Characterization of potassium transport in wild-type and isogenic yeast strains carrying all combinations of trk1, trk2 and tok1 null mutations: trk1 , trk2 and tok1 null mutant yeast

Molecular Microbiology (2003)

47

(3), 767–780

© 2003 Blackwell Publishing Ltd

Blackwell Science, LtdOxford, UKMMIMolecular Microbiology0950-382XBlackwell Publishing Ltd, 200347Original Article

trk1, trk2 and tok1 null mutant yeastA. Bertl et al.

Accepted 24 October, 2002. *For correspondence. E-mail [email protected]; Tel. (

+

46) 8 728 7108; Fax (

+

46) 8 33 28 12.

Characterization of potassium transport in wild-type and isogenic yeast strains carrying all combinations of

trk1

,

trk2

and

tok1

null mutations

Adam Bertl,

1

José Ramos,

2

Jost Ludwig,

3

Hella Lichtenberg-Fraté,

4

John Reid,

5

Hermann Bihler,

1

Fernando Calero,

2

Paula Martínez

6

and Per O. Ljungdahl

6

*

1

Universität Karlsruhe, Germany.

2

Universidad de Córdoba, Spain.

3

Universität Tübingen, Germany.

4

Universität Bonn, Germany.

5

AstraZeneca Pharmaceuticals, Wilmington, DE, USA.

6

Ludwig Institute for Cancer Research, Stockholm, Sweden.

Summary

Saccharomyces cerevisiae

cells express threedefined potassium-specific transport systems en-coded by

TRK1

,

TRK2

and

TOK1

. To gain a morecomplete understanding of the physiological functionof these transport proteins, we have constructed a setof isogenic yeast strains carrying all combinations of

trk1

DDDD

,

trk2

DDDD

and

tok1

DDDD

null mutations. The

in vivo

K++++

transport characteristics of each strain have beendocumented using growth-based assays, and the

invitro

biochemical and electrophysiological propertiesassociated with K++++

transport have been determined.As has been reported previously, Trk1p and Trk2pfacilitate high-affinity potassium uptake and appear tobe functionally redundant under a wide range of envi-ronmental conditions. In the absence of

TRK1

and

TRK2

, strains lack the ability specifically to take upK++++

, and

trk1

DDDD

trk2

DDDD

double mutant cells depend uponpoorly understood non-specific cation uptake mech-anisms for growth. Under conditions that impair theactivity of the non-specific uptake system, termedNSC1, we have found that the presence of functionalTok1p renders cells sensitive to Cs++++

. Based on thisfinding, we have established a growth-based assaythat monitors the

in vivo

activity of Tok1p.

Introduction

Potassium is accumulated by most living cells againstlarge concentration gradients.

Saccharomyces cerevisiae

cells express three defined K

+

-specific transport systemsencoded by

TRK1

,

TRK2

and

TOK1

. Each of these trans-port systems shares substantial sequence homology withK

+

transporters that are known to function in bacterial,plant and animal cells (for a review, see Rodríguez-Navarro, 2000). Trk1p and Trk2p are homologous proteinsthat derive energy from the plasma membrane proton-motive force to transport K

+

into cells (Ko and Gaber,1991). Trk1p is a high-affinity K

+

transporter that enablesgrowth at micromolar concentrations of K

+

(Ramos

et al

.,1985; Gaber

et al

., 1988). Yeast lacking

TRK1

requiremillimolar concentrations of K

+

to grow; under these con-ditions, growth is dependent upon the activity of Trk2p, aweakly expressed transporter (Ko

et al

., 1990; Ramos

et al

., 1994; Vidal

et al

., 1995). Cells carrying deletions inboth

TRK1

and

TRK2

lose the capacity to grow on mediacontaining low (

<

10 mM) concentrations of K

+

, but areviable and grow at rates indistinguishable from wild-typecells when propagated on media supplemented with K

+

concentrations

>

50 mM. The growth of

trk1

D

trk2

D

nullmutants in media containing these relatively high concen-trations of K

+

is dependent upon K

+

uptake via non-specificmechanisms (Bihler

et al

., 1998; 2002; Madrid

et al

.,1998).

The

TOK1

gene encodes a voltage-dependent potas-sium channel that displays strong outward rectification(Ketchum

et al

., 1995; Zhou

et al

., 1995; Reid

et al

.,1996). Tok1p contains two pore-forming P domains, struc-tural entities that are common to all known potassium-selective ion channels. Tok1p is the only P domaincontaining protein encoded in the yeast genome (Ahmed

et al

., 1999). The gating of the Tok1p channel is depen-dent upon the membrane voltage; the channel is open atpositive and very low negative membrane voltages (Bertl

et al

., 1993; Fairman

et al

., 1999). Additionally, the activityof Tok1p is modulated by external K

+

(Loukin

et al

., 1997;Vergani

et al

., 1997). Yeast strains carrying null mutationsin

TOK1

are viable. Despite extensive efforts to find con-ditions that specifically affect the growth of

tok1

D

null

768

A. Bertl

et al.

© 2003 Blackwell Publishing Ltd,

Molecular Microbiology

,

47

, 767–780

mutants, no growth-related phenotypes have beenreported (Zhou

et al

., 1995; Reid

et al

., 1996). Thus,although the biophysical properties of the Tok1p channelare well described, the physiological role of Tok1p is pres-ently unclear.

Tok1p is reported to be a molecular target of the yeastviral killer toxin K1 (Ahmed

et al

., 1999). Remarkably, theK1 toxin is thought to act on Tok1p from both sides of theplasma membrane (Sesti

et al

., 2001). The binding of K1toxin to the extracellular side of Tok1p channels increasesthe open probability by destabilizing the closed channelstate. Thus, under certain conditions, the binding of K1toxin to the extracellular side of Tok1p is expected to resultin the efflux of intracellular potassium (Ahmed

et al

.,1999). Lethal mutations within the

TOK1

gene have beenfound that enhance channel activity (Loukin

et al

., 1997);these

tok1

mutations may mimic the effect of the bindingof K1 toxin to the extracellular site. Conversely, the bindingof K1 toxin to an intracellular site of Tok1p has beenshown to inhibit channel opening (Sesti

et al

., 2001).When K1 is bound to the intracellular site, the Tok1pchannel is refractory towards activation by K1 binding tothe external site. The physiological significance of thebinding of K1 to Tok1p has been questioned recently(Breinig

et al

., 2002).In addition to Trk1p, Trk2p and Tok1p, yeast has been

shown to possess other mechanisms that mediate eitherthe import or the export of K

+

. Whole-cell patch clampexperiments have provided evidence for the existenceof a non-selective cation channel (Bihler

et al

., 1998).Although the identity of the protein(s) responsible for thisinwardly rectifying pathway is not known, it has been des-ignated NSC1. Recent experiments have demonstratedthat NSC1 represents the primary low-affinity uptake routefor potassium ions in yeast (Bihler

et al

., 2002). In additionto potassium, this system is also able to facilitate theuptake of a wide range of cations including Li

+

and NH

4

+

.NSC1-mediated inward currents are inhibited by extracel-lular Ca

2

+

and decrease in parallel with reduced pH. Yeastcells express at least one energy dependent K

+

effluxmechanism (Peña and Ramírez, 1991). Although themolecular components of this efflux system have not beenidentified, it has been suggested that Ena1p (also knownas Pmr2p), a putative plasma membrane Na

+

pump, mayaccount for the observed energy-dependent potassiumefflux (Bañuelos and Rodríguez-Navarro, 1998). Finally,under certain conditions, two H

+

/ion exchanger homo-logues, Kha1p and Nha1p, have also been reported tomediate K

+

movements across the plasma membrane(Bañuelos

et al

., 1998; Ramírez

et al

., 1998).Overall, the available information on potassium trans-

port in

S. cerevisiae

is abundant, but fragmentary. Muchof the information published in recent years regarding thefunction of individual potassium transport systems has

been obtained using ill-defined non-isogenic yeast strains.Consequently, the results obtained in different laborato-ries cannot be readily compared. The studies in whichisogenic strains have been used focus exclusively onpotassium uptake mediated by Trk1p and Trk2p, and aretherefore incomplete. To gain a more complete under-standing of the physiological significance of each of theindividual potassium transport systems, we have engi-neered a set of isogenic yeast strains carrying all combi-nations of

trk1

D

,

trk2

D

and

tok1

D

null mutations. Thepotassium transport characteristics of this complete set ofstrains have been documented using biochemical, physi-ological and biophysical techniques. Finally, we report thefirst growth-related phenotype conferred by

tok1

D

nullmutations and describe a growth-based assay that moni-tors the

in vivo

activity of Tok1p.

