Convenient synthesis of 5,6,11,12,17,18-hexahydrocyclononal[1,2-b:4,5-b′:7,8-b″] triindole, a Novel phytoestrogen

Article abstract of DOI:10.1021/jo020415y

An efficient one-pot synthesis is described of 5,6,11,12,17,18-hexahydrocyclononal[1,2-b:4,5-b':7,8-b ]triindole (CTr), a potent estrogen agonist from food plants. For the procedure, gramine is treated with dimethyl sulfate and sodium in ethanol at room temperature. Quenching of the reaction with water and workup of the product provides CTr in approximately 75% yield.

Full text of DOI:10.1021/jo020415y

SCHEME 1
Con ven ien t Syn th esis of 5,6,11,12,17,18-
Hexa h yd r ocyclon on a l[1,2-b:4,5-b′:7,8-b′′]tr i-
in d ole, a Novel P h ytoestr ogen
Richard E. Staub and Leonard F. Bjeldanes*
Department of Nutritional Sciences and Toxicology,
University of California, Berkeley, California 94720-3104
lfb@nature.berkeley.edu
Received J une 18, 2002
procedure for CTr is a low-yielding process that requires
HPLC purification of product from the complex acid
reaction mixture of I3C.7 We report a convenient, high-
yield, one-pot procedure for the synthesis of CTr in
quantities required for metabolic and structure-activity
studies.
Abstr a ct: An efficient one-pot synthesis is described of
5,6,11,12,17,18-hexahydrocyclononal[1,2-b:4,5-b′:7,8-b′′]tri-
indole (CTr), a potent estrogen agonist from food plants. For
the procedure, gramine is treated with dimethyl sulfate and
sodium in ethanol at room temperature. Quenching of the
reaction with water and workup of the product provides CTr
in approximately 75% yield.
Treatment of I3C under a variety of conditions pro-
vided either a complex mixture of oligomeric products
(acidic conditions) or a reasonably good yield of the
dimeric product 3,3′-diindolylmethane (basic conditions).
We then turned our attention to gramine as a source of
the presumed 3-methylene indolenine oligomerization
intermediate. Although the literature is extensive on the
use of gramine quaternary ammonium salt in the syn-
thesis of 3-methylindole analogues,10 we are aware of no
reports of the preparation of self-condensation products
from this starting material. We used dimethyl sulfate to
generate the quaternary salt of gramine in the presence
of sodium ethoxide (Scheme 1), conditions that presum-
ably produced the reactive indolenine intermediate. The
reaction was surprisingly facile and yielded primarily
CTr (∼75%) with only one significant byproduct, the
cyclic tetramer (CTet, ∼9%) (Figure 1).
5,6,11,12,17,18-Hexahydrocyclononal[1,2-b:4,5-b′:7,8-
b′′]triindole (CTr) is a novel estrogen agonist produced
under acidic conditions from the putative cancer protec-
tive agent indole-3-carbinol (I3C). I3C is an enzymatic
hydrolysis product of the indolylmethyl glucosinolate
glucobrassicin that occurs in common food plants of the
genus Brassica, including cabbage, kale, cauliflower,
Brussels sprouts, and broccoli. Under acidic conditions
in vitro, and following ingestion, I3C readily self-
condenses to form a mixture of oligomeric products
including CTr.1-3 I3C and its oligomeric products are
under study as cytostatic4-7 and tumor-suppressive
agents.8 In addition, I3C exhibits tumor-promoting activ-
ity in some assays,9 an effect that may arise in part from
estrogenic products of I3C formed in vivo. We have shown
that CTr is a strong agonist of estrogen receptor function.
Computational modeling studies indicated an excellent
fit of CTr into the ligand-binding site of the estrogen
receptor.7
Spectral analysis provided full support for the proposed
structure of CTr and its conformational preferences. The
1H NMR spectrum for CTr (Figure 2) was similar to that
previously reported2 and demonstrates that this com-
pound is composed of almost equal proportions of crown
and saddle conformations, on the basis of the resonances
of the saturated cyclononane methylene protons.11,12 This
spectral information, which is corroborated with Dreiding
models, indicates that the geminal methylene protons are
fixed in position in the crown conformation, resulting in
For our continued studies of the biological effects of
CTr, we required a high-yielding synthetic procedure for
CTr and possible analogues. Currently the synthetic
1
two doublets in the H NMR spectrum.11 In contrast, the
(1) De Kruif, C. A.; Marsman, J . W.; Venekamp, J . C.; Falke, H. E.;
Norrdhoek, J .; Blaauboer, B. J .; Wortelboer, H. M. Chem.-Biol. Interact.
