Azaviridins as new scaffolds for the development of PI-3K inhibitors

Article abstract of DOI:10.1055/s-0030-1259026

We have designed a new class of non-natural furanosteroids to serve as more chemically stable analogues of the viridins, potent inhibitors of PI-3K. Central to the design is the incorporation of a nitrogen atom into the steroidal ring system to attenuate

Full text of DOI:10.1055/s-0030-1259026

LETTER
2875
Azaviridins as New Scaffolds for the Development of PI-3K Inhibitors
T
owards the De
e
velopment
a
of Azavirid
t
ins her J. Sundstrom, Dennis L. Wright*
Department of Pharmaceutical Sciences, University of Connecticut, Storrs CT, 06269, USA
Fax +1(860)4866857; E-mail: dennis.wright@uconn.edu
Received 5 August 2010
an attenuated level of reactivity towards other biological
nucleophiles. We propose that, in general, nucleophilic
addition to the furanosteroids in the active site should be
more facile due to a combination of proximity effects and
Abstract: We have designed a new class of non-natural furanoster-
oids to serve as more chemically stable analogues of the viridins,
potent inhibitors of PI-3K. Central to the design is the incorporation
of a nitrogen atom into the steroidal ring system to attenuate the ac-
tivity of the electrophilic 2,4-diacylfuran common to these natural activation of the C3 and C7 carbonyl functionality
through hydrogen bonding to Ser806 and Asp964.7 Thus,
products. In this manuscript, we describe our initial synthetic stud-
ies on this ring system and the preparation of key intermediates for
analogue generation.
it may prove possible to attenuate the reactivity of the
electrophilic 2,4-diacylfuran to a point that it still reacts
with the active-site amine but is less prone to nonspecific
reaction towards other biological nucleophiles.
Key words: natural products, heterocycles, furans, amination, pal-
ladium
O
O
AcO
MeO
O
The phosphoinositide-3-kinases (PI-3K) are a family of
extremely important proteins involved in a variety of cell-
signaling pathways that transduce signals through the in-
termediacy of second messengers known as the phos-
phatidylinositides.1 The central role of these kinases in
cancer has made them attractive targets for the develop-
ment of antineoplastic agents.2 These lipid kinases func-
tion by regioselectively transferring a phosphate group
from ATP to the C3-hydroxy of the inositol headgroup.
Naturally occurring furanosteroids such as viridin and
wortmannin (Figure 1, a) are perhaps the most well-
known inhibitors of these enzymes and derive high poten-
cy from addition of a conserved active site lysine to the
C21-position of the 2,4-diacylfuran to give a bound viny-
logous carbamate/amide (Figure 1, b).3 The pharmacolog-
ical profile of the naturally occurring furanosteroids is
complicated; in one regard the electrophilic reactivity of
the furan nucleus produces very high potency through ir-
reversible binding. However, the reactivity of the 2,4-di-
acylfuran ring has also been shown to pose a liability to
the molecules, resulting in inhibitors with low stability
and a high potential to react nonspecifically, leading to
undesirable off-target effects.4
17
17
a)
OH
D
C
C
D
R
O
H
O
B
B
A
A
3
O
3
O
7
7
E
E
O
21
21
O
R = OMe, viridin
wortmannin
AcO
R = H, demethoxyviridin
O
MeO
b)
O
O
H
O
OH
HN
O
O
AcO
c)
MeO
17
17
D
D
R
N
N
A
H
O
A
O
O
O
O
O
1 (R = OMe or H)
2
Figure 1 a) Structure of furanosteroid natural products; b) the ad-
duct of wortmannin and an active site lysine; c) azaviridin analogues
We have been interested in the development of non-natu-
ral furanosteroids with improved properties relative to
wortmannin and viridin.5 Although the reactivity of the
furanosteroids is problematic, other protein-reactive com-
pounds such as the b-lactams have found successful clin-
ical application as they possess a balance of chemical
reactivity that largely preserves the electrophile until it
reaches its intended target.6 An attractive analogue series
would be one that retained the ability to undergo a specific
chemical reaction with the active site lysine but displayed
Towards these goals, we have designed a potential new
class of analogues that we have termed azaviridins
(Figure 1, c) that entails replacement of the C10 quaterna-
ry carbon with a nitrogen atom. This substitution should
function to increase the overall electron density of the 2,4-
diacyl furan and in effect converts the electron-withdraw-
ing C7 ketone into a vinylogous amide. The increased
electron density produced by the nitrogen should serve to
lower the chemical reactivity of this heterocycle. In addi-
tion, the substitution could serve as a simplifying modifi-
cation by replacing a synthetically challenging quaternary
carbon with a heteroatom. This substitution reveals a nov-
el furanoquinolone skeleton embedded in a steroidal
SYNLETT 2010, No. 19, pp 2875–2878
x
x
.x
x
.2
0
1
0
Advanced online publication: 10.11.2010
DOI: 10.1055/s-0030-1259026; Art ID: S04710ST
© Georg Thieme Verlag Stuttgart · New York

