Biochimica et Biophysica Acta 1833 (2013) 997–1005

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bbamcr

Yeast protective response to arsenate involves the repression of thehigh affinity iron uptake system

Liliana Batista-Nascimento a, Michel B. Toledano b, Dennis J. Thiele c, Claudina Rodrigues-Pousada a,⁎a Instituto de Tecnologia Química e Biológica, Universidade Nova de Lisboa, Av. da República, EAN, 2781-901 Oeiras, Portugalb IBITECS, LSOC, CEA-Saclay, 91191 Gif-sur-Yvette, Francec Duke University School of Medicine, Durham, NC 27710, USA

⁎ Corresponding author. Tel.: +351 214469624; fax:E-mail address: claudina@itqb.unl.pt (C. Rodrigues-P

0167-4889/$ – see front matter © 2013 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.bbamcr.2012.12.018

a b s t r a c t

a r t i c l e i n f o

Article history:Received 3 November 2012Received in revised form 19 December 2012Accepted 23 December 2012Available online 4 January 2013

Keywords:Arsenic uptakeFet3–Ftr1Iron deficiency

Arsenic is a double-edge sword. On the one hand it is powerful carcinogen and on the other it is used ther-apeutically to treat acute promyelocytic leukemia. Here we report that arsenic activates the iron responsivetranscription factor, Aft1, as a consequence of a defective high-affinity iron uptake mediated by Fet3 and Ftr1,whose mRNAs are drastically decreased upon arsenic exposure. Moreover, arsenic causes the internalizationand degradation of Fet3. Most importantly, fet3ftr1 mutant exhibits increased arsenic resistance and de-creased arsenic accumulation over the wild-type suggesting that Fet3 plays a role in arsenic toxicity. Finallywe provide data suggesting that arsenic also disrupts iron uptake in mammals and the link between Fet3, ar-senic and iron, can be relevant to clinical applications.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

Arsenic (As), the 20th most abundant element in the earth's crustis a highly toxic metalloid with respect to human health [1]. Althoughoften synonymous to a poison, it is one of the oldest drugs in historyof mankind, first used to treat cutaneous ulcers, later periodic feverand malaria, and currently as treatment for acute promyelocytic leu-kemia (PML) [2]. Due to its paradoxical biological effects, As hasbeen the object of intense scrutiny, with regards to its cellular me-tabolism and molecular biological impact. However to date, thesestudies have never established a link between As and iron (Fe)homeostasis.

Cells require Fe for a wide array of metabolic functions, which in-clude oxygen transport, cellular respiration, lipidmetabolism, gene reg-ulation and DNA replication and repair, yet Fe is toxic when present inexcess [3]. Saccharomyces cerevisiae expresses three distinct transportpathways for Fe, two reductive systems and a non-reductive one. Thereductive pathways consist of low- and high-affinity uptake systems op-erated by Fet4, and by a protein complex composed of the multicopperferroxidase Fet3 and the permease Ftr1, respectively. The Fet3–Ftr1complex is specific for Fe and is regulated both transcriptionally andpost-transcriptionally by Fe [4–6]. The non-reductive Fe uptake pathwayis mediated by the ARN family (Arn1-4) of membrane permeases thattransport siderophore-ferric iron complexes [7]. In S. cerevisiae, theseFe uptake systems are all induced under Fe scarce conditions by the

+351 214469625.ousada).

rights reserved.

Aft1 transcription factor and its homologue Aft2, as part of the Feregulon [8].

We have here conducted a genome-wide mRNA profiling of theS. cerevisiae response to arsenate (AsV), which revealed a potent induc-tion of the Fe gene regulon. However, neither FET3 nor FTR1 appearedinduced by this treatment. Strikingly, AsV caused the immediate inter-nalization of Fet3 in the endoplasmic reticulum (ER) and the apparentlack of FET3 transcriptional induction was in fact the consequence ofthe degradation of itsmRNAby themajor pathway formRNAdecayme-diated by the 5′-3′ exonuclease Xrn1. Moreover, increased AsV toler-ance in strains with null mutations of either FET3 or FTR1, togetherwith a decrease AsV cellular uptake in these mutant strains, indicate arole of Fet3 and Ftr1 in mediating AsV cellular uptake and toxicity.Taken together this work provides a molecular connection betweenAs toxicity and Fe homeostasis, which should be relevant to furtherunderstand the toxic and therapeutic effects of As at the molecularlevel.

2. Materials and methods

2.1. Yeast strains and growth conditions

The plasmids and yeast strains used in this study are listed in supple-mentary Table S3. Spot assays were carried out by spotting 5 μl of earlyexponential phase cultures (A600=0.4) sequentially diluted (approxi-mately 5×103 to 10 cells) in medium containing 2 mM Na2HAsO4

(AsV), 100 μM Fe2(SO4)3 (Fe3+) or 50 μM BPS. Growth was recordedafter incubation for 2 days at 30 °C.