1991, 80, 303-315.
cyclononane ring in the saddle conformation has consid-
erable flexibility, which collapses the 1H NMR signals
for the six methylene protons to a singlet.2,12 13C NMR
analysis of CTr prepared with 13C-2-labeled precursor
also gave two enriched signals, presumably for the crown
and saddle conformations. Further, the natural-abun-
dance 13C NMR spectrum for CTr shows splitting for the
resonances of C-2, C-3, C-9, and the methylene carbon
(2) Grose, K. R.; Bjeldanes, L. F. Chem. Res. Toxicol. 1992, 5, 188-
193.
(3) Stresser, D. M.; Williams, D. E.; Griffin, D. A.; Bailey, G. S. Drug
Metab. Dispos. 1995, 23, 965-975.
(4) Ge, X.; Yannai, S.; Rennert, G.; Gruener, N.; Fares, F. A.
Biochem. Biophys. Res. Commun. 1996, 228, 153-158.
(5) Chang, Y.-C.; Riby, J .; Chang, G. H-F.; Peng, B.-C.; Firestone,
G.; Bjeldanes, L. F. Biochem. Pharm. 1999, 58, 825-834.
(6) Riby, J . E.; Chang, G. H. F.; Firestone, G. L.; Bjeldanes, L. B.
Biochem. Pharm. 2000, 60, 167-177.
(7) Riby, J . E.; Feng, C.; Chang, Y.-C.; Schaldach, C. M.; Firestone,
G. L.; Bjeldanes, L. F. Biochemistry 2000, 39, 910-918.
(8) Chen, I.; McDougal, A.; Wang, F.; Safe, S. Carcinogenesis 1998,
19, 1631-1639.
(9) Stoner, G.; Casto, B.; Ralston, S.; Roebuck, B.; Pereira, C.; Bailey,
G. Carcinogenesis 2002, 23, 265-272.
(10) Sundberg, R. J . Indoles; Academic Press Inc.: San Diego, CA,
1996.
(11) Erdtman, H.; Haglid, F.; Ryhage, R. Acta Chem. Scand. 1964,
18, 1249-1254
(12) Bergman, J .; Ho¨gberg, S.; Lindstro¨m, J .-O. Tetrahedron 1970,
26, 3347-3352.
10.1021/jo020415y CCC: $25.00 © 2003 American Chemical Society
Published on Web 12/06/2002
J . Org. Chem. 2003, 68, 167-169
167

The 1H NMR spectrum for the CTet (Figure 2) was
similar to that previously reported12 with all eight
methylene protons represented as a singlet at 3.85 ppm
(acetone-d6) consistent with its flexible structure. 13C
NMR analysis of CTet prepared from 13C-2-labeled
precursor gave a single enriched signal, also consistent
1
with its flexible structure. Further, H NMR analysis of
the 13C-labeled CTet resulted in a broad multiplet at 3.85
ppm, thereby verifying the C-2,C-3′-methylene bridges
between the indoles in the ring. Mass spectral analysis
of the CTet product provided the expected molecular
mass of 516 (observed as 517 for the M + 1 ion), which
increased by 4 mass units when 13C-labeled precursor
was used, thereby indicating the presence of four indole
rings in the product.
F IGURE 1. Reversed-phase HPLC chromatogram (280 nm)
displaying the distribution of products from the synthesis of
cyclic trimer from gramine.
The same synthetic procedure could be applied utiliz-
ing other alcoholic solvents (methanol and 2-propanol),
but the reaction was not as efficient for the CTr product
(50-52%). Likewise, the synthesis of cyclic trimer from
indole-3-carbinol (with dimethyl sulfate and sodium) was
less efficient (44.7 ( 2.8%, n ) 3) and resulted in a
greater amount of the CTet (11.1 ( 1.7%, n ) 3), linear
trimer (7.2 ( 1.2%, n ) 3), 3,3′-diindolylmethane (3.5 (
0.5%, n ) 3), and other byproducts as determined by
HPLC analysis. The CTr product was similarly produced
when sodium ethoxide was first prepared in ethanol and
then added to the quarternary ammonium intermediate.
