Combinatorial libraries: strategies and methods for ‘lead’ discovery

Alan Spivey Department of Chemistry University of Sheffield Key sources of information • WWW: – Diversity information pages [http://www.5z.com/divinfo/] – J.Combinatorial Chem. [http://acsinfo.acs.org/journals/jcchff/index.html] – Combi. Chem. H.T.S. [http://www.bscipubl.demon.co.uk/cchts/index.html] • Books – Combinatorial peptide and nonpeptide libraries-a handbook, Ed. G.Jung, VCH, Weinheim, 1996. – Combinatorial chemistry-synthesis and application, S.R.Wilson, A.W.Czarnik, Wiley, New York, 1997. • Reviews – ‘Combinatorial chemistry’, Chem. Rev. 1997, 97(2), special issue. – ‘Combinatorial chemistry’, Curr. Opin. Chem. Biol. 1998, 2(3) & 1999, 3(3). – ‘Combinatorial chemistry’, S. Borman, Chem & Eng. News 1997, Feb24, 43. Format and scope of lecture • What is combinatorial chemistry? • The drug discovery process • Approaches to combinatorial library synthesis: – mix and split synthesis – parallel synthesis – encoded tagging • Library types: – oligomeric libraries – template based libraries • Combinatorial drug discovery! What is combinatorial chemistry?

Combinatorial chemistry is a useful tool for rapidly optimizing molecular properties, particularly ones that are difficult to design a priori…

Nature uses a combinatorial approach to generate diverse functional macromolecules such as antibodies to recognize a vast array of antigens. The drug discovery process

Idea

• The total cost of bringing a new drug 2-20 years to market is typically ~£250m (i.e. Discovery 8000-10000 Research compounds EXPENSIVE!) • Of this, £170m is spent on DISCOVERY RESEARCH. Pre-clinical candidates 2-3 years • This reflects the large amount of TIME Pre-clinical 20-30 involved in synthesising new trials compounds compounds. Regulatory/ • A typical chemist can synthesise ethical clearance ~100 compounds a year using Clinical 3-5 years trials 4-5 traditional techniques. compounds • SOLID PHASE ORGANIC Registration SYNTHESIS (SPOS) and with health COMBINATORIAL CHEMISTRY are authorities Registration 2-3 years beginning to revolutionise this and launch 1 compound situation.

Marketed Drug Discovery chemistry: stage 1

Therapeutic area identified

Discovery Biology • High Throughput Screening (HTS): – Rapid, automated screening of Biological assay established compounds for specific biological Automation of activity. assay

Assay adapted for • Role of combinatorial chemistry: High Throughput Screening (HTS) – Very large libraries. Compound Combinatorial – Maximum diversity libraries. HTS Chemistry – Mix and split libraries (& parallel Lead structure synthesis). – Mixtures of compounds (& single compounds). Medicinal Combinatorial Chemistry Chemistry

Pre-clinical candidate Discovery chemistry: stage 2

Therapeutic area identified

Discovery Biology

Biological assay established • Medicinal chemistry:

Automation of – Systematic optimisation of molecular assay and physicochemical properties of lead compound Assay adapted for High Throughput Screening (HTS) • Role of combinatorial chemistry: Compound Combinatorial HTS Chemistry – Small libraries. – ‘Targeted/focussed’ libraries. Lead structure – Parallel synthesis libraries. – Single compounds. Medicinal Combinatorial Chemistry Chemistry

Pre-clinical candidate Traditional vs. combinatorial • Traditional synthesis:

A + B AB

compounds prepared one at a time, characterised and screened • Combinatorial synthesis:

A1B1 A1B2 A1B3 + A1-3 B1-3 A2B1 A2B2 A2B3 A3B1 A3B2 A3B3

2 reaction of 3 reagents Ax with 3 reagents By provides a library of 3 (i.e. 9) compounds AxBy 3 introduction of a third set of 3 reagents Cz increases the library size to 3 (i.e. 27) compounds AxByCz Approaches to ‘combinatorial’ library synthesis • In vivo - biological methods: – Phage display, plasmids, polysomes etc.

