Journal of Natural Medicines https://doi.org/10.1007/s11418-017-1148-8

ORIGINAL PAPER

Phenolic compounds from the leaves of Breynia officinalis and their tyrosinase and melanogenesis inhibitory activities Ayano Sasaki1 · Yoshi Yamano1 · Sachiko Sugimoto1 · Hideaki Otsuka1,2 · Katsuyoshi Matsunami1 · Takakazu Shinzato3 Received: 3 October 2017 / Accepted: 27 October 2017 © The Japanese Society of Pharmacognosy and Springer Japan KK 2017

Abstract From the EtOAc-soluble fraction of a MeOH extract of the leaves of Breynia officinalis, five new compounds (1–5) along with 11 known compounds (6–16) were isolated. The structures of the new compounds were elucidated by spectroscopic methods and compounds 1–3 were found to be acylated hydroquinone apiofuranosylglucopyranosides, while compound 4 was an acylated hydroquinone glucopyranoside. Compound 5 was shown to be butyl p-coumarate and this seems to be its first isolation from a natural source. The tyrosinase inhibitory activity of all of the isolated compounds was assayed, and the activity was significant in p-coumarate derivatives. The most active compound, compound 3, also inhibited melanogenesis in an in vivo whole animal model, zebrafish. Keywords  Breynia officinalis · Euphorbiaceae · Arbutin · Tyrosinase inhibitory activity · Zebrafish

Introduction Breynia officinalis Hemsley (Euphorbiaceae) is a poisonous perennial shrub 1.5–4 m in height that grows in Okinawa, Taiwan, and southern China. It is used as an ointment to heal wounds and as a remedy for edema, and by oral administration for syphilis and intestinal hemorrhage caused by overwork [1]. In a previous report, isolation of terpenic and phenolic glycosides from a 1-butanol-soluble fraction of the MeOH extract of the leaves of B. officinalis was reported [2]. Continuing investigation of the ethyl acetate-soluble fraction afforded four new compounds (1–4) and one isolated * Hideaki Otsuka hotsuka@hiroshima‑u.ac.jp; otsuka‑h@yasuda‑u.ac.jp * Katsuyoshi Matsunami matunami@hiroshima‑u.ac.jp 1



Department of Pharmacognosy, Graduate School of Biomedical and Health Sciences, Hiroshima University, 1‑2‑3 Kasumi, Minami‑ku, Hiroshima 734‑8553, Japan

2



Department of Natural Products Chemistry, Faculty of Pharmacy, Yasuda Women′s University, 6‑13‑1 Yasuhigashi, Asaminami‑ku, Hiroshima 731‑0153, Japan

3

Subtropical Field Science Center, Faculty of Agriculture, University of the Ryukyus, 685 Aza Yona, Kunigami‑son, Kunigami‑Gun 905‑1427, Okinawa, Japan



from a natural source for the first time (5), along with 11 known compounds (6–16) (Fig.  1): (E)-2-hexenyl β-dglucopyranoside (6) [3], (Z)-2-hexenyl β-d-glucopyranoside (7) [3], phenethyl alcohol β-d-glucopyranoside (8) [4], methyl caffeate (9) [5], methyl p-coumarate (10) [6], methyl α-d-glucopyranoside 6-O-benzoate (11) [7], arbutin 6′-O-cinnamate (12) [8], eximine (13) [9], robustaside A (14) [10], breynioside (15) [2], and β-d-glucopyranose 1,6-di-O-p-coumaroyl ester (16) [11]. Isolation of arbutin derivatives prompted us to assay their tyrosinase inhibitory activity [12]. The results of this assay are also discussed.

Results and discussion Compounds (1–16) were isolated using various column chromatography (CC) methods. The structures of the new compounds (1–5) were elucidated by precise inspection of one- and two-dimensional NMR spectra together with MS, IR, and UV spectra. Sugars were identified as thiazolidine derivatives by HPLC analysis or by using a refractive index monitor. The structures of the known compounds (6–16) were identified by comparison with spectroscopic data reported in the literature. All compounds isolated were examined for their tyrosinase inhibitory activity.

13

Vol.:(0123456789)



Journal of Natural Medicines

Fig. 1  New, known, and reference compounds

R1(1"')OH2C

O O 1 1'

OH

HO

OH

O

O

1 2 3 1a

1"

R2(1"")O OH

R2 A B C H

R1 A A A H

OH

R(1")OH2C O O HO

OH

1'

OH

R D C H

4 14 17

OH

7

O OR OR

1

5 Butyl (1'– 4') 10 Methyl

2

HO

ROH2C

O

HO

OH OH O

O 1

7

1

A

7

B

O

O

1

7 1

R C

16

C

D

HO Compound 1, [𝛼]28 − 45.8, was isolated as an amorphous D powder and its elemental composition was determined as ­C31H32O13 by high-resolution (HR)-electrospray ionization (ESI)-MS. The IR spectrum exhibited absorption bands

13

assignable to hydroxy groups (3406 cm−1), ester carbonyl group(s) (1718  cm −1), and aromatic ring(s) (1603 and 1510 cm−1). The UV absorption spectrum also supported the presence of aromatic rings. The 1H NMR spectrum

