Introduction

Ammonia-oxidizing bacteria play an essential part in the cycling of reduced nitrogen compounds throughout marine, freshwater, and soil environments. Chemolithoautotrophic bacteria convert ammonia to nitrite by two key enzymes and assimilate CO2 as the major carbon source. The first enzyme, the membrane-bound ammonia monooxygenase (AMO) catalyzes the oxidation of ammonia to hydroxylamine [22, 59], which is then further oxidized to nitrite by hydroxylamine oxidoreductase (HAO) [3, 46], Two proteins were sequenced and described as possible subunits of the AMO, a 27 kDa sized membrane-associated protein (AmoA) containing, most likely, the active site of the AMO [23, 24] and a protein of 41 kDa (AmoB) [8, 33].

All known ammonia oxidizers are members of two lineages within the β- and γ-subclasses of Proteobacteria according to comparative 16S rRNA sequence analysis [30]. The marine species of the genus Nitrosococcus cluster within the γ-subclass of Proteobacteria [58]. Members of the four other genera Nitrosomonas (including Nitrosococcus mobilis, which belongs phylogenetically to Nitrosomonas), Nitrosospira, Nitrosolobus, and Nitrosovibrio, form a monophyletic assemblage within the β-subclass of the Proteobacteria [57]. The ammonia oxidizers have been divided into separate genera based on phenotypic characteristics, in particular cell shape and the arrangement of intracytoplasmic membranes (ICMs) [29]. The coccoid cells of the marine Nitrosococcus species have a centrally located stack of flattened ICMs, whereas the coccoid and rod-shaped cells of Nitrosomonas possess paired membranes running along the periphery of the cells. The pleomorphic lobes of Nitrosolobus are divided into cell compartments by cytomembranes [54]. No ICMs are found in the spiral and slender curved rods of members of the genera Nitrosospira and Nitrosovibrio, respectively [18]. Irrespective of the gross morphological differences of the genera, it has been suggested [19] and subsequently questioned [47] that the genera Nitrosospira, Nitrosolobus, and Nitrosovibrio should be reclassified in the single genus Nitrosospira because of the high levels of homology in their 16S rRNA.

The presence of certain species of ammonia oxidizers within any given environmental niche, e.g., wastewater treatment plant or agricultural soil, gives evidence of environmental conditions. However, the detection and identification of these bacteria is difficult because of their slow growth rate and aggregation to biofilms [28, 43]. Traditional counting methods like the most-probable-number (MPN) technique [31] are time-consuming and often underestimate the numbers of ammonia oxidizers in the natural environment [7, 27]. Antibodies or 16S rRNA-targeted oligonucleotide probes are used for in situ analyses in order to avoid the limitations of the MPN technique. In previous studies antibodies had been developed using whole cells of ammonia oxidizers that recognize epitopes of the cell wall [6, 36, 41, 42, 48, 52]. These antibodies were applied in ecological studies to detect and count ammonia oxidizers in bacterial communities using fluorescence microscopy, and 10 to 1000-fold greater cell numbers were found compared to MPN techniques [27, 48, 51]. However, this approach is not entirely culture-independent because ecologically relevant strains have to be isolated prior to antibody development. As ammonia oxidizers show high serological diversity within even one genus [6, 52], unculturable bacteria are consequently not detectable using antibodies raised against whole cells, and a variety of antibodies are required. In contrast to antibodies to cell envelopes, antibodies to key enzymes of N-cycling bacteria such as ammonia oxidizers [38], nitrite oxidizers [1, 4, 5], and denitrifying bacteria [10, 11, 26, 53] could be successfully applied for the detection of these bacteria in culture-independent approaches.

In this study, polyclonal antibodies were produced against the AmoA subunit of the AMO. The specificity of AmoA antibodies was tested using immunoblotting of pure cultures of members of all genera of ammonia oxidizers, and immunofluorescence (IF) labeling with the AmoA antibodies and the previously described AmoB antibodies [38] was developed to visualize and identify ammonia oxidizers in pure, mixed, and enrichment cultures.

