Preparation and a novel rearrangement reaction of 1,2,3,4-tetrahydro-9- hydroxy-β-carboline, and their applications for the total synthesis of (±)- coerulescine

Article abstract of DOI:10.3987/COM-99-8743

Novel 9-hydroxy-β-carboline derivatives were produced for the first time. A novel rearrangement reaction of 1,2,3,4-tetrahydro-9-hydroxy-β- carbolines was discovered to give 3,3-disubstituted oxindoles, which was successfully applied to the total synthesis of (±)-coerulescine.

Full text of DOI:10.3987/COM-99-8743

HETEROCYCLES, Vol. 53, No. 1, 2000, pp. 7 - 10, Received, 27th August, 1999
PREPARATIONANDANOVELREARRANGEMENTREACTIONOF1,2,3,4-
TETRAHYDRO-9-HYDROXY-β-CARBOLINE,AND THEIRAPPLICATIONS
1
FOR THE TOTAL SYNTHESIS OF (±)-COERULESCINE
Masanori Somei,* Koichi Noguchi, Ryutaro Yamagami, Yumiko Kawada, Koji
Yamada, and Fumio Yamada
Faculty of Pharmaceutical Sciences, Kanazawa University,
13-1 Takara-machi, Kanazawa 920-0934, Japan
Abstract — Novel 9-hydroxy-β-carboline derivatives were produced for the first
time. A novel rearrangement reaction of 1,2,3,4-tetrahydro-9-hydroxy-β-carbolines
was discovered to give 3,3-disubstituted oxindoles, which was successfully
applied to the total synthesis of (±)-coerulescine.
2
1-Hydroxyindoles undergo various types of reactions depending on the structures and reaction conditions.
We have designed 1,2,3,4-tetrahydro-9-hydroxy-β-carbolines (1a–d,
Scheme 1) as key substrates to elucidate whether the rearrangement of
3a
Figure 1
the 2-substituent to the 3-position
3
or vice versa, predicted by our
3
3
R
1-hydroxyindole hypotheses, occurs or no as illustrated in a general
Rearrangement
2
1
2
N
OR
R
formula A of 1-hydroxyindoles (Figure 1).
4
1
+
Quite recently, Colegate and co-workers
isolated and determined
H
A
(–)-coerulescine (2c) from Phalaris coerulescens. We have noticed
that Phalaris species generally contain 1,2,3,4-tetrahydro- β−car-
3a
5
bolines (3a,b), because our hypotheses assumed biosyntheses of the related alkaloid, (–)-horsfiline (2b)
and 2c from 3b and 3a, respectively, through the corresponding 9-hydroxy-β-carbolines (1b and 1c).
Now, we wish to report successful preparations of 1a and 1c, and their novel rearrangement achieving a
total synthesis of 3,3-disubstituted oxindole alkaloid, (±)- coerulescine (2c).
We have already developed the preparation of 1,2,3,4-tetrahydro-9-hydroxy-2-methoxycarbonyl-β-car-
6
boline (1d) from the corresponding 1,2,3,4,4a,9a-hexahydro compound (4c) by Na WO ·2H O–30%
2
H O method. According to the procedure, oxidation of 4a was examined and typical results are summa-
4
2
7
2 2
rized in Table 1. As can be seen from the Table, reaction time was found to be an important factor for pre-
paring 1a . In cases (Entries 1 and 2) where reaction times were 25 min, even a trace amount of 1a was not
detected in the reaction mixture, instead 9-hydroxy-β-carboline (5), 3,4-dihydro-9-hydroxy-β-carboline N-
oxide (6), and 9-hydroxy-β-carboline N-oxide (7) were produced. When the reaction time was longer than
25 min, total yield of the three 9-hydroxy-β-carbolines (5–7) decreased, while the production of 1a was
observed in 11% yield when it was shortened to 10 min (Entry 3). Better yield of 1a (55%) was attained by
reacting 4a for only 5 min as shown in Entry 4. As for oxidizing reagent, urea hydrogen peroxide was a
8
reagent of choice to give satisfactory result (65%, Entry 5) as discovered in our previous report. Similarly,