However, this reaction was again not as efficient as when
the sodium ethoxide was generated in the presence of
the gramine intermediate. No 3-ethoxymethylindole was
produced under our standard reaction conditions involv-
ing the formation of the quarternary ammonium inter-
mediate, even though this was reported to be the primary
product when gramine was reacted with an alkylating
agent (methyl or ethyl iodide) and sodium ethoxide in
ethanol.14
Facile cyclononane ring formation from gramine ap-
parently has some precedent in the preparation of
cyclotriveratrylene from veratrole or veratrylamine N-
tosylates.11,15,16 Further, a minor byproduct of the cyclo-
triveratrylene condensation is the corresponding cyclic
tetramer. In contrast to the cyclotriveratrylene condensa-
tion that is formed under highly acidic conditions, our
gramine-based procedure is catalyzed by strong base.
Further studies of this novel reaction should clarify its
unexpected product specificity and may prove useful in
the preparation of a unique group of estrogen receptor
agonists and antagonists of potential therapeutic useful-
ness.
F IGURE 2. 1H NMR spectra in deuterated acetone for CTr
(top) and CTet (bottom).
as would be expected for the saddle and crown conforma-
tions. The resonance for C-8 (based on the 13C NMR
spectrum for indole13) was not detected and is probably
under the C-2 resonance since the same resonances were
separated by less than 0.5 ppm in the CTet spectra. Only
single resonances were observed for the phenyl carbons,
whereas two signals were recorded for the other carbons
representing the crown and saddle conformations.
1H NMR analysis of the [13C]CTr analogue resulted
only in a broadening of the methylene proton signals,
consistent with the C-2,C-3′-methylene bridge and the
all-anti-indole ring structural assignments. Mass spectral
analysis provided the expected molecular mass of 387
(observed as 388 for the M + 1 ion), which increased by
3 mass units when 13C-labeled precursor was included
in the synthesis reaction, thereby verifying the presence
of three indole rings in the product.
Exp er im en ta l Section
Gen er a l Meth od s. 1H NMR and 13C NMR spectra were
recorded at 300 MHz, and chemical shifts are reported in parts
per million from that of the deuterated solvent. Commercial
gramine and I3C were recrystallized from toluene prior to use
in the synthesis procedures.
Preparative reversed-phase chromatography was performed
using reversed-phase HPLC (C-18, 10 × 250 mm, 5 µm). A
gradient elution was used: initially at 60% acetonitrile followed
(14) Geissman, T. A.; Armen, A. J . Am. Chem. Soc. 1952, 74, 3916-
3919.
(15) Lindsey, A. S. J . Chem. Soc. 1965, 1685-1692.
(16) Umezawa, B.; Hoshino, O.; Hara, H.; Ohyama, K.; Mitsuba-
yashi, S.; Sakakibara, J . Chem. Pharm. Bull. 1969, 17, 2240-2244.
(13) Breitmaier, E.; Voelter, W. Carbon-13 NMR Spectroscopy; VCH
Publishers: New York, 1987.
168 J . Org. Chem., Vol. 68, No. 1, 2003

by linear gradients of 60-95% acetonitrile over 40 min, 95-
100% over 5 min, and then constant 100% for 5 min before
returning to initial conditions. The eluent was monitored at 280
nm for collection of appropriate products. Retention for products
was as follows: CTr, 22.9 min; CTet, 31.3 min; 3,3′-diindolyl-
methane, 19.8 min; linear trimer, 25.9 min.
For LC/MS analysis, isocratic acetonitrile (75%) and aqueous
0.1% formic acid were pumped at a 0.2 mL/min flow rate (4.6 ×
250 mm, 5 µm), and the eluent was monitored at 280 nm. The
electrospray source was operated at 5.5 kV potential and a
heated capillary temperature of 250 °C. Nitrogen gas was used
at a sheath gas pressure of 40 psi. Auxiliary gas flow was
adjusted to achieve a total flow of 0.5 L/min of nitrogen. For
MS/MS experiments, argon (1 Torr) was used as the collision
gas with a collision energy of -25 V.
