Synthesis of hydroxycinnamoyl β-d-xylopyranosides and evaluation of their antioxidant properties

Article abstract of DOI:10.1016/j.carres.2013.12.014

Various hydroxycinnamoyl β-d-xylopyranosides were efficiently prepared from 2,3,4-tri-O-acetyl-α-d-xylopyranosyl bromide (TAXB) with amine by amine-promoted glycosylation. The resulted acetylated hydroxycinnamoyl β-d-xylopyranosides with acetoxy groups at

Full text of DOI:10.1016/j.carres.2013.12.014

Carbohydrate Research 388 (2014) 138–146
Contents lists available at ScienceDirect
Carbohydrate Research
journal homepage: www.elsevier.com/locate/carres
Synthesis of hydroxycinnamoyl b-D-xylopyranosides and evaluation
of their antioxidant properties
a
b
b
b
Yasutaka Shimotori a, , Masayuki Hoshi , Kouji Soga , Yosuke Osawa , Tetsuo Miyakoshi
⇑
a Department of Biotechnology and Environmental Chemistry, Kitami Institute of Technology, 165 Koen-cho, Kitami, Hokkaido 090-8507, Japan
b Department of Applied Chemistry, School of Science and Technology, Meiji University, 1-1-1 Higashi-mita, Tama-ku, Kawasaki, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 16 October 2013
Received in revised form 13 December 2013
Accepted 15 December 2013
Available online 21 December 2013
Various hydroxycinnamoyl b-
D
-xylopyranosides were efficiently prepared from 2,3,4-tri-O-acetyl-a-D-
xylopyranosyl bromide (TAXB) with amine by amine-promoted glycosylation. The resulted acetylated
hydroxycinnamoyl b- -xylopyranosides with acetoxy groups at C-2, C-3, and C-4 were regioselectively
deacetylated at C-4 position with Novozym 435. Antioxidant activities of free hydroxycinnamic acids
and the respective b-
-xylopyranosides were evaluated by DPPHÅ radical scavenging activity as well as
D
D
their inhibitory effect on autoxidation of bulk methyl linoleate. The radical scavenging activity on 1,
Keywords:
Hydroxycinnamic acids
Hydroxycinnamoyl b-D-xylopyranosides
Lipase
Regioselective transesterification
Antioxidant properties
1-diphenyl-2-picrylhydrazyl (DPPHÅ) decreased in the order ferulic acid > caffeic acid ꢀ caffeoyl b-
D-xylo-
pyranosides ꢀ sinapinic acid > sinapoyl b-
D
-xylopyranosides ꢀ feruloyl b-
D-xylopyranosides > p-couma-
ric acid > p-coumaroyl b-
D-xylopyranosides. In bulk methyl linoleate, the antioxidant activity order
against autoxidation was almost consistent with the scavenging activity order. The results showed that
caffeoyl b-
D
-xylopyranosides and sinapoyl b-
D-xylopyranosides were as effective as free caffeic acid,
sinapinic acid, and ferulic acid.
Published by Elsevier Ltd.
1. Introduction
without heavy metal catalyst and Novozym 435-catalyzed regiose-
lective transesterification and the evaluation of antioxidant proper-
ties of these compounds.13 In this paper, we investigated the
Hydroxycinnamic acids such as p-coumaric, caffeic, ferulic, and
sinapinic acid are very common in fruits and vegetables in the form
of esters and glycosides. These compounds have various properties
which may be of importance for the remediation of many diseases.
Their activities, for example, anti-bacterial, anti-viral, anti-inflam-
matory, anti-carcinogenic, and anti-microbial agents, are related to
their capability to act as antioxidants, metal ion chelating, radical
scavenging, and inhibition of prooxidant enzymes.1–6 Various fruits,
including these hydroxycinnamic acids have been used as tradi-
tional medicine. For instance, Rubus species have been used as anti-
microbial, anticonvulsant, muscle relaxant, and radical scavenging
agents.7–10 Hussein et al. found that the extracts provided from
Rubus sanctus included various caffeoyl sugar esters and an
ellagitannin such as 1-O-caffeoyl xylose, 3,6-di-O-caffeoyl glucose,
2,3-O-hexahydroxydiphenoyl-4,6-O-sanguisorboy glucose.11 In
synthesis of hydroxycinnamoyl b-
the antioxidant activities of hydroxycinnamic acids such as ferulic
acid, caffeic acid, and sinapinic acid and its b- -xylopyranosides.
The corresponding hydroxycinnamoyl b- -xylopyranosides were
D-xylopyranosides, and described
D
D
synthesized with high yield even with xylosyl bromide instead of
glucosyl bromide. The availability of amine-promoted glycosylation
without heavy metal catalyst was able to be shown. Their antioxi-
dant activity was evaluated on the basis of their inhibitory effects
on the autoxidation of methyl linoleate in bulk system and the rad-
ical scavenging activity against 1,1-diphenyl-2-picrylhydrazyl
(DPPHÅ). In all tested compounds, caffeoyl b-
D-xylopyranosides
showed especially high antioxidant activity and long durability of
antioxidative potency compared with free caffeic acid. We report
the synthesis of hydroxycinnamoyl b-D-xylopyranosides and its
addition, 1-O-E-cinnamoyl-b-
b-D-rhamnopyranoside and other cinnamoyl glycosides were
D
-xylopyranoside, 1-O-E-cinnamoyl-
antioxidant properties because the importance of xylosides was
confirmed.
isolated from ripe fruits of the Chilean strawberry Fragaria chiloensis
ssp. Chiloensis.12 Cheel et al. investigated antioxidant capacity of
these compounds. We previously reported the synthesis of hydroxy-
2. Experimental
2.1. General
cinnamoyl b-D-glucopyranosides by amine-promoted glycosylation
Thin layer chromatography (TLC) was performed on precoated
F-254 plates (Merck Ltd, Darmstadt, Germany). Flash column
⇑
Corresponding author. Tel.: +81 (157)269387; fax: +81 (157)247719.
E-mail address: yasu@mail.kitami-it.ac.jp (Y. Shimotori).
http://dx.doi.org/10.1016/j.carres.2013.12.014
0008-6215/Published by Elsevier Ltd.

Y. Shimotori et al. / Carbohydrate Research 388 (2014) 138–146
139
chromatography was performed on silica gel FL60D (Fuji Silysia
Chemical Ltd, Aichi, Japan). Optical rotation was determined with
a JASCO P-1010 polarimeter (JASCO corp., Tokyo Japan). IR spec-
troscopy was performed on an FT-IR 460plus spectrometer (JASCO
Corp., Tokyo Japan). 1H and 13C NMR spectra were taken on a JNM-
ECA-500 spectrometer with tetramethylsilane (Me4Si) as the inter-
nal standard, and chemical shifts were recorded as d values. Mass
spectra were recorded on an AccuTof GCv 4G (JEOL, Tokyo, Japan)
mass spectrometer.
2.2.3. Caffeoyl 2,3,4-tri-O-acetyl-b-D-xylopyranoside (1c)
Yield: 45%; white solid; mp = 172–173 °C; Rf = 0.25 (hexane–
AcOEt, 1:1); ½a 2D5
ꢂ30.7 (c 1.0, THF); IR (KBr): 3427 (O–H), 2960
ꢁ
(C–H), 1750 (C@CAC@O), 1602 (Ar, C@C), 1516 (Ar, C@C), 1370
(CAOAC@O), 1226 (CAOAC@O), 1041 (CAOAC@O) cmꢂ1
;
1H
NMR (500 MHz, DMSO-d6): d = 2.06 (m, 9H, CH3 ꢃ 3 in OAc at C-
2, C-3, and C-4), 3.58 (dd, 1H, J = 12.0, 8.0 Hz, H-5a), 4.19 (dd, 1H,
J = 12.0, 5.2 Hz, H-5b), 5.03 (ddd, 1H, J = 8.6, 8.0, 5.2 Hz, H-4),
5.15 (dd, 1H, J = 9.2, 6.9 Hz, H-2), 5.27 (dd, 1H, J = 9.2, 8.6 Hz, H-
3), 5.84 (d, 1H, J = 6.9 Hz, H-1), 6.20 (d, 1H, J = 16.0 Hz,
AC(AO)CHACHA), 6.89 (d, 1H, J = 8.0 Hz, Ph), 6.99 (dd, 1H,
J = 8.0, 2.3 Hz, Ph), 7.08 (d, 1H, J = 2.3 Hz, Ph), 7.62 (d, 1H,
J = 16.0 Hz, AC(AO)CHACHA); 13C NMR (126 MHz, DMSO-d6):
d = 20.6 (ACH3 in OAc at C-2, C-3, and C-4), 62.8 (C-5), 68.4 (C-4),
69.7 (C-2), 71.0 (C-3), 92.2 (C-1), 113.6 (AC(AO)CHACHA), 114.4,
115.5, 123.0, 127.0, 144.0, 147.0 (Ph), 147.4 (AC(AO)CHACHA),
165.3 (AC(AO)CHACHA), 169.8 (ACAO in OAc at C-2), 170.0
(AC@O ꢃ 2 in OAc at C-3 and C-4); HRMS: m/z [M]+ calcd for
C20H22O11: 438.1162; found: 438.1175.
