Subscriber access provided by Kaohsiung Medical University
Food and Beverage Chemistry/Biochemistry
Towards the authentication of European sea bass origin through a combination of biometric measurements and multiple analytical techniques Federica Farabegoli, Maurizio Pirini, Magda Rotolo, Marina Silvi, Silvia Testi, Sergio Ghidini, Emanuela Zanardi, Daniel Remondini, Alessio Bonaldo, Luca Parma, and Anna Badiani J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.8b00505 • Publication Date (Web): 08 Jun 2018 Downloaded from http://pubs.acs.org on June 8, 2018
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 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 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.
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 38
Journal of Agricultural and Food Chemistry
1
Towards the authentication of European sea bass origin through a combination of biometric
2
measurements and multiple analytical techniques
3 4
Federica FARABEGOLIa*, Maurizio PIRINIa, Magda ROTOLOa, Marina SILVIa, Silvia TESTIa,
5
Sergio GHIDINIb, Emanuela ZANARDIb, Daniel REMONDINIc#, Alessio BONALDOa, Luca
6
PARMAa, Anna BADIANIa.
7 8
a
9
Sopra 50, 40064 Ozzano dell’Emilia (BO), Italy.
Department of Veterinary Medical Science (DIMEVET), University of Bologna, Via Tolara di
10
b
11
c
12
40127 Bologna (BO), Italy.
Department of Food Science, University of Parma, Via del Taglio 10, 43126 Parma (PR), Italy.
Department of Physics and Astronomy (DIFA), University of Bologna, Viale Berti Pichat 6/2,
13 14
*
Corresponding author.
15
E-mail address:
[email protected] 16
Postal address: Department of Veterinary Medical Science, Via Tolara di Sopra 50, 40064
17
Ozzano dell’Emilia (BO), Italy Tel: (+34) 986 469 373
18 19
#
Co-corresponding author.
20
Email address:
[email protected] 21
Tel: (+39) 051 20 9 5127
22
1 ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
23
Abstract:
24
The authenticity of fish products has become an imperative issue for authorities involved in the
25
protection of consumers against fraudulent practices and in the market stabilization. The present
26
study aimed to provide a method for authentication of European sea bass (Dicentrarchus labrax)
27
according to the requirements for seafood labels (Regulation 1379/2013/EU).
28
Data on biometric traits, fatty acid profile, elemental composition, and isotopic abundance of
29
wild and reared (intensively, semi-intensively and extensively) specimens from 18 Southern
30
European sources (n = 160) were collected and clustered in 6 sets of parameters, then subjected to
31
multivariate analysis.
32
Correct allocations of subjects according to their production method, origin and stocking density
33
were demonstrated with good approximation rates (94%, 92% and 92%, respectively) using fatty
34
acid profiles. Less satisfying results were obtained using isotopic abundance, biometric traits, and
35
elemental composition. The multivariate analysis also revealed that extensively reared subjects
36
cannot be analytically discriminated from wild ones.
37 38 39 40 41
Keywords: Seafood labeling, Dicentrarchus labrax, sea bass authentication, analytical fingerprinting, stocking density, fatty acid profile, isotope analysis, elemental composition.
42
2 ACS Paragon Plus Environment
Page 2 of 38
Page 3 of 38
43
Journal of Agricultural and Food Chemistry
1 Introduction
44
The demand for sea products for human consumption is in continuous expansion, with
45
aquaculture gaining importance in compensating for the deficit of global capture fisheries 1; as a
46
result, an increasing number of new farmed species have entered the marketplace over the past
47
decade. Public interest in food quality and origin, in particular the awareness of health benefits
48
derived from fish consumption, has strongly increased in recent years. Differentiation of seafood
49
products may influence consumer preferences, especially concerning discrimination between wild
50
and farmed fish. Actually, consumer choice between wild and reared products seems to be strongly
51
affected by beliefs resulting from stereotypes; for instance, cultured fish are usually associated with
52
the use of antibiotics and growth promoters, while wild products are considered tastier and healthier
53
than reared fish 2. Thus, regulatory interventions aim to avoid mislabeling, or substituting wild fish
54
with farmed specimens, and to mitigate risks to consumer confidence and health.
