INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

Until very recently, the list of recognized methanotrophic bacteria (MB) encompassed eight genera. These genera are divided into two physiologically distinct MB groups, type I and type II methanotrophs, which form phylogenetically coherent clusters in the - and -subclasses of the class Proteobacteria, respectively. The genera Methylomonas, Methylobacter, Methylococcus, Methylomicrobium, Methylocaldum, and Methylosphaera (6, 8, 9, 32) belong to the type I MB, while the genera Methylosinus and Methylocystis (32) represent the traditionally known group of type II MB. Last year a novel acidophilic methanotroph was described, Methylocella palustris (15). Strains of M. palustris were isolated from acidic Sphagnum peat bogs (14) and classified as type II MB. However, phylogenetically they were only moderately related to the known type II MB and were more closely affiliated with the heterotrophic bacterium Beijerinckia indica subsp. indica.

The isolation of Methylocella palustris from Sphagnum bogs of different geographical locations (four different sites in west Siberia and European north Russia) suggests that these bacteria might be widely distributed in acidic wetlands of the northern hemisphere. These wetlands are considered an important source of atmospheric methane (23). However, information on the distribution and abundance of Methylocella in northern wetlands is still lacking. As the result of their profound distinctness from other known methanotrophs, these organisms have not been targeted by the culture-independent 16S ribosomal DNA (rDNA)-based molecular approaches developed for detection of type I and type II MB (12, 33). This is also true for the retrieval of the pmoA gene, which encodes the active-site polypeptide of particulate methane monooxygenase (pMMO), as M. palustris is the first MB for which pmoA could not be detected using a PCR assay considered universal for this gene (15). Although M. palustris possesses an mmoX gene, coding for the  subunit of the soluble methane monooxygenase (sMMO) and thus can be detected by the mmoX-based approach, this assay is not universal for MB, as only some of them possess sMMO. Thus, effective methods for the in situ detection of Methylocella were not available until now.

One of the most powerful tools in modern microbial ecology is fluorescence in situ hybridization (FISH). FISH allows the specific detection and enumeration of target populations directly in their natural environment without the need for cultivation (3, 4). Although FISH is potentially very useful for studies of methanotroph ecology (28), reports on the enumeration of indigenous methanotroph populations by FISH have not yet been published. So far, FISH-based techniques have only been applied for the analysis of MB in mixed and enrichment cultures (7, 20, 31). A number of 16S rRNA-targeted oligonucleotide probes have been developed for specific detection of type I and type II MB (7, 10, 16, 20, 31). However, none of the currently available probes targets Methylocella palustris. Thus, the primary goal of our study was the development of oligonucleotide probes for the specific detection of these novel acidophilic methanotrophs. The newly developed probes and those of reported group specificity for either type I MB or the Methylosinus/Methylocystis group of type II MB were further applied to determine the abundance of distinct methanotroph groups in a Sphagnum peat (pH 4.2) from west Siberia.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Bacterial strains and growth conditions. Methylocella palustris K (ATCC 700799^T), Methylosinus trichosporium OB3b (ATCC 35070^T), Methylococcus capsulatus (NCIMB 11853), Methylomicrobium album (NCIMB 11123), and Methylobacter luteus (NCIMB 11914) were used in this study. The last three strains represent subcultures of the corresponding National Collections of Industrial, Food, and Marine Bacteria strains and were kindly provided by G. Eller (Max-Planck-Institut für terrestrische Mikrobiologie, Marburg, Germany). Beijerinckia indica subsp. indica (ATCC 9039^T), Bradyrhizobium japonicum (DSM 30131^T), and Azorhizobium caulinodans (DSM 5975^T) were used as nontarget control strains. M. palustris K was grown on twice-diluted nitrogen-sufficient M1 medium supplemented with vitamins (15), under a gas headspace containing 20% methane (vol/vol). The cultures were shaken at 120 rpm at 24°C. All other methanotrophs were cultivated on NMS medium (32) under the same conditions. The only exception was Methylococcus capsulatus, which was incubated at 37°C with 40% methane (vol/vol) in the headspace. Beijerinckia indica was cultivated on nitrogen-free mineral medium supplemented with glucose (5). Azorhizobium caulinodans and Bradyrhizobium japonicum were grown on the media recommended by the Deutsche Sammlung von Mikroorganismen und Zellkulturen (Braunschweig, Germany) catalogue. To ensure constant exponential growth and a high cell ribosome content, all cultures were serially transferred at least three times prior to harvesting.

