Studies toward the synthesis of luminacin D: Assembly of simplified analogues devoid of the epoxide displaying antiangiogenic activity

Article abstract of DOI:10.1021/ol048462v

(Chemical Equation Presented) A series of structurally simplified luminacin analogues devoid of the epoxide ring are assembled in a stereocontrolled manner from 2,4-dimethoxybenzaldehyde using a syn-selective aldol reaction as the key step. The success of

Full text of DOI:10.1021/ol048462v

ORGANIC
LETTERS
2004
Vol. 6, No. 22
3909-3912
Studies Toward the Synthesis of
Luminacin D: Assembly of Simplified
Analogues Devoid of the Epoxide
Displaying Antiangiogenic Activity
Mark W. Davies,† Lesley Maskell,‡ Michael Shipman,*,† Alexandra M. Z. Slawin,§
Sandrine M. E. Vidot,† and Jacqueline L. Whatmore‡
Department of Chemistry, UniVersity of Warwick, Gibbet Hill Road,
CoVentry, CV4 7AL, UK, Cell and Molecular Biology, Peninsula Medical School,
Noy Scott House, Exeter, EX2 5EQ, UK, and School of Chemistry,
UniVersity of St. Andrews, Purdie Building, St. Andrews, Fife, KY16 9ST, UK
m.shipman@warwick.ac.uk
Received August 3, 2004
ABSTRACT
A series of structurally simplified luminacin analogues devoid of the epoxide ring are assembled in a stereocontrolled manner from 2,4-
dimethoxybenzaldehyde using a syn-selective aldol reaction as the key step. The success of the approach is critically dependent on the
nature and extent of the alcohol protecting groups. The synthetic analogues inhibit VEGF-stimulated angiogenesis in an in vitro assay indicating
that the epoxide is not essential for biological activity in this compound class.
Angiogenesis, the process by which new blood capillaries
develop from an existing blood vessel, is of fundamental
importance to the development of a cancer mass.1 Tumors
can only grow to about 1 mm in diameter in the absence of
new capillary growth.2 Today, it is widely recognized that
antiangiogenesis therapies hold considerable promise for the
treatment of cancer and other diseases.3 In 2000, a new class
of inhibitors called the luminacins were isolated from the
culture broths of Streptomyces sp. Mer-VD1207 by Naruse
et al.4 Wakabayashi demonstrated that they inhibit the initial
stages of capillary tube formation and operate by a unique
mode of action.5 The structures of luminacins C2 and D, two
of the more potent members of the family, are depicted in
Figure 1.6 Further studies by Sharma using luminacin C2
† University of Warwick.
‡ Peninsula Medical School.
§ University of St Andrews.
(1) Carmeliet, P.; Jain, R. K. Nature 2000, 407, 249.
(2) King, R. J. B. Cancer Biology; Longman: Edinburgh, 1996.
(3) Han, Z. C.; Liu, Y. Int. J. Hematol. 1999, 70, 68. Eatock, M. M.;
Scha¨tzlein, A.; Kaye, S. B. Cancer Treat. ReV. 2000, 26, 191. Matter, A.
Drug DiscoVery Today 2001, 6, 1005.
Figure 1. Structures of luminacins C2 and D and synthetic analogue
1.
10.1021/ol048462v CCC: $27.50
© 2004 American Chemical Society
Published on Web 10/06/2004

