A new pathway for vacuolar cadmium sequestration in Saccharomyces cerevisiae: YCF1-catalyzed transport of bis(glutathionato)cadmium

Abstract

The yeast cadmium factor (YCF1) gene encodes an MgATP-energized glutathione S-conjugate transporter responsible for the vacuolar sequestration of organic compounds after their S-conjugation with glutathione. However, while YCF1 was originally isolated according to its ability to confer resistance to cadmium salts, neither its mode of interaction with Cd²⁺ nor the relationship between this process and organic glutathione-conjugate transport are known. Here we show through direct comparisons between vacuolar membrane vesicles purified from Saccharomyces cerevisiae strain DTY167, harboring a deletion of the YCF1 gene, and the isogenic wild-type strain DTY165 that YCF1 mediates the MgATP-energized vacuolar accumulation of Cd·glutathione complexes. The substrate requirements, kinetics and Cd²⁺/glutathione stoichiometry of cadmium uptake and the molecular weight of the transport-active complex demonstrate that YCF1 selectively catalyzes the transport of bis(glutathionato)cadmium (Cd·GS₂). On the basis of these results—the Cd²⁺ hypersensitivity of DTY167, versus DTY165, cells, the inducibility of YCF1-mediated transport, and the rapidity and spontaneity of Cd·GS₂ formation—this new pathway is concluded to contribute substantially to Cd²⁺ detoxification.

A new class of ATP-binding cassette (ABC) transporter responsible for MgATP-energized transport of organic compounds after their conjugation with glutathione (GSH) has recently been discovered. Formerly designated the GS-X pump (1), this transporter, or family of transporters, has been implicated in the extrusion of a broad range of S-conjugated compounds from the cytosol.

To date, two closely related GS-X pumps have been identified molecularly. These are the human multidrug resistance-associated protein (MRP1) (2, 3) and the yeast cadmium factor (YCF1) protein (4, 5). MRP1 and YCF1 are 43% identical (63% similar) at the amino acid level, possess nucleotide binding folds with an equivalent spacing of conserved residues, and contain two subclass-specific structures, a central truncated cystic fibrosis transmembrane conductance regulator-like “regulatory” domain, rich in charged amino acids, and an ≈200-amino acid residue N-terminal extension (2, 4). MRP1 catalyzes the MgATP-energized transport of leukotriene C₄ and related GSH S-conjugates (GS-conjugates) across the plasma membrane of mammalian cells (3, 6, 7). YCF1 catalyzes the transport of organic GS-conjugates into the vacuole of Saccharomyces cerevisiae (5).

Given the participation of both of these integral membrane proteins in the transport of organic GS-conjugates and their implied role in the elimination and/or sequestration of cytotoxic drugs, it is intriguing that the YCF1 gene was initially identified by screening a yeast genomic library for the ability of multicopy DNA fragments to confer resistance to cadmium salts in the growth medium (4). The question of how the vacuolar sequestration of organic GS-conjugates by YCF1 is related to Cd²⁺ resistance therefore arises. Specifically, is the detoxification of Cd²⁺ by YCF1 dependent on its interaction with GSH, as is the case for organic xenobiotics (5) and, if so, how does GSH exert its effects?

In this communication we address these questions to show that YCF1 is not only competent in the MgATP-energized transport of organic GS-conjugates but also Cd²⁺ after its complexation with GSH. Our findings demonstrate a new pathway for the vacuolar sequestration of Cd²⁺ in S. cerevisiae: YCF1-mediated transport of bis(glutathionato)cadmium (Cd·GS₂).

MATERIALS AND METHODS

Yeast Strains.

The two strains of S. cerevisiae used in these studies—DTY165 (MATα ura3-52 his6 leu2-3,-112 his3-Δ200 trp1-901 lys2-801 suc2-Δ) and the isogenic ycf1Δ mutant strain, DTY167 (MATα ura3-52 his6 leu2-3,-112 his3-Δ200 trp1-901 lys2-801 suc2-Δ, ycf1::hisG)—were manipulated as described (5, 8).

Isolation of Vacuolar Membrane Vesicles.

