OUP user menu

High levels of nitrifying bacteria in intermittently aerated reactors treating high ammonia wastewater

Cesar Mota, Jennifer Ridenoure, Jiayang Cheng, Francis L. de los Reyes III
DOI: http://dx.doi.org/10.1016/j.femsec.2005.05.001 391-400 First published online: 1 November 2005

Abstract

Changes in the fractions of ammonia-oxidizing bacteria and nitrite-oxidizing bacteria in two laboratory-scale reactors were investigated using 16S rRNA probe hybridizations. The reactors were operated in intermittent aeration mode and different aeration cycles to treat anaerobically digested swine wastewater with ammonia concentrations up to 175 mg NH3-N/L. High ammonia removals (>98.8%) were achieved even with increased nitrogen loads and lower aeration: non-aeration time ratios of 1 h:3 h. Nitrosomonas/Nitrosococcus mobilis were the dominant ammonia-oxidizing bacteria in the reactors. Nitrospira-like organisms were the dominant nitrite-oxidizing bacteria during most of the investigation, but were occasionally outcompeted by Nitrobacter. High levels of nitrifiers were measured in the biomass of both reactors, and ammonia-oxidizing bacteria and nitrite-oxidizing bacterial levels adjusted to changing aeration: non-aeration time ratios. Theoretical ammonia-oxidizer fractions, determined by a mathematical model, were comparable to the measured values, although the measured biomass fractions were different at each stage while the theoretical values remained approximately constant. Stable ammonia removals and no nitrite accumulation were observed even when rRNA levels of ammonia oxidizers and nitrite-oxidizers reached a minimum of 7.2% and 8.6% of total rRNA, respectively. Stable nitrogen removal performance at an aeration: non-aeration ratio of 1 h:3 h suggests the possibility of significant savings in operational costs.

Keywords
  • Ammonia oxidizing bacteria
  • Nitrite oxidizing bacteria
  • Nitrification
  • Oligonucleotide probes
  • Intermittent aeration
  • Swine wastewater

1 Introduction

The most common method of nitrogen removal from wastewater combines nitrification (the aerobic oxidation of ammonia to nitrite by ammonia-oxidizing bacteria [AOB], followed by oxidation of nitrite to nitrate by nitrite-oxidizing bacteria [NOB]), and anoxic denitrification (the reduction of nitrate or nitrite to N2, mostly catalyzed by heterotrophic bacteria). AOB and NOB are slow growers and influenced by many environmental factors such as pH, temperature, C/N ratio, unsaturated fatty acids, dissolved oxygen, ammonia, and nitrite concentrations [16]. The number and activity of nitrifying bacteria in wastewater treatment reactors are considered the rate-limiting parameters for the bioconversion of nitrogen in domestic wastewater [7]. In addition, the fraction of nitrifying bacteria in the biomass is important in nitrogen removal systems, since nitrifiers are considered poor competitors for oxygen compared to heterotrophs [8]. Nitrifying systems have traditionally been monitored using chemical parameters rather than microbiological data. The monitoring of key microorganisms by molecular methods might permit better prediction and control of process performance and efficiency [9].

Analysis of 16S rRNA gene sequences provides evidence that AOB form two monophyletic groups, one within the Beta- and one within the Gammaproteobacteria. The betaproteobacterial AOB comprise the genera Nitrosomonas and Nitrosospira. All members of the Nitrosospira cluster are very closely related to each other, whereas the Nitrosomonas cluster reveals six distinct lineages of descent [10]. The gammaproteobacterial ammonia oxidizers include the genera Nitrosococcus, with the exception of Nitrosococcus mobilis, which is closely related to Nitrosomonas and hence belong to the group of betaproteobacterial ammonia oxidizers. Nitrite-oxidizing bacteria form four phylogenetically distinct groups (Nitrobacter, Nitrococcus, Nitrospina, and Nitrospira). All isolates of gammaproteobacterial ammonia oxidizers, as well as members of the Nitrococcus and Nitrospina genera, are obligate halophilic bacteria [11], and therefore are not expected to be dominant in municipal wastewater treatment systems. Different members of the betaproteobacterial ammonia oxidizers, as well as members of the Nitrobacter and Nitrospira genera have been found to dominate different wastewater treatment plants or natural ecosystems [1215].

