Plant tissues contain many fluorescent compounds [1
]. This natural fluorescence is called autofluorescence and can be used for imaging cells and tissues. Where fluorophores exist in spatial isolation, it is possible to identify the fluorophore by histochemistry, by comparison with purified standards, or potentially by using other forms of spectroscopy such as FTIR or Raman microspectroscopy [1
]. Autofluorescence can be used to image tissues even when the fluorophore has not been identified with the advantage of high contrast and spatial resolution [4
]. In some cases, autofluorescence may either complement or interfere with staining or labeling protocols, especially when the fluorescence is relatively bright as is the case with chlorophyll [6
]. As more plant fluorophores are characterized, autofluorescence will become an increasingly important imaging tool in plant science.
Fluorescence has the advantage of high spatial resolution for imaging applications, but fluorescence spectra have limited direct linkage to chemical characteristics. Fluorescence spectra have a fixed emission peak and intensity changes as the excitation wavelength is varied in relation to the absorption spectrum of the fluorophore [8
]. In terms of using autofluorescence as a histological tool for the microscopic characterization of tissues, the most practical approach is to use sequential excitation with a range of wavelengths over the ultraviolet (UV) and visible light spectrum. Since most modern fluorescence microscopy is performed with confocal microscopes and laser-based illumination, the optimum excitation wavelength and emission range need to be determined for a particular fluorophore in relation to the available range of laser emission lines. Although lasers with continuously variable emission are now available, there are few studies where the excitation/emission characteristics of a plant fluorophore have been fully characterized in situ [9
Autofluorescence can be characterized using three parameters: emission spectrum, intensity (quantum yield), and fluorescence lifetime [8
]. Fluorescence lifetime spectroscopy is potentially very useful in plant histology but, as of yet, its use has been rather limited [11
]. There are two main examples of fluorescence spectroscopy being used in plant science involving studies of chlorophyll or lignin. Chlorophyll fluorescence can readily be measured using handheld devices on living leaves and this measurement is routinely used in physiological studies [13
]. Lignin fluorescence has been extensively studied because of its importance in wood science. In microscopic applications, lignin autofluorescence has been used in studies of reaction wood formation [14
], cell wall porosity [16
], and plant biomass deconstruction [17
]. Lignin is characteristic of xylem tissue occurring as a complex polymer containing a variety of fluorophores [18
]. Its fluorescence is generally considered to be the combined emission of individual chemical structures, as these never occur in spatial isolation at the micro- or nanoscales. There are two types of fluorophore associated with lignin: phenolic structures that are excited by UV light and have blue emission, and conjugated structures excited by visible light with broad emission across the visible light spectrum into the red and far-red wavelengths [14
]. Lignin fluorescence is characterized by its broad range of excitation and emission wavelengths, and while relatively weak, it is certainly sufficient for imaging of cell walls in wood and other tissues [15
Discrimination of fluorophores in plant tissues can be achieved in two ways. First, by sequentially imaging tissue with multiple excitation wavelengths, different fluorophores may exhibit emission wavelengths with different relative intensities and hence may appear as different colors in fluorescence images. In other cases, fluorophores may have strongly overlapping emissions that require spectral imaging and linear unmixing to achieve separation [19
]. Some fluorophores have a pH-dependent emission that can provide discrimination [19
In the present study, we examine the autofluorescence of healthy pine needles compared to chlorotic and necrotic needles. We provide a characterization of cell wall components including lignin, suberin, ferulate, and flavonoids, extractives such as oleoresin and terpenes, and chlorophyll, including a comparison between histochemistry and autofluorescence where possible.
Our study was aimed at using autofluorescence to characterize needle histology as a potential screening tool for possible resistance to red needle cast disease in radiata pine caused by Phytophthora pluvialis
]. The advantage of using autofluorescence is to avoid the need for multiple staining protocols resulting in a single image that can be used to evaluate differences among tree genotypes or to study disease progression. Using sequential UV, blue and red excitation, we were able to image at least six distinct fluorophores in fresh vibratome sections of needles. Further information was obtained using fixed needles where some fluorophores changed their emission spectrum or became redistributed (Figure 8
). The disadvantage of imaging fresh needles is the need to complete imaging within a few days of collection.
