Synthesis of novel PPARα/γ dual agonists as potential drugs for the treatment of the metabolic syndrome and diabetes type II designed using a new de novo design program PROTOBUILD

View Article Online / Journal Homepage / Table of Contents for this issue

Organic &

Biomolecular

Chemistry

Dynamic Article Links

Cite this: Org. Biomol. Chem., 2011, 9, 1169

www.rsc.org/obc

PAPER

Synthesis of novel PPARa/c dual agonists as potential drugs for the

treatment of the metabolic syndrome and diabetes type II designed using a

new de novo design program ^PROTOBUILD†

Yushma Bhurruth-Alcor,^aTherese Røst,^bMichael R. Jorgensen,^aChristos Kontogiorgis,^aJon Skorve,^b

Robert G. Cooper,^dJoseph M. Sheridan,^dWilliam D. O. Hamilton,^dJonathan R. Heal,^dRolf K. Berge^b,cand

Andrew D. Miller*^a

Received 19th May 2010, Accepted 7th October 2010

DOI: 10.1039/c0ob00146e

Peroxisome proliferator activated receptors (PPARs) have been shown to have critical roles in fatty acid

oxidation, triglyceride synthesis, and lipid metabolism - making them an important target in drug

discovery. Here we describe the in silico design, synthesis and in vitro characterisation of a novel series

of 2,5-disubstituted indoles as PPARa/g dual agonists. PPAR activation assays are performed with

known agonists diazabenzene (WY14.643), aminopyridine (BRL49653) and bisaryl (L165.041), as

positive controls. All the indole compounds synthesized are found to be active PPARa and PPARg

agonists, with particular efﬁcacy from those with 2-naphthylmethyl substitution. This is a useful

demonstration of a new de novo design methodology implemented by the PROTOBUILD program and its

ability to rapidly produce novel modulators for a well characterized drug target.

ligands of the PPARs include a wide variety of saturated or

Introduction

unsaturated fatty acids and eicosanoid derivatives.¹⁷Synthetic

Metabolic syndrome is a complex set of disorders encompassing:

obesity, dyslipidaemia, insulin resistance, blood pressure elevation

PPAR agonists are used in clinical practice, for example ﬁbrates

(PPARa agonists) are used to improve dyslipidemia during the

and coronary heart disease.^1,2People with metabolic syndrome

have a ﬁve-fold greater risk of developing diabetes type II.³

Peroxisome proliferator-activated receptors (PPARs) have been

established as a primary drug discovery target for the treatment of

metabolic diseases because of their critical role in the regulation

of lipid metabolism and fat cell differentiation.^4–8These nuclear

receptors are transcription factors and bind to speciﬁc peroxisome

proliferator response elements (PPREs) within promoters.^9–11The

three identiﬁed subtypes (PPARa, PPARd and PPARg) are all

key regulators of lipid metabolism and act by controlling the

expression of several genes involved in peroxisomal and mito-

chondrial fatty acid oxidation, fatty acid uptake and transport,

as well as differentiation of adipocytes.^12–14The main roles of

the different PPARs are summarised in Table 1.^14,15Natural

treatment of atherosclerosis.²⁴The thiazolidinediones are a well-

founded class of anti-diabetic drugs that act as PPARg agonists,

thereby improving insulin sensitivity and decreasing the hepatic

glucose output.^25,26Unfortunately the use of thiazolidinediones

is known to have attendant side effects such as weight gain,

oedema, and anaemia with possible liver dysfunction.^18,27,28There

have thus been suggestions that dual or even pan PPAR agonists

(targeting all three subtypes) should be developed, as these may

exert complementary and synergistic actions in improving lipid

homeostasis and insulin sensitivity, while hopefully reducing

attendant side-effects.^29,30

Here we describe the computer-aided design of a set of novel

indole based compounds as dual PPARa/g agonists, making

use of a new de novo fragment-based design software program

known as PROTOBUILD. PROTOBUILD software seeks to construct

new, candidate ligands appropriate to a selected protein receptor

ligand-binding site by searching chemical space within the context

of the binding site in order to deﬁne suitable pharmacophore

structures against which series of candidate ligands can be ranked

in order of their predicted ligand-binding site afﬁnity. Although

this general concept has been implemented already in several

other design programs.^31–34PROTOBUILD differs by making use of a

genetic algorithm (GA) that controls both fragment addition and

deletions based on dynamic selection pressures. Such an approach

allows for a more thorough search of chemical space in particular

^aImperial College Genetic Therapies Centre, Department of Chemistry,

Flowers Building, Armstrong Road, Imperial College London, London, UK,

SW7 2AZ. E-mail: a.miller07@btinternet.com; Tel: +44 787 963 5513

^bInstitute of Medicine, University of Bergen, N-5021, Bergen, Norway

^cDepartment of Heart Disease, Haukeland University Hospital, N-5021,

Bergen, Norway

^dProsarix Ltd, Newton Hall Town Street, Newton, Cambridge, UK, CB22

7ZE

† Electronic supplementary information (ESI) available: Experimental

details and X-ray crystallography. CCDC reference numbers 667594 (22)

and 667595 (23). For ESI and crystallographic data in CIF or other

electronic format see DOI: 10.1039/c0ob00146e

This journal is

The Royal Society of Chemistry 2011

Org. Biomol. Chem., 2011, 9, 1169–1188 | 1169

©

View Article Online

Table 1 Summary of locations and roles of PPARs a, g and d

PPAR

Site of expression¹⁶

Role

a

g

Catabolically active tissues such as liver, heart, kidneys and skeletal muscle

Broad range of metabolically active tissues such as skeletal muscle, kidneys,

intestine and adipose tissue

Oxidation of fatty acids and lipoprotein metabolism¹⁷

Fat cell differentiation and triglyceride synthesis^15,18–21

d

Wide range of tissues and cells, high expression in the brain

Regulation of lipid metabolism and cholesterol efﬂux^22,23

by helping to avoid local minima that otherwise develop due to the

early inclusion of heavier, hydrophobic fragments at the expense

of smaller fragments. The PROTOBUILD approach is an obvious

alternative to the Metropolis criterion for Monte Carlo that is

implemented in other de novo design programs.^31,32Moreover

the use of the GA provides a convenient route for searching

different ligand-binding sites with variable afﬁnities, owing to

the fact that each binding site can compete for emerging ligand

populations via the GA. PROTOBUILD also contains several other

features that differentiate it from other previously reported design

programs. Speciﬁcally, the program was developed to exploit

a new scoring function, PROTOSCORE, that includes improved

terms for the estimation of entropy plus terms for various non-

bonded interactions not otherwise represented in design programs.

Moreover, PROTOBUILD was designed to take into account protein

receptor binding site ﬂexibilities thereby providing an even more

realistic view of protein receptor-ligand molecular recognition and

binding events. Yet other features involve:

∑ The use of detailed fragment connection data extracted from

known compounds in order to control how the same fragments

are connected together within PROTOBUILD. This approach was

introduced in order to ensure that new, candidate ligands should

be readily synthesized when required.

∑ The use of pharmacophore constraints and/or the SMILES

or SMARTS pattern language in order to direct the generation of

new ligands.

all possible low energy combinations of side chains in the

designated ligand-binding site. A diverse population of low energy

combinations is then subjected to SiteAnalysis routines that serve

as inputs to PROTOBUILD. By separating protein receptor ﬂexibility

calculations from the ligand generation algorithm, multiple recep-

tor states may thereby be included in the GA that controls the

growth of new ligands. “Cross-talk” between populations of new,

candidate ligands and ligand-binding sites associated with these

multiple receptor states then becomes an important contributor

towards the growth of new ligands. Clearly this approach does not

provide for an exhaustive solution to protein receptor ﬂexibility

but does seek to introduce an exploration of low energy ligand-

binding site ensembles in a way that is computationally feasible.

The subsequent growth of new ligands may be constrained by

the program in a number of user-deﬁned ways, as mentioned in

the Introduction. For instance, geometric constraints may also be

added in the form of allowed fragment sets for linking. Required

fragment sets are placed within a user-deﬁned volume inside

the ligand-binding site of interest and are used preferentially

by PROTOBUILD. Accordingly, if ligands are grown to within a

predeﬁned distance of the geometrically constrained fragment

sets, then a bond is created to include the required fragment sets

within growing ligand structures thereafter. Further constraints

may be applied by the speciﬁcation of pharmacophore constraints

that include locations of cationic, anionic, hydrophobic, H-bond

donor, H-bond acceptor, aromatic and ‘any’ other feature types

of interest. Constraints may also be applied through the execution

of SMILES or SMARTS queries of growing ligands that are

intended to specify molecular structural elements or templates

that must be included or excluded from a ﬁnal population of

ligands grown by the GA. Otherwise if required, a virtual ligand

(VL) command constrains ligands generated to a shape(s) deﬁned

by an existing ligand binding mode(s). In order to support the

VL function within PROTOBUILD, the GA has the ability to use

X-ray crystal structure co-ordinates of ligands co-crystallized with

protein receptors or else co-ordinates generated by the docking of

a ligand of choice into a given ligand-binding site of a protein

receptor of interest whose X-ray crystal structure is known. These

co-ordinates provide the means to imprint the preferred VL

volume of a ligand of choice into the binding site deﬁnition. The

VL volume is similar to an included pharmacophore volume but

the rules for adherence to this representation are programmable

and more complex. This allows the user to ﬁne-tune the required

compliance for any candidate ligand to this volume and thus its

ability to participate in further growth events.

∑ The use of a novel “virtual ligand” (VL) feature in order to

allow a user to embed the binding mode(s) of existing biologically

active ligands into the program and then use shape compliance as

part of an objective function during the generation of new ligands.

Herein, we describe PROTOBUILD in more detail and explain how

this program was employed to suggest a novel set of 2,5-substituted

indoles as dual PPARa/g agonists that were then evaluated in an

in vitro assay system.

Results and discussion

PROTOBUILD data input and program setup

Initially, the PROTOBUILD operation requires a design template and

a number of setup procedures. This functionality is provided by

a programming module known as SiteAnalysis, written in the

SVL language (support vector language) within MOE (Chemical

Computing Group).³⁵Initially, SiteAnalysis computes a grid-

based representation of a speciﬁed protein receptor ligand-binding

site from the given protein structure. In this process preferred site

points for hydrogen bond acceptors and donors are calculated.

Where a ﬂexible ligand-binding site is to be used, SiteAnalysis is

able to generate an ensemble of receptor conformations. Brieﬂy,

a backbone dependent- rotamer library is used to enumerate

PROTOBUILD ligand generation

In order to generate new ligands, PROTOBUILD makes use of GA-

based search and optimization logic for fragment growth under

1170 | Org. Biomol. Chem., 2011, 9, 1169–1188

This journal is

The Royal Society of Chemistry 2011

©

View Article Online

user-deﬁned constraints (as noted above). Evolutionary search

techniques have been shown to be effective at identifying energy

minima in a number of complex (NP-hard) problems within

computational chemistry³⁶and indeed have been used within other

de novo design program implementations too.³⁷Implementation of

our GA-based approach requires that initial molecular fragment

‘seeds’ should be supplied to the program in order to initiate ligand

generation. Usually these fragment seeds are extracted from data

of whole ligand docking studies or from X-ray crystallography

studies of co-crystallized ligands/protein receptors. So too seeds

may derive from fragment docking studies. In PROTOBUILD the

GA then generates and evolves ligands of increasing ﬁtness,

as determined by a predicted binding score. Starting from the

fragment seeds, ligands are grown by the steady incorporation

of new added fragments from a program-accessible fragment

library that link fragment seeds together as ligand structures are

generated. The conformational space of new added fragments is

explored by systematically evaluating all rotamers at defaulted

5^◦angles before low energy solutions are identiﬁed and retained.

New candidate ligands that result are then passed to a module for

ﬁtness evaluation. The PROTOBUILD program and associated code

libraries are written using ISO/ANSI C++ and compiled using the

Visual C++ 8.0 compiler. Command architecture and parameters

are deﬁned by XML. Within PROTOBUILD bond lengths, angles

and atomic radii are provided by the Tripos 5.5 forceﬁeld.³⁸In

a structure-based design scenario, ﬁnal PROTOBUILD designs are

minimized using the MMFF94 forceﬁeld.³⁹The latter has been

shown to perform well elsewhere for similar tasks.⁴⁰

Fitness evaluation is performed using an empirical scoring

function PROTOSCORE (see below) that ranks the effectiveness

of receptor-ligand binding interactions for each new ligand

conformers generated by program under GA guidance. A variety

of penalties are then used to ﬁlter the ligand list prior to additional

rounds of ligand growth. As noted above, the penalty criteria can

include absence of pharmacophore features, presence of chiral

centers, secondary and tertiary branch points, Lipinski parameters

(including molecular weight and log P upper limits), number of

linked ring systems and so on. The penalty criteria are particularly

helpful in removing large numbers of candidate ligands with

overcomplicated structures even though they may result previously

from valid fragment pairings. Clearly penalty criteria could vary

signiﬁcantly depending upon the protein receptor and ligand-

binding sites under investigation. Once the high scoring ligands

have been identiﬁed for the next round of ligand growth, a subset

of the remainder is still recycled in the next round allowing some

of the less desirable ligands a “wildcard entry” to the next growth

and search cycle. In particular, candidate ligands are retained with

a low degree of structural similarity in comparison to compounds

with the highest score from PROTOSCORE. In so doing, diversity

is increased in acknowledgement of the fact that some ligands

whilst being suboptimal may in fact become much better ligands

in subsequent growth and search cycles after further fragment

additions.

hydrophobic fragments to the growing molecule. This is because

these compounds score comparatively well due to the Van der

Waal’s interactions with the receptor and thus are retained for

further growth. However, such ligands are undesirable in terms

of physical properties and seldom represent a global minimum.

Fortunately, the extensive use of penalty criteria such as Log P

selection helps to overcome this problem. Also, added fragments

can be associated with a user-deﬁned selection pressure that may

be used to control the likelihood of selection at any given time

during ligand growth. In particular, PROTOBUILD is able to alter

dynamically the selection pressures of the added fragments thus

preventing the “ﬂood-ﬁlling” of the ligand-binding site early on

in the design process with hydrophobic fragments. There is also

a facility to prevent or limit the number of ring systems that

are directly connected to each other because these systems often

present challenging synthetic routes. Alternatively, SMARTS

patterns may be used to explore only ligands of certain generalized

scaffolds and force solutions that adhere to these rules.

PROTOBUILD can run in two different modes. In one ‘determin-

istic’ mode a set number of iterations of the GA are executed. In

the other ‘non-deterministic’ mode, the run continues indeﬁnitely.

In both modes the program may be paused such that emerging

ligands may be reviewed and run parameters may be changed in

accordance with a runtime programming language implemented

in XML. Both the deterministic and non-deterministic search

cycle modes for PROTOBUILD can make use of a fragmentation

capability in order to boost the exploration of diversity in chemical

space exploration during a given PROTOBUILD run. In fact, even

entire ligands may be supplied to PROTOBUILD as seeds where

they are broken up randomly and recursively in order to generate

a controllable number of discrete starting points for the design

process. This strategy is particularly useful in lead optimization

studies.

Scoring with PROTOSCORE

Central to PROTOBUILD is the empirical scoring function that was

newly developed for use in PROTOBUILD. This scoring function is

known as PROTOSCORE. PROTOSCORE incorporates a treatment for

ligand solvation, receptor deformation upon binding and calcula-

tions for aromatic and heterocyclic ring interactions. PROTOSCORE

builds upon many of the concepts implemented in the earlier

pioneering work of Bohm.⁴¹The main equation developed for

PROTOSCORE is shown (1):

pK_i= c + _x1HB + _x2SURFMATCH + _x4LIGFLEX

(1)

+

_x5PROTFLEX + _x6SOLV + _x7ARO

The terms in the equation will be described brieﬂy. Hydrogen

bonds (HB) are calculated according to the term (2):

(2)

HB = RhbF + RhbwF + RhbmF

where hb is the interaction energy of a perfect hydrogen bond and

F is a measure of geometric deviation from ideal with a value

between 0 and 1. The terms hbw and hbm are interaction energies

for water mediated hydrogen bonds and transition metal contacts

respectively. Surface lipophilic interactions and polar/apolar

interactions (SURFMATCH) are calculated by means of a surface-

matching algorithm described by the following eqn (3):

An inherent problem during de novo design is the inclination

to generate compounds with excessive hydrophobicity.³²Such

compounds are undesirable for a number of reasons such as

poor solubility and hence drug formulation. This problem arises

within ordinary growth algorithms due to the tendency of the

scoring function to favor the addition of large and usually

SURFMATCH = R(lipo–lipo) -R(alipo-lipo) -R(lipo-alipo) (3)

Org. Biomol. Chem., 2011, 9, 1169–1188 | 1171

This journal is

The Royal Society of Chemistry 2011

©

View Article Online

where lipo-lipo is the lipophilic contact area between ligand-

binding site and ligand, alipo-lipo is the contact area between

polar areas of a ligand-binding site and non-polar ligand regions

while lipo-alipo is the contact area between non-polar regions

of a ligand-binding site and polar ligand regions. SURFMATCH

expresses the protein receptor-ligand contact surface area using

a grid method in a method similar to that previously reported.⁴¹

Where ligand polar groups are buried in lipophilic environments

and ligand-binding site polar groups map close to ligand lipophilic

groups, then there is a penalty.

