6 May 2021: Takeuchi M, Ozaki H, Hiraoka S, Kamagata Y, Sakata S, et al. (2021) Correction: Possible cross-feeding pathway of facultative methylotroph Methyloceanibacter caenitepidi Gela4 on methanotroph Methylocaldum marinum S8. PLOS ONE 16(5): e0251538. https://doi.org/10.1371/journal.pone.0251538 View correction
Non-methanotrophic bacteria such as methylotrophs often coexist with methane-oxidizing bacteria (methanotrophs) by cross-feeding on methane-derived carbon. Methanol has long been considered a major compound that mediates cross-feeding of methane-derived carbon. Despite the potential importance of cross-feeding in the global carbon cycle, only a few studies have actually explored metabolic responses of a bacteria when cross-feeding on a methanotroph. Recently, we isolated a novel facultative methylotroph, Methyloceanibacter caenitepidi Gela4, which grows syntrophically with the methanotroph, Methylocaldum marinum S8. To assess the potential metabolic pathways in M. caenitepidi Gela4 co-cultured with M. marinum S8, we conducted genomic analyses of the two strains, as well as RNA-Seq and chemical analyses of M. caenitepidi Gela4, both in pure culture with methanol and in co-culture with methanotrophs. Genes involved in the serine pathway were downregulated in M. caenitepidi Gela4 under co-culture conditions, and methanol was below the detection limit (< 310 nM) in both pure culture of M. marinum S8 and co-culture. In contrast, genes involved in the tricarboxylic acid cycle, as well as acetyl-CoA synthetase, were upregulated in M. caenitepidi Gela4 under co-culture conditions. Notably, a pure culture of M. marinum S8 produced acetate (< 16 μM) during growth. These results suggested that an organic compound other than methanol, possibly acetate, might be the major carbon source for M. caenitepidi Gela4 cross-fed by M. marinum S8. Co-culture of M. caenitepidi Gela4 and M. marinum S8 may represent a model system to further study methanol-independent cross-feeding from methanotrophs to non-methanotrophic bacteria.
Citation: Takeuchi M, Ozaki H, Hiraoka S, Kamagata Y, Sakata S, Yoshioka H, et al. (2019) Possible cross-feeding pathway of facultative methylotroph Methyloceanibacter caenitepidi Gela4 on methanotroph Methylocaldum marinum S8. PLoS ONE 14(3): e0213535. https://doi.org/10.1371/journal.pone.0213535
Editor: Ivan A. Berg, University of Münster, GERMANY
Received: October 12, 2018; Accepted: February 22, 2019; Published: March 14, 2019
Copyright: © 2019 Takeuchi et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The complete genome sequences of strain Gela4 and S8 are available from DDBJ/ENA/GenBank (accession number(s) GCA_000828475.1., AP017928). RNA-Seq data are available from DDBJ Sequence Read Archive (DRA) (accession number DRA006072).
Funding: This work was funded by JSPS KAKENHI (Grant Numbers JP221S0002, JP16H06279, and JP25450265). This work was also supported by the Cooperative Program (no. 005, 2013) of the Atmosphere and Ocean Research Institute, the University of Tokyo. HO was supported by the RIKEN Special Postdoctoral Researcher Program. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Microbial methane oxidation plays an important role in the global methane cycle. Aerobic methane-oxidizing bacteria (methanotrophs) are key players in aerobic and micro-aerobic environments. The development of stable isotope probing has highlighted that methane-derived carbon is incorporated not only by methanotrophs but also by non-methanotrophic bacteria (methylotrophs or others) in diverse environments, and suggests the global importance of cross-feeding interactions in methane oxidation [1, 2, 3, 4]. Methanol is synthesized in the periplasm of methanotrophs, and therefore, may diffuse out of the cell . Hence, it is generally speculated that the co-existence of methanotrophs and non-methanotrophic bacteria mainly involves the cross-feeding of methanol excreted by methanotrophs to non-methanotrophic bacteria . Methanol-dependent cross-feeding between methanotrophs and non-methanotrophic bacteria has been studied since the 1970s [6, 7]. Using isolates from lake sediments, Krause et al.  demonstrated that methanol excreted by a methanotroph is indeed the sole carbon and energy source for the co-existing obligate methylotroph that can utilize methanol or methylamine exclusively.
On the other hand, some methanotrophs are also known to excrete some organic acids . It is speculated that these organic acids are derived from dead cells. Recently, Kalyuzhnaya et al.  demonstrated that Methylomicrobium alcaliphilum of Gammaproteobacteria release organic compounds, such as formate and acetate, via fermentation-based methanotrophy. This suggested that cross-feeding of methane-derived carbon from a methanotroph to non-methanotrophic bacteria may not be mediated solely by methanol and may involve more diverse processes. However, this has not been extensively explored yet.
Recently, we obtained a methane-oxidizing enrichment culture from marine sediments of Kagoshima Bay, Japan and isolated a novel facultative methylotroph, Methyloceanibacter caenitepidi Gela4  and a novel gammaproteobacterial methanotroph, Methylocaldum marinum S8 . M. caenitepidi Gela4 grows on a wide range of substrates and can grow syntrophically with M. marinum S8 using methane added as the sole carbon source. The co-culture of M. caenitepidi Gela4 and M. marinum S8 would be a suitable model system for addressing the nature of a facultative methylotroph cross-feeding on a methanotroph.
In this study, we aimed to reveal the metabolic pathways employed by a marine facultative methylotroph cross-feeding on a methanotroph using genome analysis, RNA sequencing (RNA-Seq), and chemical analysis.
Materials and methods
Bacterial strains and culture conditions
M. caenitepidi Gela4 was grown in NMS-SP medium  with 1% (v/v) methanol added as a carbon source at 36°C. M. marinum S8 was grown in NMS-SP medium with a headspace of methane and air (20:80, v/v). For the co-culture of M. caenitepidi Gela4 and M. marinum S8, each culture was harvested by centrifugation (4,200 × g for 20 min), washed twice with NMS-SP medium and grown together in NMS-SP medium of 40 mL at 36°C in a 200 mL glass bottle under methane/air (20:80, v/v). Culture growth was monitored by measuring the optical density at 660 nm. Cells were harvested by centrifugation at 4,200 × g for 20 min. For DNA analysis, the cells were then stored at −20°C. For RNA analysis, the supernatant was discarded after centrifugation, and 1.5 mL of RNAlater (Ambion Inc., Austin, TX, USA) was added. Re-suspended cells were stored in a refrigerator overnight. The cells were then harvested and stored at −80°C before RNA extraction, as described below.
M. caenitepidi Gela4 and M. marinum S8 cells are morphologically distinct as visualized by a microscope [11, 12]. Therefore, growth of M. caenitepidi Gela4 and M. marinum S8 in co-culture was monitored by cell count via microscopy. Cells were taken periodically and counted by a haemocytometer.
Genome sequencing, assembly, and annotation
Chromosomal DNA was extracted using an ArchivePure DNA tissue kit (5 Prime, Gaithersburg, MD, USA) according to the manufacturer’s instructions. The draft genome sequence of M. caenitepidi Gela4 comprising of two scaffolds was generated previously . The remaining gap was closed by primer-walking, using primers complementary to the contig ends, at Takara Bio Inc. (Shiga, Japan). Whole-genome shotgun sequencing of M. marinum S8 was performed using PacBio RS II (Pacific Biosciences, Menlo Park, CA, USA) according to the manufacturer’s protocol. Genomes were assembled into a single contig using Sprai pipeline version 0.9.9.19 (http://zombie.cb.k.u-tokyo.ac.jp/sprai/) and manual curation. The contig was polished with Quiver .
The Microbial Genome Annotation Pipeline (MiGAP) was used to annotate sequences . Then, gene names were then assigned to genome annotations based on the best hits of blastp searches (version 2.2.30+; E-value<1e-5)  against the SwissProt database (2017_2), with several genes annotated manually. The percentage amino acid identities between the predicted sequences determined in this study were calculated using the pairwise distances program in MEGA version 7 . The Circos package  was used to generate Circular plots. Genes were assigned to Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway modules using metabolic and physiological potential evaluator (MAPLE) . Only modules with at least three assigned genes and qValue < 0.1 were retained, leading to 92 KEGG pathway modules for M. caenitepidi Gela4 (S1 Table) and 100 modules for M. marinum S8 (S2 Table).
The in-house Julia scripts were used to define the duplicated genes in each species. The Bio.jl Julia package (https://github.com/BioJulia/Bio.jl) was used to translate the annotated CDS in each genome into amino acid sequences. All-against-all blastp search was performed using the translated amino acid sequences and blast+ (version 2.2.30+) (E-value < 1e-5). A duplicated gene network, with nodes representing gene products and edges representing pairs of gene products, with E-value < 1e-5 and > 95% identity, was constructed. Genes with at least one edge in the duplicated gene network were defined as duplicated genes.
RNA extraction and sequencing
RNA-Seq was used to compare the transcriptome profiles of M. caenitepidi Gela4 grown in pure culture on methanol with cells grown in co-culture with M. marinum S8 on methane-supplemented medium. Two replicate cultures at late exponential phase were prepared for each culture condition. The NucleoSpin RNA kit (Macherey-Nagel, Düren, Germany) was used to extract total RNA. Subsequent sample processing, including rRNA subtraction, library prep, and sequencing, was performed at Takara Bio Inc. Ribosomal RNA was removed from total RNA using the Ribo-Zero rRNA removal kit (Epicenter Biotechnologies, Madison, WI, USA). An Agilent 2100 Bioanalyzer was used to verify sample quality. RNA-Seq libraries were created using the Illumina TruSeq RNA sample prep kit version 2. All cDNA samples were individually barcoded and sequenced in the same lane on the HiSeq 2500 platform. The number of cDNA sequencing reads generated per sample varied between 22.0 and 29.9 million.
Bioinformatics analysis of the transcriptome data
For mapping of RNA-Seq reads, the genome sequences of the two species were combined into two chromosomes in a FASTA file. RNA-Seq data were mapped to the genome by STAR (version 2.5.2b)  with the parameters “—outFilterMismatchNmax 3—outFilterMultimapNmax 10—outSAMprimaryFlag AllBestScore—outSAMstrandField None—alignIntronMax 3—alignIntronMin 4”. BAM files were processed using samtools 1.3 . Specifically, uniquely mapped reads were selected with the parameter “-q 255” and primary alignment reads were selected with the parameter “-F 256”. The featureCounts option in the subread package (version 1.5.1)  was used to count the number of reads mapped to each gene. For stain S8 data, the parameters “-M–fraction” were also set to calculate fractional counts for multi-mapping reads. Using simulated RNA-Seq data composed of the sequencing reads generated from the complete genome sequences of the two species, we confirmed that the aforementioned data processing strategy enabled unambiguous separation of the sequencing reads into two species (S3 and S4 Tables). Specifically, using custom Julia scripts, we generated simulated RNA-Seq data of each strain as FASTA files by extracting 100-bp length sequences of regions with at least 1-bp overlaps with genes. The sequences were prepared for each region from both forward and reverse strands. Reads were mapped to the combined genome of two strains and the primary alignment reads were selected as described above. Then, reads were counted for each CDS using the FeatureCounts with the parameters "-M—fraction". Differentially expressed genes were identified using edgeR (version 3.16.5)  by quasi-likelihood test with genewise negative binomial generalized linear models and analysed by multivariate analysis. Only genes with ≥ 10 raw read counts in at least one sample were considered for the differentially expressed gene analysis. The reads per million mapped reads (RPM) value was calculated for each gene by dividing the read count by the total number of mapped reads. The reads per kb of transcript per million mapped reads (RPKM) value was then calculated for each gene, by dividing the RPM value by CDS length (unit = kb). The data for M. caenitepidi Gela4 were based on unique mapping analysis, while the data for M. marinum S8 were based on a primary alignment because of extensive gene duplication in the genome.
