Development of a Genetic System for the Chemolithoautotrophic Bacterium Thiobacillus denitrificans

Bacterial strains and plasmids.The plasmids and T. denitrificans (ATCC 25259) and Escherichia coli strains that were used in this study are described in Table 1.

Culturing conditions and growth media. E. coli was propagated according to established methods (20).

T. denitrificans cultures for electroporation and plasmid preparations were grown in modified M9 minimal medium (20) with additional constituents and vitamins prepared as described by Beller (5) and Widdel and Bak (23). The medium composition, per liter of water, was as follows: 6.8 g of Na₂HPO₄, 3.0 g of KH₂PO₄, 1.0 g of NH₄Cl, 2.52 g of NaHCO₃, 4.96 g of Na₂S₂O₃·5H₂O, 2.02 g of KNO₃, 0.2 g of MgSO₄·7H₂O, 0.03 g of CaCl₂·2H₂O, 2.7 mg of FeCl₃·6H₂O, 0.18 mg of CuSO₄·5H₂O, 50 μg of vitamin B₁₂, 100 μg of thiamine, and 1 ml of stock solution 6 (23). Solid medium also contained 15 g purified agar/liter (Oxoid, Hampshire, United Kingdom). Liquid cultures were grown on the bench top without aeration in capped flasks and tubes. Incubation of T. denitrificans on solid medium was performed inside an anaerobic glove box at 30°C, as colony growth of T. denitrificans was found to be limited under aerobic conditions. Ultrapurified water (18-MΩ resistance) obtained from a Milli-Q UV Plus system (Millipore, Bedford, MA) was used to prepare the growth medium and all other aqueous solutions.

Antibiotics were used where appropriate in agar plates and liquid cultures at the following concentrations (unless indicated otherwise): ampicillin, 100 μg ml⁻¹; chloramphenicol, 50 μg ml⁻¹; gentamicin, 50 μg ml⁻¹; kanamycin, 50 μg ml⁻¹; streptomycin, 50 μg ml⁻¹; spectinomycin, 100 μg ml⁻¹; tetracycline, 20 μg ml⁻¹.

Determination of plating efficiency.The cell density of suspended T. denitrificans cultures was determined by measuring the absorbance at 600 nm of washed cell suspensions and was correlated with microscopic cell counts made with a Petroff-Hausser counting chamber (Hausser Scientific, Horsham, PA). The plating efficiency was determined by serially plating 10-fold dilutions of the cell suspensions in triplicate and counting the CFU obtained.

DNA manipulations.Genomic DNA was isolated from T. denitrificans using a cetyltrimethylammonium bromide precipitation method (2). Plasmid DNA was isolated from E. coli and T. denitrificans using midi- or mini-plasmid purification kits (QIAGEN, Valencia, CA). PCR products and gel-excised plasmid fragments were purified with a QIAquick gel extraction kit (QIAGEN). Primers for amplification of T. denitrificans genomic DNA were designed from the whole genome sequence available through the Joint Genome Institute website at http://genome.ornl.gov/microbial/tbden/ and in the GenBank/EMBL database under accession no. CP000116. PCR amplification of DNA fragments containing T. denitrificans sequence was typically performed using Advantage-GC 2 polymerase (Clontech, Mountain View, CA), except where noted, to improve yields for polymerization of the high-GC (66%) T. denitrificans DNA. Manufacturer's protocols were used when working with the EZ-Tn5<KAN-2> insertion kit and the EZ-Tn5 pMOD-2<MCS> vector (EPICENTRE Biotechnologies, Madison, WI). DNA sequencing was performed by Davis Sequencing (Davis, CA).

Determination of antibiotic resistance.Various pTnMod minitransposon plasmids were prepared to 0.1 to 0.2 μg μl⁻¹ in 10 mM Tris buffer (pH 8.0), and 0.1 to 0.3 μg μl⁻¹ was electroporated into T. denitrificans as described below. Recipient cells were plated on modified M9 solid medium containing one of the following: 20, 25, or 50 μg ml⁻¹ kanamycin; 50 or 100 μg ml⁻¹ gentamicin; 50 μg ml⁻¹ streptomycin and 100 μg ml⁻¹ spectinomycin; 5, 10, or 20 μg ml⁻¹ tetracycline; or 20 or 25 μg ml⁻¹ chloramphenicol.

