BACTOX, a Rapid Bioassay That Uses Protozoa To Assess the ...

2 downloads 0 Views 2MB Size Report
A new type of toxicity test based on the protozoan Tetrahymena pyriformis ..... An early life-stage test with Daphnia magna Straus: an alternative to the 21-day.
APPLIED AND ENVIRONMENTAL MICROBIOLOGY, June 1999, p. 2754–2757 0099-2240/99/$04.0010 Copyright © 1999, American Society for Microbiology. All Rights Reserved.

Vol. 65, No. 6

BACTOX, a Rapid Bioassay That Uses Protozoa To Assess the Toxicity of Bacteria WOLFRAM SCHLIMME,1 MARCELLO MARCHIANI,1 KURT HANSELMANN,2 1 AND BERNARD JENNI * Novartis Pharma AG, CH-4002 Basel,1 and Department of Microbiology, Institute of Plant Biology, University of Zu ¨rich, 8008 Zu ¨rich,2 Switzerland Received 25 September 1998/Accepted 18 March 1999

A new type of toxicity test based on the protozoan Tetrahymena pyriformis has been developed to assess the overall toxicity of bacterial strains given as prey. This simple and rapid test is able to detect toxicant-producing bacteria, which may present a biohazard. It can also be used for the risk assessment of microbes designed for deliberate release.

Strains and preparation of microorganisms. The strains used and their sources are listed in Table 1. Besides the reference strains from official culture collections (American Type Culture Collection [ATCC], Rockville, Md.; National Collection of Type Cultures [NCTC], London, United Kingdom; and Deutsche Sammlung von Mikroorganismen [DSM], Braunschweig, Germany), a majority of the strains originated from screening programs for the isolation of bacteria producing antifungal or insecticidal substances, e.g., Novartis Biosafety Bactox Collection [NBBC], Novartis Pharma AG, Basel, Switzerland). The root-colonizing Pseudomonas fluorescens strain CHA0 (19), coded NBBC 267, was used, and so were some of its derivates, which vary in the production of secondary metabolites. For example, P. fluorescens CHA0-Rif/pME3424 (12), coded NBBC 268, contains a plasmid responsible for the overproduction of the antibiotics 2,4-diacetylphloroglucinol and pyoluteorin. All bacteria were cultured on tryptone soy agar medium at 25°C for 3 days to allow a direct comparison, though the use of other media might influence the production of bacterial metabolites. One 1-ml loopful of bacteria was resuspended into 1 ml of sterile tap water, which corresponds to a McFarland value of approximately 5, or to a concentration of 108 to 109 cells ml21 (as determined by dilution and plating on tryptone soy agar). T. pyriformis GL (ATCC 30327) (3) was grown axenically to a density of approximately 105 protozoa ml21 in 1% proteose peptone medium enriched with 0.1% yeast extract at 25°C in tissue culture flasks. The protozoa were then centrifuged at 200 3 g for 1 min. The supernatant was removed, and the protozoa were resuspended in membrane-filtered (0.22-mm pore size) and autoclaved tap water to a final density of 1 3 104 to 3 3 104 cells ml21. The number of cells was measured with a CASY cell counter (Schaerfe System, Reutlingen, Germany) calibrated to a 30 mM NaCl solution. Before the bacteria were added, the protozoa were incubated at 25°C for 30 min to adapt them to the lower osmotic pressure of tap water before the start of the assay. Filtered, autoclaved tap water was used throughout the test because tap water best mimics the osmolarity of the natural habitats of Tetrahymena and sterilization eliminates residual chlorine. If a higher level of standardization is required, however, distilled water amended with low concentrations of minerals (“artificial” freshwater) can also be used (20). Peptones from the protozoan growth medium may react with bacterial metabolites and decrease their toxicity. Therefore, the test should be performed only in freshwater,

