Identification of the bacterial community involved in methane-dependent denitrification in activated sludge using DNA stable-isotope probing

Authors


  • Editor: Michael Wagner

Correspondence: Satoshi Tsuneda, Department of Chemical Engineering, Waseda University, 3-4-1 Ohkubo, Shinjuku-ku, Tokyo 169-8555, Japan. Tel.: +81 3 5286 3210; fax: +81 3 3209 3680; e-mail: stsuneda@waseda.jp

Abstract

Methane is used as an alternative carbon source in the denitrification of wastewater lacking organic carbon sources because it is nontoxic and may be efficiently produced by anaerobic biological processes. Methane-dependent denitrification (MDD) in the presence of oxygen requires the co-occurrence of methanotrophy and denitrification. Activated sludge was incubated with 13C-labeled methane in either a nitrate-containing medium or a nitrate-free medium. Then, bacterial and methanotrophic populations were analyzed by cloning analysis and terminal restriction fragment length polymorphism analysis targeting 16S rRNA gene and cloning analysis targeting pmoA genes. DNA-based stable-isotope probing (DNA-SIP) analysis of the 16S rRNA gene revealed an association of the Methylococcaceae and the Hyphomicrobiaceae in a MDD ecosystem. Furthermore, supplementation of nitrate stimulated methane consumption and the activity of methanotrophic populations (i.e. the stimulation of uncultivated relatives of distinct groups of the Methylococcaceae). In particular, uncultured type-X methanotrophs of Gammaproteobacteria were dominant when nitrate was added, i.e. in the MDD incubations. On the other hand, most methanotrophs (types I, II, and X methanotrophs) were found to have been labeled with 13C under nitrate-free conditions. This DNA-SIP study identifies key bacterial populations involved in a MDD ecosystem.

Introduction

Biological removal of nitrate and/or nitrite from wastewater, groundwater, landfill leachate or drinking water is commonly achieved by denitrification, which involves the reduction of nitrate, via nitrite and nitric oxide, to nitrous oxide or dinitrogen gas (Zumft, 1997). Denitrification requires an oxidizing nitrogen compound as electron acceptor and organic matter, hydrogen, or sulfur as electron donor. For wastewater that has a low C/N ratio or lacks readily biodegradable carbon sources, various organic compounds as external carbon sources, such as acetate, ethanol, glucose, or methanol, may be added to achieve a satisfactory degree of denitrification (Akunna et al., 1993). However, the use of such compounds incurs significant costs in an industrial-scale plant. Recently, methane has been proposed as an alternative, inexpensive, and effective carbon source because it is nontoxic, is produced as a biogas by anaerobic treatment, and is available in numerous existing treatment plants (Davies, 1973; Werner & Kayser, 1991; Amaral et al., 1995; Thalasso et al., 1997; Eisentraeger et al., 2001).

Methane-dependent denitrification (MDD) in the presence of oxygen has been demonstrated in many studies, but the mechanism of this process and the key microbial populations responsible are not yet known (Werner & Kayser, 1991; Amaral et al., 1995; Thalasso et al., 1997; Costa et al., 2000; Waki et al., 2002; Knowles, 2005). So far, although functional genes for dissimilatory nitrite and nitric oxide reductases have been found in some methanotrophic bacteria, there has been no evidence that they can carry out denitrification (Ye & Thomas, 2000). Thalasso et al. (1997) demonstrated that nitrate depletion in MDD could not be attributed only to nitrogen assimilation, but may also be in part due to denitrification. Thus, it has been suggested that MDD in the presence of oxygen occurs due to the coexistence of methanotrophic bacteria producing organic intermediates in the metabolism of methane and denitrifiers using the organic intermediates as electron donors. Some research groups have suggested that intermediates produced by methanotrophic bacteria could be acetate (Costa et al., 2000), citrate (Rhee & Fuhs, 1978), formaldehyde, methanol (Mechsner et al., 1985), polysaccharides and proteins (Nesterov et al., 1988).

Stable-isotope probing (SIP) techniques have been developed to identify active bacterial populations in complex natural environments. These techniques are based on the incorporation of stable isotopes into the DNA of cells consuming a labeled substrate of interest (Radajewski et al., 2000). To date, the feasibility of SIP has been demonstrated for a wide range of substrates and environments (Dumont & Murrell, 2005; Neufeld et al., 2007b). SIP techniques using one-carbon compounds such as methane and methanol have been successfully applied to identifying active methanotrophic or methylotrophic populations in natural environments (Radajewski et al., 2000, 2002; Lin et al., 2004; Lueders et al., 2004a; Neufeld et al., 2007a). Recently, SIP has been used to identify the active denitrifiers in wastewater treatment systems (Ginige et al., 2004, 2005; Osaka et al., 2006).

Methanotrophic bacteria include species in the Alphaproteobacteria (type II methanotrophic bacteria) and in the Gammaproteobacteria (types I and X methanotrophic bacteria). The oxidation of methane to methanol is catalyzed by either a soluble form or a particulate form of methane monooxygenase (sMMO and pMMO, respectively) (Hanson & Hanson, 1996). The pmoA gene encoding the α-subunit of the pMMO can be used as a functional phylogenetic marker for the identification of methanotrophic bacteria and is present in almost all known methanotrophic bacteria, except in some members of the genus Methylocella (Holmes et al., 1995). The pmoA gene phylogeny is generally consistent with the 16S rRNA gene-based phylogeny of methanotrophic bacteria (Murrell et al., 1998; Costello & Lidstrom, 1999).

