Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES
Article
Investigation on Key Molecules of HLB Induced Orange Juice Off-flavor Johannes Kiefl, Birgit Kohlenberg, Anja Hartmann, Katja Obst, Susanne Paetz, Gerhard E. Krammer, and Stephan Trautzsch J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.7b00892 • Publication Date (Web): 11 Mar 2017 Downloaded from http://pubs.acs.org on March 17, 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 34
Journal of Agricultural and Food Chemistry
1
Investigation on Key Molecules of HLB Induced Orange Juice Off-flavor
Johannes Kiefl1*, Birgit Kohlenberg1, Anja Hartmann1, Katja Obst1, Susanne Paetz1, Gerhard Krammer1, Stephan Trautzsch1
1
Symrise AG, Flavors Division Research & Technology, P.O. Box 1253, D-37601
Holzminden, Germany
*
Author to whom correspondence should be addressed:
Phone:
+49 (0) 5531/901099
E-Mail:
[email protected] ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
Page 2 of 34
2 1
ABSTRACT
2
Orange fruits from Huanglongbing (HLB) infected trees do not fully mature and show
3
a severe off-flavor described as bitter-harsh, metallic and less juicy and fruity. The
4
investigation of juice from HLB infected (HLBOJ) and healthy control oranges (COJ)
5
by gas chromatography mass spectrometry showed higher concentrations of fruity
6
esters such as ethyl butyrate and ethyl 2-methylbutyrate, and soapy-waxy alkanals
7
such as octanal and decanal in the COJ whereas the HLBOJ showed higher
8
concentrations of green aldehydes like hexanal and degradation compounds of
9
limonene and linalool like α-terpineol. Application of aroma extract dilution analysis
10
on terpeneless peel oil led to the identification of long-chained aldehydes such as
11
(E,E)-2,4-decadienal,
12
decenal and octanal with the highest flavor dilution factors among 25 odor-active
13
volatiles in the peel oil of healthy oranges.
14
Taste guided fractionation and identification of the HLBOJ secondary metabolites
15
followed by sensory validation revealed that flavanoids such as hesperidin may
16
modulate the flavor to evoke the unacceptable harsh, metallic taste impression.
17
Quantitation of the bitter components showed good correlation between the limonoid
18
and flavanoid concentration with the off-flavor and quality of the oranges obtained
19
throughout the season.
(Z)-8-tetradecenal,
trans-4,5-epoxy-(E)-2-decenal,
20 21
KEYWORDS
22
Huanglongbing (HLB), Citrus Greening, Citrus sinensis, off flavor, gas
23
chromatography olfactometry, LC Taste®, limonin, flavanoids
ACS Paragon Plus Environment
(Z)-4-
Page 3 of 34
Journal of Agricultural and Food Chemistry
3 24
INTRODUCTION
25
The Florida orange industry is facing one of the most serious challenges in
26
recent decades. The orange crop is reported to be continuously decreasing from
27
104.7 million boxes in 2013-2014 to 96.8 million boxes in 2014-2015 and down to
28
81.6 million boxes in the 2015-2016 season.1,
29
infection of orange trees with the bacteria Candidatus Liberibacter ssp. which is
30
transmitted by the Asian citrus psyllid, Diaphorina citri.3 Orange fruits from infected
31
trees may drop before harvest and do not get ripe thus showing a severe off-flavor
32
described as bitter, harsh, metallic and less juicy and fruity.4, 5 Even after sorting out
33
infected small fruits before juicing, a reasonable amount of infected regular sized
34
fruits can enter the process sufficient to negatively impact the flavor.
35
2
This decline is caused by the
Approaches to prevent HLB infection by psyllid control and spraying of
36
insecticides,6,
7
nutritional foliar supplementation8,
9
37
spraying antibiotics10 are being intensely studied at the moment. These approaches
38
aim at preventing the crop decline while increasing the costs for grove management.
39
At the same time, optimization of the juice sensory quality is not targeted.
and curing of infected trees by
40
The development of resistant rootstocks is a promising field of research,11
41
however, recultivation of Florida orange groves with resistant rootstocks is estimated
42
to take 5-10 years. Orange trees are genetically modified to develop resistance as
43
well, however, regulation issues arise with selling this juice and byproducts from juice
44
production, for example, to European markets. Therefore, no short-term solution is
45
available to stop crop decline and to remove the off-flavor of orange juice and
46
products thereof yet.
