ANALYTICAL SCIENCES APRIL 2019, VOL. 35 407 2019 © The Japan Society for Analytical Chemistry

Separation of Enantiomers in by a Reversed Phase HPLC after Chiral Precolumn Derivatization

Sohei TANAKA,*1 Misaki SEKIGUCHI,*1 Atsushi YAMAMOTO,*2 Sen-ichi AIZAWA,*3 Kanta SATO,*4 Atsushi TAGA,*4 Hiroyuki TERASHIMA,*5 Yoshimi ISHIHARA,*1 and Shuji KODAMA*1†

*1 School of Science, Tokai University, 4-1-1 Kitakaname, Hiratsuka, Kanagawa 259–1292, Japan *2 Department of Biological Chemistry, College of Bioscience and Biotechnology, Chubu University, 1200 Matsumoto-cho, Kasugai, Aichi 487–8501, Japan *3 Graduate School of Science and Engineering, University of Toyama, 3190 Gofuku, Toyama 930–8555, Japan *4 Department of Pharmacy, Kindai University, 3-4-1 Kowakae, Higashi-Osaka 577–8502, Japan *5 GL Sciences Inc., 30F, Tokyo Square Tower, 6-22-1 Nishishinjuku, Shinjuku, Tokyo 163–1130, Japan

Racemic synephrine, which was transformed into diastereomers by derivatization with 2,3,4,6-tetra-O-acetyl-β-D- glucopyranosil isothiocyanate, was resolved by a reversed phase HPLC with UV detection at 254 nm. The total contents of synephrine enantiomers in citrus samples were exocarp > mesocarp > endocarp > sarcocarp, suggesting that synephrine content of outer side of citrus fruits was higher than that of the inner side. (R)-Synephrine was detected in exocarp of eleven fresh citrus fruits, except for , , and samples. (S)-Synephrine was determined in the exocarp of four citrus fruits (mikan, , , and samples) and the ratio of (S)-synephrine to total synephrine was 0.5 – 0.9%. The racemization of (R)-synephrine in aqueous solution during heating at 100°C was also examined. An increase in the heating time brought about an increase in the (S)-synephrine content in a linear fashion. The racemization was found to be significantly reduced by the addition of D-fructose, D-maltose, D-glucose, D-mannose or D-galactose, but not D-sucrose or D-mannitol. It is suggested that the reducibility of sugars may result in the inhibition of racemization.

Keywords Enantioseparation, synephrine, diastereomer, citrus, HPLC

(Received October 5, 2018; Accepted November 30, 2018; Advance Publication Released Online by J-STAGE December 14, 2018)

to those of humans. Bitter orange extracts, which contain Introduction synephrine, are widely used for weight loss/weight management, energy production, and sports performance. Synephrine is a Synephrine (Fig. 1), a phenethylamine alkaloid, is known to be chiral compound and its enantiomers have been shown to exert present in the peel and edible parts of Citrus fruits.1–5 Since different pharmacological activities on α- and β-adrenergic synephrine is structurally similar to adrenergic agonist, such as receptors. That is, (R)-synephrine is from 1 to 2 orders of , noradrenaline, and ephedrine, the effects of magnitude more active than its (S)-enantiomer.2,8 synephrine on the cardiovascular system are attributable to Synephrine has been analyzed by HPLC with an ultraviolet adrenergic stimulation. In general, vasoconstriction occurs detector,9–15 HPLC with mass spectrometry,11,16–18 and Raman when ligands bind to α-adrenergic receptors, while binding to spectrometry.19 It was reported that the amount of synephrine in β-1 adrenergic receptors result in cardiovascular contractility citrus fruits was varied according to the citrus , parts of and increased heart rate. Ligand binding to β-2 adrenergic fruits, and the maturation period.9–13,16 Arbo et al. revealed a receptors is associated with bronchodilation.6 However, variation on the content of synephrine in Citrus. sinensis synephrine binds much more poorly to α-1, α-2, β-1, and β-2 according to the maturation period and found that its content adrenergic receptors than other ligands, such as adrenaline.7 was inversely proportional to the size of the fruit.10 Synephrine Various studies have shown that synephrine binds to β-3 enantiomers in citrus fruit samples have also been separated by adrenergic receptors, resulting in an increase in the body’s HPLC with chiral columns.20–23 ability to breakdown fats. Binding to β-3 adrenergic receptors Direct and indirect methods have evolved as the main dose not influence heart rate or , but regulate lipid strategies for enantioseparation.24–26 A direct method, which and carbohydrate metabolism.5 According to Stohs,5 synephrine does not require chemical derivatization, is based on a chiral exhibits greater adrenergic receptor binding in rodents than in stationary phase or with a chiral selector in a mobile phase on humans, while data from animals cannot be directly compared an achiral stationary phase. An indirect method is based on the formation of diastereomers by derivatization of analyte † To whom correspondence should be addressed. enantiomers with a chiral reagent. Gal and Brown27 reported an E-mail: [email protected] indirect HPLC method for the chiral separation of an adrenergic 408 ANALYTICAL SCIENCES APRIL 2019, VOL. 35

