Synthesis of NK109, an anticancer benzo[c]phenanthridine alkaloid

Article abstract of DOI:10.1021/jo9718758

A total synthesis of NK109 (7-hydroxy-8-methoxy-5-methyl-2,3- methylenedioxybenzo[c]phenanthridinium hydrogensulfate dihydrate), an anticancer benzo[c]phenanthridine alkaloid, is reported. The primary structure of this compound was erroneously communicated in 1973 as fagaridine (from Fagara xanthoxyloides) which the 8-hydroxy regioisomer. NK109 has not yet been isolated from a natural source and therefore can only be obtained by synthesis. To study a wide variety of analogues, we decided to use a synthetic route via substituted benzylamine 5, which was obtained from the appropriate benzaldehyde ad naphthylamine units. The benzo[c]phenanthridine ring was conducted by radical cyclization with tri-n-octyltin hydride and 2,2'-azobis(2)-methylbutyronitrile], followed by oxidative aromatization with MnO₂. The resulting benzo[c]phenanthridine 6 was successfully methylated with methyl 2-nitrobenzenesulfonate. After deprotection of the benzyl group and subsequent hydration, NK109 was obtained. All reactions were performed under normal conditions. Purification was achieved only by recrystallization to give an overall yield of 40%.

Full text of DOI:10.1021/jo9718758

J . Org. Chem. 1998, 63, 4235-4239
4235
Syn th esis of NK109, a n An tica n cer Ben zo[c]p h en a n th r id in e
Alk a loid
Takeshi Nakanishi,* Masanobu Suzuki, Akihiro Mashiba, Keizou Ishikawa, and
Takashi Yokotsuka
Pharmaceuticals Group, Nippon Kayaku Company, Ltd., 31-12, Shimo 3-Chome, Kita-ku,
Tokyo 115, J apan
Received October 10, 1997
A total synthesis of NK109 (7-hydroxy-8-methoxy-5-methyl-2,3-methylenedioxybenzo[c]phenan-
thridinium hydrogensulfate dihydrate), an anticancer benzo[c]phenanthridine alkaloid, is reported.
The primary structure of this compound was erroneously communicated in 1973 as fagaridine (from
Fagara xanthoxyloides) which is the 8-hydroxy regioisomer. NK109 has not yet been isolated from
a natural source and therefore can only be obtained by synthesis. To study a wide variety of
analogues, we decided to use a synthetic route via substituted benzylamine 5, which was obtained
from the appropriate benzaldehyde and naphthylamine units. The benzo[c]phenanthridine ring
was constructed by radical cyclization with tri-n-octyltin hydride and 2,2′-azobis(2-methylbutyro-
nitrile), followed by oxidative aromatization with MnO2. The resulting benzo[c]phenanthridine 6
was successfully methylated with methyl 2-nitrobenzenesulfonate. After deprotection of the benzyl
group and subsequent hydration, NK109 was obtained. All reactions were performed under normal
conditions. Purification was achieved only by recrystallization to give an overall yield of 40%.
In tr od u ction
erythrine exhibited cytotoxicity in vitro but had no
significant antitumor activity in vivo.4 Various analogues
of these compounds have been synthesized, but none have
had greater antitumor activity than the respective natu-
ral compounds.5 Therefore, these compounds have not
been the subject of a clinical evaluation. Nevertheless,
there have still been many studies in this area and
several authors have reported interesting chemical find-
ings and biological effects.6
Benzo[c]phenanthridines are naturally occurring al-
kaloids with various biological activities.1 In particular,
benzo[c]phenanthridinium salts such as nitidine, fagar-
onine, sanguinarine, and chelerythrine display antitumor
properties.2 These compounds were isolated from Ruta-
Since 1987, we have sought a new benzo[c]phenan-
thridinium salt with anticancer activity. At an early
stage, 18 compounds were kindly supplied by Prof.
Hanaoka, Kanazawa University, J apan; these were
synthesized by the transformation of protoberberine
alkaloid.7 After biological screening of the samples, we
found that compound 1a had significant antitumor activ-
ity.8 Therefore, we developed a synthetic method more
practical than the protoberberine transformation. Dur-
ing this process, we discovered that 1b (7-hydroxy-8-
methoxy-5-methyl-2,3-methylenedioxybenzo[c]phenan-
ceous and Papaveraceous plants. Nitidine and fagar-
onine were expected to be useful antitumor drugs because
of their strong activity against several tumor cells.
Preclinical studies of these compounds were performed
in detail at the National Cancer Institute in the mid
1970s.3 However, they showed that these compounds
possessed a narrow antitumor spectrum as well as a
certain toxicity and instability. Sanguinarine and chel-
(4) (a) Stermitz, F. R.; Larson, K. A.; Kim, D. K. J . Med. Chem. 1973,
16, 939. (b) Caolo, M. A.; Stermitz, F. R. Heterocycles 1979, 12, 11.
