How To, Education & Training

Biomolecules at phase boundaries

Biomolecules at phase boundaries
of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
   iomolecules at Phase oundaries PETER AHLSTRC)M, l z JUKKA LAUSMAA, 1 PATRIK LOFGREN 1 and HERMAN J. C. BERENDSEN a I Department of Applied Physics Chalmers University of Technology and University of GOteborg S-41296 GOteborg Sweden 2Bioson Research Institute and Laboratory of Biophysical Chemistry University of Groningen Nijenborgh 4 NL-9747 AG Groningen the Netherlands Abstract. Experimental and theoretical studies of biomolecules at water surfaces and metal surfaces are presented. We studied lecithin molecules (monolayers) and phospholipase A2 at a water surface with molecular dynamics (MD) simulations. Results were compared with those obtained for a pure water surface. We also studied amino acids at a TiOz surface with thermal desorption spectroscopy in the presence and absence of water. Key words. Di(decanoyl)phosphatidyl choline, lecithin, surfactant monolayers, phospholipase A2 thermal desorption spectroscopy, titanium dioxide. 1. Introduction The interaction between biomolecules and phase boundaries is of great interest for several reasons. Many biomolecules, like phospholipids and bile salts, are amphiphilic and act as surfactants between a aqueous and a fatty phase. Others, like several lipases and phospholipases require an interface in order to function or to achieve their maximum activity [1]. The interactions between biomolecules and solid surfaces (e.g., metals and polymeric materials) play an decisive role in the field of biocompatibility, i.e. the interaction between man-made materials and living tissue. 2. Simulations of Water Air Interfaces We have performed molecular dynamics (MD) simulations of a series of systems having an air-water interface as a common denominator. The systems were in increasing complexity, a pure water slab, a water slab with a monolayer of phos- pholipids at each side and the same system with a phospholipase A2 molecule inserted at one of the monolayers. Results from these simulations were compared to (the scarce) experimental results for this type of systems, e.g., surface tensions. The results from the phospholipase simulation were also compared to results from a simulation of phospholipase with a monomeric substrate [2]. As far as possible the same simulation parameters were used for the different systems, cf. Table I. 2.1. PURE WATER SLABS As a reference to the simulations of phospholipid systems we performed two simulations of a 4 nm thick water slabs with a periodic repeat parallel to the slab (box size 4 x 4 nm). Two different water models, the SPC [3] and the SPC-E [4] Molecular Engineering 5: 235-243, 1995. © 1995 KluwerAcademic Publishers. Printed in the Netherlands.   36 PETER AHLSTROM ET AL. TABLE I Simulation parameters and thermodynamic averages. The solute includes NaCI, di decanoyl)lecithin and phospholipase where present). RMS deviations are given within parenthesis where appropriate. In all simulations a twin range cut-off of the interactions was used. Interactions between atoms within 0.8 nm from each other were evaluated every time step. For atoms with a distance between each other between 0.8 and 1.2 nm only electrostatic interactions were calculated. These were updated at every neighbour list update i.e. every tenth time step). In all these simulations the time step was 2 fs. Bonds were kept rigid with the SHAKE algorithm [5] and the temperature was coupled to an external bath [6] with a coupling time rr 0.1ps. For a more detailed description of the common simulation parameters, see [7] Simulation SPC SPC-E ML ML2 PLA2 Number of lecithins 0 0 2 x 42 2 x 42 2 × 42 Number of Na÷C1 0 0 8 8 7 Number of H20 2111 2111 4412 4412 3790 Simulation length ps) 200 200 155 155 155 Analyzed period ps) 100 100 75 75 50 Solvent temperature K) 307 3) 309 3) 308 2) 308 2) 308 2) Solute temperature K) - - 298 4) 298 4) 298 3) Surface tension mN/rn) 59 7) 72 6) 58 8) 62 17) 62 8) were used. They both have the same geometry and the same Lennard-Jones parameters centered on the oxygen) but differ slightly in charges, the hydrogen charge in the SPC-E model is +0.4238 vs. +0.4100 in the SPC model. This minor change in dipole moment leads to a notable change in the surface tension as calculated from the pressure in different directions see Table I); the surface tension rises from 59 mN/rn with the SPC model to 72 mN/m close to the experi- mental value) with the SPC-E model. As a