Journal of Natural Medicines

showed 14 aromatic protons, four of which [6.46 (2H, d, J  =  9.0  Hz, H-3 and 5), 6.85 (2H, d, J  =  9.0  Hz, H-2 and 6)] were correlated by 1H– 1H correlation spectroscopy (COSY). Heteronuclear single quantum coherence spectroscopy (HSQC) revealed that these protons were on δ C 116.6 and 119.0, respectively. Two anomeric proton signals [4.84 (1H, d, J = 7.7 Hz, H-1′) on δC 101.8 and 5.55 (1H, d, J = 1.1 Hz, H-1″) on δC 110.5] were observed and the presence of d-glucose and d-apiose was confirmed by HPLC analysis of the acid hydrolyzate. Two close aromatic signals at δH 7.60 and 7.65 on δC 134.36 and 134.41, coupled with adjacent aromatic protons as tt (J  =  7.5, 1.2 Hz) were assumed to be the 4-positions of the monosubstituted aromatic rings. The four 13C NMR aromatic signals appearing as a double strength peak, together with two single strength aromatic carbon signals and two carbonyl carbons, indicate that two benzoyl groups must be present in the molecule. Since mild alkaline hydrolysis of 1 gave seguinoside A (1a) [13], the structure of 1 was thus expected to be seguinoside A dibenzoate. The acylated positions were deduced by the downfield shift of C-6′ and C-5″, when compared with those of 1a, and the heteronuclear multiple bond correlations (HMBC) between ­H2-6′ and ­H2-5″, and C-1″′ and C-1″″, respectively, confirmed their presence on the hydroxy groups at these positions (Fig. 2). Therefore, the structure of 1 was elucidated as seguinoside A 6′,5″-di-O-benzoate as shown in Fig. 1. Compound 2, [𝛼]27 D − 32.5, was isolated as an amorphous powder and its elemental composition was determined as ­C33H34O13 by HR-ESI–MS. The IR and UV spectra showed similar absorptions to those of 1. The NMR spectra also showed good similarity to those of 1, except for the presence of a trans double bond [δH 6.42 (1H, d, J = 16.0 Hz) on δC 118.5 and 7.64 (1H, d, J = 16.0 Hz) on δC 146.7]. Mild

O O CH2 O O HO

O

OH

OH

O

COSY

OH OH

H C HMBC

O O CH2

Fig. 2  Selected 1H–1H COSY and HMBC correlations for compound 1 

alkaline hydrolysis of 2 also gave seguinoside A (1a). Thus, compound 2 was expected to be a seguinoside A benzoate or cinnamate. The HMBC cross-correlation of peaks between ­H2-5″ and δC 168.3, and δH 7.64 and δC 168.3 and 129.4 (C-2″″ and 7″″) confirmed that the cinnamate was esterified onto the hydroxy group at the C-5″ position. Therefore, the structure of compound 2 was elucidated as shown in Fig. 1. Compound 3, [𝛼]27 D − 52.0, was isolated as an amorphous powder and its elemental composition was determined as ­C33H34O14, having one oxygen more than compound 2. The NMR spectra of 3 were similar to those of 2 and typical signals for the trans double bond [δH 6.25 (d, J = 15.9 Hz) and 7.60 (d, J = 15.9 Hz)] were also observed. Two protons each of aromatic signals appearing in an AA′BB′ coupling manner were observed, besides those of the arbutin moiety. This evidence suggested that the cinnamate in 2 was replaced by p-coumarate, and the position of this acyl group was confirmed to be the same as that in 2 from the following two HMBC correlations, between H ­ 2-5″ (δH 4.26 and 4.31) and C-9″″ (δC 168.9), and H-7″″ (δH 7.60) and C-9″″. Therefore, the structure of 3 was elucidated as shown in Fig. 1. Compound 4, [𝛼]27 D − 56.9, was isolated as an amorphous powder and its elemental composition was determined to be ­C18H24O8. Spectroscopic data indicated that 4 was also an arbutin derivative with a six-carbon unit. This six-carbon unit comprised one methyl group, two methylene groups, one trans double bond, and a carbonyl group. The connectivity of these carbons was confirmed from the 1H–1H COSY spectrum and the position of esterification was determined to be the hydroxy group at C-6′ from the obvious downfield shift of H ­ 2-6′ protons (δH 4.30 and 4.50), and HMBC correlations ­H2-6′ and C-1″ (δC 168.1), and H-3″ (δH 7.01) and C-1″. Therefore, the structure of 4 was elucidated as shown in Fig. 1. Compound 5 was isolated as an optically inactive amorphous powder and its elemental composition was determined to be ­C13H15O3. In the 1H NMR spectrum, two protons each of aromatic signals appeared in an AA′BB′ coupling manner, and HMBC correlations indicated that 5 was a derivative of p-coumarate while the remaining four carbons comprised butyl alcohol. Therefore, the structure of 5 was elucidated as butyl p-coumarate. Some biological activities have been reported for butyl p-coumarate (5), such as larvicidal activity against Toxocara canis [14] and antifungal activity against Pythium sp. [15]. Compound 5 used in these bioassays was a synthetic material and, according to a database literature survey, this is the first report of its isolation from a natural source. Sunburn and chloasma are caused by deposition of the pigment melanin, which is formed as a result of the stimulation by sunlight of melanocytes located in the basement membrane. Arbutin is a natural skin-whitening agent, originally isolated from bearberry. Arbutin (17) itself and

