Preferential Reduction of the Thermodynamically Less Favorable Electron Acceptor, Sulfate, by a Nitrate-Reducing Strain of the Sulfate-Reducing Bacterium Desulfovibrio desulfuricans 27774

Abstract

Desulfovibrio desulfuricans strain 27774 is one of a relative small group of sulfate-reducing bacteria that can also grow with nitrate as an alternative electron acceptor, but how nitrate reduction is regulated in any sulfate-reducing bacterium is controversial. Strain 27774 grew more rapidly and to higher yields of biomass with nitrate than with sulfate or nitrite as the only electron acceptor. In the presence of both sulfate and nitrate, sulfate was used preferentially, even when cultures were continuously gassed with nitrogen and carbon dioxide to prevent sulfide inhibition of nitrate reduction. The napC transcription start site was identified 112 bases upstream of the first base of the translation start codon. Transcripts initiated at the napC promoter that were extended across the napM-napA boundary were detected by reverse transcription-PCR, confirming that the six nap genes can be cotranscribed as a single operon. Real-time PCR experiments confirmed that nap operon expression is regulated at the level of mRNA transcription by at least two mechanisms: nitrate induction and sulfate repression. We speculate that three almost perfect inverted-repeat sequences located upstream of the transcription start site might be binding sites for one or more proteins of the CRP/FNR family of transcription factors that mediate nitrate induction and sulfate repression of nitrate reduction by D. desulfuricans.

Many different types of bacteria can adapt from aerobic to anaerobic growth, but when both oxygen and an alternative electron acceptor are available, they use oxygen preferentially. Similarly, during anaerobic growth, enteric bacteria use a powerful oxidizing agent, such as nitrate, in preference to nitrite or fumarate. This has led to a widespread assumption that when more than one potential source for energy is available, bacterial regulatory mechanisms ensure that the thermodynamically most favorable electron acceptors are used first (1, 14, 29, 36). This raises the fascinating question of whether bacteria normally associated with environments in which a less-powerful electron acceptor, such a sulfate, is abundant preferentially use a thermodynamically more favorable electron acceptor, such as nitrate.

There are many examples in the literature of sulfate-reducing bacteria that are also able to reduce nitrate (14, 16, 17, 27, 31, 38), but this capacity is absent from the two strains of Desulfovibrio vulgaris for which complete genome sequence data are available (12). There is little agreement concerning how nitrate reduction is regulated, even in strains of D. desulfuricans. Baumann and Denk (2) showed that nitrate reduction occurs only in the absence of sulfate, and others subsequently reported similar conclusions (7, 17). In contrast, two groups reported that for D. desulfuricans strain Essex, the presence of sulfate is essential for nitrate reduction to occur (14, 27). However, nitrate reduction was prevented at sulfate concentrations typically used in media to propagate sulfate-reducing bacteria. In contrast, McCready et al. (16) proposed that nitrate reduction is inhibited rather than genetically repressed by sulfide generated as a product of sulfate reduction, a proposal that has since been confirmed (7). At very low concentrations, sulfate was immediately consumed, so nitrate reduction was not inhibited. They also proposed, but did not prove, that sulfate “switches off” nitrate reduction. Consequently, it is still undetermined whether sulfate represses the expression of the D. desulfuricans nitrate reductase operon.

We recently reported the sequence of the nap operon (for nitrate reduction in the periplasm) and its upstream regulatory region of D. desulfuricans strain 27774 (15). This information has provided for the first time an opportunity to investigate at the RNA level how expression of the nitrate reductase operon is regulated in a sulfate-reducing bacterium. In the present study, we first report results of experiments designed to determine whether nitrate or sulfate are reduced preferentially by D. desulfuricans strain 27774 and compare growth yields on nitrate, nitrite, and sulfate. We confirm that sulfide strongly inhibits nitrate reduction by this strain. We then report results of experiments designed to determine whether the six genes of the D. desulfuricans 27774 nap operon are cotranscribed and therefore whether they form a single operon. Finally we demonstrate that nap operon transcription is induced by nitrate but repressed by sulfate and propose a possible mechanism for this regulation.

MATERIALS AND METHODS

Bacterial strains and growth media.