• In vitro - synthetic methods: – Mix and split using Solid Phase Organic Synthesis (SPOS). • Cleavage from the solid support following ‘mix and split’ results in complex mixtures (pools) of compounds. Screening of these mixtures yields ‘hits’ whose identity must be determined by ‘deconvolution’. • If screening can be performed ‘on-bead’ (i.e. ‘one-bead one-compound’ libraries) then deconvolution can be avoided. – Parallel synthesis using Solid Phase Organic Synthesis (SPOS). • Spatially separate synthesis of single compounds whose identity is uniquely defined by their location. Mix & split synthesis: libraries of mixtures of compounds in solution

A A 1 3 Split A2 • Screening complex mixtures of compounds 1 3 =3 A1 A2 A3 in solution can give false ‘hits’ due to synergistic effects. Mix & Split B1 B3 B2

A1B1 A1B2 A1B3 • Identification of a compound within the 32=9 A2B1 A2B2 A2B3 mixture responsible for the ‘hit’ requires A3B1 A3B2 A3B3 iterative deconvolution.

Mix & Split C1 C3 C2 • Houghton Nature, 1991, 354, 84.

A1B1C1 A1B1C2 A1B1C3 A B C A B C A2B1C1 2 1 2 2 1 3 Mix & A3B1C1 A3B1C2 A3B1C3 Cleave from A1B2C1 A1B2C2 A1B2C3 Screen Deconvolution resin mixture 3 3 =27 A2B2C1 A2B2C2 A2B2C3 'Pool' of all compounds Identity of A B C A B C A B C 'Hit' 3 2 1 3 2 2 3 2 3 in solution hit A1B3C1 A1B3C2 A1B3C3

A2B3C1 A2B3C2 A2B3C3 A3B3C1 A3B3C2 A3B3C3 Mix & split synthesis: ‘one-bead one-compound’ libraries

A A 1 3 Split • Requires a very sensitive screening A2 protocol which can accommodate resin 31=3 A A A 1 2 3 bound compounds.

Mix & Split B1 B3 B2 • Identification of a ‘hit’ compound on (or A1B1 A1B2 A1B3 2 from) a single bead (~100pm) by: 3 =9 A2B1 A2B2 A2B3 A3B1 A3B2 A3B3 – analytical methods e.g. Edman sequencing of peptides, MALDI-TOF Mix & Split C1 C3 MS, single bead NMR… C2 – reading ‘encoding tags’ on beads. A1B1C1 A1B1C2 A1B1C3

A2B1C1 A2B1C2 A2B1C3 Analytical A B C A B C A B C Mix & 3 1 1 3 1 2 3 1 3 identification A B C A B C screen A1B2C1 1 2 2 1 2 3 'on-bead' of compound 3 3 =27 A B C A2B2C2 A2B2C3 2 2 1 Isolate Identity of A B C A B C A B C 'hit' 3 2 1 3 2 2 3 2 3 bead hit A1B3C1 A1B3C2 A1B3C3

A2B3C1 A2B3C2 A2B3C3 A3B3C1 A3B3C2 A3B3C3 Clark Still’s encoded tagging protocol

A + 1%T A + 1%T 1 1 3 3 Split • Still Acc. Chem. Res. 1996, 29, 155. A2 + 1%T2 T1 T2 T3 A1 A2 A3 • Each ‘monomer’ used in the library synthesis has an associated encoded Mix & Split B1 + 1% T4 B3 + 1% T6 tag.

B2 + 1% T5

T1 T4 T1 T5 T1 T6 • The tags are chlorinated aromatic A1B1 A1B2 A1B3 compound which can be analysed at T2 T4 T2 T5 T2 T6 sub-picomolar levels by Electron A2B1 A2B2 A2B3 Capture Gas Chromatography T T T 3 T4 3 T5 3 T6 (ECGC). A3B1 A3B2 A3B3 • Allows for hit identification at one- Mix & screen bead fidelity for any type of library 'on-bead'

Isolate 'hit' Identity of beads hit cleave tag by oxidation and 'read' by ECGC Mechanism of Clark Still encoded tags

• Still J. Org. Chem. 1994, 59, 4723.

