Selective extraction and determination of neonicotinoid insecticides in wine by liquid chromatography–tandem mass spectrometry

Accepted Manuscript Title: Selective extraction and determination of neonicotinoid insecticides in wine by liquid chromatography–tandem mass spectrometry Author: T. Rodr´ıguez-Cabo J. Casado I. Rodr´ıguez M. Ramil R. Cela PII: DOI: Reference:

S0021-9673(16)30894-9 http://dx.doi.org/doi:10.1016/j.chroma.2016.07.004 CHROMA 357717

To appear in:

Journal of Chromatography A

Received date: Revised date: Accepted date:

2-6-2016 1-7-2016 4-7-2016

Please cite this article as: T.Rodr´ıguez-Cabo, J.Casado, I.Rodr´ıguez, M.Ramil, R.Cela, Selective extraction and determination of neonicotinoid insecticides in wine by liquid chromatography–tandem mass spectrometry, Journal of Chromatography A http://dx.doi.org/10.1016/j.chroma.2016.07.004 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Selective extraction and determination of neonicotinoid insecticides in wine by liquid chromatography tandem mass spectrometry T. Rodríguez-Cabo, J. Casado, I. Rodríguez*, M. Ramil, R. Cela Departamento de Química Analítica, Nutrición y Bromatología, Instituto de Investigación y Análisis Alimentario (IIAA), Universidad de Santiago de Compostela, Santiago de Compostela 15782, Spain.

1

Highlights:  Single step concentration and clean-up of neonicotinoids from wine.  Quantitative extraction yields and negligible matrix effects for red and white wines.  Accurate recoveries using calibration against standard solutions.  Imidacloprid residues often detected in commercial wines.

2

Abstract A simplified, high throughput procedure for the determination of five neonicotinoid insecticides in red and white wines, using liquid chromatography (LC) tandem mass spectrometry (MS/MS), is presented. The effects of different experimental parameters (extraction sorbent, solvent elution and clean-up conditions) in the efficiency and the selectivity of the sample preparation process were assessed through calculation of the extraction yields and the matrix effects (MEs). Wines (10 mL) were concentrated using OASIS HLB cartridges, on-line connected to Florisil clean-up cartridges, with acetonitrile serving as the elution solvent. The extract (5 mL volume) was concentrated to 1 mL and injected in the LC-ESI-MS/MS system. The optimized procedure provided quantitative extraction yields at the same time that the efficiency of ESI ionization remained unchanged between standards and sample extracts. Overall recoveries, calculated against authentic standards in ACN, varied between 77 and 119% and the attained limits of quantification remained below 0.2 ng mL-1. Analysis of commercial wines revealed imidacloprid residues in more than 50% of processed samples, with a maximum level of 14 ng mL-1. Keywords:

neonicotinoids;

wine

analysis;

solid-phase

extraction;

liquid

chromatography tandem mass spectrometry *Corresponding author e-mail: [email protected] Fax: 00 34 881814468 Tel: 00 34 881814387

3

1. Introduction Neonicotinoids have been commercialized as a new generation of pesticides affecting the central nervous system of insects, whilst they exert a lower neurotoxicity towards mammals than previously developed insecticides. As a result, they have become the most popular group of insecticides with applications in agriculture and veterinary medicine. Despite these potential advantages; nowadays, the use of neonicotinoids is a matter of concern due to their high mobility in plants [1] and environmental matrices [2], having been detected in surface water samples, obtained in the vicinity of agriculture areas, from different regions of the planet [3,4]. When introduced in agriculture fields (soil, foliar and seeds treatments) neonicotinoids might reach plants blossoms and flowers affecting pollinator insects, including honey bees contaminating also honey [5,6]. Apart from their direct impact in honey production, their toxicity for bees [7] represents a threat for the biodiversity of ecosystems. Consequently, restrictions in their agriculture uses and maximum residue levels (MRLs) in some food commodities have been established [8]. Although the environmental risks associated to neonicotinoids, and/or their primary transformation products, are still under evaluation [9,10], the European Union (EU) has already included five neonicotinoids (four active ingredients and the main degradation product of thiamethoxam, THM) in the watch list of emerging pollutants to be monitored in continental waters [11].

