- Email: [email protected]
A microbial biosensor to predict bioavailable nickel in soil and its transfer to plants
Environmental Pollution 113 (2001) 19±26
www.elsevier.com/locate/envpol
A microbial biosensor to predict bioavailable nickel in soil and its transfer to plants C. Tibazarwa a,b, P. Corbisier a, M. Mench c, A. Bossus a, P. Solda, M. Mergeay a, L. Wyns b, D. van der Lelie a,* a
Environmental Technology Expertise Centre, Flemish Institute of Technological Research (Vito), Boeretang 200, B-2400 Mol, Belgium Department of Ultrastructure, Free University of Brussels, Flemish Interuniversity Institute of Biotechnology (VIB), Paardenstraat 65, B-1640 St-Genesius-Rode, Belgium c Agronomy Unit, Inra Bordeaux Aquitaine Research Center, BP 81, 33883 Villenave d'Ornon, France
b
Received 8 December 1999; accepted 8 July 2000
``Capsule'': A whole-cell biosensor was developed for the detection of bioavailable Ni2+ and Co2+ concentrations in soil samples. Abstract Ralstonia eutropha strain AE2515 was constructed and optimised to serve as a whole-cell biosensor for the detection of bioavailable concentrations of Ni2+ and Co2+ in soil samples. Strain AE2515 is a Ralstonia eutropha CH34 derivative containing pMOL1550, in which the cnrYXH regulatory genes are transcriptionally fused to the bioluminescent luxCDABE reporter system. Strain AE2515 was standardised for its speci®c responses to Co2+ and Ni2+. The detection limits for AE2515 were 0.1 mM Ni2+ and 9 mM Co2+, respectively. The signal to noise (S/N) bioluminescence response and the metal cation concentration could be linearly correlated: for Ni2+ this was applicable within the range 0.1±60 mM, and between 9 and 400 mM for Co2+. The AE2515 biosensor strain was found to be highly selective for nickel and cobalt: no induction was observed with Zn(II), Cd(II), Mn(II), Cu(III) and Cr(VI). In mixed metal solutions, the bioluminescent response always corresponded to the nickel concentrations. Only in the presence of high concentrations of Co2+ (2 mM), the sensitivity to nickel was reduced due to metal toxicity. AE2515 was used to quantify the metal bioavailability in various nickel-enriched soils, which had been treated with additives for in situ metal immobilisation. The data obtained with strain AE2515 con®rmed that the bioavailability of nickel was greatly reduced following the treatment of the soils with the additives beringite and steel shots. Furthermore, the data were found to correlate linearly with those on the biological accumulation of Ni2+ in speci®c parts of important agricultural crops, such as maize and potato. Therefore, the test can be used to assess the potential transfer of nickel to organisms of higher trophic levels, in this case maize and potato plants grown on nickel-enriched soils, and the potential risk of transfer of these elements to the food chain. # 2001 Elsevier Science Ltd. All rights reserved. Keywords: Bioavailability; Heavy metals; Nickel; Plants; Soil; Transfer
1. Introduction The development of metal-speci®c biosensor tools functioning on the basis of bioluminescent reporter systems, has been receiving increasing attention (van der Lelie et al., 1994; Brown et al., 1998; Corbisier et al., 1999). These biosensors are considered as key assets both for characterizing the extent of contaminated areas and * Corresponding author. Tel.: +32-14-335166; fax: +32-14580523. E-mail address: [email protected] (D. van der Lelie).
for following up the success or failure of bioremediation operations of large areas contaminated with heavy metals, such as the Guadiamas river valley (de Lorenzo and Kuenen, 1999). They are of particular relevance for the assessment of remediation strategies based on in situ immobilisation of heavy metals, where total metal concentrations cannot be used to determine the success or failure of the remediation operation. Some of the major advantages of biosensors are that they are relatively inexpensive to develop (especially true for whole cell biosensors), their rapidity, great sensitivity and ease of data handling. In environmental technology
0269-7491/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved. PII: S0269-7491(00)00177-9
20
C. Tibazarwa et al. / Environmental Pollution 113 (2001) 19±26
applications, the use of luminometry based assays is associated with easy-to-handle instrumentation and relatively uncomplicated data handling (Corbisier et al., 1996). Furthermore, these types of assays do not require complex and labour-intensive handling of the samples, compared with some of the more common chemical analyses used to predict availability of heavy metals, such as sequential extraction (Tessier et al., 1979). The accurate determination of the biologicallyavailable (bioavailable) heavy metal content of environmental samples is a topic which has been discussed extensively, particularly in the context of sampling techniques. The determination of the metal content in the water-extractable or Ca(NO3)2 extracted fraction, has been shown to provide a relatively good indication of the bioavailable metal content of the soil sample, depending on the soil characteristics (Weissenhorn et al., 1995). However, it has not been possible to assign one particular method for the absolute determination of the bioavailable metal fraction in a soil. The bacterial-based biosensors may prove the ideal solution since they can be used directly with the soil sample and can therefore determine bioavailable metal levels in situ (van der Lelie, 1998). A comparative study of the data obtained from a biosensor assay with those from sequential extractions of soil samples (chemical data), as well as accumulation of the metal within biological systems of higher trophic levels, such as plants (biological data), may be useful in resolving the technical merits of one method over another in determining the bioavailable metal fraction in soils and assessing their potential risk. Thus, the need for simple tests that predict uptake of heavy metal by organisms of higher trophic levels would prove invaluable in the risk assessment studies of environmental contamination. In this communication, we describe the development, standardisation and application to soil samples, of a whole cell biosensor strain based on the Ralstonia eutropha CH34 derivative, designated as strain AE2515, which responds speci®cally to nickel and cobalt. The construction of AE2515 is based on pMOL1550 (Tibazarwa et al., 2000), in which the regulatory genes, cnrYXH, of the cnr cobalt and nickel resistance determinant of Ralstonia eutropha CH34 (Liesegang et al., 1993) are transcriptionally fused upstream of the luxCDABE genes of pMOL877 (van der Lelie et al., 1997). A similar strategy has been used to construct metal-speci®c biosensor strains for Zn, Cu and Pb (Corbisier et al., 1999): these strains are referred to as BIOMET strains (Corbisier et al., 1996). Strain AE2515 was found to emit a metal-speci®c response induced by dierent concentrations of Ni2+ and Co2+. This tight metal-controlled regulation, prompted us to test for the biosensor potential of AE2515 and to assess the prospective applications of this strain in quantifying bioavailable levels of Ni2+ and Co2+ in environmental samples.
