NMR-Based Investigation of Hydrogen Bonding in a Dihydroanthracen-1(4 H)one from Rubia philippinensis and Its Soluble Epoxide Hydrolase Inhibitory Potential

Article abstract of DOI:10.1021/acs.jnatprod.8b00441

Hydrogen bonding is a vital feature of a large ensemble of chemical structures. Soluble epoxide hydrolase (sEH) has been targeted for development of the treatment for inflammation-associated diseases. Compounds 1 and 2 were purified from Rubia philippinensis, and their structures were established via physical data analysis. Compound 1 possesses intramolecular hydrogen bonding, sufficiently robust to transfer heteronuclear magnetization via a nonbonded interaction. The bonding strength was assessed using the ¹H NMR chemical shift temperature coefficients (-1.8 ppb/K), and the heteronuclear coupling constants were measured. The stereochemical details were investigated using interproton distance analysis and ECD. Purified compounds displayed moderate sEH-inhibitory activity.

Full text of DOI:10.1021/acs.jnatprod.8b00441

Article
pubs.acs.org/jnp
Cite This: J. Nat. Prod. XXXX, XXX, XXX−XXX
NMR-Based Investigation of Hydrogen Bonding in a
Dihydroanthracen-1(4H)one from Rubia philippinensis and Its
Soluble Epoxide Hydrolase Inhibitory Potential
Joonseok Oh,†,‡,# Khong Trong Quan,§,⊥,# Ji Sun Lee,§ InWha Park,§ Chung Sub Kim,†,‡
Daneel Ferreira,∥ Phuong Thien Thuong,⊥ Young Ho Kim,§ and MinKyun Na*,§
†Department of Chemistry, Yale University, New Haven, Connecticut 06520, United States
‡Chemical Biology Institute, Yale University, West Haven, Connecticut 06516, United States
§College of Pharmacy, Chungnam National University, Daejeon 34134, Republic of Korea
⊥Department of Herbal Analysis and Standardization, National Institute of Medicinal Materials, Hanoi, Vietnam
∥Department of BioMolecular Sciences, Division of Pharmacognosy, Research Institute of Pharmaceutical Sciences, School of
Pharmacy, The University of Mississippi, University, Mississippi 38677, United States
S
* Supporting Information
ABSTRACT: Hydrogen bonding is a vital feature of a large
ensemble of chemical structures. Soluble epoxide hydrolase (sEH)
has been targeted for development of the treatment for
inflammation-associated diseases. Compounds 1 and 2 were
purified from Rubia philippinensis, and their structures were
established via physical data analysis. Compound 1 possesses
intramolecular hydrogen bonding, sufficiently robust to transfer
heteronuclear magnetization via a nonbonded interaction. The
1
bonding strength was assessed using the H NMR chemical shift
temperature coefficients (−1.8 ppb/K), and the heteronuclear
coupling constants were measured. The stereochemical details
were investigated using interproton distance analysis and ECD.
Purified compounds displayed moderate sEH-inhibitory activity.
ntermolecular and intramolecular hydrogen-bonding net-
works are important features of chemical structures and
molecular hydrogen bond formation were directly measurable
utilizing several NMR experiments.5,7
I
reactivity in drug screening and development.1 One of the
most relevant examples of the former is vancomycin interacting
via a series of key intermolecular hydrogen bonding with the D-
Ala-D-Ala moieties found in the essential bacterial cell wall
precursor lipid II.2,3 The latter was formed between donor and
acceptor functionalities in the same molecule, when molecular
configuration and conformation placed them within hydrogen
bond geometry.1 Most drug-like molecules possess a number
of functional groups capable of forming these two types of
hydrogen bonds, leading ligands to specifically interact with
their biomolecular targets.4 Potential hydrogen bond for-
mation, therefore, should be considered in screening and
identification of drug prototypes.4 Among many approaches
for detecting the presence of hydrogen bonding, the NMR
