Synthesis and biological evaluation of unnatural canthine alkaloids

Article abstract of DOI:10.1016/j.tetlet.2005.02.143

Employing a 'one-pot' microwave-assisted protocol, unnatural canthine alkaloids with biological activities beyond the natural products have been prepared. This report describes unnatural canthine alkaloid analogs as selective, allosteric Akt kinase inhibitors.

Full text of DOI:10.1016/j.tetlet.2005.02.143

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 2779–2782
Synthesis and biological evaluation of unnatural canthine alkaloids
Craig W. Lindsley,a,* Michael J. Bogusky,a William H. Leister,a Ray T. McClain,a
Ronald G. Robinson,b Stanley F. Barnett,b Deborah Defeo-Jones,b
Charles W. Ross, IIIa and George D. Hartmana
aDepartment of Medicinal Chemistry, Technology, Enabled Synthesis Group, Merck Research Laboratories, Merck & Co.,
PO Box 4, West Point, PA 19486, USA
bDepartment of Cancer Research, Merck Research Laboratories, Merck & Co., PO Box 4, West Point, PA 19486, USA
Received 11 February 2005; revised 24 February 2005; accepted 24 February 2005
Abstract—Employing a Ôone-potÕ microwave-assisted protocol, unnatural canthine alkaloids with biological activities beyond the
natural products have been prepared. This report describes unnatural canthine alkaloid analogs as selective, allosteric Akt kinase
inhibitors.
Ó 2005 Elsevier Ltd. All rights reserved.
Recently, the application of diversity-oriented synthesis
to construct libraries of complex molecules, based on
natural product templates, has received a great deal of
attention.1 Importantly, Shair and co-workers have
reported on the biomimetic solid-phase synthesis of
libraries of galanthamine-like alkaloids 1 with biological
properties beyond the natural product.2 Notably, this
work resulted in the discovery, through phenotypic
screens, of secramine 2, an unnatural amaryllidaceae
alkaloid, that inhibits protein trafficking (Fig. 1).2
of b-carbolines that possess an additional D-ring in the
core canthine skeleton 4 (Fig. 2).3 Since their discovery
in 1952, only forty members of this class of alkaloids
have been reported, and little structural diversity exists
from both natural and synthetic sources. Notably, mem-
bers of this family of alkaloids have demonstrated
a wide range of pharmacological activities including
antiviral and antifungal properties as well as potent
cytotoxicity against a variety of cell lines rendering this
framework attractive for diversity-oriented organic
synthesis.4
Inspired by this work, an unrelated class of natural
products, the canthin-6-one alkaloids 3 attracted our
attention.These alkaloids represent a tetracyclic subclass
In a recent letter, our laboratory reported on Ôone-potÕ
microwave-assisted protocol for the synthesis of the can-
thine alkaloid core.5 Under this protocol, an indole-teth-
ered acyl hydrazide 5 was treated with a 1,2-dicarbonyl
compound, such as benzil 6, in HOAc with excess NH4
OAc for 40 min in a single-mode microwave synthesizer
at 220 °C to deliver the unnatural 1,2-diphenyl canthine
derivative 7 in 80% yield (Scheme 1). Further expansion
O
OH
N
O
O
OMe
S
MeO
Br
H
OH
N
Me
O
N
R1
11
1
H
R2
2
A
8
C
B
N
1
N
4
2
N
N
D
6
Figure 1. Galanthamine and the unnatural sercramine.
O
4
3
*
Corresponding author. Tel.: +1 2156522265; fax: +1 2156526345;
e-mail: craig_lindsley@merck.com
Figure 2. Canthin-6-one and canthine alkaloid core structures.
0040-4039/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2005.02.143

2780
C.W.Lindsley et al./ Tetrahedron Letters 46 (2005) 2779–2782
O
micromolar levels of inhibition against Akt1. However,
Ph
Ph
the addition of a benzylic amine to the 4-position of one
of the phenyl rings, as in 8 and 9, dramatically increased
potency, but known GPCR privileged structures, such
as the piperidinyl benzimidazolone moiety in 9, proved
optimal.9
Ph
Ph
O
6
N
N
CONHNH2
N
NH4OAc (xs)
HOAc
220oC, 40 min
(80%)
5
7
Therefore, second generation libraries of unnatural can-
thine alkaloids were designed to incorporate basic
amines into the 4-position of both the C1 and C2 phenyl
rings, a previously undescribed pattern of substitution.
