Exploring the PI3Kα and γ binding sites with 2,6-disubstituted isonicotinic derivatives

Article abstract of DOI:10.1016/j.bmcl.2009.02.115

A homology model of the p110α catalytic subunit of PI3Kα was generated from the p110γ crystal structure. Using this model, an isonicotinic scaffold was designed for chemically exploring the PI3Kα and γ binding sites. A focused library of derivatives was s

Full text of DOI:10.1016/j.bmcl.2009.02.115

Bioorganic & Medicinal Chemistry Letters 19 (2009) 2215–2219
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Exploring the PI3K
a
and
c
binding sites with 2,6-disubstituted
isonicotinic derivatives
Philip T. Cherian a, Leonid N. Koikov a, Matthew D. Wortman b, James J. Knittel a,
*
a The James L. Winkle College of Pharmacy, University of Cincinnati, Cincinnati, OH 45267, USA
b Department of Cell and Cancer Biology, College of Medicine, University of Cincinnati, Cincinnati, OH 45267, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A homology model of the p110
ture. Using this model, an isonicotinic scaffold was designed for chemically exploring the PI3K
binding sites. A focused library of derivatives was synthesized and tested. The morpholine acids 5a
and 5b proved to be the most potent analogs.
a
catalytic subunit of PI3K
a
was generated from the p110
c
crystal struc-
and
Received 26 January 2009
Revised 24 February 2009
Accepted 25 February 2009
Available online 4 March 2009
a
c
Ó 2009 Elsevier Ltd. All rights reserved.
Phosphatidylinositol 3-OH kinases (PI3Ks) are dual specific lipid
and protein kinases that influence multiple cellular processes
including cell growth, proliferation, survival and motility by gener-
ation of 3-phosphoinositides.1 PI3Ks are arranged into three classes
I, II and III depending on their structure, regulation and substrate
(Fig. 1A). In addition, the COOH in our scaffold extends one carbon
further into the phosphate binding area of ATP than the carbonyls
of LY294002 or TGX-126. Thus, derivatization of 2-chloro-6-methyl
isonicotinic acid provides easy access to three potential libraries
(Fig. 1B). In this study 6-substitution was limited to Ar = 4-F–2-
MePh and X = O, NH based on literature data.9
specificity and class I is divided into subclasses IA (
a, b and d)
and IB (
c
) based on their mode of activation.2 Aberrant activity
To evaluate the binding of these isonicotinic acid derivatives we
of the class I enzymes is observed in various pathological states
and inhibition of PI3Ks provides opportunities for treatment of
inflammation, immune diseases and cardiovascular disorders.3
built (Modeller 6v2)13 and refined (YASARA dynamics—AMBER99
force field)14 a homology model15 of p110
a
based on its sequence
–inhibitor complex
and the high resolution structure of p110
c
The PIK3CA gene encoding the p110
a
catalytic subunit is fre-
is identified as an
(1e7u) and used it for docking with CAChe16 (flexible ligand—flex-
ible protein). The docking procedure was validated by modeling
quently mutated in many cancers and PI3K
a
important chemotherapeutic target.4 Although several classes of
compounds that inhibit the class I PI3Ks have been reported, devel-
LY294002 in the p110
structure (1e7v) with RMSD 0.095. Docking of LY294002 and isoni-
cotinic derivative 5b into our p110 model revealed an orientation
similar to p110 depicting H-bonds with Lys802 (833, p110 ) and
Val851 (882, p110 ); Figure 1C–E.
c binding site which matched the X-ray
opment of inhibitors selective for PI3K
The goal of this work was to build a computer model of PI3K
based on the X-ray structure of p110 and develop a chemical scaf-
fold for easy generation of chemical libraries in order to explore
differences between the PI3K and active sites.