(5) (a) Zee-Cheng, R. K.-Y.; Cheng, C. C. J . Med. Chem. 1975, 18,
66. (b) Stermitz, F. R.; Gillespie, J . P.; Amoros, L. G.; Romero, R.;
Stermitz, T. A. J . Med. Chem. 1975, 18, 708. (c) Zee-Cheng, R. K.-Y.;
Yan, S.-J .; Cheng, C. C. J . Med. Chem. 1978, 21, 199. (d) Cushman,
M.; Mohan, P.; Smith, E. C. R. J . Med. Chem. 1984, 27, 544. (e)
Cushman, M.; Mohan, P. J . Med. Chem. 1985, 28, 1031. (f) Ishii, H.;
Ichikawa, Y.; Kawanabe, E.; Ishikawa, M.; Ishikawa, T.; Kuretani, K.;
Inomata, M.; Hoshi, A. Chem. Pharm. Bull. 1985, 33, 4139.
(6) (a) Pezzuto, J . M.; Antosiak, S. K.; Messmer, W. M.; Slaytor, M.
B.; Honig, G. R. Chem.-Biol. Interact. 1983, 43, 323. (b) Kusumoto, I.
T.; Hattori, M.; Miyaichi, Y.; Tomimori, T.; Hanaoka, M.; Namba, T.
Shoyakugaku Zasshi 1991, 45, 240. (c) Mart´ın, G.; Guiti´ın, E.; Castedo,
L.; Saa, J . M. J . Org. Chem. 1992, 57, 5907. (d) Larsen, A. K.;
Grondard, L.; Couprie, J .; Desoize, B.; Comoe, L.; J ardillier, J .-C.; Riou,
J .-F. Biochem. Pharmacol. 1993, 46, 1403. (e) Dosta´l, J .; Ta´borska´,
E.; Slav´ık, J .; Pota´cˇek, M.; Hoffmann, E. J . Nat. Prod. 1995, 58, 723.
(7) (a) Hanaoka, M. The Alkaloids; Academic: New York, 1988; Vol.
33, p 170. (b) Hanaoka, M.; Mukai, C. Stud. Nat. Prod. Chem. 1994,
14, 769.
(1) Sˇima´nek, V. The Alkaloids, Academic: New York, 1985; Vol. 26,
p 185.
(2) (a) Arthur, H. R.; Hui, W. H.; Ng, Y. L. Chem. & Ind. (London)
1958, 1514. (b) Wall, M. E.; Wani, M. C.; Taylor, H. L. Abstracts of
Papers, 162nd National Meeting of American Chemical Society;
American Chemical Society: Washington, DC, 1971; MEDI34. (c)
Messmer, W. M.; Tin-Wa, M.; Fong, H. H. S.; Bevelle, C.; Farnsworth,
N. R.; Abraham, D. J .; Troja´nek, J . J . Pharm. Sci. 1972, 61, 1858. (d)
Gadamer, J .; Stickel, A. Arch. Pharm. 1924, 262, 452. (e) Gadamer, J .
Arch. Pharm. 1920, 258, 148.
(3) (a) Phillips, S. D.; Castle, R. N. J . Heterocycl. Chem. 1981, 18,
223. (b) Suffness, M.; Cordell, G. A. The Alkaloids; Academic: New
York, 1985; Vol. 25, p 178.
(8) Hanaoka, M.; Ekimoto, H.; Kobayashi, M. F.; Irie, Y.; Takahashi,
K. Chem. Abstr. 1989, 116, 718.
S0022-3263(97)01875-6 CCC: $15.00 © 1998 American Chemical Society
Published on Web 06/05/1998

4236 J . Org. Chem., Vol. 63, No. 13, 1998
Nakanishi et al.
Ta ble 1. Str ess Test of NK109a
Sch em e 1
purity based on the area under
the HPLC curve (%) and powder color
chloride
(Cl-)
hydrogensulfate
-
time, h
(HSO4 )
0
10
99.45 (orange)
98.78 (amber)
99.47 (yellow)
99.40 (yellow)
a
Test condition: stirred at 70 °C under reduced pressure (3
thridinium hydrogensulfate dihydrate) had long-term
stability for clinical use. Compound 1b, which consists
of the biologically active benzo[c]phenanthridinium cat-
ion, hydrogensulfate as a good counteranion, and ad-
ditional dihydrate, was identified as NK109. In more
detailed studies, we found that NK109 had greater
antitumor activity than any known benzo[c]phenanthri-
dinium in several assay systems.9 NK109 retains full
activity against cisplatin- and multidrug-resistant tumor
cells.10 Therefore, clinical studies have been carried out
in J apan.
In 1973, Torto and co-workers identified a new benzo-
[c]phenanthridinium salt, fagaridine, and assigned its
structure as 1c, a different salt form of NK109.11 Sub-
sequently, we established that the true structure of
fagaridine is not 1c, but rather its isomer 2. This
assignment is based on the comparison of physical and
chemical data.12 Thus, NK109 has never been isolated
from natural sources but is a synthetic benzo[c]phenan-
thridinium salt.
mmHg).
gensulfate seemed to be completely stable under usual
storage conditions. Consequently, we selected the latter.