55
In the European context, due to the variety of sources of fish products (from seawater or
56
freshwater; from intensive, semi-intensive, or extensive rearing), Regulation (EU) No 1379/2013
57
3
58
traceability of fish products. Since wild fish are generally put on the market at a higher price,
59
there may be a temptation to falsely label farmed products to sell them at a better price. The
60
removal of specific external traits during product processing makes the identification of fish
61
species difficult, and the risk of fraudulent substitution of sought-after species with less valuable
62
ones grows. Accidental or intentional mislabeling may also occur due to the global production of
63
seafood, which allows the supply of the same products from different areas 4. Moreover, strong
64
competition among the main producing countries in the Mediterranean region (Italy, Greece,
65
Spain, and Turkey, above all) demands detailed characterization and differentiation of sea bass
66
quality5. Verifying the authenticity of fish products is therefore imperative, in order to check
67
correspondence with the label, establish the real commercial value, avoid unfair competition, and
68
ensure consumer protection against fraudulent practices.
was issued to give consumers clear and accurate information on the main features and
3 ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
Page 4 of 38
69
European sea bass (Dicentrarchus labrax) represents one of the main aquaculture fish products
70
in the European Union 6. European sea bass is a demersal opportunistic species, inhabiting coastal
71
waters (down to about 100 m depth), and shallow waters like estuaries, lagoons, and tidal flats.
72
European sea bass is intensively farmed in floating cages or in inshore ponds, and reared semi-
73
intensively or extensively in brackish lagoons 7. The different rearing systems and feeding regimes
74
in sea bass farming may affect flesh quality, especially in terms of fat level and composition 8,
75
which necessarily influences the quality of the product
76
estuarine waters, which offer dietary fatty acids (FAs) from the fauna of those areas, whereas
77
commercial feeds contain quite different FAs, typically fish products obtained from the open oceans
78
4
79
nitrogen (N); the former varies according to the nature of the dietary C sources as well, while the
80
latter depends on the trophic position of the subjects’ diet 11.
9,10
. Wild sea bass feed in inshore and
. The lipid content of the flesh also influences the ratios of the stable isotopes of carbon (C) and
81
According to some authors, the employment of chemometric techniques, in support to
82
documental routine, would be a useful tool to improve transparency and consumers’ trust in the
83
food trade. To date, the most promising approach to characterize a fish product seems to be the
84
application of different analytical techniques on the same matrix, followed by multivariate analysis
85
of the data obtained 12. The multivariate analysis applied to the FA profiles, obtained using gas
86
chromatography
87
(NIRS)
88
geographical origin. Other marine species have also been authenticated by employing stable
89
isotope analysis
90
assess origin, production method, and even characterize the species. Biometric parameters are
91
often used to authenticate sea bass type of production, since they are the easiest, quickest and
92
cheapest techniques and do not require laborious or expensive equipment, or expert knowledge
93
or specialization 12. Elemental composition of fish flesh is generally used as a chemical signature
94
of the particular water body where the subject has grown, since marine fish incorporate different
15
4,10,12,13
, nuclear magnetic resonance (NMR)
14
or near-infrared spectroscopy
, has been previously used to authenticate wild and farmed sea bass, as well as the
16
or by coupling the stable isotope ratio with multi-element analyses
4 ACS Paragon Plus Environment
17
, to
Page 5 of 38
Journal of Agricultural and Food Chemistry
95
trace elements, from the environment and the diet, into their skeletal tissues and organs 18. The
96
variation on elemental compositions has been also investigated in otoliths of different fish
97
species, for distinguishing wild from farmed environments; however, divergent results were
98
obtained due to an inter-annual variability in elemental composition of otoliths which
99
confounded the origin determinations of subjects
19
. The use of molecular genetic markers is
100
considered by Brown et al. (2015) the most suitable and informative tool for discrimination of
101
wild and farmed fish, since they are not influenced by the age of the subject 20. Unfortunately,
102
these techniques are among the most expensive and time-consuming currently available, making
103
them unavailable to many sectors 19. Previous studies have assessed the capacity of different analytical techniques to distinguish wild from
104
10
, and to determine their geographical origin 4, although these preliminary
105
farmed sea bass
106
investigations examined a limited number of sources and/or subjects, sampled in a limited time
107
span. The findings of these studies highlighted the difficulty in characterizing the farmed subjects
108
10
109
samples from various geographical locations 4, considering the inter-seasonal variability for
110
improving the robustness of the models 10.
