A systematic insight into a single-stage deammonification process operated in granular sludge reactor with high-loaded reject-water: characterization and quantification of microbiological community

Authors

  • S.T. Liu,

    Corresponding author
    1. Department of Environmental Engineering, Peking University, The Key Laboratory of Water and Sediment Sciences, Ministry of Education, Beijing, China
    • Institute of Water Quality Control, Technische Universität München, Garching, Germany
    Search for more papers by this author
  • H. Horn,

    1. Karlsruhe Institute of Technology (KIT), Engler-Bunte-Institute, Water Chemistry and Water Technology, Karlsruhe, Germany
    Search for more papers by this author
  • E. Müller

    1. Institute of Water Quality Control, Technische Universität München, Garching, Germany
    Search for more papers by this author

Correspondence

Sitong Liu, Institute of Water Quality Control, Technische Universität München, Am Coulombwall, D-85748 Garching, Germany. E-mail: s.liu@bv.tu-muenchen.de

Abstract

Aims

Reject-water from sludge dewatering was treated in a single-stage deammonification reactor. The aims were to characterize the microbiological community within deammonification granules.

Methods and Results

In situ mapping of the intact granular sludge were made with fluorescent in situ hybridization (FISH). The Planctomycetes community in the destroyed granular sludge was characterized by FISH, polymerase chain reaction-denaturing gradient gel electrophoresis (PCR-DGGE), cloning and sequence analysis.

Conclusions

AOB within Betaproteobacteria were concentrated within the first 50–100 μm of the granule, and Planctomycetes in the first 100–200 μm were responsible for nitrogen elimination. PCR-DGGE verified the higher diversity of Planctomycetes in the deammonification reactor than the anaerobic cultivation. The sequence analysis after PCR-DGGE and cloning identified the dominant Planctomycetes species for anammox reaction as Ca. Brocadia fulgida (accession no. EU478693). FISH detection using the universal probe AMX368 specific for all anammox bacteria including Ca. Brocadia fulgida failed; however, the probe BFU613 specific for Ca. Brocadia fulgida gave clear positive FISH signals. The three-dimensional structure of the ribosome may hinder binding of the universal probe to the corresponding 16S rRNA region.

Significance and Impact of the Study

A combination of multiple methods for the analysis of the microbiological community was necessary. Oligonucleotide probes should be carefully selected, and a negative FISH analysis has to be verified by other molecular biological techniques.

Introduction

As a result of the increasingly stringent effluent regulations and desired lower energy requirements for high-loaded industrial wastewater treatment, many novel nitrogen removal technologies such as completely autotrophic nitrogen removal over nitrite (CANON) (Sliekers et al. 2002), oxygen-limited autotrophic nitrification and denitrification (OLAND) (Kuai and Verstraete 1998) and single-stage nitrogen removal using anammox and partial nitritation (SNAP) (Furukawa et al. 2006) have received increased attention. Such technologies have already been implemented at the industrial scale in Rotterdam (the Netherlands), Hattingen (Germany), Werdhölzli (Zürich), Strass (Austria) and Ghent (Belgium) (Thöle et al., 2005; Wett 2007; van der Star et al. 2007; Joss et al. 2009; Desloover et al. 2011). As reported recently, successful application of the anammox technology can potentially change wastewater treatment from an energy-consuming process into an energy-yielding process (Abma et al. 2010; Kartal et al. 2010). Among these novel nitrogen removal technologies, deammonification has become a hot research point. It is cost-effective and needs less energy compared with other technologies. Deammonification was first described by Abeling and Seyfried at the beginning of 1990s (1992). It requires an aerated system and appropriate process control and is characterized by nitrogen elimination in an oxygenated environment without organic carbon consumption. The advantages of this process are that less oxygen is required, there is no need for an extra carbon source, and sludge production is greatly reduced all while maintaining the potential for high removal efficiency (Abeling and Seyfried 1992; Seyfried et al. 2001; Wett 2007).

It has been widely accepted that deammonification is achieved via sequential partial nitrification and that anammox is carried out by autotrophic aerobic ammonium oxidizers (AOB) within the Betaproteobacteria and anammox bacteria within the Planctomycetes (Hippen et al., 1997; Maslon and Tomaszek 2007). So far, there are eight species belonging to Planctomycetes that have the ability to carry out the anammox reaction. They are as follows: Ca. Brocadia anammoxidans, Ca. Kuenenia stuttgartiensis, Ca. Brocadia fulgida, Ca. Scalindua wagneri, Ca. Scalindua brodae, Ca. Scalindua sorokinii, Ca. Anammoxoglobus propionicus and Planctomycete KSU-1 (Jetten et al. 1999; Strous et al. 1999; Fujii et al. 2002; Kuypers et al., 2003; Schmid et al. 2000, 2003; Kartal et al. 2008). Deammonification has been used to treat a variety of wastewaters of differing complexities (Maslon and Tomaszek 2007; Wett 2007; Joss et al. 2009). Therefore, the microbiological information may be broad and differ from one reactor to another. The cultivation and isolation of anammox bacteria is very difficult, and consequently, rRNA-based molecular techniques for the cultivation-independent detection of these organisms are very helpful. Fluorescent in situ hybridization (FISH), polymerase chain reaction (PCR), denaturing gradient gel electrophoresis (DGGE), real-time polymerase chain reaction (real-time PCR), cloning and phylogenetic sequence analysis are widely used to detect and identify these bacteria in wastewater treatment systems (Nakajima et al. 2008; Jarvis et al. 2009; Xiao et al. 2009; Vejmelkova et al. 2011). For years, these approaches have been developed, and the number of available rRNA biomarkers for the anammox bacteria species has increased (Li and Gu 2011).

