FRAGMENT- BASED
DRUG DESIGN
Yemane Mengistu
Michigan State University
January 30, 2008
Annual Research and Development Expense
40
35
120
25
20
80
15
10
40
5
2004
2000
1996
1992
1988
1984
0
1980
Expense ($ Billions)
30
R & D Investment
Source : Pharma, FDA, Lehman :
NME ( New Medical Entities )
Annual NME Approvals
160
Drug Discovery Process
H
N
X
R1
OH
N
R2
NH
HO
N
R2
HO
Combinatorial Chemistry
Natural products
R1
N
R3
N
R4
R1
R 1 - R2 : diverse substituent
X= O, S
High-throughput
Screening (HTS)
Target molecule
eg. Enzymes
Toxicity
Lead &
Drug
optimization
Clinical trials
Creating a Library
A Natural Product
O
HO
HO
NH
O
HO
N
NH O
HO
OH
O
O
O
H
N
OH
R
O
H
O
H
OH
OH
OH
Tunicamycins
Thymidinyl nucleoside
Sun, D., Lee, R.E. Tetrahedron Lett. 2005, 46, 8497-8501
Creating a Library Using Ugi Chemistry
O
R
N
R1
R3
HN
O
N
O R
2
NH
O
OH
Thymidinyl nucleoside library
O
Creating a Library Using Ugi Chemistry
O
O
R1
R3
H
N
O
OH
R2 NH 2
R3
H
R 4 NC
R1
N
R2
Walters, W.P., Stahl, M.T., Murcko, M.A. Drug Discovery Today 1998, 3,160-178
O
R4
Creating a Thymidinyl Ugi Library
Sun, D., Lee, R.E. Tetrahedron Lett. 2005, 46, 8497-8501
Thymidinyl Ugi Library
CHO
CHO
NC
I
CHO
Cl
O
Br
COOH
O
HO 2C
CHO
CHO
HO
CHO
CHO
S
H3 C
HO2 C
CO2 H
Cl
O
CO2H
CO2H
COOH
HO2 C
CO2 H
S
CO 2H
F
CO 2H MeO
CHO
O
MeO
NC
O
NC
R
O
NH
N
NC
R3
OO
S NC
N
R1
NC
N
N
N
CO2 Me
HN
O
N
O R
2
O
O
OH
NC
Sun, D.; Lee, R.E. Tetrahedron Lett. 2005, 46, 8497-8501
CO2 H
High-throughput Screening (HTS)



A process of assaying a large number of compounds against biological
targets.
Up to 100,000 compounds can be analyzed in a day.
Robots can usually prepare and analyze many plates simultaneously.
http://www.metprog.org.uk/images/manufacturing_icon.jpg
What types of compounds become leads from an
HTS?
Lipinski’s Rules (Pfizer)
HTS Drug like (Rule of 5)
Molecular weight ≤
500
# Hydrogen Bond acceptors ≤
Lead-likeness
Molecular weight ~300
10
Sum of N and O
Fewer Hydrogen Bond Acceptors
Sum of N and O
# Hydrogen Bond Donors
≤
5
Sum of NH and OH
Lipophilicity
ClogP<5
Lipophilicity
ClogP<3
Low to high affinity for the target receptor
Drug like behavior
Lead like behavior
Congreve, M. et al. Drug Disco. Today.2003,8, 876-877
Lipinski, C.A. et al. Adv.Drug Deli.Rev.1997,23,3-25.
600
Drug
Candidates
Relative Molecular Mass
500
Drugs
400
300
200
100
0
1mM
Potency
1µM
10 nM
A Typical Drug Discovery Cascade
HTS
1,000,000
Hits actives
Lead series
Drug candidates
Drug
2000
Increased risk of failure
HTS Hits
1200
50-200
10
1
Opera ,T.I. J Comput. Aided Mol Des 2002, 16, 323-334
HTS
GlaxoSmithKline’s HTS Scoreboard
1996
1999
2003
Compounds screened
100,000
430,000
615,000 1,050,000
Average lead potency
3,000 nM
400 nM
10 nM
10 nM
Screen success
20%
50%
58%
65%
Leads per target
1.0
1.7
1.9
2.0
Chemical engineering news, 2004, 82 ,23-32
2004
Fragnomics: Fragment Based Drug Design