As proposed by Boisson et al. (1998), an attempt was made to provide an alternative test to assess the reduction of metal toxicity in contaminated soils following in situ immobilisation treatment with the soil additives beringite and steel shots. The same soils samples as described in Boisson et al. (1998) were tested so that a direct comparison could be made between the use of metal biosensors in the prediction and/or determination of metal bioavailability in the soils versus the uptake by plants. 2. Materials and methods 2.1. Construction of biosensor strain AE2515 The construction of R. eutropha AE2515, a derivative of R. eutropha CH34 (Mergeay et al., 1985) containing pMOL1550, has been described by Tibazarwa et al. (2000) and is schematically depicted in Fig. 1. Plasmid pMOL1550 is pMOL877, in which the cnrYXH regulatory region was cloned upstream of the luxCDABE reporter system. This resulted in inducible light production controlled by the cnr regulon. For biosensor purposes, cultures of AE2515 were freeze-dried and aliquoted into sterile ampoules (Corbisier et al., 1999). 2.2. Luminometry assays Luminometry assays were carried out using an ANTHOS LUCY1 luminometer (Anthos Labtech b.v., Heerhugowaard, The Netherlands) at 23 C, as previously described (Corbisier et al., 1999). For metal standards, duplicate samples were set-up per microtitre assay. As negative controls, eight reaction samples containing bidistilled water were included in the tests. The bioluminescence emitted (ALU) and the optical density (OD620 nm) of the cultures were measured over 16 h at 30 min intervals, and processed using the MIKROWIN software, as previously described (Corbisier et al. 1999). Further data processing was carried out using EXCEL 7.0 (Microsoft) or Origin 5.0 (MicrocalTMOriginTM, Microcal Software, Inc., Northampton, MA). 2.3. Soil sampling and preparation Soil samples and the experimental conditions of in situ metal immobilisation were previously described by Boisson et al. (1998) and are presented in Table 1. Brie¯y, soils from a long-term sewage sludge ®eld trial on an acid coarse sandy soils in the Bordeaux region, France, termed Louis Fargues, were treated with the soil additives beringite (BE) and steel shots (SS). The metal content of dierent parts of maize plants (cv. INRA 260) as well as potato tuber cultivars was determined using either inductively coupled plasma atomic emission spectrometry (ICP±AES) or gas ¯ame atomic
C. Tibazarwa et al. / Environmental Pollution 113 (2001) 19±26
21
Table 1 Characteristics of Louis Fargues sludge plots treated with chemical additives and analysed with the the BIOMET testa Parcel no.
Soil description
Soil additive
pH
Total (Ni2+), ppm
20 27 39 34 26 40 35 28 37 36
UNT CTRL FYM CTRL FYM+SS FYM+BE S1 CTRL S1+SS S1+BE S2 CTRL S2+SS S2+BE
None None Steel shots Beringite None Steel shots Beringite None Steel shots Beringite
6.32 6.70 6.68 7.53 6.75 7.00 7.69 7.00 7.19 7.54
2.00 51.92 15.77 10.88 44.481 59.26 66.97 179.45 215.46 190.32
a Louis Fargues sludge plots on an Arenic Udi¯uvent (sandy) soil were treated with the additives Beringite (BE) or Steel shots (SS) or left untreated (CTRL) to test for the in situ immobilisation of polluting heavy metals. The plots with FYM received farm yard manure at a loading rate of 10 t DM ha 1, with S1 received sewage sludge at loadings of 10 t DM ha 1 and those with S2 received sewage sludge at loadings of 100 t DM ha 1. The physical properties of the soils, including the pH of the water fraction, and total nickel levels (as determined by GFAAS) were determined. The texture of all soils was predominantly sandy (>90%).
tested with the nickel biosensor strain, AE2515. For each assay, metal standard solutions were included, in duplicate, while the non-induced biosensor control test was set up as eight reactions. The reconstituted soils were tested in dilution series: typically 1 to 1, 1 to 2 and 1 to 4, or depending on the apparent toxicity of the samples, and also in duplicate. 2.5. Determination of bioavailable levels of nickel in soils
Fig. 1. The BIOMET-Ni biosensor. A schematic diagram of the construction of the biosensor strain, AE2515.
absorption spectroscopy (GFAAS), depending on the chemical element and concentration. The AE2515 biosensor strain was tested on dierent soils currently being monitored as part of a larger study on the bioremediation and recovery of Ni2+ from contaminated sites (Boisson et al., 1998). 2.4. BIOMET assays on soil samples Soil samples were suspended in reconstitutive medium (RM), as described previously (Corbisier et al., 1999). The 11 soil samples of Louis Fargues were subsequently
Raw data on heavy metal availability were calibrated using metal standard solutions, as described in Corbisier et al. (1999). The signal to noise ratio (S/N) calculated for each soil sample was equated to the nickel concentration and ®nally to the Ni2+ equivalent on a weight to weight basis. 3. Results 3.1. Nickel and cobalt speci®c bioluminescence response of strain AE2515 To assess the speci®city of AE2515 as a metal biosensor, the strain was tested in the presence of varying concentrations of Ni2+, Co2+, Zn2+, Cr6+, Cr3+, Mn2+, Cd2+ and Cu2+. As shown in Fig. 2a, the response to nickel was prominent, with a maximum Signal/Noise (S/N) ratio, exceeding 1000 in the presence of 0.1 mM Ni2+. The maximal response to Co2+ corresponded to a S/N ratio of 100 with 1.6 mM Co2+
22
C. Tibazarwa et al. / Environmental Pollution 113 (2001) 19±26
Fig. 3. Correlation of Ni2+ and Co2+ concentrations with the response of the biosensor strain AE2515. Plotted is a trendline of the relationship between signal/noise and metal concentration, in mM for (a) Ni2+ and (b) Co2+. The error bars represent the average deviation from the mean value of ®ve independent experiments.
metal concentration and the bioluminescence response was perfectly linear between 9 and 400 mM (Fig. 3b).
Fig. 2. Induction of light production in presence of nickel and cobalt by the biosensor strain AE2515. The maximum signal/noise (ALU/ OD620 nm) is presented in function of the concentration of (a) Ni2+ and (b) Co2+. The error bars represent the average deviation from the mean value of three independent experiments.
