Tracking Galvoxite on Leaf Surface with Fluorescence Lifetime Imaging: Film Forming and Delivery Capabilities Towards Citrus Crop Protection

Abstract

Huanglongbing (HLB, citrus greening disease) continues to severely impact global citrus production, motivating the development of foliar treatments capable of protecting leaf surfaces and enabling targeted agrochemical delivery. Using optical and electron spectroscopy and imaging, a novel crop protection formulation, Galvoxite, was investigated. We tracked the location of Galvoxite on the leaf surface with advanced imaging and spectroscopic methods, including time-correlated single photon counting (TCSPC), fluorescence lifetime imaging (FLIM), scanning electron microscopy (SEM), and energy dispersive X-ray spectroscopy (EDS). A fluorescently labeled Galvoxite derivative containing tris(2,2′-bipyridyl)ruthenium(II) chloride (RuGalvoxite) was developed for this purpose. Steady-state fluorescence measurements revealed substantial spectral overlap between citrus leaf autofluorescence dominated by chlorophyll and the Ru-based fluorescent probe over the 550–750 nm spectral region, as is typical for organic dyes. However, the Ru-based dye also allows for temporal separation from a fluorescent background due to its long excited state lifetime. TCSPC data show that the fluorescence lifetime contrast between RuGalvoxite and the leaf fluorescence background provides excellent selectivity. Untreated citrus leaves displayed short excited-state lifetimes near 2 ns, whereas RuGalvoxite exhibited long-lived emission with average lifetimes ranging from several hundred nanoseconds up to localized microsecond-scale behavior in FLIM measurements. The RuGalvoxite decay behavior was characterized by predominantly monoexponential dynamics, attributed to dilution-induced reduction of Ru-based dye aggregation and suppression of non-radiative decay pathways. FLIM images reveal spatially localized long-lifetime domains on treated leaves that are clearly distinguishable from the short-lived autofluorescence background of untreated tissue. The bottom leaf surface exhibited more uniform film formation and preferential localization of RuGalvoxite near and within stomatal pores, whereas the waxy top surface promoted accumulation near glandular trichome-like structures. SEM imaging confirmed the formation of deposited film domains on the leaf surface, while EDS mapping identified Zn-rich regions corresponding to Galvoxite. These results demonstrate that fluorescence lifetime spectroscopy and imaging provide sensitive approaches for tracking foliar delivery systems in highly autofluorescent plant environments. The observed localization behavior and film-forming properties support the potential use of Galvoxite as a platform for targeted micronutrient or agrochemical delivery to stomatal regions of citrus leaves.

Graphical abstract

This is a visual representation of the abstract.

Introduction

The citrus industry has seen a decline in production over the past two decades. In Florida for instance, production has dropped from over 250 million boxes to only 20 million boxes.^1,2 One of the major factors that has affected the production is the presence of citrus greening disease, huanglongbing (HLB), which has no cure at his time. HLB has been found globally, with confirmed locations being North and South Americas, Southeast Asia, Eastern Africa, and the Caribbean.^3–5 In the U.S., HLB is caused by the Candidatus Liberibacter bacteria, with the asiaticus species (CLas) afflicting the U.S. citrus industry.^6,7 When a tree is infected with HLB it affects the quality of the fruit, creating smaller bitter tasting fruit, causing premature fruit drop, and stunting root growth.^8–10c HLB is spread through a vector, the Asian pitrus syllid (ACP).^11–13 As ACP feeds on the leaf of a citrus tree, it punctures into the phloem where CLas then enters and begins to incubate. As CLas propagates, it eventually blocks the movement of the nutrients in phloem, leading to severe root loss, and the eventual death of the tree. ¹⁰ Considering that virtually every tree in citrus growing regions is infected, when uninfected ACP feed on an infected tree they will contract the Clas bacteria, further exacerbating the issue. ⁷

