Mechanosensitive channel MSL8 regulates osmotic forces during pollen hydration and germination

See allHide authors and affiliations

Science  23 Oct 2015:
Vol. 350, Issue 6259, pp. 438-441
DOI: 10.1126/science.aac6014

Metered rehydration in pollen grains

When a desiccated pollen grain lands on fertile territory, it rehydrates on the way to activating its growth and metabolic processes. Studying the small plant Arabidopsis, Hamilton et al. have identified a mechanosensory ion channel that responds to the distention of the plasma membrane as the pollen grain rehydrates. With this channel damaged or absent, the pollen grains germinated overenthusiastically but then showed a tendency to burst.

Science, this issue p. 438


Pollen grains undergo dramatic changes in cellular water potential as they deliver the male germ line to female gametes, and it has been proposed that mechanosensitive ion channels may sense the resulting mechanical stress. Here, we identify and characterize MscS-like 8 (MSL8), a pollen-specific, membrane tension–gated ion channel required for pollen to survive the hypoosmotic shock of rehydration and for full male fertility. MSL8 negatively regulates pollen germination but is required for cellular integrity during germination and tube growth. MSL8 thus senses and responds to changes in membrane tension associated with pollen hydration and germination. These data further suggest that homologs of bacterial MscS have been repurposed in eukaryotes to function as mechanosensors in multiple developmental and environmental contexts.

Mechanosensitive (stretch-activated) ion channels provide an evolutionarily conserved mechanism for the perception of mechanical force at the membrane (1). The mechanosensitive channel of small conductance (MscS) from Escherichia coli belongs to a large and structurally diverse family of proteins encoded in bacterial, archaeal, plant, and fungal genomes (2, 3). Bacterial MscS homologs prevent cellular lysis upon hypoosmotic shock by releasing osmolytes from the cell in direct response to increased lateral membrane tension (4). MscS-like (MSL) proteins in plants exhibit homology to the pore-lining domain of E. coli MscS; outside of this region, they show diverse domains and topologies (3) (fig. S1, A and B). Arabidopsis thaliana mutants lacking functional MSL genes respond normally to externally applied osmotic or mechanical stresses (5).

We therefore hypothesized that MscS homologs in plants may sense and respond to rapid changes in water status (and therefore membrane tension) that are intrinsic to the plant life cycle rather than environmentally imposed. Several such events occur during the development of pollen, the multicellular haploid life stage of plants that harbors the male gametes (6). In most angiosperms, including A. thaliana, the last stage of pollen maturation is partial dehydration (<30% water content) (7). Once the desiccated pollen grain contacts the stigma cells of a compatible female flower, stigma exudate enters the grain and reactivates its metabolism (8). The pollen tube germinates, breaking through the grain cell wall and proceeding via polarized tip growth toward female gametes inside the ovaries (9). The mechanical stress exerted on pollen membranes and cell walls (10, 11) and the spatially and temporally dynamic ion fluxes known to be essential for pollen grain germination and tube growth (12) suggest a role for stretch-activated ion channels (13). Mechanosensitive cation channel activities have been detected in pollen grain and tube membranes (14), but their molecular identity and physiological functions remain unknown.

A. thaliana MSL8 (At2g17010) transcripts were detected in mRNA isolated from floral but not leaf or root tissue (fig. S1C). In transgenic plants expressing genomic MSL8 fused to green fluorescent protein (gMSL8-GFP) under the control of native sequences, fluorescence was observed inside half of the pollen grains within the anthers of the hemizygous first transformed T1 generation (Fig. 1A). gMSL8-GFP signal was observed in tricellular and mature pollen (Fig. 1, B to F) (6) but not in any other tissue. MSL8 transcripts were identified in an RNA-sequencing data set from mature, dry pollen (15) (fig. S1D). Phylogenetic analysis suggests that male-specific expression of MSL genes evolved in both monocot and dicot lineages (fig. S2).

