1.1 Spatial Considerations

The studies of our literature search comprised approximately 100 plant species (Figure 3). Three quarters of articles focused on a single family and 68% of studies (90% of the aforementioned studies) on a single species. A third of the publications covered four crops, the monocots wheat, rice, maize and the dicot tomato. The ratio of monocots and dicots was well balanced, and less than 5% analysing tree exudates. A wide range of plant ages was used for exudate collection, ranging from 2 days to 30 years, with a median age of 35 days for monocots, 28 days for dicots and 2.5 years for trees (Figure 3). Of note, the cultivar or ecotype used was only reported in 22% of publications, making the reproduction of results difficult.

1.2 Traditional Root Exudate Collection Methods

Whole root system exudate collection is a straightforward approach amenable to plants grown in a variety of conditions (e.g., in soil or hydroponic conditions). Although convenient and largely employed, no spatial information can be inferred. Yet, exudation patterns differs spatially, as observed when collecting exudates from specific root sections (Peñaloza, Corcuera, and Martinez 2002). A root-section-specific approach showed differential malate and citrate exudation in phosphate-starved white lupin roots, with malate being exuded by various root types, and citrate being exuded exclusively by cluster roots, specialized root structures involved in phosphate uptake (Peñaloza, Corcuera, and Martinez 2002). Split-root settings can also be an alternative when studying the spatial characteristics of root exudates (Gargallo-Garriga et al. 2018). In this method, the root system of a single plant is divided into two or more sections, each grown in its own compartment. Using the same conditions across compartments allows to independently characterize root exudation of different roots. Alternatively, distinct environmental conditions allows characterization of localized or systemic exudation responses. Further, sectioning specific roots to collect exudates demonstrated that exudation patterns varies with root characteristics and segments. For instance, higher malate exudation was found in axial than in lateral roots in wheat and barley (Kawasaki et al. 2018). However, the wounding caused by cutting roots, root sections, or by removing roots from their substrate can profoundly affect the metabolic profiles obtained, due to a leakage of intracellular compounds during exudate collection. Such methods may thus not reflect the plant natural exudation patterns under undisturbed conditions and should be taken into account while interpretating the obtained data.

Differential spatial exudation shapes microbiome composition: distinct exudation by specialized root structures such as cluster roots causes association with a specialized microbial community (Cheng et al. 2011; Weisskopf, Heller, and Eberl 2011; Zhang et al. 2023). Different root zones harbour distinct microbial communities (Kawasaki et al. 2016; Massalha et al. 2017). Interestingly, although root tips support highest microbial numbers (DeAngelis et al. 2009), the microbial diversity is lower compared to the root base (Acharya et al. 2023). The elongation zone is likely a major zone of exudation due to its low suberization levels, and is often associated with many bacteria (Kawasaki et al. 2016; Massalha et al. 2017). Sites of lateral root emergence release metabolites due to the wounding of the epidermal layer and thus similarly attract microbes, likely resulting in the distinct microbial profile associated with lateral roots (Kawasaki et al. 2016; Massalha et al. 2017).

While the impact of distinct exudation patterns on microbial community structure and abundance is becoming evident, understanding the spatial distribution of exudates remains in its infancy and may require advanced high-resolution techniques.

1.3 Advanced High-Resolution Techniques for Root Exudate Analysis

High-resolution techniques, such as solid-phase extraction methods, chemical imaging, matrix assisted laser desorption/ionization (MALDI), liquid extraction surface analysis (LESA) and liquid microjunction surface-sampling probe mass spectrometry (LMJ-SSP-MS), represent significant advancements in understanding the spatial specificity and temporal dynamics of root exudation at unprecedented scales.

