2015 Volume 40 Issue 3 Pages 165-172
In Japan, the Positive List System for the regulation of agricultural chemical residues in foods has been in force since May 29, 2006. Moreover, food poisoning caused by methamidophos-laced frozen gyōza dumplings came to light on Jan. 30, 2008. The news has set off alarms about the safety of not only perishables but also unexpected pesticides in processed foods. Therefore, a reliable systematic method was developed for determining pesticide residues in various foods. Firstly, ca. 200 target pesticides were selected by statistically analyzing the monitoring data in Aichi Prefecture. Secondly, a systematic method using plural separation and detection systems combining GC-MS/MS and LC-MS/MS as first priority was constructed. As the sample preparation method corresponding to the systematic method, a simple and easy acetonitrile extraction method, an auto-cleanup system combining GPC and mini-column SPE were developed. Thirdly, a new official multi-residue method that is able to determine a wide range of pesticides in various foods including fatty processed foods was developed. And finally, a comprehensive chromatographic detection system by dual-column GC-MS(/MS) and an interactive database without the use of pesticide standards was commercialized.
In Japan, a variety of foods have been imported from around the world, with the development of logistics system. At the same time, violations of the maximum residue limits (MRLs) under the Food Sanitation Law (Act No. 233 of 1947) of pesticides in imports such as frozen vegetables were increasing since 2001. Moreover, the importation and the usage of unregistered pesticides for domestic crops such as difolatan (captafol) and plictran (cyhexatin), whose registration under the Agricultural Chemicals Control law (Act No. 82 of 1948) were expired, was uncovered since 2002. Thus, pesticide residues in agricultural products were one of the highest concerns in food safety for consumers. To prohibit the distribution of foods with more than the allowable residues as a result of improper usage of pesticides, the Ministry of Health, Labour and Welfare (MHLW) introduced the Positive List System under the Food Sanitation Law for residual compositional substances of agricultural chemicals, feed additives and veterinary drugs (hereafter, agricultural chemicals) in foods on May 29, 2006. MRLs were set for ca. 800 substances by the addition of provisional MRLs, and the uniform limit was set at 0.01 ppm for agricultural chemicals in foods without MRLs.1)
To support the Positive List System, multi-quantitative residue methods of pesticides have been necessary, in addition to single and group residue methods by gas chromatography (GC) with electron-capture detection (ECD), flame photometric detection (FPD) and nitrogen-phosphorus detection (NPD), and by high performance liquid chromatography (HPLC) with ultra-violet absorbance detection (UV) and fluorescence detection (FLD). Besides, the food poisoning caused by methamidophos-laced frozen gyōza dumpling imported from China came to light on Jan. 30, 2008.2) The news has set off alarms about the safety of not only perishables but also unexpected pesticides in processed foods made from agricultural, animal and fishery products, accordingly, comprehensive systems that allowed quick identification and semi-quantification of the most pesticide residues in various foods have been also necessary.
Therefore, after the target pesticides were selected by statistically analyzing the monitoring data of pesticide residues in Aichi Prefecture, to certainly detect and accurately quantify these multiple pesticide residues in agricultural products and the processed foods consisting of complex components, a systematic method has been developed by combining an efficient method for extracting multiple pesticide residues in various samples, an effective cleanup method for removing matrixes without loss of many pesticides, and a reliable plural separation and detection method using tandem quadrupole mass spectrometry (MS/MS) as first priority.3–28)
The study described in this paper focused on the development of a new multi-quantitative residue method for determining pesticides including methamidophos at the request of the MHLW, and a comprehensive chromatographic detection system by GC-MS(/MS) without the use of pesticide standards, with an overview of a systematic method for determining multiple pesticide residues in agricultural products and the processed foods.
A systematic method that enables quantitative sequential analysis of multiple pesticide residues in a large number of various agricultural products has been constructed. Firstly, approximately 200 target pesticides were selected for cost-effective regulatory monitoring every fiscal year.6,9,15) These pesticides were commonly applied in crop protection in the country or neighboring countries, and/or were frequently found in agricultural products monitored during the past 5 fiscal years (e.g. April 2001–March 2006) in Aichi Prefecture (98% or more of the total findings), where has kept not only the number one industrial output area in Japan and also a prosperous agricultural producing area, with a population of over 7 million, and is located almost in the center of Japan. Therefore, if it is possible to analyze simultaneously throughout nearly 200 target pesticides, it was considered to be covered most of the pesticide residues.
