ABSTRACT
A method for the detection of the SOS response as measured by the liberation of resident prophages from the genomes of their hosts is described. It is based on the use of two converging oligonucleotides that flank the attP attachment site of the phage as primers for real-time PCR. Amplification was observed only after the phage DNA became excised. The system responds to both chemicals and physical conditions. Quantitative data on the concentration and/or potency of the genotoxic condition were obtained. Results can be achieved within 1 day and are less susceptible to possible toxic effects than phage generation or other methods that require DNA synthesis. The use of both gram-positive and gram-negative bacteria widens the range of compounds that can be tested because it eliminates impermeability problems derived from the particular composition of each cell wall type.
The use of microbe-based assays for genotoxicity assessment has several advantages over animal-cell-based systems, including the simplicity of the procedures, the relatively short time needed to obtain results, and, as a consequence, their inexpensiveness. Therefore, they have been used to detect the mutagenic potentials of many different classes of compounds from aliphatic epoxides to polycyclic aromatic hydrocarbons and including pesticides, mycotoxins, chemotherapeuticals, and even environmental samples like hazardous industrial wastes. Microbial genotoxicity assays can be divided into two main groups: those aimed to detect the reversion of a point mutation, such as the Salmonella mutagenicity test (2, 13), or a readily selectable forward mutation, for example, the arabinose resistance test (6), and those designed to reveal the induction of the SOS response. This coordinated reaction of bacteria against DNA damage relies on a set of genes that remain silent during normal growth but become derepressed by circumstances that affect replication and result in the synthesis of DNA repair enzymes and the inhibition of cell division (8, 21).
Two main strategies have been adopted for the visualization of the SOS response and the quantification of the inducing ability of the compounds being tested. In one of these strategies, promoters that respond to SOS induction are placed in front of appropriate reporter genes such as lacZ (12, 16), the lux cluster (4, 7, 20), or the green fluorescent protein gene (3), thus allowing colorimetric, luminescence, or fluorescence detection of the resulting activities. The other approach, the microscreen prophage induction assay, relies on an Escherichia coli strain that is lysogenic for bacteriophage lambda. Induction of the SOS response leads to the activation of the latent coprotease activity of RecA, which binds the CI repressor and induces its proteolysis (10, 21). The cleavage of CI relieves the transcriptional repression of the PR lytic promoter, which is the first step of lambda entry into the lytic cycle. Successful induction of lambda would result in the production of progeny and a loss of turbidity of the cultures, both easily scored (5, 18).
In this report, a new quantitative method for the detection of prophage induction is described, which directly measures the excision of the prophage from the host chromosome. The new method is based on the use of two converging oligonucleotides, corresponding to each side of the attP sequence of a temperate phage, to amplify the intervening DNA segment by real-time PCR. No DNA amplification is expected from an uninduced lysogenic culture (apart from the “noise” resulting from spontaneous induction) in which the attP half-sites are separated and oppositely orientated. However, liberation of the prophage will result in the generation of a circular phage genome, permitting the amplification of the regenerated attP DNA fragment. Because an early step of prophage induction is assessed, the determination is more precise than methods that score lysis and phage release, the final event of the process, which depends on successful accomplishment of all previous steps. This is particularly important when genotoxicity is to be measured, because most of the compounds that induce it directly or indirectly interfere with DNA replication and, thus, the generation of viral progeny. The procedure was developed with lysogenic cultures of E. coli and of Lactobacillus casei for two reasons: to demonstrate its general feasibility for any phage-bacterium combination and to deal with possible permeability problems for the genotoxic compound to be tested, which may be caused by the different structures of the gram-positive and gram-negative bacterial cell walls. Finally, the results can be obtained in a few hours (less than a working day) and can be confirmed, if desired, by scoring the phages (as plaques) generated in the culture as a consequence of the induction of the SOS response.
MATERIALS AND METHODS
Bacterial strains, bacteriophages, and incubation conditions.A lysogenic derivative of L. casei ATCC 393 containing an A2 prophage inserted in its tRNALeu gene (1) was propagated in liquid MRS medium (Oxoid, Basingstoke, United Kingdom) supplemented with 10 mM CaCl2 and 10 mM MgSO4 set at 30°C without aeration. A lambda lysogenic derivative of E. coli W3110 was propagated in L broth (19) at 37°C with agitation. Mixed cultures of L. casei and E. coli were done in 2× YT medium (19), with the rest of the conditions being those described above for L. casei.