2.2. General procedure for glycosylation of hydroxycinnamic
acids
A
solution of 2,3,4-tri-O-acetyl-a-D-xylopyranosyl bromide
(TAXB) (678 mg, 2.0 mmol), corresponding hydroxycinnamic acids
(6.0 mmol) and i-Pr2NEt (258 mg, 2.0 mmol) in CH3CN (10 mL) was
stirred for 24 h at room temperature under Ar atmosphere, and 1 g
4 Å molecular sieves were used for water binding. Progress of the
reaction was monitored by thin layer chromatography. On the
completion of the reaction, the reaction mixture was filtered to re-
move molecular sieves, and the solvent was evaporated under vac-
uum, and AcOEt was added to the residue. The resulting solution
was then neutralized with aqueous NaHCO3, and extracted with
AcOEt, followed by washing with brine. The combined extracts
were dried over anhydrous MgSO4, and the solvent was removed
by evaporating under vacuum. The crude product was purified by
flash chromatography (hexane–AcOEt, 6:1 for 1a, 1:1 for 1b and
1c, 3:1 for 1d, 2:1 for 1e) to give 1a–e.
2.2.4. Feruloyl 2,3,4-tri-O-acetyl-b-D-xylopyranoside (1d)
Yield: 73%; white solid; mp=141–142 °C; Rf = 0.28 (hexane–
AcOEt, 1:1); ½a 2D5
ꢂ26.4 (c 1.0, THF); IR (KBr): 3510 (O–H), 2975
ꢁ
(C–H), 1746 (C@CAC@O), 1595 (Ar, C@C), 1519 (Ar, C@C), 1372
(CAOAC@O), 1227 (CAOAC@O), 1082 (CAOAC@O) cmꢂ1
;
1H
NMR (500 MHz, CDCl3): d = 2.06 (m, 9H, CH3 ꢃ 3 in OAc at C-2, C-
3, and C-4), 3.57 (dd, 1H, J = 12.4, 9.2 Hz, H-5a), 3.94 (s, 3H, CH3
in OMe at Ph), 4.19 (dd, 1H, J = 12.4, 5.0 Hz, H-5b), 5.03 (ddd, 1H,
J = 9.2, 8.7, 5.0 Hz, H-4), 5.16 (dd, 1H, J = 9.2, 6.9 Hz, H-2), 5.27
(dd, 1H, J = 9.2, 8.7 Hz, H-3), 5.84 (d, 1H, J = 6.9 Hz, H-1), 6.09 (br
s, 1H, OH in Ph), 6.26 (d, 1H, J = 15.5 Hz, AC(@O)CH@CHA), 6.92
(d, 1H, J = 8.0 Hz, Ph), 7.04 (d, 1H, J = 2.3 Hz, Ph), 7.08 (dd, 1H,
J = 8.3, 2.0 Hz, Ph), 7.68 (d, 1H, J = 15.5 Hz, AC(@O)CH@CHA); 13C
NMR (126 MHz, CDCl3): d = 20.6 (ACH3 in OAc at C-2, C-3, and C-
4), 56.0 (CH3 in OMe at Ph), 62.9 (C-5), 68.5 (C-4), 69.7 (C-2),
71.3 (C-3), 92.2 (C-1), 109.4 (Ph), 113.5 (AC(@O)CH@CHA), 114.8,
123.7, 126.4, 146.8 (Ph), 147.3 (AC(@O)CH@CHA), 148.6 (Ph),
165.1 (AC(@O)CH@CHA), 169.5 (AC@O in OAc at C-2), 169.8
(AC@O ꢃ 2 in OAc at C-3 and C-4); HRMS: m/z [M]+ calcd for
2.2.1. Cinnamoyl 2,3,4-tri-O-acetyl-b-D-xylopyranoside (1a)
Yield: 70%; white solid; mp = 161–162 °C; Rf = 0.65 (hexane–
AcOEt, 1:1); ½a 2D5
ꢂ51.9 (c 1.0, THF); IR (KBr): 2950 (C–H),
ꢁ
1752 (C@CAC@O), 1638 (Ar, C@C), 1380 (CAOAC@O), 1230
(CAOAC@O), 1087 (CAOAC@O) cmꢂ1 1H NMR (500 MHz, CDCl3):
;
d = 2.06 (m, 9H, CH3 ꢃ 3 in OAc at C-2, C-3, and C-4), 3.59 (dd,
1H, J = 12.0, 8.7 Hz, H-5a), 4.20 (dd, 1H, J = 12.0, 5.0 Hz, H-5b),
5.03 (ddd, 1H, J = 8.7, 8.2, 5.0 Hz, H-4), 5.16 (dd, 1H, J = 8.7,
6.9 Hz, H-2), 5.27 (dd, 1H, J = 8.7, 8.2 Hz, H-3), 5.86 (d, 1H,
J = 6.9 Hz, H-1), 6.42 (d, 1H, J = 16.0 Hz, AC(@O)CH@CHA), 7.41
(m, 3H, Ph), 7.54 (m, 2H, Ph), 7.76 (d, 1H, J = 16.0 Hz,
AC(@O)CH@CHA); 13C NMR (126 MHz, CDCl3): d = 20.7 (–CH3
in OAc at C-2, C-3, and C-4), 62.8 (C-5), 68.4 (C-4), 69.5 (C-2),
71.0 (C-3), 92.2 (C-1), 116.5 (AC(@O)CH@CHA), 128.3, 129.0,
C21H24O11: 452.1319; found: 452.1347.
2.2.5. Sinapoyl 2,3,4-tri-O-acetyl-b- -xylopyranoside (1e)
D
Yield: 70%; white solid; mp = 138–139 °C; Rf = 0.20 (hexane–
130.9,
133.9
(Ph),
147.0
(AC(@O)CH@CHA),
164.7
AcOEt, 1:1); ½a 2D5
ꢂ26.7 (c 1.0, THF); IR (KBr): 3461 (O–H), 2947
ꢁ
(AC(@O)CH@CHA), 169.3 (AC@O in OAc at C-2), 169.8
(C–H), 1758 (C@CAC@O), 1603 (Ar, C@C), 1517 (Ar, C@C), 1370
(AC@O ꢃ 2 in OAc at C-3 and C-4); HRMS: m/z [M]+ calcd for
(CAOAC@O), 1221 (CAOAC@O), 1045 (CAOAC@O) cmꢂ1
;
1H
C20H22O9: 406.1264; found: 406.1281.
NMR (500 MHz, CDCl3): d = 2.07 (m, 9H, CH3 ꢃ 3 in OAc at C-2, C-
3, and C-4), 3.56 (dd, 1H, J = 12.4, 9.2 Hz, H-5a), 3.93 (s, 6H,
CH3 ꢃ 2 in OMe at Ph), 4.19 (dd, 1H, J = 12.4, 5.0 Hz, H-5b), 5.04
(ddd, 1H, J = 9.2, 8.7, 5.0 Hz, H-4), 5.17 (dd, 1H, J = 9.2, 7.3 Hz, H-
2), 5.28 (dd, 1H, J = 9.2, 8.7 Hz, H-3), 5.84 (d, 1H, J = 7.3 Hz, H-1),
5.92 (br s, 1H, OH in Ph), 6.28 (d, 1H, J = 16.0 Hz, AC(@O)CH@CHA),
6.79 (s, 2H, Ph), 7.66 (d, 1H, J = 15.5 Hz, AC(@O)CH@CHA); 13C
NMR (126 MHz, CDCl3): d = 20.6 (ACH3 in OAc at C-2, C-3, and C-
4), 56.3 (CH3 ꢃ 2 in OMe at Ph), 63.0 (C-5), 68.5 (C-4), 69.8 (C-2),
71.4 (C-3), 92.3 (C-1), 105.4 (Ph), 113.9 (AC(@O)CH@CHA), 125.3,
2.2.2. p-Coumaroyl 2,3,4-tri-O-acetyl-b-D-xylopyranoside (1b)
Yield: 67%; white solid; mp = 164–165 °C; Rf = 0.34 (hexane–
AcOEt, 1:1); ½a 2D5
ꢂ36.7 (c 1.0, THF); IR (KBr): 3405 (O–H), 3006
ꢁ
(Ar, C–H), 1752 (C@CAC@O), 1607 (Ar, C@C), 1517 (Ar, C@C),
1381 (CAOAC@O), 1239 (CAOAC@O), 1156 (CAOAC@O) cmꢂ1
;
1H NMR (500 MHz, DMSO-d6): d = 2.06 (m, 9H, CH3 ꢃ 3 in OAc at
C-2, C-3, and C-4), 3.58 (dd, 1H, J = 12.6, 8.6 Hz, H-5a), 4.19 (dd,
1H, J = 12.6, 5.2 Hz, H-5b), 5.04 (ddd, 1H, J = 9.2, 8.6, 5.2 Hz, H-4),
5.15 (dd, 1H, J = 8.0, 6.9 Hz, H-2), 5.25 (dd, 1H, J = 9.2, 8.0 Hz, H-
3), 5.84 (d, 1H, J = 6.9 Hz, H-1), 6.25 (d, 1H, J = 16.0 Hz,
AC(@O)CH@CHA), 6.86 (m, 2H, Ph), 7.42 (m, 2H, Ph), 7.69 (d, 1H,
J = 16.0 Hz, AC(@O)CH@CHA); 13C NMR (126 MHz, DMSO-d6):
d = 20.7 (ACH3 in OAc at C-2, C-3, and C-4), 62.8 (C-5), 68.4 (C-4),
69.7 (C-2), 71.1 (C-3), 92.1 (C-1), 113.3 (AC(@O)CH@CHA), 116.0,
126.4, 130.4 (Ph), 147.1 (AC(@O)CH@CHA), 158.7 (Ph), 165.3
(AC(@O)CH@CHA), 169.7 (AC@O in OAc at C-2), 170.0 (AC@O ꢃ 2
137.7,
147.2
(Ph),
147.5
(AC(@O)CH@CHA),
164.9
(AC(@O)CH@CHA), 169.5 (AC@O in OAc at C-2), 169.8 (AC@O ꢃ 2
in OAc at C-3 and C-4); HRMS: m/z [M]+ calcd for C22H26O12
:
482.1424; found: 482.1397.