Materials and Methods

Bacterial Strains and Culture Conditions

All bacterial strains used in this study are listed in Table 1. The 21 strains of ammonia oxidizers and 7 strains of nitrite oxidizers were derived from soil, sewage, a biowaste fermenter, cattle manure, a heating system, sandstone of historical buildings and sea water (for more details see [38]). All strains including the methane oxidizers Methylococcus capsulatus Bath, Methylomonas methanica Oo52006, and Methylocystis parvus 4a and the chemoorganotrophic bacteria Bacillus subtilis 019, Escherichia coli K12/067 (ATCC 23716), Methylobacterium radiotolerans, Paracoccus denitrificans 001 (ATCC 19367), and Pseudomonas sp. AM1 are stored in the culture collection of the Institut für Allgemeine Botanik, Abteilung Mikrobiologie, Universität Hamburg. All other strains were obtained from C. Coeur (University of Lyon I, Villeurbanne, France).

Table 1 Reactivity of AmoA-and AmoB-antibodies and 16S rRNA probes with ammonia oxidizers, methane oxidizers, nitrite oxidizers and different chemoorganotrophic bacteria by using immunoblot, immunofluorescence, and FISH techniques
Full size table

Terrestrial and freshwater ammonia oxidizers were grown at 28°C in mineral salts medium [28] in the presence of 10 mM ammonium. Nitrosomonas cryotolerans Nm 55, Nitrosomonas halophila Nm 1, and Nitrosococcus mobilis Nc 2 were grown in the same medium containing 10 g NaCl L−1. Nitrosomonas marina Nm 22, Nitrosococcus oceani Nc 1, and Nitrosococcus halophilus Nc 4 were cultivated in seawater medium of the following composition: 10 mM NH4Cl, 0.4 mM KH2PO4, 3 g HEPES, and 1 mL 0.05% (w/v) cresol red solution per liter of 40% seawater.

Nitrobacter hamburgensis X14, Nitrobacter winogradskyi Engel, and Nitrobacter vulgaris K48 were grown mixotrophically in the presence of 2 g of NaNO2 L−1 [9]. Nitrospira moscoviensis M-1 was cultivated in mineral medium with 0.2 g of NaNO2 L−1 [13]. Nitrospina gracilis 3, Nitrospina sp. 347, and Nitrococcus mobilis 231 were cultivated in seawater media according to Watson and Waterbury [55]. The cultures were incubated at 28°C expect Nitrospira moscoviensis M-1, which was incubated at 37°C.

The methane oxidizers were cultivated in nitrate mineral salts medium [56] containing 0.25 µM CuSO4 with a 3% methane synthetic air atmosphere. The methylotrophs were grown in mineral medium with 0.15% (w/v) methanol [15]. All other bacterial strains were cultivated according to the instructions of the American Type Culture Collection (ATCC).

Enrichment Cultures

For the enrichment of ammonia oxidizers, activated sludge, ground natural sandstone, and garden soil were added to mineral salts medium [28] containing 10 mM ammonium and incubated for 2 weeks at 28°C. The activated sludge samples originated from the aeration stage of the sewage treatment plant in Dradenau near Hamburg, Germany, and had a pH of 8.0. Garden soil samples with a pH of 7.6 were obtained from the new botanical garden in Hamburg, Germany. Stone samples were taken from specially cut specimens (5 × 5 × 1 cm) from Baumberger sandstone, which had been exposed for 5 years in the new botanical garden of Hamburg, Germany. After this exposure, the sandstones were covered with green algae biofilms and other microorganisms and showed visible signs of weathering. The sandstone had a pH of 8.4 and was ground by pestle and mortar.

Development of Antibodies

A 15-residue synthetic peptide corresponding to the AmoA N-terminal peptide sequence of Nitrosomonas eutropha Nm 57 was produced and conjugated to keyhole limpet hemocyanin (KHL)-protein (Biotrend, Cologne, Germany). The SwissProt accession numbers for N. eutropha Nm 57 amoA1 and amoA2 are U51630 and U72670, respectively. Antiserum against this synthetic peptide was raised in rabbits (Biotrend). AmoB-antibodies were recently produced by Pinck et al. [38] using the whole subunit of Nitrosomonas eutropha N904 for immunization of chickens.

Immunoblotting

Protein extracts of pure cultures (1.5 mg protein mL−1) were separated by SDS-PAGE, electroblotted onto a cellulose nitrate membrane using a discontinuous buffer system, and subsequently applied to the immunostaining. The 27-kDa AmoA protein and the 41-kDa AmoB protein were detected in separate immunoblots. Alkaline phosphatase conjugated anti-rabbit antibodies (Sigma) and alkaline phosphatase conjugated anti-chicken antibodies (Sigma) were used as secondary antibodies for the detection of AmoA and AmoB antibodies, respectively. For the enzyme reaction 0.005% 5-bromo-4-chloro-3-indolyl phosphate and 0.001% 4-nitro-blue-tetrazolium was employed. The detailed procedures of SDS-PAGE, electroblotting, and immunostaining were given by Pinck et al. [38].

