Communication

A previously developed and highly sensitive umu-microplate test system based on the nitroreductase- and O-acetyltransferase-overproducing strain Salmonella typhimurium NM3009 and the O-acetyltransferase-overproducing strain S. typhimurium NM2009 was applied to the detection of genotoxic activity in atmospheric particles in urban areas using a relatively small sample load. The results showed that the test system was able to detect slight increases in induced genotoxicity in atmospheric particles and that genotoxicity was detected mainly in the fine fraction but also partially in the coarse fraction. The present sensitive microplate test system has potential for application to the screening of various other environmental samples.

cinogenic effects (Goldsmith, 1980).It is, however, difficult to analyse most of the harmful chemicals in particulate samples because of the multiplicity of known and unknown chemicals present in the environmental complex mixture.Rapid biological screening methods are therefore useful for assessing human health risks.Many researchers have used the Ames assay to examine the ability of histidine-requesting strains of Salmonella typhimurium to revert to prototrophic mutations, and the potent mutagenic activities of numerous chemicals have been reported (Ames et al., 1975;Mortelmans & Zeiger, 2000).The Ames assay has also been used for the detection of mutagenicity in various environmental samples and novel field studies suggest that the kinds and the composition of carcinogenic substances present in environmental samples differ depending on the source of pollution (Daisey et al., 1984;Matsumoto et al., 1998).While the Ames assay is thus useful as one of the best short-term bioassay methods for detection of mutagenic activities in various known and unknown samples, it requires over 48 h and relatively large volumes of test samples.
In contrast, SOS/umu-test system has been developed for the rapid detection of DNAdamaging agents.In the system, the umu gene was cloned and a plasmid (pSK1002) carrying a umuC¢-lac¢Z fusion was constructed for the study of the regulation of umu gene expression (Shinagawa et al., 1983).In this fusion gene, the umu operon is induced by DNA-damaging agents and regulated genetically by the recA and lexA genes.The induction of umuC gene expression could be monitored by measuring the cellular b-galactosidase activity produced by the fusion gene after DNA damage.Since the umu operon has a possibility to express mutagenesis more directly than other known SOS genes (Kato & Shinoura, 1977), the umu test system has been applied to rapid screening of environ-mental mutagens and carcinogens (Oda et al., 1985).
Recently, a more sensitive SOS/umu-test system based on the nitroreductase-and O-acetyltransferase-overproducing strain S. typhimurium NM3009 and the O-acetyltransferase-overproducing strain NM2009 has been developed (Oda et al., 1995).These strains have proved to be extremely sensitive towards the genotoxic activity of nitroarenes and aromatic amines, respectively (Yamazaki et al., 1992;Oda et al., 1993), because the bacterial strains deficient in nitroreductase or O-acetyltransferase are resistant to the mutagenic action of several nitroarene compounds (McCoy et al., 1983).We have also developed a rapid and simple high-throughput umu-test system to be useful for the detection of nitroarenes and aromatic amines with quite small volume of chemicals (Oda et al., 2001).
In the present report, we have attempted to apply the microplate SOS/umu-test system based on S. typhimurium NM3009 and NM-2009 to the detection of nitroarenes and aromatic amines present in environmental atmospheric suspended particles in the urban Osaka (Japan).Our objectives were to apply the test system to actual particulate samples and to explore its applicability to other small-amount of environmental samples.

