As a first step to determine what kinds of plants have associations with CH₄ consuming methanotrophs, CH₄ consumption by plant samples was assayed by analyzing the changes of CH₄ concentration in cultures inoculated with various plant parts listed in Supplemental Table 1. About one gram of plant material was placed into a glass-vial (60 mL capacity), which was filled with 20 ml of autoclaved mineral NMS medium (Whittenbury et al., 1970) supplemented with trace vitamins and trace elements (Yoshida et al., 2007). Submerged parts of emergent aquatic plants were used for incubations after washing by shaking in distilled water three times to remove soil. The incubation vials were closed with Teflon-coated butyl rubber caps. CH₄ (5% v/v) was added to the vial by syringe and vials were incubated for 14 days at 28°C with shaking.

Eight submerged aquatic plants and two aquatic plants with floating or aerial leaves (Figure 1) were sampled from a shallow eutrophic sub-basin (614 km² of surface area and 3.5 m average depth) of Lake Biwa in Japan (35°4′ 32″N, 135°56′ 5″E), where the area covered with aquatic plants amounted to 52% of the total surface area in the basin in 2000 (Hamabata and Kobayashi, 2002). To analyze CH₄ consumption activities by the aquatic plants, 5 g of plant material (wet weight) from each macrophyte was placed into 120 mL glass vials. Rice, Oryza sativa spp. japonica, cultivator Koshihikari, which had been growing for about 2 months and was 30 cm in height, was analyzed for comparison. The glass vials were supplemented with 1% (v/v) CH₄ and closed with butyl-lubber caps. The vials were incubated for 30 h at 25°C in the dark and the CH₄ in the headspace was analyzed every hour by using GC-FID as reported previously (Iguchi et al., 2011). Ten mL of lake water was used instead of aquatic plants to compare the CH₄ consumption activities with those of aquatic plants. Aquatic plants and rice were washed by shaking three times in lake water and flood water from the rice field, respectively.

For analysis of methanotrophic microbial communities in the aquatic plants, approximately 0.5 g of plant material (wet weight) from each plant were used for DNA extraction. DNA was extracted from the aquatic plants by using an ISOIL DNA extraction kit (NIPPON GENE, Tokyo, Japan) according to manufacturer's instructions.

qPCR targeting the pmoA gene using primers A189 and A682 (Holmes et al., 1995) was performed with a LightCycler FastStart DNA Master SYBR green I kit (Roche Molecular Biochemicals, Indianapolis, IN) and the LightCycler system (Roche Applied Science, Indianapolis, IN) as described previously (Yoshida et al., 2009). The PCR profile consisted of preheating at 95°C for 10 min and 40 cycles of denaturation at 95°C for 15 s, annealing at 60°C for 10 s, and extension at 72°C for 20 s. The fluorescence signal was detected at 72°C during each cycle, and a melting curve was obtained by heating the product to 95°C and cooling it to 40°C. The calibration curves were graphed and the copy numbers of pmoA were calculated with LightCycler software version 3.5 (Roche Diagnostics) using serial dilutions of the pmoA amplicons from Methylococcus capsulatus Bath as a standard. The PCR efficiency was 90%, which was calculated from qPCR using serially diluted standard samples prepared from the pmoA amplicons of Methylococcus capsulatus Bath. Melting curve analyses demonstrated the absence of primer dimer formation and non-specific amplifications in all samples facilitating the use of 72°C as temperature of recording fluorescence.

The extracted DNA samples were used as templates for amplification of pmoA by using two pmoA-specific primer sets, A189 and A682 (Holmes et al., 1995) and A189 and mb661 (Costello and Lidstrom, 1999). PCR using KOD Fx Neo (TOYOBO, Osaka, Japan) was performed by preheating the mixture to 94°C for 2 min, followed by 30 cycles of denaturation at 98°C for 10 s, annealing at 55°C for 30 s, and extension at 74°C for 30 s. The pmoA amplicons from aquatic plants were cloned by using a pCR^®8/GW/TOPO^® TA cloning kit (Invitrogen, Carlsbad, CA, USA) and sequenced using a BigDye Terminator v3.1 cycle sequencing kit and an ABI 3130 Genetic Analyzer (Applied Biosystems). The clones were chimera-checked by using chimera detection software, Bellerophon (Huber et al., 2004) and classified to operational taxonomic units (OTUs) having <0.07 of the evolutional distance (e.d.) in amino acid sequences by using Mothur (Schloss et al., 2009). The phylogenetic tree of the OTUs and relatives was drawn by the neighborhood joining method using MEGA5.1 (Tamura et al., 2011).

