Total synthesis of (+)-kopsihainanine A

Article abstract of DOI:10.1039/c4cc01843e

Total synthesis of (+)-kopsihainanine A was accomplished on the basis of (i) Stoltz's enantioselective decarboxylative asymmetric allylation and (ii) the proposed biogenetic pathway from the related alkaloid, kopsihainanine B. In addition, HPLC analysis of the synthetic (+)-kopsihainanine A confirmed its ee to be 99% with [α]30D = 25.35. the Partner Organisations 2014.

Full text of DOI:10.1039/c4cc01843e

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Total synthesis of (+)-kopsihainanine A†
Masaya Mizutani, Shigeo Yasuda and Chisato Mukai*
Cite this: Chem. Commun., 2014,
50, 5782
Received 12th March 2014,
Accepted 14th April 2014
DOI: 10.1039/c4cc01843e
www.rsc.org/chemcomm
Total synthesis of (+)-kopsihainanine A was accomplished on the
basis of (i) Stoltz’s enantioselective decarboxylative asymmetric
allylation and (ii) the proposed biogenetic pathway from the related
alkaloid, kopsihainanine B. In addition, HPLC analysis of the synthetic
(+)-kopsihainanine A confirmed its ee to be 99% with [a]3D0 = 25.35.
In 2011, Gao and co-workers1 isolated the novel monoterpenoid
indole alkaloid, (+)-kopsihainanine A (1), together with the bio-
genetically related (+)-kopsihainanine B (2) from the leaves and
stems of Kopsia hainanesis, which is a native species historically
used in folk medicine for the treatment of rheumatoid arthritis,
pharyngitis, tonsillitis, and dropsy in the Hainan Province, China.
The unprecedented strained pentacyclic skeleton of 1 as well as its
relative stereochemistry were elucidated by spectral analysis, speci-
fically by careful 1H NMR considerations. The absolute configu-
ration of 1 was deduced to be 16R, 20R, 21S based on a comparison
of the actual electronic circular dichroism trace of (+)-1 to those
predicted for four candidates using AMI calculations and the ZINDO
method. The first total synthesis of kopsihainanine A (1) was
Scheme 1 Structures of kopsihainanines
disubstituted-tetrahydrocarbazol-4-ones (3–6).
A (1) and B (2) and 3,3,-
Gao and co-workers1 proposed a possible biogenetic pathway for
achieved from the tricyclic tetrahydrocarbazol-4-one derivative 3 in the production of (+)-kopsihainanine A (1) from (+)-kopsihainanine B
a racemic form by Xie, She and co-workers2 in 2012. Shao and (2). Namely, the initial decarboxylative carbon–carbon bond cleavage
co-workers3 very recently succeeded in the preparation of the of 2 would lead to the tetracyclic compound I with the C4-carbon
optically active (À)-3 by taking advantage of the palladium- tether at the ring juncture, which would be susceptible to the
catalyzed asymmetric decarboxylative allylation of 6 in the presence selective oxidative fission of the terminal double bond giving rise
of the PHOX ligand and completed the total synthesis of (+)-1 to the formation of the a-ketoaldehyde derivative II. The following
according to Xie and She’s procedure.2 This total synthesis unambi- condensation between the secondary amine and aldehyde moieties
guously confirmed the absolute structure of (+)-1 as depicted in would produce the hemiacetal III, which would spontaneously
Scheme 1. A similar palladium-catalyzed asymmetric decarboxylative collapse to (+)-1 via the keto–enol tautomerized intermediate IV.
allylation4 produced the optically active 4, which was converted into
During our investigation of the syntheses of indole alkaloids and
the carbamoyl derivative 5, but further transformation into the target derivatives,5 the strained and unprecedented pentacyclic framework
natural product has not yet been achieved, although the racemic 5 of (+)-1 as well as its biological background strongly intrigued us and
has already been converted into (Æ)-1 by Xie, She and co-workers.2 (+)-1 immediately became our next target molecule. We noted Gao’s
proposal1 for the biogenetic pathway from (+)-2 to (+)-1. If the
plausible biogenetic intermediate II or its equivalent could be
obtained in an optically active form, the biomimetic-type transforma-
tion of them into (+)-1 would be achieved. The retrosynthetic analysis
Division of Pharmaceutical Sciences, Graduate School of Medical Sciences,
Kanazawa University, Kakuma-machi, Kanazawa 920-1192, Japan.
