In grass plants, it was reported that a unique phenomenon, ‘pollen exudation’ occurs shortly after pollen capture2,3. In spite of the potential importance, this physiological event has received little attention2–5,10,11,35, presumably due to the two reasons. One reason might be that this event is not universal in angiosperm, being confined with grass plants2,3. Another reason is the technical difficulty of microscopic observations because of the rapidity at picolitre scale prior to pollen adhesion at grass pollination. In grasses, how mechanically pollen exudation initiates remain unclear in context of pollen-stigma interaction. In this study, we revisited pollen exudation2 to analyse the pollen behavior, directly determining metabolites in the exudates in intact rice plants. Consequently, we found that during pollen exudation pollen grains behaved like a roly-poly toy to self-position on the stigma prior to pollen adhesion (Figs. 1 and 2, Supplementary video 1). The rocking motion reported here might be closely associated with the metabolic changes induced at pollen capture, as discussed below. Furthermore, our data supports that rice exudates were originated from the grains, not stigma. The pollination pattern observed here could be systematically different from the general model of angiosperms in A. thaliana1 (Fig. 2). Therefore, the pollen exudation process reported in grass plants2,3 should be reconsidered in future research.
Water uptake from the stigma during pollen exudation has been long assumed2,5,11; however, our in-oil observation clearly shows that pollen exudation per se proceeds without any contact with stigma to germinate, but with little tube growth (Fig. S1). Also, pollen exudation occurs from the grain surface as subsequent event throughout approximately 20 s of apparent stationary state after pollen capture, and thereafter the grains did self-position on a part of ‘slippery’ stigma epidermis covered with exudates (Fig. 2B, also see roly-poly toy-like motion in Supplementary video 1). It is noted that Vpollen had been decreased over time, reaching to 85.3% of initial volume (Fig. 1G). Hence, significant water uptake from the stigma could have not initiated at least the first 2 min after pollen capture (Figs. 1 and 2B). Additionally, the observed 14.7% (= 100 − 85.3%) of Vpollen reduction might be associated with possible changes in vacuole size synchronised with pollen exudation and dehydration from the grain surface (Fig. 4C). The data demonstrate that rice pollen exudation precedes pollen adhesion, resulting in germination, but with a little pollen hydration at 2 min, at which the water potential gradient between pollen and stigma could have been established5. Most flowering plants with partially dehydrated or dry pollen lack pollen exudation, and they cause pollen adhesion (pollen foot formation) to induce ‘pollen hydration’, leading to germination and tube elongation1,7. Based on our data, the context of these events observed in rice conflicts with the early interpretation5. Taking into account the difference in number of germination apertures in the grains (i.e., three germination apertures in A. thaliana pollen grains versus a single aperture in rice), the rapidity of rice pollination appears to be strategically different from the process originally presumed2,5,11,32.
In rice, pollen swelling due to an increase in Vpollen results in anther dehiscence17. Pollen developmental stage at anther dehiscence corresponds to trinuclear pollen stage31. A reanalysis of pollen swelling in rice cultivar ‘Akihikari’21 showed that the artificial floret opening induces 15% increase in Vpollen (n = 15, 6–26% between 2–28 min). It is suggested that two contrasting physiological events, pollen swelling and exudation might be both attributed to the volumetric control of pollen vacuoles associated with changes in osmotic pressure in the vacuolar sap38. At anther dehiscence in rice, trinuclear pollen grains exhibit a polarity, exhibiting amyloplast (starch) localization33. More recently, it has been shown in root tips that amyloplast sedimentation along with the gravitational direction is affected by vacuolar function and membrane trafficking39. Hence, it is possible to interpret that amyloplast sedimentation towards the germination aperture might have occurred in the dispersed pollen grains (Fig. 4A). The direction of amyloplast sedimentation would be influenced by the orientation and exact location of the grains in the thecae emerged from each spikelet in a panicle at flowering. Along with the gravitational direction, if a single polar pollen grain with amyloplast sedimentation moves 70 mm of distance from the thecae in a spikelet to the stigma in another spikelet (at which double fertilization occurs) in the same panicle (see Fig. 4A), the gravitational potential energy, Eg (= mgh) can be estimated to be 1.39x10-7 J, assuming that the grain weight is same with rye pollen, 20.2 ng40. It is expected that Eg might be converted to the kinetic energy of pollen at capture.
