|Home | About | Journals | Submit | Contact Us | Français|
Chromate toxicity is well documented, but the underlying toxic mechanism(s) has yet to be fully elucidated. Following a Cr toxicity screen against > 6000 heterozygous yeast mutants, here we show that Cr resistance requires normal function of the cortical actin cytoskeleton. Furthermore, Cr-stressed yeast cells exhibited an increased number of actin patches, the sites of endocytosis. This was coincident with a marked stimulation of endocytosis following Cr exposure. Genetic dissection of actin nucleation from endocytosis revealed that endocytosis, specifically, was required for Cr resistance. A series of further endocytosis mutants (sac6Δ, chc1Δ, end3Δ) exhibited elevated Cr sensitivity. These mutants also showed markedly elevated cellular Cr accumulation, explaining their sensitivities. In wild-type cells, an initial endocytosis-independent phase of Cr uptake was followed by an endocytosis-dependent decline in Cr accumulation. The results indicate that actin-mediated endocytosis is required to limit Cr accumulation and toxicity. It is proposed that this involves ubiquitin-dependent endocytic inactivation of a plasma membrane Cr transporter(s). We showed that such an action was not dependent on the transporters that have been characterized to date, the sulfate (and chromate) permeases Sul1p and Sul2p.
Metals are major environmental pollutants that are implicated in a range of degenerative human conditions, including cancer and Alzheimer’s disease (Beyersmann and Hartwig, 2008; Valko et al., 2005). Prevailing theories to explain the molecular mechanism(s) of metal toxicity mostly center on the generation of reactive oxygen species (ROS). However, ROS-associated oxidative damage may be incidental rather than a cause of toxicity (Avery, 2001), and the principal cause(s) of most metals’ toxicities remains to be resolved unequivocally.
Chromium is a highly toxic metal that is widely used in leather tanning and other trades, being linked to the incidence of lung cancer among workers in these industries (Luippold et al., 2003). Cr exists predominantly in two valence states, Cr(VI) and Cr(III). Owing to its relatively rapid transport across biological membranes, Cr(VI) is considered the more toxic Cr species. However, it is intracellular Cr reduction reactions, leading to Cr(III) formation, which exerts more damage within cells. Cr induces a range of DNA lesions in cells, and this is one potential cause of Cr-associated cancer (Kirpnick-Sobol et al., 2006; Peterson-Roth et al., 2005; Reynolds et al., 2009). Protein oxidation can also be important for Cr toxicity (Sumner et al., 2005), as can further mechanisms described below. Concern over continuing industrial use of Cr and environmental Cr pollution (Salnikow and Zhitkovich, 2008; Wise et al., 2009) has reinforced the urgent need to resolve the true mode(s) of Cr toxicity. An understanding of Cr toxicity should also inform potential “biological” applications of Cr, such as in nutritional supplements for diabetes sufferers (Anderson, 1998).
Transport of chromate anions (usually the principal available Cr species) into cells is known to occur via plasma membrane sulfate transporters (Alexander and Aaseth, 1995; Pereira et al., 2008). Resultant competition between chromate and sulfate for uptake can lead to sulfur starvation in yeast (Pereira et al., 2008). Thus, sulfur starvation would represent a ROS-independent mechanism of Cr toxicity, alternative to DNA damage. A further mechanism of Cr toxicity was revealed following a recent screen of the heterozygous Saccharomyces cerevisiae deletion strain library against Cr (Holland et al., 2007). An observation in that study that proteasomal mutants were Cr sensitive led to a demonstration that Cr provokes messenger RNA (mRNA) mistranslation and that Cr toxicity was largely attributed to this loss of translational fidelity in conjunction with an associated production of toxic protein aggregates within the cells (Holland et al., 2007).
Recent reports have described a role for actin-mediated endocytosis in cellular processing of toxic protein aggregates (Ganusova et al., 2006; Meriin et al., 2007). Considering our observation that Cr induces protein aggregation (see above), a further point of interest from the data in that screening study of Cr toxicity (Holland et al., 2007) is an apparent link between the actin cytoskeleton and the Cr resistance. Thus, the annotations of genes of the Cr-sensitive strain data set from the screen were disproportionately overrepresented by actin-related functions. In this study, we investigated the basis for a novel role for the actin cytoskeleton in Cr resistance. We show that actin-mediated endocytosis is required to limit intracellular Cr accumulation and, so, to help resist Cr toxicity.
