Lead Diversification 2: Application to P38, gMTP and lead compounds

Article abstract of DOI:10.1016/j.bmcl.2011.11.033

Lead Diversification is a new technology platform developed at Pfizer for the functionalization of drug molecules using C-H activation. We describe its application to some drug programs such as P38 and gMTP and the development of some new plate based scre

Full text of DOI:10.1016/j.bmcl.2011.11.033

Bioorganic & Medicinal Chemistry Letters 22 (2012) 1255–1262
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Lead Diversification 2: Application to P38, gMTP and lead compounds
M. Abid Masood , Marc Bazin , Mark E. Bunnage, Andrew Calabrese à, Mark Cox, Sally-Ann Fancy,
⇑
Elizabeth Farrant, David W. Pearce, Manuel Perez, Laure Hitzel, Torren Peakman
Worldwide Medicinal Chemistry, Pfizer Worldwide R&D, Ramsgate Road, Sandwich, Kent CT13 9NJ, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 17 June 2011
Revised 5 November 2011
Accepted 9 November 2011
Available online 17 November 2011
Lead Diversification is a new technology platform developed at Pfizer for the functionalization of drug
molecules using C–H activation. We describe its application to some drug programs such as P38 and
gMTP and the development of some new plate based screens including a fluorination screen.
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
Biomimetic
Lead Diversification
Metalloporphyrin
Fluorination
P38
In our previous paper the concept of Lead Diversification (LD)
was outlined and some of the chemical screens developed such
as the biomimetic oxidation1 (BMO) screen were described. In this
letter further transformations such as fluorination, chlorination,
and methoxylation of drug molecules are discussed.
It should be noted that Wender et al. have also explored func-
tionalization of natural products using a Lead Diversification (LD)
approach2; however, our work explores a greater diversity of trans-
formations on drug molecules.
As part of Pfizer’s internal efforts to find a human gMTP inhibitor,
close analogues of the aminoquinoline diamide 1 lead were synthe-
sized using LD. The strategy was to make close chemical analogues
to gain further structure–activity relationship (SAR) data and ex-
plore available LD chemistry options. The compound 1 is a good sub-
strate to explore the chemoselectivity of such transformations.
Our initial LD efforts involved plate based screens for biomi-
metic oxidation (BMO) using metalloporphyrin catalyzed oxida-
tions. Two polar compounds were observed in several wells. The
more polar compound had a mass ion of M+16 amu and the less
polar compound had a mass ion of M-41 amu. This led us to believe
that the reaction had added polarity to substrate 1 or removed
lipophilicity. Isolation and structure elucidation showed the prod-
ucts to be the O-dealkylated derivative 2 and the N-oxide 3
(Scheme 1). Both compounds were subsequently found to be inac-
tive against gMTP. The O-dealkylated adduct 2 is the major human
metabolite and the N-oxide 3 is a minor human metabolite. Using
biomimetic oxidation, both metabolites were synthesized in a
short time compared to a more time consuming conventional mul-
tistep synthesis.
Over the past five years there has been exponential growth in
the field of C–H functionalisation and one of the earlier methodol-
ogies developed was by Sanford’s group using a palladium II/IV
system to catalyze C–H alkoxylation and O-acetoxylation onto aro-
matic groups.3 Many of these new transformations had been devel-
oped using simple systems like N-phenylpyrazole4 4 (Scheme 2).
Based on the published results, it was important to apply this to
more complex systems such as drug-like compounds with a molec-
ular weight of around 500 to see the scope of reactivity and
chemoselectivity of these reactions. Sanford’s work showed that
C–H activation was directed by the palladium species co-ordinat-
ing to sp2 nitrogen. It was reasoned that sp2 oxygen atoms were
also weakly Lewis basic such that a carbonyl could co-ordinate to
palladium and enable Sanford-type directed C–H activation reac-
tions to proceed.
Aminoquinoline diamide 1 was dissolved in acetic acid and pal-
ladium(II)acetate (10 mol %) and oxone were added then heated in
a microwave reactor at 150 °C for 2 h.5 LCMS analysis of the crude
reaction mixture indicated the appearance of several close running
peaks in addition to unreacted starting material (Fig. 1). The mass
ions showed that OAc (M+44) had been added to the parent
molecule.
⇑
Corresponding author. Tel.: +44 01227 780843.
E-mail addresses: mabidmasood@gmail.com, abid.masood@pfizer.com (M.A.
Masood).
Present address: Hepatochem, 185 Alewife Brook Parkway, # 410, Cambridge, MA
02138, USA.
à
Present address: Celgene San Diego, 4550 Towne Centre Court, San Diego, CA
92121, USA.