method has been recognized as an important tool given its
sensitivity for configurational and conformational changes.5
For instance, temperature coefficients of chemical shifts of
hydroxy protons, prone to forming such nonbonding
Soluble epoxide hydrolase (sEH, EC 3.3.2.10) is an enzyme
playing a predominant role in hydrolyzing epoxy fatty acids
into the related diols in humans and other mammals.8 The
epoxy fatty acids, such as the epoxides of linoleic, arachidonic,
eicosatrienoic, and docosahexaenoic acids, are generated by a
cytochrome P450-catalyzed oxidation and are endogenous
mediators for regulating inflammation,8 pain,9 and blood
pressure.10 Thus, elevating levels of endogenous epoxy fatty
acids via inhibiting sEH has been validated as a viable strategy
for mitigating inflammation and pain11,12 and lowering blood
pressure.13 During the past few decades, this was substantiated
by extensive in vivo studies and clinical trials, and such
successes have inspired exploration of sEH inhibitory
prototypes for the control of the associated clinical disorders
and complications.14,15 Despite large molecular mass ureas,
such as 12-[{(tricyclo[3.3.1.13,7]dec-1-ylamino)carbonyl}-
amino]dodecanoic acid (AUDA), exerting powerful sEH
inhibitory activity, those scaffolds lack bioavailability due to
1
interactions, were obtained via variable-temperature (VT) H
NMR experiments, and the coefficients were used to evaluate
hydrogen-bonding strength.6 In addition, scalar coupling
constants between donors and acceptors involved in intra-
Received: June 1, 2018
© XXXX American Chemical Society and
American Society of Pharmacognosy
A
DOI: 10.1021/acs.jnatprod.8b00441
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
1
Table 1. H and 13C NMR Spectroscopic Data (600 MHz, 25 °C, DMSO-d6) for Compounds 1 and 2
1
2
position
δH, mult. (J in Hz)
δC
δH, mult. (J in Hz)
δC
1
2
187.6, C
132.6, C
8.55, s
121.4, CH
134.3, C
3
4
4a
5
6
7
8
8a
9
6.76, d (6.7)
2.77, m (17.3, 2.8), 2.38, dt (17.3, 6.7)
3.06, ddd (13.6, 6.7, 3.8)
7.41, dd (7.5, 1.1)
143.0, CH
24.4, CH2
35.3, CH
129.0, CH
132.1, CH
128.5, CH
124.4, CH
128.9, C
7.32, dd (9.0, 1.8)
8.62, d (9.0)
127.9, CH
123.5, CH
123.9, C
123.5,* CH
125.0,** CH
124.7,** CH
123.3,* CH
124.4,*** C
144.7, C
8.69, m
7.44, m
7.44, m
8.69, m
7.55, td (7.5, 1.1)
7.49, td (7.5, 1.1)
7.84, dd (7.5, 1.1)
171.0, C
9a
10
10a
11
9-OH
102.5, C
66.5, CH
141.1, C
125.2, C
146.4, C
125.4,*** C
21.9, CH3
4.46, brt (3.8)
1.82, brq (1.2)
15.5, CH3
2.50, overlap
a
16.07, s
a
10-OH
5.19, d (4.4)
9-O-glucopyranosyl
1′
2′
3′
4′
5′
6′
4.85, d (7.8)
3.59, d (7.8)
3.30, overlapped
3.30, overlapped
3.00, m
106.0
74.4
76.5
69.8
76.8
61.0
3.49, dd (11.5, 1.8)
3.42, dd (11.7, 4.8)
10-O-glucopyranosyl
1″
2″
3″
4″
5″
6″
4.86, d (7.8)
3.61, d (7.8)
3.30, overlapped
3.30, overlapped
3.00, m
106.0
74.5
76.5
69.8
76.8
61.1
3.52, dd (11.5, 1.8)
3.41, dd (11.4, 4.8)
a
Chemical shift values depending on various temperatures in acetone-d6 are available in Figures S10−S12, Supporting Information. *, **,
***Values within each column may be interchanged.
limited solubility in both water and organic solvents, which has
prompted a continuing search for potential chemical scaffolds
capable of inhibiting the target hydrolase.
formula of 1 as C15H14O3 (calcd [M + H]+, m/z 243.1021;
Figure S1, Supporting Information). The 1H NMR data
indicated characteristic resonances for four aromatic methines
(δH 7.41, 7.49, 7.55, 7.84), an olefinic proton (δH 6.76), an
oxygenated methine (δH 4.46), a methine (δH 3.06), a
methylene (δH 2.77, 2.38), and an allylic methyl motif (δH
1.82). The 13C and DEPT135 NMR data exhibited 15 carbons,
including an α,β-unsaturated carbonyl moiety (δC 187.6, 102.5,
171.0) and six aromatic and two olefinic carbons, together with
an oxygenated methine, a methine, and a methylene carbon.