These molecular biological techniques enable a better understanding of the microbial community involved in the deammonification process. Helmer-Modhok et al. (2002) applied FISH to elucidate the composition and spatial structure of the microbial community responsible for the deammonification process. They verified the presence of ammonium oxidizers belonging to the beta-subclasses Proteobacteria and Planctomycetes. FISH analysis is a powerful tool that can be used to investigate the spatial distribution of AOB and anammox bacteria in complex microbial aggregates such as granular sludge (Vlaeminck et al. 2010). In a study by Jeanningros et al. (2010), it was shown that the aerobic ammonium oxidizers including Nitrosomonas halophila, Nitrosomonas eutropha and N. halophila, and the anaerobic ammonium oxidizers ‘Candidatus Kuenenia and Brocadia’ were related to nitrogen removal in the deammonification process. In addition, PCR-DGGE combined with 16S rRNA gene analysis was also shown to be a useful tool to evaluate the microbial development and diversity shifts in a deammonification reactor operated under different conditions. Using this technology, Innerebnera et al. (2007) analysed the partial 16S rRNA gene sequences obtained from the deammonification granules and the deammonification flocs. They were able to confirm differences in the community composition of the two fractions. However, the combined application of these 16S rRNA methods to obtain more detailed information about the microbiological community involved in the deammonification process still needs to be demonstrated.

In this study, a sequencing batch reactor (SBR) with granular deammonification sludge was designed to treat reject-water from sludge dewatering originating from the municipal wastewater treatment plant (WWTP) in Garching, Germany. This study reports the following results: (i) mass balance analysis for N and COD removal mechanisms; (ii) in situ distribution analysis focusing on AOB and Planctomycetes within the deammonification granular sludge and (iii) detection and phylogenetic identification of Planctomycetes at the genus and species level involved in the deammonification process by combined application of 16S rRNA methods including FISH, PCR-DGGE, cloning, sequencing and phylogenetic sequence analysis to obtain more detailed information.

Materials and methods

Reactor set-up

The SBR used had a maximum working volume of 10 l with a height of 400 mm and an inner diameter of 100 mm. The reactor was sequentially operated in an 8-h cycle consisting of filling, anoxic and aerobic phases, which were repeated five times and followed by a final process of anoxic, aerobic, settling and withdrawal phases controlled by a programmable logic controller system. A mechanical stirrer was used to mix sludge and reject-water. An aeration device having a ceramic membrane was operated at an air pressure of 0·05 bar with a maximum oxygen flow rate of 20 l h−1. The hydraulic retention time was kept constant at 3 days. Online measurement of temperature, pH and dissolved oxygen (DO) was carried out with a data acquisition system (Fluke Hydra, Everett, WA, USA). Table 1 outlines the specific reactor characteristics and operation conditions.

Table 1. Reactor characteristics and operation conditions
ItemsValues
Effective reactor volume10 l
Stirrer diameter0·1 m
Stirrer speed80–90 rev min−1
Influent math formula-N concentration 1082 ± 100 mg l−1
Influent math formula-N concentration 0 mg l−1
Influent math formula-N concentration 1 ± 1 mg l−1
C/N in the influent0·7 ± 0·1
Hydraulic retention time3 days
Ratio of anoxic time/aerobic time0·97
pH in the reactor7–8
Dissolved oxygen (DO) in the aerobic period0·6 ± 0·1 mg l−1
DO in the anaerobic period0 mg l−1
Temperature in the reactor30 ± 1°C
Operation period 3 months

Inoculum and influent of the SBR

The inoculum used in this study originated from a laboratory-scale deammonification SBR operated for 6 months, which was originally inoculated with two of three municipal sludge from the WWTP Garching and one of three deammonification sludge from a full-scale WWTP at Werdhölzli, Zürich (Joss et al. 2009). The formation of granular sludge in the SBR was achieved within 3 months. Reject-water from sludge dewatering originating from the municipal WWTP Garching, Germany, was used as the influent and mainly contained math formula-N and COD. The nitrogen and COD concentrations in the influent are shown in Table 1.

Chemical analysis

math formula-N, math formula-N, math formula-N and COD were photochemically (Photometer Dr 2800) measured using the Standard Dr Lange tests LCK303, LCK342, LCK 339 and LCK514 (Hach Lange GmbH, Düsseldorf, Germany), respectively. Total suspended solid was determined according to standard methods (APHA 1998).

