Computational protocol: Key Molecular Requirements for Raft Formation in Lipid/Cholesterol Membranes

Similar protocols

Protocol publication

[…] For all the systems the concentrations 0.34∶0.51∶0.15 were used for DPPC/DUPC/CHOL or for their respective substituent. The CHOL (or its substituent) mole fraction of 0.15 was chosen similar to the system described elsewhere . The MARTINI CG models of DPPC, DUPC and CHOL are presented in . The beads are colored according to their types. The bead type C1 presents a single-bonded carbon group while C4 bead presents the double-bond region in the chain structure of DUPC. The SCx beads (where x can be any number from 1 to 5) present a ring structure in the CHOL molecule. For all the systems the bilayers were kept perpendicular toward the z direction during the simulations. The x/y side lengths of all the CG bilayers were ∼20 nm. In average, the systems were composed of 510 DPPC, 810 DUPC lipids, and 238 cholesterols or of their respective substituent. GROMACS version 4.5.1 was used for all simulations , . The models of MARTINI potential 2.0 were used for lipids, cholesterol and standard water . The systems were simulated by using an NPT ensemble with a semiisotropic pressure tensor of 1 bar. The Berendsen coupling scheme for the temperature and pressure was used . Bond lengths were constrained by the linear constraint solver (LINCS) algorithm . The systems were simulated at 295 K. The DPPC/DOPC +15% CHOL mixture was experimentally shown to form a liquid-liquid phases at temperature below 303 K . On the other hand by interpolation of the experimental results the DPPC/DOPC +15% CHOL composition was suggested to separate into Lo/Ld phases at room temperature , . Taking into account that in our systems the DUPC is used instead of the more ordered DOPC the range of temperatures of forming liquid-liquid phases should be even lower than that of the DOPC. As a result the simulated DPPC/DUPC/CHOL system at 295 K is well in the temperature range of Lo/Ld separation which has already been reported elsewhere , . The initial random distribution of the original DPPC/DUPC/CHOL system was obtained by a 20 ns simulation at 450 K. This random configuration was used for all the other systems with modified DPPC or DUPC (by replacing the original component by its modified type). All the systems were simulated for 12 µs of effective times with a time step of 20 femtoseconds.The shifted electrostatic and van der Waals (vdW) potentials were applied with 1.2 nm cut-off distance. Although the truncation of the long-range electrostatics in the atomistic level as opposed to the Particle-Mesh Ewald (PME) or reaction field techniques has been shown to cause artifacts , the CG methods bring the possibility for certain cases to adequately account the effect of long-range interactions in the short-range potentials , , , . Despite the observation that the electrostatic interaction strength between the polar substances in the non-polarizable solvents is underestimated (due to the implicit screening with the MARTINI standard water model) , , nevertheless for simple bilayer system where no transition of a polar substance across an interface of different dielectric constants is studied the standard MARTINI water model demonstrates reasonable results . Moreover, for MARTINI bilayer systems with standard water particles it was shown that using of PME for the long-range electrostatic interactions as compared with the cut-off scheme does not affect the lipid bilayer structure and dynamics suggesting that the cut-off approach is a reasonable choice for the systems presented here.In particular, to check the relevance of DUPC lipid to raft formation we study the impact of substituting DUPC by somewhat related molecules. All the modifications of the DUPC lipid targeted one or more of the three FF properties which differentiate DPPC from DUPC. summarizes all the differences between DPPC and DUPC lipids. As shown in and listed in the second and third beads in each chain of the DUPC structure (bead numbers 6, 7, 10 and 11) are more apolar (type C4) than the other beads of the chains and reflect the double-bond regions. This difference in bead types affect the enthalpic i.e. the vdW interactions between the DPPC and DUPC lipids. The modifications of DUPC are characterized by systematic assimilations of the angles, angle force constants (stiffnesses) and beads of the DUPC chains listed in to the ones of the DPPC lipid. For the sake of clarity the DUPC modifications are named with respect to their differences from the DUPC. The lower letters ‘a’, ‘s’ and ‘b’ in the naming of the modified DUPC lipid stand for the angle, stiffness and bead, respectively. In the names of DUPC modifications which have bead type differences (i.e. DUb2, DUb3 and DUb23) the lower letter ‘b’ is followed by the number of the bead (or beads) in a chain which is different from the C4 type of the corresponding DUPC bead (or beads). shows the corresponding vdW interaction strengths between the combinations of C1 and C4 pairs.The influence of chain stiffness of DPPC is also investigated. We consider a modified DPPC lipid for which the stiffness of the chain is decreased by 60% (10 kJ/mol instead of the default 25 kJ/mol). It is denoted DPPCsoft. For the simulations with DPPCsoft the phase separated DPPC/DUPC/CHOL configuration was used as the initial structure of the new DPPCsoft/DUPC/CHOL system.It should be also noted that the interaction strength of SCx with type of Cy (where x and y may take any number from 1 to 5) is the same as the Cx-Cy interaction strength while the interaction of SCx-SCy is lower by a factor of 0.75 in respect to the strength of the corresponding Cx-Cy . This aspect is used when discussing one of the substituents of CHOL (i.e. DPPC3b). The latter (DPPC3b) is introduced as shortened DPPC-like lipid. It has only three chain beads instead of the standard four and, most importantly, the angle force constants between the chain beads are set to a high enough value (here: 300 kJ/mol) to keep the chains straight. All the three beads of the chains were set to type SC1 to suppress the aggregation of DPPC3b with itself as it is the case with CHOL. Additionally, an extra bond is added between the beads 7 and 11 to keep the chains together (see for the bead numbers).The order parameter between A and B beads of the lipid chains is calculated as where θz is the angle between the AB bond and the z direction. The order parameter of a lipid is the average over all the consecutive beads in the chain of a lipid. For CHOL the θz is the angle between the line which passes through the CHOL beads 3 and 5 () and the z direction.The domain correlation coefficients between bilayer leaflets are calculated by dividing the bilayer surface into square cells of side length 1.5 nm and calculating the correlation coefficients for the densities of the desired pairs in the corresponding cells of two leaflets.The z coordinates of the head and tail beads of lipids and CHOL and their substituent relative to the bilayer normal were calculated by dividing the bilayer surface to square cells of side length 3.0 nm and calculating the bilayer center for each of these rectangular boxes by taking into account the z coordinates of only the tail beads of both lipids and CHOL and their “analogs” except for the DPPC3b. The latter is omitted from the calculation of the bilayer center since its shorted tail would result in shifting of the center position toward the higher concentration of DPPC3b within each rectangular box.The relevance of the Umbrella model for raft formation was elucidated by recording the CHOL molecules for which the head beads (bead 1 in ) were covered by the head beads of its nearest-neighbor lipids. To consider a lipid (or two lipids) covering the head of a CHOL molecule the head of the lipid had to be within the area of 0.25 nm side square above the CHOL head (or in case of two lipids, their heads had to be located within the area of 0.47 nm side square above the CHOL head).The VMD package was used for CHOL and lipids presentation. […]

Pipeline specifications

Software tools MARTINI, GROMACS, P-LINCS, VMD
Application Protein structure analysis
Chemicals Cholesterol