Introduction

Urban wastewater (UWW) treatment is one of the major biotechnological processes used worldwide (Wagner et al. 2002). In countries of the European Union (EU), these effluents are regulated by the Central and Eastern Europe 271/91 directive. Fixed film technology is broadly used for the removal of organic matter and nitrogen (N) from UWW through the biological process of nitrification-denitrification (Gálvez et al. 2003; Hem et al. 1994; Lynga and Balmer 1992; Sakano et al. 2002). Among other advantages compared to the more generally applied activate sludge technology, submerged filters are simple to control and maintain, reduce room usage, lower costs, and minimize odors and noise (Pujol et al. 1994).

The physiology and biochemistry of both nitrifying and denitrifying bacteria are extensively known, however, the information about the ecology of these organisms is still scarce (Bothe et al. 2000). The introduction of molecular techniques in microbial ecology significantly implemented the knowledge of these organisms in their habitats as it enabled the study of the nonculturable bacteria. Muyzer et al. (1993) introduced the denaturing gradient gel electrophoresis (DGGE) technique in the field of microbial ecology. This method and its homologous temperature-gradient gel electrophoresis (TGGE) allow the separation of DNA fragments of similar size depending on their nucleotide sequence. DGGE was widely applied to the study of the population dynamics of microorganisms in natural habitats (Muyzer 1999; Muyzer and Smalla 1998) and in recent years it has considerably helped to broaden the knowledge of the microbial communities of biological systems for the depuration of UWW, particularly the ammonia oxidizers (for a review see Bothe et al. 2000; Wagner et al. 2002).

In this study, we used a pilot-scale UWW treatment plant consisting of two submerged filters working in predenitrification mode. The first one worked under anoxic conditions, keeping the denitrifying activity by the use of inlet UWW as carbon source, while the second one worked aerobically oxidizing organic nitrogen and ammonia to nitrate. Biological treatment of UWW by the submerged filter technology needs to be performed under the predenitrification mode, in order to be economically advantageous. This type of system was previously applied with success in our laboratory under different working conditions for the removal of suspended solids, organic matter, and total N from UWW (Calvo et al. 2001; Gálvez et al. 2003). We explored in this work the potential of the TGGE approach on the monitoring of the spatial and temporal variations of the microbial diversity along the described UWW treatment system. For this purpose, TGGE was used to separate DNA fragments after amplification from total DNA extracted from the biofilms by primers targeting the V3 hypervariable region of the 16S rRNA gene (Muyzer et al. 1993). In addition, the structure of the nitrifying and denitrifying microbial communities was further studied by the amplification and separation of genes encoding for the functional proteins ammonia monooxygenase (amoA) and nitrous oxide reductase (nosZ) after the subsequent identification of the bands by DNA sequencing. Characterization of specific bacterial members in the bioremediating beta subclass ammonia-oxidizing microbial community was achieved by this approach as phylogeny based on amoA/AmoA provides highly similar orders of taxa as phylogeny based on 16S rRNA (Purkhold et al. 2000). The nosZ gene was chosen as a marker in this work because its high diversity in natural habitats was consistently reported in recent literature (Scala and Kerkhof 1998, 1999; Rösch et al. 2002; Throbäck et al. 2004).

Materials and methods

Description of the pilot-scale UWW treatment plant

The pilot-scale plant used for these experiments is depicted in Fig. 1 and was based on a design already used successfully for UWW treatment in previous studies (Gálvez et al. 2003). It consisted of two methacrylate cylindrical columns (each 2 m high and 6 cm in inner diameter) both packed with clayey schists about 5–7 cm in size, 1.75 g/cm3 in density, and up to 1 m in height. The columns are connected with a valve that allows a separated cleaning of the biofilters. To avoid filter clogging, cleaning cycles of the biofilters were carried out every 3 days. Hydraulic charge was kept at 0.354 m3 m−2 h−1 and the nitrifying column was supplied with an airflow of 4.2 m3 m−2 h−1. Water flow was regulated by a peristaltic Watson Marlow 505S pump. The system was operated on a predenitrification mode with a recycle rate of 1,000 ml/h. The biofilters operate downflow (denitrifying column, anoxic) and upflow (nitrifying column, oxic) fed with UWW coming from primary treatment of the UWW treatment plant “EDAR Churriana” (EMASAGRA, Granada, Spain). The average composition of wastewater was determined by standard methods (APHA 1995) and was composed of the following: biochemical oxygen demand (5 days), 135 mg/l; chemical oxygen demand (COD), 450 mg/l; total N, 50 mg/l; and suspended solids, 120 mg/l. Under these working conditions, the removal of organic matter reached 75% and the COD in effluent water was never more than 125 mg/l, as required by the EU Directive 271/91, while a rate of 60% removal of total N was achieved. This level of N elimination is close to the theoretical maximum that can be achieved for the low COD/N ratio of the influent water (Van Loosdrecht and Jetten 1998).