2.3. General procedure for regioselective transesterification
with Novozym 435
in OAc at C-3 and C-4); HRMS: m/z [M]+ calcd for C20H22O10
:
A solution of the corresponding hydroxycinnamoyl 2,3,4-tri-O-
acetyl-b- -xylopyranosides (1a–e) (1.0 mmol), MeOH (384 mg,
422.1213; found: 422.1204.
D

140
Y. Shimotori et al. / Carbohydrate Research 388 (2014) 138–146
12.0 mmol), and Novozym 435 (0.8 g) in tert-butyl methyl ether
(MTBE) (40 mL) was stirred for 24 h at 50 °C. The progress of the
reaction was monitored with thin layer chromatography. Upon
completion of the reaction, Novozym 435 was filtered off the reac-
tion mixture, and the solvent was then removed under vacuum.
The residue was purified with flash chromatography (CHCl3–
MeOH, 100:1 for 2a, 20:1 for 2d and 2e) or by recrystallized from
CHCl3 for 2b, 2c to give 2a–e.
169.1 (AC@O in OAc at C-2), 169.5 (AC@O in OAc at C-3); HRMS:
m/z [M]+ calcd for C18H20O10: 396.1057; found: 396.1057.
2.3.4. Feruloyl 2,3-di-O-acetyl-b-D-xylopyranoside (2d)
Yield: 89%; white solid; mp = 74–75 °C; Rf = 0.22 (CHCl3–MeOH,
20:1); ½a 2D5
ꢂ60.7 (c 1.0, THF); IR (KBr): 3442 (O–H), 2941 (C–H),
ꢁ
1738 (C@CAC@O), 1592 (Ar, C@C), 1515 (Ar, C@C), 1376
(CAOAC@O), 1243 (CAOAC@O), 1080 (CAOAC@O) cmꢂ1
;
1H
NMR (500 MHz, DMSO-d6): d = 1.97 (s, 3H, CH3 in OAc at C-2),
2.03 (s, 3H, CH3 in OAc at C-3), 3.50 (dd, 1H, J = 12.0, 10.9 Hz, H-
5a), 3.73 (m, 1H, H-4), 3.83 (s, 3H, CH3 in OMe at Ph), 3.91 (dd,
1H, J = 12.0, 5.7 Hz, H-5b), 4.91 (dd, 1H, J = 8.0, 9.7 Hz, H-2), 5.08
(dd, 1H, J = 10.3, 9.7 Hz, H-3), 5.56 (br s, 1H, OH at C-4), 5.83 (d,
1H, J = 7.4 Hz, H-1), 6.47 (d, 1H, J = 16.0 Hz, AC(@O)CH@CHA),
6.82 (d, 1H, J = 8.0 Hz, Ph), 7.13 (dd, 1H, J = 8.6, 1.7 Hz, Ph), 7.36
(d, 1H, J = 1.7 Hz, Ph), 7.61 (d, 1H, J = 15.5 Hz, AC(@O)CH@CHA),
9.72 (br s, 1H, OH in Ph); 13C NMR (126 MHz, DMSO-d6): d = 20.4,
20.7 (ACH3 ꢃ 2 in OAc at C-2 and C-3), 55.7 (CH3 in OMe at Ph),
65.9 (C-5), 67.0 (C-4), 70.6 (C-2), 74.7 (C-3), 92.1 (C-1), 111.3
(Ph), 112.8 (AC(@O)CH@CHA), 115.5, 124.0, 125.3 (Ph), 147.3
(AC(@O)CH@CHA), 148.0, 149.9 (Ph), 164.9 (AC(@O)CH@CHA),
169.3 (AC@O in OAc at C-2), 169.7 (AC@O in OAc at C-3); HRMS:
m/z [M]+ calcd for C19H22O10: 410.1213; found: 410.1258.
2.3.1. Cinnamoyl 2,3-di-O-acetyl-b-D-xylopyranoside (2a)
Yield: 85%; white solid; mp = 183–184 °C; Rf = 0.25 (CHCl3–
MeOH, 20:1); ½a 2D5
ꢂ66.6 (c 1.0, THF); IR (KBr): 3416 (O–H), 2937
ꢁ
(C–H), 1747 (C@CAC@O), 1637 (Ar, C@C), 1381 (CAOAC@O),
1240 (CAOAC@O), 1164 (O–H), 1078 (CAOAC@O) cmꢂ1 1H NMR
;
(500 MHz, DMSO-d6): d = 1.97 (s, 3H, CH3 in OAc at C-2), 2.02 (s,
3H, CH3 in OAc at C-3), 3.51 (dd, 1H, J = 11.5, 10.3 Hz, H-5a), 3.74
(m, 1H, H-4), 3.90 (dd, 1H, J = 11.5, 5.7 Hz, H-5b), 4.90 (dd, 1H,
J = 9.2, 8.0 Hz, H-2), 5.08 (dd, 1H, J = 10.3, 9.2 Hz, H-3), 5.55 (d,
1H, J = 3.4 Hz, OH at C-4), 5.83 (d, 1H, J = 8.0 Hz, H-1), 6.63 (d,
1H, J = 16.0 Hz, AC(@O)CH@CHA), 7.45 (m, 3H, Ph), 7.70 (d, 1H,
J = 16.0 Hz, AC(@O)CH@CHA), 7.74 (m, 2H, Ph); 13C NMR
(126 MHz, DMSO-d6): d = 20.3, 20.7 (ACH3 ꢃ 2 in OAc at C-2 and
C-3), 65.9 (C-5), 66.9 (C-4), 70.4 (C-2), 74.5 (C-3), 92.3 (C-1),
116.6 (AC(@O)CH@CHA), 128.7, 129.0, 131.0, 133.6 (Ph), 146.6
(AC(@O)CH@CHA), 164.4 (AC(@O)CH@CHA), 169.2 (AC@O in
OAc at C-2), 169.6 (AC@O in OAc at C-3); HRMS: m/z [M]+ calcd
for C18H20O8: 364.1158; found: 364.1170.
2.3.5. Sinapoyl 2,3-di-O-acetyl-b-D-xylopyranoside (2e)
Yield: 83%; pale yellow solid; mp = 85–86 °C; Rf = 0.18 (CHCl3–
MeOH, 20:1); ½a 2D5
ꢂ52.1 (c 1.0, THF); IR (KBr): 3455 (O–H), 2944
ꢁ
(C–H), 1731 (C@CAC@O), 1605 (Ar, C@C), 1516 (Ar, C@C), 1376
2.3.2. p-Coumaroyl 2,3-di-O-acetyl-b-
D
-xylopyranoside (2b)
(CAOAC@O), 1225 (CAOAC@O), 1081 (CAOAC@O) cmꢂ1 1H
;
Yield: 86%; white solid; mp = 194–195 °C; Rf = 0.05 (CHCl3–
NMR (500 MHz, DMSO-d6): d = 1.97 (s, 3H, CH3 in OAc at C-2),
2.02 (s, 3H, CH3 in OAc at C-3), 3.50 (t, 1H, J = 10.6 Hz, H-5a),
3.73 (br s, 1H, H-4), 3.81 (s, 6H, CH3 ꢃ 2 in OMe at Ph), 3.90 (q,
1H, J = 5.7 Hz, H-5b), 4.89 (dd, 1H, J = 9.7, 8.0 Hz, H-2), 5.07 (t,
1H, J = 9.5 Hz, H-3), 5.55 (br s, 1H, OH at C-4), 5.84 (d, 1H,
J = 8.0 Hz, H-1), 6.53 (d, 1H, J = 15.5 Hz, AC(@O)CH@CHA), 7.06
(s, 2H, Ph), 7.62 (d, 1H, J = 15.5 Hz, AC(@O)CH@CHA), 9.07 (br s,
1H, OH in Ph); 13C NMR (126 MHz, DMSO-d6): d = 20.4, 20.7
(ACH3 ꢃ 2 in OAc at C-2 and C-3), 56.1 (CH3 ꢃ 2 in OMe at Ph),
65.9 (C-5), 67.0 (C-4), 70.6 (C-2), 74.8 (C-3), 92.0 (C-1), 106.7
(Ph), 113.2 (AC(@O)CH@CHA), 124.0, 138.9 (Ph), 147.7
(AC(@O)CH@CHA), 148.0 (Ph), 164.9 (AC(@O)CH@CHA), 169.3
(AC@O in OAc at C-2), 169.7 (AC@O in OAc at C-3); HRMS: m/z
[M]+ calcd for C20H24O11: 440.1319; found: 440.1344.
