Phosphorylation and Proteasome Recognition of the mRNA-Binding Protein Cth2 Facilitates Yeast Adaptation to Iron Deficiency

Iron is a vital element for many metabolic pathways, including the synthesis of DNA and proteins, and the generation of energy via oxidative phosphorylation. Therefore, living organisms have developed tightly controlled mechanisms to properly distribute iron, since imbalances lead to nutritional deficiencies, multiple diseases, and vulnerability against pathogens. Saccharomyces cerevisiae Cth2 is a conserved mRNA-binding protein that coordinates a global reprogramming of iron metabolism in response to iron deficiency in order to optimize its utilization. Here we report that the phosphorylation of Cth2 at specific serine residues is essential to regulate the stability of the protein and adaptation to iron depletion. We identify the kinase and ubiquitination machinery implicated in this process to establish a posttranscriptional regulatory model. These results and recent findings for both mammals and plants reinforce the privileged position of E3 ubiquitin ligases and phosphorylation events in the regulation of eukaryotic iron homeostasis.

IMPORTANCE Iron is a vital element for many metabolic pathways, including the synthesis of DNA and proteins, and the generation of energy via oxidative phosphorylation. Therefore, living organisms have developed tightly controlled mechanisms to properly distribute iron, since imbalances lead to nutritional deficiencies, multiple diseases, and vulnerability against pathogens. Saccharomyces cerevisiae Cth2 is a conserved mRNA-binding protein that coordinates a global reprogramming of iron metabolism in response to iron deficiency in order to optimize its utilization.
Here we report that the phosphorylation of Cth2 at specific serine residues is essential to regulate the stability of the protein and adaptation to iron depletion. We identify the kinase and ubiquitination machinery implicated in this process to establish a posttranscriptional regulatory model. These results and recent findings for both mammals and plants reinforce the privileged position of E3 ubiquitin ligases and phosphorylation events in the regulation of eukaryotic iron homeostasis.
KEYWORDS iron deficiency, phosphorylation, posttranslational regulation, protein stability, yeast I ron is an essential micronutrient for all eukaryotic organisms because it participates as a catalytic redox cofactor in numerous metabolic pathways, including energy generation through mitochondrial respiration, protein translation, DNA replication and repair, and lipid biosynthesis. Despite being very abundant, iron bioavailability is highly restricted due to its low solubility at physiological pH. Indeed, iron deficiency anemia represents the most common nutritional disorder in humans, thought to affect more than two billion people, with large impact on pregnant women and children (1). Moreover, iron imbalances lead to both frequent and rare human diseases, including hereditary hemochromatosis, Friedreich's ataxia, and aceruloplasminemia, and influence fungal and bacterial pathogenesis (2). Therefore, eukaryotic cells have developed sophisticated molecular mechanisms to properly respond to alterations in iron levels.
In response to iron deficiency, the budding yeast Saccharomyces cerevisiae activates the expression of Cth2, an mRNA-binding protein that, in coordination with its partially redundant homolog Cth1, plays a pivotal role in the adaptation to iron limitation by posttranscriptionally remodeling iron metabolism (3,4). Cth1 and Cth2 belong to a family of proteins conserved in all eukaryotes that is characterized by the presence of two tandem zinc fingers (TZFs) of the CX 8 CX 5 CX 3 H type, which directly interact with AU-rich elements (AREs) located in the 3=-untranslated region (3=-UTR) of many mRNAs to promote their rapid destabilization and inhibit their translation (3)(4)(5)(6). Humans express three TZF-containing proteins, with tristetraprolin being the most studied family member (reviewed in references 7 and 8). Upon iron scarcity, Cth1 and Cth2 coordinately promote the decay of mRNAs that encode proteins with functions in highly iron-consuming pathways, including the mitochondrial electron transport chain and the tricarboxylic acid cycle, to prioritize its utilization in essential iron-dependent processes such as deoxyribonucleotide synthesis by ribonucleotide reductase (3,4,9). Cth2 binds to its target transcripts in the nucleus. Then, the mRNA-Cth2 complexes are exported to the cytoplasm where mRNA turnover and translation inhibition take place (6,10,11). Mutagenesis of a conserved cysteine residue within Cth2 TZFs abrogates both its interaction with ARE-containing mRNAs and its nuclear export (3,11). Cells lacking a functional Cth2 protein (cth2Δ cells or Cth2-TZF mutants) grow properly under iron-replete conditions but display growth impairment under iron-deficient conditions due to defects in the downregulation of multiple iron-related mRNAs (3,6). Although the molecular reasons are not known, the overexpression of a functional Cth2 protein under iron-sufficient conditions is also detrimental for growth (5,10). Therefore, yeast cells have developed refined strategies to tightly control Cth2 expression levels to satisfy cellular iron demands. First, the partially redundant Aft1 and Aft2 iron-responsive factors specifically activate CTH2 transcription in response to iron deprivation, whereas Cth2 protein is not present under iron-sufficient conditions (3). Second, the CTH2 mRNA contains an ARE within its 3=-UTR that limits its own expression through a negativefeedback autoregulatory mechanism (6,12).
