Dustin Danowski and Professor David Busath, Department of Physiology and Developmental Biology
Botulinum Neurotoxin (BoNT) is currently being widely researched a a possible treatment for many medical conditions, as well as for cosmetic purposes. However, the mechanism describing how this family of toxins interact with their receptors to cause endocytosis is not well understood. We have developed a model of the GT1b ganglioside receptor of BoNT that may be used for Molecular Modeling and Dynamics simulations. The development of such a model is necessary to understand the mechanism of BoNT endocytosis, which then helps us determine possible uses and dangers of the toxin as a medical treatment. We had planned to complete the preliminary tests of our model before the writing of this report. However, due to lengthy delays (totaling approx. 6 months) in correcting software issues and malfunctions, the tests of the model are currently underway and have yet to be completed. However, a description of these tests and their expected outcomes is given below.
Development of a molecular dynamics model for the GT1b ganglioside was done using the program CHARMM, or Chemistry at HARvard Macromolecular Mechanics. The preliminary model was programmed using Dr. David Busath’s linux cluster WOLF, a Dell cluster with 16 core processors. We worked to develop the preliminary GT1b model within the ChARMM force field for proteins and lipids, using an established sugar force field as a guide1. Length, Angle, Dihedral, and Force Parameters for GT1b were extracted from established parameters used to model proteins and lipids. X-ray crystallography is the established method for establishing such parameters, however GT1b cannot be crystallized, and thus no exact data were available to us. Thus, our parameters that are specific to the GT1b model are approximations based on established parameters for similar molecules. The model of GT1b is shown in Figure 1.
We are currently testing this model by simulating it in a lipid monolayer, and will compute the atomic density profiles from these simulation data. Atomic density profiles enable us to estimate the configuration of the GT1b lipid while it is in solution. These tests are based on a paper by Miller et. al. and will enable us to establish the validity of our model2. Three simulations are currently being run. The first is a simulation of a 20% GT1b 80% DPPC monolayer in solution, the second is a DPPC monolayer with a single GT1b glcosphingolipid, and the third is a simulation of just the DPPC monolayer. These simulations will allow us to compare the effects of GT1b on PC head group interactions. Further, these simulations will provide initial data from which we will determine a preliminary element density profile of GT1b in a lipid system. Later, the simulations will be expanded to monolayers and bilayers of 100 DPPC or DPPE lipid molecules.
These and subsequent simulations are being performed using the Marylou4 supercomputing system, a Linux cluster consisting of 630 Dell Poweredge 1995 nodes, with each node containing 2 Dual-core Intel Xeon EM64T (@ 2.6 GHz) processors. The element density profile of the ganglioside in monolayers of DPPC, and later DPPE, will then be determined. The monolayer mixtures of lipid and ganglioside consist of 80% lipid, 20% GT1b in a square array. Systems of 100% DPPC and DPPE are and will be used for comparison of the effects of large GT1b head groups on small PC and PE head groups. Standard heating and equilibration protocols are being used, followed by molecular dynamics simulations with standard random sampling. Eventually, 150 mM NaCl solution will be included in the simulations on both sides of the lipid bilayers, and are currently included on one side of the monolayer simulations, giving conditions similar to cell membranes in biological systems. Constant temperature, normal pressure, and normal surface tension are being used in constructing the simulation environment and boundary conditions. Surface tensions of 20 mN/m and 30 mN/m are used for the monolayers, and 60 mN/m will be used for the bilayers. The molecular dynamics runs are then used to simulate the molecular motion and energy of the GT1b molecule during a 10 nanosecond time interval. These molecular dynamics simulations will enable us to calculate the atomic density profile of GT1b in various lipid bilayers and monolayers. Our calculated profiles will be compared with scattering length density profiles obtained from x-ray reflectometry with monolayers2 which can be interpreted in terms of atomic density profiles. The order parameters for the lipid tails and the GT1b ganglioside tails will also be computed for these simulation runs, allowing us to compare the average tail tilt and possible hexagonal crystallinity of our structure with published values. On comparison of our calculated data to those data published from a laboratory setting, a positive correlation will help us to establish the validity of our GT1b model. Once validity is established, the model can be used to simulate GT1b under various experimental conditions. These simulations are useful for determining possible mechanisms of interaction between botulinum neurotoxins and their GT1b receptors
References
- Miller, C.E. D.D. Busath, B. Strongin, and J. Majewski. “Integration of Ganglioside GT1b Receptor into DPPE and DPPC Phospholipid Monolayers: An X-ray Reflectivity and Grazing Incidence Diffraction Study.” Biophysical Journal. 95 (2008): 3278-3286.