Comenius University, Faculty of Natural Sciences,

Department of Organic Chemistry, Bratislava, Slovakia

Medicinal Chemistry
(Basic Principles)

24th November 2010 Andrej Boháč, Milan Remko

Molecular drugs
151 527 D
(1 356 AA)
348

1
(1 mg / 100 L)
Asn921
435.5 D
N
Leu1033
SO2Et

O
N N
O H
Cys917

IC50 = 22 nM (KDR)
J. Med. Chem. 48, 2005, 1610 - 1619.
http://www.pdb.org/
 Biogenic aminoacids COO

H3N H

• Unpolar (8) – (lipophilic) R

Alanine Valine Leucine

(Ala; A) (Val; V) (Leu; L)
Me- i
Pr- i
Bu-

Isoleucine Methionine Phenylalanine

(ILE; I) (Met; M) (Phe; F)
s
Bu- CH3S(CH2)2- PhCH2-
Tryptophan Proline
(Try; W) (Pro; P)
(indol-3-yl)CH2- -(CH2)3-

covalent bonds > ionic bonds > hydrogen bonds > van der Waals interactions
 Biogenic aminoacids COO

H3N H

• Polar (7) – (hydrophilic) R

Glycine Serine Threonine

(Gly; G) (Ser; S) (Thr; T)
H- HOCH2- HOCHCH3
│

Cysteine Tyrosine Asparagine

(Cys; C) (Tyr; Y) (Asn; N)
HSCH2- para-HOPh-CH2- NH2COCH2-
Glutamine
(Gln; Q)
NH2CO(CH2)2-

covalent bonds > ionic bonds > hydrogen bonds > van der Waals interactions
 Biogenic aminoacids COO

H3N H

• Ionized (5)– (hydrophilic) R

Lysine Arginine Histidine

(Lys; K) (Arg; R) (His; H)
H3N+(CH2)4- H2N(NH2+)CNH(CH2)3- H
N CH2

HN

Aspartic acid Glutamic acid

(Asp; D) (Glu; E)
-
OOCCH2- -
OOC(CH2)2-

covalent bonds > ionic bonds > hydrogen bonds > van der Waals interactions
Interaction analysis

Electrostatic interactions:
(5-10 kcal mol-1),
(C-C: 80 kcal /mol)

Hydrogen bonds:
vary in strenght (1-6 kcal mol-1)

Van der Waals interactions:

are very weak (0.5 -1 kcal mol-1 )
Hydrogen bonds
• The electron deficient hydrogen is called a hydrogen bond donor (HBD)
• The electron rich heteroatom is called a hydrogen bond acceptor (HBA)

- + -
- + - Drug Y H X
X H Y Target Target
Drug
HBD HBA HBA HBD

• Optimum orientation is where the X-H bond points directly to the lone
pair on Y such that the angle between X, H and Y is 180o

X H Y X H Y
Hybridised 1s Hybridised
orbital orbital orbital

HBD HBA
Hydrogen bonds
• strong hydrogen bond acceptors (HBA)
- carboxylate ion, phosphate ion, tertiary amine
RCOO-, RP(=O)(O-)2, R3N
• moderate hydrogen bond acceptors (HBA)
- carboxylic acid, amide oxygen, ketone, ester, ether, alcohol
RCOOH, RC(=O)NHR´, RC(=O)R´,RCOOR´, ROR´, ROH
• poor hydrogen bond acceptors (HBA)
- sulphur, fluorine, chlorine, aromatic ring, amide nitrogen,
aromatic amine
S, F, Cl, Ph, RC(=O)NHR´, ArNH-

• good hydrogen bond donors (HBD)

- quaternary ammonium ion R3HN+
9
Induced dipole interactions

• occur where the charge on one molecule induces a dipole on

another
• between a quaternary ammonium ion and an aromatic ring
(Lys, Arg)

d+
+
R N R3 d-

Binding site
 What is medicinal chemistry?
• definitely not a basic chemistry course for medical students

• interdisciplinary research dedicated to development of new drugs

(not only)

• drug development basic chronology:

selection of disease (cardiovascular, autoimmune, infectious, hereditary, mental, cancer …)

molecular mechanism of the pathology (medicine, molecular biology)
selection of a key biomolecule to influence
new active structure/compound identification: in Silico design, HTS,
of organic molecules possessing appropriate drug-like properties
(biologists, computer chemists)
organic synthesis (chemists)
biological or biophysical assays (biologists)
optimization of activity and other molecular properties
IP protection + clinical trials + up-scale synthesis + approval
 How many new drugs reach the
market yearly?
• DD is highly interdisciplinar, time and resources consuming process:
10 years / from 870 000 000 to 2000 000 000 USD /1 new drug

Adams C, Brantner V . Health Aff airs (Millwood) 2006 25 420–8.

