Childhood-onset movement disorders

The metabolic cofactor Coenzyme A (CoA) gained renewed attention because of its role in neurodegeneration, protein acetylation, autophagy and signal transduction. The longstanding dogma is that eukaryotic cells obtain CoA exclusively via the uptake of extracellular precursors, especially vitamin B5, which is intracellularly converted through five conserved enzymatic reactions into CoA. We demonstrate that cells and organisms possess an alternative mechanism to influence intracellular CoA levels with the use of exogenous CoA. CoA is hydrolyzed extracellularly by ecto-nucleotide-pyrophosphatases to 4’-phosphopantetheine, a biologically stable molecule, able to translocate through membranes via passive diffusion. Inside the cell, 4’-phosphopantetheine is enzymatically converted back to CoA by the bifunctional enzyme CoA synthase. Phenotypes induced by intracellular CoA deprivation are reversed when exogenous CoA is provided. Our findings answer long-standing questions in fundamental cell biology and have major implications for understanding CoA-related diseases and therapies.


INTRODUCTION
Coenzyme A (CoA) was identified more than 60 years ago 1 and as a carrier of acyl groups, CoA is essential for over 100 metabolic reactions. It is estimated that CoA is an obligatory cofactor for 9% of known enzymatic reactions 2 . CoA and acetyl-CoA influence protein acetylation levels in various model organisms [3][4][5] . Protein acetylation is an essential posttranslational modification, catalyzed by acetyltransferases that use acetyl-CoA as the source 6 . Acetyl-CoA levels also affect autophagy 7, 8 , and CoA promotes oocyte survival in Xenopus laevis by binding to and activating calcium/calmodulin-dependent protein kinase II (CaMKII) 9 . Taken together, intracellular concentrations of acetyl-CoA and CoA are critical to a broad range of cellular processes 10 .
Current thinking about how cells and organisms obtain this indispensable molecule originates from experiments performed in the 1950's 2,11 , which demonstrate how a specific sequential order of enzymatic activities result in the formation of CoA in vitro when Vitamin B5 was used as a substrate. These enzymes are, in order, pantothenate kinase (PANK); phosphopantothenoylcysteine synthetase (PPCS); phosphopantothenoylcysteine decarboxylase (PPCDC); phosphopantetheine adenylyltransferase (PPAT) and dephosphoCoA kinase (DPCK) (Figure 1a). Later, genes encoding these enzymes were identified in a wide range of organisms 2, 12-14 and references therein . In some organisms, including Drosophila melanogaster, mice and humans, PPAT and DPCK enzyme activities are combined into a single bifunctional protein, referred to as CoA synthase or COASY 12,13,15 . In vitro experiments show that in addition to Vitamin B5, pantetheine can also be phosphorylated by pantothenate kinase activity, and the formed product, 4'-phosphopantetheine, can serve as a precursor for CoA 16 . However, direct evidence that cells take up intact pantetheine and utilize it for CoA biosynthesis is still lacking.
In addition to renewed interest in the CoA molecule and its cellular roles, the biosynthetic route gained attention because of its connection with specific forms of neurodegeneration. Two enzymes in the CoA de novo biosynthetic route, PANK (first step) and COASY (combined last 2 steps) are associated with a neurodegenerative disease classified as NBIA (Neurodegeneration with Brain Iron Accumulation) 17,18 . Mutations in the gene encoding PANK2 (one of four human PANK genes) cause an NBIA disorder, called pantothenate kinase-associated neurodegeneration (PKAN) 18 . Patients experience progressive dystonia and accumulate iron in specific brain regions. Recently, patients with mutations in the gene encoding COASY were identified and they have similar clinical features and brain iron accumulation. This new NBIA disorder is referred to as CoPAN, for COASY protein-associated neurodegeneration 17 . This strongly suggests that impairment of the classic CoA biosynthetic route underlies progressive neurodegeneration in these patient groups. Currently there is no treatment available to halt or reverse the neurodegeneration in these CoA-related disorders.
CoA levels are decreased in a Drosophila model for PKAN and the neurodegenerative phenotypes and decreased CoA levels are rescued by addition of pantethine to the food 19 . Pantethine addition also rescues a ketogenic diet-induced neurodegenerative phenotype in PANK2 -/knock out mice 20 . These studies demonstrate that in a pantothenate kinase impaired background, CoA precursors other than vitamin B5 can alleviate neurodegenerative symptoms. How pantethine exerts its rescuing function (especially in the mouse study) is unclear because pantethine is highly unstable in serum and rapidly converted into Vitamin B5 and cysteamine by pantetheinases 20,21 .
The aim of this study was to determine whether alternate routes exist for cells and organisms to obtain CoA. We found that extracellular CoA levels influence intracellular CoA levels both in vitro and in vivo. We showed that CoA is not a biologically stable molecule and cells do not take up CoA directly. We presented evidence that ecto-nucleotide-pyrophosphatases hydrolyzed CoA into 4'-phosphopantetheine. In contrast to pantetheine 21 , 4'-phosphopantetheine was stable in serum, was taken up by cells via passive diffusion and was intracellularly re-converted into CoA. Via this route, exogenous CoA rescued CoAdeprived phenotypes at the cellular, developmental, organismal and behavioral level. We showed that CoA rescue was independent of the first three classic CoA biosynthetic steps (PANK, PPCS and PPCDC) and that it depended on the last bifunctional enzyme, COASY. Our data demonstrated the existence of an alternate mechanism for cells and organisms to influence intracellular CoA levels derived from an extracellular CoA source with 4'-phosphopantetheine as the key intermediate.

