Evidence for a Conserved Function of Eukaryotic Pantothenate Kinases in the Regulation of Mitochondrial Homeostasis and Oxidative Stress

Human PANK1, PANK2, and PANK3 genes encode several pantothenate kinase isoforms that catalyze the phosphorylation of vitamin B5 (pantothenic acid) to phosphopantothenate, a critical step in the biosynthesis of the major cellular cofactor, Coenzyme A (CoA). Mutations in the PANK2 gene, which encodes the mitochondrial pantothenate kinase (PanK) isoform, have been linked to pantothenate-kinase associated neurodegeneration (PKAN), a debilitating and often fatal progressive neurodegeneration of children and young adults. While the biochemical properties of these enzymes have been well-characterized in vitro, their expression in a model organism such as yeast in order to probe their function under cellular conditions have never been achieved. Here we used three yeast mutants carrying missense mutations in the yeast PanK gene, CAB1, which are associated with defective growth at high temperature and iron, mitochondrial dysfunction, increased iron content, and oxidative stress, to assess the cellular function of human PANK genes and functional conservation of the CoA-controlled processes between humans and yeast. Overexpression of human PANK1 and PANK3 in these mutants restored normal cellular activity whereas complementation with PANK2 was partial and could only be achieved with an isoform, PanK2mtmΔ, lacking the mitochondrial transit peptide. These data, which demonstrate functional conservation of PanK activity between humans and yeast, set the stage for the use of yeast as a model system to investigate the impact of PKAN-associated mutations on the metabolic pathways altered in this disease.


Introduction
Coenzyme A (CoA) is an essential cofactor that plays a critical role in the operation of major cellular processes such as the citric acid cycle, lipid metabolism, redox regulation and ergosterol biosynthesis [1][2][3][4][5] (Figure 1). The synthesis of CoA initiates with the phosphorylation of pantothenic acid (vitamin B5) by a pantothenate kinase (PanK). In humans, this activity is catalyzed by PanK isoforms (1α, 1β, 2 and 3) encoded by three PANK genes, PANK1, PANK2 and PANK3 (Figure 1).
A fourth PANK gene, PANK4, encodes a pseudo-enzyme with no kinase activity but with an active phosphatase domain [6,7]. PanK4 has recently reported to be involved in cellular signaling as a substrate of the PI3K effector kinase AKT [8]. The expression, activity, and co-regulation of human PanKs vary between tissues and cellular compartments and the importance of these differences in disease manifestation remain poorly understood [9]. Pathway for CoA biosynthesis in humans and yeast and metabolic processes controlled by CoA. PanK enzymes catalyze the phosphorylation of pantothenic acid to form 4'phosphopantothenate, which serves as a precursor for the biosynthesis of CoA. In yeast, PanK activity is catalyzed by the Cab1 enzyme, which is encoded by the CAB1 gene. In humans, 4 PanK isoforms encoded by three PANK genes (PANK1, 2 and 3) catalyze pantothenate phosphorylation. Mutations in the PANK2 gene lead to major cellular and metabolic alterations, which ultimately result in the manifestations of the clinical symptoms of PKAN.
A fourth PANK gene, PANK4, encodes a pseudo-enzyme with no kinase activity but with an active phosphatase domain [6,7]. PanK4 has recently reported to be involved in cellular signaling as a substrate of the PI3K effector kinase AKT [8]. The expression, activity, and co-regulation of human PanKs vary between tissues and cellular compartments and the importance of these differences in disease manifestation remain poorly understood [9]. PanK1 and PanK3 have been localized to the cytoplasm and nucleus [10][11][12] and their transcript levels were found to be high in the liver, kidney, brain and small intestine in the case of PANK1, and the brain, heart, thymus and skeletal muscle in the case of PANK3 [13]. PanK2 is primarily found in the mitochondrial membranes with possible sub-localization to either the intermembranous space or cristae with high expression in neurons [10][11][12]14]. The importance of PanKs in human health was further highlighted by the discovery that mutations in the PANK2 gene are associated with Figure 1. Pathway for CoA biosynthesis in humans and yeast and metabolic processes controlled by CoA. PanK enzymes catalyze the phosphorylation of pantothenic acid to form 4'phosphopantothenate, which serves as a precursor for the biosynthesis of CoA. In yeast, PanK activity is catalyzed by the Cab1 enzyme, which is encoded by the CAB1 gene. In humans, 4 PanK isoforms encoded by three PANK genes (PANK1, 2 and 3) catalyze pantothenate phosphorylation. Mutations in the PANK2 gene lead to major cellular and metabolic alterations, which ultimately result in the manifestations of the clinical symptoms of PKAN.
Unlike humans, the yeast Saccharomyces cerevisiae harbors a single gene, CAB1, which encodes PanK activity ( Figure 1). The gene is essential for cell viability as its deletion results in cell death. Several cab1 mutations have been generated and their impacts on enzyme activity and cell growth have been characterized [23]. Substitution of Glycine 351 to Serine was the first reported mutation in CAB1 [1]. This mutation leads to a thermosensitive phenotype with yeast cells unable to grow at 37 • C [1]. Further characterization of this mutation showed altered growth on non-fermentable carbon sources, consistent with altered mitochondrial function and respiration [24], as well as increased susceptibility to ergosterol biosynthesis inhibitors [25]. Purified recombinant Cab1 G351S was found to have reduced pantothenate kinase activity (~10% that of the wild-type enzyme) [5]. Recently, a yeast model for PKAN, namely yeast strains (cab1∆/cab1 N290I and cab1∆/cab1 I291T ) lacking CAB1 and expressing cab1 alleles carrying pathological variants found in the human PANK2 gene, have been shown to recapitulate all the major cellular phenotypes found in patient cells. These phenotypes include mitochondrial dysfunction, altered lipid metabolism, iron overload, and oxidative damage [24]. N290I and I291T substitutions correspond to missense mutations N500I and I501T in the PANK2 gene found in some of the PKAN patients.
The ease of handling of yeast and the amenability of this model organism to genetic manipulation makes it an ideal system to probe the role of human PanK enzymes in the regulation of cellular metabolism, in general, and in mitochondrial homeostasis, in particular. Herein, we provide evidence for the functional conservation of PanK cellular activity between humans and yeast using yeast as a model system. We show that expression of human PanKs leads to either complete (with PanK1 and PanK3) or partial (with PanK2 lacking a mitochondrial transit peptide) complementation of the growth and cellular defects caused by the G351S, N290I and I291T mutations in the CAB1 gene. The functional complementation by human PanKs of yeast mitochondrial defects caused by altered Cab1 activity suggest that cytoplasmic expression of PanKs is sufficient for recovery of a normal mitochondrial function. The establishment of a yeast model of PKAN sets the stage for future investigations into the cellular, genomic, and metabolic complexity of the disease and could expedite the evaluation of new therapeutic strategies for the treatment of the disease in humans.