comparison it can be mentioned that the generalized van der Waals theory [12] predicts that the surface tension depends linearly on the dipole moment raised to the fourth power under certain simplifying conditions [11]. The simulation results suffer from large uncertainties and, since the forces, and thus the pressures, are calculated with a truncation of the interaction at a cut-off distance, these surface tensions would need a long-range correction. Such a correction has been calculated for Lennard-Jones liquids cf., e.g. [8,9]) and at present we are developing such a correction for dipolar systems [10] based on the generalized van der Waals theory. In lack of such a correction which probably would increase the surface tension) and with the large uncertainties of the simul- ated surface tensions in mind we chose to use the SPC model in the continued simulations since it was used in simulation of phospholipase with a monomeric substrate [2]. 2.2. LECITHIN MONOLAYERS Monolayers of surfactants on airwater interfaces show several two-dimensional phases depending on the density, or long chain molecules at least the following phases are observed from low to high density): gas, liquid expanded LE), liquid compressed LC) and possibly solid. Short-chain lecithins do not appear to show any LE-LC transition but show a continuous decrease in surface tension as the  BIOMOLECULES AT PHASE BOUNDARIES 237 surface tension as the density is increased in the liquid region. We chose to simulate di(decanoyl)lecithin(DDPC) at the surface density that corresponds to the maximum rate of phospholipase A2 (0.78 nrn2/molecule). The experimental surface pressure [1], i.e. the decrease in surface tension compared to water, at this coverage of di(decanoyl)lecithin is about 12 mN/rn giving an experimental surface tension of about 60 mN/m. Three simulations of two monolayers of each 42 di(decanoyl)lecithins at each side of an approximately 4 nm thick water slab with different force fields were performed. We found that a force-field with reduced charges on the lecithin head-groups and with a Ryckaert-Bellemans potential for the tails [13,14] best reproduced experimental data. This simulation is called ML in the following. As a reference we present some results from a simulation with standard Gromos potential [15] ('ML2'). In the simulation we did not note any significant difference in the surface tension between water and a lecithin-covered surface for any of the force fields (cf. Table I). This could be due to a defiance in the force field for the lecithins, namely too small repulsion between different molecules which in its turn could be due to the united atom representation used for the tail atoms. One could note here that the halved charges in the ML simulation as compared to the ML2 simulation in this respect is compensated for by the decreased van der Waals radii in the tails. However, a simulation with halved charges and the larger van der Waals radii in the tails gave a similar (slightly higher) surface tension than both the ML and the ML2 simulations. These results seem to indicate that the effect of details in the lecithin-lecithin interaction on the surface tension is minor. On the other hand, recent results from studies of decane-water interfaces [16] show large effects on surface tension of small changes of the van der Waals parameters for the interac- tion between decane and water. A conclusion could be that the most important factor for the surface tension is the interaction between the phases and not within each phase. Both the ML and the ML2 simulations show very disordered monolayers with a large spread in the penetration depth of the lecithins into water and a disordered tail structure. The main difference between the simulations is a higher degree of ordering in the of the tails in the ML2 simulation and an increased hydration of the the phosphate group in the ML2 simulation compared to the ML simulation. This comes down to a more gel-like structure in the ML2 simulation than in the ML simulation, similar to what is observed for lipid bilayers [17]. This is also reflected in the fraction gauche dihedral interactions in the tails, about 17 in the ML2 and 27 in the ML simulation. The reorientation of the phospholipids is slow with correlation times for the head group dipole vectors in the order of 0.2 ns in the ML and 0.3 ns in the ML2 simulation. Also the diffusion of the lecithins in the surface plane is slow with diffusion coefficients in the order of 0.5 x 10 9 m 2 s -1 (ML) or 0.2 