Biological nitrogen removal from swine wastewater is especially challenging due to the presence of high ammonia concentrations and organic matter in the waste stream, with ammonia concentrations that can be higher than 1100 mg NH3-N/L [16]. The rapid growth of large-scale, confined animal production and the common use of anaerobic lagoons for storing manure for long periods have led to concerns with ammonia emissions, fish kills, and contaminated ground and surface water [17]. One promising alternative for treatment of livestock waste consists of anaerobic digestion for removal of most of the organic content, followed by intermittent-aeration (cyclic aeration/non aeration periods) for removal of nitrogen. Intermittently aerated (IA) reactors can provide appropriate environmental conditions to support growth of microorganisms carrying out nitrification and denitrification in a single reactor. Cheng and Liu [16] showed that high nitrogen removal could be achieved in IA reactors, with the added benefit of lower alkalinity requirements and reactor construction costs. The aeration to non-aeration (ANA) time ratio is a critical parameter in the operation of IA reactors, as it has a significant effect on the efficiency and total cost of the system. Ideally, these reactors should be operated using the lowest ANA time ratio possible (longer non-aerated periods), resulting in enhanced denitrification without compromising nitrification.

In this study, we assessed a limited set of ANA time ratios at different influent nitrogen loads and determined their effects on nitrogen removal performance and nitrifier populations. The fractions of nitrifying bacteria (AOB and NOB) were monitored at the genus level using slot-blot hybridization and FISH with 16S rRNA-targeted probes, and the results were compared with values from mathematical modeling.

2 Materials and methods

2.1 Reactor design and operation

Two 6-liter plexiglas reactors (reactors A and B) were operated under intermittent aeration conditions. Both reactors were inoculated with the same sludge from the Neuse River Wastewater Treatment Plant (Raleigh, NC) and were operated under fixed conditions (ANA time ratio of 1:1) for 4 months before this investigation. The reactors treated anaerobically digested swine wastewater from day 1 to day 63, and a 4:1 (by volume) mixture of anaerobically digested (AD) and raw swine wastewater from day 64 to day 78 (acclimation period). Subsequently, the ammonia and COD loads were increased by feeding a 1:1 mixture of AD and raw wastewater from day 79 to day 180 (Fig. 1). Wastewater was obtained from a swine lagoon at the NCSU Lake Wheeler Road Field Laboratory. The hydraulic retention time (τ) and target solids retention time (θc) were 3 days and 20 days, respectively, and both reactors were operated at room temperature (23 ± 2°C). The reactor design allowed biomass to settle in the clarification zone and be recycled to the aeration zone. Air cycling was controlled using a solenoid valve activated by an electronic timer (ChronTrol Corp., San Diego, CA). Compressed air was regulated to 10 psi and airflow was controlled by a gas mass flow controller at 500 mL/min (Cole-Parmer Instrument Co., Vernon Hills, IL).

1

Diagram of operational conditions showing changes in influent wastewater composition and aeration cycles.

The original experimental design consisted of operating reactor A as a control, using an ANA time ratio of 1:1. However, on day 50 the gas mass flow controller in reactor A failed and aeration had to be controlled manually: turned off during the night and repeatedly turned on during the day. Chemical data from reactor A during the period with manual control (day 50 to day 120) were not collected. Aeration in reactor B was controlled using an initial ANA of 1:2 for 64 days (stage III), and an ANA time ratio of 1:3 during stage IV for 100 days, with increased ammonia loads (Fig. 1).

In summary, this investigation can be separated into four stages with the purpose of empirically determining the effect of several ANA time ratios at different nitrogen loads (Fig. 1):

  • (a)Stage I: relatively high ammonia concentrations (average of 115 mg/L) with ANA time ratio of 1:1.

  • Stage II: higher ammonia concentrations (average of 146 mg/L). Stage IIa: biomass subjected to long (overnight) non-aerated periods; Stage IIb: biomass subjected to ANA of 1:1.

  • Stage III: relatively high ammonia concentrations (average of 115 mg/L) and biomass subjected to ANA of 1:2.

  • Stage IV: higher ammonia concentrations (average of 146 mg/L) and biomass subjected to ANA of 1:3.

2.2 Analytical methods

Influent and effluent samples from each reactor were collected twice a week (completely mixed samples during the aerated period) and analyzed for TKN (total Kjeldahl nitrogen), NH3-N, NO3-N, NO2-N, soluble COD (chemical oxygen demand), TOC (total organic carbon), pH, TSS (total suspended solids), and VSS (volatile suspended solids) using Standard Methods [18]. Dissolved oxygen (DO) was measured using a YSI 52 DO meter and a YSI 5739 oxygen probe (YSI Inc., Yellow Springs, OH).