Construction of hynL insertion mutant.In order to perform gene disruption via homologous recombination, a hynL::kan product was created. The hynL gene (Tbd_1375) was amplified from T. denitrificans genomic DNA using Vent DNA polymerase (New England BioLabs, Ipswich, MA) and hynL-1 primers containing an XbaI site in the forward primer and an SphI site in the reverse primer (Table 2). The resulting 1.8-kb PCR product was digested with XbaI and SphI and ligated into XbaI/SphI-digested pUC19. The resulting pUC19-hynL plasmid was subjected to in vitro transposition with an EZ-Tn5<KAN-2> cassette system (EPICENTRE Biotechnologies) and transformed into E. coli. Resulting kanamycin-resistant colonies were screened for Tn-kan (Table 1) placement within hynL and sequenced for the exact Tn-kan location using EZ-Tn5 primers (Table 2). Linear DNA containing hynL::kan for electroporation into T. denitrificans was generated from pUC19-hynL::kan with pUC19 primers (Table 2) and Hi-Fi Taq polymerase (Invitrogen, Carlsbad, CA).

Construction of complementation plasmid.A complementation plasmid (Fig. 1) was developed based on the IncP broad-host-range vector pRR10 (19). The kanamycin promoter, P_Kan (10), was amplified from pTnMod-OKm′ using kanP primers (Table 2) and Platinum Taq polymerase (Invitrogen). The resulting 110-bp amplicon was ligated into the SmaI site of pRR10, generating pTL1 (Table 1) and creating a unique HpaI site immediately downstream of P_Kan. In order to introduce gentamicin resistance into pTL1, a transposon containing the aminoglycoside acetyltransferase gene aacC1 was created. First, an 836-bp fragment containing aacC1 was cut from pTnMod-OGm (Table 1) with SacI. This same fragment was ligated into the SacI site of EZ-Tn5 pMOD-2<MCS>, creating pMOD-2<gent> (Table 1). The gentamicin transposon Tn-gent (Table 1) was created according to the manufacturer's instructions via PCR amplification of pMOD-2<gent> with pMOD2-PCR primers (Table 2) and transposed into pTL1, completing the complementation vector pTL2 (Table 1; Fig. 1). The transposition reaction products were electroporated into TOP10 E. coli cells. The resulting gentamicin-resistant colonies were screened for ampicillin sensitivity, and the exact location of aacC1 was confirmed by sequence analysis. To generate the final hynL expression vector, pTL3 (Table 1), a fragment containing hynL and the downstream gene hupE was amplified from T. denitrificans genomic DNA using hynL-hupE primers (Table 2) and Phusion DNA polymerase (Finnzymes, Espoo, Finland) and was ligated into the HpaI site of pTL2. The resulting clones were screened by PCR analysis for correct orientation of hynL-hupE relative to P_Kan, and the sequence was confirmed.

• Open in new tab
• Download powerpoint

FIG. 1.

Schematic diagram of the complementation vector (pTL2; Table 1) used in T. denitrificans. Plasmid construction is described in Materials and Methods. The following features are shown: oriV, the RK2 minimal vegetative origin of replication; oriT, the origin of transfer; trfA, encodes the RK2 replication initiation protein; bla, the beta-lactamase gene encoding ampicillin resistance; P_Kan, the 110-bp promoter of the kanamycin resistance gene from pTnMod-OKm′; and relevant restriction sites. The HpaI restriction site is unique and allows placement of genes for complementation. KpnI and HindIII bracket the original MCS from pRR10.

Construction of hydA insertion mutant.A disruption was created in the T. denitrificans gene encoding the large structural subunit of a group 3b hydrogenase, hydA (Tbd_1263), as follows. The hydA gene with approximately 1 kb flanking DNA on each end was amplified from T. denitrificans genomic DNA using KpnI-modified hydA primers (Table 2). The resulting 3.8-kb PCR product was digested with KpnI and ligated into KpnI-digested pUC19, yielding pUC19-hydA (Table 1). This plasmid was subjected to in vitro transposition with Tn-gent and transformed into E. coli. Resulting gentamicin-resistant colonies were screened for Tn-gent placement within hydA and sequenced for the exact Tn-gent location using pMOD2-Seq primers (Table 2). Linear DNA containing hydA::gent for electroporation into T. denitrificans strain TL001 (Table 1) was generated from pUC19-hydA::gent with pUC19 primers (Table 2).

Electrocompetent cell preparation and electrotransformation.One Shot TOP10 Electrocomp E. coli cells (Table 1; Invitrogen) were purchased and transformed according to the manufacturer's instructions.