MDD in the presence of oxygen may be hypothesized to be carried out by the interaction between two different types of microorganisms: methanotrophs, which play a role in producing some electron donors in the metabolism of methane, and denitrifiers, which use organic intermediates as electron donors for denitrification. The goal of this study was to identify microbial populations involved in MDD by DNA-SIP.

Materials and methods

Pure cultures

Hyphomicrobium denitrificans ATCC 51888 was grown on ATCC medium 784 AMS supplemented with 100 mM 12C-methanol or 13C-methanol (99%13C; Sigma, St Louis, MO) in the dark at 30 °C.

MDD assay

A sludge sample was taken from an urban wastewater treatment plant, the Ariake Water Reclamation Center in Tokyo, Japan, which uses an anaerobic–anoxic–oxic system. The basal medium contained (mg L−1): KNO3, 570; KH2PO4, 300; MgSO4·7H2O, 60; CaCl2·2H2O, 40; FeCl2·4H2O, 1; MnCl2·4H2O, 1; CoCl2·6H2O, 0.2; H3BO3, 0.12; ZnCl2, 0.02; CuCl2·2H2O, 0.02; NiCl2·6H2O, 0.02; Na2MO4·2H2O, 0.02; and Na2SeO4, 0.02, resulting in a pH of 6.8. The sludge was preincubated in a nitrate-containing medium under anoxic conditions in the dark at 20 °C for 3 days to eliminate residual organic carbon sources, and the total absence of soluble carbon sources in the supernatant was confirmed using a TOC analyzer (model TOC-5000A; Shimadzu Corp., Kyoto, Japan). The supernatant was removed from the sludge by centrifugation at 3300 g for 10 min, and the sludge was resuspended in the basal medium. The sludge incubations were conducted using 20 mL of this sludge [mixed liquor suspended solid, 3200 mg L−1; mixed liquor volatile suspended solid (MLVSS), 2500 mg L−1] and 35 mL of the basal medium in a 155 mL glass vial that was crimp-sealed with a butyl rubber stopper. Methane (10 mL, 0.4 mmol) was injected into the headspace of each vial, and each vial was incubated while being shaken at 100 r.p.m. in the dark at 30 °C. After >90% of the initial CH4 had been consumed, the vials were opened, flushed with air (500 mL) to remove evolved gas (e.g. CO2) and to replenish air in the headspaces, and were resealed. Then, methane (10 mL) was reinjected into the headspace. The assays for methane-dependent denitrification and the control test (Table 1) were performed in triplicate incubation runs.

Table 1.   Experimental set up and incubation conditions
 Carbon
source
HeadspaceNitrateAssay
Run 1CH4CH4, Air+Methane-dependent denitrification in the presence of oxygen
Run 2CH4CH4, Air+Sterilized inoculum
Run 3CH4CH4, Air+No inoculum
Run 4CH4CH4, Air Methane consumption in the absence of nitrate
Run 5 Air+Nitrate consumption under microaerophilic conditions
Run 6CH4CH4, Argon+Methane-dependent denitrification under anoxic conditions
Run 7 Argon+Endogenous denitrification under anoxic conditions
Run 8CH3OHArgon+Denitrification with methanol

Incubation for SIP analysis

Microcosms for SIP analysis were incubated under the same conditions as described above. Two types of sludge incubation were conducted using 20 mL of the sludge (i.e. 50 mg-MLVSS) placed in a 155 mL crimp-sealed glass vial: Run 1-type involved incubation with a nitrate-containing medium; and Run 4-type involved incubation with a nitrate-free medium. The 12CH4 or 13CH4 (10 mL, 0.4 mmol 99%13C; Cambridge Isotope Laboratories Inc., Andover, MA) was injected into the headspace of each vial, and each vial was incubated while being shaken at 100 r.p.m. in the dark at 30 °C. The incubation was continued in the same way until c. 4.5 mmol of CH4 had been consumed. Samplings for analysis were performed at several points during the experiment: Run 1-type at 0.42, 1.6, and 4.1 mmol of methane consumption; and Run 4-type at 0.39, 1.6 and 3.9 mmol of methane consumption. At each sampling, a microcosm was sacrificed and the sludge was stored at −80 °C before DNA extraction.

Analytical methods

Methane concentrations in the headspaces of each vial were measured using a GC380 gas chromatograph equipped with a flame ionization detector (GL Science Inc., Tokyo, Japan). Water samples were filtered through a Whatman GF/C glass microfiber filter (Whatman International Ltd, Maidstone, UK) and were stored at 4 °C until analysis. Nitrate and nitrite concentrations were measured with an ICS-3000 ion chromatograph (Dionex, Sunnyvale, CA). Ammonium concentrations were measured by colorimetry using the indophenol method (Weatherburn, 1967). Total nitrogen (T-N) and dissolved total nitrogen (DT-N) were measured at the beginning and end of incubation as nitrate-nitrogen by colorimetry using the salicylic acid method (Cataldo et al., 1975) after alkaline peroxodisulfate digestion (Ebina et al., 1983). Particulate nitrogen (P-N) was calculated as the difference between T-N and DT-N. Dissolved organic nitrogen (DO-N) was determined by subtracting ammonium-nitrogen, nitrate-nitrogen, and nitrite-nitrogen from DT-N. The contribution of denitrification to DT-N removal was calculated using the following equation.