Methane-oxidizing enrichment cultures. Four methane-oxidizing consortia enriched from acidic Sphagnum peat bogs of different geographic locations were used in FISH with newly developed oligonucleotide probes. The characteristics of these enrichments as well as the cultivation conditions used were described previously (13).

Peat sample. The Sphagnum peat sample was collected from a depth of 10 to 15 cm of an acidic peat (pH of 3.6 to 4.5) underlying a Sphagnum-Carex plant community (Bakchar bog, Plotnikovo field station in west Siberia, 56°N, 82°E). The original sample was transported to the laboratory and divided in two parts. The first subsample was fixed immediately as described below, while the other was incubated for 1 month under 10% methane and then fixed.

Fixation procedure. (i) Bacterial strains and methane-oxidizing enrichments. Cells growing in the logarithmic phase were harvested by centrifugation and resuspended in 0.5 ml of phosphate-buffered saline (PBS) containing, in grams per liter, NaCl, 8.0; KCl, 0.2; Na₂HPO₄, 1.44; and NaH₂PO₄, 0.2 (pH 7.0). Cell suspensions were mixed with 1.5 ml of 4% (wt/vol) freshly prepared paraformaldehyde solution (Sigma, Deisenhofen, Germany) and fixed for 1 h at room temperature. The cells were then collected by centrifugation (6,600 × g for 1 min) and washed twice with PBS to ensure removal of paraformaldehyde. The resulting pellet was resuspended in 0.5 ml of 50% ethanol-PBS (vol/vol), and the cell suspension was stored at 20°C until use.

(ii) Peat samples. Two grams of wet Sphagnum peat was placed into a disposable 50-ml syringe with some sterile cotton glass covering the exit hole. The plunger was replaced and pressed to extract water from the peat. This treatment allowed us to separate the peat water enriched with microbial cells from the rough Sphagnum debris. Approximately 0.5 ml of peat water was mixed with 1.5 ml of 4% (wt/vol) paraformaldehyde solution and fixed as described above. The rest of the peat material was mixed with 20 ml of sterile water and homogenized in a laboratory stomacher (model 80, Seward Medical Limited, London, United Kingdom) at 265 rpm for 1 min, and the water was again extracted with a syringe. The fraction obtained was designated the peat matrix fraction and was centrifuged (5,000 rpm) for 20 min at 4°C. The supernatant was discarded, and the pellet was resuspended with sterile water up to a final volume of 2 ml. An aliquot (0.5 ml) of this suspension was used for paraformaldehyde fixation. To assess the efficacy of the cell extraction from peat, the stomacher treatment (see above) was repeated two additional times and two additional peat matrix fractions were obtained. The rest of the peat material was cut into very small fragments (<0.5 mm) with scissors. The two additional peat matrix fractions and 50 mg of the small peat fragments were fixed as described above.

Oligonucleotide probes. Potential 16S rRNA-targeted oligonucleotide probes applicable to the specific detection of Methylocella palustris were formulated using the probe design tool of the ARB program package (developed by O. Strunk and W. Ludwig; available online at http://www.arb-home.de). Based on the 16S rRNA database included in the ARB program package, the probe design tool selects nucleotide sequence regions that allow discrimination of target sequences from all nontarget reference sequences. The final selection of suitable target sites was done using the in situ accessibility map of Escherichia coli 16S rRNA (18). According to the probe nomenclature (1), the newly designed probes were designated S-S-Mcell-1026-a-A-18 (Mcell-1026) and S-S-Mcell-0181-a-A-18 (Mcell-181). For use in FISH, the probes Mcell-1026 and Mcell-181 were labeled with indocarbocyanine dye (Cy3) and 5(6)-carboxyfluorescein-N-hydroxysuccinimide ester (FLUOS), respectively. For enhancement of the fluorescence signal of probe Mcell-181, the oligonucleotide helper probe H158 was developed and used in combination with Mcell-181. Depending on the experimental setup, the bacterial probe EUB338 fluorescently labeled with either Cy3 or FLUOS was used (2). The probes with reported group specificity for either type I MB or type II MB, i.e., probes M-84, M-705, M-450, and MA-221 (7, 16), were applied in FISH with Cy3 label. Oligonucleotide probes were purchased from MWG Biotech (Ebersberg, Germany). Probe sequences, target sites, formamide concentrations in the hybridization buffer, hybridization temperature, and sodium chloride concentrations in the washing buffer used for FISH are given in Table 1.

