In
S. cerevisiae, the role of Ppz1 is more prominent than that of Ppz2. Ppz1 regulates monovalent cation homeostasis in two ways: by inhibiting K
+ uptake in both Trk-dependent and independent manner
[14][15] and by repressing the expression of the Na
+/ K
+-ATPase encoded by the
ENA1 gene
[16][17]. A role for Ppz1 in translation fidelity has also been reported through the regulation of Tef5, a subunit of translation elongation factor EF-1A
[18]. Similarly, Ppz1 appears to be involved in the dephosphorylation of ubiquitin at S57, within the context of the regulation of endocytic trafficking and ubiquitin turnover
[19] and in the dephosphorylation of the arrestin Art1 in a process linked to the endocytosis of the methionine transporter Mup1
[20].
Ppz1 function is regulated in vivo by two proteins that act as inhibitory subunits, Hal3 and Vhs3. Both proteins bind to the catalytic domain of the phosphatase and inhibit its enzymatic activity
[21][22][23], although the role of Hal3 appears more prominent in vivo. Hal3 and Vhs3 are moonlighting proteins; they function not only as Ppz1 inhibitors but also, together with Cab3, as components of the enzyme phosphopantothenoylcysteine decarboxylase (PPCDC), which catalyzes a key step in coenzyme A biosynthesis
[24].
2. The N-Terminal Extension of Ppz1 Is an Important Factor for Toxicity
As mentioned above, the
S. cerevisiae genome encodes two PPZ paralogs,
PPZ1 and
PPZ2, which are very closely related in sequence in their C-terminal catalytic domain (85.9% identity, 89.9% similarity). However, their N-terminal extensions are far more divergent. Early evidence showed that strong overexpression of the C-terminal catalytic moiety of Ppz1 (348 residues) was also highly toxic
[25] and, more recently, it was reported that the activity of Ppz1 was a requirement for toxicity
[26]. In addition, overexpression of the Ppz1 and Ppz2 inhibitory subunit Hal3 was able to fully restore normal cell growth
[21]. Therefore, it was reasonable to test if overexpression of
PPZ2 would also be detrimental to cell growth. Yet, expression of Ppz2 under the same conditions used for Ppz1, driven from a
tetO-regulatable promoter in a multicopy plasmid, showed very little effect, if any, on cell growth
[27]. In fact, this lack of toxicity was not surprising since the work of Makanae and coworkers
[28] revealed that yeast cells could tolerate a higher load of plasmid bearing
PPZ2 than
PPZ1. It must be noted, though, that the levels of Ppz2 protein in the
tetO-based experiments were significantly lower than that of Ppz1, raising the possibility that the lack of toxicity could be due to insufficient accumulation of the phosphatase. However, a hybrid version carrying the N-terminal half of Ppz1 and the catalytic domain of Ppz2 was as toxic as Ppz1 even if it was expressed at levels similar to that of native Ppz2
[27]. This indicated that the disordered Ppz1 N-terminal region is an important determinant for toxicity and that this feature is not shared by the extension found in Ppz2. In fact, the functional relevance of the Ppz1 N-terminal extension was hinted at long ago by Clotet and coworkers
[25], who reported that particular deletions in this region affected the ability of the protein to rescue specific phenotypes of a
ppz1 mutant strain.
Current efforts in our laboratory aim to identify the precise structural features in the Ppz1 N-terminal region that are crucial for toxicity.
3. PPZ1 Overexpression Affects Protein Synthesis
It has been recently demonstrated that the excess of Ppz1 activity impairs protein synthesis, probably at the initiation step. This might be due to a possible role of Ppz1 in regulating ribosome biogenesis and function since the endogenous (non-overexpressed) Ppz1 protein has been co-purified with ribosomal proteins and localized in a wide range of ribosomal fractions covering the 40S and 60S ribosomal subunits, the 80S ribosome, and polysomes
[26]. In addition, overexpression of
PPZ1 altered the ribosomal function, causing an important reduction in the polysome content, which was normalized when the Ppz1 inhibitor Hal3 was overexpressed.