Selection of training and test sets

The deﬁnition of suitable training and test sets of protein

receptor-ligand complexes is as important as the deﬁnition of

the descriptors to be used during training. For PROTOSCORE a

diverse set of 275 receptor-ligand structures was used. This set

incorporated kinases, proteases, phosphatases, nuclear receptors,

metalloproteases and sugar binding proteins. For this dataset the

equation ﬁtted by PLS yielded an r²= 0.82, q²= 0.77, S = 6.6 kJ

mol^-1, S_press= 7.4 kJ mol^-1. On a separate test set of 50 proteins an

r²of 0.73 and S_pressof 7.9 kJ mol^-1was recorded.

Terms LIGFLEX and PROTFLEX are included to account for

protein receptor and ligand ﬂexibility. The entropic effect caused

by the requirement to reduce conformational degrees of freedom

upon binding can be expressed by a count of rotatable bonds

for both interacting ligand and protein receptor. For the ligand

the term LIGFLEX calculates the number of acyclic sp3-sp3 or

sp3-sp2 bonds but does not include terminal groups. Schemes

for protein receptors are more complicated because ﬂexibility

of the ligand-free state must be calculated and compared with

ﬂexibility in the ligand-bound state. After investigation of a

number of schemes PROTFLEX calculates the number of rotatable

bonds of accessible portions of side chains that contact the binding

ligand. A more accurate version of the function PROTFLEX2 pre-

calculates the number of low energy rotamers available to ligand-

binding site amino acid residue side chains and uses this value

to weight the number of rotatable bonds in contacting side

chains. This extra step needs only to be run once per protein

receptor. Otherwise, SOLV takes into account the entropic gain

due to desolvation when ligand binds the ligand-binding site of a

protein receptor. The term SOLV uses a ﬁnite difference method for

fast solution implementation of the Poisson-Boltzman equation.

Only the desolvation volume of bound ligand is calculated.

This term is computationally expensive requiring a Poisson-

Boltzman equation to be solved twice per ligand. With the current

implementation this adds 0.5 s per ligand to the calculation.

Finally, aromatic interactions are described by the following

equation:

Deﬁning fragments for PROTOBUILD

Ideally a de novo design program should only really output can-

didate ligands with structures that can be realistically synthesized

from available reagents. This is not always the case since proposed

ligands are constructed from randomly assembled fragments

rather than synthons connected by synthetic rules. Some de novo

design approaches to new ligands have been reported to construct

ligands from reagents although it is difﬁcult to express sufﬁcient

chemistry in silico to provide the level of diversity required in many

design projects.³⁷Automated retrosynthetic approaches have also

been used to assess ease of synthesis but often lead to solutions

that are not automatically favoured and which and are often

unreliable when dealing with complex substitution patterns on

aromatic groups.

PROTOBUILD generates new ligands according to a set of

fragment-fragment interconnectivity rules. These rules stipulate

that fragment additions can only take place if the fragment-

fragment substructure exists in a known compound. The fragment

library is constructed by decomposition of 38 000 known, patent

protected, drug-like molecules extracted from the Thomson

Pharma database. In this analysis the observed substitution

patterns for the fragments are stored as are the frequencies of

substitutions. Enforcing these rules during fragment coupling

ensures that any two combined fragments must be present as

a substructure within an existing molecule in the training set

(Scheme 1). Any library of molecules can be used as input to

the fragment deﬁnition procedures. The inclusion of these rules

does limit the chemical space being searched but signiﬁcantly

enriches the chemical ‘sense’ of the resulting molecules. The

standard fragment library in PROTOBUILD is comprised of 594

fragments, most of which are ring systems. Within this set of

fragments there are 1843 unique atom environments that can yield

a theoretical maximum of 3 396 649 different ligands composed of

only 2 fragments. Analysis shows however that only 53 426 of these

potential ligands can be generated based on the trained fragment

pairing rules since the majority of the atom environment pairings

are not observed in the training set.

(4)

ARO = AROaro + AROcat + AROs

Principally, the ARO term (4) contains calculations for aromatic-

aromatic rings AROaro, cation-pi interactions AROcat and sulfur-

pi interaction AROs. AROaro is calculated by measuring distance

and angle between the rings of a ligand and those of Trp, Phe,

Tyr or His amino acid residues belonging to a protein receptor.

The function calculates energetically favourable edge to face and

offset face to face interactions as previously described.^42,43AROcat

calculates cation-pi interactions expressed as a function of the

distance and angle of a positive charge to the face of an aromatic

ring system. The value of the term is provided from a lookup table

generated from a set of experimentally determined interactions.⁴⁴

The AROcat term compensates for the penalization of polar-apolar

interactions identiﬁed within the SURFMATCH term. The separate

term AROs describes the interaction phenomena observed when an

sp3 sulfur atom interacts with a pi system. Once again the value of

this term is trained from experimental data.⁴⁵This interaction

can be underestimated using a simple treatment of lipohilic

contacts.

Designing novel indole compounds

At the start of the design process the crystal structure of a

known potent PPARg agonist 1 (GI262570) complexed with

PPARg (PDB 1FM9) was studied (Fig. 1).⁴⁶The crystal structure

demonstrates a number of important interactions between the

ligand and receptor including hydrogen bonding between the

carboxylic acid moiety and residues His449, Tyr473 and His323

(of the AF-2 helix). The hydrophobic phenoxy moiety interacts

1172 | Org. Biomol. Chem., 2011, 9, 1169–1188

This journal is

The Royal Society of Chemistry 2011

©

View Article Online

Scheme 1 Derivation of fragments from known molecules and rules for generating new ligands. Training compounds shown in the box with different

fragments colored. Outputs 1 and 2 are examples of allowed outputs that have the same substitution patterns on the indole as observed in training

compounds and fragment A-B combinations are maintained. Outputs 3 and 4 are examples of disallowed products. Output 3 is disallowed due to

incorrect fragment pairing whereas output 4 is disallowed due to an incorrect substitution pattern.

ﬂexibility of the LBD was assessed by evaluating sidechain rotamer

ﬂexibility for all residues with at least one contact to ligand 1

in the PPARg structure using procedures implemented in MOE

(Chemical Computing Group Inc., Quebec, Canada). In general,

side chain mobility was found to be much lower in the narrow

channel of the binding site proximal to the key hydrogen bonding

interactions of the carboxyl group, than the distal regions in the

opening of the LBD. This is in agreement with published B-factor

values for the crystal structure. The lowest energy rotamer states

as determined by AMBER89 forceﬁeld were clustered and 144

Fig. 1 Structure of 1 (GI262570).

were used as inputs into the PROTOBUILD program as competing

receptor structures for evolving ligands.

Finally, PROTOBUILD was used to grow molecules in the binding

site representations driven by the PROTOSCORE scoring algorithm

as the objective function, in addition to other target ﬁlters

including molecular weight limit <500, predicted solubility >1 mM

and number of hydrogen bond donors and acceptors <10.^47,48

A total of 594 distinct fragments were utilised in the growing

procedures, each with synthetic meta-data providing instructions

as to which fragments may combine at which substitution points,

in a manner similar to the RECAP scheme.⁴⁹As shown in Fig. 2,

the input to the program was a carboxylic acid seed originally

excised from the X-ray structure of GI262570 and which was

marked as protected such that the algorithm would not delete

these atoms during rounds of optimisation. PROTOBUILD was run

primarily through van der Waals forces with residues Tyr327,

Leu330 and Cys285. Analysis of this ligand-binding domain

(LBD) and also that of PPARa (PDB 1I7G) enabled a hypothesis

for a pharmacophore required to confer both PPARg and PPARa

agonism that could be used as an input to a de novo design scheme

(Fig. 2). The coordinates of the carboxyl group from ligand 1

were used to represent the ‘seed’ structure for the ligand-growing

algorithm.

Two pharmacophore points were speciﬁed as being satisﬁed by

output structures, an aromatic feature in the narrow hydrophobic

˚

channel located 4.1 A from the carbon atom of the carboxyl

˚

seed and an included volume feature located 10.6 A from the

˚

same carbon atom and 7.1 A from the aromatic feature. The

This journal is

The Royal Society of Chemistry 2011

Org. Biomol. Chem., 2011, 9, 1169–1188 | 1173

©

View Article Online

Fig. 2 Design process. Left: Ligand 1 (stick format) in complex with PPARg with overlaid rotamer conformations from receptor ﬂexibility analysis

(pharmacophore features shown as wire spheres). Right: carboxylic acid seed for de novo design shown with exit vector but with ligand 1 removed from

site.

Table 2 2,5-disubstituted indole compounds 2 to 7 as shown in outline

under a non-deterministic mode in which iterative populations of

ligands were built and trimmed until the algorithm was manually

terminated once sufﬁcient numbers of ligands were generated

which had met the output criteria. The process resulted in several

thousand structures that satisﬁed the pharmacophore scheme

and had predicted K_dbinding afﬁnities <1 mM according to the

PROTOSCORE function. The solutions were clustered using MACCS

key ﬁngerprints and 75% similarity as the cluster deﬁnition. This

resulted in ~200 different clusters or chemotypes to be assessed for

tractability and potential novelty. These results contained known

active PPARg chemotypes as well novel structures. Some of the

known PPAR modulator scaffolds observed in the PROTOBUILD re-

sults are described (Fig. 3). The observance of known chemotypes

encourages that other compounds proposed by the algorithm may

also prove to be active. In total three different novel scaffolds were

shortlisted for synthesis. These were selected based on highest

ligand efﬁciency values rather than overall predicted afﬁnity since

generally it is considered that this often produces simpler starting

structures for initial designs. The 2,5- disubstituted indole scaffold

was the ﬁrst scaffold to be synthesised, primarily based upon

availability of reagents and the ease of synthesis, coupled with

a degree of novelty.

Recently, indole cores have featured in other known PPAR

agonists described in the past few years,⁵⁰including several that

are comprised of a 1,5-disubstituted indole core.^51–54By contrast,

PROTOBUILD output predicted that particular 2,5-disubstituted

indoles, wherein position-2 substitutions are aromatic in character,

should also result in effective PPAR agonists. Furthermore,

PROTOBUILD output also predicted an effective substitution of

hydrogen for ﬂuorine within saturated side-chain substitutions

in position-5. Therefore, we resolved to prepare a selection of

2,5-disubstituted indole variants with different aromatic groups,

R¹, in the 2-position and an ethoxypropanoic side chain in 5-

position, wherein R²is either a methyl- or a triﬂuoromethyl- group

according to the PROTOBUILD output (Fig. 4). Such compounds

each have one chiral centre a- to their respective carboxylic acid

moieties, but we elected to carry out racemic syntheses in the

structure in Fig. 4

Compound

R¹

R²

2

3

4

5

6

7

H

Benzyl

2-Naphthylmethyl

CH₃

CF₃

CH₃

CF₃

CH₃

CF₃

ﬁrst instance and these racemates were then evaluated as potential

PPAR agonists in appropriate PPAR assays in vitro.

Synthetic Chemistry

The PROTOBUILD software program suggested the synthesis of

molecules 2 to 7, with R¹= H, phenyl and naphthylmethyl groups

as potential PPARa/g modulators (Table 2). Indole structures 2

and 3 were introduced to evaluate the importance of 2-substituents

in indoles for binding to the receptor in a hydrophobic site that is

occupied by the methyl oxazole moiety of ligand 1.

For indoles 2 and 3, where R¹is hydrogen, commercially avail-

able indole-5-carboxaldehyde 8 provided an ideal template from

which the aryl substituents and PPAR-binding carboxylic acid

moiety could be synthesised (Scheme 2). A Horner–Wadsworth–

Emmons (HWE) coupling was carried out with the required

phosphonoacetates 9 and 10 to give the Z and E isomers 11 to

14. The reduction of the double bond with magnesium turnings

in dry methanol proceeded smoothly, as described by Lohray

et al.⁵⁵Finally, saponiﬁcation of the methyl ester with potassium

hydroxide in a 1 : 1 mixture of ethanol–water gave the desired

targets 2 and 3 in excellent yields.⁵⁶The HWE coupling reagents 9

and 10 were prepared as shown (Scheme 3). When R²is a methyl

group, triethylphosphite and commercially available 2-chloro-2-

ethoxyacetic acid 17 were coupled using Arbuzov chemistry as

described by Grell and Machleidt.⁵⁷This reaction proceeded

smoothly to give phosphonoacetate 9 in a quantitative yield of

1174 | Org. Biomol. Chem., 2011, 9, 1169–1188

This journal is

The Royal Society of Chemistry 2011

©

View Article Online

Fig. 4 2,5-Disubstituted indole scaffold where R¹represents the volume

feature, R²the aromatic feature and the carboxylate functional group

represents the seed.

described by Moody et al.⁵⁹and Regitz at al.⁶⁰Diazo intermediate

18 proved stable to silica chromatography. Thereafter, rhodium(II)

acetate dimer was used to generate a carbenoid from diazo

compound 18 which was inserted into the O–H bond of 2,2,2-

triﬂuoroethanol, as per Haigh et al. (Scheme 3).⁶¹

For indoles 4 to 7, where R¹is a benzyl or a naphthylmethyl

group, a suitably functionalised cyano indole 20 provided an

ideal starting material (Scheme 4). The ﬁrst step was to protect

the nitrogen of commercially available indole 20 with a benzene

sulfonyl group with the help of the phase transfer catalyst (nBu)₄Br.

Alternatively, benzene sulfonylchloride in THF was found to

protect indole 20 with the help of sodium hydride. The former

transformation gave the higher yields in our case. Protected indole

21 was puriﬁed by crystallisation from ethanol in a 96% yield.

The benzene sulfonyl group was chosen as it allows chelation

of lithium to the oxygen of the benzene sulfonyl group and

enhances substitution on the C2 position due to the inﬂuence

of the nitrogen atom. After treatment of indole 21 with LDA in

dry THF at -78 ^◦C adapted from a method by Sundberg and

Russell, the lithiated intermediate (without isolation) was treated

with either benzyl bromide or 2-(bromomethyl)-naphthalene.⁶²

After chromatography on silica gel, crystallisation from hexane–

DCM 20 : 1 yielded shiny white crystals. X-ray crystal structure

analyses were carried out to conﬁrm the selective introduction

of the R¹group on the 2-position of the indole (Fig. 5). The

next step required deprotection of the benzene sulfonyl group

in basic conditions. Cyano indoles 24 and 25 were subsequently

reduced to the corresponding aldehydes 26 and 27 respectively

using DIBAL. Next, a series of HWE couplings were performed

to couple aldehydes 26 or 27 with phosphonoacetates 9 or 10

resulting in four different Z and four different E double bond

isomers 28–35. Double bond reduction with magnesium turnings

in dry methanol was then performed giving four indole precursors

36 to 39 that were saponiﬁed as described previously to yield the

desired indole compounds 4 to 7. Hence the six desired indoles 2 to

7 were thus successfully synthesised. The purity of the compounds

was an important aspect, prior to their testing in vitro. A minimum

of 97% purity of the indoles was achieved by careful puriﬁcation of

all the intermediates, followed by crystallisation wherever possible.

In vitro PPAR activation assays

Fig. 3 Some of the previously reported chemotypes identiﬁed within

The activation of the three PPAR subtypes by the indoles 2 to 7 was

studied in vitro using a luciferase reporter gene assay in which the

reporter construct contained PPREs in the promoter of the ﬁreﬂy

luciferase gene. Speciﬁcally, a plasmid encoding for each PPAR

subtype, as well as the luciferase reporter gene was transfected into

human breast cancer MCF-7 cells. Transfection was performed

in the absence of PPAR expression vectors as negative control.

PROTOBUILD results.

92% ready for use without any puriﬁcation. For the synthesis of the

2,2,2-triﬂuoro phosphonoacetate 10, tosyl azide was prepared by

coupling sodium azide with tosyl chloride.⁵⁸This azide appeared

stable to handle and was used to synthesize a diazo intermediate

18 from phosphonoacetate 19 using sodium hydride in dry THF as

This journal is

The Royal Society of Chemistry 2011

Org. Biomol. Chem., 2011, 9, 1169–1188 | 1175

©

View Article Online

Scheme 2 Synthesis of indoles 2 and 3. i) NaH, dry THF, (0 ^◦C to) ambient temperature, 26–45 h, 93–99% ii) Mg turnings, dry MeOH, ambient

temperature, 3 h, 60–83% iii) KOH, EtOH–H₂O 1 : 1, reﬂux, 15 h, 92–93%.

Scheme 3 Preparation of the phosphonoacetates 9 and 10, i) Anhydrous DMF, reﬂux at 145 ^◦C, 15 h, 92% ii) EtOH, acetone, ambient temperature,

23 h, 98% iii) NaH, dry THF, 0 ^◦C, 16 h, 73% iv) 2,2,2-triﬂuoroethanol, [Rh(OAc)₂]₂, benzene, reﬂux, 22 h, 63%.