Since there is no cure for HLB, preventative measures have been considered to control the spread of HLB and to protect new plantings. ¹⁴ This is particularly important since ACP prefers to feed on young citrus trees with ample flush. ¹⁵ Some of the most common methods are to block ACP from feeding on citrus plants, such as growing plants in screen enclosures or using individual protective covers for plants (IPCs). ¹⁶ Insecticide spray programs were attempted to control the ACP population. However, it proved to be ineffective because the ACP would simply move in from neighboring orchards after treatment, and even spray treatments themselves were found to not be 100% effective within the treated area, leaving ACP endemic to citrus producing regions.^11,17–19 After the effects of the insecticide subside the ACP would return. ⁵ There is also the risk that eventually ACP will become resistant to the insecticide creating a need for new formulations.^20,21 Nutritional programs can help extend the lifespan and productivity of infected trees, but merely delay the inevitable decline of infected trees.^8,22,23 Recently, use of antibiotics has become a prominent strategy in the US under EPA approval, specifically the use of oxytetracycline (OTC), although effective systemic localization of OTC remains debated. ²⁴ Regardless of limited success or failure, these different strategies all bring economic burden to growers.^25,26 However, since the HLB infection cycle can be characterized by interaction between vector and plant host, bacteria and vector, and bacteria and plant host, interrupting the vector-host interaction could break this cycle. The use of spray applications to produce protective films on plants that form a physical barrier against insects has been considered to accomplish this. ²⁷

HLB has been shown to affect root mass and consequently causes nutrient deficiency. ²⁸ Given how micronutrients are utilized as cofactors in SAR metabolic pathways, ²⁹ it is expected that HLB-affected trees will not respond adequately to other plant pathogens. For instance, another important citrus disease, citrus canker, affects the canopy of the tree and causes raised lesions on the leaves and fruit. ³⁰ This disease is caused by a different phytopathogenic bacterium, Xanthomonas citri pv. citri (XCC) This disease stresses the tree, decreasing total yield and reducing even more the marketable fresh fruit. Before showing virulence, XCC colonizes epidermal cell junctions and stomata, where it starts the disease cycle.^30,31 To manage this disease, growers often maximize the application of copper-based biocides. However, the constant application of these pesticides can cause topsoil accumulation and the onset of Cu-tolerant Xanthomonas pathogens. ³² Therefore it is imperative to find an integrative solution for nutrient deficiency and crop protection against both ACP and XCC.

Recently, Pereira et al. reported a zinc (Zn)–borate (B) based formulation that can be applied as a foliar spray. The formulation locates preferentially in the junction between the epidermal cells and around the stomata on leaf surfaces. In addition, with its Zn and B content, the formulation could act as a delivery system for these micronutrients. ³³ Nutritional foliar sprays have been shown to alleviate the ACP feeding from citrus trees. ³⁴ It is hypothesized that micronutrient deficiency, namely inadequate levels of B, cause abnormalities to the cell wall structure and secondary metabolism, encouraging the ACP to feed on the foliage.^35,36 These findings further encourage the development of nutritionally targeted delivery systems. Following up on this work, we show here that a modified formulation, referred to herein as Galvoxite, forms a film on the surface of the leaf that has the potential to physically block the ACP and phytopathogenic bacteria. Given that the formulation preferentially deposits around stomata and epidermal cell junctions (areas predominantly composed of pectin) it allows for the targeted delivery of agrochemicals, such as insecticides or bactericides, through foliar spray.

Considering the negligible fluorescence of Galvoxite we incorporated a fluorescent dye, Tris(2,2'-bipyridyl)ruthenium(II) chloride (Rubyp) in the formulation (referred to as RuGalvoxite throughout the paper) to enable optical tracking of Galvoxite on leaf surfaces. Rubyp absorbs energy in the visible light spectrum and has a characteristically long lifetime owing to metal to ligand charge transfer (MLCT) between the Ru metal core and the bipyridine ligand. ³⁷ The Rubyp dye shows typical excited state lifetimes ranging from a few hundred nanoseconds to over a microsecond depending on the local environment the dye experiences, including effects of polarity, aggregation, oxygen content, dilution, or incorporation in matrices for instance.^38–44 In a previous study, we showed that Rubyp is an excellent choice for its fluorescence properties in imaging against plant background. ⁴⁵ With the long excited state lifetime of Rubyp in the hundreds of nanoseconds (ns), we have a means for sensitive and selective detection on leaf surfaces even in the presence of abundant plant autofluorescence stemming mainly from chlorophylls, which shows excited state lifetimes below 10 ns.^46–49 Fluorescence measurement techniques such as time-correlated single photon counting (TCSPC) and fluorescence lifetime imaging microscopy (FLIM) can be used to detect these properties.

In this paper, we use TCSPC and FLIM as advanced fluorescence spectroscopic and imaging methods, respectively, together with scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDS) to detect Galvoxite on leaf surfaces, to explore its film formation properties, and to observe Galvoxite and Zn localization on leaf surface. We show that Galvoxite forms partial films on the waxy top side of the leaf and high quality films on the bottom side of the leaf. In addition, RuGalvoxite was found to encircle stomata of the leaf, and the presence of RuGalvoxite and Zn within stomatal pores was confirmed. Our results indicate that Galvoxite has film forming capabilities on the leaf surface with the potential to form a barrier between leaf and CLas vector, and with the potential to precisely deliver micronutrients or therapeutics for systemic application.

