G Model

CHROMA-460431; No. of Pages 11 ARTICLE IN PRESS

Journal of Chromatography A, xxx (xxxx) xxx

Contents lists available at ScienceDirect

Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma

Ultrasound assisted combined molecularly imprinted polymer for the

selective micro-solid phase extraction and determination of aflatoxins
in fish feed using liquid chromatography-tandem mass spectrometry
G.D. Thilini Madurangika Jayasinghe, Raquel Domínguez-González,
Pilar Bermejo-Barrera, Antonio Moreda-Piñeiro ∗
Trace Element, Spectroscopy and Speciation Group (GETEE), Strategic Grouping in Materials (AEMAT), Department of Analytical Chemistry, Nutrition and
Bromatology, Faculty of Chemistry, Universidade de Santiago de Compostela, Avenida das Ciencias, s/n, 15782, Santiago de Compostela, Spain

a r t i c l e i n f o a b s t r a c t

Article history: A combined procedure based on using ultrasounds for target isolation followed by porous membrane-
Received 26 April 2019 protected micro solid phase extraction using a molecularly imprinted polymer as an adsorbent has been
Received in revised form 27 July 2019 developed as a highly selective extraction and clean-up procedure for isolating aflatoxins B1, B2, G1,
Accepted 5 August 2019
and G2 from fish feed before ultra-high-performance liquid chromatography tandem mass spectrometry
Available online xxx
determination. Polymeric adsorbent beads have been synthesized by the precipitation polymerization
method which guarantees a homogeneous particles size distribution and the integrity of the gener-
Keywords:
ated imprinted cavities. In addition, polymerization was performed using a higher proportion of organic
Molecularly imprinted polymer
Porous membrane protected
solvent (toluene) in the porogen mixture, which generates MIP particles adequate for interacting with
Aflatoxins targets dissolved in organic (hydro-organic) mixtures (extracts from fish feed). These approaches led to
Fish feed a selective and high efficient pre-concentration method for AFs. Ultrasound-assisted extraction (10 mL
Tandem mass spectrometry of 60:40 acetonitrile/0.1 M KH2 PO4 pH 6.0, 40% amplitude, continuous sonication for 7.0 min) allowed
Ultrahigh-performance liquid an efficient aflatoxins isolation from fish feed. In addition, the resulting pH of the extract (pH 7.0) has
chromatography been found to be the optimum for performing clean-up/pre-concentration (enrichment factor of 33.3)
by molecularly imprinted polymer based micro-solid phase extraction (orbital horizontal shaking speed
at 150 rpm for 10 min for loading, and 5 mL of 95:5 acetonitrile/formic acid as eluting solution using
ultrasounds 35 kHz for 15 min). The current proposal was shown to be an accurate and precise method
through relative standard deviation of intraday and inter-day tests below 20% and analytical recoveries
in the range of 80–100%. The limits of detection were within the 0.42–1.15 ␮g kg−1 range, quite lower
than those established by European Commission guidelines for aflatoxins in animal feeds.
© 2019 Elsevier B.V. All rights reserved.

1. Introduction ecotoxic reagents, and the prevention of waste generation. Sorbent-

, liquid phase- and membrane-based extraction/microextraction
Besides facing the complexity of many food/feed and envi- procedures have shown to be appealing techniques for pre-treating
ronmental materials, developed analytical methods must to be liquid samples [1–3]; whereas, assisted extraction procedures
sensitive enough to detect/determine pollutants at the low lev- based on ultrasounds, microwaves, pressurization, and matrix solid
els (maximum acceptable/residue level, MRLs) proposed by the phase dispersion (MSPD) can be successfully applied for target
international organisations. A previous sample pre-treatment stage isolation from solid matrices. Advanced assisted procedures for
is therefore needed before analysis by modern instrumental solid materials lead to high analyte extraction efficiencies, but
techniques. Sample preparation techniques should be fast and co-extraction of other matrix components is also important and fur-
repeatable procedures which avoid analytes losses and guarantee ther clean-up stages are required [4]. Solid phase extraction (SPE) is
analyte integrity. Modern trends in sample pre-treatment methods commonly used for clean-up and pre-concentration purposes [5],
imply the minimization of organic solvents and/or highly toxic or and recent trends are focused on miniaturizing the extractive pro-
cess (micro solid phase extraction, ␮-SPE), increasing the sample
throughput, and also developing adequate interfaces for online cou-
pling with the detector systems [6]. In addition, the use of new
∗ Corresponding author. adsorbents for SPE such as carbon nanotubes (CNTs), molecularly
E-mail address: [email protected] (A. Moreda-Piñeiro).

https://doi.org/10.1016/j.chroma.2019.460431
0021-9673/© 2019 Elsevier B.V. All rights reserved.

Please cite this article in press as: G.D.T.M. Jayasinghe, et al., Ultrasound assisted combined molecularly imprinted polymer for the selec-
tive micro-solid phase extraction and determination of aflatoxins in fish feed using liquid chromatography-tandem mass spectrometry,
J. Chromatogr. A (xxxx), https://doi.org/10.1016/j.chroma.2019.460431
G Model
CHROMA-460431; No. of Pages 11 ARTICLE IN PRESS
2 G.D.T.M. Jayasinghe et al. / J. Chromatogr. A xxx (xxxx) xxx

