Skip to main content
NCBI home page
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 2006 May 23;103(23):8674–8679. doi: 10.1073/pnas.0600928103

Zinc binding to a regulatory zinc-sensing domain monitored in vivo by using FRET

Wei Qiao *,†, Michelle Mooney *,†,‡, Amanda J Bird §, Dennis R Winge §, David J Eide *,
PMCID: PMC1482638  PMID: 16720702

Abstract

We have generated probes of metal binding to zinc fingers (ZFs) that provide tools to study zinc trafficking in vivo. In this study, we used these probes to examine zinc binding by the Zap1 transcription factor of Saccharomyces cerevisiae. Zap1 contains two zinc-regulated activation domains (ADs), AD1 and AD2. AD2 is located within two C2H2 ZFs, ZF1 and ZF2. Studies have indicated that apoAD2 activates transcription and zinc binding to ZF1 and that ZF2 forms an interacting-finger-pair structure that is necessary to inhibit AD function. A related structural finger pair, ZF3 and ZF4, is found in the Zap1 DNA binding domain. In vitro studies indicated that, although the ZF1/2 and ZF3/4 finger pairs bind zinc with similar affinities, zinc that was bound to ZF1/2 was much more labile. We examined the properties of Zap1 ZFs in vivo by FRET. ZF pairs were flanked by enhanced yellow fluorescent protein and enhanced cyan fluorescent protein, allowing detection of zinc-induced conformation changes by FRET. By using these reporters, we found that ZF1/2 and ZF3/4 showed similar responses to zinc under steady-state conditions in vivo. In contrast, ZF1/2 zinc binding was significantly more labile than was ZF3/4. Also, ZF1/2 accumulated in an apo form that could rapidly bind zinc, whereas the ZF3/4 pair did not. Last, we show that these properties are evolutionarily conserved indicating their importance to Zap1 function. These results indicate that the kinetic lability of ZF1/2 in vivo is a key component of Zap1 zinc responsiveness.

Keywords: regulation, transcription, zinc finger

The Cys-2His-2 (C2H2)-type zinc finger (ZF) domain is one of the most common structural motifs in biology. Sequencing of the human genome suggested that ≈3% of the encoded proteins bear C2H2 ZFs (1). These domains play a number of different roles by mediating DNA binding, RNA binding, and protein–protein interactions (2–4). The canonical C2H2 ZF has the consensus sequence of ψ/Y-X-C-X2,4-C-X3-F-X5-ψ-X2-H-X3–5-H, where ψ denotes a hydrophobic residue (5). Binding of zinc to the cysteines and histidines folds the largely unstructured apoprotein into the compact ββα form (i.e., two antiparallel β strands followed by an α-helix) (5). Many studies have addressed zinc binding to the C2H2 ZF motif in vitro (5–8). However, to our knowledge, no studies have examined directly the kinetics and other properties of zinc binding to these domains in vivo. The aim of this study was to develop genetically encoded fluorescent probes for assessing zinc binding to ZFs. We used these probes to examine the properties of zinc binding to Zap1, the zinc-responsive transcriptional activator of Saccharomyces cerevisiae in living cells.

Zap1 regulates zinc homeostasis in this yeast. It controls the expression of many genes that have critical roles in zinc metabolism, including zinc uptake and the mobilization of intracellular zinc stores (9). Zap1 increases target gene expression in zinc-limited cells, and its activity is inhibited in replete cells. The Zap1 protein has three functional domains. At the C terminus, there is a DNA-binding domain consisting of five C2H2 ZFs, which are designated ZF3–ZF7 (10). These fingers confer binding of Zap1 to zinc-responsive elements (ZREs) that are found in the promoters of its target genes. Because Zap1 DNA binding is required under low-zinc conditions, the DNA binding fingers must stably bind Zn2+ for protein function. The Zap1 DNA binding domain is similar to the five-ZF DNA binding domain of the mammalian Gli protein (11). In Gli, the first and second fingers of the DNA binding domain form an unusual ZF pair structure where the two fingers interact via hydrophobic residues lining an interface of protein–protein contact. The corresponding fingers in the Zap1 DNA binding domain, ZF3 and ZF4, likely form a similar finger pair structure because the residues lining the Gli finger–finger interface are conserved in Zap1 ZF3 and ZF4 (12).

Zap1 also contains two activation domains (ADs) that are independently regulated by zinc. Zinc-regulated AD1 maps to residues 182–502 (13). The second Zap1 AD is designated AD2. The AD2 region, residues 611–641, is located within two additional C2H2 ZFs, ZF1 and ZF2. These fingers were also proposed to form a Gli-like finger pair structure based on the conservation of interface residues (12). This hypothesis was confirmed recently by an NMR analysis of the Zap1 ZF1/2 finger pair (14). Mutational analysis indicated that the finger-pair structure is critical for the zinc responsiveness of AD2. Whereas the WT AD2 domain is shut off in high zinc, mutants that disrupt the zinc binding ligands in either of the AD2 ZFs resulted in constitutive, zinc-insensitive target gene expression (12). Also, mutations disrupting the finger–finger interaction caused constitutive activation. These results support a model in which, in the apo state, AD2 recruits basal transcription factors to increase Zap1 target gene expression. Binding of zinc to the AD2 ZFs folds the domain into the compact finger-pair structure and, thus, blocks AD function.