Saccharomyces cerevisiae BY4743 (MATa/α his3Δ1/his3Δ1 leu2Δ0/leu2Δ0 LYS2/lys2Δ0 met15Δ0/MET15 ura3Δ0/ura3Δ0) and the isogenic homozygous deletion strains abp1Δ, ede1Δ, ent1Δ, bzz1Δ, pil1Δ, sac6Δ, chc1Δ, and end3Δ, in which the specified open reading frames are replaced with the KanMX4 marker, were obtained from Euroscarf (Frankfurt, Germany). David Drubin (University of California, Berkeley, CA) kindly provided S. cerevisiae strains ARP2::URA3, arp2-G302Y::URA3, ARP3::LEU2, and arp3-G302Y::LEU2 (Martin et al., 2005); the “ARP2” and “ARP3” strains correspond to wild types: the URA3 or LEU2 markers were inserted adjacent to the native ARP2 or ARP3 genes to provide controls that are isogenic to the arp2-G302Y and arp3-G302Y expressing strains. All constructs are genome integrated. A strain expressing Abp1p as a translational fusion with green fluoresent protein (GFP) was purchased from Invitrogen (Carlsbad, CA; 95,700 strain series). A sul1Δ/sul2Δ double mutant (sul1::KanMX4 and sul2::hphNT1) was constructed in this study as follows. A homozygous diploid sul1Δ mutant (Euroscarf) was sporulated (Amberg et al., 2005) and a haploid sul1Δ strain selected (his3Δ1, leu2Δ0, ura3Δ0, MET15, LYS2, and sul1::KanMX4). The SUL2 gene was disrupted in this strain using pFA6a-hphNT1 (Janke et al., 2004) as the template for short-flanking homology (SFH)-PCR-based disruption (Longtine et al., 1998). Transformants (Gietz and Woods, 2002) were selected on yeast extract peptone dextrose (YEPD) medium supplemented with 150 μg hygromycin/ml (Invitrogen). Diagnostic PCR (Longtine et al., 1998) was used to confirm appropriate gene disruption. Where used, colony PCR for diagnosis involved heating a colony in 5 μl water for 5 min at 95°C, before using the entire sample as a template for PCR. To make a sul1Δ/sul2Δ/sac6Δ triple mutant, the SAC6 gene was disrupted in the above sul1Δ/sul2Δ haploid mutant by SFH-PCR, using pFA6a-His3MX6 (Longtine et al., 1998) as the template. Transformants were selected on yeast nitrogen base without amino acids (Formedium), supplemented as required (Ausubel et al., 2007). To generate a sac6Δ mutant isogenic to the sul1Δ/sul2Δ/sac6Δ triple mutant, diploid S. cerevisiae BY4743 was sporulated, a His, Leu, Ura auxotroph selected (his3Δ1, leu2Δ0, ura3Δ0, MET15, and LYS2), and the SAC6 gene disrupted as above.
Organisms were routinely maintained on YEPD agar and prepared for experiments by culturing in YEPD broth with shaking at 120 rpm, 30°C (Bishop et al., 2007). Experimental S. cerevisiae cultures were inoculated from overnight starter cultures grown from single colonies. These experimental cultures were preincubated as above for 4 h before the addition of CrO3 (from filter-sterilized stock solutions) or Latrunculin A (Biomol International, Plymouth Meeting, PA). For growth inhibition assays, subsequent growth [optical density 600nm (OD600)] in 300-μl aliquots of medium was monitored continuously in 48-well microplates (Greiner Bio-One, Monroe, NC), incubated with shaking at 30°C in a BioTek Powerwave microplate spectrophotometer (Smith et al., 2007).