Anaerobic cultivation

To enrich anammox bacteria, anaerobic cultivation in a 500-ml plastic bottle using the aforementioned inoculum was carried out for 2 months. The media contained 100 mg l−1 math formula-N and 100 mg l−1 math formula-N as substrate. In addition, a trace element solution was used according to the study by van de Graaf et al. (1996) and was comprised of KHCO3 1·25 g l−1, KH2PO4 0·025 g l−1, CaCl2·2H2O 0·3 g l−1, MgSO4·7H2O 0·2 g l−1, FeSO4 0·00625 g l−1 and EDTA 0·00625 g l−1. After the consumption of math formula-N or math formula-N, the supernatant was removed via centrifugation and fresh media was supplied.

Sampling for chemical and microbiological analysis

Influent samples and effluent samples were taken from the deammonification reactor twice a week and analysed for math formula-N, math formula-N, math formula-N and COD. For the tracking of nitrogen and COD concentrations, water samples were taken after each phase of one complete cycle to measure math formula-N, math formula-N, math formula-N and COD.

Sludge samples from the deammonification reactor were collected on days 20, 40 and 80 (R20, R40 and R80) and analysed with FISH to determine the composition of AOB and anammox bacteria. Simultaneously, the sludge samples were also analysed with PCR-DGGE to describe the development of anammox bacteria. At the end of the operational period (day 90), the deammonification granular sludge samples were taken and cryosections were prepared for in situ mapping of the AOB and anammox bacteria in the granules by FISH.

Sludge samples from the anaerobic cultivation were taken on days 30, 40 and 60 of incubation (An30, An40 and An60). The samples were subsequently analysed with FISH and PCR-DGGE to identify the anammox bacteria. Moreover, cloning analysis was carried out for sample An60 to obtain detailed information regarding the 16S rRNA sequence of the anammox bacteria.

Sample preparation and application of FISH

The sludge samples were fixed in 4% paraformaldehyde for 24 h at 4°C and then stored in a mixture of PBS and ethanol, with a final ethanol concentration of 50% at −20°C. Fixed granular sludge samples taken at day 90 of reactor operation were embedded in the cryosectioning medium compound (Okabe et al. 1999) and were cut into 20-μm-thick sections using a Microtome Cryostat Cryo-Star HM 560MV (Microm International GmbH, Walldorf, Germany) at −20°C. Granule sections were immobilized on microscopic slides and then dehydrated in 50, 80 and 100% ethanol. Crushed granular sludge samples R20, R40 and R80 were homogenized for 10 min with a Teflon piston (1500 rev min−1) and were immobilized on microscopic slides by air-drying and dehydrating for 3 min in an ethanol series of 50, 80 and 100%.

Hybridization was performed at 46°C for 90 min directly on the prepared slides using 1 μl 16S rRNA-targeted oligonucleotide probe and 8 μl hybridization buffer as described previously (Amann et al. 1995). The oligonucleotide probes 5′ labelled with Cy3 purchased from Eurofins MWG Operon (Ebersberg, Germany) and formamide concentrations used in the hybridization buffer are described in Table 2. Washing steps were performed at 48°C for 15 min in a specific washing buffer (Amann et al. 1995). Afterwards, the samples were stained with 4′-6-diamidino-2-phenylindole (DAPI) (5 μg ml−1) for 15 min in the dark at 4°C. Hybridized sample were analysed using an acquisition system LSM 510 META laser scanning microscope (Zeiss, LSM510 META) controlled by confocal software v3.2 (Zeiss, Göttingen, Germany). Twenty different microscopic fields were randomly selected for the homogenized granules, and positive FISH signals were quantified using class indexes according to the study by Müller et al. (2007). After the cryosection of the granule, five sections were analysed and mean values were calculated from these five sections. Class indexes from 0 (no positive FISH signal) to 5 (extensive growth representing more than 40% of the microbial populations) were used, with class index steps differing by a factor of 10.

Table 2. 16S rRNA-targeted oligonucleotide probes used for fluorescent in situ hybridization in this study
Probe nameLabelling dyeFormamide (%)Target site (probe sequence 5′–3′, 16S rRNA)Target organismReferences
AMX820Cy540AAA ACC CCT CTA CTT AGT GCC C

Ca. Brocadia anammoxidans Ca. Kuenenia stuttgartiensis

Schmid et al. (2000)
AMX368Cy515CCT TTC GGG CAT TGC GAAAll anammox bacteriaSchmidt et al. (2003)
PLA46Cy530GAC TTG CAT GCC TAA TCC Planctomycetes Neef et al. (1998)
BFU613Cy530GGA TGC CGT TCT TCC GTT AAG CGGCa. Brocadia fulgidaKartal et al. (2008)
KST157Cy525GTT CCG ATT GCT CGA AAC Ca. Kuenenia stuttgartiensis Schmid et al. (2000)
APR820Cy540AAA CCC CTC TAC CGA GTG CCC Ca. Anammoxoglobus propionicus Kartal et al. (2008)
BS820Cy540TAA TTC CCT CTA CTT AGT GCC C Ca. Scalindua wagneri Kuypers et al., (2003)
SCABR1114Cy520CCC GCT GGT AAC TAA AAA CAA G Ca. Scalindu brodae Schmid et al. (2003)
NSO190Cy555CGA TCC CCT GCT TTT CTC CAOB within the Betaproteobacteria Mobarry et al. (1996)
NIT3Cy540CCT GTG CTC CAT GCT CCG Nitrobacter Wagner et al. (1996)
NTSPA662Cy535GGA ATT CCG CGC TCC TCT Nitrospira Mobarry et al. (1996);Daims et al. (1999)

