Potato Root-Associated Microbiomes Adapt to Combined Water and Nutrient Limitation and have a Plant Genotype-Specific Role for Plant Stress Mitigation

doi:10.21203/rs.3.rs-492999/v1

Download PDF

Research Article

Potato Root-Associated Microbiomes Adapt to Combined Water and Nutrient Limitation and have a Plant Genotype-Specific Role for Plant Stress Mitigation

https://doi.org/10.21203/rs.3.rs-492999/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Due to climate change and reduced use of fertilizers combined stress scenarios are becoming increasingly frequent in crop production. In a field experiment we tested the effect of combined water and phosphorus limitation on the growth performance and plant traits of eight tetraploid and two diploid potato varieties as well as on root-associated microbiome diversity and functional potential. Microbiome and metagenome analysis targeted the diversity and potential functions of prokaryotes, fungi, plasmids and bacteriophages and was linked to plant traits like tuber yield or timing of canopy closure.

Results

The different potato genotypes responded differently to the combined stress and hosted distinct microbiota in the rhizosphere and the root endosphere. Proximity to the root, stress and potato genotype had significant effects on bacteria, whereas fungi were only mildly affected. To address the involvement of microbial functions, we investigated well and poorly performing potato genotypes (Stirling and Desirée, respectively) under stress conditions and executed a metagenome analysis of rhizosphere microbiomes subjected to stress and no stress conditions. Functions like ROS detoxification, aromatic amino acid and terpene metabolism were enriched and in synchrony with the metabolism of stressed plants. In Desirée, Pseudonocardiales had the genetic potential to take up assimilates produced in the fast-growing canopy and to reduce plant stress-sensing by degrading ethylene but overall yield losses were high. In Stirling, Xanthomonadales had the genetic potential to reduce oxidative stress and to produce biofilms, potentially around roots. Biofilm formation could be involved in drought resilience of Stirling and explain the recorded low yield losses. In the stressed rhizosphere the relative abundance of plasmids was reduced and the diversity of phages was enriched. Moreover, mobile elements like plasmids and phages were affected by combined stresses in a genotype-specific manner.

Conclusion

Our study gives new insights into the interconnectedness of root-associated microbiota and plant stress responses in the field. Functional genes in the metagenome, phylogenetic composition and mobile elements play a role in potato-stress adaption. In a poor and a well performing potato genotype grown under stressed conditions, distinct functional genes pinpoint to a distinct stress sensing, water availability and compounds in the rhizospheres.

Immunology

General Microbiology

Shotgun metagenomics

Solanum tuberosum

Solanum phureja

bacteriophage

plant-microbe interaction

plasmid

rhizosphere

endophytes

rhizobacteria

drought

Climate change and reduction in the use of mineral fertilisers can lead to the exposure of crops to combined stresses such as irregular rainfall and limited availability of mineral nutrients, thereby affecting global food production Particularly water limitation represents a major constraint, and plants must employ diverse mechanisms to mitigate drought stress and undergo numerous morphological, physiological and molecular adaptations [1]. Plant breeding or innovative agricultural management can provide solutions to cope with these challenges. In addition, the plant microbiome has the potential to alleviate plant stress.

Plant microbiomes, i.e. plant-associated microbial communities and their “theatre of activity” [2, 3] are highly complex consisting of bacteria, archaea, fungi, oomycetes, protists and viruses and can be essentially found in all plant tissues and compartments. Plant microbiota are important for plant growth and health and are involved in key functions such as nutrient mobilization, protection against pathogens or improving plant resilience to abiotic stress [4–6]. Well studied plant compartments for microbial life include the rhizosphere [7] the endosphere [8] and the phyllosphere [9]. The rhizosphere is a hotspot of microbial diversity and activity utilizing root exudates and sloughed off plant cells as nutrient sources [7]. After colonizing the rhizoplane microorganisms may also enter roots, either passively e.g. via wounds or actively using cellulolytic enzymes and may thrive in the root endosphere and/or translocate to above-ground plant tissues [8]. Microbial communities in different plant compartments are clearly different from each other, indicating strong influence of host-specific factors [8, 10]. Whereas microbiota components of the rhizosphere at a coarse level are similar for various plant species, there is a greater influence of the host plant on endophytes [10]. In addition, stress conditions, plant development and to a certain extent the plant genotype are major drivers of plant microbiota [11–13].

Plant-associated microorganisms have been shown to protect plants from abiotic stress such as drought [14]. For example, the bacterial endophyte Paraburkholderia phytofirmans strain PsJN increases drought tolerance in maize [15] and wheat [16]. Particularly bacteria producing the enzyme ACC deaminase, which is also highly active in strain PsJN, are prominent candidates for the improvement of drought stress resilience [17]. This enzyme is responsible for lowering the levels of ethylene in the plant by cleaving the plant-produced ethylene precursor 1-aminocyclopropane-1-carboxylate (ACC) to ammonia and 2-oxobutanoate, modulating ethylene signalling [18]. Other known mechanisms include the detoxification of reactive oxygen species (ROS) [19, 20] or modulating abscisic acid metabolism [14].

Recent work has also indicated that soil microbiomes alter plant fitness and competition under drought [21] and that root microbiome compositional changes correlate with drought stress tolerance across plant species [22]. Several studies have shown that drought affects the composition of root microbiota, particularly favouring monoderm bacteria such as Actinobacteria, which are known to be more resistant to desiccation than diderms [21–24]. The effect of plant nutrient starvation on plant microbiomes is less well investigated. Finkel et al. [21] recently showed that the phosphate starvation response of Arabidopsis has a large effect on the plant-associated bacterial and fungal communities, whereas different types of P-fertilizers did not show a major influence on below-ground microbial communities [25].

Potato (Solanum tuberosum) is the world´s 4th most produced staple crop, after maize, wheat and rice. Potato production has been considered to be severely impacted due to global warming and drought and a yield decline of 18–32% in the period between 2040 and 2069 has been predicted [26]. Efforts are ongoing to improve tolerance of potato to various abiotic stresses, primarily to drought, by intensive crop breeding, e.g. by enhancing photosynthetic performance [27]. Potato plants are associated with complex microbiomes, particularly in below-ground compartments [28, 29], however, the response of the potato microbiome to abiotic stress, particularly drought, and its role for stress mitigation is poorly investigated. Here, we compared ten different potato cultivars grown in the field under combined drought and P limited conditions, and particularly assessed the link of plant growth, yield traits and belowground plant microbiome composition. Furthermore, for a poorly and well performing potato genotype under combined stress conditions rhizobacterial metagenome functions were identified.

Ten potato genotypes (Supplemental Table 1) were grown in the field with a combined stress of reduced irrigation and no phosphate fertilizer (Supplemental Fig. 1A, B). For all measured time points, soil moisture was highest in the deep soil layers and reduced in the top-soil and was different between stress conditions (Supplemental Fig. 1C, D). Comparing the number of young tubers eight weeks after planting, half time to canopy closure, final yield and above ground biomass (foliage) between stressed and non-stressed potato plants revealed that different genotypes had different stress responses. The stress effect on tuber yield correlated significantly with the effect on above-ground biomass (Supplemental Fig. 1E) but a reduced number of young tubers during tuber filling correlated with stress resilience in tuber yield (Supplemental Fig. 1F). This indicates that a delay in growth under continuous but reduced water supply is beneficial for stress resilience of the potato plants.

High throughput amplicon sequencing of 166 samples revealed a total of 5.8 M 16S rRNA gene sequences and 3.6 M ITS sequences after removing plant-derived sequences. Those were grouped into 20,114 (1,302 occurring in at least three samples) different bacterial and 941 (76) fungal amplicon sequence variants (ASVs). On average the samples contained 34 k ± 19 k bacterial and 21 k ± 18 k fungal ASVs. From two contrasting genotypes (Desirée = high yield loss under stress, Stirling = low yield loss under stress), we analysed the functional potential of the microbial community in the rhizosphere under stress and no stress by shotgun metagenomics. We obtained 136 M ± 17 M per sample and in total 1,633 M paired sequences. A 98% subset of all classified reads belonged to bacteria, 1.2% to phages, 0.7% to archaea and 0.1% to fungi. Reads of plasmids summed up to 6.2% of all reads classified with the Kraken-Braken method.

Sample type and stress shape the microbial diversity and structure of potatoes.

The nonmetric multidimensional scaling (NMDS) ordination of the amplicon dataset showed that the structural differences in the microbial community composition were mostly influenced by the sample type (Fig. 1A, 1D environmental fit, bacteria: R2 = 0.79, p-value ≤ 0.001, fungi: R2 = 0.45, p-value ≤ 0.001) followed by stress (bacteria: R2 = 0.09, p-value ≤ 0.001, fungi: R2 = 0.06, p-value ≤ 0.001). The calculation of a general linear model of the values of NMDS1 resulted in a significant influence of each sample type on the bacterial composition (soil, p-value < 0.001; rhizosphere, p-value < 0.001; root, p-value < 0.001) but the fungal rhizosphere composition did not differ significantly from the other sample types (soil, p-value < 0.001; rhizosphere, p-value = 0.94 ; root, p-value < 0.001). A general linear model of the scores of NMDS2 suggested a significant influence of stress on the microbiota (bacteria: stress, p-value < 0.001; no stress, p-value < 0.001; fungi: stress, p-value < 0.001; no stress, p-value < 0.001). The highest richness (= number of different amplicon sequences) was found in the rhizosphere followed by soil and root samples (Fig. 1C, F). Both, richness and Shannon Index, revealed a reduction in microbial diversity in the rhizosphere under stress conditions. In roots a significant reduction was only observed for bacteria (Fig. 1B, 1E). Additionally, in the rhizosphere metagenomes, diversity of archaea was increased under stress (Shannon Index: stress = 3.47, no stress = 3.27, p-value = 0.015).

