Manganese-induced cellular disturbance in the baker’s yeast, Saccharomyces cerevisiae with putative implications in neuronal dysfunction

Manganese (Mn) is an essential element, but in humans, chronic and/or acute exposure to this metal can lead to neurotoxicity and neurodegenerative disorders including Parkinsonism and Parkinson’s Disease by unclear mechanisms. To better understand the effects that exposure to Mn2+ exert on eukaryotic cell biology, we exposed a non-essential deletion library of the yeast Saccharomyces cerevisiae to a sub-inhibitory concentration of Mn2+ followed by targeted functional analyses of the positive hits. This screen produced a set of 43 sensitive deletion mutants that were enriched for genes associated with protein biosynthesis. Our follow-up investigations demonstrated that Mn reduced total rRNA levels in a dose-dependent manner and decreased expression of a β-galactosidase reporter gene. This was subsequently supported by analysis of ribosome profiles that suggested Mn-induced toxicity was associated with a reduction in formation of active ribosomes on the mRNAs. Altogether, these findings contribute to the current understanding of the mechanism of Mn-triggered cytotoxicity. Lastly, using the Comparative Toxicogenomic Database, we revealed that Mn shared certain similarities in toxicological mechanisms with neurodegenerative disorders including amyotrophic lateral sclerosis, Alzheimer’s, Parkinson’s and Huntington’s diseases.


Results
To identify pathways that are influenced by Mn exposure, we screened for gene deletion strains that demonstrate increased sensitivity to Mn 2+ using the yeast non-essential gene deletion array (yGDA), Fig. 1A. These types of screens can provide a CGI profile for a target toxin and contribute to our knowledge of the cell's global stress responses to that toxin. To this end, we performed sensitivity analysis by screening approximately 4700 gene deletion strains, under two conditions (presence and absence of MnCl 2 ), for a total of approximately 28,000 individual analyses. Sensitivity was investigated by determining the relative colony growth size in the presence/absence of the target compound. In this way, we identified 68 gene deletion mutants with significantly altered growth profiles (Supplementary Material, Table SM1), of which 43 were confirmed to display high sensitivity to a sub-inhibitory concentration (1.35 mM) of Mn (a high concentration of a bioactive/toxic compound where growth of a wildtype Functional proteomic and Gene Ontology (GO) analysis of sensitive mutants identifies multiple pathways including protein synthesis. To have a comprehensive coverage of the hits identified in our sensitivity screen, the String database was used to expand the obtained CGI profile for Mn on the basis of PPI data 38 . String uses physical interactions and functional associations to study a defined set of proteins and expand it by including associated proteins. In this way, the network of functional interactors for Mn was increased to approximately 600 edges (p-value < 1.0e- 16), of which more than 85% are known interactions. For example, approximately 87% of the interactions have been experimentally verified and almost 98% are from curated databases. A schematic representation of these interactors is shown in Fig. 2.
Enrichment of cellular pathways represented by the expanded list of proteins is shown in Table 1. As expected, proteins associated with cellular development and protein metabolism were highly enriched 39 . Particularly, the ER associated activities have been highly connected to Mn toxicity 21,[40][41][42][43] . However, a direct connection between protein synthesis and Mn toxicity has not been previously reported. This led us to further investigate the influence of Mn toxicity on protein biosynthesis.
To study if Mn may also affect gene expression at the translation level, total RNA levels were analyzed. In response to the presence of Mn, we observed decreased levels of rRNA molecules (Fig. 3A). After treating the cells with 1.5 mM and 3.0 mM Mn for 45 minutes, total rRNA levels are reduced in a dose-dependent manner. We repeated this experiment by increasing the duration of Mn treatment to 3 and 24 hours. We observed similar results indicating that total levels of rRNA molecules seem to be reduced in response to Mn. Next, we investigated ribosome profiles of cells in response to Mn treatment (Fig. 3B). We observed a reduction in the pool of polysomes in response to treatment with 3 mM Mn for 1 hour, in addition to an increase in the pool of 80S ribosomes. Reduction in polysomes is interpreted as a decrease in the number of ribosomes that are active and engaged in synthesizing proteins. An increase in 80S monosomes is generally regarded as stalled initiation of translation. Treatment of the cells with Mn for 24 hours, resulted in additional reduction in polysomes in comparison to control conditions. Lastly, using an expression vector we investigated the expression of β-galactosidase, used as a reporter, in response to Mn. In a dose-dependent manner, the presence of Mn 2+ reduced the expression of β-galactosidase (Fig. 3C). Importantly, this trend differs significantly when other divalent ions such as Ca 2+ , Mg 2+ and Zn 2+ are used suggesting that the decreased rate of translation is unique to Mn 2+ stress and not a general byproduct (Supplementary Material, Fig. SM2). Altogether, these follow-up investigations connect Mn toxicity to the process of protein biosynthesis, which were identified as an enriched cellular process in our GDA analysis.
