E. P. Mazzola et al. / Carbohydrate Research 346 (2011) 759–768
767
in DMSO (Sigma–Aldrich). The cells were treated for 4 days, after
which they were incubated with MTT reagents and the absorbance
read at 600 nm as previously described.31 Actein standard for this
assay was obtained from ChromaDex (Laguna Hills, CA). Cell viabil-
ity was calculated by comparing cell counts in treated samples rel-
ative to cell counts in the untreated group.
HSQC spectra. Squared sine-bell window functions shifted by
/2 rad were used in both dimensions for most of the 2D experi-
ments, except that no shift was used for 2D COSY-30 data, and a
/4 rad shift was used for coupled 2D HSQC and HMBC. The COSY
and HMBC spectra were displayed in absolute value mode. All
other 2D spectra were acquired in the phase-sensitive, echo/anti-
echo mode.
p
p
3
.4. Acid hydrolysis
4
A solution of 4.7 mg of blighoside B in 0.3 mL of methanol-d was
prepared in a small test tube and the resulting solution transferred
to a Shigemi NMR microcell whose glass was susceptibility matched
to methanol. H, C, DEPT, HSQC, 100-ms mixing-time TOCSY, and
ROESY NMR data were acquired and processed using the same
experimental conditions as those described for blighoside A
Solutions of compounds (1–3) (2 mg each), consisting of 1 N
HCl–MeOH (1:4 v/v, 5 mL, were placed in an incubator at 50 °C
for 2 days. The solutions were then diluted with 10 mL HPLC grade
3
water and neutralized with a 10% NaHCO solution. The mixture
1
13
was then partitioned between EtOAc and water. The EtOAc layer
and the aqueous layers were dried under reduced pressure. The
EtOAc residues were analyzed by TLC using a solvent system of
NMR spectra of blighoside C at 500 MHz were acquired at 300 K
by means of a Bruker DRX-500 NMR spectrometer equipped with a
5 mm TXI HCN cryoprobe with z-gradient coil. Data acquisition and
processing were performed using Bruker TopSpin software version
1.3.4, running under Windows XP Pro SP2 on a HP xw4200 PC
workstation. A solution of 8.9 mg of blighoside C in 0.4 mL of meth-
8
:1 CHCl
3
–MeOH that revealed the presence of hederagenin
1
2
(
3b,23-dihydroxy-
D -oleanen-28-carboxylic acid) for compound
1, as compared with authentic standards.
4
anol-d was used, including tetramethylsilane as an internal chem-
ical shift reference (d = 0) for H and C NMR spectra. 1H NMR
1
13
3
.5. NMR methods
spectra were acquired using a spectral width of 4.01 kHz, 65,536
NMR spectra of blighoside A at 700 MHz were acquired at 298 K
point data sets, a 30° pulse (2.4 ls), and a pulse recycle time of
1
by means of a Bruker Avance-II-700 NMR spectrometer equipped
with a 5 mm QXI HCNP probe equipped with shielded xyz-gradient
coils. Data processing was performed using NMRPipe.32 A solution
8.2 s. The resolution of the H spectra was enhanced by Gaussian
multiplication of the FID, using a line-narrowing of –0.5 to –
0.75 Hz, and a Gaussian truncation fraction of 0.3.
1
3
of 19 mg of blighoside A in 0.3 mL of methanol-d
4
was prepared in
C NMR spectra were acquired by using 65,536 point data sets,
a small test tube and the resulting solution transferred to a Shigemi
NMR microcell whose glass was susceptibility matched to metha-
a 45° pulse (6.7 s), and a pulse recycle time of 2 s. Digital resolu-
l
tion was enhanced by forward linear prediction to 65,536 points.
1
13
nol. The residual H of the methyl resonance was used as an inter-
C spectral editing was performed by acquisition of DEPT spectra,
1
1
nal chemical shift reference (d = 3.31 ppm) for H and d = 49.1 ppm
using H read pulses of 45°, 90°, and 135°, but no linear prediction.
in the C NMR spectra. 1H NMR spectra were acquired using a
13
2 3
Separate CH, CH , and CH subspectra were generated from suit-
spectral width of 5.22 ppm, 16,384-point data sets, a 70° pulse
able combinations of the three DEPT spectra.
(
8
l
s), and a pulse recycle time of 2.5 s. No weighting or resolution
All 2D NMR spectra were acquired by pulsed field gradient-se-
lected methods. 2D COSY, TOCSY, ROESY, and TOCSY-less ROESY
1
enhancement was used for this 1D H spectrum.
