Total synthesis and biological evaluation of cortistatins A and J and analogues thereof

Article abstract of DOI:10.1021/ja902939t

Total syntheses of the highly selective antiproliferative natural products cortistatins A (1) and J (5) in their naturally occurring enantiomeric forms are described. The modular and convergent strategy employed relies on an intramolecular oxa-Michael add

Full text of DOI:10.1021/ja902939t

Published on Web 07/01/2009
Total Synthesis and Biological Evaluation of Cortistatins A
and J and Analogues Thereof
K. C. Nicolaou,* Xiao-Shui Peng, Ya-Ping Sun, Damien Polet, Bin Zou,
Chek Shik Lim, and David Y.-K. Chen*
Chemical Synthesis Laboratory@Biopolis, Institute of Chemical and Engineering Sciences
(ICES), Agency for Science, Technology and Research (ASTAR), 11 Biopolis Way,
The Helios Block, #03-08, Singapore 138667
Received April 17, 2009; E-mail: kcn@scripps.edu; david_chen@ices.a-star.edu.sg
Abstract: Total syntheses of the highly selective antiproliferative natural products cortistatins A (1) and J
(5) in their naturally occurring enantiomeric forms are described. The modular and convergent strategy
employed relies on an intramolecular oxa-Michael addition/aldol/dehydration cascade reaction to cast the
ABCD ring framework of the molecule and both Sonogashira and Suzuki-Miyaura coupling reactions to
assemble the necessary building blocks into the required heptacyclic skeleton. A divergent approach from
a late-stage epoxy ketone leads to both target molecules in a stereoselective manner. The developed
synthetic technologies were applied to the construction of several analogues of the cortistatins which were
biologically evaluated and compared to the natural products with regards to their antiproliferative activities
against a variety of cancer cells. Analogues 8 and 81, lacking both the dimethylamino and hydroxyl groups
of cortistatin A, were found to exhibit comparable biological activity as the parent compound, leading to the
conclusion that such functionalities are not essential for biological activity.
Introduction
Angiogenesis is a physiological phenomenon that involves
the generation of new blood vessels in both healthy and disease
tissues of the human body.1 And while it is an essential and
required process for growth and repair, angiogenesis is also
responsible for the transition of tumors from a dormant to a
malignant state. It is for this reason that the inhibition of
angiogenesis in general, and the search for novel molecules
possessing such inhibitory properties in particular, became an
attractive strategy of intervention for the treatment of cancer
patients.2 In 2006, the Kobayashi group disclosed the cortistatin
family of natural products (e.g., cortistatins A-D, 1-4, Figure
1) and their selective antiproliferative properties against human
umbilical vein endothelial cells (HUVECs).3 Most strikingly
and as the most potent member of the family (IC50 ) 1.8 nM),
cortistatin A (1) demonstrated a selectivity index of more than
3000 against HUVECs in comparison with normal human
dermal fibroblast (NHDF) and several other tumor cells (KB3-
1, K562, and Neuro2A). More recently, cortistatins J-L (5-7,
Figure 1) were reported by the same group.4a Among them,
cortistatin J (5) exhibited the most potent antiproliferative
activity against HUVECs (IC50 ) 8 nM) with a selectivity index
of 300-1100 as compared with NHDF, KB3-1, K562, and
Figure 1. Structures of cortistatins A-D (1-4) and J-L (5-7).
Neuro2A cells. Isolated from the sponge Corticium simplex, the
cortistatins appear to be promising drug candidates or leads,
their potential tempered only by their low natural abundance.
The impressive biological properties coupled with their unprec-
edented molecular architectures and scarcity made the cortist-
atins enticing targets for chemical synthesis.5 Culminating in
the total synthesis of cortistatin A, our studies were reported in
a preliminary communication in 2008.6 In this paper we describe
the full account of our work in the cortistatin area that includes,
in addition to the total synthesis of several natural (A and J)
and designed (e.g., 8 and 81) members of the class, their
biological evaluation as antiproliferative agents.
(1) Carmeliet, P. Nature 2005, 438, 932–936.
(2) Ziche, M.; Donnini, S.; Morbidelli, L. Curr. Drug Targets 2004, 5,
485–493.
(3) Aoki, S.; Watanabe, Y.; Sanagawa, M.; Setiawan, A.; Kotoku, N.;
Kobayashi, M. J. Am. Chem. Soc. 2006, 128, 3148–3149.
(4) For isolation of cortistatins J-L, see: (a) Aoki, S.; Watanabe, Y.;
Tanabe, D.; Setiawan, A.; Arai, M.; Kobayashi, M. Tetrahedron Lett.
2007, 48, 4485–4488. For isolation of cortistatins E-H, see: (b)
Watanabe, Y.; Aoki, S.; Tanabe, D.; Setiawan, A.; Kobayashi, M.
Tetrahedron 2007, 63, 4074–4079.
Results and Discussion
The attractiveness of cortistatins A and J and their siblings
as targets for chemical synthesis was enhanced by the prospect
9
10.1021/ja902939t CCC: $40.75  2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 10587–10597 10587

A R T I C L E S
Nicolaou et al.
Scheme 1. Retrosynthetic Analysis of Cortistatin A (1)a
of developing new synthetic technologies and strategies and
applying them to the construction of analogues for biological
evaluation as potential anticancer agents. Of particular interest
to us were the development of cascade reactions7 for total
synthesis and the discovery of selective antiproliferative agents
for cancer chemotherapy. With these aims in mind we pondered
the cortistatin molecule as a synthetic target, starting with its
structural motifs and retrosynthetic analysis.
Retrosynthetic Analysis. Being the most potent member of
the family, cortistatin A (1) became our first target for synthesis.
Inspection of its structure revealed the unique abeo-9(10-19)-
androstane-type steroidal skeleton with substitutions on rings
A and E, a structural motif common to all cortistatins.
Retrosynthetically removing the dimethylamino and C-2 hy-
droxyl groups from ring A (a) led to hypothetical precursor
ketone 8 as shown in Scheme 1. Disconnecting the isoquinoline
moiety from the main framework (b) of the molecule at this
stage (or at some stage downstream) simplified the structure
further and revealed boronic ester 11 as a potential donor of
this group in a Suzuki-Miyaura8 coupling. At this stage a
cascade involving a 1,4-addition/aldol/dehydration sequence
(c,d) was envisioned (10 f 9 f 8). Imagining an acetylenic
unit as the bridge between rings A and D of intermediate 10
allowed a Sonogashira9 disconnection to reveal vinyl triflate
12 and terminal acetylenes 13a and 13b as its potential
precursors. These intermediates were then connected to enone
14,10 itself being traceable to simple monocyclic diketone 15
via the Hajos-Parrish ketone.11 The strategy derived from this
retrosynthetic analysis had the advantages of high convergency
and flexibility for analogue construction, as well as enantiose-
lectivity options. It was with this plan that we embarked on the
cortistatin program which included both methodology develop-
ment and chemical biology studies.
a Operations: (a) epoxide opening; (b) Suzuki-Miyaura coupling; (c)
aldol condensation; (d) 1,4-addition; (e) Sonogashira coupling; (f)
Hajos-Parrish ketone construction.
to the timing of the introduction of the isoquinoline and
dimethylamino moieties within the growing molecule dictated
a number of model studies prior to its finalization. We first set
out to test the feasibility of the 1,4-addition/aldol/dehydration
cascade to form the central core of cortistatin A containing the
oxa bridge and the dienone structural motif by targeting model
system 26 as shown in Scheme 2. The key substrate for this
cascade, bicyclic hydroxy enone enal 23, was conveniently
prepared in racemic form from cyclohexenone (16) through a
short sequence. Thus, reaction of 16 with formaldehyde in the
presence of 4-DMAP gave hydroxy enone 17 in 72% yield.
Silyl protection of the latter compound (TESCl, imid.) furnished
TES ether 18 (81% yield), which reacted with lithium TMS
acetylide to afford, upon global desilylation (TBAF, 72% yield
for the two steps), dihydroxy terminal acetylene 20 via
intermediate 19. Sonogoshira9 coupling [Pd(PPh3)4 cat., CuI cat.,
Et3N] of acetylene 20 with freshly prepared enol triflate 12 (1,3-
cyclohexadione, Tf2O, Et3N) followed by DMP oxidation
afforded acetylenic aldehyde enone 22 in 58% overall yield for
the two steps. Finally, selective hydrogenation of 22 with 10%
Pd/C under carefully controlled conditions (H2, MeOH/EtOAc
2:3, 23 °C) led to hydroxy enone enal 23 (61% yield), setting
the stage for the much anticipated cascade sequence. While
pleased with the remarkable selectivity of this hydrogenation
process, we were somewhat surprised by the resistance of the
product (23) toward spontaneous cyclization. The cascade
reaction of hydroxy enone enal 23 was, therefore, investigated
under basic and acidic conditions as summarized in Table 1.
