Enzyme-Linked Immunofiltration Assay To Estimate Attachment of Thiobacilli to Pyrite

Abstract

An enzyme-linked immunofiltration assay (ELIFA) has been developed in order to estimate directly and specifically Thiobacillus ferrooxidans attachment on sulfide minerals. This method derives from the enzyme-linked immunosorbent assay but is performed on filtration membranes which allow the retention of mineral particles for a subsequent immunoenzymatic reaction in microtiter plates. The polyclonal antiserum used in this study was raised against T. ferrooxidans DSM 583 and recognized cell surface antigens present on bacteria belonging to the genus Thiobacillus. This antiserum and the ELIFA allowed the direct quantification of attached bacteria with high sensitivity (10⁴ bacteria were detected per well of the microtiter plate). The mean value of bacterial attachment has been estimated to be about 10⁵ bacteria mg⁻¹ of pyrite at a particle size of 56 to 65 μm. The geometric coverage ratio of pyrite by T. ferrooxidans ranged from 0.25 to 2.25%. This suggests an attachment of T. ferrooxidans on the pyrite surface to well-defined limited sites with specific electrochemical or surface properties. ELIFA was shown to be compatible with the measurement of variable levels of adhesion. Therefore, this method may be used to establish adhesion isotherms of T. ferrooxidans on various sulfide minerals exhibiting different physicochemical properties in order to understand the mechanisms of bacterial interaction with mineral surfaces.

The chemolithotrophic acidophilic bacterium Thiobacillus ferrooxidans, which uses sulfide mineral oxidation (S²⁻, S⁰, Fe²⁺) for growth, is of environmental and industrial interest due to its involvement in the biocycling of sulfur and iron, in the bioleaching and biovalorization of minerals, and in the acidification of waters and soils (5). Such bacterial oxidation and dissolution processes for nonsoluble minerals involve interfacial phenomena between bacteria and minerals, suggesting that microbial contact and attachment play a major role in sulfide oxidation (6, 22, 27).

The observation and measurement of bacterial attachment to mineral surfaces and the role of this process are of major interest to understand and control the mechanisms of sulfide dissolution and weathering. T. ferrooxidans attachment has been estimated by microscopic counting of nonadhering bacteria after contact between bacterial and mineral suspensions (17, 24, 29). This method has a low sensitivity and needs at least 5 × 10⁵ to 10⁶ bacteria ml⁻¹ to get a good measurement. Furthermore, cell multiplication does not allow the use of such a method to determine the dynamics of bacterial adhesion over time.

The adhesion of T. ferrooxidans was also evaluated by total protein determination with sulfide minerals collected after inoculation (32) and by epifluorescence microscopy (31). Accurate and reproducible results by these methods require large amounts of samples or high cell densities and a large number of assays (16, 32). To improve epifluorescence measurements, complementary analyses such as image analysis were recently developed. But image analysis requires regular surfaces (33) and would not be suitable for the determination of cell attachment onto mineral particles. To determine cell attachment, a spectrofluorimetric assay was also used to measure the intensity of fluorescently labeled adhering bacteria in the upper layer of sedimented inoculated mineral particles (8). The sensitivity of this measurement of adhering cells is also very low compared to that of fluorescence measurement of nonadhering cells and requires at least 10⁶ cells mg⁻¹ of particles. In addition, this method also may be prone to error due to the interference of cell treatment during the labeling reaction with the attachment ability of the treated bacteria.

The most-probable-number (MPN) technique was also used to indirectly count ferrous-iron-oxidizing adhering bacteria, after their desorption from the mineral by vortexing (30). This method was also tedious and in addition nonspecific. In fact, all these previous methods do not distinguish specifically T. ferrooxidans from other similar bacteria.

Jerez and Arredondo (13, 14) and Amaro et al. (1) have developed immunological methods for the enumeration of acidophilic bacteria. These methods are very sensitive and can also be serotype specific (11, 18). Consequently, they can allow the enumeration of bacteria belonging to one species by the selection of an antiserum which recognizes all the serotypes of that species. However, those methods have only been developed for counting nonadhering acidophiles. Muyzer et al. (21) have attempted the use of polyclonal antibodies for a combined immunofluorescence–DNA-fluorescence staining procedure to observe the abundance of T. ferrooxidans on coal particles. But unfortunately, detection of T. ferrooxidans was not observed at the end of the experiments; detection was not possible probably because this bacterium was outcompeted by other acidophilic microorganisms.

