Chemosphere xxx (2013) xxx–xxx

Contents lists available at ScienceDirect

Chemosphere

journal homepage: www.elsevier .com/locate /chemosphere

Determination of toxicity of neonicotinoids on the genome levelusing chemogenomics in yeast

0045-6535/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.chemosphere.2013.10.063

⇑ Corresponding authors. Tel.: +386 1 300 1115 (P. Trebše), +386 1 477 3250(U. Petrovic).

E-mail addresses: polonca.trebse@zf.uni-lj.si (P. Trebše), uros.petrovic@ijs.si (U. Petrovic).

1 These authors contributed equally to this work.2 Current address: University of Ljubljana, Faculty of Health Sciences, Zdravstvena

pot 5, SI-1000 Ljubljana, Slovenia.

Please cite this article in press as: Mattiazzi Ušaj, M., et al. Determination of toxicity of neonicotinoids on the genome level using chenomics in yeast. Chemosphere (2013), http://dx.doi.org/10.1016/j.chemosphere.2013.10.063

Mojca Mattiazzi Ušaj a,1, Petra Kaferle a,1, Alenka Toplak a,b, Polonca Trebše b,⇑,2, Uroš Petrovic a,⇑a Department of Molecular and Biomedical Sciences, Jozef Stefan Institute, Jamova 39, SI-1000 Ljubljana, Sloveniab Laboratory for Environmental Research, University of Nova Gorica, Vipavska 13, SI-5000 Nova Gorica, Slovenia

h i g h l i g h t s

� Additives are responsible for side effects of formulations including neonicotinoids.� Chemogenomics provided a possible mechanism of toxicity for inhibition of spermatogenesis.� Neonicotinoids negatively affect cell wall organization and biogenesis in yeast.

a r t i c l e i n f o

Article history:Received 7 June 2013Received in revised form 14 October 2013Accepted 16 October 2013Available online xxxx

Keywords:NeonicotinoidsAdditivesSide effectsChemogenomicsYeast Saccharomyces cerevisiaeSpermatogenesis

a b s t r a c t

Neonicotinoid insecticides are an important contribution to plant protection products. At the same time,their environmental impact on non-target organisms is often problematic. It has been shown that thetoxicity of formulations of neonicotinoid insecticides can originate from non-neonicotinoid additives.In the present study we used chemogenomics to analyse side effects of purified neonicotinoids, additivesand formulations on the genome-wide scale. We show that the additives in formulations have more pro-nounced effects than the active components, and that these effects could explain previously observednegative effects of neonicotinoid insecticides on spermatogenesis in animals. We also demonstrate thatcell wall organization and biogenesis in yeast is negatively affected by neonicotinoid substances.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

In spite of known harmful effects on the environment and neg-ative experiences with pesticides in the past, we can at present notimagine how sufficient quantities of high quality food could beprovided without using plant protection products. Since pests rel-atively quickly develop resistance to existing insecticides, the dis-covery of new insecticidal compounds with unique modes of actionis essential for efficient crop protection in the future. In the devel-opment of new insecticides, research is aimed at synthesizing mol-ecules that exhibit both efficient and spectrally broad protectionagainst insects, while acting upon a specific target site. Discoveriesof crop-protecting compounds with high efficiency and broad

spectrum of pest control, coupled with action at a novel target site,are however relatively rare (Cordova et al., 2006). Neonicotinoidsare relatively new systemic insecticides which are chemically sim-ilar to nicotine – the toxin present in tobacco. All commerciallyavailable neonicotinoids (imidacloprid, acetamiprid, clothianidin,dinotefuran, nitenpyram, thiacloprid and thiamethoxam) resemblenicotine and epibatidine, both of which are potent antagonists ofnicotinic acetylcholine receptors (nAChRs). Neonicotinoids havepartially positive charge and can irreversibly bind to nAChRs. Thebinding affinity of neonicotinoids is higher for nAChRs from insectsthan those from mammals (Matsuda et al., 1998; Roberts, 1999).

The toxicity of pesticides relies on their affinity to key biologicalmolecules, but other molecular targets for pesticide toxicity cannotbe excluded. In addition to the active compound, commercial for-mulations usually contain solvents or compounds which improveproperties important for easier application and better penetration.These compounds are typically considered as inert and are usuallyexcluded from the assessment of possible adverse effects on non-target organisms. It should be additionally stressed that also legis-lation regulating the placement of pesticides on the market focuses

emog-

2 M. Mattiazzi Ušaj et al. / Chemosphere xxx (2013) xxx–xxx

only on active (parent) compounds and neglects complete formu-lations and transformation products, although reports exist in theliterature that for non-target organisms the composition of formu-lations can be more important than active substances (Andersonand Roberts, 1983; Neves et al., 2001).

