The absorption of water from humid air by grass embryos during germination

Abstract

Grass embryos possess structures that do not occur in any other flowering plants. Due to the specific embryo structure and position, grass embryo surfaces may be exposed to surrounding air under partial caryopsis–soil contact conditions, but whether caryopses of the grass family (Poaceae) can sense soil air humidity to initiate successful germination under partial caryopsis–soil contact conditions remain unknown. Here, we found that grass embryos have the unique ability to absorb water from atmospheric water vapor under partial caryopsis–soil contact conditions. To absorb atmospheric moisture, grass embryos developed profuse and highly elongated hairs on the embryo surface. These hairs, classically known as coleorhiza hairs, developed only on the embryo surface exposed to humid air, and submergence of the embryo surface inhibited their development. In addition to humid air-dependent development, almost all other developmental features of coleorhiza hairs were substantially different from root hairs. However, coleorhiza hair development was regulated by ROOTHAIRLESS 1. Besides the genetic control of coleorhiza hair development, we also identified how caryopses manage to keep the hairs turgid in natural open environments as the hairs were highly sensitive to dry air exposure. Moreover, we video-documented the regulation of developmental processes. The unique humid air-dependent coleorhiza hair development and their ability to absorb water from water vapor present in microsites or soil air give grasses advantages in germination and seedling establishment. Ultimately, coleorhiza hairs may have contributed to the ecological success of the grass family.

Grass embryos have the unique ability to absorb water from humid air to promote germination under partial caryopsis–soil contact conditions via specialized hygro-adaptive hairs on the embryo surface.

Introduction

The grass family (Poaceae) is perhaps the most successful plant family regarding its global occurrence, ecological dominance, and species richness (Gibson, 2009; Linder et al., 2018). Grasses occupy a greater area of the world’s land surface than any other plant family (Bradshaw, 1998; Gibson, 2009; Linder et al., 2018). The ecological success of the grass family is often explained with traits/features like breeding system plasticity, effective seed dispersal, rapid establishment, environmental flexibility, competitiveness, etc. (Bradshaw, 1998; Gibson, 2009; Linder et al., 2018). However, grass embryos possess structures that do not occur in any other flowering plants. Do the specialized embryo structures provide grasses any advantages in caryopses germination and seedling establishment?

Before seed dispersal, grass embryo has leaves, vascular tissue, clear shoot, and root meristems (Itoh et al., 2005). Thus, the highly differentiated grass embryo looks much more like a seedling than the embryos of nongrass relatives (GPWG, 2001). Besides the radicle and plumule, several embryo-specific organs/tissues such as scutellum, coleorhiza, coleoptile, etc. are present in the grass embryo. The functions of these embryo-specific organs/tissues are described in storage, mobilization of endosperm reserves, protection of the embryonic axis, and dormancy maintenance (Barrero et al., 2009; Gibson, 2009; Kellogg, 2015; Holloway et al., 2021).

Small-scale soil surface features like cracks, depressions, or sites associated with plant litter, etc. may lead to partial seed/caryopsis–soil contact (the common term grass “seed” is technically fruit or caryopsis). Under partial caryopsis–soil contact conditions, embryo region of grass caryopsis may be exposed to surrounding air, due to the specific embryo structure and position. Grass embryo surface, particularly coleorhiza, epiblast, lateral, and ventral scales are situated just beneath the caryopsis coat. Therefore, after caryopsis coat rupture, epidermal cells of coleorhiza, epiblast, etc. may be exposed to the surrounding air under partial caryopsis–soil contact conditions.

Atmospheric humidity influences numerous physiological processes of living organisms by modifying water status. Numerous insects and cacti utilize specialized microstructures to absorb water from atmospheric humidity (Malik et al., 2014). The molecular mechanisms of humidity sensation (hygro-sensation) in insect have begun to be unraveled (Liu et al., 2007; Enjin et al., 2016; Knecht et al., 2017). The possession of hygro-sensory systems allows insects to seek out environments of appropriate humidity to avoid desiccation. Being sessile organisms, plants must have to adapt with dynamic environments with developmental plasticity. Despite extensive research on stomatal responses to atmospheric humidity, more specifically leaf-to-air vapor pressure deficit (VPD), how guard cells perceive and respond to changes in VPD have been a matter of debate, mainly whether stomatal response is passive or active (Pantin and Blatt, 2018). In that context, the sensation of air humidity and absorption of water from atmospheric water vapor by grass embryos under partial caryopsis–soil contact conditions are still in its infancy.

Water vapor present in soil air can be important source of water for imbibition (Wuest et al., 1999; Wuest, 2002, 2003). Unlike atmospheric humidity, the relative humidity (RH) of soil air, even at permanent wilting point remains very high (Wuest, 2003). During the daytime, soil air humidity may decrease <100%; at night, the RH of soil air will again rise to saturated or near-saturated level. Therefore, the aim of the study is to identify whether the specialized embryo structures, particularly embryo surface, provide any advantage in caryopses germination and seedling establishment. More specifically, the question addressed is: “Do grass embryo epidermal cells have ability to sense and absorb water from atmospheric water vapor under partial caryopsis–soil contact conditions?”

Land plants utilize root hairs for water uptake, nutrient acquisition, anchorage, and microbe interaction (Hofer, 1996). Root hairs increase the absorptive surface area of roots and improve contact between roots and the soil. Over the past several decades, substantial progress has been made in our understanding of the genetic control of root hair development, particularly in the dicot model plant, Arabidopsis (Arabidopsis thaliana). Several dozens of genes involved in root hair development have been characterized in Arabidopsis (Shibata and Sugimoto, 2019). Unlike Arabidopsis, only a few genes have been characterized with functions in root hair development in rice (Oryza sativa). It includes O. sativa cellulose synthase-like D1 (OsCSLD1; Kim et al., 2007), formin homology 1 (OsFH1; Huang et al., 2013b), sec14- nodulin domain-containing protein 1 (OsSNDP1; Huang et al., 2013a), xyloglucan 6-xylosytransferase 1 (OsXXT1; Wang et al., 2014), ROOTHAIRLESS 1 (OsRHL1; Ding et al., 2009), ALPHA-EXPANSIN 17 (OsEXPA17; Yu et al., 2011), etc. Generally, mutants of these genes display a common (i.e. short root hairs) phenotype.

The development of hairs on the epidermal cells of coleorhiza (hence called coleorhiza hairs) has been reported in dozens of grass species during caryopsis germination (Nishimura, 1922; Morita et al., 1997). The function of the coleorhiza hairs was revealed in anchorage (Morita et al., 1997). Similarly, epidermal cells of epiblast of some grasses also develop profuse and elongated hairs, like coleorhiza (Foard and Haber, 1962). The identity of epiblast was a contested topic for more than a century; some scientists considered epiblast as a second cotyledon, others as a rudimentary leaf, or an outgrowth of the coleorhiza (Coulter, 1915; Percival, 1921; Avery, 1930; McCall, 1934; Roth, 1957; Brown, 1960; Foard and Haber, 1962). Scientists who consider epiblast as leaf-like structure also consider the epiblast hairs as a specialized form of leaf trichomes (Tillich, 1977, 2007). Remarkably, the differentiation of trichomes and root hair/hairless cells is regulated by similar molecular regulators in Arabidopsis (Ishida et al., 2008). However, the effect of humidity on the development and functions of these epidermal structures is scarce.