As depicted in Scheme 2, commercially available p-bro-
momethyl benzil 10 was treated with primary and sec-
ondary amines (Fig. 4), followed by scavenging with a
fluorous-tagged thiol. After FluoroFlashTM SPE purific-
ation, the p-aminomethyl benzil analogs 11 were obtained
with an average yield and purity of 93% and 95%,
respectively.10 Then, indole-tethered acyl hydrazide 5
was treated with one of the p-aminomethyl benzil ana-
logs 11 in HOAc with excess NH4OAc for 40 min in a
single mode microwave synthesizer at 220 °C. This key
Ôone-potÕ reaction produced 12 and 13 as a 1:1 mixture
of regioisiomeric canthine alkaloids, resulting from the
unsymmetrical benzils 9 employed, in yields ranging
from 35% to 70% (un-optimized). All analogs were
purified (>98%) by mass-triggered preparative HPLC;
Scheme 1. ÔOne-potÕ microwave synthesis of the canthine core.
of this methodology for diversity-oriented organic syn-
thesis (DOS) focused on functionalization of the C1/
C2 positions of the canthine core, a historically difficult
position to analog, utilizing a diverse collection of func-
tionalized 1,2-dicarbonyl derivatives.5
Concurrently, our laboratory was engaged in a nascent
medicinal chemistry program to develop selective Akt
(PKB) kinase inhibitors. Akt is a serine/threonine kinase
that recently garnered attention as a promising target
for cancer therapy due to its critical role as a regulator
of the cellÕs apoptotic machinery.6 However, the devel-
opment of small molecule Akt inhibitors for the treat-
ment of cancer has been hindered by a lack of Akt
specific (vs the AGC family of kinases) and isozyme
selective (Akt1, Akt2, and Akt3) inhibitors due to high
sequence identity homology.7
Our lead for this program, a 2,3-diphenylquinoxaline 8,
resulted from a high throughput screening effort to iden-
tify compounds capable of inhibiting the three Akt
isozymes (Fig. 3).8 Interestingly, 8 displayed a high
degree of isozyme selectivity (IC50s: Akt1 = 3.4 lM,
Akt2 = 23.1 lM, and Akt3 >50 lM) and was found to
be specific for Akt (>50 lM vs PKA, PKC, and SGK).
In assays with Akt mutants lacking the PH domain, 8
displayed no inhibition; moreover, Akt inhibition by 8
was not competitive with ATP. The selectivity observed
has been attributed to an allosteric mode of inhibition,
for which a model has been proposed.8 Further optimi-
zation of 8 resulted in 9, a more potent dual Akt1/Akt2
inhibitor (IC50s: Akt1 = 0.06 lM, Akt2 = 0.21 lM, and
Akt3 2.2 lM) with the same allosteric mode of
inhibition.9
Br
NR1R2
(a)
O
O
O
O
5
(b)
10
11
NR1R2
NR1R2
+
N
N
N
N
12
13
Recognition of the related 1,2-diphenyl heterocyclic
motif present in 7 led us to screen this unnatural canthine
alkaloid in our Akt kinase assay. Unfortunately, 7, an
unsubstituted phenyl analog, displayed no inhibition
(>50 lM) against all three of the Akt isozymes. For
the series of inhibitors represented by 8 and 9, the
unsubstituted phenyl congeners displayed moderate
Scheme 2. Library of unnatural canthine alkaloid analogs 12:13.
Reagents and conditions: (a) (i) HNR1R2, PS-DIEA, DCM, rt,
(ii) Rf-SH, (iii) FluoroFlashTM SPE, 93%; (b) 5 (1.1 equiv), NH4OAc
(10 equiv), HOAc, 220°C, 40 min, microwave, 35–70%. All compounds
purified by mass-triggered HPLC.
HNMe2
HNEt2
NH2
NH
NH
NH
R
O
NH2
N
O
S
N
N
N
N
N
N
NH
NH2
O
RN
N
H
NH
O
O
N
R
NHR
HN
HN
HN
NH
NH
8
9
Figure 3. HTSscreening lead and optimized dual Akt1/Akt2 inhibitor.
Figure 4. Primary and secondary amines.

C.W.Lindsley et al./ Tetrahedron Letters 46 (2005) 2779–2782
2781
R1
O
NH
NH
N
O
N
N
R2
N
N
N
N
O
NH
N
N
N
16
N
N
14 R1 = H, R2 = 16
15 R1 = 16, R2 = H
14
15
Figure 6. ROE correlations to assign structures to 14 and 15.
