Diversity of Oxygenase Genes from Methane- and Ammonia-Oxidizing Bacteria in the Eastern Snake River Plain Aquifer

Sampling.The microbial communities present in the ESRPA were sampled in the fall of 2003. Two wells, ANP-9 and ANP-10, located in a pristine region of the aquifer proximal to the TCE-contaminated plume, were used for this study. A detailed site description has recently been reported by Newby et al. (40). The close proximity of the wells to each other (ca. 200 m) suggests that their respective water chemistries and microbial community structures should be similar. Previous analysis indicates comparable water chemistry characteristics between the two wells and values that are typical of the ESRPA. For example, selected water chemistry values for ANP-9 and this region of the aquifer in general include temperature of 14.1 to 14.5°C, a pH of 7.2 to 7.9, a dissolved oxygen level of 7.5 to 8.4 mg liter⁻¹, a dissolved organic carbon level of 1.5 to 2.6 mg liter⁻¹, and a methane concentration of 72 nmol liter⁻¹ (40). Well ANP-9 was used in the collection of free-living cells, while well ANP-10 was used for attached community sampling.

The free-living community in this study was defined as all cells freely dispersed within the aquifer groundwater or only loosely associated with the basalt. The attached community consisted of all cells that colonized the surfaces of basalt substrates that were incubated in well ANP-10. Methods were designed to facilitate the sampling of each community by using existing wells (18, 27-29).

A KrosFlo Pilot hollow-fiber, tangential flow filtration system with a DynaFibre microporous membrane (0.2 μm pore size and 3.9 m² in surface area; Spectrum Labs, Rancho Dominguez, Calif.) was employed for the collection of free-living organisms from ANP-9. Prior to the initiation of filtration, the well was flushed with 3 well volumes (1 volume equal to the standing water in the well) to remove built-up sediments and debris from the casing. Approximately 13,300 liters of aquifer water was then passed through the hollow-fiber apparatus and concentrated to approximately 3 liters. The concentrate was collected by reversing the pump motor and directing the retentate into sterile 1-liter containers. This concentrate was then stored at −80°C and thawed at room temperature prior to being used.

The attached communities were sampled through the use of substrate columns. These columns consisted of crushed basalt chips (4 to 7 mm sieve size) housed in a 76-cm-long nylon mesh tubing (inside diameter, 2 cm; mesh size, 0.25 cm; InterNet, Inc., Minneapolis, Minn.). Prior to use, the columns were rinsed with deionized water to remove fine particles and then autoclaved (60 min) on three successive days. The sterile columns were then suspended in the well to a depth of 80 m for 6 months. At the end of this period, the columns were removed and the attached microbial mass was harvested.

Metagenomic DNA isolation.The UltraClean Mega Prep soil DNA kit (MoBio Laboratories, Inc., Solana Beach, Calif.) was used to harvest the attached community metagenomic DNA from the basalt chips. Four 50-ml polypropylene tubes were filled with 40 g of the substrates and secured horizontally in a rotating incubator (Innova 4300; New Brunswick Scientific Co., Inc., Edison, N.J.) at 200 rpm at 60°C for 30 min to remove bacterial biomass from the basalt substrates. Metagenomic DNA was recovered as per the manufacturer's protocol. After recovery, the extractions were pooled and concentrated by using a Microcon YM-30 ultrafiltration centrifugal device (Millipore, Billerica, Mass.).

Free-living cells were recovered by filtering 50 ml of the concentrated material through a 25-mm (0.20-μm)-pore-size Nuclepore polycarbonate membrane filter (Whatman, Inc., Clifton, N.J.). DNA was then extracted from the filtered material by use of the UltraClean soil DNA isolation kit from MoBio. Briefly, the filter was placed into the bead solution and vortexed on high for 10 s to disperse the filtered material. The tubes were then placed in a 65°C water bath for 30 min. The purification of the lysed material was performed as per the manufacturer's protocol. This method was performed on three additional aliquots so that a total concentrate volume of 200 ml was processed. The resulting DNA from each extraction was pooled and stored at −20°C. DNA recovered from the substrate-attached bacteria by the modified MoBio Mega Prep method resulted in 140 μl of DNA at a concentration of 24 ng μl⁻¹ (A₂₆₀/A₂₈₀ = 1.5).