A piece (about 1 cm²) of aquatic plant material was inoculated into a glass-vial (30 mL capacity) containing 5 ml of NMS or AMS medium (Whittenbury et al., 1970) and 10% CH₄ (v/v) as described above. The cultures were incubated at 28°C with shaking and subcultured in fresh medium when CH₄ in the headspace was consumed. After 3–4 serial transfers, the cultures were serially diluted and spread onto 1.5% agar plates of NMS or AMS medium. The agar plates were incubated for 2–4 weeks at 28°C in a jar filled with a CH₄/air mixture. About 100 colonies were picked and subcultured from each culture. The colonies were purified by repeating agar cultivation until a single colony morphotype was obtained. The isolates were phylogenetically identified by sequencing 16S rRNA and pmoA genes, which were amplified from cell lysates of the isolates (Yoshida et al., 2007).

The nucleotide sequence data reported in this paper have been deposited under DDBJ/EMBL/GenBank (http://www.ddbj.nig.ac.jp/Welcome-j.html) accession no. AB844797–AB845152, AB845287–AB845309, and AB845153–AB845175.

Our previous study indicated that methanotrophs inhabit emergent parts of various plant surfaces (Iguchi et al., 2012). In this work, we aimed to determine what types of plant could potentially serve as a niche for methanotrophs. Based on the assumption that CH₄ consumption is only carried out by methanotrophic bacteria, we measured CH₄ consumption in cultures inoculated with plant parts taken from various environments, thereby evaluating CH₄ consumption activity by plant-associated methanotrophic consortia. After 14 days of incubation, none of the cultures inoculated with woody plant leaves showed detectable consumption of CH₄. In contrast, cultures inoculated with several emergent plants, macrophytic marine algae and all aquatic plants consumed significant amounts of CH₄ (Figure 1). In the emergent grass cultures, the cultures of emergent parts consumed higher amounts of CH₄ than those of submerged parts. As a result of this screening, we found that aquatic plants showed a constant high CH₄-consuming activity, and therefore the aquatic plants were subjected to further investigation.

Figure 2 shows the CH₄ consumption by eight submerged aquatic plants and two aquatic plants with floating or aerial leaves. Within 30 h of incubation at 25°C, all 10 aquatic plants consumed significant amounts of CH₄, in contrast to the vial containing lake water without aquatic plants, which did not consume CH₄ within 30 h (data not shown). The CH₄ consumption by aquatic plants was variable among the aquatic plants and sample preparations, exhibiting 3.7–37 μmol·h⁻¹·g⁻¹ dry weight of CH₄ consumption in native submerged samples. In the control using rice as an emergent plant, the CH₄ consumption was lower than that of aquatic plants with values of 0.018 ± 0.08, 0.29 ± 0.22, and 0.37 ± 0.03 μmol·h⁻¹·g⁻¹ dry weight for leaves, stems and washed roots, respectively. These results showed that aquatic plants had much higher CH₄ consumption activity than the emergent plants.

CH₄ consumption by aquatic plants declined to 0.45–4.4 μmol·h⁻¹·g⁻¹ dry weight when bacterial biofilms were removed by washing. In Egeria densa, Cabomba caroliniana, and Chara braunii, 92–99% of the CH₄ consumption was lost in the washed samples, whereas submerged parts of aquatic plants with floating or aerial leaves, Eichhomia cassipes and Trapa japonica retained 61 and 41% of the activity after washing, respectively. These results suggest that most methanotrophs were loosely associated with the surface of aquatic plants, but some associated tightly with or were within aquatic plants.

qPCR targeting a component of the pMMO-encoding gene, pmoA, produced amplicons of a single size from the extracted DNA of all CH₄ consuming samples, i.e., E. densa, C. caroliniana, C. braunii, Potamogeton maackianus, Vallisneria natans, P. crispus, and submerged parts of T. japonica and E. cassipes. The qPCR-estimated copy numbers of pmoA from those samples were well correlated (r² = 0.71) with the CH₄ consumption activities (Figure 3), clearly indicating the involvement of methanotrophs in CH₄ consumption. The populations of methanotrophs in aquatic plants were also variable. In the native sample, the copy numbers of pmoA were in the range from 4.6 ± 1.8 × 10⁵ to 6.9 ± 0.77 × 10⁷ copies·g⁻¹ dry weight. The pmoA copy number declined to 990 ± 510 to 1.0 × 10⁴ ± 890 copies·g⁻¹ dry weight after washing, corresponding to 0.001–6.5% of the pmoA before washing. As observed with CH₄ consumption activities, this result also supported a loose association of methanotrophs with the surface of aquatic plants. To date, populations of methanotrophs associated with emergent parts of plants have not been investigated in detail. Our present results indicate that methanotrophs associated with aquatic plants constitute a larger CH₄ sink than those associated with the phyllosphere of terrestrial plants, and have demonstrated an unexpected contribution of aquatic plants to CH₄ consumption.

