Radial Oxygen Loss from Plant Roots—Methods

In flooded soils, an efficient internal aeration system is essential for root growth and plant survival. Roots of many wetland species form barriers to restrict radial O2 loss (ROL) to the rhizosphere. The formation of such barriers greatly enhances longitudinal O2 diffusion from basal parts towards the root tip, and the barrier also impedes the entry of phytotoxic compounds produced in flooded soils into the root. Nevertheless, ROL from roots is an important source of O2 for rhizosphere oxygenation and the oxidation of toxic compounds. In this paper, we review the methodological aspects for the most widely used techniques for the qualitative visualization and quantitative determination of ROL from roots. Detailed methodological approaches, practical set-ups and examples of ROL from roots with or without barriers to ROL are included. This paper provides practical knowledge relevant to several disciplines, including plant–soil interactions, biogeochemistry and eco-physiological aspects of roots and soil biota.


Introduction
During flooding, the gas-filled pore spaces in the soil that normally facilitate O 2 diffusion are filled with water and the diffusion of gases (e.g., O 2 ) is highly impeded (diffusion coefficient of O 2 in water is 10,000 lower than in air; [1]). The O 2 dissolved in soil water or any entrapped air is rapidly consumed by aerobic microorganisms and roots, leading to soil anoxia or severe hypoxia [2,3]. Since O 2 is unavailable for radial entry into the root system in flooded conditions, longitudinal O 2 transport between shoot and root takes place and is essential for root respiration and plant growth [1,4]. The effectiveness of internal root aeration via longitudinal O 2 diffusion depends upon the amount of pore space resistance, the O 2 consumption by tissues along the diffusion path, the path length (root length) and the formation of tissue barriers to restrict ROL to the rhizosphere [5]. The formation of barriers to ROL in roots is paramount given its dual benefit of preventing ROL to the reduced rhizosphere (high soil O 2 demand can result in high O 2 loss from roots, [6]) and the entry into the root of phytotoxic concentrations of substances produced in flooded soils (e.g., Fe 2+ , [7]).
The presence of barriers to restrict ROL in roots was initially visualized in experiments indicating interspecific differences in root O 2 flux and marked basipetal reductions in root O 2 permeability in a number of wetland species [8,9]. Moreover, lower O 2 radial flux rates were obtained in roots of plants grown in flooded conditions in comparison to aerated controls [10]. For rice roots grown in aerated conditions, higher basipetal ROL rates decreasing towards the root tip are characteristic patterns indicating roots with 'weak' or In aerated nutrient solution, rice does not form an ROL barrier (with the exception of some wild accessions [20]), whereas the barrier is formed in stagnant, deoxygenated nutrient solution. Points indicate mean radial O2 loss at different distances behind the root tip and bands represent standard error (figure was constructed compiling published information on ROL from rice roots [11][12][13][14][15][16][17][18][19][20], n = 28). Plants were grown in aerated or stagnant, deoxygenated conditions for 5 to 31 days. Lengths of the roots were 93 to 159 mm.
In wetland plants, ROL from roots to the flooded soil occurs especially around the highly O2-permeable root tips without barriers [4,23,24]. Moreover, considerable O2 loss can originate from developing young roots and basal laterals [25,26], although the lateral roots of rice have recently been shown to also form a barrier to ROL when plants are grown in stagnant, deoxygenated solutions [27]. Given that newly formed and short roots lack barriers to ROL [13,18,28], the sediment oxygenation increases proportionally as the numbers of roots produced by plants in anaerobic soils increase [6]. Sediment oxygenation is important for the oxidation of chemically reduced compounds such as Fe 2+ , Mn 2+ , H2S, and CH4 [1,29,30]. Moreover, O2 release from roots into the flooded soils can influence the redoximorphic biogeochemistry by establishing a spatiotemporal microoxic zone of potentially high microbial activity [31,32]. The amount of O2 taken up by plant shoots and released from roots to flooded soils depends upon many above and belowground factors, including the numbers, types (e.g., adventitious, laterals) and lengths of roots, the magnitude and distribution of their pore-space and tissue respiratory demand, the degree and distribution of barriers to impede ROL, the numbers of the aerial shoots in capacity for O2 In aerated nutrient solution, rice does not form an ROL barrier (with the exception of some wild accessions [20]), whereas the barrier is formed in stagnant, deoxygenated nutrient solution. Points indicate mean radial O 2 loss at different distances behind the root tip and bands represent standard error (figure was constructed compiling published information on ROL from rice roots [11][12][13][14][15][16][17][18][19][20], n = 28). Plants were grown in aerated or stagnant, deoxygenated conditions for 5 to 31 days. Lengths of the roots were 93 to 159 mm.
In wetland plants, ROL from roots to the flooded soil occurs especially around the highly O 2 -permeable root tips without barriers [4,23,24]. Moreover, considerable O 2 loss can originate from developing young roots and basal laterals [25,26], although the lateral roots of rice have recently been shown to also form a barrier to ROL when plants are grown in stagnant, deoxygenated solutions [27]. Given that newly formed and short roots lack barriers to ROL [13,18,28], the sediment oxygenation increases proportionally as the numbers of roots produced by plants in anaerobic soils increase [6]. Sediment oxygenation is important for the oxidation of chemically reduced compounds such as Fe 2+ , Mn 2+ , H 2 S, and CH 4 [1,29,30]. Moreover, O 2 release from roots into the flooded soils can influence the redoximorphic biogeochemistry by establishing a spatiotemporal microoxic zone of potentially high microbial activity [31,32]. The amount of O 2 taken up by plant shoots and released from roots to flooded soils depends upon many above and belowground factors, including the numbers, types (e.g., adventitious, laterals) and lengths of roots, the magnitude and distribution of their pore-space and tissue respiratory demand, the degree and distribution of barriers to impede ROL, the numbers of the aerial shoots in capacity for O 2 uptake, the porosity and O 2 demand within these shoots, their lengths and the degree of submergence of the aerial shoots, the submerged soil O 2 demand, their microbial activity, their physical properties (i.e., O 2 diffusivity being lower in clay than sandy soils) and temperature [2,6].
Within the past few decades, active research has been conducted on the ecological importance of ROL to the flooded soil as well as the anatomical location, chemical characterization, interspecific differences and external triggers for barriers to impede ROL from roots [7,11,12,[15][16][17][33][34][35][36][37][38][39]. In this review, we summarize the main methodological techniques used for the visualization and quantification of root ROL. The different techniques described in here have been instrumental for our current knowledge of O 2 dynamics in submerged substrates. Understanding the strengths and weaknesses of the different methods for ROL quantification as well as the potential uses of each method will provide the needed information for an appropriate method selection and accurate quantification of ROL from roots of plants grown in diverse conditions.