In this study, we identified key amino acid residues of the TZF-containing protein Cth2 that are phosphorylated by the Hrr25 kinase. These phosphorylation events facilitate Cth2 recognition by the SCF Grr1 ubiquitin ligase complex, which promotes its proteasomal degradation. This posttranslational regulatory process is essential for the adaptation of yeast cells to iron deficiency.

RESULTS
The yeast Cth2 protein is phosphorylated at specific serine residues. In response to iron deficiency, the budding yeast S. cerevisiae expresses two iron-regulated TZF proteins named Cth1 and Cth2 (3)(4)(5). When we analyzed the electrophoretic mobility of functional, Flag epitope-tagged Cth1 and Cth2 proteins on SDSpolyacrylamide gels, we observed a slow-migrating band in addition to the bands expected according to their size (3, 12) ( Fig. 1A; see also Fig. S1 in the supplemental material). To ascertain whether the altered electrophoretic mobility was due to phosphorylation, we treated the protein extracts from yeast cells with phosphatase. We observed that the upper band disappeared when the extracts were incubated with phosphatase, but it was still present when phosphatase inhibitors were included in the assay (Fig. 1A and Fig. S1). These results strongly suggest that both Cth1 and Cth2 are phosphoproteins.
Mechanism of Cth2 Protein Turnover ® we specifically immunoprecipitated Cth2 protein. Mass spectrometry revealed that Cth2 serine residues S65, S68, and S70 were phosphorylated, with a predominance of peptides with phosphorylated S68 (schematic representation in Fig. 1B). Interestingly, Cth2 S65 and S68 residues are conserved in Cth1 protein (Fig. S1). To further ascertain the relevance of these serine residues in Cth2 phosphorylation, we mutagenized them and the nearby S64 residue to alanine. An immunoblot analysis showed that single S64A and S65A mutations did not change the electrophoretic pattern of Cth2, the S70A substitution only partially decreased Cth2 phosphorylation, and mutagenesis of S68 almost fully collapsed Cth2 phosphorylation shift (Fig. 1C). Importantly, the Cth2-S64A/ S65A, Cth2-S68A/S70A, and Cth2-S64A/S65A/S68A/S70A mutations abolished most Cth2 phosphorylation and increased Cth2 protein abundance to 1.5-to 4.4-fold ( Fig. 1C and Fig. S2A). These results suggest that all four serine residues analyzed are phosphorylated in vivo.
Cth2 phosphorylation is required for growth under iron starvation. To determine the physiological relevance of Cth2 phosphorylation in the response of yeast cells to iron deprivation, we expressed Cth2 serine mutants in a cth1Δ cth2Δ double mutant strain (to eliminate any CTH1 redundancy) and assayed growth in iron-deficient conditions, achieved in this case by the addition of the Fe 2ϩ -specific chelator Ferrozine. We observed that the single Cth2 serine mutants did not display any significant growth defect under iron-limiting conditions compared to cells expressing wild-type Cth2 (Fig. 2). Remarkably, both Cth2-S64A/S65A and Cth2-S68A/S70A double mutants exhibited a growth defect under low-iron conditions (Fig. 2). Strikingly, the Cth2-S64A/ S65A/S68A/S70A quadruple mutant partially rescued the growth defect under irondeficient conditions observed for the Cth2 double mutants (Fig. S2B). These results indicate that both Cth2-S64/S65 and Cth2-S68/S70 serine pairs are important for optimal adaptation to iron deficiency.
