INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Autotrophic nitrifying bacteria that oxidize ammonium to nitrite and nitrate are found in soils, sediments, wastewaters, freshwater, and marine water and on building facades. They are essential components of the nitrogen (N) cycle, linking the most reduced and most oxidized forms of inorganic N. Nitrification occurs as a two-step process carried out by two distinct groups of bacteria; ammonia-oxidizing bacteria convert ammonia to nitrite, and then nitrite oxidizers convert nitrite to nitrate (22, 30). Environmental factors control the rate of nitrification. The most significant environmental factors are substrate concentration, pH, temperature, and oxygen availability (12, 23). Nitrifying bacteria exhibit different substrate concentration sensitivities (26). Media containing low substrate concentrations (10 mg of NH₄⁺ liter¹) can give larger most-probable-number counts of ammonia oxidizers than media containing higher NH₄⁺ concentrations (6, 26). Also, ammonia oxidation is inhibited at high substrate concentrations. The growth rates of Nitrosomonas spp. cultures were reduced in the presence of 1,050 to 2,800 mg of NH₄⁺-N liter¹ (16). Substrate inhibition of ammonia oxidation has also been observed in studies of wastewater systems (23). Natural environments, such as soil and water, usually contain 1 to 10 mg of NH₄⁺-N liter¹ (22), yet liquid wastes from animal farms give rise to concentrations up to 1,600 or 5,600 mg of NH₄⁺-N liter¹ (5, 17). Free ammonia (NH₃) rather than the total ammonium concentration inhibits ammonia oxidizers (1). As the ratio between the ionized form and the nonionized form depends on pH, the toxicity of ammonium also depends on the environmental pH.

The pH range for growth of pure cultures of ammonia oxidizers is 5.8 to 8.5, and the pH range for growth of nitrite oxidizers is 6.5 to 8.5 (30). Nitrification was inhibited at pH values below 5.8 in our preliminary experiments performed with an enriched culture of nitrifiers obtained from wastewater. Yet in natural environments, such as soil, nitrification has been reported to occur at pH values below 4.0 (7, 29).

Limiting amounts of dissolved oxygen (concentrations below 2 mg liter¹) inhibit nitrification and cause nitrite accumulation or nitrous and nitric oxide production (9, 21). Ammonia-oxidizing bacteria are the key functional group in removing ammonium from wastewaters. Knowledge of the effect of oxygen on nitrification and nitrifying populations has economic importance since aeration of activated sludge is one of the most costly items in the operation of a wastewater treatment plant (21).

In environments with high inputs of ammonium, such as wastewaters, biooxidation of this substrate increases the oxygen uptake and lowers the pH. Such modifications of the environment not only affect the production of nitrite and nitrate but can also select a different nitrifying community that is perhaps specialized for these new conditions. Nitrification does occur in extreme environments that pure cultures of nitrifiers cannot tolerate (4). In this study we examined extreme environments in which nitrifying bacteria may be viable but have not been cultured thus far.

Because of the difficulty of obtaining nitrifier isolates, nucleic acid-based methods have greatly aided studies of the diversity of nitrifiers (11, 20, 27, 28). Recent molecular investigations have provided valuable information concerning the diversity of ammonia oxidizers in natural environments (5, 15, 20, 25). However, no previous study has focused on the structural or compositional responses of nitrifying communities to perturbations in the environment. In the present laboratory study we examined the effects of high ammonium concentrations, different pH values, and different oxygen concentrations on nitrification and on the community structure of nitrifying bacteria from wastewater. To test the abilities of the communities to regain their original structures, growth of nitrifying communities under the new conditions was followed by incubation under the original conditions.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Nitrifying culture and mineral medium. An enrichment culture of nitrifying bacteria was prepared by inoculating an aerated, continuous flow of fresh basal mineral medium containing 100 to 500 mg of NH₄⁺-N liter¹ with municipal wastewater. After a few months of growth, there was enough biomass to harvest the culture from the column. The basal mineral medium for nitrifying bacteria (30) was used, except that no phenol red was added and higher ammonium concentrations and modified buffer (0.1 M phosphate buffer) were used (17). Ammonium-N was added to the basal medium as (NH₄)₂SO₄.

