Discovery of novel PI3-Kinase δ specific inhibitors for the treatment of rheumatoid arthritis: Taming CYP3A4 time-dependent inhibition

Article abstract of DOI:10.1021/jm3003747

PI3Kδis a lipid kinase and a member of a larger family of enzymes, PI3K class IA(α, β, δ) and IB (γ), which catalyze the phosphorylation of PIP2 to PIP3. PI3Kδ is mainly expressed in leukocytes, where it plays a critical, nonredundant role in B cell receptor mediated signaling and provides an attractive opportunity to treat diseases where B cell activity is essential, e.g., rheumatoid arthritis. We report the discovery of novel, potent, and selective PI3Kδinhibitors and describe a structural hypothesis for isoform (α, β γ?) selectivity gained from interactions in the affinity pocket. The critical component of our initial pharmacophore for isoform selectivity was strongly associated with CYP3A4 time-dependent inhibition (TDI). We describe a variety of strategies and methods for monitoring and attenuating TDI. Ultimately, a structure-based design approach was employed to identify a suitable structural replacement for further optimization.

Full text of DOI:10.1021/jm3003747

Article
pubs.acs.org/jmc
Discovery of Novel PI3-Kinase δ Specific Inhibitors for the Treatment
of Rheumatoid Arthritis: Taming CYP3A4 Time-Dependent Inhibition
Brian S. Safina,*,† Stewart Baker,∥ Matt Baumgardner,† Paul M. Blaney,‡ Bryan K. Chan,†
Yung-Hsiang Chen,† Matthew W. Cartwright,‡ Georgette Castanedo,† Christine Chabot,†
Arnaud J. Cheguillaume,‡ Paul Goldsmith,∥ David M. Goldstein,§ Bindu Goyal,§ Timothy Hancox,∥
Raj K. Handa,‡ Pravin S Iyer,§ Jasmit Kaur,∥ Rama Kondru,§ Jane R. Kenny,† Sussie L. Krintel,‡ Jun Li,†
John Lesnick,† Matthew C. Lucas,§ Cristina Lewis,† Sophie Mukadam,† Jeremy Murray,† Alan J. Nadin,‡
Jim Nonomiya,† Fernando Padilla,§ Wylie S. Palmer,§ Jodie Pang,† Neil Pegg,∥ Steve Price,‡ Karin Reif,†
Laurent Salphati,† Pascal A. Savy,‡ Eileen M. Seward,‡ Stephen Shuttleworth,∥ Sukhjit Sohal,∥
Zachary K. Sweeney,† Suzanne Tay,† Parcharee Tivitmahaisoon,§ Bohdan Waszkowycz,‡ Binqing Wei,†
Qin Yue,† Chenghong Zhang,† and Daniel P. Sutherlin†
†Genentech, Inc., 1 DNA Way, South San Francisco, California 94080, United States
‡Argenta Discovery, 8/9 Spire Green Centre, Flex Meadow, Harlow CM19 5TR, United Kingdom
§Roche Research Center, 340 Kingsland Street, Nutley, New Jersey 07110, United States
∥Piramed Pharma, 957 Buckingham Avenue, Slough, Berks SL1 4NL, United Kingdom
S
* Supporting Information
ABSTRACT: PI3Kδ is a lipid kinase and a member of a larger
family of enzymes, PI3K class IA(α, β, δ) and IB (γ), which
catalyze the phosphorylation of PIP2 to PIP3. PI3Kδ is mainly
expressed in leukocytes, where it plays a critical, nonredundant
role in B cell receptor mediated signaling and provides an
attractive opportunity to treat diseases where B cell activity is
essential, e.g., rheumatoid arthritis. We report the discovery of
novel, potent, and selective PI3Kδ inhibitors and describe a
structural hypothesis for isoform (α, β, γ) selectivity gained
from interactions in the affinity pocket. The critical component
of our initial pharmacophore for isoform selectivity was strongly associated with CYP3A4 time-dependent inhibition (TDI). We
describe a variety of strategies and methods for monitoring and attenuating TDI. Ultimately, a structure-based design approach
was employed to identify a suitable structural replacement for further optimization.
and only a small portion exhibit complete remission.3
INTRODUCTION
Rheumatoid arthritis (RA) affects nearly 1% of the world
■
Furthermore, side effects tend to be serious and opportunistic
infections may arise with fatal consequences.4
population and places a significant burden upon society and
patients.1 RA is a debilitating autoimmune disorder, involving
PI3Kδ is a lipid kinase and a member of a larger family of
enzymes, PI3K class IA(α, β, δ), IB (γ), II, and III, which
chronic inflammation of the synovial membrane lining the
catalyze the phosphorylation of phosphatidylinositol-4,5-bi-
joints. Disease progression leads to pannus formation and
sphosphate (PtdIns(4,5)P2) (PIP2) to produce phosphatidyli-
eventually bone, cartilage, and ligament destruction resulting in
pain, stiffness, and deformity. Disease-modifying antirheumatic
drugs (DMARDs) remain the first line of treatment for the
nositol-3,4,5-triphosphate (PtdIns(3,4,5)P3) (PIP3), subse-
quently initiating a cascade of downstream events to modulate
a variety of biological processes such as cell growth,
proliferation, survival, and differentiation.5 Class I PI3K
isoforms exist as a heterodimer of one regulatory subunit
(p85) with the corresponding catalytic subunit (p110α, p110β,
p110δ). Activation of class IA enzymes occur through a
tyrosine-kinase-dependent mechanism thought to involve
majority of patients due to low cost and ability to retard disease,
especially at early onset. Treatment for RA has made significant
advances over the last several decades, particularly since the
introduction of biological therapies.2 In recent years, targeted
approaches such as TNF blockade, B cell depletion,
costimulation blockade, and interleukin-6 receptor blockade
have led to the control of symptoms and joint inflammation
while retarding or halting joint damage all together. However,
50% of patients still do not respond to treatment with biologics
Received: March 18, 2012
Published: May 24, 2012
© 2012 American Chemical Society
5887
dx.doi.org/10.1021/jm3003747 | J. Med. Chem. 2012, 55, 5887−5900

Journal of Medicinal Chemistry
Article
growth factors and cytokine receptors. Class IA PI3Ks have
been implicated most notably in cancer and are ubiquitously
expressed, with the exception of PI3Kδ, which is mainly
expressed in leukocytes, where in B cells it is activated
downstream of the B cell receptor.6 Studies using mice with
dominant negative mutants of PI3Kδ suggest that PI3Kδ plays a
critical, nonredundant role in B cell receptor mediated
signaling.7 The role of B cell signaling in RA has been
demonstrated by the efficacy of the B cell depleting CD20-
specific monocolonal antibody rituximab8 and in a variety of
arthritis animal models.9 Thus, targeting inhibition of PI3Kδ
with a small-molecule inhibitor provides an attractive
opportunity to treat rheumatoid arthritis. Specificity for
PI3Kδ is desired because the ubiquitously expressed PI3K
isoforms, α and β, have been shown to result in embroyonic
lethality in knockout mice and are thought to play essential
roles in insulin action and platelet aggregation.10 In contrast,
genetic ablation of PI3K isoforms δ and γ in mice yield healthy
and viable animals, with differences mainly in immune
response.10 PI3Kγ, also considered to be a target for the
treatment of inflammatory diseases,11 has not been shown to
play a prominent role in B-cell function.
Synthesis. Compounds 1−39 (Scheme 1) were prepared
from aldehyde 31.19 Intermediates 32−37 were derived from a
reductive amination with the corresponding amine to yield six
distinct aryl chloride intermediates (see Supporting Informa-
tion). Several methods were employed to further functionalize
aryl chloride intermediates 32−37. Compounds 12−14, 16, 18,
and 21 proceeded through stannane intermediates 38−39 to
afford the heteroaryl Stille coupling products. Compounds 1−
2, 4−11, 15, 17, 19−20, and 22−24 were constructed from
either Suzuki couplings with boronic ester/acid partners or
sodium hydride mediated nucleophilic additions in concen-
trated hydrochloric acid. Finally, compounds 25−30 were
accessed through Buchwald couplings or stepwise benzimida-
zole construction via 1,2-diaminobenzene and concentrated
hydrochloric acid followed by dehydrative mediated ring
closure.
RESULTS AND DISCUSSION
■
Biochemical potencies for PI3K isoforms (α, β, δ, γ) were
determined using a competitive fluorescence polarization assay
measuring PI3K mediated formation of PIP3. Time-dependent
inhibition was evaluated by employing two in vitro assays.
Initial screening was through a high throughput “IC50-shift”
dilution assay18 and was carried out by preincubating analogues
(30 min) in human liver microsomes ( NADPH) and using
CYP specific probe metabolite formation via LC-MS/MS to
assess TDI risk. The assay reports a TDI-IC50 (or the shifted
IC50 value following preincubation with NADPH) and a %
change in AUC of the two NADPH inhibition curves. More
detailed assays were performed to quantify the kinetic
parameters kinact, maximum rate of enzyme inactivation, and
KI, concentration supporting half the kinact. TDI-IC50 has been
shown to strongly correlate with kinact/KI ratio,18 thus we report
TDI-IC50 for key structure−activity relationships. Cellular
potencies, detecting pAKT(S473), were determined using an
Ri-1 cell line with added human serum albumin (HSA) and α-
1-acid glycoprotein (AAG).
Previously reported ATP-competitive PI3Kδ specific inhib-
itors12 have been shown to bind to p110δ by inducing an
allosteric “specificity” pocket.13 We describe a second unique
class of ATP-competitive PI3Kδ specific inhibitors that bind in
the affinity pocket and are derived from the morpholinyl
thienopyrimidine scaffold (Figure 1).14,19 This novel class of
Comparing a set of indole and indazole-containing analogues
(Table 1) provided pertinent clues to the origins of both PI3Kδ
specificity and CYP3A4 TDI. Indole-containing analogues 1
and 2, differing only in the piperazine substituent, were found
to possess potent CYP3A4 TDI and ranged in α/δ selectivity
from 22- to 129-fold. In contrast, replacement of the indole
moiety with an indazole in compounds 3 (GDC-0941)19 and 4
resulted in a reduction of PI3Kδ specificity (α/δ = 1- and 9-
fold, respectively) along with no observed CYP3A4 TDI. These
comparisons led us to hypothesize that the indole moiety was a
key contributor to high δ specificity and also responsible for
CYP3A4 TDI. Moreover, piperazine substitution seemed to act
synergistically with the indole moiety in increasing p110δ
specificity, which proved to be a critical observation for future
optimization.20
Residues within the ATP-binding site and affinity pocket are
highly conserved among all four isoforms (α, β, δ, γ) of class I
PI3K lipid kinases. However, several key residue differences do
exist. The difference in residues lying above the conserved
tryptophan residue is notable (Figure 2, Trp760, PI3Kδ
numbering). PI3Kδ contains a small threonine residue
Thr750 which creates a surface we refer to as the tryptophan
shelf. On the other hand, this residue is an arginine residue in
PI3Kα, and a lysine residue in PI3Kβ and PI3Kγ, and the
“shelf” is occluded. Above the hydrophobic region II lies the
hinge region where the morpholine ring makes a vital contact
Figure 1. Discovery of the benzimidazole class of PI3Kδ specific
inhibitors, free of CYP3A4 TDI.
inhibitors features an indole pharmacophore interacting with
the conserved affinity pocket that is essential for high δ
specificity. However, upon further profiling of compounds in
this class, CYP3A4 time-dependent inhibition (TDI) was
discovered, which led to unacceptable predictions for clinical
drug−drug interactions (DDI).
Evaluating the DDI risk for NMEs is especially important in
the treatment of chronic, nonlife- threatening diseases, such as
rheumatoid arthritis, where the majority of patients are being
treated with polypharmacy.15 Early monitoring of CYP3A4 TDI
and the potential of DDI in the drug discovery process has
become critical for the identification of promising clinical
candidates.16 Reversible or mechanism-based inactivation of
CYP3A4 can have significant consequences given that over half
of the marketed drugs are metabolized by this enzyme.
Moreover, mechanism-based inactivation suggests that metab-
olism results in the formation of a reactive metabolite capable
of conjugating biomolecules which may result in idiosyncratic
toxicity.17 In this manuscript, we report the discovery of novel,
potent, and selective benzimidazole containing PI3Kδ inhib-
itors that were discovered during our efforts to eliminate time-
dependent CYP3A4 inhibition.
5888
dx.doi.org/10.1021/jm3003747 | J. Med. Chem. 2012, 55, 5887−5900