Fig. 1 Reaction scheme of synephrine derivatization with TAG-ITC.

agent including synephrine based in derivatizing with spectrometer was used. LC/MS separation was performed on a 2,3,4,6-tetra-O-acetyl-β-D-glucopyranosil isothiocyanate (TAG- InertSustain C18 column (3 μm, 2.1 mm × 250 mm, GL ITC) under basic condition. But this indirect method was not Sciences) with a mobile phase consisting of acetonitrile/water/ applied for real samples. formic acid (40/60/0.1, v/v/v) at a flow rate of 0.2 mL/min at It was reported that (S)-synephrine in the peels of six citrus 40°C. ESI conditions (positive ion mode) were as follows: fruits, including mikan and orange, was not detected (less than drying gas temperature, 250°C; drying gas flow, 10 L/min; 1 mg/100 g) and that (R)-synephrine was the only enantiomer capillary voltage, 4 kV. 1H nuclear magnetic resonance (NMR) isolated from citrus fruits.22 The same results were obtained spectra were recorded on a JEOL ECS400 (400 MHz) 23 20 21 using mikan and bitter orange, but Pellati et al. determined spectrometer in CD3OD. Chemical shifts have been reported in (S)-synephrine in fresh fruits pulp of bitter orange, in which the δ ppm units with reference to the internal standard ratio of (S)- and (R)-synephrines was 7.6:92.4. They also tetramethylsilane (Si(CH3)4, 0.00 ppm). reported that the ratio of (S)-synephrine to total synephrine extended over 4.8 – 14.4% in Evodia fruits.28 Kusu et al.23 Preparation of racemic synephrine derivatized with TAG-ITC found both (S)- and (R)-synephrines were detected in two orange A stock solution of racemic synephrine (10 mM) was prepared juices and a . They suggested that (S)-synephrine with water, and stored at –15°C. TAG-ITC (20 mM) solution may possibly be formed during production steps. Thus, it was prepared with acetonitrile. Solutions of synephrine remains unclear whether (S)-synephrine is contained in citrus (100 μL), 200 mM phosphate buffer (pH 7.0) (200 μL), and fruits, or not. In order to clarify the question, we developed an acetonitrile (150 μL) and TAG-ITC (50 μL) were mixed. The HPLC method for the analysis of synephrine enantiomers in mixture was incubated at 40°C for 20 min and then the citrus fruits, citrus juices, and citrus products after synephrine synephrine derivative was analyzed by HPLC. was derivatized with TAG-ITC at neutral pH. We also studied the effect of sugars on the racemization of (R)-synephrine. Sample extraction and preparation Thirteen brands of citrus fruits, four brands of dried citrus fruits, four brands of citrus juices, two brands of orange Experimental marmalade, and a brand of canned mikan (Table 1) were purchased from local markets. A bitter orange (Citrus aurantium) Reagents and chemicals was kindly gifted by JA Aira Izu. Mikan (), orange Racemic synephrine and acetonitrile were obtained from (Citrus sinensis), and bitter orange (Citrus aurantium) were Sigma (St. Louis, MO, USA). Methanol was from Kanto divided into exocarp, mesocarp, endocarp, and sarcocarp. For Chemicals (Tokyo, Japan). (R)-(–)-Synephrine, racemic octo- other citrus fruits, exocarp was used. The citrus products were pamine, , and 2,3,4,6-tetra-O-acetyl-β-D-glucopyranosil not divided, whole products were used. Synephrine in each isothiocyanate (TAG-ITC) were from Tokyo Kasei (Tokyo, sample was extracted with a Microtec (Funabashi, Japan) Japan). Sodium dihydrogenphosphate dihydrate and other Physcotron homogenizer Model NS-52 in 10 mL of 30% chemicals (analytical grade) were obtained from FUJIFILM acetonitrile. Then, the mixture was centrifuged at 1200g for Wako Pure Chemical Corporation (Osaka, Japan). 5 min. The supernatant was filtered with a 0.2-μm filter and the filtrate was applied to a Sep-Pak Plus Short tC18 cartridge Apparatus for HPLC and mobile phase conditions (Waters, Milford, USA). The first nonbinding fraction The HPLC system consisted of a Jasco (Hachioji, Japan) (0 – 2 mL) was discarded and the next nonbinding fraction Model PU-2080 pump, a Jasco Model UV-2075 detector, a (2 – 3 mL) was collected. The nonbinding fraction was used as Rheodyne (Cotati, CA, USA) manual injector, a Shimadzu a sample solution for derivatization with TAG-ITC. (Kyoto, Japan) column oven Model CTO-10Avp, and a Shimadzu degasser Model DGU-14A. InertSustain C18 column (5 μm, 4.6 mm i.d. × 150 mm, GL Sciences, Tokyo, Japan) was Results and Discussion used. A mobile phase consisted of 20% acetonitrile, 20% methanol and 30 mM phosphate buffer (pH 7.0). Elution was Factors affecting chiral separation carried out at a flow rate of 1.0 mL/min at 40°C. Analytes were Generally, the amino group could be derivatized with detected at 254 nm. Data acquisition and processing were isothiocyanate under basic conditions. Gal and Brown reported conducted with a Chromato-PRO (Runtime Instrument, synephrine was derivatized with 2,3,4,6-tetra-O-acetyl-β-D- Kanagawa, Japan). For LC/MS analysis, an LC 7400 series glucopyranosil isothiocyanate (TAG-ITC) in the presence of (Hitachi) equipped with an Agilent 6140 quadrupole mass sodium bicarbonate. We also derivatized synephrine with ANALYTICAL SCIENCES APRIL 2019, VOL. 35 409