D. desulfuricans subsp. desulfuricans DSM 6949 (ATCC 27774) from the Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Braunschweig, Germany, was cultured in the following four media: Postgate B; Postgate C; Postgate N, which is a modified sulfate-free Postgate C liquid medium containing 14.7 mM nitrate or 5 mM nitrite; and Postgate SN, which is a modified medium containing 14.7 mM nitrate and 3.5 mM sulfate (22). The cultures were grown in sealed serum bottles (pregassed with N₂), inoculated with 10% (vol/vol) of a 48-h-old culture, and incubated at 30°C. The purity of the nitrate-grown culture was checked microscopically using the Gram stain by checking for inability of the grown culture to form colonies on nutrient agar during aerobic growth and for formation of a black FeS deposit when subcultured in Postgate B medium. Desulfovibrio vulgaris strain Hildenborough (NCIMB 8303) was also used in this study as a negative control for the growth experiments, since it lacks the periplasmic nitrate reductase operon. The bacteria for the growth yield experiments were grown in a modified Postgate zero medium that was supplemented with sulfate, nitrate, and nitrite at the required final concentrations. Postgate Zero without any terminal electron acceptor was included as a negative control for each experiment, and the final optical density at 600 nm (OD₆₀₀) was recorded as background growth.

RNA isolation.

Total RNA was isolated using the Trizol reagent, as previously described (6). Genomic DNA was removed from the purified RNA using TURBO DNase (Ambion). DNA contamination was monitored by PCR amplification of the D. desulfuricans napC gene with the primers W13-F1 (5′-ATGCACCTTTCATCGCATGCGGC-3′) and W13-GSP2 (5′-ATGACAGGCCCGGCAGTTGG-3′). The concentration and purity of the RNA were determined by measuring the absorbances at 260 and 280 nm.

RACE.

The transcription start site of the nap operon in D. desulfuricans was determined using the 5′ RACE system for the rapid amplification of cDNA ends, version 2.0 (Invitrogen). Briefly, total RNA was isolated from mid-exponential-phase cultures (OD₆₀₀, 0.4) of D. desulfuricans grown in Postgate N with 15 mM NaNO₃. The primers used were W13-P3 (5′-ATCTTACAGGACTTCATGGCCGG-3′), AAP (5′-GGCCACGCGTCGACTAGTAC-3′), W13-GSP2 (5′-ATGACAGGCCCGGCAGTTGG-3′), W13-P5 (5′-ATTGAGAACAGCCGATGGGC-3′), and UAP (5′-CUACUACUACUAGGCCACGCGTCGACTAGTAC-3′).

Reverse transcription-PCR (RT-PCR).

The SuperScript II RNase H reverse transcriptase kit (Invitrogen) was used to reverse transcribe total RNA. RNA (1 to 5 μg) was added to a tube containing random hexamers (50 ng/ml), 1 μl of 10 mM deoxynucleoside triphosphate mix, and RNase-free water to a final volume of 10 μl. A negative control reaction was also prepared by omitting the reverse transcriptase. The cDNA (2.5 μl) was added to 27 μl of sterile distilled water, 5 μl of 10× NH₄ buffer (BioLine), 1.5 μl of MgCl₂, 1 μl of deoxynucleoside triphosphate mix, 5 μl of forward primer specific to the gene of interest, 5 μl of reverse primer specific to the gene of interest, and 0.5 μl of Taq polymerase (BioLine). The products were separated on a 0.8% agarose gel.

Quantitative real-time PCR analysis of gene expression.

The RNA was reverse transcribed to cDNA using a Superscript first-strand synthesis kit (Invitrogen). Transcript levels were measured by quantitative real-time PCR using SensiMix with SYBR green detection (Quantace) and an ABI 7000 sequence analyzer (Applied Biosystems). Primers, designed using PrimerExpress (Applied Biosystems), are described in Table S1 in the supplemental material (available online). Transcript levels were quantified using the ΔΔC_t method relative to expression of the polA gene (19). For each experiment, quantitative real-time PCR was used to determine transcript levels in triplicate on two independent cDNA samples derived from three independent cultures. Error bars show standard deviations.

Rate of nitrate reduction.

The rates of nitrate reduction were measured in real time using a nitrate ion-selective electrode (Sentek, United Kingdom). D. desulfuricans was grown in 1 liter as described previously until the required OD₆₀₀ of 0.8 to 1.0 had been reached (typically after 48 h). After harvesting by centrifugation at 9,000 rpm for 20 min at 4°C was done, the cells were washed with 50 mM phosphate buffer (pH 7.4) and resuspended to a final volume of 5 ml. The bacterial suspension (1 ml) was diluted with 9 ml of 50 mM phosphate buffer (pH 7.4) and incubated in the reaction chamber. The rubber bung containing the nitrate electrode and temperature probe was inserted, and the suspension was allowed to equilibrate to 30°C. Nitrate to a final concentration of 2 mM and 3.6 mM lactate were then added, and the rate of nitrate reduction was measured. At the end of the assay, the optical density at 650 nm of a 1-in-10 dilution of the cell suspension was measured to allow the bacterial dry weight to be determined. An OD₆₅₀ of 1 has been shown experimentally to be equivalent to 0.4 g of bacterial dry mass per liter (17).

