Metaproteomic Identification of Diazotrophic Methanotrophs and Their Localization in Root Tissues of Field-Grown Rice Plants

Abstract

In a previous study by our group, CH₄ oxidation and N₂ fixation were simultaneously activated in the roots of wild-type rice plants in a paddy field with no N input; both processes are likely controlled by a rice gene for microbial symbiosis. The present study examined which microorganisms in rice roots were responsible for CH₄ oxidation and N₂ fixation under the field conditions. Metaproteomic analysis of root-associated bacteria from field-grown rice (Oryza sativa Nipponbare) revealed that nitrogenase complex-containing nitrogenase reductase (NifH) and the alpha subunit (NifD) and beta subunit (NifK) of dinitrogenase were mainly derived from type II methanotrophic bacteria of the family Methylocystaceae, including Methylosinus spp. Minor nitrogenase proteins such as Methylocella, Bradyrhizobium, Rhodopseudomonas, and Anaeromyxobacter were also detected. Methane monooxygenase proteins (PmoCBA and MmoXYZCBG) were detected in the same bacterial group of the Methylocystaceae. Because these results indicated that Methylocystaceae members mediate both CH₄ oxidation and N₂ fixation, we examined their localization in rice tissues by using catalyzed reporter deposition-fluorescence in situ hybridization (CARD-FISH). The methanotrophs were localized around the epidermal cells and vascular cylinder in the root tissues of the field-grown rice plants. Our metaproteomics and CARD-FISH results suggest that CH₄ oxidation and N₂ fixation are performed mainly by type II methanotrophs of the Methylocystaceae, including Methylosinus spp., inhabiting the vascular bundles and epidermal cells of rice roots.

INTRODUCTION

Methane is the second most important greenhouse gas in the atmosphere following carbon dioxide, and mitigation of methane emission has become a major global concern in the minimization of global climate change (1). Wetland rice fields are major sources of CH₄ emission, accounting for up to 19% of the global CH₄ budget (1). CH₄ produced from anoxic soils by methanogenic archaea is emitted to the atmosphere via the aerenchyma of the rice plant (2). The paddy rice root and rhizosphere are partially oxic, allowing the growth of aerobic methanotrophic bacteria that utilize methane and methanol as sole carbon and energy sources (3). Up to 90% of CH₄ in the rice root zone is consumed by aerobic methanotrophs (4, 5). On the basis of phylogenetic and physiological differences, aerobic methanotrophs belong to phyla Proteobacteria and Verrucomicrobia, although members of Verrucomicrobia were found as acidophilic methanotrophs in geothermal environments (6). Proteobacterial methanotrophs are classified into two groups: the family Methylocystaceae (type II methanotrophs) belongs to the class Alphaproteobacteria, and the family Methylococcaceae (type I methanotrophs) belongs to the class Gammaproteobacteria (7, 8).

Aerobic methanotrophs inhabiting the root zone and bulk soil have been detected by using culture-independent methods based on pmoA (encoding methane monooxygenase) and 16S rRNA genes in the rice rhizosphere (9–12). Nitrogen fertilization (13, 14) and rice genotype (15, 16) often influence the abundance of methanotrophs in rice fields. It was recently found that OsCCaMK, a microbial symbiosis gene of rice plants, simultaneously enhances CH₄ oxidation and N₂ fixation in a nitrogen-limited environment (17). However, little is known about the microbes in rice roots that are responsible for these processes. There are two possibilities: (i) methanotrophs mediate both CH₄ oxidation and N₂ fixation, and (ii) methanotrophs simply mediate CH₄ oxidation, whereas other diazotrophs fix N₂ by using intermediate substrates such as methanol during CH₄ oxidation (17).

Although N₂ fixation may lead to lowering methane oxidation via energy consumption as hypothesized earlier (18), in situ N₂ fixation in CH₄-enriched soil by type II methanotrophs has previously been demonstrated (19, 20). In support of possibility i, both type I and type II methanotrophs possess nif (nitrogen fixation) genes and both are able to fix N₂ under laboratory experiment conditions (21–23). Moreover, metagenomic and isotopic analyses have revealed type II methanotrophs to be the predominant root-associated bacteria (12, 24). However, these studies lack evidence as to whether methanotrophs contribute to N₂ fixation in rice roots. Therefore, the first aim of the present study was to examine whether methanotrophs simultaneously mediate both CH₄ oxidation and N₂ fixation in the roots of field rice. The second aim was to determine the location of methanotrophic microbes in the root tissues of paddy field-grown rice plants.

Metaproteomics analysis is a powerful approach to identifying microbes and their biogeochemical pathways (25). This approach has been applied to the rhizosphere (26), phyllosphere (27), and rice plants (28) under natural conditions. Knief et al. (28) found that the majority of proteins in the root zones of rice plants were derived from the Alphaproteobacteria, including Methylosinus spp., although they did not discuss the functions of these bacterial species in CH₄ oxidation and N₂ fixation.

Fluorescence in situ hybridization (FISH) has become a standard technique in environmental microbiology. FISH combined with catalyzed reporter deposition (CARD-FISH) is a powerful tool in modern microbial ecology (29). CARD-FISH allows the specific direct detection of target populations in their natural environments, such as soil, sediment, and the rhizosphere (30–33). In addition, CARD-FISH has been used to analyze methanogens in anaerobic sludge (34) and methanotrophs in the soil covering an aged municipal landfill (35). Although FISH or CARD-FISH is potentially very useful for studying methanotroph ecology (11), reports on the use of CARD-FISH to localize indigenous methanotroph populations in paddy field-grown rice roots have not yet been published.

