories of 200 ns for each replica with a 2 fs time-step using the SHAKE algorithm to constraint hydrogen atoms.32 The temperature was kept constant at 303.15K using the Nosé-Hoover thermostat,33 and Langevin dynamics during the production run of each simulation using NAMD software package.34, 35 For constant pressure control at 1 bar we used the Nosé-Hoover piston, 36 allowing the cell box size to change semi-isotropically (X=Y but not Z), and a Langevin piston on NAMD,37 which couples the piston controls to a temperature bath controlled by Langevin dynamics.37-39 van der Waals (VDW) and electrostatics


Page 7 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

were computed using a Lennard-Jones force-switching function over 8 to 12 Å

40

. All

simulations were run using periodic boundary conditions (PBC) that evaluated long-range electrostatic interactions using Particle Mesh Ewald (PME).41 2.1 Systems of study Figure 1 shows the lipid composition of all the models used in this study according to lipid headgroups; specific lipid types, number of waters, and ions for each system are listed in Table S2 in the supplementary material. Average Yeast Model (AVG-Yeast) This model was based on a previous study by Jo et al. to explain the capabilities of the Membrane Builder platform in CHARMM-GUI to construct heterogeneous lipid bilayers.15 The model contained the most representative lipids of yeast cells; however, their relative concentrations were average values between cell organelles taken from an experimental study.19 The Jo et al. study explored the effect of cholesterol concentration on membrane properties, among others. For this paper, the AVG-Yeast model is a control system to show the importance of accurate model representation for specific organelles and updates this model with C36 force field. Given the moderate concentration of sterol in this model, the results for the PM model are expected to match AVG-Yeast better than the ER or TGN models. However, discrepancies in the calculated properties will arise from the structural and unsaturation differences between cholesterol (the sterol in AVG-Yeast) and ergosterol in the new models. The lipid composition for AVG-yeast was taken from the Daum et al. experimental assays.19 Although sphingolipids and PI are analyzed in the experimental tests of this strain, they were not included in the initial theoretical model by Jo et al. as accurate CHARMM parameters were not available at the time of publication. The most representative fatty acids (FA) were experimentally determined to be 7 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 40

palmitoleic (C16:1) and oleic (C18:1) with 62.5% and 26.1% respectively. Small percentages of palmitic (C16:0) and stearic (C18:0) were also present. Complex Models Two models were developed for each organelle of study, varying in lipid concentration, diversity, and unsaturation degree (refer to Table S2). The difference between the two models for each organelle comes from the unsaturation degree of the tails to study the effect of FA saturation in the membrane physical properties; headgroup diversity was conserved in both models. The ER models used in this study were built to reflect the more fluid structure of this organelle.42 This model contains the least amount of ERG, the most representative sterol in yeast, but the highest concentration of PC lipids. The only difference between the two models is the unsaturation degree of the PS lipids; one model uses oleic (18:1) and palmitoleic acid (16:1), while the other one substitutes the latter for palmitic acid (16:0). Other characteristic ER lipids, included in the models, are PIs and PEs. PI lipids were not yet in the Membrane Builder platform library at the moment of building the first models. Place-holder lipids with corresponding FA were used to build initial setups on CHARMM-GUI, and then manipulated to add the missing coordinates on CHARMM according to the C36 FF parameters and topology for those lipids. The place-holder lipids were replaced by PI lipids (POPI and PYPI) from a PI lipid library that contained coordinates for different orientations of a given lipid. Water molecules overlapping with the inositol ring from the PIs were deleted, and neutralizing ions added to render a neutral simulation box. The TGN model was based on the comprehensive study from Klemm et al. in 2009.8 From the FA analysis, it was determined oleic (18:1) and palmitoleic (16:1) acids are the by far


Page 9 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

the most representative FA in the TGN. The PI lipids were the only ones with a different distribution, where palmitic acid (16:0) was the dominant FA. Based on these findings and specific percentages for lipid headgroups, the PC and PI lipids were selected as the dominant phospholipids for the membrane models along with a moderate concentration of ERG. The highest ERG concentration was in the PM models (40%), along with substantially more palmitic (16:0) and oleic (18:1) FA, which results in a low ratio of unsaturated to saturated fatty acids compared to other organelles.18 PS lipids are also abundant in the PM, but may be triggers for apoptosis, i.e. programmed cell death, if not properly regulated.42 Corresponding to experimental values, ERG, POPE, POPI, and POPS lipids were included in larger amounts in the PM models built for this study. 2.2 Membrane Properties The equilibrated trajectories of each system were analyzed to examine physical and structural properties of each model and compared them to each other as well as to information based on protein crystal structures in membranes and experimental trends. We computed the surface area per lipid (SA), compressibility modulus (KA), electron density profiles (EDPs), deuterium order parameters (SCD), wobble (lipid rotation) relaxation times, and sterol tilt for all the lipids in each model. Details on these calculations are provided below. Final coordinates of all models are freely available for download.43 Basic Lateral Membrane Properties Stability of simulation was determined from the overall SA for each run, a common metric that shows lipid-lipid interactions and lipid packing have reached equilibrium. The SA was defined on the x-y plane of the simulation box and divided by the number of lipids per leaflet. Using the SA data the KA is calculated directly using equation 1,


Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

= ∗

Page 10 of 40

(1)

where kB is the Boltzmann constant, AL is the SA, and σA2 its variance. KA is a measure of a membrane stiffness, or resistance to uniform compression. The SA for each individual component was calculated using a triangulation technique on the software Qhull,44 which uses the coordinates of three atoms in the phospholipid tails or one atom in sterol lipids as Vonoroi vertices to generate a polygon to estimate the volume and surface area for a particular lipid in a given plane. Electron density profiles The structure of biological lipid membranes is difficult to determine precisely due to thermal fluctuations. Bilayer structures have been studied using statistical averages based on neutron and X-ray scattering experiments such as small angle neutron scattering (SANS) and small angle X-ray scattering (SAXS).45, 46 These scattering density profiles measure the electron density of lipids in a bilayer, which can also be extracted from MD simulation data. EDPs from simulation data were plot using SIMtoEXP software package and following the procedure of Kučerka et al.47 to determine the head-to-head distance between the leaflet lipids (DHH), the Luzzati thickness (overall bilayer thickness) of the bilayer (DB), and its hydrocarbon region or core (2DC). DB and 2DC were estimated from the half point of the overall volume probability distributions of water and FA tails (CH2 and CH3 groups) respectively. Figure S2 shows an example of these curves for the TGN2 system. The hydrophobic core thickness, 2DC, was used to compare the models with estimates taken from the Orientation for Proteins in Membranes (OPM) database that reports hydrophobic lengths of transmembrane proteins.48


