Taste of microbes: the terroir explained by rhizospheric microbes

'This carefully designed experimental setup enabled a precise evaluation of how root-associated microbes impact seed flavor while accounting for variations in abiotic factors.' Microbes are everywhere, and virtually all organisms on Earth are living together with these environmental microbes. Plants are no exception and host a rich diversity of microbes on and within their tissues, forming what we call the plant microbiota (Bulgarelli et al., 2013). Over the years, numerous studies have demonstrated that plants and their associated microbiota profoundly influence each other's physiological behavior, and it is now widely recognized that microbiota is an integral part of plants living in their natural environments (Vandenkoornhuyse et al., 2015). Therefore, it is reasonable to assume that microbes have a significant influence on crop flavor, as in the terroir theory. The idea of terroir, historically, was rooted in the experiences of consumers who had tasted a wide variety of wines produced in diverse wineries across the globe, and academic field studies conducted in previous decades have corroborated these notions. For example, modern research has revealed a significant correlation between wine flavor, including its chemical profiles or pH, and abiotic factors, particularly water capacity and climate (Van Leeuwen & Seguin, 2006). However, the key challenges in field studies arise from the fact that 'soil' encompasses a multitude of multicollinear components, including chemical, physical, and biological factors, restricting the chance to uncover any causal relationship within the terroir conceptual framework. Terroir science explores more than just the taste of wines or other fruit-edible crops; it also provides insight into how plants coordinate their internal processes in response to soil and environmental conditions. Root-associated microbes apparently influence various aspects of plant physiology, including primary and specialized metabolism. For example, root colonizing Pseudomonas CH267 strongly induces the production of camalexin in roots, a known phytoalexin produced in a group of plants in Brassicaceae (Koprivova et al., 2019). In addition to eliciting a local response, root-inhabiting Pseudomonas strains systemically activate jasmonic acid pathways in the shoot (Pozo et al., 2008) and trigger the accumulation of immune-related plant specialized metabolites (PSMs), including glucosinolates (van de Mortel et al., 2012). Glucosinolates are a class of sulfur-containing PSMs produced by plants in the order Brassicales (Hopkins et al., 2009), with sinigrin (allyl glucosinolate) being one of the most well-characterized compounds in this group. Glucosinolates are glucose-conjugates whose aglycon is released by a specific class of glucosidases called myrosinases, typically resulting in the formation of isothiocyanates, which contribute to the unique flavor of Brassicaceae plants. Sinigrin, for instance, yields allyl isothiocyanates (also known as mustard oil) upon hydrolysis by myrosinases, which gives mustard seeds their characteristic spiciness. On the other hand, isothiocyanates are also acknowledged as a concern when rapeseed oil is produced from Brassica napus, given its undesired flavor and potentially harmful impact on human health at a higher dose. Breeding efforts have targeted cultivars that produce less sinigrin. Therefore, it is of agricultural and commercial significance to understand how soil microbes affect the accumulation of sinigrin in seeds. Walsh et al. addressed this question by reconstructing root–microbiota interactions using mustard plants grown in a controlled setting. They first prepared an axenic soil matrix by thoroughly autoclaving potting soils. Meanwhile, they harvested soils from five distinct locations with different characteristics, including dry and irrigated pastures, forest soils, as well as grove and brush soils, spanning roughly 100 miles at maximum distance. Microbes of each soil were extracted as 'soil slurry' by mixing soils with a buffer solution, followed by settling down soil particles. Crucially, for each soil slurry, an autoclaved sterile control was prepared to differentiate microbial effects from the effects caused by anything else in the slurries, such as their nutritional and metabolic constituents. Walsh et al. next grew mustard plants under regular external supplementation with respective soil slurries. Once all the plants had produced mature seeds, they harvested these seeds for metabolic profiling and roots and rhizosphere, the soil surrounding roots, for microbial profiling purposes. This carefully designed experimental setup enabled a precise evaluation of how root-associated microbes impact seed flavor while accounting for variations in abiotic factors. In fact, by applying a random forest model to the 16S rRNA and ITS amplicon sequencing data along with the seed sinigrin abundance data, they identified microbial taxa whose relative abundance in the rhizosphere reliably predicted the concentration of sinigrin in seeds. Interestingly, the origin of the slurries did not exhibit a clear impact on sinigrin accumulation in seeds, implying that the rhizosphere microbes influence mustard seed flavor in a manner that is independent of soil characteristics. What makes this study distinctive is that Walsh et al. went beyond merely identifying potential taxa affecting seed flavor; rather, they further aimed to explore the molecular framework underlying the association between rhizosphere microbiota compositions and seed flavor. They performed a shotgun metagenomic analysis to identify genes, rather than microbial taxa, whose normalized abundance in the rhizosphere community correlates with the sinigrin abundance in seeds. Interestingly, although no clear signature was found in glucosinolate-catabolizing genes, they observed that sulfur metabolic genes tend to be more prevalent in the rhizosphere when seeds accumulate higher levels of sinigrin. This apparently makes sense, given that a glucosinolate molecule contains two sulfur atoms, and implies that there is indeed a functional link between rhizospheric microbiota and glucosinolate production in plants. Naturally, the next step should involve testing whether the identified predictive microbial taxa are causally linked to the observed changes in sinigrin accumulation in seeds. This will require the isolation of corresponding microbial strains and subsequent experimental validation in a gnotobiotic inoculation setup to directly evaluate their influence on seed flavor. It is also important to note that the current data represents a snapshot at the end of the plant lifecycle, yet seed sinigrin abundance is likely impacted by root–soil interactions throughout the vegetative and reproductive growth stages. Thus, an intriguing experiment would be to track microbial community dynamics over time and pinpoint the time points at which the root microbial community composition is most predictive of seed sinigrin accumulation. However, the destructive nature of harvesting roots and rhizosphere imposes technical challenges toward this end, and a technical breakthrough that allows for the monitoring of root microbiota structures in a nondestructive, time-resolved manner is highly anticipated. Nevertheless, the study reinforces the growing body of evidence supporting the essential role of PSMs in root–microbiota interactions and, ultimately, in the ecological fitness of plants across generations through root–soil feedback (Nakayasu et al., 2022; Wang et al., 2023). PSM-driven plant flavor not only affects crop quality but also serves as a defense mechanism against pathogenic organisms, positioning microbial terroir as a promising area of research both in ecological and agricultural studies for the next decade. This work was supported by KAKENHI funded by JSPS to RTN (22K21367).