Aquatic toxicology has made an immense contribution to ourunderstanding of how natural and man-made substances affectthe living environment. Consequently, it has also played a centralrole in the development of policies and strategies for environmen-tal protection, providing the scientific basis for many of the stan-dards and quality objectives now applied in waste managementand water pollution control. To perform toxicity experiments,usually a base set of test species from different taxonomic groups,which are most frequently used for toxicity identification ofchemicals and biocides, is selected, e.g. bacteria Vibrio fischeri, al-gae Desmodesmus subspicatus, crustacean water flea Daphnia mag-na and fish Danio rerio. Few studies however exist where theformulations of neonicotinoids were tested on aquatic organisms(Jemec et al., 2007; Tišler et al., 2009; Malev et al., 2012). In gen-eral, much more studies have dealt with the toxicity assessmentof pure neonicotinoid on different terrestrial (Drobne et al., 2008;Trebše et al., 2009) and aquatic organisms (Jemec et al., 2007).Available data indicate that imidacloprid can be highly toxic tosome aquatic crustaceans, but generally less toxic to fish (TDCEnvironmental, 2003; Jemec et al., 2007). Also, the LD50s formammals and birds are much higher than for invertebrates (Ana-tra-Cordone and Durkin, 2005). Recent controversies regardingneonicotinoids were largely linked to their effect on pollinators,especially bees (Stokstad, 2013). Acute and chronic toxicity of imi-dacloprid and its metabolites is high in bees, as the acute toxicity(LD50) has been found to be 60 ng/organism (Suchail et al.,2001), and the chronic toxicity test revealed 50% mortality atapproximately 8 d. For acetamiprid, LD50 for bees is approximately400-fold higher compared to imidacloprid (Iwasa et al., 2004). Forthiamethoxam, LD50 toxicity data for bees in contact with theinsecticide was found to be 24 ng/organism (Senn et al., 1998),and this active compound severely affected behaviour of bees caus-ing decreased foraging success and lower survival rate (Henryet al., 2012).

At the molecular level, some research on the action of neonicot-inoids has been focused on anti-inflammatory and analgesic effectin mice (Tomizawa et al., 2001). However, the majority of studiesaddressing the environmental safety of the neonicotinoids, mainlyimidacloprid, was focused exclusively on the function of acetylcho-line receptors (Buckingham et al., 1997; Anatra-Cordone andDurkin, 2005), and the data on side effects of neonicotinoid formu-lations are lacking. In contrast to such focused approach, toxicoge-nomics and chemogenomics offer methodologies to identify alsoother potential target molecules affected by these substances (re-viewed in Dos Santos et al., 2012). Yeast Saccharomyces cerevisiaehas been successfully used as a model organism for the study ofthe toxicity of pesticides even before the advent of genomics ap-proaches (Cabral et al., 2003). In the functional genomics era thismodel organism has become extremely popular because of thesimilarity in key biological processes with homologous processesin other eukaryotes, including humans, and because of a widespectrum of available genome-wide techniques and the most thor-oughly annotated genome among all eukaryotes (Andrusiak et al.,

Table 1The neonicotinoid pesticides (‘Formulations’) addressed in our study. Active compoundspresent study.

Formulation Active compound

Mospilan 20 SP (Nippon Soda Co., Japan) Acetamiprid (20%)Actara 25 WG (Syngenta Crop Protection AG, Switzerland) Thiamethoxam (25Confidor 200 SL (Bayer, Germany) Imidacloprid (17.1%

Please cite this article in press as: Mattiazzi Ušaj, M., et al. Determinatienomics in yeast. Chemosphere (2013), http://dx.doi.org/10.1016/j.che

2012; Mattiazzi et al., 2012; Roemer et al., 2012). An example ofusing chemogenomics in yeast for the identification of bothprimary and secondary targets of a pesticide is the study of Diaset al. (2010), wherein the effects of the agricultural fungicideMancozeb were assessed. In this study, cellular processes such asoxidative stress response, protein degradation and carbohydratemetabolism have been identified as part of the response to Manco-zeb in yeast, and these results contributed to the understanding ofthe toxicity of this compound in humans.

In the present study, the focus was on the most usedneonicotinoid insecticides imidacloprid, acetamiprid, and thia-methoxam. Apart from the active substances, the effects ofcomplete formulations were analysed: imidacloprid containingformulation Confidor, acetamiprid and its formulation Mospi-lan, and thiamethoxam and its formulation Actara. Dimethylsulfoxide (DMSO) and N-methyl-2-pyrrolidone (NMP) werealso tested as the known additives to Confidor and other pesti-cide formulations. Active substances were isolated from com-mercial formulation, and their stability in media during theexperiment was confirmed by classical analytical methods.The growth rates of single deletion mutants of yeast S. cerevisi-ae were measured in the presence of the tested compounds/for-mulations to demonstrate that (i) in most cases additives in theformulations exert more pronounced effects than active sub-stances in yeast cells, and (ii) neonicotinoid active substanceshave a common negative impact on the cell wall organizationand/or biogenesis in yeast.

2. Materials and methods

2.1. Yeast strains

S. cerevisiae gene deletion collection (EUROSCARF), constructedin the BY4741 background strain (MATa his3D1 leu2D0 met15D0ura3D0 xxxD::Kanr) and representing all non-essential yeast genes,was used in the study. The strains were arrayed on 14 individualplates of 384 colonies per plate, together comprising 5376 coloniesand 4293 mutated genes, where control (‘‘wild-type’’; i.e. back-ground strain without additional deletions) strains (MATahis3D::Kanr leu2D0 met15D0 ura3D0) were placed on the rim rowsand columns.

