Cystoliths are amorphous calcium carbonate bodies that form in the leaves of some plant families. Cystoliths are regularly distributed in the epidermis and protrude into the photosynthetic tissue, the mesophyll. The photosynthetic pigments generate a steep light gradient in the leaf. Under most illumination regimes the outer mesophyll is light saturated, thus the photosynthetic apparatus is kinetically unable to use the excess light for photochemistry. Here we use micro-scale modulated fluorometry to demonstrate that light scattered by the cystoliths is distributed from the photosynthetically inefficient upper tissue to the efficient, but light deprived, lower tissue. The results prove that the presence of light scatterers reduces the steep light gradient, thus enabling the leaf to use the incoming light flux more efficiently. MicroCT and electron microscopy confirm that the spatial distribution of the minerals is compatible with their optical function. During the study we encountered large calcium oxalate druses in the same anatomical location as the cystoliths. These druses proved to have similar light scattering functions as the cystoliths. This study shows that certain minerals in the leaves of different plants distribute the light flux more evenly inside the leaf.
Leaves of higher plants contain various types of minerals. The most widespread leaf minerals are silica and calcium oxalate, and to a lesser extent calcium carbonate.1, 2 Calcium oxalate and silica are known to take part in plant calcium regulation, heavy metal detoxification, mechanical support and plant protection.3, 4
Oval-shaped and relatively large calcified bodies, known as cystoliths, occur in few, albeit large families of angiosperms, and exhibit common features in different species.5, 6 Two centuries have elapsed since cystoliths were discovered,7 and although physiological roles such as carbon or calcium reservoirs or pH regulators have been proposed,5, 8 no definitive evidence for cystolith function has been presented. Cystoliths are formed by specialized epidermal cells as mineralized extensions of the inner cell wall. The cystolith body is composed of amorphous calcium carbonate, and is connected by a silicate stalk to the peripheral cell wall.8 The facts that cystoliths from various species are transparent bodies and have morphologies that are consistent with scattering light, as well as their location, namely spanning the leaf from its surface deep into its interior, inspired us to investigate the possibility of cystoliths fulfilling an optical role in enhancing photosynthesis.
Some biominerals are known to manipulate light. Trilobites, an extinct class of marine arthropods, had calcitic lenses as part of their visual perception system.9 Calcitic lenses are found in the skeletal part of the arms of certain species of brittlestars, marine echinoderms.10 Some sponge spicules form fibers with unique optical properties,11, 12 and regularly spaced platy guanine crystals are known to constitute the basis of structural colors in fish and spiders.13 Although leaves have an intimate relation to light, and contain abundant minerals, little consideration has been given to the possible role of biominerals in leaf optics. The idea of silica ‘glass windows’ transmitting light into the leaf has been, for the most part, abandoned,14 and to our knowledge only one study to date has proposed a light scattering function for calcium oxalate crystals present in a specialized shade plant.15
The primary function of leaves is to support photosynthesis in mesophyll cells. This photosynthetic tissue is divided in most angiosperm leaves into the dense, palisade mesophyll in the upper (adaxial) side of the leaf, and the spacious, spongy mesophyll in the lower (abaxial) side. The mesophyll is enveloped by epidermal tissue and is permeated by vascular tissue. The high chlorophyll pigment concentration in the mesophyll absorbs light and hence induces a steep light gradient from the irradiated leaf surface into the tissue.16, 17 This gradient creates very different light regimes in the same tissue. In this report we use the term photosynthetic efficiency only as related to the cellular utilization of light for photosynthesis, i.e. we define photosynthetic efficiency as the yield of transforming absorbed light energy into usable chemical energy. An efficient photosynthetic tissue primarily utilizes light energy for carbon fixation, whereas an inefficient tissue loses much of the light energy in other molecular processes, or as heat.
To study the optical function of cystoliths we characterized the leaf anatomy of ten plant species containing cystoliths by light microscopy. MicroCT was performed on five species, showing common morphological characteristics and anatomical locations consistent with previous anatomical work (Figure S1).8 Based on this preliminary survey, we decided to focus on Ficus microcarpa (Moraceae), as a representative cystolith containing species.