Journal of Medicinal Chemistry
Article
a
Scheme 1. Synthesis of Compounds 1−39
a
Reagents and conditions: (a) amine, 1,2-DCE, Na(OAc)3BH, 46−95%; (b) (Bu3)2Sn2, PdCl2{PtBu2(Ph-p-NMe2)}2, dioxane, microwave, 140−160
°C, 20 min, 54%; (c) aryl bromide, CuI, Pd(PPh3)4, THF, microwave, 140 °C, 20 min, 9−63%; (d) boronic acid/ester, Pd(PPh3)2Cl2, NaHCO3,
H2O, CH3CN, microwave, 130−150 °C, 15−30 min, 26−83%; (e) indazole/indole, 0 °C, NaH, DMF, 30 min then 150 °C, 1 h, 27−49%; (f)
benzimidazole, Pd2(dba)3, XPhos, Cs2CO3, DMF, microwave, 150 °C, 30 min, 42%; (g) benzimidazole, concd HCl, 1,4-dioxane, microwave, 150 °C,
40 min, 37−63%; (h) (i) 1,2-diaminobenzene, concd HCl, 1,4-dioxane, 90 °C, 3d, (ii) triethylorthoformate, 150 °C, 6 h, 26%.
tryptophan shelf was supported by isoform residue differences,
the indole and indazole moiety both interacted with the
conserved affinity pocket. Previous reports19b of the cocrystal
structure containing 3 in PI3Kγ suggested that the indazole 2-
nitrogen participates in a hydrogen bond with the phenol of
Tyr867 and the indazole NH forms an interaction with Asp841
(Figure 3a). We theorized that the gain in selectivity resulting
from the replacement of an indazole with an indole could be
attributed to removing the strong hydrogen bonding
interaction of 3 with Tyr867 and replacing it with a methine
group (Figure 3b). Given that the residues in the affinity pocket
are conserved throughout the class I isoforms, the precise
consequences of this subtle hydrogen bonding reorganization is
not fully understood. We hypothesized, the effect of disrupting
the hydrogen bond with Tyr867 must radiate and extend past
the conserved affinity pocket, resulting in undesirable
conformational changes for the antitargets (α, β, γ), thereby
providing PI3Kδ specificity. This hypothesis laid the foundation
for future design of novel indole replacements.
Table 1. Inhibition of PI3Kα,β,δ,γ and CYP3A4-TDI for
Indole and Indazole Containing Compounds 1−4
While replacing the indazole unit with an indole moiety led
to improvements in specificity, dramatic differences in time-
dependent CYP3A4 inhibition were also observed. As
previously mentioned, initial structure−activity relationships
implicated the indole ring as the functional group responsible
for CYP3A4 TDI, independent of the N-piperazine substitution
(Table 1). Compound 5, Table 2, exhibited good affinity for
PI3Kδ (3 nM) and excellent (>100-fold) isoform selectivity
(α/δ = 194, β/δ = 102, γ/δ = 1877) along with a clean profile
across a broad kinase panel (Invitrogen 240 kinases @ 1 μM).
Compound 5 was selected for further evaluation and was found
to inactivate CYP3A4 with an IC50 of 1.5 μM when
preincubated (30 min) with human liver microsomes. Further
measurement of the maximal inactivation rate (kinact = 0.074
to Val828 and the affinity pocket where the indole moiety
makes critical interactions12 (vide infra). On a practical note,
PI3Kδ specific inhibitors cocrystallized in the PI3Kγ isoform
require additional modeling of the bound ligand in the context
of a PI3Kδ homology structure. Nonetheless, the thienopyrmi-
dine core serves as a good template for projecting substituents
toward the tryptophan shelf and affinity pocket, setting the
stage for further optimization of p110δ specificity.
min−1) and inactivator potency (KI = 6.59 μM) yielded a kinact
/
The increase in PI3Kδ specificity resulting from a single-
point replacement of a nitrogen atom with a methine group
demanded further analysis of the ligand−protein interactions.
Whereas our understanding of selectivity derived from the
KI ratio of 11.0 mL/min/μmol, confirming time-dependent
inhibition of CYP3A4. DDI assessment was performed using
Symcyp, a population-based ADME simulator.21 Simulations
were performed where 5 (500 mg BID for 5 days) was
5889
dx.doi.org/10.1021/jm3003747 | J. Med. Chem. 2012, 55, 5887−5900

Journal of Medicinal Chemistry
Article
Figure 2. Model of structure 12 (structure shown in Table 3) bound to (a) PI3Kδ (PDB entry 2WXP), highlighting the tryptophan shelf and (b)
PI3Kα, highlighting the tryptophan shelf being occluded by Arg75.
Figure 3. Model comparing indazole and indole interactions in the affinity pocket of PI3Kδ. (a) Compound 3 highlighting interactions with Tyr867
and Asp841. (b) Compound 12, lacking a hydrogen bond acceptor, resulted in greater PI3Kδ specificity.
Table 2. Time-Dependent Inhibition of CYP3A4 in Human Liver Microsomes and Pharmacokinetics in Male HW Rats, 0.5 mpk
iv and 2 mpk po for Compounds 5 and 6
TDI-CYP3A4(t)
compd
IC50 (μM)
kinact (1/min)
KI (μM)
kinact/KI
%F
CL (mL/min/kg)
Vdss (L/kg)
T1/2
5
6
1.5
7.0
0.074
0.048
6.6
78
11
75.2
75.5
31.3
25.2
2.38
2.76
1.48
1.60
0.60
coadministered with a single dose of midazolam (15 mg, day 5),
resulting in predicted 20-fold increase in midazolam AUC.
Compounding matters further, cytochrome P450 (CYP)
phenotyping indicated that 5 was predominately metabolized
by CYP3A4 and would consequently suffer from autoinhibition
and have a dramatic effect on its own clearance upon multiple
dosing. As a result of these findings, 5 was deemed unsuitable
for further development and we began to develop a strategy to
monitor and reduce TDI for indole-containing analogues.
In search of a structural hypothesis to guide target design and
identify reactive metabolites, we turned to metabolite
identification studies (MetID). Compound 5 was incubated
in human and rat liver microsomes in the presence of NADPH,
UDPGA, and GSH, followed by fragment analysis by mass
spectrometry (Figure 4). The major metabolites were found to
result from the oxidation of the morpholine ring (1.6% human,
5.2% rat (M1 and M2)), followed by oxidation of the
morpholino thienopyrimidine core and indole moiety (1.3%
human, 1.4% rat (M3)). A small fraction of glucuronide
conjugate of an oxidative metabolite was also observed (0.2%,
M4) in both human and rat liver microsomes. Glutathione
conjugate was not observed in in vitro MetID, but in vivo, a
minimal amount (<100 fold to parent) of glutathione
conjugation to M3 was identified in rat bile (30 mg/kg PO).
Struck by the lack of metabolism of the piperazine group in
compound 5, we speculated that we might redirect metabolism
from the thienopyrimidine core and indole ring (M2, Figure 4)
to the piperazine in order to diminish TDI.22 Compound 6
(Table 2), possessing a fused morpholino piperazine group, was
identified as a moderate affinity inhibitor of PI3Kδ (9 nM),
which also possessed moderate (10−100-fold) isoform
selectivity (α/δ = 42, β/δ = 69, γ/δ = 806). Upon comparison
5890
dx.doi.org/10.1021/jm3003747 | J. Med. Chem. 2012, 55, 5887−5900

Journal of Medicinal Chemistry
Article
literature.23 Consistent with observations from MetID studies
(compound 5, Figure 4) in which we observed oxidation (M3)
followed by glucuronidation (M4), we focused on several
possible sites for cytochrome P450 mediated oxidation of the
indole ring to serve as the foundation for target design.
Depending upon substitution and/or electronic effects, the
indole ring can undergo a myriad of transformations potentially
leading to reactive metabolites capable of inactivating
cytochrome P450 (Figure 6). Mechanistically, cytochrome
Figure 4. MetID of compound 5 in liver microsomes.
of metabolites of compounds 5 and 6 in rat and human liver
microsomes (Figure 5), it was found that metabolism was
Figure 6. Reactive metabolites formed from CYP450 mediated
oxidation of the indole ring, potentially causing CYP3A4 TDI.
P450 oxidation is thought to take place through several reaction
manifolds involving heme iron and oxygen.24 All of the
proposed reactive metabolites are susceptible to nucleophilic
addition, and we postulated that blocking this metabolism
would lead to a reduction in CYP3A4 TDI.
Thus, exploration began by decorating the indole ring in
order to block possible sites of aromatic hydroxylation (C5−C7
positions) without disrupting the C-2/3 position of the indole
ring critical for δ specificity and affinity. Compounds 7−9
(Table 3) demonstrate that substitution at C-5, C-6, and C-7
positions of the indole ring had little impact on PI3Kδ potency
and isoform selectivity. Unfortunately, there was also no
improvement in TDI. Moving toward the C-2 and C-3
positions of the indole ring, we examined the consequence of
methylating the C-2 position (Table 3, compound 10), further
perturbing our postulated interaction with Tyr867. Interest-
ingly, the increased size and lipophilicity of the methyl group
was well tolerated, judged by the resulting potency and
selectivity, further reinforcing our Tyr867 selectivity hypothesis.
Methylation of the indole nitrogen gave compound 11 which
possessed no TDI and thus proved to be very informative,
further implicating the indole as a reactive metabolite
progenitor. More specifically, these results led us to speculate
that N-hydroxylation could be responsible for CYP inactivation,
although the exact mechanism and structure of the reactive
metabolite responsible remained unclear. Regardless, disrupting
this key N−H interaction ablated PI3Kδ potency and thus
served no practical utility. Analogues containing C-3 sub-
stituents were not synthesized because computational analysis
predicted unfavorable torsional angles (data not shown)
between the indole ring and bicycle core and would have a
detrimental impact on potency and/or selectivity. Our attention
was then turned to a set of azaindole analogues (Table 3,
compounds 12−14) designed to reduce calculated lipophilicity
(cLogP), thereby warding off metabolism and/or altering the
ring electronics and attenuating CYP related reactivity.
Gratifyingly, 5-azaindole analogue 12 resulted in no TDI
(IC50 >10 uM) while both 6- and 7-azaindoles (13 and 14)
Figure 5. MetID of compound 6 in liver microsomes.
directed away from the thienopyrimidine core and indole
moiety and predominantly toward the fused morpholino
piperazine group, M2 (2.4% human, 0.7% rat) and M3
(10.9% human, 26.8% rat).
In kinetic measurements, the ratio of kinact/KI for 6 (Table 2)
indicated that it was 18-fold less potent than for 5, which
further strengthened our correlation between TDI IC50 and
k
inact/KI and allowed us to use TDI IC50 as a primary
optimization parameter. Further comparisons of the in vivo rat
pharmacokinetic profile for the two compounds (Table 2)
showed that the clearance (∼30 mL/min/kg) and bioavail-
ability (75%) were equivalent. The fused morpholino
piperazine group was successful in diverting metabolism away
from the thienopyrimidine core and indole ring while
maintaining moderate in vivo clearance and reducing
CYP3A4 TDI.
Ultimately, the strategy to redirect metabolism away from the
indole, while maintaining high selectivity and permeability,
proved restrictive and modification of the indole heterocycle
was deemed necessary. This presented a challenge in that the
unique specificity for PI3Kδ resulted largely from the indole
(vide supra). Cytochrome P450-catalzyed oxidative metabolism
of the indole ring has been extensively reported in the
5891
dx.doi.org/10.1021/jm3003747 | J. Med. Chem. 2012, 55, 5887−5900