Experimental design. The enriched nitrifying bacterial culture was used as a starter culture for 10 continuously fed reactors, each containing 100 ml of basal mineral medium supplemented with ammonium-N. The reactors were made from 25-cm-high glass cylinders having a diameter of 4 cm. Each reactor had a side arm with an opening that enabled outflow and maintenance of a constant volume (100 ml) of the reaction mixture. The reactor medium in each cylinder was aerated from the bottom by using coiled Teflon tubing with tiny holes. The reactors were sparged with preset gas mixtures, and the ammonium concentrations in the reactors were maintained at the levels shown in Table 1. Incubation for 74 days was conducted in the dark at 25°C. The dissolved oxygen content was monitored, and the pH was adjusted with 1 M Na₂CO₃. The reactors were not operated as chemostats; instead, the solution concentrations of ammonium and oxygen and the pH were kept approximately constant. To do this, the flow rate of medium was increased with time from 0.8 to 2.5 ml h¹ as biomass accumulated. Most of the biomass was retained in the reactor since the organisms grew as flocks. The ammonium concentration was measured initially at 3- to 5-day intervals and later daily to determine the amount of ammonium supplement needed to maintain the ammonium concentration within 15% of the starting concentration. Nitrate and nitrite concentrations were measured periodically to confirm that nitrifying activity was occurring. Samples of biomass were collected for molecular analysis after 14, 25, 38, 50, 60, and 74 days of incubation. The reactor cultures were briefly stirred prior to sampling.

1

Analytical methods. Nitrification was confirmed by measuring the nitrite and nitrate contents and the ammonium consumption in the reactors. Ammonium concentrations were determined colorimetrically with indophenol blue (14). Nitrite and nitrate contents were measured by high-performance liquid chromatography. Biomass protein contents were estimated by the biuret method (10). The contents of the heterotroph populations in the enrichment cultures were determined on R2A agar plates (Difco Laboratories, Detroit, Mich.).

Nucleic acid extraction and SSU rDNA amplification. Genomic DNAs were extracted from three pure cultures of ammonia oxidizers, from the starter culture, and from the nitrifying communities obtained from all of the reactors. Biomass subsamples were freeze-thawed three times and then processed by the DNA extraction procedure of Ausubel et al. (2). The concentration and purity of DNA in each sample were estimated by determining the ratio of absorbance at 260 nm to absorbance at 280 nm. DNAs from pure cultures of Nitrosomonas europaea ATCC 25928, Nitrosolobus multiformis ATCC 5976, and Nitrosospira strain NpAV were used as positive controls in PCR, as reference DNA in hybridization tests, and for restriction analyses.

SSU rDNA ARDRA. Amplified SSU rDNAs (19) of the pure cultures and nitrifying communities were digested individually with the following five restriction enzymes: RsaI and Sau3A, obtained from Gibco BRL, Life Technologies, Gaithersburg, Md.; and HaeIII, HinPI1, and BstU1, obtained from New England Biolabs, Ltd. The digested fragments were separated by electrophoresis on 3.5% MetaPhor agarose gels (FMC Bioproducts, Rockland, Maine) in Tris-acetate-EDTA buffer for 4 h at 4°C. Bands were visualized by UV excitation of ethidium bromide-stained gels and photographed. Individual amplified rDNA restriction analysis (ARDRA) patterns were compared by eye, and a similarity index was determined for each treatment by comparing its pattern with the pattern of the starter culture. The similarity index was the ratio of the number of common ARDRA bands after digestion with all five restriction enzymes to the total number of bands in both of the samples analyzed. Very faint ARDRA bands were counted as half bands.

