Identification of a copper‐induced intramolecular interaction in the transcription factor Mac1 from Saccharomyces cerevisiae

Mac1 is localized within the nucleus in Cu‐deficient and Cu‐replete yeast

The possibility that Cu modulation of Mac1 activity involves regulation of nuclear localization was re‐investigated. The strategy was to fuse Mac1 with the green fluorescent protein (GFP) and monitor the distribution of the GFP fluorescence. A vector containing GFP fused to the 3′ end of the MAC1 ORF controlled by the ADH1 promoter was transformed into YJJ1 cells lacking a functional Mac1. The Mac1–GFP fusion is a functional protein in that it activates expression of a CTR1–lacZ reporter gene in cells deficient in Cu due to growth in the presence of the Cu(I)‐specific chelator bathocuproine sulfonate (BCS) (Figure 1A). Addition of excess copper to the medium results in the loss of activation, as is seen in wild‐type MAC1 cells (Figure 1A).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Cellular localization of Mac1. (A) A C‐terminal fusion of Mac1 with the GFP (Mac1–GFP) activates expression of a CTR1–lacZ reporter gene. (B) Fluorescence of cells harboring the Mac1–GFP fusion cultured in Cu‐deficient (30 μM bathocuproine sulfonate, BCS) and Cu‐replete (100 μM CuSO₄) medium. Cells harboring the fusion gene and GFP under the control of the GAL1 promoter were grown at 30°C in synthetic complete media containing galactose in place of dextrose. GFP was excited at 488 nm and detected at 520 nm; DAPI was excited at 345 nm and detected at 465 nm.

To evaluate localization of the Mac1–GFP, the fusion gene was placed under the control of the GAL1 promoter and transformed into the ctr1Δ strain YPH499. Fluorescence in cells expressing the Mac1–GFP fusion was concentrated into single spots that stained positively with the DNA‐stain reagent 4′,6′‐diamidino‐2‐phenylindole dihydrochloride (DAPI). The co‐localization of Mac1–GFP and DNA suggests that Mac1–GFP is localized within the nuclei of the cells under both Cu‐limiting and Cu‐replete conditions (Figure 1B). As a control, cells harboring only GFP showed uniform fluorescence throughout the cell, independent of the Cu levels in the growth medium. It is unlikely that the nuclear localization of Mac1 in Cu‐replete cells is an artifact of overexpression of Mac1–GFP. First, the activity of the fusion protein shows the same fold Cu inhibition as does the chromosomally encoded Mac1. Secondly, there was no evidence of low‐abundance diffuse cytoplasmic fluorescence in these cells.

Mac1 is a Cu‐binding protein

Cu regulation of Mac1 activity within the nucleus may arise from direct interaction of Cu ions with Mac1. To test whether Mac1 is a Cu‐binding protein, a MAC1–GST fusion gene was constructed for bacterial expression of Mac1–GST (C‐terminal GST). The fusion gene was active in yeast and the activity was Cu modulated (data not shown). The expressed fusion protein was recovered as an insoluble inclusion body. The particulate material isolated from Escherichia coli cultured in Cu‐containing medium was highly emissive upon ultraviolet irradiation. Orange luminescence is indicative of Cu(I) coordination. The luminescent particulate material was dependent on induction of the Mac1–GST fusion. Co‐transformation of bacteria with the vector containing the MAC1–GST fusion and a second vector containing the E.coli thioredoxin vector resulted in limited expression of soluble Mac1–GST. The fusion protein, purified by affinity chromatography on glutathione–Sepharose, gave a single band on SDS–PAGE which was consistent with the calculated molecular mass of 73 kDa. Metal analysis on the purified fusion protein revealed 7.9 ± 0.2 mol eq. bound Cu and 1.9 ± 0.1 mol. eq. Zn (Table I). The two Zn ions are presumably bound within the N‐terminal DNA‐binding domain, since the N‐terminal 159 residues of Mac1 were shown previously to bind 2 mol. eq. Zn(II) (Jensen et al., 1998). The Cu content of the Mac1^up1 mutant containing a H279G substitution was only 4.4 ± 0.2 mol. eq. Significantly less soluble protein was isolated from cells cultured in low Cu‐containing medium, suggesting that Cu ions stabilized the expressed protein.