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.11.033

1256
M. A. Masood et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1255–1262
O
F
O
N
O
N
H
O
N
H
N
1
gMTP IC50: 2nM
OH
O
F
F
O
O
N
N
O
N
H
O
N
H
O
O
+
N
H
N
N
H
N
O
–
2
3
gMTP IC50: inactive
gMTP IC50: inactive
Scheme 1. Two products isolated from biomimetic oxidation scale-up reactions.
N
N
but only the products shown were isolated in sufficient purity and
quantity for structure elucidation. Biological testing showed that
the O-acetoxylated derivative 5 was less potent than the starting
compound 1.
N
i
N
O
4
The next chemical transformation attempted was Sanford’s pal-
ladium(II)acetate catalyzed O-methoxylation on the aminoquino-
line diamide lead compound 1 (Scheme 4). Again this resulted in
a mixture of products. The methoxy derivative 7 was the only
component isolated.
Scheme 2. Palladium(II) catalyzed methoxylation. Reagents and conditions: (i)
Oxone, MeOH, Pd(OAc)2 (10 mol %), 80° C, 4–6 h.
Methoxy derivative 7 was found to be a 100-fold less potent
against human gMTP. Sanford reactions gave multiple products
on drug molecules due to multiple palladium binding sites. As a
result the effective palladium concentration is much lower and
so the yield is lower. When the amount of catalyst was increased
to 20 mol % the percent conversion based on ELSD and UV did in-
crease, but adding more catalyst above this quantity did not signif-
icantly improve yield or conversion. It should be noted that
synthesis of derivatives 5, 6 and 7 from commercial fragments
using conventional synthesis would have taken a long time as well
as being very resource intensive. Using the LD approach sufficient
material was isolated to obtain data that formed valuable SAR.
While some of the chemistries developed were plate-based (oxida-
tion screens, halogenation screens), other chemistries were found
to work better as a single reaction using microwave heating.
The LD approach was applied to Pfizer’s internal program to find
an inhaled P38 inhibitor project lead, the aminopyrazole urea 8
was reasons why this compound was selected; one being submit-
ted to the plate based BMO screen. There were several reasons; the
desire to synthesize polar compounds to improve solubility, which
had previously been a problem with this compound series. In addi-
tion, the compounds in this project were large complex molecules
with heterocycles and functional groups that could react under the
screening conditions. Palladium catalyzed C–H activation chemis-
try methods were attempted and met with some success, but
purification proved impossible in all cases.
Figure 1. LCMS (ELSD) trace of Pd(II) acetate catalyzed methoxylation of 1 (crude
reaction trace).
From the ELSD and UV (254) traces of the LCMS (Fig. 1), approx-
imately 40% conversion of the starting compound 1 to products
was seen. The reaction was repeated three times and found to be
reproducible. Purification of the crude material was carried out
by preparative HPLC with two compounds being isolated in high
purity (>95%). Structure elucidation was carried out and the two
derivatives were identified as the O-acetoxylated analogue 5 and
the phenol 6 (Scheme 3).
Compound 5 is extremely interesting since to synthesize it from
commercial starting materials would be lengthy and previous
attempts at synthesis had failed because of a competing ring cycli-
zation reaction.
It was postulated that palladium coordinated to the carbonyl
oxygen and inserted into the aryl C–H bond to give the observed
regiochemistry. It should be noted that 6 products were observed,
The aminopyrazole urea 8 was reacted in the BMO chemical
screen and other oxidation screens. The BMO screen results
showed two wells that contained more polar products and another
oxidation screen showed one promising result with DuBois’ cata-
lyst6 9, urea hydrogen peroxide and catalytic diphenydiselenide
(Scheme 5). This reaction was scaled up and gave the sulfone 10
in 19% isolated yield.

M. A. Masood et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1255–1262
1257
Pd insertion is
directed by
co-ordination to
this Carbonyl
O
O
O
O
i
O
N
N
N
H
O
N
H
O
O
O
O
N
N
N
H
N
H
F
1
(6%
5
F
gMTP IC 50 2nM
gMTP IC 50 313nM
+
O
O
O
Pd
O
N
N
H
O
O
N
H
N
OH
F
6
(4% yield)
Scheme 3. Pd(II) catalyzed O-acetoxylation of 1. Reagents and conditions: (i) Oxone, acetic acid/acetic anhydride (1:1), Pd(OAc)2 (20 mol %), microwave irradiation (120 °C),
4 h.
O
O
O
O
i
N
N
N
H
O
O
N
H
O
O
O
N
H
N
N
H
N
F
F
1
7
gMTP IC50: 2nM
gMTP IC50: 150nM
40% conversion
5 products (3 mono methoxylated, 2 bismethoxylated)
Scheme 4. Reagents and conditions: (i) Oxone, MeOH, Pd(OAc)2 (20 mol %), microwave irradiation (120 °C), 4 h.
