The glycerol channel Fps1p mediates the uptake of arsenite and antimonite in Saccharomyces cerevisiae


  • †Present address: Department of Cell and Molecular Biology/Microbiology, Göteborg University, Box 462, S-405 30 Göteborg, Sweden.


The Saccharomyces cerevisiae FPS1 gene encodes a glycerol channel protein involved in osmoregulation. We present evidence that Fps1p mediates influx of the trivalent metalloids arsenite and antimonite in yeast. Deletion of FPS1 improves tolerance to arsenite and potassium antimonyl tartrate. Under high osmolarity conditions, when the Fps1p channel is closed, wild-type cells show the same degree of As(III) and Sb(III) tolerance as the fps1Δ mutant. Additional deletion of FPS1 in mutants defective in arsenite and antimonite detoxification partially suppresses their hypersensitivity to metalloid salts. Cells expressing a constitutively open form of the Fps1p channel are highly sensitive to both arsenite and antimonite. We also show by direct transport assays that arsenite uptake is mediated by Fps1p. Yeast cells appear to control the Fps1p-mediated pathway of metalloid uptake, as expression of the FPS1 gene is repressed upon As(III) and Sb(III) addition. To our knowledge, this is the first report describing a eukaryotic uptake mechanism for arsenite and antimonite and its involvement in metalloid tolerance.


Arsenic and antimony are toxic metalloids with a long history of usage as therapeutic agents. Paul Ehrlich showed at the beginning of the twentieth century that some infectious diseases such as syphilis and trypanosomiasis could be treated successfully with organic derivatives of arsenic (Xu et al., 1998). In addition, potassium arsenite was used for controlling the level of leucocytes in patients suffering from chronic myelocytic leukaemia until the 1930s (Forkner and McNair-Scott, 1931). Although these drugs have been replaced with more effective and less toxic agents, the organic arsenical melarsoprol is still in use for the treatment of sleeping sickness. Likewise, all forms of leishmaniasis are treated with drugs containing the related metalloid antimony (Borst and Ouellette, 1995).

Arsenic trioxide has been shown recently to be a highly effective and relatively safe drug in treating patients with acute promyelocytic leukaemia who are resistant to conventional chemotherapy (Shen et al., 1997; Soignet et al., 1998). In vitro studies have revealed that antimonials and melarsoprol inhibit growth and induce apoptosis of leukaemia cells in a similar way to arsenic trioxide and could therefore be used in antileukaemic therapy (Müller et al., 1998; Wang et al., 1998;Rousselot et al., 1999). There is promising evidence that arsenic trioxide causes apoptosis of both haematologic and solid tumour cell lines in vitro and that it might also be effective in vivo (König et al., 1997; Yang et al., 1999; Zhang et al., 1999). However, the development of clinical resistance in some patients treated with arsenic trioxide suggests the presence of resistance mechanisms (Soignet et al., 1998). Hence, it is of interest to elucidate the molecular mechanisms involved in resistance to both arsenic and antimony in eukaryotes.

Although tolerance to arsenic and antimony salts is well defined in bacteria, there has been little information on such mechanisms in eukaryotic cells (Silver, 1998;Rosen, 1999a, b). Several genes of the eukaryotic model organism Saccharomyces cerevisiae (baker's yeast) have been shown recently to affect metalloid tolerance, which has contributed to a rapid progress in understanding this phenomenon.

Arsenite [As(III)] and arsenate [As(V)] tolerance in yeast is mediated by a cluster of three genes: ACR1 (ARR1, YAP8), ACR2 (ARR2) and ACR3 (ARR3) (Bobrowicz et al., 1997). The ACR1 gene encodes a transcription factor belonging to the fungus-specific Yap family of bZIP proteins (Fernandes et al., 1997). Acr1p appears to be involved in controlling the induction of ACR2 and ACR3 expression by arsenite (Bobrowicz and Ulaszewski, 1998). Acr2p is a cytosolic arsenate reductase that converts arsenate to arsenite in the presence of glutaredoxin and glutathione (Mukhopadhyay and Rosen, 1998; Mukhopadhyay et al., 2000). Arsenite is the substrate for the plasma membrane transporter Acr3p that mediates the extrusion of this metalloid out from the cell (Wysocki et al., 1997;Ghosh et al., 1999). A second detoxification mechanism involves the ABC transporter Ycf1p, which transports arsenite and antimonite into the vacuole in the form of glutathione conjugates (Ghosh et al., 1999).

Active extrusion or sequestration in subcellular organelles are not the sole strategies adopted by organisms to acquire tolerance to drugs and toxic substances (Rosen, 1996;Kolaczkowski and Goffeau, 1997). Cells may also increase tolerance by altering the membrane composition or by inhibiting uptake systems.