The Cth2 serine mutants function in targeted mRNA degradation. To determine whether Cth2 serine mutants had defects in targeted mRNA degradation under conditions of iron scarcity, we determined the mRNA levels of various Cth2 target transcripts in cells grown in both iron-sufficient and iron-deficient media. In particular, we determined the mRNA levels for SDH4, which encodes a subunit of succinate dehydrogenase in the tricarboxylic acid cycle; HEM15, which encodes ferrochelatase, the last step in the heme biosynthetic pathway; RNR2, a small subunit of ribonucleotide reductase in deoxyribonucleotide synthesis; and RNase L inhibitor RLI1, which is required for ribosome recycling during protein translation. Surprisingly, cells expressing CTH2 serine mutants S64A/S65A, S68A/S70A or S64A/S65A/S68A/S70A promoted the downregulation of the four transcripts upon iron scarcity in a manner similar to that of wild-type CTH2 ( Fig. 3 and Fig. S2C). These results suggest that the growth defect displayed by the Cth2 serine mutants under iron-deficient conditions is not due to defects in targeted mRNA decay.
The patch of S64/S65/S68/S70 residues facilitates Cth2 protein degradation. Previous studies have shown that overexpression of a functional TZF-containing Cth2 protein is cytotoxic for yeast cells (5,10,12). Indeed, both the lack of a functional Cth2 protein (cth1Δ cth2Δ cells containing either empty vector or the TZF mutant allele CTH2-C190R) or the overexpression of wild-type CTH2 (cth1Δ cth2Δ cells with a multicopy CTH2 OE plasmid [where OE stands for overexpression]) led to similar growth defects under iron starvation conditions (Fig. 4A). Therefore, we considered whether the growth defect observed for the Cth2 serine mutants under iron-deficient conditions was due to their high protein levels (Fig. 1C). If this were the case, CTH2-S64A/S65A and CTH2-S68A/S70A would be gain-of-function alleles, and unlike the loss-of-function CTH2-C190R allele, these alleles should be toxic when expressed in wild-type cells. Consistent with this hypothesis, the expression of the Cth2 double serine mutants impaired the growth of wild-type cells in iron-deficient media, whereas CTH2-C190R expression did not alter growth (Fig. 4B). Furthermore, the growth defect triggered by the expression of Cth2 double serine mutants was abolished when the Cth2 TZFs were mutagenized, which indicates that Cth2-S64A/S65A and Cth2-S68A/S70A deleterious effects require binding to target mRNAs (Fig. 4B). As previously observed for CTH2 OE cells (10), these results are consistent with a gain of function for Cth2-S64A/S65A and Cth2-S68A/S70A proteins, which are unable to adapt to iron starvation.
We then explored the molecular basis for the increase of abundance of the Cth2 serine mutant proteins. We first determined CTH2-S64A/S65A and CTH2-S68A/S70A transcript levels compared to wild-type CTH2 mRNA under conditions of iron sufficiency and iron deficiency. We observed that the levels of CTH2-S64A/S65A and CTH2-S68A/ S70A mRNAs increased upon iron depletion, but their abundance was not higher than that of the wild-type CTH2 (Fig. 5A). These results indicated that the increased levels of Cth2 serine mutant proteins were not due to higher transcript abundance. Then, we ascertained whether the stability of these Cth2 serine mutant proteins was different from that of the wild-type protein. For this purpose, we grew yeast cells expressing the different Flag-tagged Cth2 proteins under iron-deficient conditions, stopped protein synthesis by the addition of cycloheximide (CHX), and determined the turnover rate of the Cth2 proteins. Our assays indicated that wild-type Cth2 protein had a half-life of approximately 14 min (Fig. 5B). A similar result was obtained when a functional amino-terminal Myc-tagged Cth2 protein was used in this assay (Fig. S3). Some of the single Cth2 serine mutants displayed a slight increase in protein stability (S65A, S68A, and S70A [ Fig. 5B]). However, mutagenesis of S64/S65 or S68/S70 residues increased the half-life of the Cth2 protein approximately sixfold (from 14 to 90 min [ Fig. 5B]). Interestingly, simultaneous mutagenesis of the four S64/S65/S68/S70 residues increased Cth2 protein stability only fourfold (from 14 to 60 min [ Fig. S2D]). These results demonstrate that the S64/S65/S68/S70 residues determine the stability of Cth2 protein under low-iron conditions. Furthermore, they also indicate that both Cth2 S64/S65 and S68/S70 serine pairs are required to promote Cth2 protein turnover under iron-deficient conditions.