Vacuolar membrane vesicles were prepared as described (5), except that the dithiothreitol (1 mM) and EGTA (1 mM) present in the standard membrane isolation medium (5, 9) were removed to prevent the attenuation of YCF1-dependent Cd²⁺ transport otherwise exerted by these compounds (see Discussion). Vesiculated vacuolar membranes were subjected to three cycles of 50-fold dilution into simplified suspension medium (1.1 M glycerol/5 mM Tris·Mes, pH 8.0), centrifugation at 100,000 × g for 35 min, and resuspension in the same medium before use.

Purification of Cadmium·Glutathione Complexes.

Singly radiolabeled ¹⁰⁹Cd·GS_n and doubly radiolabeled ¹⁰⁹Cd·[³H]GS_n complexes were prepared by sequential gel-filtration and anion-exchange chromatography of the reaction products generated by incubating 20 mM ¹⁰⁹CdSO₄ (78.4 mCi/mmol; 1 Ci = 37 GBq) with 40 mM GSH or 40 mM [³H]GSH (240 mCi/mmol) in 15 ml 10 mM phosphate buffer (pH 8.0) containing 150 mM KNO₃ at 45°C for 24 h. For gel-filtration, 2 ml aliquots of the reaction mixture were applied to a column (40 × 1.5 cm interior diameter) packed with water-equilibrated Sephadex G-15, eluted with deionized water, and ¹⁰⁹Cd and/or ³H in the fractions was measured by liquid scintillation counting. The fractions encompassed by each of the two ¹⁰⁹Cd·GS_n peaks identified were pooled, lyophilized, and redissolved in 4 ml of loading buffer (5 mM Tris·Mes, pH 8.0). For anion-exchange chromatography, 0.5 ml aliquots of the resuspended lyophilizates from gel-filtration chromatography were applied to a Mono-Q HR5/5 column (Pharmacia) equilibrated with the same buffer. Elution was with a linear gradient of NaCl (0.5 ml/min; 0–500 mM) dissolved in loading buffer. The individual fractions corresponding to the major peaks of ¹⁰⁹Cd obtained from the Mono-Q column (one each for the peaks resolved by gel-filtration chromatography) were pooled, lyophilized, and resuspended in 4 ml deionized water after liquid scintillation counting. Buffer salts were removed before transport measurements or mass spectrometry by passing the samples down a column (120 × 1.0 cm interior diameter) packed with water-equilibrated Sephadex G-15.

Measurement of ¹⁰⁹Cd²⁺ Uptake.

MgATP-energized, uncoupler-insensitive ¹⁰⁹Cd²⁺ uptake by vacuolar membrane vesicles was measured at 25°C in 200 μl reaction volumes containing 3 mM ATP, 3 mM MgSO₄, 5 μM gramicidin-D, 10 mM creatine phosphate, 16 units/ml creatine phosphate kinase, 50 mM KCl, 400 mM sorbitol, and 25 mM Tris·Mes (pH 8.0), and the indicated concentrations of ¹⁰⁹CdSO₄, GSH, or ¹⁰⁹Cd- and/or ³H-labeled purified Cd·GS_n complexes as described (5) except that the wash media contained 100 μM CdSO₄ in addition to sorbitol (400 mM) and Tris·Mes (3 mM, pH 8.0).

Pretreatment of DTY165 Cells with Cd²⁺ or 1-Chloro-2,4-Dinitrobenzene (CDNB).

For studies of the inducibility of YCF1 expression and YCF1-dependent transport, DTY165 cells were grown in yeast extract/peptone/dextrose (YPD) medium (8) for 24 h at 30°C to an OD₆₀₀ of 1.0–1.2, pelleted by centrifugation and resuspended in fresh YPD medium containing CdSO₄ (200 μM) or CDNB (150 μM). After washing in distilled water, total RNA was extracted and vacuolar membrane vesicles were prepared from the pretreated cells. Control RNA and membrane samples were prepared from DTY165 cells treated in an identical manner except that CdSO₄ and CDNB were omitted from the second incubation cycle.

RNase Protection Assays.