IF Labeling

For IF labeling cells of pure and enrichment cultures were harvested by centrifugation and stored overnight in phosphate-buffered saline (PBS)–ethanol (1:1) at −20°C. The samples were placed on slides and dehydrated by using 50, 80, and 96% ethanol (3 min each) [17]. The cells were then treated with lysozyme to enhance the permeation of the antibodies [16]. All samples were blocked with PBS containing 3% bovine serum albumin (BSA) for 30 min at room temperature. The samples were then simultaneously incubated overnight with the AmoA- and AmoB-antibodies diluted 1:10 in PBS containing 0.05% BSA and 0.025% Tween 20. Afterward, the cells were incubated with secondary antibodies diluted 1:100 in PBS containing 0.05% BSA and 0.025% Tween 20 for 1 h in the dark at room temperature. The secondary antibodies were Cy2-labeled anti-rabbit (Biotrend) and Cy3-labeled anti-chicken antibodies (Biotrend) for binding AmoA and AmoB antibodies, respectively. The reactions were stopped by washing the slides in PBS. Control preparations without primary antibodies were included in every experiment. Samples were stained with 4′,6-diamidino-2-phenylindole (DAPI) (10 µg mL−1) for 5 min to detect total cells.

FISH

For fluorescence in situ hybridization (FISH) cells were fixed in 3% formaldehyde for 1 h on ice and stored in PBS–ethanol (1:1) at −20°C. The samples were placed on slides and dehydrated by using 50, 80, and 96% ethanol (3 min each) [17]. Afterward, they were incubated with the 16S rRNA directed oligonucleotide probe EUB338, Nso 190, NEU, or Nsv 443 in hybridization buffer (0.9 M NaCl, 20 mM M Tris/HCl pH 7.2, 0.01% SDS) with a final concentration of 30% formamide (EUB338, Nsv 443) or 40% formamide (NEU, Nso 190) at a temperature of 46°C for 2 h. As a competitor for NEU, an equimolar amount of unlabeled oligonucleotide CTE was added. Probe EUB338 was used to detect members of the domain Bacteria [12], and probe Nso 190 detects all ammonia oxidizers of the β-subclass of Proteobacteria [35] except some Nitrosomonas strains (Table 1, [40]). Probe NEU is specific for the detection of only a few members of the genus Nitrosomonas (Table 1 [40, 50]). Probe Nsv 443 is directed to Nitrosolobus multiformis, Nitrosospira briensis and Nitrosovibrio tenuis (Table 1 [35, 40]). The probes were labeled with the fluorochrome Cy3 (MWG-Biotech AG). The samples were then incubated 20 min at 48°C in washing buffer (50 mM NaCl, 20 mM Tris/HCl pH 7.2, 0.01% SDS). Samples were stained with 4′,6-diamidino-2-phenylindole (DAPI) (10 µg mL−1) for 5 min to detect total cells.

Fluorescence Microscopy and Confocal Laser Scanning Microscopy

DAPI stained cells were visualized by using Leica filter set A (BP 340–380 exc.; RKP 400; LP 425 em.). The Cy2 labeling of secondary AmoA antibodies and the Cy3 labeling of secondary AmoB antibodies and 16S rRNA probes were visualized by using Leica filter set I 3 (BP 450–490 exc.; RKP 510; LP 515 em.) and Leica filter set N 2.1 (BP. 515–560 exc.; RKP 580; LP 590 em.), respectively.

IF labeling and FISH were additionally visualized with a confocal laser scanning microscope (CLSM) (model TCS 4D; Leica); excitation was supplied by an argon–krypton laser (488/568 exc.; LP 590, BP 520–560 em.). Image processing was performed with standard software (Scanware 5.1; Leica). No or only slight autofluorescence of biofilm, soil, or stone particles was found in the absence of fluorescent antibodies or probes.

Electron Microscopy

The methods used for cell fixation, embedding, ultrathin sectioning, and shadow casting were the methods described by Ehrich et al. [13]. Electron microscopy was performed with a Philips model 420 transmission electron microscope.