E-mail: mukai@p.kanazawa-u.ac.jp; Fax: +81-76-234-4410; Tel: +81-76-234-4411
of (+)-1 afforded the reasonable tactics outlined in Scheme 3. The
alkylation of the d-lactam d with the indole derivative e would
† Electronic supplementary information (ESI) available: Experimental procedures
and compound characterization data (PDF). See DOI: 10.1039/c4cc01843e
5782 | Chem. Commun., 2014, 50, 5782--5785
This journal is ©The Royal Society of Chemistry 2014

View Article Online
Communication
ChemComm
produce the condensation product c. The decarboxylative allylation
of c under suitable chiral circumstances6 would furnish the optically
active d-lactam b with all the carbon units required for (+)-1. The
Bischler–Napieralski reaction7 at the C3-position of the indole
nucleus of b followed by stereoselective reduction would result in
the production of the precursor a, an equivalent of the plausible
biogenetic intermediate II. Consecutive oxidation and condensation
of a would finally deliver the target natural product (+)-1.
Treatment of the known a-allyloxycarbonyl-N-benzoyl-d-lactam
(7)8a with 1-(tert-butoxycarbonyl)-2-(2-iodoethyl)indole (8), easily
derived from indole over 4 steps in 67% yield,9 in the presence
of potassium carbonate provided the condensation product 9 in
75% yield. The decarboxylative allylation of 9 to the optically active
10 in the presence of a suitable asymmetric catalyst must be the
most crucial step in this synthesis. However, we initially deter-
mined and optimized the synthetic route using racemic 10. Thus,
compound 9 was exposed to the achiral conditions (Pd(PPh3)4 in
THF) at room temperature to afford the racemic decarboxylated
product 10 in 80% yield; the two protecting groups on the nitrogen
atoms of which were subsequently removed by potassium hydroxide
to produce 11 in 84% yield. The Bischler–Napieralski reaction of 11
with phosphorus chloride7a in refluxing toluene smoothly occurred
to form the iminium salt. The stereoselective attack of the hydride
species of NaBH4 on the iminium salt occurred from the opposite
side of the allyl moiety resulting in the exclusive formation of the
tetracyclic compound 12 in 82% overall yield from 11. Thus, we
could easily prepare the precursor for the plausible biogenetic
intermediate II (see Scheme 2). Our next endeavor focused on
adjustment of the allyl group of 12 to the suitably oxidized form
for completion of the total synthesis. Prior to the oxidative chemical
modification of the allyl moiety, both secondary nitrogen atoms of
the indole nucleus and the piperidine of 12 were successively
protected with tert-butoxycarbonyl (13, 83%) and benzyloxycarbonyl
groups, respectively, to furnish 14.10 Upon exposure to RuO4-
mediated oxidation conditions (RuCl3/CeCl3/NaIO4),11 compound 14
underwent dihydroxylation to give 15.10 Conversion of the dihydroxy
moiety of 15 to the a-ketoaldehyde functionality was examined
under several oxidation conditions, but the a-ketoaldehyde derivative
could not be obtained. We tentatively assumed that the labile
a-ketoaldehyde functionality might immediately decompose
after its production even if the oxidation proceeded as expected.
Scheme 3 Retrosynthesis of (+)-1.
In order to intramolecularly capture the labile a-ketoaldehyde by the
amine moiety, the benzyloxycarbonyl group of 15 was removed under
hydrogenation conditions to provide 16,10 which was then treated
with IBX to give the over-oxidized pentacyclic diketone derivative 17
(in 19% overall yield from 13) instead of the expected N-protected III
and/or 1 (see Schemes 1 and 2). The direct conversion of 16 into 1 was
again examined under several oxidation conditions, but a favorable
result could not be realized. Thus, the over-oxidized product 17 was
reduced with NaBH4 in t-BuOH/THF at 0 1C to room temperature to
produce the N-protected kopsihainanine A 18 in 31% yield. The total
synthesis of (Æ)-1 was finally accomplished by removal of the tert-
butoxycarbonyl group in 80% yield (Scheme 4).
