Dehydroxylation of stilbenoid oligomers: Absolute configuration determination via comparison of experimental and theoretical electronic circular dichroic spectra

Article abstract of DOI:10.1016/j.tetlet.2013.10.106

Dehydroxylation of naturally occurring oligomeric resveratrol derivatives resulted in the formation of compounds with dramatically reduced principal stable conformers. In addition, dehydroxylation allowed the determination of the absolute configurations o

Full text of DOI:10.1016/j.tetlet.2013.10.106

Tetrahedron Letters 55 (2014) 314–318
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Dehydroxylation of stilbenoid oligomers: absolute configuration
determination via comparison of experimental and theoretical
electronic circular dichroic spectra
b
Tetsuro Ito a, , Tatsuo Nehira
⇑
a Laboratory of Pharmacognosy, Gifu Pharmaceutical University, 1-25-4 Daigaku-nishi, Gifu 501-1196, Japan
b Graduate School of Integrated Arts and Sciences, Hiroshima University, 1-7-1 Kagamiyama, Higashi-hiroshima 739-8521, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Dehydroxylation of naturally occurring oligomeric resveratrol derivatives resulted in the formation of
compounds with dramatically reduced principal stable conformers. In addition, dehydroxylation allowed
the determination of the absolute configurations of the original resveratrols via comparison of experi-
mental and theoretical electronic circular dichroic spectra. Notably, the absolute configuration of paucifl-
orol B was identified using this novel procedure, which is the first application of dehydroxylated
derivatives for the determination of the absolute configuration of naturally occurring polyphenols.
Ó 2013 Elsevier Ltd. All rights reserved.
Received 5 September 2013
Revised 19 October 2013
Accepted 23 October 2013
Available online 6 November 2013
Keywords:
Absolute configuration
Theoretical ECD spectrum
Dehydroxylation
Hydroxy groups
Resveratrol oligomers
Introduction
database useful for both elucidating their structures and identify-
ing their bioactivities.2a According to the database, as represented
Resveratrol (5-[(E)-2-(4-hydroxyphenyl)ethenyl]benzene-1,3-diol),
a key building block of stilbenoids, has drawn much attention
due to its diverse biological activities, such as cancer
chemopreventive activity and antiaging effects.1 Stilbenoids are
typically found as resveratrol oligomers (ResOligos) in limited
plant families, such as Dipterocarpaceae and Vitaceae.2 As the
increasing number of ResOligos and their attractive bioactivities
has been unveiled over the last several decades, the importance
of their structural elucidation has increased.
Different ResOligo structures result from skeletal variations and
the presence of stereoisomers. In fact, ResOligos in nature are
regioselectively biosynthesized via the oxidative coupling of resve-
ratrols to construct several specific skeletal structures bearing
condensed heterocyclic and bicyclic ring systems. The structural
elucidation of ResOligos is typically performed utilizing the follow-
ing approaches: (1) classification of biosynthetic systems from the
isolation source, (2) identification of building blockings by NMR,
and (3) specification of atomic connections using NMR correlation
techniques.
by vaticanols A–J3 and pauciflorols A–E,4 it is empirically implied
that most ResOligos are constructed from three representative sub-
structures A–C, all of which are chiral and not easily interchange-
able. The phenolic units within each of these three substructures
are typically connected via asymmetric carbon linkages, and the
number of these chiral centers is generally in proportion to the
oligomerization degree, that is, trimers have six chiral atoms in
extreme cases (Fig. 1).
With respect to the connections between these chiral substruc-
tures, the absolute configurations of ResOligos are not always
affordable solely by NMR. When NMR data is insufficient to distin-
guish enantiomers, either X-ray diffraction or electronic circular
dichroism (ECD) analysis is required. Because good crystals for
X-ray structural analysis of ResOligos are often difficult to obtain,
ECD spectra are generally more helpful. In addition, ECD measure-
ments are typically easy and fast and can be performed for
molecules of any size in low-concentration solutions.5 During the
course of our investigations, we discovered that the absolute con-
figurations of ResOligos can be reproduced by the mutual addition
or subtraction of the ECD spectra of the substructures.6 In other
words, an important but missing piece of the above mentioned
ResOligo database is solid evidence for the correlation between
the Res substructures and their ECD patterns. However, given that
ECD analyses, with the aid of theoretical calculations, have been
more widely used for the determination of the absolute
We have isolated several hundreds of ResOligos from plants in
the Dipterocarpaceae family and developed
a fundamental
⇑
Corresponding author. Tel.: +81 58 230 8100; fax: +81 58 230 8105.
