Mutagenesis Advance Access originally published online on July 26, 2007
Mutagenesis 2007 22(5):335-342; doi:10.1093/mutage/gem022
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Bioluminescent Salmonella reverse mutation assay: a screen for detecting mutagenicity with high throughput attributes
Drug Safety Research and Development, Pfizer Global Research and Development, Eastern Point Road, Groton, CT 06340, USA
Here, we describe the development and evaluation of a novel bioluminescent high-throughput Salmonella reverse mutation assay applicable to the screening of large numbers of small molecules. The bioluminescent Salmonella assay utilizes genetically engineered standard Salmonella tester strains TA98 and TA100 expressing the lux(CDABE) operon from Xenorhabdus luminescence. In principle, the assay employs bioluminescence as a sensor of changes in bacterial metabolism associated with starvation or energy depletion effectively identifying colonies of histidine-independent revertant cells in a high-throughput fashion. The assay provides highly concordant data with the outcome in the standard Salmonella plate incorporation reverse mutation assay. Since the results of the standard Salmonella plate assay are required by various regulatory agencies for approval of new drugs, the bioluminescent Salmonella assay can be effectively used for prioritization of compounds in pharmaceutical drug discovery as well as the evaluation of environmental and industrial chemicals. Because of its high throughput attributes, the assay permits effective, fast and economical screening of a large series of structural analogs enabling the investigation of structure–activity relationships.
 |
Introduction |
|---|
 Top Introduction Materials and methodsResultsDiscussionConclusionsSupplementary dataReferences |
|---|
Recent progress in combinatorial chemistry, molecular biology, genomics and automation have revolutionized pharmaceutical drug discovery. This has led to the discovery of a relatively large number of compounds capable of eliciting a desired pharmacological effect. However, 40% of these drug candidates ultimately fail during clinical development due to safety-related issues (1
). This causes delays in the introduction of potentially vital medicines to patients. Therefore, the development of relevant mechanism-based high-throughput screening technologies to assess safety at the early stages of pharmaceutical drug discovery is extremely important to both accelerate decision making earlier in the drug development process and reduce cost.
Genetic toxicology provides the necessary information for assessment of the genotoxic risk associated with the use of drugs. Based on our experience, the overall attrition of drug candidates due to genetic safety concerns includes 
12% of drug candidates. Since the early 1970s, the Salmonella reverse mutation plate incorporation assay, developed by Bruce Ames (2), has been the most widely used and most thoroughly investigated assay for the detection of mutagenic activity. The Salmonella assay has been validated by a testing program involving several hundred chemicals sponsored by the National Toxicology Program (NTP) (3–7). Positive findings in the Salmonella reverse mutation assay have a good correlation with the outcome of rodent carcinogenicity testing (8,9). The results of the Salmonella assay are currently required by regulatory agencies for drug approvals worldwide (10). A positive result in the Salmonella assay usually leads to the discontinuation of development particularly for compounds intended for non-life-threatening chronic use indications (11).
The Salmonella reverse mutation assay relies on the detection of reversions of auxotrophic mutations (point mutations in histidine synthesis pathway genes) in Salmonella typhimurium. The inability of these cells to synthesize endogenous histidine manifests as a histidine-dependent phenotype that features lack of growth in the absence of exogenous histidine. A reverse mutation event restores the endogenous histidine biosynthesis and permits cell growth even in the absence of exogenous histidine (histidine-independent phenotype). In the standard Salmonella agar plate incorporation assay, the treatment of histidine-dependent cells with mutagens leads to an increased number of histidine-independent revertants. The revertant cells are detected as bacterial colonies on a background lawn of histidine-dependent cells grown on agar plates containing trace amounts of histidine. Since the current standard plate incorporation Salmonella reverse mutation assay lacks high throughput attributes, its application as a screening tool in the early phases of drug discovery is not feasible. Thus, several attempts to increase the efficiency of this assay have been published. Both the spiral assay (13) and Miniscreen (12,14) utilize the original Salmonella tester strains grown on modified agar plate formats. The Mutascreen (15) and AmesII assays (16) utilize liquid cultures of Salmonella and the revertant cells are detected using bacterial growth curves or a micro-well fluctuation method, respectively. The most widely used screens (Spiral and Miniscreen) showed a high concordance with the outcome of the standard Salmonella assay (17,18). However, the cumbersome methods for detection of revertants, relatively high compound requirements, low throughput and low potential for automation severely limit their application as screening tools in early drug discovery.
