Wastewaters released by oil-processing and petrochemical enterprises contain large amounts of toxic derivatives, such as polycyclic and aromatic hydrocarbons, phenols, sulfides, and heavy metals. Vanadium and nickel are constituents of crude oil in the form of inorganic and metallo-organic compounds. Both metals not only are present in residual waters from the oil industry but also can be present as particulate matter in inhaled air (23) and as environmental heavy metals emitted from automotive materials (11). Both metals are also present as aerosols or in the ashes of thermal power stations operating on oil fuel (43).

The recovery of heavy metals from industrial residues is an important task for environmental and economic reasons (2). Nickel and vanadium are toxic even at low concentrations (17, 19, 26, 31) and therefore are an important source of contamination in industrial societies (30); however, heavy metals also are a valuable resource for different industrial applications. In the present work, we selected heavy-metal-resistant microorganisms from contaminated soils at an oil refinery. Among the microorganisms selected, three of them were capable of accumulating nickel and vanadium. All three microorganisms belonged to the family Enterobacteriaceae. The isolation of different nickel-accumulating microorganisms has been reported before (5, 8, 24, 44); however, to our knowledge, this is the first report of vanadium accumulation by a bacterial biomass.

Isolation of and growing conditions for microorganisms resistant to nickel or vanadium. Twenty-six samples of contaminated soils were obtained at an oil refinery. The samples were pooled, grown in brain heart infusion (BHI) medium (1) at 37°C for 4 h, and seeded in petri dishes containing BHI agar and a 10 mM concentration of either nickel chloride or vanadyl sulfate. Under these selective conditions, the control laboratory strains Escherichia coli MC4100 and Pseudomonas aeruginosa PAO1 were unable to grow. Triplicate plates were seeded in all cases and incubated at 30, 37, and 55°C for a maximum of 3 days. Since all selected organisms grew well at 37°C, further culturing was always done at 37°C. The amount of nickel chloride or vanadyl sulfate is always referred as the metal concentration present in the solutions.

Screening for microorganisms capable of accumulating metals. Metal-resistant microorganisms were seeded in sets of duplicate BHI agar plates containing a 10 mM concentration of either nickel chloride or vanadyl sulfate and incubated at 37°C. One set of plates was exposed to sulfhydric acid (resulting from the reaction of sodium sulfide with chlorhydric acid) in a sealed container. Upon formation of the metal sulfide, both sets of plates were carefully screened to detect any change in the region surrounding or inside bacterial colonies (34).

Classification of microorganisms and determination of MICs. The organisms were taxonomically identified with the commercial system PASCO (36). The taxonomic classification of the isolates was confirmed with the API 20NE system (20). The MICs of antibiotics were determined with Mueller-Hinton (MH) agar (1) by the E test (3). The MICs of nickel and vanadium were determined with petri dishes containing MH agar and increasing amounts (0 to 3,000 μg/ml) of either nickel chloride or vanadyl sulfate. The clinical strains Enterobacter cloacae RYC70770 and Escherichia hermannii RYC78330 were provided by the Hospital Ramón y Cajal, Madrid, Spain.

Determination of metal accumulation by bacterial isolates. Microorganisms were cultured overnight on Luria-Bertani agar plates without metals. Confluent bacterial lawns were collected and suspended in phosphate-buffered saline, washed once in the same buffer, divided into 1-ml samples, and pelleted by centrifugation at 12,000 rpm in a bench-top microcentrifuge (Biofuge 13; Heraeus). One of the samples was dried in a Speed Vac centrifuge to calculate the dry weight of the bacterial pellet, and the remaining samples were resuspended in sterile polypropylene tubes containing 100 μg of either nickel chloride (1 ml) or vanadyl sulfate (10 ml) per ml dissolved in water. The samples were incubated at room temperature in a roller mixer for 3 h, and the cells were harvested again by centrifugation under the same conditions. The amount of residual metal present in the supernatant was measured by atomic absorption with a Perkin-Elmer 3030 atomic absorption spectrophotometer. Values were the averages of three determinations carried out in parallel.