The morpholinylchromone LY294002 is widely used as a non-
selective PI3K inhibitor. Its X-ray structure with p110 shows H-
a
is still a major challenge.5,6
a
a
c
c
c
c
Although the pyridine ring provided a means for sequential
introduction of nucleophiles into 2- and 6-positions (benzylic bro-
mide vs 2-Cl pyridine reactivity), it also created problems for syn-
a
c
c
thesis of key intermediates 2 and 4 (Fig. 2). The issue of
bonds of the chromone C@O with Lys833, the morpholine oxygen
with Val882 and phenyl at the entrance of the binding pocket fac-
ing the solvent.7 The rigid structure of LY294002 (two rotatable
bonds) permits defining its binding mode but does not provide en-
ough conformational freedom for exploration of subtle differences
polybromination during synthesis of monobromide 2 was solved
by selective reduction of the mixture with diethyl phosphite
improving its yields from 40% to 75%.17 Conversion of 2 into 3 pro-
ceeded in high yields under standard conditions (85–90%). Nucle-
ophilic displacement of Cl in 3 proved to be sensitive to the
nature of the amine and a significant number of amines required
excess amounts (>5 equiv) and microwave heating. Under these
conditions, the amines reacted with the COOEt faster resulting in
formation of complex mixtures of mono and bis-substituted
derivatives (e.g., 12 and 13) that were directly transesterified to es-
ters (4a,b) or hydrolyzed to the respective carboxylic acids (5a,b,
g–i). Mixed anhydrides provided selective N-acylation of
N,O-nucleophiles without need for protection of the COOH (6 and
7). The Suzuki reaction was used for synthesis of 2-aryl derivatives
in the active sites of the
literature on PI3K inhibitors led us to conclude that the six-mem-
bered ring of
-pyrone acts as a pharmacological spacer.8–12
a and c
isoforms of PI3K.5 Analysis of the
c
Accordingly, we chose the 2,6-disubstituted isonicotinic acid as a
scaffold that would allow more coverage of conformational space
than LY294002 and less than its structural analog TGX1269
* Corresponding author. Tel.: +1 513 558 0733.
E-mail address: james.knittel@uc.edu (J.J. Knittel).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.02.115

2216
P. T. Cherian et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2215–2219
Figure 1. (A) 2,6-Disubstituted Isonicotnic acid scaffold. (B) Opportunities for modification. (C) The p110a docked pose of isonicotinic derivative 5b (cyan) shows a small
anti-clockwise shift for the H-bonding groups and the Ph group extending further than LY294002 (green). Docking of LY294002 (D) and 5b (E) in the p110a homology model
viewed from the same angle.
Figure 2. Synthetic scheme. Reagents and conditions: (a) NBS, AIBN, CCl4, reflux; (b) Diethyl phosphite, DIEA, THF, 0–4 °C; (c) 4-F–2-MePhOH [or] 4-F–2-MePhNH2, K2CO3,
DMF, rt; (d) NHR3R4, MW, 200–250 °C then H2SO4/EtOH (4a–b) or (g) (5a–b, g–i); (e) HNR5R6, 1,4-dioxane, 80 °C then (g) (5c–f); (f) Ar-B(OH)2, Pd(dppf), 2 M Na2CO3, 1,4-
dioxane, 80 °C then (g) (5j–n); (g) KOH, EtOH/H2O; (h) ethylchloroformate, NMM, Et2O, H2N–R, rt; (i) NH3/MeOH, 60 °C; (j) LiAlH4/Et2O; (k) Dess–Martin, DCM, rt.
(5j–n). All tested compounds were >99% pure except for 10 which
was 90% pure.18
served decrease in potency (similar to that for LY294002)8 for pro-
tonated 5c and neutral non-basic 5d was most likely caused by loss
of H-bonding with the backbone NH. While conversion of the
hydrophilic H-bond acceptor oxygen in morpholine (5b) into
hydrophilic H-bond donor +NH2 in 5c did not affect selectivity,
the replacement of +NH2 with the isosteric non-polar CH2 (5d) con-
As seen from Table 1 and Figure 3 the parent carboxylic acids 5a
and 5b inhibited PI3K
a
and
c in the low micromolar range with
sixfold selectivity for PI3Ka. Compound 5a containing the H-bond
acceptor phenyl ether oxygen is equipotent to 5b with the H-bond
donor/acceptor aniline nitrogen (Table 1). Conversion of COOH to
non-ionizable derivatives 6, 7, 8, 9 and 10 with H-bonding capabil-
ity allowed further exploration of the ATP-phosphate binding re-
gion. All compounds from this series proved to be less potent
than the parent acids 5a,b. The least drop in potency (fourfold)
was observed for alcohol 9 suggesting that the acids 5a,b interact
with the active site by H-bonding and not by salt bridge formation.
