1 Combinatorial Synthesis and Drug Discovery Random Search e.g., buying paintbrush, mop, hammer, gluestick, nail, etc. etc. for hanging a picture frame on the wall Rational Design e.g., buying a nail for hanging the frame Combinatorial Search e.g., buying a box of nails of different sizes for hanging the frame

2 ~ 5 IND ~ 250 in vivo ~ 500 in vitro > 10,000 compounds > 15 years ~ 1 NDA > $500 million Time Cost Why Combinatorial Synthesis?

3 XXXYYY + ……….. 9 products  MIXTURE SYNTHESIS Traditional and Combinatorial Synthesis X X X Y Y Y = CH 3 COOH = CH 3 CH 2 COOH = (CH 3 ) 2 CHCOOH = PhNH 2 = p-CH 3 PhNH 2 = p-NO 2 PhNH 2 For example

4 Ser Ala Tyr Arg  SPATIALLY ADDRESSABLE SYNTHESIS XY + I) XY + II) X Y + III)

5  SPLIT AND MIX STRATEGY X XX Y Y YY MIX SPLIT

6  SPLIT AND MIX STRATEGY MIX and SPLIT

7  DECONVOLUTION BY TAGGING WITH GC-ANALYZABLE MOLECULES MIX SPLIT = = = ….and many more.

8  DECONVOLUTION BY TAGGING (continued) MIX and SPLIT

9  DECONVOLUTION BY TAGGING - Real Example

10  DECONVOLUTION BY ENCODED DNA MIX and SPLIT SPLIT A1 N1 A2 N2 A3 N3 A2A2 A1A2 N1N2N2N2 A3A2 N3N2 A2A1 A1A1 N1N1N2N1 A3A1 N3N1 A2A3 A1A3 N1N3N2N3 A3A3 N3N3 A2 A3 A1

11 Parallel Synthesis and Screening of a Solid Phase Carbohydrate Library Rui Liang, Lin Yan, Jennifer Loebach, Min Ge, Yasuhiro Uozumi, Klara Sekanina, Nina Horan, Jeff Gildersleeve, Chris Thompson, Andri Smith, Kaustav Biswas, W. Clark Still, Daniel Kahne * A solid phase carbohydrate library was synthesized and screened against Bauhinia purpurea lectin. The library, which contains approximately 1300 di- and trisaccharides, was synthesized with chemical encoding on TentaGel resin so that each bead contained a single carbohydrate. Two ligands that bind more tightly to the lectin than Gal-  -1,3-GalNAc (the known ligand) have been identified. The strategy outlined can be used to identify carbohydrate-based ligands for any receptor; however, because the derivatized beads mimic the polyvalent presentation of cell surface carbohydrates, the screen may prove especially valuable for discovering new compounds that bind to proteins participating in cell adhesion. Science (1996) 274;  An Example of Combinatorial Approach in the Discovery of a Potent Ligand

12  Synthesis of the Library

13  High Throughput Screening of the Library Derivatized TentaGel beads were washed three times with 1 ml of PBST buffer and then suspended in 1 ml of PBST containing 3% bovine serum albumin. The beads were incubated at room temperature for 3 hours on a rotary shaker in 1 ml of a biotin-labeled lectin solution and then washed three times with 1 ml of TBST buffer containing 1% BSA. The beads were incubated on a rotary shaker for 20 min at room temperature in 1 ml of alkaline phosphatase-coupled streptavidin. The beads were washed three times with 1 ml of alkaline phosphatase buffer and kept in the alkaline phosphatase buffer prior to staining. A portion of the beads was transferred to a petri dish and the alkaline phosphatase buffer was replaced with 200 µl of a solution containing 5-bromo-4-chloro-3-indolyl phosphate (BCIP) and nitro blue tetrazolium (NBT). Color development was observed under a low-power microscope. The staining was terminated by washing the beads twice with 200 µl of sodium EDTA solution. The colored beads were picked out manually under the microscope for decoding.

14  High Throughput Screening of the Library A portion of the beads in the library after 5 min of staining. The dark bead in the center was identified as a hit. This level of contrast shown in the photograph was representative [60X].

15 Major Disadvantages of Combinatorial Synthesis Limited structural diversity of synthetic library Wastage of much of raw material Wastage of assay material No de novo drug designed as yet based solely on combinatorial synthesis How about combinatorial virtual library screening?

16 Combinatorial Virtual Hexasaccharide Library 19 Disaccharides as building units for a hexasaccharide 19  19  19 = 6,859 topologies!

17 New Computational Design Principle  Dual Filter Algorithm Build Virtual Library of HS Hexasaccharides (‘Average Backbone’ Hypothesis) Minimize GAG Sequences (Glycosidic Bond Torsion Constraints) Scoring (‘Affinity’) Test (10,000 GA iterations, one run per GAG) Score in top 1%? Convergence (‘Specificity’) Test (100,000 GA iterations; three runs per GAG) RMSD  2.5 Å? END DISCARD START YES NO High-Affinity, High Specificity Sequence(s)

18 Combinatorial Virtual Library Screening Raghuraman, Mosier and Desai (2006) J. Med. Chem. 49,  Application of 1 st Filter – the ‘Affinity’ Filter

19 Finding a Needle in a Haystack!  Application of 2 nd Filter – the ‘Specificity’ Filter

20 Predictability of Binding Geometry  Geometry Search with Flexible Ligand