Total Synthesis of the Proposed Structure of Ardimerin, and Proposal for its Structural Revision

Article abstract of DOI:10.1002/hlca.201600244

We report the first total synthesis of the proposed structure of ardimerin, which was achieved in 14 steps starting from 2,3,4-trimethoxybenzoic acid. The key steps include the β-selective formation of the crucial C-glycoside linkage and stepwise construction of the strained eight-membered salicylide core. The synthesis revealed that the proposed structure 1 does not match the natural product. A proposal is made for reassigning the isolated natural product to the already known structure of bergenin. Interesting properties of the synthetic eight-membered salicylides are documented, including their susceptibility toward nucleophilic ring opening and the bowl chirality.

Full text of DOI:10.1002/hlca.201600244

Received Date: 01-Aug-2016
Revised Date: 08-Sep-2016
Accepted Date: 08-Sep-2016
Article Type: Full Paper
Total Synthesis of the Proposed Structure of Ardimerin, and
Proposal for its Structural Revision
Ryota Nakayama, Eva-Maria Tanzer, Takenori Kusumi, Ken Ohmori*, and Keisuke
Suzuki*
Department of Chemistry, Tokyo Institute of Technology, 2-12-1 O-okayama, Meguro-
ku, Tokyo 152-8551, Japan (e-mail: kohmori@chem.titech.ac.jp,
ksuzuki@chem.titech.ac.jp)
Dedicated to Prof. Dr. Dieter Seebach on the occasion of his 79th birthday
We report the first total synthesis of the proposed structure of ardimerin, which was
achieved in 14 steps starting from 2,3,4-trimethoxybenzoic acid. The key steps include
the β-selective formation of the crucial C-glycoside linkage and stepwise construction of
the strained 8-membered salicylide core. The synthesis revealed that the proposed
structure 1 does not match the natural product. A proposal is made for reassigning the
isolated natural product to the already known structure of bergenin. Interesting
properties of the synthetic 8-membered salicylides are documented, including their
susceptibility toward nucleophilic ring opening and the bowl chirality.
Keywords: Ardimerin • Structure Revision • Bergenin • C-Glycosides • Lactones
This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process, which may lead to
differences between this version and the Version of Record. Please cite this article as doi:
10.1002/hlca.201600244
This article is protected by copyright. All rights reserved.

Introduction
The subtropical plant Ardisia japonica is commonly used in traditional Chinese
medicine as an antitussive or diuretic agent as well as for stopping uterine bleeding.[1]
Phytochemical tests of Ardisia have led to the discovery of numerous constituents with a
variety of biological acitvities.[2] In 2002, Jung et al. isolated a compound from the plant
with a potent DPPH radical scavenging activity, and named it “ardimerin“. The
structure was proposed to be 1, an unusual cyclic dimer of a C-glycosyl salicylic acid
derivative (Figure 1a). Moreover, the corresponding digallate 2 was reported in 2007, as
an inhibitor against HIV-1 and HIV-2 RNase H.[4]
Considerable attention has been paid to these compounds,[1][2c][5] not only because they
hold great potential as therapeutic leads, but also due to their unusual structural
features, including the presence of a rare 8-membered salicylide[6] that appeared to be
unstable. While the s-trans ester linkage is stereoelectronically preferred by the n–σ
conjugation (Figure 1b), the medium-sized ring imposes the unfavorable s-cis
conformation.[7] Thus, construction of such 8-membered salicylides is challenging, and
reported approaches have been limited, e.g., condensation of salicylic acid derivative[8]
and the Baeyer–Villiger oxidation of cyclohexanedione.[7a] Indeed, Minehan et al.
reported the synthesis of the dimethyl-protected derivative of 1, although the
deprotection failed presumably due to the instability of the 8-membered salicylide core.[5]
In view of our long-standing synthetic interest in aryl C-glycoside natural products,[9]
we were intrigued by the synthesis of 1 possessing unusual C-glycoside structure
attached to the 8-membered salicylide. In addition, during the synthetic planning
process, we noted another anomaly in 1 in that the oxygenation pattern of the aromatic
part, 2, 3, 4-trihydroxybenzoic acid, is rare in nature. This is in contrast to the
commonly-seen 3, 4, 5-oxygenation pattern of gallic acid (Figure 1c).
Our presentiments surrounding the sensitive salicylide structure as well as the unusual
aromatic substitution pattern proved true. In this paper, we wish to report the first total
synthesis of the proposed structure 1, which was found to not represent the natural
product.
Figure 1. Structural features of 1
This article is protected by copyright. All rights reserved.

Results and Discussion
1. Retrosynthetic Analysis
Scheme 1 depicts the retrosynthetic analysis of 1. The foreseen challenges included: (1)
the selective formation of the 8-membered salicylide that would have sizable strain, and
(2) the regio- and stereoselective construction of the aryl C-glycoside linkage.
We envisioned that the key salicylide core could be constructed by dimerization of the
monomeric C-glycoside A. Two routes were examined: (1) direct dimerization, and (2)
stepwise approach. The β-linked glucoside in A would be constructed via two possible
approaches. Approach A is based on the
O
C glycoside rearrangement previously
developed in our laboratory,[9] and approach B relies on the nucleophilic addition of
organometallics followed by hydride reduction.[10]
Scheme 1. Synthetic plan
2. C-Glycoside formation
2.1 Approach A Using the
formation of the C-glycoside bond (Scheme 2). An initial trial was conducted using fluoro
sugar 3, phenol 4, and BF3·OEt2 as the Lewis acid of choice (Drierite®, CH2Cl2, 0
40
O
C glycoside rearrangement procedure, we examined the
°C). However, this phenol displayed low reactivity, resulting in the isolation of O-
glycoside 5 in 52% yield (Scheme 2a). To improve the reactivity of the
This article is protected by copyright. All rights reserved.

Scheme 2. Attempted O C glycoside rearrangement (approach A)
phenolic partner, other starting materials were employed (Scheme 2b). Indeed, phenol 6
underwent the O C glycoside rearrangement with fluoro sugar 3 in the presence of BF3
OEt2 (MS 4Å, CH2Cl2, 0
Treatment of 7 with bromine afforded bromide 8, which was protected (methoxymethyl
chloride, NaH, DMF, 0 RT) to give methoxymethyl ether 9. Bromide 9 was treated
with n-BuLi (THF, –78 °C, 30 min), and the resulting aryl anion was allowed to react with
ClCO2Me (–78 °C RT), affording methyl ester 10. The crude ester 10 was subjected to
C
RT), giving the desired C-glycoside 7 in 60% yield.
C
saponification, yielding acid 11 in 34% yield in 2 steps. Finally, selective removal of the
methyl group, ortho to the carbonyl moiety, was examined using a combination of
CeCl3·7H2O and NaI, which, however, resulted in failure, giving only decarboxylaed
product of the starting material 11.
2.2 Approach B Faced with the difficulty in the O C glycoside rearrangement, we
decided to change the synthetic approach to nucleophilic addition of organometallic reagent
to the protected gluconolactone. Scheme 3 shows the synthesis of iodide 17. Following the
known protocol, 2,3,4-trimethoxybenzoic acid (13) was converted to catechol 14 in two
steps.[11] Initial attempts at the selective mono-benzylation of catechol 14 (BnBr, Cs2CO3,
acetone, 80 °C) gave poor results, affording the dibenzylated product. However, the
projected mono-benzylation became possible by employing i-Pr2NEt[12] (BnBr, cat. NaI,
CH2Cl2, RT), giving mono-benzyl ether 15 in 80% yield. The structure of 15 was assigned
by HMBC correlations as shown in Scheme 3. Iodination of 15 with N-iodosuccinimide
(NIS) in the presence of NaHCO3 gave iodide 16, which was treated with allyl bromide
and NaH (DMF, 0 °C
RT) to give allyl ether 17. The structure of 17 was verified by the
diagnostic HMBC cross peak (see Scheme 3).
This article is protected by copyright. All rights reserved.

