Isolation and Characterization of Metal-Reducing Thermoanaerobacter Strains from Deep Subsurface Environments of the Piceance Basin, Colorado

Isolation of pure bacterial strains.Thermophilic Fe(III)-reducing enrichment cultures were established by inoculation of core sediment samples obtained from the various depths of a drilling in the Piceance Basin in Colorado and drilling fluids collected after they circulated through the various depths of the geological formations. The low-porosity deeply buried sandstone environment had likely been geologically and hydrologically isolated for several million years (8). The original enrichment cultures were incubated under a H₂-CO₂ (80:20) headspace (2 atm of pressure) with 70 mM Fe(III) oxyhydroxide in a basal medium supplemented with 0.05% yeast extract. The basal medium contained the following ingredients: 2.5 g of NaHCO₃/liter, 0.08 g of CaCl₂ · 2H₂O/liter, 1.0 g of NH₄Cl/liter, 0.2 g of MgCl₂ · 6H₂O/liter, 10.0 g of NaCl/liter, 7.2 g of HEPES/liter, 1.0 g of resazurin (0.01%)/liter, 0.5 g of yeast extract/liter, trace minerals (10 ml), and vitamin solutions (1 ml) (43). No reducing agents were added to the medium. The trace mineral solution contained (per liter) 1,500 g of nitrilotriacetic acid, 200 g of FeCl₂ · 4H₂O, 100 g of MgCl₂ · 6H₂O, 20 g of sodium tungstate, 100 g of MnCl₂ · 4H₂O, 100 g of CoCl₂ · 6H₂O, 1,000 g of CaCl₂ · 2H₂O, 50 g of ZnCl₂, 2 g of CuCl₂ · 2H₂O, 5 g of H₃BO₃, 10 g of sodium molybdate, 1,000 g of NaCl, 17 g of Na₂SeO₃, and 24 g of NiCl₂ · 6H₂O. The vitamin solution contained (per liter) 0.02 g of biotin, 0.02 g of folic acid, 0.1 g of vitamin B₆ (pyridoxine) HCl, 0.05 g of vitamin B₁ (thiamine) HCl, 0.05 g of vitamin B₂ (riboflavin), 0.05 g of nicotinic acid (niacin), 0.05 g of pantothenic acid, 0.001 g of crystalline vitamin B₁₂ (cyanocobalamin), 0.05 g of p-aminobenzoic acid, and 0.05 g of lipoic acid (thioctic). Subcultures were established in a phosphate- and carbonate-buffered basal medium amended with 0.005 to 0.05% yeast extract (21). The subcultures contained 70 mM Fe(III) oxyhydroxide as an electron acceptor and different electron donors, with either pyruvate (10 mM), acetate (10 mM), or hydrogen (80% in a headspace with 2 atm of pressure) present in the media. SSU rRNA gene-based molecular analyses indicated that the enrichment with hydrogen as the electron donor was more diverse than those with pyruvate and acetate as electron donors (21, 61). Thus, the enrichment culture with hydrogen as an electron donor was used for bacterial isolation.

To isolate pure cultures, a series of 10-fold dilutions of the hydrogen-enriched Fe(III)-reducing cultures were made with cell numbers ranging from 10¹ to 10⁸ by vortexing them in an anaerobic chamber. Bacterial numbers were determined by directly counting cells, which were stained with 0.1% acridine orange, under an epifluorescence microscope (57). A 0.1-ml sample of each dilution was inoculated into 28-ml pressure tubes containing 10 ml of basal medium and 0.05% yeast extract, with ferric citrate (10 mM) as the electron acceptor and hydrogen as the primary electron donor under a H₂-CO₂ gas headspace. The tubes were incubated at 65°C for 12 to 16 h, and Fe(III) reduction was assessed visually according to color changes. The highest dilution showing iron reduction was serially diluted again and used for the second round of purification as described above. This process was repeated at least 20 times prior to initiating the isolation of pure cultures.