• global production ca 24 innovative drugs (new chemical entity) / year

(2009: 26, 2008: 25, 2007: 18, 2006: 22, 2005: 26, 2004: 24, 2003: 26, 2002: 28, 2001:
23, 2000: 26)

• Many failures have been recorded in high stages of drug development, even in clinical
trials) Where is a problem?

Drug-likeness is often missing.

Computer aided rational drug design is preferred to predict activity and
bioavailability even before the compound is prepared.
 What kind of compounds are
drugs?
• Different inorganic, more likely organic compounds and
biomolecules (proteins, antibodies, siRNA…) that activates or inhibits
the function of a target with benefit to the patient

 activity
(target binding place stereoelectronic compatibility)

 low toxicity (selectivity, antitargets: CYP, hERG, Pglycoprot...)

 bioavailablity (physico-chemical and pharmacological properties ensuring

drug-likeness)
 Basic terms in
medicinal chemistry
• TARGET (biomacromolecule to interfere with)
• BINDING POCKET – ACTIVE SITE
(part of the target appropriate to bind a small ligand)
• PHARMACOPHORE(a part of a molecule that is recognized at a
receptor site and is responsible for that molecule's biological activity)
• LIGAND organic molecule possessing target affinity, that has to be
stereo-electronically compatible with binding pocket)
 ACTIVE is a compound detected by usually HTS
 HIT active compound identified in a screen with confirmed structure and
activity, that need to be developed into lead compound
 LEAD active compound with convenient properties: drug-likeness, solubility,
synthetic feasibility, patentability
 DRUG CANDIDATE high activity, good selectivity, in vivo efficiency
 DRUG after success in clinical trials approved by FDA, EMEA
• DRUG-LIKENESS complex properties including (ADME/Tox: Absorption
Distribution Metabolism Excretion / Toxicity)
Stereoelectronic compatibility
(requirements in space and charge distribution)
Free Biological DB - UNIPROT
(gene, AA sequences, biomolecular properties)

• http://www.uniprot.org/
 From were to get active
compounds?
Micro-organisms (bacteria, fungi)
Marine chemistry (corals, bacteria, fish etc)
Plant life (flowers, trees, bushes)
A) The Natural World Animal life (frogs, snakes, scorpions)
Biochemicals (neurotransmitters, hormones)
Pure natural products, bioextracts (e.g. plant, or
microbial) Ethnopharmacology (Chinese
traditional medicine...)

LMW synthetic compounds

B) The Synthetic World
(traditional, combinatorial synthesis, historical
corporate chemical collections, commercial
sources), SelOoptSideActivities (known drugs)

C) The Virtual World Computer aided drug design (CADD)

to call them „active compounds“ evaluation through biological screening is essential

 SOSA = Selective Optimization of
Side Activities
Exploitation of Existing Drugs as Leads for the Discovery of New
Drugs

all drugs act on more than one target (known and unknown), resulting in a
several side effects

advantage: drugs and many compounds that underwent clinical

development have an established safety profile. Many
of them can be therefore safely administered to humans.

Try to transform one of the side activities into the major effect
and strongly reduce their initial pharmacological activity.

C. G. Wermuth Drug Discov Today 2006, 11, 160; J. Med. Chem. 2004, 47, 1303.
DD - flowchart
 Case story from HTS to Leads
• at Schering begins with
– HTS of 700,000 compounds, followed by focused library design
and repeat HTS of about 2,000 single compounds. Sending 200
compounds forward for IC50 assaying. The result is 100 IC50
actives. Purity and structural evaluations bring the number
down to 50 validated actives.
– subsequent in vitro efficacy, selectivity, and toxicology studies
produce 15 clusters (or single compounds) of "qualified hits".
– qualified hits must be resynthesized to yield more compound
for subsequent in vitro and in vivo evaluations. These
evaluations conclude with the identification of 13 lead
structures (or clusters) after about a 10-month period, and
these structures are then moved forward into the lead
optimization process.
– (13 / 700 000 = 1 / 53 846 enrichment)
 What should compound fulfill to
become a drug?
• Biol. active, chemically stable compound possessing
appropriate:
 pharmacodynamic properties (activity, selectivity)
 pharmacokinetic properties (bioavailability: ADME/TOX)
 others (novelty, scale up synthesis...)
• > 30% of all drug failures can be attributed to poor physiochemical
properties: Log P (Log D), pKa, and solubility all have impacts on drug
absorption and diffusion in vivo