Scheme 3. Synthesis of iodide 17.
Aryl iodide 17 was subjected to halogen–metal exchange reaction, and subsequently allowed to react with D-gluconolactone 18
(Scheme 4). The concern at this stage was the compatibility of the methoxycarbonyl group. Indeed, use of n-BuLi gave lactol 19
in 24% yield with many unidentified side products. Pleasingly, screening identified i-PrMgCl·LiCl[13] as the reagent of choice:
Treatment of iodide 17 with i-PrMgCl·LiCl (THF, –78 °C, 20 min) followed by the addition of lactone 18 (–25 °C, 4.5 h) afforded
lactol 19 in 72% yield as a sole C-β isomer, which was confirmed by the ROESY correlation (see Scheme 4).
Scheme 4. Synthesis of hydroxy acid 23
The next stage was the Lewis acid-mediated silane reduction of lactol 19 (Table 1).
Treatment of lactol 19 with Et3SiH in the presence of 3.0 equivalents of BF3 OEt2
(CH2Cl2, –78
0 °C) gave C-glycoside 20 in 71% yield as a mixture of β/α anomers
(90:10) (entry 1). The configurations were verified by the NOE studies on the separated
anomers (Figure 2a). Use of i-Pr3SiH [BF3·OEt2 (3.0 equiv.), CH2Cl2, –40 0 °C][14]
reinforced the β stereoselectivity to give 20 in 52% yield. Byproduct 21 (Figure 2b) was obtained in
22% yield, arising from the debenzylation of the phenol protecting group (entry 2). To moderate
the Lewis acidity to suppress the undesired debenzylation, a mixed solvent system (MeCN and CH2Cl2 = 3/1)
was employed, which allowed in a slight increase of the yield of 20 (69%), and the
formation of 21 was almost suppressed (8%) (entry 3). An optimal set of conditions, i.e.,
use of i-Pr3SiH in the presence of 1.5 equivalents of BF3·OEt2 (MeCN, –40 0 °C),
afforded β C-glycoside 20 in high yield without producing 21 (entry 4). The conditions
were scalable to the level employing 8.0 g of 19 as the starting material (entry 5).
This article is protected by copyright. All rights reserved.

Table 1. Reduction of lactol 19
Entry
Silanea
Et3SiH
X [equiv.]
3.0
Solvent
Temp. [°C]
Time [h]
20 [%]b
71
21 [%]b
ND
22
Ratio (β:α)
90:10
1
CH2Cl2
1.2
3.5
2.0
2.0
2.0
–78
–40
–40
–40
–40
0
0
0
0
0
2
i-Pr3SiH
i-Pr3SiH
i-Pr3SiH
i-Pr3SiH
3.0
CH2Cl2
52
>99:<1
>99:<1
>99:<1
>99:<1
c
3
3.0
CH3CN/CH2Cl2
CH3CN
69
8
4
1.5
81
7
5d
1.5
CH3CN
82
ND
a Three equivalents of silane were used. b Yield after purification by preparative TLC. C CH3CN/CH2Cl2 = 3/1. d 8.0 g scale.
Figure 2. NOE correlations of C-glycoside 20 and phenol byproduct 21
C-Glycoside 20 was treated with a catalytic amount of Pd(PPh3)4 in the presence of
morpholine (THF, RT) to give deallylated product 22, which was hydrolyzed to give
hydroxy acid 23 in 98% yield over 2 steps (Scheme 4). In the 1H-NMR spectra of 22 (300
K, 600 MHz), H(1) of the glucosyl moiety was observed as a broad singlet at δ 4.58. The
broadening suggested the slow rotation of the aryl C-glycoside bond (Figure 3). At 323 K
(400 MHz), this signal was observed as a doublet (3JHH = 9.6 Hz), confirming the β-
configuration.
This article is protected by copyright. All rights reserved.

4.56 (d, J = 9.6 Hz)
at 323 K
CDCl3
400 MHz
OBn
MeO
a
OH
O
BnO
BnO
BnO
H
O
OMe
OBn
H
J = 9.6 Hz
a
OBn
at 300 K
CDCl3
600 MHz
OH
OMe
MeO
O
BnO
BnO
BnO
OBn
O
H
slow
22
Figure 3. Variable-temperature 1H-NMR of compound 22
3. Salicylide formation
3.1 Model study In order to test the feasibility of a direct dimerization, a model study was carried out using model compound 24
(Scheme 5). After screening many different condensing agents, we identified benzenesulfonyl chloride as an effective reagent,
giving the desired salicylide 25 in 32% yield along with the undesired trimer 26 (33%) and tetramer 27 (18%). In the IR spectrum
of salicylide 25, the C=O stretching band was observed at 1758 cm–1. Compared to compound 26 (1737 cm–1), the band clearly
shifted to higher frequency, suggesting that the ester linkage of the salicylide core adopts s-cis conformation. These
results are consistent with the pioneering study by Seebach and coworkers.[7a]
Furthermore, we obtained single crystals of dimer 25 and trimer 26 suitable for X-ray
analysises. It turned out that the ester linkages of salicylide 25 adopt s-cis conformation, and
those of trimer 26 s-trans conformation, respectively (Figure 4). Interestingly, dimer 25 is
strongly distorted from planarity, while trimer 26 was almost planar.[6][7]
Scheme 5. Direct dimerization of model compound 24
dimer 25
trimer 26
top view
side view
top view
side view
Figure 4. X-Ray structures of dimer 25 and trimer 26
This article is protected by copyright. All rights reserved.

Another interesting observation was the bowl chirality in the related dimer 28 (Figure 5). Note that the protections for C(3) and
1
C(3’)-phenols in 28 are benzyl- rather than methyl-groups. In the H NMR spectra of 28 (DMSO-d6, 400MHz), the geminal
coupling between two benzyl protons was observed, suggesting the 8-membered salicylide core possesses bowl chirality.[7] To
observe the dynamic behavior, variable-temperature 1H NMR experiments were performed. During raising the temperature, the
peaks became broadened and coalesced at 355 K, and further heating sharpened the peak. The estimated barrier for flipping
was ΔG‡ = 18 kcal mol–1.[7c]
Figure 5. Dynamic NMR experiment (DMSO-d6, 400 MHz)
3.2 Direct dimerization We proceeded to explore the optimal conditions for the direct dimerization of C-glycosidic salicylic acid
23 (Table 2). It turned out that EDCI (A), trichlorobenzoyl chloride (B), and MNBA (C) were not effective for this transformation,
giving salicylide 28 in low yields with many unidentified products (entries 1–3). On the other hand, treatment of 23 with SOCl2
followed by the addition of Et3N and DMAP in toluene afforded trimer 30 in 40% yield (entry 4). Under the Mukaiyama conditions
using 2-chloro-N-methylpyridinium salt, the formation of salicylide 29 was improved (19%), although sizable amounts of trimer
30 and tetramer 31 were produced (entry 5). Use of PhSO2Cl led to a slight increase in the yield of 29 (20%), although formation
of trimer 30 and tetramer 31 still persisted (entry 6). Preferential formation of the trimer and the tetramer rather than the desired
dimer could be attributed to the low population of the s-cis conformer of the acyclic dimer 32, required for the desired
cyclodimerization to give salicylide 29 (Figure 6). In the IR spectrum of salicylide 29, the C=O stretching band was observed
again at 1763 cm–1, suggesting that the ester linkage of the salicylide core adopts s-cis conformation.
Table 2. Direct salicylide formation
This article is protected by copyright. All rights reserved.

Entry
Reagents
A, DMAP
Solvent
CH2Cl2
CH2Cl2
CH3CN
Temp
RT
Time [h]
6.0
29a [%]
30 a [%]
31 a [%]
1
2
3
4
9
18
ND
B, Et3N, DMAP
C, Et3N
RT
6.0
11
ND
trace
RT
8.5
decomposed
40
SOCl2, BnEt3NClb
then Et3N, DMAP
Cl(CH2)2Cl
RT
4.0
11
ND
then toluene
5
6
D, Et3N, DMAP
toluene
pyridine
90 °C
RT
5.0
4.0
19
20
23
37
9
PhSO2Cl
12
a Yields after purification by preparative TLC. b Acid 23 was treated with thionyl chloride and BnEt3NCl for 1.5 h followed by removal of the solvent.
The residue was added to the toluene solution of Et3N and DMAP, and stirred for 2.5 h.
Figure 6. s-Cis and s-trans conformers of the reaction intermediate
3.3 Stepwise dimerization Faced with the difficulty in the direct dimerization, we opted for a stepwise approach, that is, to form
the two ester linkages as separate steps. For this purpose, we prepared benzoic acid 34 and phenol 35 as the differentially-
protected monomer units (Scheme 6). Saponification of 20 (10% NaOH aq., 1,4-dioxane, reflux) gave benzoic acid 34 in 94%
yield, while carboxylic acid 23 was converted to allyl ester 35 in 90% yield. Acid 34 was converted to the corresponding acid
chloride by treatment with thionyl chloride in the presence of a catalytic amount of DMF (CH2Cl2, RT). After removal of the
solvents, to the residue was added a toluene solution of phenol 35 followed by Et3N and DMAP to give ester 36 in 91% yield.
Simultaneous removal of two allyl groups in 36 gave seco acid 37, which was ready for the lactonization step.
This article is protected by copyright. All rights reserved.

