Arabidopsis thaliana vacuolar TPK channels form functional K⁺ uptake pathways in Escherichia coli

Abstract

Very few vacuolar two pore potassium channels (TPKs) have been functionally characterized. In this paper we have used complementation of K⁺ uptake deficient Escherichia coli mutant LB2003 to analyze the functional properties of Arabidopsis thaliana TPK family members. The four isoforms of AtTPKs were cloned and expressed in LB2003 E. coli background.The expression of channels in bacteria was analyzed by RT-PCR. Our results show that AtTPK1, AtTPK2 and AtTPK5 are restoring the LB2003 growth on low K⁺ media. The analysis of potassium uptake exhibited elevated level of K⁺ uptake in the same three types of AtTPKs transformants.

Introduction

Plant vacuoles are intracellular compartments that can make up 90% of the cellular volume. Vacuoles play essential roles as depositories for water and minerals to provide turgor, in the storage of xenobiotics and toxic compounds and in the storage of nutrients. In addition, vacuoles can form lytic compartments for protein degradation and are crucial for cell signaling.

Several cation conductances have been recorded in the vacuolar membrane, the tonoplast, which include K⁺ channels from the two pore K (TPK) channel family. TPKs have a four transmembrane/two pore structure, each pore containing a GYGD K⁺ selectivity motif and one or two predicted C-terminal EF hands. Functional channels are believed to consist of dimers. The Arabidopsis TPK family has five members and similar numbers have been found in other species.^1-3 AtTPK4 was shown to be expressed in the plasma membrane where it potentially contributes to membrane potential regulation of pollen tubes.^4,5 Arabidopsis TPK1 was shown to be involved in K⁺ homeostasis, stomatal function and germination.^6,7 Based on fluorescent protein fusions, the three other members of the Arabidopsis TPK family (TPK2, 3 and 5) all localize to the tonoplast as do at least two rice TPK isoforms.^1,8 Expression levels of the vacuolar AtTPK2, 3 and 5 are considerably lower than that of TPK1 and show more tissue specificity. For example, TPK3 expresses in root tips and in pollen whereas TPK5 is more prevalent in vascular tissues, hydathodes and floral organs.⁹ TPK2 expression is generally very low in all tissues.¹⁰ Only very few vacuolar TPKs have been functionally characterized.^6,11,12 AtTPK1and NtTPK1 subunits readily form functional K⁺ selective channels, both when expressed homologously and in heterologous systems.^6,11-13 The latter suggests AtTPK1 and NtTPK1 subunits form homodimers.¹¹ Channel properties include inward rectification and modulation of channel activity by Ca²⁺, 14-3-3 proteins and cytosolic pH.^6,12,14 However, no functional data are available on any of the other tonoplast TPK isoforms from Arabidopsis. Whereas homologous and heterologous expression of AtTPK1 and its orthologs readily results in clearly identifiable channel activity, expression of AtTPK2, 3 and 5 in various systems does not appear to produce channel currents. This suggests that AtTPK2, 3 and 5 may encode “silent” subunits incapable of forming active channels when expressed on their own. Such silent subunits could associate with subunits from AtTPK1 to form various heterodimer combinations with slightly different channel properties. However, GUS reporter expression studies show little overlap in expression patterns of different TPK isoforms whereas bimolecular fluorescence complementation studies indicate TPKs express as homodimers.⁹ Thus, an important question remains concerning potential physiological roles of AtTPK2, 3 and 5 and this is closely connected to the question whether AtTPK2, 3 and 5 subunits can form functional proteins. Previous reports showed that NtTPK1 was capable of complementing the K⁺ uptake deficiency phenotype of the E. coli strain LB2003 which lacks both Trk and Kdp K⁺ transporters.¹⁵ We used a similar complementation strategy to test whether AtTPK2, 3 and 5 channel subunits can rescue the low K⁺ growth phenotype of these bacteria. Our data show that, apart from AtTPK1, TPK2 and TPK5 are competent in restoring E. coli growth on low K⁺ media. Furthermore, K⁺ uptake assays show expression of the same three isoforms significantly increases bacterial K⁺ uptake.

Results

Growth of E. coli LB2003 on low K⁺ media

E. coli strain LB2003 is incapable of growing on media that contain less than around 0.1 mM K⁺, due to the lack of the high and medium affinity K⁺ uptake transporters Trk (TrkG and TrkH), Kup (TrkD) and Kdp. It has been used previously to characterize plant ion channels.^15,16 It was observed that with ambient K⁺ levels of 0.1 mM K⁺ strain LB2003 transformed with empty vector (EV) has an exceedingly slow growth rate compared with the wild type strain transformed with empty vector (WT-EV) (Fig. 1A). To test whether this defect could be overcome by plant K⁺ channels, LB2003 was transformed with various Arabidopsis TPK cDNAs. RT-PCR analyses showed that in each case full length TPK transcript was present after bacterial transformation (Fig. S1). E. coli strains transformed with AtTPKs showed growth curves that were intermediate between WT-EV and EV respectively (Fig. 1A).These qualitative differences in growth rates are clearly dependent on K⁺ in the growth medium: The differences in growth rates largely disappear when external K⁺ is raised to 3 mM (Fig. 1B).

