Protein Method for Investigating Mercuric Reductase Gene Expression in Aquatic Environments

Abstract

A colorimetric assay for NADPH-dependent, mercuric ion-specific oxidoreductase activity was developed to facilitate the investigation of mercuric reductase gene expression in polluted aquatic ecosystems. Protein molecules extracted directly from unseeded freshwater and samples seeded with Pseudomonas aeruginosa PU21(Rip64) were quantitatively assayed for mercuric reductase activity in microtiter plates by stoichiometric coupling of mercuric ion reduction to a colorimetric redox chain through NADPH oxidation. Residual NADPH was determined by titration with phenazine methosulfate-catalyzed reduction of methyl thiazolyl tetrazolium to produce visible formazan. Spectrophotometric determination of formazan concentration showed a positive correlation with the amount of NADPH remaining in the reaction mixture (r² = 0.99). Mercuric reductase activity in the protein extracts was inversely related to the amount of NADPH remaining and to the amount of formazan produced. A qualitative nitrocellulose membrane-based version of the method was also developed, where regions of mercuric reductase activity remained colorless against a stained-membrane background. The assay detected induced mercuric reductase activity from 10² CFU, and up to threefold signal intensity was detected in seeded freshwater samples amended with mercury compared to that in mercury-free samples. The efficiency of extraction of bacterial proteins from the freshwater samples was (97 ± 2)% over the range of population densities investigated (10² to 10⁸ CFU/ml). The method was validated by detection of enzyme activity in protein extracts of water samples from a polluted site harboring naturally occurring mercury-resistant bacteria. The new method is proposed as a supplement to the repertoire of molecular techniques available for assessing specific gene expression in heterogeneous microbial communities impacted by mercury pollution.

Mercuric reductase (reduced NADP:mercuric ion oxidoreductase) is a flavoprotein that catalyzes the reduction of Hg²⁺ to metallic Hg⁰ in the presence of NADPH. The high vapor pressure of elemental mercury results in the volatilization of mercury from aqueous media, making the function of mercuric reductase critical for microbially mediated mercury detoxification (10, 23). There is widespread interest in understanding the contribution of aquatic bacterial processes to the biogeochemical cycling of mercury and in the controversial goal of exploiting bacterial detoxification processes for bioremediation of aquatic environments contaminated with mercury (2, 8, 10, 17). Approaches to biotechnological detoxification of mercury in aquatic environments include inoculation of water with laboratory-grown bacteria producing mercuric reductase and the immobilization of purified mercuric reductase in bioreactors, through which contaminated water is processed (2). In both cases, it is important to employ rapid and effective methods for monitoring cells, genes, and enzyme activities in the biologically complex heterogeneous media of contaminated waters.

Nucleic acid hybridization probes and PCR techniques have been useful for investigating population structure in microbial communities targeted for bioremediation of contaminated environments (1, 3, 7, 8, 13, 15, 17, 24, 27, 28). However, fluctuations in chemical bioavailability and genetic expression render information based solely on nucleic acids insufficient for evaluating bioremediation kinetics in heterogeneous environments (5, 19, 20). Moreover, traditional sensitive techniques, such as radioactive substrate (e.g., ²⁰³Hg) tracking, are expensive and not easily adaptable for evaluating bioremediation metabolic activity in open environments. Supplementary information on protein production and enzyme activity is needed to fill the gap between genetic potential and metabolic activity in situ (18, 21, 22). Techniques are now available to extract proteins directly from environmental samples and to probe such protein extracts for specific catalytic reactions relevant to bioremediation (9, 12, 16, 18, 21, 22, 30). The protein approach circumvents difficulties pertaining to nucleotide sequence variations and genetic regulation uncertainties in studies based on hybridization probes with DNA or mRNA (8, 13, 17, 23, 24, 27). This study contributes to the development of enzyme-based bioremediation assessment methods by investigating inducible mercuric reductase activity in protein molecules extracted directly from aquatic samples.

MATERIALS AND METHODS

Environmental samples.

Uncontaminated freshwater samples for controlled mercury amendment studies were obtained from San Diego Creek (SDC), a 22-km freshwater system that drains an area of approximately 254 km² through the Upper Newport Bay Ecological Reserve into the Pacific Ocean. SDC has no known history of mercury contamination, and salient features of the aquatic ecosystem have been described previously (21, 22).

