Synthetic and computational studies on liphagal: A natural product inhibitor of PI-3K

Article abstract of DOI:10.1016/j.tetlet.2010.09.058

The natural product liphagal has been shown to function as a reasonably potent and selective inhibitor of the key signaling enzyme PI-3Kα. We have been interested in developing an analog class of PI-3K inhibitors based upon this unusual terpenoid natural product. Toward that end, we have evaluated the binding of the natural product to its target protein computationally and formulated a class of simplified analogs based on the structural analysis. Utilizing the cycloadduct derived from tetrabromocyclopropene and furan, we were able to generate a key, versatile scaffold upon which to pursue this analog design.

Full text of DOI:10.1016/j.tetlet.2010.09.058

Tetrahedron Letters 51 (2010) 6120–6122
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Synthetic and computational studies on liphagal: a natural product inhibitor
of PI-3K
Yanzhong Zhang a, E. Zachary Oblak a, Erin S. D. Bolstad a, Amy C. Anderson a, Jerry P. Jasinski b,
Ray J. Butcher c, Dennis L. Wright a,
⇑
a Department of Pharmaceutical Science, University of Connecticut, Storrs, CT 06269, United States
b Department of Chemistry, Keene State College, Keene, NH 03435, United States
c Department of Chemistry, Howard University, Washington, DC 20059, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
The natural product liphagal has been shown to function as a reasonably potent and selective inhibitor of
Received 6 August 2010
Revised 10 September 2010
Accepted 14 September 2010
Available online 21 September 2010
the key signaling enzyme PI-3Ka. We have been interested in developing an analog class of PI-3K inhib-
itors based upon this unusual terpenoid natural product. Toward that end, we have evaluated the binding
of the natural product to its target protein computationally and formulated a class of simplified analogs
based on the structural analysis. Utilizing the cycloadduct derived from tetrabromocyclopropene and
furan, we were able to generate a key, versatile scaffold upon which to pursue this analog design.
Ó 2010 Elsevier Ltd. All rights reserved.
The enzyme phosphatidylinositol-3-kinase (PI-3K) is an impor-
tant mediator of key cell signaling cascades and functions by trans-
ferring a phosphate from ATP to a hydroxyl function on the cyclitol
head group of phosphoinositides.1 Dysregulation of PI-3K and
other regulators of this pathway have been linked to a wide variety
of malignancies and as such, there has been considerable interest
in the development of effective inhibitors of the enzyme.2 PI-3K
is found in several different isoforms and the development of iso-
zyme-selective inhibitors is considered to be an important step in
the development of highly effective therapeutic agents.3
OH
HO
16
CHO
HO
D
O
C
O
1
11 B
A
5
8
H
liphagal
frondosin B
Figure 1. Structures of terpenoid kinase inhibitors.
Recently, a new natural product inhibitor of PI-3K was isolated
from marine sources and named liphagal.4 This interesting terpe-
noid is structurally related to the frondosins5 which have been
identified as inhibitors of the serine/threonine kinase protein ki-
nase C (PKC) (Fig. 1).
based on this natural product. In order to assess the interactions
of liphagal with PI-3K, we first generated an accurate docking mod-
el of PI-3K bound to compounds with measured inhibition values.7
Nine inhibitors of PI-3K: liphagal, wortmannin, LY294002, querce-
tin, staurosporine, myricetin, 039, 090, 093, and QYT were flexibly
docked using Surflex-Dock as implemented by Sybyl 7.28 to crystal
structures of PI-3K9 with and without ordered water molecules in
the active site. For wortmannin, a covalent inhibitor, three forms
with different representations of the electrophilic furan (intact,
ring-opened, and a surrogate form replacing C21 with a hydro-
gen-bond accepting keto-oxygen to simulate the covalent bond
with Lys 833) were docked. Scoring was based on hydrophobic, po-
lar, electrostatic repulsive, entropic, and solvation terms.10 The
docking results were assessed for correct orientation of the ligands
in the active site and relative docking scores as compared to exper-
imentally measured inhibition values.
