Analysis of Phenolic Compounds in Rooibos Tea ... - ACS Publications

Subscriber access provided by the Henry Madden Library | California State University, Fresno

Article

Analysis of phenolic compounds in rooibos tea (Aspalathus linearis) with a comparison of flavonoid-based compounds in natural populations of plants from different regions Marietjie Stander, Ben-Erik Van Wyk, Malcolm John C. Taylor, and Helen Selma Long J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.7b03942 • Publication Date (Web): 24 Oct 2017 Downloaded from http://pubs.acs.org on October 24, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Agricultural and Food Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 35

Journal of Agricultural and Food Chemistry

Analysis of phenolic compounds in rooibos tea (Aspalathus linearis) with a comparison of flavonoid-based compounds in natural populations of plants from different regions Maria A. Stander*,a,c, Ben-Erik Van Wykb, Malcolm J.C. Taylorc, Helen S. Longb a

Department of Biochemistry, University of Stellenbosch, Private Bag X1, Matieland 7600,

South Africa. b

Department of Botany and Plant Biotechnology, University of Johannesburg, P.O. Box 524,

Auckland Park 2006, Johannesburg, South Africa. c

Mass Spectrometry Unit, Central Analytical Facility, University of Stellenbosch, Private Bag

X1, Matieland 7600, South Africa.

*Corresponding author,(Tel: +27 21 808 5825; Fax: +27 21 808 5863; E-mail:[email protected]);

ACS Paragon Plus Environment

1


Page 2 of 35

1

ABSTRACT: Tea samples from 17 populations of “wild tea” ecotypes Aspalathus linearis

2

(rooibos tea), and two populations of A. pendula were analyzed. Recent advances in column

3

technology together with high-resolution mass spectrometry were applied to improve resolution,

4

facilitating the identification of several new compounds, as well as grouping of the wild tea

5

ecotypes according to their chemical composition. The collisional cross section (CCS) data

6

obtained from ion mobility-mass spectrometry is reported for the flavonoids in rooibos for the

7

first time. Enzyme pathways for the synthesis of the unique flavonoids found in rooibos tea are

8

also proposed. Aspalathus linearis and A. pendula produce similar combinations of main

9

phenolic compounds, with no diagnostically different discontinuities between populations or

10

species. Northern resprouters (Gifberg and Nieuwoudtville) contain higher PPAG levels whilst

11

teas from Wupperthal and surrounding areas were found to contain unique dihydrochalcones

12

(phloridzin and a sieboldin analog) are reported here for the first time.

13 14 15

KEYWORDS: Aspalathus linearis, rooibos tea, flavonoids, ecotypes, wild tea profiling, liquid chromatography-mass spectrometry

16 17 18 19


2

Page 3 of 35


20

Introduction

21

The aim of this study was a comprehensive characterization of the phenolic constituents, not

22

only of the commercial (cultivated) form of rooibos tea but of the entire species complex known

23

botanically known as Aspalathus linearis (Burm.f.) Dahlg. The phenolic profile of commercial

24

rooibos tea has been thoroughly investigated, but the chemical compositions of numerous wild

25

tea types belonging to this species complex have remained poorly explored.

26

Aspalathus linearis exists as at least eight distinct sub-forms or ecotypes that have developed

27

due to geographic isolation and various mutations conferring survival advantage in response to

28

differing environments, especially exposure to fire1-3. Four of the ecotypes are reseeders (also

29

called seeders) – single-stemmed plants that are destroyed by fire and that recruit from seeds in

30

the soil seed bank, whilst four others are resprouters (also called sprouters) – multi-stemmed

31

plants arising from a persistent underground lignotuber, which regenerates (coppices) after

32

fire3,4. The ecotypes are also ecologically distinct, with some ecotypes preferring habitats at

33

higher elevations (>400–600m), higher rainfall sites in the southern areas of the range, or lower

34

rainfall sites in the northern areas2. The eight main geographical variants or ecotypes have been

35

described and illustrated by Van Wyk and Gorelik3 and comprise (1) southern sprouter; (2) grey

36

sprouter; (3) northern sprouter; (4) Nieuwoudtville sprouter; (5) red type (seeder); (6) black type

37

(seeder); (7) Wupperthal type (seeder) and the (8) tree type (seeder).