Relatively few studies have characterized plant autofluorescence [1
]. Lignin is the only cell wall-associated fluorophore that has been investigated in detail [19
], although some limited information is available for hydroxycinnamic acids such as coumarate and ferulate [20
], as well as suberin [33
In pine needles, xylem and endodermis are highly lignified and show fluorescence identical to lignin in the secondary xylem of the stem [19
]. Other tissues are less well lignified, the hypodermis has quite variable lignification, while the epidermis and transfer tissue are weakly lignified. Compared to phloroglucinol staining, autofluorescence gives comparable results but seems more sensitive to low lignin levels in the transfusion tissue, for example. Lignin is separated from other cell wall fluorophores by its characteristically wide emission range. The fluorescence spectrum of lignin has a λmax of 455 nm with UV excitation and 535 nm with blue excitation corresponding to emission from phenolic and conjugated structures respectively [14
]. Lignification of the xylem, endodermis, and hypodermis showed no obvious variation among different needles. Other studies have found differences in degree of lignification of the endodermis between different species using both lignin autofluorescence and phloroglucinol staining [34
Transfusion tissue contains a number of cell types including transfusion tracheids and parenchyma cells with variable cell wall composition. The transfusion tissue acts as a pathway for water movement from the xylem into the mesophyll spaces and the outside environment via the stomata. Despite this important physiological function, relatively little is known about its composition. Autofluorescence demonstrates that transfusion tracheids are very weakly lignified compared to xylem tracheids. Some transfusion parenchyma cell walls contain suberin, which presumably acts to restrict water loss from the xylem into the surrounding needle tissue. Some cell walls within this tissue may contain both lignin and suberin, although this is difficult to determine by autofluorescence alone.
The highly lignified endodermis may also restrict water movement into the mesophyll, forming a Casparian strip, but it also represents a lignified barrier to invasion of the vascular tissue by pathogens [35
]. We did not find evidence for suberin in the endodermis of radiata pine [4
]. Interestingly, the resin canal epithelium in needle resin canals appears to lack the suberization found in resin canals in secondary xylem [33
]. Suberin is much easier to detect by autofluorescence than by Sudan IV staining, which tends to be weak in conifer tissues, especially in comparison to cutin, which reacts strongly with this stain (Figure 1
Ferulates have been detected esterified to primary cell walls in all families of gymnosperms [37
]. In pine needles, cell wall bound ferulates are detected in phloem cell walls. This fluorophore varies considerably among needles from strong to very weak emission and may even be absent in some needles, although the reasons for this variability are not immediately apparent. Ferulate esters, like other phenylpropanoids, can be regulated as part of stress responses, so the variability we observed may reflect the stress states of needles [38
Flavonoid fluorescence did not show obvious variability among needles, but there was a prominent variation between the abaxial (curved) and adaxial surfaces of the needle with significantly less flavonoid fluorescence on the adaxial surfaces in most needles examined. This might have a significant effect on interactions between pathogens and the needle surface.
The distribution of phenolic compounds in conifer needles has been studied in some detail [2
]. Phenolic fluorophores may be bound to cell walls [26
]. In spruce needles, blue excited autofluorescent flavonoids have been localized in the lumens of needle epidermal cells, where they were described as diacylated flavonol glucosides after treating with NA [2
]. In Scots pine, analysis of enzymatically isolated needle epidermis also detected diacylated flavonol glucosides, representing 90% of the total needle content [25
]. Flavonoids act as UV absorbers and antioxidants protecting needles from damage due to sunlight exposure [39
]. Similar fluorescence was observed in pine needles in the present study with flavonoids localized in the cuticle, epidermal cell walls, and lumens, as well as in hypodermal primary cell walls. It seems that epidermal and hypodermal cell walls have a complex composition with indications of lignin, lipids, and flavonoids being present.
Terpenes and oleoresin are part of the needles defense system, so being able to detect their distribution and amount is of particular interest from a pathology perspective [36
]. However, differentiation of extractive components using autofluorescence is limited. Oleoresin in resin ducts is autofluorescent, terpenes such as pinene are very weakly fluorescent while isoprenoids in resin canals, endodermis, and transfusion tissue are non-fluorescent.
Mesophyll cells contain large amounts of faint blue fluorescent material. Based on autofluorescence and histochemistry, there were no clear indications of what this material might be. We were able to exclude oleoresin, lipids, and isoprenoids. Necrotic needles contain dense deposits of green fluorescent material in the mesophyll, which is presumably derived from the blue fluorescent material present in healthy and chlorotic needles. The green fluorophore in necrotic needles has a slightly different fluorescence spectrum from similar deposits in FAA fixed needles and is substantially different from tannin deposits in pine bark (Figure S1
). Based on this comparison and the histochemistry, we were not able to determine the exact nature of mesophyll extractives other than to say they are probably polyphenolic. Others have described the formation of tannins in physiologically stressed needles, but we could not confirm that these mesophyll deposits are tannins [39
The comparative anatomy of pine needles has been studied in detail [43
], but there have been few comprehensive histological studies of conifer needles. Investigations have focused on pathological or physiological effects such as fungal infection, acid rain, mineral deficiency, or CO2
]. Autofluorescence in combination with histochemistry has been used to explain differences in the mechanical properties of pine needles between species due to variable lignification of the endodermis [34
]. Chlorophyll autofluorescence has been used to detect lesion formation in Dothistroma
needle blight of radiata pine [47
]. We demonstrate the usefulness of autofluorescence in understanding cell wall composition in needle tissues, especially in relation to potential pathological impacts. Autofluorescence, spectral imaging, and fluorescence lifetime imaging [9
] will play an increasing role in understanding plant tissue composition.