Since sulfuric acid is not volatile, 1′ was not formed
unless it was neutralized. The hydrogensulfate form of
the acid residue was supported by an elemental analysis.
Syn th etic In vestigation . Many synthetic approaches
to benzo[c]phenanthridine alkaloid have been reported.13
However, some of these were complex, gave a low total
yield, and were limited to a particular substituent
pattern. We developed a better method for large-scale
synthesis and a route suitable for an analogue study, as
outlined in Scheme 2. The starting materials were
6-bromo-2-hydroxy-3-methoxybenzaldehyde 3a derived
from o-vanillin14 and 5-amino-2,3-methylenedioxynaph-
thalene 4 derived from 2,3-dihydroxynaphthalene.5b,15
Syn th esis of P r ecu r sor 5. Hydroxy aldehyde 3a was
protected to benzyloxy aldehyde 3b. The reaction was
carried out with benzyl chloride and potassium iodide in
a basic medium. Benzyl bromide was also available but
was not used because of its higher cost. The aldehyde
3b and the amine 4 were condensed to a Schiff base,16
and its imine portion was reduced to a corresponding
amine in a good yield. The resulting 5 is a precursor of
the benzo[c]phenanthridine ring. This reduction was
successfully performed with dimethylamine-borane17
instead of conventional sodium cyanoborohydride or
sodium borohydride. With sodium borohydride, a higher
temperature and too much reducing agent were needed.
Consequently, a large amount of hydrogen gas evolved
until the termination of the reaction. Sodium cyanoboro-
hydride caused a side reaction, that is, debromination.
Dimethylamine-borane is a preferable reducing agent
that does not give hydrogen gas and does not cause
debromination.
In this paper, we examine the effect of the acid residue
on stability and present a method for synthesizing
NK109.
Resu lts a n d Discu ssion
E ffect of Acid R esid u e on St a b ilit y. The benzo-
[c]phenanthridinium salts 1 are at equilibrium between
the phenol 1 and its ketone 1′ (Scheme 1). The pKa value
at equilibrium is 5.3. Accordingly, 1 could gradually lose
its acid residue in long-term storage, if it was the salt of
a volatile acid such as hydrochloric acid or hydroiodic
acid. Salt-free 1′ showed relatively low solubility in
aqueous solution and was unstable under an oxidative
atmosphere. Fixation to the quaternary salt 1 is impor-
tant for maintaining solubility. In addition, the acid
residue also affects stability. Experimental data are
shown in Table 1. In the case of chloride, the solid
changed from orange to amber, along with a decrease in
purity. Although this change was rather small, it would
be unsuitable for medical use. In contrast, the hydro-
Syn th esis of Ben zo[c]p h en a n th r id in e (6). Benzo-
[c]phenanthridine 6 was obtained by radical cyclization
of the precursor 5 and subsequent oxidative aromatiza-
tion of its cyclized intermediate. Cyclization to the benzo-
[c]phenanthridine ring system is a key step of the
(13) (a) Kametani, T.; Kigasawa, K.; Hiiragi, M.; Kusama, O. J .
Heterocycl. Chem. 1973, 10, 31. (b) Zee-Cheng, K.-Y.; Cheng, C. C. J .
Heterocycl. Chem. 1973, 10, 85. (c) Ishii, H.; Chen, I.-S.; Ishikawa, T.
Chem. Pharm. Bull. 1983, 31, 2963. (d) Gillespie, J . P.; Amoros, L. G.;
Stermitz, F. R. J . Org. Chem. 1974, 39, 3239. (e) Sainsbury, M.; Dyke,
S. F.; Moon, B. J . J . Chem. Soc. C 1970, 1797. (f) Kessar, S. V.; Singh,
G.; Balakrishnan, P. Tetrahedron Lett. 1974, 2269. (g) Ninomiya, I.;
Naito, T.; Ishii, H.; Ishida, T.; Ueda, M.; Harada, K. J . Chem. Soc.,
Perkin Trans. 1 1975, 762. (h) Begley, W. J .; Grimshaw, J . J . Chem.
Soc., Perkin Trans. 1 1977, 2324. (i) Cushman, M.; Cheng, L. J . Org.
Chem. 1978, 43, 286. (j) Olugbade, T. A.; Waigh, R. D. J . Pharm.
Pharmacol. 1981, 33, 81P. (k) Onda, M.; Yamaguchi, H. Chem. Pharm.
Bull. 1979, 27, 2076.
(9) Under preparation.
(14) Kametani, T.; Honda, T.; Inoue, H.; Fukumoto, K. J . Chem. Soc.,
Perkin Trans. 1 1976, 1221.
(15) Clark, J . H.; Holland, H. L.; Miller, J . M. Tetrahedron Lett.
