The two-pore channel TPK1 gene encodes the vacuolar K⁺ conductance and plays a role in K⁺ homeostasis

Abstract

The Arabidopsis thaliana genome contains five genes that encode two pore K⁺ (TPK) channels. The most abundantly expressed isoform of this family, TPK1, is expressed at the tonoplast where it mediates K⁺-selective currents between cytoplasmic and vacuolar compartments. TPK1 open probability depends on both cytoplasmic Ca²⁺ and cytoplasmic pH but not on the tonoplast membrane voltage. The channel shows intrinsic rectification and can be blocked by Ba²⁺, tetraethylammonium, and quinine. TPK1 current was found in all shoot cell types and shows all of the hallmarks of the previously described vacuolar K (VK) tonoplast channel characterized in guard cells. Characterization of TPK1 loss-of-function mutants and TPK1-overexpressing plants shows that TPK1 has a role in intracellular K⁺ homeostasis affecting seedling growth at high and low ambient K⁺ levels. In stomata, TPK1 function is consistent with vacuolar K⁺ release, and removal of this channel leads to slower stomatal closure kinetics. During germination, TPK1 contributes to the radicle development through vacuolar K⁺ deposition to provide expansion growth or in the redistribution of essential minerals.

Plant vacuoles are important in providing turgor and as a depository for minerals. The vacuolar role in providing turgor and turgor regulation is amply demonstrated in guard cells where large transtonoplast fluxes of inorganic ions such as K⁺ and Cl⁻ are essential for stomatal functioning. In addition, cell expansion and therefore plant growth are to a large extent driven by vacuolar K⁺ accumulation. Apart from the vacuolar deposition of minerals to provide turgor, vacuoles typically function as buffer compartments when excess nutrient uptake during periods of abundance requires storage in the central vacuole. This “luxury” consumption enables plants to survive during sub sequent periods of deficiency when minerals such as NO₃⁻ and K⁺ are released from the vacuole to maintain cytosolic homeostasis (1, 2).

These essential vacuolar functions can be executed only if adequate transport mechanisms are present at the tonoplast, and these have been shown to include primary pumps, carriers, and ion channels. Exhaustive characterization through patch–clamp and other electrophysiological techniques has identified at least three types of monovalent cation channel and three types of Ca²⁺-permeable channel in vacuolar membranes (3), anion channels (4), and malate channels (5).

Although the recent genomics and postgenomics approaches have allowed the identification and putative functional annotation of many genes, little advance has been made in identifying genes that encode tonoplast ion channels, in particular in the context of well characterized currents that have been observed in vacuolar membranes of many plant species. This lack of progress greatly frustrates assigning physiological function, for example through reverse genetics or heterologous expression.

So far only one, the slow vacuolar (SV), channel has been identified at the gene level and found to be encoded by the TPC1 (two-pore channel) gene (6). The SV channel is ubiquitous in plants and is believed to play a role in cation homeostasis and Ca²⁺ signaling. It forms the predominant tonoplast conductance and was one of the first plant channels to be characterized (7). At least two additional cation channels, the fast-activating vacuolar (FV) (8) and vacuolar K⁺ (VK) (9, 10) channels, are present in many plant vacuoles, and these have been postulated to be involved in guard cell turgor regulation and K⁺ nutrition. However, no evidence for the latter has been provided, nor has their function been related to specific gene products.

In contrast, genomics approaches have identified many families of putative ion channels for which functional data are often absent. These include the two-pore K channel (TPK) (11) family, which contains five members. TPKs show a four transmembrane/two-pore structure, GYGD K⁺ selectivity motifs, and one or two C-terminal EF hands. One isoform, TPK4, was recently shown to encode a plasma membrane conductance expressed in pollen tubes (11). TPK4 shows no or little voltage dependence, is K⁺ selective, exhibits intrinsic rectification, and is believed to participate in the control of the pollen-tube membrane voltage. Another isoform, TPK1, has been characterized in detail regarding membrane and tissue-specific expression. It was found to be expressed in most plant tissues and to be targeted to the tonoplast membrane, where it forms homomeric channels (12, 13). In addition, heterologous expression of TPK1 in yeast vacuoles showed that it constitutes a voltage-independent, Ca²⁺-dependent conductance with a high selectivity for K⁺ over Na⁺ (14). Nevertheless, the presence of TPK1 currents, its electrophysiological properties and its physiological role in planta have yet to be demonstrated. We therefore set out to characterize TPK1 in planta using electrophysiological, reverse genetics, and homologous expression approaches and provide evidence that it encodes the previously characterized VK channel (9), affects general K⁺ homeostasis, and has a role in stomatal functioning and seed germination.