Figure 5. Unnatural canthine alkaloids that inhibit Akt.
inhibitor, HPLC conditions were rapidly developed to
separate the two regioisomeric canthines 14 and 15.
Once separated, standard 1-D NMR techniques failed
to distinguish the two regioisomers as well as a variety
of 2-D heteronuclear NMR techniques (HMQC and
HMBC). Ultimately, the regioisomers were distin-
guished by ROE correlations (illustrated by arrows in
Fig. 6) and further confirmed by HMBC 3-bond
1H-13C correlations and structures were assigned as
shown in Figure 5.13
however, separation of the individual regioisomers
proved problematic.11 Therefore, we elected to perform
initial screening on the 1:1 mixtures of regioisomeric
canthines.
With the exception of 14:15,12 derived from the incorpo-
ration of the piperidinyl benzimidazolone moiety in 9,
all of the analogs screened against the three Akt iso-
zymes were inactive (Fig. 5). This was a surprising
result, as the analogous congeners in the quinoxaline
series 8 and 9, were active compounds. As a 1:1 mixture
of regioisomers, 14:15 displayed Akt isozyme selectivity
(IC50s: Akt1 = 1.8 lM, Akt2 = 2.8 lM, and Akt3
>50 lM) and an almost 10-fold increase in Akt2 inhibi-
tion, relative to 8. As observed with the quinoxaline
series, inhibition was PH-domain dependent and non-
competitive with ATP.8,9 Though clearly an allosteric
inhibitor, the divergent SAR from the quinoxaline series
potentially suggests an alternative allosteric binding site.
With pure regioisomers 14 and 15 in hand, they were re-
assayed in our Akt kinase screen in order to determine if
the inhibitory activity observed was due to only one or
both regiosiomers. As shown in Table 1, both 14 and
15 were essentially equipotent against both Akt1 and
Akt2. Moreover, both 14 and 15 were PH domain
dependent and noncompetitive with ATP and therefore,
represent only one of the three known templates to pro-
vide this type of allosteric inhibition of the Akt kinase.8,9
These data also suggested that the source of Akt inhibi-
tion might be solely due to the piperidinyl benzimidazo-
lone moiety, 16, in combination with a 1,2-diphenyl
heterocyclic motif. In order to evaluate this possibility,
a library of 48 compounds was prepared wherein vari-
ous heterocycles 17 with a 1,2-diphenyl motif were func-
tionalized with 16 to provide 18 (Scheme 3). When
screened in our Akt kinase assay, none of these closely
related analogs displayed any inhibition of the individ-
ual Akt isozymes.
Based on these data, 14 and 15 were studied for their
ability to induce apoptosis in a LnCaP tumor cell line,
as measured by an increase in caspase-3 activity. When
LnCaP cells were treated with 15 lM 14 or 15, there
was no observable increase in caspase-3 activity. How-
ever, when co-administered with TRAIL (TNF-a related
apoptosis inducing ligand), a modest 1.8- to 2.2-fold
increase in caspase-3 activity was observed relative to
control.8,9 This was an attractive starting point for
further studies.
With data suggesting that the canthine scaffold, in con-
cert with 16, provided a truly novel allosteric Akt kinase
The results obtained from these small libraries argue
well for the synthesis of additional libraries based on
the canthine core with increased structural diversity.
Currently, libraries incorporating our Ôone-potÕ protocol
Br
N
O
16
N
NH
Het
Het
Table 1. Akt inhibition by unnatural canthine alkaloids 14 and 15
(a)
Compound
Akt1 IC50
(lM)a
Akt2 IC50
(lM)a
Akt3 IC50
(lM)a
18
17
14
15
1.34
1.47
1.65
2.28
>50
>50
Scheme 3. Library of functionalized 1,2-diphenyl heteroaromatic
scaffolds. Reagents and conditions: (a) (i) 16, PS-DIEA, DCM, rt,
(ii) Rf-SH, (iii) FluoroFlashTM SPE, 94%.
All compounds >50 lM versus PKA, PKC, SGK, and d-PH mutants
a Average of three measurements.

2782
C.W.Lindsley et al./ Tetrahedron Letters 46 (2005) 2779–2782
Neubauer, B. L.; Lai, M. T.; Graff, J. R. Clin.Cancer Res.
2001, 7, 2475–2479.