Interestingly, we found that 23 undergoes the desired transfor-
mation to the targeted model system 26, both under basic (entries
1-3, Table 1) and acidic (entries 4 and 5, Table 1) conditions,
Model Studies. The rather daring nature of the designed
strategy toward cortistatin A and the uncertainties with regards
(5) For the first synthesis of cortistatin A, see: (a) Shenvi, R. A.; Guerrero,
C. A.; Shi, J.; Li, C.-C.; Baran, P. S. J. Am. Chem. Soc. 2008, 130,
7241–7243. For a total synthesis of cortistatin A, see: (b) Lee, H. M.;
Nieto-Oberhuber, C.; Shair, M. D. J. Am. Chem. Soc. 2008, 130,
16864–16866. For a formal synthesis of cortistatin A, see: (c)
Yamashita, S.; Kitajima, K.; Iso, K.; Hirama, M. Tet. Lett. 2009, 50,
3277–3279. For studies toward the synthesis of cortistatins, see: (d)
Yamashita, S.; Iso, K.; Hirama, M. Org. Lett. 2008, 10, 3413–3415.
(e) Simmons, E. M.; Hardin, A. R.; Guo, X.; Sarpong, R. Angew.
Chem., Int. Ed. 2008, 47, 6650–6653. (f) Dai, M.; Wang, Z.;
Danishefsky, S. J. Tetrahedron Lett. 2008, 49, 6613–6616. (g) Dai,
M.; Danishefsky, S. J. Tetrahedron Lett. 2008, 49, 6610–6612. (h)
Ku¨rti, B.; Corey, E. J. Org. Lett. 2008, 10, 5247–5250. (i) Kotoku,
N.; Sumii, Y.; Hayashi, T.; Kobayashi, M. Tetrahedron Lett. 2008,
49, 7078–7081. (j) Craft, D. T.; Gung, B. W. Tetrahedron Lett. 2008,
49, 5931–5934. (k) Dai, M.; Danishefsky, S. J. Heterocycles 2009,
77, 157–161. (l) Liu, J.; Gao, Y.; Che, C.; Wu, N.; Wang, D.; Li, C.;
Yang, Z. Chem. Commun. 2009, 662–664.
(6) Nicolaou, K. C.; Sun, Y.-P.; Peng, X.-S.; Polet, D.; Chen, D. Y.-K.
Angew. Chem., Int. Ed. 2008, 47, 7310–7313.
(7) For a recent review on cascade reactions in total synthesis, see: (a)
Nicolaou, K. C.; Montagnon, T.; Snyder, S. A. Chem. Commun. 2003,
5, 551–564. (b) Tietze, L. F.; Brasche, G.; Gericke, K. M. Domino
Reactions in Organic Synthesis; Wiley-VCH: Weinheim, 2006; p 617.
(c) Nicolaou, K. C.; Edmonds, D. J.; Bulger, P. G. Angew. Chem.,
Int. Ed. 2006, 45, 7134–7186.
(8) Kotha, S.; Lahiri, K.; Kashinath, D. Tetrahedron 2002, 58, 9633–
9695.
(9) Sonogashira, K. Handbook of Organopalladium Chemistry for Organic
Synthesis; John Wiley & Sons: New York, 2002; Vol. 1, pp 493-
529.
(10) (a) Rychnovsky, S. D.; Mickus, D. E. J. Org. Chem. 1992, 57, 2732–
2736. For the synthesis of ent-14, see: (b) Isaacs, R. C. A.; Di Grandi,
M. J.; Danishefsky, S. J. J. Org. Chem. 1993, 58, 3938–3941.
(11) Olton, Z.; Hajos, G.; Parrish, R. Org. Synth. 1990, 7, 363–368.
9
10588 J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009

Cortistatins A and J and Analogues
A R T I C L E S
Scheme 2. Synthesis of Model Dienone 26a
Scheme 3. Synthesis of Cortistatin A and J Model Compounds (30
and 33, respectively)a
a Reagents and conditions: (a) CH2O (37% aq., 5.0 equiv), 4-DMAP
(5.0 equiv), THF, 23 °C, 16 h, 72%; (b) TESCl (1.2 equiv), imidazole (2.0
equiv), DMF, 23 °C, 6 h, 81%; (c) TMSCtCLi (1.5 equiv, prepared from
n-BuLi and TMSCtCH), THF, 0 °C, 2 h; (d) TBAF (1.0 M in THF, 1.3
equiv), THF, 23 °C, 0.5 h, 72% for two steps; (e) Pd(PPh3)4 (0.05 equiv),
CuI (0.1 equiv), Et3N (3.0 equiv), 12 (1.5 equiv, from 1,3-cyclohexadione,
Tf2O and Et3N), DMF, 23 °C, 1 h, 82%; (f) DMP (1.5 equiv), NaHCO3
(6.0 equiv), CH2Cl2, 23 °C, 0.5 h, 71%; (g) Pd-C (10 wt %/wt, 0.38 equiv),
H2 (1 atm), MeOH/EtOAc (2:3), 23 °C, 25 min, 61%; (h) K2CO3 (1.2 equiv),
dioxane, 125 °C, 16 h, 50%.
a Reagents and conditions: (a) TMSOTf (1.2 equiv), Et3N (2.0 equiv),
-78 f 0 °C, 1.5 h; (b) Pd(OAc)2 (1.5 equiv), CH3CN, 23 °C, 15 h, 43%
for two steps; or IBX·MPO (0.4 M in DMSO, 6.0 equiv), DMSO, 23 °C,
6 h, 45% for two steps; (c) LiN(Me)Alloc (5.0 equiv, prepared from n-BuLi
and AllocNHMe), THF, -78 °C, 15 min, 70%; (d) KHMDS (0.5 M in
toluene, 3.0 equiv), PhNTf2 (2.0 equiv), THF, -78 f 0 °C, 0.5 h; (e)
PdCl2(PPh3)2 (0.05 equiv), n-Bu3SnH (2.0 equiv), AcOH (23 equiv), CH2Cl2,
23 °C, 1 h, 27% for two steps; (f) CH2O (37% aq., 6.0 equiv), NaCNBH3
(6.0 equiv), MeOH, 23 °C, 1 h, 56%; (g) KHMDS (0.5 M in toluene, 3.0
equiv), Davis oxaziridine (2.0 equiv), THF, -78 °C, 1 h, 60%; (h) DCC
(1.5 equiv), p-nitrobenzoic acid (1.5 equiv), 4-DMAP (0.2 equiv), CH2Cl2,
23 °C, 2 h, 71%; (i) Me4NBH(OAc)3 (8.0 equiv), AcOH/CH3CN, 1:1, 23
°C, 1 h, 25%.
Table 1. Cascade Cyclization of Hydroxy Enone-Enal 23 to
Tetracyclic Dienone 26 (Scheme 2)a
product
(yield)b
entry
conditions
1
2
3
4
5
10% KOH, MeOH, reflux, 12 h
31%
21%
50%
22%
17%
piperidine (0.1 equiv), toluene, reflux, 22 h
K2CO3 (1.2 equiV), dioxane, 125 °C, 16 h
PPTS (0.1 equiv), toluene, reflux, 14 h
p-TsOH (0.1 equiv), toluene, reflux, 14 h
either a Saegusa12 protocol [(i) TMSOTf, Et3N; (ii) Pd(OAc)2
(cat.), 43% yield] or oxidation of the intermediate silyl enol
ether with IBX · MPO13 (45% yield). Pleasantly, trienone 27
entered into a facile and chemoselective reaction with the lithio
derivative of methylamine allyloxycarbamate [generated from
AllocNHMe and n-BuLi] to afford keto carbamate 28 in 70%
yield and as a single diastereoisomer. The stereochemistry of
this product, however, was not easily discernible at this stage
and had to await further chemical transformations, which
culminated in the synthesis of model system 33 (for cortistatin
J) and 30 (for cortistatin A). To this end, ketone 28 was first
converted to its vinyl triflate (KHMDS, PhNTf2) and then to
its olefinic counterpart 32 [n-Bu3SnH, AcOH, PdCl2(PPh3)2
cat.]14 in 27% overall yield for the two steps. Finally, reductive
amination of the latter compound (CH2O aq., NaCNBH3)
furnished the cortistatin J model system 33 (56% yield), whose
stereochemical assignment as a 3-epi-cortistatin model system
became possible by NMR spectroscopic analysis (see Figure
2).
a Reactions were carried out on 0.17-0.26 mmol scale. b Yields refer
to chromatographically pure materials.
presumably through intermediates 24 and 25 (which, however,
were neither isolated nor detected). The optimum conditions
for this cyclization cascade involved heating precursor 23 in
dioxane at 125 °C in the presence of K2CO3 for 16 h, yielding
product 26 in 50% yield (entry 3, Table 1).
