Subscriber access provided by Binghamton University | Libraries
Article
Matrix effects and interferences of different citrus fruits coextractives in pesticide residue analysis using ultra high-performance liquid chromatography–high resolution mass spectrometry Natalia Besil, Veronica Cesio, Horacio Heinzen, and Amadeo R. Fernández-Alba J. Agric. Food Chem., Just Accepted Manuscript • Publication Date (Web): 25 May 2017 Downloaded from http://pubs.acs.org on May 25, 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 40
Journal of Agricultural and Food Chemistry
1
Matrix effects and interferences of different citrus fruits co-extractives in
2
pesticide residue analysis using ultra high-performance liquid
3
chromatography–high resolution mass spectrometry.
4
5
Natalia Besil1,2,3, Verónica Cesio2,3, Horacio Heinzen2,3 and Amadeo R. Fernandez-
6
Alba1
7 8
1
9
Pesticide Residues in Fruit and Vegetables. Pesticide Residue Research Group, Department of
Agrifood Campus of International Excellence (ceiA3). European Union Reference Laboratory for
10
Chemistry and Physic, University of Almeria, La Cañada de San Urbano, 04120 Almeria, Spain
11
2
12
Química. CENUR Litoral Norte. Universidad de la República (UdelaR), Uruguay
13
3
14
Facultad de Química, Universidad de la República, General Flores 2124, 11800 Montevideo,
15
Uruguay
Grupo de Análisis de Compuestos Traza. Departamento de Química del Litoral, Facultad de
Grupo de Análisis de Compuestos Traza. Cátedra de Farmacognosia y Productos Naturales,
16 17 18 19
Corresponding author: Agrifood Campus of International Excellence (ceiA3).
20
European Union Reference Laboratory for Pesticide Residues in Fruit and
21
Vegetables. Pesticide Residue Research Group, Department of Chemistry and
22
Physic, University of Almeria, La Cañada de San Urbano, 04120 Almeria, Spain.
23
E-mail address:
[email protected] , Tel: +34 950 015 034 Fax: +34 950 015 483
1 ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
Page 2 of 40
24
25
ABSTRACT
26
The matrix effects of ethyl acetate extracts from seven different citrus fruits on the
27
determination of 80 pesticide residues using liquid chromatography coupled to high
28
resolution time of flight mass spectrometry (UHPLC-(ESI)-HR-TOF) at 4 GHz
29
resolution mode were studied. Only 20% of the evaluated pesticides showed
30
noticeable matrix effects (ME) due to co-elution with natural products between tR=3-
31
11 minutes. Principal components analysis (PCA) of the co-extractives detected
32
grouped the mandarins and the orange varieties, but separated lemon, oranges, and
33
mandarins from each other. Matrix effects were different among species but similar
34
between varieties, forcing to determine pesticide residues through matrix matched
35
calibration curves with the same fruit. Twenty-three natural products (synephrine,
36
naringin, poncirin, glycosides of hesperitin, limonin, nomilin and a few fatty acids,
37
among others) were identified in the analysed extracts. Twelve of the identified
38
compounds co-eluted with 28 of the pesticides under study, causing different matrix
39
effects.
40 41
Keywords: Citrus sp; Pesticide residues; Matrix effects; Natural products, High
42
resolution time of flight mass spectrometry
43
2 ACS Paragon Plus Environment
Page 3 of 40
Journal of Agricultural and Food Chemistry
44
Introduction
45
Several approaches to determine pesticide residues in citrus fruit (1-5) and
46
processed citrus products, such as fruit soft drinks (6, 7), juices (8, 9) and essential
47
oils (10, 11), have been developed. The most notable, is the QuEChERS method
48
(12) and its different versions (13, 14), as well as the ethyl acetate method (5). After
49
chromatographic separation, the extraction methods are usually coupled to
50
sophisticated and powerful mass spectrometry platforms. Despite the available
51
analytical tools, the correct quantification of analytes during monitoring programs and
52
routine analysis is affected by matrix effects (ME). The ME is the resultant of the
53
presence of inferring matrix components at the time of measurement. The
54
International Union of Pure and Applied Chemistry (IUPAC) (15) defined it precisely,
55
as well as the SANTE guidelines, which also explicitly added that “ IThe response
56
of some determination systems (e.g. GC-MS, LC-MS/MS) to certain analytes may be
57
affected by the presence of co-extractives from the sample (matrix)0” (16). It is
58
therefore clear that the commodity, the sample treatment procedure and others
59
aspects such as the type of mass spectrometer or the acquisition mode influence
60
directly on matrix effects (17). During the quantification step the ME can play an
61
important role, such as diminishing or enhancing the signal for the analyte (18). To
62
compensate it, matrix-matched calibration curves are commonly used and
63
recommended in the DG-SANTE guidelines (16).