Heme staining of SDS-polyacrylamide gel electrophoresis (PAGE) gels.

Total cell protein samples were resuspended in sample buffer, heated at 100°C for 10 min, and separated by 15% polyacrylamide gels containing sodium dodecyl sulfate (SDS). Proteins that contain covalently attached heme groups were detected using a heme-dependent peroxidase activity stain as described previously (33).

Nitrate, nitrite, and sulfate determination.

Nitrate concentrations were determined using the nitrate electrode. The concentration of sulfate was measured using a turbidimetric method based on the reaction of sulfate with barium chloride, resulting in the precipitation of barium sulfate. A sample of culture supernatant (25 to 100 μl, filter sterilized and stored at 4°C) was added to a conditioning agent (0.85 ml glycerol, 0.5 ml concentrated HCl, 1.3 g NaCl, 17 ml ethanol, and distilled water, to a final volume of 1 liter) to a final volume of 1 ml, and 0.1 ml of 1 M barium chloride was added. The solution was then vortexed for 30 s and incubated at room temperature for 45 min to allow the precipitation of barium sulfate. The absorbance of the solution at 420 nm was determined, and the concentration of the samples was calculated from the calibration curve prepared using standard solutions of sodium sulfate.

To determine the presence of nitrite in the culture supernatant, the following solutions were mixed together using a Pasteur pipette: 3 drops of 1% sulfanilamide (920 ml of distilled water, 80 ml of 12.5 N concentrated HCl, 10 g sulfanilamide), 1 drop of 0.02% naphthylethylenediamine (0.2 g N-(1-naphthyl, ethyl-enediamine dihydrochloride, 1 liter distilled water). To the above was added 50 μl of the culture medium. In the presence of nitrite in the culture medium, the solution would change color from colorless to bright pink/purple. If no color had developed, a few grains of zinc powder were added. The development of a purple color around the zinc particles indicated the presence of nitrate in the culture medium.

Detection of NapA by Western blotting.

Proteins separated by SDS-PAGE were transferred electrophoretically using a Bio-Rad Trans-D semidry blotter onto a polyvinylidene difluoride membrane (Millipore) at 0.25 A for 1.5 h in blotting buffer. The membrane was removed and gently agitated overnight in 50 ml of 5% (wt/vol) nonfat dry milk (Bio-Rad) Tris-buffered saline (TBS) within a sealed bag at 4°C. The blocked membrane was washed three times for 5 min each with gentle agitation in TBS-Tween. The membrane was transferred to a plastic bag which contained the primary antibody in 10 ml of 5% (wt/vol) nonfat dry milk TBS-Tween and was incubated with gentle agitation for 2 h at room temperature. The primary antibody used to detect NapA was anti-NapA (rabbit) antibody raised against Escherichia coli NapA and was diluted 1:5,000. The membrane was transferred to a tray and washed three times for 5 min each with gentle agitation in TBS-Tween. The membrane was transferred to a plastic bag containing the antirabbit peroxidase-labeled secondary antibody (ECL Plus Western blotting detection reagents; Amersham Biosciences) diluted 1:5,000 in 10 ml of 5% nonfat dry milk TBS-Tween and incubated with gentle agitation for 1 h at room temperature. The membrane was then transferred to a tray and washed three times for 5 min each with gentle agitation in TBS-Tween. The membrane was then transferred to a plastic bag and treated with 5 ml of solution A and 125 μl of solution B (ECL Plus Western blotting detection reagents; Amersham Biosciences) for 5 min with gentle agitation, enabling chemiluminescent detection of the protein of interest. The membrane was then exposed to film (Hyperfilm ECL) for 2 s to 10 min and developed by autoradiography.

Detection of NapA activity.

Bacteria from 1 liter of culture were resuspended in 5 ml 50 mM Tris-HCl (pH 8.0) and passed through a French pressure cell at 10,000 lb/in². The slurry, consisting of unbroken bacteria, cell membranes, and soluble proteins, was centrifuged at 170,000 × g for 1 h to sediment the membrane-bound proteins, cell debris, and unbroken bacteria. Soluble cytoplasmic and periplasmic proteins in the supernatant were removed by decantation. The resulting membrane-containing fraction was homogenized with 1 ml of 50 mM Tris-HCl (pH 8.0).

