Pyrethroids are a class of synthetic insecticides similar to the natural chemical pyrethrins found in chrysanthemum flowers (Chrysanthemum cinerariaefolium and C. coccineum). Since pyrethroids and pyrethrins are effective broad-spectrum insecticides with low acute toxicity to birds and mammals, they have been gradually replacing organophosphate insecticides for pest control in urban environments. In 2007, 258 tons of pyrethroids were used for non-agricultural pest control in California alone. Meanwhile, their usage for agriculture has also steadily increased, with 160 tons used during the same period in California.(1) As a result, pyrethroids are increasingly found in California’s water sources from either residential or agricultural run-off.(2, 3) Pyrethroids enter human and animals via the skin, by inhalation and from the gastrointestinal tract. Although human health effects from pyrethroid pesticides at low dose environmental exposures are unknown, studies have concluded that the some pyrethroids are neurological toxins, causing prolonged nervous system depolarization and hyperexcitation.(4) It is thought that the adverse effects from large dose exposures in humans are associated with this nervous system toxicity.(5, 6) Concomitant exposure to organophosphorus insecticides and pyrethroids may increase the latter’s toxicity by slowing the metabolic clearance of the pyrethroids. A state-wide investigation in California discovered that cyfluthrin was primarily associated with respiratory irritation, dermal effects and paresthesias, which were reported in agricultural workers from 1996 to 2002. (7) Determination of pyrethroids by immunoassay has been well established in the literature with immunoassays reported for many of the individual type II pyrethroids (e.g. deltamethrin(8), cypermethrin(9), flucythrinate(10), cyhalothrin(11), fenpropathrin(12), and esfenvalerate(13)). Many of these assays have also been demonstrated to work in environmental matrices such as soil extracts(8), grain extracts(8), milk(14), fruit jucies(10), plant and fruit extracts(15), wine(15), and orange oil(16). In most cases, especially in more complex matrices, extensive cleanup procedures are required. Class determination of the pyrethroids is problematic, since immunoassay analysis is usually highly selective for its target analyte. In order to have an immunoassay that detects the entire range of pyrethroids, it must be based on an antibody that has a natural cross reactivity for the analytes of interest. Many immunoassays have been reported with cross reactivities that range across the type-II-pyrethroids (17-19) however, most often sensitivities to the different pyrethroids range across several orders of magnitude. Usually such method bias will result in assays that are more qualitative in nature. Also, since antibodies are selected for their broad selectivity instead of their sensitivity, the resulting assay will likely result in higher limits of quantitation.

Instrumental methods for the pyrethroid class of compounds have their own set of challenges. Mass spectrometry (MS) based multi-residue methods are most widely used for the determination of multiple individual pyrethroids that are separated by either gas chromatography (GC) or liquid chromatography (LC) in a variety of matrices. However, apparent analytical difficulties have been reported for deltamethrin, cyhalothrin, other pyrethroids,(20) and for pyrethrins.(21) Most of the chromatography mass spectrometry based methods offered limits of detection (LODs) for pyrethroids in the upper ppb (part-per-billion) or low ppm (part-per-million) level in complex matrices, such as vegetable oils,(22) soybean oils,(23) and blood,(24) even after multi-step sample cleanup and concentration. Experiments conducted in our laboratory also found that the sensitivity for some pyrethroids, especially cyhalothrin and beta-cyfluthrin, are much worse than other types of pesticides, by using either a GC-MS/MS or LC-MS/MS technique. For cypermethrin, the LOD obtained in our laboratory was 40 ppb in orange oil by using a very sensitive tandem mass spectrometer,(16) while an entry level tandem mass spectrometer did not have enough sensitivity to measure cypermethrin under 500 ppb. In other investigations, LODs ranging from 13 to 49 ppb in tomato extracts for nine pyrethroids were achieved by using LC post-column photoderivatization and chemiluminescence detection followed by vigorous sample extraction and a pre-concentration step.(25) Considering the IC₅₀ values of pyrethroids to aquatic organisms, such as fish, are less than 1.0 ppb,(26) these methods lack adequate sensitivity to assess the safety of the water environment.