MeOH, 20:1); ½a 2D5
ꢂ74.9 (c 1.0, THF); IR (KBr): 3410 (O–H), 3007
ꢁ
(Ar, C–H), 1732 (C@CAC@O), 1607 (Ar, C@C), 1517 (Ar, C@C),
1384 (CAOAC@O), 1263 (CAOAC@O), 1164 (O–H), 1080
(CAOAC@O) cmꢂ1 1H NMR (500 MHz, DMSO-d6): d = 1.96 (s, 3H,
;
CH3 in OAc at C-2), 2.02 (s, 3H, CH3 in OAc at C-3), 3.48 (dd, 1H,
J = 11.5, 10.3 Hz, H-5a), 3.71 (ddd, 1H, J = 10.3, 9.7, 5.7 Hz, H-4),
3.88 (dd, 1H, J = 11.5, 5.7 Hz, H-5b), 4.88 (dd, 1H, J = 9.2, 8.0 Hz,
H-2), 5.06 (dd, 1H, J = 9.7, 9.2 Hz, H-3), 5.54 (br s, 1H, OH at C-4),
5.79 (d, 1H, J = 8.0 Hz, H-1), 6.36 (d, 1H, J = 16.0 Hz,
AC(@O)CH@CHA), 6.80 (d, 2H, J = 8.6 Hz, Ph), 7.57 (d, 2H,
J = 8.6 Hz, Ph), 7.60 (d, 1H, J = 17.2 Hz, AC(@O)CH@CHA), 10.12
(br s, 1H, OH in Ph); 13C NMR (126 MHz, DMSO-d6): d = 20.3, 20.7
(ACH3 ꢃ 2 in OAc at C-2 and C-3), 65.9 (C-5), 66.9 (C-4), 70.5 (C-
2), 74.6 (C-3), 92.1 (C-1), 112.5 (AC(@O)CH@CHA), 115.8, 124.7,
130.8 (Ph), 146.9 (AC(@O)CH@CHA), 160.4 (Ph), 164.8
(AC(@O)CH@CHA), 169.2 (AC@O in OAc at C-2), 169.6 (AC@O in
OAc at C-3); HRMS: m/z [M]+ calcd for C18H20O9: 380.1107; found:
380.1125.
2.4. Evaluation of scavenging activity for DPPHÅ radicals
The antioxidant activity of hydroxycinnamoyl b-D-xylopyrano-
sides was estimated by measuring their free radical scavenging
activity using 2,2-diphenyl-1-picrylhydrazyl (DPPHÅ) as free radical
according to modification of the method of Silva et al.14 The reaction
mixture (10 mL) comprised of freshly made 0.15 mM DPPHÅ in eth-
2.3.3. Caffeoyl 2,3-di-O-acetyl-b-D-xylopyranoside (2c)
Yield: 83%; white solid; mp = 178–179 °C; Rf = 0.14 (CHCl3–
MeOH, 20:1); ½a 2D5
ꢁ
ꢂ77.6 (c 1.0, THF); IR (KBr): 3512 (O–H), 2949
anol (7000
lL), different concentrations of each hydroxycinnamoyl
(C–H), 1732 (C@CAC@O), 1609 (Ar, C@C), 1516 (Ar, C@C), 1381
b-D-xylopyranosides (1, 5, and 10
l
mol) in 300 L DMSO, and Tris–
l
(CAOAC@O), 1257 (CAOAC@O), 1035 (CAOAC@O) cmꢂ1
;
1H
HCl buffer (pH 7.4, 100 mM). The reaction mixture was kept for
30 min in a water bath at 25 °C under dark and optical density
was measured at 517 nm. DPPHÅ has an unpaired electron, which
gave purple color, and when this electron is balanced, the color is
lost. The compounds which can give an electron to the DPPHÅ can
bleached as the decrease in absorbance of the DPPHÅ solution with-
out test compounds. All analyses were carried out in triplicate.
NMR (500 MHz, DMSO-d6): d = 1.95 (s, 3H, CH3 in OAc at C-2),
2.01 (s, 3H, CH3 in OAc at C-3), 3.48 (dd, 1H, J = 11.5, 9.7 Hz, H-
5a), 3.71 (ddd, 1H, J = 9.7, 9.2, 5.2 Hz, H-4), 3.90 (dd, 1H, J = 11.5,
5.2 Hz, H-5b), 4.87 (dd, 1H, J = 9.7, 8.0 Hz, H-2), 5.05 (dd, 1H,
J = 9.7, 9.2 Hz, H-3), 5.77 (d, 1H, J = 8.0 Hz, H-1), 6.21 (d, 1H,
J = 16.0 Hz, AC(@O)CH@CHA), 6.78 (d, 1H, J = 8.0 Hz, Ph), 7.02
(dd, 1H, J = 8.0, 2.3 Hz, Ph), 7.05 (d, 1H, J = 2.3 Hz, Ph), 7.51 (d,
1H, J = 15.5 Hz, AC(@O)CH@CHA), 9.30 (br s, 1H, OH in Ph); 13C
NMR (126 MHz, DMSO-d6): d = 20.2, 20.5 (ACH3 ꢃ 2 in OAc at C-
2 and C-3), 65.8 (C-5), 66.9 (C-4), 70.4 (C-2), 74.5 (C-3), 92.1 (C-
1), 112.3 (AC(@O)CH@CHA), 115.0, 115.7, 121.7, 125.1 (Ph),
147.1 (AC(@O)CH@CHA), 148.9 (Ph), 164.5 (AC(@O)CH@CHA),
2.5. Evaluation of the inhibitory effect on autoxidation of
methyl linoleate
To 1 g of methyl linoleate, 25
lL of the acetone solution of the
test compound was mixed in a 50 mL vial, and the mixture was

Y. Shimotori et al. / Carbohydrate Research 388 (2014) 138–146
141
Table 1
agitated under ultrasonic wave for 30 s. A 25 lL aliquot of acetone
without sample was added for control. After purging the acetone
with nitrogen, the mixture was place in an oven at 40 °C in dark.
Solvent effect on glycosylation of xylosyl bromide with ferulic acida
Entry
Solvent
Donor numberb
Acceptor numberb
Yield (%)
Final concentration of each sample was 0.05 lmol/g oil. An aliquot
1
2
3
4
THF
20.0
17.0
29.8
14.1
8.0
12.5
19.3
18.9
38
44
50
53
of oil sample was dissolved in ethanol, and the conjugated diene
absorbance was measured at 234 nm with an UVmini-2400 UV–
vis spectrophotometer every 24 h at 20 °C. All tests were run in
triplicate.
Acetone
DMSO
MeCN
a
Xylosyl bromide: 2.0 mmol, ferulic acid: 4.0 mmol, i-Pr2NEt: 6.0 mmol, MS 4 Å:
1 g, solvent: 10 mL, under N2, rt, 24 h.
b
Ref. 15.
3. Results and discussion
3.1. Effect of solvent for glycosylation of ferulic acid
34%, respectively, in the case of using DBU and Ph3N (entries 7
and 8). pKa in acetonitrile of Et3N and i-Pr2NEt are 18.8 and 18.1,
respectively, there is no great difference in basicity.19,20 However,
because i-Pr2NEt is bulky structure compared with Et3N, nucleo-
philicity of Et3N is higher. Nucleophilic attack of carboxylate anion
of ferulic acid potentially competed against that of Et3N. It was as-
sumed that this caused the somewhat decrease of yield of 1d. The
yield was the lowest 14% in the case of using Ph3N (entry 7). pKa of
Ph3N was the lowest in the amine used here.21 When three equal
amounts of Ph3N for TAXB were used, it did not completely dis-
solve in acetonitrile. It was suggested that proton drawing of the
carboxyl group in ferulic acid is difficult for low basicity of Ph3N,
and the decrease of the yield resulted. DBU that had the highest ba-
sicity was used, and the yield was somewhat lower compared with
i-Pr2NEt and Et3N (entry 8). It was known that DBU caused dehy-
drogenation to give alkene.22–24 It was assumed that because the
production of carboxylate anion by dehydrogenation of ferulic acid
and the production of alkene occurred competitively, the decrease
of the yield was caused. From the above results, i-Pr2NEt is the best
as the amine. The optimization of the amine additive amount was
investigated. The yield of 1d was 53% when i-Pr2NEt was added
three equal amounts for TAXB (entry 4). However, the yield was
decreased and 36% in the case of a 0.5 equal amount (entry 1). i-
Pr2NEt improve the reactivity of ferulic acid by drawing a proton
of the carboxylic acid to the carboxylate anion. Additionally, it cap-
tures HBr produced from TAXB into i-Pr2NEtꢄHBr salt. However, the
added amount of i-Pr2NEt is less; it did not completely capture HBr,
and the reaction solution becomes acidic. Dissociation constant of
ferulic acid is higher than that of HBr. Therefore, the yield is re-
duced because the formation of carboxylate anion was inhibited,
and the reactivity of ferulic acid is reduced to lead the low yield.