Pure cultures of microorganisms were obtained by plating the enrichment cultures on 1.8% agar medium prepared with the basal medium buffer plus HEPES and phosphate. Ferric citrate (10 mM) was used as an electron acceptor, and lactate was used as an electron donor (10 mM). The plates were incubated under a N₂ gas headspace in an anaerobic transparent plastic jar at 65°C for 48 h. A preliminary assessment of the diversity of the iron-reducing bacteria was based on the differences in colony morphologies. Approximately 20 isolates were selected for further screening. Individual colonies were transferred 10 times as described above. Enterobacterial repetitive intergenic consensus (ERIC)-, repetitive extragenic palindromic (REP)-, and BOX-PCR were used to differentiate identical or closely related strains, and these assays were accomplished as described elsewhere (9, 44, 52). The isolates with different genetic fingerprinting patterns were tested again for iron reduction to ensure that they were capable of reducing ferric iron.

Physiological characterization.A variety of factors were examined for their influence on metal reduction and mineral formation, including medium pH, incubation temperature, electron donors, electron acceptors, different metal species, and headspace gas composition. To determine the range of growth substrates, the purified isolates (X513, X514, and X561) were inoculated into anaerobic basal medium with lactate (10 mM), acetate (10 mM), pyruvate (10 mM), succinate (10 mM), xylose (1%), or H₂ (100% H₂ or H₂-CO₂ at an 80:20 ratio) as an electron donor and a Fe(III) oxyhydroxide (∼70 mM) as an electron acceptor and incubated at 60°C in the dark for several days. The Fe(III) oxyhydroxide was prepared by neutralizing a solution of 0.4 M FeCl₃ with 10 M NaOH as described previously (46). Analysis of X-ray diffraction (XRD) showed that most of the Fe(III) oxyhydroxide was poorly crystalline akaganeite (FeOOH) and a small portion was amorphous. In this paper, we refer to this material as Fe(III) oxyhydroxide.

To assess the capabilities for metal reduction and mineral formation of the pure cultures (X513, X514, and X561), various electron acceptors, such as ferric citrate (10 to 20 mM), Fe(III) oxyhydroxides (∼70 mM), cobalt-EDTA [Co(III)-EDTA, 1.5 mM], potassium chromate (0.5 mM), Mn(IV) oxide (10 mM), gold chloride (AuHCl₃, 0.1 mM), and uranyl carbonate (1 mM), were examined at 60°C with lactate (10 mM) as an electron donor. The Mn(IV) oxide was prepared as described by Lovley and Phillips (29), and uranyl carbonate [UO₂(CO₃)₂²⁻] was prepared by mixing uranyl nitrate [UO₂(NO₃)₂²⁻] solution with sodium bicarbonate (45). The effects of temperatures (25 to 75°C), pH (6.5 to 9.6), and salinity (0.05 to 7% NaCl) on microbial Fe(III) reduction and mineral formation were also examined using an Fe(III) oxyhydroxide (70 mM) as an electron acceptor and lactate (10 mM) as an electron donor. In addition, the optimal growth temperature for the isolate X514 was determined under anaerobic iron-reducing conditions with Fe(III) citrate (15 mM) as an electron acceptor and lactate (10 mM) as an electron donor.

To examine the ability to produce ethanol concomitantly with Fe(III) reduction, the three isolates (X513, X514, and X561) were incubated at 60°C with xylose (10 mM) and Fe(III) oxyhydroxide (70 mM). The isolates were also used to examine their fermentative capabilities on glucose (10 mM) while simultaneously reducing Fe(III) in non-energy-generating side reactions. In addition, the ethanol-producing thermophilic anaerobe Thermoanaerobacter ethanolicus (ATCC 31938) was obtained from the American Type Culture Collection (Manassas, Va.) and grown anaerobically with Fe(III) oxyhydroxide as an electron acceptor.