Results

Immunoblotting

Polyclonal antibodies were produced against a synthetic AmoA peptide, and their specificity was tested by immunoblotting of crude extracts. Various strains were investigated including 21 ammonia oxidizers, three methane oxidizers, seven nitrite oxidizers, and 14 strains of different chemoorganotrophic bacteria, which are all summarized in Table 1. In immunoblot studies the AmoA antibodies were highly specific for the detection of a 27-kDa protein in cell extracts of Nitrosomonas strains N. europaea Freitag, N. eutropha N904, N. eutropha Dave, and N. halophila Nm 1. Cells of all the other strains of Nitrosomonas sp. including Nitrosococcus mobilis Nc 2 could not be detected by the AmoA antibodies. These results are in contrast to those previously reported results that showed AmoB antibodies detected the 41-kDa AmoB of all ammonia oxidizers of the β-subclass of Proteobacteria [38]. These results are included in Table 1 for completeness. The AmoA-antibodies did not react with the crude extracts of the tested ammonia oxidizers of the γ-subclass of Proteobacteria, the methane oxidizers, the nitrite oxidizers, and the different chemoorganotrophic bacteria.

IF Labeling of Pure Cultures

Immunofluorescence (IF) labeling using the AmoA antibodies and the previously described AmoB antibodies [38] was developed on pure cultures to visualize intact cells of ammonia oxidizers. To obtain successful antibody penetration into the bacteria, the cells had to be stored in PBS–ethanol at −20°C and were then treated with lysozyme. The AmoA antibodies detected all strains of the genus Nitrosomonas using IF labeling although fluorescence intensity varied (Table 1). Strains of Nitrosomonas sp. such as N. eutropha, N. europaea, and N. halophila that were successfully detected by immunoblotting were strongly stained, whereas cells of N. cryotolerans Nm 55 and N. marina Nm 22 showed weak fluorescence. In contrast to the AmoA antibodies, the AmoB antibodies had a broad specificity, as they detected all strains of ammonia oxidizers of the β-Proteobacteria that were tested (Table 1), all of which were stained strongly. Apart from weak IF labeling of Nitrosococcus halophilus Nc 1, Nitrosococcus oceani Nc 4, and Methylococcus methanica Oo52006 using the AmoB-antibodies, both antisera did not react with ammonia oxidizers of the γ-subclass of Proteobacteria, methane oxidizers, nitrite oxidizers, and different chemoorganotrophic bacteria. In control experiments without AmoA and AmoB antibodies the cells exhibited no fluorescence signals.

IF labeling revealed that the AmoB antibodies were associated with the cytoplasmic membranes. As a result, the signals appeared at the cell periphery of cells of Nitrosomonas sp., Nitrosovibrio sp., and Nitrosococcus sp. (Fig. 1a–c). In cells of members of the genera Nitrosospira and Nitrosolobus, the fluorescence signals were also observed in the cytoplasm along the cytoplasmic membrane, which is vaulted to the center in tightly coiled cells of Nitrosospira sp. (Fig. 1d) and partially compartmentalizes the lobate cells of Nitrosolobus sp. (Fig. 1e).

Figure 1
figure 1

IF labeling with AmoA and AmoB antibodies of pure cultures of ammonia oxidizers belonging to the β-subclass of Proteobacteria (a) Nitrosomonas eutropha N904 (zoom step 16.4). (b) Nitrosovibrio sp. Nv K7.1 (zoom step 17.0). (c) Nitrosococcus mobilis Nc 2 (zoom step 15.8). (d) Four cells of Nitrosospira sp. Nsp R6.2 (zoomstep 20.0). (e) One single cell of Nitrosolobus multiformis Nl 13 in the center (zoom step 23.0). Bars = 1 µm. The objective used was a Neoflutar objective (100×/1.4oil). The images were obtained with a CLSM (model TCS 4D microscope; Leica) by using different zoom steps. Excitation of the green Cy2-labeled anti-rabbit antibodies for the detection of AmoA and red Cy3-labeled anti-chicken antibodies for the detection of AmoB was provided by an argon–krypton laser (488/568 exc./LP 590 em. and BP 520–560 em.).