Fig. 1 MSL8 is expressed in mature pollen grains and localizes to the plasma membrane and endomembranes of pollen grains and germinating tubes.

(A) Confocal image of GFP signal in a dissected anther from stage 12 flowers of a gMSL8-GFP T1 plant. Asterisks mark transgenic pollen. The pollen cell wall is autofluorescent, but internal fluorescence is ascribed to MSL8-GFP. (B to F) Confocal images of GFP (green) and 4′,6-diamidino-2-phenylindole (DAPI) signal (blue) during pollen development. (B) Tricellular pollen from a nontransformed line. Pollen was isolated from a gMSL8-GFP–expressing line at the (C) microspore, (D) bicellular, (E) tricellular, and (F) mature stages of development. (G and H) Confocal images of ungerminated (G) and in vivo germinated (H) pollen grains in a gMSL8-GFP–expressing line. (I) A germinated pollen tube from a line expressing LAT52pMSL8-YFP. (J) MSL8-YFP (green) overlaid with CPK34-mCherry (red). (K) Magnification of box indicated in (J), green channel only. (L) MSL8-YFP overlaid with endoplasmic reticulum–mCherry (red). Arrows indicate plasma membrane. Scale bars, 10 μm [(B) to (J) and (L)] or 5 μm (K).

MSL8-GFP expressed from endogenous sequences localized both to the plasma membrane and to endomembrane compartments in pollen grains, and upon germination was mobilized to the tube periphery (Fig. 1, G and H), as did MSL8-YFP (yellow fluorescent protein) expressed from the strong pollen-specific promoter LAT52 (Fig. 1I) (16). MSL8-YFP colocalized with the pollen plasma membrane protein CPK34 (17) but not with an endoplasmic reticulum marker (maximum Pearson’s correlation coefficients of 0.66 and 0.09, respectively) (Fig. 1, J to L, and fig. S3), and there was no substantial overlap with Golgi or vacuole markers (fig. S4). A similar internal localization pattern has been observed with other pollen plasma membrane proteins (18, 19).

MSL8 produced mechanosensitive ion channel activity when expressed in Xenopus laevis oocytes (Fig. 2A and fig. S5, A and B). In this system,MSL8 (or MSL8-YFP, which was indistinguishable) (fig. S5C) had a unitary conductance of 57 pS under negative membrane potentials and 39 pS under positive membrane potentials (Fig. 2B); the conductance of MscS is ~300 pS under similar conditions (20). MSL8 exhibited a 6.3-fold preference for chloride over sodium (Fig. 2C) and is therefore more anion-selective than MscS, which has a permeability ratio (PCl:PK) of 1.2 to 3.0 (21). Finally, the gating threshold pressure for MSL8 was higher than that of MscS (–48.2 ± 14.8 mm Hg and –19.1 ± 5.1 mm Hg, respectively; n = 9 patches) (fig. S5, D and E). MSL8 activity was unaffected by treatment with MgCl2, ruthenium red, or tetramethylammonium-Cl (figs. S5C and S6, A to C). A mechanosensitive channel activity with conductance similar to MSL8 under the same conditions was occasionally detected in wild-type Col-0 pollen protoplasts (5 of 58 patches) and may correspond to the endogenous MSL8 channel (Fig. 2D and fig. S6, D to F). Final confirmation will require demonstrating the loss of channel activity in msl8 mutant pollen grains.

Fig. 2 MSL8 forms a small-conductance mechanosensitive ion channel with a preference for anions that is similar to a channel present in pollen membranes.

(A) Representative trace from an excised inside-out patch of plasma membrane from a Xenopus laevis oocyte expressing MSL8-YFP capped RNA (cRNA) (left) or water-injected (right) at –60 mV membrane potential. Channels were gated by negative tension (suction) applied to the patch pipette. The first three channel openings are labeled O1 to O3. (B) The current-voltage relationship of untagged tension-gated MSL8 in symmetric ND96 buffer. N= 8 oocytes. (C) The current-voltage relationship of tension-gated MSL8 under symmetric (filled circles, 100 mM NaCl) and asymmetric (open circles, 100 mM NaCl pipette/300 mM NaCl bath) conditions. N = 3 oocytes. (D) Representative trace from an excised inside-out patch of membrane from a Col-0 pollen protoplast at a transmembrane potential of –60 mV. Inset, four channel openings with characteristics similar to MSL8 are labeled O1 to O4.