CoA supplementation rescues CoA-depleted phenotypes
In order to answer the question of whether cells are able to obtain CoA from sources other than classic de novo biosynthesis (Figure 1a), we first sought to determine whether extracellular sources of CoA could serve as a supply for intracellular CoA. For this, we used RNA interference to induce PANK (first enzymatic step) depletion to block the de novo biosynthesis route and to create a CoA-depleted phenotype. Subsequently the rescue potential of exogenous CoA was tested. PANK depletion by RNA interference in Drosophila cultured S2 cells (Figure 1b insert) was associated with a reduction in cell count (Figure 1b,c) and histone acetylation levels ( Figure 1d-e), as previously demonstrated 4 . Addition of CoA to the medium of the cultured cells rescued the cell count in a concentration-dependent manner ( Figure 1c) and restored the histone acetylation phenotype (Figure 1f). Next, we questioned whether this rescue also applied to other cell types and systems of impaired CoA biosynthesis. Treating Drosophila S2 cells with the chemical PANK inhibitor Hopantenate (HoPan) 22 , also induced a decrease in cell count (Supplementary Results,Supplementary Figure 1a) and histone acetylation levels . This HoPan-induced phenotype was also rescued by direct supplementation of CoA to the medium of the cells (Supplementary Figure 1a,d). Next, we studied the effects of HoPan in mammalian HEK293 cells to address the possibility that the beneficial effects of exogenous CoA were insect cell-specific. When HEK293 cells were treated with HoPan, they showed a phenotype similar to Drosophila S2 cells, with decreased cell count and impaired histone acetylation. When CoA was added to the culture medium both the decreased cell count (Figure 1g) and the impaired histone acetylation phenotypes ( Figure 1h) were rescued. These in vitro results confirmed the potency of exogenous CoA to rescue phenotypes induced by impaired PANK in diverse cellular systems. d we also , suramin ,2′ disulwed that efficiently into 4′a, unlike only mild tion into (Fig. 4d). dation in ydrolases t degrade mplicated zymes to etheine in erum stathere was ls of CoA ementary that CoA eine was ypes impair-A levels, se in 4′ore, they phosphols would C analysis ls indeed phosphoentation ntetheine levels of oreover, s added h HoPan . 5c), the scued. 4′-Phosphopantetheine histone acetylation defect in NK/fbl RNAi (Supplementary of 4′-phosphopantetheine over the cell membrane. First, we incubated S2 cells cultured at 25 °C (the normal culturing temperature for these cells) and 4 °C with labeled 4′-phosphopantetheine. data are mean ± s.d.; each symbol represents an individual data point (n = 45). (b) life-span analysis of C. elegans pnk-1 mutants with (n = 90) and without (n = 96) coA treatment compared with wild types with (n = 92) and without (n = 83) coA treatment. differences in survival curves were significant (P < 0.001, log-rank (Mantel-cox) test) between untreated and coA-treated pnk-1 mutants. (c)  To test the effect of CoA supplementation in vivo, we used homozygous Caenorhabditis elegans (C. elegans) pantothenate kinase (pnk-1) mutants4, which showed decreased motility ( Furthermore, when a Drosophila w1118 control fly line was treated with HoPan, larval lethality was induced and a decreased eclosion (emerging from the pupal case) rate was observed (Figure 2c). This HoPaninduced phenotype was fully rescued by the addition of CoA to the food of the larvae (Figure 2d).
These data demonstrated that supplementation of CoA reverted the phenotypes arising from impaired de novo CoA biosynthesis, an effect that was observed in diverse eukaryotic cell types and organisms.

External supply of CoA influences intracellular CoA
The observed rescue effect could occur in several ways. Either intracellular CoA levels could have been restored, or rescue was achieved independent of the restoration of CoA levels in the cells. If the latter was true, intracellular levels of CoA would not be restored by exogenous CoA. To investigate this, a sensitive HPLC method was developed consisting of pre-column thiol-specific derivatization of samples with ammonium 7-fluorobenzofurazan-4-sulfonate (SBDF), followed by chromatographic separation by gradient elution on a C18 column and fluorescence detection (see online Methods). The HPLC-CoA analysis showed that intracellular CoA levels were significantly reduced in extracts of HoPan-treated S2 and HEK293 cells, addition of CoA to the culture medium restored the intracellular concentration of CoA (Figure 2e,f). These results suggested that extracellular CoA exerted its beneficial effects in CoA-depleted cells by increasing and thereby "normalizing" intracellular CoA concentrations.
In serum, CoA is degraded to stable 4'-phosphopantetheine The mechanism behind this alternative CoA route was not known. The observations in Figure 1 and 2 indicated that either 1) CoA entered cells directly, although such a transport process has not been described; or 2) CoA was converted to an intermediate product that entered the cell and was converted back to CoA in a PANK-independent manner. Previous research found that CoA is not stable in liver extracts and degrades to 50% at -20°C after a week 23 , however, the stability of CoA in an extracellular environment such as in aqueous or in standard cell culture medium is unknown. Moreover, these early reports did not identify specific degraded or converted products. We measured the stability of CoA in PBS, serum-free medium, medium containing fetal calf serum and in fetal calf serum (FCS) during a 3hrs incubation. HPLC analysis revealed that CoA was relatively stable in PBS and serum free medium, with >95% of the initial concentration still present after 3hrs (Supplementary Figure 3 Fig. 12a) of these enzymes. CoA and 4′-phosphopantetheine levels were also significantly reduced in all conditions ( Supplementary Fig. 12b-e), with the exception of dCOASY mutants, which showed a significant reduction of CoA but not of 4′-phosphopantetheine (Supplementary Fig. 12f).   e. CoA was added to mouse serum and concentrations of CoA and PPanSH in mouse serum over 6 hrs were determined by HPLC analysis. Data indicate mean ± SD (n = 3). f. Relative PPanSH levels in Drosophila L1 and L2 stage larvae determined by HPLC analysis under untreated conditions (100%) and after feeding CoA. Data indicate mean ± SD (n = 3), two-tailed unpaired Student's t-test was used (**P ≤ 0.01, ***P ≤ 0.001).
g. Concentration of CoA and PPanSH in mouse serum determined by HPLC analysis, 30 min after in vivo injecting various amounts of CoA intravenously. Data represent mean (n = 2), in g; Solid thick bars without error bars indicate no PPanSH or CoA was detected.
peak had to be a thiol-containing molecule, we speculated that it could be a CoA degradation product, namely dephospho-CoA, 4'-phosphopantetheine (PPanSH), or pantetheine 2 . In contrast to dephospho-CoA and pantetheine, 4'-phosphopantetheine is not commercially available and hereto, we chemically synthesized this compound (Supplementary Note) in order to complete our analysis. HPLC analysis and comparison with standards demonstrated that the thiol-containing degradation product of CoA was neither dephospho-CoA nor pantetheine (Supplementary Figure  To investigate whether this conversion also occured in vivo, Drosophila larvae were fed CoA, and L1 and L2 stage larval extracts were obtained after 2 days and 3 days of feeding, respectively. HPLC analysis showed that externally added CoA resulted in increased levels of 4'-phosphopantetheine in both L1 (>20 fold) and L2 larvae (>60 fold) ( Figure 3f). To investigate whether this conversion also occured in higher organisms, different concentrations of CoA were injected intravenously into adult mice, and plasma was collected after 30 min and 6 hrs. HPLC analysis in combination with mass spectrometry revealed the presence of low levels of endogenous 4'-phosphopantetheine in fresh mouse serum  and showed that the injected CoA was rapidly converted to 4'-phosphopantetheine after 30 min ( Figure  3g). Moreover we demonstrated using mass spectrometry that elevated levels of 4'-phosphopantetheine were still present in the plasma 6 hrs after CoA injection (Supplementary Figure 6d).
These data indicated that CoA is converted into 4'-phosphopantetheine in vitro and in vivo. Furthermore these results suggested that 4'-phosphopantetheine could be the principal molecule that was taken up by CoA-depleted cells, converted back into CoA intracellularly and this resulted in rescue of the CoAdepleted phenotypes.