2.3 Bacterial cultures

Pure cultures of Nitrosomonas europaea (ATCC 25978, ATCC medium 2265), Nitrosospira multiformis (ATCC 25196, ATCC medium 929), and Nitrobacter agilis (ATCC 25384, ATCC medium 480) were grown aerobically in 0.5-liter flasks at 30°C. The pH was kept at 8.0 by periodic addition of 20% Na2CO3. Cells were harvested by centrifugation at 3200g and cell pellets were processed for extraction of RNA.

2.4 Nucleic acids extraction

Mixed liquor samples (14 mL) from reactors were centrifuged at 3200g for 5 min and stored at −80°C until extraction. RNA was extracted using a modified low-pH hot-phenol extraction procedure [19]. RNA concentrations were measured spectrophotometrically. Samples with low RNA content indicated failure of the RNA extraction procedure and were not included in the analysis. DNA was extracted as previously described by Burrell et al. [20].

2.5 PCR, cloning, and in vitro transcription

Since pure cultures of Nitrospira were not available, in vitro-transcribed 16 S rRNA was used as reference rRNA in membrane hybridizations. DNA was extracted from a reactor sample and amplified using bacterial primers (S-D-Bact-0011-a-S-17 and S-D-Bact-1492-b-a-16). PCR products were purified using a High Pure PCR Product Purification Kit (Roche, Indianapolis, IN) before amplification with specific primers. A set of PCR primers (Fw-Ntspa-311 and Rv-Ntspa-1463, Table 2) was designed to amplify DNA of microorganisms in the Nitrospira moscoviensis subgroup (Table 1). PCR was performed using a thermal cycler (Eppendorf Scientific Inc., Westbury, NY) under the following conditions: 94°C for 5 min, 30 cycles of 92°C for 1.0 min, 61°C for 1.0 min, and 72°C for 1.0 min, and a final extension at 72°C for 7.0 min. PCR product was evaluated using agarose gel electrophoresis. Final PCR product was purified and then cloned using a TA Cloning Kit (Invitrogen, Carlsbad, CA), as previously described [21]. The rRNA was produced by transcribing the selected insert using an AmpliScribe T7 In Vitro-Transcription Kit (Epicentre, Madison, WI), according to the manufacturer's instructions. Transcripts were purified with RNase-free DNase (Ambion Inc., Austin, TX). The in vitro-transcribed RNA had a total length of 1173 bp (GenBank Accession No. AY741094), with a region perfectly matching probe S-G-Ntspa-0685-a-A-22.

View this table:
2

Oligonucleotide probes used in slot blot and in situ hybridizations

ProbeTarget siteaSequence (5′–3′)Wash temperature (°C)% FormamideTarget organismsReference
NamebAbbreviation
S-∗-Univ-1390-a-A-18Univ13901390–1407GACGGGCGGTGTGTACAA440All organisms[22]
S-G-Nsv-0444-a-A-19Nsv443444–462CCGTGACCGTTTCGTTCCG5230Nitrosospira spp.[23,24]
S-G-Nsm-0156-a-A-19Nsm156156–174TATTAGCACATCTTTCGAT465Nitrosomonas spp., Nitrosococcus mobilis[23,24]
S-G-βAOB-1224-a-A-20Nso12251224–1243CGCCATTGTATTACGTGTGA5135Betaproteobacterial ammonia-oxidizing bacteria[23,24]
S-G-Nbac-1000-a-A-15Nb10001000–1014TGCGACCGGTCATGG42Nitrobacter spp.[23,23,24]
S-G-Nbac-1035-a-A-18cNIT31035–1052CCTGTGCTCCATGCTCCG40Nitrobacter spp.[12]
S-G-Ntspa-0685-a-A-22Ntspa685664–685CACCGGGAATTCCGCGCTCCTC63Nitrospira moscoviensis, Nitrospira marina[25]
  • a E. coli numbering.

  • b Probe names have been standardized according to OPD [26].

  • c Used with unlabeled competitor probe NIT3-competitor in an equimolar ratio.