Common stress reactions of the microbial composition in various potato genotypes.

The most abundant phyla in the rhizosphere included Proteobacteria, Actinobacteria, Bacteroidetes, Ascomycota, Mortierellomycota and Basidiomycota (Supplemental Fig. 2). At the genus level the rhizosphere contained mostly Sphingomonas, Flavobacterium, Streptomyces, Mortiella, Solicoccozyma and Pseudeurotium. Roots were additionally dominated by the phyla Firmicutes as well as Olpidiomycota and by the genera Bacillus, Paenibacillus and Microdochium (Supplemental Fig. 3). Under stress conditions Actinobacteria, Sphingobacteriales and Variovorax were enriched, while Proteobacteria, Flavobacteriales and Olpidiomycota were reduced in roots and the rhizosphere. We observed sample type-specific stress reactions like the enrichment of Xanthomonadales in rhizosphere samples and Clostridia in roots under stressed conditions. In contrast to Xanthomonadales and Clostridia, the abundance of other Gammaproteobacteria and Firmicutes were reduced under stress conditions (Fig. 2). Similarly, different members of the Leotiomycetes showed different responses, Theloboales were enriched and Helotiales were reduced in the rhizosphere under stress conditions.

At the highest taxonomic resolution of the amplicon-dataset, we identified 174 ASVs showing significantly different abundance in one of the two stress treatments (Supplemental Table 2): i) root endosphere: 6 fungal and 46 bacterial ASVs; ii) rhizosphere: 4 fungal and 118 bacterial ASVs. Interestingly only three ASVs, all belonging to Actinobacteria, were significantly enriched in both root and rhizosphere samples under stressed conditions: Nonomuraea sp. ASV_90, Streptomyces sp. ASV_268 and Streptomyces sp. ASV_9.

The reduced shotgun-dataset confirmed the enrichment of Actinobacteria and other stress-specific bacterial taxa but detected more significant differences in Alphaproteobacteria compared to the amplicon dataset (Supplemental Fig. 4A). Regarding archaea, Methanococci were more abundant in non-stressed samples while Halobacteriales, Haloferacales and Methanomicrobia were enriched under stress (Supplemental Fig. 4B). In total 17 good quality metagenome assembled genomes (MAGs) were identified (Supplemental Table 3). Five MAGs (3 Actinobacteria, 2 Proteobacteria) were more abundant under stress and four (all Proteobacteria, genus Sphingobium) were more abundant under no stress (Supplemental Table 4). In general, the fold changes of MAGs were smaller (max 1.2 fold) than the fold changes of the ASVs (max. 27 fold). Consistently, a MAG and an ASV belonging to the taxon Nonomuraea were enriched in stressed samples and MAGs as well as ASVs belonging to Sphingobium were enriched in no stress samples.

Genotype-specific differences in stressed potato plants.

The Bray-Curtis distance showed that the microbial communities were more similar for samples belonging to the same genotype than to different genotypes in both stress conditions (Supplemental Fig. 5A). Concordantly, general linear models revealed significant effects for stress and genotype in root and rhizosphere samples (Supplemental Table 5). Community structures according to stress and genotype were most prominent in the subset of rhizobacteria leading to distinct clusters in the PCoA (Supplemental Fig. 5B). Noticeable is the separation of the diploid (Supplemental Fig. 5B, dark red and light red) vs. tetraploid (other colours) potatoes under stress. The fungal community in roots and rhizosphere as well as the bacterial community in roots were also significantly affected by stress and genotype but less profoundly (Supplemental Fig. 5, Supplemental Table 5). In general, the F-value of the factor stress reduced from the rhizosphere to the root microbiota (Supplemental Table 5A) indicating a lower stress effect on root microbiota as compared to the rhizosphere. Also, fungi were less affected than bacteria. In contrast, the F-value of the factor genotype was similar between subsets, indicating a constant effect of the genotype on the microbiota (Supplemental Table 5).

Correlating diversity and microbial abundance with stress responses of tetraploid potato plant growth.

Different potato genotypes showed different phenotypic stress response patterns including effects on final yield, foliage, half-time to canopy closure, number of young tubers and diameter of the largest young tuber (see also Supplemental Fig. 1E, F). The abundance of some ASVs correlated with the phenotypic stress responses of potato plants, which could be grouped in clusters (Fig. 3, Supplemental Table 6). For instance, Xanthomonadales sp. ASV_465, Chitinophaga arvensicola ASV_499 and Occallatibacter sp. ASV_869 were more abundant in the rhizosphere of potato genotypes with a stable yield (Fig. 3A, ***, cluster III). In contrast, Flavobacterium sp. ASV_30 in the rhizosphere and Streptomyces sp. ASV_168 in roots were most abundant in genotypes that suffered from high yield loss under stress. The abundance of most microbes correlated highly significantly with a faster half-time canopy closure (Fig. 3A VI, 5C IV, ***). One example is Dyadobacter sp. ASV_47 that correlated to half-time canopy closure in root and rhizosphere samples. The fungi Trichocladium opacum ASV_44 and Mortierella hyalina ASV_24 occurred in the rhizosphere of stress-resilient genotypes and correlated with tuber yield (Fig. 3B, I).

The rhizosphere metagenomes of potato plants exposed to combined stresses and no-stress have distinct functional potentials.

Shotgun metagenomic sequencing revealed a huge impact of stress on gene abundance (Supplemental Table 7). Of 17.548 genes and gene fragments in the bacterial dataset, 31% were more abundant under stress and 27% were more abundant in no-stress samples. More than 2000 stress indicator genes belonged to Actinobacteria, represented by Pseudonocardiales and Propionibacteriales, whereas 800 genes belonged to Beta- and Gammaproteobacteria represented by Xanthomonadales and Comamonadales (Fig. 4A). Merging all taxa, we identified 14 functional groups (KEGG C-level, Fig. 4B, details in Supplemental Table 8) that were more abundant in rhizosphere metagenomes under stressed conditions, including i) sugar-, ii) amino acid- and iii) vitamin/cofactor metabolism as well as iv) base excision repair. Most functions were mainly represented by Beta- and Gammaproteobacteria under no stress, whereas under stress Actinobacteria increased in proportion matching the increase of Actinobacteria under stress in the amplicon dataset. Within Actinobacteria, taurine and hypotaurine metabolism and terpenoid backbone biosynthesis were over-represented in rhizosphere metagenomes under stress conditions while within Beta- and Gammaproteobacteria biofilm formation, fatty acid biosynthesis, biotin metabolism and mismatch repair were over-represented functions. Beyond KEGG C-level we identified KEGG modules composed of stress indicator genes (Supplemental Table 9) and presented a selection in Fig. 4C. Represented by Actinobacteria, a glycine betaine/proline sugar-ABC-transporter was more abundant under stress. Furthermore, underlying genes for trehalose biosynthesis were more abundant in rhizosphere metagenomes under stress conditions.

In addition to drought, stressed plants were exposed to phosphorus limitation. Concordantly, the OmpR two-component system involved in phosphate assimilation and a phosphate ABC-transporter were more abundant under stress but only in Actinobacteria (Supplemental Table 9). The heme biosynthesis, pentose phosphate and leucine degradation pathways were more abundant under stress (Fig. 4C). Rhizosphere microbiota of stressed plants showed a higher genomic potential to produce i) isoprenoids (C5 non-mevalonate pathway, C10-C20) and ii) precursors of aromatic acids and secondary metabolites via the shikimate pathway (Fig. 4C). Summarizing the reads at the higher functional level, KEGG B, revealed an increased abundance of reads assigned to biosynthesis of secondary metabolites (Supplemental Table 8). One identified secondary metabolite with a higher abundance in metagenomes under drought was the sesquiterpenoid geosmin (Fig. 4C). Among the 12 functional groups more abundant in samples from non-stressed conditions (Fig. 4B, and more detailed in Supplemental Table 8, FDR < 0.01) was cell motility and protein export. Surprisingly, the function glutathione metabolism and four genes similar to the glutathione-S-transferase being involved in detoxification, were more abundant in metagenomes under non-stressed conditions (Supplemental Fig. 6B). Within Beta- and Gammaproteobacteria glutathione metabolism was over-represented in samples from non-stressed conditions along with carbohydrate metabolism and carbon fixation in prokaryotes (Fig. 4B). Moreover, all five orthologous gene families of urea ABC-transporters were enriched in rhizosphere metagenomes under non-stresses conditions (Supplemental Table 9A).

Functional potential in rhizosphere metagenomes differ between a good and a poor performing potato genotype.

Desirée produced 7.7 kg tubers under stress, a loss by 55% compared to non-stress conditions, while Stirling performed better with a yield of 9.3 kg, a loss by 38%. Both genotypes grew next to each other, ensuring the same pool of soil bacteria for rhizosphere enrichment. Under stress 1562 genes from Xanthomonadales (Fig. 5A, green) were more abundant in the Stirling rhizosphere metagenomes while 1703 genes from Pseudonocardiales (dark red) were more abundant in the Desirée metagenome (Supplemental Table 10). In samples from no-stress conditions Xanthomonadales were more abundant in Desirée and Pseudonocardiales more abundant in Stirling (Supplemental Fig. 7). Propinobacteriales (orange, more abundant in Stirling) and Flavobacteria (pink, more abundant in Desirée) preferred one genotype regardless of the stress treatment.

