EDX and Raman spectroscopy reveal the unexpected complexity of plant biomineralisation
H. J. Ensikat, M. Weigend
Nees Institute for Biodiversity of Plants, University of Bonn, Germany
Hans-Jürgen Ensikat has worked at the University of Bonn – as a technician in several EM laboratories – since 1971. After an apprenticeship in chemistry, he joined the TEM working group ‘Festkörperoberflächen’ in the Mineralogical Institute. In 1982 he relocated into the TEM facility of the Botanical Institute. Later he moved to the SEM lab of the Barthlott group in the Botanical Institute/Nees Institute, well-known for the 'Lotus Effect', where he is responsible for SEM and AFM. Despite retiring in 2017, he continues to take care of his SEMs.
Recent scanning electron microscopy (SEM) studies demonstrated that biomineralisation in higher plants is unexpectedly complex when reporting the co-occurrence of calcium phosphate with silica and calcium carbonate in cell walls of Loasaceae trichomes. Unlike animals, plants are able to produce several different biominerals, either in high concentrations or in combination with organic compounds forming differentially mineralized composite materials. The biominerals may be deposited in specialised cells (idioblasts) or they form hard structures such as trichome tips, protecting the plants against herbivores. SEM-based EDX analyses resolve the distribution of the elements Si, Ca, and P in great detail. It has been shown that single-celled trichomes may deposit at least three different biominerals at specific locations, and Raman-spectroscopy enables simultaneous identification of organic and organic components of the mineralized cell walls, even in situ on living samples.
We thank Thorsten Geisler-Wierwille and Martina Menneken, Institute of Geosciences of the University of Bonn, for their support in Raman spectroscopy and preparation of the manuscript.
Hans Jürgen Ensikat, Nees Institute for Biodiversity of Plants, Meckenheimer Allee 170, 53115 Bonn, Germany.
Biominerals produced by animals, e.g., calcium phosphate and carbonate in bones, teeth, egg shells, shells of mollusks, or the mineral deposits formed by corrals or diatoms in the sea are well known and widely studied. Biominerals in plants appear less impressive, as they rarely form large distinct structures or encompass entire tissues. In plants, biomineralisation is usually found at the level of individual cells, either as intracellular deposits or encrustations of the outer cell walls[1-3]. Particularly the tips of stinging hairs of stinging nettles and trichome tips of many other plants are known to contain highly-concentrated silica, which provides the required mechanical properties[4-5].
In addition, several calcium compounds, mainly Ca-based oxalates, carbonates, and phosphates, are found as biominerals, either as concentrated deposits or as composite materials in combination with organic cell wall material[1, 6-8]. (We here use the term ‘biomineral’ for solid materials with the inorganic component large enough to determine their mechanical properties. This excludes dissolved salts or traces of inorganic elements).
Our observation that calcium phosphate can replace silica in trichome tips and stinging hair apices in several plant families was the starting point for a more comprehensive study of the mineralisation of plant trichomes[5,9-10].
Mineralisation patterns were found to be far more complex than expected. Up to three different biominerals were found at specific locations in the walls of single-celled trichomes of Loasaceae and various stinging hairs. This indicates that biomineralisation of cell walls is strictly controlled by cellular processes.
The present article aims at providing an overview of the potential of the complementary use of light and scanning electron microscopy as well as confocal Raman spectroscopy for a detailed investigation of plant biomineralisation.
MATERIALS AND METHODS
The plants used for this study were cultivated in the Botanical Gardens of the University of Bonn, Germany. Scanning electron microscopy was performed with a Stereoscan S 200 (Cambridge Instruments, Cambridge, UK) and a LEO 1450 (Carl Zeiss AG, Oberkochen, Germany), equipped with secondary and backscattered electron (SE and BSE) detectors and an energy-dispersive X-ray (EDX) element analysis system (Oxford Instruments, UK).
Preparation for SEM included examination of fresh (hydrated) plant samples and cryo-SEM imaging of frozen samples at ca. -80°C, without or with very thin palladium sputter coating. It was found that this gives the best results for compositional contrast in BSE images and minimised fixation and drying artefacts.
For EDX mapping we preferred dry samples. Prior to EDX analysis, leaf pieces were fixed in ethanol (70%) with formaldehyde (4%), dehydrated and then critical-point (CP) dried in a SCD 040 Sputter Coater (Balzers Union, Liechtenstein). Isolated trichomes were rapidly dehydrated by washing in ethanol and acetone and simply dried from acetone. Dry samples were sputter-coated with a thin layer (ca. 10 nm) of palladium (Junker Edelmetalle, Waldbüttelbrunn, Germany), which, in contrast to gold, does not have X-ray lines that interfere with those of the biominerals.