An approach that uses small and relatively simple molecules to make
lead compounds
Potential
Medicinal
compounds
Look for affinity
Fragments
Target
~Enzymes, etc
Merge and
Expand
Lead
MW of Average HTS Hits and Fragments
600
Drug
Candidates
500
Drugs
Relative Molecular Mass
400
300
200
Fragments
100
Rees D.C,0Congreve
M, Murray C.W, Carr Potency
R .Nat. Rev.1µM
Drug Discov. 2004, 3:660
1mM
10 nM
Fragnomics: Fragment Based Drug Design
Conventional HTS approach
Fragment based drug design
H2 N
NH 2
H2 N
N
NH 2
N3
H2 N
N
N3
HN
H N
2
N
N N
N
N N
(CH2 )6 C
N

CH
(CH2 )6 C
Kd = 10 -100 nM
Kd = 10 -100 nM
N
N
NH 2
CH
HN
HN
N
HN
N

N
NH 2
Kd = 10 -100 M
Kd = 10 -100 M
Kd>100µM
Erlason D.A, McDowell RS, O’Brien T. J Med Chem. 2004, 47:3463-82
Lewis, W.G.et al Angew. Chem. Int. Ed. Engl. 2002, 41,1053-1057
Fragnomics: Fragment Based Drug Design
Conventional HTS approach
Fragment based drug design
O
HN
O
O
NH
O
HN
NH
HN
N
H2 N
O
N
H
HN
O
O
O
Kd = 6.5 
HO
O
Kd > 100M
Swayze, E.E, et al.J.Med.Chem. 2003, 46, 4232-4235
What Qualifies Compounds to be Fragments?



Molecular Weight Mr ~300 Da
H-bond donors (HBD) <3
H-bond acceptors (HBA) <3
O
O
HN
O
NH
O
HN
NH
HN
N
H2 N
O
N
H
HN
O
O
O
HO
O
Congreve, M. et al. Drug Discov.Today 2003,8, 876-877
What Qualifies Compounds to be Fragments?
Molecular Weight Mr ~300 Da
H-bond donors (HBD) <3
H-bond acceptors (HBA) <3
Cl
NH 2
NH 2
N
O
N
NH
O
F
O
Fragment
N
IC50 = 1.3mM
O
Mr =200
HBD= 2
HBA=3
Lead for protein kinase inhibitor
IC50 = 65 nM
Mr =456
Congreve, M. et al. Drug Discov.Today 2003,8, 876-877
What Qualifies Compounds to be Fragments?

