Genomic and Physiological Characterization of the Chromate-Reducing, Aquifer-Derived Firmicute Pelosinus sp. Strain HCF1

Pelosinus sp. strain HCF1 was isolated from the effluent of an anaerobic, chromate-reducing, flowthrough column containing aquifer sediment from the DOE Hanford 100H site and eluted with sterile, synthetic groundwater. The synthetic groundwater contained the following constituents: 1 mM phosphate buffer (0.58 mM K₂HPO₄ and 0.42 mM KH₂PO₄, pH 7), 1.25 mM NaHCO₃, 0.25 mM CaCl₂, 0.1 mM NH₄Cl, 5 mM sodium lactate (the carbon source and electron donor), 5 μM KCrO₄, and 7.5 mM Na₂SO₄. Highly purified water (18-MΩ resistance) obtained from a Milli-Q Biocel system (Millipore, Bedford, MA) was used to prepare the synthetic groundwater and all other aqueous solutions described in this article. Although sulfate was available as an electron acceptor, ion chromatographic monitoring of the effluent confirmed that lactate was fermented to acetate and propionate and that sulfate reduction was not required for this process.

Isolation of strain HCF1 from column effluent was accomplished using a modified Hungate roll-tube technique (8) with serial dilutions to 10⁻⁵. Direct microscopic cell counts showed that there were approximately 1 × 10⁷ cells/ml in the effluent samples. Basal medium (0.58 mM K₂HPO₄, 0.42 mM KH₂PO₄, 0.25 mM CaCl₂, 0.1 mM NH₄Cl) containing 2% Noble agar and amended with 20 mM sodium lactate, 10 mM NaHCO₃, and vitamin, trace element, and selenite-tungstate solutions described elsewhere (9) was used for isolation. The inoculated roll tubes were incubated at 30°C for 7 to 10 days until bacterial colonies appeared. Second and third rounds of purification were conducted until the cells and colonies were uniform based on microscopic examination. Although yeast extract is not essential for the growth of strain HCF1, 0.5 g/liter yeast extract was routinely used in the medium to facilitate growth and colony formation.

After isolation, routine cultivation of strain HCF1 was carried out under strictly anaerobic conditions in an anaerobic glove box (Coy Laboratory Products, Inc., Grass Lake, MI) with a nominal gas composition of 86% N₂–10% CO₂–4% H₂. The growth medium was the same as the medium used for isolation with the omission of Noble agar. Anaerobic techniques used in the preparation of growth medium and stock solutions were described elsewhere (10).

All cell suspension experiments described in this article were performed under strictly anaerobic conditions in an anaerobic glove box. The glass and plastic materials used to contain or manipulate the cultures were allowed to degas in the glove box for at least 1 day before use. Cell suspension assays performed to assess chromate reduction by strain HCF1 under fermentative conditions were conducted in a similar manner to those described previously for Pseudomonas sp. strain RCH2 (9). Strain HCF1 was grown anaerobically at 30°C in a glove box, harvested in mid-log phase (optical density [OD] ≈ 0.2, after 48 h of incubation) by anaerobic centrifugation (11,899 × g, 15 min, 15°C), washed with 100 ml anaerobic basal medium, and resuspended anaerobically in the glove box. Assays were performed in butyl rubber-stoppered Balch tubes and initiated by adding 1 ml of concentrated cells to 9 ml of basal medium amended with the appropriate compounds, such as 50 μM chromate and 20 mM lactate (final concentration). Assays were run in duplicate at ∼30°C. Samples (500 μl) were collected before cells were added, immediately after cell addition, and typically every 30 min thereafter. Two hundred microliters of the suspension was used for OD₆₀₀ (optical density at 600 nm) measurement, and the remainder was spun down at ∼20,800 × g (4°C, 4 min). The freshly prepared supernatants and cell pellets were used for various analyses described below. Controls included assay mixtures that did not contain lactate (normally added at 20 mM) and in which cells were inactivated by heating (under anaerobic conditions) in a boiling water bath for 15 min. The potential role of extracellular reductants in Cr(VI) reduction was investigated by conducting assays with 10 ml of filtered (0.2-μm pore size) spent medium after growth or 10 ml of filtered, spent medium that had been heat denatured anaerobically in a boiling water bath for 15 min. These tests of extracellular reductants were initiated by adding 50 μM KCrO₄.

Anaerobic cell suspension experiments to assess the reduction of Fe(III) (2 mM) under fermentative conditions were conducted as described above for Cr(VI) reduction studies, except that uncentrifuged samples (not supernatants) were used for determination of the reduced metal. Fe(III)-nitrilotriacetic acid (NTA) stock solutions (100 mM) were prepared by dissolving 0.82 g NaHCO₃, 1.28 g of trisodium nitrilotriacetic acid (Na₃NTA), and 1.35 g of FeCl₃ · 6H₂O in 50 ml anaerobic H₂O in an anaerobic glove box and filter sterilizing into an anaerobic serum bottle. Amorphous Fe(OH)₃ was prepared as described previously (11) and stored in an anaerobic glove box for at least several weeks before use to deoxygenate.