Methods for Root Radial O 2 Loss Determination
Radial O 2 loss from roots can be qualitatively visualized and/or quantitatively determined using several contrasting approaches (Table 1). Here, we describe the most widely used methods for ROL measurements reported in literature. We focus on the principles, main uses and limitations as well as practical set-ups for each method. Detailed protocols and materials needed for each method are provided as Supplementary Information (Supplementary S1-S4). Other methods and approaches for determination of related aspects to ROL including sediment redox potential, iron plaques on roots and mathematical models are summarized in [4].

Qualitative Colorimetric Methods
The colorimetric methods for ROL determination are based on the principle of the oxidation-reduction of an external molecule that reacts with molecular O 2 . Changes in coloration can be visualized and be indicative of ROL from roots. Colorimetric methods for ROL determination include the indigo-carmine [40], methylene blue [25,37], Ti 3+citrate [41,42], the anthraquinone radical anion [43] and crystal violet [44].

Principle of the Method
The methylene blue is the most widely used method for the colorimetric determination of ROL from roots. The methylene blue dye is bright blue in the oxidized state and colorless in the reduced state. This dye has been used to study the sites of ROL from roots of plats grown in different conditions including stagnant, deoxygenated solutions [25,37] and flooded soils [33]. Once the methylene blue dye is reduced in a deoxygenated medium, the sites of root ROL can be detected as an intense blue color along these regions of the roots where O 2 loss occurs ( Figure 2). for ROL determination include the indigo-carmine [40], methylene blue [25,37], Ti 3+ -citrate [41,42], the anthraquinone radical anion [43] and crystal violet [44].