The present study used (i) metaproteomics analysis to analyze the protein diversity in enriched bacterial cells from rice roots and (ii) CARD-FISH to localize methanotrophs in the root tissues of rice plants grown in a paddy field.

MATERIALS AND METHODS

Rice root sampling and bacterial cell enrichment.

The rice (Oryza sativa L.) cultivar Nipponbare was planted in the Kashimadai experimental field of Tohoku University, Japan (38°27′39″N, 141°5′33″E), on 28 May 2010. Since 2004, the field had been treated with P-K fertilizer but without nitrogen fertilizer (17). The rice roots were sampled at the flowering stage, on 22 August 2010. The CH₄ flux (14.8 to 15.4 mg C/h/m²) and N₂ fixation (88.7 nmol h⁻¹ g⁻¹ root dry weight) values have been recorded in the same paddy field where Nipponbare was cultivated in the same stage in 2013 (17). The roots were washed well with tap water until no soil particles remained. Four sets of composite root samples from at least three rice plants (over 100 g each) were independently prepared for each treatment. Approximately 100 g of each root sample was used for microbial metaproteomics analysis; the remainder was used for CARD-FISH to visualize the target bacteria residing in the rice root. For microbial metaproteomics analysis, the composite samples of roots were homogenized to prepare the root-associated bacterial cells, which consisted of both epiphytes and endophytes. The bacterial cells were extracted and purified by using a method for enriching bacterial cells from plant materials by density gradient centrifugation (36).

Protein database construction.

The metagenomic sequences of the rice root-associated microbiome obtained in a previous study (24) were used for proteomic annotation. DNA sequences of rice root microbiomes in the paddy field were generated by shotgun sequencing on 454 GS FLX titanium (454 Life Sciences, Branford, CT); the sequences are listed in DDBJ Sequence Read Archive DRA000321 (24). The metagenomic sequences were assembled with GS De Novo Assembler (454 Life Sciences). MetaGeneAnnotator (37) predicted 55,306 open reading frames (ORFs) from the assembled sequences. Similarity searches for all predicted ORFs were performed by using a BLASTP comparison against the NCBI nr database. The resultant protein database was designated the RRM (Rice Root Microbiome) database.

Proteome analysis.

Proteome analysis was performed as described previously (38). Proteins (50 μg) were separated by using 12.5% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and stained with Coomassie blue. The gel lanes were cut into 60 strips of approximately 1 mm. The gel strips were completely destained with 30% acetonitrile–25 mM NH₄HCO₃, reduced with 10 mM dithiothreitol, and alkylated with 55 mM iodoacetamide. After the gel strips were completely dried, they were digested with 40 μl of sequencing-grade modified trypsin (12.5 ng/μl in 50 mM NH₄HCO₃) at 37°C overnight. The digested peptides were extracted with 25 mM NH₄HCO₃–60% acetonitrile and then twice more with 5% formic acid–70% acetonitrile.

A nano-liquid chromatography (LC)–electrospray ionization–tandem mass spectrometry (MS/MS) analysis for peptide mixtures was performed with an LTQ ion-trap MS (Thermo Fisher Scientific, Yokohama, Japan) coupled with a multidimensional high-performance LC (HPLC) Paradigm MS2 chromatograph (AMR Inc., Tokyo, Japan) and a nanospray electrospray ionization device (Michrom Bioresources Inc., Auburn, CA). The tryptic peptides were loaded onto an octadecyl silane (ODS) L-column2 (Chemicals Evaluation and Research Institute, Tokyo, Japan) packed with C₁₈-modified silica particles (5 μm; pore size, 12 nm) and were separated by a linear gradient of 15% to 65% buffer B for 40 min, followed by a gradient of 65% to 95% buffer B for 1 min (buffer A, 2% methanol–0.1% formic acid–H₂O; buffer B, 90% methanol–0.1% formic acid–H₂O) at a flow rate of 1 μl/min. Peptide spectra were recorded in a mass range of m/z 450 to 1,800. MS/MS spectra were acquired in data-dependent scan mode. After the full-spectrum scan, one MS/MS spectrum of the single most intense peaks was also collected. The dynamic exclusion features were set as follows: a repeat count of 1 within 30 s, an exclusion duration of 180 s, and an exclusion list size of 50.

Using the protein database of the rice microbiome as described above, the MS/MS data obtained were searched against the RRM (Rice Root Microbiome) database by using Mascot program ver. 2.4 (Matrix Science, London, United Kingdom) on an in-house server to identify proteins. Search parameters were set as follows: tryptic digest with a maximum of two missed cleavage sites; fixed modification, carbamidomethyl cysteine; variable modification, methionine oxidation; peptide masses, monoisotopic, positive charge (+1, +2, +3) of peptide; and mass tolerances of 1.2 Da for the precursor ion and 0.8 Da for product ions.

CARD-FISH.