Whole-genome amplification.Purified DNA from both communities was amplified by using the Repli-g whole-genome amplification (WGA) kit (Molecular Staging, Inc., New Haven, Conn.) to produce two additional community archives, W9 and W10. Briefly, the kit utilizes φ29 polymerase in conjunction with exonuclease-resistant random hexamer primers to isothermally amplify DNA 10⁴- to 10⁶-fold with minimal bias (26). Based on multiple-displacement amplification technology, exponential amplification occurs through a hyperbranching mechanism (12, 25). Long products, averaging 12 kb in length, are obtained due to the high processivity rate of >70,000 nucleotides incorporated for each primer binding event (8). φ29 also exhibits the highest fidelity rate of any known polymerase, with an error rate of only 1 in 10⁶ to 10⁷ nucleotides incorporated (14).

For each sample, 10 ng of genomic DNA was chemically denatured by being mixed with 2.5 μl of denaturation solution (40 mM KOH, 1 mM EDTA [pH 8.0]) and incubated at room temperature for 3 min. The reaction mixture was neutralized by the addition of 5 μl of neutralization solution (a 1:10 dilution of supplied solution B in distilled water). Amplification reactions were performed in 50-μl volumes containing 12.5 μl of 4× WGA PCR mix (as supplied by the manufacturer), 0.5 μl of DNA polymerase mix, 27 μl of sterile H₂O, and 10 μl of denatured sample. The reaction mixtures were then incubated at 30°C for 16 h. Amplification of the free-living DNA was only successful after two washes with distilled water followed by concentration with a Microcon filter device. Amplified products were visualized by using 1% agarose gel electrophoresis stained with ethidium bromide. The amplified DNA was quantified by densitometry using Kodak 1D software (Kodak, Rochester, N.Y.).

Functional gene PCR amplification.Three sets of PCR primers were used in this study to elucidate the oxygenase genes associated with methane- and ammonia-oxidizing bacteria from each community (attached and free-living) and WGA material (W9 and W10) (Table 1). To date, no single primer set has been demonstrated to amplify mmoX genes from all sMMO-producing bacteria.

Primers that targeted the mmoX gene and the pmoA gene specifically were used, as was a set that allowed the corecovery of amoA gene fragments along with pmoA. All PCR products were verified on 1% agarose gels stained with ethidium bromide and visualized by using UV light. Images were captured and recorded by using Kodak 1D software. Reagents and enzymes were purchased from Fisher Scientific (Hampton, N.H.) unless otherwise specified. Oligonucleotides were synthesized by Integrated DNA Technologies, Inc. (Coralville, Iowa) with standard desalting.

pMMO diversity within the communities was assessed through the use of primers designed by Holmes et al. (19) and Costello and Lidstrom (11) (Table 1). Both groups utilized the same forward primer, A189, in conjunction with one of two reverse primers. The first reverse primer, A682, was designed to target a highly conserved region in both the pmoA and amoA genes. These genes are believed to encode the active site containing the 27-kDa polypeptide subunit of both the MMO and the AMO. The A189-A682 primer set has been shown previously to amplify a 525-bp internal section of both pmoA and amoA (19). The second reverse primer, mb661, has been shown to preferentially amplify only an internal section of approximately 470 bp from the pmoA gene (11). PCR amplifications with both primer sets were performed in 50-μl volumes containing a 200 μM concentration of each deoxynucleoside triphosphate, 1× PCR buffer, 10 μg of bovine serum albumin, 1.2 mM MgCl₂, 10 pmol of each forward and reverse primer, 2.5 U of Taq polymerase, and 20 ng of DNA template. Reaction mixtures were subjected to initial denaturation for 5 min at 94°C followed by 30 cycles of 57°C for 1 min, 72°C for 1 min, and 94°C for 1 min. The final step consisted of a 57°C annealing for 1 min, followed by a 10-min elongation at 72°C.