DNA extraction, PCR amplification and DGGE

Genomic DNA was extracted from sludge samples of the deammonification reactor (R20, R40 and R80) and the anaerobic cultivation (An20, An40 and An60) using Fast DNA SPIN extraction kit (MP Biomedicals, Illkirch Cedex, France). A combination of forward primer specific for Planctomycetes PLA40F: 5′-GGA TTA GGC ATG CAA GTC-3′ and universal bacterial reverse primer 518R: 5′-ATT ACC GCG GCT GCT G-3′ were used to amplify the partial 16S rRNA gene of Planctomycetes (Müller et al. (2007)). A 40-nucleotide GC-clamp was added to 518R at 5′ end to improve the detection of sequence variation by DGGE (Muyzer et al. 1993). PCRs were performed with HotStart Taq Master Mix (Qiagen GmbH, Hilden, Germany). The PCR protocol consisted of a 15 min activation step at 95°C, followed by 30 cycles of 1 min at 95°C, 1 min at 56°C (annealing temperature), 1 min at 72°C and a final extension for 7 min at 72°C. PCR products were then examined by DGGE electrophoresis and loaded onto a 40–70% denaturing gel with a D-CODE System Universal Mutation (Bio-Rad Laboratories, Hercules, CA, USA) according to the manufacturer's instruction. DGGE ran for 999 min at 100 V and at a constant temperature of 60°C in 1×TAE buffer. Subsequently, the gel was stained with SYBR Gold in 1×TAE buffer for 40 min, and digital images of the gel were obtained using the Gel Doc 2000 System (Bio-Rad Laboratories). Distinct DGGE bands were excised and placed in 1·5-ml tubes, incubated with TE buffer at 4°C for 12 h. One microlitre of this solution was used as a template in a second PCR with the primer set PLA40F and 518R, using the cycling PCR program described previously. The obtained PCR products were purified by QIAquick® PCR purification kit and sent to Eurofin MWG Operon (Ebersberg, Germany) for sequencing.

DNA extraction, PCR and cloning

Genomic DNA extraction of the An60 sludge sample was performed using the Fast DNA SPIN extraction kit. PCR was firstly carried out by a combination of forward primer PLA40F specific for Planctomycetes and universal bacterial reverse primer 1492R: 5′-ACG GCT ACC TTG TTA CGA CTT-3′ to amplify the 16S rRNA genes of the Planctomycetes (Lane 1991). PCR amplifications were performed as previously described. The only difference was the annealing temperature in this PCR cycling program, which was set at 53°C. The purified product was then ligated into the PCR®4-TOPO® plasmid vector and transformed into competent Escherichia coli TOP 10 cells using TOPO-TA cloning kit according to the manufacturer's instructions (Invitrogen, Darmstadt, Germany). After cultivation at 37°C for 1 h, the colonies were randomly selected and examined for the presence of the correct insert size by PCR with vector primers M13F: 5′-GTA AAA CGA CGG CCA G-3′ and M13R: 5′-CAG GAA ACA GCT ATG AC-3′. The PCR cycling program was 95°C for 15 min, 35 cycles of 45 s at 95°C, 45 s at 53°C for 45 s and 90 s at 72°C, and a final extension at 72°C for 10 min. The clone library was screened by restriction fragment length polymorphism (RFLP) using enzyme HaeIII (Fermentas, St Leon.Rot, Germany) according to the manufacturer's protocol. Clones were grouped according to their restriction patterns defining different operational taxonomic units (OTUs). For each OTU, one clone was randomly selected, PCR was performed using the primer pair 40F specific for Planctomycetes and universal bacterial primer 1429R, and PCR productions of the clone inserts were purified by QIAquick® PCR purification kit and sent to Eurofin MWG Operon using the following sequencing primers: 40F, 342F: 5′-TAC GGG AGG CAG CAG-3′ (Lane 1991), 907F: 5′-CCG TAC ATT CCT TTR AGT TT-3′ (Lane 1991) and 517R: 5′-ATT ACC GCG GCT GCT GG-3′ (Riemann et al. 2000).

Phylogenetic sequence analysis

The partial sequences obtained from the clone inserts were merged to get a nearly full-length 16S rRNA sequence. The partial 16S rRNA sequence originated from the DGGE bands and nearly full-length 16S rRNA sequences of clone inserts were analysed using NCBI databases by applying the Blast (http://www.ncbi.nlm.nih.gov/BLAST) to find the nearest published relatives and aligned by Clustal X1.8 program against the related species. A phylogenetic tree was constructed by neighbour-joining (NJ) in the mega 3.1 package containing the sequences of this study and the related published sequences (Saitou and Nei 1987). 16S rRNA full-length sequences of clone inserts were deposited in the GenBank database with the following accession numbers: JF286586, JF286587 and JF286588.