For field validation of the protein-based techniques, contaminated water samples from East Fork Poplar Creek, Tennessee (EFPC), were analyzed. The freshwater system of EFPC has a long history of contamination from an ammunition factory, and several ecological studies have been conducted on the ecosystem, including the occurrence of mercury-resistant bacteria (4, 25, 27).

The total heterotrophic bacterial population density in the environmental water samples was determined by serial dilution in phosphate-buffered saline (PBS) and spread plating on Standard Methods Agar (Becton Dickinson, Cockeysville, Md.). To test for the occurrence of mercury-resistant bacteria, Standard Methods Agar was supplemented with 25 μg of mercuric chloride ml⁻¹ (27). The pH of freshwater samples was determined by means of a microprocessor pH meter with automatic temperature compensation (Oakton model; PGC, Frederick, Md.).

Bacterial strains.

Mercury-resistant Pseudomonas aeruginosa PU21 (ilv leu Str^r Rif^r), carrying the 142.5-kb plasmid Rip64, was used as the source of mercuric reductase activity (23, 31). Strain PU21(Rip64) was routinely cultivated in medium containing 5 g of yeast extract, 10 g of tryptone, and 5 g of sodium chloride per liter of deionized water. When desired, the medium was supplemented with 25 μg of HgCl₂ per ml. To eliminate possible extraneous NADPH-oxidizing potential in the growth medium, inducible mercury-specific NADPH-dependent enzyme activity was investigated in pure cultures of strain PU21 washed three times in PBS containing 8.0 g of NaCl/liter, 0.2 g of KCl/liter, 1.15 g of Na₂HPO₄/liter, and 0.2 g of KH₂PO₄/liter. The pH of the PBS was 7.0.

To serve as a negative control, a mercury-sensitive P. aeruginosa strain, CW1, was isolated from SDC by filtering 1-liter water samples through 0.2-μm-pore-size polycarbonate membranes, which were then incubated on Pseudomonas isolation agar (Difco, Detroit, Mich.) at 42°C. Isolated colonies were purified by streaking on fresh agar plates, and species identification was done with the Biolog (Hayward, Calif.) microbial identification system. Greater than 95% match with a member of the Biolog database, consisting of over 800 strains, was required for positive identification on the basis of 95 biochemical reactions. The strain identification was further confirmed with the Analytical Profile Index rapid NFT system (Analytab, Plainview, N.Y.). For cultivation, protein extraction, and enzyme analysis, the mercury-sensitive strain, CW1, was treated similarly to the mercury-resistant strain, PU21.

Protein extraction procedure.

The method used to extract proteins from cells concentrated from aquatic environmental samples was based on that developed by Ogunseitan (22). The same method was used in this study in control experiments involving pure cultures of strains PU21 and CW1 and for freshwater samples. Briefly, cells were concentrated from the liquid phase by centrifugation at 12,000 × g for 10 min at 4°C. The cell pellets were resuspended in 1 ml of ice-cold solution containing 20 mM Tris-Cl, 1 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride (pH 7.4). The cell lysis process consisted of 3 min of pulsed sonication with a Sonic Dismembrator 550 (Fisher Scientific, Tustin, Calif.), equipped with a 3-mm microtip, and an ice bath. The sonicator was set to 20 kHz, with a power output of 195 W and a 40% duty cycle. To collect released protein molecules and to reduce interference from particulate NADPH oxidase, the lysed cell suspensions were centrifuged at 80,000 × g for 60 min. Because bacterial mercuric reductase is a cytoplasmic flavoprotein enzyme (11, 14, 26, 29, 32, 33), the aqueous protein extracts were collected for enzyme assay and the particulate fractions were discarded. The concentration of proteins was determined by the Bradford dye assay (6) (U.S. Biochemicals, Cleveland, Ohio). When required, the protein extracts were concentrated by centrifuge-driven microfiltration. The protein extracts were used immediately or stored in cryovials at −20°C.

Recovery of proteins from seeded freshwater samples.

To determine the efficiency of the protein extraction method used, the total yield of proteins from known population densities (10², 10⁴, 10⁶, and 10⁸ CFU ml⁻¹) of strain PU21 resuspended in freshwater samples was compared with the yield of proteins from an equivalent number of cells resuspended in PBS. The protein recovery efficiency experiment was conducted in triplicate for each population density, and protein extraction was initiated within 10 min of cell resuspension.