Methylocystis parvus (NCIMB11129) and Methylosinus trichosporium (NCIMB11131) were used as positive controls. Both strains were incubated in nitrate mineral salts (NMS) medium (39) at 30°C with 50% (vol/vol) methane in the headspace. Beijerinckia indica (JCM20098) was used as a negative control.

To detect methanotrophs in the rice roots, samples were prepared in two ways: (i) enriched bacterial cells extracted from rice roots (see above) were used to enumerate methanotrophs, and (ii) rice roots sectioned into ∼2-cm-thick pieces (referred to here as root samples) after being washed well with tap water were used to visualize the localization of methanotrophs. Samples were fixed at 4°C in 3% paraformaldehyde–phosphate-buffered saline (PBS; 130 mM NaCl, 10.8 mM Na₂HPO₄, 4.2 mM NaH₂PO₄ [pH 7.2]) for 2.5 to 3 h in the case of enriched bacterial samples and cultured bacteria (NCIMB11129, NCIMB11131, and JCM20098) and for 6 to 10 h in the case of root samples. All samples were stored in 50% ethanol–PBS at −20°C.

The fixed-root samples were subjected to CARD-FISH within 4 days. To detect methanotrophs in rice roots, a Ma450 probe (5′-ATC CAG GTA CCG TCA TTA TC-3′) targeting type II methanotrophs (40) was selected. The probe was labeled with horseradish peroxidase at the 5′ end (Japan Bio Services Co. Ltd., Saitama, Japan). CARD-FISH for the pure cultures and the enriched samples was done on glass slides in accordance with a method used in previous studies (31, 34). For the root samples, all steps were done in Eppendorf tubes (0.5 or 1.5 ml). The root samples were carefully transferred from tube to tube to avoid tissue damage. For permeabilization, the root samples were immersed in lysozyme solution (10 mg ml⁻¹ in 0.05 M EDTA [pH 8.0] and 0.1 M Tris-HCl [pH 8.0]) and incubated for 60 min at 37°C. After the root samples had been rinsed with ultrapure water, endogenous peroxidases were inactivated by dipping the samples in 0.15% H₂O₂ in methanol for 30 min at room temperature (RT) (31). Subsequently, the samples were dehydrated in 50%, 80%, and 96% ethanol (vol/vol) for 3, 1, and 1 min, respectively, air-dried at RT, and kept in Eppendorf tubes at 4°C until hybridization.

The root samples were equilibrated in 1.5 ml PBS buffer for 15 min at RT before hybridization. Hybridization was performed for 3 h at 40°C in 500 μl hybridization buffer (0.9 M NaCl, 20 mM Tris-HCl [pH 8.0], 10% [wt/vol] dextran sulfate, 2% [wt/vol] blocking reagent [Roche, Mannheim, Germany], 0.1% [wt/vol] SDS, and 30% [vol/vol] formamide) containing 0.1 μM labeled probe, in an incubator rotating at 12 rpm, by the use of an HB hybridization incubator (Taitec, Tokyo, Japan). For washing, the root samples were transferred into 14 ml of prewarmed washing buffer (80 mM NaCl, 5 mM EDTA [pH 8.0], 20 mM Tris-HCl [pH 8.0], and 0.01% [wt/vol] SDS) for 15 min at 42°C with 12 rpm rotation. Subsequently, the root samples were equilibrated by immersion in 14 ml of TNT buffer (100 mM Tris-HCl [pH 7.5], 150 mM NaCl, and 0.05% Triton X-100; Wako, Osaka, Japan) for 15 min at RT. For tyramide signal amplification, Alexa Fluor 488-tyramide was prepared as previously described (41). A mixture of 1 part of Alexa Fluor 488-labeled tyramide and 100 parts of amplification buffer consisting of 10% dextran sulfate, 1× PBS, 0.1% blocking reagent, and 0.0015% H₂O₂ was prepared; the root samples were dipped in 500 μl of the mixture and incubated for 30 min at 37°C in the dark. After signal amplification, the root samples were washed in 14 ml TNT buffer for 15 min and ultrapure water for 1 min. The samples were finally immersed in 96% ethanol for 1 min and air dried.

Microscopic observation.

An epifluorescence microscope (BX50F; Olympus, Tokyo, Japan) with a color charge-coupled-device (CCD) camera and phase-constant system (DP70; Olympus) was used for microscopic observation and image acquisition. To visualize and enumerate the microbes, the samples were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; 1 μg ml⁻¹ for 2 min at room temperature). More than 600 DAPI-stained cells in the enriched samples were counted to calculate the rate of detection of Ma450-positive cells. These experiments were performed in four replicates. To visualize Ma450-positive cells in the roots, the root samples were sectioned (80 to 100 μm thickness) with a Vibratome 3000 system (Vibratome, St. Louis, MO). Microscopy was done with a Zeiss LSM710 microscope (Carl Zeiss, Jena, Germany) and ZEN2011 software (Carl Zeiss).

Nucleotide sequence accession numbers.

The DNA sequences of rice root microbiomes in the paddy field are available under accession number DRX000494 in DDBJ Sequence Read Archive DRA000321.

RESULTS

Proteomic analysis.

