RESEARCH ARTICLE

Mechanisms of toxicity of triphenyltin chloride (TPTC) determinedby a live cell reporter array

Guanyong Su & Xiaowei Zhang & Jason C. Raine &

Liqun Xing & Eric Higley & Markus Hecker &

John P. Giesy & Hongxia Yu

Received: 8 August 2012 /Accepted: 24 September 2012 /Published online: 6 November 2012# Springer-Verlag Berlin Heidelberg 2012

Abstract Triphenyltin chloride (TPTC), which has been ex-tensively used in industry and agriculture, can occur at con-centrations in the environment sufficient to be toxic. Here,potency of TPTC to modulate genes in a library containing1,820 modified green fluorescent protein (GFP)-expressingpromoter reporter vectors constructed from Escherichia coliK12 strains was determined. Exposure to TPTC resulted in 22(fold change>2) or 71 (fold change>1.5) differentiallyexpressed genes. The no observed transcriptional effect(NOTEC) and median transcriptional effect concentrations(TEC50) were determined to be 0.036 and 0.45 mg/L in E.coli. These responses were 1,230 and 97 times more sensitivethan the acute median effect concentration (EC50) required toinhibit growth of cells, which demonstrated that this live cellarray represents a sensitive method to assess toxic potency ofchemicals. The 71 differentially expressed genes could beclassified into seven functional groups. Of all the altered

genes, three groups which encoded for catalytic enzymes,regulatory proteins, and structural proteins accounted for28 %, 18 %, and 14 % of all altered genes, respectively. Thepattern of differential expression observed during this studywas used to elucidate the mechanism of toxicity of TPTC. Todetermine potential relationships among genes that werechanged greater than 2.0-fold by exposure to TPTC, a corre-lation network analysis was constructed, and four genes wererelated to aroH, which is the primary target for metabolicregulation of aromatic biosynthesis by feedback inhibition inbacteria. The genes rnC, cld, and glgS were selected as po-tential biomarkers for TPTC, since their expression was morethan 2.0-fold greater after exposure to TPTC.

Keywords High throughput . NOTEC . Biomarker .

Correlation network . Toxicity assessment . Bacterial .

Genomics . Organotin

Responsible editor: Philippe Garrigues

Electronic supplementary material The online version of this article(doi:10.1007/s11356-012-1280-7) contains supplementary material,which is available to authorized users.

G. Su :X. Zhang : L. Xing : J. P. Giesy :H. YuState Key Laboratory of Pollution Control and Resource Reuse &School of the Environment, Nanjing University,Nanjing, China

G. Su : J. C. Raine : E. Higley :M. Hecker : J. P. GiesyToxicology Centre,University of Saskatchewan,Saskatoon, SK S7N 5B3, Canada

M. HeckerSchool of Environment and Sustainability,University of Saskatchewan,Saskatoon, SK S7N 5B3, Canada

J. P. GiesyDepartment of Biomedical Veterinary Sciences and ToxicologyCentre, University of Saskatchewan,Saskatoon, SK S7N 5B3, Canada

J. P. GiesyDepartment of Biology and Chemistry and State Key Laboratoryin Marine Pollution, City University of Hong Kong,83 Tat Chee Avenue,Kowloon, Hong Kong SAR, China

X. Zhang (*) :H. Yu (*)School of the Environment, Nanjing University,Nanjing 210089, Chinae-mail: howard50003250@yahoo.come-mail: yuhx@nju.edu.cn

Environ Sci Pollut Res (2013) 20:803–811DOI 10.1007/s11356-012-1280-7

Introduction

Organotin compounds are widely used with a worldwideproduction that has increased almost tenfold over the past40 years (Liu et al. 2006). These compounds are extensivelyused in industry and agriculture as biocides, fungicides, anti-fouling agents in boat paint, wood preservatives, catalysts, andstabilizers for polyvinylchloride polymers (Fent and Muller1991; Grote et al. 2007; Sano et al. 2010). Of the organotincompounds, triphenyltin chloride (TPTC) is one of the mostpotent. Concentrations less than 40 ngTPTC/ml can affect theimmune system of cultivated clams (Tapes philippinarum)(Cima et al. 1998) and enhance histone acetyltransferaseactivity, which is a type endocrine-disruption (Osada et al.2005). TPTC can also inhibit gap junctional intercellularcommunication in WB-F344 rat liver epithelial cells (Lee etal. 2010) and cause loss of post-implantation of embryos,failure to implant, and effects on body mass, size, and struc-ture of testicles and lesser fertility of Holtzmann rats (Ema2000). Because of bio-concentration and accumulation poten-tials (Horiguchi et al. 1997; Shim et al. 2000), TPTC couldalso pose a risk to human health through dietary exposure.While it has been established that TPTC is more toxic thansome other pollutants, the mechanisms by which it causestoxicity remain mostly unknown.

Genome-wide transcriptional investigations using wholecell arrays is a useful toxicogenomic approach to character-ize modes of toxic action of chemicals (Zhang et al. 2011; Suet al. 2012). Whole cell arrays consist of an assortment ofgenetically engineered microorganisms tailored to respondto activation of specific promoters. Fusion of stress pro-moters to reporter genes (such as fluorescent proteins) isthe basic concept for detection of cellular signaling (Elad etal. 2010). Compared with microarray technology, whole cellarrays avoid complex protocols of pre-treatment and high-cost experimental materials, have fewer interferences, andcan provide temporal resolution (Onnis-Hayden et al. 2009).Furthermore, the short testing time (less than 3 h) makes livecell arrays rapid, economical, high-throughput biosensorsystems for detecting toxicity and determining effects onspecific signaling pathways. Consequently, this assay is notspecific to bacteria and can represent responses of systemsthat are conserved in multiple organisms, including metazo-ans (Zhang et al. 2011).

Here, toxicity of TPTC was assessed by application of acomprehensive cell array of transcriptional fusions of GFPto each of 1,820 different gene promoters in Escherichiacoli K12. Profiles of concentration- and time-dependentexpression of genes caused by TPTC were obtained over3-h exposures to 0.1, 1, or 10 mg TPTC/L. All altered geneswere classified into seven groups according to their knownfunction (Supplementary Table 1), and the pattern of whichwas hypothesized to be indicative of the specific molecular

signaling pathways affected by exposure to TPTC. Based onthese profiles of gene expression, a correlation network wasgenerated to elucidate potential correlations of differentiallyexpressed molecular pathways.

Materials and methods

Microbial live cell array

The microbial promoter collection was produced byresearchers at the Weizmann Institute of Science (Rehovot,Israel) and includes more than 1,900, out of 2,500 pro-moters in the entire genome of E. coli K12 strain MG1655(Zaslaver et al. 2006). Each of the reporter strains iscoupled with a bright, fast-folding GFP fused to a full-length copy of an E. coli promoter in a low-copy plasmid.This enables measurement of gene expression withinminutes with high accuracy and reproducibility. All cloneswere grown at 37 °C in lysogeny broth (LB)–Lennoxmedia plus 25 mg/L kanamycin.

