Bilobalide and PC12 cells: A structure activity relationship study

Article abstract of DOI:10.1016/j.bmc.2019.115251

Ginkgo biloba extracts have been postulated to beneficial for improving cognitive function and as such they have been used as a potential treatment of Alzheimer's disease. The main active ingredients of the extract are terpene trilactones (TTLs), such as bilobalide (BB) and ginkgolides. Several structure–activity relationship (SAR) studies using ginkgolide scaffolds produced more biologically potent species by modification of the lactone moieties. However, modifications of BB scaffold have been limited, and no SAR studies on BB have been accomplished to date. Thus, the aim of this study was to elucidate how the modification of the lactone moieties of BB would affect their biological activities in a number of assays, including proliferating cell activity, neuroprotective effects against Aβ (1–40) peptides, and neurite outgrowth effects in PC12 neuronal cells. It appeared that the derivatives containing lactone groups showed similar biological activity to native BB, while those that possessed no lactone moieties exhibited lower neurite outgrowth effects. Thus, the results suggested that the lactone moieties of BB played an important role in exerting neurite outgrowth effects in PC12 cells.

Full text of DOI:10.1016/j.bmc.2019.115251

Bioorganic&MedicinalChemistryxxx(xxxx)xxxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Bilobalide and PC12 cells: A structure activity relationship study
⁎
Toyonobu Usukia, , Yukiko Yoshimotoa, Makiko Satoa, Tae Takenakaa, Ryota Takezawaa,
Yusuke Yoshidab, Masayuki Satakec, Noriyuki Suzukia, Daisuke Hashizumed, Sergei V. Dzyubae
a Department of Materials and Life Sciences, Faculty of Science and Technology, Sophia University, 7-1 Kioicho, Chiyoda-ku, Tokyo 102-8554, Japan
b Adjusted Cell Experiment Laboratory Inc, 5-4-30 Nishi-Hashimoto, Midori-ku, Sagamihara-shi, Kanagawa 252-0131, Japan
c Department of Chemistry, School of Science, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
d Materials Characterization Support Unit, RIKEN Center for Emergent Matter Science (CEMS), 2-1 Hirosawa, Wako, Saitama 351-0198, Japan
e Department of Chemistry and Biochemistry, Texas Christian University, Fort Worth, TX 76129, USA
A R T I C L E I N F O
A B S T R A C T
In memory of Koji Nakanishi.
Ginkgo biloba extracts have been postulated to beneficial for improving cognitive function and as such they have
been used as a potential treatment of Alzheimer’s disease. The main active ingredients of the extract are terpene
trilactones (TTLs), such as bilobalide (BB) and ginkgolides. Several structure–activity relationship (SAR) studies
using ginkgolide scaffolds produced more biologically potent species by modification of the lactone moieties.
However, modifications of BB scaffold have been limited, and no SAR studies on BB have been accomplished to
date. Thus, the aim of this study was to elucidate how the modification of the lactone moieties of BB would affect
their biological activities in a number of assays, including proliferating cell activity, neuroprotective effects
against Aβ (1–40) peptides, and neurite outgrowth effects in PC12 neuronal cells. It appeared that the derivatives
containing lactone groups showed similar biological activity to native BB, while those that possessed no lactone
moieties exhibited lower neurite outgrowth effects. Thus, the results suggested that the lactone moieties of BB
played an important role in exerting neurite outgrowth effects in PC12 cells.
Keywords:
Bilobalide
Terpene trilactone
Derivatization
PC12
SAR
1. Introduction
(SC)-CA1 synapses, and the facilitative effects on synaptic plasticity
were only observed at medial performant path (MPP)-dentate gyrus
The Ginkgo tree (Ginkgo biloba) is one of the oldest living species,
which is often called a “living fossil”. The extracts of G. biloba leaves are
one of the oldest documented traditional Chinese medicines, which
have been used as effective dietary supplements and phytomedicines.1,2
The extracts, such as EGb 761, have been known to have numerous
pharmacological effects including antioxidant activity,3 decreasing the
risk of peripheral artery disease,4 as well as they exhibited beneficial
effects in the treatment of the Alzheimer’s disease.5
(DG) synapses.10 BB was also suggested to exhibit anticonvulsant po-
tential,11 as well as it was found inhibit Pneumocystis carinii growth.12
However, SAR studies of BB have not been accomplished to the best of
our knowledge.
Lactone-containing natural product jiadifenin and its derivatives
were shown to enhance the neurite outgrowth effects on PC12 neuronal
cells induced by nerve growth factor (NGF).13 Therefore, it was of in-
terest to elucidate the role of BB’s lactone moieties on neurite out-
growth effects using PC12 neuronal cell lines. In addition, optimization
of the extraction and isolation procedures of TTLs, including BB, by
efficient removal of ginkgolic acids utilizing activated charcoals was
carried out. In order to elucidate the role of lactone moieties of BB,
several analogs were prepared, including non-lactone containing deri-
vatives. Next, the ability of BB’s derivatives to affect cell proliferation
and neurite outgrowth using PC12 neuronal cells was examined. Neu-
roprotective effects of BB and its derivatives against amyloid beta (Aβ1-
40) peptides, which are believed to be responsible for the occurrence
and progression of Alzheimer’s disease, were also evaluated.