2.1. Cystolith Anatomy
A F. microcarpa cystolith resembles the shape of a bunch of grapes, with many pronounced protrusions (Figures1B,C,D). Cystoliths develop in enlarged epidermal cells that protrude through a multilayered epidermis into the palisade mesophyll (Figures 1A,C). In the upper epidermis they are 60–80μm in length and 40-60μm in diameter, whereas in the lower epidermis they are smaller but also protrude into the spongy mesophyll (Figure 1E). Observations performed by high resolution microCT in whole leaves (Figure 1D) show that cystoliths are distributed homogenously inside the leaf: 5 representative scans showed that the upper epidermis contains 25–30 cystoliths/mm2, whereas the lower epidermis contains 15–20 cystoliths/mm2.
2.2. Modulated Micro-Fluorometry Detects Optical Effects of Cystoliths
To test the possibility that cystoliths act as light scatterers to enhance photosynthetic efficiency, we chose to use a non-invasive technique, namely modulated fluorometry, based on the principles of Pulse-Amplitude-Modulated (PAM) Fluorometry. PAM fluorometry is a well established method to measure photosynthetic activity through changes in the fluorescence yield of chlorophyll-a molecules.18, 19 The photosynthetic protein complex machinery resides inside chloroplasts that densely populate the mesophyll cells. Photosystem II reaction centers are the catalytic cores where light energy is converted into chemical energy by electron translocation. Photons that cannot be used in photosynthesis are either re-emitted with the characteristic fluorescence of chlorophyll-a molecules, or their energy is dissipated as heat. In a dark-adapted state all reaction centers are open, absorbed photon energy is primarily used to induce electron transport and fluorescence yield is minimal. Upon illumination the reaction centers gradually close until the electron acceptor molecules are fully reduced. This process is accompanied by a gradual increase in fluorescence yield, termed fast chlorophyll fluorescence rise,20 where the emitted fluorescence represents the light that is not used for photosynthesis, and is thus ‘wasted’. It is this ‘wasted’ light that is measured by PAM fluorometry.
We built a custom-made modulated micro-fluorometer (for a detailed description of the apparatus see Experimental and Supporting Information) that uses a laser beam illuminating an area of 60μm in diameter (Figure2A, Figure S2). The system detects the fluorescence originating from an area tenfold larger than the illuminated area, and thus collects all effective fluorescence resulting from the illumination (Figure S3). Using this modulated fluorometer we measured fluorescence yield signals from freshly detached, dark-adapted F. microcarpa leaves. For each leaf, data were collected from two locations: on–where the light is directed on a cystolith, and off–where cells devoid of cystoliths were illuminated (Figure 2A, Figure S4). Because we use red light that is strongly absorbed in the leaf, only 0.5% of the light is transmitted through the leaf in off locations, whereas 1.5% of the light is transmitted in on locations. This minute difference in transmittance was used to locate cystoliths by imaging the leaf with back illumination.
Each experimental fluorescence dataset was fitted to a simple exponential function characterized by two parameters, the kinetics coefficient τ that is inversely proportional to the steepness of the exponential rise, and the steady-state fluorescence yield level (S.S.F.Y.) achieved after the fluorescence rise (Figure 2B–D, Figure S5). The measured fast fluorescence induction kinetics generated by non-saturating light (and in this set-up also the S.S.F.Y.) correlates with the light flux in the upper mesophyll: higher light flux causes faster light saturation of the reaction centers, yielding fast fluorescence induction kinetics (Figure 2C). In addition the fluorescence steady-state (‘wasted’ light) is higher because a larger fraction of the reaction centers is closed at any moment (Figure 2D). In contrast, lower flux results both in slower kinetics, and in lower steady-state fluorescence yield, as less reaction centers are simultaneously closed.
A clear trend emerges from the modulated fluorometer results: the same light intensity, when applied through cystoliths, causes slower fluorescence yield induction kinetics and lower steady-state fluorescence yield (Figure3A–B). Since only fluorescence originating from the upper mesophyll can be detected by our epi-illumination setup the results point to the fact that the cystolith reduces the light intensity on the outer mesophyll chloroplasts. Because all the energy introduced into the system must be accounted for, the light flux in the lower mesophyll must increase. Light is subsequently more evenly distributed inside the leaf, and the net result is that more light is available for photosynthesis.
