Development of a microtiter plate version of the yeast DEL assay amenable to high-throughput toxicity screening of chemical libraries

Available online at www.sciencedirect.com

Mutation Research 634 (2007) 228–234

Short communication

Development of a microtiter plate version of the yeast DEL assay amenable to high-throughput toxicity screening of chemical libraries Nikos Hontzeas a , Kurt Hafer c , Robert H. Schiestl a,b,c,∗ a

Department of Pathology, David Geffen School of Medicine, UCLA, CA 90095, USA Department of Environmental Health, School of Public Health, UCLA, CA 90095, USA Department of Radiation Oncology, David Geffen School of Medicine, UCLA, CA 90095, USA b

c

Received 9 April 2007; received in revised form 23 June 2007; accepted 3 July 2007 Available online 10 July 2007

Abstract The yeast plate-based deletion (DEL) assay has been previously shown to detect a wide range of carcinogens. Of 60 compounds of known carcinogenic activity, 92% were correctly detectable with the DEL assay whereas 62% were correctly detectable with the Ames assay [W.W. Ku, J. Aubrecht, R.J. Mauthe, R.H. Schiestl, A.J. Fornace Jr., Why not start with a single test: a transformational alternative to genotoxicity hazard and risk assessment, Toxicol. Sci. (2007)]. In this manuscript we describe a modification of the yeast DEL assay into a colorimetric assay using the MTS tetrazolium compound (3-(4,5-dimethylthiazol-2-yl)5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt) to allow for efficient detection of chemical genotoxicity. It has been micro-scaled and can be performed in 96- or 384-well format. Chemicals previously characterized with the DEL platebased assay were utilized to test the new well-based format, and a group of cross-linking agents, previously uncharacterized by the DEL assay, were scored for genotoxicity using this new assay format. These compounds induced a range of genotoxicity detectable with the well-based DEL assay, and a lack of sensitivity was found only at extremely low genotoxic levels determined by the plate-based DEL assay. We suggest this new well-based version of the DEL assay can be used as an economical alternative to the plate-based assay to screen large numbers of compounds, such as chemical libraries in a high-throughput screening setting. © 2007 Published by Elsevier B.V. Keywords: DEL assay; High-throughput; MTS

1. Introduction DNA rearrangements including DNA deletions are involved in carcinogenesis [2,3]. An assay screening

Abbreviations: EMS, ethyl methanesulfonate; 4NQO, 4-nitroquinoline-1-oxide; MMS, methylmethane sulfonate; ActD, actinomycin D ∗ Corresponding author at: 650 Charles E. Young Drive South, Los Angeles, CA 90095, USA. Tel.: +1 310 267 2087; fax: +1 310 267 2578. E-mail address: [email protected] (R.H. Schiestl). 1383-5718/$ – see front matter © 2007 Published by Elsevier B.V. doi:10.1016/j.mrgentox.2007.07.001

for DNA deletions in yeast (DEL assay) can detect Salmonella/Ames assay negative as well as positive carcinogens [4–7]. Of 60 compounds of known carcinogenic activity, 92% were correctly detectable with the DEL assay whereas 62% were correctly detectable with the Salmonella-Ames assay [1]. In addition, carcinogens have also been reported to induce DNA deletions in related assays in vitro with human cells [8] and in vivo with mice [9]. The RS112 yeast DEL assay tester strain of Saccharomyces cerevisiae contains a plasmid with an internal fragment of the HIS3 gene integrated at the genomic