Triplicate samples for chemical analysis were taken at the beginning of the incubation period and at early and late exponential phase. Samples of 2 mL were withdrawn using a syringe and filtered through a 0.22-μm pore size filter. Samples for methanol and acetaldehyde analysis were immediately sealed in a glass ampoule. Samples for organic acids and ammonium were stored in a plastic tube. All of the samples were stored at −20°C before analysis. Methanol and acetaldehyde concentrations were analysed by the Agilent 7697A Headspace Sampler coupled to the Agilent 6890N Gas Chromatograph/MSD system with selected ion monitoring mode. One millilitre of the sample was added to a 5 mL headspace vial, heated at 80°C for 5 min, and 3 mL of headspace vapour was injected into a split injector of the gas chromatograph. The DB-WAX capillary column (60 m × 0.25 mm × 0.50 μm, Agilent J&W Scientific, CA, USA) was used as a GC column. Helium was the carrier gas at a flow rate of 1.0 mL/min. The detection limits for methanol and acetaldehyde were 310 and 570 nmol/L, respectively. Concentrations of organic acids were analysed via HPLC (Shimadzu Corp., Japan) with a conductivity detector. Shim-pack SCR-102H (300 mm × 8 mm, Shimadzu model, Shimadzu Corp., Japan) was used as a column to separate organic acids eluted in 5 mmol/L of p-toluenesulfonic acid at 40°C. The detection limits were 1 μmol/L for formate and acetate and 10 μmol/L for propionate and butyrate.
Real-time quantitative reverse transcription PCR
To support the RNA-Seq analysis results, expression levels of acetyl-CoA synthetase gene (acsA), hydroxypyruvate reductase gene (hpr), and aldehyde dehydrogenase gene (aldA) were quantified by real-time quantitative reverse transcription PCR assay. Co-cultures was incubated in triplicate and harvested at early and late exponential phases. M. caenitepidi Gela4 cells grown with methanol or sodium acetate (0.3%) were collected at late exponential phase and also analysed as reference. Total RNA was extracted as described above. For cDNA synthesis, 400–500 ng of RNA was reverse-transcribed using the PrimeScript RT reagent kit (Takara Bio, Japan) according to the manufacturer’s instruction. Primers specific for acsA, hpr and aldA genes of M. caenitepidi Gela4 were designed using Primer3 . The primers for acsA gene were GelaACS-F (5ʹ - GAACTCGACCTTTGCGTATCCG-3ʹ) and GelaACS-R (5ʹ - TGCAGTTCGCCGACACGTTAAG-3ʹ). The primers for hpr gene were GelaHPR-F (5ʹ - GATGCCGAACTTCATCGTCACG-3ʹ) and GelaHPR-R (5ʹ - GTTGCCCGCAACGAACGCATC-3ʹ). The primers for aldA gene were Gald-F (5ʹ - CGACACCGTTGCCTATCATTTC-3ʹ) and Gald-R (5ʹ - TCCAGCACGCCATCAAAATC-3ʹ). PCR amplification was not detected with these three primer sets when M. marinum S8 genomic DNA was used as a template. Real-time PCR was performed with TB Green Premix Ex Taq II (Takara Bio, Japan) using a Thermal Cycler Dice Real Time System II (Takara Bio, Japan). Each real-time PCR was carried out in a 15-μl total volume reactions that contained cDNA (1μl), Mastermix, forward and reverse primers (0.4 μM each) and nuclease-free sterile water. The thermal cycler program consisted of 40 cycles of denaturation (95°C for 5s) and annealing and extension (55°C for 30s) after an initial denaturation at 95°C for 30s. To construct calibration standard curves, dilution series of positive control DNA were used. For positive control DNA, gene fragments were amplified with the primers described above from M. caenitepidi Gela4 genomic DNA. PCR products were then separated by gel electrophoresis and extracted with a NucleoSpin Gel and PCR Cleanup kit (Macherey-Nagel, Düren, Germany).
Complete genome of M. caenitepidi Gela4 and possible metabolic pathways
We generated the complete genome of M. caenitepidi Gela4 by closing the gaps from the previously published draft genome . The M. caenitepidi Gela4 genome was obtained at a sequencing coverage of 76×. It consisted of a single circular chromosome of 3,424,964 bp, with a G+C content of 62.8% (Fig 1A). The genome was found to contain one ribosomal RNA operon (16S–23S–5S), 45 tRNA genes and 3,342 coding sequences (CDSs) (S5 Table).
Circular diagrams of the (A) M. caenitepidi Gela4 and (B) M. marinum S8 genomes. Circles (outside to inside) depict (1) contigs; (2) gene expression levels in pure culture, log10 (RPKM + 1) (only for M. caenitepidi Gela4); (3) gene expression levels in co-culture, log10 (RPKM + 1); (4) coding genes on the leading strand (red); (5) coding genes on the lagging strand (blue); (6) 16S rRNA genes (green); (7) tRNA genes (blue); (8) transposase genes (grey); (9) duplicated genes (E-value<1e-5, % identity >95.0) (red); (10) the GC content (1-kb sliding window); and (11) the GC skew (1-kb sliding window).
In M. caenitepidi Gela4, pathways for oxidization of methanol to formate were identified. The gene clusters responsible for methanol oxidation (mxaFJGIRSACKLD, GL4_0421–0431) and the synthesis of the cofactor pyrroloquinoline quinone (PQQ; pqqBCDE) were identified (GL4_3155, 2349–2351). In addition, four genes encoding a methanol dehydrogenase large subunit similar to xoxF were found (GL4_0437, 1130, 1360, and 1361). The amino acid sequences of four XoxF proteins shared 70%–94% identity with each other. All genes coding for enzymes involved in tetrahydromethanopterin (H4MPT)-mediated (fae, mtdB, mch and fhcD; GL4_1783 and 3256, 1788, 1786, 1794) and tetrahydrofolate (H4F)-mediated (mtdA, fch and ftfL; GL4_3376, 3375 and 3380) formaldehyde oxidation were identified. Genes for two formate dehydrogenases (FDH), NAD-dependent-FDH (fdhA1, B1, C1, D1, and E1; GL4_0089, 1367–1370), and tungsten-containing FDH (fdhA2 and B2; GL4_3220 and 3221) were also identified.
Methylotrophs assimilate carbon from methanol by using either the ribulose monophosphate (RuMP) pathway or the serine cycle . In M. caenitepidi Gela4, all genes necessary for the serine pathway, i.e. serine hydroxymethyltransferase (glyA; GL4_1969), serine glyoxylate aminotransferase (sga; GL4_3378), hydroxypyruvate reductase (hpr; GL4_3377), d-glycerate 2-kinase (gckA; GL4_3381), enolase (eno; GL4_2130), phosphoenolpyruvate carboxylase (ppc; GL4_3372), malate dehydrogenase (mdh; GL4_0105), malate-CoA ligase (mtk; GL4_3373 and 3374) and malyl-CoA lyase (mcl; GL4_3371) were identified. In contrast, genes for the RuMP pathway, i.e. 3-hexulose-6-P synthase (hps) and 6-phospho-3-hexuloisomerase (phi), were not detected. The gene for isocitrate lyase was not found, but genes for the ethylmalonyl-CoA (EMC) pathway were identified as a glyoxylate regeneration pathway. These gene composition patterns indicated that M. caenitepidi Gela4 assimilates carbon from methanol through the serine pathway. Moreover, genes for glycolysis and the tricarboxylic acid (TCA) cycle were also identified.
Complete genome of M. marinum S8 and potential metabolic pathways
We generated the complete genome of M. marinum S8 obtained at 102 × sequencing coverage using PacBio RS II. The genome was found to consist of a single circular chromosome of 6,088,916 bp, with a G + C content of 58.7% (Fig 1B). The genome contains 48 tRNA genes and 5,611 CDSs (Fig 1B and S6 Table). Notably, it possessed a high number of transposase genes and duplicated genes (Fig 1B); 9.5% (539) of genes were identified as duplicate when the threshold of amino acid sequence identity was set at >95% (S1 Fig, S7 Table). In addition, the genome contains two ribosomal operons (16S–23S–5S).
Among the duplicated genes were key genes involved in methane oxidization. Two sets of particulate methane monooxygenase operons (pmoABC; sS8_1787–1789 and 3975–3977) were found; pmoA and pmoC genes were identical, whereas the predicted pmoB gene products showed 99.5% amino acid identity. In addition, the genome was found to contain three additional pmoC genes (sS8_2529, 3418 and 4876), whose predicted products shared a 60%–91% amino acid identity with the PmoC protein encoded by the pmo operon.
Four genes predicted to encode the large subunit of methanol dehydrogenase were also found. One of them is located in a gene cluster encoding methanol dehydrogenase (MDH) and accessory proteins (mxaFJGIRSACKLD; sS8_1426–1436). The other three are similar to xoxF (sS8_0935, 1979 and 2523). Four genes for PQQ biosynthesis were found in the same cluster (pqqBCDE; sS8_2216–2219). All genes coding for enzymes involved in H4MPT-mediated and H4F-mediated formaldehyde oxidation were also identified. M. marinum S8 genome contains genes encoding the complete RuMP pathway, i.e. hps (sS8_1135 and 1148) and phi (sS8_1134) (S6 and S8 Tables). For the cleavage part of the RuMP pathway, genes for both the Entner-Doudoroff pathway (sS8_4013, 4180, 4181, 4179, 1800) and Embden-Meyerhof-Parmas pathway (sS8_4382, 1130) were identified. Among genes for the serine pathway, M. marinum S8 genome contains serine hydroxymethyltransferase (sS8_0137), serine glyoxylate aminotransferase (sS8_1474), hydroxypyruvate reductase (sS8_1475), D-glycerate-3-kinase (sS8_1476), enolase (sS8_1801), malate dehydrogenase (sS8_4317) and malyl-CoA lyase (sS8_2223) but lacks other genes. The M. marinum S8 genome contained all Calvin-Benson-Bassham cycle genes, such as ribulose-1,5- bisphosphate carboxylase (sS8_2432), phosphoglycerate kinase (sS8_4941), glyceraldehyde-3-phosphate dehydrogenase (sS8_1782), triosephosphate isomerase (sS8_0238), fructose-1,6-bisphosphate aldolase (sS8_1130), transketolase (sS8_1131), ribulose-phosphate 3-epimerase (sS8_ 2460) and phosphoribulokinase (sS8_4630). NAD-dependent FDH (fdhA1, B1, C1, D1 and E1; sS8_0312–0316) and tungsten-containing FDH (fdhA2 and B2–sS8_2524 and 2525) were also found. The genome was also found to contain genes for the TCA cycle.
RNA-Seq of M. caenitepidi Gela4 in pure culture and co-culture with the methanotroph
To reveal the effects of syntrophic growth with a methanotroph on the gene expression profiles of M. caenitepidi Gela4, RNA-Seq was conducted on M. caenitepidi Gela4 cells grown on methanol in pure culture and to those grown in a co-culture with the methanotroph M. marinum S8 in a medium supplemented with methane (S2 Fig). The specific growth rates of M. caenitepidi Gela4 in pure culture and the co-culture were 0.03 and 0.02 (h−1), respectively. Fig 2 presents (A) growth of the co-culture and (B) species composition at three different growth phases. Percentage of M. marinum S8 in co-culture increased from 21% to 77% during growth. The complete genome sequences of the two species enabled unambiguous separation of the sequenced reads between the two species in the co-culture samples (S3 and S4 Tables). When M. caenitepidi Gela4 was grown in co-culture with M. marinum S8, 24% of the uniquely mapped transcriptome data originated from M. caenitepidi Gela4 and 76% originated from M. marinum S8 on average. Pearson correlation coefficients of the expression levels between replicates were high in both pure culture (0.998) and co-culture (0.992) (S3 Fig), confirming high reproducibility of the RNA-Seq data. Full gene expression datasets for M. caenitepidi Gela4 and M. marinum S8 are presented in S5 Table and S6 Table, respectively.
(A) Growth of the co-culture and (B) species composition at early, middle and late exponential growth phase. The gray bar and white bars represent number of M. marinum S8 cells and M. caenitepidi Gela4 cells, respectively. Error bars represent standard deviation of triplicates.
Expression of the central metabolism genes in M. caenitepidi Gela4 for both culture conditions, presented as RPKM are shown in Table 1. The expression of central metabolism genes in M. marinum S8 in co-culture with M. caenitepidi Gela4, presented as RPKM, is shown in S8 Table. Fig 3 shows the central carbon metabolism of M. caenitepidi Gela4 and genes that were significantly up- or downregulated.
Red and blue arrows indicate genes that were upregulated (log2FC > 0.3, FDR < 0.05) and downregulated (log2FC < −0.3, FDR < 0.05) in co-culture.
