Isolation of Succinivibrionaceae Implicated in Low Methane Emissions from Tammar Wallabies

See allHide authors and affiliations

Science  29 Jul 2011:
Vol. 333, Issue 6042, pp. 646-648
DOI: 10.1126/science.1205760


The Tammar wallaby (Macropus eugenii) harbors unique gut bacteria and produces only one-fifth the amount of methane produced by ruminants per unit of digestible energy intake. We have isolated a dominant bacterial species (WG-1) from the wallaby microbiota affiliated with the family Succinivibrionaceae and implicated in lower methane emissions from starch-containing diets. This was achieved by using a partial reconstruction of the bacterium’s metabolism from binned metagenomic data (nitrogen and carbohydrate utilization pathways and antibiotic resistance) to devise cultivation-based strategies that produced axenic WG-1 cultures. Pure-culture studies confirm that the bacterium is capnophilic and produces succinate, further explaining a microbiological basis for lower methane emissions from macropodids. This knowledge also provides new strategic targets for redirecting fermentation and reducing methane production in livestock.

Kangaroos and wallabies possess adaptations to herbivory that include a forestomach supporting a cooperative host-microbe association that releases nutrients from plant biomass. The structure-function relationships inherent to these microbiomes have only recently been examined, in part because of their capacity for lignocellulose degradation coupled with low methane production, relative to domesticated livestock (1, 2). The differences in methane emissions between the macropodids and ruminants may be partially explained by the anatomical differences between the host animals, with the macropodid digestive system resulting in shorter retention time of particulate digesta within the foregut, which might prevent the establishment of methanogenic archaea (3). Methanogenic archaea have since been found in the foregut microbiomes of the Tammar wallaby (Macropus eugenii) and the eastern grey kangaroo (M. giganteus), albeit at lower levels than those found in ruminants (4). Functional gene analyses and limited cultivation studies have also suggested the presence of clades of acetogenic bacteria in the macropodid foregut (5, 6). Metagenomic analysis revealed that the foregut microbiome of the Tammar wallaby varies seasonally but is principally composed of bacteria belonging to the phyla Firmicutes, Bacteroidetes, and Proteobacteria, with the majority of the observed phylotypes only distantly related to cultivated species (7). Approximately 77% of the recovered Proteobacteria sequences [representing 9% of all sequences recovered in the 16S ribosomal RNA (rRNA) clone library] were assigned to just two deeply branched operational taxonomic units within the family Succinivibrionaceae; hereafter referred to as Wallaby Group 1 (WG-1; table S1 and fig. S1).

The closest cultured relatives of WG-1 all belong to the family Succinivibrionaceae, which includes Succinivibrio, Ruminobacter, and Anaerobiospirillum spp., although WG-1 does not share more than 93% sequence identity to the 16S rRNA genes of any of these described species. Members of the Succinivibrionaceae produce succinate as their principal fermentation end product, and for some exogenous sources of hydrogen stimulate succinate formation (8). If WG-1 possesses the same metabolic characteristics, the bacterium might underpin a fermentation scheme that helps to explain the low methane phenotype attributed to the macropodids. To examine the contributions of WG-1 to fermentation in the macropodid foregut requires its isolation, preferably in axenic culture. A combination of data sets (7) and a nucleotide composition–based binning algorithm (PhyloPythia) was used to isolate and characterize a strain representing the WG-1 lineage. PhyloPythia uses a multiclass support vector machine classifier for the taxonomic assignment of variable-length metagenome sequence fragments based on their oligonucleotide compositions (9).

The 2-Mb WG-1 composite genome (WG-1.M: 366 contigs) was analyzed using the IMG/M (10), KEGG (11), and RAST (12) databases for partial metabolic reconstruction and prediction of some of its functional capabilities (Fig. 1 and table S2). For example, like Ruminobacter amylophilus (13), WG-1.M was predicted to use starch as a carbohydrate source; three genes encoding putative GH13 alpha-amylase genes, as well as genes encoding putative glucose and maltose transporters, were identified (table S2). There was no evidence for utilization of exogenous sucrose, fructose, mannose, cellulose, cellobiose, xylose, pectin, or sorbose. The affiliation of WG-1 with the Succinivibrionaceae was confirmed with predicted genes supporting a branched, rather than complete, tricarboxylic acid cycle, the hallmark of this family. Additionally, genes encoding a phosphoenolpyruvate (PEP) carboxykinase, pyruvate-formate lyase (PFL), and acetate kinase (AK) were found (Fig. 1 and table S2). Based on these findings, we predicted that WG-1 would be similar to Anaerobiospirillum succiniciproducens (14) in that the bacterium employs an anaplerotic reaction to produce oxaloacetate from PEP, with the subsequent reduction of oxaloacetate to succinate as a principal fermentation end product. Substrate-level phosphorylation would be supported via PFL and AK. The WG-1.M assembly also included a urease gene cluster encoding all 13 genes required for urea transport and catabolism (fig. S2). Endogenous sources of urea are well recognized as a core nonprotein nitrogen source for many ungulates, enhancing nitrogen retention in animals that consume plant biomass of low protein content. Indeed, previous nutritional and physiological studies of the Tammar wallaby have suggested that as much as 84% of endogenous urea is returned via saliva and the bloodstream to the animal’s foregut (15).