View this table:
1

Forward and reverse primers used to amplify Nitrospira moscoviensis-like DNA

PrimerSequence (5′–3′)Target organisms
Fw-Nspa311ACACTGGCACTGCGACANitrospira moscoviensis
Rv-Nstpa1463TTCACCCCAATCATCGGTCANitrospira moscoviensis subgroup

2.6 Oligonucleotide probes

Oligonucleotide probes targeting the 16S rRNA of AOB and NOB, as well as universal probes were used in fluorescence in situ hybridizations (FISH) and membrane hybridizations (Table 2). Unlabeled probes were obtained from Sigma-Genosys (The Woodlands, TX) and probes labeled with Cy3 were obtained from Integrated DNA Technologies (Coralville, IA).

2.7 Slot-blot and in situ hybridizations

Mixed liquor samples (3 mL) were collected from reactors and fixed with 3 volumes of PFA fixative on ice for 2 h, as previously described [12]. Fluorescent in situ hybridization (FISH) was performed as previously described [27] using Cy-3 labeled probes (Table 2). Images were visualized with a Nikon Optiphot epifluorescence microscope (Nikon, Japan). Epifluorescence microscopy has been found to be adequate for analysis of non-granular, conventional activated sludge flocs [28]. Images were captured with a Sensys charge coupled device (CCD) camera (Photometrics, Newington, VA). To estimate the fraction of nitrifiers in the biomass using FISH, 30 pictures were taken per data point (3 pictures per well, 5 wells per slide for each of the specific probes and for DAPI). The total area of pixels with signal above background was determined using the image analysis software Metamorph (Universal Image Corporation, Downingtown, PA) after manual thresholding. The results were expressed as percentages of the DAPI stained area.

For slot-blot membrane hybridizations, oligonucleotide probes were 5′-end labeled with [γ-32P]ATP (ICN Radiochemicals, Irvine, CA) and T4 polynucleotide kinase (Promega Corp., Madison, WI) and purified with a Quickspin Oligo column (Roche Molecular Biochemicals, Indianapolis, IN). Membranes with immobilized RNA were hybridized as previously described [27] and washed at the appropriate wash temperatures (Table 2). The results were expressed as percentages of the total rRNA as measured with the universal probe.

2.8 Mathematical modeling

Modeling analysis was performed as described by Rittmann et al. [8] with the objective of estimating the theoretical biomass fractions of nitrifiers at each stage and comparing the results with the measured data. In summary, solids retention time (θx) and mixed liquor volatile suspended solids (Xv) are related in a non-linear fashion according to Eq. (1). It is assumed that VSS is composed of influent inert VSS, heterotrophs produced through oxidation of organic matter (BODL, or soluble COD), and nitrifiers produced through oxidation of ammonia

Embedded Image 1

where θ= hydraulic retention time (days), Yhet= true yield coefficient for heterotrophs (0.45 kg VSS/kg BODL), bhet= endogenous decay coefficient for heterotrophs (0.1 day−1), ΔBODL= removal of BODL across the system (mg BODL/L), Ynit= true yield for all nitrifiers (0.45 kg VSS/kg N), bnit= endogenous decay coefficient for nitrifiers (0.15 day−1), ΔTKN = TKN nitrified (mg N/L), and fd= fraction of newly synthesized biomass that is degradable by endogenous decay (0.8).

The ratio of active ammonia oxidizers to active biomass, (Xa)ao/Xa, was computed as follows:

Embedded Image 1

where Yao= 0.34 kg VSS/kg N [29]. ΔBOD was set equal to the influent soluble COD [8]. ΔTKN was set equal to the difference between influent and effluent TKN values less the ammonia assimilated into aerobic heterotrophs and not available for nitrification. The typical value of 0.124 kg N/kg VSS was assumed for the nitrogen content of heterotrophic biomass.

3 Results

3.1 Performance of reactors

The average influent ammonia and TKN concentrations for stages I and III (Fig. 1) were 115 mg NH3-N/L and 166 mg TKN/L, respectively (Table 3). During stages II and IV, nitrogen loads in the influent were increased by mixing anaerobically digested (AD) and raw wastewater, resulting in average influent concentrations of 146 mg NH3-N/L and 209 mg TKN/L. These ammonia concentrations are approximately 5 times the mean ammonia concentrations in municipal wastewater. Average influent VSS concentrations increased twofold, whereas influent TOC/TKN and soluble COD/TKN ratios did not change considerably when the influent changed from anaerobically digested (stages I and III) to a mix of anaerobically digested and raw swine wastewater (stages II and IV).