With the tetracyclic dienone structure 26 secured, we then
turned our attention to its modification into closer model systems
to cortistatins A (i.e., 30, Scheme 3) and J (i.e., 33, Scheme 3)
through functionalization of ring A as shown in Scheme 3. A
primary objective of the envisioned sequence was the stereo-
selective installment of the dimethylamino group, common to
both cortistatins A and J, into the growing molecule. Based on
manual molecular models, it was hypothesized that the trajectory
of a nucleophilic attack on the corresponding A-ring enone (i.e.,
27) would follow an antiperiplanar path to the bis-methylene
bridge of the tetrahydrofuran ring, thereby resulting in the
desired R-stereochemistry for the incoming nucleophile. To this
end, model system 26 was converted to trienone 27 through
Thus, in the reported proton NMR data for natural cortistatin
J (5),4a H2 displayed a clear doublet of 9.9 Hz solely due to
(12) Ito, Y.; Hirao, T.; Saegusa, T. J. Org. Chem. 1978, 43, 1011–1013.
(13) Nicolaou, K. C.; Gray, D. L. F.; Montagnon, T.; Harrison, S. T. Angew.
Chem., Int. Ed. 2002, 41, 996–1000.
(14) Dangles, O.; Guibe, F.; Balavoine, G.; Lavielle, S.; Marquet, A. J.
Org. Chem. 1987, 52, 4984–4993.
9
J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009 10589

A R T I C L E S
Nicolaou et al.
Figure 2. Key 1H NMR signals and NOSEY correlations for cortistatin J
(5) and model compound 33.
Figure 3. X-ray derived ORTEP drawing of 29a with thermal ellipsoids
shown at the 50% probability level.17
Scheme 4. Synthesis of Cortistatin A Model System 38a
JH1-H2 coupling, in line with a 90° H2-H3 dihedral angle. In
stark contrast, in the synthesized model system 33, the corre-
sponding H2 exhibited a dd splitting pattern with JH1-H2 ) 9.6
Hz and JH2-H3 ) 5.4 Hz. The H2-H3 dihedral angle predicted,
based on manual molecular modeling studies for compound 33,
was considerably smaller than 90° and thus supportive of the
observed JH2-H3 coupling constant. Furthermore, the notable
difference in the chemical shifts observed for H3 in model
system 33 (δ ) 2.88 ppm, CDCl3) and in natural cortistatin J
(δ ) 3.51 ppm, CDCl3) also added to our suspicion. Most
diagnostically, the signature H3-H6ab NOESY correlation that
supported the C3-R stereochemical assignment in the natural
product (5) was clearly absent in model system 33.
Despite the realization that the dimethylamino group within
intermediate 28 (Scheme 3) possessed the undesired stereo-
chemical configuration, we decided to explore a possible
pathway to a cortistatin A model system. We first attempted
the introduction of the obligatory hydroxyl group, an objective
that was nicely achieved by reaction of the potassium enolate
of ketone 28 (KHMDS) with Davis oxaziridine15 in 60% yield
and with complete stereocontrol, as shown in Scheme 3, to
afford the desired R-hydroxylated product 29 as a single
diastereoisomer. Subsequent 1,2-directed reduction of the car-
bonyl group [Me4NBH(OAc)3]16 then produced diol 30 in 25%
(unoptimized) yield. The trans relationship of the two hydroxyl
groups was supported by the observed coupling constant of the
a Reagents and conditions: (a) t-BuOOH (5.5 M in decane, 4.0 equiv),
DBU (3.0 equiv), CH2Cl2, 0 f 23 °C, 5 h, 72%; (b) NaBH4 (1.0 equiv),
CeCl3 ·7H2O (3.0 equiv), MeOH, 0 °C, 10 min, 35: 40% and 36: 40%; (c)
DMP (2.0 equiv), NaHCO3 (5.0 equiv), CH2Cl2, 23 °C, 2 h, 100%; (d)
Me2NH (2.0 M in THF, as solvent), Ti(Oi-Pr)4 (5.0 equiv), 80 °C, 5 h, 38:
44% and 41: 35%; (e) TBSOTf (2.0 equiv), 2,6-lut. (3.0 equiv), CH2Cl2, 0
°C, 1 h, 80%; (f) Me2NH (2.0 M in THF, as solvent), Ti(Oi-Pr)4 (5.0 equiv),
80 °C, 50 h, 39: 65%, 40: 16%; (g) TBAF (1.0 M solution in THF, 5.0
equiv), THF, 23 °C, 5 h, 80%; (h) TBAF (1.0 M solution in THF, 5.0
equiv), THF, 65 °C, 3 h, 85%.
1
relevant protons (J1,2 ) 5.4 Hz) in the H NMR spectrum of
30. Furthermore, and delightfully, the para-nitrobenzoate de-
rivative (29a, Scheme 3) of cortistatin A model system 29
crystallized beautifully from CH2Cl2/hexane (29a, dec 145 °C),
allowing its X-ray crystallographic analysis (see ORTEP draw-
ing, Figure 3), which confirmed all the stereochemical assign-
ments made previously on the basis of NMR spectroscopy.
While the undesired stereochemical outcome (ꢀ) of the
conjugate addition of the nitrogen nucleophile to trienone 27
was disappointing, it suggested an alternative tactic for the
introduction of the A-ring functionality of cortistatin A. Specif-
ically, it was reasoned that should the t-BuOOH-DBU18
epoxidation of 27 proceed in the same way, then the resulting
ꢀ-epoxide would be a potentially useful precursor for our desired
objective. Indeed, and as shown in Scheme 4, the t-BuOOH-
DBU epoxidation of 27 proved to be both chemo- and
stereoselective, delivering epoxide 34 in 72% yield and as a
single diastereoisomer. An X-ray crystallographic analysis of
34 (mp 186-187 °C, CH2Cl2/hexane) confirmed its stereo-
chemical assignment (see ORTEP drawing, Figure 4). Having
failed to obtain satisfactory results with Me4NBH(OAc)3,16
Zn(BH4)2, and DIBAL-H, we resorted to the use of Luche19
conditions (NaBH4-CeCl3) for the reduction of epoxy ketone
34, which furnished diastereomeric epoxy alcohols 35 and 36
in 80% yield and ca. 1:1 ratio. The two isomers were
chromatographically separated, and their stereochemistry was
assigned by NMR spectroscopy. Indeed, an X-ray crystal-
lographic analysis of 36 (mp 175-176 °C, CH2Cl2/hexane)
(15) Davis, F. A.; Chen, B. C. Chem. ReV. 1992, 92, 919–934.
(16) Evans, D. A.; Chapman, K. T.; Carreira, E. M. J. Am. Chem. Soc.
1988, 110, 3560–3578.
(17) CCDC-710300 contains the supplementary crystallographic data for
compound 29a. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif.
(19) (a) Molander, G. A. Chem. ReV. 1992, 92, 29–68. (b) Luche, J. L.
J. Am. Chem. Soc. 1978, 100, 2226–2227.
(18) Yadav, V. K.; Kapoor, K. K. Tetrahedron 1995, 51, 8573–8584.