Type II pyrethroids are those that contain an alpha cyano substituent in the alcohol moiety (Figure 1). In our approach to the analysis of type II pyrethroid insecticides in environmental samples, rather than analyzing for individual pesticides, type II pyrethroids were chemically converted to a common product, 3-PBA. To our knowledge this is the first study to use such an approach to determine a total level of type-II-pyrethroid contamination. Flumethrin and cyfluthrin, that are converted to 4-F-3-PBA, were also detected by instrumental methods and by immunoassay since 4-F-3-PBA cross reacts in the assay by 72% of 3-PBA.(27) We chose citrus oils as the study samples, because they represent one of the most difficult matrices due to their hydrophobic nature and matrix complexity. Pyrethroids have been reported to be difficult to separate from oily matrices(28) and the cleanup of the matrix or preconcentration of pyrethroids to improve sensitivity is challenging. One advantage to converting the parent compound to 3-PBA is that the 3-PBA is more hydrophilic and thus easier to separate from the matrix should future cleanup or pre-concentration be needed. A drawback is that the results may be confounded by 3-PBA existing in the sample. Indeed preliminary analysis of randomly selected citrus oil samples found levels of 3-PBA or 4-F-3-PBA from non-detectable to as high as 50 ppb. Analysis of the sample before and after hydrolysis should address this issue. The major goal of the method development reported here was to find an appropriate method to convert pyrethroids to 3-PBA in citrus oils.

18 MΩ Water was produced by a Barnstead Easypure Rodi water purification system (Thermo-Fisher, San Jose, CA). Acetonitrile (99.5+%), 3-phenoxybenzoic acid (3-PBA, 98%), ammonium acetate (99.99+%), dimethyl sulfoxide (DMSO), N,O-bis(trimethylsilyl) trifluoroacetamide (BSTFA), goat anti-rabbit-horseradish peroxidase, 3,3′,5,5′-tetramethylbenzidine (TMB), bovine serum albumin (BSA), and 80% puriss grade sodium chlorite were purchased from Sigma-Aldrich (St Louis, MO). ACS certified methanol, L-ascorbic acid (99+%), sodium acetate, ethyl acetate, hexane, 30% hydrogen peroxide, 0.5% platinum on alumina pellets, and sodium hydroxide were obtained from Thermo-Fisher and glacial acetic acid was from Mallinckrodt Baker (Phillipsburg, NJ). Mixed mode anion exchange C8 SPE cartridges (Strata-Screen A) were purchased from Phenomenex (Torrance, CA). Nitrogen gas used by the AB Sciex mass spectrometer was generated from a Parker LCMS5000 high-purity nitrogen and zero air gas generator (Parker Hannifin Corporation, Haverhill, MA).

Immunoassays were performed with 96-well Nunc microtiter plates (MaxiSorp surface, Roskilde, Denmark). Normal strength phosphate-buffered saline (PBS) (1 × PBS; 8 g/L of NaCl, 0.2 g/L of Na₂HPO₄, and 0.2 g/L of KCl, pH 7.5), PBST (PBS containing 0.05% Tween 20), 0.05 M citrate-acetate buffer (14.71 g/L Na₃C₆H₅O₇·2H₂O, pH to 5.5 using glacial acetic acid) were used for immunoassay. Substrate buffer was prepared by adding 400 l of 0.6% TMB in DMSO and 100 l of 1% H₂O₂ into 25 mL of citrate-acetate buffer.

Individual pyrethroid pesticide stock solutions (100 or 2000 ppm in acetonitrile) for deltamethrin, cypermethrin, acrinathrin, lambda-cyhalothrin, and β-cyfluthrin were obtained from either Chem Service (West Chester, PA) or Resteck (Bellefonte, PA).