Liphagal was shown to be a relatively potent inhibitor of PI-3K
(IC50 = 100 nM) and, interestingly, showed a 10-fold selectivity
over the inhibition of PI-3K as well as antiproliferative effects
a
c
against several cancer cell lines. We have been interested in the
development of isozyme-selective inhibitors of PI-3K and feel that
this natural product may serve as an excellent template for new
discovery efforts. In this Letter, we wish to report on both compu-
tational and synthetic studies6 of this natural product that will
serve as the basis for the development of new PI-3K inhibitors.
As our ultimate interest is in the development of inhibitors with
greater levels of potency and selectivity, we wanted to utilize the
rich structural information available concerning PI-3K in complex
with various inhibitors to help guide the design of new analogs
The orientation of the ligand in the active site, which deter-
mined the number of possible satisfied hydrogen bonding interac-
tions, is an important contributor to the docking score. The docking
model based on the structure of PI-3K bound to wortmannin
⇑
Corresponding author. Tel.: +1 860 486 9451; fax: +1 860 486 6847.
E-mail address: dennis.wright@uconn.edu (D.L. Wright).
0040-4039/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2010.09.058

Y. Zhang et al. / Tetrahedron Letters 51 (2010) 6120–6122
6121
(1E7U9) yielded the most accurate results. This model shows the
correct co-crystallized ligand placement, the highest number of to-
tal ligands oriented correctly in the active site, and proper ranking
of wortmannin, with the highest docking score, relative to the
other and less potent ligands. The inclusion of the ordered water
molecules did not affect the results. Figure 2 shows the docked
structure of liphagal to the kinase.
Using this model, it is possible to interpret the many interac-
tions of liphagal with PI-3K. Liphagal forms lipophilic interactions
with Ile963, Ile879, the aromatic region of Tyr867, Ile831, Ile881,
Val882, and Trp812. The cyclohexyl A-ring projects up from the
aromatic plane and into the upper pocket of the lipophilic region.
The polar region consists of several hydrogen bonds made between
the heteroatom rich aromatic D-ring and Lys833, Asp964, Lys807,
and Ser806 (Fig. 3).
Based on our computational studies, we envisioned a general-
ized tricyclic scaffold that would give access to both the polar
and hydrophobic regions of the enzyme while maintaining a
relatively rigid core structure. This would require a modular syn-
thesis whereby differentially substituted benzofuran units could
be easily fused to a suitably functionalized seven-membered ring
(Scheme 1).
Figure 3. Interaction map of liphagal with PI-3K.
We hoped to access this type of intermediate from the oxa-
bridged enone 1, a readily available and highly flexible compound
with which we have worked for several years11 and one that can be
prepared in enantiomerically pure form.12 Key to utilizing this
intermediate would be the development of a direct method for
annulations of the fused benzofuran unit, stereoselective installa-
tion of the secondary methyl group on the cycloheptyl ring, and
facile opening of the bridging ether atom.
The initial studies began with developing a straightforward
method to annulate a benzofuran heterocycle on to the bridged
synthon 1 by taking advantage of the differential reactivity of the
two halogen atoms in the dibromoenone function.13 It was antici-
pated that the increased reactivity of the b-bromide would allow
an effective sequential metal-mediated cross-coupling reaction
with a suitable aromatic partner. As a model unit, we studied the
Suzuki coupling of 2-hydroxyphenyl boronic acid with the dib-
romoenone. The coupling proceeded exclusively to deliver the phe-
nol 2 in good yield under the optimized conditions (Scheme 2).
The ability to directly cross-couple the free phenol to the dib-
romoenone set the stage for a second metal-mediated coupling
to close the benzofuran ring. Exposure of the phenol to stoichiom-
etric copper(I) iodide effected a high yielding Ullman-type cou-
pling to give 3. The regiochemical outcome of the annulation was
confirmed by crystallographic analysis of ketone 3 (Fig. 4).
