Ecotoxicology and Environmental Safety 45, 87}91 (2000)

Environmental Research, Section B

Article ID eesa.1999.1849, available online at http://www.idealibrary.com on

Sensitivity and Significance of Luminescent Bacteria in Chronic ToxicityTesting Based on Growth and Bioluminescence

Georg GellertStaatliches Umweltamt KoK ln, Aussenstelle Bonn, Friedrich-Ebert-Allee 144, D-53113 Bonn, Germany

Received April 28, 1999

This study explored the use of luminescent bacteria (Vibriofischeri) for chronic aquatic toxicity tests. The evaluated inhibi-tion of growth to Cu21, Cr61, Zn21, Hg21, Cd21, Pb21, cetyl-trimethylammonium bromide, 3,4-dichloroaniline, acetone,dimethylsulfoxide, ethanol, nitrobenzene, methanol, and 3,5-dichlorophenol was compared with results from another invest-igation, where the inhibition was determined by bioluminescence.Growth inhibition was found to indicate more reliably the pres-ence of substances with chronic toxic properties than the loss ofbioluminescence. But growth responded weaker to the majorityof the analyzed toxicants than bioluminescence. This must beconnected with the test parameters and the experimental condi-tions. But among growth experiments with freshwater bacteriaspecies the sensitivity of the growth inhibition assay with V.fischeri is competitive when a poor medium is employed. ( 2000

Academic Press

Key Words: chronic toxicity, luminescent bacteria; Vibrio fis-cheri; cell growth; bioluminescence; inhibition.

INTRODUCTION

Among bioassays for the assessment of detrimental e!ectsof pollutants in the aquatic environment, the luminescentbacteria test, according to ISO/CD 11348, which determinesshort-term inhibitory e!ects of water contents on the emis-sion of light, has achieved prominance. In the last decade ithas been established worldwide in water laboratories be-cause it has proved to be extremely sensitive, simple, rapid,inexpensive, and valuable (Steinberg et al., 1995).

Bacteria are useful for toxicity evaluations because theirresponse to chemical water contents is signi"cant, their lifecycle is short, their response to environmental change isquick, and they are readily available and easy to maintain.In light of increasing awareness of the long-term e!ects ofchemical pollutants, it would be useful to develop a lumines-cent bacteria test using a chronic test system and the samespecies, in order to get a complete picture of inhibition.

The use of luminescent bacteria in chronic toxicity testinghas not been well explored as a useful tool in environmental

87

supervision (Sweet et al., 1997). Two di!erent methods ofchronic bioassays with luminescent bacteria have been de-scribed in the scienti"c literature. One is based on theinhibition of cell multiplication. The other, chronic toxicity,is evaluated by the inhibition of bioluminescence. The ob-jective of this study was to explore to what extent thesedi!erent toxicological endpoints a!ect the toxicity level ofwater contents.

MATERIALS AND METHODS

The selected substances belong to two classes of toxicants:heavy metals and organic chemicals. From each class, chem-icals known to have low and high toxic properties wereused.

f Class I (heavy metals): copper (as CuSO4]5H

2O),

cadmium (as CdCl2]H

2O), zinc (as ZnSO

4]7H

2O), mer-

cury (as HgCl2), chromium (as K

2Cr

2O

7and as

Na2Cr

2O

7]2H

2O), and lead (as Pb(C

2H

3O

2)2]3H

2O).

f Class II (organic substances): nitrobenzene(C

6H

5NO

2), 3,5-dichlorophenol (C

6H

4Cl

2O), acetyl-

trimethylammoniumbromide CTAB (C19

H42

BrN), 3,4-dichloroaniline (C

6H

5Cl

2N), dimethylsulfoxide DMSO

(C2H

6OS), acetone (C

3H

6O), ethanol (C

2H

6O), and meth-

anol (CH4O).

All substances were dissolved in distilled water, adjustedto pH 7$0.2 and a salinity of 2% NaCl. In order to achievean homogeneous distribution, the solutions were agitatedup to 48 h at room temperature and then left for 12 h.Decanting guaranteed that only the water-soluble propor-tion was taken as stock solution. Because of possible loss byadsorption (as noted for Zn and Cd), insolubility (as notedfor nitrobenzene), or volatilization, the concentrations of thereference chemicals, were determined before the dilutionseries were prepared. The metals were measured using aninductively coupled plasma emission spectrometer. Thecontents of organic chemicals were checked by gaschromatography analysis.

