INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Denitrification is a dissimilatory process of bacteria in which oxidized nitrogen compounds are used as alternative electron acceptors for energy production. The gaseous end products NO, N₂O, and N₂ are released concomitantly. In the environment, denitrification is responsible for the release of fixed nitrogen into the atmosphere in form of N₂ (13). It causes major nitrogen losses in agricultural soils to which fertilizers are applied. Accumulation of the greenhouse gases NO and N₂O leads to the destruction of the ozone layer (3, 13). Also, denitrifying bacteria cause the removal of nitrogen compounds from waste water, where denitrification is coupled to the nitrification process (13). Bioremediation of environmental pollutants can be achieved under denitrifying conditions (5, 10, 33).

Denitrifying bacteria are phylogenetically diverse. They belong to all major physiological groups except for the Enterobacteriaceae, obligate anaerobes, and gram-positive bacteria other than Bacillus spp. (34). Defined as a physiological group, these facultative anaerobes can switch from oxygen to nitrogen oxides as terminal electron acceptors when kept under anoxic conditions. Nitrite reductase is the key enzyme in the dissimilatory denitrification process. The reduction of nitrite to NO can be catalyzed by the products of two different nitrite reductase genes: one product contains copper (the nirK product), and the other contains cytochrome cd₁ (the nirS product). The two genes seem to occur mutually exclusively in a given strain, but both types have been found in different strains of the same species (4). Although structurally different, both enzyme types are functionally and physiologically equivalent (9, 35). nirS is more widely distributed; nirK is found in only 30% of the denitrifiers studied so far. However, nirK is found in a wider range of physiological groups (4). Several different approaches were used to determine the type of nitrite reductase in laboratory pure cultures. Diethyldithiocarbamate has been used to identify nirK-containing denitrifiers (21). Very specific detection, mostly at the strain level, could be achieved with antisera against dissimilatory nitrite reductase (dNirS [4, 24, 29] and dNirK [4, 17]). Another approach was the use of gene probes for nirK (12, 32) or nirS (12, 15, 24, 29), which were generally specific for the strains investigated. Weak reactivity also occurred for the nirK gene probe with DNA from some of the other nir-type denitrifiers (32); the nirS probe, on the other hand, hybridized with a more limited variety of strains (24, 30). A PCR method with one primer pair to target the nirS nitrite reductase gene showed higher specificity than hybridization experiments (30).

In the present study, we report on the application of new primer systems for both types of nitrite reductase genes. We used several different primer pairs to determine the nir type of denitrifying strains. Using samples from aquatic habitats, we amplified nir fragments and used the most reliable primer pairs for nirK or nirS, respectively, to successfully detect, in these aquatic samples, different populations of denitrifying bacteria.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacteria and growth conditions. A variety of denitrifying and nondenitrifying bacterial strains (see Tables 2 and 3) were used to evaluate the specificity of designed PCR primers. All strains were grown aerobically at 27°C. For genomic DNA isolation, Pseudomonas, Alcaligenes, Ochrobactrum, Paracoccus, and Azospirillum strains and the denitrifying isolate IFAM 3698 were grown on nutrient broth (NB; Merck, Darmstadt, Germany). Rhizobium strains were grown on yeast extract medium (YEM [27]). Hyphomicrobium zavarzinii IFAM ZV-622^T was grown on 337-B1 medium (7) with 0.5% (vol/vol) methanol. Rhodobacter sphaeroides f. sp. denitrificans was grown on trypticase soy broth (TSB; Difco Laboratories, Detroit, Mich.), Roseobacter denitrificans was grown on oligotrophic medium (PYGV [25]) supplemented with 25 artificial seawater (16), and Blastobacter denitrificans was grown on peptone yeast extract glucose medium, i.e., PYGV without vitamins. Nondenitrifying strains of the Enterobacteriaceae were grown on Luria broth (LB [19]).

Extraction of genomic DNA. Genomic DNA was obtained from pure cultures by lysozyme-proteinase K-sodium dodecyl sulfate (SDS) treatment followed by phenol-chloroform extraction and subsequent ethanol precipitation (8). The purity and concentration of the DNA preparations were determined spectrophotometrically.