Two transfer DNA (T-DNA) insertion alleles were identified that resulted in the reduction and loss of detectable MSL8 transcripts in the flower, msl8-1 and msl8-4, respectively (fig. S7, A and B). The null msl8-4 allele was transmitted through the male germ line with reduced efficiency, while it was transmitted normally through the female germ line (Fig. 3A). Even modest transmission defects will result in rapid purification from a natural population, as pollen-specific genes are subject to strong selection against deleterious mutations (22). We detected no obvious morphological defects in the coat or cell wall of desiccated msl8-4 pollen grains (Fig. 3B).

Fig. 3 MSL8 is required for mature pollen grains to survive hypoosmotic shock during rehydration.

(A) Transmission ratio analysis of the msl8-4 allele. The progeny of reciprocal crosses between msl8-4 heterozygotes and wild-type plants were genotyped. P values were determined by a chi-square test against the expected ratio of 50:50. (B) Scanning electron micrograph of desiccated pollen from the indicated genotypes. Scale bar is 20 μm. (C)Viability of wild-type and mutant pollen after hydration. Pollen was incubated for 30 min in distilled water containing fluorescein diacetate and propidium iodide, dyes that stain viable and unviable pollen, respectively. Asterisks mark compromised pollen. Scale bar, 100 μm. (D) Hydration viability time course. The average of three experiments with N = 55 170 pollen grains is shown. Error bars, mean ± SD. Asterisks indicate significant (P < 0.05) differences from the wild type by Student’s t test. (E) Effect of a PEG series on viability (bars, 3 to 5 trials per genotype; N = 48 to 440 pollen grains per trial; error bars, mean ± SE) and pollen diameter (lines, N = 15 pollen grains per genotype; error bars, mean ± SD) after hydration. (F) Viability after hydration of nondehiscent tricellular pollen grains dissected from the anther of the indicated genotypes. N > 50 pollen grains per genotype. Error bars, mean ± SD.

To test the hypothesis that MSL8 is required for pollen to survive the osmotic downshift experienced during rehydration, we quantified the viability of mature pollen after rehydration in distilled water. Whereas wild-type pollen exhibited 83 to 95% viability over the 2-hour time course, msl8-4 pollen viability dropped from 38% to 21%, and msl8-1 pollen viability dropped from wild-type levels to 46% (Fig. 3C-D). This phenotype was rescued by the gMSL8-GFP transgene in both mutant backgrounds (fig. S7, C and D).

While the osmotic shock of in vitro hydration in distilled water is more extreme than pollen grains are likely to experience in vivo, msl8-4 pollen also shows a defect in viability when hydrated in low concentrations of polyethylene glycol (PEG) 3350 (Fig. 3E). These results are consistent with previous work showing that even slow rates of osmotic downshock are lethal to an E. coli strain lacking mechanosensitive ion channels (23). Increasing the osmolarity of the hydration medium led to corresponding increases in pollen viability. The in vivo hydration rate of msl8-4 pollen did not differ from the wild type (fig. S7E). Finally, msl8-4 pollen grains dissected from anthers before desiccation showed wild-type viability upon hydration (Fig. 3F). Thus, msl8 pollen developed normally and was fully viable before dehydration, and its loss of viability when hydrated can be attributed to the hypoosmotic challenge of water entering the desiccated grain.