Colorimetric assay for mercuric reductase activity in environmental samples.

Figure 1 presents a schematic diagram for the quantitative colorimetric assay developed for mercuric reductase activity in protein molecules extracted from pure cultures or from seeded and unseeded environmental freshwater samples. The two-step method developed is adaptable to both a highly quantitative liquid-phase spectrophotometric assay and a visual nitrocellulose membrane-based assay.

FIG. 1.

Scheme for colorimetric detection of mercuric reductase activity in protein extracts from natural microbial communities. In the first reaction, mercuric reductase activity catalyzes the reduction of Hg²⁺ to volatile Hg⁰ with concurrent oxidation of NADPH to NADP. In the second reaction, residual NADPH is oxidized via PMS and MTT or NBT is reduced to produce water-insoluble formazan. The kinetics of the first reaction was measured by recording optical density at 340 nm. The end point of the second reaction was measured by the optical density at 590 nm. Mercuric reductase activity is directly proportional to the optical density at 340 nm and inversely proportional to the final optical density at 590 nm. In the solid-support version of the assay, regions of mercuric reductase activity are colorless against a blue-stained background because NADPH is locally depleted by enzyme activity. PMSH, reduced PMS.

The liquid-phase assay was performed in 96-well quartz microtiter plates. Generally, enzyme activity was assayed by adding 100 μg of protein extract to a 100-μl solution of 50 mM sodium phosphate buffer (pH 7.5) containing (final concentrations) 100 μM NADPH, 0.2 mM magnesium acetate, 0.5 mM EDTA, 0.1% (vol/vol) β-mercaptoethanol, and 200 μM HgCl₂ for 60 min at 37°C in the dark. During this period, mercuric reductase activity in the protein extracts is responsible for oxidation of NADPH to NADP (Fig. 1, first reaction). To investigate the kinetic parameters of mercuric reductase activity, NADPH oxidation was analyzed by spectrophotometric analysis at 340 nm every 15 s with a Spectramax 250 (Molecular Devices, Sunnyvale, Calif.).

After the first reaction, a 100-μl solution of 50 mM sodium phosphate buffer (pH 8.0) containing 10 mg of nitroblue tetrazolium (NBT) or methyl thiazolyl blue (MTT) to serve as a reducible color dye and 1.5 mg of phenazine methosulfate (PMS) to serve as the catalyst for transferring electrons from NADPH to NBT or MTT was added to each reaction, and incubation was continued for 2 min. Color development is rapid due to the formation of water-insoluble formazan. The concentration of formazan was determined by measuring absorbance at 590 nm. The absorbance at this wavelength was found to be directly proportional (r² = 0.991) to the amount of NADPH remaining in the reaction mixture (Fig. 1, second reaction, and 2). Hence, the absorbance at 590 nm after the first and second reactions is inversely related to the mercuric reductase activity in a given protein sample. Two negative controls were run routinely, where either mercury or protein was omitted from the assay reaction mixture.

Mercuric reductase assays on solid support were conducted on 0.2-μm-pore-size nitrocellulose membranes (Schleicher and Schuell, Keene, N.H.) in a vacuum-driven filtration manifold (Schleicher and Schuell). Total protein extracts from each experiment were filtered through the nitrocellulose membranes. Prior to the filtration, the membrane was soaked for 15 min in a 50-ml solution containing 50 mM sodium phosphate buffer (pH 7.5). Following sample filtration through each slot, the membranes were incubated in a 50-ml solution of 50 mM sodium phosphate buffer (pH 7.5) containing 100 μM NADPH, 0.5 mM EDTA, 0.2 mM magnesium acetate, 0.1% (vol/vol) β-mercaptoethanol, and 200 μM HgCl₂ for 60 min at 37°C in the dark. After 60 min, the nitrocellulose membrane was incubated for an additional 60 min in a 50-ml solution of 50 mM sodium phosphate buffer (pH 8.0) containing 10 mg of NBT or MTT to serve as a reducible color dye and 1.5 mg of PMS to serve as the catalyst for transferring electrons from NADPH to NBT or MTT, leading to the formation of water-insoluble formazan precipitate. The incubation was continued until optimum contrast was achieved, where a dark-blue background highlighted light (unstained) regions of enzyme activity. Membrane slots with high mercuric reductase enzyme activity have low concentrations of NADPH, leading to minimal production of formazan, and are therefore colorless (Fig. 1).