Manganese-induced disturbance of processes that converge to protein biosynthesis in yeast, which mimics molecular pathways associated with neurodegeneration. It has been postulated that the ER has various active domains and membrane contact sites that are required for multiple cellular processes including protein and lipid biosynthesis, calcium regulation, and the exchange of macromolecules 22 . In this study, multiple approaches including GDA and PPI analysis (Figs 1 and 2), GO ontology enrichment (Table 1), Figure 1. Representative illustration of Mn-induced disruption in yeast gene deletion array after exposure to MnCl 2 (1.35 mM) for 24 hours. Mutants that showed a relative reduction in growth (sensitivity) of 30% or more were selected as hits (p < 0.05). Examples of hit stains are indicated using yellow circles.
www.nature.com/scientificreports www.nature.com/scientificreports/ total rRNA analysis, ribosome profiling and a β-galactosidase reporter assay (Fig. 3), together suggest that the Mn induces a significant perturbation of protein biosynthesis and associated pathways.
To augment this finding and more-closely study individual participants, we selected several genes linked to processes that converge on protein biosynthesis and then analyzed their relative transcription levels using qPCR (Fig. 4). Indeed, the presence of Mn induced alterations in the expression of these genes. For example, we observed decreased expressionof key translation initiation factor eIF4A (TIF1) and upregulation of the essential translation elongation factor eIF-5A (HYP2). Additionally, several other genes associated with translation and/ Analysis performed using the String database. Network properties are as follows: The minimum required interaction score, to be included at the predicted network, was accepted with a threshold on the high confidence equal 0.7; Number of nodes: 143; Number of edges: 594; Expected number of edges: 287; Average (avg) node degree: 8.31; avg. local clustering coefficient: 0.652; PPI enrichment p-value: <1.0e-16. Nodes and edges represent proteins and PPIs, respectively. Red nodes (protein processing in endoplasmic reticulum); blue nodes (metabolic pathways); dark green nodes (N-Glycan biosynthesis); cyan nodes (cell cycle); yellow nodes (ubiquitin-mediated proteolysis); orange nodes (meiosis); maroon nodes (DNA replication); purple nodes (amino acid biosynthesis); magenta nodes (arginine and proline metabolism); lime green nodes (alanine, aspartate and glutamate metabolism). The protein with black points in the center represent the MAPK signaling pathway. Proteins that are not connected to at least one partner are not shown.  (Fig. 1). This network was generated using the String database.
www.nature.com/scientificreports www.nature.com/scientificreports/ or ribosome biogenesis had significantly altered levels of transcription including the downregulation of NSR1, NOP1 and up-regulation of the gene RPS15. However, protein biosynthesis is a complex process that involve other pathways. For instance, we observed perturbation in the expression of genes such as UFD1, UFD2, STT3, DSK2. Additionally, OST2 and OST6, involved in post-translational processing in the ER (PPER), were significantly downregulated and upregulated, respectively. Similar alterations are inferred for the N-glycan biosynthesis pathway, which is partially regulated by OST2 and OST6 activity. We also observed decreased-expression of ARG3 involved in amino acid biosynthesis, including arginine, proline, alanine, aspartate and glutamate metabolism and increased expression of URA2.
Also of interest, genes such as CDC20 and UBC11, which are related to ubiquitin proteolysis were significantly disrupted. Protein synthesis underpins much of cell growth and multiplication 44 . Coincidently, impairment of CDC20 suggests a direct relationship among dysregulation of protein biosynthesis and alteration of MAPK signaling pathways, cell cycle and DNA replication respectively.

Mn-induced molecular impairment in yeast mimics pathways associated with neurodegeneration.
We identified alterations in various pathways that lead to impairment of protein biosynthesis, which is a conflicting topic in neurodegeneration research. Some reports have viewed this as a therapeutic target, while others suggest that it provokes the onset of certain neurodegenerative disorders 45 . Due to the conservation of the key cellular processes and genes, yeast has been used as a model organism to study human neurodegenative diseases 34 . In this sense, we conducted an additional analysis of the pathways affected by Mn using both the String database and the Comparative Toxicogenomics Database -CTD 46 , which permits the development of novel hypotheses about the relationships between chemicals and diseases 47 . The results are shown in Fig. 5.
