Correspondence: Hani A. Alhadrami, Faculty of Applied Medical Sciences, King Abdulaziz University, P.O. Box 80324, Jeddah 21589, Saudi Arabia. Tel.: +966 5055 45275; fax: +966 2 6400000 Ext. 20171; e-mail: email@example.com
The environmental fate and potency of mutagenic compounds is of growing concern. This has necessitated the development and application of rapid assays to screen large numbers of samples for their genotoxic and carcinogenic effects. Despite the development of biosensors for genotoxicity assessment, these have not been calibrated against traditional microbial bioassays. In this study, assays using the SOS-lux-marked microbial biosensors Escherichia coli K12C600 and E. coli DPD1718 were refined and optimised to screen selected mutagenic chemicals. The response of the biosensors was compared with the mutagenic response of the traditional Salmonella mutagenicity assay. For the chemicals tested (acridine, B[a]A, B[a]P, chrysene, mitomycin C and sodium azide), E. coli DPD1718 was consistently more sensitive than E. coli K12C600. The biosensors were of comparable sensitivity to the Salmonella assay but were more rapid, reproducible and easier to measure. These data validate the adoption of optimised assays making use of microbial biosensors for routine screening of test chemicals.
Chemical analysis alone is a poor strategy for assessing the mutagenicity of compounds found in the environment. Such compounds, once activated, increase the rate of mutation in cells. Chemical analysis may identify the presence and quantity of such compounds but does not reflect the potency to biological receptors. To effectively measure mutagenicity, there is a requirement for assays with a clear end points and capability of screening large numbers of samples (pure compounds, mixtures and environmental samples).
Luminescence-based bacterial biosensors can detect the mutagenic mode of action of certain chemicals (Gu & Chang, 2001) and are simple to apply, sensitive and easy to measure (Van Der Meer & Belkin, 2010). The SOS-lux-based microbial biosensors have a promoterless lux-operon (luxCDABE) under control of the SOS-dependent col promoter (Rettberg et al., 1999). When a mutagenic mode of action is activated, this leads to an increase in the concentration of luciferase and, as a consequence, bioluminescence.
The Salmonella assay (Ames assay) is the most widely used bacterial assay for screening potential mutagenic compounds (Pereira et al., 2010). There is a correlation between mutagenicity as measured by the Ames assay and carcinogenicity in mammals (Josephy et al., 1997). The assay uses a number of Salmonella strains with specific mutations that disable the cells from synthesising histidine. New mutations at the site of these pre-existing mutations can restore the gene's function and allow the cells to resynthesise histidine. Enumeration of a cultured lawn of these mutated colonies in the absence of histidine enables a quantitative assessment of mutagenicity (Mortelmans & Zeiger, 2000). The assay is simple to perform but requires high replication, a wide range of controls, extensive culturing and time-consuming enumeration.
Polycyclic aromatic hydrocarbons (PAHs) are classified by the European Union and U.S. Environmental Protection Agency as priority environmental pollutants due to their mutagenic impact on humans (Song et al., 2009). Certain PAHs are biologically inactive and require liver metabolism before activation to a mutagenic form (Shimada & Fujii-Kuriyama, 2004). Most bacteria are unable to metabolise such chemicals via cytochrome P450; therefore, the assay requires the exogenous addition of the mammalian metabolite (Maron & Ames, 1983). This metabolite is typically delivered to the assay in the presence of NADP, glucose-6-phosphate and phosphate buffer (S9 mix).
Polycyclic aromatic hydrocarbons can cause several types of mutations such as frameshift mutations, base-pair-substitution mutation and transition mutation (Pérez et al., 2003). Frameshift mutations are a consequence of the addition or deletion of one or two consecutive base pairs in the DNA sequence of a gene, resulting in misreading mRNA and synthesis of a nonfunctional protein (Sadava et al., 2006). Base-pair-substitution mutation is a result of the abnormal replacement of a single-base nucleotide with another nucleotide of the genetic materials (Mortelmans & Zeiger, 2000). Transition mutation is when a replacement of purine base (i.e. adenine and guanine) with another purine or a replacement of a pyrimidine base (i.e. thymine, cytosine and uracil) with another pyrimidine occurs (Sadava et al., 2006). The Salmonella isolates can be modified to suit specific receptors.