1.2. Development of the systematic multi-residue methodAfter selecting the targer pesticides, we proceeded with development of a systematic method that is able to avoid false-positives, false-negatives and inaccurate quantifications. As shown in Fig. 1, electron ionization (EI)-multiple reaction monitoring (MRM) mode dual-column GC-MS/MS equipped with two types of capillary column (5% phenyl-methyl series Rxi-5Sil MS and trifluoropropyl-methyl series Rtx-200MS) which were connected to two injection ports and were simultaneously installed into the ion source (Fig. 2),20,21,23) and positive/negative switching electrospray ionization (ESI)-scheduled MRM mode LC-MS/MS using two types of core-shell column (heptacosyl (C27) series CAPCELL CORE AQ and pentafluorophenyl (PFP) series CAPCELL CORE PFP) with different separation characteristics, were adopted as first priority. An example of pesticide residues analysis by ESI-MRM mode LC-MS/MS are shown in Fig. 3.24,27,28) Grapefruit was produced by natural crossing pomelo and orange, so it is necessary to note that similar interfering peaks appear over the entire product ions produced from the protonated molecule of imazalil on the chromatograms of the test solution of pomelo by LC conditions. In addition to the MS/MS system, EI-SIM/Scan mode GC-MS,9) negative chemical ionization (NCI)-SIM/Scan mode GC-MS,17) dual-column GC-ECD/ECD,3,11) dual-column GC-FPD/NPD,4–6) ESI-SIM/Scan mode LC-MS with diode-array detection (DAD),13,14) and HPLC with post-column FLD for residual N-methyl carbamate pesticides,24,29) were also adopted. Thus, the systematic method using plural separation and detection systems has been constructed.
In efficient application of the systematic method, to reduce the frequency of time consuming and troublesome maintenance such as exchange of the GC capillary column, to overcome suppression or promotion of ionization of the target pesticides and to reduction the appearance of interfering peaks on the chromatogram as much as possible, it was considered that sample purification and dilution were essential. Accordingly, base on the official multi-residue method, we took part in the drafting,29,30) we proceeded with development of a more efficient extraction and more effective cleanup method.11)
To extract multiple pesticides having a wide range of solubility and stability, it was necessary to use hydrophilic and non-protonic organic solvents. Lipids extraction rate by acetonitrile is low as compared with that by acetone etc., although the solubility of low polarity pesticides is a little low, we confirmed that it was possible to ensure sufficient extraction rates in many pesticides by homogenizing with 3 fold amount or more of acetonitrile (for example, 60 mL) to the sample amount (20 g). Moreover, the sample extract was separated into acetonitrile layer and aqueous later by salting-out. By utilizing the properties of acetonitrile, a simple and easy acetonitrile extraction method, which uses suction filter system with graduated cylinder or cylindrical separatory funnel, employs the salting-out under neutral or weak acidic conditions, and is able to exactly collect aliquot of acetonitrile layer, was developed (Fig. 4).3,13)
Then, to prevent breakthrough of solid-phase extraction (SPE) column, and to fully exhibit the performance of gel permeation chromatography (GPC) column, by removing many high polar contaminants such as salt, sugar and amino acid with the water, after the ultrasonic dehydration and re-dissolution method using ethyl acetate and anhydrous sodium sulfate was adopted as a pre-cleanup,3–8) the auto-cleanup system combining GPC and mini-column SPE was developed. The large-molecular-size contaminants such as lipids were removed by GPC. However, the distinction between the fraction containing chlorophyll, carotenoid and higher fatty acid and the fraction containing pesticide residues was not clear-cut. By the early pesticides fraction from GPC eluting with these contaminants was selectively passed through the SPE mini-column such as a graphitized carbon-PSA (ethylenediamine-N-propyl-silyl silica gel) 2-layered mini-column, it was possible to effectively remove these contaminants without loss of most pesticides (Fig. 5).9,13,17) Also, after constructing the chromatographic detection system deploying a high-sensitive and rugged equipment as far as possible, it was adopted a method for determining the test solution diluted within the quantitative limits.7,24,28)