N. Hontzeas et al. / Mutation Research 634 (2007) 228–234

HIS3 locus, yielding an integrative disruption of the HIS3 gene [10]. This disruption results in two copies of the HIS3 gene, each copy having one terminal deletion. Recombination between the two his3 deletion alleles results in reversion to HIS3+ and growth in the absence of histidine. This recombination event leads to a 6 kb DNA deletion comprising the integrated plasmid leading to deletion (DEL) events [6]. The traditional assay utilizing DEL events involves overnight growth of a single colony of the RS112 strain and subsequent subculture with the presence or absence of the chemical being tested for 17 h at 30 ◦ C under constant shaking. Yeast cells are then plated onto SC medium to determine the number of survivors (individual colonies are counted) and onto SC−HIS medium to score for DEL events. The traditional DEL assay is very powerful if one is testing a limited number of chemicals but becomes impractical for screening large numbers of chemicals and chemical libraries. In this manuscript we circumvent this problem through the design of a well-based, liquid version of the DEL assay. Yeast growth can be identified in a nonclonogenic quantitative colorimetric assay described by Mosmann [11] which measures the ability of proliferating cells to reduce MTT (3-4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide), a yellow tetrazolium salt, into a purple formazan precipitate. This reaction however, requires quenching and solubilizing the cells in order to measure the formazan precipitate. An improvement upon the MTT assay can be made by substitution with the MTS tetrazolium compound (3-(4,5-dimethylthiazol2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl))2H-tetrazolium, inner salt and an electron-coupling reagent, phenazine ethosulfate (PES). The MTS compound is reduced by cells into a colored formazan product soluble in tissue culture medium. This reaction needs not to be quenched and cell proliferation can be directly measured by recording absorbance at 490 nm. Cell proliferation is proportional to the quantity of formazan product. Thus, the MTS assay is suitable as a colorimetric assay that can be used for high-throughput applications. Here, MTS is used to construct a liquid based version of the yeast DEL assay capable of scoring DNA deletions in both 96- and 384-well plate formats. 2. Materials and methods 2.1. The DEL assay 2.1.1. Media 2.1.1.1. Synthetic complete (SC or + 13) medium. Yeast nitrogen base 0.67%, glucose 2%, agar 2% plus the following amino

229

acids and bases per 600 ml of distilled water: 40 mg each of adenine sulphate, l-isoleucine, l-leucine, l-lysine–HCl, l-tyrosine, 30 mg of l-arginine–HCl, l-histidine–HCl, lmethionine, uracil, 60 mg of l-tryptohpan. For well-based experiments, agar was not added. 2.1.1.2. Drop-out medium. SC medium lacking histidine (SC−HIS). 2.1.1.3. Inoculation (–LEU) medium. Yeast nitrogen base 0.67%, glucose 2%, plus the following amino acids and bases per 600 ml of distilled water were added after autoclaving: 12 mg uracil, 24 mg adenine sulphate, 12 mg l-histidine. 2.1.1.4. Chemicals. The following compounds were purchased from Sigma: actinomycin D (CAS No. 50-76-0) dissolved in 0.2% acetone, ethyl methanesulfonate (CAS No. 62-50-0), camptothecin (CAS No. 7689-03-4), 4nitroquinoline-1-oxide (CAS No. 56-57-5) dissolved in 0.2% DMSO, mitomycin C (CAS No. 50-07-0), CrCl3 (CAS No. 10025-73-7), K2 Cr2 O7 (CAS No. 1333-82-0), benzene (CAS No. 71-43-2), methylmethane sulfonate (CAS No. 66-273), cyclophosphamide monohydrate (CAS No. 6055-19-2), dimethyl sulfoxide (CAS No. 67-68-5) acetone (CAS No. 67-64-1). The following compounds were purchased from VWR: carmustine (CAS No. 154-93-8), chlorambucil (CAS No. 305-03-3) and cisplatin (CAS No. 15663-27-1). Stock solutions of each compound were prepared in water except for 4-nitroquinoline-1-oxide (0.4% acetone), camptothecin (DMSO), and chlorambucil (1:50 HCL–methanol). Acetone (0.1%), DMSO (1%) and HCL–methanol were tested for DEL induction since they were used as solvents. 2.1.2. Yeast strains The diploid S. cerevisiae strain RS112 was used to determine the frequency of DEL recombination: MATa/MAT␣ ura3-52/ura3-52 leu2-3,112/leu2-98 trp527/TRP5 arg4-3/ARG4 ade2-40/ade2-101 ilv1-92/ILV1 HIS3::pRS6/his3200 LYS2/lys2-801. 2.1.2.1. The recombination assay. The DEL recombination assay was adapted for microwell format from the previously described methodology [12]. For 384-well plate format, 1 ␮l of yeast (∼100,000 cells) was pipetted into eight microplate wells for each compound, four of which containing 70 ␮l SC medium and four containing 70 ␮l SC−HIS medium; each well was supplemented with 14 ␮l of MTS and 5 ␮l of compound. For 96-well plate format, the above was consistent except media, MTS, and compound volumes were 100, 20, and 7 ␮l, respectively. Control wells were treated with water in lieu of compound. The outermost columns of the 96-well plate and outer two columns of each 384-well plate were excluded from experimentation lest edge evaporative effects alter the data. Plates were incubated at 30 ◦ C at normal atmosphere during which yeast were grown in the presence of the tested compound and 490 nm absorbance was measured 10–18 h later