This model suggested that the ZF1 and ZF2 fingers of Zap1 are regulatory zinc-binding sites, whereas the related ZF3/4 fingers of the DNA binding domain are structural zinc sites. This hypothesis was examined with in vitro binding studies using peptides containing either the ZF1/2 or the ZF3/4 finger pairs (12). Both finger pairs bound two Zn2+ ions with similar affinities, with one site in each pair having an apparent KD of ≈0.2 nM and the second site having a KD of ≈4 nM. However, zinc binding to the ZF1/2 finger pair in vitro was much more labile than was binding of metal ions to the ZF3/4 pair. This difference in zinc lability supported a role for ZF1 and ZF2 as the zinc sensors for controlling AD2 function. Thus, a key prediction of this model was that the zinc lability of ZF1 and ZF2 observed in vitro would also be present in vivo. In this article, we describe the generation of fluorescent probes to assess the properties of zinc binding to these ZF domains in vivo. In addition to our insights specifically regarding Zap1, this study represents an important “proof of concept” about the potential uses of these and related probes as tools to study zinc utilization in living cells.

Results

Zap1 ZF1 and ZF2 form an interacting-finger-pair complex when zinc is bound (Fig. 1A) (12, 14). In addition to forming the canonical ββα ZF folds, these fingers also interact with each other via hydrophobic residues lining an interface of finger–finger interaction (Fig. 1B, gray residues). This compact structure suggested a possible approach to study zinc binding and finger interaction in vivo by using FRET (Fig. 1C). Our strategy was based largely on the work of Tsien and colleagues (15) on Ca2+ sensors devised by using the same approach. In those previous studies, calcium-binding peptides were flanked by variant GFP moieties that have overlapping emission/excitation spectra. Similarly, FRET has been used to study zinc binding to metallothionein in vitro and in vivo (16, 17). We used this intramolecular FRET approach to develop an in vivo Zn2+ reporter based on ZF1 and ZF2 of Zap1. Specifically, we constructed a hybrid protein in which this ZF pair was flanked by enhanced yellow fluorescent protein (eYFP) and enhanced cyan fluorescent protein (eCFP). As shown in Fig. 1C, excitation of eCFP at 440 nm results in light emission with a peak at ≈480 nm. If the eCFP and eYFP proteins are in close proximity, energy can transfer from eCFP to eYFP, resulting in emission at ≈545 nm. The degree of FRET can be determined by dividing the intensity of emission at 545 nm by the emission at 480 nm when excited at 440 nm; this “FRET ratio” reflects the distance between eCFP and eYFP (18).

Fig. 1.

Fig. 1.

Open in a new tab

Construction of the ZF1/2-FRET sensor. (A and B) Ribbon diagrams of the ZF1/2 finger pair in side (A) and top (B) view. (14). ZF1 is shown in red, and ZF2 is shown in green. The Zn2+ atoms (yellow), the C2H2 zinc ligands (blue), and the finger–finger interaction residues (gray) are shown. (C) Depiction of the ZF1/2-FRET sensor. eYFP and eCFP were fused to the N and C termini of the finger pair, respectively. (D) An alignment of ZF1/2 and ZF3/4 regions from S. cerevisiae (Sc) and C. albicans (Ca). The Zn2+ ligand residues are shown in blue, and the finger–finger interaction residues are shown in red. The locations of the β-strands and α-helices are also indicated.

The peptide that was inserted to generate the FRET reporter containing ZF1 and ZF2 (designated ZF1/2-FRET) was residues 575–643 of Zap1 (Fig. 1D). We anticipated that the zinc-bound finger pair structure would bring the flanking eCFP/eYFP moieties into close proximity such that efficient energy transfer could occur. As controls, we constructed a direct fusion of eYFP and eCFP (pYFP-CFP) as well as a ZF1/2-containing molecule in which the histidine Zn2+ ligands of both fingers (i.e., H599 and H604 of ZF1 and H636 and H641 of ZF2; Fig. 1A) were mutated to glutamines. This construct was designated ZF1/2-C2Q2-FRET.

These FRET proteins were expressed in yeast from the strong, constitutive PMA1 promoter. We assessed their subcellular distribution by using fluorescence microscopy. DAPI was used to stain the nucleus. As shown in Fig. 2A, no GFP fluorescence was detected in vector-only transformed cells. In ZF1/2-FRET-expressing cells, the reporter protein was present in both the cytosol and the nucleus but was excluded from the vacuole. ZF1/2-FRET localization was not altered by changes in zinc status, and similar results were obtained for the eYFP-eCFP and mutant ZF1/2-C2Q2-FRET proteins (data not shown).

Fig. 2.

Fig. 2.

