MeIQ‐Resorufin and 4NB‐Resorufin were synthesised as described previously (Wallabregue et al., ²⁰²³; Collins et al., ²⁰¹⁸). To yield MeIQ‐Resorufin, diisopropyl azodicarboxylate was added dropwise to a solution of 3‐(1‐hydroxyethyl)‐5‐methoxy‐1,2‐dimethyl‐1H‐indole‐4,7‐dione, resorufin and triphenylphosphine in anhydrous tetrahydrofuran at room temperature under argon. The reaction mixture was stirred vigorously for 24 h at 50°C, then concentrated under reduced pressure. The resulting residue was dissolved in ethyl acetate, washed with aqueous sodium hydrogen carbonate, water, aqueous hydrochloric acid (1 M solution) and brine, dried (sodium sulphate), filtered and concentrated under reduced pressure. The crude material was purified using silica gel flash column chromatography. The residue was dissolved in ethyl acetate and precipitated by adding hexane. The precipitation procedure was repeated twice. The precipitate was then dried under reduced pressure to give the title compound as an orange solid. The timescale for this synthesis is c. 3 wk.

To prepare 4NB‐Resorufin, 4‐nitrobenzylbromide in N,N‐dimethylformamide was added to a solution of resorufin and potassium carbonate, stirred at 40°C for 3 h and then cooled to room temperature. The solution was diluted with ethyl acetate and washed with brine. The aqueous layer was extracted with ethyl acetate. The organic components were combined, dried (sodium sulphate), filtered and concentrated under reduced pressure. The crude residue was recrystallised from dichloromethane, followed by trituration with diethyl ether to yield the title compound as a red solid. Further details are provided in the Methods S1. All the probes were dissolved in dimethyl sulfoxide (DMSO) to reach a stock concentration of 1 mM and stored at −20°C in black Eppendorf tubes to avoid light penetration. The timescale for this synthesis is c. 1–2 d.

Arabidopsis thaliana (L.) Heynh. Columbia (Col‐0) ecotype was used for seedlings. For in vitro propagation, sterilised seeds were cultivated on ½‐strength Murashige & Skoog medium (½MS; Murashige & Skoog, ¹⁹⁶²) supplemented with 1% (w/v) sucrose and 0.8% (w/v) agar, and grown vertically at 22°C, 16 h : 8 h, day : night photoperiod 100 μmol photons m⁻² s⁻¹. Two‐week‐old seedlings were used for the experiments. Arabidopsis thaliana Landsberg erecta ecotype was used for whole cell cultures; these were propagated in full‐strength MS medium, supplemented with 3% (w/v) glucose, 2.5 μg ml⁻¹ 1‐naphthaleneacetic acid (NAA) and 0.25 μg ml⁻¹ kinetin adjusted to pH 5.8, and renewed weekly. Experiments were performed with cells from cultures, which had been growing for 4 d following subculture.

For incubation with cell cultures, 1 ml of 4‐d‐old cell culture was incubated with a final concentration of 10 μM of probe for 2 h (based on studies with similar probes in mammalian cells (Wallabregue et al., ²⁰²³) as well as time‐course studies of fluorescence intensity in the presence of each probe; Fig. 3). One millilitre of cell cultures incubated with 1% DMSO and 1 ml of MS incubated with either 1% DMSO or 10 μM of probe were used as controls. For seedlings treatment, 2‐wk‐old A. thaliana seedlings were incubated with 1 ml of liquid MS medium supplemented with 100 μM of probe unless otherwise mentioned for 6 h to allow time for tissue entry and diffusion. Seedlings incubated with 10% DMSO, MS incubated with the same amount of probe or DMSO, were used as negative controls. All the experiments were performed using 12‐well plates, with three biological replicates per plate unless otherwise stated. Each biological replicate is represented by 1 ml of Arabidopsis cell suspension or one 2‐wk‐old seedling for the cell culture and seedling experiments, respectively.