38

Commercial rooibos tea, known as the Red type or Rocklands type, is a seeder with an upright

39

growth form that was selected and developed as a new crop plant by Nortier in the early part of

40

the 20th century (ca. 1920–1930). The selection criteria were taste, tea colour, growth rate, and

41

seed production2. Initially seeds were collected from the Pakhuis region and scarification of the


3


Page 4 of 35

42

seed coat was found to be necessary for successful germination. Whilst the Red type is farmed

43

extensively and represents a major industry in the Cedarberg region, there is also significant

44

small scale wild-harvesting (wild-crafting) of various wild types, especially in the Wupperthal

45

and Nieuwoudtville areas of the A. linearis range1. The main flavonoids in the leaves of rooibos

46

have been identified as aspalathin, nothofagin, orientin, isoorientin, vitexin, isovitexin,

47

isoquercetin, quercetin and rutin5-8. De Beer et al. 9 found that aspalathin and nothofagin

48

production in unfermented rooibos show slight seasonal variations, with the highest

49

concentration in mid-spring to early summer. The other flavonoids in the study did not vary

50

significantly. This agrees with a study by Joubert et al. 10 where 209 commercial fermented teas

51

of the Red type (Rocklands type) from different production regions were analysed. It was found

52

that the phenolic and phenylpropenoic acid glucoside (PPAG) content were similar with slight

53

differences between production years. PPAG is another unique compound only found in rooibos.

54

The goal of the current study was to re-examine the phenolic profile of the wild type teas using

55

high resolution chromatographic and mass spectrometric systems in order to gain a more

56

profound understanding of the phenolic diversity in A. linearis. This knowledge is not only of

57

academic interest but may also have practical value in developing or improving quality control

58

protocols and in identifying various chemotypes that can potentially be developed and

59

commercialised as unique geographical indications with novel flavours and health benefits.

60 61


4

Page 5 of 35


62

Materials and methods

63

Sampling

64

In an attempt to overcome problems due to sample non-homogeneity, representative samples

65

were taken from 10 randomly selected individual plants of each of the 17 wild-growing natural

66

populations of rooibos tea (Aspalathus linearis), together with two samples from the closely

67

related A. pendula (Table 1). These samples were combined and processed using the standard

68

methods for processing rooibos tea. They were cut, mixed, moisturised, bruised, oxidised in

69

sunlight and then rapidly dried. The vouchers specimens have been deposited in the University

70

of Johannesburg Herbarium (JRAU). Figure 1 depicts the locations of the various tea

71

populations that were sampled.

72

Extraction

73

Approximately 2 grams of dry tea was ground in a pestle and mortar and extracted with 50%

74

methanol in water containing 1% formic acid (15 ml) in a 50 mL polypropylene centrifuge tube

75

by soaking it overnight, followed by extraction in an ultrasonic bath (0.5 Hz, Integral systems,

76

RSA) for 60 minutes at room temperature. The extracts were centrifuged (Hermle Z160m, 3000

77

x g for 5 minutes) and transferred to vials.

78


5


Page 6 of 35

79

Standards

80

Standards were obtained from Sigma-Aldrich and from a kind donation by D. De Beer of the

81

Agricultural Research Council (ARC). Stock solutions were prepared quantitatively in cocktails

82

ranging from 1 to 100 µg/mL in concentration. Four different cocktails were prepared at each

83

level to enable isomers and compounds with similar elemental formulae to be distinguished. The

84

solvent used for preparation of the cocktails was 50% methanol in water containing 1% formic

85

acid, i.e., the same as that used for extraction of the tea samples.