imprinted polymers (MIPs) and (magnetic) nanoparticles ((M)NPs) uum degasser (Perkin Elmer, Waltham, MA, USA) and with a
combined with ␮-SPE have been reported to increase selectivity Flexar UHPLC autosampler (Perkin Elmer). Chromatographic sep-
and hence, applicability for treating complex samples [4]. arations were achieved on a Zorbax C18 reverse phase column
Aflatoxins (AFs) are a class of structure related mycotoxins (100 mm length, 4.6 mm i.d, 3.5 ␮m particle diameter) from Agi-
produced as secondary metabolites by fungi belonging to several lent (Santa Clara, CA, USA) connected to a C18 guard column
Aspergillus species, and AFs referred as AFB1, AFB2, AFG1 and AFG2, (4 mm length, 3.0 mm i.d) from Phenomenex (Torrance, CA, USA)
have been identified as the most dangerous and most frequently under controlled temperature (40 ◦ C) using a GECKO 2000 col-
AFs found in food/feed. AFs extraction from solid materials requires umn heater from Amchro GmbH (Hattersheim, Germany). A
aqueous mixtures of polar organic solvents (methanol, acetonitrile low-profile roller (Stovall, Greensboro, NC, USA), placed inside a
or acetone) [7], and most of developed procedures imply mechani- Boxcult temperature-controlled chamber (Stuart Scientific, Sur-
cal shaking (extraction times from 3 min to 72 h) [8–11]. Ultrasound rey, UK), was used for MIP synthesis. The same Boxcult chamber
assisted extraction (UAE) procedures with ultrasounds water-bath equipped with a Rotabit orbital rocking platform shaker (J.P.
have been also developed to speed-up AFs extraction (times within Selecta, Barcelona, Spain) was used for performing the loading step
the 5–10 min range) from rice [12] and soy-based products [13]. of the MIP-␮-SPE procedure. Elution stage was assisted by using a
In addition, microwave assisted extraction (MAE) has been also Raypa UCI-150 ultrasonic cleaner water bath (ultrasound frequen-
reported for AFs extraction from peach seed, milk powder and corn cies of 17 and 35 kHz, 325 W) from R. Espinar S.L (Barcelona, Spain).
flour (extraction time of 10 min) [14]. Although liquid-phase based A VibraCell VCx 130 ultrasound probe from Sonics (Newtown, CT,
extraction and microextraction (homogenous liquid-liquid extrac- USA) was used for AFs isolation from fish feed. A Laborcentrifugen
tion HLLE; and dispersive liquid-liquid microextraction, DLLME) 2K15 centrifuge (Sigma, Osterode, Germany) was used for extract
have been proposed for clean-up and/or pre-concentration before isolation from fish feed. A VLM EC1 metal block thermostat and
AFs isolation [8,11], most of developments are based on solid phase nitrogen sample concentrator from VLM (Leopoldshohe-Greste,
extraction (SPE) [9,10,12,13] and also on dispersive solid phase Germany) was used for extract solvent removal. MIP/NIP charac-
extraction/microextraction (D-SPE/D-␮-SPE) [14]. Novel magnetic terization was performed by using a Spectrum two FT-IR Fourier
nanocomposites sorbents are currently being used to induce dis- transform infrared spectrometer (FT-IR) from Perkin Elmer, an
persion in D-SPE/D-␮-SPE [15], and magnetite-graphene and nano Ultra Plus field emission scanning electron microscope from Zeiss
zirconia have been proposed for AFs extraction [12–14]. (Oberkochem, Germany) for SEM images, and a Micromeritics ASAP
Other group of ␮-SPE procedures imply the adsorbent enclosing 2000 (Norcross, GA, USA) for BET and porosity measurements.
inside a polypropylene (PP) membrane which allows freely ana- Soxhlet extraction systems consisted of a 200 mL glass still pot
lytes diffusion before adsorption onto the enclosed solid adsorbent attached to a glass distillation path, and a glass condenser, and
[16]. As recently reviewed by Sajid [17], conventional and new heated with a Pilz WHG2 laboratory heating mantle from Win-
adsorbents such as carbon based sorbents, zeolites, metal-organic kler (Heidelberg, Garmany). Other devices were: a vibrating ball
frameworks (MOFs), mesoporous silica based materials, and molec- mill with 15 mL zircon oxide cups and 7 mm diameter zircon balls
ularly imprinted polymers (MIPs), have been proposed in ␮-SPE. (Retsch, Haan, Germany), a Basic20 pH-meter with a glass-calomel
MIP-based adsorbents offer excellent selectivity and they can be electrode (Crison, Barcelona, Spain), a Classic ML analytical bal-
used for targets pre-concentration purposes and also decontami- ance (Mettler Toledo, Columbus, OH, USA), a Selecta 207 oven
nation of food and pharma samples [18,19]. The use of molecularly (Barcelona, Spain), and a Lauson heat-sealer (Barcelona, Spain).
imprinted polymers (MIPs) as adsorbents in ␮-SPE (molecularly UHPLCMS/MS data processing was performed with MultiQuant
imprinted micro-SPE, MIMSPE [20]) has been also reported for sev- 2.1 software (ABSciex).
eral pollutants in environmental samples [20,21], foodstuff [22],
and clinical/forensic materials [23–28]. 2.2. Reagents
Regarding AFs, magnetic-MIP composites [29] and magnetic
molecularly imprinted stir bars [30] have been recently developed Aflatoxins stock standards solutions (1000 mg L−1 ) were pre-
for corn and infant cereal-based food and infant formula analysis pared from solid AFs (AFB1, AFB2, AFG1 and AFG2) from
[29–31]. In addition, MIP technology was also used for prepar- Sigma-Aldrich (Steinheim, Germany) dissolved in methanol
ing monolithic molecularly imprinted polymeric capillary columns (LC–MS grade) purchased from Merck (Darmstadt, Germany). Sim-
for AFB1 chromatographic separations [32] and MIP@quantum ilarly, 5,7-dimethoxycoumarin (DMC) from Sigma-Aldrich was also
dot composites for fluorescent screening [33,34]. As the best prepared in methanol (stock standard solution of 1000 mg L−1 ) and
of our knowledge, MIMSPE for AFs pre-concentration has not it was used as an internal standard for sample pre-treatment opti-
been reported. MIMSPE (use of single cone-shape PP devices as mization. U-[13 C17 ]-AFB1 in acetonitrile (certified concentration
a ␮-SPE system) is therefore a novel development for AFs pre- of 0.501 ± 0.0008 ␮g mL−1 ) was purchased from LGC Standards
concentration. Improvements based on MIP particles synthesis (Wesel, Germany), and it was also used as an internal stan-
by the precipitation polymerization method in presence of large dard for method validation. All standard solutions were stored
amount of organic solvent (toluene) led to homogeneous MIP at −20 ◦ C in the dark. Retinol, 7-dehydrocholesterol, and ␤-
particles size distributions and to effective MIP particle-analyte carotene stock solutions (1000 mg L−1 ) were prepared in methanol
interactions when performing AFs pre-concentration from organic from solid reagents (Sigma-Aldrich). Other compounds used in
solutions (fish feed extracts). cross-reactivity studies, such as carrageenan (kappa and lambda)
were from CEAMSA (Porriño, Spain), agar-agar were from Alga-
mar (Redondela, Spain), and carboxymethyl cellulose (CMC) from
2. Materials and methods Scharlau (Barcelona, Spain). These reagents were dissolved in hot
water to prepare 1000 mg L−1 stock standard solutions. MIP syn-
2.1. Instrumentation thesis required methacrylic acid (MAA) and divinylbenzene (DVB)
from Sigma-Aldrich, and 2,2’-azobisisobutyronitrile (AIBN) from
Aflatoxins (AFB1, AFB2, AFG1, AFG2) determination was per- Fluka (Buchs, Switzerland). Solvents such as acetonitrile, methanol
formed with a 3200 Q TRAP LC/MS/MS system from ABSciex (HPLC grade), and reagents such as ammonium acetate, neutral
(Concord, Canada) equipped with electrospray ionization source, alumina, potassium dihydrogen phosphate, and sodium hydrox-
with a Flexar FX-15 UHPLC binary pump with integrated vac- ide were purchased from Merck. Toluene (HPLC grade), acetic

Please cite this article in press as: G.D.T.M. Jayasinghe, et al., Ultrasound assisted combined molecularly imprinted polymer for the selec-
tive micro-solid phase extraction and determination of aflatoxins in fish feed using liquid chromatography-tandem mass spectrometry,
J. Chromatogr. A (xxxx), https://doi.org/10.1016/j.chroma.2019.460431
G Model
CHROMA-460431; No. of Pages 11 ARTICLE IN PRESS
G.D.T.M. Jayasinghe et al. / J. Chromatogr. A xxx (xxxx) xxx 3

Fig. 1. Schematic of preparation of MIP-␮-SPE devices.

acid (glacial) were from Panreac (Barcelona, Spain). Certified refer- linker) and 0.1 g of AIBN (radical polymerization initiator – reagent
ence material ERM-BE376 (compound feeding stuff) was purchased which generates radicals that initiate free-radical polymerizations
from European Commission, Joint Research Centre, Institute for –) were added into the pre-polymerization mixture and stirring
Reference Materials and Measurements (Geel, Belgium). Ultra- for 1 min. The mixture was purged with nitrogen for 5 min and
pure water, 18 M cm of resistivity, was obtained from a Milli–Q immediately sealed and placed in a low-profile roller (33 rpm on
purification device (Millipore Co., Bedford, MA, USA). Other used its long axis) inside a temperature-controllable chamber (the tem-
consumables were: ACCUREL® PP membrane (Membrana, Wup- perature was ramped from room temperature to 60 ◦ C for 2 h, and
pertal, Germany), cellulose thimbles for Soxhlet extraction (33 mm then maintained at 60 ◦ C for 24 h).
i.d, 37 mm e.d., 118 mm height) from Prat-Dumas (Couze-et-Saint- DVB and AIBN are highly reactive and certain stabilizers (poly-
Front, France), Durapore 0.20 ␮m membrane filters (Millipore), merization inhibitors) are added to the commercial reagents.
nonsterile MCE syringe filters (0.45 ␮m) from Sterlitech (Kent, Therefore, DVB was purified by passing a few milliliters of the
WA, USA), and disposable syringe (sterile, 2 mL) from Dispomed reagent through a previously prepared mini column containing
(Gelnhausen, Germany). neutral alumina (approximately 0.5 g). Similarly, AIBN was purified
by crystallization at −20 ◦ C after dissolving the reagent in methanol
at 50–60 ◦ C.
2.3. Fish feed samples
After finished the polymerization, the synthesized material was
vacuum filtered and washed with acetonitrile (20 mL, 3 times) and
Fish feed samples were from local fish feed manufacturers in
oven dried overnight at 40 ◦ C.
Santiago de Compostela, Spain. All fish feed samples were ground
Non-Imprinted polymers (NIPs) were also prepared by follow-
by vibrating ball mill, and they were stored at −20 ◦ C until further
ing the same method as MIPs, but without adding the template
analysis.
(DMC). Synthesized NIPs were then subjected to the same filtering
and washing steps described above.
2.4. Synthesis of MIP particles