Many toxicity tests to determine acute and chronic effects are available for the monitoring and risk assessment of new chemicals and xenobiotics or for the evaluation of the toxicity of environmental pollutants. The choice of the best-suited organism for such bioassays is determined by the substances to be assessed and the sensitivity of the organism (5). For example, engineered cell lines are becoming a tool for the screening of hormone analogues (8), and algae are effective for testing the phytotoxicity of compounds acting on the photosynthetic pathway (16). Environmental pollutants and toxic substances are assessed with various organisms (1, 7, 9, 11, 15, 18), including protozoa, predominantly Tetrahymena (2, 7, 13, 14, 21). Usually, the test organisms are exposed to single chemicals or to environmental pollutants present in water but not to entire bacteria. Bacteria are increasingly involved in many biotechnological applications, including the production of bioactive substances, pest control, and plant protection. To provide the toxicological data for these bacteria as required by, e.g., regulatory authorities (10), the usual toxicity tests are not applicable since they are assays based on chemicals in solution. No bioassay to assess the overall toxicity of microorganisms has been described so far. We present a new type of semiquantitative bioassay to assess overall bacterial toxicity based on the protozoan Tetrahymena pyriformis fed with naturally occurring or genetically modified bacteria. Deaths among the protozoa are monitored, and the death rate is used to assess the toxicity of the bacteria. T. pyriformis was chosen because it is a well-recognized standard for toxicity testing (4, 6, 13–15, 17). The purpose of the BACTOX test is the detection of the overall toxicity of surreptitious strains which synthesize toxic secondary metabolites (toxicants) and which may constitute a biohazard. Its purpose is not the detection of specifically targeted toxins, since bacteria may produce several toxic metabolites simultaneously (synergies). This type of bioassay is of ecological relevance, since it monitors a trophic interaction at the first level of the food web. This test that uses protozoa is the first of its kind that can be used both for the detection of bacterial toxicants and for the risk assessment of bacterial strains.

* Corresponding author. Mailing address: Novartis Pharma AG, K-135.P.16, CH-4002 Basel, Switzerland. Phone: 41-61-696.58.01. Fax: 41-61-696.16.49. E-mail: [email protected]. 2754

BIOMONITORING TEST WITH TETRAHYMENA

VOL. 65, 1999 TABLE 1. Bacterial strains used in this study Bacterium

Strain(s)

Origina

B. cereus

ATCC 11145 NBBC 134–NBBC 142

ATCC H. J. Kempf

B. thuringiensis

NBBC 143–NBBC 147

H. J. Kempf

E. coli

DSM 1607, DSM 5911, DSM 2840 NBBC 172, NBBC 173 ATCC 8739, ATCC 10536, NCTC 9001

DSM J. Frey H.-J. Halfmann

L. lactis

NBBC 176–NBBC 178, ATCC 11454, ATCC 14365

K. Seeboth

P. fluorescens

NBBC 216–NBBC 266 NBBC 267–NBBC 274 NCTC 4755

H. J. Kempf K. Haas H.-J. Halfmann

S. aureus

NBBC 280–NBBC 282 ATCC 6538, ATCC 6538P, ATCC 13709, NCTC 7447

K. Seeboth H.-J. Halfmann

Streptomyces spp.

NBBC 298–NBBC 335

K. Seeboth

a

H. J. Kempf is from Novartis Crop Protection AG, Basel, Switzerland; J. Frey is from the Institut fu ¨r Vet.-Bakteriologie, University of Bern, Bern, Switzerland; H.-J. Halfmann is from Novartis Pharma AG; K. Seeboth is from Novartis Animal Health AG, Basel, Switzerland; and D. Haas is from the Laboratoire de Biologie Microbienne, Universite ´ de Lausanne, Lausanne, Switzerland.

even if T. pyriformis readily ingests bacteria in the presence of dissolved nutrients. Weaker or even false-negative responses were observed when the test was carried out in the protozoan growth medium. Standard bioassay. The test was carried out at 25°C in 12well non-surface-treated polystyrene plates which allowed for convenient repetitive microscopical examination. Each well contained 1 ml of bacterial suspension to which 500 ml of