All the hypoxic treatments were conducted using a CLARIOstar® plate reader (BMG LabTech, Ortenberg, Germany) supplemented with an atmospheric control unit (ACU) to regulate O₂ gas level. Probe sensitivity to hypoxia was evaluated at 5, 1, 0.5, 0.3 and 0.2% O₂, using the ACU, flushed with N₂ (g), for 2 h (cell culture treatment) or 6 h (seedlings), in the dark; the ACU was turned off for controls under normoxic conditions. Plates were shaken at 100 rpm every 30 s before fluorescence detection. For each cycle, 20 readings per well were taken and the mean was calculated. Resorufin was detected upon excitation at 550 nm and emission at 585 nm, using a dichroic filter of 567.5 nm. Two‐week‐old A. thaliana Col‐0 seedlings expressing a HRPE:nLUC construct (Akter et al., ²⁰²⁴) were used for in vivo luciferase detection upon treatment under normoxic or hypoxic conditions (1% O₂ v/v O₂/N₂), for 1 h in the dark, using the CLARIOstar® plate reader (BMG LabTech). Four biological replicates were employed in the experiment, and wild‐type genotype was used as a control. Luminescence was detected using the Nano‐Glo® Vivazine™ live cell substrate (Promega), following the manufacturer's instruction. Measurements were collected for 20 min, seven measurements per minute were taken, and the mean was calculated.

Four‐day‐old A. thaliana cell cultures and 2‐wk‐old seedlings were used for probe detection after hypoxic treatment (1% O₂ v/v O₂/N₂) for 2 h (cell culture) or 6 h (seedlings). For propidium iodide (PI) staining, cell cultures were incubated with 10 μg ml⁻¹ of PI solution for 5 min, then mounted on slides. Imaging was performed using either ZEISS LSM 780 (Micron, Department of Biochemistry, University of Oxford) or ZEISS LSM880 Airyscan microscope (Department of Biology, University of Oxford), equipped with a 25D7 objective lens, upon laser excitation at 561 nm and collection at 570–590 nm for Resorufin imaging, excitation at 488 nm and collection at 604–622 nm for PI, and excitation at 561 nm and collection at 650–750 nm for Chl. For lateral root primordia, A. thaliana seedlings were grown on ½MS supplemented with 0.5% sucrose for 7 d and then incubated with 100 μM MeIQ‐Resorufin for 2 h in 6‐well plates under normoxic (21% O₂) or hypoxic (1% O₂) conditions. After washing in distilled water for 10 min, seedlings were imaged using a Zeiss LSM800 Airyscan laser scanning confocal microscope (University of Utrecht) with 561‐nm excitation laser light and 565–616 nm emission detection. Fluorescence intensity was calculated using the ImageJ software (Schindelin et al., ²⁰¹²). Six measurements were taken for each image, divided between background and signal (three measurements per category). The signal intensity was quantified as the ratio of the mean signal measurement to the mean of the background measurements for each image. This process was repeated for three images per probe, under each experimental condition.

To measure plant tolerance to the probe, A. thaliana seedlings were grown vertically on squared plates for 7 d and then treated with 100 μM 4NB‐Resorufin, MeIQ‐Resorufin or resorufin, or 10% DMSO, as a negative control, under air and 16 h : 8 h photoperiod conditions. After 4 d of treatment, plates were scanned using an EPSON Perfection V750 PRO scanner with a resolution of 720 dots per inch. Primary root length was assessed using ImageJ (Schindelin et al., ²⁰¹²). Twenty‐one to twenty‐eight biological replicates were used for root length measurements. Growth rate per day was assessed as the increase in the length of the primary root after 4 d of treatment compared with the beginning of treatment. Fresh weight was measured after 4 d of treatment using four biological replicates, each consisting of seven whole seedlings.

Fluorescence spectra were recorded using a HORIBA Jobin Yvon FluoroLog3 fluorimeter (Hamamatsu R928 detector and a double‐grating emission monochromator). The standard conditions for acquiring emission and excitation spectra are room temperature and steady‐state mode unless otherwise stated. Fluorescence spectra were obtained by using the GraphPad Prism software.

Ultraviolet‐visible (UV‐Vis) spectra were recorded on a V‐770 UV–Visible/NIR Spectrophotometer equipped with a Peltier temperature controller and stirrer using disposable polystyrene cuvettes of 1 cm path length. Experiments were conducted at 25°C unless otherwise stated. Ultraviolet‐visible spectra were plotted using the GraphPad software. pH measurements of solutions were taken using an Oakton pH meter.

To a vial containing DMSO, a probe solution was added. An aliquot was taken from the reaction solution and transferred to a quartz cuvette, followed by the addition of Milli‐Q water. The solution was homogenised by pipetting for 10 s; then, the fluorescence spectra were recorded using the specified parameters, serving as the initial time point (T = 0, where T represents time).