View this table:
3

Average concentrations in the influent wastewater

StageDaysMean influent concentrations (mg/L)a
VSSTKNNH3-NNO3-NNO2-NSoluble CODTOCSoluble COD/ TKNTOC/ TKN
I and III1–64336 (±113)166.0 (±8.4)114.9 (±5.7)3.8 (±3.7)0.0 (±0.0)280.2 (±27.5)182.2 (±37.8)1.7 (±0.14)1.1 (±0.2)
II and IV81–180733 (±457)208.7 (±44.5)146.4 (±27.5)0.4 (±0.7)0.0 (±0.0)362.5 (±73.0)208.1 (±47.0)1.8 (±0.3)1.0 (±0.4)
  • Standard deviations in parentheses.

  • a A total of 10 data points each for stages I and III, and 13 points each for stages II and IV were computed.

Table 4 shows mean effluent concentrations and removal efficiencies for reactor A during stages I and IIb. For both periods, average ammonia and TKN removal efficiencies were higher than 98% and 89%, respectively; and effluent nitrite concentrations were negligible, indicating practically complete nitrification for both stages. Similar results were observed for reactor B (Table 5) using ANA ratios of 1:2 and 1:3 (stages III and IV). Average ammonia and TKN removal efficiencies were higher than 98% and 90%, respectively, with negligible nitrite accumulation. Considerably lower effluent nitrate concentrations were observed for stage IIb (reactor A) and stage IV (reactor B), both when the influent had higher nitrogen and VSS contents. For all stages, removal of soluble COD was not substantial (average of 32%), whereas average TOC removal efficiencies exceeded 57%, suggesting that degradable compounds from the conversion of particulate organic carbon were the source of energy for the denitrifying heterotrophic biomass.

View this table:
4

Average effluent concentrations and removal efficiencies for reactor A

StageDaysEffluent concentrations (mg/L)aRemoval efficiency (%)a
pHVSSTKNNH3-NNO2-NNO3-NTKNNH3-NTINTOCSoluble COD
I1–508.0 (±0.2)2475 (±881)18.0 (±7.1)1.4 (±2.6)0.4 (±0.6)66.3 (±26.8)89.1 (±4.3)98.8 (±2.3)41.9 (±23.7)57.9 (±14.0)33.2 (±12.8)
IIb121–1808.1 (±0.3)719 (±336)12.0 (±7.9)0.9 (±2.2)0.3 (±0.4)23.7 (±16.1)94.6 (±4.0)99.4 (±1.4)83.7 (±12.0)64.6 (±14.7)29.5 (±28.5)
  • Standard deviations in parentheses.

  • a A total of 8 data points were computed for each period. No chemical data were collected for reactor A during stage IIa.

View this table:
5

Average effluent concentrations and removal efficiencies for reactor B

StageDaysEffluent concentrations (mg/L)aRemoval efficiency (%)a
PHVSSTKNNH3-NNO2-NNO3-NTKNNH3-NTINTOCSoluble COD
III1–647.8 (±0.3)1202 (±358)16.1 (±8.1)1.6 (±2.6)0.2 (±0.2)54.4 (±10.4)90.34 (±4.7)98.6 (±2.3)52.6 (±9.1)58.6 (±10.4)32.9 (±9.6)
IV81–1807.8 (±0.2)1628 (±469)13.7 (±6.2)0.4 (±0.6)0.4 (±0.7)16.6 (±13.2)93.77 (±3.0)99.7 (±0.4)83.8 (±18.8)68.1 (±9.4)46.1 (±10.7)
  • Standard deviations in parentheses.

  • a A total of 9 data points were computed for each period.

ANA time ratio of 1:1 resulted in DO concentrations of up to 5 mg/L during aerated periods and approximately 1 mg/L during non-aerated periods. For ANA time ratio of 1:3, DO concentrations varied from 2.7 to 0 mg/L during aerated and non-aerated periods, respectively. DO concentrations for ANA time ratio of 1:2 were not determined, but most likely were between the values determined for the ANA time ratios presented above. pH remained relatively constant in the reactors as indicated by the small standard deviation values (Tables 4 and 5). pH in reactor A averaged 8.0 throughout stages I and II, while in reactor B pH averaged 7.8 throughout stages III and IV. Average VSS concentrations were considerably lower in stages IIb and III (Tables 4 and 5), resulting in higher ammonia loads per cell.