Exposure to TPTC

Triphenyltin chloride was purchased from Sigma Aldrich(#245712, St. Louis, MO, USA). A TPTC stock solution(20,000 mg/L) was prepared in dimethyl sulfoxide (DMSO),and other stock solutions were made by serial dilution withDMSO. Chemical-induced effects on growth of cells wereassessed by measuring optical density (OD) at 600 nm by aFluostar OPTIMA microplate reader (BMG Labtech, Offen-burg, Germany). For each concentration, three replicationswere conducted. Specifically, growth and division of E. coliwas determined after 4 h of incubation at 37 °C. The OD600value is the most commonly used to estimate the E. coli celldensity and corresponds with cell number in a given E. coliculture volume (Luo et al. 2011; Su et al. 2012).

To measure expression of genes, assay plates were preparedby adding 72 μL of LBmedium to each well in black 384-welloptical bottom plates (NUNC, Rochester, NY, USA). E. colistrains were inoculated in the 384-well plate from a 96-wellstock plate by disposable replicators (Genetix, San Jose, CA,USA). Cells were incubated at 37°C for a 2.5 h in 384-wellplates, and then 3.8 μL of DMSO (solvent control) or TPTCstock solutions were added into individual wells of the 384-well plates to make final concentrations of 0, 0.1, 1, or 10 mgTPTC/L. GFP intensity in each well was consecutively mon-itored every 10min during 3 h by use of the Fluostar OPTIMA120 microplate reader (excitation/emission, 545 nm/590 nm).Effects of TPTC on growth of cells were assessed by ODmeasurement prior to the promoter reporter assay. None ofthe three concentrations of TPTC caused significant effectson growth of cells during 10 h of exposure.

804 Environ Sci Pollut Res (2013) 20:803–811

Data analyses

All data analyses have been described previously (Zhang etal. 2011; Su et al. 2012). To select the promoter reportersthat were significantly differentially expressed in responseto exposure with TPTC, a linear regression model wasapplied. The response measured as GFP fluorescence inten-sity was fitted to a function of time for each promoterreporter strain. Classification and visualization of the geneexpression were derived by ToxClust (Zhang et al. 2009).

Determination of NOTEC

The No Observed Transcriptional Effect Concentration(NOTEC) was calculated based on the number of promoterstrains in the library of 1,820 genes that were significantlyaltered by TPTC. The No Observable Effect Concentration(NOEC) based on inhibition of growth by TPTC was deter-mined. Then, the percentage of genes differentially expressedat different concentrations relative to the NOEC was calculat-ed. Finally, a generalized linear binomial model was used toassess the concentration-dependent response curve of the per-centage of the differentially expressed genes. The NOTECwas calculated as the maximum concentration of TPTC atwhich less than 5 % of the genes were differentially expressedupon chemical exposure compared with control (Su et al.2012).

Pathway analysis

Lists of genes were developed for further analysis based onstatistical significance and 1.5- or 2.0-fold change cutoffs.Differentially expressed genes were classified into sevengroups based on their biological functions (www.ecogene.org,www.geneontology.org, and www.ecoliwiki.net; Supplementa-ry Table 1). Correlation network analyses were conducted usingthe “GeneNet” package by R software (http://cran.r-project.org/web/packages/GeneNet/). This method was used for analyzinggene expression (time series) data with focus on the inferenceof gene networks (Opgen-Rhein and Strimmer 2007).

Results

Inhibition of E. coli growth by TPTC

TPTC inhibited growth of E. coli cells in a concentration-dependent manner (Fig. 1). The median effect concentration(EC50), NOEC, and lowest observed effect concentration ofTPTC on cell growth were 43.7, 10.0, and 20.0 mg/L, respec-tively. Three concentrations, 0.1, 1, and 10 mg/L, were select-ed as exposure concentrations to assess the effects of TPTC ontranscriptional expression profiles of E. coli. At these

concentrations, cell growth would not be affected. The NOECwas included to enable determination of the NOTEC.

Gene expression profiles

Expression of genes by the microbial reporter strains wasmodulated by TPTC in a time- and concentration-dependentmanner. Exposure to TPTC resulted in fewer upregulatedpromoter strains than downregulated strains during the 3-h exposure for genes selected by both the 1.5-fold (Fig. 2b)and 2.0-fold cut-offs (Fig. 2a). Of the 22 promoter reporterstrains selected using a 2.0-fold cut-off, 2 and 20 strains wereup- and downregulated, respectively. Of the 71 promoterreporter strains selected by application of a 1.5-fold cut-off,the greatest downregulation of as much as eightfold relative tothe controls was observed for rrnC. Furthermore, this genewas separated from all other groups when data was subjectedto analysis by ToxClast. Other than rrnC, expression of 16 and54 genes were up- and downregulated, respectively (Fig. 2b).

Alteration of gene expression by TPTC was concentration-dependent. Using a 2.0-fold change as a cut-off, exposure to 0.1,1.0, or 10 mg, TPTC/L significantly altered expression of 2, 6,and 21 promoters, respectively. Only one strain was responsiveto all three concentrations (Fig. 3). Of the 1,820 genes, 71 weredifferentially expressed with a maximum absolute fold changeof at least 1.5 (Fig. 3). Among those genes, 17, 33, and 71promoters were differentially expressed after exposure to 0.1,1.0, or 10 mg TPTC/L, respectively, and 17 strains were respon-sive to all three concentrations. Exposure to 0.1 or 1.0 mgTPTC/L resulted in one gene, yhhY, being completely differentfrom those modulated by exposure to 10 mg TPTC/L (Fig. 3).

Determination of no observed transcriptional effectconcentration

Ratios of differentially expressed genes were 24 %,47 %, and 100 % after exposure to 0.1, 1.0, or 10 mg

Fig. 1 Inhibition profile of E. coli growth by different concentrationsof TPTC (data points were shown with mean values of threereplications)

Environ Sci Pollut Res (2013) 20:803–811 805

TPTC/L, respectively. The NOTEC was 0.036 mg TPTC/L, a concentration at which fewer than 5 % of geneswere differentially expressed relative to controls. Themedian transcriptional effect concentration (TEC50) was0.45 mg TPTC/L.