9
10590 J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009

Cortistatins A and J and Analogues
A R T I C L E S
Scheme 5. Synthesis of Cortistatin J Model System 45a
a Reagents and conditions: (a) Ti(Oi-Pr)4 (5.0 equiv), Me2NH (2.0 M
solution in THF, as solvent), 80 °C, 1 h, 70%; (b) thiocarbonyl diimidazole
(43) (1.5 equiv), toluene, 110 °C, 12 h, 81%; (c) P(OEt)3 (as solvent), 160
°C, 24 h, 42% (50% recovered starting material).
achieved through the Corey-Winter protocol22 [(i) thiocarbonyl
diimidazole, 81% yield; (ii) P(OEt)3, 42% yield (50% recovered
starting material, 44)] to afford cortistatin J model system 45
via intermediate thiocarbonate 44. Cortistatin J model system
45 exhibited excellent agreement with the naturally occurring
cortistatin J with regards to its 1H NMR spectral data spanning
rings A-D.4a
Figure 4. X-ray derived ORTEP drawings of compounds 34 and 36 with
thermal ellipsoids shown at the 50% probability level.21
confirmed its stereochemistry (see ORTEP drawing, Figure 4).
The undesired ꢀ-OH diastereoisomer (35) was recycled through
oxidation (DMP, 100% yield)-reduction as shown in Scheme
4. Finally, introduction of the dimethylamino group into the
molecule (36) was accomplished through a titanium-assisted20
epoxide opening [Ti(Oi-Pr)4, Me2NH] leading to an isomeric
mixture of dihydroxy dimethylamino compounds 38 (44% yield)
and 41 (35% yield), which were conveniently separated by
preparative thin-layer chromatography (silica). This ratio was
improved in favor of the desired isomer 38 by placing a bulky
protecting group (i.e., TBS) on the hydroxyl group of substrate
36 (TBSOTf, 2,6-lut., 80% yield) to afford intermediate silyl
ether 37, which upon reaction with Ti(Oi-Pr)4-Me2NH furnished
the two regioisomers 39 and 40 in ca. 4:1 ratio and 81%
combined yield. Chromatographic separation of these isomers
(preparative TLC, silica) followed by desilylation with TBAF
led to cortistatin A model systems 38 (80% yield) and 41 (85%
yield), respectively. Pleasingly, the NMR spectra of isomer 38
exhibited encouraging resemblance to those of cortistatin A.
Although hydroxy epoxide 35 possessed the wrong stereo-
chemistry for cortistatin A, it did provide a productive entry
into cortistatin J model system 45 as outlined in Scheme 5. Thus,
epoxide opening of compound 35 under the previously estab-
lished conditions [Ti(Oi-Pr)4-Me2NH] afforded dimethylamine
diol 42 in 70% yield as a single regioisomer. The exclusive
formation of regioisomer 42 is most likely a result of the Lewis
acidic titanium reagent concomitantly coordinating both the
epoxide and alcohol oxygens from the same face of epoxy
alcohol 35, thereby leading to steric bias that favors the exclusive
nucleophilic attack at C3. With the C3 stereocenter secured,
the introduction of the A-ring olefinic bond was subsequently
Initial Approach to Cortistatin A Involving Early-Stage
Installation of Isoquinoline Moiety. With the initial cortistatin
A model studies successfully completed, we then proceeded to
apply the developed synthetic technologies to the total synthesis
of cortistatin A (1). We chose first to test the approach involving
early stage installation of the isoquinoline moiety that would
allow the key cascade sequence as a subsequent event. Such an
approach would have the advantage of minimizing protecting
group manipulations and oxidation state adjustments in the late
stages of the synthesis. Toward this end, we first investigated
the route involving a Hajos-Parrish ketone synthesis11 and
starting from cyclopentadione 15, as shown in Scheme 6. Thus,
15 was converted to bicyclic enone 14 through modification of
published procedures.10b The olefinic bond of enone 14 was
dihydroxylated with OsO4 (cat.)/NMO and the resulting 1,2-
diol (46) was protected as its acetonide derivative [Me2C(OMe)2,
p-TsOH] to afford compound 47 in 64% yield. Vinyl triflate
formation (NaHMDS-PhNTf2) followed by palladium-catalyzed
carboxymethylation in DMF/MeOH [Pd(PPh3)4 cat., CO, 72%
yield for the two steps] then converted ketone 47 to R,ꢀ-
unsaturated methyl ester 48. In preparation for the attachment
of the isoquinoline moiety, compound 48 was desilylated
(TBAF, 87% yield), and the resulting alcohol (49) was oxidized
(DMP, 81% yield) to afford cyclopentanone 50.
Installment of the Isoquinoline System. With the stage now
set for the next stage of the synthesis, ketone 50 was converted
to its vinyl triflate (NaHMDS, PhNTf2), which entered the
intended Suzuki-Miyaura8 reaction with boronic ester 1123 in
the presence of Pd(PPh3)4 cat. and K2CO3 to afford isoquinoline
system 51 in 62% overall yield. For the purposes of cortistatin
A (and J-K), a stereoselective reduction of the cyclopentene
double bond was required. To this end, a number of conditions
were investigated, and as shown in Table 2, a satisfactory
solution to this problem was found. Thus, catalytic hydrogena-
(20) (a) Chong, J. M.; Sharpless, K. B. J. Org. Chem. 1985, 50, 1560–
1563. (b) Alvarez, E.; Nunez, M. T.; Martin, V. S. J. Org. Chem.
1990, 55, 3429–3431.
(21) CCDC-710299 and CCDC-710298 contain the supplementary crystal-
lographic data for compounds 34 and 36, respectively. These data can
be obtained free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
(23) Prepared from 7-bromoisoquinoline via: [Pd(dppf)Cl2], KOAc, bispina-
colato diboron, DMSO, 80 °C, 50%. For preparation of 7-bromoiso-
quinoline, see: Miller, B. R.; Frincke, J. M. J. Org. Chem. 1980, 45,
5312–5315.
(22) Corey, E. J.; Carey, F. A.; Winter, R. A. E. J. Am. Chem. Soc. 1965,
87, 934–935.
9
J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009 10591

A R T I C L E S
Nicolaou et al.
Scheme 6. Synthesis of Terminal Acetylene 13aa
Table 2. Reduction of Cyclopentenyl Isoquinoline 51 to
Cyclopentyl Isoquinoline 52 (Scheme 6)a
product
(yield)b
entry
conditions
1
2
3
Wilkinson’s catalyst (0.3 equiv), EtOH, 23 °C, 12 h
no reaction
Cy2BH (2.5 equiv), AcOH (20 equiv), THF, 12 °C, 10 h 14%
BH3 · SMe2 (10 M, 1.5 equiv), THF, 23 °C, 5 h;
20%
then AcOH (20 equiv)
4
5
PdsC (10 wt %/wt, 0.2 equiV), MeOH, 23 °C, 4 h
KOOCNdNCOOK (3.0 equiv), THF/H2O (4:1),
23 °C, 6 h; then AcOH (20 equiv)
70%
25%
a Reactions were carried out on 0.094-0.161 mmol scale under 1 atm
of hydrogen. b Yields refer to chromatographically pure materials.
Figure 5. X-ray derived ORTEP drawing of 52 with thermal ellipsoids
shown at the 50% probability level.26
52 led to alcohol 53 (71% yield), which was oxidized to
aldehyde 54 through the action of DMP (80% yield). Treatment
of the latter compound with 1,3-propanedithiol and BF3 ·OEt2
resulted in dithiane formation and concomitant acetonide
collapse to unveil the molecule’s 1,2-diol system, leading to
dihydroxy dithiane 55 (41% yield). Finally, oxidation of 55
according to the Parikh-Doering24 protocol (SO3 ·py, 70%
yield), followed by Ohira-Bestmann25 reaction (57, p-TsN3,
K2CO3) of the resulting hydroxy aldehyde (56), furnished the
targeted acetylenic compound 13a, albeit in a modest 28% yield.