The MHLW noticed multi-residue method by GC-MS and LC-MS(/MS) for agricultural products.29,30) Summary of the method is as follows, extract agricultural chemicals from the sample using acetonitrile. After salting-out under neutral conditions, dehydrate the extract with anhydrous sodium sulfate. For fruits and vegetables, clean up the extract with a graphitized carbon-NH2 (aminopropyl-silyl silica gel) 2-layered mini-column SPE. For grains, beans, nuts and seeds, clean up with a C18 (octadecyl-silyl silica gel) mini-column SPE, followed with a graphitized carbon-NH2 2-layered mini-column SPE. Perform measurement and confirmation by GC/MS or LC-MS(/MS). However, pesticides show various physicochemical behaviors, a portion of some pesticides must be lost together with matrixes in the analysis. For example, methamidophos is a water-soluble pesticide (solubility in water >200 g/L and KOW log P=−0.8 (20°C)).31) These high polarity pesticides tend to go into the aqueous layer in the salting-out step, and adsorbed on the graphitized carbon-NH2 sorbents in the cleanup step, show an extreme peak intensity due to the matrix effects in GC32) since these peaks tend to be tailing. Also, methamidophos are relatively low sensitivity in LC-MS(/MS), have a low affinity with commonly used C18 LC columns, and the suppression or promotion in ionization was often observed under the matrixes coexist. Thus, it was pointed out some problem in terms of quantitativeness.19,24) It should be noted that, in the development of the official methods, acetonitrile which does not dissolve the fat is not performed efficient distribution between the adipose tissue of animal and fishery products, so there is a possibility that pesticides are not sufficiently extracted from the sample. In principle, it is indicated that fat-soluble n-hexane and acetone etc. should be used as extraction solvent.33)
On the other hand, summary of official multi-residue method for animal and fishery products29,30) is as follows, extract agricultural chemicals from the samples using acetone–n-hexane (1 : 2) except for liquid samples such as milk, eggs and honey. After dehydrate the extract with anhydrous sodium sulfate, clean up with GPC and PSA mini-column SPE. Perform measurements and confirmation by GC/MS or LC-MS(/MS). Acetone–n-hexane (1 : 2) is able to strongly dissolve the adipose tissue of animal and fishery products. However, the extraction method, that discards the aqueous layer separated by centrifugation, and collects the resulting organic layer (mostly n-hexane), is applied to the low polarity pesticides (KOW log P>ca. 1 such as dimethoate). For the reason, it is indicated that the Japanese official multi-residue method is not applicable to high polarity pesticides in solid samples such as muscle.33) Accordingly, we started to develop a new multi-residue method, that is able to determine precisely a wide range of pesticides in foods made from various foodstuff, at the request of the MHLW.
2.2. Development of a new multi-residue methodAs described above, it was considered that some of pesticide residues in fatty foods are not sufficiently extracted with acetonitrile. Therefore, a reliable multi-residue method was developed based on the official multi-residue method for animal and fishery products. In order to extract many pesticides including water-soluble pesticides (KOW log P>ca. −1 such as methamidophos) stably and efficiently, as shown in Fig. 6, the sample was homogenized with acetic acid aqueous solution, and extracted with acetone–n-hexane (2 : 3) while melting adipose tissue under the mildly acidic conditions, and then centrifuged. The supernatant including aqueous layer were filtered together. After concentration of the combined filtrate, the aqueous residues saturated with sodium chloride was loaded onto a macroporous diatomaceous earth column, and eluted with ethyl acetate. Co-extractives were removed by a GPC automatically (Fig. 7), and then by a SAX/PSA 2-layered mini-column SPE. The cleaned sample extract was subjected to GC-MS(/MS) and LC-MS/MS.20,23)
Most of water-soluble contaminants such as sugar (KOW log P=−5–−2) and amino acids (KOW log P=−5–−1, depending on the pH) could be efficiently removed with water by the on-column liquid–liquid partition using a macroporous diatomaceous earth column. Most of lipids such as phospholipid and glyceride could be efficiently removed by the GPC using newly 12 mm i.d. columns packed with styrene divinylbenzene copolymer of 16 µm particle size (Fig. 7), moreover preparation time was shortened, and the amount of the solvent was also significantly reduced. The 2-layered mini-column packed with strong anion exchange series SAX (trimethylamino-propyl-silyl silica gel) on the PSA was used the method, because we confirmed that portion of fatty acid was eluted from PSA mini-column under the influence of contaminants such as organic acid. In developing the new multi-residue method, a less clogging macroporous diatomaceous earth column (GL Sciences InertSep K-solute), GPC/SPE cleanup system (GL Sciences G-Prep GPC8050 and 8100), SAX/PSA 2-layered mini-column (GL Sciences InertSep SAX/PSA), and GPC column (Shodex CLNpak EV2000 12F and CLNpak EV-G 12C) were commercialized.