The primer set mmoXA and mmoXB was used to amplify an approximately 1,230-bp fragment from the mmoX gene as described by Auman et al. (6). This primer set is designed to target the region of mmoX that encodes the conserved fragment of the α subunit of the hydroxylase component of sMMO. PCR amplifications were performed in 50-μl volumes containing 200 μM concentrations of each deoxynucleoside triphosphate, 1× PCR buffer, 10 μg of bovine serum albumin, 1.2 mM MgCl₂, 10 pmol of each forward and reverse primer, 2.5 U Taq polymerase, and 20 ng of DNA template. Reaction mixtures were subjected to an initial denaturation of 5 min at 94°C followed by 30 cycles of 60°C for 1 min, 72°C for 1 min, and 94°C for 1 min. The final step consisted of a 60°C annealing for 1 min, followed by 10 min of elongation at 72°C.

A PCR on a metagenomic DNA template isolated from an undefined groundwater methanotroph enrichment which had been grown on methane was used as a positive control for each primer set.

Following amplification, all PCR products were purified by a Wizard spin column PCR cleanup (Promega, Madison, Wis.). Some products required additional purification by agarose gel electrophoresis in order for clones with the proper insert size to be obtained. This procedure was done by loading approximately 1 μg of PCR product into 1% SeaPlaque GTG LMP agarose, followed by extraction via Qiaex II gel extraction (QIAGEN, Valencia, Calif.).

Archives were coded based on the well numbers used in the sampling (9 and 10 for free-living and attached communities, respectively) and the primer sets used in the PCR amplification. For instance, the archive produced from free-living cells isolated from well ANP-9 and amplified with the PCR set A189-A682 was coded as 9-682. Archives produced from whole-genome-amplified DNA were designated with a “W” preceding the well identifier.

Functional gene fragment library production.PCR amplicons for each sample were cloned by using the TOPO TA cloning kit for sequencing (version K) (Invitrogen Corporation, Carlsbad, Calif.). Briefly, 4 μl of each fresh PCR product was incubated with 1 μl of salt solution (1.2 mM NaCl, 0.06 M MgCl₂) and 1 μl of pCR4-TOPO vector. Electrocompetent Top10 Escherichia coli cells were electroporated at 2.4 kV in 0.2-cm cuvettes with 2 μl of the cloning reaction mixture. Positive transformants were selected by spread plating onto Luria-Bertani agar plates with either kanamycin (50 μg ml⁻¹) or ampicillin (60 μg ml⁻¹) used as the selective agent and incubated overnight at 37°C. Kanamycin was used when direct PCR amplicons were cloned, whereas ampicillin was used when the amplicon was purified by gel electrophoresis.

Ninety-six colonies were isolated for each library archive and inoculated into 96-well plates containing 200 μl of Luria-Bertani broth and 50 μg of kanamycin ml⁻¹ amended with 1× Haugness buffer (4% glycerol, 3.6 mM K₂HPO₄, 1.3 mM KH₂PO₄, 2 mM trisodium citrate, 1 mM MgSO₄, pH 7.5) as a cryoprotectant. These plates were then incubated in a GeneMachines HiGro microtiter-plate orbital shaker (Genomic Solutions, Ann Arbor, Mich.) for 18 h at 420 rpm and 37°C with an airflow rate of 4 standard liters per min per chamber. Ten wells were randomly sampled for insert verification by PCR using vector-specific flanking primers (M13f and M13r supplied with the kit). Amplicons were visualized on 1% agarose gels prestained with ethidium bromide. After verification, plates were sealed with aluminum tape and stored at −80°C for future use.

Restriction fragment length polymorphism (RFLP) analysis.Plasmid DNA was isolated from the 96-well cultures by using the Montage Plasmid Miniprep₉₆ kit (Millipore). Cultures were prepared by inoculating one round-bottom, 96-well cell culture block (supplied with the kit) containing 1 ml of Terrific broth with 50 μg of kanamycin ml⁻¹ per well for each archive. Blocks were incubated in the HiGro incubator for 24 h under the conditions described above. Cells were harvested by centrifugation at 1,500 × g for 5 min with subsequent removal of supernatant by decanting. The blocks were sealed and pellets were frozen at −20°C prior to processing.

Plasmid DNA was isolated from the cell pellets as per the manufacturer's full lysate protocol. Five wells were selected from each plasmid preparation and DNA quantified by measuring the absorbance at 260 nm.