Distinct functional groups were dominated i.e. were most abundant in sequence numbers by distinct taxonomic groups (Fig. 5B). In Desirée, Actinobacteria together with Beta- and Gammaproteobacteria dominated most functional groups while in Stirling mainly Beta- and Gammaproteobacteria dominated most functions More sequences were assigned to lipopolysaccharide biosynthesis in Stirling and the function was over-represented in Beta- and Gammaproteobacteria in the rhizosphere of Stirling compared to Desirée (Fig. 5B). Similarly, biofilm formation and fatty acid biosynthesis via the Raetz pathway, together with genes for ABC-transporters of lipoproteins and lipophospholipids belonged mainly to Xanthomonadales and were concordantly more abundant in Stirling (Fig. 5C, details: Supplemental Tables 11 & 12). Other Stirling indicator genes from Xanthomonadales are involved in i) the conversion of L-cysteine via glutathione to L-glutamate (glutathione metabolism), ii) the conversion of taurine to 5-glutamyltaurine (Supplemental Table 12E) iii) fructose uptake (phosphotransferase system, PTS, Fig. 5C) iv) the production of auxin by the tryptophan 2-monooxygenase (iaaM) (Supplemental Fig. 8A) and v) in the type II secretion system (Supplemental Table 12E). One rhizosphere metagenome assembled genome MeBa083 was classified as Lysobacter (order Xanthomonadales, Supplemental Table 4) and contained two bacteriocin-, two lanthipeptide-, one arylpolyene and one polyketide synthase-like region. Carbohydrate metabolism and carbon fixation pathways in prokaryotes were over-represented in Actinobacteria from Stirling compared to Actinobacteria from Desirée rhizosphere metagenomes (Fig. 5B). In total five glutathione-S-transferases from three different taxa were more abundant in Stirling metagenomes (Supplemental Fig. 8A).

In rhizosphere metagenomes genes for steroid degradation were more abundant and over-represented within Actinobacteria from Desirée compared to Actinobacteria from Stirling (Fig. 5B). Additionally, folate biosynthesis, biotin and beta-alanine metabolism were over-represented in Actinobacteria from Desirée rhizosphere metagenomes. In general, genes with assigned function that were more abundant in Desirée belonged mainly to Pseudonocardiales (Fig. 5A). This included purine degradation to urea and diverse transporter genes for i) sugars (raffinose, chitobiose, sorbitol, ribose, D-xylose) ii) oligopeptide and iii) tetrathionate (Supplemental Table 12J). Branched amino acid and C4-dicarboxylate transport genes were from Comamonadaceae and Pseudonocardiales, while more abundant amino acid urea transporter genes were only detected in Comamonadaceae (Fig. 5C). The plant growth-promoting functions, ACC-deaminases and a pyrroloquinoline quinone biosynthesis gene, were found in Pseudonocardiales (Supplemental Fig. 7B). Phosphate ABC-transporter were more abundant in the stressed rhizosphere metagenomes of both genotypes: from Pseudonocardiales in Desirée and from Xanthomonadales in Stirling. Biotin metabolism was over-represented in Actinobacteria from Desireé and in Beta- and Gammaproteobacteria from Stirling indicating that different genotype-indicator taxa can have the same function under stress.

Plasmids and phages - mobile elements in potato rhizosphere metagenomes

Besides functional genes, mobile elements varied between stress treatments: i) Shannon diversity of phages increased under stress (Fig. 6A) and ii) the relative amount of plasmid sequences was less in samples under stressed conditions (Fig. 6B). Interestingly, in rhizosphere metagenomes the Shannon diversity of antibiotic resistance genes on plasmids was higher in stressed compared to non-stressed conditions (Fig. 6C), indicating a selective advantage of bacterial plasmids harbouring antibiotic resistance genes. In total of 7010 phages detected in rhizosphere metagenomes, three were more abundant in non-stressed and 49 in stressed conditions (Supplemental Table 13A). Similarly, of 1535 plasmids detected in rhizosphere metagenomes, 49 were more abundant in non-stressed and 104 more abundant in stressed conditions (Supplemental Table 13B). Noticeably, 68 of the 104 taxa in which plasmids were more abundant under stress conditions belonged to Streptomyces. Most plasmids changed in the same ratio as bacteria (Figure D, diagonal line) but some plasmids were more abundant in rhizospheres of one stress condition although the bacterial abundance did not change (Fig. 6D, vertical line).

In addition, we observed differences in mobile elements between the genotypes under stress: i) the relative number of phages (Fig. 6B) and ii) the diversity of plasmids (Fig. 6A) was higher in metagenomes of the poorly performing genotype Desirée compared to Stirling. One hundred and six phages and 10 plasmids were more abundant in Desirée rhizosphere metagenomes (Supplemental Table 14). Three bacterial taxa, Cupriavidus nantongensis, Enterobacter asburiae and Lactobacillus plantarum, had a higher portion of plasmids despite minor changes in their whole genomes (Fig. 6E, vertical line). In rhizosphere metagenomes of Stirling compared to Desirée, 37 phages and 67 plasmids were more abundant. This included the plasmids of four Xanthomonas species (Supplemental Table 14). Two bacterial taxa, Labrenzia sp. THAF35 and Trichormus variabilis had a higher portion of plasmids despite minor changes in their whole genomes (Fig. 6E, vertical line). But in general, most plasmids were co-enriched with their bacterial hosts (Fig. 6DE, diagonal line). For phages no exact host ID is available.

Due to climate change, reduced use of fertilizers or availability of nutrients combined stress scenarios become increasingly relevant in crop production. Here, we show that for potato combined stress, i.e. water and phosphorus limitation, has a tremendous impact on the potato holobiont at all levels, from microorganisms to phages and from microbial composition to functions, all associated with severe impacts on plant traits. Overall, the responses resemble those typically reported for drought stress suggesting that water limitation had a more severe impact, however, less information is available on the impact of phosphorus limitation (e.g. on microbiomes).

Microbial composition and functions in the root environment indicate stress-adapted plant-microbe interactions

Belowground microbes suffer directly from reduced soil moisture as water lacks as a transport medium, solvent and resource. However, we and Gao et al. [30] found diversity changes due to water limitation mainly in the rhizosphere but not in bulk soil samples, indicating an impact of distinct root deposits and exudates on microbial communities under stress. In general, under drought stress monoderm bacteria like Actinobacteria were found to be enriched in most field samples while chaotropic, mobile, diderm bacteria, like Proteobacteria, were found to be reduced [11, 22]. These findings are well in line with those of our study. However, we found that under water and nutrient limitation Xanthomonadales were more abundant in the rhizosphere but not in roots suggesting a compartment-specific stress adaption. Furthermore, Micrococcales (Actinobacteria) were depleted under stress while still most Actinobacteria like Streptomyces and Pseudonocardiales were enriched in the rhizosphere. In roots, additionally the Actinobacteria Propionibacteriales and Streptosporangales were enriched, indicating an adaption to plant metabolism besides drought resistance of the bacteria by sporulation or thick cell membranes. In contrast to the bacterial root community, stress only marginally influenced the fungal root community. One exception was the phylum of the plant pathogens Olpidium (depleted in rhizosphere and roots under stress) that propagate as motile spores [31]. Also, motile bacteria and their genetic potential for flagella assembly were depleted under drought. In the rhizosphere, the fungal taxa Sordariales were favoured while Pleosporales were depleted by drought like in grassland soils [32].

Metagenome analysis of the rhizosphere revealed that half of all detected genes was differentially abundant in rhizosphere metagenomes under stressed and non-stressed conditions indicating strong stress effects on microbial functions. Various sugars are osmoprotectants, and pathways of ‘fructose-/ mannose metabolism’ and ‘ascorbate- / aldarate metabolism’ were more abundant in rhizosphere metagenomes under stress. The higher abundance of the pathways ‘base extinction repair’ and ‘fatty acid biosynthesis’ under stress suggests an adaptation of the microbial community towards a higher tolerance of DNA and membrane damaging agents such as ROS. ROS are detoxicated by antioxidant systems that utilize cofactors, like i) the pentose phosphate pathway with the potential to maintain NADPH availability ii) vitamin B6 (pyridoxine) metabolism, iii) pantothenate and coenzyme A biosynthesis and iv) heme biosynthesis, which were all enriched functions under stress conditions. Furthermore, genes in Beta- and/or Gammaproteobacteria involved in biotin metabolism were enriched. Cofactors are involved in root growth [33] and root colonization [34], defence [35] and may directly promote plant growth by alleviating osmotic and oxidative stresses [36, 37]. The enrichment of diverse cofactors in this study points to their importance for diverse plant-microbe interactions under stress.

The disaccharide trehalose is a messenger in plants that regulates carbon assimilation and sugar status at low concentrations [38]. Manipulating the potato genome in order to produce more trehalose lead to drought-resistant potato genotypes [39]. Also, the enrichment of trehalose-producing bacteria in the rhizosphere might be a promising alternative [40]. Here we identified more genes involved in the biosynthesis of trehalose under drought, indicating an interesting mechanism of rhizobacteria to modulate plant physiology under drought.

Terpenoids, also called isoprenoids, are a large class of secondary metabolites with more than 80.000 compounds descripted in the Dictionary of Natural Compounds [41]. Especially Actinobacteria are known to produce terpenoids [42], and within Actinobacteria those with a higher genomic potential in terpenoid backbone biosynthesis were over-represented in rhizosphere metagenomes under stressed conditions, suggesting a role of terpenoids in stress mitigation. We also found a higher abundance of genes responsible for geosmin production under stress. The triterpene geosmin is the typical odour of soil after a drought period followed by rain and is thought to be involved in attracting springtails for bacterial spore distribution indicating multi-kingdom interactions in the rhizosphere [43]. Besides terpenoid biosynthesis, the potential of terpenoid degradation of steroids was higher in microbial metagenomes under stress only in the poorly performing potato cultivar Desirée. This indicates a higher availability of steroids in Desirée rhizospheres produced by bacteria, fungi and/or plants. Plant-derived brassinosteroids play a role in drought tolerance but also exogenous steroid applications can boost crop yield [44]. Therefore, rhizobacteria could interfere with steroid-driven stress responses of plants.