The candidate Cu‐binding domain of Mac1 consists of the two Cys‐rich motifs comprising residues 252–341. This peptide expressed in E.coli was also isolated within insoluble inclusion bodies. Co‐expression with thioredoxin resulted in the production of soluble protein. Glutathione–Sepharose was used to affinity purify Mac1(252–341) and H279Q Mac1(252–341) peptides as N‐terminal GST fusions. The purified fusion proteins gave a single band by SDS–PAGE consistent with the calculated molecular mass of 36 kDa. Metal analysis revealed that the GST–Mac1(252–341) fusion protein contained 7.7 ± 0.4 bound Cu ions and, as expected, no bound Zn ions (Table I). The metal content of the GST–H279Q Mac1(252–341) fusion was determined to be 4.5 ± 0.5 Cu ions and, unexpectedly, two Zn ions per protein (Table 1).

The Cu complex of Mac1(252–341) was luminescent with an emission maximum near 570 nm (Figure 2). Similar emission is observed in a series of proteins with polycopper thiolate clusters (Winge et al., 1994). The emission is indicative that Mac1(252–341) and, presumably, Mac1 bind copper ions as Cu(I). With equal concentrations of copper, the emission of the H279Q protein was lower than that seen for wild‐type Mac1(252–341). The observed lower emission in the mutant molecule suggests greater solvent quenching of the Cu‐thiolate fluorophore that may arise from a less stably folded conformer.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Luminescence of CuMac1(252–341). The Cu(I) luminescence of purified Mac1(252–341) and H279Q Mac1(252–341) was measured using equimolar Cu(I) concentrations (100 nmol Cu/ml). Excitation was at 300 nm, and emission was scanned from 510–620 nm with maximal emission occurring near 570 nm.

Mac1 is stabilized in Cu‐replete cells

We attempted to show that Mac1 is a Cu‐binding protein in Cu‐replete yeast using affinity purification of Mac1–GST. Although we were unable to purify the fusion protein, we obtained indirect evidence using Western analysis showing that Cu ions enhanced the in vivo stability of Mac1. Yeast expression vectors were constructed to express either MAC1 or MAC1^up1 with 3′ myc epitope tags. The presence of the C‐terminal myc tag did not impair Mac1 function. MAC1–myc epitope fusions under the control of the ADH1 promoter were co‐transformed with the CTR1–lacZ reporter into the mac1‐1 strain YJJ1. Both Mac1–myc and H279Q Mac1–myc stimulated lacZ expression under Cu‐deficient conditions (Figure 3A). With the addition of excess copper to the growth medium, β‐galactosidase activity was substantially reduced in cells expressing Mac1–myc. In contrast, β‐galactosidase activity of cells expressing H279Q Mac1–myc was unchanged (Figure 3A).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Copper‐induced stabilization of Mac1. (A) Fusions of a myc epitope to either Mac1 or Mac1^up1 activate expression of a CTR1–lacZ reporter gene. Expression of Mac1–myc (B) or Mac1^up1–myc (C) was induced with galactose for 3 h prior to addition of glucose. Cultures were then split; one aliquot was untreated and 100 μM CuSO₄ was added to the other. Samples from each culture were removed at 45 min intervals. Cell extracts were separated by SDS–PAGE, transferred to nitrocellulose and then immunoblotted with a monoclonal anti‐c‐myc antibody and an anti‐actin antibody (to control for gel loading).

To assess protein stability in vivo, MAC1 and MAC1^up1 with 3′ myc epitope tags were subcloned into an expression vector controlled by the GAL1 promoter. The fusion genes were preinduced by a short, 3 h incubation in galactose‐containing medium. Glucose was then added to repress further transcription. Whole‐cell extracts were examined at various times after the addition of glucose (± CuSO₄) by immunoblotting, using a monoclonal anti‐c‐myc antibody. Under Cu‐deficient conditions Mac1 was significantly degraded within 90 min, but in the presence of elevated levels of CuSO₄, remained stable for at least 3 h (Figure 3B). The increase in stability of Mac1 was also seen in wild‐type cells cultured in standard medium or in ctr1Δ cells cultured in as little as 10 μM CuSO₄ (data not shown). The H279Q Mac1^up1 mutant protein was significantly more stable than the wild‐type molecule under Cu‐deficient conditions, although the turnover of the H279Q Mac1 protein was unaffected by exposure to elevated levels of CuSO₄ (Figure 3C).