Fig. 1
figure 1

Diagram of the pilot plant used in the study showing the disposition of anoxic and aerated columns and the sampling ports at different locations in the plant (25D, 75D, 25N, and 75N)

Full size image

Biofilm recovery and DNA extraction

Total DNA was periodically extracted from biofilm samples recovered from four locations of the reactor (sampling points 25DN, 75DN, 25N, and 75N as depicted in Fig. 1). Biofilm recovery was achieved by the following method: 5–50 g of clayey schists with biofilm adhered were extracted from each sampling point (see Fig. 1) and placed in flasks with 50 ml of sterile saline (0.9% NaCl). These suspensions were sonicated for 1 min and then placed in an orbital shaker at 155 rpm for 1 h. The process was repeated twice. Finally, 100 ml of saline with suspended material from the biofilm was centrifuged at 3,500 rpm for 30 min. The pellets were immediately used for DNA extraction, following a method based on the one previously described by Watanabe et al. (1998). Pellets were suspended briefly in 1 ml of buffer [10 mM Tris–HCl (pH 8.0), 1 mM EDTA, 0.35 M sucrose, and 20 mg of lysozymes per milliliter] and incubated at 37°C for 10 min. After incubation, 1.5 ml of lysis buffer [100 mM Tris–HCl (pH 8.0), 0.3 M NaCl, 20 mM EDTA, 2% (wt/vol) sodium dodecyl sulfate, 4 μg/ml proteinase K, and 2% (wt/vol) 2-mercaptoethanol] was added and the samples were incubated at 37°C for 1 h. Lysates were extracted twice with 1 vol of phenol–chloroform solution (1/1 vol/vol), incubated for 10 min with 125 μg/ml RNAase, and extracted once with 1 vol of chloroform. Aliquots of 800 μl were precipitated with 1 vol of 2-propanol (30-min incubation at −20°C followed by centrifugation for 20 min at 13,000 rpm). The pellets were washed with 500 μl of 70% ethanol solution before being resuspended in 50 μl of Tris EDTA buffer and was left to rehydrate at 4°C overnight. DNA was further purified by a new phenol–chloroform extraction and reprecipitation the following day before being stored at −20°C.

PCR

Three genes were used as target for the study of bacterial diversity: the 16S rRNA gene, the amoA gene for beta subclass ammonia-oxidizing bacteria, and the nosZ gene for denitrifying bacteria. Fragments of the three genes of a size adequate for TGGE separation (ca. 200 bp for 16S rRNA, 475 bp for amoA, and 250 bp for nosZ) were amplified by polymerase chain reaction (PCR) using nested or seminested approaches as these strategies were previously reported to avoid unspecific priming, improve amplification of species present in low numbers, and yield richer DGGE band patterns (Nicolaisen and Ramsing 2002; Ward and O’Mullan 2002). The primers used were previously described in literature and their references are listed in Table 1. All primers were purified by high-performance liquid chromatography and purchased from Sigma-Aldrich. One microliter (2–5 ng) of DNA was used as template for PCR and subsequently, 1 μl of the first PCR product was used as template for the nested PCR. Conditions for each of the PCR reactions are summarized in Table 1. Final PCR products were cleaned and/or concentrated (when required) using Microcon-YM cartridges. Two to 5 μl were loaded in each well for TGGE.

Table 1 PCR conditions applied in each experiment described in the “Material and methods”
Full size table

TGGE and DNA sequencing

TGGE was done on a TGGE Maxi system (Whatman Biometra). Denaturing gels (6% polyacrylamide gel electrophoresis with 20% deionized formamide, 2% glycerol, and 8 M urea) were made and run with 1×Tris–acetate–EDTA buffer. Temperature gradients were optimized for efficient separation of bands and were as follows: 43 to 63°C for the 16S rRNA, 38 to 52°C for the amoA, and 47.5 to 54.5°C for the nosZ gene. The gels were run at 115 v for 18 h except the gels for nosZ, which were run for 24 h. Gel bands were visualized either by staining with SYBR-green I (Sigma-Aldrich) for 20 min or by silver staining using the Gelcode silver staining kit (Pierce) following the manufacturer’s indications except for the omission of the stabilization step. SYBR-green I stained gels were analyzed with a Gel Doc 2000 System (BioRad) and silver-stained gels were photographed with a Canon digital camera.

