Introduction

Removal of nitrogen by biological means is a widely adopted method for treatment of high-nitrogen-containing wastewaters due to its effectiveness and cost benefits. This process is normally carried out by different groups of bacteria by aerobic removal of ammonium (nitrification by autotrophic bacteria) and anaerobic conversion of the nitrate and nitrite to N2 gas (denitrification by heterotrophic bacteria) under different aeration rates (Knowles 1982; Metcalf and Eddy 1991; Sharma and Ahlert 1977).

Various processes have been developed based on biological nitrification and denitrification which include short-cut nitrification and denitrification, anaerobic ammonium oxidation (Annamox), completely autotrophic nitrogen removal over nitrite (CANON) process, and oxygen-limited autotrophic nitrification–denitrification process which find widespread applications throughout the world (Peng and Zhu 2006). However, nitrification and denitrification via common intermediate nitrite (partial nitrification) has been found to be technically most feasible and economical method which has been employed successfully on a full scale in single reactor system for high ammonia removal over nitrite process (Mulder et al. 2001).

The latest advancement in biological nitrogen removal has been the development of a highly energy-efficient combination of process of partial nitrification with Annamox in a single reactor (CANON process; Third et al. 2005). This combined process consisted of addition of nitrifying culture to enriched Annamox culture in a reactor to initiate the CANON process in continuous mode which on stabilization was able to remove nitrogen for extended period at rate of 0.08 kg N m−3 day−1.

Concurrent nitrification and denitrification (called simultaneous nitrification denitrification—SND) by aerobic conversion of ammonium \( {\left( {{\text{NH}}^{ + }_{4} } \right)} \)to N2 gas has been known to occur naturally in biomass films or activated sludge flocs by virtue of the difference in dissolved oxygen (DO) levels at different points (von Munch et al. 1996). Certain groups of heterotrophic bacteria such as Alcaligenes faecalis, Thiosphaera pantotropha, Comamonas, etc. have also been known to possess activity of SND (Ferguson 1994; Joo et al. 2005; Otte et al. 1996; Patureau et al. 1997; Robertson and Kuenen 1988; van Niel et al. 1992). A novel bacterium belonging to family Comamonadaceae capable of performing SND was isolated from activated sludge and was named as Diaphorobacter nitroreducens (Khan and Hiraishi 2002).

Joo et al. (2005) demonstrated the effect of C/N ratios in removal of ammonia by an A. faecalis strain in presence of acetate and found C/N ratio of 10 to be optimum for ammonia removal, but at lower C/N ratio of 5, 40% of ammonium remained unconsumed. From the nitrogen balance, the authors found that, at C/N ratio of 10, up to half of the ammonium nitrogen was converted to intracellular nitrogen. Similar experiments on effect of C/N ratios on removal of ammonium in a sequencing batch reactor inoculated with sludge by Chiu et al. (2007) revealed that ammonium removal was better at high C/N ratio of 19.7 and reduction in C/N ratio to 6.3 resulted in decrease in ammonium removal efficiency.

In the present study, we attempted to isolate bacteria from activated biomass treating chemical and dye industries wastewater and capable of carrying our SND. A total of eight strains were characterized which showed potential for SND under aerobic conditions and belonged to the β-proteobacteria of Comamonadaceae family.

Materials and methods

Isolation and identification

The eight bacterial cultures were isolated from activated biomass from a common effluent treatment plant (CETP) treating wastes from chemical and dye industries and maintained in 30% glycerol stocks. The genomic DNA of the isolates was extracted according to Sambrook et al. (1989), and molecular taxonomic position of the isolates was determined by partial 5’ 16S rDNA sequence comparison. The 16S rDNA was amplified, and the sequences were aligned on the CLUSTAL_X program (Thompson et al. 1997), and a phylogenetic tree was constructed by the neighbor-joining method (Saitou and Nei 1987) and further evaluated by bootstrap sampling (Felsenstein 1985). The scale bar represents 0.1 fixed mutations per site, and bootstrap values were derived from 1,000 analyses. The phylogenetic tree was displayed using the TREEVIEW program (Page 1996), and the 16S rDNA sequences were deposited into the NCBI GenBank.