Open in a new tab

Expression of ZF1/2-FRET in yeast. (A) WT (DY1457) cells bearing the vector (pDR195) or pZF1/2-FRET and grown in LZM plus 1 μM ZnCl2 were analyzed by fluorescence microscopy. Nuclei were detected with DAPI, and the cells were also viewed by Nomarski optics (Nom). (B) Immunoblot analysis of lysates from cells expressing the indicated proteins and grown in low (−Zn, LZM + 1 μM ZnCl2) or high (+Zn, LZM + 1,000 μM ZnCl2) zinc. Previous studies showed that there is no loss of cell viability under these growth conditions (21). (Left) Samples of purified GFP used for quantitating in vivo abundance. (C) WT cells bearing the ZRE-lacZ reporter (pDg2L) and either pDR195 or pZF1/2-FRET were grown overnight in LZM supplemented with the indicated concentration of zinc and then assayed for β-gal activity. Each point is the mean of three replicates. Error bars represent ±1 SD.

To assess the level of expression of the FRET reporter proteins, immunoblot analyses were performed by using anti-GFP antibody and purified GFP as a quantitative standard (Fig. 2B). These immunoblot analyses indicated that the level of FRET reporter accumulation in cells was similar in both zinc-limited and -replete cells. For the direct eYFP–eCFP fusion and ZF1/2-FRET, ≈50,000 molecules of reporter accumulated per cell. For unknown reasons, the ZF1/2-C2Q2-FRET reporter accumulated to higher levels (i.e., ≈200,000 molecules per cell).

This high-level expression suggested that zinc binding by the ZF1/2-FRET protein could perturb zinc homeostasis. To determine whether this hypothesis was true, we used a Zap1-responsive ZRE-lacZ reporter as a bioassay of the labile zinc pool. Zap1 is maximally active in stimulating expression of the ZRE-lacZ reporter in cells grown in low zinc and is inhibited by increased zinc in the medium (19). We predicted that, if ZF1/2-FRET was perturbing zinc homeostasis, we would see an increase in the activity of Zap1 because of the increased zinc demand. However, the response of the ZRE-lacZ reporter was similar in vector-only transformants and cells expressing ZF1/2-FRET (Fig. 2C). Therefore, the additional zinc binding capacity associated with ZF1/2-FRET expression does not alter zinc homeostasis appreciably.

The FRET responsiveness of these constructs was assayed in vivo by spectrofluorometry. Cells expressing the direct eYFP–eCFP fusion showed a high FRET ratio regardless of zinc status (Fig. 3A). This result is consistent with the close proximity of the eCFP and eYFP moieties in this construct under all conditions. In cells expressing ZF1/2-FRET, low-zinc growth conditions resulted in a low FRET intensity, whereas high zinc resulted in a marked increase in FRET ratio (≈30%). A similar increase in FRET ratio upon ion binding in vivo was observed for the previously described calcium sensors (15). Little effect of zinc was observed with the C2Q2 zinc ligand mutant. These data indicate that the FRET response detected with ZF1/2-FRET is (i) zinc-dependent, (ii) ligand-dependent, and (iii) not an effect of zinc status on other factors (e.g., intracellular pH and GFP dimerization) that could alter eCFP or eYFP fluorescence.

Fig. 3.

Fig. 3.

Open in a new tab

Zinc-dependent response of ZF1/2-FRET. (A and B) WT (DY1457) cells expressing the indicated proteins were grown as described for Fig. 2B and assayed for FRET by spectrofluorometry. The FRET ratio was determined by dividing the FRET emission at 545 nm by the eCFP emission at 480 nm from samples excited at 440 nm. Alleles are K582A W583A (MUT1), W620A E621K (MUT2), and L600D V605D (MUT3). (C and D) WT (DY1457) (C) or zap1 mutant (ZHY6) (D) cells were transformed either with pZF1/2-FRET or with pGAL1-lacZ and pGBD-Zap1552–705 and grown in media supplemented with the indicated concentrations of ZnCl2. WT cells were grown in LZM, whereas the zap1 mutant was grown in LZM lacking EDTA. Each point is the mean of three replicates. Error bars represent ±1 SD.

One trivial explanation for the decreased FRET in zinc-limited ZF1/2-FRET-expressing cells was that the fusion protein was proteolytically cleaved within the finger region between eCFP and eYFP under these conditions. Some protease genes are up-regulated in cells grown under zinc-limiting conditions (9). This proteolysis could separate the eCFP and eYFP moieties and reduce the FRET intensity. However, immunoblot analyses indicate that no detectable proteolysis occurred in either high- or low-zinc conditions (Fig. 2B).

We also tested whether the zinc-dependent FRET response of the ZF1/2 finger pair required interaction between the fingers. Interaction between ZF1 and ZF2 is mediated by tryptophans in the β-loop regions of each finger and by hydrophobic residues lining the interface between the two α-helices (Fig. 1 B and D) (12, 14). Mutation of the tryptophan (and adjacent charged residue) of ZF1 (MUT1, K582A W583A) or ZF2 (MUT2, W620A E621K) in ZF1/2-FRET eliminated the zinc-responsive conformational change that was observed in vivo (Fig. 3B). Mutation of interaction residues in the α-helix of ZF1 (MUT3, L600D V605D) also prevented zinc-responsive FRET. Thus, the FRET reporter behaved in vivo as predicted from the protein structure of this ZF pair.

