Total synthesis of (-)-Isatisine A

Article abstract of DOI:10.1002/anie.201101621

Short and sweet: (-)-IsatisineA was constructed using common and inexpensive building blocks, such as indole and D-ribose (see scheme). The synthesis features an unprecedented intramolecular Cglycosylation of an indole and an oxidative ring contraction. C

Full text of DOI:10.1002/anie.201101621

Communications
DOI: 10.1002/anie.201101621
Natural Products
Total Synthesis of (ꢀ)-Isatisine A**
Xiao Zhang, Tong Mu, Fuxu Zhan, Lijuan Ma, and Guangxin Liang*
Isatisine A (1; Figure 1) was isolated in 2007 by Chen et al.
whilst searching for anti-HIV compounds from the leaves of
Isatis indigotica Fort. (Cruciferae).[1] This biennial herbaceous
Figure 1. Structures of isatisine A (1) and its acetonide 2 showing the
absolute configuration that was established by Kerr’s group.
plant is widely used in traditional Chinese medicine for the
Scheme 1. Total synthesis of (+)-isatisine A by Karadeolian and Kerr.
Bn=benzyl, TBS=tert-butyldimethylsilyl, Ts=p-toluenesulfonyl.
prevention and treatment of viral diseases such as influenza,
viral pneumonia, mumps, and hepatitis.[2] Interestingly, its
derivative isatisine A acetonide (2) was identified first, and
was later proved to be an artifact of the isolation process.[1]
Extensive NMR spectroscopic analyses and X-ray crystallo-
graphic analysis of 2 revealed that isatisine A (1) features an
unprecedented fused tetracyclic framework with five contig-
uous stereogenic centers, two of which are fully substituted,
and a densely functionalized tetrahydrofuran core. The
absolute configuration of these two bisindole alkaloids was
established by the first total synthesis of (+)-isatisine A (ent-
1), which was reported by Karadeolian and Kerr
(Scheme 1),[3] and later was confirmed by the total synthesis
of ent-1 by Lee and Panek (Scheme 2).[4]
Apart from the challenging structural features of isati-
sine A (1), its uncharacterized bioactivity makes this com-
pound an unusually appealing synthetic target. Isatisine A
acetonide (2), the abundant artifact which arises from the
isolation process of isatisine A, was shown to be cytotoxic
against C8166 (CC50 = 302 mm) and also displayed anti-HIV-
1IIIB activity (EC50 = 37.8 mm).[1] Given these findings, the
Scheme 2. Total synthesis of (+)-isatisine A by Lee and Panek.
[*] X. Zhang, T. Mu, F. Zhan, L. Ma, Prof. G. Liang
State Key Laboratory and Institute of Elemento-organic Chemistry
Nankai University, Tianjin 300071 (China)
Fax: (+86)22-2350-0867
plant constituent 1, was speculated to possess comparable or
even stronger antiviral activity than its derivative 2.
Unfortunately, the bioactivity of isatisine A was not
reported in the paper that described its isolation because of
the scarcity of the isolated material.[5] An efficient total
synthesis of this natural product is expected to facilitate the
charactization of its bioactivity and future medicinal chemis-
try studies.
E-mail: lianggx@nankai.edu.cn
[**] We are grateful to the State Key Laboratory of Elemento-organic
Chemistry for generous start-up financial support. We thank the
National Natural Science Foundation of China (Grant No. 21032003
and 20902049) for funding. We also thank Prof. Chris Beaudry from
Oregon State University for helping us on the manuscript
preparation.
Karadeolian and Kerr (Scheme 1), and Lee and Panek
(Scheme 2) demonstrated elegant synthetic strategies that
applied two different [3+2] annulations for the construction
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201101621.
6164
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 6164 –6166

of the tetrahydrofuran core and used a dihydroxylation to
install the cis-diol moiety found in the natural product.