Arsenate is an analogue of inorganic phosphate, and both anions share a common transport system. Bacteria with defective phosphate uptake exhibit increased tolerance to arsenate (Bennett and Malamy, 1970; Willsky and Malamy, 1980). Similarly, yeast cells bearing mutations in the PHO84, -86, -87 and -88 genes, encoding components involved in inorganic phosphate uptake, are more tolerant to arsenate than wild-type cells (Bun-ya et al., 1996; Yompakdee et al., 1996a,b). Arsenate uptake in KB carcinoma cells is inhibited by phosphate, also suggesting a common transport system in mammals (Huang and Lee, 1996).

Up to now, the identity of the protein(s) involved in arsenite and antimonite uptake has not been determined. It has been suggested that arsenite uptake may be mediated by a carrier protein in a similar way to that by which unionized organic compounds are taken up or that it enters cells through simple diffusion (Cervantes et al., 1994; Huang and Lee, 1996). Recently, the glycerol facilitator GlpF has been suggested to mediate antimonite uptake in Escherichia coli, as mutation of the gene resulted in increased tolerance (Sanders et al., 1997).

The S. cerevisiae FPS1 gene encodes a glycerol transporter that is homologous to the E. coli glycerol facilitator GlpF, and both proteins belong to the ubiquitous MIP (major intrinsic protein) family of channel proteins (Van Aelst et al., 1991). Fps1p mediates both uptake and efflux of glycerol from yeast cells (Luyten et al., 1995; Sutherland et al., 1997). We have shown recently that Fps1p plays a central role in yeast osmoadaptation by controlling the intracellular level of the osmolyte glycerol (Tamás et al., 1999). Importantly, Fps1p is gated by osmolarity changes. Fps1p is closed under hyperosmotic stress to enable intracellular accumulation of glycerol, whereas it is open under low-osmolarity conditions to allow glycerol efflux. The N-terminal extension of Fps1p is required for channel closure, and its deletion results in constitutive and unregulated glycerol transport activity. Hence, different FPS1 alleles provide an experimental tool to manipulate the transmembrane transport activity of this protein (Tamás et al., 1999; 2000).

In this work, we have studied the role of Fps1p in metalloid transport and tolerance. Physiological data and transport studies clearly demonstrate that S. cerevisiae takes up arsenite and antimonite via the Fps1p glycerol channel.


Disruption of the FPS1 gene confers arsenite and antimonite tolerance

Inactivation of the E. coli glycerol facilitator GlpF has been shown to increase bacterial tolerance to antimonite (Sanders et al., 1997). We decided to investigate whether inactivation of the yeast glycerol channel Fps1p, which is closely related to the GlpF protein, would result in a similar phenotype. Both wild-type and fps1Δ cells were able to grow in the presence of as much as 120 mM potassium antimonyl tartrate on glucose-containing medium (Table 1; Fig. 1). Antimony salts precipitate in agar at concentrations above 120 mM. Therefore, we could not exactly define the concentration required for growth inhibition, but classified the corresponding minimum inhibitory concentration (MIC) as > 120 mM (Table 1). S. cerevisiae cells are about 60 times more sensitive to antimony salts when glycerol is the only source of carbon in the medium (Wysocki et al., 1997). When we tested growth on glycerol, the fps1Δ strain was clearly more tolerant to Sb(III) than the parental strain (Fig. 1). The growth of the wild type was completely inhibited in the presence of 2 mM potassium antimonyl tartrate, whereas deletion of FPS1 increased tolerance of the mutant up to the maximum concentration tested [120 mM Sb(III)].

Table 1.   Sensitivity of yeast strains to arsenic and antimony salts.
StrainMICa (mM)
  • a.

    The MIC is the concentration at which no growth was observed on glucose medium. The MIC values were determined on the basis of three independent experiments with identical results.

  • b.

    As(III) is sodium arsenite.

  • c.

    As(V) is sodium arsenate.

  • d.

    Sb(III) is potassium antimonyl tartrate.

WT + YEplac1951.206.0> 120.0
fps1Δ + YEplac1957.002.0> 120.0
fps1Δ + YEpfps1-Δ10.802.00.7
fps1Δ + YEpGlpF7.002.0> 120.0
ycf1Δ + YEplac1950.705.00.2
ycf1Δ + YEpfps1-Δ10.405.00.1
fps1Δ ycf1Δ + YEplac1957.002.015.0
acr3Δ + YEplac1950.202.0> 120.0
acr3Δ + YEpfps1-Δ10.102.00.7
fps1Δ acr3Δ + YEplac1951.202.0> 120.0
ycf1Δ acr3Δ + YEplac1950.052.00.2
fps1Δ ycf1Δ acr3Δ + YEplac1950.302.02.0
Figure 1.