Journal of Medicinal Chemistry
Article
After an extensive effort to minimize TDI while preserving
the indole ring, we undertook the challenge of finding a suitable
replacement group, which was TDI free and had suitable
potency and α/δ isoform selectivity to serve as a starting point
for further optimization. Guided by our structurally driven
hypothesis for PI3Kδ specificity, we set out to synthesize a
variety of analogues exploring diverse indole replacements
capable of filling the affinity pocket while minimizing strong
hydrogen bonding interactions with Tyr867. Several key design
elements were extracted from a more detailed analysis of the
indole bound in the affinity pocket, Figure 7.
Table 3. Inhibition of PI3Kα, β, δ, γ, and CYP3A4 Time-
Dependent Inhibition for Compounds 7−14
The affinity pocket encapsulating the indole moiety was
generally lipophilic and displayed additional space to
accommodate substituents for further gains in affinity. We
also attempted to match ground-state conformations with
bound conformations by targeting indole replacements with
calculated torsional angles similar to that observed in the X-ray
cocrystal structure of 12 (torsional angle = 30°). In addition to
probing interactions with Tyr867, we were cognizant of Lys779
and the potentially favorable opportunities that it presented. In
search of a δ specific pharmacophore, we primarily monitored
α/δ selectivity, which tended to be a good predictor for overall
isoform selectivity. We selected the dimethylamino piperidine
side chain (Table 4) because it was suitable for making direct
analogue comparisons in α/δ selectivity, exemplified by
fluoroindole 15 (δ = 1.8 nM, α/δ = 226). Indolinone 16
challenged the Tyr selectivity hypothesis by attempting to
engage the phenol with a hydrogen bond acceptor (HBA),
which indeed translated into diminished selectivity (α/δ = 8)
and overall lower affinity (δ = 41 nM). Drastic simplification to
an o-toluene moiety resulted in 17, which had surprisingly good
selectivity (α/δ = 98) and moderate affinity for PI3Kδ.
Predictably, however, lipophilicity was greatly increased and
cellular potency was severely affected. Further modifications of
17 revealed a sharp SAR trend, and thus this series was deemed
unsuitable for a focused optimization effort.
Indole-like mimics, which preserved the general 6,5-
heterocyclic framework, such as imidazopyridine 18, were
instructive, because the sp2 C−H subunit pointing toward the
tyrosine was preserved. Replacing the indole’s N−H (HBD)
with a hydrogen bond acceptor (HBA) in 18 highlights the
important interactions with the Asp841 water network.
Moreover, compounds such as 18, which contain a H-bond
acceptor in place of the H-bond donor of the indole were
resulted in moderate to strong TDI. This trend, increasing TDI
potency, appeared to correlate with the reduction in calculated
pKa values for azaindoles 12 (8.77), 13 (7.44), and 14 (4.52).25
Unfortunately, the 5-azaindole 12 and subsequent analogues
were unsuitable for progression due to generally poor cellular
potency, low permeability, and only moderate isoform
selectivity.
Figure 7. Analysis of the affinity pocket in the X-ray cocrystal structure of indole 12 (PI3Kγ). Lipophilic (brown) and hydrophilic (blue).26.
5892
dx.doi.org/10.1021/jm3003747 | J. Med. Chem. 2012, 55, 5887−5900

Journal of Medicinal Chemistry
Article
Table 4. Inhibition of PI3Kα and δ, pAKT Ri-1 Inhibition for Compounds 15−26
generally less potent and ultimately had diminished PI3Kδ
affinity and specificity, presumably due to nitrogen lone pair-to-
Asp repulsion. Similarly, isoquinoline 19 possessed low δ
affinity (29 nM) but its high α/δ selectivity (144 fold)
warranted exploration of heteroaromatic bicycles and phenyl
groups properly oriented within the affinity pocket.
pAKT-Ri-1/PI3Kδ = 16-fold). From a structural perspective,
the affinity pocket seemed to accommodate a wide array of
ligand diversity, but the benzimidazole moiety was also
uniquely positioned to forge favorable interactions with
Lys779 (Figure 8), similar to what was observed with azaindole
12. Ultimately, the benzimidazole moiety possessed all of these
unique characteristics required for progression.
Modeling suggested that a variety of 5,6-heteroaromatic
systems (compounds 20−26) could occupy the affinity pocket
and provide sufficient potency and selectivity. In general, all of
the analogues provided moderate α/δ selectivity, which was
acceptable, especially as a starting point. Several additional key
features, however, played a role in our selection of a suitable
indole replacement. Heterocycles capable of further function-
alization at the 2-C position (i.e., 23 and 24, 25) were
investigated primarily for their ability to provide additional
handles for optimization. We envisioned that substitution at
this position would provide opportunities to make favorable
protein interactions and/or adjust the torsional angles of our
inhibitors. For example, 2-Me substituted benzimidazole 26
gave rise to a 10-fold improvement in p110δ affinity over the
unsubstituted (2-H) benzimidazole 25. When deciding which
heterocycle to subject to further SAR, we chose to avoid highly
lipophilic and potentially toxicophoric groups such as
thiophene 20 and indoles 23 and 24, capable of CYP mediated
N-hydroxylation. We favored benzimidazole 25 for its low
cellular shift (pAKT-Ri-1/PI3Kδ = 42-fold), which compared
favorably to the indole class of analogues, (e.g., compound 15,
Figure 8. Overlay of cocrystal stuctures in PI3Kγ (compounds 26,
pink, and 12, green) comparing binding modes of indole and
benzimidazole containing analogues.
5893
dx.doi.org/10.1021/jm3003747 | J. Med. Chem. 2012, 55, 5887−5900

Journal of Medicinal Chemistry
Article
Table 5. Inhibition of PI3Kα, β, δ, γ, CYP3A4 TDI and pAKT Ri-1 for compounds 26−30
While monitoring for CYP3A4-TDI, we found that the
dimethylamino piperidine side chain, possessing a very high pKa
(∼10), tended to reduce TDI, possibly masking it from
compounds 16−26 (e.g., indole 15 exhibited CYP3A4-TDI =
9.1 μM). Thus, to further strengthen our belief that a suitable
TDI-free indole replacement was being chosen, a set of
benzimidazole analogues, all possessing groups with reduced
basicity, were synthesized 27−30 (Table 5) all confirming a
TDI-IC50 >10 μM. In addition, the cellular potency was within
the desired range (<200 nM) for the majority of analogues (i.e.,
28−30). Benzimidazole compound 27 possessed a desirable
cLogP (2.3) and moderate predicted human liver microsome
stability (10 mL/min/kg), equivalent to indole compound 5
(10 mL/min/kg).
X-ray crystallographic analysis of benzimidazole 26 revealed a
shift in binding mode, compared to azaindole 12 (Figure 8,
overlay of compounds 26 and 12). The single point contact of
the morpholine hinge binder allowed for a clockwise pivot of
the ligand, affording new opportunities and interactions in both
the affinity pocket (e.g., Lys779) and “tryptophan” shelf.
Reoptimization of the central heterocycle and highly basic
functionality of the dimethylamino piperazine group was also in
order. To this end, introduction of a gem-dimethylpiperazine
moiety (compound 30), designed to fill the nearby tryptophan
shelf, led to excellent selectivity over PI3Kβ (β/δ = 815) and
moderate selectivity over PI3Kγ and α (γ/δ = 112 and α/δ =
50). Thus having fulfilled all of our requirements (reduced TDI,
excellent potency and moderate specificity), the benzimidazole
class of indole replacements was perfectly positioned for further
optimization.
foundation germane to constructing a hypothesis. We
successfully implemented several strategies such as installing
metabolic soft spots to divert metabolism, blocking potential
sites of metabolism leading to reactive metabolite formation
and modifying physiochemical properties (pKa and cLogP) to
reduce TDI. Ultimately, identification of an indole-containing
compound with reduced TDI combined with acceptable
potency, selectivity, and overall favorable drug-like properties
was elusive.
Through the use of cocrystal structures in PI3Kγ and
structure-based design, our understanding of the principles
driving PI3Kδ potency and specificity allowed for the
exploration of several novel pharmacophores to replace the
indole moiety and test our selectivity hypothesis. On the basis
of the discovery of several novel, potent PI3Kδ specific indole
replacements, we selected the benzimidazole group for its
favorable drug-like properties (i.e., metabolic stability, cellular
potency, substitution, and lack of CYP3A4 TDI). Further
cocrystal structural analysis of benzimidazole analogues
exhibited a distinct pivot about the morpholine hinge, in
contrast to the indole-containing inhibitors, providing further
opportunities for the development of PI3Kδ specific
benzimidazole inhibitors in rheumatoid arthritis, details of
which will be described in due time.
EXPERIMENTAL SECTION
■
μm All solvents and reagents were used as obtained. 1H NMR spectra
were recorded with a Bruker Avance DPX400 spectrometer or a
Varian Inova 400 NMR spectrometer and referenced to tetramethylsi-
lane. Chemical shifts are expressed as δ units using tetramethylsilane as
the internal standard (in NMR description: s, singlet; d, doublet; t,
triplet; q, quartet; m, multiplet; br, broad peak). All coupling constants
(J) are reported in hertz. Mass spectra were measured with a Finnigan
SSQ710C spectrometer using an ESI source coupled to a Waters
600MS HPLC system operating in reverse mode with an X-bridge
phenyl column of dimensions 150 mm × 2.6 mm, with 5 μm sized
particles. Analytical purity was >95% unless stated otherwise. The
following analytic method was used, unless stated otherwise, HPLC-
CONCLUSION
■
A novel class of PI3Kδ specific benzimidazole inhibitors has
been discovered through our efforts to reduce CYP3A4 time-
dependent inhibition of indole containing analogues. Initial
SAR and metabolism identification studies with compound 5
implicated the indole ring as causing CYP3A4 TDI and laid the
5894
dx.doi.org/10.1021/jm3003747 | J. Med. Chem. 2012, 55, 5887−5900