Chemical Space

Lead-like Drug-like
Chem Space: 1060 - 10200 Drugs
DB of 11 atoms C,N,O, F: 26 400 000 stable
compounds (110.9 M if stereoisomers included)
J.-L. Reimond J. Chem. Inf. Model. 47 2007 342.
 Cell Membrane – protects cell
compartment
• provides a hydrophobic barrier around the cell, preventing the
passage of water and polar molecules. Proteins (receptors, ion
channels and carrier proteins) are present, floating in the cell membrane.
• phospholipid bilayer, the hydrophobic tails interact with each other by
van der Waals interactions and are hidden from the aqueous media
• The polar head groups (phosphatidylcholine) interact with water at the
inner and outer surfaces of the membrane
 Bioavailability
• (in vitro) active compound, to perform as drug, has to reach its target
in the human body (in vivo)

• Drug-likeness is qualitative concept to estimate bioavailability from

the molecular structure before the substance is synthesized. The drug-
like molecule has to have:

 optimal MW and number of HBD, HBA (affecting solubility and absorption)

 optimal water and fat solubility logP (octanol / water) (intestinal lining,
aqueous blood, penetrate cellular membrane to rich inside the cell) The
distribution coefficient (Log D) is the correct descriptor for ionisable systems. log
D is pH dependent (e.g. at pH = 7.4 is the physiological pH of blood serum)

Lipinski's Rule of Five (Ro5)

 Lipinski Ro5
(all numbers are multiples of five, empiric rule)

for prediction of bioavailability (not activity!) to quickly

eliminate compounds that have poor physicochemical
properties for oral bioavailabilty
• an orally active drug has no more than one violation of the
following criteria:
 MW ≤ 500
 Lipophilicity (logP ≤ 5) octanol-water partition
coefficient (better log D ≤ 5 respecting the ionic states present at
physiological pH values)
 Sum of hydrogen bond donors ≤ 5 (NH,OH)
 Sum of hydrogen bond acceptors ≤ 10 (N,O)

C. A. Lipinski et al. Adv. Drug Del. Rev. 1997, 23, 3. (Ro5)

Greg M. Pearl et al., Mol. Pharmaceutics, 2007, 4, 556–560. (log D introduced)
 Additional drug-like parameters
 MW ≤ 500
 Lipophilicity (logP ≤ 5) octanol-water partition
coefficient
 Sum of hydrogen bond donors ≤ 5 (NH,OH)
 Sum of hydrogen bond acceptors ≤ 10 (N,O)

 PSA < 140 Å2(Molecular Polar Surface Area – sum of surfaces of polar
atoms (N,O...with H) that correlates with human intestinal and BBB
absorption) for good BBB penetration (PSA < 60 Å2)
 Number of rotatable bonds < 10 (flexibility factor)
(high NRB → many conformers)

Ertl, P. in Molecular Drug Properties, R. Mannhold (ed), Wiley-VCH , 2007, 111 – 126.
 Ro5 determined from 2D tructure
http://www.molinspiration.com

Ertl, P. et al., J. Med. Chem. 2000, 43: 3714-3717. (molecular property prediction toolkit )
Ro5 violations
 Absorption as f(PSA, LogP)

• AlogP http://www.vcclab.org/ (Virtual Computational Chemistry Laboratory)

(membrane transport connected with fat and water solubility)

• log pKa http://ibmlc2.chem.uga.edu/sparc/index.cfm

(influences binding Ki and also logP)

Intestinal and other absorption

• % ABS = 109 – 0.345 PSA (good when %ABS > 30 %)
(lower PSA, higher absorption)
Zao YH et al. Pharm Res 2002, 19, 1446-1457.