Table 1 Concentrations of (S)- and (R)-synephrines in each organ of fresh citrus fruit samples (R)-Synephrine (S)-Synephrine Name Species Organ (mg/100 g) (mg/100 g)

A Mikan Citrus unshiu Exocarp 84.6 ± 0.90a 0.75 ± 0.46 Mesocarp 72.7 ± 0.70 < LOQb Endocarp 33.4 ± 0.72 < LOQ Sarcocarp 6.5 ± 1.10 NDc B Orange Citrus sinensis Exocarp 24.1 ± 4.45 ND Mesocarp 15.1 ± 1.01 ND Endocarp 8.3 ± 0.22 ND Sarcocarp 1.9 ± 0.96 ND C Bitter orange () Citrus aurantium Exocarp 26.4 ± 0.87 < LOQ Mesocarp 17.4 ± 0.96 ND Endocarp 9.4 ± 0.96 ND Sarcocarp 2.1 ± 1.03 ND

a. Mean ± standard deviation, n = 3. b. Less than 0.25 mg/100 g. c. Less than 0.08 mg/100 g.

TAG-ITC at 40°C for 20 min in the presence of 0.4% (v/v) triethylamine. The derivatized synephrine was enantioseparated by reversed phase HPLC with an InertSustain C18 column using a mobile phase containing 40% (v/v) acetonitrile and 0.1% (v/v) phosphoric acid. However, many minor peaks were observed around peaks of synephrine enantiomers. This result was the same as that obtained using 10 mM tetraborate buffer (pH 9.2) instead of 0.4% (v/v) triethylamine. We tried to examine the derivatization under lower pH conditions using 80 mM phosphate buffer. The minor peaks were decreased with decreasing pH of derivatized media. A native synephrine peak was not observed above pH 7, and peak area of synephrine enantiomers at pH 7 or 8 was larger than that at pH 6. Therefore, pH 7 was selected for the derivatization of synephrine with TAG-ITC. The derivatized racemic mixture was applied to LC(ESI+)-MS. The correct mass of both ingredients gave peaks with protonation at m/z = 557, suggesting that (R)- and (S)- synephrines reacted with TAG-ITC at a ratio of 1:1. 1H NMR spectra of synephrine TAG-ITC adduct was compared with that Fig. 2 Effect of the mixing ratios of acetonitrile and methanol in of synephrine. Although most of the aryl protons of synephrine mobile phase on the enantioseparation of synephrine. HPLC was were not shifted, the N-methyl and N-methylene signals were conducted by using an InertSustain C18 column with a mobile phase shifted upfield by derivatization. It indicated that amino group consisting of 30 mM phosphate buffer (pH 7.0) and several ratios of of synephrine was reacted with TAG-ITC. acetonitrile and methanol at 40°C. R and S represent (R)-synephrine The effect of the reaction time (0 – 30 min) on the and (S)-synephrine, respectively. derivatization of synephrine with TAG-ITC was examined using a reaction mixture containing 1 mM racemic synephrine, 4 mM TAG-ITC, and 80 mM phosphate buffer (pH 7.0) at 40°C. Peak area of derivatized synephrine increased with an increase in in citrus fruit samples. It has been reported that and incubation time up to 20 min, and then kept constant up to tyramine were minor components in citrus peels.9,12–14,21 Using 30 min. Therefore, the incubation time was determined to be the above HPLC and derivatization conditions, three 20 min. phenethylamines (racemic octopamine, racemic synephrine and In our preliminary study, a mobile phase containing 40% (v/v) tyramine, each 0.3 mM) were analyzed (Fig. 3A). These acetonitrile and 30 mM phosphate buffer (pH 7.0) was used and phenethylamine derivatives were found to be well separated two peaks of synephrine enantiomers were well separated with the elution order of octopamine, synephrine, and tyramine, (Fig. 2A). However, when citrus samples were analyzed, matrix which order was the same as that of native phenethylamines peaks were overlapped with synephrine peaks. Thus, we reported by using reversed phase HPLC.9,12,13,21 Racemic investigated the effect of the mobile phase mixed with octopamine was not enantioseparated by the above HPLC acetonitrile and methanol on the separation and the retention method. Each peak area of racemic octopamine and tyramine time of derivatized synephrine (Fig. 2). Derivatized synephrine was about half of the total peak area of synephrine. The was enantioseparated by HPLC at any ratios of acetonitrile and derivatization rates of octopamine and tyramine were slower methanol concentrations. A mobile phase containing 20% (v/v) than that of synephrine, and this may be attributed to differences acetonitrile, 20% (v/v) methanol and 30 mM phosphate buffer between primary and secondary amines. The (R)-synephrine (pH 7.0) was used for further experiments considering the (0.5 mM), which is commercially available, was derivatized higher separation, a moderately short retention time of with TAG-ITC, and the derivative was analyzed by the proposed synephrine, and the better separation of contaminants contained method (Fig. 3B). Since the peak corresponding to the 410 ANALYTICAL SCIENCES APRIL 2019, VOL. 35