This allowed checking for any endogenous PPAR activity. Under

these conditions, no signiﬁcant luciferase expression was detected.

A negative control was included for each set of experiments by

treating cells with dimethyl sulfoxide (DMSO) in the absence

of any compounds. Positive controls were performed with the

following selective potent PPAR agonists at concentrations of 40

(WY14.643) at 30 mM for PPARa; 41 (BRL49653-rosiglitazone) at

1 mM for PPARg, and 42 (L165.041) at 1 mM for PPARd (Fig. 6).

Experiments were conducted with ﬁve different concentrations

of each indole 2–7 at 1 mM to 75 mM in DMSO to obtain

a preliminary indication of their effects on PPAR activation.

The lowest concentration that gave activity as well as a higher

concentration relative to the ﬁrst was selected for data presentation

(Fig. 7). All reported experiments were performed in triplicates at

these indicated concentrations. All results were normalised relative

to the control and the standard deviation was calculated from

the variability of the readings obtained for one compound at a

particular concentration.

1176 | Org. Biomol. Chem., 2011, 9, 1169–1188

This journal is

The Royal Society of Chemistry 2011

©

View Article Online

Scheme 4 i) PhSO₂Cl, (n-C₄H₉)₄NBr, NaOH(aq), toluene, H₂O, ambient temperature, 25 h, 96% ii) a) LDA, -78 ^◦C, dry THF, 55 min: b) arylBr, -78 to

0

^◦C to ambient temperature, 16 h, 30–66.5% iii) NaOH(aq), MeOH, reﬂux, 18 h, 99% iv) DIBAL, dry DCM, 0 ^◦C to ambient temperature, 16 h, then

0 ^◦C, 4.5 h, 80–86% v) NaH, dry THF, (0 ^◦C to) ambient temperature, 26–45 h, 78–90% vi) Mg turnings, dry MeOH, ambient temperature, 3 h, 73–80%

vii) KOH, EtOH–H₂O, reﬂux, 15 h, 93–99%.

Structure activity relationship

group appears to confer better indole-mediated PPARa activation

than the inclusion of the 2-benzyl group. In addition, in comparing

2-naphthylmethyl indoles 4, 5, 6 and 7, the overall functional

requirement for a PPARa/g selective agonist appears better

satisﬁed by indoles 4 and 6 so implicating the ethoxy functional

group as a potentially more suitable functional group than 2,

2, 2-triﬂuoroethoxy. The PROTOBUILD output indicated that the

indole ring could be tolerated in the ligand binding domains of

both PPARa and PPARg. The ligand binding domains of PPARa

and PPARg are similar in size, allowing such type of agonists

to be developed.^63,64The predicted binding mode of the indole 6

involves the 2-naphthylmethyl group ﬁtting into the hydrophobic

site near the entrance of the LBD pocket; the indole ring ﬁtting

into the narrow hydrophobic channel occupied by the phenoxy

moiety of ligand 1 while the carboxylic acid occupies the same

position as with 1 and other PPAR activators (Fig. 8). The

interaction of the 2-naphthylmethyl compounds with receptors

was scored highest with interacting sidechain rotamers positioned

as shown (Fig. 2), with the exception of the ﬂexible Met348 residue

whose optimal rotamer-position occupied a different confor-

mation.

When comparing the PPAR activation proﬁles of the 6 synthesised

indoles, the following patterns are apparent:

1. Indoles 2 and 3 were not capable of signiﬁcantly higher PPAR

activation than any of the three indicated positive controls 40–42.

2. Indole 4 signiﬁcantly activated PPARa relative to pos-

itive control 40 (WY14.643), equivalent to 41 (BRL49653-

rosiglitazone) with respect to PPARg-selective activation, but

much less signiﬁcant than 42 (L165.041) with regard to PPARd–

selective activation. By contrast, Indole 5 was equivalent with but

did not exhibit signiﬁcantly higher PPAR activation than any of

the three indicated positive controls 40–42.

3. Indole 6 induced PPARa selective activation more than the

positive control 40 (WY14.643), equivalent with 41 (BRL49653-

rosiglitazone) with respect to PPARg-selective activation, but also

less signiﬁcant than 42 (L165.041) with regard to PPARd–selective

activation.

4. Indole 7 also activated PPARa more than positive control

40 (WY14.643), but clearly less effective than 41 (BRL49653-

rosiglitazone) and 42 (L165.041) with regard to PPARg and

PPARd-selective activation respectively.

Hence, our initial conclusions from this limited series of

compounds are that inclusion of a 2-naphthylmethyl functional

From now, we aim to carry out some further tests on the

compounds, including toxicity and some preliminary pharma-

cokinetic measurements in order to select the compound with

This journal is

The Royal Society of Chemistry 2011

Org. Biomol. Chem., 2011, 9, 1169–1188 | 1177

©

View Article Online

Fig. 5 The molecular structures of the 2-benzyl indole intermediate 22 (both top structures) and the molecular structure of the 2-naphthylmethyl

intermediate 23 (bottom structure).

Fig. 6 Selective PPAR agonists used as positive controls and TTA 43.

the best drug-like properties. Once the best candidate is chosen,

the enantioselective synthesis can be carried out to give the desired

enantiomer and to proceed to further testing of the compound as a

Experimental

General chemistry

potential treatment of metabolic diseases as well as inﬂammatory

diseases. Thereafter we shall turn our attention to other novel

chemotypes suggested by the software program PROTOBUILD.

Unless otherwise stated all reactions were carried out under an at-

mosphere of nitrogen or argon, in oven-dried glassware. DCM was

distilled over P₂O₅, and other solvents were bought and pre-dried

1178 | Org. Biomol. Chem., 2011, 9, 1169–1188

This journal is

The Royal Society of Chemistry 2011

©

View Article Online

Fig. 7 PPAR activation with the indoles 2 to 7 in transfected human MCF-7 cells. For each indole, relative activation is given in the following order from

left to right: control cells (black column), cells incubated with a PPARa agonist 40 (WY14.643), PPARg agonist 41 (BRL49653) or a PPARd agonist

42 (L165.041) (blank columns). In between is shown the activation of PPARa, PPARg and PPARd respectively with two different concentrations of

the speciﬁc indole compound (grey-lined columns). Data are presented as mean SD (n = 3). ^*Signiﬁcantly higher activation than in control cells (P <

§

0.05), ^#signiﬁcantly higher activation than with the corresponding speciﬁc PPAR agonist (P < 0.05), signiﬁcantly higher activation than with lowest

concentration of the same indole and same PPAR isoform (P < 0.05).

This journal is

The Royal Society of Chemistry 2011

Org. Biomol. Chem., 2011, 9, 1169–1188 | 1179

©

View Article Online

([M+H]⁺, 10), 224 ([M-ethyl ether]⁺, 81), 197 ([M-(ether+ester)]⁺,

90), 167 (85), 152 (70), 139 (75), 111 (100), 103 (55), 75 (62), 65

(71); m/z HRMS (EI⁺) [M]⁺calcd 268.1076 for C₁₀H₂₁O₆P, found

268.1302; Elemental calcd C, 44.8; H, 7.9; found 44.6; H, 7.9.

(Z)-Ethyl 2-ethoxy-3-(1H-indol-5-yl)acrylate 11 and (E)-ethyl

2-ethoxy-3-(1H-indol-5-yl)acrylate 12. Procedure A: This syn-

thesis was performed in accordance to Kuhn et al.⁵¹In brief,

phosphonoacetate 9 (1.87 g, 6.97 mmol) in THF (2.0 mL) was

slowly added to an ice cold suspension of sodium hydride (60% in

oil; 310 mg, 7.75 mmol) with molecular sieves in THF (40 mL),

then stirred at 0 ^◦C for 1 h. 5-indole-carbaldehyde 8 (506 mg,

3.48 mmol) in THF (4.0 mL) was added dropwise and the resulting

orange solution stirred at 0 ^◦C for 1 h and at ambient temperature

for 44 h. After concentration in vacuo, the material was re-

dissolved in EtOAc (40 mL) and washed twice with water, with

the aqueous phase being extracted with EtOAc After drying over

MgSO₄, concentration in vacuo gave a brown liquid which was

puriﬁed by ﬂash column chromatography (hexane–EtOAc 20 : 1,

15 : 1, 10 : 1, 9 : 1, 8 : 2) to yield 844 mg (93%) of two isomers. Ratio

Z : E = 63 : 37. Z isomer 11: amorphous white powder; mp 77.5–

78.5 ^◦C (hexane–DCM 20 : 1); R_f0.35 [hexane–EtOAc 6 : 4]; n_max

(nujol)/cm^-13267, 2923–2850, 1731, 1687, 1550, 1378; ¹³C NMR

(CDCl₃): d 165.30 (C), 142.56, 136.03, 127.98 (C), 126.19 (CH),

125.60 (C), 124.90 (CH), 124.56 (CH), 123.46 (CH), 111.00 (CH),

Fig. 8 Proposed interaction of indole 6 with the active site of PPARg.

as required. All chemicals were purchased from Sigma-Aldrich,

Lancaster or Merck Biosciences. Flash column chromatography

(Merck Kieselgel 60 F₂₅₄230-240 mesh) was performed according

to the method of Still et al.⁶⁵Thin layer chromatography (TLC)

was performed on pre-coated Merck silica gel (0.2 mm, 60 F₂₅₄

)

aluminium-backed plates, and visualised with a UV lamp (254

nm) and/or stained with acidic ammonium molybdate (IV), basic

potassium manganate (VII; KMnO₄), iodine or phosphomolybdic

acid. Chromatography solvent mixture A is as follows: CH₂Cl₂–

MeOH–H₂O 65 : 25 : 4. Melting points were measured on a Stuart

Scientiﬁc SMP3 apparatus and are reported without correction.

Infrared spectra (IR) were measured on a JASCO FT/IR-620

spectrometer. ¹H NMR and ¹³C NMR spectra were recorded

on Bruker Avance 400. ¹H NMR was recorded at 400 MHz

1

103.36 (CH), 67.46 (CH₂), 60.92 (CH₂), 15.55, 14.35 (CH₃); H

NMR (CDCl₃): d 8.32 (1H, br, H1), 8.13 (1H, s, H4), 7.73 (1H, dd,

J 8.4, 1.6 Hz, H6), 7.39 (1H, d, J 8.8 Hz, H7), 7.23 (1H, t, J 2.8 Hz,

H2), 7.20 (1H, s, H1¢), 6.60–6.58 (1H, m, H3), 4.32 (2H, q, J 7.2 Hz,

H3¢), 4.03 (2H, q, J 7.0 Hz, H6¢), 1.44–1.38 (6H, m, H4¢, H7¢); m/z

(EI⁺) 259 ([M]⁺, 100), 202 ([M-CH₂CH₃-CH₂CH₃]⁺, 39), 157 (75),

129 (51), 118 ([Indole]⁺, 13); m/z HRMS (EI⁺) [M]⁺calcd 259.1208

for C₁₅H₁₇NO₃, found 259.1199; Elemental calcd C, 69.5; H, 6.6;

N, 5.4; found C, 69.6; H, 6.7; N, 5.3. E isomer 12: amorphous

white powder; mp 98.5–100.5 ^◦C (hexane–DCM 20 : 1); R_f0.28

[hexane–EtOAc 6 : 4]; n_max(nujol)/cm^-13330, 2923–2850, 1702,

1621, 1379, 1227, 1176; ¹³C NMR (CDCl₃): d 165.24 (C), 146.23,

134.93, 127.82, 126.17 (C), 124.54 (CH), 122.99 (CH), 120.54

(CH), 111.47 (CH), 110.58 (CH), 102.57 (CH), 64.62 (CH₂), 61.07

(CH₂), 14.53 (CH₃), 13.68 (CH₃); ¹H NMR (CDCl₃): d 8.23 (1H,

br, H1), 7.51 (1H, d, J 0.4 Hz, H4), 7.30 (1H, d, J 8.4 Hz, H7),

7.17 (1H, t, J 2.8 Hz, H2), 7.07 (1H, dd, J 8.4, 1.6 Hz, H6), 6.50–

6.49 (1H, m, H3), 6.32 (1H, s, H1¢), 4.16 (2H, q, J 7.0 Hz, H3¢),

3.96 (2H, q, J 6.8 Hz, H6¢), 1.43 (3H, t, J 6.8 Hz, H7¢), 1.08 (3H,

t, J 7.0 Hz, H4¢); m/z (EI⁺) 259 ([M]⁺, 100), 202 ([M-CH₂CH₃-

CH₂CH₃]⁺, 37), 157 (68), 129 (49); m/z HRMS (EI⁺) [M]⁺calcd

259.1208 for C₁₅H₁₇NO₃, found 259.1199; Elemental calcd C, 69.5;

H, 6.6; N, 5.4; found C, 69.6; H, 6.5; N, 5.5.

using residual isotopic solvent as internal reference (CDCl₃, d_H

=

7.27 ppm; CD₃OD, d_H= 3.30 ppm). ¹³C NMR spectra were

recorded at 100 MHz using residual isotopic solvent as internal

reference (CDCl₃, d_C= 77.00 ppm; CD₃OD, d_C= 49.05 ppm).

Peaks split by the presence of a phosphorus atom are indicated

with a superscript p. Mass spectra were recorded using VG

Platform II, VG-070B, Joel SX-102 or Bruker Esquire 3000 ESI

instruments. Mass accuracy is indicated to the nearest 0.1 ppm.

Elemental analysis was carried out using a Perkin Elmer 2400

CHN elemental analyser at the Science Technical Support unit,

London Metropolitan University.

Ethyl 2-(diethoxyphosphoryl)-2-ethoxyacetate 9. This synthe-

sis was performed in accordance to Grell and Machleidt.⁵⁷In brief,

triethylphosphite (1.46 mL, 8.49 mmol) was added to a colourless

solution of 2-chloro-2-ethoxyacetic acid ethyl ester 17 (1.41 g,

8.49 mmol) in anhydrous DMF (10 mL) and the reaction mixture

reﬂuxed for 15 h at 145 ^◦C. The resulting yellow solution was

cooled to ambient temperature and water (40 mL) added. The

aqueous phase was extracted with DCM, dried over MgSO₄and

concentrated in vacuo yielding 2.11 g of 9 as a pale yellow oil

(92%): R_f0.25 [hexane–EtOAc 1 : 1]; n_max(ﬁlm)/cm^-11746, 1265,

1145, 1039; ¹³C NMR (CDCl₃): d 167.33^p(C, J 2.0 Hz), 76.55^p

(CH, J 157.0 Hz), 68.29^p(CH₂, J 12.0 Hz), 63.43^p(CH₂, J 7.0 Hz),

63.34 (CH₂, J 7.0 Hz), 61.45 (CH₂), 16.12^p(CH₃, J 6.0 Hz), 14.70

(CH₃), 13.88 (CH₃); ¹H NMR (CDCl₃): d 4.24–4.09 (5H, m, H7,

H5), 3.68–3.61 (1H, m, H3), 3.54–3.47 (1H, m, H3), 1.29–1.21 (9H,

m, H8, H4 or H6), 1.19–1.16 (3H, m, H4 or H6); m/z (EI⁺) 269

Methyl 2-ethoxy-3-(1H-indol-5-yl)propanoate 15. Procedure

B: Magnesium turnings (648 mg, 26.7 mmol, 18.2 eq) were added

to a solution of 11 (380 mg, 1.47 mmol, 1.00 eq) in anhydrous

methanol (13.0 ml), under a constant ﬂow of N₂at ambient

temperature. The reaction mixture was stirred for 3 h then poured

onto water and the aqueous phase extracted with DCM. DCM

extracts were dried over MgSO₄, then concentrated in vacuo to

give a residue that was puriﬁed by ﬂash column chromatography

(hexane–EtOAc 9 : 1, 8 : 2, EtOAc) to yield 300 mg (83%) of 15 as a

yellow oil: R_f0.20 [hexane–EtOAc 7.5 : 2.5]; n_max(ﬁlm)/cm^-13408,

2977–2850, 1741, 1444, 1209, 1110; ¹³C NMR (CDCl₃): d 173.30

1180 | Org. Biomol. Chem., 2011, 9, 1169–1188

This journal is

The Royal Society of Chemistry 2011

©

View Article Online

(C), 134.80 (C), 128.19 (C), 127.98 (C), 124.38 (CH), 123.54 (CH),

121.04 (CH), 110.74 (CH), 102.23 (CH), 80.98 (CH), 66.21 (CH₂),

51.73 (CH₃), 39.51 (CH₂), 15.02 (CH₃); ¹H NMR (CDCl₃): d 8.25

(1H, br, H1), 7.51 (1H, s, H4), 7.31 (1H, d, J 8.0 Hz, H7), 7.18 (1H,

t, J 2.8 Hz, H2), 7.10 (1H, dd, J 8.4, 1.6 Hz, H6), 6.51–6.50 (1H, m,

H3), 4.13 (1H, dd, J 7.6, 1.6 Hz, H2¢), 3.71 (3H, s, H4¢¢), 3.65–3.58

(1H, m, H3¢), 3.42–3.35 (1H, m, H3¢), 3.15 (1H, s, H1¢), 3.13 (1H,

d, J 2.8 Hz, H1¢), 1.18 (3H, t, J 7.0 Hz, H4¢); m/z HRMS (EI⁺)

[M]⁺calcd 247.1208 for C₁₄H₁₇NO₃, found 247.1205; Elemental

calcd C, 68.0; H, 6.9; N, 5.7; found C, 68.2; H, 6.9; N, 5.8.

the combined organic extracts were then dried over MgSO₄and

concentrated in vacuo. The residue was puriﬁed via ﬂash column

chromatography (hexane, hexane–EtOAc 8 : 2, 6 : 4), yielding 586

mg (73%) of 18 as a pale yellow liquid: R_f0.20 [hexane–EtOAc

1 : 1]; n_max(ﬁlm)/cm^-12987, 2131, 1707, 1284, 1026; ¹³C NMR

(CDCl₃): d 163.34^p(C, d, J 12.0 Hz), 127.90^p(C, d, J 316.0 Hz),

63.56^p(CH₂, d, J 6.0 Hz), 61.61 (CH₂), 16.06 (CH₃, d, J 7.0 Hz),

1

14.25 (CH₃); H NMR (CDCl₃): d 4.25 (2H, q, J 7.2 Hz, H5),

4.23–4.10 (4H, m, H3), 1.34 (6H, dt, J 7.1, 0.7 Hz, H4), 1.29 (3H,

t, J 7.2 Hz, H6); m/z (EI⁺) 250 ([M]⁺, 16), 121 (100), 109 (98), 93

(73), 81 (68), 65 (87); m/z HRMS (EI⁺) [M]⁺calcd 250.0719 for

C₈H₁₅N₂O₅P found 250.0711; Elemental calcd C, 38.4; H, 6.0; N,

11.2; found C, 38.5; H, 6.0; N, 11.1.