Preparation of total DNA from an enrichment culture and four environmental samples. DNA was prepared from five samples. (i) A 500-ml volume of medium 337-B1 with 0.5% (wt/vol) KNO₃ for the enrichment of denitrifying methylotrophic bacteria was inoculated with 100 µl of activated sludge from a sewage treatment plant near Plön (Schleswig-Holstein, Germany). After 4 weeks at 28°C under anaerobic conditions, 500 µl of the enrichment was again inoculated and kept under the same growth conditions. Six months later, cells were harvested by centrifugation (6,000 × g for 60 min at 4°C) and resuspended in 400 µl of double-distilled water. The DNA was extracted with Chelex 100 (28).

1

PCR amplification of the nir genes. PCR amplifications from pure cultures and environmental samples were performed in a total volume of 50 µl containing 5 µl of 10× PCR buffer (500 mM KCl, 25 mM MgCl₂, 200 mM Tris-HCl [pH 8.4], 0.1% Triton X-100), 200 µM each deoxyribonucleoside triphosphate, 1.0 U of Taq polymerase (5 U µl¹; Appligene Oncor, Illkirch, France), 25 pmol (for genomic DNA) or 35 pmol (for total DNA from environmental samples) of both primers (5 pmol µl¹ each), and DNA (10 to 100 ng). After a denaturation step of 5 min at 95°C, a "touchdown" PCR was performed (Thermocycler 2400; Perkin-Elmer, Branchburg, N.J.). This consisted of a denaturation step of 30 s at 95°C, a primer-annealing step of 40 s, and an extension step of 40 s at 72°C. After 30 cycles, a final 7-min incubation at 72°C was performed. During the first 10 cycles, the annealing temperature was decreased by 0.5°C every cycle, starting at 45°C until it reached a touchdown at 40°C. The additional 20 cycles were performed at an annealing temperature of 43°C. The amplification products were analyzed by electrophoresis on 2% (wt/vol) agarose gels (Boehringer, Ingelheim, Germany) followed by a 15-min staining with ethidium bromide (0.5 mg liter¹).

Sequencing of amplified nir products. For DNA sequencing, amplified PCR products from pure cultures were purified with the QIAquick PCR purification kit (Qiagen, Hilden, Germany) as specified by the manufacturer. DNA sequences were determined by direct sequencing of purified PCR products with the cycle-sequencing kit (GATC, Konstanz, Germany) and Thermosequenase 2.0 (Amersham, Braunschweig, Germany) as specified by the manufacturers. Labeling was performed by terminating the polymerization with biotin-labeled dideoxynucleoside triphosphates. After a denaturing step of 4 min at 94°C, 30 cycles of denaturation for 30 s at 94°C and primer annealing and extension for 30 s at 55°C were performed, followed by an additional extension step of 6 min at 60°C. The sequencing products were blotted with a direct blotting apparatus (GATC) onto a nylon membrane. The separated products were visualized by an enzyme-linked streptavidin-biotin coupling assay with a streptavidin-alkaline phosphatase conjugate (GATC) and NBT/X-phosphate (Boehringer, Mannheim, Germany) as specified by the manufacturers. The sequences obtained were compared with published nirK and nirS sequences in the EMBL Nucleotide Sequence Database by FASTA analysis of the HUSAR program package based on the Genetics Computer Group sequence analysis package (6).

Hybridization analysis of nir products from total DNA of environmental samples. Approximately 100 ng (pure cultures) or 250 ng (environmental samples) of PCR product was analyzed on an agarose gel (2%, wt/vol). After electrophoresis, the DNA was transferred onto a positively charged nylon membrane (QIAbrane Nylon Plus; Qiagen) by capillary transfer (24). The DNA was cross-linked to the membrane with UV light (45 s at 302 nm).

1

Nucleotide sequence accession numbers. The nirK and nirS sequences have been deposited in the EMBL nucleotide sequence databases under accession no. AJ224902 through AJ224913.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Design of PCR primers. For the nirK gene, six sequences were available from the EMBL database, i.e., from Alcaligenes faecalis S-6 (D13155), Achromobacter cycloclastes (Z48635), Pseudomonas aureofaciens (Z21945), Pseudomonas sp. strain G-179 (M97294), Rhodobacter sphaeroides (U62291), and Rhizobium "hedysari" (U65658). For the nirS gene, six sequences were available, i.e., from Pseudomonas stutzeri JM300 (M80653), Pseudomonas stutzeri (X56813), Pseudomonas aeruginosa (X16452), Paracoccus denitrificans Pd1222 (U05002), Paracoccus denitrificans LMD29.63 (U75413), and Alcaligenes eutrophus H16 (X91394). For each gene, the available sequences were aligned by using the MULTIALIGN program (6). Five conserved regions for nirK and six conserved regions for nirS were chosen to design the primers used in this study. The sequences of the primers are shown in Table 1, and their locations within the nir genes are shown in Table 1 and Fig. 1. Comparison of the chosen primer sequences to all stored sequences in the EMBL database by using the BLASTN program (6) revealed significant similarity only to sequences of the nirK or the nirS gene.