In our recent transcriptome analysis, we have observed elevated expressions of several root hair-specific genes in rice caryopsis under partial caryopsis–water conditions (Song et al., 2021), although the presence of radicle in the samples confounded the results. Numerous previous studies suggest that a similar gene regulatory network controls the development of similar structure of plants (Ding et al., 2009; Karas et al., 2009; Tam et al., 2015; Proust et al., 2016; Moon et al., 2019). Therefore, we have hypothesized that the development of hairs on embryo epidermal surface (i.e. on coleorhiza, epiblast, lateral, or ventral scales) might be a hygro-adaptive response of grass caryopses under partial caryopsis–soil contact conditions where the hair development might be also regulated by the core regulator of root hair development.

Here, we first identify the responses of humidity sensation of grass embryos under partial caryopsis–water contact conditions. We then study the genetic control of the hairs developed on the embryo surface and test our hypothesis whether the hair development is also regulated by the core regulator of root hair development. Then we investigate how caryopses manage to keep the hairs turgid in natural open environments. Finally, we examine the influence of embryo–water contact on caryopsis germination and seedling establishment.

Results and discussion

Humidity sensation and the development of hygro-adaptive hairs on the embryo surface

Since water is the only requirement for rice caryopses germination, and the amount of water in the air (humidity) can be manipulated, therefore, humidity sensation and absorption of atmospheric water vapor can be resolved through the success/failure of caryopses germination under partial caryopsis–water contact conditions. To identify the effect of partial caryopsis–water contact on caryopses germination, opposite caryopsis–water contact conditions were set up using different methods. In these setups, half of the rice caryopsis surface was in contact with water (either directly or through agarose gel), the remaining half remained either in relatively dry air (RH ∼55%) or humid air (RH >98%) (Figure 1A). Different partial caryopsis–water contact conditions were named after the location of embryo region of caryopses and humidity conditions whether embryo placed in water (EIW), embryo placed in dry air (EIDA), or embryo placed in humid air (EIHA) (Figure 1A).

Figure 1.

Humidity sensation and the development of coleorhiza hairs on the grass embryo surface. A, Setup of partial caryopsis–water contact conditions in the glued-caryopsis technique (see “Materials and methods” for other techniques). B, The effect of opposite caryopsis–water contact orientations (EIW and EIDA) on caryopses germination in low humidity condition (RH ∼55%) (n = 15). C, Water absorption pattern of EIDA- and EIW-oriented dehulled caryopses in low humidity condition (RH ∼55%) (n = 3). D, Change of water content of imbibed caryopses after placement in EIDA and EIW orientations in low humidity condition (RH ∼55%) (n = 3). E, The effect of EIHA and EIW orientations on caryopses germination in high humidity condition (RH > 98%) (n = 15). F, Water absorption pattern of EIDA and EIHA-oriented dehulled caryopses (n = 3). G, Time-lapse images of a germinating caryopsis under EIHA condition. H, Longitudinal section view of EIHA-oriented rice embryo before radicle or coleoptile emergence. I–N, The development of coleorhiza hairs under EIHA condition across rice sub-populations. I, Indica. J, Aus. K, Ashwina. L, Rayada. M, Aromatic. N, Japonica. O–R, The development of coleorhiza hairs in cultivated and wild species of Poaceae. O, Bluegrass (Poa pratensis). P, Barley (Hordeum vulgare). Q, Ryegrass (L. perenne). R, Wheat (T. aestivum). S–T, Difference between coleorhiza and root hairs on (S) hair elongation pattern (n = 3) and (T) length of hairs (n = 25). U, The unimpaired development of coleorhiza hairs in the embryo-removed caryopsis. V, The development of coleorhiza hairs on the outer epidermal surface. W, The ability of isolated coleorhiza, epiblast, and scales to develop coleorhiza hairs. X, The retention of the hairline despite manual extension of exposed area. Y, The development of profuse and highly elongated coleorhiza hairs in the primary rootless mutant, nrtp1 (For all representative images, n ≥ 3; (B), (E), and (T), Mann–Whitney test; (C), (D), (F), and (S), Wilcoxon matched-pairs signed-rank test; NS, not significant; **P < 0.01; ***P < 0.001 and all error bars show ±sd, all scale bars = 1 mm.).

Irrespective of air humidity and the presence/absence of hull, the EIW-oriented caryopses efficiently germinated (Figure 1B), due to direct contact of embryo with water and faster absorption of water through the embryo-end (Figure 1C). Oppositely, irrespective of the presence/absence of the hull, the EIDA-oriented caryopses completely failed to germinate (Figure 1B), even after 15 days of partial caryopsis–water contact. To decipher why the EIDA-oriented caryopses failed to germinate, fully imbibed (24 h imbibition) caryopses at slight emergence of coleoptile were placed under EIDA condition and captured time-lapse images. Time-lapse microscopy showed the contraction of coleoptile (Supplemental Figure S1 and also see Supplemental Movie S1) due to loss of water. Similarly, after placement of fully imbibed caryopses under EIDA condition, rapid loss of water was observed (Figure 1D), although recovered over time.

The embryo in humid air (EIHA)-oriented caryopses, on the contrary, efficiently germinated (Figure 1E) due to the significant increase in water absorption (Figure 1F). However, the slower absorption through the chalazal-end (Horigane et al., 2006) and the presence of hull over grain, that is, barrier of evaporation, complicated the explanation of the huge difference in water absorption between the EIDA- and EIHA-oriented caryopses, especially from 72 h and onwards (Figure 1F), without the absorption of water from humid air. To identify whether the exposed embryo surface sensed high air humidity and absorbed water from humid air, time-lapse images of germinating EIHA-oriented dehulled caryopses were captured. Time-lapse microscopy showed that the entire epidermal surface of the EIHA-oriented caryopses developed profuse hairs (Figure 1G). Moreover, the same caryopsis that earlier failed to germinate under EIDA condition also developed profuse and highly elongated hairs after placement under EIHA condition (Supplemental Figure S1).

Coleorhiza hair development under EIHA conditions is a global phenomenon across Poaceae

Under EIHA condition, the development of hairs on rice embryo surface was not only limited to coleorhiza, but rather profuse hairs developed on the entire exposed epidermal surface (Figure 1, G and H). To avoid confusion, we also used the term “coleorhiza hair” for the hairs developed on the embryo surface beyond coleorhiza. Remarkably, coleorhiza hairs always developed before coleoptile or radicle emergence (Figure 1H). Therefore, coleorhiza hair was the first and foremost developmental structure in caryopses germination under EIHA condition.