1976, 41, 3361.
(10) Kanzawa, F.; Nishio, K.; Ishida, T.; Fukuda, M.; Kurokawa,
H.; Fukumoto, H.; Nomoto, Y.; Fukuoka, K.; Bojanowski, K.; Saijo, N.
Br. J . Cancer 1997, 76, 571.
(11) Torto, F. G.; Mensah, I. A.; Baxter, I. Phytochemistry 1973, 12,
2315.
(12) Submitted for publication.
(16) Castellano, J . A.; Goldmacher, J . E.; Barton, L. A.; Kane, J . S.
J . Org. Chem. 1968, 33, 3501.
(17) Billman, J . H.; McDowell, J . W. J . Org. Chem. 1961, 26, 1437.

Synthesis of NK109
J . Org. Chem., Vol. 63, No. 13, 1998 4237
Sch em e 2
Ta ble 2. Mod e of Cycliza tion w ith n -Bu 3Sn H a n d AIBNa
5 was consumed by radical species generated from the
initiator. In the absence of trialkyltin hydride, 5 was
decomposed and no cyclized compound was obtained
(entry 1). In the presence of trialkyltin hydride, stannum
radical arose and led to the formation of the cyclized
intermediate, which was easily converted to compound
6 by active manganese dioxide oxidation (entries 2-4).
A small amount of the initiator caused decomposition
rather than cyclization (entry 3). When both reagents
were divided into several portions, the decomposition rate
of 5 was high (entry 2). These results indicate that 5
was consumed throughout the reaction. Accordingly, a
sufficient amount of the initiator should be added all at
once to a hot solution of 5 and trialkyltin hydride (entry
4). Rapid conversion into the benzo[c]phenanthridine
ring was essential for avoiding extreme decomposition.
peak area
of HPLC (%)e
n-Bu3SnHb
(equiv)
AIBNb
(equiv)
temp time
others
(dec)
entry
(°C) (min) 6f
5
1
2
3
4
0
2
100
100
120
110
60
180
120
15
0
15
30 19
90
20
1
80
84
51
10
0.6 × 5c
0.5 × 5c
0.1 × 7d
2
3
3
0
a
HPLC conditions: column, LQ Pack SG-C8; mobile phase,
0.1% H3PO4-MeOH (30:70); flow rate, 1.0 mL/min; detection, UV
b
(330 nm). Added to a hot solution of 5 (246 mg) in toluene (25
d
mL). c Added at 30-min intervals. Added at 15-min intervals.
e This value does not indicate the actual yield (%), but rather shows
the area (%) under the HPLC chromatograph obtained under the
above condition. f They were analyzed after oxidation of cyclized
intermediate. Reaction conditions: the reaction mixture (100 µL)
was diluted with toluene (200 µL), MnO2 was added (10 mg), and
the solution was stirred for 5 min and then filtered through a 0.45
µm membrane filter (DISMIC-13HP, ADVANTEC, J apan).
During our study, we confirmed that AMBN, 2,2′-
azobis(2-methylbutyronitrile), and tri-n-octyltin hydride
can be used in the same manner as AIBN and tri-n-
butyltin hydride. These reagents offer some advan-
tages: AMBN is highly soluble in the reaction solvent,
that is, toluene, and tri-n-octyltin hydride is less toxic.
We next sought to determine appropriate amounts of
these reagents and a suitable reaction temperature
(Table 3). This reaction did not proceed sufficiently at
80 °C, and a large amount of 5 remained (entry 1).
Successful cyclization required a reaction temperature
of over 100 °C. Although the highest yield was obtained
at 100 °C (entry 2), this condition gave impurities that
could not be easily removed by recrystallization. There-
fore, we concluded that the best temperature for this
reaction was 110 °C (entry 3). We next examined the
appropriate amounts of the reagents. Entry 6 shows a
high yield of 6, but impurities formed as in entry 2. In
conclusion, we decided that the best condition was 2.0
equiv of trialkyltin hydride and 1.5 equiv of the initiator
at 110 °C (entry 5).
synthesis. Other cyclization methods have been de-
scribed by other investigators, for example, benzyne
cyclization18 and photoradical cyclization.19 However, the
former method requires an extremely low temperature
and the latter needs a special photoreaction apparatus.
A method similar to our synthesis was reported by Rosa
and co-workers using the radical initiator AIBN (2,2′-
azobisisobutyronitrile).20 Independently, we examined
radical cyclization for large-scale synthesis. Table 2
shows the experimental results of the mode of radical
cyclization by HPLC analysis. We found that compound
(18) (a) Kessar, S. V.; Gopal, R.; Singh, M. Tetrahedron 1973, 29,
167. (b) Kessar, S. V.; Gupta, Y. P.; Balakrishnan, P.; Sawal, K. K.;
Mohammad, T.; Dutt, M. J . Org. Chem. 1988, 53, 1708.