View larger version (18K): [in this window] [in a new window]   FIG. 1.   Diagram showing the positions of primers used to amplify nir fragments: the positions for the nirK primers are correlated with the sequence of the nirK gene from Alcaligenes faecalis S-6 (top), and the positions for the nirS primers are correlated with the sequence of the nirS gene from Pseudomonas stutzeri ZoBell (bottom).

Amplification of nirK and nirS fragments from pure cultures. The selected primers were used to amplify nirK or nirS fragments from bacterial pure cultures known to denitrify and to contain either nirK or nirS. For nirK, amplification products were obtained with the primer combinations nirK1F-nirK3R (392 bp), nirK1F-nirK4R (433 bp), nirK1F-nirK5R (514 bp), nirK2F-nirK3R (353 bp), nirK2F-nirK4R (394 bp), and nirK2F-nirK5R (475 bp). For nirS, amplification products were obtained with the primer combinations nirS1F-nirS3R (256 bp), nirS1F-nirS5R (751 bp), nirS1F-nirS6R (890 bp), nirS2F-nirS3R (164 bp), nirS2F-nirS5R (659 bp), nirS2F-nirS6R (798 bp), nirS3F-nirS5R (512 bp), nirS3F-nirS6R (651 bp), nirS4F-nirS5R (197 bp), and nirS4F-nirS6R (336 bp). Products of the expected size could be obtained from all selected primer combinations for amplification of nirK fragments with genomic DNA from Ochrobactrum anthropi and for amplification of nirS fragments with genomic DNA from Pseudomonas stutzeri. With the other pure cultures used in this study, not all but several primer combinations generated products of the expected sizes when used in PCR amplification experiments (Tables 2 and 3).

Determination of the type of nir gene of denitrifying bacteria. Five strains of denitrifying bacteria were tested to determine their type of nitrite reductase. Genomic DNA preparations from these strains reacted with either the nirK or the nirS primer combination in PCR amplification experiments. Blastobacter denitrificans, Alcaligenes xylosoxidans, and Alcaligenes sp. were determined to have the coding sequence for the nirK nitrite reductase as deduced from their amplification products with the nirK primer combinations (Table 2). The coding sequence for the nirS nitrite reductase was identified within genomic DNA from Alcaligenes eutrophus DSM 530 and the denitrifying isolate IFAM 3698 as deduced from their amplification products with the nirS primer combinations (Table 3).

Sequencing of nir gene fragments. To determine the specificity of the amplified fragments, the PCR products obtained with the primer combinations nirK1F-nirK5R and nirS1F-nirS6R from the pure cultures were partially sequenced. A total of 425 bp of the amplified fragments from nirK were sequenced; the amplified and sequenced regions for nirS were between 774 and 792 bp long. Comparison to all sequences in the EMBL database by using the FASTA program revealed that all sequences obtained were parts of the nirK or the nirS gene, respectively. The levels of homology of the fragments compared to the sequences of the nirK and nirS genes from the EMBL database are summarized in Tables 4 and 5, respectively. The calculations were done with the GAP program for the nucleic acid similarity and the TREE program for the deduced amino acid similarity.

Amplification of nitrite reductase genes from environmental samples. The development of a system of generally amplifying primer combinations for each gene permitted the qualitative detection of denitrifying bacteria in environmental samples. For these experiments, nirK1F-nirK5R or nirS1F-nirS6R were applied to total DNA preparations from an enrichment culture for denitrifying methylotrophic bacteria, from activated sludge, from the water of Lake Kleiner Plöner See, and Lake Plussee, and from sediment of Lake Kleiner Plöner See. For nirK, a weak product of the expected size (514 bp), as well as a smear of nonspecific products, was obtained from the DNA preparations from the water of Lake Kleiner Plöner See and Lake Plussee and from sediment of Lake Kleiner Plöner See. The weak band of 514 bp was extracted and reamplified at annealing temperatures of 65 to 60°C by applying the same primer combination. As a result, distinct bands of 514 bp were obtained (Fig. 2, lanes 2 to 6). For nirS, products of the expected size (about 890 bp) were generated (Fig. 3, lanes 2 to 6). The sizes of the products obtained from the environmental samples were the same as those of the positive controls for nirK of Alcaligenes xylosoxidans (Fig. 2, lane 7) and for nirS of Pseudomonas stutzeri (Fig. 3, lane 7).

