L-arabinose Resistance Research Articles

A season-dependent variation of genotoxicity of surface water samples from Taihu Lake, Yangzte delta.

The present study is to assess a year round change of genotoxicity of surface water in Taihu Lake, Yangtze delta. The water samples were collected in spring, summer, autumn and winter of 2001, respectively. The bacterial L-arabinose resistance forward mutation test (ara test), along with the Ames test, was performed for the evaluation of time profile of mutagenicity of the surface water samples. The strong genotoxic potentials were detected in the near-shore area, especially in the vicinity to the major city. All the samples collected from the lake center and areas distant from dense human activity showed genotoxicity-free in year round time. Intensified human activity along the lake may play an important role for the genotoxic pollution. By comparison of the samples from year round, the winter samples showed the strongest mutagenic potential of all seasons, which indicates that meteorological and hydrological factors also affect the water quality. The impact of dense human activity, improper management and season change should be put into consideration for designing an effective remediation project in the lake basin area.

The Mutagenicity Study of Surface Water from Meiliang Bay, Taihu Lake in Winter By E.coli UC1121 L-arabinose Resistance Forward Mutation Test

The L-arabinose resistance forward mutation test (ara test), together with Ames test, was performed for evaluation of mutagenicity of the surface water samples collected from Meiliang Bay, Taihu Lake in winter. The five of the nine surface water samples were detected as mutagen positive by ara test, most of them were collected from the estuary which flow into the lake or from the edge of shore area. The same tendency was observed with the involvement of mammalian S9 metabolic activation system. The roughly, coincident results were confirmed by Ames test when the sample doses applied were enhanced to two and a half times compared with that of ara test. The data suggest that detection sensitivity of ara test is higher if intricate environmental samples are concerned. Thus, prospect for ara test to be the primary choice is prompted when the genotoxicity screening of environmental mixtures is required.

Ogt alkyltransferase enhances dibromoalkane mutagenicity in excision repair-deficient Escherichia coli K-12.

We examined the role of the O6-alkylguanine-DNA alkyltransferase encoded by ogt gene in the sensitivity of Escherichia coli to the mutagenic effects of the dibromoalkanes, dibromoethane and dibromomethane, by comparing responses in ogt- bacteria to those in their isogenic ogt+ parental counterparts. The effects of the uvrABC excision-repair system, the adaptive response, mucAB and umuDC mutagenic processing, and glutathione bioactivation on the differential responses of ogt- and ogt+ bacteria were also studied. Mutation induction was monitored by measuring the frequency of forward mutations to L-arabinose resistance. Induced mutations occurred only in excision repair-defective strains and were totally (with dibromomethane) or substantially (with dibromoethane) dependent on the alkyltransferase (ATase) encoded by the ogt gene. An increased mutagenic response to both dibromoalkanes was also seen in ogt- bacteria that overexpressed the ogt protein from a multicopy plasmid, indicating that the differences in mutability between ogt+ and ogt- bacteria were not dependent on the ogt- null allele carried by the defective strain. The ATase encoded by the constitutive ogt gene was more effective in promoting dibromoalkane mutagenicity than the ada ATase induced by exposure to low doses of a methylating agent. The mutagenicity promoted by the ogt ATase was dependent on both glutathione bioactivation and SOS mutagenic processing. To our knowledge, this paper presents for the first time evidence that DNA ATases, in particular the ATase encoded by the ogt gene, can increase the mutagenic effects of a DNA-damaging agent. The mechanism of this effect has yet to be established.

