By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Biological samples are far from homogeneous, with complex compartmentation being the norm. Major physiological processes such as respiration do not therefore occur in a uniform manner within most tissues, and it is currently not possible to image its gradients in living plant tissues.
A compact fluorescence ratiometric-based device is presented here, consisting of an oxygen-sensitive foil and a USB (universal serial bus) microscope. The sensor foil is placed on the sample surface and, based on the localized change in fluorescence signal over time, information about the oxygen consumption (respiration) or evolution (photosynthesis) can be obtained.
Using this imaging technique, it was possible to demonstrate the spatial pattern of oxygen production and consumption at a c. 20-μm level of resolution, and their visualization in the rhizosphere, stem and leaf, and within the developing seed. The oxygen mapping highlighted the vascular tissues as the major stem sink for oxygen. In the leaf, the level of spatial resolution was sufficient to visualize the gas exchange in individual stomata.
We conclude that the novel sensor set-up can visualize gradients in oxygen-consuming and producing processes, thereby facilitating the study of the spatial dynamics of respiration and photosynthesis in heterogeneous plant tissues.
Aerobic metabolism by definition requires the presence of oxygen. In the mitochondria, oxygen serves as the terminal electron acceptor for oxidative phosphorylation, a fundamental component of respiration. In plants, oxygen is a by-product of photosynthesis. As both respiration and photosynthesis represent the basis of life on earth, an understanding of the mechanisms directing oxygen consumption, production and homoeostasis has long been a primary research goal in biology and biotechnology (Volkmer et al., 2008). An essential prerequisite for making progress in this field is the means to precisely quantify local concentrations of oxygen at the tissue and cellular levels. Conventional measurement systems such as polarography can only provide mean values which lack spatial resolution and are thus blind to any compartmentation present within the sample. This limitation has been partially overcome by the use of needle-type microsensors (Revsbech et al., 1986; Armstrong et al., 1994; Klimant et al., 1995), which have been applied to quantify oxygen concentration across a sample transect, delivering a resolution level of a few μm (Kühl, 2005; Borisjuk & Rolletschek, 2009; Rolletschek et al., 2009). However, as these devices can only measure the oxygen concentration prevailing at a discrete point at any given time, they cannot image the two-dimensional distribution of oxygen within the sample.
The development of methods for real-time imaging of physiological processes in plants facilitates progress in our understanding of plant development and have become increasingly important for experimental biologists (Reddy et al., 2007; Borisjuk et al., 2012). In particular, the targeting of specific molecules by exploiting fluorescent optical sensors has allowed the monitoring of a range of metabolic processes occurring in biologically active materials. Their particular advantages over other methodologies include their spatial coverage (ranging from mm2 to cm2), their μm level of resolution and the extended period of measurement (lasting from a few seconds to several days) that they allow. In addition, fluorescent chemical sensors use a potential-free sensing process with no analyte consumption. To date, various sensor approaches for a broad range of analytes have been reported whereby one of the most attractive analytes is oxygen (skin surfaces: Stücker et al., 1998; sediments: Holst & Grunwald, 2001; Werner et al., 2006; microbial mats: Glud et al., 1999; Kühl & Polerecky, 2008). Here, we present a means of assessing the distribution of oxygen in plant material, based on the elaboration of a transparent sensor foil. The system has been tested to image oxygen distribution across a number of plant organs of various species, and its output represents a substantial improvement in achievable spatial resolution, image processing and handling. By visualizing oxygen distribution across the sample and its changes over time, it allows inferences to be made regarding localized oxygen consumption and release (respiration and photosynthesis, respectively) at a microscopic level of resolution. This feature is novel and we anticipate that the widely applicable imaging tool will be very beneficial in furthering our understanding of plant metabolism.