E-mail address: teito@gifu-pu.ac.jp (T. Ito).
0040-4039/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tetlet.2013.10.106

T. Ito, T. Nehira / Tetrahedron Letters 55 (2014) 314–318
315
OH
OH
HO
O
O
OH
OH
OH
HO
O
*
*
HO
*
*
*
*
*
*
HO
*
*
*
*
HO
HO
OH
OH
OH
OH
OH
OH
C
A
B
* asymmetric carbons
Figure 1. Building substructures (A–C) found in resveratrol derivatives.
configurations of natural products,7 we studied the ECD spectra of
ResOligos to obtain additional practical information for the
ResOligo database (see Fig. 2).
structural elucidations was overcome by applying a ‘hybrid’ chem-
ical correlation; namely, the stoichiometric dehydroxylation12 of
1a and 2a to the target molecules was conducted experimentally,
and then standard theoretical ECD calculations were performed
for the obtained dehydroxylated structures.
Results and discussion
Dehydroxylation (or deoxygenation) was achieved via the
cleavage of the phenolic hydroxy groups of the corresponding aryl
triflates. Thus, the Pd/C-Mg-mediated dehydroxylation developed
by Mori et al.12a,13 was applied to the triflate (1b) of melanoxylin
A (1a: A) isolated from the leaves of Vateria indica,14 which suc-
cessfully resulted in dehydrogenation to yield the desired product
(1c)15 (Scheme 1). For the second substrate (2R,3R)-3-(3,5-dihy-
droxyphenyl)-6-hydroxy-2-(4-hydroxyphenyl)-2,3-dihydrobenzo-
We initially performed a conformational search using CONFLEX
6 with the MMFF94S force field8 for melanoxylin A9 [syn. (l)-amp-
elopsin B10] (1a: A) and (2R,3R)-3-(3,5-dihydroxyphenyl)-6-hydro-
xy-2-(4-hydroxyphenyl)-2,3-dihydrobenzofuran-4-carbaldehyde11
(2a: B+CHO), whose absolute configurations have already been
established. However, the calculations resulted in numerous stable
conformers [1a (23 out of 146 conformers showed >1%) and 2a (17
out of 32 conformers showed >1%)], where no conformer occupied
greater than 20% population of abundance for either compound. An
important precondition for theoretical ECD calculations is that the
furan-4-carbaldehyde (2a: B+CHO),10 to obtain 2c, (l)-
e-viniferin
(4a) isolated from the stem of Vatica arbiramis16 was selected as
the starting material because the direct dehydroxylation of 2a
was not successful. Because 4b bears an olefinic bond, dehydroxy-
lation was performed via the Pd(OAc)2-triethylammonium-for-
mate-mediated reaction developed by Cacchi et al.,17 which
succeeded in the removal of triflate groups without olefin reduc-
tion to yield 4c.15,18 The osmium tetroxide-catalyzed cleavage of
4c afforded the desired aldehyde 2c bearing two asymmetric car-
bons with the R-configuration (Scheme 2).19 The structures of all
products were elucidated via MS and NMR spectral analyses, and
then the ECD spectra of 1c and 2c were measured (solid curves
in Fig. 3).
molecule of interest should have
a finite number of stable
conformers.7 Thus, the results for 1a and 2a were not promising.
This conformational diversity for ResOligos is presumably due to
the presence of numerous flexible phenolic groups, which can lead
to many stable conformers. This limitation of such standard
For computational simulations, the established absolute config-
urations of dehydroxylated melanoxylin (1c: A) and ‘B + CHO’ (2c)
were appropriately constructed. Both conformational searches by
CONFLEX 6/MMFF94S8 for the dehydroxylated derivatives (1c
and 2c) resulted in only two stable conformers (Supplementary
data). These two conformers were further optimized by density
functional theory (DFT) calculations at the B3LYP/6-31G(d)
10% Pd/C (0.15 eq)
Mg (10 eq)
NH4OAc (5 eq)
MeOH
1b
1c
Ar (balloon), rt, 24 h
Scheme 1. Pd/C-catalyzed reductive cleavage of the TfO groups of 1b.
