Total syntheses of racemic, natural (-) and unnatural (+) glyceollin I

Article abstract of DOI:10.1021/ol802112r

(Chemical Equation Presented) The first total syntheses of racemic glyceollin I and its enantiomers are described. A Wittig approach was utilized as an entry to the appropriately substituted isoflav-3-ene so that an osmium tetroxide mediated asymmetric dihydroxylation could be deployed for stereospecific Introduction of the 6a-hydroxy group. While using triphenylphosphine hydrobromide, a novel method was found for gently removing MOM from protected phenolic hydroxyl groups present within sensitive systems.

Full text of DOI:10.1021/ol802112r

ORGANIC
LETTERS
2008
Vol. 10, No. 21
5007-5010
Total Syntheses of Racemic, Natural (-)
and Unnatural (+) Glyceollin I
Rahul S. Khupse and Paul W. Erhardt*
Department of Medicinal and Biological Chemistry, Center for Drug Design and
DeVelopment (CD3), The UniVersity of Toledo, Toledo, Ohio 43606-3390
paul.erhardt@utoledo.edu
Received September 9, 2008
ABSTRACT
The first total syntheses of racemic glyceollin I and its enantiomers are described. A Wittig approach was utilized as an entry to the appropriately
substituted isoflav-3-ene so that an osmium tetroxide mediated asymmetric dihydroxylation could be deployed for stereospecific introduction
of the 6a-hydroxy group. While using triphenylphosphine hydrobromide, a novel method was found for gently removing MOM from protected
phenolic hydroxyl groups present within sensitive systems.
Already known for its beneficial cardiovascular effects, there
is growing interest in the common Glycine max as a possible
dietary supplement for the prevention of cancer.1 Of par-
ticular note are soy’s phytoalexins that are produced as a
defense mechanism in response to various insults.2 The
glyceollins (GLYs) are pterocarpan phytoalexins3 that can
be elicited in trace amounts as a mixture having three family
members when soybean cotyledons are stressed by infection
with Aspergillus.4 In contrast to the more prevalent soy
isoflavonoids like genistein and daidzein, the GLYs exhibit
marked antiestrogenic activity in some tissues,5 as well as
demonstrating anticancer properties6 and promise for use as
selective estrogen receptor modulators or SERMs.7
(4) (a) Boue, S. M.; Carter, C. H.; Ehrlich, K. C.; Cleveland, T. E. J.
Agric. Food Chem. 2000, 48, 2167–2172. (b) Cleveland, T. E.; Boue, S. M.;
Burow, M. E.; McLachlan, J. A. U.S. Pat. Appl. Publ. 2006, 1–16.
(5) Burow, M. E.; Boue, S. M.; Collins, B. M.; Melnik, L. I.; Duong,
B. N.; Carter, C. H.; Li, S.; Wise, T. E.; Cleveland, T. E.; McLachlan,
J. A. J. Clin. Endocriol. Metab. 2001, 86, 1750–1758.
(1) (a) Persky, V.; VanHorn, L. J. Nutr. 1995, 125, 709S–712S. (b)
Herman, C.; Adlercreutz, T.; Goldin, B. R.; Gorbach, S. L.; Hockerstedt,
K. A. V.; Watanabe, S.; Hamalainen, E. K.; Markkanen, M. H.; Makela,
T. H.; Wahala, K. T.; Hase, T. A.; Fotsis, T. J. Nutr. 1995, 125, 757S–
770S.
(6) (a) Erhardt, P. W.; Khupse, R. S.; Sarver, J. G.; Cleaveland, T. E.;
Boue, S. M.; Wiese, T. E.; Burow, M. E.; McLachlan, J. A. U.S. Pat. Appl.
Submitted 2008, 1-65. (b) Salvo, V. A.; Boue, S. M.; Fonesca, J. P.; Elliott,
S.; Corbitt, C.; Collins-Burrow, B. M.; Curiel, T. J.; Srivastav, S. K.; Shih,
B. Y.; Carter-Wientjes, C.; Wood, C. E.; Erhardt, P. W.; Beckman, B. S.;
McLachlan, J. A.; Cleveland, T. E.; Burow, M. E Clin. Cancer Res. 2006,
12, 7159–7164.
(2) (a) Hammerschmidt, R. Annu. ReV. Phytopathol. 1999, 37, 285–
306. (b) Faghihi, J.; Jiang, X.; Vierling, R.; Goldman, S.; Sharfstein, S.;
Sarver, J. G.; Erhardt, P. W. J. Chromatogr. A 2001, 915, 61–74.