Analytical methods for all analytes other than Fe(II) were described in detail previously (9) and are briefly summarized here. For routine determination of cell density, the OD₆₀₀ of diluted samples was measured using a model UV160U spectrophotometer (Shimadzu Corporation). It was determined that 1 OD₆₀₀ unit was equivalent to 1 × 10⁹ cells/ml. Direct microscopic cell counts were also determined in cell suspension experiments. Lactate, acetate, propionate, nitrate, and nitrite concentrations were measured by ion chromatography (IC) using a model ICS-2000 IC (Dionex Corporation) (9). Cr(VI) in cell suspension supernatants was measured with a spectrophotometric DPC (diphenylcarbazide) microplate assay (9). Total Cr in cell suspension supernatants was measured by inductively coupled plasma-mass spectrometry (ICP-MS) (9).

Fe(II) generated from Fe(III)-NTA or Fe(OH)₃ reduction by HCF1 cells was measured with a spectrophotometric Ferrozine assay conducted in an anaerobic glove box. The microplate assay used in this study was adapted from a method described by Stookey (12). Microplates (96 wells) that had been stored in an anaerobic glove box for at least 1 day were amended with 90 μl of 1 N HCl, and a 10-μl cell suspension sample was added to the HCl immediately after sampling. One hundred microliters of Ferrozine solution (1 g/liter Ferrozine, 500 g/liter ammonium acetate) was then added to the acidified sample. After a 10-min incubation, the absorbance at 555 nm was measured using a model 550 microplate reader (Bio-Rad). Fe(II) standards (0.2, 0.5, 1, and 2 mM ferrous ammonium sulfate hexahydrate in 1 N HCl) were included on each microtiter plate.

DNA was extracted from effluent of the same fermenting columns from which strain HCF1 was originally isolated. Effluents were collected directly into 50-ml tubes containing 30 ml of an RNA preservation reagent (25 mM sodium citrate, 10 mM EDTA, 10 M ammonium sulfate, pH 5.2). After overnight incubation at 4°C, the solution was filtered (0.22-μm pore size), and DNA and RNA were coextracted from filter sections using a modification of the extraction procedure reported by Ivanov and coworkers (13). This modification involved the addition of 1 g of Chelex-100 to the extractions to chelate iron; all other procedures were as reported previously.

Extracted DNA (10 ng) was used as a template for PCR amplifications using the following primers (515F and 907R) targeting the V4-V5 hypervariable regions of the small-subunit (SSU) rRNA gene with error-correcting barcodes (14): 515F-454 (5′-CCTATCCCCTGTGTGCCTTGGCAGTCTCAGAGGTGCCAGCMGCCGCGGTAA-3′) and 907R-454 (5′-CCATCTCATCCCTGCGTGTCTCCGACTCAGGTGAGGTCGCTAACCCGTCAATTCCTTTRAGTTT-3′), where underlined text represents a spacer, italic type represents a barcode, and bold text represents a priming region.

PCRs were performed using 0.625 units of TaKaRa Ex Taq polymerase (TaKaRa, Madison, WI), 1× reaction buffer, 200 μM deoxynucleoside triphosphates (dNTPs), 0.56 mg/ml bovine serum albumin (BSA), and 1 μM each primer with the following thermocycling parameters: initial denaturation at 95°C for 1 min followed by 25 cycles of 95°C for 20 s, 30 s of annealing at 66°C, and extension at 72°C for 1 min. The final product extension was at 72°C for 10 min. Reaction primer dimers were removed from the PCR products via solid-phase reversible immobilization (SPRI) bead purification according to the manufacturer's protocol (AMPure XP; Beckman Coulter Genomics, Danvers, MA) before being checked for quality and quantity on a Bioanalyzer 2100 using a DNA 7500 chip (Agilent Technologies, Santa Clara, CA). Each PCR sample was normalized to 30 ng, and samples were combined for multiplex sequencing. Sequencing libraries were created using the LV emu-PCR kit (Lib-L; Roche, Indianapolis, IN), and sequencing was conducted on a GS-FLX sequencer (Roche, Indianapolis, IN) at the Veterans Medical Research Foundation, La Jolla, CA.

Analysis of pyrotag sequences was performed using the QIIME suite of tools (15), including quality filtering to quality score 20 (Q20), clustering at 97% sequence identity, and classification using the RDP classifier (16).

The 16S rRNA gene of strain HCF1 was amplified by the PCR using a universal bacterial 16S forward primer, 27F (5′-AGAGTTTGATCCTGGCTCAG-3′), and reverse primer 1492R (5′-TACCTTGTTACGACTT-3′) (17). PCR products were cloned using One Shot TOP10 chemically competent Escherichia coli and a TOPO TA cloning kit (Invitrogen) and were sequenced at the UC Berkeley DNA Sequencing Facility (Berkeley, CA). The 16S rRNA gene sequence of strain HCF1 was BLASTed (18) against the GenBank nr and RDPII databases. A 16S rRNA phylogenetic tree was generated using MEGA, version 4 (19).