O
O
S
O
N
CF3
9
H
N
H
N
N
H
N
H
N
i
N
N
N
S
O
O
N
S
O
N
N
N
N
Cl
N
Cl
OH
(19% yield)
OH
8
10
MW 606
MW 590
P38 Ki IC50 0.8nM
P38 Ki IC50 71.3nM
Scheme 5. Reagents and conditions: (i) 5 (5 mol %), Urea HOOH, dichloroethane, PhSeSePh (5 mol %), rt, 48 h.
The Mn(TDClPP),7 H2O2, imidazole and formic acid (catalytic)
and the second containing a mixture of the oxidant, catalytic for-
mic acid and imidazole. The rationale for doing the reaction in flow
was that the scale-up reaction was extremely exothermic and the
risk of this could be minimized in flow. Furthermore, in flow, the
amount of oxidant that the metalloporphyrin catalyst is exposed
to at any one time can be limited and so in effect ‘fresh’ catalyst
is exposed to oxidant over time and this would minimalise catalyst
bleaching.9
wells were scaled up. Problems were encountered because we
saw bleaching of the catalyst8 and no reaction of the substrate.
Bleaching is where degradation of the metalloporphyrin occurs
by direct oxidation of the resting catalyst (MIII) rather than the
reactive intermediates such as the high valent oxo-perferryl
((TF5PP+)MIV = O) and is known to be a particular problem with
hydrogen peroxide. One way to circumvent metalloporphyrin
bleaching is slow addition of the oxidant to the reaction. This reac-
tion was attempted using flow chemistry with two reservoirs. One
reservoir containing the substrate and metalloporphyrin catalyst
The reaction was carried out on a Vapourtech R4 flow appara-
tus10 using a flow rate of 200
ll/min and a 10 ml loop. The reaction
time was 40 minutes and analysis of the crude product by LCMS

1258
M. A. Masood et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1255–1262
i
Oxidants
H
N
H
N
N
N
r.t. 40mins
Vapourtech R4
S
O
N
O
N
Mn(TDClPP)
H
N
H
N
+
N
N
Cl
N
OH
S
(81%)
O
N
N
10
ClogP : 4.8
ClogD 7.4 : 4.1
P38 IC50 : 71.3nM
N
Cl
OH
8
ClogP : 6.8
ClogD 7.4 : 6.5
P38 IC50: 0.8nM
ii
Oxidants
H
N
H
N
N
N
r.t. 40mins
Vapourtech R4
S
O
O
O
N
N
Mn(TDClPP)
H
N
H
N
+
N
N
Cl
N
OH
S
O
(32% yield)
N
N
11
N
Cl
ClogP : 5.1
OH
ClogD 7.4 : 4.5
P38 IC50 : 302nM
8
ClogP : 6.8
ClogD 7.4 : 6.5
P38 IC50: 0.8nM
Scheme 6. Flow mediated biomimetic oxidation of P38 substrate. Reagents and conditions; (i) H2O2, imidazole (1 mol %) and formic acid (5 mol %), dichloroethane; (ii)
mCPBA, CH3OH/CH3CN (1:1), rt.
H
N
H
N
N
i
H
N
H
N
N
N
N
S
O
S
O
S
N
S
O
N
N
N
N
Cl
HO
HO
N
Cl
OH
OH
12
13
ClogD7.4 : 5.6
ClogD7.4: 7.7
ClogP : 7.2
P38 IC50 : 12nM
ClogP
: 5.75
p38 IC50 : 10nM
(34% yield)
Scheme 7. BMO scale-up reaction. Reagents and conditions: (i) Mn(TDClPP) (10 mol %), mCPBA, DCE, rt 8 h.
showed complete conversion of parent to the sulfoxide 10. Synthe-
used to block metabolically vulnerable positions on drug com-
pounds.13 However, introducing fluorine into a drug molecule at
a late stage through synthesis is practically challenging.14 Although
there are many commercial fluorinating reagents, the chemistry
can be problematic because fluorine is also an excellent oxidizing
agent.15 The bond dissociation energy of C–F is 452.5 KJmolÀ1, C–
H is 410 KJmolÀ1 and that of N–F is approximately 318.6 KJmolÀ1
(76.1 kcal/mol). The majority of electrophilic fluorinating agents
have a nitrogen fluorine bond for example SelectFluor16 and Accu-
fluor, and so fluorination reactions can be extremely exothermic.