We examined caryopses of dozens of rice varieties and numerous wild and cultivated species of Poaceae to ascertain whether the development of coleorhiza hairs under EIHA condition is a global phenomenon across Poaceae. The development of profuse and highly elongated coleorhiza hairs was observed in all types of rice cultivars/accessions across rice sub-populations (Figure 1, I–N). Similarly, we observed coleorhiza hair development in all grass species we examined (Figure 1, O–R), except maize (Zea mays). In some grasses like ryegrass (Lolium perenne), the length of coleorhiza hairs was often more elongated than the length of caryopses (Figure 1Q). However, the number and length of coleorhiza hairs showed wide variations across species (Figure 1, I–R). In some species like wheat (Triticum aestivum), the development of coleorhiza hairs was confined to coleorhiza (Figure 1R).

Coleorhiza hairs are several folds longer than root hairs

The duration of hair elongation and the length of coleorhiza hairs were distinct from that of root hairs. However, the elongation patterns of both of the hairs were almost the same in their initial developmental hours (Figure 1S). The duration of hair elongation was significantly different between them. The elongation of root hairs usually finished within a couple of hours (Figure 1S). However, coleorhiza hairs can elongate even after 48 h from hair initiation. Therefore, coleorhiza hairs were several folds longer than root hairs (Figure 1T). However, the diameter of both root and coleorhiza hairs was almost the same (coleorhiza hairs = 7.24 ± 0.65 µm, root hairs = 7.18 ± 0.84 µm, Mann–Whitney test, P = 0.713, n = 15).

Coleorhiza hair development is independent of the presence of the radicle

Since radicle resides only a few cell layers beneath the embryo epidermal cells, therefore, to ascertain the involvement of radicle on coleorhiza hair development, radicle-removed embryo (by micro-dissection), and a primary rootless mutant were analyzed. Micro-dissectional approach revealed that the development of coleorhiza hairs was independent of the presence of radicle (Figure 1U) and the hairs only developed on the outer epidermal cells (Figure 1V). Consequently, the unimpaired development of coleorhiza hairs was observed on the isolated coleorhiza, epiblast, and lateral and ventral scales (Figure 1W). However, the hairline remained the same, despite manual extension of exposed area (Figure 1X). The unimpaired development of coleorhiza hairs in the primary rootless mutant, nrtp1 (Yu et al., 2018) confirmed radicle-independent development of coleorhiza hairs (Figure 1Y).

The genetic control of coleorhiza hair development

To test our hypothesis, we initially tried to find out the core regulator of root hair development. Numerous previous studies suggested that a similar gene regulatory network controls the development of similar structure of plants such as rhizoids in moss and root hairs on the roots of vascular plants (Ding et al., 2009; Karas et al., 2009; Tam et al., 2015; Proust et al., 2016; Moon et al., 2019). Therefore, we not only tried to find out the key regulator of root hair development, but also searched for the core regulator of tip-growing filamentous cells with rooting functions beyond angiosperm. We found two core classes of genes, both belonging to the basic helix–loop–helix transcription factor family: LOTUS JAPONICUS ROOTHAIRLESS1-LIKE (LRL) and ROOT HAIR DEFECTIVE SIX-LIKE (RSL) that regulate tip-growing filamentous cells with rooting functions across land plants (Ding et al., 2009; Karas et al., 2009; Tam et al., 2015; Proust et al., 2016; Moon et al., 2019).

Among rice LRL and RSL type genes, only OsRHL1 (Os06g08500) and OsRSL6 (Os07g39940) were functionally characterized in root hair development (Ding et al., 2009; Moon et al., 2019). Both of their mutants showed short root hair phenotypes. The reduction of root hair lengths of rhl1 and rsl6 was 89% and 62%, respectively, to their corresponding wild-types (WTs; Ding et al., 2009; Moon et al., 2019). We also compared the reduction of root hair lengths of other rice root hair mutants (Kim et al., 2007; Yu et al., 2011; Huang et al., 2013a, 2013b; Wang et al., 2014). Remarkably, the rhl1 showed the severest reduction of root hair length among all reported rice root hair mutants, viz. fh1(84%), sndp1(80%), xxt1(73%), csld1(68%), and expa17(68%). Recently, RHL1 was also shown to interact with most of the RSL class II subfamily members (Moon et al., 2019). Moreover, LRL/RHL1 homologs showed conserved function in monocot (Oryza) (Ding et al., 2009), dicot (Arabidopsis, Lotus) (Karas et al., 2009) even in moss (Physcomitrium patens) (Tam et al., 2015). Altogether, RHL1 was the best candidate to test our hypothesis whether coleorhiza hair development is also regulated by the core regulator of tip-growing filamentous cells with rooting functions.

To examine our hypothesis, the reporter gene, β-glucuronidase (GUS) was expressed under control of RHL1 promoter. The expression of GUS was observed on the exposed epidermal tissues immediately after the hull and caryopsis coat rupture (Figure 2A). At the end of caryopsis coat rupture, almost all cells of the epidermal surface turned blue (Figure 2B). The expression of GUS in both coleorhiza and root hairs (Figure 2C) suggested the involvement of RHL1 in both coleorhiza and root hair development. To revalidate the specific staining pattern of RHL1:GUS, we also expressed another reporter gene, green fluorescent protein (GFP) under control of RHL1 promoter. In accordance with the GUS staining pattern, the entire epidermal surface exposed to humid air showed GFP fluorescence (Figure 2D), before either radicle or coleoptile protrusion (Figure 2E). Completely germinated caryopses predominantly showed GFP fluorescence on coleorhiza hairs developed around the coleoptile (Figure 2F), because the empty glume sheath (Figure 2F arrow) restricted the elongation of hairs that developed on coleorhiza and part of epiblast.

Figure 2.

RHL1 promoter-driven GUS and GFP expression. A–C, RHL1 promoter-driven GUS expression pattern. A, At the beginning of hull and caryopsis coat rupture. B, At the end of caryopsis coat rupture. C, After caryopses germination. D–F, RHL1 promoter-driven GFP expression pattern shown in three panel sets (bright field, BF; fluorescent, F, and merged, M). D, At the end of caryopsis coat rupture. E, Longitudinal section view of rice embryo before radicle and coleoptile protrusion. F, After caryopses germination. Arrow indicates the empty glume sheath. (Scale bars = 1 mm).

Although both GUS and GFP expression patterns suggested the involvement of RHL1 in coleorhiza hair development, however, we also examined the phenotype of the mutant (rhl1) for further evidence. The rhl1 mutant predominantly showed coleorhiza hairless phenotype (Figure 3, A and B). However, the development of few and/or short coleorhiza hairs was also observed in some mutant caryopses. To establish more solid evidence, we also examined the phenotype of RHL1 complemented/overexpression line. The RHL1 complemented/overexpression lines were generated by transforming WT RHL1 into rhl1-1 under control of 35S promoter (Ding et al., 2009). Therefore, transgenic rice plants of rhl1-1 over-expressing WT RHL1 served as both complemented and overexpression line. In the overexpression/complemented line, WT phenotype was not only restored but the length of coleorhiza hairs was significantly increased (Figure 3, C and D).