Results

We have constructed a series of isogenic yeast strainscarrying all possible combinations of

trk1

D

,

trk2

D

and

tok1

D

null alleles. These strains are derived from a singlewild-type strain JRY379 (

MATa his3D200 leu2-3,112trp1D901 ura3-52 suc2D9) (Reid et al., 1996). At one pointduring the construction of strains, strain PLAS132-6C(trk1D51 tok1D1::HIS3), which had previously been prop-agated on YPD without supplementing potassium, wascrossed with strain PM1 (trk2D50::kanMX). The resultingdiploid was sporulated, and tetrads were dissected onYPD + 50 mM KCl. Spore viability was excellent and, inall instances, the three deletion alleles segregated 2:2;spore-derived strains carrying all possible combinationsof deletions were recovered at the expected frequency.However, it was noticed that certain segregants carryingthe trk1D51 allele grew slightly better than other seg-regants with the same presumed genotype. The patternof growth suggested that a spontaneous suppressingmutation, which enhanced the growth of strains carryingthe trk1D null mutation, was independently segregatingin the cross. Subsequent analysis indicated that thesuppressing phenotype was closely linked to the intactTRK2 gene locus, suggesting that a mutation thatpositively affected Trk2p function had been selectedinadvertently. To eliminate the suppressor mutation, thetrk1D51 trk2D50::kanMX tok1D1::HIS3 triple mutant strainPLAS133-28D, which did not carry the suppressing muta-tion, was backcrossed to JRY379. The resulting diploidstrain was sporulated, and tetrads were dissected on YPDcontaining 50 mM KCl. All the spore-derived colonies withmatching genotypes grew at identical and indistinguish-able rates. The eight strains used in the present study,PLY232–PLY246, are meiotic segregants from thiscross.

Many laboratory yeast strains are known to carry RNA

trk1, trk2 and tok1 null mutant yeast 769

© 2003 Blackwell Publishing Ltd, Molecular Microbiology, 47, 767–780

viruses that programme the secretion of toxins that killvirus-free cells (Fink and Styles, 1972; Vondrejs et al.,1996; Wickner, 1996; Schmitt and Eisfeld, 1999). The K1killer toxin has been reported to affect the channel activityof Tok1p (Ahmed et al., 1999; Sesti et al., 2001). K1-infected strains of yeast are naturally immune to the actionof K1 toxin. We determined whether our wild-type and K+

transport deletion yeast strains were carriers of the K1killer virus. All eight strains (PLY232–PLY246) were killedwhen co-cultured in the presence of the K1 killer strain X3(Vondrejs et al., 1983), and each strain failed to inhibit thegrowth of a K1-supersensitive tester strain S6/1 (Woodsand Bevan, 1968). Thus, the results from two independenttests indicate that the strains that we have constructed arenot carriers of the K1 RNA virus.

Growth characteristics

Potassium dependence. The growth characteristics ofthe strains were examined on SDAP agar plates (pH 5.8)supplemented with 100, 10, 1 and 0.1 mM KCl (Fig. 1).All strains grew at similar rates on media containing100 mM KCl (Fig. 1A). Compared with the growth of thewild-type strain (dilution series 1), the trk1D mutant (dilu-tion series 2) grew at slightly reduced rates on mediumcontaining 10 mM KCl (Fig. 1B, most easily seen by com-paring the size of colonies forming in the most dilute,10-3, cell suspensions). The trk1D mutant grew noticeablypoorer on media containing 1 mM external potassium(Fig. 1C, dilution series 2), and the growth of this strainwas significantly impaired on media containing 0.1 mMpotassium (Fig. 1D, dilution series 2). In contrast, thetrk2D null mutant grew well even on media containing0.1 mM potassium (Fig. 1D, dilution series 3). The trk1Dtrk2D double mutant (dilution series 5) grew as well aswild type on media supplemented with 100 mM KCl(Fig. 1A); however, it grew at significantly reduced rateson media with 10 mM potassium (Fig. 1B) and was unableto form colonies on plates containing £1 mM potassium(Fig. 1C and D). These results indicate that the growthexhibited by the trk2D mutant and the trk1D mutant onplates containing 10 mM potassium was dependent uponthe presence of the other transporter, Trk1p or Trk2prespectively. The presence of the tok1D null allele, eitheralone or in combination with trk1D or trk2D mutations, didnot affect the growth of strains (compare dilution series 1with 4; 2 with 6; 3 with 7; and 5 with 8 respectively).

Monovalent cation sensitivity. The non-specific uptake ofpotassium into cells is impaired in the presence of highextracellular concentrations of competing monovalent cat-ions. Thus, when grown on media containing high concen-trations of monovalent cations, cells are expected toexhibit a greater dependence on specific potassium

uptake mechanisms. We assessed the growth character-istics of the strains on SDAP (pH 5.8, supplemented with≥50 mM K+) containing either 100 mM Li+ or 1 M Na+

(Fig. 2). Under these conditions, wild-type cells (dilutionseries 1) grew well. In the presence of 100 mM Li+, allstrains carrying the trk1D null allele (Fig. 2B, dilutionseries 2, 5, 6 and 8) exhibited uniform and equally poorgrowth. Null mutations in TRK2 or TOK1 did not result indetectable growth phenotypes (Fig. 2B, dilution series 3,4 and 7), and no additive effects were observed in strains

Fig. 1. Potassium-related growth characteristics of wild-type and isogenic yeast strains carrying all possible combinations of trk1D, trk2D and tok1D null alleles. Serial dilutions of strains with the indi-cated genotypes were grown on SDAP agar plates (pH 5.8), supple-mented with KCl at the indicated concentrations. The plates were incubated for 3 days at 30∞C, and digital images were obtained by scanning. Dilution series 1 to 8 correspond to strains PLY232 (wild type, WT), PLY234 (trk1D), PLY236 (trk2D), PLY238 (tok1D), PLY240 (trk1D trk2D), PLY242 (trk1D tok1D), PLY244 (trk2D tok1D) and PLY246 (trk1D trk2D tok1D) respectively.

770 A. Bertl et al.

© 2003 Blackwell Publishing Ltd, Molecular Microbiology, 47, 767–780

carrying trk1D in combination with trk2 or tok1 null alleles(Fig. 2B, compare dilution series 5, 6 and 8 with 2). Theseresults indicate that wild-type cells grown in the presenceof 100 mM Li+ rely exclusively on Trk1p for the specificuptake of K+ to support optimal growth.