By using the metagenomic data set, a total of 4,271 different bacterial and archaeal proteins were identified in the enriched bacterial cells from the roots of rice plants grown in the paddy field (Table 1; see also Table S1 in the supplemental material). More than 90% of the proteins were identified in the protein database constructed previously by using metagenomic sequences (24). The dominant proteins were assigned to members of the classes Alphaproteobacteria (59.6%), Betaproteobacteria (10.9%), and Deltaproteobacteria (9.6%), followed by minor members of Firmicutes (4.5%), Gammaproteobacteria (3%), Actinobacteria (2%), and Archaea (3.1%) (Table 1). Within Alphaproteobacteria, Methylocystaceae and Bradyrhizobiaceae were dominant families (Table 1), including Methylosinus and Methylocystis among the Methylocystaceae and Bradyrhizobium and Rhodopseudomonas among the Bradyrhizobiaceae as major genera with abundant detected proteins (see Table S2).

TABLE 1.

Protein identification to the microbial taxon level of enriched microbial cells from root of paddy field-grown rice plant^a

Domain	Phylum or class	Family	No. of proteins	%
Bacteria			4,138	96.9
	Alphaproteobacteriab		2,544	59.6
		Methylocystaceae	1,270	29.7
		Bradyrhizobiaceae	932	21.8
		Beijerinckiaceae	49	1.1
		Methylobacteriaceae	42	1.0
		Rhizobiaceae	44	1.0
		Xanthobacteraceae	36	0.8
		Rhodospirillaceae	29	0.7
		Hyphomicrobiaceae	23	0.5
		Others	222	5.2
	Betaproteobacteria		464	10.9
		Burkholderiaceae	399	9.3
		Comamonadaceae	19	0.4
		Others	46	1.1
	Deltaproteobacteria		409	9.6
		Myxococcaceae	108	2.5
		Geobacteraceae	59	1.4
		Bdellovibrionaceae	51	1.2
		Bacteriovoracaceae	38	0.9
		Cystobacteraceae	38	0.9
		Desulfovibrionaceae	26	0.6
		Others	89	2.1
	Firmicutes		194	4.5
		Veillonellaceae	40	0.9
		Clostridiaceae	30	0.7
		Peptococcaceae	25	0.6
		Others	99	2.3
	Gammaproteobacteria		129	3.0
	Actinobacteria		84	2.0
	Cyanobacteria		42	1.0
	Bacteroidetes		46	1.1
	Planctomycetes		44	1.0
	Acidobacteria		35	0.8
	Spirochaetes		28	0.7
	Others		119	2.8
Archaea			133	3.1
	Euryarchaeota		127	3.0
		Methanobacteriaceae	79	1.8
		Methanosarcinaceae	28	0.7
		Others	20	0.5
	Others		6	0.1

^aMicrobial taxa with more than 0.5% abundance are shown.

^bThe major genera are Bradyrhizobium and Rhodopseudomonas in addition to type II methanotrophs as shown in Table S2 in the supplemental material. With respect to peptide levels, 4,336 were assigned to Alphaproteobacteria among all 6,447.

Functionally, many proteins were associated with enzymes for CH₄ oxidation and N₂ fixation (Table 2; see also Table S1 in the supplemental material), where they were mainly affiliated with the genera Methylosinus and Methylocystis (Methylocystaceae, type II methanotrophs), Methylocella (Beijerinckiaceae), Methylobacterium (Methylobacteriaceae), and Bradyrhizobium and Rhodopseudomonas (Bradyrhizobiaceae) (Table 2).

TABLE 2.

Proteins relevant to CH₄ oxidation and N₂ fixation in enriched microbial cells from roots of paddy field-grown rice plant^a