Figure 3.

Comparative phenotypes of mutant and complemented/overexpression lines of RHL1. A, WT. B, rhl1 mutant, and (C) RHL-complemented/overexpression line after caryopses germination. D, Comparison of the length of coleorhiza hairs between WT and RHL1 complemented/overexpression line (n = 5). D, Mann–Whitney test; *P < 0.05, error bars show ±sd, scale bars = 1 mm.

The role of rice OsRHL1 and its multiple homologs [in Lotus (Lotus japonicus) (LjRHL1), Arabidopsis (A. thaliana) {AtLRL1 (At2g24260), AtLRL2 (At4g30980), and AtLRL3 (At5g58010)} and in the moss P. patens (PpLRL1 and PpLRL2)] on root hair and rhizoid development was unequivocally established in the previous studies (Ding et al., 2009; Karas et al., 2009; Tam et al., 2015). Similarly, coleorhiza hairless phenotype of the rhl1 mutant (Figure 3, A and B), increased coleorhiza hair length of the complemented/overexpression line (Figure 3, C and D), tissue-specific expression of two reporter genes (GUS and GFP) on embryo surface and coleorhiza hairs (Figure 2, A–F) altogether undoubtedly established that the OsRHL1 was also involved in the development of coleorhiza hairs.

Hair formation and epidermal patterning

Hair and hairless epidermal cells in root of rice and other cereals alternate along each longitudinal file where around 50% of epidermal cells develop root hairs and remaining 50% cells remain hairless (Dolan, 2017). However, the density of coleorhiza hair cells was more than double than that of root hairs, irrespective of locations (Figure 4, A–C). Every cell around the ventral and lateral scales often developed coleorhiza hairs (Figure 4C). There was no differentiation on hair cells and non-hair cells. Similar pattern of coleorhiza hair development (i.e. every cell developing hairs) was also observed in ryegrass.

Figure 4.

Comparative density of hair cells between root and embryo surface. A, The representative bright field (BF), fluorescent (F), and merged (M) images of transgenic rice (RHL1:GFP) roots. B, Comparison of the density of hair cells between root and embryo surface. Root (adj.) and root (tip) indicate root adjacent to coleorhiza and root-tip, respectively (n = 10). C, The development of coleorhiza hairs in every cell of the exposed embryo surface. Arrow indicates the empty glume sheath. B, Mann–Whitney test, ***P < 0.001 and error bars show ±sd, scale bars = 1 mm.

The high sensitivity of coleorhiza hairs to dry air exposure

To identify the effect of humidity, more specifically, humidity fluctuations on the structure and function of coleorhiza hairs, we exposed coleorhiza hairs to relatively dry air (RH ∼50%) by removing the lid of petri dish/humid chambers. Unlike leaf trichomes, coleorhiza hairs were highly sensitive to dry air exposure. Coleorhiza hairs shriveled/collapsed immediately after the exposure to dry air. To identify the extent of the structural changes of coleorhiza hairs, we captured videos of coleorhiza hairs under microscope after removing the lid. We also analyzed the change of humidity and temperature of the humid chambers.

Under normal growing conditions (RH >98%), coleorhiza hair looks like an elongated tube, and the normal tubular structure was considered fully turgid reference state (Figure 5A). After dry air exposure, turgid hairs deformed with rotation into a dry fibrous form (Figure 5A). Due to the shrinkage and twist, hairs lost its straight-line like structure (see Supplemental Movie S2 and Supplemental Figure S2A). We analyzed the videos to estimate the rate of shrivel/collapse. The rate of shrivel/collapse was estimated in an arbitrary unit (AU) in the form of percentage of shrivel/collapse based on time context relative to rotation numbers (Figure 5A).

Figure 5.

The high sensitivity of coleorhiza hairs to relatively dry air exposure. A, Illustration of the method used for the estimation of shrivel/collapse with an AU on time-context relative rotation numbers of coleorhiza hairs. Drawings of the hair are not to scale, nor in the order and position of rotations. B, Shrivel/collapse of coleorhiza hairs (also see Supplemental Movie S2) (r stands for Pearson correlation coefficient), (n = 10 hairs).

The progressive decline of RH of humid chamber/petri dish was observed after the removal of lid. However, the average RH remained over 90% for the first minute (Supplemental Figure S2B). The changes of RH of humid chamber depended on the changes of temperature (Supplemental Figure S2C). Video analysis showed that coleorhiza hairs shriveled/collapsed within a minute of exposure to dry air (Figure 5B and Supplemental Figure S2A, also see Supplemental Movie S2). Remarkably, the shriveled/collapsed hairs did not regain turgor after returning to high humidity conditions.

Two natural phenomena facilitate coleorhiza hair development in natural environments

We were interested to identify the natural habitat(s) where grasses get suitable humidity conditions for coleorhiza hair development. Moreover, we were also interested to identify how grasses manage to keep coleorhiza hairs turgid in natural open environments, since coleorhiza hairs are highly sensitive to dry air exposure (Figure 5B). Among the habitats of wild rice, clay-rich floodplains often naturally develop soil cracks (Figure 6A). Soil cracks, depressions, etc. may lead to partial caryopsis–soil contact. Besides soil cracks, other small-scale soil surface features like crevices, depressions, sites associated with gravel, and plant litter, etc. are microsites suitable for seed germination and seedling establishment (hence also known as safesites); this phenomenon was reported in many species including grasses (Harper et al., 1965; Winkel et al., 1991). Therefore, we examined diurnal temperature and RH patterns of air inside soil cracks.

Figure 6.

The natural occurrence of high humidity inside soil cracks/microsites and decreasing night-time temperature facilitates coleorhiza hair development in natural environments. A, Soil cracks—common microsites in clay-rich floodplain. B, Diurnal temperature and humidity patterns of air inside a soil crack situated 5 cm below soil surface on a random day. Shaded area represents nighttime. C, Diurnal temperature and humidity patterns of air inside mimicked soil cracks (made of agar-blocks) (n = 7). D, The relationship between diurnal changes of RH with that of the air temperature inside microsites (r stands for Pearson correlation coefficient) (n = 45). E, The relationship between diurnal changes of temperature with the duration of saturated humidity hours of microsites (n = 45). F, The relationship between the decrease of night temperature with the duration of saturated humidity hours inside microsites (n = 45). G and H, Induced condensation under microscope mimicking microsites and decreasing night-time temperature conditions on (G), caryopsis surface and (H), coleorhiza hairs. (Scale bar = 500 µm). I, Droplet growth pattern (n = 3) (error bar ± sd).