R2
R2
R3
R3
8. For Akt enzyme assay details and the proposed model for
the allosteric mode of inhibtion of these inhibitors, see:
Barnett, S. F.; Defeo-Jones, D.; Fu, S.; Hancock, P. J.;
Haskell, K. M.; Jones, R. E.; Kahana, J. A.; Kral, A.;
Leander, K.; Lee, L. L.; Malinowski, J.; McAvoy, E. M.;
Nahas, D. D.; Robinson, R.; Huber, H. E. Biochem.
J. 2005, 385, 399–408.
9. Lindsley, C. W.; Zhao, Z.; Leister, W. H.; Robinson, R.
G.; Barnett, S. F.; Defeo-Jones, D.; Jones, R. E.;
Hartman, G. D.; Huff, J. R.; Huber, H. E.; Duggan, M.
E. Bioorg.Med.Chem.Lett. 2005, 15, 761.
R1
R1
N
N
N
N
O
Figure 7. Generic canthine libraries with 3-position diversity.
with three-position diversity are being prepared and the
resulting biological activities of unnatural canthine alka-
loids will be reported in due course (Fig. 7).
In summary, unnatural canthine alkaloids have been
discovered with biological properties beyond the natural
product by application of microwave-assisted organic
synthesis (MAOS). Unnatural canthine alkaloids 14
and 15 were shown to be potent and selective Akt kinase
inhibitors that induced apoptosis in LnCaP cells. More-
over, these unnatural molecules were dependent on the
PH domain of the Akt kinase for enzyme inhibition
and noncompetitive with ATP, suggesting a unique,
allosteric mode of inhibition. Clearly, the melding of
natural product templates with known pharmacophores
(privileged structures) can result in attractive lead struc-
tures for medicinal chemists to pursue and refine.
10. (a) Lindsley, C. W.; Zhao, Z.; Leister, W. H. Tetrahedron
Lett. 2002, 43, 4225–4228; (b) Zhang, W.; Curran, D. P.;
Chen, C. H.-T. Tetrahedron 2002, 58, 3871–3875, Fluor-
ous scavenging proved optimal in this instance to deliver
clean material for the subsequent 1,2,4-triazene formation
and cycloaddition. Polymer-supported thiol failed to
provide material in sufficient purity.
11. Leister, W.; Strauss, K.; Wisnoski, D.; Zhao, Z.; Lindsley,
C. W. J.Comb.Chem. 2003, 5, 322–329.
12. Following the experimental procedure outlined in Refs. 5
and 9, 14 and 15 were obtained in a combined 48%
isolated yield. Analytical data for 14: Analytical LCMS, a
single peak at 2.176 min.
1
(MeCN/H2O/0.05%TFA), 4 min gradient, >99% pure; H
NMR (CDCl3, 600 MHz), d 9.09 (br s, 1H), 7.72 (t,
J = 7.78 Hz, 7.73 Hz, 1H), 7.62 (d, J = 8.48 Hz, 1H), 7.45
(m, 3H), 7.42 (d, J = 5.89 Hz, 2H), 7.41 (d, J = 5.89 Hz,
2H), 7.39 (obs d, 1H), 7.33 (d, J = 8.16 Hz, 1H), 7.31 (t,
J = 7.61 Hz, 6.98 Hz, 2H), 7.18 (t, J = 8.84 Hz, 7.79 Hz,
1H), 7.01 (br m, 3H), 4.60 (br s, 1H), 4.44 (t, J = 7.00 Hz,
5.62 Hz, 2H), 4.24 (br s, 2H), 3.71 (br t, 2H), 3.59 (br d,
2H), 2.88 (m, 4H), 2.62 (t, J = 5.46 Hz, 6.79 Hz, 2H), 1.94
(br s, 2H); 13C NMR (125 MHz, CDCl3, 25 °C) d 154.5,
143.8, 139.6, 138.8, 134.7, 134.1, 132.5, 131.8, 131.5, 131.2,
131.1, 130.1, 129.7, 129.4, 129.2, 128.4, 127.8, 125.1, 122.2,