47
Sensory analysis of fruits obtained in the 2007 and 2008 season in Florida
48
revealed that juice from infected trees is lower in orange, fatty and fresh flavor notes
49
and sweet taste but higher in bitter, metallic, pungent/peppery, salty/umami and ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
4 50
astringent flavor notes and taste. This was found to be more significant for juice
51
produced from symptomatic fruits of early season Hamlin rather than late season
52
Valencia.12 Thereby, juice from healthy Hamlin oranges was higher in aldehyde
53
volatiles such as acetaldehyde and octanal, higher in (Z)-3-hexenol and higher in
54
esters such as ethyl acetate and ethyl butanoate.4 Juice from healthy and infected
55
Valencia oranges showed no significant differences in volatiles.4 Similar results were
56
obtained by Dagulo et al. who used fruits from the same seasons and concluded that
57
HLB infection might be the major cause for the higher concentrations of terpenes and
58
alcohols and lower levels of esters such as ethyl acetate, ethyl butanoate, methyl
59
hexanoate, ethyl-3-hydroxyhexanoate and sesquiterpenes such as valencene in
60
symptomatic juices.13
61
Lower amounts of sugars and higher amounts of citric acid and limonin were
62
found to have a major effect on the flavor of juice from infected fruits.4, 13 Limonin is
63
known to cause delayed bitterness in sweet orange juices originating from the
64
conversion of the tasteless precursor limonate-A-ring lactone in the acidic juice
65
matrix in the presence of heat or enzyme limonin-D-ring-hydrolase (EC 3.1.1.36).14-16
66
During maturation limonin could be converted into its tasteless precursors down to a
67
residual level of limonin below 2.6 mg/L where bitter taste is not perceived.5
68
Therefore, limonin is an important quality parameter and is routinely analyzed in the
69
citrus industry. The role of nomilin as bitter taste enhancing component was recently
70
highlighted and nomilin (8.4 mg/L bitter threshold) as well as limonin (7.0 mg/L bitter
71
threshold) acted synergistically in a model solution by lowering the recognition
72
threshold to 3.9 mg/L (1:1 mixture). In juice, nomilin at 2 mg/L subthreshold
73
concentration reduced the limonin recognition threshold to 2.6 mg/L.5 However, in
74
juice from oranges of infected trees additional off-flavors not associated with limonin
75
or nomilin occur.12,
17
Little information is available on those additional off-flavor ACS Paragon Plus Environment
Page 4 of 34
Page 5 of 34
Journal of Agricultural and Food Chemistry
5 76
compounds because current studies are focusing more on known taste components
77
and sensory guided screening approaches like gas chromatography olfactometry or
78
LC Taste®18 are scarcely applied.
79
Commercial orange juice “not-from-concentrate” is usually made by blending
80
juice from different varieties such as the early season Hamlin with a more acidic and
81
green flavor profile and late season Valencia having a sweeter and overripe flavor
82
profile. Blending up to 25 % juice of HLB infected fruits with less than 2.5 mg/L
83
limonin with juice of healthy fruits from 2009 season did not perceivably change the
84
flavor.17 However, this way of blending is limited by the availability of juice from
85
healthy fruits which decreases year by year. Therefore, this approach is already self-
86
limiting because limonin levels in early Hamlin juice are continuously increasing19 and
87
at the same time the sugar level in late Valencia decreases which prompted us to
88
study the off-flavor in more detail during the season 2014 and 2015. It was the aim to
89
investigate the key flavor molecules of orange juice from healthy and infected fruits
90
by sensory guided analysis of the peel and juice volatiles and non-volatiles in order to
91
develop strategies for compensating juice derived off-flavor notes.
92
ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
Page 6 of 34
6 93
MATERIAL AND METHODS
94
Materials. The following reference compounds were obtained in the highest
95
available grade of purity: hesperidin, naringenin, liquiritin, and citric acid (Sigma-
96
Aldrich,
97
Vestenbergsgreuth, Germany) and sucrose (Fluka, Neu-Ulm, Germany). Nomilin,
98
tangeretin, nobiletin, eriocitrin and limonin glucoside were gifts from Dr. Baldwin and
99
Dr. Manthey (USDA-ARS, Florida, USA).