Deletion of the FPS1 gene increases tolerance of S. cerevisiae cells to arsenite and antimonite. Wild-type and fps1Δ cells were grown in liquid medium, and 10-fold serial dilutions of the cultures were spotted on agar plates with or without the addition of metalloid salts as described in Experimental procedures.

Mutation of the glpF gene did not affect bacterial tolerance to arsenite or to several metal salts, indicating that tolerance of this mutant is restricted to trivalent antimony salts (Sanders et al., 1997). Similarly, the fps1Δ mutant displayed no increased tolerance when Cd(II), Cu(II), Mn(II), Sb(V), Te(IV) or Zn(II) salts were added to the growth medium (data not shown). Surprisingly, deletion of FPS1 resulted in arsenate sensitivity (Table 1; Fig. 1). As it is unlikely that arsenate is a substrate for Fps1p, the observed phenotype could be a result of other effects caused by deletion of FPS1 (see Discussion).

The most interesting result was obtained with sodium arsenite. The fps1Δ mutant was about 5.8-fold more tolerant to As(III) than the wild type (Table 1; Fig. 1). We conclude that the loss of the glycerol channel Fps1p increases tolerance to arsenite and antimonite, possibly as a result of inhibition of the uptake system.

High osmolarity increases yeast tolerance to metalloid salts

We have shown previously that Fps1p closes during hyperosmotic stress to enable the intracellular accumulation of glycerol (Luyten et al., 1995; Tamás et al., 1999). If arsenite and antimonite enter yeast cells via Fps1p, one would expect high-osmolarity conditions to result in an increased metalloid tolerance, as was observed for cells lacking the Fps1p channel. This was indeed the case. Under low-osmolarity conditions, the growth of the wild type was severely inhibited in the presence of 1 mM arsenite. However, when the osmolarity of the growth medium was increased by the addition of 0.8 M NaCl, the wild-type strain showed an enhanced arsenite tolerance and grew as well as the fps1Δ mutant (Fig. 2).

Figure 2.

The sensitivity of the ycf1Δ and acr3Δ mutants to arsenite and antimonite can be alleviated either by (A) additional deletion of the FPS1 gene or by (B) increasing the osmolarity of the growth medium. Yeast cells were grown in liquid medium, and 10-fold serial dilutions of the cultures were spotted on agar plates. To apply hyperosmotic stress, cells were grown in the presence of 0.8 M NaCl.

Two transporters involved in arsenite and antimonite detoxification have been described recently (Wysocki et al., 1997;Ghosh et al., 1999). Ycf1p, a member of the ABC transporter superfamily, mediates uptake of As(III) and probably also of Sb(III) as glutathione conjugates into the vacuole (Ghosh et al., 1999). The ycf1Δ mutant was highly sensitive to both arsenite and antimonite (Table 1; Fig. 2A; Ghosh et al., 1999). High-osmolarity conditions or additional deletion of FPS1 largely suppressed this sensitivity, but to different extents (Fig. 2). The double fps1Δ ycf1Δ and single fps1Δ mutants were equally tolerant to arsenite (Fig. 2A). Additional deletion of FPS1 or hyperosmotic stress increased the tolerance of the ycf1Δ mutant to antimonite about 150-fold (Table 1). However, cells lacking Ycf1p were still somewhat more sensitive than the wild type (Fig. 2).

Acr3p is located in the plasma membrane and mediates arsenite extrusion from yeast cells (Wysocki et al., 1997;Ghosh et al., 1999). Deletion of ACR3 resulted in hypersensitivity to As(III) but did not affect Sb(III) tolerance (Table 1; Fig. 2A). The double ycf1Δ acr3Δ mutant showed an additive hypersensitivity to arsenite, but not to antimonite (Table 1; Fig. 2A). Deletion of FPS1 in the single acr3Δ mutant or in the double ycf1Δ acr3Δ mutant as well as hyperosmotic stress partially suppressed the sensitivity of these mutants to arsenite, albeit not to the wild-type level (Fig. 2). Interestingly, the triple fps1Δ ycf1Δ acr3Δ mutant was slightly more sensitive to antimonite than the double fps1Δ ycf1Δ mutant (Table 1; Fig. 2A). Thus, Acr3p may also be involved in antimonite tolerance. Together, these results suggest that both As(III) and Sb(III) enter yeast cells via Fps1p. However, both metalloids may possess additional entry routes.