So, in order to observe and estimate directly, more specifically, and accurately T. ferrooxidans attachment on sulfide minerals during leaching processes, experiments have been performed to develop and evaluate an enzyme-linked immunofiltration assay (ELIFA). The assessment of this method, derived from the enzyme-linked immunosorbent assay (ELISA), has been performed by comparison of the method with indirect and more classical bacterial cell counting methods. To directly perform ELIFA on mineral particles, microtiter plates with wells having a filter membrane at the bottom have been used.

MATERIALS AND METHODS

Bacterial strains and preparation of cultures.

The bacteria used in this study are listed in Table 1. T. ferrooxidans strains and isolates were grown in 72 mM ferrous sulfate or 2% (wt/vol) pyrite–M₂ basal salts medium (with compounds at the following concentrations [in grams per liter]: (NH₄)₂SO₄, 1.0; KH₂PO₄, 0.4; MgSO₄ · 7H₂O, 0.4), pH 1.8 (19). Various other bacteria were also used to determine the specificity of the antiserum raised against T. ferrooxidans DSM 583. Leptospirillum ferrooxidans CF12, Thiobacillus thiooxidans DSM 504, Acidiphilium strains, and the acidophilic heterotroph isolate T23 were grown in the basal salts medium described by Postgate (25). L. ferrooxidans CF12 was grown at pH 1.8 on a 100 mM ferrous sulfate–0.05% (vol/vol) trace element solution (15). T. thiooxidans DSM 504 and Acidiphilium strains were grown at pH 2.5 on 1% (wt/vol) elemental sulfur–0.05% (vol/vol) trace element solution and on 10 mM glycerol–0.02% (wt/vol) yeast extract, respectively. Isolate T23 was grown at pH 2.0 on 10 mM ferrous sulfate–0.02% (wt/vol) yeast extract. Leptospirillum-like strain L8 was grown in the liquid medium of Silverman and Lundgren (28) containing 144 mM ferrous sulfate and 100 mM ferric sulfate, pH 1.6 (3). Culturing was performed on a rotary shaker at 130 rpm at 28°C. Bacteria were harvested by centrifugation and then washed three times with 0.01 M H₂SO₄ and resuspended in Tris-buffered saline (TBS; pH 7.6). The strains of nonacidophilic gram-negative bacteria Pseudomonas corrugata ATCC 29736, Burkholderia cepacia ATCC 25416, and Azospirillum brasilense SP7 were also used to determine the specificity of the antiserum. They were cultivated in nutrient broth (Difco) at 28°C, harvested by centrifugation, and then washed twice and resuspended in TBS.

TABLE 1.

Bacterial strains used in these experiments

Bacterial species or strain	Code	Origin
T. ferrooxidans DSM 583	DSM 583	DSMa
T. ferrooxidans ATCC 33020	ATCC 33020	ATCCb
T. ferrooxidans Tf 2	Tf2	S. N. Groudevc
T. ferrooxidans-like isolate OP14	OP 14	D. B. Johnsond
T. thiooxidans DSM 504	DSM 504	DSM
L. ferrooxidans CF 12	CF12	D. B. Johnson
Leptospirillum-like strain L8	L8	D. Morine
Acidiphilium organovorum	Ao	D. B. Johnson
Acidiphilium isolate SJH (NCIB 12826)	SJH	D. B. Johnson
Acidophilic iron-oxidizing heterotroph T23	T23	D. B. Johnson
P. corrugata ATCC 29736	Pc	LEMIRf
B. cepacia ATCC 25416	Bc	LEMIR
A. brasilense SP7	Ab	LEMIR

^aDSM, Deutsche Sammlung von Mikroorganismen, Braunschweig, Germany. 

^bATCC, American Type Culture Collection, Manassas, Va. 

^cUniversity of Mining and Geology, Sofia, Bulgaria. 

^dSchool of Biological Sciences, University of North Wales, Bangor, United Kingdom. 

^eBRGM, Orleans, France. 