Southern blotting and SSU rDNA probe hybridization procedures. Restricted SSU rDNAs from nitrifying communities were hybridized with the following two phylogenetic probes: Ammo_Cl_2/3/4/6, which hybridizes with all known terrestrial -proteobacterial ammonia oxidizers; and All_Spira, which hybridizes with all known representatives of the Nitrosospira group (6). Pure-culture DNAs of three ammonia oxidizers were used as positive controls for hybridization. The restricted SSU rDNAs were transferred to a Hybond N+ membrane (Amersham Life Sciences Inc., Cleveland, Ohio) (2) and were cross-linked with UV light. The probes were end labeled with [³²P]ATP by using T4 polynucleotide kinase (DuPont NEN Biotechnology Division, Wilmington, Del.). After prehybridization the membranes were hybridized with the ³²P-labeled probes at 42°C for 6 to 18 h and then washed at 42°C in 2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) containing 0.1% sodium dodecyl sulfate and placed in film cassettes for exposure to autoradiogram film (Kodak, Inc., Rochester, N.Y.).

Sequencing of SSU rDNA and phylogenetic analysis. The PCR product (10 µl) from the community maintained in the presence of the highest ammonium concentration was cloned by using the manufacturer's recommended procedure (TA cloning kit; Invitrogen, San Diego, Calif.). The clones were screened for different inserts by digesting reamplified fragments with HaeIII, HinPI1, and BstU1. Five different cloned SSU rDNA fragments (lengths, ca. 1,140 to 1,180 bp) were PCR amplified by using the primers and conditions described above and were sequenced by performing automated fluorescent Taq cycle sequencing with a model 373A DNA sequencer (Applied Biosystems, Foster City, Calif.). Approximately 340 unambiguous nucleotide positions between positions 140 and 508 (Escherichia coli numbering) were used for comparison. Sequences from the nearest relatives were identified and obtained from the Ribosome Database Project (RDP) by using the SIMILARITY_RANK and SUBALIGNMENT programs of the RDP (18). Sequences were aligned manually by using both primary and secondary structures and the GDE editor obtained from the RDP. The levels of similarity of aligned sequences were determined by using the AE2 program obtained from the RDP. Phylogenetic relationships were inferred by the distance matrix method of De Soete (8) by using evolutionary distances estimated by the method of Jukes and Cantor (13).

FAME analysis. The total fatty acid contents of samples were determined by using the protocol developed by MIDI, Newark, Del. Fatty acids were quantified by comparison with known standard fatty acids by using peak width and area. The fatty acid methyl ester (FAME) data were normalized, and the major fatty acids were examined by using the principal-component analysis (PCA) portion of the statistical package S-PLUS (StatSci, Division of MathSoft, Inc., Seattle, Wash.) to find the similarities and differences in fatty acid composition among the experimental cultures.

Nucleotide sequence accession numbers. The sequences determined in this study have been deposited in the GenBank database under accession no. AF043136, AF043137, AF043138, AF043139, and AF043140.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Growth of nitrifying populations under new selective conditions. The effects of ammonium concentration, pH, and oxygen concentration were determined in continuously fed reactors (Table 1). Measurements of ammonium consumption (Table 1) and nitrite and nitrate production during incubation (data not shown) confirmed that nitrifying activity and growth occurred in all reactors. Nitrification, as determined by activity measurements, became established within 3 days after inoculation under all conditions. At the end of the 74-day incubation period the community in the reactor containing 3,000 mg of NH₄⁺-N liter¹ had the highest biomass concentration (0.259 mg of protein/ml). The low-pH reactor and the reactor containing the lowest oxygen concentration exhibited the slowest nitrification, and their biomasses were 3.5- to 6.2-fold lower than the biomass of the fastest-growing community. Dissolved oxygen measurements confirmed that the oxygen concentrations were maintained near 0.43, 3, or 9 mg liter¹ (i.e., 1, 7, or 21% O₂ in the aeration mixtures). Heterotrophic bacteria were present in all of the reactors; the concentrations of heterotrophic bacteria ranged from 1.5 × 10⁷ CFU ml¹ in the low-pH reactor to 6.8 × 10⁷ CFU ml¹ in the reactor containing 1,000 mg of NH₄⁺-N liter¹. Colony morphologies indicated that the different enrichment cultures were dominated by different heterotrophs.