To date, a number of natural products from G. biloba have been
isolated, including flavonoids, terpene trilactones (TTLs, Fig. 1) and
ginkgolic acids. TTLs, i.e., bilobalide (BB)6 and ginkgolide A, B, C, and J
(GA, GB, GC, and GJ, respectively),7 are the main components of the
extract that could affect cognitive functions, were isolated and char-
acterized by Nakanishi and co-workers. Structurally, TTLs have a cage-
like skeleton, three lactone groups, and a fairly unusual for terrestrial
natural products tert-butyl (t-Bu) group. Several structure–activity re-
lationship (SAR) studies on ginkgolides were conducted using platelet-
activating factor (PAF)8 and glycine9 receptors. In addition, it was
found that BB had some input–output relationship at Schaffer collateral
⁎ Corresponding author.
E-mail address: t-usuki@sophia.ac.jp (T. Usuki).
https://doi.org/10.1016/j.bmc.2019.115251
Received 30 October 2019; Received in revised form 2 December 2019; Accepted 4 December 2019
0968-0896/©2019ElsevierLtd.Allrightsreserved.
Pleasecitethisarticleas:ToyonobuUsuki,etal.,Bioorganic&MedicinalChemistry,https://doi.org/10.1016/j.bmc.2019.115251

T. Usuki, et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxxx
O
Table 1
HO
10
11
The amount of BB, ginkgolides, and ginkgolic acid extracted from 250 g of G.
1
biloba leavesa upon repeated treatment with charcoal.
O
17
18
12
5
2
O 4
O
19
20
O
O
7
Treatment times
Amount (mg)a
BB
OH
H
Ginkgolic acid
Ginkgolides
bilobalide
1
2
3
957
36
8
170b
158
91
150
140
77
O
HO
R1
H
O
O
a
b
Average value.
O
O
Repeated twice.
R2
Me
H
O
O
(=170 + 150) mg of TTLs. While the second washing in methanol
(MeOH) removed almost 96% of ginkgolic acid (i.e., only 36 mg re-
mained), while the amount of TTLs was relatively unaffected, i.e., less
than 10% decrease of the TTLs mass (298 (=158 + 140) mg).
Although, after the third treatment with charcoal in MeOH, the of
ginkgolic acid was reduced even further (8 mg or less than 1% of the
original amount), the amount of TTLs decreased to as well to a com-
bined 168 (=91 + 77, i.e., about 50% of the original amount) mg.
Thus, treatment with charcoal first in EtOAc and then in MeOH ap-
peared to be “optimal” in removing ginkgolic acid.
ginkgolide
ginkgolide
R1 R2
ginkgolide A (GA)
H
H
H
ginkgolide B (GB) OH
ginkgolide C (GC) OH OH
ginkgolide J (GJ)
H
H
Fig. 1. Structures of TTLs in G. biloba.
Extracts obtained by activated charcoal extraction method were
subjected to neutral silica gel column chromatography with hexane/
EtOAc as an elution solvent system (hexane/EtOAc = 5:1 → 1:1) to
give 126 mg of BB (i.e., 0.05% of the originally used 250 g of G. biloba
leaves). Ginkgolides were eluted together, and they were subsequently
separated by refractive-index reversed-phase high-performance liquid
chromatography (RI-RP-HPLC) instrument as shown in Fig. 3. Each
peak was collected and identified by mass spectrometry (MS) and 1H
NMR to obtain pure GA, GB, GC, and GJ contaminated with GC. The
isolated amount of ginkgolides from 250 g of G. biloba leaves was GC/
GJ = 3.36 mg, GC = 1.35 mg, GA = 6.73 mg, and GB = 0.67 mg,
respectively.
2. Results and discussion
2.1. Extraction and isolation of TTLs
TTLs from Ginkgo biloba leaves were isolated using modified lit-
erature protocol.8 In general, refluxing green leaves of G. biloba in ethyl
acetate (EtOAc) led to extraction of not only TTLs but also to a large
amount of ginkgolic acid (a compound that causes allergy, mutageni-
city, cytotoxicity, and neurotoxicity). Thus, in this study, the extract
was treated with activated charcoal several times to rid of the ginkgolic
acid (Fig. 2). Quantitative 1H nuclear magnetic resonance (NMR) ana-
lysis of the TTLs in the extracts was performed using the isolated singlet
peak of dimethylformamide (DMF) as an internal standard (Table 1).14
The amount of TTLs was determined through quantitative 1H NMR
using DMF as an internal standard and this manipulation was repeated
three times.