Genomic DNA library construction for Illumina sequencing of strain HCF1 was performed using the Nextera DNA Sample Prep kit (Illumina Compatible) according to the manufacturer's instructions using LMW buffer (Epicentre Biotechnologies, WI). Fragment size distribution estimated using an Agilent Bioanalyzer (Agilent) was centered at 400 bp with a distribution between 200 and 700 bp. Genome sequencing was performed on the Illumina platform (GAIIx) using a paired-end-read library with a read length of 100 bp and an insert size of 450 bp. We obtained a total of 4,183,216 pairs of reads, corresponding to an estimated 40- to 80-fold coverage of the genome assuming a genome size of 5 to 10 Mb. All raw reads were quality trimmed to an error probability of 1 in 1,000 (Q30). After trimming, reads shorter than 20 nucleotides were removed. A total of 3,957,260 reads were assembled de novo using Velvet (20) (k = 57; coverage cutoff = 24; expected coverage = 48; insert size = 450), generating 214 contigs with an N₅₀ length of 79.5 kb.

Genome annotation was first performed automatically using the JGI IMG portal (21) and then manually curated (as described in Results and Discussion). A search for the CXXCH heme-binding motif that characterizes c-type cytochromes was carried out using JGI IMG and ScanProSite (http://prosite.expasy.org/scanprosite/) search tools.

Duplicate 50-ml cultures of HCF1 were grown anaerobically using routine medium supplemented with 20 mM lactate with or without 5 mM NaNO₃. Five-milliliter samples of HCF1 cultures were collected in triplicate after 36 h and 72 h and immediately mixed with 2 volumes of the RNA preservation reagent described above. Cells were then harvested anaerobically by centrifugation (10,300 × g, 20 min, 4°C), and DNA and RNA were extracted using an AllPrep DNA/RNA minikit (Qiagen Sciences, Inc.). Cell pellets were resuspended in 600 μl RLT Plus buffer containing 1% β-mercaptoethanol and homogenized using a FastPrep-24 homogenization system (MP Biomedicals). DNA/RNA samples were then extracted according to the manufacturer's instructions except that the residual DNA in RNA samples was digested on-column using RNase-free DNase (Qiagen) for 15 min. DNA/RNA samples were finally eluted in 30 μl RNase-free water and quantified using a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific, Inc.). DNA/RNA samples were stored at −80°C.

For calibration standards, gene-specific primers were designed to amplify the rpoB, narG, nrfH1, and nrfH2 genes from strain HCF1 (see Table S1 in the supplemental material). PCR primers were designed with Primer 3 software (22). Fifty-microliter PCR mixtures contained 5 μl of 5× Ex Taq buffer, 4 μl of dNTP mix (2.5 mM each dNTP), 2.5 μM forward and reverse primers (2 μl each), 0.3 μl TaKaRa Ex Taq DNA polymerase (TaKaRa Bio, Inc.), 100 ng DNA template, and nuclease-free water. PCR amplification was performed using a Bio-Rad iCycler (Bio-Rad Laboratories) and the following thermocycling program: 2 min at 98°C; 30 cycles of 10 s at 98°C, 30 s at 55°C, and 2 min at 72°C; and 7 min at 72°C, maintained at 4°C. PCR products were purified using a QIAquick PCR purification kit (Qiagen). Amplicons of the rpoB, narG, nrfH1, and nrfH2 genes were purified, quantified (NanoDrop 1000 spectrophotometer), and serially diluted as DNA templates for calibration (10 to 10⁵ copies/reaction).

Quantitative PCR (qPCR) primers were designed with Primer 3 software. For each RNA sample, cDNA products were synthesized by SuperScript III reverse transcriptase and random hexamers (Invitrogen Corporation) according to the manufacturer's instructions, except that 2 μl 10 mM dNTP and 1 μl SuperaseIn (Ambion) were used in each reverse transcription (RT) reaction to inhibit RNA degradation. Eleven microliters of RNA was used as the template in each 20-μl RT reaction mixture. After RT, 20 μl final cDNA was 10-fold diluted by adding 180 μl nuclease-free water. For qPCR, 20-μl qPCR mixtures contained 10 μl 2× SYBR green Supermix (Bio-Rad Laboratories), 2.5 μM forward and reverse primers (see Table S1 in the supplemental material) (2 μl each), diluted cDNA template (5 μl), and 1 μl nuclease-free water. qPCR analysis was performed with a MyiQ single-color real-time PCR detection system (Bio-Rad Laboratories) using the following temperature cycling program: 2 min at 50°C; 8.5 min at 95°C; and 40 cycles of 15 s at 95°C, 30 s at 50°C, 30 s at 70°C, and 15 s at 81°C. The real-time data were acquired at the 81°C stage to exclude noise from primer dimers. For each gene, 5-point standards containing 10, 10², 10³, 10⁴, and 10⁵ copies/reaction were included in the same qPCR run to generate the calibration curve for absolute quantification.