2755

protozoan suspension was added. The total volume of 1.5 ml was large enough to prevent adverse effects of evaporation during the time course of the assay. The final concentrations of microorganisms were 3 3 104 ml21 for T. pyriformis and 108 to 109 ml21 for the bacteria. Bacterial ingestion and lysis of protozoa were observed on an inverted microscope with a 125-fold magnification. The entire surface of each well was monitored after 10 min and after 2, 4, and 8 h. Each bacterial strain was assayed several times in order to verify the reproducibility of the results and the reliability of the test. A protozoan suspension without bacteria was used as a control for every plate. Escherichia coli K-12 strain W3110 (DSM 5911) or strain HB101 (DSM 1607) was also routinely used as a negative control. P. fluorescens NBBC 267 and NBBC 268 were chosen as reference strains for weakly and strongly positive controls, respectively. The BACTOX test as a tool for monitoring bacterial toxicity. Various bacteria were tested, and in some cases a progressive lysis of the protozoa, which was a clear indication of toxicity, occurred. In the case illustrated in Fig. 1, overall lysis required 10 min with the progression presented in Fig. 1a to d. More time is needed to obtain the same response for less harmful bacteria. Microscopic observations should be made periodically in order to monitor the evolution of the protozoan population, since lysed protozoan bodies are quickly cannibalized by the survivors and released proteins may inactivate toxic substances. Approximately 10% of the protozoan population did not ingest bacteria. As a standard operating procedure, we suggest monitoring after 10 min and after 2, 4, and 6 to 8 h. Subsequent observations are unnecessary, since harmful effects have never been observed after 8 h. In order to estimate the intensity of the noxious effect of the tested bacteria, the proportion of dead protozoa was determined microscopically. We propose an easy-to-judge scale, with rankings from 1 to 5 as shown in Fig. 2. A larger scale would not be appropriate to the sensitivity of the test.

FIG. 1. Effect of harmful bacteria on T. pyriformis. The lysis of the protozoa as depicted progressively in panels a to d occurs more or less rapidly, depending on the bacterial strain fed. After only 10 min of incubation with P. fluorescens NBBC 268, the entire population of protozoa was dead (d). Images were produced by Nomarski differential interference contrast microscopy (DIC).

2756

SCHLIMME ET AL.

APPL. ENVIRON. MICROBIOL.

FIG. 2. Scale for the quantification of the BACTOX test. The different stages correspond to the following effects on the protozoan population: 1, no effect, i.e., 100% of the protozoa are healthy; 2, weak, only a few protozoa look significantly impaired compared to the control; 3, strong, with about 50% of the protozoa dead; 4, very strong, with only a few protozoa still living; and 5, maximum, i.e., all protozoa are dead. Images were produced by DIC.

Typical responses observed with different bacteria. Examples of bacteria which cause different levels of effect are presented in Table 2. No detectable effect was observed with bacteria such as E. coli K-12 or B or Lactococcus lactis. At the other end of the scale, P. fluorescens NBBC 268 killed almost all protozoa within 10 min. For such an immediate effect to take place, one has to assume that the active metabolite(s) was dissolved in the test medium, since the protozoa did not have enough time to ingest many bacteria. Other bacteria needed a few hours to develop their noxious effect. Toxicity which develops slowly may indicate that the toxic substance is less active or is synthesized in smaller amounts or that it might be liberated or activated during the digestion process. The sensitivity of the test was assessed with a total of 258 strains; among them, 106 strains were active (not shown). They belonged to gram-positive or gram-negative bacteria of 22 different genera. The high proportion of active strains obtained (41%) is an indication that the test is able to detect various toxic substances and is thus applicable to diverse bacterial types. We assessed the overall toxicity of bacteria to protozoa and not the response of protozoa to specific toxins, since bacteria commonly produce several toxic substances simultaneously. Applicable concentrations of protozoa and bacteria. It was necessary to assess the optimal concentration range of protozoa. Above a concentration of 105 cells ml21, crowding artifacts may influence the result, and cell densities below 103 ml21 are too low for a convenient observation. Variations between these concentration limits did not alter the outcome of the test results. Thus, highly standardized measurements of protozoan cell densities are not necessary. We have investigated the effects of different bacterial concentrations on the outcome of the test results. An innocuous E. coli strain applied at 10- or 100-fold the standard concentration did not produce any toxic effect (Table 2). A falsepositive result due to overcrowding at high bacterial densities can thus be excluded. Moreover, an error leading to more than a 10-fold concentration of bacteria is not realistic, because the