Full size image

Using both antibodies simultaneously, cells of Nitrosomonas sp. could be double-stained and showed bright red fluorescence of the secondary AmoB antibodies at the cell periphery as well as bright green signals of the secondary AmoA antibodies distributed over the whole cell (Fig. 1a). In contrast, cells of Nitrosococcus mobilis and cells belonging to the genera Nitrosovibrio, Nitrosospira, and Nitrosolobus were only labeled by the AmoB antibodies (Fig. 1b–e). The different IF labeling specificity of AmoA and AmoB antibodies together with the characteristic morphology of the cells enabled us to distinguish between the members of different genera of ammonia-oxidizers of β-Proteobacteria in pure cultures. Using IF labeling, the four new ammonia-oxidizing strains R1.24, G1.6, M1.3, and R6.2 isolated from historical building stones could be classified on the genus label. The straight rods of strain R1.24 were labeled with both antibodies (Table 1). In contrast the coiled rods of the strains G1.6, M1.3, and R6.2 could only detect by AmoB-antibodies (Table 1, Fig. 1d). Because of the specific IF labeling and typical morphology of the cells, strain R1.24 was classified as a member of the genus Nitrosomonas and the strains G1.6, Ml.3, and R6.2 were assigned to the genus Nitrosospira. This classification was underlined by electron microscopic investigations, where cells of strain R1.24 showed typical Nitrosomonas and cells of the strains G1.6, M1.3, and R6.2 typical Nitrosospira morphology (not shown).

In defined mixed cultures comprising members of the genus Nitrosomonas and Nitrosococcus mobilis or one strain of a Nitrosospira sp., a Nitrosovibrio sp., and a Nitrosolobus sp., respectively, cells could also be identified to the genus level by IF labeling (Fig. 2).

Figure 2
figure 2

IF labeling with AmoA and AmoB antibodies of defined mixed cultures of Nitrosomonas eutropha N904 and (a) Nitrosovibrio sp. Nv K7.1 (zoom step 5.0), (b) Nitrosococcus mobilis Nc 2 (zoom step 6.2), (c) Nitrosospira sp. Nsp G1.6 (zoom step 5.4), (d) Nitrosolobus multiformis Nl 13 (zoom step 3.8). Bars = 5 µm.The objective used was a Neoflutar objective (100×/1.4 oil). The images were obtained with a CLSM (model TCS 4D microscope; Leica) by using different zoom steps. Excitation of the green Cy2-labeled anti-rabbit antibodies for the detection of AmoA and red Cy3-labeled anti-chicken antibodies for the detection of AmoB was provided by an argon–krypton laser (488/568 exc./LP 590 em. and BP 520–560 em.). Nc, Nitrosococcus; Nl, Nitrosolobus; Nm, Nitrosomonas; Ns, Nitrosospira.

Full size image

FISH of Pure Cultures

Using FISH, the 16S rRNA probe EUB338 detected all pure cultures of ammonia oxidizers that were tested (Table 1). Probes Nso 190, NEU, and Nsv 443 were only applied on pure cultures of Nitrosomonas europaea Freitag, Nitrosomonas eutropha N904, and Nitrosospira Nsp1 and the four isolated strains R1.24, G1.6, M1.3, R6.2 (Table 1). Probe Nso 190 labeled all the strains except strain R1.24. In contrast, probe NEU could only detect Nitrosomonas europaea Freitag and Nitrosomonas eutropha N904, and probe Nsv 443 detected Nitrosospira Nsp1, but none of the isolates (Table 1). For completeness, the specificity of probe Nso 190, NEU, and Nsv 443 for nearly all of the ammonia oxidizers used in this study was included in Table 1.

IF Labeling of Enrichment Cultures

Based on the results obtained from pure cultures, enrichment cultures either derived from activated sludge, natural sandstone or soil were analyzed using IF labeling and FISH.

In enrichment cultures from activated sludge from the sewage treatment plant in Dradenau, up to 90% of the free suspended cells and about 40% of biofilm-embedded cells could be labeled in situ by the antibodies. Almost all detected ammonia oxidizers could be identified as cells of the genus Nitrosomonas according to the specific IF labeling. Except for the cells associated with the biofilm, almost all ammonia oxidizers were labeled by AmoA and AmoB antibodies. The biofilm cells, however, were only detectable by AmoA-antibodies (Fig. 3b) because of nonspecific reactions of the AmoB antibodies with sludge particles (Fig. 3c), but could be identified as Nitrosomonas sp. cells. Inside the biofilm, loosely aggregated single cells (Fig. 3d) and dense microcolonies (Fig. 3e) of Nitrosomonas sp. were found. Outside the biofilm only a few cells were exclusively stained with AmoB antibodies and could be identified as cells of Nitrosospira sp. and Nitrosococcus sp.. The IF labeling results were confirmed by FISH. Almost all DAPI-stained cells could be detected by probe EUB338. Outside the biofilm about 90% of the cells and about 40% of the cells inside the biofilm were stained with the Nitrosomonas sp. specific oligonucleotide probe Nso 190 and NEU. Only a few cells could be labeled by the Nitrosopira sp., Nitrosolobus sp., and Nitrosovibrio sp. specific probe Nsv 443.