Function	Annotationb	Accession no.	Closest organism		No. of peptides
			Genus	Family
Methane oxidation (methane monooxygenase [MMO])	Methane monooxygenase/ammonia monooxygenase, subunit B (P)	ZP_06890678	Methylosinus trichosporium OB3b	Methylocystaceae	13
	Methane monooxygenase/ammonia monooxygenase, subunit C (P)	ZP_06890680	Methylosinus trichosporium OB3b	Methylocystaceae	2
	Methane monooxygenase/ammonia monooxygenase, subunit C (P)	ZP_06888910	Methylosinus trichosporium OB3b	Methylocystaceae	1
	MmoX1 (S)	ABD46892	Methylosinus sporium	Methylocystaceae	26
	MmoX2 (S)	ABD46898	Methylosinus sporium	Methylocystaceae	21
	MmoY (S)	ABD46893	Methylosinus sporium	Methylocystaceae	19
	MmoZ (S)	ABD46895	Methylosinus sporium	Methylocystaceae	17
	MmoB (S)	ABD46894	Methylosinus sporium	Methylocystaceae	7
	MmoC (S)	ABD46897	Methylosinus sporium	Methylocystaceae	1
	MmoG (S)	ABD46891	Methylosinus sporium	Methylocystaceae	2
	MmoX	AAG31818	Methylosinus sp. LW8	Methylocystaceae	5
	MmoY (S)	AAF01269	Methylocystis sp. WI14	Methylocystaceae	10
	PmoB2 (P)	CAE48353	Methylocystis sp. SC2	Methylocystaceae	12
	PmoA2 (P)	CAE48352	Methylocystis sp. SC2	Methylocystaceae	2
	Particulate methane monooxygenase, chain A (P)	3RFR_A	Methylocystis sp. M	Methylocystaceae	2
	Soluble methane monooxygenase protein A beta subunit (S)	AAC45290	Methylocystis sp. M	Methylocystaceae	7
	Soluble methane monooxygenase protein A alpha subunit (S)	AAC45289	Methylocystis sp. M	Methylocystaceae	13
	Soluble methane monooxygenase component A alpha subunit (S)	CAD30360	Methylocystis sp. F10V2a	Methylocystaceae	3
	Particulate methane monooxygenase subunit A (P)	ABD57885	Uncultured Methylocystis sp. GSC357	Methylocystaceae	2
	Monooxygenase component MmoB/DmpM (S)	YP_002361595	Methylocella silvestris BL2	Beijerinckiaceae	3
	Methane monooxygenase (S)	YP_002361593	Methylocella silvestris BL2	Beijerinckiaceae	7
	Methane monooxygenase (S)	YP_002361596	Methylocella silvestris BL2	Beijerinckiaceae	3
	Methane monooxygenase (S)	YP_002361594	Methylocella silvestris BL2	Beijerinckiaceae	6
Methanol oxidation (methanol dehydrogenase [MDH])	PQQ-dependent dehydrogenase, methanol/ethanol family	ZP_06888489	Methylosinus trichosporium OB3b	Methylocystaceae	20
	PQQ-dependent dehydrogenase, methanol/ethanol family	ZP_06888494	Methylosinus trichosporium OB3b	Methylocystaceae	8
	Methanol dehydrogenase beta subunit	ZP_06888486	Methylosinus trichosporium OB3b	Methylocystaceae	4
	Methanol dehydrogenase beta subunit	ZP_08073772	Methylocystis sp. ATCC 49242	Methylocystaceae	4
	PQQ-dependent dehydrogenase, methanol/ethanol family	ZP_08073775	Methylocystis sp. ATCC 49242	Methylocystaceae	41
	Methanol dehydrogenase large subunit-like protein	NP_772853	Bradyrhizobium japonicum USDA 110	Bradyrhizobiaceae	3
	Methanol dehydrogenase large subunit	BAL08762	Bradyrhizobium japonicum USDA 6	Bradyrhizobiaceae	1
	PQQ-dependent dehydrogenase, methanol/ethanol family	YP_002490201	Methylobacterium nodulans ORS 2060	Methylobacteriaceae	14
	PQQ-dependent dehydrogenase, methanol/ethanol family	YP_001756860	Methylobacterium radiotolerans JCM 2831	Methylobacteriaceae	3
	Methanol dehydrogenase	P15279	Methylobacterium organophilum	Methylobacteriaceae	1
	MxaI	ACB32199	Uncultured bacterium 16A2	Uncultured bacteria	4
	MxaJ	ACB32193	Uncultured bacterium 16A2	Uncultured bacteria	2
	MxaR	ACB32198	Uncultured bacterium 16A2	Uncultured bacteria	2
Formaldehyde oxidation (formaldehyde dehydrogenase [FAD])	Formaldehyde-activating enzyme, Fae	ZP_06886820	Methylosinus trichosporium OB3b	Methylocystaceae	16
	Formaldehyde-activating enzyme, Fae	ZP_06886821	Methylosinus trichosporium OB3b	Methylocystaceae	6
	Formaldehyde-activating enzyme, Fae	AAS91609	Methylosinus sp. LW3	Methylocystaceae	2
	Formaldehyde-activating enzyme, Fae	ZP_08072607	Methylocystis sp. ATCC 49242	Methylocystaceae	5
	Formaldehyde-activating enzyme, Fae	ZP_08070745	Methylocystis sp. ATCC 49242	Methylocystaceae	2
	Formaldehyde dehydrogenase	NP_615385	Methanosarcina acetivorans C2A	Methanosarcinaceae	1
Formate oxidation (formate dehydrogenase [FDH])	Formate dehydrogenase, alpha subunit	ZP_06889979	Methylosinus trichosporium OB3b	Methylocystaceae	11
	Formate dehydrogenase, alpha subunit	ZP_08074532	Methylocystis sp. ATCC 49242	Methylocystaceae	1
	Formate dehydrogenase, alpha subunit	ZP_09437742	Bradyrhizobium sp. STM 3843	Bradyrhizobiaceae	1
	Formate dehydrogenase, alpha subunit	YP_001379195	Anaeromyxobacter sp. Fw109–5	Myxococcaceae	1
Nitrogen fixation (nitrogenase [Nif])	Nitrogenase iron protein, NifH	ZP_06886595	Methylosinus trichosporium OB3b	Methylocystaceae	7
	Nitrogenase molybdenum-iron protein alpha chain, NifD	ZP_06886596	Methylosinus trichosporium OB3b		8
	Nitrogenase molybdenum-iron protein beta chain, NifK	ZP_06886597	Methylosinus trichosporium OB3b	Methylocystaceae	8
	Nitrogenase molybdenum-iron protein beta chain, NifK	ZP_08073550	Methylocystis sp. ATCC 49242	Methylocystaceae	2
	Nitrogen fixation protein, NifH	ZP_08073546	Methylocystis sp. ATCC 49242	Methylocystaceae	2
	Nitrogenase molybdenum-iron protein alpha chain, NifD	ZP_08073551	Methylocystis sp. ATCC 49242	Methylocystaceae	1
	Nitrogenase molybdenum-iron protein beta chain, NifK	YP_002363877	Methylocella silvestris BL2	Beijerinckiaceae	5
	Nitrogenase molybdenum-iron protein alpha chain, NifD	ZP_09438590	Bradyrhizobium sp. STM 3843	Bradyrhizobiaceae	2
	Nitrogenase molybdenum-iron protein alpha chain, NifD	YP_004111075	Rhodopseudomonas palustris DX-1	Bradyrhizobiaceae	1
	Nitrogenase molybdenum-iron protein beta chain, NifK	NP_949952	Rhodopseudomonas palustris CGA009	Bradyrhizobiaceae	2
	Nitrogenase molybdenum-iron protein alpha chain, NifD	YP_004677462	Hyphomicrobium sp. MC1	Hyphomicrobiaceae	1
	Nitrogenase reductase, NifH	ZP_03515899	Rhizobium etli IE4771	Rhizobiaceae	1
	Nitrogenase iron protein, NifH	YP_001380211	Anaeromyxobacter sp. Fw109–5	Myxococcaceae	2
	Nitrogenase iron protein 1, NifH1	P51602	Methanobacterium ivanovii	Methanobacteriaceae	1
	Nitrogenase alpha chain, VnfD	YP_004291600	Methanobacterium sp. AL-21	Methanobacteriaceae	1
	Nitrogenase molybdenum-iron protein beta chain, NifK	NP_618770	Methanosarcina acetivorans C2A	Methanosarcinaceae	1