^fLEMIR, Laboratoire d’Ecologie Microbienne de la Rhizosphère, CEA, Saint Paul Lez Durance, France. 

Sulfide minerals.

The pyrites used in these experiments originated from a hydrothermal Peruvian deposit (pyrite K₄) and a sedimentary deposit in Spain (pyrite ES). The mineral particles were collected by wet sieving after dry grinding in a tungsten carbide mill. The particle size distribution was determined by laser scattering particle size analysis (Malvern Master Sizer). The mean particle diameter was 65 μm for pyrite K₄ and 56 μm for pyrite ES. From these data, and considering the particles as spheres, particle surface areas were estimated to be 190 and 220 cm² g⁻¹, respectively. Infrared spectrometry indicated more pronounced mineral-oxidized species [FeSO₄, Fe₂(SO₄)₃] at the surface of pyrite K₄ than at the surface of pyrite ES. Chemical analyses of the pyrites are presented in Table 2.

TABLE 2.

Main chemical properties of the pyrites

Pyrite	Chemical content (Fe/S/As; % [wt/wt])	Purity (% [wt/wt])
K4	45.88/53.65/0.0089	99.53
ES	45.36/52.97/0.34	98.33

Preparation of antiserum and specificity determination.

A polyclonal antiserum was raised against the pure strain of T. ferrooxidans DSM 583. The bacteria were grown in the presence of pyrite, as the sole source of energy, and harvested at exponential and early stationary phases. Bacteria were prepared as described by Muyzer et al. (21), and a mixture of equal amounts of cells from these culture stages (10⁹ cells ml⁻¹) was used for the immunization of a rabbit (A. Dorier, Institut de Biologie Appliquée, Villeurbanne, France).

Determination of attachment of T. ferrooxidans to pyrite.

T. ferrooxidans DSM 583 was used for attachment experiments. Cells adapted either to pyrite K₄ (Tf/K₄ cells) or to ferrous sulfate (Tf/Fe²⁺ cells) by three successive cultures were inoculated onto pyrite ES and pyrite K₄. A total of 2 × 10⁹ bacteria ml⁻¹ were resuspended in 200 ml of M₂ basal salts medium containing 4 g of pyrite and incubated for 24 h at 30°C with stirring (750 rpm) in batch reactors described previously (20).

Bacterial attachment was evaluated directly by ELIFA performed on adhering cells and indirectly by enumeration of nonadhering cells by either the MPN method or direct microscopic counting with a Thoma cell. Counting the nonadhering bacteria allowed the indirect determination of attachment by calculating the difference between the numbers of cells in the medium before and after contact with pyrite. Measurement of attachment was performed before cell multiplication, as the generation times of T. ferrooxidans on pyrite ES and pyrite K₄ were 30 to 40 h and 40 to 60 h, respectively. Moreover, cell growth is preceded by a lag phase of 24 h for pyrite ES and by a longer one for pyrite K₄.

ELIFA.

ELIFA is a modified ELISA method using a microtiter plate with wells having a 0.2-μm-pore-size Durapore filter membrane at the bottom (Multiscreen MAGVN; Millipore). Inoculated pyrite particles were retained on the filter. Prior to deposition in the wells of the microtiter plate, pyrite samples were washed in a tube two times with 10 volumes of M₂ basal salts medium to remove nonattached bacteria, i.e., bacteria present in the interstitial solution between pyrite particles. For comparison and estimation of standard curves, the reaction was also performed at the same time with bacteria remaining in the supernatant after harvesting the pyrite. The ELIFA was derived from the method described by Gouzou et al. (10).