Scheme 6. Synthesis of seco acid 37
Table 3 summarizes the results of the salicylide formation. Treatment of 37 with EDCI (A) and DMAP (CH2Cl2, RT) caused only
decomposition (entry 1). Use of MNBA (C)[15] or 2-chloro-N-methylpyridinium iodide (D)[16] as a condensing agent gave the
desired salicylide 29 in moderate yields (entries 2 and 3). By contrast, an excellent result was obtained by the Yamaguchi
lactonization.[17] The initial trial using the standard conditions, such that treatment of acid 37 with 2,4,6-trichlorobenzoyl chloride
(B) (Et3N, toluene, RT), failed to convert 37 to the mixed anhydride. The attempted lactonization under the Yonemitsu conditions
[the seco acid 37 was added to a toluene solution of 2,4,6-trichlorobenzoyl chloride, Et3N, and 2.0 equivalents of DMAP (5.0
mM, RT)] gave rise to decomposed products, resulting from the DMAP-induced cleavage of the ester linkage (entry 4).[18]
However, reducing the amount of DMAP to 0.2 equivalents gave salicylide 29 in high yield (entry 5). Use of 4-pyrrolidinopyridine
(4-PPY) led to a slight increase of the yield (entry 6).
Table 3. Salicylide formation
Entry
Reagents
Solvent
Temp.
29
1
2
3
4
5
A, DMAP (1.2)
CH2Cl2
toluene
CH3CN
toluene
toluene
RT
ND
65
C, Et3N (2.2), DMAP (0.2)
D, Et3N (8.0)
RT
90 °C
RT
69
B, Et3N (1.2), DMAP (2.0)
B, Et3N (2.2), DMAP (0.2)
ND
85
RT
This article is protected by copyright. All rights reserved.

6
B, Et3N (2.2), 4-PPY (0.2)
toluene
RT
88
4. Structure elucidation
Finally, global deprotection of 29 [H2, Pd(OH)2, THF, room temperature] afforded the targeted structure 1 (Scheme 7). To our
surprise, the product turned out to be extremely unstable, easily prone to solvolytic ring opening. For example, when the crude
material was suspended in water, partial cleavage of the salicylide occurred, giving hydroxy acid 38. When dissolved in CD3OD,
the salicylide was immediately solvolyzed, giving d3-methyl ester 39. Attempted isolation of 1 by the reported protocol (reversed-
phase HPLC, MeOH:H2O = 2:8)[3] also caused decomposition, giving a trace amount of monomeric methyl ester 40. The crude
material of 1 was obtained by careful filtration under an argon atmosphere through a membrane filter (THF), and then
1
concentrated in vacuo. Using this crude material, the structure of compound 1 was unambiguously confirmed by H and 13C
NMR spectroscopy and high-resolution mass spectroscopy (HRMS). The HRMS (ESI, negative) of the synthetic material 1
indicated the molecular formula of C28H31O18 ([M–H]– m/z 655.1510). Thanks to the generous donation of the NMR spectra
measured in DMSO-d6 from Prof. Jung, we compared the 13C-NMR spectra of the synthetic 1 with that of the natural product.
The data did not match, suggesting the structure incongruity (Table 4).
Scheme 7. Synthesis of 1 and its instability to H2O and MeOH
This article is protected by copyright. All rights reserved.

Table 4. Comparison of 13C-NMR reported for the natural product and the synthetic material 1a
Reported b
Synthetic c
Reported b
Synthetic c
δC
δC major (minor)
δC
δC major (minor)
163.5
151.2
148.2
140.7
118.2
116.0
109.6
164.5
81.9
79.9
73.8
72.3
70.8
61.2
59.9
81.9 (81.7)
78.3 (78.5)
74.6
151.4 (152.0)
142.4 (142.7)
137.7 (1137.9)
133.4 (133.1)
119.1 (118.8)
117.7 (118.3)
73.7 (72.7)
70.3
61.4 (61.3)
61.21 (61.24)
a Measured in DMSO-d6. b 400 MHz. c 600 MHz.
Figure 7. Partial 1H NMR spectrum of compound 1 (600 MHz, DMSO-d6)
This article is protected by copyright. All rights reserved.

Interestingly, the NMR spectrum of the synthetic sample suggested the presence of two diastereomers that is due to the bowl
chirality relative to the glycoside chirality (Figure 7).
At this stage, we asked ourselves what the real structure of the natural product was. We found a clue: the differences in the 13C
NMR spectra between the synthetic material and the reported data mostly centered on the aromatic peaks, strongly suggesting
the incongruity of the aromatic substitution pattern. A notion came to us that the 2,3,4-trihydroxybenzoic acid derivatives are
relatively rare compared to the 3,4,5-trihydroxycounterpart (gallic acid). A literature search indicated the ratio of roughly 1:15.
1
The original report mentioned “the H and 13C NMR spectra of the natural product were reminiscent of bergenin (41)”,[3] which
led us to a naive assumption that the bergenin dimers might be the true natural product (Figure 8). In order to prepare the
assumed dimers, we purchased a sample of 41 (Tokyo Chemical Industry Co., Inc.) and were surprised to notice that the 1H and
13C NMR coincided with the reported data for “ardimerin”,1 which was also the case with the per-acetyl derivative (Ac2O,
pyridine, RT). In view of the fact, we had a strong suspicion that “ardimerin” is identical to bergenin itself.
Figure 8. Structures of bergenin (41) and its hypothetical dimers
The question was why the misassignment had occurred. The basis of the original assignment of the dimeric structure came
from the high-resolution FAB-MS data (m/z 657.4563, C28H32O6). Surprisingly, the FAB-MS spectra by our measurement of
bergenin (41) showed the cluster ion peak observed at m/z 657 as [2M+H]+ rather than the parent peak m/z 329 [M+H]+
(Figure 9).2, [20]
Figure 9. FAB-MS spectra of bergenin
We identified another pitfall leading to the misassignment in the NMR interpretation (Figure 10). In the HMBC spectrum of
4
bergenin (41), the JHC correlation between the aryl proton and C(1’) of the glucosyl moiety was clearly observed (red arrow,
3
Figure 10a). In the original paper,[1] this HMBC correlation was erroneously assigned as JHC between the aryl proton and C(1’)
1 Further supporting fact is that bergenin was previously isolated from Ardisia japonica.[19] Strangely, however, the specific
optical rotation of ardimerin (+48.2°) was opposite of the sign for the reported data of bergenin (–38.9°).
2 The cluster ion peak was more clearly observed in the ESI-MS spectra (see Supporting Information).
This article is protected by copyright. All rights reserved.

as shown by the red arrow in Figure 10b, which led them to conclude the vicinal relationship of the aryl proton and the C-
3
glycoside linkage. Moreover, unfortunately, bergenin showed no JHC correlation between H(2’) and the carbonyl carbon, which
gave a further support to the misassignment.
Figure 10. HMBC correlation of bergenin and the original misassignment
Conclusions
In conclusion, the first total synthesis of 1, the proposed structure of ardimerin, was
accomplished via a stepwise dimerization of the C-glycoside. The analytical study
revealed that the natural product isolated by Jung et al. and named as “ardimerin“ does
not correspond to compound 1. We propose that the identity is bergenin.
Experimental Section
General
All reactions utilizing air- or moisture-sensitive reagents were performed in dried glassware under an atmosphere of dry argon.
The authentic sample of bergenin was obtained from Tokyo Chemical Industry Co., Inc. Ethereal solvents (anhydrous; Kanto
Chemical Co., Inc.) were used as received. DMF and acetonitrile were distilled over CaH2, hexane was distilled from Na, and
stored over 4Å molecular sieves. Other reagents were used without further purification as received from commercial suppliers.
For thin-layer chromatography (TLC) analysis, Merck pre-coated plates (TLC silica gel 60 F254, Art 5715, 0.25 mm) were used.
Silica-gel preparative thin-layer chromatography (PTLC) was performed using plates prepared from Merck Silica gel 60 PF254
(Art7747). For flash column chromatography, silica gel 60N (Spherical, neutral, 63-210 m) from Kanto Chemical was used.
Higher-accuracy purifications were performed by Smart Flash EPCLC W-Prep 2XY system (YAMAZEN SCIENCE, inc.). Melting
point (mp) determinations were performed by using a Yanaco MP-500 instrument or a METTLER TOLEDO MP70 melting point
1
system, and are uncorrected. H- and 13C-NMR were measured on a JEOL JNM Lamda-400 (400 MHz), JEOL JNM ECX-500
(500 MHz), or Bruker AV-600 (600 MHz) spectrometer in the solvent indicated; Chemical shifts (δ) are expressed in parts per
million downfield from internal standard (tetramethylsilane), and coupling constants are reported as hertz (Hz). Splitting patterns
are indicated as follows: br, broad; s, singlet; d, doublet; t, triplet; q, quartet; quint, quintet; m, multiplet. Infrared (IR) spectra
were recorded on a Perkin-Elmer Spectrum 100 FTIR spectrometer. Attenuated total reflectance Fourier transform infrared
(ATR-FTIR) spectra were recorded by using a Perkin-Elmer 100 FTIR spectrometer equipped with a universal ATR sampling
accessory. High-resolution mass spectra (HR-MS) were obtained with Bruker Daltonics micrOTOF-Q II. X-Ray crystallographic
data were recorded with a Rigaku RAXIS-RAPID diffractometer.
This article is protected by copyright. All rights reserved.