Materials and Methods

Galvoxite Preparation and Application

The formulation was prepared based on a previously published protocol with some modifications. ³³ Initially, 3.71 g of boric acid (Fisher Scientific, USA) and 4.58 g of sodium gluconate (Thermo Scientific, USA) were dissolved in 13.5 mL of 2.2 M aqueous sodium hydroxide (Fisher Scientific, USA) solution. In parallel, 8.85 g of zinc nitrate hexahydrate (Fisher Scientific, USA) were also dissolved in 2.5 mL of deionized water and added to the boric acid solution. Afterwards, the pH was adjusted to 7 by adding 4.7 mL of 5 M aqueous sodium hydroxide and the suspension was left to stir for 24 h. The final volume was adjusted with deionized water to obtain a final concentration of 57 000 parts per million (ppm) of Zn. This formulation is referred to in this work as Galvoxite. A variant of the material containing a Ru(II) complex was also prepared using the same protocol. For this variation, the same quantities of the reagents and procedure were used, but 30.0 mg of Tris(2,2'-bipyridyl)ruthenium(II) chloride hexahydrate (Acros Organic, USA) were dissolved alongside the zinc nitrate. This material is referred to as RuGalvoxite.

Leaves were acquired from a Ray Ruby grapefruit tree sapling for the experiments. Both Galvoxite and RuGalvoxite were diluted to a concentration of 800 ppm of Zn with DI water and sprayed with a hand operated pump on to the citrus leaves until saturation (10 mL). The leaves were then dried using nitrogen gas and examined for instant readings or left overnight to see visual spread.

Photographs Under UV Light

The leaves to be examined were cleaned using a 1% detergent solution of Alconox and deionized (DI) water. The leaves were placed in a centrifuge tube containing 50 mL of the 1% solution and shaken for 5 min to clean residual fertilizer or debris off the plant. The leaves were placed on paper towels and allowed to air dry for 30 min. Afterwards, the leaves were sprayed with RuGalvoxite via a hand pump and dried overnight. They were examined under a UVP UVLMS-38 EL Series 3UV Lamp at 302 nm wavelength. Photographs were then taken to record the macroscopic coverage of the formulation.

Emission Spectra

Emission spectra were obtained using a custom inverted confocal microscope described previously. ⁵⁰ A 466 nm pulsed diode laser (PicoQuant LDH-P–C-470 connected to a PDL 800-D laser driver) was used to excite the sample. The spectra were recorded using a PI Acton SP-2156 spectrograph coupled to an iXon EM + DU-897 BI EM-CCD. Andor SOLIS software was used to record spectra utilizing an acquisition time of 1 s and collecting three frames per spectrum.

TCSPC Decay Curves and FLIM Images

TCSPC and FLIM used the same setup as described above for the emission spectra. A 466 nm pulsed diode laser (PicoQuant LDH-P–C-470 connected to PDL 800-D laser driver) was used to excite the sample with a repetition rate of either 5 MHz, 250 kHz or 125 kHz, depending upon the fluorescence properties of the sample, and recorded with the custom fluorescence microscope. PicoQuant Picoharp software was used to record TCSPC decay curves and Fluofit was used to fit the collected data. All data were tail-fitted with a mono- or biexponential function to determine the excited state lifetimes. Each imaged area was 100 µm ² using a dwell time of 10 ms per pixel for FLIM imaging using the PicoQuant Symphotime software. A PicoQuant Micro Photon Devices (PDM series) photon detector was used to detect the photon counts from each pixel. Images were constructed as 100 × 100 pixels. Data were binned in 10 × 10 pixels to maintain appropriate photon statistics for fitting.

Scanning Electron Microscopy–Energy Dispersive Spectroscopy (SEM-EDS)

Leaves were sprayed with a solution of RuGalvoxite at a concentration of 800 ppm Zn until the leaf surface was saturated. The samples were left to dry overnight and examined the next day. SEM images were acquired using a Hitachi 3000 tabletop microscope, operated at 15 kV using standard imaging mode at 1000x magnification. The EDS spectra were collected using Quantax 70 on the elemental mapping mode at a voltage of 15 kV and 1000x magnification.