In this study, the performance of the SOS-lux-based microbial biosensors for the screening of mutagenic chemicals was compared with the traditional Salmonella assay (the Ames assay), and a critical comparative evaluation of assays sensitivity and performance was made.
Materials and methods
Mutagenicity testing was performed against a range of doses of acridine, benzo (a) anthracene (B[a]A), benzo (a) pyrene (B[a]P), chrysene, mitomycin C (MMC) and sodium azide. All chemicals were purchased from Sigma (St. Louis, MO). Standards (except MMC) were dissolved in 100% v/v spectrophotometric grade DMSO. MMC was dissolved in MilliQ water. Dilutions were performed to obtain appropriate assay concentrations, and these were tested against appropriate controls (in equivalent to DMSO concentration and MilliQ water). The selection of the chemicals doses was based upon data published by McCann et al. (1975) and Madill et al. (1999).
Mutagenicity assessment using the Salmonella mutagenicity assay
The standard plate incorporation procedure described by Maron & Ames (1983) was used for the Ames assay. Salmonella typhimurium TA98, TA100 and TA102 were obtained from Molecular Toxicology Inc. (Boone, NC) and maintained according to standard protocols (Mortelmans & Zeiger, 2000). Each sample was tested in triplicate in the absence and the presence of the mammalian liver homogenate (S9 mix derived from Aroclor 1254-exposed rats and obtained from Molecular Toxicology Inc.). In brief, Salmonella strains TA98, TA100 and TA102 were grown overnight in 150-mL Erlenmeyer flask containing 25 mL Oxoid nutrient broth at 37 °C in an orbital shaker at 150 r.p.m. with appropriate antibiotics (25 μg mL−1 ampicillin for TA98 and TA100, and 2 μg mL−1 tetracycline for TA102). The cultures were incubated until they reached an absorbance of 1.0 at 660 nm (corresponding to 1–2 × 109 CFU mL−1). Two millilitre of melted top agar supplemented with histidine and biotin solution was distributed into sterile glass tubes and placed in a 45 °C water bath. For the mutagenicity assay conducted without S9 mix, 100 μL of the test chemical and 100 μL of the tester strain were added, gently mixed by vortexing and poured onto the surface of minimal glucose agar plate. The plates were gently tilted and rotated to obtain an even distribution, placed onto a level surface to solidify and incubated at 37 °C for 48 h. Following the incubation, the revertant colonies were enumerated on a Gallenkamp colony counter. To conduct the assay with the S9 mix, 500 μL of the previously prepared S9 mix (Maron & Ames, 1983) was added to the top agar along with the test chemical and the tester strain. Appropriate reagent and negative controls (Maron & Ames, 1983) were included to enumerate the spontaneous revertants.
Mutagenicity assessment using the SOS-lux microbial biosensors
An aliquot (25 mL) of biosensor strains Escherichia coli K12C600 and E. coli DPD1718 was grown overnight on LB media at 37 °C on an orbital shaker at 150 r.p.m. in the presence of the appropriate antibiotics (50 μg mL−1 ampicillin for E. coli K12C600 and 30 μg mL−1 chloramphenicol for E. coli DPD1718) (Rettberg et al., 1999; Vankemmelbeke et al., 2005). Overnight cultures were diluted 1:50 in LB broth and grown at 37 °C until they reached their preoptimised CFU (1–2 × 109 CFU mL−1). A negative control of 100 μL MilliQ water, a DMSO reagent control or 100 μL tested chemical were mixed with 900 μL of overnight culture in 3mL luminometer cuvettes. Bioluminescence was measured using a Jade bench-top luminometer (Labtech International, Uckfield, UK), and each sample was tested in triplicate. To determine whether mutagenicity occurred, the test sample was incubated for a 300min period, and the maximum induction (relative to controls) that occurred between 180 and 300 min was recorded.
Statistical analysis was performed using Minitab 15 for Windows. Normality testing and equal variances were carried out to assess normality. Analysis of variance (anova) and t-test were applied to assess the differences between the bioluminescence readings of the biosensors. Correlation tests (Pearson product-moment or Spearman rank-order) assessed the correlation between the Ames assay and the biosensors. A result with P ≤0.05 was considered significant. The ‘twofold increase rule’ was applied to the Salmonella assay data to define whether a test compound was mutagenic (Maron & Ames, 1983). For the biosensors, a compound was considered a mutagen if there was a significant difference between the bioluminescence readings of the negative control and the test compound during the incubation period. If the bioluminescence values decreased during the incubation time, the compound was considered more likely to be a cytotoxic (Rettberg et al., 1999).