The genotypic variance in between the different isolates was determined by random amplification of polymorphic DNA (RAPD) analysis using standard primer (Amersham) with the following sequence 5′ d[AAGAGCCCGT] 3′. The RAPD reaction mixture consisted of the following components in a total volume of 50 μl: autoclaved distilled water—31.0 μl; 10X polymerase chain reaction (PCR) Buffer (with 15 mM of MgCl2)—5.0 μl; dXTPs (200 μmol)—2.5 μl; primer (25 pmol)—5.0 μl; Taq polymerase (3 U μl−1)—1.5 μl; template DNA—5.0 μl; mineral oil—two drops. RAPD PCR program was carried out in Thermocycler (Perkin Elmer) for 40 cycles with each cycle consisting of three major steps, denaturation at 94°C for 1 min, annealing at 55°C for 1 min, extension at 72°C for 1 min. The different band patterns were observed by agarose gel electrophoresis (2% gel).

Growth medium and culture conditions

A complex mineral medium as described by Blaszczyk (1993) was used for cultivation and studies on simultaneous nitrification and denitrification by the isolates. The composition of the medium (C/N ratio = 8) was as follows (g l−1): K2HPO4—7.0; KH2PO4—3.0; sodium citrate·2H2O—0.5; MgSO4·7H2O—0.1; (NH4)2SO4—1.0 (7.5 mM \({\text{NH}}_4^ + \)–N); FeSO4·7H2O—0.05; KNO3—1.8 (20.9 mM \({\text{NO}}_{\text{3}}^{\text{ - }} \)–N); CH3COONa—10.0. Initial pH was set at 7.0 ± 0.2, and 250-ml capacity conical flasks containing 100-ml medium were used in the experiments, which were incubated at 30°C in an orbital shaker at 150 rpm. The experimental conditions remained the same in all the subsequent experiments unless specifically stated.

Studies on growth and SND by the isolates

One hundred microliters of the different Diaphorobacter strains viz HPC 805, HPC 815, HPC 820, HPC 821, HPC 834, HPC 847, HPC 848, and HPC 856 was inoculated from glycerol freezer stocks in 5-ml Luria–Bertani broth tubes. After 24 h of incubation, the cultures were centrifuged at 7,000 rpm for 10 min, washed twice with sterile distilled water, and resuspended in minimum amount of distilled water. The optical density was measured at 600 nm against distilled water blank and 1-O.D. cells were inoculated into experimental flasks. Samples were withdrawn at 24-h intervals for estimating the changes in pH, biomass, chemical oxygen demand (COD), nitrate, nitrite, and ammonium concentration. All physico-chemical analyses were performed as per standard methods (APHA, AWWA, WPCF 2005). Specific growth rates (μ) were calculated by the following formula:

$$\mu {\text{ = }}\frac{{{{d}}\left[ {\text{X}} \right]}}{{\left[ X \right] \times d\left[ T \right]}}$$

where [X] is the actual cell density expressed as O.D. at time [T], [dX] is change in cell density over a fixed interval of time d[T].

Enzyme assay

Cell-free extracts were prepared by subjecting pre-grown cultures to lysis by ultrasonication in presence of 20-mM phosphate buffer (disodium hydrogen phosphate + potassium dihydrogen phosphate, pH 7.0). The ultrasonicated samples were placed in a sonication bath for 30 min to detach the membrane-bound enzymes which were separated from the cell debris by centrifugation. Formation of nitrite from nitrate in presence of centrifuged cell-free extract in the reaction mixture was taken as a measure of nitrate reductase activity (Sanderson and Cocking 1964). Similarly, the disappearance of nitrite from reaction mixture in presence of cell-free extract was considered as a qualitative measure for nitrite reductase activity (Wray and Filner 1970). Protein concentration in the cell-free extract was determined by Folin-phenol reagent (Lowry et al. 1951) to determine the specific activity of the enzymes. One unit of enzyme activity was defined as the amount of enzyme which catalyzed the transformation of 1 μmol of substrate per minute. The amount of enzyme activity per unit of protein was defined as the specific activity of a given preparation of material.