Fig. 1

Selected metabolic features of the WG-1 phylogroup as inferred from genome comparisons. The assumption that phosphenolpyruvate (PEP) serves as the branch point in WG-1 to the formate-, acetate-, and lactate-producing C3 pathway and the succinate-producing C4 pathway is based on data from McKinlay et al. (21). Broken border lines indicate annotations identified only in the WG-1 isolate genome sequence. Abbreviations are as follows: AcCoA, acetyl-coenzyme A; AK, acetate kinase; APS, adenylylsulfate; CL, citrate lyase; E4P, erythrose-4-phosphate; F6P, fructose-6-phosphate; Fm, fumarase; FR, fumarate reductase; G3P, glyceraldehyde-3-phosphate; G6P, glucose-6-phosphate; ICL, isocitrate lyase; MDH, malate dehydrogenase; Menz, malic enzymes; OAA, oxaloacetate; PEPCK, PEP carboxykinase; PFL, pyruvate formate-lyase; PK, pyruvate kinase; PTS, phosphotransferase system; R5P, pentose-phosphates; S7P, sedoheptulose-7-phosphate. Gene identification numbers (Integrated Microbial Genomics gene object identifiers) can be found in table S2.

The partial metabolic reconstruction of WG-1.M was used to develop a strategy for the enrichment and axenic cultivation of WG-1. We first tried peptone-yeast-starch (PYS) broth (13), aiming to produce a culture enriched with WG-1, but quantitative PCR showed low numbers of the target bacterium (<0.1% total). Using the knowledge of the urease gene cluster and alpha-amylases in the WG-1.M, we developed a defined medium (TGFI; table S3), which contained starch and urea as the sole carbohydrate and nitrogen sources, respectively, as well as the antibiotic bacitracin, because a putative bacitracin resistance gene was identified in the WG-1.M assembly. Digesta samples were used to inoculate TFGI media, and subsequent WG-1 specific quantitative PCR (qPCR) screens indicated that the TFGI enrichment became dominated by WG-1 (>50% total population). We predicted that this medium would specifically favor the growth of WG-1, so we used semicontinuous batch culture methods, which qPCR indicated was axenic. Serial dilutions of the culture were then plated onto TGFI-agar plates; single colonies were picked and used to start fresh cultures. These cultures were used to produce PCR amplicons of the 16S rRNA gene, which were sequenced and confirmed to be 100% identical to the WG-1 phlyogroup identified from the original 16S rRNA gene libraries and from the WG-1.M data set (fig. S1). Oligonucleotide primers specific for the WG-1 16S rRNA gene were used in PCR reactions and only produced a strong amplicon from the Tammar wallaby digesta samples and not from digesta samples from other animals, including western grey kangaroos (M. fuliginosus) harvested from drought-affected land in western Australia (table S1). These latter results are not surprising given the specificity of the primers for WG-1, its preferred carbohydrate source, and the nutritional ecology of these animals at the time of sampling.

Carbohydrate utilization of the WG-1 isolate was consistent with the metabolic predictions of WG-1.M (table S4). We also monitored growth and fermentation when the bacterium was cultured with different headspace gases. There was no measurable growth of WG-1 cultures with N2 alone, suggesting that hexoses are necessary but not sufficient to support WG-1 energy transduction and growth. The addition of CO2 to the headspace gas (N2) did result in apparently concentration-dependent growth (fig. S3). Succinate, formate, and acetate were the principal fermentation products, and CO2 concentration appeared to have little effect on the molar ratio produced (succinate:formate:acetate being 1.5:1:1). Replacing some of the headspace gas with H2 neither stimulated nor inhibited the growth of WG-1 and had no measurable effects on the fermentation profile (fig. S3). These results suggest that WG-1 is a capnophilic (CO2-loving) bacterium, being dependent on CO2 to support its metabolism via succinate biosynthesis. The protons and electrons required for the reductive steps leading to succinate biosynthesis are derived from hexose catabolism via the Embden-Meyerhof-Parnas pathway.