resulting turbidity would make reading the test too difficult. A 10-fold dilution of a moderately toxic strain which develops its activity during the digestion cycle (e.g., P. fluorescens NBBC 267) did not change the result. This might be due to feeding TABLE 2. Effects of bacterial strains taken as references and of various bacterial concentrations on the protozoaa Bacterium

Strain

Concn

Control

Effectb after: 10 min

2h

4h

8h

1*

1*

1*

1*

E. coli B

DSM 2840

13

1*

1*

1*

1*

E. coli K-12

DSM 5911

13

1*

1*

1*

1*

L. lactis

ATCC 11454

13

1*

1*

1*

1*

P. fluorescens

NBBC 268

13

4

5

5*

5*

P. fluorescens

NBBC 267

13

1

2

2

2

P. fluorescens

NBBC 233

13

1

3

4

4

E. coli

NCTC 9001

13

1*

1*

1*

2

E. coli

ATCC 8739

13 103 1003

1* 1* 1*

1* 1* 1*

1* 1* 1*

1* 1* 1*

P. fluorescens

NBBC 267

13 0.13

1 1

2 2*

2 2*

2 2*

P. fluorescens

NBBC 268

13 0.13

4 2*

5 3

5* 3

5* 2

a The noxious effect on T. pyriformis is ranked for intensity on a scale from 1 to 5, as pictured in Fig. 2. The standard concentration (13) of bacteria was 108 to 109 cells ml21. The experiments were repeated at least 10 times for the strains tested with the 13 concentration and three times when strains were tested with other concentrations. b *, no variation was observed for any repetition of the experiment.

BIOMONITORING TEST WITH TETRAHYMENA

VOL. 65, 1999 TABLE 3. Results of the BACTOX test with some relevant taxonomic groups of bacteriaa Bacterium

RG (NIH/EU)

No. of tested strains

% of active strains

Range on the toxicity scale

E. coli K-12 or B E. coli (wild types) P. fluorescens L. lactis B. cereus B. thuringiensis S. aureus Streptomyces spp.

1 2 1 1 1/2 1 2 1

3 5 60 5 10 5 7 38

0 80 57 0 0 0 100 37

1 1–2 1–5 1 1 1 2–4 1–5

a By definition, EU RG1 contains only nonpathogenic bacteria, while NIH RG1 does not exclude opportunistic pathogens. A broad range on the toxicity scale indicates that different strains gave different responses, thus revealing the heterogeneity of the taxonomic group considered.