2-Ethoxy-3-(1H-indol-5-yl)propanoic acid 2. Procedure C:

Potassium hydroxide (15.1 mg, 0.269 mmol, 1.25 eq) was added

to a solution of 15 (53.4 mg, 0.216 mmol, 1.00 eq) dissolved in

a 1 : 1 mixture of ethanol and water, and the mixture was then

Ethyl 2-(diethoxyphosphoryl)-2-(2,2,2-triﬂuoroethoxy)acetate

10. This conversion was performed according to the method

of Haigh et al.⁶⁶In brief, a mixture of diazo ester 18 (586 mg,

2.34 mmol, 1.00 eq), triﬂuoroethanol (0.340 ml, 4.68 mmol,

2.00 eq) and Rh(II) acetate dimer (10.5 mg, 0.024 mmol, 0.01 eq)

◦

stirred at 78 C for 15 h. Thereafter the whole was concentrated

in vacuo, and the concentrate subjected to vigorous stirring at 0 ^◦C

as the solution pH was adjusted to 1 with additions of aliquots

of 1 M HCl. The resulting precipitate was dissolved in DCM

and this organic solution was washed with water. Subsequently,

the combined aqueous washes were re-extracted with DCM,

and the combined DCM extracts were dried over MgSO₄then

concentrated in vacuo to give 46.2 mg (92%) of 2 as a pale yellow

oil: R_f0.45 [Solvent A]; n_max(DCM ﬁlm)/cm^-13411, 2978–2929,

1724, 1265, 1105; ¹³C NMR (CDCl₃): d 177.33 (C), 134.76 (C),

127.84 (C), 127.49 (C), 124.63 (CH), 123.25 (CH), 120.95 (CH),

110.93 (CH), 101.80 (CH), 80.27 (CH), 66.52 (CH₂), 38.91 (CH₂),

14.82 (CH₃); ¹H NMR (CDCl₃): d 10.8 (1H, br, OH), 8.47 (1H, br,

H1), 7.57 (1H, s, H4), 7.29 (1H, d, J 8.4 Hz, H7), 7.15–7.13 (2H,

m, H2, H6), 6.52 (1H, br, H3), 4.22 (1H, dd, J 7.8, 4.6 Hz, H2¢),

3.70–3.62 (1H, m, H3¢), 3.49–3.41 (1H, m, H3¢), 3.29 (1H, dd, J

14.2, 4.2 Hz, H1¢), 3.19 (1H, dd, J 14.0, 7.6 Hz, H1¢), 1.19 (3H, t, J

7.0 Hz, H4¢); m/z (EI⁺) 233 ([M]⁺, 22), 130 (M-(ethoxypropanoic

acid side chain)]⁺, 100); m/z HRMS (EI⁺) [M]⁺calcd 233.1052 for

C₁₃H₁₅NO₃, found 233.1047; Elemental calcd C, 66.9; H, 6.5; N,

6.0; found C, 67.0; H, 6.4; N, 5.9.

◦

in benzene (7.00 ml) was heated with stirring at 85 C for 22 h.

Thereafter, the solution was concentrated in vacuo to a residue

that was puriﬁed by ﬂash column chromatography (hexane–

EtOAc 8 : 2, 7 : 3, 6 : 4, 1 : 1) to yield 474 mg (63%) of 10 as a

colourless liquid: R_f0.20 [hexane–EtOAc 4 : 6]; n_max(ﬁlm)/cm^-1

1743, 1265, 1041; ¹³C NMR (CDCl₃): d 166.09 (C), 123.35 (C,

q, J 277.5 Hz), 77.11^p(CH, d, J 156.0 Hz), 68.52^p(CH₂, dq, J

35.0, 11.5 Hz), 63.99^p(CH₂, t, J 6.0 Hz), 62.21 (CH₂), 16.23^p

(CH₃, d, J 6.0 Hz), 13.99 (CH₃); ¹H NMR (CDCl₃): d 4.50 (1H,

d, J 14.4, H2), 4.37–4.08 (7H, m, H5, H7, H3), 3.98–3.89 (1H,

m, H3), 1.36–1.30 (9H, m, H6, H8); m/z (EI⁺) 322 ([M]⁺, 1),

249 ([M-ethyl]⁺, 20), 183 ([M-(triﬂuoroether + 2 ethyl groups)]⁺,

61), 155 ([M-(triﬂuoroether + 3 ethyl groups)]⁺, 100), 65 (64);

m/z HRMS (EI⁺) [M]⁺calcd 322.0793 for C₁₀H₁₈F₃O₆P found

322.0787; Elemental calcd C, 37.3; H, 5.6; found C, 37.3; H, 5.6.

(Z)-Ethyl 3-(1H-indol-5-yl)-2-(2,2,2-triﬂuoroethoxy)acrylate 13

and (E)-ethyl 3-(1H-indol-5-yl)-2-(2,2,2-triﬂuoroethoxy)acrylate

14. Procedure D: phosphonoacetate 10 (60.0 mg, 0.186 mmol,

1.06 eq) in THF (0.600 ml) was added dropwise to a suspension

of sodium hydride (9.80 mg, 0.245 mmol, 1.40 eq) in anhydrous

Ethyl 2-diazo-2-(diethoxyphosphoryl)acetate 18. The forma-

tion of tosyl azide was accomplished according to the published

procedure procedure.⁵⁸In brief, p-toluene sulfonyl chloride (14.9 g,

78.2 mmol, 1.00 eq) in acetone (70 ml) and sodium azide (5.10 g,

78.4 mmol, 1.00 eq) in ethanol (25 ml) were combined yielding

15.1 g of the desired reagent as a colourless liquid (98%) R_f

◦

THF (0.800 ml) at 0 C under a constant ﬂow of N₂. Thereafter

the reaction mixture was stirred at 0 ^◦C for 25 min, then 5-indole-

carbaldehyde 8 (25.4 mg, 0.175 mmol, 1.00 eq) was added in

dry THF (0.900 ml). Post stirring for a further 25.5 h at 0 ^◦C,

the solution was concentrated in vacuo and the residue dissolved

in EtOAc for washing with water. Combined aqueous washes

were extracted with EtOAc and the combined organic extracts

were then washed with brine, dried over MgSO₄, concentrated

in vacuo leaving a residue that was puriﬁed by ﬂash column

chromatography (hexane–EtOAc 20 : 1, 15 : 1, 10 : 1, 8 : 2) to yield

54.3 mg (99%) of two isomers. Ratio Z : E = 59 : 41. Z isomer 13:

amorphous white powder; mp 98.0–99.0 ^◦C (hexane–DCM); R_f

0.20 [hexane–EtOAc 7.5 : 2.5]; n_max(nujol)/cm^-13343, 2977–2850,

1691, 1459, 1371, 1261, 1056; ¹³C NMR (CDCl₃): d 164.05 (C),

140.54 (C), 136.37 (C), 128.08 (C), 123.43 (C, q, J 277.0 Hz),

127.34 (CH), 125.06 (CH), 124.84 (CH), 124.84 (C), 124.19 (CH),

111.16 (CH), 103.62 (CH), 67.86 (CH₂, q, J 34.5 Hz), 61.34 (CH₂),

1

0.40 [hexane–EtOAc 8 : 2]; H NMR (CDCl₃): d 7.85–7.83 (2H,

m, H2), 7.42–7.40 (2H, m, H3), 2.48 (3H, br, H1¢); m/z (EI⁺)

197 ([M]⁺, 8), 155 ([M-azide]⁺, 70), 91 ([methylbenzene]⁺, 100);

m/z HRMS (EI⁺) [M]⁺calcd 197.0259 for C₇H₇N₃O₂S, found

197.0256; Elemental calcd C, 42.6; H, 3.6; N, 21.3; found C, 42.7;

H, 3.6; N, 21.3. This next stage was accomplished according to the

procedure of Regitz et al.⁶⁰In brief triethylphosphonoacetate 19

(0.63 ml, 3.20 mmol, 1.00 eq) was added dropwise to an ice-cold

suspension of sodium hydride (157 mg, 3.92 mmol, 1.23 _◦eq) in

dry THF (8.00 ml). The reaction mixture was stirred at 0 C for

45 min after which time tosyl azide (630 mg, 3.20 mmol, 1.00 eq)

in dry THF (2 ¥ 1.00 ml) was added, under a constant ﬂow of

N₂. The reaction mixture was further stirred for 16 h at 0 ^◦C

then quenched by the addition of ether and water. Thereafter,

the aqueous phase was extracted with ether and the combined

ethereal extracts were then washed with aq NaOH (0.5 M), water

and brine. Finally the aqueous phase was extracted with ether and

1

14.29 (CH₃); H NMR (CDCl₃): d 8.26 (1H, br, H1), 8.11 (1H,

s, H4), 7.68 (1H, dd, J 8.6, 1.4 Hz, H6), 7.41 (1H, d, J 8.8 Hz,

H7), 7.27 (1H, br, H1¢), 7.24 (1H, t, J 2.8, H2), 6.61–6.60 (1H,

This journal is

The Royal Society of Chemistry 2011

Org. Biomol. Chem., 2011, 9, 1169–1188 | 1181

©

View Article Online

m, H3), 4.41–4.31 (4H, m, H3¢, H6¢), 1.40 (3H, t, J 7.2 Hz,

H7¢); m/z (EI⁺) 313 ([M]⁺, 100), 230 ([M-CH₂CF₃]⁺, 6), 202

([M-CH₂CH₃-CH₂CF₃]⁺, 39), 157 (33), 129 (44), 118 ([Indole]⁺,

13); m/z HRMS (EI⁺) [M]⁺calcd 313.0926 for C₁₅H₁₄F₃NO₃found

313.0925; Elemental calcd C, 57.7; H, 4.5; N, 4.5; found C, 57.6; H,

4.6; N, 4.5. E isomer 14: amorphous pale yellow powder; mp 88.0–

89.0 ^◦C (hexane–DCM); R_f0.15 [hexane–EtOAc 7.5 : 2.5]; n_max

(nujol)/cm^-13372, 2950–2850, 1711, 1461, 1378, 1269, 1157; ¹³C

NMR (CDCl₃): d 163.51(C), 143.55 (C), 135.52 (C), 127.76 (C),

124.77 (CH), 124.54 (C), 123.42 (CH), 123.24 (C, q, J 277.0 Hz),

121.67 (CH), 121.56 (CH), 110.62 (CH), 103.00 (CH), 68.16 (CH₂,

1-Benzenesulfonyl-1H-indole-5-carbonitrile 21. This reaction

was performed according to a procedure of Iwanowicz et al.⁶⁷

In brief, a solution of benzenesulfonyl chloride (0.850 ml,

0.688 mmol, 1.10 eq) in toluene (4.00 ml) was added dropwise

to a vigorously stirred biphasic system of indole-5-carbonitrile

20 (864 mg, 6.08 mmol, 1.00 eq) and n-tetrabutylammonium

bromide (197 mg, 0.609 mmol, 0.10 eq) in 50% aqueous NaOH

(5.90 ml), toluene (5.00 ml) and water (9.00 ml). Post 25 h

stirring at ambient temperature, the aqueous phase was removed

and the organic phase washed with 0.1 M NaHCO₃, water and

saturated brine. All aqueous extracts were extracted with EtOAc

then combined organic extracts were dried over MgSO₄and

concentrated in vacuo to give an off-white powder that was puriﬁed

via recrystallisation from hexane–DCM 20 : 1 to yield 1.65 g (96%)

of 21 as white needles: mp 131.0–133.5 ^◦C (hexane–DCM 20 : 1);

R_f0.35 [hexane–EtOAc 6 : 4]; n_max(nujol)/cm^-12222, 1458, 1375,

1080, 721; ¹³C NMR (CDCl₃): d 137.73 (C), 136.43 (C), 134.42

(CH), 130.65 (C), 129.54 (CH), 128.37 (CH), 127.59 (CH), 126.77

(CH), 126.36 (CH), 119.18 (C), 114.27 (CH), 108.65 (CH), 107.03

(C); ¹H NMR (CDCl₃): d 8.10–8.08 (1H, d, J 8.4 Hz, H2¢), 7.90–

7.89 (3H, m, H2¢, H4, H7), 7.71 (1H, d, J 3.6 Hz, H2), 7.62–7.56

(2H, m, H3¢), 7.49 (2H, t, J 7.8 Hz, H4¢, H6), 6.74 (1H, d, J 3.6 Hz,

H3); m/z (EI⁺) 282 ([M]⁺, 50), 141 ([M-BzSulfonyl group]⁺, 66),

114 ([unsubstituted indole]⁺, 12), 77 (100); m/z HRMS (EI⁺) [M]⁺

calcd 282.0463 for C₁₅H₁₀N₂O₂S found 282.0474; Elemental calcd

C, 63.8; H, 3.6; N, 9.9; found C, 63.7; H, 3.5; N, 9.9.

1

q, J 35.0 Hz), 61.29 (CH₂), 13.75 (CH₃); H NMR (CDCl₃): d

8.19 (1H, br, H1), 7.62 (1H, d, J 0.8 Hz, H4), 7.34 (1H, d, J

8.4 Hz, H7), 7.21 (1H, t, J 2.8 Hz, H2), 7.15 (1H, dd, J 8.4,

1.6 Hz, H6), 6.74 (1H, s, H1¢), 6.55–6.53 (1H, m, H3), 4.27 (2H,

q, J 8.4 Hz, H3¢), 4.20 (2H, q, J 7.2 Hz, H6¢), 1.14 (3H, t, J

7.2 Hz, H7¢); m/z (EI⁺) 313 ([M]⁺, 100), 230 ([M-CH₂CF₃]⁺,6), 202

([M-CH₂CH₃-CH₂CF₃]⁺, 38), 157 (34), 129 (48), 118 ([Indole]⁺,

15); m/z HRMS (EI⁺) [M]⁺calcd 313.0926 for C₁₅H₁₄F₃NO₃found

313.0926; Elemental calcd C, 57.7; H, 4.5; N, 4.5; found C, 57.6;

H, 4.6; N, 4.4.

Methyl 3-(1H-indol-5-yl)-2-(2,2,2-triﬂuoroethoxy)propanoate

16. This transformation was carried out as described in

procedure B using starting material 13. After puriﬁcation via

ﬂash column chromatography (hexane–EtOAc 20 : 1, 10 : 1, 9 : 1,

8 : 2) 16 was obtained as a pale yellow oil 653 mg (60%): R_f

0.20 [hexane–EtOAc 7.5 : 2.5]; n_max(DCM ﬁlm)/cm^-13377, 2933,

1741, 1513, 1278, 1164; ¹³C NMR (CDCl₃): d 171.50 (C), 134.91

(C), 128.05 (C), 123.56 (C, q, J 277.0 Hz), 127.42 (C), 124.43

(CH), 123.62 (CH), 121.25 (CH), 110.84 (CH), 102.47 (CH),

82.02 (CH), 67.89 (CH₂, q, J 34.0 Hz), 52.11 (CH₂), 39.17 (CH₂);

¹H NMR (CDCl₃): d 8.13 (1H, br, H1), 7.51 (1H, s, H4), 7.33

(1H, d, J 8.4 Hz, H7), 7.20 (1H, t, J 2.8 Hz, H2), 7.09 (1H, dd,

J 8.4, 1.6 Hz, H6), 6.52–6.51 (1H, m, H3), 4.26 (1H, dd, J 7.6, J

4.8 Hz, H2¢), 4.02–3.93 (1H, m, H3¢), 3.75 (3H, s, H6¢), 3.73–3.66

(1H, m, H3¢), 3.24 (1H, dd, J 14.0 4.8 Hz, H1¢), 3.17 (1H, dd,

J 14.2, 7.8 Hz, H1¢); m/z (EI⁺) 301 ([M]⁺, 100), 242 (65), 201

([M-CH₂CH₃-CH₂CF₃]⁺, 39), 131 (80); m/z HRMS (EI⁺) [M]⁺

calcd 301.0926 for C₁₄H₁₄F₃NO₃found 301.0923; Elemental calcd

C, 55.8; H, 4.7; N, 4.7; found C, 55.9; H, 4.7; N, 4.6.