Null thioredoxin and glutaredoxin Escherichia coli K-12 mutants have no enhanced sensitivity to mutagens due to a new GSH-dependent hydrogen donor and high increases in ribonucleotide reductase activity

This work investigates whether a mutator phenotype is associated to the simultaneous deficiency in thioredoxin and glutaredoxin, the two known hydrogen donors of ribonucleotide reductase. To this end, new Escherichia coli K-12 strains carrying delta trxA and/or grx::kan null mutations were constructed to monitor mutagenesis by selecting forward mutations to L-arabinose resistance. Highly sensitive and specific enzyme-linked immunoassays were developed to confirm that trx-grx- cells lacked thioredoxin and glutaredoxin. A number of remarkable properties were observed in the newly constructed thioredoxin- and glutaredoxin-deficient bacteria compared with the wild type cells. Thus, they (i) grew on minimal medium plates, suggesting that the presence of thioredoxin and glutaredoxin may not be absolutely essential for sulfate reduction; (ii) showed normal mutagenic sensitivities toward a wide variety of DNA-damaging agents, as compared with wild type cells and trx- or grx- single mutants; (iii) displayed 14% of GSH-dependent and 30% of NADPH-dependent ribonucleotide reduction capacity with CDP as substrate in the presence or the absence of exogenous ribonucleotide reductase, respectively; and (iv) showed very high levels of ribonucleotide reductase activity, which was increased from 19- to 23-fold. The existence of a new glutathione-dependent hydrogen donor for ribonucleotide reductase and the high activity levels of this enzyme in trx-grx- defective cells could explain that thioredoxin and the first discovered glutaredoxin are not essential for deoxyribonucleotide synthesis, even under mutagenic stress.

Effect of ogt expression on mutation induction by methyl-, ethyl- and propylmethanesulphonate in Escherichia coli K12 strains.

We have previously reported the isolation of an Escherichia coli K12 mutant that is extremely sensitive to mutagenesis by low doses of ethylating agents. We now show by Southern analysis that the mutation involves a gross deletion covering at least the ogt and fnr genes and that no O6-alkylguanine-DNA-alkyltransferase activity is present in cell-free extracts of an ada::Tn10 derivative of these bacteria. Confirmation that sensitisation to ethylation-induced mutagenesis was attributable to ogt and not to any other loci covered by the deletion was obtained by constructing derivatives. Thus an ogt::kanr disruption mutation was introduced into the parental ogt+ bacteria, and the ogt::kanr mutation was then eliminated by cotransduction of ogt+ with the closely linked Tetr marker (zcj::Tn10). The delta(ogt-fnr) deletion or ogt::kanr disruption mutants were highly sensitive to ethyl methanesulphonate-induced mutagenesis, as measured by the induction of forward mutations to L-arabinose resistance (Arar). Furthermore, the number of Arar mutants increased linearly with dose, unlike the case in ogt+ bacteria, which had a threshold dose below which no mutants accumulated. Differences in mutability were even greater with propyl methanesulphonate. Overproduction of the ogt alkyltransferase from a multicopy plasmid reduced ethylmethanesulphonate-induced mutagenesis in the ogt- mutant strains and also methylmethanesulphonate mutagenesis in ada- bacteria. A sample of AB1157 obtained from the E. coli K12 genetic stock centre also had a deletion covering the ogt and fnr genes. Since such deletions greatly influence the mutagenic responses to alkylating agents, a survey of the presence of the ogt gene in the E. coli K12 strain being used is advisable.

Fpg protein protects Escherichia coli K-12 from mutation induction by the carcinogen 4-nitroquinoline 1-oxide.

