Aerobic denitrifiers isolated from an alternating activated sludge system


  • Lone Frette,

    1. Department of General Microbiology, University of Copenhagen, Sølvgade 83H, 1307 Copenhagen K, Denmark
    Search for more papers by this author
  • Bo Gejlsbjerg,

    1. Department of General Microbiology, University of Copenhagen, Sølvgade 83H, 1307 Copenhagen K, Denmark
    Search for more papers by this author
  • Peter Westermann

    Corresponding author
    1. Department of General Microbiology, University of Copenhagen, Sølvgade 83H, 1307 Copenhagen K, Denmark
      Corresponding author. Tel.: +45 35 32 20 46; Fax: +45 35 32 20 40; E-mail:
    Search for more papers by this author

Corresponding author. Tel.: +45 35 32 20 46; Fax: +45 35 32 20 40; E-mail:


One hundred and sixty-nine bacterial strains were isolated from activated sludge from a waste water treatment basin operating under alternating aerobic/anaerobic conditions. Sixteen strains from a subsample of 23 nitrogen oxide reducers were true respiratory denitrifiers, and all denitrified under both anaerobic and aerobic conditions. REP-PCR band analysis showed different patterns for all strains. One strain (strain 1) produced large amounts of N2O and was studied in detail. Nitrous oxide was the major end product of denitrification by this strain, and NO2 was reduced more efficiently than NO3. The aerobic denitrification was most pronounced with NO2 as electron acceptor, and the reduction of NO2 was not coupled to NH+4 oxidation.


Denitrifying bacteria belong to a biochemically and taxonomically diverse group of facultatively anaerobic bacteria characterised by the ability to use nitrogen oxides (NO3 and NO2) as electron acceptors producing gaseous nitrogen, mainly N2[1]. The oxidation of organic material coupled to reduction of oxygen leads to a higher energy yield than reduction of nitrate, and oxygen is therefore commonly accepted to be the first choice as electron acceptor [2]. Thus, oxygen is known to repress the denitrification reductases [1].

In the last decade, however, denitrification under aerobic conditions in bacteria from at least four different genera has been reported [3–5], suggesting that aerobic denitrification does occur and might have evolved several times. Environments rich in nitrogen oxides and with frequent shifts in aerobic/anaerobic conditions could be expected to favour bacteria with constitutive denitrification enzymes, since reactivation and/or reinduction of the reductases after a period of aerobiosis would be competitively disadvantageous in the anaerobic period that followed. One of the most vigorous aerobic denitrifiers described, Thiosphaera pantotropha (now reclassified as Paracoccus denitrificans) was indeed isolated from an anaerobic waste water treatment plant receiving oxygen in some of the influents [4].

Alternating activated sludge treatment plants are specially designed to remove nitrogen from waste water by recirculating activated sludge containing nitrifying and denitrifying bacteria into basins, where the waste water is aerated at intervals. Therefore, the presence of aerobic denitrifiers could be expected in this environment. The aim of the present study was to investigate whether aerobic denitrification is common among denitrifiers in an alternating activated sludge treatment plant.

2Materials and methods


For enrichment and isolation both mineral and complex media were used, and for denitrification experiments only mineral medium was used. Mineral medium contained (per liter of distilled water): 50 mg CaCl2·2H2O, 1 mg Na2MoO4·2H2O, 1 mg CuSO4·5H2O, 216 mg MgSO4·7H2O, 10 mg FeSO4·7H2O, 10 mg disodium-EDTA (C10H14O8Na2·2H2O), 414 mg KCl and 10 mmol of phosphate buffer (pH 7.0), modified after Blösl and Conrad [6]. The complex medium was the nutrient broth glucose medium (NBG) described by Blösl and Conrad [6].

2.2Enrichment and isolation

Enrichment of denitrifying organisms was carried out in anaerobic NBG and mineral medium with 20 mM NaNO2 or NaNO3 as electron acceptor. Mineral medium was furthermore supplied with 10 mM of ethanol, sodium acetate or sodium succinate as electron donor and 5 mM NH4Cl to minimise assimilatory reduction of NO2 and NO3.

Bacteria were isolated from an alternating activated sludge basin at Måløv waste water treatment plant, Denmark. Activated sludge (1 ml) was used to inoculate autoclaved 30 ml tubes containing 9 ml enrichment medium, and 10-fold serial dilutions were made. To ensure anaerobiosis, the tubes were flushed with N2, closed with rubber stoppers and sealed with aluminium caps.