In previous publications, acute toxicity of TPTC rangedfrom 2.0 to 4,455 μg/L in nine aquatic species thatincluded algae, fish, mollusks, worms, daphnids, andmysid shrimp (Fig. 4) (Goel and Prasad 1978; Goel andSrivastava 1981; Wong et al. 1982; Devries et al. 1991;Nagase et al. 1991; Fargasova 2002, 1997). Of thesespecies, green algae (Ankistrodesmusfalcatus ssp. acicul)were the most sensitive species with an EC50 of 2.0 μgTPTC/L. Based on effects of TPTC on cell growth, E. coliwas less sensitive than all other species and endpoints(Supplementary Table 3). However, the transcriptionalendpoints, NOTEC and TEC50, in E. coli measured inthis study were 1,230- and 97-fold more sensitive than theacute EC50 for inhibition of cell growth. Aligned onto thesensitive distribution curve of aquatic species, the NOTECand TEC50 are equivalent to the 47th and 74th centilespecies, respectively.

Classification of differentially expressed genes

All differentially expressed genes were classified into one ofseven groups: enzyme (pepE, aroH, serB, add, gcvT, lacZ,cysK, serA, dcd, dmsA, ligB, fpr, galU, gadB, metA, icd, lipA,cueO, aroK, and ttdA), regulatory protein (cysB, cspD, ftsK,gadX, phoP, mcrA, rmf, rsd, wrbA, fdhD, flgM, manX, and cld),structural protein (rpsT, rpsU, mglB, ytfF, b1403, ompC, ompA,skp, osmC, and aqpZ), rRNA or tRNA (argW, rrnB, rrnD, andrrnC), stress responsive pathway (evgA, uspB, uspA), aminoacids formation (menG), and unclear function (yfbV, mntR,yhgF, yjdI, yhcG, yafK, insA_7, yaiE, b3007, yhhY, yeaU, glgS,proQ, yedW, yedP, ymcC, yfiE, uspF, yifE, and ybcW), whichaccounted for 28 %, 18 %, 14 %, 6 %, 4 %, 1 %, and 28 % ofthe 71 genes altered by TPTC, respectively. Among thesedifferentially expressed genes, 22 (31 % of 71) genes’ foldchange was greater than 2.0 (Supplementary Table 4).

Correlation network

A correlation network was constructed to analyze high-dimensional data from E. coli gene expression (fold

11

aroH

serB

rrnC

yhcG

yhhY

mcrA

aqpZ

galU

add

wrbA

glgS

cld

yifE

uspF

menG

insA_7

mglB

yjdI

lacZ

ybcW

skp

rmf

Fold Change

0.1

0.2

0.5 1 2

0 30 60 90 120 150 180

0.1 mg/L1 mg/L

10 mg/L

Time (min)

Con

c.

aFig. 2 Real-time geneexpression profiles ofdifferentially expressed gens inE. coli after exposure to 0.1,1.0, or 10 mg TPTC/L asrepresented by the lower,middle, and upper bands ineach gene column, respectively.Classification and visualizationof the gene expression werederived by ToxClust.Dissimilarity between geneswas calculated by theManhattan distance between thegene expressions at all theconcentration versus timecombinations. Fold change ofgene expression is indicated bycolor gradient, and the timecourse of expression changes isindicated from left to right. aClustering of the time-dependent expression of theTPTC altered genes selected by2.0-fold change cut-off. bClustering of the time-dependent expression of theTPTC altered genes selected by1.5-fold change cut-off

806 Environ Sci Pollut Res (2013) 20:803–811

change>2) after exposure to TPTC based on graphicalGaussian models (Opgen-Rhein and Strimmer 2007)(Fig. 5). From their expression profiles after 3 h, correlationcoefficients between each two related nodes (genes) werecalculated. Six of the 22 genes that were altered by morethan 2.0-fold were directly related with aroH, and fourgenes (wrbA, cld, rmf, and glgs) exhibited significant

correlations with aroH protein (r>0.89). And four genes(uspF, aqpZ, galU, aroH) were associated with rrnC. Func-tions of these genes are given in Supplementary Table 1.

rrnCgalUdmsAyhcGrpsTrrnBaroKompCrsdevgAcspDrpsUlipAyedWompAyifEuspFproQmanXgadBrrnDmcrAaqpZybcWrmfglgScldmenGskpwrbAaroHserBcysKb3007fprphoPmetAyhgFgcvTligBcueOyfbVpepEuspAyeaUytfFinsA_7mglBuspBftsKserAosmCyfiEyhhYttdAlacZyjdIcysByafKflgMicdargWdcdaddgadXyaiEyedPmntRymcCfdhDb1403

Fold Change

0.1

0.2

0.5 1 2

bFig. 2 (continued)

Fig. 3 Concentration-dependent promoter activity of reporter strainsin TPTC exposure. (Venn diagram displayed the differentiallyexpressed genes selected by 1.5- or 2.0-fold change cut-off at threedifferent TPTC concentrations including 0.1, 1, and 10 mg/L, whichwere marked with red, green, and blue, respectively)

Fig. 4 Species sensitivity distribution using the ecotoxicity data ofTPTC and the data acquired in the present study. Probit model wasfitted for different species. NOTEC and TEC 50 are represented as blueand red asterisks on the fitted curve, respectively. Detailed ecotoxicitydata are available in the Supplementary Table 2)

Environ Sci Pollut Res (2013) 20:803–811 807

Discussion

The fact that transcriptional endpoints measured in this studywere more sensitive than the acute EC50 based on inhibitionof growth of cells demonstrated that effects of TPTC ontranscription as expressed by the NOTEC represent a sensitiveendpoint to assess toxicity of this chemical. This result isconsistent with previous reports that have demonstrated thatthe NOTEC is more sensitive than conventional endpoints,since it reflects sub-lethal and molecular level responses to atoxicant (Lobenhofer et al. 2004; Poynton et al. 2008). Therelatively inexpensive live cell array could provide a sensitivetool to assess the toxicity of environmental chemicals in ashort time (3 h). Future studies should evaluate the sensitivityof live cell array relative to several commonly used testspecies for additional chemicals.

The gene expression profile suggested that TPTC can causetoxicity toE. coli throughmodulation of enzymes in biochemicalreactions, including pepE, aroH, serB, add, gcvT, lacZ, cysK,serA, dcd, dmsA, ligB, fpr, galU, gadB, metA, icd, lipA, cueO,aroK, and ttdA, formation of regulatory proteins including cysB,cspD, ftsK, gadX, phoP, mcrA, rmf, rsd, wrbA, fdhD, flgM,manX, and cld, and structural proteins, including rpsT, rpsU,mglB, ytfF, b1403, ompC, ompA, skp, osmC, and aqpZ, rRNA ortRNA, including argW, rrnB, rrnD, and rrnC, stress responsivepathways, evgA, uspB, and uspA, and formation of amino acids(menG). To our knowledge, this is the first report using the E.coli whole cell assay to assess toxicity of TPTC. However, thetoxic action of TPTC can be grouped into four general catego-ries: (1) effects on immune function (Nishida et al. 1990), (2)

clastogens (Sasaki et al. 1994), (3) cytotoxicity (Snoeij et al.1985), and (4) inhibition of intercellular gap junctions.