image

Moreover, the contribution of assimilation to the DT-N removal was calculated using the following equation.

image

DNA extraction, CsCl density gradient centrifugation, fractionation and PicoGreen assay

Total DNA was extracted from 0.15 g (wet weight) of a sludge pellet using Isoplant (Nippon Gene Inc., Toyama, Japan) according to the manufacturer's instructions. The DNA was purified using a phenol/chloroform/isoamyl alcohol (25/24/1) solution, and was precipitated by adding ethanol and sodium acetate. DNA (c. 5 μg) was spun in CsCl gradients with an average density of 1.740 g mL−1. The density was determined with an AR200 digital refractometer (Reichert Inc., Depew, NY). Centrifugation was conducted at 184 400 g and 20 °C for >36 h (Lueders et al., 2004b). Centrifuged gradients were fractionated from bottom to top into 12 fractions (c. 400 μL) by displacement with water using a syringe pump (Harvard Apparatus Inc., Holliston, MA) at a flow rate of 800 μL min−1. The density of each fraction was determined with an AR200 digital refractometer (Leica Microsystems Inc., Buffalo, NY). DNA was precipitated from each fraction by adding two volumes of polyethylene glycol at 20 °C for 2 h and centrifuging at 20 400 g at 4 °C for 15 min. DNA pellets were washed once with 70% ethanol and were dissolved in 40 μL TE buffer. Subsequently, the total DNA concentration of each fraction was determined using the PicoGreen dsDNA Assay kit (Molecular Probes Inc., Eugene, OR).

PCR characterization

The following primer sets were used for PCR amplification: (1) forward primer 8f (Amann et al., 1995) and reverse primer 926r (Muyzer et al., 1995) for the amplification of 16S rRNA gene fragments, (2) forward primer A189 and reverse primer A682 (Holmes et al., 1995) for the amplification of pmoA gene fragments. The PCR mixture contained 0.5 μM concentrations of each primer, 200 μM concentrations of dNTP, 1.5 mM concentrations of MgCl2 for the 16S rRNA gene and pmoA gene, 2.5 U of rTaq DNA polymerase (Toyobo, Osaka, Japan), and 5 μL of 10 × PCR buffer for rTaq. The PCR amplifications of the 16S rRNA gene and pmoA gene were performed in a total volume of 50 μL in 0.2 mL reaction tubes using a model 9700 thermal cycler (Applied Biosystems, Foster City, CA) using the following programs: (1) 16S rRNA gene, 3 min at 94 °C, 30 cycles (40 s at 94 °C, 40 s at 56 °C, 40 s at 72 °C), and 3 min at 72 °C; (2) pmoA gene, a touchdown PCR program with annealing temperatures decreasing from 62 °C to 52 °C (decrease by 1 °C after two cycles), 3 min at 94 °C (40 s at 94 °C, 40 s at 62–52 °C, 40 s at 72 °C), 15 cycles (40 s at 94 °C, 40 s at 52 °C, 40 s at 72 °C), and 3 min at 72 °C. The presence of PCR products was confirmed by 2% agarose gel electrophoresis and the subsequent staining of the gels with ethidium bromide.

Cloning, sequencing and phylogenetic analysis

PCR products were purified by eluting the bands from 2% agarose gels using a Wizard SV gel and a PCR clean-up system (Promega, Madison, WI). The PCR amplicons were cloned using the pGEM-T Easy Vector System (Promega) according to the manufacturer's instructions. Then, colonies were randomly picked up with a needle and transferred to Insert Check Ready Solution (Toyobo). Clones were sequenced by an ABI PRISM 3100-Avant DNA sequencing system (Applied Biosystems) using a DYEnamic ET Terminator Cycle Sequencing kit (Amersham Biosciences, Freiburg, Germany) according to the manufacturer's instructions. The 16S rRNA gene sequences with more than 97% identity were considered to belong to the same operational taxonomic unit (OTU). The pmoA gene sequences exhibiting more than 98% identity were considered to belong to the same OTU (Yan et al., 2006). A database search was conducted using blast from the DNA Data Bank of Japan (DDBJ). Sequences determined in this study and those retrieved from the database were aligned using clustal w (Thompson et al., 1994). Phylogenetic trees were constructed using a neighbor-joining algorithm (Saitou & Nei, 1987).

Terminal restriction fragment length polymorphism (T-RFLP) analyses

T-RFLP analysis of 16S rRNA genes was carried out using the forward primer 8f labeled at the 5′ end with the dye 6-carboxy-fluorescein and the reverse primer 926r. After purification of PCR products with a Wizard SV gel and a PCR clean-up system (Promega), 4 μL of the PCR-products were digested with 10 U of the restriction enzyme MspI (TaKaRa) in the manufacturer's recommended reaction buffers for 4 h at 37 °C. The enzyme was subsequently inactivated by incubation at 65 °C for 20 min. Aliquots of the digested amplicons were desalted by ethanol precipitation. Desalted digests were suspended in 15 μL of Hi-Di formamide (Applied Biosystems) containing GeneScan-1000 size standard (Applied Biosystems), denatured (5 min at 94 °C), cooled on ice, and resolved by an ABI PRISM 3100-Avant Genetic Analyzer automated sequence analyzer (Applied Biosystems) using the genescan software (Applied Biosystems).