Cells were fixed with formaldehyde (4% vol/vol) with shaking at 30°C for 15 min, then harvested by centrifugation, resuspended in PBS, and incubated with 4% (vol/vol) formaldehyde for a further 1 h at room temperature. The fixed cells were washed five times in PBS and resuspended in 50 μl PBS. An 8-μl aliquot was removed and 2 μl Alexa Fluor 488 phalloidin (Invitrogen) added, followed by incubation in the dark at room temperature with vortexing every 15 min for 90 min. After five washes with PBS, the actin-stained cells were visualized using either a Deltavision Personal DV (Imsol, Preston, UK) microscope with SoftWoRx deconvolution software (Applied Precision, Issaquah, WA) or a Zeiss Axioskop fluorescence microscope equipped with an HB0 50 illuminator.
To monitor FM4-64 uptake by flow cytometry, cells were harvested by centrifugation from 16-h experimental cultures and resuspended in the supernatant to OD600~0.5. FM4-64 (SynaptoRed reagent; Calbiochem, EMD Biosciences, San Diego, CA) was added to 3 ml cell suspension to a final concentration of 8μM. Samples (250 μl) were removed at intervals, the cells fixed with formaldehyde (10%, vol/vol), and then incubated for 40 min at room temperature before washing three times with PBS. Cells were analyzed with a FACScanto flow cytometer (Becton Dickinson, Franklin Lakes, NJ), with excitation at 488 nm and emission with a 780/60 bandpass filter. Fluorescence readings were corrected for cell size by division with forward scatter readings. Data were collected for 30,000 cells in each sample. For visualization, FM4-64 was added to culture samples to a final concentration of 40μM. Samples were removed at intervals and kept on ice until visualization by deconvolution microscopy (Deltavision, Applied Precision, Issaquah, WA). To measure Lucifer yellow (LY) (Sigma, St Louis, MO) uptake, the dye was added to cell suspensions (prepared as described above) to a final concentration of 8.7mM and incubated at 30°C with shaking. Samples (100 μl) were removed at intervals, washed three times with PBS, and kept on ice until analysis either by fluorescence microscopy (Zeiss Axioskop) or by fluorimetry (Cary Eclipse Fluorescence spectrophotometer), with excitation at 428 nm and the emission read at 535 nm.
Exponential phase cultures were supplemented with CrO3 as specified. Samples were removed at intervals and cells harvested by centrifugation. After five washes with dH20, cells were resuspended in 1 ml dH20 and the OD600 recorded. Cells were pelleted (16,000 × g, 2 min), resuspended in 3 ml dH2O, and 0.5 ml of 70% (vol/vol) HNO3 was added (Fisher, trace analysis grade). Cell samples were digested during heating with microwave irradiation (Anton Paar Multiwave, fitted with a 48-slot carousel): the microwave was ramped up to 1400 W over 10 min, the samples were held for a further 10 min, and then ramped down to 0 W over 20 min. Digested samples were diluted to 25 ml with milli-Q water (18.2 MΩ·cm) and stored until analysis. Multi-element analysis was undertaken using inductively coupled plasma atomic emission spectroscopy (ICPMS) (Thermo-Fisher Scientific X-SeriesII) with appropriate standard solutions. Levels of Cr were normalized to cell biomass (OD600) in each sample and corrected for background Cr measured in untreated cells. The background-equivalent concentration of the ICPMS analyses was ~0.025 ng Cr/ml, corresponding to an instrument detection limit of ~0.0025 ng Cr/ml. The background Cr content of untreated cells was ≤ 1.0 ng Cr per OD600 unit of cells.
Previously, a screen of the heterozygous S. cerevisiae deletion strain collection for CrO3 sensitivity led to the finding that Cr provokes mRNA mistranslation and protein aggregation and that mistranslation is a primary cause of Cr toxicity (Holland et al., 2007). An additional observation from the data of that screen is that strains defective in actin- and cytoskeleton-related functions are overrepresented among the group of Cr-sensitive mutants. Thus, of 97 gene ontology (GO) terms that were significantly overrepresented in the annotation of genes in the Cr-sensitive data set, 18 were in actin-related functions. No actin-related functions were overrepresented in the annotation of genes in the Cr-resistant data set. In the Cr-sensitive data set, the most significantly overrepresented of all the GO terms was the Arp2/3 complex, an actin nucleation center required for the motility and integrity of actin patches (Table 1). To corroborate this apparent requirement for the actin cytoskeleton in Cr resistance in the present study, cells were treated simultaneously with Cr and the actin monomer–binding drug latrunculin A (LatA); LatA is an inhibitor of actin polymerization (Ayscough et al., 1997). Treatment with Cr and LatA together had a more marked effect on net yeast growth than the sum of the two independent treatments (Fig. 1), indicating a synergistic action. This supported the hypothesis that the actin cytoskeleton, and actin polymerization specifically, is important for Cr resistance.
To determine whether Cr treatment perturbs the F-actin cytoskeleton, cells were stained with Alexa Fluor 488 phalloidin. Exponential control cells (minus Cr) exhibited a staining pattern that is typical of actively growing cells, with discrete actin patches localized to the surface of growing buds, and actin cables polarized within mother cells (Fig. 2A). Actin patches were somewhat delocalized in cells treated with Cr for 2 h, with the mother cells (indicated by arrows) showing some actin patches and fewer cables. After 16-h growth, control cells entering the diauxic shift ahead of stationary phase exhibited a small number of intensely staining F-actin foci (Fig. 2B), consistent with the “actin bodies” of yeast exiting exponential growth (Sagot et al., 2006). In contrast, the Cr-treated cells maintained a number of small actin patches. These differing phenotypes were not attributable simply to a slowed growth cycle during Cr treatment, as the differences were maintained through progression to stationary phase.
As actin patches are the initiation sites of endocytosis, we used the dyes FM4-64 and LY to compare endocytic activity in control and Cr-treated cells. The FM4-64 dye fluoresces when it is incorporated into membranes. The plasma membrane stains immediately upon FM4-64 addition to cells, and the labeled membrane is subsequently internalized by endocytosis (Vida and Emr, 1995). Intracellular redistribution of the dye to the vacuolar membrane (the ultimate destination of internalized plasma membrane) was more rapid in the Cr-treated than in untreated cells (Fig. 3A), indicating that Cr treatment increases membrane trafficking. This was substantiated with LY, a common marker of fluid phase endocytosis. LY uptake was increased markedly in the Cr-treated cells (Fig. 3B). The LY staining also indicated that the Cr-treated cells had enlarged vacuoles compared with the control cells. The resultant proximity of the vacuolar and plasma membranes of Cr-treated cells precluded our attempts to quantify endocytic rate precisely by observation of the internalization of individual Abp1-GFP foci (Smythe and Ayscough, 2006). Nonetheless, results with FM4-64 and LY showed a Cr-induced stimulation of endocytosis.
In conjunction with the sensitization to Cr observed in actin-impaired cells (Table 1 and Fig. 1), the increase in actin-mediated endocytosis observed in response to Cr exposure (Fig. 3) suggested that endocytosis may play a role in Cr resistance. To test whether there was a requirement in Cr resistance specifically for functional endocytosis, among the range of actin-related functions, we used strains expressing alleles for modified versions of the actin nucleation complex proteins, Arp2 and Arp3. Expression of the arp3-G302Y allele in place of wild-type ARP3, which confers defects in endocytosis as well as actin nucleation (Martin et al., 2005), gave some Cr sensitivity (Fig. 4B). In contrast, the arp2-G302Y allele renders yeast defective for actin nucleation but not endocytosis (Martin et al., 2005) and conferred a slight Cr resistance (Fig. 4A). The reason for the latter resistance is not known, but the data indicated that it is specifically endocytosis that is important for actin-mediated Cr resistance.
Cr toxicity was examined in a further series of endocytosis-defective mutants. In the presence of Cr, three of the tested mutants (chc1Δ, end3Δ, and sac6Δ) exhibited markedly decreased growth compared with the wild type, indicating Cr sensitivity (Fig. 5A). Chc1p, End3p, and Sac6p are involved in the initial internalization step of endocytosis. Other tested mutants (abp1Δ, ede1Δ, ent1Δ, bzz1Δ, and pil1Δ) were not discernibly Cr sensitive (data not shown). We reasoned that if the synergy between Cr and LatA shown in Figure 1 was linked specifically to endocytosis, then that synergy should be suppressed in a (Cr sensitive) endocytosis-defective mutant. Therefore, we treated end3Δ cells with Cr and LatA individually or together. Unlike the effect in wild-type cells (Fig. 1), the combined treatment did not have a greater effect on growth of the end3Δ mutant than the sum of the two independent treatments (Fig. 5B). Collectively, the results assign the link between Cr toxicity and actin specifically to endocytosis, with Cr toxicity being increased in a subset of endocytosis mutants.
