Lignin is an abundant phenylpropanoid polymer produced by the oxidative polymerization of p-hydroxycinnamyl alcohols (monolignols). Lignification, i.e., deposition of lignin, is a defining feature of secondary cell wall formation in vascular plants, and provides an important mechanism for their disease resistance; however, many aspects of the cell wall lignification process remain unclear partly because of a lack of suitable imaging methods to monitor the process in vivo. In this study, a set of monolignol analogs γ-linked to fluorogenic aminocoumarin and nitrobenzofuran dyes were synthesized and tested as imaging probes to visualize the cell wall lignification process in Arabidopsis thaliana and Pinus radiata under various feeding regimens. In particular, we demonstrate that the fluorescence-tagged monolignol analogs can penetrate into live plant tissues and cells, and appear to be metabolically incorporated into lignifying cell walls in a highly specific manner. The localization of the fluorogenic lignins synthesized during the feeding period can be readily visualized by fluorescence microscopy and is distinguishable from the other wall components such as polysaccharides as well as the pre-existing lignin that was deposited earlier in development.
Lignin is an abundant phenylpropanoid polymer produced by the oxidative polymerization, primarily, of p-hydroxycinnamyl alcohols (monolignols), mainly coniferyl alcohol 1G and sinapyl alcohol 1S, and typically minor amounts of p-coumaryl alcohol 1H (Figure 1a). It is a major component of plant cell walls, particularly in vascular tissues, where it is essential for plant growth and development. Lignin biosynthesis is a highly conserved trait within vascular plants and its variability is considered to correlate closely with the diversity and evolution of land plants. Research on lignin biosynthesis and bioengineering has been a major focus particularly because lignin hinders many agro-industrial processes such as those that generate pulp and paper or biofuels from lignocellulosic plant biomass (Boerjan et al., 2003; Bonawitz and Chapple, 2010; Vanholme et al., 2012; Weng and Chapple, 2010).
Cell wall lignification successively involves the biosynthesis of monolignols in the cytosol, translocation of the monolignols to the cell wall and its preformed polysaccharide matrix and, finally, oxidative polymerization of the monolignols to form the lignin polymer. Extensive studies have elucidated the prominent biochemical pathways for the synthesis of monolignols, although there are pathway differences evident among the various plant lineages and new details continue to emerge; most of the enzymes involved and the corresponding genes have been identified and characterized. In parallel, functional genomic studies have demonstrated that perturbation of the monolignol pathway can drastically change not only lignin content but also lignin composition and subunit character (Bonawitz and Chapple, 2010; Vanholme et al., 2012). Such plasticity in lignin biosynthesis opens new directions in lignin bioengineering to manipulate the polymer's structure for better plant biomass utilization (Ralph et al., 1997, 1998; Sederoff et al., 1999; Ralph et al., 2004; Weng et al., 2008, Vanholme et al., 2012).
Despite our detailed knowledge of monolignol biosynthesis, many of the molecular events that occur during the biosynthesis of the lignin polymer remain to be diagnostically elucidated (Fagerstedt et al., 2010; Li and Chapple, 2010; Liu, 2012; Schuetz et al., 2013). For example, our knowledge regarding monolignol transport from the cytosol to the apoplast is limited, although recent studies suggest that specific membrane proteins including ABC transporters may be involved in this process (Ehlting et al., 2005; Kaneda et al., 2008; Miao and Liu, 2010; Alejandro et al., 2012). It is currently unknown how or if lignin deposition, i.e., polymerization of monolignols, is spatially regulated within the cell walls. Metabolic supply of the monolignols and/or localization of polymerization catalysts may partially dictate this process (Olson and Varner, 1993; Wi et al., 2010; Sasaki et al., 2006; Berthet et al., 2011; Lee et al., 2013). Although much information has accumulated on the above topics, our improved understanding demands novel techniques that facilitate genetic, biochemical, and cell biological analyses of the events associated with cell wall lignification processes.
