(307f) De Novo Protein Design of Multimeric Proteins with Flexible Templates for Application to Designing Aggregating Peptides

As protein design methods develop and computing power improves it becomes increasingly possible to shift focus from the tertiary structure of the proteins to the quaternary structure of protein complexes. In such structures any number of identical (homo) or non-identical (hetero) proteins can come together to form a specified macromolecular structure. Such multimeric protein systems are prevalent in biological systems and can be either essential (e.g. enzymatic complexes) or disrupting (e.g. amyloid fibrils) to correct cell function. Taking into account all parts of such multimeric systems is important for accurate design. In recent years there has been much effort in producing design frameworks for the design of multimeric systems and protein-protein interfaces [1-3]. Here we introduce a novel de novo design method capable of designing for general multimeric peptides/proteins. Such a framework is important for the study of amyloid aggregation, as well as the design of peptides capable of forming self-assembling superstructures or disrupting such formation [4]. The framework is applied to the design of novel aggregating tripeptides and hexapeptides and validated experimentally.

The novel protein design method capable of improving binding affinity (aggregation) of a multimeric system involves three stages. The first stage is a Mixed Integer Linear Programming model [5-7] implemented using a framework enhanced for use in the peptide aggregation design problem by allowing for the possibility of identical amino acid positions in homo-multimer systems. A flexible template structure is initially determined through Molecular Dynamics simulations and a centroid-centroid 8-bin force field is employed to assess the potential energy of the system. This potential energy is minimized to generate a rank-ordered list of designed peptide sequences. The second stage calculates the Fold Specificity of each designed sequence for the flexible template structure.

The third stage is a novel framework that employs Molecular Dynamics simulations in CHARMM [8] to predict the assembly of multiple peptides into a macromolecular structure. These dynamic runs are used to generate an ensemble of peptide structures in solution, as well as ensembles of aggregate structures of varying size. From these ensembles we calculate approximate molecular partition functions of the structures, which are used to calculate an approximate binding affinity [9]. The more accurate the partition functions are, the more precise the binding affinity will be, and so this process is iterated, sampling more and more conformations, until the partition functions converge.

The work detailed in the presentation shows the development of a general de novo protein design method for multimer systems and its application to the design of aggregating tripeptide and hexapeptide systems. The first stage of the method optimizes protein-protein interactions that take place in an aggregate system based on a template generated through molecular dynamics. The rank ordered list of sequences are then run through a molecular dynamics framework to generate an ensemble of structures for approximate binding affinity calculations. The top sequences have been experimentally tested for aggregation through the detection of hydrogel formation.

[1] Kortemme, T. and D. Baker. Computational design of protein–protein interactions. Current Opinion in Chemical Biology, 8:91–97, 2004.

[2] Bellows, M. L., H. K. Fung, C. A. Floudas, A. Lopez de Victoria, and D. Morikis. New Compstatin Variants Through Two De Novo Protein Design Frameworks. Biophysical Journal,98:2337–2346, 2010.

[3] Bellows, M. L., M. S. Taylor, P. A. Cole, L. Shen, R. F. Siliciano, H. K. Fung, and C. A. Floudas. Discovery of entry inhibitors for HIV-1 via a new de novo protein design framework. Biophysical Journal, 99:3445–3453, 2010.

[4] Colombo, G., P. Soto, and E. Gazit. Peptide self-assembly at the nanoscale: a challenging target for computational and experimental biotechnology. Trends in Biotechnology, 25(5):211-218, 2007.

[5] Fung, H. K., M. S. Taylor, and C. A. Floudas. Novel formulations for the sequence selection problem in de novo protein design with flexible templates. Optimization Methods and Software, 22:51-71, 2007.

[6] Fung, H. K., C. A. Floudas, M. S. Taylor, L. Zhang, and D. Morikis. Toward Full-Sequence De Novo Protein Design with Flexible Templates for Human Beta-Defensin-2. Biophysical Journal94:584-599, 2008.

[7] Fung, H. K., W. J. Welsh, and C. A. Floudas. Computational De Novo Peptide and Protein Design: Rigid Templates versus Flexible Templates. Industrial & Engineering Chemistry Research, 47:993-1001, 2008.

[8] Brooks, B. R., et al. CHARMM: the biomolecular simulation program. Journal of Computational Chemistry, 30(10):1545-1614, 2009.

[9] Lilien, R. H., B. W. Stevens, A. C. Anderson, and B. R. Donald. A novel ensemble-based scoring and search algorithm for protein redesign and its application to modify the substrate specificity of the Gramicidin Synthetase A Phenylalanine Adenylation enzyme. Journal of Computational Biology, 12:740-761, 2005.