A Tobacco Syntaxin with a Role in Hormonal Control of Guard Cell Ion Channels

See allHide authors and affiliations

Science  22 Jan 1999:
Vol. 283, Issue 5401, pp. 537-540
DOI: 10.1126/science.283.5401.537


The plant hormone abscisic acid (ABA) regulates potassium and chloride ion channels at the plasma membrane of guard cells, leading to stomatal closure that reduces transpirational water loss from the leaf. The tobacco Nt-SYR1 gene encodes a syntaxin that is associated with the plasma membrane. Syntaxins and related SNARE proteins aid intracellular vesicle trafficking, fusion, and secretion. Disrupting Nt-Syr1 function by cleavage with Clostridium botulinum type C toxin or competition with a soluble fragment of Nt-Syr1 prevents potassium and chloride ion channel response to ABA in guard cells and implicates Nt-Syr1 in an ABA-signaling cascade.

The size of stomatal guard cells in higher plant leaves is rapidly reversible and is crucial to maintaining the hydrated environment within the leaf. In dry conditions, guard cells respond to the hormone abscisic acid (ABA) to regulate plasma membrane K+ and Cl channels which facilitate solute efflux. The concurrent decrease in turgor and cell volume closes the stomatal pore to reduce transpirational water loss (1). Response to ABA depends on guanosine triphosphatases (GTPases), protein (de-)phosphorylation, and changes in cytosolic-free Ca2+ concentration and pH (2, 3) and is associated with substantial alterations in intracellular membrane structure in the guard cells (4).

We isolated elements that contribute to ABA signaling in vivo, adapting a strategy similar to that used to identify mammalian receptor and ion channel proteins (5, 6). Polyadenylated [poly(A)+] RNA from leaves of drought-stressedNicotiana tabacum was injected into Xenopus laevis oocytes. Expression of the Nicotiana mRNA led to a cross-coupling between exogenous ABA-sensitive elements and the endogenous signaling pathways of the oocyte, evidenced by activation of the Xenopus Ca2+-dependent Clcurrent in the presence of 20 μM ABA (n = 16; Fig. 1A). Current activation was specific to mRNA-injected oocytes and was observed in response to ABA, but not to acetate or kinetin, a plant hormone that stimulates cell division and stomatal opening (1). After sucrose gradient fractionation of Nicotiana mRNA, the active fraction (mean size, 1.3 kb) was used to construct a cDNA library for expression and screening. Subdivision of library pools yielded clones that promoted the ABA-evoked current and a copurifying background current with similar characteristics but independent of ABA. The background current was isolated to a single clone (Fig. 1B). After depleting this transcript from Nicotiana mRNA, no ABA-sensitive current was observed (7), indicating that the gene carried by the clone was necessary to evoke the ABA-sensitive current (8).

Figure 1

Expression cloning of Nt-SYR1. (A) Voltage clamp recordings of Xenopus oocytes expressing poly(A)+ RNA from drought-stressedNicotiana leaves. Current-voltage (I-V) curves from one set of injections (means ± SE, n = 3 cells) recorded at the end of 1-s voltage steps before (•) and 30 s after (○) adding 20 μM ABA. (Inset) Currents from one cell cross-referenced by symbol. Voltage protocol (above): conditioning voltage, –120 mV; test voltages (8 cycles), –180 to +60 mV. Scale: horizontal, 1 s; vertical, 1 μA. The ABA-evoked current was identified with Cl channels by tail current analysis (ECl = –26 mV, n = 4) (8). (B) Sib-selection cloning of Nt-SYR1.Nicotiana leaf poly(A)+ RNA was size-fractionated on a 10 to 30% sucrose gradient. The fraction yielding the ABA response was used to construct a cDNA library in the pSPORT vector with the Superscript plasmid system (Gibco-BRL). DNA derived from pools of clones was linearized with Not I and transcribed in vitro with the T7 RNA polymerase. Xenopus oocytes were injected with this cRNA and assayed for ABA sensitivity. Currents before (left) and during (right) challenge with 20 μM ABA are shown from representative oocytes injected with progressively smaller cRNA pools. Scale: horizontal, 1 s; vertical, 500 nA.