2.2. Media and reagents

Control and master plates contained 1% (w/v) yeast extract(Sigma, USA), 2% (w/v) peptone (Sigma, USA), 2% (w/v) D-(+)-glu-cose (Fluka, Germany), 2% (w/v) agar (Sigma, USA) and 200 mg L�1

antibiotic geneticin (Gibco-Invitrogen, USA). In the test plates, dif-ferent formulations or active substances were added to the med-ium of the upper composition, one per test.

Mospilan 20 SP (Nippon Soda Co., Japan), Confidor 200 SL(Bayer, Germany) and Actara 25 WG (Syngenta Crop ProtectionAG, Switzerland) pesticides were analysed in this study. Purifiedneonicotinoids and Mospilan, Confidor and Actara were added be-fore autoclaving since prior testing (see Section 2.4 below) demon-strated stability of their active substances under autoclavingconditions. Dimethyl sulfoxide (DMSO; molecular biology grade)

and additives are listed. Substances in bold have been experimentally tested in the

Additives

Surfactants (not specified; 5%) Carrier (not specified; 75%)%) Diatomaceous Earth (N/A) Crystalline Silica, Quartz (N/A) Starch (N/A)) Dimethyl sulfoxide (N/A) N-Methyl 2 pyrrolidone (N/A)

on of toxicity of neonicotinoids on the genome level using chemog-mosphere.2013.10.063

M. Mattiazzi Ušaj et al. / Chemosphere xxx (2013) xxx–xxx 3

(Sigma, USA) and N-methyl-2-pyrrolidone (NMP; biotech. grade)(Sigma, USA) were added after autoclaving. Additives and activesubstances for the respective pesticides are given in Table 1.

2.3. Crystallization and purification

Active substance acetamiprid was isolated from Mospilan 20 SP,thiamethoxam was isolated from Actara 25 WG and imidaclopridwas isolated from Confidor 70 WG (Bayer, Germany), followingthe procedure of Obana et al. (2003). The purity of the obtainedcrystals was confirmed by 1H NMR spectra of molecules and bymeasuring the melting points (Cernigoj et al., 2007).

2.4. Stability tests

To ensure reliable toxicity data, the stability of acetamiprid,imidacloprid and thiamethoxam in media before and after auto-claving (20 min, 121 �C, 15 psi) was checked. Imidacloprid, ace-tamiprid, and thiamethoxam in the concentrations of 140 mg L�1

for imidacloprid, 120 mg L�1 for acetamiprid and 170 mg L�1 forthiamethoxam were extracted from media according to Baskaranet al. (1997) with slight modifications. Each solution was aliquotedin six flasks (50 mL), three of them were extracted prior incubation(2 h after preparation of the reaction mixture) and three after incu-bation. The solid phase extraction (SPE) of imidacloprid, acetami-prid and thiamethoxam from media solutions was performed onStrata C18-E columns (200 mg). The columns were initially precon-ditioned with 5 mL of methanol followed by 5 mL of distilledwater. The neonicotinoid sample (50 mL aliquots diluted to150 mL of distilled water) was loaded on the column and the re-tained neonicotinoid was eluted with 2 mL of methanol. In the nextstep methanol was removed by rotary evaporation in vacuum at30 �C (Büchi–Waterbath B-480; Germany, Flawil, Switzerland)and dried leftover was rediluted in 1 mL of acetonitrile–water(20:80 v/v) solution (HPLC solvent mixture). Prepared sampleswere stored at 4 �C until subjected to liquid chromatograph (HPLCHP 1000 Series) equipped with diode array detection (DAD). AllHPLC-DAD analyses were performed in duplicates on Zorbax C8(4.6 � 250 mm, 5 lm particle size) column at 25 �C using an iso-cratic separation with mobile phase of acetonitrile–water (20:80v/v) at a flow rate of 1.25 mL min�1. The stability of imidacloprid,acetamiprid and thiamethoxam was followed from the peak areasat 270 nm for imidacloprid, 245 nm for acetamiprid and 254 nmfor thiamethoxam, which were identified on the basis of retentiontimes comparison with authentic standards.

2.5. Chemogenomic assay

The assay for measuring the growth fitness of arrayed colonieson agar was performed as described in Mattiazzi et al. (2010).Modifications are described below (Sections 2.5.1-2.5.3).

2.5.1. ScreeningCells from master plates were pinned to test and control plates.

In addition to media components as described in Section 2.2, thetest plates contained, 15 mM acetamiprid, 10 mM thiamethoxam,or 0.8% (v/v) DMSO and 0.9% (v/v) NMP. Concentrations ofMospilan, Actara and Confidor in screening experiments wereequivalent to 5 mM acetamiprid, 20 mM imidacloprid and 20 mMthiamethoxam. The concentrations were selected based on thepre-determined semi-inhibitory (i.e. 20% growth inhibition in thewild-type strain) concentration and solubility. Pinning procedurewas performed using a custom-made pinning robot (Adept plus,Slovenia). Strains were grown for 96 h on test plates with Mospi-lan, Confidor and Actara, for 72 h on test plates with acetamiprid,

Please cite this article in press as: Mattiazzi Ušaj, M., et al. Determinatioenomics in yeast. Chemosphere (2013), http://dx.doi.org/10.1016/j.che

for 48 h on test plates with DMSO + NMP and thiamethoxam, aswell as on control plates, all at 30 �C.