Whole-cell hybridization. Hybridization was done on 70% ethanol-rinsed and dried Teflon-coated slides with eight wells for independent positioning of the samples. Approximately 2 µl of the fixed cell suspension was spread on each well, air dried, and dehydrated by successive passages through an ethanol series (50, 80, and 100% [vol/vol]) for 3 min each. A 50-ml polypropylene screw-top Falcon tube containing a slip of Whatman filter paper soaked in hybridization buffer was used as a hybridization chamber as described by Stahl and Amann (29). The chamber was allowed to equilibrate for at least 30 min at the hybridization temperature. A 9-µl aliquot of hybridization buffer (0.9 M NaCl, 20 mM Tris-HCl [pH 7.2], 0.01% sodium dodecyl sulfate [SDS], and formamide concentrations as given in Table 1) was placed on each spot of fixed cells. The slide was transferred to the equilibrated chamber and prehybridized for 30 min. Following prehybridization, 1 µl of fluorescent probe solution (50 ng of probe per µl in double-distilled water) was added to each spot, and the slide was returned to the hybridization chamber for 2 h. Then, slides were washed at the hybridization temperature for 10 min in washing buffer (20 mM Tris-HCl, 0.01% SDS, and NaCl at the concentrations given in Table 1) and rinsed with twice-distilled water. The slides were air dried, stained with the universal DNA stain 4',6'-diamidino-2-phenylindole (DAPI; 2 µM) for 10 min in the dark, rinsed again with distilled water, and finally air dried. Each well of the slide was mounted with a drop of Citifluor AF1 antifadent (Citifluor Ltd., Canterbury, United Kingdom), covered with a coverslip and viewed immediately.

Optimization of hybridization conditions. Two approaches were used to optimize the hybridization conditions. These were (i) gradually increasing the hybridization stringency by the addition of formamide to the hybridization buffer in 5% (vol/vol) steps (22) and (ii) increasing the hybridization temperature from 30 to 70°C in 5°C steps without formamide in the hybridization buffer. In each case hybridization was performed with M. palustris and nontarget organisms displaying the smallest number of mismatches within the target region (Fig. 1). Beijerinckia indica subsp. indica and Azorhizobium caulinodans were used as nontarget organisms for probes Mcell-1026 and Mcell-181, respectively.

View larger version (32K): [in this window] [in a new window]   FIG. 1.   Alignment of 16S rRNA target regions of probes Mcell-1026 and Mcell-181. The three currently known strains of M. palustris, strains K, S6, and M131 (14, 15), exhibit identical target regions for the two probes. The alignment shows those bacterial reference species and environmental clone sequences which display the smallest number of mismatches in the target region of Mcell-1026 and Mcell-181. The GenBank accession numbers for the environmental clones MC5, MC8, MC10, MC12, MC16, and Mph14 are X65581, X65576, X65579, X65577, X65575, and U58013, respectively.

Quantitative FISH procedure. The slides with fixed peat fractions were simultaneously hybridized with Cy3-labeled Mcell-1026 and FLUOS-labeled Mcell-181. Special attention was paid to make sure that target cells were hybridized with both probes. However, the high background fluorescence due to a large amount of semidecomposed Sphagnum material made it difficult to use the FLUOS-labeled probe for enumeration of MB. In contrast, bacteria hybridized with the Cy3-labeled probe stood out sharply against the background. Thus, cell counting of M. palustris was performed on preparations hybridized with Cy3-labeled Mcell-1026. In addition, the number of cells stained with the bacterial probe EUB338, type I MB probes M-84 and M-705, and type II MB probes M-450 and MA-221 were determined. Cell counting was performed on 100 randomly chosen fields of view (FOV) for each test sample. The number of target cells per gram of wet peat was determined from the area of the sample spot, the area of the FOV, the volume of the fixed sample used for hybridization, and the volume of the peat water extracted (PWE) from the sample as follows: total number of target cells per gram of peat = (mean target cell number per FOV × area of sample spot × total volume of the fixed sample × total volume of the PWE)/(area of FOV × volume of the fixed sample applied × dilution × PWE aliquot taken for fixation × peat sample weight).