The above results showed that certain components of the endocytosis apparatus are required for cellular Cr resistance. One potential explanation was that these components might be required for processing Cr-induced cytotoxic protein aggregates (see “Introduction”). However, two of the endocytosis mutants that showed elevated Cr sensitivity here (chc1Δ and end3Δ) give decreased aggregation (of heterologous proteins with expanded polyglutamine domains) when functioning in the processing pathway (Meriin et al., 2007). On the other hand, and in the context of metal toxicity, the same three endocytosis functions required for Cr resistance here were shown previously to be required for endocytosis of the high-affinity zinc transporter, Zrt1p, in response to Zn stress (Gitan et al., 1998). Consequently, we hypothesized that a similar response directed at a Cr-transporting protein could act to decrease Cr accumulation in response to Cr stress, accounting for the present phenotypes. Therefore, we compared Cr accumulation by the wild-type and endocytosis-defective mutants. The Cr-sensitive endocytosis mutants (sac6Δ, chc1Δ, and end3Δ) accumulated high levels of Cr (Fig. 6A). Indeed, the most sensitive strains, end3Δ and sac6Δ, accumulated greater than fourfold more Cr than the wild type at an external concentration of 50μM CrO3. Non-Cr–sensitive endocytosis mutants exhibited comparable intracellular Cr to the wild type. It was inferred that the increased Cr sensitivities of the sac6Δ, chc1Δ, and end3Δ mutants were a likely result of increased Cr accumulation.
To further support the idea that endocytosis may be required for internalization of a Cr transporter, akin to the Zrt1p-targeted Zn response (Gitan et al., 1998), we monitored the initial kinetics of Cr uptake (Fig. 6B). If Cr uptake occurs through a transporter that can be downregulated by endocytic internalization in response to Cr, then an initial endocytosis-independent phase of Cr uptake should be followed by an endocytosis-dependent decline in the Cr accumulation rate. We tested this by comparing Cr uptake kinetics in wild-type and end3Δ cells, the latter mutant selected as it exhibited the strongest phenotype in the long-term Cr accumulation assays (above). An initial rapid uptake of Cr within 10 min was similar in wild-type and end3Δ cells, indicating independency from endocytosis. A subsequent slower phase of Cr accumulation was markedly faster in the endocytosis mutant than in the wild type. Cr accumulation was continuing in end3Δ cells after 100 min, by which time, net Cr accumulation by the wild type had virtually ceased (Fig. 6B). We noted that this difference in Cr accumulation, evident within 40–70 min, occurred before any discernible difference in growth rate of the two strains in the same experiments. This ruled out the possibility that the difference in Cr accumulation was due to differing growth rates and associated dilution of cellular Cr. The data supported the idea that the end3Δ mutant may fail to internalize a transporter(s), allowing Cr to enter the cells continuously.
Plasma membrane proteins of yeast are commonly marked for internalization by ubiquitination. Specifically, the ubiquitin isopeptidase encoded by DOA4 is known to be important for internalization of several transporters in yeast (Galan and Haguenauer-Tsapis, 1997; Horak and Wolf, 2001). Therefore, we compared Cr accumulation in wild-type and doa4Δ mutant cells. As in the endocytosis mutant (Fig. 6B), deletion of the ubiquitin isopeptidase partly suppressed the decline in Cr accumulation over time (Fig. 6C). The doa4Δ mutant was also Cr hypersensitive (data not shown). The data further supported the hypothesis that Cr uptake is limited by internalization of a Cr transporting activity.