2.5.2. Data acquisition and image analysisPlates were scanned in transmission mode with Umax Power-

Look 112 scanner (Amersham Biosciences). The obtained 600 dpi8-bit grayscale images were transformed into 16-bit images usingPhotoshop (Adobe Systems) software, and further analysed usingGenePix Pro 5.1 (Axon Instruments) software, as described inMattiazzi et al. (2010).

2.5.3. Data analysisLog2 of test/control fitness ratio was computed for each strain,

and the mean and standard deviation (SD) of log2 of test/controlfitness ratios were computed on all ratio values. Strains with ratiosbelow (mean � 1SD) were considered sensitive to the tested chem-ical. All sensitive strains were divided into two groups. The strainsscoring lower than (mean � 2SD) were placed in the first group(more sensitive), while those with scores between (mean � 1SD)and (mean � 2SD) were placed in the second group (less sensitive).For gene set enrichment analysis, Gene Ontology (GO) Term Findertool (http://go.princeton.edu/cgi-bin/GOTermFinder) was used.

2.6. Plate dilution assay

Overnight cultures of wild-type cells were grown in liquid med-ium of the composition described in Section 2.2 with agar omitted.A 10-fold dilution series was prepared and plated on agar plateswith the medium of the same composition as described above,with or without the addition of 1 M NaCl, and Mospilan equivalentto 5 mM acetamiprid, Confidor equivalent to 6 mM imidacloprid orActara equivalent to 20 mM thiamethoxam.

3. Results

We isolated active neonicotinoid substances from their respec-tive formulations, and confirmed the purity of the obtained crystalsby 1H NMR spectra of molecules and by measuring the meltingpoints and comparing them with data for analytical standards.We monitored SPE extraction with methanol by HPLC-DAD andyielded the extraction recoveries of (90 ± 10)% for imidacloprid,(95 ± 10)% for acetamiprid and (90 ± 10)% for thiamethoxam. Ourstability tests showed that autoclaving performed as described isSection 2.4 had no effect on imidacloprid, acetamiprid or thiameth-oxam, pure or in the yeast growth media.

We used chemogenomics to assay the molecular effects of com-mercially available neonicotinoid-based pesticides (formulations),isolated neonicotinoids (active compounds) and the combinationof two known additives alone, DMSO and NMP, on a whole genomescale (Table 1). For each tested substance, we identified a chemog-enomic interaction profile which consists of relative sensitivities tothe addition of the substance for each viable yeast strain with a de-leted gene. Since yeast cells do not express nicotinic acetylcholinereceptors, the obtained profile is an indication of the secondary tar-gets of a given neonicotinoid and/or the additives in theformulation.

We found a relatively large number of mutants to be sensitive(i.e. values lower than (mean � 1SD); see Section 2.5.3) to eachtested substance (Fig. 1), from 17.0% of the tested strains in thecase of thiamethoxam to 6.9% of the strains in the case of Actara.The overlap between genes found in the analysis of a formulationand its respective neonicotinoid or additive showed that the mostsimilar pair was that of Confidor and the additives in this formula-tion, DMSO and NMP, while the least similar pair was that of Actaraand thiamethoxam. When we used a more conservative threshold

n of toxicity of neonicotinoids on the genome level using chemog-mosphere.2013.10.063

Fig. 1. Similarity of chemogenetic profiles of the neonicotinoid insecticide formulations and their respective active compound (neonicotionid) or additive. Venn diagramsshow the overlap of the number of mutant strains with significantly higher sensitivity to the tested substances. Results are shown for the genes with relative fitness valuesless than Mean – 1 SD (upper panel), and for those with relative fitness values less than Mean – 2 SD (lower panel).

4 M. Mattiazzi Ušaj et al. / Chemosphere xxx (2013) xxx–xxx

(i.e. values lower than (mean � 2SD)), the high similarity of theprofiles of Confidor and DMSO and NMP became even more obvi-ous (Fig. 1).

We then compared pair-wise all the experimentally obtainedprofiles using Pearson correlation as the similarity metric. Thehighest correlation was again observed between the profile ofConfidor and the mixture of DMSO and NMP, additives present inConfidor, indicating that additives in the formulations have morepronounced effects than the active substances in yeast cells.

To investigate further which of the two substances in the for-mulation-neonicotinoid or formulation-additive pairs contributesto the biological effects, we performed gene set enrichment analy-sis of genes whose mutations resulted in an increased sensitivity,using the biological process Gene Ontology (GO) for gene annota-tions (Table 2). This approach enables a more detailed, as well asbiologically relevant analysis of the chemogenomic interactionprofiles, since the identified enriched biological processes areindicative of the biological effects of the tested substance. Forexample, two of the most significantly enriched GO terms of theConfidor’s set of sensitive mutants were histone exchange, where7 genes, out of the 9 annotated to this GO term in the set ofscreened mutant strains, when mutated conferred sensitivity toConfidor (corrected p-value 3.8 � 10�4), and histone acetylation

Table 2Enriched terms for the biological process GO for the performed experiments. p-Valuesfor all the terms, with the exception (�) of histone acetylation in the dimethylsulfoxide + N-methyl 2 pyrrolidone (DMSO + NMP), which was added for comparisonto the Confidor results, denote statistical significance of enrichment.