The direct monitoring of cell wall lignification processes in live plants and cells is a promising strategy to gain insight into the elusive aspects of lignin biosynthesis. Unlike proteins, and similarly to polysaccharides or lipids, lignins are not amenable to being genetically tagged for visualization in living systems. Instead, numerous imaging techniques, including light and electron microscopy coupled with lignin stains (Lewis and Yamamoto, 1990; Donaldson, 2001; Fromm et al., 2003), direct microspectrophotometric detection (Singh et al., 2009; Sun et al., 2011; Ding et al., 2012; Gierlinger et al., 2012; Donaldson, 2013), mass spectrometry-based chemical imaging (Saito et al., 2005, 2012; Jung et al., 2012), and immunochemical labeling using lignin-specific antibodies (Ruel et al., 2009; Tranquet et al., 2009; Kiyoto et al., 2013), have been developed and used to visualize lignins in plant tissues. However, these methods typically involve several disadvantages either in sensitivity, specificity, sample-preparation times, artifacts from sample fixation, and/or applicability to living plant systems.
In a preliminary study, we exploited the utility of some fluorescence-tagged coniferyl alcohol probes (Figure 1a) for studying the in vitro lignin polymerization process (Tobimatsu et al., 2011). We showed that these probes can be incorporated integrally into synthetic lignin polymers produced via in vitro peroxidase-catalyzed polymerization resulting in lignin polymers brightly labeled with the fluorogenic dyes (Figure 1b). Given the pronounced plasticity of lignification in vivo, the developed fluorescence-tagged monolignols may provide an opportunity to image lignification processes by incorporating them metabolically into living plants and potentially offering new tools to visualize the dynamics of the cell wall lignification processes in vivo via fluorescence microscopy. In this study, we have expanded upon the above work and have included newly synthesized sinapyl and p-coumaryl alcohols tagged with blue dimethylaminocoumarin (DMAC) and green nitrobenzofuran (NBD) fluorophores. To test their applicability for in vivo imaging and to provide an indication of the types of in-depth studies that might be contemplated, we performed a series of feeding experiments in various plant systems.
Results and Discussion
Synthesis of fluorescence-tagged monolignols
Our previous study (Tobimatsu et al., 2011) demonstrated that the fluorogenic derivatives of coniferyl alcohol γ-coupled by ethylenediamine spacers to DMAC or NBD fluorophores (2G and 3G) are compatible with peroxidase-catalyzed polymerization in vitro, a process that mimics the typical lignification process in vivo (Fournand et al., 2003; Barakat et al., 2007; Demont-Caulet et al., 2010; Tobimatsu et al., 2010, 2012). Importantly, whereas plant peroxidases oxidatively modify a wide range of aromatic compounds, both DMAC and NBD fluorophores remain intact during the typical lignin polymerization process and brightly label the resultant lignin polymers (Figure 1b). In this study, the fluorescence-tagged monolignols consisted of all the three major types of monolignols in combination with DMAC and NBD fluorophores (Figure 1a) were synthesized via slightly modified synthetic protocols (Figure 1c, see also Appendix S1 for details). In brief, commercially available hydroxycinnamic acids (6G, 6S, and 6H) were converted into γ-carboxymethylated monolignols (9G, 9S, and 9H) by γ-selective reductions and alkylations, and then conjugated with DMAC or NBD fluorophores bearing ethylenediamine linkers (dye 4a or 5a) (Cotté et al., 1999; Joullie et al., 2003; Sandler et al., 2005) via carbodiimide-promoted amide coupling. Optical properties of the probes were verified by UV-vis absorption and fluorescence spectroscopies (Table 1). As expected, the photophysical characteristics of the probes primarily depend on the fluorophore introduced and are not particularly affected by differences in the monolignol aromatic substitution. We expect that these monolignol analogs could reflect the properties and functions of the corresponding monolignols, and thus prove valuable to monitor their different behaviors in the lignification process in vitro and in vivo.