Steinheim,
Germany);
limonin,
poncirin
and
didymin
(Phytolab,
100
LC/MS grade acetonitrile and formic acid were from Fluka (Neu-Ulm,
101
Germany) and sulfuric acid was obtained from Sigma-Aldrich (Steinheim, Germany).
102
Deionized water was prepared by an arium mini UV lab water system (Sartorius,
103
Göttingen, Germany). Silica gel for column chromatography (silica gel 60, 63−
104
200 µm) was obtained from Merck KGaA (Darmstadt, Germany).
105
Orange juices were spiked with hesperidin (> 98 %) obtained from Blue
106
California Ingredients (Rancho Santa Margarita, USA), a polymethoxylated flavone
107
extract from orange containing 1.32 % tangeretin and 28.6 % nobiletin (Miritz,
108
Kirchgandern, Germany) and limonin and poncirin obtained from Phytolab (> 98 %
109
and > 95 % respectively, Vestenbergsgreuth, Germany) for sensory evaluation. Stock
110
solutions of these compounds of 1-10 % concentration in 1,2-propanediol and
111
ethanol mixtures were prepared at 50 °C and finally diluted in juice.
112
Hamlin and Valencia oranges (Citrus sinensis (L.) Osbeck) were collected on
113
a monthly basis from commercial groves in southern Florida during the season 2014
114
and 2015. The trees were selected based on visual symptoms: blotchy-mottle leaves,
115
twig die back and small fruits with diameter less than 6 cm were characteristic for
116
HLB infected symptomatic trees.3 Regular sized fruits from HLB infected symptomatic
117
trees were classified as asymptomatic and the other small, lopsided fruits as
118
symptomatic HLB fruits. Fruits were taken from the same number of healthy and ACS Paragon Plus Environment
Page 7 of 34
Journal of Agricultural and Food Chemistry
7 119
infected trees in a random manner. The fruits were washed, hand juiced and lightly
120
pasteurized using a dish washer with 71 °C for 15 s. Before squeezing fruits were
121
checked for aborted seeds which is another criteria for classification of symptomatic
122
fruits.3
123 124
Preparation and analysis of terpeneless peel oil. The contribution of
125
hydrocarbon terpenes such as limonene, myrcene and pinenes to odor perception in
126
juice is limited because of their high odor thresholds.20 For this reason, column
127
chromatography was used to reduce hydrocarbons from 2 g cold pressed peel oil.
128
Chromatography was performed on a 45 cm × 40 mm i.d. glas column filled with
129
silica gel. First, the column was loaded and then eluted with 1 L of 100 % pentane
130
and finally rinsed with 1 L of 100 % diethyl ether. The pentane fraction was discarded
131
and the diethyl ether fraction with a concentration of limonene below 400 ppm was
132
used for GC-MS analysis after concentration to 1 mL by Turbovap evaporation
133
(Turbovap II, Biotage, Sweden). One microliter of the sample solution was injected
134
with a Gerstel MPS XL autosampler (Gerstel, Muelheim, Germany) in a cooled
135
injection system containing an empty glass liner and then desorbed at 180 °C in
136
splitless injection mode. GC-MS analysis was performed as described in sensory
137
guided analysis of volatiles part.
138 139
Preparation of orange juice extract. Twenty liter juice from HLB
140
symptomatic fruits was centrifuged and processed by solid phase extraction. A
141
column with 250 mm × 25 mm size filled with polystyrene resin was rinsed with the
142
juice and medium polar components were trapped. The loaded column was washed
143
with water to remove residual polar components such as sugars and organic acids
144
and was then eluted with ethanol. The eluate was concentrated by evaporation to ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
8 145
dryness and finally freeze dried to obtain 19 g extract. The pulp obtained by
146
centrifugation and the polar and medium polar eluates were evaluated by five trained
147
panelists on sucrose and citric acid taste solution and the medium polar fraction
148
showed highest bitter taste intensity.
149 150
Juice basic chemical parameters. The basic chemical parameters such as
151
sucrose, total soluble solids (°Brix), Scott oil and citric acid of the juice were
152
determined as follows: sucrose was quantified via an external standard method using
153
an Agilent 1100 series HPLC system (Agilent, USA) comprising a binary pump,
154
autosampler and RI detector. Chromatographic separation was carried out at 80 °C
155
on a Grom Resin ZP column (250 mm x 8 mm, 8 µm, Dr. Maisch, Germany) in
156
isocratic mode with an aqueous mobile phase. Flow rate was 0.4 mL/min and
157
injection volume was 10 µL. The Brix value was determined with an Abbe
158
refractometer at 20 °C (Atago, Tokyo, Japan). Citric acid was quantified via an
159
external standard method using an Agilent 1290 Infinity HPLC system (Agilent, USA)
160
comprising a binary pump, autosampler and DAD detector. Chromatographic
161
separation was carried out at 40 °C on an Aminex HPX-87H column (300 mm x
162
7.8 mm, 9 µm, Bio Rad, USA) in isocratic mode with 0.004 M sulfuric acid water.