Cu induces an intramolecular interaction in Mac1

In vivo DNA‐binding and transactivation activities of Mac1 are inhibited in Cu‐replete cells (Georgatsou et al., 1997; Graden and Winge, 1997; Labbe et al., 1997). The two activities are independent and separable. The minimal DNA‐binding domain (DBD) of Mac1 consists of residues 1–159 (Jensen et al., 1998). Although the exact boundaries of the transactivation domain are not known, residues 240–417 contain transactivation activity (Graden and Winge, 1997). Neither of these minimal modules are regulated by Cu ions. Cu‐modulation of transactivation activity requires sequences upstream of the actual activation domain (AD) (Graden and Winge, 1997). These results led to the hypothesis that Cu binding to Mac1 may result in direct inhibition of both activities through a Cu‐induced intramolecular interaction.

To test for an intramolecular interaction between the N‐terminal DBD and the C‐terminal activation domain, a two‐hybrid‐based assay was used. The Mac1 DBD (residues 1–159) was fused to the minimal Gal4 DBD (lacking any transactivation activity) in the first expression vector [Mac1(1–159)–Gal4]. The second vector consisted of the Mac1 AD fused to the VP16 activation domain of the herpes simplex virus (Sadowski et al., 1988) (Figure 4). Two constructs were engineered for the AD vector, one consisting of Mac1(240–417), and the second H279Q Mac1(240–417). The Mac1(1–159)–Gal4 DBD vector was co‐transformed into yeast strain Y190 with AD vectors encoding Mac1(240–417)–VP16, H279Q Mac1(240–417)–VP16 or the empty plasmid. Expression of the GAL1–lacZ reporter gene was assessed.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Identification of a Cu‐induced intramolecular interaction in Mac1. An intramolecular interaction between the N‐terminal DNA‐binding domain (DBD) and the C‐terminal activation domain of Mac1 was identified using a two‐hybrid approach. The Mac1 DBD Mac1(1–159) was fused to the Gal4 DBD, while Mac1(240–417) and H279Q Mac1(240–417) were fused to the VP16 activation domain. The dots shown in the fusion construct are cysteinyl residues in Mac1. Cells were grown under Cu‐deficient conditions (30 μM BCS), or with excess copper added to the growth media. The β‐galactosidase activity of a GAL1–lacZ reporter was measured in each case. Each sample was assayed in triplicate.

If a Cu‐induced Mac1 intramolecular interaction occurs, Cu may promote the heterodimerization of the Mac1 DBD and AD expressed as separate fusion proteins and yield expression of lacZ. Therefore, β‐galactosidase activity would be observed in Cu‐treated cells containing both Mac1(1–159)–Gal4 and Mac1(240–417)–VP16. No β‐galactosidase activity would be expected in cells containing Mac1(1–159)–Gal4 and H279Q Mac1(240–417)–VP16 as the H279Q substitution in Mac1^up1 abrogates Cu‐repression (Jungmann et al., 1993; Graden and Winge, 1997; Labbe et al., 1997).

β‐galactosidase was assayed in co‐transformants cultured with either the Cu‐specific chelator BCS added to generate Cu‐deficient conditions or with CuSO₄ added to the growth media. Only low levels of β‐galactosidase were observed in Cu‐deficient cells (Figure 4). However, with the addition of CuSO₄, substantial β‐galactosidase activity was observed in cells containing Mac1(1–159)/Gal4 and Mac1(240–417)/VP16. As predicted, the Cu‐induced interaction was precluded with either H279Q Mac1(240–417)VP16 or the VP16 control plasmid.

β‐galactosidase activity in Cu‐treated cells with the latter two combinations was no greater than the activity observed for Cu‐deficient cells (BCS‐treated) (Figure 4).