2.2. Derivatization of BB
With BB in hand, the chemical derivatization including modification
of lactone moieties was carried out. It was previously shown that dii-
sobutylaluminium hydride (DIBAL-H) reduction of GA afforded lactone-
free ginkgolides.15 Thus, we attempted to prepare lactone-free BB by
The first treatment with activated charcoal in EtOAc left behind
some ginkgolic acid (957 mg) along with
a combined 320
Fig. 2. Partial 1H NMR (300 MHz) spectra of the Ginkgo biloba extract in methanol-d4. (a) After initial treatment with activated charcoal; (b) After second treatment
by activated charcoal. Peaks at around 6 ppm are attributed to H12 of ginkgolides. Peak at 6.3 ppm derived from H12 proton of BB. Peaks at 6.5 and 7.1 ppm are
attributed to aromatic protons of ginkgolic acid. Singlet at 8 ppm is attributed to the aldehyde proton of DMF, which was used as an internal standard.
2

T. Usuki, et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxxx
Fig. 3. RI-RP-HPLC chromatogram; mixture of GJ and GC (Rt = 22.7 min); GC (Rt = 25.5 min); GB (Rt = 49.7 min); GA (Rt = 56.3 min). Conditions: column –
Cosmosil 5C-18-AR-II, 10 × 250 mm; solvent: MeOH/H2O = 3/7; flow rate: 2.0 mL min−1; detection: refractive index; temperature: 37 °C. Rt – retention time.
O
OH
O
HO
Et3SiH (15 eq)
BF3·Et2O (7.5 eq)
HO
HO
HO
O
DIBAL-H (25 eq)
O
1
O
2
HO
O
HO
HO
CH2Cl2
-78 °C to rt, 19 h
23%
THF, -78 °C
2 h, 38%
O
O
O
O
OH
O
OH
OH
H
H
H
1
bilobalide (BB)
BB-diether (2)
Scheme 1. Preparation of BB-diether 2.
O
treatment with 25 equivalents of DIBAL-H in tetrahydrofuran (THF) at
−78 °C for 2 h. A complex mixture of compounds was obtained as
judged by thin-layer chromatography (TLC) analysis, and BB-trilactol
(1) in 38% yield along with 5% of recovered BB were isolated by
column chromatography (Scheme 1). Isolated 1 was then reacted with
triethylsilane (Et3SiH, 15 eq) and boron trifluoride diethyl etherate
(BF3·Et2O, 7.5 eq) in dichloromethane (CH2Cl2) to produce BB-diether
(2) in 23% yield, whose structure was determined by mass spectro-
metry (MS) and two-dimensional (2D) NMR spectroscopy. While cross
peak was observed between H1 and H2 by 1H-1H correlation spectro-
scopy (COSY) analysis, relative stereochemistry of C2 was not estab-
lished. Reduction of two acetals of C4 and C11 proceeded instead. Ether
bond of C12 was likely to be relatively unstable and easy to form the
spirocycles, and thus the ether bond was cleaved during silane reduc-
tion.
O
H
HO
Ac2O (10 eq)
iPr2EtN (10 eq)
AcO
O
O
OAc
O
O
O
O
O
CH2Cl2, rt
18 h, 32%
O
O
O
OH
H
H
BB
diAc-isoBB (3)
O
H
OAc
H
AcO
O
10
12
O
O
O
7
O
NOE
H
H
H
Next, acetylation of BB was carried out using acetic anhydride
(Ac2O, 10 eq) in the presence of diisopropylethylamine (i-Pr2NEt,
10 eq) in CH2Cl2 to give the diacyl (diAc) derivative diAc-isoBB (3) with
the spiro skeleton in 32% yield (Scheme 2),16 along with 12% of re-
covered BB. The structure of new compound 3 was confirmed using
NMR and MS measurements. Specifically, nuclear Overhauser effect
spectroscopy (NOESY) measurements exhibited the cross peaks be-
tween t-Bu and H7, 10, and 12.
3
Scheme 2. Preparation of diAc-isoBB 3 and the NOE correlations.
O
O
HO
p-BrBzCl (10 eq)
BrBzO
O
pyridine
O
OBzBr
In order to investigate the effect of bulky functional group such as
benzoate by means of biological activity, p-bromobenzoylation (p-BrBz)
of BB was carried out using p-BrBzCl (10 eq) in the presence of pyridine
(Scheme 3) at 60 °C for 26 h.16 Two products were obtained and se-
parated by silica gel column chromatography. Major product was iso-
lated in 63% yield and the structure was unambiguously established by
X-ray crystallography (Fig. 4) as diBrBz-isoBB (4).
60 °C, 26 h
63%
O
O
O
O
O
O
O
OH
O
H
H
BB
diBrBn-isoBB (4)
Scheme 3. Preparation of diBrBz-isoBB 4.
3

T. Usuki, et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxxx
described in Scheme 4.
The other chemical derivatizations of the two hydroxyl groups of BB
were also attempted. When BB was treated with methoxymethyl
chloride (MOMCl, 20 eq) in the presence of i-Pr2NEt in dichloroethane
(CH2Cl)2 under reflux condition for 24 h, a complex mixture was ob-
tained. In addition, treatment with pivaloyl chloride (PivCl, 10 eq) and
4-dimethylaminopyridine (DMAP, 0.5 eq) in pyridine proved to be
unsuccessful as well, as BB was recovered in 89% yield, most likely due
to the steric bulkiness of t-Bu group. Overall, chemical modifications of
BB afforded four derivatives 2 – 5 that were used for the SAR study.