Principle of the Method
The methylene blue is the most widely used method for the colorimetric determination of ROL from roots. The methylene blue dye is bright blue in the oxidized state and colorless in the reduced state. This dye has been used to study the sites of ROL from roots of plats grown in different conditions including stagnant, deoxygenated solutions [25,37] and flooded soils [33]. Once the methylene blue dye is reduced in a deoxygenated medium, the sites of root ROL can be detected as an intense blue color along these regions of the roots where O2 loss occurs ( Figure 2).

Practical Set-Up
Root systems of intact plants are completely submerged in a container filled with deoxygenated agar solution and the "colourless" reduced methylene blue ( Figure 2A). The root-shoot junction is submerged a few centimeters (i.e., 2 cm) down into the reduced solution to avoid O 2 leaks in that zone while the shoots remain in air. The plant is fixed and ROL is visualized after c. 30 min by blue coloration in the rhizosphere or on the root surfaces (see protocol in Supplementary Information S1). Roots with a weak barrier, or no barrier at all, will display blue coloration longitudinally across the entire root ( Figure 2B). In contrast, roots with a strong barrier to ROL only show blue halos at the root tip where there is no barrier formation ( Figure 2C). The resolution of the technique is sufficient to indicate the sites of ROL produced by root 'windows' of developing laterals ( Figure 2D).

Generalities
The methylene blue method is appropriate for a rapid and initial screening of the main sites of root ROL or the visualization of a barrier to ROL in roots. This inexpensive method has been used to determine differences in ROL among genotypes from a large population (e.g., recombined inbred lines of maize, [37]) or for a rapid determination of ROL on different roots from the same genotype (e.g., rice, [18,44]). The main limitations of the methylene blue method are (1) its qualitative nature and (2) the unknown O 2 detection limit ( Table 1). This will restrict the method use to a simple characterization of the sites of ROL and/or the presence/absence of barriers to ROL instead of detailed physiological studies on root ROL barrier strength.

Principle of the Method
The polarography is an analytical technique based on the electro-oxidation or electroreduction of substances in solution under an applied electromotive force. This technique utilizes the characteristics of the current-voltage curve for O 2 identification and the current output to quantify its amounts. The polarographic O 2 determination in a deoxygenated medium requires an electrical circuit where a cathode, a cylindrical platinum electrode (syn. sleeving-electrode), is connected to an anode (e.g., reference electrode; usually Calomel or Ag/AgCl electrode) and these two to an amperometer ( Figure 3A). The sleeving-electrode is polarized in the deoxygenated solution before measuring ROL from a root (see protocol in Supplementary Information S2). Polarization voltage is applied to the circuit and a current plateau placed somewhere between −0.2 and −0.8 V indicates that the rate of O 2 reduction is independent of voltage and only depends on the rate of O 2 diffusion to the cylindrical electrode surface ( [9]; Figure 3B). The current plateau depends upon the amount of O 2 present in the solution. Under the appropriate polarizing voltage, the sleeving-electrode acts as a sink for O 2 , (the O 2 concentration at the electrode surface is effectively maintained at zero) and a diffusion gradient is maintained between root and electrode [1,45].

Practical Set-Up
The root-sleeving electrode and the root system of an intact plant are immersed in an acrylic chamber filled with stagnant, deoxygenated solution ( Figure 3A). Once the sleeving-electrode is polarized, a root of an intact plant is inserted through the cylindrical Pt electrode ( Figure 3C). The electrode can be positioned at different distances along the root and the diffusion current at a given position on the root is measured after the current has equilibrated (c. after 25 min, Figure 3D). After measuring the root, the residual current in the deoxygenated solution is obtained by placing the electrode away from the root and measuring the current at equilibrium. The rate of ROL from the root at equilibrium can be calculated using the equation proposed by Armstrong and Wright [45]: where ROL is the radial O2 loss (O2, nmol m −2 root surface s −1 ), is the diffusion current (μA) with the root in the electrode minus the residual current (μA) in the deoxygenated solution without a root and is the surface area of the part of the root within the elec-