Cd²⁺ and CDNB-elicited increases in YCF1 mRNA levels were assayed by RNase protection using 18S rRNA as an internal control. YCF1-specific probe was generated by PCR amplification of the full-length YCF1::HA gene, encoding human influenza hemagglutinin 12CA5 (HA) epitope-tagged YCF1, using plasmid pYCF1-HA (5) as template. The forward, YCF1-specific, primer and backward primer, containing the HA-tag coding sequence, had the sequences 5′-AAACTCGAGATGGCTGGTAATCTTGTTTC-3′ and 5′-GCCTCTAGATCAAGCGTAGTCTGGGACGTCGTATGGGTAATTTTCATTGA-3′, respectively. 18S rRNA-specific probe was synthesized by PCR of S. cerevisiae genomic DNA using sense and antisense primers with the sequences 5′-AGATTAAGCCATGCATGTCT-3′ and 5′-TGCTGGTACCAGACTTGCCCTCC-3′, respectively. Both PCR products were individually subcloned into pCRII vector (Invitrogen) to generate plasmids pCR-YCF1 and pCR-Y18S. After linearization of pCR-YCF1 and pCR-Y18S with AflI and NcoI, a 320-nucleotide YCF1-specific RNA probe and 220-nucleotide 18S rRNA-specific probe were synthesized using T7 RNA polymerase and SP6 RNA polymerase, respectively. Aliquots of total RNA, prepared as described (10), from control, CdSO₄-, or CDNB-pretreated DTY165 cells were hybridized with a mixture of ³²P-labeled YCF1 antisense probe (1 × 10⁶ cpm) and 18S rRNA antisense probe (5 × 10² cpm), and RNase protection (11) was assayed using an RPAII kit (Ambion).

Matrix-Assisted Laser Desorption Mass Spectrometry (MALD-MS).

The ¹⁰⁹Cd·GS_n complexes purified by gel-filtration and anion-exchange chromatography were adjusted to a final concentration of 2–5 mM (as Cd) with deionized water, mixed with an equal volume of sinapinic acid (10 mg/ml) dissolved in acetonitrile/H₂O/trifluoroacetic acid (70:30:0.1%, vol/vol) and applied to the ion source of a PerSeptive Biosystems (Cambridge, MA) Voyager RP Biospectrometry Workstation. The instrument, which was equipped with a 1.3-m flight tube and variable two-stage ion source set at 30 kV, was operated in linear mode. Mass/charge (m/z) ratio was measured by time-of-flight after calibration with external standards.

Protein Assays.

Protein was estimated by a modification of the method of Peterson (12).

Chemicals.

[³H]GSH [(glycine-2-³H)-l-Glu-Cys-Gly; 4.4 Ci/mmol] was from DuPont/NEN, and ¹⁰⁹CdSO₄ (78.44 Ci/mmol) was from Amersham. All other reagents were of analytical grade and purchased from Fisher, Fluka, Research Organics, or Sigma or synthesized as described (5).

RESULTS

ycf1Δ Mutants Are Defective in GSH-Dependent Cd²⁺ Transport.

Physiological (1 mM) concentrations of GSH (13) promoted Cd²⁺ uptake by vacuolar membrane vesicles purified from the wild-type strain DTY165 but not the ycf1Δ mutant strain DTY167 (Fig. 1). Addition of Cd²⁺ (80 μM) to GSH-containing media elicited MgATP-dependent, uncoupler-insensitive ¹⁰⁹Cd²⁺ uptake rates of 4.5 and 0.8 nmol/mg per min by DTY165 and DTY167 membranes, respectively (Fig. 1 A and B). Uptake by DTY165 membranes was diminished more than 9-fold by the omission of GSH (Fig. 1A) whereas uptake by DTY167 membranes was slightly stimulated (Fig. 1B).

Figure 1.

Uptake of Cd²⁺ into vacuolar membrane vesicles purified from DTY165 and DTY167 cells. Uptake of ¹⁰⁹Cd²⁺ by DTY165 membranes (A) or DTY167 membranes (B) was measured in the absence of MgATP plus (○) or minus GSH (1 mM) (□) or in the presence of MgATP (3 mM) plus (•) or minus (▪) GSH. ¹⁰⁹CdSO₄ and gramicidin-D were added at concentrations of 80 μM and 5 μM, respectively. (C) Rate of ¹⁰⁹Cd²⁺ uptake by DTY165 membranes plotted as a function of the total concentration of Cd²⁺ ([Cd²⁺]) added to uptake media containing 1 mM GSH, 3 mM MgATP, and 5 μM gramicidin-D. Values shown are means ± SE (n = 3–6).

GSH maximally stimulated uptake within minutes (t½ < 5 min) of the addition of Cd²⁺ to the uptake medium (data not shown) and uptake was sigmoidally dependent on Cd²⁺ concentration, achieving half-maximal velocity at 120 μM (Fig. 1C).