The heat generated during fluorination is sufficient for fluorine
radical formation which can destroy the substrate through uncon-
trolled reactions.17 As a result methods for finding conditions and
reagents to fluorinate organic molecules would be a useful tool
for both the synthetic and medicinal chemist. A plate based combi-
natorial method was used to develop such a fluorination screen
(Fig. 2) as a quick technique for finding optimum conditions for flu-
orinating drug molecules using a combinatorial array of reagents18
and solvents. The fluorinating reagents were made up as stock
sis of the sulfoxide 10 had been unsuccessfully attempted previ-
ously using conventional oxidation reagents. Scale-up of the
second well using Mn(TDClPP) and a protic solvent gave the
sulfone 11 in 32% yield (Scheme 6).
A closely related P38 compound 12 was reacted in the BMO
chemical screen and one well showed clean conversion to a more
polar product (Scheme 7).
The procedure was scaled up to 150 mmol of reactant using
Mn(TDClPP), mCPBA in dichloroethane for 8 h at room tempera-
ture. The more polar product isolated was the sulfoxide 13 on
the alkyl thioether portion in 34% isolated yield (Scheme 7). Biolog-
ical testing against MAPK P38 showed that the alkyl sulfoxide 13
was equipotent to the starting dithioether.
Fluorination strategies are common in medicinal chemistry,
serving to influence both pharmacological and metabolic processes
whilst avoiding detrimental effects on favorable properties.11 Fluo-
rine is small, electronegative (4.0 on the Pauling scale) and can
form hydrogen bonds.12 In drug discovery, fluorine is most often

M. A. Masood et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1255–1262
1259
Figure 2. Plate map of fluorination screen.
F
H
N
H
N
H
N
H
N
N
N
N
i
N
O
S
O
S
N
N
N
N
N
N
Cl
Cl
14
P38 Ki : 0.8nM
8
OH
OH
(28%)
P38 Ki : 0.8nM
ii
H
N
H
N
N
N
O
S
O
N
N
N
Cl
10
OH
P38 Ki 320nM
(19%)
Scheme 8. Fluorination screen scale-up reactions. Reagents and conditions: (i) AccufluorÒ, dichloroethane, rt, 50 h; (ii) SynfluorÒ, dichloroethane, rt, 30 h.
solutions in dichloroethane19 or acetonitrile. Four solvents were
used in the screen, based on literature searches20 and Pfizer inter-
nal examples.
The reactions were easy to scale-up and were always carried
out in plastic or Teflon vessels due to the reactivity of fluorinating
reagents with glass.14
The pyrazole adduct 8 was subjected to the plate-based fluori-
nation screen for 72 h.21 After quenching and analysis by LCMS,
two wells were found to have generated products. Scale-up, isola-
tion of products and then structure elucidation, identified the
fluoropyrazole derivative 14 and the sulfoxide 10 (Scheme 8).
The fluoropyrazole derivative 14 was isolated in 28% yield and
showed that selective fluorination of large drug molecules rich in
heteroatoms and functional groups could be carried out. Biological
testing against MAPK P38 showed that the fluoropyrazole deriva-
tive 14 was equipotent with the lead compound 8 and opened
up new chemical space.
Using a conventional multistep synthetic approach to make the
fluoropyrazole derivative 14 would have proved difficult at the
urea forming step with the fluoro-aminopyrazole coupling being
poor-yielding due to the low reactivity of the electron deficient
aminopyrazole.
The sulfoxide 10 was an interesting result and showed the oxi-
dizing power of fluorine. Sulphur has a great affinity for fluorine

1260
M. A. Masood et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1255–1262
Through the screening approach, three close-in analogues had
N
S
N
been synthesized and the problem of improving the half life
without losing potency while keeping the clogP down had been re-
solved. These three halogenated derivatives were shown to have
interesting pharmacological and metabolic properties. The impact
on metabolism is most striking and suggests that LD has been
successful in blocking metabolically vulnerable sites.
N
S
N
N
N
N
N
F
17
18
Cl
N
N
N
In a second example, compound 21 was the lead compound in
the melanocortin-4 receptor antagonist (MC-4) program. Com-
pound 21 is lipophilic and also required modification to improve
pharmacological properties. Compound 21 was tested in the
biomimetic oxidation screen (BMO) and while several wells
showed promise, scale-up reactions were unsuccessful. In an ex-
tended oxidation screen, DuBois’ catalyst6 was found to give a
clean reaction profile. The DuBois scale-up reaction on 21 was car-
ried out by microwave heating of a solution of 21, DuBois catalyst,
diphenyl-diselenide and urea hydrogen peroxide in 1,2-dichloro-
ethane at 100 °C for 2 h. The product was isolated in 9% yield
and identified as the pyrrolidinone 22 (Scheme 9). Alternative syn-
thetic strategies to this interesting target using conventional
chemistry were very lengthy.