An efficient method to investigate the spatial specificity of root exudation is to use solid-phase extraction methods in rhizotrons. For example, microtubes made from polydimethylsiloxane (PDMS) can be placed at different locations close to a root system, and solvent passing through the tubing collects and extracts metabolites for downstream analytical procedures. PDMS selectively sorbs nonpolar compounds and has been successfully used in different forms to extract lipophilic compounds such as thiophenes, quinone, or sorgoleone from the rhizosphere (Dayan, Howell, and Weidenhamer 2009; Mohney et al. 2009; Weidenhamer, Boes, and Wilcox 2009; Weidenhamer et al. 2014). To our knowledge, no similar method has been developed for polar compounds. One may consider developing a microtube consisting of polar polymer with similar porosity than PDMS to absorb charged, polar compounds using a polar solvent (e.g., water, alcohol). However, the main limitation will arise from the polarity characteristics of the soil, as the migration of polar compounds from the soil matrix to a polar mobile phase may be limited.

Chemical imaging through MALDI can be employed to characterize the exudate profile of a specific root. With this technique, a laser moves across a surface of an object of interest (such as a root with exudates), releasing ions that are detected with mass spectrometry (MS). Limitations for exudate studies are the 3D-structure of roots distorting the image, low molecule concentrations in exudates (improving with MALDI2), and compound identification (no ion fragmentation, thus sometimes couples with liquid chromatography mass spectrometry (LC-MS). Despite these technical limitations, MALDI is a valuable tool to detect heterogenous metabolite signals in roots. Thin roots can be imaged directly: they can be incubated on a thin slice of agar before drying to image root surface and exudation. For Brachypodium distachyon seedlings, distinct metabolite patterns were associated with root surface and surrounding, root tip and base after 2 h incubation (Sasse et al. 2020). Also, Arabidopsis plants grown in phosphate-sufficient and -deficient conditions exhibited distinct organic acid exudation when incubated on a wet nylon membrane for 7 h. Malate was mostly exuded around the root tip, and citrate from older tissues (Gomez-Zepeda et al. 2021). Radial metabolite signals can be studied with root sections. Distinct signals were detected in Bacillus amyloliquefaciens biofilms of 3-week-old roots of Arabidopsis, tomato, and tobacco roots (Debois et al. 2014). Sectioning of Medicago truncatula roots connected to nodules elucidated distinct spatial patterns for primary metabolites (organic acids, amino acids, sugars and lipids) and secondary metabolites (flavonoids), aligning with different stages of nitrogen fixation (Ye et al. 2013).

A method to create high-resolution data faster than with MALDI and allowing probing live tissue is liquid extraction surface analyses (LESA) using nanoESI coupled to MS (Robert et al. 2012). The technique consists in depositing a 1 μL solvent droplet on a specific target of the root (resolution of 1 mm²) for 5 s, followed by aspiration and MS analysis (Robert et al. 2012). The technique permits a detailed characterization of spatial exudation patterns, given that the droplet stays on the roots, which is challenging for small structures such as fine roots (< 2 mm diameter). The solvent choice is critical, as some solvents may not only extract root surface chemicals, but also trigger the release of internal compounds. Distinct benzoxazinoid signals were detected on maize crown and primary roots (Robert et al. 2012). However, whether the collection method would be sufficiently robust to collect low concentrations of exudates or exudates from fine roots remains to be investigated. A variation of LESA is liquid micro junction-surface sampling probe MS (LMJ-SSP-MS), in which roots are grown in a microfluidic framework. Nanoliter samples are removed from denoted areas for MS analysis. Poplar seedlings were grown for several weeks in this device. They exhibited differential distribution of amino acids across the root surface (Walton et al. 2022).