Because mature rice pollen grains had positive turgor (see Results), the interpretation on water relations with ‘dehydrated’ pollen grains proposed in rye5 might not be applicable to the hydrating rice pollen grains. Considering rice stigma morphology with its smaller cell width than the pollen diameter, together with its geometry at pollination (see Supplementary video 2), it is pointed that collision to the stigma surface would cause elastic deformation, which instantly alters turgor to generate a pressure pulse in the cell (Fig. 4B). Microscopic observation also shows that pollen grains had no bounce on stigma surface at pollen capture (see Supplementary video 1), arising another possible effect of electrostatic force upon pollen capture (Fig. 4B). Importantly, a significant lapse exists between pollen capture and exudation followed by the rocking motion (Supplementary video 1). The lapse lasts for about 20 s, at which pollen grains apparently had remained in the stationary state (Fig. 4B to C). What was happening in the grains during the lapse? We interpret this to that majority of impact energy would have stored (consumed) in the pollen grains, without generating significant heat, aside from the elastic response. In vitro rice pollen analysis reported that vegetative nucleus migrated towards the vicinity of germination aperture and then to the center of the pollen as the central vacuole shrunk, showing a close interaction between vacuolar dynamics and nuclear migration34.
Compared with the metabolites between exudates and stigmatic papillae, much greater content of ATP has been detected in the pollen grains (Table S1). The energy gained upon collision might be rapidly converted to the chemical energy to be consumed by motor proteins. During the pollen exudation, it is presumed that vegetative nucleus and twin sperm nuclei would move towards the single germination aperture, and then amyloplast aggregation with starch degradation might occur along with the direction of gravitational force (Fig. 4C), actively consuming ATP by mortar proteins (see Fig. 4C). These putative intracellular logistics might be closely associated with pollen exudation due to the vacuolar volumetric regulation (see Fig. 4). And consequently, this cellular modification should alter the center of gravity to the lower position, where the germination aperture is located, causing roly-poly toy motion on lubricant stigma surface to be fixed for the subsequent steps (Fig. 4C to E). Once the position is fixed, viscous foot structure is formed, and thereafter pollen hydration slightly occurs, as Vpollen slightly increased (Fig. 1G), resulting in germination and tube growth (Fig. 4F to G).
Single-cell metabolomics has been widely used in life science and contributed to the various discoveries in cellular responses in plants. In plant reproduction, pharmacological and genetical studies have been succeeded to deal with the pollen-stigma interaction in A. thaliana or self-incompatible plants, such as Brassica family1,7,41. While pollen adhesion is a critical factor for heat-induced spikelet sterility in grass plants, no attempt has been made to directly assay metabolites related to pollen-stigma interaction. In this study, conducting picoPPESI-MS assay in rice let to find clear spatial variations in chemical compositions between pollen exudates, pollen grains and stigmatic papilla cells. Our results strongly suggest that high concentration of sugars, fatty acids and redox-related metabolites observed in exudates are responsible for the rapid foot formation that is completed within 2 min after pollen capture (Fig. 1), optimising the physical viscosity of pollen foot as well as cellular redox status and involving in the signaling cascades at pollen-stigma recognition (Fig. 3). The saturated fatty acids detected in the exudates would be used for synthesizing triacylglycerol that restores hydraulic contact in the foot structure42 and plays a crucial role as signaling molecules for the subsequent pollen-tube growth, as pointed previously43. Hence, it is pointed that such concentrated solutes in the exudates might alter the apoplastic water relations (throughout a reduction in apoplastic osmotic potential) in the stigma papillae cells in contact (see Supplementary discussion, Reinterpretation on water relations of pollination process in grass plants). The contact angle of pollen foot shows high wettability, which refers to the presence of larger adhesive forces between the exudate and stigmatic surface than cohesive forces, suggesting the contribution of electrostatic force at pollen adhesion, rather than hydrogen bonds and Van der Waals forces.