O O Rh(O2CCF3)3 N2 OMe [Rh] OMe CH2Cl2, Δ Cl n m n Clm O O O O

10 different α,ω-diols (n = 3-12) 40 different tags 4 different chlorophenols (m = 2-5) Buchner reaction OR CH2Cl2

OR

1) ceric ammonium O OTMS nitrate (CAN) n OO Clm n 2) OTMS Clm MeO 'read' using Me NTMS Electron Capture Gas Chromatography O Parallel synthesis: spatial separation gives single compounds A A 1 3 Split A2 • Suitable for the synthesis of relatively small libraries as each compound requires its own A1 A2 A3 reaction ‘well’. Split • Each reaction ‘well’ may be anything from a B1 B2 B3 small flask to a radio-frequency tagged ‘tea A1B1 A1B2 A1B3 bag’ to an etched region on a silicon chip! • Once screening has identified a hit no further A2B1 A2B2 A2B3 work is required to deduce the identity of the A3B1 A3B2 A3B3 active compound although it is routine practice Split ± cleaveage to independently verify the structure. from resin C1 C2 C3 Screen Spatial location A1B1C1 A1B1C2 A1B1C3 each of hit defines 'position' its identity etc (x27)... Identity of 'Hit' or hit

A1B1C1 A1B1C2 A1B1C3 etc (x27)...

spatially separated libraries Keeping track of ‘tea-bag’ parallel synthesis: Irori radio-frequency tagging • http://www.irori.com/ Library types

• There are essentially two strategically distinct types of library at the molecular level:

Building Building Building Building block 1 block 2 block 1 block 2

Template

Building Building Building Building block 3 block 4 block 3 block 4

Oligomeric Template based

peptides/peptoids drug-like molecules oligonucleotides natural product-like molecules oligosaccharides heterocycle based molecules unnatural oligomers polyaromatics Balasubramanian’s peptide library

• Protein tyrosine phosphatase substrate library (oligomeric). • Balasubramanian J. Am. Chem. Soc. 1997, 119, 9568. • Library synthesis:

1) Fmoc-Aaa-OH, PyBOP, n cycles HOBt, DIPEA Tx 2) piperidine K K N-G-Q-Q-P-I-L--Xaa-Xaa-A-Xaa-Fmoc NH2 amino methyl mix and split functionalised with encoding Kieselguhr OPO3H

one-bead one-peptide encoding via dummy peptide sequence incorporating encoded library glycine in place of phosphotyrosine Balasubramanian’s peptide library

• Protein tyrosine phosphatase substrate library (oligomeric). • Balasubramanian J. Am. Chem. Soc. 1997, 119, 9568. • Library screening:

Tx K N-G-Q-Q-P-I-L--X3aa-X2aa-A-X1aa-Fmoc 1) leukocyte antigen receptor protein tyrosine phosphatase (PTP) 2) α-chymotrypsin OPO3H 3) fluorescent labelling of N-terminus with carboxyfluorescein: F one-bead one-peptide encoded library

Identity of PTP sequence T encoding tag peptide x substrate sequences Hits = K N-G-Q-Q-P-I-L F "preference for at least two acidic residues in variable positions and a glutamic acid residue at X1aa" Beck-Sickinger’s cyclic peptide library • Neuropeptide Y analogue library (oligomeric). • Beck-Sickinger J. Org. Chem. 1999, 64, 4353. • Library synthesis:

HO NHFmoc CO2Me w CO2Me w NHFmoc OH O Wang resin PPh3, DEAD 1) piperidine n Mitsunobu parallel synthesis 2) Fmoc-Aaa-OH, DIC, HOBt cycles 3) piperidine

CO2Me

w NH-R-Q-R-Xaa-K-S-P-Y-Xaa-H O 1) LiOH 2) TBTU, HOBt, DIPEA O Xaa-Y-P-S-K-Xaa-R-Q-R TFA w NH-R-Q-R-Xaa-K-S-P-Y-Xaa O spatially separated solution library Beck-Sickinger’s cyclic peptide library

• Neuropeptide Y analogue library (oligomeric). • Beck-Sickinger J. Org. Chem. 1999, 64, 4353. • Library screening:

Xaa-Y-P-S-K-Xaa-R-Q-R competitive binding assay in solution with radiolabelled neuropeptide Y Hits identity defined by spatial location spatially separated solution library "weak competitive binding at μM level by range of derivatives" Oligonucleotide libraries

• These oligomeric libraries are generally prepared using in-vivo ‘biological’ methods and screened using Systematic Evolution of Ligands by Exponential enrichment (SELEX) procedures.