This paper investigates the role of the fpg gene product in protecting Escherichia coli cells against the lethal and mutagenic effects of 4-nitroquinoline 1-oxide (4NQO). To this end, the araD81 mutation which make the cells sensitive to L-arabinose was combined with an fpg-1::Knr allele, in either an uvrA+ or uvrA-, umuC+ or umuC-, genetic background. Mutation induction was monitored by selecting forward mutations to L-arabinose resistance (Arar). The formamidopyrimidine-DNA glycosylase (Fpg protein) protected bacteria from 4NQO-induced mutagenesis since Fpg- defective cells showed greater Arar mutation induction than fpg+ bacteria did. This was confirmed since the increased sensitivity of the fpg- cells to mutagenesis by 4NQO was suppressed when the Fpg protein was overproduced by placing the fpg gene in a multicopy plasmid vector. The fpg- mutation had no detectable influence on 4NQO mutagenesis in a uvrA- genetic background, but its effect was magnified in umuC- cells. No influence on cell survival was observed after 4NQO treatment. Our data suggest that 8-hydroxyguanine, a non-lethal, non-bulky and directly miscoding lesion, might be responsible for the detected influence of Fpg protein expression on mutation induction by 4NQO. This is in agreement with the reported in vivo formation of 8-hydroxyguanine in cellular DNA after 4NQO exposure. The increased 4NQO-induction of GC to TA transversions on fpg- bacteria further support such a possibility. This work reinforces the role of Fpg protein in the bacterial defense against the mutagenicity by genotoxic agents.

A method for selection of forward mutations in supF gene carried by shuttle-vector plasmids.

The supF gene of Escherichia coli has been widely used as a mutagenic target in several shuttle-vector plasmids. Mutations in this gene are usually screened by a colony colour assay based on the suppression of a lacZ amber mutation in an appropriate bacterial indicator strain. This screening method cannot measure the low mutation frequencies usually detected in prokaryotes, and therefore precludes the use of supF gene for studying mutational spectra in bacteria. In this paper we report the development of a simple method for the selection of supF forward mutations in shuttle-vector plasmids. The method has implied the construction of an araD- araC(Am) mutant strain (MBL50) of E.coli. The L-arabinose sensitivity caused by the accumulation of a toxic intermediate in araD- mutants is abolished in MBL50 because the araC(Am) mutation blocks the L-arabinose catabolic pathway. Strain MBL50 becomes sensitive to L-arabinose when transformed with a supF+ plasmid but remains resistant upon transformation with a supF- mutant. This new L-arabinose resistance selection method was able to detect supF- mutant fractions up to three orders of magnitude below those determined with the colony colour screening assay. The method was further validated by carrying out in vivo mutagenesis experiments with N-methyl-N-nitrosourea (MNU) and a shuttle-vector-bearing strain (UC2109) completely defective in O6-methylguanine (O6meG) alkyltransferase repair capacity. The DNA sequence alterations of 22 independent supF- mutants induced by MNU were determined. All mutations were G:C-->A:T transitions in agreement with the predicted significance of the mispairing potential of the O6meG lesion. A preference for the sequence 5'-GG-3' was detected, revealing a 5'-flanking base influence. The accumulation of all 22 MNU-induced mutations in three sites of the supF genes might be related to the lack of O6meG alkyltransferase repair capacity of strain UC2109. The L-arabinose resistance method described in this paper allows rapid scoring and sequencing of forward mutations in the supF gene on shuttle-vectors, thus permitting its use as a genetic target for repair and mutagenesis studies in bacteria. Since shuttle-vectors replicate both in bacteria and mammalian cells, this method makes it possible to compare prokaryotic and eukaryotic mutational spectra.

Quercetin and the mutagenicity of wines

Various studies have shown the mutagenicity of red wine. The major mutagens identified in red wine have been flavonoids, i.e. rutin and its aglycone quercetin. Besides flavonoids, however, it has recently been reported that H2O2 may account for the mutagenicity of red wine in the L-Arabinose resistance test. In the present study we report on the role of flavonoids in the mutagenicity of red wine in the Ames assay. Different wines from Portugal and Spain have been tested after concentration in XAD-2 columns in strains TA98 and TA104 of Salmonella typhimurium concurrently with the determination of the respective content of quercetin by HPLC. A similar approach was used for pilot scale productions of red wines. In all cases quercetin could be demonstrated as the major mutagen in red wines. The levels of quercetin in finished wines and during the wine-making process showed a good fit with the levels of mutagenicity detected. Catalase had no effect whatsoever on the mutagenicity of wines in both TA98 and TA104. These results do not rule out a role for H2O2 in the mutagenicity of wines, detected in other genetic end-points, because H2O2 can be formed from the auto-oxidation of quercetin.