Electrocompetent T. denitrificans cells were prepared from 48-h (late log phase), 200-ml cultures (optical density at 600 nm = 0.1 to 0.25; 1 × 10⁸ to 2.5 × 10⁸ cells ml⁻¹). All manipulations were carried out on ice, and all solutions were sterile and ice cold. Cells were harvested by centrifugation at 4°C for 10 min at 3,220 × g. The cells were washed twice in 100 ml ultrapure H₂O and resuspended in 500 μl ultrapure H₂O. Cell integrity of T. denitrificans was monitored throughout the wash procedure by phase-contrast microscopy. Cells were freshly prepared immediately before electroporation.

All electrotransformations were performed in 0.2-cm-gap Bio-Rad Gene Pulser cuvettes using a Bio-Rad Gene Pulser II equipped with a Pulse Controller Plus. Electrocompetent T. denitrificans cells (50 μl) were pulsed at 12.5 kV cm⁻¹ for ∼5 ms (resistance = 200 Ω; capacitance = 25 μF). Typically, 2 to 5 μl containing ∼100 to 500 ng DNA was used for electroporation. Cells were recovered in 1 ml ice-cold, modified M9 medium immediately following electroporation and transferred into sterile 1.7-ml Microfuge tubes. Electroporated cells were allowed to recover for 24 h at room temperature prior to plating on solid medium containing antibiotics. Each T. denitrificans electrotransformation experiment included a negative control in which no DNA was added. In all cases, no background growth of T. denitrificans on selective medium was observed.

Hydrogenase activity assay.In vivo assays were conducted with wild-type, mutant, and complemented mutant strains of T. denitrificans to assess specific hydrogenase activity. Cells (100 to 200 ml) were grown with thiosulfate and nitrate under strictly anaerobic conditions as described elsewhere (5), except that antibiotics were included in the growth medium (e.g., 50 μg ml⁻¹ kanamycin for the hynL mutant, 50 μg ml⁻¹ gentamicin for the complemented mutant). Cultivation and handling of cells for the in vivo assay were conducted in an anaerobic glove box with headspace containing ∼3% hydrogen (5). T. denitrificans cells in late exponential phase (∼2 × 10⁸ cells ml⁻¹) were harvested anaerobically by centrifugation in sealed polycarbonate bottles and washed once with an anaerobic resuspension buffer described previously (5). Cells were then resuspended in the anaerobic buffer at a final density of 1 × 10⁹ to 2 × 10⁹ cells ml⁻¹. One milliliter of this cell suspension was added to a quartz cuvette with a polytetrafluoroethylene stopper (1-cm path length; Spectrocell, Oreland, PA). Then, 10 μl of an anaerobic 100 mM benzyl viologen solution was added to the cuvette. The cuvette was then immediately sealed with the stopper, gently mixed, and quickly removed from the glove box for analysis. Reduction of benzyl viologen was monitored spectrophotometrically by measuring absorbance at 555 nm every 20 s for 5 min. All materials used in the assays were stored in the anaerobic glove box for at least 1 day prior to use. Negative controls included assays without benzyl viologen or without hydrogen. The hydrogen-free negative controls involved a butyl rubber stopper rather than a polytetrafluoroethylene stopper and were prepared in the same manner as regular samples except that, after sealing the cuvette, the glove box atmosphere was replaced by alternately evacuating the cuvette headspace under vacuum and flushing with an H₂-free, anaerobic mixture of 90% N₂-10% CO₂.

Growth on solid medium and characterization of antibiotic sensitivity.In order to allow the isolation of clonal populations of genetically modified T. denitrificans, plating onto solid growth medium was optimized. Medium solidified by the addition of purified agar (Oxoid) yielded the best results (see Materials and Methods), with visible colonies observed after 7 to 10 days of growth under anaerobic conditions at 30°C. When late logarithmic liquid medium cultures containing (1.3 ± 0.2) × 10⁸ cells ml⁻¹ (mean ± standard deviation; n = 16)—as determined by direct cell counts in a Petroff-Hausser counting chamber—were plated on this medium, (1.6 ± 0.6) × 10⁸ CFU ml⁻¹ (n = 3) were recovered, demonstrating a high plating efficiency.