Growth was visually verified as turbidity, and to achieve pure cultures, streakings were made on plates containing solid enrichment medium, which were then incubated in an anaerobic (5% H2 and 95% N2) glove box (Coy Laboratory Products Inc.). Colonies were picked and transferred to anaerobic tubes containing 10 ml enrichment medium. Strains removing NO2 or NO3 in tubes, as measured by NO2 and NO3 strips (Reflectoquant, Merck) were considered to be denitrifiers. Purity was determined microscopically, and strains were stored at −80°C in 10% (v/v) glycerol and 90% (v/v) enrichment medium.

2.3Screening for N2O production

Aliquots of cell suspension, for each strain, were transferred to four 30 ml tubes containing 9 ml acetate mineral medium. Two tubes were flushed with N2 and two were incubated with an atmospheric headspace. Both aerobic and anaerobic tubes were amended with either 1 mM NaNO2 or 20 mM NaNO3. Tubes were mounted horizontally on a rotary shaker and incubated at room temperature for 24 h while shaking vigorously at 250 rpm. At 24 h, 1 ml of gas was sampled from the headspace and analysed for O2, CO2 and N2O on a gas chromatograph (HP 5890) equipped with an electron capture detector. Uninoculated tubes served as controls.

2.4Characterisation by REP-PCR

Strains were characterised by REP-PCR [7]. Template DNA was prepared as described by Johnsen et al. [8] by boiling 200 μl of cell suspension in MilliQ treated water (OD600=0.6) for 10 min in Eppendorf tubes sealed with cap locks. The tubes were immediately cooled on ice and centrifuged (20 000×g, 10 min, <9°C), and the supernatants were subsequently kept on ice. The REP-PCR was performed by Anita Jørgensen (Geochemical Department, Geological Survey of Denmark and Greenland, Denmark) as described by Johnsen et al. [8]. REP-PCR band patterns were compared manually to determine whether strains were multiplications of single clones.

2.5Denitrification by strain 1

Experiments were carried out in autoclaved 500 ml serum bottles closed with rubber stoppers. Aerobic experiments were carried out in bottles with an atmospheric headspace, and anaerobiosis was ensured by flushing bottles with N2. Experiments were carried out in mineral medium amended with 10 mM sodium acetate as electron donor and 10 mg N l−1 NaNO2 or 20 mg N l−1 NaNO3 as electron acceptor. Furthermore, 10 mg l−1 glucose, 50 mg l−1 yeast extract (Difco Laboratories, Detroit, MI), 100 mg l−1 bacto tryptone (Difco Laboratories) and 100 mg l−1 NaCl were added to all bottles to prevent nutrient deficiency.

Two aerobic experiments were carried out in 400 ml medium, one with 10 mg N l−1 NaNO2 and one with 20 mg N l−1 NaNO3. The O2 concentration in the medium was measured continuously with a Clarke-type O2 electrode. The medium was mixed by adjusting the stirring intensity on a magnetic stirrer.

Triplicate experiments were carried out in 300 ml medium. Half of the bottles amended with NaNO2 were also amended with 20 mg N l−1 NH4Cl. The bottles were shaken vigorously by incubating bottles horizontally on a rotary shaker at 250 rpm. To one set of anaerobic bottles amended with NO3 was added 10% (v/v) acetylene (C2H2) to inhibit the reduction of N2O to N2.

Every 30 min, 1 ml of gas was sampled for analysis of N2O, CO2 and O2 on a gas chromatograph (HP 5890) equipped with an electron capture detector. Simultaneously 10 ml of medium was sampled for immediate absorbance (OD590) measurements on a spectrophotometer (Pharmacia LKB Ultrospec III), and for later analysis of NH+4, NO2 and NO3 on a Aquatech autoanalyser. The samples were frozen until analysis.

Bottles were inoculated with resting cells to an OD590 of 0.14 corresponding to 3.5×108 cells ml−1, as measured by microscopical cell counts.

Resting cells were obtained by growing cultures in aerobic Luria-Bertani (LB) medium [9] supplemented with 1 g l−1 glucose and 20 mM NaNO3. At the mid-exponential phase, cells were centrifuged and washed twice in mineral medium. Resting cells were used within 3.5 h of harvest and kept on ice until used.


The amount of N2O in the liquid was calculated using the Ostwald coefficient [10]. For experiments performed in triplicate, the initial N2O production rates were calculated by linear regression as long as the NO3 or NO2 concentrations followed zero order kinetics. That is, until t=60 min for anaerobic bottles amended with NO2 or NO2+NH+4, t=90 min for aerobic bottles amended with NO2 or NO2+NH+4, t=180 min in anaerobic bottles amended with NO3, and t=270 min in aerobic bottles amended with NO3. The percentage of N2O produced was calculated from the amount of N2O produced when all NO3 or NO2 was consumed, relative to the amount of NO3 or NO2 added. For bottles amended with NO2 or NO2+NH+4, this was carried out at 90 min for anaerobic bottles, at 120 min for aerobic bottles, and at 270 min for anaerobic bottles amended with NO3. Student's t-test for unrelated samples was used to test if differences in N2O production between treatments were significant.