Table 1. Optical properties of fluorescence-tagged monolignols.a
At the onset of our imaging studies, fluorescence-tagged monolignol probes were fed into live stems of Arabidopsis thaliana via a transpiration stream. To first determine whether exogenous monolignols could be fed to live plant tissues, we initially fed the non-labeled syringyl monolignol, sinapyl alcohol 1S, to excised stems of the Arabidopsis fah1 mutant, a mutant that lacks the ability to synthesize the syringyl monolignol 1S and therefore syringyl lignins (Chapple et al., 1992; Marita et al., 1999). The use of the syringyl lignin-free fah1 mutant allows us to conveniently evaluate the incorporation of monolignol 1S and the new production of syringyl lignin by Mäule histochemical staining; the absence of syringyl lignins in this mutant was evident as its stem section does not exhibit the characteristic positive red coloration upon Mäule staining (Figure 2a). Stems of the mutant were allowed to take up medium containing 200 μm monolignol 1S for 3 days, sectioned, washed with 50% MeOH for 1 day and then stained with Mäule reagent. Newly produced syringyl lignins in xylem and interfascicular fiber tissues were clearly evident by the positive observation of a red coloration upon Mäule staining (Figure 2a). The result suggests that monolignol 1S administered through the transpiration stream leads to the specific incorporation of sinapyl alcohol in the lignifying tissues. These data indicate that exogenously provided monolignols can be incorporated when administered via the transpiration stream, thus validating this system as one with which the incorporation of fluorescence-tagged monolignols can be evaluated. We then introduced the DMAC-tagged guaiacyl monolignol probe 2G into the fah1 mutant similarly via the transpiration stream. Probe incorporation was evident by DMAC fluorescence that could be clearly observed in stem sections when visualized by epifluorescence microscopy under UV illumination (Figure 2b). As we observed in the feeding experiments using non-labeled monolignol 1S, the incorporation of DMAC-tagged probe 2G was likewise specific to the walls of xylem and interfascicular fiber cells. These data indicate that probe 2G can be delivered to these lignifying cell walls via a transpiration stream and metabolically incorporated into lignins. The observation that the staining with the labeled monolignol probe 2G accurately reflected staining with non-labeled monolignol 1S suggests that these molecules accurately report on the incorporation of natural lignin precursors.
In parallel, we also tested NBD-tagged guaiacyl monolignol probe 3G with another feeding system in which the feeding solution was administrated to Arabidopsis stems via a peristaltic micro-pump. A polysaccharide stain, Congo red (CR) was first used to test this feeding method in Arabidopsis thaliana Col-0 ecotypes. The red fluorescence observed by confocal laser scanning microscopy (CLSM) in the epidermis, pith, xylem and fiber tissues of the stem sections fed with CR (Figure 3a) revealed that the fed solution containing CR was able to diffuse through the entire stem and stain polysaccharides independently of the tissue type. However, lower fluorescent intensity was detected in the interfascicular fiber regions, which might be caused by reduced polysaccharide accessibility or reduced CR diffusion due to intensive lignification of these tissues. The labeling with CR through all tissues validated this feeding approach and confirmed that fed compounds can flow and diffuse into the various tissues in the stem. Using this approach, 6-week-old stems were fed with 2.5 ml of either 1% NBD-tagged guaiacyl monolignol probe 3G (1 μm) or 1% of non-conjugated (but protected with a t-butoxycarbonyl group) NBD dye 5b (1 μm) as control along with a 99% unlabeled monolignol 1G and 1S mixture (49.5 μm each) to promote incorporation of probe 3G into lignifying tissues. As expected, no fluorescent signal was detected in any tissue from sections derived from stems fed with NBD dye 5b (Figure 3d). In contrast, when monolignol probe 3G was fed, NBD fluorescence was detected only in lignifying tissues such as xylem and fibers and not in non-lignified tissues such as pith and cortex (Figure 3b,c), supporting the hypothesis that probe 3G is incorporated into the lignin polymer only when and where lignification occurs. Xylem vessels exhibit a high fluorescent signal as compared to interfascicular fibers; similar results were seen in the feeding experiments using the transpiration stream as described above. This difference could be explained by the higher substrate availability for peroxidases and laccases present in the xylem. In addition, as the solution containing probe 3G is mainly transported through the vessel elements and starts diffusing toward the other tissues from the vessels, it is conceivable that the concentration of probe 3G that reaches interfascicular fibers is reduced because of the higher hydrophobicity of lignified cell wall as observed with CR feeding (Figure 3a) or the ability of the lignifying xylem cells to deplete the probe before it is able to reach the fibers.