Conversion of CoA into 4'-phosphopantetheine by ENPPs
Next we questioned which factors could be responsible for the conversion of CoA into 4'-phosphopantetheine in serum. To identify candidate enzymes, serum from various species (fetal calf, mouse and human) was pre-conditioned, and CoA conversion into 4'-phosphopantetheine was assessed. First, the effect of heat inactivation of the serum was studied. HPLC analysis showed that heating the serum at 56°C for 30 min completely abolished the conversion of CoA to 4'-phosphopantetheine ( Figure  4a), this indicated the involvement of enzymes or proteins in the process. Second, the conversion of CoA to 4'-phosphopantetheine requires the hydrolysis of a phosphoanhydride bond, which is typically catalyzed by (pyro)phosphatases or hydrolases. The majority of enzymes in the known family of (pyro)phosphatases and hydrolases depend on metal ions for their activity. To test these candidates, ethylenediaminetetraacetic acid (EDTA) was added to serum to chelate metal ions. Treatment of serum with EDTA completely prevented the formation of 4'-phosphopantetheine ( Figure 4b). This strongly © 2015 Nature Am nature CHeMICaL BIOLOGY | AdvAnce online publicAtion | www.nature.com/naturechemicalbiology

5
We aimed to test this hypothesis. In the Drosophila genome, we identified single orthologues for all the enzymes involved in CoA biosynthesis 12 , including dPANK/fbl, dPPCDC and dCOASY.
We obtained a set of Drosophila strains carrying either mutations in genes encoding these enzymes or an upstream activation sequence (UAS)-RNAi construct. Homozygous mutants or flies ubiquitously expressing the RNAi construct showed downregulation of mRNA levels ( Supplementary Fig. 11) or protein levels lines, hypomorphic or null mutants) used. This has been reported previously not only for Drosophila but also for other organisms 12,37 .
Regardless of the severity of the phenotypes and the developmental stage in which they first arose, the determination of the rescue potential of CoA in the available mutants was a valuable tool for testing our hypothesis. A scheme of the hypothesis, the Drosophila life span and the phenotypes of the fly lines used are presented in Supplementary Figure 10.  CoA was incubated in heat-inactivated fetal calf serum, mouse serum and human serum for 3 hrs and CoA stability was measured using HPLC analysis. b. CoA stability was determined in fetal calf serum, mouse serum and human serum pre-treated with EDTA (10mM) and CoA levels were measured after 3 hrs using HPLC analysis. c. CoA was incubated in fetal calf serum, mouse serum and human serum pre-treated with ATP or ADP (both 10mM) and CoA levels were measured after 3 hrs. d. CoA stability was determined in fetal calf serum, mouse serum and human serum pre-treated with sodium fluoride (NaF), levamisole, suramin or 4,4'-diisothiocyanatostilbene-2,2' disulphonic acid (DIDS) (all 10mM) and CoA levels were measured after 3 hrs. Data in all the above represent mean ± SD (n = 3), two-tailed unpaired Student's t-test was used for statistical analysis to compare indicated subsets (*P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001). In all the above experiments CoA was added to the indicated sera to a final concentration of 10µM, and percentages relative to CoA stability for 3 hrs in PBS (100%) are indicated (see Online methods for detailed protocol).
suggested that metal ions were required for the CoA conversion. The most likely hydrolase or (pyro) phosphatase candidates, which possess the ability to convert CoA and which are metal-ion dependent for their activity, are nudix hydrolases, alkaline phosphatases and ecto-nucleotide pyrophosphatases (ENPPs) 24-28 . These candidate enzymes are also known for their ability to hydrolyze adenosine 5'-triphosphate (ATP) and adenosine 5'-diphosphate (ADP) 29-31 . Therefore, we tested the conversion of CoA into 4'-phosphopantetheine in serum after addition of excess ATP and ADP. Both competitively blocked the conversion in all sera tested, further underscoring the involvement of one of these enzymes ( Figure 4c). Alkaline phosphatase and ENPPs are excreted by cells and are present in serum 29, 32 . Nudix hydrolases are intracellular hydrolases of CoA 25, 30 ; however, an additional possible extracellular role for this class of hydrolases cannot be excluded.
Next, we used sodium fluoride (NaF), and levamisole to inhibit nudix hydroloases, and alkaline phosphatase respectively, and in addition, we also used two different ENPP inhibitors, suramin and 4,4'-diisothiocyanatostilbene-2,2' disulphonic acid (DIDS) [33][34][35] . Our data showed that only suramin and DIDS were able to efficiently abolish the degradation of CoA into 4'-phosphopantetheine in all the sera, unlike levamisole, and sodium fluoride (NaF) which showed only mild or no inhibition of CoA degradation into 4'-phosphopantetheine, respectively ( Figure 4d). Sodium fluoride (NaF) did not influence CoA degradation in serum, which indicated that either nudix hydrolases were not present or did not degrade CoA in serum. These experiments implicated ENPPs as the most likely class of enzymes to hydrolyze CoA into 4'-phosphopantetheine in serum. Moreover, in all of the CoA serum stability experiments listed above, there was an inverse correlation between the levels of CoA and 4'-phosphopantetheine (Supplementary Figure 7a-c), which underscored that CoA degradation into 4'-phosphopantetheine was mediated by ENPPs.