2-Benzyl-1-(phenylsulfonyl)-1H-indole-5-carbonitrile 22. This

reaction was performed by adaptation ofaprocedure ofIwanowicz

et al.⁶⁷In brief, a solution of 2 M LDA solution in pentane

(0.24 ml, 0.470 mmol, 1.20 eq) was added dropwise to a solution

of phenylsulfonyl indole 21 (109 mg, 0.39 mmol, 1.00 eq) in

anhydrous THF (4 ml) at -78 ^◦C, under a constant ﬂow of

N₂. The reaction mixture was stirred at -78 ^◦C for 55 min,

followed by the dropwise addition of a solution of benzyl bromide

(70 mL, 0.585 mmol, 1.5 eq) in dry THF (2 ml). The reaction

mixture was allowed to warm up to 0 ^◦C with stirring over

50 min, then to ambient temperature over 15 h. Thereafter, the

mixture was quenched with 5% citric acid and the organic phase

washed with water, then aqueous extracts were extracted with

ethylacetate. Combined organic extracts were dried over MgSO₄

and concentrated in vacuo, giving a residue that was puriﬁed

via ﬂash column chromatography (hexane–EtOAc 10 : 1) to yield

56.2 mg (66.5%) of the product as yellow oil. Crystallisation

from hexane–DCM 20 : 1 yielded 22 as white needles: mp 130.5–

131.5 ^◦C (hexane–DCM 20 : 1); R_f0.20 [hexane–EtOAc 8 : 2]; n_max

(nujol)/cm^-12220, 1460, 1371, 1053; ¹³C NMR (CDCl₃): d 143.63

(C), 138.98 (C), 138.59 (C), 136.99 (C), 134.19 (CH), 129.46 (CH),

129.38 (C), 129.33 (CH), 128.69 (CH), 127.18 (CH), 127.03 (CH),

126.39 (CH), 125.04 (CH), 119.32 (C), 115.35 (CH), 110.07 (CH),

3-(1H-indol-5-yl)-2-(2,2,2-triﬂuoroethoxy)propanoic acid 3.

The reaction was performed according to procedure C, using

starting material 16. This process afforded 38.1 mg (93%) of 3 as

a pale yellow oil: R_f0.25 [Solvent A]; n_max(DCM ﬁlm)/cm^-13415,

2933, 1726, 1279, 1167; ¹³C NMR (CDCl₃/MeOD 2 : 1): d 173.97

(C), 135.68 (C), 128.53 (C), 124.18 (C, q, J 277.0 Hz), 127.67

(CH), 125.20 (CH), 123.52 (CH), 121.31 (CH), 111.40 (CH),

101.73 (CH), 82.44 (CH), 68.18 (CH₂, q, J 34.5 Hz), 39.63 (CH₂);

¹H NMR (CDCl₃/MeOD 2 : 1): d 9.85 (1H, br, OH), 7.45 (1H, s,

H4), 7.28 (1H, d, J 8.0 Hz, H7), 7.14–7.13 (1H, m, H2), 7.03 (1H,

dd, J 8.4, 1.6 Hz, H6), 6.38–6.37 (1H, m, H3), 4.19 (1H, dd, J

8.4, 4.0 Hz, H2¢), 3.97–3.87 (1H, m, H3¢), 3.70–3.60 (1H, m, H3¢),

3.18 (1H, dd, J 14.2, 4.2 Hz, H1¢), 3.06 (1H, dd, J 14.2, 8.2 Hz,

H1¢); m/z (EI⁺) 287 ([M]⁺, 18), 130 ([M-(ethoxypropanoic acid

side chain)]⁺, 100); m/z HRMS (EI⁺) [M]⁺calcd 287.0769 for

C₁₃H₁₂NO₃F₃found 287.0763; Elemental calcd C, 54.4; H, 4.2; N,

4.9; found C, 54.7; H, 4.2; N, 4.8.

1

107.10 (C), 35.12 (CH₂); H NMR (CDCl₃): d 8.28 (1H, d, J

8.8 Hz, H7), 7.71 (1H, s, H4), 7.66 (2H, d, J 7.6 Hz, H2¢), 7.58

(1H, t, J 7.6 Hz, H4¢), 7.54 (1H, dd, J 8.8, 1.2 Hz, H6), 7.42 (2H,

t, J 8.8 Hz, H3¢), 7.34–7.29 (3H, m, H8¢, H9¢), 7.20–7.18 (2H, m,

H7¢), 6.16 (1H, s, H3), 4.37 (2H, s, H5¢); m/z (EI⁺) 372 ([M]⁺, 57),

284 ([M-Bn group]⁺, 30), 230 ([M-Phsulfonyl group]⁺, 100), 141

(11), 115 ([indole]⁺, 5), 91 (12), 77 (47); m/z HRMS (EI⁺) [M]⁺

calcd 372.0932 for C₂₂H₁₆N₂O₂S found 372.0934; Elemental calcd

C, 71.0; H, 4.3; N, 7.5; found C, 71.0; H, 4.2; N, 7.6. Crystal data

1182 | Org. Biomol. Chem., 2011, 9, 1169–1188

This journal is

The Royal Society of Chemistry 2011

©

View Article Online

for 22: C₂₂H₁₆N₂O₂S, M = 372.43, monoclinic, P2₁/n (no. 14), a_◦=

167 (30), 158 (42), 149 (30), 91 (14); m/z HRMS (EI⁺) [M]⁺calcd

235.0997 for C₁₆H₁₃NO found 235.0997; Elemental calcd C, 81.7;

H, 5.6; N, 6.0; found C, C, 81.6; H, 5.5; N, 5.9.

˚

10.8591(3), b = 10.0425(3), c = 17.0542(4) A, b = 100.888(3) ,

3

V = 1826.32(9) A , Z = 4, D_c= 1.354 g cm^-3, m(Mo-Ka) =

˚

0.197 mm^-1, T = 173 K, colourless blocks, Oxford Diffraction

Xcalibur 3 diffractometer; 4231 independent measured reﬂections,

F²reﬁnement, R₁= 0.044, wR₂= 0.101, 2710 independent observed

absorption-corrected reﬂections [|F_o| > 4s(|F_o|), 2q_max= 58^◦],

245 parameters. CCDC 667594.

(Z)-Ethyl 3-(2-benzyl-1H-indol-5-yl)-2-ethoxyacrylate 28 and

(E)-ethyl 3-(2-benzyl-1H-indol-5-yl)-2-ethoxyacrylate 29. This

reaction was carried out according to procedure A using the

starting materials 26 and 9. After puriﬁcation via ﬂash column

chromatography (hexane–EtOAc 15 : 1, 12 : 1) the product could

be obtained as a colourless oil, yielding 66.0 mg (90%) of two

isomers. Ratio Z : E = 63 : 37. Z isomer 28: colourless oil: R_f0.20

[hexane–EtOAc 8 : 2]; n_max(DCM ﬁlm)/cm^-13441, 3055, 1743,

1646, 1051, 740; ¹³C NMR (CDCl₃): d 165.31 (C), 142.43 (C),

138.67 (C), 138.21 (C), 136.56 (C), 128.81 (CH), 128.77 (CH),

126.82 (CH), 126.30 (CH), 125.53 (C), 124.08 (CH), 122.71 (CH),

110.43 (CH), 101.73 (CH), 67.41 (CH₂), 60.87 (CH₂), 34.69 (CH₂),

15.56 (CH₃), 14.36 (CH₃); ¹H NMR (CDCl₃): d 8.03 (1H, br, H4),

7.89 (1H, br, H1), 7.63 (1H, dd, J 8.4, 1.6 Hz, H6), 7.37–7.33

(2H, m, H4¢), 7.30–7.22 (4H, m, H3¢, H7, H5¢), 7.17 (1H, s, H1¢¢),

6.35 (1H, d, J 0.8 Hz, H3), 4.32 (2H, q, J 7.2 Hz, H6¢¢), 4.14

(2H, s, H1¢), 4.02 (2H, q, J 7.1 Hz, H3¢¢), 1.42–1.37 (6H, m, H7¢¢,

H4¢¢); m/z (EI⁺) 349 ([M]⁺, 100), 292 ([M-ethyl]⁺, 17), 292 (32),

247 (36), 219 (26), 91 (50); m/z HRMS (EI⁺) [M]⁺calcd 349.1678

for C₂₂H₂₃NO₃found 349.1672; Elemental calcd C, 75.6; H, 6.6;

N, 4.0; found C, 75.6; H, 6.5; N, 3.9. E isomer 29: pale yellow oil:

R_f0.15 [hexane–EtOAc 8 : 2]; n_max(DCM ﬁlm)/cm^-13441, 3055–

2850, 1725, 1644, 1548, 1266, 1043, 738; ¹³C NMR (CDCl₃): d

165.20 (C), 146.16 (C), 138.41 (C), 138.16 (C), 135.44 (C), 128.81

(CH), 128.74 (CH), 126.76 (CH), 126.26 (C), 122.40 (CH), 119.96

(CH), 111.60 (CH), 110.00 (CH), 101.20 (CH), 64.63 (CH₂), 61.05

(CH₂), 34.73 (CH₂), 14.58 (CH₃), 13.77 (CH₃); ¹H NMR (CDCl₃):

d 7.75 (1H, br, H1), 7.40 (1H, s, H4), 7.36–7.32 (2H, m, H4¢), 7.28–

7.25 (3H, m, H3¢, H5¢), 7.15 (1H, d, J 8.4 Hz, H7), 6.98 (1H, dd,

J 8.4, 1.2 Hz, H6), 6.29 (1H, s, H1¢¢), 6.27 (1H, br, H3), 4.18–4.12

(4H, m, H6¢¢, H1¢), 3.96 (2H, q, J 6.9 Hz, H3¢¢), 1.44–1.41 (3H, t,

J 7.0 Hz, H4¢¢), 1.09 (3H, t, J 7.0 Hz, H7¢¢); m/z (EI⁺) 349 ([M]⁺,

16), 232 ([M-(benzyl+ethyl)]⁺, 100), 155 (77), 84 (42), 49 (56);

m/z HRMS (EI⁺) [M+H]⁺calcd 349.1678 for C₂₂H₂₃NO₃found

349.1671; Elemental calcd C, 75.6; H, 6.6; N, 4.0; found C, 75.7;

H, 6.8; N, 3.9.

2-Benzyl-1H-indole-5-carbonitrile 24. This transformation

was performed according to procedure E, adapted from protocols

of Iwanowicz et al.⁶⁷In brief 2-benzyl indole 22 (141 mg,

0.379 mmol, 1.00 eq) was reﬂuxed at 110 ^◦C for 18 h in methanolic

2 M NaOH, then the reaction mixture was allowed to cool to

ambient temperature before acidiﬁcation to pH 2 using aliquots

of 0.1 M HCl. The aqueous phase was extracted with DCM, then

organic extracts were dried using MgSO₄, concentrated in vacuo

leaving a residue that was puriﬁed via ﬂash column chromatog-

raphy (hexane, hexane–EtOAc 10 : 1, 9 : 1, 8 : 2) giving 87.0 mg

(99%) of 24 as an off-white solid: mp 119.5–120.5 ^◦C (hexane–

DCM 20 : 1); R_f0.25 [hexane–EtOAc 7 : 3]; n_max(nujol)/cm^-13300,

3056–2850, 2229, 1373, 1056, 740; ¹³C NMR (CDCl₃): d 140.38

(C), 137.94 (C), 137.53 (C), 128.94 (CH), 128.82 (CH), 128.49

(C), 127.10 (CH), 125.40 (CH), 124.49 (CH), 120.82 (C), 111.26

(CH), 102.84 (C), 101.66 (CH), 34.59 (CH₂); ¹H NMR (CDCl₃):

d 8.12 (1H, br, H1), 7.88 (1H, s, H4), 7.38–7.25 (6H, m, H4¢, H7,

H5¢, H6, H3¢), 6.41 (1H, s, H3), 4.16 (2H, s, H5¢); m/z (EI⁺) 232

([M]⁺, 100), 231 (64), 155 (71), 91 (16), 84 (21), 49 (23), 43 (25); m/z

HRMS (EI⁺) [M]⁺calcd 232.1000 for C₁₆H₁₂N₂found 232.1002;

Elemental calcd C, 82.7; H, 5.2; N, 12.1; found C, 82.9; H, 5.1; N,

11.9.

2-Benzyl-1H-indole-5-carbaldehyde 26. This reaction was car-

ried out according to procedure F: in brief a DIBAL solution,

1.0 M in DCM (2.20 ml, 2.20 mmol, 1.24 eq) was added dropwise

under a constant ﬂow of N₂to a solution of 2-benzyl-1H-indole-5-

carbonitrile 24 (411 mg, 1.77 mmol, 1.00 eq) in anhydrous DCM

(8.00 ml) at 0 ^◦C. The reaction mixture was allowed to warm

with stirring to ambient temperature over 16 h. Thereafter, the

mixture was cooled to 0 ^◦C and Rochelle solution (1.0 M, 17.0 ml)

was added. Post 4.5 h of stirring at ambient temperature, the two

phases were separated and the aqueous phase was extracted with

DCM. Combined organic extracts were concentrated in vacuo and

the resulting orange slurry was dispersed in further DCM for

acidiﬁcation with HCl (1 M, 2.90 ml). After stirring for 1.7 h

at ambient temperature, the mixture was ﬁnally quenched using

NaOH (2 M, 2.90 ml) and extracted with DCM. Combined

organic extracts were washed with water and brine, then dried

over MgSO₄and concentrated in vacuo to give an orange powder

that was puriﬁed by ﬂash column chromatography (hexane–

EtOAc 10 : 1, 9 : 1) to yield 332 mg (80%) of 26 as a pale orange

amorphous powder: mp 126.5–127.5 ^◦C (hexane–DCM 20 : 1); R_f

0.20 [hexane–EtOAc 7.5 : 2.5]; n_max(nujol)/cm^-13324, 3056–2850,

1671, 1373, 1056, 739; ¹³C NMR (CDCl₃): d 192.47 (CH), 139.95

(C), 139.88 (C), 137.76 (C), 129.67 (CH), 128.87 (CH), 128.81

(CH), 128.54 (C), 127.00 (CH), 125.05 (CH), 122.10 (CH), 111.04

(CH), 102.58 (CH), 34.65 (CH₂); ¹H NMR (CDCl₃): d 10.01 (1H,

s, H1¢¢), 8.08 (2H, br, H1, H4), 7.70 (1H, dd, J 8.6, 1.4 Hz, H6),

7.38–7.28 (6H, m, H4¢, H7, H5¢, H3¢), 6.49 (1H, br, H3), 4.17

(2H, s, H5¢); m/z (EI⁺) 235 ([M]⁺, 100), 206 ([M-aldehyde]⁺, 27),

Methyl 3-(2-benzyl-1H-indol-5-yl)-2-ethoxypropanoate 36.

The reaction was performed according to procedure B using 28.

After puriﬁcation via ﬂash column chromatography (hexane–

EtOAc 15 : 1, 10 : 1) 40.6 mg (76%) of 36 was obtained as a

pale yellow oil: R_f0.25 [hexane–EtOAc 7.5 : 2.5]; n_max(DCM

ﬁlm)/cm^-13392, 3027–2900, 1741, 1644, 1446, 1115, 705; ¹³C

NMR (CDCl₃): d 173.28 (C), 138.51 (C), 138.03 (C), 135.28 (C),

128.83 (C), 128.80 (CH), 128.68 (CH), 128.24 (C), 126.69 (CH),

122.91 (CH), 120.44 (CH), 110.16 (CH), 100.86 (CH), 81.04

(CH), 66.19 (CH₂), 51.73 (CH₂), 39.55 (CH₂), 34.72 (CH₂), 15.06

(CH₃); ¹H NMR (CDCl₃): d 7.77 (1H, br, H1), 7.41 (1H, br, H4),

7.36–7.32 (2H, m, H4¢), 7.29–7.26 (3H, m, H3¢, H5¢), 7.16 (1H, d,

J 8.4 Hz, H7), 7.01 (1H, dd, J 8.4, 1.8 Hz, H6), 6.28 (1H, s, H3),

4.12 (2H, s, H1¢), 4.09 (1H, t, J 6.6 Hz, H2¢¢), 3.71 (3H, s, H6¢¢),

3.64–3.56 (1H, m, H3¢¢), 3.41–3.34 (1H, m, H3¢¢), 3.11–3.09 (2H,

m, H1¢¢), 1.18 (3H, t, J 6.8 Hz, H6¢¢); m/z (EI⁺) 337 ([M]⁺, 37), 234

([M-(benzyl + methyl groups)]⁺, 81), 220 ([M-(benzyl+2 methyl

groups)]⁺,100), 91 (29); m/z HRMS (EI⁺) [M]⁺calcd 337.1668 for

This journal is

The Royal Society of Chemistry 2011

Org. Biomol. Chem., 2011, 9, 1169–1188 | 1183

©

View Article Online

C₂₁H₂₃NO₃found 337.1674; Elemental calcd C, 74.8; H, 6.9; N,

4.2; found C, 79.9; H, 5.2.; N, 4.2.