Study on the mutagenicity of brandy with the Ara test.

The forward mutation assay to L-arabinose resistance (Ara test) in Salmonella typhimurium was used to demonstrate that evaporated residues of brandy had direct-acting mutagenic activity. The mutagenicity covered a 100-fold range, from 13482 to 127 AraR induced mutants/ml brandy equivalent. Rat liver S9 mix suppressed the mutagenic activity of brandy in the Ara test. The inactivating capacity was independent of microsomal monoxygenase enzymes and appeared to be mediated through a heat stable component of the S9 fraction. Catalase was identified as the putative S9 component responsible for its inactivating capacity. The implication of reactive oxygen species in the direct-acting mutagenicity of brandy was supported by the higher sensitivity of Escherichia coli bacterial strains deficient in two major cellular antioxidant defense (glutathione and/or catalase) compared to their parental wild-type. Phenolic compounds of a polar nature could be responsible for the mutagenicity through the production of reactive oxygen intermediates. Non-matured beverages (gin and non-matured rum) were non-mutagenic. It is conceivable that mutagenic phenolics might be extracted from the wood during maturation in the barrel. Autoxidation of phenolic compounds could be a common mechanism in the mutagenicity of complex mixtures of plant origin.

Influence of S9 mix on the expression of mutants in the l-arabinose resistance test of Salmonella typhimurium

S9 mix produces an effect similar to that of d-glucose in the l-arabinose resistance test of Salmonella typhimurium, releasing the growth inhibition exerted by l-arabinose. Two elements are responsible for this effect: the glucose-6-phosphate present in the cofactors of the S9 mix and the S9 fraction itself. UV light was used as a mutagen to compare the efficiency of S9 mix and d-glucose in allowing phenotypic expression of mutants in selective plates with l-arabinose; 0.5 ml of S9 mix per plate showed and efficacy similar to that of 0.5 mg of d-glucose per plate. To verify that the S9 mix is equivalent to d-glucose traces in selective plates with respect to the number of induced mutants in compounds requiring metabolic activation, we utilized 2 direct-acting nitrofurans. Our conclusion is that activation of agents could be erroneously attributed to the S9 mix, when plates with 0.5 mg of d-glucose are compared to plates with 0.5 ml of S9 mix plus 0.5 mg of d-glucose. Our results suggest that d-glucose traces be omitted in experiments requiring the presence of the S9 mixture.

Quantitative relationship between mutagenic potency in the Ara test of Salmonella typhimurium and carcinogenic potency in rodents. A study of 11 direct-acting monofunctional alkylating agents

This work attempted to derive a quantitative relationship between mutagenicity and carcinogenicity by examining the association between mutagenic potency in the Ara test of Salmonella typhimurium and carcinogenic potency in rodents. Mutagenesis was monitored by selecting forward mutations to L-arabinose resistance. Lethality was measured at equivalent experimental conditions to those of mutant yield by using a mixed population of a pair of isogenic strains distinguished by their differential nutritional requirements. The study was carried out with a group of 11 direct-acting monofunctional alkylating agents, which failed to show any quantitative correlation in the histidine reverse-mutation test. Our data suggest that the mutagenic efficiency of the compounds is directly proportional to the magnitude of the maximum yield of L-arabinose resistance mutants and inversely proportional to the dose and the number of lethal hits at which the maximum yield occurs. A highly significant correlation (r10 = 0.86, P less than 0.01) was found between the mutagenic efficiencies of the compounds in the Ara test and their carcinogenic potencies in rodents, expressed as TD50 ('tumor dose' 50) values. The result suggests that the Ara forward-mutation test of S. typhimurium might be capable of reflecting the relative potency of animal carcinogens, at least when confined to particular chemical classes. A more generic and definitive conclusion about the predictive value of the Ara test would require this analysis to be extended to other types of genotoxic carcinogens.