(Fig. 2b). The maximal bioluminescent response was reached within 4±6 h after the start of experimentation. The data were found to be highly reproducible for a given batch of freeze-dried vials, and this over a period exceeding 8 months with storage at 4 C. The batch-tobatch variability was not determined. The metals Zn2+, Cr6+, Cr3+, Mn2+, Cd2+, Cu2+ Ð at non-toxic concentrations in the range of 0.1 mM Ð did not signi®cantly induce bioluminescence (S/N ratio <2, data not shown). The detection limits for Ni2+ and Co2+ were found to be 0.1 and 9 mM, respectively, and give an indication of the sensitivity of the biosensor for these metals. From recent data on the metal-controlled regulation of the cnr genes, it has been determined that the regulatory genes respond predominantly to Ni2+: therefore the sensitivity of the biosensor is inherently superior for Ni2+. For Ni2+, a perfect linear correlation was found to exist between the S/N ratio and metal concentrations from 0.1 to 60 mM (Fig. 3a): the equation of the ®tted trendline is also indicated. For Co2+, the correlation between
3.2. Selectivity of strain AE2515 bioluminescence responses in mixed-metal solutions The performance of the biosensor in complex samples was tested using mixed metal standards of the combination of Ni2+ and Co2+. As shown in Fig. 4, the Ni2+ concentrations were maintained at a constant level, while the Co2+ levels were varied, and the biosensor response monitored. As shown, the Ni2+-speci®c response was markedly reduced in the presence of high Co2+ levels. The greatest reduction was observed with 1.6 mM Co2+ causing a 62.5% reduction in the S/N for 0.3 mM Ni2+. From the slopes of the inhibition curves, it can be concluded that inhibition of the biosensor performance for nickel in the presence of high levels of Co2+ is due the toxicity of the latter to cellular growth of the biosensor strain. Thus, even in these mixed-metal solutions, the biosensor response was clearly targetted to Ni2+, thereby con®rming the selectivity of the biosensor for Ni2+. Similar observations were made for combinations of Ni2+ and other heavy metals (data not shown). 3.3. Application of strain AE2515 as a biosensor for the detection of bioavailable concentrations of nickel in soils The eectiveness of strain AE2515 to detect the bioavailable fraction of Ni2+ in soil samples was tested on the Louis Fargues soils, which continue to be monitored
C. Tibazarwa et al. / Environmental Pollution 113 (2001) 19±26
Fig. 4. Performance of the AE2515 biosensor strain in mixed metal solutions. The signal/noise values were compared for the biosensor in the presence of various ®xed Ni2+ concentrations and increasing Co2+ concentrations. Data is an average of two independent experiments, carried out in duplicate.
as part of an ongoing bioremediation study (Boisson et al., 1998). The BIOMET values for bioavailable nickel were compared with the calcium nitrate exchangeable fraction of the soils (chemical data). As shown in Fig. 5, a signi®cant correlation (R=0.97) was observed between the two types of data. According to the correlation equation, 50±60% of the the Ca(NO3)2 exchangeable nickel fraction is measured as bioavailable with the biosensor strain AE2515. These ®ndings corroborate those of Boisson et al. (1998), that Ca(NO3)2 fractions typically give a value greater than the bioavailable metal content. Since the soils were being monitored for the eect of the additives beringite and steelshots on the nickel bioavailability in the soil, we used the BIOMET data to determine the change in metal bioavailability in such treated soils, and compared these data with those of the calcium nitrate exchangeable fractions, as determined by Boisson et al. (1998). As shown in Fig. 6, the reduction in nickel bioavailability in the presence of the soil additives as determined with the biosensor strain, corroborates the ®ndings of Boisson et al., although does not duplicate their data. For the farm yard manure soils treated with steelshots, no Ca(NO3)2 data was available. However, the BIOMET data revealed that this treatment resulted in a clear reduction in bioavailability. In general, the Ca(NO3)2 fractions give higher percentage reduction values than the AE2515 biosensor strain. 3.4. Correlation between bioavailable nickel levels determined using the BIOMET nickel biosensor test versus metal accumulation in plants The ecological relevance of the BIOMET-nickel test was evaluated by examining its correlation with other tests used to quantify metal accumulation in organisms of higher trophic levels, such as plants. In general, heavy metal accumulation tests in plants are used to assess the potential transfer of heavy metals to the food chain. In
23
Fig. 5. Correlation of concentration of bioavailable nickel, as determined with the biosensor strain AE2515 and the calcium nitrate exchangeable nickel fraction in dierent Louis Fargues soils. The AE2515-determined data on nickel bioavailability dierent samples of Louis Fargues soils are plotted against the calcium nitrate data obtained by Boisson et al., 1998.
Fig. 7, the BIOMET data for nickel-contaminated soils Ð with or without additive treatment Ð are compared with the levels of nickel in dierent parts of maize or potato plants grown on the same soils. Interestingly, the BIOMET quantitation of Ni2+ levels correlates linearly with the levels obtained for nickel accumulation in the grains (Fig. 7a), the 6th (Fig. 7b) and ear leaves (Fig. 7c) of maize plants (determined by GFAAS spectrophotometry analyses; see Boisson et al., 1998). Nickel accumulation in potato tubers also showed a signi®cant, although not linear, correlation, (Fig. 7d). 4. Discussion This communication represents a comparison of the performance of the nickel-speci®c biosensor strain AE2515 with physicochemical and biological methods used to determine bioavailable heavy metal. Heavy metal-speci®c biosensors based on gene fusions are now being widely applied in many laboratories (Corbisier et al., 1999), and this has resulted in the development of an amalgam of dierent biosensor strains, termed BIOMET strains, speci®c for dierent metal substrates such as Zn(II), Cd(II), Cu(II), Pb(II) and Cr(VI). A prerequisite for the construction of these heavy metal speci®c biosensors is a thorough understanding of bacterial heavy metal resistance systems, especially their metal speci®c regulation. Our recent understanding of the regulatory mechanism encoded by the cobalt and nickel resistance cnr determinant, has allowed the construction of a biosensor responsive to induction by these metals based on CH34 (Tibazarwa et al., 2000). The high selectivity for nickel was demonstrated in a mixed metal assay. The
24
C. Tibazarwa et al. / Environmental Pollution 113 (2001) 19±26
Fig. 6. The eects of dierent treatments of additives on nickelenriched Louis Fargues soils on nickel bioavailability, as determined with the AE2515 nickel biosensor strain. The bioavailable Ni2+ levels determined using the AE2515 biosensor were used to calculate the change in bioavailability of the metal in three Louis Fargues soils diering in composition Ð S1, S2 and farm yard manure (FYM) Ð following treatment by in situ immobilisation with the additives, Steelshots (SS) or Beringite (BE) or no treatment (CTRL), as described in Boisson et al., 1998. (a) S1 soil, which had received doses of sewage sludge at loadings of 10 t DM ha 1; prior to; (b) S2, which had received doses of sewage sludge at loadings of 100 t DM ha 1; and (c) FYM, which had received doses of farm yard manure at loadings of 10 t DM ha 1. A comparison with chemical data on nickel bioavailability, as obtained by calcium nitrate extraction (Boisson et al.) is also included.