The low yielding conversion of hydroxy aldehyde 56 to
terminal acetylene 13a prompted us to investigate the reaction
mixture further. Indeed, a second product was discovered (22%
yield) whose structure was determined to be that of enone 64
(Scheme 7). This intriguing observation can be explained by
the mechanism presented in Scheme 7. Thus, while in the normal
mode of the Ohira-Bestmann reaction (57 f 58 f 59), the
initially formed hydroxy phosphonate (59 + 56 f 60) cyclizes
rapidly to form the transient intermediate betaine 61, which
undergoes spontaneous collapse to generate the terminal acety-
lene 13a (path a), in this instance, intermediate 60 is presumed
to rearrange through proton transfer to isomeric alkoxy phos-
phonate 62 (path b), which could collapse as shown (62 f 63
f 64) to afford the observed truncated molecule (64).
a Reagents and conditions: (a) OsO4 (2.5 wt %/wt in t-BuOH, 0.02 equiv),
NMO (2.5 equiv), acetone/H2O (1:1), 23 °C, 16 h, 73%; (b) Me2C(OMe)2
(5.0 equiv), p-TsOH (0.04 equiv), acetone, 23 °C, 1 h, 87%; (c) NaHMDS
(1.0 M in THF, 1.2 equiv), PhNTf2 (1.1 equiv), THF, 0 °C, 2 h; (d) Pd(PPh3)4
(0.05 equiv), MeOH (30 equiv), CO, DMF, 70 °C, 4 h, 72% for the two
steps; (e) TBAF (1.0 M in THF, 1.2 equiv), THF, 23 °C, 16 h, 87%; (f)
DMP (2.0 equiv), NaHCO3 (5.0 equiv), CH2Cl2, 23 °C, 3 h, 81%; (g)
NaHMDS (1.0 M in THF, 1.5 equiv), PhNTf2 (1.2 equiv), THF, -78 f 0
°C, 2 h; (h) Pd(PPh3)4 (0.1 equiv), 11 (1.3 equiv), K2CO3 (3.0 equiv), THF,
80 °C, 3 h, 62% for two steps; (i) Pd-C (10 wt %/wt, 0.5 equiv), H2,
MeOH, 23 °C, 4 h, 70%; (j) DIBAL-H (1.0 M in toluene, 3.0 equiv), toluene,
-78 °C, 3 h, 71%; (k) DMP (1.5 equiv), NaHCO3 (5.0 equiv), CH2Cl2, 23
°C, 0.5 h, 80%; (l) HS(CH2)3SH (6.0 equiv), BF3 ·OEt2 (10.0 equiv), CH2Cl2,
-40 °C, 3 h, 41%; (m) SO3 ·py (4.0 equiv), CH2Cl2/DMSO (2:1), 23 °C,
1 h, 70%; (n) p-TsN3 (1.5 equiv), dimethyl-2-oxopropylphosphonate (57)
(1.5 equiv), K2CO3 (4.0 equiv), MeOH/CH3CN/THF (1:2:1), 23 °C, 12 h,
13a: 28%, 64: 22%.
tion of compound 51 with 10% Pd/C in MeOH at ambient
temperature selectively delivered isoquinoline derivative 52 in
70% yield and as a single diastereoisomer. That this crystalline
compound (mp ) 204-206 °C, CH2Cl2/hexanes) possessed the
desired ꢀ-stereochemistry was determined through X-ray crys-
tallographic analysis (see ORTEP drawing, Figure 5). Although
this stereochemical outcome was expected on the basis of
manual molecular modeling, its experimental verification was
both gratifying and comforting at this stage. Prior to installing
the planned acetylenic moiety, the ester group of the growing
molecule was adjusted to the aldehyde oxidation state and
protected as a dithiane system. Thus, DIBAL-H reduction of
With the final step of the synthesis of acetylenic compound
13a pending improvement, and with sufficient quantities of this
(24) Parikh, J. R.; Doering, W. V. E. J. Am. Chem. Soc. 1967, 89, 5505–
5507.
(25) (a) Ohira, S. Synth. Commun. 1989, 19, 561–564. (b) Mu¨ller, S.;
Liepold, B.; Roth, G. J.; Bestmann, H. J. Synlett 1996, 521–522. (c)
Roth, G. J.; Liepold, B.; Mu¨ller, S. G.; Bestmann, H. J. Synthesis 2004,
59–62.
(26) CCDC-684134 contains the supplementary crystallographic data for
compound 52. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif.
9
10592 J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009

Cortistatins A and J and Analogues
A R T I C L E S
Scheme 7. Postulated Mechanism for the Formation of Acetylenic
Compound 13a and Byproduct 64
Scheme 8. Attempted Synthesis of Dienone 8a
a Reagents and conditions: (a) Pd(PPh3)4 (0.1 equiv), CuI (0.1 equiv),
Et3N (3.0 equiv), 12 (1.5 equiv), DMF, 23 °C, 1.5 h, 61%; (b) IBX (6.0
equiv), DMSO, 23 °C, 9 h, 70%.
Scheme 9. Synthesis of Acetylenic Compound 13ba
intermediate in hand, we proceeded to explore its possible
conversion to the core structure of cortistatin A (1) through the
proposed cascade depicted in Scheme 1. Our attempted sequence
to reach cortistatin structure 8 is shown in Scheme 8. Thus,
Sonogoshira9 coupling of 13a with freshly prepared vinyl triflate
12 [Pd(PPh3)4 cat., CuI cat., Et3N], followed by removal of the
dithiane protecting group from the resulting coupling product
65 (IBX),27 led to enal enynone 66 in 43% overall yield for the
two steps. Reduction of the acetylenic bond within the latter
compound was then investigated as a necessary step prior to
the proposed cyclization cascade. Unfortunately, however, only
trace amounts of the desired precursor enal enone 10a were
obtained under several tried conditions, notably those that proved
successful in the synthesis of model system 26 (22 f 23,
Scheme 2) previously discussed. From analysis of the reaction
mixtures of the various reductions, it was surmised that the
difficulties arose from the presence of the isoquinoline moiety
within the employed substrate. For these reasons, and due to
the previous low-yielding steps already mentioned, this approach
to cortistatin A was abandoned in favor of one in which the
isoquinoline installment was reserved for a later stage and after
the crucial cyclization cascade.
a Reagents and conditions: (a) DIBAL-H (1.0 M in toluene, 3.0 equiv),
toluene, -78 °C, 3 h, 79%; (b) DMP (1.5 equiv), NaHCO3 (4.6 equiv),
CH2Cl2, 23 °C, 0.5 h, 86%; (c) HS(CH2)3SH (3.0 equiv), BF3 ·OEt2 (3.5
equiv), CH2Cl2, -78 °C, 1.5 h, 70%; (d) SO3 ·py (3.0 equiv), Et3N (5.0
equiv), CH2Cl2/DMSO (4:1), 23 °C, 1.5 h, 72%; (e) p-TsN3 (1.5 equiv),
dimethyl-2-oxopropylphosphonate (57) (1.5 equiv), K2CO3 (3.5 equiv),
CH3CN, 23 °C, 2 h; then aldehyde 70, THF/CH3CN/MeOH (1:2:1); 23 °C,
16 h, 13b: 45% (10% recovered starting material), 64a: 20%.
cortistatin A, we retreated back to intermediate 48 (see Scheme
6) with the intention of reaching cortistatin A through inter-
mediate compounds 13b (Scheme 9) and 74 (Scheme 10). Such
an approach would ensure the construction of the mainframe
of the molecule prior to the attachment of the isoquinoline group
and leave only a few steps in which to deal with its presence.
The preparation of acetylenic compound 13b from intermedi-
ate 48 (Scheme 9) involved an initial reduction (DIBAL-H, 79%
yield)-oxidation (DMP, 86% yield) protocol to afford aldehyde
68 via alcohol 67. Subsequent exposure of 68 to 1,3-pro-
panedithiol in the presence of BF3 ·OEt2 resulted in dithiane
formation with concomitant acetonide removal to furnish
Revised Strategy and Total Synthesis of Coristatin A.
Recognizing the problematic nature of the isoquinoline moiety
in several of the reaction steps of our first attempt toward
(27) Nicolaou, K. C.; Mathison, C. J. N.; Montagnon, T. J. Am. Chem.
Soc. 2004, 126, 5192–5201.
9
J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009 10593

A R T I C L E S
Nicolaou et al.