Here, we describe the development and evaluation of a bioluminescent higher throughput Salmonella reverse mutation assay applicable to screening of large numbers of chemicals in early drug discovery and for evaluation of large sets of environmental and industrial chemicals. The bioluminescent Salmonella assay utilizes genetically engineered standard Salmonella tester strains TA98 and TA100 expressing the lux(CDABE) operon from Xenorhabdus luminescence. The assay employs bioluminescence as a sensor of changes in bacterial metabolism associated with starvation or energy depletion effectively identifying colonies of histidine-independent revertant cells. It provides an economical, higher throughput tool for assessment of mutagenicity with high concordance to outcomes in the current standard plate incorporation method.
|
Materials and methods |
|---|
|
TopIntroductionMaterials and methodsResultsDiscussionConclusionsSupplementary dataReferences |
|---|
Construction of plasmids
The plasmid pSB417 carrying the whole lux(CDABE) operon was obtained from Dr S. Swift (University of Nottingham). The coding sequence of lux(CDABE) was placed under the control of a constitutive promoter residing on kanamycin transposon EZ::TN<KAN-2> (Epicentre). Briefly, the transposon was ligated with PvuII fragment of pUC19 (NEB) that contained the origin of replication and ß-lactamase as selection marker. The reaction resulted in a plasmid named pTN-kan. To allow a read-through control of the lux(CDABE) from kanamycin-resistance gene promoter, the lux(CDABE) sequence was placed downstream of the kanamycin resistance gene. Because the absence of appropriate cloning sites, the lux(CDABE) operon was isolated from pBS417 as EcoRI fragment and inserted into the EcoRI site of pBK-CMV (Stratagene, La Jola, CA). The resulting plasmid, pLUX, was used as a source of the 5.7-kb promoterless BamHI- and PstI-flanked lux(CDABE) sequence. The promoterless lux(CDABE) fragment was subsequently ligated into the pTNKan. The resulting plasmid, named pUC-TNlux, contained the functional lux(CDABE) expression cassette consisting of kanamycin phosphotransferase and lux(CDABE). The pUC19-based pUC-TNlux plasmid replicates in the cytoplasm in high copy numbers (1000–3000 copies per cell). The medium (100–300 copies per cell) and low (2–6 copies per cell) copy number plasmids were constructed using vectors pBR322 (NEB) and pFN467 (ATCC number 86962), respectively. The lux(CDABE) expression cassette was excised from pUC-TNlux as PvuII-flanked fragment and inserted into EcoRV site of pBR322 or PvuII site of pFN467 creating pBR-TNlux and pFN-TNlux, respectively.
Construction of bioluminescent Salmonella tester strains
Salmonella tester strains TA98 and TA100 (Molecular Toxicology, Bonne, NC) were used for preparation of electrocompetent cells. Briefly, the bacterial cultures were grown in NB medium (Life Technologies) at 37°C until reaching the cell density OD600 = 0.5–0.8. Salts were removed by series of washes and finally re-suspended in 10 ml of ice-cold 10% glycerol in water. Eighty microlitres aliquots of electrocompetent cells were flash frozen in liquid nitrogen and stored at –80°C. The transformation of S. typhimurium was performed using a BTX electroporator (BTX, Holliston, MA). Frozen aliquots of electrocompetent cells were slowly thawed on ice. Eighty microlitres of the cell suspension and 5 µl of lux(CDABE) containing plasmid DNA solution (10 µg) were placed into an electroporation cuvette (0.2-cm electrode gap). The electroporation conditions consisted of three 99-µsec pulses with field strength 2.5 KV/cm. The electroporation mix was re-suspended in 1 ml SOC medium (Life Technologies, Geithesburg, MD) and incubated at 37°C for 1 h. The cell suspension was plated on agar plates supplemented with kanamycin to select for transformed cells. At least four independent bioluminescent colonies from each strain were further characterized and their capability of detecting mutagens was evaluated.