13



Journal of Natural Medicines

Table 1  Tyrosinase inhibitory activity Sample

IC50 (μM)

1 2 3 4 5 6 7 8 9 0 11 12 13 14 15 16 Arbutin (17) p-Coumaric acid

> 200 > 200 16.9 ± 2.3 > 200 75.4 ± 6.1 > 200 > 200 > 200 > 200 143.6 ± 15.1 > 200 > 200 > 200 69.1 ± 9.6 > 200 156.6 ± 16.5 86.1 ± 6.8 > 200

Each value represents the mean ± SEM (N = 3)

synthetic α-arbutin (hydroquinone α-d-glucopyranoside) are available commercially as skin-whitening agents and they inhibit tyrosinase, which is involved at an early stage of melanin production [12]. B. officinalis was found to be a rich source of arbutin derivatives. The compounds (1‒16) isolated in this study were assayed for their tyrosinase inhibitory activity, and the results are summarized in Table 1. Of the compounds tested, compounds 3, 5, 10, 14, and 16 showed significant activity, and the activity of compounds 3, 5, and 14 was stronger than that of the positive control, arbutin, with ­IC50 values of 16.9 ± 2.3, 75.4 ± 6.1, 69.1 ± 9.6, and 86.1  ±  6.8  μM, respectively. These compounds contain p-coumarate as a common unit. The first step in the reaction of tyrosinase may be hydroxylation of the aromatic ring to form a cathechol-like compound. However, methyl caffeate (9) did not show inhibitory activity and nor did p-coumaric acid itself (Table 1). It is noteworthy that alkyl esters of p-coumaric acid show diverse biological activities, such as larvicidal [14], antifungal [5], as well as tyrosinase inhibitory activities. The cytotoxicity of these compounds was evaluated using the human lung epithelial cancer cell line A549, and none of the compounds showed cytotoxicity at 100 μg/ mL. Finally, in vivo inhibition of melanogenesis by 3, the most potent compound, was evaluated using a vertebrate whole animal model, zebrafish (Figs. 3 and 4). A reduction of pigmentation was clearly observed for 3 and the positive control, arbutin without any morphological differences (Fig. 4). Green fluorescence of the transgene flk1:EGFP indicated the normality of the network of vascular system. These results

13

indicated that compound 3 will be a promising candidate whitening agent, and further mechanistic considerations of its inhibitory activity will be of future interest.

Conclusion Five new phenolic compounds were isolated from the leaves of Breynia officinalis. Intensive spectroscopic investigation revealed that they were derivatives of arbutin and butyl p-coumarate. In an assay for tyrosinase inhibitory activity, compound 3 was found to have promising properties. These results suggested that compound 3 would likely inhibit the final stage of melanin pigment deposition, which was surveyed using an in vivo zebrafish method. Compound 3 actually showed significant anti-melanogenesis activity as shown in Fig. 4. Therefore, compound 3 will be a promising candidate for a skin-whitening agent.

Materials and methods General Optical rotations were measured on JASCO P-1030 spectropolarimeters. IR and UV spectra were measured on Horiba FT-710 and JASCO V-520 UV/Vis spectrophotometers, respectively. 1H and 13C NMR spectra were taken on a Bruker Avance III spectrometer at 600 and 150 MHz, respectively, with tetramethylsilane as an internal standard. HR-ESI mass spectra were performed with an LTQ Orbitrap XL spectrometer (Thermo Fisher Scientific). Silica gel CC and reversed-phase (ODS) open CC were performed on silica gel 60 (Merck, Darmstadt, Germany) and Cosmosil ­75C18-OPN (Nacalai Tesque, Kyoto, Japan) [Φ = 50 mm, L  =  20  cm, stepwise gradient (5%) from 10% MeOH in ­H2O (400 mL) to 100% MeOH (400 mL) and then acetone (400 mL)] to give 20 fractions. HPLC was performed on Inertsil ODS-3 column (GL Science, Tokyo, Japan; Φ = 10 mm, L = 25 cm), Cosmosil πNAP (Nakalai Tesque, Kyoto, Japan, Φ = 10 mm, L = 25 cm), and Cosmosil Cholester columns (Φ = 10 mm, L = 25 cm) and the eluate was monitored with a refractive index monitor. Mushroom tyrosinase was purchased from Sigma–Aldrich Co. LLC. (St. Louis, MO, USA).

Plant material Leaves of B. officinalis Hemsl. (Euphorbiaceae) were collected in Kunigami-son, Kunigami-gun, Okinawa, Japan, in August 2000, and a voucher specimen was deposited in the Herbarium of the Department of Pharmacognosy, Graduate School of Biomedical Sciences, Hiroshima University (00-BO-Okinawa-0828).