(3) Al-Hazimi, H. M. G.; Alkhathlan, H. Z. J. King Saud UniV. 2000,
12, 93–122.
10.1021/ol802112r CCC: $40.75
Published on Web 09/26/2008
2008 American Chemical Society

Efforts in our laboratory have been directed toward
synthesizing the individual GLYs.8 To date, there have been
four GLYs isolated from natural sources but none have
undergone total synthesis. GLY I is the most prevalent family
member and its synthesis is described herein.
The isoflav-3-ene is a key intermediate in our overall
synthesis, and devising a robust method for establishment
of its double bond was important. We imagined that if a
Wittig olefination reaction was set up to occur in an
intramolecular fashion, the desired isoflav-3-ene could be
obtained as a consequence of the ring-closure.9 Our synthetic
route to this key intermediate is depicted in Scheme 1.
nation with 0.5 equiv of iodine and 0.6 equiv of Selectfluor
gave the desired halide 5 in 70% yield.11 Unlike 1, initial
attempts to regioselectively protect 3 resulted in significant
amounts of dibenzylated material (30%). The latter can be
circumvented by specifically deploying sodium bicarbonate
in acetonitrile, whereupon we observed less than 5% diben-
zylation.12 Reduction to the salicylalcohol derivative 6 by
using sodium borohydride in MeOH followed by conven-
tional workups also proved to be problematic, perhaps owing
to ready formation of the quinone-methide under either basic
or even acidic conditions. Review of the literature indicates
that alternate methods are typically used for making salicyl-
alcohol from salicylaldehyde.13 In the end, we were able to
conduct a standard sodium borohydride reduction in MeOH
by adopting a workup where after evaporation of solvent,
0.1 N sulfuric acid was carefully added so as to achieve an
acidic pH of not less than 6.0. Addition of water then
conveniently precipitated high-purity product in nearly 80%
yield. Coupling of 5 and 6 was accomplished by using
potassium carbonate in acetone14 to obtain product 7 in high
yield (72%).
Scheme 1
When formation of phosphonium salt 8 was attempted by
treatment of 7 with triphenylphosphine hydrobromide
(TPP·HBr) in refluxing acetonitrile, a mixture of products
was produced wherein inadvertent loss of the MOM group
was also observed. Alternatively, when this reaction was
done in freshly distilled acetonitrile at room temperature, a
quantitative yield of 8 was obtained.15 Intramolecular
condensation of 8 was then accomplished by using sodium
tert-butoxide as base in refluxing methanol, after which
product precipitated from the reaction medium.16 Crystal-
lization of this material from methanol afforded pure 9 in
78% yield for the two-step process. To avoid the possibility
for acid-catalyzed 1,2-elimination of the 6a-hydroxy group
during removal of MOM as the last step of the overall
synthesis, at this point we replaced MOM with a TBDMS
group. Since we repeatedly found 9 to be sensitive to the
various acidic conditions typically deployed to remove
MOM, we exploited our earlier observation wherein the
MOM group was inadvertently lost during treatment of 7
with TPP·HBr. Thus, refluxing 9 in acetonitrile:water (20:1)
in the presence of TPP·HBr produced the sensitive phenolic
intermediate, which after a quick column chromatographic
purification was used immediately in the next step. Treatment
with TBDMS-chloride and triethylamine in DCM gave the
stable TBDMS-protected isoflav-3-ene 10 as a white solid
that was crystallized from DCM and methanol to afford a
Because the o-hydroxy group in 1 forms a tight hydrogen
bond with the adjacent carbonyl, regioselective protection
of the p-hydroxy group as the MOM ether occurs cleanly.
This is followed by benzyl-protection of the o-hydroxy group,
using more rigorous conditions. R-Bromination of the
orthogonally protected ketone by using conventional methods
resulted in a complex mixture of products, namely due to
loss of MOM and ring bromination.10 Alternatively, iodi-
(7) (a) Collins-Burow, B. M.; Burow, M. E.; Duong, B. N.; McLachlan,
J. A. Nutr. Cancer 2000, 38, 229–244. (b) LaVecchia, C.; Gallus, S.;
Fernandez, E. J. Br. Menopause Soc. 2005, 11, 166–172. (c) Newcomb,
P. A.; Zheng, Y.; Chia, V. M.; Morimoto, L. M.; Doria-Rose, V. P.;
Templeton, A.; Thibodeau, S. N.; Potter, J. D. Cancer Res. 2007, 67, 7534–
7539.