Practical Set-Up
The root-sleeving electrode and the root system of an intact plant are immersed in an acrylic chamber filled with stagnant, deoxygenated solution ( Figure 3A). Once the sleeving-electrode is polarized, a root of an intact plant is inserted through the cylindrical Pt electrode ( Figure 3C). The electrode can be positioned at different distances along the root and the diffusion current at a given position on the root is measured after the current has equilibrated (c. after 25 min, Figure 3D). After measuring the root, the residual current in the deoxygenated solution is obtained by placing the electrode away from the root and measuring the current at equilibrium. The rate of ROL from the root at equilibrium can be calculated using the equation proposed by Armstrong and Wright [45]: where ROL is the radial O 2 loss (O 2 , nmol m −2 root surface s −1 ), I is the diffusion current (µA) with the root in the electrode minus the residual current (µA) in the deoxygenated solution without a root and A is the surface area of the part of the root within the electrode (cm 2 ). The aboveground parts of the plants remain above water and photosynthetically active radiation is provided to ensure stomatal opening and O 2 transport from shoot to roots. The ROL is usually measured at different positions along the root. ROL decreasing towards the root-shoot junction (the source of the O 2 ) indicates the presence of a barrier to ROL (Figure 1, cf. [12,24]). In contrast, higher basipetal ROL rates decreasing towards the root tip are characteristic patterns of roots with 'weak' or absent barriers to ROL (Figure 1).

Generalities
The polarographic determination of ROL is appropriate to accurately detect very small amounts of O 2 leaking from roots. Moreover, this technique can be used to detect values of high ROL rates (i.e., c. 21 kPa pO 2 ). However, incomplete reduction of O 2 at the electrode surface can occur when artificially perfusing root sections with high pO 2 (i.e., pO 2 > 38.7 kPa, [46]). Therefore, care must be taken when evaluating ROL at different O 2 concentrations surrounding the shoot. The main limitations for this method include the use of custom-built equipment and the long time required to obtain ROL rates ( The O 2 microelectrodes are amperometric sensors where an anode and a cathode are fused into one sensor [47], resembling a miniaturized version of the Clark-type sensors [48]. The O 2 quantification at the microelectrode tip relies upon O 2 diffusion from the medium to an O 2 -reducing cathode, which is polarized against an internal anode. An O 2 microelectrode contains a gold-coated sensing cathode, a silver or gold wire guard cathode (reducing cathode) and an Ag/AgCl anode, all inserted into a glass capillary filled with an electrolyte solution and sealed at the tip with a gas-permeable silicone membrane ( [47]; Figure 4A). The electrodes inside the glass capillary form a circuit where the resulting signal is in the pA (10-12 A) range. A picoammeter is used to polarize the circuit and to amplify the current output. The guard cathode removes O 2 in the electrolyte, minimizing zero current, response time and polarization time. When the zero current has stabilized following a period of polarization, the current in the measuring circuit depends only on the diffusional supply of O 2 from the medium to the surface of the microelectrode membrane [47]. The response to O 2 is linear all the way up to pure O 2 , i.e., 101 kPa pO 2 .

Principle of the Method: O 2 Micro-Optodes
The O 2 micro-optodes are basically a fiber glass coated with an oxygen-quenchable luminophore ( Figure 4B). The luminophore is excited when illuminated at a particular wavelength. Oxygen can act as a dynamic luminescence quencher, decreasing the luminescence quantum yield of the luminophore and therefore the O 2 concentration can be determined by the difference resulting from the amount of emitted light by the luminophore in presence and absence of O 2 [49,50]. The response to O 2 is non-linear but modern meters handle the non-linear response based on a 2-point calibration, typically at zero O 2 and with O 2 at air equilibrium. Oxygen optodes lose resolution with increasing O 2 partial pressure and most sensors should not be used for pO 2 exceeding 40 kPa.  . Arrow in C points to O2 microelectrode, arrowhead to temperature sensor and asterisk key to O2 micro-optode. Arrow in D shows the tip of an O2 microelectrode.

Principle of the Method: O2 Micro-Optodes
The O2 micro-optodes are basically a fiber glass coated with an oxygen-quenchable luminophore ( Figure 4B). The luminophore is excited when illuminated at a particular wavelength. Oxygen can act as a dynamic luminescence quencher, decreasing the luminescence quantum yield of the luminophore and therefore the O2 concentration can be determined by the difference resulting from the amount of emitted light by the luminophore in presence and absence of O2 [49,50]. The response to O2 is non-linear but modern meters handle the non-linear response based on a 2-point calibration, typically at zero O2 and with O2 at air equilibrium. Oxygen optodes lose resolution with increasing O2 partial pressure and most sensors should not be used for pO2 exceeding 40 kPa.