Pd (OAc)2 (0.1 eq)
phosphine ligand (0.2 eq)
Et3N (15 eq)
HCOOH (10 eq)
DMF
OsO4 (cat)
NaIO4 (2.0 eq)
dioxane
4b
4c
2c
Ar (balloon), 80°C, 2 h
rt, 2 h
Figure 2. Structures of resveratrol derivatives (1aꢀ4a), their triflates (1b, 3b, 4b),
and dehydroxylated derivatives (1cꢀ4c).
Scheme 2. Pd(OAc)2-catalyzed reductive cleavage of 4b followed by osmium-
tetroxide-catalyzed oxidation of the olefinic bond.

316
T. Ito, T. Nehira / Tetrahedron Letters 55 (2014) 314–318
approximation level. The two optimized stable conformers were
then subjected to energy calculations by the time-dependent den-
sity functional theory (TDDFT) at the B3LYP/TZVP approximation
level.20 All calculated rotational strengths were converted into
Gaussian-type curves and added for the calculated ECD spectrum
of each conformer. After correction using the Boltzmann distribu-
tion of the conformers, the weighted ECD spectra were added to
afford the ECD spectrum for both 1c and 2c (Fig. 3 a and b).
A comparison of the experimental and theoretical ECD spectra
for both 1c and 2c proved the feasibility of dehydroxylation. Since
the relative configurations of the molecules in focus have been
established, we simply compared the overall shapes of the experi-
mental ECD with the theoretical one that is derived from one enan-
tiomer. The two ECD curves for 1c were similar (Fig. 3a); they
clearly showed a negative peak at 225 nm and a positive peak at
205 nm. In the case of 2c, the two peaks at 205 and 225 nm were
both negative (Fig. 3(b)), and the difference in the sign of the peak
at 205 nm for 1c and 2c was obviously attributable to the extra
chiral carbon in 1c. Considering the additivity rules for ECD
spectra, this result implied that computational simulations were
practically helpful because the spectral differences were assumed
to be due to the complex interactions between all transition
moments of the aromatic groups.
Having achieved the successful reproduction of the ECD spectra
of ResOligos with known absolute configurations, our attention
naturally turned to the determination of the absolute configura-
tions of other ResOligos using their dehydroxylated derivatives.
Accordingly, the simple interplay of a chemical conversion and a
computational simulation was further applied to the determina-
tion of the absolute configuration of other representative ResOligo
substructures.
Dehydroxylation was first applied to (l)-pauciflorol B4a (3a: C),
whose absolute configuration has not yet been assigned. The dehy-
droxylated product (3c) was obtained via triflate (3b) using the
same protocol as described for 1c. The ECD spectrum was then
measured (Fig. 3(c)). For the conformational search using the CON-
FLEX 6/MMFF94S method, a temporary absolute configuration of
authentic 3a bearing the (2R,3R)-2,3-dihydrobenzofuran skeleton
was assumed. As a result, more than 6500 stable conformers were
identified, with no conformer occupying over 2% population of
abundance. On the other hand, the conformational search for dehy-
droxylated 3c yielded only two stable conformers (Supplementary
data). After further optimization by DFT calculations at the B3LYP/
6-31G(d) approximation level and TDDFT energy calculations at
the B3LYP/TZVP approximation level, all rotational strengths of
each conformer were converted into Gaussian-type curves and
added to afford the ECD spectrum. After correction according to
Boltzmann’s law, the ECD curves for the two conformers were
combined to afford the calculated ECD spectrum of 3c (Fig. 3(c)).
The calculated ECD curve for 3c was then compared to the
experimentally obtained ECD spectrum. Because the relative con-
figuration of 3c was previously determined from NMR data, by
considering the signs of the peaks at 225 and 210 nm, the absolute
configuration of 3c, dehydroxylated derivative of 3a, was deter-
mined to be (3S,4S,4aR,5R,9bR,10R)-3,4,5,10-tetrakisphenyl-
3,4,4a,5,9b,10-hexahydrobenzo[5,6]azuleno[7,8,1-cde]benzofuran,
which deductively corresponded to the absolute configuration
of 3a.