4'-phosphopantetheine rescues CoA-depleted phenotypes
Our data so far predicted that PANK impairment not only induced decreased CoA levels but also decreased levels of 4'-phosphopantetheine. Furthermore, it predicted that addition of 4'-phosphopantetheine to CoA-depleted cells could rescue the induced phenotypes. HPLC analysis of HoPan treated Drosophila S2 cells indeed showed reduced levels of 4'-phosphopantetheine, and external supplementation with either CoA or 4'-phosphopantetheine significantly increased intracellular levels of 4'-phosphopantetheine ( Figure 5a). Moreover, when 4'-phosphopantetheine was added to Drosophila S2 cells treated with HoPan ( Figure Figure 8i). Next we investigated whether intact 4'-phosphopantetheine entered cells and whether it was subsequently converted into CoA. First we treated intact cultured Drosophila S2 cells with stable isotopelabelled 4'-phosphopantetheine under various conditions, and mass spectrometry analysis was used to Cell count was determined in control (100%) and dPANK/ fbl RNAi treated Drosophila S2 cells with an without addition of PPanSH (100µM) to the medium. d. S2 cells, with and without HoPan (0.5mM) were incubated with stable-isotope labelled PPanSH(D4) (100µM) and levels of both unlabeled CoA and labelled CoA(D4) were measured using mass spectrometry. Cumulative CoA and CoA(D4) levels were used for statistical analysis. e. Stable-isotope labelled PPanSH(D4) (100µM) was added to S2 cells at 4°C and 25°C, incubated for various time intervals and mass spectrometry was used to measure levels of labelled compound in harvested cell extracts. f. Stable-isotope labelled PPanSH(D4) (10, 100, 1000µM) was added to S2 cells for 30min and mass spectrometry was used to measure levels of labelled compound in harvested cell extracts. Data in all the above represent mean ± SD (n = 3), two-tailed unpaired Student's t-test was used for statistical analysis to compare indicated subsets (*P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001).

NATure CHemICAl BIOlO
We first tested two available mutants for dPANK/fbl: the hypomorphic 19 dPANK/fbl 1 and the null mutant dPANK/fbl null . Homozygous dPANK/fbl 1 mutants showed reduced levels of dPANK/Fbl protein, and in homozygous dPANK/fbl null mutants levels of dPANK/Fbl protein were below detection (Supplementary Fig. 12a). Correlating with this, homozygous dPANK/fbl 1 mutants showed a reduced adult life span ( Fig. 6a and Supplementary Fig. 13a) 12,19 , whereas homozygous dPANK/fbl null mutants developed only until an early L2 larval stage, and pupae were not observed (Fig. 6b). Addition of CoA to the food of the homozygous dPANK/fbl 1 mutants increased the life span from 20 d to 40 d ( Fig. 6a and Supplementary Fig. 13a (c) cell counts in control (100%) and dPANK/fbl RnAi-treated Drosophila S2 cells with and without ppanSH (100 µM) added to the medium. (d) levels of unlabeled coA and labeled coA(d4) in S2 cells with and without Hopan (0.5 mM) incubated with stable isotope-labeled ppanSH(d4) (100 µM), as measured via mass spectrometry. cumulative coA and coA(d4) levels were used for statistical analysis. (e) levels of labeled compound in cell extracts from S2 cells incubated at 4 °c and 25 °c with stable isotope-labeled ppanSH(d4), as measured by mass spectrometry. (f) levels of stable isotope-labeled ppanSH(d4) in S2 cell extracts after the addition of various concentrations to the cells, as measured by mass spectrometry. data represent mean ± s.d. (n = 3). *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, two-tailed unpaired Student's t-test.
conditions of impaired CoA biosynthesis by HoPan. Next we investigated the characteristics of the passage of 4'-phosphopantetheine over the cell membrane. First, within 30 min after the incubation of cells with labelled 4'-phosphopantetheine, the intracellular presence of labelled 4'-phosphopantetheine was detected in cells cultured at 25°C (normal culturing temperature of S2 cells) and 4°C. There was no significant difference in the intracellular levels of labelled 4'-phosphopantetheine between these two conditions ( Figure 5e). Next we investigated whether under these conditions, the levels of intracellular 4'-phosphopantetheine increased to the same extent as the externally added increased concentrations of 4'-phosphopantetheine. Hereto an increasing concentration series (10-100-1000µM) of labelled 4'-phosphopantetheine was added to the cells. This appeared to be the case (Figure 5f). These results indicated that the capacity of cells to accumulate the externally provided 4'-phosphopantetheine was not influenced by temperature and was determined by extracellularly provided concentrations. Finally we investigated the membrane permeating efficiency of 4'-phosphopantetheine using a Parallel Artificial Membrane Permeability Assay (PAMPA assay)36. Based on this assay 4'-phosphopanteheine but not CoA showed membrane permeating properties (Supplementary Figure 9e-f). Altogether, these results pointed to a capacity of 4'-phosphopanteheine to permeate membranes via passive diffusion.