Qualitatively, the pattern of growth observed in the pres-ence of 100 mM Li+ was reiterated on media containing1 M Na+ (Fig. 2C), and all strains carrying the trk1D muta-tion grew at reduced rates (dilution series 2, 5, 6 and 8).However, on this media, strains carrying both trk1D andtrk2D alleles exhibited increased sensitivity to sodium(Fig. 2C, compare dilution series 5 and 8 with 2). Theseresults indicate that both Trk1p and Trk2p contribute inde-pendently to K+ uptake in the presence of high extracellu-lar sodium concentrations. The deletion of TOK1 had nodetectable effect on yeast growth in the presence of highconcentrations of sodium.

pH dependence. It has been shown recently that themajor non-specific potassium uptake system, known as

NSC1, is inhibited at low pH (Bihler et al., 2002). Conse-quently, when grown on media at low pH, cells areexpected to exhibit a greater dependence on specificpotassium uptake mechanisms. We assessed the growthcharacteristics of our strains on SDAP containing 100 mMK+ at various pHs. At pH 5.8, all eight strains grew atindistinguishable rates (Fig. 3A). However, when the pHof the growth medium was lowered to 3.1, the strainscarrying trk1D null alleles exhibited less robust growth(Fig. 3, compare dilution series 2 with 1, and dilutionseries 6 with 4), and strains lacking both TRK1 and TRK2did not grow (Fig. 3B, dilution series 5 and 8). The growthrates of strains carrying trk2D null alleles, alone or incombination with tok1D, could not be distinguished fromstrains with an intact TRK2 locus (Fig. 3, compare dilutionseries 3 with 1, and dilution series 7 with 4). However, thefinding that both single trk1D (Fig. 3, dilution series 2) andtrk2D (Fig. 3, dilution series 3) mutant strains are able togrow at pH 3.1 indicates that either Trk1p or Trk2p isindependently able to catalyse sufficient specific K+

Fig. 2. Growth of the yeast strains in the pres-ence of Li+ or Na+. Dilution series were pre-pared from wild-type and isogenic yeast strains carrying all possible combinations of trk1D, trk2D and tok1D null alleles and spotted on solid SDAP, pH 5.8, containing 100 mM KCl (A and B) or 50 mM KCl (C). The agar plates were supplemented with 100 mM LiCl (B) or 1 M NaCl (C) as indicated. The plates were incu-bated for 3 days at 30∞C, and digital images were obtained by scanning. Dilution series 1 to 8 correspond to strains PLY232 (wild type, WT), PLY234 (trk1D), PLY236 (trk2D), PLY238 (tok1D), PLY240 (trk1D trk2D), PLY242 (trk1D tok1D), PLY244 (trk2D tok1D) and PLY246 (trk1D trk2D tok1D) respectively.

trk1, trk2 and tok1 null mutant yeast 771

© 2003 Blackwell Publishing Ltd, Molecular Microbiology, 47, 767–780

uptake to support growth. The deletion of TOK1 had nodetectable effect on pH-dependent growth (Fig. 3, com-pare dilution series 4 with 1, dilution series 6 with 2,dilution series 7 with 3, and dilution series 8 with 5).

Rubidium influx and potassium efflux

The rate of rubidium influx into wild-type and isogenicyeast strains carrying all possible combinations of trk1D,trk2D and tok1D null alleles was determined. Potassium-depleted cells, containing a residual concentration of K+

corresponding to 220 ± 19 nmol K+ mg-1 dry weight of

cells, were incubated in the presence of 50 mM RbCl. Thekinetic parameters derived from a curve-fitting analysis ofthe rates of Rb+ uptake are presented in Table 1. Com-pared with the wild-type strain (PLY232), the trk1D strain(PLY234) exhibited significantly reduced Rb+ uptake(Table 1, compare the Vmax values, row 2 with 1). Theresidual rate of Rb+ uptake observed in the trk1D strainexhibited biphasic kinetics, a finding that suggests thatthese cells express two additional uptake systems: ahigh-affinity low-capacity system (Km = 0.22 mM, Vmax =5 nmol min-1 mg-1); and a low-affinity moderate-capacitysystem (Km = 30 mM, Vmax = 10 nmol min-1 mg-1). The Rb+

Fig. 3. trk1D trk2D double mutants exhibit pH-dependent growth. Wild-type and isogenic yeast strains carrying all possible combinations of trk1D, trk2D and tok1D null alleles were grown on SDAP (supplemented with 100 mM KCl) agar plates at pH 5.8 and adjusted to pH 3.1 as indicated. The plates were incubated for 2 days at 30∞C, and digital images were obtained by scanning. Dilution series 1 to 8 correspond to strains PLY232 (wild type, WT), PLY234 (trk1D), PLY236 (trk2D), PLY238 (tok1D), PLY240 (trk1D trk2D), PLY242 (trk1D tok1D), PLY244 (trk2D tok1D) and PLY246 (trk1D trk2D tok1D, respectively.

Table 1. Kinetics of Rb+ uptake into K+ depleted cells.

Genotype (strain) Vmaxa Km (mM) Repetitions

1. WT (PLY232) 30 ± 2.5 0.20 ± 0.02 32. trk1D (PLY234)b 5 ± 0.5/10 ± 1 0.22 ± 0.02/30 ± 2.5 33. trk2D (PLY236) 30 ± 2.3 0.20 ± 0.02 34. tok1D (PLY238) 30 ± 2.8 0.20 ± 0.02 55. trk1D trk2D (PLY240) 10 ± 0.8 35 ± 3 36. trk1D tok1D (PLY242)b 1.5 ± 0.5/10 ± 0.65 0.22 ± 0.02/30 ± 2 57. trk2D tok1D (PLY244) 30 ± 2.2 0.20 ± 0.02 58. trk1D trk2D tok1D (PLY246) 10 ± 0.55 35 ± 3.2 5

a. Vmax is expressed as nmol min-1 mg-1 dry weight cells.b. Rb+ transport showed a biphasic pattern; the kinetic parameters calculated for both phases are listed

772 A. Bertl et al.

© 2003 Blackwell Publishing Ltd, Molecular Microbiology, 47, 767–780

uptake into trk1D trk2D (PLY240) cells exhibited monopha-sic uptake kinetics, and only the low-affinity moderate-capacity system remained. Together, these resultssuggest that the residual high-affinity system observed intrk1D cells results from the action of Trk2p. The kinetics ofRb+ uptake into the trk2D null mutant (PLY236) strain wereindistinguishable from those observed for the wild-type(PLY232) strain (Table 1, compare rows 3 with 1). Theseresults are consistent with previous reports (Gaber, 1992;Ramos et al., 1994; Madrid et al., 1998). The presence ofthe tok1D null allele, either alone or in combination withtrk1D or trk2D mutations, did not affect Rb+ uptake(Table 1, compare rows 4 with 1; 6 with 2; 7 with 3; and8 with 5 respectively). In additional experiments, the rateof K+ uptake was determined directly using a potassium-selective electrode (data not shown). The results fromthese experiments confirmed that disruption of TRK1 hadthe most significant effect on K+ uptake.

The rate of rubidium-induced potassium efflux fromeach strain was measured (Fig. 4). The potassium contentof cells grown in the presence of high potassium (100 mM)was determined and, regardless of their genotype, cellscontained similar amounts of potassium (380 ± 25 nmolK+ mg-1 dry weight of cells). Potassium efflux was inducedby the suspension of cells in buffer containing 50 mMRbCl. All mutant cells carrying an intact copy of TRK1exhibited identical rates of potassium efflux that wereindistinguishable from the efflux rate observed from wild-type (WT) cells (Fig. 4; for purposes of clarity, we haveonly included the data for the single mutant strains). TheTRK1-carrying cells lost 50% of their intracellular potas-sium about 50 min after the addition of Rb+. In contrast,mutants carrying the trk1D allele exhibited significantlyslower rates of potassium efflux; trk1D null mutantsrequired about 70 min to efflux 50% of their intracellularpotassium. No additive effects were observed in eithertrk1Dtrk2D or trk1Dtok1D double mutants (data notshown).

Electrophysiological characterization

The electrophysiological attributes of wild-type andisogenic yeast cells carrying all possible combinations oftrk1D, trk2D and tok1D null alleles were analysed usingthe patch-clamp technique in the whole-cell configuration.In the presence of millimolar external divalent cations,wild-type cells displayed two characteristic currents: alarge and slowly activating outward current; and a smallinward current that appeared almost instantaneously inresponse to large negatively directed voltage steps(Fig. 5A). The slowly activating outward current wasobserved in all strains carrying an intact TOK1 gene(Fig. 5A–C and E) and was absent in strains carrying thetok1D null allele (Fig. 5D and F–H). The large and slowly

activating outward current has previously been ascribedto the activity of Tok1p (Zhou et al., 1995; Reid et al.,1996; Bertl et al., 1998a).