Time-lapse datalogging of temperature and RH of air inside wet soil cracks (microsites) (Figure 6B) and mimicked soil cracks (made of agar blocks) (Figure 6C) showed similar humidity patterns. We observed high air humidity inside both natural and mimicked soil cracks (microsites). The fluctuations of RH mostly occurred during the daytime, whereas RH largely remained at saturated level at night. The change of diurnal RH inside wet microsites (soil cracks) was associated with the change of diurnal air temperature (Figure 6D). However, the duration of saturated (RH 100%) humidity hours was not associated with the change of diurnal temperature (Figure 6E), but with the decrease of night temperature (Figure 6F).

At night, air temperature inside wet soil cracks/microsites continued to decrease for couple of degrees after reaching saturated humidity (around 2.7°C during the experimental periods). The continued decrease of temperature after reaching saturated humidity will force condensation/dew formation as the air inside soil cracks/microsites cannot hold any more water vapor; consequently, dew will appear. The mimicked decrease of night-time temperature (∼3°C) of humid chamber under microscope showed condensation or dew formation on both grain surface and coleorhiza hairs (Figure 6, G and H). The condensation speed was rather fast as within a few minutes, visible water droplets condensed on the tip of coleorhiza hairs (Figure 6I). Besides the specialized embryo structures, grasses utilized two natural phenomena for the development of coleorhiza hairs under partial caryopsis–soil contact conditions in microsites: first, the natural occurrence of high humidity inside wet microsites and second, the natural night-time temperature decrease.

The extent of embryo–water contact influences coleorhiza hair development

We designed four micromanipulated embryo–water contact conditions (Figure 7A) to identify how the extent of embryo–water contact influences coleorhiza hair development. In these conditions, embryo surface was placed either under humid air or under water. Caryopsis placement under only EIHA (OEIHA) condition was identical to that of EIHA set-up in the water-tube method, except the angular position of caryopsis and the status of the stopper (Figure 7A, also Supplemental Movies S3 and S4). Under OEIHA condition, therefore, embryo surface of caryopsis was only under humid air, whereas the rest of the caryopsis surface was in contact with water. In the other two conditions, that is, embryo in contact with water (EICWW) and embryo under water (EUW), embryo surface of caryopsis was under water, either by enveloping with thin layer of water or submergence, respectively (Figure 7A, see “Materials and methods” and Supplemental Movies S5 and S6).

Figure 7.

The extent of embryo–water contact influences coleorhiza hair development. A, Schematic outline of the set-up of micromanipulated embryo–water contact conditions. (see Supplemental Movies S3, S4, S5, and S6). B, The effect of embryo–water contact on number of coleorhiza hairs. For box plots, the center line, tiny squares, limits, whiskers, and points correspond to the median, mean, upper and lower quartiles, 1.5× interquartile range and outliers, respectively. (n = 10). C, The effect of embryo–water contact on length of coleorhiza hairs (n = 10). D and E, Revalidation of the inhibition of coleorhiza hair development after submergence of embryo surface under (D) EICWW and (E) EUW conditions. (For representative images, n ≥ 2. B and C, Mann–Whitney test; ***P < 0.001, scale bars = 1 mm).

Interestingly, the exposure of embryo surface to humid air always developed coleorhiza hairs, irrespective of embryo–water contact areas (Figure 7B). However, the length of coleorhiza hairs depended on the extent of the embryo–water contact (Figure 7C; Supplemental Figure S3, A and B). Under OEIHA condition, the length of coleorhiza hairs at radicle emergence stage was remarkably shorter than that of EIHA condition (Figure 7C; Supplemental Figure S3, A and B and see Supplemental Movies S3 and S4). Irrespective of grass species, humid air-grown roots continue to develop root hair after contact with water. However, submergence of embryo surface inhibited coleorhiza hair development under EICWW and EUW conditions (Figure 7, B and C; Supplemental Figure S3, C and D, and see Supplemental Movies S5 and S6).

The development of few coleorhiza hairs was occasionally observed in EICWW-oriented caryopses (Figure 7C). However, completely submerged caryopses under EUW condition never developed coleorhiza hairs (Figure 7, B and C; Supplemental Figure S3, C and D). Therefore, we revalidated the underwater inhibition of coleorhiza hair development using dehulled RHL1:GUS caryopses (Figure 7, D and E). Under EICWW condition, a few coleorhiza hairs developed on lateral/ventral scale margin of RHL1:GUS caryopses only if cells of the lateral/ventral scale margin lost direct contact with water during coleoptile emergence (Figure 7D). Remarkably, submergence did not restrict root hair development (Figure 7, D and E) although the same RHL1 regulated the development of both coleorhiza and root hairs.

We also examined the phenotype of caryopses of RHL1 overexpression line under EUW condition to see whether overexpression of RHL1 is sufficient to induce underwater coleorhiza hair development. Like WT caryopses, caryopses of RHL1 overexpression line did not develop coleorhiza hair under EUW condition (Supplemental Figure S4). Complete submergence of embryo surface (Figure 7, B, C and E; Supplemental Figure S4), even covering/enveloping of embryo surface with thin layer of water (Figure 7, B and C; Supplemental Movie S5), inhibited coleorhiza hair development. Simply, the extent of embryo–water contact significantly influenced coleorhiza hair development. Overall, the development of coleorhiza hairs was largely regulated by environmental factors like air humidity, caryopsis–water contact, etc.

Absorption of atmospheric water vapor/dew accelerates caryopses germination and seedling establishment

Since coleorhiza hairs significantly increased the exposed embryo surface (Figure 8A), the increased surface area may harvest more dew water or capture more water from atmospheric water vapor. To see whether coleorhiza hairs do actually harvest dew water or capture water from atmospheric moisture, saturated humidity and both steady and decreasing night-time temperature conditions were mimicked under microscope. Under steady temperature and saturated humidity conditions, dew droplets often condensed on the tip of coleorhiza hairs. Condensed droplet size gradually increased, occasionally coalesced. The fate of the condensed dew droplets was revealed after conversion of time-lapse images into fast-motion videos (see Supplemental Movie S7), because the absorption of dew droplet was very slow process. We estimated the droplet absorption time (by disappearance), and usually it took >30 min for a droplet to disappear after reaching its peak size (Figure 8B, see Supplemental Movie S7).

Figure 8.