Subsequently, the second indole moiety was installed by a
nucleophilic addition to a rather reactive acyl iminium ion to
prepare ent-1.[3,4,6]
Herein, we report our total synthesis of (ꢀ)-isatisine A
(1), the natural enantiomer of the alkaloid. Our synthetic
strategy was inspired by the multiple hydroxy groups and
their relative stereochemistry on the tetrahydrofuran core of
1, which reminded us of d-ribose. We decided to take
advantage of this feature by synthesizing 1 from this furanose.
A retrosynthetic analysis for 1 from d-ribose (20) is shown in
Scheme 3. The final installation of the indole onto the indoxyl
Scheme 4. Initial synthetic efforts starting with 18. a) LDA, ꢀ788C,
THF; then 19 (76%); b) BzCl, Et3N, DMAP, CH2Cl2 (87%);
c) BF3·OEt2, MeNO2, 08C to RT (61%); d) SeO2, 1,4-dioxane, 1358C
(48%). Bz=benzoyl, DMAP=4-dimethylaminopyridine, LDA=lithium
diisopropylamide, THF=tetrahydrofuran.
smoothly afforded the cyclized product 23 in 61% yield.
To the best of our knowledge, this reaction represents the
first C glycosylation at the 2-position of an indole with an
unactivated anomeric center of a furanose.[10] With the key
intermediate 23 in hand, we chose to use the well-
established SeO2 oxidation to oxidize the position a to
the carbonyl group in 23.[11] However, when 23 was treated
with SeO2 in 1,4-dioxane in a sealed tube at 1358C, not only
did oxidation occur at the a position to give the desired
carbonyl group, but it also occurred at the anomeric center
(C1) to give 24. Variation of the reaction conditions did not
improve the result. To solve the problem, we decided to try
a different substrate, which contained a chlorine atom at
the 3-position of the indole moiety, in the hope that the
adjustment of the electronic properties of the indole would
change the reactivity at the anomeric center.
Scheme 3. Retrosynthetic analysis of (ꢀ)-isatisine A (1). Bn=benzyl,
PG=protecting group.
Indeed, this adjustment fulfilled our expectation and
allowed us to finish the total synthesis of (ꢀ)-isatisine A (1;
Scheme 5). The aldol reaction between the enolate of 25[12]
and 19, and subsequent benzoylation, produced the C-
glycosylation substrate 27 in good yield. Gratifyingly, com-
pared with 22, the chloroindole 27 showed a 23% improve-
ment on the yield of the C-glycosylation product. Treatment
of 28 with SeO2 in 1,4-dioxane in a sealed tube at 1358C
afforded the desired dicarbonyl product 29 without any
oxidation at the anomeric center. Interestingly, when 29 was
treated with hydrogen peroxide and potassium carbonate in
THF, a ring contraction was observed to afford the tetracyclic
intermediate 31 directly, in good yield. Presumably, this
transformation produced intermediate 30, which is the usual
product for an oxidative dicarbonyl cleavage reaction.[13]
Debenzoylation and ring opening occurred simultaneously
when 31 was subjected to NaOMe in methanol/THF (4:1).
Hydrogenolysis of 32 removed the two benzyl protecting
groups to produce a triol, which was then selectively
protected as a ketal to give 33 in excellent yield. Oxidation
of the chloroindole moiety in 33 with mCPBA[14] afforded 34
as a pair of diastereomers, the enantiomeric counterparts of
which were used in the total synthesis of ent-1 by Karadeolian
and Kerr. By using their procedure,[3,6] we were able to obtain
2. A mild hydrolysis of 2 with 1n HCl in MeOH readily
afforded 1 in 88% yield.
14 could be achieved through an acid-catalyzed nucleophilic
addition, which was used in the synthesis of ent-1 reported by
Karadeolian and Kerr.[3,6] It also has been previously dem-
onstrated that the indole precursor 15 can be oxidized to give
indoxyl 14.[3,6,7] The tetracyclic indole precursor 15 could arise
from a ring contraction of compound 16, which should be
readily prepared by an intramolecular C-glycosylation reac-
ꢀ
tion. The substrate 17 for this key C C bond formation could
be formed by an aldol addition of the enolate of 18 to the d-
ribose-derived ketone 19.