Bioluminescent Salmonella reverse mutation assay
Salmonella cells were initially inoculated into 35 ml of LB medium (Life Technologies) supplemented with 50 µg/ml kanamycin (Sigma, St Louis, MO) and grown at 30°C for 15–18 h (over night) under agitation (250 r.p.m.). The overnight culture was diluted 3:1 with fresh kanamycin supplemented LB medium and grown at 37°C until the cell density reached OD600 = 0.5–0.6. Then, cells were washed twice in PBS (Dulbecco's phosphate-buffered saline, Life Technologies). The final cell density was adjusted to OD600 = 0.5 and the cell suspension was kept on ice. Aliquots of a sterile top agar solution (7.25 ml) were prepared using following components: 50x Vogel–Bonner salt concentrate (Life Technologies) 0.2 ml, biotin (170 µg/ml) 0.44 ml, histidine (6.7 mg/ml) 0.01 ml, kanamycin (50 µg/ml) 0.01 ml, glucose (2.6%) 1.50 ml, cells (OD600 = 0.5) 2.5 ml, agar (2.4%) 2.59 ml. The top agar was prepared before use and kept at 50°C. Since the exposure of Salmonella cells to a high temperature might decrease their viability, the unused top agar was discarded after 10 min. The assay was performed in minimal agar 24-well plates prepared ahead of time. Each well contained 1 ml of minimal agar. The minimal agar was prepared using 15 g Bactoagar (DIFCO, Franklin Lakes, NJ), 4 g glucose and 20 ml of 50x Vogel–Bonner salt concentrate (1% MgSO4 x H2O, 10% citric acid, monohydrate, 50% K2HPO4, 17.5% Na2NH2PO4 x 4H2O) in total volume of 1000 ml of distilled water. The sterilized agar was allowed to cool down to 50°C in the water bath prior to dispensing appropriate volumes into each well. The assay procedure was initiated by pipetting 17.5 µl of PBS or in case of metabolic activation (rat S9 Aroclor induced liver homogenate) 17.5 µl of complete S9 mix onto the surface of minimal agar. Then, 10 µl of the compound solution was added followed by 70 µl of the top agar containing cells was laid over the mix. Alternatively, the assay was also feasible in 48-well plate format. In this case, the amount of reagents and cells were reduced 50% to accommodate the smaller well size. To allow for a proper distribution of all components, the plate was agitated with a vortexer at 950 r.p.m. while being plated. The final Salmonella cell density was 2.5 x 107 cells per well in 24-well plate format. The Salmonella cells were exposed to 3- to 5-fold serial dilutions of tested compounds. The maximum concentration tested in the bioluminescent assay was 1.35 mg/ml in the final overlay, which translates to 0.135 mg per well. This corresponds to 5 mg per plate in the standard Salmonella assay assuming 3.7 ml of agar overlay. To facilitate the comparison of bioluminescent and plate assay data, all concentrations in the bioluminescent assay were expressed as corresponding milligrams per plate eq. (equivalent) or micrograms per plate eq. The positive controls for bioluminescent strains TA100lux and TA98lux in the absence of metabolic activation were nitrofurantoin at 0.007 mg per plate eq. and 2-nitrofluorene at 0.018 mg per plate eq., respectively. The positive control for strains TA100lux and TA98lux in the presence of metabolic activation was 2-anthramine at 0.018 mg per plate eq. and 0.037 mg per plate eq., respectively. Plates were incubated at 37°C for 48 h. Luminescent colonies of revertants were visualized using photon counting CCD camera (Lumi-Imager, Boehringer, Germany). For comparison of bioluminescent and standard assay formats, at least three independent experiments were performed and statistically analysed using t-test. In a screening mode of testing, a compound producing at least a 2- or 2.5-fold increase of bioluminescent colonies over the untreated controls with evidence of a dose response was considered mutagenic for TA100lux or TA98lux, respectively.
Standard Salmonella reverse mutation assay
The assay was performed in a plate incorporation format using S. typhimurium tester strains TA98 and TA100 according to the previously described protocols (19). Briefly, the Salmonella cells were treated with tested compounds dissolved in dimethyl sulfoxide (DMSO) in soft agar overlays at concentrations ranging from 0.015 to 5 mg per plate or an appropriate amount of DMSO in the presence or absence of aroclor 1254-induced rat liver S9 mixture and an NADPH-regenerating system. The number of visible revertant colonies present after 72 h incubation at 37°C was recorded and the fold of change over DMSO-treated control plates was calculated. At least three independent experiments with each Salmonella tester strain were performed. The analysis of the data included calculating average and SEM. The statistical significance was determined using the t-test. A 2-fold statistically significant increase of the number of revertant colonies over the DMSO-treated controls was considered as a positive response.
Spiral modification of Salmonella reverse mutation assay
The assay was conducted with tester strains TA98 and TA100 in the absence and presence of metabolic activation (rat liver S9) according to methods described previously (13). Concentrated stocks of bacterial cultures were applied in a spiral pattern uniformly across minimal agar plates containing trace amounts of histidine using an Autoplate 4000 (Spiral Biotech, Bethesda, MD). Stock solutions of test article (34.5 mg/ml) or control were then applied to each plate in a spiral pattern over the bacteria using a variable cam setting of the stylus to achieve an approximate 240-fold concentration gradient (0.021–5 mg per plate eq.) on the plate. For plates treated in the presence of the metabolic activation system, a mixture of the rat S9 and NADPH-regenerating system was applied in a spiral pattern over the treated bacteria using the variable cam setting. Duplicate plates were treated for each stock solution or control article with each strain and test condition. After incubation (40–48 h) at 37°C, the plates were scanned using an automatic laser counter instrument to determine the number of revertant colonies from discrete portions of the spiral pattern. An estimated dose was calculated by multiplying the compound concentration in the stock solution by the volume of stock solution deposited along the spiral pattern in which the revertant count was obtained. Each corrected count and dose was extrapolated to the equivalent pour-plate value based on the ratio of the area of the spiral pattern to the area of a standard 150-mm plate. For each strain, a mutagenic response was evaluated by comparing the estimated average number of revertants per plate of the negative control plate to the test article plates. A positive response was determined as a dose-related increase in revertant counts with a minimum of a 3-fold increase over background in least three consecutive dose levels.