Production of vinification grapes has a relevant impact in the economies of template regions all around the world. Although fungicides are the main group of pesticides employed in vineyards, the consumption of insecticides is raising in order to protect plants from pest, which might decrease their productivity [12] and/or act as vectors of serious viric diseases, such as the leafroll disease [13]. Several neonicotinoids, particularly imidacloprid (IMI) and THM, have been approved to be used in vineyards in European Union countries. From fumigated plants, these compounds, and/or their transformation products, might reach grapes and pass to wine. Some preliminary

4

studies, using grapes sprayed with these insecticides after harvest, have calculated processing factors around 0.4 for IMI and acetamiprid (ACE) [14]. That is, 40% of the compounds existing in grapes are transferred to final wine. Such elevated factors are in agreement with the high polarities of neonicotinoids and thus, with their poor affinity to remain retained in peels, lees and other solid wastes during vinification operations.

Most neonicotinoids are not amenable to gas chromatography techniques; thus, liquid chromatography (LC), usually combined with tandem mass spectrometry (MS/MS) and less often with high resolution MS, is the most resorted determination technique [15]. Sample preparation methods for their extraction and concentration from wine are mainly based on generic techniques, such as QuEChERS followed by dispersive solidphase extraction clean-up [16-17], liquid-liquid extraction [18] and solid-phase (SPE) extraction on non-selective polymeric sorbents [19-20]. Less often, microextraction techniques, such as membrane-assisted solvent extraction (MASE), have been also proposed [21]. In most cases, neonicotinoids have been included in multiresidue methods dealing with dozens, or even hundreds, of pesticides. Thus, little attention was paid to the overall performance of sample preparation for these insecticides, usually representing not more than 2-3% of the targeted analytes. As a result, the determination of neonicotinoids in wines, particularly in red wines, is severely affected by matrix effects (normally signal suppression). Also, conversely to the profusion of fungicide residue data, little information could be found in relation to detection frequencies and concentration levels of neonicotinoids in commercial wines [16]. The aim of this research is to develop a high throughput and sensitive analytical method for the determination of 4 neonicotinoid insecticides and the main transformation product of THM, clothianidin (CLO), in wine samples and to provide a first overview of their occurrence and levels in commercial wines produced in Spain. Efforts were mostly focussed on optimizing a selective sample preparation

5

methodology, permitting to obtain accurate concentration values without the timeconsuming matrix-matched calibration strategy.

2. Experimental 2.1. Standards, solvents and sorbents Standards of THM, CLO, IMI, ACE and thiacloprid (THC) were purchased from Sigma Aldrich (Milwaukee, WI, USA). Individual solutions of each compound were prepared in methanol and stored at -20 ºC. Diluted solutions and mixtures were made in acetonitrile. Deuterated IMI (IMI-d4) was also provided by Sigma Aldrich and used as internal surrogate (IS) throughout the sample preparation procedure. Chemical structures of target compounds and the IS are provided as supplementary information, Fig. S1. Calibration standards, containing a fixed value (50 ng mL-1) of the IS and increased concentrations of target compounds (1-500 ng mL-1), were prepared in acetonitrile: water (1:1). Methanol and acetonitrile, gradient LC quality; dichloromethane and ethyl acetate, for trace analysis, were purchased from Merck (Darmstadt, Germany). Ultrapure water was obtained from a Milli-Q system (Millipore, Billerica, MA, USA). Formic acid was also provided by Merck. The reversed-phase OASIS HLB (60 and 200 mg) cartridges and the mixed-mode OASIS MCX and OASIS MAX (150 mg sorbent each) ones were purchased from Waters (Milford, MA, USA). Florisil (900 mg) and PSA (500 mg) clean-up cartridges were also provided by Waters.

2.2. Samples and sample preparation conditions Samples of red and white wines, produced in different areas of Spain, were purchased in local markets. After reception, wine bottles were kept at room temperature (for a maximum of 1 month) before being used for analysis or method development. After opening, wine samples were extracted in the next 24 h.

6

Several sample preparation strategies, based on the use of SPE cartridges, were considered for the extraction and the concentration of neonicotinoids from wine. Under optimized conditions, 10 mL wine samples were diluted with the same volume of ultrapure water and concentrated using a 200 mg OASIS HLB cartridge. This sorbent was previously conditioned with acetonitrile and an ethanol: water (12:88) solution (5 mL each). Thereafter, the SPE cartridge was rinsed with 5 mL of the above ethanolic solution and the sorbent dried with a gentle stream of nitrogen. Before elution, the HLB cartridge was on-line connected to the Florisil clean-up one. In a first step, 2 mL of acetonitrile were passed through both cartridges. After discarding the reversed phase sorbent, additional acetonitrile was pushed through the normal phase one to complete an elution volume of 5 mL. This extract was evaporated to 1 mL, and stored at 4 ºC until analysis. In addition to the above protocol, a simplified methodology was also proposed for white wines extraction. In this case, 2 mL of wine, diluted with the same volume of ultrapure water, were concentrated using a 60 mg OASIS HLB cartridge. Conditioning and washing steps were performed under same conditions as those reported for 200 mg cartridges. After sorbent drying, analytes were recovered using just 1 mL of acetonitrile. Whatever the sample preparation strategy, before LC-ESI-MS/MS analysis, acetonitrile extracts were diluted (1:1) with ultrapure water.