Fast activating inward currents, observed upon extra-cellular acidification, were recorded in all cells carrying anintact TRK2 gene (Fig. 5A, B, D and F); cells carrying thetrk2D null allele did not display significant acid-activatedinward currents (Fig. 5C, E, G and H). These results areconsistent with earlier observations (Bihler et al., 1999).In each case, the Trk2p-related inward currents were sen-sitive to external pH and insensitive to external potassium,indicating that this inward current is mediated by H+ ratherthan by K+ (data not shown). The Trk2p-related acid-activated inward currents in these yeast strains are muchsmaller compared with the Trk2p-associated currentsreported using cells derived from other non-isogenic strainbackgrounds (Bertl et al., 1995; Bihler et al., 1998). Theactivity of the TRK1 gene product was not detected underour experimental conditions.

Fig. 4. Potassium efflux from wild-type (PLY232, open diamonds) and isogenic yeast strains carrying either trk1D (PLY234, solid trian-gles) trk2D (PLY236, open circles) and tok1D (PLY238, solid squares) null alleles. Cells were grown and prepared as described in Experi-mental procedures. Potassium efflux was initiated by the addition of 50 mM RbCl, and the intracellular potassium content was monitored using atomic absorption spectrophotometry.

trk1, trk2 and tok1 null mutant yeast 773

© 2003 Blackwell Publishing Ltd, Molecular Microbiology, 47, 767–780

In experiments in which divalent cations were excludedfrom the external medium, large and slowly activatinginward currents were observed in all strains. The tracesfrom recordings obtained from the trk1Dtrk2Dtok1D(PLY246) strain are shown in Fig. 5I. The inward currentsdisplayed under these conditions exhibited all the charac-teristics of the non-specific cation channel (NSC1) (Bihleret al., 1998; Roberts et al., 1999), confirming that theactivity of NSC1 is completely independent of the specificK+ transport systems.

TOK1-dependent growth – Tok1p facilitates Cs+ import

Cs+ effectively inhibits the growth of yeast in a pH-dependent manner (Rodríguez-Navarro, 2000). Cells areexquisitely sensitive to Cs+ when grown under conditionsin which the non-specific cation transport system (NSC1)is active, i.e. pH ≥ 5.5 and in the absence of Ca2+. We

examined the growth characteristics of wild-type andisogenic yeast strains carrying all possible combinationsof trk1D, trk2D and tok1D null alleles on SDAP (pH 4.1,supplemented with 100 mM KCl) in the presence of100 mM Cs+ (Fig. 6). The low pH was chosen to inhibit thenon-specific cation conductance. Under these conditions,the growth of the wild-type strain was clearly inhibited(Fig. 6B, dilution series 1). Note that papillation wasobserved (most easily seen at the 10-1 dilution) in strainscarrying an intact copy of TRK1. Papillation should not beconfused with bona fide growth and is probably the resultof heterogeneous levels of TRK1 expression. The strainscarrying the trk1D and trk2D alleles, alone or in combina-tion, were also unable to grow (Fig. 6B, dilution series 2,3 and 5). These results suggest that Trk1p and Trk2p donot transport Cs+. In contrast, the strain carrying the tok1Dnull allele exhibited robust growth (Fig. 6B, dilution series4). The results indicate that, under these experimental

Fig. 5. Current recordings from wild-type and isogenic yeast cells carrying all possible combinations of trk1D, trk2D and tok1D null alleles using the patch-clamp technique in the whole-cell configuration. Whole-cell currents in response to 2.5 s voltage pulses ranging from 100 mV to -200 mV (left) and the corresponding I-V plots (right) are displayed from cells of the following genotypes: (A) wild type (PLY232); (B) trk1D (PLY234); (C) trk2D (PLY236); (D) tok1D (PLY238); (E) trk1D trk2D (PLY240); (F) trk1D tok1D (PLY242); (G) trk2D tok1D (PLY244); (H and I) both trk1D trk2D tok1D (PLY246). The 100 pA scale bar shown below (D) applies to all traces in (A–H); the 400 pA scale bar applies only to traces in (I). To show more clearly the acid-activated inward currents exhibited by strains carrying an intact TRK2 gene (A, B, D and F), the I-V traces at pH 5 (black traces) are superimposed over the corresponding I-V traces at pH 7.5 (grey traces). Internal and external solutions were as described in Experimental procedures, except in (I), where the external medium contained no divalent cations, and the pH was adjusted to pH 7.5. Note, as described in Experimental procedures, the left part of (I) is a steady-state I-V plot. Recordings were obtained at 10 kHz, filtered at 280 Hz and data sampled at 1 kHz.

774 A. Bertl et al.

© 2003 Blackwell Publishing Ltd, Molecular Microbiology, 47, 767–780

growth conditions, Tok1p facilitates the transport of toxicquantities of Cs+ across the plasma membrane. Therobust growth of cells lacking Tok1p function was depen-dent on the presence of TRK1 (Fig. 6B, compare dilutionseries 4 with 6). The absence of TRK2 did not affect theCs+-related growth phenotypes (Fig. 6B, compare dilutionseries 4 with 7, and 6 with 8). Thus, in the presence of100 mM Cs+, cells lacking the capacity specifically to takeup K+ via Trk1p do not obtain sufficient K+ to supportgrowth.

Discussion

We have constructed a set of isogenic yeast strains car-rying all combinations of trk1, trk2 and tok1 null mutationsand have conducted a detailed analysis of the potassiumtransport properties of each strain. In the process of con-structing these strains, strain PLAS132-6C (trk1D tok1D)was found to carry a suppressing mutation closely linkedto the intact TRK2 locus. This suppressing mutation sig-nificantly improves the growth of strains carrying the trk1Dnull allele on YPD media. Although we do not know theprecise nature of the suppressing mutation, it is likely thatthis mutation increases the level of Trk2p expression; suchmutations within the TRK2 gene have been describedpreviously (Vidal et al., 1990). It is known that mutationsin a large number of genes can suppress the potassium

uptake defects of strains lacking specific uptake systems.For example, suppressing mutations have been mappedto genes encoding a broad spectrum of metabolite trans-porters (Ko et al., 1993; Wright et al., 1996; 1997; Lianget al., 1998) and to genes affecting histone acetylation(Vidal et al., 1990). These findings are consistent with ourexperience that, unless proper precautions are taken toreduce selection pressure (see Experimental proce-dures), potassium transport-defective laboratory strainsoften pick up suppressing mutations that obfuscateprecise phenotypic analysis. The use of genetic crossescombined with tetrad analysis provides a very sensitivemethod of detecting the presence of suppressingmutations.

Our phenotypic analyses confirm that TRK1 and TRK2encode proteins that are, or comprise vital componentsof, transport systems that specifically import potassium.These physiologically important transport systems exhibita high degree of functional redundancy. Under a widevariety of growth conditions, single trk1D or trk2D mutantsare able to grow, whereas strains carrying both trk1D trk2Dnull mutations do not. We have observed that Trk1p andTrk2p can functionally substitute for one another whencells are grown in the presence of potassium concentra-tions ≥1 mM (Fig. 1), on media additionally supplementedwith 1 M NaCl (Fig. 2C) and when the pH of the media islow (Fig. 3C). However, at K+ concentrations £1 mM, it

Fig. 6. tok1D mutant strains grow on media in the presence of 100 mM CsCl. Dilution series were prepared from wild-type and isogenic yeast strains carrying all possible combinations of trk1D, trk2D and tok1D null alleles and spot-ted on solid SDAP (pH 4.1, containing 100 mM KCl) (A) and the same medium supplemented with 100 mM CsCl (B). The plates were incu-bated for 4 days at 30∞C, and digital images were obtained by scanning. Dilution series 1 to 8 correspond to strains PLY232 (wild type, WT), PLY234 (trk1D), PLY236 (trk2D), PLY238 (tok1D), PLY240 (trk1D trk2D), PLY242 (trk1D tok1D), PLY244 (trk2D tok1D) and PLY246 (trk1D trk2D tok1D) respectively. Note that dilu-tion series 1 and 3 exhibit a high degree of papillation on media containing 100 mM CsCl (observed most easily at the 10-1 dilution).