Methyl 3-benzyl-2-hydroxy-4-methoxybenzoate (15).
To a solution of 14 (101 mg, 511 μmol), DIPEA(180 μL, 1.03 mmol) and NaI (15.3 mg, 102 μmol) in CH2Cl2 (2.0 mL) was added
BnBr (61.0 μL, 512 μmol) at room temperature. After stirring for 20 h, the reaction was stopped by adding saturated aqueous
NH4Cl, and the mixture was extracted with EtOAc (×3). The combined organic extracts were washed with brine, dried (Na2SO4),
and concentrated in vacuo. The residue was purified by preparative TLC (hexane/Et2O = 2/1) to afford 15 (118 mg, 80%) as a
1
3
white solid. Rf 0.54 (hexane/Et2O 2:1); H-NMR (600 MHz, CDCl3) 3.88 (s, Me); 3.95 (s, Me); 5.11 (s, CH2); 6.48 (d, J(H,H) =
9.0, H–C); 7.32 (t, 3J(H,H) = 7.4, H–C); 7.38 (t 3J(H,H) = 7.4, H–C(2)); 7.55 (d, 3J(H,H) = 7.4, H–C(2)); 7.63 (d, 3J(H,H) = 9.0, H–
C); 10.96 (s, OH). Anal. calc. for C16H16O5 (288.30): C 66.66, H 5.59; found: C 66.78, H 5.45. Spectral data matched the
reported data.[21]
Methyl 3-benzyloxy-2-hydroxy-5-iodo-4-methoxybenzoate (16).
To a suspension of 15 (751 mg, 2.60 mmol), NaHCO3 (2.21 g, 26.3 mmol) in MeOH (26
mL) was added NIS (884 mg, 3.93 mmol) at 0 °C. After stirring for 1 h at room
temperature, the resulting mixture was diluted with CH2Cl2 and H2O was added. The
aqueous phase was separated and extracted with CH2Cl2 (×3). The combined organic
extracts were washed with 15% aqueous Na2S2O3 and water, dried over anhydrous
Na2SO4, filtered and then concentrated in vacuo. The residue was purified by flash
column chromatography (silica gel, hexane/EtOAc = 95/5) to afford 16 (1.01 g, 94%) as a
1
white solid. Rf = 0.60 (hexane/EtOAc 4:1). M.p. 88–90 °C. H-NMR (600 MHz, CDCl3):
3.95 (s, Me); 3.96 (s, Me); 5.07 (s, CH2); 7.33–7.39 (m, H–C(3)), 7.51 (d, 3J(H,H) = 7.4, H–
C(2)); 8.03 (s, H–C), 10.97 (s, 1H, OH). 13C-NMR (150 MHz, CDCl3): 52.5; 61.3; 75.3; 78.7;
111.1; 128.2; 128.39 (2C); 128.44 (2C); 134.0; 136.9; 139.9; 157.4; 158.0; 169.3. IR (ATR):
3204, 3029, 2951, 1683 (C=O), 1598, 1439, 1418, 1319, 1255, 1201, 1019, 956, 739, 694.
+
HR-ESI-MS: 436.9845 ([M+Na]+, C16H15INaO5 ; calc. 436.9856).
Methyl 2-allyloxy-3-benzyloxy-5-iodo-4-methoxybenzoate (17).
To a suspension of NaH (63% dispersion in mineral oil, 316 mg, 8.30 mmol) in DMF (20 mL) was added 16 (2.85 g, 6.88 mmol)
in DMF (50 mL) followed by allyl bromide (1.15 mL, 13.5 mmol) at 0 °C. After stirring for 13 h at room temperature, the reaction
was quenched by adding saturated aqueous NH4Cl. The resulting mixture was extracted with Et2O (×3). The combined organic
extracts were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated in vacuo. The residue was purified by
flash column chromatography (silica gel, hexane/Et2O = 90/10) to afford 17 (2.81 g, 89%) as a white solid. Rf = 0.48
(hexane/EtOAc 4:1). M.p. 47–49 °C. 1H-NMR (600 MHz, CDCl3): 3.89 (s, Me); 3.93 (s, Me); 4.60 (d, 3J(H,H) = 5.9, CH2); 5.02 (s,
3
2
3
2
3
CH2); 5.24 (dd, J(H,H) = 10.4, J(H,H) = 1.1, H–C); 5.35 (dd, J(H,H) = 17.1, J(H,H) = 1.1, H–C); 6.89 (ddt, J(H,H) = 17.1,
10.4, 5.9, H–C); 7.34–7.41 (m, H–C(3)); 7.48 (d, 3J(H,H)= 7.14, H–C(3)); 8.02 (s, H–C). 13C-NMR (150 MHz, CDCl3): 52.2; 61.2;
75.5; 75.9; 85.1; 118.3; 123.2; 128.3; 128.4 (2C); 128.9 (2C); 133.5; 135.3; 136.7; 133.5; 135.3; 136.7; 146.3; 154.3; 157.3;
164.7. IR (ATR): 3069, 3031, 2979, 2944, 2894, 1731 (C=O), 1574, 1415, 1280, 1190, 1068, 1017, 967, 926, 692. HR-ESI-MS:
+
499.0172 ([M+Na]+, C19H19INaO5 ; calc. 499.0169).
Lactol 19.
Iodide 17 (2.98 g, 6.56 mmol) was azeotropically dried with toluene (5 mL × 3) and dissolved in THF (70 mL), to which was
added i-PrMgCl·LiCl (10.8 mL, 0.95 M THF soln.) at –78 °C. After stirring for 10 min at this temperature, gluconolactone 18
[azeotropically dried with toluene (5mL × 3), 1.3 M in THF, 5.0 mL, 6.57 mmol] was added at –78 °C, and the reaction
temperature was immediately raised to –25 °C. After stirring for 4 h at this temperature, the reaction was quenched by adding
AcOH (1.2 M solution in THF, 10.0 mL, 12.0 mmol) dropwise at –78 °C. After stirring for 30 min at this temperature, the mixture
was warmed to room temperature. To it was added saturated aqueous NaHCO3, and the aqueous phase was extracted with
This article is protected by copyright. All rights reserved.