Scheme 1
reveal that it is an inhibitor of Src signal transduction, acting
by disruption of proline-rich ligand-mediated protein-protein
interactions rather than by direct inhibition of Src kinase
activity.7 On the basis of these observations, the luminacins
have become important targets for angiogenesis research. To
further probe the biological function of the luminacins,
methods for the synthesis of these natural products and
related structures are required. Elegant approaches to lumi-
nacins C8 and D9 have recently been reported. Our endeavors
have focused on exploring if a syn-selective aldol reaction
can be used to construct the central C-2′/C-3′ bond in a
stereoselective manner. In this Letter, we report the viability
of this approach by realizing the synthesis of luminacin
analogue (()-1 containing many of the structural elements
of the natural products. Furthermore, we show that com-
pounds of this type inhibit vascular endothelial growth factor
(VEGF)-stimulated angiogenesis in human umbilical vein
endothelial cells (HUVECs), a discovery that establishes that
the epoxide is not essential for inhibiting angiogenesis in
this compound class.
Our initial synthetic approach to 1 is outlined in Scheme
1. 2,4-Dimethoxybenzaldehyde 2 was converted into trisub-
stituted benzene derivative 3 in quantitative yield by Wittig
olefination and reduction of the resulting alkene by catalytic
hydrogenation. Friedel-Crafts acylation of 3 using pentanoyl
chloride gave ketone 4 as a single regioisomer. The C-1
hydroxymethyl group was introduced by removal of the
methyl ethers from 4 with BBr3, which was further reacted
with formaldehyde under basic conditions to give benzyl
alcohol 5.10 Selective methylation of the phenolic hydroxyl
groups of 5 provided pentasubstituted aryl ketone 6. Treat-
ment of this ketone with titanium tetrachloride and tributyl-
amine according to the Evans protocol11 furnished the
tetrachlorotitanium enolate, which was condensed with
5-hexenal to give the syn-aldol 7 in 81% yield. This reaction
(4) Naruse, N.; Kageyama-Kawase, R.; Funahashi, Y.; Wakabayashi, T.;
Watanabe, Y.; Sameshima T.; Dobashi, K. J. Antibiot. 2000, 53, 579.
(5) Wakabayashi, T.; Kageyama-Kawase, R.; Naruse, N.; Funahashi Y.;
Yoshimatsu, K. J. Antibiot. 2000, 53, 591.
(6) Luminacin C2 is identical to UCS15A isolated from another Strep-
tomyces sp. (see ref 7a). In this article, we use the term luminacin C2 to
refer to both luminacin C2 and UCS15A.
(7) (a) Sharma, S. V.; Oneyama, C.; Yamashita, Y.; Nakano, H.;
Sugawara, K.; Hamada, M.; Kosaka N.; Tamaoki, T. Oncogene 2001, 20,
2068. (b) Oneyama, C.; Nakano H.; Sharma, S. V. Oncogene 2002, 21,
2037. (c) Oneyama, C.; Agatsuma, T.; Kanda, Y.; Nakano, H.; Sharma, S.
V.; Nakano, S.; Narazaki, F.; Tatsuta, K. Chem. Biol. 2003, 10, 443.
(8) Tatsuta, K.; Nakano, S.; Narazaki, F.; Nakamura, Y. Tetrahedron
Lett. 2001, 42, 7625.
(9) Shotwell, J. B.; Krygowski, E. S.; Hines, J.; Koh, B.; Huntsman, E.
W. D.; Choi, H. W.; Schneekloth, J. S.; Wood J. L.; Crews, C. M. Org.
Lett. 2002, 4, 3087. Fang, F.; Johannes, C.; Yao Y.; Zhu, X. Patent Appl.
WO 03/057685, 2003.
Figure 2. X-ray crystal structure of 8.
3910
Org. Lett., Vol. 6, No. 22, 2004

Scheme 2
was highly stereoselective, with no trace of the anti dia-
We suggest that 10 is produced because it is thermodynami-
cally more stable as a result of an intramolecular hydrogen
bond between the ketone and the phenolic hydroxyl group.
The downfield chemical shift of the -OH group in the 400
stereomer being detected in the crude reaction mixture by
1H NMR spectroscopy. Esterification of diol 7 with 4-nitro-
benzoyl chloride furnished dibenzoate 8, whose relative
stereochemistry was established by X-ray crystallography
(Figure 2).12 Further conversion of diol 7 into 9 was realized
by oxidation of the benzylic hydroxyl group with manganese
dioxide followed by cleavage of the terminal alkene to the
aldehyde, which immediately resulted in lactol formation.
Lactol 9 existed as a 54:46 mixture of anomers (in CDCl3).
Unfortunately, at this late stage efforts to effect demethylation
to 1 using Lewis acids (e.g., BBr3) met with failure.
The evidence gathered during the attempts to deprotect 9
suggested that one of the methyl ethers was much more
readily cleaved than the other. Since carbonyl groups are
known to assist in demethylation at adjacent sites,13 we
speculated that removal of the C-2 methyl ether was more
facile than at C-6. Thus, a revised strategy was conceived
in which only the phenol at C-2 would be protected.14
To realize this strategy, chemoselective protection of triol
5 was undertaken, which furnished phenol 10 in 93% yield
(Scheme 2). The structure of 10 has been unambiguously
established by X-ray crystallography (data not presented).
1
MHz H NMR spectrum (δ 12.90 in CDCl3) supports this
proposal. Methylation of 10 and deprotection of the acetal
provided 11 in 60% yield over the two steps. Titanium-
mediated condensation between 11 and 5-hexenal furnished
the aldol products in 76% yield. High diastereoselectivity
was again observed (syn:anti ) 87:13). By analogy with 7,
the major product was assigned as the syn stereoisomer 12.
Oxidation of the hydroxymethyl group of syn-12 to the
corresponding aldehyde proceeded cleanly. Interestingly, this
oxidation was much more rapid than that of 7 (4 h instead
of 7 days). Tetrahydropyran formation was uneventful,
yielding 13 in good yield as a 57:43 mixture of anomers (in
CDCl3). Gratifyingly, demethylation of 13 to 1 could be
achieved by heating with lithium chloride in DMF at 80 °C
for 14 h. While considerable amounts of 13 were recovered
under these conditions (27%), efforts to drive the reaction
to completion led to degradation and the production of an
inseparable impurity that we tentatively assign to be the C-2′
epimer of 1. Thus, luminacin analogue (()-1 can be obtained
in 13 steps from 2,4-dimethoxybenzaldehyde in 6% overall
yield using this strategy.
Compounds 1, 9, and 13 were evaluated for their effects
on the inhibition of VEGF-induced angiogenesis in HUVECs
using an established fibrin matrix assay.15 For comparison
purposes in the biological assays, 14 was prepared by
oxidation of 5 (100%) with manganese dioxide (Scheme 2).
At a drug concentration of 50 µM, VEGF-stimulated
angiogenesis was almost completely inhibited (97.6% (
2.1%) using bismethyl ether 9. The inhibitory effects of this
compound on tube formation are clearly visible in Figure 3.
Surprisingly, phenols 1 and 13, which more closely resemble
luminacin D, were less effective at this drug concentration
(10) Saimoto, H.; Yoshida, K.; Murakami, T.; Morimoto, M.; Sashiwa
H.; Shigemasa, Y. J. Org. Chem. 1996, 61, 6768.
(11) Evans, D. A.; Rieger, D. L.; Bilodeau, M. T.; Urpi, F. J. Am. Chem.
Soc. 1991, 113, 1047. Yoshida, Y.; Hayashi, R.; Sumihara, H.; Tanabe, Y.
Tetrahedron Lett. 1997, 38, 8727.
(12) Crystallographic Data for 8. X-ray diffraction studies on a colorless
crystal grown from methanol were performed at 125 K using a Bruker
SMART diffractometer with graphite-monochromated radiation (λ )
0.71073 Å). The structure was solved by direct methods. C38H44N2O11, M
) 704.75, space group P-1, triclinic, a ) 8.9292(17), b ) 13.439(3), c )
16.334(3) Å, R ) 109.518(3), â ) 99.424(4), γ ) 93.348(4)°, U )
1809.0(6) Å3, Z ) 2, 10 840 reflections were measured, 6424 were unique
(Rint ) 0.0338) which were used in all calculations. The final R was 0.0732
for I > 2σ(I), and wR(F2) was 0.1660. Some disorder was observed in the
alkenyl chain, which was modeled in two equal occupancy orientations.
(13) Kocienski, P. J. Protecting Groups; Thieme: Stuttgart, 1994.
(14) Aldol reactions using triol 5 were unsuccessful. Other protecting
groups (MOM, Bn, and allyl) for the simultaneous protection of the C-2
and C-6 phenolic groups were examined, but these proved to be less
practical.
(15) Whatmore, J. L.; Swann, E.; Barraja, P.; Newsome, J. J.; Bunderson,
M.; Beall, H. D.; Tooke, J. E.; Moody, C. J. Angiogenesis 2002, 5, 45.
Org. Lett., Vol. 6, No. 22, 2004
3911