0147-6513/00 $35.00Copyright ( 2000 by Academic Press

All rights of reproduction in any form reserved.

88 GEORG GELLERT

The chronic toxicity tests were conducted only with thegrowth inhibition assay, using the marine Gram-negativeluminescent bacterium <ibrio ,scheri (DSMZ 7151/NRRLB-11177) as test organism. A complete description ofthe test procedure (performed according to the test protocol,standardized and validated 1996 by DIN) was publishedrecently (Gellert and Stommel, 1999). In an abridged versionthe test runs as follows: an inoculum is introduced ina growth medium (control culture). Similar cultures (testcultures) are prepared in parallel, containing inoculum, nu-trients, and graded amounts of dissolved test substances. Allcultures are incubated for 6$1 h at 203C on microplates ina Labsystems iEMS incubator and continuously agitated ata speed of 1150 rpm. To avoid interactions betweenlipophilic toxicants and the polystyrene material of thecommonly used plates, only microplates made of quartzglass were used (Gellert and Stommel 1999). To restrictfaults by volatility, the plates were covered with standardmicroplate lids without perforation. At the end of the gen-eration time the cell concentrations were measured byphotometrically determining the turbidity (at a wavelengthof 450 nm by a Labsystems iEMS reader). Chronic e!ectsare calculated by quantifying the extent of growth inhibitionin the test cultures in comparison with the control cultures.Each experiment consisted of one control and four testconcentrations, indicating measurable e!ects up to minor100% inhibition. For each tested chemical at least six repli-

TABLChronic Toxicity of Reference Chemicals to V

Chemical

LOEC (mg L~1)

Growth Bioluminescence

Cu2` 7.90 0.020a

Cr6` 2.80 0.035aZn2` 8.80 0.10aHg2` 0.04 0.05aCd2` 3.80 0.06aPb2` 80.0 1.00a

Na2Cr

2O

72H

2O

CTAB`

3,4-DichloroanilineAcetone 1380 600aDMSO* 1870 800cEthanol 1035 6300aNitrobenzene 15.0 1.56cMethanol 3990 2500c3,5-Dichlorophenol 0.70 0.20a

Note. `, CTAB, cetyltrimethylammoniumbromide; *, DMSO, dimethylsua Bulich et al., 1996.b Zieseniss et al., 1995.c Huynh et al., 1995.d Backhaus et al., 1997.

cate tests were carried out. The bioluminescence inhibitionassays were performed in two variations. The essential dif-ference concerned the temperature during the incubationtime. In one procedure it amounted to 153C for 24 h(Zieseniss et al., 1995; Backhaus et al., 1997) and in another,273C for 20 to 24h (Huynh et al., 1995; Bulich et al., 1996).

The results were reported as the quantity of substance inmg L~1 required to reduce growth to 2 standard deviations(LOEC), to 10% (EC

10), 20% (EC

20), and 50% (EC

50) of

the mean growth level, calculated from the control cultures.This convention was used to enable a comparison betweenthe EC values with the data achieved with the biolumines-cence inhibition assay presented in the literature.

RESULTS

Table 1 demonstrates the di!erent grade of sensitivitybetween both assays. For heavy metals the data obtained bythe bioluminescence inhibition assay were lower (i.e., re#ect-ing higher toxicity) than those recorded with the growthinhibition assay. When bioluminescence inhibition wasmeasured at 153C, the toxicity values (as LOEC and EC

10)

were between 26 (for Na2Cr

2O

7]2H

2O) and 395 times (for

Cu2`) lower, while measurements at 273C indicated lowerEC

20values, between 2.6 (for Zn2`) and 119 times (for

Pb2`) lower than the corresponding values evaluated with

E 1. fischeri Evaluated with Di4erent Methods

Toxicological endpoint

EC10

(mg L~1 ) EC20

(mg L~1 )

Growth Bioluminescence Growth Bioluminescence

22.7 5.60b

8.06 0.09b17.4 6.68b

0.09 0.12b14.5 0.36b

191 1.61b

19.7 0.749d

0.50 1.780d 0.59 1.20b5.41 0.549d

8820 30000b2.01 2.20b

lfoxide.