The pathways of sulfonic amino acid taurine metabolism and the sulfur relay system were mainly represented by Actinobacteria and more abundant under stress in the rhizosphere. Even within Actinobacteria the taurine metabolism was over-represented in the rhizosphere under stressed conditions, indicating the recruitment of Actinobacteria active in sulfur metabolism. This links water or phosphor limitations to sulfur metabolism, as suggested by Kaya et al. [45]. Phosphor-limitations decrease selenate and selenite adsorption of soils [46] and promote selene uptake in wheat [47]. Interestingly, within Beta-/and Gammaproteobacteria the selenocompound metabolism was over-represented in the rhizosphere of Stirling under stressed conditions, indicating that recruitment of Beta- and/or Gammaproteobacteria active in selenocompound metabolism in Stirling rhizospheres might be a reaction to reduced phosphor availability.

Functions in rhizobacteria match distinct growth strategies of potato genotypes under stress

Various bacterial phyla and bacterial or fungal ASVs correlated with high and small yield losses as well as other plant traits. Bacteria may influence root growth [48] and shift maturity [49]. This may affect the balance of energy investments in canopy cover and tuber bulking, which differs between potato genotypes and influences potato performance under drought [50]. In our study microorganisms might have contributed to the minimal delay of the canopy closure of the variety Desirée under stress. Potential functions involved in this effect include i) ACC deaminase enabling the bacteria to degrade the plant stress hormone ethylene and ii) a gene known as plant growth promotion factor involved in pyrroloquinoline quinone biosynthesis [51] from Pseudonocardiales. Microbes profit from a fast canopy closure as they feed from plant assimilates produced in leaves but do not profit from other traits like higher tuber yield in the short run. Furthermore, the Pseudonocardiales potentially associated with fast canopy closure in Desirée had genes encoding sugar and branched amino acid ABC transporters as well as C4-dicarboxylate two component systems, matching photosynthesis-derived plant exudates [52]. Although the mechanism must be elucidated in further experiments, the association of Pseudonocardiales and canopy closure under stress is reinforced by five Pseudonocardiales ASVs that were more abundant in genotypes with a minimal halftime canopy delay.

Most microbes in the roots and rhizosphere correlated with a minimal canopy closure delay but none of those indicated also genotypes with low yield losses. However, under drought a Streptomyces strain ASV_62 was abundant in the rhizospheres and roots of potato plants with delayed young tuber development, a trait which correlated with stable yields under drought. Concordantly in Arabidopsis thaliana bacterial inoculation delayed plant development to overcome long-term water deficits [49], indicating that selecting beneficial microbes delaying plant development at early stages under stress might be interesting for agricultural applications.

In the rhizosphere of the good performer Stirling compared to Desirée we detected more abundant biofilm pathways and modules like lipopolysaccharide production including the Raetz pathway, the transport of lipoproteins and phospholipids [53, 54]. Xanthomonadales are known to produce biofilms in soils and rhizosphere, alone but also in consortia [55, 56]. In this study candidates for multispecies biofilms are the genera Granulicella, Streptomyces and Leifsonia that co-occurred with Xanthomonadales (Fig. 3, cluster III) and were described to produce exopolysaccharides [57] or to be present in biofilms [58, 59]. Sponge-like biofilms can maintain moisture in the rhizosphere under drought [60, 61], may contain antimicrobial substances [55] and facilitate transport of some minerals and nutrients, which could explain the high abundance of Xanthomonadales in the good performing genotypes in this study. Moreover, Stirling rhizosphere metagenomes hosted genes encoding tryptophan 2-monooxygenase (iaaM) from Xanthomonadales potentially involved in plant-microbe interactions [62]. Interestingly, Na et al. [63] found that under drought the slower growing genotype of Panicum milianceaum L., just like the slower growing Stirling in this study, harboured more Lysobacter (Xanthomonadales) at flowering, indicating a potential role of some Xanthomonadales in stress-induced growth delay. Under no-stress, Xanthomonadales were more abundant in the faster growing genotype Desirée, excluding only vegetative growth stage- or genotype-specific colonization preferences but showing a stress-dependent enrichment of Xanthomonadales in the rhizosphere of Stirling.

The role of mobile genetic elements in rhizobacterial adaptions to stress.

Plasmids and bacteriophages (phages) in bacteria often harbour operons or genes for virulence, quorum sensing, antibiotic resistance and secondary metabolism. Mobile elements may be exchanged between bacteria to accumulate in populations if needed, thereby enabling a fast adaption to environmental conditions. However, there are high energy costs to maintain plasmids and the lysis of infected bacteria for phage propagation [64]. While plasmids and phages within plant microbiomes have been rarely described, particularly those involved in drought or nutrient stress tolerance, we found significant differences in plasmids and phages occurring under stress and unstressed conditions. Overall, the relative abundance of plasmids was reduced under stress, probably because of the high maintenance costs of plasmids. However, various bacterial plasmids were more abundant under drought. These plasmids either multiplied in their parent strain or spread to other bacteria. This includes plasmids from Variovorax sp. and Cupriavidus sp., both known for comprising strains which improve tolerance of plants to drought [65, 66]. Strikingly, the plasmids of Cupriavidus sp. were over-represented in Desirée, indicating a genotype-specific advantage under stress encoded on those plasmids. Despite the lower relative abundance, we found a higher diversity of plasmid-encoded antibiotic resistance genes, indicating the presence of antimicrobial substances, potentially produced by other microorganisms, in the rhizosphere.

Interestingly, phages, in contrast to plasmids, increased in relative abundance under stress in a genotype-specific manner. Phages occur in higher relative amounts in populations if they are more abundant within one bacterial cell or persist in more cells, which is likely if phage-infected bacteria have a fitness advantage over uninfected ones. One fitness advantage of phage-infected bacterial cells is the resistance to other phage infections, a phenomenon called “superinfection immunity” [64] which results in reduced phage-induced bacterial lysis despite an increased relative abundance of temperate phages and might be the case in the rhizosphere of Desirée. In the rhizosphere of the good performer Stirling, the suggested biofilm could limit the relative abundance of phages [67]. So far, very little information exists on the role of phages in the plant environment. However, we know from the human gut that phages play a major role in microbiome development and adaptation [68] and think that plant-associated phages merit further investigation to understand their role in microbiome modulation and adaptation.

Distinct functions and taxa in rhizobacteria match the distinct phenotypic potato stress responses but to which extent they manipulate plant growth or whether they react to the plant-chosen growth strategy remains to be elucidated. Besides intertaxon stress adaptions, we identified changes in plasmid and phage diversity and relative abundance showing intrataxon genome adaptions to combined stresses in a genotype-specific manner. Mobile elements act faster on genomes than genome adaption through propagation and might play an important role in bacterial stress adaption over the time period of a growing season. In this study potatoes were cultivated under constant but reduced water supply, therefore, a slower potato growth increased the total amount of water available over the complete life cycle. If rhizobacteria and phages shape the phenotypic stress pattern of potato plants, engineering root-associated microorganisms and phages could be used to ensure a plant stress response matching the needs and watering regime of the farmer.

Experimental site, set-up and sampling.

On the 16th of May 2018, at the James Hutton Institute in Dundee, Scotland, ten Solanum tuberosum genotypes (eight tetraploids belonging to group Tuberosum; two belonging to the diploid group Phureja) were planted in clay soil (edaphic soil factors, April 2018, pH: 6; Lime req, Arable: 2.5t/ha; Lime req, Grass: 0.0 t/ha; Extractable Phosphorus: 12.2 mg/l; Extractable Potassium: 242.0 mg/l; Extractable Magnesium: 174.0 mg/l; Extractable Calcium: 1700 mg/l; Extractable Sodium: 12.80 mg/l; Extractable Sulphur: 1.8 mg/l; Extractable Copper: 20.9 mg/l; Extractable Manganese: 3.9 mg/l; Extractable Boron: 0.83 mg/l; Extractable Zinc: 11 mg/l; Organic Matter, LOI: 7.04 %). The genotypes were grown in polytunnels (9.2 m width x 100 m length; c.f. [69] and ploidy as well as maturity classes are listed in the Supplemental Table 1. Two treatments were established: i) conventional fertilizer application (Defra RB209, ORIGIN 14-14-21, 1050 kg ha^− 1, N = ammonium nitrate) at planting with supplemental irrigation (two to three 30 min. applications per week as required) and ii) conventional fertilizer application but without phosphorus fertilizer and with reduced irrigation. Soil humidity was measured during the experiment. For microbial analysis soil and plant samples were taken from the 7th to 10th July at constant warm weather. Each plot consisted of two genotype and two border rows (Supplemental Fig. 1A,B). Bulk soil was sampled at random sites between the plots. Out of eight plants per row in a plot, four with a similar growth but not the border plants, were selected for analysis. Using a potato fork, whole plants with root systems were removed and a representative collection of 6–10 dirty but shaken root branches (no stolons) were collected into 50 ml tubes, stored in a cooling box and processed at the same day. At destructive microbiome sampling, the largest diameter of each tuber (all called young tubers) was recorded. In an adjacent experiment, the half-time to canopy development as well as shoot biomass (Supplemental Fig. 1C,D; indicated by haulm) and yield (indicated by tuber) were determined by the James Hutton Institute.

Sample preparation and DNA extraction.

Soil samples were homogenized and particles larger than 0.5 cm were removed. Tubes containing the roots and 25 ml of sterile water were shaken for 3 min. Centrifugation of the suspension for 10 min at 4000 x g sedimented the rhizosphere soil. Root samples were i) washed under running tap water, ii) surface sterilized by submerging them for 5 min in 2.5% NaOCl enriched with one drop of Tween 20, iii) washed three times in sterile water and iv) dried in the oven (85°C) overnight. Cut roots (length 0.5 cm) were frozen at -80°C and homogenized twice for 1.5 min in a TissueLyser at 30 Hz in two different orientations. Root powder (40 ± 5 mg), rhizosphere soil (200 ± 50 mg) and bulk soil (250 ± 10 mg) were stored at -20°C in aliquots till DNA extraction, which was performed according to the Qiagen DNeasy Power Soil Kit. The 2 ml reaction tubes were shaken for 10 min twice in a TissueLyser at 20 Hz.