View larger version (63K): [in this window] [in a new window]   FIG. 2.   Southern blot hybridization of nirK fragments obtained from environmental samples with the primer combination nirK1F-nirK5R to the digoxigenin-labeled fragment from Alcaligenes xylosoxidans. Lanes: 1 and 8, digoxigenin-labeled DNA size standard VII (Boehringer); 2, sediment from Lake Kleiner Plöner See (5 µl of PCR product), 3, water from Lake Plussee (7 µl); 4, water from Lake Kleiner Plöner See (7 µl); 5, activated sludge from the sewage treatment plant at Plön (10 µl); 6, enrichment culture for denitrifying methylotrophic bacteria (12 µl); 7, A. xylosoxidans positive control (5 µl).

View larger version (59K): [in this window] [in a new window]   FIG. 3.   Southern blot hybridization of nirS fragments obtained from environmental samples with the primer combination nirS1F-nirS6R to the digoxigenin-labeled fragment from Pseudomonas stutzeri ZoBell. Lanes: 1 and 8, digoxigenin-labeled DNA size standard VII (Boehringer); 2, sediment from Lake Kleiner Plöner See (15 µl), 3, water from Lake Plussee (12 µl); 4, water from Lake Kleiner Plöner See (20 µl); 5, activated sludge from the sewage treatment plant at Plön (10 µl); 6, enrichment culture for denitrifying methylotrophic bacteria (20 µl); 7, P. stutzeri ZoBell positive control (1 µl).

Hybridization analysis with probes for nirK and nirS. To confirm the specificity of these amplified nirK and nirS fragments, hybridization experiments were performed. The DIG-labelled PCR product from Alcaligenes xylosoxidans (nirK), obtained with the primer combination nirK1F-nirK5R, and from Pseudomonas stutzeri (nirS), obtained with the primer combination nirS1F-nirS6R, were used as probes. All the amplified fragments hybridized, and the products of the positive controls reacted with the specific probe (Fig. 2 and 3). Previous results had shown that the probes did not hybridize with negative controls (PCR without template), with nonspecific PCR products, or nonlabelled DNA standard (data not shown).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The genetic diversity of denitrifying bacteria in environmental samples can be investigated by different molecular methods. We describe herein the first steps required to detect denitrifying bacteria in aquatic habitats by the use of two distinct PCR systems for the nitrite reductase genes, nirK and nirS. Since denitrifiers are not defined by close phylogenetic relationship, an approach involving 16S rRNA molecules is not suitable for general detection of this physiological group in the environment. The use of rRNA-targeted probes has been successfully applied so far only for strains and specific groups to explore the denitrifying community of activated sludge (18).

A more general approach to the detection of all denitrifying bacteria in environmental samples could be the use of a physiological gene or of an enzyme as a molecular marker. For this purpose, nitrite reductase and its genes have been used by several authors, since this is the key enzyme in the denitrification process. Antisera against the dissimilatory nitrite reductase (dNir) from Pseudomonas stutzeri ATCC 14405 were highly specific and reacted with the immunizing strain and few other closely related bacteria (14, 29). Less specific reactions could be obtained with combinations of antisera against heme-type dNirs from P. stutzeri JM300 and P. aeruginosa (4). When a variety of approximately 150 denitrifying strains of uncharacterized dNir type were screened with this combination and an antiserum against Cu dNir from Alcaligenes cycloclastes, 90% of the strains could be identified as possessing either the heme-type or Cu dNir (4). Due to the inducible nature of the enzyme, antisera could be useful in detecting conditions of active denitrification (29). In contrast, approaches targeting the nirK or nirS gene would detect the denitrifiers irrespective of the denitrifying conditions. Compared to the antisera, a broader response to different pure cultures possessing the nirS gene was achieved by use of the specific gene probes (15, 24). By using hybridization with a gene probe for nirK, this type of nitrite reductase could be always confirmed in pure cultures (32). In enrichment cultures, different populations of denitrifiers could be detected by restriction enzyme-digested preparations (HindIII) of total DNA with a probe for nirS (23).