64
Since 1992, the occurrence of ME in pesticide residue analysis using either
65
conventional and/or mass spectrometer detectors was investigated by many
66
researchers (19-26), as well as the influence of the clean-up step. (27-29).
67
Actually, accurate high resolution mass spectrometry (HR-MS) coupled to
68
chromatographic separation systems are being increasingly incorporated in the 3 ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
Page 4 of 40
69
routine work of pesticide residues monitoring laboratories. Nevertheless, few studies
70
have shown the performance of the different analytical methods using HR-MS for the
71
determination of pesticides in citric matrices where the ME was briefly discussed (25,
72
30). Recently, Gomez-Ramos et al (31) studied the ME caused by three different
73
dispersive sample preparation methods in one orange variety among a wide variety
74
of other fruits and vegetables. Although the phytochemical composition of citrus fruits
75
has been largely described, the role of specific co-extractives on the ME in trace
76
contaminant analysis has been only scarcely studied. Previous studies of ME in
77
orange extracts using GC-MS reported by Sugitate et al (34) allowed the
78
identification of 1-monopalmitin, 1-monolinolein, squalene, α-tocopherol and β-
79
sitosterol.
80
polyphenols as well as terpenoids were determined using liquid chromatography
81
(HPLC) coupled to different detectors (32, 33), but the simultaneous identification
82
and influence of specific co-extractives on MEs during citrus pesticide residues
83
analysis using UHPLC-HR-MS-TOF has not been judiciously studied yet.
84
Aiming to contribute to the understanding of the overall ME in the pesticide residues
85
analysis in citrus fruits, four oranges, two mandarins and one lemon varieties were
86
studied using an ethyl acetate extraction method (35) followed by UHPLC-HR/MS.
87
The behaviour of different citrus species and varieties were compared. Furthermore,
88
the influence of specific co-extracted matrix components was evaluated, taking into
89
account the chemical diversity of the material under study.
Flavonoids,
coumarins,
permethoxylated
flavones
among
other
90
91
Material and methods
92
Standards and reagents
4 ACS Paragon Plus Environment
Page 5 of 40
Journal of Agricultural and Food Chemistry
93
High purity pesticides standards (Table 1) purchased from Dr. Ehrenstorfer GmbH
94
(Augsburg, Germany), Sigma-Aldrich (Steinheim, Germany) and Riedel-de Häen
95
(Selze, Germany) were stored at -30 °C. Individual pesticide stock solutions (1000-
96
2000 µg/mL) were prepared by dissolving reference standards in the appropriate
97
solvent. Each solution was stored in amber screw-capped glass vials in the dark at -
98
20 °C. The working standard mix solution for spiking purposes was prepared at 10
99
µg/mL concentration in acetonitrile (MeCN).
100
Flavonoids standards were from Fluka AG and the other natural products reference
101
standards were isolated in Pharmacognosy and Natural Product at the Faculty of
102
Chemistry, UdelaR, and characterized using nuclear magnetic resonance (NMR) and
103
electronic ionization mass spectrometry (EI-MS).
104
HPLC-grade acetonitrile (MeCN) was obtained from Sigma-Aldrich (Steinheim,
105
Germany) and analytical grade ethyl acetate (EtOAc) from Pharmco Products Inc.