Total cell counts. Total cell counts from peat samples were obtained by DAPI staining (see above). Dilutions that resulted in approximately 20 to 200 cells per FOV were used. The total cell number in 100 randomly chosen FOV was determined as described above.

Microscopy. Cells were counted with a Zeiss Axioplan 2 microscope (Zeiss, Jena, Germany) equipped with the following filter sets: HQ light filter AHF/AF 41001 (AHF Analysentechnik, Tübingen, Germany) for FLUOS-labeled probes (excitation 460 to 500 nm, emission 510 to 560 nm), AHF/F 41007 for Cy3-labeled probes (excitation 510 to 560 nm, emission 572.5 to 642.5 nm), and Zeiss Filter 02 for DAPI staining (excitation 365 nm, long-pass emission 420 nm). Color micrographs were made on Fujichrome Provia 1600 ASA color reversal film. Exposure times were 0.04 to 0.06 s for phase contrast and 30 to 60 s for epifluorescence micrographs.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

Probe design. Two probes, Mcell-1026 and Mcell-181, each targeting a unique region of the 16S rRNA of Methylocella palustris, were developed based on the public domain 16S rRNA data set and the probe design tool of the ARB program package. The specificity of the two newly designed probes was also confirmed using the probe match tool of the Ribosomal Database Project (21). The sequence of probe Mcell-1026 matched the sequences of all three known strains of Methylocella palustris but exhibited at least three mismatches to all other currently available 16S rRNA reference sequences (Fig. 1). The probe Mcell-181 had one mismatch (A:G at position 187) with the 16S rRNA sequence of Azorhizobium caulinodans, while all other nontarget reference sequences from cultured organisms displayed at least two mismatches to this probe (Fig. 1). Consequently, Beijerinckia indica subsp. indica and Azorhizobium caulinodans, whose 16S rRNA sequences showed the smallest number of mismatches in the target region to the probes Mcell-1026 and Mcell-181, respectively, were chosen as negative control strains for optimization of the probe hybridization conditions.

Helper probe. Initial tests showed that Cy3-labeled probe Mcell-1026 provided higher signal intensity with M. palustris K than Cy3-labeled EUB338, yielding a high resolution for environmental studies. In contrast, the FLUOS-labeled Mcell-181 probe showed a lower intensity signal than did FLUOS-labeled EUB338. In order to enhance the fluorescence signal of Mcell-181, three potential helper oligonucleotides (H158, H159, and H160) were designed (17). Each of the three oligonucleotides targeted a 16S rRNA region that was adjacent to the target site of probe Mcell-181. Their use in conjunction with Mcell-181 showed that one of them, H158, caused a detectable increase in the probe-conferred signal (results not shown). Thus, in the following studies, probe Mcell-181 was always used in combination with H158. The H158 probe sequence and the corresponding target site are shown in Table 1.