Changes in community structures of nitrifying populations. The ARDRA revealed that the community in the reactor receiving 3,000 mg of NH₄⁺-N liter¹ was most different from the starter culture and that the communities grown at low and high pH values were somewhat different (Fig. 1). We observed new fragments at approximately 900 and 200 bp in the restriction pattern of the community receiving a high concentration of ammonium, while the two fragments at 530 and 370 bp disappeared. The change in the restriction patterns was evident on day 38 of incubation and was greater on subsequent sampling days. The new faint fragment at 200 bp was also present in the patterns of the communities grown at low and high pH values.

View larger version (46K): [in this window] [in a new window]   FIG. 1.   Agarose gel electrophoresis (3% MetaPhor) of restriction digests of the starter culture and the resulting communities (restriction enzyme HinPI1) after 38, 50, and 74 days of incubation under selective conditions. Lanes 100 bp contained the molecular weight standard. For an explanation of the abbreviations see Table 1.

1

1

View larger version (21K): [in this window] [in a new window]   FIG. 2.   Similarity indices of nitrifying communities after incubation under selective conditions, as determined by comparison with the starter culture, which had a similarity index of 1.0. The ARDRA similarity index is the ratio of the number of common electrophoretic bands after digestion with five restriction enzymes (RsaI, HaeIII, HinPI1, BstU1, and Sau3A) to the total number of bands.

Ability of the communities to regain their structures. The three communities that showed structural changes under selective conditions (3,000 mg of NH₄⁺-N liter¹, pH 8.2, and pH 6.0) produced 60 to 70% less (NO₂ + NO₃)-N than the other communities in the first week after they were returned to the standard conditions (data not shown). The amounts of ammonium consumed after 8 days were almost equal in all reactors, ranging from 37 to 75 mg of NH₄⁺-N liter¹ day¹ (Table 1).

1

View larger version (73K): [in this window] [in a new window]   FIG. 3.   Agarose gel electrophoresis (3% MetaPhor) of restriction digests of the communities (restriction enzyme HinPI1) after 34 days of incubation under the original conditions. Lanes 100 bp contained the molecular weight standard. For an explanation of the abbreviations see Table 1.

SSU rDNA probe hybridization. The All_Spira probe, which was designed to detect ammonia oxidizers belonging to the Nitrosospira group (including the genus Nitrosolobus), hybridized with pure-culture DNA of Nitrosospira strain NpAV and Nitrosolobus multiformis ATCC 5976 and not with the restriction products of the cultures from the 10 experimental treatments. The general probe Ammo_Cl_2/3/4/6, which was designed to hybridize with all known terrestrial ammonia oxidizers belonging to the -proteobacteria, produced strong hybridization signals with the restriction fragments from all of the pure cultures tested and from cultures subjected to all of the experimental treatments except 3,000 mg of NH₄⁺-N liter¹. The PCR primers used in this study could have generated amplification products from nonammonia oxidizer DNA, or this community contained one or more ammonia oxidizers that are different from the ammonia oxidizers already known. On the basis of the results of the two hybridization tests (All_Spira and Ammo_Cl_2/3/4/6), we estimated that the predominant members of all of the reactor communities except the community grown in the presence of 3,000 mg of NH₄⁺-N liter¹ were ammonia oxidizers belonging to the Nitrosomonas group.

Cloning and sequencing of cloned SSU rDNA PCR fragments. The restriction patterns of 87 cloned SSU rDNA fragments obtained from the community grown in the presence of 3,000 mg of NH₄⁺-N liter¹ revealed five different clones of ammonia oxidizers. The predominant pattern, pattern Al-7K, accounted for 74% of the clones. A clone that was representative of each of the five ARDRA patterns was partially sequenced. All of the sequences clustered in the Nitrosomonas group (Fig. 4). The sequence that was most dissimilar to the database among these five clones was Al-7K; this sequence was 92.6% similar to the sequence of Nitrosomonas eutropha C-91, the closest relative in the database. Thus, the phylogenetic analysis revealed a previously unknown sequence type produced by a member of the -proteobacterial ammonia oxidizers that has not been cultured. The neighboring clones, Al-8H and Al-8B1, as well as Al-9K3, differed by only 0.5 to 1.5% from the dominant clone Al-7K. The rDNA sequence of minor clone Al-8N was very similar to the rDNA sequence of Nitrosomonas eutropha C-91 (98.5% similarity).