This direct, two-step method for the annulation of a benzofuran
onto the dibromoenone should allow access to a wide variety of
Y
X
OH
HO
Z
CHO
Br
Br
O
O
simplify
1
R1
R2
O
O
11
5
8
H
R3
liphagal
generalized
scaffold
1
Scheme 1. Synthetic strategy for analog generation.
Br
OH
Br
HO
1 equiv. CuI
Et3N, 72%
O
Br
B(OH)2
O
O
O
O
PdCl2(PPh3)2, Ag2CO3
87%
O
O
1
2
3
Scheme 2. Two-step benzofuran annulation.
analogs in the aromatic region. We were also pleased to see that
a small modification to the protocol allowed the annulation to be
carried out with a more oxidized aromatic precursor to install a
key phenolic oxygen at C16 (Scheme 3).
It was found that incorporation of the additional methoxy group
rendered the standard boronic acid coupling partner prone to pro-
to-deboronation. Coupling of 1 to an excess of 2-hydroxy, 4-
methoxyphenyl boronic acid produced low yields (<50%) of the ex-
pected Suzuki product. Moreover, the coupled product and the
deboronated phenol were very tedious to separate. Fortunately,
switching to the trifluroboronate derivative developed by Moland-
er and co-workers14 allowed for a productive coupling. Using only
a slight excess, the previously unreported phenol 415 was effec-
tively cross-coupled in good yield and in sufficient purity to be ta-
ken on without the benefit of further purification to give 5 in 79%
overall yield for the two-steps. With a successful route to the ben-
zofuran in place, our attention turned to installing key functional-
ity for liphagal analogs on to the central seven-membered ring.
Figure 2. Docked complex of liphagal and PI-3K.

6122
Y. Zhang et al. / Tetrahedron Letters 51 (2010) 6120–6122
catalyst led to saturation of both olefins to give 7 as the exclusive
diastereomer where reduction of the exocyclic methylene group
occurred selectively from the exo-face of the bicyclic framework.
This suggests that simply using a single isomer of the starting dib-
romoenone will allow us to transfer the stereochemistry from the
initial cycloaddition to the secondary methyl group on the B-ring, a
key stereogenic center in liphagal and the sole stereogenic center
in frondosin B. To further elaborate ring-B toward the desired
intermediates, the ability to effectively open the ether bridge was
evaluated. Treatment of 7 with boron triflouride etherate led to a
very facile opening of the bridged ether to give 8. The regiochem-
istry of the opening is controlled by the stabilization of an interme-
diate carbonium ion by the annulated benzofuran. The olefinic and
hydroxyl functionality proved versatile for the introduction of var-
ious appendages needed for the generation of analogs.
In conclusion, we have made efforts toward using the structure
of the natural product liphagal to develop isozyme-selective inhib-
itors of PI-3K. Utilizing high-resolution structures of the kinase, a
docked complex of liphagal and the protein was created. Analysis
of this complex led to the design of a basic simplified tricyclic scaf-
fold for development of new inhibitors. A route to these types of
scaffolds was developed using a key bridged bicyclic intermediate.
Current efforts are directed toward synthesis and evaluation of
simplified liphagal analogs.
Figure 4. X-ray structure of intermediate 3.
MeO
MeO
4
OH
(1.1eq.)
Br
1)
Br
BF3K
O
Acknowledgment
O
Pd(PPh3)4,Cs2CO3
2) CuI (1 eq),
MeCN:Et3N, 80 ºC
O
79% for two steps
O
O
The authors are grateful to the American Cancer Society (RSG-
04-267-01) for generous support of this work.
1
5
Scheme 3.
References and notes
1. Fruman, D. A.; Meyers, R. E.; Cantley, L. C. Ann. Rev. Biochem. 1998, 67, 481.
2. Vivanco, I.; Sawyers, C. L. Nat. Rev. Cancer 2002, 2, 489.
3. Sundstrom, T.; Anderson, A.; Wright, D. L. Org. Biomol. Chem. 2009, 7, 840.
4. Marion, F.; Willaims, D. E.; Patrick, B. O.; Hollander, I.; Mallon, R.; Kim, S. C.;
Roll, D. M.; Fledberg, L.; Van Soest, R.; Andersen, R. J. Org. Lett. 2006, 8, 321.