230

N. Hontzeas et al. / Mutation Research 634 (2007) 228–234

Fig. 1. The DEL assay was simulated by adding different dilutions of RS112 His+ revertant yeast to 100,000 background RS112 cells in −His media. 12, 18, and 24 h time points are charted. At 12 h, RS112 His+ additions corresponding to 250, 500, and 1000 DEL events were discernibly significant from 100,000 background RS112 cells. By both 18 and 24 h, as few as 25 DEL events per 100,000 cells were significantly detectable. Yet at 24 h, whilst 25–1000 RS112 cells are still significantly different than background, growth in −His media becomes saturated and response pattern is lost. The experiment was carried out using at least six repeats for each treatment group, and the results are presented as means ± S.D. Significance (* p < 0.05).

using a Molecular Devices SpectraMax M5 microplate reader (Sunnyvale, CA). 2.1.2.2. Sensitivity measurement. A mock experiment was set-up to measure the sensitivity of the well-based DEL assay. On a 96-well plate, wells containing 100,000 RS112 yeast cells in 100 ␮l SC−HIS medium were supplemented in six-plicate with 5 ␮l of RS112 His+ revertant cells ranging from 0 to 1000 cells. Twenty microliters MTS was added to each well, plates were incubated at 30 ◦ C, and 490 nm absorbance was read hourly between 12 and 24 h using a Molecular Devices SpectraMax M5 microplate reader (Sunnyvale). 2.1.3. Data analysis Proliferation for a given compound treatment was assessed by averaging the absorbance across the four treated wells in SC medium and dividing that by the average absorbance of control wells. DEL induction was measured by dividing the absorbance of each SC−HIS well by the corresponding paired SC well. Thus, for each compound done in quadruplicate, each plate contains a set of four different measurements of DEL induction. For each compound, the ratio of average DEL induction in treated versus control cultures, was taken as the fold-increase in HIS+ growth. A Student’s t-test was performed on the four measurements of DEL induction and the same measurements recorded from control wells to determine significance of the fold-increase. It should be noted that this value of

fold-increase is analyzed statistically for significance and not absolute; thus two different chemicals may have the same foldincrease but varying levels of significance. In plots and tables, data from a single experiment performed in quadruplicate is presented; all experiments were repeated independently in at least three separate experiments in both 96- and 384-well plate formats.

3. Results and discussion In the present study the DEL assay has been modified to allow rapid determination of DEL recombination effects. It has been micro-scaled to 96- or 384-well format using the colorimetric agent MTS. To determine experimental sensitivity a mock experiment was performed which simulated the DEL induction using 96-well plates. The assay was most sensitive at 18 h post incubation at which time as few as 25 RS112 His+ revertant cells could be significantly differentiated from spontaneous background levels in 100,000 RS112 cells. 250 RS112 His+ revertant cells were significantly detected 12 h after dispensing into microwells (Fig. 1) suggesting that strong inducers of DEL recombination which induce 250 or more DEL events/100,000 cells are rapidly detectable using the well-based DEL assay.