TABLE 2Reproducibility of the Growth Inhibition Assay with V. fischeri

Organic compounds Heavy metal ions

Substrate na CVb Substance n CV

CTAB 9 23.8 Cu2` 9 20.73,4-Dichloroaniline 9 17.9 Cr6` 9 6.9Acetone 6 40.8 Zn2` 9 4.7DMSO 6 8.0 Hg2` 9 5.8Ethanol 6 7.1 Cd2` 9 14.4Nitrobenzene 9 4.4 Pb2` 6 2.4Methanol 6 29.2 Na

2Cr

2O

7]2H

2O 9 6.3

3,5-Dichlorophenol 9 11.5

a number of determinationsb CV, coe$cient of variation"standard variation; mean]100.

LUMINESCENT BACTERIA IN CHRONIC TOXICITY TESTING 89

the growth inhibition assay. The only exception was Hg2`.For this ion all applications indicated similar toxicity levels.

In the case of the organic substances the di!erence intoxicity was not so obvious. The sensitivity of both methodswas found to be nearly of the same order of magnitude.Methanol, acetone, DMSO, and ethanol did not displaytoxicity to <. ,scheri (LOEC'500 mg L~1), regardless ofthe methodology used. Remarkable di!erences in toxicitycould be noted for nitrobenzene and 3,4-dichloroaniline.For these toxic chemicals growth inhibition seemed to bea 10 times weaker toxicological signal than inhibition ofbioluminescence. The results produced by the growth inhi-bition assay were highly reproducible. The precision is re-#ected in the coe$cient of variation (CV) data, presented inTable 2, which were obtained for the tested substances. For10 of 15 compounds the CV did not exceed 15%. The highCV for acetone is likely a result of variable loss rates on themicroplates caused by volatility.

DISCUSSION

The lower sensitivity of the growth inhibition assay forthe majority of the tested heavy metal ions can be attributedto the toxicological endpoint and the conditions used. Cellmultiplication may activate fewer metabolic pathways with-in the organism, which are available for inhibition, than cellmultiplication and luciferase production. On the part of thetest conditions, it must be assumed that the sensitivity ofgrowth inhibition is weakened by the presence of nutrientbroth, which masks toxicity by complexation. This phenom-enon could be noted for Cu2`, Zn2`, Cd2`, and Pb2`. Therelatively high toxicity values for nitrobenzene and 3,4-dich-loroaniline, generated by the growth inhibition assay, mustbe linked to the test conditions. The necessity to agitate thecultures constantly (in contrast to the bioluminescence inhi-bition assay) may have caused evaporation, which led toconstant decreases in concentration during incubation. Ap-

parently the microplates could not be covered as hermeti-cally as needed to achieve conditions of a closed test system.Sensitivity is an important criterion in the selection of chro-nic toxicity tests. In this report the application of bio-luminescence inhibition turned out to be a much moresensible method for nearly all substances than growth inhi-bition. But concerning the meaningfulness of bioassays,there are further aspects to be considered.

f The growth inhibition assay allows toxicants to a!ectall metabolic aspects of cell growth and division throughoutthree to four cell cycles. The bioluminescence inhibitionassay, however, re#ects more the activities of the electrontransport system. Its reproduction rate amounts approxim-ately only to 1.5. Therefore, growth inhibition must beconsidered more ecologically relevant for the assessment ofenvironmental risks than inhibition of bioluminescence.

f Growth rate was found to be the metabolic test (incontrast to experiments based on intracellular ATP leveland respiration rate) that always responded negatively totoxicants (regardless of chemical class), that did not providemisleading indications (Pill et al., 1991), and that achievedreliable results (Table 2).

f Finally, the growth inhibition assay o!ers the advant-age of a short test duration. The incubation period of6$1 h does not allow adaptations to the toxicants and,subsequently, regrowth.

To form an idea of the sensitivity of the growth inhibitionassay with <. ,scheri, the obtained results are compared inTable 3 with results evaluated with the same method butwith other species of bacteria. It should be emphasized thatthere was remarkable agreement between the results of theseexperiments with <. ,scheri and those of two other laborat-ories. Toxicity data for 3,5-dichlorophenol, Zn2`, and Cd2`

are examples supporting this conclusion.For many substances the toxic response of <. ,scheri did

not agree well with those generated with freshwater bac-teria. Apart from Hg2`, the EC

50values for heavy metal

ions, obtained with <. ,scheri, were higher. This fact mustbe connected with the addition of NaCl as osmotic regula-tor. The saline milieu is responsible for the higher toleranceto Cd2`, Pb2`, and Cu2` (Carlson-Ekvall et al., 1995) andto Zn2` (Ankley et al., 1990). Also, the sample pH adjust-ment to a neutral range, which is part of the test prepara-tion, may have contributed to the detection of toxicity ata lower level. Other than Dutka and Kwan (1981), who usedP. -uorescens at a pH of 6.7, the cited authors performedtheir experiments without pH adjustment. Di!erences in pHof 1 or 2 units can drastically change the speciation ofcertain metals and result in a change in bioavailability. ForCr6`, Villaescusa et al. (1997) proved that its lowest toxicitycan be noted at the pH range 7.0 to 7.5.