Further evidence about the effect of an excess of Ppz1 activity on protein synthesis came from the identification of changes in the phosphorylation state of diverse proteins involved in this process
[29]. 5 proteins required for translation initiation were found hyperphosphorylated, while residues in 10 proteins were significantly dephosphorylated, including those in Rps6, a conserved component of the small (40S) ribosomal subunit
[29]. Regarding Rps6, overexpression of Ppz1 triggered fast dephosphorylation of S232 and S233 (decrease to 45% after 1 h), which continued to decline until very low levels after 4 h (16%)
[29]. It is known that phosphorylation of S232 and S233 in response to nutrients, among other stimuli, occurs in a TORC1- and TORC2-dependent manner
[30][31]. Phosphorylation of these sites is mediated by the Sch9 and Ypk3 protein kinases
[30][32][33], and it has been reported that Glc7, together with its regulatory subunit Shp1 dephosphorylate S232 and S233
[30]. Moreover, expression of an unphosphorylatable S232/S233 mutant version of Rps6 impeded cell growth with an almost 30% decrease in the cells proliferation rate and, although it did not affect the polysome to 80S monosome ratio, a defect in the biogenesis of the 40S small subunit was observed
[30]. Therefore, it is plausible that aberrant dephosphorylation of Rps6 could contribute to Ppz1 toxicity.
4. An Excess of Ppz1 Interferes with Normal Adaptation to Low Glucose
The effects of mild expression of Ppz1 in carbon sources other than galactose were investigated by placing the
PPZ1 ORF under the control of the doxycycline-repressed promoters
tetO2 and
tetO7, present in strains MLM03 and MLM04
[26]. Remarkably, the deleterious effect on cell growth under these circumstances was only evident in medium containing low glucose concentrations but not in the presence of 2% glucose. The toxic effect was even more dramatic when cells were grown in other carbon sources, such as galactose, wherein growth was completely abolished even with 2% of the sugar in the medium
[26]. These results suggested that the defect in cell proliferation induced by increased levels of Ppz1 was, to some extent, dependent on the availability of glucose.
Several repressors, such as Mig2, Mig1, Nrg1 and Nrg2, are known to downregulate gene expression in the presence of glucose. Under glucose depletion conditions, however, the Snf1 kinase is activated by phosphorylation at T210 and, in turn, phosphorylates Mig1 at S311. Under these conditions, Snf1 also phosphorylates Hxk2 at S15, thus preventing its nuclear localization and its interaction with Mig1, promoting the exit of both proteins from the nucleus and releasing their target genes from transcriptional repression
[34]. Dephosphorylation of Mig1 S311 and Hxk2 S15 was attributed to the activity of the Glc7-Reg1 complex
[34]. Notably, phosphoproteomic experiments demonstrated fast and sustained dephosphorylation of Mig1 at S311 and of Hxk2 at S15 when PPZ1 expression was induced, as well as concomitant dephosphorylation of Snf1 T210. These changes were accompanied by substantial retention of Mig1 in the nucleus
[29].
Snf1 is also dephosphorylated and inactivated by the Glc7-Reg1 complex, and, as a result, Snf1 can no longer phosphorylate the Mig transcriptional repressors
[35][36]. About 60 phosphorylated residues have been detected in Reg1 according to the data collected by the SGD
[37]. Residues in its N-terminal segment (1–400 aa) were phosphorylated under glucose-limiting conditions in an Snf1-dependent manner, thus regulating the Reg1-Bmh1/2 interaction
[35][38], and this N-terminal region acts as an in vitro substrate for Ppz1
[39]. It is suggestive that
ourthe phosphoproteomic data show that Reg1 is significantly dephosphorylated (about 50%) in the N-terminal residues S346 and S349 after 1 h from the induction of the Ppz1 overexpression, remaining dephosphorylated for at least 4 h
[29]. The relevance of these phosphorylation sites of the Reg1 N-terminal segment in the glucose repression process is still unknown. Taken together, these data show that the excess of Ppz1 might disrupt the proper adaptation of yeast cells to conditions of low glucose by acting on several targets of the pathway involved in the regulation of the Mig transcriptional repressors. This may proceed by affecting physiological Ppz1 targets or by interfering with natural substrates for Glc7.