One-hundred-milliliter cultures of HCF1 were grown anaerobically using routine medium supplemented with 20 mM lactate or 20 mM fructose. Anaerobically, cells were then spun down, washed, resuspended in 10 ml basal medium supplemented with the same substrates (20 mM lactate or 20 mM fructose) and 10 mM NaHCO₃, and fixed with a headspace of 80% N₂–20% CO₂. In addition, 100 ml of lactate-grown HCF1 cells was resuspended in 10 ml basal medium with 10 mM NaHCO₃ and fixed with a headspace of 90% H₂–10% CO₂. After 1 h of incubation, triplicate 3-ml samples (∼3 × 10⁹ cells) were mixed with 6 ml of the RNA preservation reagent described above prior to RNA isolation.

PCR amplification to make calibration standards and RT-qPCR conditions were the same as those described above, except that PCR and qPCR primers for the six target hydrogenase genes from strain HCF1 are given in Table S2 in the supplemental material.

This whole-genome shotgun project has been deposited at GenBank under the accession number AMZI00000000. The version described in this paper is the first version, AMZI01000000. 16S rRNA gene pyrotag sequence data for the flowthrough column from which strain HCF1 was isolated have been deposited in the MG-RAST database under accession number 4502869.3.

Pelosinus sp. strain HCF1 ferments a range of sugars and short-chain mono- and dicarboxylic acids to propionate and acetate. For example, lactate, which has been used in the polymeric form (glycerol polylactate) to stimulate in situ reductive immobilization of chromium in groundwater (23), is fermented to propionate and acetate in a 2:1 ratio (see Fig. S1 in the supplemental material). In addition to lactate, carboxylic acids and sugars that serve as growth substrates and are fermented to propionate and acetate include the following (propionate/acetate ratios are in parentheses): fumarate (1.5:1), fructose (2:1), glucose (2:1), cellobiose (2:1), and mannitol (4:1). These ratios were determined during growth in the absence of yeast extract or H₂ in the initial headspace.

Anaerobic suspensions of lactate-grown cells of strain HCF1 reduced chromate in a minimal, anaerobic buffer containing 20 mM lactate as the electron donor and no electron acceptor (other than ∼45 μM chromate) (Fig. 1A). The initial, specific rate of Cr(VI) reduction was ∼1.5 × 10⁻¹¹ μmol h⁻¹ cell⁻¹. Chromate reduction was much slower in the absence of lactate than in its presence and was negligible in controls with killed cells and lactate (Fig. 1A). Whereas Fig. 1A depicts dissolved Cr(VI) concentrations based on the DPC assay, which is specific to Cr(VI), Fig. 1B depicts total dissolved Cr concentrations as determined by ICP-MS. Comparison of the results shown in Fig. 1A and B, which are based on aliquots from the same samples, makes it clear that strain HCF1 reduced Cr(VI) but that the reduced Cr did not precipitate, even though Cr(III) would be expected to form poorly soluble oxides or hydroxides at pH values of >5 (24). The bacterial production of reduced, soluble Cr has been observed previously, for example, in Pseudomonas stutzeri strain RCH2 (9), and may be attributable to the formation of Cr(III)-organic complexes [e.g., Cr(III)-NAD⁺ complexes formed during Cr(VI) reduction by flavin reductases (25, 26)] or to kinetic limitations on Cr(III) precipitation. Notably, the Cr(VI) and total Cr concentration profiles for the no-lactate control were similar to each other (Fig. 1), suggesting that the small amount of Cr(VI) reduced under these conditions precipitated. It is not clear why reduced Cr precipitated more readily in the cell suspensions without lactate, but it is possible that a soluble Cr(III)-propionate complex formed in lactate-containing cultures. In the absence of lactate, the electron donor may have been hydrogen, which is typically present in the glove box atmosphere at 1 to 2%. Additional experiments suggested that H₂ could serve as an electron donor for Cr(VI) reduction, albeit not a very favorable one (see Fig. S2 in the supplemental material). Experiments were also conducted to test the possibility that soluble chemical reductants or redox-active proteins were released by strain HCF1 during metabolism of lactate that might cause extracellular Cr(VI) reduction. These cell-free experiments used filtrates of spent medium after growth of strain HCF1 with lactate; the filtrates were tested with and without heat denaturation under anaerobic conditions. In all cases, no Cr(VI) reduction was detected (data not shown), ruling out the role of extracellular chromate reduction by biogenic chemical reductants. These experiments did not address the possibility of chromate-inducible extracellular enzymes, because the filtrates were derived from cells grown in the absence of chromate.