Journal of Medicinal Chemistry
Article
Agilent 1200, water with 0.05% TFA, acetonitrile with 0.05% TFA,
Agilent Zorbax SD-C18, 1.8 μM, 2.1 mm × 30 mm, 40 °C, 3−95%B in
8.5 min, 95% in 2.5 min, 400 μL/min, 220 nm and 254 nm, equipped
with Agilent quadrupole 6140, ESI positive, 110−800 amu. Microwave
reactions were performed using a Biotage initiator reactor. Compound
3 and intermediates 31 and 33 have been previously described.19
Procedures for unreported compounds and intermediates 32−39 are
described in the Supporting Information.
(R)-8-((2-(5-Fluoro-1H-indol-4-yl)-4-morpholinothieno[3,2-
d]pyrimidin-6-yl)methyl)octahydropyrazino[2,1-c][1,4]oxazine
(6). A mixture of (R)-8-(2-chloro-4-morpholin-4-ylthieno[3,2-d]-
pyrimidin-6-ylmethyl)octahydropyrazino[2,1-c][1,4]oxazine (34)
(200 mg, 0.49 mmol), Cs2CO3 (318 mg, 0.98 mmol), 1-(tert-
butyldimethylsilanyl)-5-fluoro-4-(4,4,5,5-tetramethyl-[1,3,2]-
dioxaborolan-2-yl)-1H-indole (238 mg, 0.63 mmol), and tetrakis-
(triphenylphosphine)palladium (56 mg, 0.05 mmol) in dioxane/water
(8 mL/4 mL) was heated at 140 °C for 25 min in a microwave reactor.
The resulting cooled mixture was loaded onto an SCX column, and
the column was washed with DCM and methanol. The product was
eluted with 2 M NH3 in methanol, and the product fractions were
concentrated in vacuo. The residue obtained was purified by column
chromatography on silica using 1−10% 2 M NH3 in EtOAc as eluant
2-(4-((2-(1H-Indol-4-yl)-4-morpholinothieno[3,2-d]-
pyrimidin-6-yl)methyl)piperazin-1-yl)-2-methylpropanamide
(1). A 20 mL microwave vial was charged with a suspension of 2-(4-
((2-chloro-4-morpholinothieno[3,2-d]pyrimidin-6-yl)methyl)-
piperazin-1-yl)-2-methylpropanamide (32) (see Supporting Informa-
tion for procedure) (0.293 g, 0.67 mmol), 4-(4,4,5,5-tetramethyl-
[1,3,2]dioxaborolan-2-yl)-1H-indole (0.21 g, 0.87 mmol), cesium
carbonate (0.472 g, 1.34 mmol), and Pd(PPh3)4 (0.077 g, 0.067
mmol) in dioxane (8 mL) and water (2 mL). The reaction mixture was
heated in a microwave for 15 min at 140 °C. The crude reaction
mixture was loaded onto an SCX-2 cartridge, and the cartridge was
washed with methanol; the product was eluted with 2N ammonia in
methanol. The solution was concentrated, and the residue was further
purified by chromatography (Si-PPC) using 0−5% MeOH in DCM as
eluant. The fractions containing the product were concentrated in
vacuo. The residue was crystallized from EtOH to give the product as a
white crystalline solid (100 mg, 28%). 1H NMR (CDCl3) δ 8.71 (br s,
1H), 8.16 (d, J = 7.5, 1.0 Hz, 1H), 7.49−7.51 (m, 1H), 7.44 (dt, J =
8.0, 1.0 Hz, 1H), 7.35 (s, 1H), 7.25−7.30 (m, 2H), 7.09 (br d, J = 5.5
Hz, 1H), 5.65 (br d, J = 5.5 Hz, 1H), 4.07 (t, J = 4.0 Hz, 4H), 3.89 (t, J
= 4.0 Hz, 4H), 3.81 (s, 2H), 2.50−2.64 (m, 8H), 1.22 (s, 6H). LCMS
m/z [M + H]+ 520.
1
to give compound 6 as a yellow solid (207 mg, 83%). H NMR (400
MHz, CDCl3) δ 1.91 (m, 1 H), 2.36−2.47 (m, 4 H), 2.63−2.78 (m, 3
H), 2.95 (d, J = 8.6 Hz, 1 H), 3.25 (m, 1 H), 3.62−3.74 (m, 2 H),
3.81−3.89 (m, 7 H), 4.05 (t, J = 4.7 Hz, 4 H), 6.88−6.90 (m, 1 H),
7.05 (dd, J = 10.9, 8.8 Hz, 1 H), 7.28 (m, 1 H), 7.34−7.39 (m, 2 H)
and 8.22 (br s, 1H). RT: 6.61 min. LCMS m/z [M + H]+ 509.2.
Analytical LC-MS on an Agilent (1200/G6140) using a 1.8uM, 2.1
mm × 30 mm ZORBAX SD-C18 analytical column, 40 °C, 400 μL/
min, and H2O/MeCN modified with 0.05% trifluoroacetic acid
running a linear gradient from 3% MeCN to 95% MeCN in 10 min
monitoring by UV wavelength 254 nm and ESI+ TIC MS showed 97%
purity.
2-(4-((2-(6-Fluoro-1H-indol-4-yl)-4-morpholinothieno[3,2-d]-
pyrimidin-6-yl)methyl)piperazin-1-yl)-2-methylpropanamide
(7). The title compound was prepared from 32 in a manner similar to
the Suzuki procedure described for compound 1 to give compound 7
1
as an off-white solid (95 mg, 65%). H NMR (400 MHz, CDCl3) δ
4-(2-(1H-Indol-4-yl)-6-((4-(methylsulfonyl)piperazin-1-yl)-
methyl)thieno[3,2-d]pyrimidin-4-yl)morpholine (2). The title
compound was prepared from 33 in a manner similar to the Suzuki
1.26 (s, 6H), 2.63 (m, 8H), 3.87 (s, 2H), 3.93 (t, J = 4.8, 4H), 4.10 (t, J
= 4.8, 4H), 5.18 (br s, 1H), 7.10 (br s, 1H), 7.20 (dd, J = 1.1, 6.5, 1H),
7.32 (m, 1H), 7.38 (s, 1H), 7.56 (s, 1H), 8.99 (dd, J = 2.3, 8.8, 1H),
8.26 (br s, 1H). LCMS m/z [M + H]+ 538.4.
2-(4-((2-(6-Cyano-1H-indol-4-yl)-4-morpholinothieno[3,2-d]-
pyrimidin-6-yl)methyl)piperazin-1-yl)-2-methylpropanamide
(8). The title compound was prepared from 32 in a manner similar to
the Suzuki procedure described for compound 1 to give compound 8
as a cream solid (41 mg, 47%). 1H (400 MHz, CDCl3) δ 1.27 (s, 6H),
2.63 (m, 8H), 3.88 (s, 2H), 3.94 (t, J = 4.7, 4H), 4.11 (t, J = 4.7, 4H),
5.21 (br s, 1H), 7.12 (br s, 1H), 7.39 (s, 1H), 7.55 (t, J = 5.6, 1H),
7.69 (s, 1H), 7.82 (s, 1H), 8.47 (s, 1H), 8.60 (br s, 1H). LCMS m/z
[M + H]+ 545.4.
1
procedure described for compound 1 to give compound 2. H NMR
(400 MHz, CDCl3) δ 2.67−2.71 (4H, m), 2.81 (3H, s), 3.29−3.33
(4H, m), 3.89 (2H, s), 3.89−3.93 (4H, m), 4.08−4.12 (4H, m), 7.28−
7.33 (m, 2H), 7.39 (s, 1H), 7.50 (d, J = 8.2 Hz, 1H), 7.53−7.54 (m,
1H), 8.19 (d, J = 8.0 Hz, 1H), 8.28 (br s, 1H). LCMS m/z [M + H]+
513.2. LCMS 30 min. CAD GRADIENT, 0.7 mL/min on Agilent
1200/G6140 system). Purity is 100% by UV 254.
2-(4-((2-(1H-Indazol-4-yl)-4-morpholinothieno[3,2-d]-
pyrimidin-6-yl)methyl)piperazin-1-yl)-2-methylpropanamide
(4). The title compound was prepared from 32 in a manner similar to
the Suzuki procedure described for compound 1 to give compound 4
(40 mg, 40%). 1H NMR (500 MHz, CDCl3) δ 10.25 (s, br, 1H), 9.01
(s, 1H), 8.26 (d, 1H), 7.62 (m, 1H), 7.51 (m, 1H), 7.36 (s, 1H), 7.12
(m, 1H), 5.25 (m, 1H), 4.11 (m, 4H), 3.90 (m, 4H), 3.83 (s, 2H), 2.62
(s, br, 8H), 1.25 (s, 6H). RT: 1.94 min. LCMS m/z [M + H]+ 521.1.
Analytical LC-MS on an Agilent (1200/G6140) using a 1.8 μM, 2.1
mm × 30 mm ZORBAX SD-C18 analytical column, 40 °C, 400 μL/
min, and H2O/MeCN modified with 0.05% trifluoroacetic acid
running a linear gradient from 3% MeCN to 95% MeCN in 10 min
monitoring by UV wavelength 254 nm and ESI+ TIC MS showed
100% purity.
2-(4-((2-(5,7-Difluoro-1H-indol-4-yl)-4-morpholinothieno-
[3,2-d]pyrimidin-6-yl)methyl)piperazin-1-yl)-2-methylpropa-
namide (9). To a solution of 2-(4-((2-chloro-4-morpholinothieno-
[3,2-d]pyrimidin-6-yl)methyl)piperazin-1-yl)-2-methylpropanamide
(32) (66.06 mg, 0.1505 mmol) in acetonitrile (3 mL) in a microwave
reaction vessel was added 1-(tert-butoxycarbonyl)-5,7-difluoro-1H-
indol-4-ylboronic acid (89.4 mg, 0.301 mmol) (see Supporting
Information for procedure), bis(triphenylphosphine)palladium(II)
chloride (5.28 mg, 0.00752 mmol), and a 1 M solution of sodium
carbonate (47.8 mg, 0.452 mmol) in water. The tube was sealed and
heated at 130 °C under microwave irradiation for 30 min. The solvent
was evaporated, and the residue was purified by column chromatog-
raphy (Si-PPC, MEOH/DCM, gradient 0:100 to 20:80) followed by
reverse phase HPLC (Phenomenex Gemini 5 μ C18, 20 mM
triethylamine in water on a gradient of acetonitrile 70:30 to 40:60).
2-(4-((2-(5-Fluoro-1H-indol-4-yl)-4-morpholinothieno[3,2-d]-
pyrimidin-6-yl)methyl)piperazin-1-yl)-2-methylpropanamide
(5). The title compound was prepared from 32 and 1-(tert-
butyldimethylsilyl)-5-fluoro-4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-
2-yl)-1H-indole (see Supporting Information for procedure) in a
manner similar to the Suzuki procedure described for compound 1 to
1
LCMS (10 min run): RT 3.17 min, [M + H]+ 556.2. H NMR (400
1
MHz, DMSO-d6) δ 11.77 (br s, 1H), 7.52−7.47 (m, 1H), 7.72−7.36
(m, 1H), 7.10−6.99 (m, 2H), 6.97−6.90 (s, 1H), 6.83−6.78 (m, 1H),
3.96−3.89 (m, 4H), 3.89−3.83 (br s, 2H), 3.80−3.73 (m, 4H), 2.67 (s,
1H), 2.34−2.31 (s, 1H), 2.30−2.27, m, 2H), 2.62−2.54 (m, 4H),
1.14−1.03 (m, 6H). Analytical LC-MS on an Agilent (1200/G6140)
using a 1.8 μM, 2.1 mm × 30 mm ZORBAX SD-C18 analytical
column, 40 °C, 400 μL/min, and H2O/MeCN modified with 0.05%
trifluoroacetic acid running a linear gradient from 3% MeCN to 95%
MeCN monitoring by UV wavelength 254 nm and ESI+ TIC MS
showed 100% purity.
give compound 5 obtained as a white solid (77 mg, 78%). H NMR
(400 MHz, CD3OD) δ 1.90 (s, 6H), 2.56−2.64 (m, 8H), 3.82 (m,
4H), 3.87 (m, 2H), 4.02 (m, 4H), 6.58 (dd, J = 0.8, 3.7, 1H), 6.95 (dd,
J = 8.8, 11.0, 1H), 7.28 (s, 1H), 7.30 (s, 1H) and 7.41 (ddd, J = 0.8,
4.0, 8.8, 1H). RT: 7.44 min. LCMS m/z [M + H]+ 538.4. Analytical
LC-MS on an Agilent (1200/G6140) using a 1.8uM, 2.1 mm × 30 mm
ZORBAX SD-C18 analytical column, 40 °C, 700 μL/min, and H2O/
MeCN modified with 0.05% trifluoroacetic acid running a linear
gradient from 2% MeCN to 98% MeCN in 30 min monitoring by UV
wavelength 254 nm and ESI+ TIC MS showed 100% purity.
5895
dx.doi.org/10.1021/jm3003747 | J. Med. Chem. 2012, 55, 5887−5900