Incorporation into Pinus radiata stems
We also performed lignin-specific staining of Pinus radiata stems with the monolignol probes 2G and 3G, in which case live stems were pre-sectioned and immediately exposed to the probes for staining. Freshly cut stem sections (60 μm) from Pinus radiata seedlings (~6-month-old) were incubated for 1 h in aqueous solutions containing 1 μm probe 2G or 3G in combination with 100 μm non-labeled monolignol 1G to promote incorporation of fluorescence-tagged monolignol probes, after which they were washed with ethanol to remove non-incorporated precursors and observed by CLSM.
The application of the DMAC probe 2G or the NBD probe 3G resulted in the appearance of fluorescence specifically in lignifying xylem tissues on the inner side of the cambial zone, suggesting that substantial lignification was occurring in those tissues during the feeding period (Figure 4a,b); phenol oxidase activity for lignification in those tissues was confirmed by histochemical staining with tetramethylbenzidine (Figure S1) (Imberty et al., 1984; Ros-Barcelo et al., 2006). Co-visualization with blue DMAC fluorescence and green cell autofluorescence allowed simultaneous imaging of newly labeled lignin and pre-existing lignins that were formed before the feeding period (Figure 4c). In all the staining experiments, the tracheids at different developmental stages clearly showed different probe incorporation patterns that most likely reflect the developmental shifts in the lignification sites during the process of xylem formation. In the cell layers composed of early tracheids adjacent to the cambium, the probe incorporation is only seen in the cell corners (Figure 4a–c, images i). As the tracheid cells are further distant from the cambium, i.e., as the cells become more mature, the probe fluorescence gradually spreads to the compound middle lamella (Figure 4a–c, images ii), and then to the secondary cell walls (Figure 4a–c, images iii). These observations indicate that the lignification first occurs in the cell corner, subsequently in the compound middle lamella and finally in the cell walls, which is essentially consistent with the earlier results obtained by other techniques (Terashima and Fukushima, 1988; Terashima et al., 1988).
To further evaluate the ability of probes 2G and 3G to specifically label lignin-producing sites, a series of control experiments was carried out, and the labeled and control sections were imaged under identical microscopic conditions for comparison (Figure 5). Heat treatment to quench enzyme activity (Figure 5b and Figure S1) considerably reduced incorporation of the probes. In addition, treatment with non-conjugated fluorescent dyes 4b or 5b, which lack the ability to participate in lignin polymerization, showed only weak fluorescence under the same microscopic conditions used for probes 2G or 3G (Figure 5c). These results confirmed that the probes are primarily incorporated via their participation in lignin polymerization occurring in vivo. Treatment only with non-fluorescent monolignol 1G showed very weak fluorescence in lignifying xylem tissues, suggesting that contributions by lignin autofluorescence in imaging with probes 2G and 3G are minimal (Figure 5d). A close comparison between the labeled and unlabeled section images suggested that blue autofluorescence under UV excitation may still contribute in the sections imaged with DMAC probe 2G particularly in unlabeled regions, presumably where pre-existing lignins are abundant. On the other hand, green autofluorescence under the conditions used for NBD fluorescence was almost absent in the control unlabeled sections. The results indicate a significant brightness advantage for the NBD probes over the DMAC probes in overcoming the autofluorescence of plant materials (Figure 5 and Figure S2).
Incorporation into Arabidopsis thaliana live seedlings
To further illustrate the use of fluorescence-tagged monolignols in live plants, we performed feeding experiments using live Arabidopsis seedlings, an experimental system that has been used frequently in live plant imaging studies (Brandizzi et al., 2002; Paredez et al., 2006; Dhonukshe et al., 2008; Okumoto et al., 2008; Ckurshumova et al., 2011; Swanson et al., 2011; Anderson et al., 2012). Seedlings of Arabidopsis thaliana Col-0 ecotypes were incubated in liquid Murashige Skoog (MS) medium that contained between 10 and 100 μm NBD-tagged monolignol probe 3G, and then directly observed by CLSM. The seedlings treated with probe 3G exhibited robust NBD fluorescence throughout the root tissues, clearly showing the probe diffusion into the root whereas, in contrast, no fluorescence was observed in the seedlings treated with non-labeled monolignol 1G (Figure 6a). In the elongation zone near the root tip, NBD fluorescence was observed inside undifferentiated cells (Figure 6b), a finding that indicated that probe 3G is either plasma-membrane-permeable or is actively imported into the cells. In the differentiation zone, the fluorescence was primarily observed in the apoplastic compartment. In particular, strong fluorescence was observed in the developing protoxylem and metaxylem (Figure 6c), and suggests that probe 3G is well incorporated into the cell wall lignins developed in the vascular cylinder.