Figure 3. Experimental and calculated CD spectra of dehydroxylated derivatives
(a) 1c, (b) 2c, and (c) 3c. CD calculations were performed at the B3LYP/TZVP level.
This determination can strengthen the ResOligo database
because the structure of 3a is identical to one of the three repre-
sentative building substructures (C in Fig. 1) of the ResOligos. Anal-
yses with the ResOligo database involve the mutual addition or
subtraction of ECD spectra, mainly at 220–240 nm and optionally
at 205–215 nm. Thus, the present method can verify the database
analysis results in these wavelength regions. There are generally
two types of information for each ResOligo in the database: relative
configurations established by NMR and combination of the three
required building substructures A–C. Therefore, the comparison
of experimentally obtained and calculated ECD spectra may enable
the determination of the absolute configuration of each building
substructure and provide additional (or interactional) ECD results
among the building structures of interest.

T. Ito, T. Nehira / Tetrahedron Letters 55 (2014) 314–318
317
We know that compounds 1a–3a are the simplest forms of skel-
References and notes
etal variations of ResOligos in the family Dipterocarpaceae. This
study provided two important findings. First, because the absolute
configuration of 1a and 2a was previously established, the results
obtained by dehydroxylation can be confirmed to agree with the
preceding structural determinations. Second, the absolute configu-
ration of 3c has been unambiguously determined using the dehy-
droxylation-computation (‘hybrid’) method based on a chemical
correlation, allowing the determination of the absolute configura-
tion of 3a.
The dehydroxylation (or deoxygenation) of phenols was origi-
nally applied in structure–activity relationship (SAR) studies, be-
cause the desired products of dehydroxylation enable the
identification of the contribution of hydroxy groups to biological
activity.12a Recently, we also applied this method to the SAR study
of resveratrol to identify the role of hydroxy groups in expressing
anti-androgenic activity.12b In this study, for the first time, dehydr-
oxylation was used for the determination of absolute configura-
tion. It should be emphasized that dehydroxylation can
dramatically reduce the number of dominant conformations and
enable ECD calculations. The principal consideration in this proce-
dure is the complete removal of phenolic hydroxy groups from
mother compounds, because any remaining flexible hydroxy
groups will dramatically increase the number of stable
conformers.
1. (a) Jang, M.; Cai, L.; Udeani, G. O.; Slowing, K. V.; Thomas, C. F.; Beecher, C. W.;
Fong, H. H.; Farnsworth, N. R.; Kinghorn, A. D.; Mehta, R. G.; Moon, R. C.;
Pezzuto, J. M. Science 1997, 275, 218–220; (b) Howitz, K. T.; Bitterman, K. J.;
Cohen, H. Y.; Lamming, D. W.; Lavu, S.; Wood, J. G.; Zipkin, R. E.; Chung, P.;
Kisielewski, A.; Zhang, L. L.; Scherer, B.; Sinclair, D. A. Nature 2003, 425, 191–
196.
2. (a) Ito, T. Yakugaku Zasshi 2011, 131, 93–100; (b) Riviere, C.; Pawlus, A. D.;
Merillon, J. M. Nat. Prod. Rep. 2012, 29, 1233–1317.
3. (a) Tanaka, T.; Ito, T.; Nakaya, K.; Iinuma, M.; Riswan, S. Phytochemistry 2000,
54, 63–69; (b) Ito, T.; Tanaka, T.; Nakaya, K. I.; Iinuma, M.; Takahashi, Y.;
Naganawa, H.; Ohyama, M.; Nakanishi, Y.; Bastow, K. F.; Lee, K. H. Tetrahedron
2001, 57, 7309–7321.
4. (a) Ito, T.; Tanaka, T.; Nakaya, K. I.; Iinuma, M.; Takahashi, Y.; Naganawa, H.;
Ohyama, M.; Nakanishi, Y.; Bastow, K. F.; Lee, K. H. Tetrahedron 2001, 57, 7309–
7321; (b) Ito, T.; Tanaka, T.; Iinuma, M.; Iliya, I.; Nakaya, K. I.; Ali, Z.; Takahashi,
Y.; Sawa, R.; Shirataki, Y.; Murata, J. Tetrahedron 2003, 59, 5347–5363.
5. Berova, N.; Nakanishi, K.; Woody, R. W. Circular Dichroism Principles and
Applications, 2nd ed.; Wiley-VCH: New York, 2000.