Journal of Medicinal Chemistry
Article
2 -M et hyl -2 -(4- ((2- (2- met hyl -1 H -i ndol -4- yl )-4-
morpholinothieno[3,2-d]pyrimidin-6-yl)methyl)piperazin-1-
yl)propanamide (10). The title compound was prepared from 32
and 2-methyl-4-(4,4,5,5-tetramethyl-[1,3,2]dioxaborolan-2-yl)-1H-in-
dole (see Supporting Information for procedure) in a manner similar
to the Suzuki procedure described for compound 1 to give compound
10 as a white solid (0.295 g, 62%). 1H NMR (400 MHz, DMSO-d6) δ
8.00 (d, J = 8 Hz, 1H), 7.37 (s, 1H), 7.32 (d, J = 8 Hz, 1H), 7.09 (Br s,
1H), 7.03 (t, J = 8 Hz, 1H), 7.02 (br s, 1H), 6.90 (br s, 1H), 3.93 (t, J
= 5 Hz, 4H), 3.81 (s, 2H), 3.77 (t, J = 5 Hz, 4H), 2.47−2.56 (m, 4H),
2.39 (s, 3H), 1.03 (s, 6H). LCMS RT 5.17 min; m/z 534.7 [M + H]+.
Analytical LC-MS on a Waters Micromass (ZQ2000), using a 3 mm ×
100 mm Higgins Clipeus 5 μm C18 analytical column and H2O/
MeCN modified with 0.1% formic acid running a linear gradient from
5% MeCN to 95% MeCN monitoring by DAD UV (200−600 nm)
detection and ESI+ TIC MS showed 99% purity.
MeCN monitoring by DAD UV (200−600 nm) detection and ESI+
TIC MS showed 95% purity.
2-Methyl-2-(4-((4-morpholino-2-(1H-pyrrolo[2,3-b]pyridin-
4-yl)thieno[3,2-d]pyrimidin-6-yl)methyl)piperazin-1-yl)-
propanamide (14). The title compound was prepared from 38 and
4-bromopyrrolo[2,3-b]pyridine-1-carboxylic acid tert-butylester in a
manner similar to the Stille procedure described for compound 12 to
1
give compound 14 as a white powder (0.049 g, 20%). H NMR (400
MHz, CDCl3) δ 8.93 (s, 1 H), 8.43 (d, J = 5.2 Hz, 1 H), 8.08 (d, J =
5.2 Hz, 1 H), 7.43 (d, J = 2.2 Hz, 2 H), 7.38 (s, 1 H), 7.09 (d, J = 5.2
Hz, 1 H), 5.24 (d, J = 5.2 Hz, 1 H), 4.10 (t, J = 4.7 Hz, 4 H), 3.92 (t, J
= 4.7 Hz, 4 H), 3.86 (s, 2 H), 2.65−2.58 (m, 4 H), 1.60−1.55 (m, 4
H), 1.24 (s, 6 H). LCMS: RT 5.07 min, m/z [M + H]+ 521.3.
Analytical LC-MS on a Waters Micromass (ZQ2000), using a 3 mm ×
100 mm Higgins Clipeus 5 μm C18 analytical column and H2O/
MeCN modified with 0.1% formic acid running a linear gradient from
5% MeCN to 95% MeCN monitoring by DAD UV (200−600 nm)
detection and ESI+ TIC MS showed 99% purity.
2 -M et hyl -2 -(4- ((2- (1- met hyl -1 H -i ndol -4- yl )-4-
morpholinothieno[3,2-d]pyrimidin-6-yl)methyl)piperazin-1-
yl)propanamide (11). The title compound was prepared from 32 in
a manner similar to the Suzuki procedure described for compound 9 to
1-((2-(5-Fluoro-1H-indol-4-yl)-4-morpholinothieno[3,2-d]-
pyrimidin-6-yl)methyl)-N,N-dimethylpiperidin-4-amine (15).
The title compound was prepared from 35 and 1-(tert-butyldime-
thylsilyl)-5-fluoro-4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1H-
indole (see Supporting Information for procedure) in a manner similar
to the Suzuki procedure described for compound 9 to give compound
1
give the title compound 11 (113 mg, 62%). H NMR (400 MHz,
DMSO-d6) 8.16 (d, J = 8, 1H), 7.57 (d, J = 8, 1H), 7.42 (m, 3H), 7.28
(t, J = 8, 1 H), 7.06 (br d, 1H), 6.93 (br d, 1H), 3.99 (m, 4H), 3.86
(m, 9H), 2.56 (br m, 8H), 1.08 (s, 6H). RT: 3.26 min. MS (ESI): m/z
(M + H)+ 534.2. Analytical LC-MS on an Agilent (1200/G6140) using
a 1.8 μM, 2.1 mm × 30 mm ZORBAX SD-C18 analytical column, 40
°C, 400 μL/min, and H2O/MeCN modified with 0.05% trifluoroacetic
acid running a linear gradient from 3% MeCN to 95% MeCN in 10
min, monitoring by UV wavelength 254 nm and ESI+ TIC MS showed
100% purity.
1
15 (0.753 g, 26%). H NMR (400 MHz, DMSO-d6) δ 11.22 (s, 1H),
7.47−7.40 (m, 2H), 7.36 (s, 1H), 7.03−6.83 (m, 1H), 6.69−6.65 (m,
1H), 3.93−3.90 (m, 4H), 3.84 (s, 2H), 3.81−3.74 (m, 4H), 2.99−2.91
(m, 2H), 2.17 (s, 6H), 2.07 (t, J = 10.7 Hz, 3H), 1.73 (d, J = 11.6 Hz,
2H), 1.49−1.35 (m, 2H). RT: 6.61 min. MS (ESI): m/z (M + H)+
495.2. Analytical LC-MS on an Agilent (1200/G6140) using a 1.8 μm,
2.1 mm × 30 mm ZORBAX SD-C18 analytical column, 40 °C, 700
μL/min, and H2O/MeCN modified with 0.05% trifluoroacetic acid
running a linear gradient from 2% MeCN to 98% MeCN in 30 min,
monitoring by UV wavelength 254 nm and ESI+ TIC MS showed
100% purity.
2-Methyl-2-(4-((4-morpholino-2-(1H-pyrrolo[3,2-c]pyridin-4-
yl)thieno[3,2-d]pyrimidin-6-yl)methyl)piperazin-1-yl)-
propanamide (12). A mixture of 2-[4-(4-morpholin-4-yl-2-
(tributylstannanyl)thieno[3,2-d]pyrimidin-6-ylmethyl)piperazin-1-yl]-
isobutyramide 38 (150 mg, 0.22 mmol), 1-benzenesulfonyl-4-bromo-
1H-pyrrolo[3,2-c]pyridine (90 mg, 0.27 mmol), Pd(PPh3)4 (25 mg, 10
mol %), and CuI (50 mg, 0.25 mmol) in dioxane (2.5 mL) was purged
with argon gas and then heated at 140 °C, for 20 min, in a microwave
reactor. The reaction mixture was loaded onto an Isolute SCX-2
cartridge, washed with MeOH/DCM, and then eluted with 2 M NH3
in MeOH. The resulting residue was dissolved in IMS/dioxane (1:1
mL) and 12.5 M aqueous NaOH solution (0.1 mL) added. After
stirring for 1 h, the reaction mixture was loaded onto an Isolute SCX-2
cartridge, washed with MeOH/DCM, and then eluted with 2 M NH3
in MeOH/DCM. The resulting residue was purified by flash
chromatography (Si-PPC, MeOH:DCM, gradient 0:100 to 05:95) to
give compound 12 (28 mg, 24%) as a pale-yellow solid. 1H NMR (400
MHz, CDCl3,) δ 9.11 (s, 1 H), 8.54 (d, J = 5.6 Hz, 1 H), 7.45 (s, 1 H),
7.41 (dd, J = 3.3, 0.9 Hz, 1 H), 7.38 (dd, J = 5.6, 0.95 Hz, 1 H), 7.36
(d, J = 3.3 Hz, 1 H), 7.10 (d, J = 5.3 Hz, 1 H), 5.25 (d, J = 5.2 Hz, 1
H), 4.10 (t, J = 4.75 Hz, 4 H), 3.91 (t, J = 4.7 Hz, 4 H), 3.85 (s, 2 H),
2.60 (s, 8 H), 1.24 (s, 6 H). LCMS RT 4.86 min; m/z [M + H]+ 521.3.
Analytical LC-MS on a Waters Micromass (ZQ2000), using a 3 mm ×
100 mm Higgins Clipeus 5 μm C18 analytical column and H2O/
MeCN modified with 0.1% formic acid running a linear gradient from
5% MeCN to 95% MeCN monitoring by DAD UV (200−600 nm)
detection and ESI+ TIC MS showed 99% purity.
4-(6-((4-(Dimethylamino)piperidin-1-yl)methyl)-4-
morpholinothieno[3,2-d]pyrimidin-2-yl)indolin-2-one (16). The
title compound was prepared from 39 in a manner similar to the Stille
procedure described for compound 12 to give compound 16 as a
1
yellow powder (32 mg, 14%). H NMR (400 MHz, DMSO-d6) δ
10.47 (s, 1H), 7.96 (d, J = 8.1 Hz, 1H), 7.38 (s, 1H), 7.31 (t, J = 7.8
Hz, 1H), 6.92 (d, J = 7.6 Hz, 1H), 3.88−3.99 (m, 6H), 3.76−3.86 (m,
6H), 2.93 (d, J = 11.6 Hz, 2H), 2.16 (s, 6H), 2.06 (t, J = 10.9 Hz, 3H),
1.73 (d, J = 11.6 Hz, 2H), 1.33−1.51 (m, 2H). MS (ESI): m/z (M +
H)+ 493.
N,N-Dimethyl-1-((4-morpholino-2-m-tolylthieno[3,2-d]-
pyrimidin-6-yl)methyl)piperidin-4-amine (17). The title com-
pound was prepared from 35 in a manner similar to the Suzuki
procedure described for compound 9 to give compound 17 as an off-
white powder (250 mg, 55%). 1H NMR (DMSO-d6) δ 8.14−8.25 (m,
2H), 7.32−7.41 (m, 2H), 7.24−7.32 (m, 1H), 3.92−4.03 (m, 4H),
3.72−3.86 (m, 6H), 2.92 (d, J = 11.7 Hz, 2H), 2.40 (s, 3H), 2.16 (s,
6H), 2.05 (t, J = 11.1 Hz, 3H), 1.72 (d, J = 11.3 Hz, 2H), 1.40 (qd, J =
11.8, 3.8 Hz, 2H). MS = 452 [M + H]+.
1-((2-(Imidazo[1,2-a]pyridin-5-yl)-4-morpholinothieno[3,2-
d]pyrimidin-6-yl)methyl)-N,N-dimethylpiperidin-4-amine (18).
The title compound was prepared from 39 in a manner similar to the
Stille procedure described for compound 12 to give compound 18 as a
2-Methyl-2-(4-((4-morpholino-2-(1H-pyrrolo[2,3-c]pyridin-4-
yl)thieno[3,2-d]pyrimidin-6-yl)methyl)piperazin-1-yl)-
propanamide (13). The title compound was prepared from 38 and
tert-butyl 4-bromo-1H-pyrrolo[2,3-c]pyridine-1-carboxylate in a man-
ner similar to the Stille procedure described for compound 12 to give
1
white solid (72 mg, 63%). H NMR (400 MHz, CDCl3) δ 9.23 (s, 1
H), 7.95 (dd, J = 8.6, 1.3 Hz, 1 H), 7.77 (d, J = 8.6 Hz, 1 H), 7.75 (d, J
= 1.3 Hz, 1 H), 7.35 (s, 1 H), 7.32−7.29 (m, 1 H), 4.07 (t, J = 4.8 Hz,
4 H), 3.91 (t, J = 4.8 Hz, 4 H), 3.84 (s, 2 H), 3.04 (d, J = 11.25 Hz, 2
H), 2.33−2.25 (m, 6 H), 2.21−2.10 (m, 3 H), 1.84 (d, J = 12.5 Hz, 2
H), 1.61 (m, 2 H). LCMS: RT 3.89 min; m/z [M + H]+ 478.
Analytical LC-MS on a Waters Micromass (ZQ2000), using a 3 mm ×
100 mm Higgins Clipeus 5 μm C18 analytical column and H2O/
MeCN modified with 0.1% formic acid running a linear gradient from
5% MeCN to 95% MeCN monitoring by DAD UV (200−600 nm)
detection and ESI+ TIC MS showed 95% purity.
1
compound 13 as a pale-yellow solid (242 mg, 39%). H NMR (400
MHz, DMSO-d6): δ 11.75 (s, 1 H), 9.13 (s, 1 H), 8.81 (s, 1 H), 7.71
(t, J = 2.7 Hz, 1 H), 7.45 (s, 1 H), 7.40−7.37 (m, 1 H), 7.07 (d, J = 3.5
Hz, 1 H), 6.95 (d, J = 3.5 Hz, 1 H), 4.01 (t, J = 4.6 Hz, 4 H), 3.87 (s,
2H), 3.83 (t, J = 4.6 Hz, 4 H), 2.58 (s, 4 H), 2.49 (s, 4 H), 1.10 (s, 6
H). LCMS: RT 4.51 min, m/z [M + H]+ 521.2. Analytical LC-MS on a
Waters Micromass (ZQ2000), using a 3 mm × 100 mm Higgins
Clipeus 5 μm C18 analytical column and H2O/MeCN modified with
0.1% formic acid running a linear gradient from 5% MeCN to 95%
5896
dx.doi.org/10.1021/jm3003747 | J. Med. Chem. 2012, 55, 5887−5900