EtOAc (× 3). The combined organic extracts were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated in
vacuo. The residue was purified by flash column chromatography (silica gel, hexane/EtOAc = 90/10 69/31) to afford lactol 19
(4.11 g, 72%) as a colorless oil. Rf = 0.27 (toluene/Et2O 85:15). [α]36520 = –44.4 (c = 0.289, CHCl3) 1H-NMR (600 MHz, CDCl3):
3.71 (d, 2J(H,H) = 10.7, CH2(1)); 3.81–3.87 (m, CH2(1), Me (6)); 3.89–3.95 (m, H–C, CH2(1)); 4.02 (dd, 3J(H,H) = 9.1, 9.1, H–C);
2
2
2
4.12 (brt, H–C(2)); 4.48 (d, J(H,H) = 11.5, CH2(1)); 4.49 (d, J(H,H) = 9.1, CH2(1)); 5.54 (d, J(H,H) = 11.9, CH2(1)); 4.59 (dd,
3
2
3
2J(H,H) = 11.6, J(H,H) = 5.7, CH2(1)); 4.67 (dd, J(H,H) = 11.6, J(H,H) = 5.7, CH2(1)); 4.69 (d, 2J(H,H) = 11.0, CH2(1)); 4.89–
3
4.93 (m, CH2(4)); 4.98 (d, 2J(H,H) = 10.7, CH2(1)); 5.24 (d, 3J(H,H) = 10.3, CH2(1)); 5.37 (d, J(H,H) = 17.1, CH2(1)); 6.14 (ddt,
3J(H,H) = 17.1, 10.3, 5.7, H–C); 6.93 (d, 3J(H,H) = 7.2, H–C(2)); 7.16 (dd, 3J(H,H) = 14.3, H–C(3)); 7.24–7.33 (m, H–C(19)); 7.44
3
(d, J(H,H) = 7.2, H–C(2)); 7.93 (s, H–C). 13C-NMR (150 MHz, CDCl3, several signals overlapped): 29.7; 52.0; 62.4; 68.9; 72.1;
73.2; 75.0; 75.1; 75.4; 75.7; 78.0; 83.2; 83.7; 97.8; 118.2; 120.8; 125.3; 127.5; 127.66; 127.69; 127.76; 127.78; 127.8; 128.2;
128.3; 128.4; 128.5; 130.7; 130.7; 133.8; 136.9; 137.4; 138.2; 138.4; 138.6; 146.8; 154.2; 156.1; 165.6. IR (neat) 3441, 3063,
+
3030, 2947, 2867, 1728 (C=O), 1598, 1454, 1310, 1212, 1086, 1013, 736, 697. HR-ESI-MS: 889.3525 ([M+Na]+, C53H54NaO11
;
calc. 889.3558).
C-Glucoside 20.
Lactol 19 (7.90 g, 9.11 mmol) was azeotropically dried with toluene (5 mL × 3) and dissolved in CH3CN (80 mL), to which was
added i-Pr3SiH (5.59 mL, 27.2 mmol) and BF3·OEt2 (1.70 mL, 13.4 mmol) at –40 °C. After stirring for 15 min at this temperature,
the reaction mixture was warmed to 0 °C. After stirring for 65 min, the reaction was quenched by adding saturated aqueous
NaHCO3. The two phases were separated and the aqueous phase was back-extracted with EtOAc (×3). The combined organic
extracts were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated in vacuo. The residue was purified by
flash column chromatography (silica gel, hexane/EtOAc = 80/20) to afford C-glycoside 20 (6.35 g, 82%) as a colorless oil. Rf =
0.28 (hexane/EtOAc 4:1). [α]36523 = –24.1 (c = 1.10, CHCl3). 1H-NMR (600 MHz, CDCl3): 3.58–3.64 (br, H–C); 3.77–3.85 (m, H–
C(3), CH2); 3.87 (s, Me); 3.88 (s, Me); 4.02 (d, 2J(H,H) = 10.7, CH2(1)); 4.52 (d, 2J(H,H) = 10.7, CH2(1)); 4.54 (d, 2J(H,H) = 12.2,
2
2
CH2(1)); 4.61 (d, J(H,H) = 12.2, CH2(1)); 4.62–4.64 (m, H–C, CH2(3)); 4.87 (d, J(H,H) = 10.7, CH2(1)); 4.91–5.00 (m, CH2(4));
5.24 (d, 2J(H,H) = 10.4, CH2(1)); 5.37 (dd, 3J(H,H) = 17.2, 2J(H,H) = 1.2, CH2(1)); 6.14 (ddt, 3J(H,H) = 17.2, 10.4, 5.9, H–C); 6.89
(d, 3J(H,H) = 6.1, H–C(2)); 7.12 (m, H–C(3)); 7.20 (m, H–C(2)); 7.25–7.40 (m, H–C(16)); 7.48 (d, 3J(H,H) = 7.3, H–C(2)); 7.73 (s,
H–C). 13C-NMR (150 MHz, CDCl3, several signals overlapped): 52.0; 61.9; 69.1; 73.3; 75.0; 75.1; 75.45; 75.50; 75.53; 78.4;
79.4; 83.0; 87.0; 118.1; 121.2; 127.51; 127.52; 127.6; 127.67; 127.72; 128.0; 128.07; 128.10; 128.2; 128.3; 128.4; 128.5; 128.6;
133.9; 137.1; 137.6; 138.19; 138.24; 138.71; 153.6; 156.5; 165.6. IR (neat) 3064, 3031, 2941, 2903, 2868, 1728 (C=O), 1599,
1454, 1322, 1209, 1070, 1028, 994, 736, 697. HR-ESI-MS: 873.3596 ([M+Na]+, C53H54NaO10+; calc. 873.3609).
C-Glycoside 22.
To a solution of 20 (3.17 g, 3.73 mmol) in THF (35 mL) was added Pd(PPh3)4 (435 mg, 0.376 mmol) and morpholine (3.08 ml,
35.3 mmol) at room temperature. After stirring for 2 h, the reaction was quenched by adding saturated aqueous NH4Cl, and the
resulting mixture was extracted with EtOAc (×3). The combined organic extracts were washed with brine, dried over anhydrous
Na2SO4, filtered and concentrated in vacuo The residue was purified by flash column chromatography (silica gel, hexane/EtOAc
20
= 90/10 80/20) to afford 22 (3.03 g, quant.) as a colorless oil. Rf = 0.35 (hexane/Et2O 6:4). [α]365 = –125 (c = 1.10, CHCl3).
2
1H-NMR (600 MHz, CDCl3): 3.57–3.62 (brd, H–C); 3.69 (brs, H–C); 3.73 (d, J(H,H) = 11.0, CH2(1)); 3.75–3.85 (m, CH2(1), H–
2
2
2
C(2)); 3.91 (s, Me); 3.92 (s, Me); 4.09 (d, J(H,H) = 10.8, CH2(1)); 4.51 (d, J(H,H) = 10.8, CH2(1)); 4.52 (d, J(H,H) = 12.2,
2
2
2
CH2(1)); 4.58 (br, 1H); 4.62 (d, J(H,H) = 12.2, CH2(1)); 4.65 (d, J(H,H) = 10.7, CH2(1)); 4.87 (d, J(H,H) = 10.7, CH2(1)); 4.92
2
2
2
2
(d, J(H,H) = 11.0, CH2(1)); 4.95 (d, J(H,H) = 11.0, CH2(1)); 5.02 (d, J(H,H) = 11.0, CH2(1)); 5.05 (d, J(H,H) = 11.0, CH2(1));
6.89 (d, 3J(H,H) = 7.0, H–C(2)); 7.08–7.15 (m, H–C(3)); 7.20 (d, 3J(H,H) = 7.0, H–C(2)); 7.25–7.39 (m, H–C(16)); 7.52 (d, 3J(H,H)
= 7.4, H–C(2)); 7.61 (s, H–C); 10.99 (s, OH). 13C-NMR (150 MHz, CDCl3, several signals overlapped): 61.8; 69.1; 73.4; 74.8;
74.0; 75.1; 75.5; 78.4; 79.3; 82.8; 87.1; 109.0; 124.0; 127.4; 127.52; 127.55; 127.62; 127.65; 128.0; 128.1; 128.2; 128.29;
128.33; 128.4; 137.3; 137.6; 138.2; 138.3; 138.7; 139.6; 156.5; 157.5; 170.3. IR (neat) 3030, 2923, 2855, 1673 (C=O), 1435,
1342, 1211, 1070, 736, 697. HR-ESI-MS: 833.3329. ([M+Na]+, C50H50NaO10+; calc. 833.3296).
This article is protected by copyright. All rights reserved.