Results

The nitrogen and COD concentrations during one representative SBR cycle

The reject-water having an math formula-N concentration of 1 082 ± 100 mg N l−1 and C/N of 0·7 ± 0·1 was used as the influent for a single-stage deammonification reactor. Ten days after start-up, stable operation was reached. The nitrogen removal efficiency was 90·5 ± 3·5%, and COD removal efficiency was 78 ± 3% during the reactor operation. Full granulation was achieved within 3 months where the average granule diameter was 0·6–0·8 mm.

Figure 1 presents the nitrogen and COD concentrations in one SBR cycle. Additionally, Table 3 shows the nitrogen and COD development according to the cycle steps. The last two cycle steps without filling led to significantly lower nitrogen and COD concentrations in the final effluent. From the data presented, it can be concluded that most of the nitrogen removal was observed in the aerobic phase. The ratio of △math formula-N/△COD in this period was 7, with a small amount of math formula-N production.

Table 3. Fate of N and COD in the anoxic and aerobic phases during one sequencing batch reactor cycle
N course
(1) math formula-N in reactor after influent phase: 102·3 ± 6 mg l−1
(2) math formula-N in reactor after anoxic phase: 102·6 ± 10 mg l−1
(3) math formula-N in reactor after aerobic phase: 75·6 ± 9 mg l−1
(4) math formula-N in effluent: 47·9 mg l−1
(5) math formula-N in reactor after influent phase: 1·2 ± 0·6 mg l−1
(6) math formula-N in reactor after anoxic phase: 0·95 ± 0·4 mg l−1
(7) math formula -N in reactor after aerobic phase: 1·25 ± 0·7 mg l−1
(8) math formula-N in effluent: 0·2 mg l−1
(9) math formula-N in reactor after influent phase: 48·9 ± 3 mg l−1
(10) math formula-N in reactor after anoxic phase: 43·4 ± 5 mg l−1
(11) math formula-N in reactor after aerobic phase: 50·5 ± 5 mg l−1
(12) math formula-N in effluent: 51·1 mg l−1
COD course
(13) COD in reactor after influent phase: 193·0 ± 9 mg l−1
(14) COD in reactor after anoxic phase: 175·5 ± 14 mg l−1
(15) COD in reactor after aerobic phase: 173·2 ± 5 mg l−1
(16) COD in effluent: 153 mg l−1
△COD/△N, g COD/g N
(17) Overall: 0·79
(18) In anoxic phase: 3·21
(19) In aerobic phase: 0·12
Figure 1.

The N and COD concentrations during one sequencing batch reactor (SBR) cycle. (image_n/jam12042-gra-0001.png) math formula-N concentration; (image_n/jam12042-gra-0002.png) math formula-N concentration; (image_n/jam12042-gra-0003.png) math formula-N concentration and (image_n/jam12042-gra-0004.png) COD concentration.

A slight COD removal was also observed during the aerobic phase, whereas an average concentration of 177·5 ± 14 mg l−1 was still present. However, contrary to the nitrogen removal, the main COD removal took place in the anoxic phase, with a △COD/△math formula-N of 3. A stoichiometric ratio of 4·2 g COD (g N)−1 is postulated for denitrification (Carrera et al. 2004), and consequently, there may have been too little organic carbon in the system to achieve complete math formula-N reduction under anoxic conditions. Other potential electron donors such as inactive biomass may have been available.

The deammonification reactor was characterized by an math formula-N elimination rate of 399 ± 12 mg N (l day)−1, and 78% of math formula-N was removed in the aerobic phase. The COD removal rate was 227·5 ± 12·5 mg COD (l day)−1, and 88% was consumed in the anoxic phase.

Characterization of the microbiological population involved in deammonification: FISH analysis of the sludge originated from SBR and anaerobic cultivation

Application of probes PLA46 and NSO190 indicated the co-existence of Planctomycetes and AOB in granular sludge from the deammonification reactor. The high class indexes of 4·5 for AOB and 4 for Planctomycetes remained stable during the whole operation. This showed that each group represented about 20–40% of the sludge population in the SBR. The detection of Nitrospira and Nitrobacter using the specific probes NTSPA662 and NIT3, respectively, failed, indicating that the nitrite oxidizers (NOB) were not able to survive in this deammonification reactor.

Although strong cauliflower-aggregate signals characteristic of anammox bacteria (Jetten et al. 2003) were detected, the hybridization signals with the universal probe AMX368 for anammox bacteria were very weak. Figure 2 shows the quantitative FISH analysis of sludge from the SBR and anaerobic cultivation, where specific probes were used for the Planctomycetes phylum, AMX368 for anammox bacteria, as well as for different anammox bacteria species. The Planctomycetes phylum with a class index of 4 represents one dominant group within the deammonification granular sludge. However, the anammox group-specific probe AMX368 detected only very few bacterial cells, resulting in weak fluorescent signals. Moreover, it failed to detect the dominant Planctomycetes species Ca. Brocadia fulgida that only led to positive FISH signals with probe BFU613 (class index 3·5). Additionally, Ca. Brocadia anammoxidans (probe AMX820) was present in very low amounts (class index 0·8), and other anammox bacteria species such as Ca. Anammoxoglobus propionicus, Ca. Kuenenia stuttgartiensi, Ca. Scalindua wagneri and Ca. Scalindua brodae were not found in either the SBR or the anaerobic cultivation.