(19) (a) Kessar, S. V.; Gupta, Y. P.; Dhingra, K.; Sharma, G. S.;
Narula, S. Tetrahedron Lett. 1977, 17, 1459. (b) Sˇmidrkal, J . Collect.
Czech. Chem. Commun. 1984, 49, 1412.
(20) Rosa, A. M.; Prabhakar, S.; Lobo, A. M. Tetrahedron Lett. 1990,
31, 1881.

4238 J . Org. Chem., Vol. 63, No. 13, 1998
Nakanishi et al.
Ta ble 3. Cycliza tion Con d ition s w ith n -Oct3Sn H a n d
simultaneously. With methyl o-nitrobenzenesulfonate
7,22 compound 6 was successfully methylated to benzo-
[c]phenanthridinium 8. Reagent 7 was easily prepared
from o-nitrobenzenesulfonyl chloride and sodium meth-
oxide in methanol. Compound 7 is not commonly used
in methylation reactions but has 6-fold greater activity
than conventional dimethyl sulfate.23 Moreover, since
the resulting benzo[c]phenanthridinium o-nitrobenzen-
sulfonate 8 is barely soluble in toluene, it gradually
precipitated and was removed from the reactive liquid
phase. Thus, there was no further reaction.
AMBNa
yield (%)b
n-Oct3SnH
(equiv)
AMBN
(equiv)
temp
(°C)
purity
(%)c
entry
6
5
1
2
3
4
5
6
7
2.0
2.0
2.0
2.0
2.0
1.5
1.2
2.0
2.0
2.0
2.0
1.5
1.2
1.2
80
100
110
120
110
110
110
5.6
61.6
55.2
51.5
56.6
57.6
47.1
56.3
0
0
0
0
80.0
98.9
95.3
97.8
8.4
12.2
a
Reaction conditions: 5 (492 mg) and n-Oct3SnH in toluene (49
mL) were heated a each experimental temperature. An AMBN
toluene solution (minimum volume needed to dissolve) was then
added to the hot reaction mixture and heated for 1 h. The
subsequent solution was oxidized with MnO2 at room temperature
for 1 h. b Calculated yield by HPLC analysis using isolated samples
as a standard. HPLC conditions: column, LQ Pack ODS-M15;
mobile phase, 0.1% H3PO4-CH3CN (4:6); flow rate, 1.5 mL/min;
detection, UV (270 nm). c Purity of precipitate 6 after primary
purification.
Syn th esis of NK109 (1b). Compound 8 was neutral-
ized with sodium hydroxide in ethanol to remove the acid
residue and gave the pseudobase 9.24 This was treated
with sulfuric acid for deprotection of the benzyl group to
give hydrogensulfate 10. Subsequent hydration yielded
NK109 (1b).
Con clu sion
Sch em e 3
We have developed a versatile and convenient method
for the synthesis of NK109, which exhibits superior
antitumor activity and is expected to be an invaluable
anticancer agent. Cyclization with n-Oct3SnH-AMBN
is an excellent method for constructing the benzo[c]-
phenanthridine ring system and does not require freezing
conditions. Methyl o-nitrobenzenesulfonate is an active,
useful, and readily available agent for methylation.
Furthermore, none of the steps required extreme condi-
tions. All purification could be carried out only by
recrystallization. The overall yield was approximately
40%. Thus, our procedure shows good productivity and
is both safe and simple. We have already produced
NK109 at a 2-kg scale for clinical studies. This method
may be adaptable to the synthesis of other benzo[c]-
phenanthridine alkaloids. In these synthetic studies we
also discovered several advantages of the hydrogensulfate
form of the acid residue of NK109.
Exp er im en ta l Section
We also studied the cyclization of 5 with LDA18
(Scheme 3). After lithiation of 5 at -78 °C, the reaction
vessel was rapidly brought to room temperature and 5
was converted into the cyclized product 6 (34% yield).
When lithiated 5 was warmed slowly, only the starting
material 5 was obtained (63% recovery). This result
suggests that a lithiated intermediate would be converted
into benzyne at a higher temperature to form a benzo-
[c]phenanthridine ring, whereas the lithium cation would
be slowly replaced by a proton at lower temperature. In
a large-scale synthesis, it is difficult to control the rate
of an increase in temperature. Therefore, this procedure
was not suitable for our purpose.
Syn th esis of Ben zo[c]p h en a n th r id in iu m (8). N-
Methylation of a pyridine ring or fused aromatics is easily
achieved by considering the intrinsic pKa value but is
often difficult in a steric environment.21 In benzo[c]-
phenanthridine 6, we met resistance to methylation due
to steric hindrance by the 4-position hydrogen. Methyl
p-toluenesulfonate, MeOTs, required extreme conditions,
that is, heating with neat MeOTs at 120 °C. Methyl
trifluoromethanesulfanate, MeOTf, was much more reac-
tive, but the undesired debenzylation of 6 proceeded
Gen er a l. 2,2′-Azobis(2-methylbutyronitrile) and tri-n-oc-
tyltin hydride were purchased from J apan Hydrazine Co. Inc.
and Sankyo Organic Chemicals Co. Ltd., respectively. Other
materials were obtained from commercial suppliers and were
used without further purification. NMR spectra were obtained
on a Varian Gemini-200 spectrometer. All chemical shifts are
reported in ppm relative to tetramethylsilane as an internal
standard (TMS, δ 0.00). Mass spectra were recorded on a
Micromass Limited Auto Spec-Q spectrometer. Elemental
analyses were performed on a Yanako CHN Corder (C, H, N
analysis) and a Yokogawa IC-7000D (Br, S analysis).