The SCF Grr1 ubiquitin ligase complex facilitates Cth2 protein degradation by the proteasome. To determine the mechanism that yeast cells utilize to regulate Cth2 protein stability, we expressed Flag 2 -Cth2 in cells lacking the drug efflux pump gene PDR5, which are highly sensitive to the proteasome inhibitor MG132 (13). Yeast cells were cultivated in the presence or absence of MG132 under both iron-replete and iron-depleted conditions. As previously reported, Cth2 protein levels were dramatically upregulated in response to iron scarcity (3) (Fig. 6A). The addition of the MG132  Fig. 1C were grown in SC-ura containing 100 M BPS (ϪFe). Then, 50 g/ml cycloheximide (CHX) was added to stop translation, and aliquots were isolated at the indicated times. Total proteins were extracted, and Cth2 protein levels were determined by immunoblotting with anti-Flag antibody. Equal amounts of total proteins were loaded in each lane. Cth2 protein levels were determined, and the average half-life was calculated (n ϭ 3 for Cth2-S64A/S65A and Cth2-S68A/S70A and n ϭ 2 for the rest of experiments). A representative immunoblot is shown.
Mechanism of Cth2 Protein Turnover ® proteasome inhibitor to the iron-deficient cells elevated Cth2 protein abundance, whereas little increase was observed under iron-replete conditions (Fig. 6A). To further address the relevance of the proteasome in Cth2 protein turnover, we used a temperature-sensitive pre1-1 strain, which is defective in an essential subunit of the proteasome when grown at 37°C. We determined Cth2 protein stability at the restrictive temperature in the pre1-1 strain and an isogenic wild-type strain. Cth2 protein displayed a half-life of 10 min at 37°C when expressed in the control strain, whereas it was highly stabilized in the pre1-1 strain (Fig. 6B). Taken together, these data demonstrate that Cth2 protein is degraded via the proteasome.
A previous genome-wide screen devoted to isolate ubiquitinated substrates in complex with their E3 ubiquitin ligase identified Cth2 protein as a potential substrate for Grr1, the substrate adaptor component of the SCF Grr1 E3 ubiquitin ligase complex (14). To ascertain whether Grr1 regulates Cth2 protein abundance, we determined the steady-state Cth2 protein levels in grr1Δ cells grown under iron-sufficient and irondeficient conditions. Again, we observed that Cth2 protein was barely detected under iron-sufficient conditions but was highly expressed upon iron depletion (Fig. 7A). Furthermore, grr1Δ cells displayed an increase in Cth2 protein abundance compared to GRR1-expressing cells grown in both iron sufficiency and starvation (Fig. 7A). Importantly, Cth2 protein upregulation in iron-deficient grr1Δ cells was not due to an increase in CTH2 mRNA levels. Instead, under iron starvation conditions, CTH2 transcript abundance diminished in grr1Δ cells compared to GRR1-expressing cells (Fig. 7B). Overall, these results were consistent with Grr1 regulating Cth2 protein abundance at a posttranslational level. Therefore, we determined Flag 2 -Cth2 protein stability in wildtype cells and cells lacking Grr1 protein. We observed that the half-life of the Cth2 protein increased from 10 min to more than 2 h when the GRR1 gene was deleted (Fig. 7C), demonstrating that the Grr1 F-box protein is a major factor that determines Cth2 protein stability under iron-deficient conditions. The Grr1 F-box protein is the substrate recognition module of the SCF Grr1 ubiquitin ligase complex, which is composed of the cullin scaffold Cdc53 that bridges the adaptor protein Skp1 and the RING finger protein Rbx1 (15, 16). Thus, we hypothesized that cells Wild-type (YWO0607) and pre1-1 (YWO0608) yeast strains transformed with pRS416-Flag 2 -Cth2 plasmid were grown in SC medium for 3 h at 37°C. Then, 100 M BPS was added, and the cells were incubated for 6 additional hours at 37°C (ϪFe). At this point, cycloheximide (CHX) was added to stop translation, and aliquots were isolated and analyzed at the indicated times (in minutes). A representative experiment of two independent biological assays is shown. The asterisk indicates a nonspecific band. defective in any of the SCF Grr1 components would increase Cth2 protein stability. To test this, we used the skp1-11 and rbx1-1 yeast strains, which express temperaturesensitive alleles of the essential genes SKP1 and RBX1. We observed that the stability of Cth2 protein dramatically increased in skp1-11 and rbx1-1 mutants at the restrictive temperature (Fig. 7D). Taken together, these results demonstrate that the SCF Grr1 E3 ubiquitin ligase promotes the degradation of Cth2 protein during iron deficiency.