View larger version (27K): [in this window] [in a new window]   FIG. 4.   Phylogenetic tree of the clones from the nitrifying community incubated in the presence of 3,000 mg of NH4-N liter1 and their nearest relatives. The predominant clone in this community was Al-7K.

Fatty acid analysis of the nitrifying enrichment cultures. The fatty acid compositions of the starter culture and reactor communities incubated under selective conditions, as well as under standard conditions, varied with respect to the levels of three fatty acids that are commonly found in nitrifiers, 16:0, 16:17c, and 18:1 (data not shown) (3). Some reactor communities grown under the selective conditions contained small amounts of up to nine other fatty acids as well. We used multivariate PCA to expose differences in major fatty acid fractions among the experimental cultures (Fig. 5). This analysis showed that the nitrifying communities of the reactors under selective conditions produced different FAME profiles than the starter culture (the data for the pH 6.0 treatment is not included in the PCA because the sample was lost during extraction). The profiles that diverged the most were obtained for the community grown at a high pH and for the community grown in the presence of 500 mg of NH₄⁺-N liter¹. As expected, the community grown at pH 7.0 had a FAME profile similar to the FAME profile of the starter culture. The PCA also showed that all of the reactor communities returned to the starter culture FAME profile after the original growth conditions were restored.

View larger version (17K): [in this window] [in a new window]   FIG. 5.   PCA of FAMEs from the starter culture and from nitrifying communities incubated under selective conditions and after restoration of the standard conditions. The ellipse indicates the 95% confidence limit of selected data points. For an explanation of the abbreviations see Table 1.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Ammonium concentration, oxygen concentration, and pH are thought to be the environmental parameters most important to the nitrification rate and also likely to determine the nitrifier community selected. We found that ammonium at a very high concentration (3,000 mg of NH₄⁺-N liter¹) selected a novel nitrifier population and that the pH extremes tested, pH 6.0 and 8.2, selected a somewhat altered community, but the other conditions did not result in community shifts detectable by ARDRA. The community shift caused by a high ammonium concentration occurred gradually, increasing at each sampling time up to 74 days (Fig. 2). Apparently, a nitrifier population better adapted to the high ammonium concentration slowly outgrew the original members of the community. The shift in population composition was not apparent from the nitrification rates since the ammonium consumption rates were rapid and equal before and after the structural change. This shows that the original, probably more conventional nitrifiers were also quite active in the presence of a high ammonium concentration. Nonetheless, they were eventually replaced by a more adapted strain.

The structural changes brought about by the pH extremes took longer to develop. The nitrification rate in the low-pH environment was retarded, but the nitrification rate in the high-pH environment was not retarded. This suggests that in the former environment the shift was delayed by the generally unfavorable conditions even for the newly dominant nitrifiers, while in the latter environment the selective advantage of the newly dominant group must have been minor.

The ammonia oxidizer population selected in the reactor containing the high ammonium concentration seemed to be substantially different from the nitrifiers in the other reactors and from the nitrifiers described previously since the shift was detected with each of the five restriction enzymes used. This difference in the populations was confirmed by the lack of hybridization to the general ammonia oxidizer rDNA probe and by the finding that the rDNA sequence similarity between the dominant operational taxonomic unit (OTU) and all other ammonia oxidizers in the database was only 92.6%. The sequence of this clone and the sequences of clones belonging to three other minor OTUs varied by 0.4 to 1.5%. Together, these clones appear to represent a cluster of organisms specialized for very high ammonium concentrations (Fig. 4). Clone Al-7K of the dominant OTU might be a member of a new species since strains with SSU rRNA evolutionary distances that differ by more than 2.5 to 3.0% have been found to be members of different species (24). The evidence that clone Al-7K is actually a nitrifier rests on the facts that it branches within the family of ammonia oxidizers in the -proteobacteria and that it was the dominant clone recovered by primers for this family from a highly active nitrifying community. No nonnitrifiers have been found yet in this family. The high ammonium concentration which we used (3,000 mg of NH₄⁺-N liter¹), although unusual for natural environments, can be found in animal wastewaters (5). Biological treatment of such waste streams is an important practical problem. Hence, finding nitrifier strains adapted to high ammonium concentrations may have some value in treatment of high-strength ammonium wastes.