Practical Set-Up
Radial O2 loss from roots to the external medium can be quantified by inserting roots of intact plants (or detached root sections) into stagnant, deoxygenated solution. The O2 flux is measured at the root surface as well as a few μm (typically 200 μm) away from the root surface (deoxygenated medium) by either O2 microelectrodes or micro-optodes (see protocol in Supplementary Information S4). For this purpose, the microsensor (O2 microelectrode or micro-optode) needs to be fixed to a micro-manipulator fitted on a motorized stage to control micrometrical movement from the deoxygenated solution towards de root ( Figure 4C). The root is mounted horizontally in a fine woven wire mesh and held by

Practical Set-Up
Radial O 2 loss from roots to the external medium can be quantified by inserting roots of intact plants (or detached root sections) into stagnant, deoxygenated solution. The O 2 flux is measured at the root surface as well as a few µm (typically 200 µm) away from the root surface (deoxygenated medium) by either O 2 microelectrodes or microoptodes (see protocol in Supplementary Information S4). For this purpose, the microsensor (O 2 microelectrode or micro-optode) needs to be fixed to a micro-manipulator fitted on a motorized stage to control micrometrical movement from the deoxygenated solution towards de root ( Figure 4C). The root is mounted horizontally in a fine woven wire mesh and held by plastic bands attached to the mesh and mesh with the root mounted is submerged into the deoxygenated solution. The microsensor movement is controlled remotely and visualized by stereo microscope to guarantee that the microsensor tip reaches the root surface at the target point for ROL determination ( Figure 4D). Stirring near the root must be uniform in order to reproduce accurate determinations of O 2 fluxes from roots to medium (ROL) or from medium to roots (O 2 consumption by roots). Radial O 2 loss from roots can be calculated from the slope of the O 2 gradient from the root surface into the deoxygenated medium using the equation proposed by Henriksen [51]: (2) where D is the O 2 diffusion coefficient in the agar medium (approximates that in water at the experimental temperature), C1 is the O 2 concentration at the root surface (mol m −3 ) calculated from the partial pressure, C2 is the O 2 concentration at a given radial distance away from the root surface r2 (m) and r1 is the radius of the root (m). ROL measurements can be conducted along the root when the aboveground parts of the plant are above water (i.e., as explained before for ROL measurements using root-sleeving electrodes) or can be conducted on excised root segments sealed with lanoline in both cut ends [52]. Provided that the main source for O 2 is O 2 from root aerenchyma (deoxygenated solutions still have traceable amounts of O 2 ), the slope of the O 2 gradient from the deoxygenated medium into the root surface determine the presence of barriers to ROL. Oxygen gradients increasing towards the root surface indicate O 2 release and thus that no barrier to ROL is present, or if present, the barrier is weak ( Figure 4E). In contrast, a strong barrier to ROL is present when dissolved O 2 decreases towards the root surface ( Figure 4E) as no O 2 is being released by roots. The patterns of ROL obtained with O 2 microsensors might be similar to those obtained using the root-sleeving electrode (e.g., Figure 1; [28]), but in many cases the measured fluxes are lower when obtained by a microsensor. The root-sleeving electrodes reduce all O 2 at the Pt surface and therefore the O 2 gradient is steeper inside the ring, resulting in a higher flux.

Generalities
Oxygen microelectrodes and O 2 micro-optodes are very sensitive to low O 2 concentrations (nanomolar range), making them appropriate tools for ROL determinations. Moreover, the narrow tip diameters (3-10 µm for Clark-type microelectrodes and 50 µm for micro-optodes) allow for the spatial resolution of O 2 gradients to tissue and cell levels. The microelectrodes have been instrumental to obtaining O 2 profiles inside plant tissues, including the visualization of anoxic cores inside roots [53,54]. The higher spatial resolution of the O 2 microelectrode and O 2 micro-optodes, however, can also be a weakness for root ROL characterization as permeable cells (windows) and emerging laterals can cause O 2 leaks to the rhizosphere ( Figure 2D; [7,20,53,55]).
The O 2 microelectrodes with guard cathodes have almost undetectable zero currents of <1 pA and 90% response times in less than 200 ms [47,56]. O 2 microelectrodes have a faster response and smaller tip diameter than O 2 micro-optodes, but O 2 microelectrodes require calibration more often, especially when changes in temperature occur ( Table 1). The development of micro-optodes with a smaller diameter than 40 µm is limited due to a dramatic decrease in luminescence signal reducing O 2 sensitivity [49]. Due to its working principle, an O 2 microelectrode consumes an extremely low amount of O 2 (ca. 4 × 10 −4 nmol O 2 h −1 , [57]), causing a ca. 1-2% signal difference between the signal in stagnant media compared to turbulent media or media with high diffusivity such as gasses. In contrast, the O 2 micro-optode is insensitive to stirring of the media due to its non-oxygen consuming detecting principle. The stiffness of the glass capillary makes the microelectrode suitable to penetrate plant tissues, whereas the flexible nature of the optical fiber of the O 2 micro-optode makes it unsuitable for such measurements [50]. The signal of both microsensors can be affected by some other chemical compounds and strong organic solvents (Table 1).