Modulation of enzymes in biochemical reactions might beone of the most important TPTC-induced toxic pathways,since genes falling into this functional group accounted for28 % of the total differentially expressed genes. After expo-sure to TPTC, 19 enzymes were altered, including α-aspartyldipeptidase (pepE), 3-deoxy-D-arabino-heptuloso-nate-7- phosphate synthase (aroH), 3-phosphoserinephosphatase (serB), adenosine deaminase (add), aminome-thyltransferase (gcvT), β-galactosidase (lacZ), cysteine syn-thase A (cysK), D-3-phosphoglycerate dehydrogenase(serA), deoxycytidine triphosphate deaminase (dcd), dimeth-yl sulfoxidereductase (dmsA), DNA ligase (ligB),ferredoxin-NADP reductase (fpr), glucose-1-phosphate uri-dylyltransferase (galU), glutamate decarboxylase B subunit(gadB), homoserinetranssuccinylase (icd), isocitrate dehy-drogenase (metA), isocitrate dehydrogenase (metA), lipoatesynthase (lipA), shikimate kinase I (aroK), L-tartrate dehy-dratase (ttdA). Glucose-1-phosphate uridylyltransferase(galU) is an enzyme associated with glycogenesis, and itsdownregulation by TPTC would inhibit synthesis of UDP-glucose from glucose-1-phosphate and UTP (Thoden andHolden 2007). As a senescence-associated enzyme, down-regulation of β-galactosidase (lacZ) might imply that E. coliis senescent after exposure to TPTC (Pardee et al. 1959).Transcriptional activities of these two genes were bothdownregulated more than 2.0-fold by TPTC. Expression of3-phosphoserine phosphatase (serB) was upregulated bymore than 2.0-fold, which indicates that a phosphoserinephosphatase process might be disturbed by exposure toTPTC (Veiga-da-Cunha et al. 2004). 3-Phosphoserine phos-phatase has been shown to be a breast cancer marker mole-cule (Pestlin et al. 2005), and its altered expression might beindicative of another new toxicity mechanism by TPTC.

The fact that nearly 18% of all altered genes were classifiedas “regulatory proteins” suggested that TPTC can also causetoxicity through its disturbance of transcriptional regulators,activators, or inhibitors, especially for processes involvingDNA. Four genes, cld, mcrA, rmf, and wrbA, were downregu-lated more than 2.0-fold. Based on downregulation of cld, thelength of the O-antigen component of lipopolysaccharideswould be disturbed after exposure of E. coli to TPTC (Raetzand Whitfield 2002). As a nuclease, downregulation of mcrAprotein suggested the potential damage to DNA of bacteriaafter exposure of TPTC (Anton and Raleigh 2004). The factthat the ribosome modulation factor was shown to influencesurvival of E. coli under acid stress has already been shown byothers (Yamagishi et al. 1993). Thus, downregulation of rmfcould be indicative of damage to this bacterium. Disturbanceof wrbA was proposed to be implicated in protection againstoxidative stress (Burnett et al. 1974).

Structural proteins, such as membrane proteins and watermajor intrinsic protein (MIP) channels, represented another

Fig. 5 Sparse graphical Gaussian model for 22 genes inferred from anE. coli live cell array data set with 19 data points

808 Environ Sci Pollut Res (2013) 20:803–811

protein group altered by TPTC. aqpZ encodes for water MIPchannels (Hovijitra et al. 2009), which was downregulatedby less than 0.5-fold. aqpZ protein’s downregulation woulddisturb the cell’s osmoregulatory capacity since it allows E.coli to adapt to osmotic variations by rapid diffusion ofwater molecules. Otherwise, the fact that both skp and ompAwere downregulated by TPTC demonstrated that skp canbind outer membrane proteins, such as ompA. Lack of skpprotein would lead to accumulation of protein aggregates inthe periplasm, which also implies that skp can recognizeearly folding intermediates of outer membrane proteins(Schafer et al. 1999). The gene mglB, which encodes forthe D-galactose-binding periplasmic protein, was downregu-lated by less than 0.5-fold. Downregulation of expression ofthis protein would inhibit transport of galactose and glucose.

TPTC elicited transcriptional alteration of a group of“rRNA or tRNA” genes, which produced RNA after expres-sion. All genes in this category were downregulated byTPTC. rrnC, as one of seven ribosomal RNA operons (Yeonet al. 2008), was suppressed less than 0.2-fold. This sug-gested that inhibition of decoding mRNA into amino acidsmight be a mechanism of toxicity for TPTC. The genemenG, which is related to adenosylmethionine and belongsto the group of “amino acids formation,” was downregulatedby TPTC. This suggested that inhibition of adenosylmethio-nine formation was another toxicity pathway of TPTC.

TPTC altered transcription of a group of stress responsivegenes, which can be divided into three categories accordingto previous studies (Onnis-Hayden et al. 2009): detoxifica-tion (uspB), drug resistance/sensitivity (evgA), and generalstress (uspA). Both uspB and evgA are related to compound-/chemical-induced mortality or stress, such as response toantibiotics, and were downregulated after exposure toTPTC. However, uspA was upregulated by TPTC, whichimplied that biochemical and biophysical homeostasis of thecell were disturbed after a 3-h exposure to TPTC.

After construction of the correlation network, significantrelationships among genes altered by TPTC were observed.Through the process of the network construction, a graphicalGaussian model, also known as covariance selection or con-centration graph, was employed. Based on the constructednetwork, two genes, aroH and rrnC, seemed to have beenvery actively involved in the response to the exposure withTPTC and were related with six and four other genes, respec-tively. The gene aroH encodes 3-deoxy-D-arabino-heptuloso-nate-7-phosphate synthase (DAHPS), which is feedback-regulated by tyrosine and phenylalanine (Shumilin et al.2004) and is the primary target for metabolic regulation ofaromatic biosynthesis by feedback inhibition in bacteria andfungi (Keith et al. 1991). Based on the observed significantcorrelations, it is hypothesized that disturbance of DAHPSwould affect four genes including wrbA, cld, rmf, and glgs.Fold change of expression of rrnC was less than 0.2, and

downregulation of expression of this gene might contribute toor is affected by four related genes (uspF, aqpZ, galU, aroH)through the gene network.

