Prof. Amiram Goldblum's laboratory of molecular modeling and drug design develops algorithms and computer techniques and employs them, as well as other computational methods, for solving problems in structural biology of proteins and associated problems in ligand binding and drug design.
A central activity in the laboratory is the development of novel optimization methods for solving complex combinatorial problems that are part of most research subjects in structural biology and in drug design. A large scope may be addressed by our novel core technology, called Iterative Stochastic Elimination (ISE) which is used to find optimal solutions to diverse issues such as molecular conformational ensembles, multiple loops in proteins, protein-ligand docking, protein design, focused libraries, cheminformatics and others.
ISE was recently devised in our lab, and is now extended by PhD and MSc students to:
- Constructing conformational ensembles of large (>8 amino acids) loops
- Predicting the preexisting ensembles multiple protein loops
- Building populations and computing properties of large ligands
- Structure based design from molecular scaffolds and fragments
- Predicting molecular properties from database comparisons
- Assigning Drug likeness and molecular bioactivity to molecules
- Docking molecules to their targets
- Design of new functions in proteins
- Designing protein sequence for inhibiting protein – protein interactions
- Differentiating between folding and non-folding sequences
- Assigning Mollecular Bioactivity Indexes and Drug Like Indexes
- Chemoinformatics and QSAR applications
Recent Experimental Success with ISE:
- Acetylcholinesterase inhibition: An ISE model for distinguishing between active inhibitors and non-inhibitors was constructed and applied to the ZINC database with more than 2 million molecules. Top 800 molecules were docked (ISE-dock) and 9 best ones by MBI and docking results were purchased for in vitro studies at the lab of Prof. Marta Weinstock-Rosin (Excellon (rivastigmine) inventor). Only 5 could be dissolved, and 3 displayed IC50 values of 3.25, 3.5 and 3.75 micromolars.
- Computational Protein Design: we designed 10 peptides (12 positions in each) to block the dimerization/tetramerization of the N-terminal of Bcr-Abl, that triggers the proliferation of Chronic Myeloid Leukemia (CML) cells. The peptide sequences were sent to Prof. Martin Rutahrdt at Goethe University, Frankfurt, and 6 of them inhibited the proliferation of CML cells in vitro more than a reference peptide that was used for the design.
- Differential Toxicity: We applied ISE to distinguish between molecules that displayed toxicity to cancer cells (HCT116 Colon Cancer cell line) more than to normal cell culture (human), and used the results to discover other toxic molecules (about 150 molecuels were picked out of 30,000, with 10% of them showing differential toxiciy, measured both at Targetex company in Budapest, our partners in the CancerGrid consortium (www.cancergrid.eu). Further molecules (~250) were picked and purchased from the enamine catalog of about 1 million molecules, and some 25 of them showed differential cytotoxicity, with about 12 showing it in concentrations < 2 micromolar.