Expression of a constitutively open glycerol channel causes hypersensitivity to arsenite and antimonite

Deletion of the N-terminal extension of Fps1p has been shown to result in a constitutively open channel with a high level of unregulated glycerol transport (Tamás et al., 1999). We examined the growth of yeast cells expressing the open Fps1p channel (encoded by the fps1-Δ1 allele) in the presence of metalloid salts (Fig. 3A). Transformants containing the YEpfps1-Δ1 plasmid were hypersensitive to sodium arsenite, potassium antimonyl tartrate and antimony(III) chloride (Table 1; Fig. 3A). As this sensitivity was similar irrespective of the metalloid salt used, the form of the As(III) or Sb(III) compound is probably not significant for the phenotype.

Figure 3.

Expression of the constitutively open Fps1p glycerol channel causes metalloid hypersensitivity.

A. The sensitivity to arsenite [As(III)] and to various trivalent antimony salts [Sb(III)] is increased in fps1Δ cells transformed with the YEpfps1-Δ1 plasmid containing the open Fps1p channel.

B. The ycf1Δ and acr3Δ mutants transformed with the YEpfps1-Δ1 plasmid are more sensitive to trivalent arsenic and antimony salts than the same mutants transformed with an empty vector.

C. Yeast cells transformed with the YEpGlpF plasmid containing the bacterial glycerol facilitator GlpF grow as well as the control strain with an empty plasmid in the presence of both arsenite and antimonite. The transformants were spotted onto plates containing either metalloid salts or 0.8 M NaCl as a control. Yeast cells were transformed with the indicated plasmids, and the transformants were grown as described in Experimental procedures.

Expression of the constitutively open Fps1p channel in the ycf1Δ or acr3Δ mutants further increased their hypersensitivity to both metalloids (Table 1; Fig. 3B). The enhanced sensitivity is probably a consequence of increased intracellular levels of the toxic metalloids.

We have previously expressed the E. coli glycerol facilitator and Fps1p homologue GlpF in yeast cells. Similar to the constitutively open Fps1p channel, GlpF mediates unregulated glycerol transport in yeast (Tamás et al., 1999). Unexpectedly, expression of glpF in yeast did not result in an increased sensitivity to either arsenite or antimonite (Table 1; Fig. 3C). In spite of this, GlpF is functional and active, as fps1Δ cells transformed with the YEpGlpF plasmid grew poorly on plates containing 0.8 M NaCl (Fig. 3C; Tamás et al., 1999).

The intracellular glycerol level does not influence metalloid tolerance

The polyol glycerol plays an important protective role in the cellular response of yeast to various stress conditions (Albertyn et al., 1994; Siderius et al., 2000;Påhlman et al., 2001). Deletion of FPS1 or expression of the constitutively open Fps1p channel has been shown to alter the intracellular glycerol level both in the presence and in the absence of osmotic stress (Philips and Herskowitz, 1997; Tamás et al., 1999). Therefore, the cellular glycerol content might influence tolerance to metalloid salts. We have monitored the growth of a gpd1Δ gpd2Δ mutant, lacking the glycerol 3-phosphate dehydrogenase-encoding genes that is devoid of glycerol biosynthesis, in the presence of arsenite and antimonite. As none of the metalloid salts affected the growth of the mutant (data not shown), the intracellular glycerol level is not likely to influence cellular tolerance to arsenite and antimonite.

Fps1p mediates As(III) uptake in yeast

The physiological data described above strongly suggested that As(III) and Sb(III) enter yeast cells via the Fps1p glycerol channel. Therefore, we have measured the uptake of the [74As]-As(III) isotope in various yeast strains.

Wild-type cells consistently took up 20–30% more [74As]-As(III) than cells lacking FPS1(Fig. 4A). On the other hand, cells expressing the open Fps1p channel exhibited a high level of [74As]-As(III) uptake. These cells took up three to five times more [74As]-As(III) than the wild type and the fps1Δ mutant after 60 min (Fig. 4A). The reason for the rather small difference in As(III) uptake between the wild type and the fps1Δ mutant is most probably the salt content of the isotope preparation. The NaCl concentration was determined using the k0 method (De Corte et al., 1987) to about 200 mM in the transport assay. Such a concentration of NaCl increases osmolarity and leads to a reduction in the transport activity of Fps1p (Tamás et al., 1999; unpublished data).

Figure 4.

Fps1p mediates arsenite uptake in S. cerevisiae.

A. The uptake of 10 µM [74As]-As(III) was measured for wild-type cells (black circles), for the fps1Δ mutant (white circles) and for fps1Δ cells in which the fps1-Δ1 allele had been integrated (black triangles).