Scheme 10. Construction of Dienone 74a
Table 3. Hydrogenation of Alkyne 72 To Form Enone-Enal 10b
(Scheme 10)a
product(s)
(yield)b
entry
conditions
1
2
3
4
5
6
Wilkinson’s catalyst (0.2 equiv),
EtOH, 23 °C, 12 h
75(30%) + 72(50%)
complex mixture
complex mixture
no reaction
Pd-C (10 wt %/wt, 0.04 equiv),
EtOAc/MeOH (4:1), 23 °C, 15 min
Pd(OH)2 (10 wt %/wt, 0.03 equiv),
EtOAc/MeOH (4:1), 23 °C, 15 min
Pd-BaSO4 (5 wt %/wt, 0.05 equiv),
quinoline (1 drop), MeOH, 23 °C, 12 h
Pd-BaSO4 (5 wt %/wt, 0.05 equiv),
MeOH, 23 °C, 6 h
75(50%)
Pd-BaSO4 (5 wt %/wt, 0.05 equiv),
MeOH/THF (anhydrous)
10b(20%) + 75(30%)
(5:1), 23 °C, 1 h
7
8
9
Pd-BaSO4 (5 wt %/wt, 0.1 equiv),
MeOH/THF (3:1), 23 °C, 1 h
Pd-BaSO4 (5 wt %/wt, 0.2 equiv),
MeOH/THF (3:2), 23 °C, 45 min
Pd-BaSO4 (5 wt %/wt, 0.24 equiV),
MeOH/THF (1:1), 23 °C, 0.5 h
10b(30%) + 75(30%)
10b(55%)
10b(64%)
a Reactions were carried out on 0.014-0.116 mmol scale under 1 atm
of hydrogen. b Yields refer to chromatographically pure materials.
1, 5-7, Table 3), and resubjecting the isolated material to further
hydrogenation under more forceful conditions led to unattractive
mixtures containing over-reduced products.
Substrate 10b performed well under the intended conjugate
addition/aldol/dehydration cascade as previously developed
(K2CO3, dioxane, reflux) to afford the targeted dienone 74 in
52% yield, presumably through the transient and non detectable
intermediates 9b and 73, as shown in Scheme 10.
a Reagents and conditions: (a) Pd(PPh3)4 (0.1 equiv), CuI (0.1 equiv),
Et3N (3.0 equiv), 12 (1.4 equiv, freshly prepared from 1,3-cyclohexadione,
Tf2O and Et3N), DMF, 23 °C, 1 h, 85%; (b) IBX (4.0 equiv), DMSO, 0 f
23 °C, 4 h, 81%; (c) Pd-BaSO4 (5 wt %/wt, 0.24 equiv), H2, MeOH/THF
(1:1), 23 °C, 0.5 h, 64%; (d) K2CO3 (1.2 equiv), dioxane, 125 °C, 12 h,
52%.
With the construction of the cortistatin dienone 74 completed,
we then turned our attention to the final stages of the synthesis
of cortistatin A (1) that required introduction of the isoquinoline
moiety and functionalization of ring A. The successful drive to
the target molecule (1) is summarized in Scheme 11. Installation
of the isoquinoline moiety became our first priority, an objective
that required temporary protection of the carbonyl group of
dienone 74. After considerable experimentation it was found
that ketal 76 could be generated from 74 through the action of
TMSOCH2CH2OTMS in the presence of TMSOTf in CH2Cl2
at -50 f -10 °C together with small amounts (ca. 10%) of
the next desired intermediate, hydroxy ketal 77. The latter
compound was formed from 76 by treatment of the crude
reaction mixture with TBAF in 63% overall yield for the two
steps. Oxidation of 77 with SO3 ·py furnished ketone 78 (80%
yield) which was converted to its vinyl triflate (KHMDS,
Comins reagent)28 and thence to vinyl isoquinoline 79 through
a Suzuki-Miyaura coupling8 reaction with boronic ester deriva-
tive 11 [Pd(PPh3)4 cat., K2CO3] in 60% overall yield. Subsequent
removal of the protecting group from 79 with p-TsOH gave
dienone vinyl isoquinoline 80 (88% yield), whose desired stereo-
and chemoselective reduction to dienone isoquinoline 8 was
accomplished by catalytic hydrogenation using 10% Pd/C as
the catalyst and MeOH as the solvent (50% yield, 30%
recovered starting material). The stereochemistry of 8 was
tentatively assigned based on the result obtained with the
truncated isoquinoline 52 discussed above (Scheme 6) and
ultimately confirmed by the successful synthesis of 1 (see
below). The next step required introduction of unsaturation in
ring A, an operation in which a stark contrast between the
dihydroxy dithiane 69 in 70% yield. Oxidation of 69 with
SO3 ·py led to hydroxy aldehyde 70 in 72% yield, setting the
stage for forging the required acetylenic moiety. Despite
attempts to obtain the targeted acetylene through other means,
the Ohira-Bestmann25 method was the most efficient. We were,
however, able to improve the yield of this reaction to 45% (10%
recovered starting material) after optimization of the reaction
conditions [1.5 equiv of 58 generated in situ, 3.5 equiv of
K2CO3, THF/CH3CN/MeOH (1:2:1), 23 °C, 16 h], accompanied
with enone byproduct 64a in 20% yield.
Acetylenic substrate 13b entered the Sonogashira9 coupling
reaction [Pd(PPh3)4 cat., CuI cat., Et3N] with vinyl triflate 12
with significantly improved efficiency (over that in the first
generation study) to afford coupling product 71 in 85% yield.
The IBX27-induced dithiane removal from the latter compound
also proved superior to that involving the isoquinoline containing
compound (65 f 66, Scheme 8), affording enal enone 72 in
81% yield. The next step involving selective hydrogenation of
72 to the desired cascade precursor 10b was also expected to
proceed more efficiently than the corresponding transformation
of the isoquinoline compound (66 f 10a, Scheme 8). Indeed,
exploration of various protocols (Table 3) led to the identifica-
tion of optimum conditions for the preparation of 10b [entry 9,
Table 3, 5% Pd-BaSO4, MeOH/THF (1:1), 23 °C, 0.5 h, 64%
yield]. Interestingly, partially hydrogenated and cyclized com-
pound 75 (stereochemistry supported by NOE studies) resisted
further hydrogenation under certain reaction conditions (entries
(28) Comins, D. L.; Dehghani, A. Tetrahedron Lett. 1992, 33, 6299–6302.
9
10594 J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009

Cortistatins A and J and Analogues
A R T I C L E S
Scheme 11. Completion of the Total Synthesis of Cortistatin A (1)a
Scheme 12. Completion of the Total Synthesis of Cortistatin J (5)a
a Reagents and conditions: (a) Me2NH (2.0 M in THF, as solvent), Ti(Oi-
Pr)4 (5.0 equiv), 80 °C, 2 h, 60%; (b) thiocarbonyl diimidazole (43) (2.0
equiv), toluene, 110 °C, 12 h, 81%; (c) P(OEt)3 (as solvent), 160 °C, 24 h,
40% (45% recovered starting material).
of IBX ·MPO to this enol ether furnished, upon chromatographic
purification, the desired trienone 81 in 46% overall yield from
8. Similar results were observed with other isoquinoline
intermediates. Chemo- and stereocontrolled epoxidation of
trienone 81 with t-BuOOH-DBU afforded epoxy ketone 82 in
70% yield. Luche19 reduction (NaBH4-CeCl3) of the latter
compound furnished isomeric hydroxy compounds 83 and 84
in 80% combined yield and ca. 1:1 ratio. Chromatographic
separation of the two isomers allowed recycling of the undesired
isomer (83) through an oxidation (DMP, 100% yield)-reduction
protocol and conversion of the desired isomer 84 to cortistatin
A (1) and its regioisomer 85. This final step of the total synthesis
was brought about through the action of Me2NH-Ti(Oi-Pr)4
and delivered 1 in 45% yield and 85 in 36% yield, after
chromatographic separation. All physical properties (1H and 13
C
NMR spectra, MS, and [R]D25) of synthetic cortistatin A (1)
were in accordance to those reported for the natural substance.3,5a
Total Synthesis of Cortistatin J. Having completed the total
synthesis of cortistatin A from hydroxy epoxide 84, we then
proceeded to convert the latter intermediate to cortistatin J (5)
employing our developed technologies as discussed above for
the construction of model system 45 from model hydroxy
epoxide 35 (see Scheme 5). Thus, exposure of 83 (Scheme 12)
to Me2NH-Ti(Oi-Pr)4 in refluxing THF led to dimethylamino
diol 86 in 60% yield. Reaction of the latter with thiocarbonyl
diimidazole (toluene, 110 °C) gave thiocarbonate derivative 87
(81% yield), whose treatment with P(OEt)3 (as solvent, 160 °C)
led to cortistatin J (5) in 40% yield (plus 45% recovered starting
material). Synthetic 5 exhibited identical physical properties to
those reported for the natural substance.4a
Biological Evaluation of Cortistatins A and J and Analogues.