Chromate transport into yeast cells is mediated, at least partly, by the sulfate permeases encoded by SUL1 and SUL2 (Pereira et al., 2008). These transporters are expected to be repressed in rich medium such as that used here (Cherest et al., 1997; Pereira et al., 2008). Nevertheless, to discount the possibility that endocytosis of either one or both of these transporters was responsible for the decreased Cr accumulation and toxicity in the wild-type versus sac6Δ, chc1Δ, and end3Δ mutants, we constructed a sul1Δ/sul2Δ double mutant and a sul1Δ/sul2Δ/sac6Δ triple mutant (our attempts to construct a sul1Δ/sul2Δ/end3Δ mutant were unsuccessful). This was to determine whether the Cr resistance of wild-type versus sac6Δ cells (Fig. 5) depends on the expression of Sul1p and Sul2p. Deletion of SAC6 yielded a Cr-sensitive phenotype in the sul1Δ/sul2Δ background, which was comparable to that observed in the wild-type background (Fig. 7). This indicated that endocytosis-dependent Cr resistance does not require (endocytosis of) Sul1p or Sul2p. Sul1p and Sul2p are the only characterized Cr transporters in yeast, but they only partly account for the Cr uptake observed (Pereira et al., 2008) indicating the presence of an alternative uncharacterized Cr transporter(s). Having ruled out Sul1p and Sul2p, our findings suggest that this uncharacterized Cr transport activity in yeast may be one that is endocytosed in response to Cr stress.
This work has elucidated a role for the actin cytoskeleton in mediating resistance to chromate. We observed that cortical actin patches adopted an unusual conformation during Cr stress, coincident with a marked stimulation of endocytosis, and that endocytosis is required for Cr resistance. Proteins involved in the initial internalization step of endocytosis—Chc1, End3, and Sac6—were shown to limit the accumulation of intracellular Cr. Thus, loss of any of these proteins led to higher levels of Cr accumulation and associated toxicity. These findings rationalize one of the most striking observations from our previous screen of the heterozygous deletion strain collection: a marked overrepresentation of actin-related functions among the Cr-sensitive gene data set (Holland et al., 2007). The most significantly overrepresented category in that data set was the Arp2/3 protein complex, required for actin nucleation. By exploiting arp2 and arp3 alleles that allow these genes’ roles in endocytosis to be dissected from more generic aspects of actin nucleation–dependent function (Martin et al., 2005), as well as testing other endocytosis mutants and the existence of synergy between Cr and latrunculin A action, we were able to assign the actin dependency of Cr resistance specifically to endocytic function.
While the novel stimulation of endocytosis by Cr described here would appear to have immediate benefits for Cr resistance, incidental effects could limit these in the longer term. Thus, the response to Cr was associated with the maintenance of multiple actin patches rather than the normal development of “actin bodies” as cells progressed toward stationary phase. Actin bodies (or actin aggregates) have been proposed to act as an actin store, used by cells for reestablishing polarity upon reentry to growth from stationary phase (Sagot et al., 2006). We have noted that cells treated overnight with Cr, i.e., lacking actin bodies, show a Cr concentration–dependent delay in resumption of growth when inoculated to fresh medium lacking Cr (S.H and S.V.A, unpublished data). Furthermore, the timing of reentry to growth was correlated with the time from which polarized actin could again be observed in cells. These observations are consistent with a role for actin bodies as actin stores (Sagot et al., 2006). Moreover, they highlight one potential reason why the actin rearrangements described in this paper may be a shorter-term rather than long-term response to Cr stress.
We considered different explanations for the endocytosis dependency of Cr resistance. The specific phenotypes of the mutants we tested did not fit with observations elsewhere on endocytosis-mediated aggregation of proteins with expanded polyglutamine domains (Meriin et al., 2007). Strikingly, the same three endocytosis proteins identified here as being required for Cr resistance were shown previously to be important for the internalization of the high-affinity Zn transporter Zrt1p (Gitan et al., 1998). The higher levels of Cr accumulated by the sac6Δ, chc1Δ, and end3Δ mutants, accounting for their Cr sensitivities, lends further support to the idea that endocytosis may play a similar role in Cr as in Zn resistance, i.e., internalization of a plasma membrane transporter(s) responsible for metal uptake. A similar response also targets high-affinity Fe transporters during exposure of cells to elevated Fe (Felice et al., 2005). Furthermore, an initial endocytosis-independent phase of Cr uptake here was followed by an endocytosis- and recycled ubiquitin–dependent decline in uptake, in keeping with Cr-transporter internalization as a response to initial Cr stress.