The vial was then sealed and placed in an aluminium box, positioned 20 cm away from a white LED and illuminated for 4 h. After illumination, another aliquot was collected, transferred to a quartz cuvette and diluted with Milli‐Q water. The fluorescence spectrum was subsequently recorded under the specified parameters. Further details are provided in the Methods S1.

To a quartz cuvette containing buffer with the desired pH or NaCl concentration, a probe solution was added. The solution was homogenised by pipetting for 10 s, and then, the cuvette was sealed and left at room temperature for 10 min (NaCl), and 5 or 60 min (pH buffers). The fluorescence spectra were then recorded using the specified parameters. Further details are provided in the Methods S1.

Statistical analyses were performed using the GraphPad Prism 10.3.0 or R Statistical Software (v.4.1.3; Foundation for Statistical Computing, Vienna, Austria). Normal distribution of data was evaluated through the Shapiro–Wilk test. Based on the outcome of the normality test, the appropriate tests were performed to evaluate differences. Additional details are provided in the legend of the corresponding figure.

We first set out to determine whether the 4NB‐ and MeIQ‐Resorufin probes could permeate A. thaliana cells. Four‐day‐old whole autotrophic suspension‐cultured cells were incubated with either 10 μM probe or 1% DMSO loading control before shaking in the dark for 2 h under normoxia (21% O₂) or hypoxia (1% O₂). Cells were visualised using a confocal microscopy to detect fluorescence arising from resorufin (λ_ex/em 561/570–590 nm), Chl (λ_ex/em 561/650–750 nm) and PI (λ_ex/em 488/604–622 nm). Propidium iodide staining was exploited for live/dead cell imaging, since this dye is not cell‐permeable in intact membranes (Riccardi & Nicoletti, 2006).

The arising images (Fig. 2a, Supporting Information Fig. S1) and signal quantification (Fig. 2b) show that when cells were incubated with either 4NB‐Resorufin or MeIQ‐Resorufin in hypoxic conditions, a significant increase in fluorescence in the 570–590 nm range was observed in cytosolic regions of the cell, consistent with resorufin fluorescence (Fig. 2b,c). No fluorescence was observed at this wavelength range when cells were incubated with either 4NB‐Resorufin or MeIQ‐Resorufin in normoxia (Fig. 2a). Similarly, no fluorescence was observed if cells were incubated in hypoxia and treated with 1% DMSO only, nor if either probe was incubated in MS media, in hypoxia and in the absence of cells (Fig. S2). Chl autofluorescence was also observed in the presence of both 4NB‐Resorufin and MeIQ‐Resorufin and when cells had been incubated in both normoxia and hypoxia. PI fluorescence was used to demonstrate that the probes were entering cells and that the cells remained viable. Propidium iodide fluorescence emission was only observed at cell surfaces, indicating the probes were not toxic to the cell under the conditions tested. Propidium iodide staining was then used to determine the proportion of dead or damaged cells as a function of increasing concentrations of each probe (Fig. S3a,b). Across the different concentrations tested, there was no significant increase in PI staining with increasing concentration of probe, relative to DMSO control. There was some increase in PI staining in hypoxic control cells compared with normoxic control cells, although this was not significant. Overall, the results indicate that, under the conditions used, the presence of the probes does not cause significant damage to the cell membranes and is no more damaging to cells than hypoxia itself.

The results therefore indicate that both probes are able to permeate Arabidopsis whole cells and enter the cytoplasm. The observed fluorescence in hypoxia but not in normoxia indicates that the probes have been bioreductively cleaved to form resorufin in a manner that is dependent on a sustained decrease in oxygen availability. This is consistent with the bioreductive properties of these probes in bacterial and mammalian systems (Collins et al., ²⁰¹⁸; Wallabregue et al., ²⁰²³). Although the reductase enzyme(s) responsible for bioreduction in the probes have not been identified in plants, the same probes are reduced by nitroreductase and CYP450 enzymes in mammalian and bacterial cells (O'Connor et al., ²⁰¹⁷; Collins et al., ²⁰¹⁸; Wallabregue et al., ²⁰²³), suggesting that similar enzymes may be enabling probe activation in plant cells. These probes are therefore able to detect whether whole plant cells grown in culture are hypoxic.