Plaque enumeration was done on lawns of L. casei ATCC 393 or E. coli W3110 growing on the same medium used for the propagation of each bacterium, which was solidified with 1.5% agar (Roko, Llanera, Spain) and covered by semisolid medium (0.7% agar).
Prophage excision assays.Cultures of E. coli and L. casei were grown to the early exponential phase (optical density at 600 nm of 0.1), the cultures were aliquoted (25 ml), the genotoxic agent was added at concentrations that ranged from 50 nM to 5 μM, and incubation continued under the same conditions. Control tests of the solvents used were performed to discard any induction ability. To determine the effect of UV irradiation, the cultures were centrifuged, suspended in a 1/10 volume of 0.85% (wt/vol) NaCl, and exposed to 1 J/m2 for different periods of time. The cell suspensions were then diluted 1:10 into prewarmed broth, and incubation was continued in the dark.
Two samples of 100 μl were taken from the cultures every 30 min. One of them, to be used in PCR experiments, was immediately frozen, while the other was centrifuged at 4°C, and the supernatant was diluted and used to determine the phage concentration. Each experiment was carried out in triplicate and repeated at least three times.
Real-time PCR was performed with primers that matched phage sequences placed at each side of the attP attachment site: 5′-TTGTGTGCCCATATTTCTGAACTCT-3′ and 5′-GCAAGAATGCCGGTTTAAAGCC-3′ for A2 (1) and 5′-GTTGATTCATAGTGACTGC-3′ and 5′-CTGATAGTGACCTGTTCG-3′ for lambda (14). The PCRs were done in a final volume of 15 μl containing 1.25 μl of the cell culture using the PCR Q SYBR Green Supermix (Bio-Rad) according to the instructions of the manufacturer. Amplifications were carried out using an iCycler iQ Multicolor real-time PCR detection System (Bio-Rad) with a thermocycle profile as follows: stage 1 consisted of a cycle at 95°C (3 min), and stage 2 consisted of 30 cycles of 95°C (45 s), 67°C (45 s), and 72°C (45 s) for A2 and 95°C (35 s), 57°C (35 s), and 72°C (35 s) for lambda. The data obtained were recorded as threshold cycle (CT) values; i.e., the cycle number during which the fluorescence signal crossed the threshold set by the manufacturer of the thermocycler.
RESULTS
Evaluation of the genotoxic effect of mitomycin C.The degree of liberation of prophage genomes from the bacterial chromosome as a consequence of induction by different concentrations of mitomycin C is shown in Fig. 1a and 2a for lambda and A2 lysogenic cultures, respectively. The generation of new phages (Fig. 1b and 2b) was recorded as well for comparison.
(a) Influence of mitomycin C on generation of autonomous lambda genomes by a lysogenic culture of E. coli W3110 as measured by real-time PCR of the attP segment. Note that in the ordinate axis, the CT values are depicted in descending order to make a more intuitive representation (the lower the CT value, the higher the initial concentration of the DNA segment being amplified). (b) Production of viable phage lambda virions from the same cultures. •, no drug; ▪, 37.5 nM; ▴, 300 nM; □, 1.5 μM; ○, 9 μM; ▵, 30 μM; ×, 60 μM. Error bars represent the standard deviations of three independent experiments.
(a) Influence of mitomycin C on the generation of autonomous A2 genomes by a lysogenic culture of L. casei ATCC 393 as measured by real-time PCR of the attP segment. (b) Production of viable phage A2 virions from the same cultures. •, no drug; ▪, 37.5 nM; ▴, 150 nM; ×, 900 nM; ○, 9 μM; □, 15 μM; ▵, 30 μM.
Drug concentrations as low as 37.5 nM induced significant responses for both organisms. However, the behaviors of the cultures were different. In the case of E. coli, there was a continuous increase in the detection of free phage lambda DNA, which is indicative of a gradual induction of the whole population. Further increases in the concentration of mitomycin C made the induction process faster up to 30 μM. When 60 μM was used, the induction rate started to decline, probably as a consequence of the inhibitory effect of the genotoxic compound on DNA synthesis. In the case of L. casei, a plateau was reached, which became higher as the concentration of the drug was increased, probably indicating that induction was not achieved in the whole population until the cultures were treated with 900 nM mitomycin C, above which induction of A2 was progressively impeded, the CT values obtained at 30 μM being similar to those of the uninduced control. Concordant data were obtained when bacteriophage release was recorded (Fig. 1b and 2b), although this process seemed to be more susceptible to the toxic effect of mitomycin C than DNA liberation; for example, no A2 production above the uninduced control was observed at 15 μM mitomycin C, while significant generation of free phage DNA was still observed (Fig. 2a and b).