GO biological process Corrected p-value

ConfidorHistone exchange 3.8 � 10�4

Histone acetylation 5.5 � 10�8

DMSO + NMPHistone exchange 6.8 � 10�4

Histone acetylation 1 (�)ActaraProtein retention in the Golgi apparatus 2.5 � 10�3

AcetamipridAminoglycan biosynthetic process 1.0 � 10�3

Chitin biosynthetic process 1.0 � 10�3

Cell wall biogenesis 4.7 � 10�3

Please cite this article in press as: Mattiazzi Ušaj, M., et al. Determinatienomics in yeast. Chemosphere (2013), http://dx.doi.org/10.1016/j.che

(14/23 genes annotated to this term; corrected p-value5.5 � 10�8), indicating that Confidor interferes with these two cel-lular processes in yeast. Mutations in all of the above mentioned 7genes involved in histone exchange, SWR1, SWC3, SWC5, VPS71,VPS72, ASF1 and ARP6, resulted in significantly increased sensitivityalso to DMSO and NMP (Table 3a). Strikingly, in the analysis of theinteraction profile of DMSO and NMP, the GO term histone ex-change was the most significantly enriched one as well (7/9 genesannotated to this GO term – the same 7 genes as with Confidor;corrected p-value 6.8 � 10�4). The proteins encoded by these 7genes – together with Swc4 protein, which is encoded by theessential SWC4 gene and was therefore not included in our study– build the SWR1 complex, which is specifically involved in the ex-change of the histone variant H2AZ (Htz1p). Importantly, alsomutation in the HTZ1 gene resulted in significantly increasedsensitivity to both Confidor and DMSO and NMP (Table 3a).

On the other hand, mutations in the 13 genes encoding the his-tone acetyltransferase (RTT109) or components of acetyltransferasecomplexes (EAF3, AHC1, SGF29, SGF73, SPT3, EAF7, EAF6, VPS75,SPT8, YNG1, ELP3, and AHC2; while ASF1, the 14th gene from theprofile annotated to histone acetylation is primarily involved inhistone exchange, see above) conferred sensitivity to Confidor,but did not significantly increase the sensitivity to DMSO andNMP (Table 3b). Notably, however, even with Confidor, more se-vere effects on the fitness values were determined for the mutantsof genes involved in histone exchange than those involved in his-tone acetylation (compare values in Table 3a and b).

In the case of Actara and its active component thiamethoxam,sensitivity to Actara was significantly increased in mutants ofgenes involved in the process of protein retention in the Golgiapparatus (6/8 genes annotated to this GO term; corrected p-value2.5 � 10�3), while no significantly enriched GO terms were identi-fied in the thiamethoxam experiment. The process of proteinretention in the Golgi apparatus is a part of the multidrug responseand mutations of genes annotated to it are sensitive to many sub-stances with different modes of action, indicating a non-specifictoxic response (Hillenmeyer et al., 2008). Since mutations in noneof these 6 genes from the Actara profile (VPS27, VPS35, VPS4, VPS1,VPS13, PEP8) caused increased sensitivity to the active substancethiamethoxam (Table 3c), it can be concluded that the toxic effectwas caused by the additives in the formulation.

on of toxicity of neonicotinoids on the genome level using chemog-mosphere.2013.10.063

Table 3Relative fitness values of strains with significantly increased sensitivity in theConfidor/DMSO and NMP experiments, (A) mutated in genes encoding the SWR1complex components and histone variant H2AZ, (B) mutated in genes encodinghistone acetylases and histone acyltransferases, (C) in the Actara/thiamethoxamexperiments of genes annotated to the ‘Protein retention in the Golgi apparatus’ geneontology term. Bold, italics – relative fitness values less than Mean-2 SD; Italics –relative fitness values between Mean-1 SD and Mean-2 SD.

Confidor DMSO + NMP

(A) SWR1 complex componentsVPS71 �4.64 �4.84ARP6 �5.20 �4.85SWC3 �5.25 �4.69SWC5 �5.62 �4.69VPS72 �5.29 �4.44SWR1 �5.48 �2.83ASF1 �2.19 �0.41HTZ1 �4.12 �4.27

(B) Histone acetylase and histone acyltransferase complex componentsRTT109 �1.44 �0.27SPT3 �1.60 �0.18SGF73 �2.27 �0.18SGF29 �2.10 �0.02EAF3 �2.17 0.10AHC1 �1.49 0.16EAF7 �0.86 �0.23EAF6 �0.76 0.30VPS75 �0.90 �0.14SPT8 �0.85 �0.29YNG1 �0.86 0.14ELP3 �1.69 �0.14AHC2 �1.02 �0.22

Actara Thiamethoxam

(C) Protein retention in the Golgi apparatusVPS27 �3.50 �0.26VPS35 �2.91 0.11VPS4 �3.41 0.30VPS1 �3.77 0.43VPS13 �0.94 �0.10PEP8 �0.89 0.14

Fig. 2. Neonicotinoid formulations affect the cell wall of yeast cells. Results of theplate dilution assay are shown of the wild-type strain yeast cells grown on mediawith or without the addition of Actara equivalent to 20 mM thiamethoxam,Confidor equivalent to 6 mM imidacloprid, Mospilan equivalent to 5 mM acetam-iprid and 1 M NaCl, as marked in the legend. Ten-fold serial dilutions were used andthe plates were incubated at 30 �C for 96 h.