998 L. Batista-Nascimento et al. / Biochimica et Biophysica Acta 1833 (2013) 997–1005

2.2. DNA microarray analysis

Wild-type (WT) cells were grown in triplicate in media containing2 mM of AsV for 1 hour, and RNAwas extracted, labeled, and hybridizedto Affymetrix Yeast Genome S98 arrays. (For further information visit theDuke Microarray Core Facility at http://www.genome.duke.edu/cores/microarray/). All data were analyzed using both Partek® GenomicsSuiteTM and dChip softwares. Data have been deposited in the NCBIGene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) and areaccessible through GEO Series accession number GSE33427.

2.3. RNA blot analysis

Total yeast RNA was isolated from cells untreated or treatedwith either 2 mM of AsV, 100 μM of Fe2(SO4)3 (Fe3+), 100 μM ofBPS or 400 μM H2O2. PCR-amplified fragments were radiolabeledwith 32P-dCTP to be used as probes. U3, a small nuclear RNA(SNR17A), was used as loading control.

2.4. Real-time PCR analysis

RNA was extracted from early log-phase cultures that were eitheruntreated or exposed during 60 min to 2 mM AsV. DNA was removedby on-column DNAse I digestion (RNase-Free DNase Set; Qiagen).Total RNA (1ìg) was reverse transcribed with Transcriptor ReverseTranscriptase (Roche Diagnostics). qPCR reactions were performedin the LightCycler 480 Instrument (Roche), using LightCycler 480Green I Master (Roche) and the oligonucleotides are listed in supple-mentary Table S4. Actin (ACT1) was used as a reference gene. All as-says were made in triplicate.

2.5. β-Galactosidase assay

WT cells and the isogenic aft1 mutant were transformed with aβ-galactosidase reporter plasmid containing the promoter sequenceof CTH2 previously described [9], pCM64-CTH2-FeRE-CYC1-LacZ. WTwas also co-transformed with the plasmids pEG202LexA-AFT1 andpSH18-34. β-Galactosidase was measured in triplicate following thedegradation of the colorimetric substrate ONPG (o-nitrophenyl-b-D-galactopyraniside) at A420 and normalized against total proteinconcentration.

2.6. Measurement of total iron and arsenic

Strains were grown in YPD media with 1 mM of AsV for 2 or4 hours, collected by centrifugation and washed with 10 mM EDTAand metal-free water. The total Fe and As were measured by induc-tively coupled plasma (ICP) atomic emission spectroscopy.

2.7. Protein analysis

Protein extracts were generated from cell cultures using cell lysisbuffer supplemented with protease inhibitors (Roche). Proteinwas re-solved by SDS–PAGE, and immunoblotted with Anti-HA and Anti-cMyc(Roche) Anti-TAP-tag (Life Technologies). Anti-Pgk1 (Life Technologies)was used as loading control.

2.8. Fluorescence microscopy

Microscopy experiments were carried out on live cultures usingLEICADMRA2Microscope coupledwith a CoolSNAPTMHQPhotometricscamera (Roper Scientific). The analysis of fluorescence intensity wasdone using theMetaMorph software package (MDS Analytical Technol-ogies). Overnight liquid cultures expressing Aft1-GFP, Fet3-GFP orSec63-GFP were re-inoculated to an optical density A600=0.1 in YPDmedium containing 1 mM of AsV, 100 μMof BPS (Bathophenanthroline

disulfonate), 100 μM BSC (Bathocuproine disulphonate) or 10 μMCu2SO4. After washing with phosphate-buffered saline (PBS) cellswere resuspended in DABCO solution and visualized.

2.9. Cells and media

Mousse embryonic fibroblasts (MEF) cells and HeLa cells were Feloaded by addition of Ferric ammonium citrate (Fe3+, 100 μM) andFe depleted by addition of desferroxamine 5 μM and treated with ar-senic trioxide (AsIII, 5 μM) for 24 hours. To extract protein, cells weresolubilized in 1.0% of Triton X-100, 150 mM NaCl, 10 mM EDTA, and10 mM Tris (pH7.4) with a protease inhibitor cocktail (Roche). Pro-tein was resolved by 12% SDS–PAGE and transferred to a nitrocellu-lose membrane (GE Healthcare). Ferritin levels were detected usingMouse Anti-Ferritin (Abcam), Ccs1 levels using Rabbit Anti-Ccs1(Santa Cruz Biotechnology, Inc), and Rat Anti-Tubulin (Novus Biolog-icals) was used as a loading control.