Specific Requirement for GSH.

The stimulatory action of GSH was abolished by the omission of MgATP from the assay medium (Fig. 1 and Table 1), and 1 mM concentrations of oxidized GSH (GSSG), S-methylglutathione, cysteinylglycine, cysteine, or glutamate did not promote MgATP-dependent, uncoupler-insensitive Cd²⁺ uptake by vacuolar membrane vesicles from either strain (Table 1).

Table 1.

Effects of different GSH-related compounds on uncoupler-insensitive ¹⁰⁹Cd uptake by vacuolar membrane vesicles purified from DTY165 and DTY167 cells

Compound	109Cd uptake, nmol/mg/10 min
	DTY165		DTY167
	− MgATP	+ MgATP	− MgATP	+ MgATP
Cd2+	5.8 ± 2.4	5.6 ± 1.5	4.3 ± 1.3	4.6 ± 2.1
Cd2+ + GSH	4.2 ± 1.2	37.4 ± 4.5	3.3 ± 1.1	8.3 ± 2.7
Cd2+ + GSSG		5.1 ± 3.2		3.8 ± 2.3
Cd2+ + GS-CH3		4.5 ± 1.9		3.7 ± 3.1
Cd2+ + Cys-Gly		5.6 ± 3.2		6.9 ± 1.4
Cd2+ + Cys		7.0 ± 1.2		3.9 ± 1.0
Cd2+ + Glu		5.7 ± 1.1		5.2 ± 1.3

GSH, oxidized glutathione (GSSG), S-methylglutathione (GS-CH₃), cysteinylglycine, cysteine, and glutamate were added at concentrations of 1 mM. MgATP, ¹⁰⁹CdSO₄, and gramicidin-D were added at concentrations of 3 mM, 80 μM, and 5 μM, respectively. Values shown are means ± SE (n = 3–6). 

Purification of Transport-Active Complex.

To determine the mode of action of GSH and the form in which Cd²⁺ is transported, reaction mixtures initially containing Cd²⁺ and GSH were fractionated and YCF1-dependent uptake was assayed.

Incubation of ¹⁰⁹Cd²⁺ with GSH and gel-filtration of the mixture on Sephadex G-15 yielded two major ¹⁰⁹Cd-labeled peaks: a low molecular weight peak [LMW-mono(glutathionato)cadium (Cd·GS)] and a high molecular weight peak (HMW-Cd·GS) (Fig. 2A). When rechromatographed on Mono-Q, LMW-Cd·GS and HMW-Cd·GS eluted at 0 (Fig. 2C) and 275 mM NaCl, respectively (Fig. 2B). Of these two ¹⁰⁹Cd-labeled components, HMW-Cd·GS, alone, underwent YCF1-dependent transport. MgATP-dependent, uncoupler-insensitive HMW-¹⁰⁹Cd·GS uptake by DTY165 membranes increased as a single Michaelian function of concentration (K_m, 39.1 ± 14.1 μM; V_max, 157.2 ± 30.4 nmol/mg/10 min) (Fig. 3A). By contrast, uptake of LMW-¹⁰⁹Cd·GS by DTY165 membranes was negligible at all of the concentrations examined (Fig. 3B). Vacuolar membranes from DTY167 cells transported neither HMW-¹⁰⁹Cd·GS nor LMW-¹⁰⁹Cd·GS (Fig. 3).

Figure 2.

Purification of cadmium·glutathione complexes by gel-filtration (A) and anion-exchange chromatography (B and C). ¹⁰⁹CdSO₄ (20 mM) was incubated with 40 mM GSH at 45°C for 24 h, and the mixture was chromatographed on Sephadex G-15 to resolve a high molecular weight ¹⁰⁹Cd-labeled component (HMW-¹⁰⁹Cd·GS) from a low molecular weight component (LMW-¹⁰⁹Cd·GS) (A). The peaks corresponding to HMW-¹⁰⁹Cd·GS and LMW-¹⁰⁹Cd·GS were then chromatographed on Mono-Q and eluted with a linear NaCl gradient (- — -) (B and C). ¹⁰⁹Cd (cpm × 10⁻³) was determined on 5 μl aliquots of the column fractions by liquid scintillation counting. Solid bars denote fractions subjected to further manipulations.