Sensitivity to heavy metal ions can be raised byusing a minimal medium. As recently reported (Gellert andStommel, 1999), Cu2` was 11.9 times and Zn2` 2.8 times

TABLE 3EC50 Values (mg L21) of Anorganic and Organic Substances on Growth of Vibrio fischeri and on Other Bacterial Species Reported

in the Literature

Chemical Growth inhibition with Bacterial growth inhibition with<. ,scheri 6 (h) other single species

Present study Values from P. putida P. -uorescens P. -uorescens A. hydrophila P. aeruginosaother studies 6 hf; 16 h 16}18h 2h 18h 18}20h

Cu2` 36.8 39.0a 0.30a 12.7b 18.0c 21.0d

0.61e 17.0d

1.05f

Cr6` (as K2Cr

2O

7) 15.9 6.30h 8.20i 2.7a 2.20c 2.28g

Zn2` 24.7 26.5h 28.7i 2.10e 360d 4.10c 500d

7.15b

17.3a

Hg2` 0.12 0.39a 0.03a 0.03d 0.05d

Cd2` 28.9 18.1h 25.4i 0.40e 8.2c

0.72f

1.58a

Pb2` 272 460h 14.0d 140c 710d

CTAB` 0.81 2.66g

Acetone 14910 4385f 22000c

DMSO* 25470 26000e

Ethanol 13410 10500e 24000c

Nitrobenzene 380 '100i

3,5-Dichlorophenol 3.20 2.8a 3.60i 16.1e

Note. `, CTAB, cetyltrimethylammoniumbromide; *, DMSO, dimethylsulfoxide.a Schmitz et al., 1998.b Torslov, 1993.c Paran et al., 1990.d Dutka and Kwan, 1981.e Reinke et al., 1995.fSlabbert, 1986.g Huber et al., 1994.h Schmitz et al., 1999.iSpillmann, 1995.

90 GEORG GELLERT

more toxic in a medium containing only yeast extract as thenutrient component than in a complete de"ned medium.For the organic compounds there is a fairly good agreementbetween all bioassays. On the part of the test conditions, thismay be due to the fact that interactions between lipophilictoxicants and NaCl or nutrient components can normallybe excluded.

CONCLUSIONS

The evaluated data and analyses lead to the followingconclusions:

1. The results of this study indicate that both assays arepotentially useful for the detection of chronic chemicaltoxicity.

2. Growth seems to be a less sensible toxicological end-point than bioluminescence. On the conditions part thismust be ascribed to complexation mechanisms, caused bynutrient components and to evaporation supported by the

test conditions. When a poor medium is employed, thegrowth inhibition assay with <. ,scheri is competitive withtests using freshwater bacteria species.

3. The advantage of the growth inhibition assay is itsgreater ecological signi"cance for the detection of hazard-ous water contents, its sensitivity in comparison with otherprocedures based on cell multiplication, and its precisionwithin one and between di!erent laboratories.

4. The growth inhibition assay can play a vital role in thetoxicity assessment of industrial wastewater. The saline en-vironment suppresses extraneous bacterial growth, thusminimizing false negative results.

REFERENCES

Ankley, G. T., Peterson, G. S., Amato, J. R., and Jensen, J. J. (1990).Evaluation of sucrose as an alternative to sodium chloride in the Micro-tox assay: Comparison to "sh and cladoceran tests with freshwatere%uents. Environ. ¹oxicol. Chem. 9, 1305}1310.

LUMINESCENT BACTERIA IN CHRONIC TOXICITY TESTING 91

Backhaus, T., Froehner, K., Altenburger, R., and Grimme, L. H.(1997). Toxicity testing with <ibrio ,scheri: A comparison between thelong term (24 h) and the short term (30min) bioassay. Chemosphere 35,2925}2938.

Bulich, A., Huynh, H., and Ulitzur, S. (1996). The use of luminescentbacteria for measuring chronic toxicity. In ¹echniques in Aquatic ¹oxi-cology (G. K. Ostrander, Ed.). CRC, Boca raton, FL.