2.3. Scattering Properties of F. elastica Cystoliths
To exclude the possibility of artifacts in the fluorescence measurements introduced by the physical presence of the light scatterers, we performed the same modulated fluorometry measurements in Ficus elastica leaves, where the cystoliths are located in an extended multilayered epidermis and do not reach the mesophyll (Figure 1F–H). Thus in F. elastica the light is scattered mostly in the multilayered epidermis and penetrates less into the mesophyll. This anatomy is expected to reduce the scattered light effect. Indeed the modulated fluorometer measurements performed on and off cystoliths showed minor differences, barely exceeding the experimental uncertainty (Figure 3C–D). This minor effect is probably due to the reduced light flux on the outer mesophyll because of cystolith scattering in the outer epidermis, and the thinner mesophyll geometry under the cystolith facilitating some light penetration into the lower mesophyll.
2.4. Calcium Oxalate Druses with Optical Function
During the study, while seeking a cystolith-free species to be used as a control, we examined Carya illinoinensis (Juglandaceae) (pecan) leaves with the microCT and noted the presence of internal bodies anatomically located in the upper part of the palisade mesophyll. This anatomical location is similar to that of the cystoliths (Figure4). Those bodies were identified as relatively large calcium oxalate monohydrate crystal druses located in specialized mesophyll cells.21 The size of these crystal aggregates is similar to that of cystoliths (in the average 60–80 μm, Figure 4B, C), and their average distribution is 20–25 druses/mm2 (Figure 4D). We note that the calcium oxalate crystals in C. illinoinensis differ from other calcium oxalate druses present in leaves, which are typically smaller (less than 20μm), unevenly distributed, and are present in different anatomical locations.3
We performed modulated fluorometry experiments on C. illinoinensis leaves, and found that indeed light irradiated through the calcium oxalate druses causes slower fluorescence yield induction kinetics and lower steady-state fluorescence yield (Figure5), similar to what was observed for cystoliths in F. microcarpa (Figure 3A–B). Thus the calcium oxalate druses also scatter light into the mesophyll, reducing the light flux in the outer tissue and enriching the lower tissue with photons.
2.5. Ex vivo Light Scattering Properties
To characterize the light scattering capabilities of the isolated calcified bodies outside the leaves, we tested the optical characteristics of ten extracted F. microcarpa cystoliths and ten C. illinoinensis druses. Both bodies are transparent in visible light. Their refractive indices were determined to be 1.55 ± 0.01 for cystoliths and 1.58 ± 0.02 for calcium oxalate druses. These values are significantly higher than the refractive index of the surrounding cytoplasm and cell wall, estimated to be around 1.35–1.42.22 Both mineral structures have pointed protrusions and angular surfaces, suggesting high scattering capability. To quantify this scattering activity, extracted minerals were embedded in 1.44 ± 0.02 refractive index resin, irradiated with collimated light and the resultant angular scattering was measured (Figure6, Figure S6). The two biominerals scatter light in a similar manner. The impinging collimated light is transmitted with an effective scattering angle distributed between 0° and 30° (which should be even higher in leaves where the refractive index of the surrounding medium is somewhat lower). These measurements also confirmed that the minerals are transparent in visible light as the integrated light intensity on the detector with and without mineral scattering was the same.
We show here that cystoliths and certain calcium oxalate druses function as light scatterers in the leaf. These mineral bodies shift light from the upper mesophyll to the lower mesophyll, and by so doing reduce the effect of photoinhibition in the upper mesophyll, while at the same time provide more light to the darker lower mesophyll.
Photoinhibition reduces the effective exploitation of light by the photosynthetic apparatus.23 It has been reported that even at light fluxes lower than 10% of direct sunlight, the outermost chloroplasts of a leaf are photoinhibited and actively dissipate some of the light energy into non-photosynthetic pathways.24 Thus under normal illumination conditions the carbon fixation rate in the upper mesophyll is only slightly higher than in the lower mesophyll, even though it absorbs 3–5 times more irradiation.16, 25–27 The mineral light scatterers contribute to the exploitation of light by taking some of the light that would have been absorbed by the photoinhibited, upper mesophyll and scattering it into the lower mesophyll.