For soil-grown roots, exudation can be investigated by adding a membrane to a region of interest. In switchgrass, MALDI/LC-MS of polyvinylidene fluoride (PVDF) membranes revealed that purines, pyrimidines and phytohormones were associated with ellipse-like zones in proximity of the main root (Veličković et al. 2020). Further, in situ, label-free imaging methods are being developed for native environments. Laser ablation-isotope ratio mass spectrometry (LA-IRMS) allows for studying C12/C13 ratios in solid samples, such as plants in soil, down to a resolution of 10 µm (Bruneau et al. 2002; Rodionov et al. 2019). The depth that can be imaged is currently in the 10–60 µm range, which is sufficient to investigate microbial, symbiotic and root structures in a flat experimental setup (Lee et al. 2022). Imaging of roots in soil remains challenging, due to the background of soil and microbial metabolites, and the solid soil matrix. The choice of the ideal imaging method depends on the experimental setup and must be carefully considered.

Overall, spatial studies reveal a large heterogeneity for primary and secondary metabolites around roots, and in various plant–microbe interactions. Exudate studies with spatial resolution are an exciting new avenue of research and could provide information about spatial microbiome structure, biofilm formation and soil heterogeneity to get a comprehensive picture of below-ground metabolite distributions. However, the use of high-resolution techniques requires a specialized equipment, expertise, and may be costly to establish and maintain. Additionally, investigating the interconnectedness of spatial exudation and differential microbiome requires matching scales for exudation and microbiome analyses to allow drawing any conclusions. The choice of the appropriated exudate collection method must thus be carefully considered according the biological question and available analytical capabilities.

1.4 Growth Environment: Lab Versus Field

Laboratory and field setups differ in many abiotic and biotic factors, causing considerable changes in exudate profiles. A laboratory environment provides consistent lighting, temperature, humidity/watering and nutrient control, whereas fields are natural environments with high fluctuations in the aforementioned factors within one experiment, between seasons or years (Table 1). Additionally, in natural soils, factors such as pesticides, pests, or heavy metals need to be considered. The plant nutritional status, symbiont colonization, and defences against pathogens may differ in laboratory versus field setups.

Systematic studies comparing exudate profiles of plants grown in laboratory versus field setups are lacking. To shed some light on potential differences in exudation between these environments, we identified six studies investigating exudation of one wheat variety in growth chambers, greenhouses and in the field (Rehman et al. 2018; McManus et al. 2018; Chen et al. 2019; O'Neal, Vo, and Alexandre 2020; Guo et al. 2021; Qu et al. 2021). The studies featured distinct environmental conditions, growth media, age at harvest, collection media, collection time and analytical methods, which all impact the exudate profile in addition to the mentioned difference in growth environment. All studies included a targeted analysis of organic acids, with 16 organic acids identified overall (Rehman: nine quantitative; Chen: six quantitative, Guo: four qualitative). Three organic acids were detected in two studies, and tartaric acid was quantified in both (Rehman: 922 µg mL⁻¹ exudate, Chen: 0.6 mg C g⁻¹ root). Conversion of the former value with root mass results in 3.5 mg tartaric acid g⁻¹ root, a value five times higher in these Pseudomonas-inoculated, greenhouse-grown plants compared to the field-grown plants. Apart from the mentioned differences in experimental design, additional factors such as temperature, light, water, humidity, soil properties, microbial community distinct between greenhouse and field could cause differences. Several conclusions can be drawn from this comparison. First, it is crucial to first report all biological, environmental and experimental factors in studies so that they can be compared. Second, even if parameters are recorded, they usually differ between studies to a degree that a direct comparison is not possible. Thus, it would be of high interest to systematically compare exudates of plants grown in different environments including sterile hydroponic setups to determine to which degree results from controlled environments can be extrapolated to field conditions. Overall, the various plant growth conditions should be carefully reported in studies as proposed in Table 1.

1.5 Plant Growth Matrix

Here, the plant growth setup is defined as the liquid and solid substances plants grow in. These media comprise natural matrices such as soil, sand, and clay mixtures and artificial setups based on sterile or nonsterile hydroponic nutrient media, and semihydroponic systems with for example, glass beads for physical support. In our literature review, half of the studies used soil and a third a nutrient medium for plant growth (Table S1, Figure 3). Three quarters of studies were conducted in nonsterile environments and thus also contained nonplant metabolites (such as microbial or soil-derived compounds), an aspect rarely discussed in the articles. The use of sterile conditions is most feasible for seedlings, but few studies grew plants to several weeks old in sterile conditions as shown for Arabidopsis thaliana, cucumber, B. distachyon, rice and tomato (Figure 3).