, suggesting the need to introduce new variables and to apply this method on a wide range of fish
111
Prompted by all these considerations, the present study aimed to provide a method for
112
authentication of the origin of European sea bass specimens by combining data from a number of
113
chemical analyses. Being one of the main European aquaculture products, European sea bass
114
(Dicentrarchus labrax) was chosen due to the availability of multiple product typologies in the
115
Mediterranean market. For the first time, a vast number of samples have been analyzed, from
116
sources properly scattered during a one-year sampling among the southern European area, from
117
both aquaculture and fisheries sectors; moreover, a “stocking density factor” has been introduced
118
as a variable in the multivariate statistical analysis.
119 120
2 Materials and Methods 5 ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
121
2.1 Sampling design
122
Wild and farmed European sea bass subjects were sampled from 18 different Italian and foreign
123
sources. Batches of 10 specimens were ordered from every source, equally split into two
124
“macroseasons” (autumn-winter, AW; spring-summer, SS) to guarantee a perfect balance (equal
125
number of specimens from each macroseason for each source). The geographic distribution for each
126
sampling site is reported in Fig. 1: wild subjects (n = 45) came from four main areas, three in the
127
Mediterranean Sea and one close to the French Atlantic coast. Intensively reared subjects (IR, n =
128
85) were collected from fish farms equipped with either floating or submersible cages; in two cases,
129
sea bass belonged to semi-intensive farms (SIR, n = 20), and in a single case they came from an
130
extensive farm (ER, n = 10). Supplementary material Table S1 reports the respective FAO fishing
131
area, subarea and division for each fishing site, the country of origin, and the locality of sea bass
132
farms, as required by Regulation (EU) No 1379/2013, equipment features, and stocking density for
133
breeding. Since the studies of Xiccato et al. (2004)5 and Carbonara et al. (2015)21 suggested that
134
stocking density in production influences fish composition, an additional categorical variable was
135
used, taking one of the following labels: “0” for wild specimens, “1” for extensively farmed
136
specimens (up to 0.0025 kg m−3), “2” for semi-intensively farmed specimens (up to 1 kg m−3), and
137
“3” for intensively farmed specimens (up to 30 kg m−3). Stocking density of production was
138
recently recognized as a potential chronic stress factor in several species of fish; high density of
139
rearing can adversely affect the quality of the farmed product, due to a negative effect on fish
140
growth rate and on survival and feeding rates
141
was based on the statements of individual farmers, through the compilation of a special form
142
attached to every consignment.
21,22
. Source allocation according to stocking density
143 144 145 146
2.2 Biometric traits and sample preparation Each of 10 subjects from every batch was analyzed following the US EPA procedure for fish sampling
23
. Every subject, after measurement of weight and length, was eviscerated, and data on 6 ACS Paragon Plus Environment
Page 6 of 38
Page 7 of 38
Journal of Agricultural and Food Chemistry
147
the weight of eviscerated fish, viscera, liver, gonads (if any), and perivisceral fat were recorded.
148
Wild and reared subjects, both national and foreign, were purchased from large-scale retail trade;
149
and they were of commercial size. Fillets from each subject were homogenized by means of a food
150
processor (Multiquick System ZK 100, Braun, Kronbergim Taunus, Germany) and quartered by
151
hand twice; thereafter, samples were put into plastic bags and compacted, then frozen and stored at
152
−20 °C. The complete description of the procedure is reported in section 2.2 of supporting
153
information on Materials and Methods.
154 155
2.3 Moisture determination, lipid analysis and fatty acid profile
156
The moisture content of flesh was measured following AOAC procedure 950.46B 25, according
157
to which 10 g of homogenized flesh were put into a porcelain crucible and heated at 100–102 °C
158
during 18 h to determine dry weight before calculating the moisture.
159
The extraction of total lipids (TL) was done in duplicate for each sample, following a procedure
160
proposed by Bligh and Dyer (1959) 26, and adopting some adjustments, briefly described as follows:
161
around 4 g of homogenized fillets were put into a tube immersed in an ice bath, and mixed with 18
162
mL of a solution of chloroform and methanol 1:1 for 1 min using an Ultra-Turrax T25 (IKA-Werke
163
GmbH & Co., Staufen, Germany); then 12 mL of a solution of chloroform and deionized water 1:1
164
were added to the sample and mixed. After centrifugation (4000 rpm at 4 °C for 10 min), the lower
165
phase containing chloroform and dissolved TL, was separated and passed through a layer of
166
anhydrous sodium sulfate. One milliliter of the lipid extract was completely desolvated on a hot
167
plate at 50 °C, the lipid residue was weighed and the percentage of TL was calculated.