Portions of individual bands on silver-stained TGGE gels were picked up with sterile pipette tips, placed in 10 μl of sterile PCR water, and directly used for reamplification with the appropriate primers. PCR products were purified by gel running and extraction with the QIAEX II kit (Qiagen). The DNA recovered was used for automated sequencing in an Applied Biosystems 3100 capillary automatic sequencer using the appropriate primers.

Biocomputing and phylogenetic analysis

DNA sequences were analyzed using the biocomputing tools provided online by the European Bioinformatics Institute (http://www.ebi.ac.uk). The BLASTn, BLASTx (Altschul et al. 1997), and FASTA (Pearson and Lipman 1988) programs were used for preliminary sequence similarity analysis. Phylogenic inference from amoA/AmoA and nosZ/NosZ was calculated using the CLUSTALX V.1.8 software (Jeanmougin et al. 1998) for the aligning of sequences, calculation of distance matrixes, and construction of bootstrapped neighbor-joining trees (Felsenstein 1985; Saitou and Nei 1987). Gaps were excluded from the analysis and the Kimura correction was applied (Kimura 1983). Bootstrap values below 50% are not shown in trees.

Results

Bacterial community-profiling by TGGE

TGGE profiles which were based on the V3 region of the 16S rRNA gene were produced from samples taken during three consecutive weeks and after 1 year of plant operation. These profiles reflected a high uniformity of the populations especially in the anoxic biofilm while the aerated column showed significant variation of the band patterns in the samples taken with a 1-year interval (data not shown). Samples taken at different locations of the column system (25DN, 75DN, 25N, or 75N) also reflected a spatial variation only in the aerated part of the system. As expected, the profiles obtained had a high complexity but it was possible to reamplify and sequence eight dominant bands whose affiliations are shown in Table 2. It is remarkable that the sequence of a band which predominated in samples from the aerated biofilm, but not from the anoxic biofilm, was found related to the nitrite oxidizer genus Nitrospira, confirming a major presence of these bacteria in the aerated part of the biofilm and suggesting that these group of organisms are the most likely candidates responsible for the oxidation of nitrite to nitrate in the system used for this study.

Table 2 Comparison of the 16S rDNA sequences of major bands isolated from TGGE gels (ca. 160 bp) and sequences of the European Molecular Biology Laboratory database
Full size table

Composition of microbial communities of functional microbial groups—phylogeny inferred from NosZ and AmoA

TGGE profiles of amoA and nosZ genes showed the presence of ammonia-oxidizing and denitrifying bacteria in both the anoxic and aerobic filters. The community profiles of denitrifying bacteria remained very stable over space and time (Fig. 2a), while profiles of ammonia oxidizers greatly varied at different sampling times and locations in the reactor (Fig. 2b). Several bands separated by TGGE were reamplified and sequenced to analyze the composition of these microbial communities in the biofilm.

Fig. 2
figure 2figure 2

TGGE diversity profiles and phylogenetic neighbor-joining trees based on partial NosZ (a) and AmoA (b) proteins (150 and 70 AA-aligned, respectively). Lanes A, B, and C: samples taken from the anoxic column during three consecutive weeks. Lanes G, H, and I: samples taken from the aerated column during three consecutive weeks. Lanes D and J: samples taken after 1 year of plant functioning (anoxic and aerated columns, respectively). Lanes E, F, K, and L: samples taken from 25DN, 75DN, 25N, and 75N points, respectively, after 1 year of plant functioning. Bands numbered pointed with arrows in each gel (Amo or Nos) correspond to bands numbered in each phylogenetic tree

Full size image

A total of 43 different bands were extracted from nosZ TGGE gels from both the anoxic and the aerated parts of the submerged filter biofilm, although only 18 partial nosZ sequences were successfully read in their entirety, the rest either did not amplify successfully or showed the presence of overlapped sequences. Several individual bands yielded identical DNA sequences (Nos19 and Nos23; Nos6 and Nos6B; Nos9 and Nos10; and Nos1, Nos2, Nos3B, Nos22, and Nos24). These high occurrence of redundancies were most probably due to PCR artifacts (heteroduplexes), favored by the combination of a mixed template and the use of degenerated primers for amplification. None of the 11 deduced NosZ sequences showed a 100% identity with the sequences on databases, but most of them was related to those of gamma-proteobacteria (genus Pseudomonas), while the rest clustered in the periphery of previously known denitrifying alpha-proteobacteria (Bradyrhizobium and Azospirillum) (Fig. 2a).

A total of 21 bands were extracted from the amoA gels, where 14 of them was successfully reamplified and fully sequenced. Again, some bands physically separated on the gel yielded identical DNA sequences, but with a much lower incidence compared to the nosZ gels (Amo27 and Amo28; Amo7, Amo25, and Amo26). Several bands have different DNA sequences but translated into identical protein sequences (i.e., Amo2, Amo3, Amo4, and Amo5 bands).