^aBoldface entries indicate members of the family Methylocystaceae, including Methylosinus and Methylocystis.

^b“P” and “S” in parentheses indicate particulate methane monooxygenase (pMMO) and soluble methane monooxygenase (sMMO), respectively.

The biochemical processes of conversion from CH₄ to methanol, from methanol to formaldehyde, from formaldehyde to formate, and from formate to CO₂ are mediated by methane monooxygenase (MMO), methanol dehydrogenase (MDH), formaldehyde dehydrogenase (FADH), and formate dehydrogenase (FDH), respectively (3). In relation to CH₄ oxidation, MMOs were assigned to Methylosinus (MmoX1, MmoX2, MmoY, MmoZ, MmoB, MmoC, MmoG, MmoX, and methane monooxygenase), Methylocystis (MmoY, PmoA2, PmoB2, particulate-form MMO [pMMO] subunits, and soluble-form MMO [sMMO] subunits), and Methylocella (MmoB and methane monooxygenase) (Table 2). Biochemically, MMO generally has two forms, a membrane-bound particulate form (pMMO) and a cytoplasmic soluble form (sMMO) (3, 8). The detected MMO proteins included both sMMO and pMMO (Table 2).

MDHs (14 proteins) were detected in Methylosinus (pyrroloquinoline quinone [PQQ]-dependent dehydrogenase and methanol dehydrogenase beta subunit), Methylocystis (PQQ-dependent dehydrogenase), Bradyrhizobium (MDH), and uncultured bacteria (MxaI, MxaJ, and MxaR). The formaldehyde-activating enzymes (FADHs) were matched with Methylosinus and Methanosarcina, whereas the FDHs were matched with Methylosinus, Bradyrhizobium, and Anaeromyxobacter (Table 2).

Biological nitrogen fixation is the process by which atmospheric N₂ gas is converted to ammonium by a nitrogenase complex containing nitrogenase reductase (NifH) and the alpha subunit (NifD) and beta subunit (NifK) of dinitrogenase. These nitrogenase proteins were detected in Methylosinus, Methylocystis, Methylocella, Bradyrhizobium, Rhodopseudomonas, Hyphomicrobium, Rhizobium, Anaeromyxobacter, Methanobacterium, and Methanosarcina (Table 2).

The total numbers of peptides detected were summarized into their respective biochemical categories (Table 3). Proteins relevant to CH₄ oxidation (MMO, MDH, FADH, and FDH; 285 of a total of 343 peptides) and N₂ fixation (NifH, NifD, and NifK; 26 of a total of 42 peptides) were detected frequently in members of the family Methylocystaceae. Notably, Methylosinus species showed both biochemical processes. The three structural components of nitrogenase (NifH, NifD, and NifK) were consistently detected in Methylosinus as well (Table 3). Therefore, our proteome analysis suggested that both CH₄ oxidation and N₂ fixation were mediated mainly through Methylocystaceae members, including Methylosinus and Methylocystis (type II methanotrophs), in rice roots. However, we could not exclude the possibility that other diazotrophs such as Bradyrhizobium, Rhodopseudomonas, and Anaeromyxobacter fix N₂ as minor contributors in rice microbiome (Table 3).

TABLE 3.