trk1, trk2 and tok1 null mutant yeast 775

© 2003 Blackwell Publishing Ltd, Molecular Microbiology, 47, 767–780

becomes obvious that Trk1p is the more effective trans-porter. Under these conditions, trk1D strains exhibitseverely reduced growth (Fig. 1D). Thus, at low externalK+ concentrations, the K+ import mediated by Trk2p occursat rates too low to maintain robust growth. The observedpattern of growth is consistent with the characteristics ofRb+ transport (Table 1). Trk1p and Trk2p exhibit nearlyidentical affinities for Rb+ (Km ª 0.2 mM); however, Trk1pclearly possesses a higher capacity to transport Rb+

(K+).We have found three additional ways to distinguish the

activities of Trk1p and Trk2p experimentally. First, in con-trast to the trk2D mutant, the trk1D mutant exhibitedreduced rates of growth in the presence of 100 mM Li+

(Fig. 2B), and no additive effects were observed in strainscarrying the trk1D allele in combination with trk2D and/ortok1D null mutations (Fig. 2B). Thus, Trk1p, but not Trk2p,maintains the ability to transport K+ under these condi-tions. The inability of Trk2p to function in the presence of100 mM Li+ is probably the consequence of competitionby Li+ for transport or, alternatively, Li+ may directly inhibitthe activity of Trk2p. Secondly, the presence of Trk1p,directly or indirectly, increases the rate at which cells effluxK+ when challenged with high concentrations of Rb+

(Fig. 4). The mechanistic basis of this efflux is presentlyunclear. Finally, the activity of Trk2p, but not that of Trk1p,can be observed readily in whole-cell patch-clamp exper-iments (Fig. 5) (Bihler et al., 1999).

In cells lacking intact copies of TRK1 and TRK2, potas-sium uptake occurs via low-affinity non-selective uptakemechanisms. This non-specific uptake of K+ enables yeastcells to grow, albeit suboptimally, when the concentrationof extracellular K+ is ≥10 mM (Fig. 1B, dilution series 5),and is entirely independent of the presence of Tok1p(Fig. 1B, dilution series 8). We have found that, underconditions that decrease the efficacy of the major non-specific uptake system NSC1 (Bihler et al., 1998), yeastcells become more dependent upon Trk1p and Trk2p forgrowth. Yeast cells normally tolerate environments withhigh concentrations of sodium and exhibit optimal growthin acidic growth conditions. However, at 1 M externalsodium, strains lacking both TRK1 and TRK2 grow verypoorly, even when external concentrations of K+ are high(Fig. 2C). The reason for this growth inhibition is thatsodium competes with potassium for transport; thus, cellscannot internalize sufficient potassium to maintain essen-tial metabolic processes. Similarly, because of the inacti-vation of NSC1 (Bihler et al., 2002), trk1D trk2D doublenull mutants are hypersensitive to low pH (Fig. 3).

We have established a growth-based assay that moni-tors the in vivo activity of Tok1p (Fig. 6). This assay isbased on growth conditions that effectively inactivateNSC1, and that challenge cells with growth-inhibiting con-centrations of Cs+. For growth under these conditions,

cells must eliminate the influx of Cs+ and simultaneouslymaintain sufficient rates of potassium-specific transport.We have found that tok1D null mutants carrying an intactTRK1 gene have the correct requisites for growth (Fig. 6).The robust growth of these strains can most easily beexplained by the elimination of a Tok1p-mediated pathwayfor Cs+ uptake. Consistent with this interpretation, it hasbeen shown in patch-clamp experiments that the replace-ment of K+ by Cs+ in the external medium does not alterthe steady-state current voltage characteristics of Tok1p(Bertl et al., 1998b). Thus, Cs+ not only mimics K+ asthe gating particle, it appears to translocate across theplasma membrane in a Tok1p-mediated manner. Thepresence of TRK1 enables cells to maintain the specifictransport of potassium.

It is important to note that the isogenic strains con-structed in this study are not carriers of the K1 RNA killervirus and, consequently, do not express or secrete the K1killer toxin. All our strains, including the triple trk1D trk2Dtok1D strain (PLY246), displayed an equal degree of sen-sitivity to K1 toxin as the corresponding wild-type strain(PLY232). These results are not consistent with thehypothesis that the cytotoxic effect of killer toxin is basedon its capacity to increase the open probability of theTok1p channel (Ahmed et al., 1999). Nor are our resultsconsistent with the suggestion that the binding of K1 toan intracellular site of Tok1p, which inhibits channel open-ing, is the mechanism underlying the specific immunity ofcells expressing K1 toxin (Sesti et al., 2001). If either ofthese possibilities were true, strains carrying tok1D nullalleles would be resistant to K1 toxin. Our data fully sup-port those presented by Breinig et al. (2002) and clearlydemonstrate that the Tok1p channels are not importantor essential to the physiologically relevant action of K1toxin.

Finally, inward-rectifying K+ channels from other organ-isms have been cloned based on their ability to comple-ment the potassium-related growth defects of strainscarrying trk1D and trk2D deletions (Anderson et al.,1992; Sentenac et al., 1992). However, many interestinginwardly rectifying K+ channels are inhibited by high con-centrations of external divalent cations. Therefore, to anal-yse heterologously expressed K+ channels in yeast, it isdesirable to reduce concentrations of external divalentcations. However, these conditions favour the activity ofthe non-specific uptake system NSC1 (Bihler et al., 2002).In electrophysiological experiments, the activity of NSC1has the potential of masking inwardly rectifying signalsoriginating from heterologously expressed channels(Fig. 5I). We have found that, under conditions that inac-tivate NSC1, at pHs < 4, yeast cells rely exclusively on thespecific potassium transport catalysed by Trk1p and Trk2p(Fig. 3B) for growth. The low pH has the added benefitthat most inwardly rectifying K+ channels from plants are

776 A. Bertl et al.

© 2003 Blackwell Publishing Ltd, Molecular Microbiology, 47, 767–780

activated by low extracellular pH (Hoth et al., 1997).Therefore, by minimizing external Ca2+ and lowering pHbelow pH 4, the experimental conditions are optimal forthe functional analysis of heterologously expressed cal-cium-sensitive inward-rectifying K+ channels. Owing to theexcellent signal-to-noise ratio, we expect that, under theseconditions, the trk1D trk2D double mutant and the trk1Dtrk2D tok1D triple mutant will be good hosts for the anal-ysis of heterologous ion transporters.

Experimental procedures

Strains, media and microbiological techniques

Yeast strains are listed in Table 2, and oligonucleotides usedto construct deletion cassettes specific for TRK1, TRK2 andTOK1 are listed in Table 3. All K+ transport deletion yeaststrains are derived from JRY379 (Reid et al., 1996). Themating type of JRY379 was switched by transformation withplasmid pGAL-HO (Herskowitz and Jensen, 1991), resulting

Table 2. Saccharomyces cerevisiae strains.

Name Genotype Reference

JRY379 MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 Reid et al. (1996)JRY381 MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 tok1D1::HIS3 Reid et al. (1996)PM1 MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 trk2D50::lox-kanMX-lox This workPM6 MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 trk1D50::lox-kanMX-lox This workPM7 MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 This workPM9 MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 trk1D51 tok1D1::HIS3 This workPM10 MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 trk2D1::lox-kanMX-lox tok1D1::HIS3 This workPM17 MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 trk1D51 This workPLAS132-6C MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 trk1D51 tok1D1::HIS3 This workPLAS133-28D MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 trk1D51 trk2D50::kanMX tok1D1::HIS3 This workPLY232 MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 This workPLY234 MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 trk1D51 This workPLY236 MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 trk2D50::lox-kanMX-lox This workPLY238 MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 tok1D1::HIS3 This workPLY240 MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 trk1D51 trk2D50::lox-kanMX-lox This workPLY242 MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 trk1D51 tok1D1::HIS3 This workPLY244 MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 trk2D50::lox-kanMX-lox tok1D1::HIS3 This workPLY246 MATa his3D200 leu2-3,112 trp1D901 ura3-52 suc2D9 trk1D51 trk2D50::lox-kanMX-lox tok1D1::HIS3 This work

Table 3. List of oligonucleotides.