6. Ito, T.; Ito, H.; Oyama, M.; Sawa, R.; Yakahashi, Y.; Nehira, T.; Darnaedi, D.;
Tanaka, T.; Murata, J.; Iinuma, M. In 54th Symposium on the Chemistry of Natural
Products: Symposium Papers, 2012; pp 519–524.
7. (a) Berova, N.; Di Bari, L.; Pescitelli, G. Chem. Soc. Rev. 2007, 36, 914–931; (b)
Bringmann, G.; Bruhn, T.; Maksimenka, K.; Hemberger, Y. Eur. J. Org. Chem.
2009, 17, 2717–2727.
8. All conformational searches were performed with CONFLEX 6 (Ver. 6.89 by
CONFLEX, Tokyo) using
a commercially available PC (operating system:
Windows7 Professional SP1 64-bit, CPU: QuadCore Xeon E3-1225 processor
3.10 GHz, RAM 8 GB). The initial structure was constructed on a graphical user
interface considering the absolute configuration of interest and was subjected
to conformational search adopting MMFF94S (2010–12-04HG) as the force
field, where initial stable conformers were generated for up to 50 kcal/mol.
9. Matsuda, H.; Asao, Y.; Nakamura, S.; Hamao, M.; Sugimoto, S.; Hongo, M.;
Pongpiriyadacha, Y.; Yoshikawa, M. Chem. Pharm. Bull. 2009, 57, 487–494.
10. Oshima, Y.; Ueno, Y.; Hikino, H.; Ling-Ling, Y.; Kun-Ying, Y. Tetrahedron 1990,
46, 5121–5126.
11. Ito, T.; Furusawa, M.; Iliya, I.; Tanaka, T.; Nakaya, K. I.; Sawa, R.; Kubota, Y.;
Takahashi, Y.; Riswan, S.; Iinuma, M. Tetrahedron Lett. 2005, 46, 3111–3114.
12. (a) Mori, A.; Mizusaki, T.; Ikawa, T.; Maegawa, T.; Monguchi, Y.; Sajiki, H. Chem.
Eur. J. 2007, 13, 1432–1441; (b) Iguchi, K.; Toyama, T.; Ito, T.; Shakui, T.; Usui,
S.; Oyama, M.; Iinuma, M.; Hirano, K. J. Androl. 2012, 33, 1208–1215.
13. (a) Representative procedure for the preparation of triflates of oligomeric
resveratrol derivatives: Excessive trifluoromethanesulfonic anhydride (0.1 mL,
The assignment of the absolute configuration of new
compounds is conducted by either X-ray crystallography, which
requires single crystals, NMR-modified Mosher’s method for sec-
ondary alcohols, or ECD exciton chirality analysis for which the
complete relative configuration must be known. However, these
methodologies are difficult to apply to ResOligos because these
compounds are typically amorphous solids that lack a secondary
alcohol2 and often some relative configurations are unknown.
Other approaches have been undertaken for the elucidation of oli-
gostilbenoids, including the comparative analysis of the absolute
1.2 mmol) was added to a solution of 3a (27.2 mg, 40 lmol) in pyridine
(1.0 mL) at 0 °C. The reaction mixture was kept at room temperature for 12 h.
The mixture was diluted with ice water (20 mL) and extracted with EtOAc
(20 mL ꢁ 3). The organic phase was washed with saturated NaHCO3, saturated
CuSO4, water, and brine, and dried over NaSO4. Filtration and evaporation in
vacuo yielded a dark brown solid (49 mg). Preparative TLC on silica gel (n-
configuration of the b-D
-glucopyranosyl group,21 and semisyn-
thetic preparation using authentic mother compounds.22 The pres-
ently suggested approach is a simple and practical combination of
established reactions and nondestructive simulations, and thus is
an eligible option for ResOligos and presumably numerous other
polyphenols.
hexane-EtOAc 6:1) yielded 3b as a clear solid (38.8 mg, 22.4
Procedure for Pd/C-Mg-mediated dehydroxylation: After two vacuum/Ar
cycles to remove air from the reaction tube, mixture of 3b (30.0 mg,
17.3 mol), 10% Pd/C (6.0 mg, 10 wt % of 3b), NH4OAc (10.8 mg, 140 mol
(1 equiv for OTf groups)), and Mg metal (6.7 mg, 275 mol (2 equiv for OTf
lmol, 56%); (b)
a
l
l
Due to the lack of chiroptical properties of ResOligos in the
database, an empirical ECD approach (determination of the struc-
ture and connection patterns by addition and/or subtraction of
experimental ECD spectra) is not completely feasible at present.