The polyclonal anti-3-PBA antibody was developed as previously described.(29)

In this study, several procedures were tested to convert pyrethroids to the intermediate alcohol and aldehyde. The procedures tested were based on chemical hydrolysis(30) or photolysis(31, 32). The conversion scheme of pyrethroid hydrolysis and oxidation is illustrated in Figure 1. Photolysis resulted in uncontrollable conversion to other unknown products (data not shown), so chemical hydrolysis and oxidation was the preferred method to produce 3-PBA from the type II pyrethroids. Since citrus oils have high levels of volatile components the matrix was initially simplified by volatilization of 95% of the mass under a steady stream of nitrogen. Analysis confirmed that there was no significant loss of the pyrethroid during this process (supporting information). Hydrolysis was accomplished with sodium hydroxide, which yields an alpha-cyano alcohol. The alcohol rearranges rapidly in the presence of water to the corresponding aldehyde. The main challenge was to select the proper oxidant that could convert the aldehyde to the carboxylic acid in the presence of other functional groups that could compete for the oxidant, including competing aldehydes.(33) Since, the concentration of total aldehyde in the oils ranged from 1 to 5 wt%, the selected method should oxidize all aldehydes in the citrus oil selectively over the other functional groups. The oxidation of the other aldehydes has another benefit by creating natural surfactants to help solubilize the remaining citrus oil components. Of the oxidants that were evaluated; household bleach, sodium chlorite, and hydrogen peroxide all demonstrated the ability to perform the oxidation in the presence of citrus oil (Figures S2-S4).

Optimization of the chemical conversion was critical to the success of this oxidation method. We varied the concentrations of oxidants and sodium hydroxide in a test mixture of 10 μM deltamethrin and 10 μM C¹³-3-PBA. We discovered that with increasing concentrations of household bleach and sodium chlorite the efficiency of the conversion of pyrethroid to 3-PBA also increased. However, if the oxidant concentration was too high, then chlorination occurred resulting in a loss of the product and the C¹³-3-PBA internal standard. Based on these data hydrogen peroxide and 10% chlorite were the most effective oxidants to use for the process. Household bleach was not studied further because it is difficult to accurately assess the concentration of oxidant and thus would be hard to prevent chlorination.

Since the method would be most useful without clean-up steps, we developed methods that would allow for immunoassay analysis after the oxidation with limited sample workup. Options for neutralization of the large excess of oxidant prior to analysis that would not require extraction were explored. Many different antioxidants were considered and tested with the discovery that ascorbic acid was of sufficient reactivity to neutralize the chlorite, and a solid phase platinum catalyst could neutralize the hydrogen peroxide. The solid phase catalyst was preferred in this case, because it could be easily removed before analysis. After neutralization the mixture was diluted in PBS prior to analysis by immunoassay, or cleaned up using a mixed mode SPE prior to instrument analysis. Matrix effects were observed at dilutions lower than 400 fold, with less of a matrix effect observed for the samples that were oxidized by hydrogen peroxide (supporting information).

Samples of orange oil were then spiked with 10 μM of different type-II-pyrethroids (deltamethrin, cypermethrin, lambda-cyhalothrin, acrinathrin, fenpropathrin) to determine if the conversion rate was similar for the different pyrethroids. All were oxidized with 10% sodium chlorite and were spiked with C¹³-3-PBA as a recovery standard. Following the oxidation the samples were cleaned up with a mixed mode SPE and spiked with 2-PBA as an internal standard prior to GC-MS analysis. The selected pyrethroids ranged from 40-60% conversion to 3-PBA with cypermethrin being the most efficient and acrinathrin being the least when using chlorite oxidation, hydrogen peroxide ranged from 50-70% conversion and recovery the difference between the two oxidation methods was not significant (Figure 2). This method was also applied to 10-fold concentrated orange oil, lemon, and grapefruit oils (Figure 3), overall, the conversion ranged between 30-50% in other citrus oil matrices with a decrease in conversion to the 3-PBA in lemon, grapefruit, and 10x orange oil compared to the conversion efficiency in orange oil.

