Viridin analogs derived from steroidal building blocks

Article abstract of DOI:10.1016/j.bmcl.2012.09.015

Naturally occurring furanosteroids such as viridin and wortmannin have long been known as potent inhibitors of the lipid kinase PI-3K. We have been interested in directly accessing analogs of these complex natural products from abundant steroid feedstock materials. In this communication, we describe the synthesis of viridin/wortmannin hybrid molecules from readily available building blocks that function as PI-3K inhibitors and maintain their electrophilic properties. The compounds also show anti-proliferative effects against a breast cancer line.

Full text of DOI:10.1016/j.bmcl.2012.09.015

Bioorganic & Medicinal Chemistry Letters 22 (2012) 6919–6922
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Viridin analogs derived from steroidal building blocks
Kishore Viswanathan, Sophia N. Ononye, Harold D. Cooper, M. Kyle Hadden, Amy C. Anderson,
⇑
Dennis L. Wright
Department of Pharmaceutical Sciences, University of Connecticut, Storrs, CT 06269, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Naturally occurring furanosteroids such as viridin and wortmannin have long been known as potent
inhibitors of the lipid kinase PI-3K. We have been interested in directly accessing analogs of these com-
plex natural products from abundant steroid feedstock materials. In this communication, we describe the
synthesis of viridin/wortmannin hybrid molecules from readily available building blocks that function as
PI-3K inhibitors and maintain their electrophilic properties. The compounds also show anti-proliferative
effects against a breast cancer line.
Received 3 August 2012
Accepted 4 September 2012
Available online 13 September 2012
Keywords:
PI-3K
Wortmannin
Ó 2012 Elsevier Ltd. All rights reserved.
Viridin
Steroid
Irreversible enzyme inhibitors
Over the past decade, there has been an intense interest in the
development of potent and specific inhibitors of protein kinases
that phosphorylate serine, threonine or tyrosine residues in their
target proteins for the treatment of cancer and other diseases. In
addition to this large class of protein kinases, there has also been
interest in targeting the complementary lipid kinases.1 Perhaps
the most well-studied of these kinases is the family of phosphoin-
ositide kinases, most notably phosphatidyl-inositol-3-kinase
(PI-3K), which is responsible for the conversion of phosphatidyl-
inositol-4,5-bisphosphate (PIP2) to the corresponding phosphati-
dyl-inositol-3,4,5-triphosphate (PIP3).2
compound was limited by unfavorable pharmacokinetic properties
and hepatotoxicity associated with the protein-reactive inhibitor.
Since these initial studies, a variety of reversible, ATP-competitive
PI-3K inhibitors have been developed and several are currently in
clinical trials against a range of malignancies.6 There has been
renewed interest in the furanosteroid class because of the develop-
ment of a ring-opened prodrug, PX-866, formed by the simple reac-
tion of wortmannin and diallyl amine.7 It has been shown that this
compound reforms wortmannin in vivo but with an improved
pharmacokinetic profile.8 PX-866 is currently in Phase I clinical tri-
als in patients with solid tumors.
Dysregulation of this pathway has been tied to a variety of
cancers.3 There are three PI-3K isoforms; the ability to selectively
target specific isozymes remains a high priority.4 The strong links
between over activation of this signaling pathway and cancer has
prompted the development of a wide range of small molecule
inhibitors with varying degrees of potency and selectivity.4
Two of the earliest known PI-3K inhibitors, wortmannin and
viridin (Fig. 1), are naturally occurring furanosteroids derived from
fungal sources. These unusual steroidal natural products are very
As these compounds are highly complex and synthetically
demanding there have been limited attempts to generate analogs
in order to improve selectivity and other properties.9–12 We have
been attracted to the use of readily available steroids as starting
points for direct and concise syntheses of viridin analogs. Although
typical steroids are clearly related to these natural products, there
are key differences, specifically in rings A and C. Whereas steroids
typically have a carbocyclic A-ring, wortmannin possess a hetero-
cyclic lactone system. Conversely, the viridins maintain the A-ring
carbocycle but possess an aromatic C-ring. We envisioned that the
use of a readily available steroid such as androstenedione 1 would
provide direct access to a class of analogs such as 2 that would
represent a hybrid structure between wortmannin and viridin.
These analogs would incorporate the saturated (non-aromatic)
C-ring found in wortmannin with the carbocyclic A-ring as found
in viridin (Scheme 1).
potent, albeit non-selective, inhibitors of the enzyme (IC50
=
2–10 nM)3 and function as quasi-irreversible enzyme inhibitors,
alkylating a key lysine residue in the enzyme active site through
the electrophilic C20-position. In addition to targeting PI-3K, these
natural products have also been shown to inhibit structurally
related protein kinases such as mTOR, DNA-PK and ATM.5 Despite
the promising anti-proliferative activity of wortmannin, the
Key to developing these steroid building blocks was the instal-
lation of the furan E-ring at the junction of rings A and B, necessi-
tating the introduction of an additional carbon atom on ring A and
vicinal oxygenation on the B-ring (Scheme 2).
⇑
Corresponding author.
E-mail address: dennis.wright@uconn.edu (D.L. Wright).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.09.015