The supernatant containing the periplasmic and cytoplasmic proteins was concentrated, and the soluble proteins were separated on native 7.5% polyacrylamide gels. After electrophoresis, gels were placed in a solution of dithionite-reduced methyl viologen containing 20 mM nitrate. Nitrate reductase activity was detected as a colorless band against a dark purple background (32).

RESULTS

Growth of Desulfovibrio desulfuricans with sulfate, nitrate, or nitrite as the only electron acceptor.

In preliminary experiments, bacteria that had been grown in the presence of sulfate as the only terminal electron acceptor were used to inoculate fresh media in which only nitrate or nitrite was present. Because nitrite above 5 mM was toxic, only a limiting concentration, 2 mM, was used in these experiments. This nitrite was reduced within 48 h, resulting in an immediate but short burst of growth, consistent with a previous report that the nitrite reductase genes are expressed constitutively in D. desulfuricans (27). In contrast, in the presence of nitrate there was an initial short burst of growth due to a limited amount of sulfate introduced with the inoculum, followed by a lag that lasted for up to 80 h. Nitrate-dependent growth then continued until all of the available nitrate had been reduced.

In subsequent experiments, bacteria that had been adapted to nitrate-dependent growth were used as the inoculum for cultures in which only nitrate, nitrite, or sulfate was present. Growth in all three cultures started without a lag, and while the growth rate was similar under all three conditions, the growth yield was higher in the presence of nitrate than in that of sulfate (Fig. 1A). This conclusion was confirmed in further experiments in which only limiting concentrations of the electron acceptor was supplied (Fig. 1B). It should be noted that to avoid growth inhibition because of nitrite toxicity in these experiments, nitrite was added sequentially in increments only after the nitrite from each previous addition was reduced completely. The growth yield in the presence of nitrite was 35% of the yield in the presence of nitrate, assuming that nitrate and nitrite were both reduced completely to ammonia, while the growth yield in the presence of sulfate was 45% of that observed in the presence of nitrate as the terminal electron acceptor. Since eight electrons are required to reduce both nitrate to ammonia and sulfate to sulfide, the same conclusion applies on a per-electron basis. However, since only six electrons are required to reduce nitrite to ammonia, the growth yield per electron in the presence of nitrite was 47% of that on nitrate. These experiments established that both nitrate and nitrite could productively support growth of D. desulfuricans even under conditions in which a transfer of contaminating sulfate from an inoculum had been avoided.

FIG. 1.

(A) Growth of nitrate-adapted D. desulfuricans in the presence of sulfate (30 mM), nitrate (15 mM), or nitrite (2 mM) as alternative electron acceptors. (B) Growth yield at the stationary phase of growth of D. desulfuricans in the presence of sulfate, nitrate, and nitrite as alternative electron acceptors. The inoculum was a culture that had been adapted to growth with nitrate as the only electron acceptor. The concentration of the carbon source, lactate, was 48 mM. Due to its toxicity, nitrite was added in sequential aliquots of 3 mM after the previously added nitrite had been reduced. This resulted in considerable variations in yield, as indicated by the larger error bars. The error bars represent standard deviations from the mean values for two replicates from three independent sets of cultures.

Effect of growth conditions on rates of nitrate reduction by bacterial suspensions.

Bacteria were harvested toward the end of exponential growth with nitrate, nitrite, or sulfate as the only electron acceptor, and rates of nitrate reduction in the presence of lactate as the electron donor were determined using a nitrate-specific electrode (Fig. 2A and B). Only the nitrate-grown bacteria reduced nitrate at a significant rate (Fig. 2B), a result that clearly established that the presence of nitrate during growth is required for the D. desulfuricans nap operon to be expressed. To confirm this conclusion, proteins from lysed bacteria were separated by SDS-PAGE, transferred electrophoretically to a polyvinylidene difluoride membrane, and probed with an anti-E. coli NapA antiserum. A strong band of cross-reacting antigen was detected only in the sample of bacteria grown in the presence of nitrate (Fig. 2C). Proteins from similar samples were also separated on a nondenaturing gel. After electrophoresis, the gel was stained with methyl viologen in the presence of sodium dithionite and nitrate to reveal bands of NapA activity. A single band of viologen-oxidizing activity was detected only in the extract from nitrate-grown bacteria (Fig. 2D).