The results indicate that the E. coli whole cell array hasthe potential to identify novel biomarkers for determinationof specific chemical classes in environmental media (Gou etal. 2010; Watson and Mutti 2004). Three general principlesare proposed for selection of biomarkers of chemical pollu-tion based on use of the E. coli array: (1) The endpoints needto be chemical-specific; (2) the magnitude of changes in geneexpression should be related to the concentration of chemi-cal; and (3) the change in gene expression should be greatenough that it can be monitored easily. Based on thesecriteria, rrnC, cld, and glgS are recommended as potentialbiomarkers for TPTC, as their fold changes in expressionwere greater than 2 and proportional to concentrations ofTPTC between 0.1 and 10 mg/L (Supplementary Figures 1A,1B, 1C). cld and glgS genes encode proteins that regulate thelength of the O-antigen component of lipopolysaccharidechains and can serve as a predictor of synthesis of glycogen.The gene rrnC directly encodes rRNA. However, to date,there were no reports that expression of these three genes isregulated by some specific chemicals. Suitability of thesegenes as biomarkers remains to be validated and requiresfurther investigation into the consistency and TPTC-specificity of their responses. For this purpose, field TPTC-sample tests will be performed with these genes in our futurework.

Acknowledgments The research was supported by grants from Na-tional Natural Science Foundation of China (NSFC) (grant no.21007025), Jiangsu Provincial Key Technology R&D Program(#BE2011776), Jiangsu Provincial Environment Monitoring Station(Project # 1012), National Science and Technology Major Project (No.2008ZX08526-003), and a Discovery Grant from the National Scienceand Engineering Research Council of Canada (Project # 326415-07), anda grant from the Western Economic Diversification Canada (Projects #6578 and 6807). The authors wish to acknowledge the support of aninstrumentation grant from the Canada Foundation for Infrastructure.Prof. Giesy was supported by the program of 2012 "High Level ForeignExperts" (#GDW20123200120) funded by the State Administration ofForeign Experts Affairs, P.R. China. He was also supported by theCanada Research Chair program, an at-large Chair Professorship at theDepartment of Biology and Chemistry and State Key Laboratory inMarine Pollution, City University of Hong Kong, and the Einstein Pro-fessor Program of the Chinese Academy of Sciences. Prof. Zhang wassupported by a Program for New Century Excellence Talents in Univer-sities (Ministry of Education, China).Mr. Guanyong Su was supported bythe Shanghai Tongji Gao Tingyao Environmental Science and Technol-ogy Development Foundation (STGEF).

References

Anton BP, Raleigh EA (2004) Transposon-mediated linker insertion scan-ning mutagenesis of the Escherichia coli McrA endonuclease. J

Environ Sci Pollut Res (2013) 20:803–811 809

Bacteriol 186(17):5699–5707. doi:jb.186.17.5699-5707.2004/jb.186.17.5699-5707.2004

Burnett RM, Darling GD, Kendall DS, Lequesne ME, Mayhew SG,Smith WW, Ludwig ML (1974) Structure of oxidized form ofclostridial flavodoxin at 1.9-A resolution—description of flavinmononucleotide binding-site. J Biol Chem 249(14):4383–4392

Cima F, Marin MG, Matozzo V, Da Ros L, Ballarin L (1998) Immu-notoxic effects of organotin compounds in Tapes philippinarum.Chemosphere 37(14–15):3035–3045

Devries H, Penninks AH, Snoeij NJ, Seinen W (1991) Comparativetoxicity of organotin compounds to rainbow-trout (Oncorhynchusmykiss) yolk-sac fry. Sci Total Environ 103(2–3):229–243

Elad T, Lee JH, Gu MB, Belkin S (2010) Microbial cell arrays. Wholecell sensing systems I: reporter cells and devices 117:85–108.doi:10.1007/10_2009_16

Ema M (2000) Reproductive and developmental toxicity of triphenyl-tin chloride in rats. Congenit Anomalies 40(1):8–13

Fargasova A (1997) Comparative study of ecotoxicological effect oftriorganotin compounds on various biological subjects. EcotoxEnviron Saf 36(1):38–42

Fargasova A (2002) Structure-affected algicidal activity of triorganotincompounds. B Environ Contam Tox 69(5):756–762. doi:10.1007/s00128-002-0125-3

Fent K, Muller MD (1991) Occurrence of organotins in municipalwaste-water and sewage-sludge and behavior in a treatment-plant. Environ Sci Technol 25(3):489–493

Goel HC, Prasad R (1978) Action of molluscicides on freshly laid eggs ofsnail Indoplanorbis exustus (Deshayes). Indian J Exp Biol 16(5):620–622

Goel HC, Srivastava CP (1981) Laboratory evaluation of some mollusci-cides against french water snails, Indoplanorbis and Lymnaea spe-cies. J Commun Dis 13(2):121–127

Gou N, Onnis-Hayden A, GuAZ (2010) Mechanistic toxicity assessmentof nanomaterials by whole-cell-array stress genes expression analy-sis. Environ Sci Technol 44(15):5964–5970. doi:10.1021/es100679f

Grote K, Hobler C, Andrade AJM, Grande SW, Gericke C, TalsnessCE, Appel KE, Chahoud I (2007) Effects of in utero and lacta-tional exposure to triphenyltin chloride on pregnancy outcomeand postnatal development in rat offspring. Toxicology 238:177–185. doi:10.1016/j.tox.2007.05.033

Horiguchi T, Shiraishi H, Shimizu M, Morita M (1997) Effects oftriphenyltin chloride and five other organotin compounds on thedevelopment of imposex in the rock shell, Thais clavigera. Envi-ron Pollut 95(1):85–91

Hovijitra NT, Wuu JJ, Peaker B, Swartz JR (2009) Cell-free synthesisof functional aquaporin Z in synthetic liposomes. BiotechnolBioeng 104(1):40–49. doi:10.1002/bit.22385

Keith B, Dong XN, Ausubel FM, Fink GR (1991) Differential induc-tion of 3-deoxy-D-arabino-heptulosonate 7-phosphate synthasegenes in Arabidopsis thaliana by wounding and pathogenic at-tack. P Natl Acad Sci USA 88(19):8821–8825

Lee C-H, Chen IH, Lee C-R, Chi C-H, Tsai M-C, Tsai J-L, Lin H-F(2010) Inhibition of gap junctional intercellular communication inWB-F344 rat liver epithelial cells by triphenyltin chloride throughMAPK and PI3-kinase pathways. J Occup Med Toxicol 5:17

Liu SM, Hsia MP, Huang CM (2006) Accumulation of butyltin com-pounds in cobia Rachycentron canadum raised in offshore aqua-culture sites. Sci Total Environ 355(1–3):167–175. doi:10.1016/j.scitotenv.2005.02.025

Lobenhofer EK, Cui XG, Bennett L, Cable PL,Merrick BA, Churchill GA,Afshari CA (2004) Exploration of low-dose estrogen effects: identifi-cation of No Observed Transcriptional Effect Level (NOTEL). TToxicol Pathol 32(4):482–492. doi:10.1080/01926230490483324