We could develop the synthetic route towards the target molecule
in a racemic form. Our next effort, therefore, was directed toward the
preparation of the optically active 10 with a high enantiomeric excess
(ee), which is mandatory for the total synthesis of (+)-1. Stoltz and
Scheme 4 Total synthesis of (Æ)-kopsihainanine A. (a) 8, K2CO3, DMF,
50 1C, 75%. (b) Pd(PPh3)4 (8 mol%), THF, rt, 80%. (c) KOH, MeOH, reflux,
84%. (d) POCl3, toluene, reflux, then NaBH4, MeOH, 0 1C to rt, 82%.
(e) (Boc)2O, Et3N, DMAP, DCM, rt, 83%. (f) CbzCl, Na2CO3, DCM, rt.
(g) RuCl3ÁnH2O, CeCl3Á7H2O, AcOEt/CH3CN/H2O (3/3/1), 0 1C. (h) H2 (1 atm),
Pd(OH)2, EtOH, rt. (i) IBX, DMSO, rt, 19% from 13. (j) NaBH4, t-BuOH/THF (1/1),
0 1C to rt, 31%, (k) KOH, MeOH, reflux, 80%.
Scheme 2 Proposed biogenetic pathway for production of (+)-1 from (+)-2.
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 5782--5785 | 5783

View Article Online
ChemComm
Communication
co-workers8a recently developed the palladium-catalyzed highly
enantioselective decarboxylative allylation of 1-acyl-3-alkyl-3-(allyl-
oxycarbonyl)lactams to 3-alkyl-3-allylpyrrolidinone, piperidinone,
and caprolactams. Thus, application of Stoltz’s procedure using
(S)-t-BuPHOX (19) to 9 was examined. A solution of racemic 9 in
toluene was heated at 40 1C in the presence of Pd2(dba)3 (5 mol%)
and (S)-t-BuPHOX (19) (12.5 mol%) for 6 h to obtain (+)-10 in 66%
yield with 71% ee (Table 1, entry 1). According to the literature,8a
several solvents were screened to get the best results. Neither
toluene/hexane or tetrahydrofuran afforded results better than that
of toluene alone (entries 2 and 3). Methyl tert-butyl ether (MTBE)
was found to produce a better chemical yield as well as a better ee
(entry 4). A slight improvement in the ee was attained when a
combination of MTBE with hexane or cyclohexane was employed
(entries 5 and 6). A similar result was observed upon treatment with
a half amount of the palladium catalyst and the ligand 19 (entry 7).
Finally, a more electron-poor catalyst, (S)-(CF3)3-t-BuPHOX (20),8,12
became the most powerful catalyst for our purpose. Indeed, treat-
ment of 9 with Pd2(dba)3 (5 mol%) and 20 (12.5 mol%) in MTBE
effected the highly enantioselective asymmetric allylation to pro-
duce (+)-10 in 80% yield with 98% ee (entry 8). These tendencies
Scheme 5 Total synthesis of (+)-kopsihainanine A (1). (a) KOH, MeOH,
reflux, 77%. (b) POCl3, toluene, reflux, then NaBH4, MeOH, 0 1C to rt, 92%.
(c) (Boc)2O, Et3N, DMAP, DCM, rt, 83%. (d) CbzCl, Na2CO3, DCM, rt.