Figure 3
figure 3

IF labeling with AmoA and AmoB antibodies of an enrichment culture from activated sludge samples from the sewage treatment plant of Dradenau. (a) DAPI staining; (b, d, e) IF labeling with AmoA-specific antibodies; (c) IF labeling with AmoB-specific antibodies. Ammonia oxidizers belonging to the Nitrosomonas europaea cluster occurring as (d) single cells (zoom step 4.9) or (e) microcolonies (zoom step 4.7). Bars a–c = 20 µm, d, e = 5 µm. The objective used was a Neoflutar objective (100×/1.4 oil). The images were obtained with aCLSM (model TCS 4D microscope; Leica) by using different zoom steps. DAPI was visualized with Leica filter set A (PB 340–380 exc.; RKP 400; LP 425 em.). IF labeling of the green Cy2-labeled anti-rabbit antibodies for the detection of AmoA and red Cy3-labeled anti-chicken antibodies for the detection of AmoB was visualized by filter set I 3 (BP 450–490 exc.; RKP 510; LP 515 em.) and filter set N 2.1 (BP 515–560 exc.; RKP 580; LP 590 em.), respectively. Details in d, e were provided by an argon–krypton laser (488/568 exc./LP 590 em. and BP 520–560 em.).

Full size image

In enrichment cultures prepared from samples of Baumberger sandstone, cells were mostly bound to particles and appeared mainly as single cells. About 68% of the cells were identified by IF labeling with AmoA and AmoB antibodies as cells of Nitrosomonas. About 17% of the cells were only stained by AmoB antibodies and had typical Nitrosospira sp. and Nitrosovibrio sp. morphologies (Fig. 4a). In contrast to the IF labeling, the detection of cells with probes EUB338, Nso 190, NEU, and Nsv 443 failed because of nonspecific reactions of the probe with the stone particles.

Figure 4
figure 4

IF labeling with AmoA and AmoB antibodies of enrichment cultures from (a) Baumberger sandstone (zoom step 1.9), (b) garden soil samples (zoom step 2.3). Bars = 5 µm. The objective used was a Neoflutar objective (100×/1.4 oil). The images were obtained with a CLSM (model TCS 4D microscope; Leica) by using different zoom steps. Excitation of the green Cy2-labeled anti-rabbit antibodies for the detection of AmoA and red Cy3-labeled anti-chicken antibodies for the detection of AmoB was provided by an argon–krypton laser (488/568 exc./LP 590 em. and BP 520–560 em.). Nc, Nitrosococcus; Nl, Nitrosolobus; Nm, Nitrosomonas; Ns, Nitrosospira; Nv, Nitrosolobus.

Full size image

In the enrichment cultures of the soil samples, cells were freely suspended or particle bound and were generally not aggregated. About 85% of the cells could be labeled with the AmoB antibodies, and only a few were additionally stained with the AmoA antibodies. Because of the specific IF labeling and the morphology the ammonia oxidizers, cells could be identified as members of the genera Nitrosospira and Nitrosolobus. Coccoid ammonia oxidizers were also detected that might be small cells of members of the genus Nitrosolobus or belong to Nitrosococcus sp. (β-Proteobacteria) (Fig. 4b). The dominance of Nitrosolobus sp. was confirmed by electron microscopic analyses (Schnier, personal communication). FISH with the probes EUB338, Nso 190, NEU, and Nsv 443 did not succeed in the soil sample because probes bound nonspecifically to soil particles so that ammonia oxidizers could not be detected.