Differential gene expression analysis revealed that 689 genes of M. caenitepidi Gela4 were significantly upregulated [log2FC (fold-change) > 0.30, false discovery rate (FDR) < 0.05], and 764 genes were downregulated in co-culture when compared with pure culture (log2FC < −0.30, FDR < 0.05; Fig 4). To identify pathway modules enriched for differentially expressed genes, genes in M. caenitepidi Gela4 were assigned to KEGG pathway modules using metabolic and physiological potential evaluator , and gene set enrichment analysis (GSEA)  was performed using the pathway module annotations. Despite the low number of genes assigned to KEGG pathway modules (656 of 3,342 coding genes), 5 and 7 modules were significantly upregulated and downregulated, respectively (Fig 5, FDR < 0.01). Among the upregulated pathway modules were ‘Citrate cycle, second carbon oxidation, 2-oxoglutarate = > oxaloacetate’, and ‘Citrate cycle (TCA cycle, Krebs cycle)’, whereas among the downregulated modules were ‘Formaldehyde assimilation, serine pathway’ and ‘Cytochrome c oxidase, cbb3-type’. We also performed the hypergeometric test for association of KEGG pathway modules to up- or downregulated genes, which yielded similar results (S4 Fig, FDR < 0.05).
Red and blue points represent genes whose expression was upregulated (log2FC > 0.30, FDR < 0.05) and downregulated (log2FC < −0.30, FDR < 0.05), respectively.
Down-regulation of methanol assimilation pathways in M. caenitepidi Gela4 in co-culture
In M. marinum S8, genes involved in methane oxidization, including the methane monooxygenase (pmo operon) genes, methanol dehydrogenase (mxaFJ), hexulose-6-phosphate synthase, NAD-dependent formate dehydrogenase, and fae, showed high expression (S8 Table). This suggested that methane oxidization occurs in M. marinum S8 as expected during the co-culture.
In M. caenitepidi Gela4, genes encoding enzymes for the transformation of methanol to formaldehyde (mxa operon) and formaldehyde oxidation (fae) were among the most highly expressed in both pure culture and co-culture, and no significant changes in the expression levels of mxaF and mxaJ genes were observed (false discovery rate (FDR) > 0.05) (Table 1). This implied that methanol is a primary carbon source of the methylotroph in co-culture with the methanotroph. However, surprisingly, methanol was below the detection limit in both pure culture of the methanotroph or co-culture (Fig 6A). Considering that methanol dehydrogenase subunit 1 precursor is a major protein in both acetate-grown and methanol-grown Methylobacter extorquens AM1 , high expression of mxa operon genes and fae may not necessarily imply growth on methanol.
Concentrations of (A) methanol, (B) acetate, and (C) formate in culture supernatants taken at the beginning of the incubation, and early and exponential phase of M. caenitepidi Gela4, M. marinum S8, and co-culture. ND: not detected. Error bars represent standard deviation of triplicate experiments. Culture supernatants of M. caenitepidi Gela4 at the beginning of the incubation and early exponential phase was not analysed due to high concentration (1% in the liquid phase at the beginning) for the analysis.
In line with undetectable concentration of methanol, many genes in pathways involved in carbon assimilation through formaldehyde (the serine pathway (sga, hpr, ppc and mtk), and dissimilation (H4MPT pathway and H4F pathway) were significantly downregulated in M. caenitepidi Gela4 grown in co-culture compared to that in pure culture (Fig 3). In addition, ‘Formaldehyde assimilation, serine pathway’ was among the downregulated KEGG modules (Fig 5). These results highlighted a downregulation of methanol utilization in M. caenitepidi Gela4 in co-culture compared to that in pure culture.
Up-regulation of TCA cycle in M. caenitepidi Gela4 in co-culture
Next, we focussed on additional genes in central carbon metabolism pathways that were found to be upregulated in M. caenitepidi Gela4 in co-culture. The most highly upregulated gene was aldehyde dehydrogenase (aldA; GL4_0813; log2FC 4.83). The gene encoding the acetate oxidation enzyme, acetyl-CoA synthetase (acsA, GL4_0367), was also significantly upregulated (log2FC 3.57, Table 1), which would lead to increased acetyl-CoA synthesis. The gene encoding the proton-translocating pyrophosphatase (hppa) (GL4_1976), which is involved in the cleavage of pyrophosphate, the by-product of acetyl-CoA synthetase, was also upregulated.
Some genes in the EMC pathway, which converts acetyl-CoA to glyoxylate and succinyl-CoA, were upregulated in M. caenitepidi Gela4 in co-culture. Genes encoding acetoacetyl-CoA reductase (phaB; GL4_0015), 3-hydroxybutyryl-CoA dehydrogenase (hbdA; GL4_3328), crotonyl-CoA carboxylase/reductase (ccr; GL4_3337), ethylmalonyl-CoA mutase (ecm; GL4_3338) and l-malyl-CoA/beta-methylmalyl-CoA lyase (mcl2; GL4_0468)  were significantly upregulated (Fig 3, Table 1).
In the TCA cycle, genes for citrate synthase (citA, GL4_2103), malate dehydrogenase (mdh, GL4_0105), succinyl-CoA synthethase (sucC and sucD; GL4_0106 and GL4_0107) and 2-oxoglutarate dehydrogenase (sucA and sucB; GL4_0108 and GL4_0109) were upregulated (Fig 3, Table 1). The TCA cycle was also found to be upregulated according to the pathway enrichment analysis (Fig 5 and S4 Fig).
Importantly, many of the above genes that were differentially expressed in M. caenitepidi Gela4 were consistent with the expression profile of the facultative methylotroph Methylobacterium extorquens AM1 grown on acetate as compared with growth on methanol . Specifically, upregulation was observed in the genes involved in the TCA cycle. Furthermore, genes encoding the proton-translocating pyrophosphatase (hppa) (GL4_1976) and the NAD(P) transhydrogenase (GL4_3073–3075), which is used to balance redox equivalents in the cell, were upregulated in both M. extorquens AM1 grown on acetate and M. caenitepidi Gela4 in co-culture. PEP carboxykinase (GL4_0586), which is operating in an opposing fashion to the serine cycle PEP carboxylase, was also upregulated in both cases. In addition, downregulation of NAD-dependent FDH (GL4_0089, 1367 and 1368) and upregulation of tungsten-containing FDH (GL4_3220 and 3221) was observed in both cases. Collectively, these results suggested that carbon assimilation via the EMC pathway and energy generation via the TCA cycle also occur in M. caenitepidi Gela4 co-cultured with M. marinum S8, consistent with observations using 13C steady-state metabolic flux analysis in M. extorquens AM1 grown on acetate .