The inhibitory activities of the synthesized compounds were
tested against human umbilical vein endothelial cells (HUVECs),
and the active candidates were further tested against a panel of
cancer cells, including breast (MCF-7), CNS (SF268), lung
(NCI-H460), ovarian (1A9), Taxol-resistant ovarian (PTX22),29
and epothilone-resistant ovarian (A8)30 cells using doxorubicin
a Reagents and conditions: (a) TMSO(CH2)2OTMS (5.0 equiv), TMSOTf
(1.5 equiv), CH2Cl2, -50 f -10 °C, 4 h; (b) TBAF (1.0 M in THF, 3.5
equiv), THF, 23 °C, 2 h, 63% for the two steps; (c) SO3 ·py (6.0 equiv),
Et3N (10.0 equiv), CH2Cl2/DMSO (3:1), 23 °C, 3 h, 80%; (d) KHMDS
(0.5 M in toluene, 2.0 equiv), Comins reagent (2.0 equiv), THF, -78 °C,
1 h; (e) 11 (3.0 equiv), Pd(PPh3)4 (0.1 equiv), K2CO3 (3.0 equiv), THF, 80
°C, 3 h, 60% for two steps; (f) p-TsOH (1.5 equiv), acetone/H2O (10:1),
23 °C, 1 h, 88%; (g) Pd-C (10 wt %/wt, 0.3 equiv), H2, MeOH, 23 °C,
1 h, 50% (30% recovered starting material); (h) TMSOTf (14 equiv), Et3N
(30 equiv), THF, -78 f 0 °C, 1.5 h; (i) IBX·MPO (0.4 M in DMSO, 6.0
equiv), DMSO, 23 °C, 6 h, 46% for two steps; (j) t-BuOOH (6.0 equiv, 5.5
M in decane), DBU (3.0 equiv), CH2Cl2, 0 f 23 °C, 5 h, 70%; (k) NaBH4
(1.0 equiv), CeCl3 ·7H2O (4.0 equiv), MeOH, 0 °C, 10 min, 80% (ca. 1:1
mixture of diastereoisomers); (l) DMP (5.0 equiv), NaHCO3 (10.0 equiv),
CH2Cl2, 23 °C, 2 h, 100%; (m) Me2NH (2.0 M in THF, as solvent), Ti(Oi-
Pr)4 (5.0 equiv), 80 °C, 5 h, 1: 45%, 85: 36%.
Saegusa12 and the IBX·MPO13 oxidation methods was ob-
served. Thus, TMS enol ether formation from 8 followed by
Pd(OAc)2 addition resulted in significant complexation of
palladium to the isoquinoline moiety (as deduced by baseline
material on TLC) and no productive reaction, whereas addition
(29) Giannakakou, P.; Sackett, D. L.; Kang, Y.-K.; Zhan, Z.; Buters,
J. T. M.; Fojo, T.; Poruchynsky, M. S. J. Biol. Chem. 1997, 272,
17118–17125.
9
J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009 10595

A R T I C L E S
Nicolaou et al.
Table 4. Antiproliferative Effects of Cortistatin A (1), Cortistatin J (5), and Selected Synthetic Analogues (GI50 Values in µM)a
cell line/
compound
MCF-7
(SI)
NCI-H460
(SI)
SF268
(SI)
IA9
(SI)
PTX22
(SI)
A8
(SI)
entry
HUVEC
1
synthetic
cortistatin A
(1)
0.002 ( 0.001
>10
(>5000.00)
7.786 ( 1.692
(3893.00)
4.429 ( 0.664
(2214.50)
6.518 ( 0.756
(3259.00)
6.543 ( 0.165
(3271.50)
5.523 ( 0.045
(2761.50)
2
3
synthetic
cortistatin J
(5)
0.070 ( 0.013
0.007 ( 0.001
27.439
(391.99)
6.197 ( 0.122
4.340 ( 0.161
43.22
(617.43)
7.055 ( 0.013
8.538 ( 0.588
(88.53)
(62.00)
(100.79)
(121.97)
8
7.893 ( 0.991
(1127.57)
4.693 ( 0.826
6.326 ( 0.698
5.236 ( 0.039
4.905 ( 0.488
3.714 ( 0.072
(670.43)
(903.71)
(748.00)
(700.71)
(530.57)
4
5
6
10b
13a
13b
>10
>10
-
-
-
-
-
-
-
-
-
-
-
-
5.745 ( 0.173
5.085 ( 0.170
5.26 ( 0.358
6.681 ( 0.729
8.400 ( 1.302
>10
7.379 ( 0.698
(0.89)
-
(0.92)
-
(1.17)
-
(1.46)
-
(>1.74)
(1.28)
-
7
8
26
27
>10
6.373 ( 0.847
-
6.505 ( 0.211
5.700 ( 0.173
5.142 ( 0.337
6.251 ( 0.071
6.009 ( 0.401
5.281 ( 0.305
(1.02)
(0.89)
(0.81)
(0.98)
(0.94)
(0.83)
9
28
5.753 ( 0.252
6.487 ( 0.235
5.448 ( 0.278
5.558 ( 0.163
(0.97)
6.338 ( 0.318
6.050 ( 0.029
5.499 ( 0.038
(1.13)
(0.95)
(1.10)
(1.05)
(0.96)
10
11
12
13
14
15
16
17
18
19
30
32
33
35
36
38
41
44
45
53
>10
>10
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
>10
>10
>10
>10
>10
>10
>10
0.253 ( 0.099
>10
>10
>10
(39.53)
>10
>10
9.861 ( 0.565
(>39.53)
6.032 ( 0.324
(16.90)
-
(>39.53)
4.744 ( 0.453
(13.29)
(>39.53)
5.851 ( 0.349
(16.39)
(>39.53)
5.845 ( 0.057
(16.37)
(38.99)
20
54
0.357 ( 0.047
5.632 ( 1.152
6.008 ( 0.281
(16.83)
(15.78)
21
22
55
56
>10
8.040 ( 1.756
-
-
-
-
-
>10
(>1.24)
-
>10
>10
>10
6.456 ( 0.110
>10
(>1.24)
(>1.24)
(>1.24)
(0.80)
-
(>1.24)
23
24
64
65
>10
5.804 ( 1.059
-
-
-
-
5.652 ( 0.111
5.218 ( 0.646
5.322 ( 0.471
5.971 ( 0.292
5.660 ( 0.071
5.609 ( 0.033
(0.97)
(0.90)
(0.92)
(1.03)
(0.98)
(0.97)
25
26
27
66
71
72
4.542 ( 0.853
3.995 ( 1.030
0.544 ( 0.052
5.060 ( 0.081
(1.11)
5.225 ( 0.132
(1.15)
5.161 ( 0.199
(1.14)
3.719 ( 0.905
(0.82)
4.477 ( 1.054
(0.99)
4.057 ( 0.643
(0.89)
0.745 ( 0.110
4.64 ( 0.429
5.452 ( 0.260
5.121 ( 0.102
5.622 ( 0.250
3.518 ( 0.887
(0.19)
(1.16)
(1.37)
(1.28)
(1.41)
(0.88)
0.532 ( 0.024
(0.98)
0.729 ( 0.180
(1.34)
0.524 ( 0.030
(0.96)
0.519 ( 0.013
(0.95)
0.512 ( 0.005
(0.94)
0.578 ( 0.004
(1.06)
28
29
30
31
74
77
78
79
>10
>10
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
>10
0.486 ( 0.031
6.760 ( 0.117
(13.90)
5.504 ( 0.398
(141.13)
0.676 ( 0.031
(84.5)
2.18 ( 0.778
(57.37)
7.421 ( 0.286
(927.63)
7.860 ( 0.146
7.803 ( 0.935
(229.50)
46.026
4.323 ( 0.265
5.100 ( 0.392
5.683 ( 0.036
5.806 ( 0.168
5.416 ( 0.004
(8.90)
(10.49)
(11.69)
(11.95)
(11.14)
32
33
34
35
80
81
82
83
0.039 ( 0.003
0.008 ( 0.001
0.038 ( 0.007
0.008 ( 0.000
4.750 ( 0.946
(121.79)
0.511 ( 0.033
(63.88)
5.363 ( 0.098
(141.13)
7.552 ( 2.068
(944.00)
5.498 ( 0.060
5.589 ( 1.015
(164.38)
55.163
(5014.80)
8.539 ( 1.420
(142.32)
0.006 ( 0.001
(1.20)
4.383 ( 0.868
(112.38)
0.384 ( 0.029
(48)
5.291 ( 0.697
(135.66)
0.561 ( 0.001
(70.13)
4.999 ( 0.366
(128.18)
0.566 ( 0.009
(70.75)
4.026 ( 1.068
(103.23)
0.531 ( 0.002
(66.38)
0.627 ( 0.018
0.638 ( 0.003
0.656 ( 0.030
0.906 ( 0.069
(16.5)
(16.79)
(17.26)
(23.84)
7.636 ( 1.989
(954.50)
4.375 ( 0.022
3.799 ( 0.573
(111.74)
6.918 ( 0.550
(628.91)
5.856 ( 0.631
(97.60)
6.137 ( 0.327
(767.13)
6.020 ( 0.033
7.303 ( 0.712
(214.79)
7.535 ( 1.320
(685.00)
45.571
7.118 ( 0.237
(889.75)
6.057 ( 0.553
7.638 ( 0.248
(224.65)
7.379 ( 0.221
(670.82)
36.669
7.364 ( 1.163
(920.5)
36
37
84
85
0.076 ( 0.002
0.034 ( 0.004
5.356 ( 0.251
5.963 ( 0.211
(175.38)
6.439 ( 0.373
(585.36)
47.213
(786.88)
0.052 ( 0.001
(10.4)
38
39
40
41
86
87
0.011 ( 0.001
0.060 ( 0.008
0.005 ( 0.001
0.007 ( 0.001
(4184.18)
35.584
(593.07)
0.007 ( 0.001
(1.4)
(759.52)
0.005 ( 0.000
(1.00)
(611.15)
0.058 ( 0.001
(11.6)
Taxol
0.006 ( 0.001
(1.2)
Doxorubicin
0.040 ( 0.018
0.052 ( 0.019
0.045 ( 0.008
(6.429)
0.018 ( 0.006
0.053 ( 0.0002
0.037 ( 0.006
(5.71)