Page 11 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Deuterium order parameters (SCD) NMR experiments have been largely used to characterize the structure and order of lipids in pure and mixed bilayers, including mixtures with sterols.49 SCDs are segmental parameters that measure the order inside a bilayer, and are calculated from simulation data according to equation (2),

S = cos β −

(2)

where β is the angle between each C-H vector (bond) in the lipid tail and the bilayer normal, see Figure 2 (left). Stereospecific nomenclature classifies the FA tails in lipids by their position with respect to the glycerol group. By convention, the FA in the sn-2 position is the one attached to the oxygen atom of the second carbon of the glycerol group; the other tail is named sn-1 chain. The hydrogens attached to the second carbon in the FA tails in the sn-2 position give two different experimental signals that are also computed from the simulation data.50-54 Sterol tilt To complete the structural analysis of each organelle-specific bilayer, we computed the sterol tilt angle for each model as described by Lim et al.55 The tilt angle was defined as the angle between the z-axis and the vector between C17, the carbon at the base of the sterol, and C3, the carbon attached to the hydroxyl group. Lipid Relaxation Times A metric on a bilayer’s lipid motions can be found from reorientational correlation functions that can be related to NMR relaxation parameters, if available.49 Axial relaxation time constants are of special interest in computational models to determine the environment inside a bilayer. These relaxation times are measured by the second rank reorientational correlation


Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 40

function C2(t), equation (3), for specific atoms in the simulation data, and could be validated against experimental data from NMR experiments if available,56 = 〈 !̂ 0 ∙ !̂ %〉

(3)

where P2 is the second Legendre polynomial, and !̂ is a particular heavy atom-hydrogen vector. The correlation function from simulation was computed using vectors between selected atoms in the molecule of interest to determine its axial relation time rates. The vectors in this study were selected to estimate the slow relaxation time (found to be classified as wobble vs. true rotational motion)57 between the first carbons of each FA tail (C22-C32), and the second carbon of the glycerol group with its corresponding hydrogen (C2-HS), refer to Figure 2 (right). Two and three exponential fits were obtained to determine time constants for each lipid type in the membranes; equation 4 shows the general form of the fitting function. = '( + ∑+2 '+ , -.⁄/0

(4)

The independent coefficient ao in equation 4, was estimated from the average value of the plateau reached by the correlation function over at least 100 ns (see Figure S3). The time constants in this equation represent the fast and slow relaxation times associated with the lipid’s fast isomerization and wobble respectively.53 3. Results The following results allowed for comparisons between organelle-specific models and the general AVG-Yeast model properties (a total of 4.2 µs of simulation data). In addition, we evaluated each calculated property between the two models for each organelle to determine the influence of lipid chain length and unsaturation degree on the mechanical behavior of each bilayer. All properties were computed based on the last 150 ns of equilibrated trajectories for each replica, and the results are independent of initial lipid configuration on the x-y plane. The 12 ACS Paragon Plus Environment

Page 13 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

ER and TGN membranes reached equilibrium after 25 ns, shown by plateau of the average surface area per lipid (SA) on Figure S4. The plateau is especially noticeable on the PM membranes and AVG-yeast, which took longer simulation time to equilibrate – nearly 40 ns, and more than 50 ns for PM2 – due to their high sterol content. All property values were averaged over three replicate runs for each model. Surface Area, Compressibility Modulus and Sterol Tilt Table 1 summarizes the results for the overall SA, KA, and sterol tilt for each membrane. In addition, Table S3 lists the component areas for each lipid in the system. It is easy to see from these values, as well as Figure S4, that the PM and AVG-yeast models have a lower surface area (46.8-47.4 Å2), but higher KA values when compared to the other membranes. This occurs due to the presence of ERG, or cholesterol (CHOL) in the case of AVG-yeast. Sterols position themselves in between other lipid tails, creating a more ordered environment and a bulkier more rigid bilayer. It is interesting to note, however, that PM1, PM2, and AVG-yeast membranes have very different KA values despite their similar SA. To this fact, the unsaturation fraction, 0.70, 0.64, and 0.66 respectively (listed in Table S2) and sterol type may be a cause for this change; this will be discussed in more detail in the next section. Figure 3 shows the sterol tilt angle probability distributions for all the systems in this study, and the last column in Table 1 summarizes the average tilt angles for each bilayer. There is no statistical difference in the average tilt angle of the sterol molecules in the ER and TGN systems. In both cases the sterol molecules are about 27º inclined with respect to the bilayer normal (z-axis). In a more saturated environment such as the AVG-yeast or PM models, the tilt angles are 20.0±0.8º and 17.5±0.7º respectively. Electron density profiles


Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 40

Figure 4 shows the overall EDPs for each system, calculated by adding the total EDPs for each lipid in the systems; DHH was estimated from the peak values of each curve in this figure. All these values (bilayer thicknesses) are reported in Table 2, detailed EDPs showing the specific location of phosphate, glycerol, and lipid headgroups as well as the cumulative FA tails are found in Figure S5. The DHH for the ER and TGN membranes are very similar, between 37.4 and 38.6 Å. Moreover, the difference in unsaturation fraction between ER1 (0.86) and ER2 (0.69), or between TGN1 (0.74) and TGN2 (0.79), does not play a relevant role in the head-to-head distance and bilayer thickness of these membranes. This result was also observed in the SA values for these models, and especially their KA had no statistical difference (refer to Table 1) possibly due to the unsaturation degree of TGN (0.74 and 0.79 for models 1 and 2 of this membrane) falling in between that of the ER models (0.69 and 0.86). Deuterium order parameters (SCD) The SCD parameters for our models followed expected trends, decreasing in the presence of double bonds, and increasing with increasing concentration of sterol. Figure 5 illustrates more clearly the increase in order when sterol is present in higher concentrations, and Figure 6 shows two FA tails from different lipids in different bilayer environments. At first glance the SCD values for the PM membranes, which contain the highest ERG concentration (40%), are higher than those of the ER or TGN models. This happens due to the presence of ERG in the leaflets that affects the lateral organization of lipids resulting in a more symmetric and thicker bilayer.58 Figures S6 to S9 show the SCD values for each lipid in all the models. Overall, the PS and PI lipids tend to have higher overall order compared to the other lipids across all organelles. The difference between lipids varies no more than 0.05 and these differences occur at the end of tails


Page 15 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

C11-C15 for the ER and TGN membranes and also for the upper carbons (C3-C6) of the PM models. Lipid Relaxation Times Table 3 lists fits to a metric for axial reorientational motion, i.e. the cross-chain (CC) vector. Provided in this table is the average time constant τ3 for the CC of each model from three-exponential fitting to the correlation function C2(t) along with the first and second constants for the same fitting. The values from the corresponding two-exponential fitting are listed in Table S4 as well as the time constants to the two and three exponential fittings to the C2(t) for the second carbon of glycerol (CG). Previously, it was determined via motional models that motions on the 10ns timescale for DPPC lipids was classified as lipid wobble and full rotational motion is an order of magnitude longer.57 The relaxation times for the CC vector would best match this axial motion but CG motions will also result from this slow motion. As expected, the PM membranes have longer relaxation times compared to other membranes, while the ER membranes have the fastest times. This occurs due to the ordering effect of ERG molecules on each membrane. The slow relaxation times for the ER are slower than previous values reported for DPPC,57 which is likely the result of more acyl chain tangling with chain unsaturation. Note also that there is no statistical difference between the lipid wobble of the two models for each organelle. Overall, the CC and CG vectors results in similar fits with slightly slower times for the CC vectors. The slowest relaxation, which is likely mostly representing lipid wobble motion, increases from 9 ns for the ER to 12 ns for the TGN and 16.5 ns for the PM. In addition, 2D radial distribution functions (RDFs) of the phosphate atoms were generated for all possible lipid-lipid interactions in the organelle models, 15 for each model. As expected, lipid-lipid interactions also depend on membrane structure and composition; Figure 7


Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 40

shows the RDFs for the most relevant lipid-lipid interaction in all six organelle models, PE-PS. Figure S10 shows RDFs for other important lipid-lipid interactions in all the systems, PE-PE, PS-PS, PA-PE, and PC-PE. All our RDFs show two major peaks, which correspond to two interaction conformations; one where the oxygens in the phosphate group point towards each other (4.9-6.6 Å) and one where they point way from each other (6.7-8.9 Å) (see Figures 7 and S10). In turn, these conformations allow for the formation of one or two hydrogen bonds between the lipid headgroups depending upon steric hindrance introduced by the size of a particular head group. The bottom snapshots in Figures 7 and S10 illustrate this for lipid-lipid interactions of the ER2 model. Blocked 2D RDFs were also computed for all systems at 20ns intervals to show convergence during the simulation time scale. Figure S11 shows the blocked 2D RDFs of the three replicates for the PC and PI lipids of TGN1 model. These plots show convergence within the first 100-140ns of simulation in all systems, and also consistency between replicates even when the simulation started from different lipid distributions in the xyplane as in the case of both TGN models. Our systems converge much quicker than those presented by Hong et al,59 but are consistent with the trend that higher sterol content slows down the convergence. 4. Discussion Our membrane models are in agreement with data trends available in the OPM database that contains estimates for the hydrophobic thicknesses of transmembrane proteins based on their crystal structures.48 The measurements in the database can be directly compared to our estimates of 2DC. Values for the ER hydrophobic core in the database are more reliable because the hydrophobic length of the 4cad protein was estimated based on 8 helical subunits,60 while the PM transmembrane protein 2k9p had only 2 helical subunits.61 Despite limited experimental


Page 17 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

data, the ER estimates statistically agree with the experimental value of 30.8±1.3 Å, and the PM estimates are a slight overestimate but fall within 5% error with respect to experimental values. Although there were no experimental measurements for direct comparison of the TGN and AVG-yeast models, they follow expected trends based on sterol content. The relative composition of DOPC-DPPC-CHOL from AVG-yeast in a ternary phase diagram lies in the liquid-disordered (Ld) region at 303.15 K,62 but the other lipids in the model introduce interesting dynamics due to having a monounsaturated tail. The use of CHOL vs. ERG also contributed to drive the overall mixture in AVG-yeast closer to the liquid-ordered (Lo) region, resulting in a more ordered bilayer with similar behavior to the PM models in this aspect.63 The environment in a membrane is determined by its lipid components, unsaturation degree, and sterol fraction. The SA, commonly used to determine simulation equilibrium, also provides information about the lipid packing in a bilayer. High SA values as those of the ER membranes, 64.0±0.4 and 63.7±0.4 Å2, suggest the lipids occupy more space compared to those in the PM or AVG-yeast membranes. Sterols introduce more order in a lipid bilayer making it thicker and more rigid (lower SA and higher KA). The EDPs for the PM and AVG-yeast show the increase in bilayer thickness, with a DHH 5-6 Å larger than the ER and TGN models. As a result, lipid packing is more condensed, sterol tilt angle with respect to the bilayer is smaller (molecules are more aligned with the bilayer normal), SCD parameters are higher, and lipid relaxation times are longer. FA composition between models for the same organelle has little effect on membrane mechanical properties, but some differences can be observed. For example, PM1 and PM2 models have the same headgroup diversity, but the degree of unsaturation for these models is 0.71 and 0.64, respectively. In other words, there are more unsaturated FA tails in the lipids of


Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 40

PM1, which introduce disorder in the bilayer as their conformation has kinks at the location of double bonds. Since PM2 has a more saturated environment than PM1, one would expect its KA to be also higher than that of PM1; however, the results shows the opposite trend suggesting the combination between FA unsaturation and lipid diversity plays a role in the mechanical properties of a bilayer. In the PM1 model the palmitoleic (C16:1) and oleic (C18:1) FAs (YOlipids) had the longest relaxation times for the CC correlation, and the PI headgroup had the longest relaxation time for the CG correlation. On the other hand, the PM2 lipids had all similar relaxation times for both correlation functions. No significant statistical difference was observed between the relaxation times of the two models for each organelle, and these values increase with the increase of sterol content in the bilayer as expected. Our relaxation times are in between those reported by Venable et al;64 they are 3-10 times larger than those of pure DPPC bilayers because the present membranes have a sterol component, but are approximately five times smaller than the values of the sphingolipid system since our systems do not have strong interlipid hydrogen bonding near the hydrophobic interface as the sphingolipids. The cell environment is very dynamic, and the instantaneous composition of a given bilayer will vary. However, a healthy cell maintains an equilibrium composition of sterol and lipids that enables its proper function and growth. The type of sterol in a membrane also reproduced different properties in a bilayer. AVGyeast, with lower KA than the PM models (0.34±0.04 N/m), has a similar unsaturated fraction to PM2 (listed in Table S2) but its sterol composition is 22.2% – a little more than half the percentage of sterol in the PM membranes (40%). Studies show CHOL induces more order and reduces fluidity in a bilayer compared to ERG;63 it is, after all, more saturated than ERG. Despite this, the lower value in KA is expected in the AVG-yeast model because the sterol content is not


Page 19 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

as high as that of the PM membranes, thus the bilayer is more fluid than the PMs. This is also observed in the order parameters for these three models, the highest value for the sn-2 chains in the PM models is 0.30, while the corresponding values for the mono- or di-unsaturated lipids in the AVG-yeast model are not higher than 0.27, see Figures S9 and S10. Although subtle, the ordering effect of the high sterol concentration in the PMs vs. the AVG-yeast model is noticeable. A more robust comparison on the effects of CHOL versus ERG could be made if the membranes had the same sterol composition, but that was not in the scope of this work. AVG-yeast is a good model to understand the ordered structure and thickness of the PM of yeast; however, it better describes internal dynamics of the TGN or ER membranes as it was shown through the KA values and lipid relaxation times, Tables 1 and 3 respectively. This occurs in part because of its lower sterol content compared to that of the PM, and also because AVGyeast is a simpler model than PM as far as lipid diversity (6 vs. 11 lipid types, respectively). Bilayer structure as described by the order parameters was driven by sterol content and unsaturation degree. The order parameters for the ER and TGN were at least 20% lower with respect to the PM and AVG-yeast models. The unsaturation fraction between models for the same organelle was irrelevant for this property, but decreased in opposite direction to the order parameters among all the models; the ER models had 0.86 and 0.69 unsaturation degree, the TGN 0.74 and 0.79, the PM 0.70 and 0.64, and the AVG-yeast 0.65. With higher concentration of unsaturated FAs, the bilayer core was more disordered because of the kinks double bonds introduce in the lipid tail conformations. AVG-yeast could be used to understand general trends that directly correspond to the sterol content in a bilayer, but was not detailed enough to understand internal dynamics of the membrane and cannot give us specific detail of membraneprotein interactions.


Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 40

Our models provide a better representation of the PM, TGN, and ER membranes in yeast, but lipid composition has also a particular signature in the inner and outer leaflet of some bilayers. A symmetric bilayer is an accurate representation of the ER membrane; however, the TGN and PM have asymmetric leaflets that were not modeled in this study.5 Sphingolipids and ceramides play an important role in the PM, especially in protein-lipid interactions and cell signaling processes.1, 20, 45 However, the current PM model lacks fully saturated lipids and longchained inositol phosphoceramide (IPC). Similarly, the TGN did not contain long-chained IPCs. These lipid types introduce leaflet interdigitation and promote coexistence of lipid domains.5, 65 Important membrane dynamics arise from these lipids and high sterol concentrations, forming Lo sterol-ceramide-rich domains and Ld sterol-ceramide-poor phases.5, 10, 66, 67 Asymmetric bilayers were not built for the present study because the ceramides and sphingolipid parameters were under development and were only recently published.64 The short simulation timescale in this study (200 ns) and the lack of fully saturated lipids in the symmetric leaflets did not allow us to see distinct phase separation in our models. More simulations will be run to study the effect of ceramides in domain formation as well as asymmetry in both the PM and TGN models. 5. Conclusions The yeast organelle models provided insight into the effects of lipid saturation and diversity on the structural, mechanical and dynamical properties of a bilayer. Two membranes per organelle with different lipid unsaturation degrees, but identical headgroup distribution, were built to emulate the environment of the ER, TGN, and PM bilayers. Each membrane composition mimics available experimental data that characterizes the lipid content. Data analysis for each organelle model was evaluated against each other, and the hydrophobic bilayer thicknesses 2DC as estimated from the EDP were compared against experimental data available at the OPM