linear correlation between the concentration of Ni2+ and the bioluminescent (S/N ratio) response at low metal concentrations has enabled quantitative analyses of bioavailable nickel levels to be made in environmental samples. Furthermore, the bioavailable fraction of the metal detected in nickel-enriched soils could be correlated to the biological accumulation of Ni2+ in several organs of maize plants, notably the grains, as well as potato tubers. Plant-based tests are typically used to give an indication of the relative toxicity and danger of a metal-contaminated soil (Mench et al., 1994) as well as its potential health risks. Comparing the application of the BIOMET Ni-biosensor test versus the growth of maize and potato plants on nickelcontaminated soil, it can be argued that the BIOMET test oers the advantage of low cost and rapidity (results within 6 h), whereas the plants must be monitored over 6 months, after which heavy metal levels must be measured on pretreated plant material with techniques such as ICP±AES. The rapid assessment of heavy metal bioavailability and uptake by plants is an important parameter in several respects. First, it allows a rapid, cost-ecient and early assessment of the potential risks of heavy metal transfer linked to crop species grown on heavy metal enriched soils. The monitoring of this risk commonly involves the growth of crops on such soils. Secondly, a correlation between heavy metal bioavailability and plant uptake data can be used for modelling and in predicting the applicability of dierent phytoremediation techniques as ecient remediation strategies for heavy metal polluted soils. This is of particular relevance for both in situ immobilisation, aiming at reducing the plant-available heavy metal fraction and its uptake by plants, and phytoextraction, where information on the plant-available fraction is a prerequisite to predict the eciency of the remediation. We show here that the BIOMET nickel biosensor can be used to rapidly (and cost eectively) assess plant availability of the heavy metal. Studies are presently being carried out to demonstrate this concept for other BIOMET heavy metal biosensors. In this same context of heavy metal bioavailability, we show also that the BIOMET nickel biosensor can be used to assess the eects of dierent in situ immobilisation treatments, aiming to decrease the heavy metal bioavailability to plants. However, a shortcoming of the BIOMET test might be that it is not able to predict the eects of treatment with chemical chelators, which are sometimes used to increase uptake of heavy metals by plants (Huang and Cunningham, 1996). For instance, metal-EDTA complexes are not detected by BIOMET, as was shown for Cu-EDTA complexes (Corbisier et al., unpublished data). On a cost-to-cost basis, the use of the BIOMET would outweigh that of the use of plant-based toxicity tests. It is therefore proposed that the BIOMET test is
C. Tibazarwa et al. / Environmental Pollution 113 (2001) 19±26
25
Fig. 7. Correlation between nickel concentrations detected in crops with those detected by the nickel biosensor test. The bioavailable Ni2+ levels determined (in ppm) using the AE2515 biosensor were compared with the nickel accumulation in dierent parts of maize and potato tubers (in ppm). Nickel concentrations in plants were determined by ICP. Panels are: (a) grains of maize, (b) ear leaf of maize, (c) 6th leaf of maize and (d) potato tubers.
applied as an alternative management tool for the assessment of heavy metal polluted soils and the followup of their (bio)remediation. Although the use of BIOMET should reduce the number of plant-based tests, it should nevertheless be used in conjunction with plant studies, which are necessary for the assessment of potential toxicity or de®ciency eects due to the remediation strategy, such as the application of soil additives (Boisson et al., 1998). Acknowledgements We thank Ann Provoost and Sa®eh Taghavi for help with molecular biology procedures. This work was supported by a grant to C. T. from the Flemish Government and the VIB (Vlaamse Instelling voor Biotechnologie). References Boisson, J., Mench, M., Sappin-Didier, V., Solda, P., Vangronsveld, J., 1998. Short-term in situ immobilisation of Cd and Ni by
beringite and steel shots application to long-term sludged plots. Agronomie 18, 347±359. Brown, N.L., Lloyd, J.R., Jakeman, K., Hobman, J.L., Bontidean, I., Mattiasson, B., CsoÈregi, E., 1998. Heavy metal resistance genes and proteins in bacteria and their application. Biochem. Soc. Trans. 26 (4), 662±665. Corbisier, P., van der Lelie, D., Borremans, B., Provoost, A., de Lorenzo, V., Brown, N., Lloyd, J., Hobman, J., CsoÈregi, E., Johansson, G., Mattiasson, B., 1999. Whole cell- and protein-based biosensors for the detection of bioavailable heavy metals in environmental samples. Anal. Chim. Acta 387, 235±244. Corbisier, P., Thiry, E., Diels, L., 1996. Bacterial biosensors for the toxicity assessment of solid wastes. Environ. Toxicol. Water Qual. 11, 171±177. de Lorenzo, V., Kuenen, G., 1999. Scienti®c basis for the bioremediation of the toxic spill of the Aznalcollar mine: combining bacteria and plants to address an intractable kind of pollution. Environ. Microbiol. 1 (3), 275±278. Huang, J.W., Cunningham, S.D., 1996. Lead phytoextraction: species variation in lead uptake and translocation. New Phytol. 134, 75±84. Liesegang, H., Lemke, K., Siddiqui, R.A., Schlegel, H.G., 1993. Characterization of the inducible nickel and cobalt resistance determinant cnr from pMOL28 of Alcaligenes eutrophus CH34. J. Bacteriol. 175, 767±778. Mench, M., Martin, E., Solda, P., 1994. After eects of metals derived from a highly metal-polluted sludge on maize (Zea mays). Water Air Soil Pollution 75, 277±291.
26
C. Tibazarwa et al. / Environmental Pollution 113 (2001) 19±26
Mergeay, M., Nies, D.H., Schlegel, H.G., Gerits, J., Charles, P., Van Gijsegem, F., 1985. Alcaligenes eutrophus CH34 is a facultative chemolithotroph with plasmid bound resistance to heavy metals. J. Bacteriol. 162, 328±334. Tessier, A., Campbell, P.G.C., Bissonm, M., 1979. Sequential extraction procedure for the speciation of particulate trace metals. Anal. Chem. 51, 844±850. Tibazarwa, C., Wuertz, S., Mergeay, M., Wyns, L., van der Lelie, D., 2000. Regulation of the cnr cobalt and nickel resistance determinant of Ralstonia eutropha-like CH34 (ex Alcaligenes eutrophus CH34). J. Bacteriol. 182 (5), 1399±1409. van der Lelie, D., 1998. Biological interactions: the role of soil bacteria in the bioremediation of heavy metal-polluted soils. In: Van-
gronsveld, J., Cunningham, S. (Eds.), From Metal-contaminated soils: in situ inactivation and phytorestoration. Landes Bioscience, Springer-Verlag, Berlin, pp. 31±50. van der Lelie, D., Regniers, L., Borremans, B., Provoost, A., Verschaeve, L., 1997. The VITOTOX test, a SOS bioluminescence Salmonella typhimurium test to measure genotoxicity kinetics. Mut. Res. 2-3, 279±290. van der Lelie, D., Corbisier, P., Baeyens, W., Wuertz, S., Diels, L., Margeay, M., 1994. The use of biosensors for environmental monitoring. Res. Microbiol. 145, 67±74. Weissenhorn, I., Mench, M., Leyval, C., 1995. Bioavailability of heavy metals and arbuscular mycorrhiza in a sewage-sludge amended sandy soil. Soil Biol. Biochem. 27 (3), 287±296.