B. The uptake of [14C]-glycerol was measured for wild-type cells in the absence (black circles) and in the presence (white circles) of 1 mM sodium arsenite as well as in the presence of 1 mM NaCl (black triangles).

We have also monitored Fps1p-mediated glycerol transport in wild-type cells in the presence and absence of sodium arsenite (Fig. 4B). The presence of 1 mM sodium arsenite inhibited glycerol uptake by about 30% after 10 min (Fig. 4B), whereas 1 mM NaCl did not have any effect (Fig. 4B). Thus, the reduction in glycerol transport is likely to be caused by the As(III) in the medium and not by osmotic closure of the Fps1p channel.

As no radioactive Sb(III) isotope is commercially available, we could not test transport of this metalloid directly. Moreover, neither potassium antimonyl tartrate nor antimony(III) chloride caused any inhibition of glycerol transport (data not shown).

Together, these results demonstrate that arsenite enters yeast cells via the Fps1p glycerol channel. The lack of effect of Sb(III) on glycerol transport could result from the poor solubility of antimony salts at the concentration required for inhibition.

Expression of FPS1 is repressed upon arsenite and antimonite addition

All organisms have evolved mechanisms to acquire tolerance to toxic substances. S. cerevisiae controls the cellular level of As(III) and Sb(III) by active extrusion via Acr3p and vacuolar sequestration via Ycf1p (Wysocki et al., 1997;Ghosh et al., 1999). An additional level of control could be by regulating the uptake of these metalloids. We therefore analysed the expression of the FPS1 gene after the addition of 1 mM sodium arsenite or 1 mM potassium antimonyl tartrate to the growth medium. Metalloid addition resulted in a strong repression of the FPS1 gene (Fig. 5). In the case of both arsenite and antimonite, the signal was significantly reduced within the first 15 min after metalloid addition. The repression of the FPS1 gene is likely to be specific, as the metalloids stimulated the expression of certain stress-responsive genes, whereas the expression of genes whose products are not known to be involved in metalloid tolerance was weakly or not affected (M. J. Tamás, unpublished data). Thus, the metalloid entry pathway encoded by FPS1 appears to be regulated by the presence of arsenite and antimonite in the growth medium.

Figure 5.

Expression of FPS1 is regulated by metalloid salts. Northern blot analysis of total RNA extracted from wild-type cells before and after the addition of 1 mM sodium arsenite or 1 mM potassium antimonyl tartrate. The filter was hybridized to 32P-labelled DNA fragments recognizing FPS1 and 18S rRNA, which serves as the loading standard. The graph represents quantification of the mRNA levels of FPS1 relative to those of 18S rRNA after the addition of As(III) (black circles) and Sb(III) (white circles). The highest relative mRNA level was set to 100.


Until recently, relatively little was known about the mechanisms of tolerance to arsenic and antimony in eukaryotes (Rosen, 1999a). Most mechanistic studies of tolerance have previously focused on the extrusion and/or detoxification of arsenite and antimonite, whereas the molecular identity of the proteins mediating the uptake of these toxic metalloids was unknown.

In this study, we have demonstrated that arsenite and antimonite enter yeast cells via the glycerol channel Fps1p. Inactivation of the uptake system, either by deletion of FPS1 or by increasing the osmolarity of the growth medium, resulted in enhanced tolerance. Moreover, cells with a constitutively open and unregulated Fps1p channel were hypersensitive to both arsenite and antimonite. These observations are consistent with a direct role for Fps1p in metalloid uptake. This conclusion is further supported by the fact that the intracellular glycerol level, which increases under hyperosmotic stress, did not appear to affect metalloid tolerance.

Curiously, deletion of FPS1 not only caused arsenite [As(III)] tolerance but also resulted in arsenate [As(V)] sensitivity. Arsenate probably enters yeast cells via the phosphate transporters (Bun-ya et al., 1996; Yompakdee et al., 1996a,b) and is unlikely to be a substrate for Fps1p. Instead, the observed phenotype may be the result of other effects caused by FPS1 deletion, such as an altered membrane composition (Sutherland et al., 1997). Mutants lacking Fps1p also grow poorly when respiration is blocked. Glycerol production is then enhanced for redox balancing, but it cannot leave the cells, resulting in an osmotic imbalance and poor growth (Tamás et al., 1999). The sensitivity of the fps1Δ strain to arsenate is consistent with the fact that pyruvate dehydrogenase is the central target for arsenate and that inhibition of this enzyme blocks respiration (Szinicz and Forth, 1988).

We have demonstrated by transport assays that As(III) is a substrate for Fps1p. The uptake of [74As]-As(III) was higher for wild-type cells than for the fps1Δ mutant, whereas cells with the open Fps1p channel exhibited a very high level of [74As]-As(III) uptake. Furthermore, the presence of As(III) inhibited Fps1p-mediated glycerol uptake. Hence, the transport data are in agreement with the observed phenotypes.