5. Ppz1 Overexpression Affects Diverse Signaling Pathways
Genome-wide transcriptomic and phosphoproteomic analyses of the short-term response to Ppz1 overexpression using the ZCZ01 strain
[29] revealed a widespread impact on the yeast cell. Nearly 1300 genes (about 20% of the genome) suffered a significant modification in their expression. These included cyclins
CLN1,
CLN2,
CLB1, and
CLB5, which are repressed, consistently with the previously reported halt in the G1 phase caused by excess of Ppz1
[26][40]. Similarly, the mRNA levels of many genes involved in ribosome biogenesis showed a very early decline (30 min after Ppz1 induction). Strikingly, the entire Bas1-Bas2-dependent ADE pathway, required for de novo purine biosynthesis, was also repressed. Metabolomic analyses revealed a marked increase in ATP and GTP levels, as well as in the adenylate pools and in the adenylic energy charge, thus suggesting that the transcriptomic effect could be due to a negative feedback loop caused by the accumulation of its final products. Likewise, the levels of dNTPs were also augmented. These results indicated that the halt in growth caused by high levels of Ppz1 is not due to the scarcity of building blocks for DNA synthesis or a lack of energy.
Cells overexpressing Ppz1 suffer oxidative stress. This conclusion was deduced from the transcriptomic data (strong induction of known oxidative stress-responsive genes) and further confirmed by the determination of reactive oxygen species (ROS). Because it is well known that oxidative stress is responsible for DNA damage
[41], this could explain the formation of Rad52 foci, a common response to DNA damage, observed in Ppz1-overexpressing cells
[29]. In fact, it has been reported that respiration is activated in response to DNA damage, leading to increased ATP production and to elevated dNTP levels
[42], which are required for efficient DNA repair and cell survival. This link may explain the effects on nucleotide metabolism described above. In addition, oxidative stress constitutes one of the cues that activate the Gcn2 kinase, leading to the phosphorylation of eIF2α and to the reduction in protein synthesis
[43]. Thus, it is conceivable that oxidative stress may contribute to the above-mentioned alteration in translation found in Ppz1-overexpressing cells.
6. The Alteration of Monovalent Cation and pH Homeostasis Contributes to Ppz1 Toxicity
The relationship between Ppz1 and monovalent cation homeostasis was explored long ago, prompted by the finding that the
ppz1 and
ppz1 ppz2 deletion strains were highly tolerant to sodium and lithium cations
[16]. This effect was initially attributed to a depression of the
ENA1 Na
+,K
+-ATPase gene, involved in response to salt stress
[16][17], but shortly afterward, the contribution of negative regulation of potassium influx through the Trk1/Trk2 high-affinity potassium transport system was proposed
[14]. Because a strain lacking the Trk1 and Trk2 transporters cannot grow in standard media unless supplemented with potassium
[44], it was reasonable to assume that the growth blockage caused by an excess of Ppz1 might derive from a strong inhibition of Trk-mediated potassium influx.
However,
we recent
ly research showed
[45] that cells overexpressing Ppz1 do not grow even in the presence of potassium amounts able to sustain the growth of a
trk1 trk2 strain. This result clearly argued against Trk inhibition being the basis of Ppz1 toxicity. Interestingly, deletion of the gene encoding the Na
+,K
+/H
+ plasma membrane antiporter Nha1 resulted in improvement of growth in Ppz1-overexpressing cells, placing again the focus on monovalent cation homeostasis. Nha1 is a housekeeping protein that exports Na
+ and K
+ in exchange for protons. In conjunction with the Ena1 ATPase, it enables cell growth in the presence of high concentrations of toxic monovalent cations
[46][47]. This antiporter also plays a key role in the regulation of internal pH and membrane potential
[46][48] and is involved in the very early response to osmotic stress
[49]. Indeed, overexpression of Ppz1 promoted intracellular acidification and depletion of intracellular potassium content, and these effects were partially counteracted in cells lacking Nha1
[45]. In addition, the beneficial effect of the
NHA1 deletion vanishes when cells are grown at pH near neutrality. All this evidence suggests that high levels of Ppz1 lead to the hyperactivation of Nha1 and a subsequent exacerbated influx of H
+ in exchange for K
+ ions. Several pieces of evidence support this scenario: (i) Nha1-mediated potassium efflux activity in cells overexpressing Ppz1 is far higher than in control cells, and (ii) as expected, expression of native Nha1 eliminates the growth improvement derived from the
NHA1 deletion, whereas expression of mutated versions of Nha1 (such as the D77N variant), unable to mediate H
+/cation exchange, did not.