CoA rescues dPANK/fbl, dPPCDC but not dCOASY phenotypes
Our data showed that CoA from external sources can replenish intracellular CoA levels through its hydrolysis product 4'-phosphopantetheine and subsequent conversion back to CoA. The most likely candidate for the latter conversion is the last bifunctional enzyme, COASY, of the classic CoA biosynthetic pathway. This hypothesis (Supplementary Figure 10) predicts that CoA but not Vitamin B5 can rescue phenotypes caused by mutations in genes encoding enzymes upstream of 4'-phosphopantetheine in the CoA pathway. As a corollary, CoA would not be predicted to rescue COASY mutant phenotypes.
We aimed to test this hypothesis. In the genome of Drosophila single orthologs were identified for all the enzymes involved in CoA biosynthesis 12 , including dPANK/fbl, dPPCDC and dCOASY. A set of Drosophila strains was obtained, carrying either mutations in genes encoding these enzymes or carrying a UAS-RNAi construct. Homozygous mutants or flies ubiquitously expressing the RNAi construct showed a downregulation of mRNA levels ( Supplementary Figure 11a It should be stressed that not all mutants with defects in CoA biosynthesis enzymes showed an identical phenotype, which can be explained by the type of flylines (RNAi expressing lines, hypomorphic or null mutants) used. This has been reported previously not only for Drosophila but also for other organisms 12,37 . Regardless of the severity and developmental stage in which the phenotypes manifested, the determination of the rescue potential of CoA in the available mutants was a valuable tool to test our hypothesis.   (9 mM  Finally we tested a mutant line of the bifunctional enzyme dCOASY, downstream of 4'-phosphopantetheine. Homozygous dCOASY mutants developed until first instar larval stage and addition of CoA to the food did not result in a significant rescue (Figure 6g). As a negative control for all rescue experiments, vitamin B5 was added to the food, and this did not result in any significant rescue of the phenotypes. A summary of the rescue with CoA in all flylines is presented in Supplementary Figure 10.
To test our hypothesis further, COASY was downregulated with RNAi in mammalian HEK293 cells. Under these conditions levels of COASY protein, CoA and histone acetylation were significantly reduced (Supplementary Figure 14g- unaltered in COASY-compromised mammalian cells (Supplementary Figure 14g). Addition of CoA to the medium neither rescued the COASY RNAi-induced decrease in intracellular CoA levels (Supplementary Figure 14g) nor restored histone acetylation levels (Supplementary Figure 14f). These results were also in agreement with our hypothesis.
Taken together, these results demonstrated that impairment of the CoA biosynthetic pathway by genetic manipulation can give rise to highly complex pleiotropic effects affecting lifespan, development and fecundity. These phenotypes can be (partially) rescued by the addition of CoA to the food of the animals, which is then hydrolyzed to 4'-phosphopantetheine which crosses the plasma membrane via passive diffusion before being converted back to CoA intracellularly, a step requiring COASY (Figure 6h).

DISCUSSION
In our study we addressed the basic question of whether cells and organisms possess alternative ways to obtain the essential molecule CoA in addition to the canonical pathway utilizing Vitamin B5. We demonstrate that cells and organisms are able to acquire exogenous CoA, which is converted into the stable molecule 4'-phosphopantetheine, which enters cells and is converted again into CoA. These newly identified characteristics of 4'-phosphopantetheine suggest that this molecule can serve as a transport form of CoA or stable reservoir for rapid access and conversion. The proposed mechanism hypothetically allows a net flow of CoA or 4'-phosphopantetheine between cells and between membrane-bound cellular compartments. Our data further suggest that not all cells or organelles within an organism need to harbor all CoA biosynthetic enzymes in order to obtain CoA and that the route to CoA does not necessarily need to follow the archetypal direction starting from the uptake of Vitamin B5. Moreover, these observations hold promise for therapeutic intervention for PKAN because 4'-phosphopantetheine is stable in serum and passes through cell membranes, thereby allowing for restoration of intracellular CoA levels in cells with defective CoA synthesis. The stability of 4'-phosphopantetheine is in strong contrast to characteristics of the dephosphorylated form, pantetheine, which is degraded rapidly in serum by pantetheinases into vitamin B5 and cysteamine 20, 21 . These results show that the phosphate group protects the molecule from degradation and allows 4'-phosphopantetheine to serve as an effective substrate for CoA biosynthesis from its ready reserve in the circulation.
One intriguing question is whether the proposed route shown here has a physiological function or whether it is artificially provoked by manipulating concentrations of extracellular CoA. Compared to CoA concentrations in cytoplasm [0,02-0,14 mM] and mitochondria [2,2-5 mM] 38 , the concentrations used in our study (_M range) are relatively low. Answers may come from previous studies demonstrating that bacteria are able to excrete, but not take up 4'-phosphopantetheine from their environment, suggesting that bacteria-derived 4'-phosphopantetheine may be present in the digestive system 39 . The presence of endogenous 4'-phosphopantetheine in mouse serum and in dPANK/fbl null mutants are consistent with a possible source of extracellular 4'-phosphopantetheine. The function of such a 'ready' pool of CoAprecursor may be for transport from one organ to another. In addition to being a source for intracellular CoA, extracellular CoA or 4'-phosphopantetheine may have signaling functions based on reports of an effect of CoA on platelet aggregation and vasoconstriction 40, 41 . Our results suggest that these effects, which have been attributed to CoA, may in fact be from 4'-phosphopantetheine. Future experiments are required to demonstrate the presence and possible impact of a net flow of CoA between organelles, cells and organisms (such as between intestine bacteria to the host).
The ability of 4'-phosphopantetheine to translocate across membranes answers long-standing questions regarding the intracellular distribution of CoA and its biosynthetic enzymes. CoA is present in the cytoplasm and in organelles including mitochondria 38 . All CoA biosynthetic enzymes are present in cytoplasm 42 but only a subset have been found in mitochondria. It remains unclear how mitochondria obtain CoA, and the localization of COASY (but not the other CoA biosynthetic enzymes) to the mitochondrial matrix, is also unexplained 17, 43 . It has been postulated that CoA synthesized in the cytosol can be transported into the mitochondrial matrix by specific CoA transporters localized in the mitochondrial inner membrane 44 . Indeed evidence for the presence of such CoA transporters has been presented 45 . Based on our observations we hypothesize that 4'-phosphopantetheine is able to passage over the mitochondrial inner membrane into the matrix and be subsequently converted into CoA by matrix COASY. This may explain the localization of COASY in the mitochondrial matrix 17, 43 .
The presence of a 4'-phosphopantetheine uptake mechanism may have large public health implications. Pathogens and parasites acquire resistance to current treatments, and species-specific inhibitors of CoA biosynthetic enzymes are attractive candidates for a new class of antibiotics and anti-malarial drugs 46, 47 . Such inhibitors may be more effective anti-microbials when 4'-phosphopantetheine uptake is blocked as well. Alternatively, differences in the uptake capacity of 4'-phosphopantetheine by eukaryotic cells (this manuscript) and bacteria 39 may be further explored as possible targets for antimicrobial strategies.
CoA is essential for coordinating key aspects of cell function. It is therefore not surprising that an extracellular pool exists to facilitate swift replenishment and that it relies on the formation of a stable intermediate. While these novel observations raise many new questions about CoA metabolism, they also suggest therapeutic approaches for a range of life-threatening human diseases.