3. Results

3.1. The Saccharomyces cerevisiae genome-wide response to arsenicexposure

In S. cerevisiae, resistance to AsV has been ascribed to Yap1 andYap8, two AP-1 like transcription factors that regulate the expressionof genes involved in redox homeostasis and As detoxification process-es, respectively [10]. To identify novel As tolerance pathways, we an-alyzed the genome-wide mRNAs profile of yeast cells exposed to AsV.Fig. 1 summarizes the results of these experiments (see data depositin http://www.ncbi.nlm.nih.gov/geo/). In total, 1608 genes had theirexpression significantly altered upon AsV exposure. Classification ofthese genes by Gene Ontology showed that they belonged to manyfunctional categories, including the response to metals and oxidativestress, cell cycle, mRNA processing, mitochondrial related-processes,transcriptional and translational activity, transporter activity, endo-plasmic reticulum related-processes, protein folding, among others(Fig. 1A). Many of these genes have already been shown to respondto As [11,12]. However expression of genes responsive to metals,such as, zinc, iron (Fe) and copper (Supplementary Fig. S1A) hasnever been shown to be altered by As. Within the cellular Fe homeo-stasis category, 82 genes showed changes of at least 1.2-fold in theirexpression levels in response to AsV stress (Fig. 1B and Supplementa-ry Table S1). Among these, many of the genes of the Aft1/2 regulon[13–15] were present, such as CTH2, the ARN family genes and FIT3that encode cell wall mannoproteins involved in siderophore-Fe3+ up-take (Fig. 1C and Supplementary Table S2). Surprisingly, despite a largeinduction of the Aft1/2 regulon, FET3 and FTR1 mRNA levels appearedsignificantly decreased upon AsV exposure (Fig. 1C and SupplementaryTable S1). Moreover the samewas observed for FET3mRNA levels uponarsenite (AsIII) exposure (Supplementary Fig. S1B).

3.2. Arsenic activates Aft1 and causes Fe scarcity

Except for FET3 and FTR1, the mRNA profiling analysis indicatedthat AsV exposure mimics the genomic response to Fe deficiency.Yeast cells respond to Fe deprivation by the Aft1/2-dependent tran-scriptional activation of the Fe regulon that encodes proteins involvedin Fe uptake, trafficking and utilization [16]. Aft1 is controlled at thelevel of its nucleocytoplasmic distribution. To confirm that AsV canactivate Aft1/2, we monitored by fluorescence microscopy the kinet-ics cellular localization of an Aft1-GFP fusion in cells exposed to AsV(Fig. 2A). After 40 minutes of exposure to AsV, Aft1-GFP fluorescencewas indeed shifted from a diffuse to a nuclear pattern, similar to whatis seen in cells treated with the iron chelator BPS. Subcellular localiza-tion of a Myc-Aft1 fusion by cellular fractionation similarly showedthat in AsV-treated lysates, Aft1 was almost exclusively seen in the

Fig. 1.Genome-wide response of Saccharomyces cerevisiae upon arsenate exposure. (A) Representation of the results from triplicate Affymetrix DNAmicroarray studies showing the GeneOntology of the main categories with at least 50% of the mRNAs changed. (B) Representation of the total number of mRNAs corresponding to Fe-dependent processes. (C) Microarrayvalidation by RT-PCR.

999L. Batista-Nascimento et al. / Biochimica et Biophysica Acta 1833 (2013) 997–1005

nucleus, as also seen with BPS, whereas in the absence of treatment itwas evenly distributed between the cytosolic and nuclear fraction(Fig. 2B). AsV also induced the expression of the Aft1-dependentgene CTH2 to levels similar to those triggered by BPS, as monitoredusing a promoter-LacZ fusion (Fig. 2C). In the aft1 mutant, AsV-induced CTH2 expression was abolished. To question the physiologi-cal link between As and the Aft1/2 regulon, we next evaluated As tol-erance of aft1 and aft2 single mutants by growth in the presence ofAsV (2 mM). The aft1 mutant was unable to grow under this condi-tion, while the aft2 mutant displayed a normal growth (Fig. 2D). In-terestingly AsV-dependent growth inhibition of aft1 mutant couldbe totally reversed by addition of Fe3+ (Fig. 2D), which indicatesthat AsV not only activates the Aft1/2-dependent Fe-deprivation re-sponse, but also causes Fe scarcity.

3.3. Arsenic triggers degradation of FET3mRNAby the 5′-3′ exonuclease Xrn1

As shown above, AsV potently induced the genes of the Fe regulon,with the notable exception of FET3 and FTR1 that were paradoxicallydown-regulated (Fig. 1C, and Supplementary Table S2). Fet3 is themulticopper oxidase that oxidizes ferrous (Fe2+) to ferric iron (Fe3+),for subsequent cellular uptake by the transmembrane permease Ftr1

[17,18]. Aft1 activates the Cth1 and Cth2 mRNA-binding proteins thatpromote decay of the Fe binding proteins-encoded mRNAs containingAREs in their 3′ UTR [3]. However, similar to AsV-treated WT cells, theFET3 mRNAs remained undetectable in a cth2 mutant exposed to AsV(Fig. 3A), consistent with the notion that FET3 does not carry anyCth1/2 AREs in its 3′ UTR.

