Correspondence: James I. Prosser, School of Biological Sciences, University of Aberdeen, Cruickshank Building, St Machar Drive, Aberdeen AB24 3UU, UK. Tel.: +44 1224 273254; fax: +44 1224 272703; e-mail: email@example.com
Ammonia-oxidizing bacteria are believed to be an important source of the climatically important trace gas nitrous oxide (N2O). The genes for nitrite reductase (nirK) and nitric oxide reductase (norB), putatively responsible for nitrous oxide production, have been identified in several ammonia-oxidizing bacteria, but not in Nitrosospira strains that may dominate ammonia-oxidizing communities in soil. In this study, sequences from nirK and norB genes were detected in several cultured Nitrosospira species and the diversity and phylogeny of these genes were compared with those in other ammoniaoxidizing bacteria and in classical denitrifiers. The nirK and norB gene sequences obtained from Nitrosospira spp. were diverse and appeared to be less conserved than 16S rRNA genes and functional ammonia monooxygenase (amoA) genes. The nirK and norB genes from some Nitrosospira spp. were not phylogenetically distinct from those of denitrifiers, and phylogenetic analysis suggests that the nirK and norB genes in ammonia-oxidizing bacteria have been subject to lateral transfer.
Autotrophic ammonia-oxidizing bacteria (AOB) are responsible for the oxidation of ammonia to nitrite and consequently play an important role in the nitrogen cycles of diverse ecosystems (Prosser, 1989). Ammonia oxidation is accompanied by production of the greenhouse gases nitric oxide (NO) and nitrous oxide (N2O), which are of great environmental concern due to their roles in global warming and destruction of the stratospheric ozone layer (Dickinson & Cicerone, 1986). These gases are produced in soil through several microbial processes, including denitrification, nitrification and nitrifier denitrification, the latter two being carried out by autotrophic AOB (Webster & Hopkins, 1996; Kester et al., 1997; Jiang & Bakken, 1999; Colliver & Stephenson, 2000; Dundee & Hopkins, 2001; Wrage et al., 2001, 2005; Shaw et al., 2006). Nitrifier denitrification, whereby AOB reduce NO2− to N2O and possibly further to N2, appears to be a universal trait in betaproteobacterial ammonia oxidizers (Shaw et al., 2006) and a greater understanding of the factors controlling this process is required for predictions of NO and N2O emission rates.
The probable enzymes involved in nitrifier denitrification are ammonia monooxygenase, hydroxylamine oxidoreductase, nitrite reductase, nitric oxide reductase and nitrous oxide reductase (Wrage et al., 2001). Nitrite reductase reduces nitrite to NO and is encoded by two structurally different genes: nirK, encoding a copper-containing enzyme, and nirS, encoding cytochrome cd1. These genes are functionally and physiologically equivalent and appear to be mutually exclusive (Braker et al., 1998). Nitric oxide reductase is encoded by the norB gene and catalyzes the reduction of NO to N2O. Genes encoding both the small and large subunits of nitric oxide reductase, norC and norB, have been retrieved from denitrifying bacteria (Braker & Tiedje, 2003) and cultured AOB, including Nitrosomonas and Nitrosococcus spp. (Casciotti & Ward, 2005), but homologous sequences have not been obtained from Nitrosospira. The functional genes involved in denitrification have been detected in a variety of phylogenetically unrelated microorganisms, suggesting acquisition by lateral gene transfer, which is supported by phylogenetic analysis of nir and nor gene sequences (Philippot, 2002). Many questions regarding the genes involved in the nitrifier denitrification pathway remain open. Additional nirK and norB sequence information, in particular from the Nitrosospira lineage, can help in understanding the evolution of AOB and whether these functional genes have been acquired or distributed through lateral gene transfer.
The aim of this study was to determine the diversity and relationships of nirK and norB genes in Nitrosospira spp., Nitrosomonas spp. and denitrifying bacteria and to compare the phylogenies of nirK and norB genes in AOB with those based on ammonia monooxygenase (amoA) and 16S rRNA genes.
Material and methods
Bacterial isolates and growth conditions
The betaproteobacterial AOB strains studied (Table 1) were grown in static liquid batch culture at 28oC, as described by Keen & Prosser (1987) in modified Skinner and Walker medium (Powell & Prosser, 1986). All cultures were checked for heterotrophic bacterial contamination by plating on nutrient agar and by PCR amplification and denaturing gradient gel electrophoresis analysis.