Two and five equal amounts of i-Pr2NEt caused 52% and 47% yield,
respectively (entries 3 and 5). These yields were almost the same
with the use of three equal amounts regardless of added amounts.
However, the maximum yield of 65% was shown in the case of one
equal amount (entry 2). These results showed that the optimum
amounts of i-Pr2NEt was one equal amount for TAXB.
We had previously reported on the synthesis of hydroxycinna-
moyl b-
feruloyl 2,3,4,6-tetra-O-acetyl-b-
fully obtained with 74% yield.13 Therefore, synthesis of feruloyl
2,3,4-tri-O-acetyl-b- -xylopyranoside (1d) was attempted under
D-glucopyranosides by amine-promoted glycosylation and
D-glucopyranoside was success-
D
the same conditions (Scheme 1). However, the yield was 53% and
it was 20% lower compared with the case of feruloyl glucoside. This
result showed that optimum reaction conditions are different
depending on the type of sugar. Effect of solvent was confirmed
by using various types of solvent (Table 1). In the case of using ace-
tone and THF, the yields were 44% and 38%, respectively. These re-
sults was about 10% lower compared with those of DMSO and
CH3CN, which were 50% and 53%, respectively. The donor number
of DMSO was highest 29.8. However, there was no great difference
of donor numbers between other solvents.15 The acceptor numbers
of THF and acetone were smaller compared with those of DMSO
and CH3CN. Generally, Lewis acid as a catalyst such as Ag2CO3,
Hg(CN)2 and Cs2CO3 was used in glycosylation by Koenigs–Knorr
method.16–18 Because solvents that had large acceptor number acts
Lewis acid, reactivity of TAXB was improved by bonding between
solvent and Br in TAXB. On the other hand, solvents with large do-
nor number was bonded with hydrogen atom in carboxyl group,
and promoted the production of carboxylate anion. It was sug-
gested that the reactivity of carboxylate anion was decreased, be-
cause solvents had large acceptor number solvated with
carboxylate anion and stabilized. Large donor number did not re-
duce its reactivity by electrostatic repulsion between carboxylate
anion. Therefore, the yields of DMSO and CH3CN were higher than
those of THF and acetone.
3.2. Optimization of amine and amount added
The equal amount and the type of amines were investigated
(Table 2). When i-Pr2NEt was used as an amine, 1d was obtained
at the highest 53% yield compared with other types of amines (en-
try 4). There was no great difference in the case of using Et3N (en-
try 6). In contrast, the yields was decreased, and showed 14% and
Scheme 1. Glycosylation of hydroxycinnamic acids and regioselective transesterification using Novozym 435.

142
Y. Shimotori et al. / Carbohydrate Research 388 (2014) 138–146
Table 2
p-coumaric acid, the yield was increased compared with i-Pr2NEt
when Et3N was used (entries 2 and 3). On the other hand, when
caffeic acid was used, there was no difference in the yield even
using Et3N compared with i-Pr2NEt (entries 4 and 5). Basicity of
Et3N is slightly lower than that of i-Pr2NEt. i-Pr2NEt was easy to
salt formation with phenolic hydroxyl that had high acidity, and
salt formation of Et3N was difficult. The yield was increased by
these reasons (entries 2 and 3). Acidity of phenolic hydroxy group
in caffeic acid was higher than that of p-coumaric acid because caf-
feic acid had hydroxyl group at ortho position, the difference in the
yield did not occur to form salt regardless of the type of amines. All
1 were obtained in about 70% yields except caffeic acid.
Effect of amine and its amount on glycosylation of ferulic acida
Entry
Amine/amountb (equiv)
pKa (MeCN)
Yield (%)
1
2
3
4
5
6
7
8
i-Pr2NEt/0.5
i-Pr2NEt/1.0
i-Pr2NEt/2.0
i-Pr2NEt/3.0
i-Pr2NEt/5.0
Et3N/3.0
18.1c
36
65
52
53
47
49
14
34
18.8d
1.3e
Ph3N/3.0
DBU/3.0
24.3d
a
Xylosyl bromide: 2.0 mmol, ferulic acid: 4.0 mmol, MS 4 Å: 1 g, MeCN: 10 mL,
under N2, rt, 24 h.
b
Amine amount was based on xylosyl bromide.
Ref. 19.
Ref. 20.
Ref. 21.
c
3.4. Novozym 435-catalyzed regioselective transesterification of
1
d
e
It was known that high polar antioxidants exhibit high antiox-
idative activity for autoxidation of bulk oil such as methyl linole-
ate.26,27 It was seemed that deprotection of acetyl group at C-2,
C-3, and C-4 positions without transesterification of aglycon at C-
1 position caused the increase of molecular polarity, and antioxi-
dant activities in aqueous and lipophilic medium might been
developed. The reaction using base such as NaOH and NaOMe
hydrolyzed aglycon at C-1 position because this position had high
reactivity. It was reported that acetyl xylan esterase catalyzed the
deacetylation in sugar moieties.28–30 Although acetyl xylan ester-
ase was useful enzyme, the purchase of it was difficult. In constant,
lipase is a same category as hydrolase with acetyl xylan esterase,
and it can be purchased comparatively cheaply. Novozym 435
had wide substrate specificity. Therefore, Novozym 435-catalyzed
transesterification was investigated for the deprotection of acetyl
group regioselectively. Additionally, the reactivity and selectivity
3.3. Optimization of ferulic acid amount added and
glycosylation using other hydroxycinnamic acids
It could be performed to improve the yield of 1d by optimizing
equal amounts of i-Pr2NEt. Optimization of equal amounts of feru-
lic acid added was investigated to increase the yield of 1d (Table 3).
1d was obtained at 65% yield when ferulic acid of two equal
amounts to TAXB added (entry 6). The yields were 73% and 69%,
respectively, in the case of the addition of three and four equal
amounts. There was no great difference in the yield of 1d com-
pared with the addition of two equal amounts (entries 7 and 8).
However, ferulic acid of three equal amounts gave the maximum
yield, and these conditions were most suitable. Therefore, glycosyl-
ation was subjected using cinnamic acid, p-coumaric acid, caffeic
acid, and sinapinic acid in addition to ferulic acid. All hydroxycin-
namic acid derivatives except p-coumaric acid and caffeic acid gave
1 with about 70% yields (entries 1, 7, and 9). Galland et al. reported
that glycosylation of hydroxycinnamic acid derivatives that were
of Novozym 435 for hydroxycinnamoyl b-
was compared with that of hydroxycinnamoyl b-
sides. We previously reported lipase-catalyzed regioselective
D-xyropyranosides (1)
D-glucopyrano-
hydrolysis of feruloyl 2,3,4,6-tetra-O-acetyl-b-D-glucopyrano-
methylated at carboxyl group with 2,3,4,6-tetra-O-acetyl-b-
copyranosyl bromide was investigated.25 Methyl esters of 4-
(20,30,40,60-tetra-O-acetyl)-b-
-glucopyranosyl coumarate, ferulate,
D-glu-
side.13 Regioselective deacetylation at C-4 and C-6 position was
successfully achieved by Novozym 435 without transesterificating
ferulic acid ester at C-1 position. Therefore, deacetylation of
D
and caffeate were obtained at 75%, 65%, and 30%, respectively.
These results indicated that the competition of carboxylate anion
and phenoxy anion caused the yield to decrease when p-coumaric
acid and caffeic acid reacted with TAXB. It was suggested that the
yield of 1 in caffeic acid was decreased compared with p-coumaric
acid because hydroxy group at ortho position stabilized phenoxy
anion and the reactivity of phenoxy anion was increased. Ferulic
acid and sinapinic acid had methoxy groups unlike p-coumaric acid
and caffeic acid. It was seemed that steric hindrance of these meth-
oxy groups inhibited the reaction of TAXB and phenoxy anion, and
the decrease of the yield was not occurred. In the case of
hydroxycinnamoyl 2,3,4-tri-O-acetyl-b-D-xylopyranosides was
also performed using Novozym 435 (Table 4). 1d, 1b, and 1c did
not dissolve completely in MTBE (entries 2, 3, and 4). Mixed sol-
vent (MTBE; 30 mL and benzene; 10 mL) was used. Reaction pro-
gressed with a yield of approximately 80% in all substrates at
24 h. These results indicated that there was no affect for the reac-
tion by mixing of benzene. High resolution mass analysis and 2D
NMR of the product obtained was analyzed, and these results
showed that these products were hydroxycinnamoyl 2,3-di-O-
acetyl-b-D-xylopyranosides (2) deacetylated at only C-4 position.