Freshwater microcosms experimentally contaminated with mercury.

In order to standardize the new assay method for mercuric reductase activity in freshwater samples, two triplicate sets of 100-ml SDC freshwater microcosms were inoculated with strain PU21 to a total population density of either 0, 10², 10⁶, or 10⁸ cells per ml. One set of microcosms was amended with 25 μg of mercuric chloride per ml, and the other set was not amended with mercury. The freshwater microcosms were incubated at 21 ± 1.5°C for 48 h, after which the contents was centrifuged and the pelleted cells were processed for protein extraction and mercuric reductase assays as described above. To determine whether other unknown metal contaminants in the freshwater interfered with the enzyme assay, kinetics experiments were conducted in the absence of EDTA in the reaction mixture (26). Pure cultures of strains PU21 and CW1 were assayed in parallel with the freshwater microcosm contents to serve as positive and negative controls, respectively.

Determination of mercuric ion concentration in natural environmental samples.

The concentration of mercuric ions in aquatic samples from EFPC and SDC used to standardize and verify the protein assay method was determined by cold vapor atomic absorption spectrophotometry (CVAAS) with the Bacharach mercury analyzer system (model MAS-50B). The method consists of adding 1 ml of 10% stannous chloride to 40 ml of environmental water samples in 250-ml-capacity biological oxygen demand bottles. The stannous chloride chemically reduces Hg²⁺ to volatile Hg⁰. The released Hg⁰ was flushed with air into the CVAAS system for specific spectrophotometric determination. Standard solutions containing various concentrations of HgCl₂ in distilled, deionized water were used to calibrate the CVAAS system. Five replicates were run for each water sample.

Validation of the method with EFPC samples.

Protein molecules were extracted from the 100-ml water samples by the same technique as previously described for the freshwater microcosms. Similarly, a mercuric reductase assay was performed exactly as described for pure culture and freshwater microcosm contents.

RESULTS

Background data for environmental samples.

The total heterotrophic bacterial population density was (5.6 ± 2.3) × 10⁵ CFU ml⁻¹ for SDC water and (3.2 ± 1.1) × 10⁴ CFU ml⁻¹ for EFPC water. No mercury-resistant bacteria were found in SDC water, whereas (2.5 ± 1.2) × 10³ CFU ml⁻¹ were found to be resistant to mercuric ion toxicity in EFPC water. The pH of SDC water was found to be 7.8, and the pH of EFPC water was 6.9. Inorganic mercury was not detected in SDC water (the detection limit was 0.01 μg per 40 ml). The concentration of mercuric ions in EFPC water was found to be 2.8 ± 0.9 μg liter⁻¹.

Specificity of mercuric reductase assay.

The kinetics of mercury reductase activity in proteins extracted from strain PU21 is shown in Fig. 3 and 4. The biphasic kinetics is characterized by an initial rapid-reaction velocity followed by a slower steady-state rate. The specificity of the assay technique is supported by data reported in Fig. 3. NADPH was not oxidized in the absence of mercury substrate from the reaction mixture. NADPH was also not oxidized in the absence of protein from the reaction mixture nor in the presence of proteins extracted from the mercury-sensitive negative control strain, CW1 (Fig. 3). Although NADPH-dependent mercuric ion reduction occurred in proteins extracted from both the mercury-grown strain PU21 and cells grown without mercury, the assay kinetics differentiated between specific induction and constitutive expression of mercuric reductase activity in strain PU21 (Fig. 3).

FIG. 3.

Specificity of the assay used for mercuric reductase activity shown by kinetics of the catalysis process. Open circles, no enzyme activity in protein from the negative control, the mercury-sensitive strain CW1; open diamonds, no enzyme activity when no protein was added to the assay reaction mixture; filled circles, no enzyme activity when mercury substrate was omitted from the assay reaction mixture containing protein from the positive control strain, PU21; open triangles, constitutive enzyme activity in protein from strain PU21 grown in the absence of mercury; open squares, induced enzyme activity in protein from strain PU21 grown in the presence of mercuric ions. The concentration of protein in each reaction mixture was 0.4 μg μl⁻¹. O.D., optical density.

FIG. 4.

