Visualisation of calcium oxalate crystal macropatterns in plant leaves using an improved fast preparation method

Leaves of the majority of plants contain calcium oxalate (CaOx) crystals or druses which often occur in spectacular distribution patterns. Numerous studies on CaOx in plant tissues across many different plant groups have been published, since it can be visualised readily under a light microscope (LM). However, there is surprisingly limited knowledge on the actual, precise distribution of CaOx in the leaves of quite ordinary plants such as common native and exotic trees. Traditional sample preparation for the documentation of the distribution of CaOx crystals in a given sample – including overall distribution – requires time‐consuming clearing procedures. Here we present a refined fast preparation method to visualise the overall CaOx complement in a sample: The plant material is ashed and the ash viewed under the polarising microscope. This is a rapid method which overcomes many shortcomings of other methods and permits the visualisation of the entire CaOx content in most leaf samples. Pros and cons in comparison with the conventional clearing technique are discussed. Further aspects for CaOx investigations by micro‐CT and scanning electron microscopy are discussed.


INTRODUCTION
Calcium oxalate (CaOx) crystals and druses (crystal aggregates) are very striking, but still quite enigmatic structures in higher plants. They are visible under the light microscope, ideally the polarising microscope, but for an investigation of the morphological details a scanning electron microscope (SEM) is usually required. [1][2][3][4] Apart from the shape and size of the individual CaOx crystals and clusters, they also show characteristic distribution patterns in plant tissues (Figures 1 and 2). interpretations, it has also been convincingly argued that CaOx crystals are often simply inert deposits of excess Ca. 11 The uncertainty regarding functionality may be part of the reason, why there are relatively few critical studies providing an overview on CaOx distribution patterns in plant tissues. Larger studies were performed mainly by a limited number of botanists specialised on biomineralisation and often focused on certain plant groups only. 1 CaOx crystals in leaves can be visualised well in transparent samples, either by light microscopy (LM; Figure 2A) or by microcomputed tomography (micro-CT) with X-rays ( Figure 2B). Scanning electron microscopy (SEM) as a surface-imaging technique is able to visualise the CaOx in great detail on exposed layers and is additionally able to provide EDX element analyses ( Figure 2C). Remnants of CaOx crystals may also be visible in the ash of incinerated leaves ( Figure 2D).
The most established preparation technique for LM examinations uses clearing of the leaves with bleach, such as sodium hypochlorite, and immersion in xylene or a different mounting medium to provide transparency for polarisation microscopy (Pol-LM) ( Figure 2A). 3,12 If samples are too thick and transparency cannot be achieved, then leaves may have to be dissected or cleaved. A complication arises from the fact that cellulose is not effectively cleared which may then interfere with the visualisation of CaOx.
In our recent study of patterns of granular structures in fossil leaves, resembling CaOx druses in related fresh leaves, we identified the necessity to get a reference image database for better comparison of structures in fossil and fresh leaves. 13 Thus, we tried to improve on existing methods. We had previously often used incineration as a quick test to look for biomineral structures in plants: simply, we incinerated the sample and examined the ash with a stereo microscope, and we found that cystoliths, CaOx druses and mineralised trichomes usually were clearly visible within amorphous ash remnants from cell walls. It is not new that biomineral structures such as silica phytoliths and CaOx crystals withstand incineration and can be analysed in ash samples. 14,15 Controlled incineration leads to carbonisation of the leaf sample which retains its basic shape and structure and the biominerals are preserved in their original shape and location. We initially examined incinerated leaf samples under a LM with surface illumination. Ashed leaves that remained too intransparent were separated into an upper and lower part with the help of adhesive tape, so that the inside (mesophyll) of the leaf could be viewed ( Figure 2D). 13 However, the results were not entirely satisfactory and further improvements were necessary for satisfying results. These were obtained by viewing and imaging ashed leaves under the polarising microscope. Here we present some of these results in comparison with those of the established clearing methods, and we discuss advantages and disadvantages of both techniques. Additional examinations and analyses with microcomputer tomography (micro-CT) and SEM will be mentioned, providing complementary information on localisation and composition of biomineral structures in leaves.

Plant material
The new preparation method was successfully used for leaf samples from numerous trees and shrubs from the Botanical Gardens, University of Bonn, and from other locations. Fresh, fully developed leaves of different ages as well as air-dried leaves were prepared successfully.