6920
K. Viswanathan et al. / Bioorg. Med. Chem. Lett. 22 (2012) 6919–6922
Figure 1. (a) Structure of naturally occurring furanosteroids and a prodrug form; (b) X-ray structure of wortmannin bound to PI-3K
3K
c, labeled residues are conserved in PI-
a.
O
O
OTBS
OTBS
O
a-c
d-f
H
H
O
O
O
O
O
A
H
H
O
Br
B
E
OTBS
7
8
9
O
O
1
2
OTBS
OH
Scheme 1. Design for androgen-derived furanosteroid analogs.
i-j
g-h
O
HO
OH
O
O
OH
10
O
11
O
Scheme 3. Synthesis of
a truncated furanosteroid analog; (a) NaBH4, EtOH;
a,b
c
(b)TBSCl, Imidazole, CH2Cl2; (c) Br2/AcOH, Ethylene oxide, À35 °C; (d) NBS, benzoyl
peroxide, CCl4, reflux; (e) LiCO3, LiBr, DMF, 90 °C; (f) Bu3Sn-CH2-OTBS, PdCl2(PPh3)2,
dioxane, 90 °C; (g) AcOH:THF:H2O(4:1:1); (h) AD-mix b, 5 mol % OsO4, 5 mol %
(DHAD)2PHAL, MeSO2NH2, K2S2O8, t-BuOH/ H2O(1:1), 0 °C to RT; (i) COCl2, DMSO,
Et3N, CH2Cl2, À78 °C; (j) HF-Pyridine, THF.
H
H
H
H
O
O
Br
O
Br
1
3
O
O
d
H
H
a second bromination with N-bromosuccinimide in the presence
of a radical initiator led to a smooth conversion to the dibromide
3.13 Elimination of the allylic bromide under basic conditions14
produced bromodienone 4 that contains key ring A/B functionality
that allowed direct installation of the diacylfuran moiety. Introduc-
tion of the C20-carbon as a formyl equivalent was possible using a
palladium catalyzed coupling with an oxygenated stannane to give
silylether 5.15 Completion of the diacylfuran unit would require
selective oxidation of the dienone followed by cyclization to the
heterocycle. Deprotection of the silylether was followed by regio-
selective osmylation of the B-ring olefin to produce the triol 6.
Exposure of the triol to an excess of the Swern reagent16 effected
sequential oxidation and heterocycle formation to produce the
key wortmannin analog 2.
H
H
O
OTBS
Br
4
5
O
O
H
H
OH
e,f
H
H
O
g
O
O
OH
O
6
2
OH
Scheme 2. A-ring functionalization; (a) Br2/AcOH, Ethylene oxide, À35 °C, 60%; (b)
NBS, benzoyl peroxide, CCl4, reflux, 82%; (c) LiCO3, LiBr, DMF, 90 °C; (d) Bu3Sn-CH2-
OTBS, PdCl2(PPh3)2, dioxane, 90 °C, 68% (e) HF pyr (f) AD-mix b, 5 mol % OsO4,
5 mol % (DHAD)2PHAL, MeSO2NH2, K2S2O8, t-BuOH/ H2O(1:1), 0 °C to RT; 45% insted
of just OsO4 (g) COCl2, DMSO, Et3N, CH2Cl2, À78 °C; 68%.
It was also possible to adapt this route to simpler steroid-like
building blocks such as the Wieland–Miescher ketone 8 which
led to a CD-truncated system with an intact diacylfuran substruc-
ture 12 (Scheme 3). This compound was envisioned as a minimal
After considerable experimentation, we were pleased to find
that bromination of androstenedione 1 with bromine followed by