A major route for mRNA degradation in eukaryotic cells begins withpoly-A deadenylation by Ccr4, followed by decapping by Dcp1/2, and5′-3′ exonucleotidic decay by Xrn1 [19,20]. To test the possible role ofthis pathway in mediating AsV-dependent FET3 mRNA decay we usedccr4 and xrn1 mutants. Importantly, the xrn1 mutant treated with AsVexpressed FET3 at levels similar to those of untreated wild-type cells(Fig. 3B), indicating that AsV triggers degradation of this messengersby the 5′-3′ exonuclease Xrn1 pathway. Because oxidative stress hasbeen associated with changes in the mRNA stability of the Aft1regulon [21], and because As compounds induce oxidative stress,we also analyzed FET3 expression in the presence of H2O2. Asshown in Fig. 3C we did not observe any decrease of the FET3mRNA upon addition of H2O2, suggesting that AsV-mediated FET3degradation is independent of ROS formation. Taken together,these results demonstrate that As triggers FET3 mRNA degradationby Xrn1.

Fig. 2. Aft1 activation in response to arsenate stress. (A) Time-course representation of AFT1-GFP fusion expressed diffusely throughout the cytoplasm while in presence of arsenate(AsV) after 40 minutes localizes to the nucleus. (B) Nuclear fractionation of wild-type cells expressing Aft1-Myc. In presence of arsenate (AsV) Aft1 is mainly found in the nuclearfraction. Cytosolic fraction was identified by incubation with an antibody against Pgk1. (C) Arsenate (AsV) induces the expression of CTH2-lacZ reporter construct. (D) aft1mutant issensitive to arsenate (AsV) and recovers upon supplementation with iron (Fe3+).

1000 L. Batista-Nascimento et al. / Biochimica et Biophysica Acta 1833 (2013) 997–1005

Fig. 3. Exonuclease Xrn1 mediates FET3 mRNA degradation upon arsenate addition. (A) FET3 mRNA down-regulation is Cth2 independent. (B) Upon arsenate (AsV) addition FET3expression levels remain stable only in the absence of XRN1. (C) FET3 mRNA degradation is independent of the ROS formation.

1001L. Batista-Nascimento et al. / Biochimica et Biophysica Acta 1833 (2013) 997–1005

3.4. Arsenic impairs high affinity Fe uptake by decreasing Fet3 proteinlevels and plasma membrane expression

To explore the consequences of the decrease of FET3 mRNA levels,we first checked Fet3 protein levels. AsV caused a decrease in theselevels that started 4 hours after drug exposure; for comparison, thedecrease and increase in Fet3 levels observed under Fe repletion(Fe3+) and depletion (BPS) are shown (Fig. 4A). Felice et al. reportedthat high Fe levels cause not only the repression of the Aft1 Feregulon, but also the internalization and degradation of the Fet3/Ftr1 transport system [6]. Thus as another way of checking Fet3 ex-pression, we examined its cellular localization (Fig. 4B). In untreatedcells and in cells treated with copper, Fet3 was largely localized at theplasma membrane. However, after 5 minutes exposure to AsV, Fet3was essentially detected in an internal compartment resembling theendoplasmic reticulum (ER), as indicated by the localization of theER marker SEC63. As expected, Fet3 was similarly localized incopper-depleted cells (BCS) [22]. After 60 minutes exposure to AsV,Fet3 was still in the ER. To evaluate the functional consequence ofthese changes in Fet3p expression, we examined the effect of AsVon the Fe total cellular content. A 4 hours treatment with AsV at1 mM caused a 70% decrease on the Fe cellular content (Fig. 4C). Toevaluate whether the decrease in Fe was due to impairment of theFet3–Ftr1 Fe reductive pathway, we checked whether addition of var-ious siderophores to the medium could restore the cellular Fe contentof AsV-treated cells by recruitment of the non-reductive pathway ofthe Arn proteins family. Addition of a cocktail of siderophores indeedfully restored the Fe cellular content of AsV-treated cells, and this ef-fect was dependent upon the presence of Aft1 (Fig. 4C).

By decreasing Fet3 expression, AsV thus causes defective high affin-ity Fe uptake and eventually Fe deficiency. Further AsV toxicity appears,at least in part, due to Fe deficiency, as suggest the improvement of AsVtolerance by exogenous Fe3+ (see Fig. 2D). We checked whetherAsV-induced Aft1 activation was also the consequence of AsV-inducedFe deficiency. To this end, we compared the time-dependent expressionof CTH2 and FET3 during 1 hour after As exposure (Fig. 4D). Interesting-ly, FET3 mRNA levels started to decrease 10 minutes after drug expo-sure and became undetectable at 1 hour, whereas CTH2 mRNA levelsstarted to increase at 50 min, long after FET3 mrRNA disappearance.These data indicate that AsV-induced Aft1 activation is a consequenceof AsV-induced defective high-affinity Fe uptake and Fe deficiency.