One specific goal for generating this reporter was to test the correlation between zinc binding and AD2 AD function. Our hypothesis was that zinc binding to ZF1 and ZF2 is the regulatory zinc-sensing event and results in a conformational change shutting off AD2 function. An alternative hypothesis was that these fingers are structural zinc sites. For example, ZF1 and ZF2 could be structural ZFs acting as site for docking of another protein and this second protein is the zinc-sensing component of the regulatory system. In this latter model, ZF1/2 zinc binding (and the resulting conformational change) and zinc sensing could be separable events. To test these models further, we examined the correlation between zinc binding to ZF1 and ZF2 in vivo with AD2 function. AD2 function was assessed by using a fusion of the ZF1/2 domain to the Gal4 DNA-binding domain (12). Zinc regulation was then assessed by using a Gal4-responsive GAL1 promoter–lacZ fusion. As shown in Fig. 3C, there was a close correlation between the increase in FRET intensity and the decrease in AD function in WT cells with rising zinc added to the medium. This correlation was also observed in zap1 mutant cells (Fig. 3D). Because zap1 mutants are defective for zinc uptake, these cells require much more exogenous zinc to become zinc-replete (20). Thus, whereas WT cells were assayed in low-zinc medium containing 1 mM EDTA as a metal buffer, the zap1 mutant cells were grown in low-zinc medium (LZM) that was prepared without EDTA where the available zinc levels are much higher relative to the total added zinc. These results indicate that zinc binding by ZF1/2 and zinc sensing controlling AD2 function are closely correlated.

With the ZF1/2-FRET reporter, we could also probe the kinetics of zinc binding to ZF1 and ZF2 in vivo. One situation that we examined was zinc shock (i.e., when zinc-deficient cells are resupplied with zinc). Because zinc-limited cells up-regulate their zinc-uptake transporters, these cells rapidly accumulate zinc when treated with the metal (21). As shown in Fig. 4A, when zinc-limited WT cells were resupplied with 1,000 μM zinc, the FRET intensity starting at a low basal level reached its maximum within 10 s of zinc treatment. The short response time we observed is indicative of very rapid zinc delivery to the accumulated apo form of these ZFs. When lower levels of zinc were used, the response was more gradual consistent with the lower rates of zinc uptake occurring under these more moderate treatment conditions. As shown in Fig. 4B, this response depended highly on uptake via the high-affinity zinc-uptake transporter Zrt1. Mutants that were defective in zrt1 showed a slower FRET response to 1,000 μM Zn and no response to concentrations ≤100 μM, consistent with uptake occurring by means of lower-affinity transporters in this strain (20).

Fig. 4.

Fig. 4.

Open in a new tab

Response of ZF1/2-FRET to transitions in zinc availability. WT (DY1457) (A) or zrt1 mutant (ZHY1) (B) cells were grown in LZM plus 1 μM ZnCl2 to exponential phase. The cells were then harvested and transferred to a temperature-controlled cuvette in the spectrofluorometer. After 10 min, ZnCl2 was added to the indicated concentration and fluorescence was measure every 10 s. Zinc availability was buffered by the presence of 20 mM citrate in the assay buffer, and previous studies have shown that there is no loss of cell viability under these zinc-shock conditions (21). (C) WT (DY1457) cells were grown in zinc-replete LZM plus 100 μM ZnCl2. At time 0, the cells were inoculated into LZM + 1 μM ZnCl2 with (filled symbols) or without (open symbols) 100 μg/ml cycloheximide and incubated at 30°C for additional time. At the indicated times, cells were harvested and assayed for FRET. The values are normalized to the FRET ratio at time 0, and representative experiments are shown.

We also examined the behavior of the FRET reporter during zinc withdrawal. As shown in Fig. 4C, cells pregrown in a zinc-replete medium had a steadily decreasing FRET intensity after transfer to zinc-limiting conditions. The loss of FRET is likely to be due to the loss of zinc binding to ZF1/2 as the metal is released and bound by other newly synthesized metalloproteins. This hypothesis was indicated by the observation that the protein synthesis inhibitor cycloheximide blocked the decrease in FRET (Fig. 4C). Thus, ZF1/2 may not compete well for zinc relative to newly synthesized zinc-binding proteins.

In vitro studies of a Zap1 ZF1/2 peptide indicated that these fingers bind zinc with nanomolar affinity but show remarkable lability of metal binding (12). In comparison, the Zap1 ZF3/4 finger pair bound zinc with an affinity similar to ZF1/2 but with much less kinetic lability. To determine whether these differences held true in vivo, we constructed an analogous FRET reporter containing Zap1 ZF3 and ZF4. Also, we generated similar FRET reporters for the ZF1/2 and ZF3/4 finger pairs from the Candida albicans Zap1 ortholog.