(e) RuCl3ÁnH2O, CeCl3Á7H2O, AcOEt/CH3CN/H2O (5/3/1), 0 1C. (f) H2
(1 atm), Pd(OH)2, EtOH, rt. (g) IBX, DMSO, rt, 33% from 13. (h) NaBH4,
t-BuOH/THF (1/1). 0 1C to rt, 45%. (i) KOH, MeOH, reflux, 85%.
regarding the solvent and ligand are in complete agreement with which was then transformed into (+)-kopsihainanine A (1) in 85%
those reported by Stoltz.8a Furthermore, the enantioselective allyl- yield under basic conditions (Scheme 5). The spectral data for the
ation could be scaled up to the 1 mmol scale (entry 9).
synthetic (+)-1 was in accordance with those of the natural one.1
The total synthesis of (+)-kopsihainanine A (1) was accomplished However, the synthetic (+)-1 showed its specific rotation as [a]D30
from (+)-10 (98% ee) according to the procedure described in a 25.35 (c = 0.33, CHCl3), while that of the natural product was [a]D25
=
=
racemic series. Compound (+)-10 was subsequently converted into 60 (c = 0.10, CHCl3).1 Furthermore, Shao reported the specific
(À)-11 (77%), (+)-12 (92%), and (+)-13 (83%). The N-benzyloxy- rotation of the synthetic (+)-1 as [a]2D5 = 55.0 (c = 0.5, CHCl3)3 which
carbonylation of (+)-13 was followed by dihydroxylation, debenzyloxy- is very similar to that of the natural (+)-1. We measured the specific
carbonylation, and oxidation to produce (+)-17 in 33% overall yield rotation of the synthetic (+)-1 by changing the solvent, but similar
from (+)-13. Reduction of (+)-17 afforded (À)-18 in 45% yield, low values were again observed [[a]1D6 = 39.44 (c = 0.10, MeOH) and
[a]1D9 = 10.00 (c = 0.10, AcOEt)]. The simplest interpretation about
these results is the assumption that racemization must have
Table 1 Asymmetric decarboxylative allylation of d-lactam 9a
occurred during the conversion of (+)-10 into the final product.
Therefore, the HPLC analysis of (+)-1 was carried out. It was not an
easy task to determine the suitable conditions for the HPLC analysis
of 1 due to its poor solubility. After screening various conditions
using the racemic 1, we finally found the optimal conditions
Entry
Ligand
Solvent
Yield (%)
eeb (%)
(CHIRALPAK IE, hexane/MeOH/CH2Cl2/ethylenediamine
=
70/6/24/0.06, 280 nm) showing two diagnostic peaks due to (+)-
and (À)-1 in the ratio of 50 to 50. Thus, the HPLC analysis of (+)-1
by applying the optimal conditions unambiguously demonstrated
that a high ee (99%)13 is still maintained. Currently we do not have
any clues to explain the difference between the specific rotation of
our compound and those of the natural and Shao’s ones.
Stoltz’s enantioselective decarboxylative asymmetric allylation of
a-allyloxycarbonyl-N-benzoyl-a-(2-indolyl-2-ethyl)-d-lactam produced
the corresponding optically active d-lactam derivative with 98% ee,
which was exposed to the Bischler–Napieralski conditions and
NaBH4 to afford the indoloperhydroquinoline framework. According
to the proposed biogenetic pathway from kopsihainanine B (2) to A,
several chemical manipulations were executed resulting in the
completion of total synthesis of (+)-kopsihainanine A. HPLC analysis
of the synthetic (+)-kopsihainanine A confirmed its ee to be 99% with
[a]3D0 = 25.35 (c = 0.33, CHCl3). We have completed the total synthesis
of (+)-kopsihainanine A (1) from indole through 15 steps in 3.0%
1
2
3
4
19
19
19
19
19
19
19
20
20
Toluene
Hexane–toluene (2 : 1)
THF
66
65
65
76
77
75
75
80
82
71
27
39
81
84
85
86
98
98
MTBE
5
Hexane–MTBE (2 : 1)
Cyclohexane–MTBE (2 : 1)
Cyclohexane–MTBE (2 : 1)
MTBE
6
7c
8
9d,e
MTBE
a
b
The reactions were performed on the 0.12 mmol scale. ee was
c
determined by HPLC analysis (Diacel CHIRALCEL OD-H). 2.5 mol%
of Pd2(dba)2 and 6.3 mol% of 19 were used. The reaction was
performed on the 1.0 mmol scale. Reaction time was 15 h.
d
e
5784 | Chem. Commun., 2014, 50, 5782--5785
This journal is ©The Royal Society of Chemistry 2014

View Article Online
Communication
ChemComm
7 (a) S. Blechert, R. Knier, H. Schroers and T. Wirth, Synthesis, 1995,
592–604; (b) M. G. Banwell, B. D. Bissett, S. Busato, C. J. Cowden,
D. C. R. Hockless, J. W. Holman, R. W. Read and A. W. Wu, J. Chem.
Soc., Chem. Commun., 1995, 2551–2553; (c) K. C. Nicolaou, S. M. Dalby
and U. Majumder, J. Am. Chem. Soc., 2008, 130, 14942–14943.