We next wanted to determine whether the probes could be exploited to detect cellular hypoxia in a higher throughput plate‐based format by measuring fluorescence intensity, and whether such an assay can be used to quantify fluorescence levels as a function of O₂ concentration. By exposing cells to different concentrations of O₂, we also wanted to determine whether bioreduction and probe fluorescence were the same for both probes. One millilitre of autotrophic A. thaliana 4‐d‐old cell cultures was transferred to 12‐well plates and incubated for 2 h in a plate reader where the atmospheric environment was controlled at concentrations of O₂ ranging from 0.2% to normoxia (21% O₂). Plates were kept in dark conditions throughout the experiment to avoid O₂ production through photosynthesis; therefore, glucose was added to the culture medium to maintain cells in a heterotrophic state. Fluorescence intensity was measured over time for each condition (Fig. 3a,b) and average intensity (from 15 to 120 min) was used to determine overall fluorescence levels (Fig. 3c,d) at each O₂ concentration.

As detected using microscopy, resorufin fluorescence was observed when cells were equilibrated with both probes in hypoxia compared with cells exposed to normoxia. Cells incubated with 4NB‐Resorufin at O₂ concentrations of 1% and below showed a sustained increase in fluorescence over the course of 2 h; when the O₂ concentration was 0.2%, a maximum and sustained level of fluorescence was observed after c. 15 min (Fig. 3a). There was an overall increase in fluorescence intensity as O₂ concentration reduced from 1% to 0.2% (Fig. 3c). When cells were incubated with MeIQ‐Resorufin, a more rapid increase in fluorescence was observed at all O₂ concentrations measured except 21% (Fig. 3b). The absolute levels of fluorescence attained appeared to be dependent on O₂ concentration, with the highest levels of fluorescence observed for 1% and 0.5% O₂ (Fig. 3d). This fluorescence was not sustained, however, and over the course of 2 h, fluorescence levels decreased, most notably in the presence of 0.2% O₂.

MeIQ‐Resorufin‐derived fluorescence appeared to be detected at higher O₂ concentrations compared with fluorescence observed in the presence of 4NB‐Resorufin at equivalent O₂ levels (Fig. 3b,d), consistent with the comparison between these probes in mammalian cells (Wallabregue et al., ²⁰²³), suggesting that MeIQ‐Resorufin may be capable of detecting milder hypoxia than 4NB‐Resorufin. However, absolute levels of fluorescence were approximately fivefold higher in the presence of MeIQ‐Resorufin than of 4NB‐Resorufin, which may reflect a greater uptake of this probe into the cell, a faster intracellular diffusion or a greater propensity for MeIQ‐Resorufin bioreduction in the A. thaliana cell environment. In line with this, the observed decline in fluorescence signal with time for MeIQ‐Resorufin suggests that it is sequestered more rapidly: Resorufin fluorescence is quenched by acidic conditions (pH 6.5 and below; Fig. S4), suggesting that the decline in fluorescence could be a result of vacuolar sequestration of high levels of resorufin derived from the MeIQ‐conjugated probe. Alternatively, resorufin fluorescence quenching could derive from reduced intracellular pH associated with lactate fermentation under hypoxia (Wagner et al., ²⁰¹⁹). Overall, the results indicate that the probes are capable of detecting intracellular hypoxia in a semi‐quantitative manner in a plate‐based assay, thus extending their utility as hypoxia‐sensing probes.

To confirm the biological relevance of 4NB‐ and MeIQ‐Resorufin as hypoxia probes, we next sought to determine whether cell permeability and bioreduction of the probes could be observed in A. thaliana whole seedlings under hypoxic conditions. We anticipated that additional incubation time would be required for diffusion and uptake of the probe in seedlings compared with that in cells; given we also observed a decrease in resorufin fluorescence in whole cells after prolonged incubation with MeIQ‐Resorufin at low O₂ concentrations, we first determined whether a stable fluorescence signal could be observed in seedlings incubated with 4NB‐ and MeIQ‐Resorufin at severe hypoxia (0.1% O₂) over the course of 6 h. The fluorescent signal intensity was not strong in cultured cells (Fig. 2); therefore, we increased the probe concentration in these experiments to 100 μM to ensure adequate signal intensity in a system with a more complex tissue structure (including potential barriers to uptake). The arising data (Fig. S5) revealed a time‐dependent increase in resorufin fluorescence intensity over 6 h, suggesting that the probes were sufficiently stable to elicit a detectable signal in the whole‐seedling environment.