Compound 2 transesterified only C-4 positioned acetyl group were
obtained with about 80% yields in all substrate. These results
showed that structural difference in hydroxycinnamic acid deriva-
tives bound to C-1 position did not affect the substrate specificity
of Novozym 435. Hydroxycinnamoyl 2,3-di-O-acetyl-b-D-glucopyr-
anosides were obtained in a yield of 80–90% at 24 h, and the
Table 3
Effect of amount of ferulic acid on glycosylation and applied the optimal conditions
for synthesis of other hydroxycinnamoyl b-D-xylopyranosidesa
Entry
Carboxylic acid/amount (equiv)
Amine
Product
Yield (%)
1
2
3
4
5
6
7
8
9
Cinnamic acid/3.0
p-Coumaric acid/3.0
i-Pr2NEt
i-Pr2NEt
Et3N
i-Pr2NEt
Et3N
1a
1b
70
53
67
45
46
65
73
69
70
Table 4
Novozym 435-catalyzed regioselective transesterification of 1a
Caffeic acid/3.0
1c
Entry
Substrate
Solvent
Product
Yield (%)
Ferulic acid/2.0
Ferulic acid/3.0
Ferulic acid/4.0
Sinapinic acid/3.0
i-Pr2NEt
1d
1
2
3
4
5
1a
1b
1c
1d
1e
MTBE (40 mL)
2a
2b
2c
2d
2e
85
86
71
89
83
MTBE/PhH (30 mL/10 mL)
MTBE/PhH (30 mL/10 mL)
MTBE/PhH (30 mL/10 mL)
MTBE (40 mL)
i-Pr2NEt
1e
a
Xylosyl bromide: 2.0 mmol, amine: 6.0 mmol, MS 4 Å: 1 g, MeCN: 10 mL, under
N2, rt, 24 h.
a
1: 1.0 mmol, MeOH: 12.0 mmol, Novozym 435: 0.8 g, 50 °C, 24 h.

Y. Shimotori et al. / Carbohydrate Research 388 (2014) 138–146
143
transesterification did not further progress even extending the
reaction time. These results indicated that Novozym 435 recog-
nized the whole molecular of substrates, and transesterified only
acetyl groups at C-4 and C-6 position. In other words, hydroxycin-
assigned to H-1. The doublet–doublet at dH 4.91 correlated with
H-1 in the H–H COZY spectrum was assigned to H-2. Similarly,
the signal of H-3 was correlated with H-2, the signal of H-4 was
correlated with H-3, the signal of H-5a and H-5b was correlated
with H-4. From the H–H COZY spectrum of xylopyranoside 2d,
the doublet–doublet at dH 5.08, the multiplet at dH 3.73, the dou-
blet–doublet at dH 3.50 and the doublet–doublet at dH 3.91 were
assigned to H-3, H-4, H-5a, and H-5b, respectively. The assignment
of 13C NMR was based on the HMQC spectrum. The assignment of
signals for C-1 (dC 92.1), C-2 (dC 70.6), C-3 (dC 74.7), C-4 (dC 67.0), C-
5 (dC 65.9), and –OCH3 (dC 55.7) was based on the cross-peak of the
1H and 13C NMR spectra. The assignments of acetoxyl groups of the
carbon signals at dC 169.3 and 169.7 were based on the HMQC and
HMBC spectra. The signal at dC 169.3 correlated with H-2 (dH 4.91)
in the HMBC spectrum was assigned to AC@O in OAc at C-2, while
the signal at dH 1.97 correlated with AC@O at dC 169.3 and was as-
signed to –CH3 in OAc at C-2. The signal in dC 20.4 correlated with –
CH3 in OAc at H-2 (dH 1.97) in the HMQC spectrum was assigned to
–CH3 in OAc at C-2. In addition, the signal at dC 169.7 was corre-
lated with H-3 (dH 5.08) in the HMBC spectrum, and was assigned
to AC@O in OAc at C-3. The signal at dH 2.03 correlated with AC@O
in OAc at C-3 (dC 169.7) in the HMBC spectrum was assigned to –
CH3 in OAc at C-3, while the signal at dC 20.7 correlated with –
CH3 in OAc at C-3 (dH 2.03) in the HMQC spectrum was assigned
to –CH3 in OAc at C-3. Additionally, the dH of H-4 were most shifted
5.03 to 3.73 between 1d and 2d compared with the dH of H-2 and
H-3. Thus, it was elucidated that xylopyranoside 1d was regiose-
lectively deacetylated at C-4 with Novozym 435 to afford to
xylopyranoside 2d. Other xylopyranosides were elucidated in the
same way.
namoyl 2,3-di-O-acetyl-b-D-glucopyranosides was not fully com-
plying with the substrate specificity of Novozym 435. Reaction
mechanism of lipase-catalyzed transesterification was sug-
gested.31–33 The crystal structure of Candida antarctica lipase B
(Novozym 435) has been resolved by Uppenberg et al.34 The active
site of Candida antarctica lipase B is composed by Ser105, His187,
and Asp224. Transesterification was started by attack oxygen atom
of serine to carbonyl carbon at acetyl group in substrate. Naoshima
et al. reported that it is easy to react as oxygen of serine that is ac-
tive center of lipase and carbonyl carbon of ester group in sub-
strates is close.35 It was previously considered that only acetyl
groups of hydroxycinnamoyl glucosides at C-4 and C-6 position
were selectively transesterified because steric hindrance around
these acetyl groups was small. Hydroxycinnamoyl 2,3,4-tri-O-acet-
yl-b-D-xylopyranosides had no O-acetyl carbinol group at C-5 posi-
tion unlike glucopyranosides. The vicinity of C-5 position was
vacant spatially. Carbonyl carbon in acetyl group at C-4 position
was lower steric hindrance compared with other acetyl groups,
and only C-4 acetyl group was transesterified. Acetyl group at C-
3 position was close to acetyl group at C-2 position, and acetyl
group at C-2 position was close to hydroxycinnamic acids moiety.
Distance between carbonyl carbon and oxygen of serine was long
because phenyl propane structure of ferulic acid was high steric
hindrance, which ferulic acid ester at C-1 position was not reacted.
Ferulic acid ester and acetyl groups at C-2 and C-3 position did not
transesterify even extend for more than 24 h. Lemke et al. sug-
gested that lipase recognized not size but shape of substrates,
and this recognition related to reactivity and selectivity.36 C-2, C-
3, and C-4 position were the same acetyl groups in hydroxycinna-
3.6. Radical scavenging of hydroxycinnamic acids and
respective b-D-xylopyranosides
moyl b-D-xylopyranosides (1) although C-1 position was ferulic
acid ester. In spite of that, Novozym 435 was transesterified only
acetyl group at C-4 position. From these results, it was assumed
that Novozym 435 recognized the shape of substrate molecule.
The antioxidant properties of hydroxycinnamic acids and its b-
D
-xylopyranosides synthesized were evaluated by the demonstra-
tion of DPPHÅ radical scavenging activity and the inhibitory effect
on autoxidation of methyl linoleate in bulk phase. Figure 1 shows
the DPPHÅ radical scavenging activity of hydroxycinnamic acids
Hydroxycinnamoyl 2,3,4-tri-O-acetyl-b-D-xylopyranosides (1)
were consistent with substrate specificity of Novozym 435, and
hydroxycinnamoyl 2,3-di-O-acetyl-b-
consistent was indicated.
D
-xylopyranosides (2) do not
and the respective b-
tested compounds increased with concentration in the range of
0.1–1.0
mol/mL to compare the previous data.13 Ferulic acid had
D-xylopyranosides. Scavenging activity of all
l
3.5. Structural elucidation of xylopyranoside 1d and 2d
the best radical scavenging activity compared with other tested
compounds, with the activity about 90% in all concentration. The
activity decreased in the order ferulic acid > caffeic acid ꢀ caffeoyl
The structural elucidation of the product 1d was based on the
2D NMR spectrum. The assignment of sugar part was described.
Because two oxygen atoms bond to the C-1 position of glycoside,
H-1 shifts to the downfield compared with the corresponding sig-
nals in simple carbohydrates. In addition, the doublet at dH 5.84
was assigned to H-1 in the 1H NMR spectrum. The doublet–doublet
at dH 5.16 correlated with H-1 signal and was assigned to H-2 from
the H–H COZY spectrum. In the same way, H-2 was correlated with
H-3, H-3 was correlated with H-4 and H-4 was correlated with H-
5a and H-5b, respectively. Therefore, the signals at dH 5.27, 5.03,
3.57, and 4.19 were assigned to H-3, H-4, H-5a, and H-5b, respec-
tively. The assignment of signals in 13C NMR was based on the
HMQC spectrum. The signals C-1 (dC 92.2), C-2 (dC 69.7), C-3 (dC
71.3), C-4 (dC 68.5), C-5 (dC 62.9), and –CH3 in OAc (dC 20.6) were
determined by the corresponding cross-peaks of the 1H and 13C
NMR spectra. In addition, the assignment of the carbon signals at
dC 169.5 and 169.8 were based on the HMBC spectrum. The signal
at dC 169.8 was correlated with H-3 at dH 5.27 and H-4 at dH 5.03,
and was assigned to AC@O in OAc on C-3 and C-4, respectively. The
signal at dC at 169.5 was correlated with H-2 at dH 5.16, and was
assigned to AC@O in OAc on C-2. The sugar part of xylopyranoside
2d was elucidated in the same way. The doublet at dH 5.83 was
b-
D-xylopyranosides (1c ꢀ 2c) ꢀ sinapinic acid > sinapoyl b-
D
-xylo-
-xylopyranosides (1d ꢀ 2d) >
-xylopyranosides (1b ꢀ 2b).