MATERIALS AND METHODS
In order to collect airborne particles, a low-volume Andersen cascade impactor (AN-200; Dylec Co., Tokyo) was operated continuously for approximately 10 days from Nov. 1 to 10 and from Nov. 10 to Nov. 20, 2000, at a site at Hirano in the eastern central area of Osaka, Japan.The sampler has the ability to collect separately on nine filter stages with aerodynamic diameters of more than 11 (stage 0), 7.0-11 (stage 1), 4.7-7.0(stage 2), 3.3-4.7 (stage 3), 2.1-3.3 (stage 4), 1.1-2.1 (stage 5), 0.65-1.1 (stage 6), 0.43-0.65 (stage 7) and less than 0.43 mm (back-up stage).The filter (80 mm f, QR100, silica fiber; Toyo Advantec Co., Japan) was preheated at 800°C for 2 h to avoid background contamination and kept in a room for over 48 h at 20°C and 50% relative humidity before use.At the end of each sampling, the filter of the sampler was removed, kept for over 48 h at 20°C and 50% relative humidity, weighed and then stored at -20°C before pretreatment.A quarter of each filter in two independent sampling days was cut into pieces and extracted with 100 ml of methylene chloride by ultra-sonication for 30 min at 0°C according to a previous method (Sweetman et al., 1982;Ando et al., 1991) with some modification.The extracted solution was filtered with anhydrous sodium sulfate using a cellulose filter (No. 5C; Toyo Advantec Co., Japan).The sample solution was then evaporated, dissolved in hexane, transferred to a handmade PTFE micro-concentration tube, concentrated almost to dryness, and thoroughly dissolved in 20 ml of Me 2 SO in order to obtain a wide range dose-response curve.Aliquots of 4 ml of the sample solution were then carefully diluted with Me 2 SO on a PTFE 96-well microplate.After incubation for 3 h in an incubation shaker (IS-963; Tomy, Japan) with 96 ml of exponential bacterial culture of S. typhimurium NM3009 without S9 mix, or with strain NM2009 plus S9 mix (4.8 : 1), the whole culture was transferred to a new polycarbonate microplate and bacterial growth was assesed as turbidity (A 600 ) for each well measured by microplate reader (Multiskan MS; Labsystems).Then, 10 ml of the culture was transferred to a new microplate and 90 ml of Z-buffer and 50 ml of 0.1% SDS solutions were added to the well, followed by 10 ml of chlorophenol red-b-D-galactopyranoside (Boehringer Mannheim, U.S.A.) solution (4 mg/ml in 0.1 M P-buffer) as a colorimetric reagent for b-galactosidase activity induced by umuC gene expression.The plate was then gently mixed and left to stand at 37°C for 15 min in the absence of S9 mix or for 5 min in the presence of S9 mix.After incubation, 100 ml of stopping solution (1 M Na 2 CO 3 ) was added to each well and gently mixed and b-galactosidase activity (A 570 ) was measured.Relative b-galactosidase activity (units), which reflects relative umuC gene induction, was determined by A 570 /A 600 .The value of a blank filter was subtracted and averaged wide range dose-response curves were obtained for a couple of the independent sampling days.Finally, in order to evaluate the circumstances of the air involving particles, we calculated the average minimal reactive air volume and minimal particulate concentration defined as the dose showing the relative b-galactosidase activity double that induced by Me 2 SO alone.