Table 1. Strains of ammonia oxidizing bacteria used in this study
Ammonia oxidizer strains
Nitrosospira briensis strain128
Nitrosospira lineage, cluster 3
Nitrosospira multiformis ATCC25196
Nitrosospira lineage, cluster 3
Soil, Surinam, South America
Nitrosospira tenuis strain NV12
Nitrosospira lineage, cluster 3
Nitrosospira sp. strain 40KI
Nitrosospira lineage, cluster 4
Clay loam soil (pH 6.5), Norway
Nitrosospira sp. strain B6
Nitrosospira lineage, cluster 2
Pebbles (pH 6.3), nitrifying reactor, Norway
Nitrosospira sp. strain En13
Nitrosospira lineage, cluster 3
Craibstone soil (pH 4.5), Aberdeen, UK
Nitrosospira sp. strain NpAV
Nitrosospira lineage, cluster 3
Nitrosospira sp. strain AF
Nitrosospira lineage, cluster 3
Sandi soil (pH 4), Zambia
Nitrosomonas europaea ATCC19718
Nitrosomonas europaea lineage
Cell lysis and PCR amplification
Approximately 1 mL of a stationary-phase AOB culture was centrifuged at 13 000 g for 5 min. The pellet was resuspended in 50 μL of 5% sterile Chelex® 100 Resin (BioRad, Hemel Hempstead, Hertfordshire, UK) by light vortexing, boiled (100°C for 10 min) and frozen (−20°C for 10 min) and thawed twice. After centrifugation for 2 min at 13 000 g, 1 μL of supernatant was used for PCR amplification. The PCR primers are presented in Table 2, and PCR amplification was carried out with 50 μL reaction mixtures containing 1.25 U of Taq polymerase (BIOTAQ™ DNA Polymerase, Bioline, London, UK) with 1 × manufacturer's reaction buffer, 1.5 mM MgCl2, 200 μM of each deoxynucleotide triphosphate (Bioline), 50 pmol of each primer and 400 ng of bovine serum albumin.
The program used to amplify the norB gene (norB1f-norB8r primer set) was: denaturation at 94°C for 4 min, 30 cycles at 94°C for 30 s; 55°C for 45 s, 72°C for 45 s and a final extension at 72°C for 7 min. For amplification of the nirK gene with the primer set F1aCu-R3Cu, the thermal cycling conditions were: denaturation at 94°C for 3 min, 35 cycles at 94°C for 30 s, 57°C for 1 min, 72°C for 1 min and a final extension at 72°C for 7 min. The nirK gene was amplified with primer sets nirK1F-nirK5R and Cunir3-Cunir4 using the PCR conditions described by Braker et al. (1998) and Casciotti & Ward (2001), respectively. PCR products were analyzed by electrophoresis of 5 μL aliquots of the reaction mixtures on 1.0% agarose gel using Bioline Hyperladder DNA fragment size markers (Bioline).
Cloning and sequencing
PCR products were purified by precipitation with 50 μL of 20% PEG/2.5 M NaCl (incubated for 15 min at 37°C), washed twice with 125 μL ice-cold 80% ethanol, cloned using the pGEM-T Easy vector system (Promega Ltd, Southampton, UK), according to the manufacturer's instructions, and transformed into XL-Blue MRF Kan supercompetent Escherichia coli cells (Stratagene Inc., Cambridge, UK). White colonies were selected for amplification with vector primers M13f-M13r (Promega), and PCR products were purified as described above. Clones with the correct insert (as judged by size) were subjected to sequencing with universal SP6 or T7 vector primers (Promega) with a PE Biosystems 377 DNA sequencer, using the sequencing service of the University of Dundee.
The sequences obtained were compared against database sequences using tblastx provided by NCBI (http://www.ncbi.nlm.nih.gov) and were aligned and clustered using clustal-w. Maximum likelihood analyses were performed on this alignment using tree-puzzle (version 5.2) with the JTT model of substitution and gamma distribution for rate of heterogeneity among sites. Phylogenetic trees of nirK and norB genes were constructed from the same aligned amino acid sequences using seqboot, prodist, kitsch and consense in the phylip programs and the JTT rate matrix, and were visualized using treeview software, version 1.6.6. Phylogenetic trees comparing 16S rRNA, amoA, nirK and norB genes were constructed with treecon for Windows (version 1.3b Yves van de Peer, Ghent, Belgium) with the neighbor-joining algorithm using the BLOSUM 62 substitution matrix.