Functional Analysis of Human PANK1, PANK2, and PANK3 in Yeast
Previous studies have shown that expression of a full-length human PanK2, with an intact mitochondrial targeting motif, does not complement the loss of the yeast CAB1 gene [24].
To gain insights into the function of the active human PANKs, full length PANK1 and PANK3 as well as a truncated PANK2, lacking the mitochondrial targeting sequence (Pank2 mtm∆ ), were codon-optimized, using the yeast codon usage, and cloned into the pESC-URA3 expression vector for expression as protein fusions with an N-terminal c-Myc tag. In these constructs, PANK expression is under the regulatory control of the GAL1 inducible promoter, allowing transcription of the genes under galactose but not glucose conditions. The resulting plasmids were then introduced into yeast mutant strains lacking the chromosomal CAB1 gene but expressing either cab1 G351S or the hypomorphic mutants cab1 N290I , or cab1 I291T . The ability of human PanKs to restore the cellular and metabolic defects of these mutants was examined ( Figure 2). Consistent with published data, both G351S and N290I substitutions caused a thermosensitive phenotype [24] and reduced growth on galactose both at 28 • C and 37 • C (Figure 2A,B,D,E). On the other hand, I291T showed a mild thermosensitive phenotype at 37 • C on both glucose [24] and galactose and this phenotype was further exacerbated in the presence of galactose at 37 • C ( Figure 2C). In the cab1 G351S genetic background, all human isoforms rescued the growth phenotype of the mutant on galactose at 37 • C; however, the extent of the complementation varied between PanK isoforms with optimal complementation achieved with hPanK3. In the yeast strain harboring the cab1 N290I allele, which displays a more severe growth defect on galactose at 37 • C compared to the cab1 G351S mutant, expression of human PanK1 and PanK3 restored growth to levels similar to those seen in the cab1 G351S mutant background. However, expression of human PanK2 mtm∆ resulted in only partial complementation ( Figure 2B). Finally, in the mutant strain carrying the cab1 I291T variant, which displays only a mild growth defect on galactose at 37 • C, expression of human PanKs restored growth to levels comparable to those of the wild-type strain ( Figure 2C). Unlike the PanK2 mtm∆ isoform, expression of the full length human PANK2 (harboring the mitochondrial transit peptide) in the hypomorphic mutants G351S and N290I failed to complementation their growth defect at 37 • C (data not shown). . Yeast cells were inoculated in minimal medium without uracil and tryptophan but containing either 2% glucose or 2% galactose at a cell density of 10 4 cells/ml. Optical density measurements at 600 nm (OD600) were determined at the indicated time points. Data for liquid assays represent three biological replicates (n = 3) and the plotted graphs represent the average ± SEM.
Consistent with published data, both G351S and N290I substitutions caused a thermosensitive phenotype [24] and reduced growth on galactose both at 28 °C and 37 °C (Figure 2A,B,D,E). On the other hand, I291T showed a mild thermosensitive phenotype at 37 °C on both glucose [24] and galactose and this phenotype was further exacerbated in the presence of galactose at 37 °C ( Figure 2C). In the cab1 G351S genetic background, all human isoforms rescued the growth phenotype of the mutant on galactose at 37 °C; however, the extent of the complementation varied between PanK isoforms with optimal complementation achieved with hPanK3. In the yeast strain harboring the cab1 N290I allele, which displays a more severe growth defect on galactose at 37 °C compared to the cab1 G351S mutant, expression of human PanK1 and PanK3 restored growth to levels similar to those seen in the cab1 G351S mutant background. However, expression of human PanK2 mtmΔ resulted in only partial complementation ( Figure 2B). Finally, in the mutant strain carrying the cab1 I291T variant, which displays only a mild growth defect on galactose at 37°C, expression of human PanKs restored growth to levels comparable to those of the wild-type strain ( Figure 2C). Unlike the PanK2 mtmΔ isoform, expression of the full length human PANK2 (harboring the mitochondrial transit peptide) in the hypomorphic mutants G351S and N290I failed to complementation their growth defect at 37 °C (data not shown).
To further quantify the complementation level by human PanKs in these yeast mutants, the same assays were conducted in liquid media and cell growth was measured , and cab1∆/cab1 I291T (C) mutant strains carrying pFL38/CAB1 and pESC-URA3 vector containing the different hPANK gene were spotted onto SC medium supplemented with the indicated carbon sources and incubated at 28 • C and 37 • C. Yeast growth was monitored over time and images collected 2 or 3 days postinoculation. (D-F) Growth rates of cab1∆/cab1 G351S (D), cab1∆/cab1 N290I (E), and cab1∆/cab1 I291T (F) strains carrying the empty vector (EV), CAB1 and human PANKs in liquid media on glucose or galactose at 30 • C vs. 37 • C (from top to bottom). Yeast cells were inoculated in minimal medium without uracil and tryptophan but containing either 2% glucose or 2% galactose at a cell density of 10 4 cells/mL. Optical density measurements at 600 nm (OD 600 ) were determined at the indicated time points. Data for liquid assays represent three biological replicates (n = 3) and the plotted graphs represent the average ± SEM. To further quantify the complementation level by human PanKs in these yeast mutants, the same assays were conducted in liquid media and cell growth was measured over a 96 h time period as shown in Figure 2D-F. Overall complementation efficiency by the human PanKs measured in liquid-based media was similar to that measured in solid growth media ( Figure 3A-C).