Methyl

3-(2-benzyl-1H-indol-5-yl)-2-(2,2,2-triﬂuoroethoxy)-

propanoate 37. The reaction was performed according to

procedure B using 30. After puriﬁcation via ﬂash column

chromatography (hexane–EtOAc 15 : 1, 10 : 1, 9 : 1, 8 : 2) the

desired product 48.0 mg (73%) of 37 was obtained as a very pale

yellow oil: R_f0.15 [hexane–EtOAc 8 : 2]; n_max(DCM ﬁlm)/cm^-1

3361, 2920–2846, 1741, 1448, 1275, 1157, 790; ¹³C NMR (CDCl₃):

d 171.50 (C), 138.43 (C), 138.16 (C), 135.37 (C), 128.86 (C),

128.81 (CH), 128.70 (CH), 123.74 (C, q, J 278.0 Hz), 127.28 (C),

126.72 (CH), 122.85 (CH), 120.54 (CH), 110.29 (CH), 100.88

(CH), 82.02 (CH), 67.85 (CH₂, q, J 34.5 Hz), 52.08 (CH₃), 39.17

(CH₂), 34.69 (CH₂); ¹H NMR (CDCl₃): d 7.76 (1H, br, H1), 7.40

(1H, s, H4), 7.36–7.32 (2H, m, H4¢), 7.28–7.25 (3H, m, H3¢, H5¢),

7.17 (1H, d, J 8.4 Hz H7), 6.99 (1H, dd, J 8.2, 1.4 Hz, H6), 6.28

(1H, br, H3), 4.26–4.23 (1H, m, H2¢¢), 4.13 (2H, s, H1¢), 3.99–3.93

(1H, m, H3¢¢), 3.74–3.66 (4H, m, H6¢¢, H3¢¢), 3.21–3.11 (2H, m,

H1¢¢); m/z (EI⁺) 391 ([M]⁺, 27), 220 ([M-propanoic acid side

chain]⁺, 100), 49 (23); m/z HRMS (EI⁺) [M]⁺calcd 391.1395 for

C₂₁H₂₀F₃NO₃found 391.1394; Elemental calcd C, 64.4; H, 5.2; N,

3.6; found C, 64.5; H, 5.1; N, 3.5.

3-(2-Benzyl-1H-indol-5-yl)-2-ethoxy-propanoic acid 4. The

synthesis was performed according to procedure C using 36. This

resulted in 29.8 mg (93%) of 4 as a light yellow oil: R_f0.20 [hexane–

EtOAc 1 : 1]; n_max(DCM ﬁlm)/cm^-13620, 3398, 1710, 1646, 1461,

1108; ¹³C NMR (CDCl₃): d 175.40 (C), 138.45 (C), 138.15 (C),

135.35 (C), 128.86 (C), 128.81 and 128.70 (CH each), 127.69 (C),

126.71 (CH), 122.95 (CH), 120.63 (CH), 110.27 (CH), 100.89

(CH), 80.36 (CH), 66.82 (CH₂), 38.85 (CH₂), 34.71 (CH₂), 15.05

(CH₂); ¹H NMR (CDCl₃): d 7.77 (1H, br, H1), 7.42 (1H, s, H4),

7.36–7.32 (2H, m, H4¢), 7.29–7.25 (3H, m, H3¢, H5¢), 7.17 (1H,

d, J 8.4 Hz, H7), 7.02 (1H, dd, J 8.2, 1.4 Hz, H6), 6.28 (1H, s,

H3), 4.14–4.12 (3H, m, H2¢¢, H1¢), 3.63–3.56 (1H, m, H3¢¢), 3.49–

3.42 (1H, m, H3¢¢), 3.22 (1H, dd, J 14.2, 4.2 Hz, H1¢¢), 3.10 (1H,

dd, J 14.2, 7.8 Hz, H1¢¢), 1.18 (3H, t, J 7.0 Hz, H4¢¢); m/z (EI⁺)

323 ([M]⁺, 16), 220 ([M-(benzyl+methyl group)]⁺, 88), 199 (29),

171 (M-(benzyl + ether side chain + OH)]⁺, 100), 127 (51), 57

(53); m/z HRMS (EI⁺) [M]⁺calcd 323.1521 for C₂₀H₂₁NO₃found

323.1520; Elemental calcd C, 74.3; H, 6.6; N, 4.3; found C, 74.2;

H, 6.5; N, 4.4.

3-(2-Benzyl-1H-indol-5-yl)-2-(2,2,2-triﬂuoroethoxy)propanoic

acid 5. The synthesis was performed according to procedure C

using 37. This gave product 5 14.0 mg (99%) as a pale yellow oil:

R_f0.20 [hexane–EtOAc 1 : 9]; n_max(DCM ﬁlm)/cm^-13289, 2950–

2850, 1702, 1651, 1425, 1053; ¹³C NMR (CDCl₃): d 174.74 (C),

138.37 (C), 138.31 (C), 135.46 (C), 128.93 (C), 128.83 (CH), 128.74

(CH), 123.42 (C, q, J 277.0 Hz), 126.92 (C), 126.77 (CH), 122.86

(CH), 120.67 (CH), 110.43 (CH), 100.93 (CH), 81.63 (CH), 68.13

(CH₂, q, J 34.5 Hz), 38.98 (CH₂), 34.72 (CH₂); ¹H NMR (CDCl₃):

d 7.76 (1H, br, H1), 7.43 (1H, s, H4), 7.34–7.32 (2H, m, H4¢), 7.28–

7.25 (3H, m, H3¢, H5¢), 7.18 (1H, d, J 8.4 Hz, H7), 7.02 (1H, dd,

J 8.2, 1.4 Hz, H6), 6.29 (1H, d, J 0.8 Hz, H3), 4.29 (1H, dd, J 8.0,

4.4 Hz, H2¢¢), 4.13 (2H, s, H1¢), 3.96–3.87 (1H, m, H3¢¢), 3.77–3.70

(1H, m, H3¢¢), 3.28 (1H, dd, J 14.2, 4.2 Hz, H1¢¢), 3.15 (1H, dd, J

14.0, 8.0 Hz, H1¢¢); m/z (EI⁺) 377 ([M]⁺, 30), 220 ([M-propanoic

acid side chain]⁺, 100), 142 (9), 91 (12); m/z HRMS (EI⁺) [M]⁺

calcd 377.1239 for C₂₀H₁₈F₃NO₃found 377.1244; Elemental calcd

C, 63.7; H, 4.8; N, 3.7; found C, 63.5; H, 4.7; N, 3.7.

(Z)-Ethyl 3-(2-benzyl-1H-indol-5-yl)-2-(2,2,2-triﬂuoroethoxy)-

acrylate 30 and (E)-ethyl 3-(2-benzyl-1H-indol-5-yl)-2-(2,2,2-

triﬂuoroethoxy)acrylate 31. The synthesis was carried out ac-

cording to general procedure D using the starting materials 26 and

10. After puriﬁcation via ﬂash column chromatography (hexane–

EtOAc 15 : 1, 10 : 1), 107 mg (85%) of two isomers was obtained.

Ratio Z : E = 39 : 61. Z isomer 30: yellow amorphous powder;

mp 120.5–121.5 ^◦C (hexane–DCM 20 : 1); R_f0.30 [hexane–EtOAc

7.5 : 2.5]; n_max(DCM ﬁlm)/cm^-13441, 1743, 1548, 1371, 1265,

1055; ¹³C NMR (CDCl₃): d 164.08 (C), 140.33 (C), 138.91 (C),

138.10 (C), 136.91 (C), 128.88 (C), 128.81 (CH), 128.80 (CH),

123.42 (C, q, J 277.0 Hz), 127.52 (CH), 126.87 (CH), 124.33

(CH), 124.24 (C), 123.36 (CH), 110.61 (CH), 101.85 (CH), 67.80

1

(CH₂, q, J 35.0 Hz), 61.30 (CH₂), 34.66 (CH₂), 14.28 (CH₃); H

NMR (CDCl₃): d 8.02 (1H, s, H4), 7.90 (1H, br, H1), 7.59 (1H, dd,

J 8.6, 1.4 Hz, H6), 7.37–7.34 (2H, m, H4¢), 7.30–7.24 (5H, m, H3¢,

H5¢, H1¢¢, H7), 6.37 (1H, d, J 1.2 Hz, H3), 4.40–4.30 (4H, m, H3¢¢,

H6¢¢), 4.14 (2H, s, H1¢), 1.40 (3H, t, J 7.0 Hz, H7¢¢); m/z (EI⁺) 403

([M]⁺, 45), 252 (25), 221 ([M-propanoic acid side chain]⁺, 100),

91 (27); m/z HRMS (EI⁺) [M]⁺calcd 403.1395 for C₂₂H₂₀F₃NO₃

found 403.1392; Elemental calcd C, 65.5; H, 5.0; N, 3.5; found C,

65.4; H, 4.9; N, 3.5. E isomer 31: pale yellow oil: R_f0.25 [hexane–

EtOAc 7.5 : 2.5]; n_max(DCM ﬁlm)/cm^-13441, 1743, 1548, 1265,

1053; ¹³C NMR (CDCl₃): d 163.53 (C), 143.28 (C), 138.55 (C),

138.26 (C), 136.03 (C), 128.79 (CH), 128.74 (CH), 128.54 (C),

123.23 (C, q, J 277.0 Hz), 126.79 (CH), 124.31 (C), 122.75 (CH),

121.76 (CH), 120.95 (CH), 110.08 (CH), 101.32 (CH), 67.64 (CH₂,

2-(Naphthalen-2-ylmethyl)-1-(phenylsulfonyl)-1H-indole-5-car-

bonitrile 23. This reaction was performed by adaptation of a

procedure of Iwanowicz et al.⁶⁷in a similar way to the synthesis of

22 from 21, but replacing a solution of benzyl bromide with a solu-

tion of 2-(bromomethyl)-naphthalene. After the work up, the oily

residue was puriﬁed via ﬂash column chromatography (hexane–

EtOAc 15 : 1) to yield 23 (30%) as yellow oil. Crystallisation

from hexane–DCM 20 : 1 afforded shiny white needles: mp 142.0–

143.0 ^◦C (hexane–DCM 20 : 1); R_f0.30 [hexane–EtOAc 7 : 3]; n_max

(nujol)/cm^-12950–2850, 2222, 1651, 1548, 1379, 1055; ¹³C NMR

(CDCl₃): d 143.41 (C), 139.07 (C), 138.57 (C), 134.41 (C), 134.09

(CH), 133.47 (C), 132.44 (C), 129.34 (CH), 128.36 and 127.95

(CH), 127.65, 127.61, 127.45, 127.25, 126.29, 126.25, 125.92,

125.08 (CH), 119.32 (C), 115.38 (CH), 110.30 (CH), 107.14 (C),

1

q, J 35.0 Hz), 61.26 (CH₂), 34.65 (CH₂), 13.78 (CH₃); H NMR

(CDCl₃): d 7.84 (1H, br, H1), 7.53 (1H, s, H4), 7.37–7.33 (2H, m,

H3¢), 7.30–7.25 (3H, m, H4¢, H5¢), 7.17 (1H, d, J 8.4 Hz, H7),

7.09 (1H, dd, J 8.4, 1.6 Hz, H6), 6.71 (1H, s, H1¢¢), 3.61 (1H, br,

H3), 4.27 (2H, q, J 16.8, 8.4 Hz, H3¢¢), 4.20 (2H, q, J 7.2 Hz,

H6¢¢), 4.12 (2H, s, H1¢), 1.17 (3H, t, J 7.0 Hz, H7¢¢); m/z (EI⁺) 403

([M]⁺, 45), 252 (25), 221 ([M-propanoic acid side chain]⁺, 100),

91 (27); m/z HRMS (EI⁺) [M]⁺calcd 403.1395 for C₂₂H₂₀F₃NO₃

found 403.1392; Elemental calcd C, 65.5; H, 5.0; N, 3.5; found C,

65.5; H, 5.0; N, 3.4.

1

35.24 (CH₂); H NMR (CDCl₃): d 8.31 (1H, d, J 8.8 Hz, H7),

7.86–7.83 (1H, m), 7.79 (1H, d, J 8.4 Hz), 7.73–7.71 (2H, m), 7.63

(2H, dd, J 8.6, 1.0 Hz), 7.59 (1H, br, H4), 7.56 (1H, dd, J 8.8,

1.6 Hz, H6), 7.50–7.46 (3H, m, H3¢), 7.32–7.28 (3H, m), 6.21 (1H,

br, H3), 4.53 (2H, s, H5¢); m/z (EI⁺) 422 ([M]⁺, 61), 280 ([M-methyl

naphthyl group]⁺, 100), 141 (34), 77 (34); m/z HRMS (EI⁺) [M]⁺

1184 | Org. Biomol. Chem., 2011, 9, 1169–1188

This journal is

The Royal Society of Chemistry 2011

©

View Article Online

calcd 422.1089 for C₂₆H₁₈N₂O₂S found 422.1087; Elemental calcd

C, 73.9; H, 4.3; N, 6.6; found C, 74.0; H, 4.2; N, 6.6. Crystal data

for 23: C₂₆H₁₈N₂O₂S, M = 422.48, monoclinic, P2₁/n (no. 14), a =

(1H, dd, J 8.6, 1.0 Hz, H6), 7.52–7.47 (2H, m), 7.38 (1H, dd, J 8.4,

1.2Hz), 7.23–7.21 (1H, d, J 8.4, H7), 7.19 (1H, s), 6.41 (1H, br, H3),

4.34–4.29 (4H, m, H6¢¢, H1¢), 4.03 (2H, q, J 7.0 Hz, H3¢¢), 1.43–

1.37 (6H, m, H4¢¢, H7¢¢); m/z (EI⁺) 399 ([M]⁺, 78), 342 ([M-both

ethyl groups]⁺, 25), 297 (24), 141 (100); m/z HRMS (EI⁺) [M]⁺

calcd 399.1834 for C₂₆H₂₅NO₃found 399.1830; Elemental calcd C,

78.2; H, 6.3; N, 3.5; found C, 78.3; H, 6.2; N, 3.6. E isomer 33:

colourless oil: R_f0.30 [hexane–EtOAc 7 : 3]; n_max(DCM ﬁlm)/cm^-1

3392, 3050–2900, 1712, 1635, 1228, 1153, 740; ¹³C NMR (CDCl₃):

d 165.19 (C), 146.09, 138.06, 135.90, 135.46, 133.52, 132.33, 128.67

(C), 128.39, 127.66, 127.52, 127.19, 127.07, 126.23 (CH), 126.20

(C), 125.69 (CH), 122.37 (CH), 119.92 (CH), 111.61 (CH), 110.05

(CH), 101.25 (CH), 64.60 (CH₂), 61.04 (CH₂), 34.70 (CH₂), 14.53

(CH₃), 13.73 (CH₃); ¹H NMR (CDCl₃): d 7.85–7.79 (4H, m), 7.71

(1H, s), 7.52–7.45 (2H, m), 7.42 (1H, s, H4), 7.37 (1H, dd, J 8.4,

1.6 Hz), 7.13 (1H, d, J 8.4 Hz, H7), 6.99 (1H, dd, J 8.4, 1.6 Hz,

H6), 6.33 (1H, d, J 0.8 Hz, H3), 6.29 (1H, s, H1¢¢), 4.28 (2H, s, H1¢),

4.15 (2H, q, J 7.2 Hz, H6¢¢), 3.95 (2H, q, J 13.8, 7.0 Hz, H3¢¢), 1.42

(3H, t, J 6.8 Hz, H4¢¢), 1.09 (3H, t, H7¢¢); m/z (EI⁺) 399 ([M]⁺,

72), 342 ([M-both ethyl groups]⁺, 24), 282 (34), 141 (100); m/z

HRMS (EI⁺) [M]⁺calcd 399.1834 for C₂₆H₂₅NO₃found 399.1832;

Elemental calcd C, 78.2; H, 6.3; N, 3.5; found C, 78.3; H, 6.2; N,

3.6.