Sequencing and phylogenetic analysis of nirK genes
NirK genes were amplified using three primer sets (Table 2). Amplification with the primer set Cunir3-Cunir4 (Casciotti & Ward, 2001) generated multiple, nonspecific PCR products with all the tested AOB strains. Primer set nirK1-nirK5R, which is commonly used to survey nirK gene diversity in environmental samples (Braker et al., 1998; Thröback et al., 2004), generated products from all AOB strains tested but sequences showed no significant similarity to database nirK genes. The closest relationships were 59% similarity to a conserved hypothetical protein, 61% to methylenetetrahydrofolate reductase and 57% similarity to the amoC gene. The most successful amplification of nirK genes was obtained with the F1aCu-R3Cu primer set (Hallin & Lindgren, 1999), which generated sequences from Nitrosospira strains with 91–96% similarity to nirK genes from Nitrosomonas strains deposited in the NSBI database by Casciotti & Ward (2001).
Only sequences obtained from PCR products amplified with the F1aCu-R3Cu primer set were used for further phylogenetic study, and analysis of predicted amino acid sequences of 14 AOB and seven representatives of classical denitrifiers yielded the tree in Fig. 1. This analysis grouped four Nitrosospira strains (Nitrosospira sp. AF; Nitrosospira sp. En13; Nitrosospira sp. NpAV and Nitrosospira briensis) in a distinct cluster (I), supported by high bootstrap values (>90), together with five Nitrosomonas spp. obtained previously by Casciotti & Ward (2001). Interestingly, the nirK sequences from Nitrosospira sp. 40KI, Nitrosospira sp. B6 and Nitrosospira tenuis NV12 grouped with Nitrosomonas sp. TA-921i-NH4, and fell within the denitrifier cluster (II) with well-supported bootstrap values. This indicates that nirK sequences from some AOB are not phylogenetically distinct from those of denitrifiers. In contrast, nirK sequences from Nitrosospira multiformis were phylogenetically distinct from all other AOB and did not fall within the two main clusters. The nirK sequence from Nitrosomonas europaea was different from all other nirK sequences, fell outside the main clusters and was not included in this phylogenetic tree.
Sequencing and phylogenetic analysis of the norB gene
All tested AOB, except Nitrosospira multiformis and Nitrosospira sp. B6, yielded norB PCR products of the expected size of c. 669 bp. Two isolates, Nitrosospira tenuis and Nitrosospira briensis, yielded an additional nonspecific PCR band. tblastx analysis revealed that sequences obtained from Nitrosospira strains had 73–86% similarity to the norB genes from a variety of denitrifying bacteria. For example, sequences from Nitrosospira sp. AF showed 77% similarity to the norB genes from Sinorhizobium meliloti and Silicibacter pomeroyi; Nitrosospira briensis showed 78% similarity to norB gene from Roseobacter denitrificans; and Nitrosospira sp. NpAV showed 73% similarity to norB genes from Paracoccus halodenitrificans. Interestingly, the norB genes amplified from Nitrosospira strains showed a higher similarity to norB genes from classical denitrifiers than to norB genes from Nitrosomonas strains available in the NCBI database, suggesting different evolutionary origins for norB genes of these two genera.
Phylogenetic analysis of norB-deduced, 116-amino acid sequences from eight AOB strains obtained in this study and 16 norB sequences selected from the NCBI database generated the tree presented in Fig. 2. While the norB sequences from Nitrosomonas spp. grouped together in one cluster (II), separate from denitrifier sequences, the Nitrosospira sequences obtained in this study mainly grouped with cluster (I) denitrifiers. The norB sequences obtained from Nitrosospira spp. AF, En13, NpAV, Nitrosospira tenuis and Nitrososipra briensis grouped in cluster (I) with alphaproteobacterial denitrifiers, such as Roseobacter denitrificans, Paracoccus denitrificans, Rhodopseudomonas palustris and Sinorhizobium meliloti. This cluster formed two subclusters, containing norB sequences from AOB and denitrifiers, respectively, each with well-supported bootstrap values. NorB gene sequences of Nitrosospira multiformis (obtained from NCBI as no PCR product was obtained), were phylogenetically distinct from all other AOB, as for nirK genes. One norB sequence obtained from Nitrosospira spp. 40KI grouped with Silicibacter pomeroyi with high bootstrap support and formed a separate cluster (III) with norB genes from Alcaligenes faecalis and Nitrosospira multiformis.