2.5. Template removal procedure

Because of the high toxicity of AFs, a dummy template is pre-
ferred for the synthesis of the imprinted polymers [30]. However,
DMC template was removed from the synthesized MIP by Soxh-
dummy templates must be chemical structurally similar to analytes
let extraction. Approximately, 300 mg of dried MIP beads were
under assessment so that the generated recognition cavities can be
placed into a cellulose thimble and were treated with 200 mL of
complementary to those analytes. Since AFs are difuranocoumarin
85:10:5 methanol/water/acetic acid mixture until template was
derivatives, AFs are chemical structurally similar to DMC (5,7-
not detected in the washing solutions (UHPLC-MS-MS analysis).
dimethoxycoumarin), and they exhibit the same methoxycoumarin
After Soxhlet extraction (total template removal), the material
structure (Fig. S1, electronic supplementary information ESI). DMC
inside the cellulose thimble was fully rinsed with ultrapure water
is therefore a dummy template of choice when preparing MIPs for
and then oven-dried at 40 ◦ C for 12 h before use.
AFs recognition [29,30,32–34].
DMC (0.0699 g) was mixed with 115 ␮L of MAA and 25 mL of
porogen (1:3 acetonitrile/toluene) into 30 mL glass tube. The mix- 2.6. Preparation of the MIP- SPE device
ture was stirred for 5 min and covered with aluminium foil and
kept in the dark overnight to allow self-assembly of the template MIP-␮-SPE device was prepared by cutting the PP membrane
and monomer. A large porogen volume (25 mL) leads to further MIP in circle shape (12 cm diameter), and by folding 3 times to obtain
synthesis by precipitation polymerization, which allows homoge- a quarter circle (cone) as shown in Fig. 1. MIP particles (approxi-
neous and dispersed MIP particles and guarantees the integrity of mately 50 mg) were placed into either one of the closed-end folds of
the generated imprinted cavities. In addition, the high proportion of the cone, and the upper part of the device was heat-sealed. Before
toluene in the porogen mixture allows MIP particles with recogni- use, all prepared MIP-␮-SPE devices were conditioned by sonica-
tion cavities more available for target retention in organic solvents tion with 5 mL of 0.1 M KH2 PO4 /NaOH buffer solution (pH 6) for
(acetonitrile/water extracts from fish feed samples). 10 min. The MIP-␮-SPE devices are soaked in buffer solution mean-
After template-monomer self-assembly, 1.25 mL of previously while they were stored. Control NIP-␮-SPE devices were prepared
purified DVB (reagent containing vinyl groups that acts as a cross- using the same procedure as described above. Several MIP-␮-SPE

Please cite this article in press as: G.D.T.M. Jayasinghe, et al., Ultrasound assisted combined molecularly imprinted polymer for the selec-
tive micro-solid phase extraction and determination of aflatoxins in fish feed using liquid chromatography-tandem mass spectrometry,
J. Chromatogr. A (xxxx), https://doi.org/10.1016/j.chroma.2019.460431
G Model
CHROMA-460431; No. of Pages 11 ARTICLE IN PRESS
4 G.D.T.M. Jayasinghe et al. / J. Chromatogr. A xxx (xxxx) xxx

Fig. 2. UAE-MIMSPE workflow for AFs extraction from fish feed.

devices (at least 20) can be arranged inside the temperature- Multi reaction monitoring mode (MRM) was used for data acquisi-
controlled chamber (Fig. 1). tion, and m/z (precursor ions) → m/z (product ions) transitions were
acquired using positive electrospray ionization under optimized
2.7. AFs isolation from fish feed by ultrasound assisted extraction ion source potentials and collision energies listed in Table 1. Fig. 3
shows a chromatogram for a mixture of AFs (standard) and also
Dried/homogenized fish feed (0.250 g) was mixed into a 30 mL for a fish feed the sample under optimized sample pre-treatment
centrifuge tube with 10 mL of 60:40 acetonitrile/0.1 M KH2 PO4 conditions.
buffer pH 6 (pH adjusted by using 0.1 M NaOH). The pH of the final Standard addition graphs were prepared in duplicate (10 mL of
extracting solution is 7.0. The tube containing the sample and the a fish feed extract) by spiking with AFs standards covering concen-
extractant was placed in an ice-bath, and the mixture was then trations of 0.075, 0.15, 0.30, 0.60, 1.5, and 3.0 ␮g L−1 . Taking into
ultrasonicated for 7.0 min at 40% amplitude (40% of the ultrasonica- account a pre-concentration factor of 33.3 AFs concentrations were
tor power/frequency (130 W/20 kHz)) using continuous sonication. 2.5, 5.0, 10, 20, 50, and 100 ␮g L−1 in the reconstituted extract. DCM
Ultrasound dissipation in the solid sample-extractant mixture leads (1.5 ␮g L−1 ) or U-[13 C17 ]-AFB1 (1.5 ␮g L−1 ) were used as internal
to an efficient AFs extraction in short times and also a temperature standards for sample pre-treatment optimization and for valida-
increase which is minimized by placing the test tube in an ice-bath. tion, respectively. By assuming a pre-concentration factor of 33.3,
Finally, the extract was isolated by centrifugation (3000 rpm, 4 ◦ C, the concentration of internal standards in the reconstituted solu-
10 min). tion was 50 ␮g L−1 DCM or U-[13 C17 ]-AFB1.

2.8. AFs pre-concentration by MIMSPE 2.10. Statistical analysis

Statgraphics Centurion XVI v16.1.15 (Manugistics Inc., Rockville,

The isolated extract (10 mL) was directly subjected to MIMSPE
MD, USA) software was used for statistical (Cochran C-test and
by mechanical orbital-horizontal shaking at 150 rpm and at room
ANOVA) evaluation for comparing standard deviation and mean
temperature for 10 min. The MIP-␮ SPE devices was then separated
values, respectively.
with tweezer and rinsed with 5 mL of 0.1 M/0.1 M KH2 PO4 /NaOH
(pH 6.0). Elution stage was performed by sonication with 5 mL
of 95:5 acetonitrile/formic acid for 15 min (water-bath sonication, 3. Results and discussion
35 kHz, 325 W). Finally, the eluate was dried under N2 gas and it
3.1. Characterization
was again dissolved in 300 ␮L of methanol. Taking into account
the volume of the fish feed extract (10 mL) and the volume of the Synthesized materials (MIP and NIP) have been characterized
re-dissolved eluate after MIMSPE and drying (300 ␮L), the enrich- by SEM and FT-IR. High degree of agglomeration of MIP/NIP spher-
ment factor was 33.3. Retention properties of MIP beads remained ical particles was observed in SEM images (Fig. S2(a, b)), electronic
constant at least after 20 uses. supplementary information ESI), and beads of approximately 5 ␮m
Fig. 2 shows a workflow summarising the steps involved in the and 3 ␮m in diameter were obtained for MIP and NIP, respectively.
UAE-MIMSPE procedure. FT-IR spectra (Fig. S3, ESI) provided similar information concern-
ing the functioning groups present in MIP and NIP beads [bands
2.9. UHPLC-MS/MS measurement at 1350 and 2900 cm−1 (C H stretching) and at 1650 cm−1 (C O
stretching)]. These findings demonstrate that MIP (after template
Chromatographic separation was performed under gradient removal) and NIP have similar characteristic bonds, which implies
elution [0.1% formic acid in ultrapure water (mobile phase A) and a successful template removal without MIP damage.
0.1% formic acid in methanol (mobile phase B)] at a flow rate of BET and porosity studies (Table S1, ESI) show that MIP (before
250 ␮L min−1 . MS/MS acquisition settings are listed in Table 1. and after template removal) and NIP can be classified as meso-

Please cite this article in press as: G.D.T.M. Jayasinghe, et al., Ultrasound assisted combined molecularly imprinted polymer for the selec-
tive micro-solid phase extraction and determination of aflatoxins in fish feed using liquid chromatography-tandem mass spectrometry,
J. Chromatogr. A (xxxx), https://doi.org/10.1016/j.chroma.2019.460431
G Model
CHROMA-460431; No. of Pages 11 ARTICLE IN PRESS
G.D.T.M. Jayasinghe et al. / J. Chromatogr. A xxx (xxxx) xxx 5

Table 1
Instrument parameters for UHPLC-MS/MS.