Outer membrane isolation and SDS-PAGE. Outer membranes were obtained as described by Fukuoka et al. (14) but with 25 mM Tris-Cl (pH 7.2) instead of HEPES as the washing buffer. Outer and inner membranes were separated by incubation for 30 min on ice with 1.5% Triton X-100 instead of Sarkosyl NL-97. The amount of protein in the extracts was determined with a bicinchoninic acid protein assay kit (Pierce) in accordance with the manufacturer’s instructions. Ten micrograms of the outer membrane extracts was suspended in 5 μl of water, the suspension was mixed with the same volume of 2× sodium dodecyl sulfate (SDS) gel loading buffer (40), and the mixture was incubated at 100°C for 10 min in a dry bath. The outer membrane proteins (OMPs) were analyzed by SDS–12% polyacrylamide gel electrophoresis (PAGE) as previously described (14). The gels were stained with GELCODE blue stain reagent (Pierce) in accordance with the manufacturer’s instructions.

Isolation of microorganisms capable of accumulating heavy metals. Samples of contaminated soils were screened for the presence of microorganisms resistant to nickel and/or vanadium. Fifty-one microorganisms able to grow in the presence of nickel chloride or vanadyl sulfate concentrations higher than 10 mM were isolated. A preliminary screening of metal accumulation (34) was carried out as described in Materials and Methods. Isolates forming either a clear halo around the colony (possible sequestration of the metal) or a dark color around or inside the colony (possible reduction and precipitation of the metal) were considered to be possible metal biosorbents (34).

Three isolates producing dark colonies when grown in the presence of vanadium (Fig. 1) were selected for further characterization. The API 20NE biotype of two isolates was 1144113; these isolates were classified as E. hermannii CNB50 and E. hermannii CNB52. The biotype of the third isolate was 3305573; this isolate was classified as E. cloacae CNB60. Identification with the PASCO system confirmed the taxonomic classifications of the isolates. Two clinical strains were cultured as controls. The clinical strain E. hermannii RYC78330 resisted up to 0.5 mM nickel and 1 mM vanadium, and the clinical strain E. cloacae RYC70770 resisted up to 1 mM each metal; neither clear halos nor dark colors were observed around the colonies when the clinical strains were grown in the presence of metals.

FIG. 1 Metal accumulation by bacterial colonies. A dark coloration of the bacterial colony was considered a preliminary marker of metal accumulation. C, bacterial cells grown without metal; V, bacterial cells grown in the presence of 100 μg of vanadium (more ...)

The MICs of different antibiotics and the two metals used for the selection of the bacterial isolates are shown in Table 1. The different values observed for E. hermannii CNB50 and CNB52 indicate that these isolates are different strains.

Strain	Metal	MIC, μg/ml, ofa:
		V	Ni	Ery	Cip	Nor	Chl	Tet	Am/C	Pip	Cef	Imi	Ami	Gen
E. hermannii CNB50	None	750	1,000	8	0.008	0.047	4	0.5	0.75	2	0.094	0.094	1.5	2
	NI			12(1.5)	0.012(1.5)	0.047(1)	6(1.5)	0.5(1)	0.75(1)	2(1)	0.094(1)	0.125(1.3)	1.5(1)	2(1)
	V			128(16)	0.125(15.6)	0.75(16)	6(1.5)	3(6)	0.75(1)	2(1)	0.094(1)	0.125(1.3)	1.5(1)	3(1.5)
E. hermannii CNB52	None	1,000	2,000	8	0.032	0.047	6	2	8	2	0.19	0.047	1.5	2
	Ni			12(1.5)	0.032(1)	0.047(1)	6(1)	2(1)	8(1)	2(1)	0.19(1)	0.047(1)	1.5(1)	2(1)
	V			48(6)	0.094(3)	0.75(16)	8(1.3)	16(8)	8(1)	3(1.5)	0.19(1)	0.047(1)	1.5(1)	2(1)
E. cloacae CNB60	None	1,000	2,000	48	0.012	4	2	8	1.5	0.19	0.125	1	2
	Ni			48(1)	0.012(1)	0.047(1)	4(1)	2(1)	8(1)	1.5(1)	0.19(1)	0.125(1)	1(1)	2(1)
	V			64(1.3)	0.16(13.3)	1(21.2)	4(1)	8(4)	2(0.25)	1(0.7)	0.19(1)	0.125(1)	1(1)	4(2)