Although there are a few reports of the structural elucidation of
naturally occurring ResOligos based on computational
approaches,23 as far as we know, neither conformational analyses
nor the contribution of hydroxy groups to conformational flexi-
bility have been evaluated. We have clarified that the flexible
phenolic hydroxy groups of ResOligos not only control the bioac-
tivity but also enhance their conformational stabilization in prot-
ic solvents.
In summary, the current Letter reports the first application of
the dehydroxylation of ResOligos, the building blocks of stilbe-
noids, to the configurational assignment of these important struc-
tures. As demonstrated with 1a–3a, the method is a simple and
practical combination of chemical conversion and computational
simulation. Thorough study of the applications of this method to
compounds in the ResOligo database is underway.
l
groups)) in methanol (1.0 mL) was stirred at ordinary pressure (balloon) and
temperature (ca. 20 °C) for 24 h. The reaction mixture was filtered using a
membrane filter (Millipore, Millex-LH, 0.45 l), and the filtrate was partitioned
between EtOAc (10 mL) and water (10 mL). The aqueous layer was extracted
with EtOAc (10 mL ꢁ 3), and the combined organic layers were washed with
brine (10 mL), dried with anhydrous Na2SO4, filtered, and concentrated under
reduced pressure to yield a brown solid (12 mg). Preparative TLC (n-hexane-
EtOAc 6:1) yielded 3c as a clear solid (6.9 mg, 12.5 lmol, 72%).
14. Ito, T.; Masuda, Y.; Abe, N.; Oyama, M.; Sawa, R.; Takahashi, Y.; Chelladurai, V.;
Iinuma, M. Chem. Pharm. Bull. 2010, 58, 1369–1378.
15. All triflated and dehydroxylated products were characterized by their 1H NMR
and MS spectral data. 1b: 1H NMR (400 MHz, CDCl3, TMS int.): d 7.31 (1H, d,
J = 2.0 Hz, ArH), 7.27 (2H, d, J = 9.0 Hz, ArH), 7.23 (2H, d, J = 9.0 Hz, ArH), 7.17
(2H, d, J = 8.8 Hz, ArH), 7.12 (1H, br s, ArH), 7.09 (2H, d, J = 8.8 Hz, ArH), 6.86
(1H, d, J = 2.0 Hz, ArH), 6.64 (1H, d, J = 2.4 Hz, ArH), 5.93 (1H, d, J = 11.2 Hz, CH),
5.15 (1H, br s, CH), 4.10 (1H, d, J = 11.2 Hz, CH), 3.83 (1H, dd, J = 18.0, 4.0 Hz,
CH), 3.55 (1H, br d, J = 18.0 Hz, CH) ppm. ESIMS: m/z 1112.8803 [MꢀH]ꢀ (calcd
for C33H16O16F15S5, 1112.8808); 1c: 1H NMR (400 MHz, CDCl3, TMS int.): d 7.44
(1H, br d, J = 8.0 Hz, ArH), 7.30–7.09 (14H, m, ArH), 6.85 (1H, d, J = 8.0 Hz, ArH),
6.63 (1H, d, J = 8.0 Hz, ArH), 5.96 (1H, d, J = 12.0 Hz, CH), 4.53 (1H, br s, CH),
4.34 (1H, d, J = 12.0 Hz, CH), 3.82 (1H, dd, J = 18.0, 4.0 Hz, CH), 3.51 (1H, br d,
J = 18.0 Hz, CH) ppm. ESIMS: m/z 375.1692 [M+H]+ (calcd for C28H23O,
375.1743). CD nm (c = 26.7 lM, i-PrOH) (De): 205 (+28.5), 209 (+22.6), 225
(ꢀ76.8); 2c: 1H NMR (500 MHz, CDCl3, TMS int.): d 9.78 (1H, s, CHO), 7.45 (1H,
t, J = 7.7 Hz, ArH), 7.42 (1H, dd, J = 7.7, 1.7 Hz, ArH), 7.38–7.25 (9H, m, ArH),
7.17 (2H, m, ArH), 5.68 (1H, d, J = 5.2 Hz, CH), 4.98 (1H, d, J = 5.2 Hz, CH) ppm.