Planar Optodes 2.4.1. Principle of the Method
The planar optode is a relatively recent tool for the 2D visualization and quantification of O 2 . The planar optode uses the same principle of an O 2 quenchable luminophore, as explained above for the O 2 micro-optode, but the signal intensity (O 2 -dependent) is photographed and analyzed, allowing for both the visualization and quantification of O 2 dynamics in 2D.

Practical Set-Up
The practical set-up for the planar optode method consists of a commercial regular RBG camera with an emission filter (530 nm) attached, an O 2 -sensitive luminophore (indicator dye), a light-harvesting dye (antenna), and a high-power blue LED light with an excitation filter (435 nm; [58] Figure 5A). The indicator and antenna dyes are dissolved in a permeable matrix and coated into a dust-free, transparent polyester foil to form the O 2 -sensing planar optode (see details on methodological aspects in Supplementary  Information S4). A black silicone layer is added to reduce light scattering and reflection from any background behind the planar optode [58]. In water-saturated systems, the planar optode is placed near the surface of the soil/sediment or next to roots inside a transparent-walled aquarium. The RGB camera is positioned perpendicular to the planar optode and LED lights are orientated at a 45 • angle to the camera ( Figure 5A; [58,59]).

Generalities
The majority of studies using planar optodes have a focus on benthic interactions in marine sediments [30,58,59,63,64]. This technique is also used for evaluating spatio-temporal variability in O2 dynamics in waterlogged crops (e.g., rice; [65]). Spatial resolution in planar optodes is largely determined by image format, pixel size, optical lenses and the camera's field of view [60]; using an appropriate set-up, maximum spatial resolutions can be comparable to that obtained with the smaller O2 microelctrode (e.g., <10 μm 2 ). Nonetheless, planar optodes can resolve O2 dynamics in larger areas of the root and soil interfaces as compared to O2 microsensors [66,67]. Response time of the planar optodes depends upon the thickness of the layers and the silicone coating [59]. Moreover, the optode response is nonlinear, with maximum sensitivity at low O2 concentrations for most of the indicators, but still sufficient accuracy from 50% to 100% of air equilibrium (up to 20.6 kPa pO2) [58]. The main constraint for using planar optodes is that O2 quantification and imaging are being conducted throughout and along a glass wall that interferes with the root radial O2 diffusion. Placing an O2-impermeable barrier along a root will enlarge the oxy- Oxygen distribution within the planar optode is evaluated by illuminating the planar optode with the specific LED light, and the luminescence emitted is recorded in a photograph. Measurements are conducted in dark conditions to avoid light interference ( Figure 5B). The main approaches for image analysis include luminescence intensity, ratiometric imaging and luminescence lifetime. Luminescence intensity measurements have some disadvantages as they are sensitive to variations in background reflection, inhomogeneous distribution of the luminophore and variations in the homogeneity of the excitation light [60]. Luminescence lifetime imaging offers some advantages, as this technique allows for excellent contrast enhancement and background suppression of unwanted luminescence in the image [61,62]. However, luminophores with relatively long lifetimes and complex hardware and software for triggering and synchronizing image acquisition and sensor excitation are required [60]. The ratiometric imaging approach is therefore a preferred technique, as this method is simple, inexpensive and reduces the noise of the image using differences in luminescence intensity between the antenna dye (does not react with O 2 ; therefore, luminescence intensity is unaffected) and the indicator dye (emission quenched in the presence of O 2 ) as an internal reference for the O 2 sensor [58]. Color patterns in the planar optode image indicate differences in O 2 concentrations at the interface between roots and anoxic sediment, highlighting longitudinal gradients of ROL ( Figure 5C). Radial O 2 loss can be quantified in selected regions of interest in the picture, based on O 2 extinction/depletion between root surface and the anoxic sediment using open-source software ( Figure 5D; see details in Supplementary Information S4).