Effect of UV radiation on prophage induction.Even the lowest intensity used (1 J/m2) induced the liberation of the phage genomes to a level comparable to the values obtained with the most inducing concentrations of mitomycin C (Fig. 3a for lambda and b for A2). Radiation intensities of 30 J/m2 or more lowered the final degree of induction of A2 but did not affect the final outcome for lambda, although the kinetics of induction were slower from 10 J/m2 on.
Effect of UV radiation (260 nm) on generation of autonomous lambda (a) and A2 (b) genomes by lysogenic cultures of E. coli W3110 and L. casei ATCC 393, respectively, as measured by real-time PCR of their attP segments. •, 0 J/m2; ▪, 1 J/m2; ▵, 10 J/m2; ×, 30 J/m2; ○, 50 J/m2; □, culture treated with 9 μM (E. coli) or 0.9 μM (L. casei) mitomycin C.
Differential effect of other genotoxic compounds.The inducing abilities of diverse genotoxic molecules are shown in Table 1. Some of them, such as epichloridrin, mediated DNA liberation of both phages (although the maximal induction was reached at 50 nM and 5 μM for lambda and A2, respectively), while others exerted their effect on only one of them. For example, DNA gyrase inhibitors such as ciprofloxacin (but also nalidixic acid and norfloxacin) were effective elicitors of the SOS response for E. coli but did not have a significant effect on L. casei (Fig. 4a and b, respectively), presumably because the DNA gyrase of this last organism is not recognized by these drugs. In accordance with this, L. casei, but also more than 50 other Lactobacillus strains tested, is completely resistant to quinolones (unpublished data). The degree of induction of lambda was dose dependent and gave plateaus similar to those resulting from mitomycin C treatment of L. casei (Fig. 2a). This may indicate that E. coli is more permeable to quinolones than to mitomycin C, which is corroborated by the very low concentrations of ciprofloxacin (1 to 10 nM) that induced lambda DNA excision. It was also noted that prophage liberation occurred later than with mitomycin C, possibly reflecting the indirect effect of quinolones on DNA synthesis. Conversely, other compounds, such as etopoxide, preferentially induced the lytic cycle of phage A2 (Fig. 4a versus b). Again, the degree of induction was dose dependent, and in general, it was slower than that with mitomycin C.
Influence of ciprofloxacin and etopoxide on generation of autonomous lambda (a) and A2 (b) genomes from pure cultures of lysogenic E. coli W3110 and L. casei ATCC 393, respectively, as measured by real-time PCR of their attP segments. •, cultures with no antibiotic; □, cultures treated with 9 μM (E. coli) or 0.9 μM (L. casei) mitomycin C. Concentrations of etopoxide are indicated as follows: ▵, 500 nM; +, 1.5 μM; ×, 5 μM. (a) Concentrations of ciprofloxacin are indicated as follows: ⧫, 1 nM; ▪, 10 nM; ▴, 50 nM; —, 500 nM. (b) Concentrations of ciprofloxacin are indicated as follows: ⧫, 50 nM; ▪, 500 nM; ▴, 5 μM.
Degree of induction of A2 and lambda genome liberation elicited by a series of genotoxic agents compared with mitomycin C
Detection of genotoxic effects in mixed cultures of lysogenic E. coli and L. casei.The mixed cultures were incubated without forced aeration at 30°C to favor the development of L. casei and in liquid 2× YT medium to avoid the inhibitory effect of MRS medium on the growth of E. coli. Under these conditions and at the densities used for the experiments (optical density at 600 nm of 0.1), no interaction of one organism with the other was observed: the viable counts were similar to those obtained from individual cultures (data not shown). Discrimination of the effect of any genotoxic compound on the response of each organism was provided by the use of specific pairs of oligonucleotides for lambda and A2, respectively, and, when phage production was being scored, by the susceptible bacteria on which plaques were generated.
When mitomycin C was added to mixed cultures, a continuous increase in the lambda signal was observed, with the slopes of the curves being steeper as the drug concentration was raised. However, phage induction took longer and was not as pronounced as it was in aerated cultures (compare Fig. 1a and Fig. 5a), possibly reflecting a limitation in the energy available to the cells due to the virtually anaerobic atmosphere created by the static cultivation. In the case of A2, the responses to different concentrations of the drug were similar to those described above (see Fig. 2a and 5b for comparison); i.e., plateaus indicative of partial induction or toxic effects were obtained after an initial increase of the response. These data (and others obtained with different genotoxic compounds) indicate that the procedure may be simplified to allow the simultaneous testing of the genotoxic effects occurring on both lysogenic cultures.