To assess the relative affinities of these reporters for zinc in vivo, we first examined the FRET response in cells grown over a range of zinc concentrations. As shown in Fig. 5 A and B, the S. cerevisiae and C. albicans ZF1/2-FRET and ZF3/4-FRET reporters behaved similarly in their zinc responsiveness. These data are consistent with the similar affinities of the S. cerevisiae domains established in vitro (12) and suggest conservation of these affinities of the C. albicans domains. To examine the lability of zinc binding, we first compared the FRET response of these constructs during zinc withdrawal. Although both S. cerevisiae and C. albicans ZF1/2 reporters showed a decrease in FRET over time, the FRET intensities of the two ZF3/4 constructs were unaffected for several hours (Fig. 5C and D). This result indicates that zinc remains stably bound to the ZF3/4 pairs over this period of zinc withdrawal. Zinc binding was also assessed during zinc shock. Given that ZF1/2 and ZF3/4 peptides in vitro showed greatly different off-rates but similar affinities, we predicted that the on-rates would also be slower for the ZF3/4 finger pair. This result was predicted because the dissociation constant KD = Koff/Kon. Consistent with this expectation, whereas the S. cerevisiae and C. albicans ZF1/2-FRET constructs responded rapidly to zinc shock, the ZF3/4 constructs increased in FRET activity much more slowly (Fig. 5 D and E). An increase in response to 1,000 μM zinc was detected only after several minutes, as opposed to <10 s for the ZF1/2-FRET reporters. All of these responses were posttranslational; cycloheximide treatment did not alter the FRET response (data not shown). These results show that the same difference in lability observed in vitro exists also in vivo. Also, these results indicate that the lability of ZF1 and ZF2 is conserved between the evolutionarily distant S. cerevisiae and C. albicans Zap1 orthologs.

Fig. 5.

Fig. 5.

Open in a new tab

Behavior of orthologous ZF FRET constructs. WT (DY1457) cells bearing S. cerevisiae- or C. albicans-derived pZF1/2-FRET or pZF3/4-FRET plasmids were assayed for FRET after growth in LZM supplemented with a range of zinc concentrations (A and B), during zinc withdrawal (without cycloheximide) as described in Fig. 4C (C and D), and for zinc-shock responsiveness to 1,000 μM ZnCl2 as described in Fig. 4A (E and F). (Insets) ZF3/4-FRET assayed over longer time. (A–D) Values normalized to the maximum FRET ratio (n = 3; error bars represent ±1 SD). (E and F) Values normalized to the FRET ratio before zinc addition.

Discussion

FRET is very sensitive to the distance between the flanking fluorophores and, therefore, is useful to the study of conformational changes associated with metal binding to intervening domains. The apo forms of ZFs are thought to be largely unstructured (22), and zinc is required for forming the folded ZF domain. These considerations suggested that we could detect the conformational changes associated with zinc binding by using FRET. The laboratories of Berg (23) and Tsien (24) have both demonstrated the usefulness of FRET for observing Zn2+ binding to ZFs in vitro. Here, we show the application of this technique to characterize zinc binding to Zap1 ZF pair domains in vivo. We also tested a FRET reporter containing a single finger domain, the consensus CP-1 ZF peptide (25), and the FRET response in vivo was similar to the finger-pair domains that were used here (data not shown).

In vitro studies of zinc binding by an S. cerevisiae Zap1 ZF1/2 peptide showed that these fingers could bind Zn2+ (12, 14). Zinc binding by ZF1/2 in vivo was indicated by the high FRET intensity observed in zinc-replete vs. -limited cells. This FRET response depended on the zinc-binding ligands of ZF1 and ZF2. Also, the FRET response depended on the finger–finger interaction that occurs between ZF1 and ZF2. No partial increase in FRET intensity was observed when the finger interaction was mutationally disrupted, suggesting that cooperativity of zinc binding may occur between the fingers. Whereas most tandem ZFs are thought to bind zinc and fold independently of each other (26), the ZF pair interaction that occurs between ZF1 and ZF2 could foster a cooperative binding/folding process (14).

The remarkable lability of ZF1/2 zinc binding observed in vitro was also detected in vivo. After the transition from replete to deficient conditions, zinc occupancy of the ZF1/2 finger pair steadily decreased. The rate at which zinc is lost from these fingers is likely slowed by the mobilization of vacuolar zinc stores that occurs during this transition to zinc deficiency. This mobilization was indicated by the observation that the loss of FRET intensity during zinc withdrawal occurs more rapidly in cells bearing zrc1 and cot1 mutations that reduce vacuolar zinc storage (data not shown) (27). Although other factors in the cell may be catalyzing removal of zinc from ZF1/2, we propose that the lability of these fingers in vivo is an intrinsic property of their metal-binding properties, as observed in vitro.