At present, the only transporters in yeast known to mediate Cr(VI) entry are the sulfate permeases, Sul1p and Sul2p. Our genetic analysis showed that endocytosis-dependent Cr resistance cannot be related solely to Sul1p or Sul2p function, e.g., arising from endocytic inactivation of Sul1p and Sul2p during Cr stress. Besides, if there was such a response, it may only undermine the transcriptional upregulation of SUL1 and SUL2 that occurs during Cr stress (Pereira et al., 2008). That upregulation is thought to occur as a response to counter S starvation, which results from CrO42− competition with SO42− for uptake via Sul1p and Sul2p. Moreover, Sul1p and Sul2p do not account exclusively for Cr uptake, as ~40% of Cr uptake activity is retained in a sul1Δ/sul2Δ double mutant via an unknown mechanism (Pereira et al., 2008). The relative contribution of this unknown mechanism to Cr uptake should be increased in the YEPD medium used here in which SUL1 and SUL2 expression is expected to be repressed (Cherest et al., 1997; Pereira et al., 2008). Therefore, if endocytic inactivation of a Cr-transporting protein does occur, as we suggest, our evidence indicates that it must be targeted at this uncharacterized Cr uptake activity(s). We cannot discount the possibility that endocytosis-dependent Cr resistance requires (internalization of) three or more Cr transporters, leaving open a potential partial role for Sul1p and Sul2p. Moreover, testing these possibilities will await the discovery first of the alternative Cr uptake system(s). We have discounted the possibility that transporters already known to be internalized (Zrt1p, Ftr1p, and Fet3p) may play a role here as the corresponding deletion strains are not significantly affected for Cr uptake (S.H. and S.V.A., unpublished data). Note that no novel candidate Cr transporter has emerged from further analysis of our existing screening data (Holland et al., 2007), although that is not unexpected as those data are ideally suited to mining for broader trends and effects (e.g., Cr sensitivity of actin-related mutants) rather than definitive assessment of each strain’s individual behavior (Holland et al., 2007; Khozoie et al., 2009).
The novel physiological response to Cr stress described here is unexpected, as it suggests a selective pressure due to Cr during the organism’s evolutionary history. However, this would contradict an observation from our previous genome-wide screen—that more yeast deletion strains are Cr resistant than Cr sensitive (Holland et al., 2007)—as when selection occurs in the environment to which an organism is adapted, then most mutations will be deleterious; whereas, when selection occurs in increasingly suboptimal conditions, then an increasing proportion of mutations will be beneficial (Fisher, 1930). Therefore, it seems unlikely that the present response has evolved specifically to Cr. The possibility that enhanced endocytosis might be part of a general stress response is undermined by the fact that there have been almost no reports of a net increase in endocytic rate among the many previous studies of cellular stress responses. An alternative explanation, therefore, is that chromate may share a specific response pathway with a close chemically related species, e.g., sulfate.
In conclusion, this study has reinforced the value of the Cr toxicity database we established previously (Holland et al., 2007), by rationalizing the most striking outcome from that screen: The marked overrepresentation of actin-related functions among cellular activities required for Cr resistance. The new insight to Cr toxicity revealed here has particular resonance considering the continuing episodes of environmental Cr pollution and the current widespread promotion of Cr(III) supplements to potential Alzheimer’s and diabetes sufferers (Anderson, 1998; Wang and Yao, forthcoming). Appreciation of any long-term impact of Cr exposure in organisms requires a full understanding of Cr toxicity. That aim may now be a step closer with the link between endocytosis, Cr accumulation, and Cr toxicity established here.
Natural Environment Research Council (NE/E005969/1) and the National Institutes of Health (R01 GM57945).
We thank David Drubin (University of California, Berkeley) for his kind gifts of yeast strains. We also thank Kathryn Ayscough (University of Sheffield) and Nick Read (University of Edinburgh) for helpful discussions.