Nucleotide sequence accession numbers

16S rRNA gene sequences determined in this study were deposited under accession numbers AB280265AB280414. PmoA sequences determined in this study were deposited under accession numbers AB280415AB280427.

Results

Consumption of methane and nitrate

The consumption of methane was greatly affected by supplementation with nitrate (Fig. 1a). In Run 1, a sludge with nitrate supplementation consumed >90% of the initial concentration of methane after 54 h of incubation. In contrast, in Run 4, a sludge without nitrate supplementation took 113 h of incubation to consume >90% of the initial CH4. The addition of nitrate resulted in differences in the incubation time it took for c. 4.5 mmol of methane to be consumed: in Run 1, it took 134 h, whereas in Run 4, it took 415 h. In addition, consumption of methane was never observed under anoxic conditions during experimental periods. This might be due to the extremely slow growth of a microbial consortium exhibiting anoxic methane oxidation (AOM) coupled to denitrification (Raghoebarsing et al., 2006) or to the absence of such a consortium in the sludge. Nitrate concentrations obviously decreased with consumption of methane in Run 1 (Fig. 1b). In addition, decrease of nitrate was not observed in the absence of external carbon sources (Run 5). In Runs 6 and 7, the decrease of nitrate may have been due to denitrification with organic substrates produced by self-digestion of the sludge or endogenous denitrification. In the assay with methanol as the carbon source, the sludge exhibited the strongest denitrification activity (Run 8).

Figure 1.

 Time course of consumption of methane (a) and NO3-N and NO2-N (b). In a panel (b), NO3-N is shown by the solid line and NO2-N is shown by the broken line; NO2-N was detected only in Run 8. Error bars, which indicate SDs (n=3), are smaller than the symbols.

Nitrogen balance

Nitrogen balance at the beginning and at the end of incubation is shown in Table 2. When the vials were incubated with methane as the external carbon source (Run 1), T-N concentrations in the medium decreased more than in incubated vials that were free of external carbon (Run 5). This suggested that the sludge in Run 1 exhibited MDD activity. However, the T-N removal rates by MDD (22.1±2.5%; n=3) were less than that by denitrification with methanol as a carbon source (52.8±1.5%; n=3). These differences might be attributed to the different fates of depleted nitrate in each of the incubations. Denitrification with methane contributed 40.1±3.1% (n=3) to the dissolved total nitrogen (DT-N) removal from the medium, whereas 59.9±3.1% (n=3) of the removed DT-N accumulated in the sludge (Run 1). This suggests that most of the nitrate in the medium was assimilated for bacterial growth in MDD. On the other hand, denitrification with methanol contributed 87.5±5.4% (n=3) to the DT-N removal from the medium, whereas 12.5±5.4% (n=3) of the removed DT-N was accumulated as biomass in the sludge (Run 8).

Table 2.   Nitrogen balance at beginning and end of incubation
IncubationT-N (mg L−1)DT-N (mg L−1)P-N (mg L−1)*NO3-N (mg L−1)NO2-N (mg L−1)NH4-N (mg L−1)DO-N (mg L−1)
  • *

    Particulate nitrogen (P-N) was calculated as the difference between total nitrogen (T-N) and dissolved total nitrogen (DT-N).

  • Dissolved organic nitrogen (DO-N) was calculated from subtraction of ammonium-nitrgen, nitrate-nitrogen, and nitrite-nitrogen from DT-N.

Initial521.9 ± 4.8321.5 ± 10.0200.3 ± 12.2295.4 ± 7.900.3 ± 0.325.8 ± 15.3
Run 1406.5 ± 13.834.2 ± 5.9372.3 ± 12.718.9 ± 1.606.0 ± 0.89.3 ± 8.0
Run 5496.4 ± 10.0382.1 ± 6.9114.3 ± 12.8360.5 ± 1.500.8 ± 0.020.8 ± 9.9
Run 8246.3 ± 8.35.9 ± 0.5240.4 ± 10.7000.3 ± 0.05.6 ± 0.6

CsCl density gradient centrifugation for separation of 12C-DNA and 13C-DNA

Characterization for separation of 12C-DNA and 13C-DNA was carried out by fluorometric determination of total DNA of each fraction ranging from 1.69 to 1.78 g mL−1 CsCl buoyant density. The fractionation using an isopycnic centrifugation is the most important step in SIP analysis. Therefore, we demonstrated successful separation of 12C-DNA and 13C-DNA extracted from pure cultures of H. denitrificans (Fig. 2a). Furthermore, we also showed the profiles of the incubation controls with 12CH4 (4.1 mmol consumption) (broken lines in Fig. 2b and c). We analyzed samples at several points during the experiment: Run 1 at 0.42, 1.6 and 4.1 mmol; and Run 4 at 0.39, 1.6 and 3.9 mmol. Although the position of DNA extracted from sludge consuming the initial CH4 was almost the same as that of the incubation controls with 12CH4, DNA peaks were steadily shifted to the ‘heavy’ density fraction with the consumption of 13CH4 (Fig. 2b and c).

Figure 2.