Colorised SEM images were rendered with a standard image processing software, Paint Shop Pro, either Jasc PSP 6 or Corel Paint Shop Pro X9. The images from different detectors were combined with the software function ‘Combine HSL’. For combined SE-BSE images (e.g., Fig. 1), the SE image was used for the lightness (L) channel, the inverted BSE image for the hue (H) channel; color saturation (S) was uniformized by a picture without contrast. The combination of SE and element-mapping images is shown in Fig. 2: the lightness channel contains the SE image; the saturation channel uses the sum of the selected elements, for example Si and Ca; the hue channel contains the positive Ca image mixed with the inverted Si image, producing a color shift towards green for Ca and towards red for Si. Areas without Si or Ca remain grey (colourless).
Figure 1. SEM images of silica and calcium oxalate deposits in plants. A) SE, BSE and combined color image of a fresh Pampas grass leaf with silica phytoliths and silicified trichome tips. In the color image both, topography and compositional contrast, are distinctly visible. Phytoliths are covered by cutin and wax and therefore hardly visible in the SE image. B) Single phytolith isolated with piranha solution. C, D) Combined SE-BSE color images of intercellular spaces inside a CP-dried Lotus leaf, faced with calcium oxalate lithocysts. D) Detail image, showing different development stages of crystal druses, growing in a kind of ‘suicide cells’. Finally, the sharp tips of the druses destroy the wall of the lithocyst, which collapses, and the crystals are set free. Scale bars: A: 40µm; B: 10µm; C: 100µm; D: 20µm.
Figure 2. SEM images and Raman spectra of calcium carbonate deposits in Urtica dioica (A) and Saxifraga hostii (B-D). A) SE and EDX element mapping images of isolated cystoliths of a stinging nettle (Urtica dioica) leaf, with combined silicon and calcium images, illustrating the image processing for rendering color images using the ‘combine HSL’function. The color image shows clearly the Si and Ca distribution. B) Combined SE-BSE image of a fresh leaf of Saxifraga hostii with a crust of calcium carbonate (red) on the surface, mainly at leaf margin. C) Detail of the leaf margin after removal of calcium carbonate and wax crusts with hydrocloric acid and, after dehydration, with chloroform. The CP-dried sample shows a pore complex for calcium carbonate segregation. D) Calcium carbonate deposit on the fresh leaf, showing crystalline shape. E) Raman spectra of Urtica cystolith and Saxifraga mineral crust. The sharp Raman band at 1086 cm-1 and two less intense bands at 713 and 283 cm-1 indicate a crystalline state of the Saxifraga deposits. The broader Raman band of the cystolith with a maximun at 1080 cm-1 is typical for amorphous or less ordered calcium carbonate. Scale bars: A: 50µm; B: 500µm; C: 100µm; D: 10µm.
For Raman spectroscopy isolated trichomes and stinging hairs were carefully washed to remove any remnants of cell liquids. Raman spectra were then collected with a confocal Horiba Scientific LabRam HR800 Raman spectrometer at the Institute of Geosciences of the University of Bonn, Germany. The Raman spectra were excited with a Nd:YAG (532.09 nm) or a diode (783.976 nm) laser through a Olympus BX microscope.
In vivo Raman analyses of fresh trichomes were performed with the 784 nm excitation on sample regions that were immersed in water to avoid any sample heating and to obtain a clear microscopic image using a 50x long-distance objective lens.
Intracellular biomineral deposits: Biominerals may be found either as intracellular and/or external deposits or as cell wall components, both may occur simultaneously within the same plant. Intracellular deposits occur with different compositions:
Silica bodies (phytoliths) occur in numerous plant species and are particularly common in Poaceae such as Cortaderia hieronymi (Pampas grass, Fig. 1 A, B), where they are found in the leaf epidermis next to small, unicellular trichomes with a mineralized tip. Due to the characteristic shapes and chemical stability of phytoliths they are of great importance as microfossils and useful for the reconstruction of paleohabitats.
Calcium oxalate crystals are found in different parts of many plant species and occur in many different shapes, from single crystals over crystal sand to so-called druses. They develop in vacuoles of specialized cells, sometimes termed ‘lithocysts’ or ‘crystal idioblasts’, and are usually not found in intercellular spaces. The images of a Lotus (Nelumbo nucifera) leaf show large intercellular spaces with crystal druses (Fig. 1 C) and illustrate how the druses grow in lithocysts and are expelled into the intercellular space (Fig. 1 D) when the cell wall ruptures; a process that has been descibed previously.
‘Cystoliths’ is the common designation for intracellular calcium carbonate bodies, which occur in numerous plant species. EDX mapping of cystoliths from an Urtica dioica leaf (Fig. 2 A) indicate that they are unexpectedly made up of both calcium carbonate and silica, where silica appears to represent the nucleus of cystolith formation. Calcium phosphate has elsewhere been reported to occur in cystoliths – we found it in the leaves of Urtica mairei (Fig. 3 C) together with calcium carbonate.