We noted that the seedlings treated with probe 3G showed dot-like NBD fluorescence in an optical section through the endodermis (Figure 6d), as would be expected for signals from the Casparian strip – a specialized cell wall modification that serves as a physical barrier to apoplastic diffusion (Petricka et al., 2012). It has long been controversial, but recently well established by Naseer et al. (2012), that Arabidopsis Casparian strips are primarily made of lignins. To further determine whether the probe is incorporated into Casparian strips, we used an apoplastic tracer, propidium iodide (PI), that cannot penetrate into the stele past the Casparian strips (Naseer et al., 2012; Lee et al., 2013). Seedlings were treated successively with probe 3G and PI and then co-localized by CLSM with green and red fluorescence from NBD and PI (Figure 6d). Consequently, and as expected, the appearance of the dot-like NBD fluorescence coincided with the sites where the diffusion of PI is blocked by the Casparian strips. Therefore, the observed dot-like localization of probe 3G is most likely to be due to its incorporation into the lignins actively produced in the Casparian strips. Collectively, these results further demonstrate that probe 3G is able to penetrate live intact tissues and accumulate in lignin-containing structures in vivo.
Subcellular localization in Arabidopsis protoplast cells
Finally, to identify where in individual plant cells the fluorescence-tagged monolignols might localize, NBD-tagged monolignol probe 3G was incubated with Arabidopsis protoplast cells and subcellular localization was examined by CLSM. FM4-64 was used as a marker for the plasma membrane (Figure 7). The NBD fluorescence was clearly observed within the cells treated with 50 μm probe 3G for 4 h, whereas no fluorescence was observed in the control cells without the treatment (Figure S3). The observation of the cell treated with probe 3G revealed that the probe was specifically localized within the cytoplasm; this cytoplasmic localization remained the same after prolonged incubation (24 h), with no particular localization in the vacuole. Thus, at least under the used conditions, probe 3G was not permeable through the vacuolar membrane, whereas it readily penetrated the plasma membrane. Within our survey, probe 3G did not associate particularly with plasma membrane and accumulated mainly in the cytosolic compartment of the cell (Figure 7e). Overall, these results indicate that probe 3G can easily penetrate the plasma membrane and diffuse into the cytoplasm in cells.
In summary, we have synthesized a set of fluorescence-tagged monolignols that can be used for a variety of in vivo imaging studies aimed at understanding the plant cell wall lignification process. The fluorescent monolignol probes can be introduced into intact plant tissues and cells under various feeding regimens. Metabolic labeling using either DMAC- or NBD-tagged monolignol probes in Arabidopsis and Pine stems demonstrated their compatibility with cell wall lignification in vivo; incorporation of the probes specifically occurs in the walls and compound middle lamella of xylem and fiber cells where lignins are being actively produced. The resultant labeled lignins can be directly visualized by microscopy, and are readily distinguishable from the other wall components such as polysaccharides and the pre-existing lignins that are deposited prior to the feeding period. These monolignol probes could therefore be particularly useful to monitor active lignification in vivo without unnecessarily disturbing plant physiology and environment. Labeling and direct imaging experiments using NBD-tagged monolignol probes in Arabidopsis seedlings further illustrated the utility of the probes in live plant imaging studies. It was shown that the probe can penetrate intact live tissues and participate in cell wall lignification processes in the root system, e.g., in developing xylem tissues in the stele and Casparian strips in the endodermis. Subcellular localization of the NBD-tagged monolignol probe in Arabidopsis protoplast cells clearly showed that the probe localization is cytoplasmic and that the probe itself does not notably associate with the plasma membrane. Taken together with our observations in other feeding systems, the interference by background fluorescence due to non-specific probe binding to the components in plant cells/tissues should be minimal. Related studies in this area are expected to expand our understanding of plant cell wall lignification processes and ultimately enable more efficient manipulation of cell wall structure and architecture.
Detailed synthetic protocols and characterization data for fluorescence-tagged monolignols and synthetic lignins are described in the Supporting Information (Appendix S1 and Data S1). UV-vis absorption spectra were recorded on a Shimadzu BioSpec-nano spectrophotometer equipped with a quartz cell adapter. Fluorescence spectroscopy was conducted with a PTI QuantaMaster Model C-60/2000 spectrofluorometer (Photon Technology International) at 25 ± 0.1°C and data acquisition used FelixGX software (Photon Technology International). Fluorescent quantum yields (Φf) were determined according to the method described in the literature (Fery-Forgues and Lavabre, 1999), using anthracene (Sigma-Aldrich, http://www.sigmaaldrich.com/; λem = 350 nm, Φf = 0.27 in EtOH, η = 1.36) or fluoresceine (Sigma-Aldrich; λem = 450 nm, Φf = 0.92 in 0.1 N NaOH aq., η = 1.33) as standards.
Arabidopsis thaliana stems fed via a transpiration system
Primary inflorescence stems of Arabidopsis thaliana (Col-0) ecotype and fah1-2 mutants were harvested from 6-week-old plants while submerged in water and kept in water until ready for incubation with feeding solutions. Monolignol 1S or probe 2G was dissolved in dimethyl sulfoxide and diluted to a final concentration of 200 μm 1S or 2G in water. Each stem was incubated in water containing 200 μm of 1S or 2G under continuous light. Mäule staining was performed as previously described (Chapple et al., 1992) with minor modifications. Hand-sectioned stem sections were fixed in 4% (v/v) glutaraldehyde for 1 h. After rinsed with water three times, sections were incubated in 0.5% (w/v) potassium permanganate solution for 10 min. Sections were rinsed with water multiple times until the solution was cleared, and then were incubated in 10% (w/v) HCl for 5 min and rinsed with water twice. Sections were then mounted in concentrated ammonium hydroxide and examined by bright field microscopy. DMAC fluorescence was detected using a Nikon E800 epifluorescence microscope using filters with an excitation wavelength of 360 ± 20 nm and emission wavelength of 420 nm.
Arabidopsis thaliana stems fed via a micro-pump system
Stems from 6-week-old Arabidopsis thaliana ecotype Col-0 plants were used to study de novo lignification of the green probe 3G in planta. Stem fragments of approximately 8–9 inches were collected at mid-day from the base of freshly cut Col-0 inflorescences. Immediately after harvesting, the stem base was attached to the sleeve of a 3 ml syringe and fed with 2.5 ml of feeding solution, containing probe 3G (1 μm) or dye 5b (1 μm) along with unlabeled monolignols 1G and 1S mixture (49.5 μm each), using a KDS230 micro-pump (KD Scientific) at a continuous rate of 2 μl min−1 at room temperature. In the case of CR feeding, the stem was fed with 2.5 ml of 0.5% CR (Sigma-Aldrich) solution. In order to extend the lignification period, fed-stems rested at room temperature for an additional 16 h after the end of the feeding prior to any measurement. Then the stem was embedded in 7% agarose, sectioned to 100 μm using a Leica VT1000S vibratome, washed with 96% ethanol for 3–5 min, and then placed in water. The imaging used a Carl Zeiss 710 LSM CLSM with excitation/emission at 488/492–534 nm (NBD) and 514/519–650 nm (CR).