 Quantitative profiles of CsCl density gradient fractions by PicoGreen assays. (a) Quantitative distribution of total DNA extracted from fully 13C-labeled (•) and unlabeled Hyphomicrobium denitrificans (○). (b) Quantitative distribution of total DNA extracted from 13CH4- or 12CH4-incubated sludge samples at different consumption amounts of methane (0.42 mmol of 13CH4, ▪; 1.6 mmol of 13CH4, ▴; 4.1 mmol of 13CH4, •; 4.1 mmol of 12CH4, ○) in Run 1. (c) Quantitative distribution of total DNA extracted from 13CH4-incubated sludge samples at different consumption amounts of methane (0.39 mmol of 13CH4, ▪; 1.6 mmol of 13CH4, ▴; 3.9 mmol of 13CH4, •; 4.1 mmol of 12CH4, ○) in Run 4. The density-range characteristics for ‘light’ DNA are shaded in gray.

T-RFLP fingerprinting of bacterial populations in gradient fractions

T-RFLP fingerprinting of bacterial populations was performed from gradient fractions at three different points in time of 13CH4 consumption (supplementary Fig. S1). At an early point in time of 13CH4 consumption in Run 1 (0.42 mmol of 13CH4), some characteristic T-RFs (i.e. 135, 141, 436, 437, 439, 455, 487 and 490 bp) were detected from the ‘heavy’ fraction (1.741 g mL−1), indicating key bacterial populations involved in MDD in sewage sludge. T-RFLP profiles also showed that the diversity of T-RFs in this ‘heavy’ fraction was clearly reduced in comparison to those of the ‘light’ fractions. After 1.6 mmol of 13CH4 consumption in Run 1, some T-RFs (135, 141, 437, 439 and 455 bp) were detected at a further ‘heavy’ fraction (1.758 g mL−1). After 4.1 mmol of 13CH4 consumption in Run 1, some additional characteristic T-RFs (e.g. 529 bp) other than the above-mentioned T-RFs, were contained in the ‘heavy’ fractions (1.741 and 1.751 g mL−1). In T-RFLP analysis of sludge samples consuming 13CH4 under nitrate-free conditions (Run 4), two major T-RFs (436 and 487 bp) and some minor T-RFs (e.g. 135, 141 and 152 bp) were detected in the ‘heavy’ fraction (1.740 g mL−1) at an early time point (0.39 mmol of 13CH4 consumption). T-RFLP profiles of the ‘heavy’ fractions changed dramatically due to the further 13C labeling in Run 4, which resulted in many T-RFs (e.g. 114, 455 bp) in the ‘heavy’ fractions. At the point in time of c. 4 mmol of 13CH4 consumption, T-RFLP profiles of the ‘heavy’ fraction in Run 1 (1.751 g mL−1) and Run 4 (1.748 g mL−1) were quite different. Additionally, it was evident that the above-mentioned characteristic T-RFs shifted to ‘heavy’ fractions in comparison with fingerprints generated as 12CH4 incubation controls (supplementary Figs S1c and S1d).

Phylogenetic analysis of bacterial populations and assignment to T-RFs

After c. 4 mmol of 13CH4 consumption, 16S rRNA gene clone libraries (denoted R1.13 and R4.13, respectively) were generated from the ‘heavy’ fraction (Run 1, 1.751 g mL−1; Run 4, 1.758 g mL−1) in order to identify the likely key players in each incubation and to assign phylogenetic groups to distinct T-RFs. Additionally, the 16S rRNA gene clone library (denoted R1) was also constructed from DNA of Run 1 (4.1 mmol 13CH4 consumption) before isopycnic centrifugation. The phylogenetic affiliations of sequenced clones and their affiliations with distinct T-RFs are summarized in Table 3. In contrast to the clone library R1 (52 OTUs; n=62), the clone library R1.13 (33 OTUs; n=84) had a completely different bacterial composition and diversity. The clone library R1.13 showed a clear predominance of sequences related to the Methylococcaceae of the Gammaproteobacteria (e.g. T-RF of 135, 141, 455, and 487 bp) and to the Hyphomicrobiaceae of the Alphaproteobacteria (T-RF of 436, 437, and 439 bp). Of these clones, 24% of all clones were affiliated with the novel uncultured type X methanotrophs related to the Methylocaldum (Fig. 3). Two of the clones were related to the Methylophilaceae of the Betaproteobacteria (T-RF of 490 and 493 bp), known as the methylotrophs. Moreover, some sequences of nonmethanotrophs and nonmethylotrophs (e.g. Comamonadaceae, Myxococcales, Rhodocyclaceae) were also detected within this library (34%).

Table 3.   Phylogenetic affiliations and numbers of 16S rRNA gene clones
Phylogenetic groupR1
No. of clones
R1.13§R4.13§
No. of clonesT-RF length (bp)No. of clonesT-RF length (bp)
  • Characteristic T-RFs for different clone groups are given.

  • Clone library ‘R1’ was constructed from DNA of Run 1 without CsCl density gradient centrifugation.

  • §

    Clone libraries ‘R1.13’ and ‘R4.13’ were constructed from 13C-DNA separated by CsCl gradient centrifugation.

  • Terminal restriction fragment length of each clone is shown in base pairs. T-RFs detected for more than one clone within one phylogenetic group are indicated in boldface. T-RFs detected in more than one phylogenetic group are marked with an asterisk (*).