Our synthesis commenced with the d-ribose derivative
ketone 19, which was readily prepared from d-ribose (20) in
four steps by using published procedures.[8] Initially, we used
acetylindole 18[9] as the substrate for the aldol reaction
(Scheme 4). The reaction between the enolate of 18 and
ketone 19 successfully afforded the adduct 21 in 76% yield as
a single diastereomer. Protection of the tertiary alcohol 21
with a benzoyl group smoothly produced the C-glycosylation
precursor 22. Investigation of several Lewis acidic conditions
for this reaction revealed that BF3·OEt2 in nitromethane
produced 23 in a useful yield. When 22 was subjected to these
reaction conditions, the intramolecular C glycosylation
Angew. Chem. Int. Ed. 2011, 50, 6164 –6166
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6165

Communications
The success of the synthetic strategy allowed us to construct
the rather complex alkaloid from common and inexpensive
building blocks such as indole and d-ribose. Biological
assessment of (ꢀ)-isatisine A and its analogues is underway
in our laboratory.
Received: March 5, 2011
Published online: May 23, 2011
Keywords: alkaloids · glycosylation · natural products ·
.
ring contraction · total synthesis
[1] J.-F. Liu, Z.-Y. Jiang, R.-R. Wang, Y.-T. Zeng, J.-J. Chen, X.-M.
Zhang, Y.-B. Ma, Org. Lett. 2007, 9, 4127 – 4129.
[2] H. Z. Zheng, Z. H. Dong, Q. Yu in Modern Study of Traditional
Chinese Medicine, Vol. 1, Xueyaun, Beijing, 1997, pp. 328 – 334.
[3] A. Karadeolian, M. A. Kerr, Angew. Chem. 2010, 122, 1151 –
1153; Angew. Chem. Int. Ed. 2010, 49, 1133 – 1135.
[4] J. Lee, J. S. Panek, Org. Lett. 2011, 13, 502 – 505.
[5] Personal communication with the corresponding author of
Ref. [1].
[6] A. Karadeolian, M. A. Kerr, J. Org. Chem. 2010, 75, 6830 – 6841.
[7] K. Higuchi, Y. Sato, M. Tsuchimochi, K. Sugiura, M. Hatori, T.
Kawasaki, Org. Lett. 2009, 11, 197 – 199.
[8] N. Li, J. Lu, J. A. Piccirilli, Org. Lett. 2007, 9, 3009 – 3012.
[9] O. Olivia, C. Rosimeire, A. Robledo, Tetrahedron 1998, 54,
13915 – 13928.
[10] There are only a limited number of 2-indolylfuranoses reported
in the literature. Most of them were prepared in 2 or 3 steps
through the addition of a 2-indolyllithium compound or a 2-
indolylmagnesium compound to a reducing sugar or lactone and
a subsequent ring closure. For examples, see: a) S. Matsuda,
A. A. Henry, P. G. Schultz, F. E. Romesberg, J. Am. Chem. Soc.
2003, 125, 6134 – 6139; b) D. Guianvarcꢀh, J. Fourrey, M. T. H.
Dau, V. Guerineau, R. Benhida, J. Org. Chem. 2002, 67, 3724 –
3732; c) D. Guianvarcꢀh, R. Benhida, R. J.-L. Fourrey, Tetrahe-
dron Lett. 2001, 42, 647 – 650. The only reports of C glycosylation
at the 2-position of an indole are with an activated ribosyl
fluoride, see; d) M. Yokoyama, M. Nomura, H. Togo, H. Seki, J.
Chem. Soc. Perkin Trans. 1 1996, 2145 – 2149; e) M. Yokoyama,
M. Nomura, T. Tanabe, H. Togo, Heteroat. Chem. 1995, 6, 189 –
193.