The lowest observed adverse effective concentration (LOAEC) was the dose at which there was a significant (P ≤0.05) measure of mutagenicity relative to the control values for both of the assays.
Results and discussion
Mutagenicity response using the Salmonella mutagenicity assay
A significant mutagenic response of Salmonella strain TA98 was recorded in response to B[a]A tested at 500 μg mL−1 in the absence of S9 mix (Table 1). In contrast to this, B[a]P and chrysene only had a significant mutagenic impact on the Salmonella strains tested in the presence of S9 mix (Table 1). When TA100 was exposed to acridine at concentrations of 3 μg mL−1, 5 μg mL−1 and 10 μg mL−1, a significant mutagenic response was observed (Table 1). A significant mutagenicity was reported for sodium azide tested with TA98 and TA100 (at 3 μg mL−1, 5 μg mL−1 and 10 μg mL−1) and TA102 (at 10 μg mL−1) (Table 1). This remark substantiated that sodium azide is as direct-acting potent mutagen when tested with the Salmonella assay, which correlates with the observation of Nilan et al. (1973) who reported that sodium azide was a base-pair-substitution mutagen.
Table 1. The number of reverse mutants of Salmonella typhimurium TA98, TA100 and TA102 by B[a]A, B[a]P, chrysene, sodium azide and acridine at various doses tested with and without S9 mix
Number of histidine revertants per plate: mean values of at least three plates ± standard error (SE).
Concentration based on 100 × 15-mm Petri dish containing 20–25 mL of MG agar.
Numbers in italic, boldface and underlined represent the number of spontaneous revertants colonies for each strain. The spontaneous revertant ranges (revertants/plate without S9 mix) were as follows: TA98 (13-77), TA100 (67-173) and TA102 (117-530).
Numbers in boldface represent twofold increase or more in the number of revertant colonies over the solvent controls (spontaneous revertants), which was an indication of a significant mutagenic response.
NT: compound was not tested against the strain(s), B[a]P was recommended to be tested only with TA98 and TA100 (McCann et al., 1975). Acridine and sodium azide were recommended to be assayed in the absence of the S9 mix (Madill et al., 1999).
For the Salmonella strains to be responsive to B[a]A, B[a]P and chrysene, the exogenous addition of S9 mix was required (Table 1). Significant dose-dependent mutagenic effects occurred for B[a]P when exposed to Salmonella strains TA98 and TA100 (Table 1). This confirmed that B[a]P caused a base-pair-substitution mutation as it reverted TA100, and a frameshift mutation as it reverted TA98. These observations were in agreement with data published by Unger & Guttenplan (1980). The exogenous addition of S9 is used because B[a]P requires activation by cytochrome P450 (CYPs) enzymes to form bay-region epoxides through two activation steps: The first step would involve the formation of B[a]P-7,8-oxide by the action of CYP. The second step would be the formation of B[a]P-7,8-diol by microsomal epoxide hydrolase (Gelboin, 1980). The recombinant human CYP1B1 is acknowledged to be more active than CYP1A1 in catalysing oxidation of B[a]P to B[a]P-7,8-diol with human liver epoxide hydrolase being present in the reaction mixture. Data previously published (Shimada & Fujii-Kuriyama, 2004) confirmed that CYP1B1 was essential for the first-step oxidation of B[a]P at the 7,8-position; hence, it played a more critical role than CYP1A1 in the tumorigenesis caused by B[a]P.
Salmonella strains TA98 and TA102 were responsive to B[a]A in the presence of the S9 mix (Table 1). It has been acknowledged that B[a]A causes frameshift mutations and transition mutation (McCann et al., 1975).
Chrysene had a significant mutagenic effect on TA98 at all tested doses except 5 μg mL−1 (Table 1). Chrysene at concentration of 50 μg mL−1 was mutagenic to Salmonella strain TA100 (Table 1). Chrysene is a causative agent of frameshift mutations, and this was approved by the reversion of TA98 only in the presence of the S9 mix. Similar results were reported by Malachova (1999).