Although plant growth in a soil matrix is arguably preferable because it is a natural environment, this poses many technical challenges for in situ or ex situ exudate collection, for example, the metabolite background found in soils, or the root damage imposed when removing plants from the matrix (Table 1, see ‘collection’ section for full discussion). Growth in a hydroponic medium is artificial, and exudates produced and/or transported in the rhizosphere likely do not mirror natural plant behaviour (Oburger and Jones 2018). However, hydroponic systems make in situ exudate collection and downstream analysis considerably easier due to the low metabolite background of the system. Also, a sterile setup as amenable in many hydroponic systems permits focus on plant metabolites without interference of for example, microbial metabolism (Figure 1, Table 1). The effect of microbial presence on root exudates can be tested by comparing inoculated versus sterile hydroponic plants. Further, to better mimic physical soil structure, a substrate such as glass beads can be added to create a semihydroponic system (McLaughlin, Joller, et al. 2023).

The choice of the plant growth medium within the experimental setup depends on the biological question and should be defined in conjunction with the choice of an exudate collection method (see next sections). It has to be determined whether the presence of microbes is desirable or not to choose a sterile or nonsterile setup. Further, it should be determined whether the compound(s) of interest are detectable in a soil matrix background or not. If not, plants can either still be grown in the desired substrate and exudates are collected ex situ which results in tissue wounding, or the substrate can be adapted by, for example, lowering the percentage of soil, avoiding clay (binds exudates), and adding inert material such as sand or glass beads.

1.6 Experimental Blanks

For metabolomics studies, a central part of the experimental setup are control samples, as these allow to investigate the systems metabolite background. For metabolites extracted from tissues, these blanks consist of empty tubes ‘extracted’ with the same solvents and procedures as the tissue samples. The signal from these tubes comprises of contaminants from plastic and water, and of residues remaining on glass beads or containers from previous experiments. For hydroponic setups, experimental blanks consist of containers without plants set up analogous and in parallel to plant-containing containers. Their metabolite signal similarly reflects the background of the system. Such blank signals can be subtracted from biological samples during data analysis.

Preparation of experimental blanks for plants grown in a (soil) matrix is more challenging. Analogously, empty containers with, for example, soil are prepared and containers are flushed with collection medium as are the plant-containing containers. However, these experimental blanks are not true negative controls, as plants shape their microbiome and the surrounding soil, altering this signal in their containers. Preparation of these nonplant controls is still valuable, as it allows to assess the bulk soil metabolite signal. Bulk soil metabolite profiles can be compared to exudate metabolite profiles to understand which signals were increased by plant presence, and which ones deleted. To expand these datasets, an ex situ exudate collection step can be added (see ‘in situ or ex situ’ section).

Including empty containers without plants in the experimental setup is mandatory, and depending on the setup, yields different kinds of information. In a setup with low metabolite background, experimental blank signals can be subtracted from exudate samples to determine which signals stem from exudates. In a setup with a complex metabolite background such as soil, comparing ‘bulk soil’ blank samples against exudate samples yields information on exudation and depletion of compounds.

1.7 In situ or Ex Situ Collection

Exudates can be collected within the experimental setup in situ or ex situ by transferring plants from the growing matrix (hydroponic, agar, or soil matrix) into a container with collection medium (Table 1) (Jaitz et al. 2011). In our literature review, two-thirds of exudates were collected ex situ, one-third in situ (Figure 3).