C NMR spectra were acquired by using 65,536 point data sets,
a 70° pulse (10 s), and a pulse recycle time of 2.5 s. Exponential
1
3
1
(T-ROESY) were used to confirm H assignments. 2D HSQC, HMBC,
1
3
l
and 1D DEPT were used to assign C spectra. 2D COSY, ROESY,
T-ROESY, and HMBC NMR spectra were obtained by using 2048
multiplication with a line broadening of 1 Hz and an automatic
baseline correction were used.
(t
2048 (F
(F
) Â 4096 (F
taken from 2048 (t
and transformed to 8192 (F
ing with linear prediction was performed in the Topspin program
by using the forward complex mode, with 128 coefficients and
the final data sizes stated in the descriptions of the individual
2
) Â 512 (t
) Â 2048 (F
) points (HMBC), whereas 2D HSQC spectra were
) Â 512 (t ) point data sets, linearly predicted
) Â 2048 (F ) points. Spectral process-
1
) point data sets, zero-filled and transformed to
3
4
33
2
D ROESY and TOCSY spectra, using a DIPSI-2 mixing se-
2
1
) points (COSY, ROESY, and T-ROESY) or 8192
3
5
1
quence, were used to confirm H assignments. Two TOCSY exper-
iments were run: one with a 10-ms mixing time, and the other a
2
1
2
1
1
00-ms mixing time. These spectra were taken with the carrier
2
1
set to 2.88 ppm, a spectral window of 5.22 ppm, the initial incre-
ment in the indirect dimension set to one-half the dwell-time
(
increment) in this dimension, 8192 (t
2
) Â 512 (t
1
) point data sets,
) points. HSQC
zero-filled and processed to 16,384 (F
2
) Â 1024 (F
1
experiments. For 2D COSY, the read pulse was 30° (2.4
SY data were recorded using an isotropic mixing time of 300 ms,
and 16,384 (t ) point data sets, zero-filled and processed
) Â 512 (t
to 16,384 (F ) points. Three 2D ROESY spectra were ac-
ls). 2D TOC-
and HMBC spectra were used to correlate the hydrogens to their
respective or nearby carbons. In addition, coupled HSQC spectra
were taken to help assign the types of carbohydrate residues at-
tached to the aglycone. HSQC and coupled HSQC NMR spectra were
2
1
2
) Â 2048 (F
1
quired, using spin-lock pulse times of 100, 250, and 500 ms, and for
1
obtained by setting the carriers to 2.883 ppm and 100.4 ppm in H
three 2D T-ROESY spectra, spin-lock times of 250, 500, and
1000 ms were used. 2D HSQC and HMBC were acquired with
1
3
1
and
.22 ppm and 90 ppm, respectively, as well as 8192 (t
) point data sets; quadrature detection in the 13C dimension
C
dimensions, respectively, and spectral windows to
H
1
3
5
(t
1
2
) Â 512
and
respectively.
Values of 1JC-1,H-1 were measured from 1H-coupled 2D HSQC
spectra obtained from 2048 (t ) point data sets, linearly
) Â 512 (t
predicted as previously described and transformed to 8192
(F ) points. The CH coupling constants were extracted
) Â 2048 (F
from F slices of the coupled HSQC spectra. Sine-bell squared win-
dow functions shifted by /2 rad were used in both dimensions for
most of the 2D experiments, except that no shift was used for 2D
COSY-30 data, and a /4 rad shift was used for coupled HSQC. 2D
2 1
C spectral widths of 4.01 kHz (F ) and 25.1 kHz (F ),
was obtained using gradients for coherence selection. The coupled
HSQC data reported in this paper were measured by removing the
composite-pulse decoupling during FID detection. HMBC spectra
2
1
1
were acquired with the carrier set to 4.77 ppm in H and
2
1
1
3
1
6
1
9.6 ppm in
C, spectral windows of 6.0 ppm in H, and
2
1
3
20 ppm in C. These spectra were acquired in magnitude mode.
p
The raw FIDs were zero-filled and transformed to 1024
(
F
2
) Â 512 (F
1
) points.
J
p
1
C-1,H-1 were measured from 1H-coupled 2D HSQC
) Â 512 (t ) point data sets. The
CH coupling constants were extracted from F slices of the coupled
Values of
COSY and HMBC spectra were displayed in magnitude mode. All
other 2D spectra were acquired in the phase-sensitive, echo/anti-
echo mode.
spectra obtained from 2048 (t
2
1
2
Mazzola, Eugene P.