Journal of Medicinal Chemistry
Article
N,N-Dimethyl-1-((2-(3-methylisoquinolin-4-yl)-4-
morpholinothieno[3,2-d]pyrimidin-6-yl)methyl)piperidin-4-
amine (19). The title compound was prepared from 39 in a manner
similar to the Suzuki procedure described for compound 9 to give
1-((2-(1H-Benzo[d]imidazol-1-yl)-4-morpholinothieno[3,2-
d]pyrimidin-6-yl)methyl)-N,N-dimethylpiperidin-4-amine (25).
A mixture of 1-((2-chloro-4-morpholinothieno[3,2-d]pyrimidin-6-yl)-
methyl)-N,N-dimethylpiperidin-4-amine (35) (50 mg, 0.13 mmol),
benzimidazole (32 mg, 0.27 mmol), and concentrated HCl (53 μL,
0.64 mmol) in 1,4-dioxane (1 mL) was subjected to microwave
irradiation at 150 °C for 40 min. The reaction mixture was cooled to
ambient temperature and loaded onto an Isolute SCX-2 cartridge (2
g). The cartridge was then washed with MeOH, and the desired
product was eluted using 2 M NH3 in MeOH. The eluent was
collected and concentrated in vacuo. The resultant residue was purified
by flash chromatography (Si-PPC, DCM/2 M NH3 in MeOH 100:0 to
98:2 to 95:5 to 90:10) to afford compound 25 as a white solid (38 mg,
63%). 1H NMR (400 MHz, CDCl3) δ 9.04 (s, 1 H), 8.61−8.57 (m, 1
H), 7.86−7.82 (m, 1 H), 7.43−7.31 (m, 2 H), 7.27 (s, 1H), 4.07 (t, J =
4.8 Hz, 4 H), 3.92 (t, J = 4.8 Hz, 4 H), 3.82 (s, 2 H), 3.08−2.99 (m, 2
H), 2.30 (s, 6 H), 2.20−2.10 (m, 3 H), 1.88−1.78 (m, 2 H), 1.66−1.54
(m, 2 H). LCMS: RT = 5.12 min, m/z [M + H]+ 478.2. Analytical LC-
MS on a Waters Micromass (ZQ2000), using a 3 mm × 100 mm
Higgins Clipeus 5 μm C18 analytical column and H2O/MeCN
modified with 0.1% formic acid running a linear gradient from 5%
MeCN to 95% MeCN monitoring by DAD UV (200−600 nm)
detection and ESI+ TIC MS showed 99% purity.
1
compound 19 (0.0383 g, 14%) as a light-brown foam. H NMR (300
MHz, DMSO-d6) δ 9.29 (s, 1H), 8.09−8.17 (m, 1H), 7.56−7.72 (m,
2H), 7.48−7.54 (m, 1H), 7.37 (s, 1H), 3.86−3.97 (m, 4H), 3.85 (s,
2H), 3.70−3.78 (m, 4H), 2.96 (d, J = 11.68 Hz, 2H), 2.46 (s, 3H),
2.18 (s, 6H), 2.08 (t, J = 10.74 Hz, 3H), 1.92 (s, 3H), 1.75 (d, J =
12.06 Hz, 2H), 1.34−1.52 (m, 2H). MS (ESI): m/z (M + H)+ 503.0.
N,N-Dimethyl-1-((2-(2-methylbenzo[b]thiophen-3-yl)-4-
morpholinothieno[3,2-d]pyrimidin-6-yl)methyl)piperidin-4-
amine (20). The title compound was prepared from 39 in a manner
similar to the Suzuki procedure described for compound 9 to give
1
compound 20 (0.0037 g, 26%) as an off-white foam. H NMR (300
MHz, CDCl3) δ 8.28 (dd, J = 1.13, 7.18 Hz, 1H), 7.72−7.81 (m, 1H),
7.23−7.39 (m, 3H), 3.98−4.17 (m, 4H), 3.85−3.92 (m, 4H), 3.83 (d, J
= 0.76 Hz, 2H), 3.05 (d, J = 11.71 Hz, 2H), 2.82 (s, 3H), 2.30 (s, 6H),
2.06−2.20 (m, 3H), 1.83 (d, J = 11.71 Hz, 2H), 1.50−1.72 (m, 2H).
MS (ESI): m/z (M + H)+ 508.0.
1-((2-(1H-Indazol-3-yl)-4-morpholinothieno[3,2-d]pyrimidin-
6-yl)methyl)-N,N-dimethylpiperidin-4-amine (21). The title
compound was prepared from 39 and 1-boc-3-bromo-1H-indazole in
a manner similar to the Suzuki procedure described for compound 9 to
N,N-Dimethyl-1-((2-(2-methyl-1H-benzo[d]imidazol-1-yl)-4-
morpholinothieno[3,2-d]pyrimidin-6-yl)methyl)piperidin-4-
amine (26). A mixture of 1-((2-chloro-4-morpholinothieno[3,2-
d]pyrimidin-6-yl)methyl)-N,N-dimethylpiperidin-4-amine (35) (500
mg, 1.26 mmol), 1,2-diaminobenzene (550 mg, 5.09 mmol), and
concentrated HCl (1 mL, 12 mmol) in 1,4-dioxane (10 mL) was
heated at 90 °C for 3 days. The reaction mixture was cooled to
ambient temperature and loaded onto an Isolute SCX-2 cartridge (25
g). The cartridge was then washed with MeOH, and the desired
product was eluted using 2 M NH3 in MeOH. The eluent was
collected and concentrated in vacuo. The resultant residue was purified
by flash chromatography (Si-PPC, DCM/2 M NH3 in MeOH 100:0 to
98:2 to 95:5 to 90:10 to 80:20) to afford the title compound as a white
solid (220 mg, 37%). LCMS: RT = 1.90 min, m/z [M + H]+ 468.4.
A mixture of N-[6-(4-dimethylaminopiperidin-1-ylmethyl)-4-mor-
pholin-4-yl-thieno[3,2-d]pyrimidin-2-yl]benzene-1,2-diamine (80 mg,
0.17 mmol) and triethyl orthoformate (1 mL, 6.0 mmol) in DMF (1
mL) was heated at 150 °C for 6 h. The reaction mixture was cooled to
ambient temperature and loaded onto an Isolute SCX-2 cartridge (10
g). The cartridge was then washed with MeOH, and the desired
product was subsequently eluted using 2 M NH3 in MeOH. The
eluent was collected and concentrated in vacuo. The resultant residue
was purified by flash chromatography (Si-PPC, DCM/2 M NH3 in
MeOH 100:0 to 98:2 to 95:5) followed by reverse phase HPLC
(Phenomenex Luna C18, 20 mM Et3N in water on a gradient of 20
mM Et3N in acetonitrile 95:5 to 2:98) to afford compound 26 as a
1
give compound 21 (0.012 g, 9%) as a white solid. H NMR (300
MHz, DMSO-d6) δ ppm 13.37 (br s, 1H), 8.54 (d, J = 8.3 Hz, 1 H),
7.61 (d, J = 8.3 Hz, 1 H), 7.42 (s, 1 H), 7.40 (t, J = 8.3 Hz, 1 H), 7.24
(t, J = 8.3 Hz, 1 H), 3.92−4.13 (m, 4 H), 3.64−3.92 (m, 6 H), 2.96 (d,
J = 11.3 Hz, 2 H), 2.08 (t, J = 10.76 Hz, 2 H), 1.78 (d, J = 11.33 Hz, 1
H), 1.45 (q, J = 11.2 Hz, 2 H). MS (ESI): m/z (M + H)+ 477.
1-((2-(1H-Indazol-1-yl)-4-morpholinothieno[3,2-d]pyrimidin-
6-yl)methyl)-N,N-dimethylpiperidin-4-amine (22). To a 0 °C
solution of indazole (118 mg, 1.0 mmol) in 8 mL of DMF was added
sodium hydride (60 mg, 1.5 mmol, 60% dispersion in oil), stirred at 0
°C for 15 min, then 1-((2-chloro-4-morpholinothieno[3,2-d]-
pyrimidin-6-yl)methyl)-N,N-dimethylpiperidin-4-amine (35) (395
mg, 1.0 mmol) was added, and the reaction mixture heated to 150
°C for 1 h. The cooled reaction mixture was diluted with water, the
resulting suspension was filtered, and the collected solids were purified
by flash chromatography (from 100% DCM to 20:80 MeOH w/2%
ammonium hydroxide/dichloromethane) and the solids were tritu-
rated with ethyl acetate, solids filtered and dried in a 50 °C vacuum
oven overnight to give compound 22 as a white powder (233 mg,
1
49%). H NMR (400 MHz, DMSO-d6) δ: 8.68 (dd, J = 8.7, 0.8 Hz,
1H), 8.39 (d, J = 0.8 Hz, 1H), 7.83−7.93 (m, 1H), 7.56 (ddd, J = 8.5,
7.0, 1.1 Hz, 1H), 7.39 (s, 1H), 7.26−7.37 (m, 1H), 3.95−4.09 (m,
4H), 3.77−3.91 (m, 6H), 2.94 (d, J = 11.7 Hz, 2H), 2.17 (s, 6H), 2.07
(t, J = 10.7 Hz, 3H), 1.73 (d, J = 10.9 Hz, 2H), 1.41 (qd, J = 11.9, 4.0
Hz, 2H). MS (ESI): m/z (M + H)+ 478.
1
1-((2-(1H-Indol-3-yl)-4-morpholinothieno[3,2-d]pyrimidin-6-
yl)methyl)-N,N-dimethylpiperidin-4-amine (23). The title com-
pound was prepared from 35 in a manner similar to the Suzuki
procedure described for compound 9 to give compound 23 (0.161 g,
white solid (20 mg, 24%). H NMR (400 MHz, CDCl3) δ 8.09−8.04
(m, 1 H), 7.74−7.68 (m, 1 H), 7.29−7.24 (m, 3 H), 4.08−4.02 (m, 4
H), 3.91−3.86 (m, 4 H), 3.83 (s, 2 H), 3.08−3.00 (m, 2 H), 2.93 (s, 3
H), 2.30 (s, 6 H), 2.21−2.10 (m, 3 H), 1.88−1.79 (m, 2 H), 1.68−1.53
(m, 2 H). LCMS: RT = 4.24 min, m/z [M + H]+ 492.3 Analytical LC-
MS on a Waters Micromass (ZQ2000), using a 3 mm × 100 mm
Higgins Clipeus 5 μm C18 analytical column and H2O/MeCN
modified with 0.1% formic acid running a linear gradient from 5%
MeCN to 95% MeCN monitoring by DAD UV (200−600 nm)
detection and ESI+ TIC MS showed 99% purity.
1
27%) as a cream powder. H NMR (400 MHz, CDCl3) δ 8.75 (br s,
1H), 8.63−8.70 (m, 1H), 8.13 (d, J = 2.53 Hz, 1H), 7.37−7.44 (m,
1H), 7.26 (s, 3H), 4.05 (d, J = 5.56 Hz, 4H), 3.92 (d, J = 5.05 Hz, 4H),
3.80 (s, 2H), 2.97−3.08 (m, 2H), 2.29 (s, 6H), 2.05−2.20 (m, 3H),
1.77−1.86 (m, 2H), 1.53−1.66 (m, 2H). Elemental analysis: Theory:
C, 65.5%; H, 6.77%; N, 17.6%. Found: C, 65.2%; H, 6.68%; N, 17.4%.
MS (ESI): m/z (M + H)+ 477.
1-((2-(1H-Indol-1-yl)-4-morpholinothieno[3,2-d]pyrimidin-6-
yl)methyl)-N,N-dimethylpiperidin-4-amine (24). The title com-
pound was prepared from 35 in a manner similar to the procedure
described for compound 22 to give compound 24 as an off-white
powder (157 mg, 33%). 1H NMR (400 MHz, DMSO-d6) δ 8.78 (d, J
= 7.9 Hz, 1H), 8.30 (d, J = 3.8 Hz, 1H), 7.62 (d, J = 7.5 Hz, 1H), 7.36
(s, 1H), 7.25−7.34 (m, 1H), 7.12−7.23 (m, 1H), 6.70 (d, J = 3.8 Hz,
1H), 3.92−4.04 (m, 4H), 3.75−3.87 (m, 6H), 2.93 (d, J = 11.7 Hz,
2H), 2.16 (s, 6H), 2.06 (t, J = 11.1 Hz, 3H), 1.72 (d, J = 11.7 Hz, 2H),
1.40 (qd, J = 11.8, 3.8 Hz, 2H). MS (ESI): m/z (M + H)+ 477.
2-Methyl-2-(4-((2-(2-methyl-1H-benzo[d]imidazol-1-yl)-4-
morpholinothieno[3,2-d]pyrimidin-6-yl)methyl)piperazin-1-
yl)propanamide (27). To a microwave vial was added 150 mg of 2-
(4-((2-chloro-4-morpholinothieno[3,2-d]pyrimidin-6-yl)methyl)-
piperazin-1-yl)-2-methylpropanamide (32), 7.7 mg palladium acetate,
17 mg bis(tri-t-butylphosphine)palladium, 68 mg 2-methylbenzimida-
zole, and 66 mg sodium tert-butoxide. The vial was capped and purged
with nitrogen before addition of 2.6 mL of toluene. The reaction was
heated for 105 min at 155 °C. The resulting cooled mixture was
loaded onto an SCX column, and the column was washed with
methanol. The product was eluted with 2 M NH3 in methanol, and
5897
dx.doi.org/10.1021/jm3003747 | J. Med. Chem. 2012, 55, 5887−5900