3.5. The Fet3–Ftr1 high affinity Fe uptake system contributes to arsenatecellular uptake

To test the possible relevance of Fet3 in mediating As biological ef-fects, we assayed the As tolerance of the fet3ftr1 double mutant.

Surprisingly, this mutant was significantly more resistant to AsVthan its WT counterpart, indicating that Fet3 and Ftr1 are requiredto mediate As toxicity. In keeping with this idea, the xrn1 mutationthat prevents AsV-induced FET3 mRNA degradation and hence Fet3levels down-modulation, is more sensitive to AsV than the WT strain(Fig. 5A). fet3ftr1 was also strongly protected from AsV-inducedgrowth inhibition in liquid medium and from AsV-induced celldeath (~20% vs. 90 % at 1 mM AsV), (Fig. 5B and C). To understandthe basis of the As resistance of the fet3ftr1mutant, we comparativelymeasured the As content in these and in wild-type cells treated with1 mM of AsV for 2 hours. Fig. 5D shows that As levels in the fet3ftr1mutant were only 30% of those measured in wild-type cells. These re-sults strongly suggest that the Fet3–Ftr1 high affinity Fe uptake sys-tem contributes to AsV cellular uptake.

3.6. Arsenic trioxide destabilizes Fe uptake in mammalian cells

Arsenic trioxide (AsIII) is a small molecule used in the treatment ofPML. Although this compoundwas shown to cause damage to/or degra-dation of the PML protein/retinoic acid receptor-alpha (PML/RARα) fu-sion protein, thereby inducing apoptosis of acute PML cells [23], itsmechanism of action is not fully understood. To extend the resultsobtained in yeast, we checked whether AsIII could also perturb Fe up-take in mammalian cells, using mouse embryo fibroblasts (MEFs) andHeLa cells. In mammals, ferritin synthesis is post-transcriptionally reg-ulated by the Iron Regulatory Protein (IRP) cytosolic Fe availability. InFe deplete cells, IRP binds the ferritin mRNA 5′ Iron Response Element(IRE) preventing translation, whereas in Fe replete cells IRP no longerbinds the 5′ ferritin mRNA IRE, thereby allowing translation, [24]. Asshown in Fig. 6 expression of ferritin was indeed reduced under lowFe conditions, and increased in the presence of Fe3+. Interestingly thepresence of AsIII significantly prevented the Fe3+-induced increase inferritin protein levels. Moreover, this effect of AsIII was specific for Fesince levels of the copper chaperone for superoxide dismutase Ccs1,which are inversely correlated with the intracellular copper (Cu) con-centration [25], were unaffected by AsIII. This result indicates that asin yeast, As interferes with Fe metabolism, possibly by preventing Feuptake.

4. Discussion

Several mechanisms have been proposed to explain As toxicity andcarcinogenicity [26] however, a link between As and Fe homeostasishas so far never been described. We provide here evidence that in thesingle-celled eukaryote S. cerevisiae, the high affinity Fe uptake systemFet3–Ftr1 is a major determinant of As toxicity by contributing to AsVcellular uptake. We also show that yeast cells respond to AsV by

1002 L. Batista-Nascimento et al. / Biochimica et Biophysica Acta 1833 (2013) 997–1005

Fig. 5. fet3ftr1 mutant is tolerant to arsenate exposure. (A) Sensitivity of fet3ftr1, xrn1 mutants and the respective wild-type cells to 1 mM AsV. (B) and (C) Arsenic cell survival andgrowth curves, respectively. Data are expressed as percentages of colonies formed compared to the control cultures.

1003L. Batista-Nascimento et al. / Biochimica et Biophysica Acta 1833 (2013) 997–1005

decreasing Fet3 and Ftr1 expression through degradation of their tran-scripts, thereby preventing AsV uptake, at the price of becoming Fedeficient.