MS-MS Compounda Precursor ion (m/z) Product ion (m/z) DP (V)b EP (V)b CE (V)b CXP (V)b

DMC 206.800 121.100 46.590 7.630 34.150 3.000

AFB1 313.000 241.000 70.390 4.800 52.340 7.230
AFB2 315.000 259.200 84.900 2.820 39.320 2.240
AFG1 329.000 200.120 59.940 4.180 55.140 4.500
AFG2 331.000 213.200 68.780 4.110 35.000 3.000
U-[13 C17 ]-AFB1 312.800 268.100 41.700 3.370 20.070 3.870
HPLC
Column Zorbax Eclipse C18 reverse phase column (100 mm length ×4.6 mm i.d, 3.5 ␮m particle diameter)
Injection volume 20 ␮L
Flow rate 250 ␮L min−1
Mobile phase composition 0.1% formic acid in ultrapure water (A) and 0.1% formic acid in methanol (B)

Total time (min) A% B%

0.1 50 50
3.0 50 50
8.0 0 100
12.0 0 100
13.0 50 50
15.0 50 50
a
Electrospray operation conditions are: Ion spray voltage (IS), 5500 kV; Ion source temperature, 300 ◦ C; nebulizer gas and curtain gas (N2 ), 40 psi; collision gas (N2 ), high.
b
DP-Declustering Potential, EP-Entrance Potential, CE-Collision Energy, CXP-Collision cell Exit Potential.

pores (pores higher than 0.2 Å). Surface area and pore volume of tral pHs. Since the pH of the polymerization mixture was close to
MIP before and after template removal are quite similar and higher 7.0, interactions between analytes and -recognition cavities in MIP
than those values found in NIP. These findings suggest that the tem- particles are favoured at neutral pHs. It must be said that a 60:40
plate removal procedure does not change surface are and porosity acetonitrile/aqueous 0.1 M KH2 PO4 (pH 7.0) extract is obtained
of the prepared materials. In addition, higher pore diameters have when mixing acetonitrile (60%) and aqueous 0.1 M KH2 PO4 at pH
been observed in MIP than in NIP which agrees to the imprinting 6.0 (40%). In addition, strong acidic or basic conditions enhance
effect of templates in MIPs. AFs decomposition [35], which also leads to lower recoveries. Our
findings agree with those reported by other authors that applied
3.2. Optimization of MIMSPE conditions extraction techniques such as hollow fiber-solid phase microex-
traction (HF-SPME) [35] and DLLME [36] and that have reported
MIMSPE operating conditions have been optimized by using fish a neutral environment as the best conditions for enhancing AFs
feed extracts obtained from un-spiked and spiked (1.5 ␮g L−1 of extraction. Therefore, an extracting solution consisted of 60:40 ace-
each AFs and DCM as an internal standard) fish feed sample under tonitrile/aqueous 0.1 M KH2 PO4 , pH 6.0 was finally selected, which
non-optimized UAE conditions: 60:40 acetonitrile/aqueous 0.1 M gives an extract of pH 7.0 (optimum pH for MIMSPE).
KH2 PO4 (mixture pH of 6.0), 60% amplitude, for 10 min (continuous After selecting the optimum pH of the extract several exper-
sonication). Un-spiked samples were always analysed in order to iments were performed at different orbital-horizontal stirring
subtract the AFs naturally occurring in the fish feed sample and to speeds (Fig. 4(b)). High recoveries for AFB2 and AFG2 were obtained
evaluate properly the recovery of the spikes. for high speeds (within 100 and 200 rpm). However, AFB1 and AFG1
All parameters affecting the MIMSPE loading and elution condi- recoveries were gradually increased until 150 rpm, and they were
tions, as well as those affecting UAE (Section 3.3) have been studied slightly lower at 200 rpm. Recovery impairment at high speeds
by using a univariate optimization approach. agree with previous data regarding MIMSPE [24,24,28] and other
␮-SPE procedures [37,38], and it is attributed to back-diffusion
3.2.1. Loading conditions phenomena when using high shaking speed and also when using
Variables affecting AFs adsorption through the MIP’s cavities long loading times. Therefore, 150 rpm was selected as best orbital-
such as pH of the extract, loading time and orbital horizontal horizontal shaking speed for further experiments.
stirring speed were evaluated using fish feed sample (0.250 g) As shown in Fig. 4(c), back-diffusion phenomena were observed
spiked with 1.5 ␮g L−1 of each AF and DMC (50 ␮g L−1 after when using large loading times, independently of the elution (son-
pre-concentration). Elution was performed under non-optimized ication) time used. In accordance to these findings, a loading time
conditions using sonication and 5 mL of 95:5 acetonitrile/formic of 10 min was finally selected.
acid (eluting solution) for 10 min (water-bath sonication, 35 kHz,
325 W). All experiments were performed in triplicate and recover- 3.2.2. Elution conditions
ies were calculated using a matched calibration graphs. Experiments regarding elution were tested by using several elu-
Several pH values were tested to attain the best interaction tion mixtures based on acetonitrile or methanol at acid pHs (pH
between AFs and MIP recognition cavities. Since the pH of extract given by the small proportion of formic acid) and under sonication
is dependent on the final pH of the extracting solution used in UAE, (water-bath sonication, 35 kHz, 325 W). Moderate acid pHs pro-
the pH of the extracting solution for UAE and the best pH of the mote AFs protonation and hence AFs desorption from MIP particles.
extract for an efficient MIMSPE pre-concentration were studied at After preliminary studies, acetonitrile (95%) and formic acid (5%)
the same time. Several extracting solutions consisting of 60:40 ace- was the selected mixture since it offered the higher elution effi-
tonitrile/aqueous 0.1 M KH2 PO4 at four pHs 5.0, 6.0, 7.0 and 8.0 were ciency for all AFs. Regarding the elution (sonication) time, Fig. 4(c)
tested. As shown in Fig. 4(a), the highest recoveries for all AFs were shows that AFs are conveniently eluted when using the highest son-
obtained when the fish feed extract exhibits a pH of 7.0. Oxygen ication time tested (15 min) independently of the loading time fixed
atoms in AFs are responsible of hydrogen acceptor properties of AFs, for AFs loading. Therefore, an elution time of 15 min was finally
and a low protonation degree is expected when working at neu- selected.

Please cite this article in press as: G.D.T.M. Jayasinghe, et al., Ultrasound assisted combined molecularly imprinted polymer for the selec-
tive micro-solid phase extraction and determination of aflatoxins in fish feed using liquid chromatography-tandem mass spectrometry,
J. Chromatogr. A (xxxx), https://doi.org/10.1016/j.chroma.2019.460431
G Model
CHROMA-460431; No. of Pages 11 ARTICLE IN PRESS
6 G.D.T.M. Jayasinghe et al. / J. Chromatogr. A xxx (xxxx) xxx

Fig. 3. MRM chromatogram of a 200 ␮g L−1 AFB1, AFB2, AFG1 and AFG2 standard solution (a) and a fish feed sample (b).