The growth of T. denitrificans on solid medium was inhibited by a variety of common antibiotics. Growth of 1 × 10⁹ cells ml⁻¹ plated on solid medium was inhibited by chloramphenicol (50 μg ml⁻¹), gentamicin (50 μg ml⁻¹), kanamycin (50 μg ml⁻¹), streptomycin (50 μg ml⁻¹) plus spectinomycin (100 μg ml⁻¹), and tetracycline (5 μg ml⁻¹). The growth of T. denitrificans in liquid medium was similarly inhibited, with one difference: T. denitrificans grown in liquid medium also displayed sensitivity to ampicillin (100 μg ml⁻¹). As T. denitrificans takes much longer to grow on solid medium (10 days, as opposed to 2 days in liquid medium), it is likely that the ampicillin had degraded during the 10 days it took for growth to occur on solid medium. These results indicated that several different antibiotic resistance markers could be used for selection of genetic variants.

Identification of potential expression vectors.Both broad- and narrow-host-range plasmids (13) were tested for the ability to replicate within T. denitrificans. The plasmids were introduced by electroporation, and viability after electroporation was demonstrated by both N₂ bubble formation (evidence of denitrification) and growth on solid media.

An IncP plasmid containing the RK2 minireplicon, pPA20 (1), and an IncQ plasmid, pANT4 (15), were evaluated for the ability to replicate in T. denitrificans. Electroporation of pPA20 into T. denitrificans resulted in multiple kanamycin-resistant isolates (5 to 10 isolates μg⁻¹ DNA). To ensure that pPA20 was successfully replicating in T. denitrificans, the plasmid was purified from one of the kanamycin-resistant transformants, checked for correct size by gel electrophoresis (data not shown), and reintroduced into T. denitrificans via electroporation. A significant increase in the number of transformants μg⁻¹ of plasmid DNA was seen with pPA20 isolated from T. denitrificans (1.4 × 10⁶ transformants μg⁻¹) when compared to pPA20 isolated from E. coli (5 to 10 transformants μg⁻¹), a 10⁵-fold increase. The low number of transformants resulting from the E. coli-purified plasmid is likely a result of plasmid degradation by host restriction/modification systems, of which T. denitrificans has several (6). Electroporation of pANT4 into T. denitrificans yielded no kanamycin-resistant isolates after several attempts.

Analysis of antibiotic resistance determinants in T. denitrificans.A series of minitransposon vectors (Table 1) were used to confer functional antibiotic resistance in T. denitrificans, including pTnMod-OKm′ (Kan^r), pTnMod-SmO (Str^r Spcm^r), pTnMod-OGm (Gent^r), pTnMod-OCm (Cam^r), and pTnMod-OTc (Tet^r), containing Tn5 inverted repeats, the transposase external to the repeats, and a conditional origin of replication (11). Evaluation of these five different antibiotic resistance determinants contained on pTnMod vectors showed that three constructs resulted in functional antibiotic resistance after transposition into the T. denitrificans genome, including kanamycin, gentamicin, and streptomycin-spectinomycin resistance. Electroporation of T. denitrificans with pTnMod-OTc and pTnMod-OCm did not result in tetracycline- or chloramphenicol-resistant colonies, respectively, using 5, 10, and 20 μg tetracycline ml⁻¹ or 25 and 50 μg chloramphenicol ml⁻¹. Since no colonies were obtained, it is not clear whether the lack of resistance resulted from lack of gene expression or lack of functionality of the antibiotic resistance marker in T. denitrificans.

Development of a gene disruption system for hynL and hydA.In order to disrupt the genes encoding the hydrogenase large structural subunits within T. denitrificans, gene replacement of hynL by hynL::kan, and of hydA by hydA::gent, was performed by using homologous recombination. Electroporation of wild-type T. denitrificans with 250 ng of hynL::kan linear DNA resulted in 10 CFU on modified M9 Km₅₀ plates after 7 to 10 days. PCR analysis with hynL-2 primers (Table 2), which anneal outside of the hynL gene (Fig. 2), was used to confirm homologous recombination within the T. denitrificans genome. Production of the expected 3.1-kb band (Fig. 2, lane 3) signified that hynL::kan recombined properly within the genome. Sequence analysis of the PCR products confirmed that the Tn-kan insertion was located 801 bases downstream from the hynL translation start site in the T. denitrificans genome. PCR analysis of pUC19-hynL::kan (Table 1; the plasmid used to generate the linear DNA fragment used for electroporation) with hynL-2 primers generated no product, as expected (data not shown).

• Open in new tab
• Download powerpoint

FIG. 2.