Journal of Natural Medicines

Fig. 3  Melanogenesis inhibition in zebrafish in  vivo model. a Treatment with 3 did not alter zebrafish development: upper specimen, no treatment; lower specimen, treatment with 3 (1  mM). b Magnifica-

tion of upper image in a (no treatment) clearly shows black melanin spots on the dorsal side and on egg sac. c Treatment with 3 (1 mM). d Treatment with positive control, arbutin (1 mM)

Fig. 4  Morphological analysis by transgenic GFP expression in blood vessel. Internal morphological structure was examined by transgenic zebrafish with expression of GFP in blood vessel driven by kdr/

flk1/VEGF-R2 promoter. a, c No treatment. b, d Treatment with 3 (1  mM). The vascular development and network are essentially the same. The melanin is also observed as dark absorbed spots in a and c 

Extraction and fractionation

gum. The gummy residue was suspended in 3.0 L of H ­ 2O and then extracted successively with 3.0 L each of EtOAc and 1-BuOH to afford 111.5 and 46.6 g of the EtOAc- and 1-BuOH-soluble fractions, respectively. Evaporation of the ­H2O layer gave 313 g of residue. The EtOAc-soluble fraction (111 g) was subjected to silica gel (550 g) CC with n-hexane-CHCl3 (1:1, 3 L) and

Air-dried leaves of B. officinalis (4.11 kg) were extracted three times with MeOH. The MeOH extract was concentrated to 3.0 L and then 150 mL of ­H2O was added. This solution was washed with 3.0 L of n-hexane (159 g) and then the methanolic layer was concentrated to a viscous

13



then increasing amounts of MeOH in ­CHCl3 ­[CHCl3 (3 L), ­CHCl3–MeOH (50:1, 3 L), (40:1, 3 L), (30:1, 3 L), (20:1, 3 L), (15:1, 3 L), (10:1, 3 L), (7:1, 3 L), (5:1, 3 L), (3:1, 3 L), (2:1, 3 L) and then MeOH (3 L)] and 3-L fractions were collected to give 14 fractions. Fraction 2 (7.76 g) was subjected to ODS CC and fraction 11 (70.5 mg) was purified by HPLC [Cholester column, MeOH–H2O (13:7), flow rate 2.0 mL/min] to give 4.7 mg of 5 from the peak at 46 min. Fraction 4 (4.11 g) was subjected to ODS CC and fractions 7 (97.6 mg) and 8 (89.3 mg) were purified by HPLC [πNAP column, MeOH–H2O (1:1) and (11:9), respectively, flow rate 2.0 mL/min] to give 4.4 mg of 9 and 5.6 mg of 10 from the peaks at 25 min and 28 min, respectively. Fraction 6 (3.10 g) was subjected to ODS CC and fraction 6 (85.4 mg) was purified by HPLC [Cholester column, MeOH–H2O (19:31), flow rate 2.0 mL/min] to give 5.2 mg of 11 from the peak at 32 min. Fraction 7 (4.05 g) was subjected to ODS CC to give two fractions: fraction 4 (115.1 mg), fraction 10 (154.1 mg). The former fraction was purified by HPLC [ODS column, MeOH–H2O (1:3), flow rate 2.8 mL/min] to give 9.8 mg of 8, 5.3 mg of 6, and 3.4 mg of 7 from the peaks at 19 min, 25 min, and 29 min, respectively. The latter fraction was purified by HPLC [Cholester column, MeOH–H2O (29:21), flow rate 2.0 mL/min] to give 316 mg of 1 and 6.6 mg of 2 from the peaks at 29 min and 47 min, respectively. Fraction 8 (6.72 g) was subjected to ODS CC to give four fractions: fraction 1, 550 mg; fraction 2, 329 mg; fraction 3, 390 mg; and fraction 4, 305 mg. The first fraction (164 mg out of 550 mg) was purified by HPLC [πNAP column, MeOH–H2O (43:67), flow rate 2.0 mL/min] to give 13.7 mg of 15 from the peak at 19 min. The second fraction (163 mg out of 329 mg) was purified by HPLC [πNAP column, MeOH–H2O (3:2), flow rate 2.0 mL/min] to give 13.7 mg of 14, 37.7 mg of 13, and 19.5 mg of 4 from the peaks at 18 min, 24 min, and 31 min, respectively. The third fraction (52 mg out of 390 mg) was purified by HPLC [πNAP column, MeOH–H2O (3:2), flow rate 2.0 mL/min] to give 14.7 mg of 12 from the peak at 23 min. The fourth fraction was purified by HPLC [πNAP column, MeOH–H2O (59:41), flow rate 2.0 mL/min] to give 2.9 mg of 16 and 10.5 mg of 3 from the peaks at 22 min and 33 min, respectively. [ ] Compound 1 Colorless amorphous powder, 𝛼D28 − 45.8 (c = 0.78, MeOH); IR νmax (film) ­cm−1: 3406, 2932, 1718, 1603, 1510, 1451, 1279, 1072, 829; UV λ max (MeOH) nm (log ε): 278 (3.57), 227 (4.13); 1H NMR (­CD3OD, 600 MHz): δ 8.05 and 7.99 (2H each, dd, J = 8.2, 1.2 Hz, ­H2-2″′, 6″′ and ­H2-2″″, 6″″), 7.65 and 7.60 (1H each, tt, J = 7.5, 1.2 Hz, H-4″′ and 4″″), 7.52 and 7.44 (2H each, dd, J = 8.2, 7.5 Hz, ­H2-3″′, 5″′ and H ­ 2-3″″, 5″″), 6.85 (2H, d, J = 9.0 Hz, H-2 and 6), 6.46 (2H, d, J = 9.0 Hz, H-3 and 5), 5.55 (1H, d, J = 1.1 Hz, H-1″), 4.84 (1H, d, J = 7.7 Hz, H-1′), 4.73 (1H, dd, J = 11.8, 2.2 Hz, H-6′a), 4.44 (1H, d, J = 11.3 Hz, H-5″a), 4.41 (1H, dd, J = 11.8, 7.7 Hz,