The Grr1 F-box protein is necessary for optimal growth under iron-deficient and respiratory conditions. We have shown here that, similarly to Cth2 serine mutants, grr1Δ cells accumulate Cth2 protein ( Fig. 1 and Fig. 7). Given that Cth2 serine Mechanism of Cth2 Protein Turnover ® mutants display a growth defect under low-iron conditions due to elevated levels of functional Cth2 protein ( Fig. 2 and Fig. 5), we decided to assay the growth of grr1Δ cells under iron-deficient conditions (Fig. 8A). Remarkably, we observed a dramatic growth defect for grr1Δ cells when iron bioavailability was limited ( Fig. 8B and C). In agreement with the elevated levels of functional Cth2 being the reason for this growth defect, deletion of CTH2 gene rescued grr1Δ growth in low-iron media (Fig. 8C). This rescue was only partial because yeast cells solely lacking CTH2 display a growth defect under iron-depleted conditions (3) (Fig. 8C). Among Cth2 target mRNAs, genes within the mitochondrial electron transport chain are considered preferential substrates (3). Therefore, we considered the possibility that, in response to iron deficiency, Cth2 would promote the repression of mitochondrial respiration, which consumes elevated amounts of iron but is dispensable for yeast cell growth. Thus, we tested the growth of grr1Δ cells under respiratory conditions by using a medium containing the nonfermentable ethanol and glycerol carbon sources instead of glucose. Yeast strains without GRR1 exhibited a growth defect under respiring (ethanol and glycerol [ethanol/ glycerol]) conditions compared to fermentative (glucose) conditions ( Fig. 8D and E). Again, CTH2 deletion partially rescued growth in ethanol/glycerol probably due to a release in the repression of respiration. All together, these results demonstrate that Grr1 protein plays a relevant function in the adaptation of yeast cells to iron deficiency and respiratory conditions by limiting Cth2 protein levels.
The Grr1 ubiquitin ligase preferentially interacts with phosphorylated Cth2. Given that the Grr1 F-box protein targets only phosphorylated substrates through its leucine-rich motif (17), we ascertained whether Cth2 phosphorylation was necessary for Grr1 recruitment. For this purpose, we expressed a truncated version of Grr1 protein lacking its F-box domain (Grr1ΔFbox), which recognizes its substrates but is unable to promote their degradation due to lack of interaction with the Skp1 linker protein (14). Flag-tagged Grr1ΔFbox protein was coexpressed with Myc-tagged Cth2, Cth2-S64A/ S65A, or Cth2-S68A/S70A proteins, and Cth2 protein levels were determined by immu- noblotting before (I lanes) and after (P lanes) Grr1 immunoprecipitation (Fig. 9). Since these cells lacked a fully functional Grr1 protein, they preferentially accumulated slow-migrating Cth2 forms that may correspond to phosphorylated Cth2 (Fig. 9, input [I] lanes). More important, Grr1ΔFbox pulldown recruited wild-type Cth2, mostly in its phosphorylated form, whereas the interaction was considerably reduced when the Cth2 S64/S65 or S68/S70 residues were mutagenized to alanine (Fig. 9, immunoprecipitated [P] lanes). These results strongly suggest that Grr1 specifically recognizes phosphorylated Cth2 and promotes its degradation by the proteasome.