122.0, 121.9, 120.9, 110.4, 110.3, 109.9, 60.0, 51.7, 47.3,
41.2, 26.0, 23.7, 21.8; HRMScalcd for C 39H36N5O
(M+H), 590.2914; found 590.2926; Analytical data for
15: Analytical LCMS, a single peak at 2.389 min (MeCN/
H2O/0.05%TFA), 4 min gradient, >99% pure; 1H NMR
(CDCl3, 600 MHz), d 8.40 (br s, 1H), 7.72 (t, J = 7.95 Hz,
8.1 Hz, 1H), 7.60 (d, 8.79 Hz, 1H), 7.57 (d, 7.35 Hz, 2H),
7.47 (d, J = 7.95 Hz, 1H), 7.43 (d, J = 7.35 Hz, 2H), 7.36
(d, J = 8.21 Hz, 1H), 7.31 (d, J = 7.91 Hz, 2H), 7.28 (obs,
1H), 7.23 (t, J = 7.60 Hz, 8.97 Hz, 2H), 7.21 (obs, 1H),
7.12 (t, J = 7.30 Hz, 1H), 7.09 (t, J = 7.90 Hz, 1H), 7.07 (d,
J = 8.21 Hz, 1H), 4.71 (m, 1H), 4.44 (t, J = 5.43 Hz, 2H),
4.33 (s, 2H), 3.73 (t, J = 5.56 Hz, 2H), 3.69 (br d,
J = 11.4 Hz, 2H), 2.98 (q, J = 13.36 Hz, 12.65 Hz, 2H),
2.85 (t, J = 12.4 Hz, 2H), 2.62 (m, 2H), 2.03, (d,
J = 13.17 Hz, 2H); 13C NMR (125 MHz, DMSO-d6,
25 °C) d 153.2, 142.0, 140.7, 140.5, 137.0, 131.5, 131.3,
130.2, 129.4, 128.4, 128.0, 127.9, 127.8, 127.5, 125.6, 122.9,
120.6, 120.2, 119.9, 119.5, 110.8, 108.8, 108.2, 58.5, 50.8,
46.5, 40.3, 25.2, 24.7, 21.1; HRMScalcd for C 39H36N5O
(M+H), 590.2914; found 590.2921.
References and notes
1. (a) Schreiber, S. L. Chem.Eng.News 1992 (October 26),
220–232; (b) Tan, D. S.; Foley, M. A.; Shair, M. D.;
Schreiber, S. L. J.Am.Chem.Soc. 1998, 120, 8565–8566;
(c) Lindsley, C. W.; Chan, L. K.; Goess, B. C.; Joseph, R.;
Shair, M. D. J.Am.Chem.Soc. 2000, 122, 422–423.
2. Pelish, H. E.; Westwood, N. J.; Feng, Y.; Kirchhausen, T.;
Shair, M. D. J.Am.Chem.Soc. 2001, 123, 6740–6741.
3. For an excellent review on the canthine and canthin-6-one
alkaloids see: Ohmoto, T.; Koike, K. In The Alkaloids;
Brossi, A., Ed.; Academic: New York, 1989; Vol. 36, pp
135–170.
4. (a) Haynes, H. F.; Nelson, E. R.; Price, J. R. Aust.J.Sci.
Res.Ser.A 1952, 5, 387; (b) Nelson, E. R.; Price, J. R.
Aust.J.Sci.Res.Ser.A
Price, J. R. Aust.J.Sci.Res.Ser.A
1952, 5, 563; (c) Nelson, E. R.;
1952, 5, 68; (d)
Anderson, L. A.; Harris, A.; Phillipson, J. D. J.Nat.Prod.
1983, 46, 374; (e) Arisawa, M.; Kinghorn, A. D.; Cordell,
G. A.; Farnsworth, N. R. J.Nat.Prod. 1983, 46, 222; (f)
Ohmoto, T.; Nikaido, T.; Koide, K.; Kohda, K.; San-
kawa, U. Chem.Pharm.Bull. 1988, 36, 4588; (g) Ouyang,
Y.; Koide, K.; Ohmoto, T. Phytochemistry 1994, 36, 1543–
1546.
5. Lindsey, C. W.; Wisnoski, D. D.; Wang, Y.; Leister, W.
H.; Zhao, Z. Tetrahedron Lett. 2003, 44, 4495–4498.
6. For excellent reviews see: (a) Graff, J. R. Expert Opin.
Ther.Targets 2002, 6, 103–113; (b) Nicholson, K. M.;
Anderson, N. G. Cell.Signalling 2002, 14, 381–395; (c) Li,
Q.; Zhu, G.-D. Curr.Top.Med.Chem. 2002, 2, 939–971.
7. (a) Hanks, S.; Hunter, T. FASEB 1995, 9, 576–596; (b)
Zinda, M. J.; Johnson, M. A.; Paul, J. D.; Horn, C.;
Konicek, B. W.; Lu, Z. H.; Sandusky, G.; Thomas, J. E.;
13. All NMR experiments conducted on a Varian Inova
600 MHz spectrometer. Additional details available upon
request.