Data Analysis

Fluorescence spectra were averaged for each type of sample studied for statistical analysis. The spectra were processed using Matlab software. FLIM images were refitted after collection of the on-the-fly fitted data. Data were binned in 10 × 10 pixels to maintain appropriate photon statistics for fitting. TCSPC and FLIM data were tail-fitted using the following formula: ⁵¹

$$\mathrm{I}(\mathrm{t})={\sum }_{\mathrm{i}}{\mathrm{a}}_{\mathrm{i}}{\mathrm{e}}^{\frac{-\mathrm{t}}{\tau }}$$

(1)

where τ is the excited state lifetime of one component and a_i is the amplitude of that excited state lifetime. From these fits multiple excited state lifetimes (τ) can be extracted in the case of a multiexponential decay, as well as the relative amplitudes (a_i) of each component. Lifetimes were calculated by Symphotime for FLIM imaging data and by Fluofit software for TCSPC data, in both cases from decay curves fit with an exponential decay from which average lifetimes were calculated using the formula below:

$${\tau }_{\mathrm{avg}}={\displaystyle \frac{\mathrm{\Sigma }({\mathrm{A}}_{\mathrm{i}}\times {\tau }_{\mathrm{i}}^2)}{\mathrm{\Sigma }({\mathrm{A}}_{\mathrm{i}}\times {\tau }_{\mathrm{i}})}}$$

(2)

where A_i is the amplitude of the lifetime component and τ_i is the respective lifetime value.

Statistical Analysis

Experiments were completed for a minimum of three independent biological replicates. Results were evaluated with student's two-tailed t-test. Statistically significant differences were determined for the P < 0.05 confidence interval.

Results

Inspection Under UV Excitation

For an initial characterization of RuGalvoxite, citrus leaves were sprayed with RuGalvoxite until saturation and dried overnight to study the basic behavior of the treatment on a leaf surface. The leaves were visually inspected under standard lighting as well as under a UV lamp to observe the adherence and spread of RuGalvoxite on the citrus leaf surface. The bottom and the top of a treated leaf are shown in Figure 1 under standard room fluorescent lighting (Figure 1a,c) and UV light (Figure 1b,d). Comparing Figure 1a and Figure 1c it can be seen that RuGalvoxite films are easier to distinguish on the bottom of a leaf under standard room fluorescent lighting, displaying the white color of the zinc borate. From the images, it is also apparent that RuGalvoxite spreads better on the bottom surface of the leaf compared the top side. To visualize the fluorescence of the treatments, the leaves were placed under a 302 nm UV light. Consequently, the orange fluorescence of the encapsulated Rubyp dye in the RuGalvoxite treatment is visible in Figure 1b and Figure 1d. The areas with fluorescence match up with the areas observable under standard room lighting Also, additional areas can be seen under UV light that were too thin for visualization otherwise. The images in Figure 1b and 1d confirm that the bottom of the treated leaf shows a better distribution of RuGalvoxite spray treatment compared to the top of the leaf. For the latter, it appears that the treatment pools near the edges of the leaf.

Figure 1.

Digital photographs of leaves sprayed with RuGalvoxite that were left to dry overnight. Panels (a) and (c) show photographs taken under standard room lighting. Panels (b) and (d) show photographs under UV (302 nm) lighting. The encircled areas demonstrate visible RuGalvoxite residue under standard room lighting. On the other hand, under UV light, RuGalvoxite appears to fluoresce orange as expected for Rubyp. The photographs indicate that RuGalvoxite spreads well on the bottom side of leaves, but not on the top side.

Fluorescence Spectroscopy

The recorded emission spectrum of the leaf samples matches the Photosystem II emission with peak wavelength at 690 nm, as shown in Figure 2a.^37,52 Rubyp yields an emission peak wavelength at 560 nm with a shoulder at 620 nm (Figure 2b). ⁵³ The fluorescence spectrum collected from Rubyp sprayed on a leaf surface results in a superposition of the Rubyp dye and leaf emission spectra. The spectrum obtained from the bottom of the leaf (Figure 2e) has a smaller ratio of Rubyp to leaf emission compared to the top of the leaf (Figure 2f), which could be an effect of the better spreading of spray treatment at the bottom of the leaf versus the pooling of the spray at the top of the leaf. Furthermore, Galvoxite emission was also measured, which presented a very faint emission spectrum with peak around 550 nm (Figure 2c). The low level of fluorescence given off by the Galvoxite is not present in the fluorescence measurements for RuGalvoxite (see Figure 2d) because of the high emission intensity of the leaf and Rubyp dye. Therefore, any fluorescence given off by Galvoxite was completely masked by the other emissions, including the leaf emission as illustrated in Figures 2g and 2h.

Figure 2.