(8) (a) Khupse, R. S.; Erhardt, P. W. 30th Natl. Med. Chem. Symp. Seattle,
WA 2006, Abstract 87. (b) Khupse, R. S.; Erhardt P. W. Abstracts of Papers,
234th National Meeting of the American Chemical Society, Boston, MA,
2007; American Chemical Society: Washington, DC, 2007; Abstract MEDI-
182.
(11) Jereb, M.; Stavber, S.; Zupan, M. Synthesis 2003, 853–858.
(12) (a) Dewick, P. M. Synth. Commun. 1981, 11, 853–857. (b) Kevin;
T. Mendelson, W. W.; Zhang, W. Y.; Lantos, I.; Matsuoka, R. T.; LynnM.;
Liu, L.; Procter, G. World patent WO9619438 1996, 40 pp.
(13) (a) Nagata, W.; Okada, K.; Aoki, T. Synthesis 1979, 365–368. (b)
Belyanin, M. L.; Filimonov, V. D.; Krasnov, E. A. Russ. J. Appl. Chem.
2001, 74, 103–105. (c) Brink, M. Acta UniV. Lund., Sect. 2 1965, 16, 1–
11, CAN64:51716.
(9) (a) Conti, C.; Desideri, N.; Orsi, N.; Sestili, I.; Stein, M. L. Eur.
J. Med. Chem. 1990, 25, 725–730. (b) McAllister, G. D.; Hartley, R. C.;
Dawson, M. J.; Knaggs, A. R. J. Chem. Soc., Perkin Trans. 1 1998, 20,
3453–3458. (c) Begasse, B.; Corre, M. Tetrahedron 1980, 36, 3409–3412.
(10) (a) Fougerousse, A.; Gonzalez, E.; Brouillard, R. J. Org. Chem.
2000, 65, 583–586. (b) King, L. C.; Ostrum, G. K. J. Org. Chem. 1964,
29, 3459–3461. (c) Cacchi, S.; Caglioti, L.; Cernia, E. Synthesis 1979, 64–
66. (d) Horiuchi, C. A.; Kiji, S. Bull. Chem. Soc. Jpn. 1997, 70, 421–426.
(14) Kraus, G. A.; Kim, I. Org. Lett. 2003, 5, 1191–1192.
(15) Yuan, Y.; Men, H.; Lee, C. J. Am. Chem. Soc. 2004, 126, 14720–
14721.
(16) Burali, C.; Desideri, N.; Stein, M. L.; Conti, C.; Orsi, N. Eur.
J. Med. Chem. 1987, 22, 119–123.
5008
Org. Lett., Vol. 10, No. 21, 2008

69% yield for the two-step process. The overall yield of 10
from 1 by this route consistently occurs in ca. 20%.
The remaining portions of the syntheses followed the path
shown in Scheme 2. Distinct stereochemistry was established
By using the Sharpless model, the cinchona alkaloids can
be deployed as chiral ligands to dictate the stereospecificity
of this step.19 Our analysis indicated that use of (DHQD)2-
PHAL as a chiral catalyst should cause the two oxygen atoms
to be delivered from the ꢀ-face so as to ultimately provide
the (6aS,11aS) stereochemistry desired for the natural (-)-
GLY I. Alternatively, use of (DHQ)2PHAL should provide
the (6aR,11aR) stereochemistry desired for the unnatural (+)-
enantiomer.
Scheme 2
However, when we now added the DHQD reagent to our
procedure, the anticipated enantioselectivity was not ob-
served. We speculate that the chiral ligand needs to become
tightly bound with the osmate ester complex to display its
enantioselectivity, such that disruption of this arrangement
under the conditions of our modified OT reaction also leads
to a loss of asymmetric induction. Thus, while our method
having reduced OT levels remains very useful for the
production of large quantities of racemic material, to pursue
the pure enantiomers of GLY I we returned to the more
traditional method of using stoichiometric amounts of OT
and each chiral ligand. These higher loading conditions gave
each enantiomer in greater than 95% ee. Enantiomeric purity
and absolute configuration were assessed by NMR chiral shift
reagent studies and by CD studies.