A total of 169 colonies were isolated from the plates. Of these, 32 strains were chosen by differences in colony morphology and cell appearance in a phase contrast microscope, and tested for removal of nitrogen oxides. Of the 32 strains, all bacteria isolated on mineral medium (22 strains), and only one strain isolated on NBG, were able to consume nitrogen oxides. These 23 nitrogen oxide consumers were subsequently screened for N2O production under aerobic and anaerobic conditions.

Twenty-one of the 23 strains produced N2O aerobically, while the two remaining strains (strain 3 and strain 20) did not produce N2O at all (Table 1). Especially strains with a high aerobic CO2 production consumed O2, but O2 concentrations at the end of the experiments always exceeded 17% of atmospheric saturation (data not shown). Table 1 gives a summary of the results. Eight strains (strains 1, 2, 6, 12, 14, 15, 16 and 22) produced N2O and CO2 under aerobic and anaerobic conditions and with either NO2 or NO3 as electron acceptor, and one strain (strain 8) did not produce CO2 when incubated anaerobically with NO2. Seven strains (strains 4, 5, 10, 11, 13, 17 and 18) produced CO2 under all incubation conditions, but only N2O under some. Five of these strains (strains 5, 10, 11, 13 and 18) did not produce N2O when incubated anaerobically with NO2, one strain (strain 4) produced only N2O when incubated anaerobically with NO3 or aerobically with NO2, and one strain (strain 17) only produced N2O when incubated aerobically with NO2. Three strains (strains 19, 20 and 23) produced N2O under all conditions, but only CO2 under one or two conditions, and two strains (strains 7 and 9) did not produce CO2 at all despite N2O production in all incubations. Two strains (strains 3 and 20) did not produce N2O and only produced CO2 in one incubation. One of these (strain 3) was originally isolated on NBG (data not shown).

Table 1.  Production of N2O and CO2 by 23 strains capable of consuming nitrogen oxides
StrainN2O production (ppm)aCO2 production (%)b
  1. Incubations were performed with either 20 mM NaNO3 or 1 mM NaNO2 anaerobically or with an atmospheric headspace.

  2. a+=1–100 ppm, ++=100–500, +++=≥500 ppm.

  3. b*=0–1%, **=≥1%.


This implies that 16 of the 23 strains screened for N2O and CO2 production were true respiratory denitrifiers. These strains were also able to denitrify aerobically, although two strains were only found to denitrify aerobically from NO2. The remaining seven strains could not be characterised as true respiratory denitrifiers due to the lack of CO2 production, even though five of these produced N2O.

Strains 3, 19 and 23 were not characterised by REP-PCR due to poor growth. Of the 20 strains characterised by REP-PCR, strains 7, 13 and 21 did not show any REP-PCR bands after one trial, but the remaining 17 strains all showed different REP-PCR patterns (Fig. 1).

Figure 1.

REP-PCR patterns for the different strains. The numbers below the bands indicate strains, and the numbers at the sides indicate REP-PCR product sizes (in bp). L indicates 0.1 kb ladder as marker.

Because of the interest in sources of N2O emission [11], strain 1 was chosen for further denitrification experiments due to its high N2O production under aerobic and anaerobic conditions (Table 1).

These experiments showed that strain 1 was able to reduce NO3 and NO2 to N2O under aerobic conditions (Fig. 2Fig. 3Fig. 4) as denitrification occurred while the O2 concentration in the medium varied between 6.4 and 7.7 mg O2 l−1 (Fig. 2) or the headspace O2 concentrations in triplicate aerobic experiments were 80% or more of atmospheric concentration (data not shown).

Figure 2.

Aerobic denitrification by strain 1. A: N2O production and O2 concentration in the liquid phase. B: Growth and CO2 production. A: ▴ N2O production from NO2, ■ N2O production from NO3, ▵ O2 concentration in the bottle amended with NO2, □ O2 concentration in the bottle amended with NO3. B: ▴ growth (NO2), ■ growth (NO3), ◯ CO2 production (NO2), • CO2 production (NO3). 4 mg NO2-N or 8 mg NO3-N was added.

Figure 3.