Materials and Methods
Pea (Pisum sativum L. cv Erbi) plants were grown in a growth chamber under a 16-h photoperiod, with a lit temperature of 19°C and an unlit temperature of 16°C, and seeds were harvested 25 d after pollination (DAP). For the analysis of oilseed rape (Brassica napus L. cv Reston) seedling roots, seedlings were grown on 0.7% phyto-agar for 14 d under a 16-h photoperiod at a constant temperature of 22°C. Barley (Hordeum vulgare L. cv Barke), potato (Solanum tuberosum L. cv Desiree) and maize (Zea mays L. cv PH 207) plants were grown in a glasshouse under a 16-h photoperiod. Barley caryopses were harvested at 14 DAP. Leaves of a sycamore tree (Platanus occidentalis L.) and of a maple tree (Acer platanoides L.) infected with powdery mildew (Uncinula tulasnei, Fuckel) were collected from a specimen growing at the institute, and those of hibiscus (Hibiscus rosa-sinensis L.) and wandering Jew (Tradescantia fluminensis Velloso) were harvested from indoor-grown potted plants. Escherichia coli (strain DH5α) was grown on LB agar at a temperature of 37°C.
Planar oxygen sensor
The sensor consisted of an oxygen-sensitive foil (SF-RPSU4; PreSens GmbH, Regensburg, Germany) and a detection unit (USB (universal serial bus) microscope ‘VisiSens’; PreSens GmbH; for technical details see Tschiersch et al., 2011). The VisiSens software (PreSens GmbH) provided control of all camera settings, but the camera needed to be pre-calibrated with respect to colour channel sensitivity. The 24-bit, 1280 × 1024 (1.3-megapixel) images were stored as colour PNG (portable network graphics) files. The working distance was c. 2 cm and the 15 mm × 12 mm field of view resulted in a virtual pixel resolution of c. 12 μm. The actual spatial resolution, governed by the sensor and its lateral oxygen diffusion within the sensitive layer, is c. 20 μm. The software managed the image acquisition process and computed oxygen maps from the raw sensor response signal (in combination with two point calibration; see the following paragraph). The optical properties of the light source, and the excitation and emission filters were customized to the fluorescence response signals of the oxygen sensor foil. Please note that the standard response time of the sensor foil is in the range of 20 s. As a precaution against microbial artefacts, the sensor can be sterilized with 70% ethanol or ethylene oxide.
Working principle and calibration of the sensor
The device is based on the well-established fluorescence quenching technique (Holst & Grunwald, 2001; Meier et al., 2011; Schreml et al., 2011), and operates via a wavelength ratiometric calibration scheme, in which the sensor comprises an indicator dye and a reference dye both immobilized in a polymer matrix to form a ‘sensitive’ layer. The interaction between each indicator dye molecule and oxygen resulted in the quenching of the dye's red fluorescence, which entailed the energy of the excited dye being transferred to the oxygen molecule by collision. As a consequence, the luminescence signal of the dye was reduced as the sample's oxygen content increased, while the reference dye was unaffected by oxygen and produced a consistent level of signal. When used for measurement, the sensitive dye and reference dye were simultaneously excited with an identical light source (using LEDs incorporated into the USB microscope), but their emission spectra differed (while both dyes were excited with blue light, the sensitive dye emitted red light and the reference dye green light). As the emission wavelengths matched the red and green channel sensitivity of a colour RGB chip, it was possible to capture the reference and the sensor signals within a single image. Quantification was obtained by ratio-ing the red and green channel of the RGB image in order to reference out the main interferences of intensity-based measurements, namely, inhomogeneous light field and dye concentration, including varying sensor layer thickness. The relevant oxygen concentrations were computed from this ratio by applying a calibration function derived from the sensor output obtained from exposure to a known oxygen concentration. The Stern–Vollmer plot (incorporated into the software) leads to a linear relation which was used for calculating the respective oxygen concentrations of the measurement images (Tschiersch et al., 2011). At the start and end of the experiments the sensor foil was calibrated by obtaining calibration images both from a sodium sulfite solution (0% air saturation) and from air-saturated distilled water (100% air saturation). To this purpose, a drop of the calibration solution was placed on a glass plate and immediately covered with the sensor foil. Changes in the luminescence of the indicator dye were recorded for at least 10 min until stable luminescence values were reached. Generally these values represent a mean value of an area of 0.5–1.0 cm2. The obtained calibration values were corrected for the backscattering signal of the particular sample or directly used by the VisiSens software to create images of oxygen distribution from the recorded luminescence signal. In some cases, perfluorodecalin (PFD; Sigma-Aldrich, St. Louis, USA) was used instead of water.
Chlorophyll fluorescence imaging
Estimates of key fluorescence parameters were obtained using an IMAGING-PAM chlorophyll fluorometer (Heinz Walz GmbH, Effelrich, Germany). The imaged leaf area was 7 mm × 9 mm. Leaf segments were dark-incubated for 20 min followed by a 260-s exposure to blue light (peak 470 nm) at an irradiance of 315 μmol quanta m−2 s−1 both to monitor changes in chlorophyll fluorescence and to estimate electron transport activity during the induction of photosynthesis. Measurements of minimum fluorescence (F0) in the dark and of steady-state chlorophyll fluorescence (Ft) under actinic illumination were acquired and the peak fluorescence from dark- (Fm) and light-adapted () leaf segments was measured during a 800-ms exposure to saturating light intensity. Estimates of the effective quantum yield of photosystem II (PSII) (ΦII), the maximum quantum yield of PSII in the dark (Fv/Fm), and the photosynthetic electron transport rate (ETR) were obtained from digitized images of Ft (or F0) and Fm (or ) via a pixel-by-pixel calculation (Baker et al., 2007). For clarity, fluorescence parameter images were displayed in false colour.
Image capture and microscopy
Two- and three-dimensional images were captured using, respectively, a digital microscope VK-X100 (Keyence, Osaka, Japan) and a three-dimensional laser scanning microscope VHX-100K (Keyence, Osaka, Japan).
Sensor design and experimental set-up
The sensor described here is flexible and may be bent or cut into any desired shape using scissors. It consists of two layers, a transparent polyester support and the sensitive layer (Fig. 1). The former (125 μm thick) acts as a transparent, flexible and oxygen-blocking base layer. Although polyester is not completely impermeable to oxygen, the diffusion of oxygen through the polyester support is very slow and therefore negligible for the experiments presented here. The advantage of using polyester (instead of oxygen-impermeable glass support) is its flexibility. The transparent polyester layer (glossy side of the sensor) is turned towards the camera (USB microscope) during measurement. The sensitive layer incorporates the fluorescent dyes and responds to oxygen. The dyes are immobilized within a polymer matrix which is highly permeable to oxygen. In this way oxygen can penetrate into the sensitive layer and interact with the fluorescent indicator dye while the dyes themselves cannot leach out. The sensitive layer is dull, slightly red-coloured, c. 6–8 μm thick and strongly adhesive to the polyester support. This side must be in direct contact with the sample, because the sensor must equilibrate.
The general experimental set-up is shown in Fig. 1. In measurement mode, the sensor foil is carefully laid on the sample surface, so that the sensitive layer is in contact with the sample. At intervals of at least 1 min the indicator dye is excited by illuminating the sample for 10 s with weak blue light (using LEDs incorporated into the USB microscope), and the emitted signal is measured from the reverse side via the USB microscope. The time required to reach a near steady-state of oxygen distribution depends on the nature of the sample and measurement conditions (5–90 min). The transparency of the sensor foil allows the continuous illumination of the sample, sufficient to drive photosynthesis.
Notably, the imaging technique relies on there being a restricted liquid layer between the sensor foil and the surface. This allows the oxygen-consuming and -producing reactions of the sample to deplete or enhance the concentration of oxygen within this unstirred liquid layer and the unstirred nature of the layer allows spatial resolution. The time-dependent changes in oxygen concentrations represent the localized change in oxygen in the liquid adjacent to the foil and are very much dependent on the balance between consumption (evolution) by the tissue and diffusion from the surrounding media/air. Thus, the oxygen concentration measured under these conditions in the respective plant tissue essentially represents the localized consumption (evolution) of oxygen but does not represent the steady-state oxygen concentration operating during regular conditions (when the tissue is not covered with the sensor foil).
A restricted liquid layer between the sensor foil and the sample is easily created by adding a few μl of water to the surface of the sample before applying the foil. Instead of distilled water, any kind of buffer can be used, supplemented with any kind of (respiratory) substrate or inhibitor. In addition, hydrogels, PFD or agar gels can be used (as indicated in the text).
The transparent carrier foil allows the measurement of local oxygen concentrations, which can then be correlated to structural data, thereby considerably improving the explanatory power of the oxygen images. The same transparency did not prevent the backscattering and reflection of light from the sample surface, an unavoidable effect which must be considered in calculating oxygen concentrations by subtracting the starting image from that at the desired time-point of incubation. This feature is already implemented in the VisiSens software.
Mapping the oxygen consumption across stem sections
To monitor the stem's interior, the sensor foil was placed on the cut surface of a stem. For this study, we have chosen a range of plants with different architecture regarding position and size of the vascular bundles in the stem (Figs 2, 3). In potato, most of the vascular tissue is concentrated in the corners of the angular stem (Fig. 2a). Over a monitoring period of 50 min, the kinetics of the decline in oxygen concentration varied topologically (Fig. 2b–d; Supporting Information Movie S1), reaching its lowest level within the vasculature. Under steady-state conditions, an oxygen concentration equivalent to 65 μM was obtained in the vascular tissue and one equivalent to 110 μM in the parenchyma. In the inner cavity, only minor changes in oxygen concentration were observed. Based on the measured decline in oxygen concentration over time, the estimated respiration rate in the vascular bundle was about six-fold that obtaining in the parenchyma.
The vasculature of the maize stem is more dispersed than that of potato (Fig. 2e), and, as a result, the steady-state oxygen concentration was more homogeneous, with no pronounced differences between the vasculature and the parenchyma (Fig. 2f). In oilseed rape, the stem section is dominated by a pronounced peripheral ring of vascular tissue (Fig. 3a). When the sensor foil was placed on the stem section, the concentration of oxygen first declined in the vascular ring, and the resulting gradient in oxygen concentration which formed over time identified the ring of vascular tissue as the major oxygen sink (Fig. 3b–f). Overall, therefore, the vasculature appeared to have the highest respiratory demand, a feature that is readily attributable to its physiological function, which probably requires a substantial level of metabolic activity.
Mapping the oxygen consumption within the developing seed
Microsensor-based studies have already demonstrated hypoxia to be a characteristic condition within the interior of the seed (Borisjuk & Rolletschek, 2009). Here, the planar sensor allowed for the first time the elaboration of a two-dimensional plot of oxygen distribution across the developing seed. For this purpose, the barley and pea samples (Fig. 4a,b) were sliced through their centre and covered with the sensor foil, allowing the oxygen distribution within the seed to be monitored (Fig. 4c,d). Within an incubation time of 15 min, the lowest concentrations were obtained in the innermost part of the seed (Fig. 4e,f), a result expected from earlier microsensor-based experiments (Borisjuk & Rolletschek, 2009; Rolletschek et al., 2009). Notably, it was possible to demonstrate that the hypocotyl in pea was also characterized by low oxygen concentrations, indicating a locally elevated respiration rate, clearly distinct from that obtaining in its surrounding tissue (Fig. 4d).
Oxygen exchange in the leaf
The diffusion of oxygen through air is c. 10 000 times faster than through water (Stumm & Morgan, 1996), so the sensor is unable to map the oxygen distribution in the air space above a sample. However, by replacing the air between the sensor foil and the leaf with a liquid medium – in this case PFD – it was possible to circumvent this limitation. PFD is colourless and inert, but has a high capacity for dissolving both oxygen and carbon dioxide. In the case of oxygen, solubility is 17 times higher in PFD than in water and in the case of carbon dioxide it is at least 2.3 times higher (King et al., 1989), allowing PFD to have been developed as a blood substitute used in medicine and biotechnology (Lowe, 2003). PFD has been previously used to improve the resolution of Arabidopsis thaliana mesophyll cells imaged by confocal microscopy, and has little, if any, impact on the physiology of the leaf (Littlejohn et al., 2010). This assumption was confirmed by our own measurements using a chlorophyll fluorescence imaging technique elucidating photosynthetic key parameters (Fig. S1a). The maximum quantum yield of PSII (Fv/Fm) in dark-adapted leaves of sycamore was unaffected by the presence of PFD (0.779 ± 0.009 in its absence, and 0.778 ± 0.016 in its presence). Subsequent illumination of the leaf segments induced photosynthetic ETR whether the leaf had been exposed to air or to PFD (Fig. S1b). The ETR remained stable also over a prolonged period of illumination (data not shown). These data let us conclude that immersion of the leaf in PFD has no significant impact on photosynthetic properties, and is thus applicable for our purpose.
Before oxygen imaging, we immersed the leaf of hibiscus in PFD for 30 min by enclosing it along with PFD. The recording of the oxygen data itself was performed in a custom-made chamber (to avoid the volatilization of the PFD), consisting of two glass plates and sealed with gas-tight clay (Blu-Tack; Bostik Limited, Leicester, UK). This simple procedure allowed PFD to fill the internal air spaces of the leaf while avoiding the volatilization of PFD almost completely. Note that, for experiments with a time period not longer than 60 min, the sealing with Blu-Tack was dispensable. The sensor foil was applied to the abaxial surface of the hypostomatic leaf, placed in the dark, and the emission of signal was followed over the next 60 min (Fig. 5a,b). The oxygen concentration above the intercostal leaf area declined much more rapidly compared with the veins, and incorporated all the stomatal oxygen signals. Thus, oxygen distribution was aligned to the leaf structure (Fig. 5c). Please note that veins usually have much fewer stomata and the cuticle has much reduced oxygen permeability (Lendzian, 1982; Frost-Christensen et al., 2003). Positioning the sensor on the adaxial surface (which lacks stomata) showed that the light–dark switch induced no change in oxygen concentration (data not shown).
For the continuously illuminated (15 min at 23 μmol photons m−2 s−1) leaf, there was a rapid increase in oxygen concentration above the leaf surface as a result of photosynthetic oxygen evolution (Fig. 5d). The addition of DCMU, a potent inhibitor of photosynthetic electron transport, abolished the light-induced increase in oxygen concentration at the leaf surface (Fig. S2).
A peculiarity of the monocotyledonous species wandering Jew is that its leaf carries only c. 20 stomata mm−2 (in contrast to the c. 300 in hibiscus), although each stoma is large (Fig. 6a). Stomata are spatially well separated and sited within a small cavity of the epidermis at a depth of c. 30–40 μm (depth is colour-coded; see Fig. 6b), which allows the monitoring of gas exchange at the level of a separate stoma. In leaves initially exposed to light, transfer to the dark induced an overall decline in oxygen concentration, visualized as spatially separated spots, each corresponding to an individual stoma (blue spots in Fig. 6d). The oxygen concentration in these areas was estimated to be c. 25 μM (after correction for backscattering from the sample surface; see Fig. 6c). Subsequent illumination depressed the signal rapidly, indicating an increase in the oxygen concentration above the leaf surface as a result of photosynthetic activity (data not shown). Thus, the imaging system can monitor the oxygen dynamics of leaves, achieving a level of spatial resolution sufficiently high to enable monitoring of the oxygen exchange of individual stomata. Alternative platforms able to replicate this level of resolution (e.g. scanning electrochemical microscopy; see Tsionsky et al., 1997) are technically much more complex.
Mapping the oxygen dynamics – other applications
To demonstrate the wide applicability of the novel sensor approach, we performed oxygen mapping on various objects. First, we quantitatively visualized the oxygen consumption of an intact root system, and how this is reflected in the oxygen status of the immediate surroundings (rhizosphere). For this purpose, seeds of Brassica napus were allowed to germinate on agar plates, and the growing root was covered with the sensor foil (Fig. 7a). During the following incubation, the oxygen concentration decreased markedly, as recorded for various regions across the root system, reflecting root respiration. The image shown in Fig. 7(d) illustrates the oxygen status in the rhizosphere after 45 min. Peak signal (i.e. the lowest oxygen concentration) was associated with the primary root. The respiratory activity of the root clearly influenced its immediate environment; the signal was most strongly affected close to the primary root, but much less so at a distance of a few mm (Fig. 7g). Thus, the imaging method permitted the two-dimensional mapping of oxygen concentration in the rhizosphere of intact, respiring roots. Notably, the plant was completely intact during the analysis, and roots/rhizosphere can thus be monitored over longer time periods if needed. Secondly, the new imaging system permits the monitoring of the oxygen consumption by pathogenic fungi. Powdery mildew (Uncinula tulasnei) settles on the adaxial surface of the maple leaf and grows along the veins. Fig. 7(b) shows the growing mycelium of the fungus captured through the transparent sensor foil. Using PFD as an interfering medium, the simultaneous visualization of oxygen concentration above an infected leaf area in comparison to noninfected leaf areas was feasible (Fig. 7e). Whereas the oxygen concentration above the adaxial surface of the noninfected leaf area was almost unaffected during 30 min of dark incubation (because of the lack of stomata), the oxygen concentration above the infected area rapidly decreased (mainly because of the respiration of the fungus). Hence, we believe that imaging of oxygen concentration with planar sensors is a promising method for studies of the impact of biotic stress on plants. As a third application of the planar oxygen sensor we demonstrate the visualization of oxygen consumption by single bacteria colonies (Fig. 7c). Application of the transparent sensor to a set of Escherichia coli colonies grown on an agar plate allows localization and separate measurement of respiratory activity of single bacterial colonies. Short-term incubation (5–20 min) with the oxygen sensor resulted in clear images of the oxygen distribution, where single colonies were identifiable by their oxygen consumption (Fig. 7f). The oxygen concentration above the colonies decreased very rapidly, reaching 20 μM within 20 min (Fig. 7i). In this way, the sensor technique enabled the simultaneous screening of the respiratory activity of hundreds of bacterial colonies. Another application would be to screen for respiratory phenotypes in yeast or any other microorganism.
Advantages and limitations of the method
The novel feature of the sensor set-up is the two-dimensional mapping of oxygen concentration at a microscopic level of resolution. The present experiments have demonstrated for a wide range of species its potential for imaging oxygen gradients in plant tissue. Their temporal changes offer the capability of spatially resolved tracking of both respiration and photosynthesis. For interpretation of the acquired images, it should be noted that the planar sensor also has some limitations which are inherent in the underlying principle. First, the technique provides information on oxygen distribution across the sample at a given time-point but the measured local oxygen concentration depends on a number of factors and does not necessarily reflect the steady-state concentration in vivo. Thus, we cannot draw conclusions about the existence of hypoxia/anoxia from the oxygen images, but rather where and why oxygen depletion and evolution are most likely to occur. Secondly, during long-term experiments the generated oxygen image may not only depend on the localized consumption (evolution) of oxygen but also be affected by lateral diffusion of oxygen. Thereby, strongly oxygen-consuming regions can affect oxygen concentrations in their surroundings. Thirdly, although the generated oxygen images are quantitative (based upon external calibration of the sensor foil), real quantification of oxygen fluxes is limited. This is because the underlying volume, in which the change in oxygen concentration over time is monitored, remains unknown. Fourthly, oxygen diffusion resistance is known to affect oxygen flux and steady-state oxygen concentrations. Local differences in diffusive resistance may therefore affect the acquired oxygen images and one should, if possible, consider this tissue property for data interpretation. Finally, in the case of tissue pieces and sliced probes, wounding might also affect the measured oxygen concentrations. Based on experiments with oxygen-sensitive, needle-type sensors in seeds we did not observe large wound-induced increases in local respiration rates. As long as any wounding effect is spatially homogenous, this does not represent a problem for the imaging approach. Additional experiments might be advisable here, depending on the sample type. Taking these limitations into account, the new method clearly offers the exciting ability to measure oxygen spatially, thus allowing the monitoring and evaluation of the localized consumption and evolution of oxygen in plants. Its compact design and easy handling makes the planar oxygen sensor a promising tool in plant science.
The vasculature is the major oxygen sink of stems
The stem segment measurements produced a series of spatially resolved oxygen maps, which indicated that the stem vasculature is respiratorily more active than its surrounding parenchyma, thereby identifying it as the major oxygen sink in the stem. This result is consistent with the microsensor-based study by van Dongen et al. (2003). Please note that the vascular region of the stem is characterized by the presence of few intercellular air spaces, a feature certainly increasing local diffusive resistance. In addition, the level of metabolic activity (and thus respiration) in the phloem must be relatively high, given that the transport of assimilate is an active process (van Bel & Knoblauch, 2000; Opik & Rolfe, 2005). The combination of these factors would explain the basis for the hypoxic state which seems to prevail in the vasculature (van Dongen et al., 2003). This conclusion needs further support, but the literature already provides some (indirect) evidence for hypoxic metabolism in the vasculature of plants. The common plant response to oxygen limitation (hypoxia/anoxia) involves, inter alia, the induction of alcohol dehydrogenase (ADH) and sucrose synthase activity (Geigenberger, 2003; Borisjuk & Rolletschek, 2009; Thiel et al., 2011). In maize, ADH transcript accumulates to a higher degree in the vasculature than in the epidermis (Nakazono et al., 2003), while in various tree species, the oxygen concentration declines from the outer towards the inner sapwood (Spicer & Holbrook, 2005). Ethanol accumulates and a high ADH activity obtains in the vascular cambium and xylem sap of various species (Kimmerer & Stringer, 1988; MacDonald & Kimmerer, 1991), while the ADH protein has been detected within the xylem sap of a Populus trichocarpa × P. deltoides (poplar) hybrid (Dafoe & Constabel, 2009). The phloem also carries a number of enzymes associated with redox regulation and the anti-oxidation response (Walz et al., 2002; Lin et al., 2009; Rodriguez-Medina et al., 2011). These enzymes belong to a prominent class of hypoxia-induced genes, and may play a role in counteracting anoxic injury (Geigenberger, 2003). Nitric oxide (NO) has also been proposed as an important metabolite in hypoxic plant cells (Dordas et al., 2003; Borisjuk et al., 2007; Borisjuk & Rolletschek, 2009), while NO concentrations are particularly high in stem cross-sections, particularly in the region of the vasculature (Gabaldón et al., 2005; Corpas et al., 2006). Such findings have indeed provided indirect evidence for oxygen depletion in the vasculature, but the use of a planar oxygen sensor has managed to directly demonstrate the high respiratory demand in this tissue type.
As further examples of the applicability of planar oxygen sensors we have chosen leaves, seeds, roots and even nonplant organisms. The oxygen maps in developing seeds confirmed earlier results (Rolletschek et al., 2004; Borisjuk & Rolletschek, 2009), but can visualize for the first time the two-dimensional pattern of oxygen concentration and its consumption in high resolution for the complete seed. Analysis of the root's oxygen status under varying environmental conditions has already attracted much attention because of its relevance in plant ecology and agriculture (e.g. resistance of crop plants to waterlogging). Using needle-type sensors, the oxygen flux into roots was accurately measured (Armstrong et al., 1994; Bidel et al., 2000; McLamore et al., 2010). The planar oxygen sensor permits an additional analysis of the effects of root respiration on the oxygen status in the immediate rhizosphere. This feature could produce a number of new insights into the plant–soil interaction (Frederiksen & Glud, 2006) and could facilitate improvements in our current understanding of the dynamic interplay between root metabolism and the rhizosphere microbial community (Blossfeld et al., 2011). Another application is the use of the planar sensor to monitor changes in oxygen consumption upon local pathogen infection.
The novel feature of planar oxygen sensing is that the rate of oxygen production and/or consumption can be visualized in specific parts of the plant. The resulting oxygen maps are quantitatively based, and enjoy a resolution level in the sub-millimetre range. Their utilization will facilitate the study of the respiratory dynamics within a complex tissue, as well as oxygen homeostasis in the plant's immediate surroundings. As a result, it should become possible to investigate some of the mechanisms underlying cellular growth and the interaction of the plant with its immediate environment. Beyond its application to plants, the planar oxygen sensor system could also be of interest in other fields such as microbiology and medicine. In the medical field it provides the means to document the oxygen status of cells during pathogen infection or carcinogenesis, which could aid in the identification of appropriate drug targets (Babilas et al., 2005).
For this study, a decision was made to initially target oxygen as an analyte because advantage could be taken of a well-established fluorescence quenching technique (Holst & Grunwald, 2001; Rolletschek et al., 2009; Tschiersch et al., 2011), but also because oxygen is such a key analyte in plant biology. Future development of planar sensor technology will probably concentrate on the quantification of carbon dioxide and ammonium, and the measurement of local pH. In addition, multi-analyte sensors able to detect oxygen and other analytes simultaneously can in principle be designed by combining an oxygen-sensitive dye (and its reference dye) with dyes specific for other analytes. Such a tool would generate an increasingly holistic view of a number of important physiological processes.
We acknowledge funding by the Bundesministerium für Wirtschaft und Technologie within the framework of Zentrales Innovationsprogramm Mittelstand (ZIM). We also wish to thank Steffen Wagner and Stefan Ortleb for excellent technical assistance.