Phylogeny inferred from the partial amoA/AmoA sequences revealed the prevalence of Nitrosomonas species in the UWW submerged filter treatment system with five band sequences (Amo2, Amo3, Amo4, Amo5, and Amo14) affiliated to Nitrosomonas oligotropha, six band sequences (Amo7/Amo25/Amo26, Amo8/Amo9, and Amo22) affiliated to Nitrosomonas europaea, and three band sequences (Amo27, Amo28, and Amo29) that showed only 75.7–76.1% similarity of the DNA sequence with the closest described species (Nitrosomonas nitrosa). However, the Amo27/Amo28 and Amo29 protein sequences showed a similarity of 100 and 99%, respectively, which are identical to the sequence of AmoA of unidentified bacteria GLII-20 (accession No. AAG39794.1) that was also isolated from a nitrifying wastewater treatment plant (Purkhold et al. 2000) and to sequence ST-A-mRNA-1 (accession No. BAD12250.1) that was isolated by DGGE from mRNA extracted from a septic tank (Ebie et al. 2004). Another two sequences also isolated in the aforesaid study (BAD12251.1 and BAD12252.1) gave high identity values with sequences Amo27/Amo28 and Amo29. According to Purkhold et al. (2000), the low identity values with any previously known Nitrosomonas indicate that this branch may represent a new species.

Discussion

Research on the use of biofilters for UWW management has traditionally focused on changes on water parameters, mostly organic matter disappearance and N elimination. However, the study of the bacterial ecology of these systems is needed to get a more detailed picture of the interaction between its biology and chemistry as a way to improve the necessary knowledge to progress in the design and management of treatment plants.

In this study, we appraised the use of the TGGE approach for the profiling and characterization of the bacterial communities in a UWW treatment plant based on submerged biofilters whose efficiency for the removal of COD, total N, and pathogenic bacteria was proven in previous studies (Calvo et al. 2001; Gálvez et al. 2003). When the method was founded on the amplification of the hypervariable region V3 of 16S rRNA gene, this approach has proven to be a suitable tool for the spatial and temporal monitoring of the community profiles of the fixed biofilm used in the study. Community structure monitoring is particularly required by UWW plants expected to suffer shock loads of pollutants, which often cause the breakdown of the community structure of the biological systems and the loss of its normal mineralization activity (Eichner et al. 1999; Watanabe et al. 1999). The results reported here support the value of TGGE as a method that could be applied in future research to follow the effects of relevant external factors, such as fluctuations on inlet water composition, on the community structure of the UWW submerged filter system used in this study.

The wide distribution of the denitrification activity within phylogenetically distant bacterial species makes the design of universal primers suitable for molecular studies of this functional microbial group unfeasible. The degenerated primers used in this study were previously designed by Scala and Kerkhof (1998) who suggested their potential utility for future DGGE/TGGE studies. We encountered several difficulties in the use of these degenerated primers for TGGE-based diversity studies. The approach yielded a number of new sequences of the nosZ gene, but the frequent finding of redundant sequences pointed out that the band patterns can heavily overestimate the diversity of nosZ copies in the analyzed samples, hence, such patterns must be interpreted cautiously in terms of evaluation of diversity. The occurrence of nosZ copies phylogenetically close to those of well-known denitrifyers was also found in a previous study using an equivalent fixed biofilm based on the denitrification-nitrification process (Sakano et al. 2002). However, copies of nosZ recovered from natural environmental samples are often unrelated to any previously known sequences (Scala and Kerkhof 1999).

Based on the amoA sequences obtained from reamplified TGGE bands, we identified fairly diverse ammonia-oxidizing populations in the biofilm of the UWW system, although all partial sequences were affiliated to Nitrosomonas-like amoA, and no Nitrosospira-related sequences were found. The predominance of ammonia oxidizers of the genus Nitrosomonas over Nitrosospira was commonly reported in previous studies of the ammonia oxidizer communities on UWW plants based on the activated sludge technology (Purkhold et al. 2000).

TGGE coupled to reamplification and sequencing of bands allowed the temporal and spatial monitoring of the community structure of the UWW plant used in this study, and provided characterization of the predominant populations of ammonia-oxidizing and denitrifying communities of the fixed biofilm involved in the process. The TGGE approach offers advantages over cloning methods, which require a greater amount of time and money, providing an outcome of equivalent validity on the characterization of the populations. The major disadvantage of TGGE is the limitation of the size of the DNA fragments and the number of bands that can be resolved on the surface of the gel. A key drawback met in this particular study were the artifacts due to mixed template PCR reactions, especially when the use of degenerated primers could not be avoided (nosZ–based studies).