We could not measure Sb(III) uptake directly, as no Sb(III) isotope is commercially available. We also could not detect any inhibition of Fps1p-mediated glycerol transport by potassium antimonyl tartrate or antimony(III) chloride. This could result from the poor solubility of antimony salts at the concentration that may be required for the inhibition of glycerol transport. As the different FPS1 mutants displayed similar phenotypes when As(III) or Sb(III) was present in the growth medium, we also assume that Sb(III) is transported by Fps1p.

Although Fps1p appears to be the main entrance for As(III) and Sb(III) into yeast cells, additional routes of entry are likely to exist. Strains lacking the two transporters Acr3p and Ycf1p were hypersensitive to As(III) and Sb(III) (Wysocki et al., 1997; Ghosh et al., 1999; this work). As neither deletion of FPS1 nor hyperosmotic stress could fully suppress the sensitivity of the ycf1Δ and acr3Δ mutants, alternative entry routes must exist. Fps1p is a member of the ubiquitous MIP family of water and solute transporters found in bacteria, fungi, plants and animals. The MIP proteins have a widespread role in osmoregulation and metabolism, and several clinical disorders have been ascribed to defective MIP channels (Borgnia et al., 1999; Kjellbom et al., 1999;Hohmann et al., 2000). The S. cerevisiae genome contains four genes encoding MIP channels: FPS1 encodes a glycerol channel protein; AQY1 and AQY2 encode aquaporin water channels; and YFL054c encodes a protein with homology to other MIP glycerol channels (Hohmann et al., 2000). YFL054c-like genes are found in other yeasts as well as in higher eukaryotes and could constitute an additional metalloid entry pathway. However, the single yfl054cδ and the double yfl054cδ fps1Δ mutant did not affect metalloid tolerance compared with wild-type and fps1Δ cells respectively (M. J. Tamás and R. Wysocki, unpublished data). Likewise, the metalloid hypersensitivity of the ycf1Δ mutant was not suppressed by the additional deletion of YFL054c. Thus, YFL054c does not seem to be involved in metalloid uptake under the conditions tested. Interestingly, human organs such as kidney and lung contain abundant MIP water channels that are responsible for transcellular water flow (Borgnia et al., 1999). As arsenic severely affects these organs, it is tempting to speculate that human aquaporin water channels may be involved in arsenic transport.

Inactivation of another MIP channel, the E. coli glycerol facilitator GlpF, has been shown to increase bacterial tolerance to Sb(III) but not to As(III) (Sanders et al., 1997). However, GlpF did not appear to transport either As(III) or Sb(III) when present in yeast. GlpF may possibly need some bacterial modification to recognize these metalloids as substrates. Yeast also differs from bacteria in tolerance to Sb(III), as S. cerevisiae strains are about 100-fold more resistant to potassium antimonyl tartrate than E. coli strains (Table 1; Sanders et al., 1997; Wysocki et al., 1997). It is therefore possible that the transport capacity of GlpF is not sufficient to produce metalloid sensitivity in yeast.

Arsenic and antimony formally belong to the same group of the periodic table; however, they differ from one another chemically. Furthermore, one finds various As(III) and Sb(III) species in aqueous media. As(III) is mainly present as the oxo acid HAsO2 at neutral pH (Santhanam and Sundaresan, 1985), which might be the form transported by Fps1p. As yeast cells exhibited similar sensitivities to potassium antimonyl tartrate and antimony(III) chloride, it is unlikely that Fps1p mediates transport of antimonite in the form of a salt complex. In solution, the predominant form of Sb(III) at neutral pH would be Sb(OH)3 (Baes and Mesmer, 1976). Similar to what has been suggested for GlpF (Sanders et al., 1997), Fps1p might recognize Sb(OH)3 as the inorganic equivalent of the polyol glycerol.

Although the form of As(III) and Sb(III) entering yeast cells via Fps1p remains elusive, Ycf1p transports glutathione conjugates of both metalloids, whereas Acr3p probably transports the arsenite anion coupled to the membrane potential (Wysocki et al., 1997; Ghosh et al., 1999; Rosen, 1999a; Fig. 6). Interestingly, we have found that the triple fps1Δ ycf1Δ acr3Δ mutant was more sensitive to Sb(III) than the double fps1Δ ycf1Δ mutant. In addition, expression of the ACR3 gene appears to be induced by both As(III) and Sb(III) (Bobrowicz and Ulaszewski, 1998). This could also point to a role for Acr3p in Sb(III) extrusion. No eukaryotic Sb(III) efflux system has been identified to date.