FIG. 2.

Nitrate reductase activity of D. desulfuricans cells grown in the presence of sulfate or nitrite (A) as the terminal electron acceptor or after growth in the presence of nitrate (B) as the terminal electron acceptor. The bacteria were harvested and washed, and the rate of nitrate reduction was determined using the nitrate electrode. The presence of the NapA protein after growth in the presence of sulfate, nitrate, or nitrite (C) was detected using a polyclonal antiserum to E. coli NapA. (D) Nitrate reductase activity of periplasmic proteins from bacteria grown in the presence of sulfate, nitrate, or nitrite. Periplasmic proteins were separated by nondenaturing PAGE and stained for nitrate reductase activity.

The first two genes of the nap operon are predicted to encode c-type cytochromes. We therefore attempted to detect and identify the NapC and NapM cytochromes by staining SDS-polyacrylamide gels for covalently bound heme (Fig. 3A). At least 10 different bands of protein from the sulfate-grown cultures stained for covalently bound heme (Fig. 3A, lane 1), including three weakly staining bands that migrated with apparent molecular masses of about 50 kDa, 26 kDa, and 16 kDa. Based on control experiments that showed the absence of the 26-kDa and 16-kDa bands in D. vulgaris and the extensive literature characterizing the NrfAH complex (25), these cytochromes were tentatively identified as NrfA, NapC, and NapM, respectively. All three of these cytochromes were more abundant after growth in the presence of nitrite (Fig. 3A, lane 3) and much more abundant in nitrate-grown cultures (Fig. 3A, lane 2). Although both NrfA and NrfH are more strongly expressed in nitrate-grown cells, the level of NrfH in nitrite-grown cells was below the detection limit.

FIG. 3.

(A) Detection of c-type cytochromes during growth by use of SDS-PAGE and staining for peroxidase activity of covalently bound heme in cultures grown in the presence of alternative terminal electron acceptors. Lane 1, sulfate-grown cells; lane 2, nitrate-grown cells; lane 3, nitrite-grown cells. Each lane was loaded with total proteins from 120 μg of bacterial dry mass. (B) Heme stain to detect c-type cytochromes induced during growth with constant flushing with 5% CO₂-N₂ of the cultures in the presence of both 5 mM nitrate and 5 mM sulfate as the terminal electron acceptors. The inoculum had been adapted to growth in the presence of sulfate, but essentially identical data were obtained with inocula that had been preadapted to growth with nitrate. Lane 1, whole-cell protein at an OD₆₀₀ of 0.3; lane 2, whole-cell protein at an OD₆₀₀ of 0. 65. Each lane was loaded with total proteins from 120 μg of bacterial dry mass.

Effect of sulfate on nitrate reduction.

Although far more energy is released during the reduction of nitrate than during that of sulfate, previous studies suggested that in the presence of both terminal electron acceptors, D. desulfuricans preferentially uses sulfate. To determine the mechanism of this apparent repression, bacteria were grown in the presence of both sulfate and nitrate and both the biomass and concentration of nitrate in the medium were measured during growth (Fig. 4A). Growth stopped when all of the sulfate had been consumed and no nitrate was reduced. However, growth resumed if the cultures were then flushed with nitrogen to expel hydrogen sulfide that had saturated the culture. Control experiments showed that growth was not inhibited by higher concentrations of sulfide generated by sulfate-sufficient cultures. These results confirmed previous reports that sulfide produced during sulfate reduction inhibits, represses, or both inhibits and represses nitrate reduction (7, 16).

FIG. 4.

(A) Growth of sulfate-adapted D. desulfuricans in the presence of both 5 mM sulfate and 2 mM nitrate as electron acceptors. The culture medium (180 ml) was inoculated with 20 ml of D. desulfuricans that had been grown in the presence of sulfate. The OD₆₀₀ and the concentrations of nitrate and sulfate were measured by removing samples with a syringe during growth. The data are from a single representative experiment. (B and C) Growth with constant flushing with 5% CO₂-95% N₂ of sulfate-adapted (B) or nitrate-adapted (C) D. desulfuricans in the presence of both 5 mM sulfate and 5 mM nitrate as electron acceptors. The data are means of data from three independent growth experiments. The error bars represent standard deviations from the means.