The esters 4a,b showed less inhibition than the corresponding car-
boxylic acids 5a,b probably due to their limited solubility (Fig. 3).
The active sites within 4 Å of ATP or LY294002 ligands for all
four isoforms of PI3Ks have very similar sequence with a difference
siderably reduced potency at PI3Ka but not at PI3Kc (Fig. 3). Com-
pounds 5d,e showed similar potency and selectivity (Fig. 3), that is,
the 4-CH2 in piperidine is not essential. At the same time, addition
of 3-OH (5f) to the pyrrolidine 5e which enables polar interactions
does not affect the activity at PI3Kc but increases activity at PI3Ka
(Fig. 3). This biological profile agrees with the above difference in
properties of the Ser/Ala pairs and potential involvement of the
hydrophobic Ile881 in p110c versus corresponding Val850 in
p110 located in the morpholine binding pocket.
a
Compounds with ‘opened’ morpholine ring (monoethanolam-
ines 5g,h and bis-derivative 12) had potency similar to 5c–f while
the much more polar and flexible diethanolamine 5i was practi-
cally inactive (Fig. 3). The use of rigid flat phenol rings allowed
for further exploration of the H-bonding sites in the morpholine
binding pocket. Replacement of morpholine in 5b by 4-OHPh and
3-OHPh (5l,m) led to only ca. twofold drop in potency compared
to 5b, that is, the presence of the morpholine ring is not vital
(Fig. 3) and it can be replaced by an aromatic ring with a polar sub-
stituent in 3 or 4 positions. Pyridine mimics properties of both phe-
of 7 amino acids between the
a
and
c
isoforms and 5 or 3 between
and show the
b and d relative to a 10
.
The active sites of PI3K
a
c
largest difference not only in sequence but also in polarity and
H-bonding properties of amino acids next to the H-bonded oxygen
of morpholine—non-polar Ala885 in p110c corresponds to the po-
lar H-bonding Ser854 in p110 . The same difference exists be-
a
tween Ala805 and Ser774. Since it was logical to assume that
addition or removal of H-bonding fragments to the morpholine
template might impact interaction with either PI3K
a
or
c
isoforms,
nol (p-bonding aromatic ring with polar region) and morpholine
we synthesized a set of flexible (5g–i), semi-rigid (5c–f) and rigid
(5j–n) analogs of morpholine in 5b. Replacement of the morpho-
line oxygen with NH (5c) or CH2 (5d), its complete removal (pyrrol-
idine derivative 5e) or removal with addition of OH group (3-OH
pyrrolidine 5f) led to a significant drop in potency (Table 1 and
Fig. 3) indicating lack of direct interaction with Ser854. The ob-
(H-bond acceptor weakly basic nitrogen) and permits additional
examination (5j,k) of the polar portion of the active site that inter-
acts indiscriminately with the 3- or 4-OH in 5l,m. In contrast to
these equipotent 3- and 4-OH derivatives 5l,m, the 4-pyridine 5j
is several times more potent than 3-pyridine 5k (Fig. 3), that is,
the presence of nitrogen in pyridine in the same position as oxygen

P. T. Cherian et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2215–2219
2217
Table 1
Properties and biological activity of the synthesized compounds
No.
R1
R2
X
Yield (%)
Mp (°C)
IC50 (lM)
PI3K
a
PI3Kc
6
O
O
O
O
O
50
24
30
82
53a
179–181
209–211
235–237
117–119
168–171
P50
P10
P50
7
ꢀ10
c
8
>10
c
c
4a
5a
7 ( 1)
23 ( 5)
c
c
4b
9
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
78
78–80
38
139–141
Decomp.