Sequencing the transcript cDNA (9) revealed an open-reading frame encoding a syntaxin-related protein (Nt-Syr1; GenBank number AF112863) of 300 amino acids with a predicted molecular mass of 34.01 kD and an isoelectric point of 7.95. Alignments of Nt-Syr1 protein (Megalign, DNAstar, Madison, Wisconsin) showed similarities to the syntaxin-like Knolle gene product ofArabidopsis thaliana (38% identity) (10), human syntaxin-1A (23% identity) (11), and the yeast syntaxins SSO1p and SSO2p (22% identity each) (12). Syntaxins are essential for synaptic transmission, they coordinate cellular growth, and are implicated in vesicle trafficking in yeast, plants, and animals (10, 13–15). Features of Nt-Syr1 common to syntaxin proteins (Fig. 2A) include three domains (H1 through H3) with high probabilities for forming coiled-coil structures in protein-protein interactions, a putative membrane-spanning (hydrophobic) domain, and an adjacent domain (within H3) of 84% identity (92% homology) with the epimorphin consensus sequence (11). Nt-Syr1 also showed partial conservation of the three sites necessary for binding and cleavage by Clostridium botulinum type C neurotoxin (BotN/C) (16). Unlike other syntaxin proteins, Nt-Syr1 harbors a putative EF-hand, Ca2+-binding sequence and nucleotide binding site. Southern blot analyses using Nt-SYR1 cDNA indicated a low number of homologous genes in the Nicotiana genome and yieldedAt-SYR1 from Arabidopsis (GenBank numberAF112864) encoding a predicted protein with an overall identity of 72% with Nt-Syr1 (8, 17).

Figure 2

Structure and expression analysis of Nt-Syr1 (37). (A) Key features of Nt-Syr1 include putative Ca2+- (EF-hand) and nucleotide-binding (NBS) sites, partially conserved domains for recognition (*) and cleavage (⇕) by BotN/C, and putative coiled-coil domains (H1 through H3). High amino acid conservation with Knolle (10) and syntaxin-1A (Syn1A) (11) is found in the epimorphin domain and hydrophobic COOH-terminus. (B) Northern blot (above) and protein immunoblot (Western; below) analyses showing transient enhancement of Nt-SYR1 transcript and protein levels by treatment with 20 μM ABA and drought stress. Total RNA (15 μg/lane) and crude protein extracts (10 μg/lane) were isolated fromNicotiana leaves. Ribosomal RNA was used as a loading control (8). Northern blots probed with fullNt-SYR1 cDNA and protein immunoblots probed with Anti-Sp2 antiserum (18). Contr, control; 1/2 to 24, hours of ABA exposure; Drought, 48 hours. (C) Subcellular localization of the Nt-Syr1 by protein immunoblot analysis of fractionatedNicotiana leaves (MM, 50,000g microsomal membrane; S, soluble fraction; PM, two-phase partitioning plasma membrane; IM, inner membrane).

Northern (RNA) blot and protein immunoblot analyses showed that the Nt-SYR1 transcript and translation product (34 kD) were present in low abundance in leaves of well-watered plants (Fig. 2, B and C). Virtually all Nt-Syr1 was found in the 50,000g microsomal pellet and, after two-phase partitioning, was localized primarily to the plasma membrane, paralleling the distribution of the plasma membrane H+–adenosine triphosphatase (ATPase) (18). Remarkably, Nt-SYR1 [andAt-SYR1 (8)] transcript levels rose transiently approximately ninefold within 30-min exposure to ABA and after 48 hours drought stress (n = 3). Nt-Syr1 protein showed a parallel, albeit delayed, transience in ABA as expected for de novo translation and protein accumulation.

We explored syntaxin-related function of Nt-Syr1 by complementation of the H440 strain of Saccharomyces cerevisiae, which harbors the lethal deletion of plasma membrane syntaxin genes SSO1 and SSO2, and carriesSSO1 on a plasmid behind the GAL1galactose-inducible promoter (19). The H440 strain will grow on galactose, but not on glucose (12). After transformation, constitutive expression of Nt-Syr1 failed to rescue yeast growth on glucose (8). To examine Nt-Syr1 function in the plant, we used BotN/C which disrupts secretion by cleavage of syntaxins containing specific recognition sites (16, 20). Western blot analysis (21) of Nicotiana leaf microsomal proteins showed that Nt-Syr1, which contains homologs of the BotN/C-recognition sites, was cleaved by BotN/C, but not by BotN/D toxin (Fig. 3A), which targets the vesicle-associated protein synaptobrevin (16). Loss of BotN/C cleaved fragments (30 kD) was probably related to product instability and breakdown by endogenous proteases. The specificity of BotN/C action was indicated by the fact that cleavage was observed only when protein extracts were pretreated with ATP, which in synaptic protein complexes is required to expose syntaxin through complex disassembly by NSF ATPase (22). These results, and observations of antibody binding to high molecular weight bands (8), implicate Nt-Syr1 in similar complexes in planta.