We verified that approximately 31% (44 proteins/genes, Supplementary Material - Fig. SM 3) of the inferred network for hits (genes) affected by Mn (Fig. 2) have homologues in human, of which approximately 73% (32 proteins/genes, Supplementary Material - Fig. SM 3) are potentially linked to neurodegeneration, according to the CTD 46 . The genes affected by Mn suggest that this cation-induced toxicity in yeast involves disruption of several pathways which together lead to impairment of protein biosynthesis (Fig. 5A). These alterations shared characteristics with pathways involved in neurodegenerative diseases (Fig. 5B). For example, we identified that the MnCl 2 affects the CDC20 involved in the metabolism of proteins 46 ,the cell cycle and MAPK signaling pathways and is potentially involved in the development of neurodegenerative disorders such as AD, ALS, HD and PD 46 . At the same time, UFD1 is associated with protein processing in endoplasmic reticulum and potentially linked to the evolution of AD, ALS and PD 46 . Altogether; our findings and subsequent inferences suggest that the developmental impairment induced by Mn, according to cell cycle disruption, is mainly influenced by collective perturbation of pathways that converge to disturbance of protein biosynthesis. We demonstrated a decrease in total RNA, www.nature.com/scientificreports www.nature.com/scientificreports/ polysome and β-galactosidase activity as well as potential alterations in the expression of genes directly associated with translation (HPY2 and TIF1) and ribosome biogeneses (RPS15, NSR1 and NOP1). Together findings suggest a plausible hypothesis for Mn-induced neurotoxicity and neurodegeneration (Fig. 5B). Further analysis in higher-order animal models is needed to confirm this theory.

Discussion
The role of Mn in toxicity, particularly in relation to neurotoxicity and neurodegeneration disorders, remains unclear with several proposed hypotheses 12 . The CTDdescribes Mn as an essential trace element, with possible connections to approximately 570 biological processes and/or pathways 48 . In this work, using a functional genomics and systems biology approach, we observe a connection between Mn and cellular processes such as cell cycle progression, cell signaling, and protein metabolism. Agreeably, previous studies have suggested the possibility that Mn may disturb cellular development processes 49,50 as well as the flow of genetic information that could influence protein synthesis [14][15][16] , including ER stress 21,[40][41][42][43] .
Our global chemical-genetic sensitivity screen, followed by GO term enrichment of interaction network of participants, suggest that Mn disturbed anabolic metabolism pathways In line with this, we provide evidence to suggest that disruptions in the biosynthesis of amino acids through decreased expression of ARG3 which is involved in the biosynthesis of arginine from ornithine carbamoyltransferase 51 . Previous works have suggested that ornithine deficiency causes hyperammonemia and neurotoxicity in humans 52 . Specifically, alteration of arginine and proline metabolism has been associated with development of ALS 53 . The impairment of amino acid biosynthesis can directly disrupt translation efficiency 54 , a process that is energetically very costly 23,24 . Furthermore, these events appear to be associated with inactivation of MAPK pathways that can lead to translation repression 23,55 , antiapoptotic activities 56 and/or cell cycle arrest 23,44,56 . Interestingly, we identified and inferred significant impairment of genes involved in the cell cycle and MAPK pathways (Figs 1, 2 and 4) such as CDC20 and CLB2, which correlated with our qPCR results. Cell cycle disruption has been associated with AD, PD and ALS 57 . , β-galactosidase expression assay and ribosome profile, which are an evidence of perturbation of protein biosynthesis and other associated pathways (protein processing in endoplasmic reticulum, metabolic pathways, N-Glycan biosynthesis, cell cycle, ubiquitin mediated proteolysis, amino acid biosynthesis, MAPK signaling pathway and translation control analysis) was performed by qPCR. Bars represent the mean value of at least 3 independent experiments and error bars represent (mean ± SEM). Preliminarily, we verified some trends to be different between Mn treatment and the control using t-test ( + p < 0.05). Then, we confirmed several significant differences by ANOVA two-way, followed of Bonferroni post-test (*p < 0.05).
www.nature.com/scientificreports www.nature.com/scientificreports/ At the same time, aberrations from strictly controlled of MAPK signaling pathway have also been implicated in the development of different human diseases including AD, PD and ALS 56,58 .