Results

The eight cultures were isolated from activated biomass of a CETP treating dye and chemical industry wastewater and tentatively numbered as HPC 805, HPC 815, HPC 820, HPC 821, HPC 834, HPC 847, HPC 848, and HPC 856. The genotypic characterization of the eight isolates was carried out by 16S-rDNA-sequence-based phylogenetic analysis. Partial sequencing of the 16S rDNA of the isolates (674–675 bases) was carried out and deposited in the NCBI GenBank with the accession numbers AY996989, AY996993, AY996996, AY996997, DQ131848, AY997013, DQ131850, and DQ131852, respectively (Kapley et al. 2007). Homology search revealed that all the isolated cultures shared sequence similarity of 99–100% to other Diaphorobacter strains such as ZY 2006b, F2, NA5, PCA039, D. nitroreducens KSP4, KSP3, and Acidovorax sp. 3DHB1. In addition, the isolates showed 98–99% sequence homology to D. nitroreducens NA10B (type strain JCM 11421) and Acidovorax sp. LW1 (type strain DSM 13225). With respect to other members of the family Comamonadaceae, the isolates showed sequence similarities of 95–98% to A. aerodenitrificans, A. avinae C1, Acidovorax sp. 1916, Alicycliphilus sp. R24604, R24611, A. denitrificans K601, Alicycliphilus sp. A8, Aquaspirillum sp. EMB 325, and Simplispira sp. R28033.

Figure 1 shows a neighbor-joining phylogenetic tree as prepared from the partial 16S rDNA sequences. Because complete sequences of closely related strains were available in the NCBI GenBank, while partial sequences of our isolates were available, neighbor-joining tree was prepared by taking partial lengths (800 bases) of the related sequences from NCBI GenBank from 5′ end. Partial sequences of type strains from the same group viz β-proteobacteria and other groups (one each) were included for demonstration of genetic relatedness. All our isolates were clustered together with HPC 820, HPC 821, HPC 856, HPC 815, and HPC 848 forming a tight cluster while HPC 805, HPC 847, and HPC 834 showing a slight divergence from this cluster. The eight strains also formed a close cluster with other Diaphorobacter strains and Acidovorax sp. LW1, in addition to strains belonging to Alicycliphilus, Aquaspirillum, and Simplispira. The tree showed a clear evolutionary divergence of this group of β-proteobacteria from type strains belonging to other groups such as Acinetobacter, Citrobacter, Rhizobium, Campylobacter, Geobacter, Bacillus, and Methanobacterium. Genetic variance between the eight different strains was confirmed by RAPD, which showed different band patterns in the different strains (Fig. 2).

Fig. 1
figure 1

Neighbor-joining matrix tree based on partial 16S rDNA sequences showing phylogenetic relationships between our laboratory isolates, other members of Comamonadaceae, and type members from other groups. Names of the different cultures along with the accession numbers for the 16S rDNA sequences are given in the parentheses. Bootstrap values are mentioned at branching points

Full size image
Fig. 2
figure 2

RAPD profile of different Diaphorobacter strains (lane 1 HPC 805, lane 2 HPC 815, lane 3 HPC 820, lane 4 HPC 821, lane 5 HPC 834, lane 6 HPC 847, lane 7 HPC 848, lane 8 HPC 856, lane 9 1Kb ladder)

Full size image

The ability of the different Diaphorobacter strains to grow under conditions of SND was studied, and the growth profiles were evaluated in a time series analysis and expressed graphically in the form of specific growth rate (μ). It is seen from Fig. 3, that the growth rates were higher for isolates HPC 821 (1.016 O.D. ml−1 h−1) and HPC 815 (0.762 O.D. ml−1 h−1) with a decrease observed in the order HPC 820 > HPC 847 > HPC 848 > HPC 805 > HPC 856 > HPC 834.