3.2 Fraction of nitrifiers in the biomass

The fraction of nitrifiers in the biomass on different days was monitored using membrane hybridizations and the average values and respective standard deviations during each period are shown in Figs. 2 and 3 (results expressed as % 16S rRNA). Additionally, a sample from each period was randomly selected and the fraction of nitrifiers in the biomass was estimated using FISH (results expressed as % DAPI area). Both membrane hybridization and FISH results for the selected days are also shown in Figs. 2 and 3.

2

(a) Average fractions of total betaproteobacterial AOB, Nitrosomonas and Nitrosospira; (b) average fractions of Nitrospira and Nitrobacter for reactor A (stages I and II). Bars represent average data from membrane hybridizations (expressed as % 16S rRNA), with respective standard deviations. The bar with a day label corresponds to the membrane hybridization data for that specific day, and the error bar represents the standard deviation for the triplicate measurements. (●) correspond to FISH data for the same day (expressed as % DAPI area), with respective error bar. (♦) correspond to average theoretical AOB biomass fractions.

3

(a) Average fractions of total betaproteobacterial AOB, Nitrosomonas and Nitrosospira; (b) average fractions of Nitrospira and Nitrobacter for reactor B (stages III and IV). Bars represent average data from membrane hybridizations (expressed as % 16S rRNA), with respective standard deviations. The bar with a day label corresponds to the membrane hybridization data for that specific day, and the error bar represents the standard deviation for the triplicate measurements. (●) correspond to FISH data for the same day (expressed as % DAPI area), with respective error bar. (♦) correspond to average theoretical AOB biomass fractions.

In reactor A, total AOB in the betaproteobacterial group (measured with probe Nso1225) averaged 18% of total rRNA in the biomass during stage I. Mathematical modeling predicted an average active AOB biomass fraction of 21.5% (±0.7%). Nitrosomonas/Nitrosococcus mobilis (measured with probe Nsm156) were the dominant AOB with an average of 9.5% of total 16S rRNA, and Nitrosospira (measured with probe Nsv443) were only occasionally detected at low levels, averaging 1.4% of total 16S rRNA. Nitrospira was the dominant NOB, with an average of 15% 16S rRNA. Nitrobacter was also detected at an average of 7% of 16S rRNA. High standard deviations indicate significant variability in the fraction of nitrifiers in reactor A during stage I. During stage IIa, aeration was controlled manually: turned off during the night and repeatedly turned on during the day. Although the biomass was exposed to long periods (overnight) without aeration, fractions of nitrifiers remained high, as indicated by averages of 22% for total betaproteobacterial ammonia-oxidizer 16S rRNA and 16% and 8% for Nitrospira and Nitrobacter, respectively. Changing the aeration to an ANA time ratio of 1:1 had a positive effect on the fraction of AOB, as suggested by an average of 28% of betaproteobacterial ammonia-oxidizer rRNA, with Nitrosomonas still as the dominant betaproteobacterial ammonia-oxidizers. During the same period, mathematical modeling predicted the average active AOB fraction to be 19.8% (±4.6%) of total active biomass. Nitrosospira levels remained low, with an average of 3% of 16S rRNA. Levels of Nitrospira decreased to an average of 7%, and Nitrobacter levels increased to an average of 11% of 16S rRNA, indicating a shift in the NOB population.

Longer non-aeration periods slightly impacted the fraction of nitrifiers in reactor B (Fig. 3). Total fraction of betaproteobacterial ammonia-oxidizers decreased from an average of 28% in stage III (ANA time ratio of 1:2) to an average of 22% in stage IV (ANA time ratio of 1:3). Theoretical values for the same periods averaged 22.4% (±1.2%) and 23.2 (±2.9%), respectively. Fractions of NOB did not appear to be affected, as levels of Nitrospira averaged 11% and 8% and Nitrobacter levels averaged 3% and 5% for ANA time ratios of 1:1 and 1:3, respectively.

The combined level of Nitrosomonas and Nitrosospira remained within one standard deviation of the level of total betaproteobacterial AOB for most of the samples. However, there were exceptions, suggesting that there might be ammonia-oxidizers in the biomass that hybridized with probe Nso1225, but not with either probe Nsm156 or probe Nsv443. The lowest AOB and NOB levels measured were 7.2% and 8.6% of total rRNA (6.1% as Nitrospira, and 2.5% as Nitrobacter), respectively. The lowest theoretical active ammonia-oxidizer biomass fraction was 11.9%. These fractions were sufficient to allow stable nitrification performance, as there was no significant ammonia or nitrite accumulation during the entire investigation.