RFLP patterns were produced for each clone by the digestion of 1 μg of plasmid DNA with 5 U each of MspI and HinPII (New England Biolabs, Beverly, Mass.) in a final reaction mixture volume of 50 μl. Reaction mixtures were incubated at 37°C for 2 h, followed by enzyme inactivation at 65°C for 10 min. The manual grouping of fragmentation patterns was performed by analyzing 10 μl of each digest on 4% NuSieve 3:1 agarose (Cambrex Bio Science, Baltimore, Md.). Gels were stained with ethidium bromide and rinsed with deionized H₂O prior to imaging. One representative from each RFLP group was chosen for further analysis.

Sequencing and phylogenetic analysis.Sequencing of insert DNA was performed on an ABI Prism model 3730 sequencer (Applied Biosystems, Inc., Foster City, Calif.) by using BigDye version 3 chemistry. Contiguous sequences were produced by using ContigExpress software of the Vector NTI suite, version 9. Nucleotide-nucleotide BLAST (blastn) was used to search GenBank for nearest relative sequences. BLAST results and representatives for each archive were aligned by using AlignX software from the Vector NTI suite, version 9.

Edited alignments were evaluated with the maximum likelihood method by using the PAUP package (version 4.0b10; Sinauer Associates, Inc., Sunderland, Mass.). Evolutionary distance calculations were generated by using the model best describing the data for each tree as determined using the DT-ModSel program (38). Confidence estimates for the nodes within phylogenetic trees were performed by bootstrap analysis (100 replicates). Trees were visualized with TreeView software, version 1.6.6 (43).

Sampling.Free-living cell collection produced 3 liters of concentrate at 4.2 × 10⁶ cells ml⁻¹ (1.26 × 10¹⁰ cells in total), as determined by direct microscopic counts of 4′,6-diamidino-2-phenylindole (DAPI)-stained cells. After taking into account a 1:4,433 dilution factor, these results suggest a concentration of approximately 10³ cells ml⁻¹ in the original ground water. This number agrees with previous cell density estimates from this same well (40). Biomass from the substrate columns was not quantified.

Metagenomic DNA isolated from sampling of free-living cells by the heat-lysis method used in conjunction with the MoBio kit produced a total of 160 μl of DNA at 12.5 ng μl⁻¹ (A₂₆₀/A₂₈₀ = 1.61). The DNA was ∼25 kb in size.

Whole-genome amplification.In an effort to assess the applicability of using whole-genome amplification as a means to generate additional metagenomic template DNA for analyses, archives generated from amplified and unamplified templates were compared. WGA samples resulted in quantities of 550 ng μl⁻¹ and 800 ng μl⁻¹ (quantified by densitometry) for the free-living and attached communities, respectively. These concentrations represent nearly 10⁴-fold amplifications. Both amplified samples migrated between 20 and 25 kb when analyzed by agarose gel electrophoresis.

PCR amplification.DNA extracted from the free-living and attached communities resulted in PCR amplicons of the expected size for a given primer pair. The only exception was the absence of a PCR product from DNA isolated from the attached community with the mmoX-specific primers. These primers were also unsuccessful at amplifying the targeted gene fragment from WGA attached community DNA. In some instances, the products did not resolve as single, isolated bands even after PCR optimization.

Gene fragment library production.It was found initially that high transformation rates could be obtained, although a large percentage (30 to 50%) of those clones contained nonspecific inserts. In order to increase the percentage of clones containing the proper insert size, it was necessary to gel purify the PCR product prior to cloning. This was the case even in the absence of other visible contaminating DNA, such as multiple bands or primer dimers (data not shown). After gel purification, the percentage of clones not containing inserts of the proper size dropped well below 10%.

In general, DNA from all communities transformed at a very low efficiency compared to positive controls supplied with the kit, though some efficiencies were far lower than others. The reason for these differences remains unclear, though we speculate that it may be due to the nature of the oligonucleotide primers used in the PCR. In most cases, multiple cloning and transformation reactions were necessary to produce 96 colonies for the archives.