Braker & Tiedje (2003) showed that norB genes from traditional denitrifiers did not cluster strictly according to the 16S rRNA gene-based phylogeny or to nitrite reductase genes, and an incongruence between norB and 16S rRNA gene phylogenies has also been reported by Casciotti & Ward (2005) and Philippot (2002).
Comparative phylogenetic analysis of 16S rRNA, amoA, nirK and norB genes
The phylogenies of nirK and norB genes obtained from AOB in this study were compared with those based on ammonia monooxygenase (amoA) and 16S rRNA genes (Fig. 3). For each phylogenetic tree, corresponding sequences from Nitrosomonas europaea were used as outgroups, as this species differs significantly from all other AOB (Casciotti & Ward, 2001). The phylogeny of AOB based on 16S rRNA gene sequences is generally believed to reflect the evolutionary history of AOB well, as it is highly consistent with species identities (Aakra et al., 2001). The phylogenies of 16S rRNA genes and amoA genes were very similar, but differed greatly from those generated using nirK and norB gene sequences (Fig. 3). For example, phylogenetic analysis based on the nirK gene clearly differentiated Nitrosospira tenuis and Nitrosospira briensis, which were phylogenetically very close in terms of 16S rRNA, amoA and norB genes. Phylogenetic analysis based on the norB gene of Nitrosospira sp. 40KI was very different from that of all other Nitrosospira spp. included in this study, and from 16S rRNA, amoA and nirK gene phylogenies. Overall, the functional nirK and norB genes appeared to be less well conserved than the 16S rRNA genes and functional amoA genes.
Phylogenetic analyses based on 16S rRNA, amoA, nirK and norB genes revealed the phylogeny of Nitrosospira multiformis to be very different from all other AOB examined here, despite belonging to the same genera and performing the same function in soils. The nitrite and nitric oxide reductases are traditionally thought to be genetically linked (Zumft, 1997). However, our results indicate that the norB and nirK gene phylogenies for AOB are not congruent. The results obtained here therefore provide strong evidence that denitrification genes were acquired or distributed through lateral gene transfer. They suggest that the nirK and norB genes, putatively involved in the nitrifier denitrification pathway, are commonly present in AOB, thereby supporting the findings of Beaumont et al. (2002) and Casciotti & Ward (2001; 2005).
All of the Nitrosospira strains used in this work were able to reduce nitrite to N2O and carry out nitrifier denitrification as shown previously by Shaw et al. (2006). The rate of N2O production varied between strains, but there was no correlation between production rate and norB and nirK gene diversity. Moreover, variation in N2O production may be due to variation in cell concentrations and/or the activity of different batches of cells (Shaw et al., 2006). To date, nirK and norB genes have been studied predominantly in heterotrophic denitrifying bacteria, where they are involved in the production of NO and N2O. The presence of the fragments of nirK and norB genes in cultured AOB does not necessarily confirm that they are functional, and the mechanisms of denitrification in AOB are still unclear (Beaumont et al., 2002; Casciotti & Ward, 2005). Schmidt et al. (2004) postulated a key role of nirK and norB genes in nitrifier denitrification as Nitrosomonas europaea ATCC 19718 nirK- and norB-deficient mutants were found to be incapable of nitrogen gas production. However, physiological evidence suggests that these bacteria might recruit nirK as a protection against nitrite toxicity resulting from ammonia oxidation, rather than functioning in a respiratory pathway (Stein & Arp, 1998; Beaumont et al., 2002).
More information regarding the mechanism and regulation of genes involved in the nitrifier denitrification pathway can be obtained from techniques such as whole-genome microarray (Cho et al., 2006). However, attempts to understand the genetics of nitrifier denitrification have included only one representative of AOB, Nitrosomonas europaea, which appears to be different from all other AOB, based on both structural and functional genes. In addition, the significance of the nirK and norB sequences obtained from Nitrosospira spp. for NO and N2O production during nitrifier denitrification requires further investigation.
This work was funded by a Biotechnology and Biological Sciences Research Council, UK, grant (32/D19035) awarded to EMB and JIP. We thank Zena Smith for providing the Nitrosospira cultures.