106
(Brookfield, CT. USA). Formic acid (98% purity for mass spectrometry) was from
107
MACRON Chemicals (Netherlands). Sodium chloride from J.T. Mallinckrodt Baker
108
Inc. (Phillipsburg, NJ, USA) and magnesium sulphate anhydrous from Scharlau
109
(Spain) were employed throughout the whole study.
110
Instrumentation
111
Liquid chromatography high resolution mass spectrometry was used to separate and
112
identify pesticides and natural products compounds. The UHPLC analysis was
113
performed with an Agilent 1290 Infinity, equipped with a C8 Zorbax Eclipse Plus
114
column (100×2.1mm, 1.8µm) from Agilent Technologies. A gradient program at 30
115
°C, using of two solvents: A (0.1 % of HCOOH in ultrapure water) and B (0.1 % of
116
HCOOH and 5 % of water in acetonitrile) was developed for the chromatographic 5 ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
Page 6 of 40
117
separations. It started at 80: 20 (A: B) relationship for the first 2 min. Then, B was
118
lineally increased to 100% from 2 to 15 min. Solvent B was maintained at 100% for 2
119
min, decreased to 20% and kept from 17 to 19 min to re-equilibrate the column. The
120
total run time was 19 min per sample. The flow rate was 0.3 mL/min with an injection
121
volume of 4 µL.
122
The liquid chromatograph was connected to an Agilent 6550 iFunnel QTOF-MS
123
(Palo Alto, Agilent Technologies). The acceptable mass accuracy (within ± 2 ppm) of
124
the QTOF-MS (used as TOF only) was calibrated before each analysis with a
125
reference solution of HP-0921 (commercially available from Agilent) for scanning up
126
to 1700 mass-to-charge ratio (m/z) and operated at 4 GHz(12000-16000 FWHM).
127
The ionization mode was positive electrospray (+ESI), with the fragmentation voltage
128
at 360 V. The nebulizer and drying gas was nitrogen at 350 °C.. In each analysis, the
129
reference mass standard solution was constantly infused into the QTOF-MS for the
130
monitoring and measurement of its mass accuracy with the reference masses of
131
121.0509 and 922.0098 m/z.
132
Sample preparation
133
The oranges, mandarins and lemons samples used in the study, were obtained from
134
several local growers in Uruguay. One of them came from a non-commercial,
135
pesticide free orange tree grown in a familiar garden. Each sample employed
136
weighted 2 kg of the different citrus fruit variety. The fruits were cut in four pieces,
137
the two opposite ones were separated for analysis and the other two were discarded.
138
The selected fruit pieces were chopped and homogenized using a kitchen blender.
139
After composition and homogenization the final laboratory sample was stored at -20
140
°C until analysis. The extraction method procedure chosen was a previously
141
validated one by our group (35). Ten grams of the homogenized sample were 6 ACS Paragon Plus Environment
Page 7 of 40
Journal of Agricultural and Food Chemistry
142
weighed and placed into a 50 mL teflon centrifuge tube. Then, 10 mL of ethyl acetate
143
were added and the tube was hand shaked vigorously for 1 min. Afterwards, 8 g of
144
anhydrous magnesium sulphate and 1.5 g of sodium chloride were added. The
145
mixture was shaked vigorously by hand for 1 min. The centrifuge tubes were placed
146
in an ultrasonic bath for 15 min and centrifuged during 10 min at 5000 rpm. The ethyl
147
acetate supernatant solution of 1 g/mL of sample was separated and concentrated
148
under a N2 stream to dryness. The residue was re-dissolved in MeCN:H2O (1:9). The
149
final extract was injected in the UHPLC system using the conditions described
150
above.
151 152
Matrix matched standards preparation
153
To each citrus species and variety extract, the exact volume of pesticides working
154
standard solution and MeCN:H2O (1:9) was added to re-dissolved it (final volume of
155
100 µL). The fortified extracts were diluted 5 times with MeCN:H2O (1:9), yielding a
156
final concentration in the fortified extract from 10 to 150 µg/L. The calibration curve
157
used as reference was prepared in solvent at the same studied levels. The resulting
158
concentrations in vial were from 2 to 30 µg/L respectively, ending up with the
159
equivalent of 0.2 g of sample per mL of MeCN:H2O (1:9).