^aMICs were determined as described in Materials and Methods. V, vanadium; Ni, Nickel, Ery, erythromycin; Cip, ciprofloxacin; Nor, norfloxacin; Chl, chloramphenicol; Tet, tetracycline; Am/C, amoxicillin-clavulanate; Pip, piperacillin; Cef, ceftazidime; Imi, imipenem; Ami, amikacin; Gen, gentamicin. Numbers in parentheses indicate the ratio of the MIC in the presence of the metal to the MIC in the absence of the metal.

OMPs of heavy-metal-accumulating bacteria. Analysis of OMP profiles has been widely used for examining strain variation within bacterial species (6, 9, 29). To analyze the similarity of the three environmental isolates to clinical isolates of the same species, OMPs from cells grown on Luria-Bertani liquid broth in the absence of metals were extracted and analyzed by SDS-PAGE. E. cloacae CNB60 lacked the 37-kDa porin OmpF (25), which was present in E. cloacae RYC70770 (Fig. 2, lanes 1 and 2). No clear differences were observed in the OMP profiles of E. hermannii isolates (Fig. 3, lanes 1 to 3). OMPs from this bacterial species have not been analyzed before; five major OMPs with molecular masses of 23 kDa (Omp23), 36 kDa (Omp36), 40 kDa (Omp40), 43 kDa (Omp43), and 48 kDa (Omp48) were detected.

FIG. 2 OMP profiles for environmental and clinical isolates of E. cloacae. Lanes 1 and 2, E. cloacae RYC70770 and E. cloacae CNB60 grown in the absence of metals, respectively; lane 3, E. cloacae CNB60 grown in the presence of nickel; lane 4, E. cloacae CNB60 (more ...)

FIG. 3 OMP profiles for environmental and clinical isolates of E. hermannii. Lanes 1 to 3, E. hermannii RYC78330, E. hermannii CNB52, and E. hermannii CNB50 grown in the absence of metals, respectively; lane 4, E. hermannii CNB50 grown in the presence of nickel; (more ...)

No major differences in OMP profiles were observed when E. cloacae CNB60 was grown in the presence of metals (Fig. 2, lanes 3 and 4), although some minor differences were observed for small proteins (28 to 30 kDa) when the bacteria were grown in the presence of vanadium (Fig. 2, lane 4). The presence of nickel in the culture medium seemed to have no effect on the OMP profiles of E. hermannii CNB50 (Fig. 3, lane 4). However, the presence of vanadium resulted in the appearance of a new, 45-kDa OMP (Omp45) and a severe reduction in the amounts of Omp48 and Omp43 (Fig. 3, lane 5), as was also the case for E. hermannii CNB52 (data not shown).

Effect of nickel and vanadium on the susceptibility to antibiotics of E. hermannii and E. cloacae. To evaluate the effect of nickel and vanadium on the antibiotic susceptibility of the environmental isolates, the MICs of antibiotics belonging to different structural families were determined with petri dishes containing MH agar in the absence of the metal or in the presence of nickel or vanadium (100 μg/ml). The results are shown in Table 1. Growth in the presence of nickel did not significantly affect the MICs of the different antibiotics. The presence of vanadium increased the MICs of quinolones (ciprofloxacin and norfloxacin) and tetracycline for the three isolates and of erythromycin for the E. hermannii isolates. The MIC of amoxicillin-clavulanate for E. cloacae decreased slightly when the cells were grown in the presence of vanadium.