2.3. Biological activity
With BB and the derivatives BB-diether 2, diAc-isoBB 3, diBrBz-
isoBB 4, and diBn-BB-aldehyde 5 in hand, three biological assays using
rat pheochromocytoma PC12 cell lines were utilized. In addition to BB
and the prepared derivatives, a commonly used G. biloba extract
BioGinkgo 27/7, which has been standardized to contain 7% TTLs and
27% ginkgo flavone glycosides, was used as a reference.
Cell culture conditions for PC12 were optimized at 1% concentra-
tion of serum and 2,000 cell per well in the presence of 10 ng/mL of
NGF. Cell viabilities were measured by the cell count reagent SF (WST-
8). BB, BB-diether 2, and diAc-isoBB 3 exhibited the proliferating cell
activity (Fig. 5). Significantly, compound 2, i.e., the lactone-free con-
gener, showed the highest activity. Neuroprotective effects against
1 µM of Aβ (1–40) peptides were also examined (Fig. 6). BB and its
derivatives, except 2, all showed neuroprotective effects against Aβ
(1–40). In contrast, BioGinkgo 27/7 did not show the appropriate ac-
tivity in both experiments, likely due to the low concentration of BB.
Next, studies on the effect of the standard extract, BB and BB’s de-
rivatives on neurite outgrowth of PC12 cells were conducted (Figs. 7
and 8). BioGinkgo 27/7 was found to show the highest neurite out-
growth effects. The result might probably be attributed to the sy-
nergistic effects of the complex mixture of all TTLs and/or flavonoids.
More detailed studies are ongoing and they will be reported elsewhere.
BB, 3, 4, and 5, which all have lactone moieties, also showed positive
effects. However, BB-diether 2, the compound that did not possess
lactone groups, did not exert the effect. These results strongly indicated
the importance of the BB’s lactone moieties on the neurite outgrowth
effects. It should be noted that in general, halted cell proliferation leads
to differentiation, such as a formation of neurite outgrowth in neuronal
Fig. 4. View of diBrBz-isoBB (4) by X-ray crystallography.
Ag2O (4 eq)
TBAI (0.1 eq)
BnBr (10 eq)
O
H
O
28
HO
25
O
29
30
16
27
O
18-24
BnO
1
3
O
O
DMF
rt, 17 h, 43%
O
O
4
5
7
OH
9-15
OBn
O
H
BB
diBn-BB-aldehyde (5)
- H2O
decarboxylation benzylation
O
O
O
HO
HO
OH
O
acetal cleavage
formation of
HO
O
O
O
O
cyclopentadiene
OH
H
O
Scheme 4. Synthesis of diBn-BB-aldehyde (5) with reaction mechanism.
In addition, benzylation of BB was conducted using silver oxide
(Ag2O, 4 eq), tetrabutylammonium iodide (TBAI, 0.1 eq), and benzyl
bromide (BnBr, 10 eq) in DMF at room temperature for 17 h (Scheme
4). Unexpectedly, dibenzylated compound diBn-BB-aldehyde (5) was
obtained in 43% yield, along with 21% of recovered BB.17,18 Structure
of 5 was determined by NMR and MS analysis. Plausible reaction
pathway might include the following steps: (1) dehydration at C8 leads
to formation of the double bond; (2) acetal cleavage and formation of
the cyclopentadiene moiety; (3) decarboxylation and benzylation, as
Fig. 5. Effects of substances (0.26 µg/mL) on cell proliferation. Relative values
of cell viability are calculated by comparison with control (0.004% DMSO).
4

T. Usuki, et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxxx
suggested that lactone moieties were critical for biological activities. By
comparison with BB, both 3 and 4 exhibited better neuroprotective
effects, arguably indicating the beneficial effects of the OH-substitu-
tions. diBn-BB-aldehyde 5 also exhibited better neurite outgrowth ef-
fects as compared to BB, however, due to the lack of BB’s scaffold, direct
correlations should be avoided.
3. Conclusion
We disclose a new extraction method of TTLs from G. biloba using
activated charcoal that removes substantial amounts of ginkgolic acid,
while retaining appreciable quantities of TTLs. Four derivatives, namely
BB-diether 2, diAc-isoBB 3, diBrBz-isoBB 4, and diBn-BB-aldehyde 5
were synthesized from BB isolated from the Ginkgo biloba leaves.
Bioassays of BB, its derivatives, and BioGinkgo 27/7 were conducted on
PC12 neuronal cells. BioGinkgo 27/7 showed strong neurite outgrowth
effects, while BB, 3, and 4 showed appreciable neuroprotective effects
against Aβ (1–40) and neurite outgrows effects. However, lactone-free
congener 2 did not show these effects. These results of the SAR study
strongly suggest that the lactone moieties of BB play an important role
in these effects for PC12 cells. In addition, functionalization of hy-
droxyl-groups appeared to play a role, although more detailed studies
are required to clarify these effects.
Fig. 6. Effects of substances (0.26 µg/mL) on neuroprotection against Aβ
toxicity. Relative values of cell viability are calculated by comparison with
control (0.004% DMSO).