RFLP and phylogenetic analysis.A total of 96 clones were screened for each of the 10 functional gene archives. Initially, many RFLP patterns were observed in all archives, and a majority of those groups contained only a single representative. Subsequent sequencing revealed that many of these patterns were a result of false-positive clones that did not contain the proper insert. These RFLP groups were discarded from the archives. Figure 1 shows a representative gel containing multiple RFLP patterns from 9-661, 9-682, W9-661, W9-682, and W10-682.

• Open in new tab
• Download powerpoint

FIG. 1.

Representative RFLP analysis. Plasmid DNA isolated from clones produced in the functional gene archives was digested with HinPII. Two hundred fifty nanograms of each digest was separated by using 2% LMP GTG agarose run at 5 V cm⁻¹ for 1.75 h in 1% Tris-acetate-EDTA. Lanes 1 and 12 contain 500 ng of 1-kb Plus DNA molecular weight marker. Lanes 2 and 3, 9-661; lanes 4 and 5, 9-682; lanes 6 and 7, W9-661; lanes 8 and 9, W9-682; lanes 10 and 11, W10-682. The gel was stained with ethidium bromide. Images were captured and recorded by using Kodak 1D software and UV transillumination.

Archives produced with primer set A189-A682 resulted in a combined total of 327 clones. From these clones, only type II methanotroph and AOB sequences that corresponded to α-Proteobacteria and β-Proteobacteria, respectively, were found (Fig. 2). Archive compositions of whole-genome amplified and unamplified communities were similar except that group I uncultured (pmoA) sequences were obtained with less frequency in the WGA free-living community archive than in the unamplified archive, while Nitrosospira spp. (amoA) were represented to a much greater degree in the WGA-derived archive than in the unamplified archive. Interestingly, this difference was not observed for the attached community. A comparison of the free-living and attached communities showed differences between the two communities. Group II uncultured (pmoA) sequences were more prevalent in the attached community than in the free-living community. Furthermore, representative sequences from group I and III uncultured (pmoA) bacteria and Nitrosospira were not found in the attached community archive (Table 2).

• Open in new tab
• Download powerpoint

FIG. 2.

Phylogenetic relationships among partial amoA/pmoA gene sequences produced from archives of unamplified and WGA free-living and attached communities. Gene fragments were obtained by using the primer set A189-A682 (540 base positions were considered). The bar represents 10% sequence divergence as determined by the length of the horizontal lines connecting any two species. The phylogenetic tree was constructed by using PAUP based on maximum likelihood analysis using the K80+G correction model. Bootstrap values greater than 75% are represented by asterisks. Unique clones obtained from the archives are indicated by boldface type. Each clone is named in the following manner: the first number (9 or 10 in bold) refers to the well number, the second number (e.g., 682 in bold) refers to the reverse primer that was used to obtain this sequence, the third number (in bold) is the RFLP group number, and the numbers in parentheses refer to the fraction of clones represented by this phylotype.

Archive 9-682 produced a total of 83 clones from which 13 unique RFLP patterns were obtained (Fig. 2). Just over half (59%) of the clones were found to contain oxygenase genes that group within α-Proteobacteria, while the remainder (41%) were from subclass β-Proteobacteria. BLAST results for groups 1, 3 to 7, 11, 13, 16, and 19 revealed sequences closely (95 to 98%) related to those of previously identified pmoA/amoA-containing bacteria. However, groups 2, 9, and 17 showed only 88% sequence identity to the oxygenase genes of uncultured bacterium M84-P36. These three groups branched separately from M84-P36 and formed a distinctive cluster with W9-682 7 (Fig. 2).

W9-682 produced a total of 91 clones that resulted in 10 unique RFLP patterns, of which 14% were α-Proteobacteria and 86% were β-Proteobacteria (Fig. 2). All groups, with the exception of group 7, yielded high sequence identity (94 to 99%) to previously identified pmoA/amoA-containing bacteria similar to those in 9-682. Group 7, like groups 2, 9, and 17 of 9-682, had a highest identity value of 88% for the uncultured bacterium M84-P36. Group 11 formed a distinct branch within the tree.