Fluorescence characterization of materials on different substrates when excited at 466 nm. (a) Demonstrates the fluorescence emission of the leaf surface. (b-d) Shows the emission spectra of materials deposited on a glass slide. (e-f) Demonstrates the emission spectra of Rubyp deposited on a leaf surface. The spectrum (e & f) is a superposition of spectra in panels (a) and (b). The ratio of Rubyp emission to leaf emission is indicative of the amount of Rubyp on the area of the leaf we studied. (g-h) Show the emission spectrum of Galvoxite deposited on leaf surface. Since Galvoxite has significantly lower emission compared to the plant's autofluorescence, only leaf fluorescence can be observed.

The fluorescence spectra collected for the treatment (RuGalvoxite) on leaf surface shows a superposition of their spectra as shown in Figure 3. The leaf emission with peak at 690 nm dominates the spectrum, however, the emission of the Rubyp embedded in Galvoxite is clearly visible as well. Similar observations as those described for the control experiments in Figure 2e and Figure 2f (see above) can be made here with respect to the appearance of the Rubyp contribution to the emission spectrum. Again, the spectrum obtained from the top of the leaf (Figure 3a) has a larger ratio of Rubyp to leaf emission compared to the bottom of the leaf (Figure 3b), likely to be caused by the better spreading of spray treatment at the bottom of the leaf versus the pooling of the spray at the top of the leaf. At the top of the leaf more material accumulates in a smaller area, leading to an increase in observed fluorescence intensity of the RuGalvoxite treatment.

Figure 3.

Fluorescence characterization of RuGalvoxite sprayed onto the leaf surface. Both the RuGalvoxite and chlorophyll emission can be observed in the spectra obtained for (a) Leaf top and (b) Leaf bottom. The ratio of RuGalvoxite emission to chlorophyll emission is greater for the spray treatment on top of the leaf when compared to the bottom of the leaf. This effect is likely caused by the pooling of the droplets in smaller areas, leading to more material and higher brightness. This would be consistent with qualitative results in Figure 1 and quantitative data in Figure 2 e-f.

Time-Correlated Single Photon Counting (TCSPC)

The fluorescence spectra obtained for leaves, Rubyp dye, and RuGalvoxite reveal that there is significant overlap between 550 nm and 750 nm, which covers most of the available spectrum. Spectral separation of the Rubyp dye emission from the leaf fluorescence, caused mainly by chlorophyll, to locate the RuGalvoxite spray treatment on the leaf surface is therefore precarious. Due to the limitations of using fluorescence emission to distinguish Rubyp from leaf chlorophyll emission we took advantage of advanced optical analysis, in particular with TCSPC and its imaging form FLIM (see below), to exploit the long excited state lifetime of Rubyp compared to chlorophyll. With this approach, we can temporally separate the chlorophyll (sub 10 nanoseconds excited state lifetime) from the Rubyp dye in RuGalvoxite (several hundred nanoseconds lifetime).^{46–48,54–56} We have previously demonstrated this approach successfully for in planta tracking. ⁴⁵

We first examined the fluorescence decays of the untreated leaf surface, Galvoxite, Rubyp, and RuGalvoxite to obtain baseline data before investigating the spray treatment on the leaf surface. The decay curves for these experiments are shown in the insets in Figure 4. To extract the excited state lifetimes from these data, the curves were fitted with a biexponential decay, except for RuGalvoxite, which could be fitted with a monoexponential decay. We attribute the latter to a dilution effect of Rubyp by Galvoxite, leading to limited aggregation of Rubyp dye and hence a more straightforward monoexponential behavior.^57,58 Fitting parameters for the data shown in Figure 4 are reported in Table S1. The excited state lifetime of the leaves of the citrus plant were found to have an average lifetime of approximately 2 ns.^46–48 Galvoxite has an average lifetime of 9.4 ns, which is also too close to that of the leaf to be reliably distinguishable. Rubyp dye has an average lifetime of 210 ns, which is approximately 100 and 22 times longer than that of the leaf or Galvoxite, respectively. Finally, RuGalvoxite has an average excited lifetime of 541 ns with a monoexponential decay. The longer lifetime and appearance of a monoexponential decay can again be attributed to a dilution effect of Rubyp by Galvoxite, leading to limited aggregation of Rubyp dye, and potential changes in local environment.^54,55 The reduced aggregation removes non-radiative decay pathways that can shorten the excited state lifetime, and the reduction in available pathways results in less complex photophysics as observed in the monoexponential behavior. The large contrast between the RuGalvoxite excited state properties and the leaf excited state properties allows us to easily distinguish the RuGalvoxite spray treatment from any other contributing components in observed fluorescence emissions from our treated samples.