4. Materials and Methods
Young three-year-old Pinus radiata
D. Don clonal trees growing in pots outdoors at the Scion campus in Rotorua, New Zealand, were examined to determine needle histology comparing healthy needles with chlorotic and necrotic needles resulting from natural senescence. Mature needle fascicles from 10 different trees were collected sequentially and maintained in a fresh condition by immersing the base of the fascicle in water in plastic tubes stored at 4 °C for no more than 5 days before discarding. Fresh needles and needles fixed in FAA (90:5:5 absolute ethanol, glacial acetic acid, formaldehyde) for at least 6 weeks were sectioned transversely at a thickness of 30 μm using a Leica vibratome. Sections were collected from the basal 1 cm of the needle and were stored briefly in water before mounting in 50% glycerol in phosphate buffer at pH9 [19
]. Comparisons were also made by mounting in glycerol–buffer at pH 5 or 7. In some cases, sections were treated with or mounted in 0.1 M ammonium hydroxide to enhance localization of cell wall-bound ferulates [20
]. To detect fluorescence of flavonoids, fresh sections were mounted in NA reagent [25
], and a comparison was made under identical conditions with similar sections in glycerol buffer. Treatment with NA results in a significant brightening of flavonoid fluorescence where present. Sections were examined using a Leica SP5 II confocal fluorescence microscope using sequential excitation at 355, 488, and 633 nm, with respective fluorescence emission at 400–500 (blue), 500–570 (green), and 650–750 (red) nm. In some cases, imaging was performed with 488 and 561 nm excitation, with respective emission at 500–570 and 570–700 nm. A UV corrected 20× immersion objective was used with 80% glycerol in water as the lens immersion medium. Individual images rendered as maximum intensity projections were montaged to provide a high-resolution view of the whole needle cross-section using Microsoft Image Composite Editor software.
For spectral imaging, fresh un-fixed sections were excited at 355 nm and spectral image sequences were acquired from 400–800 nm using a slit width of 10 nm and a scan interval of 5 nm. Some spectra were acquired from fixed sections using 488 nm excitation with emission from 500–800 nm. Regions of interest corresponding to cuticle, hypodermis, mesophyll, phloem, and xylem from spectral image sequences were converted into spectra using Digital Optics V++ software (Auckland, New Zealand) using an image threshold based on the average brightness [48
]. This procedure was intended to preserve the fluorophore brightness relative to other tissues by excluding non-fluorescent cell lumens from intensity measurements. Intensity comparisons were made relative to lignin fluorescence in the xylem of the vascular bundle from the same image sequence. Spectral measurements were performed using three replicates which were found to be highly repeatable. Differences between spectra were determined using a
= non-significant, * = p
< 0.05, ** p
< 0.01, *** p
< 0.001). Spectral unmixing was performed using the Poisson NMR application in ImageJ software (Bethesda, MD, USA) [50
Some fresh needle sections were treated with phloroglucinol to localize lignin. A 2% solution of phloroglucinol in 95% ethanol was freshly prepared and stored in the dark at 4 °C. Immediately prior to use, several drops of concentrated HCl were added to 0.5 mL of phloroglucinol, and sections were treated for 2 min followed by washing in 95% ethanol. Sections were mounted in 95% ethanol for microscopy [51
]. Sudan IV was prepared as a saturated solution in 95% ethanol. Sections were stained for 1 h before mounting in glycerol–buffer at pH7 [52
]. NADI stain was used to identify isoprenoids and lipids in fresh sections by staining for 1h at room temperature. NADI reagent selectively stains isoprenoids purple and lipids blue. The reagent was prepared by mixing 0.5 mL of a 1% α-naphtol solution in 40% alcohol, 0.5 mL of 1% dimethyl-p-phenylenediamine chloride in water, and 49 mL of phosphate buffer 0.05 M (pH 7.2) [28
]. The nitrosylation test for catechols was performed by sequentially treating sections of fresh healthy needles with 1 volume of 10% sodium nitrite, 1 volume of 20% urea, 1 volume of 10% acetic acid, and 2 volumes of 2 M sodium hydroxide. In the presence of catechols, this reagent forms a red color which later turns brown after 1 h [54
]. Sections of necrotic needles and healthy FAA fixed needles were treated with a 20% w/v
solution of vanillin in ethanol for 3 min followed by concentrated HCl to test for the presence of condensed tannins, which produce a red coloration [55