Integrating Microflow NMR into Fragment-based Drug Discovery Daniel

Integrating Microflow NMR into Fragment-based Drug Discovery Daniel S. Sem Chemical Proteomics Facility at Marquette (CPFM) Department of Chemistry Marquette University Protasis Webinar 7/15/08

Integrating Microflow NMR into Fragmentbased Drug Discovery Research Focus: Drug discovery using NMR l l l Background on Fragment-based Drug Discovery NMR Equipment in the CPFM Flow Probe Applications – Routine quality control of compound collections – Protein 1 H-15 N HSQC screening of labeled proteins – Fragment-based screening using STD (saturation transfer difference) NMR

New Drug Design Paradigm: Fragment-Assembly Companies using fragment assembly approaches: Advanced Medicines / Theravance Sunesis (Thiol tethering) Structural Genomi. X Abbott (SAR by NMR) Vertex (SHAPES) And from a Chemical Proteomics slant: Triad Therapeutics l San Diego, California l Initially funded June, 1999 l Raised $42. 5 M in A and B rounds; $15 M in C round l 50+ employees

Modular Drug Design / Fragment Assembly 1 SAR by NMR 2 SHAPES 3 1. Reviewed by: Pellecchia, Sem & Wuthrich (2002) Nat. Rev. Drug Disc. 1, 211 -219. 2. Shuker, Hajduk, Meadows & Fesik (1996) Science 274, 1531 -1534. . 3. Fejzo et al. (1999) Chem. Biol. 6, 755 -769.

Combinatorial Library for Dehydrogenases Chemical inhibitors discovered across a gene family (dehydrogenases) Co mm on Va ria ble scaffold Oxidoreductase-1 Oxidoreductase-2 Oxidoreductase-N scaffold Reviewed by: Proof of concept: Pellecchia, Sem & Wuthrich (2002) Nat. Rev. Drug Disc 1, 211 -219. Sem et al. (2004) Chemistry and Biology 11, 185.

Screen NMR-designed Library for Inhibitors Specific for Target versus Antitarget: Assay results TB Target LDH scaffold 55 m. M 42 n. M Malaria Target DHPR DOXPR 26 m. M >50 m. M TB target proteomic leverage target specificity 12 m. M Malaria target > 50 m. M 10 m. M > 25 m. M 202 n. M (IC 50) 620 n. M Sem et al. (2004) Chemistry and Biology 11, 185. 100 n. M 7. 9 m. M

Triad l Technology Platform – Internal use only – Proprietary NMR-based drug design platform to generate drug leads – Chemicals, proteins, software, databases, methods l Proof of Principle completed (Series B, $30 M) l Leads for infectious disease targets l Dual Business Strategy (Drug Discovery) – Internal drug discovery & licensing early-stage leads – Evolved into late-stage licensing (IND candidates) Triad ceased operations on 3/19/04: Drug licensing business model not practical

Drug Discovery in Academics Can focus on developing enabling methods rather than drug leads. The longer view … l Focus on diseases with smaller markets; third world diseases. l The CPFM has resources to aid in drug discovery: compounds; databases; software; NMR screening capability – automation & labeled probes. l

NMR Equipment in the CPFM 600 MHz Varian NMR System Cryogenic probe (1 H/2 H/{13 C}/{15 N}) z-axis gradients 4 channels 300 MHz Varian NMR System Broadband probe z-axis gradients 60 sample change

NMR Equipment in the CPFM 400 MHz Varian NMR System x, y, z-axis gradients (imaging capability) 2 channels BB and inverse probes Protasis Cap. NMR Microflowprobe: TXI triple resonance, 1 H/2 H/{13 C}/{15 N} detection z-gradient, variable temp. , 10 u. L flowcell Automated sample introduction using LEAP Technologies (CTC Analytics) liquid handler Automation managed via Protasis One-Minute NMR (OMNMR) software

NMR-based Drug Discovery at Marquette’s CPFM l Focus on developing new methods l Blending: chemistry, NMR screening, informatics l Integrating use of microflow NMR l Focus on infectious disease www. marquette. edu/cpfm

Could we discover a new version of this CR-based biligand more easily? a) b) Avoiding extensive synthesis (for linking)? Using microflow NMR (speed; conserve samples)? CR = catechol rhodanine Privileged scaffold Strategy: Combine thiol tethering and STD-based screening, using a Flow NMR platform

STD-based screening (STD = saturation transfer difference) CF-STD NMR 1, 2: Cofactor fingerprinting with saturation-transfer-difference NMR STD NMR 3 1. 2. 3. Stockman & Dalvit (2002) Prog. NMR Spectr. 41, 187 -231. Yao & Sem (2005) FEBS Lett. , 579, 661 -666. Mayer & Meyer (1999) Angew. Chem. Int. Ed. 38, 1784 -1788. Cofactor structures

STD-based screening (STD = saturation transfer difference) CF-STD NMR 1, 2: Cofactor fingerprinting with saturation-transfer-difference NMR STD NMR 3 STD: PKA + c. AMP, c. CMP, c. GMP 1 D: c. AMP, c. CMP, c. GMP STD: RSP 2 + c. AMP, c. CMP, c. GMP 1. 2. 3. Stockman & Dalvit (2002) Prog. NMR Spectr. 41, 187 -231. Yao & Sem (2005) FEBS Lett. , 579, 661 -666. Mayer & Meyer (1999) Angew. Chem. Int. Ed. 38, 1784 -1788. 600 MHz; 25 m. M protein; 1 m. M cofactors

Flow Probe Applications

Using Flow Probe for HSQC Experiments To screen for folding conditions (ex. structural proteomics) To screen for fragment binding (ex. SAR by NMR) Our model protein: GB 1 (Ig. G binding domain from protein G) Well studied (ex. Frank et al. (2002) NSB 9, 877 -885)

Using Flow Probe for HSQC Experiments Our model protein: GB 1 (Ig. G binding domain from protein G; 56 AA) ~1 m. M, p. H 7 Spectra taken on the 400 MHz flow probe (10 u. L sample volume) Acquisition time varied 2 hrs. 5 hrs. Main advantage: automation and low sample consumption 10 hrs.