2.3. LC-ESI-MS/MS conditions Compounds were determined by LC-ESI(+)-MS/MS in the multiple reaction monitoring (MRM) mode. The employed instrument consisted of two binary high-pressure mixing pumps (Varian 212-LC), an autosampler and a Varian 320 MS triple quadrupole furnished with an ESI source, operated in the positive mode (ESI+). Chromatographic separations were developed in a Zorbax Eclipse SDB C18 column (100 mm x 2 mm, 3.5 m) acquired from Agilent. The column was connected to a C18 guard cartridge (4 mm x 2 mm) provided by Phenomenex (Torrance, CA, USA). Column and pre-column were

7

maintained at 40 ºC and operated at a constant flow of 0.2 mL min-1. The mobile phases were water (A) and acetonitrile (B), both 0.01% in formic acid. The chromatographic gradient was programmed as follows: 0-5 min, 2% B; 17 min, 40% B; 20-25 min, 100% B; 26-35 min, 2% B. The injected volume for calibration standards and wine extracts was 5 L. Nitrogen was used as nebulization and drying gas in the ESI source at 55 and 18 PSI, respectively. The drying gas temperature was set at 200 ºC, and the ESI needle voltage fixed at 5000 V. Argon was employed as collision gas (pressure 2.0 mTorr) in the CID cell of the LC-MS/MS instrument. Efficiencies of the different extraction approaches evaluated in this study were evaluated as the ratios between responses (peak areas) measured for each compound in the extracts of wine samples spiked before and after SPE extraction, multiplied by 100. Matrix effects in the ESI+ source were calculated as the difference between responses for each compound in spiked and non-spiked extracts, for aliquots of the same wine sample, divided by the signal for the addition standard and multiplied by 100 [22]. Unless otherwise stated, sample preparation conditions were optimized with aliquots of red and white wines produced in Ribera de Duero and Rías Baixas geographic areas (Spain). The addition level employed in these optimization assays experiments was 100 ng mL-1 per compound. The overall recoveries of the method were calculated as the difference between concentrations in spiked and non-spiked fractions of each sample divided by the added value. Concentrations in wine extracts were established by comparison with calibration standards, which contained increasing levels of target analytes and the same amount of IMI-d4 as that added to wine samples (100 ng). In the same way, the levels of investigated compounds in non-spiked wines were established by comparison with authentic standards after IS normalization.

3. Results and discussion 3.1. LC-ESI(+)-QqQ parameters

8

Table 1 summarizes the most relevant features of the LC-ESI(+)-MS/MS method for the determination of the neonicotinoid compounds involved in this study, including the IS. The linearity in the response of the system was investigated with standard solutions, prepared in acetonitrile: water (1:1), at 8 different levels in the range between 1 and 500 ng mL-1. The IS (IMI-d4) was added to calibration standards at a constant concentration of 50 ng mL-1. Determination coefficients (R2), corresponding to the plots of responses (peak area or peak area/IS peak area) versus concentration, stayed above 0.991 for all compounds. The limits of quantification (LOQs) of the LC-QqQ system, without considering the sample preparation step, varied between 0.2 ng mL-1 for ACE and THC to 0.8 ng mL-1 for the rest of compounds. These values are calculated as the concentration of each species providing a peak area 10 times higher than the standard deviation (SD) of the baseline (measured for the quantification transition). The repeatability in the response of the LC-MS/MS system was evaluated with consecutive injections (n=4 replicates) of a mixture of standards at 50 ng mL -1. The relative SDs of their peak areas varied between 1.6 and 3.8%.

3.2. Sample preparation conditions The first of the considered approaches for the determination of neonicotinoids in wines was direct injection of filtered (0.45 m pore size) samples. This choice led to significant signal attenuation (MEs below 100%) for THM, CLO and IMI in red wines, Fig. 1A. A poor selectivity, due to the presence of extra peaks in the MRM chromatograms for most compounds in red and white wines (figure not shown), was also noticed. Thus, solid-phase extraction (SPE) was investigated as extraction and concentration technique in order to minimize the above reported MEs, and also to reduce the LOQs of the method, by increasing the ratio between sample and extract volumes. Efficiency and selectivity of SPE methods are mainly controlled by the characteristics of (1) the sorbent and (2) the elution solvent. As regards the first parameter, mixed-mode

9

materials have been found to provide cleaner extracts than reversed-phase ones when target compounds and matrix components establish different interactions with the sorbent

polymer

[23].