Oligonucleotide Sequence (5¢-3¢) Length (bp)

-83a -44TRK1-lox-disruption-sense GAAGGAGGGTATTCTATTGGCTCTCAAGGAAGTCATTCCTGCTGAAGCTTCGTACGCTGC 60

+123 +83TRK1-lox-disruption-antisense ACGTAGGCAAATGTATATCACAATGTTACTAATCAAGGTTGCATAGGCCACTAGTGGATCTG 62

-117 -77TRK2-lox-disruption-sense CCCCTCGGCCGTGCGGCTGAAAAAGAGAAATGATATTGGAGCTGAAGCTTCGTACGCTGC 60

+44 +5TRK2-lox-disruption-antisense AATTACGTTGGCTCTTATGTAGGTAAAGAGGGGTAAACTTGCATAGGCCACTAGTGGATCTG 62

110 91Kan-antisenseb CCTCAGTGGCAAATCCTAAC 20

-446 -427TRK1-START CGTTCGGGGCTGACAACGCA 20

+259 +238TRK1-STOP GGGACAATGTACTAATGGCGT 21

-479 -460TRK2-START CCCGTCCATTGAGTGCCCGT 20

+173 +154TRK2-STOP CGGATAGGATTCGTTGTGCT 20

-325 -306TOK1-START CGCATTCGCGTCTCGTTACC 20

+152 +133TOK1-STOP GCGAGACGCAACGGGTGCAT 20

a. Position numbers refer to bases 5¢ of the start codon (–) or 3¢ of the stop codon (+) in each ORF; the sequences corresponding to the pUG6plasmid are underlined.b. Güldener et al. (1996).

trk1, trk2 and tok1 null mutant yeast 777

© 2003 Blackwell Publishing Ltd, Molecular Microbiology, 47, 767–780

in strain PM7. Polymerase chain reaction (PCR)-generatedtrk1D50::lox-kanMX-lox, trk2D50::lox-kanMX-lox andtok1D50::lox-kanMX-lox disruption constructs were pre-pared as described previously (Güldener et al., 1996)using the oligonucleotide pairs, TRK1-lox-disruption-sense/antisense, TRK2-lox-disruption-sense/antisense and TOK1-lox-disruption-sense/antisense respectively (Table 3). Thesedeletion constructs contain ª 40 bp of homology to theupstream 5¢ and downstream 3¢ regions of each of therespective genes. Strain JRY379 was separately transformedwith each of the PCR-generated deletion constructs, andtransformants were selected on YPD media containing G418(kanamycin sulphate) as described previously (Güldeneret al., 1996). Genomic DNA was isolated from G418-resistantstrains. To confirm that deletion constructs had integrated intothe correct chromosomal location and for generating probesfor Southern analysis, the oligonucleotides TRK1-START andTRK2-START (sense primers) and the Kan antisense primer,which lies 358 bp within the lox-kanMX-lox cassette in pUG6,were used to prime PCRs. Southern analysis was used toverify that transformants carried single copies of the respec-tive deletion constructs. Strains PM6 and PM1 (Table 2) havethe entire TRK1 and TRK2 coding sequences replaced bythe kanMX gene. Strain PM6 (trk1D50::lox-kanMX-lox) wastransformed with plasmid pSH47 containing the gene for theCre-loxP recombinase under the control of the GAL1 pro-moter (Güldener et al., 1996). Transformants were grownon media containing galactose, and single cell-derivedcolonies that exhibited kanamycin sensitivity were identified.Kanamycin-sensitive strains were propagated on media con-taining 5-fluoroorotic acid (5-FOA) to select for the loss ofplasmid pSH47. The resulting strain, PM17, carries theunmarked trk1D51 deletion allele.

TOK1-START/STOP oligonucleotide pairs (Table 3) wereused to obtain a PCR-generated tok1D1::HIS3 disruptionconstruct with extended regions of homology flanking theTOK1 open reading frame (ORF; 328 bp 5¢ of the initiationcodon and 152 bp 3¢ of the termination codon). The templateDNA was isolated from strain JRY381. The amplified deletionfragment with extended flanking homology facilitated the con-struction of the double K+ transporter mutants PM9 andPM10. Double mutant strains PM9 (trk1D51 tok1D1::HIS3)and PM10 (trk2D50::kanMX tok1D1::HIS3) were obtained bytransforming strains PM17 and PM1 to His+ respectively. Ineach case, the deletions were confirmed by PCR and South-ern analysis.

Strains PM7 (wild type) and PM9 (trk1D51 tok1D1::HIS3)were crossed, the resulting diploid was sporulated, and tet-rads were dissected on YPD media supplemented with50 mM KCl. Spore viability was excellent and, in allinstances, the deletion alleles segregated 2:2. As expected,linkage between trk1D51 and tok1D1::HIS3 was observed;these genes lie ª26 cM apart on chromosome X (46 parentalditype; one non-parental ditype and 48 tetratype). Theobserved linkage provided additional confirmation that thedeletion constructs were integrated at their appropriate chro-mosomal locations. A trk1D51 tok1D1::HIS3 spore-derivedstrain (PLAS132-6C) identified in this analysis was crossedwith PM1 (trk2D50::kanMX), the resulting diploid was sporu-lated and tetrads dissected (YPD + 50 mM KCl). Again, sporeviability was excellent and, in all instances, the three deletion

alleles segregated 2:2. Spore-derived strains carrying allpossible combinations of deletions were recovered at theexpected frequency. Owing to a spontaneously arising sup-pressor mutation linked to the wild-type TRK2 locus (seeResults), a trk1D51 trk2D50::kanMX tok1D1::HIS3 spore-derived triple mutant strain (PLAS133-28D) that lacked thesuppressing mutation was backcrossed to JRY379. Theresulting diploid strain was sporulated, and tetrads were dis-sected (YPD + 50 mM KCl). Strains PLY232–PLY246 aremeiotic segregants from this cross. To reduce the likelihoodof the inadvertent selection of interfering suppressing muta-tions, all strains were maintained as frozen stocks, and work-ing cultures of each strain were revived from frozen stockson YPD supplemented with 50 mM KCl. Phenotypic analy-sis was carried out on strains within 2 weeks after beingresuscitated.

Standard yeast media were prepared and yeast geneticmanipulations were performed as described by Guthrie andFink (1991). Potassium-dependent phenotypes were analy-sed using strains grown in SDAP (synthetic dextrose argininephosphate; Rodríguez-Navarro and Ramos, 1984). Briefly,the media components of SDAP were prepared as follows: a10 mM arginine stock was adjusted with phosphoric acid topH 6.5 and autoclaved. Glucose (40%) and a 100¥ stocksolution of the other ingredients (trace elements, vitamins,salts, histidine, leucine and uracil) were filter sterilized andadded to the arginine/phosphate after cooling. The pH of theprepared media without additional adjustments was 5.8.Where required, media was made solid with the addition of2% purified agar (Sigma, A7921). The pH of agar plates wasmeasured with a surface pH electrode. Note that, for thepreparation of agar plates with pH < 4.5, the pH had to beadjusted after autoclaving by the addition of an appropriateamount of phosphoric acid. Media containing agar will notsolidify if the pH is adjusted to acidic pH (<4.5) beforeautoclaving.

Growth tests on agar plates

For a qualitative assessment of growth phenotypes associ-ated with the deletion of single and multiple K+ transporters,cells were grown for 18 h at 30∞C in SDAP supplementedwith 50 mM KCl. Cells were harvested by centrifugation of1 ml samples at 4000 r.p.m. resuspended in sterile double-distilled water at a cell density corresponding to OD600 =1.0 ± 0.05. Tenfold serial dilutions were prepared, and 7 mlaliquots of each dilution were spotted on SDAP agar platessupplemented as indicated. As prepared SDAP media has apH of 5.8, when growth was examined at lower pHs, themedia was adjusted to the desired pH with phosphoric acid.The plates were incubated at 30∞C, and digital greyscaleimages were obtained at the times indicated using a standardflatbed scanner set to a resolution of 300 d.p.i.