2-Ben zyloxy-6-br om o-3-m eth oxyben zaldeh yde (3b). To
a mixture of 6-bromo-2-hydroxy-3-methoxybenzaldehyde 3a
(350 g, 1.51 mol), KI (252 g, 1.51 mol), and anhydrous K2CO3
(523 g, 3.78 mol) in methanol (23 L) was added 671 g of benzyl
chloride (5.30 mol). After 7 h of reflux, anhydrous K2CO3 (209
g, 1.51 mol) was added and the solution was refluxed for
another 2 h and then cooled to room temperature. After
evaporation of methanol, the residue was partitioned between
toluene and water. The organic extract was washed with
water and dried over MgSO4 in a toluene solution containing
(22) Kiprianov, A. I.; Tolmachev, A. I. Zh. Obshch. Khim. 1959, 29,
2868: Chem. Abstr. 1960, 54, 12126.
(23) Kiprianov, A. I.; Tolmachev, A. I. Zh. Obshch. Khim. 1957, 27,
142: Chem. Abstr. 1957, 51, 12912.
(21) Schug, J . C.; Viers, J . W.; Seeman, J . I. J . Org. Chem. 1983,
48, 4892.
(24) (a) Sˇima´nek, V.; Preininger, V. Heterocycles 1977, 6, 475. (b)
Bunting, J . W. Adv. Heterocycl. Chem. 1979, 25, 1.

Synthesis of NK109
J . Org. Chem., Vol. 63, No. 13, 1998 4239
3b (net 467 g, 97% yield25). This solution was used for the
next reaction.
13C NMR (CDCl3) δ 57.6, 124.8, 128.9, 131.3, 132.2, 134.9; FAB-
MS m/z 218 (M+ + H). Anal. Calcd for C7H7NO5S: C, 38.71;
H, 3.25; N, 6.45; S, 14.76. Found: C, 38.57; H, 3.04; N, 6.28;
S 14.90.
Chromatography on silica gel with hexanes-ethyl acetate
(4:1) yielded an analytical sample of 3b, as a pale yellow
1
7-Ben zyloxy-8-m eth oxy-5-m eth yl-2,3-m eth ylen ed ioxy-
ben zo[c]p h en a n th r id in iu m 2-Nitr oben zen esu lfon a te (8).
A mixture of 6 (230 g, 0.56 mol) and 7 (244 g, 1.12 mol) in
toluene (4.6 L) was refluxed for 35 h to yield a yellow
suspension, which was then cooled, filtered, and washed with
toluene. The crude product was recrystallized from DMF to
semicrystalline solid: mp 49 °C; H NMR (CDCl3) δ 3.91 (s,
3H), 5.12 (s, 2H), 6.98 (d, J ) 8.8 Hz, 1H), 7.33-7.48 (m, 6H),
10.25 (s, 1H); 13C NMR (CDCl3) δ 56.3, 76.5, 112.5, 117.4, 128.5
× 3, 128.7 × 2, 129.1, 129.6, 136.3, 150.4, 152.9, 190.4; FAB-
MS m/z 320 and 322 (M+•), 321 and 323 (M+ + H). Anal.
Calcd for C15H13BrO3: C, 56.10; H, 4.08; Br, 24.88. Found:
C, 55.76; H, 3.77; Br, 24.75.
1
give 8 (280 g, 80%) as a yellow powder: mp 227 °C; H NMR
(DMSO-d6) δ 4.13 (s, 3H), 4.96 (s, 3H), 5.44 (s, 2H), 6.35 (s,
2H), 7.31-7.48 (m, 3H), 7.48-7.60 (m, 3H), 7.63 (dd, J ) 7.8,
1.6 Hz, 2H), 7.72 (s, 1H), 7.80?7.87 (m, 1H), 8.25 (br d, J )
9.2 Hz, 2H), 8.27 (s,1H), 8.75 (d, J ) 9.2 Hz, 1H), 8.77 (d, J )
9.2 Hz, 1H), 9.92 (s, 1H); 13C NMR (DMSO-d6) δ 52.2, 56.9,
75.5, 102.7, 104.1, 105.7, 118.5, 119.5, 119.5, 119.9, 122.2,
125.0, 125.7, 127.8, 128.3 × 2, 128.4, 128.8 × 2, 128.9, 129.8,
130.5, 131.0, 131.4, 132.1, 136.3, 139.3, 143.6, 147.7, 148.6,
148.6, 150.4, 150.7; FAB-MS m/z 424 (M+ + H). Anal. Calcd
for C33H26N2O9S‚DMF: C, 61.79; H, 4.75; N, 6.01; S, 4.58.
