A support vector machine algorithm was used to define 200 genes whose expression in diagnostic leaves allowed the most accurate discrimination between P-replete and P-starved plants grown hydroponically in the glasshouse. Table 1 lists the twenty genes whose expression in diagnostic leaves discriminates with the greatest power between P-replete and P-deficient potato plants growing hydroponically in the glasshouse. As an independent test of the diagnostic transcriptome, diagnostic leaves were sampled from plants growing in the field at HRI-Wellesbourne. An analysis of gene expression in these leaves correctly identified that this potato crop was P-deficient. Thus, the diagnostic transcriptome can identify P-deficiency in field-grown potato crops and, therefore, has potential as a viable management tool for precision agriculture. It will allow farmers to monitor the immediate physiological P status of their crops and, thereby, optimise Pi-fertiliser applications and reduce P losses to the environment and commercial losses due to P-deficiency. Plant roots shape their environment in various ways and are in turn shaped by physiochemical properties of the surrounding soil. Roots affect soil by dislocating particles, by polymer production,vertical growing and by the release of a wide variety of small molecules.Overall, these processes result in the formation of larger soil aggregates which increase water-holding capacity and influence the stability of soil organic carbon . Plant induced changes in chemistry can lead to weathering of minerals and alter the composition of microbial communities . Further, root exudates can enhance root penetration of soils .

Although the effects of plant presence on soils and microbial communities have been a major research questions for over a century , relatively few studies have sought to understand the physiochemical effects of plant growth substrates on plant physiology and exudation. Plant morphology and exudations can be influenced by both physical and chemical properties of soil particles. Typically, particles range from small to large and determine physical parameters such as water-holding capacity of soils . It has been shown that 1-mm beads reduce root and shoot growth, elevate root:shoot ratios, and alter root morphology of maize when compared to hydroponic growth . Root morphology can be altered by adsorption of root exudate metabolites to substrate such as activated carbon . Natural environments, such as soils, can also affect both root exudate profiles and morphology ; however, determination of causal factors may be confounded by variables such as granule size, chemistry, and microbial community composition. Soil minerals differ in structure and surface charge, thereby governing their interaction with dissolved organic compounds . Results suggest that substrate chemistry can alter exudation. For example, aluminum ions present in stone wool are thought to increase exudation of organic acids in maize . In contrast to particle chemistry, the effect of particle size on exudation is less clear. Particle size could alter exudation in multiple ways. Particle size reduces root growth rates, which limits exudate dispersal. In addition, diffusion depends on substrate size and can thus be a limiting factor for dispersal of exuded compounds . Exudates are mainly produced by root tips : when root morphology is altered by particle size, the number and structure of root tips can be altered, possibly changing the quality and quantity of exudates. Particle size also influences the composition of microbial communities .

Microbes differentially metabolize exudates and secrete secondary metabolites , further changing the presence of exuded compounds. d) Particle size can modulate plant chemistry . Here, we investigate the effect of particle size and chemistry on root morphology and exudation in Brachypodium distachyon. Specifically, we asked three questions: a) whether root morphology of a model grass is altered in physically restricted conditions as observed in other species, b) if and how the exudate profile changes with particle size, and c) how root morphology and exudation are influenced by substrate chemistry. To facilitate these studies, we used a sterile environment enabling us to focus on plant metabolism without the additional layer of microbial metabolism present in a natural environment. B. distachyon growth and exudation profiles were comparted on various inert substrates in a range of defined particle sizes relative to hydroponic solution. We found that particle size had a significant effect on root weight and root length; however, particle size did not influence composition of root exudates. A defined mixture of soil metabolites was used to evaluate sorption to particles. We further demonstrated that clay sorbed a large degree of exudates, altering the amount of exudates freely available around root. These clay-sorbed exudates could further support growth of a rhizobacterium. Our results highlight the importance of considering soil structure and chemistry when studying plant–soil interactions.Different particle sizes and surface chemistries were chosen to investigate how root morphology and exudation is affected in various plant growth substrates. The particle sizes chosen corresponded to large, intermediary, and small particles . Glass beads were chosen as an inert system with defined sphere diameter, sand as an inert and natural system, and clay as a natural system with a reactive surface. The mineral composition of the sand substrates was determined as more than 98% quartz, whereas the clay was a mixture of 51% opalCT, 37% mica-illite, 10% quartz, and trace amounts of K-feldspar and calcite . The chemical properties of the substrates were assessed by determining the sorption of a mixture of more than forty metabolites belonging to various chemical classes that have also been found in root exudates .