Figure 3.

Kinetics of MgATP-dependent, uncoupler-insensitive ¹⁰⁹Cd·GS₂ (HMW-¹⁰⁹Cd·GS; A) and ¹⁰⁹Cd·GS (LMW-¹⁰⁹Cd·GS; B) uptake. S-(2,4-dinitrophenyl)glutathione (DNP-GS) was added at the concentrations (μM) indicated to DTY165 membranes (•, ○, ▪, □, ▵) or DTY167 membranes (⋄). A secondary plot of the apparent Michaelis constants for Cd·GS₂ uptake (K_m^app/Cd·GS₂) as a function of DNP-GS concentration is shown (C). The kinetic parameters for Cd·GS₂ transport by DTY165 membranes were K_m, 39.1 ± 14.1 μM, V_max, 157.2 ± 30.4 nmol/mg/10 min, and K_i(DNP-GS), 11.3 ± 2.1 μM. Kinetic parameters were computed by nonlinear least squares analysis (14). Values shown are means ± SE (n = 6).

Cd·GS₂ Is the Transport-Active Complex.

The transport-active complex, HMW-Cd·GS, was identified as Cd·GS₂ by three criteria. (i) The average Cd/GS molar ratio of the transported species, estimated from the ¹⁰⁹Cd/³H ratios of the HMW-Cd·GS peaks obtained after chromatography of reaction mixtures initially containing ¹⁰⁹Cd²⁺ and [³H]GSH on Sephadex G-15 and Mono-Q were 0.44 ± 0.09 and 0.49 ± 0.17, respectively (Table 2). (ii) DTY165 membranes accumulated ¹⁰⁹Cd and [³H]GS in a molar ratio of 0.49 ± 0.01 when incubated in media containing HMW-¹⁰⁹Cd·[³H]GS, MgATP, and gramicidin-D (Table 2). (iii) The principal ion peak detected after MALD-MS of HMW-Cd·GS had an m/z ratio of 725.4 ± 0.7, consistent with the molecular weight of Cd·GS₂ (724.6 Da, Fig. 4). The transport-inactive complex, LMW-Cd·GS, on the other hand, was tentatively identified as Cd·GS on the basis of its smaller apparent molecular size (Fig. 2A), failure to bind Mono-Q (Fig. 2C), and Cd/GS ratio of 0.67 ± 0.04 and 0.86 ± 0.07 after chromatography on Sephadex G-15 and Mono-Q (Table 2), respectively.

Table 2.

Molar Cd/GS ratios of LMW-Cd·GS and HMW-Cd·GS complexes fractionated by Sephadex G-15 and Mono-Q chromatography (Fig. 2) before and after MgATP-dependent, uncoupler-insensitive uptake by vacuolar membrane vesicles purified from DTY165 and DTY167 cells

Fraction	109Cd uptake, nmol/mg/10 min		Molar ratio Cd/GS
	DTY165	DTY167	Before uptake	After uptake
Sephadex G-15
HMW-Cd·GS			0.44 ± 0.09
LMW-Cd·GS			0.67 ± 0.04
Mono-Q
HMW-Cd·GS	66.3 ± 2.7	5.6 ± 2.6	0.49 ± 0.17	0.49 ± 0.01
LMW-Cd·GS	4.4 ± 0.8	3.9 ± 1.4	0.86 ± 0.07
After 2-ME
HMW-Cd·GS	11.9 ± 2.4	4.4 ± 3.0

Cd/GS ratios were estimated from the ¹⁰⁹Cd/[³H] radioisotope ratios of samples prepared from ¹⁰⁹CdSO₄ and [³H]GSH. HMW-¹⁰⁹Cd·[³H]GS was pretreated with 2-mercaptoethanol (2-ME) by heating a 1:4 mixture of HMW-¹⁰⁹Cd.[³H]GS with 2-ME at 60°C for 10 min before measuring ¹⁰⁹Cd²⁺ uptake. Uptake was measured using 50 μM concentrations (as Cd) of the complexes indicated in standard uptake medium containing 5 μM gramicidin-D. Values shown are means ± SE (n = 3–6). 

Figure 4.

Matrix-assisted laser desorption mass spectrometry of HMW-Cd·GS. MALD-MS was performed on Sephadex G-15-, Mono-Q-purified HMW-Cd·GS as described. The molecular structure inferred from a mean m/z ratio of 725.4 ± 0.7 (n = 9) and average Cd/GS stoichiometry of 0.5 (Cd·GS₂, molecular weight 724.6 Da) is shown.