Sensitivity of the assay for mercuric reductase activity shown by kinetic curves for different concentrations of proteins extracted from the positive control strain, PU21. Solid diamonds, negative control with no mercury substrate in the reaction mixture. The approximate concentrations (in micrograms per microliter) of protein in reaction mixtures are as follows: solid triangles, 0.015; solid squares, 0.031; solid circles, 0.063; open diamonds, 0.125; open triangles, 0.25; open squares, 0.375; open circles, 0.5. The total volume of the reaction mixture was 200 μl. O.D., optical density.

Sensitivity of mercuric reductase assay.

The sensitivity of the assay method is reflected by the detection of mercuric reductase activity in 2.9 μg of protein extracted from strain PU21 cells (Fig. 4). The level of mercury-dependent NADPH oxidation is also positively correlated with the specific protein concentration in the reaction mixture (Fig. 4). This observation formed the basis for quantitative assessment of mercuric reductase activity in a given protein sample because there is a linear correlation between the residual NADPH concentration and optical density at 590 nm after titration with MTT-PMS to produce formazan (residual [NADPH] = 0.032 + 0.002 × A₅₉₀ [r² = 0.992] [Fig. 2]). The quantitative assessment thus derives from the complete scheme presented in Fig. 1. For pure cultures of strain PU21, approximately 30 μM of NADPH was oxidized by 100 μg of protein from uninduced cultures whereas approximately 78 μM of NADPH was oxidized by an equivalent amount of protein from induced cultures (Table 1).

FIG. 2.

Standard curve showing the correlation between residual NADPH concentration and formazan production after addition of PMS and MTT. Formazan is a water-insoluble colored dye with maximum spectrophotometric absorbance at 590 nm. Error bars represent standard deviations.

TABLE 1.

Liquid-phase assay for mercuric reductase activity in pure cultures

Bacterial strain (108 CFU ml−1)	Amt of Hg2+ added to media (μg ml−1)	Total protein extracted			[NADPH] oxidizeda			Specific activity (μM NADPH mg−1 m−1)
		Amt (mg ml−1)	SD	CV (%)	Amt (μM)	SD	CV (%)
P. aeruginosa PU21 (Rip64)	0	1.081	0.134	1.5	30.60	2.39	3.1	5.1
P. aeruginosa PU21 (Rip64)	0	1.081	0.134	1.5	15.49	5.38	6.6b	2.6
P. aeruginosa PU21 (Rip64)	25	0.999	0.078	0.8	77.96	1.52	6.9	13.0
P. aeruginosa CW1	0	1.714	0.063	3.7	−0.31	3.68	3.4	0

^aOne hundred micrograms of each protein extract was used in the assay. Values refer to cumulative micromoles of NADPH oxidized after 1 h of incubation. The reported statistics are based on three experimental replicates. CV, coefficient of variance. Negative values are due to optical density measurement fluctations and are assumed to be zero. 

^bNegative control experiment performed for assay specificity. Mercury substrate was absent from the assay reaction mixture. 

Reproducibility of mercuric reductase assay in protein extracts from freshwater microcosms.

The background protein concentration extracted from unseeded SDC water samples is reported in Table 2. The efficiency of protein recovery from experimentally seeded SDC freshwater samples was determined to be (97 ± 2)%.

TABLE 2.

Liquid-phase assay for mercuric reductase activity in microcosm experiments

No. of P. aeruginosa PU21(Rip64) (CFU ml−1)a	Amt of Hg2+ added to freshwater (μg ml−1)	Total protein extracted			[NADPH] oxidizedb			Specific activity (μM NADPH mg−1 m−1)
		Amt (mg ml−1)	SD	CV (%)	Amt (μM)	SD	CV (%)
0 2	0	0.062	0.002	3.8	9.49	0.83	0.9	0.01
0 2	25	0.059	0.004	7.5	−17.48	9.25	7.9	0
102	0	0.077	0.005	6.3	8.40	1.67	1.8	1.4
102	25	0.182	0.002	1.3	20.87	3.17	4.0	3.5
106	0	0.671	0.017	2.5	29.59	6.15	8.7	4.9
106	25	0.888	0.122	13.8	60.32	3.83	9.7	10.1
108	0	0.992	0.050	5.0	28.33	5.88	8.2	4.7
108	25	0.971	0.057	4.7	87.05	1.86	14.4	14.5

^aInoculated into freshwater microcosms (100 ml total volume). Proteins were extracted after 48 h of incubation at 21 ± 1.5°C. 