In subsequent experiments, cultures were flushed continuously with nitrogen to prevent the accumulation of hydrogen sulfide, and concentrations of nitrate and sulfate were measured throughout the growth cycle. Analysis of the culture medium revealed slight differences in the utilization of nitrate and sulfate depending upon the source of the inoculum used. When the inoculum had been grown with sulfate, little or no nitrate was reduced until virtually all of the sulfate had been consumed (Fig. 4B). In contrast, bacteria that had been adapted to nitrate-dependent growth reduced the sulfate preferentially but also simultaneously reduced nitrate (Fig. 4C).

Proteins in whole-cell lysates from cultures inoculated with bacteria adapted to sulfate were also separated by SDS-PAGE and stained for covalently bound heme (Fig. 3B). Although the NrfA band was clearly visible even at the early stages of growth (Fig. 3B, lane 1), the concentration of NrfA increased about fivefold as the bacteria adapted to nitrate-dependent growth. Low concentrations of NapC and NapM were also present during the early stages of growth, but the concentrations of both of these cytochromes increased more than 10-fold during subsequent growth.

Transcription start site and cotranscription of napM with napA.

The RACE method was used to identify the napC transcription start site, which was located 112 bases upstream of the first base of the translation start codon (Fig. 5A). Upstream from this deoxyadenosine was a potential −10 sequence, TATAAT, that matched perfectly the E. coli consensus −10 sequence of promoters recognized by the σ⁷⁰ form of RNA polymerase. A −35 sequence was identified 17 bases upstream of the putative −10 sequence. The sequence around the transcription start site was searched for inverted repeat sequences that might be potential binding sites for regulatory proteins. Starting at position +13 is the sequence 5′-AACAAG-3′ separated by two bases from the inverted repeat 5′-CTTGTT-3′: this sequence does not match the consensus sequence for any previously reported transcription factor and was not considered further. However, centered between bases −56 and −57, between bases −75 and −76, and between bases −128 and −129, are three almost perfect inverted repeat sequences separated by 6 bases, two of which are identical at 9 bases out of 10 and the third of which is identical at 8 bases out of 10 to the core of the E. coli consensus sequence recognized by the cAMP receptor protein, CRP (consensus TGTGA-N6-TCACA, where N is any base).

FIG. 5.

(A) Transcription start site of the first gene of the nap gene cluster, napC, as determined by the RACE system (TS, transcription start). The putative −35 and −10 regions are indicated (bold), as is the ribosome binding site (RBS). Analysis of the upstream regulatory region revealed three inverted repeat (IR) sequences that match in 9 out of 10 positions the E. coli consensus sequence binding site for CRP. (B) RT-PCR of the nap gene cluster of D. desulfuricans. L, 1-kb DNA ladder; lane 1, RT-PCR products using primers for the detection of the napA transcript (positive control); lane 2, RT-PCR products using primers napA (no reverse transcriptase; negative control); lane 3, RT-PCR products using primers for the napC transcript; lane 4, RT-PCR products using primers for the detection of the napCMA transcript.

RT-PCR was then used to demonstrate that transcripts initiated at the napC promoter are extended across the napM-napA boundary into the rest of the operon (Fig. 5B). This confirmed that the six nap genes can be cotranscribed as a single operon, though the possibility that there might be a secondary promoter in the 88-bp gap between napM and napA was not excluded.

Nitrate induction of transcription from the napC promoter.

To confirm that the changes in nitrate reductase activity and the abundance of NapC and NapM under different growth conditions were due to regulation of nap operon transcription, quantitative real-time PCR was used to determine the relative quantities of nap operon mRNA present in cultures grown in the presence of sulfate, nitrate, and nitrite. For these experiments, the polA transcript was used as the internal control, and expression of the dsrA (encoding a subunit of the dissimilatory sulfite reductase) and nrfA (encoding the cytochrome c nitrite reductase, NrfA) transcripts was also monitored. First, mRNA was isolated from cultures growing in the presence of only nitrate, sulfate, or nitrite. Levels of the napC transcript were greatly elevated during growth in the presence of nitrate but were not induced by nitrite (Fig. 6A). Only small differences in levels of the dsrA and nrfA transcripts were detected under the three conditions tested, although dsrA mRNA was slightly less abundant during nitrate-dependent growth and nrfA mRNA was slightly elevated during nitrite or nitrate-dependent growth in high nitrate concentrations. These results are in agreement with those of previous microarray studies of the D. vulgaris transcriptome under nitrite stress, showing elevation of nrfA mRNA posttreatment and downregulation of the genes involved in sulfate reduction (10, 11).

FIG. 6.