While analyzing the S. cerevisiae genome-wide response to AsV byDNA microarrays, we observed that the genes of the Fe regulon consti-tute an important component of the AsV genomic response, but para-doxically the FET3 and FTR1 transcripts were very low in AsV-treatedcells (Fig. 1C). As the transcription factor Aft1, which in response to Fedeprivation induces expression ofmany genes of the Fe regulon, includ-ing FET3 and FTR1, is potently activated in AsV-treated cells (Fig. 2), wesearched for a post-transcriptional regulatory mechanism that couldexplain the decrease of the later transcripts. We found that in AsV-treated cells the FET3 and FTR1 transcripts were degraded by the 5′-3′exonuclease Xrn1 (Fig. 3B), thereby causing a significant decrease in

Fig. 4. Arsenate exposure reduces Fet3 protein levels and causes mislocalization at the endoment. (B) Representation of FET3-GFP strain expressed diffusely throughout the plasmamemER. SEC63-GFP was used as a ER-marker and addition of Bathocuproine disulphonate (BSC)(AsV) exposure yeast cells redirect Fe internalization through the non-reductive transportupon exposure of arsenate (AsV) 1 mM and addition of siderophores 5 μM (ferrichrome,100 in untreated conditions. Bars represent the average plus standard deviation of three inof arsenate (AsV) 2 mM by RT-PCR. The assay was made using biological and technical trip

the levels of the proteins that they encode (Fig. 4A). We also observedthat a GFP-Fet3 fusion protein, which was mainly seen at the plasmamembrane in untreated cells, appeared predominantly at the ER inAsV-treated cells (Fig. 4B), further indicating a decrease in Fet3 cellularactivity. Fet3, that oxidizes Fe2+ to Fe3+ and Ftr1, constitute the yeasthigh affinity Fe uptake system. Decrease in the activity of either orboth proteins should therefore cause Fe deficiency, which is what weobserved in AsV-treated cells, based on the decrease of total cellularFe, on the lower AsV tolerance of cells lacking Aft1, and on the signifi-cant improvement of AsV tolerance by exogenous Fe3+ in both wildtype and aft1 cells (Fig. 2D). Additional proof that the observed cellularFe deficiency of AsV-treated cells is the consequence of a defect of thehigh affinity Fe uptake systemwas obtained by the correction of this de-ficiency by addition of siderophores,which turn on an alternate yeast Fe

plasmic reticulum (ER). (A) Fet3 levels decrease after 4 hours of arsenate (AsV) treat-brane while in presence of arsenate (AsV) 1 mM after 5 minutes localizes mainly to the100 μM or Copper (Cu2+) 10 μM were used as control conditions. (C) Upon arsenatesystem. The intracellular Fe content levels determined by ICP in the wild-type strainferrirubin, enterobactin and desferroxamin) for 4 hours. Values are made relative todependent experiments. (D) Kinetic study of CTH2 and FET3 expression upon additionlicates.

Fig. 6. Arsenic trioxide destabilizes Fe uptake in mammals. (A) MEF and HeLa cellswere incubated with an Fe chelator, 5 μM desferroxamine (−Fe), 5 μM arsenic trioxide(AsIII), and 100 μM ferric ammonium citrate (Fe3+). After 24 hours, media were col-lected and cells were lysed. Ferritin and Ccs1 levels were detected by western blotusing mouse anti-Ferritin and rabbit anti-Ccs1 antibodies. Anti-tubulin was used asloading control.

1004 L. Batista-Nascimento et al. / Biochimica et Biophysica Acta 1833 (2013) 997–1005

uptake pathway (Fig. 4C). The activation of Aft1 observed in AsV-treated cells should therefore result from AsV-induced Fe deficiency.The kinetics of FET3 and CTH2 expression showing that the CTH2 tran-script started to increase only after the FET3 message had alreadybeen fully repressed support this idea (Fig. 4D).

We also show that yeast cells lacking the Fet3–Ftr1 high affinity Feuptake system are much more resistant to AsV than their wild typecounterpart, while those lacking Xrn1, and therefore unable to de-grade the corresponding transcripts are much sensitive to this chem-ical (Fig. 5A). These results together with the observation that yeastcells lacking Fet3–Ftr1 accumulate much lower amounts of As thanwild type cells (Fig. 5D) strongly suggest that additionally to the phos-phate transporters [27], the Fet3–Ftr1 system also constitute a routeof AsV uptake in S. cerevisiae. We do not yet understand how Fet3–Ftr1 complex mediates uptake of AsV. It has been however shownthat Fe in its different chemical forms has strong affinity towards inor-ganic arsenic ions [28]. Hence a possible explanation is that Fet3–Ftr1might serve as a channel to uptake arsenate bound to iron. To furtherexplore this hypothesis a structural study will be required.

In summary, our study has uncovered a surprisingly adapted yeastcell response to AsV that prevents As toxicity,wherebyAsV initially trig-gers the Xrn1-dependent degradation of the FET3 and FTR1 transcripts,which encode a major route of AsV cellular uptake, thereby preventingsubsequent As uptake. The trade off of such a protective response to Astoxicity is the occurrence of a relative Fe cellular deficiency, which inturn activate the Aft1-dependent Fe regulon as a corrective measureto the later deficiency.Whether such a protective response to As is con-served in higher eukaryotes is unknown. However, impairment by AsIIIof the Fe3+-induced increase in ferritin protein levels observed mainlyin MEF but with less extent in HeLa cell lines indicate that here also Asinterferes with Fe metabolism, possibly by inhibiting Fe cellular uptake(Fig. 6). A possible explanation for this could be related to the fact that

AsIII has the capacity to co-precipitate with Fe3+ [29], which resultsin a less soluble compound, therefore inhibiting Fe uptake. Neverthelessto further explore these observations it should be used amore appropri-ated cell model with relevance for acute PML, such as NB4 cell line(acute PML carrying the t(15;17) translocation) [30].