The rest of the syntheses (Scheme 2) were conducted in
parallel for each enatiomer of 11 after refining the chemistry
for the individual steps by deploying the racemic mixture
as a model system. Absolute stereochemistry and ee were
assesed at each step by using CD and NMR studies. The
next step involved debenzylation of diol 11 to the tetrol 12
with the likelihood for spontaneous cyclization to the
protected glycinol 13 via the quinone-methide depicted in
Scheme 2. A “one-pot” method for reductive cyclization has
been described20 and used successfully during a synthesis
of pisatin.18a Using this procedure, we observed only small
amounts of cyclized product. Alternatively, we found tetrol
12 to be stable and conveniently isolable, and therefore
performed the cyclization as a separate step. To avoid
polymerization of the reactive quinone-methide and thus
promote intramolecular ring closure, we used a dilute
solution. In addition, use of molecular sieves to remove water
formed as byproduct helped to drive this cyclization forward.
In the end, polymer-bound 1,3,4,6,7,8-hexahydro-2H-py-
rimido(1,2-a)pyrimidine21 was deployed as a base in anhy-
drous ethanol over molecular sieves, and a 64% overall yield
across these two steps was achieved. The NMR shift of the
C-11a proton is particularly diagnostic for defining a cis
versus a trans ring closure within these types of systems.
Our C-11a signal at 5.2 ppm was in accord with those of
the cis isomers for variabilin and for pisatin as reported by
others.18
during the osmium tetroxide (OT) dihydroxylation step. This
approach has been utilized for other 6a-hydroxypterocarpans
wherein the dihydrobenzopyran system then dictates that
subsequent formation of the dihydrobenzofuran ring results
in the desired 6a,11a-cis-arrangement.17,18
Stoichiometric amounts of OT are typically required
because the OT generally forms a stable osmate ester during
the reaction that precludes recycling of osmium. Exploring
the possibility for reducing the OT load by adding a
secondary oxidant, as well as by surveying different condi-
tions, we observed that when 1 equiv of methanesulfonamide
(MSA) is added, breaking of the osmate ester appears to be
facilitated such that this can allow for recycling of
osmium.19a From our survey, N-methylmorpholine oxide
(NMO) proved to be the most efficient secondary oxidant,
while acetone:water proved to be the best solvent media.
Remarkably, this combination of MSA and NMO enables
the desired transformation to take place in high yield (90%)
for racemic 11 while using only 10 mg of OT for 10 mmol
of reactant compared to the typical stoichiometric amount
that would have required 2.5 g of OT.
Two approaches were contemplated for assembly of the
final isoprenyl-ring systems, namely a modified aldol con-
(17) (a) Schoning, A.; Freidrichsen, W. Z. Naturforsch. B: Chem. Sci.
1989, 44, 975–980. (b) Riedrichsen, W.; Schoening, A. Z. Naturforsch. B:
(19) (a) Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, B. K. Chem.
ReV. 1994, 94, 2483–2547. (b) Norrby, P.; Kolb, H. C.; Sharpless, K. B.
J. Am. Chem. Soc. 1994, 116, 8470–8478.
Chem. Sci. 1990, 45, 1049–1054
.
(18) (a) Pinard, E.; Gaudry, M.; Henot, F.; Thellend, A. Tetrahedron
Lett. 1998, 39, 2739–2742. (b) van Aardt, T. G.; van Rensburg, H.; Ferreira,
D. Tetrahedron 2001, 57, 7113–7126. (c) Mori, K.; Kisida, H. Liebigs Ann.
(20) Prasad, A. V. K.; Kapil, R. S.; Popli, S. P. J. Chem. Soc., Perkin
Trans. 1 1986, 1561–1563.
(21) Xu, W.; Mohan, R.; Morrissey, M. M. Tetrahedron Lett. 1997, 38,
7337–7340.
Chem. 1989, 1, 35–39
.
Org. Lett., Vol. 10, No. 21, 2008
5009

densation22 and a Harfenist-Thom rearrangement of the
propargyl ether.23 In either strategy, ring closure to position
4 of 13 was expected to give the desired 14a as the major
product, while closure to the alternative position 2 would
provide regioisomer 14b that can be additionally elaborated
to produce glyceollin II. Introduction of a dimethylpropargyl
group proved to be tedious and appeared to cause collapse
of the GLY skeleton. Alternatively, the aldol approach
proved to be quite useful upon condensing the tautomeric
keto-form of the phenol with the unsaturated aldehyde
masked as its acetal. This was done in refluxing p-xylene
by using 0.25 equiv of picoline as base with continuous
removal of the liberated alcohol. As anticipated, the TBDMS-
protected forms of both GLY I and GLY II were obtained,
with the significantly more abundant being the GLY I
regioisomer when assessed at the crude product stage by
NMR. Separation of these regioisomers was accomplished
by using silica column chromatography to afford ca. 50%
yields of the racemic and enantiomeric forms of TBDMS-
GLY I (14a). A 10% yield of the (() form of TBDMS-
GLY II (14b) was also obtained during the model chemistry
when larger quantities of the racemic mixture were deployed.
