INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Methane oxidation by methanotrophic bacteria occurs in soil and aquatic environments; this process reduces CH₄ emission and may be a negative feedback mechanism for increases in atmospheric CH₄ levels (17, 50). The atmospheric CH₄ budget is a concern since CH₄ is one of the important greenhouse gases and affects the Earth's climate (15, 49). Methanotrophs are important regulators of CH₄ emission from rice fields. Only part of the CH₄ produced in rice field soil is released into the atmosphere; the remainder is oxidized by methanotrophic bacteria living in oxic niches in flooded fields (i.e., the surface soil layer and the rhizosphere) (7, 18, 25, 33, 55). Slurries of anoxic rice field soil change from production of CH₄ to consumption of CH₄ when they are aerated (29). In nonsaturated rice field soil CH₄ oxidation is induced when the soil is moistened and exposed to CH₄ concentrations higher than 1,000 parts per million by volume (ppmv) (3, 4). The most-probable-number counts of methanotrophic bacteria also increase under these conditions. Type II methanotrophs (organisms belonging to the  subdivision of the class Proteobacteria [-Proteobacteria]) have been isolated from rice roots (27). However, the structure of the methanotrophic community and the possible changes in this community that occur during induction of CH₄ oxidation in rice field soil are not known.

In general, little is known about the structure of the methanotrophic community in soil, but it seems that type II methanotrophs are found more frequently than type I methanotrophs are found (28). An immunofluorescent analysis of tundra soils revealed the presence of both type I and type II methanotrophs (62). A phospholipid analysis of soil microorganisms demonstrated that type II methanotrophs were dominant in boreal peatland soil (58). A hybridization analysis of 16S rRNA extracted from Alaskan soil also indicated that type II methanotrophs were dominant (11). Type II methanotrophs were also found to be the dominant methanotrophs in peat bogs (20), whereas type I methanotrophs seem to prevail in aquatic environments, such as lake water (28, 52) and lake sediments (6).

The introduction of denaturing gradient gel electrophoresis (DGGE) to microbial ecology provided a valuable molecular fingerprinting technique for studying microbial community structure (31, 44-46). DGGE facilitates separation of mixtures of PCR-amplified gene fragments based on sequence differences (47) and allows large numbers of samples to be analyzed simultaneously. Thus, this technique is ideally suited for monitoring the dynamics of microbial communities influenced by environmental changes.

Recently, DGGE was used to analyze oxic agricultural soils incubated with high CH₄ mixing ratios, and the results revealed that type I methanotrophs were present in soil extracts but type II methanotrophs were present in enrichment cultures prepared from the same soils (35, 48). In these studies 16S ribosomal DNA (rDNA) primer sets were used to detect a wide range of species belonging to the domain Bacteria. However, in another study methanotrophs could not be resolved by DGGE when universal 16S rDNA primers were used (61).

Methanotrophs are a phylogenetically heterogeneous group belonging to the - or -Proteobacteria (28). Seven genera of type I (-Proteobacteria) and type II (-Proteobacteria) methanotrophs have been proposed (8, 9). While the genera Methylococcus (also classified as type X [28]), Methylomonas, Methylobacter, Methylomicrobium, and Methylosphaera belong to the type I methanotroph group, the genera Methylocystis and Methylosinus make up the type II methanotroph group. Scientists have developed 16S rDNA probes that target methanotrophic bacteria belonging to either type I or type II (11, 40, 60). Hybridization of the total extracted environmental DNA in a blanket peat bog by using genus-specific probes suggested that members of the genera Methylosinus and Methylococcus were the dominant methanotrophs in this environment, while the genera Methylomonas and Methylobacter were not detected (20).

Methanotrophs have similar physiological characteristics. The key enzymes particulate methane monooxygenase (pMMO), soluble methane monooxygenase (sMMO), and methanol dehydrogenase (MDH) are highly conserved, so that enzyme-based gene markers may offer the possibility of detecting all known methanotrophs (28). Gene probes that target functional genes have been developed for the pmoA gene (41) coding for the -subunit of the pMMO present in all known methanotrophs and the mxaF gene (40, 38) coding for the -subunit of the MDH present in all methylotrophs (28, 39). The pMMO is closely related to the ammonium monooxygenase (AMO) of the ammonium oxidizers, and the degenerate pmoA primer set (primers A189f and A682r) also detects the homologous sequence of the amoA gene for the -subunit of AMO (32). Scientists have also developed gene probes that target the mmoB gene coding for the sMMO (39) that is present in most type II methanotrophs and in members of the genus Methylococcus (type X) but not in most type I methanotrophs (28). A community analysis of environmental DNA from a blanket peat bog in which cloning and sequencing of these gene fragments were performed revealed a distinct phylogenetic cluster of type II methanotrophs that are probably new, uncultured acidophilic methanotrophs (38-41).

We studied changes in microbial community structure during induction of CH₄ oxidation in rice field soil by performing a DGGE analysis of small-subunit (SSU) rRNA-based and functional gene markers. A universal SSU rRNA-based primer set and two primer sets that target methylotrophic members of the - and -Proteobacteria were used. In addition, we employed primer sets that target functional genes, such as the pmoA, mmoB, and mxaF genes (39-41). This approach facilitated detection of changes in microbial community structure during induction of CH₄ oxidation in a nonsaturated rice field soil.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Soil. The rice field soil used in this study has been described previously in detail (29). This soil had a maximum water-holding capacity (WHC) corresponding to a gravimetric water content of 44% (wt/wt). In the experiments described below, the rice field soil was moistened with demineralized water until the gravimetric water content was 19% (wt/wt), which corresponded to 43% of the WHC. The moist soil was sieved with a 2-mm sieve in order to ensure homogeneity and to minimize the number of anoxic microsites.

CH₄ oxidation. Soil (4.1 ± 0.1 g [fresh weight]) was placed in 120-ml serum bottles, and the bottles were closed with butyl rubber stoppers. Fourteen bottles were flushed with moistened synthetic air (20.5% O₂ in N₂), and subsequently 50,000 ppmv of CH₄ was added. CH₄ was not added to a second set of eight bottles. All of the bottles were incubated at 25°C in the dark. Three bottles were used only for gas analysis, while the remaining bottles were used for DNA extraction. Four bottles were replenished three times with 5 ppmv of CH₄ after the soil had consumed enough of the CH₄ so that the mixing ratio was <1.8 ppmv. Bottles to which CH₄ was not added were used as controls for DNA extraction to provide community data for rice soil not supplemented with CH₄. Rice soil samples used for DNA extraction (one bottle per sample) were removed at different times (see Fig. 1) and stored at 20°C.

Gas analysis. Methane and O₂ contents were periodically measured by gas chromatography as described previously (29). Oxygen was repeatedly added to the bottles to keep the O₂ concentration constant at about 17 to 20% (vol/vol).

Bacterial strains. For molecular analyses nine methanotrophic reference strains obtained from the collection at our institute were used as previously described by Gilbert and Frenzel (25). The bacteria were cultured in nitrate mineral salts medium (pH 6.8) with 15% CH₄ in the headspace (25). The cultures were harvested after 2 to 3 days by centrifugation, and the cell pellets were stored at 20°C until DNA was extracted.

DNA extraction. The method used to extract DNA from rice field soil and from pure cultures of methanotrophs was a modification of the method of Moré et al. (43). Cell pellets or 0.6-g (fresh weight) portions of soil were placed in 2-ml screw-cap tubes. Approximately 1 g of sterilized (170°C, 4 h) zirconia-silica beads (diameter, 0.1 mm; Biospec Products Inc., Bartlesville, Okla.), 800 µl of sodium phosphate buffer (120 mM; pH 8), and 260 µl of a sodium dodecyl sulfate solution (10% sodium dodecyl sulfate, 0.5 M Tris-HCl [pH 8.0], 0.1 M NaCl) were added to the soil, which was resuspended by vortexing. The cells were lysed by shaking the preparations with a cell disruptor (model FP120 FastPrep; Savant Instruments Inc., Farmingdale, N.Y.) for 45 s at a setting of 6.5 m s¹. After centrifugation (3 min, 12,000 × g) the supernatant was collected, and the soil-bead mixture was extracted a second time by resuspension in 700 µl of phosphate buffer. Proteins and debris were precipitated from the supernatant by adding 0.4 volume of 7.5 M ammonium acetate and then incubating the preparation on ice for 5 min. After centrifugation at 12,000 × g for 3 min, the nucleic acids were precipitated by adding 0.7 volume of isopropanol and then centrifugating the preparation at 12,000 × g and 4°C for 45 min. Subsequently, the DNA pellet was washed with 70% ethanol at 4°C and dried under a vacuum. Finally, the DNA was resuspended in 200 µl of Tris-EDTA buffer (10 mM Tris base, 1 mM EDTA; pH 8). Soil DNA was purified further by using a Prep-A-Gene kit (Bio-Rad, Munich, Germany) as specified by the manufacturer.

PCR amplification. For PCR amplification we used the following three SSU rRNA-based primer sets: a universal primer set that targeted all life, which was modified from the primers described by Weisburg et al. (63) by using primers 533f (GTGCCAGCAGCCGCGGTAA) and 907r (AATTCCTTTGAGTTT) (Escherichia coli positions 515 to 533 and 907 to 922 [10]) as forward and reverse primers, respectively; and the MB10 and MB9 primer sets. The latter two primer sets were formulated by utilizing hybridization probes 10 and 9, which target methylotrophic - and -Proteobacteria (60), respectively, as forward primers and primer 533 as the reverse primer. In addition, three functional primer sets, the pmoA, mmoB, and mxaF primer sets, originally designated A189f/A682r, 77f/369r, and f1003/r1561, respectively, were utilized (38, 39, 41).

Cloning. 16S rDNA PCR products of soil DNA amplified with the MB9 primer set were cloned by using a pGEM-TEasy cloning kit (Promega, Madison, Wis.). Randomly selected clones were sequenced as described previously (53). Cloned inserts were amplified with primers targeting vector sequences. For DGGE analysis of MB9 clones, the PCR products were amplified with the MB9 primer set.

DGGE. DGGE was carried out as described previously, with minor modifications (45). PCR products were separated by using a DCode System (Bio-Rad) and 1-mm-thick polyacrylamide gels (6.5% [wt/vol] acrylamide-bisacrylamide [37.5:1]; Bio-Rad) prepared with and electrophoresed in 0.5× TAE (pH 7.4) (0.04 M Tris base, 0.02 M sodium acetate, 1 mM EDTA) at 60°C and constant voltage. A denaturing gradient consisting of 80% (vol/vol) denaturant corresponded to 6.5% acrylamide, 5.6 M urea, and 32% deionized formamide. Gels were poured on GelBond PAG film (FMC Bioproducts, Rockland, Maine) to avoid gel distortion. The DGGE conditions for the various PCR products were optimized by the perpendicular DGGE method (44). The conditions used for electrophoresis are summarized in Table 1. The gels were stained with 1:50,000 (vol/vol) SYBR Green I (Biozym, Hessisch-Oldendorf, Germany) for 30 min and scanned with a model Storm 860 PhosphorImager (Molecular Dynamics, Sunnyvale, Calif.). Some gels were silver stained (12), dried, and recorded with an overhead scanner (model Scanjet 4c/T; Hewlett-Packard, Palo Alto, Calif.).

Extraction of PCR products from DGGE gels. Due to its spectral characteristics, SYBR Green I bound to double-stranded DNA is maximally excited at 497 nm, and fluorescence emission is centered around a wavelength of 520 nm (Molecular Probes, Eugene, Oreg.). Thus, double-stranded DNA in gels can be detected with non-UV light sources, which is a prerequisite for avoiding UV light-induced DNA damage. We visualized DGGE bands in SYBR Green I-stained gels with blue light (, >400 nm) by using a Dark Reader transilluminator (Clare Chemical Research, Ross on Wye, United Kingdom). Samples of individual DGGE bands were obtained by excising a small core with a sterile 200-µl pipette tip; these samples were reamplified and reanalyzed by DGGE to verify that the bands were pure. Bands with the same mobility were excised from different lanes to check for sequence identity.

Sequencing of DGGE bands. Reamplified PCR products of excised DGGE bands were purified by using an EasyPure DNA purification kit (Biozym). The concentrations and purities of PCR products were determined by measuring the absorption at 260 and 280 nm of 1:20 dilutions in H₂O with a GeneQuant spectrophotometer (Pharmacia Biotech, Uppsala, Sweden). Sequencing reactions were performed with an ABI dye terminator cycle sequencing kit (Perkin-Elmer Applied Biosystems) by using 30 to 180 ng of template DNA, as specified by the manufacturer. Cycle sequencing products were separated from excess dye terminators and primers by using Microspin G-50 columns (Pharmacia Biotech, Freiburg, Germany) and were analyzed with a model ABI 373 DNA sequencer (Perkin-Elmer Applied Biosystems).

Nucleotide sequence accession numbers. Sequences of partial pmoA and mxaF gene fragments and of 16S rRNA gene fragments of excised DGGE bands have been deposited in the GenBank database under accession no. AF126295 to AF126297 and AF126908 to AF126945.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Induction of CH₄ oxidation in rice field soil. When moist rice field soil was incubated in the presence of 50,000 ppmv of CH₄, oxidation started after a lag phase of about 24 h. The methane concentration decreased linearly for approximately 48 h, and this was followed by a second phase of faster CH₄ consumption (Fig. 1). The CH₄ supplied was consumed until the residual concentration was 1 ppmv (i.e., less than the atmospheric mixing ratio, about 1.8 ppmv) (Fig. 1).

View larger version (22K): [in this window] [in a new window]   FIG. 1.   Oxidation of CH4 at a mixing ratio of 50,000 ppmv in bottles containing rice field soil at 43% of WHC. CH4 (5 to 6 ppmv) was replenished on days 14, 18, and 24 after the initial CH4 had been consumed so that the CH4 level was below the atmospheric CH4 level. The soil samples used for molecular analyses were removed at the times indicated. The data are means ± standard deviations (n = 3). d, days.

PCR amplification. Soil samples were taken at the times indicated in Fig. 1 from CH₄-supplemented soil, and DNA was extracted and used for PCR amplification. Soil samples were also periodically taken from controls to which no CH₄ was added. PCR products of the expected sizes were obtained after amplification with the functional pmoA and mxaF primer sets and the 16S rDNA MB9, MB10, and universal primer sets when the template DNAs were isolated from nine different methanotrophic reference strains and from soil samples. PCR products of the expected sizes were also obtained with the mmoB primer set that targets the mmoB gene coding for the -subunit of sMMO for the appropriate reference strains. However, only a very weak PCR product was obtained from soil samples that oxidized CH₄. PCR amplification with the mmoB primer set containing a GC clamp failed, and thus, DGGE analysis of mmoB PCR products was not possible.

View larger version (28K): [in this window] [in a new window]   FIG. 2.   PCR yields of amplification reactions performed with different primer sets and template DNA isolated from soil samples without CH4 oxidation (i.e., dry soil, control soil, lag-phase soil collected on days 1 and 2) and with CH4 oxidation (days 4 to 23). The data are means ± standard deviations (n = 6 or 7).

SSU rDNA DGGE and sequence analysis of members of the domain Bacteria. The DGGE analysis of PCR products amplified with the universal primer set revealed complex band patterns, which were similar for control soil without CH₄ and for soil supplemented with CH₄ (Fig. 3). The DGGE profiles obtained for dry rice field soil and control soil for the first 8 days of incubation contained fewer DGGE bands than the profiles obtained at later sampling times. The numbers and intensities of the bands increased after the soil was moistened. CH₄-supplemented soil obtained at zero time also produced fewer DGGE bands than soil obtained later. The sequences of the major DGGE bands (Fig. 3, bands I to V) grouped either with the gram-positive branch close to Bacillus species or with the Cytophaga-Flavobacterium-Bacteroides group (Fig. 3). However, sequences of methanotrophic or methylotrophic bacteria were not detected. Slight temporal changes in band patterns (Fig. 3) probably reflected CH₄-independent activation of bacteria by increases in the soil water content. The observation that even gram-positive, spore-forming Bacillus species known to be notoriously recalcitrant to lysis were detected by DGGE indicated that the lysis and DNA extraction protocols were effective (42, 43). The sequences of DGGE bands with the same mobility were identical when they were obtained at different times (Fig. 3). This was also the case when we used primers other than the universal primers (see below).

View larger version (49K): [in this window] [in a new window]   FIG. 3.   (A) DGGE band pattern obtained with the universal SSU rDNA primer set targeting all life. PCR product (20 µl; ~160 ± 41 ng) was loaded into each lane for all soil samples. The marked bands were sequenced. Mcs, Methylocystis; Ms, Methylosinus. (B) Phylogenetic tree constructed with partial 16S rDNA sequences, showing the relationship of the labeled DGGE bands to closely related members of the Bacteria. The sequences of the individual bands were obtained from the following lanes: band I, lanes 9d control, 4d, and 14d; band II, lanes 9d control and 14d; band III, lane 9d control; band IV, lane 4d; and band V, lane 9d control. The scale bar indicates the estimated number of base changes per nucleotide sequence position.

16S rDNA DGGE and sequence analysis of type I methanotrophs. The DGGE profiles of the PCR products amplified with the MB10 primer set contained three major bands, which demonstrated that type I methanotrophs were less diverse than the total bacterial community (Fig. 4). The PCR product yields obtained from control and lag-phase soil samples (samples obtained at zero time and 2 days after CH₄ was added) were consistently lower than the PCR product yields obtained from soil samples consuming CH₄ (Fig. 4), which confirmed the PCR product yields shown in Fig. 2. Although twice the volume of the PCR product was loaded onto DGGE gels, the band intensities obtained with control soil were still lower than the band intensities obtained with CH₄-oxidizing soil (Fig. 4, samples obtained after days 4 to 23). Three bands were retrieved from DGGE gels. The DNA sequences of these bands grouped closely with the DNA sequences of Methylobacter species within the radiation of the type I methanotrophs (-Proteobacteria) (Fig. 4).

View larger version (41K): [in this window] [in a new window]   FIG. 4.   (A) DGGE band pattern obtained with the MB10 16S rDNA primer set targeting type I methanotrophs. For soil without CH4 oxidation (lanes dry soil to 2d) 40 µl (~63 to 100 ng) of PCR product was loaded into each lane, and for soil with CH4 oxidation (lanes 4d to 23d) 25 µl (~397 ± 80 ng) of PCR product was loaded into each lane. Mc, Methylococcus; Mm, Methylomonas. (B) Phylogenetic tree based on 16S rDNA sequences showing the relationship of the labeled DGGE bands to the most closely related members of the -Proteobacteria. The sequences of the individual bands were obtained from lane 18d. The scale bar indicates the estimated number of base changes per nucleotide sequence position.

16S rDNA DGGE and sequence analysis of type II methanotrophs. The DGGE band patterns of PCR products obtained with the MB9 primer set contained about 12 major bands after incubation with CH₄. Control soil samples produced about seven major bands, all of which were also present in the soil oxidizing CH₄ (Fig. 5, samples obtained after 18 and 23 days). The intensities of six bands increased after the soil started to consume CH₄ (Fig. 5), like the results obtained for PCR products obtained with the MB10 primer set. Pure excised and reamplified MB9 DGGE bands could not be retrieved from DGGE gels, as indicated by the appearance of several DGGE bands when they were electrophoresed again on a DGGE gel. Therefore, the MB9 PCR products from control soil (after 13 days) and soil consuming CH₄ (after 14 days) were cloned to obtain sequence data for type II methanotrophs (n = 30). The electrophoretic mobilities of the PCR products of the MB9 clones were similar to those of the original soil DGGE bands. However, several DGGE bands of clones did not correspond to bands in the original DGGE band pattern, indicating that community analysis by cloning and community analysis by DGGE were subject to different biases (results not shown).

View larger version (39K): [in this window] [in a new window]   FIG. 5.   (A) DGGE band pattern obtained with the MB9 16S rDNA primer set targeting type II methanotrophs. For soil without CH4 oxidation (lanes dry soil to 2d) 40 µl (~360 ± 118 ng) of PCR product was loaded into each lane, and for soil with CH4 oxidation (lanes 4d to 23d) 40 µl (~572 ± 162 ng) of PCR product was loaded into each lane. Mcs, Methylocystis; Ms, Methylosinus. (B) Phylogenetic tree based on 16S rDNA sequences, showing the relationship of the MB9 clones to the most closely related members of the -Proteobacteria. The scale bar indicates the estimated number of base changes per nucleotide sequence position. c, clones from control soil (clones obtained after 13 days); ox, clones from soil with CH4 oxidation (clones obtained after 14 days).

DGGE and sequence analysis of the pMMO gene. Several methanotrophs contain at least two copies of the pMMO gene (56). This could explain the appearance of multiple strong DGGE bands for the pmoA PCR product of Methylococcus capsulatus, whereas Methylosinus trichosporium produced a very faint second band that was visible only on silver-stained DGGE gels (Fig. 6). It should be noted that the pmoA primer set is degenerate and also amplifies the amoA gene from ammonium oxidizers (32). The DGGE profiles of pmoA PCR products were markedly different for control soil and soil supplemented with CH₄ (Fig. 6). Control soil produced only one major band, whereas CH₄-oxidizing soil produced at least six additional bands, which appeared after 4 to 7 days of incubation and increased in intensity. At this time, the soil was oxidizing CH₄ at the maximal rate. Simultaneously, the intensities of the control soil bands decreased.

View larger version (43K): [in this window] [in a new window]   FIG. 6.   (A) DGGE band pattern obtained with the pmoA primer set targeting the gene of the -subunit of the pMMO. For soil without CH4 oxidation (lanes dry soil to 2d) 35 µl (~181 ± 59 ng) of product was loaded into each lane, and for soil with CH4 oxidation (lanes 4d to 23d) 35 µl (~707 ± 203 ng) of PCR product was loaded into each lane. Mc, Methylococcus; Ms, Methylosinus; Mcs, Methylocystis. (B) Phylogenetic tree based on derived amino acid sequences of pmoA fragments, showing the relationship of the labeled DGGE bands to pmoA sequences of other methanotrophs. The sequences of the individual bands were obtained from the following lanes: band I, lanes dry soil, 9d control, 13d control, 4d, and 7d; bands II and III, lanes 7d and 14d; and bands IV through VI, lane 7d. The scale bar indicates the estimated number of changes per amino acid sequence position. Bootstrap values are given for 1,000 replicate trees.

DGGE and sequence analysis of the MDH gene. The mxaF PCR products amplified from DNA extracted from control soil and soil supplemented with CH₄ were very weak and sometimes even undetectable at the beginning of the incubation period. Control soil produced only one DGGE band, whereas CH₄-supplemented soil produced three bands after the onset of CH₄ oxidation (i.e., 4 days after CH₄ was added) (Fig. 7). Interestingly, the control soil band apparently was absent in the soil consuming CH₄ (Fig. 7). The mxaF gene sequences of purified DGGE bands from control soil and soil consuming CH₄ grouped with the cluster containing the type II methanotrophs (Fig. 7).

View larger version (40K): [in this window] [in a new window]   FIG. 7.   (A) DGGE band pattern obtained with the mxaF primer set targeting the gene of the -subunit of the MDH. For soil without CH4 oxidation (lanes dry soil to 2d) 35 µl (~22 ± 5 ng) of PCR product was loaded into each lane, and for soil with CH4 oxidation (lanes 4d to 23d) 35 µl (~75 ± 56 ng) of PCR product was loaded into each lane. Mc, Methylococcus; Ms, Methylosinus; Mb, Methylobacter. (B) Phylogenetic tree based on derived amino acid sequences of mxaF fragments, showing the relationship of the labeled DGGE bands to mxaF sequences of other methylotrophs. The sequences of Hyphomicrobium spp. were obtained from reference 24. The sequences of the individual bands were obtained from the following lanes: band I, lane 23d; band II, lanes 13d control and 35d control; and band III, lanes 14d and 18d. The scale bar indicates the estimated number of changes per amino acid sequence position. The bootstrap values shown are the bootstrap values for 1,000 replicate trees.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Rice field soil moistened to 43% of WHC and initially supplemented with 5% CH₄ oxidized CH₄ after a lag phase. The soil then consumed CH₄ at concentrations below atmospheric levels but was not able to maintain this capacity for a prolonged period of time if it was only supplemented with low CH₄ mixing ratios (<6 ppmv). We assumed that the initial CH₄ oxidation activity (V_max) was high enough to allow consumption of CH₄ at atmospheric trace gas concentrations but that later CH₄ oxidation activity ceased, probably because oxidation of low CH₄ concentrations did not result in generation of the maintenance energy necessary for enzyme synthesis (16). Similar observations of decreasing CH₄ oxidation rates after induction of CH₄ oxidation at high CH₄ concentrations have been made with other soils (4, 51, 54).

The molecular analysis of the soil microbial community during incubation in the presence of CH₄ revealed the following: (i) PCR amplification with the specific 16S rDNA MB10 and MB9 primer sets, as well as with the functional pmoA and mxaF primer sets, resulted in significantly higher PCR product concentrations with soil consuming CH₄ than with control soil; (ii) the intensities of the DGGE bands of PCR products obtained with the MB10, MB9, pmoA, and mxaF primer sets increased when soil consumed CH₄; and (iii) DGGE analysis of PCR products obtained with the MB10, MB9, pmoA, and mxaF primer sets revealed differences in band patterns between control soil and CH₄-consuming soil.

We hypothesize that the increased PCR product concentrations observed after amplification with the specific MB10 and MB9 primer sets and with primers targeting pmoA and mxaF from soil consuming CH₄ indicate that there are more target sites in soil oxidizing CH₄ than in control soil due to stimulation by CH₄. It is unlikely that the observed differences in PCR product concentrations were due to different concentrations of PCR-inhibiting compounds, such as humic acids, in the template DNA extracted from the soil. Since all soil samples were treated in the same way for DNA extraction, the concentrations of potential PCR inhibitors should a priori have been the same in all soil samples. This presumption was supported by PCR amplification results obtained with the universal primer set. These amplification results showed that the yield of PCR products was the same regardless of the soil sample. However, it was surprising that the concentration of the mxaF PCR product was less than the concentration of the pmoA PCR product, since methanotrophs should contain both genes. We speculate that either there were more copies of the pmoA gene than of the mxaF gene per cell, the degenerate pmoA primer set also amplified the amoA gene from ammonium oxidizers present in the soil (see below), or PCR amplification operated more efficiently with the pmoA primer set than with the mxaF primer set.

Determination of microbial abundance based solely on PCR product concentrations is impossible (13, 21). However, Ferris and Ward (23) concluded that a change in the intensity of a particular DGGE band in a temporal or spatial environmental gradient may be used to infer that there has been a change in the size of a population. However, the intensities of different bands are not comparable to each other. Therefore, we concluded that the observed increases in the intensities of certain DGGE bands with longer exposure of soil to CH₄ indicated that there was an increase in the size of the methanotrophic population. Since the intensities of DGGE bands amplified with primers that target genes specific for methanotrophs were higher in the presence of CH₄ than in the absence of CH₄, our data suggest that the sizes of CH₄-oxidizing populations increased in rice field soil after it received CH₄ at high mixing ratios. A similar increase in the size of a methanotrophic population has also been demonstrated by most-probable-number counting by using incubation mixtures having initial CH₄ mixing ratios of >7,000 ppmv (4).

The DGGE band pattern for 16S rDNA templates amplified with the universal primer set was as complex as expected for a habitat such as soil. Changes in the DGGE band pattern occurred only at the beginning of the experiment when the soil water content changed from dry to moist. Addition of CH₄, however, had no effect on the DGGE band pattern. We assumed that the soil microorganisms responded to the difference in the soil water content and that this resulted in population shifts. The DNA sequences of the major DGGE bands grouped closely with the genera Flexibacter and Flavobacterium in the Cytophaga-Flavobacterium-Bacteroides group and with the genus Bacillus in the gram-positive bacteria with low G+C contents. Isolation of abundant fermentative rice field soil bacteria and a molecular survey of microbial diversity performed by cloning 16S rDNA genes showed that, indeed, bacteria belonging to these two groups represent a major portion of the bacterial community in rice field soil (14, 30).

The DGGE band patterns of PCR products obtained with the 16S rDNA MB10 and MB9 primer sets that target methylotrophic bacteria, as well as the functional primer sets that target pmoA and mxaF, contained fewer bands than the DGGE band patterns obtained with the universal SSU rDNA primer set. This observation is consistent with the assumption that the community of specialized CH₄ oxidizers was smaller than the community of microorganisms in general. DNA sequences of MB10 and pmoA primer set-generated DGGE bands grouped closely with DNA sequences of Methylobacter sp. and Methylococcus capsulatus within the type I methanotrophs. However, mxaF primer set-generated DGGE bands, some pmoA primer set-generated DGGE bands, and clone sequences obtained with the MB9 primer set grouped with the type II methanotrophs. The presence of type II methanotrophs (or type X methanotrophs, including Methylococcus capsulatus) in soil consuming CH₄ was also revealed by the PCR that targeted mmoB, but only soil samples that oxidized CH₄ at high rates produced an mmoB PCR product.

Differences between the DGGE band patterns obtained with soil incubated in the presence of CH₄ and soil incubated in the absence of CH₄ were observed when the pmoA and mxaF PCR products were examined. The major DGGE band detected and sequenced in control soil grouped close to Nitrosospira species. The pmoA primer set targeting pmoA also amplifies the amoA gene of ammonium oxidizers coding for the -subunit of the AMO (32). pmoA primer set-generated DGGE bands that were related to bands of methanotrophs, as determined by sequencing, intensified after the onset of CH₄ oxidation, while the pmoA primer set-generated DGGE band closely related to the Nitrosospira band became fainter. Apparently, ammonium-oxidizing populations were outnumbered or outcompeted by methanotrophs when CH₄ consumption started. Our observation is in agreement with recent results of Bodelier and Frenzel (5), who demonstrated that ammonium oxidizers did not contribute to CH₄ oxidation in rice microcosms.

16S rDNA sequence analysis of clones revealed that the MB9 primer set was not specific for type II methanotrophs and detected other -Proteobacteria as well. Therefore, changes in the DGGE band pattern observed with MB9 primer set-generated PCR products after the onset of CH₄ oxidation cannot be linked to methanotrophs exclusively. However, 16S rDNA sequences that clustered with sequences of methanotrophic bacteria were related to type II methanotrophs, which supported the data obtained in the mxaF and pmoA analyses. The number and diversity of sequence types obtained by cloning MB9 primer set-generated PCR products indicated that the diversity was greater than the diversity observed in the DGGE analysis of the same PCR products. DGGE analysis revealed only the most abundant populations, while cloning probably also detected less abundant populations.

Two MB9 primer set-generated clones grouped close to B. indica and novel, recently isolated, acidophilic methanotrophs (19). Detection of these closely related populations in neutral rice field soil suggests that this new group of methanotrophs may not be limited to acidic peat bogs. However, these novel methanotrophs must be isolated from rice field soil to support our molecular findings.

Our knowledge of the community structure of methanotrophs in rice field soil has been limited. Recently, two type II methanotrophs (strains Rp1 and Rp2) were isolated from high dilutions of most-probable-number count preparations obtained from the rhizoplane of rice roots (27). Generally, type II methanotrophs have been detected more frequently in soil environments than type I methanotrophs (see above). Several factors that influence competition between type I and type II methanotrophs, such as CH₄ and O₂ concentrations and nitrogen and copper availability, are currently being examined (28). Amaral and Knowles (1, 2) studied the distribution of soil and sediment methanotrophs in agarose diffusion columns with opposing gradients of CH₄ and O₂. These authors concluded that type I methanotrophs may be favored at low CH₄ concentrations and high O₂ concentrations, whereas type II methanotrophs may be favored at high CH₄ concentrations and low O₂ concentrations (2). The prevalence of type II methanotrophs in soils and the presence of type I methanotrophs in aquatic sediments support this hypothesis (28). With respect to O₂ and CH₄ availability, flooded rice field soil provides at least two different niches for methanotrophs: (i) the soil surface of flooded rice field soil (bulk soil) and (ii) the rhizosphere. The soil surface is comparable to aquatic sediments and is characterized by steep opposing O₂ and CH₄ gradients (26). At the interface of these gradients, the concentrations of O₂ and CH₄ are both very low, which could favor both types of methanotrophs. In the rhizosphere, O₂ and CH₄ concentrations can both be very low, but this region is characterized by spatial and temporal heterogeneity of O₂ and CH₄ concentrations due to the influence of the rice roots (26). The results of in situ probing with 16S rDNA-based probes suggested that type II methanotrophs were numerically dominant in the rhizosphere of aquatic macrophytes (37), and two numerically relevant type II methanotrophs were isolated from the rhizoplane of rice (27).

In this study, we used rice field soil originating from all of the habitats described above. Our results show that both type I and type II methanotrophs were present in the original rice field soil. When the soil was incubated under moist conditions with high mixing ratios of CH₄, we obtained DGGE bands of both type I and type II methanotrophs, which indicated that these bacteria were active. Apparently, factors other than O₂ and CH₄ availability may determine the composition of the methanotrophic community in rice field soil.

The active methanotrophs were apparently not able to consume CH₄ at atmospheric CH₄ mixing ratios, since CH₄ oxidation activity eventually ceased when CH₄ was supplied at low concentrations (<5 ppmv). Nevertheless, the populations of methanotrophs seemed to persist, as indicated by the results of the DGGE analyses. Recent experiments showed that some CH₄ production takes place even in drained, nonsaturated rice field soil, probably because there are anoxic niches in which active methanogenesis occurs (34). Therefore, it is possible that the methanotrophs were supplied with additional CH₄ that was produced inside soil aggregates.