160
Determination of matrix effect
161
Matrix effects were evaluated injecting matrix-matched calibration curves and
162
reference calibration curves in solvent. Matrix effects were quantified as the
163
percentage of deviation of the response of the analyte in matrix vs. its signal intensity
164
in the reference curve (IUPAC defines sensitivity as the slope of the calibration
165
curve) calculated as:
7 ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
Page 8 of 40
166
=
( ℎ ) − 1 × 100 ( )
167
Positive values represent enhancements and negative values suppressions of the
168
analyte signal induced by the matrix. In this work ME was classified, as currently, in
169
3 levels; a) weak ME is considered when ME values is less or equal to 20 %. It can
170
also be assumed as no matrix effects because the shift is between the repeatability
171
values (36); b) medium ME is considered between 20 to 50%, and c) strong ME
172
occurred when the values were above 50%.
173
Results and discussion
174
The ME caused by ethyl acetate extracts of three different citrus species, lemon,
175
oranges, mandarins and varieties from the latter two, during pesticide residues
176
analysis were characterized. ME calculation is usually performed using, at least, two
177
different procedures. The first one compares the signal response between an analyte
178
in a determined matrix and in solvent. This procedure is often done at different
179
concentration levels as reported by Cervera (37).
180
Other authors (5, 22) reported the determination of ME as the comparison between
181
the slopes of matrix-matched standard curves against calibration standards curves in
182
solvent. The requirements to accomplish this criteria are that: either the sum of least
183
squares at each point must not exceed 20%, or the regression coefficient (r2) of both
184
curves is higher than 0.99 The main drawback of this approach is that it can be
185
inaccurate for pesticides with an r2 < 0.99, or whose sum of least squares at each
186
point exceeds 20% (38). This was the criteria followed in the present study, and thus
187
the slope adjustment to r2 ≥ 0.999 was verified throughout the experiment.
8 ACS Paragon Plus Environment
Page 9 of 40
Journal of Agricultural and Food Chemistry
188
Two types of matrix interferences are observed using ESI ionization followed by
189
mass spectrometry detection. The most commonly effect of matrix co-extractives on
190
residue determination occurs when they co-elute with the analytes from the
191
chromatographic column in the ionization step. The pesticides of interest and the co-
192
extractives placed in the surface of the solvent droplet compete in ion generation
193
during the Coulomb explosion, causing in most cases a drop in the signal intensity
194
(i.e., “suppression”). In addition to this effect, if the co-eluting compound generates
195
ions with similar mass to those of the analyte (interferences), the ion ratio could
196
change, leading to miss-interpretation of the results. The later is seen when
197
determining pesticide residues through single quadrupole analyzers due to their poor
198
mass resolution, and even in QqQ ones despite the selectivity improvement they
199
have in these instruments. The problem could be partially avoided when high
200
resolution spectrometers are employed, such as TOF or Orbitrap™.
201
Linearity
202
An important drawback of TOF systems is their short dynamic range due to rapid
203
saturation of the detector. The results obtained in this work confirmed previous
204
reports (39, 40). In some cases pesticide levels as little as 0.01 mg/kg were enough
205
to saturate the detector. The linear ranges are variable for each pesticide due to the
206
combination of specific analytes and matrix components. This is the case of
207
cyprodinil, which has a linear range up to 0.05 mg/kg in lemons and oranges, but
208
reached 0.10 mg/kg in mandarins.
209
Matrix effects
210
All citrus varieties caused a differential response for a number of analytes when
211
compared with the corresponding signal in solvent. ME values for each combination 9 ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
Page 10 of 40
212
of pesticide and citrus variety are shown in Table 1. The three citrus species showed
213
differential matrix effects for some pesticides under study. This fact stresses the
214
need for protocol validation of each citrus species matrix to perform accurate
215
residues determinations, despite the generally accepted statement that it is sufficient
216
to validate a single matrix as representative for all the components in a matrix
217
category.
218 219 220