Fig. 3 Chromatograms of racemic octopamine, racemic synephrine, and tyramine (A), and (R)-synephrine (B) derivatized with TAG-ITC. HPLC was conducted by using an InertSustain C18 column with a mobile phase consisting of 20% acetonitrile, 20% methanol and 30 mM phosphate buffer (pH 7.0) at 40°C. R and S represent (R)- synephrine and (S)-synephrine, respectively. Fig. 4 Chromatograms of standard solution (A) and derivatized synephrine extracted from exocarp (B), mesocarp (C), endocarp (D), and sarcocarp (E) of mikan sample A in Table 1. HPLC was conducted by using an InertSustain C18 column with a mobile phase consisting of 20% acetonitrile, 20% methanol and 30 mM phosphate buffer (pH 7.0) (S)-synephrine derivative was not detected, it was found that no at 40°C. R and S represent (R)-synephrine and (S)-synephrine, racemization of (R)-synephrine occurred during derivatization respectively. process. Synephrine was subjected to the proposed HPLC method using the above optimum conditions. The limit of detection (LOD) of each synephrine enantiomer defined as a signal-to- (Citrus unshiu Citrus sinensis) and a ponkan (Citrus reticulate) noise ratio of 3 was 0.0005 mM (0.0836 mg/100 g) and the and the ratio of (S)-synephrine to total synephrine was limit of quantification (LOQ) of each enantiomer defined as a 0.5 – 0.9%. (S)-Synephrine was detected in exocarps of bitter signal-to-noise ratio of 10 was 0.0015 mM (0.251 mg/100 g). orange, , , and hassaku samples, but the contents Linearity (r2 > 0.999) was demonstrated in the concentration of their samples were significantly low (0.08 – 0.25 mg/100 g). range of 0.0015 – 1 mM by each standard curve (9 points) for Synephrine enantiomers in products of citrus fruits were also (R)- and (S)-synephrines. The repeatability of five consecutive analyzed (Table 3). (R)-Synephrine was detected in three dried determinations was evaluated at 0.005, 0.05, and 0.5 mM for citrus fruits, four citrus juices, two orange , and a (R)- and (S)-synephrine. Good repeatabilities of peak areas canned mikan, but (S)-synephrine was not detected. (RSD < 1.2%) and retention times (RSD < 0.2%) were obtained for both enantiomers. Interday repeatabilities of peak areas and The effect of sugars on the racemization of (R)-synephrine retention time in three days (n = 5, each day) were less than 0.4 Pellati et al.29 studied the racemization of synephrine and 0.3% RSD, respectively, at 0.5 mM for each enantiomer. enantiomer. They found that the racemizaton hardly occurred in When a standard solution (0.3 mL) containing 83.6 mg/L each buffer solutions at pH values between 3 and 8 at 80°C after 48 h enantiomer of synephrine (final concentration: 50.2 mg each (less than 10% isomerization) and showed a possible mechanism enantiomer per g samples) was added to mikan, orange, and of racemization of (R)-synephrine in acidic and basic media. bitter orange peel samples (0.5 g), recoveries of (R)- and (S)- We also