13

Journal of Natural Medicines

H-6′b), 4.39 (1H, d, J  =  11.3  Hz, H-5″b), 4.37 (1H, d, J = 9.7 Hz, H-4″a), 4.06 (1H, d, J = 1.1 Hz, H-2″), 3.96 (1H, d, J = 9.7 Hz, H-4″b), 3.74 (1H, ddd, J = 9.7, 7.7, 2.2 Hz, H-5′), 3.71 (1H, dd, J = 9.1, 7.7 H, H-2′), 3.66 (1H, dd, J  =  9.1, 9.1  Hz, H-3′), 3.45 (1H, dd, J  =  9.7, 9.1  Hz, H-4′); 13C NMR ­(CD3OD, 150  MHz): Table  2; HR-ESI–MS (positive-ion mode): m/z: 635.1735 [M+Na]+ (Calcd ­C31H32O13Na: 635.1735). Compound 2 Colorless amorphous powder, [𝛼]27 D − 32.5 (c = 0.61, MeOH); IR νmax (film) ­cm−1: 3395, 2930, 1715, 1635, 1604, 1510, 1449; UV λmax (MeOH) nm (log ε): 274 (4.08), 262 (4.04), 221 (4.12); 1H NMR (­CD3OD, 600 MHz): δ 8.02 (2H, dd, J = 8.3, 1.3 Hz, H-2″′ and 6″′), 7.64 (1H, d, J = 16.0 Hz, H-7″″), 7.63 (1H, tt, J = 7.5, 1.3 Hz, H-4″′), 7.53 (2H, dd, J = 6.9, 3.2 Hz, H-2″″ and 6″″), 7.39 (3H, overlapped, H-3″″, 4″″ and 5″″), 7.50 (2H, dd, J = 8.3, 7.5 Hz, H-3″′ and 6″′), 6.86 (2H, d, J = 9.0 Hz, Table 2  13C NMR spectroscopic data for compounds 1–4 and 1a ­(CD3OD, 150 MHz) C

1

1a

2

3

4

1 2, 6 3, 5 4 1′ 2′ 3′ 4′ 5′ 6′ 1″ 2″ 3″ 4″ 5″ 6″ 1′″ 2′″, 6″′ 3′″, 5″′ 4′″ 7′″ 1″″ 2″″, 6″″ 3″″, 5″″ 4″″ 7″″ 8″″ 9″″

151.9 119.0 116.6 153.7 101.8 78.3 78.8 72.2 75.3 65.5 110.5 78.7 79.2 75.4 68.4

152.4 119.2 116.7 153.8 102.3 78.1 78.8 71.5 78.7 62.6 110.8 78.0 80.8 75.4 66.1

152.0 119.1 116.7 153.8 101.9 78.2 78.8 72.2 75.4 65.5 110.5 78.7 79.2 75.5 68.0

152.1 119.1 116.7 153.7 101.9 78.3 78.74 72.2 75.3 65.5 110.5 78.71 79.2 75.5 67.8

131.3 130.7 129.7 134.4 167.8 135.7 129.4 130.1 131.6 146.7 118.5 168.3

131.29 130.7 129.7 134.4 167.9 127.1 131.32 116.9 161.3 147.0 114.8 168.9

152.3 119.6 116.6 154.0 103.7 75.0 77.9 71.9 75.4 64.7 168.1 122.0 151.4 35.3 22.4 14.0