While an m/z ratio of 725 for HMW-Cd·GS would be equally compatible with the transport of Cd·GSSG, this is refuted by two findings: (i) GSSG, alone, does not promote YCF1-dependent uptake (Table 1), and (ii) the transport-active complex is probably a mercaptide. Pretreatment of HMW-Cd·GS with 2-mercaptoethanol inhibits MgATP-dependent, uncoupler-insensitive Cd²⁺ uptake by DTY165 membranes by more then 80% (Table 2), and S-methylation abolishes the stimulatory action of GSH (Table 1).

Cd·GS₂ Transport Is Directly Energized by MgATP.

Purification of Cd·GS₂ enabled the energy requirements of YCF1-dependent transport to be examined directly and confirmed that more than 83% of the MgATP-dependent, uncoupler-insensitive Cd²⁺ transport measured using DTY165 membranes was mediated by YCF1. Agents that dissipate both the ΔpH and Δψ components of the H⁺-electrochemical gradient established by the vacuolar H⁺-ATPase [V-ATPase; carbonylcyanide p-trifluoromethoxyphenylhydrazone (FCCP), gramicidin-D] or directly inhibit the V-ATPase, itself (bafilomycin A₁), decreased MgATP-dependent Cd·GS₂ uptake by vacuolar membrane vesicles from DTY165 cells by 22% (Table 3). Ammonium chloride, which abolishes ΔpH while leaving Δψ unaffected, on the other hand, inhibited uptake by only 15% (Table 3). From these results and the inability of uncouplers to markedly increase the inhibitions caused by V-ATPase inhibitors, alone (Table 3), Cd·GS₂ uptake by wild-type membranes is inferred to proceed via a YCF1-dependent, MgATP-energized pathway that accounts for most of the transport measured and a YCF1-independent pathway, primarily driven by the H⁺-gradient established by the V-ATPase, that makes a minor contribution to total uptake.

Table 3.

Effects of uncouplers and V-ATPase inhibitors on uptake of Cd·GS₂ by vacuolar membrane vesicles purified from DTY165 and DTY167 cells

Addition	Cd·GS2 uptake, nmol/mg/10 min
	DTY165	DTY167
Control	105.8 ± 12.4 (100)	17.3 ± 2.7 (100)
Gramicidin-D	77.8 ± 6.4 (73.5)	9.8 ± 2.0 (56.6)
FCCP	62.2 ± 11.4 (58.8)	10.2 ± 1.6 (59.0)
NH4Cl	89.8 ± 8.2 (84.8)	10.0 ± 1.7 (57.8)
NH4Cl + gramicidin-D	69.8 ± 12.0 (66.0)	8.8 ± 2.2 (50.9)
Bafilomycin A1	81.8 ± 6.0 (76.6)	12.8 ± 3.6 (74.0)
Bafilomycin A1 + gramicidin-D	70.2 ± 12.2 (66.4)	7.2 ± 2.4 (41.6)

Uptake was measured in standard uptake medium containing 50 μM purified ¹⁰⁹Cd·GS₂. Bafilomycin A₁, carbonylcyanide p-trifluoromethoxyphenylhydrazone (FCCP), gramicidin-D, and NH₄Cl were added at concentrations of 0.5 μM, 5 μM, 5 μM, and 1 mM, respectively. Values outside parentheses are means ± SE (n = 3–6); values inside parentheses are rates of uptake expressed as percentage of control. 

Cd·GS₂ Competes with DNP-GS for Transport.

As would be predicted if Cd·GS₂ and the model organic GS-conjugate, DNP-GS, follow the same transport pathway, the K_i for inhibition of MgATP-dependent, uncoupler-insensitive Cd·GS₂ uptake by DNP-GS (11.3 ± 2.1 μM; Fig. 3 A and C) coincided with the K_m for DNP-GS transport (14.1 ± 7.4 μM, ref. 5).

Pretreatment with Cd²⁺ or CDNB Increases YCF1 Expression.

