INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Wetland rice fields are an important source of the greenhouse gas CH₄. They contribute up to 20% (20 to 50 Tg year¹) of the global CH₄ emissions (28, 34). It is thought that due to the demands of the growing human population, cultivation of wetland rice will increase during the next 3 decades by 60% (14). Therefore, a reduction in CH₄ emission from this source due to microbial CH₄ oxidation could play an increasingly important role in the global CH₄ budget.

Methane-oxidizing bacteria (methanotrophs) oxidize CH₄ with molecular O₂ and use it as a carbon and energy source (11, 23). Methane is produced as an end product of anaerobic degradation of organic matter, whereas O₂ in flooded rice fields occurs only at the soil-floodwater interface and, due to O₂ leakage from the rice roots, in a thin soil layer around the roots (3). Methanotrophs have been detected in rice field soil and on rice roots (7, 9, 19, 43), but until now not much has been known about the population structure and dynamics of these microorganisms over the growth period of rice.

The Methanotrophs have been divided into two families: the Methylococcaceae, belonging to the  subclass of the class Proteobacteria, and the Methylocystaceae, belonging to the  subclass of the Proteobacteria. The family Methylococcaceae includes the genera Methylobacter, Methylomonas, Methylomicrobium, Methylococcus, Methylocaldum, and Methylosphaera, and the family Methylocystaceae includes the genera Methylosinus and Methylocystis (11). Physiological differences between the two families are correlated with different ecological preferences. Graham et al. (21), for example, have shown that Methylosinus trichosporium outcompetes Methylomonas albus in chemostat cultures under N-limiting conditions. Amaral and Knowles (2) found that in CH₄-O₂ countergradients members of the Methylococcaceae grow preferentially at low CH₄ concentrations and members of the Methylocystaceae grow preferentially at high CH₄ concentrations. In the complex rice field ecosystem temporal and spatial changes in the microbial environment occur and might result in shifts in the methanotrophic community, which could influence overall CH₄ oxidation activity. A better understanding of the methanotrophic community structure in rice fields is therefore important to understand this dynamic ecosystem. It could also facilitate future attempts to activate or affect CH₄ oxidation in this or other systems with the aim of reducing CH₄ emission. A combination of process measurements and molecular community analysis was used in this study to compare community structure and actual activity of methanotrophs.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Compartmented microcosms. The microcosm system used in this study was kindly provided by P. Bodelier (Netherlands Institute of Ecology, Centre for Limnology, Maarssen, The Netherlands) and has been described previously (6). In short, microcosms were built of stainless steel cylinders (height, 12 cm; diameter, 9 cm). In the center of each microcosm, the root compartment was separated by a perforated stainless steel cylinder (diameter, 4 cm; hole diameter, 1 mm) that was covered with two layers of nylon gauze (mesh size, 0.45 µm). In each compartment, one porewater sampler was installed vertically from below. Each porewater sampler consisted of a hydrophilic porous membrane (pore diameter, 0.1 µm) fitted to a steel needle and connected with plastic tubing for sampling. Porewater samples were taken weekly by using evacuated vacuum tubes. The sampling procedure and porewater analysis method used have been described previously (30).

Soil. Rice field soil was taken after plowing from an Italian rice field in spring 1998 (Istituto Sperimentale per la Cerealicoltura, Vercelli, Italy). This site has been used repeatedly to study the microbial ecology of rice fields (30). The soil is a sandy loam (Cambisol), and it was dried at room temperature, ground with a jaw crusher, and sieved (hole size, 2 mm) before it was used in the microcosms. Each bulk soil compartment was filled with 495 g (dry weight) of soil, and each root compartment was filled with 105 g (dry weight). The microcosms were flooded with demineralized water and incubated at 25°C for 10 days before rice seedlings were transplanted.

Rice plants and growth conditions. Rice seeds (Oryza sativa type japonica variety KORAL) were germinated at 25°C for 12 days before transplanting. Planted microcosms were incubated in a growth chamber with 12-h days and 12-h nights and a light intensity of 0.22 to 0.55 mmol m² s¹. The temperature was 25°C during the light phase and 20°C in the dark. The humidity was adjusted to 70%.