Although strain HCF1 is primarily characterized by its fermentative metabolism, it is capable of anaerobic Fe(III) reduction under the tested growth conditions (data not shown) and in cell suspensions. Anaerobic suspensions of lactate-grown cells of strain HCF1 reduced soluble Fe(III) in a minimal, anaerobic buffer containing 20 mM lactate as the electron donor and Fe(III)-NTA as the electron acceptor (Fig. 2). In the absence of cells or when cells were anaerobically autoclaved before resuspension, Fe(III) reduction was negligible (Fig. 2). Analogous to observations for Cr(VI) reduction studies, a small amount of Fe(III) was reduced when cells were suspended in the absence of lactate. Presumably, reducing equivalents were supplied by H₂ in the glove box atmosphere, as suggested by cell suspension studies in which H₂ was supplied as the sole electron donor at a range of concentrations (see Fig. S3 in the supplemental material). Clearly, H₂ is a more effective electron donor for Fe(III) reduction than Cr(VI) reduction (Fig. S2 and S3). The specific rate of Fe(III) reduction in the presence of lactate was ∼1.5 × 10⁻⁹ μmol h⁻¹ cell⁻¹ [approximately 100-fold higher than the rate of Cr(VI) reduction]. When solid-phase Fe(III), such as synthetic, amorphous Fe(OH)₃, was used rather than Fe(III)-NTA, the Fe(III) reduction rate was orders of magnitude lower (data not shown).

Strain HCF1 is capable of reducing millimolar quantities of nitrate when growing with lactate (Fig. 3A). However, under these conditions, nitrate reduction did not closely mirror lactate consumption; the most rapid lactate consumption occurred between 24 and 36 h, whereas the most rapid nitrate reduction occurred between 36 and 48 h (Fig. 3A). Beginning at 36 h, nitrite began to accumulate, and propionate and acetate production began to deviate from that observed during growth without nitrate (specifically, acetate production was higher and propionate production was lower in the presence of nitrate, despite similar lactate consumption rates in the presence and absence of nitrate). Since most nitrite accumulation occurred after lactate was depleted, it is not clear what electron donor was driving nitrate reduction after 48 h. Possibilities include lactate-derived formate and H₂, although these two compounds should also have been present during lactate fermentation. Indeed, cell suspension studies in which H₂ was the only electron donor supplied (90% H₂–10% CO₂ in the headspace) resulted in nitrate reduction to nitrite, whereas controls with no electron donor supplied (80% N₂–20% CO₂ in the headspace) did not generate nitrite (see Fig. S4 in the supplemental material). For the study represented in Fig. 3, it is likely that nitrite was reduced further to ammonium, based upon two observations: (i) the amount of accumulated nitrite accounted for only approximately half of the nitrate reduced, and (ii) an ammonium-forming nitrite reductase (nrfA; two copies) is encoded in the genome (see below). More insight on nitrate/nitrite reduction was provided by targeted transcriptional analysis (reverse transcription-quantitative PCR [RT-qPCR]) of three putative nitrate or nitrite reductase genes conducted for samples collected at 36 and 72 h in the presence and absence of nitrate (Fig. 3B); these data are addressed in the section on c-type cytochromes below.

Based on its 16S rRNA gene sequence, strain HCF1 belongs to the Pelosinus-Sporotalea group in the Firmicutes (Fig. 4). It is most closely related to Pelosinus fermentans R7 (27) and Sporotalea propionicum TmPN3 (28), whose 16S rRNA gene sequences share 98.0% and 95.0% identities, respectively, with that of strain HCF1. Notably, two distinct 16S rRNA gene copies of differing lengths were amplified from strain HCF1 using universal forward (27F) and reverse (1492R) primers. The longer 16S rRNA gene of strain HCF1 contains an additional ∼100-bp insertion region at the 5′ end, whereas the rest of the two sequences are identical. Similarly, it has been reported that two length-variable copies of the 16S rRNA gene exist in Pelosinus sp. UFO1 and Sporotalea sp. TM1 (29), suggesting that this is not an unusual phenomenon in the Pelosinus-Sporotalea group (Fig. 4). Whether or not the longer copy in the strain HCF1 genome is a pseudogene remains to be determined.

A draft sequence of the genome of Pelosinus sp. strain HCF1 was obtained by paired-end sequencing on an Illumina GAIIx platform. After quality control, the 100-bp paired-end reads were assembled into 214 contigs using Velvet (20) with a coverage of 48×, an N₅₀ of 79.5 kb, and a maximum contig length of 288 kb. The genome size is estimated at 4.98 Mb with a GC content of 39.8% (Tables 1 and 2).

Analysis of the genome sequence of strain HCF1 indicates that it ferments lactate to propionate via the methylmalonyl-CoA pathway. The proposed enzymatic reactions involved in lactate fermentation by strain HCF1 are shown in Fig. 5A. As proposed, lactate is converted to pyruvate by lactate dehydrogenase (multiple copies are present in the genome). Pyruvate can be converted to acetyl-CoA by pyruvate formate-lyase (PFL) (two copies are present in the genome, Hcf1DRAFT_00614 and Hcf1DRAFT_04611, with adjacent genes encoding a PFL activase). In addition, pyruvate can be converted to acetyl-CoA by pyruvate:ferredoxin/flavodoxin oxidoreductase (PFOR) (Hcf1DRAFT_02505), which can generate reduced ferredoxin as an electron donor for hydrogenases. Acetyl-CoA can be converted to acetate (with ATP generation from substrate-level phosphorylation) via phosphotransacetylase (Pta) (Hcf1DRAFT_00266) and acetate kinase (Ack) (Hcf1DRAFT_03144).