Figure 2.

Class indexes of the specific probe/DAPI according to the fluorescent in situ hybridization (FISH) analysis with probes PLA46 for Planctomycetes, BFU613 for Ca. Brocadia fulgida, APR820 for Ca. Anammoxoglobus propionicus, KST157 for Ca. Kuenenia stuttgartiensi, BS820 for Ca. Scalindua wagneri and SCABR1114 for Ca. Scalindua brodae (R20, R40 and R80 represent the samples on days 20, 40 and 80 of the deammonification reactor operation, respectively. An30, An40 and An60 represent the samples on days 30, 40 and 60 of the anaerobic cultivation, respectively). (image_n/jam12042-gra-0005.png) Sample R20; (image_n/jam12042-gra-0006.png) sample R40; (image_n/jam12042-gra-0007.png) sample R80; (image_n/jam12042-gra-0008.png) sample An30; (image_n/jam12042-gra-0009.png) sample An40 and (image_n/jam12042-gra-0010.png) sample An60.

During SBR operation, the anammox population mostly represented by the Ca. Brocadia fulgida, remained stable and an increasing number of these organisms was observed during anaerobic cultivation. This demonstrates that anammox bacteria were enriched under anaerobic conditions.

In situ mapping of deammonification granules using FISH analysis

In situ distribution of deammonification populations was carried out on 20-μm granule cryosections using specific FISH probes for AOB within the Betaproteobacteria (NSO 190) and for the Planctomycetes phylum (PLA 46) representing mainly anammox bacteria as shown before (see section “Characterization of the microbiological population involved in deammonification: FISH analysis of the sludge originated from SBR and anaerobic cultivation”). Figure 3 shows the localization of AOB and anammox bacteria within the granules. It can be seen that NSO190-positive AOB were mainly detected in the outer region of the granules, close to the surface (Fig. 3a,b), and PLA46-positive anammox bacteria were distributed in the inner region of the granules (Fig. 3c,d). Cells stained with DAPI colonized the outer edges of the granules. However, after applying the probes NSO190 and PLA46, there were no positive FISH signals in this region. These populations only stained with DAPI were probably comprised of heterotrophic bacteria that may have been responsible for the COD removal as described previously (see section ‘The nitrogen and COD concentrations during one representative SBR cycle’).

Figure 3.

Fluorescent in situ hybridization (FISH)-CSLM micrographs of one granular cryosection hybridized with NSO190 probe (magenta in colour, light grey in white and black) specific for AOB and counterstained with DAPI (blue in colour, dark grey in white and black, a and b) and hybridized with PLA46 probe specific for the Planctomycetes phylum (magenta in colour, light grey in white and black) and DAPI (blue in colour, dark grey in white and black, c and d).

The quantitative data using class indexes (Müller et al. 2007) of AOB and anammox bacteria distribution within the granules are shown in Fig. 4. The highest numbers of NSO190-positive AOB (class index 3·75) were concentrated at a distance of 50 μm (left area) and 150 μm (right area) from the granule surface. At deeper regions within the granule, a decrease in their concentration was observed. In the presence of high numbers of AOB, only some anammox bacteria (class index 0·5–1) were found. These anammox bacteria were more concentrated in deeper regions of the granule, at a distance of 100–200 μm from the surface (class index 3·0). In this inner region, oxygen concentrations were probably very low because of diffusion limitation and consumption by aerobic bacteria. The centre of the granule is characterized by a DAPI-negative hole with a diameter of 100 μm, demonstrating no microbial populations were present in the middle of the granule. The granule centre may possibly have been comprised of extracellular polymeric substances (EPS). A similar observation was made by McSwain et al. (2005) who showed that a protein core improved granular formation and stability.

Figure 4.

In situ mapping of the distribution of NSO190-positive AOB within Betaproteobacteria and PLA46-positive Planctomycetes in deammonification granule. (image_n/jam12042-gra-0011.png) NSO190-positive/DAPI and (image_n/jam12042-gra-0012.png) PLA46-positive/DAPI.

PCR-DGGE analysis for Planctomycetes community

PCR-DGGE was performed as an additional method to characterize in detail the Planctomycetes community diversity in both the deammmonification SBR and the anaerobic cultivation. The DGGE profiles of the SBR sludge samples (R20, R40 and R80) were characterized by 12 distinct bands (Fig. 5, bands a–l). No significant changes of the DGGE fingerprints were observed at day 20 (R20), day 40 (R40) and day 80 (R80) of reactor operation. However, not all of the detected Planctomycetes species in the SBR survived during the 2-month anaerobic cultivation. This resulted in a continuous decrease in DGGE bands at day 30 (An30), day 40 (An40) and day 60 (An60) of anaerobic cultivation. The DGGE profile of sample An60 was characterized by only three stained bands (Fig. 5, bands a, b and h). The signal intensity of both DGGE bands a and b increased under long-term anaerobic cultivation, and the highest staining signal was observed for band a (see Fig. 5a,b).