C-Glycoside 23.
To a solution of 22 (4.94 g, 3.69 mmol) in 1,4-dioxane (30 mL) was added 10% aqueous NaOH (20 mL) at room temperature.
The reaction mixture was refluxed for 1.5 h. After cooling to room temperature, the reaction was quenched by adding 3M HCl
(20 mL), and the mixture was extracted with EtOAc (×3). The combined organic extracts were dried over anhydrous Na2SO4,
filtered and concentrated in vacuo. The residue was purified by flash column chromatography (silica gel, hexane/EtOAc = 60/40
20
+ TFA 1%) to afford 23 (2.88 g, 98%) as a purple amorphous solid. Rf = 0.34 (MeOH/CHCl3 1:9). [α]365 = –118 (c = 0.923,
1
CHCl3); H-NMR (600 MHz, CDCl3): 3.62–3.71 (br, H–C(2)); 3.73–3.79 (m, H–C, CH2(1)); 3.79–3.87 (m, H–C, CH2(1)); 3.94 (s,
2
2
2
2
Me); 4.05 (d, J(H,H) = 10.7, CH2(1)); 4.49 (d, 1H, J(H,H) = 10.7, CH2(1)); 4.55 (d, J(H,H) = 12.2, CH2(1)); 4.58(d, J(H,H) =
2
2
2
10.8, CH2(1)); 4.65 (brd J(H,H) = 12.2, H–C, CH2(1)); 4.87 (d, J(H,H) = 10.8 Hz); 4.91 (d, J(H,H) = 11.0, CH2(1)); 4.96 (d,
3
3
2J(H,H) = 11.0 CH2(1)); 5.03 (s, CH2); 6.88 (d, J(H,H) = 8.0, H–C(2)); 7.09–7.12 (m, H–C(3)); 7.17 (d, J(H,H) = 5.8, H–C(2));
3
7.22–7.37 (m, H–C(16)); 7.49 (d, J(H,H) = 7.4, H–C(2)); 7.65 (s, H–C); 10.91 (s, OH). 13C-NMR (150 MHz, CDCl3, several
signals overlapped): 61.8; 68.9; 73.2; 74.95; 74.98; 75.1; 75.6; 78.4; 78.6; 83.1; 87.1; 108.2; 123.5; 125.3; 127.6; 127.7; 127.8;
127.98; 128.03; 128.14; 128.16; 128.28; 128.39; 128.43; 128.45; 137.3; 137.46; 137.51; 138.1; 138.7; 139.5; 156.9; 157.8;
172.0. IR (neat) 3064, 2924, 2859, 1673 (C=O), 1615, 1497, 1454, 1442, 1343, 1211, 1070, 1028, 798, 736, 697. HR-ESI-MS:
819.3110 ([M+Na]+, C49H48NaO10+; calc. 819.3140). Anal. calc. for C49H48O10 (796.91): C 73.85, H 6.07; found: C 73.64, H 5.89.
C-Glycoside 34.
To a solution of 20 (1.09 g, 1.28 mmol) in 1,4-dioxane (10 mL) was added 10% aqueous NaOH (6.0 mL) at room temperature.
The reaction mixture was refluxed for 1.5 h. After cooling to room temperature, the reaction was quenched by adding 3 M HCl (7
mL), and the mixture was extracted with EtOAc (× 3).The combined organic extracts dried over anhydrous Na2SO4, filtered and
concentrated in vacuo The residue was purified by flash column chromatography (silica gel, hexane/EtOAc = 70/30 + TFA 1%)
to afford 34 (1.01 mg, 94%) as a white amorphous solid. Rf 0.67 (CHCl3/MeOH 9:1). [α]36520 = –63.8 (c = 1.03, CHCl3). 1H-NMR
(600 MHz, CDCl3): 3.61–3.65 (br, H–C); 3.68–3.77 (m, H–C(2), CH2); 3.83 (t, 3J(H,H) = 8.9, H–C); 3.89 (s, Me); 4.09 (d, 2J(H,H)
= 11.2, CH2(1)); 4.53 (d, 2J(H,H) = 12.2, CH2(1)); 4.54–4.59 (m, CH2); 4.61 (d, 2J(H,H) = 10.7, CH2(1)); 4.65 (brd, H–C); 4.82 (d,
2J(H,H) = 6.4, CH2); 4.86 (d, 2J(H,H) = 10.7, CH2(1)); 4.90–4.97 (m, CH2(4)); 5.35 (d, 2J(H,H) = 11.7, CH2(1)); 5.39 (d, 2J(H,H) =
17.0, CH2(1)); 6.03 (ddt, 3J(H,H) = 17.0, 11.7, 6.4, H–C); 6.89 (d, 3J(H,H) = 7.3, H–C(2)); 7.10–7.15 (m, H–C(3)); 7.18–7.19 (m,
H–C(2)); 7.24–7.46 (m, 18H); 8.01 (s, H–C); 10.88 (br, OH). 13C-NMR (150 MHz, CDCl3, several signals are overlapped): 61.9;
69.2; 73.4; 75.0; 75.2; 75.5; 75.9; 76.4; 78.5; 79.5; 82.8; 87.1; 118.0; 121.5; 127.8; 127.96; 128.03; 128.14; 128.37; 128.42;
128.56; 128.66; 130.3; 131.5; 136.4; 137.7; 138.12; 138.16; 138.7; 144.3; 151.6; 157.5; 167.7. IR (neat) 3202, 3031, 2917,
+
2868, 1741, 1745, 1693 (C=O), 1454, 1361, 1312, 1070, 1028, 992, 737, 697. HR-ESI-MS: 859.3429 ([M+Na]+, C52H52NaO10
;
calc. 859.3453).
C-Glycoside 35.
To a solution of 23 (989 mg, 1.24 mmol), and NaHCO3 (161 mg, 1.92 mmol) in DMF (13 mL) was added 1-bromo-2-propene
(0.160 mL, 1.89 mmol) at room temperature. The resulting reaction mixture was refluxed for 2.5 h. After cooling to room
temperature, the reaction was quenched by addition of water. The mixture was extracted with Et2O (× 3). The combined extracts
were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated in vacuo The residue was purified by flash
column chromatography (silica gel, hexane/EtOAc = 90/10 69/31) to afford 35 (939 mg, 90%) as a purple amorphous solid. Rf
20
1
0.54 (hexane/EtOAc 4:1). [α]365 = –128 (c = 1.06, CHCl3). H-NMR (600 MHz, CDCl3): 3.59 (brd, H–C), 3.72–3.77 (m, H–C,
CH2(1)), 3.78–3.86 (m, H–C(2), CH2(1)), 3.92 (s, Me), 4.08 (brd, 2J(H,H) = 10.1, CH2(1)), 4.50–4.52 (m, CH2(2)), 4.56 (br, H–C),
2
2
4.62 (d, 2J(H,H) = 12.2, CH2(1)), 4.65 (d, 2J(H,H) = 10.7, CH2(1)), 4.81 (d, J(H,H) = 5.7, CH2), 4.88 (d, J(H,H) = 10.7, CH2(1)),
2
2
2
4.92 (d, J(H,H) = 11.0, CH2(1)), 4.95 (d, J(H,H) = 11.0, CH2(1)), 5.02 (d, J(H,H) = 10.6, CH2(1)), 5.05 (d, 2J(H,H) = 10.6,
2
2
3
CH2(1)), 5.29 (d, J(H,H) = 10.5, CH2(1)), 5.39 (d, J(H,H) = 17.1, CH2(1)), 5.99 (ddt, J(H,H) = 17.1, 10.5, 5.7, H–C), 6.89 (d,
3J(H,H) = 6.8, H–C(2)), 7.09–7.14 (m, H–C(3)), 7.21 (d, 3J(H,H) = 6.7, H–C(2)), 7.25–7.38 (m, H–C(16)), 7.52 (d, 3J(H,H) = 7.4,
H–C(2)), 7.64 (s, H–C), 11.00 (s, OH). 13C-NMR (150 MHz, CDCl3) 61.9; 65.9; 69.1; 73.4; 74.8; 75.1; 75.6; 78.4; 79.4; 81.4;
82.7; 108.9; 119.1; 123.9; 124.6; 127.48; 127.59; 127.62; 127.69, 127.73, 137.73, 127.93, 128.06, 128.10, 128.14, 128.28;
128.34; 128.39; 128.41; 131.5; 137.3; 137.3; 137.7; 138.2; 138.3; 138.7; 156.7; 157.6; 169.6. IR (neat) 3064, 3030, 2867, 1671
This article is protected by copyright. All rights reserved.