Comparison of its 1H and 13C NMR data with rubiasin A from
R. cordifolia17 suggested that 1 is a dihydroanthracen-
1(4H)one-type molecule. This was confirmed via the COSY
correlations between H-3/H2-4, H-4/H-4a, and H-4a/H-10
(Figure 1). The HMBC cross-peaks of H3-11 (δH 1.82) and H-
3 (δH 6.76) with C-1 (δC 187.6) placed the ketocarbonyl
functionality at C-1, while HMBC cross-peaks of H-5 (δH
7.41) with C-10 (δC 66.5), H-4 (δH 2.77, 2.38) with C-2 (δC
132.6), and H3-11 (δH 1.82) with C-2 established the locations
of the hydroxy and methyl groups at C-10 and C-2,
respectively. Interestingly, a highly deshielded hydroxy proton
In an ongoing phytochemical study aimed at the
identification of bioactive secondary metabolites from Rubia
philippinensis Elmer (Rubiaceae), compounds 1 and 2 were
identified via a combinatorial approach of computation with
1
conventional and advanced NMR techniques. VT H NMR
studies ranging in temperature from 200 to 300 K permitted
the identification of intramolecular hydrogen bonding in 1.
Interestingly, this nonbonding interaction was sufficiently
robust to generate nonbonding scalar heterocoupling constants
(2JC‑1,OH and 3JC‑2,OH), which were directly measured using the
excitation-sculptured indirect-detection experiment (EX-
SIDE).16 The configurational analysis was carried out utilizing
comparison of experimental and simulated electronic circular
dichroism (ECD) spectra. These anthracene-type compounds
were also evaluated for their sEH inhibitory potential.
RESULTS AND DISCUSSION
■
Compound 1 was obtained as a yellow powder. The
protonated HRESIMS ion detected at m/z 243.1022 along
with the 13C NMR data (Table 1) established the molecular
1
resonance (δH 16.07) was observed in multiple H NMR
experiments using variable spectrum widths and anhydrous
B
DOI: 10.1021/acs.jnatprod.8b00441
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
16.07) to the C-1 carbonyl oxygen, transferring magnetization
from the hydroxy proton to C-1 and C-2 in the HMBC
experiment even with routine heteronuclear transfer delay and
coupling constant settings (nJCH = 8 Hz; Figure 1 and Figure
S7, Supporting Information). The long-range heteronuclear
coupling constants were measured utilizing EXSIDE to confirm
3
2JC‑1,OH (1.3 Hz) and JC‑2,OH (0.8 Hz) values (Figure 3). The
13C NMR chemical shift of C-1 along with the respective
heterocoupling constants from HO-9 to C-8a (4 Hz), C-9 (4
Hz), and C-9a (5 Hz) (Figures S13−S15, Supporting
Information), larger than those via the hydrogen bonding,7
confirmed the presence of the 4a,10-dihydroanthracen-1(4H)-
one (1) rather than a 4a,10-dihydroanthracen-9(4H)-one (1a)
motif (Figure 1). Collectively, these criteria defined the 2D
structure of 1 as shown in Figure 1.
The relative configuration was established based upon the
3
homonuclear coupling constant JH‑4a,H‑10 (3.8 Hz) and NOE-
derived interproton distances from H-10 to H-4a and the
diastereotopic protons H2-4. To accomplish the configura-
tional analysis via accurate interproton distances, the 1D
NOESY sequence based on the double-pulse field gradient
spin−echo NOE pulse and the peak amplitude normalization
for improved cross-relaxation (PANIC) method were
employed (Figure 4 and Figures S16 and S17, Supporting
Information).20,21 The PANIC-based calibrated interproton
distances from H-10 to H-4a, H-4α, and H-4β were 2.2, 2.7,
and 2.9 Å, respectively. These spatial distances exhibited an
excellent agreement with those simulated for a diastereomer of
1 possessing a (4aR*,10R*) configuration. For the establish-
ment of the absolute configuration of 1, Mosher’s method22
was initially attempted. Unlike the identical approach used for
rubiasin A,17 only the S-MTPA derivative could be formed
(Figure S18, Supporting Information). Alternatively, the
absolute configurational assignment was done via comparison
of the experimental and simulated ECD spectra. As shown in
Figure 5, both experimental and computed ECD spectra
exhibited the respective negative and positive Cotton effects at
ca. 250 and 280 nm originating from the UV bands of the
dihydroanthracen-1(4H)one architecture.17 Ultimately, the full
structure of 1 was established as (4aR,10R)-9,10-dihydroxy-2-
methyl-4a,10-dihydroanthracen-1(4H)-one.