www.elsevier.com/locate/envpol
A microbial biosensor to predict bioavailable nickel in soil and its transfer to plants C. Tibazarwa a,b, P. Corbisier a, M. Mench c, A. Bossus a, P. Solda, M. Mergeay a, L. Wyns b, D. van der Lelie a,* a
Environmental Technology Expertise Centre, Flemish Institute of Technological Research (Vito), Boeretang 200, B-2400 Mol, Belgium Department of Ultrastructure, Free University of Brussels, Flemish Interuniversity Institute of Biotechnology (VIB), Paardenstraat 65, B-1640 St-Genesius-Rode, Belgium c Agronomy Unit, Inra Bordeaux Aquitaine Research Center, BP 81, 33883 Villenave d'Ornon, France
b
Received 8 December 1999; accepted 8 July 2000
``Capsule'': A whole-cell biosensor was developed for the detection of bioavailable Ni2+ and Co2+ concentrations in soil samples. Abstract Ralstonia eutropha strain AE2515 was constructed and optimised to serve as a whole-cell biosensor for the detection of bioavailable concentrations of Ni2+ and Co2+ in soil samples. Strain AE2515 is a Ralstonia eutropha CH34 derivative containing pMOL1550, in which the cnrYXH regulatory genes are transcriptionally fused to the bioluminescent luxCDABE reporter system. Strain AE2515 was standardised for its speci®c responses to Co2+ and Ni2+. The detection limits for AE2515 were 0.1 mM Ni2+ and 9 mM Co2+, respectively. The signal to noise (S/N) bioluminescence response and the metal cation concentration could be linearly correlated: for Ni2+ this was applicable within the range 0.1±60 mM, and between 9 and 400 mM for Co2+. The AE2515 biosensor strain was found to be highly selective for nickel and cobalt: no induction was observed with Zn(II), Cd(II), Mn(II), Cu(III) and Cr(VI). In mixed metal solutions, the bioluminescent response always corresponded to the nickel concentrations. Only in the presence of high concentrations of Co2+ (2 mM), the sensitivity to nickel was reduced due to metal toxicity. AE2515 was used to quantify the metal bioavailability in various nickel-enriched soils, which had been treated with additives for in situ metal immobilisation. The data obtained with strain AE2515 con®rmed that the bioavailability of nickel was greatly reduced following the treatment of the soils with the additives beringite and steel shots. Furthermore, the data were found to correlate linearly with those on the biological accumulation of Ni2+ in speci®c parts of important agricultural crops, such as maize and potato. Therefore, the test can be used to assess the potential transfer of nickel to organisms of higher trophic levels, in this case maize and potato plants grown on nickel-enriched soils, and the potential risk of transfer of these elements to the food chain. # 2001 Elsevier Science Ltd. All rights reserved. Keywords: Bioavailability; Heavy metals; Nickel; Plants; Soil; Transfer
1. Introduction The development of metal-speci®c biosensor tools functioning on the basis of bioluminescent reporter systems, has been receiving increasing attention (van der Lelie et al., 1994; Brown et al., 1998; Corbisier et al., 1999). These biosensors are considered as key assets both for characterizing the extent of contaminated areas and * Corresponding author. Tel.: +32-14-335166; fax: +32-14580523. E-mail address: [email protected] (D. van der Lelie).
for following up the success or failure of bioremediation operations of large areas contaminated with heavy metals, such as the Guadiamas river valley (de Lorenzo and Kuenen, 1999). They are of particular relevance for the assessment of remediation strategies based on in situ immobilisation of heavy metals, where total metal concentrations cannot be used to determine the success or failure of the remediation operation. Some of the major advantages of biosensors are that they are relatively inexpensive to develop (especially true for whole cell biosensors), their rapidity, great sensitivity and ease of data handling. In environmental technology
0269-7491/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved. PII: S0269-7491(00)00177-9
20
C. Tibazarwa et al. / Environmental Pollution 113 (2001) 19±26
applications, the use of luminometry based assays is associated with easy-to-handle instrumentation and relatively uncomplicated data handling (Corbisier et al., 1996). Furthermore, these types of assays do not require complex and labour-intensive handling of the samples, compared with some of the more common chemical analyses used to predict availability of heavy metals, such as sequential extraction (Tessier et al., 1979). The accurate determination of the biologicallyavailable (bioavailable) heavy metal content of environmental samples is a topic which has been discussed extensively, particularly in the context of sampling techniques. The determination of the metal content in the water-extractable or Ca(NO3)2 extracted fraction, has been shown to provide a relatively good indication of the bioavailable metal content of the soil sample, depending on the soil characteristics (Weissenhorn et al., 1995). However, it has not been possible to assign one particular method for the absolute determination of the bioavailable metal fraction in a soil. The bacterial-based biosensors may prove the ideal solution since they can be used directly with the soil sample and can therefore determine bioavailable metal levels in situ (van der Lelie, 1998). A comparative study of the data obtained from a biosensor assay with those from sequential extractions of soil samples (chemical data), as well as accumulation of the metal within biological systems of higher trophic levels, such as plants (biological data), may be useful in resolving the technical merits of one method over another in determining the bioavailable metal fraction in soils and assessing their potential risk. Thus, the need for simple tests that predict uptake of heavy metal by organisms of higher trophic levels would prove invaluable in the risk assessment studies of environmental contamination. In this communication, we describe the development, standardisation and application to soil samples, of a whole cell biosensor strain based on the Ralstonia eutropha CH34 derivative, designated as strain AE2515, which responds speci®cally to nickel and cobalt. The construction of AE2515 is based on pMOL1550 (Tibazarwa et al., 2000), in which the regulatory genes, cnrYXH, of the cnr cobalt and nickel resistance determinant of Ralstonia eutropha CH34 (Liesegang et al., 1993) are transcriptionally fused upstream of the luxCDABE genes of pMOL877 (van der Lelie et al., 1997). A similar strategy has been used to construct metal-speci®c biosensor strains for Zn, Cu and Pb (Corbisier et al., 1999): these strains are referred to as BIOMET strains (Corbisier et al., 1996). Strain AE2515 was found to emit a metal-speci®c response induced by dierent concentrations of Ni2+ and Co2+. This tight metal-controlled regulation, prompted us to test for the biosensor potential of AE2515 and to assess the prospective applications of this strain in quantifying bioavailable levels of Ni2+ and Co2+ in environmental samples.