168
Quantification of phospholipids (PL) in TL was carried out by colorimetric determination of
169
phosphorus, following the procedure reported by Marinetti (1962) 27. Briefly, 5 µg of TL extracted
170
sample was mineralized with perchloric acid at 250 °C. Afterwards, ammonium heptamolybdate
171
and 1-amino-2-hydroxy-4-naphthalene sulfonic acid were added to the samples; they were kept at
172
100 °C for 7 min, till a colored complex was obtained. The absorbance of this complex was 7 ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
Page 8 of 38
173
measured at 830 nm to estimate the amount of phosphorus in the samples; PL quantification was
174
obtained by multiplying the calculated amount of phosphorus by 25.
175
Neutral lipids (NL) were separated from PL following a normal phase SPE (solid phase
176
extraction) method, described by Bayır et al. (2010) 28, which uses 12 mL Strata SI-1 Silica (55 µm,
177
70 Å) columns, with 2 g of substrate (Phenomenex, Torrance, CA, USA). Briefly, after
178
equilibration with 3 mL of chloroform, silica columns were loaded with TL sample, and diluted in 3
179
mL of chloroform. After that, the NL fraction was separated by charging the cartridge with 3 mL of
180
chloroform eight times. After vacuum drying of the column, PL were eluted with four aliquots of 3
181
mL of methanol and then four aliquots of 3 mL of a solution of chloroform/methanol 3 : 7 (v : v).
182
Fatty acid methyl esters (FAME) from TL, NL and PL were obtained by transmethylation 29
183
involving sulfuric acid as catalyst, following the method proposed by Christie (1989)
184
samples were dried under a gentle nitrogen stream and diluted with 100 µL of toluene and 1 mL of
185
the methylating solution of 1% sulfuric acid in methanol (96% purity). Samples were kept at 50 °C
186
in a heater for 12 h. Then, 1 mL of 5% NaCl buffer and 900 µL of hexane were added to the
187
samples, which were successively vortex-mixed and centrifuged for 10 min at 2000 rpm. The
188
supernatant containing FAME was then collected and injected into a Varian 3380 gas
189
chromatograph (Agilent Technologies, Palo Alto, USA) fitted with a CP-8200 Varian autosampler,
190
a split injector set at 230 °C, and a flame ionization detector system set at 300 °C. Chromatographic
191
separation of FAME was attained by means of a 30 m × 0.32 mm (i.d.) × 0.25 µm (film thickness)
192
fused silica-bonded phase column (DB-23, J&W Scientific); the oven temperature was programmed
193
from 150 to 230 °C at a rate of 5 °C min−1, with a final isotherm. FAs were identified by comparing
194
the retention times of unknown FAME with those of known FAME standard mixtures, their content
195
being reported as a percentage of the sum of FAME (% FAME) of TL, NL and PL, respectively.
196
The complete description of the procedures for moisture determination, lipid analysis and fatty acid
197
profile is reported in section 2.3 of supporting information on Materials and Methods.
198 8 ACS Paragon Plus Environment
. Briefly,
Page 9 of 38
199
Journal of Agricultural and Food Chemistry
2.4 Determination of macro-, microelements and toxic elements
200
Mineralization was performed by adding 6 mL of nitric acid (Ultrapure Merck, Darmstadt,
201
Germany, 67% purity) and 2 mL of hydrogen peroxide (Ultrapure Merck, 31% purity) to 1 g of
202
sample, through a MULTIWAVE 3000 microwave system (Perkin Elmer). Detection of the
203
elements was obtained by means of an Optima 2100 inductively coupled plasma atomic emission
204
spectrometer (ICP-OES; PerkinElmer, Waltham, MA, USA). For determination of the
205
macroelements a Meinhard cyclonic spray chamber was employed, with radial viewing
206
configuration. Microelements and toxic elements were determined with a CETAC U5000 ultrasonic
207
nebulizer (Thermo Fisher Scientific, Waltham, MA, USA) in axial view configuration. The
208
complete description of the procedure is reported in section 2.4 of supporting information on
209
Materials and Methods.