Fig. 2. (a–c) Well-based DEL assay evaluating the genotoxicity of 13 carcinogens by measuring o.d. (490 nm) 14 h after the addition of MTS. Genotoxicity is qualified by increased growth in −HIS medium relative to growth in +13 medium compared to untreated yeast. In panels A and B, are compounds previously tested using the DEL assay; panel C, cross-linking agents previously uncharacterized by DEL assay. Experiments performed in panels A, B, and C were, respectively, done on a single 384-well plate; four repeats for each treatment group were used and the results are presented as means ± S.D. The experiment was repeated at least three times in both 96- and 384-well plate formats each time yielding similar results.

N. Hontzeas et al. / Mutation Research 634 (2007) 228–234

231

232

N. Hontzeas et al. / Mutation Research 634 (2007) 228–234

To validate the well-based DEL assay, nine chemicals previously characterized by the plate-based DEL assay were used in both 96- and 384-well plate formats. In as little as 10–12 h after the addition of MTS all of the high and moderate genotoxic treatments (comparable to >250 DEL events induced per 100,000 cells as observed by the plate-based assay) were readily distinguishable from controls. For most compounds tested, the concentration range was analogous to those ranges reported in previous studies with the plate-based DEL assay, yet for MMS the concentration range was decreased to avoid elevated cytotoxicity. For the concentrations used here, these compounds showed varied levels of induction of DEL recombination in yeast [6]. In this way, the sensitivity of the well-based assay can be validated using data previously acquired using the plate-based assay. In addition, a group of cross-linking agents previously uncharacterized by the DEL assay were tested for genotoxicity induction. The growth measurement of 13 different chemicals including chemicals tested previously by the well-based DEL assay is plotted in Fig. 2a and b, and the calculated genotoxicity is presented in Table 1. Yeast cells exposed to 2, 5 and 10 ␮g/ml EMS showed diminished growth in +13 medium and increased growth in −HIS medium. The effect of EMS on cell proliferation is dose-dependent as greater proliferation was observed in samples treated with 2 ␮g/ml EMS, while those treated with 10 ␮g/ml produced the lowest survival. Although for each EMS treatment, the growth in −HIS medium was significantly greater than that observed in untreated controls, no quantitative genotoxicity relationship with dose was detectable unlike seen in the plate-based assay [5,6]. 4-Nitroquinoline-1-oxide (4NQO) induced significant growth in −HIS medium even at concentrations as low as 0.08 ␮g/ml (Table 1) and all concentrations tested between 0.08 and 1.0 ␮g/ml diminished growth in +13 medium (Fig. 2a). Camptothecin induced significant genotoxicity at 1.7–13.9 ␮g/ml indicated by increased growth in −HIS medium relative to growth in +13 medium at each of the concentrations tested. Solvents DMSO and acetone, used to dissolve camptothecin and 4NQO, respectively, were scored for DNA deletion potential; neither solvent generated any different response from control treated yeast cells (Table 1). Methylmethane sulfonate caused a significant increase in growth in −HIS medium even at the lowest dose tested of 0.0011 ␮g/ml, and genotoxicity was detected in each treatment up to 0.11 ␮g/ml. Twenty-five micrograms per millilitre actinomycin D also caused significant growth in −HIS medium.