Figure 1.E. coli growth. E. coli wildtype cells (WT-EV), the LB2003 mutant (EV) and bacteria transformed with AtTPK1, AtTPK2, AtTPK3 and AtTPK5 were grown in liquid yeast extract/tryptone medium containing either 0.1 mM (A) or 3 mM (B) K⁺ in the medium. OD 600 values were obtained for each genotype for a period of 8 h (0.1 mM K⁺) or 6 h (3 mM K⁺).

To get a more detailed picture of how AtTPKs affect E. coli growth as a function of external K⁺ we grew all genotypes in liquid media without added K⁺ (“0K”), with 0.1 mM K⁺, 3 mM K⁺ and 150 K⁺ mM in the solution. The growth doubling time in minutes, determined during the exponential growth phase, was substantially larger for EV and TPK3 transformed E. coli in the presence of 0 and 0.1 mM K⁺ in the medium when compared with WT-EV, TPK2 or TPK5 (Fig. 2). When external K⁺ is 3 mM, there is no significant difference between WT-EV and EV indicating that transport mechanisms other than Trk and Kdp mediate sufficient K⁺ uptake for growth. Not surprisingly, in this condition all four strains expressing Arabidopsis TPKs showed growth rates comparable to that of WT-EV and EV. Although the doubling times of TPK1, TPK2 and TPK5 expressing strains were larger than that of WT-EV, statistical analysis did not show significant differences between samples (Fig. 2).

Figure 2.E. coli doubling times. E. coli wildtype cells (WT-EV), the LB2003 mutant (EV) and bacteria transformed with AtTPK1, AtTPK2, AtTPK3 and AtTPK5 were grown in liquid yeast extract/tryptone medium without added K⁺ (0K⁺) or with 0.1 mM, 3 mM or 150 mM K⁺ in the medium. Doubling times were derived from OD 600 values during the exponential growth phase. Asterisks indicate significant differences based on unpaired two-tailed t-tests (p < 0.05).

K⁺ uptake in E. coli LB2003 transformants

To test whether the AtTPK dependent growth complementation was due to actual K⁺ uptake we performed 1 h K⁺ uptake assays in cells previously starved for K⁺. Our results show that with 0.1 mM K⁺ in the assay buffer the net K⁺ influx in WT-EV cells is around 2 nmol/gFW/h (Fig. 3). In the EV cells uptake is “negative”, i.e., cells are actually loosing K⁺ during the assay. Net K⁺ flux in TPK3 expressing cells is similar to that in EV cells with a loss of around 0.5 nmole per hour. Cells that express AtTPK1 or TPK2 show net uptake albeit smaller than that observed for WT-EV cells. With around 3 nmole per hour, the largest uptake was recorded for cells that expressed AtTPK5. The pattern of net K⁺ flux changed greatly when assays were performed with 3 mM K⁺ in the external buffer. At this external concentration, a net influx was recorded for EV cells but it was significantly smaller than that of WT-EV cells. In this condition there was also net influx for TPK3 expressing cells but this flux was not significantly different (p < 0.05) from that in EV cells. Net influx in TPK1 and TPK2 expressing cells mirrored that measured in WT-EV cells whereas the net influx in TPK5 expressing cells was not significantly different (p < 0.05) from EV (Fig. 3).

Figure 3.

K⁺ uptake in E. coli expressing Arabidopsis TPKs. K⁺ starved E. coli cells were exposed to 0.1 or 3 mM K⁺ for one hour and the difference in K⁺ content before and after was measured to calculate net uptake for wildtype E. coli (WT), the LB2003 mutant (EV) and bacteria transformed with AtTPK1, AtTPK2, AtTPK3 and AtTPK5. Negative values denote net K⁺ loss from cells. Asterisks indicate significant differences based on unpaired two-tailed t-tests (p < 0.05).

Concluding Remarks

Taken together our data show that several Arabidopsis TPKs can complement the E. coli LB2003 phenotype. Growth data (Figs. 1 and 2) show that in particular TPK1, 2 and 5 are competent in restoring growth to almost wild type levels in conditions where EV cells grow only very slowly (0 and 0.1 mM K⁺). In contrast, expression of TPK3, as confirmed by RT-PCR (Fig. S1), did not have a substantial impact on growth in these conditions. Expression of TPK1, 2 and 5 led to a substantially larger K⁺ uptake in short-term assays (Fig. 3) compared with EV cells. This occurred irrespective of using 0.1 or 3 mM K⁺ in the uptake buffer. Again, no or only a very small increment in net K⁺ uptake was found in cells expressing AtTPK3. Clearly, AtTPK3 does not complement the LB2003 growth defect, nor does it augment K⁺ uptake in these cells and this leads us to conclude that AtTPK3 does not form functional K⁺ channels when expressed in E. coli. In contrast to AtTPK3, AtTPK1, TPK2 and TPK5 not only complement the LB2003 growth defect but their expression leads to substantially increased K⁺ influx (Figs. 1–3). Our results therefore strongly suggest that in addition to TPK1, AtTPK2 and AtTPK5 can form functional K⁺ transport systems in E. coli and as such may also form functional channels in planta.