The effect of sulfate or nitrate on the expression of the napC gene in D. desulfuricans. (A) Quantitative RT-PCR analysis of gene expression of D. desulfuricans grown to an OD₆₀₀ of ∼0.25 in the presence of sulfate, nitrate, or nitrite as the terminal electron acceptor. RNA was isolated, and quantitative RT-PCR was used to quantify dsrA, nrfA, and napC transcripts. Quantities of the transcript level for each gene are normalized against growth in the presence of sulfate. (B) Quantitative RT-PCR analysis of gene expression of sulfate-adapted D. desulfuricans grown in the presence of both sulfate and nitrate as the terminal electron acceptors. RNA was isolated from early (sulfate-reducing stage) and late (nitrate-reducing stage) phases of growth, and quantitative RT-PCR was used to quantify the dsrA, nrfA, and napC transcripts. Quantities of each transcript are normalized against the transcript level in the early phase of growth. Each level of expression shown is the mean for duplicate assays of RNA from three independent growth experiments. The error bars represent standard deviations from the means from the six assays. (C and D) Bacteria were grown in the presence of either sulfate (C) or nitrate (D) to an OD₆₅₀ of ∼0.4, and RNA was isolated from a sample of each culture. The preculture grown in the presence of sulfate (40 ml) was transferred to 60 ml of prewarmed medium containing either sulfate alone (control) or both sulfate and nitrate (C); conversely, the nitrate-adapted culture was transferred to 60 ml of fresh medium containing either nitrate alone or nitrate plus sulfate (D). After growth for a further 2.5 h, RNA was isolated from each culture and the concentration of napC transcript was determined by quantitative RT-PCR. Relative amounts of transcript are normalized against the transcript level before subculture. Each level of expression shown is the average of three independent determinations. The error bars represent standard deviations from the means.

RNA was also isolated during the early and late stages of growth in the presence of both nitrate and sulfate. Only low levels of napC mRNA were present during the early stages of growth, while sulfate was still present in the culture but was much more abundant during the subsequent nitrate-dependent growth (Fig. 6B). As in the cultures with only a single terminal electron acceptor, no significant changes in abundance of dsrA or nrfA mRNA were detected after all of the sulfate had been reduced.

To confirm that sulfate represses nap operon expression, bacteria were adapted either to growth with sulfate in the absence of nitrate or to growth with nitrate alone. Each culture was then transferred into fresh medium containing the original electron acceptor, sulfate or nitrate (controls), or into a mixture of 15 mM nitrate and 30 mM sulfate. RNA was isolated both immediately before and 2.5 h after transfer to the new medium, and levels of napC mRNA were determined. Only very low levels of napC transcript were detected in bacteria grown in the presence of sulfate alone, but nitrate induced napC expression even in the presence of sulfate (Fig. 6C). Conversely, high levels of napC transcript were present in bacteria that had been adapted to growth in the presence of only nitrate, and these levels were sustained on transfer to new medium containing only nitrate. However, napC expression decreased fivefold when bacteria were transferred into medium containing both nitrate and sulfate (Fig. 6D). These data suggest that napC expression is regulated by at least two mechanisms: nitrate induction and sulfate repression.

Further attempts to define a lower limit of the sulfate concentration required to repress nap operon expression revealed up to 80% repression of nap operon expression 30 min after the addition of only 1 mM sulfate, during which time less than 25% of the sulfate had been reduced. Although error bars resulting from combining data from independent experiments designed in this way were large, the data strongly suggest that repression of nitrate reduction by sulfate is physiologically significant under conditions found in natural environments.

DISCUSSION

Relatively little information was available about how transcription is regulated in sulfate-reducing bacteria until the recent availability of microarrays based upon the published genome sequences of Desulfovibrio vulgaris strain Hildenborough (12). This has led to a series of elegant studies, usually using microarrays constructed specifically to study a subgroup of genes (3, 4, 10, 21, 30, 39, 40). Nitrate reduction and its regulation in sulfate-reducing bacteria have been topics of extended controversy, first because nitrate reductase genes are found in only some sulfate-reducing bacteria, second because culture conditions used to test for nitrate reduction lead to the accumulation of sulfide, a potent inhibitor of nitrate reduction, and finally because until recently no direct method was available for measuring levels of the nap operon transcripts in these bacteria. However, like many other sulfate-reducing bacteria, both the Hildenborough and DP4 strains of D. vulgaris, as well as a second closely related sulfate-reducing bacterium, Desulfovibrio strain G20, for which the complete genome sequence is known, are among the many sulfate-reducing bacteria that lack genes for nitrate reduction.