Fusion of the minimal DBD and AD segments creates a functional Mac1

To determine whether a functional copper‐regulated Mac1 could be generated by using the minimal DBD and Cu‐binding domains, the minimal Mac1 DBD (residues 1–159) was fused in‐frame to the minimal Cu‐binding activation domain (residues 252–341) yielding Mac1‐ (DBD+AD). This minimal Mac1(DBD+AD) contains 248 residues, unlike the 417 residues in the intact Mac1. The ability of MAC1 (DBD+AD) and wild‐type MAC1 both under control of the ADH1 promoter to activate a CTR1–lacZ reporter were compared (Figure 5). The activity of Mac1(DBD+AD) was similar to that of wild‐type Mac1. Mac1(DBD+AD) expressed high levels of β‐galactosidase under Cu‐deficient conditions. The level of β‐galactosidase expression was substantially reduced in Cu‐replete cells.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Minimal Mac1. A minimal Mac1 construct was generated by fusing the Mac1 minimal DNA‐binding domain (DBD) (residues 1–159) with the Cu‐binding activation domain (residues 252–341) and placed under control of the ADH1 promoter and terminator. The dots in the scheme of the constructs represent cysteinyl residues in Mac1. β‐galactosidase activities of Mac1 and Mac1(DBD+AD) were assayed in the mac1‐1 strain YJJ1 using a CTR1–lacZ reporter. Each sample was assayed in triplicate.

Strains and culture conditions

The strains used in this study include YPH499Δ ctr1 (MATa, ade2‐101, his3‐Δ200, leu2‐Δ1, lys2‐801, trp1‐Δ63, ura3‐52, Δctr1::LEU2) (Dancis et al., 1994c), YW02 (MATa his3‐Δ200 leu2‐3,112, lys2‐801 trp1‐1, ura3‐52), YJJ1 (MATa, his3‐Δ200, leu2‐3, 2–112, lys2‐801, trp1‐1, ura3‐52, mac1‐1) (Jungmann et al., 1993) and Y190 (MATa, ade2‐101, cyh2, his3, leu2‐3,112, trp1‐901, ura3‐52, gal4, gal80, LYS2:pGAL‐HIS3, URA3:pGAL‐lacZ) (Bai and Elledge, 1997). DNA transformations were performed using the lithium acetate procedure. Cells were propagated either in YPD or synthetic complete medium plus dextrose agar plates. Transformed cells were grown in liquid complete medium lacking the appropriate nutrients.

Plasmid construction

Mac1–GFP fusions. Sequences coding for the GFP were amplified by PCR using vector pEGFP–C1 (Clontech) as a template creating 5′ BamHI and ClaI sites and 3′ NotI and SalI sites. The PCR product was digested with BamHI–NotI and subcloned into vector pYeF2 which uses the GAL1 promoter generating GFP–pYeF2. MAC1 and MAC1^up1 were amplified by PCR creating 5′ BamHI and AflIII sites, 3′ ClaI and HindIII sites, and removing the stop codon. The PCR product was digested BamHI–HindIII, subcloned into pHolly (a pBluescript derivative kindly provided by R.D.Palmiter with a small insert creating sites for NdeI, NcoI and StuI) and sequenced. MAC1 was excised from pHolly by digestion with BamHI and ClaI and subcloned into GFP–pYeF2 generating Mac1GFP–pYeF2.

Mac1–myc tag fusions. Six copies of the c‐myc epitope were excised from vector CS2+MT (Rupp et al., 1994) as a BamHI–StuI fragment and subcloned into pHolly to pick up a stop codon and a 3′ SalI site. The myc epitope was then removed from pHolly as a BamHI–SalI fragment and ligated into GFP–pYeF2 digested BamHI–SalI, thus removing sequences for GFP. MAC1 and MAC1^up1 were excised from pHolly by digestion with BamHI and ClaI and subcloned into Myc–pYeF2 generating Mac1–myc–pYeF2 and H279Q Mac1–myc–pYeF2.