Figure 3

Neurotoxin BotN/C, but not BotN/D, targets Nt-Syr1 and blocks ion channel response to ABA in Nicotianaguard cells. (A) Microsomal protein fractions were isolated from Nicotiana leaves, pretreated either with or without 1 mM ATP, incubated with BotN/C and BotN/D, separated by SDS-PAGE (6 μg/lane), and assayed by protein immunoblot analysis. (B) Voltage clamp analysis of Cl channel response to ABA in guard cells with and without BotN/C. Voltage steps (above): conditioning voltage (5 s), +30 mV (8); test voltages (6 cycles), –160 mV to +30 mV. Currents are from one guard cell loaded with 0.1 μM BotN/C before (top) and 8 min after (center) adding 20 μM ABA. Data from a second cell in ABA (bottom) is shown for comparison. No significant difference in current characteristics were observed without ABA between nonloaded cells and cells loaded with either toxin. Scale: vertical, 100 μA cm–2; horizontal, 2 s. Zero current level is on the left. (C) Means ± SE of the ABA response of the Cl current (ICl; top) and inward-rectifying K+ current (IK,in; bottom) from nonloaded (contr) and BotN/C- and BotN/D-loaded (0.1 μM) guard cells (n ≥ 5). Currents were taken at –200 mV and normalized to the corresponding measurements taken before ABA treatments.

We tested the effects of BotN/C and BotN/D on ABA-mediated control of guard cell K+ and Cl channels inNicotiana. ABA treatment normally results in a 40 to 60% inactivation of inward-rectifying K+ channel current (IK,in), a two- to fourfold stimulation of current through the Cl channels (ICl) and slowing of ICl gating (2, 23, 24). Voltage clamp recordings (Fig. 3B) (25) showed that cytosolic loading with BotN/C, but not with BotN/D, prevented ABA action on Cl channel gating, and a similar loss of sensitivity to ABA was found for IK,in after BotN/C loading (Fig. 3C). Equivalent results were obtained in recordings from Vicia faba guard cells (8).

In separate experiments we used the C-truncated Nt-Syr1 protein Sp2 (18) to “poison” Nt-Syr1 functioning. By analogy with the action of C-truncated Syntaxin-1A in secretion (20), we reasoned that if a protein complex with Nt-Syr1 was necessary for ABA signaling, adding the truncated protein—including the protein-protein interaction domains, but lacking the COOH-terminal membrane anchor—might compete with Nt-Syr1 for partners and prevent normal complex functioning. Voltage clamp records (Fig. 4A) (25) showed that current through the outward-rectifying K+ channels (IK,out) was enhanced two- to threefold in ABA, while IK,in was reduced, at –200 mV, to roughly 25% of the control. ABA also shifted the voltage-sensitivity of IK,in (Fig. 4C), consistent with its Ca2+-sensitivity and ABA action on its gating (2). In guard cells loaded 20 or 100 μM Sp2 protein IK,in and IK,out showed complete loss of sensitivity to ABA, and a similar loss of sensitivity was found for ICl (Fig. 4B).

Figure 4

Truncated Nt-Syr1 protein (Sp2) blocks ion channel response to ABA in Nicotiana guard cells. (A) Voltage clamp analysis of K+ channel response to 20 μM ABA. Voltage protocol (bottom right): conditioning voltage, –100 mV; test voltages (16 cycles), –250 mV to +30 mV; tailing voltage, –100 mV. Currents are shown from one guard cell loaded with 20 μM Sp2 before (○) and 10 min after (•) adding ABA. Data from a second cell in ABA (▴) for comparison show characteristic slowing of inward current by ABA (2). No significant difference in current characteristics were observed without ABA between nonloaded cells and cells loaded with Sp2. Scale: vertical, 100 μA cm–2; horizontal, 1 s. Zero current levels are on the left. (B) Means ± SE of steady-state currents before and 10 min after adding ABA in nonloaded (Control) and Sp2-loaded (+Sp2) cells. Data are for IK,in (open bars, down) and IK,out (open bars, up) and for ICl (shaded bars, down). Currents were recorded at –200 mV (IK,in), +30 mV (IK,out), and –100 mV (ICl). (C) Steady-state current-voltage relationships for IK,in and IK,out from (A) and cross-referenced by symbol showing characteristic shift in voltage-sensitivity for IK,in in ABA (▴). (Inset) Conductances (gK,rel) of IK,in ± Sp2 in ABA (•, ▴) relative to conductances without ABA.