3.3. Optimization of the UAE conditions solid samples as a consequence of sample wetting which leads a
further extractive effect by organic solvents [7,39]. As reported by
AFs are low molecular mass polar compounds, and hence extrac- Rodríguez-Cervantes et al., 40% is an adequate water proportion
tion is favoured when polar organic solvents (also water/polar when combining with the organic solvent for AFs extraction from
organic solvents mixtures) such as methanol, acetonitrile or ace- dried materials [39], and an extracting solution consisting of 60:40
tone are used [7]. Sample matrix also conditions the selection of acetonitrile/phosphate aqueous was selected in the current study.
the extracting solution. Previous reports have stated that a little The pH of the mixture was varied and AFs extractions from fish feed
amount of water enhances AFs extraction when pre-treating dried sample was successful when using slightly acid pHs. These findings

Please cite this article in press as: G.D.T.M. Jayasinghe, et al., Ultrasound assisted combined molecularly imprinted polymer for the selec-
tive micro-solid phase extraction and determination of aflatoxins in fish feed using liquid chromatography-tandem mass spectrometry,
J. Chromatogr. A (xxxx), https://doi.org/10.1016/j.chroma.2019.460431
G Model
CHROMA-460431; No. of Pages 11 ARTICLE IN PRESS
G.D.T.M. Jayasinghe et al. / J. Chromatogr. A xxx (xxxx) xxx 7

Fig. 4. Effect of the pH of extracting solution (a), the orbital horizontal shaking speed (b), and the loading time/ elution time (c) on the AFs recovery.

are explained taking into account the hydrogen acceptor capac- solutions containing 0.6 ␮g L−1 of each AF and DMC (template
ity of AFs. However, the pH of the extracting solution was fixed at molecule) prepared in an acetonitrile/aqueous phosphate mix-
7.0 because this pH was the optimum pH for achieving a further ture. Similarly, other compounds present in fish feed such as
quantitative pre-concentration by MIMSPE. Moreover, neutral pHs vitamin A and D, ␤-carotein, CMC, agar-agar, and carrageenan
guarantee AFs integrity [35]. Therefore, a 60:40 acetonitrile/0.1 M (kappa and lambda) were investigated for cross-reactivity (acetoni-
KH2 PO4 (pH 6.0), which leads to an extracting mixture of pH 7.0 trile/aqueous phosphate mixtures containing 1.0 mg L−1 of each
was selected for UAE.’ compound). Results in triplicate for MIMSPE experiments using MIP
Fig. 5(a, b) shows effect of UAE conditions (ultrasound ampli- and NIP and after UHPLC-MS-MS assessment showed extraction
tude, sonication time) on AFs extraction from fish feed samples. efficiencies closed to 100% for DMC (template) and AFs when using
MISMPE operating conditions were fixed at the optimum values, MIP-␮-SPE (Table 2). Extraction efficiencies for AFs were how-
and AFs recoveries were assessed by using matched calibra- ever lower than 30% for experiments involving NIP-␮-SPE. These
tions and after performing the experiment in triplicate. Regarding findings show that AFs (and DMC) adsorption occurs through the
ultrasound amplitude (Fig. 5(a)), the recoveries for all AFs were generated recognition cavities in the synthesized MIP. The extrac-
gradually increased until ultrasonicating at 40%; whereas, recov- tion efficiencies for other fish feed ingredients were within the
eries decreased when sonicating at high amplitudes. In addition, 12–48% range for MIP-␮-SPE, and from 14 to 50% for NIP-␮-SPE.
high recoveries were obtained when sonicating for 7 min and it These results show that retention of these compounds in MIP is not
was decreased when increasing the sonication time (Fig. 5(b)). attributed to the selective imprinting cavities but is attributed to
Although higher extraction yield could be expected when using nonspecific adsorption (surface adsorption).
high ultrasound amplitudes and extraction times, high ultrasound As listed in Table 2, high distribution ratios and low selectivity
amplitudes and sonication times affect the integrity of the extracted coefficients were obtained for DMC and AFs. Regarding the other
targets. In addition, high ultrasound amplitude and sonication time fish feed ingredients, low distribution ratios and high selectivity
increase the temperature of the mixture, which also affect AFs coefficients were obtained. In general, high distribution ratios and
integrity. low selectivity coefficients prove that MIP offers imprinting prop-
Finally, continuous (ultrasonication for 7.0 min) and discontinu- erties and high selectivity for DMC and AFs.
ous (seven cycles of ultrasonication for 59 and six cycles of relaxing
in between each ultrasonication cycle) ultrasonication modes were
tested. Results (Fig. 5(c)) showed higher extraction efficiencies 3.5. Analytical performance
when using a continuous ultrasonication (AFs recoveries close to
100%) compared to discontinuous ultrasonication (AFs recoveries 3.5.1. Calibration and matrix effect
within the 60–80% range). Discontinuous sonication is not able to Matrix effect has been estimated by comparing the slopes of
promote AFs releasing because of the short ultrasound amplitude methanol calibration curves and standard addition curves cover-
selected. Better extraction can be obtained by continuous ultrason- ing concentrations of 2.5, 5, 10, 20, 50, 100 ␮g L−1 for each AF and
ication at short ultrasonication amplitudes and short times. using 50 ␮g L−1 of U-[13 C17 ]-AFB1 as an internal standard. Standard
additions were prepared by spiking the fish feed sample (0.250 g)
3.4. Cross-reactivity and imprinting effect mixed with 10 mL of the extractant with increasing AFs concentra-
tions of 0.075, 0.15, 0.3, 0.6, 1.5, 3.0 ␮g L−1 (concentrations within
Parameters such as extraction efficiency (analytical recovery), the 2.5–100 ␮g L−1 range after MIMSPE pre-concentration).
distribution ratio (D) and selectivity coefficient (SDMC/D ) were Mean slopes (also the standard deviations) for seven methanol
calculated to study the selectivity (imprinting effect) of the syn- calibrations and seven standard addition curves obtained in differ-
thesized material for AFs. The experiment was performed using ent days are listed in Table 3. Matrix effect expressed as the ratio

Please cite this article in press as: G.D.T.M. Jayasinghe, et al., Ultrasound assisted combined molecularly imprinted polymer for the selec-
tive micro-solid phase extraction and determination of aflatoxins in fish feed using liquid chromatography-tandem mass spectrometry,
J. Chromatogr. A (xxxx), https://doi.org/10.1016/j.chroma.2019.460431
G Model
CHROMA-460431; No. of Pages 11 ARTICLE IN PRESS
8 G.D.T.M. Jayasinghe et al. / J. Chromatogr. A xxx (xxxx) xxx

Fig. 5. Effect of the ultrasound amplitude (a), the ultrasonication time (b), and the continuous/discontinuous ultrasonication mode (c) on the AFs recovery.

(percentage) between calibration and standard addition slopes was average slopes (95% confidence interval), and statistically signifi-
higher than 20% for all AFs (28% for AFB1, 21% for AFB2, 34% of AFG1 cant differences were not found between the slopes of calibration
and 49% for AFG2). A statistical evaluation was performed by com- and standard addition for AFB1 and AFB2 (p-values of 0.0999
paring the standard deviation and average values of the slopes of and 0.5854, respectively, values >0.050 at 95% significance level).
methanol calibration and standard addition at a confidence inter- However, ANOVA showed that slopes for calibration and standard
val of 95%. The application of the Cochran’C test (comparison of addition of AFG1 and AFG2 were statistically significant different at
the standard deviation of slopes) led to conclude that there were a 95% significance level (p-values of 0.0308 and 0.0318 for AFG1 and
not statistically significant differences between the standard devi- AFG2, respectively, < 0.050). Therefore, matrix effect is important
ation of the average slope for calibration and standard addition when assessing AFG1 and AFG2, and accurate results are obtained
for each AFs. Therefore, ANOVA was applied for comparing the when performing the standard addition technique for determina-

Please cite this article in press as: G.D.T.M. Jayasinghe, et al., Ultrasound assisted combined molecularly imprinted polymer for the selec-
tive micro-solid phase extraction and determination of aflatoxins in fish feed using liquid chromatography-tandem mass spectrometry,
J. Chromatogr. A (xxxx), https://doi.org/10.1016/j.chroma.2019.460431
G Model
CHROMA-460431; No. of Pages 11 ARTICLE IN PRESS
G.D.T.M. Jayasinghe et al. / J. Chromatogr. A xxx (xxxx) xxx 9

Table 2 Table 4
Extraction efficiency (%), distribution ratio (D) and selectivity coefficient for MIP-␮- Intraday and inter-day precision (RSD%), and intraday and inter-day analytical
SPE and NIP-␮-SPE. recovery (AR%) of the method.