7.52 (br m, ArH), 7.87 (br m, ArH), 10.04 (br, NH from crown
and saddle conformers). 13C NMR ((CD3)2CO): δ 136.8 and 136.1
(br, C-2, crown and saddle), 129.8 and 129.5 (br, C-9, crown and
saddle), 121.3 (C-4), 119.5 (C-5), 118.5 (C-6), 111.4 (C-7), 109.2
and 108.6 (br, C-3, crown and saddle), 22.65 and 21.72 (br, CH2,
crown and saddle). LC/MS (m/z): 388 (M + 1, 100). LC/MS/MS
(m/z): products of 388; 388 (M + 1, 51), 271 (100), 257 (32), 245
(4), 130 (C8H6NdCH2+, 37). Mp: chars at 290 °C.
Da ta for CTet. 1H NMR ((CD3)2CO): δ 3.85 (s, CH2), 6.83 (t,
J ) 7.0 Hz, ArH), 6.96 (t, J ) 7.0 Hz, ArH), 7.22 (d, J ) 7.7 Hz,
ArH), 7.30 (d, J ) 8.0 Hz, ArH), 9.91 (s, NH). 13C NMR (pyridine-
d5): δ 137.3 (C-8), 136.9 (C-2), 130.6 (C-9), 121.3 (C-4), 119.6
(C-5), 118.7 (C-6), 111.9 (C-7), 108.3 (C-3), 22.81 (CH2). LC/MS
(m/z): 517 (M + 1, 100). Mp: chars at over 300 °C.
[
13C-2]Indole was used to synthesize [13C-2]I3C as previously
Syn th esis of CTr . Dimethyl sulfate (1.2 mL, 12.6 mmol) was
added to gramine (0.5 g, 2.88 mmol) dissolved in 30 mL of
absolute ethanol. Solid sodium (100 mg, 4.3 mmol) was added,
and the solution was constantly stirred at room temperature for
4 h. The reaction mixture was poured into 100 mL of water and
partitioned with dichloromethane (50 mL, 2×). The combined
organic layers were back-extracted with 5% sodium carbonate
(75 mL) and water (100 mL), dried over Na2SO4, filtered, and
evaporated in vacuo. Before dryness was reached, 25 mL of
hexane was added to precipitate the product and produce a
yellow-white final powder when dry. The mass yield was nearly
quantitative (97.1 ( 2.7%, n ) 3) for a crude product that
contained ∼75% CTr and ∼9% CTet as determined by HPLC
analysis. Trace amounts of linear trimer and 3,3′-diindolyl-
methane were also detected and verified by LC/MS as previously
described.2 Pure CTr for spectral analysis was only obtained by
preparative reversed-phase HPLC. The CTr and CTet could not
be separated using normal-phase chromatography (HPLC and
TLC). Attempts to crystalize CTr were unsuccessful and were
likely inhibited by the mixture of the saddle and crown confor-
mations. CTet was crystallized from dimethyl sulfoxide as fine
white crystals.
described.17 13C-2]I3C was then reacted under the conditions
[
described above to produce 13C-labeled cyclic products for
spectral analysis.
Da ta for [13C-2]CTr . 13C NMR ((CD3)2CO): δ 136.1 (s), 135.3
(s). LC/MS (m/z): 391 (M + 1, 100). LC/MS/MS (m/z): products
of 391; 391 (M + 1, 9), 389 (12), 273 (100), 258 (36), 247 (17),
131 (C8H6NdCH2+, 70).
Da ta for [13C-2]CTet. 13C NMR ((CD3)2CO): δ 136.0 (s). LC/
MS (m/z): 521 (M + 1, 100).
Ack n ow led gm en t. We thank Nanjing Zhang and
Professor Casida at the University of California at
Berkeley for their assistance in obtaining the 13C NMR
spectra. This research was supported in part by the
Department of Defense, Army Breast Cancer Research
Program Grant DAMD17-96-1-6149, National Institutes
of Health Grant CA69056, and a National Institutes of
Health, NRSA Trainee appointment on Grant 5-T32
ES07075.
J O020415Y
Da ta for CTr . 1H NMR ((CD3)2CO): δ 3.99 (br s, CH2 saddle),
4.12 (br d, J ) 15.5 Hz, CH2 crown), 4.83 (br d, J ) 15.0 Hz,
CH2 crown), 6.91 (br m, ArH), 7.12 (br m, ArH), 7.24 (br m, ArH),
(17) Staub, R. E.; Feng, C.; Onisko, B.; Bailey, G.; Firestone, G. L.;
Bjeldanes, L. F. Chem. Res. Toxicol. 2002, 15, 101-10.
J . Org. Chem, Vol. 68, No. 1, 2003 169