According to the results discussed above, Mn-induced toxicity in yeast appears to be associated with essential pathways linked to protein metabolism. In the current study we inferred that Mn may induce ER stress (Fig. 2), which was demonstrated through qPCR analysis showing up-regulation of OST6 and downregulation of OST2 (Fig. 4). This is in agreement with previous in vitro and yeast studies that suggested that the ATPase activity of ER gene SPF1 is compromised under exposure to Mn resulting in severe ER stress 43 . Other works have suggested that Mn-induced ER stress can be mediated through iron depletion, increased phosphorylation of the eukaryotic translation initiation factor 2α (phospho-eIF2α) 59 , activation of PERK and IRE1 signaling pathways 41,42 and ER tumefaction 60 . An RNA-Seq approach in Caenorhabditis elegans revealed that Mn induced both up and down-regulation of ER-related protein families (FKB and ABU) which are both implicated in ER stress 40 . ER stress can trigger a signaling reaction known as the unfolded protein response (UPR), which induces adaptive programs that improve protein folding. In certain neurodegenerative diseases such as AD, ALS, HD and PD, when the cell damage is irreversible, UPR can also activate apoptosis 61,62 .
Moreover, we found that Mn could potentially influence glycosylation throughOST2 and OST6 63,64 . Aminoglycoside antibiotics have been proposed to introduce errors in post-translational modifications such as glycosylation and protein misfolding that can lead to destabilized membranes and chronic stress 65 . Other studies suggest that alterations of SLC39A8 links Mn deficiency to inherited glycosylation disorders, specifically impairment of Mn-dependent enzymes activity, most notably the Golgi enzyme β-1,4-galactosyltransferase, which is essential for biosynthesis of the carbohydrates in glycoproteins 63 . Moreover, Golgi glycosylation defects may also be the result of Gdt1p/TMEM165 deficiencies that stem from Golgi Mn homeostasis defects 64 . Collectively, this evidence suggest that ER stress in yeast treated with Mn, may be associated with the impairment of N-glycan biosynthesis 66 , which could consequently lead to arrest the protein biosynthesis.
Additionally, ER stress can be exacerbated by the impairment of endosome-to-Golgi retrograde trafficking 67 . Since the retromer complex, comprised of vacuolar protein sorting, is essential to the bidirectional transport between the trans-Golgi network and endosomes. It is one of the key vesicular trafficking pathways in the cell 68 , particularly the transport of protein to endoplasmic reticulum 67,69,70 .
Vacuole protein sorting appears disrupted in the presence of Mn. VPS5 mutants are hypersensitive to Mn (Fig. 1B), and the PPI network analysis implicatedprotein/genes with similar function such as Vps35, Vps29, Vps17 and PEP8 (Fig. 2). Interestingly, previous studies have reported that Mn is linked to yeast VPS1, VPS53 and PEP8 71 . Vesicle transport is considered to play an important role in yeast and mammalian models of ALS as well 72   www.nature.com/scientificreports www.nature.com/scientificreports/ Alternatively, Gitler et al. 76 identified that YPK9 overexpression significantly rescued the ability of proteins to leave the ER and traffic to the Golgi, which reduced the toxic effects of α-syn intracellular accumulation and Mn toxicity, suggesting a close connection between genetic and environmental causes of neurodegeneration. Ypk9 is a yeast orthologue of human PARK9/ATP13A2, whose expression in animal models of PD is capable of rescuing neurodegeneration 76 . At the same time, Golgi dysfunction can lead to the rapid repression of rRNA and ribosomal proteins 23 , affecting protein biosynthesis. Indeed, we observed that translation arrest in yeast was notably increased after 24 hours of exposure to Mn (decreasing of β-galactosidase activity), suggesting that long or chronic exposure of Mn 2+ appears more effective in yeast than short or acute exposure which is similar to observations made by others 11,77 . We conducted the same analyses using other divalent cations such as Ca 2+ , Mg 2+ and Zn 2+ at equivalent or higher concentrations than used with Mn to determine if this was a general effect of metal stress and did not observe decreased rates of expression like the dose-dependent response to Mn seen in Fig. 3C. In addition, our ribosome profile analysis revealed a reduction in heavy polysomes fractions in response to Mn suggesting that Mn reduces efficiency of translation (Fig. 3B). This may be in agreement with a recent study in human SH-SY5Y cells that identified Mn-induced ER stress associated with increased phosphorylation of translation initiation factor eIF2α 59 . Similar profiles have been reported when using anti-translation drugs including pactamycin and harringtonine 78 .