MSL8 also plays a role in pollen germination. During in vitro germination assays, msl8-4 pollen grains and germinated tubes burst 26% of the time, whereas only 3% of the wild type burst (Fig. 4, A to C). Pollen bursting is associated with cell-wall defects and lesions in ion channel genes and is thus thought to result from a failure to balance osmotic pressure with the strength of the cell wall (19). msl8-4 and msl8-1 pollen germinated at a higher rate than the corresponding wild type (both burst and intact tubes: 43% in msl8-4 and 57% in msl8-1; compare to 20% and 23% in Ler and Col-0, respectively) (Fig. 4, A to C). Conversely, overexpression of LAT52pMSL8-YFP inhibited germination, and three independent homozygous lines exhibited only 4 to 39% of wild-type germination rates (Fig. 4D). The germination ratewas inversely correlated with the level of MSL8-YFP transcript in these lines (Fig. 4E), confirming that MSL8 negatively regulates in vitro germination. MSL8-YFP overexpression lines were impaired in transmission of the transgene to the next generation (Fig. 4F and fig. S8A), but only through the male parent (Fig. S8B-C). Because these lines exhibited wild-type hydration survival (fig. S8D), this defect can be attributed to reduced rates of pollen germination.

Fig. 4 MSL8 negatively regulates germination and is required for cellular integrity during germination.

Wild-type (A) and msl8-4 (B) pollen germinated for 4 hours in liquid germination media and stained with aniline blue for callose, a marker of germination. Examples of ungerminated (single asterisk), ungerminated and burst (double asterisk), germinated (white arrow), or germinated and burst (black arrow) pollen are indicated. Scale bar, 50 μm. (C) Germination rate and bursting frequency in the indicated genotypes. N ≥ 396 pollen grains per genotype. Asterisks indicate significant (P < 0.05) differences by Student’s t test (D) Percentage germination overnight on solid media of pollen from wild-type and three independent homozygous LAT52pMSL8-YFP transgenic lines. Error bars, mean ± SE. (E) Quantitative reverse-transcription polymerase chain reaction of MSL8-YFP and MSL8 transcripts relative to ACTIN in flowers from Ler, msl8-4, and the LAT52pMSL8-YFP lines in (D). Two technical replicates of three biological replicates are presented. Error bars, mean ± SE. (F) Survival of selection for the Bialophos resistance gene in offspring of the LAT52pMSL8-YFP lines in (D). P values were determined by a chi-square test against 75% expected survival.

Changes in osmotic potential are part of normal pollen function, in addition to being environmental stresses that must be tolerated. Here, we show that MSL8 is required for a tuned response to developmentally normal osmotic challenges. During pollen rehydration, MSL8 maintains cellular integrity upon osmotic downshift, playing a role analogous to that of E. coli MscS. During pollen germination, however, MSL8 maintains the optimal osmotic potential required to drive germination yet prevent lysis of the nascent pollen tube. MSL8 may accomplish these functions by releasing osmolytes directly in response to membrane tension and/or function indirectly in pathways that regulate pollen desiccation, membrane trafficking, or cell-wall dynamics. This study illustrates how MSL8, a eukaryotic homolog of the bacterial osmotic safety valve MscS, has been repurposed to help pollen cope with predictable osmotic changes that are characteristic of pollen development. It also contributes to a growing body of evidence that mechanical signaling plays a critical role in plant and animal development (24, 25).

Supplementary Materials

Materials and Methods

Figs. S1 to S8

Table S1

References (2632)

References and Notes

  1. Acknowledgments: This work was funded by NSF MCB1253103 and NIH 2R01GM084211 to E. S. Haswell and by the Monsanto Excellence Fund Fellowship to E. S. Hamilton. We acknowledge M. Wilson and the Donald Danforth Plant Science Center imaging facility for technical assistance, J. Harper for plasmid ps819, M. Johnson for line ms1-1, and the Arabidopsis Biological Resource Center for T-DNA insertion lines and ER-rk (CD3-959) and G-rk (CD3-967) membrane markers. The authors do not have any conflicts of interest. The supplementary materials contain additional data.
View Abstract


Navigate This Article