163
Flow rate was 0.8 mL/min and injection volume was 10 µL. The eluent was monitored
164
at 210 nm. Peel oil content was quantified by the Scott bromate titration method.21
165
Briefly, 25 g of the juice sample and 25 g of ethanol were distilled with a few boiling
166
stones until the solvent ceased to reflux. Then, 10 mL of 4 N HCl with a drop of 0.1 %
167
methyl red indicator were added and peel oil content was determined by titrating the
168
distilled fraction with 0.025 N bromide-bromate solution until the color disappeared.
169
ACS Paragon Plus Environment
Page 8 of 34
Page 9 of 34
Journal of Agricultural and Food Chemistry
9 170
Sensory guided analysis of volatiles. Extraction of juice components was
171
performed using Gerstel® PDMS Twister with 1 mm layer thickness and 10 mm
172
length (Gerstel, Muelheim, Germany). One gram of juice was weighed in a 10 mL
173
headspace vial and 5 mg/kg of 2-nonanol as internal standard was added and the
174
vial was finally filled with 10 grams of deionized water. The Twister was added to this
175
solution and stirred at room temperature maintained at 20 °C for one hour. The
176
Twister was then taken out of the solution and rinsed with few milliliters of deionized
177
water and carefully dried with clean paper towels.
178
Sample volatiles were transferred to the GC by desorbing the Twister at
179
180 °C for 8 min in the Gerstel® thermal desorption unit (TDU), then the analytes
180
were cryofocussed at -50 °C in the cooled injection system inlet containing a Tenax
181
filled glass liner and then desorbed at 200 °C for 10 min onto the GC column, where
182
the analytes were separated on a ZB-Wax plus column (60 m x 0.32 mm, 0.25 µm,
183
Phenomenex, Torrance, California, USA) placed in a Agilent 6890A gas
184
chromatograph (Waldbronn, Germany) using helium as the carrier gas.
185
The effluent was transferred by a Graphpack®-3D/2 sulfonated crosspiece
186
(Gerstel, Muelheim, Germany) into three deactivated fused silica capillaries directed
187
to a sniffing port set at 250 °C (ODP 3, Gerstel, Germany), FID set at 300 °C, and
188
MS (Agilent 5973 MSD) for simultaneous detection. A constant flow of 2 mL/min was
189
applied. The GC temperature was programmed starting at 40 °C, held for 2 min, then
190
raised at 4 °C/min to 240 °C, and held for 38 min. Retention indices were calculated
191
by cochromatography of a homologous series of n-alkanes from C6-C30. The effluent
192
was transferred to the MS via a 2.2 m deactivated capillary column held at 280 °C
193
and the MS was set in scan mode (EI ionization 70 eV, m/z 25 – 400, 3.4 scans/sec)
194
The volatile analytes were identified by matching the retention index (constraint of ±5
195
set for positive match) and mass spectrum (constraint of >800/1000 match factor set ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
10 196
for positive match) with an in-house database which was built by analysis of
197
authentic references on standard instruments with defined protocols. The peak area
198
of 2-nonanol obtained by the FID signal was used for peak response normalization.
199
Data were acquired by Agilent MSD ChemStation version E.02.02.
200
Aroma extraction dilution analysis22 of the cold pressed and terpeneless peel
201
oils was performed as follows: the oil was stepwise diluted 1:1 with diethyl ether, and
202
aroma-active compounds were located by sniffing each dilution (HRGC-O) and by
203
calculating retention indices on a ZB-Wax plus column (60 m x 0.32 mm, 0.25 µm)
204
(Phenomenex, Torrance, California, USA). Evaluation of the dilutions was performed
205
until no odorant could be sniffed in the diluted extract and the flavor dilution (FD)
206
factor was, thus, denoted for each compound. The extract with FD 1 was sniffed by
207
three trained assessors independently to define a list of odor active components.
