Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Deltaproteobacteria (Pelobacter) and Methanococcoides are responsible for choline-dependent methanogenesis in a coastal saltmarsh sediment


Coastal saltmarsh sediments represent an important source of natural methane emissions, much of which originates from quaternary and methylated amines, such as choline and trimethylamine. In this study, we combine DNA stable isotope probing with high throughput sequencing of 16S rRNA genes and 13C2-choline enriched metagenomes, followed by metagenome data assembly, to identify the key microbes responsible for methanogenesis from choline. Microcosm incubation with 13C2-choline leads to the formation of trimethylamine and subsequent methane production, suggesting that choline-dependent methanogenesis is a two-step process involving trimethylamine as the key intermediate. Amplicon sequencing analysis identifies Deltaproteobacteria of the genera Pelobacter as the major choline utilizers. Methanogenic Archaea of the genera Methanococcoides become enriched in choline-amended microcosms, indicating their role in methane formation from trimethylamine. The binning of metagenomic DNA results in the identification of bins classified as Pelobacter and Methanococcoides. Analyses of these bins reveal that Pelobacter have the genetic potential to degrade choline to trimethylamine using the choline-trimethylamine lyase pathway, whereas Methanococcoides are capable of methanogenesis using the pyrrolysine-containing trimethylamine methyltransferase pathway. Together, our data provide a new insight on the diversity of choline utilizing organisms in coastal sediments and support a syntrophic relationship between Bacteria and Archaea as the dominant route for methanogenesis from choline in this environment.


Coastal saltmarsh sediments represent a highly productive environment, which are predominantly anaerobic and characterized by a high rate of carbon cycling [1]. These sediments represent a significant source of natural methane emissions, resulting from the degradation of organic matter, facilitated by the microorganisms inhabiting these sediments. It has been estimated that between 35 and 90% of the methane production in intertidal mudflats and saltmarshes originates from trimethylamine (TMA) [2, 3]. Quaternary amines are precursors of TMA and are ubiquitous in marine microbes, where they act as osmolytes and essential cellular components. Along with the common marine osmolyte glycine betaine (GBT), choline has been identified as an important precursor of TMA [2, 4]. Choline is a widely distributed component of membrane lipids and is essential for the formation of polar membrane lipids (such as phosphatidylcholine) in all eukaryotes and some bacteria [5]. Standing concentrations of choline have not been assessed in saltmarsh sediments due to the lack of a suitable method, however the concentrations of TMA range from low nanomolar (nM) in oceanic samples to low micromolar (mM) in marine and coastal sediments [6,7,8]. In marine and coastal systems, cycling of quaternary amines, such as choline, leads to atmospheric fluxes of methylated amines and methane, both of which are important climate-active trace gases [2, 7, 8].

In anaerobic marine and coastal sediments, methanogenesis is a key final step in organic matter degradation. However, the key microbes and the metabolic pathways responsible for methanogenesis from choline are yet to be established. Early studies have focused on the competition between sulphate reducers and methanogens, since sulphate reducers can utilize hydrogen and acetate at much lower concentrations, therefore outcompeting methanogens occupying the same environmental niche [3, 9, 10]. Hence it was thought that methanogens could only thrive in anaerobic sediments depleted of sulphate or utilize alternative non-competitive substrates, such as TMA [2, 4, 11]. Indeed, many methanogens isolated from these sediments, notably Methanococcoides of the family Methanosarcinales [11, 12], are able to use non-competitive substrates such as TMA, but not acetate, formate nor H2/CO2 for methanogenesis. These methylotrophic methanogens are known to form strong interactions with bacterial choline degraders [13, 14]. The bacterial choline-to-TMA degradation pathway, through a choline-TMA lyase (encoded by cutC), was only elucidated recently and we now know that cutC is widely distributed in many marine and coastal sediments [15,16,17]. It is therefore likely that methanogenesis from choline in anaerobic saltmarsh sediments requires the coupling of the bacterial degradation of choline, with subsequent methanogenesis from TMA. However, whether or not TMA is indeed a key intermediate in choline-dependent methanogenesis in saltmarsh sediments warrants investigation. Indeed, very recently, direct demethylation of choline for methanogenesis has been identified in Methanococcoides and Methanolobus strains [18,19,20,21], although TMA was the preferred substrate over choline and direct demethylation of choline only occurred in the absence of TMA in Methanococcoides sp. AM1 [21].

DNA-SIP (stable-isotope probing) is a powerful tool to link microbial identity to metabolic function, through the incorporation of a 13C label into the DNA of active microbes, in a culture-independent manner [22]. This technique has recently been applied in several studies to understand biogeochemical cycles in saltmarsh sediments [23,24,25]. Combining DNA-SIP with shotgun metagenomics and metagenomic binning to retrieve bins from 13C-labelled heavy DNA, offers the unique opportunity to uncover the metabolic potentials that are encoded by population genomes from metabolically active microbes [26,27,28]. In this study, we have used DNA-SIP with 13C2-labelled choline, followed by amplicon sequencing of 16S rRNA genes, to reveal the active microbial populations responsible for choline degradation. Furthermore, we used metagenomics to retrieve microbial genomes from 13C-labelled DNA and demonstrated that methanogenesis from choline in Stiffkey saltmarsh sediments is a two-step process, involving bacterial degradation of choline to TMA by Pelobacter, using the choline-TMA lyase pathway, followed by methanogenesis from TMA by the methylotrophic Archaea Methanococcoides.

Materials and methods

Sampling and microcosm set-ups

Sediment samples were collected from a pond at Stiffkey saltmarsh, Norfolk, UK (latitude 52.96, longitude 00.93) using an acrylic corer on 15/04/2014. The sediment core was transported to the laboratory and stored at 4 °C overnight, until microcosm incubations were set up. The sediment core was then sectioned into discrete layers and the most active, anaerobic 4–6 cm layer was used for analysis.

Microcosm set-up for DNA-SIP experiments were carried out as previously described by Neufeld et al. [29] using 13C2 choline (the two carbons in the acetyl group of choline was labelled with 13C, Sigma-Aldrich). Six sets of triplicate microcosms were used, consisting of 5 g of sediment, 20 ml of sea salts and an initial concentration of 5 mM choline (time point 0; T0). Microcosms were divided equally between 12C and 13C choline addition (18 each). Choline and TMA concentrations were monitored twice a day. Upon choline depletion one set of triplicate microcosms were sacrificed for 12C and 13C choline amended microcosms (time point 1, 166.5 h; T1) and an additional 5 mM 12C or 13C choline was provided to the remaining microcosms. This process of sacrifice and choline addition was repeated (time point 2, 214.5 h; T2). The remaining triplicate microcosms were sacrificed when the third addition of choline was depleted (time point 3, 261 h; T3).

Ion-exchange chromatography and gas chromatography

The concentrations of choline and TMA were determined twice a day using ion exchange chromatography (IC) on an 881 Compact IC Pro (Metrohm, Herisau, Switzerland) as described previously [30]. 100 μl aliquots of liquid medium from each microcosm was filter-sterilized with a 0.2 μm pore size centrifuge filter and diluted tenfold in MillQ water prior to IC-analysis.

Methane concentration in the head-space of the microcosms was monitored daily. Gas chromatography (GC) was carried out to quantify methane, using an Agilent 6890 FID instrument with a Porapak Q column with N2 carrier gas flowing at 20 ml min−1. The temperature set up was as follows: injector 150 °C, column 125 °C and detector 200 °C. An injection volume of 100 µl was used for all measurements. Methane concentrations were determined by peak area against a set of standards of known concentrations covering the measured range.

DNA extraction and gradient fractionation

DNA was extracted from ~500 mg of sample from the sacrificed microcosms, using the Fast DNA soil extraction kit (MP Bioscience, UK) according to the manufacturer’s instructions. DNA was extracted from the unamended sediment (T0), and the sacrificed microcosm sediments obtained at 166.5 h (T1), at 214.5 h (T2) and at 261 h after choline addition (T3). DNA concentrations were estimated using a spectrophotometer (NanoDrop ND-1000) and found to be in a range of 150–200 ng/μl. Aliquots of DNA (3 μg of DNA, 15–20 μl) were subjected to ultracentrifugation in CsCl. After centrifugation, between 12 and 13 CsCl fractions were collected by piercing the top and the bottom of the tube using a 23-Gauge needle. The density of each fraction was measured using a digital refractometer and DNA was recovered using polyethylene glycol as described previously [29].

Amplicon sequencing of 16S rRNA genes

Microbial community analyses of the 16S rRNA gene amplicons were performed using unfractionated DNA from T0, T1, T2 and T3 and DNA extracted from SIP gradient heavy and light fractions. The primers used for amplifying the 16S rRNA genes were designed to amplify both Bacteria and Archaea [31]. Amplicon sequencing of the 16S rRNA gene was carried on an Illumina Miseq platform, as described by Caporaso et al. [31], on the fractionated T0 ‘heavy’ and ‘light’ fractions, and for T1, T2 and T3, both 12C and 13C, heavy and light fractions. As a control against amplification of DNA originating from the laboratory and the DNA extraction kit, we ran nuclease-free water through the soil DNA extraction kit and Illumina PCR steps, referred to as the negative control. Multiplex 515F/806R paired-end 16S rRNA bacterial and archaeal community sequencing primers described by Caporaso et al. [31] were used.

16S rRNA gene amplicon reads were joined, de-multiplexed, trimmed and filtered. Singletons and chimeras were removed, reads were normalized and Operational Taxonomic Unit (OTU) binning was performed using the open-source bioinformatics pipeline Quantitative Insights Into Microbial Ecology (QIIME), MacQIIME version 1.8.0 [31]. Sequences were joined using fastq-join [32], singletons were removed, then the sequences were trimmed (240 bp) and quality filtered (maximum expected error threshold of 0.5) using USEARCH [33]. UCHIME was used for chimera detection and removal, with the Broad Microbiome Utilities ‘Gold.fa’ reference database [34]. For rarefaction analysis, approximately-maximum-likelihood phylogenetic trees were constructed using FastTree 2.1.3 [35]. OTU binning was performed with the UCLUST method with a cut off of 97% sequence identity using the RDP Classifier 2.2 against the Greengenes 13_8 dataset [33, 36, 37]. The QIIME processing yielded sequences of 240 bp, rarefied to a sequencing depth of 36,372 reads per sample (normalized to the smallest sample).

Metagenome sequencing and binning

Metagenomic sequencing was carried out on triplicate unfractionated time-point 0 and triplicate biological replicates of fractionated 13C-labelled time-point 3 DNA from both ‘heavy’ and ‘light’ fractions. The metagenomic libraries were constructed using the NEBNext Ultra DNA Library Prep Kit for Illumina (NEB, Hitchin, UK) and NEBNext Multiplex Oligos for Illumina (Index Primers Set 1), then sequenced using the MiSeq Reagent Kit v3 with 600 cycles.

The removal of adapters and quality trimming was carried out using sickle (quality > 20 average per kmer; [38]). Reads were assembled into contigs using Megahit v1.1.2, assigned to bins using Metabat v0.32.5 and annotated with Prokka [39,40,41,42]. Reads were mapped against contigs using BWA-MEM and resultant SAM and BAM files manipulated with Samtools v1.6 [43, 44]. To validate the completeness of each bin (assuming each bin represents a single organism), the percentage of the 40 core prokaryotic Clusters of Orthologous Groups of proteins (COGs) identified within each bin was calculated, based on previous approaches [45,46,47]. Estimation of contamination in each bin was carried out using the CheckM program [48]. Taxonomy was first assigned for each bin using specI [49]. Secondly, for the bins lacking a specI taxonomic assignment, blastp was carried out using the core COGs against the NCBI Nucleotide collection database (August 2017). Finally, for the remaining bins with too few core genes identified, blastp of all the contigs was carried out against the NCBI Nucleotide collection (nt) with an Hsp e-value cut-off of 1e–30.

Read data have been submitted to the European Nucleotide Archive (ENA) under the study accession number PRJEB23843.

Statistical and bioinformatics analysis

PRIMER v.6.8 (PRIMER-E, Plymouth, UK) was used for statistical analyses of the taxonomic data [50, 51]. The patterns of relative abundance for OTUs (16S rRNA gene amplicons) or bins (metagenome) from each sample were ordinated to each other using non-metric multidimensional scaling (MDS) analysis, allowing the comparison of compositional similarity between samples. Resemblance matrices were calculated on untransformed standardized OTU/bin relative abundance data using Bray–Curtis similarity analysis. MDS analysis was applied to the matrices using the default settings in PRIMER, with Kruskal’s stress formula 1, a minimum stress of 0.01 and 50 restarts. Similarity percentage analysis (SIMPER) was used to determine the percentage of similarity and dissimilarity between sets of microbial communities; i.e. time-points, 12C and 13C samples and fractions. SIMPER was applied to the resemblance matrices in PRIMER using the default settings.

To understand the relative prevalence of selected functional metabolic pathways in the metagenome data, BLAST analysis of both the binned and the un-binned metagenome data was carried out. The representative protein sequence queries were selected because they had proven role in choline and TMA degradation. Additionally, a control query for GBT degradation was selected. The functional metabolic genes used were; cutC encoding choline-TMA lyase [16, 52], mttB encoding the pyrrolysine-containing TMA methyltransferase [53], and the non-pyrrolysine containing GBT: corrinoid methyltransferase, mtgB [19]. After BLAST analysis, the resultant hits were aligned with MUSCLE 3.5 [54] and maximum likelihood phylogenetic trees were reconstructed using PhyML 3.0 [55] with default settings (i.e. HKY85 model of nucleotide substitutions, ts/tv ratio = 4.00, BioNJ starting tree). This phylogenetic analysis predicted true-positive hits, based on their clustering with representative, validated functional genes, whilst false-positive hits (those that clustered closest to genes with alternative functions) were rejected. The positive BLAST hits were normalized to gene length and sample size.


Methanogenesis from choline involved TMA as a key intermediate in saltmarsh sediments

We have hypothesized that methanogenesis from choline in saltmarsh sediments is a two-step process, involving the bacterial degradation of choline to TMA and subsequent methane formation from TMA by methylotrophic methanogens. Triplicate choline enrichment microcosms (with 13C2-labelled choline and unlabelled choline), were monitored for choline and its metabolites using IC and methane formation was quantified by GC. The IC data presented in Fig. 1 showed that the addition of choline (100 μmol) at T0 was depleted and an equivalent amount of TMA was produced after 60 h. The released TMA was fully degraded after a further 7 days (186 h) and approximately two times the amount of methane was produced (202 μmol). The time taken to degrade all the choline decreased with the two subsequent choline additions, which were 48 and 25 h respectively (Fig. 1). TMA was not detectable after the second and third choline additions, suggesting a rapid consumption by the microbial community (Fig. 1). Following the second choline addition, the methane production was roughly equal to the choline degradation rate (Fig. 1). The cumulative amount of choline after the three additions was 300 μmol (per incubation) and the average amount of methane produced was 469 μmol (per incubation).

Fig. 1

Choline, trimethylamine (TMA) and methane dynamics during the microcosm incubations. Choline and TMA were determined by ion-exchange chromatography; methane was determined by gas chromatography. Time-points at which the microcosms were sampled are indicated by the labels T0, T1 (166.5 h), T2 (214.5 h) and T3 (261 h). Error bars represent standard deviations from six to nine biological replicates

Microbial community analyses of the 16S rRNA genes from the choline-amended microcosms and the SIP gradient fractions by amplicon sequencing

To uncover the microbial community response to choline amendment and to uncover the microbes involved in choline transformation, we sequenced the 16S rRNA genes from both the light and heavy fractions of the DNA-SIP microcosms over the time course experiment. Sequencing using the Illumina Miseq platform yielded 4,806,794 reads. OTUs were assigned to 3,353,870 sequences, with an average sequencing depth of 74,869 reads per sample. The QIIME processing yielded sequences of 240 bp, rarefied to the lowest sequencing depth of 36,372 reads per sample. Sequencing reads were assigned to 940 OTUs at 97% sequence identity cut-offs.

Over the SIP time course there was a progressive community shift from T0 through T1 to T2 and T3. MDS analysis showed that T2 and T3 heavy fractions of the 13C2-choline microcosms were significantly different from the unlabelled microcosms, an indication that the labelling was successful for the enriched 13C2-choline metabolizing bacteria (Fig. 2a). The light fractions from 13C2-microcosms clustered closely with T0 samples (<25% dissimilar; Fig. 2a). The negative control was most dissimilar to all other samples (92.9–99.8%), and clustered separately from all other samples in the MDS plot (Table 1, Fig. 2a).

Fig. 2

Multidimensional scaling (MDS) plots of fractionated choline DNA-SIP samples taken before enrichment (T0) and at time-points 1 (T1), 2 (T2) and 3 (T3) incubated with either 12C-choline (12C light and 12C heavy) or 13C2-choline (13C light and 13C heavy). a 16S rRNA gene amplicon data based on rarefied absolute abundance of OTUs (97% cut-off), with an additional blank negative control (−ve) sample. b Metagenome bins abundance, normalized to contig length and sample size

Table 1 Dissimilarity matrix of the SIMPER analysis showing average dissimilarity between triplicate samples for 16 S rRNA gene OTUs and metagenome bins for each time point (T0, T1, T2 and T3) and light versus heavy fractions

The SIMPER analysis for the T1 the heavy and light 13C2-choline fractions showed these two communities had 79% dissimilarity, which increased to 96% for T2 and 93% for T3 (Table 1), suggesting a clear separation of an active microbial population from the community in response to 13C2-choline amendment. Yet the unlabelled 12C-choline microcosms showed <40% dissimilarity between the heavy and light fractions (Table 1). To determine the key OTUs responsible for the observed dissimilarity of the microbial communities between the heavy and light fractions, SIMPER analysis was applied. Data presented in Fig. 3 indicate that there was an increase in the Archaea Methanococcoides, and Deltaproteobacteria (Desulfuromonas and Pelobacter), whilst other taxa such as Epsilonproteobacteria (Helicobacteraceae) and Gammaproteobacteria declined in relative abundance after choline additions (Fig. 3). These same taxonomic groups (i.e. Pelobacter and Methanococcoides) also showed enrichment when much lower levels of choline (150 μM, 1.5 mM) were added to microcosms using the Stiffkey saltmarsh sediments (Fig. S1).

Fig. 3

Relative abundance of the top 9 taxonomically assignments as identified by similarity percentage analysis (SIMPER). a 16S rRNA gene amplicon OTUs. b Metagenome bins

Two Deltaproteobacteria OTUs of the Desulfuromonadaceae family (OTU 514001824 and OTU 958716325), classified as Pelobacter and Desulfuromonas respectively (Fig. 4), were significantly enriched in the 13C-choline heavy fractions which, together, accounted for >80% reads in 13C-heavy fractions (Table S1). The SIMPER analysis indicated that these two taxonomic groups, Pelobacter and Desulfuromonas OTUs, accounted for 19% and 12%, respectively, of community shift between all fractions and time points (Table S3). These OTUs accounted for <0.1% of the sequences in T0 samples before 13C2-choline enrichment (Fig. 3). Methanococcoides was enriched in all fractions, i.e. ‘heavy’ and ‘light’ for both 12C and 13C-choline microcosms (Table S1).

Fig. 4

Neighbour-joining phylogenetic tree of 16S rRNA genes of Desulfuromonadaceae. Sequence alignment was carried out using the ClustalW program based on partial 16S rRNA genes (1320 bp) from representative species from the Geobacteraceae and Desulfuromonadaceae families and the two most prevalent OTUs derived from amplicon sequencing of 16S rRNA genes from the choline DNA-SIP heavy fractions (OTU 514001824 and OTU 958716325). Sedimentation selenatireducens and Desulfoluna spongiiphila were used as the outgroup to root the tree. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (500 replicates) are shown next to the branches. Only bootstrap values greater than 50% are shown. The evolutionary distances were computed using the Maximum Composite Likelihood method in the MEGA7 package [62]. The scale bar indicates evolutionary distance in mutations per residue

Metagenome sequencing of heavy isotope labelled DNA from 13C2-choline DNA-SIP and bins

To gain a better understanding of the metabolic potential in these enriched taxa (i.e. Pelobacter), we sequenced triplicate 13C light and triplicate 13C heavy fractions of time point 3 (T3) DNA-SIP fractions, together with the triplicate unfractionated T0 samples. The dataset was sequenced using an Illumina MiSeq with 2 × 300 paired-end cycles, resulting in ~186 Mb sequencing data from 11 runs. After quality trimming and filtering, ~116 Mb were mapped to contigs. For the T0, T3 light and T3 heavy samples the percentage of reads that mapped to contigs accounted for 49%, 65% and 80% of total reads, respectively. These resulted in 230,960 contigs, with a minimum length cut-off of 1 kb, which were assigned to 270 bins (Table S2).

The MDS analysis of the metagenome bins showed that the T0 and T3 heavy fraction samples clustered closely with themselves, with >70% similarity (Fig. 2B). The T3 light fraction samples showed lower similarity. The T0, T3 light and T3 heavy clusters were distinct from each other. The MDS analysis of metagenome bins therefore agrees with the results from amplicon sequencing data on 16S rRNA genes in that significant enrichment of active microbes involved in choline metabolism had occurred in the SIP microcosms.

SIMPER analyses revealed that the T3 13C-choline heavy labelled samples were highly dissimilar to both the T0 (87%) and the T3 light samples (80%; Table 1). In agreement with the 16S rRNA gene amplicon data (Table S3), the SIMPER analyses of the metagenome revealed an increase in the Archaea Methanococcoides between the T0 samples and T3 light fraction samples and confirmed enrichment of Pelobacter in the T3 heavy fraction samples (Table S4).

The taxonomy assigned to the 270 metagenome bins was determined to the greatest possible resolution, which varied between class and species level (Table S2). The most prevalent taxonomic assignment among the 270 bins (irrespective of relative abundance) was Pelobacter (40 bins), Desulfovibrio (19 bins) and Methanococcoides (13 bins; Table S2).

Functional gene profiling of the metagenomes and microbial genomes retrieved from DNA-SIP

Although each of the metagenome bins only represents a partial genome of the assigned species, they provided a valuable source for mining the functional genetic potential involved in choline degradation and methane formation. Specifically, these bins were screened for the presence of the metabolic genes involved in converting choline to TMA (cutC) and TMA to methane (mttB). Data presented in Table 2 and Table S2 revealed that cutC was indeed present in bins that were assigned to Pelobacter (e.g. 5, 134, 174, 227 and 252). Similarly, the mttB gene was also found in bins that were assigned to Methanococcoides (e.g. 68, 76, 117 and 267). Additionally, a complete set of proteins (mtaA, mtaB and mtaC) required for methanogenesis from methanol are also found in these Methanococcoides bins (e.g. 68 and 267). Analysis of the recovered contigs containing these key functional genes showed that for Bin 174 the cutC containing contig showed high gene synteny as well as sequence similarity to Pelobacter isolates (Fig. 5).

Table 2 Selected bins containing homologues of functional genes (cutC and mttB) involved in choline-dependent methanogenesis
Fig. 5

Alignment of the cutC-homologue containing contig from the 13C choline enriched bin 174 with the closest genome matches—Pelobacter acetylenicus and Pelobacter carbinolicus

BLAST analysis of the key genes in the functional metabolic pathways against the unassembled metagenome reads revealed an increase in the two expected pathways involved in methanogenesis from choline (Fig. 6). The cutC gene was detected at very low levels (<0.05 hits per million reads) in both the T0 and T3 light metagenome samples, but we detected 12.4 hits per million reads in the T3 heavy samples. Phylogenetic analysis of the cutC genes extracted from the T3 heavy metagenome samples showed that the majority (~60%) of the cutC sequences originated from Pelobacter. The mttB gene was also detected at low levels (<0.4 hits per million reads) in the T0 samples, then increased to 2.8 in T3 light and 17.6 hits per million reads in T3 heavy samples. Phylogenetic analyses of mttB sequences retrieved from these metagenomes confirmed that they were from Methanococcoides. Interestingly, our analyses of the nonpyrrolysine-containing GBT methyltransferase mtgB gene, which is responsible for the direct demethylation of GBT [19], showed little variation in abundance (0.2–0.8 hits per million reads) between T0 and T3 samples.

Fig. 6

Analyses of functional gene abundance in the un-binned metagenome datasets obtained from saltmarsh sediments before (T0) and after choline DNA-SIP (L-, light and H-, heavy fractions of T3). The cutC gene encodes a choline-TMA lyase ([17, 16]). mttB encodes a pyrrolysine-containing TMA methyltransferase [63, 64] and mtgB encodes a non-pyrrolysine glycine betaine (GBT) methyltransferase responsible for direct demethylation of GBT to dimethylglycine (DMG) and methane [19]. Relative abundance was achieved by normalizing to the length and abundance of the recA gene. Note that the normalized mttB counts are not exclusive to methanogens and a complete separation of bacterial pyrrolysine-containing mttB from their archaeal counterparts through phylogeny analysis was not possible [64]. The pie charts give a breakdown of the phylogeny of the hits against a reference tree for cutC or mttB, and the size of the pie charts reflects relative abundance of sequences retrieved from the metagenomes. The cutC sequences grouped with either Pelobacter (P. acetylenicus and P. carbinolicus) or Desulfovibrio salexigens whereas the archaeal mttB sequences clustered within Methanococcoides


Coastal saltmarshes represent a significant natural source of methane in the global methane budget. Our understanding of the key microbes and metabolic pathways responsible for methanogenesis by microbial populations inhabiting these sediments is still very limited. In this study, we have shown that the Deltaproteobacteria genera Pelobacter and the Archaea, Methanococcoides were the predominant species involved in methanogenesis from Stiffkey saltmarsh sediments. These microbes were identified through a combination of DNA-SIP, high-throughput sequencing analyses of the 16S rRNA genes, and 13C-enriched metagenomics, on which binning and subsequent reconstruction of the metagenome bins was performed.

DNA-SIP and amplicon sequencing analyses clearly showed that the Desulfuromonadaceae, particularly Pelobacter and Desulfuromonas were virtually exclusively found in the 13C heavy fractions, suggesting that these obligate anaerobes incorporated heavy carbon from the 13C2-choline into their DNA. Indeed, cultivated representatives of Pelobacter have previously been isolated from various marine and coastal sediments and can grow on choline while producing TMA [56, 57]. This was further supported by subsequent analyses of the retrieved population genomes that were assigned to Pelobacter (5, 134, 174, 227 and 252), showing the presence of cutC and associated genes involved in the choline-TMA lyase pathway (Table 2, Fig. 5). Desulfuromonas isolates such as Desulfuromonas acetoxidans [58] and Desulfuromonas svalbardensis [59], on the other hand, are not known to grow on choline and no cutC homologues were found in the genomes of the two aforementioned strains. Yet, our data presented in Fig. 3 appears to support the role of as-yet uncultivated novel Desulfuromonas strains in these sediments in choline degradation. However, it is also likely that Desulfuromonas was labelled through cross-feeding of 13C-acetate released by Pelobacter. Indeed, both Pelobacter carbinolicus and Pelobacter acetylenicus constitutively expressed an acetaldehyde dehydrogenase, converting acetaldehyde to acetate in pure culture or co-culture with a methanogen [60]. Interestingly, a significant proportion of cutC sequences retrieved from the 13C-heavy metagenomes are classified as Desulfovibrio salexigens (Fig. 6). Furthermore, several metagenome-assembled bins related to this sulphate-reducing bacterium were also found although the SpecI matches of these bins were <90% identical to those of Desulfovibrio salexigens (Table S2). However, the 16S rRNA gene amplicon sequencing indicated that Desulfovibrio (Family Desulfovibrionaceae) was not abundant (<1%) in this saltmarsh sediment and our SIMPER analyses did not support their role in contributing to community shift during choline SIP incubations (Fig. 3, Table S1). Clearly, the role of Desulfuromonas and Desulfovibrio in choline degradation warrants further investigation.

Amplicon sequencing of 16S rRNA genes showed that Methanococcoides were also enriched over time, however, they showed no preferential enrichment in the heavy 13C fractions compared to either the light fractions or the heavy fractions from the 12C-choline control microcosms (Fig. 3). Methanococcoides cannot use acetate and therefore could not incorporate 13C label from the acetyl group of 13C2-choline [14]. They can, however, use TMA for methanogenesis [14]. TMA, a metabolite of bacterial choline degradation [16, 17], was indeed found in our choline-amended microcosms (Fig. 1). We therefore postulate that the enrichment of Methanococcoides in these microcosms was due to their growth on TMA generated by bacterial choline degradation. To further support our hypothesis, our analyses of the population genomes retrieved from metagenomics sequences showed the presence of mttB, TMA-methyltransferase, a key gene involved in methanogenesis from TMA, in Methanococcoides bins (Table 2).

Metagenomic sequencing of the 13C-enriched DNA and binning of the metagenomic data not only provided a better understanding of the metabolic potential of the key functional microbes identified by SIP, Desulfuromonadaceae (e.g. Pelobacter) and Methanococcoides, but also provided an opportunity to elucidate major pathway for methanogenesis from choline in the sediments. There are several possible routes for choline transformation to methane in anaerobic sediments (Fig. 7): (1) direct demethylation of choline to methane (the enzyme responsible has yet to be characterized [11, 21]); (2) TMA formation from choline followed by methanogenesis from TMA, or (3) demethylation of GBT to methane. The occurrence of TMA during the anaerobic degradation of choline in our microcosms (Fig. 1) suggests that direct choline demethylation was not the dominant route for methanogenesis. Instead, our data suggest that choline was initially converted to TMA, which then served as the substrate for methanogenesis. TMA formation from choline can be achieved through either a choline-TMA lyase encoded by cutC or, alternatively, GBT as the intermediate (Fig. 7). We, however, did not observe any significant accumulation of GBT in the microcosms although the method is capable of quantifying GBT, TMA and choline simultaneously [61]. To further support our hypothesis, we compared the relative abundance of key metabolic genes in the un-amended salt-marsh sediment and after choline enrichment. The relative capacity of the microbial community to degrade choline to TMA and subsequently TMA to methane markedly increased after SIP incubations, as evidenced by a substantial increase in relative abundance of both cutC (choline to TMA) and mttB (TMA to methane) (Fig. 6). In contrast, the relative abundance of mtgB, encoding the GBT methyltransferase responsible for direct demethylation of GBT [19], was more abundant before enrichment and showed no response to choline amendment in microcosms, suggesting that GBT was not a major intermediate in methanogenesis from choline.

Fig. 7

Currently known pathways for methanogenesis from choline. Choline can be converted to trimethylamine (TMA) by either a bacterial choline-TMA lyase (encoded by cutC [17, 16]) or indirectly through glycine betaine (GBT) as the intermediate. Methanogenesis from TMA by methanogenic Archaea requires the key enzyme, pyrrolysine-containing TMA methyltransferase encoded by mttB [53]. Direct demethylation of choline and GBT can also support methanogenesis and the GBT methyltransferase (mtgB) has been identified very recently [19, 21] whereas genes responsible for direct choline demethylation to methane have not yet been identified [20, 21]

Using a combination of DNA-SIP with 13C2-labelled choline, 16S rRNA gene sequencing and metagenome sequencing we have identified Deltaproteobacteria, of the genera Pelobacter as the major choline-utilizers and TMA producers, whilst the methanogenic Archaea Methanococcoides was also enriched and involved in methane formation from TMA. Metagenome and metabolite data showed a significant enrichment in the choline degradation pathway via TMA to methane and a correlating intermediate release of TMA and a final accumulation of methane in the choline enrichment microcosms. This all indicated that a syntrophic relationship between Bacteria and Archaea was the dominant route for methanogenesis from choline in Stiffkey saltmarsh sediments.


  1. 1.

    Nedwell DB. The input and mineralization of organic carbon in anaerobic aquatic sediments. In: Marshall KC (eds). Advances in Microbial Ecology. Advances in Microbial Ecology, vol 7. Springer, Boston, MA. 1984 p. 93–131.

  2. 2.

    King GM. Metabolism of trimethylamine, choline, and glycine betaine by sulfate-reducing and methanogenic bacteria in marine sediments. Appl Environ Microbiol. 1984;48:719–25.

    CAS  PubMed  PubMed Central  Google Scholar 

  3. 3.

    Oremland RS, Marsh LM, Polcin S. Methane production and simultaneous sulphate reduction in anoxic, salt marsh sediments. Nature. 1982;296:143–5.

    CAS  Article  Google Scholar 

  4. 4.

    King GM. Distribution and metabolism of quaternary amines in marine sediments. In: Blackburn TH, SØrensen J (eds). Nitrogen cycling in coastal marine environments. John Wiley and Sons Ltd, Chichester, UK. 1988. p. 143–73.

  5. 5.

    Sohlenkamp C, López-Lara IM, Geiger O. Biosynthesis of phosphatidylcholine in bacteria. Progress Lipid Res. 2003;42:115–62.

    CAS  Article  Google Scholar 

  6. 6.

    Fitzsimons M, Kahni-Danon B, Dawitt M. Distributions and adsorption of the methylamines in the inter-tidal sediments of an East Anglian Estuary. Environ Exp Bot. 2001;46:225–36.

    CAS  Article  Google Scholar 

  7. 7.

    Gibb SW, Mantoura RFC, Liss PS. Ocean‐atmosphere exchange and atmospheric speciation of ammonia and methylamines in the region of the NW Arabian Sea. Glob Biogeochem Cycles. 1999;13:161–78.

    CAS  Article  Google Scholar 

  8. 8.

    Gibb SW, Hatton AD. The occurrence and distribution of trimethylamine-N-oxide in Antarctic coastal waters. Mar Chem. 2004;91:65–75.

    CAS  Article  Google Scholar 

  9. 9.

    Schönheit P, Kristjansson JK, Thauer RK. Kinetic mechanism for the ability of sulfate reducers to out-compete methanogens for acetate. Arch Microbiol. 1982;132:285–8.

    Article  Google Scholar 

  10. 10.

    Thauer RK, Jungermann K, Decker K. Energy conservation in chemotrophic anaerobic bacteria. Bacteriol Rev. 1977;41:100.

    CAS  PubMed  PubMed Central  Google Scholar 

  11. 11.

    Parkes RJ, Brock F, Banning N, Hornibrook ER, Roussel EG, Weightman AJ, et al. Changes in methanogenic substrate utilization and communities with depth in a salt-marsh, creek sediment in southern England. Estuar, Coast Shelf Sci. 2012;96:170–8.

    Article  Google Scholar 

  12. 12.

    Munson MA, Nedwell DB, Embley TM. Phylogenetic diversity of Archaea in sediment samples from a coastal salt marsh. Appl Environ Microbiol. 1997;63:4729–33.

    CAS  PubMed  PubMed Central  Google Scholar 

  13. 13.

    Fiebig K, Gottschalk G. Methanogenesis from choline by a coculture of Desulfovibrio sp. and Methanosarcina barkeri. Appl Environ Microbiol. 1983;45:161–8.

    CAS  PubMed  PubMed Central  Google Scholar 

  14. 14.

    Sowers KR, Ferry JG. Isolation and characterization of a methylotrophic marine methanogen, Methanococcoides methylutens gen. nov., sp. nov. Appl Environ Microbiol. 1983;45:684–90.

    CAS  PubMed  PubMed Central  Google Scholar 

  15. 15.

    Jameson E, Doxey AC, Airs R, Purdy KJ, Murrell JC, Chen Y. Metagenomic data-mining reveals contrasting microbial populations responsible for trimethylamine formation in human gut and marine ecosystems. Microb Genomics. 2016a;2.

  16. 16.

    Jameson E, Fu T, Brown IR, Paszkiewicz K, Purdy KJ, Frank S,. et al. Anaerobic choline metabolism in microcompartments promotes growth and swarming of Proteus mirabilis. Environ Microbiol. 2016b;18:2886–98.

  17. 17.

    Craciun S, Balskus EP. Microbial conversion of choline to trimethylamine requires a glycyl radical enzyme. Proc Natl Acad Sci USA. 2012;109:21307–12.

  18. 18.

    L’Haridon S, Chalopin M, Colombo D, Toffin L. Methanococcoides vulcani sp. nov., a marine methylotrophic methanogen that uses betaine, choline and N,N-dimethylethanolamine for methanogenesis, isolated from a mud volcano, and emended description of the genus Methanococcoides. Int J Syst Evolut Microbiol. 2014;64:1978–83.

    Article  Google Scholar 

  19. 19.

    Ticak T, Kountz DJ, Girosky KE, Krzycki JA, Ferguson DJ. A nonpyrrolysine member of the widely distributed trimethylamine methyltransferase family is a glycine betaine methyltransferase. Proc Natl Acad Sci USA. 2014;111:E4668–76.

  20. 20.

    Ticak T, Hariraju D, Arcelay MB, Arivett BA, Fiester SE, Ferguson DJ. Isolation and characterization of a tetramethylammonium-degrading Methanococcoides strain and a novel glycine betaine-utilizing Methanolobus strain. Arch Microbiol. 2015;197:197–209.

    CAS  Article  Google Scholar 

  21. 21.

    Watkins AJ, Roussel EG, Webster G, Parkes RJ, Sass H. Choline and N,N-dimethylethanolamine as direct substrates for methanogens. Appl Environ Microbiol. 2012;78:8298–303.

    CAS  Article  Google Scholar 

  22. 22.

    Radajewski S, Ineson P, Parekh NR, Murrell JC. Stable-isotope probing as a tool in microbial ecology. Nature. 2000;403:646–9.

    CAS  Article  Google Scholar 

  23. 23.

    Darjany LE, Whitcraft CR, Dillon JG. Lignocellulose-responsive bacteria in a southern California salt marsh identified by stable isotope probing. Front Microbiol. 2014;5.

  24. 24.

    Seyler LM, McGuinness LM, Kerkhof LJ. Crenarchaeal heterotrophy in salt marsh sediments. ISME J. 2014;8:1534–43.

    CAS  Article  Google Scholar 

  25. 25.

    Webster G, Rinna J, Roussel EG, Fry JC, Weightman AJ, Parkes RJ. Prokaryotic functional diversity in different biogeochemical depth zones in tidal sediments of the Severn Estuary, UK, revealed by stable-isotope probing. FEMS Microbiol Ecol. 2010;72:179–97.

    CAS  Article  Google Scholar 

  26. 26.

    Eyice Ö, Namura M, Chen Y, Mead A, Samavedam S, Schäfer H. SIP metagenomics identifies uncultivated Methylophilaceae as dimethylsulphide degrading bacteria in soil and lake sediment. ISME J. 2015;9:2336–48.

    CAS  Article  Google Scholar 

  27. 27.

    Ziels RM, Sousa DZ, Stensel HD, Beck DA. DNA-SIP based genome-centric metagenomics identifies key long-chain fatty acid-degrading populations in anaerobic digesters with different feeding frequencies. ISME J. 2017;12:112–23.

    Article  Google Scholar 

  28. 28.

    Coyotzi S, Pratscher J, Murrell JC, Neufeld JD. Targeted metagenomics of active microbial populations with stable-isotope probing. Curr Opin Biotechnol. 2016;41:1–8.

    CAS  Article  Google Scholar 

  29. 29.

    Neufeld JD, Vohra J, Dumont MG, Lueders T, Manefield M, Friedrich MW, et al. DNA stable-isotope probing. Nat Protoc. 2007;2:860–6.

    CAS  Article  Google Scholar 

  30. 30.

    Zhu Y, Jameson E, Crosatti M, Schäfer H, Rajakumar K, Bugg TD, et al. Carnitine metabolism to trimethylamine by an unusual Rieske-type oxygenase from human microbiota. Proc Natl Acad Sci USA. 2014;111:4268–73.

  31. 31.

    Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods. 2010;7:335–6.

    CAS  Article  Google Scholar 

  32. 32.

    Aronesty E. ea-utils: Command-line tools for processing biological sequencing data. Durham, NC: Expression Analysis; 2011.

    Google Scholar 

  33. 33.

    Edgar RC. Search and clustering orders of magnitude faster than BLAST. Bioinformatics. 2010;26:2460–1.

    CAS  Article  Google Scholar 

  34. 34.

    Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics. 2011;27:2194–200.

    CAS  Article  Google Scholar 

  35. 35.

    Price MN, Dehal PS, Arkin AP. FastTree 2–approximately maximum-likelihood trees for large alignments. PLoS ONE. 2010;5:e9490.

    Article  Google Scholar 

  36. 36.

    McDonald D, Price MN, Goodrich J, Nawrocki EP, DeSantis TZ, Probst A, et al. An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea. ISME J. 2012;6:610–8.

    CAS  Article  Google Scholar 

  37. 37.

    Wang Q, Garrity GM, Tiedje JM, Cole JR. Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl Environ Microbiol. 2007;73:5261–7.

    CAS  Article  Google Scholar 

  38. 38.

    Joshi N, Fass J. Sickle: a sliding-window, adaptive, quality-based trimming tool for FastQ files (Version1.33) [Software]. 2011.

  39. 39.

    Kang DD, Froula J, Egan R, Wang Z. MetaBAT, an efficient tool for accurately reconstructing single genomes from complex microbial communities. Peer J. 2015;3:e1165.

    Article  Google Scholar 

  40. 40.

    Li D, Liu C-M, Luo R, Sadakane K, Lam T-W. MEGAHIT: an ultra-fast single-node solution for large and complex metagenomics assembly via succinct de Bruijn graph. Bioinformatics. 2015;31:1674–6.

    CAS  Article  Google Scholar 

  41. 41.

    Seemann T. Prokka: rapid prokaryotic genome annotation. Bioinformatics. 2014;30:2068–9.

    CAS  Article  Google Scholar 

  42. 42.

    Vollmers J, Wiegand S, Kaster A-K. Comparing and evaluating metagenome assembly tools from a microbiologist’s perspective-not only size matters! PLoS ONE. 2017;12:e0169662.

    Article  Google Scholar 

  43. 43.

    Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, et al. The sequence alignment/map format and SAMtools. Bioinformatics. 2009;25:2078–9.

    Article  Google Scholar 

  44. 44.

    Li H. Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. Preprint arXiv:13033997. 2013.

  45. 45.

    Hess M, Sczyrba A, Egan R, Kim T-W, Chokhawala H, Schroth G, et al. Metagenomic discovery of biomass-degrading genes and genomes from cow rumen. Science. 2011;331:463–7.

    CAS  Article  Google Scholar 

  46. 46.

    Ciccarelli FD, Doerks T, Von Mering C, Creevey CJ, Snel B, Bork P. Toward automatic reconstruction of a highly resolved tree of life. Science. 2006;311:1283–7.

    CAS  Article  Google Scholar 

  47. 47.

    Wrighton KC, Thomas BC, Sharon I, Miller CS, Castelle CJ, VerBerkmoes NC, et al. Fermentation, hydrogen, and sulfur metabolism in multiple uncultivated bacterial phyla. Science. 2012;337:1661–5.

    CAS  Article  Google Scholar 

  48. 48.

    Parks DH, Imelfort M, Skennerton CT, Hugenholtz P, Tyson GW. CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes. Genome Res. 2015;25:1043–55.

    CAS  Article  Google Scholar 

  49. 49.

    Mende DR, Sunagawa S, Zeller G, Bork P. Accurate and universal delineation of prokaryotic species. Nat Methods. 2013;10:881–4.

    CAS  Article  Google Scholar 

  50. 50.

    Clarke K. Non‐parametric multivariate analyses of changes in community structure. Aust J Ecol. 1993;18:117–43.

    Article  Google Scholar 

  51. 51.

    Clarke K, Ainsworth M. A method of linking multivariate community. Mar Ecol Progress Ser. 1993;92:205–19.

    Article  Google Scholar 

  52. 52.

    Kalnins G, Kuka J, Grinberga S, Makrecka-Kuka M, Liepinsh E, Dambrova M, et al. Structure and function of CutC choline lyase from human microbiota bacterium Klebsiella pneumoniae. J Biol Chem. 2015;290:21732–40.

    CAS  Article  Google Scholar 

  53. 53.

    Ferguson D, Krzycki JA. Reconstitution of trimethylamine-dependent coenzyme M methylation with the trimethylamine corrinoid protein and the isozymes of methyltransferase II from Methanosarcina barkeri. J Bacteriol. 1997;179:846–52.

    CAS  Article  Google Scholar 

  54. 54.

    Edgar RC. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 2004;32:1792–7.

    CAS  Article  Google Scholar 

  55. 55.

    Guindon S, Dufayard J-F, Lefort V, Anisimova M, Hordijk W, Gascuel O. New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Syst Biol. 2010;59:307–21.

    CAS  Article  Google Scholar 

  56. 56.

    Aklujkar M, Haveman SA, DiDonato R, Chertkov O, Han CS, Land ML, et al. The genome of Pelobacter carbinolicus reveals surprising metabolic capabilities and physiological features. BMC Genomics. 2012;13:690.

    CAS  Article  Google Scholar 

  57. 57.

    Schink B. Fermentation of acetylene by an obligate anaerobe, Pelobacter acetylenicus sp. nov. Arch Microbiol. 1985;142:295–301.

    CAS  Article  Google Scholar 

  58. 58.

    Pfennig N, Biebl H. Desulfuromonas acetoxidans gen. nov. and sp. nov., a new anaerobic, sulfur-reducing, acetate-oxidizing bacterium. Arch Microbiol. 1976;110:3–12.

    CAS  Article  Google Scholar 

  59. 59.

    Vandieken V, Mußmann M, Niemann H, Jørgensen BB. Desulfuromonas svalbardensis sp. nov. and Desulfuromusa ferrireducens sp. nov., psychrophilic, Fe (III)-reducing bacteria isolated from Arctic sediments, Svalbard. Int J Syst Evolut Microbiol. 2006;56:1133–9.

    CAS  Article  Google Scholar 

  60. 60.

    Schmidt A, Frensch M, Schleheck D, Schink B, Muller N. Degradation of acetaldehyde and its precursors by Pelobacter cabinolicus and P. acetylenicus. PLoS ONE. 2014;9:e115902

    Article  Google Scholar 

  61. 61.

    Lidbury I, Murrell JC, Chen Y. Trimethylamine and trimethylamine N-oxide are supplementary energy sources for a marine heterotrophic bacterium: implications for marine carbon and nitrogen cycling. ISME J. 2015;9:760–9.

    CAS  Article  Google Scholar 

  62. 62.

    Kumar S, Stecher G, Tamura K. MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol Biol Evol. 2016;33:1870–4.

    CAS  Article  Google Scholar 

  63. 63.

    Burke SA, Lo SL, Krzycki JA. Clustered genes encoding the methyltransferases of methanogenesis from monomethylamine. J Bacteriol. 1998;180:3432–40.

    CAS  PubMed  PubMed Central  Google Scholar 

  64. 64.

    Prat L, Heinemann IU, Aerni HR, Rinehart J, O’Donoghue P, Söll D. Carbon source-dependent expansion of the genetic code in bacteria. Proc Natl Acad Sci USA. 2012;109:21070–5.

Download references


This work was supported by the Natural Environment Research Council (NERC), UK, through research grants (NE/I027061/1 and NE/I025077/1) and a Ph.D. studentship to H.J. and a Warwick Integrative Synthetic Biology (WISB) early career fellowship, funded jointly by BBSRC and EPSRC to E.J.

Author information



Corresponding authors

Correspondence to Eleanor Jameson or Yin Chen.

Ethics declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Electronic supplementary material

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Jameson, E., Stephenson, J., Jones, H. et al. Deltaproteobacteria (Pelobacter) and Methanococcoides are responsible for choline-dependent methanogenesis in a coastal saltmarsh sediment. ISME J 13, 277–289 (2019).

Download citation

Further reading


Quick links