based colorimetric assay using HUVECs.16 Crucially, it was
not cytotoxic (IC50 > 100 µM) at the concentrations used in
the angiogenesis assay.
These findings establish that an aldol strategy is suitable
for the construction of “luminacin-like” structures, although
judicious choice of protecting groups is required. Further-
more, we have determined that in vitro VEGF-stimulated
angiogenesis can be efficiently inhibited by simplified
analogues. These data establish that free phenolic groups at
C-2 and C-6 are not essential for inhibition, and neither is
the presence of the C-6′/C-8′ epoxide ring or the C-5′
hydroxyl group of the natural products. Current work is
focused on applying these findings to the synthesis of more
potent inhibitors of angiogenesis and to completing the total
synthesis of luminacin D.
Figure 3. Inhibition of VEGF-stimulated angiogenesis of HUVECs
on a fibrin matrix using synthetic luminacin analogue 9:15 (a) control
(no added VEGF); (b) added VEGF (100 ng/mL); (c) added 9 (50
µM); (d) added VEGF (100 ng/mL) and 9 (50 µM). All experiments
were conducted on at least two different cell populations.
Acknowledgment. We are indebted to EPSRC (GR/
S14429/01), Cancer Research UK, and FORCE Cancer
Research (Exeter) for their generous financial support of this
work. We thank the EPSRC Chemical Database Service at
Daresbury.17
(1, 27.5% ( 26.5%; 13, 28.4% ( 31.8%). Analogue 14,
devoid of the pyran ring, failed to inhibit angiogenesis at 50
µM. Good correlation between drug concentration and
percentage inhibition was witnessed with these compounds
{e.g., 1: 100% (100 µM); 53.0 ( 9.4% (80 µM); 27.5 (
26.5% (50 µM); 2.9 ( 3.8% (20 µM)}. The cytotoxicity of
the most potent analogue 9 was determined using an MTT-
Supporting Information Available: Detailed experi-
mental procedures and 1H and 13C NMR spectra for
compounds 1, 3-7, and 9-14. This material is available
free of charge via the Internet at http://pubs.acs.org.
OL048462V
(17) Fletcher, D. A.; McMeeking R. F.; Parkin, D. J. Chem. Inf. Comput.
Sci. 1996, 36, 746.
(16) Mosmann, T. J. Immunol. Methods 1983, 65, 55.
3912
Org. Lett., Vol. 6, No. 22, 2004