208 209
LC Taste® fractionation18 and taste dilution analysis. Orange juice bitter
210
components were studied by chromatographic fractionation of an extract of HLB
211
infected fruits followed by taste analysis of three sequential dilutions. LC Taste®
212
fractionation was performed using high temperature liquid chromatography (HTLC)
213
on a polymer-based PRP-1 column in a semipreparative scale (250 mm × 10 mm;
214
10 µm particle size; Hamilton, Bonaduz, Switzerland) at elevated temperature (80 °C
215
isotherm) and detection on a diode array detector (SunChrom SpectraFlow;
216
wavelength range 200-550 nm, SunChrom, Friedrichsdorf, Germany). The mobile
217
phase consisted of water (A) and ethanol (B) and a gradient elution of 100 % A at
218
0 min to 25 % B in 25 min to 100 % B at 25-40 min was used. Flow rate was
219
10 mL/min, injection volume 1 mL and the eluent was monitored at 230 nm.
220
The fractions were dissolved in 1,2-propanediol and diluted in sucrose 3 %
221
and citric acid 0.1 % taste solution and sequentially diluted with taste solution 1:10, ACS Paragon Plus Environment
Page 10 of 34
Page 11 of 34
Journal of Agricultural and Food Chemistry
11 222
1:50 and 1:500. The genuine concentration of the bitter components in orange juice
223
was mimicked by the dilution of 1:500. The serial dilutions of each of these fractions
224
were then presented to the sensory panel in order of ascending concentration. The
225
intensity of the perceived bitterness of each dilution and fraction was evaluated on a
226
scale of 0-10.
227 228
Identification of taste active components. The composition of the fractions
229
obtained by LC Taste® analysis was analyzed first by liquid chromatography high
230
resolution mass spectrometry (LC-HRMS). Mass spectra were recorded using a
231
mass spectrometer Bruker microTOFQII (Bruker, Bremen, Germany) with ESI+ and
232
ESI- ionization and scan range of 50-1600 Da in combination with a Waters Acquity
233
UPLC system (Waters, Eschborn, Germany). Chromatographic separation was
234
carried out on a C-18 column (Kinetex, 1.7 µm, 100 mm × 2.1 mm; Phenomenex,
235
Torrance, California, USA) at a temperature of 50 °C and a flow rate of 0.55 mL/min
236
using an acetonitrile/water gradient. Two microliter was injected and the gradient
237
started with 100 % water containing 0.1 % formic acid, increasing to 95 % acetonitrile
238
within 22 min. This concentration was kept for 5 min. Peak identity was confirmed
239
either by cochromatography of the reference compounds or by NMR analysis. 1H,
240
13C, COSY, HMQC, and HMBC measurements were performed on a Bruker
241
Avance-III 600 spectrometer (Bruker, Rheinstetten, Germany). Data processing was
242
performed with MNova (version 9, Mestrelab Research, Santiago de Compostela) or
243
Topspin (version 3.2; Bruker, Rheinstetten). Spectra were recorded in DMSO-d6 and
244
referenced towards tetramethylsilane.
245 246
Quantitative HPLC analysis of hesperidin. Hesperidin was quantified via an
247
external standard method using an Agilent 1290 Infinity HPLC system (Agilent, USA), ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
12 248
consisting of binary pump, autosampler and DAD detector. Chromatographic
249
separation was carried out at 40 °C on a poroshell 120 SB-C18 column (100 mm x
250
2.1 mm, 2.7 µm, Agilent, USA). The mobile phase consisted of 0.1 % formic acid
251
water (A) and acetonitrile (B) and a gradient elution of 5-5 % B at 0-0.1 min; 5-50 %
252
B at 0.1-10 min; 50-100 % B at 10-12 min and 100-100 % at 12-15 min was used.
253
Flow rate was 0.4 mL/min and injection volume was 2 µL. The eluent was monitored
254
at 280 nm.