4. Experimental section
cells; thus, the cell proliferation could be viewed as the contrasting
process to the cell differentiation. The relationships between the cell
viability and the neurite outgrowth were reasonable in this study. On
the other hand, the results of the neuroprotection showed linearity with
the results of the neurite outgrowth.
4.1. General
All reagents were obtained from commercial suppliers and used
without further purification unless otherwise stated. Anhydrous MeOH
was purchased from Kanto Chemicals (Tokyo, Japan). HPLC grade
solvents for HPLC runs were obtained from commercial suppliers.
BB-diether 2 showed neither neuroprotective effect against Aβ
(1–40) peptides nor neurite outgrowth effects. These results strongly
Fig. 7. Effects of substances (0.26 μg/mL) on neurite outgrowth.
5

T. Usuki, et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxxx
Fig. 8. Neurite outgrowth effects on PC12 cells. (A) control (0.004% DMSO); (B) 0.26 μg/mL BB; (C) 0.26 μg/mL 2; (D) 0.26 μg/mL 3; (E) 0.26 μg/mL 4; (F) 0.26 μg/
mL 5; (G) 0.26 μg/mL BioGinkgo 27/7.
CH2Cl2, i-Pr2NEt, and pyridine were dried by distillation and stored
over activated molecular sieves. DMF was distilled over MgSO4 and
stored over activated molecular sieves. Analytical TLC was performed
on Silica gel 60 F254 plates (Merck). Column chromatography was
performed with acidic Silica gel 60 (spherical, 40–50 μm) or neutral
Silica gel 60 N (spherical, 40–50 μm) (Kanto Chemicals). All non-aqu-
eous reactions were conducted under an atmosphere of nitrogen with
magnetic stirring using freshly distilled solvents unless otherwise
6

T. Usuki, et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxxx
indicated.
EtOAc = 1:5) afforded 1 as a colorless oil (5.6 mg, 0.017 mmol, 38%);
Rf 0.47 (hexane/EtOAc = 1:5); 1H NMR (300 MHz, CD3OD) δ 5.98 (1H,
s, H12), 5.60 (1H, s, H4), 5.35–5.31 (2H, m, H2, 11), 4.62 (1H, d,
J = 2.7, H10), 4.32 (1H, d, J = 5.4 Hz, H6), 2.67 (1H, d, J = 15.9 Hz,
H7), 2.20 (2H, s, H1), 1.98 (1H, dd, J = 15.9, 5.4 Hz, H7), 1.14 (9H, s,
tBu); ESI-MS (m/z) calcd for C15H24NaO8 [M+Na]+ 355.14, found
355.15.
Optical rotations were measured on a JASCO P-2200 digital po-
larimeter at the sodium lamp (λ = 589 nm) D line and are reported as
follows: [α]D (c g/100 mL, solvent). 1H and 13C NMR spectra were
T
recorded on a JEOL JNM-EXC 300 spectrometer (300 MHz) or on a
JEOL JNM-ECA 500 spectrometer (500 MHz). 1H NMR data are re-
ported as follows: chemical shift (δ, ppm), integration, multiplicity (s,
singlet; d, doublet; t, triplet; q, quartet; m, multiplet), coupling con-
stants (J) in Hz, assignments. 13C NMR data are reported in terms of
chemical shift (δ, ppm). FAB-MS spectra were obtained on a JEOL JMS-
700 instrument. ESI-MS spectra were recorded on a JEOL JMS-T100LC
instrument. X-ray single crystal analysis was obtained using a Rigaku
Saturn70 diffractometer using multi-layer mirror monochromated Mo-
Kα radiation (λ = 0.71073 Å) at 91 K. The HPLC analyses were per-
formed using a JASCO instrument equipped with a multiwavelength
detector (MD-2010), semi-micro HPLC pump (PU-2085), autosampler
(AS-2057), column thermostat (CO-2060), and refractive-index detector
(RI-2031).
Bilobalide-diether (2): To a solution of 1 (11.5 mg, 0.035 mmol,
1.0 eq) in CH2Cl2 (1.7 mL) was added Et3SiH (87.6 μL, 0.549 mmol,
15 eq) and BF3·Et2O (34.5 μL, 0.274 mmol, 7.5 eq) at −78 °C. After
stirring at room temperature (rt) for 19 h, the reaction mixture was
diluted with EtOAc and quenched with sat NaHCO3. The aqueous layer
was then extracted with EtOAc. The combined organic layers were
washed with brine, dried over Na2SO4, and concentrated in vacuo.
