(6ki) Computational Protein Design Using Optimization Programs and Force-Field Calculations | AIChE

(6ki) Computational Protein Design Using Optimization Programs and Force-Field Calculations


Chowdhury, R. - Presenter, The Pennsylvania State University
Maranas, C., The Pennsylvania State University
Kumar, M., The University of Texas at Austin
Research Interests:

Computational protein design enables us to generate a focused library of “designer” proteins with a desired phenotype. Computational techniques typically aid to scan thousands of amino acid type and spatial orientation combinations in reasonable time as a lever for directed evolution of proteins up a fitness landscape along the desired property – under polynomial type. These in turn, serve as precedents for experimental validation. Herein, we have laid out mixed-integer linear programming (MILP) as a method of efficiently performing more than five simultaneous amino acid substitution on a polypeptide backbone to tailor its functionality. Our work is divided into three main paradigms: (a) computational redesign of channel proteins for precise sub-nm tuning of pore size and chemistry for desired bioseparations, (b) altering substrate specificity of enzymes by exploring novel scaffolds and binding chemistry, and (c) de novo design of variable antibody domains targeted against a specific antigen epitope. First, monodispersed angstrom-size pores embedded in a suitable matrix are promising for highly selective membrane-based separations. They can provide substantial energy savings in water treatment and small molecule bioseparations. Such membrane proteins (primarily aquaporins) are commonplace in biological membranes but difficult to implement in synthetic industrial membranes due to their modest and non-tunable selectivity. PoreDesigner, a computational design workflow for the redesign of the robust beta-barrel Outer Membrane Protein F as a scaffold targeting of any specified pore diameter (spanning 3–10 Å), internal geometry and chemistry. PoreDesigner uses a mixed-integer linear program to optimally place long side -chain hydrophobic amino acids at the pore constriction region that yield a smaller and more hydrophobic pore by maximizing the interaction energy between the pore wall and the permeating water wire. This lays a promising foundation for using biomimetic membrane materials as viable filtration assemblies for performing precise angstrom-scale separations. Secondly, Iterative Protein Redesign and Optimization Suite of programs (IPRO) uses a MILP to optimally alter binding pocket residues of an enzyme to alter substrate affinity. A new algorithm has been devised that besides predicting amino acid substitutions, also predicts the possibility of introducing an insertion or deletion at the binding pocket to affect substrate recognition – thus using protein length as a lever for directed evolution. Finally, we use OptMAVEn-2.0 algorithm to rapidly design highly humanized antibody variable fragments targeted against an antigen epitope protein. This serves as a one-stop solution for designing antibody-mediated therapeutics during disease outbreaks upon experimental assessment of the best designs using immune-sorbent assays. Overall, in all cases the “designed” proteins are selected and ranked depending on thermodynamics of interactions by computing stability scores or interaction energy terms using force-field calculations that account for non-covalent forces (electrostatics, van der Waals, and solvation effects).

Teaching Interests:

While an integrative research and education proposal is always a priority of NSF CAREER program, developing a platform to maximize the public benefits of any proposed work can be tricky. To this end, advances in information technology provide a growing set of tools that can be applied to get the message out to a wide range of audience. To this end, I will pursue two different objectives: (a) Prepare workshops to engage structure biology mentorship to middle school, high school and undergraduate students, and (b) Introduce a graduate level coursework titled “Optimization in Structural Biology” to provide a relevant knowhow of linear algebra with tantamount interfacing with relevant biochemistry and chemistry.

For the first, I am already a part of NSF-CAREER proposal of Dr.Rajib Saha, Asst. Professor, department of Chemical Engineering, University of Nebraska, Lincoln and will have partial support for organizing “fun-camps” for young individuals from middle and high school which will use “Lego” and “flexible polymers”, and hands on computer simulations as models to provide visual understanding of protein-protein interaction, lock and key hypothesis of enzymes and protein folding dynamics. For undergraduate students, I will be able to arrange visits to the headquarters of “Applied Biomimetic Inc.” where Dr. Mariusz Grzelekaowski -the Chief Technical Officer, has expressed interest in showing how they express and purify the designed channel proteins (from PoreDesigner work – where we have an agreement over our Intellectual Property) and use them for various state-of-the art bioseparations. Additionally, I plan to organize popular lecture series where each student from my lab will provide short “layman’s” description of their research dring these camps.

Regarding the coursework - I have an Online Teaching Certificate from World Campus Penn State (OSL-2050) which has trained me how to frame coursework material compliant with university policies. I think, currently there exist no single coursework which talks about (a) optimization programs, (b) structural biology, and (c) force-field based molecular mechanics under the same umbrella. This course work will provide students engaged in experimental pursuits a clear understanding of the computational paradigm, and computational students a deeper dive into how to write algorithms that could be employed for in silico protein evolution for any desired biochemical objective. The course will have two guest lectures (a) one by an experimental professor and (b) one by another computational scientist, who will tie the course content to real-world projects they have performed thus providing insights about systems that they have worked on.