Brain Blood Barrier penetration

• LogBB = -0.0148 PSA + 0.152 CLogP + 0.139
(to estimate CNS penetration and possible side effects)
BB = C-brain / C-blood
CNS: logBB > -0.5 (side effects)
non CNS: logBB < -1
 Other considerations
• despite good druglikeness some compounds should
be avoided as drug candidates:

 If contain substructures with known reactive, toxic, mutagenic

or teratogenic properties (RCOX, (RCO)2O, Michael acceptors, epoxides,
-NO2, -NO, -N3, NH-NH, N=N...)

 bad metabolic parameters, e.g. fast metabolism can quickly

destroy the pharmacological activity of the compound (metabolic
half life, metabolic clearance should be determined)

 Inhibit antimetabolites (CYP, hERG, P-glycoprotein)

 Case study
(phy-chem properties optimization)

LogP round 2.5 allows to penetrate compounds to CNS → side effects

MeO
N
R
N N
H
Cardiotonicum
R: MeO- log P (2.59) FW: 255.27 (CNS side eff ects bright visions)
MeSO- lop P (1.17) FW: 287.34 (Sulmazol)

2-Aryl-3H-imidazo[4,5-b]pyridine
 pKa universal measure of acidity
and also basicity

Ionised form

pKa = -log {[H+].[B] / [HB+]} = pH - log [B]/[HB+] = pH – log 1 = pH

pKa = pH if concentration of base and its protonated form is equal (pH at

which the base is on 50 % ionized)

the higher pKa the stronger base

or the weaker acid
 at more basic solution

In more acidic environment

pKa = -log {[H+].[B] / [HB+]} = pH - log [B]/[HB+]

-----------------------------------
5.1 = 4.1 - log [B]/[HB+] two equiations wit two variables
[B] + [HB+] = 100%
-----------------------------------
9% [B] and 91% [HB+]
 Rational drug design

frequently relies on computer modelling techniques

(computer-aided (assisted) drug design CADD)
AIM:
to discover, enhance, or study biologically active molecules that will
bind to a selected target and if so how strongly before a compound
is synthesized

to estimate drug-like properties and use them for elimination of undesirable
structures

it still takes several iterations of design, synthesis, and testing before an optimal
molecule is discovered
 Rational methods in DD
• Structure-based drug design SBDD
requires 3D information about biomolecules (X-ray crystallography or
NMR spectroscopy, PDB database) http://www.rcsb.org/pdb/

• Ligand-based drug design LBDD

• Fragment-based drug design FBDD

 Ligand Based Drug Design
LBDD (indirect DD)
• relies on knowledge of known molecules that bind to the biological
target (known: structure and bioactivity IC50)

 These molecules (ligands) may be used to derive a pharmacophore model

which defines the minimum necessary structural characteristics a molecule
must possess in order to bind to the target.
Virtual screening (based on pharmacophore models; high-throughput
docking) including drug property filtering (Zinc 13 000 000 / Lipinski)
http://zinc.docking.org/

 Alternatively, a quantitative structure-activity relationship (QSAR) in which a

correlation between calculated properties of molecules and their
experimentally determined biological activity may be derived. These may be
used to predict the activity of new analogues.
 Structure Based Drug Design
SBDD (direct DD)
• relies on knowledge of the 3D structure of the biological
target obtained through X-ray crystallography or NMR
spectroscopy (http://www.rcsb.org/pdb/)
• SBDD be divided roughly into two categories
 “finding” ligands for a given receptor (database searching). A large
number of potential ligand molecules are screened to find those fitting
the binding pocket of the receptor. (this method is sometimes referred
as ligand-based drug design) It saves synthetic effort to obtain new
active compounds.
 “building” ligands (receptor-based drug design). In this case, ligand
molecules are built up within the constraints of the binding pocket by
assembling small pieces (atoms, fragments) in a stepwise manner. The
key advantage is that novel structures, not contained in any database,
can be suggested.
 Virtual SBDD screening
(„finding ligands“)

• 500 potentially actives / 1 500 000 mol. in library

1 : 3000
 SBDD „finding ligands“ method
• relies on 3D target structure with docked ligand (PDB complex)
• active side identification
 ligand and protein atoms need to be classified and their atomic properties
• hydrophobic atom: all carbons
• H-bond donor: OH,NH
• H-bond acceptor: O,N
• Polar atom: O,N,S,P,X,M,C-HETATOM