(C=O), 1613, 1454, 1363, 1330, 1207, 1069, 1028, 991, 736, 697. HR-ESI-MS: 859.3431 ([M+Na]+, C52H52NaO10+; calc.
859.3453).
Ester 36.
To a solution of 34 (615 mg, 735 μmol) in CH2Cl2 (7.0 mL) was added DMF (11.5 μ , 149 μmol) and SOCl2 (105 μL, 1.47
mmol) at room temperature. After stirring for 1.5 h at this temperature, the solution was concentrated in vacuo and the volatile
materials were removed by co-evaporation with toluene (5 mL ×1). The residue was taken up in toluene (2.5 mL) and 35 (0.1 M
in toluene, 7.45 mL, 745 μmol) followed by Et3N (115 μL, 829 μmol), and DMAP (22.0 mg, 180 μmol) were added. After stirring
for 1.5 h at room temperature, the reaction was stopped by adding saturated aqueous NH4Cl. The mixture was extracted with
EtOAc (×3). The combined organic extracts were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated in
vacuo. The residue was purified by flash column chromatography (silica gel, hexane/EtOAc = 80/20) to afford 36 (1.11 g, 91%)
as a colorless oil. Unreacted phenol 35 (33.2 mg, 5%) was recovered as a colorless oil. Rf 0.31 (hexane/EtOAc 4:1). [α]36520 = –
1
98.1 (c = 0.435, CHCl3). H-NMR (600 MHz, CDCl3): 3.64 (brt, H–C(2)); 3.70–3.88 (m, H–C(6); CH2(4), Me); 3.89 (s, Me); 4.01
2
2
2
(d, J(H,H) = 10.4, CH2(1)); 4.10 (d, J(H,H) = 10.6, CH2(1)); 4.47–452 (m, CH2(2)); 4.56 (d, J(H,H) = 11.4, CH2(2)); 4.60–4.67
2
2
(m, H–C, CH2(8)); 4.72 (brs, H–C); 4.85–5.02 (m, CH2(11)); 5.05 (d, J(H,H) = 10.5, CH2(1)); 5.08 (d, J(H,H) = 10.3, CH2(1));
5.19 (d, 2J(H,H) = 17.1, CH2(1)); 5.26 (d, 2J(H,H) = 17.2, CH2(1)); 5.80 (ddt, 3J(H,H) = 17.1, 10.5, 5.7, H–C); 6.07 (ddt, 3J(H,H) =
17.2, 10.3, 5.9, H–C); 6.91–6.96 (m, H–C(4)); 7.08–7.39 (m, H–C(44)); 7.49 (d, 3J(H,H) = 7.2, H–C(2)); 7.95 (s, H–C); 8.11 (brs,
H–C). 13C-NMR (150 MHz, CDCl3, several peaks are overlapped): 61.8; 62.0; 65.6; 69.0; 69.1; 73.3; 73.4; 75.1; 75.16; 75.25;
75.30; 75.5; 75.6; 78.35; 78.40; 79.4; 79.5; 86.9; 87.0; 117.8; 118.4; 119.9; 120.1; 127.48; 127.51; 127.57; 127.65; 127.66;
127.70; 127.75; 127.79; 127.9; 128.0; 128.1, 128.2; 128.30; 128.33; 128.38; 128.40; 128.44; 128.46; 128.48; 128.7; 130.9;
132.0; 134.0; 136.8; 137.1; 137.4; 137.7; 138.1; 138.2; 138.6; 138.7; 145.1; 145.6; 154.7; 156.5; 157.2; 162.6; 163.5. IR (neat)
3030, 2867, 1752 (C=O), 1723 (C=O), 1453, 1360, 1320, 1205, 1154, 1067, 990, 736, 697. HR-ESI-MS: 1677.6898 ([M+Na]+,
C
104H102NaO19+; calc. 1677.69075).
Seco acid 37.
To a solution of ester 36 (104 mg, 62.9 μmol) in THF (3.0 mL) was added Pd(PPh3)4 (7.0 mg, 6.1 μmol) and dimedone (43.0
mg, 307 μmol) at room temperature. After stirring for 2.0 h, the reaction was quenched upon addition of saturated aqueous
NH4Cl, and the mixture was extracted with EtOAc (×3). The combined organic extracts were washed with brine, dried over
anhydrous Na2SO4, filtered and concentrated in vacuo The residue was purified by preparative TLC (silica gel, hexane/EtOAc =
50/50) to afford 37 (96.6 mg, 96%) as a white amorphous solid. Rf = 0.19 (hexane/EtOAc 7:3). [α]36520 = –160 (c = 1.79, CHCl3).
1H-NMR (600 MHz, CDCl3): 3.60 (br, H–C); 3.64 (brd, 3J(H,H) = 9.6, H–C); 3.71–3.84 (m, H–C(6), CH2(4)); 3.86 (s, Me); 3.93 (s,
Me); 4.02 (d, 2J(H,H) = 10.8, CH2(1)); 4.08 (d, 2J(H,H) = 10.8, CH2(1)); 4.47 (m, CH2(2)); 4.53 (d, 2J(H,H) = 12.0, CH2(1)); 4.56–
4.62 (m, CH2(5), H–C(1)); 4.69 (brs, H–C); 4.84 (d, 2J(H,H) = 4.2, CH2(1)); 4.86 (d, 2J(H,H) = 4.2, CH2(1)); 4.88–4.96 (m,
CH2(6)); 4.99 (d, 2J(H,H) = 11.4, CH2(1)); 5.04 (d, 2J(H,H) = 11.4, CH2(1)); 6.93 (m, H–C(4)); 7.08–7.18 (m, H–C(16)); 7.21–7.36
3
(m, H–C(28)); 7.51 (d, J(H,H) = 7.8, H–C(2)); 7.95 (s, H–C); 8.00 (brs, H–C(1)); 10.50 (s, OH). 13C-NMR (150 MHz, CDCl3,
several peaks are overlapped): 61.7; 61.9; 69.0; 69.1; 73.38; 73.40; 75.01; 75.07; 75.12; 75.21; 75.4; 75.5; 75.6; 78.4; 78.5;
79.3; 79.5; 82.9; 86.9; 87.1; 108.4; 118.9; 124.4; 127.57; 127.63; 127.66; 127.71; 127.77; 127.96; 127.99; 128.05; 128.09;
128.16; 128.19; 128.23; 128.26; 128.28; 128.33; 128.32; 128.39; 128.42; 128.44; 131.7; 136.4; 137.3; 137.4; 137.9; 138.08;
138.12; 138.14; 138.21; 137.70; 137.72; 139.8; 144.6; 144.9; 156.95; 157.05; 158.2; 165.4; 167.6. IR (neat): 3183, 3030, 2869,
1720 (C=O), 1691 (C=O), 1604, 1454, 1339, 1067, 1028, 736, 697. HR-ESI-MS: 1597.6265 ([M+Na]+, C98H94NaO19+; calc.
1597.6282).
Salicylide 29.
To a solution of hydroxy acid 37 (156 mg, 99.0 μmol), Et3N (30.5 μL, 220 μmol), and 2,3,4-trichlorobenzoyl chloride (19.0 μL,
122 μmol) in toluene (6.0 mL) was added 4-pyrrolidinopyridine (2.3 mg 15.5 μmol) at room temperature. After stirring for 1.0 h,
the reaction was quenched by adding saturated aqueous NaHCO3, and the mixture was extracted with EtOAc (×3). The
combined organic extracts were washed with brine, dried over anhydrous Na2SO4, filtered and concentrated in vacuo The
This article is protected by copyright. All rights reserved.

residue was purified by preparative TLC (silica gel, hexane/EtOAc = 90/10 69/31) to afford 29 (137 mg, 88%) as a white
20
1
amorphous solid. Rf (hexane/EtOAc 4:1). [α]D = +15.0 (c = 0.953, CHCl3). H-NMR (600 MHz, CDCl3, signals from the minor
3
diastereomer are marked with an asterisk): 3.25–3.40* (br); 3.35–3.64 (m); 3.72 (s, Me); 3.80 (t, J(H,H) = 8.6, H–C); 4.13 (d,
2
2
2J(H,H) = 11.1, CH2(1)); 4.19–4.16* (br); 4.33 (d, J(H,H) = 12.3, CH2(1)); 4.39 (d, J(H,H) = 12.3, CH2(1)); 4.56 (m, CH2(2));
4.62 (d, 3J(H,H) = 9.6, H–C); 4.69 (d, 2J(H,H) = 10.0, CH2(1)); 4.88–4.94 (m, CH2(3)); 5.03–5.09 (m, CH2); 6.60–6.75* (br); 6.93
3
(d, J(H,H) = 5.9, H–C); 7.05–7.46 (m, H–C(29)); 7.47–7.59 (m, H–C(2)). 13C-NMR (150 MHz, CDCl3, signals from the minor
diastereomer are marked with an asterisk*): 61.8; 68.7*; 69.0; 73.3; 73.9*; 75.05; 75.47; 75.8; 77.9*; 78.4; 78.9*; 79.6; 83.1;
86.8*; 87.0; 118.9*; 119.1; 123.9; 127.5–128.9 (Bn); 128.9; 133.1; 136.1; 137.17*; 137.4; 137.9; 138.0; 138.6; 144.1; 144.5;
156.5; 163.1*; 163.4. IR (neat) 3063, 3030, 2867, 1763 (C=O), 1600, 1454, 1325, 1160, 1093, 1067. HR-ESI-MS: 1579.6195
([M+Na]+, C98H92NaO18+; calc. 1579.6176).
Compound 1.
A two-necked flask, thoroughly purged with argon, was charged with Pd(OH)2 (17.9 mg), to which was added a solution of 29
(199 mg, 127 μmol) in THF (6.0 mL). The atmosphere was changed from argon to H2 (1 atm), and the mixture was stirred for 7 h
at room temperature. After changing the atmosphere from H2 back to argon, the mixture was filtered through a membrane filter
(HLC-DISK 13, 0.45 μm, Kanto Chemical Co., Inc.). The filtrate was concentrated in vacuo to afford the crude material of 1 (98.5
mg) as a white solid.
1H-NMR (600 MHz, CDCl3, signals from the minor diastereomer are marked with an
asterisk*): 3.14–3.16 (m, H–C(2)); 3.24–3.26 (m, H–C(2)); 3.38 (m, CH2(1)); 3.63 (m,
3
CH2(1); 3.72* (s, Me(0.6)); 3.74 (s, Me(2.4); 4.26* (d, J(H,H) = 9.7, H–C(0.2)); 4.29 (d,
3J(H,H) = 8.6, H–C(0.8)); 4.38–4.60 (br, OH); 4.35–5.10 (br, OH(3)); 6.98 (s, H–C(0.8));
6.99* (s, H–C(0.2)); 10.06* (s, OH(0.2)); 10.12 (s, OH(0.8)). 13C-NMR (150 MHz, CDCl3,
signals from the minor diastereomer are marked with an asterisk*): 61.21; 61.24*;
61.30*; 61.41; 70.3; 72.7*; 73.7; 74.6; 78.3; 78.5*; 81.7*; 81.9; 117.7; 118.3*; 118.8*; 119.1;
133.1*; 133.4; 137.7; 137.9*; 142.4; 142.7*; 151.4; 152.0*; 164.5. HR-ESI-MS: 655.1510
([M–H]–, C28H31O18–; calc. 655.1516).
Supplementary Material
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/xxxxxx.
Acknowledgements
This work was supported by JSPS Research Fellowship (No. 13F13703) for Eva-Maria
Tanzer and JSPS KAKENHI Grant Numbers JP23000006, JP16H06351, JP16H01137.
We thank Prof. Duk Sang Jung (Hanbat National University) for sending a copy of the
NMR spectra of the natural product.
This article is protected by copyright. All rights reserved.