x 10 m 2 s -I (ML2) in the surface plane. These values are, however, very approximate since the mean square displacement does not fully approach a straight line during the short analyzable period. Measurements on tracer diffusion in di(palmitoyl)leci- thin(DPPC) monolayers [18] yields a ten times lower diffusion coefficient at low surface pressures. In part his fact represent the difference between DDPC and   38 PETER AHLSTROM ET AL DPPC but also the uncertain determination of the diffusion coefficient in the simulation is important. It has also been noted in several MD simulations that diffusion coefficients tend to come out higher than the experimental ones. This could be due to inaccuracies in the force fields, notably the lack of polarisability, cf. [191. 2 3 PHOSPHOLIPASE A2 AT LECITHIN MONOLAYERS Phospholipase Az ( PLA2 ) is an enzyme that degrades phospholipids at the so- called A position. Phospholipase A2 is highly stereospecific and degrades only 3- sn-phosphoglycerides. The mechanism reminds of the mechanism of serine pro- teases but is assumed to involve a water molecule as the nucleophile. Close to the catalytic site, a calcium ion is bound which is essential for the activity of the enzyme. The rate of degradation of the phospholipids increase by orders of magni- tude for an aggregated substrate if its surface density is not too high, i.e. PLA2 will degrade micelles and monolayers with a moderate density but not bilayers. The last simulation in this series was to a simulation in which we inserted a phospholipase molecule with its so-called inteffacial recognition site ( IRS ) di- rected towards one of the monolayers from the ML simulation, cf. Figure 1. The phospholipase was placed such that one protruding lecithin molecule was halfway into the active site of the phospholipase. During the course of the simulation the attraction between the PLA2 and the phospholipids due to van der Waals interac- tions steadily increased whereas the electrostatic attraction was more fluctuating but on average slowly increasing. The properties of the lecithin monolayer during this simulation were analyzed similarly to the ML simulation above. The lateral diffusion coefficient of the lecithins might be somewhat lower than in the ML simulation but with the large uncertainties the difference is barely significant. Also the inclination angles of the tails and and head groups (compared to the monolayer normal) remained grossly unchanged. The protein mainly retains its secondary structure and the R-factor (the r.m.s deviation of the structure as compared to the X-ray structure after optimal super- position) of the backbone is about 0.19 nm. Three regions show larger fluctuations around the average structure than the other regions, namely: (1) the/3-sheet around residue 80 that was poorly defined in the crystal structure of the porcine enzyme; (2) the surface loop around residue 65 which is known experimentally to be very flexible and not crucial to the enzyme stability; (3) The region around the calcium-binding loop. The large fluctuations in the last region are combined with the calcium ion losing most of its carbonyl ligands. This is obviously due to problems with the force field and has been noted in other simulations of phospholipase as well. We believe that the reason is that the carbonyl ligands to the calcium ion are highly polarized in the real protein but not in the simulation. The charges used in this simulation were adopted for protein simulations in which the electric fields on the carbonyl  BIOMOLECULES AT PHASE BOUNDARIES ,.a 239 Fig. 1. One configuration of the phospholipase and the monolayers in the PLA2 simulation thick lines: lecithins; thin lines: phospolipase; for clarity water molecules are not drawn). groups are by far not as high as when they are ligands to a calcium ion. We tried to change the van der Waals parameters in the line with earlier successful simula- tions of calcium binding proteins [20] but this did not show to be sufficient to keep the calcium ligands. It has been proposed to introduce a much higher charge on the carbonyl oxygens -0.58e instead of -0.38e) and correspondingly on the carbonyl carbons in order to correctly describe ion binding [21]. We believe that the best most physical) solution is to include polarizabitities for all atoms in the system. This requires, however, a complete reparametrisation of the force field and work is on its way to find the optimal solution to this problem.
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!