pVT102‐U vectors. The MAC1 ORF was PCR‐amplified with 5′ BamHI and 3′ SalI sites retaining the stop codon. The PCR product digested with BamHI–SalI was subcloned into pHolly and sequenced. MAC1, MAC1–myc and MAC1–GFP were excised from pHolly using BamHI and SalI and subcloned into pVT102‐U resulting in vectors Mac1–pVT102‐U, Mac1–myc–pVT102‐U and Mac1–GFP–pVT102‐U. GST was excised from pAED4–GST as a ClaI–NotI fragment and subcloned into Mac1–myc–pVT102‐U replacing the myc‐tag with sequences for GST generating Mac1–GST–pVT102‐U. Sequences coding for codons 252–341 of MAC1 were amplified by PCR, creating a 5′ BamHI and NotI sites and 3′ SalI and HindIII sites and stop codon after codon 341. The PCR product was subcloned into pVT102‐U using BamHI and HindIII sites generating vector Mac1(252–341)pVT102‐U. MAC1 (codons 1–159) was excised from Mac1(1–159)–myc (Jensen et al., 1998) as a BamHI–NotI fragment and subcloned into Mac1(252–341)pVT102‐U generating Mac1(DBD+AD)pVT102‐U. The expression of constructs in vector pVT102‐U is constitutive under the control of the ADH1 promoter.

Escherichia coli expression vectors. MAC1 and MAC1^up1 were excised from pHolly by digestion with NcoI and ClaI and subcloned into a T7‐based vector Mac1(1–159) pAED4–GST (Jensen et al., 1998) to generate C‐terminal GST fusion expression vectors. The codons for the first 24 amino acids were optimized for expression in E.coli (Anderson and Kurland, 1990). DNA sequences coding for residues 252–341 of Mac1 as well as H279Q Mac1 were amplified by PCR, generating a 5′ BamHI site with a ATG start codon and a 3′ NotI site with an upstream stop codon. The PCR products were subcloned into pGEX‐4T‐1 (Pharmacia) to generate N‐terminal GST fusions.

Two‐hybrid vectors. Sequences coding for codons 240–417 of the MAC1 ORF were PCR‐amplified. The PCR products contained 5′ BamHI and 3′ ClaI sites and the stop codon was removed. The PCR products were subcloned into vector pVT–VP16 (Jensen et al., 1998), generating vectors Mac1(240–417)–VP16 and H279QMac1(240–417)–VP16 with a fusion of Mac1(240–417) with the minimal activation domain (AD) from the herpes simplex virus‐1 VP16 protein (Sadowski et al., 1988). Sequences coding for amino acids 1–159 were excised from vector Mac1(1–159)–VP16 as a BamHI–XhoI fragment and subcloned into pBG4D‐1 cut BamHI–SalI, destroying both the SalI and XhoI sites creating vector Mac1(1–159)pBG4D‐1.

Immunofluorescence

Mac1–GFP–pYeF2 and GFP–pYeF2 in YPH499Δctr1 were grown in low‐copper synthetic complete media lacking uracil with 2% raffinose. Cells were grown at 30°C to OD₆₀₀ of 1.0, diluted 100‐fold and then split into three, 10 ml aliquots. One aliquot was the untreated control, galactose (2% final) and 30 μM BCS (Cu‐specific chelator bathocuproine sulfonate) were added to the second, and galactose and 100 μM CuSO₄ were added to the third. After growth to OD₆₀₀ of 1.0, 0.5 ml of 37% formaldehyde was added and the cells were incubated for an additional 10 min. The cells were harvested, resuspended in 5 ml of buffer A (40 mM KPO₄ pH 6.5, 0.5 M MgCl₂), plus 0.5 ml of 37% formaldehyde and grown for an additional 1 h. The cells were then washed twice in buffer A and once in buffer A plus 1.2 M sorbitol. Spheroplasts were prepared and spotted onto polylysine‐coated slides and incubated at 25°C for 10 min. Mount solution (1 mg/ml p‐phenylenediamine, 20 mM sodium carbonate pH 9, 90% glycerol, 50 ng/ml DAPI) was overlayed on the slides. Cells were viewed using a fluorescence microscope. The GFP was excited at 488 nm and detected at 520 nm; DAPI was excited at 345 nm and detected at 465 nm.