Optimization of hybridization conditions for Methylocella-specific probes. An attempt was made to determine optimal hybridization conditions for the probes Mcell-1026 and Mcell-181 by the conventional approach. In this approach, the hybridization stringency is increased by increasing the concentration of formamide in the hybridization buffer (22). The range of formamide concentrations used was from 0 to 40%, in 5% increments. However, we observed that formamide clearly inhibited the whole-cell hybridization of Methylocella palustris K with all fluorescent probes tested, including EUB338. Simultaneous application of fluorescently labeled probes Mcell-1026 and EUB338 confirmed that the higher the formamide concentration used in the hybridization buffer, the fewer cells were detectable by FISH. However, without formamide in the hybridization buffer all cells were stained with both probes Mcell-1026 and EUB338 (Fig. 2). At formamide concentrations above 20%, most cells of M. palustris K remained unstained regardless of the temperature used for hybridization. This observation was confirmed using several batches of formamide purchased from different companies (Fluka, Merck, Serva, and Sigma). However, we did not observe any negative effect of formamide on whole-cell hybridization of the other methanotroph species tested (Methylosinus trichosporium OB3b, Methylococcus capsulatus, Methylomicrobium album, and Methylobacter luteus). Similarly, the use of formamide did not show any negative effect on whole-cell hybridization of nontarget control strains (Beijerinckia indica subsp. indica and Azorhizobium caulinodans) or of some other bacteria (E. coli, Corynebacterium sp., and Bacillus sp.) with EUB338.

View larger version (90K): [in this window] [in a new window]   FIG. 2.   Whole-cell hybridization of Methylocella palustris K with and without formamide in the hybridization buffer. The phase-contrast images are shown on the right side, the epifluorescence micrographs of whole-cell hybridization with probes Mcell-1026 (yellow) and EUB338 (green) are on the left side and in the middle, respectively. (A) No formamide; (B) 10% formamide in the hybridization buffer. The scale bar (10 µm) applies to each row of images.

Evaluation of 16S rRNA probes with reported group specificity for type II MB. We assessed whether the probes MA-221 and M-450 (Table 1) showed any nonspecific hybridization with M. palustris. The probes MA-221 and M-450 exhibit three and two mismatches, respectively, with the 16S rRNA sequence of M. palustris and, as expected, did not show any binding to cells of M. palustris under any of the hybridization conditions used. A weak nonspecific signal was observed if probe M-450 was applied to cells of Beijerinckia indica subsp. indica under the hybridization conditions reported in the original publication (20% formamide, 46°C) (16). Thus, we used probe M-450 under slightly higher stringency conditions (30% formamide, 46°C). The probe MA-221 had one terminal mismatch (G:U at position 240) to 16S rRNA sequences of numerous nonmethanotrophic bacteria, including Bradyrhizobium japonicum, Rhodopseudomonas palustris, Nitrobacter winogradskyi, Blastobacter denitrificans, Agromonas oligotrophica, and Photorhizobium thompsonianum. Whole-cell hybridization of Bradyrhizobium japonicum with MA-221 showed that the stringency conditions reported in the original publication (10% formamide, 37°C) (7) did not allow discrimination of B. japonicum from type II MB of the Methylosinus/Methylocystis group. Target specificity for probe MA-221 was achieved at 35% formamide and 55°C.

Analysis of methanotrophic consortia enriched from Sphagnum peat bogs. The Methylocella-specific probes Mcell-1026 and Mcell-181 and probes with reported group specificity for type I MB (M-84 and M-705) and type II MB (MA-221 and M-450) were used to analyze four methanotrophic consortia enriched from different Sphagnum bogs on nitrogen-sufficient and nitrogen-free media (13). Whole-cell hybridization revealed that cells of Methylocella palustris were present only in those enrichments obtained on nitrogen-sufficient media (enrichments from Sosvyatskoe, Kyrgyznoye, and Krugloye peat bogs). This result is to be expected, given that the three known strains of Methylocella palustris (S6, K, and M131) were isolated from these three enrichment cultures (14). Likewise, failure to detect Methylocella cells in the fourth methanotrophic enrichment, which was obtained on nitrogen-free medium from Bakchar bog in west Siberia, is reasonable given our failure to isolate Methylocella from the Bakchar consortium or to detect mmoX in DNA of this enrichment. Target cells of type I MB-specific probes (M-84 and M-705) and of type II MB-specific probe MA-221 were not observed in any of the four acidophilic enrichments. The application of probe M-450 detected some type II MB in only the enrichment obtained from Krugloye peat bog. These findings are in good agreement with the results of mmoX-based analyses reported previously (13). Only the DNA of the Krugloye bog enrichment yielded mmoX clones related to the Methylocystis group. The fact that none of the methanotroph-targeted probes used in this study detected MB in the Bakchar enrichment may indicate that hitherto unknown acidophilic methanotrophs are present in this consortium.