Mineral light scattering would be beneficial for the leaf under almost all environmental conditions, except for extremely dark environments. The scatterers cover about 5% of the leaf surface. Thus under light limiting conditions, the most light that could be lost to photosynthesis as scattered, non-effective flux is 5%. The plant species containing mineral light scatterers identified here rarely grow under such low illumination conditions, except for short periods of time close to sunrise and sunset, or under extremely overcast weather conditions. Under all other light conditions there will be a photosynthetic gain from this 5% incident light because it can be effectively used in the lower mesophyll. For example, under direct sun light the flux in the lower mesophyll is ∼10% of the impinging light, and adding another 5% of the impinging light by mineral scattering will elevate the light flux of the lower mesophyll by 50%; a pronounced gain for a tissue responsible for almost half of the leaf carbon fixation. This is similar to the well known ‘sieve effect’ caused by chloroplasts at the cellular level.28 Because the loss under very dim light is 5% of very slow photosynthetic activity, whereas the gain under stronger light regimes is 5% of fully active photosynthesis, changing illumination conditions as well as changes in leaf orientation are not important as long as the leaf, on average, is illuminated by a reasonable flux. It is impossible to quantify from our data the net photosynthetic gain for the tissue, but unless extremely dim light conditions persist throughout long periods, the overall contribution is positive.
Mineral light scattering is only one component of the structural aspects of leaf optics.29, 30 The general anatomy of the leaf is designed to transmit light from the surface into the tissue and then scatter it to minimize losses. The elongated palisade cells were suggested to act as ‘light pipes’ directing light into the tissue.31 The spongy mesophyll is a highly scattering tissue due to the optical interfaces between air spaces and cells, which minimize light transmittance through the leaf. Light scattering by minerals functions in an analogous manner, namely by taking light from the leaf surface and distributing it inside the mesophyll. Since the minerals are internal structures and their scattering activity is roughly independent of the illumination direction, their scattering role is affected by the directionality of the external light regime in a similar way to the whole leaf optics. In collimated illumination perpendicular to the leaf the mineral scattering effect is pronounced. With collimated angular illumination or diffuse light the mineral scatterers will take the same fraction of light and scatter it, in tilted or wider directions, into the leaf.
Analysis of the cystolith and calcium oxalate druse distributions, using the leaf tomograms (Figures 1+4), shows that the average distance between cystoliths is ∼180 μm, and ∼210 μm between druses. The model in Scheme1 is drawn to scale and shows schematically the area affected by a scatterer using the appropriate anatomical dimensions and scattering angles. In this model a distance of ∼200μm between scatterers results in an overlapping continuum of scattered light in the lower mesophyll. This is consistent with the measured distances and the notion of a functional light scattering system. The presence of cystoliths in the lower epidermis of F. microcarpa may reflect the varied nature of irradiation in a tree canopy. These cystoliths, though smaller and less abundant, may fulfill the same role as those in the upper mesophyll in some leaves or at some time of the day, when light is mostly abaxial.
Although we demonstrate here that the scattered light interacts with photosynthetic pigments, additional experiments are necessary to quantify the resulting increase in photosynthetic efficiency, such as measurements of electron transport, carbon fixation, or oxygen evolution.
Here we demonstrate that some minerals in leaves scatter light to the photosynthetic pigments. In the case of cystoliths this is the first demonstrated mechanism for their function in leaves. They may well perform other functions, but these still need to be elucidated. The similarity in the anatomical locations and optical functions of cystoliths and certain calcium oxalate druses, raises the possibility that light scattering by biominerals in leaves may be a widespread phenomenon.
5. Experimental Section
Leaf biomineral extraction: Fresh leaves were collected from Ficus microcarpa, Ficus elastica, and Carya illinoinensis trees at the Weizmann Institute of Science. The biominerals were extracted based on the method of Arnott6. Fresh clean leaves were cut into small pieces and blended with a Waring laboratory blender in absolute ethanol for one minute. The mixture was filtered through cheesecloth. The heavier biominerals were separated from lighter leaf fragments by a series of decantations in ethanol. The extracted biominerals were kept in ethanol at -18 °C.