Table 1 Fold-increase in DEL induction Compound

DEL (foldincrease)a

DEL significanceb

EMSc (2 ␮g/ml) EMS (5 ␮g/ml) EMS (10 ␮g/ml) 4NQOc (0.08 ␮g/ml) 4NQO (0.2 ␮g/ml) 4NQO (1 ␮g/ml) MMSc (0.011 ␮g/ml) MMS (0.011 ␮g/ml) MMS (0.055 ␮g/ml) MMS (0.11 ␮g/ml) Camptothecin (1.74 ␮g/ml) Camptothecin (5.23 ␮g/ml) Camptothecin (8.71 ␮g/ml) Camptothecin (13.93 ␮g/ml) ActDc (6.28 ␮g/ml) ActD (12.55 ␮g/ml) ActD (25.11 ␮g/ml) Cr (3) (31.67 ␮g/ml) Cr (3) (110.85 ␮g/ml) Cr (3) (221.7 ␮g/ml) Cr (6) (19.99 ␮g/ml) Cr (6) (69.99 ␮g/ml) Cr (6) (139.99 ␮g/ml) Benzene (250 ␮g/ml) Benzene (400 ␮g/ml) Cyclophosphamide (13.96 ␮g/ml) Cyclophosphamide (27.91 ␮g/ml) Cyclophosphamide (55.82 ␮g/ml) Mitomycin C (1.67 ␮g/ml) Mitomycin C (3.34 ␮g/ml) Mitomycin C (6.69 ␮g/ml) Mitomycin C (13.37 ␮g/ml) Chlorambucil (0.304 ␮g/ml) Chlorambucil (3.04 ␮g/ml) Chlorambucil (9.13 ␮g/ml) Carmustine (2.14 ␮g/ml) Carmustine (6.42 ␮g/ml) Carmustine (10.70 ␮g/ml) Cisplatin (27 ␮g/ml) Cisplatin (300 ␮g/ml) DMSO 1%c Acetone 0.4%c HCL–methanol (1:50) 0.5%c

1.59 3.19 2.65 3.39 3.05 1.78 1.54 1.68 2.16 3.09 1.28 1.53 1.57 1.77 0.89 0.96 1.26 0.93 0.98 1.44 2.13 3.78 3.12 1.77 2.31 0.77 0.74 0.95 1.07 1.07 1.19 1.29 1.18 2.37 3.58 1.18 3.35 7.62 1.92 4.57 0.96 0.96 0.95

*** *** *** *** *** *** ** *** *** *** ** *** *** ***

ns ns **

ns ns *** *** *** *** *** ***

ns ns ns ns ns * ** * *** *** * *** *** *** ***

ns ns ns

a Fold DEL increase was calculated by dividing the DEL induction measured for the respective compound concentration by that of the controls performed on the same plate. The concentration reported is the final concentration of treatment in each microwell. Each experiment was repeated at least three times on separate plates and similar results were attained in each measurement. b Significance * (p < 0.05), ** (p < 0.01), *** (p < 0.005); ns-not significant (p > 0.05). c These are controls for solvents used to dissolve compounds as described in materials and methods.

N. Hontzeas et al. / Mutation Research 634 (2007) 228–234

Both chromium(III) (CrCl3 ) and chromium(VI) (K2 Cr2 O7 ) induced DNA deletions in previous experiments [13] and were also tested for DEL induction in the well-based format (Table 1 and Fig. 2b). Chromium(III) did not induce any genotoxic events at concentrations of 32 or 111 ␮g/ml, yet a significant increase in growth on −HIS medium was detected at 222 ␮g/ml. Chromium(VI) showed a very potent induction with the DEL assay and severely decreased survival with increased dose between 20 and 140 ␮g/ml. At benzene concentrations of 250 and 400 ␮g/ml significant DEL induction was observed. The DEL assay previously was reported to detect cyclophosphamide genotoxicity by the plate-based assay at the lowest significance p < 0.05, without the addition of the metabolic activator S9 but showed a much more powerful induction in the presence of S9 [12]. The HTS assay was unable to detect any significant HIS+ growth for cyclophosphamide since the HTS assay is not as sensitive as the plate-based assay at very low genotoxic levels and S9 was not used in this experiment. The chemotherapeutic agent mitomycin C, a crosslinking agent previously uncharacterized by the yeast DEL assay, caused significant increases in HIS+ growth in yeast at concentrations of 6.7 ␮g/ml and above whereas diminished growth in complete +13 medium was only significantly observed at 13.4 ␮g/ml (Fig. 2c and Table 1). Chlorambucil, also a drug used for chemotherapy, caused significant genotoxicity at concentrations as low as 0.3 ␮g/ml, while carmustine caused more pronounced genotoxic effects at concentrations above 2.1 ␮g/ml. HCL–methanol, used as a solvent for chlorambucil, by itself induced no DEL events compared to control treated yeast (Table 1). Cisplatin, another widely used chemotherapeutic agent, caused significant genotoxicity at concentrations 27 and 300 ␮g/ml. A comparison can be made between the sensitivity of the well-based DEL assay and the traditional platebased assay. With chromium(III), the lowest dose which induced significant genotoxicity with the traditional agar plate version was 111 ␮g/ml [13] whereas in the wellbased format the lowest detectable concentration was 222 ␮g/ml. With the plate-based assay it was previously observed that 222 ␮g/ml Cr(III) corresponded to 33.6 DEL events per 100,000 cells [13]. If one considers that the background as tested in the present study is about 20 DEL events per 100,000 cells, this is a small increase of approximately 13.6 cells above background. The fact that this genotoxicity level was significantly detectable here (Table 1) corresponds well with the mock experiment done to identify the sensitivity of the well-based DEL assay (Fig. 1) in which the sensitivity was found