Quantitative analysis of the complementation efficiency of human PanKs was determined by comparing the growth rate of mutant cells harboring the human genes to that of cells carrying the empty vector or the wild-type CAB1 gene under similar growth conditions as described in the Material and Methods. Expression of the human PanKs resulted in partial complementation of the growth defect of the cab1 G351S mutant at 37 • C on galactose, which was estimated to be~22% (PanK1),~36% (PanK2) and~51% (PanK3) that of the complementation achieved with the wild-type CAB1 ( Figure 3A). In the yeast strain harboring the cab1 N290I allele, the estimates were~47%,~8%, and~62 % for PanK1, PanK2, and PanK3, respectively, under the same conditions ( Figure 3B). Although the cab1 I291T mutation has only a mild growth defect at 37 • C under galactose conditions, expression of human PanKs restored growth of the mutant to~94% (PanK1),~88% (PanK2), and~94% (PanK3) that of the same strain carrying the wild-type CAB1 gene ( Figure 3C).  Figure 2D-F. Overall complementation efficiency by the human PanKs measured in liquid-based media was similar to that measured in solid growth media ( Figure 3A-C). Representation of the mid-log phase growth of yeast strains cab1Δ/cab1 G351S (A), cab1Δ/cab1 N290I (B), and cab1Δ/cab1 I291T (C) carrying EV, CAB1 and human PANKs from data shown in Figure 2. The bar graphs show the OD600 values of the strains expressing PANKs at the time point when the strains expressing the CAB1 wild-type allele were at the late log phase (36 h for glucose and 60 h for galactose conditions, respectively). (D-F) Oxygen consumption rate was measured in the same strains grown at 28 °C or 37 °C (especially for the strain expressing the cab1 I291T mutant allele) in SC medium without uracil and tryptophan and containing 2% galactose. OCR was normalized to the corresponding mutant strain transformed with the CAB1 wild-type allele. OCR was conducted in triplicates (n = 3) and the plotted graphs represent the average ± SEM. *: p < 0.05 and **: p < 0.01 relative to mutant strain transformed with CAB1 allele and #: p < 0.05 and ##: p < 0.01 relative to mutant strain transformed with the empty vector. Using ANOVA followed by Bonferroni's post hoc test.
Quantitative analysis of the complementation efficiency of human PanKs was determined by comparing the growth rate of mutant cells harboring the human genes to that of cells carrying the empty vector or the wild-type CAB1 gene under similar growth conditions as described in the Material and Methods. Expression of the human PanKs resulted in partial complementation of the growth defect of the cab1 G351S mutant at 37 °C on galactose, which was estimated to be ~22% (PanK1), ~36% (PanK2) and ~51% (PanK3) that of the complementation achieved with the wild-type CAB1 ( Figure 3A). In the yeast strain harboring the cab1 N290I allele, the estimates were ~47%, ~8%, and ~62 % for PanK1, PanK2, and PanK3, respectively, under the same conditions ( Figure 3B). Although the cab1 I291T mutation has only a mild growth defect at 37 °C under galactose conditions, expression of human PanKs restored growth of the mutant to ~94% (PanK1), ~88% (PanK2), and ~94% (PanK3) that of the same strain carrying the wild-type CAB1 gene ( Figure 3C).
The ability of human PANKs to complement a null cab1Δ mutant was also determined by examining the loss of the CAB1-expressing plasmid in the yeast strain carrying a deletion of the chromosomal CAB1 gene (data not shown). Although a few clones were de- Representation of the mid-log phase growth of yeast strains cab1∆/cab1 G351S (A), cab1∆/cab1 N290I (B), and cab1∆/cab1 I291T (C) carrying EV, CAB1 and human PANKs from data shown in Figure 2. The bar graphs show the OD 600 values of the strains expressing PANKs at the time point when the strains expressing the CAB1 wild-type allele were at the late log phase (36 h for glucose and 60 h for galactose conditions, respectively). (D-F) Oxygen consumption rate was measured in the same strains grown at 28 • C or 37 • C (especially for the strain expressing the cab1 I291T mutant allele) in SC medium without uracil and tryptophan and containing 2% galactose. OCR was normalized to the corresponding mutant strain transformed with the CAB1 wild-type allele. OCR was conducted in triplicates (n = 3) and the plotted graphs represent the average ± SEM. *: p < 0.05 and **: p < 0.01 relative to mutant strain transformed with CAB1 allele and #: p < 0.05 and ##: p < 0.01 relative to mutant strain transformed with the empty vector. Using ANOVA followed by Bonferroni's post hoc test.
The ability of human PANKs to complement a null cab1∆ mutant was also determined by examining the loss of the CAB1-expressing plasmid in the yeast strain carrying a deletion of the chromosomal CAB1 gene (data not shown). Although a few clones were detected, the rate of plasmid loss was found to be very low, suggesting lack of complementation of the null mutant. It is possible that the few clones detected were selected due to suppressor mutations.