Mutagenesis in Escherichia coli lacking catalase

Escherichia coli K-12 strains completely lacking catalase activity due to mutations in katG, katE, and katF genes were constructed in order to assess the role of hydrogen peroxide in mutagenesis. Mutagenesis was monitored by selecting forward mutations to L-arabinose resistance. Lethality was measured at experimental conditions equivalent to those of the mutant yield by using a mixed culture of pairs of isogenic strains distinguished by their differential nutritional requirements. Deficiency in katG, katE, and katF genes leads to an enhanced spontaneous mutation rate as well as an enhanced sensitivity to both the lethal and mutagenic effects of hydrogen peroxide or an H2O2-generating mixture of compounds, such as coffee. To compare further the responses of the catalase-deficient bacteria to those of catalase-proficient counterparts, other genotoxins were analyzed. Both catalase-deficient and catalase-proficient strains were equally mutated by MMS, 4-NQO, and ultraviolet light. It is concluded that the bacterial strains and the mutagenicity tests described in the paper represent a useful tool to study the role of H2O2 in mutagenesis.

Mutagenicity of alcoholic beverages in the l-arabinose resistance test with Salmonella typhimurium

Mutagenicity of alcoholic beverages in the l-arabinose resistance test with Salmonella typhimurium

Simple method for precise determination of chemical lethality in the l-arabinose resistance test of Salmonella typhimurium

A simple method is described for the determination of the lethal effects of chemicals assayed with the l-arabinose resistance test of Salmonella typhimurium. The method uses a mixed culture of 2 isogenic bacterial strains which are distinguished on the basis of their different nutritional requirements: sensitivity or resistance to l-arabinose, auxotrophy or prototrophy to histidine and leucine. BA13 (the mutation indicator strain) is the strain for routine screening of mutagens and allows the selection of forward mutation from l-arabinose sensitivity to l-arabinose resistance. BAL13 (the survival indicator strain) is a derivative of BA13. Both bacterial strains are found to be equally sensitive to the lethal effects of mutagens. The method described permits the measurement of cell survival at the same high cell concentration as used in the measurement of the mutant yield and in the same type of minimal medium with l-arabinose and glycerol, except for the histidine supplement in the mutant plates or the leucine in the survival plates.

Study on the mutagenicity of nifurtimox and eight derivatives with the l-arabinose resistance test of Salmonella typhimurium

The mutagenicity of nifurtimox (nfx) and 8 nfx analogues has been investigated with the l-arabinose forward-mutation assay of Salmonella typhimurium. The nfx analogues tested were obtained by replacing the 3-methyl-4-yl-tetrahydro-1,4-thiazine-1,1-dioxide group of the parent compound with the following other groups: indazol-1-yl (1); pyrazol-1-yl (2); benzimidazol-1-yl (3); 1,2,4-triazol-4-yl (4); 1-methyl-3-methylthio-1,2,4-triazol-4-yl-5-thione (5); 3,5-bis(methylthio)-1,2,4-triazol-4-yl (6); 1-adamantyl (7); 4,6-diphenylpyridin-1-yl-2-one (8). The mutagenic activity of each chemical was determined by the standard plate-incorporation test, in the presence or absence of the S9 activation mixture. The 9 compounds were mutagenic and exhibited linear dose—mutagenic response relationships. They were direct-acting mutagens and showed a nearly 1000-fold range in mutagenic potency from chemical 1 to nfx. In most cases, the addition of S9 mixture to the test plates decreased the mutagenicity of compounds. This effect was particularly noticeable in the case of chemical 1–3, 5 and 7 where a more than 70% decrease in mutagenic activity was observed in the presence of the S9 mixture. The mutagenic potency of compounds in the Ara test showed a negative linear correlation with previously reported antitrypanosomal activity. Thus, chemicals 6 and 8 with in vitro activities against Trypanosoma cruzi clearly superior to that of nfx showed 2 of the lowest mutagenic potencies in the Ara test and these were only somewhat higher than the mutagenicity of the reference drug.