M. Mattiazzi Ušaj et al. / Chemosphere xxx (2013) xxx–xxx 5

A different result was however obtained in the experiment ana-lysing Mospilan and its active component acetamiprid: both exper-iments, testing the effect of the neonicotinoid acetamiprid and ofthe formulation Mospilan, resulted in a profile for which the GOterms related to cell wall organization and biogenesis were foundamong the most significantly enriched and with the highest cover-age of the genes. This indicated that acetamiprid, and not the addi-tives in the formulation, causes the major side-effect of Mospilanin yeast cells, namely affecting the biogenesis and/or organizationof the cell wall. To further investigate the relation between cellwall organization and biogenesis and neonicotinoids, cells in thepresence of all three tested formulations were challenged withan osmotic stress caused by the addition of NaCl to the growthmedium. Osmotic stress is detrimental to cells with impaired cellwall and therefore inhibits growth when combined with a sub-stance that interferes with cell wall organization or biogenesis.The plate dilution assay (Fig. 2) showed growth inhibition in allcombinations of formulations and NaCl, and thus revealed thatall tested formulations of neonicotinoid insecticides inhibit the for-mation and/or function of the yeast cell wall. We thus revealed acommon side effect of neonicotinoid insecticides in yeast cells,and further studies are required to establish whether this effectextends also to other organisms.

4. Discussion

Yeast, a model eukaryotic organism with highly conserved corecellular processes, was used to investigate the side effects of neon-icotinoid insecticides. To achieve this, the sensitivity to purified

Please cite this article in press as: Mattiazzi Ušaj, M., et al. Determinatioenomics in yeast. Chemosphere (2013), http://dx.doi.org/10.1016/j.che

neonicotinoids, formulations and additives from formulationswas determined for all viable deletion mutants of yeast, generatingchemogenomic interaction profiles. Pair-wise similarity compari-sons of the chemogenomic interaction profiles of the tested sub-stances revealed that the most similar profiles are those ofConfidor, and the additives of this formulation, DMSO and NMP.On the other hand, the chemogenomic interaction profiles of thetwo tested pairs of formulation and corresponding purified activecompound (Mospilan – acetamiprid and Actara – thiamethoxam)showed much less overlap, indicating that additives may largelycontribute to the biological off-target effects of the tested formula-tions. In line with this conclusion, the comparison of chemogenom-ic interaction profiles of Confidor and DMSO + NMP, based on theenriched biological processes, revealed a common and most severeeffect on H2AZ histone exchange process, indicating that the addi-tives, and not the neonicotinod active substance, were responsiblefor this activity, since DMSO and NMP entirely explain this side ef-fect (see below). Conversely, the less severe effect on histone acet-ylation can be assigned to the effect of other components ofConfidor, either the active component imidacloprid, other addi-tives than DMSO and NMP, or the synergistic effects thereof.

DMSO perturbs cellular membranes, and ample data existshowing that mutation in the HTZ1 gene, encoding the yeastH2AZ histone, renders yeast cells sensitive to membrane perturba-tions, such as those caused by lipid metabolism modifications(Smith et al., 2006; Lockshon et al., 2007; Wan et al., 2009). There-fore, the observed biological effect of a DMSO-containing sub-stance was not surprising. What was somewhat surprising,however, was that this additive-mediated effect was the predomi-nant one. It should be noted that the observed effect of Confidor ongenes involved in histone exchange could explain the inhibitory ef-fect of this pesticide on spermatogenesis in animals (Luo et al.,1999; Najafi et al., 2010), since the replacement of somatic his-tones with sperm-specific histones is an important step in sper-matogenesis (Loppin et al., 2001). Our results suggest that thesedetrimental effects are actually caused predominantly by additivesin the formulations, but, unfortunately, due to too low solubility inyeast growth media pure imidacloprid could not be tested. Addi-tional experiments, preferably directly in animal models, will berequired to test this hypothesis.

The comparisons of chemogenomic interaction profiles thusindicated that additives contribute more importantly to the side ef-fects of formulations than the neonicotinoids, at least in the yeastmodel system. Additional studies will be needed to address the

n of toxicity of neonicotinoids on the genome level using chemog-mosphere.2013.10.063

6 M. Mattiazzi Ušaj et al. / Chemosphere xxx (2013) xxx–xxx

potential toxicity of additives, but such studies are not easily feasi-ble since the nature of the additives of the formulations is most of-ten regarded as confidential business information and thereforenot known.

Not all observed biological effects could be explained by addi-tives alone, however. One such example is a larger number ofsensitive mutants at the 1 SD threshold identified in the thiameth-oxam experiment, compared to Actara, the formulation containingthis active ingredient. However, given the fact that at the morestringent threshold of 2 SD a very similar number of mutantswas identified in the corresponding screening experiments, it isalso possible that the former difference occurred due to unknowntechnical reasons. In the case of the formulation Mospilan, its ac-tive ingredient acetamiprid was found as a likely causative factorfor the effect on the cell wall of yeast cells. The other tested sub-stances did not show significant enrichment of cell wall biogenesisrelated genes as the most prominent hits, presumably becauseother biological processes (e.g. histone exchange) are morestrongly affected. Nonetheless, in the follow-up experiment, weconfirmed that formulations with neonicotinoids in general inhibitthe formation and/or function of the yeast cell wall, and we foundevidence suggesting that this effect is most likely the consequenceof the neonicotinoid active substances. It remains to be discoveredwhat is the molecular mechanism behind this effect, and whetherpotentially also other chitin-based biological structures are af-fected by neonicotinoids. The main importance of this result isthe identification of a common biological effect of neonicotinoids,which represents an important lead in the identification of theirside effects in target organisms and the mechanisms behind them.