Clog P <3


A measure of Lipophilicity of a compound
Polar Surface Area (PSA) <60

A measure of permeability through the cell membrane.
Cl
NH 2
N
NH 2
N
NH
O
F
O
O
N
Fragment
O
Lead for protein kinase inhibitor
Clog P=1.92
PSA=48.14
Clog P=3.07
PSA= 77.6
Congreve, M. et al. Drug Discov.Today 2003,8, 876-877
Ertl, P.et. al. J.Med.Chem. 2000, 43,3714-3717
Some Common Drug-Based Fragments
Ring system from drug
Heterocyclic system
H
N
H
N
N
N
N
O
OH
H
N
N
NH
N
N
N
H
N
NH
O
O
S NH 2
O
O
N
N
Side chains
N
N
NH
Hartshorn, M.J., Murray, C.W.et.al. J. Med. Chem. 2005, 48, 403-413
NH2
OH
Conventional HTS vs. Fragonomics Based on Central Scaffold
R2
CH3
R2
CH3
NH
NH
NH
NH
N
N
N
N
N
R1
R1
HN
CH3
O
O
100 X R1
100 X R2, and
100 X R3 yields
100
A library with 1 million
compounds
N
H3C
HN
R3
N
N
H3C
HN
CH3
HN
O
100
R3
O
100
Variations yield a library of only 300 compounds
Carr, R, and Hann, M. Modern Drug Discov. 2002, 45-48
Conventional (HTS) Drug Design
Erlanson, D.A, Hansen, K.S. Curr Opin Chem Biol. 2004, 8,399-406.
Conventional (HTS) and Fragment Based Drug Design
Erlanson, D.A, Hansen, K.S. Curr Opin Chem Biol. 2004, 8,399-406.
Fragment Based Drug Design
1. Prepare set of potential binding elements
with a common chemical linkage group
Maly, D.J., Choong, D.J., and Ellman, J.A. Proc.Natl.Acad.Sci.USA. 2000,97,2419-2424
Fragment Based Drug Design
1. Prepare set of potential binding elements
with a common chemical linkage group
2.Screen Potential binding elements
Maly, D.J., Choong, D.J., and Ellman, J.A. Proc.Natl.Acad.Sci.USA. 2000,97,2419-2424
Fragment Based Drug Design
1. Prepare set of potential binding elements
with a common chemical linkage group
2.Screen Potential binding elements
3. Prepare library of all possible combinations
of linked binding elements.
Maly, D.J., Choong, D.J., and Ellman, J.A. Proc.Natl.Acad.Sci.USA. 2000,97,2419-2424
Fragment Based Drug Design
1. Prepare set of potential binding elements
with a common chemical linkage group
2.Screen Potential binding elements
3. Prepare library of all possible combinations
of linked binding elements.
4.Screen library of linked
binding elements
Maly, D.J., Choong, D.J., and Ellman, J.A. Proc.Natl.Acad.Sci.USA. 2000,97,2419-2424
Application of Fragment based drug design
1. Protein kinase inhibitors
Tyrosine kinase (Src) activate numerous signaling pathways within cells, leading
to cell proliferation, differentiation , migration and metabolic changes
Src kinases have been implicated in the pathology of tumors, osteoclast-mediated
Bone resorption and disorders associated with T-cell proliferation
Scapin,G. Drug Discovery today. 2002, 7,2002
Maly, D.J., Choong, D.J., and Ellman, J.A. Proc.Natl.Acad.Sci.USA. 2000,97,2419-2424
Library for the Protein kinase inhibitor
N OCH3
NOCH3
N OCH3
N OCH 3
OH
N
H
O
O
3
2
OCH3
N
1
4
N OCH3
N OCH3
OH
O
N OCH3
O
O S N
OH
OH
N
6
5
N OCH 3
7
N
Cl
OCH 3
Cl
8
N OCH3
N OCH3
OH
S
9
Cl
N
10
11
O
12
D.J.Maly, I.C.Choong and J.A, Ellman,Proc.Natl.Acad.Sci.USA,2000,97,2419-2424
Library for the Protein kinase inhibitor
N OCH3
NOCH3
N OCH3
N OCH 3
OH
N
H
O
O
3
2
OCH3
N
1
N OCH3
N OCH3
N OCH3
Ki =41µM
OH
O
4
O S N
O
OH
OH
N
6
5
N OCH 3
7
N
Cl
N OCH3
OCH 3
Cl
8
N OCH3
OH
S
9
Cl
N
10
11
O
12
D.J.Maly, I.C.Choong and J.A, Ellman,Proc.Natl.Acad.Sci.USA,2000,97,2419-2424
Library of protein kinase inhibitor
N OCH3
N OCH 3
N OCH 3
N OCH 3
O
F
F
F
F
F
13
OH
O
N
N
Br
Br
14
Ki= 40 M
15
16
N OCH 3
N OCH 3
O
HO
N OCH 3
N OCH 3
O
O
S
HO
O
18
OCH 3
OCH 3
N
N
17
N OCH 3
S
OCH 3
N
20
19
C 6H 5H 2 CO
H3 CO
21
22
23
24
D.J.Maly, I.C.Choong and J.A, Ellman,Proc.Natl.Acad.Sci.USA,2000,97,2419-2424
Fragment-Based Design : Protein Kinase Inhibition
N OCH3
H3CO
N
N
O
O
N
N
OH
HO
N
OH
OH
7
16
Compound
IC50µM
C-Src
Fyn
Lyn
Lck
[7]
41± 5
>1000
>1000
>1000
[16]
40± 16
64± 50
400± 170
>500
[7,16]
0.064± 0.038
5.0± 2.4
13 ± 2.4
>250
D.J.Maly, I.C.Choong and J.A, Ellman,Proc.Natl.Acad.Sci.USA,2000,97,2419-2424
Correlation of linker structure with IC50 values for c-Src Inhibition
Entry
Compound
Linker
linker
O
O
N
N
1
2
HO
N
OH
3
0.0640.038
7,16, n=2
(CH2 )n
7,16, n=3
(CH2 )n
1.10.2
(CH2 )n
6.53.0
7,16, n=4
(CH2 )n
4
c-Src IC50,  M
7,16, n=5
6.50.8
(CH2 )n
6
7
8
7,16, n=6
5.32.1
7,16, cis
1.20.6
7,16, trans (1R,2R)
0.620.02
Application of Fragment Based Drug Design
2. Matrix Metalloproteinase inhibitors

Matrix Metalloproteinases is a family of zinc-dependent endopeptidases.