Generalities
The majority of studies using planar optodes have a focus on benthic interactions in marine sediments [30,58,59,63,64]. This technique is also used for evaluating spatiotemporal variability in O 2 dynamics in waterlogged crops (e.g., rice; [65]). Spatial resolution in planar optodes is largely determined by image format, pixel size, optical lenses and the camera's field of view [60]; using an appropriate set-up, maximum spatial resolutions can be comparable to that obtained with the smaller O 2 microelctrode (e.g., <10 µm 2 ). Nonetheless, planar optodes can resolve O 2 dynamics in larger areas of the root and soil interfaces as compared to O 2 microsensors [66,67]. Response time of the planar optodes depends upon the thickness of the layers and the silicone coating [59]. Moreover, the optode response is nonlinear, with maximum sensitivity at low O 2 concentrations for most of the indicators, but still sufficient accuracy from 50% to 100% of air equilibrium (up to 20.6 kPa pO 2 ) [58]. The main constraint for using planar optodes is that O 2 quantification and imaging are being conducted throughout and along a glass wall that interferes with the root radial O 2 diffusion. Placing an O 2 -impermeable barrier along a root will enlarge the oxygenated area as compared to a situation without such a wall cf. [68]. Moreover, light scattering properties can be affected by several factors including the wall material, the planar optode itself, background light, a thin water film or air bubbles trapped between optode and wall and the silicone coating; potentially generating a skewed and smeared image of the O 2 distribution around the root [68,69]. These issues can be minimized or even eliminated by an appropriate selection of materials and calibration procedures [69]. A comprehensive list of existing O 2 indicators, including critical photophysical and sensing properties and the criteria for indicator selection, is summarized in Quaranta et al. [70]. Another important aspect for planar optode use is that good contact between the sensor foil and the root is needed to ensure optimal luminescence and O 2 quantification. Imaging in focal planes is ideal, as the further away from the actual roots the picture is taken, the lower the obtained O 2 concentration [71], but obtaining full contact of roots with the optode wall is often challenging as the root architecture and geometry varies. However, recent developments in optical O 2 sensor nanoparticles have allowed for a full visualization and quantification of O 2 at different focal planes of the whole rizhosphere level when using artificial transparent sediment [71].

Conclusions and Further Perspectives
Much remains to be learnt regarding radial O 2 loss from roots and the associated redox processes occurring in submerged substrates. The contribution of radial O 2 loss to important bio-geochemical processes, including methane production/consumption, the oxidation of metals, the solubilization of metals, pH dynamics, and the development of microoxic zones for microbial activity, as well as possible drawbacks of the presence of barriers to radial O 2 loss in agricultural crops, remains unclear. Different methods for qualitative and/or quantitative determinations of ROL from roots have been fundamental to our current understanding of this important biological process. Technical and practical knowledge, as well as the production of custom-made equipment, has been the main limitation for a broader use of these techniques for ROL determination in ecological studies.
However, recent research progress, method development and the commercially available equipment needed for ROL determinations have opened the possibility to increase our current understanding of this topic. The integration of different techniques and analyses appears to be an excellent option to unravel complex chemical processes in submerged substrates. For example, the combined use of O 2 planar optodes with diffusive gradients in thin films has been successfully used to demonstrate higher metal mobilization in a geochemical niche of higher O 2 concentration and low pH adjacent to root tips of rice [72]. Moreover, the recent development of hyperspectral luminescence imaging, together with signal deconvolution analysis, enabled the development of a multi-indicator, high-resolution approach with simultaneous imaging of multiple analytes [73]. The techniques described here have been proved to be instrumental for a wide range of uses, from large population screening studies to single cell O 2 determination. The combined use of the techniques described here would be key for understanding internal O 2 dynamics and oxidation processes in wetland ecosystems. Moreover, the integration of the techniques for ROL determination with other cutting-edge technologies for plant phenotyping would be essential to study how plants develop and respond to adverse environmental conditions.

Conflicts of Interest:
The authors declare no conflict of interest.