Journal of Medicinal Chemistry
Article
the product fractions were concentrated in vacuo. The residue
using a fluorescence polarization assay that monitors formation of the
product 3,4,5-inositoltriphosphate molecule as it competes with
fluorescently labeled PIP3 for binding to the GRP-1 pleckstrin
homology domain protein. An increase in phosphatidyl inositide-3-
phosphate product results in a decrease in fluorescence polarization
signal as the labeled fluorophore is displaced from the GRP-1 protein
binding site. Class I PI3K isoforms were purchased from Millipore or
were expressed and purified as heterodimeric recombinant proteins.
Tetramethylrhodamine-labeled PIP3 (TAMRA-PIP3), di-C8-PIP2,
and PIP3 detection reagents were purchased from Echelon
Biosciences. PI3K isoforms were assayed under initial rate conditions
in the presence of 10 mM Tris (pH 7.5), 25 μM ATP, 9.75 μM PIP2,
5% glycerol, 4 mM MgCl2, 50 mM NaCl, 0.05% (v/v) Chaps, 1 mM
dithiothreitol, and 2% (v/v) DMSO at the following concentrations
for each isoform: PI3Kα,β at 60 ng/mL; PI3Kγ at 8 ng/mL; PI3Kδ at
45 ng/mL. After assay for 30 min at 25 °C, reactions were terminated
with a final concentration of 9 mM EDTA, 4.5 nM TAMRA-PIP3, and
4.2 μg/mL GRP-1 detector protein before reading fluorescence
polarization on an Envision plate reader. IC50 or Ki apparent values
were calculated by fitting the dose−response curves to a four-
parameter equation or Morrison tight-binding equation for com-
petitive inhibitors, respectively. All data points are an average of a
minimum of three runs.
The human B-cell line, Ri-1, was plated in 384-well tissue culture
plates in RPMI1640. Test compound was added to replicate wells with
a final DMSO concentration not to exceed 0.5%. After 30 min, cells
were treated with 25 μg/mL F(ab′)2 Frag Goat antihuman IgM Fc5u
(Jackson 109−006−129) for 1 h at 37 °C with 5% CO2. Following all
treatments, cells were lysed in lysis buffer supplemented with
phosphatase and protease inhibitors under nondenaturing conditions
at 4 °C. Quantitative levels of pAkt Ser473 protein lysates were
measured using Meso Scale Discovery (MSD cat. no. K1211CAD)
electrochemiluminescence technology. Briefly, MSD Multi-Spot pAkt
Ser473 384-well plates were blocked with 35 mL of 5% BSA blocking
buffer for 60 min at room temperature (RT) and washed four times in
Tris wash buffer. The plates were incubated with 35 mL of pAkt
Ser473/protein lysate on the plate shaker overnight at 4 °C. Plates
were then washed with Tris wash buffer four times. The wells were
then incubated with 10 mL of detection antibody per well for 2 h on
the shaker at RT. Plates were then washed again with Tris wash buffer
four times after, and 35 mL of read buffer was added and the plates
were read using a MSD Sector Imager 6000. All data points are an
average of a minimum of three runs.
obtained was purified by reverse phase HPLC to afford 29.3 mg of 27
1
as a white solid (100% pure). H NMR (400 MHz, DMSO) δ 8.12−
7.95 (m, 1H), 7.60 (dt, J = 6.6, 2.8 Hz, 1H), 7.44 (s, 1H), 7.32−7.17
(m, 2H), 7.06 (d, J = 3.1 Hz, 1H), 6.94 (d, J = 2.9 Hz, 1H), 4.04−3.93
(m, 4H), 3.88 (s, 2H), 3.84−3.75 (m, 4H), 2.83 (s, 3H), 2.56 (brm,
4H), 2.47 (brm, 4H), 1.08 (s, 6H). LCMS m/z [M + H]+ 535.2.
4-(2-(2-Ethyl-1H-benzo[d]imidazol-1-yl)-6-((4-
(methylsulfonyl)piperazin-1-yl)methyl)thieno[3,2-d]pyrimidin-
4-yl)morpholine (28). To a microwave vial was added 1 g of 4-(2-
chloro-6-((4-(methylsulfonyl)piperazin-1-yl)methyl)thieno[3,2-d]-
pyrimidin-4-yl)morpholine (33), 95 mg of tris(dibenzylideneacetone)-
dipalladium(0), 95 mg 2-dicyclohexylphosphino-2′,4′,6′ triisopropylbi-
phenyl, 0.36 g 2-ethylbenzimidazole, and 1.51 g cesium carbonate.
N,N-Dimethylformamide (16.1 mL) was added, and the reaction was
heated for 20 min at 145 °C. The reaction was purified by reverse
phase HPLC to afford 283.8 mg of 28 as a white solid (99.5% pure).
1H NMR (400 MHz, DMSO) δ 8.04−7.95 (m, 1H), 7.68−7.59 (m,
1H), 7.47 (s, 1H), 7.29−7.20 (m, 2H), 4.03−3.91 (m, 5H), 3.86−3.76
(m, 4H), 3.26 (q, J = 14.9, 7.5 Hz, 3H), 3.17 (br m, 4H), 2.91 (s, 3H),
2.62 (br m, 4H), 1.32 (t, J = 7.4 Hz, 3H). LCMS m/z [M + H]+ 542.
2-(1-((2-(2-Ethyl-1H-benzo[d]imidazol-1-yl)-4-
morpholinothieno[3,2-d]pyrimidin-6-yl)methyl)piperidin-4-
yl)propan-2-ol (29). A microwave vial was charged with 2-(1-((2-
chloro-4-morpholinothieno[3,2-d]pyrimidin-6-yl)methyl)piperidin-4-
yl)propan-2-ol (36) (1.5 g, 3.65 mmol) and 2-ethyl-1H-benzo[d]-
imidazole (0.60 g, 4.1 mmol), XPhos (0.17 g, 0.36 mmol),
tris(dibenzylideneacetone)dipalladium (0.17 g, 0.19 mmol), cesium
carbonate (2.4 g, 7.4 mmol), and 1,4-dioxane (15 mL). The reaction
was sealed and heated in a Biotage microwave at 140 °C for 40 min.
The reaction was then filtered through Celite and concentrated. The
crude product was purified by flash chromatography (0−10% MeOH
gradient in EtOAc) followed by reverse phase HPLC to give the title
compound 29 (0.795 g, 42%). 1H NMR (400 MHz, DMSO-d6) δ 8.00
(m, 1H), 7.65 (m, 1H), 7.42 (s, 1H), 7.25 (m, 2H), 4.07 (s, 1H), 3.97
(m, 4H), 3.83 (s, 2H), 3.81 (m, 4H), 3.27 (m, 2H), 2.99 (d, J = 10.9,
2H), 1.99 (t, J = 11.3, 2H), 1.67 (d, J = 11.3, 2H), 1.32 (t, J = 7.43,
3H), 1.11−1.29 (m, 3H), 1.04 (s, 6H). RT: 7.57 min. MS (ESI): m/z
(M + H)+ 521.3. Analytical LC-MS on an Agilent (1200/G6140) using
a 1.8 μM, 2.1 mm × 30 mm ZORBAX SD-C18 analytical column, 40
°C, 700 μL/min and H2O/MeCN modified with 0.05% trifluoroacetic
acid running a linear gradient from 2% MeCN to 98% MeCN in 30
min, monitoring by UV wavelength 254 nm and ESI+ TIC MS showed
99% purity.
4-(6-((2,2-Dimethyl-4-(oxetan-3-yl)piperazin-1-yl)methyl)-2-
(2-ethyl-1H-benzo[d]imidazol-1-yl)thieno[3,2-d]pyrimidin-4-
yl)morpholine (30). A mixture of 2-chloro-6-(2,2-dimethyl-4-oxetan-
3-ylpiperazin-1-ylmethyl)-4-morpholin-4-yl-thieno[3,2-d]pyrimidine
(37) (136 mg, 0.31 mmol), 2-ethylbenzimidazole (48 mg, 0.33 mmol),
tris(dibenzylideneacetone)dipalladium (8 mg, 3 mol %), XPhos (14
mg, 9 mol %), and Cs2CO3 (143 mg, 0.44 mmol) in DMF (4 mL) was
purged with argon then heated at 150 °C for 30 min in a microwave
reactor. The reaction mixture was loaded onto an Isolute SCX-2
cartridge which was washed with MeOH and the product eluted with 2
M NH3/MeOH. The resulting residue was purified by column
chromatography (Si-PCC, MeOH/DCM, 0−5%) then (Si-PCC, 2 M
NH3/MeOH/DCM, 0−2%) affording the 30 as a cream solid (45 mg,
26%). 1H NMR (400 MHz, CDCl3) δ 7.97−7.96 (1 H, m), 7.75−7.74
(1 H, m), 7.32−7.22 (3 H, m), 4.66 (2 H, t, J = 6.47 Hz), 4.59 (2 H, t,
J = 6.08 Hz), 4.04 (4 H, t, J = 4.72 Hz), 3.91−3.86 (6 H, m), 3.46−
3.45 (1 H, m), 3.33 (2 H, q, J = 7.48 Hz), 2.64 (2 H, t, J = 4.98 Hz),
2.33 (2 H, s), 2.14 (2 H, s), 1.42 (3 H, t, J = 7.47 Hz), 1.20 (6 H, s).
LCMS: RT 2.78 min; m/z [M + H]+ 548.3. Analytical LC-MS on a
Waters Micromass (ZQ2000), using a 100 mm × 2.1 mm Acquity
BEH C18 1.7 μM analytical column and H2O/MeOH modified with
0.1% formic acid running a linear gradient from 5% to 95% MeOH
monitoring by PDA UV detection and ESI+ TIC MS showed 94%
purity.
Computational Methods. The sequences of the kinase domains
of the four PI3K isoforms from multiple species were aligned using
PSI-BLAST (NCBI). A human PI3K.α H1047R mutant structure
(PDB code 3HIZ) was chosen as the template because, among all
structures available then, it shared with human PI3Kδ the highest level
of sequence homology both overall and around the active site.
Sequence alignment was manually refined before a model of human
PI3Kδ was built using PRIME (Schrodinger) with standard
̈
parameters. Subsequently, a mouse PI3Kδ structure (PDB 2WXP)
was published. The structure supported the previous homology model
and was adopted for the docking calculation. Structure of the target
protein was processed using Maestro (Schrodinger). Ligands were
̈
prepared using LigPrep (Schrodinger) and docked against the target
̈
using GLIDE (Schrodinger) in SP or XP modes. Torsional angle
̈
analysis was carried out as described. Substructures representing the
torsional angles in question were minimized at molecular mechanics
(MacroModel, Schrodinger) and then at quantum mechanics level
̈
(Jaguar, Schrodinger). The resulting conformer was used to perform a
̈
“relaxed coordinate scan” of the dihedral in 15° increment using
density functional theory (B3LYP) with 6-31G** basis set (Jaguar,
Characterization of Biochemical and Cellular Activity. In
vitro enzymatic activity of the class I PI3K isoforms was measured
Schrodinger).
̈
5898
dx.doi.org/10.1021/jm3003747 | J. Med. Chem. 2012, 55, 5887−5900