Fertilization. Each of the microcosms was fertilized weekly with 2 ml of a fertilizer solution containing (per liter) 5.5 g of urea, 4.2 g of KH₂PO₄, and 3.4 g of KCl. The overall amount of fertilizer added during a 13-week growth period was the amount of fertilizer added to rice fields during rice growth, which was (per hectare) 160 kg of N, 140 kg of P₂O₅, and 155 kg of K₂O₂ in 1998 (30).

Flux measurements. For CH₄ flux measurement, the microcosms were covered with glass cylinders (volume, 2 liters). In the top of each glass cylinder a fan was installed to ensure complete mixture of the headspace. The increase in the CH₄ concentration was monitored for 2 h by taking 1-ml gas samples (n = 18) through a butyl septum. The samples were analyzed with a gas chromatograph equipped with a flame ionization detector. The in situ CH₄ oxidation activity was determined by comparing flux measurements before and after addition of difluoromethane (CH₂F₂), a specific inhibitor of methane oxidation (33). The inhibitor was injected into the gas phase (final concentration, 1% [vol/vol]) immediately after the microcosms were covered with the glass cylinders. Methane concentrations were monitored for 2 h as described above. After this the glass covers were removed, and the microcosms were allowed to equilibrate with the amibient air for approximately 1 h outside the growth chamber to remove the CH₂F₂. The in situ oxidation value was calculated from the difference between CH₄ emission with the inhibitor and CH₄ emission without the inhibitor. Methane flux was measured weekly in four parallel microcosms. In situ CH₄ oxidation was determined every second week in an additional four microcosms of the same plant age in order to allow direct comparison of fluxes with and without inhibitor. The effect of the inhibitor was detectable within 20 to 40 min after it was added by the increase in CH₄ emission compared to that of the untreated microcosms. The recovery of the system after CH₂F₂ treatment was studied by measuring the CH₄ fluxes without the inhibitor 1 day and 2 weeks after the inhibitor treatment. After 1 day the CH₄ emission from CH₂F₂-treated microcosms was about 70% of the CH₄ emission from the controls. After 14 days the fluxes of the controls and the inhibitor-treated microcosms were equal, which allowed repeated inhibitor measurements with the same microcosms.

Division into compartments. The bulk soil was taken directly from the non-root-containing compartments of the microcosms and mixed 1:3 (by weight) with sterile tap water. The root-containing soil (rhizosphere) was strongly attached to the roots and therefore was removed together with the plants. To separate the soil, the roots were washed twice in 125 ml of sterile tap water. The resulting soil suspensions were subsequently pooled and are referred to below as the rhizosphere. The roots were washed again with tap water to remove the remaining soil particles. The rhizoplane was detached by adding sterile glass beads and tap water to the roots (ratio of roots to glass beads to water, 1:10:10 [wt/wt/wt]) and shaking the preparation at 180 rpm for at least 30 min. The resulting suspension was decanted and is referred to below as the rhizoplane. The remaining roots were thoroughly washed with sterile tap water and cut into small pieces. Root pieces and sterile tap water (1:10, wt/wt) were homogenized for 2 min (Standard Blender 260: catalog no. FD-216; Buddeberg, Mannheim, Germany). The resulting suspension is referred to below as the homogenate.

Potential methane oxidation. Four microcosms containing plants of each age (28, 57, and 92 days after transplanting [dap]) were used. Slurries of the soils in the two soil compartments of each microcosm were prepared as described above. To study the potential CH₄ oxidation rates (MOR) on the rice roots, whole washed roots were used. All samples were measured in three parallel incubations. Soil slurries (20 ml) and roots were transferred into sterile glass bottles (150 ml for slurries; 50 ml for roots), which were closed with stoppers and supplemented with 10,000 ppm (by volume) of CH₄ each. The bottles were incubated at 25°C in the dark and shaken at 120 rpm. Methane depletion was monitored by sampling the headspace after the bottles were shaken thoroughly and then performing an analysis with a gas chromtograph equipped with a flame ionization detector. The first sample was taken 30 min after CH₄ was added, and this was followed by sampling at 2-h intervals during the first 8 h of the experiments. Then the bottles were left overnight and sampled the next day at 2-h intervals again. From the mean CH₄ depletion curves for the three parallel samples from one microcosm, two linear regressions were calculated. The first regression described the initial CH₄ oxidation, which corresponded to the CH₄ depletion at the beginning of the measurements until the onset of fast (induced) CH₄ oxidation. The second regression described the fast (induced) CH₄ oxidation period. The initial and induced MOR were calculated from the slopes of these regression lines. The time lag until the onset of induced CH₄ oxidation was determined from the point of intersection of the two regression lines.