As proposed by Boisson et al. (1998), an attempt was made to provide an alternative test to assess the reduction of metal toxicity in contaminated soils following in situ immobilisation treatment with the soil additives beringite and steel shots. The same soils samples as described in Boisson et al. (1998) were tested so that a direct comparison could be made between the use of metal biosensors in the prediction and/or determination of metal bioavailability in the soils versus the uptake by plants. 2. Materials and methods 2.1. Construction of biosensor strain AE2515 The construction of R. eutropha AE2515, a derivative of R. eutropha CH34 (Mergeay et al., 1985) containing pMOL1550, has been described by Tibazarwa et al. (2000) and is schematically depicted in Fig. 1. Plasmid pMOL1550 is pMOL877, in which the cnrYXH regulatory region was cloned upstream of the luxCDABE reporter system. This resulted in inducible light production controlled by the cnr regulon. For biosensor purposes, cultures of AE2515 were freeze-dried and aliquoted into sterile ampoules (Corbisier et al., 1999). 2.2. Luminometry assays Luminometry assays were carried out using an ANTHOS LUCY1 luminometer (Anthos Labtech b.v., Heerhugowaard, The Netherlands) at 23 C, as previously described (Corbisier et al., 1999). For metal standards, duplicate samples were set-up per microtitre assay. As negative controls, eight reaction samples containing bidistilled water were included in the tests. The bioluminescence emitted (ALU) and the optical density (OD620 nm) of the cultures were measured over 16 h at 30 min intervals, and processed using the MIKROWIN software, as previously described (Corbisier et al. 1999). Further data processing was carried out using EXCEL 7.0 (Microsoft) or Origin 5.0 (MicrocalTMOriginTM, Microcal Software, Inc., Northampton, MA). 2.3. Soil sampling and preparation Soil samples and the experimental conditions of in situ metal immobilisation were previously described by Boisson et al. (1998) and are presented in Table 1. Brie¯y, soils from a long-term sewage sludge ®eld trial on an acid coarse sandy soils in the Bordeaux region, France, termed Louis Fargues, were treated with the soil additives beringite (BE) and steel shots (SS). The metal content of dierent parts of maize plants (cv. INRA 260) as well as potato tuber cultivars was determined using either inductively coupled plasma atomic emission spectrometry (ICP±AES) or gas ¯ame atomic
C. Tibazarwa et al. / Environmental Pollution 113 (2001) 19±26
21
Table 1 Characteristics of Louis Fargues sludge plots treated with chemical additives and analysed with the the BIOMET testa Parcel no.
Soil description
Soil additive
pH
Total (Ni2+), ppm
20 27 39 34 26 40 35 28 37 36
UNT CTRL FYM CTRL FYM+SS FYM+BE S1 CTRL S1+SS S1+BE S2 CTRL S2+SS S2+BE
None None Steel shots Beringite None Steel shots Beringite None Steel shots Beringite
6.32 6.70 6.68 7.53 6.75 7.00 7.69 7.00 7.19 7.54
2.00 51.92 15.77 10.88 44.481 59.26 66.97 179.45 215.46 190.32
a Louis Fargues sludge plots on an Arenic Udi¯uvent (sandy) soil were treated with the additives Beringite (BE) or Steel shots (SS) or left untreated (CTRL) to test for the in situ immobilisation of polluting heavy metals. The plots with FYM received farm yard manure at a loading rate of 10 t DM ha 1, with S1 received sewage sludge at loadings of 10 t DM ha 1 and those with S2 received sewage sludge at loadings of 100 t DM ha 1. The physical properties of the soils, including the pH of the water fraction, and total nickel levels (as determined by GFAAS) were determined. The texture of all soils was predominantly sandy (>90%).
tested with the nickel biosensor strain, AE2515. For each assay, metal standard solutions were included, in duplicate, while the non-induced biosensor control test was set up as eight reactions. The reconstituted soils were tested in dilution series: typically 1 to 1, 1 to 2 and 1 to 4, or depending on the apparent toxicity of the samples, and also in duplicate. 2.5. Determination of bioavailable levels of nickel in soils
Fig. 1. The BIOMET-Ni biosensor. A schematic diagram of the construction of the biosensor strain, AE2515.
absorption spectroscopy (GFAAS), depending on the chemical element and concentration. The AE2515 biosensor strain was tested on dierent soils currently being monitored as part of a larger study on the bioremediation and recovery of Ni2+ from contaminated sites (Boisson et al., 1998). 2.4. BIOMET assays on soil samples Soil samples were suspended in reconstitutive medium (RM), as described previously (Corbisier et al., 1999). The 11 soil samples of Louis Fargues were subsequently
Raw data on heavy metal availability were calibrated using metal standard solutions, as described in Corbisier et al. (1999). The signal to noise ratio (S/N) calculated for each soil sample was equated to the nickel concentration and ®nally to the Ni2+ equivalent on a weight to weight basis. 3. Results 3.1. Nickel and cobalt speci®c bioluminescence response of strain AE2515 To assess the speci®city of AE2515 as a metal biosensor, the strain was tested in the presence of varying concentrations of Ni2+, Co2+, Zn2+, Cr6+, Cr3+, Mn2+, Cd2+ and Cu2+. As shown in Fig. 2a, the response to nickel was prominent, with a maximum Signal/Noise (S/N) ratio, exceeding 1000 in the presence of 0.1 mM Ni2+. The maximal response to Co2+ corresponded to a S/N ratio of 100 with 1.6 mM Co2+
22
C. Tibazarwa et al. / Environmental Pollution 113 (2001) 19±26
Fig. 3. Correlation of Ni2+ and Co2+ concentrations with the response of the biosensor strain AE2515. Plotted is a trendline of the relationship between signal/noise and metal concentration, in mM for (a) Ni2+ and (b) Co2+. The error bars represent the average deviation from the mean value of ®ve independent experiments.
metal concentration and the bioluminescence response was perfectly linear between 9 and 400 mM (Fig. 3b).
Fig. 2. Induction of light production in presence of nickel and cobalt by the biosensor strain AE2515. The maximum signal/noise (ALU/ OD620 nm) is presented in function of the concentration of (a) Ni2+ and (b) Co2+. The error bars represent the average deviation from the mean value of three independent experiments.