Drosophila S2 Cell Culture, RNA Interference, and CoA and 4'-phosphopantetheine treatment
Drosophila Schneider's S2 cells were maintained at 25°C in Schneider's Drosophila medium (Invitrogen) supplemented with 10% fetal calf serum (Gibco) and antibiotics (penicillin/streptomycin, Invitrogen) under laboratory conditions. Synthesis of RNAi constructs and RNA interference (dsRNA) treatment was carried out as described previously 4 . Non-relevant (human gene; hMAZ) dsRNA was used as control.
The cells were incubated for 4 days to induce an efficient knock-down. Cells were then subcultured, with or without CoA (Sigma-Aldrich, Cat. No: C4780, 95% -which is used for all the experiments wherever stated below) or 4'-phosphopantetheine (PPanSH) (Acies Bio, >92%) at different concentrations and were maintained for additional 3 days until analysis for rescue efficiency of the compounds was performed. The stock solutions of compounds were made in sterile water and stored in -20°C until use.

HoPan treatment of Drosophila S2 Cell in combination with CoA or 4'-phosphopantetheine treatment
Drosophila Schneider's S2 cells were maintained at standard conditions as explained above. Cells in the exponential phase of growth were used for all the experiments. Different indicated concentrations of CoA or 4'-phosphopantetheine (deuterium labelled PPanSH(D4) or unlabelled PPanSH) were added to S2 cells either in the presence or absence of 0.5mM HoPan (Zhou Fang Pharm Chemical,; 99%) for 48hrs. Similarly, Drosophila S2 cells were treated with different concentrations of PPanSH(D4) at either 25°C or 4°C and cells were then harvested at various time points to access transport of PPanSH(D4). Stable isotope labelled PPanSH containing 4 deuterium atoms was purchased as a sodium salt (from Syncom; synthesized as previously described 48 , 99.7%). As a read out, cell count, intracellular total CoA and PPanSH levels (both labelled and unlabeled levels wherever appropriate) and histone acetylation levels were analyzed as explained below.

Western blot analysis and Antibodies
For Western blot analysis, cells were collected and washed with PBS, followed by centrifugation. The cells were lysed and sonicated in 1X Laemmli Sample Buffer and boiled for 5min with 5% β-mercaptoethanol (Sigma). Protein content was determined using DC protein assay (BioRad). Equal amounts of protein were loaded on a 10 or 12.5% SDS-PAGE gel, transferred onto PVDF membranes and blocked with 5% milk in PBS/0.1% Tween, followed by overnight incubation with primary antibodies. The primary antibodies used were: rabbit-anti dPANK/fbl, 1:4000 Eurogentec custom made as described previously 12 , mouse anti-tubulin (Sigma Aldrich Cat no: T5168, 1:5000), anti-acetyl-Histone3 (Active Motif Cat no: 39139, 1:2000), anti GAPDH (Fitzgerald Cat no: 10R-G109a, 1:10000), rabbit anti COASY (Abcam Cat no: AB129012, 1:1000). Appropriate HRP-conjugated secondary antibodies (Amersham) were used and detection was performed using enhanced chemi-luminescence (Pierce cat nog: 32106) and Amersham hyperfilm (GE healthcare). Band intensities were quantified with Image-studio software. Full uncut gel images for all Westerns displayed in this paper are shown in Supplementary Figures 15 and 16.

C. elegans lifespan assay
After synchronization, C. elegans L1 animals were grown on control NMG plates or NGM plates supplemented with 400µM CoA. The life span experiments were started by transferring 100 one-day old adults per condition on NGM plates, which contained 5-fluoro-2′deoxy-uridine (FUDR) to inhibit growth of offspring. Once a day surviving animals were counted, the worms that disappeared or crawled out of the plate were excluded from the analysis.

C. elegans motility assay
After synchronization, L1 C. elegans were grown on control NMG plates or NGM plates containing various concentrations of CoA. One-day old adults were placed in a drop of M9 buffer and allowed to recover for 30sec. During the following 30sec, the number of body bends were counted. A movement was scored as a bend when both the anterior and posterior ends of the animal turned to the same side. At least 15 worms were scored per condition and each experiment was repeated thrice. The sequential light microscopy images demonstrating movements of C. elegans in M9 buffer were captured using Leica MZ16 FA microscope at 32x magnification within the time frame of 1sec and processed using ImageJ (National Institutes of Health, Maryland, USA) and Adobe Photoshop (Adobe Systems Incorporated, San Jose, California, USA).

Drosophila maintenance and crosses
Drosophila melanogaster stocks/crosses were raised on standard cornmeal agar fly food (containing water, agar 17 g/L, sugar 54 g/L, yeast extract 26 g/L and nipagin 1.3 g/L) at 25°C.

Drosophila larval collection and larval count experiment
One week old flies (in the ratio 10 females and 5 males) were kept on 5ml of standard fly food in a vial at 25°C with or without various concentrations of CoA or Vitamin B5 (Sigma, Cat. No. P5155). The flies were allowed to lay eggs for 2 days and parent flies were then discarded. The L1, L2 and L3 larvae were collected from the food with 20% sucrose at appropriate time (day 4, 6 and 8 respectively) for larval counting and stored in -80°C until analysis. The pupal count was performed between 10-12 days.

Drosophila HoPan Toxicity and CoA Rescue Experiment
One week old w1118 flies (in the ratio 10 females and 5 males) were kept in vials containing standard fly food with or without HoPan and CoA at indicated concentrations. The flies were allowed to lay eggs for 2 days, after which the adults were discarded. The resulting offspring were allowed to develop. The numbers of flies which eclosed were counted to evaluate HoPan toxicity and CoA rescue efficiency.