Found: C, 61.78; H, 4.56; N, 5.79; S, 4.67.
N-(6-Br om o-2-b en zyloxy-3-m et h oxyb en zyl)-6,7-m et h -
ylen ed ioxy-1-n a p h th yla m in e (5). To the above-mentioned
toluene solution (14 L) of 3b (net 467 g, 1.45 mol) was added
2,3-methylenedioxy-5-naphthylamine 4 (298 g, 1.59 mol), and
the mixture was refluxed for 1.5 h. Removal of toluene gave
a Schiff base as an oil. The volume was brought back to 14 L
with toluene, dimethylamineborane (65 g, 1.10 mol) was added,
and the mixture was cooled to 20 °C. To this mixture was
slowly added acetic acid (940 mL), and the resulting solution
was stirred for 1 h. After reduction was complete, the reaction
mixture was quenched with 1 N HCl (7 L). After being stirred
for 1 h, the mixture was neutralized with 5 N NaOH and the
organic layer was separated. The aqueous layer was extracted
with toluene (6 L). The organic layers were combined, washed
with water, dried over anhydrous MgSO4, and concentrated
in vacuo. The resulting residue was recrystallized with EtOH
to give 5 (664 g, 92%) as a white powder: mp 126 °C; 1H NMR
(CDCl3) δ 3.90 (s, 3H), 4.37 (br s, 1H), 4.48 (s, 2H), 5.04 (s,
2H), 5.99 (s, 2H), 6.74 (dd, J ) 7.4, 1.2 Hz, 1H), 6.82 (d, J )
8.8 Hz, 1H), 7.05 (s, 1H), 7.11 (br d, J ) 8.0 Hz, 1H), 7.33 (d,
J ) 8.8 Hz, 1H), 7.18?7.37 (m, 6H); 13C NMR (CDCl3) δ 56.0,
43.8, 75.9, 97.6, 100.9, 104.7, 105.5, 113.0, 115.5, 117.6, 120.2,
125.1, 128.2 × 2, 128.4 × 2, 131.2, 132.9, 136.9, 143.0, 147.0,
147.2, 147.6, 152.5; FAB-MS m/z 491 and 493 (M+•), 492 and
494 (M+ + H). Anal. Calcd for C26H22BrNO4: C, 63.43; H,
4.50; N, 2.84; Br, 16.23. Found: C, 63.29; H, 4.37; N, 2.67;
Br, 16.30.
7-Ben zyloxy-8-m eth oxy-2,3-m eth ylen ed ioxyben zo[c]-
p h en a n th r id in e (6). A solution of 5 (529 g, 1.07 mol) and
n-Oct3SnH (989 g, 2.15 mol) in toluene (53 L) was heated to
110 °C. A solution of AMBN (310 g, 1.61 mol) in toluene (0.53
L) was then added to the above solution. After 70 min, the
mixture was cooled to room temperature, MnO2 (636 g) was
added, and the resultin solution was stirred well for 1 h. After
oxidation was complete, the reaction mixture was filtered
through Celite, and the filtrate was concentrated in vacuo. The
residue was recrystallized with CHCl3-hexane to give 6 (251
g, 57%) as a white powder: mp 172 °C; 1H NMR (CDCl3) δ
4.07 (s, 3H), 5.32 (s, 2H), 6.13 (s, 2H), 7.26 (s, 1H), 7.34-7.46
(m, 3H), 7.58 (dd, J ) 8.0, 1.6 Hz, 2H), 7.61 (d, J ) 9.0 Hz,
1H), 7.84 (d, J ) 9.0 Hz, 1H), 8.34 (d, J ) 9.0 Hz, 1H), 8.36 (d,
J ) 9.0 Hz, 1H), 8.69 (s, 1H), 9.75 (s, 1H); 13C NMR (CDCl3) δ
50.0, 76.2, 101.6, 102.5, 104.7, 118.6, 118.7, 119.0, 120.3, 122.5,
127.4, 128.4, 128.7, 128.9 × 2, 129.0 × 2, 129.5, 130.1, 137.6,
140.4, 144.4, 147.2, 148.7, 148.9, 149.9; FAB-MS m/z 410 (M+
+ H). Anal. Calcd for C26H19NO4: C, 76.27; H, 4.68; N, 3.42.
Found: C, 76.02; H, 4.41; N, 3.25.