In situ collection is advantageous due to minimal disturbance. In hydroponic and semihydroponic setups, exudates can be collected directly from the growth medium, or by exchanging the growth for a collection medium. Exudates from plants growing in a soil matrix can be collected in situ by flushing the system with collection medium, or by removing plants for ex situ collection. For the latter, a washing step is usually added. While ex situ collection considerably lowers metabolite background, such exudate profiles are biased due to the stress and damage imposed during transfer: Wheat and pea roots released multiple times higher amino acid levels when they were previously ‘damaged’ by swirling roots in a container containing sand compared to undamaged plants (Ayers and Thornton 1968). To reduce damage signal, it was proposed to let root systems recover for some time after transfer. For three grassland species grown in soil for 3 months, the initial exudate profiles after transfer to the hydroponic solution were similar to the metabolite profile of crushed roots, but profiles were similar to hydroponically-grown plants when recovering for 3 days in water (Williams et al. 2021).

When designing the experimental setup, the mode of exudate collection should be determined carefully. For hydroponic systems, in situ collection is desirable. For soil-based systems, a combination of in situ followed by ex situ collection typically renders most information.

1.8 Temporal Considerations

The timing and duration of exudate collection is crucial for several reasons. First, some metabolites have a diurnal signature (de Barros Dantas et al. 2023). Thus, the time of day for exudate collection can impact the exudation profile. Second, the duration of exudate collection can create a bias towards certain metabolites, as metabolite concentrations around roots are a balance between active and passive release and re-uptake (Jones and Darrah 1995). In our literature review, exudates were collected in a wide temporal range, from a few seconds up to 1 month (collection in growth medium without previous washing). In 10 studies, exudates were collected for less than 1 h, in 23 studies between 1 and 12 h, in 24 studies between 1 and 3 days, and in 10 studies for more than 3 days. On average, exudates were collected for 3 h (Figure 3).

Here, we discuss the duration of exudate collection: Presence of a compound outside of roots depends on its release via diffusion or passive/active transport, and on its re-uptake by the plant (usually by active transport). In a nonsterile system, metabolite uptake/degradation and exudation by other organisms further complicate the picture, and the relevance of metabolite re-uptake by plants is rather unclear.

Active release of plant exudates via transporters was shown for secondary metabolites such as strigolactones, phenolics, for protons, and for primary metabolites such as organic acids such as malate and citrate (Sasse, Martinoia, and Northen 2018). To date, only few transporters involved in exudation have been described, and the mode of exudation remains to be discovered for most compounds. Amino acid exudation for example might be driven by diffusion, by channels, or by ATP-driven transporters, but likely not by not proton-driven transport, as it cannot be blocked by specific drugs (Agorsor, Kagel, and Danna 2023).

Metabolite uptake systems are described for sugars and amino acids, both featuring proton-driven, active transport systems (Lee et al. 2007; Svennerstam, Ganeteg, and Näsholm 2008; Dündar and Bush 2009; Tegeder and Ward 2012). Wheat roots were shown to take up 6% of labelled lysine, glycine, and glutamate from soils, with the remaining label being detected in soil microbial biomass (Owen and Jones 2001). Blocking of one amino acid uptake system increased abundance of associated microbes and reduced plant growth (Agorsor, Kagel, and Danna 2023). Treatment with microbial products altered amino acid exudation and uptake levels (Phillips et al. 2004), highlighting the relevance of studying metabolite balance at the root-rhizosphere interface. Aside from amino acids, many other low-molecular weight compounds are taken up by plants (Jones and Darrah 1995; Phillips et al. 2004; Warren 2016; Sasse et al. 2020). Thus, when studying exudation, metabolite uptake by roots should be considered.