Suspended bacteria (100 μl of suspension) or inoculated and washed pyrite particles were added to each well of the microtiter plate. The medium was removed by filtration using an accessory vacuum filter holder (Multiscreen filtration system; Millipore) which allows the simultaneous processing of 96 individual samples. The wells were washed three times with 200 μl of TBS. To saturate the remaining binding sites on the filter and on pyrite particles, the wells were incubated at 37°C for 1 h with 100 μl of 1% (wt/vol) bovine serum albumin (BSA; Sigma) in TBS. They were then rinsed three times with 200 μl of TBS–0.05% (vol/vol) Tween 20 (Prolabo). Then, 100 μl of diluted antiserum (1:10,000) in TBS containing 1% BSA was added to each well. After 1 h of incubation at 37°C, the wells were again rinsed three times with TBS-Tween 20 and then incubated with 100 μl of goat anti-rabbit globulin (Sigma; A 75-39) alkaline phosphatase conjugate, diluted (1:2,000) in TBS-Tween-BSA. The wells were rinsed two times with TBS-Tween and one time with 10% (vol/vol) diethanolamine (Prolabo) buffer (pH 9.8). Thereafter, the samples were incubated with 100 μl of p-nitrophenylphosphate (Boehringer GmbH, Mannheim, Germany; 1 mg/ml, in diethanolamine buffer) as a substrate. Absorbance at 405 nm was measured with a MR 700 RS Dynatech spectrophotometer. Controls in each experiment consisted of a 0.2-μm filtrate of the bacterial suspension and of noninoculated pyrite particles.

To determine the number of bacteria fixed per milligram of pyrite, the weight of pyrite particles used in the ELIFA reaction was determined as follows. Pyrite particles present in each well were dissolved in 10 ml of nitric acid (69%; analytical reagent; Prolabo). Total solubilized iron, determined by inductive coupled plasma analysis (Jobin Yvon JY 38) using an internal yttrium standard (40 ppm), was used to calculate the amount of pyrite according to the chemical composition of the pyrite.

The statistical analysis of the results obtained by the ELIFA has been performed by using the Student t test. The mean and its confidence limits have been calculated considering that the probability of the results to be excluded from these limits was 5%.

Enumeration of nonadhering thiobacilli by the MPN technique.

The MPN technique was performed with microtiter plates as described by Rouas (26). Each bacterial dilution was tested in 40 wells. Twenty-five microliters of culture dilution was added to each well, which contained 200 μl of M₂ basal salts medium supplemented with 72 mM ferrous sulfate. After 14 days of incubation at 30°C, the wells presenting an orange color due to bacterial iron oxidation were counted, and the MPN and its confidence limits were determined by using a statistical program (12).

RESULTS

Specificity of the antiserum.

The antiserum raised against T. ferrooxidans DSM 583 showed a strong positive reaction with cells of T. ferrooxidans DSM 583 and Tf 2 and T. ferrooxidans-like isolate OP14 but not with cells of T. ferrooxidans ATCC 33020 (Fig. 1). It also exhibited a strong reaction with cells of T. thiooxidans DSM 504. Only weak and nonsignificant reactions were obtained with L. ferrooxidans CF12 and Leptospirillum-like strain L8 and with the acidophilic heterotrophs including Acidiphilium isolates Ao and SJH. The iron-oxidizing heterotroph T23 and the gram-negative neutrophiles P. corrugata, B. cepacia, and A. brasilense SP7 also exhibited no significant reaction (Fig. 1). These results indicate that the antiserum raised against T. ferrooxidans DSM 583 can recognize some cell surface antigens present on bacteria belonging to the genus Thiobacillus.

FIG. 1.

Reactivities of the antiserum raised against T. ferrooxidans DSM 583 to various bacterial strains and isolates. The ELISA reaction of exponential-phase bacteria (5 × 10⁶ cells per well) was determined for a dilution of the antiserum of 1:10,000. OD (405 nm), optical density at 405 nm. The data are means of three determinations with standard deviations. See Table 1 for strain designations.

The weak reaction obtained with T. ferrooxidans ATCC 33020 indicates that this strain exhibits antigenic surface components different from those of the other T. ferrooxidans strains tested. As Koppe and Harms (18) have described different serotypes for T. ferrooxidans that are characterized by a specific lipopolysaccharide (LPS) banding pattern, our result suggests that the cells of T. ferrooxidans ATCC 33020 and DSM 583 have different cell surface LPSs and therefore belong to different serotypes.

No cross-reaction between cells belonging to the serotype of T. ferrooxidans DSM 583 (which is the same as that for T. ferrooxidans ATCC 19859 and ATCC 23270) and cells of T. thiooxidans DSM 504 has previously been reported (18, 21). Koppe and Harms (18) have shown that these strains have different LPS banding patterns, but our results suggest that the antiserum raised against T. ferrooxidans DSM 583 recognizes common antigens of T. ferrooxidans and T. thiooxidans, which are probably common antigenic proteins or glycoproteins.