Complementation Analysis of Respiratory Activity Deficiency of cab1 Mutant Strains
Previous studies have demonstrated that strains expressing the hypomorphic alleles cab1 G351S , cab1 N290I, and cab1 I291T exhibit mitochondrial dysfunction [24]. To assess whether hPANKs could rescue this phenotype, the oxygen consumption rate (OCR) of the mutants carrying an empty vector, yeast CAB1 or the human PANKs was measured on galactosebased growth medium. As expected, optimal respiration rates were observed for strains carrying the wild-type CAB1 allele whereas the mutants carrying the empty vector all showed reduced respiration rates (~60% that of the wild-type) ( Figure 3D-F). In each of the three yeast backgrounds, expression of human PANK1 and PANK3 restored the respiration efficiency in the mutants to levels similar to those achieved with the wildtype allele ( Figure 3D-F). Expression of the PANK2 mtm∆ gene, however, resulted in partial restoration of the respiration rates with measured rates estimated at~80% those of the mutants carrying the wild-type CAB1 gene.

Human PanKs Complement Iron Susceptibility of Yeast cab1 Mutants
Iron dyshomeostasis is an important feature of PKAN disorder [11,18,19,24]. In yeast, iron accumulation results in increased susceptibility to FeSO 4 [26]. Interestingly, the strain harboring the cab1 G351S or cab1 N290I mutant alleles showed greater sensitivity to this ferrous salt as well as iron accumulation [24]. To determine whether expression of hPanKs could also rescue the iron-sensitivity phenotype of the cab1 mutants, growth assays were conducted in both solid and liquid galactose-based media lacking or supplemented with 7 mM FeSO 4 . As shown in Figure 4, the severe growth defect of all three yeast mutants (cab1 G351S , cab1 N290I , and cab1 I291T ) on FeSO 4 was complemented by the expression of hPanKs under both permissive and restrictive temperatures. Expression of hPanK3 or hPanK1 resulted in complete rescue of the growth defect whereas that of hPANK2 mtm∆ resulted in partial complementation of this phenotype (Figures 4 and S1). cab1 G351S , cab1 N290I, and cab1 I291T exhibit mitochondrial dysfunction [24]. To assess whether hPANKs could rescue this phenotype, the oxygen consumption rate (OCR) of the mutants carrying an empty vector, yeast CAB1 or the human PANKs was measured on galactosebased growth medium. As expected, optimal respiration rates were observed for strains carrying the wild-type CAB1 allele whereas the mutants carrying the empty vector all showed reduced respiration rates (~60% that of the wild-type) ( Figure 3D-F). In each of the three yeast backgrounds, expression of human PANK1 and PANK3 restored the respiration efficiency in the mutants to levels similar to those achieved with the wild-type allele ( Figure 3D-F). Expression of the PANK2 mtmΔ gene, however, resulted in partial restoration of the respiration rates with measured rates estimated at ~80% those of the mutants carrying the wild-type CAB1 gene.