Although other models are possible, the decrease in metallated ZF1 and ZF2 during zinc withdrawal may be due to competition for zinc with newly synthesized metalloproteins. Our results suggest that zinc is constantly bound and released by ZF1 and ZF2. The released zinc may then be bound by apometalloproteins and no longer available to Zap1. This concept fits well with how we believe a metal-sensing regulator might function in vivo. The kinetics of zinc binding by ZF1 and ZF2 is also apparent in their ability to bind zinc during zinc shock. Under zinc-shock conditions, ZF1 and ZF2 can bind newly imported zinc within seconds. Thus, the ZF1/2 domain can accumulate in an unmetallated form in zinc-deficient cells that is competent for rapid zinc binding. Again, this property makes sense for a metalloregulatory protein in that it can respond quickly to sudden increases in cellular zinc levels.

ZF3 and ZF4 of the Zap1 DNA binding domain likely form a similar ZF pair structure as that described for ZF1 and ZF2. ZF1/2 and ZF3/4 peptides bind zinc with similar affinities in vitro (12). In vivo, we observed that zinc occupancy of the different finger pairs was similar when cells were grown for several hours in a medium containing a range of zinc concentrations. This observation is consistent with the similar affinities of these domains for zinc. However, as in vitro, these finger pairs showed very different behavior in their kinetics of zinc binding and release in vivo. In zinc withdrawal, although zinc was lost from ZF1 and ZF2 over a 5-h period, there was no change in the FRET intensity of the ZF3/4-FRET sensor over the same period. This finding suggests that ZF3 and ZF4 are much more resistant to the loss of zinc to other metalloproteins. Also, under zinc-shock conditions, the ZF3/4 finger pair was slow in binding newly imported zinc. Thus, whereas ZF1 and ZF2 accumulate in the cell in a form capable of rapid zinc binding, this behavior is not true for ZF3 and ZF4. These behaviors were also found for the ZF1/2 and ZF3/4 domains of the C. albicans Zap1 ortholog. The Zap1 ZF1/2 domains of S. cerevisiae and C. albicans differ greatly in their amino acid sequences (Fig. 1D) but have conserved lability of binding and the ability to accumulate in the cell in an apo form that can rapidly bind zinc. The conserved stability of zinc binding by the ZF3/4 pair in vivo is more consistent with its structural role in the DNA-binding domain.

Several studies have sought to use zinc-responsive fluorescent probes to estimate the free or labile zinc levels within cells. For example, Bozym et al. (28) argued that the zinc concentration in PC-12 cells is ≈5 pM, based on a FRET reporter that is analogous to ours by using carbonic anhydrase as the zinc sensor domain. However, these studies are difficult to interpret because, at high probe concentrations, the dose–response of the probe is dominated by the intracellular fluorophore concentration and not by the free zinc level (29). Similar studies using small molecule zinc probes are subject to the same caveat. Therefore, although a determination of free Zn2+ concentrations within cells would be a major breakthrough, neither ours nor any other current fluorophore technology of which we are aware is sufficient to achieve this important goal.

ZF-FRET proteins and related probes provide tools to study zinc homeostasis and metabolism in living cells. These sensors can be used to map the residues within ZF1 and ZF2 that confer zinc-binding lability. These sensors can be used also to examine the impact of oxidative stress or other in vivo factors on the ability of ZFs to bind their zinc. Also, it may be possible to target these proteins to specific subcellular compartments, such as the nucleus and mitochondria, to examine the zinc status of subcellular compartments. Last, it should be possible to express these proteins in the cells of other organisms to address similar questions of zinc metabolism and homeostasis.

Materials and Methods

Yeast Strains and Culture Conditions.

The yeast strains DY1457 (MATα ade6 can1 his3 leu2 trp1 ura3), ZHY1 (DY1457 zrt1Δ::LEU2), and ZHY6 (DY1457 zap1Δ::TRP1) were grown in YP medium supplemented with 2% glucose (YPD) or synthetic-defined medium supplemented with 2% glucose. LZM was prepared as described in ref. 30. LZM is zinc-limiting because it contains 1 mM EDTA and 20 mM Na3-citrate to control metal availability. ZHY1 and ZHY6 were grown in LZM that was prepared without EDTA (LZM − EDTA) because of their zinc-uptake defects. ZnCl2 was used as the zinc salt for all experiments.

Plasmid Constructs.

pYFP-CFP was constructed by inserting an XhoI- and BamHI-digested PCR fragment from pDH18 (18) into XhoI- and BamHI-cut pDR195 (31). pZF1/2-FRET, pZF3/4-FRET, and pZF1/2-C2Q2-FRET were generated by first partially digesting pYFP-CFP with NarI. PCR fragments of ZAP1 were then cloned between eYFP and eCFP by homologous recombination (32). To simplify additional plasmid constructions, pYFP-LEU2-CFP was generated by homologous recombination. This plasmid has the LEU2 gene inserted between eYFP and eCFP and flanked by NotI and SacII sites. Inserts were cloned into NotI- and SacII-digested pYFP-LEU2-CFP. This method was used for constructing the interface mutant plasmids pZF1/2-K582A/W583A-FRET, pZF1/2-W620A/E621K-FRET, and pZF1/2-L600D/V605D-FRET. The template for all mutant constructs was pGBD-Zap1552–705 bearing the indicated mutation (12).

β-Gal Assays.