^bOne hundred micrograms of each protein extract was used in the assay. Values reported are cumulative micromoles of NADPH oxidized after 1 h of incubation. The statistics refer to three experimental replicates. CV, coefficient of variance. Negative values are due to optical density measurement fluctuations and are assumed to be zero. 

The specificity and sensitivity of the mercuric reductase assay were reproducible in protein extracts from SDC freshwater samples seeded with the mercury-resistant strain, PU21, at 0, 10², 10⁶, or 10⁸ CFU per ml in the presence or absence of mercuric ion inducer (Fig. 5 and Table 2). Mercuric reductase activity was not detected in the background population of SDC freshwater samples used to standardize the protocol (Fig. 5 and Table 2). The levels of mercuric reductase activity detected in the microcosms correlated positively with the population density of strain PU21 used to inoculate the freshwater microcosms (Table 2). There was a two- to threefold increase in the level of mercury-dependent NADPH oxidation due to amendment of the freshwater microcosms with mercuric ions as an inducer (Table 2). The effect of mercury amendment was most noticeable in microcosms seeded with 10⁸ CFU of strain PU21 per ml, where the relative amount of NADPH oxidized increased from approximately 28 μmol in protein extracts from mercury-free microcosms to 87 μmol in freshwater microcosms amended with 25 μg of HgCl₂ per ml (Table 2).

FIG. 5.

Kinetic curves for mercuric reductase activity in proteins extracted from 48-h-old freshwater microcosms seeded with 0 (open squares), 10² (open circles), 10⁶ (open diamonds), and 10⁸ (solid circles) CFU ml⁻¹ of strain PU21. For comparison, enzyme activity from an equivalent amount of protein in a pure culture of strain PU21 is plotted (open triangles), as well as that from a culture of the negative control strain, CW1 (solid squares). O.D., optical density.

The standard kinetics of mercuric reductase activity in proteins extracted from seeded freshwater samples is presented in Fig. 5. The data show a minor reduction in the reaction rate when kinetic parameters of mercuric reductase activity in proteins extracted from freshwater seeded with 10⁸ CFU of strain PU21 per ml are compared with those of proteins extracted from pure culture (Fig. 5). However, a distinctly different kinetic pattern was observed when EDTA was omitted from the reaction mixture to investigate the possibility that unknown contaminants in the environmental sample interfere with the enzyme assay (Fig. 6). In proteins extracted from SDC water seeded with 10⁸ CFU of strain PU21, a two-stage catalytic kinetic reaction was observed in the absence of EDTA. Initially high rates of enzyme activity declined rapidly after 10 min. In SDC water amended with mercury, the linear enzyme kinetics quickly recovered after the decline, but no such quick recovery was observed in freshwater microcosms that were not amended with mercury (Fig. 6).

FIG. 6.

Kinetic characteristics of mercuric reductase activity in protein extracts from seeded SDC freshwater samples in the absence of EDTA in the reaction mixture. Open circles, positive control, consisting of protein extracts from a pure culture of strain PU21; open diamonds, protein extract from a freshwater sample seeded with 10⁸ CFU of strain PU21 per ml, with mercury amendment; open triangle, same as for open diamonds, but without mercury amendment; open squares, negative control with protein extracted from unseeded SDC water. All assays were carried out without EDTA. O.D., optical density.

The results of the nitrocellulose membrane version of the assay is shown in Fig. 7 for the freshwater microcosm study. The data show no mercuric reductase activity from the negative control strain, CW1, and no activity detected from unseeded freshwater samples. The membrane assay signals also show positive correlations with seeded bacterial population densities and with amendment with mercuric ion inducer (Fig. 7). No signals were observed when the membrane reaction mixture contained no mercury substrate.

FIG. 7.

Picture of stained nitrocellulose membrane filter showing mercuric reductase activity in SDC freshwater microcosms. Lanes A, B, and C depict replicate experiments. Row 1, proteins from 10⁸ cells of the negative control strain, CW1; row 2, proteins from unseeded freshwater without mercury; row 3, proteins from unseeded freshwater with mercury; row 4, proteins from 10² cells of strain PU21 seeded per ml of freshwater without mercury; row 5, same as row 4, but with mercury inducer; row 6, proteins from 10⁶ cells of strain PU21 seeded per ml of freshwater without mercury; row 7, same as row 6, but with mercury inducer; row 8, proteins from 10⁸ cells of strain PU21 seeded per ml of freshwater without mercury; row 9, same as row 8, but with mercury inducer.