Amplicon and shotgun metagenomic sequencing and data processing.

Amplicon library was prepared in a two-step PCR approach according to Samad et al [70]. The following primers bind to targeted DNA in the first PCR: i) for bacteria 799f-illumina 5’-TCG TCG GCA GCG TCA GAT GTG TAT AAG AGA CAG AAC MGG ATT AGA TAC CCK G-3’; 1175r-illumina 5’-GTC TCG TGG GCT CGG AGA TGT GTA TAA GAG ACA GAC GTC RTC CCC DCC TTC CTC − 3’ and ii) for fungi ITS1f-illumina 5’-TCG TCG GCA GCG TCA GAT GTG TAT AAG AGA CAG CTT GGT CAT TTA GAG GAA GTA A -3’ and ITS2-illumina 5’-GTC TCG TGG GCT CGG AGA TGT GTA TAA GAG ACA G GCT GCG TTC TTC ATC GAT GC -3’. Each sample was processed in three technical replicates to reduce random PCR effects. Bacterial amplicons of 480 bp from root-DNA were extracted from 2% agarose gels. enriched via gel extraction. In the second PCR we indexed each sample with primers of the Nextera XT Index Kit (Illumina, Inc, USA). A blank DNA-isolation without material and a control library (D6305, ZymoBIOMICS, USA) were included. The samples were sequenced on an Illumina MiSeq at the Competence Unit Bioresources of the AIT Austrian Institute of Technology in Tulln. Illumina MiSeq reads were filtered with Bowtie2 v2.3.4.3 [71] to avoid the presence of Illumina PhiX contamination and quality was preliminarily checked with FastQC v0.11.9 [72]. Primers were stripped using Cutadapt v1.18 [73]. Sequences were quality filtered, trimmed, denoised and amplicon sequence variants (ASVs) were generated with DADA2 v1.14 [74]. Denoised forward and reverse ASV sequences were merged and chimeras were removed. Filtered ASVs were checked using Metaxa2 v2.2.1 [75] and ITSx v1.1.2 [76], respectively, for targeting the presence of V3-V4 16S rRNA and ITS1 region, in bacterial sequences and fungal sequences. Taxonomic assignment of 16S-rRNA-gene ASVs and ITS-based ASVs was performed using the RDP classifier [77] of DADA2 against the SILVA v138 [78] database and UNITE 8.2 [79] database, respectively. BIOM objects (i.e. count matrices equipped with taxonomic information) with bacterial and fungal counts were built and imported into the R statistical environment.

For shotgun metagenomics the DNA was cleaned following the Microcon Ultracel YM-30 (Merck, Germany) protocol. At the Vienna Bio Core Facility, the NGS DNA Library Prep Kit (Westburg, Netherlands) was used and 2x150 bp sequencing was conducted on a NovaSeq (Illumina, USA). Illumina’s PhiX reads were filtered out of the sequencing data with Bowtie2 v2.3.4.3 [71]. Filtered reads were processed using fastp v0.20.1 [80] with a cutting-by-quality sliding-window approach from 5´ to tail and 3´ to the front of each read. The selected window size was 4 bp with a minimum quality of Q20. Adapters were auto-detected and removed. A quality check was carried out with FastQC v0.11.9 [72]. Fastp and FastQC output summaries, respectively, were inspected using MultiQC v1.9 [81]. The metagenomic reads were used to generate four different datasets. i) BIOM tables of archaea, bacteria, fungi, phages and plasmids; ii) metagenome assembled genomes (MAGs); iii) abundance table of reads mapping against the antibiotic resistance gene database; and iv) annotated genes and gene-fragments with their abundance.

Ad i) In more detail, filtered reads were classified with Kraken2 v2.0.9 [82] (confidence = 0.1) against archaeal, bacterial and fungal genomes downloaded from the NCBI Reference Sequence Database (RefSeq), using the kraken-build routine. The fungal sequence database was then integrated with all fungal genomes available in GenBank, downloaded using the “ncbi-genome-download” script [83]. The plasmidic sequence database was built upon data available from NCBI’s RefSeq repository. Bacteriophagous viral sequences were downloaded from GenBank using the NCBI’s E-utilities [84]. All sequences were downloaded, and databases were built between July and August 2020. Abundance estimation of Kraken2 results was inferred using Bracken v2.6.0 [85] and BIOM tables were generated using the kraken-biom v1.0.1 [86] utility.

Ad ii) Classified archaeal, bacterial and fungal reads, respectively, were then assembled using MEGAHIT v1.2.9 [87]. Metagenomic assemblies were performed following the recommended settings for low-depth soil metagenomic data and only assembled contigs with a minimum length of 1000 bp were kept. Contigs were later checked by BLAST search against the entire NCBI nt database (downloaded in August 2020) and hits were processed employing BlobTools v1.1.1 [88]. Gene prediction and annotation of archaeal and bacterial contigs were carried out with MetaProdigal v2.6.3 [89] and Prokka v1.14.5 [90]. Gene prediction of fungal contigs was performed with GeneMark-ES Suite v4.33 [91]. The binning of metagenomic contigs was carried out with MetaBat 2 [92] and MaxBin v2.2.7 [93]. For MetaBat 2, an iterative strategy was adopted by looping the binning with all possible combinations of values for “—maxP” (percentage of contigs for binning) and “—minS” (minimum edge score for binning) in a range of min = 60 and max = 95, with an increment of 5, whereas for the “—maxEdges” parameter (maximum number of edges per node), the values ranged between min = 200 and max = 500, with an increment of 50 at each loop. Each resulting binning set was than evaluated with CheckM [94] by considering a completeness ≥ 50%, a contamination < 10% and total number of bins. For MaxBin, a probability threshold of 0.8 was chosen. For both MetaBat 2 and MaxBin, the minimum required length of each MAG was set to 1500 bp. MetaBat 2 and MaxBin outputs, respectively, were then combined using DAS Tool v1.1.2 [95], with a score threshold of 0.25. Final MAGs quality was then assessed using CheckM and taxonomy classification was assigned with GTDB-Tk v1.3.0 [96].

Ad iii) The antimicrobial resistance (AMR) reference gene data (PRJNA313047) were downloaded on 11th of June 2020 and sequences were used to build a BLAST database. Magic-BLAST [97] was utilized to map previously classified Kraken-Bracken reads against the AMR database, with a similarity of 99%. Alignment files were processed with SAMtools v1.10 [98]and BamTools v2.5.1 [99] and mapping results were used to build a table for statistical analysis.

Ad iv) Magic-BLAST [97] was used to map classified reads against a database built upon previously predicted genes from bacterial metagenomic contigs. BAM alignment files were processed, and a table was generated for further analysis. Functional annotation of predicted proteins was carried out with eggNOG v2.0.3 [100] against the eggNOG DB v5.0.1 [101] and the output was filtered as follows: we excluded i) non-bacterial tax levels ii) unspecific gene functions (matching to more than three different KEGG orthologous) and iii) KEGG functions related to human diseases. The annotation of gene clusters for secondary metabolite biosynthesis was performed with antiSMASH v5.2.0 [102].

Data analysis

Data analysis was done in R studio using the packages phyloseq [103], tidyverse [104] to organize data and vegan [105] and RAM for ecological measurements. Bray-Curtis distance estimated the bacterial community dissimilarities between the individual samples. The resulting beta-diversity was visualized through nonmetric multidimensional scaling (NMDS), principal coordinate analysis (PCoA) or boxplots. A general linear model evaluated the coefficients of stress-treatment, sample-type or genotype for the NMDS or PCoA-scores. Analysis of variance (ANOVA) tested the relevance of this model for our data. Species richness corresponds to the number of distinct ASVs and alpha diversity to the Shannon index. Significant differences in the alpha diversity and microbial species richness between sample types were calculated using the Wilcoxon test for p-value smaller 0.05. For further analyses reproducible reoccurring core microbiota (detected in at least three out of four replicates) were considered. The taxonomic networks determined by metacoder R-package [106] used the Wilcoxon-test for colouring significant differences in taxonomic abundance at each taxonomic rank. Fold changes for ASVs were calculated by DESeq2 [107] and distinct ASVs selected by a false discovery rate (FDR) smaller than 0.05. Core ASVs in samples under stressed conditions that correlated in normalized abundance with phenotypic stress responses of tetraploid potatoes were determined via spearman correlations and a p-value smaller than 0.05. Stress response per genotype are the percentual change between yield, foliage mass, half-time canopy closure, number young tubers and diameter of largest tuber under stressed and non-stressed conditions. Abundance tables of the shotgun dataset for taxa, MAGs, antibiotic resistance genes and eggNOG annotated-genes were analysed in the same way as ASV-tables but differentially filtered. We considered taxa that occurred in two out of three replicates, antibiotic resistance genes with a coverage larger than 20% in two out of three replicates, eggNOG gene fragments with a coverage larger than 50% in two out of three replicates. To identify distinct abundances at a higher functional level the associated reads were summarized and used for Foldchange analysis [107]. To identify over-represented functions within one taxonomic group we combined per sample sequences assigned either to i) Actinobacteria, ii) Alphaproteobacteria or iii) Beta-/and Gammaproteobacteria and their lower taxonomic ranks by eggNOG (best taxonomic level). Normalized reads were DESeq2’s median of ratios. In the KEGG-database each pathway consists of several KEGG-orthologues. Distinct abundant pathways were considered if we detected at least five differentially abundant gene-fragments that represented at least five distinct KEGG-orthologues. Additionally, KEGG functions at Level C were removed if less than 20% of known KEGG-orthologues were present in our dataset. All significant differentially abundant assembled genes were assigned to pathways by the web-tool KEGG Mapper-Search&Color Pathway [108] to identify submodules. Analyses were visualized in R by ggforce [109], pheatmat [110], formattable [111] and ggplot2 [112]. Layout was adapted in Inkscape (https://inkscape.org).