Pinus radiata stems
Stems of 6-month-old Pinus radiata stems were harvested in early autumn, a period when secondary cell wall formation in developing xylem is dominated by lignification. Stem sections with secondary growth (60 μm) were sectioned immediately after harvest using a sledge microtome and placed into water. Enzymatic activity in control sections was quenched by incubating sections for 1 h at 85°C in water. Untreated sections and controls were subsequently transferred to aqueous solutions containing 1 μm monolignol probe 2G or 3G in combination with 100 μm non-labeled monolignol 1G for 1 h. After staining, sections were washed with 50 ml 96% ethanol and 50 ml water under gentle agitation for 1 h each to remove non-incorporated precursors. Peroxidase activity in stem sections was monitored as described (Ros-Barcelo et al., 2006). Both the staining of sections in the presence of 0.1 mm ferulic acid and heat treatment of sections (see above) completely abolished peroxidase-dependent staining (Figure S1). Sections were imaged using a Leica SP5 II CLSM. For DMAC probe 2G, sequential imaging was used with excitation at 355 nm and 496 nm, and emission was acquired sequentially from 400–485 nm (DMAC) and 510–600 nm (autofluorescence). Some images were also made with only UV excitation and emission from 400–500 nm. A control section showed some weak lignin autofluorescence in the 400–485/500 nm range at the gain used for imaging the treated sections. For NBD probe 3G, excitation and emission settings were 488 and 500–650 nm. Autofluorescence in control sections was negligible in this emission range at the same gain. Sections treated with TMB to detect peroxidase were mounted in 50% glycerol and imaged using a Leica MZ12.5 stereomicroscope using transmitted light.
Arabidopsis thaliana seedlings
Seeds of Arabidopsis ecotype Col-0 were surface-sterilized and transferred to 12-well plates (6 seeds per well) with 2 ml liquid ½MS medium in each well. The plates were left overnight at 4°C for stratification after which they were transferred to a shaker in a temperature-controlled chamber (21°C, 130 rpm, 16 h light/8 h dark light cycle). After 3 days, probe 3G was added (to a total concentration of 100 μm) and then the seedlings were further grown for 1 day after which the live seedlings were imaged. For co-staining with PI, seedlings were treated with 10 μm probe 3G for 2 days under the conditions described above, washed, and then treated with 10 μm PI for 10 min after which the seedlings were imaged. For CLSM, a Carl Zeiss LSM5 EXCITER with 488 (NBD) and 543 (PI) nm excitation, and 505–530 (NBD) and >560 (PI) nm emission.
Arabidopsis thaliana protoplast cells
Protoplasts were isolated from fully expanded rosette leaves of Arabidopsis ecotype Col-0 (60 days old) according to literature (Wu et al., 2008) with minor modifications. The buffers used for protoplast isolation procedure, i.e., protoplast enzyme solution and wash buffer, were prepared according to literature (Robert et al., 2007). Peeled leaves were placed in a Petri dish with protoplast enzyme solution for 90 min in the dark, on an Excella shaker (Eppendorf) at 70 rpm. The remainder of the leaf tissue with tape was removed and a Cellector Tissue Sieve (EC Apparatus, USA) was used to clean the protoplasts from the remaining undigested tissue. Protoplasts were spun down for 20 min (20°C, 80 g) and the protoplast enzyme solution was replaced with the wash buffer. The washing step was repeated. Probe 3G was dissolved in wash buffer and as such added to the protoplast suspension; the final concentration of the probe was 50 μm. The protoplast suspension was gently shaken for 4 h. In order to stain the plasma membrane of protoplasts, a 5-min treatment with 4 μm FM4-64 (Invitrogen) was performed. All treatments were carried out in the wash buffer at room temperature in the dark and at least in triplicate, with a minimum of 20 protoplasts checked for each treatment. For CLSM, a Carl Zeiss 710 LSM with excitation/emission at 488/505–530 nm (NBD) and 543/560 nm (FM4-64) was used.
We thank Darrel McCaslin (UW Biochemistry, Biophysics Instrumentation Facility) for assistance with fluorescence spectroscopy. The authors acknowledge partial funding from US Department of Energy, the Office of Science (DE-SC0006930), University of Wisconsin Vilas Associate Award, Stanford University Global Climate and Energy Project (GCEP), and Scion CORE funding. This work was also supported, in part, by the Great Lakes Bioenergy Research Center (GLBRC), Joint BioEnergy Institute (JBEI), and Center for Direct Catalytic Conversion of Biomass to Biofuels (C3Bio), an Energy Frontier Research Center, funded by the US DOE's Office of Science (DE-FC02-07ER64494, DE-AC02-05CH11231, and DE-SC0000997, respectively). YT gratefully acknowledges Postdoctoral Fellowship support from the Japan Society for the Promotion of Science (JSPS).