Discussion

Except for the two marine strains of Nitrosococcus sp., ammonia oxidizers form a monophyletic assemblage within the β-subclass of Proteobacteria [47]. Therefore, it is assumed that their key enzyme, the ammonia monooxygenase (AMO), is highly conserved. However, little is known about the structure and function of AMO since the enzyme has not been purified [14, 44, 45]. Inactivation experiments suggested that the proteins AmoA and AmoB were possible subunits of AMO [33]. Previously AmoB antibodies had been developed using the whole purified 41-kDa AmoB protein of Nitrosomonas eutropha N904, which detected all ammonia oxidizers of the β-subclass of Proteobacteria that were tested [38]. In this study, polyclonal antibodies were produced against the AmoA using a small 15-residue peptide corresponding to the AmoA N-terminal peptide sequence of Nitrosomonas eutropha Nm 57. Consequently, the antibodies detect only a small hydrophilic part of the 27-kDa AmoA [33]. The antibodies that recognized the AmoA of Nitrosomonas eutropha also detected the AmoA of other Nitrosomonas species. Using immunoblot analysis, only members of the Nitrosomonas europaea cluster (N. europaea, N. eutropha, N. halophila) [39] reacted with AmoA antibodies, but when IF labeling was used, the AmoA antibodies also detected members of the remaining species of Nitrosomonas. However, the species that were not detected in the immunoblot showed only weak fluorescence. The IF method was obviously more sensitive than the Western blot technique, analogous to the findings of Ward and co-workers [53], who used polyclonal antibodies targeting the nitrite reductase. The different IF labeling intensities of Nitrosomonas species might result from the differences between the N-terminal AmoA sequences, which could have caused the different binding efficiency of the antibodies. All Nitrosomonas strains showing high AmoA-sequence similarity to N. eutropha, such as N. europaea and N. halophila, were intensively stained, whereas strains with low AmoA-sequence similarity such as N. cryotolerans and N. marina were weakly stained [37, 40]. Only Nitrosococcus mobilis, close relative of Nitrosomonas eutropha [39, 40], could not be detected by AmoA antibodies. The N-terminal AmoA sequence of N. eutropha is identical with the sequence of N. europaea [37]. Aside from the two Nitrosomonas species, only the sequence of the N terminus of AmoA of one strain of N. cryotolerans is known [37], whereas the N-terminal AmoA sequences of the other species of Nitrosomonas are known only partially (Nitrosococcus mobilis, SwissProt accession number AF037108) or not at all. The N-terminal AmoA sequence of N. cryotolerans differs from the sequence of N. eutropha [37], which might be the reason for the weak IF labeling using AmoA-antibodies. Recently, Norton and co-workers [37] described the AmoA-sequence of two new isolated strains of Nitrosomonas with N-terminal AmoA sequences, which further deviate from the sequence of N. eutropha and therefore might not be detected by AmoA antibodies. Although the N-terminal AmoA sequences of strains of the genera Nitrosospira, Nitrosovibrio, and Nitrosolobus correspond with the described sequence of N. cryotolerans [37], they did not react with the AmoA antibodies. The third and fourth amino acids of their AmoA are deleted and the glutamic acid in position 7 is replaced by aspartic acid [37]. The reason for this disagreement might be that little strain-specific deviations between the N-terminal AmoA-sequence exist, as described for Nitrosolobus multiformis [37], and the N. cryotolerans strain we used might correspond more with N. eutropha than the used strains of Nitrosospira sp., Nitrosovibrio sp., and Nitrosolobus sp. The different specificity of AmoA antibodies will only be completely explained when the N terminus of the AmoA of the ammonia oxidizer strains that were used in this study has been sequenced. If instead of the N-terminal sequence a more conserved internal or C-terminal region of AmoA had been chosen, these antibodies might be reacted with all ammonia oxidizers of the β-subclass of Proteobacteria. In Nitrosococcus oceani, a member of the γ-subclass of Proteobacteria, only three amino acids of the 15 N-terminal AmoA sequence correspond with the AmoA of N. eutropha [2, 37], and therefore it did not react with the antibodies. Further, no reaction of AmoA antibodies was found with the closely related particulate methane monooxygenase (pMMO) of methane oxidizing bacteria, because the N-terminal pMMO sequence of Methylococcus capsulatus is almost identical to the AmoA of Nitrosococcus oceani [2] and therefore clearly differs from the AmoA sequence of N. eutropha. The fluorescent signals of all cells correlated with the locations of the AMO in the cytoplasmic and intracytoplasmic membranes [23].