Ethics approval and consent to participate

Not applicable

Consent for publication

Not applicable

Competing interests

The authors declare that they have no competing interests

Authors’ contributions

AS conceived and designed the research plan. HF, PW, SS and FT performed research. LA supported the bioinformatics analysis, which was mostly done by HF. HF wrote the paper with contributions of AS and FT. All authors read and approved the final manuscript.

Funding

This study is part of the SolACE (https://www.solace-eu.net/) project and has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement N°727247.

Availability of data and materials

The datasets generated and analysed during the current study are available in the NCBI repository, in the bioproject PRJNA720816.

Acknowledgements

We thank Branislav Nikolic and Theresa Ringwald form AIT for their support in laboratory work. We also acknowledge Natascha Bodenhausen from the Research Institute of Organic Agriculture (FiBL) for help in designing the sampling protocol. We thank our colleagues from the James Hutton Institute (Lawrie Brown, Timothy George, Susan Mitchell, David Roberts, Jacqueline Thompson and Gladys Wright) for potato cultivation, stress application, phenotypic potato trait measurements and the support in the laboratory. Finally, we are very grateful to the whole SolACE-team, who gave feedback at early stages of the study.

Khan A, Pan X, Najeeb U, Tan DKY, Fahad S, Zahoor R, et al. Coping with drought: stress and adaptive mechanisms, and management through cultural and molecular alternatives in cotton as vital constituents for plant stress resilience and fitness. Biol Res. 2018;51(1):47. doi:10.1186/s40659-018-0198-z.
Berg G, Rybakova D, Fischer D, Cernava T, Verges MC, Charles T, et al. Microbiome definition re-visited: old concepts and new challenges. Microbiome. 2020;8(1):103. doi:10.1186/s40168-020-00875-0.
Singh BK, Trivedi P, Egidi E, Macdonald CA, Delgado-Baquerizo M. Crop microbiome and sustainable agriculture. Nat Rev Microbiol. 2020;18(11):601–2. doi:10.1038/s41579-020-00446-y.
Berendsen RL, Pieterse CM, Bakker PA. The rhizosphere microbiome and plant health. Trends Plant Sci. 2012;17(8):478–86. doi:10.1016/j.tplants.2012.04.001.
Bulgarelli D, Schlaeppi K, Spaepen S, Ver Loren van Themaat E, Schulze-Lefert P. Structure and functions of the bacterial microbiota of plants. Annu Rev Plant Biol. 2013;64:807–38. doi:10.1146/annurev-arplant-050312-120106.
Hardoim PR, van Overbeek LS, Berg G, Pirttila AM, Compant S, Campisano A, et al. The Hidden World within Plants: Ecological and Evolutionary Considerations for Defining Functioning of Microbial Endophytes. Microbiol Mol Biol Rev. 2015;79(3):293–320. doi:10.1128/MMBR.00050-14.
Philippot L, Raaijmakers JM, Lemanceau P, van der Putten WH. Going back to the roots: the microbial ecology of the rhizosphere. Nat Rev Microbiol. 2013;11(11):789–99. doi:10.1038/nrmicro3109.
Compant S, Cambon MC, Vacher C, Mitter B, Samad A, Sessitsch A. The plant endosphere world - bacterial life within plants. Environ Microbiol. 2020. doi:10.1111/1462-2920.15240.
Vorholt JA. Microbial life in the phyllosphere. Nat Rev Microbiol. 2012;10(12):828–40. doi:10.1038/nrmicro2910.
Trivedi P, Leach JE, Tringe SG, Sa T, Singh BK. Plant-microbiome interactions: from community assembly to plant health. Nat Rev Microbiol. 2020;18(11):607–21. doi:10.1038/s41579-020-0412-1.
Xu L, Coleman-Derr D. Causes and consequences of a conserved bacterial root microbiome response to drought stress. Curr Opin Microbiol. 2019;49:1–6. doi:10.1016/j.mib.2019.07.003.
Pfeiffer S, Mitter B, Oswald A, Schloter-Hai B, Schloter M, Declerck S, et al. Rhizosphere microbiomes of potato cultivated in the High Andes show stable and dynamic core microbiomes with different responses to plant development. FEMS Microbiol Ecol. 2017;93(2); doi:10.1093/femsec/fiw242.
Inceoglu O, Al-Soud WA, Salles JF, Semenov AV, van Elsas JD. Comparative analysis of bacterial communities in a potato field as determined by pyrosequencing. PLoS One. 2011;6(8):e23321. doi:10.1371/journal.pone.0023321.
Ngumbi E, Kloepper J. Bacterial-mediated drought tolerance: Current and future prospects. Appl Soil Ecol. 2016;105:109–25. doi:10.1016/j.apsoil.2016.04.009.
Naveed M, Mitter B, Reichenauer TG, Wieczorek K, Sessitsch A. Increased drought stress resilience of maize through endophytic colonization by Burkholderia phytofirmans PsJN and Enterobacter sp. FD17. Environ Exp Bot. 2014;97:30–9. doi:10.1016/j.envexpbot.2013.09.014.
Naveed M, Hussain MB, Zahir ZA, Mitter B, Sessitsch A. Drought stress amelioration in wheat through inoculation with Burkholderia phytofirmans strain PsJN. Plant Growth Regul. 2013;73(2):121–31. doi:10.1007/s10725-013-9874-8.
Glick BR, Cheng Z, Czarny J, Duan J. Promotion of plant growth by ACC deaminase-producing soil bacteria. Eur J Plant Pathol. 2007;119(3):329–39. doi:10.1007/s10658-007-9162-4.
Glick BR, Penrose DM, Li J. A model for the lowering of plant ethylene concentrations by plant growth-promoting bacteria. J Theor Biol. 1998;190(1):63–8. doi:10.1006/jtbi.1997.0532.
Sessitsch A, Hardoim P, Doring J, Weilharter A, Krause A, Woyke T, et al. Functional characteristics of an endophyte community colonizing rice roots as revealed by metagenomic analysis. Mol Plant Microbe Interact. 2012;25(1):28–36. doi:10.1094/MPMI-08-11-0204.
Sheibani-Tezerji R, Rattei T, Sessitsch A, Trognitz F, Mitter B. Transcriptome Profiling of the Endophyte Burkholderia phytofirmans PsJN Indicates Sensing of the Plant Environment and Drought Stress. mBio. 2015;6(5):e00621-15. doi:10.1128/mBio.00621-15.
Fitzpatrick CR, Mustafa Z, Viliunas J. Soil microbes alter plant fitness under competition and drought. J Evol Biol. 2019;32(5):438–50. doi:10.1111/jeb.13426.
Fitzpatrick CR, Copeland J, Wang PW, Guttman DS, Kotanen PM, Johnson MTJ. Assembly and ecological function of the root microbiome across angiosperm plant species. Proc Natl Acad Sci U S A. 2018;115(6):E1157-E65. doi:10.1073/pnas.1717617115.
Naylor D, DeGraaf S, Purdom E, Coleman-Derr D. Drought and host selection influence bacterial community dynamics in the grass root microbiome. ISME J. 2017;11(12):2691–704. doi:10.1038/ismej.2017.118.
Santos-Medellin C, Edwards J, Liechty Z, Nguyen B, Sundaresan V. Drought Stress Results in a Compartment-Specific Restructuring of the Rice Root-Associated Microbiomes. mBio. 2017;8(4); doi:10.1128/mBio.00764-17.
Leech FJ, Richardson AE, Kertesz MA, Orchard BA, Banerjee S, Graham P. Comparative effect of alternative fertilisers on pasture production, soil properties and soil microbial community structure. Crop Pasture Sci. 2019;70(12); doi:10.1071/cp19018.
Hijmans RJ. The effect of climate change on global potato production. Am J Potato Res. 2003;80(4):271–9. doi:10.1007/bf02855363.
Dahal K, Li XQ, Tai H, Creelman A, Bizimungu B. Improving Potato Stress Tolerance and Tuber Yield Under a Climate Change Scenario - A Current Overview. Front Plant Sci. 2019;10:563. doi:10.3389/fpls.2019.00563.
Aleti G, Nikolic B, Brader G, Pandey RV, Antonielli L, Pfeiffer S, et al. Secondary metabolite genes encoded by potato rhizosphere microbiomes in the Andean highlands are diverse and vary with sampling site and vegetation stage. Sci Rep. 2017;7(1):2330. doi:10.1038/s41598-017-02314-x.
Xiao H, Rodrigues RR, Bonierbale M, Veilleux R, Williams M. Foliar application of Fe resonates to the belowground rhizosphere microbiome in Andean landrace potatoes. Appl Soil Ecol. 2018;131:89–98. doi:10.1016/j.apsoil.2018.08.006.
Gao C, Montoya L, Xu L, Madera M, Hollingsworth J, Purdom E, et al. Fungal community assembly in drought-stressed sorghum shows stochasticity, selection, and universal ecological dynamics. Nat Commun. 2020;11(1):34. doi:10.1038/s41467-019-13913-9.
Teakle DS, Thomas BJ. Effect of heat on zoospore motility and multiplication of Olpidium radicale and O. brassicae. Ann Appl Biol. 1985;107(1):11–5. doi:10.1111/j.1744-7348.1985.tb01542.x.
de Oliveira TB, de Lucas RC, Scarcella ASA, Contato AG, Pasin TM, Martinez CA, et al. Fungal communities differentially respond to warming and drought in tropical grassland soil. Mol Ecol. 2020;29(8):1550–9. doi:10.1111/mec.15423.
