Bacteria form complex and dynamic associations with plants that range from mutually beneficial to commensal or pathogenic (1
) and play key roles in soil quality, host productivity, and host health through direct or indirect mechanisms, such as mineralizing soil organic matter, activating plant defense mechanisms, and even producing antibiotics against phytopathogens (3–6
). Plant-associated bacteria colonize both exterior surfaces (epiphytes) and interior surfaces (endophytes), and the soil surrounding plants is the most likely origin for many of these organisms. Plant soil-microbe interactions are complex, with both plant species and soil type being factors exerting their effect on soil microbial community and so playing key roles in soil pathogen suppression (7
). Plants can be inoculated with growth-promoting bacteria that colonize via roots to confer pathogen resistance or improve productivity (8
). However, when not inoculated, plant roots are known to influence rhizosphere microbial community composition through both root morphology and selective enrichment via root exudation (9
). Building on previous work, Bulgarelli and colleagues (10
) proposed a model of root selection for soil microbiota whereby the composition of the soil microbiome is defined by edaphic factors, followed by substrate-driven community selection within the rhizosphere, and finally by host genotype-dependent tuning of endophytic bacteria that colonize roots and, eventually, leaves and reproductive structures. In potatoes and poplars, the rhizosphere and the root microbial composition are cultivar specific and developmental stage dependent, while the microbiome of Arabidopsis spp. is mostly influenced by the edaphic and physical properties of soil (11–13
). Aboveground epiphytic microbiota are mostly selected by nutrient-poor conditions and highly variable conditions of temperature, humidity, and UV radiation intensity (14
) of the plant environment, and yet some leaf epiphytes have plant species-specific associations that change little with geographic location (16
). Elucidating the origin of bacteria that colonize crops and, hence, shape the disease resistance, productivity, and stress tolerance of their hosts may provide biological targets for future biotechnological application. However, model crop systems, such as grapevines, also provide an opportunity to unravel the complex influence of soil type, edaphic factors, climate, plant variety, and vineyard management practices on bacterial colonization and stability within the crop. Understanding these ecological principles is vital if we are to translate this knowledge into practical application.
Viticulturists rely on the assumption that the land from which grapes are grown imparts a unique quality to the wine, called the terroir. The physiological response of the vines to soil type, climatic conditions, and vineyard characteristics and management is believed to give rise to particular organoleptic properties of the wine defining its terroir.
The existence and impact of region-
specific microbiota in defining the terroir have been more controversial, as some studies have been inconclusive (17
), while other studies found evidence for a biogeographic pattern (see, e.g., references 18
). Microbial spatial pattern analyses have mainly been focused on studying the grape microbiota, but the relative importance of the microbiome of local soils in defining the terroir, which is also influenced by local climate and vineyard characteristics (20
), is not known. The potential importance of bacteria and fungi associated with soils and growing plant organs has been characterized (21
), but to date, research has mostly focused on the fermentative stages whereby fungi (17
) and bacteria may have a direct influence on the wine.
In the present study, we characterized the bacteria associated with Merlot grapevines using 16S rRNA amplicon sequencing and shotgun metagenomics to disentangle the influences of host cultivar, soil edaphic parameters, and grapevine developmental stage (dormancy, flowering, and preharvest) on the bacterial community composition, structure, and function from three grapevine parts (roots, leaves, and grapes or flowers) and associated soil (bulk soil and root zone soil). Specifically, we aimed to determine whether the soil was the main source of bacteria colonizing these grapevines and how edaphic factors, climate, and vineyard characteristics might influence the phylogeny and functional potential of this source.
Plant parts maintain distinct microbial communities, with sample type explaining the greatest proportion of variation in microbial community structure of all factors. Overall, Merlot microbial diversity was greater in belowground (roots, root zone, and bulk soil) than aboveground (leaves, flowers, and grapes) samples, as described for other plant systems (28
). This decrease in diversity may be due to the extremes of temperature, UV exposure, and humidity experienced by epiphytes on aboveground organs (15
). Species of Pseudomonas and Sphingomonas taxa found in aboveground plant parts could influence plant health and productivity (30
), while Methylobacterium spp. found in leaves and grapes may stimulate plant development through phytohormone production (31
). Sphingomonas and Methylobacterium spp. can survive the wine fermentation process (27
), and yet their impacts on wine organoleptic properties remain unknown. Steroidobacter spp., which were found in higher abundance in roots than in soil, may play an essential role in plant development and physiology, as brassinosteroids have been shown to control seed germination, stem and root elongation, vascular differentiation, fruit ripening, leaf expansion, and stress protection in plants (32
). The enrichment of Actinobacteria
, and Proteobacteria
spp. in roots (Fig. 2B
) is likely driven by chemoattraction via photoassimilates secreted by root cells (10
). The taxa enriched in root samples (Xanthomonadales
, Bradyrhizobium, Saprospirales
, and Actinomycetales
spp.) comprised 64% of the core OTUs (OTUs present in ≥95% of samples), which suggests that they outcompeted other bacteria for colonization of the roots or that the plant selects for these taxa or both. For example, the enrichment of roots in taxa belonging to the Rhizobiales
order, such as Bradyrhizobium spp. (both part of the root core microbiota), may result in N fixation and antibiotic production (33
) that would promote plant growth and disease suppression.