Supplementary data
EIMS: m/z 300.1155 [M]+ (calcd for C21H16O2, 300.1150). CD nm (c = 33.3
lM, i-
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.tetlet.2013.
10.106.
PrOH) (
D
e
): 228 (ꢀ34.1), 265 (+8.2); 3b: 1H NMR (500 MHz, CDCl3, TMS int.): d
7.40 (2H, d, J = 8.6 Hz, ArH), 7.32 (2H, d, J = 8.6 Hz, ArH), 7.27 (2H, d, J = 8.6 Hz,
ArH), 7.26–7.19 (8H, m, ArH), 7.17 (1H, t, J = 2.0 Hz, ArH), 7.02 (1H, d, J = 1.8 Hz,

318
T. Ito, T. Nehira / Tetrahedron Letters 55 (2014) 314–318
ArH), 7.00 (2H, br s, ArH), 6.07 (1H, d, J = 11.5 Hz, CH), 4.98 (1H, d, J = 2.9 Hz,
organic layers were washed with brine (20 mL), dried (Na2SO4), and
evaporated. Chromatography of the residue (56 mg) on silica gel (20 g, n-
hexane–EtOAc system) provided 23.6 mg (44%) of 4c.
CH), 4.73 (1H, d, J = 10.3 Hz, CH), 4.33 (1H, d, J = 12.0 Hz, CH), 4.14 (1H, br d,
J = 12.0 Hz, CH), 3.81 (1H, t, J = 12.0 Hz, CH) ppm. ESIMS: m/z 1734.8102
[MꢀH]ꢀ (calcd for C51H27O22F24S9, 1734.8103); 3c: 1H NMR (500 MHz, CDCl3,
TMS int.): d 7.41–7.39 (2H, m, ArH), 7.36–7.31 (4H, m, ArH), 7.28–7.24 (6H, m,
ArH), 7.23–7.12 (11H, m, ArH), 7.03 (1H, d, J = 7.5 Hz, ArH), 6.79 (1H, d,
J = 8.0 Hz, ArH), 6.70 (1H, d, J = 8.0 Hz, ArH), 6.03 (1H, d, J = 12.6 Hz, CH), 4.61
(1H, d, J = 12.6 Hz, CH), 4.47 (1H, d, J = 10.3 Hz, CH), 4.36 (1H, d, J = 3.5 Hz, CH),
4.09 (1H, br d, J = 11.5 Hz, CH), 3.99 (1H, dd, J = 11.5, 10.3 Hz, CH) ppm. ESIMS:
19. Osmium tetroxide-catalyzed cleavage of 4c: A mixture of dioxane (3 mL), 4c
(37.4 mg, 0.1 mmol), and osmium tetroxide (2.0 mg) was stirred for 5 min and
maintained at 24–26 °C. During this process, sodium metaperiodate (42.8 mg,
0.2 mmol) was added over a period of 50 min. The tan slurry was then stirred
for an additional 1.5 h to yield degradation products 2c (13 mg, 43 lmol, 43%).