Rubidium influx and potassium efflux

Rubidium influx was measured in cells from exponentiallygrown cultures (OD660 of 0.3–0.4) in SDAP medium supple-mented with 100 mM KCl. Cells were depleted of potassiumby incubating cells for 5 h in potassium-free SDAP

778 A. Bertl et al.

© 2003 Blackwell Publishing Ltd, Molecular Microbiology, 47, 767–780

(Rodríguez-Navarro and Ramos, 1984; Ramos et al., 1985).Starved cells were harvested and resuspended in KTRANSbuffer [10 mM MES (2-(N-morpholino) ethanesulphonic acid)adjusted to pH 5.8 with Ca(OH)2, 0.1 mM MgCl2 and 2%glucose]. After 2 min of incubation in KTRANS buffer, 50 mMRbCl was added, and subsamples were removed at regulartime intervals during the first 10 min. The subsamples wereimmediately filtered, the filters were washed with 20 mMMgCl2, treated with HCl, and the cation concentrations wereanalysed by atomic absorption spectrophotometry (Ramoset al., 1994; Prista et al., 1997; Madrid et al., 1998). The initialrate of Rb+ uptake was determined from the time course ofthe net accumulation.

Potassium efflux was measured in cells grown asdescribed for rubidium influx experiments. Cells were har-vested and resuspended in KTRANS buffer, and potassiumefflux was triggered by the addition of 50 mM RbCl. Subsam-ples were removed at regular time intervals and filteredimmediately. The filters were treated as described for Rb+

influx, and the cation concentrations were analysed by atomicabsorption spectrophotometry (Ramos et al., 1994; Pristaet al., 1997; Madrid et al., 1998).

Protoplast preparation

Procedures for growing yeast cells and isolating protoplastsfor patch-clamp recordings were as described previously(Bertl et al., 1993). Cells were grown overnight at 30∞C in100 ml Erlenmeyer flasks with continuous rotary shaking at250 r.p.m. A 10 ml aliquot of the cell suspension was centri-fuged at 500 g for 5 min. The supernatant was discarded, thepellet was resuspended in 3 ml of a solution containing50 mM potassium phosphate buffer, pH 7.2, and 40 mM b-mercaptoethanol, and the cell suspension was incubated ona rotary shaker for 15 min at room temperature. Three millil-itres of buffer containing 150 mg of bovine serum albumin(BSA) and 0.1–1 mg of zymolyase (varied with the enzymebatch) was added. The mixture was vortexed quickly andincubated at room temperature for 45 min. Protoplasts wereharvested by centrifugation for 3–5 min at 500 g and resus-pended in 10 ml of stabilizing buffer (220 mM KCl, 10 mMCaCl2, 5 mM MgCl2, pH 7.2) supplemented with 1% glucose.

Whole-cell current recording

General procedures for whole-cell current recordings havebeen described by Hamill et al. (1981) and modified protocolsfor use on yeast protoplasts were reported by Bertl et al.(1995; 1998b). For data acquisition and analysis, the PULSE/PULSEFIT software package was used in combination with theEPC9 patch-clamp amplifier (Heka). Data were exported toIGOR for leak correction (using a self-written routine) andconstruction of the current voltage (I-V) plots. All experimentsreported here were performed in the whole-cell recordingconfiguration, which was obtained by combining slight suc-tion and a voltage pulse (0.9 V of 80 ms duration) after atight seal formed. Seal resistances varied between 5 and20 GOhm, and all whole-cell recordings were corrected for alinear leak (50–250 pS). Current recordings emphasizing thekinetic aspects were done in response to 2.5 s voltage steps

progressing from +100 mV to -200 mV in -20 mV incre-ments, with a 1 s holding interval at -40 mV between eachpair. For the steady-state I-V plot (Fig. 5I), a stretch of 100 msof current recording obtained at the end of each test pulsewas averaged and plotted versus the applied voltage. Tocompare I-V plots obtained from one protoplast under differ-ent experimental conditions (pH 7.5 vs. pH 5.5), voltageramps of 5 s duration were applied to clamp the membranefrom +100 mV to -200 mV, and the resulting current wasplotted versus the applied voltage. Standard pipette solutions(cytosolic solution) contained 175 mM KCl, 4 mM MgCl2,4 mM K-ATP, 1 mM EGTA, 0.152 mM CaCl2, adjusted topH 7.0 with KOH. Extracellular (bath) solution contained150 mM KCl, 10 mM CaCl2, 5 mM MgCl2, buffered to pH 5with Tris/MES. Deviations from this composition are given inthe respective figure legends.

Acknowledgements

This work was supported by grant BIO-CT97-2210 from theEuropean Union. The authors thank Blanka Janderová for thekind gift of K1 killer strain (X3) and killer supersensitive strain(S6/1).

References

Ahmed, A., Sesti, F., Ilan, N., Shih, T.M., Sturley, S.L., andGoldstein, S.A. (1999) A molecular target for viral killertoxin: TOK1 potassium channels. Cell 99: 283–291.

Anderson, J.A., Huprikar, S.S., Kochian, L.V., Lucas, W.J.,and Gaber, R.F. (1992) Functional expression of a proba-ble Arabidopsis thaliana potassium channel in Saccharo-myces cerevisiae. Proc Natl Acad Sci USA 89: 3736–3740.

Bañuelos, M.A., and Rodríguez-Navarro, A. (1998) P-typeATPases mediate sodium and potassium effluxes inSchwanniomyces occidentalis. J Biol Chem 273: 1640–1646.

Bañuelos, M.A., Sychrová, H., Bleykasten-Grosshans, C.,Souciet, J.L., and Potier, S. (1998) The Nha1 antiporter ofSaccharomyces cerevisiae mediates sodium and potas-sium efflux. Microbiology 144: 2749–2758.

Bertl, A., Slayman, C.L., and Gradmann, D. (1993) Gatingand conductance in an outward-rectifying K+ channel fromthe plasma membrane of Saccharomyces cerevisiae. JMembr Biol 132: 183–199.

Bertl, A., Anderson, J.A., Slayman, C.L., and Gaber, R.F.(1995) Use of Saccharomyces cerevisiae for patch-clampanalysis of heterologous membrane proteins: characteriza-tion of Kat1, an inward-rectifying K+ channel from Arabi-dopsis thaliana, and comparison with endogenous yeastchannels and carriers. Proc Natl Acad Sci USA 92: 2701–2705.

Bertl, A., Bihler, H., Reid, J.D., Kettner, C., and Slayman, C.L.(1998a) Physiological characterization of the yeast plasmamembrane outward rectifying K+ channel, DUK1 (TOK1),in situ. J Membr Biol 162: 67–80.

Bertl, A., Bihler, H., Kettner, C., and Slayman, C.L. (1998b)Electrophysiology in the eukaryotic model cell Saccharo-myces cerevisiae. Pflügers Arch 436: 999–1013.

trk1, trk2 and tok1 null mutant yeast 779

© 2003 Blackwell Publishing Ltd, Molecular Microbiology, 47, 767–780

Bihler, H., Slayman, C.L., and Bertl, A. (1998) NSC1: a novelhigh-current inward rectifier for cations in the plasma mem-brane of Saccharomyces cerevisiae. FEBS Lett 432: 59–64.

Bihler, H., Gaber, R.F., Slayman, C.L., and Bertl, A. (1999)The presumed potassium carrier Trk2p in Saccharomycescerevisiae determines an H+-dependent, K+-independentcurrent. FEBS Lett 447: 115–120.

Bihler, H., Slayman, C.L., and Bertl, A. (2002) Low-affinitypotassium uptake by Saccharomyces cerevisiae is medi-ated by NSC1, a calcium-blocked non-specific channel.Biochim Biophys Acta 1558: 109–118.

Breinig, F., Tipper, D.J., and Schmitt, M.J. (2002) Kre1p, theplasma membrane receptor for the yeast K1 viral toxin. Cell108: 395–405.

Fairman, C., Zhou, X., and Kung, C. (1999) Potassium uptakethrough the TOK1 K+ channel in the budding yeast. JMembr Biol 168: 149–157.