Microbial metabolic capabilities enriched in roots compared to soils included traits previously identified as favoring the movement or attraction of bacteria toward root exudates (10
) and genes encoding the metabolism of macro- and micronutrients (K, S, and Fe) as well as plant stress tolerance. Iron acquisition is both a known virulence factor (35
) and a potential pathogen biocontrol signature, as nonpathogenic bacteria producing siderophores reduce iron availability, restricting pathogen growth (3
); therefore, it may benefit the plant to enhance siderophore production in the root and root zone.
Whereas plant endophytic colonization from soil microbiota is a quite well-established concept, the origin of plant epiphytes is believed to be more complex. In the present study, where epiphytes and endophytes were analyzed simultaneously, soil was more diverse and contained the majority of OTUs found in aboveground samples, suggesting that soil serves as a primary reservoir for potential plant-associated bacteria. Regarding endophytes, Rhizobium, Staphylococcus, and Agrobacterium spp. were observed in grapes and leaves but were not previously observed as epiphytes in the same plant parts of Merlot vines (22
), suggesting that they likely originated in the soil and migrated through roots to aerial parts, where they flourished because of their potential adaptation to improved nutrient supply and reduced bacterium-bacterium competition due to the selective environment. In concordance, previous studies on grapevine endophytes demonstrated the migration of endophyte Burkholderia phytofirmans
strain PsJN from the rhizosphere to young grapes of cutting systems (37
). In addition, it is likely that the rootstock affects further endophytic colonization, as it has been shown to play a key role in pathogen resistance, and it may be expected that different rootstocks differentially select microbes from the surrounding soil. Further research is needed to explore the effects of rootstock selection on the final microbiota of grapevine tissues.
Referring to epiphytes, soil has been previously proposed as a potential microbial inoculum for grapevine aboveground organs. Martins and colleagues (22
) showed that some epiphytic bacteria were common among aerial plant parts and soil, which led them to propose that the physical proximity between soil and grapevine parts might facilitate microbial migration through rain splash, high winds, insects, etc. In addition, Compant and colleagues (21
) demonstrated that aboveground organs may also be colonized by epiphytes from other plant surfaces (38
), which may explain the presence of species of taxa that were abundant and unique to aboveground grapevine parts in this study, e.g., Actinobacillus spp. were present in leaves and grapes but not in the rest of the sample types. Furthermore, while endophytes could colonize surface tissues, epiphytes could establish as endophytes through wounds and natural openings, suggesting that epiphytes and endophytes are part of a continuum (39
). Thus, none of these means of microbial colonization of plant organs are mutually exclusive and many or all of them may act in concert to influence the final grapevine organ microbial community structure.
Biogeographic trends in bacterial community structure and composition were observed only in bulk soil. Bacterial and fungal communities associated with Chardonnay and Cabernet Sauvignon musts have previously been shown to exhibit significant biogeography (19
). In the current study, however, edaphic factors within vineyards outweighed biogeographic trends, suggesting that soil bacterial assemblages have extensive local heterogeneity even across small distances, and soils with similar edaphic variables and homogenous climatic conditions. This local heterogeneity could necessitate highly specific regionalization of vineyard management practices to support beneficial bacterial taxa, both above ground (e.g., Pseudomonas, Sphingomonas, and Methylobacterium spp.) and below ground (e.g., Rhizobiaceae
and Bradyrhizobium spp.). The vineyard-associated localization of bacterial taxa has implications for viticulturists who rely on the assumption that the “land” imparts a unique quality to the wine specific to that growing site, called the terroir, but to date, only grapes and must have been implicated. These results suggest that the local biogeography of the soil microbiota may also have an indirect influence.
The origin of the microbes in wine ferments is still poorly understood, but, when not purposely inoculated, they are commonly assumed to come from the grapes themselves (41
); however, some particular microbes might come from the winery equipment and barrels as well (42
). The present study results evidenced that microbial community structures were most similar between grape and must samples, suggesting that the community present on grapes prior to fermentation remains relatively stable or, at least, more stable than differences among plant organs. Despite this, in comparisons of the phylotypes shared among the different sample types, California must samples shared more species of bacterial taxa with the soil from this study than with the grape samples. A plausible explanation might be related to the harvest practices and transport procedures, where harvested grapes ultimately forming part of the must might have received a significant input of microorganism phylotypes from soil when manually harvested, as these grapes are usually stored in boxes located on the ground. Mechanical harvesting is not exempt from this, as the machinery generates dust from the ground, which could colonize grapes. Martins and colleagues (22
) proposed a similar idea, suggesting that dust generated in tillage (commonly used to control weeds in vineyards) might contribute to the migration of microorganisms from soil to aerial parts, where they would establish as grape epiphytes. All these results imply that soil microbiota may have a more direct role in winemaking and that soil microbiota are potentially involved in the formation of the regional wine terroir.
In conclusion, microbial community structures were significantly different between plant parts and soils, and yet aboveground samples shared more OTUs with belowground environments than they did with each other, suggesting that soil is a major microbial reservoir. Differences in the soil bacterial communities of different vineyards are reflected in the roots, with the abundances of species of several taxa showing specific vineyard regionalization. This suggests that, even within a single region, there can be significant variation in the community structure in both soils and vines, pointing to a potential microbial seed bank in soils that is differentially selected for by the Merlot vine and that might shape the microbial assemblages of aboveground organs, such as grapes, thereby indirectly influencing wine characteristics.