Figure 4.

TCSPC decay curves for RuGalvoxite sprayed on the bottom and top surfaces of the leaf are shown in panel (c). Panels (a) and (b) show data for the controls. Panel (a) on the top left contains the decay curves for Rubyp and RuGalvoxite acquired in the absence of leaf fluorescence, while panel (b) on the top right shows the decay curves for Galvoxite and the leaf surface.

The decay curves for treated leaves (RuGalvoxite sprayed on the leaf surface) are shown in the main panel of Figure 4 for both leaf top and leaf bottom. RuGalvoxite sprayed on the top and bottom of the leaf has a lifetime of 455 and 432 ns, respectively. The emissions, again, exhibit a monoexponential behavior, akin to RuGalvoxite in the absence of leaf surface. We also show representative decay curves with fit parameters for stomatal and glandular trichome regions of untreated and RuGalvoxite treated leaves in Figure S1 and Table S2, respectively. These areas of interest are discussed further below. The results clearly demonstrate that RuGalvoxite emission is distinguishable from leaf background when using TCSPC for optical analysis. This opens up the opportunity to map the localization of RuGalvoxite on leaf surfaces through Fluorescence Lifetime Imaging (FLIM).

Fluorescence Lifetime Imaging Microscopy (FLIM)

Control FLIM images of untreated leaf top and bottom were first taken and are shown in Figure 5a and 5b. Observed excited state lifetimes in the FLIM images are generally below 10 ns. Leaves were sprayed with RuGalvoxite at a concentration of 800 ppm Zn to mimic the farm foliar application rates. Figure 5c and 5d reveal that leaf areas were the RuGalvoxite treatment localized show excited state lifetimes nearing 1 μs in the FLIM images, noting the difference in scale bars between treated and untreated samples. The FLIM contrast in the images is obtained from mono- or biexponential tail fitting using the procedures described in the Methods section. In Figures 5c and 5d the wide FLIM contrast compared to the untreated controls is derived from the substantially longer excited-state lifetime of RuGalvoxite relative to leaf autofluorescence. In terms of photon counts, the dark regions in the images show a typical range of 50 to 90 photon counts per pixel, while brighter areas range from around a hundred to several hundreds photon counts per pixel. As discussed above, the treatment appears to spread better on the bottom of the leaf than on the top surface.

Figure 5.

Fluorescence intensity (grey scale) and FLIM images (color map) of (a, b) leaf controls and (c, d) leaves sprayed with RuGalvoxite at 800 ppm of Zn. The fluorescence images show integrated intensity. For the FLIM images note that the controls in Figures 5a and 5a use the same color scale as the treated samples in Figures 5c and 5d, but the full range of values is drastically different (< 0.01 microsecond lifetime) compared to the treated samples (1.0 microseconds). Compared to the controls, the treated leaves show regions with significantly higher excited state lifetimes, indicating the presence of RuGalvoxite. Panels c and d show areas of RuGalvoxite that have built up on the leaf surface (predominantly green and red colors) and regions of low excited state lifetime where the leaf surface is visible (dark blue). In panels c and d the yellow arrows point to the same sample region in both the fluorescence and FLIM images. In Panel c, these indicate the location of presumed glandular trichomes on the top leaf surface where RuGalvoxite accumulated, while there is limited RuGalvoxite coverage in between these structures due to the waxy nature of the top leaf surface. In panel d, the yellow arrows indicate the location of stomatal pores in fluorescence and FLIM maps for the bottom of the treated leaf. RuGalvoxite is organized near stomata and located within the stomatal pores.

Interestingly, RuGalvoxite at the top of the leaf tends to accumulate at specific structures on the leaf surface, which we speculate are glandular trichomes. At the bottom of the leaf RuGalvoxite is found in between stomata and inside stomatal pores. This suggests Galvoxite has a greater affinity to not only accumulate near stomata at the bottom leaf surface as previously reported, but also to localize at the stomatal pores themselves. We speculate that Galvoxite is attracted to the rhamnogalacturonan II and homogalacturonan polysaccharides found in pectin.^33,59 These polysaccharides are found in the cell walls (guard cells) and are also found in areas surrounding stomata and trichomes.