Microscopy equipment
Light microscopy (LM) was carried out with a standard microscope (Müller Optronic, Erfurt, Germany) with an additional polarisation filter attachment. Scanning electron microscopy (SEM) was performed with a LEO 1450 SEM (Cambridge Instruments, Cambridge, UK), and a Tescan Vega 4 SEM (Tescan GmbH; www.tescan.com), both equipped with secondary electron (SE) and backscattered electron (BSE) detectors and EDX element analysis systems. X-ray images and microcomputed tomography (micro-CT) scans of dry leaves were obtained with a SkyScan 1272 Micro-CT system (Bruker microCT, Kontich, Belgium) in the Institute of Evolutionary Biology and Ecology at the University of Bonn.

Standard preparation of fresh leaves for SEM and micro-CT
Pieces of fresh leaves were fixed in 70% v/v ethanol + 4% v/v formaldehyde in water for at least 20 h and dehydrated with ethanol. For freeze-fracturing, ethanolinfiltrated samples were immersed in liquid nitrogen and broken randomly. After unfreezing, all samples were critical point-dried (CPD 020, Balzers Union, Liechtenstein) and mounted on sample holders for SEM or micro-CT. SEM samples were sputter-coated with an approximately 10 nm thin layer of palladium.

2.3.2
Leaf clearing with sodium hypochlorite-based household bleach Fresh leaf samples were immersed in boiling ethanol for 30 sec in order to dissolve epicuticular wax and improve permeability of the cuticle, followed by immersion in distilled water for 1 h. Then samples were transferred to household bleach (DanKlorix, www.colgate.com; 2.8% NaClO) until they appeared colourless, usually 2 to 12 h, depending on leaf permeability. Then the samples were rinsed with water, followed by dehydration with ethanol and acetone. After immersion in xylene or immersion oil, the samples were ready for Pol-LM.

Ashing
Pieces of fresh or dry leaves were heated to approximately 600 • C to 750 • C until all organic components were oxidised and the remains were almost white. We used two incineration methods: either a temperature-controlled oven (Brennofen Uhlig U15, Efco GmbH, Rohrbach, Germany). Leaf pieces were placed between glass slides, with a spacer that ensured entrance of air ( Figure 3A-C). At 600 • C−650 • C, the oxidation was accomplished after 5 to 15 min. Much faster was incineration in a gas flame: leaf samples were placed between two nickel meshes (screen printing meshes 'Rotamesh'; kindly provided by the print-ing company Frintrup, Bonn, Germany), and burnt in a low-intensity gas flame at red heat for 2 to 3 min ( Figure 3D-F). The remaining piece of ash is quite fragile. For polarisation-LM, the ash piece was immersed in a thin layer of immersion oil on a microscopy glass slide, or an embedding medium with higher viscosity. A drop of the liquid was dispersed on the glass slide over ca. 4 cm 2 , and the ash piece was dropped into it. A coverslip was omitted to avoid fragmentation of the fragile specimens.