2876
T. J. Sundstrom, D. L. Wright
LETTER
framework. In addition to this structural change, these an- 3,4-Dibromofuran 6 could be directly deprotonated at the
alogues also necessitate modification to the peripheral C2-position with LDA followed by addition to o-ni-
functionality on the A-ring, an area that is largely solvent trobenzaldehyde to give the labile nitro alcohol 7. No
exposed in the bound state7 and likely tolerant of slight al- products arising from reaction of the nitro functionality
teration. Our initial investigations have focused on devel- were noted. Oxidation of the secondary alcohol to the ke-
oping methods to construct this complex heterocyclic ring tone 8 could be accomplished in high yield upon treatment
system and in this report, we report the preparation of a with activated manganese dioxide. The next key transfor-
key intermediate that can be used to pursue a new class of mation involved selective reduction of the aromatic nitro
PI-3K inhibitors.
group to the corresponding amine without over-reduction
of the two furyl bromides. After considerable optimiza-
tion, it was found that use of a low valent iron reagent9 led
to a very clean reduction to produce the desired aniline 9.
With the key dibromide in hand, we were ready to study
the ring-closure reaction to form the central 4-pyridone
ring system. It was envisioned that a palladium-catalyzed
cyclization could be highly effective as it was assumed
that the more electron-deficient b-keto bromide would be
the most reactive toward insertion. Pleasingly, it was
found that under optimized conditions [Pd(OAc)2/Xant-
phos]10 an effective cyclization could be realized, leading
to furanoquinolone 10 in very good yield. With a direct
access to 10 secured, we turned our attention to incorpo-
ration of the remaining ring system corresponding to the
A-ring of the natural product by sequential reactions of
the vinylogous amide nitrogen and the remaining furyl
bromide (Scheme 3).
Our retrosynthetic analysis of azaviridin analogues 3 indi-
cated a key A-ring olefin 4 as a late-stage intermediate.
This was an attractive target as Sorensen and co-workers
have previously shown in their synthesis of viridin8 that a
related olefin could be elaborated to the a-hydroxy ketone
found in the A-ring of the natural product (Scheme 1).
O
O
R
N
N
O
O
O
21
O
O
3
4
R = H or OH
O
Initially, the annulation was envisioned to proceed by
alkylation of the pyridone nitrogen followed by ring clo-
sure onto the aromatic bromide. However, all attempts to
produce an effective alkylation were frustrated by com-
petitive N- versus O-alkylation such as in the allylation of
10 to give an almost equal mixture of 11 and 12. Changes
of solvent, base, or other reaction conditions had little im-
pact on the outcome of the alkylation. Reversing the order
of elaboration of the two groups became necessary to de-
velop an effective annulation protocol. Sonogashira cou-
pling of the furyl bromide to protected propynyl alcohol
gave the alkyne 13 that could be converted to the allylic
alcohol 14 through reduction and deprotection. Ring clo-
sure could be effective in moderate yield by initial conver-
sion to the highly labile mesylate followed by reaction of
the crude material with sodium hydride to give tetracycle
15.
Br
Br
H2N
Br
O
O
Br
O
5
6
Scheme 1 Retrosynthetic analysis of azaviridins
The unsaturated A-ring in 4 indicated a variety of poten-
tial annulation strategies yielding a tetracyclic furanoqui-
nolone that we hoped to access from the furyl aniline 5
through a chemoselective palladium insertion reaction.
Readily available 3,4-dibromofuran was seen as a conve-
nient starting material to the furan portion of these hetero-
cycles. In this communication, we report our efforts to
access these types of furan-based heterocycles by access-
ing sequential annulations on dibromide 6 (Scheme 2).
Upon successful completion of this model study, we in-
vestigated the synthesis of a more functionalized system
that would contain key carbonyl functionality on the aro-
matic ring that corresponds to C17 of the natural furano-
steroids. High-resolution crystal structures of wortmannin
bound to PI-3K7 shows that this carbonyl functionality
provides a key interaction by functioning as a hydrogen-
bond acceptor with Val882 of the enzyme, an interaction
we would hope to maintain in the synthesis of these ana-
logues.
The known nitrobenzaldehyde 1611 emerged as a useful
starting material that incorporates an additional car-
boalkoxy handle on the aromatic C-ring. Conversion into
the analine 17 proceeded uneventfully and upon exposure
to low-valent palladium was converted to the pyridone
Br
NO2
Br
NO2
MnO2
95%
LDA then
O2N
6
OH
O
Br
Br
O
O
OHC
85%
7
8
Fe3(CO)12
BnNEt3Cl
Br
NH2
Pd(OAc)2
HN
Xantphos
Cs2CO3
70%
O
Br
1 N NaOH
O
Br
86%
O
O
9
10
Scheme 2 Synthesis of a furanoquinolone intermediate
Synlett 2010, No. 19, 2875–2878 © Thieme Stuttgart · New York