Sensitivity of ELIFA reaction for T. ferrooxidans detection.

Bacterial suspensions of T. ferrooxidans DSM 583, previously grown in the presence of either ferrous sulfate (Tf/Fe²⁺ cells) or pyrite K₄ (Tf/K₄ cells), were incubated for 23 h in the presence of pyrite ES or K₄. Standard curves for ELIFA have been determined from bacteria remaining in the supernatant after the harvesting of the pyrite. The optical density resulting from the ELIFA reaction for nonadhering bacteria was compared with the number of bacteria estimated by the MPN method (Fig. 2). The MPN method was chosen after controlling the viability of nonadhering bacteria at about 100%.

FIG. 2.

Intensity of ELIFA reaction against nonadhering bacteria for different inocula of T. ferrooxidans DSM 583 after 23 h of incubation in the presence of pyrite. (A) Ferrous sulfate-adapted bacteria (Tf/Fe²⁺; □) and pyrite K₄-adapted bacteria (Tf/K₄; ⧫) incubated in the presence of pyrite ES. (B) Ferrous sulfate-adapted bacteria after incubation with either pyrite K₄ (⧫) or pyrite ES (□). The log of the number of bacteria per well was determined by the MPN technique, and the horizontal confidence limits are those of that determination. OD (405 nm), optical density at 405 nm. Each OD value is the mean of replicates of the ELIFA for the corresponding number of cells, and the vertical confidence limits were calculated from these replicates. The data are given as mean values ± 95% confidence intervals.

The lower threshold for the detection of bacteria by ELIFA was 0.7 × 10⁵ to 10⁵ bacteria ml⁻¹ (Fig. 2). Therefore, this method is more sensitive than the microscopic counts, for which a minimum of 5 × 10⁵ to 10⁶ bacteria ml⁻¹ were necessary. The optical densities obtained at 405 nm were significantly different for bacterial concentrations differing by 0.5 log units and showed the accuracy of this reaction.

Better sensitivity of detection was obtained for pyrite-adapted bacteria (Tf/K₄) than for ferrous sulfate-adapted bacteria (Tf/Fe²⁺) (Fig. 2A). This difference could be explained by the preparation of antibodies with T. ferrooxidans grown on pyrite.

After 1 day of incubation of the bacteria with two different pyrites (K₄ and ES), similar ELIFA reactions were observed (Fig. 2B).

Indirect measurement of attached bacteria by enumeration of nonadhering bacteria.

The concentrations of nonadhering bacteria in the medium before and after contact with pyrite were significantly different (data not shown). The numbers of attached bacteria, estimated from this difference by either MPN or direct counting, were similar (Fig. 3). Furthermore, similar levels of attachment were observed with Tf/Fe²⁺ and Tf/K₄ inocula in the presence of both pyrites ES and K4 (3 × 10⁵ to 9 × 10⁵ bacteria mg⁻¹ of pyrite).

FIG. 3.

Estimation of attachment of T. ferrooxidans DSM 583 to pyrites by various methods. Panels A and B represent the attachment of ferrous sulfate-adapted bacteria (Tf/Fe²⁺) and pyrite K₄-adapted bacteria (Tf/K₄), respectively, to pyrite ES. Panels C and D show the attachment of Tf/Fe²⁺ and Tf/K₄ bacteria, respectively, to pyrite K₄. The inoculum was 10⁷ bacteria ml⁻¹. Bars represent 95% confidence intervals of the mean. The counting methods used were direct microscopic counts with a Thoma cell (THOMA), MPN, and ELIFA.

This indirect method of attachment estimation was possible only for inocula with concentrations of 10⁷ bacteria ml⁻¹ or lower, because above this value the difference between the concentrations of nonadhering bacteria before and after contact was not significant. These results indicated the need for a direct method to measure bacterial attachment.