The HRESIMS analysis of 2 (yellowish, amorphous powder)
exhibited a sodium-adduct ion at m/z 571.1787 (Figure S20,
Supporting Information), and the molecular formula was
assigned as C27H32O12 in conjunction with the 13C NMR data
(Table 1). The 1D NMR data of 2 shared close similarities
with those of typical 9,10-anthraquinones except for the
presence of the 13C NMR resonances for two oxygenated
aromatic tertiary carbons (δC 144.7, 146.4) and the absence of
two ketocarbonyl functionalities (ca. δC 180). This is indicative
of the two ketocarbonyl functionalities being reduced to the
corresponding phenolic moieties, which was confirmed by the
respective HMBC cross-peaks from H-1, H-8 and H-4, H-5 to
C-9 and C-10 (Figure 1). The C-2 methyl motif was
determined via the HMBC cross-peaks from H-1 and H-3 to
the methyl carbon and from H3-11 to C-1, C-2, and C-3
(Figure 1). The HMBC cross-peaks of the H-1′ and H-1″
anomeric protons (δH 4.85, 4.86) with the oxygenated
aromatic tertiary carbons (δC 144.7, 146.4) established the
locations of the monosaccharide moieties at C-9 and C-10.
The large coupling constants of the anomeric protons (J = 7.8
Hz) both indicated β-glucosidic linkages. The D-configuration
of the glucopyranose moieties was established using chiral
Figure 1. (A) 2D structures of compounds 1 and 2 and (B) key 2D
NMR correlations. The respective 2D correlations from hydroxy
protons were detected in anhydrous acetone-d6 or DMSO-d6.
aprotic NMR solvents (DMSO-d6 or acetone-d6) (Figure 2 and
Figures S2 and S12, Supporting Information). This suggests
that the exchangeable proton is involved in strong intra-
molecular hydrogen bonding. The hydroxy proton displayed
HMBC cross-peaks to C-9 (δC 171.0), C-9a (δC 102.5), and C-
8a (δC 128.9), as shown in Figure 1, rendering the deshielded
proton to be assigned tentatively to HO-9 (Figure 1). The
deshielded proton, however, also displayed HMBC cross-peaks
with the carbonyl carbon (δC 187.6) and C-2 (δC 132.6)
(Figure 1B). It was speculated initially that keto−enol
tautomerism involving the β-hydroxyenone functionality
might lead to the generation of such heteronuclear NMR
correlations on the NMR time scale. To verify this assumption,
VT 1H NMR experiments (200 to 300 K, acetone-d6, 10 mM)
were conducted in an attempt to observe the footprints of each
tautomer.18 However, based on the resultant NMR spectra
(Figure 2 and Figures S10−S12, Supporting Information),
such an interconversion apparently does not occur. Alter-
natively, a temperature coefficient [Δδ/ΔT(K)] of the hydroxy
proton was calculated to evaluate its hydrogen-bonding
strength since heteronuclear couplings from an exchangeable
hydrogen atom to heteroatoms (i.e., 13C, 15N, or 121Sb and
123Sb) were feasible via localized hydrogen bonding.5−7,19 The
temperature coefficient of the hydroxy proton in acetone-d6
was −1.8 ppb/K (200 to 300 K, acetone-d6, 10 mM; Figure 2),
much stronger than those measured for the hydrogen bonds
between HO-5 and the ketocarbonyl moieties in the flavonols
kaempferol (−3.1 ppb/K) and quercetin (−2.9 ppb/K) and
the flavone luteolin (−2.3 ppb/K) under similar VT NMR
experimental conditions.6 This evidenced the presence of a
robust intramolecular hydrogen bonding from HO-9 (δH
C
DOI: 10.1021/acs.jnatprod.8b00441
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
1
Figure 2. (A) VT H NMR experiments of 1 (200−300 K, acetone-d6; HO-9 and HO-10 are indicated in red and blue, respectively) and (B)
1
temperature coefficients of the two hydroxy protons HO-9 (−1.8 ppb/K) and HO-10 (−7.3 ppb/K) (10 mM). Some ranges for the stacked H
NMR spectra were snipped for better presentation. Another labile proton resonating at 3−4 ppm throughout the VT experiments is HDO.26
(Tokyo, Japan), IR data were generated on a Thermo Electron US/
Nicolet380 (Madison, WI, USA), and ECD spectra were recorded on
a Chirascan qCD (Applied Photophysics, UK). NMR experiments
were performed on a Bruker Avance III (600 MHz) spectrometer
(Billerica, MA, USA). HRESIMS data were generated on a Waters
SYNAPT G2 high-resolution mass spectrometer (Milford, MA, USA).