86

LCMS analysis

87

A Waters Synapt G2 Quadrupole time-of-flight (QTOF) mass spectrometer (MS) connected to a

88

Waters Acquity ultra performance liquid chromatograph (UPLC) (Waters, Milford, MA, USA)

89

was used for high resolution UPLC-MS analysis. Electrospray ionization was used in negative

90

mode with a cone voltage of 15 V, desolvation temperature of 275 °C, desolvation gas at 650

91

L/hr and the rest of the MS settings optimized for best resolution and sensitivity. Data were

92

acquired by scanning from m/z 150 to 1500 in resolution mode as well as in MSE mode. In MSE

93

mode two channels of MS data were acquired, one at a low collision energy (4 V) and the second

94

using a collision energy ramp (40 to 100 V) to obtain fragmentation data as well. Leucine

95

enkaphalin was used as lock mass (reference mass) for accurate mass determination and the

96

instrument was calibrated with sodium formate. Separation was achieved on a Waters HSS T3,

97

2.1 x 100 mm, 1.7 µm column. An injection volume of 2 µL was used and the mobile phase

98

consisted of 0.1% formic acid (Solvent A) and acetonitrile containing 0.1 % formic acid as

99

solvent B. The gradient started at 100% solvent A for 1 minute and changed to 28 % B over 22

100

minutes in a linear way. It then went to 40% B over 50 seconds and a wash step of 1.5 minutes


6

Page 7 of 35


101

at 100% B, followed by re-equilibration to initial conditions for 4 minutes. The flow rate was

102

0.3 mL/min and the column temperature was maintained at 55 ºC.

103

Ion mobility data was obtained using the same UPLC gradient and column as above and IMS

104

Wave velocity was set at 332 m/s and wave height at 20.2 V. Polyalanine was used for the

105

calibration and calculations and the rest of the settings were according to Rautenbach et al.

106

(2017).11

107 108 109 110

Results and discussion Sample preparation and Extraction

111

Samples were fermented to compare the compounds found with those present in commercial

112

rooibos tea (Red type). Walters et al. 12 have done a large amount of work to optimize the

113

extraction of rooibos tea. An extraction solvent of 50% methanol with 1% formic was used in

114

this study since it showed a slightly higher extraction efficiency for the more polar, early eluting

115

molecules (as well as PPAG and aspalathin) than the 40% acetonitrile solvent of Walters12. This

116

solvent is also more environmentally friendly than acetonitrile. The extraction solvent of Walters

117

was better for the more non-polar and later eluting molecules (when using sonication) (Figure

118

1S).

119

LCMS method

120

There are a few critical separations that have to be achieved when analyzing the phenolic

121

compounds in rooibos using liquid chromatography (Walters et al. 12), since several compounds

122

are present with the same nominal mass. Figure 2 shows the extracted mass chromatograms at

123

m/z 447, 449 and 609 representing separations between orientin/isorientin (447), R/S-eriodictyol-


7


Page 8 of 35

124

6-C-glucopyranoside and R/S-eriodictyol-8-C-glucopyranoside (449), quercetin-3-O-

125

robinobioside and rutin (609). This separation improves if the photo diode array (PDA) detector

126

is taken out of the system since peak broadening caused by the additional path length and flow

127

cell is eliminated (Figure 2S). The aim of the study was to compare the relative concentrations

128

of the flavonoids to each other in the various tea samples. It was noted that the extracted mass

129

chromatograms of the standards contained in our cocktails showed similar responses in ESI

130

negative mode whilst the peak areas produced by the UV detector were more variable due to the

131

compounds exhibiting different UV maxima and absorbance at set wavelengths (Figure 3S).

132

Integration of the peaks in the extracted mass chromatograms [M-H]-1 for the standards (Figure

133

3S) for example vicenin-2 (compound 22), iso-orientin (30), vitexin (36), isoquercitrin (40),

134

kaempferol glucoside (47) and phloridzin (51) resulted in peak areas that were very close in

135

magnitude, showing a relative standard deviation (RSD) of only 4.5% at an injected

136

concentration of 100 µg/mL. It was thus decided to use rutin as reference calibrant for

137

compounds where no standards were available and to quantify compounds based on the areas of

138

their extracted mass chromatograms, since the UV response was shown to be less selective and

139

less sensitive (Figure 4S). A typical commercial rooibos tea sample was used as a control

140

sample at the beginning and end of the run. The reproducibility of the extraction method was

141

tested by extracting the same commercial tea sample five times and relative standard deviations

142

(RSDs) between 6 and 13% were observed for the six main peaks (data not shown). The

143

intraday variation was determined by reinjecting the same sample extracts a week later. The

144

variation was between 16.5 and 29%, with aspalathin at 16.5 and the eriodictyol-C-b-D-

145

glucopyranosides at the higher end. This is due to the instability of flavonoids in solution, all

146

samples was analyzed for quantitative purposes in a 24 hour period.