RESULTS
We used the incineration method successfully for leaves of more than 100 species, mostly mature leaves of dicotyledonous trees and shrubs. In most cases, the remaining piece of ash was quite fragile, but stable enough to withstand the immersion in oil without heavy disaggregation. The ash consisted of amorphous calcium compounds as remnants of cell walls, sometimes silica, and the compact incinerated remnants of druses and crystals, which retained their original shape. Under the polarisation-LM between crossed pol filters, amorphous ash and silica appeared almost invisible (dark) whereas the incinerated CaOx druses and crystals appeared bright. Any other structures, such as residual carbon remnants and even air bubbles were invisible, so that the CaOx remnants could be seen with excellent contrast and without interference from any other structures. The method reached its technical limitations when the ashed samples became intransparent as a result of melting processes, for example, in presence of phosphates, or when the ashed tissue was so thin and fragile, that it disaggregated during immersion in oil. The incineration in the oven took approximately 20 min. At ca. 400 • C the samples became black; at 600 • C to 650 • C it took usually 10 to 15 min until the samples became white and all carbon was oxidised. Some samples rich in Si or P (e.g., Moraceae, Urticaceae leaves) required up to 20 min at 700 • C to turning white. The enclosure between glass slides prevented ignition of the samples so that the oxidation proceeded under controlled conditions. It was found useful to interrupt the heating after reaching ca. 450 • C and verify that the sample does not stick to the glass slide, so that the ash finally could be transferred easily into the immersion oil. Burning in a gas flame takes usually 2-3 min for complete oxidation and the overall results were similar to the results of oven heating. The risk of fragmentation during immersion in oil was a slightly increased, so that visualisation of large areas proofed to be problematic. The entire preparation took no more than 5 min from cutting the leaf to the visualisation under the microscope. On the other hand, the bleaching procedure took between 3 and 10 h if the leaves had been pretreated with hot ethanol in order to dissolve wax layers; longer without previous pretreatment. After washing in water, dehydration, and immersion in xylol or immersion oil, the samples were transparent and CaOx crystals and druses were clearly visible in the Pol-LM. The cleared specimens were stable, and the positions of the CaOx particles and leaf veins could be visualised precisely. In stereoscopic viewing, their spatial distribution was clearly visible. Figure 4A-D shows a comparison of Quercus and Parrotia leaves prepared by ashing and by clearing. The low magnification images of Quercus (4× objective) show the leaf vasculature and numerous druses in the areoles. Small CaOx crystals are present in the bundle sheaths and can be seen clearly at higher magnification (10× objective; Figure 4C and D). In the incinerated samples ( Figure 4A and C) only the CaOx appears bright, so that the individual particles can be seen clearly. Stereoscopic views revealed that the height positions of the particles were still preserved in sufficiently stable ash pieces. In the cleared sample ( Figure 4B and D), the crystals and also the cell wall structures of the veins appear bright, so that the crystals in the bundle sheaths are obscured and only partly visible. Pol-LM images with crossed polaris-ers show only polarisation-active structures. Amorphous ash, silica, and carbon remnants remained invisible (dark), but they became visible in brightfield imaging mode, simply by removing one polarising filter. Figure 4E and F shows the same area of an incinerated Pterocarya leaf in both modes: Pol-LM shows numerous druses in the parenchyma and few small crystals along a vein. The corresponding brightfield image shows the druses, but also the pattern of medium and small veins with remnants of amorphous ash and traces of carbon. Only larger veins are decorated with CaOx crystals.
Both preparation methods, clearing and ashing, have advantages and shortcomings and should be seen as complementary techniques. The ashing method had been developed as a quick test, but has the potential to help in generating excellent images, not least by eliminating the organic background signal. The clearing method gives best structural preservation and stable specimens, which can be used subsequently, for example, for SEM examinations and analyses. Thick leaf tissue may result in technical limitations for both methods, but to different extent. Incinerated samples have a higher transparency than cleared ones, so that thicker leaves can be examined. Figure 5A shows a Pol-LM image of the ash of a Ceanothus leaf with a strong background signal. On the other hand, leaves, which produce too little ash, may be unsuitable for ashing, because the remnants disintegrate or collapse ( Figure 5B).
In many young leaves with low cellulose content, the background signal in Pol-LM of chemically cleared samples may be negligible, so that perfect images are possible. However, several different cellulose structures can be present, for example, at stomata or trichomes, and may cause misinterpretations ( Figure 5C). In some cases, we observed a chemical preparation artefact during clearing. Clouds of small crystals occurred irregularly in some samples, most likely CaOx, but possibly calcium carbonate, which were never observed in incinerated samples (Juglans regia; Figure 5D). By contrast, incineration caused sometimes alterations in samples with a high content of alkali (K, Na) salts and Si or phosphate, which form melting compounds (Banksia serrata; Figure 5E). This may be avoided by a pretreatment, which removes these soluble components, for example, fixation in formaldehyde.
As demonstrated above, the two-dimensional distribution of CaOx crystals and druses in plant leaves and their morphology can be studied successfully by light microscopy, particularly Pol-LM. For a more comprehensive understanding, additional microscopic techniques such as SEM and micro-CT are required. SEM, with its high resolution and depth of field, reveals minute details and provides impressive images from sample surfaces. EDX element analyses are a valuable feature for the chemical identification of biominerals. Druse-like structures are not always CaOx; Silica, calcium carbonate, calcium sulphate and other salts may display similar morphology. Raman spectroscopy can discriminate between different anions such as carbonate, oxalate or phosphate. Micro-CT is additionally valuable for the determination of the 3-D distribution of biominerals, but it has limitations. The resolution is sufficient for CaOx particles of 10 µm or more, but we could not visualise biominerals of smaller dimensions, such as those found, for example, in conifers. Micro-CT is applicable for correlative studies of chemically cleared Pol-LM samples, after suitable preparation including CP drying.