The VT experiments were conducted by cooling the probe using
liquid N2. TLC was performed on glass plates precoated with silica gel
60 F254 or RP-18 F254 (20 × 20 cm, 200 μm, 60 Å, Merck,
Kenilworth, NJ, USA). Vacuum-liquid chromatography (VLC) was
conducted on Merck silica gel (70−230 mesh), and MPLC was
carried out using a Biotage Isolera equipped with a reversed-phase C18
SNAP Cartridge KPC18-HS (120 or 400 g, Uppsala, Sweden). HPLC
was performed on a Gilson system (Middleton, WI, USA) with a
Phenomenex Kinetex 5 μm C18 column (250 × 21.20 mm, 5 μm,
Torrance, CA, USA).
derivatization followed by LC analyses (Figure S26, Supporting
Information).23 Consequently, the structure of 2 was
established as 9,10-di-O-β-D-glucopyranosyl-2-methylanthra-
cene.
The purified compounds were tested for their inhibitory
potential against sEH and exhibited inhibition with IC50 values
of 1.1 and 0.3 μM, respectively. These inhibitory activities are
relatively moderate in comparison with the positive control
AUDA (IC50 0.6 nM); however, 1 and 2 might be employed
for the generation/expansion of chemical libraries to imple-
ment preliminary lead optimization and SAR studies for the
development of sEH inhibitors, given the interesting scaffold
possessing robust hydrogen bonding and observed inhibitory
activity.
Plant Material. The roots of R. philippinensis were collected from
Bidoup-Nui Ba National Park, Lamdong Province, Vietnam, in August
2014, and identified by two of the authors, P.T.T. and M.N. A
voucher specimen was deposited at the herbarium of the Vietnam
EXPERIMENTAL SECTION
General Experimental Procedures. Optical rotations were
obtained on a JASCO DIP-1000 automatic digital polarimeter
■
D
DOI: 10.1021/acs.jnatprod.8b00441
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 3. Heteronuclear coupling constants for HO-9 to C-1 (2JC‑1,OH = 1.3 Hz) (A) and to C-2 (3JC‑2,OH = 0.8 Hz) obtained from EXSIDE spectra,
where those values were acquired via dislocation in Hz along with the f1 dimension divided by the scaling factor 15. See the Experimental Section
for details regarding the NMR parameters.
National Institute of Medicinal Materials, Hanoi, Vietnam
(VDL20140801), as well as the Laboratory of Pharmacognosy at
the College of Pharmacy, Chungnam National University, Daejeon,
Korea (CNU1409).
Extraction and Isolation. The dried roots of R. philippinensis (1.5
kg) were extracted with EtOH (3 L × 3) at room temperature for 4
days. The crude extract was concentrated under vacuum to yield a
reddish-brown slurry (150 g). The slurry was suspended in H2O (1.5
L) and sequentially partitioned with CH2Cl2 (3 × 2 L) and EtOAc (3
× 2 L) to yield the CH2Cl2, EtOAc, and H2O extracts. The CH2Cl2
fraction (50 g) was subjected to silica gel VLC and eluted with n-
hexane/EtOAc (20:1, 10:1, 5:1, 3:1, 2:1) and CHCl3/MeOH (8:1) to
yield six fractions (D-1−D-6). Fraction D-4 (6.1 g) was divided into
10 subfractions (D-4−1−D-4−10) using MPLC (column size 400 g)
with a gradient of MeOH/H2O (10:90 → 100:0, 7 L). Fractions D-4-
2 (180 mg) and D-4-3 (120 mg) were purified utilizing MeOH/H2O
(60:40, 4 mL/min, UV 254 nm) to yield 1 (tR 37.0 min, 10 mg). The
EtOAc fraction (14.0 g) was subjected to silica gel VLC and eluted
with n-hexane/EtOAc/MeOH (2:1:0.2) and CHCl3/MeOH (8:1,
5:1, 3:1, 0:1) to produce five fractions (EA-1−EA-5). Fraction EA-4
(2.6 g) was divided into 11 subfractions (EA-4-1−EA-4-11) using
MPLC (column size 120g) with a gradient of MeOH/H2O (10:90 →
E
DOI: 10.1021/acs.jnatprod.8b00441
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Compound 2: yellowish, amorphous powder; [α]2D2 −175 (c 0.3,
MeOH); UV (MeOH) λmax (log ε) 219 (4.59), 247 (4.73), 358
(4.17), 376 (4.29), 397 (4.24) nm; IR (neat) νmax 3349, 1058, 1018
cm−1; 1H and 13C NMR, see Table 1; HRESIMS m/z 571.1787 [M +
Na]+ (calcd for C27H32O12Na, 571.1791).