Denitrification by strain 1. A: N2O production, B: growth and C: CO2 production. ▴ NO2 aerobic; ▵ NO2 anaerobic; ■ NO3 aerobic; □ NO3 anaerobic; × NO3 anaerobic+C2H2 (n=3). Where invisible, the standard deviations are within the symbols; 3 mg NO2-N or 6 mg NO3-N was added.

Figure 4.

Production of N2O by strain 1. ▴ NO2 aerobic, ▵ NO2 anaerobic, • NH+4+NO2 aerobic; ◯ NH+4+NO2 anaerobic; 3 mg NO2-N or 3 mg NO2-N+6 mg NH+4-N was added (n=3). Where invisible, the standard deviations are within the symbols.

Under both aerobic and anaerobic conditions NO2 and NO3 were reduced to N2O without any accumulation of NH+4 (or NO2 when NO3 was reduced) (data not shown).

The anaerobic N2O production from NO3 was slightly higher when N2O reduction was inhibited by C2H2 (P<0.02) compared to uninhibited bottles (Fig. 3A), and 89% and 85% of the NO3 was converted to N2O, respectively. The initial N2O production rate from NO3 was higher when NO3 was reduced under anaerobic conditions (27 μg N min−1, r2=1.00) compared to aerobic conditions (2 μg N min−1, r2=0.95) (Fig. 3A).

Under both aerobic and anaerobic conditions the N2O production rates from NO2 exceeded the anaerobic N2O production rate from NO3 (Fig. 3A). The N2O production rate from NO2 was only slightly higher under anaerobic conditions (46 μg N min−1, r2=1.00), than under aerobic conditions (38 μg N min−1, r2=1.00), but the amount of N2O produced from NO2 was significantly higher under aerobic conditions (98%) compared to anaerobic conditions (91%) (P<0.01). Under anaerobic conditions a minor N2O consumption (0.5 μg N min−1) took place after all NO2 was removed (Fig. 3A). The rate was different from 0 (P<0.02). Consumption of N2O was not observed under aerobic conditions (Fig. 3A).

Under aerobic conditions, growth measured as OD590 continued throughout the experiment, even when NO3 or NO2 had been consumed, but under anaerobic conditions the growth was coupled to NO3 or NO2 reduction, since growth stopped when all NO3 or NO2 was consumed (Fig. 3B). Under anaerobic conditions, growth yield was 3.4 times higher when 6 mg NO3-N was used than when 3 mg NO2-N was used (Fig. 3B).

Amendment with NO3 or NO2 under aerobic conditions did not affect growth, compared to growth in aerobic bottles with no NO3 or NO2 amendment (data not shown).

The CO2 production was coupled to growth (Fig. 3C), and amendment with NO3 or NO2 did not affect the CO2 production under aerobic conditions (data not shown).

Neither N2O nor CO2 production was observed in controls amended with NO3 or NO2, but without strain 1, and no N2O was produced in bottles inoculated with strain 1, but without NO3 or NO2 (data not shown).

The initial N2O production rate from NO2 under anaerobic conditions was slightly higher when NH+4 was added as well (49 μg N min−1, r2=1.00 and 46 μg N min−1, r2=1.00, respectively). Likewise, the amount of N2O produced was significantly higher (105% and 91%, respectively; P<0.05) (Fig. 4). The opposite was the case under aerobic conditions, where additions of NH+4 resulted in a slightly lower initial N2O production rate (31 μg N min−1, r2=1.00 and 38 μg N min−1, respectively) and a smaller amount of N2O produced (94% and 98%, respectively; P<0.02) (Fig. 4).


Under anaerobic conditions, strain 1 showed a coupling between NO3 or NO2 reduction, growth and CO2 production. In our experiments, the theoretical energy yield from the reduction of NO3 to N2O is calculated to be 3.8 times higher than the energy yield from the reduction of NO2 to N2O. This is in correspondence with our findings that growth coupled to reduction of NO3 was 3.4 times higher than when coupled to reduction of NO2. Furthermore, the majority of the available carbon was nonfermentable, and most of the reduced nitrogen was recovered as gaseous nitrogen, which is characteristic of respiratory denitrifiers [12]. Strain 1 was able to reduce NO3 and NO2 under both anaerobic and aerobic conditions. The major end product of denitrification was N2O. Inhibition of N2O reductase by C2H2 resulted in a small increase in the N2O production, and the N2O production in bottles only amended with NO2 was highest under aerobic conditions. Therefore, we suggest that a minor N2O reduction took place under anaerobic conditions, but that this reduction was inhibited by oxygen. This corresponds well with earlier findings that N2O reductase is the most O2 sensitive denitrification enzyme in Pseudomonas nautica[13].