Luo Z, Wu Q, Zhang M, Li P, Ding Y (2011) Cooperative antimicrobialactivity of CdTe quantum dots with rocephin and fluorescence mon-itoring for Escherichia coli. J Colloid Interface Sci 362(1):100–106

Nagase H, Hamasaki T, Sato T, Kito H, Yoshioka Y, Ose Y (1991)Structure-activity-relationships for organotin compounds on thered killifish Oryzias latipes. Appl Organomet Chem 5(2):91–97

Nishida H, Matsui H, Sugiura H, Kitagaki K, Fuchigami M, Inagaki N,Nagai H, Koda A (1990) The immunotoxicity of triphenyltinchloride in mice. J Pharmacobiodyn 13(9):543–548

Onnis-Hayden A, Weng HF, He M, Hansen S, Ilyin V, Lewis K, Gu AZ(2009) Prokaryotic real-time gene expression profiling for toxicityassessment. Environ Sci Technol 43(12):4574–4581. doi:10.1021/es803227z

Opgen-Rhein R, Strimmer K (2007) From correlation to causationnetworks: a simple approximate learning algorithm and its appli-cation to high-dimensional plant gene expression data. BMC SystBiol 1. doi:10.1186/1752-0509-1-37

Osada S, Nishikawa J, Nakanishi T, Tanaka K, Nishihara T (2005) Someorganotin compounds enhance histone acetyltransferase activity.Toxicol Lett 155(2):329–335. doi:10.1016/j.toxlet.2004.10.009

Pardee AB, Jacob F, Monod J (1959) Genetic control and cytoplasmicexpression of inducibility in the synthesis of beta-galactosidase byE. coli. J Mol Biol 1(2):165–178

Pestlin G, Andres H, Berndt P, Hagmann M, Karl J, Langen H, Zolg W(2005) Diagnosing breast cancer from a liquid sample (e.g. serum,plasma, whole blood or nipple aspirate fluid) derived from anindividual comprises measuring the amount of L-3-phosphoserinephosphatase in the sample. WO2005050217-A1

Poynton HC, Loguinov AV, Varshavsky JR, Chan S, Perkins EI, VulpeCD (2008) Gene expression profiling in Daphnia magna part I:concentration-dependent profiles provide support for the no ob-served transcriptional effect level. Environ Sci Technol 42(16):6250–6256. doi:10.1021/es8010783

Raetz CRH, Whitfield C (2002) Lipopolysaccharide endotoxins.A n n u R e v B i o c h em 71 : 6 3 5 – 7 0 0 . d o i : 1 0 . 11 4 6 /annurev.biochem.71.110601.135414

Sano T, Takagi H, Nagano K, Nishikawa M (2010) Analysis of tri-organotin compounds in water samples by hydrophilic interactionliquid chromatography-electrospray ionization-mass spectrome-try. J Chromatogr A 1217(26):4344–4346. doi:10.1016/j.chroma.2010.04.046

Sasaki YF, Sakaguchi M, Yamada H, Miyahara T, Kozuka H (1994)Antagonizing effect of triphenyltin chloride on cytosine-1-β-D-arabinofuranoside potentiation of chromosome aberrations in-duced by mitomycin C. Mutat Res 323(3):99–104

Schafer U, Beck K, Muller M (1999) Skp, a molecular chaperone ofgram-negative bacteria, is required for the formation of solubleperiplasmic intermediates of outer membrane proteins. J BiolChem 274(35):24567–24574

Shim WJ, Kahng SH, Hong SH, Kim NS, Kim SK, Shim JH (2000)Imposex in the rock shell, Thais clavigera, as evidence of organotincontamination in the marine environment of Korea. Mar EnvironRes 49(5):435–451

Shumilin IA, Bauerle R,Wu J,Woodard RW,Kretsinger RH (2004) Crystalstructure of the reaction complex of 3-deoxy-D-arabino-heptulosonate-7-phosphate synthase from Thermotoga maritima refines the catalyticmechanism and indicates a new mechanism of allosteric regulation. JMol Biol 341(2):455–466. doi:10.1016/j.jmb.2004.05.077

Snoeij NJ, Vaniersel AAJ, Penninks AH, Seinen W (1985) Toxicity oftriorganotin compounds—comparative in vivo studies with a se-ries of trialkyltin compounds and triphenyltin chloride in male-rats. Toxicol Appl Pharm 81(2):274–286

Su G, Zhang X, Liu H, Giesy JP, Lam MH, Lam PK, Siddiqui MA,Musarrat J, Al-Khedhairy A, Yu H (2012) Toxicogenomic mecha-nisms of 6-HO-BDE-47, 6-MeO-BDE-47, and BDE-47 in E. coli.Environ Sci Technol 46(2):1185–1191. doi:10.1021/es203212w

Thoden JB, Holden HM (2007) Active site geometry of glucose-1-phosphate uridylyltransferase. Protein Sci 16(7):1379–1388.doi:10.1110/ps.072864707

810 Environ Sci Pollut Res (2013) 20:803–811

Veiga-da-Cunha M, Collet JF, Prieur B, Jaeken J, Peeraer Y, Rabbijns A,Van Schaftingen E (2004)Mutations responsible for 3-phosphoserinephosphatase deficiency. Eur J Hum Genet 12(2):163–166.doi:10.1038/sj.ejhg.5201083

Watson WP, Mutti A (2004) Role of biomarkers in monitoring expo-sures to chemicals: present position, future prospects. Biomarkers9(3):211–242. doi:10.1080/13547500400015642

Wong PTS, Chau YK, Kramar O, Bengert GA (1982) Structure–toxicity relationship of tin-compounds on algae. Can J Fish AquatSci 39(3):483–488

Yamagishi M, Matsushima H, Wada A, Sakagami M, Fujita N,Ishihama A (1993) Regulation of the Escherichia-coli Rmfgene encoding the ribosome modulation factor—growthphase-dependent and growth rate-dependent control. EMBO J12(2):625–630

Yeon SM, Choi BS, Kim YC (2008) Organization of three rRNA (rrn)operons from Sphingobium chungbukense DJ77. J Microbiol 46(6):697–703. doi:10.1007/s12275-008-0193-0

Zaslaver A, BrenA, RonenM, Itzkovitz S,Kikoin I, Shavit S, LiebermeisterW, SuretteMG,AlonU (2006)A comprehensive library of fluorescenttranscriptional reporters for Escherichia coli. NatMethods 3:623–628.doi:10.1038/nmeth895

Zhang XW, Newsted JL, Hecker M, Higley EB, Jones PD, Giesy JP(2009) Classification of chemicals based on concentration-dependent toxicological data using ToxClust. Environ Sci Tech-nol 43(10):3926–3932