These results also suggested that Hrr25 kinase could influence Cth2 protein stability in vivo. To test this possibility, we grew wild-type and hrr25Δ mutant cells under iron-depleted conditions and performed Cth2 protein stability assays. We observed that the deletion of the HRR25 gene augmented Cth2 protein half-life from 12 min to more than 2 h (Fig. 10B). To further address Cth2 regulation by Hrr25, we expressed wild-type HRR25 and its catalytically inactive HRR25-K38A allele under the control of the GAL1 promoter under iron-depleted conditions and determined Cth2 stability (18). In agreement with our previous observations, the expression of HRR25 promoted the degradation of Cth2 (Fig. 10C, HRR25 cells in galactose-containing medium), whereas Cth2 protein half-life was more than 2 h in the absence of HRR25 expression (Fig. 10C, HRR25  cells in glucose-containing medium). Importantly, the expression of the HRR25-K38A kinase-dead allele did not enhance Cth2 protein turnover in galactose-containing medium (Fig. 10C, HRR25-K38A). These results demonstrate that Cth2 phosphorylation by the casein kinase Hrr25 is essential to promote Cth2 protein degradation.

DISCUSSION
The adaptation of eukaryotic cells to iron deficiency requires sophisticated regulatory mechanisms. In S. cerevisiae, the posttranscriptional factor Cth2 is tightly regulated to allow the appropriate utilization of the iron cofactor in specific metabolic pathways (3,11,12). Cth2 is expressed at very low levels in iron-replete conditions, whereas the Aft1 and Aft2 transcription factors strongly activate CTH2 expression upon iron limitation (3,19). However, the iron-deregulated CTH2 overexpression achieved with a constitutive or galactose-inducible promoter causes toxicity (5,10). Therefore, CTH2 expression is limited by a negative-feedback regulation in which Cth2 protein binds to AU-rich elements (AREs) within its own transcript and promotes its degradation (12). In the absence of this autoregulatory mechanism, yeast cells increase Cth2 protein abundance and diminish fitness in iron-limiting conditions (12). Here, we have shown that Cth2 expression is also controlled through its specific phosphorylation and degradation by the proteasome. When Cth2 serine residues are mutagenized or a component of the E3 ubiquitin ligase complex SCF Grr1 that recognizes phospho-Cth2 is absent, Yeast hrr25Δ (SPY892) cells transformed with either pGAL1-3HA-HRR25 or pGAL1-3HA-HRR25-K38A plasmid were grown overnight in SCRaf-ura-leu, and then cells were transferred to either SC-ura-leu (Glucose) or SC medium with galactose, and without uracil and leucine (SCGal-ura-leu) (Galactose) with 100 M BPS to exponential phase. For panels B and C, CHX was added to stop protein translation, and Cth2 protein levels were extracted and analyzed as described in the legend of Fig. 1C. The average half-life was calculated (n ϭ 2), and a representative immunoblot is shown.
Cth2 protein stability increases, leading to strong growth defects in iron-deficient cells. Therefore, the strict control of Cth2 expression levels is crucial during iron deficiency.
The molecular reasons underlying Cth2 overexpression toxicity have not been fully elucidated. This work and previous studies have shown that mutagenesis or deletion of Cth2 tandem zinc fingers (TZFs), which mediate specific binding to AREs within the target mRNAs, rescues the slow growth phenotype of Cth2-overexpressing cells (5,10). These results suggest that the increased levels of a functional Cth2 protein could be accelerating the degradation of target mRNAs. We do not observe an enhanced turnover of Cth2 targets, including SDH4, HEM15, and the essential transcripts RNR2 and RLI1 (Fig. 3). However, we cannot dismiss the possibility that changes in mRNA levels of other transcripts are responsible for detrimental growth. The nucleocytoplasmic shuttling of Cth2 protein is required to promote targeted mRNA turnover (11). However, Cth2 does not contain any nuclear export signal and relies on its cotranscriptional binding to ARE-containing mRNAs to exit the nucleus, and consequently, Cth2 TZF mutants are trapped in the nucleus (11). These observations raise the possibility that the overexpression of Cth2 could be titrating components of the mRNA degradation machinery during its nucleocytoplasmic shuttling, whereas the disruption of its TZF would trap Cth2 in the nucleus and rescue cellular viability. Deciphering strategies that separate shuttling from mRNA degradation would help to differentiate between these two possibilities.