Cell counting by the MPN technique. All four compartments (bulk soil, rhizosphere, rhizoplane, and homogenate) of one (20 and 72 dap), two (85 dap), or four (28, 57, and 92 dap) parallel microcosms were used for most-probable-number (MPN) determinations. Each sample was diluted in twofold steps in 2 × 8 parallel dilution series (microtiter plates) by using ammonia mineral salt medium (modified as described by Whittenbury et al. [47]) containing 0.5 g of NH₄Cl per liter and 0.54 g of KH₂PO₄ per liter; the pH was adjusted to 6.8. Trace element solution SL10a (2 ml liter¹), MgSO₄ (final concentration, 0.2 g liter¹), and CaCl₂ (final concentration, 0.015 g liter¹) were added after autoclaving (19). Inoculated plates were incubated in gas-tight jars at 25°C for 4 weeks in the dark. One of the two parallel microtiter plates was incubated in an atmosphere containing 20% CH₄ in synthetic air (20.5% O₂ in N₂): the other plate was incubated in synthetic air as a control for heterotrophic growth. Cell numbers and standard deviations for the eight parallel dilution series from one sample were calculated from positive dilutions by using the code number system and table of Rowe et al. (38). To illustrate the importance of the compartments for the whole system, the number of cells per total mass of each compartment was calculated. On average, the total root biomasses per plant were 0.24, 0.98, 4.46, 3.06, 2.74, and 3.37 g (dry weight) for the plants 20, 28, 57, 71, 85, and 92 dap, respectively.

16S rRNA targeting oligonucleotide probes. Different oligonucleotide probes with high specificity for the two families of methanotrophs have been described recently (17). The following probes were used in this study (target sites refer to Escherichia coli numbering [12]): M450 (5'ATCCAGGTACCGTCATTATC3'), M84 (5'CCACTCGTCAGCGCCCGA3'), and M705 (5'CTGGTGTTCCTTCAGATC3'). For probe M450 no nonmethanotrophic pure culture with sequence identity to the probe was found in the EMBL database. The sequence of Methylocella (15), a new genus of acidotolerant -protebacterial methanotrophs, exhibited two mismatches with this probe and therefore was not detected under the hybridization conditions used. For probe M84 organisms like Thiorhodovibrio sibirica, Thiothrix eikelboomii, and Thiobacillus neapolitanus (a halophile), as well as Ectothiorodospira strains and phototrophic bacteria, exhibited no mismatches. However, parallel hybridization with probe M84 and probe M705, which showed no identity to cultures other than methanotrophic cultures, and adjustment of specificity with Thiobacillus thiooxidans allowed in situ quantification of both families of methanotrophs with these probes. The probes were synthesized and labeled with fluorochromes (fluorescein, Texas red, CY5) by MWG Biotech (Ebersberg, Germany).

Fixation and whole-cell hybridization. Samples of diluted soil slurries (2 ml of a 1:100 dilution), rhizoplane (1 to 2 ml), and homogenate (1 ml) were centrifuged at 13,000 × g for 8 min. The pellets were resuspended in 100 µl of phosphate-buffered saline (PBS) (pH 7.0) and were fixed with 300 µl of 4% paraformaldehyde (in PBS) for 2 h at room temperature. After they were washed five times with PBS, the pellets were resuspended in ethanol-PBS (1:1) and stored at 20°C until hybridization.