8 (a) D. C. Behenna, Y. Liu, T. Yurino, J. Kim, D. E. White, S. C. Virgil
and B. M. Stoltz, Nat. Chem., 2012, 4, 130–133. For reviews on related
studies, see: (b) J. T. Mohr and B. M. Stoltz, Chem. – Asian J., 2007, 2,
1476–1491; (c) A. Y. Hong and B. M. Stoltz, Eur. J. Org. Chem., 2013,
2745–2759.
overall yield (previous synthesis reported by Shao: from N-benzyl-
carbazolone through 10 steps in 3.6% overall yield3).
This work was supported in part by a Grant-in Aid for
Scientific Research from the Ministry of Education, Culture,
Sports, Science and Technology, Japan, for which we are
thankful. We thank Daicel Corporation for help with screening
the conditions for HPLC analysis of kopsihainanine A (1).
9 See ESI† for details.
10 The structures of compounds 14, 15 and 16 could not be clearly
established by 1H NMR analysis because (i) compounds 14 and 15
showed the complex spectra due to their rotamers and (ii) 16 was
obtained as a mixture of two diastereoisomers. Therefore, the yields
for each step were not calculated.
11 (a) B. Plietker and M. Niggemann, J. Org. Chem., 2005, 70, 2402–2405;
(b) P. Tiwari and A. K. Misra, J. Org. Chem., 2006, 71, 2911–2913;
(c) N. M. Neisius and B. Plietker, J. Org. Chem., 2008, 73, 3218–3227.
Notes and references
1 J. Chen, J.-J. Chen, X. Yao and K. Gao, Org. Biomol. Chem., 2011, 9,
5334–5336.
2 P. Jing, Z. Yang, C. Zhao, H. Zheng, B. Fang, X. Xie and X. She,
Chem. – Eur. J., 2012, 18, 6729–6732.
3 Z. Li, S. Zhang, S. Wu, X. Shen, L. Zou, F. Wang, X. Li, F. Peng,
H. Zhang and Z. Shao, Angew. Chem., Int. Ed., 2013, 52, 4117–4121.
4 C. J. Gartshore and D. W. Lupton, Angew. Chem., Int. Ed., 2013, 52, 12 For the synthesis of ligand 20, see: (a) M. R. Krout, J. T. Mohr and
4113–4116.
B. M. Stoltz, Org. Synth., 2009, 86, 181–193; (b) N. T. McDougal,
J. Streuff, H. Mukherjee, S. C. Virgil and B. M. Stoltz, Tetrahedron
Lett., 2010, 51, 5550–5554.
5 (a) C. Mukai and Y. Takahashi, Org. Lett., 2005, 7, 5793–5796;
(b) M. Mizutani, F. Inagaki, T. Nakanishi, C. Yanagihara, I. Tamai
and C. Mukai, Org. Lett., 2011, 13, 1796–1799.
6 For reviews on transition-metal-catalyzed decarboxylative allylation,
see: (a) J. D. Weaver, A. Recio III, A. J. Grenning and J. A. Tunge,
Chem. Rev., 2011, 111, 1846–1913; (b) S. Oliver and P. A. Evans,
Synthesis, 2013, 3179–3198.
13 HPLC analysis of both compounds (+)-12, derived from the Bischler–
Napieralski reaction, and the precursor (À)-18 for the final product
was also carried out just for confirmation. As predicted, no racemi-
zation was observed at these stages: (+)-12 shows 98% ee, while
(À)-18, 99% ee.
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 5782--5785 | 5785