By estimating the number of attached bacteria per milligram of pyrite, it was possible to determine the coverage ratio of pyrite by bacteria. It was assumed that the average bacterial cell surface was about 0.5 μm². On the other hand, the specific surface area of pyrite particles calculated after laser scattering particle size analysis was 200 cm² g⁻¹. Thus, the area covered by attached bacteria represented about 0.75 to 2.25% of the total geometric pyrite surface. As a consequence, the number of bacteria attached to each mineral particle was estimated to be 170 to 500 cells, assuming that the mean particle diameter was 60 μm. This low coverage ratio value suggests that an immunological method can allow an easier direct estimation of the adhesion level of these bacteria.

ELIFA reaction on T. ferrooxidans attached to pyrite particles.

Controls consisted of autoclaved mineral particles. The optical density obtained with sterile mineral particles was lower than 0.2 units after a 30-min reaction. In the presence of inoculated and washed pyrite, a fast and positive reaction was noted (data not shown). This result demonstrated the presence of attached bacteria on inoculated mineral particles.

In measurements performed on samples of pyrite ES and K₄ inoculated with T. ferrooxidans and having a weight less than 1.5 mg, a linear relationship between the amount of pyrite and the optical density was obtained (Fig. 4). These results confirmed that attached bacteria can be measured by ELIFA. As for the observed reaction with nonadhering bacteria, a better sensitivity was obtained for bacteria previously cultivated in the presence of pyrite (Tf/K₄) than in the presence of ferrous sulfate (Tf/Fe²⁺).

FIG. 4.

Intensity of ELIFA reaction for T. ferrooxidans DSM 583 attached to particles of pyrite ES after 23 h of incubation. The optical density (OD) was measured after a 20-min reaction. T. ferrooxidans was previously cultivated in the presence of ferrous sulfate (□) or pyrite K₄ (⧫). The correlation coefficients of the linear regression curves are r² = 0.962 and r² = 0.939, respectively.

To estimate the bacterial attachment, the optical density at 405 nm, measured with inoculated pyrite particles, was compared to standard curves (Fig. 2) obtained with nonadhering bacteria. This method allowed a direct and accurate estimation of the attached bacteria (95% confidence intervals lower than 0.3 log unit) (Fig. 3). Furthermore, the method was sensitive, as about 10⁴ bacteria could be detected per well of microtiter plate. As with indirect measurements of attached bacteria, similar adhesion levels (10⁵ bacteria mg⁻¹ of pyrite) were observed for the different inocula (Tf/Fe²⁺ and Tf/K₄) on both pyrites, showing that this ELIFA method gave dependable and homogeneous results. From this measurement of attachment level, the coverage ratio of pyrite by bacteria could be estimated to be about 0.25%, corresponding to a mean number of 57 attached bacteria per mineral particle.

The measurement of the attachment of T. ferrooxidans on pyrite was determined in another set of experiments. A comparison of the different methods was made with two different amounts of inoculum (Table 3). With an inoculum of about 10⁷ bacteria ml⁻¹, the adhesion level determined was similar to the previous one. For a larger inoculum (6.9 × 10⁷ bacteria ml⁻¹) high attachment levels were observed by indirect methods (1.6 × 10⁶ to 3.6 × 10⁶ bacteria mg⁻¹). This high number of attached bacteria was confirmed by ELIFA (up to 10⁶ bacteria mg⁻¹), showing that this method allows the determination of variable levels of adhesion.

TABLE 3.

Estimation of T. ferrooxidans DSM 583 attachment to pyrite for two different inoculum sizes by various counting methods

Culture of T. ferrooxidans	Avg inoculum size determined by MPN (no. of bacteria ml−1)	Attachment estimationa (no. of bacteria mg−1 of pyrite) by:
		Microscopic counting	MPN	ELIFA
Ferrous sulfate adapted	1.2 × 107	5.8 × 105	2.9 × 105–6.0 × 105	2 × 105–4 × 105
Pyrite adapted	6.9 × 107	1.6 × 106	1.7 × 106–3.6 × 106	8 × 105–106

^aThese results have been obtained from only two determinations when microscopic counts and ELIFA were used as the estimation methods; they are given without standard deviations or confidence intervals. 

DISCUSSION

The knowledge of reactions at the interface between sulfide minerals (e.g., pyrite) and bacteria such as T. ferrooxidans and of other phenomena such as chemotaxis and attachment is of major interest to understand and control the mechanisms involved in the bacterial dissolution and weathering of minerals and in bioleaching processes of metals.