Scheme 1
5
4
1
H
H
R1
NR2
R1
R
3
a
b
NR2
2
NMe
9
N
OH
NR
N
H
N
H
O
N
H
1
3 a) R = H
4
2
b) R = OMe
a) R1 = H, R2 = H
a) R1 = H, R2 = H
a) R = H
b) R = Me
c) R = COOMe
b) R1 = OMe, R2 = Me
c) R1 = H, R2 = Me
d) R1 = H, R2 = COOMe
b) R1 = OMe, R2 = Me
c) R1 = H, R2 = Me
c
d
d) R1 = H, R2 = COOMe
N
N
H
O
8
b
1a
1c
e
1d
2a + 9 + 10
2c + 11 + 15a
15b
b
b or e
N
N
H
9
NMeAc
f
Cl
NMe
N
NR
N
R
CHO
N
H
H
N
OH
N
H
15
11
12
a) R = H
a) R = Me
b) R = COOMe
10
b) R = Ac
a) Na WO ·2H O, 30% H O ; b) MeOH-c-HCl (1:3, v/v), reflux;
2
4
2
2 2
c) HCHO/AcOH/NaBH CN; d) ClCOOMe, Et N; e) MeOH-c-HCl,
3
3
under various reaction conditions; f) Ac O, pyridine.
2
Table 1
MCPBA
DDQ
+
Na2WO4・2H2O
(0.2 mol eq.)
+
+
4a
NH
N
N
N
N
OH
Oxidizing
Reagent
N
OH
N
OH
N
OH
O
O
7
1a
5
6
Reaction Conditions
Yield (%) of
Oxidizing
Reagent
(mol eq)
Solvent
Temp.
(°C)
Time
(min)
Entry
System
1a
5
6
7
30% H2O2
(10)
MeOH-H2O
(10:1, v/v)
1
2
3
4
5
19
22
21
27
7
25
25
0
0
14
5
21
25
t-BuOH-H2O
(4:1, v/v)
21
15
5
"
"
"
10
5
trace
trace
11
55
22
5
"
MeOH-H2O
(9:1, v/v)
98% Urea·H2O2
(10)
5
65
trace
5
2
25.5
"

oxidation of 1,2,3,4,4a,9a-hexahydro-2-methyl-β-carboline (4b) afforded 1,2,3,4-tetrahydro-9-hydroxy-
2-methyl-β-carboline (1c) in 69% yield.
On the other hand, oxidation of 5 with MCPBA afforded 7 in 29% yield in addition to β-carboline N-ox-
9
ide (8) and β-carboline (9) in 31 and 11% yields, respectively. Dehydrogenation of 6 with DDQ provided
75% yield of 7. Interestingly, treatment of 6 with chloroacetyl chloride produced 8 and 9 in 60 and 33%
yields, respectively. Thus, four novel 9-hydroxy-β-carboline compounds (1a and 5–7) and 8 are now
readily available. Since these compounds are useful building blocks for β-carboline alkaloids syntheses,
their applications are currently in progress.
With 1a,c and 1d in hand, we treated them with MeOH-c-HCl (1:3,v/v). Although no reaction took place at
room temperature, the desired reaction of 1a occurred under refluxing for 1 h resulting in the formation of
2a in 47% yield together with 9 and 10 in 2 and 36% yields, respectively. Subsequent methylation of 2a
with HCHO-AcOH-NaBH CN afforded (±)-2c in 91% yield, while the reaction of 2a with methyl
3
chloroformate gave 2d in 99% yield. Similarly, treatment of 1cwith refluxing MeOH-c-HCl (1:3,v/v) for 3
h produced (±)-2c, 11, and 6-chloro-1,2,3,4-tetrahydro-2-methyl-β-carboline (15a) in 42, 46, and 9%
yields, respectively.
The structure of (±)-2c was confirmed by the direct comparison with the authentic sample prepared
4c
alternatively according to the reported procedure. Spectral data of our samples, authentic sample, and the
reported ones of (–)-2c were identical in every respects. The structure of relatively unstable 11 was deduced
based on its spectral data and the fact that it yielded formyl compounds (12a and 12b) by the reaction with
Ac O-pyridine.
2
The mechanism of the rearrangement would be explained as shown in Scheme 2. Initial protonation of 9-
hydroxy group, followed by its elimination and chloride attack at the 4a carbon generates chloroindolenine
(13). After addition of water to the 9a imine carbon atom of 13 giving 14, concerted elimination of chloride
and rearrangement of the alkyl side chain attached to the 9a carbon atom result in 3,3-disubstituted oxindole
structure(2a).
Scheme 2
Cl–
+
NH2
Cl
Cl
H3O+
Cl–
4a
9a
NH2
+
NH2
+
NH2
+
N
N
H
N
H
O
N
OH
OH
OH2
H+
H+
Protonated1a
13
14
Protonated 2a
It should be noted that 2d was not formed in the reaction of 1d with methanolic HCl under various examin-
ed reaction conditions. In these cases, nucleophilic substitution reaction took place affording 6-chloro-1,2,
3,4-tetrahydro-2-methoxycarbonyl-β-carboline (15b, 8–16%) together with 1,2,3,4-tetrahydro-2-meth-
10
oxycarbonyl-β-carboline (21–30%).
In conclusion, we could prepare novel 9-hydroxy-β-carbolines which would be utilized as useful building
blocks for the syntheses of β-carboline compounds. We have also discovered a novel rearrangement