M. marinum S8 excretes acetate and formate
To search for possible compounds cross-fed by the methanotroph to complement our RNA-Seq analyses, we measured the concentrations of organic acids in the culture supernatants. Acetate was produced at <15.6 μmol/L in the pure culture of M. marinum S8, while that in the co-culture and M. caenitepidi Gela4 was <12.4 and <1.9 μmol/L, respectively (Fig 6B). Formate was detected at a concentration of <26.2, <8.4, and <12.4 μmol/L in culture supernatants of M. caenitepidi Gela4, M. marinum S8, and the co-culture (Fig 6C). Acetaldehyde was below the detection limit in most of the culture supernatants, but a trace amount (590 nmol/L) was detected in one of the triplicate samples from the late exponential phase of M. marinum S8. Propionate was below the detection limit in all of the culture supernatants.
Quantitative analyses of acsA, hpr, and aldA genes
RNA-Seq analysis suggested that M. caenitepidi Gela4 utilizes compounds other than methanol, possibly acetate, in co-culture. To further investigate and to monitor if the metabolic pathway changes with growth phase, real-time quantitative reverse transcription PCR was performed for acsA, hpr of the serine pathway and aldA genes of M. caenitepidi Gela4 in co-culture at early and late exponential growth phases. Pure cultures of M. caenitepidi Gela4 grown with methanol and acetate were also analysed as references. Student t-test revealed that the ratio of acsA and hpr expression patterns in M. caenitepidi Gela4 were higher in cells grown with acetate and in cells in co-cultured cells at late exponential phase than in cells grown with methanol (p < 0.05) (Table 2). Similar results were obtained for the ratio of aldA and hpr.
Divergent response of xoxF genes in M. caenitepidi Gela4 in co-culture
The M. caenitepidi Gela4 genome contains four xoxF genes that encode methanol dehydrogenases. XoxF is a protein that has gained recent attention regarding methylotrophy , and shares about 50% amino acid identity with MxaF . The protein has been shown to be involved in methanol oxidation  and regulation of methanol dehydrogenase genes . Among the xoxF genes, the expression of GL4_1360 and 1361 was significantly downregulated, whereas the expression of GL4_0437 and GL4_1130 was upregulated with log2FC values of 0.29 in co-culture (S5 Fig, Table 1). These results highlighted the divergent response of xoxF genes in M. caenitepidi Gela4 upon co-culture; this was in contrast to mxaF, which remained highly expressed.
In the current study, we compared gene expression of the facultative methylotroph M. caenitepidi Gela4 grown in pure culture with methanol as the sole carbon source or co-cultured with the methanotroph M. marinum S8 with methane as the sole carbon source. The cross-feeding compound from methanotrophs to non-methanotrophic bacteria has been often considered as methanol excreted by methanotrophs. Krause et al. (2017)  has examined this interrelationship in detail. The authors conducted transcriptome analysis of two obligate methylotroph strains in co-culture with the methanotroph Methylobacter tundripaludum and demonstrated that the obligate methylotroph relies solely on methanol in co-culture with M. tundripaludum. In the current study, we examined for the first time the response of a facultative methylotroph, M. caenitepidi Gela4, which can assimilate a wide range of substrates in addition to methanol, such as acetate, ethanol, formate and succinate , grown in co-culture with a methanotroph by transcriptome analysis.
We discovered that the expression levels of nearly half of the genes of M. caenitepidi Gela4 were altered during co-culture (Fig 4, Table 1). Of note, genes for acetyl-CoA synthesis, the EMC pathway and TCA cycle were upregulated during co-culture cultures (Fig 3, Table 1). Considering that acetyl-CoA is a starting compound of the EMC pathway and TCA cycle, compounds that can be used for the acetyl-CoA synthesis are cross-fed from M. marinum S8 to M. caenitepidi Gela4.
One possible source of acetyl-CoA in co-culture is acetate. Indeed, expression of acetyl-coenzyme A synthetase increased with a log2FC value of 3.57. Some methanotrophs are known to produce organic acids, e.g. formate and acetate [10, 32]. For example, gammaproteobacterial methanotroph, Methylomicrobium alcaliphilum, has been shown to produce organic acids under oxygen-limited conditions . The pure culture of M. marinum S8 produced up to 16 μmol/L of acetate during growth (Fig 4B). Pyruvate dehydrogenase (sS8_5306) and acetate kinase (sS8_2955) were found in the M. marinum S8 genome, suggesting that M. marinum S8 is also capable of ‘fermentation-based methanotrophy’ as M. alcaliphilum. These results suggested that acetate is a carbon source for M. caenitepidi Gela4 in co-culture. Increased relative expression level of acsA to hpr in M. caenitepidi Gela4 grown with acetate as well as in co-culture was confirmed by real-time PCR analysis (Table 2), supporting that M. caenitepidi Gela4 utilizes acetate in co-culture. Two genes (GL4_0878 and 2347) were found to encode proteins homologous to acetate transporter of M. extorquens (ActP) with 26–27% amino acid identities. However, there was no significant difference in the expression levels of these two genes between pure culture and co-culture conditions. Functional acetate transporter in M. caenitepidi Gela4 should be identified in future studies.
Another possible source of acetyl-CoA is acetaldehyde. Indeed, the most highly upregulated gene of M. caenitepidi Gela4 co-cultured with M. marinum S8 was that encoding the aldehyde dehydrogenase gene. However, relatively high expression of aldA was observed not only in co-culture but also in M. caenitepidi Gela4 grown with acetate (Table 2), suggesting that aldA is involved in acetate metabolism. Substrate specificity of aldA in M. caenitepidi Gela4 and its role in acetate metabolism should be investigated in future studies.
While formate was also produced in M. marinum S8 at 8 μmol/L, it was not considered as a cross-feeding compound as M. caenitepidi Gela4 produced a higher concentration of formate (23 μmol/L).