N
N
N
S
S
N
N
Cl
Cl
19
20
∞
Alk5 ActR2B : cLogP
PLM
R/
T1/2
min
IC50
(nM)
1.19
1.36
0.63
IC50
(nM)
541
Clint
Compound
ul/min/mg
122
17
18
19
20
3.65
3.81
4.13
4.81
11.4
1340
1060
>6000
12.7
109
<11.6
NA
>120
NA
77.8
∞
Pig liver microsomes (PLM)
In our previous communication the biomimetic oxidation of the
3-methyl of an indazole core for the corticosteroid releasing factor
(hCRF-1) project was described. The SAR generated helped to im-
prove solubility for the azaindole series. The heterocyclic com-
pound 23 was a key intermediate in a more polar series for
Pfizer’s CRF-1 program. The majority of the synthetic efforts were
focused on adding polarity to the dialkylamine portion of the inda-
zole and also onto the 2-aryl position as this takes the phenyl out
of the plane. According to Miller et al. this conformational change
is key for potency.24 Also, it should be noted that no compound had
been made where the phenyl had been in the plane (i.e., held in
plane by an internal hydrogen bond) to confirm that the ‘flat’ mol-
ecule is inactive. At this time, Sanford’s group had published the
palladium catalyzed methoxylation of N-aryl pyrazoles using
methanol via C–H activation.25 It was felt that this methodology
would prove invaluable to ortho-functionalize the phenyl group
in 23 and generate structure–activity relationships (SAR). In
addition, copper-catalyzed, directed aromatic hydroxylation had
been reported by JinQuan Yu’s group,26 however in our hands these
conditions were unsuccessful. The reaction was then attempted
using a methanol/water mixture (2:1) as solvent with microwave
heating (Scheme 10). This is a new method for aromatic hydroxyl-
ation, however the methoxy derivative 25 is also formed as the
major adduct.
The reaction was heated at 100 °C and after 4 h, LCMS analysis
showed two more polar products alongside the unreacted starting
material. Purification using preparative HPLC followed by structure
elucidation showed the products to be the 2-hydroxyphenyl 24
isolated in 11% yield and the 2-methoxyphenyl derivative 25,
which was isolated in 13% yield. Biological testing against CRF-1
showed the 2-methoxyphenyl compound 25 had an IC50 of
52 nM.27 This is more potent than the starting compound 23, one
explanation for this is that the methoxy pushes the phenyl out of
the plane and this change in conformation is important for activity.
The 2-hydroxyphenol forms an internal hydrogen bond with the
diazaindazole core and so the phenyl is held in the plane and this
change in conformation results in the compound being inactive.
The 2-hydroxyphenyl adduct 24 has an internal hydrogen bond
that holds the phenyl ring in the plane and this change in confor-
mation results in the compound being inactive. Another compound
in the series was the 2-fluorophenyl derivative 26; this compound
had an IC50 of 72 nM.
Figure 3.
O
O
O
N
DuBois
catalyst
S
i
Cl
N
F
F
Cl
F
O
O
HO
N
HO
N
N
O
N
O
N
O
Cl
Cl
N
N
N
N
[9% yield}
21
22
MC4 IC50 96nM
cLogP 2.33
clogD7.4 : 2.10
cLogP 1.82
clogD7.4 : 0.41
Scheme 9. Reagents and conditions: (i) PhSeSePh (5 mol %), Urea.HOOH, dichloro-
ethane, microwaves 100 °C, 2 h.
and had reacted to give the ArSF2Ar2 species,22 which was hydro-
lyzed to the sulfoxide 10 on work-up.
Compound 1 was a lead compound for Pfizer’s activin-like
kinase 5 receptor (Alk5) antagonist program (for veterinary medi-
cine). Pyrazole 17 had a short half life and one of the project objec-
tives was to see if fluorination or chlorination could solve this
issue. The pyrazole 17 was submitted to the fluorination screen.
Several wells showed promising product profiles (Fig. 3). Taking
the best, scale-up using accufluorÒ in acetonitrile gave the fluoro-
pyrazole derivative 18 in 53% isolated yield. The half life in pig liver
microsomes (PLM) increased tenfold as well as twofold increase in
selectivity over ActR2B23 as a result of this small change in molec-
ular structure. Surprisingly the chlorinated adduct 19 was twofold
more potent and had a longer half-life as well as having lower
clearance in the in vivo pig model. The dichlorinated compound
20, formed as an impurity in the chlorination reaction, was signif-
icantly less potent, but more selective than the mono-chloro
adduct 19.
In a similar plate based approach to chlorination, N-chlorosuc-
cinimide (NCS) in acetonitrile was found to achieve chlorination
generating two products as well as residual starting material. After
scale-up, structure elucidation and testing it was found that the
two products were the mono-chloro adduct 19 and the dichloro-
adduct 20 where chlorination had occurred on both the pyrazole
and the pyridine.