20 ( 5)
ꢀ 100
5 ( 1)
P100
ꢀ100
34 ( 3)
>100
10
5b
5c
5d
5e
5f
30
44a
87a
53a
43a
48a
35a
33a
20a
68a
56a
55a
86–88
b
>100
c
153–155
>100
c
93–95
>100
b
>100
>100
ꢀ100
>100
29 ( 7)
>100
ꢀ100
>100
b
b
b
5g
5h
5i
>100
>100
>100
5j
186–188
252–254
122–124
42 ( 3)
>100
5k
5l
ꢀ100
(continued on next page)

2218
P. T. Cherian et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2215–2219
Table 1 (continued)
No.
5m
5n
11
R1
R2
X
Yield (%)
47a
6a
Mp (°C)
156–158
227–229
IC50 (lM)
PI3K
a
PI3Kc
NH
NH
NH
NH
NH
ꢀ100
>100
>100
>100
>100
42 ( 6)
>100
>100
>100
>100
Cl
70
191–193
b
12
61
b
13
10
14a
14b
R = H
R = Et
26
21
187–189
53–55
36 ( 8)
51 ( 32)
>100
>100
15
LY294002
0.2 ( 0.03)
1.2 ( 0.3)
IC50 values are the mean of three experiments SEM.19
a
Isolated yield over two steps from 3.
TFA salts were very hygroscopic.
Less than 10% inhibition.
b
c
Figure 3. Percent inhibition of PI3K
50 M and 7, 8, 4a and 4b at 10 M due to limited solubility.
a and c lM; 6 at
by synthesized compounds. Average of three experiments + SD (4a,b is average of two).19 All compounds tested at 100
l
l
in morpholine is favorable. Binding of pyridine (as a part of the
condensed imidazo-quinoline, NVP-BEZ235) in morpholine pocket
with p110c20 Although this paper describes only the PI3K
a and c
isoforms and a limited number of 6-substituents, this scaffold
and methodology can be easily applied for mapping the active sites
of other PI3Ks.
of p110
in 5b with 4-pyridine (5j) does not affect interaction with the
PI3K but results in a sixfold drop in potency on PI3K (Table 1).
a
has been reported recently.15c Substitution of morpholine
c
a
Thus, in accordance with our modeling the above set of 2,6-
disubstituted isonicotinic derivatives with limited conformational
flexibility demonstrate the importance of morpholine in the 2-po-
sition and a polar substituent at the 4-position of the pyridine
spacer with the morpholino acids 5a and b being the most potent
Acknowledgment
This work was funded in part by The James L. Winkle College of
Pharmacy Dean’s Pilot Projects Grant.
(Table 1). PI3Ka and c are not sensitive to the nature of –CH2X– in
the 6-position and morpholine can be replaced with a polar aro-
matic ring with some loss of potency and selectivity (5j, l and
m). The suggested structural modifications led to some discrimina-
References and notes
1. Vanhaesebroeck, B.; Leevers, S. J.; Khatereh, A.; Timms, J.; Katso, R.; Driscoll, P.
C.; Woscholski, R.; Parker, P. J.; Waterfield, M. D. Annu. Rev. Biochem. 2001, 70,
535.
2. Fruman, D. A.; Meyers, R. A.; Cantley, L. C. Ann. Rev. Biochem. 1998, 67, 481.
3. Ward, S. G.; Sotsios, Y.; Dowden, J.; Bruce, I.; Finan, P. Chem. Biol. 2003, 10,
tion between PI3Ka and c but less than what was expected based
on the difference in their polarity/H-bonding, most likely due to
the inherent flexibility of their active sites (exemplified by the lat-
est X-ray data on two structurally different ligands complexed
207.

P. T. Cherian et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2215–2219
2219
4. Velculescu, V. E.; Vogelstein, B.; Silliman, N.; Ptak, J.; Szabo, S.; Yan, H.; Gazdar,
A.; Powell, S. M.; Riggins, G. J.; Willson, J.; Markowitz, S.; Kinzler, K. W.; Victor,
E. Science 2004, 304, 554.
5. Knight, Z. A.; Gonzalez, B.; Feldman, M. E.; Zunder, E. R.; Goldenberg, D. D.;
Williams, O.; Loewith, R.; Stokoe, D.; Balla, A.; Toth, B.; Balla, T.; Weiss, W. A.;
Williams, R. L.; Shokat, K. M. Cell 2006, 125, 733.