20. All DFT calculations were performed with Gaussian 09 (Revision A.02 by
Gaussian, Wallingford, CT) with a PC (operating system: CentOS a Linux, CPU: 2
Intel Xeon X5550 processors 2.67 GHz, RAM 24 GB). Stable conformers with
distributions over 1% obtained from the conformational searches were further
optimized by the DFT method assuming no solvent with B3LYP functional and
6-31G(d) basis set. The obtained conformers analyzed their populations by
considering their Boltzmann distribution at 298 K on the basis of their internal
energies. All conformations of distributions over 1% were subjected to time-
dependent simulations at B3LYP/TZVP approximation level. For each
conformer, the resultant rotational strengths were converted into Gaussian-
m/z 575.2325 [M+Na]+ (calcd for C42H32ONa, 575.2345). CD nm (c = 18.1
lM, i-
PrOH) (
D
e
): 205 (+130.0), 229 (ꢀ97.0); 4b: 1H NMR (400 MHz, acetone-d6, TMS
int.): d 7.69 (2H, d, J = 9.0 Hz, ArH), 7.69 (3H, br s, ArH), 7.53 (2H, d, J = 9.0 Hz,
ArH), 7.52 (2H, d, J = 9.0 Hz, ArH), 7.46 (1H, d, J = 2.2 Hz, ArH), 7.33 (2H, d,
J = 9.0 Hz, ArH), 7.32 (1H, d, J = 15.8 Hz, ACH@), 7.11 (1H, d, J = 2.2 Hz, ArH),
6.96 (1H, d, J = 15.8 Hz, ACH@), 6.01 (1H, d, J = 6.0 Hz, CH), 5.43 (1H, d,
J = 6.0 Hz, CH) ppm. FABMS: m/z 1112.8785 [MꢀH]ꢀ (calcd for C33H16O16F15S5,
1112.8802); 4c: 1H NMR (400 MHz, acetone-d6, TMS int.): d 7.40–7.15 (15H, m,
ArH), 7.36 (1H, t, J = 2.4 Hz, ArH), 7.19 (1H, d, J = 6.6 Hz, ArH), 7.07 (1H, d,
J = 16.8 Hz, ACH@), 6.92 (1H, dd, J = 6.6, 2.4 Hz, ArH), 6.88 (1H, d, J = 16.8 Hz,
ACH@), 5.62 (1H, d, J = 5.8 Hz, CH), 4.86 (1H, d, J = 5.8 Hz, CH) ppm. EIMS: m/z
374.1677 [M]+ (calcd for C28H22O2, 374.1671).
type curves (bandwidth sigma = 2000 cmꢀ1
) and added to give the ECD
spectrum. The theoretical ECD spectrum for the compound was composed
after correction based on the Boltzmann distribution of the stable conformers
with over 1%.
16. Abe, N.; Ito, T.; Ohguchi, K.; Nasu, M.; Masuda, Y.; Oyama, M.; Nozawa, Y.; Ito,
M.; Iinuma, M. J. Nat. Prod. 2010, 73, 1499–1506.
17. Cacchi, S.; Ciattini, P. G.; Morera, E.; Ortar, G. Tetrahedron Lett. 1986, 27, 5541–
5544.
18. Procedure for Pd(OAc)2–triethylammonium–formate-mediated dehydr-
oxylation: After two vacuum/Ar cycles to remove air from the reaction tube,
21. (a) Ito, T.; Hoshino, R.; Hara, Y.; Oyama, M.; Iinuma, M. Phytochemistry Lett.
2013, 6, 193–197; (b) Abe, N.; Ito, T.; Oyama, M.; Sawa, R.; Takahashi, Y.;
Chelladurai, V.; Iinuma, M. Chem. Pharm. Bull. 2011, 59, 239–248.
22. (a) Ito, T.; Oyama, M.; Sajiki, H.; Sawa, R.; Takahashi, Y.; Iinuma, M. Tetrahedron
2012, 68, 2950–2960; (b) Takaya, Y.; Yan, K. X.; Terashima, K.; He, Y. H.; Niwa,
M. Tetrahedron 2002, 58, 9265–9271; (c) Takaya, Y.; Yan, K. X.; Terashima, K.;
Ito, J.; Niwa, M. Tetrahedron 2002, 58, 7259–7265.
23. (a) He, C.-N.; Peng, Y.; Xu, L.-J.; Liu, Z.-A.; Gu, J.; Zhong, A.-G.; Xiao, P.-G. Chem.
Pharm. Bull. 2010, 58, 843–847; (b) Qin, Y. H.; Zhang, J.; Cui, J. T.; Guo, Z. K.;
Jiang, N.; Tan, R. X.; Ge, H. M. RSC Advances 2011, 1, 135–141.
a
mixture of 4b (114 mg, 0.1 mmol), Pd(OAc)2 (2.0 mg, 0.01 mmol), Et3N
(0.21 mL, 1.5 mmol), and triphenylphosphine (5.0 mg, 0.02 mmol) in DMF
(2.0 mL) was added to 99% formic acid (37.5 L, 1.0 mmol). The mixture was
l
stirred at ordinary pressure (balloon) and temperature (ca. 80 °C) for 2 h. The
reaction mixture was partitioned between EtOAc (20 mL) and water (20 mL).
The aqueous layer was extracted with EtOAc (20 mL ꢁ 3) and the combined