Using both antisera simultaneously, the different IF labeling specificity of AmoA and AmoB antibodies together with the characteristic morphology of the cells enabled us to distinguish between the members of different genera of ammonia oxidizers of β-Proteobacteria in pure cultures as well as in mixed cultures. They could be applied for classification of new isolates of ammonia oxidizers. By IF labeling the isolates could assigned to the genus Nitrosomonas or Nitrosospira. This clear classification was not possible by using the 16S rRNA probes NEU, Nso 190, and Nsv 443. This might be due to the choice of the probes. Although probe Nso 190 detected many ammonia oxidizers of the β-subclass of Proteobacteria [35], some Nitrosomonas strains were not included [40], and probe NEU is only specific for the detection of a few members of the genus Nitrosomonas [40, 50]. Therefore it is not remarkable that strain R1.24, which was classified by IF labeling to genus Nitrosomonas, could not be detected by probes Nso 190 and NEU. For a clear classification of the Nitrosomonas isolate the probe Nm0 should be used, which cover all strains of Nitrosomonas but might not detect the other four genera of ammonia oxidizers [39, 40]. In contrast, no 16S rRNA probe is thus far known that exclusively detects the genus Nitrosospira [40]. Therefore no probe exists that enables the clear classification of the isolates G1.6, M1.3, and R6.2, which were assigned by IF labeling to genus Nitrosospira. Although probe Nsv 443 has a broader specificity, as it is designed for the detection of Nitrosolobus sp., Nitrosospira sp., and Nitrosovibrio sp. [35, 40], it could not detect the isolates. In contrast to the antibodies, thus far no set of 16S rRNA probes is available for the classification of ammonia oxidizers to the five genera [40].

In addition, the specific immunoreactions were suitable for the identification of ammonia oxidizers enriched from different natural samples. In enrichment cultures of activated sludge from the sewage treatment plant in Dradenau, cells of Nitrosomonas sp. were mainly detected, which could be confirmed using FISH analysis with the Nitrosomonas- specific oligonucleotide probes NEU and Nsv 443. This finding is in agreement with previous in situ IF and FISH analysis, and Nitrosomonas sp. was mostly isolated from activated sludge samples [30, 35, 48]. In enrichment cultures derived from sandstone, members of the genus Nitrosomonas were also mainly found. However, additional cells were exclusively detected by the AmoB antibodies, and combined with the characteristic morphology the cells could be identified as members of the genera Nitrosovibrio and Nitrosospira. In previous studies Nitrosovibrio sp. and Nitrosospira sp. were predominantly isolated from sandstone samples [34, 43]. In enrichment cultures from neutral garden soil (pH 7.6) from the new botanical garden in Hamburg, members of the genera Nitrosospira and Nitrosolobus were the most abundant ammonia oxidizers. In these cultures, coccoid cells were also detected by AmoB antibodies that might have been small cells of Nitrosolobus sp. or might have belonged to members of the genus Nitrosococcus (β-Proteobacteria). Electron microscopic investigations of the same cultures revealed characteristic cells with a morphology and ultrastructure of small Nitrosolobus sp. cells (Schnier, personal communication), which is in agreement with recent studies in which Nitrosolobus [32, 49] and Nitrosospira [6, 19, 21, 25] were the most abundant ammonia oxidizers in loamy and arable soils with neutral pH.

In contrast to the AmoA antibodies and the oligonucleotide probes, the AmoB antibodies could not be applied to detect ammonia oxidizers within the biofilm of the activated sludge sample because of nonspecific reactions. However, only slight nonspecific reaction of the antibodies appeared in the presence of stone and soil particles, and the IF cell signals could be easily distinguished from the background, whereas with probes EUB338, Nso 190, NEU, and Nsv 443 the detection of ammonia oxidizers failed in these samples.

The antibodies against the AmoA and AmoB subunits of the AMO detected all ammonia oxidizers of the β-Proteobacteria that were tested and therefore have an advantage over previous antibody detection methods, which were limited in the application to specific serological groups [6, 52] as they recognize specific epitopes of the cell wall [41, 42, 48]. Presumably, the antibodies against the key enzyme, especially the AmoB antibodies, might be used for the detection of all ammonia oxidizers of the β-subclass of Proteobacteria even those that have not been described. Beside the enrichment cultures, the antibodies may be further used for the direct in situ detection of ammonia oxidizers in natural samples with a sufficient number of cells. The AmoB antibodies might provide the only tool for detecting all species of ammonia oxidizers of the β-subclass of Proteobacteria simultaneously and exclusively [40].