Results

A Specific Vacuolar Current Is Absent in TPK1 Loss-of-Function Mutants.

To assess whether conductance properties of tonoplasts depend on TPK1 expression, four independent TPK1 insertion mutants were obtained from The Salk Institute. Lines salk_131790, salk_146903, sail_167a03, and salk_040082 carry insertions in the 5′ region, first exon, and 3′ region respectively of At5g55630 (Fig. 1A). RT-PCR analysis of the homozygous T-DNA insertion lines showed only lines 146903 (tpk1-1) and 167a03 (tpk1-2) to be devoid of any TPK1 mRNA transcript (Fig. 1B), indicating they were genuine loss-of-function mutants.

Fig. 1.

Identification of T-DNA insertion mutants for TPK1. (A) Schematic view of the four T-DNA insertions in TPK1 (At5g55630). Lines SALK_146903 and SAIL_167A03 both locate to the first exon at positions 503 and 620, respectively. (B) RT-PCR data obtained from WT and mutant leaf mRNA, using AtTPK1 full length primers and actin primers as control showing the absence of transcript in lines tpk1-1 and tpk1-2 (SALK_146903 and SAIL_167A03, respectively).

Fig. 2A shows typical slowly activated, voltage-dependent, vacuolar (SV) currents in tonoplasts derived from wild-type mesophyll vacuoles with a unitary conductance of ≈55 pS in symmetrical 100 mM KCl solutions. These currents are exclusively activated at positive tonoplast voltage and are derived from TPC1 cation channels (6). A second type of current that is far less frequently observed is present in the same membrane patch and activated instantaneously at both positive and negative tonoplast voltages. Both types of current with the same unitary conductances can also be recorded from guard cells (Fig. 2B) and epidermal cells (data not shown). In contrast, patch–clamp analysis of a large number of vacuoles derived from tpk1-1 (n = ≈200) and tpk1-2 (n = ≈100) shoot tissue showed the presence of SV-type currents only in mesophyll (Fig. 2C) or guard cells (Fig. 2D). SV channels in the tpk1-1 and tpk1-2 background showed characteristics similar to those observed in wild-type vacuoles showing that TPK1 is not part of the SV channel as suggested (15). The instantaneous current was never observed in the tpk1-1 and tpk1-2 genetic background (Fig. 2 C and D), strongly suggesting TPK1 expression is essential for its function.

Fig. 2.

Cation channel activity in Arabidopsis mesophyll and guard cell vacuoles. (A) Excised luminal side-out patch from WT mesophyll vacuole showing voltage-dependent outward current (cytoplasm to vacuole) mediated by SV channels, at positive membrane potential. A second, smaller conductance is active at both positive and negative membrane potentials. (B) Excised luminal side-out patch from a WT guard cell vacuole showing similar voltage-dependent and independent currents, as in A. (C and D) TPK1 loss-of-function mutants (tpk1-1 and tpk1-2) show only the presence of SV channel activity in mesophyll (C) and guard cell vacuoles (D). C and D show recordings from tpk1-1 vacuoles, and similar current profiles were obtained from tpk1-2 vacuoles (data not shown). Both bath and pipette solution contained 100 mM KCl, 2 mM CaCl₂, and 5 mM Mes/Tris (pH 6.5). Membrane voltages are denoted at the right, and arrow symbols on the left denote closed levels.

TPK1-GFP Expression Locates to the Tonoplast and Restores Instantaneous Currents.