It has been reported that Nha1 S481 is responsible when phosphorylated by an unknown kinase for binding to yeast 14-3-3 proteins and that this decreases Nha1 activity
[50][51]. This is suggestive because it is known that the mutation S481A, which prevents phosphorylation of this Ser residue, significantly increases Nha1-mediated cation efflux
[50]. In agreement,
wescholars identified Nha1 S481 as one of the early targets for dephosphorylation in Ppz1 overexpressing cells
[29], raising the possibility that such dephosphorylation may activate Nha1 and contribute to the growth arrest phenotype.
The fact that deletion of
NHA1 did not fully normalize growth rate and K
+ and H
+ intracellular contents in cells overexpressing Ppz1 indicates that additional targets affected by Ppz1 overexpression and relevant for monovalent cation homeostasis must exist. Although their nature is still obscure, several pieces of evidence suggest that the essential H
+-ATPase Pma1 could be one of them.
WeScholars observed that cells overexpressing Ppz1 failed to acidify the medium properly even in cells where
NHA1 was deleted and that this effect was not due to lower-than-normal Pma1 levels
[45]. This suggests that Pma1 could be inhibited in Ppz1-overexpressing cells. Such inhibition could contribute to the above-mentioned cytosolic acidification. Pma1 can be phosphorylated in vivo at multiple sites, and it has been demonstrated that this modification controls Pma1 function. Thus, phosphorylation of S911 and T912 and, to a lesser extent, of S899 has been deemed important for the activation of Pma1
[52][53]. In this regard,
we reported that all three residues were rapidly dephosphorylated upon overexpression of Ppz1
[29]. Because the ATPase activity of Pma1 is the major consumer of cellular ATP, its inactivation may explain the increase in ATP levels in cells overexpressing Ppz1.
7. Hal3 and the subcellular localization of Ppz1
In agreement with the inhibitory role of Hal3 on Ppz1 function, expression of Hal3 from a multicopy plasmid fully restored normal growth in cells overexpressing Ppz1 either from its own promoter or from the strong GAL promoter [21][26]. Ppz1 is mainly located at the cell periphery, likely interacting with the plasma membrane by means of its N-myristoylated Gly2 [20][25][27][54][55]. Consequently, a G2A Ppz1 variant is fully cytosolic. Thus, the simplest scenario was that Hal3 interacts with the excess of Ppz1 and inhibits its phosphatase activity, thus avoiding harmful dephosphorylation of key cellular targets. However, we have found very recently [56] that episomal expression of Hal3 in cells overexpressing Ppz1 triggers the translocation of the phosphatase from the plasma membrane to intracellular structures that could be identified as the vacuolar membrane and, less often, the endoplasmic reticulum. Colocalization experiments showed that Hal3 accompanies Ppz1 in these intracellular locations.
Translocation of Ppz1 to internal membranes is a crucial event for the ability of Hal3 to counteract Ppz1 toxicity. Indeed, mutation of the VPS27 gene aggravated the cell growth defect even when Ppz1 was overexpressed at moderate levels and greatly impaired the ability of Hal3 to normalize the growth of these cells [56]. VPS27 encodes a component of the ESCRT-0 complex required for various cellular functions, including the trafficking of membrane proteins to the vacuole for degradation .Whereas the mechanism underlying Ppz1 relocalization is still obscure, current evidence suggests that this is more an exceptional instrument to avoid Ppz1 toxicity, linked to excessive phosphatase activity, than a normal regulatory process. Relocalization of the phosphatase does not occur with a non-catalytic version of Ppz1 bearing the R451L mutation, which normally binds to Hal3, suggesting that critical toxic events must occur. The nature of these events is not known. However, as described above, overexpression of Ppz1 results in a fast and strong drop in intracellular pH [45], and this may, in turn, lead to increased Ppz1-Hal3 interaction [54]. Therefore, a reasonable hypothesis would be that the internal acidification caused by deregulation of Nha1, and perhaps Pma1, could trigger this phenomenon. While Ppz1 is directed in most cases to the vacuole, it is rarely detected in the vacuolar lumen, and the levels of the phosphatase are very high even many hours after induction [56]. Therefore, the translocation event is a safeguard mechanism that is not based on the degradation of the toxic protein. It is likely that localization at vacuolar and ER membranes occurs by means of the N-terminal myristoyl moiety since the G2A variant is not recruited to internal membranes [56] while exhibiting notable toxicity [27].