|
Results |
|---|
|
TopIntroductionMaterials and methodsResultsDiscussionConclusionsSupplementary dataReferences |
|---|
Construction and characterization of bioluminescent Salmonella tester strains
The bioluminescent Salmonella tester strains were derived from commonly used strains TA98 and TA100 that are sensitive to frame-shift and base-pair substitution mutagens, respectively. The bioluminescent phenotype of Salmonella cells was achieved via transformation with episomal plasmids (pUC-TNlux, pBR-TNlux or pFN-TNlux) that carry the luxCDABE operon expression cassette. At least four independent strains derived from each plasmid and parental strain were characterized. The sensitivity of the luminescent reporter to disruption of cellular metabolism and metabolic changes elicited by starvation were evaluated. The disruption of mitochondrial metabolism via treatment with 1 mM potassium cyanide instantly abolished the luminescent phenotype. Although there is a substantial difference among plasmid copy numbers in cells for individual plasmid backbones, the luminescent output was independent of plasmid backbone used. The intensity of the bioluminescence correlated with the cell concentration only during the late log phase of the bacterial growth reaching peak levels at the beginning of the stationary phase followed by a sharp decline (Figure 1). The bacterial strains TA100 and TA98 carrying pUC-TNlux and pBR-TNlux, respectively, were designated as tester strains TA100lux and TA98lux and used in further experiments.
|
Principle of the bioluminescent Salmonella assay
The standard Salmonella assay relies on the detection of reversions of auxotrophic mutations in histidine synthesis pathway (2
). The treatment of histidine-dependent cells with mutagens leads to an increased number of histidine-independent colonies of revertant cells on plates containing only a trace amount of histidine (Figure 2A).
|
The bioluminescent Salmonella assay exploits bioluminescence for the detection of histidine-independent revertant cells. In contrast, the available luminescence-based assays such as Vitotox (20
) utilize DNA damage promoter–reporter biosensor for detection of activation of DNA damage pathway. In the bioluminescent Salmonella assay, the lack of histidine in the medium over a time period leads to a starvation of histidine-dependent cells eliciting changes in bacterial metabolism, and making cells incapable of maintaining high level of bioluminescent output. On the other hand, histidine-independent revertant cells are not starving and capable of producing measurable bioluminescence even in the absence of histidine (Figure 2B). Such cells are detected as small luminescent colonies on a background lawn of histidine-dependent cells using a photon counting camera (Figure 3). Since the size of revertant colonies in a microplate format is affected by the cytotoxicity of tested chemicals and likely also by bacterial quorum sensing mechanisms, the bioluminescent Salmonella assay requires measuring the incidence of revertant colonies instead of monitoring the overall luminescence output. The application of image analysis algorithms would then enable detecting even small colonies of revertants cells in a high throughput format in an automated fashion.
|
A representative example of the bioluminescent detection of revertant colonies is presented in Figure 3. The figure represents an inverted image of cells treated with increasing concentrations of 2-anthramine in a 24-well microplate format. The colonies of revertants were visualized using a photon counting camera after 48-h incubation at 37°C and counted. According to expectations, the treatment of cells with 2-anthramine increased the incidence of revertant colonies in a dose-dependent fashion. Furthermore, the size of revertant colonies decreased with increased incidence of revertant colonies, making them invisible without a dissecting microscope. To confirm the histidine-independent phenotype of cells forming bioluminescent colonies, we have evaluated 20 small colonies formed after treatment with 2-anthramine via replica plating into medium lacking histidine. All colonies tested were capable of growing in the absence of histidine.
The residual bioluminescence of the bacterial lawn provided the opportunity for evaluating cytotoxicity of tested chemicals and differentiating microcolonies from colonies of true revertants (Figure 4). Microcolonies, a known artifact of the Salmonella assay, consist of histidine-dependent cells (non-revertants) that feed on the histidine made available as a consequence of cytotoxicity of tested chemical. They appear on the background of a reduced bacterial lawn and, if replica plated, they do not grow in a histidine-deficient medium. In our bioluminescent assay, the reduction of the bacterial lawn detected microscopically correlated with the loss of the residual bioluminescence of the bacterial lawn. The microcolonies in the bioluminescent assay were detected as small colonies present in wells with reduced bioluminescence of the bacterial lawn (Figure 4, wells A3 and A4). The histidine-dependent nature of microcolonies was confirmed by replica plating of 20 microcolonies into histidine-deficient medium. In this experiment, none of the microcolonies was capable of growing in the absence of histidine.