Calcium carbonate deposition outside of the tissue is uncommon, but has long been known from the stonecrops (Saxifraga spp.). Figs. 2 B-D show leaves of Saxifraga hostii that secrete the mineral through special pores at the leaf margin. Raman spectra (Fig. 2 E) demonstrate that the excreted calcium carbonate in Saxifraga represents crystalline calcite, as evidenced by the sharp Raman ν1(CO3) band at 1086 cm-1 and less intense bands at 713 and 283 cm-1, whereas the cystoliths of Urtica are composed of less ordered, possibly amorphous calcium carbonate, characterized by a much broader ν1(CO3) band with a maximum at 1080 cm-1.
Mineralized trichome walls
Plants can be covered by many different types of hairs (trichomes) with different functions. Some of them have mineralized outer walls, providing mechanical stability or other special properties. In some cases their function for protection against herbivorous animals is obvious, e.g., for stinging hairs that act as hypodermic syringes, or for small trichomes with silicified hard and sharp tips. Stinging hairs, which inject a pain-inducing liquid under the skin when touched, occur in several different plant families. It is well known that those of the common stinging nettle are mineralized with silica and calcium carbonate (Fig. 3A) and form a stable container for the caustic liquid. EDX analyses of Loasa stinging hairs (Fig. 3B) surprisingly showed that they had calcium phosphate instead of silica in their apices. Calcium phosphate had been rarely reported from plants – we found it also in stinging hairs of other species of Urticaceae, such as Urtica mairei (Fig. 3C) and stinging hairs of other plant families. Other Loasaceae species were found to have silica and calcium phosphate in the stinging hair tips, and calcium carbonate in the shaft.
Figure 3. Stinging hairs with mineralized walls. Combined SE and EDX element mapping images of stinging hairs of different species. Color codes: Si (red); Ca (yellow); Ca with P (green); non-mineralized (gray, colorless). A) Apex of Urtica dioica stinging hair, with silicified tip and calcium carbonate in the shaft. B) Loasa pallida stinging hair, with calcium phosphate in the tip and calcium carbonate in the shaft; silica was not present. C) Surface of a CP-dried leaf of Urtica mairei with stinging hairs, small trichomes, and cystoliths. Stinging hair tip is silicified, whereas the shaft is mineralized with calcium phosphate. EDX spectra show element contents of selected spots defined in C). Cystoliths (Spot 4) have a lower P-to-Ca ratio than stinging hair shaft and base. Scale bars: A-B: 100µm; C: 300µm.
A dense cover of the plant surface with stiff trichomes represents a mechanical barrier for small animals, and an encrustation of the outer wall with silica particles, found in many grasses, bamboo, sugar cane, or in Horsetail (Equisetum), is strongly abrasive for animal teeth.
The highest complexity of mineralisation patterns is found in the glochidiate trichomes of Loasaceae. Some species contain three biominerals at certain locations of these single-celled structures.
While several Loasa species contained calcium phosphate and carbonate, but no silica, trichomes of Caiophora lateritia had silica and calcium phosphate in the barbs and calcium carbonate in the shafts, together with organic compounds.
EDX spectra and mapping show the distribution in detail (Fig. 4 A,B). Raman spectra provide detailed information about inorganic components, such as calcium carbonate and phosphate, including their crystallinity (or lack thereof), as well as organic components, such as cellulose and pectin.
Unfortunately, the main Raman bands of carbonate and cellulose overlap broadly. Calcium compounds, however, can easily be removed by a treatment with hydrochloric acid, providing a differentiated picture of the distribution of cellulose and other organic components (Fig. 4C). Also other organic cell wall components such as lignin and waxes have characteristic Raman spectra and can be identified if present.
Figure 4. Glochidiate trichomes of Loasaceae. A) Cryo-SEM image of Caiophora lateritia leaf surface; combined SE-BSE image. B) EDX mapping images of an isolated glochidiate trichome and spectra of selected spots, showing the distribution of three biominerals: silica (red), calcium phosphate (green), calcium carbonate (yellow / brown). C) EDX mapping image and Raman spectra of a Loasa pallida glochidiate trichome, which contains calcium phosphate in barb tips, but no silica. Both, inorganic and organic components can be identified. The position of the carbonate band indicates amorphous calcium carbonate (see Fig. 2). After dissolving the calcium compounds with HCl, the remaining material shows the spectrum of cellulose and a pectin peak. Scale bars: A-C: 40µm.