K. Viswanathan et al. / Bioorg. Med. Chem. Lett. 22 (2012) 6919–6922
6921
Figure 2. Competitive ring opening study; (a) an equal mixture of 2 and wortmannin and the diagnostic C20 protons (b) addition of n-butylamine to the mixture shows
preferential attack on 2 relative to wortmannnin, monitored by following the change in diagnostic proton.
pharmacophore model that retains the central furan, activated
by the C3 and C7 carbonyl groups but lacks the hydrophobic
CD-system of the parent compound.
O
O
NH2
The racemic enone 7 was processed as previously described to
produce 9. Selective deprotection of the primary silylether was
followed by regioselective osmylation of the disubstituted olefin
to produce the triol 10. Exposure of the triol to an excess of the
Swern reagent effected sequential oxidation and heterocycle
formation to produce the truncated wortmannin analog 11.
A central feature of the mechanism of inhibition of PI-3K by
wortmannin is the formation of an enzyme adduct through attack
of an active site lysine toward the activated, electrophilic furan.
Using a simple NMR-based competition experiment, we evaluated
the reactivity of 2 towards a limiting amount of a model biological
nucleophile (n-butylamine) and compared the relative reactivity
with that of wortmannin (Fig. 2).
The NMR spectra showed a rapid reaction between the furanos-
teroid analog and the amine with comparatively less conversion of
the natural product. The increase in relative reactivity is likely
influenced by the greater withdrawing power of the A-ring ketone
relative to the C3-lactone found in wortmannin. Based on the reac-
tivity of 2, it was straightforward to prepare a ring-opened prodrug
analogous to PX-866 by reaction with allyl amine (Scheme 4).17
At this point, it was important to conduct a preliminary evalu-
ation of the ability of the three viridin analogs to inhibit the
enzyme target and the growth of a cancer cell line. Recombinant
H
H
O
H
H
O
O
O
E
OH
O
NH
2
12
Scheme 4. Ring opening of furanosteroid 2.
natural product. Analysis of the crystal structure18 of wortmannin
bound to PI-3K (PDB ID; 1E7U) provided a rationale for the
c
relative changes in affinity for the synthetic analogs. The key differ-
ences between wortmannin and 2 include the absence of the polar
functionality at C1 and C11 of wortmannin, the change in hybrid-
ization at the BC-ring junction and use of a ketone as a surrogate
for the C3-lactone functionality. Major interactions between
Tyr867, Ser806, Asp964 and Val882 are predicted to be maintained
with compound 2 as well as several hydrophobic interactions with
the steroid core. The primary loss of affinity may be ascribed to loss
of hydrophobic interactions between the C11 acetate with Trp812/
Met804. Interestingly, the minimal, racemic pharmacophore model
11 showed stronger enzyme inhibition than the derivatives with
the full steroid core. This compound is still poised to make the
strong H-bonding interactions with Tyr867, Ser806, Asp964 but
would also lose the hydrophobic interactions with Trp812/
Met804 as well as the interaction between Val882 and the D-ring
ketonic oxygen. The increased affinity relative to 2/12 may result
human PI-3Ka is a frequently targeted isoform and was chosen
for enzyme inhibition assays (Table 1).
All three of the viridin analogs maintained sub-micromolar lev-
els of enzyme inhibition albeit with reduced potency relative to the