233

to be somewhere between 12.5 and 25.0 DEL events per 100,000 cells. One should be mindful when using the DEL assay to detect extremely low genotoxic treatments which induce DEL events faintly above background. The plate-based assay being slightly more sensitive is better suited for detecting the genotoxicity of such low-potency exposures, whereas the well-based assay is better suited for rapidly assessing genotoxicity of many treatments simultaneously. When, evaluating the well-based assay and comparing it to the plate-based assay, there are both benefits and drawbacks. The well-based assay is economically superior to the plate-based assay and substantially less labor intensive. To perform the plate-based assay, 4–5 days are needed to perform the entire assay and score colonies. In the well-based assay, significant results can be collected in as few as 10–12 h, and moreover there is no requirement to count and score colonies. Unlike the plate-based assay, the amount of DEL events observed cannot be quantified in the well-based assay using an absolute number and results are not quantitative. Rather, the wellbased assay is intended as an easy method to determine the binary presence or lack of genotoxicity. For example, EMS was previously observed by the plate-based assay to potently induce DEL events which correlated with the concentration of EMS used in exposure [5,6]. Although here in the well-based assay a decrease in yeast proliferation in complete medium was observed to correlate quantitatively with dose, no such correlation of growth in −HIS medium was observable with EMS dosage (Fig. 2a and Table 1) although the HIS+ growth increase was significant at each dosage. Furthermore, exposing yeast cells to extremely high cytotoxic doses of nongenotoxic compounds have the potential to yield a false-positive report of genotoxicity using the well-based assay. Such high cytotoxic treatments can cause so much cell killing that the absorbance measured in both +13 and −HIS wells is reduced near to background levels; thus when the ratio of growth in −HIS to +13 is taken, it nears unity. In such an instance, the well-based DEL assay would accurately report cytotoxicity yet fail to accurately report the induction of genotoxicity. One must be mindful of this when interpreting the data of the well-based DEL assay in order to get an accurate assessment of the genotoxic potential for a given compound. This well-based version of the DEL assay is amenable to multiple formats. For example, yeast cells could be treated in 5 ml liquid cultures for 17 h (as done for the plate-based assay) and afterwards scored in +13 and −HIS liquid media using MTS. When this format was used, a similar qualification of genotoxicity was