8

Page 9 of 35


147

Table 2 contains the compounds that were detected in the rooibos tea samples. Many of these

148

constituents were previously identified by other authors (Joubert5, Krafczyk et al.6, Iswaldi et al.

149

7

150

fragments, UV maxima and relative retention times reported in these previous papers.

151

The collisional cross section (CCS) values were validated against those of Gonzales et al. 13. All

152

of the CCS values that were also measured in the paper of Gonzales et al. 13 corresponded with

153

the values in this paper. For example, a value of 230.6 Ȧ2 was reported for rutin; our value was

154

230.8 Ȧ2. The highest deviation from Gonzales et al. 13 was found in the smaller molecules, but

155

it was still within 2.2% of what they reported. There are a few compounds in Table 2 that has

156

more than one CCS value. Figure 3 shows an extracted mass chromatograms (C) of m/z 449,

157

447, 451 that corresponds to compounds 20,23,26 and 27 the eriodictyol glucopyranosides (m/z

158

449) as well as m/z 451 (aspalathin) and 447 (orientin, isoorientin). It also shows the 2D ion

159

mobilogram on top showing the peaks corresponding to the eriodictyol glucopyranosides having

160

two spots (B) and a mobilogram of compound 20 showing two peaks at two drift times.

161

Something similar was observed for chlorogenic acid. Kuhnert et al., 2015 reported this

162

phenomenon as a differentiation of the prototropic ions, which means that the ions can carry their

163

charge on two different positions that results in different shapes of the molecules (ions) and thus

164

different collisional cross sections 14.

165

Interestingly, no naringenin was detected in any of the samples, neither as the chalcone nor as the

166

flavanone. This indicates that naringenin, which represents the starting point of the classical

167

flavonoid pathway, is rapidly converted either to eriodictyol or to one of the dihydroflavonols

168

(Figure 4). Another possibility is that it may not be formed at all due the dominance of the

169

alternate pathway involving the production of phloretin via the carbon double bond reductase

and Walters et al.12) and were identified in this study according to their accurate mass, MSMS


9


Page 10 of 35

170

(CDBR) enzyme (Figure 5). A number of new compounds were detected, with tentative

171

identifications being made based on the accurate mass determinations and comparison of the

172

elemental compositions to databases such as Metlin and Chemspider. If possible, further

173

confirmation was achieved by injecting reference standards. These compounds include 3-O-

174

caffeoylquinic acid (compound 15, chlorogenic acid), catechin (compound 13), citric acid

175

(compound 1), dihydroxy benzoic acid (compound 5), epicatechin (compound 19), p-coumaric

176

acid (compound 29) and hesperetin (compound 58) (Figure 5S). All except citric acid were

177

minor constituents. Phloretin-2-β-D-glucoside (compound 51, phloridzin) and phloretin-4-O-

178

glucoside (compound 54, trilobatin) are newly identified compounds that were present in some

179

of the tea samples and these determinations were confirmed with reference standards. A new

180

dihydrochalcone glycoside (compound 44) that has the same accurate mass (m/z 451.1237, UV

181

maximum (284 nm) and MS fragment ions (m/z 289,167,123) as sieboldin was also detected in

182

some of the samples (Figure 6S). The retention time and the CCS value (Table 2) did not

183

correspond to the sieboldin standard and it was tentatively identified as an isomer of sieboldin.

184

Since sieboldin is an O-glucoside of luteolin dihydrochalcone, this compound is likely to also be

185

a glucoside with the glucose attached at another position, most probably at the same position as

186

in phloridzin, which is the 2’-β-D-glucoside (Figure 6).