|
Evaluation of bioluminescent Salmonella assay using model agents
The performance of the bioluminescent Salmonella assay was compared with the standard Salmonella plate incorporation assay using a set of six model agents (Figure 5). The bioluminescent Salmonella assay with strains TA98lux and TA100lux in 24-well plates was performed according to the protocol described in Materials and Methods. The 24-well microplate assay format seemed to be most advantageous because it allowed testing of one compound at five concentrations in duplicate on a single plate. The bioluminescent colonies were visualized using a photon counting camera (Lumimager F-1, Roche, Palo Alto, CA) after incubation at 37°C for 48 h. The standard plate incorporation Salmonella assay was performed according to standard protocol (19
) using the parental TA98 and TA100 strains in 10-cm agar plates. Each compound was tested using both assay formats in parallel in the absence and presence of metabolic activation (rat liver S9). To facilitate the comparison of assay formats, the concentrations of compounds in the bioluminescent assay were expressed as milligrams per plate eq. corresponding to ones used in the standard plate assay (see Materials and Methods). At least three independent experiments were performed and statistical significance of the data was evaluated using t-test.
|
The background incidences of revertant colonies per well for TA100lux and TA98lux strains were 10.1 ± 2.3 and 2.1 ± 1.2, respectively. The treatment of TA100lux with positive controls, nitrofurantoin at 0.007 mg per plate eq. in the absence of S9 and 2-anthramine at 0.018 mg per plate eq. in the presence of S9, produced 4.9 ± 0.8-fold and 7.0 ± 1.6-fold increases of revertant colonies over the controls, respectively. In the case of TA98lux, the positive controls, 2-nitrofluorene at 0.018 mg per plate eq. in the absence of S9 and 2-anthramine at 0.037 mg per plate eq. in the presence of S9, elicited 35.3 ± 6.1-fold and 38.5 ± 8.4-fold increases of revertant colonies over the controls, respectively. In the standard Salmonella assay, the background incidences of revertant colonies for TA100 strain in the absence and presence of S9 were 124 ± 18 and 138 ± 21, respectively. In case of the strain TA98, the background in the absence and presence of S9 were 24 ± 3 and 34 ± 4, respectively. The treatment of TA100 strain with positive controls, nitrofurantoin at 0.007 mg per plate in the absence of S9 and 2-anthramine at 0.018 mg per plate in the presence of S9, produced 10.6 ± 2-fold and 9.9 ± 2-fold increases of revertant colonies over the controls, respectively. The positive controls in the case of TA98, 2-nitrofluorene at 0.018 mg per plate in the absence of S9 and 2-anthramine at 0.037 mg per plate in the presence of S9, elicited 40 ± 5-fold and 66 ± 8-fold induction of mutation frequency.
The set of model agents consisted of three Salmonella assay positive and three negative compounds and the results are summarized in Figure 5. Both assays provided highly reproducible data. Though the results from the bioluminescent Salmonella assay showed 100% concordance with the standard plate incorporation assay, a comparison of the results between the two assays revealed differences in the shape of the dose–response curves. Generally, the maximum magnitude of the response was lower in the bioluminescent assay. This was most apparent in the case of busulfan and 1,8-dihydroxy-4,5-dinitroanthracene-9,10-dione (Figure 5). Furthermore, the cytotoxicity of several compounds (nitrofurazone and fuchsin), detected as thinning of bacterial lawn, was more pronounced in the standard plate assay.
Validation of the bioluminescent Salmonella assay for screening
The potential of the bioluminescent Salmonella assay as a screening tool for mutagens was evaluated using 105 compounds previously tested through the NTP (3–7,21). The selected set included compounds with published responses evenly distributed across a wide range. To compare the bioluminescent Salmonella assay with a commonly used screen, the compounds were also tested in parallel using the spiral Salmonella assay (13). The results of the bioluminescent and spiral Salmonella assays were compared with each other and with data produced previously using the standard Salmonella pre-incubation plate assay format and published by the NTP (3–7,21). The bioluminescent assay was performed using the 24-well microplate format with strains TA98lux and TA100lux. The spiral assay was performed according to published protocol (13) using strains TA98 and TA100. To emulate screening conditions, a single bioluminescent and spiral Salmonella assay was performed with each compound. The results of both assays including the published NTP data were expressed as maximum fold of increases of revertant colonies over the untreated controls and summarized graphically in Figure 6. A tabular format of all data points for each tested compound is available as supplemental material on the Mutagenesis Online or can be requested from the authors. The statistical analysis of the data included evaluation of concordance, sensitivity, specificity and predictive values positive and negative (Figure 6).
|
The concordance between the bioluminescent and spiral Salmonella assay for the set of 105 compounds was in the range of 84.7–89.5%. The bioluminescent Salmonella assay showed 93.3–95.2% concordance with the published NTP data (Figure 6). On the other hand, the spiral assay displayed only an 81.0–87.8% concordance with the published NTP results. The majority of discordant spiral assay results were caused by enhanced compound toxicity seen under the conditions of the spiral assay and comprised mainly false-negative results.