We interpret these results to indicate a central role for Nt-Syr1 in early steps of ABA signaling and to implicate its functioning in a heteromultimeric complex with other proteins. This idea accords with the homology of Nt-Syr1 to other syntaxins, its presence in high–molecular weight components, and the action of Sp2 on the ABA response of guard cell ion channels. The target (t-)SNARE syntaxin takes part in a number of protein-protein interactions essential for vesicle trafficking, secretion and endocytosis (13,26). Syntaxin binding partners at the presynaptic membrane include SNAP-25 and the vesicle (v-)SNARE synaptobrevin, which form a stable ternary complex for vesicle fusion (13). Less is known of syntaxin function in stimulus perception, although its interaction with other elements is likely to be important for signaling (27,28). Syntaxins do interact with other proteins that may not be related to secretion processes directly. Syntaxin-1A copurifies with the orphan G-protein–coupled receptor CIRL (29) and interacts with tomosyn, a microfilament-associated protein (30). Syntaxins also bind Ca2+ channels and CFTR Cl channels, influencing their regulation (13, 31,32). Significantly, Ca2+ channels and SNARE proteins interact with the syntaxin-1A epimorphin domain, and peptides synthesized to this domain, corresponding to that in Sp2, prevent secretion (13, 20, 32).

In plants, SNARE proteins have been implicated in cell growth and development (33). During stomatal movements, the plasma membrane surface area of guard cells can change by 50%, much more than can be accommodated by lateral expansion and compression of the bilayer (1, 33). Guard cell membrane structure undergoes substantial change during stomatal movements (4), and protoplast volume is coupled to membrane trafficking (34). Thus, we anticipate a crucial role for SNARE proteins in ABA-evoked changes in secretion and endocytosis. It is interesting that Nt-Syr1 harbors putative nucleotide and Ca2+-binding domains, especially in light of Ca2+ action as a second messenger during ABA signaling (2, 3). For synaptic transmission Ca2+facilitates the later stages of vesicle fusion, but binds to synaptotagmin, not to syntaxin (13). These domains may therefore indicate regulatory activities unique to Nt-Syr1.

Most importantly, our data point to a new function for this putative SNARE protein as a key element in a hormonal signal cascade, and not simply as a component of the response mechanics. ABA control of K+ and Cl channel gating cannot be linked readily to membrane trafficking per se. It is equally difficult to explain concurrent regulation of all three ion channels through a direct interaction with each of the channel proteins, especially as syntaxins are not known to associate with K+ channels. Hence, disruption of ABA signaling by BotN/C and Sp2 in vivo implies an action of Nt-Syr1 upstream, possibly close to the primary event of ABA binding. How might Nt-Syr1 contribute to ABA signaling? At present, there is almost no data that bears on SNARE function in intracellular (nonsynaptic) signal transmission. Nonetheless, several lines of evidence suggest more intimate roles for SNARE proteins beyond the mechanics of vesicle trafficking, including scaffolding and nucleation (27, 30, 35) that are essential features of many signaling cascades. SNARE proteins have also been suggested to contribute to the sensing of osmotic stress (36). At present our data do not speak directly to the possibility of ABA binding to Nt-Syr1, or to one or more possible roles as a scaffolding protein, a second messenger element, or a modulator to the poise of one or more signaling elements or ion channels. Identifying proteins that interact with Nt-Syr1 will help to resolve this issue and to gain further insights into its function.

  • * Present address: Instituto de Recursos Naturales y Agrobiologia, Consejo Superior de Investigaciones Cientificas, Post Office Box 1052, Sevilla 41089, Spain.

  • To whom correspondence should be addressed. E-mail: mblatt{at}


Stay Connected to Science

Navigate This Article