(Fig. 2b). The maximal bioluminescent response was reached within 4±6 h after the start of experimentation. The data were found to be highly reproducible for a given batch of freeze-dried vials, and this over a period exceeding 8 months with storage at 4 C. The batch-tobatch variability was not determined. The metals Zn2+, Cr6+, Cr3+, Mn2+, Cd2+, Cu2+ Ð at non-toxic concentrations in the range of 0.1 mM Ð did not signi®cantly induce bioluminescence (S/N ratio <2, data not shown). The detection limits for Ni2+ and Co2+ were found to be 0.1 and 9 mM, respectively, and give an indication of the sensitivity of the biosensor for these metals. From recent data on the metal-controlled regulation of the cnr genes, it has been determined that the regulatory genes respond predominantly to Ni2+: therefore the sensitivity of the biosensor is inherently superior for Ni2+. For Ni2+, a perfect linear correlation was found to exist between the S/N ratio and metal concentrations from 0.1 to 60 mM (Fig. 3a): the equation of the ®tted trendline is also indicated. For Co2+, the correlation between
3.2. Selectivity of strain AE2515 bioluminescence responses in mixed-metal solutions The performance of the biosensor in complex samples was tested using mixed metal standards of the combination of Ni2+ and Co2+. As shown in Fig. 4, the Ni2+ concentrations were maintained at a constant level, while the Co2+ levels were varied, and the biosensor response monitored. As shown, the Ni2+-speci®c response was markedly reduced in the presence of high Co2+ levels. The greatest reduction was observed with 1.6 mM Co2+ causing a 62.5% reduction in the S/N for 0.3 mM Ni2+. From the slopes of the inhibition curves, it can be concluded that inhibition of the biosensor performance for nickel in the presence of high levels of Co2+ is due the toxicity of the latter to cellular growth of the biosensor strain. Thus, even in these mixed-metal solutions, the biosensor response was clearly targetted to Ni2+, thereby con®rming the selectivity of the biosensor for Ni2+. Similar observations were made for combinations of Ni2+ and other heavy metals (data not shown). 3.3. Application of strain AE2515 as a biosensor for the detection of bioavailable concentrations of nickel in soils The eectiveness of strain AE2515 to detect the bioavailable fraction of Ni2+ in soil samples was tested on the Louis Fargues soils, which continue to be monitored
C. Tibazarwa et al. / Environmental Pollution 113 (2001) 19±26
Fig. 4. Performance of the AE2515 biosensor strain in mixed metal solutions. The signal/noise values were compared for the biosensor in the presence of various ®xed Ni2+ concentrations and increasing Co2+ concentrations. Data is an average of two independent experiments, carried out in duplicate.
as part of an ongoing bioremediation study (Boisson et al., 1998). The BIOMET values for bioavailable nickel were compared with the calcium nitrate exchangeable fraction of the soils (chemical data). As shown in Fig. 5, a signi®cant correlation (R=0.97) was observed between the two types of data. According to the correlation equation, 50±60% of the the Ca(NO3)2 exchangeable nickel fraction is measured as bioavailable with the biosensor strain AE2515. These ®ndings corroborate those of Boisson et al. (1998), that Ca(NO3)2 fractions typically give a value greater than the bioavailable metal content. Since the soils were being monitored for the eect of the additives beringite and steelshots on the nickel bioavailability in the soil, we used the BIOMET data to determine the change in metal bioavailability in such treated soils, and compared these data with those of the calcium nitrate exchangeable fractions, as determined by Boisson et al. (1998). As shown in Fig. 6, the reduction in nickel bioavailability in the presence of the soil additives as determined with the biosensor strain, corroborates the ®ndings of Boisson et al., although does not duplicate their data. For the farm yard manure soils treated with steelshots, no Ca(NO3)2 data was available. However, the BIOMET data revealed that this treatment resulted in a clear reduction in bioavailability. In general, the Ca(NO3)2 fractions give higher percentage reduction values than the AE2515 biosensor strain. 3.4. Correlation between bioavailable nickel levels determined using the BIOMET nickel biosensor test versus metal accumulation in plants The ecological relevance of the BIOMET-nickel test was evaluated by examining its correlation with other tests used to quantify metal accumulation in organisms of higher trophic levels, such as plants. In general, heavy metal accumulation tests in plants are used to assess the potential transfer of heavy metals to the food chain. In
23
Fig. 5. Correlation of concentration of bioavailable nickel, as determined with the biosensor strain AE2515 and the calcium nitrate exchangeable nickel fraction in dierent Louis Fargues soils. The AE2515-determined data on nickel bioavailability dierent samples of Louis Fargues soils are plotted against the calcium nitrate data obtained by Boisson et al., 1998.
Fig. 7, the BIOMET data for nickel-contaminated soils Ð with or without additive treatment Ð are compared with the levels of nickel in dierent parts of maize or potato plants grown on the same soils. Interestingly, the BIOMET quantitation of Ni2+ levels correlates linearly with the levels obtained for nickel accumulation in the grains (Fig. 7a), the 6th (Fig. 7b) and ear leaves (Fig. 7c) of maize plants (determined by GFAAS spectrophotometry analyses; see Boisson et al., 1998). Nickel accumulation in potato tubers also showed a signi®cant, although not linear, correlation, (Fig. 7d). 4. Discussion This communication represents a comparison of the performance of the nickel-speci®c biosensor strain AE2515 with physicochemical and biological methods used to determine bioavailable heavy metal. Heavy metal-speci®c biosensors based on gene fusions are now being widely applied in many laboratories (Corbisier et al., 1999), and this has resulted in the development of an amalgam of dierent biosensor strains, termed BIOMET strains, speci®c for dierent metal substrates such as Zn(II), Cd(II), Cu(II), Pb(II) and Cr(VI). A prerequisite for the construction of these heavy metal speci®c biosensors is a thorough understanding of bacterial heavy metal resistance systems, especially their metal speci®c regulation. Our recent understanding of the regulatory mechanism encoded by the cobalt and nickel resistance cnr determinant, has allowed the construction of a biosensor responsive to induction by these metals based on CH34 (Tibazarwa et al., 2000). The high selectivity for nickel was demonstrated in a mixed metal assay. The
24
C. Tibazarwa et al. / Environmental Pollution 113 (2001) 19±26
Fig. 6. The eects of dierent treatments of additives on nickelenriched Louis Fargues soils on nickel bioavailability, as determined with the AE2515 nickel biosensor strain. The bioavailable Ni2+ levels determined using the AE2515 biosensor were used to calculate the change in bioavailability of the metal in three Louis Fargues soils diering in composition Ð S1, S2 and farm yard manure (FYM) Ð following treatment by in situ immobilisation with the additives, Steelshots (SS) or Beringite (BE) or no treatment (CTRL), as described in Boisson et al., 1998. (a) S1 soil, which had received doses of sewage sludge at loadings of 10 t DM ha 1; prior to; (b) S2, which had received doses of sewage sludge at loadings of 100 t DM ha 1; and (c) FYM, which had received doses of farm yard manure at loadings of 10 t DM ha 1. A comparison with chemical data on nickel bioavailability, as obtained by calcium nitrate extraction (Boisson et al.) is also included.