In an earlier study, we proposed that sulfate represses nitrate reduction by D. desulfuricans, but data presented in Fig. 4 show that this might have been due to inhibition of nitrate reduction by sulfide rather than a direct effect of sulfate on nap operon transcription. By exploiting knowledge of the DNA sequence of the nap operon in D. desulfuricans strain 27774 (15), we were able to avoid these problems in the current study. Consequently, the current study is the first in which regulation of nap operon transcription in a sulfate-reducing bacterium has been reported. We established unequivocally that nitrate reduction is induced by nitrate at the level of transcription initiation and that at least the first three of the six nap genes of the nap operon for nitrate reduction are cotranscribed as an operon. There are no significant intergenic regions or candidates for promoter sequences between napA and the last three genes of the nap cluster, and translation initiation codons are located sufficiently close to, or overlap, the termination codons of the preceding gene, typical of an operon organization optimized for translational coupling. This might also indicate that NapA, -D, -G, and -H are synthesized in stoichiometrically equal amounts, but this is not necessarily true. The simplest interpretation of our results is that the six nap genes form a single operon, and the RT-PCR results clearly establish not only that nap transcription is inducible by the substrate, nitrate, but also that it is repressed by the preferred electron acceptor, sulfate. This is contrary to the dogma that bacteria have evolved regulatory systems that enable them to use preferentially the thermodynamically most favorable source of energy for growth.

During nitrate reduction by E. coli, most of the nitrite nitrogen accumulates in the growth medium before it is reimported for reduction to ammonia (5, 9, 26, 28, 37). In contrast, during nitrate reduction by the strain of D. desulfuricans used in the current study, only traces of nitrite accumulate in the medium during growth in the presence of 5 mM nitrate, though there is a transient accumulation of nitrite during growth in the presence of 15 mM nitrate. In contrast to the case with E. coli, in which synthesis of both of the nitrite reductases is induced by nitrite (8, 13, 23, 34, 35), this is achieved in various Desulfovibrio species by the synthesis of a very active, constitutive nitrite reductase, NrfHA, a phenomenon documented in both the present and previous studies (10, 14, 18, 20). Note, however, that although NrfHA is always present even in sulfate-grown bacteria, its synthesis is induced during growth in the presence of nitrite, but only in a minority of strains (10, 17). Whether the minimal accumulation of nitrite during nitrate reduction by sulfate-reducing bacteria is a general phenomenon is unknown, but a relevant factor might be the much greater sensitivity of D. desulfuricans to nitrite toxicity than is the case for strains of E. coli. It is possible that this toxicity would provide a selective pressure for the evolution of regulatory mechanisms that limit nitrite accumulation. However, an alternative explanation is that the constitutive synthesis of a nitrite reductase allows more-effective use of a limited supply of electron acceptors in a carbon-rich environment; but if so, why is the nitrate reductase so clearly inducible?

Possible mechanisms of control of nap gene expression.

The repression of nap operon transcription by the preferred electron acceptor, sulfate, is reminiscent of catabolite repression mediated by inducer exclusion and the CRP-mediated response to cAMP in E. coli in response to the availability of glucose or to the repression of fumarate and dimethyl sulfoxide reductase synthesis by nitrate-activated NarL. In both cases, transcription factors bind specifically to DNA target sequences located close to the transcription start site of the regulated operon. Since in both of these examples the target DNA sequence is an inverted repeat, we searched upstream of the nap operon transcription start site for inverted repeat sequences. Figure 5A shows that there are three inverted repeats in the nap operon regulatory region that closely resemble the consensus binding site for CRP in E. coli. Furthermore, multiple binding sites for CRP and FNR (and also for nitrate-activated NarL) are found in E. coli promoters that under different environmental conditions are either activated or repressed in response to environmental signals. Significantly, identical sequences have previously been noted upstream of D. vulgaris genes implicated in the management of reactive nitrogen species, and these sequences have been proposed as consensus binding sites for regulatory proteins of the CRP-FNR family (24). One might speculate that one Desulfovibrio transcription factor responds to the presence of nitrate to induce nap operon transcription but another responds to sulfate to repress it. It will therefore be fascinating to determine whether genes encoding proteins of the Crp/Fnr family similar to those annotated in the genomes of D. vulgaris strains Hildenborough (12) and DP4 are also present in the genome of D. desulfuricans 27774.