13

131.3 130.67 a 129.65b 134.41c 167.84d 131.1 130.68a 129.58b 134.36c 167.80d

 C NMR chemical shifts with the same superscript letters may be exchanged

Journal of Natural Medicines

H-2 and 6), 6.53 (2H, d, J = 9.0 Hz, H-3 and 5), 6.42 (1H, d, J = 16.0 Hz, H-8″″), 5.51 (1H, d, J = 1.1 Hz, H-1″), 4.82 (1H, d, J = 6.0 Hz, H-1′), 4.70 (1H, dd, J = 11.7, 2.2 Hz, H-6′a), 4.37 (1H, dd, J = 11.7, 7.7 Hz, H-6′b), 4.30 (1H, d, J = 11.3 Hz, H-5″a), 4.28 (1H, d, J = 9.7 Hz, H-4″a), 4.26 (1H, d, J = 11.3 Hz, H-5″b), 4.00 (1H, d, J = 1.1 Hz, H-2″), 3.89 (1H, d, J = 9.7 Hz, H-4″b), 3.69 (1H, ddd, J = 9.8, 7.7, 2.2 Hz, H-5′), 3.68 (1H, dd, J = 8.7, 6.0 Hz, H-2′), 3.63 (1H, dd, J = 8.7, 8.7 Hz, H-3′), 3.42 (1H, dd, J = 9.8, 8.7  Hz, H-4′); 13C NMR ­(CD3OD, 150  MHz): Table  2; HR-ESI–MS (positive-ion mode): m/z: 661.1887 [M+Na]+ (Calcd ­C33H34O13Na: 661.1892). Compound 3 Colorless amorphous powder, [𝛼]27 D − 52.0 (c = 1.05, MeOH); IR νmax (film) ­cm−1: 3384, 2948, 1712, 1696, 1631, 1604, 1512; UV λmax (MeOH) nm (log ε): 304 (4.13), 291 (4.07), 226 (4.09); 1H NMR ­(CD3OD, 600  MHz): δ 8.04 (2H, dd, J  =  8.1, 1.2  Hz, H-2″′ and 6″′), 7.64 (1H, tt, J  =  7.5, 1.2  Hz, H-4″′), 7.60 (1H, d, J = 15.9 Hz, H-7″″), 7.52 (2H, dd, J = 8.1, 7.5 Hz, H-3″′ and 5″′), 7.42 (2H, d, J = 8.7 Hz, H-2″″ and 6″″), 6.89 (2H, d, J = 8.9 Hz, H-2 and 6), 6.82 (2H, d, J = 8.7 Hz, H-3″″ and 5″″), 6.56 (2H, d, J = 8.9 Hz, H-3 and 5), 6.25 (1H, d, J = 15.9 Hz, H-8″″), 5.54 (1H, d, J = 1.1 Hz, H-1″), 4.84 (1H, d, J = 7.7 Hz, H-1′), 4.72 (1H, dd, J = 11.7, 2.1 Hz, H-6′a), 4.41 (1H, dd, J = 11.7, 7.7 Hz, H-6′b), 4.31 (1H, d, J = 11.3 Hz, H-5″a), 4.30 (1H, d, J = 9.7 Hz, H-4″a), 4.26 (1H, d, J = 11.3 Hz, H-5″b), 4.03 (1H, d, J = 1.1 Hz, H-2″), 3.92 (1H, d, J = 9.7 Hz, H-4″b), 3.75 (1H, ddd, J = 9.8, 7.7, 2.1 Hz, H-5′), 3.71 (1H, dd, J = 8.9, 7.7 Hz, H-2′), 3.67 (1H, dd, J = 8.9, 8.9 Hz, H-3′), 3.46 (1H, dd, J = 9.8, 8.9  Hz, H-4′); 13C NMR ­(CD3OD, 150  MHz): Table  2; HR-ESI–MS (positive-ion mode): m/z: 677.1838 [M+Na]+ (Calcd ­C33H34O14Na: 667.1841). Compound 4 Colorless amorphous powder, [𝛼]27 D − 56.9 (c = 1.95, MeOH); IR νmax (film) ­cm−1: 3384, 2961, 2931, 1701, 1651, 1608, 1511; UV λmax (MeOH) nm (log ε): 282 (3.80), 211 (4.35); 1H NMR ­(CD3OD, 600 MHz): δ 7.01 (1H, dt, J = 15.9, 7.2 Hz, H-3″), 6.95 (2H, d, J = 9.0 Hz, H-2 and 6), 6.70 (2H, d, J = 9.0 Hz, H-3 and 5), 5.89 (1H, dt, J = 15.7, 1.5 Hz, H-2″), 4.73 (1H, d, J = 7.4 Hz, H-1′), 4.50 (1H, dd, J = 11.8, 2.1 Hz, H-6′a), 4.30 (1H, dd, J = 11.8, 6.9 Hz, H-6′b), 3.63 (1H, ddd, J = 9.7, 6.9, 2.1 Hz, H-5′), 3.48 (1H, dd, J = 8.7, 8.7 Hz, H-3′), 3.45 (1H, dd, J = 8.7, 7.4 Hz, H-2′), 3.39 (1H, dd, J = 9.7, 8.7 Hz, H-4′), 2.25 (2H, qd, J = 7.2, 1.5 Hz. H ­ 2-4″), 1.53 (2H, sextet, J = 7.2 Hz, ­H2-5″), 0.98 (3H, t, J = 7.2 Hz. ­H3-6″); 13C NMR ­(CD3OD, 150 MHz): Table 2; HR-ESI–MS (positive-ion mode): m/z: 391.1364 [M+Na]+ (Calcd ­C18H24O8Na: 391.1363). Compound 5 Colorless amorphous powder; IR νmax (film) ­cm−1: 3368, 2959, 1682, 1603, 1513, 1442, 1275, 1168, 982; UV λmax (MeOH) nm (log ε): 300 (4.35), 282 (4.40), 1H NMR ­(CD3OD, 600 MHz): δ 7.62 (1H, d, J = 16.0 Hz, H-7), 7.48 (2H, d, J = 8.7 Hz, H-2 and 6), 6.83 (2H, d, J = 8.7 Hz,

H-3 and 5), 6.34 (1H, d, J = 16.0 Hz, H-8), 4.20 (2H, t, J = 6.6 Hz, ­H2-4′), 1.71 (2H, tt, J = 7.4, 6.6 Hz, ­H2-3), 1.47 (2H, qt, J = 7.4, 7.4 Hz, H ­ 2-2′), 1.00 (1H, t, J = 7.4 Hz, ­H3-1′); 13C NMR (­ CD3OD, 150 MHz): δ 169.4 (C-9), 161.3 (C-4), 131.1 (C-2 and 6), 127.2 (C-1), 116.8 (C-3 and 5), 115.3 (C-8), 65.3 (C-4′). 32.0 (C-3′), 20.3 (C-2′), 14.1 (C-1′); HR-ESI–MS (negative-ion mode): m/z: 219.1026 [M–H]− (Calcd ­C13H15O3: 219.1027).