DNA extraction. Soil slurries (2 ml, undiluted) or 5-ml portions of either rhizoplane or homogenate suspensions were pelleted by centrifugation in 2-ml screw-cap tubes (13,000 × g, 5 min). The pellets were stored at 20°C until DNA was extracted. The DNA extraction protocol used was based on cell lysis with 10% sodium dodecyl sulfate in a cell disrupter, followed by DNA purification with ammonium acetate precipitation and isopropanol precipitation. The procedure was described in detail by Henckel et al. (24). To avoid interference from humic acids during PCR amplification, DNA derived from soil samples was purified further by using polyvinylpolypyrrolidine (Sigma-Aldrich Chemie GmbH, Steinheim, Germany), columns and a modified protocol described by Holben et al. (26). Polyvinylpolypyrrolidine was suspended overnight in 5 M HCl, and then washed with Tris-EDTA buffer (10 mM Tris, 1 mM EDTA; pH 8.0), and the pH was adjusted to 8.0 with NaOH. Spin columns (Micro Bio-Spin chromatography columns: BioRad, Munich, Germany) were filled with 2 ml of this suspension and packed and dried by centrifugation (375 × g, 1 min) before they were loaded with 150 µl of DNA extract. DNA concentrations were estimated by spectrometry at a wavelength of 260 nm (GeneQuant spectrophotometer; Pharmacia Biotech, Uppsala, Sweden) by using 1:10 dilutions in water. For PCR amplification, all DNA concentrations were adjusted to 2 ng of DNA µl¹.

PCR amplification. The DNA was amplified by using three primer sets targeting 16S rRNA genes: a universal eubacterial primer set (primers 533f and 907r [45]), the 10 primer set (primers 197f and 533r [42]) targeting methylotrophs that use the ribulose monophosphate (RuMP) pathway for carbon assimilation (including members of the Methylococcaceae), and the 9 primer set targeting methylotrophs that use the serine pathway for carbon assimilation (including members of the Methylocystaceae) (42). Additionally, one primer set for the -subunit of the methanol dehydrogenase gene, which is present in all gram-negative (proteobacterial) methylotrophs, mxaF (primers 1003f and 15621r [32]), was used. For all primer pairs a GC clamp was attached to the 5' end of one primer for subsequent denaturing gradient gel electrophoresis (DGGE) analysis. The PCR protocols and primers used have been described in detail previously (24).

DGGE. DGGE analysis was carried out as described by Henckel et al. (24) at 60°C and 150 V for 5 h (Dcode system; BioRad). Denaturing gradients from 40 to 70% were used for the primer set 9 amplification products, whereas for the amplification products of all other primer sets gradients from 35 to 70% were used (80% corresponded to 6.5% acrylamide, 5.6 M urea, and 32% deionized formamide). Gels were stained after electrophoresis in 1:30,000-diluted SYBR green I (Biozym, Hess Oldendorf, Germany) for 30 min and were scanned with a Storm 860 phosphor imager (Molecular Dynamics).

Reamplification and sequencing of DGGE bands. DGGE bands were illuminated with a Dark Reader transilluminator (Clare Chemical Research, Ross on Wye, United Kingdom), excised, suspended in 200 µl of PCR grade water, and stored at 20°C. Band purity was controlled by reamplification and subsequent DGGE. Only reamplification products which resulted in a single band with the predicted electrophoretic mobility were sequenced. For sequencing, PCR products were purified by using QIAQuick PCR purification columns (QIAGEN GmbH, Hilden, Germany). Sequencing reactions were performed by using an ABI-Dye terminator cycle sequencing kit (Perkin Elmer Applied Biosystems, Weiterstadt, Germany) as specified by the manufacturer. Cycle sequencing products were purified by using Microspin G-50 columns (Pharmacia Biotech, Freiburg, Germany) and were analyzed with an ABI 373 DNA sequencer (Perkin Elmer Applied Biosystems).

Nucleotide sequence accession numbers. The accession numbers of the nucleotide sequences determined in this study are as follows: sequences obtained from 10 amplifications, AJ300106 to AJ300120; sequences obtained from 9 amplifications, AJ300121 to AJ300127; and methanol dehydrogenase gene sequences, AJ300128, AJ300129, AJ300159, and AJ300160.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Methane oxidation activity and cell numbers. The level of in situ CH₄ oxidation in the microcosms decreased from 70% at the beginning of plant growth to negligible values around 60 dap (Fig. 1). In parallel, the level of CH₄ emission increased from 25 to 55 mg of CH₄ m² h¹ and then remained stable until the end of the season (Fig. 1). The in situ CH₄ oxidation activity was correlated with the plant growth phase, as the decline coincided with the onset of the reproductive growth phase.