The recovery rate of the various metabolites from the glass beads and the 4-mm sand was comparable to the defined medium control without substrate, whereas the recovery rate from the250-µm sand and from clay were lower by approximately 30% and 70%, respectively . Consistently, differences between clay and other substrates explained 84% of the variance in a principal component analysis, and only 8% of the variance accounts for differences between the control, glass beads, and sand . The metabolites depleted by clay belonged to a variety of chemical classes, among them charged compounds, such as organic acids and ammonium salts , and other nitrogenous compounds , and of comparatively polar compounds such as sugars . We confirm that as expected, clay particles sorb a variety of metabolites from the defined medium.Exudation could be limited by diffusion in our experimental systems with small particle sizes. A diffusion test with a dye resulted in fastest diffusion in controls without substrate added . The diffusion rate of the dye decreased with lower diameters in glass beads and followed a logarithmic trend. In 4-mm sand, the diffusion rate was initially similar to 1-mm glass beads, but then resembled more 0.5-mm beads. For clay, diffusion similar to 1-mm or 0.5-mm beads was observed initially, but subsequently, the dye front ceased moving, likely due to sorption of the negatively charged dye. Based on this analysis, in our experimental setup, exudates would require a minimum diffusion rate of 1.25 cm/h to reach the edge of the glass jar in which the plants were grown . Thus, diffusion was not limiting in glass beads with a diameter equal to or >1 mm, but might be limited in substrates with smaller diameters . This confirms, that as expected, sand and glass beads are inert substrates, whereas clay strongly sorbs a variety of metabolites. In addition, exudation may be limited by diffusion in substrates with particle sizes smaller than 1 mm.The aforementioned substrates were used to investigate how Brachypodium distachyon root morphology and exudation was affected in these experimental systems compared to a hydroponic control. Plants were grouped according to their behavior in the different substrates: plants with weights and root morphology similar to hydroponic controls were termed “big” , and plants with distinct weight and root morphology were termed “small” . The root fresh weight of plants grown in 3-mm glass beads, 4-mm sand, 250-µm sand, and clay was comparable to the hydroponic control, whereas roots grown in 2-mm, 1-mm, and 0.5-mm glass beads and 5-µm sand were significantly smaller . The shoot fresh weight of plants grown in 5-µm sand were significantly smaller compared with plants grown in hydroponics, and all other conditions . The altered root and shoot weights resulted in decreased root/shoot ratios for clay,vertical farm system and 2-mm and 1-mm glass beads-grown plants, and an increased ratio of 5-µm sand-grown plants . Root length and number were assessed for first-order roots , second-order roots , and higher order roots. The total root length correlated with particle size, with maximal length for hydroponically and clay grown roots, approximately 30% shorter root systems for 3-mm beads-, 4-mm sand- and 250-µm sand-grown roots, and 50% or shorter root systems for 1-mm beads-, 2-mm beads-, 0.5-mm beads-, and 5-µm sand-grown roots . First-order root lengths were significantly decreased by more than half for all substrates except for clay and 4-mm sand , whereas the second-order root length was decreased by 40%–70% in 2-mm beads-, 1-mm beads-, and 0.5-mm bead-grown roots, and by ~85% in 5 µm sand-grown roots . Higher order root lengths varied more within one experimental treatment, with a trend for higher total lengths in hydroponics and clay compared with glass beads and sand, and significantly lower lengths in 5-µm sand . Interestingly, root length had a higher Pearson correlation coefficient when correlated with particle size than root numbers. Only roots grown in 1-mm beads and 5-µm sand showed a statistically significant reduction in root number compared with hydroponic controls, which is a result of the large variability in total root number of hydroponically grown plants . The observed reduction in root number originated from a reduced number of secondary and higher order roots.

A correlation analysis between root and shoot weight, total root length, and total root number of all samples showed a significant correlation of all parameters investigated . Root weight and length, and to a lesser degree root number, correlated with particle size. Overall, clay-grown plants were most similar to hydroponically grown plants regarding tissue weight and root morphology. Plants grown in 3-mm glass beads or 4-mm sand had comparable fresh weight compared to the aforementioned plants, but slightly reduced total root length driven by a reduction in first-order root length. Plants grown in 1-mm and 0.5-mm glass beads exhibited reduced root weight and root length, caused by a reduction in first and second-order root length. Plants grown in 5-µm sand exhibited the largest reduction in tissue weight, root length, and number.To investigate whether changes in root morphology might affect exudation profiles, spatial patterns of exudation were investigated using matrix-assisted laser desorption/ionization mass spectrometry . A total of 24 ions were detected in the vicinity of roots . It was not possible to confidently identify these ions given that the MALDI used is not suited for fragmentation of low m/z ions and the fact that MALDI often results in different ions versus the electrospray ionization used in our liquid chromatography–mass spectrometry analyses. However, despite this lack of identifications, our results suggest differences in spatial patterns of chemical components. Some ions showed higher abundances around root tip and elongation zone, supporting a role of these young root tissues in exudation. Other ions were detected along most of the root axis, suggesting exudation also from older root parts, whereas the location of other ions overlapped with the location of the root, which could either indicate short diffusion distances, or association with the cell wall. Overall, these data suggest that multiple tissues are involved in exudation.To investigate whether altered root morphology and various spatial exudation patterns altered overall exudation profiles, exudates were collected in situ and in vitro . The first collection approach generated exudation profiles shaped by plant metabolism and particle chemistry, whereas the second collection approach generated exudate profiles shaped only by plant metabolism. With both collection steps, approximately 100 metabolites were identified based on comparison of retention times, exact mass, and MS/MS fragmentation patterns compared with authentic standards. The metabolites included organic acids, amino acids and derivatives, sugars and other carbohydrates, nucleic bases, nucleosides, and derivatives. Multivariate statistical analysis revealed that 46% of in situ exuded metabolites were significantly different in pairwise comparisons, compared with 30% of in vitro exuded metabolites . A similar result is evident in hierarchical clustering analysis, with in situ-collected exudates clustering according to biological replicates, and in vitro-collected exudates clustering in less distinct patterns .