To understand the effect of headspace atmosphere on iron mineralogy, the anaerobic media were prepared and incubated under three different headspace atmospheres, N₂, N₂-CO₂ (80:20), and H₂-CO₂ (80:20), with Fe(III) oxyhydroxide (∼70 mM) as an electron acceptor. In the case of the N₂ and N₂-CO₂ headspaces, lactate (10 mM) was used as an electron donor, while H₂ was used as an electron donor when a H₂-CO₂ headspace atmosphere was used. Incubation time ranged from 2 to 3 weeks. Abiotic controls were established for each treatment. To determine whether iron reduction and mineral formation with three strains (X513, X514, and X561) were enzyme-linked biological processes, we used various inhibitors (final concentrations are indicated within parentheses), including the ferric reductase inhibitor p-chloromercuriphenyl sulfonate (pCMPS) (0.1 mM), the hydrogenase inhibitor quinacrine dihydrochloride (1 mM), the protonophore carbonyl cyanide m-chlorophenylhydrazone (CCCP) (0.02 mM), and the electron transport inhibitors 2-n-heptyl-4-hydroxyquinoline-N-oxide (HOQNO) (0.05 mM) and dicumarol (0.05 mM) (1, 15, 21, 35, 36, 37)

Phylogenetic analyses of bacterial isolates.Genomic DNA was recovered from the isolates as described previously (62). SSU rRNA genes were amplified from genomic DNA by using conserved primers 27f and 1592r (56, 62). Amplified products were purified with the WizardPrep kit (Promega, Madison, Wis.). Nucleic acid sequences were determined by the automated Taq-cycle sequencing method with an ABI 373A sequencer (Applied Biosystems, Foster, Calif.) by using 12 primers as described previously (62).

Analyses of SSU rRNA sequences were initially conducted by searching current databases with the FASTA program from the Genetics Computer Group software package (10). The genetic data environment was used to manually align the sequences from the isolates with representative sequences from the Ribosomal Database Project (RDP) (33). The phylogenetic relationships of these isolates with previously described bacteria were established with the neighbor-joining DNA distance program in the ARB package (32) and the maximum-likelihood method with the program fastDNAml from the RDP. The maximum-likelihood tree was constructed with a transition/transversion ratio of 2.0 by using jumbled orders of 10 for the addition of taxa.

Chemical and mineralogical characterization.To examine the chemical conditions of metal reduction and mineral formation by the thermophilic Fe(III)-reducing bacteria, subsamples (1 ml) of bacterial cultures and abiotic controls were taken from the culture bottles at different times and measured for redox potential (E_h) and pH at room temperature in an anaerobic chamber (58). Fe(II) concentrations (0.5 N HCl, soluble) were determined by measuring the absorbance at 562 nm on a Hewlett-Packard model 8453 spectrophotometer by the ferrozine method with anaerobic water for sample dilution (50, 57).

A JSM-35CF (JEOL, Ltd., Tokyo, Japan) scanning electron microscope (SEM) with an energy-dispersive X-ray (EDX) detector was used for the analysis of the morphology, mineralogy, and chemistry of the mineral phases precipitated or transformed by the thermophilic Fe(III)-reducing bacteria (46). The mineralogical compositions of the precipitated or transformed phases were determined by using XRD. All XRDs were performed on an XDS 2000 diffractometer (40 kV, 35 mV; Scintag, Inc., Sunnyvale, Calif.) equipped with Co-Kα radiation. To understand the morphology of the transformed crystalline iron minerals, mineralogical characteristics of the precipitates were also examined by transmission electron microscopy (TEM). Culture media containing bacterial cells, organic matter, and inorganic solids [magnetite and Fe(III) oxyhydroxides] were fixed with 2.5% glutaraldehyde in 0.1 M cacodylate. After being washed with a HEPES buffer and an alcohol-water solution (1:1 ratio), samples were dehydrated with propylene oxide and embedded in a low-viscosity, thermally curing epoxy resin. Ultrathin sections (70 to 80 nm) were cut from resin blocks with a diamond knife and transferred to a 30-mesh Formvar-coated Cu TEM grid for image analysis on an FZ 2000 TEM (JEOL, Ltd.) equipped with an EDX detector (59).

Nucleotide sequence accession numbers.The sequences obtained by this study have been submitted to GenBank under accession numbers AF542517 (X514), AF542518 (X561), AF542519 (X521), AF542520 (X513), and AF542521 (X524).

Molecular analyses of bacterial isolates.After a series of dilution and plating, five different pure cultures were obtained and designated X513, X514, X521, X524, and X561. These strains were closely related, as indicated by the fingerprinting patterns revealed by BOX-, ERIC-, and REP-PCR. Strains X513, X514, X524, and X561 showed the same fingerprinting patterns by BOX- and REP-PCR, but they showed different fingerprinting patterns by ERIC-PCR (data not shown). This suggests that the ERIC-PCR method may have a higher level of resolution in differentiating these anaerobic thermophilic Fe(III)-reducing bacteria.

Preliminary sequence analysis based on the FASTA search agreed with results of microscopic examinations indicating that these isolates were gram-positive bacteria. To further determine their phylogenetic relationships, their SSU rRNA gene sequences were compared with those of gram-positive bacteria listed in the RDP. Maximum-likelihood analysis indicated that these isolates were closely related to Thermoanaerobacter ethanolicus 39E, with 99.4 to 99.5% similarity (Fig. 1), and to the clones P-5, H-92, and H-112 from pyruvate- and hydrogen-enriched cultures (61). These isolates were 81.4 to 83.2% similar to the thermophilic metal-reducing bacterium Bacillus infernus and 81.0 to 83.5% similar to Thermoterrabacterium ferrireducens. A similar topology for the SSU rRNA phylogenetic tree was obtained with the neighbor-joining distance method (data not shown).

• Open in new tab
• Download powerpoint

FIG. 1.

Phylogenetic relationships of the thermophilic iron-reducing bacteria from the deep subsurface environments. The tree was constructed by the maximum-likelihood method. Escherichia coli was used as the out-group.

Physiological characteristics.Since these isolates were very closely related, only three isolates (X513, X514, and X561) were selected for further physiological and biogeochemical characterization. All three isolates examined were able to use lactate, acetate, succinate, xylose, and glucose as electron donors in reducing Fe(III) oxyhydroxide and producing magnetite under a N₂ atmosphere. Although these isolates are closely related phylogenetically, they showed some physiological differences. For example, X514 was able to use both pyruvate and hydrogen in reducing Fe(III) oxyhydroxide and producing magnetite, whereas X561 was not. X513 was able to utilize pyruvate but not hydrogen for iron reduction and magnetite production. The non-hydrogen utilizers, including X513 and X561, may use yeast extract (0.05%) as an energy source. No growth, Fe(III) reduction, or magnetite formation was observed in controls without microbial inocula. These strains were capable of producing ethanol from glucose (<0.1%), and their growth was unaffected in media containing 1% ethanol. However, a much slower growth was observed in incubated cultures containing 2% ethanol, and no growth was evident in incubated cultures containing 3% or more ethanol.

The effects of incubation temperature on microbial growth rates and Fe(III) reduction rates were examined with the isolate X514 (Fig. 2). The maximum Fe(III) reduction rates and bacterial growth were observed at 60°C (Fig. 2A), and generation times of 6 to 10 h were typically observed (data not shown). The initial concentration of Fe(III) citrate (15 mM) was reduced biotically by 36% after 24 h of incubation and by 50% after 48 h of incubation at 60°C. The cell density increased from 5 × 10⁴ cells/ml at time zero to 7.5 × 10⁵ cells/ml at 24 h after incubation at 60°C (Fig. 2A). Although the abiotic reduction of Fe(III) increased as the temperature increased, the Fe(III) reduction occurring without cells was significantly less than that occurring with cells (Fig. 2B). The abiotic reduction of Fe(III) at 55 and 75°C after 48 h of incubation accounted for 1.5 and 38% of the total reduced Fe(II), respectively. No significant increase in abiotic Fe(III) reduction [<3% of the initial Fe(III) concentration] was detected at 45°C after 24 or 48 h of incubation. These results indicate that the optimal temperature for the reduction of Fe(III) citrate and for bacterial growth was about 60°C, and the minimum bacterial population doubling time of the cultures was approximately 6 h.

• Open in new tab
• Download powerpoint

FIG. 2.