187

In some instances where authentic standards were not available in our laboratory, the tentative

188

identifications were made from accurate mass and elemental composition with further

189

confirmations using MS fragmentation spectra from literature. Compounds identified in this way

190

were α-1-caffeoylglucose (compound 11) and piscidic acid (compound 6, Figure 7S). α-1-

191

Caffeoylglucose (M-H of 341.0873) gave a fragment at m/z 179, indicative of caffeic acid and

192

also a neutral loss of 162, characteristic of the loss of a hexose, its secondary fragment ion of m/z


10

Page 11 of 35


193

135 indicated that the position of the glucose is most probably at position 1 according to Jaiswal

194

et al.15. Piscidic acid was confirmed by its fragment ions m/z 165,193,179, and its accurate mass.

195

Furthermore, it eluted early in the run, similar to what Mata et al. 16 has found in Opuntia juice

196

(Figure 7S). Piscidic acid belongs to the family of phenylpyruvic acid derivatives.

197

The compound (52) that eluted at 20.77 minutes had an accurate mass of 493.1331 and fragment

198

ions similar to those of aspalathin (which eluted earlier). It had a mass difference of m/z 42,

199

characteristic of an acetyl group. The compound, was tentatively identified as an acetylated form

200

of aspalathin, most probably an acetylglucoside.

201

The compound (46) that eluted at 18.9 minutes had a molecular ion of m/z 469.1345 (C21H25O12)

202

with the same fragments as PPAG (M-H: C15H17O8) (325,163,119) and was thus tentatively

203

identified as a PPAG derivative (Figure 8S). The compound (18) that eluted at 13.0 minutes had

204

a molecular ion of m/z 327.107 (M-H: C15H19O8) and therefore differs with two protons from

205

PPAG, as does its first fragment (165,147,103). It has a similar UV spectrum to that of PPAG

206

and was thus tentatively identified as a PPAG derivative, most probably the phenyl propanoic

207

glucoside.

208 209

Comparison between samples

210

Initially the data was subjected to principle component analysis, but no clear groupings were

211

observed, except for some outliers that had higher levels of PPAG (compound 29, sample 18 and

212

19, Data in Table 3, Figure 9S, 10S, 11S).

213


11


Page 12 of 35

214

Upon closer inspection of the list of major compounds in the tea in Table 2, it is clear that there

215

are groups of structurally related compounds (Figure 6, 7 and 8). The first group is related to the

216

dihydrochalcone 3-hydroxyphloretin (produced from phloretin via chalcone-3-hydroxylase), that

217

has been glycosylated via a C or O bond on different positions and with different sugars, of

218

which glucose is the most abundant. The second group is related to phloretin dihydrochalcone

219

and includes vitexin (compound 36), isovitexin (compound 39) and vicenin-2 (compound 22).

220

Phloridzin (compound 51) is an O-glucoside and nothofagin (compound 45) a C-glucoside of

221

phloretin.

222

A probable pathway from 3-hydroxyphloretin via eriodictyol to quercetin (compound 56) is

223

shown in Figure 8. This constitutes the third group of related compounds. In addition to

224

quercetin itself, three different glycosylated quercetin derivatives were detected.

225

The fourth group is related to coumaric acid (compound 28) and include E and Z

226

phenylpropenoid acid glucoside (compounds 16,29, PPAG) and piscidic acid (compound 6). In

227

the shikimic pathway, shikimic acid is converted to chorismate and the pathway is followed to

228

either phenylalanine or tyrosine. Phenylalanine is converted to phenyl propenoic acid which is

229

hydroxylated to hydroxy phenyl propenoic acid. We postulate that rooibos has an enzyme that

230

can glycosylate hydroxy phenyl propenoic acid to form phenyl propenoic acid glucoside and its

231

stereoisomer E-PPAG. When the concentrations (in mMol) of these four main groups were

232

added up (after the concentrations were converted to mM), the relations between some of the

233

samples from the same areas became clear. Figure 9 shows that there is still a large degree of

234

variation in the different flavonoid profiles. The northern resprouters (samples 18 and 19, the

235

Gifberg and Niewoudtville populations) showed much higher levels of PPAG when compared to

236

plants from other areas. This possibly points to some sort of drought tolerance strategy


12

Page 13 of 35


237

developed by the plant to survive in the low rainfall (

Analysis of Phenolic Compounds in Rooibos Tea ... - ACS Publications

Recommend Documents