The genome of the WG-1 isolate was sequenced and assembled into 43 scaffolds comprising 2.9 Mbp of unique sequence. This draft genome assembly (WG-1.G) was directly compared with the WG-1.M assembly, providing a direct evaluation of the phylogenetic binning of metagenomic data. Using NUCmer (16), we found that almost 90% of WG-1.M reads and assemblies were shared with WG-1.G, with an average percent identity of 98.9% (Table 1; see also fig. S4 and table S5). Despite the high sequence similarity, examples of breakpoints within the alignment of the two data sets were observed (fig. S4 and table S5), suggesting instances of misassembly and/or misclassification of PhyloPythia binning. Each of the 1657 open reading frames (ORFs) identified from the WG-1.M data was compared to the 2403 putative ORFs obtained from the WG-1.G data set using RAST (12). About 75% of the WG-1.M–derived ORFs were >99% identical with the corresponding ORF in WG-1.G. However, ~10% of WG-1.M–derived ORFs had no homologs in the WG-1.G draft genome (Table 1 and fig. S5). There are two possible explanations for this finding: Either ~10% of the WG-1.M ORFs are false positives and not present in the genome, or there are several gaps in the draft WG-1.G data set. However, 95% of the Pfam and RAST subsystems (the collection of functional roles that make up a metabolic pathway, complex, or class of proteins) identified in WG-1.M were also represented in WG-1.G (Table 1), and the two data sets also possess a high degree of synteny (table S5). For these reasons, we believe that PhyloPythia-directed binning of WG-1.M data produced a detailed and accurate representation of the genome of a previously uncultured bacterium, which ultimately contributed to its successful axenic culture.

Table 1

Summary and comparison of genomic data recovered from the WG1 metagenome assemblage (WG-1.M) and WG-1 genome (WG-1.G).

View this table:

There are now numerous metagenomic data sets reported in the literature, but the use of those data to facilitate the targeted isolation and cultivation of a component species has only been reported in two instances to our knowledge, both from low-complexity communities composed of only a few prokaryotic species (17, 18). In the present study, a composite genome was recovered from a microbiome with a much higher degree of complexity [approximately 500 species (7)]. Nucleotide binning methods such as PhyloPythia and PhyloPythiaS (19) will play a critical role in future (meta)genomic-directed isolation studies (“reverse metagenomics”), including microbial communities that exhibit very complex species diversity; provided that genomic fragments of the targeted population can be unambiguously identified and that they comprise genes which can be leveraged for selective cultivation strategies.

The genomic and physiological characterization of WG-1 further distinguishes the attributes of the foregut microbiomes of Australian macropodids relative to other herbivores, and further implicates the foregut microbiota as a contributing factor to why these animals are low methane emitters as compared with ruminants. Although the abundance of WG-1 is variable in samples collected from animals in winter and spring (7), our results show that these bacteria will be numerically dominant when the plane of nutrition is rich in starches and soluble sugars. An environment favoring large numbers of WG-1 would not only contribute to substrate oxidations and reductions remaining closely coupled, with little methane being formed; it would also ensure that more digestible energy is available to the host animal for nutrition. Such metabolism is not unlike what has been predicted for ruminant livestock fed ionophores, such as monensin (20). For these reasons, studies of WG-1 and the Succinivibrionaceae in general should be intensified in pursuit of new strategies to redirect fermentation in livestock toward pathways that do not favor methanogenesis.

Supporting Online Material

Materials and Methods

Figs. S1 to S5

Tables S1 to S5

References (2246)

References and Notes

  1. Acknowledgments: The Tammar wallaby project is partially supported by the Commonwealth Scientific and Industrial Research Organisation’s (CSIRO’s) Office of the Chief Executive (OCE) Science Leader and Transformational Biology Capability Platform grant programs (awarded to M.M.), a CSIRO OCE Postdoctoral Fellowship (awarded to P.B.P.), and the U.S. Department of Energy–Joint Genome Institute Community Sequencing Program. We are especially grateful for the support from L. Hinds (CSIRO Ecosystem Sciences), A.-D. Wright (University of Vermont), and P. Evans (CSIRO Livestock Industries), who assisted in sample collections. The work conducted by the U.S. Department of Energy–Joint Genome Institute is supported by the Office of Science of the U.S. Department of Energy under contract no. DE-AC02-05CH11231. The whole-metagenome shotgun project and whole-genome shotgun project (for the WG-1 isolate) has been deposited at DDBJ/EMBL/GenBank under accession numbers ADGC00000000 and AFAK00000000, respectively. 16S rRNA gene sequences are deposited under accession numbers GQ358225 to GQ358517.
View Abstract

Navigate This Article