DISCUSSION
The present article compares different methods to visualise CaOx crystals especially in leaf tissue -traditional bleaching and clearing for Pol-LM, SEM, incineration and micro-CT. Neither of these approaches provides a magical bullet for the specific task -each has its advantages and drawbacks. From the point of view of simply obtaining a comprehensive overview over the CaOx in a leaf sample, the incineration method described here has the clear advantage of providing an unobstructed view of the crystals in their original shape and largely their original position. The crystals are not obscured by organic matter (esp. cellulose) under polarised light and can therefore be imaged precisely. This advantage is particularly striking in the extensive biomineralisation associated with, for example, leaf veins -which may be completely obscured by sclerenchymatic and xylem cells in cleared samples. The TA B L E 1 A Comparison of the advantages and limitations of the incineration and clearing methods for viewing CaOx in leaves (more advantageous method bold type).

Incineration Clearing
Sample preparation time incineration method is highly suitable for mature leaves of normal thickness, including those that are too thick for clearing to provide satisfactory results. Both cleared and incinerated leaves can be used for subsequent SEM-studies with compositional contrast using the backscattered electrons signal (BSE), or EDX for identification of the element composition. Another major advantage of the incineration method are the very short sample preparation times, which permit the screening of larger sample numbers (without the generation of chemical waste) in a relatively short time span for comparative studies. However, the incineration method also has certain disadvantages: It cannot be used for thin and fragile leaf tissue, as incineration of those samples yields fragile ash fragments without sufficient stability for subsequent mounting and investigation. These are, on the other hand, the type of samples most suited for the clearing method. Also, since the organic material is reduced to ash, a certain degree of shrinkage and dislocation of the crystals is inevitable with the incineration method. As the images and Table 1 show, the differences may be negligible in most material but need to be kept in mind for critical comparative studies.

(burner) to 20 (furnace) min
In the LM view the three-dimensional arrangement of crystals in the leaf tissue is only represented as a twodimensional image. This shows precisely, for example, the differential distribution of crystal druses in the intercostal area versus tiny crystals along the leaf veins. A stereo-microscopic view provides an impression of the three-dimensional location of the crystals, but it is difficult to illustrate it sufficiently by LM micrographs. The traditional method of sectioning can be used for this purpose, but leads to notoriously unpredictable results especially in heavily mineralised leaves, including the laceration of sections and crystal dislocation by the microtome blade. Alternatively, micro-CT can be used for this purpose and the images shown in Figure 6 beautifully demonstrate how the precise location of CaOx in the leaf tissue can be visualised with this technique.

CONCLUSION
The addition of another tool for our study of plant biomineralisation presented here should be greatly helpful for deepening our understanding of this -to date surprisingly poorly documented and understood -phenomenon. Recent studies have been able to provide some unexpected insights into the range of functions that biominerals may play in plant tissues, adding to the widely accepted role in herbivore defence. The demonstration of the lightscattering mechanism, for example, in Ficus leaves, is an example. 10,16 To the best of our knowledge, no explanation has so far been provided for the peculiar patterns of thick layers of crystals along the leaf veins, or the differentiation in CaOx deposits in the intercostal areas versus those along the leaf veins. However, in recent years, a large number of studies has started to elucidate the physiological role of CaOx in more detail, demonstrating a complex role in carbon dioxide storage and associations with drought and other abiotic stressors. This indicates that far from being an inert mineral deposit, CaOx plays an active role in both the carbon and calcium cycles. 17,18 We therefore argue that incineration provides a valuable complementary tool for the study of leaf biomineralisation, with the particular benefit of being widely applicable to most leaf tissues and permitting the rapid screening of multiple samples. A comprehensive picture of CaOx in leaves can then be obtained by the complementary use of leaf clearing, SEM and micro-CT, leading to a genuine, three-dimensional picture and permitting even the identification of the element composition of the biominerals present. In view of these promising insights, we believe that the increased ease of visualisation, for example, by the incineration method described in detail here, will provide a tool for the study of CaOx and other biominerals, opening the field to functional and physiological studies elucidating both the causes and mechanisms of CaOx deposition in leaf tissue.

A C K N O W L E D G E M E N T S
This project was supported by a DFG grant (University of Bonn, DFG Research Unit 2685) with project number: 396637283.
Open access funding enabled and organized by Projekt DEAL.