Extraction Distribution Selectivity coefficient Intra-daya Inter-daya

efficiency (%)a ratio (D)b (SDMC/X )c
RSD AR% RSD AR%
MIP
AFB1
DMC 98 41 —
2.5 7 103 ± 1 9 95 ± 2
AFB1 95 18 2
5 —b —b 8 89 ± 2
AFB2 93 14 3
10 —b —b 11 90 ± 5
AFG1 93 13 3
20 8 100 ± 2 9 93 ± 5
AFG2 90 9 5
50 —b —b 8 90 ± 4
Vit A 12 0.14 293
100 7 101 ± 7 7 98 ± 7
Vit D 48 0.94 44
␤-Carotein 10 0.11 373 AFB2
CMC 10 0.12 342 2.5 3 100 ± 1 12 96 ± 4
Agar 37 0.59 69 5 —b —b 6 92 ± 3
Lambda 40 0.66 62 10 —b —b 13 84 ± 4
Kappa 23 0.29 141 20 14 97 ± 3 12 97 ± 2
50 —b —b 11 90 ± 5
NIP
100 8 83 ± 7 7 95 ± 7
DMC 13 2 21
AFB1 26 0.35 117 AFG1
AFB2 25 0.34 121 2.5 14 88 ± 5 9 82 ± 2
AFG1 23 0.31 132 5 —b —b 9 88 ± 3
AFG2 22 0.28 146 10 —b —b 13 88 ± 4
Vit A 14 0.17 241 20 15 92 ± 3 9 90 ± 5
Vit D 50 1 41 50 —b —b 10 86 ± 5
␤-Carotein 8 0.09 456 100 13 98 ± 8 9 92 ± 8
CMC 11 0.12 342
Agar-agar 37 0.58 71 AFG2
Lambda 44 0.79 52 2.5 11 90 ± 11 10 88 ± 2
Kappa 29 0.42 98 5 —b —b 5 89 ± 6
10 —b —b 11 86 ± 6
A1 =Amount of analyte in aqueous solution at equilibrium. 20 11 95 ± 3 10 88 ± 2
A2 =Amount of analyte enriched by MIP/NIP at equilibrium. 50 —b —b 12 93 ± 5
AT =Total amount of analyte used in extraction. 100 3 100 ± 3 7 98 ± 6
DDMC = Distribution ratio of dimethoxycoumarine (template).
a
DX =Distribution ratio of other compounds (X = vitamin A, vitamin D, ␤-carotein, n = 7.
b
carboxymethyl cellulose (CMC), agar-gar, carrageenan Lambda, carrageenan Kappa). Not evaluated.
a
% =(A2 /AT ) × 100.
b
D =(A2 /A1 ).
c
SDMC/X =DDMC /DX . blanks. LOD/LOQ values, expressed as ␮g kg−1 and assuming a pre-
concentration factor of 33.3, are listed in Table 3. These values are
Table 3
lower than the EU regulation limit for animal feed (20 ␮g/kg) [40],
Mean slopes of calibration and standard addition, and LOD and LOQ values. although slighter higher than those previously reported (some of
them within the ng kg-1 range) [29–31].’
Slope (mean ± SD) LOD (␮g LOQ (␮g
a a,b
kg–1 ) b kg–1 ) b
Calibration Standard addition
3.5.3. Precision and accuracy
AFB1 0.0088 ± 0.0015 0.011 ± 0.0029 0.42 1.3 Intraday and inter-day precision and analytical recovery were
AFB2 0.0033 ± 0.0019 0.0039 ± 0.0021 0.88 2.7
established by spiking several aliquots of the same fish feed sam-
AFG1 0.0049 ± 0.0013 0.0066 ± 0.0013 1.1 3.3
AFG2 0.0013 ± 0.0009 0.0019 ± 0.0024 1.2 3.5
ple with AFs concentrations at several concentration levels (0.075,
a
0.15, 0.3, 0.6, 1.5, 3.0 ␮g L−1 ), which implies 2.5, 5.0, 10, 20, 50
n = 7.
b and 100 ␮g L−1 after extract re-dissolution. Intraday assays con-
Pre-concentration factor of 33.3.
sisted of preparing three standard addition in three different days
by replicating the lowest concentration of all four AFs (0.075 ␮g
tions. Matrix effect was not evaluated in previous reported methods L−1 ) seven times; whereas, the other concentration levels were
based on MIP-based SPE and HPLC using MS/MS [29,30] and fluo- replicated twice. Likewise, the second and third standard additions
rescent [31] detection. were obtained by replicating the intermediate (0.6 ␮g L−1 ) and the
Finally, the regression coefficients were higher than 0.9979 for highest (3.0 ␮g L−1 ) concentration levels also seven times.
AFB1, 0.9988 for AFB2, 0.9986 for AFG1 and 0.9959 for AFG2, which Similarly, inter-day assays (precision and analytical recovery)
shows good linearity for all cases. were obtained by preparing six standard additions in seven dif-
ferent days but replicating each concentration level twice. Results
3.5.2. Limit of detection (LOD) and limit of quantification (LOQ) (Table 4) show RSD values lower than 15% for AFs concentrations
LOD and LOQ were established by subjecting eleven blank close to the LOQ of the method, and lower than 7% for intermediate
samples (blanks obtained after applying UAE and MIP-␮-SPE) to and high AFs levels. Precision is quite similar than those reported by
HPLC-MS-MS and integrating the noise at the retention times and other authors using MIT (lower than 10%) [29–31]. Good intraday
m/z (precursor ion) → m/z (product ion) for each AF. Then, LOD/LOQ and inter-day precision in the measurements are therefore demon-
values were established by using the following equations. strated. Table 4 also lists the analytical recoveries of the intraday
and inter-day assays, which were within the 80–100% range for all
S s
LOD = 3 X and LOQ = 10 X cases. The proposed method offers quantitative analytical recover-
b b
ies, which is a clear advantage over other reported method which
where b is the mean slope of the standard addition calibra- showed analytical recoveries lower than 60% [30] or lower than 80%
tions, and s is the standard deviation of the measurements of the for low AFs concentration [29]. Uncorrected matrix effect could be

Please cite this article in press as: G.D.T.M. Jayasinghe, et al., Ultrasound assisted combined molecularly imprinted polymer for the selec-
tive micro-solid phase extraction and determination of aflatoxins in fish feed using liquid chromatography-tandem mass spectrometry,
J. Chromatogr. A (xxxx), https://doi.org/10.1016/j.chroma.2019.460431
G Model
CHROMA-460431; No. of Pages 11 ARTICLE IN PRESS
10 G.D.T.M. Jayasinghe et al. / J. Chromatogr. A xxx (xxxx) xxx