Alphaproteobacteria95150, 152, 4031379, 113, 132, 152*, 403, 441
 Hyphomicrobiaceae424436, 437, 4392436*, 439
 Methylocystaceae (type II MOB)111515150, 152*
Betaproteobacteria2  11140, 141*, 430*, 488,
 Comamonadaceae341392139*, 490*
 Methylophilaceae12490, 4933114
 Rhodocyclaceae345291370
Gammaproteobacteria46123, 224, 238, 487*1087, 193, 240, 436*, 490*
 Methylococcaceae (type I & X MOB)128135, 141, 444, 455, 487*19135, 141*, 144, 444, 455, 456, 487
Deltaproteobacteria, Myxococcales4577, 78, 491,678, 151, 152*, 506
Acidobacteria1  5171, 201, 285
Actinobacteria2    
Bacteroidetes12  991, 93, 181, 541, 542
Chloroflexi6  4147, 164, 512, 518
Cyanobacteria32321495
Gemmatimonadetes13130, 129277, 131
Nitrospirae1    
Planctomycetes   3207, 268, 291*
Verrucomicrobia1    
Candidate division OP11   1166
Candidate division Termite group 1   2291
Unidentified bacteria3    
Figure 3.

 Phylogenetic affiliation of clones derived from methanotrophs and methylotrophs by neighbor-joining analysis. The partial 16S rRNA gene sequences obtained from the 13C-DNA fraction in Run 1 and 13C-DNA fraction in Run 4 are labeled ‘R1.13-’, and ‘R4.13-’, respectively. The number of clones assigned to each sequenced OTU with greater than 97% identity is shown in parentheses. Geobacter metallireducens (accession no. L07834) is used as the outgroup. Closed circles (i.e. bootstrap values, >75% derived from 1000 replicates) and open circles (i.e. bootstrap values, 50–75% derived from 1000 replicates) are indicated at branch points. The scale bar represents 5% sequence divergence.

Also in the clone library R4.13 (65 OTUs; n=99), clones related to the Methylococcaceae of the Gammaproteobacteria (e.g. T-RF of 135, 141, 455, and 487 bp) were predominant (19% of all clones). Furthermore, some sequences detected in the clone library R4.13 were related to other methanotrophs and methylotrophs, i.e. Methylocystaceae (T-RF of 150 and 152 bp), Hyphomicrobiaceae (T-RF of 436 and 439 bp), and Methylophilaceae (T-RF of 114 bp). However, most of clones in the clone library R4.13 (73%) were related to sequences other than those of methanotrophs and methylotrophs.

Identification of active methanotrophs based on pmoA gene clone library

Our results show that supplementation with nitrate affected the rate of methane consumption by the sludge (Fig. 1a). To characterize how the methanotrophic populations in the sludge were affected by the supplementation with nitrate, pmoA gene clone libraries were constructed from the same DNA templates as the 16S rRNA gene clone libraries: DNA template before isopycnic centrifugation of Run 1 (denoted ‘R1.PmoA’, 44 clones), the ‘heavy’ gradient fraction from Run 1 (denoted ‘R1.13.PmoA’, 46 clones), and the ‘heavy’ gradient fraction from Run 4 (denoted ‘R4.13.PmoA’, 45 clones). Of these clones, three OTUs were obtained from the ‘R1.PmoA’ library, one OTU was obtained from the ‘R1.13.PmoA’ library, and nine OTUs were obtained from the ‘R4.13.PmoA’ library. The amino acid sequences of each OTU were divided into six clusters, and the phylogenetic analysis of three pmoA clone libraries revealed that the methanotrophic populations were strongly affected by the supplementation with nitrate (Fig. 4).

Figure 4.

 Neighbor-joining analysis of partial pmoA gene products (177 amino acids) from each clone library. The partial pmoA sequences obtained from non-SIP treated DNA sample and 13C-DNA fraction in Run 1 and 13C-DNA fraction in Run 4 are labeled ‘R1.PmoA-’, ‘R1.13.PmoA-’, and ‘R4.13.PmoA-’, respectively. The number of clones assigned to each sequenced OTU with greater than 98% identity is shown in parentheses. The amoA from Nitrosomonas europaea (accession no. AF037107) is used as the outgroup. Closed circles (i.e. bootstrap values, >75% derived from 1000 replicates) and open circles (i.e., bootstrap values, 50% to 75% derived from 1000 replicates) are indicated at branch points. The scale bar=10% amino acid substitution.

Four OTUs from the ‘R4.13.PmoA’ library belonged to cluster I, which were closely related to the pmoA of type II methanotrophs (i.e. Methylocystis species). In cluster II, two OTUs were closely related to the pmoA of type I methanotrophs (i.e. Methylomonas sp.). The pmoA of type X methanotrophs might belong to clusters III to VI. The clones of Cluster VI (R1.PmoA-2, R1.13.PmoA-1 and R4.13.PmoA-1) were most abundant in all clone libraries, and they were closely related to the pmoA of an uncultured bacterium (AF211874).

Discussion

In this study, DNA-SIP was used to identify the active microorganisms in wastewater that were involved in MDD in the presence of oxygen. The clone library R1.13, which was generated from a ‘heavy’ fraction after c. 4 mmol of 13CH4 consumption, contained a high number of clones related to the Methylococcaceae and Hyphomicrobiaceae. Additionally, clones other than those of these bacteria were also present in a ‘heavy’ fraction (e.g. Betaproteobacteria). This was probably due to the prolonged supply of 13CH4, i.e. heavy 13C labeling led to enrichment and selection of certain bacterial populations. Thus, time-course T-RFLP analysis combined with DNA-SIP was conducted in order to obtain information on key bacteria at early points in time. T-RFLP profiles of a ‘heavy’ fraction at an early point in time showed the predominance of the T-RFs derived from methanotrophs (135, 141, 455 and 487 bp) and methylotrophs (i.e. 436, 437 and 439 bp), which supported the idea that the Methylococcaceae and Hyphomicrobiaceae are key bacterial populations in the MDD ecosystem in this study.

DNA-SIP results showed an association of methanotrophs (i.e. Methylococcaceae) and methylotrophs (e.g. Hyphomicrobiaceae and Methylophilaceae) in the MDD ecosystem. The 13C labeling of methylotrophs might be due to indirect labeling effects via organic metabolites from methanotrophs because methylotrophs cannot directly utilize methane. Although it is difficult to identify key substrates for this trophic link from only the ecological data of DNA-SIP, we would like to suggest the following. In the metabolism of methanotrophs, possible one-carbon organic metabolites are methanol, formaldehyde, and formate (Hanson & Hanson, 1996). Methanol is formed from methane by methane monooxygenases when methanotrophs utilize NADH as a reducing agent. Then, methanol is further oxidized to formaldehyde, which plays a central role as an intermediate in assimilation for biosynthesis of cell materials and in dissimilation. Formaldehyde is oxidized sequentially to formate and carbon dioxide, when most of the reducing power (i.e. NADH) required for the metabolism of methane is produced. From the point of view of energetics in methanotrophs, the oxidation of methanol is an important step in the production of formaldehyde to gain reducing power. Therefore, formate might be a key substrate for trophic links in MDD ecosystems, because it is implausible that these compounds were transferred from methanotrophs to methylotrophs. However, it is known that methanol and formaldehyde accumulate in methanotrophic cultures (Hanson & Hanson, 1996). Therefore, we cannot rule out the possibility that methanol and formaldehyde are key substrates involved in trophic interactions of methanotrophs and methylotrophs in MDD ecosystems. Nevertheless, the identities of key substrates for trophic links must be confirmed by experiments using pure cultures and cocultures of methanotrophs and methylotrophs.

Nitrogen balance in Run 1 showed the occurrence of MDD in the presence of oxygen, i.e. 40% of the initial dissolved total nitrogen, mostly nitrate-nitrogen, was removed from the medium in some gas form. Known methanotrophs have not been observed to perform denitrification, although some of them have functional genes involved in denitrification (Ye & Thomas, 2000). This, in turn, suggests that denitrification in MDD ecosystems may be carried out by some of the 13C-labeled bacteria other than the methanotrophs. The major phylotype within nonmethanotrophs was related to the Hyphomicrobiaceae (T-RF of 436, 437 and 439 bp), which are restricted facultative methylotrophs and grow using one-carbon compounds such as methanol and formate as a carbon source and oxygen or nitrate/nitrite as the terminal electron acceptor. To date, some members of the Hyphomicrobiaceae have been identified as methylotrophic denitrifiers in sewage sludge (Holm et al., 1996; Osaka et al., 2006). Therefore, our results indicate that these bacteria may play a major role in denitrification in MDD ecosystems. Additionally, the uncultured Betaproteobacteria (e.g. Methylophilaceae, Rhodocyclaceae) were also detected in 13C clone library profiles and T-RFLP profiles in ‘heavy’ fractions. In some previous studies, these bacteria were shown to be important denitrifying bacterial populations in wastewater treatment systems (Wagner & Loy, 2002; Ginige et al., 2005; Osaka et al., 2006). Therefore, these Betaproteobacteria are also candidate denitrifiers in MDD ecosystems. However, we cannot rule out that these 13C-labeled nonmethanotrophs may have assimilated intermediates aerobically, because denitrifiers are usually facultative anaerobes. For example, isolates of the Methylophilaceae are known to be unable to perform denitrification, i.e. methylotrophic aerobes (Jenkins et al., 1984; Urakami & Komagata, 1986), although the uncultured Methylophilaceae have been identified as important methylotrophic denitrifiers in sewage sludge in other recent SIP studies (Ginige et al., 2004; Osaka et al., 2006).

It has long been believed that denitrifying activity and enzyme synthesis are completely suppressed by oxygen because denitrifiers normally gain energy by oxygen-dependent respiration under aerobic conditions and conduct denitrifying metabolism only when in oxygen-depleted environments (Knowles, 1982; Zumft, 1997). Thus, it is conceivable that denitrification was suppressed under our experimental conditions, i.e. the initial oxygen concentrations (18% v/v) of 90% air saturation in headspaces. This amount of oxygen in headspaces was allowed to consume about 0.53 mmol of CH4, i.e. 12% v/v CH4 (data not shown). Although the oxygen in the vials was mostly consumed to oxidize methane (10% v/v) by the methanotrophs, microorganisms are required to perform denitrification in the presence of oxygen. Nevertheless, previous studies on MDD (Werner & Kayser, 1991; Amaral et al., 1995; Thalasso et al., 1997; Costa et al., 2000; Waki et al., 2002; Knowles, 2005) have shown evidence of methane-dependent denitrification in the presence of oxygen. We will now discuss possible explanations for denitrification in the presence of oxygen. First, the spatial arrangement of methanotrophs and denitrifiers may be due to a dissolved oxygen gradient in sludge flocs (Rittmann & Langeland, 1985; Schramm et al., 1999), which has also been reported in sludge in which there is co-occurring methanotrophy and denitrification (Waki et al., 2002). Another possibility is that aerobic denitrification was conducted by certain species of bacteria capable of corespiring oxygen and nitrogen oxides and producing N2 (Meiberg et al., 1980; Robertson & Kuenen, 1984; Patureau et al., 1994). In contrast to common denitrifiers, aerobic denitrifiers appear to have an ecological advantage in niches with frequent aerobic/anoxic shifts (Frette et al., 1997). Additionally, it has been suggested that some aerobic denitrifiers prefer one-carbon sources such as methanol and formate (Zhao et al., 1999; Takaya et al., 2003). In fact, aerobic denitrification is observed in some methylotrophs, such as H. denitrificans (Meiberg et al., 1980) and Paracoccus denitrificans (Robertson & Kuenen, 1984). Furthermore, Knowles (2005) showed that some methanotrophs produce gaseous nitrogen oxides (e.g. NO and N2O) during growth on nitrate when oxygen is depleted through methane oxidation. Therefore, MDD ecosystems might exhibit these phenomena.