The following reporter plasmids were used in this study: pDg2L (i.e., ZRE-lacZ) (19), pGAL1-lacZ, and pGBD-Zap1552–705 (33). Cells were grown for 14–20 h to exponential phase (OD600, 0.3–0.6), and β-gal activity was measured as described in ref. 34.

FRET Measurements.

Cells were grown in media for ≈16 h to exponential phase. Cultures were then harvested for assay or were shifted to LZM supplemented with 1 μM Zn for the Zn withdrawal experiments. All cultures were adjusted to the same cell density and kept chilled throughout processing. Cells were harvested by centrifugation and washed twice with 5 ml of FRET buffer [1 mM EDTA/2% glucose/20 mM Na3 citrate/5 mM MgSO4/1 mM NaCl/1 mM CaCl2/38 mM (NH4)2SO4/1 mM KH2PO4, pH 4.7]. Cells were then resuspended in FRET buffer to an OD600 of 0.4. The eYFP, eCFP, and FRET emission intensities were measured by spectrofluorometry with the following excitation (Ex) and emission (Em) wavelengths: eYFP, 500 nm (Ex)/545 nm (Em); eCFP, 440 nm (Ex)/480 (Em); and FRET, 440 nm (Ex)/545 nm (Em). Three readings were taken for each wavelength combination, and the mean value was used for subsequent calculations. In Zn-shock experiments, the cells were suspended in FRET buffer at 30°C throughout the experiment. The FRET ratio was determined by dividing the FRET emission intensity by the eCFP emission intensity (18).

Immunoblotting and Microscopy.

Total protein extracts were prepared by trichloroacetic acid precipitation. SDS/PAGE fractionation was performed with a 12% acrylamide resolving gel. Proteins were detected by using mouse anti-GFP (Sigma) or anti-Pgk1 (Molecular Probes) mAbs and ECL Plus chemiluminescence (Amersham Biosciences). Fluorescence microscopy was performed by using filter sets for DAPI and GFP.

Acknowledgments

We thank Mark Milanick for assistance with fluorometry and Wolf Frommer (Stanford University, Stanford, CA) and Trisha Davis (University of Washington, Seattle) for providing plasmids. This work was supported by National Institutes of Health Grants R01-GM056285 and R01-GM069786.

Abbreviations

AD

activation domain

eYFP

enhanced yellow fluorescent protein

eCFP

enhanced cyan fluorescent protein

LZM

low-zinc medium

ZF

zinc finger

ZRE

zinc-responsive element.

Footnotes

Conflict of interest statement: No conflicts declared.

This paper was submitted directly (Track II) to the PNAS office.

