Guard cell SLAC1-type anion channels mediate flagellin-induced stomatal closure

Abstract

During infection plants recognize microbe-associated molecular patterns (MAMPs), and this leads to stomatal closure. This study analyzes the molecular mechanisms underlying this MAMP response and its interrelation with ABA signaling. Stomata in intact Arabidopsis thaliana plants were stimulated with the bacterial MAMP flg22, or the stress hormone ABA, by using the noninvasive nanoinfusion technique. Intracellular double-barreled microelectrodes were applied to measure the activity of plasma membrane ion channels. Flg22 induced rapid stomatal closure and stimulated the SLAC1 and SLAH3 anion channels in guard cells. Loss of both channels resulted in cells that lacked flg22-induced anion channel activity and stomata that did not close in response to flg22 or ABA. Rapid flg22-dependent stomatal closure was impaired in plants that were flagellin receptor (FLS2)-deficient, as well as in the ost1-2 (Open Stomata 1) mutant, which lacks a key ABA-signaling protein kinase. By contrast, stomata of the ABA protein phosphatase mutant abi1-1 (ABscisic acid Insensitive 1) remained flg22-responsive. These data suggest that the initial steps in flg22 and ABA signaling are different, but that the pathways merge at the level of OST1 and lead to activation of SLAC1 and SLAH3 anion channels.

Keywords: ABA; Arabidopsis thaliana; S-type anion channel; flg22; guard cells; innate immunity; microbe-associated molecular pattern (MAMP); stomata.

Figure 1

Nanoinfusion of flg22 and ABA triggers rapid stomatal closure in intact Arabidopsis leaves. (a) Illustration of the nanoinfusion technique used to induce flg22‐ and ABA‐dependent stomatal closure. A microcapillary was moved into the substomatal cavity of an open stoma and used to infuse solutions into the intercellular space. Movement of neighboring stomata was monitored on an upright microscope. (b) Images of a stoma in the abaxial epidermis of an Arabidopsis leaf stimulated by nanoinfusion of 20 nM flg22. Images were obtained just before (left panel), directly after (middle panel), and 35 min after stimulation with flg22 (right panel). Note that the leaf becomes transparent because of solution infused into the intercellular space. (c) Time‐dependent changes in average stomatal aperture before and after stimulation with control solution (closed circles, n = 8), 10 μM ABA (open circles, n = 13), or 20 nM flg22 (open triangles, n = 13); the arrow indicates the time point of nanoinfusion. Data are presented as average values of 8 to 13 stomata of at least three independent experiments, and error bars represent ± SE.

Figure 2

Nanoinfusion of flg22 stimulates S‐type anion channels in guard cells. (a) Illustration of the nanoinfusion technique, combined with the voltage‐clamp technique, using intracellular double‐barreled microelectrodes in a guard cell of a neighboring stoma. (b) Guard cells of Arabidopsis thaliana acc. Landsberg erecta in intact plants were impaled with double‐barreled microelectrodes, filled with 300 mM KCl, and clamped at a holding potential of −100 mV. A bipolar step protocol (upper panel) was used to obtain current traces within 30 min after nanoinfusion of control solution (left lower panel) or 20 nM flg22 (right lower panel). (c) Current–voltage plots of guard cells clamped from a holding potential of −100 mV to test pulses ranging from −200 to 20 mV, as shown in (b). Data are presented as average values of eight guard cells, and error bars represent ± SE. (d) K⁺ efflux channels were blocked by filling electrodes with 300 mM CsCl and clamping guard cells to a holding potential of 0 mV. A bipolar step protocol (lower panel) was used to obtain current traces within 30 min after nanoinfusion of control solution (left upper panel) or 20 nM flg22 (right upper panel). (e) Current–voltage relationships for guard cells clamped from a holding potential of 0 mV to test pulses ranging from 60 to −140 mV, as shown in (d). Data are presented as average values of 10 guard cells, and error bars represent ± SE.

Figure 3

Both SLAC1 and SLAH3 contribute to the S‐type anion conductance of guard cells. (a) Guard cells of Arabidopsis thaliana acc. Columbia 0 (Col‐0) were impaled with double‐barreled electrodes filled with 300 mM CsCl and clamped to a holding potential of 0 mV. Experiments were carried out with epidermal strips to ensure identical extracellular ion concentrations for all guard cells. (b) Bipolar step protocol used to obtain current traces of the wild‐type, and the slac1‐3, slah3‐1, and slac1‐3/slah3‐1 loss‐of‐function mutants as shown in (a). (c) Current–voltage relationship for guard cells of Col‐0 (closed circles), slah3‐1 (open triangles), slac1‐3 (closed triangles), and slac1‐3/slah3‐1 (open circles). Data are presented as average values of eight to 11 cells, and error bars represent ± SE.