210 211
2.5 Stable isotopes analysis
212
Determination of C and N isotopic composition in sea bass muscle was carried out through the
213
procedure described below: 0.07 ± 0.01 and 0.7 ± 0.1 mg, respectively for δ13C and δ15N analysis,
214
of freeze dried fillets (< 0.6 mm grain size), were analyzed by means of an EA/NA-1100 elemental
215
analyzer with CHN configuration (Thermo Finnigan), in helium continuous flow mode and coupled
216
to a Finnigan Delta Plus XP mass spectrometer. Sample isotope ratios were then calculated
217
according to international reference standard V-PDB for C, and atmospheric nitrogen for N. The
218
complete description of the procedure is reported in section 2.5 of supporting information on
219
Materials and Methods.
220 221
2.6 Quality of analytical data
222
Analyses of samples were performed in duplicate and data were constantly monitored by means
223
of standard reference materials (SRMs). During analysis, each SRM was processed in duplicate
224
three times, following the same procedures. The accuracy and precision of the measurements, 9 ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
225
calculated on the basis of repeated analysis of SRMs, was ± 0.2‰. The descriptions of the SMRs
226
used are reported in section 2.6 of supporting information on Materials and Methods.
227 228
2.7 Data processing and statistical analysis
229
MATLAB software was used for data processing, PCA analysis and for sample classification
230
through Discriminant Analysis (MATLAB “classify” function embedded into ad-hoc built scripts to
231
apply crossvalidation procedures and to estimate classification performance).
232
Every source was characterized using biometric parameters and compositional characteristics of
233
the entire sea bass fillets (as edible part), grouped into the following sets of variables: biometric
234
parameters, FA composition of TL, FA composition of NL, FA composition of PL, elemental
235
composition, and isotopic abundances (Table 1). The whole set of measures, provided by each
236
analytical technique, was statistically analyzed in order to obtain an “analytical signature” for every
237
source; first of all, average, range, standard deviation, and coefficient of variation were calculated
238
for each source. Therefore, for the automatic classification of samples, mono- and multivariate
239
statistical analyses were employed, with the aim of predicting the production method (wild or
240
farmed sea bass), geographical origin (FAO fishing subarea for wild subjects, and country of origin
241
for farmed ones), and the so-called “stocking density” factor.
242
Analyses of single parameters were carried out for each classification target, to determine if
243
different groups of subjects were significantly different from each other. To compare wild and
244
farmed sea bass, Student’s t-test was applied, accepting as significant P values ≤ 0.05; to compare
245
wild subjects from different FAO fishing subareas, farmed subjects from different countries, and
246
sea bass classified according to stocking density, one-way analysis of variance (ANOVA) was
247
conducted, accepting as significant P values ≤ 0.05, plus a post hoc Bonferroni correction due to the
248
high number (175) of parameters analyzed. Multiparametric analyses (principal component analysis
249
and classification) were conducted on each aforementioned set of variables, separately. The
250
principal component analysis (PCA) procedure was applied to give an overview of the ability of 10 ACS Paragon Plus Environment
Page 10 of 38
Page 11 of 38
Journal of Agricultural and Food Chemistry
251
selected parameters to discern samples according to production method, geographical origin, and
252
stocking density, by visual inspection of the parameter space. The aim of the multivariate statistical
253
analysis was to assess the classification performance of each set of variables, and to determine an
254
optimal and restricted parameter set for sample classification. For these purposes, samples were
255
categorized according to quadratic discriminant classifier. The “leave-one-out cross-validation”
256
procedure allowed the assessment of method robustness during classification. This validation
257
method consists of removing from the whole sample set, one at time, every single subject, which is
258
going to be classified on the basis of information concerning all the other samples 30.
259 260
3 Results and Discussions
261
Each sample was processed to collect data about biometric characteristics and chemical
262
composition (moisture, lipid content, FA composition, macro- and microelements, and isotopic
263
abundance). Data obtained are reported in Supplementary Material Table S2.
264 265 266
3.1 Biometric traits Weight and length of wild subjects analyzed in this study were similar to those reported in other 4,8,24
. Condition factor is considered a good indicator of dietary condition of fish species 32;
267
studies
268
likewise what reported in other works
269
hepatosomatic index were found among subjects grouped according to the stocking density (P