Acknowledgement

The work has been funded by Grant J1-9808 from the SlovenianResearch Agency.

References

Anatra-Cordone, M., Durkin, P., 2005. Imidacloprid. Human health assessment andecological risk assessment – Final report. Syracuse Environmental ResearchAssociates Inc., New York.

Anderson, T.E., Roberts, D.W., 1983. Compatibility of beauveria-bassiana isolateswith insecticide formulations used in Colorado Potato Beetle (Coleoptera,Chrysomelidae) control. J. Econ. Entomol. 76, 1437–1441.

Andrusiak, K., Piotrowski, J.S., Boone, C., 2012. Chemical-genomic profiling:systematic analysis of the cellular targets of bioactive molecules. Bioorg. Med.Chem. 20, 1952–1960.

Baskaran, S., Kookana, R.S., Naidu, R., 1997. Determination of the insecticideimidacloprid in water and soil using high-performance liquid chromatography.J. Chromatogr., A 787, 271–275.

Buckingham, S.D., Lapied, B., LeCorronc, H., Grolleau, F., Sattelle, D.B., 1997.Imidacloprid actions on insect neuronal acetylcholine receptors. J. Exp. Biol.200, 2685–2692.

Cabral, M.G., Viegas, C.A., Teixeira, M.C., Sa-Correia, I., 2003. Toxicity of chlorinatedphenoxyacetic acid herbicides in the experimental eukaryotic modelSaccharomyces cerevisiae: role of pH and of growth phase and size of the yeastcell population. Chemosphere 51, 47–54.

Cernigoj, U., Lavrencic Štangar, U., Trebše, P., 2007. Degradation of neonicotinoidinsecticides by different advanced oxidation processes and studying the effectof ozone on TiO2 photocatalysis. Appl. Catal. B-Environ. 75, 229–238.

Cordova, D., Benner, E.A., Sacher, M.D., Rauh, J.J., Sopa, J.S., Lahm, G.P., Selby, T.P.,Stevenson, T.M., Flexner, L., Gutteridge, S., Rhoades, D.F., Wu, L., Smith, R.M.,Tao, Y., 2006. Anthranilic diamides: a new class of insecticides with a novelmode of action, ryanodine receptor activation. Pestic. Biochem. Physiol. 84,196–214.

Dias, P.J., Teixeira, M.C., Telo, J.P., Sa-Correia, I., 2010. Insights into the mechanismsof toxicity and tolerance to the agricultural fungicide mancozeb in yeast, assuggested by a chemogenomic approach. OMICS 14, 211–227.

Dos Santos, S.C., Teixeira, M.C., Cabrito, T.R., Sa-Correia, I., 2012. Yeasttoxicogenomics: genome-wide responses to chemical stresses with impact inenvironmental health, pharmacology, and biotechnology. Front Genet. 3, 63.

Please cite this article in press as: Mattiazzi Ušaj, M., et al. Determinatienomics in yeast. Chemosphere (2013), http://dx.doi.org/10.1016/j.che

Drobne, D., Blazic, M., Van Gestel, C.A.M., Lešer, V., Zidar, P., Jemec, A., Trebše, P.,2008. Toxicity of imidacloprid to the terrestrial isopod Porcellio scaber (Isopoda,Crustacea). Chemosphere 71, 1326–1334.

Environmental, T.D.C., 2003. Insecticide market trends and potential water qualityimplications. TDC Environmental LLC., San Mateo, CA, USA.

Henry, M., Beguin, M., Requier, F., Rollin, O., Odoux, J.F., Aupinel, P., Aptel, J.,Tchamitchian, S., Decourtye, A., 2012. A common pesticide decreases foragingsuccess and survival in honey bees. Science 336, 348–350.

Hillenmeyer, M.E., Fung, E., Wildenhain, J., Pierce, S.E., Hoon, S., Lee, W., Proctor, M.,St Onge, R.P., Tyers, M., Koller, D., Altman, R.B., Davis, R.W., Nislow, C., Giaever,G., 2008. The chemical genomic portrait of yeast: uncovering a phenotype for allgenes. Science 320, 362–365.

Iwasa, T., Motoyama, N., Ambrose, J.T., Roe, R.M., 2004. Mechanism for thedifferential toxicity of neonicotinoid insecticides in the honey bee, Apismellifera. Crop Prot. 23, 371–378.

Jemec, A., Tišler, T., Drobne, D., Sepcic, K., Fournier, D., Trebše, P., 2007. Comparativetoxicity of imidacloprid, of its commercial liquid formulation and of diazinon toa non-target arthropod, the microcrustacean Daphnia magna. Chemosphere 68,1408–1418.