The pathway from pyruvate to propionate through methylmalonyl-CoA appears similar to that used by the close relative Veillonella parvula strain DSM 2008 (GenBank accession number NC_013520); to illustrate, a similar organization for selected genes in the methylmalonyl-CoA pathway in both organisms is shown in Fig. 5B. However, strain HCF1 and V. parvula may differ in the enzymes catalyzing pyruvate carboxylation to oxaloacetate. For V. parvula, this is likely catalyzed by pyruvate carboxylase (Vpar_0752), but BLASTP searching (18) for this gene product did not result in any hits in the HCF1 genome. Instead, it is possible that pyruvate is converted to oxaloacetate by oxaloacetate decarboxylase (OAD) (Hcf1DRAFT_02350 shares 50% sequence identity to the alpha subunit of OAD in Vibrio cholerae [PDB accession number 2NX9]). However, this same gene (Hcf1DRAFT_02350) was also tentatively annotated as pyruvate carboxylase subunit B and does not occur in a cluster with the beta and gamma subunits of OAD as does the alpha subunit of OAD in Vibrio cholerae. Subsequent steps (Fig. 5A) could be catalyzed by malate dehydrogenase (Hcf1DRAFT_00468), fumarase (e.g., Hcf1DRAFT_00270), and succinate dehydrogenase (Hcf1DRAFT_02750-02752, i.e., Hcf1DRAFT_02750 to Hcf1DRAFT_02752). Propionate formation is proposed to occur via propionyl-CoA:succinate CoA transferase (Hcf1DRAFT_02208) (this shares 41% sequence identity with E. coli YgfH, which was demonstrated to have this activity [30]). Other enzymes in the proposed pathway (Fig. 5A) include methylmalonyl-CoA mutase (Hcf1DRAFT_00246-00247 and Hcf1DRAFT_02209-02210), methylmalonyl-CoA epimerase (Hcf1DRAFT_03696 and Hcf1DRAFT_03033, which share 97% sequence identity), and the Na⁺ ion-translocating methylmalonyl-CoA decarboxylase (the alpha subunit is likely Hcf1DRAFT_03697, which shares 79% sequence identity with Vpar_1244 in V. parvula; the beta subunit is likely encoded by Hcf1DRAFT_03584, Hcf1DRAFT_04139, and Hcf1DRAFT_01795, which collectively share 55 to 79% sequence identity with Vpar_1240). Note that the last enzyme, methylmalonyl-CoA decarboxylase, is not used by a better-studied lactate-fermenting bacterium, Propionibacterium freudenreichii, which instead uses methylmalonyl-CoA carboxytransferase to simultaneously decarboxylate (S)-methylmalonyl-CoA to propionyl-CoA and carboxylate pyruvate to oxaloacetate (31, 32). Many of the genes cited in this section appear likely to have been misannotated by automated annotation pipelines. Our reannotation of some of these genes was based in part on gene context. As shown in Fig. 5B, a number of the genes putatively encoding the methylmalonyl-CoA pathway in V. parvula and strain HCF1 are organized in similar gene clusters. Experimental validation of the functions of some of these genes is required for more definitive annotation.

The genome of fermentative strain HCF1 encodes an assortment of [NiFe]- and [FeFe]-hydrogenases, which are proposed to catalyze H₂ oxidation and H₂ production, respectively. The two [NiFe]-hydrogenases are likely membrane-associated respiratory uptake (H₂-oxidizing) hydrogenases (classified as group 1 hydrogenases by Vignais and coworkers [33]). The structural genes of the two [NiFe]-hydrogenases are localized in clusters that include genes for the three subunits (i.e., large, small, and b-type cytochrome). One [NiFe]-hydrogenase includes the genes Hcf1DRAFT_03506-03510, and the other includes Hcf1DRAFT_02596-02598 (with some “adjacent” genes on opposite DNA strands). Genes encoding proteins required for [NiFe]-hydrogenase assembly and maturation were identified, including hypFCDE (Hcf1DRAFT_00423-00420), hypAB (Hcf1DRAFT_00123-00122), and Hcf1DRAFT_03507-03506.

There is evidence for at least four [FeFe]-hydrogenases encoded in the genome of strain HCF1. Hcf1DRAFT_00661 was annotated as a ferredoxin hydrogenase large subunit and appears to belong to the group B1 monomeric [FeFe]-hydrogenases (as described by Calusinska and coworkers [34]). Hcf1DRAFT_02066 was also annotated as a ferredoxin hydrogenase large subunit but shares only 21% amino acid sequence identity with Hcf1DRAFT_00661. Hcf1DRAFT_02066 is part of a gene cluster that includes an alpha subunit of formate dehydrogenase (Hcf1DRAFT_02068), as has been observed for [FeFe]-hydrogenases identified in Eubacterium acidaminophilum and a variety of Clostridium spp. (34, 35). Considering the colocation of [FeFe]-hydrogenase and formate dehydrogenase genes, it is possible that the formate dehydrogenase interacts with the hydrogenase in a complex linking formate oxidation to H⁺ reduction.