RNase protection assays of YCF1 expression in DTY165 cells and measurements of MgATP-dependent, uncoupler-insensitive ¹⁰⁹Cd·GS₂ and [³H]DNP-GS uptake by vacuolar membranes prepared from the same cells after 24 h of growth in media containing CdSO₄ (200 μM) or the cytotoxic DNP-GS precursor, CDNB (150 μM), demonstrated a parallel increase in all three quantities. YCF1-specific mRNA levels were increased by 1.9- and 2.5-fold by pretreatment of DTY165 cells with CdSO₄ and CDNB, respectively (Fig. 5). The same pretreatments increased MgATP-dependent, uncoupler-insensitive ¹⁰⁹Cd·GS₂ uptake into vacuolar membrane vesicles by 1.4- and 1.7-fold and [³H]DNP-GS uptake by 1.6- and 2.8-fold (Fig. 5).

Figure 5.

Induction of YCF1 expression and YCF1-dependent Cd·GS₂ and DNP-GS transport by pretreatment of DTY165 cells with CdSO₄ (Cd²⁺, 200 μM) or CDNB (150 μM) for 24 h. YCF1-specific mRNA and 18S rRNA were detected in the total RNA extracted from control or pretreated cells (10 μg/lane) by RNase protection. Uptake of ¹⁰⁹Cd·GS₂ (50 μM) or [³H]DNP-GS (66.2 μM) by vacuolar membrane vesicles was measured in standard uptake medium containing 5 μM gramicidin-D. Values shown are means ± SE (n = 3).

DISCUSSION

These investigations provide an indication of the mechanism by which YCF1 confers Cd²⁺ resistance in S. cerevisiae and its relationship to the transport of organic GS-conjugates by demonstrating that the integral membrane protein encoded by this gene specifically catalyzes the MgATP-energized uptake of Cd·GS₂ by vacuolar membrane vesicles. The codependence of Cd·GS₂ and organic GS-conjugate transport on YCF1 is evident at multiple levels. (i) The ycf1Δ mutant strain, DTY167, is hypersensitive to Cd²⁺ and CDNB in the growth medium, and both hypersensitivities are alleviated by transformation with plasmid-borne YCF1 (4, 5). (ii) Vacuolar membrane vesicles purified from DTY167 cells are grossly impaired for MgATP-energized, uncoupler-insensitive organic GS-conjugate and GSH-promoted Cd²⁺ uptake. (iii) Cd·GS₂ and organic GS-conjugate compete for the same uptake sites on YCF1. (iv) Factors that increase YCF1 expression elicit a parallel increase in Cd·GS₂ and organic GS-conjugate transport. Thus, a number of ostensibly disparate observations—the strong association between cellular GSH levels and Cd²⁺ resistance (e.g., ref. 15), the markedly increased sensitivity of vacuole deficient S. cerevisiae strains to Cd²⁺ toxicity (M.S. and D.J.T., unpublished results), and the coordinate regulation of the yeast YCF1 and GSH1 genes, the latter of which encodes γ-glutamylcysteine synthetase (16, 17)—are now explicable in terms of a model in which YCF1 catalyzes the GSH-dependent vacuolar sequestration of Cd²⁺.

While Tommasini et al. (18) have recently confirmed our earlier findings (5) by showing that YCF1 is a vacuolar GS-conjugate transporter, and gone on to demonstrate that transformation with human MRP1 suppresses the ycf1Δ mutation, their results differ from ours in their inability to detect MgATP-energized, GSH-dependent Cd²⁺ uptake by microsomes from MRP1-transformed yeast. Because of this and their failure to inhibit DNP-GS transport by the inclusion of GSH (1 mM) and Cd²⁺ (100–300 μM) in the uptake medium, these authors attribute the Cd²⁺ resistance conferred by MRP1 to a two-phase process: the initial transport of Cd²⁺ into the vacuole by H⁺/Cd²⁺ antiport and its subsequent complexation with an unknown binding molecule transported by MRP1. Two alternative conclusions therefore follow. Either MRP1 and YCF1 do not have precisely the same transport mechanism or Tommasini et al.’s inability to measure GSH-promoted Cd²⁺ transport has a methodological basis. Of these, the second explanation is the more likely. First, the idea that MRP1 and YCF1 use identical mechanisms for the uptake of organic GS-conjugates but different mechanisms for Cd²⁺ seems unnecessarily complicated when both transporters must encounter the same microenvironment in yeast. Second, the two-phase mechanism does not explain how MRP1 phenocopies YCF1 and how DNP-GS inhibits YCF1-dependent Cd·GS₂ transport in a simple competitive manner. Third, inspection of the procedures employed by Tommasini et al (18)—methods originally developed in our laboratory for a different purpose (9)—reveals a critical technical error: systematic contamination of the Cd²⁺ uptake media with approximately 140 μM concentrations of dithiothreitol and EDTA. We know from preliminary experiments that these concentrations are more than sufficient to cause a near total abolition of GSH-dependent Cd²⁺ uptake by YCF1 while leaving organic GS-conjugate uptake unaffected. Dithiothreitol reacts with Cd²⁺ to form a transport-inactive dimercaptide (Z.-S.L. and P.A.R., unpublished results); EDTA chelates free Cd²⁺ to quench its reaction with GSH.