|
Discussion |
|---|
|
TopIntroductionMaterials and methodsResultsDiscussionConclusionsSupplementary dataReferences |
|---|
Since a positive finding in the Salmonella reverse mutation assay historically exhibits a good correlation with the outcome of the rodent 2-year bioassay for carcinogenicity (8
,9,22), the Salmonella assay has become an essential component of the safety assessment paradigm for drugs and chemicals required by regulatory agencies worldwide. Although several attempts to develop higher throughput formats of the Salmonella assay were published, their application as screening tools are severely limited due to a lack of higher throughput attributes. Here, we describe the development and evaluation of a novel higher throughput bioluminescent Salmonella assay format suitable to screening of large numbers of compounds in drug discovery or alternatively large sets of environmental and industrial chemicals.
In the Salmonella assay, the histidine-dependent cells are treated with tested compounds in a soft agar overlay containing only a trace amount of histidine sufficient for cells to undergo two to three divisions that is necessary for induction and expression of mutations. Compound treatment-induced reversion mutation events are detected as an increased number of histidine-independent colonies grown on a lawn of histidine-dependent cells (2).
The bioluminescent Salmonella assay utilizes bioluminescence as a sensor for the detection of colonies of histidine-independent revertant cells (Figure 1B). The lack of histidine in the medium over time leads to starvation of histidine-dependent cells that makes them incapable to maintain the bioluminescent phenotype. Under the same conditions, the histidine-independent revertant cells are capable of maintaining luminescence forming small luminescent colonies easily detectable via a photon counting camera (Figure 3). The bioluminescent sensor was achieved via the expression of the lux(CDABE) operon from Xenorhabdus luminescence (ATCC number 29999) in the Salmonella cells. Since the lux(CDABE) operon includes genes encoding a thermostable luciferase (A and B) and a substrate producing fatty-acid reductase (C, D and E) (23), the expression of the lux(CDABE) provides a closed system that generates bioluminescence in living cells at 37°C. To achieve the bioluminescent phenotype, the cells have to synthesize components of the lux(CDABE) pathway and provide an energy source in the form of reduced flavine mononucleotide (FMNH2) (24). The fact that the luminescent phenotype of Salmonella cells was affected by starvation (Figure 1) and disruption of cellular metabolism via potassium cyanide treatment indicate that the constitutive expression of lux(CDABE) in Salmonella provides a functional biosensor for changes associated with starvation in living Salmonella cells. This biologically engineered feature makes it suitable for detecting histidine-independent revertant cells grown among histidine-dependent cells in our assay system (Figure 3).
The application of bioluminescence for the detection of revertant cells permitted unprecedented miniaturization of the format of the Salmonella assay. In general, the background mutation frequency in various bacterial systems is relatively stable (between 10-6 and 10-9 mutants per cell). Therefore, all miniaturization approaches require treating cells at a high population density. However, this results in a reduced number of macroscopically discernible revertant colonies, particularly at concentrations producing a large number of revertants (Figure 3). This fact effectively limits options for assay miniaturization. We hypothesize that this phenomenon, forming of only small pinhead colonies of revertant cells in high-density populations, is most likely a product of quorum sensing-controlled behaviours occurring only when bacteria are at high cell population densities (25). Since in the bioluminescent Salmonella assay the histidine-independent revertant cells are capable of producing enough FMNH2 and components of lux(CDABE) pathway to maintain bioluminescent phenotype, even small bioluminescent colonies that are not visible macroscopically can be easily visualized using a photon counting camera. In addition, the digitized images of colonies enable potentially fully automated data analysis (J. Aubrecht, J. Osowski and K. Lam manuscript in preparation).
In addition, the association of a state of starvation or energy depletion with the bioluminescent phenotype enabled a relatively easy assessment of cytotoxicity of tested chemicals and simplified differentiation of histidine-dependent microcolonies (non-revertants) arising as a consequence of cytotoxicity and colonies of histidine-independent cells (true revertants). Since the starving histidine-dependent cells are capable of only residual bioluminescence, the bacterial lawn was easily detectable as a shadow background on the inverted image (Figures 3 and 4). The application of toxic chemicals resulted in a complete loss of the residual bioluminescence of the bacterial lawn (Figure 4). The microcolonies of histidine-dependent cells (non-revertants) feeding on unutilized histidine were distinguishable as bioluminescent colonies in wells that lack residual background bioluminescence of the bacterial lawn (Figure 4, wells A3 and A4). This finding was in agreement with microscopic evaluation of wells.