Archives 10-682 and W10-682 were both very similar in composition and showed less oxygenase diversity than the free-living community archives, as only five RFLP patterns were seen for 10-682 and six patterns were seen for W10-682 (Fig. 2). Both archives contained a large majority (89 and 97%, respectively) of β-Proteobacteria, of which nearly all showed high levels of similarity to oxygenase genes of the uncultured bacterium gp22 and are distantly related to the AOB Nitrosospira (43 of 45 nucleotides were identical). Two of the groups of α-Proteobacteria, 10-682 4 and W10-682 10, clustered near the MOB Methylocystis sp. 42/22, while the third group of W10-682 15 clustered near other uncultured α-Proteobacteria.

Archives produced with primer set A189-mb661 resulted in a combined total of 317 clones. An overall high percentage of α-Proteobacteria was found in each of these archives. Archive 9-661 and W9-661 contained 97 and 77% Methylocystis, respectively, and both 10-661 and W10-661 were made up of 100% Methylocystis-like sequences (Table 3).

Archives 9-661 and W9-661 were very similar in composition, with a high percentage of each corresponding to Methylocystis sp. (Fig. 3). Seven RFLP patterns were seen in archive 9-661, whereas W9-661 resulted in 13 patterns. Four of the groups, 9-661 9, W9-661 5, W9-661 11, and W9-661 13, branched with the type I methanotroph Methylomonas sp., though they formed a divergent cluster. Another type I methanotroph Methylobacter sp. was distantly related (81% sequence identity) to the groups 9-661 3 and W9-661 12. The nearest identity of these two groups is to the uncultured bacterium LP20 (89%), though the sequences were divergent enough to branch separately.

• Open in new tab
• Download powerpoint

FIG. 3.

Phylogenetic relationships among partial pmoA gene sequences produced from archives of unamplified and WGA free-living and attached communities. Gene fragments were obtained by using the primer set A189-mb661 (550 base positions were considered). The bar represents 10% sequence divergence as determined by the length of the horizontal lines connecting any two species. The phylogenetic tree was constructed by using PAUP based on maximum likelihood analysis using the HKY+G correction model. Bootstrap values greater than 75% are represented by asterisks. Unique clones obtained from the archives are indicated by boldface type. Clones are named according to the manner described in the legend to Fig. 2.

Archives 10-661 and W10-661 each produced six unique RFLP patterns and were similar in composition (Fig. 3). All of the clones screened in each archive were highly related to oxygenase genes of Methylocystis, with >98% sequence identities in most cases and >95% sequence identities in all cases. These clones clustered very closely to the Methylocystis sp. used as the reference in the phylogenetic tree. No type I methanotrophs or AOB were found.

Archives produced from primer set mmoXA-mmoXB resulted in a total of 143 clones. Of these clones, 12 RFLP patterns were produced from 9-mmoX, and 7 patterns were produced from W9-mmoX. Overall, both archives resulted in similar compositions, with Methylocystis-like oxygenase sequences resulting in 95 and 100% of the total clones for 9-mmoX and W9-mmoX, respectively (Table 4).

Phylogenetic analysis of the two archives revealed three distinct clusters, two closely related to Methylocystis sp. and the other to Methylomonas sp. Type I methanotrophs found only in 9-mmoX (groups 5, 8, and 11) were related to Methylomonas sp. LW15 (>91% sequence identity). These three groups collectively account for 5% of the overall archive. Type II methanotrophs represented in 9-mmoX include 87% Methylocystis sp. IMET 10486 (groups 1, 2, 4, and 14), 7% Methylocystis sp. 51 (groups 3, 13, 15, and 16), and 1% Methylocystis sp. SE12 (group 12). Type II methanotrophs represented in W9-mmoX include 68% Methylocystis sp. SE12 (groups 1, 5 and 7), 30% Methylocystis sp. IMET 10486 (groups 6, 10, and 12), and 2% Methylocystis sp. 51 (group 14) (Fig. 4).

• Open in new tab
• Download powerpoint

FIG. 4.