As the molecule contained an N-phenylpyrazole, it was felt that
the system was set up to attempt Sanford’s palladium-catalyzed
C–H functionalization. The transformation selected was a palla-

M. A. Masood et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1255–1262
1261
O
O
O
N
N
N
N
N
N
i
N
N
N
N
+
N
N
N
N
N
O
OH
(11% yield)
(13% yield)
25
23
CRF1 Ki: 300nM
24
inactive
CRF-1 Ki: 52nM
Scheme 10. Reagents and conditions: (i) Oxone, MeOH/H2O (2:1), Pd(OAc)2 (10 mol %), microwave heating 120 °C, 3 h.
O
In a different project, pyrimidone 28 was a lead compound in
the melanin concentrating receptor 1 (MCHR-1) program. The
pyrimidone 28 was tested in the BMO screen and found to be unre-
acted under all conditions. Pyrimidone 28 was then tested in the
halogenation screen. Chlorination using NCS in acetonitrile/
tetrahydrofuran (4:1) were optimum conditions and gave two
close-running products which, after purification and structure elu-
cidation, were found to be the monochloro- and dichloro-adducts,
29 and 30, respectively (Scheme 12).
O
N
N
N
N
N
i
N
N
N
N
N
F
F
Cl
(15% isolated
yield)
These reactions on 28 demonstrated that selective halogenation
of heterocyclic groups in complex drug molecules was possible.
Examples of compounds where the introduction of oxygen, fluo-
rine and chlorine, have blocked metabolically vulnerable sites have
been highlighted in this communication. However, biomimetic oxi-
dation (BMO) has been previously used effectively to synthesize
drug metabolites.1 Metalloporphyrins are efficient catalysts for
synthesizing drug metabolites because cytochrome p450’s contain
an iron metalloporphyrin in the active site of the enzyme and this
is responsible for the oxidations.1 As part of Pfizer’s oxytocin
antagonist program, a human specific metabolite 32 (Scheme 13)
was identified and the synthesis of the metabolite 32 was needed
for safety studies. The parent compound 31 was tested in the BMO
screen and after scale-up; the hydroxymethyl metabolite 32 was
isolated in 6% yield.
27
CRF-1 : IC50: 52nM
26
CRF-1 : IC50: 72nM
Scheme 11. Reagents and conditions: (i) Pd(OAc)2 (20 mol %), CH3CN, NCS,
microwave heating 160 °C, 2 h.
dium-catalyzed directed chlorination using N-chlorosuccinnimide
(NCS) in acetonitrile.27 The 2-chloro-6-fluoro analogue 27 was iso-
lated in 15% yield (Scheme 11). As predicted, this chloro-derivative
27 was slightly more potent than the starting compound, primarily
due to the slight change in the conformation of the phenyl; that is,
the phenyl would be more perpendicular in the 2,6-disubstituted
compound than the 2-substituted compound 26. The activity of
the products and 2-hydroxy adduct 24 demonstrated that confor-
mation of the phenyl was key SAR.
Cl
Cl
Cl
O
Cl
i
O
N
O
+
N
N
N
Cl
O
N
O
N
N
N
Cl
O
N
OH
OH
29
30
inactive
28
OH
(16%)
(11%)
MCH1: IC50 >9.80 uM
MCH-1: IC 50 0.120 uM
Scheme 12. Reagents and conditions: (i) MeCN/THF (4:1), NCS, rt, 19 h.
O
OH
O
F
F
N
N
N
N
N
i
N
N
N
N
N
N
N
(6%)
31
hOxytocin IC50 : 49.2nM
Scheme 13. Reagents and conditions: (i) Mn(TDClPP) (5 mol %), mCPBA, CH3OH/CH3CN (1:1), rt, 15 h.
32
hOxytocin IC 50 : 919nM

1262
M. A. Masood et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1255–1262
The vial was irradiated in the microwave reactor at 130 °C for 3600–7200 s (1–
Biological testing of the metabolite 32 showed it to be less po-
tent than the starting compound 31.
2 h). Typical yields on drug molecules were ca. 5–20%, it had not been shown
conclusively whether adding more palladium catalyst and reheating improves
the yield.
The solvent was evaporated under reduced pressure in the Genevac to give the
crude reaction mixture, this was taken up in dichloromethane (40 ml) and
washed with brine (saturated, 30 ml), water (30 ml), dried over magnesium
sulphate, filtered and the solvent evaporated to give the crude material. This
was then purified using preparative HPLC.
Transformations on drug molecules and the impact on some
Pfizer medicinal chemistry projects have been elaborated. With
the MCH lead compound, a simple but synthetically-difficult
N-demethylation gave a more potent and more polar compound.