Figure 4

Flg22 stimulates the activity of SLAC1, as well as SLAH3. (a) Guard cells of intact Arabidopsis thaliana acc. Col‐0 plants were impaled with double‐barreled electrodes, filled with 300 mM CsCl, and clamped from a holding potential of 0 mV. Bipolar step protocols (upper panel) were applied to obtain current traces for the single slah3‐1 and slac1‐3 loss‐of‐function mutants, as well as the slac1–3/slah3–1 double mutant. Data were obtained within 30 min after stimulation by nanoinfusion with flg22. (b) Current–voltage relationship for guard cells stimulated by nanoinfusion with control solution (open circles) or 20 nM flg22 (closed circles), obtained with pulse protocols as shown in (a). Data are presented as average values of eight cells, and error bars represent ± SE.

Figure 5

The S‐type anion channels SLAC1 and SLAH3, and the protein kinase OST1 are essential for rapid flg22‐induced stomatal closure. (a) Time‐dependent stomatal movement induced by nanoinfusion of 20 nM flg22 in slac1‐3 (closed circles, n = 9), slah3‐1 (open circles, n = 16), slac1‐3/slah3‐1 (open triangles, n = 11), and wild‐type Col‐0 (closed triangles, data from Fig. 1c). Data are given as average values of nine to 16 stomata from at least four independent experiments, and error bars represent ± SE; arrows indicate the time point of nanoinfusion. (b) Time‐dependent stomatal movement of Arabidopsis thaliana acc. Landsberg erecta (Ler) stimulated by nanoinfusion of control solution (closed circles, n = 20), 10 μM ABA (open circles, n = 20), and 20 nM flg22 (closed triangles, n = 21), as well as the response of the ost1‐2 mutant to flg22 (open triangles, n = 17). Data are given as average values of 17–21 stomata from at least four independent experiments, and error bars represent ± SE.

Figure 6

Calcium‐dependent protein kinases 3 (CPK3), 5, 6, and 11 are essential for the flg22‐induced reactive oxygen species (ROS) production of mesophyll cells, but not for their depolarization or for stomatal closure. (a) Expression level of CPKs in wild‐type Arabidopsis thaliana acc. Col‐0 and the cpk3/5/6/11 mutant, analyzed by reverse transcription polymerase chain reaction (RT‐PCR). The expression of CPK3,CPK5,CPK6, and CPK11 is abolished in the quadruple mutant, while that of the control CPK4 is not affected. (b) Hydrogen peroxide (H₂O₂) production of mesophyll tissue measured with a platinum microdisc electrode in Col‐0 (open circles) and the cpk3/5/6/11 mutant (closed circles) stimulated with flg22, as well as Col‐0 exposed to flg22‐Δ2 (open triangles). Error bars represent ± SE of six experiments. (c) Membrane potential recordings of mesophyll cells of Col‐0 (open circles) and cpk3/5/6/11 (closed circles) stimulated with flg22, as well as Col‐0 exposed to flg22‐Δ2 (open triangles). Error bars represent ± SE of six experiments. (d) Normalized data of stomatal closure, induced by nanoinfusion of control solution, 10 μM ABA, or 20 nM flg22. The stomatal apertures at the start of the experiments were set to 1 and final apertures, measured after 40 min for ABA responses or 50 min for flg22 and control experiments, are shown relative to the starting values. Data are shown for the cpk3/5/6/11 quadruple mutant, the rboh‐D/F double loss‐of‐function mutant, Wassilewskija (Ws‐0), and Ws‐0 transformed with a functional FLS2 receptor. In cpk3/5/6/11, three out of 16 stomata did not respond to flg22, whereas in rbohD/F, two out of 11 and four out of 12 stomata did not close in response to flg22 and ABA, respectively. Data are given as normalized values (aperture before nanoinfusion = 1) of at least eight experiments. Error bars represent ± SE.

Figure 7

Abscisic acid and flg22 signaling pathways merge at OST1. (a) Time‐dependent stomatal movement of the Arabidopsis thaliana acc. Landsberg erecta (Ler) abi1‐1 mutant, stimulated by nanoinfusion of control solution (closed circles, n = 11), 10 μM ABA (open circles, n = 14), and 20 nM flg22 (closed triangles, n = 13); for comparison, data of the Ler wild‐type stimulated with flg22 (open triangles, n = 21) from Fig. 5(b) are shown. Data are given as average values of 11–14 stomata from at least four independent experiments, and error bars represent ± SE; arrows indicate the time point of nanoinfusion. (b) Schematic representation of the signaling pathway for flg22‐induced membrane responses in guard cells. Flg22 binds to receptor‐like kinase FLS2 in the plasma membrane, which interacts with the BRI1‐associated kinase 1 (BAK1) and somatic embryogenesis‐related kinase (SERK) coreceptors. The interaction between both receptors leads to OST1 activation, either through the inhibition of PP2Cs or by an alternative mechanism. OST1 can directly phosphorylate and activate SLAC1, which releases anions into the guard cell wall. OST1 may activate Ca²⁺‐permeable channels, causing a cytosolic Ca²⁺ signal. The cytosolic Ca²⁺ signal will activate calcium‐dependent protein kinases (CPKs), which can activate SLAC1, as well as SLAH3. Likewise, Ca²⁺ can bind to calcineurin B‐like (CBL) proteins that interact with CBL‐interacting protein kinases (CIPKs), which in turn can activate SLAC1 and SLAH3. Alternatively, OST1 could be capable of activating CPKs or CIPKs through a Ca²⁺‐independent mechanism.