Under aerobic conditions the reduction of NO3 and NO2 did not affect growth and CO2 production, hence it seems that the overall carbon oxidation was the same, even though NO3 or NO2 and O2 respiration occurred concurrently. This is in contrast to the aerobic denitrifier Thiosphaera pantotropha, which increased its growth rate when both O2 and NO3 served as oxidants [4].

Even though aerobic denitrification does not result in increased aerobic growth, it can have other advantages in an alternating activated sludge system. Oxygen is known to inhibit denitrification [14, 15], and this can occur at the level of NO3 uptake [16] or by direct inhibition of the reduction enzymes [16, 17]. However, the fluctuating O2 concentrations in the sludge system could favour bacteria possessing constitutive denitrification enzymes. Strain 1 was characterised by a more effective NO2 reduction than NO3 reduction, especially under aerobic conditions. In contrast, Bonin et al. [13] found that NO2 reductase in Pseudomonas nautica was more O2 sensitive than the NO3 reductase. Under aerobic conditions in activated sludge, NH+4 will be oxidised to NO2 by nitrifying bacteria, and an effective NO2 removal will prevent a possible toxic effect of NO2[12]. Also, the NO2 removed during the aerobic phase in the activated sludge system will not be further oxidised to NO3, leaving a minor amount of oxidants for anaerobic denitrification. This will constitute a disadvantage to denitrifying bacteria, in which the denitrifying system is induced by anaerobiosis and the presence of NO3 or NO2[17].

Robertson et al. [3] discussed whether there is a universal link between aerobic denitrification and heterotrophic nitrification. Pseudomonas sp., Pseudomonas aureofaciens, Alcaligenes faecalis and Thiosphaera pantotropha have already been shown to act both as denitrifiers and as heterotrophic nitrifiers under aerobic conditions [3, 18, 19]. In contrast to these bacteria, strain 1 was not able to nitrify, since additions of NH+4 under aerobic conditions did not result in NH+4 removal (data not shown) or in an increase in either N2O production rate or amount of N2O produced. Under anaerobic conditions, however, NH+4 slightly stimulated the N2O production, perhaps because NH+4 prevented a minor assimilatory reduction of NO2.

Preliminary work indicates that strain 1 does not belong to any of the reported genera combining aerobic denitrification and heterotrophic nitrification. Further characterisation of this bacterium is in progress.

All 16 denitrifiers isolated in this work were able to denitrify under aerobic conditions. Of these, strain 13 did not show a REP-PCR pattern for unknown reasons, and since the other denitrifiers all showed different REP-PCR patterns, there is an indication that aerobic denitrification is a widespread capability among culturable denitrifiers in the alternating activated sludge system.

Probably because of overgrowth by fermentative bacteria, we had no success in isolating denitrifiers on complex medium, which has the consequence that all the true denitrifiers described above were isolated on mineral medium with a nonfermentative carbon source. This may mean that the isolated denitrifiers are nonrepresentative of the population of denitrifiers present in the activated sludge. However, laboratory studies on nitrogen transformation in activated sludge from the Måløv waste water treatment plant strongly indicated that NO2 was reduced to N2O aerobically. A coupling between NH+4 oxidation and NO2 reduction seemed responsible for the majority of the N2O produced under aerobic conditions (manuscript in preparation). Strain 1 did show more efficient aerobic reduction of NO2 than NO3, although no NH+4 oxidation was observed. In general, the isolated strains produced relatively more N2O from NO2 than from NO3 under aerobic conditions, indicating that the reduction of NO2 is more important than the reduction of NO3 under aerobic conditions in this environment. Normally, no N2O reduction is expected under aerobic conditions [1], but if such a reduction takes place, it cannot be excluded that NO3 is a better inducer of N2O reductase than NO2, as shown for Pseudomonas stutzeri[17]. A less sufficient induction of the N2O reductase by NO2 than by NO3 could therefore be responsible for the relatively high N2O production.

Aerobic denitrification is a rather newly discovered metabolic activity which obviously can be carried out by many different bacteria. Our study has shown that an alternating activated sludge system, by exerting selective pressure on the population of denitrifiers towards aerobic denitrification, can lead to a high diversity of bacteria carrying out this previously unaccepted metabolism. Further research is needed to investigate whether the presence of aerobic denitrifiers is a general trait in environments exposed to frequent aerobic/anaerobic shifts. Furthermore, the abundance of bacteria capable of aerobic denitrification and their in situ denitrifying activity have to be elucidated.

All the strains isolated in this study are now under more thorough characterisation, and probably many other aerobic denitrifiers await isolation from activated sludge systems.


We thank Anita Jørgensen, Geochemical Department, Geological Survey of Denmark and Greenland, Denmark for performing the REP-PCR.