Another putative [FeFe]-hydrogenase is Hcf1DRAFT_01773, which occurs in a 3-gene cluster (Hcf1DRAFT_01771-01773). Considerable sequence similarity exists between these three genes and NuoEFG, which are subunits of complex I (NADH:ubiquinone oxidoreductase); such similarity has been observed previously for certain [FeFe]-hydrogenases (33).

A fourth putative [FeFe]-hydrogenase, Hcf1DRAFT_02349, appears to belong to the group B2 monomeric [FeFe]-hydrogenases (as described by Calusinska and coworkers [34]). Notably, an adjacent gene, Hcf1DRAFT_02350, is putatively pyruvate carboxylase or oxaloacetate decarboxylase, which is involved in lactate fermentation to propionate (as discussed above).

Transcriptional studies of strain HCF1 were undertaken to explore differential expression of structural genes representing the two [NiFe]-hydrogenases and four [FeFe]-hydrogenases under several anaerobic conditions: (i) H₂ (a 90% H₂–10% CO₂ mixture) as the sole electron donor, (ii) lactate with no H₂ (80% N₂–20% CO₂ headspace), and (iii) fructose with no H₂ (80% N₂–20% CO₂ headspace). Although definitive trends were not apparent to distinguish expression among three conditions tested, one clear trend was that one [NiFe]-hydrogenase (represented by HCF1DRAFT_03510) was expressed at significantly higher levels than the other five hydrogenases tested (on average, ∼8-fold higher, normalized to rpoB and averaged across all conditions examined) (see Table S3 in the supplemental material).

Although many chromate-reducing bacterial species have been reported, only for a relatively small subset of these bacteria have the proteins catalyzing Cr(VI) reduction been identified and characterized biochemically, as reviewed elsewhere (36, 37). To date, the best-studied chromate reductase is ChrR, a soluble, dimeric, NADH-dependent flavoprotein reported in Pseudomonas putida (38, 39). BLASTP searches of the strain HCF1 genome were conducted for the following chromate reductases: ChrR from P. putida (GenBank accession number AF375642); an Old Yellow Enzyme (OYE) homolog from Thermus scotoductis (GenBank accession number AM902709 [40]); YieF, a ChrR homolog from E. coli (GenBank accession number AAK62985.2 [38]); the Fre flavin reductase from E. coli (EcoCyc accession number EG10334 [25]); FerB from Paracoccus denitrificans (putatively GenBank accession number YP_917833.1 [41, 42]); and an azoreductase from Bacillus sp. OY1-2 (GenBank accession number BAB13746.1 [43]). The best match to any of these proteins was Hcf1DRAFT_04683, whose predicted protein sequence shared 51% identity with ChrR from P. putida. Alignment of these two proteins (see Fig. S5 in the supplemental material) revealed high conservation of the LFVTPEYNXXXXXXLKNAIDXXS motif in Hcf1DRAFT_04683 (with a leucine-to-isoleucine deviation at position 90), indicating that it is a member of the NADH_dh2 family of putative flavin-binding quinone reductases (39). Notably, Hcf1DRAFT_04683 also shared 42% sequence identity with the FerB flavoprotein, which has been shown to be effective at reducing Fe(III)-NTA as well as Cr(VI) (41, 42) and thus could be relevant to Fe(III) reduction in strain HCF1. The only other BLASTP match of note was Hcf1DRAFT_00464, which shared 45% protein sequence identity with the NAD(P)H:flavin oxidoreductase (OYE homolog) from T. scotoductis.