To further characterize the TPK1-associated conductance, we transiently expressed a TPK1-GFP fusion product in either the tpc1-2 (6), tpk1-1, or tpk1-2 genetic background. The tpc1-2 genotype lacks the SV-type current that tends to obscure less frequently observed currents and thus greatly facilitates channel characterization. Patch–clamp measurements on highly fluorescent tonoplasts showed instantaneous currents (Fig. 3A) in ≈40% of transformed vacuoles. The instantaneous current is not a linear function of membrane voltage but exhibits rectification (compare current at 120 and −120 mV; Fig. 3B), yielding an inward unitary conductance of ≈45 ± 5 pS and an outward conductance of ≈20 ± 10 pS at voltages 60 mV positive or negative of E_K⁺.

Fig. 3.

Single-channel characteristics of the Arabidopsis TPK1-dependent conductance. (A) Single-channel recordings in an excised luminal side-out tonoplast patch from a tpc1-2 mesophyll protoplast transiently transformed with pA7-TPK1-GFP. Note that inward currents are significantly larger than outward currents at comparable membrane voltages. (B) Current–voltage relationship of TPK1 current, showing inward unitary conductance of ≈45 pS and outward unitary conductance of ≈22 pS. (C) Open probability (P_o) as a function of tonoplast membrane voltage. Open probabilities were determined in excised tonoplast patch recordings of 50 sec for each membrane voltage. Solutions were as described for Fig. 2.

In contrast to the strong voltage dependence of SV gating, TPK1-induced currents were prevalent at all measured tonoplast potentials. Fig. 3C shows that the membrane potential did not have any significant effect on the open probability of the TPK1-associated channel. Transient or stable expression of TPK1-GFP in the tpk1-1 or tpk1-2 background [supporting information (SI) Fig. 6 A and B] restored the occurrence of instantaneous currents that were identical to those observed in wild-type vacuoles. Thus, the combined data clearly show that TPK1 is expressed in the tonoplast, and that its expression is essential for the presence of this type of tonoplast current.

TPK1-Encoded Current Is K⁺ Selective, and Its Gating Is Modulated by Cytoplasmic Ca²⁺ and Cytoplasmic pH.

Replacing the 100 mM KCl luminal solution with one containing 20 mM KCl resulted in a 33 mV negative shift of the reversal potential (Fig. 4A), which closely reflects the theoretical reversal potential of −36 mV for a perfectly selective K⁺ channel. Biionic conditions showed the channel has some permeability for NH₄⁺ and Rb⁺ but virtually none for Na⁺, Cs⁺, or Li⁺ (Fig. 4B). On the basis of extrapolated reversal potentials, the following relative permeability sequence was derived: K⁺ (1) > Rb⁺ (0.48 ± 0.13) >NH₄⁺ (0.35 ± 0.10) ≫ Li⁺ (0.02 ± 0.004) ≫ Na⁺ (0.01 ± 0.003) ≫ Cs⁺ (0.01 ± 0.002).

Fig. 4.

TPK1-dependent currents are K⁺-selective and modulated by cytoplasmic [Ca²⁺] and cytoplasmic [H⁺]. (A) Current–voltage plot of TPK1 currents in symmetrical KCl (100 mM) solution (filled symbols) or when luminal KCl was reduced to 20 mM (open symbols). (B) When the luminal KCl was replaced with equimolar solutions of RbCl (open squares), NH₄Cl (filled circles), LiCl (filled squares), NaCl (open circles), or CsCl (filled triangles), the extrapolated reversal potentials varied from −15 (RbCl) to −115 (CsCl) mV, indicating residual permeability for Rb⁺ and for NH₄⁺ but virtually none for other monovalent cations. (C) When the cytoplasmic Ca²⁺ levels drop below ≈50 μM, TPK1 channel open probability steeply declines, and channel activity totally disappears when Ca²⁺ concentrations are less than ≈500 nM. Recordings were made over 30 sec on a cytoplasmic side-out tonoplast patch in 100 mM KCl on tpc1-2 vacuoles overexpressing TPK1. P_o was calculated as total open time divided by total time normalizing the pCa_cyt²⁺ = 3–100%. (D) Representative recording on a cytoplasmic side-out excised patch showing that the TPK1 channel open probability has a cytoplasmic pH optimum that is slightly acidic at ≈pH 6.7. P_o was calculated as described above, normalizing the maximum value to 100%.