To compare the performance of the bioluminescent format with the standard assay, a set of 10 model agents were tested in both assays in parallel. Although, the bioluminescent assay data were in agreement with the standard plate incorporation assay (Figure 5), several trends such as lower magnitude of the response in the bioluminescent assay and more pronounced cytotoxicity for several compounds in the standard assay were observed. Similar discrepancies have been found previously when comparing pre-incubation and plate incorporation assay protocols for the standard Salmonella assay (26). Therefore, the discrepancies between the bioluminescent and the standard assay formats are most likely due to differences in assay protocols such as layering the assay components onto the top agar layer in the case of the bioluminescent format versus mixing the assay components in the tube and subsequently plating onto the top agar in the plate incorporation format.
The suitability of the bioluminescent Salmonella assay as a screening tool was evaluated further using 105 previously tested compounds (3–7). Furthermore, the bioluminescent assay format was compared with the spiral modification of the Salmonella assay, another commonly used screening format in drug discovery (13). There was 93–95% concordance between the bioluminescent assay format and the published NTP data (Figure 6). The discrepancy between the bioluminescent and plate assay results published by the NTP could be attributed to variable purity grades of tested compounds, variability of the test systems and/or differences in the experimental protocol (bioluminescent versus standard preincubation plate assay formats). Since the set of tested compounds included agents covering a whole range of mutagenic responses, the bioluminescent Salmonella assay seems suitable for screening. In contrast, the spiral assay showed only 81–88% concordance with the NTP results. The discordant results were mainly false negatives due to increased cytotoxicity suggesting differences in compound availability to the tester strain under the conditions of the spiral assay. The comparison of the spiral and bioluminescent formats revealed that the amount of test compound sample required for the spiral assay, 16 mg per strain, is almost 10-fold higher than that required for the bioluminescent assay. Because of the relative simplicity of the bioluminescent assay, the bioluminescent format is expected to be amenable to full automation and reaching throughput of hundreds of compounds per week.
|
Conclusions |
|---|
|
TopIntroductionMaterials and methodsResultsDiscussionConclusionsSupplementary dataReferences |
|---|
In summary, this study shows that the bioluminescent Salmonella reverse mutation assay provides highly concordant data with the outcome of the current standard Salmonella plate incorporation assay and is an efficient screening format capable of testing large numbers of small molecules. Additional gains in assay efficiency could be realized through further assay automation steps. Since the conduct of the Salmonella plate assay is recommended in Organization for Economic Cooperation and Development and International Conference of Harmonization regulatory testing guidelines, the bioluminescent Salmonella assay can be effectively used for prioritization of compounds in pharmaceutical drug discovery and alternatively for screening large sets of environmental and industrial chemicals requiring some initial prioritization of concern level.
|
Supplementary data |
|---|
|
TopIntroductionMaterials and methodsResultsDiscussionConclusionsSupplementary dataReferences |
|---|
Supplementary data are available at Mutagenesis Online.
|
Acknowledgments |
|---|
We thank Dr Robert Schiestl and members of the Aubrecht laboratory for comments on the manuscript, Dr Jim Mcloughlin for help with obtaining recombinant material and Dr Simon Swift for providing the plasmid pSB417 containing the lux(CDABE) operon.
|
Notes |
|---|
* To whom correspondence should be addressed. Tel: 860 715 3384; Fax: 860 715 7884; Email: Jiri_Aubrecht{at}groton.pfizer.com
|
References |
|---|
|
TopIntroductionMaterials and methodsResultsDiscussionConclusionsSupplementary dataReferences |
|---|
-
1. DiMasi JA. Success rates for new drugs entering clinical testing in the United States. Clin. Pharmacol. Ther. (1995) 58:1–14.[CrossRef][Web of Science][Medline]
2. Ames BN, McCann J, Yamasaki E. Methods for detecting carcinogens and mutagens with the Salmonella/mammalian-microsome mutagenicity test. Mutat. Res. (1975) 31:347–364.[Web of Science][Medline]
3. Haworth S, Lawlor T, Mortelmans K, Speck W, Zeiger E. Salmonella mutagenicity test results for 250 chemicals. Environ. Mutagen. (1983) 5:3–142.[Medline]
4. Mortelmans K, Haworth S, Lawlor T, Speck W, Tainer B, Zeiger E. Salmonella mutagenicity tests: iI. Results from the testing of 270 chemicals. Environ. Mutagen. (1986) 8:1–119.[Web of Science][Medline]
5. Zeiger E, Anderson B, Haworth S, Lawlor T, Mortelmans K. Salmonella mutagenicity tests: IV. Results from the testing of 300 chemicals. Environ. Mol. Mutagen. (1988) 11:1–157.[Web of Science][Medline]
6. Zeiger E, Anderson B, Haworth S, Lawlor T, Mortelmans K. Salmonella mutagenicity tests: V. Results from the testing of 311 chemicals. Environ. Mol. Mutagen. (1992) 19:2–141.[CrossRef][Web of Science][Medline]
7. Zeiger E, Anderson B, Haworth S, Lawlor T, Mortelmans K, Speck W. Salmonella mutagenicity tests: III. Results from the testing of 255 chemicals. Environ. Mutagen. (1987) 9:1–109.[Medline]
8. Kim BS, Margolin BH. Prediction of rodent carcinogenicity utilizing a battery of in vitro and in vivo genotoxicity tests. Environ. Mol. Mutagen. (1999) 34:297–304.[CrossRef][Web of Science][Medline]
9. Zeiger E. Identification of rodent carcinogens and noncarcinogens using genetic toxicity tests: premises, promises, and performance. Regul. Toxicol. Pharmacol. (1998) 28:85–95.[CrossRef][Web of Science][Medline]
10. Muller L, Kikuchi Y, Probst G, Schechtman L, Shimada H, Sofuni T, Tweats D. ICH-harmonised guidances on genotoxicity testing of pharmaceuticals: evolution, reasoning and impact. Mutat. Res. (1999) 436:195–225.[CrossRef][Web of Science][Medline]
11. Snyder RD, Green JW. A review of the genotoxicity of marketed pharmaceuticals. Mutat. Res. (2001) 488:151–169.[CrossRef][Web of Science][Medline]
12. Burke DA, Wedd DJ, Burlinson B. Use of the Miniscreen assay to screen novel compounds for bacterial mutagenicity in the pharmaceutical industry. Mutagenesis (1996) 11:201–205.