Absorption of atmospheric water vapor/dew accelerates germination and seedling establishment. A, Comparison of surface area with (+ CH) and without (− CH) coleorhiza hairs. (Mann–Whitney test; ***P < 0.001, n = 25, error bar ±s.d). B, The fate of dew drop condensed on tip of coleorhiza hair in steady temperature condition. Arrow indicates the direction of droplet movement. Number on each figure represents minutes. The time point of the highest droplet diameter was considered zero (0) minute. Scale bar = 100 µm (also see fast-motion Supplemental Movie S7). C, The accumulation of large dew drops on coleorhiza hairs after mimicked decreasing night-time temperature condition. Scale bar = 1 mm (also see fast-motion Supplemental Movie S8). D, The effect of two caryopsis–water contact conditions (EIHA and OEIHA) and two RHL1 genotypes (WT and rhl1 mutant) on caryopsis coat rupture to radicle emergence time. For box plot, the center line, tiny squares, limits, whiskers, and points correspond to the median, mean, upper and lower quartiles, 1.5× interquartile range and outliers, respectively (n = 5) (Also see Supplemental Movies S3, S4, S9, and S10). Two-way analysis of variance (caryopsis–water contact × RHL1) showed a significant main effect of caryopsis–water contact (F_{(1, 16)} = 78.60, P < 0.0001) and RHL1 (F_{(1, 16)} = 10.35, P = 0.0054) and a significant interaction between factors (F_{(1, 16)} = 5.904, P = 0.0273), followed by Turkey’s multiple comparison tests (WT EIHA versus rhl1 EIHA, P = 0.0052; WT OEIHA versus rhl1 OEIHA, P = 0.9432; WT EIHA versus WT OEIHA, P = 0.0017; rhl1 EIHA versus rhl1 OEIHA, P < 0.0001).

The progressive accumulation of dew drops on coleorhiza hairs was observed in the mimicked decreasing night-time temperature conditions (Figure 8C and see Supplemental Movie S8). The condensed dew drops accumulated on coleorhiza hairs (Figure 8C) were more than sufficient for seedling establishment as the slow progression of caryopses germination of coleorhiza hairless mutant (rhl1) was immediately ramped up after addition of tiny droplet of water adjacent to embryo surface (see Supplemental Movie S9).

Finally, we compared germination time, more specifically, time required for caryopsis coat rupture to radicle emergence between two caryopsis–water contact conditions (EIHA and OEIHA) and two RHL1 genotypes (WT and rhl1 mutant) (Figure 8D and see Supplemental Movies S3, S4, S9 and S10). Two-way analysis of variance showed highly significant main effect of caryopsis–water contact (71% of total variation), followed by RHL1 (9% of total variation) and a significant interaction between factors (Figure 8D).

Coleorhiza hairs: the same but different

Numerous developmental features of coleorhiza hairs were substantially distinct from root hairs. These included (1) the order of development: the first and foremost developmental structure of caryopses germination, even before coleoptile or radicle emergence (Figure 1, G and H), (2) site of the development: on the entire exposed embryo surface (Figure 1, G and H and U–Y), unlike root hairs that develop certain distance apart from root tip, (3) duration of hair elongation and length of hair: longer elongation period, hence several folds more elongated than root hairs (Figure 1, S and T), (4) radicle-independent development (Figure 1, U–Y), etc. Besides the developmental distinctiveness, the regulation of the developmental processes of coleorhiza hairs was also distinct from root hairs such as caryopsis–water contact-dependent elongation (Figure 7C; Supplemental Figure S3, A and B), submergence/water contact-dependent inhibition (Figure 7, B–E; Supplemental Figure S3, C and D), no differentiation of hair and hairless cells (Figure 4, B and C), etc. Despite all these differences, surprisingly, rice utilized the same OsRHL1 for coleorhiza hair development.

The high sensitivity of coleorhiza hairs to dry air exposure (Figure 5B), the natural occurrence of high humidity inside wet microsites (Figure 6B), the exceptional humid air-dependent coleorhiza hair development (Figure 1, G–R and U–Y), their developmental inhibition after direct water contact (Figure 7, B–E; Supplemental Figure S3, C and D), and the delayed germination of coleorhiza hairless rhl1 mutant (Figure 8D) altogether established the fact that coleorhiza hairs were specialized hygro-adaptive hairs developed to harvest water from atmospheric humidity and/or dew water to accelerate caryopses germination and seedling establishment. The development of coleorhiza hairs in every cell and their excessive elongation were perhaps to increase the water harvesting efficiency.

The unique ability to develop specialized hairs on embryo surface before germination and their ability to absorb water from water vapor present in microsites or soil air provide grasses advantages in germination and seedling establishment. Ultimately, coleorhiza hairs may have contributed, at least to some extent, to the ecological success of the grass family. Unlike other traits/features that have contributed to the ecological dominance, coleorhiza hair is the only feature that is limited to the grass family and common across the family members. The identification of hygro-adaptive hairs and their developmental regulations will improve our understanding how plants adapt in natural environments utilizing the core gene system, capitalizing on natural phenomena. In future, coleorhiza hairs and their developmental regulations can be also manipulated for practical applications.

Materials and methods

Plant materials and growth conditions

Rice (O. sativa), the model plant of the grass family, was selected for the study. Caryopses of aus variety, Kasalath were used in most of the experiments. Both of the mutants (nrtp1 and rhl1) were also isolated from the ethyl methylsulfonate mutagenized population of Kasalath (Ding et al., 2009; Yu et al., 2018). Similarly, all the transgenic plants (RHL1:GUS, RHL1:GFP, 35S:RHL1) were also developed in the Kasalath background. The RHL1-overexpression line (35S:RHL1) also served as the complemented line as WT RHL1 (under control of 35S) was transformed in to rhl1-1 mutant (Ding et al., 2009). Long-grain aromatic rice variety—Basmati, indica variety—Yangdao6, a special deep-water rice—Rayada (Bin Rahman and Zhang, 2013, 2018) were also used for the time-lapse microscopy. Dozens of rice varieties (from our preexisting collections) were examined for the global nature of coleorhiza hair development. These varieties and accessions were earlier collected from different rice research institutes/universities, and from our research collaborators. The caryopses of the wild and other cultivated species of Poaceae were purchased from local seed stores.

To avoid unintended interference of surface sterilizing agents, especially the oxidative agents on the embryo epidermal surface, no surface sterilizing agent was used to sterilize caryopses. Instead, caryopses were thoroughly rinsed with sterile water. The initial experiments of the effect of caryopsis–water contact orientations on caryopses germination were conducted in the climate chamber (HPP110, Memmert). The percentage of caryopses germination capacity was usually scored 5 days after caryopsis–water contact.

Setup of caryopsis–water contact conditions

A number of caryopsis–water contact conditions were designed for the study. The general outline of the micromanipulated caryopsis–water contact conditions was depicted in Figures 1, A and 7, A. The degree of caryopsis/embryo–water contact was manipulated under stereomicroscopes (SZ51 and SZX 16; Olympus, Tokyo, Japan). Generally, partial caryopsis–water contact conditions were set-up in a way that half of the rice caryopsis surface was in contact with water (either directly or through agarose gel), the remaining half remained in relatively dry air or humid air.

To specify the extent of dryness/wetness of air, we measured RH of air surrounding the caryopsis. The RH of air surrounding the caryopsis was either low (RH ∼55%) or high (RH >98%), depending on the experiments. Low humidity experiments were conducted in multiple open environmental conditions such as on lab bench (RH ∼55%), in growth room (RH ∼55%), microscopy room (RH ∼50%), and greenhouse (RH 55%–70%) conditions. The experiments of high humidity conditions were conducted in enclosed petri-dish or humid chambers. The RH inside the enclosed petri-dish or humid chambers was usually very high (RH >98%); mostly remaining at saturated or near-saturated level.