Further work should address the following unsolved question ofhow As triggers the Xrn1-dependent degradation of the FET3 and FTR1transcripts, whether other important transcripts are targeted by Xrn1in As-treated cells, and what is the biochemical basis by which thehigh affinity Fe uptake Fet3–Ftr1 system operate AsV cellular uptakemechanism. Finally, although the Fet3 homologue inmammals, Cerulo-plasmin, participates on the Fe export system by oxidizing Fe2+

transported by Ferroportin to Fe3+, before release into the extracellularmedium making it available for the Transferrin receptors and DMT1(divalent metal transporter 1) to import back inside the cell [31], itwould be important to address whether As uptake bymammal cells in-volves, as in yeast, the Fe system.

In conclusion, our data describe a hitherto unknown mechanism ofAsV uptake in yeast whichmight be conserved inmammals and it shouldbe useful considering that As is not only an important toxic chemical, butis also used in the treatment of PML and other human diseases.

Acknowledgments

We thank Júlia Costa from the Glycobiology Laboratory at ITQB-UNLfor the resources to perform the mammalian experiments. We wouldlike to thank Jerry Kaplan for providing the FET3-GFP strain and CalorineC. Philpott for providing arn1-4 mutant.

We are also grateful to Tom Rapoport from Harvard Medical Schoolfor providing us with Sec63-GFP plasmid. This work was supported bygrants from the Fundação para a Ciência e a Tecnologia (No. SFRH/BD/39389/2007 to L.B.-N., No. PTDC/BIAMIC/108747/2008 to CR-P) and toD.J.T. by United States National Institutes of Health grant GM41840.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.bbamcr.2012.12.018.

References

[1] B.K. Mandal, K.T. Suzuki, Arsenic round the world: a review, Talanta 58 (2002)201–235.

[2] L.K. Hillestad, Acute promyelocytic leukemia, Acta Med. Scand. 159 (1957)189–194.

[3] S. Puig, E. Askeland, D.J. Thiele, Coordinated remodeling of cellular metabolismduring iron deficiency through targeted mRNA degradation, Cell 120 (2005) 12.

[4] A. Dancis, R.D. Klausner, A.G. Hinnebusch, J.G. Barriocanal, Genetic evidence thatferric reductase is required for iron uptake in Saccharomyces cerevisiae, Mol.Cell. Biol. 10 (1990) 2294–2301.

[5] E. Georgatsou, D. Alexandraki, Two distinctly regulated genes are required for fer-ric reduction, the first step of iron uptake in Saccharomyces cerevisiae, Mol. Cell.Biol. 14 (1994) 3065–3073.

[6] M.R. Felice, I. De Domenico, L. Li, D.M. Ward, B. Bartok, G. Musci, J. Kaplan,Post-transcriptional regulation of the yeast high affinity iron transport sys-tem, J. Biol. Chem. 280 (2005) 22181–22190.

[7] C.C. Philpott, O. Protchenko, Response to iron deprivation in Saccharomycescerevisiae, Eukaryot. Cell 7 (2008) 20–27.

[8] C.W. Yun, T. Ferea, J. Rashford, O. Ardon, P.O. Brown, D. Botstein, J. Kaplan, C.C.Philpott, Desferrioxamine-mediated iron uptake in Saccharomyces cerevisiae. Evi-dence for two pathways of iron uptake, J. Biol. Chem. 275 (2000) 10709–10715.

[9] S. Puig, S.V. Vergara, D.J. Thiele, Cooperation of two mRNA-binding proteinsdrives metabolic adaptation to iron deficiency, Cell Metab. 7 (2008) 555–564.

[10] R.A. Menezes, C. Amaral, L. Batista-Nascimento, C. Santos, R.B. Ferreira, F. Devaux,E.C. Eleutherio, C. Rodrigues-Pousada, Contribution of Yap1 towards Saccharomycescerevisiae adaptation to arsenic-mediated oxidative stress, Biochem. J. 414 (2008)301–311.

[11] A.C. Haugen, R. Kelley, J.B. Collins, C.J. Tucker, C. Deng, C.A. Afshari, J.M. Brown, T.Ideker, B. Van Houten, Integrating phenotypic and expression profiles to maparsenic-response networks, Genome Biol. 5 (2004) R95.