Parameters for EXSIDE and PANIC.20,21 The key EXSIDE
NMR experiment was carried out using the Bruker pulse
hsqcetgplrjcsp with t1 increments (indirect dimension) of 900 and
a scaling factor of 15, to ensure sufficient resolution to observe
splitting of correlations along the f1 dimension. According to the
original author’s recommendation,16 a selective pulse width (i.e., the
HO-9 resonance in this study) for the direct dimension (f2) should be
carefully chosen as wide as possible to shorten the relaxation delay for
a generated shaped pulse. To increase the digital resolution, the f1
spectrum width was reduced to ∼100 ppm. The 1D NOESY for the
PANIC analysis was implemented using a double-pulse field gradient
spin−echo NOE (DPFGSENOE) excitation sculpted selective
sequence incorporated with a zero-quantum filter element (500 ms
mixing time, 2 s relaxation time, 64 scans). The equation below was
utilized to acquire accurate interproton distances (“runknown”). The
resonance for H-8 in compound 1 was selectively irradiated using 1D
NOESY, and the integration for the irradiated resonance was
arbitrarily normalized to −1000. The generated NOE intensity for
H-7 was integrated with reference to −1000 (Figure S16, Supporting
Information), and the integration of H-7 was used as “NOEreference”.
The typical interproton distance between the aromatic protons (i.e.,
H-7 to H-8) of 2.5 Å was used as “rreference”. To calibrate interproton
distances from H-10 to H-4a and from H-4α to H-4β (“runknown”), H-
10 was irradiated using the aforementioned pulse. Its integration was
normalized to −1000, and the resultant NOE integrations for H-4a
and the two diastereotopic protons (Figure S17, Supporting
Information) were used to obtain respective “runknown” values with
the following equation.
Figure 4. HMBC via the intramolecular hydrogen bonding from HO-
9 and quantitative NOE-distance measurement for relative configura-
tional analysis of 1. NOE-derived interproton distances are denoted
with their respective predicted distances shown in parentheses: x =
2.9 Å (3.1 Å); y = 2.2 Å (2.3 Å); z = 2.7 Å (2.7 Å). The 3D structure
was optimized at the B3LYP/6-31+G(d,p) level in the PCM mode
(CH3CN).
Figure 5. Overlay of experimental and simulated ECD spectra of
compound 1. The simulation was carried out at the B3LYP/6-
31+G(d,p) level in the PCM mode (CH3CN).
6
NOEunknown/NOEreference = (rreference)6/(runknown
)
100:0, 5 L). Compound 2 (20 mg) was obtained from fraction EA-4-3
(51 mg) employing HPLC, eluting with MeOH/H2O (40:60, 2.5
mL/min, UV 254 nm, tR 40.6 min).
Computational Details. All conformers of 1 were identified using
the Macromodel (version 2015-2, Schrodinger LLC) module with
“mixed torsional/low mode sampling” in the MMFF94 force field.
The search process was initially performed in the gas phase with a 50
kJ/mol energy window limit and 10 000 maximum number of steps to
identify all potential conformers. The Polak−Ribiere conjugate
gradient protocol was utilized to minimize conformers with 10 000
maximum iterations and a 0.001 kJ (mol Å)−1 convergence threshold
on the rms gradient. Two conformers were found within 10 kJ/mol of
each global minimum and subjected to geometry optimization using
the Gaussian 09 package (Gaussian Inc.) at the B3LYP/6-31+G(d,p)
level with the polarizable continuum model (PCM) mode with a
dielectric constant representing CH3CN. The optimized coordinates
were utilized for ECD simulation at the identical functional basis set,
and the Boltzmann-averaged spectrum was visualized using SpecDis.24
The interproton distance analysis for the establishment of the relative
configuration of 1 was performed using the most abundant
conformers of each plausible diastereomer (i.e., 4aR*,10R* and
4aS*,10R*) at the aforementioned basis set. In the epimer
(4aS*,10R), the predicted distances from H-10 to H-4a, H-4α, and
H-4β were 3.0, 3.0, and 2.4 Å.