Zhang XW, Wiseman S, Yu HX, Liu HL, Giesy JP, Hecker M (2011)Assessing the toxicity of naphthenic acids using a microbialgenome wide live cell reporter array system. Environ Sci Technol45(5):1984–1991. doi:10.1021/es1032579

Environ Sci Pollut Res (2013) 20:803–811 811

Mechanisms of Toxicity of Triphenyltin Chloride (TPTC)

Determined by a Live Cell Reporter Array

Guanyong Su1,2, Xiaowei Zhang1,*, Jason C. Raine2, Liqun Xing1, Eric Higley2, Markus Hecker2,3, John P. Giesy1,2,,4,5, Hongxia Yu1,*

1 State Key Laboratory of Pollution Control and Resource Reuse & School of the Environment, Nanjing University, Nanjing, China 2 Toxicology Centre, University of Saskatchewan, Saskatoon, SK S7N 5B3, Canada 3 School of Environment and Sustainability, University of Saskatchewan, Saskatoon, SK, S7N 5B3Canada 4 Department of Biomedical Veterinary Sciences and Toxicology Centre, University of Saskatchewan, Saskatoon, SK S7N 5B3, Canada 5 Department of Biology and Chemistry and State Key Laboratory in Marine Pollution, City University of Hong Kong, 83 Tat Chee Avenue, Kowloon, Hong Kong SAR, China

Corresponding author:

School of the Environment, Nanjing University, Nanjing, 210089, China Tel: 86-25-89680623 Fax: 86-25-83707304 E-mail:

howard50003250@yahoo.com (Xiaowei Zhang) yuhx@nju.edu.cn (Hongxia Yu)

Supporting Table 1: Description of Gene Function Groups

Function Group Description

Enzyme Genes encoded enzymes, which played an important role in biochemical reaction.

Regulatory Protein Genes produced protein after gene expression, which were

transcriptional regulators, activators or inhibitors, especially for DNA process.

Structure Protein Genes produced protein after gene expression, which were

part of organism structure, for example, membrane protein, water MIP channel,

rRNA or tRNA Genes in the “RNA” group didn't produce protein but RNA, which might be rRNA or tRNA.

Stress Responsive Pathway Stress responsive genes.

Amino Acids Formation Genes were related to the formation of amino acids. Function Unclear Genes' functions were still not very clear right now.

Supporting Table 2: Functions of 71 differentially expressed genes

Type function Classification 1 pepE Protein (alpha)-aspartyl dipeptidase Enzyme

2 aroH Protein 3-deoxy-D-arabino-heptulosonate-7-phosphate synthase, tryptophan

repressible Enzyme

3 serB Protein 3-phosphoserine phosphatase Enzyme

4 add Protein adenosine deaminase Enzyme

5 gcvT Protein aminomethyltransferase Enzyme

6 lacZ Protein beta-galactosidase, lac operon Enzyme

7 cysK Protein cysteine synthase A,O-acetylserine sulfhydrolase A subunit Enzyme

8 serA Protein D-3-phosphoglycerate dehydrogenase Enzyme

9 dcd Protein Deoxycytidine triphosphate deaminase Enzyme

10 dmsA Protein dimethyl sulfoxide reductase, chain A Enzyme

11 ligB Protein DNA ligase Enzyme

12 fpr Protein ferredoxin-NADP reductase Enzyme

13 galU Protein glucose-1-phosphate uridylyltransferase Enzyme

14 gadB Protein glutamate decarboxylase B subunit Enzyme

15 metA Protein homoserine transsuccinylase Enzyme

16 icd Protein isocitrate dehydrogenase Enzyme

17 lipA Protein lipoate synthase Enzyme

18 cueO Protein multicopper oxidase with role in copper homeostasis Enzyme

19 aroK Protein shikimate kinase I Enzyme

20 ttdA Protein L-tartrate dehydratase Enzyme

21 cysB Protein Cys regulon transcriptional activator Regulatory Protein

22 cspD Protein DNA replication inhibitor Regulatory Protein

23 ftsK Protein DNA-binding membrane protein required for chromosome resolution and

partitioning Regulatory Protein

24 gadX Protein DNA-binding transcriptional dual regulator Regulatory Protein

25 phoP Protein PhoP transcriptional regulator PhoP transcriptional dual regulator Regulatory Protein

26 mcrA Protein restriction of DNA at 5-methylcytosine residues Regulatory Protein

27 rmf Protein ribosome modulation factor Regulatory Protein

28 rsd Protein regulator of sigma D stationary phase protein, binds sigma 70 RNA

polymerase subunit Regulatory Protein

29 wrbA Protein The purified WrbA protein has NAD(P)H:quinone oxidoreductase activity Regulatory Protein

30 fdhD Protein affects formate dehydrogenase-N Regulatory Protein

31 flgM Protein Negative regulator of flagellin synthesis Regulatory Protein

32 manX Protein PTS system mannose-specific EIIAB component Regulatory Protein

33 cld Protein regulator of length of O-antigen component of lipopolysaccharide chains Regulatory Protein

34 rpsT Protein 30S ribosomal subunit protein S20 Structure Protein

35 rpsU Protein 30S ribosomal subunit protein S21 Structure Protein

36 mglB Protein D-galactose-binding periplasmic protein Structure Protein

37 ytfF Protein inner membrane protein Structure Protein

38 b1403 Protein IS21 protein 2 Structure Protein

39 ompC Protein outer membrane porin protein C Structure Protein

40 ompA Protein outer membrane protein 3a (II*Gd) Structure Protein

41 skp Protein periplasmic molecular chaperone for outer membrane proteins Structure Protein

42 osmC Protein resistance protein, osmotically inducible Structure Protein

43 aqpZ Protein water MIP channel Structure Protein

44 rrnB rRNA

The 5S and 23S rRNAs are the RNA components of the large subunit (50S

subunit) of the E. coli ribosome.There are seven ribosomal RNA (rRNA)

operons, called rrnA, rrnB, rrnC, rrnD, rrnE, rrnG, and rrnH

RNA

45 rrnD rRNA

The 5S and 23S rRNAs are the RNA components of the large subunit (50S

subunit) of the E. coli ribosome.There are seven ribosomal RNA (rRNA)

operons, called rrnA, rrnB, rrnC, rrnD, rrnE, rrnG, and rrnH

RNA

46 rrnC rRNA

The 5S and 23S rRNAs are the RNA components of the large subunit (50S

subunit) of the E. coli ribosome.There are seven ribosomal RNA (rRNA)

operons, called rrnA, rrnB, rrnC, rrnD, rrnE, rrnG, and rrnH

RNA

47 argW tRNA tRNA(argW) is one of seven arginine tRNAs RNA

48 evgA Protein response regulator in two-component regulatory system with EvgS，

involved in acid resistance, osmotic adaption, and drug resistance

Environmental

Stress Response

49 uspB Protein ethanol tolerance protein Component of

Organism

50 uspA Protein universal stress global stress response regulator Environmental

Stress Response

51 menG Protein

S-adenosylmethionine: 2-demethylmenaquinone methyltransferase proteinE.