Acridine, chrysene, B[a]A and B[a]P caused a mutagenic response to the Ames assay at concentrations between 20 and 200 μg mL−1 (Madill et al., 1999). Previous data (McCann et al., 1975; Pérez et al., 2003) reported the mutagenic effects of these compounds in the presence of S9 mix.
There was a degree of inconsistency in the dose response of the Ames assay, but using the collated data, the LOAEC was predicted (Table 2). Fundamentally, there may not be a clear dose response such as that observed for chronic and acute exposure assays, because once the critical mutagenic dose is reached, the effect may remain constant. At elevated doses, a cytotoxic effect may occur, and this could inhibit or indeed cause confusion in the detection of mutagenicity. In addition to problems in interpreting the data, the actual assay could cause artefacts because the concentration of hydrophobic compounds present in the agar may not correlate with the causal dose because the agar could disrupt the bioavailability of the test sample. For future assays, a measure of the diffusion of the test chemical may be a better indicator of the exposure dose rather than the amendment concentration added.
Table 2. Lowest observed adverse effective concentration values (μgmL−1) for each of the chemicals tested for individual assays (n = 3)
Salmonella mutagenicity strains
E. coli K12C600
E. coli DPD1718
The values represent the concentrations at which there was a significant difference from the reagent control (P ≤0.05). For the Ames assay, each of the assays was conducted using S9 mix except for this preceded with *. Values with the same letter following the value are not significantly different from one another (P ≤0.05). NR designates no mutagenic response at the doses tested.
Mutagenicity response using the SOS-lux biosensors E. coli K12C600 and E. coli DPD1718
Mitomycin C caused a significant induction of the luminescence of the SOS-lux-marked biosensor E. coli K12C600 and was considered a mutagen. This was adopted as a positive control (Fig. 1). The responses of the biosensor to acridine tested at 1 and 3 μg mL−1 and to 5 μg mL−1 chrysene were significantly different (P =0.01 and 0.02, respectively) to the reagent control (MilliQ water) defining these as having mutagenic modes of action (Fig. 1). There was a significant mutagenic response (P =0.02) to B[a]A tested at 50 μg mL−1 and to sodium azide (P =0.01) tested at 0.5, 1, 1.5, 3 and 5 μg mL−1 (Fig. 1). There was no response to B[a]P in the presence or the absence of S9 mix (data not shown).
Mitomycin C, acridine and chrysene caused significant induction of E. coli DPD1718 (Fig. 2) defining these chemicals as mutagenic. B[a]P tested at 1 μg mL−1 in the presence of S9 mix also caused a significant increase in luminescence. These findings confirmed that E. coli DPD1718 responded to the DNA cross-link mode of action of MMC, acridine and chrysene (Fig. 2).
The biosensor responses reflected those of the Ames assay but were less responsive to B[a]P. This failure to respond to B[a]P and other such chemicals (that are acknowledged to be mutagenic) could be on account of several reasons central to the operational activities of these biosensors and the assay procedures adopted. B[a]P and similar molecules may be incapable of inducing the SOS-DNA repair system of E. coli K12C600. The SOS responses are stimulated by disrupting and arresting the DNA, which in turn inhibits the cell division and leads to the accumulation of single-stranded DNA (ssDNA) (Janion, 2008). Chemicals that are direct inducers of the SOS system such as MMC and methyl methane sulphonate (MMS) will be more mutagenic to the biosensor.
The biosensor E. coli K12C600, unlike the Salmonella mutagenicity strains, does not carry rfa (increase of cell wall permeability) and uvrB (the excision repair system for DNA) mutations. Therefore, the cell wall of E. coli K12C600 is less permeable than the cells in the Ames assay resulting in reduced transportation of hydrophobic molecules. Indeed, this was the least sensitive of the assays tested. As with the Ames assay, the bioavailability of hydrophobic molecules within the assay procedure could be a key constraint on its performance.
Rettberg et al. (2001) used the pPLS1 plasmid in E. coli K12C600 and constructed another SOS-lux biosensor by introducing pPLS1 plasmid into S. typhimurium TA1535 (which carries rfa mutation). The resultant S. typhimurium TA1535 had greater sensitivity for the detection of hydrophobic compounds when compared with E. coli K12C600 (Rettberg et al., 2001).