We therefore proceeded to examine seedling tissues exposed to normoxia (21% O₂) or hypoxia (1% O₂) by confocal microscopy following incubation with both probes (Fig. 4). When seedlings were incubated with 4NB‐Resorufin (Fig. 4(I)), hypoxia‐dependent fluorescence was observed in root tissue (Fig. 4(I)a,b) and leaves (Fig. 4(I)c,d). In the root tissue, fluorescence was observed in the cortex and vascular tissue; in the leaf tissue, fluorescence was observed, albeit not homogenously. Overall, these images indicate that 4NB‐Resorufin was able to permeate the seedling tissue and translocate through the plant, likely via the vasculature. MeIQ‐Resorufin was similarly able to permeate seedling tissue and fluoresce in a hypoxia‐dependent manner. MeIQ‐Resorufin‐derived fluorescence was observed in vascular regions of the root and near‐homogenously in the leaf tissue (Fig. 4(II)a–d). These data suggest translocation of the probe throughout the plant tissue, including to the leaf mesophyll layer. As observed in cells, a significant increase in fluorescence was only observed in the presence of either probe (and hypoxia; Fig. 4e,f); seedlings treated with equivalent volumes of DMSO did not demonstrate fluorescence at 570–590 nm in either hypoxia or normoxia, while Chl autofluorescence was observed in these seedlings in leaf tissue (Fig. S6).

To confirm that seedling intracellular O₂ concentrations were hypoxic under the experimental hypoxia settings, we employed an A. thaliana line containing a luciferase reporter cassette under the control of the HRPE. This HRPE is known to promote ERF‐VII‐mediated gene upregulation under hypoxic conditions (Gasch et al., ²⁰¹⁶; Akter et al., ²⁰²⁴). When seedlings were exposed to normoxia or hypoxia under the experimental conditions used for probe incubation, we detected a significant difference in luciferase activity, commensurate with elevated luciferase expression in hypoxia (Fig. S7), confirming that the seedlings were experiencing a hypoxic environment.

We next checked whether incubation of seedlings with the probes could affect plant growth. Seven‐day‐old Arabidopsis seedlings were treated with 100 μM 4NB‐Resorufin, MeIQ‐Resorufin or resorufin (Fig. S8). No treatment, or treatment with 10% DMSO, was used as negative controls. Physiological measurements were taken after 4 d of the treatment, during which seedlings were kept under 16 h : 8 h photoperiod conditions, in air. Consistent with our results for cell culture, we did not observe any apparent phenotypical differences between seedlings treated with the probes compared with the DMSO negative controls (Fig. S8a). No statistically significant differences were measured in primary root length, growth rate and fresh weight compared with DMSO‐treated seedlings, albeit the comparison with untreated seedlings indicates a mild negative effect due to DMSO (Fig. S8b–d). These results suggest, therefore, that 4NB‐Resorufin, MeIQ‐Resorufin and their reduced form do not affect biomass or root growth in Arabidopsis, in the conditions used in this study. Collectively, these results demonstrate that the probes are capable of acting as fluorescent markers of hypoxia in live plant tissue as well as in cells.

Finally, we sought to determine whether these probes could be used to detect physiologically relevant hypoxia in an otherwise normoxic environment. Arabidopsis root lateral primordia represent a physiological hypoxic niche, arising due to rapid cell division and high metabolism, and demonstrated by localised ERF‐VII stabilisation and hypoxic‐response gene upregulation (Shukla et al., ²⁰¹⁹). We tested whether MeIQ‐Resorufin could detect hypoxia in root lateral primordia by incubating 7‐d‐old Arabidopsis seedlings with 100 μM probe for 2 h under either normoxic or hypoxic condition. We selected to use MeIQ‐Resorufin for this experiment given the higher levels of fluorescence observed for this probe in the whole cell cultures. Under normoxic conditions, MeIQ‐Resorufin‐derived fluorescence was not typically observed in primary root tissue, as expected (Fig. 4), while fluorescence was consistently observed in root lateral primordia regions at all stages of development (Figs 5, S9). When these tissues were incubated in hypoxia, fluorescence was usually observed in all regions of the root, including the lateral primordia (Figs 5, S9), consistent with previous experiments. These results confirm that the MeIQ‐Resorufin probe can be used to visualise regions of hypoxia in otherwise normoxic Arabidopsis tissues.

Data acquired in this study are available in the Supporting Information of this article (Figs S1–S14; Table S1; Methods S1). The time required to synthetically prepare each probe is given in the Methodology section. Requests for resources (including 4NB‐Resorufin and MeIQ‐Resorufin) can be made to the corresponding author.