Journal of Medicinal Chemistry
Article
Nonredundant Role for the Phosphoinositide 3-Kinase p110δ in
Signaling by the B-Cell Receptor Complex. Mol. Cell. Biol. 2002, 22,
8580−8591. (c) Okkenhaug, K.; Bilancio, A.; Farjot, G.; Priddle, H.;
Sancho, S.; Peskett, E.; Pearce, W.; Meek, S. E.; Salpekar, A.;
Waterfield, M. D.; Smith, A. J. H.; Vanhaesebroeck, B. Impaired B and
T Cell Antigen Receptor Signaling in p110δ PI 3-Kinase Mutant Mice.
Science 2002, 297, 1031−1034.
ASSOCIATED CONTENT
* Supporting Information
Characterization of synthetic intermediates and compounds
32−39. This material is available free of charge via the Internet
at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
*Phone: (650) 467-7450. Fax: (650) 467-8922. E-mail:
bsafina@gene.com.
(8) Edwards, C. W. J. Cambridge, G. B-cell targeting in rheumatoid
arthritis and other autoimmune disease. Nature Rev. Immunol. 2006, 6,
394−403.
■
(9) Monach, P. A.; Benoist, C.; Mathis, D. The Role of Antibodies in
Mouse Models of Rheumatoid Arthritis, and Relevance to Human
Disease. Adv. Immunol. 2004, 82, 217−248.
Notes
The authors declare no competing financial interest.
(10) Vanhaesebroeck, B.; Ali, K.; Bilancio, A.; Geering, B.; Foukas, L.
C. Signalling by PI3K isoforms: insights from gene-targeted mice.
Trends Biochem. Sci. 2005, 30, 194−204.
ACKNOWLEDGMENTS
■
(11) Ruckle, T.; Schwarz, M. K.; Rommel, C. PI3Kγ inhibition:
̈
The authors wish to thank Mengling Wong, Martin Struble,
Chris Hamman and Michael Hayes for compound purification.
We thank Baiwei Lin, Yutao Jiang and Deven Wang for
towards an ‘aspirin of the 21st century’? Nature Rev. Drug Discovery
2006, 5, 903−918.
(12) (a) Ameriks, M. K.; Venable, J. D. Small Molecule Inhibitors of
Phosphoinositide 3-Kinase (PI3K) δ and γ. Curr. Top. Med. Chem.
2009, 9, 738−753. (b) Herman, S. E. M.; Gordon, A. L.; Wagner, A. J.;
Heerema, N. A.; Zhao, W.; Flynn, J. M.; Jones, J.; Andritsos, L.; Puri,
K. D.; Lannutti, B. J.; Giese, N. G.; Zhang, X.; Wei, L.; Byrd, J. C.;
Johnson, A. J. The phospatidylinositol 3-kinase-δ inhibitor CAL-101
demonstrates promising pre-clinical activity in chronic lymphoctic
leukemia by antagonizing intrinsic and extrinsic cellular survival
signals. Blood 2010, 116, 2078−2087. (c) Williams, O.; Houseman, B.
T.; Kunkel, E. J.; Aizenstein, B.; Hoffman, R.; Knight, Z. A.; Shokat, K.
M. Discovery of Dual Inhibitors of the Immune Cell PI3Ks p110δ and
p110γ: A Prototype for New Anti-inflammatory Drugs. Chem. Biol.
2010, 17, 123−134.
1
determination of purity by HPLC, mass spectroscopy, and H
NMR.Å, angstrom(s); °C, degrees Celsius; AAG, alpha 1-acid
glycoprotein; Ac, acetyl; Acac, acetylacetonate; ADME,
absorption, distribution, metabolism and excretion; AKT,
protein kinase B; ATP, adenosine 5′-triphosphate; AUC, area
under the curve; ASP, aspartic acid; Bn, benzyl; BOC, boc, tert-
butoxycarbonyl; cLogP, calculated logP; compd, compound;
CYP, cytochrome P; DDI, drug−drug interaction; DCE,
dichloroethane; DCM, dichloromethane; DMF, dimethylfor-
mamide; ESI, electrospray ionization; EtOAc, ethyl acetate;
HBA, hydrogen bond acceptor; HBD, hydrogen bond donor;
HCl, hydrochloric acid; HPLC, high-performance liquid
chromatography; high-pressure liquid chromatography; HSA,
human serum albumin; iv, intravenous; LC-MS, liquid
chromatography−mass spectrometry; LYS, lysine; MeCN,
acetonitrile; MetID, metabolite identification; MeOH, meth-
anol; mpk, mg per kg; MS, mass spectrometry; NADPH,
reduced nicotinamide adenine dinucleotide phosphate; NME,
new molecular entity; NMR, nuclear magnetic resonance;
pAKT, phosphorylated AKT; PI3K, phosphoinositide 3-kinase;
PIP3, phosphatidylinositol (3,4,5)-triphosphate; po, oral
adminstration; RA, rheumatoid arthritis; SAR, structure−
activity relationships; TDI, time-dependent inhibition; THF,
tetrahydorfuran; THR, threonine; TRP, tryptophan; TYR,
tyrosine; UV, ultraviolet; VAL, valine; Phos, 2-dicyclohex-
ylphosphino-2′,4′,6′-triisopropylbiphenyl
(13) Berndt, A.; Miller, S.; Williams, O.; Le, D. D.; Houseman, B. T.;
Pacold, J. I.; Gorrec, F.; Hon, W.; Liu, Y.; Rommel, C.; Gaillard, P.;
Ruckle, T.; Schwarz, M. K.; Shokat, K. M.; Shaw, J. P.; Williams, R. L.
̈
The p110δ structure: mechanisms for selectivity and potency of new
PI(3)K inhibitors. Nature Chem. Biol. 2010, 6, 117−124.
(14) Sutherlin, D. P.; Sampath, D.; Berry, M.; Castanedo, G.; Chang,
Z.; Chuckowree, I.; Dotson, J.; Folkes, A.; Friedman, L.; Goldsmith,
R.; Heffron, T.; Lee, L.; Lesnick, J.; Lewis, C.; Mathieu, S.; Nonomiya,
J.; Olivero, A.; Pang, J.; Prior, W. W.; Salphati, L.; Sideris, S.; Tian, Q.;
Tsui, V.; Wan, N. C.; Wang, S.; Wiesmann, C.; Wong, S.; Zhu, B.-Y.
Discovery of (Thienopyrimidine-2-yl)aminopyrimidines as Potent,
Selective, and Orally Available Pan-PI3-Kinase and Dual Pan-PI3-
Kianse/mTOR Inhibitors for the Treatment of Cancer. J. Med. Chem.
2010, 53, 1086−1097.
(15) Bode, C. The nasty surprise of a complex drug−drug
interaction. Drug Discovery Today 2010, 15, 391−395.
(16) (a) Venkatakrishnan, K.; Obach, R. S. Drug−Drug Interactions
via Mechanism-Based Cytochrome P450 Inactivation: Points to
Consider for Risk Assessment from in Vitro Data and Clinical
Pharmacologic Evaluation. Curr. Drug Metab. 2007, 8, 449−462.
(b) Riley, R. J.; Grime, K.; Weaver, R. Time-Dependent CYP
inhibition. Expert Opin. Drug Metab. Toxicol. 2007, 3, 51−66.
(17) (a) Walsh, J. S.; Miwa, G. T. Bioactivation of Drugs: Risk and
Drug Design. Annu. Rev. Pharmacol. Toxicol. 2011, 51, 145−167.
(b) Kalgutkar, A. S.; Obach, R. S.; Maurer, T. S. Mechanism-Based
Inactivation of Cytrochrome P450 Enzymes: Chemical Mechansims,
Structure−Activity Relationships and Relationship to Clinical Drug−
Drug Interactions and Idiosyncratic Adverse Drug Reactions. Curr.
Drug Metab. 2007, 8, 407−447.
(18) (a) Obach, R. S.; Walsky, R. L.; Venkatakrishnan, K.
Mechanism-Based Inactivation of Human Cytochrome P450 Enzymes
and the Prediction of Drug−Drug Interactions. Drug Metab. Dispos.
2006, 35, 246−255. (b) Zimmerlin, A.; Trunzer, M.; Faller, B. CYP3A
Time-Dependent Inhibition Risk Assessment Validated with 400
Reference Drugs. Drug Metab. Dispos. 2011, 39, 1039−1046.
(c) Krippendorff, B.-F.; Neuhaus, R.; Lienau, P.; Reichel, A.;
Huisinga, W. Mechanism-Based Inhibition: Deriving KI and kinact
REFERENCES
■
(1) Isaacs, J. D. The changing face of rheumatoid arthritis: sustained
remission for all? Nature Rev. Immunol. 2010, 10, 605−611.
(2) Majithia, V.; Geraci, S. A. Rheumatoid Arthritis: Diagnosis and
Management. Am. J. Med. 2007, 120, 936−939.
(3) Van Vollenhoven, R. F. Treatment of rheumatoid arthritis: state
of the art 2009. Nature Rev. Rheum. 2009, 5, 531−541.
(4) Allison, M. PML problems loom for Rituxan. Nature Biotechnol.
2010, 28, 105−106.
(5) Cantley, L. C. The phosphoinositide 3-kinase pathway. Science
2002, 296, 1655−1657.
(6) Rommel, C.; Camps, M.; Ji, H. PI3Kδ and PI3Kγ: partners in
crime in inflammation in rheumatoid arthritis and beyond? Nature Rev.
Immunol. 2007, 7, 191−201.
(7) (a) Clayton, E.; Bardi, G.; Bell, S. E.; Chantry, D.; Downes, C. P.;
Gray, A.; Humphries, L. A.; Rawlings, D.; Reynolds, H.; Vigorito, E.;
Turner, M. A Crucial Role for the p110δ Subunit of Phosphatidyli-
nositol 3-Kinase in B Cell Development and Actication. J. Exp. Med.
2002, 196, 753−763. (b) Jou, S.; Carpino, N.; Takahashi, Y.; Piekorz,
R.; Chao, J.; Carpino, N.; Wang, D.; Ihle, J. N. Essential,
5899
dx.doi.org/10.1021/jm3003747 | J. Med. Chem. 2012, 55, 5887−5900

Journal of Medicinal Chemistry
Article
Directly form Time-Dependent IC50 Values. J. Biomol. Screening 2009,
14, 913−923. (d) BurtH. J.GaletinA.HoustonJ. B.IC50-based
approaches as an alternative method for assessment of time-dependent
inhibition of CYP3A4 Xenobiotica 2010, 40, 331−343. (e) Mukadam,
S.; Tay, S.; Tran, D.; Wang, L.; Delarosa, E. M.; Khojasteh, S. C.;
Halladay, J. S.; Kenny, J. R. Evaluation of Time-Dependent
Cytochrome P450 Inhibition in a High-Troughput, Automated
Assay: Introducing a Novel Area Under the Curve Shift Approach.
Drug Metab. Lett. 2012, 6, 43−53.
(19) (a) Raynaud, F. I.; Eccles, S. A.; Patel, S.; Alix, S.; Box, G.;
Chuckowree, I.; Folkes, A.; Gowan, S.; De haven Brandon, A.; Di
Stefano, F.; Hayes, A.; Henley, A. T.; Lensun, L.; Pergl-Wilson, G.;
Robson, A.; Saghir, N.; Zhyvoloup, A.; McDonald, E.; Sheldrake, P.;
Shuttleworth, S.; Valenti, M.; Wan, N. C.; Clarke, P. A.; Workman, P.
Biological properties of potent inhibitors of class I phosphatidylinosi-
tide 3-kinase: from PI-103 through PI-540, PI-620 to the oral agent
GDC-0941. Mol. Cancer Ther. 2009, 8, 1725−1738. (b) Folkes, A. J.;
Ahmadi, K.; Alderton, W. K.; Alix, S.; Baker, S. J.; Box, G.;
Chuckowree, I. S.; Clarke, P. A.; Depledge, P.; Eccles, S. A.;
Friedman, L. S.; Hayes, A.; Hancox, T. C.; Kugendradas, A.; Lensun,
L.; Moore, P.; Olivero, A. G.; Pang, J.; Patel, S.; Pergl-Wilson, G. H.;
Raynaud, F. I.; Robson, A.; Saghir, N.; Salphati, L.; Sohal, S.; Ultsch,
M. H.; Valenti, M.; Wallweber, H. J. A.; Wan, N. C.; Wiesmann, C.;
Workman, P.; Zhyvoloup, P.; Zvelebil, M. J.; Shuttleworth, S. J. The
identification of 2-(1H-indazol-4-yl)-6-(4-methanesulfonyl-piperazin-
1-ylmethyl)-4-morpholin-4-yl-thieno-[3,2-d]pyrimidine (GDC-0941)
as a potent, selective, orally bioavailable inhibitor of class I PI3 kinase
for the treatment of cancer. J. Med. Chem. 2008, 51, 5522−5532.
(20) Sweeney, Z. K.; Murray, J.; Chan, B. K.; Balazs, M.; Bradley, E.;
Castanedo, G.; Chabot, C.; Chantry, D.; Flagella, M.; Goldstein, D.
M.; Kondru, R.; Lesnick, J.; Li, J.; Lucas, M. C.; Nonomiya, J.; Pang, J.;
Price, S.; Salphati, L.; Safina, B.; Savy, P. P. A.; Seward, E. M.; and
Sutherlin, D. P. unpublished results.
(21) Jamei, M.; Marciniak, S.; Feng, K.; Barnett, A.; Tucker, G.;
Rostami-Hodjegan, A. The Simcyp Population-Based ADME Simu-
lator. Exp. Opin. Drug Metab. Toxicol. 2009, 5, 211−223.
(22) (a) Chen, A.; Dube,
Gaudreault, M.; Grimm, E.; Houle, R.; Lacombe, P.; Laliberte,
S.; MacDonald, D.; Mackay, B.; Martin, D.; McKay, D.; Powell, D.;
Lev
́
D.; Dube,
́
L.; Gagne,
́
S.; Gallant, M.;
́
S.; Liu,
́
esque, J.-F. Bioorg. Med. Chem. Lett. 2010, 20, 5074−5079.
(b) Walsh, J. S.; Miwa, G. T. Annu. Rev. Pharmacol. Toxicol. 2011, 51,
145−167.
(23) (a) Dalvie, D. K.; Kalgutkar, A. S.; K-Bakht, S. C.; Obach, R. S.;
O’Donnell, J. P. Biotransformation Reactions of Five-Membered
Aromatic Heterocyclic Rings. Chem. Res. Toxicol. 2002, 15, 269−299.
(b) Argikar, U. A.; Mangold, J. B.; Harriman, S. P. Strategies and
Chemical Design Approaches to Reduce the Potential for Formation
of Reactive Metabolic Species. Curr. Top. Med. Chem. 2011, 11, 419−
449.
(24) (a) Guengerich, F. P. Mechanism of Cytochrome P450
Substrate Oxidation: MiniReview. J. Biochem. Mol. Toxicol. 2007, 21,
163−168. (b) Ortiz de Montellano, P. R. Chem. Rev. 2010, 110, 932−
948.
(25) Milletti, F.; Storchi, L.; Sforna, G.; Cruciani, G. New and
original pKa prediction method using grid molecular interaction fields.
J. Chem. Inf. Model. 2007, 47, 2172−2181.
(26) Benchware 3D Explorer (v2.6, Tripos).
5900
dx.doi.org/10.1021/jm3003747 | J. Med. Chem. 2012, 55, 5887−5900