examined the racemization of (R)-synephrine (0.5 mM) synephrines were between 93 and 107%. in aqueous solution at 100°C (Fig. 5). pH of the solution was 9.6. An increase in the heating time (2 h) brought about an Separation of synephrine enantiomers in citrus fruit samples increase in the ratio of (S)-synephrine to the total synephrine Kusu et al.23 studied an HPLC method using a chiral column enantiomers in a linear fashion. The total peak areas of (S)- and for synephrine in Citrus unshiu and reported that the synephrine (R)-synephrines were not affected during heating, showing no content was mesocarp > endocarp > exocarp > sarcocarp. On decomposition. Eroglu et al. reported the effect of thermal the other hand, Arbo et al.10 reported that its content was treatment on synephrine content of bitter along peel > albedo > pulp. The (S)- and (R)-synephrine contents in with ascorbic acid and sugar contents, where no significant citrus fruit samples were analyzed by the proposed HPLC effect on the total synephrine content was obserbed.30 It has method (Table 1). The representative chromatograms are shown been reported that orange juices contain 4 – 6% (w/v) sucrose, in Fig. 4. The total contents of synephrine enantiomers in 2 – 3% (w/v) glucose, and 2 – 3% (w/v) fructose.31–33 When mikan were exocarp > mesocarp > endocarp > sarcocarp, and 0.5 mM (R)-synephrine was heated at 100°C for 2 h in the the same results were also obtained with orange and bitter presence of 5% (w/v) D-sucrose, the production of (S)- orange samples, suggesting that synephrine content of outer side synephrine by the racemization slightly decreased. However, of citrus fruits was higher than that of the inner side. Synephrine the racemization was found to be significantly reduced by the was detected in the exocarp of eleven citrus fruits except for addition of D-glucose or D-fructose instead of sucrose. Since lemon, lime, and grapefruit (Tables 1 and 2). (S)-Synephrine both D-glucose and D-fructose are reducing sugars, the was detected in exocarp of a mikan (Citrus unshiu), two reducibility of sugars may result in the inhibition of the ANALYTICAL SCIENCES APRIL 2019, VOL. 35 411

Table 2 Concentrations of (S)- and (R)-synephrines in the exocarp of fresh citrus fruit samples (R)-Synephrine (S)-Synephrine Name Species (mg/100 g) (mg/100 g)

D Citrus fruit hybrid () Citrus unshiu × Citrus sinensis 64.7 ± 0.92a 0.31 ± 2.73 E Citrus fruit hybrid () Citrus unshiu × Citrus sinensis 92.6 ± 1.28 0.64 ± 0.97 F Ponkan Citrus reticulate var poonensis 64.5 ± 0.60 0.47 ± 2.03 G Iyokan Citrus iyo 50.4 ± 0.76 < LOQb H Dekopon Citrus dekopon 25.3 ± 1.04 < LOQ I Hassaku Citrus hassaku 12.2 ± 0.36 < LOQ J Amanatsu Citrus natsudaidai 8.8 ± 1.30 ND K Citrus junos 12.1 ± 1.03 ND L Lemon Citrus limon ND ND M Lime Citrus aurantifolia ND ND N Grapefruit Citrus paradisi ND ND

a. Mean ± standard deviation, n = 3. b. Less than 0.25 mg/100 g. c. Less than 0.08 mg/100 g.