13. Houk VS, Schalkowsky S, Claxton LD. Development and validation of the spiral Salmonella assay: an automated approach to bacterial mutagenicity testing. Mutat. Res. (1989) 223:49–64.[CrossRef][Web of Science][Medline]
14. Brooks TM. The use of a streamlined bacterial mutagenicity assay, the MINISCREEN. Mutagenesis (1995) 10:447–448.
15. Falck K, Partanen P, Sorsa M, Suovaniemi O, Vainio H. Mutascreen, an automated bacterial mutagenicity assay. Mutat. Res. (1985) 150:119–125.[Web of Science][Medline]
16. Gee P, Maron DM, Ames BN. Detection and classification of mutagens: a set of base-specific Salmonella tester strains. Proc. Natl Acad. Sci. USA (1994) 91:11606–11610.
17. Diehl M, Fort F. Spiral Salmonella assay: validation against the standard pour-plate assay. Environ. Mol. Mutagen. (1996) 27:227–236.[CrossRef][Web of Science][Medline]
18. Diehl MS, Willaby SL, Snyder RD. Comparison of the results of a modified miniscreen and the standard bacterial reverse mutation assays. Environ. Mol. Mutagen. (2000) 36:72–77.[CrossRef][Web of Science][Medline]
19. Maron DM, Ames BN. Revised methods for the Salmonella mutagenicity test. Mutat. Res. (1983) 113:173–215.[CrossRef][Web of Science][Medline]
20. van der Lelie D, Regniers L, Borremans B, Provoost A, Verschaeve L. The VITOTOX test, an SOS bioluminescence Salmonella typhimurium test to measure genotoxicity kinetics. Mutat. Res. (1997) 389:279–290.[Web of Science][Medline]
21. Zeiger E, Drake JW. An environmental mutagenesis test development programme. IARC Sci. Publ. (1980) 303–313.
22. Ames BN, Durston WE, Yamasaki E, Lee FD. Carcinogens are mutagens. Simple test system combining liver homogenates for activation and bacteria for detection. Proc. Natl Acad. Sci. USA (1973) 70:2281–2285.
23. Frackman S, Anhalt M, Nealson KH. Cloning, organization, and expression of the bioluminescence genes of Xenorhabdus luminescens. J. Bacteriol. (1990) 172:5767–5773.
24. Stewart GS, Williams P. Lux genes and the applications of bacterial bioluminescence. J. Gen. Microbiol. (1992) 138:1289–1300.
25. Bassler BL. Small talk. Cell-to-cell communication in bacteria. Cell (2002) 109:421–424.[CrossRef][Web of Science][Medline]
26. Schneider M, Quistad GB, Casida JE. Glutathione activation of chloropicrin in the Salmonella mutagenicity test. Mutat. Res. (1999) 439:233–238.[Web of Science][Medline]
Received on January 12, 2007; revised on April 26, 2007; accepted on May 10, 2007.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. Ackerman, R. Sharma, J. Hitchcock, T. Hayashi, Y. Nagai, S. Li, S. Lu, J. Miret, K. Tang, F. Spence, et al. Inter-laboratory evaluation of the bioluminescent Salmonella reverse mutation assay using 10 model chemicals Mutagenesis, September 1, 2009; 24(5): 433 - 438. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. W. Knight, L. Birrell, and R. M. Walmsley Development and Validation of a Higher Throughput Screening Approach to Genotoxicity Testing Using the GADD45a-GFP GreenScreen HC Assay J Biomol Screen, January 1, 2009; 14(1): 16 - 30. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||