These biosensors in which mutagenic compounds damage DNA and activate the reporter genes fused with the SOS repairing genes may be prone to the occurrence of false-negative responses when inhibition of metabolism is a consequence of cytotoxicity (Sørensen et al., 2006). To address this, the user must consider adopting a wide selection of chemical doses to ‘range-find’ the responses and differentiate cytotoxicity from mutagenicity. In this study, LOAEC values were calculated to drive a more relative comparative value of assay sensitivity (Table 2).
Comparative evaluation of the SOS-lux biosensors and the Ames assay
The SOS-lux biosensors were adequate for the detection of the mutagenicity of DNA-damaging substances at the range of concentrations associated with the Ames assay. An example is given for B[a]P: in the Ames assay, the LOAEC was between 0.02 and 0.05 μg mL−1 for S. typhimurium TA98 and TA100 (Table 2). Similarly, the LOAEC for E. coli DPD1718 was 0.07 μg mL−1 (Fig. 2 and Table 2). There was no significant difference between any of these assays. With acridine, the LOAEC for the biosensors E. coli DPD1718 and E. coli K12C600 was 0.06 and 0.31 μg mL−1, respectively, both significantly lower than the value for the Ames assay (1.24 μg mL−1 for S. typhimurium TA100) (Table 2). The biosensor E. coli DPD1718 was never less sensitive to the test chemicals than the Ames assay (Table 2).
For the collated data, correlation analysis confirmed a relatively close correlation between the induction of the SOS system as measured in the biosensor E. coli K12C600 and mutagenesis as measured in the Ames assay (P ≤0.05). Nevertheless, no significant correlation was reported between E. coli DPD1718 and the Ames assay (P =0.98). Nor did a physicochemical factor such as Kow, aqueous solubility or Henry's Law Constant aid in explaining the relationship between the mutagenic values for each of the assays (Debnath et al., 1992). It has, however, been shown by several authors (i.e. Quillardet & Hofnung, 1993; Mersch-Sundermann et al., 1994) that the results of SOS chromotest are comparable to those of the Ames assay with regard to the genotoxic mutagenic potency and that 60–70% of genotoxic substances identified by the SOS bioassays are also carcinogenic to mammals.
The SOS-lux-marked biosensors may have several key benefits when compared with the Ames assay including the following: (1) an assessment of mutagenicity can be made in 3 h, while the Ames results take at least 48 h; (2) the SOS-lux biosensors can detect a range of mutagenic chemicals with different DNA-damaging mechanisms using the same biosensor strain; (3) the simultaneous measurement of cell number/CFU and light emission of the biosensors allow discrimination between mutagenicity and cytotoxicity of the test substance; (4) temporal monitoring of the biosensor allows a kinetic interpretation of the response, which can be related to the form and mode of action of the chemical; and (5) the SOS-lux biosensors are capable of rapid screening of large sample numbers in a microplate format.
While there is a need for a sensitive, rapid and cost-effective bioassay for monitoring mutagenicity, it is clearly shown that no single bioassay can be applied in isolation to provide a definite and comprehensive assessment of the mutagenic hazard in samples under investigation. The SOS-lux biosensors cannot replace the role of direct measurement of carcinogenic effects in animals or detection of chromosomal aberrations in humans. However, such an approach can be employed as a cost-effective screening tool prior to a more established technique.
The Ames Salmonella assay has allowed a sensitive and accurate screening of molecules for their potential mutagenicity. Although it is laborious, it can be streamlined through the use of automated colony counting to enumerate histidine revertants colonies. However, Salmonella strains are classified as human pathogens, and in some countries, there are significant restrictions in their use and applications. The two SOS-lux biosensors had several practical advantages over the traditional assays including procedural simplicity, a rapid and unambiguous result, ease of measurement, rapid tabulation of exposure concentration and in vivo analysis without cell disruption. If the cell wall permeability of the biosensors can be enhanced, through assay refinement, then such an approach could transform rapid screening of potentially mutagenic samples.
The authors would like to thank Dr. Petra Rettberg and Dr. Ying Zhang for providing the biosensors strains. Dr. Matt Aitkenhead is also acknowledged for developing the colony counter software. The government of Saudi Arabia is acknowledged for funding this project.