Furthermore, we verified direct impairment of protein biosynthesis, including disruption of ribosome biogenesis due to down-regulation of the genes NOP1, NSR1; although this can be partially composed through up-regulation of the gene RPS15. A review at the CTD 48 revealed that RPS15 is a marker of Disease Progression, including memory impairment in transgenic mice modelingAD treated with copper 79 as well as RPL14, potentially affected by Mn, is a marker of PD), which have been observed in case of residential exposure to maneb,. It is very interesting because, while paraquat is a derived of bipyridine, maneb is a polymeric complex of Mn. Unpublished studies from our group have identified maneb-induced impairment of protein biosynthesis in cerebellar granule neurons.

Final Considerations
Literature in this field has consolidated robust hypotheses regarding Mn induced-neurotoxicity and neurodegeneration that include mitochondrial dysfunction, energy impairment, oxidative stress, disruption of neurotransmitters, ER stress, neuroinflammation, DNA damage and epigenetic alterations, apoptosis, autophagy, and many others 9,10,80,81 . However, occasionally, the accuracy of these hypotheses is challenged in different models. All processes cited above either occur after the process of protein synthesis is completed and/or are directly linked to it. In this study we identified protein synthesis as a key target of Mn-induced toxicity in S. cerevisiae. Defects in protein synthesis have been documented in different neurodegenerative disorders in humans. Since Mn-induced toxicity has been linked to human neurodegenerative disorders, the data presented in the current study may provide a connection between Mn-induced toxicity and human neurodegenerative disorders through the process of protein synthesis. Altogether, our findings provide strong evidence that Mn-toxicity can occur at multiple levels simultaneously, which appear to be associated with disruption of orchestrated essential pathways including metabolism and protein biosynthesis. In this way the presented study adds to our current understanding of the Mn-induced mode of toxicity. Additional experiments with mammalian models must be conducted to validate if these findings apply to other systems. experimental Gene expression analysis. Manganese sensitivity/resistance screening using yeast gene deletion array.
Approximately 4700 MATa haploid yeast, S. cerevisiae strains (BY4741, MATa ura3Δ0 leu2Δ0 his3Δ1 met15Δ0) from the non-essential Gene Deletion Array (yGDA) were manually arrayed onto agar plates as previously described by Alamgir,et al. 82 in the presence or absence of sub-inhibitory concentration (a high concentration of a bioactive/toxic compound where growth of a WT strain is not completely inhibited) of MnCl 2 (1.35 mM). Plates were incubated at 30 °C overnight. Finally, digital images of plates were used to analyze the growth of individual colonies, by automatized visual density comparison between control and their respective Mn treatment through of the available in-house and online software "SGAtools" from the University of Toronto 83 . The experiment was repeated between three and five times (Supplementary Material, Table SM 1). Colonies that showed 30% reduction or more in at least three repeats were considered hits [84][85][86] .
Transcriptomics experiment using q-PCR. The quantitative PCR (q-PCR) assay has been used to study the effects of a gene deletion on expression under specified conditionsgene deletion in previous works 87,88 . Primers of selected genes were synthetized (Table 2). Total RNA was isolated from each strain, using a Qiagen RNA isolation kit; followed by cDNA construction, using an iScript cDNA synthesis kit and finally used SYBR green supermix (Bio-Rad) for the multiplex real time PCR assay; according to the instructions of the manufacturer. The quantification of mRNA was performed by q-PCR on a Rotor-Gene RG-300 from Corbett research, according to Samanfar et al. 85 .
protein synthesis analysis. Total rRNA analysis. The yeast wild type strains were preincubated and grown overnight, then grown on YPD media at 30 °C to an OD 600 of 0.8-1.0, in either the absence or presence of Mn (1.5 mM or 3 mM), for 0.75, 1 and 24 hours respectively. Total RNA was isolated from each strain using a Qiagen RNA isolation kit (RNeasy mini kit). RNA electrophoresis was carried out in 1X MOPS running buffer diluted from 10X MOPS buffer [40.8 g 3-(N-morpholino) propanesulfonic acid (MOPS); 6.8 g sodium acetate; 3.8 g ethylenediamine tetraacetic acid (EDTA)]. Volume was completed up to 1000 ml by the addition of ultrapure water treated with diethyl pyrocarbonate (DEPC), and the pH was adjusted to 7.0 using sodium hydroxide (NaOH)). The concentration of agarose used in the RNA gels was 1.2% (w/v). Samples were prepared in 80% v/v deionized formamide, heated at 65 °C for 5 minutes, then immediately cooled on ice. Before loading the samples on the gel, 1/10th of sample volume 10X RNA loading dye (0.0125 g Bromophenol Blue; 10 µl 0.5 M EDTA; 2.5 ml www.nature.com/scientificreports www.nature.com/scientificreports/ www.nature.com/scientificreports www.nature.com/scientificreports/ 100% glycerol; 2.5 ml DEPC-treated water; mixed by vortexing and autoclaved) was added to the samples for a final concentration of 1 × (1 µg).