A number of channel blockers was applied to investigate channel inhibition (Table 1). The typical Ca²⁺ channel blockers Gd³⁺, La³⁺, and Pb²⁺ did not have discernible effects on the TPK1 current at 100 and 500 μM, respectively. However, millimolar additions of these compounds led to reversible channel blockage. Similar results were obtained with the K⁺ channel inhibitors Ba²⁺ and TEA, both of which almost completely inhibited TPK1-mediated current at a concentration of 2 mM. The nonspecific channel blocker Zn²⁺ had only a slight effect on channel conductance even at 7 mM. The most potent channel blockers were quinine and quinidine, well known for their capacity to block animal Ca²⁺-activated K⁺ channels, which both inhibited TPK1 current with a K_i in the micromolar range.

Table 1.

Pharmacology of TPK1 current

Channel inhibitor	Range tested, mM	Ki, mM
La3+	0.1–5.0	1.7 ± 0.4
Gd3+	0.1–5.0	4.3 ± 1.1
Pb2+	0.1–5.0	3.6 ± 1.3
Zn2+	0.5–10.0	ND
TEA+	1.0–5.0	1.7 ± 0.4
Ba2+	1.0–5.0	1.3 ± 0.3
Quinidine	0.02–1.0	0.11 ± 0.03
Quinine	0.02–1.0	0.07 ± 0.02

Compounds were added directly to the bath solution from concentrated stocks, and the effect on TPK1-mediated current was recorded. Data are given as the concentration where 50% inhibition of the TPK1 current was observed (K_i) and are averages ± SEM of three or four independent experiments.

Both luminal and cytoplasmic Ca²⁺ concentrations were varied to study whether the TPK1-associated instantaneous currents are sensitive to this divalent cation. Changes in luminal [Ca²⁺] from 2 to 0.05 mM had no significant effect on either channel conductance or channel open probability. However, the cytoplasmic Ca²⁺ had a profound effect on the TPK1 open probability (Fig. 4C and SI Fig. 7). Channel open probability steeply decreased when cytoplasmic Ca²⁺ was lowered to <50 μM, and TPK1 currents completely ceased to be observed at cytoplasmic Ca²⁺ concentrations of less than ≈0.5 μM.

In addition to cytoplasmic Ca²⁺, cytoplasmic pH affected TPK1 open probability (Fig. 4D). Open probability showed a maximum at around pH 6.7, steeply dropped off toward more alkaline pH, and moderately decreased at more acidic pH. These results suggest that at physiological pH (≈7.5–7.8), the TPK1 open probability is only ≈20–30% of the maximum open probability.

TPK1-Associated Tonoplast Current Strongly Resembles VK Currents.

VK-type currents were first described in Vicia faba guard cell vacuoles (9). Their characterization and subsequent reports (e.g., refs. 10 and 16) showed the VK channel to be voltage-independent and highly selective for K⁺ with residual Rb⁺ and NH₄⁺ permeability. Conductance for other monovalents such as Na⁺, Li⁺, and Cs⁺ was negligible. The V. faba VK channel shows intrinsic rectification with an ≈3-fold larger inward than outward current in symmetrical solutions (10). Activation of the V. faba VK channel requires a cytoplasmic [Ca²⁺] of ≈1 μM and showed maximum open probability at a cytoplasmic pH of ≈6.5 (9).

The V. faba VK channel characteristics are remarkably similar to those presented in this study for the Arabidopsis TPK1 channel; both the Vicia VK and TPK1 channels are highly selective for K⁺, show intrinsic rectification, and are voltage-independent. Furthermore, channel activity in both cases depends on cytosolic Ca²⁺ concentration and requires a minimum concentration of between 0.5 and 1.0 μM. Maximum channel current occurs at cytosolic pH optima of 6.5 and 6.7 for Vicia VK and Arabidopsis TPK1, respectively. Thus, our data strongly suggest that TPK1-associated currents form the basis for the previously characterized guard cell VK channels.

VK currents were hitherto described only in Vicia guard cells. However, Fig. 2A clearly shows that Arabidopsis mesophyll cells also show VK-type currents. A more in-depth analysis of different shoot cell types showed that VK-type single-channel currents were indeed most prevalent in guard cell vacuoles (≈30%, n = 37) followed by mesophyll cell vacuoles (≈10%, n = 212) and occurred at the lowest frequency in epidermal cell vacuoles (≈6%, n = 43). No significant differences were observed regarding the unitary conductance of the VK currents in the various cell types.

TPK1 Is Involved in Inter- and Intracellular K⁺ Distribution.

WT, TPK1 loss-of-function (KO) mutants (tpk1-1 and tpk1-2) and TPK1-overexpressing lines (TPK1ox) were grown in a large range of different conditions to assess the function of TPK1 in planta. In a number of conditions, significant differences were observed between the genotypes. When plants were grown on media containing high (80 mM) or low (10 μM) levels of K⁺, both high and low K⁺ conditions led to reduced growth of the KO lines compared with WT plants (Fig. 5A), whereas the opposite phenotype was apparent in the TPK1ox lines. This result was obtained for both total seedling fresh weight and root length (data not shown). However, the amount of K⁺ measured in the seedlings was not significantly different in the genotypes. The observation that seedling growth is negatively affected at both high and low K⁺ conditions and the absence of significant differences in overall tissue K⁺ concentrations suggests that TPK1 is not involved in K⁺ uptake per se but functions in inter- and intracellular K⁺ homeostasis.

Fig. 5.

TPK1 expression affects K⁺ homeostasis, stomatal function, and seed germination. (A) When loss-of-function mutants (tpk1-1 and tpk1-2), WT, and TPK1-overexpressing (TPK1ox) plants are grown on plates containing low (10 μM) or high (80 mM) KCl, growth of the KO and TPK1ox lines is affected (Upper), but no significant difference in K⁺ contents is apparent (Lower). Data represent fresh weights and tissue K⁺ concentrations on fresh-weight basis of 14-day-old plants for four independent experiments. (B) Whole-leaf conductance rapidly decreases when leaves are exposed to 10 μM ABA but significantly slower in tpk1 KO lines (triangles) than in WT leaves (circles). Conductance reduction was quickest in the TPK1ox lines. Data show averages ± SEM for five independent experiments by using four leaves per experiment for each genotype. Conductances in control conditions were: 140 ± 13, 133 ± 12, 124 ± 10, 115 ± 11 mol m⁻²·s⁻¹ for tpk1−1, tpk1−2, WT, and TPK1ox, respectively. (C) On control agar plates (Top), germination in tpk1 KO lines is slower than in WT seeds, whereas that of TPK1ox seeds is faster. The differences in germination rates become more significant in the presence of 2.5 μM ABA (Upper Middle), and even more pronounced in the presence of 5 μM ABA (Lower Middle). Bottom shows representative seedling growth of the three genotypes 2 days after imbibition on agar plates containing 2.5 μM ABA. Germination was scored on the basis of radicle protrusion through the endosperm layer after 24 h (control), 48 h (2.5 μM ABA), and 36 h (5 μM ABA). Approximately 100 seeds per plate per genotype were used, and germination percentages were derived from five to six independent experiments. Data are means ± SEM, and Student's t tests identify significant differences at the P < 0.1 (∗) or P < 0.05 (∗∗) level.

Because the VK current was hypothesized to participate in stomatal functioning, we examined this process in the three genotypes by analyzing whole-leaf conductance in response to stimuli known to alter stomatal apertures. No significant difference in leaf conductance was apparent after leaves had been preincubated in an “opening” buffer in the light. However, when abscisic acid (ABA) was added to the buffer, leaf conductance dropped considerably slower in leaves derived from KO plants compared with WT, whereas stomatal closure in TPK1ox lines was slightly faster than that recorded in WT (Fig. 5B). Nevertheless, after ≈2 hours of incubation with ABA, the final leaf conductances showed no significant differences. When leaves of tpk1-1 and tpk1-2 were exposed to the fungal toxin fusicoccin, a strong opening stimulus, no significant differences were observed in the kinetics of increasing conductance (SI Fig. 8). These data suggest that the loss of TPK1 in guard cells decreases vacuolar K⁺ release and hence slows the kinetics of stomatal closure but does not affect K⁺ uptake during stomatal opening.

TPK1 is highly expressed in the embryo (12), and we therefore examined whether its expression had any effect on germination. Interestingly, germination in KO lines was consistently slower than in WT, whereas that of TPK1ox lines was faster (Fig. 5C). This difference is apparent only during the first 2 days after stratification and disappears after 3–4 days (SI Fig. 9). However, the disparity in germination rates became far more pronounced in the presence of increasing concentrations of ABA (Fig. 5C and SI Fig. 9), suggesting that the KO lines are more sensitive to ABA. We examined whether this difference in germination rates was correlated to external and/or internal levels of K⁺. However, germination rates in the presence (20 or 50 mM) and absence of K⁺ were not significantly different (data not shown), suggesting that the early stages of germination do not require uptake of external K⁺. The K⁺ levels in seeds, germinated seeds (1–2 days) in control conditions and in the presence of ABA (2–4 days) were also comparable between genotypes (SI Fig. 10), indicating that tissue K⁺ levels were not responsible for the observed differences in germination.

Discussion

Plant vacuoles are dynamic compartments that function in turgor maintenance and cell signaling and as a depository for both xenobiotics and beneficial compounds. Essential minerals such as K⁺ are stored in vacuoles to provide turgor and cell movement and as a buffer for cytosolic K⁺ homeostasis (1, 17, 18). All these functions require adequate transport mechanisms, for example, to mediate the large transtonoplast fluxes of K⁺ during stomatal regulation or cell expansion.

Three K⁺ permeable conductances have been described at the tonoplast: the FV-, SV-, and VK-type channels. No genes have so far been associated with the FV channel, and its role in plants is largely unknown. The SV channel was recently identified as being encoded by the two-pore channel TPC1 and may function in guard cell and ABA signaling during seed germination (6).

Here we show that a loss of function in TPK1 leads to the loss of a tonoplast current, which has all of the hallmarks of the previously characterized guard cell VK current (9, 10); it is voltage-independent, K⁺-selective, intrinsically rectifying, and regulated by cytoplasmic Ca²⁺ and pH. Expression of TPK1 is crucial for the occurrence of this tonoplast conductance and in conjunction with a recent report on heterologously expressed TPK1, which showed currents very similar to those described here (14), it is tempting to conclude that TPK1 expression is not only essential, but it is also sufficient to establish VK currents in planta.

TPK1 is highly selective for K⁺ and when plants are exposed to low (10 μM) or high (80 mM) K⁺ levels, changes in seedling growth were correlated to TPK1 expression. However, the overall tissue K⁺ levels were not. These data suggest that in such conditions, TPK1 is involved in intracellular K⁺ redistribution and/or affects K⁺ retranslocation between different tissues. Both processes are likely to become increasingly important in conditions of very high or very low ambient K⁺, to maintain turgor for cell expansion but also to maintain adequate cytosolic K⁺ levels (1).

Our patch–clamp experiments show the presence of VK currents most frequently in guard cells. Regulation of stomatal aperture depends on the movement between the guard cell vacuole and apoplast of K⁺ and other osmotica (19). Our data support a role for the VK channel in this process, as suggested (9). In light conditions, all three genotypes showed comparable whole-leaf stomatal conductances. At 10 μM, the stress hormone ABA induced a rapid decline in WT leaf conductance approximately halving the initial conductance. However, the response to ABA was far slower in the tpk1 KO lines and marginally faster in the TPK1ox lines. In contrast, the kinetics of stomatal opening in response to fusicoccin were not statistically different in the three genotypes.

ABA triggers guard cell K⁺ release that ultimately derives from guard cell vacuoles (19). Thus, our results are consistent with a model where ABA-induced TPK1-mediated K⁺ release from the vacuole is important in stomatal closure. The process contains sufficient redundancy in the form of alternative tonoplast K⁺ transporters to allow stomatal closure in tpk1 KO lines to proceed, albeit with considerably slower kinetics.

Our data derived from stomata provide good evidence for a role of TPK1 in vacuolar K⁺ release. TPK1 also is important during seed germination, a process that is compromised in KO mutants when seeds are grown on control plates but far more so when ABA is present. This suggests that TPK1 function is relatively insensitive to ABA and compensates other processes that are essential for germination and negatively regulated by ABA.

After the loss of dormancy, a range of processes is set into motion, many of which involve vacuoles. Protein storage vacuoles are the main seed depositories for minerals that need to be mobilized to sustain the growing embryo (20). Most minerals are stored as complex salts in vacuolar inclusion bodies and may be released from vacuoles through cation channels (20). At the same time, formation of de novo vacuoles and coalescence of existing small vacuoles occurs, which require deposition of inorganics such as K⁺ to promote H₂O influx. These two processes would therefore require opposite K⁺ fluxes, both of which can be mediated by TPK1, depending on local transtonoplast K⁺ gradients and membrane voltages. Our germination data do not provide conclusive evidence to specify a role for TPK1 in these processes. However, a number of observations point to a role in cell expansion and thus vacuolar K⁺ accumulation; TPK1 expression levels are similar in embryonic and endosperm tissues (20) but are rapidly induced to ≈10-fold 4–5 days after seed imbibition (I. Graham and S. Penfield, University of York, personal communication). During this phase, cellular expansion becomes increasingly important. Furthermore, Arabidopsis seed germination is biphasic (20, 21), with an initial seed-coat rupture and a subsequent rupture of the endosperm layer. The first phase of germination is relatively little affected by TPK1 expression (SI Fig. 9), whereas the second phase, associated with rapid radicle expansion, is affected. In accordance with the observed phenotype, the second germination phase is severely inhibited by the presence of ABA, whereas seed-coat rupture is not inhibited (21). These results suggest TPK1 participates in generating turgor by vacuolar K⁺ deposition. With tonoplast potentials in the region of −10 to −20 mV (1) and low vacuolar K⁺ concentrations, TPK1 would be capable of mediating vacuolar K⁺ accumulation when activated. The necessary K⁺ does not derive from external sources but could originate from globoid inclusions (20, 21) in various seed tissues, because the overall K⁺ levels in seeds are considerably higher than those observed in 1- to 2-day-old seedlings (ref. 22; SI Fig. 10).

Interestingly, the TPK1 loss-of-function germination phenotype is the opposite of that observed with TPC1 where loss of function resulted in decreased ABA sensitivity. Both channels are capable of conducting transtonoplast K⁺ fluxes, but TPC1 can also conduct other cations. The contrasting phenotypes suggest that TPC1 is involved not in K⁺ transport but possibly in Ca²⁺ release as an intermediate in ABA signaling (6).

In conclusion, our studies show that TPK1 encodes a native tonoplast channel that can be recorded in mesophyll, epidermal, and guard cells and is equivalent to the previously identified VK conductance. The voltage independent gating properties of TPK1 make it an ideal system to have a role in intracellular K⁺ distribution. In guard cells, TPK1 participates in vacuolar K⁺ release during stomatal closure and may be directly or indirectly activated by ABA. During seed germination, TPK1 may be involved in the vacuolar release of mobilized K⁺ reserves but is more likely to participate in the latter stages of radicle growth through vacuolar K⁺ loading to provide turgor driven cell expansion.

Methods

Plant Growth and Treatments.

Arabidopsis thaliana (L) ecotype Columbia (0) WT, TPKox, and insertional mutant seeds were surface-sterilized and placed on agar plates. Growth medium composition was as described (11) and contained 1.25 mM KNO₃, 0.5 mM Ca(NO₃)₂, 0.5 mM MgSO₄, and 0.625 mM KH₂PO₄ as macronutrients. After stratification at 4°C for 2 days, seeds were transferred to a growth cabinet or growth room with the following conditions: 12 h, 120 μmol·m⁻²·s⁻¹ light intensity, 24/20°C day/night temperature, relative humidity 70–80%).

TPK1 Loss of Function, Overexpression, and Complementation.