Effect of mitomycin C on generation of autonomous lambda (a) and A2 (b) genomes in mixed lysogenic cultures of E. coli W3110 and L. casei ATCC 393 as measured by real-time PCR of their attP segments. •, no antibiotic; ▪, 900 nM; ▴, 1.5 μM; ⧫, 3 μM.
DISCUSSION
Combinatorial chemistry is used to generate new compounds, such as antimicrobials or anticancer agents, from lead molecules (11). Exploitation of the resulting libraries of compounds is limited by the problems of testing a high number of products in a short time at a low cost. Complementarily, the ever-increasing concern about environmental pollution also requires methods to monitor the presence of xenobiotics and their evolution once bioremediation measures are put into place (9, 17). Surprisingly, most methods to test the potential genotoxicity of new compounds or contaminated media that are in use today were developed more than 20 years ago (2, 6, 15). This is a nice indication of their reliability and their inexpensiveness. However, the application of new technologies, such as quantitative PCR, might help to improve the detection of genotoxic activities. In this report, we describe a procedure that uses lysogens of gram-positive (L. casei) and gram-negative (E. coli) bacteria to measure genotoxicity by checking the excision of phage DNA from the genomes of lysogenic hosts. The methodology is based on two principles. DNA excision happens early after the induction of the SOS response and does not depend on de novo DNA synthesis. Consequently, it should be less susceptible to the possible inhibitory effect of the genotoxic compound being tested than the production of new phages (18) or the generation of mutant bacteria (2, 6). This principle was proven after observing that concentrations of mitomycin C that did not allow the generation of A2 virions still induced significant liberation of its DNA from the L. casei genome (Fig. 2a and b). The second principle is based on the different compositions and structures of the cell walls of gram-positive and gram-negative bacteria. All microbial tests in use employ Salmonella enterica serovar Typhimurium or E. coli strains, whose external membranes might act as an effective barrier to some compounds, especially those that are hydrophilic or of high molecular mass. However, these products are more likely to pass easily through the multiple pores left by the peptidoglycan-based wall of gram-positive bacteria, thus widening the range of products susceptible to testing (it is well known, for example, that several antineoplasic drugs are effective inhibitors of the DNA metabolism of gram-positive but not of gram-negative bacteria due to their inability to reach the cytoplasm of the latter organisms). The results obtained indicate that this may be the case. Several of the compounds tested elicited the SOS response in one microorganism but not in the other. Even in the cases of some molecules that operated on both organisms, such as mitomycin C, the data obtained indicate that entry into L. casei was easier than entry into E. coli.
The method was developed with wild-type strains because they tend to be more robust than mutants with weakened walls to make them more permeable, but it can be set, in principle, with any lysogenic bacterium whose attB/attP sequences are known. This may be very useful in particular situations such as xenobiotic testing of very polluted samples, because bacteria that stand their inherent toxicity might be used. In addition, no pathogenic or recombinant microorganisms were used to improve the security and general acceptability of the test (this contrasts with mutant generation methods, which are based on S. enterica serovar Typhimurium strains, and with those that measure enzymatic activities, where the bacteria usually harbor a recombinant plasmid in which an SOS-inducible promoter is placed in front of the reporter gene).
In addition, application of the methodology is simple. The whole treatment can be made in a single tube containing a mixed culture of both bacteria, and real-time PCR can be done simultaneously by using probes specific for each phage DNA labeled with different fluorophores. Besides, the data obtained are quantitative: they can be matched with those obtained for a reference molecule (such as mitomycin C or a lead compound in the case of combinatorial libraries) or compared with a plot obtained with known concentrations of the compound being tested. Finally, the whole determination, from setting the cultures to completion of the real-time PCR, can be finished in 5 to 7 h, i.e., less than a working day.
ACKNOWLEDGMENTS
This work was supported by CICYT grant SAF2004-0033 from the Ministry of Science and Technology (Spain) and the FEDER Plan. N.S. and R.M. are holders of a fellowship associated with this grant and a scholarship from FICYT (Principado de Asturias), respectively. The support of Mercadona SA is also acknowledged.
We thank K. F. Chater for critical reading of the manuscript and M. Sierra for providing some of the genotoxic compounds used.
FOOTNOTES
- Received 21 February 2007.
- Accepted 22 February 2007.
- Copyright © 2007 American Society for Microbiology