RESULTS AND DISCUSSION
Figure 1 shows average size distribution of particles collected in the sampling area from Nov. 1 to Nov. 10 and from Nov. 10 to 20, 2000.There seemed to be a temporal bimodal size distribution of particles in early wintertime in this urban area and the concentration of fine particles (less than 2 mm) was 1.6 times higher than that of coarse particles (2 mm or more).Additional analytical results indicated that the fine particles included a large amount of carbonaceous components and water-soluble ionic species (not shown).We therefore speculate that the fine particles in the present sample are temporally affected by various direct anthropogenic activities including diesel exhaust (Ross et al., 1987) and indirect secondary atmospheric conversion of gas to particles (Mizohata & Mamuro, 1980).The new microplate test system based on S. typhimurium NM3009 without S9 mix and NM2009 with S9 mix was applied to the samples fractionized by typical particle size.
Figure 2 shows relative b-galactosidase activity for particulate samples fractionized by particle size (indicated by stage number) as an average value within the two sampling periods, with plots drawn as a function of atmospheric air volume.The average value for relative b-galactosidase activity induced by Me 2 SO alone was 0.61±0.17(n=8).The fine particulate group (stages 5 to 7 and back-up filter) shows a markedly higher level of b-galactosidase activity than the coarse particulate group at the sampling air volumes of over 10 m 3 .The results indicate strong SOS/umugenotoxicity in fine particles and weak genotoxicity in coarse particles in the absence of metabolic activation.The urban atmospheric particulate thus possesses potent umu-genotoxicity present mainly in the fine fraction, where it originates from direct anthropogenic sources such as diesel exhaust particles, and partly in the coarse fraction, where it originates through indirect particle re-suspension (Nicholson, 1988).
The results for the same samples using S. typhimurium NM2009 with metabolic activation by S9 mix are shown in Fig. 3 in the same manner as in Fig. 2. The indirect SOS/umu-genotoxicity was not as pronounced as the direct genotoxicity shown in Fig. 2, but slight increases in b-galactosidase activity were again found in the fine and to a lesser extent in the coarse particle fractions.There seemed to be almost no SOS/umu- The value of a blank filter was subtracted and the plots represent the average of two independent sampling days as a function of air volume.The value of a blank filter was subtracted and the plots represent the average of two independent sampling days as a function of air volume.
genotoxicity in the stage 0 (more than 11 mm), stage 2 (4.7-7.0 mm) and stage 4 (2.1-3.3 mm) fractions.The average relative b-galactosidase activity induced Me 2 SO alone was 2.9 ± 1.3 (n = 8), 4.8 times higher than in direct SOS/umu-genotoxicity.We attributed this to the effect of the S9 mix itself in the assay with metabolic activation.Minimal reactive air volume and minimal concentration, which represent atmospheric circumstances of the particles, are calculated and summarized in Table 1.The minimal concentrations of 1-nitropyrene and 2-aminoanthracene are also given as a positive standard.Minimal reactive air volume and minimal particulate concentration are low in the case of S. typhimurium NM3009 without S9 mix, with the lowest values commonly detected in the back-up filter fraction (< 0.43 mm).The results suggest that strong SOS/umu genotoxicity in the fine fraction de-pends on the presence of nitroarenes in the fine particles although detailed chemical analysis might be necessary.Minimal concentration in the fractionized particulate samples is higher than in the positive standard solutions both with and without metabolic activation.This may indicate that the suspended particles also contain other substances such as non-genotoxic chemicals originating in the earth's crust or elsewhere.
The present report shows that the new highthroughput microplate SOS/umu-test system with S. typhimurium NM3009 and NM2009 is able to rapidly detect slight increases in induced SOS/umu genotoxicity in atmospheric particles.Although it is necessary to investigate seasonal changes of particulate genotoxic capacity adding statistical analysis, the present test system is applicable to other environmental samples such as water, soil and mud with quite small sample amounts.

Figure 1 .
Figure 1.Average size distribution of atmospheric suspended particulate collected in early winter (from Nov. 1 to Nov. 10 and from Nov. 10 to Nov. 20, 2000) at Hirano in central (Osaka, Japan) using an Andersen cascade impactor.

Figure 2 .
Figure 2. Relative b-galactosidase activity in sizefractionized particulate samples using highthroughput microplate umu-test with S. typhimurium NM3009 in the absence of S9 mix.

Figure 3 .
Figure 3. Relative b-galactosidase activity in sizefractionized particulate samples using highthroughput microplate umu-test with S. typhimurium NM2009 in the presence of S9 mix.

Table 1 . Minimal reactive air volume and minimal particulate concentrations of the size fractionized par- ticulate samples umu-microplate test system with S. typhimurium NM3009 without S9 mix and NM2009 with S9 mix.
The values represent the average of two independent sampling days.Minimal concentrations of 1-nitropyrene and 2-aminoanthracene are also shown in the bottom.