Our results show that supplementation with nitrate had strong effects on methane consumption (Fig. 1a) and methanotroph community structure (Figs 3 and 4). The pmoA phylogeny is generally consistent with the 16S rRNA gene phylogeny of methanotrophs because it is said that there is no evidence of horizontal gene transfer of pmoA among methanotrophic bacteria (Murrell et al., 1998; Costell & Lidstrom, 1999). Thus, six clusters derived from pmoA clones obtained in this study were classified in three different groups (i.e. types I, II, and X) according to the phylogenetic indications of pmoA genes and 16S rRNA genes. The phylogenetic analyses showed that types I and X methanotrophs were stimulated by supplementation with nitrate, whereas type II methanotrophs were minor components of the population in the presence of nitrate (Run 1). On the other hand, a variety of methanotrophs, including type II methanotrophs, took up methane in the absence of nitrate (Run 4). Mohanty et al. (2006) demonstrated that nitrogenous fertilizers stimulated methane consumption and the growth of type I methanotrophs, whereas type II methanotrophs were generally inhibited. Other previous studies describe type I methanotrophs as tending to dominate in eutrophic lakes, whereas type II methanotrophs tend to dominate in oligotrophic lakes (Saralov et al., 1984; Hanson & Wattenberg, 1991). Growth of the type II methanotrophs was favored under nitrogen-limiting conditions due to the ability to fix nitrogen, although type X methanotrophs (e.g. Methylococcus capsulatus) are also able to fix nitrogen (Hanson & Hanson, 1996). Therefore, it was suggested that type I methanotrophs tend to dominate under nutrient-rich conditions that allow rapid growth, whereas type II methanotrophs tend to be abundant under nutrient-poor conditions that limit growth. It should be noted, however, that the types of methanotrophs in each habitat are not only determined by nutrient conditions, but also by various other factors, such as the concentrations of methane and oxygen (Amaral et al., 1995; Henckel et al., 2000) and the temperature (Sundh et al., 2005). However, there were discrepancies in the results for the 16S rRNA genes and the pmoA genes; i.e. there were high frequencies of putative pmoA sequences derived from type X methanotrophs. This might be indicative of small sample size of the clone, PCR bias or cloning bias (Horz et al., 2001; Bodrossy et al., 2003; Cebron et al., 2007). Methanotrophic populations should also be investigated using PCR-independent methods (e.g. FISH analysis and PLFA analysis).

In addition, we detected bacterial groups with unique functions (Myxococcales of the Deltaproteobacteria). This order of bacteria contains groups described as ‘micropredators’, which are able to degrade other organisms (e.g. bacteria, yeast) by means of a variety of hydrolytic exoenzymes, such as proteases, lipases, and cell wall lytic enzymes (Shimkets et al., 2005). In recent studies, the members of the Myxococcales have been found to play a significant role in the turnover of biomass carbon in soil (Reichenbach, 1999; Lueders et al., 2006). Therefore, it has been suggested that these bacteria incorporate 13C by degrading active or dead 13C-labeled cells (e.g. methanotrophs) in sludge exhibiting MDD activity, although it is not possible to determine whether there is the ability to lyse and digest other cells merely from 16S rRNA sequence similarity. These indirect labelings of microorganisms other than methanotrophs and methylotrophs could be due to heavy 13C labeling of sludge. The clone library R4.13 consisted of diverse taxonomic bacterial groups. In spite of similar levels of methane consumption, there were significant differences in the diversity levels between the clone libraries R1.13 and R4.13. This might not be due so much to the supplementation of nitrate as to differences in incubation time: Run 1, 134 h; Run 4, 415 h. Therefore, most bacteria may have been labeled due to the heavy labeling process and the long incubations.

In conclusion, our results show the occurrence of MDD in the presence of oxygen. Nitrate consumption was attributed not only to nitrogen assimilation for growth, but also to denitrification. However, it was also found that 60% of the initial DT-N accumulated as biomass in sludge, which was attributed to bacterial growth. This indicated that the assimilation of nitrate for the growth of active bacteria (methane-dependent assimilation: MDA) was a primary pathway of nitrate-depletion in this experiment. Thus, MDA may be thought of not as removing nitrogen from wastewater, but rather as moving nitrogen from the main treatment to the sludge treatment. The application of DNA-SIP provides insights into key bacterial players involved in MDD, and supplementation with nitrate was observed to have strong effects on methane consumption and community structure of methanotrophs. Advances in understanding the metabolomics of methanotrophic bacteria under various conditions and further research into the determinants of methanotroph community structure will significantly improve the characterization of trophic links in MDD ecosystems.

Ancillary