No matter the mode of release, an equilibrium of release and re-uptake is reached at different times for distinct compounds. For example, rates of carbohydrate and organic acid exudation for excavated rice were highest when collected in water for 2 h and dropped when collected for 4 or 6 h. The authors thus suggested to use a 2 h collection window to not underestimate exudation rates (Aulakh et al. 2001). A linear increase was discovered for amino acids exuded between few hours—1 day by Medicago, wheat and maize, whereas concentrations remained the same after 2 days, suggesting that these compounds reached an equilibrium after 1 day (Phillips et al. 2004). In general, different kinetics can be observed for metabolites when collected between 0.5 h and 3 days, without clear trends for specific chemical classes (McLaughlin, Zhalnina, et al. 2023).

From the data available, it can be concluded that exudation collections from a few hours to 1 day usually do not saturate metabolite signals. Incubation times of multiple days can increase the overall metabolite intensity further, but for many compounds, a balance between exudation and re-uptake is reached (Table 1). Thus, with access to sensitive instrumentation, it is advised to choose rather shorter than longer collection times. Metabolite signals can be further improved by, for example, concentrating collected exudates (see ‘Exudate processing’ section).

1.9 Collection Medium

The choice of a solvent in which to collect exudates in is a critical determinant in exudate collection, as the collection medium properties determine compound solubility. The selection should consider the osmotic balance between roots and solvent, the chemical properties of the compounds of interest, and potential enzymatic activity. In our literature review, two-thirds of the studies used water as collection medium, with a significantly smaller amount used nutrient or buffer (Figure 3).

Exudate collection in growth medium is straightforward regarding osmotic balance but comes with the disadvantage that high salt levels cause downstream issues in sample processing and analysis. Ammonium hydrogen carbonate or ammonium acetate-based solvents can be prepared equimolar to the growth medium and are compatible with many analytical workflows. The chemical properties of the compounds of interest should also be considered while collecting exudates. The ions present in the collection medium may lead to interactions/complexation between nutrients and exuded phytosiderophores. For instance, the exuded benzoxazinoid DIMBOA spontaneously chelates with iron, resulting in an equilibrium between free and complexed DIMBOA in the solvent (Hu et al. 2018). Because such complexes may not be detected with usual analytical methods, the interactions between exudates and nutrients represent a considerable bias in the exudate metabolome analysis. In addition, the presence of such macronutrients is interfering with several analytical techniques.

Tap water might be a suitable collection medium especially for matrix-based plant growth setups, given that water quality is well controlled (major solutes present, pH). Tap water reduces osmotic stress, limits interactions of exudates with nutrients, and is a good proxy to identify soluble deposits that would diffuse in the rhizosphere. Also, it mimics rainfall in a soil-based system. As tap water might contain microbes, sterilization is recommended before use. Further, the pH of the collection medium should be considered carefully: Soil water pH vary generally in a wide range between 3 and 9, whereas for hydroponic growth solutions, a typical choice of pH is 5.7–5.9. pH determines, for example, the ionization of organic acids, and using a solvent at a nonnatural pH may compromise the interpretation of the analysis.

Collected exudates may contain plant and microbial enzymes that convert the exuded compounds. Deactivating these enzymes may be crucial to distinguish between root-exuded compounds from their metabolization products. For example, collecting maize root exudate with water showed benzoxazinoid profiles dominated by DIMBOA (Cotton et al. 2019). However, when the obtained water extracts were immediately mixed with acidified methanol (final H₂O:MeOH:FA of 50:50:0.5, FA: formic acid), DIMBOA-Glc, but not DIMBOA, was detected as the major constituent (Hu et al. 2018). It is thus likely that maize roots release DIMBOA-Glc in the rhizosphere, where it is deglycosylated by either root or microbial enzymes. Adding a deactivating or precipitating solvent to the collected water may therefore be relevant when investigating actual root exudation rather than the rhizosphere metabolome.

The choice of collection medium should follow the general experimental design. For a hydroponic or semihydroponic setup, an equimolar solution compatible with downstream analytics is desirable. For soil matrix systems, tap water might be a good alternative, as it mimics rain, and osmotic balance is less affected in these systems due to the high amount of matrix metabolites. The pH of the collection medium should be controlled, and the use of enzyme deactivation in collected exudates should be considered. The collection medium impacts exudate processing and analytical considerations, as outlined below.

