Multiscale Modeling of Lipid Bilayer Interactions with Solid Substrates

1
Multiscale Modeling of Lipid Bilayer Interactions
with Solid Substrates

• David R. Heine, Aravind R. Rammohan, and Jitendra
  Balakrishnan
• October 23rd, 2008
• RPI High Performance Computing Conference

2
Outline

• Background
• structure of lipid bilayers
• applications of supported lipid bilayers
• Modeling challenges
• Atomistic modeling
• Mesoscale modeling
• Experimental work
• Conclusions

3
Lipids and Bilayers
4
Technological Relevance of Supported Lipid
Bilayers

• SLBs are important for various biotech
  applications
• Biological research
• Model systems to study the properties of cell
  membranes
• Stable, immobilized base for research on membrane
  moieties
• Biosensors for the activity of various biological
  species
• Cell attachment surfaces
• Pharmaceutical research
• Investigation of membrane receptor drug targets
• Membrane microarrays High throughput screening
  for drug discovery
• How does bilayer-substrate interaction affect
  bilayer behavior?

5
Supported Lipid Bilayers at Corning

• Applications Membrane-protein microarrays for
  pharmaceutical drug discovery
• Substrate texture is important in the adhesion
  and conformation of bilayers on the surface
• Crucial for the biological functionality of
  bilayers
• Objective Quantify the effect of substrate
  topography and chemical composition on bilayer
  conformation and dynamics

6
Bilayer Length Time Scales

• Bilayer dynamics vary over large length and time
  scales, suggesting a multiscale approach.

Length Scales
Stokes Radius 2.4 nm
Bilayer Thickness 4 nm Area per lipid 60 /- 2
Å2
Undulations 4 Å 0.25 mm
7
Multiscale Approach

• Atomistic model
• capture local structure and short term dynamics
• Mesoscale model
• capture longer length and time scales
• sufficient to look at interaction with rough
  surfaces

8
Atomistic Model

• The bilayer is composed of 72 DPPC lipid
  molecules described in full atomistic detail
  using the CHARMM potential
• Water uses the flexible SPC model to allow for
  bond angle variations near the substrate
• The substrate is the 100 face of a-quartz with
  lateral dimensions of 49 x 49 Å described by the
  ClayFF potential

9
Simulation Technique

• System is periodic in x and y directions with a
  repulsive wall above the water surface in the z
  direction
• NVT ensemble must be used since pressure control
  is prohibited by the solid substrate
• Temperature is maintained at 323K with a
  Nose-Hoover thermostat
• Total energy and force on the bilayer are
  extracted during the simulation.

Heine et al. Molecular Simulations, 2007,
33(4-5), pp.391-397.
10
Simulation Technique

• System is periodic in x and y directions with a
  repulsive wall above the water surface in the z
  direction
• NVT ensemble must be used since pressure control
  is prohibited by the solid substrate
• Temperature is maintained at 323K with a
  Nose-Hoover thermostat
• Total energy and force on the bilayer are
  extracted during the simulation.

Heine et al. Molecular Simulations, 2007,
33(4-5), pp.391-397.
11
Comparison with Experimental Measurements
Bilayer-Substrate Interaction Energy from
Simulations
Simulations show an energy minimum at a
separation of 3 to 3.5 nm
SFA Measurements Between Substrate and Bilayer
Experimental measurements show a repulsion
starting around 4 nm and pullout at 3 nm
separations
courtesy J. Israelachvili, UCSB
12
Bilayer structure near the substrate

• Lower monolayer is compressed in the vicinity of
  substrate
• Upper monolayer seems relatively unaffected

13
Effect of substrate on lateral lipid diffusion

• Reduction in lateral diffusivity observed,
  compared to free bilayers
• Bulk simulations match diffusivity of free
  bilayers
• Suppression of transverse fluctuations near
  substrate inhibit a key mechanism for lateral
  diffusion

Transverse lipid motion enables lateral diffusion
Substrate reduces transverse motion reduces
diffusivity
Experimental value For free bilayers
14
Atomistic Simulation Results

• MD simulations show bilayer-substrate equilibrium
  separation of 3 3.5 nm, in agreement with SFA
  experiments
• Lateral diffusion of the lipid head groups
  decreases as the bilayer approaches the substrate
• Suppression of transverse fluctuations may be
  responsible for reduced lateral diffusion

15
Mesoscopic Model

• Conservative force
• Elastic stretching of bilayer
• Bending modes of bilayer
• Surface interactions
• Other (electrostatic, etc.)

• Dissipative force
• Formulation based on Newtonian solvent viscosity

• Random force
• Formulation based on fluctuation-dissipation
  theorem

16
Mesoscopic Modeling of Supported Lipid Bilayers

• Continuum representation to study large length
  and time scales
• 1 mm2, 1 ms
• Allows study of bilayer behavior on textured
  substrates
• Dynamic model that includes effect of solvent and
  environment

All dimensions in nanometers z axis not to scale
17
Mesoscopic Model Results
Substrate topography contours
Membrane topography contours
18
Mesoscopic Model Results
19
Mesoscopic Model Results

• Allows study of bilayer on micron and microsecond
  scales
• Minimum surface roughness of 4-5 nm required for
  membrane spanning conformation
• Spanning configuration important for maintaining
  bilayer mobility

20
AFM measurementsSpreading of Bilayer on
Synthetic Substrates
AFM image measurements courtesy Sergiy Minko,
Clarkson University
Ref Nanoletters, 2008, 8(3), 941-944
21
AFM measurementsSmoothening of membrane on rough
substrates
AFM image measurements courtesy Sergiy Minko,
Clarkson University
22
Lipid membrane conformationNumerical and
Experimental Results
AFM images courtesy Sergiy Minko, Clarkson U.
Macroscopic model predictions
Roiter et al. Nanoletters 8, 941 (2008)

• Model shows membrane coating up to about 4-5 nm
• AFM images show membrane coating 5 nm particles

23
Conclusions

• MD simulations show bilayer-substrate separation
  of 3 3.5 nm, in agreement with SFA experiments
• MD simulations show reduced lateral diffusion in
  lipids as the bilayer approaches the substrate
• Mesoscopic model shows membranes coat particles
  up to 4 5 nm in diameter, in agreement with AFM
  observations
• Larger surface features are needed to achieve
  separation between bilayer and substrate
• High-performance computing has opened up new
  approaches for understanding biomolecule-substrate
  interactions, which aids design
• There is still plenty of room to grow as these
  models are still restricted in terms of size,
  timescale, and complexity

24
Acknowledgements

• Professor Sergiy Minko his group at Clarkson U.
• Professor Jacob Israelachvili his group at U.
  C. Santa Barbara

25
(No Transcript)
26
Lipid Behavior on Nanoparticles

• Bilayer conforms to Nanoparticles lt 1.2 nm
• Bilayer undergoes structural re-arrangement
  involving formation of holes between 1.2 22 nm
• Beyond 22 nm bilayer envelops the particle

Ref Nanoletters, 2008, 8(3), 941-944