Table 5 of the targets. In addition, the selective MIMSPE procedure allows

Found AFs concentrations in ERM-BE376 (compound feeding stuff) CRM.
effective pre-concentration of AFs (an enrichment factor of 33.3)
Certified value (␮g/kg) Found value (␮g/kg) Recovery % and sensitive AFs assessment (LODs within the 0.42–1.2 ␮g kg−1
AFB1 12.9 ± 1.8 12.1 ± 0.35 93 range, values lower than EU regulation limits for AFs in animal
AFB2 0.68 ± 0.10 <0.88 feed). Matrix effect was found also to be negligible for some AFs
AFG1 5.2 ± 0.8 4.9 ± 0.5 94 (AFB1 and AFB2), which implies that clean extracts are obtained.
The high clean-up efficiency is mainly attributed to the high MIP’s
selectivity through AFs and also to the PP membrane that acts
Table 6
Concentrations (expressed as ␮g kg−1 ) of AFB1, AFB2, AFG1 and AFG2 in fish feed as the barrier and avoids the interaction of large biomolecules
samples. from the fish feed extract with the MIP’s particles. The procedure
was found to be precise (RSD values lower than 20%) and accu-
[AFB1] (␮g kg−1 ) [AFB2] (␮g kg−1 ) [AFG1] (␮g kg−1 ) [AFG2] (␮g kg−1 )
rate (analytical recoveries within the 80–100% range, and good
S1 27.0 ± 1.4 16.9 ± 4.9 7.7 ± 0.1 1.5 ± 0.04
accordance between certified and found AFs concentrations in the
S2 23.1 ± 0.4 27.8 ± 0.4 13.9 ± 1.0 28.1 ± 1.4
S3 23.9 ± 4.0 4.6 ± 1.1 3.7 ± 0.04 —a
ERM-BE 376 CRM). MIP-␮-SPE devices can be re-used 20 times
S4 22.9 ± 1.4 70.7 ± 5.8 4.9 ± 1.0 —a (20 loading/eluting cycles) which offers a practical advantage over
S5 26.7 ± 2.6 42.4 ± 1.9 9.1 ± 0.2 28.2 ± 0.9 commercially available SPE cartridges which are single-use devices.
S6 24.6 ± 0.3 18.4 ± 9.4 2.2 ± 0.2 4.5 ± 0.9 The proposed material can be useful for AFs enrichment from food-
a
<1.2 ␮g kg−1 . stuff after targets extraction. The application of the method to fish
feed shows the presence of AFs in this feedstuff. Fish feed is scarcely
studied for AFs and these findings open new insights for the control
the reason of the non-quantitative analytical recoveries in these
of these substances in aquaculture products.
proposals.’
In addition, accuracy was also assessed by determining the AFs
concentrations (AFB1, AFB2, and AFG1) in the ERM-BE 376 (com- Declaration of Competing Interest
pounds in feeding stuff) CRM. The material was thoroughly mixed
and homogenized by rolling it on a low-profile roller for 30 min The authors declare that they have no known competing finan-
before use. AFs extraction and pre-concentration (n = 6) was car- cial interests or personal relationships that could have appeared to
ried out as previously detailed in the experimental section, and influence the work reported in this paper.
extracts were analyzed by UHPLC-MS-MS. As shown in Table 5,
sensitivity of the method was high enough for assessing AFB1 and Acknowledgments
AFG1 (12.05 ± 0.35 and 4.90 ± 0.54 ␮g kg−1 for AFB1 and AFG1,
respectively); however, sensitivity was not good enough to assess This work was supported by the Dirección Xeral de I+D
AFB2, which certified concentration was lower than the LOD of the - Xunta de Galicia (Spain), Grupos de Referencia Competitiva
method for this AF. (6RC2014/2016 and ED431C2018/19), and Development of a
Strategic Grouping in Material- AEMAT (ED431E2019/08). The
3.5.4. Reusability of MIMSPE devices authors also wish to thank Dr. M. José Pazos-Guldrís (Unidade
Reusability of the MIMSPE devices was studied by using de Microscopía) at Rede de Infraestruturas de Apoio á Investigación
three independent MIMSPE devices for pre-concentrating sev- e ao Desenvolvemento Tecnolóxico – Universidade de Santiago de
eral aliquots of an extract spiked with AFB1 at 3.0 ␮g L−1 . After Compostela) for SEM technical support, and Dr. Carmen Álvarez-
each loading/elution cycle, MIP beads were sonicated with 5 mL of Lorenzo (I + D Farma – Universidade de Santiago de Compostela)
0.1 M/0.1 M KH2 PO4 /NaOH (pH 6.0) for 5 min, and then soaked in a for BET and porosity measurements.
clean 0.1 M/0.1 M KH2 PO4 /NaOH (pH 6.0) solution until a new use
(Section 2.8). Results (AFB1 analytical recovery) after each succes- Appendix A. Supplementary data
sive loading/elution cycle showed good retention properties since
analytical recoveries were within the 80–120 % range after twenty Supplementary material related to this article can be found, in
successive loading/elution cycles (Fig. S4, ESI). Reusability of MIM- the online version, at doi:https://doi.org/10.1016/j.chroma.2019.
SPE devices can be therefore established at twenty times. 460431.

3.5.5. Application References

Applicability of the proposed UAE-MIP-␮-SPE and HPLC-MS-MS
was demonstrated by analyzing six fish feed samples commercially [1] T. Hyötyläinen, M.-L. Riekkola, Sorbent- and liquid-phase microextraction
available in the Spanish market. Each sample was subjected to the techniques and membrane-assisted extraction in combination with gas
chromatographic analysis: a review, Anal. Chim. Acta 614 (2008) 27–37.
optimized procedure in triplicate. Results listed in Table 6 show the [2] F. Pena-Pereira, I. Lavilla, C. Bendicho, Liquid-phase microextraction
presence of AFB1, AFB2 and AFG1 in all analysed fish feed samples; techniques within the framework of green chemistry, Trends Anal. Chem. 29
whereas, AFG2 was quantified in four samples. Fish feed samples (2010) 617–628.
[3] M. He, B. Chen, B. Hu, Recent developments in stir bar sorptive extraction,
have shown AFB1 and AFB2 concentrations higher than the Euro- Anal. Bioanal. Chem. 406 (2014) 2001–2026.
pean Commission (EC) regulation value, although AFG1 and AFG2 [4] J. Moreda-Piñeiro, A. Moreda-Piñeiro, Recent advances in combining
were lower than EU MRL (20 ␮g kg−1 ) in most fish feed samples. microextraction techniques for sample pre-treatment, Trends Anal. Chem. 71
(2015) 265–274.
[5] M. Tobiszewski, A. Mechlińska, B. Zygmunt, J. Namieśnik, Green analytical
4. Conclusions chemistry in sample preparation for determination of trace organic
pollutants, Trends Anal. Chem. 28 (2009) 943–951.
[6] J. Pawliszyn, Sampling and Sample Preparation for Field and Laboratory:
Ultrasonication combined with porous membrane protected Fundamentals and New Directions in Sample Preparation, Elsevier,
MIMSPE has been shown to be a convenient and cost-effective sam- Amsterdam, 2002.
ple pre-treatment for AFs isolation and selective pre-concentration [7] G.S. Shephard, Aflatoxin analysis at the beginning of the twenty-first century,
Anal. Bioanal. Chem. 395 (2009) 1215–1224.
from fish feed samples. Selected ultrasonication conditions allowed [8] N. Sheijooni-Fumani, J. Hassan, S.R. Yousefi, Determination of aflatoxin B1 in
an efficient AFs extraction and guaranteed the stability (integrity) cereals by homogeneous liquid–liquid extraction coupled to high

Please cite this article in press as: G.D.T.M. Jayasinghe, et al., Ultrasound assisted combined molecularly imprinted polymer for the selec-
tive micro-solid phase extraction and determination of aflatoxins in fish feed using liquid chromatography-tandem mass spectrometry,
J. Chromatogr. A (xxxx), https://doi.org/10.1016/j.chroma.2019.460431
G Model
CHROMA-460431; No. of Pages 11 ARTICLE IN PRESS
G.D.T.M. Jayasinghe et al. / J. Chromatogr. A xxx (xxxx) xxx 11