View larger version (22K): [in this window] [in a new window]   FIG. 1.   In situ CH4 oxidation and emission from rice microcosms during the growth period. In situ oxidation was calculated from differences in the fluxes with and without the inhibitor CH2F2. The error bars indicate standard errors based on four parallel microcosms.

View larger version (23K): [in this window] [in a new window]   FIG. 2.   Methane concentrations in the porewater of rhizosphere and bulk soil of rice microcosms during the growth period of rice. The error bars indicate standard errors based on three to six parallel microcosms.

1

1

1

1

1

1

1

1

View larger version (42K): [in this window] [in a new window]   FIG. 3.   Initial and induced rates of potential CH4 oxidation (MO) in soil slurries and on roots from rice microcosms containing plants of different ages. Bulk, bulk soil. The error bars indicate standard errors for samples from four parallel microcosms. The standard error for the initial MOR in bulk soil at 57 dap was too small to be seen. gdw, grams (dry weight).

1

1

View larger version (24K): [in this window] [in a new window]   FIG. 4.   Ammonia concentrations in the porewater of rice microcosms during the growth period. The in situ CH4 oxidation from Fig. 1 is given for comparison. The error bars indicate standard errors based on four parallel microcosms.

View larger version (26K): [in this window] [in a new window]   FIG. 5.   MPNs of methanotrophs (MOB) in the different compartments of rice microcosms during the growth period. Cell numbers were calculated based on the total masses in the compartments. The error bars indicate standard errors based on two to four parallel microcosms. At 20 and 72 dap only one microcosm was counted.

Community analysis. 16S rDNA amplification with the universal eubacterial primer set resulted in PCR products for all of the samples investigated. The sequences of the DGGE bands obtained did not cluster with methanotroph sequences, indicating that this group was not dominant among the eubacteria. However, DGGE analysis based on amplifications with the 9 and 10 primer sets and fluorescent in situ hybridization (FISH) revealed the presence of members of both families of methanotrophs in all compartments during the whole season. The DGGE band patterns of the 9 amplification products indicated that there was a stable population of serine pathway methylotrophs (including members of the Methylocystaceae) during the whole growth period and in all compartments (Fig. 6A). The major bands in different DGGE lanes (Fig. 6A, bands 1, 2, 4, and 5) were excised and sequenced, and they clustered with the group of Methylosinus and Methylocystis spp.; the most closely related sequences were those found by Henckel et al. (24) (accession no. AF126928 and AF126914) in rice field soil and those found by Wise et al. (48) (accession no. AF177318 and AF177320) in landfill soil (Fig. 7). Minor bands with lower electrophoretic mobility (Fig. 6A, bands 6 and 7) clustered distantly with Methylobacterium sp. and were most closely related to sequences found in the rape rhizosphere (O. Kaiser, A. Puehler, and W. Selbitschka, unpublished data) (EMBL Data Library accession no. AJ295558) and in rice field soil (25) (accession no. AF283225) (Fig. 7).

View larger version (53K): [in this window] [in a new window]   FIG. 6.   DGGE band patterns retrieved after 16S rDNA amplification from soil and root compartments of rice microcosms. The bands that were excised and sequenced are indicated. B, bulk soil; Rh, rhizosphere; RP, rhizoplane; H, homogenate. (A) Band patterns of primer set 9 amplification products. Mss, amplification product obtained with DNA from a Methylosinus sporium culture. (B) Band patterns of primer set 10 amplification products. Mcc, amplification product obtained with DNA from a Methylococcus capsulatus culture; Ac, sequence that clustered with Acinetobacter; Az, sequence that clustered with Azoarcus; Mx, sequence that clustered with Myxococcales; Th, sequence that clustered with Thiothrix.