Non-protic

solvents

(e.g.

acetonitrile,

ethyl

acetate,

dichloromethane) avoid the extraction of water soluble compounds, existing in wine and concentrated in the SPE sorbent, when used in the elution of reversed-phase polymers, providing colourless, cleaner extracts than methanol from red wines [24]. In this research, the OASIS HLB (200 and 60 mg sorbent loading) reversed-phase cartridges and the mixed-mode OASIS MAX and MCX ones (both containing 150 mg of sorbent) were tested for the retention of selected compounds from 10 mL red wine samples, considered as the most complex wine matrix. Except the 200 mg HLB sorbent, the rest of cartridges presented breakthrough problems, particularly for THM, which is the most polar of the investigated species, and in some cases also for CLO. The 200 mg OASIS HLB cartridges were selected for further experiments and the effects of different elution solvents in the yield and in the selectivity of the elution process were compared. ACN was a stronger eluent than dichloromethane and ethyl acetate, with cumulative responses (peak areas) above 95% for the 1st fraction (1 mL volume) of solvent obtained from the HLB cartridge, data not shown. Methanol displayed a similar elution profile to that of ACN; however, ACN rendered less complex extracts than methanol. Despite the transparent appearance of ACN extracts, important MEs were still noticed during LC-ESI-QqQ determination, with significant signal suppression for red wine extracts versus pure standard solutions. Fig. 1 compiles these effects for 10 and 2 mL wine samples concentrated using the 200 and 60 mg HLB cartridges, respectively. In both cases, the volume of the final extract was 2 mL. In order to reduce the above ME, reversed-phase extracts were submitted to a further clean-up either using Florisil or PSA cartridges. The 1st material is supposed to retain species with a higher polarity than analytes through normal-phase interactions. The 2nd is considered a weak anionic exchanger, and normal-phase material, being reported for the clean-up of SPE extracts during analysis of neonicotinoids in raw extracts from

10

pollen samples obtained by QuEChERs [25]. As appreciated in Fig. 1, both sorbents effectively removed sources of signal attenuation, attaining ME around 100% for the two investigated wine matrices.

Table 2 summarizes the extraction efficiencies, independent of MEs, for SPE of different sample volumes, without and with on-line clean-up, either using Florisil or PSA sorbents. Final elution volumes were 1 and 2 mL for 60 and 200 mg OASIS HLB cartridges, respectively. When combining the larger HLB sorbent with a 2nd clean-up cartridge a total elution volume of 5 mL was used. In all the cases, acetonitrile extracts were evaporated to 1 mL and diluted with the same volume of ultrapure water before injection. All the investigated conditions rendered extraction efficiencies between 96 and 110 %, with the exception of the combination of OASIS HLB with PSA purification, with extraction recoveries 61 and 86%.

Taking into account MEs and extraction efficiency data provided in Fig. 1 and Table 2, SPE of 10 mL volume samples with on-line clean-up, using a Florisil cartridge, emerged at the most suitable methodology for the extraction of red and white samples. Moreover, for this 2nd matrix, SPE of 2 mL samples (using the 60 mg HLB cartridges) provides quantitative recoveries without changes in the efficiency of the ESI+ process for sample extracts versus pure standard solutions. Fig. 2 shows the overlayed LCESI(+)-MS/MS chromatograms for aliquots of the same white wine, fortified at 20 ng mL-1, using the two reported sample preparation methodologies. The volume of the final extract was adjusted to 2 mL (acetonitrile:water, 1:1) in both cases. As appreciated, concentration of larger wine volumes followed by on-line clean-up not only increased the intensities of the chromatographic peaks, but also reduced the presence of interferences in the MRM chromatograms, particularly in the case of THM, CLO and IMI transitions.