Page 21 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

database.48 A reference membrane model was constructed based on the composition of Jo et al. 15

to compare its bilayer properties to those calculated for organelle-specific models. As the

results in this study confirm, the lipid composition of a membrane alters its environment. One can decide to use simplified models to understand general trends in the structure, mechanical, and dynamical behavior of a membrane, but tailored models are needed to study detailed processes in specific compartments in the cell, especially those that involve transmembrane proteins. Our goal with this study was threefold; to study the effects of lipid composition in membrane structure, to develop accurate models for the study of lipid-protein interactions in the future, and to motivate experimental groups to continue lipidomics studies given the need for more accurate molecular level studies. Our current models are an improvement in the field, these are the first all-atom simulations to focus on organelle-specific lipid diversity that include representative lipid headgroups with a variety of fatty acid tails. However, in order to understand more complex events like the dynamics of lipid domain formation better models are needed to describe the bilayer environment. We hope experimental studies become available to build more detailed models for a variety of organelles, especially given the significance of membranemembrane and membrane-protein interactions for cell processes. The formation of sterol-rich domains in bilayers containing fully saturated lipids has gained attention in the past twenty years, and has been more recently associated with membrane curvature and signaling processes.68, 69 A future study will examine extended simulations of the PM models (microsecond trajectories) to determine the time scale and dynamics of sterol-rich domains and sterol flip-flop in a bilayer. In addition, new models will be built to reflect the asymmetry in the PM and TGN bilayers and include IPC lipids, whose parameters have been recently made available.64 The current, and


Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 40

future, models allow us to study in more detail interactions of the bilayer with peripheral proteins, which sometimes interact with a bilayer only when a lipid gradient is present. Acknowledgments We appreciate the work of the Dr. Wonpil Im lab at the University of Kansas for their collaboration in the development of PI lipid parameters, updates to the CHARMM-GUI Membrane Builder, and general discussions. Funding This study was funded in part by NSF grant DBI-1145652, XSEDE computational resources MCB-100139, and the High Performance Computing Clusters at the University of Maryland, College Park (Deepthought1 and Deepthought2) supported by Division of Information Technology. Viviana Monje-Galvan received partial funding by the National Science Foundation Louis Stokes Alliances for Minority Participation (LSAMP) Bridge to the Doctorate Fellowship to UMBC under subaward 0000012127 with the University of Maryland College Park. Associated Content Included in the supporting material are three tables that describe the composition of lipids in the membrane, individual lipid component surface areas, and fits to reorientatial correlations functions of the CC and CG vectors. The following figures are also included: chemical structures of lipids, example calculation of the bilayer thicknesses, correlation functions, time series of averages surface area for each model and respective replicas, component density profiles for each model, and SCD for all membrane models..


Page 23 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

References 1. 2. 3. 4. 5. 6. 7. 8.

9. 10. 11. 12. 13. 14. 15. 16.

17.

Klug, L., and Daum, G. (2014) Yeast lipid metabolism at a glance, FEMS Yeast Research 14, 369-388. Santos, A. X. S., and Riezman, H. (2012) Yeast as a model system for studying lipid homeostasis and function, FEBS Letters., 586,2858-2867. Engelman, D. M. (2005) Membranes are more mosaic than fluid, Nature 438, 578-580. Nelson, D. L., and Cox, M. M. Principles of Biochemistry, 5th ed., W.H. Freeman and Company, New York. van Meer, G., Voelker, D. R., and Feigenson, G. W. (2008) Membrane lipids: where they are and how they behave, Nature Reviews Molecular Cell Biology 9, 112-124. Maxfield, F. R. T., Ira. (2005) Role of cholesterol and lipid organization in disease, Nature 438, 612-621. Bankaitis, V. A., Garcia-Mata, R., and Mousley, C. J. (2012) Golgi Membrane Dynamics and Lipid Metabolism, Current Biology 22, R414-424. Klemm, R. W., Ejsing, C. S., Surma, M. A., Kaiser, H.-J., Gerl, M. J., Sampaio, J. L., Robillard, Q. d., Ferguson, C., Proszynski, T. J., Shevchenko, A., and Simons, K. (2009) Segregation of sphingolipids and sterols during formation of secretory vesicles at the trans-Golgi network, The Journal of Cell Biology 185, 601-612. Leber, A., Hrastnick, C., and Daum, G. (1995) Phospholipid-synthesizing enzymes in Golgi membranes of the yeast, Saccharomyces cerevisiae, FEBS Letters 377, 271-274. Simons, K., and Ikonen, E. (2000) How Cells Handle Cholesterol, Science 290, 17211726. Pandit, K. R., and Klauda, J. B. (2012) Membrane models of E. coli containing cyclic moieties in the aliphatic lipid chain, Biochim. Biophys. Acta, Biomembr. 1818, 12051210. Lim, J. B., and Klauda, J. B. (2011) Branching at the Iso- and Anteiso- Positions in Complex Chlamydia Membranes: A Molecular Dynamics Study, Biochim. Biophys. Acta, Biomembr. 1808, 323-331 Wu, Emilia L., Engström, O., Jo, S., Stuhlsatz, D., Yeom, Min S., Klauda, Jeffery B., Widmalm, G., and Im, W. (2013) Molecular Dynamics and NMR Spectroscopy Studies of E. coli Lipopolysaccharide Structure and Dynamics, Biophys. J. 105, 1444-1455. Wu, Emilia L., Fleming, Patrick J., Yeom, Min S., Widmalm, G., Klauda, Jeffery B., Fleming, Karen G., and Im, W. (2014) E. coli Outer Membrane and Interactions with OmpLA, Biophys. J. 106, 2493-2502. Jo, S., Lim, J. B., Klauda, J. B., and Im, W. (2009) CHARMM-GUI Membrane Builder for mixed bilayers and its application to yeast membranes, Biophys J 97, 50-58. Ingólfsson, H. I., Melo, M. N., van Eerden, F. J., Arnarez, C., Lopez, C. A., Wassenaar, T. A., Periole, X., de Vries, A. H., Tieleman, D. P., and Marrink, S. J. (2014) Lipid Organization of the Plasma Membrane, Journal of the American Chemical Society 136, 14554-14559. Koldsø, H., Shorthouse, D., Hélie, J., and Sansom, M. S. P. (2015) Lipid Clustering Correlates with Membrane Curvature as Revealed by Molecular Simulations of Complex Lipid Bilayers, PLOS Computational Biology 10,e1003911.


Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

18. 19.

20.

21. 22. 23. 24.

25. 26.

27. 28. 29.

30.