Drosophila life span
One-day old female adult flies of Drosophila homozygous mutants or RNAi-constructs expressing lines, were collected with appropriate controls and were kept on standard fly food at 25°C with or without CoA or Vitamin B5 (Sigma) at necessary concentration (in 50µl added on top of the fly food and dried before flies were added). The flies were counted every 12-24hrs and flipped to new fly food vials with or without CoA or Vitamin B5.
Drosophila ovary rescue experiment UAS-dPPCDC RNAi constructs were ubiquitously expressed under the control of Actin-Gal4. The crosses were raised at 25°C. F1 RNAi-construct expressing females and control females were collected shortly after eclosion and transferred to standard fly food or food containing Vitamin B5 or CoA (18mM). Flies were maintained for 2 days on this food at 25°C. After this period extra yeast and w1118 control males were added and the crosses were kept at 25°C for another 2 days. After this 4 day period ovaries were dissected and stained for further analysis. The vials (or plates) from the crosses (with eggs that were being laid during the 4 day period of CoA treatment) were kept for another 10 days and offspring numbers were counted after eclosion.

RNA isolation, quantitative Real-Time PCR, and primers
Drosophila larvae and samples of 1-day old adult flies or larvae were collected for homozygous dPPCDC mutants, dPPCDC RNAi-construct expressing lines and for homozygous dCOASY mutants, followed by brief washing with PBS. The samples were lysed in TRIZOL (Invitrogen) for RNA extraction and reverse transcribed using M-MLV (Invitrogen) and oligo(dt) 12-18 (Invitrogen). SYBR green (Bio-Rad) and Bio-Rad Real-Time PCR with specific primers were used for gene expression level analysis. The expression levels were normalized for rp49 (house-keeping gene). The Primer sequences used were dPPCDC (TGCACCTGCGATGAATACCC; TCGGCTGAAAGGCGGATAAC), dCOASY (GGCTGTGCGGCGGATTATTG; CGGGTTAAAGGCTGCTCTGG) and rp49 (GCACCAAGCACTTCATCC; CGATCTCGCCGCAGTAAA) (Biolegio).

Drosophila ovary dissection and staining
Drosophila ovaries were collected in cold PBS and fixed in 4% formaldehyde (from methanol-free 16% Formaldehyde Solution, Thermo Scientific) for 45min at RT. The fixed tissue was washed in PBS + 0.1% Triton-X-100 for 1hr at RT and afterwards permeabilized in PBS + 0.2% Triton-X-100 for 1hr. Finally the ovaries were stained with Rhodamin-Phalloidin (20U/ml) to detect F-actin and DAPI (0.2 _g/ml) for DNA. Finally the samples were mounted in 80% glycerol and analyzed on a Zeiss-LSM780 NLO confocal microscope with Zeiss Zen software. Adobe Photoshop and Illustrator (Adobe Systems Incorporated, San Jose, California, USA) were used for image assembly.

PAMPA assay procedure
Parallel Artificial Membrane Permeability Assay (PAMPA) was performed and processed according to manufacturer's instructions (BD Gentest Pre-coated PAMPA plates). Briefly, two superimposed wells are separated by an artificial lipid-oil-lipid membrane. The compound of interest (PPanSH, CoA, caffeine, amiloride) was added to the bottom well in phosphate-buffered saline, whereas the top well was filled with phosphate-buffered saline alone. After 5hrs of incubation at RT, concentrations of the different compounds were measured using UV-VIS absorption spectroscopy (BMG Labtech SPECTROstar Omega) along with calibration curves for all compounds. The permeability efficiency was further calculated according to manufacturer's instructions (Supplementary figure 9e). For caffeine and amiloride, four replicates were performed; for PPanSH and CoA twelve replicates were performed. Caffeine and amiloride were obtained from Sigma.

Serum collection from PKAN patients
Serum was collected from the blood of PKAN patients and respective healthy family members (control) using standard protocols. Briefly, venous blood was collected in commercially available red topped Vacutainer tubes (Becton Dickinson) and allowed to remain at RT for 15-30min undisturbed for the blood clotting. The tubes were then centrifuged at 2,000 g for 10min at 4°C. The resulting supernatant serum was immediately transferred to 2ml cryovials and maintained at -80°C until CoA stability assessments were performed. Blood samples and clinical data were obtained under OHSU's IRB-approved repository protocol #7232 following informed consent.

CoA and pantethine serum stability measurements
CoA serum stability studies were conducted in commercially obtained serum and in serum collected from PKAN patient and healthy family members as controls. Human and Mice sera were purchased from Sigma and Fetal calf serum was purchased from Gibco. Additionally, dMEM medium with or without 10% fetal calf serum was used for evaluating CoA stability. Briefly, all sera and samples were incubated for 30min at 37°C with or without pre-conditioning compounds at final concentration 10mM [Adenosine 5′-triphosphate (ATP), Adenosine 5′-diphosphate (ADP), Ethylenediaminetetraacetic acid (EDTA), Levamisole, Suramin, 4,4′-Diisothiocyanatostilbene-2,2′-disulfonic acid disodium (DIDS) and sodium fluoride (NaF), all purchased from Sigma] followed by addition of CoA 20µM in the ratio of 1:1 and incubated at 37°C after brief vortex for indicated time intervals. For heat inactivation, all sera were incubated for 30min at 56°C after which CoA was added. Serum samples at different time points were collected, deproteinized and analyzed by HPLC as described below. For pantethine serum stability, pantethine (Sigma) was incubated in fetal calf serum, mice serum and human serum for 15min in 37°C and total levels of pantetheine and cysteamine was measured using HPLC.

Mice and CoA intravenous injection study
Adult male mice of C57BL/6J 129/SvJ mixed genetic background were used for this study. Two mice (approximately 25-30g wt) were used for each condition. 0.1mg or 0.5mg CoA in 0.25ml saline solution was injected intravenously (i.v) into the tail vein. Saline solution (0.25ml) was injected to control groups. After 30min and 6hrs blood samples were collected and further processed to obtain plasma followed by sample preparation for HPLC or LC-MS analysis as indicated below. All animal studies were approved by the Ethics Committee of the Foundation IRCCS Neurological Institute C. Besta, in accordance with guidelines of the Italian Ministry of Health: Project no. BT4/2014. The use and care of animals followed the Italian Law D.L. 116/1992 and the EU directive 2010/63/EU.