Purification using silica gel column chromatography (hexane/
EtOAc = 3:1) afforded 2 as a colorless oil (2.5 mg, 0.008 mmol, 23%);
Rf 0.56 (hexane/EtOAc = 1:1); 1H NMR (500 MHz, CD3OD) δ 5.46 (1H,
d, J = 3.5 Hz, H2), 4.43 (1H, dd, J = 6.3, 3.5 Hz, H10), 4.37 (1H, dd,
J = 9.8, 6.3 Hz, H11), 4.16 (1H, d, J = 9.5 Hz, H12), 4.12 (1H, dd,
J = 6.8, 1.8 Hz, H6), 3.75 (1H, dd, J = 10.0, 3.5 Hz, H11), 3.63 (1H, d,
J = 9.5 Hz, H12), 3.60 (2H, dd, J = 13.0, 7.5 Hz, H4), 2.51 (1H, dd,
J = 14.3, 1.8 Hz, H7), 2.16 (1H, dd, J = 14.3, 6.8 Hz, H7), 2.01 (1H, d,
J = 11.5 Hz, H1), 1.88 (1H, dd, J = 12.0, 3.5 Hz, H1), 1.05 (9H, s, t-
Bu); 13C NMR (125 MHz, CD3OD) δ 103.0 (C2), 96.0 (C8), 85.0 (C6),
76.9 (C10), 76.2 (C11), 68.5 (C12), 67.6 (C4), 60.6 (C9), 57.4 (C5),
41.1 (C7), 36.0 (17), 33.9 (C1), 27.1 (C18/19/20); ESI-MS (m/z) calcd
for C15H26O6 [M + H]+ 303.18, found 303.14.
4.2. Materials of Ginkgo biloba leaves and chemicals
Dried green, female leaves of G. biloba were collected at 20–30 m
height from two or three trees with permission inside the campus of
Sophia University, Tokyo in early October 2008. The collected leaves
were stored at −30 °C before use.
4.3. Extraction and isolation of BB and ginkgolides
10,12-Di-O-acetyl-isobilobalide (3): To a solution of bilobalide
(13.7 mg, 0.042 mmol, 1.0 eq) in CH2Cl2 (450 μL) was added i-Pr2NEt
(74.9 μL, 0.419 mmol, 10 eq) and Ac2O (40.9 μL, 0.419 mmol, 10 eq).
After stirring at rt for 18 h, the reaction mixture was diluted with EtOAc
and quenched with 1 N HCl. The aqueous layer was then extracted with
EtOAc. The combined organic layers were washed with brine, dried
over Na2SO4, and concentrated in vacuo. Purification using silica gel
column chromatography (hexane/EtOAc = 3:1) afforded 3 as white
crystals (5.1 mg, 0.012 mmol, 32%); Rf 0.33 (hexane/EtOAc = 2:1); 1H
NMR (300 MHz, CDCl3) δ 6.93 (1H, s, H12), 5.93 (1H, s, H10), 4.43
(1H, dd, J = 7.8, 4.5 Hz), 3.21 (2H, dd, J = 29.3, 18.2 Hz), 2.81–2.63
(2H, m, 1H), 2.19 (3H, s, Me), 2.18 (3H, s, Me), 1.20 (9H, s, t-Bu); 13C
NMR (75 MHz, CDCl3) δ 168.9, 168.1, 167.3 (C2, C4, C11, C21, C23),
100.5 (C8), 93.1 (C12), 78.3 (C6), 66.3 (C10), 63.2 (C9), 62.7 (C5),
35.6 (C17), 34.8 (C1), 32.1 (C7), 26.5 (C18/19/20), 20.6, 20.4 (C22/
24); ESI-MS (m/z) calcd for C15H24NaO8 [M+Na]+ 433.11, found
433.10.
250 g of G. biloba leaves were treated with liquid nitrogen and
crushed in a mortar to small pieces of around 0.5 × 0.5 cm in size.
Subsequently, the crushed leaves together with 0.8 L of EtOAc were
placed in a round bottom flask equipped with a condenser, and sub-
jected to reflux for 1 h. After cooling, the mixture was filtered through
Celite 545 using a glass fritted filter. The EtOAc solution was washed
three times with sat. Na2SO3 aq. (200 mL), followed by brine (200 mL).
The activated charcoal (25 g) was added to the dried EtOAc solution
and the resulting mixture was stirred for 1 h at room temperature.
Subsequently, the mixture was filtred through Celite 545 (1.5 cm
thickness on a glass fritted filter), and the volatiles were removed in
vacuo. The residue was dissolved in MeOH and activated charcoal
(15 g) was added, and the resulting mixture was stirred for 1 h at room
temperature. Subsequently, the mixture was filtrered through Celite
545 (1.5 cm thickness on a glass fritted filter), and the volatiles were
removed in vacuo to afford the crude TTL extract.
The crude TTL extract was dissolved in a minimum amount of
EtOAc and loaded on a neutral silica gel column, and eluted using
hexane/EtOAc mixtures. At hexane/EtOAc = 5/1 → 3/1 ginkgolic acid
was isolated; at hexane/EtOAc = 1/1, BB and ginkgolides mixtures
were isolated. Each fraction was identified using 1H NMR.