behavior, since protozoa track their prey actively and concentrate it into their digestive vacuoles. A 10-fold dilution of the fast-acting strain NBBC 268, however, weakened the toxic effect by two levels on the scale. This result strengthens the hypothesis that in this case, the toxic substance was dissolved in the solution. To obtain the highest sensitivity, the BACTOX test should be carried out at the optimal bacterial concentration. Reliability and reproducibility of the test. The levels given in Table 2 are median values obtained from several repetitions of the experiment. For all innocuous bacterial strains which showed a value of 1 even after 8 h of incubation, higher levels of toxicity were never observed during at least 10 repetitions of the experiment. Variations between toxicity levels observed during repetitions of the experiment were seldom more than one level, and they were due to differences both in visual estimations and in the time it took for harmful effects to develop. Results with bacteria belonging to different taxonomic groups. The test was carried out with various types of bacteria in order to assess its efficiency and to compare the results with their classification in biohazard risk groups (RGs). The BACTOXpositive strains were classified in RG1 and RG2 of either the National Institutes of Health (NIH) or the European (EU) legislation. Some relevant examples are presented in Table 3. In line with their classification, the K-12 and B strains of E. coli showed no activity, while four out of five wild-type isolates did. No active strains of the food-grade L. lactis bacteria were detected, but all the pathogenic Staphylococcus aureus strains tested were harmful (Table 3). The strains of the highly specific insect pathogen Bacillus thuringiensis were innocuous for Tetrahymena, and so were the food-poisoning Bacillus cereus strains. The latter species is classified in EU RG2, but the tested strains were not of clinical origin. A significant number of strains classified in EU RG1 and akin to the biotechnologically important groups P. fluorescens and Streptomyces spp. presented a very high activity. The large range obtained on the toxicity scale in those groups reveals that the production of active metabolites is more strain specific than species specific. The toxic properties of newly isolated strains must therefore be assessed prior to growth on a large scale (10), in order to determine the appropriate biosafety level of physical containment. Refined risk assessment or additional safety measures can then be implemented for BACTOX-positive strains classified in EU RG1. Negative BACTOX results, however, do not exclude the possibility of harmful effects. Conclusion. We present a new type of biosafety test based on the protozoan T. pyriformis, which is not comparable to other toxicity tests. It allows the monitoring of potentially hazardous bacteria taken as such instead of the monitoring of single chemicals. The test organism is also naturally bacteri-