Sugar analysis of compound 4 Compounds 4 (1.2  mg) was hydrolyzed with 1  M HCl (0.1 mL) at 80 °C for 2 h. The reaction mixture was partitioned with an equal amount of EtOAc (0.1 mL), and the water layer was analyzed by HPLC with a chiral detector (JASCO OR-2090plus) on an amino column [Shodex Asahipak N ­ H2P-50 4E (Φ = 4.6 mm, L = 250 mm), ­CH3CN–H2O (4:1), flow rate 1 mL/min]. The hydrolyzate gave a peak for d-glucose at 9.5 min with a positive optical rotation sign. The peak was identified by co-chromatography with an authentic sample.

Alkaline hydrolysis Compound 1 (6.5  mg) was dissolved in 2  mL of 0.1  M ­NaOCH3 in MeOH and stood for 30 min at 25 °C. The reaction mixture was neutralized with Amberlite IR-120B (­ H+) and concentrated. The residue thus obtained was subjected to preparative TLC with ­CHCl3–MeOH–H2O (15:6:1) and the collected due zone was eluted with ­CHCl3–MeOH (1:1) to give 2.2 mg of 1a. 1a: Colorless amorphous powder, [𝛼]27 D − 95.9 (c = 0.22, MeOH); 1H NMR ­(CD3OD, 600 MHz): identical to chemical shifts reported for seguinoside A; 13C NMR ­(CD3OD, 150 MHz): Table 2; HR-ESI–MS (positiveion mode): m/z: 427.1210 [M+Na]+ (Calcd ­C17H24O11Na: 427.1211). Compounds 2 and 3 were also hydrolyzed in a similar manner to give 1a.

Sugar analysis of compound 1a Compound 1a (1.0  mg) was hydrolyzed with 1  M HCl (0.2 mL) at 80 °C for 2 h. The reaction mixture was partitioned with an equal amount of EtOAc (0.2 mL), and the separated water layer was neutralized with Amberlite IRA96SB ­(OH−). The neutralized water was evaporated to dryness and the residue was dissolved in pyridine (0.1 mL), containing 0.5 mg of l-cysteine methyl ester. The reaction mixture was kept for 1 h at 60 °C and then, after addition of 1.4 mg of o-tolylisothiocyanate in 70 μL of pyridine, the reaction temperature was maintained at 60 °C for further 1 h. The thiazolidine derivatives thus formed were analyzed by HPLC with the refractive index detector on an ODS column [Cosmosil 5C18ARII (Nakalai Tesque, Kyoto, Japan),

13



Journal of Natural Medicines

Φ = 4.6 mm, L = 250 mm, ­CH3CN–50 mM ­H3PO4 (1:4), flow rate 0.8 mL/min] to give two peaks at 19 and 33 min. Authentic d-apiose and d-glucose were derivatized with d- and l-cysteine methyl esters and the thiazolidine derivatives of d-apiose gave peaks at 17 and 19 min, respectively; whereas those of d-glucose gave peaks at 37 and 33 min, respectively.

Tyrosinase inhibitory activity To the sample solution (10 μL), 40 μL of l-tyrosine solution (0.25 mg/mL) in 50 mM potassium phosphate buffer (pH 6.80) was added and the optical density (OD) at 475 nm was measure as background. Mushroom tyrosinase (100 U/mL) in 50 μL of the same buffer was added followed by incubation at 25 °C for 10 min [16]. The amounts of dopachrome formed were photometrically determined by measuring the OD at 475 nm using a microplate reader. Arbutin was used as a positive control. The inhibition of tyrosinase activity was calculated as follows:

Inhibition ratio (%) [ ( )/ ] = 1 − Asample − Abackground Acontrol − Abackground × 100 Acontrol: without test sample.

a photoperiod of 14 h light and 10 h dark. Breeding was carried out at 7:00 AM in the morning and fertilized eggs were collected and washed with embryo medium [18]. Compound 3 and positive control, arbutin, were dissolved in embryo medium with different concentrations and dispensed into a 96-well plate with the volume of 100 μL. Five eggs were transferred to each well by microspatula and were incubated for 48 h at 26 °C until pigmentation. The eggs were dechorionated using forceps and photographed by fluorescent microscope (KEYENCE BZ-8000) with bright field and fluorescence observation. Acknowledgement  The authors are grateful for access to the superconducting NMR instrument (Brucker Avance III 600) at the Analytical Center of Molecular Medicine of the Hiroshima University Faculty of Medicine, and HR-ESI mass spectra were performed with LTQ Orbitrap XL spectrometer at the Analysis Center of Life Science of the Graduate School of Biomedical Sciences, Hiroshima University. This work was supported in part by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan, and the Japan Society for the Promotion of Science (Nos. 22590006, 23590130, 25860078, 15H04651, and 17K08336). Thanks are also due to the Research Foundation for Pharmaceutical Sciences and the Takeda Science Foundation for the financial support. One of the authors (S.S.) thanks the Research Foundation for Pharmaceutical Sciences and the Cosmetology Research Foundation for the financial support.