Here we demonstrate that the SCF Grr1 E3 ubiquitin ligase complex plays a key function in yeast iron homeostasis by recognizing phosphorylated Cth2 protein and targeting it for degradation by the proteasome. The physiological relevance of SCF Grr1 regulation is corroborated by the severe grr1Δ growth defect under iron-deficient conditions, which is rescued by CTH2 deletion (Fig. 8). In addition to Cth2, a recent global study has also identified Dre2, a component of the cytosolic iron-sulfur protein assembly machinery, as a Grr1 target (14). The implication of E3 ubiquitin ligases in iron metabolism extends to mammals, where the SCF FBXL5 complex promotes the proteasomal degradation of the iron regulatory proteins IRP2 and apo-IRP1 (20,21), and higher plants, where the RING-type BRUTUS and HRZ proteins play central roles in the iron homeostasis of Arabidopsis and rice, respectively (22,23). Collectively, these data reinforce the notion that E3 ubiquitin ligase complexes occupy a privileged position in the regulation of eukaryotic iron homeostasis.
The shuttling of Cth2 protein is essential for targeted mRNA degradation (11). Cth2 enters the nucleus in order to cotranscriptionally associate with the mRNAs that will be specifically degraded later (11). Regarding subcellular distribution, we have previously reported that Cth2 proteins lacking their 89 amino-terminal amino acids still localize to both the nucleus and cytoplasm (11), which suggests that the mutagenesis of the S64/S65/S68/S70 residues does not alter Cth2 nucleocytoplasmic shuttling. Consistent with this, we show here that the phosphorylation of yeast Cth2 protein at a patch of serine residues (S64/S65/S68/S70) does not alter targeted mRNA turnover but Cth2 protein stability. Interestingly, the Cth2-C190R mutant protein, which cannot bind mRNAs and is trapped in the nucleus, is still phosphorylated (12). These results suggest that phosphorylation occurs before Cth2 associates with its mRNA targets. Furthermore, we previously observed that Cth2 protein accumulated in its phosphorylated form when forced to localize to the cytoplasm, whereas it was lost when Cth2 import into the nucleus was enhanced (11). Altogether, these data suggest that Cth2 is synthesized in the cytoplasm and that at least a fraction of it is phosphorylated even before its entrance into the nucleus. These results reinforce phosphorylation as a mechanism to control TZF-type proteins in different ways. Although much progress has been made, additional approaches are necessary to completely understand the regulatory role and the impact that phosphorylation exerts on TZF-containing proteins, particularly in an iron-deficient situation.
By using both in vitro and in vivo approaches, we have revealed that Cth2 is phosphorylated by the conserved casein kinase Hrr25, which is involved in a wide range of cellular functions, including ribosome maturation, vesicle trafficking and endocytosis, autophagy, DNA repair, and chromosome segregation during meiosis (24)(25)(26). We have observed that Cth2 protein is highly stable in hrr25Δ mutant cells, although its electrophoretic mobility is not altered after the addition of cycloheximide ( Fig. 10B and  C). These results suggest that Cth2 protein is still phosphorylated in hrr25Δ cells, implicating other kinases in Cth2 phosphorylation. Consistent with this, the separate mutagenesis of Cth2 S64/S65 and S68/S70 residues has uncovered that Hrr25 is responsible for the phosphorylation of only the S64 and S65 sites, whereas the other pair of serine residues seems more promiscuous (Fig. 10). Further observations highlight the interplay existing between the two pairs of serine residues. In particular, the stabilization observed for the Cth2 quadruple serine mutant was not as pronounced as the one displayed for the Cth2 double mutants (Fig. 5; see also Fig. S2D in the supplemental material). Consistent with this, the Cth2-S64A/S65A/S68A/S70A steadystate protein levels increased only 1.5-fold, and the growth defect was not as severe as the growth defect observed for the cells expressing the Cth2-S64A/S65A and Cth2-S68A/S70A double mutants (Fig. S2). In any case, these results indicate that the destabilization of Cth2 protein requires the simultaneous phosphorylation at both serine pairs by different kinases. Although we still do not understand its physiological significance, we could speculate that this dual targeting would allow the stabilization of Cth2 protein in an active state in response to certain situations by altering only one of these phosphorylation events. The further characterization of phosphorylation events within TZF-containing proteins would help to understand how these regulatory factors are posttranslationally controlled.

