Computer simulations of peptide folding, aggregation, and adsorption to solid substrates
Biological and Biomimetic Chemistry
Final Report Abstract
In this project, we have investigated solvent properties and adsorption behavior of short synthetic proteins. From recent experiments, it is known that some of these peptides possess a specific adsorption characteristics at semiconductor surfaces like, for example, silicon (Si) and gallium-arsenide (GaAs) with certain crystal orientations at the surface. It could be shown in these experiments that permutations and pointwise mutations of the amino acid sequences can cause noticeably changes in the adsorption behavior of the peptides. In our study, our main interest was devoted to an amino-acid sequence that exhibits good binding propensity to the (100) surface of GaAs and poor binding strength to Si(100) and to a peptide with the same amino acid contents but with randomly permuted sequence. In the latter case, the adsorption strength to Si(100) improved noticeably. It was one of the main objectives of our project to get deeper insights into specific properties of peptide binding to semiconductor substrates by means of single-molecule Monte Carlo computer simulations of a hybrid model that enables a detailed analysis of the thermodynamics of folding and adsorption properties on atomic scales. On the other hand, it is required that the model is sufficiently simple allowing for an efficient simulation. The particular complexity of the problem lies in the competition of the folding and the adsorption transitions, both affecting the conformational changes of the peptides under the influence of thermal fluctuations. In a first step, we analysed solvent properties of these peptides. For that purpose, we performed simulated tempering computer simulations employing a simplified implicit-water all-atom protein model that had recently been developed in the Lund group. Although unstructured conformations dominate at room temperature, we found surprisingly clear evidence that the peptide with good Si(100) binding strength and the peptides with small Si(100) binding propensities exhibit different tendencies in forming secondary-structure (i.e., a-helical and ß-stranded) segments. We could also identify the amino acid proline and its different position in the sequences we compared with each other as being relevant for the different trends in structure formation. By a pairwise mutation regarding the proline positions, we could also show that then the trends of secondary-structure formation reverse. In order to check our experimentally not yet verified prediction that this trend reversal also changes the binding propensity to Si(100) substrates, we have developed and analysed in the second part of the project an extension of the peptide model, where the interaction with a Si(100) substrate has been incorporated. In our model, the Si substrate is considered as bare and flat. This approach is justified, as experiments performed in Semiconductor Physics Group at the Universität Leipzig revealed that the Si(100) substrate is not yet noticeably oxidized while the peptide adsorption process proceeds. The multicanonical computer simulations we performed with this model not only confirmed qualitatively the experimentally observed Si(100) binding specificity of the different peptides; we also found that the position changes of proline in the sequences actually reverse also the binding propensity of the peptides as predicted from our studies of the peptides' solvent behavior. The project will be continued. Possible next steps include studies of specific properties of peptide adsorption to other semiconductors, e.g., GaAs, the dependence of the binding strength from the crystal orientation at the surface, and cluster formation (aggregation) near the surface.
Publications
- Differences in Solution Behavior among Four Semiconductor-Binding Peptides, J. Phys. Chem. B 111, 4355 (2007)
S. Mitternacht, S. Schnabel, M. Bachmann, W. Janke, and A. Irbäck
- Sequence-Specific Peptide Adsorption at (100) Silicon Surfaces, Lund/Leipzig (2007)
M. Bachmann, K. Goede, S. Mitternacht, M. Grundmann, W. Janke, and A. Irbäck