Compliance with ethical standard  Conflict of interest  The authors declare no conflict of interest.

Cytotoxicity assay The cytotoxicity was evaluated using human lung epithelial carcinoma by MTT assay. In brief, the cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% heat-inactivated fetal bovine serum and antibiotics. A549 cells (5 × 103 cells/well) were cultured in a 96-well plate with solutions of the test compounds within a ­CO2 incubator at 37 °C for 72 h. The medium was replaced with fresh medium containing MTT and the plate was further incubated for another 1.5 h. Absorbance was measured at 540 nm by microplate reader after removal of medium and solubilization with DMSO. The viability was compared to that of control cells incubated in the same medium without test compounds.

In vivo zebrafish melanogenesis model Melanogenesis assay was performed by the method described previously [17]. In brief, transgenic zebrafish having enhanced green fluorescent protein (EGFP) and blood vessel specific flk1 promoter, Tg(flk1:EGFP), was purchased from Riken National BioResource Project. Approximately 10 male and female adult zebrafishes were maintained separately in two 10-L aquariums at 26 °C with

13

References 1. Chang PC, Wu CC (1979) Herbs of Taiwan in color, vol 1. Taipei, Nan Cun (South Villa) Trade, p 99 2. Morikawa H, Kasai R, Otsuka H, Hirata E, Shinzato T, Aramoto M, Takeda Y (2004) Terpenic and phenolic glycosides from leaves of Breynia officinalis HEMSL. Chem Pharm Bull 52:1086–1090 3. Mizutani K, Yuda M, Tanaka O, Saruwatari Y, Fuwa T, Jia M-R, Ling Y-K, Pu X-F (1998) Chemical studies on Chinese traditional medicine, dangshen. I.: isolation of (Z)-3- and (E)-2-hexenyl β-dglucosids. Chem Pharm Bull 36:2689–2690 4. Ma S-J, Mizutani M, Hiratake J, Hayashi K, Yagi K, Watanabe N, Sakata K (2001) Substrate specificity of β-primeverosidase, a key enzyme in aroma formation during oolong tea and black tea manufacturing. Biosci Biotechnol Biochem 65:2719–2729 5. He Z-D, Yang C-R (1991) Brandioside, a phenylpropanoid glycoside from Brandisia hancei. Phytochemistry 30:701–702 6. Oyama K, Watanabe N, Yamada T, Suzuki M, Sekiguchi Y, Kondo T, Yoshida K (2015) Efficient and versatile synthesis of 5-O-acylquinic acids with a direct esterification using a p-methoxybenzyl quinate as a key intermediate. Tetrahedron 71:3120–3130 7. Evtushenko EV (2012) Regioselective benzoylation of glycopyranosides by benzoic anhydride in the presence of Cu(CF3COO)2. Carbohyd Res 359:111–119 8. Yang R-L, Li N, Ye M, Zong M-H (2010) Highly regioselective synthesis of novel aromatic esters of arbutin catalyzed by

Journal of Natural Medicines

9. 10. 11. 12.

13.

immobilized lipase from Penicillium expansum. J Mol Catalysis B: Enzymatic 67:41–44 Verotta L, Orsini F, Pelizzoni F, Torri G, Rogers CB (1999) Polyphenolic glycosides from African Proteaceae. J Nat Prod 62:1526–1531 Ahmed AS, Nakamura N, Meselhy MR, Makhboul MA, El-Emary N, Hattori M (1988) Phenolic constituents from Grevillea robusta. Phytochemistry 53:149–154 Shimomura H, Sashida Y, Adachi T (1988) Phenylpropanoid glucose esters from Prunus buergeriana. Phytochemistry 27:641–644 Sugimoto K, Nishimura T, Nomura K, Sugimoto K, Kuriki T (2003) Syntheses of arbutin-α-glycosides and a comparison of their inhibitory effects with those of α-arbutin and arbutin on human tyrosinase. Chem Pharm Bull 51:798–801 Zhong X-N, Otsuka H, Ide T, Hirata E, Takushi A, Takeda Y (1998) Hydroquinone glycosides from leaves of Myrsine seguinii. Phytochemistry 49:2149–2153

14. Kiuchi F, Nakamura N, Tsuda Y, Kondo K, Yoshimura H (1988) Studies on crude drugs effective on visceral larva Migrans. II.: larvicidal principles in Kaempferiae rhizoma. Chem Pharm Bull 36:412–415 15. Tawata S, Taira S, Kobamoto N, Zhu J, Ishihara Toyama S (1996) Synthesis and antifungal activity of cinnamic acid esters. Biosci Biotech Biochem 60:909–910 16. Pomerants SH (1963) Separation, purification, and properties of two tyrosinases from hamster melanoma. J Biol Chem 238:2351–2357 17. Choi T-Y, Kim J-H, Ko DH, Kim C-H, Hwang J-S, Ahn S, Kim SY, Kim C-D, Lee J-H, Yoon T-J (2007) Zebrafish as a new model for phenotype-based screening of melanogenic regulatory compounds. Pigment Cell Res 20:120–127 18. Westerfield M (2007) The zebrafish book: a guide for the laboratory use of zebrafish (Danio rerio), 5th edn. University of Oregon Press, Eugene

13