1.10 Exudate Processing, Storage and Normalization

Exudates are often processed and stored before analysis (Table 1). A first processing step usually involves removal of plant and microbial cells and other debris in the collection medium, either by filtering through a 0.45 or 0.22 µm filter, or by centrifuging at high speed to pellet cells and debris (Oburger et al. 2013; Zhalnina et al. 2018). In our literature review, three quarters of studies included a filtration step, 6% performed centrifugation, and 16% no debris removal step (Figure 3).

After, a concentration step may be added depending on the concentration of the compound of interest and the sample volume. Aqueous exudates are generally lyophilized (Zhalnina et al. 2018; Sasse et al. 2020; Lopez-Guerrero et al. 2022), whereas exudates in organic solvents are concentrated by evaporation under vacuum at room temperature to minimize metabolite breakdown (Ziegler et al. 2016). In our review, 40% of exudates were lyophilized, 41% not concentrated, and 13% evaporated under vacuum (Figure 3). Drying and concentrating exudates changes metabolite solubility. To assess this, standards with known concentrations should be treated with the same protocol as the samples.

Water or ammonium acetate as collection media do not interfere with concentration steps or analytical procedures and are thus a straightforward choice to circumvent issues with salts (McLaughlin, Joller, et al. 2023). When nutrient media with salts are concentrated, high salt levels can cause ion suppression depending on the analytical procedure chosen. Salts can be precipitated with organic solvents to lower the concentration before analysis (Sasse et al. 2020). Alternatively, salts can be removed by using an extraction cartridge binding the compound of interest (Sasse et al. 2016; Ziegler et al. 2016; Li et al. 2020; Lopez-Guerrero et al. 2022). If a column designed for retention of apolar compounds (e.g., C18) is chosen for analysis, salts pose less of an issue as they generally elute with the solvent front (Perez de Souza et al. 2021).

Exudates should generally be stored in an ultra-low freezer (−80°C) to avoid metabolite breakdown (Zhalnina et al. 2018; Sasse et al. 2020; Lopez-Guerrero et al. 2022). In collaboration with the analytical facility, the solvent in which the samples are injected into the analytical instrument is determined. Often, these instruments work best with an organic solvent such as methanol, but also other solvents such as acetonitrile:isopropanol:water 3:3:2 or methanol:water 1:1 might be acceptable (Sasse et al. 2016; Pétriacq et al. 2017; Zhalnina et al. 2018). To dissolve samples in these solvents, they need to be dried in a first step as outlined above. Dissolving pellets in the solvent of choice requires sonication in a water bath. Remaining insoluble debris must be removed by filtration or centrifugation to avoid clogging the injection needle of the analytical instrument (Pétriacq et al. 2017; Zhalnina et al. 2018).

Exudate volumes should be normalized to a root parameter of choice before determining metabolite profiles (Table 1). This is based on the hypothesis that larger root systems exude more compared to smaller systems, although this has not been shown systematically. Straightforward procedures include normalizing by root weight or total root surface (Sasse et al. 2020). Alternative normalization methods are used, for example, by normalizing to total organic carbon in exudates (Zhalnina et al. 2018).

In summary, removal of debris from exudates is generally advised, as is storage in an ultra-low freezer and exudate volume normalization according to root weight or a similar measure. Exudate concentration and choice of solvent for injection should be chosen with the analytical facility of choice and be determined in a pilot experiment setup, as outlined below.