Accumulation of nickel and vanadium. Metal accumulation by E. cloacae CNB60, E. hermannii CNB50, and E. hermannii CNB52 was tested. Cells from overnight cultures were suspended in water containing 100 μg of either nickel or vanadium per ml (a value usually found in residual waters from the oil industry) (3a). Metal accumulation was measured as described in Materials and Methods and expressed as nanomoles of accumulated metal per milligram of bacterial dry weight. E. hermannii CNB50, E. hermannii CNB52, and E. cloacae CNB60 accumulated 687.71 ± 29.55 nmol of vanadium and 130.57 ± 3.81 nmol of nickel per mg, 918.07 ± 58.21 nmol of vanadium and 175.12 ± 4.56 nmol of nickel per mg, and 671.70 ± 13.03 nmol of vanadium and 117.12 ± 2.87 nmol of nickel per mg, respectively.

Three bacterial strains, two E. hermannii strains and one E. cloacae strain, that are capable of accumulating nickel and vanadium have been analyzed. It has been shown that environmental gasoline-utilizing isolates of P. aeruginosa are indistinguishable from clinical isolates of the same species (12). Our results indicate that this may be the case for E. hermannii as well. Since this organism was first reported in 1982 (4), a few cases of infections by E. hermannii have been described, but the presence of this bacterial species in environmental habitats has not been reported. The isolation of E. hermannii from contaminated soils at an oil refinery suggests that this bacterium also has an environmental habitat. E. cloacae is known to be present in soil samples, to be able to colonize the rhizosphere of plants (35), and to survive strong environmental desiccation (33). Enterobacter species also have been found in heavy-metal-contaminated soils (37), one zinc-tolerant Enterobacter sp. isolate capable of accumulating this metal has been described (18), and the reduction of selenium oxyanions by E. cloacae has been described (21). However, to our knowledge, this is the first time that the accumulation of nickel and vanadium by E. cloacae has been reported.

Multidrug resistance (MDR) has been described for several bacterial species and is currently associated with the expression of cryptic efflux pump systems and with the reduced expression of OMPs involved in the transport of antibiotics inside bacterial cells. Similar types of pumps are involved in heavy-metal resistance (41, 42). Most MDR systems so far described for gram-negative bacteria are three-component pumps (28, 32, 39); the regulation of their expression still is not completely understood. Some environmental signals can induce an MDR phenotype; for example, the presence of salicylate induces the expression of the efflux pump system encoded by the marRAB operon and simultaneously represses the synthesis of the OmpF porin (7). Vanadium may have a similar role in the induction of the MDR phenotype, since the MICs of different antibiotics for the three isolates were higher in the presence of vanadium. Additionally, increased amounts of Omp45 and decreased amounts of Omp43 and Omp48 were observed in E. hermannii, and some changes in the amounts of minor bands (in the range of 28 to 30 kDa) were detected in E. cloacae (Fig. 2 and 3). The results obtained suggest the possible existence of functional efflux pump systems in these bacteria.

Metal accumulation or biotransformation is an alternative mechanism for metal detoxification in bacteria (15). The presence of heavy metals affects the ecology of sediments (27) and the biodegradation of organic chemicals by microorganisms (38). The removal of toxic components from industrial effluents is of great importance, not only because of the decontamination effect but also because this removal protects the activated sludge of sewage treatment plants from the action of toxic compounds and ensures the functioning of the plants (10). The bacterial isolates described here may be of use for removing nickel and/or vanadium from contaminated effluents.

We thank A. Suárez and E. Campanario for technical assistance. We also thank the Microbiology Service of the Hospital Ramón y Cajal (Madrid, Spain) for the taxonomic identification of the bacterial isolates.

This work was supported by REPSOL S.A. and by grant BIO94-0792 from the Spanish CICYT.