The interaction of RraA with the degradosome is facilitated by protein-RNA

remodeling via the ATPase activity of RhlB

Amino Acids

Formation

52 yfbV Protein conserved inner membrane protein Function Unclear

53 mntR Protein conserved protein Function Unclear

54 yhgF Protein conserved protein (3rd module) Function Unclear

55 yjdI Protein conserved protein;Uncharacterized protein yjdI Function Unclear

56 yhcG Protein function unknow Function Unclear

57 yafK Protein hypothetical protein Function Unclear

58 insA_7 Protein hypothetical protein Function Unclear

59 yaiE Protein hypothetical protein，UPF0345 family，function unknown Function Unclear

60 b3007 Protein unknown CDS Function Unclear

61 yhhY Protein predicted acetyltransferase Function Unclear

62 yeaU Protein predicted dehydrogenase D-malate dehydrogenase (decarboxylating) Function Unclear

63 glgS Protein predicted glycogen synthesis protein Function Unclear

64 proQ Protein predicted structural transport element Function Unclear

65 yedW Protein putative 2-component transcriptional regulator（yedV） Function Unclear

66 yedP Protein Putative mannosyl-3-phosphoglycerate phosphatase Function Unclear

67 ymcC Protein putative synthetase Function Unclear

68 yfiE Protein putative transcriptional regulator (LysR family) Function Unclear

69 uspF Protein putative universal stress protein F Function Unclear

70 yifE Protein Similar to Yersinia pestis KIM, hypothetical protein y0333 Function Unclear

71 ybcW Protein Uncharacterized protein Function Unclear

Supporting Table 3: References of NOEC for different species exposure to TPTC

a NOEC was calculated following “Guidelines for Deriving Numerical National Water Quality Criteria for the Protection of

Aquatic Organisms and their Uses”, which was drafted by U.S. Environmental Protection Agency.

References: De Vries H, Penninks AH, Snoeij NJ, Seinen W (1991) COMPARATIVE TOXICITY OF ORGANOTIN

COMPOUNDS TO RAINBOW TROUT ONCORHYNCHUS-MYKISS YOLK SAC FRY. Science of the Total Environment 103 (2-3):229-244

Fargasova A (1997) Comparative study of ecotoxicological effect of triorganotin compounds on various

Species Scientific Name

Species Group

Endpoint Duration

(Days) Concentration

(ng/mL)

Calculation of NOEC (ng/mL)a

Mean of (ng/mL)

Ankistrodesmus falcatus ssp.

acicul (Wong et al.

1982)

Algae Moss or Fungi

EC50 8 2.0 4.0×10-1 4.0×10-1

Scenedesmus quadricauda (Fargasova 1997)

Algae Moss or Fungi

EC50 NR 9.1×10-1 1.8×10-1

1.0×102 Scenedesmus quadricauda (Fargasova 2002)

Algae Moss or Fungi

EC50 12 1.1×103 2.3×102

Scenedesmus quadricauda (Fargasova 2002)

Algae Moss or Fungi

EC50 12 3.5×102 7.0×101

Oryzias latipes (De Vries et al. 1991)

Fish LC50 2 6.4×101 1.4 1.4

Oncorhynchus mykiss (Nagase et al.

1991) Fish LOEC 110 2.3×10-1 9.3×10-2 9.3×10-2

Indoplanorbis exustus (Goel and

Prasad 1978) Molluscs LC50 1 6.2×10-4 1.3×10-5

3.7 Indoplanorbis exustus (Goel and

Srivastava 1981) Molluscs LC50 1 3.5×102 7.4

Lymnaea acuminate (Goel and

Srivastava 1981) Molluscs LC50 1 4.3×101 9.1×10-1 9.1×10-1

Tubifex tubifex (Fargasova 1997)

Worms LC50 4 2.4 5.1×10-2 5.1×10-2

biological subjects. Ecotoxicology and Environmental Safety 36 (1):38-42 Fargasova A (2002) Structure-affected algicidal activity of triorganotin compounds. Bulletin of

Environmental Contamination and Toxicology 69 (5):756-762. doi:10.1007/s00128-002-0125-3

Goel HC, Prasad R (1978) Action of Molluscicides on Freshly Laid Eggs of Snail Indoplanorbis-Exustus (Deshayes). Indian Journal of Experimental Biology 16 (5):620-622

Goel HC, Srivastava CP (1981) Laboratory evaluation of some molluscicides against french water snails, Indoplanorbis and Lymnaea species. J Commun Dis 13 (2):121-127

Nagase H, Hamasaki T, Sato T, Kito H, Yoshioka Y, Ose Y (1991) Structure-Activity-Relationships for Organotin Compounds on the Red Killifish Oryzias-Latipes. Applied Organometallic Chemistry 5 (2):91-97

Wong PTS, Chau YK, Kramar O, Bengert GA (1982) Structure-Toxicity Relationship of Tin-Compounds on Algae. Canadian Journal of Fisheries and Aquatic Sciences 39 (3):483-488

Supporting Table 4: Classification of differentially expressed genes and proportion of each

groups basing on the function of genes (see in Supplementary Table 1) and fold change.

Fold Change Genes Number

1.5 - 2.0

cysK, dcd, dmsA, gadB, fpr, gcvT, pepE, metA, serA, ttdA, icd, ligB, lipA, cueO, aroK, fdhD, flgM, ftsK, gadX, manX, rsd, phoP, cysB, cspD, ompA, ompC, osmC, ytfF, rpsT, rpsU, b1403, rrnB, rrnD,

argW, uspA, uspB, evgA, yaiE, yeaU, yedP, yedW, yfbV, proQ, yfiE, yhgF, ymcC, b3007, yafK, mntR

49

2.0 - 3.0 lacZ, serB, galU, aroH, add, wrbA, cld, rmf, mglB, skp, aqpZ, glgS,

insA_7, uspF, yhcG, yhhY, yifE, yjdI 18

3.0 - 4.0 mcrA, menG, ybcW 3 7.0 - 8.0 rrnC 1

Each color represents one toxic pathway induced by TPTC, and detailed description of each toxic pathway can be found in Supplementary Table 1. О “enzyme”; О “Regulatory Protein”; О “Component of Organism”; О “rRNA or tRNA”; О “Environmental Stress Response”; О “Amino Acids Formation”; О “Function Unclear”.

Supporting Figure 1: Gene expression profiles of three potential biomarkers.