Temperature profiles of growth (a) and iron reduction (b) rates by the thermophilic iron-reducing bacteria (X514) at 24 and 48 h. The results shown are the means of results from triplicate cultures.

Three isolates (X513, X514, and X561) fermented glucose (10 mM) and reduced Fe(III) oxyhydroxide (70 mM) and produced magnetite simultaneously. An ethanol-producing anaerobe, Thermoanaerobacter ethanolicus (ATCC 31938), also fermented glucose while simultaneously reducing Fe(III) citrate (15 mM). However, Thermoanaerobacter ethanolicus (ATCC 31938) did not grow with Fe(III) citrate as an electron acceptor and lactate (10 mM) as an electron donor. Three isolates (X513, X514, and X561) fermented xylose and produced ethanol (<0.1%) concomitantly with the reduction of Fe(III) oxyhydroxide and magnetite formation. The ability of three strains (X513, X514, and X561) to reduce metals besides iron was also examined. These isolates were able to reduce Co(III) [Co(III)-EDTA, 1 mM] to Co(II) and Cr(VI) (potassium chromate, 0.5 mM) to Cr(III) with lactate (10 mM) as an electron donor, as indicated by the color change of the culture solution from purple [Co(III)-EDTA] and yellow (potassium chromate) to colorless at 60°C. Three strains were also able to reduce U(VI) (UO₂CO₃²⁻, 1 mM) and produced black precipitate with lactate as an electron donor. XRD analysis showed that the precipitated phase was uraninite (UO₂) (Fig. 3C), thereby demonstrating the reduction of U(VI) to U(IV). SEM and EDX analyses showed that the crystalline phase contained significant quantities of U (Fig. 3B) and was ball shaped (Fig. 3A). Three strains were also able to reduce Mn(IV) oxide (10 mM), producing a reddish-brown precipitate with lactate (10 mM) as an electron donor. XRD analysis showed that the precipitated phase was rhodochrosite (MnCO₃) (Fig. 3F); therefore, Mn(IV) was reduced to Mn(II). SEM with EDX analysis showed that this crystalline phase contained significant quantities of Mn (Fig. 3E) and was shaped like a rhombohedron with curved faces (Fig. 3D). These results indicated that these thermophilic Thermoanaerobacter ethanolicus strains can reduce several other transition metals besides iron as long as their concentrations are held below the toxic levels. Gold (0.1 mM) was reduced from Au(I) to Au(0), as evidenced by a color change from yellow to brownish black, but it likely proved lethal to the cells, as transfers from gold-containing cultures did not grow.

• Open in new tab
• Download powerpoint

FIG. 3.

SEM (A), EDX analysis (B), and XRD analysis (C) of uranium precipitated by X514. SEM (D), EDX analysis (E), and XRD analysis (F) of rhodochrosite formed by the reduction of Mn(IV) oxide by X514.

Three strains (X513, X514, and X561) were examined for Fe(III) reduction with Fe(III) oxyhydroxide as an electron acceptor and lactate (10 mM) as an electron donor under different salinities (0.05 to 7% NaCl) and pHs (6.5 to 9.6). Thermophilic iron reduction and magnetite formation occurred at salinities ranging from 0.05 to 5% NaCl (wt/vol) with lactate as an electron donor and Fe(III) oxyhydroxide as an electron acceptor. Magnetite formation was not observed at 6% or higher NaCl. The thermophilic bacteria exhibited optimum growth at 3% NaCl with Fe(III) citrate (15 mM) as an electron acceptor and lactate (10 mM) as an electron donor. These results are consistent with the effect of salinity on the growth of the thermophilic metal-reducing bacteria isolated at the subsurface environments (4, 21). Microbial transformation of Fe(III) oxyhydroxides (FeOOH) to magnetite (Fe₃O₄) was observed at a pH range of 6.9 to 8.9. The E_h of the initial medium (−40 to −50 mV) was decreased to −200 mV (under N₂ headspace gas) and −450 mV (under H₂-CO₂ headspace gases), and pH was also decreased from 8.0 to 7.4 upon incubation. Measured E_h and pH were consistent with the thermodynamic stabilities of magnetite and siderite (3, 46, 58). In other experiments at pH values slightly below neutral (6.6 to 6.9), minor amounts of magnetite were observed upon incubation with the thermophilic microorganisms.