11

3.3. Performance of the method The accuracy of the recommended method (OASIS HLB concentration of 10 mL samples followed by Florisil clean-up) was investigated with aliquots of red and white wines spiked at three different concentrations levels: 5, 10 and 20 ng mL-1. Samples fortified at lower and higher concentrations were extracted in the same day (n=3 replicates). Those spiked at 10 ng mL-1 were processed in three consecutive days, in order to assess the intra-day accuracy and precision of the procedure. IMI-d4 was added to all samples at a concentration of 10 ng mL-1. In the case of red wines, the obtained recoveries varied between 77 and 110 % with SDs in the range from 2 to 14%, Table 3. For the white wine matrix, the overall recoveries and SDs were 82-119% and 2-8%, respectively. Globally, results obtained during accuracy assessment confirmed the suitability of calibration using authentic standard solutions (prepared in acetonitrile: water, 1:1) for quantification purposes. Procedural blanks were performed using synthetic wine samples, spiked with the IS and concentrated as reported above, without detecting any significant contamination problem. The LOQs of the method were calculated from the instrumental LOQs compiled in Table 1, considering the concentration factor provided by the sample preparation method (10 mL samples and 2 mL final extract volume) and the lack of signal attenuation problems. The obtained values varied between 0.1 and 0.2 ng mL-1, depending on the compound. The linear response range of the method extended up to 100 ng mL-1. Table 4 compares the LOQs of the developed method procedure with those previously attained in the literature for these compounds, indicating also MEs (when available). Despite using an old-fashioned, medium-range, LC-ESI-MS/MS instrument, the LOQs attained in this work are lower than those reported in previous studies. On the other hand, the volume of sample and the effort devoted to sample preparation are similar to those required by other procedures, such as QuEChERS or microextraction techniques, at the same time that changes in the efficiency of ESI(+) are minimized, Table 4. Furthermore, the excellent yield of the extraction process,

12

combined with non-significant MEs during ESI ionization, avoided the use of matrixmatched calibration standards for quantification as recommended in previous SPE protocols [20], and in those methods based on microextraction techniques [21]. Concentration of 2 mL white wine samples, without clean-up, attained overall recoveries between 91 and 110% with SDs below 6 %, Table S1. The LOQs of this straightforward approach, valid only in case of white wines, varied between 0.5 and 2 ng mL-1.

3.4. Wine samples analysis The optimized method was applied to a total of 18 wines produced in four different geographic areas of Spain (Galicia, Rioja, Castilla-León and Castilla-La Mancha). Four compounds remained undetected in all the processed samples. On the other hand, 10 white wines, from two different geographic areas, contained concentrations of IMI, between 0.40 and 14.2 ng mL-1, Table 5. The levels of IMI in samples codes 1 and 4 were determined with both sample preparation methods, with consistent results; however, the better selectivity attained by concentration of large sample volumes combined with Florisil clean-up is evident, Fig. 3. The concentrations of the insecticide IMI compiled in Table 5 stay in the same range of values as the residues of this compound reported in California wines [16]. In comparison with those of fungicides, they remain one order of magnitude below the residues of anti-mildium and anti-botrytis agents measured in wines from the same areas [26]. Data obtained in this study prove the use of IMI based formulations in vineyards and the possibility of this insecticide to reach final commercialized wines. Using the processing factor published by Pazzirota et al. [14] for IMI (0.38), grapes employed during elaboration of these wines contained IMI levels up to 37 ng g-1. This value remains far below the MRL established by the EU for this insecticide in vinification grapes (1000 ng g-1) [27].

13

4. Conclusions Reversed-phase SPE combined with on-line clean-up with a Florisil cartridge rendered quantitative recoveries for the extraction of 5 neonicotinoid compounds from red and white wine samples, with moderate solvent consumption and non-significant matrix effects during ESI(+) ionization. Thus, the proposed method provides a high sample throughput since quantification against pure standard solutions can be used to obtain accurate concentration results. Analysis of commercial wines revealed the existence of moderate to low residues of IMI in around 50% of the processed samples. Field studies are required to investigate transfer factors of this group of insecticides to final wine, depending on (1) soil or foliar fumigation, (2) application doses and (3) time from application to grapes harvest.

Acknowledgements This study has been financially supported by the Spanish Government, Xunta de Galicia and E.U. FEDER funds (projects CTQ2015-68660-P and GRC2013-020). T.R.C. acknowledges a FPI fellowship to the Spanish Government.