(7.43)
(2.57)
(7.57)
(5.29)
a Antiproliferative effects of cortistatin A (1), cortistatin J (5), and its synthetic analogues on human umbilical vein endothelial cells (HUVECs),
human tumor cell lines, and drug-resistant tumor cell lines in a 96 h growth inhibition assay using WST-8 colorimetric method. The human cancer cell
lines included were breast (MCF-7), lung (NCI-H460), CNS (SF268), ovarian (IA9) and its drug-resistant mutants PTX10 (Taxol-resistant) and A8
(epithilone resistant). Growth inhibition of 50% (GI50) is calculated as the drug concentration which caused 50% inhibition as compared to untreated
control. GI50 values for each compound represent mean of 2-5 independent experiments ( SE; Selectivity index (SI ) GI50 cancer cells/GI50 HUVECs
cells).
9
10596 J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009

Cortistatins A and J and Analogues
A R T I C L E S
and Taxol as standards. The results are summarized in Table 4.
As expected, cortistatin A (1) exhibited potent and selective
inhibition toward HUVECs, with the selectivity index ranging
between 2000 to 5000 against the cancer cell lines examined
(entry 1). Synthetic cortistatin J (5) was found under the
conditions employed in this study to be somewhat less active
than originally reported4a against the HUVECs and less selective
against the cancer cell lines examined (entry 2). Cortistatin
model systems 26-28 (entry 7-9), 30 (entry 10), 32 (entry
11), 33 (entry 12), 35 (entry 13), 36 (entry 14), 38 (entry 15),
41 (entry 16), 44 and 45 (entries 17 and 18) containing only
the ABCD ring systems of the cortistatin core structure showed
complete loss of activities against all cell lines tested. Com-
pounds 53 (entry 19) and 54 (entry 20) containing the DE ring
systems and the isoquinoline domain, despite their significant
loss of activity compared to natural cortistatin A, showed a
notable level of selectivity toward the HUVECs. While com-
pounds 74, 77, and 78 (entry 28-30) containing solely the
ABCDE ring systems of the cortistatins were completely
inactive, interestingly, their activities were restored upon
attachment of the isoquinoline moiety, as illustrated by com-
pounds 79-87 (entry 31-39). Stereochemistry of the isoquino-
line appendage seems to play somewhat of a role, although not
dramatic, in the observed inhibitory activity toward the HU-
VECs, as seen by comparing compounds 8 (entry 3) and 80
(entry 32). Compounds 8 (entry 3), 81 (entry 33), and 83 (entry
35) were the most active analogues synthesized. In particular,
the activities of advanced intermediates 8 (entry 3) and 81 (entry
33), lacking most of the A-ring functionalities, were of interest.
These results lead to the rather surprising conclusion that neither
the dimethyl amino nor the hydroxyl functionalities on ring A
of the cortistatin molecule are absolutely essential for biological
activity. In contrast, the isoquinoline moiety appears to be
consistently essential and required for the biological action of
these molecules. Considering the easier access through chemical
synthesis to such simplified analogues, this finding may prove
valuable and pathpointing for future explorations within the field.
(5). These syntheses demonstrate the power of cascade reactions
and palladium-catalyzed couplings in complex molecule con-
struction and the importance of chemical synthesis in rendering
scarce biologically active natural products readily available for
chemical biology investigations. Indeed, through application of
the developed synthetic routes, a number of analogues of the
cortistatins were synthesized and tested, providing important
insights into their structure activity relationships (SARs).31 Of
particular interest is the discovery that simplified analogues such
as 8 and 81 exhibit comparable antiproliferative activities against
HUVECs as the most active naturally occurring cortistatins
themselves. These findings are expected to catalyze and facilitate
further developments in the area of anticancer research.
Acknowledgment. Professor K. C. Nicolaou is also at the
Department of Chemistry and The Skaggs Institute for Chemical
Biology, The Scripps Research Institute, 10550 North Torrey Pines
Road, La Jolla, California 92037 (USA) and Department of
Chemistry and Biochemistry, University of California, San Diego,
9500 Gilman Drive, La Jolla, California 92093 (USA). We thank
Professor Paraskevi Giannakakou (Cornell University) for kindly
providing us with the IA9, PTX10, and A8 cell lines. We thank
Mr. Rong-Ji Sum (CSL-ICES) for his assistance in the biological
studies, Ms. Doris Tan (ICES) for high resolution mass spectro-
metric (HRMS) assistance, and Dr. Tommy Wang Chern Hoe
(Singapore Immunology Network (SIgN)-A*STAR) for X-ray
crystallographic analysis. Financial support for this work was
provided by A*STAR, Singapore.
Supporting Information Available: Experimental procedures
and compound characterization (PDF, CIF). This material is
available free of charge via Internet at http://pubs.acs.org.
JA902939T
(30) Giannakakou, P.; Gussio, R.; Nogales, E.; Downing, K. H.; Zaharevitz,
D.; Bollbuck, B.; Poy, G.; Sackett, D.; Nicolaou, K. C.; Fojo, T. Proc.
Natl. Acad. Sci. U.S.A. 2000, 97, 2904–2909.
(31) (a) Aoki, S.; Watanabe, Y.; Tanabe, D.; Arai, M.; Suna, H.; Miyamoto,
K.; Tsujibo, H.; Tsujikawa, K.; Yamamoto, H.; Kobayashi, M. Bioorg.
Med. Chem. 2007, 15, 6758–6762. (b) Sato, Y.; Kamiyama, H.; Usui,
T.; Saito, T.; Osada, H.; Kuwahara, S.; Kiyota, H. Biosci. Biotechnol.
Biochem. 2008, 72, 2992–2997. (c) Shi, J.; Shigehisa, H.; Guerrero,
C. A.; Shenvi, R. A.; Li, C.-C.; Baran, P. S. Angew. Chem., Int. Ed.
2009, 48, 4328–4331. (d) Czako, B.; Kurti, L.; Mammoto, A.; Ingber,
D. E.; Corey, E. J. J. Am. Chem. Soc. 2009, 131, 9014–9019.
Conclusion
In this article we described the evolution of a modular,
convergent synthetic strategy culminating in the total synthesis
of the antiangiogenic natural products cortistatins A (1) and J
9
J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009 10597