5. (a) Patil, A. D.; Freyer, A. J.; Kilmer, L.; Offen, P.; Carte, B.; Jurewicz, A. J.;
Johnson, R. K. Tetrahedron 1997, 9, 3837; (b) Hallock, Y. F.; Cardellina, J. H.;
Boyd, M. R. Nat. Prod. Res. 1998, 11, 153.
H2
Ph3P=CH2
96%
O
O
(PPh3)3RhCl
90%
O
O
O
CH2
6. For other synthetic approaches see Ref. 5a and (a) Mehta, G.; Likhite, N. S.;
Kumar, C. S. A. Tertahedron Lett. 2009, 50, 5260; (b) George, J. H.; Baldwin, J. E.;
Adlington, R. M. Org. Lett. 2010, 12, 2394; (c) Alvarz-Manzaneda, E.; Chahboun,
R.; Alvarez, E.; Cano, M. J.; Haidour, A.; Alvarez-Manzaneda, R. Org. Lett.
doi:10.1021/ol101173w.
3
6
BF3 OEt, DBU
79%
O
O
7. Bolstad, E. S.; Anderson, A. Proteins 2008.
8. SYBYL 7.3, Tripos Inc., 1699 South Hanley Rd., St. Louis, Missouri 63144, USA.
9. Walker, E. H.; Pacold, M. E.; Perisic, O.; Stephens, L.; Hawkins, P. T.; Wymann,
M. P.; Williams, R. L. Mol. Cell 2000, 6, 909.
CH3
O
CH3
OH
10. Jain, A. N. J. Comput. Aided Mol. Des. 1996, 10, 427.
8
one diastereomer
11. (a) Orugunty, R. S.; Wright, D. L.; Battiste, M. A.; Abboud, K. A. Org. Lett. 2002, 4,
1997; (b) Orugunty, R. S.; Wright, D. L.; Battiste, M. A.; Helmich, R. J.; Abboud,
K. J. Org. Chem. 2004, 69, 406; (c) Pelphrey, P.; Jasinski, J.; Butcher, R. J.; Wright,
D. L. Org. Lett. 2005, 7, 423; (d) Pelphrey, O. M.; Bolstad, D. B.; Wright, D. L.
Synlett 2007, 2647.
12. Pelphrey, P. M.; Abboud, K. A.; Wright, D. L. J. Org. Chem. 2004, 69, 6931.
13. Pelphrey, P. M.; Orugunty, R. S.; Helmich, R. J.; Battiste, M. A.; Wright, D. L. Eur.
J. Org. Chem. 2005, 4296.
14. (a) Molander, G. A.; Figueroa, R. Alrichim. Acta 2005, 38, 49; (b) Park, Y. H.; Ahn,
H. R.; Canturk, B.; Jeon, S. I.; Lee, S.; Kang, H.; Molander, G. A.; Ham, J. Org. Lett.
2008, 10, 1215.
15. Prepared from 4-methoxyphenol by bromination and protection of the phenol
as a THP ether (80% overall) followed by lithiation (t-BuLi) and reaction with
B(OiPr)3 and treatment with 2 N HCl then KHF2 (85% overall).
7
Scheme 4.
The B-ring ketone provides convenient functionality for the
installation of the C8-secondary methyl group common to both
liphagal and the frondosins (Scheme 4).
We anticipated that the facial bias introduced through the
bridging ether could be exploited to control the stereoselectivity
during reduction of an exocyclic olefin. Treatment of ketone 3 with
a Wittig reagent gave rise to the exo-methylene derivative 6. Pleas-
ingly, catalytic hydrogenation of 6 in the presence of Wilkinson’s