T3-segregating T-DNA mutants for TPK1 were obtained from the SALK institute (http://signal.salk.edu/cgi-bin/tdnaexpress) and genotypically tested through PCR to identify homozygous mutants using 5′-ctagtatgtcgagtgatgcagctcgtac-3′ and 5′-acggtccctttgaatctgagacgtgg-3′ forward and reverse primers for line SALK146903 and 5′-gacgaaccacctcctcatcc-3′ and 5′-ctcgaactcatccattatccca-3′ forward and reverse primers for line SAIL167A03. Amplified fragments were cloned (TA-cloning vector) and sequenced to verify the position of the T-DNA insertion. Homozygous lines were subjected to RT-PCR with gene-specific primers described above to confirm the absence of transcript in lines tpk1-1 and tpk1-2. For transient complementation, Arabidopsis mesophyll protoplasts were isolated from tpk1-1, tpk1-2, and tpc1-2 (6) loss-of-function mutants and transformed with pA7-TPK1-GFP (13) plasmids as described (6). After 24–48 h of incubation, GFP fluorescence of protoplasts and vacuoles was observed by using an epifluorescence microscope, and fluorescent vacuoles were used for electrophysiology experiments.

All plant transformations were done with Agrobacterium tumefaciens (GV3101) by using the floral dipping method. Overexpressors of TPK1 were generated by transferring p35S-TPK1 (12) into Arabidopsis WT. Experiments were conducted with T2 or T3 plants of three independent lines (nos. 3, 5, and 8) showing 14- to 23-fold increased expression of TPK1. Complemented plants were generated by subcloning TPK1 into pBART27-EYFP as a C-terminal fusion and transformation of tpk1-1 with this construct.

Electrophysiological Experimentation.

A. thaliana leaf protoplasts were isolated as described in ref. 17 for root protoplasts, with the only alteration being the presence of 0.3% Macerase during enzymatic tissue digestion. After transfer of leaf protoplasts to the recording chamber, vacuoles were released by washing protoplasts with a solution containing 10 mM EDTA, 10 mM EGTA (pH 8), with an osmolarity of 350 mOsm. Data acquisition occurred through a CED (Cambridge, U.K.) A/D converter at 0.5–3 kHz (depending on experimental conditions), whereas data were typically filtered at 0.5 kHz. Results were analyzed by using CED software (version 6). For open-probability analyses, the total open time of all channels in the respective membrane patch was obtained during 50-sec recordings as a function of membrane voltage, cytoplasmic pH, or cytoplasmic and luminal Ca²⁺ concentration. To determine the voltage dependence of the open probability, the number of single channels per patch was established, and constant opening for all channels was set as 100%. For the pH and Ca²⁺ dependence of the open probability, the maximum conductance was set as 100%. Experimental solutions are described in the figure legends and were all adjusted to 360 mOsm with sorbitol.

Whole-Leaf Conductance Measurements.

Leaves of mature plants were incubated in “opening” buffer consisting of 10 mM KCl, 10 mM Mes–KOH (pH 6.15) for 2 h in the light to induce maximum opening. Subsequently, H₂O gas exchange of four leaves per treatment was determined by removing individual leaves from the experimental solution and placing them in the measuring chamber of an Infrared Gas Analyser, Li-Cor 6400 (LI-COR, Cambridge, U.K.) to serve as a control value (t = 0). The same leaves were then treated with ABA by adding concentrated ABA to the “opening” buffer to a final concentration of 10 μM. Reduction in leaf conductance was followed over time on the same set of leaves by repeatedly measuring leaf conductance at the time points indicated. Control experiments on non-ABA-treated leaves showed <10% average variation in leaf conductance over the measurement period. Four to five independent experiments were carried out by using different batches of plants.

Germination Assays.

Five- to 8-week-old sterilized, stratified seeds (48 h, 4°C) derived from parallel-grown WT, tpk1-1, tpk1-2, and three lines of TPK1 overexpressors were placed on agar plates with the medium, as described above. Seeds were scored as “germinated” when, at ×10 magnification, a clearly visible radicle protruded through the seed coat.

Tissue Ion Analysis.

Imbibed seeds and seedlings were dried for 48 h at 80°C, and subsequently K⁺ was extracted with a 20 mM CaCl₂ solution for 5 h. K⁺ concentrations in the CaCl₂ solution were determined by using a flame photometer (Sherwood, Cheadle, U.K.).

Data Analysis.

Where appropriate, data are represented as means ± SEM and Student's t tests were applied to identify significant differences at the P < 0.1 (∗) or P < 0.05 (∗∗) level. Data for TPK1-overexpressing lines are averages of three independent lines.

Abbreviations

SV

slow vacuolar

FV

fast-activating vacuolar

TPK

two-pore K⁺ channel

ABA

abscisic acid.