The space inside the ligand binding region would be

studied with virtual probe atoms of the four types above
to determine what kind of chemical fragments can be put
into their corresponding spots in the ligand binding region
of the receptor. (latice, grids, intersections)
 Fragment based drug discovery
FBDD
• Screening of small (MW < 300), low potency fragments
(epitops, “seed templates”), which are subsequently
developed into higher potency structures

 we need a fragments database to choose fragments from

 although the diversity of organic structures is infinite, the
number of basic fragments is rather limited
 seed is put into the binding pocket, and add other fragments
one by one
 new molecules can be regarded as combinations of two or
more individual binding epitopes
 FBDD: Ligand efficiency / Ro3
• Ligand Efficiency LE
LE = - ΔG/HAC ≈ - RTln(IC50)/HAC
(HAC = Number of heavy atoms)

fragments typically exhibit higher ligand efficiency than higher MW HTS not ideal
hits 

I. D. Kuntz et al. Proc. Natl. Acad. Sci. USA 1999, 96, 9997.
 
• The “rule of three” for appropriate fragment design:
– MW < 300
– Number of H-bond donors ≤ 3 (NH,OH)
– Number of H-bond acceptors ≤ 3 (N,O)
– clogP ≤ 3
 Case studies - PKI
Protein kinases (tyrosine, serin-threonine and histidine
kinases) phosphorylate specific amino acids in protein
substrates. Growth factors trigger the start of signalling
cascade that control transcription of specific genes in
DNA leading cell growth and cell division. In many
cancers excess of growth factor or PK receptor has been
observed. Therefore PK inhibitors would be useful as
anticancer agents. All PK use ATP as the
phosphorylation agent. ATP is bound quite loosely, there
are areas which remain unoccupied. There are also
differences in AA within PKs in this binding region. Therefore
it is possible to design selective ATP competing inhibitors.
 EGFR inhibitor gefinitib (IRRESA)
(approved for refractory lung cancer, AstraZeneca)

EGFR (ErbB, HER1) tyrosine kinase receptor: abnormal or over-expressed in the breast, lung,
brain, prostate, gastrointestinal tract, ovaries cancer. EGFR is a receptor for EGF proteins (1986
Nobel Prize). Upon activation by its growth factor, EGFR forms active homodimer possesing
intracellular TK activity that initiate several signal transduction cascades leading to DNA
synthesis and cell proliferation. Gefitinib inhibits EGFR by binding to the ATP-binding site. Thus
receptor and therefore also malignant cells are inhibited.
4-anilinoquinazoline SAR optimized
Lead I, good in vitro activity, in vivo
hampered by rapid metabolism
caused by cytochrome P450
enzymes
main metabolic products I, II both met. routes
blocked
Cl- similar in size and lipophilicity as Me- group
F- almost the same size as H- (no steric effect)
both groups are resistant to oxidation, better in vivo
activity, pharmacokinetic properties improved by
morpholino group, because of basic nitrogen that
enhanced water solubility
 Substrate binding - Induced fit
(different conformers of the same protein)
• Binding of ligand alters the shape of the protein active side
to maximise intermolecular interactions. This will result 3D
changes in protein binding site: induced fit

Phe
Phe
S S
H
O
H
O Ser
Ser CO2
CO2 Induced
Asp fit Asp

Intermolecular bonds not optimum Intermolecular bond lengths optimised

length for maximum bonding (conformational changes in AA residues
and in backbone observed in the protein)
Example: Binding of pyruvic acid in Lactate Dehydrogenase LDH

C O
H3C C
H3N
O
 O

C O
H3C C
H3N
O
 Prove of VEGFR2 mechanism –
impact of induced fit
(different induced fit derived conformers of the same protein accommodate differently the same
set of 16 compounds, induced fit is a complication for CADD)

in this docking system, the best nM

inhibitors should rich the relative
level of interaction energy ca 50
kcal/mol. 3CJF is the one of the two
receptor conformers that almost
reached this level for the best of our
16 compounds. This allow us to select
the promising compound DMI-76H
(both enantiomers -46.96 and -46.74)
that has the capability to inhibit
VEGFR-2 signalling pathway. Next
values (44, 41, ... kcal/mol).
Thank you for your attention