MATERIALS AND METHODS
Yeast strains and growth conditions. The Saccharomyces cerevisiae strains used in this work are listed in Table S1 in the supplemental material. To delete CTH2 and HRR25 genes in particular strains, we generated cth2::hphB, cth2::HisMX6, and hrr25::KanMX4 cassettes as previously described (27,28). Yeast precultures were grown to exponential phase at 30°C in synthetic complete medium (SC) lacking specific requirements. To regulate iron availability, the cells were incubated in SC (with Fe) or SC supplemented with 100 M concentration of the Fe 2ϩ -specific chelator bathophenanthroline disulfonic acid disodium or bathophenanthroline disulfonic acid (BPS) (Merck) (without Fe). To perform growth assays in solid media, yeasts were spotted at an optical density at 600 nm (OD 600 ) of 0.1 and after 1:10 and 1:100 dilutions in SC containing 0.7 to 1.1 mM concentrations of the Fe 2ϩ -specific chelator Ferrozine (Merck).
Plasmids. The plasmids used in this work are listed in Table S2. Flag-tagged CTH2 was amplified from pSP413 plasmid and cloned into pRS426 to obtain pSP421 plasmid. Two copies of the Myc epitope were introduced into a NotI restriction site after the CTH2 start codon to generate a Myc 2 -tagged Cth2 allele, which was cloned into pRS413 to obtain pSP886 plasmid. Plasmids pSP413 and pSP886 were used as the templates to mutagenize Cth2 residues S64, S65, S68, S70, and C190 by using either the overlapping extension method (3) or the Phusion site-directed mutagenesis kit (Thermo Fisher Scientific). In all cases, BamHI and XhoI restriction sites were used to clone CTH2 amplified sequences. Other plasmids used in this work have been described previously (3,4,12,14,29) (Table S2). All PCR amplifications were performed with the Phusion polymerase (Thermo Fisher Scientific).
Mass spectrometry. Immunoprecipitates were precipitated by the addition of ice-cold acetone and digested by the sequential addition of lys-C and trypsin proteases as previously described (30). Peptide digests were then analyzed by liquid chromatography coupled to tandem mass spectrometry (LC-MS/ MS) on an LTQ-Orbitrap XL instrument (Thermo Fisher Scientific) (30). Phosphorylation sites were identified with the ProLuCID database search algorithm using a differential modification search that considered a mass shift of 79.9663 on serines, threonines, and tyrosines (31). Phosphosite validation and localization were performed using Debunker and Ascore, respectively (32,33).
Western blotting. Total protein extracts were obtained by using the alkali method (34). Equivalent amounts of protein were resolved on SDS-polyacrylamide gels and transferred to nitrocellulose membranes. Ponceau S staining was used to assess protein transfer and loading. The Flag epitope was detected using a horseradish peroxidase (HRP)-conjugated antibody. The primary antibodies used in this study included anti-Flag (Merck), anti-c-myc (Roche Applied Science), anti-Pgk1 (Thermo Fisher Scientific), anti-glutathione S-transferase (anti-GST) (Merck), and anti-peroxidase-antiperoxidase (anti-PAP) (Merck). Immunoblots were developed with horseradish peroxidase (HRP)-labeled secondary antibodies and the ECL Select Western blotting detection kit (GE Healthcare Life Sciences). Images were scanned with an ImageQuant LAS 4000 mini biomolecular imager (GE Healthcare Life Sciences), and specific signals were quantified and processed with ImageQuant TL analysis software (GE Healthcare Life Sciences).