Figure 6.

Mechanisms involved in arsenic and antimony transport and tolerance in S. cerevisiae. Pentavalent arsenate enters cells via phosphate transporters such as Pho84p (Bun-ya et al., 1996), whereas the uptake of trivalent arsenite and antimonite is mediated by the glycerol channel Fps1p (this work). Inhibition of any of these entry routes results in increased metalloid tolerance. In the cytosol, arsenate is reduced to arsenite by the arsenate reductase Acr2p (Mukhopadhyay et al., 2000) and immediately extruded from the cell by Acr3p (Wysocki et al., 1997; Ghosh et al., 1999). Arsenite and antimonite tolerance also involves the ABC transporter Ycf1p, which mediates sequestration of the glutathione conjugates As(GS)3 and Sb(GS)3 into the vacuole (Ghosh et al., 1999).

The mechanism of tolerance to arsenic and antimony in both prokaryotic and eukaryotic organisms has generally been linked to active extrusion or vacuolar sequestration. Here, we have shown that reducing the influx of toxic substances provides an additional level of control. Reduction in metalloid influx may be a mechanism by which yeast cells respond to metalloids in nature. The presence of arsenite and antimonite in the growth medium resulted in a rapid and sustained repression of the FPS1 gene. Hence, yeast cells apparently diminish the uptake of toxic metalloids by reducing transcription of the gene encoding the uptake system.

In conclusion, this work has identified the glycerol channel Fps1p as the entry route for As(III) and Sb(III) in yeast. To our knowledge, this is the first report describing a eukaryotic uptake mechanism for these metalloids. We have also shown that this entry pathway is regulated, as yeast cells reduced expression of the uptake system in the presence of metalloids, presumably to increase tolerance.

Experimental procedures

Strains, plasmids and growth conditions

The S. cerevisiae strains and the plasmids used in this study are described in Table 2. The E. coli strain DH5α[supE44 ΔlacU169 (φ80 lacZ ΔM15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1] was used for plasmid selection and propagation.

Table 2.   Yeast strains and plasmids used in this study.
 W303-1A MAT a ura3-1 leu2-3/112 trp1-1 his3-11/15 ade2-1 can1-100 GAL SUC2 mal0 Thomas and Rothstein (1989)
 YSH294W303-1A fps1Δ::LEU2 Van Aelst et al. (1991)
 YSH642W303-1A gpd1Δ::TRP1 gpd2Δ::URA3 Ansell et al. (1997)
 IRW100W303-1A ycf1Δ::loxP-kanMX-loxPThis study
 IRW101W303-1A fps1Δ::LEU2 ycf1Δ::loxP-kanMX-loxPThis study
 RW104W303-1A acr3Δ::loxP-kanMX-loxPThis study
 RW105W303-1A ycf1Δ::loxP acr3Δ::loxP-kanMX-loxPThis study
 RW106W303-1A fps1Δ::LEU2 acr3Δ::loxP-kanMX-loxPThis study
 RW107W303-1A fps1Δ::LEU2 ycf1Δ::loxP acr3Δ::loxP-kanMX-loxPThis study
 YMT101W303-1A fps1Δ::LEU2 YIp-URA3-fps1-Δ1 Tamás et al. (1999)
 pUG6Vector containing the loxP–kanMX–loxP deletion cassette Güldener et al. (1996)
 YEplac1952 µS. cerevisiae–E. coli shuttle vector, AmpR, URA3 Gietz and Sugino (1988)
 YEpFPS1YEplac195 containing the FPS1 gene Van Aelst et al. (1991)
 YEpfps1-Δ1YEplac195 with the fps1-Δ1 allele encoding the constitutively open Fps1p channel Tamás et al. (1999)
 YEpGlpF E. coli glpF in YEplac195 behind the yeast PGK1 promoter Luyten et al. (1995)

Yeast strains were routinely cultivated on complete YPAD medium (1% yeast extract, 2% peptone, 2% glucose, 0.004% adenine sulphate) or on minimal YNB medium (0.67% yeast nitrogen base) supplemented with auxotrophic requirements and with either 2% glucose or 2% glycerol as a carbon source (Kaiser et al., 1994). To test the growth of various yeast strains in the presence of metalloids, sodium arsenite (Sigma), sodium arsenate (Sigma), potassium antimonyl tartrate (Acros), antimony(III) chloride (Aldrich) and other metal salts were added to minimal YNB medium. E. coli cells were grown in LB medium supplemented with 100 µg ml−1 ampicillin when necessary (Sambrook et al., 1989).