Phylogenetic relationships among partial mmoX gene sequences produced from archives of unamplified and WGA free-living communities. Gene fragments were obtained by using the primer set mmoXA-mmoXB (1,271 base positions were considered). The bar represents 10% sequence divergence as determined by the length of the horizontal lines connecting any two species. The phylogenetic tree was constructed by using PAUP based on maximum likelihood analysis using the TrN+G correction model. Bootstrap values greater than 75% are represented by asterisks. Unique clones obtained from the archives are indicated by boldface type. Clones are named according to the manner described in the legend to Fig. 2.

Sampling.It has been reported that most bacteria in an aquifer are attached to the aquifer solids, and only a small fraction exist as free-living cells (18, 29). The sampling of attached biomass from aquifer boreholes is usually accomplished through the collection of core samples. This method can be expensive, as well as challenging, depending on the aquifer to be sampled. In this study, we chose to use porous columns filled with material simulating the existing fractured basalt of the aquifer and incubated in a borehole in the aquifer.

The use of artificial substrata incubated in wells to sample attached communities close to the study site described herein has been described recently by Lehman et al. (29). The study by Lehman et al. found that the use of dialysis chambers containing basalt (similar to the substrate columns used here) resulted in the recovery of a greater range of microbial diversity compared to that from authentic core samples. Therefore, boreholes may not be ideal for collecting representative samples of microbes colonizing aquifer solids. Free-living communities sampled by filtration of groundwater are also subject to biases. Due to the large volume of ground water used in this study (13,300 liters), it is likely that many different microbial niches were sampled, not all necessarily corresponding to the free-living community. Sediment particles were present in the filtrate, and bacteria colonizing these sediment layers would have been extracted along with free-living cells. It has been observed that bacteria display a dynamic equilibrium between attached and free-living phases (51). These bacteria form biofilms as part of their normal life cycle that culminates with the release of free-living cells capable of colonizing new habitats. Thus, it is difficult to draw a distinct line between attached and free-living communities. Though the communities may differ in overall composition, it is not surprising that we found no distinct disparity between the attached and free-living communities sampled. It is likely that free-living cells present in the free-living samples are also represented, to some extent, in the attached samples.

Whole-genome amplification.WGA technology is based on the unbiased amplification of limited DNA samples (21, 26). This technology was able to amplify our starting DNA to produce a product equal in size to the starting template when visualized by agarose gel electrophoresis (data not shown). PCR of WGA products also resulted in amplicons equal in size to unamplified samples. The differences noted between archives were most likely due not to the WGA process but rather to bias introduced by the PCR. Since multiple PCRs were not pooled, bias generated through a single PCR amplification may have skewed the resulting archives.

Functional gene library production and phylogenetic analysis.The primer sets A189-A682, A189-mb661, and mmoXA-mmoXB were used successfully as functional gene probes in the assessment of MOB and AOB in most of the communities tested. It is unclear why no amplification was seen from the attached and WGA attached DNA with the mmoXA-mmoXB primer set. Sequences were amplified from both attached and WGA attached communities by using the primer sets A189-A682 and A189/mb661, and these were closely related to sequences of the sMMO-producing MOB Methylocystis.

An explanation for our observations may lie in the distinct difference in the species of Methylocystis found in free-living versus attached communities. None of the sequences amplified from the free-living samples with A189-A682 and A189-mb661 was found in the mmoXA-mmoXB archives and vice versa. It appears that the three primer sets used exhibit specificity towards different Methylocystis spp. This fact supports the conclusion that the methanotroph species targeted by mmoXA-mmoXB were not present in the attached samples in sufficient numbers to be detected.

The number of available mmoX sequences on which to base primer design is small. The primer set mmoX f882-mmoX r1403, widely used for mmoX amplification, is based on sMMO gene clusters from only two methanotrophs (34). The mmoXA-mmoXB primer set, which was designed 5 years after the introduction of mmoX f882-mmoX r1403, is based on only six known full mmoX sequences (6). Deduced amino acid alignments show a high percentage of conservation from sequences amplified by Auman et al. (6) with these primers. This result could be due to the fact that the primer set is specific for only highly unique genes. Another study done by Auman and Lidstrom failed to amplify any type II mmoX genes from their samples, even though type II-specific 16S rRNA PCR showed these bacteria to be present (5). Since we observed a similar result in a different environment, it may be that genes found to be conserved in cultivable organisms may not be sufficiently similar to genes in the environment, as noted by Hanson and Hanson (17), and therefore must be considered when interpreting data generated by such primer sets. To date, no single primer set has been demonstrated to amplify mmoX genes from all sMMO-producing bacteria.

