2973 Environmental Toxicology and Chemistry, Vol. 24, No. 11, pp. 2973–2982, 2005  2005 SETAC Printed in the USA 0730-7268/05 $12.00 + .00 Hazard/Risk Assessment BIOTESTS AND BIOSENSORS IN ECOTOXICOLOGICAL RISK ASSESSMENT OF FIELD SOILS POLLUTED WITH ZINC, LEAD, AND CADMIUM ANNE KAHRU,*² A NGELA IVASK,² K AJA KASEMETS,² L EE PO ˜ LLUMAA,² I MBI KURVET,² M ATTHIEU FRANC ¸ OIS,‡ and HENRI-CHARLES DUBOURGUIER‡ ²National Institute of Chemical Physics and Biophysics, Akadeemia tee 23, Tallinn 12618, Estonia ‡Institut Supe ´rieur d’Agriculture, 48 Boulevard Vauban, 59046 Lille Cedex, France ( Received 5 January 2005; Accepted 3 May 2005) Abstract—The combined chemical and ecotoxicological hazard evaluation study was conducted on 60 smelter-influenced soils containing 1 to 13, 50 to 653, and 100 to 1,198 mg/kg of Cd, Pb, and Zn, respectively. For these soils (liquid-to-soil ratio = 10), water extractability of Zn, Cd, and Pb was less than 0.19% (median values). Acetic acid (0.11 M) extracted 23, 9.7, and 0.7% of Cd, Zn, and Pb, respectively. Although heavy metal concentrations in the studied soils were high, the toxic effects of water extracts were observed only in few samples and in few biotests (algae Selenastrum capricornutum and metal detector assay). For most of the aquatic test organisms (e.g., crustaceans, photobacteria), the bioavailable concentrations of metals in soil–water extracts were either subtoxic, or the adverse effects were compensated by soil nutrients, etc. However, analysis of the soils with recombinant Cd sensor Bacillus subtilis (pTOO24) showed that about 65% of these apparently subtoxic samples contained bioavailable Cd when analyzed in the suspension assay (detection limit 1.5 mg Cd/kg soil), indicating the desorption of Cd induced by direct contact of bacteria with soil particles. The median bioavailable fraction of Cd (1%) was 23-fold lower than the fraction extracted by acetic acid. The Pb–Cd sensor Staphylococcus aureus (pT0024) detected bioavailable Pb only in the suspensions of five of the most lead- polluted soils (417 mg Pb/kg): the median bioavailability of Pb was 0.42%. Consequently, the hazard assessment relying on total metal levels in soils should be revised by critical comparison with data obtained from bioassays. Development and use of biosensors (excellent tools for mechanistic studies and signaling hazard already at subtoxic level) should be encouraged. Keywords—Bioavailability Ecotoxicity Heavy metals Polluted soils Biosensors INTRODUCTION In parallel to the increase of the human population, the production of metals increases and, due to the badly managed emission control and sorptive capacity of soils and sediments, trace metals (including toxic heavy metals Pb, Cd, Hg, etc.) are accumulating steadily in the environment. The environ- mental risk assessment of heavy metal polluted soils and sed- iments usually is based on the total amounts of metals, i.e., amounts quantified after digestion of soils with strong acids. The shortcomings of this approach, i.e., inability to discrim- inate between the hazardous and not hazardous fractions of metals, have been discussed in almost every scientific paper dealing with bioavailability of heavy metals in soils [1]. Be- cause the bioavailability of a certain pollutant is its uptake potential by living organisms, it depends on the nature and the physico-chemical properties of this toxicant, on the environ- mental matrix (groundwater, soil, sediment); the exposure route for the biological recipient; and its physiology. The Or- ganization for Economic Cooperation and Development (OECD) toxicity tests for soils include assays with earth- worms, bacteria, and plants. These tests currently are applied mainly for testing the environmental hazard of new chemicals. For large-scale soil pollution studies, the tests with earthworms and plants are labor and space consuming. Also, it has been shown that earthworms are insensitive to heavy metal pollu- tion: In spiked OECD artificial soil, the 14-d Eisenia fetida half-lethal concentration (LC50) values were 4,480, 1,010, and 300 mg/kg of soil for Pb, Zn, and Cd, respectively [2]. Anal- ogously to earthworm tests, collembola Folsomia candida re- * To whom correspondence may be addressed (anne@kbfi.ee). production-inhibition tests in spiked OECD artificial soil re- sulted in 50% effective concentration (EC50) values quite sim- ilar to those obtained with earthworms: The 28-d EC50 values were 2,970, 900, and 590 mg/kg of soil for Pb, Zn, and Cd, respectively [3]. Because microbes are a crucial part of the soil microflora, bacterial activity often is measured to char- acterize the toxicity of polluted soils. Soil microbes have prov- en more sensitive to heavy metal pollution than other members of soil biota [4]. The toxicity for soil microbial activity, mea- sured as 50% inhibition in dehydrogenase assay of spiked loess soil [5] (652 mg Pb/kg soil, 115 mg Zn/kg, and 90 mg Cd/ kg), is indeed somewhat higher compared to the toxicity to earthworms and collembolas, but could be explained by the soil properties (particularly 1.1% of organic matter in loess soil compared to 10% of organic matter in artificial soil). In addition, natural microbial communities of metal-polluted soils adapt to the pollution. In soils influenced by metal smelters, the toxicity of Pb and Zn to microbial soil respiration was very low [6]; a 50% reduction rate occurred only at as high concentrations as 25,000 mg Pb/kg and 4,000 mg Zn/kg. Thus, it may be concluded that the bioaccessibility and bioavail- ability of heavy metals in soils (both spiked and naturally polluted) apparently is rather low. The latter is reflected by relatively high permitted limit values (PLVs) for heavy metals in agricultural soils according to European Council Directive 86/278/ (European Economic Community): 1 to 3, 50 to 300, and 150 to 300 mg/kg for Cd, Pb, and Zn, respectively [7]. The potential transfer of pollutants to groundwater mainly is determined by their aqueous solubility. Lately it has been shown that, even in the case of earthworms, the main exposure to heavy metals occurs via dermal contact with pore water and