14

References [1] G. Tanner, C. Czerwenka, LC-MS/MS Analysis of Neonicotinoid Insecticides in Honey: Methodology and Residue Findings in Austrian Honeys, J. Agric. Food Chem. 59 (2011) 12271-12277. [2] A. R. Main, N. L. Michel, J. V. Headley, K. M. Peru, C.A. Morrissey, Ecological and Landscape Drivers of Neonicotinoid Insecticide Detections and Concentrations in Canada’s Prairie Wetlands, Environ. Sci. Technol. 49 (2015) 8367−8376. [3] A. Ccanccapa, A. Masiá, A. Navarro-Ortega, Y. Picó, D. Barceló, Pesticides in the Ebro River basin: occurrence and risk assessment, Environ. Pollut. 211 (2016) 414424. [4] M.L. Hladik, D.W. Kopin, First national-scale reconnaissance of neonicotinoid insecticides in streams across the USA, Environ. Chem. 13 (2016) 12-20. [5] N. Campillo, P. Viñas, G. Férez-Melgarejo, M. Hernández-Córdoba, Liquid Chromatography with Diode Array Detection and Tandem Mass Spectrometry for the Determination of Neonicotinoid Insecticides in Honey Samples Using Dispersive Liquid-Liquid Microextraction, J. Agric. Food Chem. 61 (2013) 4799−4805. [6] M. Gbylik-Sikorska, T. Sniegocki, A. Posyniak, Determination of neonicotinoid insecticides and their metabolites in honey bee and honey by liquid chromatography tandem mass spectrometry, J. Chromatogr. B 990 (2015) 132-140. [7] M. Henry, M. Béguin, F. Requier, O. Rollin, J.-F. Odoux, P. Aupinel, J.Aptel, S. Tchamitchian, A. Decourtye, A common pesticide decreases foraging success and survival in honey bees, Science 336 (2012) 348-350. [8] Regulation (EC) 396/2005 of 23 February 2005. Off. J. Eur. Union, L70/1, 16/3/2005. [9] S. Pitam, I. Mukherjee, A. Kumar, Evaluation of the environmental fate of acetamiprid in the laboratory, Environ. Monit. Assess. 185 (2013) 2807-2816.

15

[10] M.J. Hilton, T.D. Jarvis, D.C. Ricketts, The degradation rate of thiamethoxam in European field studies, Pest. Manag. Sci. 72 (2015) 388-397. [11] Commission Implementing Decision (EU) 2015/495 of 20 March 2015. Off. J. Eur. Union, L78/40, 24/3/2015. [12] S. van Timmeren, J. C. Wise, C. Van der Voort, R. Isaacs, Comparison of foliar and soil formulations of neonicotinoid insecticides for control of potato leafhopper, Empoasca fabae (Homoptera: Cicadellidae), in wine grapes, Pest Manag. Sci. 67 (2011) 560-567. [13] E. Allsopp, Transmission of grapevine leafroll-associated virus 3 by vine mealybug, Planococcus ficus (Signoret) to grapevines treated with imidacloprid, S. Afr. J. Enol. Vitic. 36 (2015) 252-255. [14] T. Pazzirota, L. Martin, M. Mezcua, C. Ferrer, A.R. Fernández-Alba, Processing factor for a selected group of pesticides in a wine-making process: distribution of pesticides during grape processing, Food Addit. Contam. 30 (2013) 1752-1760. [15] C. Hao, D. Morse, X. Zhao, L. Sui, Liquid chromatography/tandem mass spectrometry analysis of neonicotinoids in environmental water, Rapid Commun. Mass Spectrom. 29 (2015) 2225-2232. [16] K. Zhang, J. W. Wong, D. G. Hayward, P.Sheladia, A. J. Krynitsky, F. J. Schenck, M. G. Webster, J. A. Ammann, S. E. Ebeler, Multiresidue pesticide analysis of wines by dispersive solid-phase extraction and ultrahigh-performance liquid chromatographytandem mass spectrometry, J. Agric. Food Chem. 57 (2009) 4019-4029. [17] R. Romero-González, A. Garrido Frenich, J.L. Martínez Vidal, O.D. Prestes, S.L. Grio, Simultaneous determination of pesticides, biopesticides and mycotoxins in organic products applying a quick, easy, cheap, effective, rugged and safe extraction procedure

and

ultra-high

performance

liquid

chromatography-tandem

mass

spectrometry, J. Chromatogr. A 1218 (2011) 1477-1485. [18] F. Cus, H. Basa Cesnik, S. Velikonja Bolta, A. Gregorcic, Pesticide residues and microbiological quality of bottled wines, Food Control 21 (2010) 150-154.