Light microscopy: Fresh leaves were cut into 1mm2 squares. Leaf sections were embedded in warm (∼50 °C), 3.5% w/v LE agarose (BMA) solution that was still liquid, and left to harden for 30 minutes. An agarose gel block containing a cut leaf was sectioned in water into 50μm slices using an OTS-4000 vibrotome (Electron Microscopy Sciences). The sections were placed in a water drop on a microscope slide, and images were taken using a Nikon E600 Pol microscope and a Nikon DS Fi1 camera.
Electron microscopy: Fresh leaves were cut into 1mm2 squares, dehydrated in a series of ethanol solutions, and critical-point-dried using a CPD-030 critical-point dryer (Bal-Tec). When dried, leaf sections were broken to expose the inner cross-section and coated with gold. Images of extracted biominerals, as well as leaf cross-sections, were taken using a SUPRA55 scanning electron microscope (Zeiss).
MicroCT experiments: A fresh leaf slice was positioned inside a plastic pipette tip. The slice was covered with water and placed in a microCT (MICRO XCT-400, Xradia). Tomography was carried out using source parameters of 40 kV and 200μA. 800 images were taken with 15s exposure time. Raw data were reconstructed with the XRadia software that uses a filtered backprojection algorithm. 3D surface rendering and measurements were carried out with Avizo software (VSG).
MicroCT data analysis: Mineralized bodies were counted in tomograms of 2mm2 leaf sections. These counts were used to estimate the density of mineralized bodies/leaf area. Distance distributions were calculated using the square root of the area of Voronoi cells constructed from the microCT images.
Modulated fluorometer set-up: A modulated fluorometer was home-built for the experiments. A computer-controlled analog-modulated diode laser at 636nm (Coherent CUBE) provided both the continuous actinic light and the 20KHz modulated measurement light. The actinic light was generated by a 0.2V DC signal. The measuring light was generated by 2.5V pulses. The pulse duration was 3 μs with 47 μs between pulses superposed on the actinic light level. These parameters resulted in ∼50% of the light flux originating from the actinic light and ∼50% from the modulation pulses (Figure S7).
The expanded laser beam was activated using a fast mechanical shutter (Stanford Research Systems SR475) and directed onto the leaf through a dichroic beamsplitter (Semrock FF650) and a 10X microscope objective, providing a Gaussian excitation spot with a 60μm waist (characterized by replacing the leaf with a CCD camera). Chlorophyll fluorescence was epi-detected, spectrally filtered using both a color glass (Schott RG650) and a bandpass interference filter (Semrock 697/75) to eliminate the residual excitation light, and measured using a high-sensitivity near-infrared enhanced photomultiplier tube (Hamamatsu R10699) and a lock-in amplifier (Stanford Research Systems SR830) to which it was coupled through a 100kΩ resistor. Anatomical characterization of the observed leaf was performed by transmission imaging using either an incandescent lamp or a near-infrared LED emitting at 850nm (Mightex LCS-850) as a light source and a CCD camera (Mintron 13V1C) detector. The entire experiment was computer controlled by dedicated LabView software (see figure S2 for an image of the modulated fluorometer).
This set-up can measure typical fluorescence induction kinetic curves of the leaf (Figure S3). These results are similar to the conventional PAM fluorometer data and confirm that our set-up measures the known changes in chlorophyll-a fluorescence yield (Kautsky effect)18, 19. Because of the very small illumination spot, the dim light intensities, and the instruments limitations in this set-up, we do not have the dynamic range needed to accurately measure the parameters of F0 and Fm which are available in commercial PAM fluorometers.
Modulated fluorometry experiments: Mature leaves were detached from trees and kept in a moist, dark box for 1 hour. The measurement location was selected by moving the fluorometer stage and illuminating for 40 seconds with back illuminating light and laser beam. The camera image gave information about cystolith location and the laser spot on the image indicated the location to be measured (Figure S4). After the location was fixed the leaf was kept in the dark for 5 minutes, allowing a full dark acclimatization again, followed by the modulated fluorometry measurement. The measurement duration was 4 seconds from the shutter opening, in order to capture the induction kinetics and the initial fluorescence decline (only the first 2.5 seconds are shown in the results). On each leaf 6 locations on and off were measured alternately. During the ∼1hour of measurements on each leaf no detectable change occurred in the consecutive measurement results indicating that the viability of the photosynthetic apparatus was unchanged.