Fig. 3
figure 3

Growth profiles of Diaphorobacter strains in denitrifying medium (filled circle HPC 805, open circle HPC 815, filled inverted triangle HPC 820, open triangle HPC 821, filled square HPC 834, open square HPC 847, filled diamond HPC 848, open diamond HPC 856)

Full size image

In addition to growth, the removal of COD and ammonia by the different strains was monitored and is shown in Fig. 4. Strains HPC 805, HPC 815, HPC 821, and HPC 856 gave highest COD removal (85–93%) and ammonia removal (92–96%), while lower values were observed for remaining strains (70–80% COD removal, 80–85% ammonia removal). This correlated well with the higher growth rates seen in case of the above strains.

Fig. 4
figure 4

Percentage COD and ammonia removal by Diaphorobacter strains at the end of 96 h of incubation

Full size image

The corresponding nitrite released due to simultaneous nitrification of ammonia and denitrification of nitrate supplied in the medium increased with time, and a maximum \({\text{NO}}_2^ - \)–N concentration of 19.85 mM was detected after 96 h in case of strain HPC 815. The \({\text{NO}}_2^ - \)–N concentration observed in case of strains HPC 805 was slightly lower (17.85 mM), which decreased to 14.3 mM for HPC 821, HPC 848, and HPC 856, while lower values were observed for the remaining strains (Fig. 5). Further denitrification of nitrite was indicated by the decrease in levels of \({\text{NO}}_2^ - \)–N after 96 h; however, no \({\text{NO}}_3^ - \)was detected in the medium. Final pH increased to 9.0–9.5 at the end of incubation for all the isolates due to the utilization of acetate from the medium.

Fig. 5
figure 5

Nitrite release by Diaphorobacter strains in denitrifying medium (filled circle HPC 805, open circle HPC 815, filled inverted triangle HPC 820, open triangle HPC 821, filled square HPC 834, open square HPC 847, filled diamond HPC 848, open diamond HPC 856)

Full size image

Maximum activity of nitrate reductase was observed in HPC 856 (0.39 U ml−1) with a specific activity of 0.075 U mg−1 protein followed by 0.17, 0.09, and 0.01 U activity in case of HPC 805, HPC 815, and HPC 821, respectively. Nitrate reductase activities were very low in case of the remaining isolates. Reduction in the concentration of nitrite from reaction mixture in presence of cell-free extract of all the isolates indicated the existence of nitrite reductase activity in all the Diaphorobacter strains.

Discussion

The efficiency of wastewater treatment plants for nitrogen removal by SND can be judged from the amount of ammonia oxidized to nitrite and further to nitrate followed by reduction of nitrate under anoxic conditions. The process of SND can be considered economical provided nitrification stops at nitrite formation level (partial nitrification), which instead of nitrate serves as substrate for denitrification and is reduced to N2, thus resulting in reduction in usage of oxygen and reducing power. Conventionally, this process has been carried out in presence of organic matter or added substrates such as acetate, methanol, ethanol, citrate, etc. as electron donors, in addition to intracellular polymers such as polyhydroxybutyrate (Joo et al. 2005; Khan and Hiraishi 2001; Yoo et al. 1999; Zhao et al. 1999).