Implicated in a variety of diseases including arthritis and tumor metastasis.

Conventional high-throughput screening failed to get non-peptide inhibitor.
Haiduk, P.J. et al. JACS. 1997,119, 5818-5827
Application of Fragment Based Drug Design
2. Matrix Metalloproteinase Inhibitors
HO
HO
H
N
O
Kd=17 mM
N
Kd=0.2 mM
HO HN
O
O
IC50=57 nM
N
Haiduk, P.J. et al. JACS. 1997,119, 5818-5827
Puerta, D.T, Lewis J.A. JACS. 2004, 126, 8389
Application of Fragment Based Drug Design
2 Matrix Metalloproteinase Inhibitors
O
O
O
HO
O
O S
O
O
F
F
IC 50 = 0.5 nM
F
ABT-518 , a drug candidate in clinical trial by Abbot Pharmaceutical Company
Wada, C.K, et al. J.Med.Chem. 2002, 45, 219-232
Application of Fragment Based Drug Design
3. Thymidylate synthase (TS)



Is the sole source for production of thymidine monophosphate (dTMP).
dTMP plays a central role in DNA synthesis .
It has been a target for dividing cancer cells.
Banerjee D, Mayer-Kuckuk P, Capiaux G, et al. Biochim. Biophys. Acta, 2002, 1587,:164-73.
Application of Fragment based Drug Design
Site Directed Ligand Discovery for TS
TS
Screen against library of
Disulfide-containing small
Molecules
O
R
N
H
S
S
NH2
Erlanson, D.A, Braisted, A.C .Proc.Natl.Acad.Sci.USA. 2000, 97 ,9367–9372
Preparation of Disulfide-Containing Library Members
O
O
O
N O
COOH
O
N O
1,3-disopropylcarbodiimide (DIC)
DMF
OH
O
O
O
NH
THF
S S
H2N
HN
O
S
HN
TFA
S
NH2
DCM
O
O
S
O
S
NH
Parlow, J.J. & Normansell, J.E.Mol.Diversity 1995,1, 266-269
O
Synthesis of Sulfonyl Libraries
HN
S
O
O
O
HN
S
R S Cl
O
S
O
NH2
O
S
O
NH S R
O
Poly vinyl pyridine
N
filteration
HN
H2N
O
S
S
O
R
NH S
O
TFA
S
O
S
DCM
Erlanson, D.A, Braisted, A.C .Proc.Natl.Acad.Sci.USA. 2000, 97 ,9367–9372
O
NH S R
O
Thymidylate Synthase Inhibitor
Selected
Non selected
F
O
linker
O
O
N
O
S
N
S
N
O
S
linker
N
O
linker
linker
O
O
Cl
O
linker
HN
O
S
NH
linker
S
O
O
S
O
O
S
N
N
linker
linker
O
linker =
N
H
Erlanson, D.A, Braisted, A.C .Proc.Natl.Acad.Sci.USA. 2000, 97 ,9367–9372
S
S
NH 2
R
O
O
N
CO2H
O
S
NH
N S
O
O
O
CO2H
R'
HO
Compund
R
Ki
CH3
1
O
1100
CO2 H
2
O
O
6
CO2H
N
H
O
O
CO2NH2
61
8
N
H
CO2H
373
N
H
0.33
12
N
H
9
O
5
CO2H
O
CO2NH2
4
noncompetitive
H
35
CO2H
3
37
NH2
7
N
H
Ki
R'
Compound
O
N
H
CO2H
246
10
N
H
N
Erlanson, D.A, Braisted, A.C .Proc.Natl.Acad.Sci.USA. 2000, 97, 9367–9372
> 100 only
Thymidylate Synthase Inhibitor
HOOC
H
HOOC
COOH
HN
O
HOOC
H
HN
H
COOH
HN
O
H2N
S
O
N
CO2 H
N
HN
O
O
S
N
O
O
N
O
COOH
O
N
H
Methylenetetrahydrofolate
Km= 14 M
N-tosyl-D-proline
ki= 1.1 0.25 mM
N
CO2H
ki= 24  7 M
O
S
O
N
COOH
NH
ki= 330 40 nM
Erlanson, D.A, Braisted, A.C .Proc.Natl.Acad.Sci.USA. 2000, 97 ,9367–9372
Application of Fragment Based Drug Design
4. Cysteine Aspartyl Protease-3 ( Caspase-3)

Mediator of apoptosis ( programmed cell death).