linear correlation between the concentration of Ni2+ and the bioluminescent (S/N ratio) response at low metal concentrations has enabled quantitative analyses of bioavailable nickel levels to be made in environmental samples. Furthermore, the bioavailable fraction of the metal detected in nickel-enriched soils could be correlated to the biological accumulation of Ni2+ in several organs of maize plants, notably the grains, as well as potato tubers. Plant-based tests are typically used to give an indication of the relative toxicity and danger of a metal-contaminated soil (Mench et al., 1994) as well as its potential health risks. Comparing the application of the BIOMET Ni-biosensor test versus the growth of maize and potato plants on nickelcontaminated soil, it can be argued that the BIOMET test oers the advantage of low cost and rapidity (results within 6 h), whereas the plants must be monitored over 6 months, after which heavy metal levels must be measured on pretreated plant material with techniques such as ICP±AES. The rapid assessment of heavy metal bioavailability and uptake by plants is an important parameter in several respects. First, it allows a rapid, cost-ecient and early assessment of the potential risks of heavy metal transfer linked to crop species grown on heavy metal enriched soils. The monitoring of this risk commonly involves the growth of crops on such soils. Secondly, a correlation between heavy metal bioavailability and plant uptake data can be used for modelling and in predicting the applicability of dierent phytoremediation techniques as ecient remediation strategies for heavy metal polluted soils. This is of particular relevance for both in situ immobilisation, aiming at reducing the plant-available heavy metal fraction and its uptake by plants, and phytoextraction, where information on the plant-available fraction is a prerequisite to predict the eciency of the remediation. We show here that the BIOMET nickel biosensor can be used to rapidly (and cost eectively) assess plant availability of the heavy metal. Studies are presently being carried out to demonstrate this concept for other BIOMET heavy metal biosensors. In this same context of heavy metal bioavailability, we show also that the BIOMET nickel biosensor can be used to assess the eects of dierent in situ immobilisation treatments, aiming to decrease the heavy metal bioavailability to plants. However, a shortcoming of the BIOMET test might be that it is not able to predict the eects of treatment with chemical chelators, which are sometimes used to increase uptake of heavy metals by plants (Huang and Cunningham, 1996). For instance, metal-EDTA complexes are not detected by BIOMET, as was shown for Cu-EDTA complexes (Corbisier et al., unpublished data). On a cost-to-cost basis, the use of the BIOMET would outweigh that of the use of plant-based toxicity tests. It is therefore proposed that the BIOMET test is
C. Tibazarwa et al. / Environmental Pollution 113 (2001) 19±26
25
Fig. 7. Correlation between nickel concentrations detected in crops with those detected by the nickel biosensor test. The bioavailable Ni2+ levels determined (in ppm) using the AE2515 biosensor were compared with the nickel accumulation in dierent parts of maize and potato tubers (in ppm). Nickel concentrations in plants were determined by ICP. Panels are: (a) grains of maize, (b) ear leaf of maize, (c) 6th leaf of maize and (d) potato tubers.
applied as an alternative management tool for the assessment of heavy metal polluted soils and the followup of their (bio)remediation. Although the use of BIOMET should reduce the number of plant-based tests, it should nevertheless be used in conjunction with plant studies, which are necessary for the assessment of potential toxicity or de®ciency eects due to the remediation strategy, such as the application of soil additives (Boisson et al., 1998). Acknowledgements We thank Ann Provoost and Sa®eh Taghavi for help with molecular biology procedures. This work was supported by a grant to C. T. from the Flemish Government and the VIB (Vlaamse Instelling voor Biotechnologie). References Boisson, J., Mench, M., Sappin-Didier, V., Solda, P., Vangronsveld, J., 1998. Short-term in situ immobilisation of Cd and Ni by
beringite and steel shots application to long-term sludged plots. Agronomie 18, 347±359. Brown, N.L., Lloyd, J.R., Jakeman, K., Hobman, J.L., Bontidean, I., Mattiasson, B., CsoÈregi, E., 1998. Heavy metal resistance genes and proteins in bacteria and their application. Biochem. Soc. Trans. 26 (4), 662±665. Corbisier, P., van der Lelie, D., Borremans, B., Provoost, A., de Lorenzo, V., Brown, N., Lloyd, J., Hobman, J., CsoÈregi, E., Johansson, G., Mattiasson, B., 1999. Whole cell- and protein-based biosensors for the detection of bioavailable heavy metals in environmental samples. Anal. Chim. Acta 387, 235±244. Corbisier, P., Thiry, E., Diels, L., 1996. Bacterial biosensors for the toxicity assessment of solid wastes. Environ. Toxicol. Water Qual. 11, 171±177. de Lorenzo, V., Kuenen, G., 1999. Scienti®c basis for the bioremediation of the toxic spill of the Aznalcollar mine: combining bacteria and plants to address an intractable kind of pollution. Environ. Microbiol. 1 (3), 275±278. Huang, J.W., Cunningham, S.D., 1996. Lead phytoextraction: species variation in lead uptake and translocation. New Phytol. 134, 75±84. Liesegang, H., Lemke, K., Siddiqui, R.A., Schlegel, H.G., 1993. Characterization of the inducible nickel and cobalt resistance determinant cnr from pMOL28 of Alcaligenes eutrophus CH34. J. Bacteriol. 175, 767±778. Mench, M., Martin, E., Solda, P., 1994. After eects of metals derived from a highly metal-polluted sludge on maize (Zea mays). Water Air Soil Pollution 75, 277±291.
26
C. Tibazarwa et al. / Environmental Pollution 113 (2001) 19±26
Mergeay, M., Nies, D.H., Schlegel, H.G., Gerits, J., Charles, P., Van Gijsegem, F., 1985. Alcaligenes eutrophus CH34 is a facultative chemolithotroph with plasmid bound resistance to heavy metals. J. Bacteriol. 162, 328±334. Tessier, A., Campbell, P.G.C., Bissonm, M., 1979. Sequential extraction procedure for the speciation of particulate trace metals. Anal. Chem. 51, 844±850. Tibazarwa, C., Wuertz, S., Mergeay, M., Wyns, L., van der Lelie, D., 2000. Regulation of the cnr cobalt and nickel resistance determinant of Ralstonia eutropha-like CH34 (ex Alcaligenes eutrophus CH34). J. Bacteriol. 182 (5), 1399±1409. van der Lelie, D., 1998. Biological interactions: the role of soil bacteria in the bioremediation of heavy metal-polluted soils. In: Van-
gronsveld, J., Cunningham, S. (Eds.), From Metal-contaminated soils: in situ inactivation and phytorestoration. Landes Bioscience, Springer-Verlag, Berlin, pp. 31±50. van der Lelie, D., Regniers, L., Borremans, B., Provoost, A., Verschaeve, L., 1997. The VITOTOX test, a SOS bioluminescence Salmonella typhimurium test to measure genotoxicity kinetics. Mut. Res. 2-3, 279±290. van der Lelie, D., Corbisier, P., Baeyens, W., Wuertz, S., Diels, L., Margeay, M., 1994. The use of biosensors for environmental monitoring. Res. Microbiol. 145, 67±74. Weissenhorn, I., Mench, M., Leyval, C., 1995. Bioavailability of heavy metals and arbuscular mycorrhiza in a sewage-sludge amended sandy soil. Soil Biol. Biochem. 27 (3), 287±296.