Plasmid purification and E. coli transformation was performed according to the method of Sambrook et al. (1989). S. cerevisiae cells were transformed with plasmids by a modified lithium acetate method using PLATE mixture (Kaiser et al., 1994). Total yeast DNA was isolated according to the method of Rose et al. (1990).

Disruption of the YCF1 and ACR3 genes

Disruption of the YCF1 and ACR3 genes was achieved using a short flanking homology method (Wach et al., 1998). The disruption cassettes containing the loxP–kanMX–loxP marker flanked by short homology regions to the target gene were prepared by polymerase chain reaction (PCR) using the pUG6 plasmid as a template (Güldener et al., 1996) with the following primers: ACR3KAN (5′-CGTTGTAATTCAAGAGAACCCAACCAACAAATCATCAGGT
AGCATAGGCCACTAGTGGATCTG-3′) for YCF1. The cycling conditions were as follows: hot start for 10 min; 5 min denaturation at 95°C; three cycles of 45 s at 95°C, 45 s at 45°C, 2 min at 72°C; 25 cycles of 45 s at 95°C, 45 s at 60°C, 2 min at 72°C; 7 min final elongation at 72°C. The PCR products were purified and transformed into yeast according to the method of Gietz and Woods (1998) with the following modifications: after heat shock, the cells were washed with water, cultivated in 1 ml of liquid YPAD for 3 h and spread on a YPAD plate containing 200 mg l−1 geneticin (Gibco BRL). After incubating the cells for 2 days at 30°C, the lawn of transformants was replica plated onto a new YPAD–geneticin plate. After another 3 days at 30°C, only large colonies were isolated from the background and streaked out for single colonies on a YPAD–geneticin plate. The correct replacement of the YCF1 and ACR3 genes with the loxP–kanMX–loxP cassette was verified by amplifying novel DNA junctions as described previously (Pearson et al., 1998).

Metalloid sensitivity assays

Yeast cells were grown overnight in liquid YNB medium. The OD600 of the cultures was adjusted to about 0.3, and 10-fold serial dilutions of the cell suspensions were prepared. Each dilution (3 µl) was spotted on YNB agar plates containing various concentrations of metalloid salts. The growth was monitored for 2–3 days at 30°C.

Transport assays

Yeast cells were grown in liquid YNB medium at 30°C to an OD600 of about 2.0. The cells were harvested, washed and suspended in ice-cold MES buffer (10 mM MES, pH 6.0) to a density of 40–60 mg of cells ml−1. Arsenite uptake was measured by adding a mixture of ‘cold’ sodium arsenite (NaAsO2) and the [74As]-As(III) isotope to give a final concentration of 10 µM As(III). Samples were withdrawn at the indicated time points, filtered, washed in water and the radioactivity was counted in a 5 × 5 inch NaI (Tl) well-type scintillation detector (Canberra). The radioactivity retained on the filter was taken as a measure of the arsenite taken up by the cells. Radioactive 74As-labelled arsenite was prepared as described by De Kimpe et al. (1993).

To determine the inhibition of glycerol transport, cells were prepared as above. Glycerol uptake (100 mM ‘cold’ glycerol plus 40 µM [14C]-glycerol; 160 mCi mmol−1; Amersham Pharmacia) was measured with wild-type cells or with cells expressing the fps1-Δ1 allele according to the method of Tamás et al. (1999). Sodium arsenite, potassium antimonyl tartrate or antimony(III) chloride (various concentrations from 0.1 mM to 100 mM) was added to the cell suspension 10 min before starting the glycerol uptake measurement. All transport assays were repeated at least three times to confirm the trends observed. Representative data are shown.

Northern analysis

Sodium arsenite (1 mM) or potassium antimonyl tartrate (1 mM) was added to exponentially growing cells on YNB medium. Total RNA was isolated at the time points indicated in Fig. 5 and separated by electrophoresis as described previously (de Winde et al., 1996). Blots were hybridized with 32P-labelled fragments of FPS1 and 18S rRNA in buffer containing 7% SDS, 0.5 M sodium phosphate buffer, pH 7.0, and 1 mM EDTA. The signal was quantified using a phosphorimager (Fuji, BAS-1000).


We thank L. Mees (Gent) for preparing the arsenite isotope, as well as S. Hohmann and R. Bill (Göteborg) for suggestions and critical reading of the manuscript. R.W. is supported by a young scientist fellowship from the Foundation of Polish Science, C.C.C. is Research Assistant of the Fund for Scientific Research, Flanders (FWO), M.V.H. is supported by the Flemish Institute for the Promotion of Scientific–Technological Research in Industry (IWT), and M.J.T. was supported by a postdoctoral fellowship (PDM) from the Katholieke Universiteit Leuven.