Scheme 5. Total synthesis of (ꢀ)-isatisine A. a) LDA, ꢀ788C, THF;
then 19 (80%); b) BzCl, Et3N, DMAP, CH2Cl2 (93%); c) BF3·OEt2,
MeNO2, 08C to RT (84%); d) SeO2, 1,4-dioxane, 1358C (67%);
e) H2O2, K2CO3, THF (62%); f) NaOMe, MeOH/THF (4:1), 08C to RT
(84%); g) 20% Pd(OH)2/C, H2 (1 atm), THF; h) 2,2-dimethoxypro-
pane, p-toluenesulfonic acid, acetone (98%, 2 steps); i) mCPBA,
CH2Cl2; j) indole, CSA, CH2Cl2 (36%, 2 steps); k) 1n HCl, MeOH
(88%). CSA=10-camphorsulfonic acid, mCPBA=m-chloroperbenzoic
aci.
[11] a) S. Goswami, A. C. Maity, H. Fun, S. Chantrapromma, Eur. J.
Org. Chem. 2009, 1417 – 1426; b) S. Belsey, T. Danks, G. Wagner,
Synth. Commun. 2006, 36, 1019 – 1024; c) W. Aelterman, N.
De Kimpe, V. Kalinin, J. Nat. Prod. 1997, 60, 385 – 386; d) H. L.
Riley, J. F. Morley, N. A. C. Friend, J. Chem. Soc. 1932, 1875 –
1883.
The specific rotations of the synthetic acetonide 2 and 1
19
we obtained were ½aꢁD ¼ꢀ242 degcm3 gꢀ1 dmꢀ1 (c =
19
0.7 gcmꢀ3, MeOH) and ½aꢁD ¼ꢀ240 degcm3 gꢀ1 dmꢀ1 (c =
0.43 gcmꢀ3, MeOH) respectively. These specific rotations
are close to the value but opposite to those reported by
[12] C. S. Chien, T. Suzuki, T. Kawasaki, M. Sakamoto, Chem.
Pharm. Bull. 1984, 32, 3945 – 3951.
25
Karadeolian and Kerr (½aꢁD ¼ + 271 degcm3 gꢀ1 dmꢀ1 and
[13] For the cleavage of dicarbonyl compounds with hydrogen
peroxide under basic conditions, see: a) J. Yin, K. Zhang, C.
Jiao, J. Li, C. Chi, J. Wu, Tetrahedron Lett. 2010, 51, 6313 – 6315;
b) H. Langhals, G. Schoenmann, K. Polborn, Chem. Eur. J. 2008,
14, 5290 – 5303; c) T. Gꢁngꢂr, Y. Chen, R. Golla, Z. Ma, J. R.
Corte, J. P. Northrop, B. Bin, J. K. Dickson, T. Stouch, R. Zhou,
S. E. Johnson, R. Seethala, J. H. M. Feyen, J. Med. Chem. 2006,
49, 2440 – 2455; d) M. Tesmer, H. Vahrenkamp, Eur. J. Inorg.
Chem. 2001, 1183 – 1188.
25
½aꢁD ¼ + 274 degcm3 gꢀ1 dmꢀ1, respectively).[3] In addition,
the specific rotation of our synthetic 2 compares well to that
reported in the report describing the isolation of
2
14
(½aꢁD ¼ꢀ283 degcm3 gꢀ1 dmꢀ1 (c = 0.46 gcmꢀ3, MeOH)).[1]
Therefore, the isatisine A (1) we prepared is the natural
enantiomer.
In conclusion, we have achieved a total synthesis of the
natural enantiomer of isatisine A (1) in 11 steps with a 6.8%
overall yield from the readily available d-ribose derivative 19.
The synthesis features an unprecedented intramolecular
C glycosylation of an indole and an oxidative ring contraction.
[14] There are no previous reports of the oxidation of 3-chloroindoles
to give indoxyls in the literature. We applied the conditions for
oxidizing an indole to an indoxyl. See Ref. [3], [6], and [7].
6166 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 6164 –6166