7-Ben zyloxy-6-eth oxy-8-m eth oxy-5-m eth yl-2,3-m eth yl-
en ed ioxy-5,6-d ih yd r oben zo[c]p h en a n th r id in e (9). To a
yellow suspension of 8 (280 g, 0.45 mol) in ethanol (4.2 L) was
added 0.1 N NaOH (4.7 L), and the mixture was stirred for
several hours until its color changed to white. The resulting
white suspension was filtered, washed with ethanol-water (1:
1), and dried in vacuo to give 9 (208 g, 99%) as a white
1
powder: mp 141-149 °C; H NMR (CDCl3) δ 1.05 (t, J ) 7.1
Hz, 3H), 2.61 (s, 3H), 3.60, 3.90 (each dq, J ) 9.6, 7.1 Hz, 1H
× 2), 3.94 (s, 3H), 5.09, 5.21 (AB, J ) 10.9 Hz, 1H × 2), 5.64
(s,1H), 6.04 (s, 2H), 7.06 (d, J ) 8.6 Hz, 1H), 7.11 (s, 1H), 7.33-
7.58 (m, 5H), 7.46 (d, J ) 8.6 Hz, 1H), 7.63 (s, 1H), 7.64 (d, J
) 8.6 Hz, 1H), 7.78 (d, J ) 8.6 Hz, 1H); minor isomer at the
6-position δ 2.52 (s), 3.96 (s), 5.21 (s), 5.90 (AB); 13C NMR
(CDCl3) δ 15.2, 40.6, 56.0, 61.6, 75.7, 84.7, 100.6, 101.0, 104.6,
113.0, 119.1, 120.1, 122.6, 123.3, 125.0, 126.1, 126.7, 128.0,
128.3 × 2, 128.4 × 2, 131.0, 138.0, 138.7, 145.5, 147.3, 147.8,
152.2; FAB-MS m/z 424 (M+ + H - EtOH). Anal. Calcd for
C
29H27NO5: C, 74.18; H, 5.80; N, 2.98. Found: C, 74.36; H,
5.53; N, 2.84.
7-H yd r oxy-8-m et h oxy-5-m et h yl-2,3-m et h ylen ed ioxy-
ben zo[c]p h en a n th r id in iu m Hyd r ogen su lfa te d ih yd r a te:
NK109 (1b). A mixture of 9 (1.05 kg, 2.24 mol) and 6 N H2SO4
(53 L) was heated (100 °C) for 1 h. After debenzylation was
complete, the mixture was cooled to 50 °C and acetone (53 L)
was added. The resulting orange solid was cooled, filtered,
washed with acetone, and dried in vacuo. Hydration at 75%
relative humidity for several days gave NK109 (1b) (1.07 kg,
1
98%) as an orange solid: mp 256 °C (dec); H NMR (DMSO-
d6) δ 4.08 (s, 3H), 4.92 (s, 3H), 6.34 (s, 2H), 7.73 (s, 1H), 8.13
(d, J ) 9.0 Hz, 1H), 8.235 (d, J ) 9.0 Hz, 1H), 8.245 (s, 1H),
8.46 (d, J ) 9.0 Hz, 1H), 8.73 (d, J ) 9.0 Hz, 1H), 10.05 (s,
1H), 10.5-12.3 (br s, 1H); 13C NMR (DMSO-d6) δ 51.7, 56.8,
102.6, 104.0, 105.6, 113.6, 115.1, 118.7, 120.1, 124.3, 125.0,
127.3, 130.6, 131.1, 132.0, 145.8, 145.9, 148.4 × 2, 150.9; FAB-
MS m/z 334 (M+). Anal. Calcd for C20H17NO8S‚2H2O: C,
51.39; H, 4.53; N, 3.00; S, 6.86. Found: C, 51.34; H, 4.13; N,
2.66; S, 6.52.
Meth yl 2-Nitr oben zen esu lfon a te (7). Sodium methoxide
(73 g, 1.42 mol) in methanol (0.36 L) was added in several
portions to a solution of o-nitrobenzenesulfonyl chloride (315
g, 1.42 mol) in methanol (2.55 L) at 10 °C. After 0.5 h, the
mixture was acidified with 0.1 N HCl (0.6 L), and then water
(2.11 L) was added. After an additional 0.5 h of stirring, the
resulting pale yellow crystal was filtered, washed with enough
water to remove sodium chloride, and dried in vacuo to give 7
(285 g, 97%) as a pale yellow crystal: mp 59 °C; 1H NMR
(CDCl3) δ 3.98 (s, 3H), 7.73-7.89 (m, 3H), 8.09-8.17 (m, 1H);
Ack n ow led gm en t. We are very grateful to Prof.
Miyoji Hanaoka (Kanazawa University, J apan) for
providing several benzo[c]phenanthridine alkaloids.
Su p p or t in g In for m a t ion Ava ila b le: 1H and 13C NMR
spectra of compounds 1b, 3b, and 5-9 (14 pages). This
material is contained in libraries on microfiche, immediately
follows this article in the microfilm version of the journal, and
can be ordered from the ACS; see any current masthead page
for ordering information.
(25) As determined by HPLC analysis using the analytical sample
as a standard.
J O9718758