2.1 Pilot Experiment Setup

A pilot experiment should be conducted to determine the exudate processing procedure, and the suitability of the analytical approach for the experimental question. The pilot should be set up with the analytical facility of choice and include: (i) blank and experimental samples to determine exudate signal above background, (ii) an experimental treatment to determine if changes in the exudate profiles can be detected, (iii) a minimum of four biological replicates (up to eight replicates) to assess the variability within the treatments. Processing steps such as exudate concentration, different collection solvents and collection times can be tested also. The exudate signal of samples should clearly be above the signal of the experimental blanks (generally, a few 100 features are significantly enriched in exudate samples (Zhalnina et al. 2018; Sasse et al. 2020; Lopez-Guerrero et al. 2022). Further, biological replicates of experimental treatments should cluster together and apart from a different treatment/genotype. This can be tested for example, with a principal component analysis.

2.2 Internal Standards (IS)

IS are compounds that are added to samples at equimolar levels to compare intensities across samples during analysis. They allow the identification of effects resulting in lower-than-average metabolite abundance, such as loss of sample during processing steps, or ion suppression during mass spectrometry measurements caused for example by differing salt levels in samples. If IS are added after exudate collection, they identify heterogeneity between samples during downstream processing steps. If IS are added right before analytical procedures, they allow detection of peak intensity decreases from the first to the last samples, and systemic differences between sample types.

From the variety of IS available, the IS chosen should be detected by the method of choice and not be present in the samples. For many compounds of interest such as such as phytohormones, sugars, amino acids, organic acids and isotopically labelled IS are available (Oburger et al. 2013; Gomez-Zepeda et al. 2021; Williams et al. 2021; Lopez-Guerrero et al. 2022; Seitz et al. 2022). Alternatively, IS of various chemical classes can be mixed, or plant metabolites can be labelled for example, by growing them in a ¹³CO₂ environment, followed by extraction (Strehmel et al. 2014; Ziegler et al. 2016; Lopez-Guerrero et al. 2022).

2.3 Description of the Analytical Method

Root exudates have been analysed with various analytical methods in the past decades (Dundek et al. 2011). Colorimetric assays determined the abundance of carbohydrates, amino acids or proteins (Dundek et al. 2011). Total organic carbon/nitrogen (TOC/TON) quantification is still frequently used (Keiluweit et al. 2015; Sun, Ataka, et al. 2021; Oburger et al. 2022). The implementation of MS for plant metabolite analysis allows detection of a vast array of compounds in low concentrations (Escolà Casas and Matamoros 2021). In our literature review, more than half of studies relied on LCMS and GCMS detection, with a smaller fraction employing TOC/TON only (Figure 3). Nine studies collected exudates for growth assays without performing metabolite analysis. These studies investigated growth of microbes, nematodes, or plants grown in previously collected (and concentrated) exudates. In studies comprising metabolite analysis, primary metabolites were identified in exudates of all species, with differences between species or conditions being rather quantitative than qualitative. About one-third of studies also investigated secondary metabolites mostly on a qualitative basis, which are more difficult to identify due to their vast diversity.

Excellent reviews are available discussing analytical approaches in detail (Escolà Casas and Matamoros 2021; Oburger et al. 2022). Here, we summarize briefly the factors important for experimental design. MS-based analyses can be targeted or untargeted. Targeted approaches use chemical standards to reliably identify compounds, also allowing quantification when measured in a dilution series in the appropriate sample background. If a specific compound of interest is present, a standard should be included in the study if possible. Untargeted analyses either report mass, retention time, and if available fragmentation spectra of features directly, or report putative compound identification by database queries. This approach is suitable to compare two experimental treatments in an unbiased manner. Generally, the abundance of unidentified peaks illustrates the vast diversity of still uncharacterized secondary metabolites of different plant species (van Dam and Bouwmeester 2016).

From an experimental design perspective, experimental blanks are imperative as mentioned (see ‘experimental blanks’), and the use of IS is advised (see ‘internal standards’). Finally, quality control (QC) samples including the compounds or chemical classes of interest should be included at the beginning and end of a run, as well as interspersed between samples to monitor stability of peak height and retention time throughout the experiment.