In the initial experiments, hulled and dehulled caryopses were vertically glued (using cyanoacrylate) on the margin of glass slide (25 × 76 × 1 mm) with four to eight caryopses per slide (see Figure 1A) where embryo region of half of the caryopses was in upward direction, and other half of the caryopses was in the opposite direction. The glued-caryopses slide was then horizontally placed on the margin of petri dish. Then the water-filled petri dishes with glued-caryopses slides were placed on undisturbed lab bench and in humid chambers. Thus, partial caryopsis–water contact conditions were set-up in the glued-caryopsis method, where half of the caryopsis surface was immersed in water (see Figure 1A) and the other half of the surface was in relatively dry air (RH ∼55%) or humid air (RH >98%). These partial caryopsis–water contact conditions were named after the location of embryo region of caryopses and the RH conditions of air, that is, EIW, EIDA, or EIHA.

The EIW, EIDA, and EIHA conditions were also replicated in gel set-up where agarose gel (0.4%, w/v) was kept in clear plastic zip-lock bag to avoid rapid desiccation in low humidity experiments. Caryopses were inserted through the prepunctured holes in a way that half of the caryopsis surface was in the solidified agarose gel, and the remaining half of the caryopsis remained in relatively dry air (RH ∼55%) or humid air (RH >98%).

Besides the slide- and gel-based set-up, horizontal partial caryopsis–water contact condition was also set-up using water tube for time-lapse imaging. In this set-up, a caryopsis was partially inserted into a customized tube end, keeping the embryo-end outside the tube. Initially, 1-mL tips (with appropriate cut-end to hold the caryopsis) were used as improvised tube. The tube was then filled with water and the flow of water was regulated with a stopper. The water-filled tube holding a caryopsis under partial caryopsis–water contact condition was inserted horizontally into the custom-made stage-top humid chamber. The Supplemental Movie S3 displayed the placement of rice caryopsis in tube-setup for EIHA condition.

The set-up of OEIHA condition was almost identical to the tube set-up of EIHA, except angular position of caryopsis and the status of the stopper. The caryopsis-inserted tube was slightly rotated (clockwise 30°–40°), and the stopper was released such that microflow of water was over the caryopsis surface, keeping only the emerging embryo surface in humid air. The Supplemental Movie S4 displayed the placement of rice caryopsis in tube-setup with microflow of water over the caryopsis for OEIHA condition.

For the EICWW condition, caryopsis was fixed horizontally on its sagittal plane on the edge of a hole of agarose gel. Then the hole was carefully filled with water. Thus, the bottom half of the caryopsis remained in direct contact with water. However, entire emerging embryo surface also remained in contact with water because of the capillary action due to the empty glume sheath over embryo region. Therefore, under EICWW condition, entire embryo surface was enveloped by thin layer of water. The Supplemental Movie S5 displayed the placement of rice caryopsis for EICWW condition.

The EUW condition was maintained by submerging caryopsis. The depth of submergence under EUW conditions varied depending on the experiments. For the experiments of time-lapse imaging and GUS staining, the depth of water was around 1 cm. For other EUW experiments, water depth was 5 cm. The effect of direct water contact on coleorhiza hair development was verified under EICWW and EUW conditions (i.e. enveloped and submerged, respectively) using dehulled RHL1:GUS caryopses. To keep embryo surface of dehulled caryopsis enveloped by thin layer of water, caryopsis placement was slightly modified due to lack of empty glume sheath. Instead of sagittal plane, the embryo region of horizontally oriented caryopsis was kept in facing-up position and water level was maintain up to the embryo surface.

Only agarose-based EIHA gel setup was used for the analysis of other grass species. For species with large caryopsis (like wheat, etc.), caryopsis was inserted through the prepunctured holes in the same way used for rice. For species with tiny caryopsis, like bluegrass (P. pratensis), etc., caryopsis was placed on the edge of a hole of agarose gel in the same way of EICWW set up. However, a small piece of filter paper was placed over the area of caryopsis located on the gel to ensure direct water contact on both sides of the caryopsis.

Absorption of water by caryopses

To estimate the amount of water absorbed by caryopses under different partial caryopsis–water contact conditions, we measured caryopses weight before the experiments and after several fixed intervals up to 72 h, using a high-precision semi-micro analytical balance (AUW220D, Shimadzu, Japan). The differences in caryopses weight from the start of the experiments to the fixed intervals were considered the amount of water absorbed by the caryopses. Similar approach was used to estimate the change of water content after the placement of imbibed caryopses under EIDA or EIW conditions. To reduce the experimental errors, the measurement of water absorption or the change of water content were analyzed from batches of 10 caryopses with three independent replications.

Histochemical GUS analysis

Histochemical GUS analysis was performed as described previously (Ding et al., 2009) with slight modification. Caryopses of the desired stage of germination were incubated with GUS staining solution (100-mM NaH₂PO₄ buffer pH 7.0, 1% (v/v) Triton X-100, 1-mg/mL X-Gluc, 5-mM K₃[Fe(CN)₆]), and 5 mM K₄[Fe(CN)₆]) overnight at room temperature. After staining, the caryopses were mounted on slides and images were captured by digital camera (DP71, Olympus) attached with stereomicroscope (SZX 16, Olympus).

Fluorescence microscopy

GFP fluorescence was observed under fluorescence stereomicroscope (SZX16; Olympus) coupled with reflected light fluorescence illuminator. The GFPHQ filter (Olympus) was used for GFP observation. The excitation and emission wavelengths were 460–480 nm and 595–540 nm, respectively. A digital camera (DP71; Olympus) recorded the fluorescence images with the help of the image acquisition software (DP Manager; Olympus). Both brightfield and fluorescent images of germinating caryopses were taken at different stages during germination.

Time-lapse microscopy

Customized stage-top humid chambers with heated lid surface were set up on stereomicroscopes (Olympus SZX 16 and SZ 51). The distance between caryopsis and lid surface was maintained >4 cm. Long-working distance objective (SDFPLAPO1XP; Olympus) was used in SZX 16 system. The addition of another perforated lid and the vertical temperature gradient ensured that the heated upper lid played no effect on caryopsis germination. Temperature around the embryo surface inside the humid chamber was 25.5°C ± 0.5°C. Time-lapse images were captured by the CCD camera (DP 71; Olympus) attached with the stereomicroscope. The image acquisition interval varied from 15 s to 60 min depending on the experiments.

Measurements of coleorhiza and root hairs

Both root and coleorhiza hair-related measurements such as length, diameter, density, elongation, etc. were analyzed from digital images with an image analysis software (ImageJ). The diameter of hair was measured at the mid-length of hairs. The surface area of exposed embryo surface was also estimated from digital images, ignoring the curvature of embryo surface. The surface area of coleorhiza hairs was estimated by the summation of lateral surface area of cylindrical hairs (2 πrh where r and h are the radius and length of coleorhiza hairs, respectively) and the curved surface of the hair tip (πr² where r is the radius of coleorhiza hairs).

The shrivel/collapse of coleorhiza hairs

The shrivel/collapse of coleorhiza hairs after the exposure to relatively dry air (RH ∼50%) was captured immediately after the removal of the lid of the petri dish. The extent of shrivel/collapse of coleorhiza hairs was analyzed from the recorded videos using ImageJ. The rate of shrivel/collapse was estimated in an AU. The AU was based on time context relative to rotation numbers of coleorhiza hairs (Figure 5A). Ten clearly visible outlying coleorhiza hairs were scored for their number of rotations along with the time points. Then, the rotation number was converted into percentage of the shrivel/collapse of coleorhiza hairs in context of the total rotation number of the respective hair.

Soil cracks and data acquisition

Water-saturated clay-rich field soil blocks, dimension 45 cm (L) × 33 cm (W) × 13 cm (H) were stacked and left undisturbed for desiccation in greenhouse (temperature 20°C–35°C and RH 55%–70%) to develop natural soil cracks. After soil-crack development, soil moisture was increased by submergence and water level was drained to the bottom soil block after 2 h of submergence. The bottom soil block was always kept in water-saturated condition. Two to three days after the initial submergence, dataloggers were inserted carefully into the suitable crack sites of upper block, preferably facing the dead end of cracks, keeping at least 5 cm distance from the soil surface. After insertion of datalogger, crack site was stabilized by toothpicks. Data of completely collapsed crack-sites were discarded as the RH of completely collapsed crack-sites always remained at the saturated level. Temperature and RH Datalogger, TM-305U (Tenmars, Taiwan) was selected for the study after analyzing numerous dataloggers, because the humidity sensor of TM-305U is directly exposed to the air. Moreover, TM-305U can reproducibly record the duration of saturated humidity hours. Usually, the data acquisition interval was 1 min.

Simplified soil-crack pattern was mimicked in agar blocks, ignoring the complexity of crack geometry. Agar blocks (dimension 12 × 12 × 4 cm) were cast in a specific design creating two cavities inside. The upper cavity represented a simplified crack in the shape of inverted triangular prism where the mimicked crack opening was 1.5 cm, depth 3.5 cm and length 12 cm. At the bottom of the mimicked crack-end, another cavity was made to accommodate the datalogger (TM-305U), keeping sensors toward the crack opening. Agar block was enclosed by two square petri dishes, keeping the crack opening facing upward.

Micro-droplet growth pattern and the fate of condensed droplets

Micro-droplet growth pattern was analyzed from the time-lapse images of induced condensation on the tip of coleorhiza hairs. The radius of the micro-droplet was analyzed with the ImageJ. Natural soil crack was mimicked in agar blocks and condensation was induced by decreasing the temperature of the humid chamber. After coleorhiza hair development, temperature inside the mimicked crack was decreased. Except the rate of decline, the extent of temperature decrease of natural cracks (∼3°C) was also replicated under microscope. Besides the addition of another perforated lid, the specific area of lid was coated with transparent antifog coating and kept in heated condition to capture clear image. To mimic decreasing night-time temperature, room temperature was gradually decreased and the thermal isolation layer over the humid chamber was removed. To identify the fate of condensed droplets, time-lapse images were converted into fast-motion video using ImageJ software.

Statistical analysis and reproducibility

No statistical methods were used to predetermine sample size. The experiments were not randomized, and the investigators were not blinded to allocation during the experiment and outcome assessment. All statistical tests were analyzed using Prism version 8. Appropriate statistical tests, P-values and n numbers, etc. were mentioned throughout the figures or figure legends. Similarly, box elements and error bars were also defined throughout.

The feature of coleorhiza hair development was observed in tens of thousands of rice caryopses and dozens of species across Poaceae, except Z. mays. To ensure the reproducibility, the low humidity experiments were repeated in multiple open environmental conditions beside climate chamber. Diurnal temperature and humidity data patterns of wet soil cracks were accumulated from soil cracks developed on same or different soil blocks on same or different dates. Sample size of air humidity of soil cracks was 45 and was large enough since the data logging intervals was 1 min (1,440 data points per site/day). Data of completely collapsed crack-sites were discarded as RH of completely collapsed crack-sites always remained at saturated level. To ensure the reproducibility of micromanipulated caryopsis–water contact experiments, particularly in the EIHA condition, besides the typical agarose gel-based set-up, experiments were also carried out in the tube-set-up conditions (like OEIHA set-up, except the closed stopper). Similarly, the occasional coleorhiza hair development in the EICWW condition was also verified using RHL1:GUS caryopses. The caryopsis coat rupture to radicle emergence time was derived from parallel experiments (set-up on SZX 16 and SZ 51). However, mutant caryopses that developed few/short hairs were excluded from the analysis. Time-lapse images of all major developmental processes were captured and presented in fast-motion videos.

Accession numbers

Sequence data from this article can be found in the GenBank/EMBL data libraries under accession numbers XM_015789291.2, BT026512.1, NM_001342047.1, NP_200609.1, and FJ375304.1 for Os06g08500, At2g24260, At4g30980, At5g58010, and LjRHL1, respectively.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. The contraction of embryo and emerging coleoptile after exposure to low air humidity (RH ∼55%) under EIDA condition.

Supplemental Figure S2. The effect of lid removal of humid chamber on coleorhiza hair, air temperature, and RH.

Supplemental Figure S3. Representative images of germinating caryopses of micromanipulated caryopsis–water contact conditions at the radicle emerging stage.

Supplemental Figure S4. Comparative phenotypes of germinated caryopses of the WT and RHL1 complemented/overexpression line at the radicle emerging stage under EUW conditions.

Supplemental Movie S1. The response of a germinating seed after exposure to low air humidity under EIDA condition.

Supplemental Movie S2. The sensitivity of coleorhiza hairs to dry air exposure.

Supplemental Movie S3. The response of rice embryo surface under EIHA condition.

Supplemental Movie S4. The response of rice embryo surface under OEIHA condition.

Supplemental Movie S5. The response of rice embryo surface under EICWW condition.

Supplemental Movie S6. The response of rice embryo surface under EUW condition.

Supplemental Movie S7. The fate of dew drops condensed on the tip of coleorhiza hairs.

Supplemental Movie S8. The progressive accumulation of dew drops in conditions mimicking decreasing night-time temperature.

Supplemental Movie S9. The response of rhl1 embryo surface under EIHA condition.

Supplemental Movie S10. The response of rhl1 embryo surface under OEIHA condition.

 

A.N.M.R.B.R. and J.Z. conceptualized the study; methodology, investigation, and writing of the original draft were done by A.N.M.R.B.R.; transgenic materials by W.D.; reviewing and editing were done by A.N.M.R.B.R., W.D., and J.Z.; J.Z. supervised the study.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/general-instructions) is Jianhua Zhang (jzhang@hkbu.edu.hk).