Further verification of the one-phase model is provided by the functional equivalence of YCF1 to MRP1 in all other regards. At the biochemical level, YCF1 specifically catalyzes the transport of Cd·GS₂. This is the most straightforward explanation of the mass and 1:2 Cd/GS ratio of the transport-active complex and the finding that the concentration-dependence of uptake assumes a Michaelian, rather than sigmoidal, function with a moderately low K_m when Cd·GS₂ is purified from the other components present in a mixture of Cd²⁺ and GSH. Analogously, MRP1 catalyzes the transport of bis(glutathionato)platinum (Pt·GS₂) (19). MALD-MS and ¹H-NMR spectroscopy of the transport-active complex formed when the anticancer drug, cisplatin, is combined with GSH indicate that Pt spontaneously complexes with GSH in a 1:2 ratio. Two GSH molecules coordinate each Pt²⁺ ion via their cysteinyl sulfur and amino nitrogen atoms to generate the transport-active complex. At the cellular level, YCF1 and MRP1 confer resistance to and are induced by a similar spectrum of xenobiotics. YCF1 confers cross-resistance to organic xenobiotics and heavy metals (4, 5). Enhanced expression of MRP1 in mammalian cells is associated with cross-resistance to cisplatin and heavy metals, such as cadmium and arsenite, as well as GSH-conjugable drugs (7, 20). Expression of YCF1 is increased by exposure of cells to GSH-conjugable xenobiotics and Cd²⁺. Expression of MRP1 is increased by exposure of cells to arsenite, Cd²⁺ and Zn²⁺ (20).

The existence of phytochelatins (PCs), peptides consisting of repeating units of γ-glutamylcysteine followed by a C-terminal glycine [(γ-Glu-Cys)_n-Gly], in the fission yeast, Schizosaccharomyces pombe (21) and the isolation and characterization of the heavy metal tolerance (HMT1) gene, a six transmembrane span-single nucleotide binding fold (“single half”) ATP binding cassette (ABC) transporter, responsible for vacuolar uptake of PC and Cd·PC, from the same organism (21, 22), prompts two related questions. Do similar mechanisms of Cd²⁺ sequestration operate in S. cerevisiae? Is YCF1 competent in the transport of Cd·PC as well as Cd·GS₂? The first question has not been resolved but PCs are more widespread in fungi than was once thought. Although heavy metal detoxification in fungi has been largely attributed to metallothioneins (23), recent studies of Candida glabrata, S. cerevisiae, and Neurospora crassa also implicate PCs. Exposure of C. glabrata to copper salts stimulates metallothionein formation but exposure to cadmium salts stimulates PC formation (24). In both S. cerevisiae and N. crassa, PCs have been detected and in the latter synthesis is not only activated by Cd²⁺ but also Zn²⁺ (25). Interestingly, in all three organisms, PC₂ is the dominant PC species found. Hence, the second question or a modification thereof. Given the structural resemblance between Cd·PC₂ [Cd·(γ-Glu-Cys)₂-Gly] and Cd·GS₂, might YCF1 be able to transport either complex? The answer to this question is no: YCF1 does not transport PC₂. PC₂ fractions purified from S. pombe show no evidence of YCF1-dependent transport by vacuolar membrane vesicles isolated from S. cerevisiae irrespective of whether the PC preparations are complexed with Cd²⁺ or not (Z.-S.L. and P.A.R., unpublished results). The novelty of YCF1-mediated Cd·GS₂ transport as a mechanism for the vacuolar sequestration of Cd²⁺ is therefore substantiated.