Partial nitrification is essential for economical operation of wastewater treatment systems by promoting accumulation of heterotrophic ammonia-oxidizing bacteria (AOB) and the washout of nitrite-oxidizing bacteria (NOB) responsible for oxidation of nitrite to nitrate (Laanbrock and Gerards 1993). The above explanation holds true for mixed bacterial systems in wastewater treatment plants where these two groups of bacteria coexist and their numbers varied based on the surrounding levels of DO, pH, ammonium concentration, temperature, sludge age, etc., with higher levels of ammonium favoring faster growth of AOB (Peng and Zhu 2006). Peng and Zhu have also reviewed the different growth patterns which were known to occur within the anaerobic AOB belonging to Planctomycetes (fast growth pattern at lower substrate and high ammonium concentrations and slow growth pattern in presence of high substrate concentration). However, similar growth patterns were also observed among the aerobic Diaphorobacter strains in our study because, according to Wong-Chong and Loehr (1978) and Gibbs et al. (2004), ammonium concentrations higher than 3.5 mg NH3–N l−1 inhibited further nitrification of nitrite to nitrate and allowed the denitrification to proceed under aerobic conditions. The lack of nitrate formation was attributed to the fact that pure cultures of heterotrophic AOB did not possess the ability of forming nitrate from nitrite, a reaction which was carried out by NOB in wastewater treatment systems (Peng and Zhu 2006).

Our results supported this theory, and simultaneous nitrification of ammonium at high rates by the eight Diaphorobacter isolates leading to formation of nitrite was observed, followed by denitrification resulting in a decrease in nitrite concentration on further incubation. In our study, presence of 7.5 mM \({\text{NH}}_4^ + \)–N (212 mg l−1 \({\text{NH}}_4^ + \)–N) allowed the complete reduction of 20.9 mM \({\text{NO}}_3^ - \)–N (162 mg l−1 \({\text{NO}}_3^ - \)–N) under shake flask conditions indicating efficient SND through common intermediate nitrite. The percent conversion of total available nitrogen (N; 7.5 mM \({\text{NH}}_4^ + \)–N + 20.9 mM \({\text{NO}}_3^ - \)–N = 28.4 mM \({\text{NO}}_3^ - \)) to nitrite varied between 17 and 70% for the different strains, and the absence of ammonium and nitrate indicated incorporation of the remaining amount of supplied nitrogen (30–83%) into the newly synthesized bacterial biomass. However, studies are required for evaluating the quality of off-gas with respect to the composition of gaseous products of denitrification especially to ascertain the absence of long-lived greenhouse gas such as N2O. Thus, from the above studies, it was concluded that this group of bacteria belonged to the heterotrophic AOB capable of performing partial nitrification with characteristic absence of nitrate formation in the medium.

A periplasmic nitrate reductase (NAP) has been suggested to be involved in the aerobic conversion of nitrate to nitrite. NAP activity was not repressed by either ammonium or oxygen (Bedzyk et al. 1999; Bell et al. 1990). Celen and Kilic (2004) isolated 39 cultures with NAP activity in presence of different nitrogen sources (nitrate alone, nitrate + ammonium, ammonium alone), and it was found that, when ammonium and nitrate were present together, aerobic nitrate reduction did not lead to high nitrite accumulation in contrast to higher nitrite accumulation when nitrate was the sole nitrogen source. From our studies, it was clear that both nitrate and nitrite reductase activities were present in all strains of isolated Diaphorobacter; however, further purification and characterization of the enzymes was required.

Very few reports are available on the diversity of this genus with three strains being isolated by Khan and Hiraishi (2002), which were named D. nitroreducens NA10B, KSP3, and KSP4, respectively. Similarly, Heylen et al. (2006) studied the diversity of denitrifying bacteria in sludge and of the 68 isolates belonging to the β-proteobacteria of family Comamonadaceae; only a single isolate belonging to genus Diaphorobacter (D. nitroreducens) was found to occur. Considering the rarity of this genus in effluent treatment plants, our study revealed a relatively high number of isolates of genus Diaphorobacter with capability of carrying out SND at different rates. Accordingly, further studies have to be carried out for SND by these bacteria in presence of alternative electron donors such as high-organic-containing wastewaters especially effluents from distilleries and dye industries. In addition, the characterization of purified nitrate and nitrite reductases from these organisms needs to be performed along with studies on nitrogen balance during the SND process.