Page 24 of 40

Tuller, G., Nemec, T., Hrastnik, C., and Daum, G. (1999) Lipid composition of subcellular membranes of an FY1679‐derived haploid yeast wild‐type strain grown on different carbon sources, Yeast 15, 1555-1564. Daum, G., Tuller, G., Nemec, T., Hrastnik, C., Balliano, G., Cattel, L., Milla, P., Rocco, F., Conzelmann, A., Vionnet, C., Kelly, D. E., Kelly, S., Schweizer, E., Schüller, J., Hojad, U., Greiner, E., and Finger, K. (1999) Systematic analysis of yeast strains with possible defects in lipid metabolism, Yeast 15, 601-614. Schneiter, R., Brügger, B., Sandhoff, R., Zellnig, G., Leber, A., Lampl, M., Athenstaedt, K., Hrastnik, C., Eder, S., Daum, G., Paltauf, F., Wieland, F. T., and Kohlwein, S. D. (1999) Electrospray Ionization Tandem Mass Spectrometry (Esi-Ms/Ms) Analysis of the Lipid Molecular Species Composition of Yeast Subcellular Membranes Reveals Acyl Chain-Based Sorting/Remodeling of Distinct Molecular Species En Route to the Plasma Membrane, The Journal of Cell Biology 146, 741-754. Jo, S., Kim, T., and Im, W. (2007) Automated builder and database of Protein/Membrane complexes for molecular dynamics simulations, PLoS One 2, e880. Jo, S., Kim, T., Iyer, V. G., and Im, W. (2008) CHARMM-GUI: A web-based graphical user interface for CHARMM, J. Comput. Chem. 29, 1859-1865. Jo, S., Lim, J. B., Klauda, J. B., and Im, W. (2009) CHARMM-GUI Membrane Builder for Mixed Bilayers and Its Application to Yeast Membranes, Biophys. J. 97, 50-58. Wu, E. L., Cheng, X., Jo, S., Rui, H., Song, K. C., Lee, J., Davila-Contreras, E. M., Beaven, A. H., Monje-Galvan, V., Venable, R. M., Klauda, J. B., and Im, W. (2014) CHARMM-GUI Membrane Builder Toward Realistic Biological Membrane Simulations, J. Comput. Chem. 35, 1997-2004. Jorgensen, W. L., Chandrasekhar, J., Madura, J. D., Impey, R. W., and Klein, M. L. (1983) Comparison of simple potential functions for simulating liquid water, The Journal of Chemical Physics 79, 926. Brooks, B. R., Brooks, C. L., Mackerell, A. D., Nilsson, L., Petrella, R. J., Roux, B., Won, Y., Archontis, G., Bartels, C., Boresch, S., Caflisch, A., Caves, L., Cui, Q., Dinner, A. R., Feig, M., Fischer, S., Gao, J., Hodoscek, M., Im, W., Kuczera, K., Lazaridis, T., Ma, J., Ovchinnikov, V., Paci, E., Pastor, R. W., Post, C. B., Pu, J. Z., Schaefer, M., Tidor, B., Venable, R. M., Woodcock, H. L., Wu, X., Yang, W., York, D. M., and Karplus, M. (2009) CHARMM: The biomolecular simulation program, J. Comput. Chem. 30, 1545-1614. Jo, S., Kim, T., and Im, W. (2007) Automated Builder and Database of Protein/Membrane Complexes for Molecular Dynamics Simulations, PLOS one, e880. Phillips, J. C., Braun, R., Wang, W., Gumbart, J., Tajkhorshid, E., Villa, E., Chipot, C., Skeel, R. D., Kale, L., and Schulten, K. (2005) Scalable molecular dynamics with NAMD, J. Comput. Chem. 26, 1781-1802. Klauda, J. B., Venable, R. M., Freites, J. A., O'Connor, J. W., Mondragon-Ramirez, C., Vorobyov, I., Tobias, D. J., MacKerell, A. D., and Pastor, R. W. (2010) Update of the CHARMM all-atom additive force field for lipids: Validation on six lipid types., J. Phys. Chem. B 114, 7830-7843. Li, Z., Venable, R. M., Rogers, L. A., Murray, D., and Pastor, R. W. (2009) Molecular dynamics simulations of PIP2 and PIP3 in lipid bilayers: determination of ring orientation, and the effects of surface roughness on a Poisson-Boltzmann description, Biophys J 97, 155-163. 24 ACS Paragon Plus Environment

Page 25 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.

43. 44. 45. 46. 47. 48. 49.