16

[19] P. Pérez-Ortega, B. Gilbert-López, J. F. García-Reyes, N. Ramos-Martos, A. Molina-Díaz, Generic sample treatment method for simultaneous determination of multiclass pesticides and mycotoxins in wines by liquid chromatography-mass spectrometry, J. Chromatogr. A 1249 (2012) 32-40. [20] A. Economou, H. Botitsi, S. Antoniou, D. Tsipi, Determination of multi-class pesticides in wines by solid-phase extraction and liquid chromatography-tandem mass spectrometry, J. Chromatogr. A 1216 (2009) 5856-5867. [21] M. Moeder, C. Bauer, P. Popp, M. van Pinxteren, T. Reemtsma, Determination of pesticide residues in wine by membrane-assisted solvent extraction and highperformance liquid chromatography-tandem mass spectrometry, Anal. Bioanal. Chem. 403 (2012) 1731-1741. [22] Matuszewski, B.K., Constanzer, M.L., Chavez-Eng, C.M., Strategies for the assessment of matrix effect in quantitative bioanalytical methods based on HPLCMS/MS, Anal. Chem. 75 (2003) 3019-3030. [23] I. Carpinteiro, M. Ramil, I. Rodríguez, R. Cela, Determination of fungicides in wine by mixed-mode solid phase extraction and liquid chromatography coupled to tandem mass spectrometry, J.Chromatogr. A 1217 (2010) 7484-7492. [24] T. Rodríguez-Cabo, I. Rodríguez, M. Ramil, R. Cela, Liquid chromatography quadrupole time-of-flight mass spectrometry selective determination of ochratoxin A in wine, Food Chem. 199 (2016) 401-408. [25] O. López-Fernández, R. Rial-Otero, J. Simal-Gándara, High-throughput HPLC– MS/MS determination of the persistence of neonicotinoid insecticide residues of regulatory interest in dietary bee pollen, Anal. Bioanal. Chem. 407 (2015) 7101-7110. [26] T. Rodríguez-Cabo, I. Rodríguez, M. Ramil, A. Silva, R. Cela, Multiclass semivolatile compounds determination in wine by gas chromatography accurate time-offlight mass spectrometry, J. Chromatogr. A 1442 (2016) 107-117. [27] http://ec.europa.eu/food/plant/pesticides/eu-pesticides-database

17

Captions to figures Fig. 1. Matrix effects (MEs) observed for LC-ESI(+)-MS/MS determination of neonicotinoids using different sample preparation conditions, n=3 replicates. A, red wine. B, white wine.

Fig 2. LC-QqQ chromatograms for aliquots of a spiked (20 ng mL-1) white wine sample. Solid line, concentration on 60 mg HLB cartridges without additional clean-up (2 mL of wine, final extract volume 2 mL). Dotted line, concentration on 200 mg HLB cartridges followed by Florisil clean-up (10 mL of wine, final extract volume 2 mL).

Fig. 3. LC-ESI(+)-MS/MS chromatograms for IMI in a non-spiked wine sample (code 1, Table 5) after SPE of a 2 mL aliquot without clean-up (A) and after SPE of a 10 mL aliquot combined with Florisil clean-up (B). Solid line, quantification transition. Dotted line, confirmation transition.

18

Table 1. Summary of LC-ESI(+)-QqQ conditions for determination of neonicotinoid insecticides.

Linearity

Quantification Retention

Capillary

Compound

Qualification transition

a

Response

Transition time (min)

Voltage (V)

(collision energy, eV)

2

-1

(R , 1-500 ng mL )

-1

ratio

(collision energy, eV)

LOQ (ng mL )

Without IS

After IS

THM

13.98

44

292 > 211 (10)

292 > 181 (18)

0.39

0.997

0.998

0.8

CLO

15.04

48

250 > 169 (9)

250 > 132 (12)

0.73

0.999

0.999

0.8

IMI

15.51

56

256 > 209 (14)

256 > 175(15)

0.80

0.995

0.999

0.6

ACE

16.11

56

223 > 126 (18)

223 > 90 (32)

0.23

0.995

0.995

0.2

THC

17.48

48

253 > 126 (22)

253 > 99 (41)

0.14

0.991

0.991

0.2

IMI-d4

15.48

70

260 > 213 (13)

260 > 179 (16)

-

-

-

a

Qualification/quantification transition

19

Table 2. Recoveries (%) of the extraction step with SDs, depending on the type of wine and sample preparation conditions. Data for samples spiked at 100 ng mL-1, n=3 replicates.

Red wine b

Compound

SPE +

a

SPE

b

White wine b

b

SPE +

SPE +

a

SPE

SPE

Florisil

PSA

b

b

SPE of

SPE + PSA

Florisil

THM

96  8

95  7

103  3

61  1

101  5

99  13

104  2

81  9

CLO

99  10

98  9

104  5

73  2

102  8

103  8

108  1

86  1

IMI

100  13

93  9

104  3

64  1

103  6

101  9

103  4

79  1

ACE

99  15

97  10

108  3

65  1

102  8

99  8

96  3

84  3

THC

97  2

95  6

105  5

65  1

99  7

95  9

104  4

82  7

a

Sample volume 2 mL, 60 mg HLB cartridges. Sample volume 10 mL, 200 mg HLB cartridges.

b

20

Table 3. Recoveries (%) of the overall method (SPE of 10 mL samples with Florisil clean-up) and SDs for spiked aliquots of red (Tempranillo variety) and white (Albariño variety) wines.