A total of 358 cloned pmoA gene sequences were determined and classified into 21 OTUs (Figure 4 and Supplemental Table 2). These sequences were similar to those in members of the genus Methylosinus, which are type II methanotrophs in the Alphaproteobacteria, and the genera Methylosarcina, Methylomicrobium, Methylomonas, Methylovulum, Methylocaldum, and Methylococcus, which are type I methanotrophs in the Gammaproteobacteria. Three OTUs, OTU3, OTU19, and OTU21 were detected in all aquatic plants and represented 28, 20, and 11% of the total clones, respectively. The sequences of OTU3, the most abundant OTU, were most closely related (94–99% amino acid sequence identity) to pmoA sequences of Methylosarcina lacus LW14 (Kalyuzhnaya et al., 2005) among the available isolates, and were most closely related to pmoA-clones from rice roots (Shrestha et al., 2010), lake littoral wetlands (Siljanen et al., 2011), lake littoral sediments (Deutzmann et al., 2011), and rice field soil (Mayumi et al., 2010) (Figure 5). The sequences of OTU19, the second most abundant OTU, were 87–89% similar to the pmoA sequence from Methylocaldum gracile VKM-14L (Bodrossy et al., 1997) among isolated strains, and all of the clones were most closely related to pmoA clones from rice roots (Shrestha et al., 2008). The sequence of the third most abundant OTU (OTU21), which was 91–95% similar to the pmoA sequence of Methylococcaceae sp. strain OS501 (Iguchi et al., 2011) as the closest related isolate, was most closely related to a variety of environmental clones from littoral sediments of Lake Constance (Bussmann et al., 2004), CH₄-consuming sludge (Osaka et al., 2008), and rice roots (Shrestha et al., 2010). In addition to those highly abundant OTUs, six OTUs had more than 5% frequency among the total clones and were detected in more than four aquatic plants as follows; the sequences of OTU1 were closely related to clones from lake sediment (Costello and Lidstrom, 1999; Pester et al., 2004), rice roots (Lüke et al., 2010), plant leaves (Iguchi et al., 2012) and onshore oil and gas field soil (Xu et al., 2013), the sequences of OTU2 were related to clones from landfill soils (Henneberger et al., 2012), the sequences of OTU4, OTU8, and OTU14 were related to clones from lake sediments (Costello and Lidstrom, 1999; Lin et al., 2004; Nercessian et al., 2005; Bussmann et al., 2006; Deutzmann et al., 2011; Siljanen et al., 2011), and the sequences of OTU15 were related to clones from rice roots (Shrestha et al., 2008; Lüke et al., 2010) and rice field soil (Ho et al., 2011).

Among the identified OTUs, all sequences of OTU19 as well as most sequences of OTU15 formed a cluster with the sequences of the clones obtained from rice roots, but not with other clusters. In particular, methanotrophs corresponding to OTU19 were widely and abundantly distributed in seven different aquatic plants and the sequences formed a cluster only with the sequences of the clones obtained from two different rice roots in different places. These results suggest that methanotrophs represented by OTU19 live in association with aquatic plants.

We attempted to enrich and isolate methanotrophs associated with C. demersum, E. densa, E. crassipes, T. japonica, P. crispus, and C. caroliniana. Since the use of NMS and AMS media could result in biased enrichment of particular methanotrophs (Hoefman et al., 2014), our results are not expected to directly reflect the population of methanotrophs on plant surfaces. Nevertheless, we wanted to isolate a variety of methanotrophs from plant surfaces, and therefore, NMS and AMS media were applied as the first trial. Indeed, our previous study also showed that a novel genus of type-I methanotroph, Methylovulum miyakonense (Iguchi et al., 2011) could be isolated by using NMS medium. Each culture that had been inoculated with a piece of an aquatic plant consumed all of the CH₄ within a week, and maintained CH₄ consumption activity after several serial transfers. For each sample, 1–9 colonies among 100 colonies grown on agar plates consumed CH₄, and these colonies were further purified. The 100 picked colonies included 50 colonies from NMS medium and 50 colonies from AMS medium. Finally, 23 strains were isolated; these belonged to the genera Methylosinus, and Methylocystis based on sequencing of 16S rRNA and pmoA genes (Supplemental Figure 1). However, no type-I methanotrophs were isolated despite their abundance in aquatic plants. The nitrogen source used did not seem to result in isolation of specific methanotrophs, as similar strains were obtained using both NMS and AMS. Although there seems to have been a bias in the isolation of the dominant methanotrophs from aquatic plants due to the enrichment method, isolation of methanotrophic strains from aquatic plants would be the first step to reconstructing and establishing an ecological system of aquatic plants and methanotrophs having high CH₄ consuming activity.