They are responsible for the cleavage of the key cellular proteins such
as cytoskeleton proteins.

Reducing the apoptotic response in diseases with dysregulated
apoptosis such as myocardial infarction, stroke, traumatic brain,
Alzheimer’s disease, and Parkinson diseases could benefit .
Hotchkiss, R.S. et al. Nat. Immunol. 2000, 1 , 496-501
Tethering with Extenders-dynamically Assembling Fragments
Caspase-3 using extender A
Caspase-3 using extender B
Erlanson, D.A, Hansen, K.S. Curr Opin Chem Biol. 2004, 8,399-406.
Assembly of the Extender with Enzyme and with Fragment Library
Cl
O
Cl
O
O
H
N
O
Cl
O
O
Cl
HOOC
H
N
O
S
O
Extender B
Caspase-3
O
O
H
N
Casp
H
N
S
SH
O
O
HOOC
O
HOOC
Caspase-3
S
S
O
O
Extender A
Casp
S
HOOC
SH
S
O
Assembly of the Extender with Enzyme and with Fragment Library
O
O
Casp
H
N
S
Casp
H
N
S
SH
S
SH
O
O
O
HOOC
HOOC
FRAGMENTS
O

H 2N
S
O
S
S
COOH
N
H
A

OH
O
H 2N
O
S
S
S

COOH
N
H

B

O
O
H 2N
S
S
c
S
s
SO 2

N
H
Erlanson, D.A, Lam, J.W, Wesmann ,C. Nat. Biotechnol. 2003, 21, 308-314
Assembling the inhibitor
O
Caspase
S
O
O
H
N
S
S
NH
COOH
S
O
OH
HOOC
Extender A + Fragment A
Ki( M)
Compound
O
1
H
O
H
N
O
S
NH
COOH
2.8
O
OH
HOOC
O
O
H
N
H
NH
2
O
15.3
S
COOH
O
HOOC
O
3
H
O
H
N
NH
O
>100
S
O
HOOC
O
O
4
H
NH
H
N
O
S
COOH
0.2
OH
O
HOOC
Erlanson, D.A, Lam, J.W, Wesmann ,C. Nat. Biotechnol. 2003, 21, 308-314
Assembling the inhibitor
O
Caspase
S
O
H
N
O
O
S
S
S
S
NH
S
O
O
HOOC
Extender B + Fragment C
Ki( M)
Compound
O
O
5
H
O
O
H
N
S
S
O
O
S
0.33
S
NH
O
O
HOOC
O
O
6
H
H
N
O
HOOC
O
O
S
S
O
O
S
1.8
NH
O
Erlanson, D.A, Lam, J.W, Wesmann ,C. Nat. Biotechnol. 2003, 21, 308-314
Caspase-3
Superimposition of Inhibitor 1(Gray) and compound 4 (salmon) with Capsase 3
Erlanson, D.A, Lam, J.W, Wesmann ,C. Nat.Biotechnol.2003 , 21,308-314
Summary

The use of fragment based drug design accompanied by different
means of detection could increase the chance of finding new medical
entities.

Site directed ligand discovery and fragment based lead discovery are
still in their infancy, but the success of these emerging approaches could
success.

No single technology will suffice, and the combination of HTS, site
directed , and fragment-based lead discovery will likely become
increasingly important.
Acknowledgements





Prof. Kevin D. Walker
Prof. Babak Borhan
Prof. Bill Wulff
Prof. Bob Hausinger
Dr. Philip J. Hajduk , Abbott Laboratories

Lab members:,Mark, Irosha, Washington,
Danielle, Behnaz

Friends: Khassay, Mercy, Rahman, Anil,
Munmun, Luis