PCR amplification systems, on the other hand, are limited neither to actively denitrifying cells nor to cultivated strains. Application to pure cultures of a primer pair derived from three nirS sequences flanking a conserved central region of the nirS gene revealed a reactivity broader than that with the use of antisera but not as satisfactory as that with the use of gene probes (30). A strain-specific reaction such as that obtained with the use of antisera was not achieved, although amplification failed with some nirS-containing strains as detected by hybridization. This may be because for PCR amplification only the homology of the primer hybridization region is decisive whereas hybridization of gene probes can be detected if any region of the probe shows sufficient homology. In the present study, a PCR system based on six sequences each for nirK or nirS available from the database promised an even more general approach. Regions that are conserved for both genes could not be found, because the enzymes are structurally different. When the sequences for each of the genes were aligned separately, conserved regions for each became evident. Degenerated primers that flanked regions at the more highly conserved C terminus of the two genes were designed. Primers were not positioned within the nirS region that codes for heme binding, because this is homologous to highly conserved regions in other heme-binding proteins (11). By using different combinations of primers and low-temperature stringency conditions in the PCR assays, amplification of nir fragments from both genes was possible for all denitrifying strains tested. The specificity of the PCR procedure was confirmed, since specific products were not obtained when the primer combinations were used with nondenitrifying strains that could perform assimilatory respiration of NO₃ or with strains possessing the gene coding for the other nir type. This is consistent with the finding that the two genes are mutually exclusive in a given strain (4). Combinations containing nirK4R or those containing nirS5R resulted in the smallest number of specific products; this may be due to the degree of conservation in the region where the primer should hybridize. From every pure culture tested, at least two different primer combinations were applied successfully for nirK and at least three were applied successfully for nirS. The specificity of amplifications was confirmed by sequencing the largest products. Generally, sequencing revealed that the products from the pure cultures with the primer pairs nirK1F-nirK5R and nirS1F-nirS6R were specific fragments of the genes coding for copper-containing and cytochrome cd₁-containing nitrite reductase, respectively. Calculation of the homology revealed that with these primer pairs, PCR products could be obtained from genes showing homology as low as 65.5% for nirK and 59.5% for nirS to the sequences available for the design of the primers. This indicates that these PCR systems could be reasonable means for a more general detection of denitrifying bacteria.

The fragments of nirS investigated here were more heterogeneous than were those of nirK. In-frame deletions or insertions of up to 18 bp could be observed within the sequenced region of nirS. Since these results are not in agreement with the findings that the Cu dNirs were more heterogeneous than the cd₁ dNirs (4), the different molecular weights of the Cu dNir subunits may be a result of processing of the enzyme.

The degrees of conservation of the nir genes are very variable. The nirK fragments from eight strains of Ochrobactrum anthropi were 98.8 to 100.0% homologous (data not shown). On the other hand, the sequence of the fragment from two of these strains was 100% homologous to that of the same fragment from Alcaligenes faecalis S-6 (data not shown). However, both types of nir genes are distributed among closely related Pseudomonas (RNA group I [35]) and Alcaligenes (4) species. Even among strains of Alcaligenes faecalis, both types could be detected. The distribution of nir genes among denitrifying bacteria could be explained in different ways. The occurrence of the same nir type among phylogenetically different groups (34) might be caused by a common denitrifying ancestor. During evolution, the ability to denitrify may have been lost in some branches, resulting in closely related nondenitrifying and denitrifying strains (2). The occurrence of different nir types among the same species could be an indication of a horizontal gene transfer (4).

The PCR systems for the nirK and the nirS genes for nitrite reductase, using one generally amplifying primer combination for each nir gene, could be applied successfully to detect populations of denitrifying bacteria in aquatic systems. Due to higher cell densities, detection of such populations from soils by using these PCR systems should be possible as well.

nirS fragments could be amplified directly from DNA preparations of environmental samples, whereas for nirK a reamplification step was necessary when total DNA from water or sediment of lakes was the target for amplification. This might be in agreement with a numerical preponderance of heme type nitrite reductase, as reported for the distribution of nir types in numerically dominant isolates from soils (4). Further investigations will be necessary to determine the abundance of nirK and nirS type denitrifiers and to obtain more information on the genetic diversity of denitrifying bacteria in the environment.