Our results demonstrated that methanol assimilation pathways were downregulated in M. caenitepidi Gela4 in co-culture conditions (Fig 3). It is possible that low concentrations with a high turnover would be sufficient to maintain cross-feeding. However, the concentration of methanol, both in pure culture of M. marinum S8 and co-culture, was below the detection limit (< 310 nmol/L), which is more than 1,000 times lower than the previous report on M. tundripaludum cross-feeding methanol (1.24 mmol/L) . Surprisingly, the methanol concentration in the culture supernatant of M. marinum S8 was also much lower than the values reported previously for other methanotrophic species (0.01–1.00 mmol/L) . Considering that a higher concentration of acetate (16 μmol/L) was produced by M. marinum S8, methanol may not be a primary carbon source for M. caenitepidi Gela4 in co-culture with M. marinum S8, if any, in contrast to the previous study on an obligate methylotroph .
M. alcaliphilum produces acetate under oxygen-limited conditions . In this study, relative expression of acsA of M. caenitepidi Gela4 in co-culture was also high in the early exponential phase (Table 2), which is not considered to be an oxygen-limited condition, although the difference was not significant by t-test. Methanol was not detected, but 8.5 μmol/L of acetate was detected in early exponential phase of M. marinum S8, suggesting that M. caenitepidi Gela4 may utilize acetate not only in the late exponential phase. Further study is required to elucidate the mechanism of acetate production and effects of environmental conditions such as oxygen concentration on acetate excretion by M. marinum S8.
XoxF is a rare-earth element-dependent methanol dehydrogenase . Although we did not supplement the growth medium with rare-earth elements (Inductively coupled plasma-mass spectrometry analysis confirmed that the concentration of La, Ce, Nd, and Pr in the growth medium was < 1 μg/kg), M. caenitepidi Gela4 expressed three xoxF genes at relatively high RPKMs, and their expression was altered in co-culture with the methanotroph (S5 Fig). This was in contrast to mxaF genes that remained highly expressed. The expression of xoxF genes of Methylotenera mobilis 13 was also altered in co-culture . Although further studies are required to elucidate the functions of these genes, the results suggested that xoxF genes may be involved in the response to a co-culture conditions, such as, low methanol concentration.
It has long been considered that the cross-feeding compound between methanotrophic and non-methanotrophic bacteria is primarily methanol. Our results presented here suggest that this is not always the case. Organic acids such as acetate can also be an important cross-feeding compound depending on the species and perhaps on environmental conditions. M. marinum S8 and M. caenitepidi Gela4 is a good co-culture model to further study methanol-independent cross-feeding from methanotrophs to non-methanotrophs.
S1 Fig. Number of duplicated genes in the M. marinum S8 genome.
S2 Fig. Schematic showing the transcriptome experiment.
S3 Fig. Pairwise comparison of gene expression levels for M. caenitepidi Gela4 cells in pure culture and co-culture.
The results of hypergeometric tests for overlaps of KEGG modules with (a) upregulated or (b) downregulated genes. The modules with FDR < 0.05 and > = 3 annotated M. caenitepidi Gela4 genes are shown according to the value of log10 (FDR).
S5 Fig. Heat maps showing the expression levels of four xoxF genes of M. caenitepidi Gela4 in pure culture and co-culture with M. marinum S8.
Scaled RPKM was calculated as a z-score of RPKM values for each gene per culture.
S1 Table. KEGG pathway modules detected for M. caenitepidi Gela4.
S2 Table. KEGG pathway modules detected for M. marinum S8.
S3 Table. Evaluation of cross mapping between M. caenitepidi Gela4 and M. marinum S8.
S5 Table. Transcriptomics data of M. caenitepidi Gela4 grown in pure culture with methanol as well as in co-culture with M. marinum and methane.
S6 Table. Transcriptomics data of M. marinum S8 grown in co-culture with M. caenitepidi Gela4 and methane.
S7 Table. Duplicated genes identified in the M. caenitepidi Gela4 genome (upper table) and M. marinum S8 genome (lower table) using 95% identity as a threshold.
The PacBio sequencing was kindly performed by TOMY Digital Biology Co., Ltd. (Tokyo, Japan). Some of the computations were performed using the NIG supercomputer at the ROIS National Institute of Genetics. We thank Jianqiang Sun for comments on RNA-seq analysis. We also acknowledge Dr. M. Zhang, AIST for assistance with the inductively coupled plasma-mass spectrometry analysis.
- 1. Jensen S, Neufeld JD, Birkeland NK, Hovland M, Murrell JC. Methane assimilation and trophic interactions with marine Methylomicrobium in deep-water coral reef sediment off the coast of Norway. FEMS Microbiol Ecol. 2008;66: 320–330. pmid:18811651
- 2. Paul BG, Ding H, Bagby SC, Kellermann MY, Redmond MC, Andersen GL, et al. Methane-oxidizing bacteria shunt carbon to microbial mats at a marine hydrocarbon seep. Front Microbiol. 2017;8: 186. pmid:28289403
- 3. Mayumi D, Yoshimoto T, Uchiyama H, Nomura N, Nakajima-Kambe T. Seasonal change in methanotrophic diversity and populations in a rice field soil assessed by DNA-stable isotope probing and quantitative real-time PCR. Microbes Environ. 2010;25: 156–163. pmid:21576868
- 4. He R, Wooller MJ, Pohlman JW, Catranis C, Quensen J, Tiedje JM, et al. Identification of functionally active aerobic methanotrophs in sediments from an arctic lake using stable isotope probing. Environ Microbiol. 2012;14: 1403–1419. pmid:22429394
- 5. Corder RE, Johnson ER, Vega JL, Clausen EC, Gaddy JL. Biological production of methanol from methane. Abs Pap Am Chem Soc. 1986;196: 231–234.