The scavenging activities of both caffeoyl b- -xylopyranosides (1c
and 2c) were almost the same as that of quercetin. Feruloyl b-
xylopyranosides (1d and 2d) and sinapoyl b- -xylopyranosides
(1e and 2e) had high DPPHÅ radical scavenging activity similar to
that of -tocopherol. Antioxidants that have many phenolic hydro-
pyranosides (2e > 1e) ꢀ feruloyl b-
D
p-coumaric acid > p-coumaroyl b-
D
D
D-
D
a
xy groups show high antioxidant activities. DPPHÅ radical scaveng-
ing activities of ferulic and sinapinic acids showed similar activities
compared with that of caffeic acid. Ortho methoxy groups for hy-
droxyl group at benzene ring improve antioxidative activity in or-
der to stabilize phenoxy radical. Ferulic acid and sinapinic acid
showed high radical scavenging activity compared with p-couma-
ric acid because these carboxylic acids had ortho methoxy groups.
Although the order of DPPHÅ radical scavenging activity was almost
the same as the case of glucopyranosides, radical scavenging activ-
ity of xylopyranosides was generally low compared with that of
glucopyranosides. Glucopyranosides had hydroxyl group and rela-
tively high polar acetoxy group at C-6 position. Xylopyranosides
did not have C-6 position. Therefore, the water solubility of

144
Y. Shimotori et al. / Carbohydrate Research 388 (2014) 138–146
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
0.1 μmol/mL
0.5 μmol/mL
1.0 μmol/mL
Figure 1. DPPHÅ radical scavenging activity of free hydroxycinnamic acids and its b-
D-xylopyranosides.
glucopyranosides was higher than that of xylopyranosides. Gluco-
pyranosides were easy to scavenge the water-soluble DPPHÅ radical,
and DPPHÅ radical scavenging activity of xylopyranosides was low-
Figure 2 shows the inhibitory effect on autoxidation of bulk methyl
linoliate using hydroxycinnamic acids and respective b-
D
-xylopyr-
-xylopyr-
anosides. The activity decreased in the order caffeoyl b-
D
er. The radical scavenging activities of all hydroxycinnamoyl b-
D
-
anosides (1c ꢀ 2c) > caffeic acid ꢀ sinapinic acid > sinapoyl
xylopyranosides except caffeic acid were lower than those of the
respective free hydroxycinnamic acids. Fukumoto et al. suggested
that the size of the steric hindrance of the sugar moiety reduced
the antioxidant properties.37 These results were in good agreement
with their suggestion. There was no great difference in the DPPHÅ
b-
D
-xylopyranosides (2e > 1e) > ferulic acid > feruloyl b-
D
-xylopyr-
-xylopyr-
anosides (1d ꢀ 2d) > p-coumaric acid ꢀ p-coumaroyl b-
D
anosides (1b ꢀ 2b). Both caffeoyl b-
D-xylopyranosides (1c and 2c)
showed the highest antioxidant capacities, which were higher than
that of quercetin. Free caffeic acid showed the antioxidant activity
scavenging activities between p-coumaroyl b-
D
-glucopyranosides
similar to quercetin. Free sinapinic acid, sinapoyl b-
sides (1e and 2e), and free ferulic acid showed a higher antioxidant
capacity compared with -tocopherol. Free caffeic acid and caf-
feoyl b- -xylopyranosides (1c and 2c) that had two phenolic hy-
D-xylopyrano-
and b- -xylopyranosides. The difference between acetoxy group
D
(1b) and hydroxy group (2b) at C-4 position did not affect the
a
DPPHÅ radical scavenging activities because DPPHÅ radical scaveng-
D
ing activities of p-coumaroyl b-
were originally low. The activities of caffeoyl b-
D
-xylopyranosides (1b and 2b)
-xylopyranosides
droxyl groups showed the highest activities against autoxidation
of methyl linoleate as well as DPPHÅ radical scavenging activity.
Sinapinic acid showed the highest antioxidant capacity, and p-cou-
maric acid was the lowest antioxidant capacity by above men-
tioned reason. Slight decrease of antioxidant capacity in
xylopyranosides was observed compared with that of glucopyr-
anosides as with DPPH radicalÅ scavenging activity. Glucopyrano-
sides had acetoxy group or hydroxyl group at C-6 position.
Xylopyranosides show comparatively high solubility for methyl
linoleate compared with glucopyranosides. Porter et al. advanced
the so-called ‘polar paradox’ to describe the observation that polar
antioxidants are more effective in nonpolar lipids, whereas nonpo-
lar antioxidants are more effective in polar lipid emulsions.38 To
explain this theory, a mechanism was previously postulated on
the basis of interfacial properties of different antioxidants.39 It
D
(1c and 2c) were almost the same compared with that of free caf-
feic acid. Glycosylation did not affect for DPPHÅ radical scavenging
activity for caffeic acid. Both free caffeic acid and caffeoyl b-D-xylo-
pyranosides (1c and 2c) did not show concentration dependence.
Additionally, deacetylation at C-4 position did not affect the activ-
ity. It was suggested that there was no effect on DPPHÅ radical scav-
enging activity because of the originally high activity of free caffeic
acid. Feruloyl b-
dent on concentration than feruloylb-
reduction of DPPHÅ radical scavenging activity of feruloyl b-
pyranosides (1d and 2d) in 0.1 mol/mL was observed. Although
the concentration dependence of DPPHÅ radical scavenging activity
was hardly seen in sinapoyl b- -glucopyranosides, that in sinapoyl
b-
D
-glucopyranosides (1d and 2d) were more depen-
-glucopyranosides. Great
-xylo-
D
D
l
D
D
-xylopyranosides (1e and 2e) was slightly observed. Xylose
was seemed that hydroxycinnamoyl b-
pressed low antioxidant capacity compared with hydroxycinna-
moyl b- -glucopyranosides for ‘polar paradox’ theory. Feruloyl
and sinapoyl b- -xylopyranosides showed low antioxidant capac-
D-xylopyranosides ex-
moiety reduced the DPPHÅ radical scavenging activity compared
with glucose moiety by above mentioned reason, and it seemed
that the concentration dependency was expressed. Difference be-
tween acetoxy group (1d) and hydroxyl group (2d) of feruloyl
D
D
ity compared with the respective free carboxylic acids. Solubility
of free hydroxycinnamic acids for methyl linoleate was low be-
cause of high polar carboxy group, and it was seemed that their
b-D
-xylopyranosides in C-4 position did not affect the DPPHÅ radical
scavenging activity. However, DPPHÅ radical scavenging activity of
2e was slightly increased compared with that of 1e. It was assumed
that DPPHÅ radical scavenging became easy by increase of water
solubility.
antioxidant capacity was higher than that of the respective b-D-
xylopyranosides. The antioxidant capacity of 2e was higher than
that of 1e. Antioxidant capacity of 2e increased because solubility
for methyl linoleate reduced by deacetylation at C-4 position.
These results were in good agreement with the ‘Polar paradox’.
3.7. Inhibitory effect on autoxidation of bulk methyl linoleate
Both caffeoyl b-D-xylopyranosides (1c and 2c) showed higher anti-
Formation of hydroperoxides in bulk methyl linoleate at 40 °C
was estimated based on the absorbance at 234 nm due to a conju-
gated methyl linoleate yielded during the initial oxidation stage.
oxidant capacities compared with free caffeic acid. Solubility for
methyl linoleate decreased compared with free caffeic acid be-
cause these caffeoyl b-D-xylopyranosides had high polar functional

Y. Shimotori et al. / Carbohydrate Research 388 (2014) 138–146
145
3.00
2.50
2.00
1.50
1.00
0.50
0.00
Control
α-Tocopherol
Quercetin
p-Coumaric acid
p-Coumaroyl xyloside 1b
p-Coumaroyl xyloside 2b
Caffeic acid
Caffeoyl xyloside 1c
Caffeoyl xyloside 2c
Ferulic acid
Feruloyl xyloside 1d
Feruloyl xyloside 2d
Sinapinic acid
Sinapoyl xyloside 1e
Sinapoyl xyloside 2e
0
5
10
15
20
25
30
Oxidation time [days]
Figure 2. Inhibition of the formation of hydroperoxide in bulk methyl linoleate by hydroxycinnamic acids and its b-
D
-xylopyranosides.
groups such as acetoxyl and hydroxyl groups. The degree of autox-
idation of methyl linoleate in the case of adding caffeoyl b- -xylo-
was shown from these results. It expects to be used by many
researchers as environmentally friendly glycosylation method.