reaction on 1-hydroxyindole skeletons providing valuable means for the preparation of 3,3-disubstituted
oxindole alkaloids. The present findings suggest chemists to be very careful to use acids for isolation of
indole alkaloids from natural resources, otherwise oxindole alkaloids might be obtained as artifacts of
1b,2b,3a
genuine 1-hydroxy- or 1-methoxyindole alkaloids, if they happen to exist.
ACKNOWLEDGMENT
This work is supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education,
Science, Sports, and Culture, Japan, which is gratefully acknowledged.
REFERENCES AND NOTES
1. a) This is Part 95 of a series entitled “The Chemistry of Indoles”. b) Part 94: M. Hasegawa, K. Yamada,
Y. Nagahama, and M. Somei, Heterocycles, 2000, submitted. All new compounds gave satisfactory
spectral and elemental analysis or high-resolution MS data for crystals or gums, respectively. 1a, mp
171—175°C (decomp); 1c, mp 181—183°C (decomp); 1d, gum; 2a, gum; 2c, mp 111—112°C; 5, mp
205.0—206.5°C; 6, mp 193.5—195.0°C (decomp); 7, mp 254.0—265.5 (decomp); 11, mp 179.5—
181.0°C; 15a, mp 207—209°C (decomp); 15b, mp 194—195°C.
2. a) M. Somei, N. Oshikiri, M. Hasegawa, and F. Yamada, Heterocycles, 1999, 51, 1237; b) Review;
M. Somei, ibid., 1999, 50, 1157 and references cited therein.
3. a) Review; M. Somei, J. Synth. Org. Chem. Jpn., 1991, 49, 205; b) M. Somei and M. Natsume,
Tetrahedron Lett., 1973, 2451; c) M. Somei and Y. Fukui, Heterocycles, 1993, 36, 1859. See also
reference 2b.
4. a) N. Anderton, P. A. Cockrum, S. M. Colegate, J. A. Edgar, K. Flower, I. Vit, and R. I. Willing,
Phytochemistry, 1998, 48, 437. b) S-I. Bascop, J. Sapi, J-Y. Laronze, and J. Levy, Heterocycles,
ꢀ
1994, 38, 725. c) M. E. Kuehne, D. M. Roland, and R. Hafter, J. Org. Chem., 1978, 43, 3705.
5. C. Pellegrini, C. Strassler, M. Weber, and H-J. Borschberg, Tetrahedron: Asymmetry, 1994, 5, 1979;
ꢀ
ꢀ
G. Lakshmajah, T. Kawabata, M. Shang, and K. Fuji, J. Org. Chem., 1999, 64, 1699. See also
reference 4b.
6. M. Somei, K. Kobayashi, K. Tanii, T. Mochizuki, Y. Kawada, and Y. Fukui, Heterocycles, 1995, 40,
119.
7. M. Somei and T. Kawasaki, Heterocycles, 1989, 29, 1251; M. Somei, T. Kawasaki, K. Shimizu, Y.
ꢀ
ꢀ
Fukui, and T. Ohta, Chem. Pharm. Bull., 1991, 39, 1905. See also references 2b and 8. Our synthetic
method for 1-hydroxy- and 1-alkoxytryptophans has been utilized by chemists without citing us fairly.
One recent example is the following. D. L. Boger, H. Keim, B. Oberhauser, E. P. Schreiner, and C. A.
Foster, J. Am. Chem. Soc., 1999, 121, 6197.
8. M. Somei and T. Kobayashi, Heterocycles, 1992, 34, 1295.
9. Synthesis of 8 from 9 in 78% yield: M. Natsume and R. Tanabe, Ann. Rept. ITSUU Lab., 1968, 15,
ꢀ
21 and references cited therein.
10. Remarkable differences observed in the reactivities of 1a,c and 1d might be explained by the protonation
of β-nitrogen. The detailed elucidation of the results is an interesting future subject.