Figure 5.

Denaturing gradient gel electrophoresis (DGGE) analysis of PCR amplified 16S rRNA gene fragments using 40F/518R GC primer. R20, R40 and R80 represent the samples of days 20, 40 and 80, respectively, of the deammonification reactor operation. An30, An40 and An60 represent the samples of days 30, 40 and 60 of the anaerobic cultivation, respectively. The letters a–l represent the bands excised for sequencing and phylogenetic analysis; (a) and (b) indicate the signal intensity of DGGE bands for samples R20 and An60, respectively.

Twelve distinct DGGE bands of the R20 sample (a–l) were excised and sequenced. The partial 16S rRNA sequences of these excised DGGE bands (480 bp) and their close relatives were analysed by Blast using the NCBI database and were included in a phylogenetic tree (Fig. 6). All sequences were affiliated within the Planctomycetes phylum, and only two sequences were related to known anammox bacteria (sequences of DGGE bands a and b). The other sequences either branched with uncultured Planctomycetes (DGGE bands c, d, e, f, g, h and j) or formed individual clades (DGGE bands i, k and l). The DGGE bands c and d with a sequence similarity of 100% and DGGE bands g and h with a sequence similarity of 99% were alike and formed an individual cluster in the phylogenetic tree (Fig. 6). The sequences of the DGGE bands a and b found in the SBR and enriched in the anaerobic cultivation were affiliated with the genus Ca. Brocadia. They showed a 99% sequence similarity to Ca. Brocadia fulgida (accession no. EU478693) and a 99% sequence similarity to Ca. Brocadia sp. (accession no. AM285341).

Figure 6.

Neighbour-joining tree of partial 16S rRNA sequences (489 bp) obtained by denaturing gradient gel electrophoresis (samples R20), nearly full-length 16S rRNA sequences (1452 bp) obtained by cloning (sample An60) and published Planctomycetes sequences registered in the NCBI database. Scale bar represents 0·05 estimated sequence divergences. Bootstrap values (%) from 500 bootstraps are indicated.

Cloning analysis to identify the functional Planctomycetes using anaerobic culture

Both FISH and PCR-DGGE analyses identified Ca. Brocadia fulgida as the dominant anammox bacteria. Cloning analysis providing nearly full-length rRNA sequences was carried out to prove whether Ca. Brocadia spp. actually represented the only anammox bacteria species within the Planctomycetes phylum.

A clone library was constructed from sludge sample An60 using the primer set PLA40F and 1492R. Forty clones containing the correct insert size were examined by RFLP and classified into three OTUs. The dominant OTU was composed of 32 clones (80% of the clone library) and the sequence of one representative clone named uncultured Planctomycete clone ST_2 (accession no. JF286587) was closely related to Ca. Brocadia fulgida (accession no. EU478693, 99% similarity). The sequence of the representative clone named uncultured Planctomycete clone ST_3 (accession no. JF286588) of the second OTU (four clones, 10% of the clone library) matched the sequence from Ca. Brocadia sp. (accession no. AM285341) with a similarity of 99%. The representative clone named uncultured Planctomycete clone ST_1 (accession no. JF286586) in the third OTU (four clones, 10% of the clone library) showed a 99% similarity to an uncultured Planctomycetes clone (accession no. GQ356109) and 96% similarity to Ca. Brocadia sp. (accession no. AM285341). The phylogenetic relationship of the clone sequences from this study and their closest relatives are shown in the phylogenetic tree together with the PCR-DGGE sequences (Fig. 6).

Phylogenetic analysis of the clone sequences showed a low microbial diversity, indicating that only a few Planctomycetes species could survive the anaerobic conditions. The dominant anammox bacteria as verified by FISH, PCR-DGGE and cloning investigations were identified as Ca. Brocadia fulgida.

Discussion

In this study, the reject-water from sludge dewatering at WWTP Garching, Germany, was successfully treated in a single-stage deammonification SBR. A 90·5 ± 3·5% of N and 78 ± 3% of COD were removed from the influent characterized by math formula-N of 1 082 ± 100 mg N l−1 and C/N of 0·7 ± 0·1. The successful granulation of the deammonification sludge was achieved within 3 months after inoculation. Sludge granulation usually occurs under aerobic conditions (Li et al. 2011). However, Muda et al. (2010) reported that granulation occurred in a reactor operated under intermittent anaerobic and aerobic conditions. As previously reported, a Ca2+ concentration of 150–300 mg l−1 can enhance the granulation process and reduce the granulation time by binding effects (Jiang et al. 2003; de Graaff et al. 2011). In this study, a high concentration of Ca2+ (100–150 mg l−1) in the influent might has supported granulation of the deammonification sludge.