D
pyranosides (1c and 2c) was smaller compared with that of caffeic
acid. Especially, the degree was noticeable from 11 to 28 day. Caf-
feoyl b-D-xylopyranosides can expect further prolonged antioxida-
Acknowledgement
tive potency. These results showed that xylose moiety was
important.
We are grateful to the Novozymes A/S for the generous gift of
Novozym 435.
4. Conclusion
References
Hydroxycinnamoyl
were effectively prepared from various hydroxycinnamic acids ex-
cept caffeic acid in about 70% yields. Acetoxy group of 2,3,4-tri-O-
acetyl-b-D-xylopyranosides was transesterified at C-4 position
regioselectively by Novozym 435. The regioselective deacetylation
of only acetoxyl group at C-4 position in these compounds with
Novozym 435 is primarily interesting. The antioxidant properties
of hydroxycinnamic acids and the respective hydroxycinnamoyl
2,3,4-tri-O-acetyl-b-D-xylopyranosides
1. Kishimoto, N.; Kakio, Y.; Iwai, K.; Mochida, K.; Fujita, T. Biocontrol Sci. 2005, 10,
155–161.
2. Kanno, K.; Sagakami, H.; Kawazoe, Y.; Yamamoto, N. Eur. Pat. Appl. Cpden:
EPXXDW EP 544321 A1 19,930,602, 1993.
3. Mattila, P.; Kumpulainen, J. J. Agric. Food Chem. 2002, 50, 3660–3667.
4. Silva, F. A. M.; Gorges, F.; Guimarães, C.; Lima, J. L. F. C.; Motos, C.; Reis, S. J.
Agric. Food Chem. 2000, 48, 2122–2126.
5. Faulds, C. B.; Mandalari, G.; LoCurto, R.; Bisignano, G.; Waldron, K. W. Appl.
Microbiol. Biotechnol. 2004, 64, 644–650.
6. Amorati, R.; Pedulli, G. F.; Cabrini, L.; Zambonin, L.; Landi, L. J. Agric. Food Chem.
2006, 54, 2932–2937.
b-
sides (1c and 2c) showed the highest DPPHÅ radical scavenging
activities in all hydroxycinnamoyl b- -xylopyranosides, and its
activities were almost the same as quercetin. Feruloyl (1d and
2d) and sinapoyl (1e and 2e) b- -xylopyranosides showed high
DPPHÅ radical scavenging activity as same as -tocopherol. DPPHÅ
radical scavenging activity of hydroxycinnamoyl b- -xylopyrano-
D-xylopyranosides were investigated. Caffeoyl b-D-xylopyrano-
7. Richards, R. M. E.; Durham, D. G.; Liu, X. Planta Med. 1994, 60, 471–473.
8. Rauha, J.-P.; Remes, S.; Heinonen, M.; Hopia, A.; Kähkönen, M.; Kujala, T.;
Pihlaja, K.; Vuorela, H.; Vuorela, P. Int. J. Food Microbiol. 2000, 56, 3–12.
9. Nogueira, E.; Vassilieff, V. S. J. Ethnopharmacol. 2000, 70, 275–280.
10. Costantino, L.; Albasini, A.; Rastelli, G.; Benvenuti, S. Planta Med. 1992, 58, 342–
344.
11. Hussein, S. A. M.; Ayoub, N. A.; Nawwar, M. A. M. Phytochemistry 2003, 63, 905–
911.
12. Cheel, J.; Theoduloz, C.; Rodríguez, J.; Saud, G.; Caligari, P. D. S.; Schmeda-
Hirschmann, G. J. Agric. Food Chem. 2005, 53, 8512–8518.
13. Shimotori, Y.; Tsutano, K.; Soga, K.; Osawa, Y.; Aoyama, M.; Miyakoshi, T.
Carbohydr. Res. 2012, 359, 11–17.
14. Silva, F. A. M.; Borges, F.; Guimaraes, C.; Lima, J.; Matos, C.; Reis, S. J. Agric. Food
Chem. 2000, 48, 2122–2126.
15. Gutmann, V. The Donor–Acceptor Approach to Molecular Interactions; Plenum:
New York and London, 1978.
16. Mönch, B.; Gebert, A.; Emmerling, F.; Becker, R.; Nehls, I. Carbohydr. Res. 2012,
352, 186–190.
D
D
a
D
sides except caffeic acid and p-coumaric acid derivatives slightly
decreased compared with the respective free carboxylic acids. In
the case of sinapoyl b-
of 2e deacetylated was higher than that of 1e. Caffeic acid and caf-
feoyl b- -xylopyranosides (1c and 2c) had the most effective anti-
oxidant activity among tested hydroxycinnamic acid derivatives
against autoxidation of methyl linoleate. Caffeoyl b- -xylopyrano-
D-xylopyranosides (1e and 2e), the activity
D
D
sides (1c and 2c) had high antioxidant activity and long durability
of antioxidative potency. The importance of xylose moiety was
confirmed from these results. The antioxidant activities of 1c and
2c were almost the same as quercetin and higher than that of
17. Batista, R.; Humberto, J. L.; Chiari, E.; Oliveira, A. B. Bioorg. Med. Chem. 2007, 15,
381–391.
18. Kimmel, R.; Kafka, S.; Košmrlj, J. Carbohydr. Res. 2010, 345, 768–779.
19. Kelly-Rowley, A. M.; Lynch, V. M.; Anslyn, E. V. J. Am. Chem. Soc. 1995, 111,
3438–3447.
a
-tochopherol. When xylosyl bromide was used instead of glucosyl
20. Ivari, K.; Agnes, K.; Lilli, S.; Toomas, R.; Vahur, M.; Ivo, L.; Ilmar, A. K. J. Org.
Chem. 2005, 70, 1019–1028.
21. Haav, K.; Saame, J.; Kütt, A.; Leito, I. Eur. J. Org. Chem. 2012, 2167–2172.
bromide, the corresponding hydroxycinnamoyl b- -xylopyrano-
sides was able to be synthesized with high yield. The availability
of amine-promoted glycosylation without heavy metal catalyst
D
ˇ
22. Gazák, R.; Trouillas, P.; Biedermann, D.; Fuksová, K.; Marhol, P.; Kuzma, M.;
ˇ
Kren, V. Tetrahedron Lett. 2013, 54, 315–317.

146
Y. Shimotori et al. / Carbohydrate Research 388 (2014) 138–146
23. Montalban, A. G.; Baum, S. M.; Cowell, J.; Mckillop, A. Tetrahedron Lett. 2012, 53,
4276–4279.
31. Ema, T.; Kobayashi, J.; Maeno, S.; Sakai, T.; Utaka, M. Bull. Chem. Soc. Jpn. 1998,
71, 443–453.
24. Zhang, Q.; Fan, Z.; Dong, J.; Shi, X.-X.; Lu, X. Tetrahedron: Asymmetry 2013, 24,
633–637.
25. Galland, S.; Mora, N.; Abert-Vian, M.; Rakotomanomana, N.; Dangles, O. J. Agric.
Food Chem. 2007, 55, 7573–7579.
26. Pekkarinen, S. S.; Stöckmann, H.; Schwarz, K.; Heinonen, I. M.; Hopia, A. I. J.
Agric. Food Chem. 1999, 47, 3036–3043.
32. Schulz, T.; Pleiss, J.; Schmid, R. D. Protein Sci. 2000, 9, 1053–1062.
33. Raza, S.; Fransson, L.; Hult, K. Protein Sci. 2001, 10, 329–338.
34. Uppenberg, J.; Hansen, M. T.; Patkar, S.; Jones, T. A. Structure 1994, 2, 293–308.
35. Naoshima, Y.; Mori, Y.; Yagi, Y. In Proceeding of the 15th International
Conference on Genome Informatics Poster and Software Demonstrations,
P129-1–P129-2, 2004.
27. Kikuzaki, H.; Hisamoto, M.; Hirose, K.; Akiyama, K.; Taniguchi, H. J. Agric. Food
Chem. 2002, 50, 2161–2168.
28. Morley, K. L.; Chauve, G.; Kazlauskas, R.; Dupont, C.; Shareck, F.; Marchessault,
R. H. Carbohydr. Polym. 2006, 63, 310–315.
29. Pouvreau, L.; Jonathan, M. C.; Kabel, M. A.; Hinz, S. W. A.; Gruppen, H.; Schols,
H. A. Enzyme Microb. Technol. 2011, 49, 312–320.
36. Lemke, K.; Lemke, M.; Theil, F. J. Org. Chem. 1997, 62, 6268–6273.
37. Fukumoto, L. R.; Mazza, G. J. Agric. Food Chem. 2000, 48, 3597–3604.
38. Porter, W. L.; Black, E. D.; Drolet, A. M. J. Agric. Food Chem. 1989, 37, 615–624.
39. Frankel, E. N.; Huang, S.-W.; Kanner, J.; German, J. B. J. Agric. Food Chem. 1994,
42, 1054–1059.
30. Poutanen, K.; Sundberg, M.; Korte, H.; Puls, J. Appl. Microbiol. Biotechnol. 1990,
33, 506–510.