◦

˚

18.0309(4), b = 5.88641(15), c = 19.5220(4) A, b = 98.636(2) , V =

3

-3

-1

˚

2048.52(19) A , Z = 4, D_c= 1.370 g cm , m(Cu-Ka) = 1.615 mm ,

T = 173 K, colourless needles, Oxford Diffraction Xcalibur PX

Ultra diffractometer; 3230 independent measured reﬂections, F²

reﬁnement, R₁= 0.040, wR₂= 0.076, 1894 independent observed

absorption-corrected reﬂections [|F_o| > 4s(|F_o|), 2q_max= 126^◦],

280 parameters. CCDC 667595.

2-(Naphthalen-2-ylmethyl)-1H-indole-5-carbonitrile 25. The

reaction was carried out according to procedure E using starting

material 23. This procedure gave 335 mg (99%) of 25 as an

amorphous white powder: mp 152.0–153.0 ^◦C (hexane–DCM

20 : 1); R_f0.25 [hexane–EtOAc 7 : 3]; n_max(nujol)/cm^-13206, 2980–

2850, 2215, 1461, 1371, 1053; ¹³C NMR (CDCl₃): d 140.25 (C),

137.95 (C), 134.96 (C), 133.54 (C), 132.47 (C), 128.76 (CH), 128.51

(C), 127.74, 127.54, 127.30, 127.23, 126.99, 126.47, 125.99, 125.44,

124.53 (CH), 120.81 (C), 111.28 (CH), 102.88 (C), 101.77 (CH),

34.79 (CH₂); ¹H NMR (CDCl₃): d 8.13 (1H, br, H1), 7.89 (1H, s),

7.86–7.80 (3H, m), 7.73 (1H, s, H4), 7.54–7.47 (2H, m), 7.37–7.34

(2H, m, H6), 7.27 (1H, d, J 8.4 Hz, H7), 6.46 (1H, d, J 0.8 Hz,

H3), 4.33 (2H, s, H1¢); m/z (EI⁺) 282 ([M]⁺, 100), 155 (52); m/z

HRMS (EI⁺) [M]⁺calcd 282.1157 for C₂₀H₁₄N₂found 282.1142;

Elemental calcd C, 85.1; H, 5.0; N, 9.9; found C, 85.1; H, 4.9; N,

9.9.

Methyl 2-ethoxy-3-(2-(naphthalen-2-ylmethyl)-1H-indol-5-yl)-

propanoate 38. The reaction was performed according to pro-

cedure B using 32. After puriﬁcation via ﬂash column chromatog-

raphy (hexane–EtOAc 10 : 1, 9 : 1, 8 : 2) 85.0 mg (80%) of 38 was

obtained as a light yellow oil: R_f0.15 [hexane–EtOAc 8 : 2]; n_max

(DCM ﬁlm)/cm^-13392, 3050–2900, 1741, 1643, 1273, 1114, 740;

¹³C NMR (CDCl₃): d 173.30 (C), 137.89, 135.97, 135.30, 133.53,

133.18, 132.33, 128.90 (C), 128.41 (CH), 128.26 (C), 127.72,

127.52, 127.24, 127.09, 126.35, 125.69 (CH), 122.95 (CH), 120.47

(CH), 110.19 (CH), 101.02 (CH), 81.02 (CH), 66.20 (CH₂), 51.77

(CH₃), 39.53 (CH₂), 34.92 (CH₂), 15.07 (CH₃); ¹H NMR (CDCl₃):

d 7.84–7.79 (4H, m), 7.71 (1H, s, H1), 7.51–7.45 (2H, m), 7.42 (1H,

s, H4), 7.38 (1H, dd, J 8.4, 1.6 Hz), 7.14 (1H, d, J 8.4 Hz, H7), 7.00

(1H, dd, J 8.0, 1.6 Hz, H6), 6.33 (1H, d, J 1.2 Hz, H3), 4.28 (2H,

s, H1¢), 4.09 (1H, dd, J 7.2, 6.0 Hz, H2¢¢), 3.71 (3H, s, H6¢¢), 3.63–

3.58 (1H, m, H3¢¢), 3.40–3.33 (1H, m, H3¢¢), 3.11 (1H, s, H1¢¢), 3.09

(1H, d, J 1.6 Hz, H1¢¢), 1.17 (3H, t, J 6.8 Hz, H4¢¢); m/z (EI⁺) 387

([M]⁺, 27), 270 ([M-side chain on 5-position]⁺, 100), 146 (67); m/z

HRMS (EI⁺) [M]⁺calcd 387.1834 for C₂₅H₂₅NO₃found 387.1835;

Elemental calcd C, 77.5; H, 6.5; N, 3.6; found C, 77.5; H, 6.4; N,

3.7.

2-(Naphthalen-2-ylmethyl)-1H-indole-5-carbaldehyde 27. The

reaction was carried out according to procedure F using 25. After

ﬂash column chromatography (hexane–EtOAc 12 : 1, 10 : 1, 9 : 1,

7 : 3), 272 mg (86%) of 27 was obtained as an amorphous pale

yellow powder: mp 162.5–163.5 ^◦C (hexane–DCM 20 : 1); R_f0.20

[hexane–EtOAc 7.5 : 2.5]; n_max(nujol)/cm^-13221, 2980–2850, 1660,

1306, 796; ¹³C NMR (CDCl₃/MeOD 1 : 2): d 194.21 (CH), 141.47

(C), 141.20 (C), 136.49 (C), 134.11 (C), 132.84 (C), 129.29 (C),

129.08 (C), 128.65, 128.03, 127.96, 127.63, 127.51, 126.54, 126.02

(CH), 125.93 (CH), 121.94 (CH), 111.83 (CH), 102.50 (CH), 35.13

1

(CH₂); H NMR (CDCl₃/MeOD 1 : 2) d 9.86 (1H, s, H1¢¢), 8.00

(1H, s, H4), 7.78–7.73 (3H, m), 7.69 (1H, s, H1), 7.60 (1H, d, J

8.4 Hz, H6), 7.44–7.33 (4H, m, H7), 6.36 (1H, br, H3), 4.25 (2H,

s, H1¢); m/z (EI⁺) 285 ([M]⁺, 100), 254 ([M-aldehyde]⁺, 38), 158

(50), 158 (42); m/z HRMS (EI⁺) [M]⁺calcd 285.1154 for C₂₀H₁₅NO

found 285.1146; Elemental calcd C, 84.2; H, 5.3; N, 4.9; found C,

84.1; H, 5.3; N, 4.9.

(Z)-Ethyl 2-ethoxy-3-(2-(naphthalen-2-ylmethyl)-1H-indol-5-yl)-

acrylate 32 and (E)-ethyl 2-ethoxy-3-(2-(naphthalen-2-ylmethyl)-

1H-indol-5-yl)acrylate 33. The synthesis was conducted accord-

ing to procedure A using the staring materials 27 and 9. After ﬂash

column chromatography (hexane–EtOAc 12 : 1, 9 : 1, 8 : 2, EtOAc)

112 mg (79%) of two isomers was obtained. Ratio Z : E = 63 : 37.

Z isomer 32: colourless oil: R_f0.30 [hexane–EtOAc 7 : 3]; n_max

(DCM ﬁlm)/cm^-13352, 3053–2900, 1701, 1620, 1252, 1095, 740;

¹³C NMR (CDCl₃): d 165.32 (C), 142.39, 138.60, 136.60, 135.70,

133.52, 132.36, 128.81 (C), 128.45, 127.67, 127.53, 127.16, 127.12,

126.34, 126.26, 125.74 (CH), 125.48 (C), 124.07 (CH), 122.70

(CH), 110.50 (CH), 101.81 (CH), 67.40 (CH₂), 60.87 (CH₂), 34.83

(CH₂), 15.54 (CH₃), 14.32 (CH₃); ¹H NMR (CDCl₃): d 8.06 (1H,

s, H4), 7.98 (1H, br, H1), 7.86–7.78 (3H, m), 7.72 (1H, s), 7.65

2-Ethoxy-3-(2-(naphthalen-2-ylmethyl)-1H-indol-5-yl)propa-

noic acid 6. The synthesis was performed according to procedure

C using 38. This procedure resulted in 58.5 mg (98%) of 6 as a

yellow oil: R_f0.25 [hexane–EtOAc 1 : 9]; n_max(DCM ﬁlm)/cm^-1

3403, 3053–2900, 1720, 1644, 1267, 1110, 739; ¹³C NMR (CDCl₃):

d 174.59 (C), 138.03, 135.90, 135.39, 133.53, 132.35, 128.87 (C),

128.44 (CH), 127.68 (CH), 127.65 (C), 127.53, 127.24, 127.12,

126.25, 125.72 (CH), 123.01 (CH), 120.69 (CH), 110.32 (CH),

101.04 (CH), 80.31 (CH), 66.89 (CH₂), 38.75 (CH₂), 34.93 (CH₂),

15.07 (CH₃); ¹H NMR (CDCl₃): d 7.85–7.79 (4H, m), 7.71 (1H, s),

7.51–7.45 (2H, m), 7.43 (1H, s, H4), 7.38 (1H, dd, J 8.4, 1.6 Hz),

7.15 (1H, d, J 8.4 Hz, H7), 7.01 (1H, dd, J 8.2, 1.4 Hz, H6), 6.34

(1H, br, H3), 4.28 (2H, s, H1¢), 4.12 (1H, dd, J 7.6, 4.4 Hz, H2¢¢),

3.62–3.54 (1H, m, H3¢¢), 3.50–3.42 (1H, m, H3¢¢), 3.23 (1H, dd, J

This journal is

The Royal Society of Chemistry 2011

Org. Biomol. Chem., 2011, 9, 1169–1188 | 1185

©

View Article Online

14.2, 4.2 Hz, H1¢¢), 3.08 (1H, dd, J 14.2, 7.8 Hz, H1¢¢), 1.18 (3H,

t, J 7.0 Hz, H4¢¢); m/z (EI⁺) 373 ([M]⁺, 33), 270 ([M-side chain on

5-position]⁺, 100), 142 (36); m/z HRMS (EI⁺) [M]⁺calcd 373.1678

for C₂₄H₂₃NO₃found 373.1676; Elemental calcd C, 77.2; H, 6.2;

N, 3.8; found C, 73.2; H, 6.2; N, 3.0.

(2H, s, H1¢), 4.25 (1H, dd, J 7.6, 4.8 Hz, H2¢¢), 4.01–3.92 (1H, m,

H3¢¢), 3.74 (3H, s, H6¢¢), 3.74–3.66 (1H, m, H3¢¢), 3.23–3.18 (1H,

m, H1¢¢), 3.17–3.12 (1H, m, H1¢¢); m/z (EI⁺) 441 ([M]⁺, 56), 270

([M-propanoate side chain]⁺, 100); m/z HRMS (EI⁺) [M]⁺calcd

441.1552 for C₂₅H₂₂F₃NO₃found 441.1548; Elemental calcd C,

68.2; H, 5.0; N, 3.2; found C, 72.7; H, 4.9; N, 3.8.

(Z)-Ethyl 3-(2-(naphthalen-2-ylmethyl)-1H-indol-5-yl)-2-(2,2,2-

triﬂuoroethoxy)-acrylate 34 and (E)-ethyl 3-(2-(naphthalen-

2-ylmethyl)-1H-indol-5-yl)-2-(2,2,2-triﬂuoro-ethoxy)-acrylate 35.

The synthesis was carried out according to procedure D using the

starting materials 27 and 10. After puriﬁcation via ﬂash column

chromatography (hexane–EtOAc 15 : 1, 10 : 1), 183 mg (78%) of

two isomers was obtained. Ratio Z : E = 38 : 62. Z isomer 34:

almost colourless oil: R_f0.35 [hexane–EtOAc 7 : 3]; n_max(DCM

ﬁlm)/cm^-13408, 2975, 1702, 1620, 1263, 1167, 742; ¹³C NMR

(CDCl₃): d 164.08 (C), 140.38, 138.80, 136.95, 135.56, 133.56,

132.42, 128.93 (C), 128.58 (CH), 127.71 (CH), 123.43 (C, q, J

277.0 Hz), 127.55, 127.48, 127.18, 127.15 126.34, 125.82 (CH),

124.38 (CH), 124.30 (C), 123.39 (CH), 110.64 (CH), 102.00 (CH),

67.82 (CH₂, q, J 35.0 Hz), 61.30 (CH₂), 34.88 (CH₂), 14.28 (CH₃);

¹H NMR (CDCl₃): d 8.06 (1H, s, H4), 7.93 (1H, br, H1), 7.88–7.82

(3H, m), 7.75 (1H, s), 7.61 (1H, dd, J 8.6, 1.4 Hz, H6), 7.56–7.48

(2H, m), 7.43 (1H, dd, J 8.4, 1.6 Hz), 7.29 (1H, s, H1¢¢), 7.24 (1H, d,

J 8.4 Hz, H7), 6.45 (1H, d, J 0.8 Hz, H3), 4.43 (4H, m, H3¢¢, H6¢¢),

4.32 (2H, s, H1¢), 1.41 (3H, t, J 7.2 Hz, H7¢¢); m/z (EI⁺) 453 ([M]⁺,

100), 342 (17), 141 (93); m/z HRMS (EI⁺) [M]⁺calcd 453.1552 for

C₂₆H₂₂F₃NO₃found 453.1553; Elemental calcd C, 68.9; H, 4.9; N,

3.1; found C, 68.8; H, 4.8; N, 2.9. E isomer 35: off-white powder:

mp 123.0–124.5 ^◦C (hexane–DCM 20 : 1); R_f0.30 [hexane–EtOAc

7 : 3]; n_max(DCM ﬁlm)/cm^-13441, 2970–2850, 1709, 1645, 1267,

1132, 740; ¹³C NMR (CDCl₃): d 163.51 (C), 143.33, 138.42, 136.08,

135.73, 133.54, 132.38, 128.58 (C), 128.49, 127.70, 127.53 (CH),

123.25 (CH₂, q, J 277.0 Hz), 127.16, 127.13, 126.31, 125.78 (CH),

124.37 (C), 122.82 (CH), 121.82 (CH), 121.00 (CH), 110.11 (CH),

101.49 (CH), 67.68 (CH₂, q, J 35.0 Hz), 61.26 (CH₂), 34.87 (CH₂),

3-(2-(Naphthalen-2-ylmethyl)-1H-indol-5-yl)-2-(2,2,2-triﬂuoro-

ethoxy)propanoic acid 7. The synthesis was performed according

to procedure C using 39, giving 5.60 mg (98%) of 7 as a pale

yellow oil: R_f0.40 [Solvent A]; n_max(DCM ﬁlm)/cm^-13441,

3000–2700, 1706, 1646, 1267, 1049, 740; d_C(CDCl₃) 174.21 (C),

138.19, 135.85, 135.50, 133.55, 132.38, 128.96 (C), 128.48, 127.69

(CH), 123.43 (C, q, J 277.0 Hz), 127.54, 127.23, 127.14 (3xCH),

126.98 (C), 126.27, 125.74 (CH), 122.91 (CH), 120.70 (CH),

110.46 (CH), 101.07 (CH), 81.65 (CH), 68.13 (CH₂, q, J 34.5 Hz),

38.98 (CH₂), 34.93 (CH₂); ¹H NMR (CDCl₃/MeOD 2 : 1): d 9.53

(1H, br, H6¢¢), 7.76–7.71 (3H, m), 7.65 (1H, s), 7.51 (1H, s, H4),

7.42–7.38 (2H, m), 7.34 (1H, dd, J 8.4, 1.2 Hz), 7.16 (1H, d, J

8.4 Hz, H7), 7.09–7.07 (1H, m, H6), 6.59 (1H, s, H1¢¢), 6.18 (1H,

br, H3), 4.26–4.20 (4H, m, H3¢¢, H1¢); m/z (EI⁺) 427 ([M]⁺, 47),

270 (100), 141 (53); m/z HRMS (EI⁺) [M]⁺calcd 427.1395 for

C₂₄H₂₀F₃NO₃found 427.1389; Elemental calcd C, 67.4; H, 4.7; N,

3.3; found C, 67.3; H, 4.6; N, 3.2.

PPAR activation assay

A plasmid encoding for each subtype of PPARs, as well as the

luciferase reporter gene was transfected into human breast cancer

MCF-7 cells. The ﬁreﬂy luciferase catalyses the bioluminescent

oxidation of luciferin in the presence of ATP, magnesium and

oxygen.⁶⁸The amount of luciferase protein expressed was mea-

sured on a luminometer and expressed in relative light units

(RLU), used to quantify the efﬁciency of the PPAR activation.

The indole compounds 2 to 7 were dissolved in DMSO, 0.1%

(v/v). DMSO concentration was kept constant in all samples.

Human MCF-7 breast cancer cells were seeded at a density of

85,000 cells per well of a 12-well plate. The following day, they were

transiently transfected by the SuperFect transfection procedure

according to the manufacturer¢s protocol (Qiagen), using 0.9 mg

PPREx3-LUC reporter plasmid and 0.15 mg of the plasmid vector,

pcDNA3.1 humanPPARa/g/d.