Validation of assay in naturally contaminated aquatic samples.

The mercuric reductase activity of proteins extracted from mercury-contaminated EFPC water is reported in Table 3. The concentration of proteins extracted from polluted EFPC water was comparable to the protein yield from relatively pristine SDC water (Tables 2 and 3). Thirty-six micromoles of NADPH was oxidized due to mercuric reductase activity in protein extracts from EFPC water compared to none in the control experiment, where mercury was omitted from the reaction mixture (Table 3).

TABLE 3.

Liquid-phase assay for mercuric reductase activity in EFPC protein samples

[Hg2+] (μg liter−1) (± SD)	No. of Hgr bacteria (CFU ml−1) (± SD)	Total protein yield			[NADPH] oxidizeda			Specific activity (μM NADPH mg−1 m−1)
		Amt (mg ml−1)	SD	CV (%)	Amt (μM)	SD	CV (%)
2.8 ± 0.9	(2.5 ± 1.2) × 103	0.141	0.008	5.7	36.40	15.31	11.2	6.1
Noneb					−9.94	12.89	14.3	0

^aOne hundred microliters of each protein extract was used in the assay. Values refer to interpolated cumulative micromoles of NADPH oxidized after 1 h of incubation. The reported statistics are based on three experimental replicates. CV, coefficient of variation calculated based on the original mean optical density (at 590 nm) data. 

^bNegative control experiment performed for assay specificity. Mercury substrate was absent from the assay reaction mixture. Negative values are due to optical density measurement fluctuations and are assumed to be zero. 

DISCUSSION

The goal of this study was to integrate three critical biochemical processing techniques into a comprehensive system capable of providing direct information on the induction and activity of mercuric reductase in microbial communities inhabiting contaminated ecosystems. The first technique is the conservative recovery of microbial proteins directly from heterogeneous biological systems (22). The second technique exploits the NADPH-dependent kinetics of mercuric reductase activity (26), and the third technique is the titration of residual NADPH through the stoichiometric reduction of tetrazolium salts (16). Successful integration of these techniques has produced a quantitative liquid-phase assay and a visual solid-support assay for mercuric reductase activity in protein extracts from seeded and unseeded aquatic microbial communities containing mercury-resistant bacteria.

Biogeochemical variations across environmental samples could presumably limit the efficiency of protein recovery from natural microbial communities. This was not a significant problem in the freshwater samples used in this investigation because the calculated efficiency of protein recovery in seeded samples was high (97% ± 2%) compared to protein extraction from pure cultures. Other investigators have detected enzymes in protein fractions of agricultural soils with comparable protein recovery efficiencies (9). The sonication technique, when used at low temperatures, optimizes cell lysis and particle disintegration while preserving protein integrity for catalytic functions. The extraction technique has been used previously for freshwater samples and for highly turbid sewage samples (22). Some reduction in protein recovery efficiency can be expected in soils or sediments with high concentrations of compounds that can adsorb protein molecules or that reduce cell lysis due to interaction with soil particles. This interference phenomenon has been documented for studies on direct extraction of nucleic acids from natural environmental samples (7, 8, 13, 17). The present investigation focused on aqueous samples containing suspended particulate materials, but sonication parameters can be adjusted to improve protein recovery from complex environmental samples, such as soils. The yield of proteins from unseeded environmental samples in this study (Tables 2 and 3) was consistent with the bacterial population density in the samples and with the yield in previous studies (18, 21, 22). After 48 h of incubation of strain PU21 in freshwater microcosms, the recovery of total proteins remained comparable to that from an equivalent number of cells in pure culture (Table 2).

Given an optimized protein yield from environmental samples, it is possible to assay for multiple enzymes in the extract, provided that the inhibitors are absent or inactivated and that the enzyme is specific for substrate molecules supplied in the assay. Notable inhibitors of mercuric reductase activity reported in the literature are other metal ions, such as cadmium, silver, copper, and gold salts (26). The freshwater sample used to standardize the present method has no history of contamination with any of these metals. However, it is highly likely that polluted environmental samples contain heterogeneous chemical mixtures, some of which may interfere with enzyme assays. The presence of EDTA in the mercuric reductase assay reaction mixture reduces interference from other metals (26). Hg-EDTA is a terminal substrate for mercuric reductase, and the product concentration under this condition is given by the following equation: P = v_O [1 − exp(−kt)]/k, where v_O is the initial observed velocity and k is an apparent first-order rate constant (26). The kinetic parameters of mercuric reductase from strain PU21 observed in this study (Fig. 4) was similar to previously investigated enzymes with high specificity for mercuric ions (2, 11, 26). No other metal apart from mercury has been shown to be reduced by mercuric reductase. Therefore, the product in terms of oxidized NADPH is a good index for mercuric reductase activity even in the presence of other polluting metals possibly coextracted with the protein.