[12] M. Thorsen, G. Lagniel, E. Kristiansson, C. Junot, O. Nerman, J. Labarre,M.J. Tamas, Quan-titative transcriptome, proteome, and sulfur metabolite profiling of the Saccharomycescerevisiae response to arsenite, Physiol. Genomics 30 (2007) 35–43.

1005L. Batista-Nascimento et al. / Biochimica et Biophysica Acta 1833 (2013) 997–1005

[13] P.L. Blaiseau, E. Lesuisse, J.M. Camadro, Aft2p, a novel iron-regulated transcriptionactivator that modulates, with Aft1p, intracellular iron use and resistance to oxi-dative stress in yeast, J. Biol. Chem. 276 (2001) 34221–34226.

[14] J.C. Rutherford, S. Jaron, D.R. Winge, Aft1p and Aft2p mediate iron-responsivegene expression in yeast through related promoter elements, J. Biol. Chem. 278(2003) 27636–27643.

[15] M. Shakoury-Elizeh, J. Tiedeman, J. Rashford, T. Ferea, J. Demeter, E. Garcia, R. Rolfes,P.O. Brown, D. Botstein, C.C. Philpott, Transcriptional remodeling in response to irondeprivation in Saccharomyces cerevisiae, Mol. Biol. Cell. 15 (2004) 1233–1243.

[16] J. Ihrig, A. Hausmann, A. Hain, N. Richter, I. Hamza, R. Lill, U. Muhlenhoff, Ironregulation through the back door: iron-dependent metabolite levels contributeto transcriptional adaptation to iron deprivation in Saccharomyces cerevisiae,Eukaryot, Cell 9 (2010) 460–471.

[17] R. Stearman, D.S. Yuan, Y. Yamaguchi-Iwai, R.D. Klausner, A. Dancis, A permease-oxidase complex involved in high-affinity iron uptake in yeast, Science 271 (1996)1552–1557.

[18] J. Kaplan, T.V. O'Halloran, Iron metabolism in eukaryotes: Mars and Venus at itagain, Science 271 (1996) 1510–1512.

[19] U. Sheth, R. Parker, Decapping and decay of messenger RNA occur in cytoplasmicprocessing bodies, Science 300 (2003) 805–808.

[20] S.F. Newbury, Control of mRNA stability in eukaryotes, Biochem. Soc. Trans. 34(2006) 30–34.

[21] L. Castells-Roca, U.Muhlenhoff, R. Lill, E. Herrero, G. Belli, The oxidative stress responsein yeast cells involves changes in the stability of Aft1 regulon mRNAs, Mol. Microbiol.81 (2011) 232–248.

[22] A. Singh, S. Severance, N. Kaur, W. Wiltsie, D.J. Kosman, Assembly, activation, andtrafficking of the Fet3p.Ftr1p high affinity iron permease complex in Saccharomycescerevisiae, J. Biol. Chem. 281 (2006) 13355–13364.

[23] H. de The, Z. Chen, Acute promyelocytic leukaemia: novel insights into the mech-anisms of cure, Nat. Rev. Cancer 10 (2010) 775–783.

[24] T.A. Rouault, The role of iron regulatory proteins in mammalian iron homeostasisand disease, Nat. Chem. Biol. 2 (2006) 406–414.

[25] E.C. West, J.R. Prohaska, Cu, Zn-superoxide dismutase is lower and copper chap-erone CCS is higher in erythrocytes of copper-deficient rats and mice, Exp. Biol.Med. (Maywood) 229 (2004) 756–764.

[26] K.T. Kitchin, S. Ahmad, Oxidative stress as a possible mode of action for arseniccarcinogenesis, Toxicol. Lett. 137 (2003) 3–13.

[27] B.P. Rosen, Families of arsenic transporters, Trends Microbiol. 7 (1999) 207–212.[28] W. Tang, Q. Li, S. Gao, J.K. Shang, Arsenic (III, V) removal from aqueous solution

by ultrafine alpha-Fe2O3 nanoparticles synthesized from solvent thermalmethod, J. Hazard. Mater. 192 (2011) 131–138.

[29] Y. Jia, D. Zhang, R. Pan, L. Xu, G.P. Demopoulos, A novel two-step coprecipitationprocess using Fe(III) and Al(III) for the removal and immobilization of arsenatefrom acidic aqueous solution, Water Res. 46 (2012) 500–508.

[30] P.J. Dilda, G.G. Perrone, A. Philp, R.B. Lock, I.W. Dawes, P.J. Hogg, Insight into theselectivity of arsenic trioxide for acute promyelocytic leukemia cells by character-izing Saccharomyces cerevisiae deletion strains that are sensitive or resistant tothe metalloid, Int. J. Biochem. Cell Biol. 40 (2008) 1016–1029.

[31] M.W. Hentze, M.U. Muckenthaler, N.C. Andrews, Balancing acts: molecular con-trol of mammalian iron metabolism, Cell 117 (2004) 285–297.