Table 3 Concentrations of (S)- and (R)-synephrines in citrus fruit products (R)-Synephrine (S)-Synephrine Sample (mg/100 g) (mg/100 g)

O Dried citrus fruits Round slice of orange 1.6 ± 1.02a NDb P Round slice of citrus fruit hybrid () 3.5 ± 1.20 ND Q Orange peel 2.1 ± 1.31 ND R Segment of mikan 3.3 ± 0.98 ND S Juices Orange 2.5 ± 1.03 ND T Orange 0.5 ± 0.60 ND U Mikan 4.7 ± 1.77 ND V Mikan + orange 2.2 ± 0.56 ND W Orange marmalade 3.6 ± 0.56 ND X 0.7 ± 0.20 ND Y Canned mikan 1.4 ± 0.66 ND

a. Mean ± standard deviation, n = 3. b. Less than 0.08 mg/100 g.

racemization of (R)-synephrine was studied (Fig. 6). Five reducing sugars (D-fructose, D-maltose, D-glucose, D-mannose, and D-galactose) suppressed the racemization of (R)-synephrine, but two non-reducing sugars (D-sucrose and D-mannitol) did not affect the racemization. Among the five sugars, D-fructose and D-maltose showed the strong suppression at lower concentration less than 30 mM compared to other three sugars. Further studies are necessary to clarify the mechanism. In conclusion, using a derivatization with TAG-ITC, an HPLC method for separation of the synephrine enantiomers was developed. Generally, primary amino group could be derivatized with isothiocyanate under basic conditions. However, the optimum pH for the derivatization of synephrine, having a secondary amino group, with TAG-ITC was found to be 7. Using the proposed method, synephrine enantiomers in 14 citrus fruits and 11 citrus fruit products were analyzed. A small amount of (S)-synephrine was found to be detected in the Fig. 5 Effect of heating on the racemization of (R)-synephrine. exocarp of a mikan (Citrus unshiu), two tangors (Citrus unshiu (R)-Synephrine (0.5 mM) (closed circles) or that with 5% of D-sucrose Citrus sinensis) and a ponkan (Citrus reticulate), where the ratio (open square), D-glucose (open triangle) or D-fructose (open circle) were heated at 100°C, and then, derivatized with TAG-ITC. HPLC of (S)-synephrine to total synephrine was less than 1%. The was conducted by using an InertSustain C18 column with a mobile effect of sugars on the racemization of (R)-synephrine during phase consisting of 20% acetonitrile, 20% methanol and 30 mM heating at 100°C was also studied. D-Sucrose and D-mannitol phosphate buffer (pH 7.0) at 40°C. did not affect the racemization. However, an increase in the concentration of D-fructose, D-maltose, D-glucose, D-mannose or D-galactose caused a decrease in the racemization. Therefore, it is suggested that the reducibility of sugars may result in the racemization. In order to confirm this suggestion, the effect of inhibition of racemization. the concentrations of several sugars (3 – 100 mM) on the 412 ANALYTICAL SCIENCES APRIL 2019, VOL. 35