Response of the l-arabinose forward mutation assay of Salmonella typhimurium to frameshift-type mutagens

The present study was designed to evaluate the capacity of the l-arabinose resistance test of Salmonella typhimurium in the detection of frameshift-type mutagens. To this end the response of the Ara test was examined with respect to 15 chemicals which had been previously described as able to revert the Ames tester strain TA97. The mutagenicity of each compound was determined by the liquid test under experimental conditions which optimize the mutagenic response of the Ara test with the tester strain BA9. Strain TA97 was used simultaneously with BA9. The Ara forward-mutation assay efficiently detected the mutagenic activity of 14 out of the 15 chemicals assayed. PR toxin was the only compound which gave a weak dose response without doubling the spontaneous mutant level. In comparison with the Ara test, a total of 3 chemicals (HZ, PE and PR toxin) were not found to be mutagenic with strain TA97. In most cases ( the Ara test was comparatively greater than that of strain TA97. Three chemicals (DEO, PRF and 9-AA) were detected with quite similar degrees of sensitivity by both mutation assays. ICR-191, which seems highly specific in reverting frameshift mutations with added cytosines in a run of cytosines, was the only chemical with a lower mutagenic activity in the Ara test than in strain TA97. The results enhance the interest of the l-arabinose forward-mutation assay as an alternative to the set of specific tester strains used by the histidine reverse-mutation assay in massive, general and primary screening for genotoxic agents.

Mutagenicity of red wine in the L-arabinose resistance test with Salmonella typhimurium.

The forward mutation assay to L-arabinose resistance (Ara test) in Salmonella typhimurium detected a Spanish red table wine (Rioja) as a direct-acting mutagen. The best mutagenic response was obtained by preincubating strain BA13 with the wine sample in the presence of sodium phosphate buffer and in the absence of any external metabolic activation. In fact, the S9 mixture abolished most of the mutagenic activity of red wine in the Ara test. Such an inactivating capacity seems to be independent of microsomal enzymes and mediated through some kind of heat-stable component(s) in the S9 fraction. Both regular wine (directly from the bottle) and lyophilized wine were strong mutagens in the Ara test, inducing 4914 and 2739 AraR mutants/ml. Both pKM101 and delta uvrB were critical factors in the detection of the mutagenicity of wine, exhibiting a synergic effect in strain BA13. The mutagenicity of red wine was somehow pH-dependent, increasing with the pH value of the preincubation mixture. In comparison with the Ara test, the His reverse mutation assay (Ames test) was much less sensitive to the mutagenicity of lyophilized red wine, TA102 being the most (448 His+/ml) and TA98 the least (38 His+/ml) sensitive strain. TA100, TA104 and TA97 manifested intermediate mutagenicities to red wine. Previous reports have identified strain TA98 as the His strain most sensitive to the apolar fraction (e.g. XAD-2-bound) of red wine. Based on these results we propose that TA98 mainly detects glycosides of mutagenic flavonols present in red wine (quercetin, rutin, etc.), which do not constitute the major direct-acting mutagens detected with the Ara test.(ABSTRACT TRUNCATED AT 250 WORDS)