Application of FISH to native Sphagnum peat. The Cy3-labeled Mcell-1026 probe in conjunction with FLUOS-labeled EUB338 and DAPI staining was applied to the peat water extracted from a Sphagnum sample. Numerous cells of Methylocella palustris were detected in peat water, and the rRNA content of cells was high enough to allow their detection (Fig. 3). The target cells were present in almost every FOV, allowing a statistically valid counting procedure.

View larger version (52K): [in this window] [in a new window]   FIG. 3.   Specific detection of Methylocella palustris in a peat sample by FISH. Pictured are epifluorescence micrographs of in situ hybridization with probes Mcell-1026 (a) and EUB338 (b), DAPI staining (c), and the phase-contrast image (d). The scale bar (10 µm) applies to all images.

Evaluation of cell extraction procedure. Peat is one of the most problematic environmental samples for FISH-based studies due to intensive autofluorescence of organic compounds (19). In addition, Sphagnum peat contains a large amount of nondecomposed organic material and cannot be prepared as a completely homogenized slurry. We attempted to circumvent these problems by a newly developed procedure, which is based on serial cell extraction from peat (see Materials and Methods).

Quantification of acidophilic MB in peat by FISH. The number of microscopic FOV that need to be examined for a statistically valid quantification of MB in peat was determined experimentally. The water fraction extracted from the peat subsample enriched with MB (see above) was used for this assessment. The numbers of cells detected with Mcell-1026 were determined for 300 FOV, and both the average cell numbers and the corresponding standard error values were calculated for the randomly chosen 25, 50, 100, 200, and 300 FOV (Fig. 4). Low error values and almost the same average cell numbers were obtained for the 100, 200, and 300 FOV treatments, while the corresponding values calculated for the 25 and 50 FOV treatments showed significant variations and high error values. Thus, it was concluded that examination of 100 FOV is sufficient for a statistically valid quantification of methanotroph cells in a peat sample.

View larger version (12K): [in this window] [in a new window]   FIG. 4.   Average cell numbers of Methylocella palustris and the corresponding standard error values as a function of the number of microscopic FOV examined.

Effect of formamide on enumeration of acidophilic MB in native peat. The negative effect of formamide on whole-cell hybridization of Methylocella palustris K in pure culture was also observed for its in situ detection in native peat. A peat sample was used for FISH without and with formamide (5 and 10%) in the hybridization buffer. The three hybridization treatments were performed using the same stringency conditions: the treatment without formamide was carried out at 50°C, while the two treatments with formamide in the hybridization buffer were carried out at 47 and 43°C. The hybridization treatment with probe Mcell-1026 and without formamide led to the detection of 1.2 (±0.1) × 10⁶ cells per g of wet peat. An increasing concentration of formamide in the hybridization buffer resulted in a decreasing number of acidophilic MB detectable by FISH. The number of M. palustris cells detected by FISH with 5 and 10% formamide was only 24 and 2%, respectively, of the cell number detected without formamide in the hybridization buffer.

FISH-based analysis of indigenous methanotrophs inhabiting Sphagnum peat. The Methylocella-specific probes and several fluorescent probes targeting distinct methanotroph groups (Table 1) were applied to a native peat sample. In situ hybridization with probes Mcell-1026 and Mcell-181 revealed that M. palustris was present at a relatively high abundance (1.2 (±0.1) × 10⁶ cells per g of wet peat), and that the cells were present in roughly equal numbers in the peat water and peat matrix fractions. By contrast, probes M-84 and M-705, which target most known genera of type I MB, failed to detect any numerically significant population of these organisms in acidic peat. The number of cells targeted by these probes was low (i.e., close to the detection limit) and comprised only 1.6 (±1.9) × 10³ cells per g of wet peat. The application of probes with reported group specificity for type II MB of the Methylosinus/Methylocystis group, i.e., probes MA-221 and M-450, gave conflicting results for the number of type II MB in acidic peat. The probe MA-221 revealed only 1.8 (±0.8) × 10⁴ cells per g of wet peat, while the use of probe M-450 led to the detection of 1.8 (±0.1) × 10⁶ methanotroph cells g¹.