- 6. Wilkinson TG, Topiwala HH, Hamer G. Interactions in a mixed bacterial population growing on methane in continuous culture. Biotechnol Bioeng. 1974;16: 41–59. pmid:4813164
- 7. Lamb SC, Garver JC. Interspecific interactions in a methane-utilizing mixed culture. Biotechnol Bioeng. 1980;22: 2119–2135. pmid:29345770
- 8. Krause SM, Johnson T, Samadhi Karunaratne Y, Fu Y, Beck DA, Chistoserdova L, et al. Lanthanide-dependent cross-feeding of methane-derived carbon is linked by microbial community interactions. Proc Natl Acad Sci USA. 2017;114: 358–363. pmid:28028242
- 9. Bothe H, Jensen KM, Mergel A, Larsen J, Jørgensen C, Bothe H, et al. Heterotrophic bacteria growing in association with Methylococcus capsulatus (Bath) in a single cell protein production process. Appl Microbiol Biotechnol. 2002;59: 33–39. pmid:12073128
- 10. Kalyuzhnaya MG, Yang S, Rozova ON, Smalley NE, Clubb J, Lamb A, et al. Highly efficient methane biocatalysis revealed in a methanotrophic bacterium. Nat Commun. 2013;4: 2785. pmid:24302011
- 11. Takeuchi M, Katayama T, Yamagishi T, Hanada S, Tamaki H, Kamagata Y, et al. Methyloceanibacter caenitepidi gen. nov., sp. nov., a facultatively methylotrophic bacterium isolated from marine sediments near a hydrothermal vent. Int J Syst Evol Microbiol. 2014;64: 462–468. pmid:24096357
- 12. Takeuchi M, Kamagata Y, Oshima K, Hanada S, Tamaki H, Marumo K, et al. Methylocaldum marinum sp. nov., a thermotolerant, methane-oxidizing bacterium isolated from marine sediments, and emended description of the genus Methylocaldum. Int J Syst Evol Microbiol. 2014;64: 3240–3246. pmid:24981325
- 13. Chin CS, Alexander DH, Marks P, Klammer AA, Drake J, Heiner C, et al. Nonhybrid, finished microbial genome assemblies from long-read SMRT sequencing data. Nat Methods. 2013;10: 563–569. pmid:23644548
- 14. Sugawara H, Ohyama A, Mori H, Kurokawa K. Microbial Genome Annotation Pipeline (MiGAP) for diverse users. 20th International Conference on Genome Informatics, Yokohama, Japan. 2009; S001-1–2.
- 15. Camacho C, Coulouris G, Avagyan V, Ma N, Papadopoulos J, Bealer K, et al. BLAST+: architecture and applications. BMC Bioinformatics. 2009;10: 421. pmid:20003500
- 16. Kumar S, Stecher G, Tamura K. MEGA7: Molecular Evolutionary Genetics Analysis Version 7.0 for Bigger Datasets. Mol Biol Evol. 2016;33: 1870–1874. pmid:27004904
- 17. Krzywinski M, Schein J, Birol I, Connors J, Gascoyne R, Horsman D, et al. Circos: An information aesthetic for comparative genomics. Genome Res. 2009;19: 1639–1645. pmid:19541911
- 18. Takami H, Taniguchi T, Arai W, Takemoto K, Moriya Y, Goto S. An automated system for evaluation of the potential functionome: MAPLE version 2.1.0. DNA Res. 2016;23: 467–475.
- 19. Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, et al. STAR: Ultrafast universal RNA-seq aligner. Bioinformatics. 2013;29: 15–21. pmid:23104886
- 20. Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics. 2009;25: 2078–2079. pmid:19505943
- 21. Liao Y, Smyth GK, Shi W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics. 2014;30: 923–930. pmid:24227677
- 22. McCarthy DJ, Chen Y, Smyth GK. Differential expression analysis of multifactor RNA-Seq experiments with respect to biological variation. Nucleic Acids Res. 2012;40: 4288–4297. pmid:22287627
- 23. Rozen S, Skaletsky HJ. primer3 on the WWW for general users and for biologist programmers. In: Krawetz S, Misener S, editors. Bioinformatics Methods and Protocols: Methods in Molecular Biology. Humana Press, Totowa, NJ; 2000. pp. 365–386.
- 24. Hanson RS, Hanson TE. Methanotrophic bacteria. Microbiol Rev. 1996;60: 439–471. pmid:8801441
- 25. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci. 2005;102: 15545–50. pmid:16199517
- 26. Schneider K, Peyraud R, Kiefer P, Christen P, Delmotte N, Massou S, et al. The ethylmalonyl-CoA pathway is used in place of the glyoxylate cycle by Methylobacterium extorquens AM1 during growth on acetate. J Biol Chem. 2012;287: 757–766. pmid:22105076
- 27. Erb TJ, Frerichs-Revermann L, Fuchs G, Alber BE. The apparent malate synthase activity of Rhodobacter sphaeroides is due to two paralogous enzymes, (3S)-malyl-coenzyme A (CoA)/β-methylmalyl-CoA lyase and (3S)-malyl-CoA thioesterase. J Bacteriol. 2010;192: 1249–1258. pmid:20047909
- 28. del Rocío Bustillos-Cristales M, Corona-Gutierrez I, Castañeda-Lucio M, Águila-Zempoaltécatl C, Seynos-García E, Hernández-Lucas I, et al. Culturable Facultative Methylotrophic Bacteria from the Cactus Neobuxbaumia macrocephala Possess the Locus xoxF and Consume Methanol in the Presence of Ce3+ and Ca2+. Microbes Environ. 2017;32: 244–251. pmid:28855445
- 29. Chistoserdova L, Lidstrom ME. Molecular and mutational analysis of a DNA region separating two methylotrophy gene clusters in Methylobacterium extorquens AM1. Microbiology. 1997;143: 1729–1736. pmid:9168622
- 30. Nakagawa T, Mitsui R, Tani A, Sasa K, Tashiro S, Iwama T, et al. A catalytic role of XoxF1 as La3+-dependent methanol dehydrogenase in Methylobacterium extorquens Strain AM1. PLoS One. 2012;7: e50480. pmid:23209751
- 31. Firsova YE, Torgonskaya ML, Trotsenko YA. Functionality of the xoxF gene in Methylobacterium dichloromethanicum DM4. Microbiology. 2015;84: 796–803.
- 32. Vecherskaya M, Dijkema C, Saad HR, Stams AJ. Microaerobic and anaerobic metabolism of a Methylocystis parvus strain isolated from a denitrifying bioreactor. Environ Microbiol Rep. 2009;1: 442–449. pmid:23765898
- 33. Patel SKS, Mardina P, Kim SY, Lee JK, Kim IW. Biological methanol production by a Type II methanotroph Methylocystis bryophila. J Microbiol Biotechnol. 2016;26: 717–724. pmid:26838340
- 34. Hibi Y, Asai K, Arafuka H, Hamajima M, Iwama T, Kawai K. Molecular structure of La3+-induced methanol dehydrogenase-like protein in Methylobacterium radiotolerans. J Biosci Bioeng. 2011;111: 547–549. pmid:21256798