Red wine

White wine

Compound a

5 ng mL-1

b

10 ng mL-1

a

20 ng mL-1

a

5 ng mL-1

b

10 ng mL-1

a

20 ng mL-1

THM

92  6

103  3

92  6

98  5

119  7

97  6

CLO

97  7

97  7

84  6

96  5

95  5

82  5

IMI

102  4

110  7

100  4

109  8

107  5

90  6

ACE

85  7

86  4

77  5

99  2

89  8

83  6

THC

94  2

89  6

88  7

102  3

96  5

92  5

a

Intra-day repeatability, n=3 replicates.

b

Inter-day repeatability, n=9 replicates, 3 days.

21

Table 4. Summary of matrix effects (MEs), limits of quantification (LOQs) and additional features of different sample preparation methods for LC-ESI-MS/MS determination of neonicotinoid compounds in wine.

Sample

Extraction

Clean-up

a

MEs (%)

LOQs Compound

Ref.

-1

volume (mL)

technique

sorbents

Red

White

(ng mL )

20

QuEChERS

PSA,

77-159

84-123

2.0-3.3

IMI, ACE

[16]

carbon 10

QuEChERS

none

50-70

n.a.

4.5-5.4

THM, IMI, ACE, THC

[17]

10

SPE

none

n.a.

n.a.

8-20

IMI, ACE, THC

[20]

b

SPE

none

18-45

n.a.

1.6-7.8

THM, CLO, IMI, ACE

[19]

MASE

none

n.a.

n.a.

0.15-1.3

IMI, THC

[21]

SPE

Florisil

83-93

104-110

0.1-0.2

THM, CLO, IMI, ACE, THC

This work

4 10 10

c

a

MEs defined as slope ratios for matrix-matched and authentic standards for ref. 16,17 and 19. LC-ESI-TOF-MS determination c MASE, membrane assisted solvent extraction n.a., not available b

22

Table 5. Concentrations of IMI in non-spiked, commercial white wines, n=3 replicates. Sample code

Production year

Grape variety

Average conc. (ng mL-1)  SD

1

2014

Albariño

14.2  0.4

a

1

2014

Albariño

13.6  0.8

2

2012

Albariño

3.6  0.1

3

2014

Albariño

3.7  0.3

4

2014

Albariño

4.4  0.4

a

4

2014

Albariño

5.2  0.7

5

2014

Albariño

1.8  0.1

6

2014

Albariño

6.4  0.3

7

2014

Albariño

2.3  0.1

8

2014

Albariño

2.0  0.1

9

2015

Viura, Verdejo,

0.40  0.06

Tempranillo blanco 10

2015

Viura, Malvasía

0.52  0.01

a

Values obtained for SPE of 2 mL samples without clean-up.

23

Figure

Direct injection

SPE of 2 mL samples

SPE of 10 mL samples+Florisil

SPE of 10 mL samples+PSA

A

SPE of 10 mL samples

120

MEs (%)

100 80 60 40

20 0 THM

CLO

IMI

ACE

THC

Compound

Direct injection

SPE of 2 mL samples

SPE of 10 mL samples+Florisil

SPE of 10 mL samples+PSA

B

SPE of 10 mL samples

140 120 MEs (%)

100 80 60 40 20 0 THM

CLO

IMI

ACE

THC

Compound

Fig. 1

Figure

MCounts

MCounts

MCounts

7

8

6 4

7

THM

IMI

CLO 5

6

3 5

4

4 3

2

3

2 2 1

1 1

0

0

5

10

15

20

25

30

0

5

minutes

10

15

20

25

30

5

minutes

MCounts

10

15

20

25

30

MCounts

25

20

ACE

THC

20

15

15

10 10

5 5

0

0

5

10

15

20

25

30

minutes

5

10

15

20

25

30

minutes

Fig. 2

minutes

Figure

A

B

kCounts MCounts

900

IMI

IMI

800

4

700

600

3

500

2

400

300 1

200

100 0 0

7.5

10.0

12.5

15.0

17.5

20.0

22.5

minutes

7.5

10.0

12.5

15.0

17.5

20.0

22.5 minutes

Fig. 3