LETTER
Towards the Development of Azaviridins
2877
HN
N
N
base
+
Br
O
O
Br
O
Br
Br
O
O
O
10
11
12
HCCCH2OTBS
Pd(PPh3)2Cl2
Et3N, DMF
75%
HO
HN
N
HN
MsCl, Et3N
1) H2, Lindlar, 95%
2) HF–py, quant.
TBSO
then NaH
40%
O
13
O
O
O
O
O
14
15
Scheme 3 Annulation of ring A
18.12 The increased acidity of the aniline nitrogen due to moiety. Current efforts are directed at the synthesis and
the second carbonyl group had no obvious impact on the biological evaluation of azaviridin analogues.
cyclization. The second annulation proceeded through a
similar route to give the tetracycle 2013 retaining key func-
Acknowledgment
tionality at a position equivalent to C17 of the natural
products (Scheme 4).
We would like to thank the American Cancer Society (RSG-04-
267-01) for generous support of this work.
EtO2C
CO2Et
CO2Et
References and Notes
a–c
Br
Br
NH2
d
HN
(1) Fruman, D. A.; Meyers, R. E.; Cantley, L. C. Ann. Rev.
Biochem. 1998, 67, 481.
OHC
16
O
O
Br
NO2
(2) (a) Sundstrom, T.; Anderson, A.; Wright, D. L. Org. Biomol.
Chem. 2009, 7, 840. (b) Samuels, Y.; Wang, Z.; Bardelli,
A.; Silliman, N.; Ptak, J.; Szabo, S.; Yan, H.; Gazdar, A.;
Powell, S. M.; Riggins, G. J.; Willson, J. K. V.; Markowitz,
S.; Kinzler, K. W.; Vogelstein, B.; Velculescu, V. E. Science
2004, 304, 554. (c) Vivanco, I.; Sawyers, C. L. Nat. Rev.
Cancer 2002, 2, 489. (d) Hu, L.; Hofmann, J.; Jaffe, R. B.
Clin. Cancer Res. 2005, 11, 8208. (e) Cully, M.; You, H.;
Levine, A. J.; Mak, T. W. Nat. Rev. Cancer 2006, 6, 184.
(3) Norman, B. H.; Shih, C.; Toth, J. E.; Ray, J. E.; Dodge, J. A.;
Johnson, D. W.; Rutherford, P. G.; Schultz, R. M.; Worzalla,
J. F.; Vlahos, C. J. J. Med. Chem. 1996, 39, 1106.
(4) Schultz, R. M.; Merriman, R. L.; Andis, S. L.; Bonjouklian,
R.; Grindey, G. B.; Rutherford, P. G.; Gallegos, A.; Massey,
K.; Powis, G. Anticancer Res. 1995, 15, 1135.
O
17
18
O
CO2Et
CO2Et
HO
HN
N
O
e–g
h
O
O
O
19
20
Scheme 4 Reagents and conditions: (a) 2-lithio-3,4-dibromofuran,
THF, –78 °C, 97%; (b) MnO2, CH2Cl2, r.t., 98%; (c) Fe3(CO)12,
BnEt3NCl, PhH, NaOH, 68%; (d) Pd(OAc)2, Xantphos, Cs2CO3,
dioxane, 120 °C, 71%; (e) TBSOCH2CCH, Pd(PPh3)2Cl2, CuI, DMF,
65 °C, 64%; (f) H2, Lindlar catalyst, CH2Cl2, r.t., 85%; (g) HF–pyri-
dine, THF, r.t., 90%; (h) MsCl, Et3N, CH2Cl2; (i) NaH, THF, 40%
over two steps.
(5) Wright, D. L.; Robotham, C. V.; Aboud, K. Tetrahedron
Lett. 2002, 43, 953.
(6) Drahl, C.; Cravatt, B. F.; Sorensen, E. J. Angew. Chem. Int.
Ed. 2005, 44, 5788.
(7) Walker, E. H.; Pacold, M. E.; Perisic, O.; Stephens, L.;
Hawkins, P. T.; Wymann, M. P.; Williams, R. L. Mol. Cell
2000, 6, 909.
(8) Anderson, E. A.; Alexanian, E. J.; Sorensen, E. J. Angew.
Chem. Int. Ed. 2004, 43, 1998.
(9) Des Abbayes, H.; Alper, H. J. Am. Chem. Soc. 1977, 99, 98.
(10) Begouin, A.; Hesse, S.; Queiroz, M. R. P.; Kirsch, G.
Synthesis 2006, 2794.
(11) Mee, S. P. H.; Lee, V.; Baldwin, J. E.; Cowley, A.
Tetrahedron 2004, 60, 3695.
The azaviridins have been designed as analogues of the
naturally occurring viridin furanosteroid with an intrinsi-
cally lower level of chemical reactivity through the inclu-
sion of an electron-donating nitrogen atom to attenuate the
electrophilicity of the annulated 2,4-diacylfuran. In this
manuscript, we have described the synthesis of a key tet-
racylic scaffold for the synthesis and evaluation of this
class of potential new inhibitors. Central to the synthesis
of these aza-steroid analogues are the selective, sequential
cross-coupling reactions of an acylated dibromofuran
(12) Selected Characterization Data for Compound 18
1H NMR (500 MHz, CD3OD): d = 9.10 (s, 1 H), 8.29 (d,
J = 8.8 Hz, 1 H), 8.25 (s, 1 H), 7.84 (d, J = 8.8 Hz, 1 H), 4.43
Synlett 2010, No. 19, 2875–2878 © Thieme Stuttgart · New York