234

N. Hontzeas et al. / Mutation Research 634 (2007) 228–234

measured for each of the compounds although concentrations used to expose yeast cells were generally higher than those reported in Table 1. The well-based assay is also adaptable for high-throughput screening. Generally high-throughput screens use one compound per well, yet however in order to utilize the DEL assay, a minimum of two wells is required per compound as one is used for growth in +13 and another in −HIS medium, yet more wells should be used to increase measurement sensitivity. Also, the toxicity of many compounds is only discernable within a specific dosage range. Thus, if a high-throughput screen is performed at a single concentration for each compound, the genotoxicity of some compounds may be overlooked. Furthermore, any cytotoxicity observed might not necessarily be due to the toxicity of a chemical, but could be the result of MTS interference or interaction with the chemical compound. A potential artefact may be due to interaction of the compound with the MTS, leading to apparent toxicity. However, since both the −HIS as well as the complete medium has the same concentration of MTS both endpoints would be affected the same way and even so there may be an artefactual effect on toxicity. There should be no such artefactual effect on genotoxicity. To make sure of the effect of any particular compound on DEL recombination one should still rely on the regular plating assay counting distinct colonies. In summary, the DEL assay has been micro-scaled for use in a 96- or 384-well format, adept for highthroughput screen-based assays. This format is sensitive enough to detect at least as few as 25 DEL events per 100,000 cells and was used to assess the genotoxicity of 13 different compounds tested at various concentrations. Cross-linking agents previously uncharacterized with the DEL assay were strong inducers of DNA deletions using this assay. The well-based DEL assay described here is ergonomically superior and can report genotoxicity much more rapidly than the traditional plate-based assay. We submit that this well-based DEL assay is well suited for rapidly qualifying the genotoxicity of a large

number of compounds and is amenable to automation in its current format for high-throughput purposes. Acknowledgement This research was supported by project 1 to RHS of NIH grant 1 U19 AI 67769-01 to William McBride. References [1] W.W. Ku, J. Aubrecht, R.J. Mauthe, R.H. Schiestl, A.J. Fornace Jr., Why not start with a single test: a transformational alternative to genotoxicity hazard and risk assessment, Toxicol. Sci. 99 (2007) 20–25. [2] A.J. Bishop, R.H. Schiestl, Homologous recombination and its role in carcinogenesis, J. Biomed. Biotechnol. 2 (2002) 75–85. [3] A.J. Bishop, R.H. Schiestl, Role of homologous recombination in carcinogenesis, Exp. Mol. Pathol. 74 (2003) 94–105. [4] A.J. Bishop, R.H. Schiestl, Homologous recombination as a mechanism for genome rearrangements: environmental and genetic effects, Hum. Mol. Genet. 9 (2000) 2334–2427. [5] R.H. Schiestl, Nonmutagenic carcinogens induce intrachromosomal recombination in yeast, Nature 337 (1989) 285–288. [6] R.H. Schiestl, R.D. Gietz, R.D. Mehta, P.J. Hastings, Carcinogens induce intrachromosomal recombination in yeast, Carcinogenesis 10 (1989) 1445–1455. [7] R.J. Brennan, R.H. Schiestl, Detecting carcinogens with the yeast DEL assay, Methods Mol. Biol. 262 (2004) 111–124. [8] J. Aubrecht, R. Rugo, R.H. Schiestl, Carcinogens induce intrachromosomal recombination in human cells, Carcinogenesis 16 (1995) 2841–2846. [9] R.H. Schiestl, J. Aubrecht, F. Khogali, N. Carls, Carcinogens induce reversion of the mouse pink-eyed unstable mutation, Proc. Natl. Acad. Sci. U.S.A. 94 (1997) 4576–4581. [10] R.H. Schiestl, S. Igarashi, P.J. Hastings, Analysis of the mechanism for reversion of a disrupted gene, Genetics 119 (1988) 237–247. [11] T. Mosmann, Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays, J. Immunol. Methods 65 (1983) 55–63. [12] Z. Kirpnick, M. Homiski, E. Rubitski, M. Repnevskaya, N. Howlett, J. Aubrecht, R.H. Schiestl, Yeast DEL assay detects clastogens, Mutat. Res. 582 (2005) 116–134. [13] Z. Kirpnick-Sobol, R. Reliene, R.H. Schiestl, Carcinogenic Cr(VI) and the nutritional supplement Cr(III) induce DNA deletions in yeast and mice, Cancer Res. 66 (2006) 3480–3484.