Biomimetic oxidation has been reported as a method for metabo-
lite synthesis, for example the oxytocin lead compound 31 was
converted into the key metabolite 32 using this approach. LD is a
viable and often practical approach for opening up chemical space
and synthesizing key compounds. While the oxidation and BMO
screens may not be successful on most substrates, other screens
such as the halogenation screens can often work. As research into
C–H activation increases and more transformations are discovered,
LD will potentially have a more mainstream use in drug discovery.
Additional chemical transformations on drug molecules via C–H
activation and oxidation have been described and their impact on
drug discovery projects highlighted. The fluoropyrazole P38 com-
pound 14 is another example of selective remote functionalization
of a highly functionalized molecule. The conditions for fluorination
6. Litvinas, N. D.; Brodsky, B. H.; DuBois, J. Angew. Chem., Int. Ed. Engl. 2009, 48,
4513.
7. Doro, F. G.; Lindsay-Smith, J. R.; Ferreira, A. G.; Assis, M. D. J. Mol. Catal. A: Chem.
2000, 164, 97.
8. Cunningham, I. D.; Danks, T. N.; Hay, J. N.; Hamerton, I.; Gunathilagan, S.;
Januczak, C. J. Mol. Catal., A: Chem. 2002, 185, 25.
9. Gonzalves, D’A. A. M. R. J. Heterocycl. Chem. 2002, 34, 499.
10. Ley, S. V.; Baxendale, I. R. Chimia 2008, 62, 162.
11. Park, B. K.; Kitteringham, N. R.; O’Neill, P. M. Annu. Rev. Pharmacol. Toxicol.
2001, 41, 443; Fisher, M. B.; Henne, K. R.; Boer, J. Curr. Opin. Drug Discov. Devel.
2006, 9, 101.
12. Bondi, A. J. Chem. Phys. 1964, 68, 441.
13. Ismail, M. D. J. Fluorine Chem. 2002, 118, 27.
14. Ramsden, C. A.; Smith, R. G. J. Am. Chem. Soc. 1998, 120, 6842.
15. Organofluorine Chemistry. Principles and Commercial Applications; Banks, R. E.,
Smart, B. E., Tatlow, J. C., Eds.; Plenum Press: New York, London, 1996;
Advances in Organic Synthesis Modern Organofluorine Chemistry—Synthetic
Aspects; Laali, K. K., Rahman, A.-U., Eds.; Bentham Science: Amsterdam, 2006;
Vol. 2, Merrit, R. F.; Johnson, F. A. J. Org. Chem. 1966, 31, 1859; Purrington, S. T.;
Kagen, B. S.; Patrick, T. B. Chem. Rev. 1986, 86, 997; Rozen, S. Acc. Chem. Rev.
1988, 21, 307; Chambers, R. D.; Parsons, M.; Sandford, G.; Thomas, E.; Trimcic,
J.; Moilliet, J. S. Tetrahedron 2006, 62, 7162; Chambers, R. D.; Skinner, C. J.;
Thomson, J.; Hutchinson, J. C. S. Chem. Commun. 1995, 17; Chambers, R. D.;
Holling, D.; Sanford, G.; Batsanov, A. S.; Howard, J. A. K. J. Fluorine Chem. 2004,
125, 661.
were found using
efficiently provided optimal methodology.
LD has demonstrated the difficulty in predicting selectivities
and served to highlight purification as a significant hurdle, as the
molecular weight of the drug increases.
a combinatorial plate based screen that
16. Banks, R. E. J. Chem. Soc., Chem. Commun. 1992, 595; Singh, R. P.; Shreeve, J. M.
Acc. Chem. Res. 2004, 37, 31.
Acknowledgments
17. Montgomery, J. A., Jr.; Petersson, G. A.; Al-Laham, M. A.; Mantzaris, J. Chem.
Phys. Lett. 1990, 169, 497.
The authors are greatly indebted to Professor S. V. Ley (Cam-
bridge), Professor J. T. Groves (Princeton), Professor M. P. Doyle
(Maryland), Professor G. Sandford (Durham), Professor E.J. Corey
(Harvard), Professor John Brown (Oxford), Dr Dafydd Owen (PGRD),
Mrs. Rachel Osborne, Dr. David Pryde (PGRD), Dr. Nawaz Khan and
Dr Simon Mantell for very helpful discussions.
18. Lal, G. S.; Pez, G. P.; Syvret, R. G. Chem. Rev. 1996, 96, 1737; Gerstenberger, M. R.
C.; Haas, A. Angew. Chem., Int. Ed. Engl. 1981, 20, 647; Welch, J. T. Tetrahedron
1987, 43, 3123; Borodkin, G. I.; Zaikina, P. A.; Shubin, V. G. Tetrahedron Lett.
