Comparative metatranscriptomics reveals kingdom level changes in the rhizosphere microbiome of plants

The ISME Journal (2013) 7, 2248–2258 & 2013 International Society for Microbial Ecology All rights reserved 1751-7362/13 www.nature.com/ismej ORIGINAL ARTICLE Comparative metatranscriptomics reveals kingdom level changes in the rhizosphere microbiome of plants Thomas R Turner1,5, Karunakaran Ramakrishnan1, John Walshaw2, Darren Heavens3, Mark Alston3, David Swarbreck3, Anne Osbourn4, Alastair Grant5 and Philip S Poole1,5 1 Department of Molecular Microbiology, John Innes Centre, Norwich, UK; 2Institute of Food Research, Norwich, UK; 3The Genome Analysis Centre, Norwich, UK; 4Department of Metabolic Biology, John Innes Centre, Norwich, UK and 5Earth Life Systems Alliance, School of Environmental Sciences, University of East Anglia, Norwich, UK Plant–microbe interactions in the rhizosphere have important roles in biogeochemical cycling, and maintenance of plant health and productivity, yet remain poorly understood. Using RNA-based metatranscriptomics, the global active microbiomes were analysed in soil and rhizospheres of wheat, oat, pea and an oat mutant (sad1) deficient in production of anti-fungal avenacins. Rhizosphere microbiomes differed from bulk soil and between plant species. Pea (a legume) had a much stronger effect on the rhizosphere than wheat and oat (cereals), resulting in a dramatically different rhizosphere community. The relative abundance of eukaryotes in the oat and pea rhizospheres was more than fivefold higher than in the wheat rhizosphere or bulk soil. Nematodes and bacterivorous protozoa were enriched in all rhizospheres, whereas the pea rhizosphere was highly enriched for fungi. Metabolic capabilities for rhizosphere colonisation were selected, including cellulose degradation (cereals), H2 oxidation (pea) and methylotrophy (all plants). Avenacins had little effect on the prokaryotic community of oat, but the eukaryotic community was strongly altered in the sad1 mutant, suggesting that avenacins have a broader role than protecting from fungal pathogens. Profiling microbial communities with metatranscriptomics allows comparison of relative abundance, from multiple samples, across all domains of life, without polymerase chain reaction bias. This revealed profound differences in the rhizosphere microbiome, particularly at the kingdom level between plants. The ISME Journal (2013) 7, 2248–2258; doi:10.1038/ismej.2013.119; published online 18 July 2013 Subject Category: Microbial population and community ecology Keywords: rhizosphere; metatranscriptomics; microbiome; wheat; oat; pea Introduction Interactions between plants and microbes in the rhizosphere are of global importance to biogeochemical cycling (Philippot et al., 2009), plant health and productivity (Bloemberg and Lugtenberg, 2001). Colonisation of the rhizosphere, the region of soil influenced by plant roots, is necessary for both plant pathogens and plant growth-promoting rhizobacteria. The latter aid plants by providing nutrients, modulating growth, defending against diseases (Lugtenberg and Kamilova, 2009) and contributing to disease-suppressive soils (Mendes et al., 2011). Many plant-associated microbes are known and Correspondence: PS Poole, Department of Molecular Microbiology, John Innes Center, Norwich Research Park, Norwich, Norfolk NR93AD, UK. E-mail: Received 22 March 2013; revised 10 May 2013; accepted 21 May 2013; published online 18 July 2013 well-studied, including the symbiotic nitrogenfixing Rhizobium leguminosarum (Young et al., 2006), and both beneficial and pathogenic Pseudomonas spp. (Feil et al., 2005; Paulsen et al., 2005). Also, association of mycorrhizal fungi with most land plants is fundamental to acquisition of mineral nutrients such as phosphate (Bonfante, 2010). However, little is known about how these organisms interact at the community level. Every gram of soil is estimated to contain in excess of 50 000 species of bacteria (Roesch et al., 2007), the vast majority of which are uncultured (Handelsman, 2004). Sequencing of polymerase chain reaction (PCR)-amplified 16S rDNA has been extensively used to examine rhizosphere bacterial communities of various plants, and recently highthroughput pyrosequencing (Margulies et al., 2005) has revealed these communities in previously unobtainable detail. Plants studied include the important crop potato (Solanum tuberosum) Rhizosphere communities of crop plants TR Turner et al 2249 (Inceoglu et al., 2011), plants of the Antarctic (Teixeira et al., 2010) and recently the model dicot Arabidopsis thaliana (Bulgarelli et al., 2012; Lundberg et al., 2012). However, PCR amplification of genomic DNA is inherently biased by primer design (Hong et al., 2009; Pinto and Raskin, 2012) and is limited to the targeted division of life (Bacteria, Archaea, Eukarya or a smaller taxonomic group). Studies to date have largely focused on either bacteria or fungi, often neglecting other eukaryotes and archaea. High-throughput sequencing has also enabled the use of metagenomic strategies where total genomic DNA from the rhizosphere is sequenced (Tett et al., 2012). While encompassing all domains of life and indicating the metabolic potential of a microbiome, a metagenome contains relatively few rRNA genes, reducing the strength of taxonomic assignments. Metatranscriptomics, where total RNA from the environment is sequenced, reveals active community members and metabolic pathways (Urich et al., 2008). Many applications of metatranscriptomics are focussed on the latter but have a significant challenge in the requirement for enrichment of mRNA (Stewart et al., 2010; Yi et al., 2011). However, the dominance of rRNA in a metatranscriptomic sample allows robust assessment of the entire microbiome, without prior selection of taxonomic groups for study. This is technically much less challenging than enrichment of mRNA, avoids PCR bias and can be carried out straightforwardly on multiple samples. In this study, comparative metatranscriptomics was used to study the rhizosphere microbiomes of three crop plants grown in the same soil: wheat (Triticum aestivum), a major world food staple; oat (Avena strigosa), a cereal that produces anti-fungal avenacins (Maizel et al., 1964); and pea (Pisum sativum), a widely grown crop legume nodulated by N2-fixing bacteria R. leguminosarum. In addition, we compared the rhizosphere microbiome of the wild-type oat with that of an avenacin-deficient mutant, sad1 (Haralampidis et al., 2001). Avenacins are triterpenoid saponins that protect oat from root pathogens (Papadopoulou et al., 1999) including Gaeumannomyces graminis, the causative agent of take-all (Osbourn et al., 1994). Although metatranscriptomic analysis of total rRNA has been used to profile microbial communities in soil (Urich et al., 2008) and oceans (Ottesen et al., 2011; Shi et al., 2011), this is, to our knowledge, the first application of it to study the rhizosphere microbiomes of important crop plants. Materials and methods Plant growth and samplin (...truncated)