2757

otrophic, which makes it more relevant than conventional toxicity tests for environmental-impact studies involving deliberately released bacteria. This assay can be applied in the risk assessment of genetically modified or wild-type bacteria which may exhibit toxic properties. We thank D. Haas, H. J. Halfmann, H. J. Kempf, U. Schnider, and K. Seeboth for kindly providing bacteria; R. Peck for the protozoan strain; and C. M. Fischer for corrections. This work was supported by the Swiss National Science Foundation (Priority Programme Biotechnology, grant no. 5002-35145) and by Novartis Pharma AG. REFERENCES 1. Baird, D. J., I. Barber, A. M. V. M. Soares, and P. Calow. 1991. An early life-stage test with Daphnia magna Straus: an alternative to the 21-day chronic test? Ecotoxicol. Environ. Saf. 22:1–7. 2. Benitez, L., A. Martin-Gonzalez, P. Gilardi, T. Soto, J. Rodrigez de Lecea, and J. C. Gutie´rrez. 1994. The ciliated protozoa Tetrahymena thermophila as a biosensor to detect mycotoxins. Lett. Appl. Microbiol. 19:489–491. 3. Borden, D., G. S. Whitt, and D. L. Nanney. 1973. Electrophoretic characterization of classical Tetrahymena pyriformis strains. J. Protozool. 20:693–700. 4. Bringhmann, G., and R. Kuhn. 1980. Comparison of the toxicity thresholds of water pollutants to bacteria, algae and protozoa in the cell multiplication inhibition test. Water Res. 14:231–241. 5. Fawell, J. K., and S. Hedgecott. 1996. Derivation of acceptable concentrations for the protection of aquatic organisms. Environ. Toxicol. Pharmacol. 2:115–120. 6. Greenberg, A. E., L. S. Clesceri, and A. D. Eaton, (ed.). 1992. Standard methods for the examination of water and wastewater, 18th ed. American Public Health Association, Washington, D. C. 7. Huber, H. C., W. Huber, and U. Ritter. 1990. Simple bioassays for evaluating toxicity of environmental chemicals using microcultures of human peripheral lymphocytes and monoxenic cultures of the ciliate Tetrahymena pyriformis. Zentbl. Hyg. Umweltmed. 189:511–526. 8. Joyeux, A., P. Balaguer, P. Germain, A. M. Boussioux, M. Pons, and J. C. Nicolas. 1997. Engineered cell lines as a tool for monitoring biological activity of hormone analogs. Anal. Biochem. 249:119–130. 9. Lahti, K., J. Ahtiainen, J. Rapala, K. Sivonen, and S. I. Niemela ¨. 1995. Assessment of rapid bioassays for detecting cyanobacterial toxicity. Lett. Appl. Microbiol. 21:109–114. 10. Lelieveld, H. L., H. Bachmayer, B. Boon, G. Brunius, K. Bu ¨rki, A. Chmiel, C. H. Collins, P. Crooy, O. Doblhoff-Dier, A. Elmquist, W. Frommer, C. Frontali-Botti, R. Havenaar, H. Haymerle, C. Hussey, O. Ka ¨ppeli, M. Lex, S. Lund, J. L. Mahler, R. Marris, C. Mosgaard, C. Normand-Plessier, F. Rudan, R. Simon, M. Logtenberg, and R. G. Werner. 1995. Safe biotechnology. Part 6. Safety assessment, in respect of human health, of microorganisms used in biotechnology. Appl. Microbiol. Biotechnol. 43:389–393. 11. McFeters, G. A., P. J. Bond, S. B. Olson, and Y. T. Tchan. 1983. A comparison of microbial assays for the detection of aquatic toxicants. Water Res. 17:1757–1762. 12. Natsch, A., C. Keel, N. Hebecker, E. Laasik, and G. De´fago. 1997. Influence of biocontrol strain Pseudomonas fluorescens CHA0 and its antibiotic overproducing derivative on the diversity of resident root colonizing pseudomonads. FEMS Microbiol. Ecol. 23:341–352. 13. Nilsson, J. R. 1989. Tetrahymena in cytotoxicity: with special reference to effects of heavy metals and selected drugs. Eur. J. Protistol. 25:2–25. 14. Sauvant, M. P., D. Pepin, J. Bohatier, and C. A. Groliere. 1995. Microplate technique for screening and assessing cytotoxicity of xenobiotics with Tetrahymena pyriformis. Ecotoxicol. Environ. Saf. 32:159–165. 15. Sauvant, M. P., D. Pepin, J. Bohatier, C. A. Groliere, and J. Guillot. 1997. Toxicity assessment of 16 inorganic environmental pollutants by six bioassays. Ecotoxicol. Environ. Saf. 37:131–140. 16. Schafer, H., H. Hettler, U. Fritsche, G. Pitzen, G. Roderer, and A. Wenzel. 1994. Biotests using unicellular algae and ciliates for predicting long-term effects of toxicants. Ecotoxicol. Environ. Saf. 27:64–81. 17. Slabbert, J. L., and J. P. Maree. 1986. Evaluation of interactive toxic effects of chemicals in water using a Tetrahymena pyriformis toxicity screening test. Water SA 12:57–62. 18. Snell, T. W., and B. D. Moffat. 1992. A 2-d life cycle test with the rotifer Brachionus calyciflorus. Environ. Toxicol. Chem. 11:1249–1257. 19. Stutz, E. W., G. De´fago, and H. Kern. 1986. Naturally occurring fluorescent pseudomonads involved in suppression of black rot of tobacco. Phytopathology 76:181–185. 20. Weber, C. I. 1993. Methods for measuring the acute toxicity of effluents and receiving waters to freshwater and marine organisms, 4th ed. EPA/600/4-90/ 027 August. Office of Research and Development, U. S. Environmental Protection Agency, Cincinnati, Ohio. 21. Yoshioka, Y., Y. Ose, and T. Sato. 1985. Testing for the toxicity of chemicals with Tetrahymena pyriformis. Sci. Total Environ. 43:149–157.