The Fe(III) reduction process was sensitive to oxygen and metabolic inhibitors as well. Three isolates (X513, X514, and X561) were viable and fermented glucose on exposure to O₂. However, exposure of the thermophilic iron-reducing bacteria to oxygen decreased the Fe(III) reduction rate, indicating the anaerobic nature of the Fe(III)-reducing microorganisms. The formation of magnetic iron oxides by three thermophilic strains was also prevented by the addition of the ferric reductase inhibitor pCMPS, the hydrogenase inhibitor quinacrine dihydrochloride, the protonophore CCCP, and the electron transport inhibitors HOQNO and dicumarol. The susceptibility of the thermophilic Fe(III)-reducing cultures to these metabolic inhibitors indicated that the iron reduction by these bacteria appears to be an enzymatic process.

Mineralogical characteristics.All three strains were able to reduce Fe(III) oxyhydroxide and produce magnetite (Fe₃O₄) under temperatures ranging from ∼45°C to 70°C (Fig. 4). TEM analysis revealed elongated prismatic magnetite particles in X514 samples after 21 days of incubation at 60°C (Fig. 5B). The crystal morphology of the elongated prismatic magnetite particles formed by the X514 strain differed from that of the magnetite crystals formed by the Thermoanaerobacter ethanolicus strain (TOR-39) isolated from Taylorsville Triassic Basin, because TOR-39 produced sharp and well-formed octahedral crystals (59). The magnetite particles were associated with acicular goethite (Fe₂O₃) and thin, needle-like Fe(III) oxyhydroxides. Also, degradation of the texture of magnetite, such as rounded margins, was observed at temperatures of 70°C after 21 days of incubation (Fig. 5C). In addition, poorly crystalline magnetite crystals associated with fine-grained Fe(III) oxyhydroxides were observed at 50°C (Fig. 5A). Goethite (FeOOH) was also formed at 45, 65, and 70°C during the microbial transformation of the Fe(III) oxyhydroxides. The presence of goethite in some samples suggests that Fe(III) oxyhydroxides may be abiotically transformed into goethite through a dehydration process during incubation (47). This likely occurred due to the experimental temperature (>45°C) used in the incubations (47). Neither Fe(III) reduction nor magnetic iron mineral formation was observed below 40°C or above 75°C. No magnetic mineral formation was observed in the abiotic control experiments.

• Open in new tab
• Download powerpoint

FIG. 4.

XRD analysis of iron minerals formed by the thermophilic iron-reducing bacteria (X514) at different temperatures.

• Open in new tab
• Download powerpoint

FIG. 5.

TEMs of magnetite formation in X514 cultures at 50°C (A), 60°C (B and C), and 70°C (D).

The gas composition of the headspace atmosphere appeared to affect the mineral formation. All reddish-brown Fe(III) oxyhydroxides in culture bottles under a N₂ headspace changed to black, with the black precipitate being magnetite, as determined by XRD analysis. All reddish-brown solids under N₂-CO₂ and H₂-CO₂ changed to brownish black. Tubes incubated under a N₂-CO₂ atmosphere contained a mixture of magnetite and siderite, as determined by XRD analysis, whereas under a H₂-CO₂ atmosphere, siderite (FeCO₃) was the principal reduced iron mineral. It appeared that a H₂-CO₂ atmosphere provides a favorable energetic environment, allowing the complete reduction of Fe(III) oxyhydroxides plus considerable carbon dioxide and facilitating the formation of siderite rather than less-reduced magnetite (3, 46, 58). This study indicated that geochemical and environmental factors of iron biomineralization include atmospheric composition, pH, incubation temperature, and salinity. The presence of CO₂ gas was an important factor allowing the complete reduction of Fe(III) oxyhydroxide and the formation of siderite in subsurface environments.