Motivated in part by a desire to identify candidate c-type cytochromes that might be catalyzing Fe(III) or Cr(VI) reduction in strain HCF1, we searched all strain HCF1 open reading frames (ORFs) for the CXXCH heme-binding motif that characterizes c-type cytochromes. Of the 37 ORFs found to contain one or more instances of the CXXCH motif, most were clearly not c-type cytochromes (including many Fe-S proteins, the molecular chaperone DnaJ, and ribosomal proteins L31 and L32, among others). A list of encoded proteins containing a CXXCH motif is shown in Table 3; all genes with annotations clearly indicating that they do not encode c-type cytochromes have been excluded from the table, leaving only 13 genes, most of which are annotated as “hypothetical proteins.” There was no evidence of the multiheme c-type cytochromes that have been reported to catalyze Fe(III) reduction in Geobacter and Shewanella species (44). Instead, the five encoded c-type cytochromes that could be confidently annotated are all related to the reduction of nitrogen oxides (nitrate, nitrite, or nitric oxide). Included in the HCF1 genome are two highly similar (sharing 58 to 68% amino acid sequence identity) pairs of c-type cytochrome genes encoding a nitrite reductase complex composed of NrfH and NrfA (formate-dependent nitrite reductase): Hcf1DRAFT_02324-02323 and Hcf1DRAFT_00451-00450. In both cases, these genes are clustered with upstream cytochrome c biosynthesis genes. The NrfHA nitrite reductase complex could play a key role in the dissimilatory reduction of nitrate to ammonium coupled to the oxidation of formate or H₂, as has been reported in certain proteobacteria, such as Wolinella succinogenes (45). Notably, the genome of strain HCF1 also encodes a membrane-bound, respiratory nitrate reductase (narGHJI [Hcf1DRAFT_02301-02298]), which could reduce nitrate to nitrite, thereby supplying nitrite to the NrfHA nitrite reductase complex. Thus, it is possible that formate- or H₂-dependent dissimilatory reduction of nitrate to ammonium catalyzed by NarG, NrfA, and NrfH (among other proteins) is responsible for the above-discussed observation of nitrate reduction occurring during growth of strain HCF1 with lactate (Fig. 3A). Transcriptional (RT-qPCR) analyses of narG and the two putative nrfH copies in cultures of strain HCF1 growing with lactate in the presence and absence of nitrate provide some insight on which nitrate and nitrite reductases may be involved. It is clear (Fig. 3B) that both narG (Hcf1DRAFT_02301) and “nrfH2” (Hcf1DRAFT_02324) are nitrate inducible, whereas “nrfH1” (Hcf1DRAFT_00451) is apparently not and is less likely to be catalyzing nitrite reduction under the conditions tested.

Finally, to complete the list of well-defined c-type cytochromes encoded in the HCF1 genome, we identified a membrane-anchored nitric oxide reductase, which includes the c-type cytochrome NorC (Hcf1DRAFT_04668) and the b-type cytochrome NorB (Hcf1DRAFT_04669). The presence of nitric oxide reductase in strain HCF1 is unexpected, as this enzyme is typically associated with denitrification, and the genome sequence does not indicate that strain HCF1 has this capability (i.e., we could not identify genes encoding nitric oxide-forming nitrite reductases, such as NirS or NirK, or the nitrous oxide reductase NosZ). BLASTP searches of the closely related fermenter V. parvula and a range of Clostridium species did not reveal any matches to the NorC of strain HCF1.

Analysis by PCR amplification and pyrosequencing of SSU rRNA genes from fermenting column effluents demonstrated that a 16S rRNA gene taxon, identical to strain HCF1 over the region amplified (V4-V5 hypervariable region), was the dominant taxon, representing almost 80% of the pyrotag reads (Fig. 6).

Pelosinus sp. strain HCF1 and close relatives have been reported as dominant microbial community members in experimental systems inoculated with Hanford 100H aquifer sediment (Fig. 6) and groundwater (2) as well as laboratory (46) and field (5) aquifer systems associated with the Integrated Field Research Challenge site located at Oak Ridge National Laboratory (Tennessee). Considering its ability to become a prominent member of aquifer microbial communities and its diverse metabolic capabilities, Pelosinus sp. strain HCF1 (and close relatives) could play an important role in multiple biogeochemical cycles in aquifer environments. For example, in addition to catalyzing in situ Cr(VI) reduction under fermentative conditions in contaminated aquifers, it could potentially also reduce U(VI) (2, 47), nitrate (to nitrite and ammonium), and Fe(III). Although these reductive processes may not all be coupled to energy conservation in strain HCF1, they nonetheless could directly or indirectly mediate the mobility, toxicity, and bioavailability of several elements of environmental relevance. Strain HCF1 can also ferment a range of substrates relevant to both native and engineered conditions and could thereby release important metabolites for interspecies transfer, including acetate, propionate, H₂, and formate. Overall, the metabolic activities of strain HCF1 could mediate biogeochemical processes affecting the physicochemical and microbial components of its local aquifer environment.

Genomic and focused transcriptional analyses of strain HCF1 reported here have provided some insights into the underlying genetic basis of some of this organism's metabolic capabilities, including the methylmalonyl-CoA pathway and associated enzymes that catalyze the fermentation of lactate (and other substrates) to propionate and acetate, a variety of [NiFe]- and [FeFe]-hydrogenases that likely catalyze both H₂ production and consumption, flavoproteins that may catalyze Cr(VI) and Fe(III) reduction, and nitrate and nitrite reductases (NrfH, NrfA, and NarGHI) that likely play a role in dissimilatory reduction of nitrate to ammonium. These nitrite reductases (nrfH and nrfA), along with nitric oxide reductase (norCB), are among the unexpected genes identified in the genome of strain HCF1 (e.g., BLASTP searches revealed no homologs in the V. parvula genome). To date, identification of nrfA in any firmicute genome is rare (as indicated by the FunGene database, release 6/7/2012 [http://fungene.cme.msu.edu/]). The metabolic context relevant to some of the genes identified in strain HCF1, such as nitric oxide reductase, is not currently understood. Understanding should improve as a result of future studies that will focus on characterizing the gene expression of strain HCF1 in an environmentally relevant context, such as in flowthrough columns constructed with Hanford 100H aquifer sediment and eluted with synthetic groundwater that simulates in situ conditions.