Numerous studies have addressed the effects of soil type and plant genetic make-up (species, genotype or cultivar) on the assembly of indigenous root-associated bacterial communities (Badri & Vivanco, 2009; Aira et al., 2010; Weinert et al., 2010; Bulgarelli et al., 2012; Lundberg et al., 2012); still, it remains unclear which factor (genotype or soil) dominates the assembly process (Berg & Smalla, 2009). While plant species effects are substantial, effects of transgenic modifications in crops, which, for example, facilitate the release of certain substances into the rhizosphere (e.g. phytases, which allow for better phosphate uptake) or the
accumulation of molecules in the root (e.g. zeaxanthin, a plant pigment, which is target for enhanced production in crops to improve human visual capacity) are marginal and comparable to cultivar-specific differences (George et al., 2009; Weinert et al., 2009; Chun et al., 2011). On the other hand, the salicylic acid and JA signaling pathways have been reported to shape the indigenous leaf and rhizosphere microbial community of A. thaliana plants (Kniskern et al., 2007; Hein et al., 2008). Another study, using the model plant A. thaliana and a transgenic line overexpressing defensive glucosinolates in roots, reported effects on the microbial rhizosphere composition (Bressan et al., 2009). From these reports, it appears that manipulating evolutionary and ecologically relevant traits in plant defense affects the structure of plant microbial communities, while the expression of “novel”, evolutionarily irrelevant compounds does not affect this assembly.
In our study (Manuscript I), we assessed the root microbial communities of three different N. attenuata genotypes grown in four different native soil types. Specifically, we were interested in the role of ET signaling in the recruitment of endophytic bacteria, since endophytes are thought to establish a more stable interplay with their host (Hardoim et al., 2008). The gaseous plant hormone ET has been shown to play an important role in plant-pathogen interactions (van Loon et al., 2006), but effects of ET signaling on a plant’s indigenous microbial community remained so far unknown.
Signaling effects are best studied using transgenic lines, which vary only in the expression of the gene of interest (although pleiotropic effects can occur) (Bergelson et al., 1996). Hence, an untransformed wild type (WT) control, one transgenic line impaired in ET biosynthesis (ir-aco1) and one impaired in ET perception (35S-etr1) were used. Our study revealed that the soil type the plants were cultured in, generally determined the microbial assembly in the root endosphere. Although we also observed genotype-dependent changes in diversity, namely that ET signaling- impaired plants harbored a less diverse microbial community than WT, these were marginal. Hence, we concluded that soil factors are more important in shaping N. attenuata’s root endosphere bacterial community than ET signaling. Furthermore, great variability in total colonization and species composition across the samples supports the hypothesis that stochastic processes prevail in community assembly, as discussed by Hardoim et al. (2008).
Recent advances in high-throughput sequencing technologies revolutionized research in plant-microbe interactions, and new technologies like pyrosequencing might be applied to obtain a more complete view of the ET-associated microbial community assembly, including non- culturable bacteria and fungi. The approach we followed (extracting culturable bacteria and assaying them on one culture medium) selects for fast-growing, easy-to-culture and very abundant bacterial isolates (i.e. Bacillus and Pseudomonas sp.), and draws an incomplete picture. Hence, conclusions about the magnitude of ET effects, we reasoned, should be treated with caution. Nevertheless, a recent study by Doornbos et al. (2010), using ET-insensitive tobacco plants (Tetr18) and a culture-independent approach, came to a similar finding: the plant’s ET- insensitivity had only marginal effects on the rhizosphere microbial community composition. They attributed the minor changes in community structure to an increased susceptibility of Tetr18 plants to opportunistic microbes than to impairments in ET-signaling (Doornbos et al., 2010). Furthermore, we found, as other studies before, that the overexpression of a mutated version of
the A. thaliana ET receptor (AtETR1) resulted in pleiotropic effects, giving rise to a severe root phenotype (lack of root hairs, few lateral braches, lack of gravitropism) (Luschnig et al., 1998; Clark et al., 1999), which might in turn affect the microbial communities to a greater extent than ET signaling itself. Innovative techniques, like inducible and transient manipulation of gene expression, might help to separate ET signaling from the pleiotropic effects of the mutant receptor (Koo et al., 2009).
Although ET-signaling-mediated changes in bacterial species diversity were marginal (Manuscript I), they might nevertheless impact plant health, as it is often the case for the occurrence of a single pathogenic microbe or the inoculation of a PGP strain. Several studies working with single microbial isolates reported on the importance of ET in modulating mutualistic plant-microbe interactions; and the hormone was shown to determine the balance between mutualism and pathogenicity (Dong et al., 2003; Camehl et al., 2010). Performing in
vitro inoculation assays using selected bacterial isolates, we confirmed plant genotype-specific
colonization by Pseudomonas thivervalensis, which exclusively entered the ET-insensitive line
35S-etr1 (Manuscript I). We infer that changes the root biochemistry, e.g. in the metabolome or
root exudates of 35S-etr1 plants might promote this specific interaction. Similar effects have been shown before by Oger et al. (2004). They found that the production of opines in transgenic Lotus
corniculatus roots led to 104-fold enrichment of opine-utilizing bacterial isolates (e.g.
Agrobacterium sp.) in the rhizosphere. Ongoing research on N. attenuata root exudates suggests
that ET signaling indeed affects root exudate composition and hence might facilitate these specific plant genotype-microbe interactions (M. Bonilla, unpublished results).
From our findings and the related literature, it becomes clear that the role of ET in mutualistic plant-microbe is not well understood. In-depth analyses using state-of-the-art technologies, like culture-independent analyses of plant microbial communities as well as innovative plant genetic manipulation strategies, should be employed to disentangle ET-mediated effects at the microbial community level as well as specific (single strain) plant-microbe interactions.