Carlson-Ekvall, C. E. A., and Morrison, G. M. (1995). Contact toxicity ofmetals in sewage sludge: Evaluation of alternatives to sodium chloride inthe microtox assay. Environ. ¹oxicol. Chem. 14, 17}22.

DIN 38 412 - L 37 (1996). Determination of the Inhibitory E+ect of =aterConstituents on Bacteria (Photobacterium phosphoreum Cell Multiplica-tion Inhibition ¹est).

Dutka, B. J., and Kwan, K. K. (1981). Comparison of three microbialtoxicity screening tests with the microtox test. Bull. Environ. Contam.¹oxicol. 27, 753}757.

Gellert, G., and Stommel, A. (1999). In#uence of microplate material on thesensitivity of growth inhibition tests with bacteria assessing toxic sub-stances in water and waste water. Environ. ¹oxicol. =ater Qual. 14, inpress.

Gellert, G., Stommel, A., and Trujillano, A. (1999). Development of anoptimal bacterial medium based on the growth inhibition assay with<ibrio ,scheri. Chemosphere 39, 467}476.

Huber, H. Ch., Ritter, U., Huber, W., and MuK cke, W. (1994): Photomet-rische Bestimmung der ZytotoxizitaK t von Chemikalien. Z. Umweltchem.OG kotox. 6, 251}253.

Huynh, H., Bulich, A., and Ulitzur, S. (1995). Measuring Chronic ¹oxicity;sing ¸uminescent Bacteria, Azur Environmental, Carlsbad, CA.

ISO/CD 11348 (Revised version, water quality) Determination of the Inhibit-ory E+ect of =ater Samples on the ¸ight emission of <ibrio ,scheri(¸uminescent Bacteria ¹est) 1994.

Paran, J. H., Sharma, S., and Qureshi, A. A. (1990). A rapid and simpletoxicity assay based on growth rate inhibition of Pseudomonas -uores-cence. ¹ox. Assess. 5, 351}365.

Pill, K. G., Kupillas, G. E., Picardal, F. W., and Arnold, R. G. (1991):Estimating the toxicity of chlorinated organic compounds using a multi-parameter bacterial bioassay. Environ. ¹oxicol.=ater Qual. 6, 271}291.

Reinke, M., Kalnowski, G., and Dott, W. (1995). Evaluation of an auto-mated, miniaturized Pseudomonas putida growth inhibition assay. <om=asser 85, 199}213.

Schmitz, R. P. H., Kretkowski, C., EisentraK ger, A., and Dott, W. (1999).Ecotoxicological testing with new kinetic Photorhabdus luminescensgrowth and luminescence inhibition assays in microtitration scale.Chemosphere 38, 67}78.

Schmitz, R. P. H., EisentraK ger, A., and Dott, W. (1998). Miniaturized kineticgrowth inhibition assays with <ibrio ,scheri and Pseudomonas putida(application, validation and comparison). J. Microbiol. Methods. 31,159}166.

Slabbert, J. L. (1986). Improved bacterial growth test for rapid watertoxicity screening. Bull. Environ. Contam. ¹oxicol. 37, 565}569.

Spillmann, B. (1995). <ergleich verschiedener Biotestsysteme zur Bestim-mung der Hemmung von ¸ichtemission und Zellvermehrung bei ¸euchtbak-terien. Diplomarbeit der Fachhochschule Rheinland-Pfalz, Binngen,Germany.

Steinberg, S. M., Poziomek, E. J., Englemann, W. H., and Rogers, K. R.(1995). A review of environmental applications of bioluminescencemeasurements. Chemosphere 30, 2155}2197.

Sweet, L. I., Travers, D. F., and Meier, P. G. (1997). Chronic toxicityevaluation of wastewater treatment plant e%uents with bioluminescentbacteria: A comparison with invertebrates and "sh. Environ. ¹oxicol.Chem. 16, 2187}2189.

Torslov, J. (1993). Comparison of bacterial toxicity tests based on growth,dehydrogenase activity and esterase activity of Pseudomonas -uorescens.Ecotoxicol. Environ. Saf. 25, 33}40.

Villaescusa, I., Marti, S., Matas, C., Martinez, M., and Ribo, J. M. (1997).Chromium (VI) toxicity to luminescent bacteria. Environ. ¹oxicol. Chem.16, 871}874.

Zieseniss, K., and Grabert, E. (1995). LUMIStox Leuchtbakterientest-Bestimmung der chronischen ToxizitaK t. BIOforum 18, 455}457.