Implication of active oxygen species in the direct-acting mutagenicity of tea

The present study shows that the l-arabinose resistance test with Salmonella typhimurium detects that freshly infused tea is highly mutagenic in the absence of mammalian microsomal activation. Both the mutagenesis protocol (preincubation test) and the additional genetic characteristics of the bacterial tester strain (excision repair deficiency, normal lipopolysaccharide barrier and the presence of plasmid pKM101) were critical factors in the optimal induction by tea of forward mutations to l-arabinose resistance. The TA104 strain — a histidine auxotroph specific to oxidative mutagens — was the most sensitive tester strain of the Ames test to the direct-acting mutagenicity of tea. In comparison with strain TA104, the sensitivity of the Ara forward mutation test was 18 times higher, one cup of tea (200 ml) inducing 3 × 10 6 Ara R mutants. More than 90% of the mutagenicity of 150μl of a fresh tea infusion, or that of the equivalent amount (1.32 mg) of the corresponding lyophilized residue, was suppressed by 10 units of catalase. In contrast to catalase, superoxide dismutase was rather ineffective. These results indicate that hydrogen peroxide is produced in tea solutions, playing an essential role in its mutagenicity. In comparison, the role of superoxide anion seems negligible. Like catalase, the chelating agent DETAPAC showed a protective effect with respect to the mutagenicity of tea, suggesting the additional implication of hydroxyl radicals.

Coffee is highly mutagenic in the L-arabinose resistance test in Salmonella typhimurium.

The present study shows that the L-arabinose resistance test in Salmonella typhimurium detects coffee as a strong mutagen in the absence of mammalian microsomal activation. The response of the Ara forward mutation assay was 8.5 times higher than that of TA104, which is the most sensitive to coffee of the tester strains of the Ames test. Both the mutagenesis protocol (preincubation test) and the additional genetic characteristics of the bacterial tester strain (excision repair deficiency, normal lipopolysaccharide barrier, and the presence of plasmid pKM101) were critical factors in the optimal induction by coffee of forward mutations to L-arabinose resistance. All ten samples of roasted coffee analyzed with the Ara assay were highly mutagenic: one cup of coffee (150 ml) was calculated to induce 3-4 X 10(6) AraR mutants. In contrast, coffee prepared from unroasted beans (green coffee) had no mutagenic activity. Regular- and sugar-roasted coffees showed similar mutagenicities, but the specific mutagenic activity of instant coffees (1559 AraR mutants/mg) was almost 2 times that of noninstant ones (834 AraR mutants/mg). The Ara assay allowed the direct testing of coffee, although it was demonstrated that lyophilization has no effect on the mutagenicity of this beverage. Like roasted coffee, roasted barley induced a large number of AraR mutants per mg (227), though its specific mutagenic activity was approximately 4 and 7 times lower than that of noninstant and instant coffees, respectively.

Conditions for the optimal use of the L-arabinose-resistance mutagenesis test with Salmonella typhimurium.

Different conditions of mutagenesis have been compared in order to optimize the use of the L-arabinose resistance test with Salmonella typhimurium. The mutagenesis protocols compared were the plate-incorporation, the pre-incubation and the liquid tests. Fourteen chemicals were used in the comparison: six direct-acting mutagens and eight pre-mutagens. Five concentrations of S9 (3, 7.5, 10, 15 and 33% v/v) were compared in the liquid test with pre-mutagens, and three densities of bacteria (ranging from 10(8) to 10(6) cells) were used in the comparison between the plate-incorporation and the pre-incubation mutagenesis test. In general, the liquid test proved the most sensitive mutagenesis protocol. When carrying out this test in a mass screening of mutagens, we propose to select the L-arabinose-resistant mutants in plates supplemented with 0.5 mg of D-glucose, and to express the mutagenic response as the absolute number of induced mutants. The plate-incorporation and the pre-incubation mutagenesis protocols could be considered as alternative procedures in the case of previous negative results with the liquid test. Two recommendations can be finally made in order to avoid false negative results: (a) a large population of bacteria (ranging from 10(8) to 10(7) cells) must be exposed to the mutagens in both the plate-incorporation and the pre-incubation mutagenesis tests, and (b) ideally, several S9 concentrations (ranging from 3 to 30% v/v) would be employed for testing in liquid samples of unknown mutagenicity.