• e.g. The discovery of very high-affinity RNA and DNA ligands to human IgE which inhibit binding to the Fcε receptor I.

• Wiegand J. Immunology 1996, 157, 221 (and references therein). Unnatural backbone oligomer libraries Backbone Monomers

R O 1 H R O N Fmoc OOPNP N O N H H O R2 O oligocarbamate

R1 OO NPr2 R O O P NC P ODMTr O P O O O O O R2 oligo-phosphodiester

R 1 H OO R NS Boc N S N SO Cl H O 2 O R2 H vinylogous sulfonamidopeptide OOO O NNN + R-NH2 Cl R1 R2 R3 vinylogous sulfonamidopeptide Schultz’s purine library • Kinase inhibitor library (template based). • Schultz Science 1998, 281, 533. • Library synthesis: Cl R1 N N Cl NH 6 N N N N N N F R NH H 1 2 2 OH N N F N N F O O O Mitsunobu reaction PS resin with acid labile alcohol linker R2NH2 pre-functionalised with hydroxyethyl Nucleophilic group parallel synthesis substitution R 1 NH (C-2) N R N 1 NH N N NH N N TFA R2 N N NH OH O R2 spatially separated solution library Schultz’s purine library • Kinase inhibitor library (template based). • Schultz Science 1998, 281, 533. • Library screening:

Screen each position for inhibition of R1 h-CDK2-cyclin A kinase complex NH identity defined by Hits spatial location N N

N N NH

R2 3 & 4 substituted anilines OH & benzylamines Ala, Val or Ile derived spatially separated amino solution library alcohols HN Cl 6 N N small alkyl R4 e.g. Me 2 9 N N N OH H R3 Schreiber’s ‘natural product’ library • Protein epitope binding library (template based). • Schreiber J. Am. Chem. Soc. 1998, 120, 8565. • Library synthesis: O I Me I OH O O O N HO H Me O O N N NH2 H O O H T PyBOP, DIPEA HATU, DIPEA N PS with NMP, rt. O DMAP, rt H H photolabile O amide linker O mix and split R1 with Clark Still encoding Pd(II)/Cu(I) R1 R1 O R1 O H R HN O 2 N R3CO2H R2NH2 R2HN N N R O O H 3 T O O H T N HO O H T N H DCC, DMAP N O H H H O O O O O O one-bead one-compound encoded library Schreiber’s ‘natural product’ library

• Protein epitope binding library (template based). • Schreiber J. Am. Chem. Soc. 1998, 120, 8565. • Library screening:

R1 O compound from single R2HN N bead in 'nano-droplet' Hits R3 O O NH2 R H screen for 1 O protein binding O O O hv R2HN N "...several members of this R O O H 3 N T library activate a reporter H gene in mink lung cells!" O O O T Identity of hit one-bead one-compound 'read' using encoded library Electron Capture Gas Chromatography Nicolaou’s sarcodictyin library • Tubulin-microtubule disruptant library (template based). • Nicolaou J. Am. Chem. Soc. 1998, 120, 10814. • Library synthesis: Me OAc H Me

O OAc Me Me Me OR1 H H Me 1) NaOMe H OH O 2) R -LG Me Me 1 O OH OTIPS O H O H 3 O PPTS, CH2Cl2 Me Me Me Me α,ω-diol functionalised OTIPS OTIPS hydroxymethyl-PS 1) TBAF parallel synthesis 2) Dess-Martin 3) NaClO2 4) R2-LG Me OR1 H Me Me OR1 spatially separated H Me library in IRORI O PPTS, R3-OH 'tea bags' O H O R3 Me Me OR H O O 2 Me Me OR O 2 Nicolaou’s sarcodictyin library