The nitrifier populations in all of the other reactors, including the starter culture, appeared to consist of Nitrosomonas-like nitrifiers since their rDNA hybridized strongly to the general ammonia oxidizer family probe but not to the Nitrosospira family probe. This was expected since Nitrosomonas strains are the most common type of ammonia oxidizers found in wastewaters (28, 30). Seven phylogenetic clusters of ammonia oxidizers belonging to the -proteobacteria are currently recognized (25), and many new ammonia oxidizer sequences from different environments have recently been described (6, 11, 15, 25). These investigations showed that Nitrosospira types are the most common ammonia oxidizers in soil and freshwater, not Nitrosomonas types, as was previously thought based on culture-based studies.

The evidence of Suwa et al. (26) suggests that there is some correspondence between ammonia oxidizer sensitivity or tolerance to ammonia and phylogeny, a result also noted in this study at least for very high ammonium concentrations (26). So far, there is little evidence for a similar correspondence between ammonia oxidizer type and pH or oxygen status. In a recent study researchers found closely related Nitrosospira sequences in both neutral and acid soils, although some sequences might have been more common in one soil type than in the other (25). In contrast, Kowalchuk et al. found different nitrifier sequence types in acid and alkaline Dutch dune sites (15).

An important finding of this study is that the community which exhibited the greatest structural shift (similarity index, 0.39) was able to reacquire its original structure after the selection conditions were eliminated. The time needed for recovery (8 to 15 days) was relatively short. The speed of recovery was probably aided by the fact that we cleaned and reinoculated the reactors when the conditions were changed, which reduced the residual biomass of the community grown in the presence of the high ammonium concentration. Recovery depended on the selection conditions, however, since the communities selected at pH 6.0 and 8.2 did not return to their original structures when the original conditions were reestablished. Restoration of the original community structure is not usually expected in microbial ecology because the high diversity in most habitats usually leads to many community structures with virtually the same functions. In this case, however, we were dealing with a community from wastewater that was already highly selected before the experiment was started. Hence, the probability of restoring the original structure of a simpler community is higher.

In addition to the nucleic acid-based methods used in this study, we also used a biochemical method to analyze the nitrifying communities. The three major fatty acids which we found are common but not unique to nitrifiers; 16:0 and 16:1 are found in nitrite oxidizers, and 18:1 is found in ammonia oxidizers (31). The PCA of these three fatty acids, as well as the eight major fatty acids, showed that the communities in all of the reactors incubated under selective conditions diverged from a starter culture but that all of the communities, including the communities in the reactors containing high ammonium concentrations and the reactors at extreme pH values, returned to their original states (Fig. 5). The FAME analysis, however, encompassed the entire community, including the nitrite oxidizers and heterotrophs. Since the numbers of heterotrophs in all of the communities were similar (10⁷ CFU ml¹), the FAME profiles for the treatments which resulted in low biomasses (e.g., 1% oxygen and 50 mg of NH₄⁺-N liter¹) could have resulted from relatively high proportions of heterotrophs. The FAME analysis appeared to be much more sensitive than ARDRA for revealing the community shifts since the shifts were detected by the former method under all treatment conditions. This is consistent with the finer level of resolution (e.g., species level resolution) of microbial taxa provided by FAME analysis. The fact that ARDRA did not detect recovery in the two reactors incubated at the pH extremes but FAME analysis did could be due either to physiological adaptation to pH by the existing populations or to the fact that the FAME analysis reflected the fatty acids of the entire community, including nitrite oxidizers and heterotrophs, and that these organisms and not the ammonium oxidizers returned to the original composition.