Levels of rRNA from membrane hybridizations and area percentages from FISH remained within one standard deviation for most of the samples, except for betaproteobacterial ammonia-oxidizer data points for stage III. ANOVA showed that the theoretical and measured values of active ammonia-oxidizer biomass fractions were significantly different (α= 0.05) in stages IIb and III, with theoretical values about 25% less than measured values. Mean theoretical ammonia-oxidizer biomass fractions did not change significantly throughout all the stages (ANOVA, α= 0.05).

4 Discussion

The average levels of nitrifying bacteria measured in the IA reactors using membrane hybridizations ranged from 18% to 30% for betaproteobacterial ammonia-oxidizers; 6% to 15% for Nitrospira; and 5% to 12% for Nitrobacter. These levels of nitrifiers are generally higher than most fractions reported. Previous studies in bench and full-scale activated sludge systems using membrane and in situ hybridizations reported AOB and NOB fractions in the ranges of 5–20% and 1.5–12 %, respectively [68,3032]. Although FISH and membrane hybridization quantify biomass fractions in different ways, both methods yielded comparable measurements of the fraction of nitrifiers for most samples analyzed.

Stoichiometric calculations based on operating parameters (influent COD and TKN removed) and previously determined kinetic parameters [8] indicated that the AOB levels should not change significantly among the different stages, with a predicted average of 21.7% of total biomass. The relatively constant predicted values were expected, since the model was based on the removal of TKN (including ammonia assimilated into biomass) and influent soluble COD, and the system showed stable ammonia oxidation performance. In contrast, experimentally measured ammonia-oxidizer biomass fractions were significantly different throughout stages. The highest average ammonia-oxidizer biomass fractions were measured for stages IIb and III, which corresponded to the stages with the lowest total VSS concentrations. Lower VSS concentrations resulted in increased ammonia-oxidizer biomass fractions, which may be due to higher ammonia loads per cell. These results highlight limitations of the model when there are significant changes in VSS concentrations. However, other explanation for the discrepancies between measured and estimated values comes from limitations with probe specificity. Nso1225, the probe used to target betaproteobacterial AOB in this study, has a mismatch with cultured representatives of the N. mobilis group [33]. In addition, novel AOB members might be present in the reactors.

Stoichiometrically, the oxidation of ammonia to nitrite (catalyzed by AOB) yields more electron equivalents than the oxidation of nitrite to nitrate (catalyzed by NOB). Therefore, one would expect a single sludge system receiving ammonia and organic N as the sole source of nitrogen to comprise higher levels of AOB than NOB. However, NOB levels (including Nitrospira and Nitrobacter) were at least the same as ammonia-oxidizer levels in 21% of the samples analyzed (data not shown). We hypothesize that low levels of oxygen during the initial period of non-aerated cycles and the existence of anoxic microenvironments within flocs could allow for concurrent nitrate reduction to nitrite, and nitrite oxidation by NOB. This hypothesis is supported by the low Ks of NOB for oxygen and nitrite. Thus, NOB could obtain substrate from ammonia oxidation during the aeration period, and from nitrate reduction during the initial period of non-aerated cycles. In addition, pure culture studies have indicated that N. moscoviensis[34] and N. agilis[35] have the ability to simultaneously incorporate CO2 and pyruvate in the presence of oxygen. This ability might give NOB competitive advantages in wastewater treatment systems. Alternatively, the low levels of AOB measured, as compared to NOB levels, could indicate the presence of novel AOB in the reactors that were undetected by the probes used, as previously mentioned.