6922
K. Viswanathan et al. / Bioorg. Med. Chem. Lett. 22 (2012) 6919–6922
Table 1
Acknowledgment
Biological evaluation of viridin analogs
Compound
IC50,p110
a/p85
a
(nM)
IC50, MCF-7 (
lM)
We thank the American Cancer Society (RSG-04-267-01) for
support of this work.
2
11
12
338.9
177.5
270.5
11.9
22.9
>100
37.7
>100
Supplementary data
Wortmannin
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2012.09.
015.
from increased flexibility in the binding mode and the placement
of an additional polar hydroxyl group. This truncated analog may
serve as an excellent starting point for new inhibitor development
as the small size would easily allow the introduction of a wide
range of substituents to increase potency or selectivity.
References and notes
1. Willems, L.; Tamburini, J.; Chapuis, N.; Lacombe, C.; Mayeux, P.; Bouscary, D.
Curr. Oncol. Rep. 2012, 14, 129.
The finding that these analogs maintained sub-micromolar lev-
els of enzyme inhibition prompted us to evaluate them for poten-
tial antiproliferative effects in a breast cancer cell line (MCF-7).
Although there is a strong link between PI-3K dysregulation and
breast cancer,19–21 our results show that wortmannin was not
effective against this particular cell line. This observation is in
accord with previous reports.22,23 However, the steroid-derived
analog 2 and the ring-opened variant 12 both showed reasonable
levels of activity against this line with IC50 values of 22.9 and
2. Fruman, D. A.; Meyers, R. E.; Cantley, L. C. Ann. Rev. Biochem. 1998, 67, 481.
3. Chen, Y.; Wang, B. C.; Xiao, Y. J. Cell. Physiol. 2011, 227, 2818.
4. Sundstrom, T.; Anderson, A.; Wright, D. L. Org. Biomol. Chem. 2009, 7, 840.
5. 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.
6. Nuss, J. M.; Tsuhako, A.; Anand, N. K. Ann. Rep. Med. Chem. 2009, 44, 339.
7. Ihle, N. T.; Williams, R.; Chow, S.; Chew, W.; Berggren, M. I.; Paine-Mrrieta, G.;
Minion, D. J.; Halter, R. J.; Wipf, P.; Abraham, R.; Kirkpatrick, L.; Powis, G. Mol.
Cancer Ther. 2004, 3, 763.
8. Zask, A.; Kaplan, J.; Toral-Barza, L.; Hollander, I.; Young, M.; Tischler, M.;
Gaydos, C.; Cinque, M.; Lucas, J.; Yu, K. J. Med. Chem. 2008, 51, 1319.
9. Sato, S.; Nakada, M.; Shibasaki, M. Tetrahedron Lett. 1996, 37, 6141.
10. Mizutani, T.; Honzawa, S.; Tosaki, S.; Shibasaki, M. Angew. Chem., Int. Ed. 2002,
41, 4680.
37.7 lM respectively. The truncated analog 11 proved inactive in
the cellular assay which may relate to the poor permeability of this
relatively polar compound (logP = 0.58 versus 2.65 for compound
2). Although the origins of the activity differences between wort-
mannin and 2 require further investigation, it does illustrate that
subtle structural changes can have a significant impact on cellular
activity.
11. Shigehisa, H.; Mizutani, T.; Tosaki, S.; Ohshima, T.; Shibasaki, M. Tetrahedron
2005, 61, 5057.
12. Anderson, E. A.; Alexanian, E. J.; Sorensen, E. J. Angew. Chem., Int. Ed. 2004, 43,
1998.
13. Hodson, H. F.; Madge, D. J.; Widdowson, D. A. J. Chem. Soc. Perkin Trans. 1 1995,
1, 2965.
The viridins are potent and highly complex inhibitors of PI-3K
that act by promoting the addition of a nucleophilic active site
lysine to an unusual electrophilic furan. As the potential of PI-3K
as a drug target in cancer has received increased attention, there
is a need for direct and flexible routes to new inhibitors. We have
described the use of abundant steroid and steroid-like building
blocks for the preparation of viridin analogs. These first-generation
prototype molecules maintain sub-micromolar levels of enzyme
inhibition and interestingly show improved anti-proliferative
activity against a breast cancer line. The easy availability of the
starting materials and relatively short routes should allow for the
preparation of several new analogs to improve enzyme potency
and anti-proliferative activity.
14. Ali, H.; Van Lier, J. E. J. Chem. Soc. Perkin Trans. 1 1991, 2485.
15. Majeed, A. J.; Antonsen, O.; Benneche, T.; Undheim, K. Tetrahedron 1989, 45,
993.
16. Kraus, G. A.; Wan, Z. Synlett 2000, 363.
17. Wipf, P.; Minion, D. J.; Halter, R. J.; Berggren, M. I.; Ho, C. B.; Chiang, G. G.;
Kirkpatrick, L.; Abraham, R.; Powis, G. Org. Biomol. Chem. 1911, 2004, 2.
18. Walker, E. H.; Pacold, M. E.; Perisic, O.; Stephens, L.; Hawkins, P. T.; Whymann,
M. P.; Williams, R. Mol. Cell 2000, 6, 909.
19. Osaki, M.; Oshimura, M.; Ito, H. Apoptosis 2004, I, 667.
20. Gershtein, E. S.; Scherbakov, A. M.; Shatskaya, V. A.; Kushlinsky, N. E.;
Krasilnikov, M. A. Anticancer Res. 2007, 27, 1777.
21. Granville, C. A.; Memmott, R. M.; Gills, J. J.; Dennis, P. A. Clin. Cancer Res. 2006,
12, 679.
22. Lemke, L. E.; Paine-Murrieta, G. D.; Taylor, C. W.; Powis, G. Cancer Chemother.
Prmacol. 1999, 44, 491.
23. Sabnis, G. J.; Jelovac, D.; Long, B.; Brodie, A. Cancer Res. 2005, 65, 3903.