β-Galactosidase expression assay. The efficiency of translation was quantified using an inducible β-galactosidase reporter gene in the p416 plasmid 82,89 . β-galactosidase is a model of intracellular protein synthesis that can provide a profile of aberrancy in the rate of protein synthesis and an estimate of gene expression. Thus, yeast cells transformed with p416 plasmid were preincubated for 1 hour and then exposed to a crescent toxicological curve of Mn (0.25-3 mM) and other divalent ions (Ca 2+ , Mg 2+ and Zn 2+ at 0.5, 1.5, 5 mM and a higher 15 mM concentration for Mg 2+ ) for 3 hours at 30 °C, followed of spectrophotometric determination of β-galactosidase activity 82 . Metal ion concentrations were suggested by previous works [90][91][92][93] .
Ribosome profile analysis. Ribosome profiling 94 , allows for the monitoring of translation dynamics in vivo. Yeast wild type strains were preincubated 1 hr and then grown on YPD media at 30 °C to an OD 600 of 0.8-1.0, in the absence or presence of 3 mM Mn, for 1 hr and 24 hrs respectively at 30 °C. Immediately before harvest, cycloheximide was added to all samples, to a final concentration of 100 μg/ml, and the culture was incubated again at 30 °C for 15 minutes, followed by a cold snap in an ice water bath. Cells were harvested, washed with a cycloheximide/ water solution (100 μg/ml) and centrifuged at 4000 rpm for 4 min at 4 °C using a Sorvall SLA-1500 rotor to separate the supernatant. Cell pellets were resuspended in 10 ml of ice-cold lysis buffer A (YA buffer: 10 mM Tris-HCl [pH 7.4], 100 mM NaCl, 30 mM MgCl 2 , cycloheximide 50 μg/ml, heparin 200 μg/ml) and centrifuged at 4000 rpm for 4 min at 4 °C (Sorvall SS34 rotor) twice. Pellets were resuspended in 750 µl of YA buffer, lysed by vortexing with glass beads, transferred to microtubes, and centrifuged at 13000 rpm for 10 minutes at 4 °C. The supernatant was preserved for the quantitative determination of total RNA, followed by fractionation on 10-50% sucrose gradients containing 50 mM Tris-acetate [pH 7.0], 50 mM NH 4 Cl, 12 mM MgCl 2 , and 1 mM dithiothreitol. The extract was centrifuged for 2 h at 40,000 rpm using a SW40-Ti rotor in a Beckman LE-80 K at 4 °C. The polysome profiles were analyzed via a Biocomp gradient station and the absorbance was recorded at 254 nm using a spectrophotometer (Bio-Rad Econo UV monitor) coupled with the Biocomp station. In this method, free mRNAs from the top fractions were separated from polysome-associated mRNAs from the bottom fractions 95 .

protein-protein interaction (ppI) prediction and gene ontology (Go) analysis. A PPI network
can be described as a heterogeneous network of proteins joined by interactions as edges. Protein network and GO enrichment analysis were based on the data from the current project and analyzed using the STRING database (http://string-db.org) 38 . Additional GO analysis was conducted at the Comparative Toxicogenomic Database -CTD (http://ctdbase.org/) 48 to test the hypothesis of a conserved mode of action of Mn between yeast and humans. Both STRING and the CTD database were accessed on November 18 th , 2018. Data analysis. The results were expressed as mean ± sem of at least three independent experiments. To detect statistically significant differences, ANOVA (analysis of variance) followed by Bonferroni's tests was be used; preceded of single t-test analysis between pairs of treatments. Fitting and statistical analyses were performed using GraphPad Prism (GraphPad 4.0 Software Inc, San Diego, CA, USA).

Data Availability
All data generated and/or analyzed during this study are included in this published article and/or its Supplementary Material Files).  Table 2. Genes selected for q-PCR analysis. Primer sequences and associated parameters are included.