Shi H, Zhu JK. SOS4, a pyridoxal kinase gene, is required for root hair development in Arabidopsis. Plant Physiol. 2002;129(2):585–93. doi:10.1104/pp.001982.
Streit WR, Joseph CM, Phillips DA. Biotin and other water-soluble vitamins are key growth factors for alfalfa root colonization by Rhizobium meliloti 1021. Mol Plant Microbe Interact. 1996;9(5):330–8. doi:10.1094/mpmi-9-0330.
Zhang S, Yang X, Sun M, Sun F, Deng S, Dong H. Riboflavin-induced priming for pathogen defense in Arabidopsis thaliana. J Integr Plant Biol. 2009;51(2):167–74. doi:10.1111/j.1744-7909.2008.00763.x.
Palacios OA, Bashan Y, de-Bashan LE. Proven and potential involvement of vitamins in interactions of plants with plant growth-promoting bacteria—an overview. Biol Fert Soils. 2014;50(3):415–32. doi:10.1007/s00374-013-0894-3.
Havaux M, Ksas B, Szewczyk A, Rumeau D, Franck F, Caffarri S, et al. Vitamin B6 deficient plants display increased sensitivity to high light and photo-oxidative stress. BMC Plant Biol. 2009;9:130. doi:10.1186/1471-2229-9-130.
Ponnu J, Wahl V, Schmid M. Trehalose-6-phosphate: connecting plant metabolism and development. Front Plant Sci. 2011;2:70. doi:10.3389/fpls.2011.00070.
Stiller I, Dulai S, Kondrak M, Tarnai R, Szabo L, Toldi O, et al. Effects of drought on water content and photosynthetic parameters in potato plants expressing the trehalose-6-phosphate synthase gene of Saccharomyces cerevisiae. Planta. 2008;227(2):299–308. doi:10.1007/s00425-007-0617-9.
Vilchez JI, Garcia-Fontana C, Roman-Naranjo D, Gonzalez-Lopez J, Manzanera M. Plant Drought Tolerance Enhancement by Trehalose Production of Desiccation-Tolerant Microorganisms. Front Microbiol. 2016;7:1577. doi:10.3389/fmicb.2016.01577.
Reddy GK, Leferink NGH, Umemura M, Ahmed ST, Breitling R, Scrutton NS, et al. Exploring novel bacterial terpene synthases. PLoS One. 2020;15(4):e0232220. doi:10.1371/journal.pone.0232220.
Yamada Y, Kuzuyama T, Komatsu M, Shin-Ya K, Omura S, Cane DE, et al. Terpene synthases are widely distributed in bacteria. Proc Natl Acad Sci U S A. 2015;112(3):857–62. doi:10.1073/pnas.1422108112.
Becher PG, Verschut V, Bibb MJ, Bush MJ, Molnar BP, Barane E, et al. Developmentally regulated volatiles geosmin and 2-methylisoborneol attract a soil arthropod to Streptomyces bacteria promoting spore dispersal. Nat Microbiol. 2020;5(6):821–9. doi:10.1038/s41564-020-0697-x.
Vriet C, Russinova E, Reuzeau C. Boosting crop yields with plant steroids. Plant Cell. 2012;24(3):842–57. doi:10.1105/tpc.111.094912.
Kaya C, Senbayram M, Akram NA, Ashraf M, Alyemeni MN, Ahmad P. Sulfur-enriched leonardite and humic acid soil amendments enhance tolerance to drought and phosphorus deficiency stress in maize (Zea mays L.). Sci Rep. 2020;10(1):6432. doi:10.1038/s41598-020-62669-6.
Eich-Greatorex S, Krogstad T, Sogn TA. Effect of phosphorus status of the soil on selenium availability. J Plant Nutr Soil Sci. 2010;173(3):337–44. doi:10.1002/jpln.200900004.
Li HF, McGrath SP, Zhao FJ. Selenium uptake, translocation and speciation in wheat supplied with selenate or selenite. New Phytol. 2008;178(1):92–102. doi:10.1111/j.1469-8137.2007.02343.x.
Baldan E, Nigris S, Romualdi C, D'Alessandro S, Clocchiatti A, Zottini M, et al. Beneficial Bacteria Isolated from Grapevine Inner Tissues Shape Arabidopsis thaliana Roots. PLoS One. 2015;10(10):e0140252. doi:10.1371/journal.pone.0140252.
Bresson J, Varoquaux F, Bontpart T, Touraine B, Vile D. The PGPR strain Phyllobacterium brassicacearum STM196 induces a reproductive delay and physiological changes that result in improved drought tolerance in Arabidopsis. New Phytol. 2013;200(2):558–69. doi:10.1111/nph.12383.
Aliche EB, Oortwijn M, Theeuwen TPJM, Bachem CWB, Visser RGF, van der Linden CG. Drought response in field grown potatoes and the interactions between canopy growth and yield. Agricult Water Man. 2018;206:20–30. doi:10.1016/j.agwat.2018.04.013.
Choi O, Kim J, Kim JG, Jeong Y, Moon JS, Park CS, et al. Pyrroloquinoline quinone is a plant growth promotion factor produced by Pseudomonas fluorescens B16. Plant Physiol. 2008;146(2):657–68. doi:10.1104/pp.107.112748.
Zhalnina K, Louie KB, Hao Z, Mansoori N, da Rocha UN, Shi S, et al. Dynamic root exudate chemistry and microbial substrate preferences drive patterns in rhizosphere microbial community assembly. Nat Microbiol. 2018;3(4):470–80. doi:10.1038/s41564-018-0129-3.
Castiblanco LF, Sundin GW. New insights on molecular regulation of biofilm formation in plant-associated bacteria. J Integr Plant Biol. 2016;58(4):362–72. doi:10.1111/jipb.12428.
Tetz VV, Korobov VP, Artemenko NK, Lemkina LM, Panjkova NV, Tetz GV. Extracellular phospholipids of isolated bacterial communities. Biofilms. 2005;1(3):149–55. doi:10.1017/s147905050400136x.
Berendsen RL, Vismans G, Yu K, Song Y, de Jonge R, Burgman WP, et al. Disease-induced assemblage of a plant-beneficial bacterial consortium. ISME J. 2018;12(6):1496–507. doi:10.1038/s41396-018-0093-1.
Ren D, Madsen JS, Sorensen SJ, Burmolle M. High prevalence of biofilm synergy among bacterial soil isolates in cocultures indicates bacterial interspecific cooperation. ISME J. 2015;9(1):81–9. doi:10.1038/ismej.2014.96.
Kielak AM, Barreto CC, Kowalchuk GA, van Veen JA, Kuramae EE. The Ecology of Acidobacteria: Moving beyond Genes and Genomes. Front Microbiol. 2016;7:744. doi:10.3389/fmicb.2016.00744.
Guo Q, Shi M, Chen L, Zhou J, Zhang L, Li Y, et al. The biocontrol agent Streptomyces pactum increases Pseudomonas koreensis populations in the rhizosphere by enhancing chemotaxis and biofilm formation. Soil Biol Biochem. 2020;144; doi:10.1016/j.soilbio.2020.107755.
Weber M, Liedtke J, Plattes S, Lipski A. Bacterial community composition of biofilms in milking machines of two dairy farms assessed by a combination of culture-dependent and -independent methods. PLoS One. 2019;14(9):e0222238. doi:10.1371/journal.pone.0222238.
Flemming HC, Wingender J, Szewzyk U, Steinberg P, Rice SA, Kjelleberg S. Biofilms: an emergent form of bacterial life. Nat Rev Microbiol. 2016;14(9):563–75. doi:10.1038/nrmicro.2016.94.
Kumar M, Mishra S, Dixit V, Kumar M, Agarwal L, Chauhan PS, et al. Synergistic effect of Pseudomonas putida and Bacillus amyloliquefaciens ameliorates drought stress in chickpea (Cicer arietinum L.). Plant Signal Behav. 2016;11(1):e1071004. doi:10.1080/15592324.2015.1071004.
Costacurta A, Mazzafera P, Rosato YB. Indole-3-acetic acid biosynthesis by Xanthomonas axonopodis pv. citri is increased in the presence of plant leaf extracts. FEMS Microbiol Letters. 1998;159(2):215–20. doi:10.1111/j.1574-6968.1998.tb12863.x.
Na X, Cao X, Ma C, Ma S, Xu P, Liu S, et al. Plant Stage, Not Drought Stress, Determines the Effect of Cultivars on Bacterial Community Diversity in the Rhizosphere of Broomcorn Millet (Panicum miliaceum L.). Front Microbiol. 2019;10:828. doi:10.3389/fmicb.2019.00828.
Obeng N, Pratama AA, Elsas JDV. The Significance of Mutualistic Phages for Bacterial Ecology and Evolution. Trends Microbiol. 2016;24(6):440–9. doi:10.1016/j.tim.2015.12.009.
Pereira SIA, Abreu D, Moreira H, Vega A, Castro PML. Plant growth-promoting rhizobacteria (PGPR) improve the growth and nutrient use efficiency in maize (Zea mays L.) under water deficit conditions. Heliyon. 2020;6(10):e05106. doi:10.1016/j.heliyon.2020.e05106.
Sharp RG, Chen L, Davies WJ. Inoculation of growing media with the rhizobacterium Variovorax paradoxus 5C-2 reduces unwanted stress responses in hardy ornamental species. Sci Hortic. 2011;129(4):804–11. doi:10.1016/j.scienta.2011.03.016.
Hansen MF, Svenningsen SL, Roder HL, Middelboe M, Burmolle M. Big Impact of the Tiny: Bacteriophage-Bacteria Interactions in Biofilms. Trends Microbiol. 2019;27(9):739–52. doi:10.1016/j.tim.2019.04.006.
Mirzaei MK, Maurice CF. Menage à trois in the human gut: interactions between host, bacteria and phages. Nat Rev Microbiol. 2017;15(7):397–408. doi:10.1038/nrmicro.2017.30.