255 256
Quantitative LC-MSMS analysis of bitter components. The LC-MSMS
257
system consisted of an Agilent HP 1200 system and a 3200 QTRAP MSMS (Applied
258
Biosystems, Darmstadt, Germany) with an electrospray ionization (ESI) source. The
259
data were processed using the Analyst 1.5 software. Chromatographic separation
260
was carried out at 50 °C on a Luna C18(2) column (150 mm x 2.0 mm, 3 µm,
261
Phenomenex, USA). The mobile phase consisted of 0.1 % formic acid water (A) and
262
acetonitrile (B) using a gradient elution for method 1 (M1) of 5-30 % B at 0-5 min; 30-
263
95 % B at 5-30 min; 95-95 % B at 30-35 min and for method 2 (M2) of 5-70 % B at
264
0-15 min; 70-95 % at 15-19 min; 95-95 % at 19-20 min. Flow rate for both methods
265
was 0.2 mL/min and injection volume was 5 µL. Mass spectrometer was operated in
266
negative mode for M1 including liquiritin (internal standard, IS), eriocitrin, limonin
267
glucoside, didymin, poncirin, naringenin, limonin, nomilin and for M2 in two MSMS
268
scan segments: first segment 0-13 min in negative mode for liquiritin (IS) and second
269
segment 13-20 min in positive mode for nobiletin and tangeritin. In the multiple
270
reaction monitoring (MRM) mode, the compounds and IS were analyzed using the
271
mass transition and MSMS operating conditions described in Table 1. The optimum
272
instrumental parameters were determined by analysis of analytes and IS (10 µg/mL)
273
solutions into ESI source via a syringe pump. The following MS parameters were ACS Paragon Plus Environment
Page 12 of 34
Page 13 of 34
Journal of Agricultural and Food Chemistry
13 274
used for the measurements: source temperature: 500 °C, ion spray voltage: -4.5 kV
275
in negative mode and 5.5 kV in positive mode, nitrogen served as curtain gas
276
(20 psi).
277
For quantitative analysis, aliquots (0.5 g/2 g, depending on analyte concentration) of
278
the OJ spiked with 150 µL of IS solution (60 µg/mL), were extracted with methanol
279
(up to 4 g) in an ultrasonic bath for 15 min. The extract was separated from solids by
280
centrifugation; the supernatant was passed through a 0.2 µm membrane filter and
281
analyzed by means of LC-MSMS. For calibration purpose, reference standards
282
(100 µg/mL for eriocitrin, limonin glucoside, didymin, poncirin, naringenin, limonin,
283
nomilin, nobiletin and tangeretin) and IS (60 µg/mL for liquiritin) were dissolved in
284
methanol. A series of working calibration solutions spiked with IS solution (60 µg/mL)
285
were prepared by appropriate dilution of standard stock solution with methanol to
286
yield final concentrations of 0.2, 0.5, 1, 3, 7 µg/mL for eriocitrin; 0.8, 2, 5, 12,
287
38 µg/mL for limonin glucoside; 0.4, 1, 3, 6, 15 µg/mL for didymin; 0.05, 0.2, 0.5, 1,
288
3 µg/mL for poncirin and tangeretin; 0.06, 0.1, 0.2, 0.5, 1, 3 µg/mL for naringenin and
289
nobiletin; 3, 7, 17, 50 µg/mL for nomilin and 1, 3, 6, 16 µg/mL for limonin. The limits
290
of quantitation (LOQ) ranged from 0.01 to 5 µg/mL (Table 1). Recovery experiments
291
were performed in a commercial OJ by spiking the OJ with standard stock solution
292
and IS solution (60 µg/mL). Mean recovery rates ranging from 90 % (nobiletin) to
293
115 % (eriocitrin) were found.
294 295
Comparative sensory analysis. Sensory tests were carried out with a panel of 10
296
trained panelists without known taste disorders. Panelists were informed about the
297
procedure and intention of the project, and written consents were obtained. Four
298
samples including a reference juice were presented and the panelists were asked to
299
rank and score the attributes sweet, acidic and HLB bitter on a scale of 0-10. The ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
14 300
samples were tested in sensory panel rooms under single-blinded and standardized
301
conditions and a fully randomized presentation design among product, panelist and
302
replications was chosen. Only four samples including the reference juice were tested
303
in one session to allow proper rinsing of the mouth and to avoid carry-over of taste
304
impressions.
305
The comparative sensory test was independently run three times with two
306
different juice bases spiked with hesperidin, limonin, poncirin or polymethoxylated
307
flavons. Significance levels were not calculated because less than 10 panelists
308
attended. First, panelist were trained by descriptive profiling of orange juices with low
309
250
54
1,6
255
140
22
16 250
49
13
0,3
2,3
3,5
230
240
15
10
0,1
0,5
3,3
270
>250
16
11