In the final step, attempted deprotection with basic reagents
like silica-supported TBAF or TAS-F24 largely resulted in
degradation of the formed products. After isolating small
quantities of racemic GLY I and conducting a stability study
by using HPLC, we deduced that this ring system is not
stable in basic pH above 8, or in acidic pH below 4. Thus,
we explored mildly acidic reagents that might be used for
deprotection. In the end, NEt3·3HF,25 which provides a pH
in the range of 4-5 when run on the 0.1 mmol scale in
anhydrous acetonitrile, was utilized for this transformation
and allowed for 70% to 80% yields. Starting from building
block 1, the overall yield for the total stereoselective synthesis
of natural (-)-GLY I (15) was 3.4%.
nature, all of which bear cis (6aS,11aS) stereochemistry.26
Most conspicuous are the protons at C-6 which appear as
two separate doublets with the C-6 equatorial proton ap-
pearing downfield compared to the C-6 axial proton, and
wherein W coupling occurs with the C-11a and C-6 equato-
rial proton, an event not possible for the trans system. Both
COSY and NOSEY spectra also confirm this relationship
between the C-6 equatorial and C-11a protons. Absolute
stereochemistry was confirmed by CD and chiral HPLC
studies which deployed an authetic sample of (-)-GLY I
from nature as a standard. High-resolution mass spectral data
were in agreement with theoretical values for each enanti-
omer, as well as for racemic GLY I, and correctly support a
common molecular formula of C20H18O5.
In addition to conveying the first total syntheses of racemic
GLY I and its enantiomers, this report describes a useful
entry to substituted isoflav-3-ene systems via a Wittig
approach, as well as the use of sulfonamides for reducing
OT load during dihydroxylation reactions when control of
stereochemistry is not essential. Of particular note is a novel
method for gently removing MOM from protected phenolic
hydroxyl groups present within sensitive structures by using
triphenylphosphine hydrobromide. The scope and selectivity
for this type of deprotection procedure is presently being
surveyed within the CD3 laboratories.
Acknowledgment. This work was supported by a grant
from the United Sates Department of Agriculture (USDA)
Southern Regional Research Center (SRRC) in New Orleans.
We thank Dr. Steve Boue of the SRRC for a natural GLY I
sample isolated from elicited soy, Dr. Michael Ogawa and
Mr. Fei Xie of the Bowling Green State University Depart-
ment of Chemistry for assistance during the CD studies, and
Dr. Jared Anderson of The Univeristy of Toledo Department
of Chemistry for discussions during the chiral HPLC studies.
The proton NMR spectra of our synthesized GLYs
resemble the data reported for the materials obtained from
Supporting Information Available: Experimental pro-
cedures, NMR and CD studies including selected spectra,
chiral HPLC, and NMR spectra for new intermediates and
final compounds. This material is available free of charge
via the Internet at http://pubs.acs.org.
(22) North, J. T.; Kronenthal, D. R.; Pullockaran, A. J.; Real, S. D.;
Chen, H. Y. J. Org. Chem. 1995, 60, 3397–3400.
(23) Harfenist, M.; Thom, E. J. Org. Chem. 1972, 37, 841–848.
(24) (a) Gambacorta, A.; Turchetta, S.; Botta, M. Synth. Commun. 1989,
19, 2441–2448. (b) Barrett, A. G. M.; Pena, M.; Willardsen, J. A. J. Org.
Chem. 1996, 61, 1082–1100.
OL802112R
(25) Julian, L. D.; Newcom, J. S.; Roush, W. R. J. Am. Chem. Soc.
2005, 127, 6186–6187.
(26) Lyne, R. L.; Mulheirn, L. J.; Leworthy, D. P. J. Chem. Soc., Chem.
Commun. 1976, 13, 497–498.
5010
Org. Lett., Vol. 10, No. 21, 2008