2878
T. J. Sundstrom, D. L. Wright
LETTER
(q, J = 7.1 Hz, 2 H), 1.44 (t, J = 7.1 Hz, 3 H). 13C NMR (125
MHz, CD3OD): d = 167.5, 152.6, 150.4, 133.2, 129.8, 129.7,
125.9, 125.3, 119.9, 119.5, 103.4, 79.6, 62.6, 14.8. IR (KBr):
3087, 2964, 1633, 1598, 1573 cm–1. HRMS–FAB: m/z calcd
for C14H11BrNO6 [M + H]+: 335.9866; found: 335.9866.
1 H), 6.63 (dt, J = 2.2, 10.1 Hz, 1 H), 6.00 (dt, J = 3.3, 10.1
Hz, 1 H), 4.99 (s, 2 H), 4.42 (q, J = 7.1 Hz, 2 H), 1.44 (t,
J = 7.1 Hz, 3 H). 13C NMR (125 MHz, CDCl3): d = 166.2,
164.7, 142.0, 141.8, 137.1,136.5, 132.4,130.2, 127.2, 124.5,
123.1, 115.9, 113.6, 111.9, 61.4, 46.6, 14.7. ESI-HRMS:
m/z calcd for C17H14NO4 [M + H]+: 296.0923; found:
296.0917.
(13) Selected Characterization Data for Compound 20
1H NMR (500 MHz, CDCl3): d = 9.26 (d, J = 1.9 Hz, 1 H),
8.29 (dd, J = 2.0 Hz, 1 H), 7.52 (s, 1 H), 7.35 (d, J = 8.9 Hz,
Synlett 2010, No. 19, 2875–2878 © Thieme Stuttgart · New York