Fink, G.R., and Styles, C.A. (1972) Curing of a killer factor inSaccharomyces cerevisiae. Proc Natl Acad Sci USA 69:2846–2849.

Gaber, R.F. (1992) Molecular genetics of yeast ion transport.Int Rev Cytol 137: 299–353.

Gaber, R.F., Styles, C.A., and Fink, G.R. (1988) TRK1encodes a plasma membrane protein required for high-affinity potassium transport in Saccharomyces cerevisiae.Mol Cell Biol 8: 2848–2859.

Güldener, U., Heck, S., Fiedler, T., Beinhauer, J., andHegemann, J. (1996) A new efficient gene disruptioncassette for repeated use in budding yeast. Nucleic AcidsRes 24: 2519–2524.

Guthrie, C., and Fink, G.R. (1991) Guide to yeast geneticsand molecular biology. Methods Enzymol 194: 3–37.

Hamill, O.P., Marty, A., Neher, E., Sakmann, B., andSigworth, F.J. (1981) Improved patch-clamp techniques forhigh-resolution current recording from cells and cell-freemembrane patches. Pflügers Arch 391: 85–100.

Herskowitz, I., and Jensen, R.E. (1991) Putting the HO geneto work: practical uses for mating-type switching. MethodsEnzymol 194: 132–146.

Hoth, S., Dreyer, I., Dietrich, P., Becker, D., Müller-Röber, B.,and Hedrich, R. (1997) Molecular basis of plant-specificacid activation of K+ uptake channels. Proc Natl Acad SciUSA 94: 4806–4810.

Ketchum, K.A., Joiner, W.J., Sellers, A.J., Kaczmarek, L.K.,and Goldstein, S.A. (1995) A new family of outwardly rec-tifying potassium channel proteins with two pore domainsin tandem. Nature 376: 690–695.

Ko, C.H., and Gaber, R.F. (1991) TRK1 and TRK2 encodestructurally related K + transporters in Saccharomyces cer-evisiae. Mol Cell Biol 11: 4266–4273.

Ko, C.H., Buckley, A.M., and Gaber, R.F. (1990) TRK2 isrequired for low affinity K + transport in Saccharomycescerevisiae. Genetics 125: 305–312.

Ko, C.H., Liang, H., and Gaber, R.F. (1993) Roles of multipleglucose transporters in Saccharomyces cerevisiae. MolCell Biol 13: 638–648.

Liang, H., Ko, C.H., Herman, T., and Gaber, R.F. (1998)Trinucleotide insertions, deletions, and point mutations inglucose transporters confer K+ uptake in Saccharomycescerevisiae. Mol Cell Biol 18: 926–935.

Loukin, S.H., Vaillant, B., Zhou, X.L., Spalding, E.P., Kung,C., and Saimi, Y. (1997) Random mutagenesis reveals aregion important for gating of the yeast K+ channel Ykc1.EMBO J 16: 4817–4825.

Madrid, R., Gómez, M.J., Ramos, J., and Rodríguez-Navarro,A. (1998) Ectopic potassium uptake in trk1 trk2 mutants ofSaccharomyces cerevisiae correlates with a highly hyper-polarized membrane potential. J Biol Chem 273: 14838–14844.

Peña, A., and Ramírez, J. (1991) An energy-dependent effluxsystem for potassium ions in yeast. Biochim Biophys Acta1068: 237–244.

Prista, C., Almagro, A., Loureiro-Dias, M.C., and Ramos, J.(1997) Physiological basis for the high salt tolerance ofDebaryomyces hansenii. Appl Environ Microbiol 63: 4005–4009.

Ramírez, J., Ramírez, O., Saldana, C., Coria, R., and Peña,A. (1998) A Saccharomyces cerevisiae mutant lacking aK+/H+ exchanger. J Bacteriol 180: 5860–5865.

Ramos, J., Contreras, P., and Rodríguez-Navarro, A. (1985)A potassium transport mutant of Saccharomyces cerevi-siae. Arch Microbiol 143: 88–93.

Ramos, J., Alijo, R., Haro, R., and Rodríguez-Navarro, A.(1994) TRK2 is not a low-affinity potassium transporter inSaccharomyces cerevisiae. J Bacteriol 176: 249–252.

Reid, J., Lukas, W., Shafaatian, R., Bertl, A., Scheurmann-Kettner, C., Guy, R., and North, R. (1996) The S. cerevi-siae outwardly-rectifying potassium channel (DUK1)identifies a new family of channels with duplicated poredomains. Receptors Channels 4: 51–62.

Roberts, S.K., Fischer, M., Dixon, G.K., and Sanders, D.(1999) Divalent cation block of inward currents and low-affinity K+ uptake in Saccharomyces cerevisiae. J Bacteriol181: 291–297.

Rodríguez-Navarro, A. (2000) Potassium transport in fungiand plants. Biochim Biophys Acta 1469: 1–30.

Rodríguez-Navarro, A., and Ramos, J. (1984) Dual systemfor potassium transport in Saccharomyces cerevisiae. JBacteriol 159: 940–945.

Schmitt, M.J., and Eisfeld, K. (1999) Killer viruses in S. cer-evisiae and their general importance in understandingeucaryotic cell biology. Recent Res Dev Virol 1: 525–545.

Sentenac, H., Bonneaud, N., Minet, M., Lacroute, F.,Salmon, J.M., Gaymard, F., and Grignon, C. (1992) Clon-ing and expression in yeast of a plant potassium ion trans-port system. Science 256: 663–665.

Sesti, F., Shih, T.M., Nikolaeva, N., and Goldstein, S.A.(2001) Immunity to K1 killer toxin: internal TOK1 blockade.Cell 105: 637–644.

Vergani, P., Miosga, T., Jarvis, S.M., and Blatt, M.R. (1997)Extracellular K+ and Ba2+ mediate voltage-dependent inac-tivation of the outward-rectifying K+ channel encoded bythe yeast gene TOK1. FEBS Lett 405: 337–344.

Vidal, M., Buckley, A.M., Hilger, F., and Gaber, R.F. (1990)Direct selection for mutants with increased K+ transport inSaccharomyces cerevisiae. Genetics 125: 313–320.

Vidal, M., Buckley, A.M., Yohn, C., Hoeppner, D.J., andGaber, R.F. (1995) Identification of essential nucleotidesin an upstream repressing sequence of Saccharomycescerevisiae by selection for increased expression of TRK2.Proc Natl Acad Sci USA 92: 2370–2374.

780 A. Bertl et al.

© 2003 Blackwell Publishing Ltd, Molecular Microbiology, 47, 767–780

Vondrejs, V., Psenicka, I., Kupková, L., Dostalová, R.,Janderová, B., and Bendová, O. (1983) The use of killerfactor in the selection of hybrid strains. Folia Biol 29: 372–384.

Vondrejs, V., Janderová, B., and Valasek, L. (1996) Yeastkiller toxin K1 and its exploitation in genetic manipulations.Folia Microbiol 41: 379–394.

Wickner, R.B. (1996) Double-stranded RNA viruses ofSaccharomyces cerevisiae. Microbiol Rev 60: 250–265.

Woods, D.R., and Bevan, E.A. (1968) Studies on the natureof the killer factor produced by Saccharomyces cerevisiae.J Gen Microbiol 51: 115–126.

Wright, M.B., Howell, E.A., and Gaber, R.F. (1996) Aminoacid substitutions in membrane-spanning domains of Hol1,a member of the major facilitator superfamily of transport-ers, confer nonselective cation uptake in Saccharomycescerevisiae. J Bacteriol 178: 7197–7205.

Wright, M.B., Ramos, J., Gómez, M.J., Moulder, K., Scherrer,M., Munson, G., and Gaber, R.F. (1997) Potassium trans-port by amino acid permeases in Saccharomyces cerevi-siae. J Biol Chem 272: 13647–13652.

Zhou, X.L., Vaillant, B., Loukin, S.H., Kung, C., and Saimi, Y.(1995) YKC1 encodes the depolarization-activated K+

channel in the plasma membrane of yeast. FEBS Lett 373:170–176.