Human PanKs Complement Iron Susceptibility of Yeast cab1 Mutants
Iron dyshomeostasis is an important feature of PKAN disorder [11,18,19,24]. In yeast, iron accumulation results in increased susceptibility to FeSO4 [26]. Interestingly, the strain harboring the cab1 G351S or cab1 N290I mutant alleles showed greater sensitivity to this ferrous salt as well as iron accumulation [24]. To determine whether expression of hPanKs could also rescue the iron-sensitivity phenotype of the cab1 mutants, growth assays were conducted in both solid and liquid galactose-based media lacking or supplemented with 7 mM FeSO4. As shown in Figure 4, the severe growth defect of all three yeast mutants (cab1 G351S , cab1 N290I , and cab1 I291T ) on FeSO4 was complemented by the expression of hPanKs under both permissive and restrictive temperatures. Expression of hPanK3 or hPanK1 resulted in complete rescue of the growth defect whereas that of hPANK2 mtmΔ resulted in partial complementation of this phenotype (Figure 4 and Figure S1). (A-C) The iron susceptibility of cab1∆/cab1 G351S (A), cab1∆/cab1 N290I (B), and cab1∆/cab1 I291T (C) yeast strains carrying an empty vector or either yeast CAB1 or human PANK genes was determined using plate-based assays. Cells were spotted at the indicated cell densities onto SC medium without uracil and tryptophan with 2% galactose in the absence or presence of 7 mM FeSO 4 . Cell growth was measured after 3 days of incubation at 28 • C and 37 • C. (D-F) Depit results of assays similar to those described in (A-C) but conducted in liquid culture with growth measured at 60 h post-inoculation. Liquid growth assays were conducted in triplicates (n = 3) and the plotted graphs represent the average ± SEM.

Complementation of the Oxidative Damage of CAB1 Defective Yeast Mutants by Human PanKs
It is well-described that iron accumulation is associated with an altered cellular oxidative state [27][28][29]. Accordingly, and as previously reported, the cab1 G351S and cab1 N290I, mutant strains exhibited a 3.5-fold increase in reactive oxygen species (ROS) compared to the parental strain. The mutants also showed an increase of malondialdehyde (MDA) content, suggesting lipid peroxidation and impairment of membrane function [24].Thus, ROS production was measured in all yeast mutant strains carrying human PanKs or the empty vector by cytofluorometric analyses using the fluorescent ROS indicator dihydrorhodamine 123 (DHR123) (Figure 5A-F).
graphs represent the average ± SEM.

Complementation of the Oxidative Damage of CAB1 Defective Yeast Mutants by Human PanKs
It is well-described that iron accumulation is associated with an altered cellular oxidative state [27][28][29]. Accordingly, and as previously reported, the cab1 G351S and cab1 N290I, mutant strains exhibited a 3.5-fold increase in reactive oxygen species (ROS) compared to the parental strain. The mutants also showed an increase of malondialdehyde (MDA) content, suggesting lipid peroxidation and impairment of membrane function [24].Thus, ROS production was measured in all yeast mutant strains carrying human PanKs or the empty vector by cytofluorometric analyses using the fluorescent ROS indicator dihydrorhodamine 123 (DHR123) (Figure 5A-F). Figure 5. Analysis of reactive oxygen species (ROS) yeast PanK-deficient mutants expressing human PanK isoforms. ROS measurements were performed for cab1Δ/cab1 G351S (A), cab1Δ/cab1 N290I (C), and cab1Δ/cab1 I291T (E) yeast strains transformed with pFL38/CAB1, EV, and pESC-URA3 vector containing different human PanK isoforms. ROS production was determined as the percentage of fluorescent cells using the DHR123 probe. Cytofluorometry analysis of the cab1Δ/cab1 G351S (B), cab1Δ/cab1 N290I (D), and cab1Δ/cab1 I291T (F) yeast strains transformed with either pFL38/CAB1, empty Figure 5. Analysis of reactive oxygen species (ROS) yeast PanK-deficient mutants expressing human PanK isoforms. ROS measurements were performed for cab1∆/cab1 G351S (A), cab1∆/cab1 N290I (C), and cab1∆/cab1 I291T (E) yeast strains transformed with pFL38/CAB1, EV, and pESC-URA3 vector containing different human PanK isoforms. ROS production was determined as the percentage of fluorescent cells using the DHR123 probe. Cytofluorometry analysis of the cab1∆/cab1 G351S (B), cab1∆/cab1 N290I (D), and cab1∆/cab1 I291T (F) yeast strains transformed with either pFL38/CAB1, empty vector or pESC-URA3 vector containing different human PanK isoforms following staining with DHR123 probe. Each curve represents the distribution of the measured events (counts) according to their fluorescence intensity expressed in a log unit. The grey spectrum represents background autofluorescence. The strains were grown in YP medium supplemented with 2% galactose at 28 • C. This assay was conducted in triplicate (n = 3) and the plotted graphs represent the average ± SEM. *: p < 0.05 and **: p < 0.01 relative to mutant strain transformed with CAB1 allele and #: p < 0.05 and ##: p < 0.01 relative to mutant strain transformed with the empty vector.
All yeast mutant strains carrying the empty vector showed high fluorescence signals, indicative of intense oxidative stress associated with the loss of PanK activity ( Figure 5A-F). Conversely, cells expressing the wild type CAB1 gene showed low levels of ROS indiating low oxidative stress levels. Interestingly, ROS content was significantly reduced by the expression of each of the hPanKs in all mutant genetic backgrounds ( Figure 5A-F) with expression of human PANK3 or PANK1 in the mutants causing significantly greater reduction in oxidative stress compared to PANK2 mtm∆ .

Correlation between Complementation Ability and Human PanK Expression and Activity
To assess whether the degree by which human PanKs complement the growth defect, respiration rate and oxidative stress of yeast mutants altered in PanK activity could be linked to the level of expression of each enzyme, immunoblot analyses were conducted. Since human PanKs were expressed with a C-terminal Myc Tag, their expression in the cab1 G351S , cab1 N290I, and cab1 I291T mutants following growth on galactose-based medium was monitored using an anti-Myc monoclonal antibody ( Figure 6A,B).
indicative of intense oxidative stress associated with the loss of PanK activity ( Figure 5A-F). Conversely, cells expressing the wild type CAB1 gene showed low levels of ROS indiating low oxidative stress levels. Interestingly, ROS content was significantly reduced by the expression of each of the hPanKs in all mutant genetic backgrounds ( Figure 5A-F) with expression of human PANK3 or PANK1 in the mutants causing significantly greater reduction in oxidative stress compared to PANK2 mtmΔ .

Correlation between Complementation Ability and Human PanK Expression and Activity
To assess whether the degree by which human PanKs complement the growth defect, respiration rate and oxidative stress of yeast mutants altered in PanK activity could be linked to the level of expression of each enzyme, immunoblot analyses were conducted. Since human PanKs were expressed with a C-terminal Myc Tag, their expression in the cab1 G351S , cab1 N290I, and cab1 I291T mutants following growth on galactose-based medium was monitored using an anti-Myc monoclonal antibody ( Figure 6A,B). Whereas PanK1 and PanK3 proteins were strongly induced in these strains, the levels of PanK2 mtmΔ were very low in the three genetic backgrounds. As a control, the expression of the endogenous mitochondrial outer membrane protein porin (Por1) was also exam- Whereas PanK1 and PanK3 proteins were strongly induced in these strains, the levels of PanK2 mtm∆ were very low in the three genetic backgrounds. As a control, the expression of the endogenous mitochondrial outer membrane protein porin (Por1) was also examined using antiPor1 monoclonal antibody and similar levels of this protein could be detected in all strains ( Figure 6A,B). Consistent with these data, endogenous PanK activity in PanK-defective cells measured using D-[1-14 C] pantothenic acid also showed correlation between levels of expression and enzyme activity. As expected, cab1 G351S cells harboring an empty vector exhibited low PanK-specific activity [5,25], whereas cab1 G351S cells expressing PanK1 or PanK3 showed comparatively higher PanK activity with PanK1 leading to 19.4-fold increase in the endogenous PanK activity of the mutant cells and PanK3 tõ 104-fold increase in the endogenous PanK activity ( Figure 6C,D). PanK2 mtm∆ expression in cab1 G351S cells resulted in only a modest (~1.5-fold) increase in the endogenous PanK activity. Together, these findings suggest that differences in expression levels and activity between human PanKs in yeast likely account for their ability to achieve complete or partial complementation of the cellular defects caused by the specific CAB1 mutations.

Discussion
In this study, we provided the first evidence for the functional conservation of PanK function between humans and yeast. We found that growth and cellular and mitochondrial defects of yeast strains carrying cab1 G351S , cab1 N290I and cab1 I291T mutated alleles of the pantothenate kinase gene CAB1, but not the null cab1∆ mutant, can be complemented by expression of active human PanK-encoding genes PANK1, PANK2 and PANK3. In the case of human PanK2, initial attempts to express the full-length protein with its mitochondrialtargeting motif (MTM) failed to complement these defects, whereas a PanK2 isoform lacking this motif was functional in yeast although its expression could not reach levels sufficient to achieve full complementation. These studies are of particular importance as expression and regulation of PANK2 are important factors in PKAN and the establishment of a reliable model to investigate the function of human PanKs would make it possible to express individual alleles found in patients and assess their cellular and metabolic impacts using yeast as a model system.
To date, the molecular mechanism underlying PKAN manifestation remains poorly understood. A simplistic model suggests that the loss of PanK2 activity leads to reduced levels of CoA, which in turn impacts a series of metabolic events that rely on this cofactor to achieve the correct cellular functions [30,31]. The validity of this model has, however, been challenged because CoA levels do not seem to be altered, if at all, in cells lacking PanK2 in vitro or mice [15,32,33]. Other molecular mechanisms have also been proposed including altered mitochondrial ACP, which controls the biosynthesis of fatty acids and protein lipoylation [15,33]. A better understanding of how loss of PanK2 causes PKAN may help in the development of optimal therapeutic strategies for this disease. Therefore, establishment model organisms, especially those that are amenable to high throughput genetic and cell biological analyses, could play a central role in advancing this knowledge. Mice, zebrafish (Danio rerio) and fruit flies (Drosophila melanogaster) have been used as models of PKAN and have helped answer important questions about the role of pantothenate phosphorylation in the onset of various physiological defects reminiscent of PKAN. Similarly, yeast, with its plethora of genetic and cell biological tools, represents an ideal system to probe the molecular and cellular mechanisms that lead to PKAN or PKAN-like phenotypes. In fact, the same yeast mutants used in this study have previously been shown to recapitulate several of the metabolic defects seen in PKAN [24].
Our data showed that PanK3 and PanK1 could be expressed at high levels in yeast and achieved the highest degree of complementation of all the phenotypes exhibited by the three yeast mutants defective in PanK activity, while a partial complementation was obtained by human PanK2 lacking a mitochondrial transit peptide.
Immunoblot analyses also showed that the expression level of PanK2 is dramatically lower than that of PanK1 and PanK3, which may account for the differences in complementation of the cellular defects caused in yeast by different Cab1 mutations. It remains unclear why the expression of PanK2 is restricted by yeast cells. One possibility is that overexpression of PanK2 is detrimental to yeast cells through the modification of other unknown metabolites. This raises the possibility that this enzyme might play a role in other yet to be discovered metabolic pathways. This hypothesis could also explain the complex molecular and metabolic profile of PKAN.
Unlike in Drosophila, where the localization of PanK to the mitochondria was crucial for optimal rescue of the PanK fbl mutant phenotype [22], the yeast Cab1 is a cytoplasmic enzyme [24,34] and optimal rescue is likely to be achieved with enzymes that are also expressed in the yeast cytoplasm. This could also explain the previously reported lack of complementation of yeast cab1 null mutant and of hypomorphic mutants with the fulllength PanK2, which carries the native mitochondrial targeting motif [24]. Research aimed to determine the localization of human PanKs in yeast and investigate the mechanism of complementation is warranted. Future molecular designs to express different forms of PanK2 as well as metabolomic analyses in yeast are warranted and may shed further light on the impact of PanK2 expression on yeast cellular metabolism.
Current therapeutic approaches for PKAN include activation of active cytosolic PanKs (PanK3 in particular) to compensate for the altered activity of the mutated hPanK2 [35], metabolite supplementation using intermediates in the CoA pathway such as 4'-phosphopa ntetheine [15] as well as the utilization of iron chelators such as deferiprone [36]. Moreover, extracellular CoA supplementation has been demonstrated to be an effective bypass treatment in cultured neuronal cells derived from PKAN patient fibroblasts [37]. Because none of these approaches has shown efficacy in PKAN patients, there is an urgent need to unravel the cellular and metabolic mechanism of the disease for the design of improved therapies. Yeast is thus an attractive model to study PKAN and may contribute to the discovery of an effective treatment for this disorder.
In summary, this study demonstrates complementation of pantothenate kinase deficiency in yeast by human PanK isoforms. These findings, which highlight the functional conservation in pantothenate phosphorylation across species, could set the stage for future biochemical and cell biological analyses in yeast to probe the effect of individual PanK2 mutations on cellular metabolism and mitochondrial activity. Strains carrying these mutations might serve as tools to evaluate new therapeutic strategies that reverse the phenotypes resulting from PanK loss of function.

Yeast Complementation and Iron Sensitivity Spotting Assays
Spot assay analysis was performed by spotting serial dilutions of the different strains on SC-U-W agar plates, supplemented with 2% galactose (inducing condition) or 2% glucose. When mentioned, the agar plates were supplemented with 7 mM FeSO 4 . Plates were incubated at 28 • C or 37 • C and captures were taken every 24 h.

Oxygen Consumption Rate (OCR) and Reactive Oxygen Species (ROS) Measurements
The oxygen consumption rate (OCR) was measured at 30 • C in whole cells cultured for 18 h at 28 • C or for 16 h at 37 • C in SC-U-W medium supplemented with 2% galactose using a Clark-type oxygen electrode (Oxygraph System, Hansatech Instruments, King's Lynn, UK) with 1 ml of air-saturated respiration buffer (0.1 M phthalate-KOH, pH 5.0) and 0.5% glucose. The reaction started by the addition of 20 mg of wet-weight cells as described [39]. Values for OCR were normalized to the dry weight of the cells.
ROS content was determined by flow cytometric analysis, using dihydrorhodamine 123 (DHR123; Sigma-Aldrich®, Darmstadt, Germany) that inside the cell can be oxidized by ROS (mainly H 2 O 2 ), producing green, fluorescent R123 (excitation/emission spectra of 488/530 nm) detected by the fluorescence channel (FL-1) with a 530/30 nm band pass filter. For flow cytometry analysis, cells were pre-grown at 28 • C in SC liquid medium supplemented with 2% glucose and then inoculated in YP supplemented with 2% galactose. After 18 h of growth at 28 • C or 16 h at 37 • C, samples were loaded with 1.25 µg/mL of DHR123 for 2 h at 28 • C. At the end of the incubation time, cells were harvested (30 s at 14,000 rpm) and re-suspended in H 2 O. The fluorescence signal was quantified by a NovoCyte flow cytometer (NovoCyte ® , ACEA Biosciences, Inc., San Diego, CA, USA). An unstained sample, without DHR123, was prepared as a negative control for each strain to set the threshold index, thus providing an auto-fluorescence background against which to measure positive fluorescence. For each sample, 10,000 cellular events were analyzed. Data achieved from flow cytometer were analyzed using NovoExpress software (NovoExpress ® , ACEA Biosciences, Inc., San Diego, CA, USA). ROS generation was measured as the percentage of fluorescent cells (PFC) corresponding to cells that produced ROS-level increments of at least one log unit [40,41]. Values were normalized to the mutant strains harboring the CAB1 wild-type allele.

Monitoring of hPanK Steady-State Expression in Yeast
To determine the steady-state level of the hPanKs isoforms, all the mutant strains ∆cab1 expressing PANK1, PANK2 or PANK3 were pregrown in SC glucose 2% medium and then inoculated in SC medium supplemented with 2% galactose until 2-3 OD 600 was reached at 28 • C or 37 • C depending on the mutated strain analyzed. Then, 10 OD 600 cells for each strain were collected and total protein extraction was performed using the TCA method by chilling the cells in the presence of 20 mM NaOH, 0.5% β-mercaptoethanol, 650 µM PMSF and 25% TCA on ice. The protein extracts were then resuspended in 150 µl of Laemmli Sample Buffer at a pH of 6.8. An equal amount of protein suspension, corresponding to 1 OD 600 of the original cell suspension was loaded on 12% SDS-PAGE and electroblotted on a nitrocellulose filter. The filter was incubated with mouse α-Myc (1:2000) and mouse α-Por1 (1:10,000) primary antibodies. Secondary antibody anti-mouse StarBright 700 (BioRad, Hercules, CA, USA, 1:10,000 dilution) was used to detect both primary antibodies. Fluorescent signals were revealed using Bio-Rad ChemiDoc Imagers and analyzed with Image Lab Software (Bio-Rad). The ratios between Myc and Por1 were calculated.

Liquid Yeast Growth Assays
For yeast growth assays, precultures of S. cerevisiae cells were grown overnight at 30 • C in SC-URA-TRP media (2% glucose). Cells were diluted to OD of 0.2 in the fresh media, grown 4 h to mid-log phase, and then harvested, washed, and adjusted in sterile water to OD 600nm = 1. Yeast cells were inoculated in polypropylene tubes at a final volume of 3.5 mL. The initial cell density was 10 4 cells/mL in SC-URA-TRP media (2% glucose or galactose). A total of 6 mM FeSO 4 was added to the media when needed. Cells were incubated at 30 • C and 37 • C and optical density (OD 600 ) measurements were taken using a spectrophotometer.

PanK Enzymatic Activity in Yeasts
Pantothenate kinase activity assay was performed as previously described [1]. Briefly, the assay tracked the ATP-dependent conversion of D-[1-14 C] pantothenate to phosphopantothenate and binding to the DEAE cellulose ion exchange filter. Cell-free extracts from yeast expressing human PanK proteins were obtained by homogenization using glass bead beating, followed by centrifugation at 1000 rpm for 5 min. The 40 µL enzyme reaction contained reaction buffer (100 mM Tris HCl, 2.5 mM MgCl 2 , 2.5 mM ATP, pH 7.4), D-[1-14 C] pantothenate (2 nmol, 0.1 µCi), and 72 µg cell-free extracts. Protein content of the cell-free extracts was determined using the Bradford assay. The reaction was done at 30 • C for 10 min following the addition of 4 µL of 10% acetic acid to stop the reaction. The reaction mixture was spotted on a DE-81 filter (0.6mm in diameter) placed within a spin column with a 2 mL collection tube. Following 5 min incubation, the spotted filters were centrifuged for 20 s at 1000 rpm, washed twice with 1% acetic acid in ethanol, and collected for liquid scintillation spectrometry.