10,12-Di-O-(p-bromobenzoyl)-isobilobalide (4): To a solution of
bilobalide (13.0 mg, 0.040 mmol, 1.0 eq) and p-bromobenzoyl chloride
(87.4 mg, 0.398 mmol, 10 eq) was added pyridine. After stirring at
60 °C for 26 h, the reaction mixture was diluted with EtOAc and
quenched with sat NaHCO3. The aqueous layer was then washed with
EtOAc. The combined organic layers were washed with brine, dried
over Na2SO4, and concentrated in vacuo. Purification using silica gel
column chromatography (hexane/EtOAc = 5:1) afforded 4 as a color-
less oil (17.7 mg, 0.025 mmol, 63%); Rf 0.41 (hexane/EtOAc = 3:1); 1H
NMR (300 MHz, CD3OD) δ 8.01–7.85 (4H, m, Bz), 7.71–7.26 (4H, m,
Bz), 7.33 (1H, s, H12), 6.34 (1H, s, H10), 4.46 (1H, dd, J = 7.8, 3.9 Hz,
H6), 3.27 (1H, d, J = 18.9 Hz, H1), 3.28 (1H, d, J = 18.9 Hz, H1), 2.96
(1H, dd, J = 15.0, 3.6 Hz, H7), 2.63 (1H, d, J = 15.3, 8.4 Hz, H7), 1.32
(9H, s, t-Bu); ESI-MS (m/z) calcd for C29H24Br2O10 [M+Na]+ 712.96,
found 713.05.
4.4. Separation of ginkgolides by RI-HPLC
RI-RP-HPLC system [Cosmosil 5C18-AR-II, 10 × 250 mm; solvent,
MeOH/H2O = 3/7; flow rate, 2.0 mL/min; detection, refractive index;
temperature, 37 °C] was used to separate the mixture of ginkgolides.
Each fraction was identified using NMR and MS.
4.5. Derivatization of bilobalide
Bilobalide-trilactol (1): To a solution of bilobalide (14.4 mg,
0.044 mmol, 1.0 eq) in THF (1.5 mL) was added DIBAL-H (1.1 mL,
1.10 mmol, 25 eq). After stirring at −78 °C for 2 h, the reaction mixture
was diluted with EtOAc and quenched with 3 N HCl. The aqueous layer
was then extracted with EtOAc. The combined organic layers were
washed with brine, dried over Na2SO4, and concentrated in vacuo.
Purification using silica gel column chromatography (hexane/
diBn-bilobalide-aldehyde (5): To a solution of bilobalide
(29.1 mg, 0.089 mmol, 1.0 eq) and TBAI (3.3 mg, 0.009 mmol, 0.1 eq)
in DMF (1 mL) was added BnBr (0.106 mL, 0.892 mmol, 10 eq). After
stirring at rt for 20 min, Ag2O (82.7 mg, 0.357 mmol, 4.0 eq) was added
in a single portion and the reaction mixture was stirred at rt for 17 h.
The mixture was filtered through a glass fritted filter covered with
Celite 535 and the filter cake was washed with Et2O. The filtrate was
7

T. Usuki, et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxxx
concentrated in vacuo. Purification using silica gel column chromato-
graphy (hexane/EtOAc = 4:1) afforded 5 as a yellow oil; Rf 0.80
(hexane/EtOAc = 1:1); [α]D20 −6.1 (c 0.1, CHCl3); 1H NMR (300 MHz,
acetone‑d6) δ 10.24 (1H, s, H25), 7.37–7.21 (10H, m, Bn), 7.02 (1H, d,
J = 5.7 Hz, H4), 6.81 (1H, d, J = 5.7 Hz, H5), 5.03 (2H, d, J = 2.1 Hz,
Bn), 5.00 (2H, d, J = 3.8 Hz, Bn), 3.46 (1H, d, J = 14.9 Hz, H6), 2.67
(1H, d, J = 14.9 Hz, H6), 1.38 (9H, s, t-Bu); 13C NMR (125 MHz,
acetone‑d6) δ 186.3 (C25), 171.7 (C7), 170.6 (C16), 170.0 (C3), 145.3
(C4), 137.2 (C5), 129.3, 129.1, 128.9, 128.8, 128.7, 128.2 (C11/12/
13/19/21/22/23), 67.2 (C18), 66.6 (C9), 64.5 (C1), 38.1 (C6), 36.1
Tesque, Kyoto, Japan), modified method in MTT assay. Briefly, the
culture supernates were replaced by a conditioned medium including
10% WST-8 reagent and incubated at 37 °C in 5% CO2/95% air for
1–2 h. The absorbance was measured at 450 nm by microplate reader.
The cell viabilities were indicated as relative values for vehicle (0.004%
DMSO). In neurite outgrowth assay, the cell images were observed by a
microscope before cell viability assay. The numbers of cells showing
neurite outgrowth of the same, twice and more than twice the length of
a cell, were determined on the images. These rates of neurite outgrowth
were indicated as %.
(C27), 31.7 (C28/31/32); FAB-HRMS (m/z) calcd for
C27H29O5
[M + H]+ 433.2015, found 433.1986.
Declaration of Competing Interest
4.6. X-ray diffraction analysis of 4
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influ-
ence the work reported in this paper.