References

  • 1.Lander E. S., Linton L. M., Birren B., Nusbaum C., Zody M. C., Baldwin J., Devon K., Dewar K., Doyle M., FitzHugh W., et al. Nature. 2001;409:860–921. doi: 10.1038/35057062. [DOI] [PubMed] [Google Scholar]
  • 2.Rhodes D., Klug A. Sci. Am. 1993;268:56–65. doi: 10.1038/scientificamerican0293-56. [DOI] [PubMed] [Google Scholar]
  • 3.Brown R. S. Curr. Opin. Struct. Biol. 2005;15:94–98. doi: 10.1016/j.sbi.2005.01.006. [DOI] [PubMed] [Google Scholar]
  • 4.Mackay J. P., Crossley M. Trends Biochem. Sci. 1998;23:1–4. doi: 10.1016/s0968-0004(97)01168-7. [DOI] [PubMed] [Google Scholar]
  • 5.Berg J. M., Godwin H. A. Ann. Rev. Biophys. Biomol. Struct. 1997;26:357–371. doi: 10.1146/annurev.biophys.26.1.357. [DOI] [PubMed] [Google Scholar]
  • 6.Buchsbaum J., Berg J. Inorg. Chim. Acta. 2000;297:217–219. [Google Scholar]
  • 7.Lachenmann M. J., Ladbury J. E., Qian X., Huang K., Singh R., Weiss M. A. Protein Sci. 2004;13:3115–3126. doi: 10.1110/ps.04866404. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Giedroc D. P., Chen X., Pennella M. A., LiWang A. C. J. Biol. Chem. 2001;276:42322–42332. doi: 10.1074/jbc.M106517200. [DOI] [PubMed] [Google Scholar]
  • 9.Lyons T. J., Gasch A. P., Gaither L. A., Botstein D., Brown P. O., Eide D. J. Proc. Natl. Acad. Sci. USA. 2000;97:7957–7962. doi: 10.1073/pnas.97.14.7957. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Bird A. J., Evans-Galea M., Blankman E., Zhao H., Luo H., Winge D. R., Eide D. J. J. Biol. Chem. 2000;275:16160–16166. doi: 10.1074/jbc.M000664200. [DOI] [PubMed] [Google Scholar]
  • 11.Pavletich N. P., Pabo C. O. Science. 1993;261:1701–1707. doi: 10.1126/science.8378770. [DOI] [PubMed] [Google Scholar]
  • 12.Bird A., McCall K., Kramer M., Blankman E., Winge D., Eide D. EMBO J. 2003;22:5137–5146. doi: 10.1093/emboj/cdg484. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Herbig A., Bird A. J., Swierczek S., McCall K., Mooney M., Wu C. Y., Winge D. R., Eide D. J. Mol. Microbiol. 2005;57:834–846. doi: 10.1111/j.1365-2958.2005.04734.x. [DOI] [PubMed] [Google Scholar]
  • 14.Wang Z., Feng L., Matskevich V., Venkataraman K., Parasuram P., Laity J. J. Mol. Biol. 2006;357:1167–1183. doi: 10.1016/j.jmb.2006.01.010. [DOI] [PubMed] [Google Scholar]
  • 15.Miyawaki A., Llopis J., Heim R., McCafferty J. M., Adams J. M., Ikura M., Tsien R. Y. Nature. 1997;388:882–887. doi: 10.1038/42264. [DOI] [PubMed] [Google Scholar]
  • 16.Pearce L. L., Gandley R. E., Han W., Wasserloos K., Stitt M., Kanai A. J., McLaughlin M. K., Pitt B. R., Levitan E. S. Proc. Natl. Acad. Sci. USA. 2000;97:477–482. doi: 10.1073/pnas.97.1.477. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Hong S. H., Hao Q., Maret W. Protein Eng. Des. Sel. 2005;18:255–263. doi: 10.1093/protein/gzi031. [DOI] [PubMed] [Google Scholar]
  • 18.Hailey D. W., Davis T. N., Muller E. G. Methods Enzymol. 2002;351:34–49. doi: 10.1016/s0076-6879(02)51840-1. [DOI] [PubMed] [Google Scholar]
  • 19.Zhao H., Butler E., Rodgers J., Spizzo T., Duesterhoeft S., Eide D. J. Biol. Chem. 1998;273:28713–28720. doi: 10.1074/jbc.273.44.28713. [DOI] [PubMed] [Google Scholar]
  • 20.Zhao H., Eide D. Proc. Natl. Acad. Sci. USA. 1996;93:2454–2458. doi: 10.1073/pnas.93.6.2454. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.MacDiarmid C. W., Milanick M. A., Eide D. J. J. Biol. Chem. 2003;278:15065–15072. doi: 10.1074/jbc.M300568200. [DOI] [PubMed] [Google Scholar]
  • 22.Miura T., Satoh T., Takeuchi H. Biochim. Biophys. Acta. 1998;1384:171–179. doi: 10.1016/s0167-4838(98)00015-6. [DOI] [PubMed] [Google Scholar]
  • 23.Godwin H. A., Berg J. M. J. Am. Chem. Soc. 1996;118:6514–6515. [Google Scholar]
  • 24.Shaner N. C., Campbell R. E., Steinbach P. A., Giepmans B. N. G., Palmer A. E., Tsien R. Y. Nat. Biotechnol. 2004;22:1567–1572. doi: 10.1038/nbt1037. [DOI] [PubMed] [Google Scholar]
  • 25.Kim C. A., Berg J. M. Nature. 1993;362:267–270. doi: 10.1038/362267a0. [DOI] [PubMed] [Google Scholar]
  • 26.Krizek B. A., Zawadzke L. E., Berg J. M. Protein Sci. 1993;2:1313–1319. doi: 10.1002/pro.5560020814. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.MacDiarmid C. W., Gaither L. A., Eide D. EMBO J. 2000;19:2845–2855. doi: 10.1093/emboj/19.12.2845. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Bozym R. A., Thompson R. B., Stoddard A. K., Fierke C. A. Am. Chem. Soc. Chem. Biol. 2006;1:103–111. doi: 10.1021/cb500043a. [DOI] [PubMed] [Google Scholar]
  • 29.Dineley K. E., Malayandi L. M., Reynolds I. J. Mol. Pharmacol. 2002;62:618–627. doi: 10.1124/mol.62.3.618. [DOI] [PubMed] [Google Scholar]
  • 30.Gitan R. S., Luo H., Rodgers J., Broderius M., Eide D. J. Biol. Chem. 1998;273:28617–28624. doi: 10.1074/jbc.273.44.28617. [DOI] [PubMed] [Google Scholar]
  • 31.Rentsch D., Laloi M., Rouhara H., Schmelzer E., Delrot S., Frommer W. FEBS Lett. 1995;370:264–268. doi: 10.1016/0014-5793(95)00853-2. [DOI] [PubMed] [Google Scholar]
  • 32.Ma H., Kunes S., Schatz P. J., Botstein D. Gene. 1987;58:201–216. doi: 10.1016/0378-1119(87)90376-3. [DOI] [PubMed] [Google Scholar]
  • 33.Bird A. J., Zhao H., Luo H., Jensen L. T., Srinivasan C., Evans-Galea M., Winge D. R., Eide D. J. EMBO J. 2000;19:3704–3713. doi: 10.1093/emboj/19.14.3704. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Guarente L. Methods Enzymol. 1983;101:181–191. doi: 10.1016/0076-6879(83)01013-7. [DOI] [PubMed] [Google Scholar]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES