The essential liaison of two copper proteins: the Cu-sensing transcription factor Mac1 and the Cu/Zn superoxide dismutase Sod1 in Saccharomyces cerevisiae

Although copper is an essential trace element for cell function and viability, its excess can lead to protein oxidation, DNA cleavage, and ultimate cell damage. Cells have established a variety of regulatory mechanisms to ensure copper ion homeostasis. In Saccharomyces cerevisiae, copper sensing and response to copper deficiency are regulated by the transcription factor Mac1. Our group has previously reported that in addition to copper, several chromatin proteins modulate Mac1 functionality. In this study, based on a synthetic growth deficiency phenotype, we showed that the Cu/Zn superoxide dismutase Sod1 plays an important role in Mac1 transcriptional activity, in unchallenged nutrient-rich growth conditions. Sod1 is a multipotent cytoplasmic and mitochondrial enzyme, whose main known function is to detoxify the cell from superoxide ions. It has been previously reported that Sod1 also enters the nucleus and affects the transcription of several genes, some of which are involved in copper homeostasis under Cu-depleted (Wood and Thiele in J Biol Chem 284:404-413, 2009) or only under specific oxidative stress conditions (Dong et al. Mol Cell Biol 33:4041-4050, 2013; Tsang et al. Nar Commun 8:3446, 2014). We have shown that Sod1 physically interacts with Mac1 transcription factor and is important for the transactivation as well as its DNA-binding activities. On the other hand, a constitutively active mutant of Mac1 is not affected functionally by the Sod1 ablation, pointing out that Sod1 contributes to the maintenance of the copper-unchelated state of Mac1. In conclusion, we showed that Sod1–Mac1 interaction is vital for Mac1 functionality, regardless of copper medium deficiency, in unchallenged growth conditions, and we suggest that Sod1 enzymatic activity may modify the redox state of the cysteine-rich motifs in the Mac1 DNA-binding and transactivation domains.


Introduction
Transition metal ions are essential in all organisms, being cofactors for many proteins responsible for vital biological functions, also involved in the biology of reproduction, the nervous and immune systems, in carcinogenesis, cell death, tissue regeneration, and epigenetic modifications. However, transition metal ions can also catalyze cytotoxic reactions, producing active free radicals (ROS) when in excess. Therefore, all organisms require mechanisms to tightly regulate the levels of these metal ions. Many human genetic diseases (e.g., Menkes, Wilson's, Parkinson's, Alzheimer's, myelodysplasias, and hemochromatosis) and severe malnutrition cause aberrant Cu/Fe homeostasis with detrimental health problems (Arnal et al. 2013;Bandmann et al. 2015;El-Youssef 2003;Gaggelli et al. 2006; González-Domínguez 1 3 et al. 2014; Kaler et al. 2008). Copper is included among these essential transition metals, and once inside the cells, it is incorporated as a catalytic or structural cofactor into a variety of proteins. In eukaryotes, including Saccharomyces cerevisiae, one of the primary ways in which copper levels are regulated is through changes in the expression of genes required for copper uptake, cellular distribution, storage, and export.
In S. cerevisiae, copper is transported in its cuprous form (Cu + ) following reduction by the metallo-reductases (Fre1, Fre7) and by high-affinity (Ctr1) and low-affinity (Fet4, Smf1) transporters (Vest et al. 2019). Once in the cytosol, copper is assembled, among other proteins, into the superoxide dismutase Sod1 by the copper chaperone Ccs1 (Furukawa et al. 2004;Yuan et al. 1995). The Cu/Zn superoxide dismutase (Sod1) is a homodimeric free radical scavenging enzyme that catalyzes the conversion of O 2 − to H 2 O 2 and O 2 (Miao and St. Clair, 2009). More recent evidence indicates that Sod1 acts as a regulatory protein for diverse cellular processes such as regulation of respiration in response to glucose and oxygen levels via interaction with the casein kinase pathway, it plays an important role in various human diseases, and starvation stimulates its activity in order to prevent oxidative damage and enhance survival and finally, it acts as a nuclear transcription factor to control the general oxidative stress transcriptional response (Che et al. 2016;Chung 2017;Glasauer et al. 2014;Reddi and Culotta 2013;Tsang et al. 2014Tsang et al. , 2018. Different yeast mutants with Sod1 deficiencies have shown amino-acid auxotrophies and reduced growth rate under aerobic conditions (Liu et al. 1992). Structurally, the Sod1 monomer consists of an eightstranded beta-barrel that binds a copper ion and a zinc ion that play both structural and catalytic roles (Tainer et al. 1983). Sod1 interacts with Ccs1 as a heterodimer. The structure of Ccs1 contains three defined domains: (1) βαββαβfolded Atx1-like domain, (2) central Sod1-like domain, and (3) carboxy-terminal domain (Lamb et al. 1999). Domains 1 and 3, bind Cu(I), whereas domain 2 is crucial for the interaction and activation of Sod1 (Culotta et al. 1997;Lamb et al. 2000;Schmidt et al. 2000). Ccs1 chaperone is required for Sod1 maturation and enzymatic activity unless cells lacking Ccs1 are cultured with high levels of copper ions (Boyd et al. 2019;Fetherolf et al. 2017;Schmidt et al. 1999). Both Sod1 and Ccs1 are widely distributed throughout the cytosol, mitochondrial intermembrane space, and nucleus (Sturtz et al. 2001;Tsang et al. 2014;Wood and Thiele 2009).
By tightly regulating the expression of the copper handling genes mentioned above, S. cerevisiae is able to balance copper levels despite the extracellular environment. Two transcription factors regulate copper gene expression. Ace1 is activated in response to high-copper conditions and activates the expression of genes that encode copperbinding metallothioneins and Sod1 superoxide dismutase (Labbé et al. 1997;Yamaguchi-Iwai et al. 1997). Conversely, in response to copper deficiency, Mac1 transcription factor, in synergy with other chromatin regulators, activates genes such as CTR1, FRE1, and FRE7 essential for copper reduction and import (Georgatsou and Alexandraki 1999;Gkouskou et al. 2019;Gross et al. 2000;Jungmann et al. 1993;Labbé et al. 1997;Voutsina et al. 2019). The activity of these two transcription factors is coordinated so that their target genes are oppositely regulated in response to copper levels (Keller et al. 2005;Peña et al. 1998). In other organisms, such as Drosophila melanogaster and mammals, copper homeostasis is controlled at the transcriptional level by the zinc-responsive factor MTF-1 (Günther et al. 2012;Zhang et al. 2001).
Transcription factor Mac1 contains an amino-terminal DNA-binding domain and a carboxyl-terminal activation domain and is localized to the nucleus in both copperreplete and copper-starvation conditions (Graden and Winge 1997;Jungmann et al. 1993;Serpe et al. 1999). The amino-terminal (1-40) amino acids include a copperfist domain, which undergoes a conformational change on a copper-binding motif that allows DNA-binding. This domain contains a conserved, among fungal species, array of zinc-binding residues (Cys-X 2 -Cys-X 8 -Cys-X-His) (Turner et al. 1998). The carboxy-terminal activation domain contains two cysteine-rich, REPI and REPII copper-binding motifs (Brown et al. 2002;Jensen and Winge 1998;Voutsina et al. 2001). Each one of these motifs binds four copper ions in a poly-copper cluster. Under copperreplete conditions, an intramolecular interaction between the DNA binding and transactivation domain results in the loss of both DNA binding and transactivation activity of Mac1 (Graden and Winge 1997;Labbé et al. 1997). A point mutation, from histidine to glutamine (H279Q), in the REPI domain, leads to a constitutively active Mac1 up protein that binds fewer copper ions per molecule than the wildtype protein (Graden and Winge 1997;Jensen and Winge 1998;Keller et al. 2000;Serpe et al. 1999). While this copper-induced allosteric switch explains how Mac1 activity is regulated by copper, the mechanism for copper translocation to the nucleus for binding to Mac1 remains unknown.
Previous studies suggested that activation of Mac1regulated genes under copper-limiting conditions requires a functional Sod1 and that Mac1 activity can be affected by the DNA damaging agent MMS or by oxidative stress in a Sod1-dependent manner, suggesting that other environmental factors alter, and/or are critical to copper sensing/ binding (Dong et al. 2013;Tsang et al. 2014;Wood and Thiele 2009). Many questions have been raised, regarding the mechanism by which Sod1 accomplishes Mac1 activation and whether additional proteins affect copper sensing and homeostasis.
In this study, based on a newly identified synthetic growth deficiency phenotype of the double mutant mac1Δ sod1Δ, we extended previous findings on the Sod1-Mac1 liaison, independently of prolonged copper deficiency or oxidative stress conditions. We present a weak, possibly transient physical interaction between Sod1 and Mac1 proteins, important for the specificity of the Sod1 enzymatic activity on Mac1 function. We showed that Sod1 modulates both DNA binding and transactivation functions of Mac1, while this relationship becomes Sod1-independent in the constitutively active mutant (copper-independent) Mac1 up . Therefore, our findings point to the copper-related conformational alteration that the transcription factor may undergo due to the catalytic activity of Sod1. This is a housekeeping property of Mac1 protein regulation, under nutrient rich, physiological growth conditions.
Cells were propagated at 30 ℃ in enriched yeast extract, peptone-based medium with 2% glucose (YPD) and in 0.67% yeast nitrogen base 20 amino acids, uracil, adenine, and 2% glucose-synthetic complete (SC) medium was used when needed.

Plasmids
The pACTII:MAC1 and pGBT9:MAC1 plasmids were previously described ). The pGBT9:SOD1 plasmid was created by cloning a PCR fragment containing the SOD1 gene as an EcoRI-XhoI fragment as an in-frame C-terminal fusion with the Gal4-DBD gene into the pGBT9 vector. The pGBT9:CCS1 plasmid was created by cloning a PCR fragment containing the CCS1 gene as an XmaI-PstI fragment and in-frame C-terminal fusion with the GAL4-DBD gene into the pGBT9 vector. For the yeast one-hybrid experiment, the PCR fragment of MAC1 (+ 1, + 1254) was cloned as an NcoI-NheI fragment and in-frame C-terminal fusion with the LexA gene, into the pAS64F2 vector. For overexpression of Sod1 under the control of TPI1 promoter, a PCR fragment containing the SOD1 gene as an NcoI-XhoI fragment was cloned into pYX142 vector. For the Sod1 and Ccs1 protein co-immunoprecipitation experiment, the PCR fragment of CCS1 (+ 1, + 747) was cloned as an XmaI-PstI fragment and in-frame with the 9MYC epitope, into the pYX142 vector. As a negative control of the coimmunoprecipitation experiment, the PCR fragment of CCS1 (+ 1, + 750) was cloned as an XmaI-EcoRV fragment, into the pYX142 vector. The pRS315:MAC1 up plasmid was a gift from Graden and Winge (Graden and Winge 1997). The DNA polymerase used for fragment amplification was Vent (NEB). All PCR primer sequences are available upon request.
β-Galactosidase, two-hybrid assays 5 ml of yeast cultures was diluted and regrown to an OD600 of 1.0 in SC and β-galactosidase activity was measured (Ausubel 1987). Yeast two-hybrid assays were performed as previously described (Georgakopoulos et al. 2001;Voutsina et al. 2001).

Chromatin immunoprecipitation (ChIP) assays
ChIP assays were performed as previously described in (Kuo and Allis 1999) with some adaptations. Briefly, overnight yeast cultures (FT5 MAC1-9MYC, FT5 sod1Δ MAC1-9MYC, and FT5 MAC1-9MYC transformed with pYX142:SOD1) grown in YPD or SC medium for plasmid retention, subsequently inoculated 50 ml cell cultures, which were grown in nutrient-rich conditions (YPD) until a final OD 600 of 1.0, cross-linked with formaldehyde for 20 min at RT, lysed by vortexing for 1 h at 4 ℃ with glass beads, sheared by sonication in fragments peaking at ∼400 bp. IP was performed using the Red Anti-c-Myc Affinity Gel (Ezview™) antibody. The ratio of IP/INPUT sample used in real-time quantitative polymerase chain reaction (qPCR) was 1/25; the dye used was SYBR green. The enrichment values were normalized to those of INPUT samples (nonimmune) and are presented as fold change over the enrichment value obtained by amplification of the control region of CYT1 as mentioned in the figure legends. The polymerase used for PCR was Taq (MINOTECH Biotechnology). All PCR primer sequences are shown in Table 1 in Supplementary data.

Reverse transcriptase-qPCR analyses
RNA was extracted using the hot acid phenol method and RT was performed as previously described (Ausubel 1987) and transcript enrichment was calculated by qPCR. Normalization of the expression levels was done over the constitutively expressed gene TAF10. All PCR primer sequences are shown in Table 1 in Supplementary data.

Mac1 transcriptional activator interacts weakly with Sod1 protein in vivo
Investigating the Mac1-Sod1 reported functional connection, we observed that Mac1 and Sod1 exhibited a synthetic sick phenotype, not previously reported. We found that under nutrient-rich conditions (YPD), the concomitant deletion of MAC1 and SOD1 genes had a severe growth defect compared to the wild-type strain, as well as the singly deleted strains mac1Δ and sod1Δ (Fig. 1A). The growth assay phenotypes were supported by the growth curves of the strains above in nutrient-rich liquid cultures without induced copper deficiency. The mac1Δ sod1Δ strain exhibited a slow-growth phenotype compared to the wildtype and single deletant strains (Fig. 1B). This finding confirmed a close collaborating function or effect of the Mac1 and Sod1 proteins.
We have further detected a Sod1-Mac1 protein physical interaction in vivo both by yeast two-hybrid assays and co-IP experiments. Regarding the yeast two-hybrid assay, we used a plasmid-borne Sod1-Gal4 DNA-binding domain (DBD) hybrid protein as a bait, and a Mac1-Gal4 A Cultures of the strains were grown in rich medium (YPD) to an OD 600 ~ 1.0. Five serial 1/10th dilutions of the cells were spotted on a YPD plate. The plate was incubated for 2 days at 30 ℃. B Growth curves of wild type, mac1Δ, sod1Δ, and mac1Δ sod1Δ strains grown in YPD medium. Time points of the exponentially growing cells (starting from an ΟD 600 ~ 0.25) were taken and measured every 2 h. The experiments were performed twice activation domain (AD) hybrid protein as a prey (due to the transactivation function of Mac1) ( Fig. 2A). Taking advantage of the homodimerization ability of Mac1, we used the Mac1-Gal4DBD and Mac1-Gal4AD hybrids as positive control of the experiment, and combinations of empty plasmids and hybrids as negative controls. The testing on SC plates lacking tryptophan and leucine (complemented by the TRP1 and LEU2 plasmid gene markers) and lacking histidine or adenine or X-gal supplemented (to challenge the genomic HIS3, ADE2, and LacΖ reporters) reconfirmed the viability of Mac1-Gal4DBD and Mac1-Gal4AD expressing cells (Fig. 2B). The Sod1-Gal4DBD and Mac1-Gal4AD expressing cells exhibited viability, not as strong as the Mac1-Gal4DBD and Mac1-Gal4AD expressing cells. This may indicate a probably weak Sod1-Mac1 interaction. In fact, since we have not been able to identify β-galactosidase activity on the same plating assay, we have further investigated the potential Sod1-Mac1 interaction using the more sensitive yeast two-hybrid assay in liquid cultures, since it allows realtime in vitro quantitative measurements. In consistency, the cells transformed with plasmids expressing Sod1-Gal4DBD and Mac1-Gal4AD exhibited a slightly higher enzymatic activity of β-galactosidase, compared to the negative control cultures transformed with combinations of an empty plasmid and one expressing a hybrid protein (Fig. 2C). We have also examined, but we were unable to detect, a possible interaction between the Sod1 copper chaperone Ccs1 and Mac1 proteins, by the yeast twohybrid assay in liquid cultures (Fig. 2C).

Fig. 2
Mac1 and Sod1 in vivo interaction. A Schematic representation of the yeast two-hybrid system. SOD1 gene was cloned into pGBT9 vector and was co-expressed as a hybrid with the DNA-binding domain of the Gal4 transcription factor. MAC1 gene was cloned into pACTII vector and was co-expressed as a hybrid with the activation domain of the Gal4 transcription factor. Sod1-Mac1 protein interaction activates the transcription of three reporter genes. B Testing of Sod1 and Mac1 protein interaction using the yeast two-hybrid assay. The co-transformed yeast strains were plated on the control medium (SC -Leu, -Trp) and selective media (SC -Leu, -Trp, -Ade, or SC -Leu, -Trp, -His, +3-aminotriazole). The combinations with empty plasmids were used as negative controls and Mac1-Gal4BD/ Mac1-Gal4AD was used as positive control. C Yeast cells were cotransformed with pACTII:MAC1 and pGBT9:SOD1 plasmids and pACTII:MAC1 and pGBT9:CCS1 plasmids. The combinations with empty plasmids were used as negative controls. Liquid cultures were grown to mid-log phase in SC media and β-galactosidase in vitro assays were performed. Samples were analyzed in duplicates and data are representative of at least three independent experiments. NS, not significant; ***p < 0.001, one-way ANOVA test. D Co-immunoprecipitation (co-IP) of Sod1 and Mac1 proteins was performed in whole cell lysates from the FT5 Mac1-9MYC Sod1-3HA strain. The FT5 Sod1-3HA strain was used as a negative control. Sod1-Ccs1 protein interaction was used as a positive control. The FT5 Sod1-3HA strain was transformed with pYX142 expressing either Ccs1-9MYC or Ccs1 as a negative control. Anti-MYC agarose beads were used for the precipitation and anti-HA antibodies for immunoblotting. Wholecell extract (WCE) lanes contain 20% of the extract before immunoprecipitation. The IP/WCE ratios of the samples, obtained by the Fiji-ImageJ processing package are shown. The co-IP experiment was performed twice ▸ For further confirmation of the in vivo Sod1-Mac1 interaction, we performed a co-IP assay of Sod1-3HA and Mac1-9MYC proteins that were both epitope-tagged in the genome. We have used the known Sod1-Ccs1 interaction, as a positive control of the experiment, where Sod1-3HA was tagged in the genome and Ccs1-9MYC was overexpressed by the pYX142 plasmid. As it is shown in Fig. 2D, a weak Mac1-Sod1 interaction was detected also at the co-IP assay in cells grown in the nutrient-rich growth medium YPD. This direct interaction corroborated the observed functional interaction between Mac1 and Sod1.

Superoxide dismutase protein, Sod1 is required for both transactivation and DNA-binding activity of the transcription factor Mac1 in nutrient-rich conditions
It has been previously reported that the Sod1 enzyme is required for full Mac1 function, although, according to the existing literature, the conditions are contradictory. First, it had been revealed that in induced copper deficiency, using the copper chelator BCS, transcription of the Mac1-driven gene CTR1 was Sod1 dependent (Wood and Thiele 2009). Later, another study supported that in MMS-induced DNA damage conditions, Sod1 was required for an unknown aspect of the mechanism controlling the CTR1 expression (Dong et al. 2013). Finally, it was reported that in DNA damage-inducing conditions, following H 2 O 2 , paraquat, and menadione but not MMS treatment, Sod1 was localized in the nucleus and it was required for the induction of oxidative response and some Cu/Fe homeostasis-related genes (Tsang et al. 2014).
We aimed to investigate the Mac1-Sod1 association in physiological, nutrient-rich conditions without copper deficiency or oxidative stress induction, since that was the condition of the identified synthetic sick phenotype. First, performing western blot analysis, we verified that Mac1 protein was not affected quantitatively in a sod1Δ strain (Supplementary data, Fig. 1). Next, we explored the requirement for Sod1 protein in Mac1 activation, analyzing the Mac1-driven mRNA levels of CTR1 and FRE1 genes in the absence of functional Sod1. Particularly, we used a sod1Δ strain, a ccs1Δ strain, and the double-deletant sod1Δ ccs1Δ strain. Sod1 maturation and enzymatic activity require the chaperone Ccs1 (Boyd et al. 2019;Fetherolf et al. 2017). We found that CTR1 and FRE1 mRNA levels in all three deletion strains displayed similar defects in the function of the Mac1 transcription factor compared to the wild-type strain ( Fig. 3A and B). These results indicated that the enzymatically active Sod1 is required for full Mac1 activity in nutrient-rich, unchallenged conditions. Since Mac1 appeared to physically interact with Sod1, we examined whether the Sod1 enzyme affects the DNA-binding ability of the Mac1 transcription factor under nutrientrich conditions. We investigated the localization of Mac1 on the promoters of target genes, in the absence of Sod1 by performing chromatin immunoprecipitation of the tagged Mac1-9ΜYC protein and assayed the genome in sod1Δ and wildtype cells. Our results strongly demonstrated that there is no significant enrichment of the Mac1 transcription factor on the CTR1 and FRE1 promoters in the absence of Sod1. Thus, Sod1 protein is required for the proper binding of Mac1 on these promoters in nutrient-rich conditions (Fig. 3C).
Furthermore, to address the question of whether Sod1 deficiency affects Mac1 transactivation ability, we performed yeast one-hybrid assay and β-galactosidase activity measurements. The plasmid-borne hybrid protein LexA-Mac1 was used along with the pJK103 plasmid which contains the β-galactosidase gene (LacZ) under the control of the LexA promoter. The functionality of this assay relies on the constitutive DNA binding of the LexA-Mac1 hybrid on the LexA promoter and the Mac1-dependent transactivation of the reporter gene. We transformed the mac1Δ and mac1Δ sod1Δ strains with the corresponding plasmids to investigate the transactivation ability of Mac1 in the absence of Sod1. We used mac1Δ instead of the wild-type strain because we have previous evidence that LexA-Mac1 hybrid is not fully active in a wild-type strain, probably due to homodimerization (and squelching) with the native Mac1 protein (Supplementary data, Fig. 2). According to our results, in the mac1Δ sod1Δ strain, Mac1 displayed defective transactivation ability, resulting in reduced transcription of the β-galactosidase gene, compared to the mac1Δ strain (Fig. 3D). Thus, we found that the Sod1 enzyme is important for both the DNA binding and the transactivation functions of Mac1.
Finally, we overexpressed Sod1 using the pYX142 plasmid, to investigate whether it further stimulated Mac1 function. First, we explored the mRNA levels of CTR1 and FRE1 transcripts in wild type and Sod1 overexpressing cultures grown in the nutrient-rich medium YPD. The mRNA levels of both genes indicated that Sod1 overexpression enhanced the transcriptional activity of Mac1 (Fig. 4A). Furthermore, Mac1-9ΜYC binding demonstrated significant enrichment on the CTR1 and FRE1 promoters in the Sod1 overexpressing strain, as was indicated by the chromatin immunoprecipitation experiment (Fig. 4B). Despite the requirement for Sod1 in Mac1 function, we were unable to detect any localization of the Sod1 protein on CTR1 and FRE1 promoters with chromatin immunoprecipitation experiments in SOD1-3HA cells grown in YPD medium (Fig. 3C) implying either a weak Sod1 binding or transient Sod1 localization via protein-protein interaction. Fig. 3 Sod1 is required for both DNA-binding activity and transactivation function of Mac1. A mRNA levels of the Μac1-driven gene CTR1, of wild-type, sod1Δ, ccs1Δ, and sod1Δ ccs1Δ strains, grown in rich medium (ΥPD). **p < 0.01, one-way ANOVA test. B mRNA levels of the Mac1-driven gene FRE1, of wild-type, sod1Δ, ccs1Δ, and sod1Δ ccs1Δ strains, grown in YPD. *p < 0.05, **p < 0.01, oneway ANOVA test. All samples in A and B were analyzed in duplicates and in three independent experiments. C Chromatin immunoprecipitation experiment detecting the localization of the tagged Mac1-9MYC in wildtype and sod1Δ strains and Sod1-3HA in wildtype strain. ChIP assay was followed by real-time PCR analysis of the immunoprecipitated (IP) and input DNA using primers specific for CTR1 and FRE1 promoter region and CYT1 coding region. Immunoprecipitation efficiency is represented by the ratios: CTR1 or FRE1 IP DNA/input DNA/CYT1 IP DNA/CYT1 input DNA. ***p < 0.001, two-way ANOVA test. Samples were analyzed in triplicates and in two independent experiments. D mac1Δ and mac1Δ sod1Δ cells transformed with the LacZ reporter plasmid pJK103, and the LexA-Mac1 plasmid pAS64F2. mac1Δ and mac1Δ sod1Δ strains transformed with the pYX103, and the LexA plasmid pAS64F2 were used as negative controls. The liquid cultures were grown to mid-log phase in synthetic complete medium (SC) and β-galactosidase assays were performed. **p < 0.01, unpaired t-test Samples were analyzed in triplicates in three independent experiments

Sod1 deficiency does not affect the constitutively active Mac1 up mutant protein
The carboxyl-terminal Mac1 activation domain (AD) contains two cysteine-and histidine-rich domains, REPI and REPII, that each bind four Cu + ions in a tetranuclear copper cluster. A point mutation in the REPI domain, converting histidine to glutamine (H279Q), results in the Mac1 up protein, a constitutively active form, independently of copper concentration (Fig. 5A) (Serpe et al. 1999). Having proved that Sod1 is required for both DNA binding and transactivation functions of Mac1, we investigated whether the ablation of Sod1 affects the function of the reduced and constitutively active mutant protein Mac1 up under nutrient-rich conditions. To test that, we transformed the mac1Δ and mac1Δ sod1Δ strains with the plasmid expressing native Mac1 protein or the constitutively active mutant Mac1 up protein. While mRNA levels of Mac1-dependent genes CTR1 and FRE1 driven from the plasmid-borne Mac1 were significantly lower in the absence of SOD1, mRNA levels driven by the plasmid-borne Mac1 up mutant protein were unaffected in that strain ( Fig. 5B and C). This result confirmed that while Sod1 protein is required for physiological Mac1 activation, this requirement can be bypassed by a constitutively active variant of the Mac1 protein suggesting that, regardless of growth conditions, a weak or a transient interaction between Mac1 and Cu/Zn Sod1 contributes to the transcriptionally active form of the transcription factor.

Discussion
In S. cerevisiae, copper uptake occurs by two high-affinity Cu transporters, Ctr1 and Ctr3, or the low-affinity transporter Fet4, following its reduction at the plasma membrane to Cu + by cell-surface metallo-reductases (Dancis et al. 1994;Georgatsou and Alexandraki 1999;Hassett et al. 2000;Knight et al. 1996). Mac1 is a nuclear Cu-sensing transcriptional regulator that is responsible for activating several genes including CTR1, CTR3, and FRE1 genes under copper insufficient conditions (Georgatsou and Alexandraki 1999;Gross et al. 2000;Jungmann et al. 1993;Labbé et al. 1997). Copper-dependent changes in Mac1 activity have been demonstrated that contribute not only to its proper binding on DNA, through the ion-binding at the DNA-binding domain, but also to the inactivation of the transcription factor (Heredia et al. 2001;Jensen and Winge 1998). Copper binding at the transactivation domain of Mac1 results in a poly-copper cluster formation that initiates a conformational switch which favors an intramolecular interaction between the N-terminal DNA-binding domain and the C-terminal transactivation domain (Brown et al. 2002). This conformation in turn prevents Mac1 from binding to DNA and activating gene expression when copper ions are in excess (Graden and Winge 1997;Jensen and Winge 1998).
Even though this copper-induced allosteric switch explains how Mac1 activity can be regulated by copper, recent studies have revealed that activation of gene Fig. 4 Overexpression of Sod1 induces Mac1 activity. A mRNA levels of the Mac1-driven genes, CTR1 and FRE1, in cells grown in rich medium (YPD), with or without overexpression of Sod1. BY4742 strain was transformed with either a pYX142 or a pYX142:SOD1 plasmid. Liquid cultures were grown to mid-log phase in YPD medium and RNA extraction was performed. *p < 0.05, **p < 0.01, two-way ANOVA test. B Chromatin immunoprecipitation experiment detecting the localization of the tagged Mac1-9MYC on the CTR1 and FRE1 promoters in wild-type and Sod1 overexpressing strains grown as in A. ChIP assay was followed by real-time PCR analysis of the immunoprecipitated (IP) and input DNA using primers specific for CTR1 and FRE1 promoter region and CYT1 coding region. Immunoprecipitation efficiency is represented by the ratios: CTR1 or FRE1 IP DNA/input DNA/CYT1 IP DNA/CYT1 input DNA. *p < 0.05, **p < 0.01, two-way ANOVA test. All samples were analyzed in triplicates in two independent experiments 1 3 expression under copper-limiting conditions requires the functional Cu/Zn superoxide dismutase Sod1 and that Mac1 activity can be affected by the DNA damaging agent MMS or other oxidative agents in a Sod1-dependent manner (Dong et al. 2013;Tsang et al. 2014;Wood and Thiele 2009). These studies suggested that other environmental factors modify or may be crucial to copper sensing. A plethora of questions have been raised regarding the mechanism by which Sod1 acts upon Mac1 activation, the role of Sod1 in copper sensing also in other organisms, and the possibility that additional proteins and metabolites affect copper sensing and homeostasis. Our group has previously reported that in addition to copper, several chromatin proteins modulate Mac1 functionality Voutsina et al. 2019).
In this study, we reexamined the Sod1 involvement in the Mac1-regulated transcription, but in unchallenged nutrientrich growth conditions and revealed the following new findings. First, we found that Mac1 and Sod1 proteins exhibited a collaborating function or effect as was demonstrated by the synthetic sick phenotype of the simultaneous ablation of both genes/proteins (Fig. 1). Second, we succeeded in detecting an interaction between Sod1 and Mac1 proteins. This interaction is not robust and irreversible but weak and potentially transient (Fig. 2). According to our presented yeast two-hybrid assay, the Mac1-Sod1 association exhibited survival under histidine and adenine deficiency, multiple times but not always. However, β-galactosidase activity measurement consistently resulted in a slightly higher enzymatic activity of β-galactosidase, over the negative controls. The weak potential in vivo Mac1-Sod1 interaction was confirmed by the co-IP experiment. In that assay, the interaction was detectable only following formaldehyde crosslinking of the extract before assaying the interaction. These findings established the intimate relationship between the two proteins. This is quite important because, although we did not detect any Sod1 protein localization on Mac1 target promoters, a close Mac1-Sod1 protein interaction is a prerequisite for the specificity of the Sod1 enzyme function on a has a point mutation in the REPI region, converting histidine to glutamine (H279Q), resulting in a constitutively active form independently of copper concentration. B CTR1 mRNA levels driven from Mac1 or Mac1 up in mac1Δ and mac1Δ sod1Δ strains, grown in rich medium (YPD). mac1Δ and mac1Δ sod1Δ strains were transformed with pRS315:MAC1 and pRS315:MAC1 up plasmids. NS, not significant; **p < 0.01, two-way ANOVA test. C FRE1 mRNA levels driven from Mac1 or Mac1 up in mac1Δ and mac1Δ sod1Δ strains, grown as in B. NS, not significant; **p < 0.01, two-way ANOVA test. All samples were analyzed in two independent experiments ▸ selected target protein (Finkel 2011). Third, in consistency with a previously reported work, we confirmed that the single deletions of SOD1 and CCS1 and the double gene deletion resulted in the reduced transcriptional ability of Mac1 (Wood and Thiele 2009). In addition, we established that the Sod1 requirement in Mac1-driven transcription occurred under physiological, nutrient-rich growth conditions and not only in the induced copper-deficiency environment, pointing out that the interplay is independent of copper-deficiency conditions ( Fig. 3A and B). Fourth, we demonstrated that Sod1 is required not only for the binding of Mac1 to CuRE elements on CTR1 and FRE1 promoters (as previously shown) but also for its transactivation function ( Fig. 3C and D). The latter was revealed in contradiction to previous conclusions (Wood and Thiele 2009) because we performed the one-hybrid assay in a strain lacking the endogenous MAC1 gene avoiding molecular competition. This is quite important, since it indicates that all cysteine-rich areas of Mac1 (DNA binding and transactivation) may be affected by the Sod1 enzymatic activity. In accordance with this result, the transcriptional activity of the Mac1 up allele was unaffected by Sod1 ablation (Fig. 5). The Mac1 up mutant that binds fewer copper ions, does not respond to copper ion concentration and is constitutively transcriptionally active (Serpe et al. 1999). We found that under nutrient-rich conditions, Mac1 up conformation does not require Sod1 activity to function, implying that the catalytically active Sod1 can only modulate Mac1 activity when Mac1 is functionally responsive to copper levels/binding. Finally, the necessity of Sod1 in the Mac1 function was further verified through the overexpression of the enzyme which resulted in increased mRNA levels of the Mac1-driven genes CTR1 and FRE1 and in enriched Mac1 localization on the corresponding gene promoters (Fig. 4). This also implied a stoichiometric relation between Mac1 protein and Sod1 protein/activity in the final transcriptional outcome. In fact, addition of copper ions in the culture of Sod1 overexpressing strain, reduced the Sod1 effect on Mac1 transactivation function (Supplementary data, Fig. 3). This indicated that copper ions compete with Sod1 action on Mac1 conformational alterations.
Taking together our findings, we suggest that Sod1 interaction with Mac1 is vital for the transcription factor Fig. 6 Proposed model for Mac1 and Sod1 association in unchallenged conditions. Via Sod1 interaction, the transcription factor Mac1 undergoes changes in the redox state of cysteine-rich motifs at the transactivation and DNA-binding domains, important for the main-tenance of its transcriptionally active form (left panel). Upon Sod1 depletion, Mac1 is prone to the copper-chelated form, promoting an intramolecular interaction between the DBD and AD, leading to its inactivation (right panel). Created with BioRender.com functionality, probably due to a copper-related conformational alteration that the transcription factor undergoes, not only under induced copper-deprivation or oxidative stress conditions but also in a nutrient-rich environment as a housekeeping property of Mac1 regulation. Sod1 is the major antioxidant protein in all living organisms. The role of the Cu/Zn SOD1 is to catalyze the conversion of the potentially toxic superoxide anion, O 2 ·− , to the less toxic substance hydrogen peroxide, H 2 O 2 . Mac1 likely undergoes changes in its redox state that correlate with its activity in a Sod1-dependent manner. It is possible that Sod1-generated H 2 O 2 serves as an important signal for Mac1 activation, as it does in a variety of other cysteine-rich proteins (Finkel 2011;Hancock et al. 2006;Scandalios 2005). It has been previously reported, that H 2 O 2 acts as a signaling agent to regulate transcription factor activity, among other functions (Delaunay et al. 2000;Veal et al. 2007). For instance, the tumor suppressor p53-regulated genes, exhibit distinct patterns in response to different levels of H 2 O 2, and many studies provide evidence of the involvement of SOD in p53 regulation (Rieber and Strasberg-Rieber 2012;Sablina et al. 2005;Watanabe et al. 2013). Similarly to Mac1, due to its copper-fist motif at the DNA-binding domain, the transcription factor p53 is also a Cu-and Zn-binding protein (Hainaut et al. 1995;Hainaut and Mann 2001). It binds copper to form Cu(I)-cysteinyl thiolates, rendering it incapable of DNA binding (Furuta et al. 2002;Hainaut et al. 1995).
For S. cerevisiae, we propose that the Sod1 catalytic reaction serves conformational changes of the Mac1 transcription factor, vital for its transcriptional activity, independently of copper deficiency or other medium reactive agents (Fig. 6). Mac1 exists in different redox forms; the transcriptionally active form and the copper-chelated at REP regions inactive form. A potential model is that H 2 O 2 , which is the product of Sod1 enzymatic catalysis, could modify the redox state of the reactive cysteine residues within the REP regions of the transactivation domain. Sod1 catalytic inability, by either Sod1 or the copper chaperone Ccs1 ablation, results in a shift to the copper-binding redox state of Mac1. Sod1 may also contribute to the conformational change on the copperfist motif that allows the DNA-binding of the transcription factor. The function of Mac1 up independently of the Sod1 presence together with the reduction of Sod1 effect by the addition of excess copper ions, are in accordance with our proposed model; Sod1 function follows copper ion effect on Mac1 conformation/activity.
Copper-responsive transcription factors play a critical role in copper homeostasis in a variety of species including fungi, green algae, plants, flies, and humans (Günther et al. 2012;Rutherford and Bird 2004;Sommer et al. 2010;Yamasaki et al. 2009;Zhang et al. 2001). Thus, full understanding of the conserved copper regulatory and sensing pathways in yeast under physiological conditions could reveal novel aspects of the copper regulon with potential applications in human health and copper-related diseases and pathologies.