performance liquid chromatography-fluorescence detection, J. Sep. Sci. 34 spectrometry after molecularly imprinted polymer microsolid-phase
(2011) 1333–1337. extraction, Anal. Bioanal. Chem. 409 (2017) 1207–1220.
[9] M. Alcaide-Molina, J. Ruiz-Jiménez, J. Mata-Granados, M.L. de Castro, High [26] J. Sánchez-González, S. Odoardi, A.M. Bermejo, P. Bermejo-Barrera, F.S.
through-put aflatoxin determination in plant material by automated Romolo, A. Moreda-Piñeiro, S. Strano-Rossi, Development of a
solid-phase extraction on-line coupled to laser-induced fluorescence micro-solid-phase extraction molecularly imprinted polymer technique for
screening and determination by liquid chromatography–triple quadrupole synthetic cannabinoids assessment in urine followed by liquid
mass spectrometry, J. Chromatogr. A 1216 (2009) 1115–1125. chromatography–tandem mass spectrometry, J. Chromatogr. A 1550 (2018)
[10] B. Huang, Z. Han, Z. Cai, Y. Wu, Y. Ren, Simultaneous determination of 8–20.
aflatoxins B1, B2, G1, G2, M1 and M2 in peanuts and their derivative products [27] J. Sánchez-González, S. Odoardi, A.M. Bermejo, P. Bermejo-Barrera, F.S.
by ultra-high-performance liquid chromatography–tandem mass Romolo, A. Moreda-Piñeiro, S. Strano-Rossi, HPLC-MS/MS combined with
spectrometry, Anal. Chim. Acta 662 (2010) 62–68. membrane-protected molecularly imprinted polymer micro-solid-phase
[11] Z. Es’ haghi, H. Reza Beheshti, J. Feizy, Extraction of aflatoxins from food extraction for synthetic cathinones monitoring in urine, Drug Test. Anal. 11
samples using graphene-based magnetic nanosorbents followed by (2019) 33–44.
high-performance liquid chromatography: a simple solution to overcome the [28] J. Sánchez-González, S. García-Carballal, P. Cabarcos, M.J. Tabernero, P.
problems of immunoaffinity columns, J. Sep. Sci. 37 (2014) 2566–2573. Bermejo-Barrera, A. Moreda-Piñeiro, Determination of cocaine and its
[12] X.W. Lai, D.L. Sun, C.Q. Ruan, H. Zhang, C.L. Liu, Rapid analysis of aflatoxins B 1, metabolites in plasma by porous membrane-protected molecularly imprinted
B 2, and ochratoxin A in rice samples using dispersive liquid–liquid polymer micro-solid-phase extraction and liquid chromatography—Tandem
microextraction combined with HPLC, J. Sep. Sci. 37 (2014) 92–98. mass spectrometry, J. Chromatogr. A 1451 (2016) 15–22.
[13] N. Nouri, H. Sereshti, A. Farahani, Graphene-coated magnetic-sheet [29] L. Tan, R. He, K. Chen, R. Peng, C. Huang, R. Yang, Y. Tang, Ultra-high
solid-phase extraction followed by high-performance liquid chromatography performance liquid chromatography combined with mass spectrometry for
with fluorescence detection for the determination of aflatoxins B1, B2, G1, determination of aflatoxins using dummy molecularly imprinted polymers
and G2 in soy-based samples, J. Sep. Sci. 41 (2018) 3258–3266. deposited on silica-coated magnetic nanoparticles, Microchim. Acta 183
[14] L.-J. Du, C. Chu, E. Warner, Q.-Y. Wang, Y.-H. Hu, K.-J. Chai, J. Cao, L.-Q. Peng, (2016) 1469–1477.
Y.-B. Chen, J. Yang, Q.-D. Zhang, Rapid microwave-assisted dispersive [30] M. Díaz-Bao, P. Regal, R. Barreiro, C.A. Fente, A. Cepeda, A facile method for
micro-solid phase extraction of mycotoxins in food using zirconia the fabrication of magnetic molecularly imprinted stir-bars: a practical
nanoparticles, J. Chromatogr. A 1561 (2018) 1–12. example with aflatoxins in baby foods, J. Chromatogr. A 1471 (2016) 51–59.
[15] A. Chisvert, S. Cárdenas, R. Lucena, Dispersive micro-solid phase extraction, [31] O. Semong, B.S. Batlokwa, Development of an aflatoxin B1 specific
Trends Anal. Chem. (2019), http://dx.doi.org/10.1016/j.trac.2018.12.005. molecularly imprinted solid phase extraction sorbent for the selective
[16] C. Basheer, A.A. Alnedhary, B.S. Madhava Rao, S. Valliyaveettil, H.K. Lee, pre-concentration of toxic aflatoxin B1 from child weaning food, Tsabana,
Development and application of porous membrane-protected carbon Mol. Impr. 5 (2017) 1–15.
nanotube micro-solid-phase extraction combined with gas [32] M. Szumski, D. Grzywiński, W. Prus, B. Buszewski, Monolithic molecularly
chromatography/mass spectrometry, Anal. Chem. 78 (2006) 2853–2858. imprinted polymeric capillary columns for isolation of aflatoxins, J.
[17] M. Sajid, Porous membrane protected micro-solid-phase extraction: a review Chromatogr. A 1364 (2014) 163–170.
of features, advancements and applications, Anal. Chim. Acta 965 (2017) [33] A. Chmangui, M.R. Driss, S. Touil, P. Bermejo-Barrera, S. Bouabdallah, A.
36–53. Moreda-Piñeiro, Aflatoxins screening in non-dairy beverages by Mn-doped
[18] R. Viveiros, K. Karim, S.A. Piletsky, W. Heggie, T. Casimiro, Development of a ZnS quantum dots – molecularly imprinted polymer fluorescent probe,
molecularly imprinted polymer for a pharmaceutical impurity in supercritical Talanta 199 (2019) 65–71.
CO2 : rational design using computational approach, J. Clean. Prod. 168 (2017) [34] P. Guo, W. Yang, H. Hu, Y. Wang, P. Li, Rapid detection of aflatoxin B1 by
1025–1031. dummy template molecularly imprinted polymer capped CdTe quantum dots,
[19] R. Sulc, G. Szekely, S. Shinde, C. Wierzbicka, F. Vilela, D. Bauer, B. Sellergren, Anal. Bioanal. Chem. 411 (2019) 2607–2617.
Phospholipid imprinted polymers as selective endotoxin scavengers, Sci. Rep. [35] Z. Es’haghi, H. Sorayaei, F. Samadi, M. Masrournia, Z. Bakherad, Fabrication of
7 (2017) 44299. a novel nanocomposite based on sol–gel process for hollow fiber-solid phase
[20] Q. Feng, L. Zhao, J.-M. Lin, Molecularly imprinted polymer as micro-solid microextraction of aflatoxins: B1 and B2, in cereals combined with high
phase extraction combined with high performance liquid chromatography to performance liquid chromatography–diode array detection, J. Chromatogr. B
determine phenolic compounds in environmental water samples, Anal. Chim. 879 (2011) 3034–3040.
Acta 650 (2009) 70–76. [36] D. Afzali, M. Ghanbarian, A. Mostafavi, T. Shamspur, S. Ghaseminezhad, A
[21] F. Tan, M. Deng, X. Liu, H. Zhao, X. Li, X. Quan, J. Chen, Evaluation of a novel novel method for high preconcentration of ultra trace amounts of B1, B2, G1
microextraction technique for aqueous samples: porous membrane envelope and G2 aflatoxins in edible oils by dispersive liquid–liquid microextraction
filled with multiwalled carbon nanotubes coated with molecularly imprinted after immunoaffinity column clean-up, J. Chromatogr. A 1247 (2012) 35–41.
polymer, J. Sep. Sci. 34 (2011) 707–715. [37] C. Basheer, A. Jayaraman, M.K. Kee, S. Valiyaveettil, H.K. Lee, Polymer-coated
[22] T.P. Lee, B. Saad, W.S. Khayoon, B. Salleh, Molecularly imprinted polymer as hollow-fiber microextraction of estrogens in water samples with analysis by
sorbent in micro-solid phase extraction of ochratoxin A in coffee, grape juice gas chromatography–mass spectrometry, J. Chromatogr. A 1100 (2005)
and urine, Talanta 88 (2012) 129–135. 137–143.
[23] X.-Y. Yin, Y.-M. Luo, J.-J. Fu, Y.-Q. Zhong, Q.-S. Liu, Determination of [38] C. Basheer, J. Lee, S. Pedersen-Bjergaard, K.E. Rasmussen, H.K. Lee,
hyperoside and isoquercitrin in rat plasma by membrane-protected Simultaneous extraction of acidic and basic drugs at neutral sample pH: a
micro-solid-phase extraction with high-performance liquid chromatography, novel electro-mediated microextraction approach, J. Chromatogr. A 1217
J. Sep. Sci. 35 (2012) 384–391. (2010) 6661–6667.
[24] J. Sánchez-González, M.J. Tabernero, A.M. Bermejo, P. Bermejo-Barrera, A. [39] C. Rodríguez-Cervantes, A. Ramos, M. Robledo-Marenco, V. Sanchis, S. Marín,
Moreda-Piñeiro, Porous membrane-protected molecularly imprinted polymer M. Girón-Pérez, Determination of aflatoxin and fumonisin levels through
microsolid-phase extraction for analysis of urinary cocaine and its ELISA and HPLC, on tilapia feed in Nayarit, Mexico, Food Agric. Immunol. 24
metabolites using liquid chromatography - tandem mass spectrometry, Anal. (2013) 269–278.
Chim. Acta 898 (2015) 50–59. [40] COMMISSION REGULATION (EC), No 1881/2006 of 19 December 2006 Setting
[25] J. Sánchez-González, R. Salgueiro-Fernández, P. Cabarcos, A.M. Bermejo, P. Maximum Levels for Certain Contaminants in Foodstuffs, 2006, R1881— EN—
Bermejo-Barrera, A. Moreda-Piñeiro, Cannabinoids assessment in plasma and 01.07.2010 — 004.001— 1/26.
urine by high performance liquid chromatography–tandem mass

Please cite this article in press as: G.D.T.M. Jayasinghe, et al., Ultrasound assisted combined molecularly imprinted polymer for the selec-
tive micro-solid phase extraction and determination of aflatoxins in fish feed using liquid chromatography-tandem mass spectrometry,
J. Chromatogr. A (xxxx), https://doi.org/10.1016/j.chroma.2019.460431