Light intensity was calibrated for each leaf species by a set of neutral density filters applied to the laser beam to yield an induction kinetics time of about 0.5 second. Light flux used to measure F. microcarpa leaves was 350 μmol(photon) m−2 s−1, to measure C. illinoinensis leaves was 220 μmol(photon) m−2 s−1, and to measure F. elastica leaves was 700 μmol(photon) m−2 s−1. In order to prove that the illumination conditions are far from saturation, we verified that increasing light intensities indeed produce higher fluorescence.
Modulated fluorometry data analysis: Generally, the fluorescence rise after the onset of illumination is non-exponential. However, free energetic coupling between different types of PSII, mainly between PSII α and β allows for this rise to become close to exponential32–34. Thus, the fluorescence yield data from the first 2.5 seconds after the shutter opening were fitted using Matlab to the function:
Fluorescence yield = f0 + (fsteady-state - f0)*(1 - e−time\induction kinetics coefficient) The fitted parameters that characterize the induction kinetics are the fluorescence steady state (fsteady-state) and the induction kinetics coefficient. This simple exponential function describes fully and accurately the differences in the induction kinetics relevant to our experiments (figure S5).
Leaf transmittance measurements: The modulated fluorometer set-up was used to measure light transmittance at different locations in the leaf. A large area photoreceiver (New Focus 2031) was placed 2 mm from the abaxial side of the leaf and light intensities were recorded on and off mineral. The laser intensity without a leaf was used to calculate the transmittance value.
Refractive index measurements: Extracted biominerals were placed in refractive index calibration oils, observed under the microscope and photographed. The biomineral refractive index was determined using steps of 0.01 in the refractive index of the oils. The oil refractive index resulting in minimal observable contrast, was determined.
Light scattering experiments: Extracted biominerals were mounted on a microscope slide with Neo-Mount (Merck) (refractive index 1.44 ± 0.02). Optical characterization was performed on a Zeiss Axiovert 200 inverted microscope. The white light source was collimated using two irises, and scattered light was collected by a 40X objective (Zeiss EC Plan Neofluar) having a numerical aperture of 0.75. The back focal plane was imaged onto a CCD camera (Mintron 13V1C) to obtain the angular light intensity distribution following scattering. Each scatterer was also imaged onto another CCD camera (Acton Photonmax 512) to identify the fraction of the illuminated area filled by it (quantified with ImageJ software). For each plant species 10 individual mineral scatterers were measured (cystoliths with close to upright position were chosen).
Light scattering data analysis: Images of light scattered by the mineral scatterers were analyzed using ImageJ software. The center of the image was determined as the center of the circle created by the objective lens (Figure S6B), and pixel intensities were integrated radially from the center to the outer image radius. The outer radius, corresponding to the projection of the back aperture of the objective lens represents the maximal scattering angle through the objective lens, 48.6 degrees in the lens used. This was used to calibrate the conversion of the radial distance (in pixels) to angular scattering (degrees). Locations devoid of minerals in the sample were used to determine the background signal. When measuring through a mineral the outer 10 pixels closest to the outer radius had no detectable signal in all measurements and were used to correct for camera dark counts and readout noise. Since the mineral did not cover the entire illuminated area some fraction of light did not pass through the mineral and was not scattered (Figure S6A). The fraction of light that did not pass through the mineral was calculated as the fraction of area in the illumination beam that was devoid of mineral. This unscattered light that did not pass through the mineral was subtracted from the signal. The resulting data representing the angular scattering of the biomineral were summed up into 5.4° steps, the resolution for detecting scattered light (Figure S6C–D).
Supporting Information is available from the Wiley Online Library or from the author.
We thank Osip Schwartz and Dana Charuvi for their help with the experiments. This work was supported by Crown center of Photonics, ERC starting researcher grant SINSLIM 258221, Department of Energy Award DE-FG02-07ER15899 and the Schmidt Minerva Center for Supramolecular Architecture. D.O. is the incumbent of the Recanati Career Development Chair of Energy Research, L.A. of the Dorothy and Patrick Gorman Professorial Chair of Biological Ultrastructure, and S.W. of the Dr. Trude Burchardt Professorial Chair of Structural Biology.