After 48 h of incubation in freshwater, the kinetic parameters of mercuric reductase activity from strain PU21 remained comparable to that of extracts from pure culture, showing no apparent inhibition due to SDC freshwater components that could overcome the EDTA protection (Fig. 5). In the absence of EDTA, however, there was a noticeable but reversible inhibition of enzyme activity in proteins extracted from seeded SDC water (Fig. 6). No such inhibition was observed in proteins from pure cultures when the assay was performed without EDTA, supporting the conclusion that the interference was due to naturally occurring components of the water samples (Fig. 6). The exact nature of this reaction warrants further investigation in terms of its potential contribution to mercury detoxification in situ.

To control for other cellular components in addition to mercuric reductase that may oxidize NADPH, assays were routinely conducted without mercury substrate, and an insignificant optical density at 340 nm was recorded under these circumstances (Fig. 3 and Table 1). The data show that mercuric reductase was responsible for the NADPH oxidation measured by the assay. The high specificity and sensitivity of the assay protocol for mercuric reductase in the protein extracts are supported by the evidence showing more intense signals in the induced state, where mercury was added to the microcosm contents (Tables 1 and 2 and Fig. 7).

Enzymatic redox reactions coupled to color product formation have been used extensively in many biochemical ecology studies, such as those employing the popular Biolog microbial identification database. In the present study, the amount of formazan produced (determined by absorbance at 590 nm) consistently correlated with the residual NADPH (Fig. 2). The correlation gave an excellent opportunity for quantitative assessment of enzyme activity and also for the color differential in the nitrocellulose membrane-based version of the assay. Several enzymes have been assayed on electrophoretic gels with a similar approach (16, 30).

The biphasic nature of the kinetic parameters of mercuric reductase activity (Fig. 1, first reaction) suggests that incubation periods longer than approximately 60 min could reduce the ability to distinguish between high and low enzyme activity because of the slow but increasing cumulative oxidation of NADPH at low protein concentrations over extended incubation periods (Fig. 4). However, the enzymatic reaction is effectively stopped by the addition of MTT-PMS solution, which rapidly oxidizes residual NADPH to produce formazan. It is also possible, if desired, to stop the enzymatic reaction after the initial rapid-reaction velocity, during the uniformly linear stage of the process (10 min in Fig. 4) by addition of 10% trichloroacetic acid to precipitate proteins prior to the color development process of the second reaction (Fig. 1). The possible interference of trichloroacetic acid with the kinetics of color development was not investigated in this study.

The measurement of mercuric reductase activity in proteins extracted from EFPC water (Table 3) validates the method proposed. Other investigators have demonstrated that insoluble mercuric sulfide and metallic mercury are the predominant forms of mercury found in the floodplains of EFPC (4). Certainly, the biogeochemical dynamics of mercury speciation will affect the bioavailability of mercury and, consequently, the level of genetic expression due to mercury exposure. It is not clear at present how much the biogeochemical dynamics contributed to the relatively large variance observed for mercuric reductase activity in EFPC protein (Table 3). Further studies aimed at linking stream geochemical characteristics and mercury bioavailability determined by enzyme production are currently ongoing in our laboratory. It should also be interesting to investigate the limits of enzyme kinetics as a bioindicator of exposure in terms of various concentrations of organic and inorganic forms of mercury in heterogeneous microbial communities.

In conclusion, this study provides a simple, rapid, and specific technique for screening aquatic samples directly for mercuric reductase enzyme activity. The technique is both sensitive and reproducible, and it can be used to monitor the activity of deliberately released organisms in bioreactors or in natural aquatic environments (1). The environmental sample size can be increased if necessary to increase sensitivity. The technique will be particularly useful for circumventing molecular divergence problems typically faced by investigations based on nucleic acid probes, because the assay accounts directly for the mercury detoxification phenotype in microbial communities without depending on bacterial isolation.