Scanning Electron Microscopy (SEM)

For SEM, leaves were washed in a 1% detergent solution in water to remove any residual fertilizer or treatments used by the nursery on the plants. The stomata at the bottom of the leaf can be seen in SEM images as the dark circular areas (Figure S2). The RuGalvoxite film applied at a spray rate of 800 ppm of Zn on leaf surface is easily observed in SEM, as depicted in Figure 6. The film morphology consists of domains of RuGalvoxite with obvious domain boundaries in between them. Similar to earlier notions, the material is more densely clumped together on the top of the leaf compared to the bottom of the leaf. While at the top of the leaf (Figure 6a) no stomata are visible in the field of view, they are abundant at the bottom of the leaf (Figure 6b). For the latter, the RuGalvoxite near stomata forms domains that encircle the stomata while leaving several tens of micrometers of distance from the region where the stoma are located. We speculate that the interactions between the guard cells around the stoma and RuGalvoxite are more favorable for deposition of the material compared to the cells adjacent to the stomata. Although this is not well understood at this time, we propose that the content of rhamnogalacturonan II and homogalacturonan polysaccharides around the guard cells plays a role. These observations are consistent with those made from the FLIM images, where RuGalvoxite was found close to and within the stoma, but with a gap between the domains nearest the stomata (Figures 5c and 5d). The RuGalvoxite films observed in SEM give an indication of how Galvoxite can spread across the leaf surface, which opens up the possibility to use it as a physical barrier that protects the plant from ACP puncturing the leaf into the phloem with potential infection with citrus greening. This will require further engineering of Galvoxite to improve its spread on the waxy top of the leaf in the future.

Figure 6.

EDS maps of 800 ppm of Zn for RuGalvoxite on leaf with corresponding SEM image for (a) top and (b) bottom surfaces of leaf. From the micrographs the light gray is identified as RuGalvoxite, which appears to deposit around the stomata on the bottom side of the leaf (b). The EDS maps for Ru, Zn, and Cu are colored in red, green, and blue, respectively. The blue scale bar represents 100 µm.

We also completed EDS on treated leaves to localize Zn contained in Galvoxite and Ru contained in the Rubyp dye. The elemental mapping is shown in Figure 6a for the top of the leaf and Figure 6b for the bottom of the leaf. As expected, RuGalvoxite displays an increased signal in Zn on the treated leaf surface compared to the controls (see Figure S2). In the Zn map, the breaks between RuGalvoxite domains are not obvious. This might imply that in these areas, there still was a thin film of RuGalvoxite present. Overall, there is a good qualitative correlation between areas where RuGalvoxite can clearly be seen in the SEM image and the Zn map, supporting successful deposition of the formulation on the leaf surface. For Ru, the EDS map appears as slightly elevated with respect to the untreated control study (see Figure S2) in the EDS map but lower than directly applying Rubyp dye directly to the leaves (see Figure S3), which makes sense considering that Rubyp is diluted into Galvoxite for RuGalvoxite treatment. The EDS spectra in Figure S4 clearly show a Zn peak at around 1 keV for the samples that were treated with Galvoxite and RuGalvoxite. There is no prominent Ru peak for RuGalvoxite, which would be expected around 2.6 keV, due to the low amount of Rubyp dye in the RuGalvoxite. While the Ru maps in Figure 6 present a slightly elevated signal compared to the control samples and the presence of unapplied trace metals, such as the Cu map (Figure 6), the Ru-associated signal is nonetheless weak. This correlates with Figure S4 where we note the absence of a prominent Ru peak for RuGalvoxite as an indication that the Ru concentration is substantially diluted within the formulation relative to the Rubyp dye-only control, thereby limiting the strength of the EDS spectral signal for the treated samples. We therefore caution against overinterpretation of the Ru EDS data.

Discussion

Citrus agriculture has been severely affected by HLB, a systemic bacterial crop disease vectored by the Asian citrus psyllid (Diaphorina citri Kuwayama). Moreover, this disease causes nutrient deficiency which can affect the tree's response to biotic stress, such as citrus canker. In this study, we found that Galvoxite, a novel crop protection formulation based on a zinc (Zn)-borate (B) composition, has film forming capabilities on citrus leaf surfaces and is able to localize in stomatal pores when applied as a foliar spray. The films have the potential to act as a physical barrier between vector and host. Basic visual inspection of the films (Figure 1) shows that the bottom of the leaf is more accommodating to the RuGalvoxite spray treatment. On the waxy side of the leaf, Galvoxite tends to pool in concentrated areas. This pooling causes an increase in the observed fluorescence intensity at the site of measurement due to the presence of more Rubyp dye. In Figure 3a the spectrum obtained from the top of the leaf has a larger ratio of Rubyp to leaf emission compared to the bottom of the leaf (Figure 3b). The SEM images in Figure 6 give a clear picture of the film morphology as it appears on the leaf surface. The elemental mapping of Zn gives further support to the successful deposition of the formulation.