As expected, the primer set A189-A682 amplified sequences corresponding to both pmoA and amoA from all four communities. The abundance of amoA-related sequences compared to pmoA found in all four archives except 9-682 suggests that nitrifying bacteria might dominate the samples. Of particular interest is the presence of the large subset of clones that cluster with the uncultured bacterium M84-P36. These clones make up a total of 35% of the archive and are only distantly related to other uncharacterized pmoA clones (20), suggesting the presence of a potentially significant novel group of pmoA-carrying methanotrophs. The proportion of sequences in clone libraries, however, cannot be used to accurately predict the relative abundance of the AOB or the novel pmoA group in the natural communities, due to biases that are inherent in PCR amplification and the number of sequences analyzed. Such proportions, however, are useful for comparisons of archive data within a given data set.

Characterization of the communities with pmoA-specific primers (A189-mb661) suggested that bacteria related to the type II methanotroph Methylocystis dominate free-living as well as attached methanotroph communities. Newby et al. obtained similar results in their study of the methanotroph communities from the ESRPA (40). However, their study sampled a much smaller volume of groundwater and characterized only free-living cells. The free-living communities (amplified and unamplified) also showed the presence of type I methanotrophs related to Methylomonas sp. LW15, a known sMMO-containing bacterium (6). The small percentage of clones from archives 9-661 and W9-661 found to be related to the genus Methylobacter, which has not been shown to produce sMMO, suggests that a broad range of niches may have been sampled through the collection of free-living cells. Again, the Newby et al. study documented the presence of these type I methanotrophs by using a different molecular approach (40). A direct comparison of the gene fragments could not be made, due to differences in the targeted areas of the genes in question.

It has been proposed that growth of type II methanotrophs is favored over growth of type I methanotrophs under low-oxygen, high-methane conditions (17). The analysis of groundwater sampled from well ANP-9 showed a significant concentration of methane but also revealed an oxic environment (40). It is expected that an environment such as this should favor the growth of type I methanotrophs. The presence of copper, however, is another key factor responsible for regulating the growth of type I and type II methanotrophs. Copper is necessary for pMMO activity and plays a key role in the regulation of pMMO and sMMO expression (30). Under low-copper conditions, bacteria with the ability to produce sMMO are able to outcompete those able to produce only pMMO (17, 39). This mechanism is due to the fact that the active pMMO enzyme contains approximately 12 to 15 copper atoms mol⁻¹ (41). While the exact nature of these Cu²⁺ ions in the function of pMMO is not clear, they are postulated to play a role in the active site of the protein (41, 44). Samples of groundwater from ANP-9 show that the copper levels are below the detection limit of 1.0 μg ml⁻¹ (23). Taking into account the conditions present in the wells sampled (the presence of dissolved methane and low copper concentrations), the community composition observed in this study is not surprising.

Our results, obtained through the molecular assessment of oxygenase genes, demonstrate that in both the attached and free-living communities, the majority of the methanotrophic populations present show a high degree of similarity to other type II methanotrophs. The natural attenuation of TCE observed at the TAN site may be due in part to the sMMO-producing communities described here. However, we did not attempt to enumerate methanotrophs from either free-living or attached communities. Because of this fact, the organisms represented in this study constitute an undetermined percentage of the overall bacterial population. Recently, Lehman et al. found that very few MOB or AOB could be recovered from authentic core or groundwater samples taken from within the TCE plume (29). The results from their study may reflect the fact that the enumerations were based on most-probable number assays performed with medium selective for type I methanotrophs and would not have detected the type II populations described here.

The presence of sMMO-producing bacteria alone does not account for the rate of TCE natural attenuation observed at the TAN site. It was our goal to provide evidence that the in situ bacterial populations within the pristine aquifer have the potential for TCE cometabolism. Future work to quantify type II methanotrophs specifically and monitor the production of sMMO is necessary in order to correlate the presence of these bacteria with the TCE attenuation rate observed at the TAN site within the contaminated plume.