The new multi-residue method had been validated in accordance with Japanese guidelines for residual agricultural chemicals in foods by the MHLW undertaking since FY 2012,34,35) and was noticed as the Japanese official multi-residue method for agricultural chemicals by LC-MS II (animal and fishery products) on Feb. 26, 2015 (Fig. 6).36)
In order to deal with urgent incidents such as the frozen gyōza dumpling cases, comprehensive system for quick identifying and semi-quantifying of most of the pesticides including unexpectedly has been also necessary. Therefore, we evaluated a simultaneous method for determining environmental pollutants including pesticides by EI-Scan mode GC-MS coupled with three kinds of database: relative retention time, mass spectra and calibration curve, and reported about the usefulness with the problems.16,37) That is, the conventional calibration curve database uses deuterated stable polycyclic aromatic hydrocarbons such as naphthalene-d8, phenanthrene-d10 and fluoranthene-d10 as internal standards, and assigns by the relative retention time. It was considered that the conventional database was useful for screening of relatively stable environmental pollutants, however was unsuitable for determining pesticides showing various physicochemical behaviors, for example, unstable pesticides must be broken down in the GC injection port, and show an extreme rate of recovery such as >200% due to the matrix effects (Fig. 8-A).32)
Therefore, to improve accuracy and reproducibility of the calibration curve database, after 19 or more stable isotopically labeled pesticides (surrogates) such as acephate-d6, carbaryl-d7, chlorpyrifos-d10, imazalil-d5, isoxathion-d10 and es-fenvalerate-d7 were selected as the appropriate internal standards for GC-MS analysis,9) the surrogate mixture was commercialized by a manufacturer for an inexpensive and stable supply. Then, we invented the technology “Multi-quantitative analysis using the chromatography” that makes it possible to exactly assign these surrogates to hundreds of pesticides on the basis of the physicochemical property data obtained experimentally (Fig. 8-B). Incidentally, some substances strictly regulated by the Chemical Substance Control Law (Act No. 117 of 1973) such as p,p′-DDT-d8 and α-HCH-d6 were not selected in the present surrogates.
Finally, a comprehensive chromatographic detection system by GC-MS without the use of pesticide standards was announced in Nov. 2013.38,39) The system uses an EI-SIM/Scan alternate switching mode GC-MS, dual-column confirmation named the Shimadzu Twin Line MS system (Fig. 2), and a semi-quantitative database named the Shimadzu Quick DB. The Quick DB includes retention indices, mass spectra and calibration curves created with pesticide surrogates for over 450 pesticides (Fig. 9). Approximately 160 multi-class priority pesticides were selected for SIM mode, and additional pesticides were selected for Scan mode, with consideration for detection frequency and sensitivity, then 19 surrogates were assigned for these pesticides based on the technology described above. The system was applied to fatty foods such as cheese, butter, and miso (a traditional Japanese seasoning produced by fermenting soybeans) to demonstrate its use in routine analysis. Not a few interfering peaks appeared on the chromatograms of some foods, and in some instances the interfering peaks could not be resolved by using alternative target ion chromatograms (SIM ions). To overcome this problem, GC-MS was equipped with two different columns. As a result, pesticides that could not be detected on the first chromatogram because of interference and/or the retention times were shifted to the rear under the influence of matrixes could be detected and quantified by the second chromatogram. Scan mode was useful to confirm compound identity, while conscious of the existence of the matrixes. The capability of the new system allows quick identification and semi-quantification of many pesticides in various foods at the ppb level (e.g. 0.01 mg/kg), as required under the Japanese Positive List System.
Recently, GC-MS/MS has been generally accepted in the pesticide residue analysis, because it provides low detection limits as a consequence of high selectivity by the use of the MRM mode. The Shimadzu Quick DB for an EI-MRM mode GC-MS/MS was also commercialized at the same time.
In Japan, GC-MS(/MS) and LC-MS/MS has become first priority of the separation and detection equipment with the introduction of the Positive List System and improvements of the performance of recent MS(/MS). Especially, MS/MS is able to obtained selective MRM chromatograms and product ion scan spectra, accordingly, simplified sample preparation method is likely to be used. However, Some of the matrixes have a great influence on MS(/MS) during ionization. The performance of recent GC and HPLC with selective detection has also improved. Therefore, when precise quantification is required, if necessary, the sample must be confirmed by GC and HPLC after adequate cleanup. Besides, the food poisoning caused by malathion-laced frozen foods manufactured in Japan came to light on Dec. 2013. The news has set off alarms about the safety of not only imports but also domestic foods. The MS(/MS) system incorporating the present comprehensive chromatographic detection method is expected to use for not only urgent incidents but also regulatory monitoring of pesticide residues in agricultural products and inspection before shipping various foods
This work was supported by “Analytical method development project for the introduction of the Positive List System regarding residual agricultural chemicals” by the Department of Food Safety/Pharmaceutical and Food Safety Bureau/Ministry of Health, Labor and Welfare, and “Knowledge Hub AICHI, priority research project” by Aichi Science & Technology Foundation.