• Tubulin-microtubule disruptant library (template based). • Nicolaou J. Am. Chem. Soc. 1998, 120, 10814. • Library screening:

screen for induction of tubulin polymerisation Me OR1 H Me and identity defined by cytotoxicity with ovarian cancer cells Hits spatial location O

H O R3 side chain crucial Me Me OR2 O o NMe spatially separated Me H Me N both N's library in IRORI important 'tea bags' O ketals tolerated H OH Me Me OMe esters prefered over O amides; reduction to alcohol not tolerated DeWitt’s quinolone library • Ciprofloxazin analogue library (template based). • DeWitt Tet. Lett. 1996, 37, 48115, and patent: WO 94/08711, 1994. • Library synthesis:

NMe2 DMAP toluene OO OO Et-ester 1) MeO OMe THF F 2) R -NH F 1 2 W O W OH W O 3) Δ

Wang resin F N F N N F R1

parallel synthesis R2-NH2 NMP, Δ OO F HO OO R TFA F N N N 2 W O H R R 1 N N N 2 H spatially separated R1 library in DIVERSOMER reaction tubes DeWitt’s quinolone library • Ciprofloxazin analogue library (template based). • DeWitt Tet. Lett. 1996, 37, 48115, and patent: WO 94/08711, 1994. • Library screening:

OO Screen for gyrase inhibition in solution identity defined by F Hits HO spatial location R N N N 2

R1 R3 OO spatially separated F library in DIVERSOMER HO reaction tubes N N N NH

ciprofloxazin Gallop’s mercaptoacyl proline library • Angiotensin Converting Enzyme (ACE) inhibitor library (template based). • Gallop J. Am. Chem. Soc. 1995, 117, 7029. • Library synthesis:

O 1) 20% piperidine, DMF O DCC-DMAP 2) ArCHO NHFmoc N Ar T OH T O T O 3) CH(OMe)3 R R1 1 TentaGel 4) Ac2O, DIPEA resin Z mix and split AgNO3, Et3N MeCN O R2 SH O O N Ar 1) AcS R2 COCl H HO N Ar R1 T O R Z 2) TFA 1 3) ethylenediamine Z 480 member pool of compounds Gallop’s mercaptoacyl proline library

• Angiotensin Converting Enzyme (ACE) inhibitor library (template based). • Gallop J. Am. Chem. Soc. 1995, 117, 7029. • Library screening:

O R2 SH Screened for HS O ACE inhibition N Ar O HO Me R1 HO2C N Z DECONVOLUTION VIA FOUR ITERATIONS OF SUB-LIBRARY RE-SYNTHESIS CO Me 480 member pool AND SCREENING 2 of compounds Ki ~ 160pM 3 x more potent than captopril Ellman’s benzodiazepine library • Benzodiazepine library (template based). • Ellman J. Org. Chem. 1997, 62, 2885. • Library synthesis:

1) R COCl, Pd(0) NHBpoc 1 NH2 2) TFA NH2 O Si SnMe3 Si aminomethyl-PS Me Me Me Me R1

1) R2

parallel synthesis FOC NHFmoc 2) 20% piperidine DMF 3) 5% AcOH, Δ R3 O N R 3 O 1) RLi H O R2 N N 2) R3-Br, DMF H N HF R R2 2 R1 Si N Si N Me Me Me Me R R1 spatially separated 1 solution library Ellman’s benzodiazepine library

• Benzodiaxepine library (template based). • Ellman J. Org. Chem. 1997, 62, 2885. • Library screening:

screening of this library was not reported because it pertained that some of the compounds in the library still R3 O contained silicon due to an anomolous cleavage N mechanism which was particularly troublesome when R1 was an electron withdrawing substituent R2 H N R3 O R1 N

R2 spatially separated F N solution library Si Me Me EWG Summary

• What is combinatorial chemistry? • The drug discovery process • Approaches to combinatorial library synthesis: – mix and split synthesis – parallel synthesis – encoded tagging • Library types: – oligomeric libraries – template based libraries • Combinatorial drug discovery.