Number of peptides relevant to CH₄ oxidation and N₂ fixation in enriched microbial cells from root of paddy field-grown rice plant^a

Domain	Family	Genus	CH4 oxidation				N2 fixation
			MMO	MDH	FADH	FDH	NifH	NifD	NifK	VnfD
Bacteria
	Methylocystaceae	Methylosinus	114	32	24	11	7	8	8
	Methylocystaceae	Methylocystis	51	45	7	1		1	2
	Beijerinckiaceae	Methylocella	19						5
	Methylobacteriaceae	Methylobacterium		18
	Bradyrhizobiaceae	Bradyrhizobium		9		1		2
	Bradyrhizobiaceae	Rhodopseudomonas						1	2
	Hyphomicrobiaceae	Hyphomicrobium				1		1
	Rhizobiaceae	Rhizobium					1
	Cystobacterineae	Anaeromyxobacter				1	2
Archaea
	Methanobacteriaceae	Methanobacterium					1			1
	Methanosarcinaceae	Methanosarcina			1				1	1
	Other	Uncultured bacteria		8
Total peptides			184	112	32	15	11	13	18	2

^aBoldface entries indicate members of the family Methylocystaceae, including Methylosinus and Methylocystis. MMO, methane monooxygenase; MDH, methanol dehydrogenase; FADH, formaldehyde dehydrogenase; FDH, formate dehydrogenase; NifH, nitrogenase reductase; NifD, alpha subunit of dinitrogenase; NifK, beta subunit of dinitrogenase; VnfD, alpha subunit of vanadium dinitrogenase.

Detection of type II methanotrophs in enriched bacterial cells.

The Ma450 probe specific for Methylocystis and Methylosinus type II methanotrophs was used to detect these members of the Methylocystaceae in bacterial cells enriched from rice roots (40). This probe has often been used to detect and enumerate type II methanotrophs in natural environments (35, 40, 42). Clear signals indicated that type II methanotrophs in the enriched samples were successfully visualized by CARD-FISH (Fig. 1). Ma450-positive cells accounted for 6.6% ± 0.3% (range, 6.2% to 7.0%) of DAPI-stained cells. Note that signals were not detected from the strain used as a negative control (data not shown).

FIG 1.

Catalyzed reporter deposition-fluorescence in situ hybridization (CARD-FISH) detection of Methylocystaceae members (Methylosinus and Methylocystis) in enriched bacterial cells from rice roots. Micrographs of in situ hybridization with probe Ma450 (A), 4′,6-diamidino-2-phenylindole staining (B), and phase-contrast image (C) are shown. Scale bar = 10 μm.

Localization of type II methanotrophs in rice roots.

Both the proteome analysis and the CARD-FISH analysis of the enriched bacterial cells suggested that type II methanotrophs were present in the rice root microbiome. We applied CARD-FISH to the root tissues of rice plants during the maximum-tillering stage to localize type II methanotrophs in the rice root. Type II methanotrophic signals were clearly observed not only in and around the epidermal cells (Fig. 2A, B, C, and J) but also in the vascular cylinder (Fig. 2G, H, I, and L) of the rice roots. CARD-FISH signals were observed in and around the epithelial cells in both cross sections (Fig. 2A, B, C, and J) and vertical sections (Fig. 2D, E, F, and K) of the roots. In the vascular cylinder, CARD-FISH signals were often detected in and around the phloem around the xylem vessels (Fig. 2I and L).

FIG 2.

Catalyzed reporter deposition-fluorescence in situ hybridization (CARD-FISH) detection of Methylocystaceae members (Methylosinus and Methylocystis) in roots of field-grown rice (Oryza sativa Nipponbare) by confocal laser scanning microscopy. A, B, C, and J, cross sections of epidermis of rice root; D, E, F, and K, vertical sections of rice root; G, H, I, and L, cross sections of stele of rice roots. A, D, and G, autofluorescence of cell wall of rice root (blue); B, E, and H, Alexa Fluor 488 of Ma450 probe (green); C, F, I, J, K, and L, their overlay. Scale bars = 10 μm.

DISCUSSION

Combined analysis by using metaproteomics and CARD-FISH suggested that CH₄ oxidation and N₂ fixation were simultaneously performed mainly by type II methanotrophs (Methylocystaceae) inhabiting the vascular bundles and epidermal cells of the roots of rice plants grown in a paddy field.

In paddy ecosystems, many studies of aerobic methanotrophs in rice roots have focused on the diversity and activity of methane oxidation (11, 14, 43). However, no study has yet emphasized N₂ fixation by methanotrophs in the rice and soil microbiomes of paddy fields. Several of the methanotrophs have been reported to fix nitrogen in culture experiments (21, 22). In the present study, metaproteomics analysis indicated that nitrogenase was also expressed in field-grown rice roots, as shown by bacterial protein levels (Tables 2 and 3). Indeed, major structural components of nitrogenase (NifH, NifD, and NifK) were detected at the protein level (Table 3). In addition, type II methanotrophs were likely the major contributors to N₂ fixation (Tables 2 and 3). This supposition is partially supported by the detection of nitrogenase of type II methanotrophs in rice root-associated microbes by transcriptome analysis (44), although the authors of that paper did not discuss this aspect of their findings. Moreover, this result is also consistent with the observation that type II methanotrophs fix N₂ in CH₄-enriched soil (19).

The present report supports the hypothesis that methanotrophs are involved in N₂ fixation and CH₄ oxidation in paddy fields (17, 18). A recent observation has shown that N₂ fixation for peatland development has been mediated by methanotrophs as well (45). In a metagenomic analysis of the bacterial community associated with the roots of rice grown in a paddy field, type II methanotrophs had a relative abundance of approximately 10% in metagenomic reads and were the dominant bacterial group (24). This is consistent with the detection of approximately 7% relative abundance of type II methanotrophs by CARD-FISH in the present study (Fig. 1). However, it is difficult to compare the protein abundance of type II methanotrophs (approximately 29.7% in total proteins in Table 1) with the abundance in cell or DNA levels (approximately 7% to 10%), because metatranscriptome analysis depends on its database and detects the highly expressed proteins in microbial cell community.