Tracking foliar or systemic treatment is challenging due to the chlorophyll fluorescence that covers most of the visible spectrum. ⁴⁹ This emission interferes with that of fluorescent markers that could be used to localize or track plant treatments. In our work, the negligible fluorescence of Galvoxite is a key reason for adding a fluorescent dye to enable optical tracking of Galvoxite on leaf surfaces. Due to the difficulty of spectrally separating marker emission from the chlorophyll background we used Rubyp as a fluorescent marker due to its long excited state lifetime compared to traditional organic chromophores. TCSPC and FLIM were used as advanced spectroscopy and imaging methods to differentiate these emissions. The TCSPC data demonstrate that the Rubyp fluorescent marker is clearly distinguishable from the chlorophyll background by its long excited state lifetime (Figures 4a,b). FLIM is a combination of TCSPC and mapping, where each pixel of a FLIM image has its own TCSPC decay curve collected, and the pixel value represents the average lifetime. ⁶⁰ We observed that RuGalvoxite has affinity for the local environment at the bottom of the leaf, particularly near stomata, which possess a higher content of rhamnogalacturonan II and homogalacturonan polysaccharides, instead of the waxy nature of the top of the leaf. This was confirmed with SEM analysis (Figure 6) based on the images and the Zn EDS maps. Our findings suggest that Galvoxite has affinity for these polysaccharide moieties found in pectin.^33,59

Based on the strong FLIM Rubyp dye lifetime contrast we also found that RuGalvoxite accumulates within stomatal pores (Figure 5d). This is important for the potential application of Galvoxite as a delivery vehicle for systemic and localized plant treatments in disease scenarios such as citrus greening and canker. For instance, inorganic copper biocides are commonly used for managing citrus canker, which is caused by XCC colonizing the stomata. ³⁰ Considering Rubyp as a model cargo, Galvoxite could then improve the performance of cargoes, such as repellents or copper-based biocides, by localizing them around and inside the pathogen entry point. Similar assumptions can be constructed for the delivery of micronutrient chelates into the plant tissue, to prevent their immobilization on the leaf cuticle and enhance their absorption. This is where the strength of advanced optical spectroscopy in the detection and mapping of a diluted cargo against a complex background comes into focus. If SEM analysis had been used independently to map the Rubyp dye content of Galvoxite the results would have been inconclusive due to limitations on its sensitivity in elemental mapping. TCPSC and FLIM as optical fluorescence/phosphorescence based methods can detect extremely low fluorophore concentrations, representing a substantial sensitivity mismatch with EDS mapping, while in addition the lifetime contrast can provide selective discrimination against a complex background. In the application demonstrated here, the leaf autofluorescence occurs around 2 ns, while the Rubyp dye lifetime is roughly 0.5 µs. In this scenario TCSPC/FLIM dramatically outperforms elemental mapping for trace detection.

Conclusion

Lifetime-resolved measurements enabled clear differentiation of treated and untreated regions through the large disparity in excited-state decay behavior between the leaf autofluorescence background and the long-lived Ru-associated emission. Our results support the film-forming capabilities of Galvoxite, which may contribute to the formation of a protective surface barrier. The observed differences between the upper and lower leaf surfaces further indicate that leaf microstructure strongly influences formulation deposition and localization behavior. Preferential accumulation near stomatal regions and glandular trichomes suggests that Galvoxite films interact non-uniformly across the heterogeneous leaf environment, potentially affecting transport and retention pathways relevant to foliar delivery applications. These observations suggest potential utility of Galvoxite for delivering agrochemicals such as insecticides or repellents for enhanced protection against ACP. Overall, these findings establish fluorescence lifetime imaging as a sensitive tool for evaluating foliar formulations on plant surfaces and provide insight into how formulation morphology and local leaf microenvironments influence the spatial behavior of crop protection materials.

Supplemental Material

sj-docx-1-app-10.1177_27551857261465003 - Supplemental material for Tracking Galvoxite on Leaf Surface with Fluorescence Lifetime Imaging: Film Forming and Delivery Capabilities Towards Citrus Crop Protection

Supplemental material, sj-docx-1-app-10.1177_27551857261465003 for Tracking Galvoxite on Leaf Surface with Fluorescence Lifetime Imaging: Film Forming and Delivery Capabilities Towards Citrus Crop Protection by Carlos Flores, Jorge Pereira, Anthony Congelosi, Peter Wallace, Sydney McKenna, Swadeshmukul Santra and Andre J. Gesquiere in Applied Spectroscopy Practica