Raman spectroscopy even permits in vivo analysis of growing trichomes. Fig. 5 shows spectra of very young stinging hair tips of Urtica dioica and Nasa amaluzensis that were acquired with a 784 nm laser under immersion of the sample region in water. Active cytoplasmic streaming confirmed their living state.
Figure 5. In vivo Raman analyses and light microscopy of trichomes. In vivo acquired Raman spectra of tips of very young stinging hairs of Nasa amaluzensis and Urtica dioica, and paraffin wax for comparison. Samples were immersed in water. In Nasa, carbonate and phosphate as well as organic components can be identified. The Urtica spectrum shows only faint traces of wax and cellulose; silica – if present – would not contribute a distinct signal in the spectrum.
The spectrum of the Nasa stinging hair tip shows distinct Raman scattering signals of cellulose, wax, pectin, and the mineral components phosphate and carbonate. The Raman spectrum of Urtica, on the other hand, reveals only small traces of wax and cellulose, whereas the suspected amorphous silica content is almost invisible in the spectrum as silica has a comparatively low Raman scattering cross section.
Intracellular depositions of calcium oxalate crystals and silica phytoliths have attracted interest since long for their remarkable morphology.
Nevertheless, their purpose remains often unclear, and some biominerals may represent little more than convenient deposits of physiological waste products. Conversely, biomineral formation in defensive plant structures appears to be highly ordered and site-specific. Biomineralisation in general and the formation of calcium phosphate-rich wall regions is anything but random, but rather appears to be under very strict developmental control.
Calcium phosphate has until recently rarely been reported as a structural biomineral in plants, despite its tremendous importance in medicine and zoology. The biomineralized cell walls here studied are various composite materials with cellulose and pectin as main organic component and various inorganic components. It thus differs profoundly from biomineraliziation in animals which leads to protein-based composites.
The study of biomineralisation in plant trichomes can benefit from the fact that the living cells are accessible to in vivo microscopy, and Raman spectroscopy can be used on a living system, as well as other modern analytical techniques in light microscopy.
Biomineralized plant structures in themselves are only the tip of a very complex physiological and ecological phenomenon: It seems worthwhile investigating both, regulatory mechanisms – starting with a selective ion uptake, transport and integration – and biomechanical and ecological functions of these complex plant structures.
SUMMARY AND CONCLUSIONS
Biomineralisation in plants is more than a simple deposition of excess material or incrustation of cell walls. It appears to be a precisely controlled formation of composite materials of various inorganic and organic components which are deposited at specific locations by a single cell. The localisation of mineralisation processes right on the plant surface – in trichomes – means that they can be studied in vivo with suitable methods.
1. Skinner, H. C. W., Jahren, A. H. Biomineralization. In: Biogeochemistry: treatise on Geochemistry. W. H. Schlesinger, Elsevier Science 8: 117-184, 2003.
2. Bauer, P., Elbaum, R., Weiss, I. M. Calcium and silicon mineralization in land plants: Transport, structure and function. Plant Sci. 180: 746-756, 2011.
3. Marafon, A. C., Endres, L. Silicon: fertilization and nutrition in higher plants. Rev. Ciec. Agrar., 56: 380-388, 2013.
4. Thurston, E. L., Lersten, N. R. The morphology and toxicology of plant stinging hairs. Bot. Rev. 35: 393-412, 1969.
5. Mustafa, A., Ensikat, H. J., Weigend, M. Stinging hair morphology and wall biomineralization across five plant families: conserved morphology versus divergent cell wall composition. Am. J. Bot. 105: 1109-1122, 2018.
6. Franceschi, V. R., Horner, H. T. Calcium oxalate crystals in plants. Bot. Rev. 46: 361-427, 1980.
7. Hartl, W.P. et al. Diversity of calcium oxalate crystals in Cactaceae. Can. J. Bot. 85: 501-517, 2007.
8. Ensikat, H. J., Geisler, T., Weigend, M. A first report of hydroxylated apatite as structural biomineral in Loasaceae – plants’ teeth against herbivores. Sci. Rep. 6: 26073, 2016.
9. Mustafa, A., Ensikat, H. J., Weigend, M. Mineralized trichomes in Boraginales: complex microscale heterogeneity and simple phylogenetic patterns. Ann. Bot. 121: 741-751, 2018.
10. Weigend, M., Mustafa, A., Ensikat, H. J. Calcium phosphate in plant trichomes: the overlooked biomineral. Planta 247: 277-285, 2018.
11. Ensikat, H. J., Mustafa, A., Weigend, M. Complex patterns of multiple biomineralization in single-celled plant trichomes of the Loasaceae. Am. J. Bot. 104: 195-206, 2017.
12. Kuo-Huang, L. L. Calcium oxalate crystals in the leaves of Nelumbo nucifera and Nymphaea tetragona. Taiwania 35: 178-190, 19