Lim, J. B., Rogaski, B., and Klauda, J. B. (2012) Update of the Cholesterol Force Field Parameters in CHARMM, J. Phys. Chem. B 116, 203-210. Ryckaert, J.-P., Ciccotti, G., and Berendsen, H. J. C. (1977) Numerical integration of the cartesian equations of motion of a system with constraints: molecular dynamics of nalkanes, Journal of Computational Physics 23, 327–341. Hoover, W. G. (1985) Canonical dynamics: Equilibrium phase-space distributions, Physical Review A 31, 1695-1697. Frenkel, D., and Smit, B. (2001) Understanding Molecular Simulation From Algorithms to Applications, Vol. 1, 2 ed., Academic Press. Leach, A. R. (2001) Molecular Modeling Principles and Applications, 2 ed., Pearson Education, Great Britain. Nosé, S., and Klein, M. L. (1983) A study of solid and liquid carbon tetrafluoride using the constant pressure molecular dynamics technique, The Journal of Chemical Physics 78, 6928. Feller, S. E., Zhang, Y., Pastor, R. W., and Brooks, B. R. (1995) Constant pressure molecular dynamics simulation: The Langevin piston method, The Journal of Chemical Physics 103, 4613-4621. Phillips, J. C., Braun, R., Wang, W., Gumbart, J., Tajkhorshid, E., Villa, E., Chipot, C., Skeel, R. D., Kale, L., and Schulten, K. (2005) Scalable molecular dynamics with NAMD, Journal of Computational Chemistry 26, 1781-1802. Martyna, G. J., Tobias, D. J., and Klein, M. L. (1994) Constant pressure molecular dynamics algorithms, The Journal of Chemical Physics 101, 4177-4189. Steinbach, P. J., and Brooks, B. R. (1994) New spherical‐cutoff methods for long‐range forces in macromolecular simulation, Journal of Computational Chemistry 15, 667-683. Darden, T., York, D., and Pedersen, L. (1993) Particle mesh Ewald: An N⋅log(N) method for Ewald sums in large systems, The Journal of Chemical Physics 98, 10089. Zinser, E., Sperka-Gottlieb, C. D., Fasch, E. V., Kohlwein, S. D., Paltauf, F., and Daum, G. (1991) Phospholipid synthesis and lipid composition of subcellular membranes in the unicellular eukaryote Saccharomyces cerevisiae, Journal of Bacteriology 173, 20262034. Klauda, J. B. (2015) Free Programs, Molecular Coordinates, & Files, University of Maryland. Barber, C. B., Dobkin, D. P., and Huhdanpaa, H. (1996) The quickhull algorithm for convex hulls, ACM Transactions on Mathematical Software (TOMS) 22, 469-483. Kučerka, N., Nagle, J. F., Sachs, J. N., Feller, S. E., Pencer, J., Jackson, A., and Katsaras, J. (2008) Lipid bilayer structure determined by the simultaneous analysis of neutron and X-ray scattering data, Biophys J 95, 2356-2367. Pan, J., Marquardt, D., Heberle, F. A., Kucerka, N., and Katsaras, J. (2014) Revisiting the bilayer structures of fluid phase phosphatidylglycerol lipids: Accounting for exchangeable hydrogens, Biochim. Biophys. Acta 1838, 2966-2969. Kučerka, N., Katsaras, j., and Nagle, J. F. (2010) Comparing Membrane Simulations to Scattering Experiments: Introducing, Journal of Membrane Biology 235, 43-50. Lomize, M. A., Lomize, A. L., Pogozheva, I. D., and Mosberg, H. I. (2006) OPM: orientations of proteins in membranes database, Bioinformatics 22, 623-625. Leftin, A., and Brown, M. F. (2011) An NMR database for simulations of membrane dynamics, Biochim Biophys Acta 1808, 818-839. 25 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61.

62. 63. 64.

65.

Page 26 of 40

Seelig, A., and Seelig, J. (1974) Dynamic structure of fatty acyl chains in a phospholipid bilayer measured by deuterium magnetic resonance, Biochemistry 13, 4839-4845. Seelig, A., and Seelig, J. (1975) Bilayers of dipalmitoyl-3-sn-phosphatidylcholine. Conformational differences between the fatty acyl chains, Biochim Biophys Acta 406, 15. Seelig, J., and Waespe-Sarcevic, N. (1978) Molecular order in cis and trans unsaturated phospholipid bilayers, Biochemistry 17, 3310-3315. Klauda, J. B., Eldho, N. V., Gawrisch, K., Brooks, B. R., and Pastor, R. W. (2008) Collective and noncollective models of NMR relaxation in lipid vesicles and multilayers, J Phys Chem B 112, 5924-5929. Perly, B., Smith, I. C., and Jarrell, H. C. (1985) Acyl chain dynamics of phosphatidylethanolamines containing oleic acid and dihydrosterculic acid: 2H NMR relaxation studies, Biochemistry 24, 4659-4665. Lim, J. B., Rogaski, B., and Klauda, J. B. (2011) Update of the Cholesterol Force Field Parameters in CHARMM, The Journal of Physical Chemistry B 116, 203-210. Klauda, J. B., Roberts, M. F., Redfield, A. G., Brooks, B. R., and Pastor, R. W. (2008) Rotation of Lipids in Membranes: Molecular Dynamics Simulation, 31P Spin-Lattice Relaxation, and Rigid-Body Dynamics, Biophysical Journal 94, 3074-3083. Klauda, J. B., Pastor, R. W., Brooks, B. R., Roberts, M. F., and Redfield., A. G. (2008) Rotation of Lipids in Membranes: MD Simulation, 31P Spin-Lattice Relaxation, and Rigid-Body Dynamics, Biophys. J. 94, 3074–3083. Simons, K., and Ikonen, E. (2000) How Cells Handle Cholesterol, Science 290, 17211726. Hong, C., Tieleman, D. P., and Wang, Y. (2014) Microsecond Molecular Dynamics Simulations of Lipid Mixing, Langmuir 30, 11993-12001. Neumoin, A., Cohen, L. S., Arshava, B., Tantry, S., Becker, J. M., Zerbe, O., and Naider, F. (2009) Structure of a double transmembrane fragment of a G-protein-coupled receptor in micelles, Biophys J 96, 3187-3196. Manolaridis, I., Kulkarni, K., Dodd, R. B., Ogasawara, S., Zhang, Z., Bineva, G., O’Reilly, N., Hanrahan, S. J., Thompson, A. J., Cronin, N., Iwata, S., and Barford, D. (2013) Mechanism of farnesylated CAAX protein processing by the intramembrane protease Rce1, Nature 504, 301-305. Davis, J. H., Clair, J. J., and Juhasz, J. (2009) Phase equilibria in DOPC/DPPCd62/cholesterol mixtures, Biophys J 96, 521-539. Arora, A., Raghuraman, H., and Chattopadhyay, A. (2004) Influence of cholesterol and ergosterol on membrane dynamics: a fluorescence approach ☆, Biochimica et Biophysica Acta (BBA) Reseach Communications 318, 920–926. Venable, R. M., Sodt, A. J., Rogaski, B., Rui, H., Hatcher, E., MacKerell, A. D., Pastor, R. W., and Klauda, J. B. (2014) CHARMM All-Atom Additive Force Field for Sphingomyelin: Elucidation of Hydrogen Bonding and of Positive Curvature, Biophysical Journal 107, 134-145. Holthuis, J. C. M., Pomorski, T., Raggers, R. J., Sprong, H., and Van Meer, G. (2001) The Organizing Potential of Sphingolipids in Intracellular Membrane Transport, Physiological Reviews 81, 1689-1723.