A comparison of the recovery of the various methods of analysis was determined at two concentrations by spiking orange oil samples with different combinations of pyrethroids. Four sets of samples, labeled A through D, were created with each set containing one sample with low and one sample with high pyrethroid concentration. Set A was spiked with deltamethrin, set B contained β-cyfluthrin for 4-F-3-PBA, set C had a combination of deltamethrin and acrinathrin, and set D was spiked with a combination of deltamethrin, acrinathrin and lambda-cyhalothrin. The individual pyrethroids were converted to 3-PBA or 4-F-PBA for analysis by using either sodium chlorite for instrumental analysis or sodium chlorite and hydrogen peroxide for the immunoassay Table 1. Each sample was analyzed in triplicate by the different quantitation methods. Overall, the GC-MS had the lowest sensitivity to the lower concentration of pyrethroid with sample A2 and B2 being lower than the limit of quantitation and sample C2 and D2 showing an artificially high recovery because they were very close to the limit of quantitation. Both ELISA and the LC-MS/MS methods were able to reliably detect the 3-PBA in the samples, although some of the lower concentration samples were very close to the limit of detection with the ELISA method. The relatively low recovery at high concentrations and high recovery at low concentrations for sets C and D may reflect the variation among individual pyrethroids and the effect of the concentration on the conversation rate. Of note is that the spiked orange oil was assayed by LC-MS/MS to determine that the spiking concentration was accurate yet three of the four low concentration samples were below the detectable limit when assayed for individual pyrethroids by this instrumental method; however, when those samples were oxidized and converted to 3-PBA the LC-MS/MS was able to detect them easily.

This method described above was also applied to a water matrix by addition of hydrogen peroxide to laboratory and environmental waters samples. Hydrogen peroxide was selected because it eliminated the loss of 3-PBA to chlorination unlike the sodium chlorite oxidized samples. Much lower dilution (4 fold dilution overall from the original sample) was needed for the immunoassay since there was little matrix effect observed for the oxidation in water. Nanopure water was spiked with 20nM of several individual pyrethroids, which was then hydrolyzed and oxidized prior to analysis by GC-MS and immunoassay (Figure 4). Overall conversion efficiencies of each of the individual pyrethroids were above 50% in all cases.

An environmental water sample was collected from Putah Creek (a local creek that flows through agricultural lands) and analyzed before and after oxidation with hydrogen peroxide. No contamination from type-II-pyrethroids or 3-PBA were detected by immunoassay method using either oxidized or non oxidized Putah Creek water at a limit of quantification, determined using the IC80 of the immunoassay standard curve, of 1 nM for 3-PBA in non-oxidized samples and using a lower threshold of 50% conversion we estimate a 2 nM limit of detection for type-II-pyrethroid determination in oxidized water samples. To see the effectiveness of this procedure in environmental water it was spiked with different pyrethroids at 20 and 100 nM and the recovery was determined (Table 2). The average conversion and recovery rate from all replicates was 67% of the pyrethroids in water. This was similar in efficiency to the conversion observed earlier with citrus oils where we saw an average conversion and recovery of 58%. However, limits of detection using this method with immunoassay are much lower than citrus oils due to the lower matrix effect, which requires less dilution.

Here we have described a new approach for type II pyrethroid analysis and demonstrated its feasibility for use in both environmental water and citrus oils. This approach allows for quantitation with LC-MS/MS, GC-MS, or immunoassay. This method provides a semi quantitative analysis of the total amount of type II pyrethroids and is relatively quick, requires no extraction steps (except for GC-MS analysis), and is fairly robust. The authors acknowledge that a limiting factor is the endogenous 3-PBA in the samples would give a higher value for pyrethroid contamination, but this can be corrected by analysis of the sample before and after performing the oxidation procedure. In many cases the 3-PBA would be from the breakdown of pyrethroids in the sample and would be indicative of pyrethroid use. Another limiting problem is the inherent variability of the pyrethroid conversion but the benefits in the time and sensitivity of analysis may outweigh this drawback. This method is likely most useful for the immunoassay determination of pyrethroid levels and is a good method for first pass screening of samples before more detailed quantitation as would be the case in quality control screening. Finally, the sensitivity achieved using the LC-MS/MS technique for 3-PBA in the citrus oil matrix is lower than the limit of detection when assaying individual pyrethroids in the citrus oil.