10. M. D. Arbo, E. R. Larentis, V. M. Linck, A. L. Aboy, A. L. Pimentel, A. T. Henriques, E. Dallegrave, S. C. Garcia, M. B. Leal, and R. P. Limberger, Food Chem. Toxicol., 2008, 46, 2770. 11. K. Miyagi, T. Fujise, N. Koga, K. Wada, M. Yano, and H. Ohta, Food Sci. Technol. Res., 2009, 15, 389. 12. D. He, Y. Shan, Y. Wu, G. Liu, B. Chen, and S. Yao, Food Chem., 2011, 127, 880. 13. R. M. Uckoo, G. K. Jayaprakasha, S. D. Nelson, and B. S. Patil, Talanta, 2011, 83, 948. 14. C. D. Lorenzo, A. D. Santos, F. Colombo, E. Moro, and M. Dell’Agli, Food Control, 2014, 46, 136. 15. P. Li, S.-L. Zeng, L. Duan, X.-D. Ma, L.-L. Dou, L.-J. Wang, P. Li, Z.-M. Bi, and E-H. Liu, J. Chromatogr. A, 2016, 1469, 96. 16. M. Bader, T. Lang, R. Lang, and T. Hofmann, J. Agric. Food Chem., 2017, 65, 4853. 17. P. A. S. Tette, L. R. Guidi, E. M. A. F. Bastos, and C. Fig. 6 Effect of concentrations of sugars on the racemization of (R)- Fernandes, Food Chem., 2017, 229, 527. synephrine by heating. (R)-Synephrine (0.5 mM) with 3 – 100 mM 18. P. Paiga, M. J. E. Rodrigues, M. Correia, J. S. Amaral, M. sugars were heated at 100°C, and then, derivatized with TAG-ITC. B. P. P. Oliveira, and C. Delerue-Matos, Eur. J. Pharm. Sci., HPLC was conducted by using an InertSustain C18 column with a 2017, 99, 219. mobile phase consisting of 20% acetonitrile, 20% methanol and 30 mM 19. Y. Zhang, C. Zhao, G. Tian, C. Lu, Y. Li, L. He, H. Xiao, phosphate buffer (pH 7.0) at 40°C. D-Mannitol (closed squares); and J. Zheng, Food Chem., 2018, 240, 743. D-sucrose (open squares); D-galactose (open rhombs); D-mannose (closed rhombs); D-glucose (open circles); D-maltose (closed triangles); 20. F. Pellati, S. Benvenuti, and M. Melrgari, J. Pharm. Biomed. D-fructose (open triangles). Anal., 2005, 37, 839. 21. F. Pellati, S. Benvenuti, M. Melrgari, and F. Firenzuoli, J. Pharm. Biomed. Anal., 2002, 29, 1113. 22. F. Kusu, K. Matsumoto, and K. Takamura, Chem. Pharm. Bull., 1993, 43, 1158. References 23. F. Kusu, K. Matsumoto, K. Arai, and K. Takamura, Anal. Biochem., 1996, 235, 191. 1. F. Pellati and S. Benvenuti, J. Chromatogr. A, 2007, 1161, 24. N. R. Srinivas, Biomed. Chromatogr., 2004, 18, 207. 71. 25. R. Bhushan and H. Bruckner, Amino Acids, 2004, 27, 231. 2. S. J. Stohs and H. G. Preuss, J. Funct. Foods, 2012, 4, 2. 26. I. Ilisz, R. Berkecz, and A. Peter, J. Pharm. Biomed. Anal., 3. J.-H. Kim and Y. Park, J. Exerc. Nutr. Biochem., 2016, 20, 2008, 47, 1. 20. 27. J. Gal and T. R. Brown, J. Pharmacol. Methods, 1986, 16, 4. N. Bakhiya, R. Ziegenhagen, K. I. Hirsch-Ernst, B. 261. Dusemund, K. Richter, K. Schultrich, S. Pevny, B. Schafer, 28. F. Pellati, S. Benvenuti, F. Yoshizaki, and M. Melegari, J. and A. Lampen, Mol. Nutr. Food Res., 2017, 61, 1601020. Sep. Sci., 2006, 29, 641. 5. S. J. Stohs, Phytother. Res., 2017, 31, 1463. 29. F. Pellati, G. Canazza, and S. Benvenuti, J. Chromatogr. A, 6. M. A. Inchiosa Jr., J. Obesity, 2011, doi:10.1155/2011/764584. 2010, 1217, 3503. 7. S. J. Stohs and V. A. Badmaev, Phytother. Res., 2016, 30, 30. E. C. Eroglu, Z. Baysal, M. Unlu, U. Seday, and Z. 732. Karasahin, Indian. J. Pharm. Educ. Res., 2017, 51, S412. 8. C. M. Brown, J. C. McGrath, J. M. Midgley, A. G. B. Muir, 31. H. Kelebek, S. Selli, A. Canbas, and T. Cabaroglu, J. W. O’Brien, C. M. Thonoor, C. M. Williams, and V. G. Microchem. J., 2009, 91, 187. Wilson, Br. J. Pharmacol., 1988, 93, 417. 32. H. L. Peters, K. E. Levine, and B. T. Jones, Anal. Chem., 9. F. Pellati and S. Benvenuti, J. Chromatogr. A, 2007, 1165, 2001, 73, 453. 58. 33. J. Velterop and F. Vos, Phytochem. Anal., 2001, 12, 299.