(A) Electropherogram of PCR products from wild-type (WT) T. denitrificans, the hynL mutant, and the complemented (Compl.) hynL mutant, as well as digested plasmid DNA from the complemented mutant. Lane 1, HyperLadder III, Bioline; lane 2, wild-type DNA, primers hynL-2-f and -r; lane 3, hynL mutant (strain TL001) DNA, primers hynL-2-f and -r; lane 4, complemented-mutant genomic DNA, primers hynL-2-f and -r; lane 5, complemented-mutant plasmid DNA, primers hynL-2-r and pUC19-r; lane 6, Hi-Lo Marker, Bionexus (the arrow indicates 8 kb); lane 7, complemented-mutant pTL3 plasmid DNA, NdeI digested. (B) Maps of primer positions and amplicon sizes corresponding to lanes 2 to 5 in panel A. Note that the hynL-2-f primer anneals with genomic DNA upstream of the hynL gene, whereas the pUC19-r primer anneals with pTL3 plasmid DNA, rendering these primers specific to the T. denitrificans genome and the complementation plasmid pTL3, respectively.

The confirmed hynL knockout strain (TL001; Table 1) was used to generate a second knockout in a group 3b hydrogenase, hydA. Homologous recombination of wild-type hydA with hydA::gent was successful, resulting in 60 to 75 colonies μg⁻¹ of transformed linear DNA. One strain, TL002 (Table 1), showed the correct-size product for hydA::gent amplification (data not shown) and was also confirmed by sequence analysis, which showed that the disruption was located 182 bp from the translation start site.

For both strains TL001 and TL002, no growth defects were apparent during cultivation with thiosulfate and nitrate.

Development of a complementation system to allow for in-trans expression of the hynL gene.To demonstrate that the phenotype of the mutant strain TL001 was due to disruption of the hydrogenase large structural subunit (hynL), and not due to a secondary gene mutation, an expression vector containing hynL was constructed. While pPA20 was able to replicate in T. denitrificans, it is not ideally suited as an expression vector due to its relatively large size (8 kb) and lack of a multiple cloning site (MCS). Since the RK2 minireplicon of plasmid pPA20 was sufficient for replication in T. denitrificans, other plasmids containing the same replicon were considered. The broad-host-range plasmid pRR10 (19) was used as the basis for the complementation vector since it contains the same replicon as pPA20, is relatively small, and contains the pUC19 MCS and lacZα fragment (for insert screening). As ampicillin resistance is not an effective selectable marker for T. denitrificans on solid medium, a gentamicin-resistant variant of pRR10 was generated. The aacC1 gene from pTnMod-OGm served as the gentamicin resistance marker in pRR10; after transposition, Tn-gent (containing aacC1) (Table 1) was determined by sequence analysis to be located 400 bp into the bla gene (Fig. 1).

The gene cluster containing the group 1 hydrogenase gene disruption has five genes between hynL and the promoter, with one additional gene, hupE, located immediately downstream of hynL (6). Microarray expression data (7 and unpublished data), as well as the short intergenic gaps between genes in this cluster, are consistent with the transcription of both hynL and hupE being dependent on a promoter located at the beginning of this cluster, approximately 4 kb upstream of hynL. Since identification of the hynL promoter was uncertain, an alternate promoter from the aph(3′)-I gene encoding kanamycin resistance on pTnMod-OKm′ was used instead of the native hynL promoter.

Approximately 100 ng of the complementation vector pTL3 was electroporated into TL001, and two to five gentamicin-resistant isolates were obtained. An 8.6-kb plasmid purified from a gentamicin-resistant TL001 isolate (Fig. 2, lane 7) was confirmed to contain hynL-hupE by both PCR analysis (Fig. 2, lane 5) and sequence analysis of the recovered plasmid. In addition, hynL-2 primers were used to confirm the maintained presence of hynL::kan within complemented strain TL001 (Fig. 2, lane 4).

Effect of genetic manipulations on hydrogenase activity.Interruption of the hynL gene, which encodes the large, active-site-bearing subunit of a periplasmic [NiFe]hydrogenase (6), resulted in a dramatic decrease in specific hydrogenase activity (Fig. 3). Compared to the rate of specific hydrogen oxidation in wild-type cells, the rate in mutant strain TL001 was, on average, 75% reduced in replicate experiments. The hydrogenase assay proved to be generally robust, as specific rates in replicate experiments with wild-type cells agreed within 1%. Negative control assays that excluded hydrogen or benzyl viologen resulted in rates that were negligible relative to the wild type (Fig. 3). Complementation of the hynL mutation (strain TL001/pTL3; Table 1) resulted in specific hydrogenase activity that exceeded that of the wild type (Fig. 3). Overall, the assay results for the hynL::kan mutant (strain TL001) and the complemented mutant indicate that the genetic system described here enabled successful manipulation of a targeted gene in T. denitrificans.