Page 27 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

66. 67. 68.

69.

Vanegas, J. M., Contreras, M. F., Faller, R., and Longo, M. L. (2012) Role of unsaturated lipid and ergosterol in ethanol tolerance of model yeast biomembranes, Biophys J 102, 507-516. Collins, M. D. (2008) Interleaflet Coupling Mechanisms in Bilayers of Lipids and Cholesterol☆, Biophys J 94, L32-34. Bastos, A. E. P., Marinho, H. S., Cordeiro, A. M., de Soure, A. M., and de Almeida, R. F. M. (2012) Biophysical properties of ergosterol-enriched lipid rafts in yeast and tools for their study: characterization of ergosterol/phosphatidylcholine membranes with three fluorescent membrane probes, Chemistry and Physics of Lipids 165, 577-588. Sadeghi, S., Muller, M., and Vink, R. L. C. (2014) Raft Formation in Lipid Bilayers Coupled to Curvature, Biophysical Journal 107, 1591-1600.


Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 40

Table 1 - Average and standard error of the surface area per lipid, compressibility modulus, and sterol tilt angle (values calculated from the time series of equilibrated data). Sterol tilt (º) model SA/lipid (Å2) KA (N/m) 27.3 ± 0.9 er1 64.0 ± 0.4 0.29 ± 0.02 27.9 ± 0.8 er2 63.6 ± 0.4 0.28 ± 0.02 26.3 ± 0.8 tgn1 60.6 ± 0.4 0.28 ± 0.06 26.6 ± 0.9 tgn2 60.9 ± 0.4 0.27 ± 0.02 17.9 ± 0.8 pm1 47.4 ± 0.2 0.57 ± 0.06 17.2 ± 0.8 pm2 46.8 ± 0.3 0.47 ± 0.08 20.0 ± 0.8 avg-yeast 47.3 ± 0.3 0.34 ± 0.04


Page 29 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Table 2 – Bilayer thicknesses of yeast membranes: head-to-head distance (DHH ± 0.1), bilayer thickness (DB ± 0.2), and hydrophobic core (2DC ± 0.2, compared to crystal structure estimates in the Orientation of Protein in Membranes database, OPM 48) Model

DHH (Å)

DB (Å)

2DC (Å)

OPM (Å)

er1 er2 tgn1 tgn2 pm1 pm2 avg-yeast

37.8 37.4 38.6 38.4 43.4 44.4 43.0

34.4 34.1 36.4 35.8 39.3 39.6 39.9

29.1 29.3 26.5 29.7 34.4 34.9 34.0

30.8 ± 1.3 (4cad)* n/a 29.8 ± 3.1 (2k9p)** n/a

(*http://opm.phar.umich.edu/protein.php?pdbid=4cad; **http://opm.phar.umich.edu/protein.php?pdbid=2k9p)


Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 40

Table 3 – Average of 3-exponential fitting to the cross-chain (CC) correlation function. Time constants are reported in ns with their respective standard error. The 2-exponential fittings for the CC and CG as well as the 3-exponential fitting for the CG are reported in Table S4. Cross Chain –C2(t) model τ1 τ2 τ3 er1 0.126 ± 0.025 1.526 ± 0.252 8.515 ± 1.219 er2 0.153 ± 0.037 1.878 ± 0.474 9.701 ± 3.285 tgn1 0.168 ± 0.053 2.162 ± 0.662 12.048 ± 2.696 tgn2 0.174 ± 0.172 2.141 ± 1.079 11.504 ± 2.590 pm1 0.162 ± 0.129 2.207 ± 1.044 16.854 ± 5.390 pm2 0.160 ± 0.058 2.156 ± 0.950 16.271 ± 4.413 avg-yeast 0.165 ± 0.052 1.814 ± 0.604 11.716 ± 2.543


Page 31 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Figure 1 – Model lipid composition (by headgroup); “*-exp” refer to available experimental data. The sterol used in Avg-yeast was cholesterol, but its composition is shown under ERG in this graph Figure 2 – (A) Deuterium order parameters (SCD) diagram. Each parameter was computed using equation 2; β is the angle between the C-H bond and the bilayer normal computed for each carbon position in the FA tails. (B) Lipid wobble diagram showing the bond vectors used to compute the glycerol carbon (atoms shown in yellow) and the cross-chain (atoms shown in purple) correlation functions for all the lipids in each model. Colored atoms represent carbon (cyan), nitrogen (blue), oxygen (red), and phosphorus (gold). Figure 3 – Average sterol tilt angles for all membrane models (angle between the bilayer normal and the sterol vector defined from the carbon at the base of the molecule to the carbon with the hydroxyl group, C17-C3) Figure 4 - Overall electron density profiles for each model, the major peaks were used to estimate the head-to-head bilayer thickness. Figure 5 – SCD parameters for the sn1 chains of the ER2 and PM2 model membranes. Notice the same lipids have higher order parameters in the PM2 model that has higher sterol content than the ER2; even unsaturated carbons in the chain (9 and 10) have increased order parameters when more sterol is present in the bilayer. Figure 6 – SCD parameters comparisson between an unsaturated (top) and saturated (botom) FA chain from two different lipid types across all models. Figure 7 – 2D RDFs for the PS-PE lipid interaction in all the organelle models; the bottom snapshot shows a YOPE (left) and a POPS (right) from the ER2 model interacting through two 31 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 40

hydrogen bonds between the anime group of each lipid and an oxygen atom in the oposite lipid. Colored atoms represent carbon (cyan), nitrogen (blue), oxygen (red), and phosphorus (gold).


Page 33 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Figure 1


Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 40

Figure 2


Page 35 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Figure 3


Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 40

Figure 4


Page 37 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Figure 5


Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 38 of 40

Figure 6


Page 39 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Figure 7


Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 40 of 40

TOC Graphic


Modeling Yeast Organelle Membranes and How Lipid Diversity

Recommend Documents