Because the fluorescence intensity of the 16S rRNA probe for type II methanotrophs is too low in FISH (46), it was expected that direct visualization of methanotrophs in rice roots might be difficult by using FISH. This study therefore used the CARD-FISH approach to localize type II methanotrophs in rice roots. Signals of type II methanotrophs were visualized not only in enriched bacterial cells from the rice roots (Fig. 1) but also in epidermal cell walls and around the xylem of rice roots in situ (Fig. 2). These results suggested that type II methanotrophs inhabit the surface and interior of the root tissues as endophytes.

An early inoculation experiment detected antibody-labeled methanotrophs on the surface and in the xylem of rice roots (46). However, the present report is the first to give direct evidence that type II methanotrophs reside on the surface of, and inside, the root tissues of field-grown rice plants. This finding is supported by the fact that type II methanotrophs have often been isolated from surface-sterilized rice roots (47, 48). In addition, the abundance of type II methanotrophs is positively correlated with rice plant growth (49). Ecologically, symbiosis of rice plants with N₂-fixing methanotrophs is probably beneficial for both organisms, because methanotrophs in the root zone are able to use the CH₄ that is continuously produced from paddy soils by methanogens under anaerobic conditions (17).

Generally, endophytic bacterial cells are difficult to extract and separate from plant cells, because inside the plant tissues they are tightly attached to the host cells (50). A high relative abundance of type II methanotrophs has been observed in enriched bacterial cells from rice roots, not only by a previous metagenome analysis (24) but also by the proteomic analysis and CARD-FISH approach used in the present study. The method used here to extract and enrich bacterial cells may therefore be suitable for studies of endophytic methanotrophic bacteria in plants (36, 51). These results are in agreement with detection of type II methanotrophs as dominant in root materials by using DNA–stable-isotope probing (SIP) analysis (12).

Apart from type II methanotrophs, other proteins were matched with Methylocella, Bradyrhizobium, Methylobacterium, Rhodopseudomonas, Hyphomicrobium, Rhizobium, Anaeromyxobacter, Methanobacterium, Methanosarcina, and uncultured bacteria. The metaproteomic data (Tables 2 and 3) revealed that their functions were also related to methanol oxidation or nitrogen fixation or both. Methylocella have been isolated from acidic environments as facultative methanotrophs and have N₂-fixing activity (52, 53). A recent study has shown that Methylocella are widely distributed in natural environments and are not restricted to acidic environments (54). The proteome data from the present study suggest that Methylocella also inhabit rice roots and may contribute to CH₄ oxidation and nitrogen fixation. Methylobacterium and Bradyrhizobium have been detected in rice roots (24, 27, 28) and often possess the MDH-encoding xoxF gene for methanol oxidation (55–57). In addition, nifH has been detected in Bradyrhizobium in rice roots by metagenomic and proteomic analyses (28, 50). Thus, it is likely that Bradyrhizobium species contribute in a minor way to C1 compound metabolism and N₂ fixation. Moreover, nitrogenase proteins of Rhodopseudomonas, Hyphomicrobium, Rhizobium, and Anaeromyxobacter were detected in the present study as minor components which were also similar to those in the phyllosphere or rhizosphere of rice plants shown by previous proteomic analysis (28). Proteins of archaea were also detected from the root microbiome (Tables 1 to 3; see also Table S1 in the supplemental material), which is in agreement with a previous study (28). It is likely that Methanobacteriaceae and Methanosarcinaceae contribute not only to methane production (58) but to N₂ fixation as well.

Methane oxidation by type II methanotrophs generally occurs in aerobic environments, but anaerobic and microanaerobic conditions are generally required for N₂ fixation (59, 60). This oxygen paradox warrants a brief discussion. An early report indicated that oxygen concentration is a limiting factor for growth and nitrogenase expression of methanotrophs under N-free conditions (60). In that work, the nitrogenase activity of Methylosinus trichosporium OB3b (type II) was less sensitive to the partial pressure of O₂ than was that of the Methylococcus capsulatus Bath (type I) methanotroph. In addition, a Methylocystis (type II) isolate from soil was able to grow and fix N₂ under microaerobic conditions (<0.5% to 4% oxygen) in N-free medium (59). Moreover, in the center (i.e., the stele, including the xylem and phloem) and surface (epidermis) of the rice root, O₂ concentrations are lower than atmospheric saturation concentrations (61). Therefore, both the epidermis and the xylem of the rice root are under low-O₂ conditions, and this may enable methanotrophs to express CH₄ oxidation and N₂ fixation simultaneously. Because the O₂ concentration in the center (55% of atmospheric saturation) was higher than that in the surface (5% of atmospheric saturation) (61), it is possible that N₂ fixation mainly occurs on the surface and CH₄ oxidation in the center. To investigate the paradox further, it will be crucial in the future to characterize the biochemical activity of Methylosinus isolates from rice plants and to observe the activities of these bacteria microscopically in rice tissues by using ¹⁵N/¹³C tracers.