It has been noted that on-site metabolomics shows active accumulation of two glycosylated flavonoids, astragalin (Kaempferol-3-O-glucoside) and sambicyanin (Cyanidin-3-xyloglucoside) in cellular fluids of dry stigmatic papillae prior to pollen capture (Figs. 3B and S3). These endogenous effective antioxidants most likely prevent reactive oxygen and nitrogen species to mediate cellular redox state in stigmatic papillae44. Since the microcapillary tip is assumed to be located in the vacuoles of stigma cells during the pressure probe operation37, these glycosylated flavonoids are likely present in the vacuoles. External flavonoid (Kaempferol) treatment is capable of cancelling self-incompatible response to cause self-pollination in Brassica45, and hence greater accumulation of astragalin and sambicyanin in stigma cells likely mediates rapid pollination in rice with allogamy of anemophilous flowers24.
The grain-to-grain variations in pollen exudation might be partially attributed to the variations in Eg. Based on our findings, pollen exudates might have three advantages, i) rapid pollen foot formation (pollen adhesion), ii) prevention of dehydration from the pollen surface by forming the film containing lipids and proteins, iii) lubricant working at the receptive zone of the papillae, reducing friction resistance at the roly-poly toy motion. The differences between large and small amounts of exudates occasionally and repeatedly observed during exudation (see Figs. 1, 2B, and Supplementary video 1) might be attributed to the differences in source organelles (central large vacuole or small vacuoles/lipid vesicles) (Fig. 4C). The small exudates might come from vacuoles or lipid vesicles of vegetative cells due to the rapid increase in permeability, as pointed previously2 (Fig. 4D). The rice exudates likely pass through the empty space in the exine in pollen grains (see inset in Fig. 4C). And thereafter, the exudates were stored at the interface, causing a reduction in Vpollen (Fig. 4D).
Successful pollen adhesion and subsequent pollen hydration are ensured, then greater number of germinated grains would result in higher spikelet fertility. The amount of pollen exudation might depend on the vacuolar size, intercellular logistic, and kinetic (potential) energy at shedding, together with pollen fertility suggested by Watanabe2,3. If this were the case in grass plants, pollen weight, pollen velocity, and the travel distance at pollen dispersal should be the critical factors, which highly influenced by environmental conditions. Heslop-Harrison5 interpreted that pollen exudation might be attributed to the alterations in the membrane properties in the vegetative cells. Direct measurement of cell hydraulic conductivity is technically demanding at present partially due to the high viscosity; however, it should be emphasised that collision to certain object, such as stigma or oil appears to be essential for the pollen exudation, independent from stigmatic water flow (Figs. 1 and S1). Elastic deformation pointed above might instantly alter turgor in the vegetative cells upon pollen capture (Fig. 4B). Given the fact that cytoplasmic streaming velocity maintains even when turgor is changed at mild salt stress26, the presumed intracellular logistic might not be influenced by the pressure difference. Considering that pollination occurs even with the distance of few millimeters inside of the closed rice spikelet24, Eg threshold could be considerably small, and individual pollen grains are likely to perceive such a small kinetic energy. Further research will be required for understanding the pollen exudation process.
It can be pointed that pollen exudation might be a trait as the early step of pollination in the cultivated rice that have attained to achieve rapid self-pollination by selective breeding during long-term domestication. In addition to the major roles on foot formation and tube growth of lipids identified in the exudates, lipids in rice exudates may have another impact on allergic inflammation46. It is anticipated that pollen exudation might aggravate pollen allergy symptom, which is currently predicted to be more serious under the influence of climate change47. Currently, establishing steady production of grass plants as staple food is urgent in global food security under climate change. Whether heat tolerant cultivars have similar exudation patterns remains unknown. These cultivars may have a potential to germinate without relying on the stigmatic water. Much efforts are needed to answer these questions. In this study, we unveiled the dynamics and chemical composition during rice pollen exudation. We conclude that pollen exudation followed by the related roly-poly toy-like motion accounts for the optimal pollen adhesion that ensures rapid germination in rice. Considering the early reports2,3 and our findings, similar mechanisms may broadly exist in other grass plants with hydrating pollen grains. Therefore, further work from this perspective might contribute to the development of breeding in grass plants under climate change.