Wishart J, George TS, Brown LK, White PJ, Ramsay G, Jones H, et al. Field phenotyping of potato to assess root and shoot characteristics associated with drought tolerance. Plant Soil. 2014;378(1–2):351–63. doi:10.1007/s11104-014-2029-5.
Samad A, Trognitz F, Compant S, Antonielli L, Sessitsch A. Shared and host-specific microbiome diversity and functioning of grapevine and accompanying weed plants. Environ Microbiol. 2017;19(4):1407–24. doi:10.1111/1462-2920.13618.
Langmead B, Salzberg SL. Fast gapped-read alignment with Bowtie 2. Nat Methods. 2012;9(4):357–9. doi:10.1038/nmeth.1923.
Andrews S. FastQC: a quality control tool for high throughput sequence data. Available online at: http://www.bioinformatics.babraham.ac.uk/projects/fastqc. 2010.
Martin M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnetjournal. 2011;17(1):10–2. doi:10.14806/ej.17.1.200.
Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJA, Holmes SP. DADA2: High resolution sample inference from Illumina amplicon data. Nat Methods. 2016;13(7):581–3. doi:10.1038/nmeth.3869.
Bengtsson-Palme J, Thorell K, Wurzbacher C, Sjöling Å, Nilsson RH. Metaxa2 Diversity Tools: Easing microbial community analysis with Metaxa2. Ecological Informatics. 2016;33:45–50. doi:10.1016/j.ecoinf.2016.04.004.
Bengtsson-Palme J, Ryberg M, Hartmann M, Branco S, Wang Z, Godhe A, et al. Improved software detection and extraction of ITS1 and ITS2 from ribosomal ITS sequences of fungi and other eukaryotes for analysis of environmental sequencing data. Meth Ecol Evol. 2013;4(10):914–9. doi:https://doi.org/10.1111/2041-210X.12073.
Wang Q, Garrity GM, Tiedje JM, Cole JR. Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl Environ Microbiol. 2007;73(16):5261–7. doi:10.1128/AEM.00062-07.
Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucl Acids Res. 2013;41(Database issue):D590-D6. doi:10.1093/nar/gks1219.
Nilsson RH, Larsson K-H, Taylor AFS, Bengtsson-Palme J, Jeppesen TS, Schigel D, et al. The UNITE database for molecular identification of fungi: handling dark taxa and parallel taxonomic classifications. Nucl Acids Res. 2019;47(D1):D259-D64. doi:10.1093/nar/gky1022.
Chen S, Zhou Y, Chen Y, Gu J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics. 2018;34(17):i884-i90. doi:10.1093/bioinformatics/bty560.
Ewels P, Magnusson M, Lundin S, Käller M. MultiQC: summarize analysis results for multiple tools and samples in a single report. Bioinformatics. 2016;32(19):3047–8. doi:10.1093/bioinformatics/btw354.
Wood DE, Lu J, Langmead B. Improved metagenomic analysis with Kraken 2. Genome Biol. 2019;20(1):257. doi:10.1186/s13059-019-1891-0.
Blin K. kblin/ncbi-genome-download, https://github.com/kblin/ncbi-genome-download. In.; 2019.
Sayers E. A General Introduction to the E-utilities. National Center for Biotechnology Information (US); 2010.
Lu J, Breitwieser FP, Thielen P, Salzberg SL. Bracken: estimating species abundance in metagenomics data. PeerJ Comput Sci. 2017;3:e104. doi:10.7717/peerj-cs.104.
Dabdoub S. smdabdoub/kraken-biom https://github.com/smdabdoub/kraken-biom. In.; 2021.
Li D, Liu C-M, Luo R, Sadakane K, Lam T-W. MEGAHIT: an ultra-fast single-node solution for large and complex metagenomics assembly via succinct de Bruijn graph. Bioinformatics. 2015;31(10):1674–6. doi:10.1093/bioinformatics/btv033.
Dominik RL, Georgios K, Tim B, Jason S, Sujai K. BlobTools v1.0.1. In.: Zenodo; 2017.
Hyatt D, Chen G-L, LoCascio PF, Land ML, Larimer FW, Hauser LJ. Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinformatics. 2010;11:119. doi:10.1186/1471-2105-11-119.
Seemann T. Prokka: rapid prokaryotic genome annotation. Bioinformatics. 2014;30(14):2068–9. doi:10.1093/bioinformatics/btu153.
Borodovsky M, Lomsadze A. Eukaryotic Gene Prediction Using GeneMark.hmm-E and GeneMark-ES. Curr Protoc Bioinformatics. 2011;CHAPTER:Unit-4.610; doi:10.1002/0471250953.bi0406s35.
Kang DD, Li F, Kirton E, Thomas A, Egan R, An H, et al. MetaBAT 2: an adaptive binning algorithm for robust and efficient genome reconstruction from metagenome assemblies. PeerJ. 2019;7; doi:10.7717/peerj.7359.
Wu Y-W, Simmons BA, Singer SW. MaxBin 2.0: an automated binning algorithm to recover genomes from multiple metagenomic datasets. Bioinformatics. 2016;32(4):605–7. doi:10.1093/bioinformatics/btv638.
Parks DH, Imelfort M, Skennerton CT, Hugenholtz P, Tyson GW. CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes. Genome Res. 2015;25(7):1043–55. doi:10.1101/gr.186072.114.
Sieber CMK, Probst AJ, Sharrar A, Thomas BC, Hess M, Tringe SG, et al. Recovery of genomes from metagenomes via a dereplication, aggregation and scoring strategy. Nat Microbiol. 2018;3(7):836–43. doi:10.1038/s41564-018-0171-1.
Chaumeil P-A, Mussig AJ, Hugenholtz P, Parks DH. GTDB-Tk: a toolkit to classify genomes with the Genome Taxonomy Database. Bioinformatics. 2020;36(6):1925–7. doi:10.1093/bioinformatics/btz848.
Boratyn GM, Thierry-Mieg J, Thierry-Mieg D, Busby B, Madden TL. Magic-BLAST, an accurate RNA-seq aligner for long and short reads. BMC Bioinformatics. 2019;20(1):405. doi:10.1186/s12859-019-2996-x.
Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics. 2009;25(16):2078–9. doi:10.1093/bioinformatics/btp352.
Barnett DW, Garrison EK, Quinlan AR, Strömberg MP, Marth GT. BamTools Bioinformatics. 2011;27(12):1691–2. doi:10.1093/bioinformatics/btr174.
Huerta-Cepas J, Forslund K, Coelho LP, Szklarczyk D, Jensen LJ, von Mering C, et al. Fast Genome-Wide Functional Annotation through Orthology Assignment by eggNOG-Mapper. Mol Biol Evol. 2017;34(8):2115–22. doi:10.1093/molbev/msx148.
Huerta-Cepas J, Szklarczyk D, Heller D, Hernandez-Plaza A, Forslund SK, Cook H, et al. eggNOG 5.0: a hierarchical, functionally and phylogenetically annotated orthology resource based on 5090 organisms and 2502 viruses. Nucl Acids Res. 2019;47(D1):D309-D14. doi:10.1093/nar/gky1085.
Medema MH, Blin K, Cimermancic P, de Jager V, Zakrzewski P, Fischbach MA, et al. antiSMASH: rapid identification, annotation and analysis of secondary metabolite biosynthesis gene clusters in bacterial and fungal genome sequences. Nucl Acids Research. 2011;39(Web Server issue):W339-W46; doi:10.1093/nar/gkr466.
McMurdie PJ, Holmes S. phyloseq: An R Package for Reproducible Interactive Analysis and Graphics of Microbiome Census Data. Plos One. 2013;8(4); doi: ARTN e6121710.1371/journal.pone.0061217.
Wickham H, Averick M, Bryan J, Chang W, McGowan L, François R, et al. Welcome to the Tidyverse. Journal of Open Source Software. 2019;4(43); doi:10.21105/joss.01686.
Oksanen J, Blanchet FG, Friendly M, Kindt R, Legendre P, McGlinn D, et al.: vegan: Community Ecology Package. In.; 2020: R package version 2.5-7 https://CRAN.R-project.org/package=vegan.
Rice P, Longden I, Bleasby A. EMBOSS: the European Molecular Biology Open Software Suite. Trends Genet. 2000;16(6):276–7. doi:10.1016/s0168-9525(00)02024-2.
Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550. doi:10.1186/s13059-014-0550-8.
Kanehisa M, Sato Y. KEGG Mapper for inferring cellular functions from protein sequences. Protein Sci. 2020;29(1):28–35. doi:10.1002/pro.3711.
Pedersen TL: ggforce: Accelerating 'ggplot2'. In.; 2020: R package version 0.3.2. https://CRAN.R-project.org/package=ggforce.
Kolde R: pheatmap: Pretty Heatmaps. In.; 2019: R package version 1.0.12. https://CRAN.R-project.org/package=pheatmap.
Ren K, Russell K: formattable: Create 'Formattable' Data Structures. In.; 2021: R package version 0.2.1. https://CRAN.R-project.org/package=formattable.
Ginestet C. ggplot2: Elegant Graphics for Data Analysis. J R Stat Soc a Stat. 2011;174:245-; doi: DOI 10.1111/j.1467-985X.2010.00676_9.x.

Download PDF

Version 1

posted

You are reading this latest preprint version

Potato Root-Associated Microbiomes Adapt to Combined Water and Nutrient Limitation and have a Plant Genotype-Specific Role for Plant Stress Mitigation

Status:

Version 1

Abstract

Background

Results

Conclusion

Figures

Background

Results

Discussion

Conclusion

Methods

Data analysis

Declarations

Ethics approval and consent to participate

Consent for publication

Competing interests

Authors’ contributions

Funding

Availability of data and materials

Acknowledgements

References

Supplementary Files

Status:

Version 1