Single crystals suitable for X-ray analysis were obtained from a
methanol solution by slow evaporation of the solvent. A colorless
crystal (0.090 × 0.060 × 0.040 mm) was mounted on a glass fiber. All
measurements were made on a Rigaku Saturn70 diffractometer using
multi-layer mirror monochromated Mo-Kα radiation (λ = 0.71073 Å)
at 91 K. The structure was solved by direct methods and expanded using
Fourier techniques. The non-hydrogen atoms were refined aniso-
tropically. Hydrogen atoms were refined using the riding model. The
asymmetric units consist of three crystallographically independent
molecules of 4, which have almost the same structure. The final cycle of
full-matrix least-squares refinement on F2 was based on 19,253 ob-
served reflections and 1117 variable parameters. All calculations were
performed using the CrystalStructure crystallographic software package
except for refinement, which was performed using SHELXL-97.
Crystallographic data are summarized in the supplementary data. CIF
data were deposited in Cambridge Structural Database (CCDC-
1937209).
Acknowledgments
This paper is dedicated to Professor Koji Nakanishi (Columbia
University), who passed away on March 28, 2019. T.U. and S.V.D. are
grateful for having an opportunity of knowing and working with
Professor Koji Nakanishi, who introduced us to the unique chemistry
and biology of Gingko biloba. We thank Dr. Yali Yong (Columbia
University) and Ms. Sasha Roudenko MD (Columbia University) for the
preliminary experiments. We also thank Ms. Manami Kurita (Sophia
University), Ms. Akiho Imura (Sophia University), and Ms. Aisya
Syahmina (Sophia University) for helpful suggestions about the
manuscript.
Appendix A. Supplementary material
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.bmc.2019.115251.
4.7. Cell culture
Rat pheochromocytoma PC12 cell line was obtained from the RIKEN
BioResource Center (RIKEN, Ibaraki, Japan). The undifferentiated cells
were grown in a culture medium, Dulbecco’s modified Eagle’s medium
(DMEM) supplemented with 10% fetal bovine serum (NICHIREI BIOS-
CIENCE INC., Tokyo, Japan) and 10% horse serum (Invitrogen, N.Y.,
USA) at 37 °C in 5% CO2/95% air.
References
1. Strømgaard K, Nakanishi K. Angew. Chem. Int. Ed. 2004;43:1641–1655.
2. van Beek TA. Ginkgo biloba. Netherland: Harwood Academic Publishers; 2000.
3. Bridi R, Crossetti FP, Steffen VM, Henriques AT. Phytother. Res. 2001;15:449–451.
4. Gardner CD, Taylor-Piliae RE, Kiazand A, Nicholus J, Rigby AJ, Farquhar JW. J.
Cardiopulm. Rehabil. Prev. 2008;28:258–265.
5. Kanowski S, Hermann WM, Stephan K, Wierich W, Hörr R. Pharmacopsychiatry.
1996;29:47–56.
4.8. Cell proliferation, neuroprotective and neurite outgrowth effects
6. Nakanishi K, Habaguchi K, Nakadaira Y, et al. J. Am. Chem. Soc. 1971;93:3544–3546.
7. Nakanishi K. Pure Appl. Chem. 1967;14:89–113.
8. Jaracz S, Strømgaard K, Nakanishi K. J. Org. Chem. 2002;67:4623–4626.
9. Jaracz S, Nakanishi K, Jensen AA, Strømgaard K. Chem. Eur. J. 2004;10:1507–1518.
10. Suzuki E, Sato M, Takezawa R, Usuki T, Okada T. J. Physiol. Sci. 2011;61:421–427.
11. Sasaki K, Wada K, Hatta S, Ohshika H, Haga M. Res. Commun. Mol. Pathol. Pharmacol.
1997;96:45–46.
PC12 cells (2,000 cells/well) were plated on type I collagen coated
cell culture plate (96 wells, 6 mm diameter) with the culture medium
and incubated at 37 °C in 5% CO2/95% air for 24 h. The culture
medium were substituted for the conditioned medium [DMEM sup-
plemented with 1% fetal bovine serum (FBS), 1% horse serum and
10 ng/mL mNGF 2.5S (Almone labs, Jerusalem, Israel)] including
substances (0.26 µg/mL) and cultured at 37 °C in 5% CO2/95% air for
24 h. The culture medium were substituted for the conditioned medium
including with 1 µM Aβ (1–40) peptide (Sigma-Aldrich Inc, St. Louis,
USA) (neuroprotection assay) and without Aβ (cell proliferation and
neurite outgrowth assay). After 24 h of culture, the cell viability of the
cultures was determined by colorimetric assay with WST-8 (Nacalai
12. Atzori C, Bruno A, Chichino G, Bombardelli E, Scaglia M, Ghione M. Antimicrob.
Agents Chemother. 1993;37:1492–1496.
13. Cho YS, Carcache DA, Tian Y, Li Y-M, Danishefsky SJ. J. Am. Chem. Soc.
2004;126:14358–14359.
14. Goto H, Usuki T. Phytochem. Anal. 2011;23:84–87.
15. Ishii H, Dzyuba SV, Nakanishi K. Org. Biomol. Chem. 2005;3:3471–3472.
16. Weinges K, Hepp M, Huber-Platz U, Irngartinger H. Liebigs Ann. Chem.
1987:1079–1085.
17. Sakaitani M, Ohfune Y. J. Org. Chem. 1990;55:870–876.
18. Brummond KM, Hong S. J. Org. Chem. 2005;70:907–916.
8