References
[1]
[2]
E. G. De Mejía, E. V. Ramírez-Mares, ‘Ardisia: health-promoting properties and toxicity of phytochemicals and extracts’, Toxicol.
Mech. Method 2011, 21, 667–674.
a) S. Piacente, C. Pizza, N. De Tommasi, ‘Constituents of Ardisia japonica and Their in Vitro Anti-HIV Activity’, J. Nat. Prod. 1996,
59, 565–569; b) Y. Fukuyama, Y. Kiriyama, J. Okino, M. Kodama, H. Iwaki, S. Hosozawa, K. Matsui, ‘Naturally Occurring 5-
Lipoxygenase Inhibitor. II. Structures and Syntheses of Ardisianoses A and B, and Maesanin, Alkenyl-1, 4-benzoquinones from the
Rhizome of Ardisia japonica’, Chem. Pharm. Bull. 1993, 41, 561–565; c) H. Kobayashi, E. De Mejía, ‘The genus Ardisia: a novel source
of health-promoting compounds and phytopharmaceuticals’, J. Ethnopharmacol. 2005, 96, 347–354.
[3]
[4]
G. Ryu, S. Y. Lee, B. S. Kim, S. Y. Ryu, H. J. Hwang, B. W. Choi, B. H. Lee, D. S. Jung, ’Ardimerin, a New Dimeric Lactone from the
Herb of Ardisia japonica’, Nat. Prod. Sci. 2002, 8, 108–110.
a) N. T. Dat, K. Bae, A. Wamiru, J. B. McMahon, S. F. J. Le Grice, M. Bona, J. A. Beutler, Y. H. Kim, `A Dimeric Lactone from
Ardisia japonica with Inhibitory Activity for HIV-1 and HIV-2 Ribonuclease H’, J. Nat. Prod. 2007, 70, 839–841; b) I. P. Singh, H. S.
Bodiwala, ‘Recent advances in anti-HIV natural products’, Nat. Prod. Rep. 2010, 27, 1781–1800.
[5]
M. Mavlan, K. Ng, H. Panesar, A. Yepremyan, T. G. Minehan, ‘Synthesis of 3,3’-Di-O-methyl Ardimerin and Exploration of Its DNA
Binding Properties’, Org. Lett. 2014, 16, 2212–2215.
[6]
[7]
P. G. Edgerley, L. E. Sutton, ‘The Stereochemistry of the Salicylides and Some Related Compounds’, J. Chem. Soc. 1951, 1069–1074.
a) D. Seebach, T. Hoffmann, F. N. M. Kühnle, J. N. Kinkel, M. Schulte, `Preparation, Structure, and Properties of All Possible Cyclic
Dimers (Diolides) of 3-Hydroxybutanoic Acid’, Helv. Chim. Acta 1995, 78, 1525–1540; b) W. D. Ollis, J. S. Stephanatou, J. F.
Stoddart, ’Conformational behaviour of medium-sized rings. Part 9. Disalicylides and trisalicylides’, J. Chem. Soc., Perkin Trans. 1
1982, 1629–1636; c) W. D. Ollis and J. F. Stoddart, ‘Conformational behaviour of di-o-thymotide and di-o-carvocrotide’, J. Chem. Soc.,
Chem. Commun. 1973, 571–572.
[8]
[9]
a) J. M. Gnaim, B. S. Green, R. Arad-Yellin, P. M. Keehn, ‘Improved preparation of the clathrate host compound tri-o-thymotide and related
trisalicylide derivatives’, J. Org. Chem. 1991, 56, 4525–4529; b) T. D. Harris, S. R. Oruganti, L. M. Davis, P. M. Keehn, B. S. Green,
‘Substituent modification in tri-o-thymotide and its effects on host geometry and guest enclathration .1. synthesis’, Tetrahedron 1987, 43,
1519–1540.
a) K. Suzuki, ‘Total synthesis of aryl C-glycoside antibiotics’, Pure & Appl. Chem. 1994, 66, 2175–2178; b) K. Suzuki, `Lessons from total
synthesis of hybrid natural products’, Chem. Rec. 2010, 10, 291–307.
[10] M. D. Lewis, J. K. Cha, Y. Kishi, ‘Highly stereoselective approaches to α- and β-C-glycopyranosides’, J. Am. Chem. Soc. 1982, 104,
4976–4978.
[11] A. Yanagisawa, M. Taga, T. Atsumi, K. Nishimura, K. Ando, T. Taguchi, H. Tsumuki, I. Chujo, S. Mohri, ‘Process Development of the
PDE4 Inhibitor K-34’, Org. Process. Rev. Dev. 2011, 15, 376–381.
[12] J. W. Gathirwa, T. Maki, ‘Benzylation of hydroxy groups with tertiary amine as a base’, Tetrahedron 2012, 68, 370–375.
[13] A. Krasovskiy, P. Knochel, ‘A LiCl-Mediated Br/Mg Exchange Reaction for the Preparation of Functionalized Aryl- and
Heteroarylmagnesium Compounds from Organic Bromides’, Angew. Chem. Int. Ed. 2004, 43, 3333–3336.
[14] B. A. Ellsworth, A. G. Doyle, M. Patel, J. Caceres-Cortes, W. Meng, P. P. Deshpande, A. Pullockaran, W. N. Washburn, ‘C-
Arylglucoside synthesis: triisopropylsilane as a selective reagent for the reduction of an anomeric C-phenyl ketal’, Tetrahedron:
Asymmetry 2003, 14, 3243–3247.
[15] I. Shiina, M. Kubota, H. Oshiumi, M. Hashizume, ‘An Effective Use of Benzoic Anhydride and Its Derivatives for the Synthesis of
Carboxylic Esters and Lactones: A Powerful and Convenient Mixed Anhydride Method Promoted by Basic Catalysts’, J. Org. Chem.
2004, 69, 1822–1830.
[16] M. Usui, K. Saigo, T. Mukaiyama, ‘The Facile Synthesis of Lactones’, Chem. Lett. 1976, 49–50.
[17] J. Inanaga, K. Hirata, H. Saeki, T. Katsuki, M. Yamaguchi, ‘A Rapid Esterification by Means of Mixed Anhydride and Its Application
to Large-ring Lactonization’, Bull. Chem. Soc. Jpn. 1979, 52, 1989–1993.
[18] a) M. Hikota, Y. Sakurai, K. Horita, O. Yonemitsu, ‘Synthesis of erythronolide A via a very efficient macrolactonization under usual
acylation conditions with the Yamaguchi reagent’, Tetrahedron Lett. 1990, 31, 6367–6370; b) M. Hikota, H. Tone, K. Horita, O.
Yonemitsu, ‘Chiral synthesis of polyketide-derived natural products. 27. Stereoselective synthesis of erythronolide A via an
extremely efficient macrolactonization by the modified Yamaguchi method’, J. Org. Chem. 1990, 55, 7–9.
[19] W. Xing, L. Chen, ‘Micelle-mediated extraction and cloud point preconcentration of bergenin from Ardisia japonica’, Separation and
Purification Technology 2013, 110, 57–62.
This article is protected by copyright. All rights reserved.

[20] H. Pan, ‘A non-covalent dimer formed in electrospray ionisation mass spectrometry behaving as a precursor for fragmentations’,
Rapid Commun. Mass Spectrom. 2008, 22, 3555–3560.
[21] V. S. Parmar, R. Jain, S. Singh, ‘Synthesis of new highly substituted flavones, occurring in Ageratum houstonianum’, J. Chem. Research
(M), 1987, 2301–2309.
Entry for the Table of Contents
This article is protected by copyright. All rights reserved.