Compared to previous quantitative indirect methods using counting of cells remaining in suspension (4, 19) and to qualitative methods using a combined immunofluorescence–DNA-fluorescence staining procedure (21), the immunoenzymatic ELIFA method performed on filtration membranes and presented in this paper allowed the direct quantification of T. ferrooxidans cells attached to pyrite particles. This method, performed with microtiter plates, is sensitive (10⁴ bacteria were detected per well) and accurate. A 95% confidence interval lower than 0.3 log units in the optical density at 405 nm was observed regularly.

The mean values of attached bacteria estimated by ELIFA and by the indirect determinations were 10⁵ and 5 × 10⁵ bacteria mg⁻¹, respectively. Compared with indirect measurements, ELIFA was shown to slightly underestimate bacterial adhesion. Such a result suggests that the accessibility of antiserum to bacterial surfaces was limited by the contact with the mineral. It can be considered that some antigenic determinants recognized by the antiserum may also be cell surface components that play a role in bacterial adhesion. Using immunoblotting of T. ferrooxidans lysates, Koppe and Harms (18) identified LPS and membrane proteins as antigenic determinants. Independently, bacterial adhesion studies performed by Arredondo et al. (2) suggested that both components are involved in the attachment of T. ferrooxidans to solid surfaces.

The ELIFA reaction obtained with T. ferrooxidans grown on pyrite was more sensitive than that observed with ferrous-iron-grown cells. This result suggests that the mineral can induce a modification in the expression of bacterial surface components. The presence of a solid substrate such as pyrite enhances the excretion of exopolymeric substances (EPS), and bound ferric ions in these EPS are responsible for the attachment of T. ferrooxidans to mineral particles (9).

The antiserum produced against the strain of T. ferrooxidans DSM 583 has been shown to recognize cell surface antigens present on bacteria belonging to the genus Thiobacillus. As ELIFA enables direct measurement of attached bacteria, the preparation and selection of an antiserum which recognizes the serotypes of one bacterial species may allow the determination of the attachment dynamics of this bacterial species during a complete bioleaching process. ELIFA can also be very useful to examine how various microorganisms interact with each other and with the surfaces of mineral particles. Attached T. ferrooxidans cells were recently evaluated by measuring their ability to oxidize ferrous iron (7). But this method is based on the assumption that the specific rate of ferrous iron oxidation does not vary significantly during the bioleaching process. In fact, ferrous iron oxidation is both a biological and chemical process which varies significantly between exponential and stationary phases, and according to growth conditions (4, 20).

The measurements of coverage ratio for T. ferrooxidans attached to pyrite ranged from 0.25 to 2.25% for inoculum concentrations promoting optimal bioleaching (10⁷ bacteria ml⁻¹). In other experiments not reported in this paper (4), attachment measurements indicated that for inocula ranging from 10⁷ to 10⁸ bacteria ml⁻¹, the surfaces of mineral particles were not saturated by bacteria. Under saturation conditions and for similar-size mineral particles, the coverage ratio was estimated to range from 20 to 45% (23, 24, 29). From these observations, it can be proposed that T. ferrooxidans could adhere only to specific and limited sites on the pyrite surface. It can also be suggested that bacteria need a well-defined area with proper electrochemical properties to attach to and oxidize a sulfide mineral surface. Finally, it may also be possible that only a reduced proportion of the bacterial population is able to attach at the surface due to the presence of specific cell surface components which allow efficient bacterial adhesion.

The ELIFA method presented here was shown to be compatible with the measurement of different levels of adhesion. It would be possible to use ELIFA to determine and study adhesion isotherms of T. ferrooxidans on various sulfide minerals exhibiting different physicochemical properties to better understand the mechanisms of bacterial interaction with mineral surfaces. In the present experiments performed with two pyrites exhibiting different surface properties, equivalent attachment levels were observed despite different bio-oxidation rates for both sulfides. Measurements of bacterial attachment and of its dynamics, complemented by the determination of bacterial activity and mineral solubilization, may contribute to a better understanding of the role of direct bacterial contact with mineral particles in biological oxidation processes.