STD based screening with our drug target: DHPR (Dihydrodipicolinate reductase) Optimizing concentrations for flow-based screening (generally, we need > 50 u. M protein, and high ligand concentration) This requires use of reporter ligands to detect binding! Of course, [competitor] > [enzyme target] 100 u. M DHPR + ligand (NAD+): 80 m. M NAD+ 40 m. M NAD+ 20 m. M NAD+ 10 m. M NAD+ Acquisition time = 47 minutes (96 -well plate in < 4 days; this might be a secondary assay) Samples in D 2 O, 20 m. M K-phosphate, p. H 7. 6, 298 K

STD based screening with our drug target: DHPR (Dihydrodipicolinate reductase) Optimizing concentrations for flow-based screening – binding 2 ligands (generally, we need > 50 u. M protein, and high ligand concentration) 100 u. M DHPR + 10 m. M PDC + ? : 10 m. M PDC + 10 m. M NADH 10 m. M PDC + 10 m. M NAD+ 10 m. M PDC Acquisition time = 47 minutes Samples in D 2 O, 20 m. M K-phosphate, p. H 7. 6, 298 K

What’s going on? Flow-probe-based STD screening sees NAD but not NADH binds too tightly NADH seemed to increase the PDC STD effect? They bind to different sites, so possible synergy? Follow-up titration (this one is at 600 MHz w/cryoprobe):

Next: combine STD-based screening w/ thiol-tethering Search for fragments that bind in cofactor (NAD or CRAA) and substrate (PDC) sites {CRAA = catechol rhodanine acetic acid} What is thiol tethering? Bring weak binding fragments together (link) using disulfide bonds Pioneered by Erlanson, Wells and other at Sunesis Erlanson et al. (2000) PNAS 15, 9367 -9372. Erlanson et al. (2003) Nat. Biot. 21, 308 -314. Erlanson et al. (2004) Curr. Opin. Chem. Biol. 11, 730 -737.

Next: combine STD-based screening w/ thiol-tethering Search for fragments that bind in cofactor (NAD or CRAA) and substrate (PDC) sites Our approach Bring weak 2 weak binding thiol-containing fragments together in the cofactor and substrate sites of DHPR, then link them later Start with a first fragment that binds (CRAA) and screen for the second Detect binding based on competition STD and reporter ligands (NAD, PDC) CRAA

Use STD Screening to Identify Thiol Fragments that Fit in the PDC Site Relative STD for 99 thiols (screened in pools of 5) Flow probe, using 10 m. M PDC as a reporter (1 m. M thiol; 100 u. M DHPR; 45 min acquisition) Thiol Fragment Database: (99 thiols) www. marquette. edu/cpfm

Why do some thiol fragments cause an increase in the PDC STD signal? Perhaps binding at other sites in the tetramer. PDC Relative STD for 99 thiols (screened in pools of 5) Flow probe, using 10 m. M PDC as a reporter (1 m. M thiol; 100 u. M DHPR; 45 min acquisition)

Discovery that TNB (5 -thio-2 -nitrobenzoic acid) binds in the PDC site 20 m. M TNB, 200 u. M DHPR + 4 m. M PDC STD with 400 MHz microflowprobe, 1 hr. acquisition time => 25% decrease in TNB STD signal due to PDC

Proof that PDC and TNB (a thiol fragment) occupy the same site => Titration @ 600 MHz Competition of PDC (varied) against 2 m. M TNB (reporter) (100 u. M DHPR) Competition of TNB (varied) against 2 m. M PDC (reporter) (100 u. M DHPR) Note: STD doesn’t go to zero - perhaps because there are 4 active sites that are not equivalent?

A New Strategy: In situ thiol tethering and competition STD screening a same time The Goal: Screen various thiols (RS-) to see which can form a higher affinity biligand, blocking both sites (thereby decreasing STD signals for NAD and PDC reporters) In this way, we discovered a biligand with TNB (5 -thio-2 -nitrobenzoic acid)

Results of in situ thiol tethering / competitive STD

Synthesis of the Thiol-Tethered Biligand (to verify in situ hit)

Secondary Assay: In-gel binding to the colored CRAA 2 -TNB biligand

Secondary Assay: Steady-state inhibition: CRAA 2 -TNB biligand is competitive vs. NADH (Ki < 7+2 m. M)

Acknowledgements Graduate Students: Aurora Costache Huili Yao Xia Ge NMR Facility Manager: Sheng Cai, Ph. D. Chemistry Dept. American Heart Association Biomedical Technology Alliance NIH (600 MHz spectrometer) Marquette University www. marquette. edu/cpfm Fragment database, created with Sci. Tegic software