The total amount of plasmid was kept constant at 2.65 mg

by compensating with empty vector pCMV-5. Post 24 h after

transfection, the cells were treated with the compounds to be tested

or the PPAR speciﬁc agonists. 48 h after transfection, the cells

were washed once with PBS, lysed in 80 mL lysis buffer containing

25 mM TAE, 2 mM DTT, 10% (v/v) glycerol, 1% (v/v) Triton

X-100 and de-ionised water. 35 mL of the cell extracts were used

for luciferase determination on a LUCY-1 luminometer (Anthos,

Austria). The luciferase assay was performed in accordance with

the protocol of the Luciferase Assay Kit (BIO Thema AB,

Sweden).

1

13.79 (CH₃); H NMR (CDCl₃): d 7.86–7.79 (4H, m), 7.71 (1H,

s, H1), 7.55 (1H, s, H4), 7.53–7.49 (2H, m), 7.36 (1H, dd, J 8.4,

1.6 Hz), 7.16 (1H, d, J 8.4 Hz, H7), 7.09 (1H, dd, J 8.4, 1.6 Hz),

6.72 (1H, s, H1¢¢), 6.36 (1H, d, J 0.8 Hz, H3), 4.30–4.23 (4H, m,

2xH3¢¢, 2xH1¢), 4.20 (2H, q, J 7.2 Hz, H6¢¢), 1.17 (3H, t, J 7.2 Hz,

H7¢¢); m/z (EI⁺) 453 ([M]⁺, 61), 342 (10), 210 (100), 141 (46),

105 (69); m/z HRMS (EI⁺) [M]⁺calcd 453.1552 for C₂₆H₂₂F₃NO₃

found 453.1548; Elemental calcd C, 68.9; H, 4.9; N, 3.1; found C,

69.0; H, 5.0; N, 3.0.

Methyl 3-(2-(naphthalen-2-ylmethyl)-1H-indol-5-yl)-2-(2,2,2-

triﬂuoroethoxy)-propanoate 39. The reaction was performed

according to the procedure B using 34. After puriﬁcation via

ﬂash column chromatography (hexane–EtOAc 9 : 1, 8 : 2) 11.3 mg

(79%) of 39 was obtained as a colourless oil: R_f0.25 [hexane–

EtOAc 7.5 : 2.5]; n_max(DCM ﬁlm)/cm^-13403, 3054–2850, 1743,

1645, 1279, 1165, 739; ¹³C NMR (CDCl₃): d 171.49 (C), 138.05,

135.90, 135.44, 133.56, 132.38, 128.92 (C), 128.46 (CH), 123.58

(C, q, J 277.0 Hz), 127.69 (CH), 127.55 (CH), 127.36 (C), 127.24,

127.13, 126.26, 125.73 (CH), 122.93 (CH), 120.60 (CH), 110.33

(CH), 101.07 (CH), 82.04 (CH), 67.89 (CH₂, q, J 34.5 Hz), 52.09

(CH₃), 39.18 (CH₂), 34.94 (CH₂); ¹H NMR (CDCl₃): d 7.85–7.79

(3H, m), 7.77 (1H, br, H1), 7.72 (1H, s), 7.52–7.46 (2H, m), 7.42

(1H, s, H4), 7.38 (1H, dd, J 8.6, 1.8 Hz), 7.16 (1H, d, J 8.0 Hz,

H7), 6.99 (1H, dd, J 8.4, 1.6 Hz, H6), 6.34 (1H, m, H3), 4.29

Statistical Analysis

Statistical analyses were carried out using GraphPad Prism

version 5.0a. All P values were two-sided and values <0.05 were

considered statistically signiﬁcant. Values are given as means

with standard deviations (SD). One-way analysis of variance

1186 | Org. Biomol. Chem., 2011, 9, 1169–1188

This journal is

The Royal Society of Chemistry 2011

©

View Article Online

and Tukey¢s multiple comparison test were used to demonstrate

signiﬁcant differences.

17 K. Schoonjans, M. Watanabe, H. Suzuki, A. Mahfoudi, G. Krey, W.

Wahli, P. Grimaldi, B. Staels, T. Yamamoto and J. Auwerx, J. Biol.

Chem., 1995, 270, 19269.

18 S. M. Rangwala and M. A. Lazar, TIPS, 2004, 25, 331.

19 P. Kirkpatrick, Nat. Rev. Drug Discovery, 2002, 1, 169.

20 M. W. Schwartz and S. E. Kahn, Nature, 1999, 402, 860.

21 G. Martin, K. Schoonjans, A. M. Lefebvre, B. Staels and J. Auwerx, J.

Biol. Chem., 1997, 272, 28210.

22 W. Wahli, Swiss Medical Weekly, 2002, 132, 83.

23 W. R. Oliver, J. L. Shenk, M. R. Snaith, C. S. Russell, K. D. Plunket,

N. L. Bodkin, M. C. Lewis, D. A. Winegar, M. L. Sznaidman,

M. H. Lambert, H. E. Xu, D. D. Sternbach, S. A. Kliewer, B. C.

Hansen and T. M. Willson, Proc. Natl. Acad. Sci. U. S. A., 2001, 98,

5306.

Conclusion

A novel series of 2,5-disubstituted indoles, designed using a soft-

ware program PROTOBUILD, have been synthesised successfully and

shown to be dual PPARa/g agonists using a luciferase reporter

gene activation assay. Further tests are now required in order to

assess in vivo efﬁcacy and ADME-toxicology properties in order

to allow selection of the best candidate for future development.

The data reported here demonstrate the potential utility of a new

software program PROTOBUILD to predict alternative, novel drug

leads for a target that has been the subject of intense interest. Our

next step will be to explore the design, synthesis and evaluation

of potential drug leads based upon completely novel chemotypes

generated by our software program PROTOBUILD.

24 B. Staels, J. Dallongeville, J. Auwerx, K. Schoonjans, E. Leitersdorf and

J. C. Fruchart, Circulation, 1998, 98, 2088.

25 P. Raskin, E. B Rappaport, S. T Cole, Y. Yan, R. Patwardhan and M.

I. Freed, Diabetologia, 2000, 43, 278.

26 J. Berger, P. Bailey, C. Biswas, C. A. Cullinan, T. W. Doebber, N. S.

Hayes, R. Saperstein, R. G. Smith and M. D. Leibowitz, Endocrinology,

1996, 137, 4189.

27 A. R. Miller, Drug Dev. Res., 2006, 67, 574.

28 B. Gross and B. Staels, Best Pract. Res. Clin. Endocrinol. Metab., 2007,

21, 687.

Abbreviations

29 A.-M. Lefebvre, J. Peinado-Onsurbe, I. Leitersdorf, M. R. Briggs, J. R.

Paterniti, J. C. Fruchart, C. Fievet, J. Auwerx and B. Staels, Arterioscler.

Thromb. Vasc. Biol., 1997, 17, 1756.

EI

electron ionisation

30 P. Gervois, J. C. Fruchart and B. Staels, Int. J. Clin. Pract., 2004, 58,

PPAR

PPRE

LBD

VLDL

peroxisome proliferator-activated receptor

peroxisome proliferator response element

ligand binding domain

22.

31 H. J. Bohm, J. Comput.-Aided Mol. Des., 1992, 6, 61.

32 M. Stahl, N. P. Todorov, T. James, H. Mauser, H.-J. Bohm and P. M.

Dean, J. Comput.-Aided Mol. Des., 2002, 16, 459.

33 B. A. Grzybowski, A. V. Ishchenko, C. Y. Kim, G. Topalov, R. Chap-

man, D. W. Christianson, G. M. Whitesides and E. I. Shakhnovich,

Proc. Natl. Acad. Sci. U. S. A., 2002, 99, 1270.

very low-density lipoprotein

Acknowledgements

34 R. S. Bohacek and C. McMartin, J. Am. Chem. Soc., 1994, 116,

5560.

We wish to thank Mitsubishi Chemical Corporation, IC-Vec Ltd,

and the EPSRC for their ﬁnancial support of the Imperial College

Genetic Therapies Centre. The authors also wish to acknowledge

Dr Andrew White for the X-ray crystallography structures, John

Barton for high resolution mass spectrometry and Stephen Boyer

for elemental analysis.

35 The Molecular Operating Environment (MOE) available under license

from Chemical Computing Group Inc., 1010 Sherbrooke St. W. Suite

910, Montreal, Quebec, Canada H3A 2R7.

36 A. R. Organov and C. W. Glass, J. Chem. Phys., 2006, 124, 244704.

37 H. M. Vinkers, M. R. de Jonge, F. F. Daeyaert, J. Heeres, L. M.

Koymans, J. H. van Lenthe, P. J. Lewi, H. Timmerman, K. Van Aken

and P. A. Janssen, J. Med. Chem., 2003, 46, 2765.

38 M. Clark, R. D. Cramer and N. Van Opdenbosch, J. Comput. Chem.,

1989, 10, 982.

39 T. A. Halgren, J. Comput. Chem., 1996, 17, 490.

40 P. A. Nielsen and T. Liljefors, J. Comput.-Aided Mol. Des., 2001, 15,

753.

References

1 A. M. Rouhi, Metabolic Syndrome, C & E News, 2004, 83.

2 C. L. Scott, Am. J. Cardiol., 2003, 92, 35i.

3 M. P. Stern, K. Williams, C. Gonza´lez-Villalpando, K. J. Hunt and S.

M. Haffner, Diabetes Care, 2004, 27, 2676.

4 I. Issemann and S. Green, Nature, 1990, 347, 645.

5 C. Dreyer, G. Krey, H. Keller, F. Givel and G. Helftenbein, Cell, 1992,

68, 879.

6 T. M. Willson, P. J. Brown, D. D. Sternbach and B. R. Henke, J. Med.

Chem., 2000, 43, 527.

7 B. Desvergne and W. Wahli, Endocr. Rev., 1999, 20, 649.

8 W. Wahli, O. Braissant and B. Desvergne, Chem. Biol., 1995, 2,

261.

41 H-J. Bohm, J. CAMD., 1998, 12, 309.

42 F. J Carver, C. A. Hunter and E. M. Seward, Chem. Commun., 1998,

775.

43 C. A. Hunter, J. Singh and J. M. Thornton, J. Mol. Biol., 1991, 218,

837.

44 M. L. Verdonk, G. J. Boks, H. Kooijman, J. A. Kanters and J. Kroon,

J. Comput.-Aided Mol. Des., 1993, 7, 173.

45 R. J. Zauhar, C. L. Colbert, R. S. Morgan and W. J. Welsh, Biopolymers,

2000, 53(3), 233.

46 J. E. Cobb, B. R. Henke and S. G. Blanchard, Diabetologia, 2000, 43,

A189.

47 J. Xu and J. Stevenson, J. Chem. Inf. Comp. Sci., 2000, 40, 1177.

48 C. A. Lipinski, F. Lombardo, B. W. Dominy and P. J. Feeney, Adv. Drug

Delivery Rev., 1997, 23, 3.

49 X. Q. Lewell, D. B. Judd, S. P. Watson and M. M. Rhann, J. Chem. Inf.

Comp. Sci., 1998, 38, 511.

9 J. Berger and D. E. Moller, Annu. Rev. Med., 2002, 53, 409.

10 A. K. Hihi, L. Michalik and W. Wahli, Cell. Mol. Life Sci., 2002, 59,

790.

11 T. Lemberger, B. Desvergne and W. Wahli, Annu. Rev. Cell Dev. Biol.,

1996, 12, 335.

50 J. Rudolph, L. Chen, D. Majumdar, W. H. Bullock, M. Burns, T. Claus,

F. E. Dela Cruz, M. Daly, F. J. Ehrgott, J. S. Johnson, J. N. Livingston,

R. W. Schoenleber, J. Shapiro, L. Yang, M. Tsutsumib and X. Ma, J.

Med. Chem., 2007, 50, 984.

51 B. Kuhn, H. Hilpert, J. Benz, A. Binggeli, U. Grether, R. Humm, H.

P. Marki, M. Meyer and P. Mohr, Bioorg. Med. Chem. Lett., 2006, 16,

4016.

52 N. Mahindroo, C. F. Huang, Y. H. Peng, C. C. Wang, C. C. Liao, T.

W. Lien, S. K. Chittimalla, W. J. Huang, C. H. Chai, E. Prakash, C.

P. Chen, T. A. Hsu, C. H. Peng, I. L. Lu, L. H. Lee, Y. W. Chang, W.

12 G. D. Barish, V. A. Narkar and R. M. Evans, J. Clin. Invest., 2006, 116,

590.

13 P. Lefebvre, G. Chinetti, J. C. Fruchart and B. Staels, J. Clin. Invest.,

2006, 116, 571.

14 R. K. Semple, V. K. K. Chatterjee and S. O’Rahilly, J. Clin. Invest.,

2006, 116, 581.

15 K. Schoonjans, B. Staels and J. Auwerx, Biochim. Biophys. Acta., 1996,

1302, 93.

16 O. Braissant, F. Foufelle, C. Scotto, M. Dauca and W. Wahli,

Endocrinology, 1996, 137, 354.

This journal is

The Royal Society of Chemistry 2011

Org. Biomol. Chem., 2011, 9, 1169–1188 | 1187

©

View Article Online

C. Chen, Y. C. Chou, C. T. Chen, C. M. Goparaju, Y. S. Chen, S. J.

Lan, M. C. Yu, X. Chen, Y. S. Chao, S. Y. H. Wu and P. Hsieh, J. Med.

Chem., 2005, 48, 8194.

60 M. Regitz, W. Anschutz and A. Liedhegener, Chem. Ber., 1968, 101,

3734.

61 D. Haigh, H. C. Birrell, B. C. C. Cantello, R. M. Hindley, A.

Ramaswamy, H. K. Rami and N. C. Stevens, Tetrahedron: Asymmetry,

1999, 10, 1335.

62 R. J. Sundberg and H. F. Russell, J. Org. Chem., 1973, 38, 3324.

63 R. T. Nolte, G. B Wisely, S. Westin, J. E. Cobb, M. H. Lambert, R.

Kurokawa, M. G. Rosenfeld, T. M. Willson, C. K. Glass and M. V.

Milburn, Nature, 1998, 395, 137.

64 H. E. Xu, M. H. Lambert, V. G Montana, K. D. Plunket, L. B. Moore,

J. L. Collins, J. A. Oplinger, S. A. Kliewer, R. T. Gampe Jr, D. D. McKee,

J. T. Moore and T. M. Willson, Proc. Natl. Acad. Sci. U. S. A., 2001,

98, 13919.

53 N. Mahindroo, C. C. Wang, C. C. Liao, C. F. Huang, I. L. Lu, T. W.

Lien, Y. H. Peng, W. J. Huang, Y. T. Lin, M. C. Hsu, C. H. Lin, C.

H. Tsai, J. T. Hsu, X. Chen, P. C. Lyu, Y. S. Chen, S. Y. Wu and H. P.

Hsieh, J. Med. Chem., 2006, 49, 1212.

54 N. Mahindroo, Y. H. Peng, C. H. Lin, U. K. Tan, E. Prakash, T.

W. Lien, I. L. Lu, H. J. Lee, J. T. Hsu, X. Chen, C. C. Liao, P. C.

Lyu, Y. S. Chao, S. Y. Wu and H. P. Hsieh, J. Med. Chem., 2006, 49,

6421.

55 B. B. Lohray, V. B. Lohray, A. C. Bajji, S. Kalchar and R. Ramanujam,

US patent, 2000, 6(054), 453.

56 W. Zhang, X. Y. Cao, H. Zi and J. Pei, Org. Lett., 2005, 7, 959.

57 W. Grell and H. Machleidt, Annal. Chem., 1966, 699, 53.

58 L. McElweewhite, W. A Goddard and D. A. Dougherty, J. Am. Chem.

Soc., 1984, 106, 3461.

59 C. J. Moody, E. Sie and J. J. Kulagowski, Tetrahedron, 1992, 48,

3991.

65 W. C. Still, M. Kahn and A. Mitra, J. Org. Chem., 1978, 43, 2923.

66 D. Haigh, Tetrahedron, 1994, 50, 3177.

67 E. J. Iwanowicz, W. F. Lau, J. Lin, D. G. M. Roberts and S. M. Seiler,

Bioorg. Med. Chem. Lett., 1996, 6, 1339.

68 J. R. de Wet, K. V. Wood, M. Deluca, D. R. Helinski and S. Subramani,

Mol. Cell. Biol., 1987, 7, 725.

1188 | Org. Biomol. Chem., 2011, 9, 1169–1188

This journal is

The Royal Society of Chemistry 2011

©

Article Doi

Synthesis of novel PPARα/γ dual agonists as potential drugs for the treatment of the metabolic syndrome and diabetes type II designed using a new de novo design program PROTOBUILD

DOI: 10.1039/c0ob00146e

Source and publish data:

Authors:

Article abstract of DOI:10.1039/c0ob00146e

Full text of DOI:10.1039/c0ob00146e

Products guided by the article

R&D Labs maybe for 1206726-86-4

Relevant to this article

Hot Product