The betaproteobacterial ammonia-oxidizing population was dominated by Nitrosomonas/Nitrosococcus mobilis in both reactors for all operating conditions. Nitrosospira were only occasionally detected in low levels, which suggests that this group of AOB probably played a negligible role in ammonia oxidation in the reactors. These results are in agreement with a number of previous studies that have suggested that Nitrosomonas can outcompete Nitrosospira in environments with high nitrogen and dissolved oxygen concentrations because of Nitrosomonas’ higher maximum growth rates [34,35]. Several molecular ecological investigations have led to the suggestion that members of the Nitrosospira cluster generally are the most ubiquitously distributed ammonia-oxidizers in nature [1,11]. Nitrosospira are thought to act as Ks strategists, competing well in environments with limited ammonia concentrations [35]. There is evidence that niche differentiation occurs not only at the genus level of AOB, but also at the species level, as the substrate affinities (Ks values) differ significantly among the AOB species. Within the genus Nitrosomonas, different Ks values reflect well the phylogenetically definable groups [11]. Nitrosomonas/Nitrosococcus mobilis are well known to occur in wastewater treatment plants. Purkhold et al. [36] analyzed samples from 11 nitrifying wastewater treatment plants and determined that in all but two plants only nitrosomonads could be detected. N. europaea has frequently been reported as the dominant ammonia-oxidizer in activated sludge [15,35]. Dionisi et al. [30] used competitive PCR to study nitrifiers in a full-scale wastewater treatment plant and determined that Nitrosomonas oligotropha was the dominant ammonia-oxidizer. Juretschko et al. [32] used FISH to investigate nitrifying bacteria in an industrial wastewater treatment plant with high ammonia concentrations and determined that Nitrosococcus mobilis was the dominant ammonia-oxidizer in the activated sludge.

Although levels of Nitrospira were generally higher than those of Nitrobacter, the results show that both genera of NOB coexisted, suggesting that both contributed to nitrite oxidation in the reactors. Recent studies have also reported coexistence of Nitrospira and Nitrobacter in biofilms and activated sludge. Daims et al. [14] used FISH to study the NOB community of a sequencing biofilm batch reactor treating wastewater with high ammonia and salt concentrations. In addition to Nitrospira, they detected smaller numbers of Nitrobacter. Liebig et al. [37] investigated the nitrifier community composition of a chemostat treating sludge reject water using FISH and detected Nitrospira and Nitrobacter at similar levels. Coskuner and Curtis [38] detected both NOB genera in a full-scale activated sludge plant. Nitrobacter is thought to outcompete Nitrospira in environments with high substrate concentrations (such as in culture media) due to their higher maximum growth rates, whereas Nitrospira are better competitors in environments with low substrate concentrations as a result of their higher affinity for nitrite and oxygen [31]. Nitrospira has been regarded as the key NOB in wastewater treatment due to frequent reports of quantitative dominance in activated sludge and biofilms [13,32,34]. However, factors selecting for Nitrospira or Nitrobacter remain unknown and deserve further investigation [14]. From the perspective of performance stability, it might be prudent to choose conditions favoring a more complex community of nitrifiers [6,14,15,39], as differences in environmental sensitivity among functionally similar species give stability to ecosystem processes [40]. In reactor A, the shifts in NOB community seemed to be related to the change in aeration conditions. During the period with manual control, the air supply was off during long periods (overnight), resulting in anoxic conditions for extended periods. These conditions favored Nitrospira presumably because of their higher affinity for oxygen. When automated aeration control was restored at 1 h ON:1 h OFF, higher oxygen levels seemed to favor Nitrobacter.

Nitrification performance was not negatively affected by the shifts in the AOB and NOB levels, as high ammonia removals and no significant nitrite accumulation were observed. Under the conditions evaluated in this study, the lowest measured ammonia-oxidizer levels of 7.2% of total rRNA and lowest NOB levels of 8.6% of total rRNA (6.1% as Nitrospira, and 2.5% as Nitrobacter) were sufficient for stable nitrification performance. However, it has not been determined whether these relatively low levels of nitrifiers can sustain the same performance for extended periods.

Our results provide an insight into the nitrifier populations in reactors treating high-nitrogen wastewater in response to changing operating conditions. High levels of nitrifiers were measured in the biomass of both reactors using ANA time ratios with non-aerated times up to 3 h. Low C/N ratio and high ammonia in the influent are likely the most important factors contributing to the high fraction of nitrifiers and consequently the stability of the nitrification process. Although slightly lower levels of total betaproteobacterial AOB were measured after the increase in the non-aeration period and change in the influent characteristics, ammonia removal efficiencies were not significantly affected. This illustrates that stable nitrification performance at lower aeration requirements can be achieved, resulting in significant savings in operational costs.

Acknowledgements

We thank Zhengzheng Hu for assistance with reactor operation and chemical analysis, and Daniel Noguera and Michael Hyman for providing bacterial cultures. This research was supported by the United States Department of Agriculture National Research Initiative Program (Grant 2001-35102–10783).

References

  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
View Abstract