Lockshon, D., Surface, L.E., Kerr, E.O., Kaeberlein, M., Kennedy, B.K., 2007. Thesensitivity of yeast mutants to oleic acid implicates the peroxisome and otherprocesses in membrane function. Genetics 175, 77–91.

Loppin, B., Berger, F., Couble, P., 2001. The Drosophila maternal gene sesame isrequired for sperm chromatin remodeling at fertilization. Chromosoma 110,430–440.

Luo, Y., Zang, Y., Zhong, Y.A., Kong, Z.M., 1999. Toxicological study of two novelpesticides on earthworm Eisenia foetida. Chemosphere 39, 2347–2356.

Malev, O., Klobucar, R.S., Fabbretti, E., Trebše, P., 2012. Comparative toxicity ofimidacloprid and its transformation product 6-chloronicotinic acid to non-target aquatic organisms: Microalgae Desmodesmus subspicatus and amphipodGammarus fossarum. Pestic. Biochem. Physiol. 104, 178–186.

Matsuda, K., Buckingham, S.D., Freeman, J.C., Squire, M.D., Baylis, H.A., Sattelle, D.B.,1998. Effects of the alpha subunit on imidacloprid sensitivity of recombinantnicotinic acetylcholine receptors. Br. J. Pharmacol. 123, 518–524.

Mattiazzi, M., Jambhekar, A., Kaferle, P., DeRisi, J.L., Krizaj, I., Petrovic, U., 2010.Genetic interactions between a phospholipase A(2) and the Rim101 pathwaycomponents in S. cerevisiae reveal a role for this pathway in response to changesin membrane composition and shape. Mol. Genet. Genomics 283, 519–530.

Mattiazzi, M., Petrovic, U., Krizaj, I., 2012. Yeast as a model eukaryote in toxinology:a functional genomics approach to studying the molecular basis of action ofpharmacologically active molecules. Toxicon 60, 558–571.

Najafi, G., Razi, M., Hoshyar, A., Shahmohamadloo, S., Feyzi, S., 2010. The effect ofchronic exposure with imidacloprid insecticide on fertility in mature male rats.Int. J. Fertil. Steril. 4, 9–16.

Neves, P.M.O.J., Hirose, E., Tchujo, P.T., Moino, J.A., 2001. Compatibility ofentomopathogenic fungi with neonicotinoid insecticides. Neotrop. Entomol.30, 263–268.

Obana, H., Okihashi, M., Akutsu, K., Kitagawa, Y., Hori, S., 2003. Determination ofneonicotinoid pesticide residues in vegetables and fruits with solid phaseextraction and liquid chromatography mass spectrometry. J. Agric. Food Chem.51, 2501–2505.

Roberts, T., 1999. Metabolic Pathways of Agrochemicals, Pt. 2: Insecticides andFungicides, vol. 2. Royal Society of Chemistry Ingram Publisher Services[Distributor], Cambridge LaVergne, p. 1500.

Roemer, T., Davies, J., Giaever, G., Nislow, C., 2012. Bugs, drugs and chemicalgenomics. Nat. Chem. Biol. 8, 46–56.

Senn, R., Hofer, D., Hoppe, T., Angst, M., Wyss, P., Brandl, F., Maienfisch, P., 1998. CGA293’343: a novel broad-spectrum insecticide supporting sustainable agricultureworldwide. Brighton Crop Protection Conference. Pests and Diseases, Brighton,UK. pp. 27–36.

Smith, J.J., Sydorskyy, Y., Marelli, M., Hwang, D., Bolouri, H., Rachubinski, R.A.,Aitchison, J.D., 2006. Expression and functional profiling reveal distinct geneclasses involved in fatty acid metabolism. Mol. Syst. Biol. 2 (2006), 0009.

Stokstad, E., 2013. Pesticides under fire for risks to pollinators. Science 340, 674–676.

Suchail, S., Guez, D., Belzunces, L.P., 2001. Discrepancy between acute and chronictoxicity induced by imidacloprid and its metabolites in Apis mellifera. Environ.Toxicol. Chem. 20, 2482–2486.

Tišler, T., Jemec, A., Mozetic, B., Trebše, P., 2009. Hazard identification ofimidacloprid to aquatic environment. Chemosphere 76, 907–914.

Tomizawa, M., Cowan, A., Casida, J.E., 2001. Analgesic and toxic effects ofneonicotinoid insecticides in mice. Toxicol. Appl. Pharmacol. 177, 77–83.

Trebše, P., Drobne, D., Stanek, K., Blazic, M., Jemec, A., Lešer, V., 2009. Comparativetoxicity of three pesticides affecting non-target terrestrial isopod Porcellioscaber (organophosphorus pesticide diazinon, neonicotinoids imidacloprid andacetamiprid). Nova Science Publishers, New York. pp. 1–13.

Wan, Y., Saleem, R.A., Ratushny, A.V., Roda, O., Smith, J.J., Lin, C.H., Chiang, J.H.,Aitchison, J.D., 2009. Role of the histone variant H2A.Z/Htz1p in TBPrecruitment, chromatin dynamics, and regulated expression of oleate-responsive genes. Mol. Cell. Biol. 29, 2346–2358.

on of toxicity of neonicotinoids on the genome level using chemog-mosphere.2013.10.063