Materials and Methods

Plasmids and E. coli Strains

E. coli strain LB2003 [∆trkA kup1 (trkD1) ∆kdpABC5 rpsL metE thir ha gal] was kindly provided by Evert Bakker (University of Osnabruck).¹⁵

The cDNA from RNA isolated from Arabidopsis seedlings and buds was used to obtain full length clones of AtTPK1, AtTPK2, AtTPK3 and AtTPK5. Clones were amplified with the following primers:

AtTPK1BamHI_for GCGGATCCTGATGTCGAGTGATGCAGCTCG and AtTPK1SmaI_rev GCCCCGGGCCTTTGAATCTGAGACGTGG for AtTPK1; AtTPK2BamHI_for GCGGATCCTGATGGCTAACGACGGTAACGG and AtTPK2SmaI_rev GCCCCGGGAATAGAAGTTGCAGTGGGTA for AtTPK2; AtTPK3BamHI_for GCGGATCCTGATGGCCAACGAAGGAAGTGA and AtTPK3KpnI_rev GCGGTACCGCATCGCCACTGCCACCTTC for AtTPK3; AtTPK5SacI_for GCGAGCTCGCATGGAACCACTCATCAGCCC and AtTPK5Pst1_rev GCCTGCAGGCCAAAGGATCCCCCAAAAGATCAGG for AtTPK5. As PCR template, 20 ng of cDNA was used. PCR was performed in 50 μl volume with 1× Phusion HF PCR buffer, 200 μM of dNTP, 3% DMSO and 1× Phusion polymerase (Finnzymes). The cycling profile was: 95°C for 30 sec; 36 cycles at 95°C for 10 sec, 72°C for 30 sec; final extension at 72°C for 10 min. The full length clones were subcloned into the pQE-32 (Qiagen) expression vector using different restriction sites: AtTPK1 and AtTPK2 by BamHI and SmaI; AtTPK3 by BamHI and KpnI; AtTPK5 by SacI and KpnI.

Bacterial growth assays

Bacteria were transformed with pQE32-AtTPKs and pQE32-EV plasmids and grown overnight in high K⁺ medium (KLM) containing 5 g/liter yeast extract, 10 g/liter tryptone, 150 mM KCl and 100 mg/liter ampicillin. From the overnight culture, 100 μl was diluted in 5 ml of KLM and grown to OD 600 = 0.5. Channel expression was induced by the addition of 1 mM IPTG. After 2 h of induction, the cell density was measured and the culture was diluted (normalized) to an OD 600 of 0.5. five-μl drops of the diluted cell suspension were spotted on ampicillin plates containing 5 g/liter yeast extract, 10 g/liter tryptone and either 0, 0.1, 3 or 150 mM KCl with IPTG and ampicillin. Reproducibility of complementation assays was confirmed in independent replicate experiments and data are shown for 3–5 assays for each condition and genotype. In order to assay bacterial growth in liquid medium, the culture density of induced transformants was normalized as described above. After centrifugation for 2 min at 10,000×g and wash with fresh medium, cells were inoculated into liquid medium containing 5 g/liter yeast extract, 10 g/liter tryptone, with IPTG and ampicillin and different concentrations of K⁺ (0, 0.1, 3 or 150 mM KCl). Experiments were repeated 3–5 times and data are expressed as means ±SD.RT-PCR RNA from pelleted samples of induced transformants was isolated with the NucleoSpin RNA II kit (Macherey-Nagel) according to manufacturer’s instructions. cDNA obtained from RNA preps was used as template for PCR amplification of full length AtTPKs clones with the same set of primers described above.

K⁺ uptake assay

Bacteria were grown on liquid medium containing 5 g/liter yeast extract, 10 g/liter tryptone, 10 mM KCl with IPTG and ampicillin to an OD600 = 0.5. After centrifugation for 5 sec at 4,000 RPM, bacteria were resuspended in liquid medium with 0 mM KCl and IPTG and left for 3 h for K⁺ starvation. After 3 h of K⁺ starvation the cultures were split in four and spun down for 5 sec at 4,000 RPM. One fraction was incubated on ice as reference. Another three fractions of bacterial cultures were resuspended in 10 ml of liquid medium containing either 0.1 mM or 3 mM KCl and IPTG, and left for 1h for K⁺ uptake. After measuring OD 600 of each tube, the samples (including reference samples) were washed with 0 mM KCl liquid medium. After spinning of samples for 5 sec at 4,000 RPM and removing supernatant, bacterial pellets were resuspended in 2 ml of 5% TCA (Trichloroacetic acid) and left for 0.5 h. The K⁺ measurement of bacterial samples were performed by flame photometer. Net K⁺ uptake (μmol/gFW/h) was calculated from the difference between K+ content in each sample minus that found in the reference sample. Experiments were repeated 3 times and data are expressed as means ±SD.

Statistical analyses

Statistical differences between measurements on different times and variants were analyzed by on unpaired two-tailed t-tests. Differences were considered significant at a probability level of p < 0.05.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.