The deammonification process, including aerobic ammonium oxidization and anammox reaction, took place during the aerobic phase as proven by the N and COD removal performance during the SBR cycle. This observation appears to contradict the widely accepted concept, in which aerobic ammonium oxidization occurs in the aerobic phase and the anammox reaction takes place in the anaerobic phase. An oxygen gradient within a granule could possibly lead to the occurrence of ammonium oxidation at the outer surface, where oxygen concentrations are higher. In the inner region of a granule, where oxygen concentrations are lower because of diffusion limitation, it is plausible that the anammox reaction could occur, even during the aerobic phase of an SBR cycle. This was confirmed by the distribution of the relevant bacteria groups within the deammonification granules studied in this research. Autotrophic Ca. Brocadia fulgida species within the Planctomycetes, characterized by a low assimilation rate, were found at distances far from the surface of the granules Autotrophic AOB within Betaproteobacteria, showing higher growth rates were located near to the granule's surface. This spatial arrangement of different kinds of bacteria according to growth rate and substrate availability might stabilize the granule.

COD removal in the system was performed by heterotrophic organisms, that is denitrifiers requiring anoxic conditions. Low COD removal rates during the aerobic phase indicate that there was no denitrifying activity. This might have been caused (i) by a low penetration of organic compounds into the inner anoxic zone of the granules or (ii) because there were no denitrifying bacteria in this anoxic region of the granule. During the anoxic phase, the denitrifying activity was apparent from both COD removal and math formula-N utilization. The △COD/△math formula-N throughout this phase was 3. This points to a large flexibility of the deammonification granules in carrying out various reactions in a single unit. Moreover, it underlines their capability to maintain high substrate removal rates.

In this study, FISH analysis of the deammonification granules showed the co-existence of both AOB, concentrated close to the granule's surface, and Planctomycetes dominated by the anammox bacteria Ca. Brocadia fulgida, growing in the inner region. Similar results were documented by Vlaeminck et al. (2010), who investigated granules from OLAND reactors. The architecture of these granules was characterized by AOB at the rim and internal layers embedded with anoxic anammox bacteria. Additionally, irregularly shaped voids or channels were contained in the inner anoxic zone where no cells were detected. In the present study, a cell-free hole was identified in the middle of the granules. Internal, bacteria-free gaps might consist of an aggregation of EPS. Such findings were seen by McSwain et al. (2005), who showed that a protein core improved granular formation and stability.

Further characterization of the relevant anammox bacteria species within the Planctomycetes group was carried out using the deammonification granular sludge and enriched anaerobic cultures by a combination of different molecular biological techniques. The PCR-DGGE analysis revealed a higher diversity of Planctomycetes in the deammonification reactor compared with anaerobic cultivation supplied with specific nutrients aimed at boosting the growth of anammox bacteria. Phylogenetic analysis of the DGGE bands present in the deammonification sludge showed that some Planctomycetes species were closely related with uncultured Planctomycete clones previously detected in soils, a desert ecosystem and a wastewater treatment reactor (Jangid et al. 2008; Kwon et al. 2010). They, however, did not show any relationship with the published anammox bacteria. The Planctomycetes phylum is a large phylogenetic group found in different ecosystems, where some members might be heterotrophs who prefer aerobic conditions and lack the skills to perform anammox reaction (Neef et al. 1998; Miskin et al. 1999). The anammox bacteria Ca. Brocadia fulgida, identified as the dominant Planctomycetes species in both the deammonification reactor and enriched anaerobic cultivation, might be responsible for the anammox reaction in this study. This was verified by sequence analysis of both PCR-DGGE bands, clones and quantitative FISH analysis using the specific probe BFU613. However, hybridization with the universal probe AMX368, specific for all published anammox bacteria including Ca. Brocadia fulgida (Schmidt et al. 2003), failed. Blast analysis showed that the AMX368 probe sequence has no mismatch with the target region of Ca. Brocadia fulgida sequence, and consequently, the probe AMX368 might hybridize to this organism. The reason for the failure of FISH using AMX368 might be the inaccessibility of the site targeted by the probe because of a higher order structure of the ribosome (Fuchs et al. 2001; Behrens et al. 2003). This observation indicates that the choice of the right oligonucleotide probe has to be done very carefully and a negative FISH analysis has to be verified with other molecular biological techniques.

Anammox versatility has been recently reported, identifying Ca. Brocadia and Ca. Kuenenia dominating the WWT systems and Ca. Scalindua as the predominant anammox bacteria in anoxic marine habitats. Ca. Brocadia fulgida has been described to be capable of co-oxidizing organic compounds, for example, acetate and ammonium more efficiently than other anammox bacteria (Kartal et al. 2008). The different physiological characteristics could be a key for determining the predominant anammox bacteria in both natural environments and engineered systems (Terada et al. 2011). As observed previously, the presence of organic carbon could encourage Ca. Brocadia fulgida to compete with other Planctomycetes (Winker et al. 2012;). This finding could support the present results regarding the domination of Ca. Brocadia fulgida in comparison with other anammox bacteria when treating high-loaded reject-water with a C/N ratio around 0·7.

Acknowledgements

This project was supported by Alexander von Humboldt Stiftung/Foundation (09.2009–08.2011).

Ancillary