View larger version (35K): [in this window] [in a new window]   FIG. 7.   Phylogenetic placement of sequences retrieved from DGGE gels after PCR amplification of rice microcosm samples with the 9 and 10 primer sets for 16S rDNA. The tree was constructed by the neighbor-joining method, using the maximum-parsimony algorithm and the Jukes-Cantor correction. Sequences are indicated as follows: running number (DGGE band number in Fig. 6), followed by the compartment (B, bulk soil; Rh, rhizosphere; RP, rhizoplane; H, homogenate), the age of the microcosm (in dap), and the primer set used for DNA amplification (primer set 9 or 10). proteobact., proteobacterium; bact., bacterium.

View larger version (79K): [in this window] [in a new window]   FIG. 8.   (A) DGGE band patterns of mxaF amplification products from soil and root samples of rice microcosms. The bands that were excised and sequenced are indicated. Mss, amplification product obtained with DNA from a Methylosinus sporium culture; B, bulk soil; Rh, rhizosphere; RP, rhizoplane; H, homogenate. (B) Unrooted phylogenetic tree for mxaF sequences from rice microcosm samples compared to previously published sequences. Sequences are indicated as follows: running number (DGGE band number in Fig. 8A), followed by the compartment (Rh, rhizosphere; RP, rhizoplane), and the age of the microcosm (in dap). uncul. putative methanotr., uncultivated putative methanotroph.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Limitation of in situ CH₄ oxidation activity. The short period of high levels of in situ CH₄ oxidation at the beginning of the growth period is in contrast to the results of previous studies, which showed that in situ CH₄ oxidation occurred during the whole growth period of rice (10, 19). However, the latter in situ oxidation was determined by using other methods (e.g., CH₃F inhibition or N₂ headspace), another rice variety, and different growth conditions for the plants. All these factors might affect the percentage of in situ CH₄ oxidation measured. The time pattern for in situ CH₄ oxidation and CH₄ emission detected in our study was confirmed by obtaining measurements in a rice field (30) using the inhibitor CH₂F₂ and the same rice variety that was used in the microcosms (KORAL).

Community structure of methanotrophs. The DGGE band patterns of the 9 primer set PCR amplification products showed that the community of serine pathway methylotrophs (including members of the Methylocystaceae) remained stable during the whole growth period and in all compartments of the system. Most sequences derived from the DGGE gel clustered with the known genera of the Methylocystaceae, Methylocystis and Methylosinus, and were most closely related to sequences detected previously in rice field soil after cloning of PCR amplification products (24). No differences in band intensities or numbers were detected after the decrease in in situ CH₄ oxidation; thus, DGGE band patterns showed no direct correlation with in situ activities. Henckel et al. (24), in contrast, found that there were increases in the number and intensities of bands after the onset of CH₄ oxidation in rice field soil with a nonsaturated water content in their experiment. In our system, the soil was flooded for 10 days before the first samples were taken. Therefore, the effects of increasing CH₄ concentrations in the soil porewater and the onset of CH₄ oxidation in the unplanted system on the methanotrophic community could not be monitored.

In situ dominance of Methylocystaceae. This is the first study to show the in situ dominance of either family of methanotrophs with family-specific probes (17) and FISH. Despite the low percentage of methanotrophic cells compared to the overall DAPI counts, there was evidence of in situ dominance of members of the Methylocystaceae during the whole season and in all compartments. Members of the Methylococcaceae occurred at higher relative proportions only in the rhizoplane, illustrating the importance of the rice roots for the methanotrophic diversity of the rice field ecosystem. In previous studies strains belonging to the Methylocystaceae have been isolated from the highest positive MPN dilutions of rice root and rice field soil incubation mixtures (7, 43), indicating the numerical dominance of this family. Our FISH results indicate that these data were not influenced by cultivation biases but reflected the numerical ratio of members of the methanotrophic families in situ.

Conclusions. Both families of methanotrophs occur in the rice microcosm system, but drying of the soil selected for genera with a desiccation-resistant resting stage. The main population growth occurred during the first 2 months of the growth period, accounting for the losses during the remainder of the year. The rhizoplane supported the greatest diversity of methanotrophs. FISH analysis provided evidence that in situ dominance by members of the family Methylocystaceae occurred during the whole season and in all compartments. The methanotrophic activity in the rice microcosms seemed to be influenced mainly by the rice roots and the availability of N sources.