Enzymatic Hydrolysis of 2 and Establishment of the
Absolute Configuration of Sugar Moieties. Compound 2 (2
mg) was treated with β-glucosidase from almonds (5 mg; Sigma-
Aldrich, St. Louis, MO, USA) in 1 mL of distilled H2O at 37 °C for 20
h. The mixtures were passed through filters (PTFE 0.45 μM) to
obtain the aqueous solutions containing the resultant monosacchar-
ides. The solutions were dried and subjected to chiral derivatization.23
The sugar-containing residue was dissolved in pyridine (0.5 mL)
containing L-cysteine methyl ester hydrochloride (3 mg) and heated
at 60 °C for 1 h. Phenyl isothiocyanate (50 μL) was added, and the
mixtures were heated at 60 °C for another 1 h. The reaction mixture
was dried under a stream of N2 gas to remove residual pyridine. The
residue was dissolved in 1 mL of HPLC-grade MeOH and analyzed
employing an LC-10AD series HPLC system (Shimadzu, Kyoto,
Japan) equipped with a Phenomenex column (250 × 4.6 mm, 5 μm),
mobile phase CH3CN/H2O (25:75) with a flow rate of 0.6 mL/min.
The sample solution (10 μL) was injected onto the system, and the
derivatized sugar moiety was detected at the UV wavelength 250 nm.
Authentic D- and L-glucopyranose (each 1 mg) were derivatized and
analyzed using the identical protocol. The D-configuration of the sugar
motif was established based upon the comparison of the retention
time of the sugar derivative of 2 (tR 15.398 min) with those of the
standard D-glucose (tR 15.391 min) and L-glucose (tR 14.596 min)
derivatives (Figure S26, Supporting Information).
sEH Assay. The assay was conducted with reference to a
published protocol.25 A 50 μL amount of sEH (140 ng/mL) and
20 μL of different concentrations of compounds 1 and 2 in MeOH
were mixed in a 96-white-well plate containing 80 μL of 25 mM bis-
Tris-HCl buffer (pH 7.0) and 0.1% bovine serum albumin. The
mixture was incubated at 37 °C, and the products of hydrolysis were
monitored at excitation and emission wavelengths of 330 and 465 nm
after 1 h.
Compound 1: yellow, viscous oil; [α]2D2 +236 (c 0.2, MeOH); UV
(MeOH) λmax (log ε) 246 (4.08), 372 (4.12) nm; IR (neat) νmax
1
3382, 1668, 1615, 1595, 1557, 1284, 1065, 1047, 758 cm−1; H and
13C NMR, see Table 1; HRESIMS m/z 265.0837 [M + Na]+, m/z
243.1022 [M + H]+ (calcd for C15H14O3Na, 265.0841, and C15H15O3,
243.1021).
F
DOI: 10.1021/acs.jnatprod.8b00441
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
(14) Https://clinicaltrials.gov/ct2/show/NCT00847899?term=
NCT00847899.&rank=1, accessed on July 15, 2015.
(15) Imig, J. D.; Hammock, B. D. Nat. Rev. Drug Discovery 2009, 8,
794−805.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jnat-
prod.8b00441.
■
S
(16) Krishnamurthy, V. J. Magn. Reson., Ser. A 1996, 121, 33−41.
́
(17) Chang, L. C.; Chavez, D.; Gills, J. J.; Fong, H. H. S.; Pezzuto, J.
M.; Kinghorn, A. D. Tetrahedron Lett. 2000, 41, 7157−7162.
1D and 2D NMR and HRESIMS spectra of 1 and 2, VT
experiments of 1, and computational details (PDF)
̌
́
̌
́
̌
́
́
(18) Seckarova, P. N.; Marek, R.; Malinakova, K.; Kolehmainen, E.;
́
́
̌
Hockova, D.; Hocek, M.; Sklenar, V. Tetrahedron Lett. 2004, 45,
6259−6263.
AUTHOR INFORMATION
Corresponding Author
*Tel: +82 42 821 5925. Fax: +82 42 823 6566. E-mail: mkna@
cnu.ac.kr (M. Na).