• Open in new tab
• Download powerpoint

FIG. 3.

Hydrogen oxidation (measured as in vivo benzyl viologen reduction) versus time for wild-type T. denitrificans, strain TL001 (hynL mutant), strain TL002 (hynL hydA double mutant), and the complemented hynL mutant (TL001/pTL3). Negative controls (averaged results for controls with no H₂ and controls with no benzyl viologen) are also shown. Each datum point represents the average of duplicate or triplicate assays. Linear-regression fits are plotted.

The observation that interruption of the hynL gene did not completely eliminate hydrogenase activity merits further discussion. Contamination with a wild-type copy of hynL is largely ruled out because we grew the TL001 culture in medium containing 50 μg ml⁻¹ kanamycin and could not detect the wild-type hynL gene in PCR analyses (including more sensitive real-time PCR analyses [data not shown]) of TL001 cultures that were subjected to the hydrogenase assay (Fig. 2, lane 3). We first hypothesized that the other hydrogenase in T. denitrificans, putatively a cytoplasmic, group 3b [NiFe]hydrogenase (6), was responsible for this residual activity in strain TL001. This would have been a noteworthy phenomenon, as homologous group 3b hydrogenases (e.g., in the well-studied hyperthermophile Pyrococcus furiosus) are characteristically H₂-evolving enzymes involved in disposing of excess reductant under fermentative conditions (22) rather than H₂-oxidizing enzymes that would function under the denitrifying conditions of this study. However, hydrogenase assays conducted with a double mutant (TL002; with mutations in the large subunits of both [NiFe]hydrogenases) indicated very similar activity to that of the single hynL mutant (Fig. 3), suggesting that the group 3b hydrogenase was not catalyzing significant hydrogen oxidation activity. The residual hydrogenase activity in the double mutant leaves open the possibility that an additional hydrogenase is encoded in the T. denitrificans genome but has not yet been identified. However, BLASTP searches for other kinds of hydrogenases (e.g., [Fe]hydrogenases, H₂-forming methylenetetrahydromethanopterin dehydrogenase, E. coli hydrogenase 3) against the T. denitrificans genome did not reveal convincing evidence of additional hydrogenases.

An alternative explanation for residual hydrogenase activity in the hynL mutant is that the translated product of the disrupted gene was still capable of catalyzing hydrogen oxidation, albeit at a greatly reduced level. This seems unlikely for two reasons: (i) the EZ-Tn5<KAN-2> cassette contains stop codons in all three reading frames, which should preclude expression of a large portion (more than half) of the hynL gene, and (ii) the assembly and maturation of the active site of the hydrogenase should not proceed properly in the absence of key features at the C-terminal end of HynL, including the two cysteine residues in the very highly conserved C-terminal motif DPCxxCxxH/R and the endolytically formed peptide with a cleavage site located immediately downstream of this motif (8). The two conserved cysteine residues in this motif (along with two conserved cysteine residues in the N-terminal region of HynL) ligate the metal center to the large subunit (17).

The physiological role of hydrogenases in T. denitrificans remains unclear. Although the bacterium can oxidize H₂ as a sole electron donor under denitrifying conditions, this metabolism does not support growth (6) and appears to be relatively slow. Under the assay conditions used in this study (i.e., cells grown under denitrifying conditions with thiosulfate in the presence of H₂ and then resuspended with H₂ as the sole electron donor), in vivo, specific activity was on the order of 3 nmol·min⁻¹·mg protein⁻¹ (based on a molar absorption coefficient for benzyl viologen of 7,780 M⁻¹·cm⁻¹) (21). The use of hydrogenases in an H₂-evolving mode to dispose of excess reducing equivalents under fermentative conditions cannot be reconciled with the obligate chemolithotrophic lifestyle of T. denitrificans, which has never been shown to function under fermentative conditions. Now that a genetic system is available for T. denitrificans, mutants defective in each of the two encoded hydrogenases can be compared under a range of physiological conditions, which should enhance our understanding of the roles of each hydrogenase. In addition, traditional biochemical studies can be conducted with the double mutant (strain TL002) to investigate the possibility that there is a third, as-yet-unidentified hydrogenase in T. denitrificans.