15. (a) The model proved to be in good agreement with the first X-ray structure of
PI3K
[2rd0, Ref. (b)] published later in December 2007; RMSD 0.942 for 5 Å
a
binding site.; (b) Huang, C.; Mandelker, D.; Schmidt-Kittler, O.; Samuels, Y.;
Velculescu, V. E.; Kinzler, K. W.; Vogelstein, B.; Gabelli, S. B.; Amzel, L. M.
Science 2007, 318, 1744; During review of the manuscript one of the referees
indicated that a model of p110a kinase domain has been previously reported
6. Marone, R.; Cmiljanovic, V.; Giese, B.; Wymann, M. P. Biochim. Biophys. Acta
2008, 1784, 159.
7. Walker, E. H.; Pacold, M. E.; Perisic, O.; Ried, R.; Stephens, L.; Williams, R. L. Mol.
Cell. 2000, 6, 909.
in: (c) Maira, S.-M.; Stauffer, F.; Brueggen, J.; Furet, P.; Schnell, C.; Fritsch, C.;
Brachmann, S.; Chene, P.; Pover, A. D.; Schoemaker, K.; Fabbro, D.; Gabriel, D.;
Simonen, M.; Murphy, L.; Finan, P.; Sellers, W.; Garcıa-Echeverrıa, C. Mol.
Cancer Ther. 2008, 7, 1851.
8. Vlahos, C. J.; Matter, W. F.; Hui, K. Y.; Brown, R. F. J. Biol. Chem. 1994, 269, 5241.
9. Roberson, A.; Jackson, S.; Kenche, V.; Yaip, C.; Parbaharan, H.; Thompson, P. WO
01/53266 A1, Thrombogenix, 2001.
10. Knight, Z. A.; Chiang, G. G.; Alaimo, P. J.; Kenski, D. M.; Ho, C. B.; Coan, K.;
Abraham, R. T.; Shokat, K. M. Bioorg. Med. Chem. 2004, 12, 4749.
11. Knight, Z. A.; Shokat, K. M. Biochem. Soc. Trans. 2007, 35, 245.
12. Maronea, R.; Cmiljanovicb, V.; Giese, B.; Wymann, M. P. Biochim. Biophys. Acta
2008, 1784, 159.
16. www.cache.fujitsu.com/biomedcache.
17. Liu, P.; Chen, Y.; Deng, J.; Tua, Y. Synthesis 2001, 14, 2078.
18. The synthesized compounds were purified by column chromatography
or preparative HPLC and/or recrystallized from following solvents: 6,
Hex/EtOAc (1:1); 7/8, acetone; 9, Hex/EtOAc (3:1); 5n, triturated with ether;
rest of compounds, MeOH/H2O. Structures were confirmed by NMR and
HRMS.
19. The percent inhibition and IC50 curves were determined using the Promega
TM
13. Eswar, N.; John, B.; Mirkovic, N.; Fiser, A.; Ilyin, V. A.; Pieper, U.; Stuart, A. C.;
Marti-Renom, M. A.; Madhusudan, M. S.; Yerkovich, B.; Sali, A. Nucleic Acids Res.
2003, 31, 3375.
14. Krieger, E.; Darden, T.; Nabuurs, S. B.; Finkelstein, A.; Vriend, G. Proteins 2004,
57, 678.
Luminescent Kinase Glo assay. The IC50 values were calculated using
GraphPad PRISM V4.0—sigmoidal dose–response curve.
20. Apsel, B.; Blair, J. A.; Gonzalez, B.; Nazif, T. M.; Feldman, M. E.; Aizenstein, B.;
Hoffman, R.; Williams, R. L.; Shokat, K. M.; Knight, Z. A. Nat. Chem. Biol. 2008, 4,
691.