ORCID
Chung Sub Kim: 0000-0001-9961-4093
Daneel Ferreira: 0000-0002-9375-7920
Young Ho Kim: 0000-0002-5212-7543
MinKyun Na: 0000-0002-4865-6506
(19) Tokunaga, T.; Seki, H.; Yasuike, S.; Ikoma, M.; Kurita, J.;
Yamaguchi, K. Tetrahedron Lett. 2000, 41, 1031−1034.
(20) Butts, C. P.; Jones, C. R.; Song, Z.; Simpson, T. J. Chem.
Commun. 2012, 48, 9023−9025.
■
(21) Hu, H.; Krishnamurthy, K. J. Magn. Reson. 2006, 182, 173−177.
(22) Hoye, T. R.; Jeffrey, C. S.; Shao, F. Nat. Protoc. 2007, 2, 2451−
2458.
(23) Tanaka, T.; Nakashima, T.; Ueda, T.; Tomii, K.; Kouno, I.
Chem. Pharm. Bull. 2007, 55, 899−901.
(24) Bruhn, T.; Hemberger, Y.; Schaumloffel, A.; Bringmann, G.
SpecDis, Version 1.51; University of Wuerzburg: Germany, 2011.
(25) Jones, P. D.; Wolf, N. M.; Morisseau, C.; Whetstone, P.; Hock,
B.; Hammock, B. D. Anal. Biochem. 2005, 343, 66−75.
(26) Gottlieb, H. E.; Kotlyar, V.; Nudelman, A. J. Org. Chem. 1997,
62, 7512−7515.
Author Contributions
#J. Oh and K. T. Quan contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This study was supported by the Basic Science Research
Program (NRF-2017R1A2A2A05001340) through the Na-
tional Research Foundation of Korea (NRF) grant funded by
the Korean Government. This work was also supported by the
Korea Polar Research Institute (grant no. PE18210). We thank
Drs. E. Kwan (Harvard University), C. P. Butts (University of
Bristol), K. Krishnamurthy (Chempacker LLC), and Agilent
Spinsights Forum for practical advice regarding NMR
techniques used in this study. We also recognize the Korea
Institute of Science and Technology Information (KISTI) for
supercomputing access.
REFERENCES
■
(1) Jeffrey, G. A.; Saenger, W. Hydrogen Bonding in Biological
Structures; Springer-Verlag: Berlin, 1991.
(2) Yim, G.; Thaker, M. N.; Koteva, K.; Wright, G. J. Antibiot. 2014,
67, 31−41.
(3) Butler, M. S.; Hansford, K. A.; Blaskovich, M. A.; Halai, R.;
Cooper, M. A. J. Antibiot. 2014, 67, 631−644.
(4) Kuhn, B.; Mohr, P.; Stahl, M. J. Med. Chem. 2010, 53, 2601−
2611.
(5) Huang, X. S.; Liu, X.; Constantine, K. L.; Leet, J. E.; Roongta, V.
Magn. Reson. Chem. 2007, 45, 447−450.
(6) Exarchou, V.; Troganis, A.; Gerothanassis, I. P.; Tsimidou, M.;
Boskou, D. Tetrahedron 2002, 58, 7423−7429.
(7) Borisov, E.; Zhang, W.; Bolvig, S.; Hansen, P. E. Magn. Reson.
Chem. 1998, 36, S104−S110.
(8) Morisseau, C.; Hammock, B. D. Annu. Rev. Pharmacol. Toxicol.
2013, 53, 37−58.
(9) Kitamura, S.; Morisseau, C.; Inceoglu, B.; Kamita, S. G.; De
Nicola, G. R.; Nyegue, M.; Hammock, B. D. PLoS One 2015, 10,
e0117438.
(10) Lin, W.-K.; Falck, J.; Wong, P. Y. Biochem. Biophys. Res.
Commun. 1990, 167, 977−981.
(11) Inceoglu, B.; Jinks, S. L.; Schmelzer, K. R.; Waite, T.; Kim, I.
H.; Hammock, B. D. Life Sci. 2006, 79, 2311−2319.
(12) Rose, T. E.; Morisseau, C.; Liu, J.-Y.; Inceoglu, B.; Jones, P. D.;
Sanborn, J. R.; Hammock, B. D. J. Med. Chem. 2010, 53, 7067−7075.
(13) Imig, J. D.; Zhao, X.; Capdevila, J. H.; Morisseau, C.;
Hammock, B. D. Hypertension 2002, 39, 690−694.
G
DOI: 10.1021/acs.jnatprod.8b00441
J. Nat. Prod. XXXX, XXX, XXX−XXX