Western blot analysis

Strain YPH499Δctr1 transformed with either Mac1–myc–pYeF2 or H279Q Mac1–myc–pYeF2 under the control of the GAL1 promoter were grown in low‐copper synthetic complete media lacking uracil containing 2% raffinose in place of dextrose to an OD₆₀₀ of 1.0. Mac1–myc and H279Q Mac1–myc expression was induced by the addition of galactose (2%) to the growth medium. Cells were grown for 3 h, at which time glucose was added to 2% to stop induction, and the culture was split. To one aliquot was added 100 μM CuSO₄ (at the same time as glucose), while the other aliquot was untreated. At 45 min intervals, 10 ml of each culture was removed and cells were harvested, washed and flash‐frozen. The cells were lysed with glass beads and centrifuged at 4°C for 5 min at 10 000 g. Total protein was quantified by the method of Bradford (1976). SDS–PAGE was carried out with 20 μg protein using a 7.5% gel. Proteins were transferred to nitrocellulose and the blots incubated with monoclonal anti‐c‐myc and anti‐actin antibodies in blocking buffer (10% non‐fat dry milk, phosphate‐buffered saline, 0.1% Tween‐20). Detection by enhanced chemiluminescence (ECL) was performed after incubation with a horseradish peroxidase‐conjugated anti‐mouse antibody.

Purification of Mac1 and Mac1(252–341)

Mac1 expressed in E.coli was insoluble and present in inclusion bodies, as was Mac1(252–341). Co‐expression of E.coli thioredoxin increases the solubility of some proteins expressed in E.coli (Lutsenko and Kaplan, 1995). Expression of either Mac1–GST or GST–Mac1(252–341) fusions with thioredoxin increased the level of soluble protein. GST fusions were co‐expressed with thioredoxin in BL21 (DE3) cells pregrown at 30°C to an OD₆₀₀ of 0.5 prior to the addition of IPTG. At 30 min after IPTG induction, 1.4 mM CuSO₄ was added and the cells were grown for an additional 4 h. The harvested cells were washed with 0.25 M sucrose and resuspended in buffer (20 mM NaH₂PO₄ pH 7.5, 250 mM NaCl, 10% glycerol, 5 mM DTT) for sonication. The lysate was clarified at 100 000 g for 30 min at 4°C. Triton X‐100 was added to 1% and the extract was purified on a glutathione–Sepharose column. The Mac1–GST fusion proteins were eluted with 20 mM glutathione in sonication buffer. The eluted proteins were monitored by SDS–PAGE and a single band at the expected molecular weight was observed for each fusion.

Protein and metal analyses

Protein samples were quantified by amino acid analysis following hydrolysis in 5.7 M HCl in vacuo. Metal analyses were performed on a Perkin‐Elmer Analyst 100 atomic absorption spectrometer. Cu(I) luminescence measurements were performed on a Perkin‐Elmer 650‐10S fluorometer with a band‐pass filter. Excitation was at 300 nm.

β‐galactosidase assays

Plasmids Mac1–pVT102‐U, Mac1–myc–pVT102‐U, H279Q Mac1–myc–pVT102‐U, Mac1–GFP–pVT102‐U and Mac1(DBD+AD)–pVT102‐U were co‐transformed with a CTR1–lacZ reporter into the mac1‐1 yeast strain YJJ1. Plasmids Mac1(240–417)/VP16, H279Q Mac1(240–417)–VP16 or pVT–VP16 were co‐transformed with Mac1(1–159)pBG4D‐1 into yeast strain Y190 which contains an integrated GAL–lacZ reporter gene. Yeast transformants were grown in low‐copper synthetic complete media with 2% dextrose and lacking either leucine and uracil for YJJ1 transformants or leucine and tryptophan for Y190 transformants. The cells were grown to an OD₆₀₀ of 1.0. The cultures were then diluted 100‐fold with the addition of either 30 μM BCS or 100 μM CuSO₄ added to the growth medium and grown to a final OD₆₀₀ of 1.0. Harvested cells were washed with buffer (85 mM Na₂HPO₄, 45 mM NaH₂PO₄, 10 mM KCl, 85 mM β‐mercaptoethanol, 1 mM MgSO₄) and frozen at −70°C. Cells were resuspended in the same buffer and lysed by vortexing with glass beads. β‐galactosidase activities were assayed using o‐nitrophenyl β‐d galactopyranoside as a substrate. The absorbance at 420 nm was recorded and protein concentrations were determined by the method of Bradford (1976). Each sample was assayed in triplicate.