The genome sequence of the rumen methanogen Methanobrevibacter ruminantium reveals new possibilities for controlling ruminant methane emissions

Abstract

Background: Methane (CH(4)) is a potent greenhouse gas (GHG), having a global warming potential 21 times that of carbon dioxide (CO(2)). Methane emissions from agriculture represent around 40% of the emissions produced by human-related activities, the single largest source being enteric fermentation, mainly in ruminant livestock. Technologies to reduce these emissions are lacking. Ruminant methane is formed by the action of methanogenic archaea typified by Methanobrevibacter ruminantium, which is present in ruminants fed a wide variety of diets worldwide. To gain more insight into the lifestyle of a rumen methanogen, and to identify genes and proteins that can be targeted to reduce methane production, we have sequenced the 2.93 Mb genome of M. ruminantium M1, the first rumen methanogen genome to be completed.

Methodology/principal findings: The M1 genome was sequenced, annotated and subjected to comparative genomic and metabolic pathway analyses. Conserved and methanogen-specific gene sets suitable as targets for vaccine development or chemogenomic-based inhibition of rumen methanogens were identified. The feasibility of using a synthetic peptide-directed vaccinology approach to target epitopes of methanogen surface proteins was demonstrated. A prophage genome was described and its lytic enzyme, endoisopeptidase PeiR, was shown to lyse M1 cells in pure culture. A predicted stimulation of M1 growth by alcohols was demonstrated and microarray analyses indicated up-regulation of methanogenesis genes during co-culture with a hydrogen (H(2)) producing rumen bacterium. We also report the discovery of non-ribosomal peptide synthetases in M. ruminantium M1, the first reported in archaeal species.

Conclusions/significance: The M1 genome sequence provides new insights into the lifestyle and cellular processes of this important rumen methanogen. It also defines vaccine and chemogenomic targets for broad inhibition of rumen methanogens and represents a significant contribution to worldwide efforts to mitigate ruminant methane emissions and reduce production of anthropogenic greenhouse gases.

Figure 1. Genome atlas of M1.

Single circles in the top-down- outermost-innermost direction are described. Outermost 1^st ring: DBA between the nr database (Ring 3) and dbMethano, a custom methanogen database (Ring 2). Regions in green indicate protein sequences highly conserved between M1 and methanogens but not found in the nr database beyond methanogen genomes. Regions in red indicate protein sequences conserved between M1 and the nr database but not present in other methanogen genomes. 2^nd ring: gapped BlastP results using dbMethano. 3^rd ring: gapped BlastP results using the nr database minus published methanogen genome sequences. In both rings, regions in blue represent unique proteins in M1, whereas highly conserved features are shown in red. The degree of colour saturation corresponds to the level of similarity. 4^th ring: G+C content deviation: green shading highlights low-GC regions, orange shading high-GC islands. Annotation rings 5 and 6 indicate absolute position of functional features as indicated. 7^th ring: ORF orientation. ORFs in sense orientation (ORF+) are shown in blue; ORFs oriented in antisense direction (ORF-) are shown in red. 8^th ring: prediction of membrane bound and cell surface proteins. White: no transmembrane helices (TMH) were identified, Black: ORFs with at least one TMH, Red: ORFs predicted to encompass a signal peptide sequence and Blue: ORFs predicted to incorporate both TMH and a signal peptide sequence. 9^th ring: COG classification. COG families were assembled into 5 major groups: information storage and processing (yellow); cellular processes and signalling (red); metabolism (green); poorly characterized (blue); and ORFs with uncharacterized COGs or no COG assignment (grey). 10^th ring: GC-skew. Innermost ring: genome size (Mb). Selected features representing single ORFs are shown outside of circle 1 with bars indicating their absolute size.

Figure 2. Methanogenesis pathway.

The predicted pathway of methane formation in M1 based on the scheme of Thauer et al. for methanogens without cytochromes is shown. The diagram is divided into three parts to show the capture of reductant, the reduction of CO₂, and conservation of energy at the methyltransfer step. The main reactions are indicated by thick arrows and enzymes catalysing each step are coloured green. Protein subunits coloured red signify the corresponding genes that were up-regulated during co-culture with Butyrivibrio proteoclasticus. Cofactor participation is indicated with thin arrows. For simplicity, protons are not shown and the overall reaction is not balanced. Membrane-located proteins are contained in light brown boxes and potential vaccine and chemogenomic targets are labelled with a circled V or C, respectively. Full gene names and corresponding locus tag numbers can be found in Table S1. H₄MPT; tetrahydromethanopterin; MF, methanofuran; F₄₂₀, coenzyme F₄₂₀ oxidised; F₄₂₀H₂, coenzyme F₄₂₀ reduced; Fd_ox?, unknown oxidised ferredoxin; Fd_red?, unknown reduced ferredoxin; HSCoM, reduced coenzyme M; HSCoB, reduced coenzyme B, CoMS-SCoB, coenzyme B-coenzyme M heterodisulphide; NADP⁺, nicotinamide adenosine dinucleotide phosphate non-reduced; NADPH, nicotinamide adenosine dinucleotide phosphate reduced.

Figure 3. Stimulation of growth of M1 by alcohols.

The inclusion of (A) 20 mM methanol or (B) 5 or 10 mM ethanol when M1 was grown on H₂ resulted in an increase in culture density (measured as OD₆₀₀ nm) compared to cultures grown on H₂ alone. H₂ was added once only, at the time of inoculation, by gassing the cultures with H₂ plus CO₂ (4∶1) to 180 kPa overpressure. Higher concentrations of ethanol (20 mM) resulted in some inhibition of growth (not shown), and there was no stimulation by isopropanol (5 to 20 mM; not shown). No growth occurred when cultures were supplemented with methanol (A), ethanol (B), or isopropanol (not shown) when no H₂ was added, and no methane was formed by those cultures. The symbols in panel are means of 4 replicates, and the thin vertical bars in panel (A) represent one standard error on either side of the mean. Error bars are omitted from panel (B) for the sake of clarity.

Figure 4. Putative cell envelope topography of M1.

Ultrastructural studies of M1 , show that the cell wall is composed of three layers and is comparable to the organization seen in Gram positive bacteria . The three layers can be described as: (1) a thin electron-dense inner layer composed of compacted newly synthesised pseudomurein, (2) a thicker less-electron-dense middle layer which is also composed of pseudomurein, and (3) a rough irregular outer layer that is distal to the pseudomurein layers and assumed to be composed of cell wall glycopolymers, wall-associated proteins and possibly other components. Representative adhesin-like proteins with different cell-anchoring domains are shown. The number of these proteins predicted in the M1 genome is shown in brackets. OT, oligosaccharyl transferase; Sec, Sec protein secretion pathway; PMBR, pseudomurein binding repeat (PF09373); M1-C, M1 adhesin-like protein conserved C-terminal domain.

Figure 5. Observation of interspecies interactions between M1 and B. proteoclasticus B316.

Graph displays growth rate of M1 in co-culture with B316. Microscopic images taken at 2, 8 and 12 h post innoculation of B316 (lighter, rod-shaped organism) into BY⁺ (+0.2% xylan) media containing a mid-exponential M1 culture (darker, short ovoid rod-shaped organism). Growth as determined by Thoma slide enumeration is shown along with sampling time.

Figure 6. M1 peptide vaccine results.

Sheep antibody responses to (A) vaccination with peptides designed against M1 genes (B) binding of antibodies to immobilised M1 cells. In the antibody-binding experiment a negative control (NC) serum from a sheep which had not had colostrum as a lamb was included, as was a sample without added serum which served as a blank, B.

Figure 7. Effect of the lytic enzyme PeiR on M1 growth in vitro.

(A) Addition of PeiR to growing cultures at 73 h resulted in a dramatic drop in culture density, indicative of cell lysis. At a low concentration of PeiR (final concentration of 2.5 mg per litre), the cultures were able to recover, indicated by the increase in culture density after 100 h, and (B) by production of methane at levels similar to that of cultures receiving no PeiR. Addition of higher concentrations of PeiR (7.5 and 22.5 mg per litre) resulted in a lasting effect, with (A) no subsequent recovery of growth and (B) a reduced methane yield. Chloroform, a known potent inhibitor of methanogens, resulted in a similarly reduced methane yield (B), but the decrease in culture density was less (A), as expected since it halts metabolism rather than lysing cells. PeiR was added to 10 ml cultures in 0.1 ml of buffer. The buffer alone had no effect. The symbols (A) and solid bars (B) are means of 3 replicates, and the thin vertical bars represent one standard error on either side of the mean.

Figure 8. Organization of three gene clusters proposed to be involved in M1 NRP biosynthesis.

Figure 9. Chemogenomic and vaccine gene targets within M1.

The number of genes identified by each analysis is shown in the Venn diagram and a selection of the gene targets are summarized in the boxes grouped by functional category. (A) Chemogenomic gene targets were defined by identification of genes that occurred across three separate analyses: the Functional Genome Distribution (FGD), Differential BLAST Analysis (DBA), and metabolic profile of M1 (b) Vaccine target genes were defined as described in Materials and Methods. TMH, transmembrane helices, SP, signal peptide.