2006, 47, 2639; Zhang, Y.; Shibatomi, K.; Yamamoto, H. Synlett 2005, 2837.
19. Dichloroethane was used as an alternative to dichloromethane as it is less
volatile and can be used with liquid handlers.
20. Chambers, R. D.; Spink, R. C. H. Chem. Commun. 1999, 883; Chambers, R. D.; Fox,
M. A.; Sanford, G. Lab Chip 2005, 5, 1132.
21. The plate was sealed up and agitated at room temperature for between 48 and
72 h as this was found to be the optimum reaction time. The reactions were
quenched by the addition of aqueous methanol (1:1) and then analyzed by
LCMS.
22. Marat, R. K.; Janzen, A. F. Can. J. Chem. 1977, 55, 3031.
23. Matsuzaki, T.; Hanai, S.; Kishi, H.; Liu, Z.-H.; Bao, Y.-L.; Kikuchi, A.; Tsuchida, K.;
Sugino, H. J. Biol. Chem. 2002, 277, 19008.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2011.11.033.
References and notes
24. Miller, D. C.; Klute, W.; Calabrese, A.; Brown, A. D. Bioorg. Med. Chem. Lett. 2009,
19, 6144.
25. Chen, X.; Hao, X.-S.; Goodhue, C. E.; Yu, J.-Q. J. Am. Chem. Soc. 2006, 128, 6790.
26. CRF-1 potency is expressed as functional activity measured using CHO cells
(Cell Sciences SNB0000377) expressing recombinant human CRF-1 receptor
grown in DMEM:F12 (1:1) media containing 10% (V/V) Foetal Bovine Serum
(PAA), 400 lg/ml Geneticin (GIBCO-BRL 10131-027) and 1% (V/V) Glutamax in a
cell incubator at 37 °C, 5% CO2 to 70% confluence. FAC hCRF (10 nM) was added
with test compound to 10,000 cells/well in Phosphate Buffered Saline
containing 500 lM FAC IBMX. The functional response was measured using
DiscoveRx HitHunter cAMP II Assay kit (Amersham Biosciences—90-0034-03).
Each compound was tested multiple times using a 0.5 log serial dilution dose–
response with a top final assay concentration of 20 lM. The % response of the
test compound at different test doses were then fitted to a 4-parameter logistic
curve to determine the compound IC50. Ki values were determined from the
IC50 using the Cheng and Prussoff relationship and the EC50 for the agonist
dose–response curve which was determined on the same day.
27. Stowers, K. J.; Sanford, M. S. Org. Lett. 2009, 11, 4584; Kalyani, D.; Dick, A. R.;
Anani, W. Q.; Sanford, M. S. Org. Lett. 2006, 8, 2523.
1. Chorghade, M. S.; Hill, D. R.; Lee, E. C.; Pariza, R. J.; Dolphin, D. H.; Hino, F.;
Zhang, L.-Y. Pure Appl. Chem. 1996, 68, 753; Bernadou, J.; Bonnafous, M.; Labat,
G.; Loiseau, P.; Meunier, B. Drug Met. Dispos. 1991, 19, 360; Segrestaa, J.; Verite,
P.; Estour, F.; Menager, S.; Lafont, O. Chem. Pharm. Bull 2002, 50, 744; Mohajer,
D.; Karimipour, G.; Bagherzadeh, M. New J. Chem. 2004, 28, 740.
2. Wender, P. A.; Hilinski, M. K.; Mayweg, A. V. W. Org. Lett. 2005, 7, 79.
3. Yoneyama, T.; Crabtree, R. H. J. Mol. Catal. A 1996, 108, 35; Sanford, M. S.;
Deprez, N. R. Inorg. Chem. 2007, 46, 1924; Eberson, L.; Gomez-Gonzales, L. J.
Chem. Soc. Chem. Commun. 1971, 263; Zhang, R.; Nagraj, N.; Lansakara, D. S. P.;
Lowell, P. H.; Newcombe, M. Org. Lett. 2006, 8, 2731.
4. Desai, L. V.; Stowers, K. J.; Sanford, M. S. J. Am. Chem. Soc. 2008, 130, 13285;
Kalyani, D.; Deprez, N. R.; Desai, L. V.; Sanford, M. S. J. Am. Chem. Soc. 2005, 127,
7330.
5. Palladium acetate catalyzed O-acetoxylation:
Substrate (2.86 mmol) is suspended/dissolved in glacial acetic acid (2–3 ml) in
a microwave vial. PhI(OAc)2 (3.0 equiv) was added portion-wise followed by
Palladium acetate (10–20 mg, excess). The vial is sealed by crimping on the lid.