HPLC sample preparation protocol for total CoA and 4'-phosphopantetheine measurement
Samples were briefly washed with ice-cold PBS solution. Samples were sonicated thoroughly in 100µl ice-cold PBS and centrifuged for 10-15min at 4°C to collect supernatant. Tris(2-carboxyethyl)phosphine hydrochloride (Sigma) (50mM; 10µl) was added to 50µl sample supernatant and were incubated at RT for 15min after vortex-mixing. Saturated ammonium sulfate solution or Millipore 3KD centrifugal filter units were used to remove proteins. The samples were centrifuged at 14,000 rpm for 15min at 4°C. The clear supernatant (50µl) or the filtrate was derivatized with 45µl of ammonium 7-flurobenzo-2-oxa-1,3-doazole-4-sulfonate (SBD-F, Sigma) (1mg/ml in borax buffer -0.1M containing 1mM EDTA disodium, pH 9.5), and 5µl ammonia solution (12.5% v/v, Merck Millipore) at 60°C for 1hr. The derivatized samples were placed in a refrigerated autosampler (10°C) in the Shimadzu HPLC system, and injected for total CoA and PPanSH analysis using optimized chromatographic separation conditions combined with fluorescence detection (described below).

Chromatography separation condition
Chromatographic analysis was performed with a Shimadzu LC-10AC liquid chromatograph, SCL-10A system controller, SIL-10AC automatic sample injector and LC-10AT solvent delivery system. Shimadzu RF-20Axs fluorescence detector was used for derivatized sample extract analysis. The fluorescence detector was set at excitation and emission wavelengths of 385nm and 515nm, respectively. Signal output was collected digitally with Shimadzu Labsolution software and postrun analysis were performed. Chromatographic separation of the analytes was achieved with a Phenomenex Gemini C18 guard column (4 x 3mm) connected to a Phenomenex Gemini NX-C18 analytical column (4.6 x 150mm; 3µm particles) at 45°C. The two mobile phases consisted of A: 100mM ammonium acetate buffer (pH 4.5) and B: acetonitrile. Flow rate was maintained at 0.8ml/min with a slow gradient elution: 0% B till 7min, 20% B at 20min, 20% B at 22min, 50% B at 23min, maintained at 50% B till 27min, 0% B at 28min and 7-10min for column re-equilibration.

Sample preparation for mass spectrometry and instrumental parameters
Samples were briefly washed with ice-cold PBS solution. Samples were then sonicated thoroughly in 100µl ice-cold milliQ (MQ) water containing 50mM Tris(2-carboxyethyl)phosphine hydrochloride. Subsequently 100µl saturated ammonium sulfate was added to each sample and centrifuged for 20 min at 10°C, 16,100 rcf to collect supernatant. To 150µl of supernatant, 15µl of ammonium hydroxide (12.5%) was added and 20µl was injected for LC-MS (liquid chromatography-mass spectrometry) analysis. For mouse plasma analysis, 50µl of MQ water containing 50mM Tris(2-carboxyethyl)phosphine hydrochloride was added to 50µl of plasma and processed further as mentioned above. Appropriate dilution series of standard CoA, PPanSH and PPanSH(D4) was processed similarly before analysis. The LC separation of metabolites were obtained using Phenomenex Gemini NX-C18 analytical column (4.6 x 150mm; 3µm particles) at 45°C. The flow was maintained at 1ml/min with optimized mobile phase gradient of MQ water (A), 200mM ammonium acetate (NH 4 Ac) in 95/5 MQ water/acetonitrile adjusted to pH 4.5 with acetic acid (B), and acetonitrile (C). The separated analytes were detected with positive mode mass spectrometry (Sciex API5500 Q-trap) under unit resolution.

Statistical Analysis
All experimental results are presented as mean of at least 3 independent experiments ± SD, unless otherwise stated. Statistical significance was determined by a two-tailed unpaired Student's t test between appropriate groups wherever applicable. For life span survival curve, more than 80 flies or C.elegans were used in each group and statistical significance was determined using Log-rank (Mantel-Cox) test (See figure legends for exact number or flies/C.elegans used in survival analysis). Statistical P values ≤ 0.05 were considered significant (*P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001). Data were analyzed using GraphPad Prism    Pantethine is rapidly degraded in serum. Pantethine was incubated for 15min at 37°C in fetal calf serum, mice serum and human serum and levels of total pantetheine and cysteamine were measured using HPLC. b. CoA hydrolysis to PPanSH in serum derived from PKAN patients and their healthy family members as a control. CoA (20µM) was incubated for 3 hrs at 37°C in serum derived from PKAN patient and in serum derived from their healthy family members as a control. Levels of CoA and PPanSH were measured using HPLC analysis. In patient's serum CoA was as efficiently hydrolysed into PPanSH as in serum derived from healthy family members. Error bars represent ± SD where applicable for mean values (n = 3). Genders were indicated as female=F and male=M. c. CoA was added to human serum and concentrations of CoA and PPanSH in human serum over 6 hrs were determined by HPLC analysis. Data points indicate mean value ± SD (n = 3).     a-f. Immunofluorescence was used to visualize protein acetylation levels in control (a,d), dPANK/fbl RNAi treated (b) and HoPan treated (e) S2 cells with and without PPanSH (c,f). An antibody against acetylated Lysine (green), Rhodamin-Phalloidin (red), marking F-actin, and DAPI (blue, DNA) were used. Addition of PPanSH rescues acetylation defects of dPANK/fbl RNAi and HoPan treated S2 cells. Scale bars indicate 20µm. g. Cell count of mammalian HEK293 control cells (100%), cells treated with HoPan with and without CoA or PPanSH added to the medium. Data indicate mean values ± SD (n = 3) and two-tailed unpaired Students t-test was used for statistical analysis. h. Relative CoA levels were determined by HPLC of control (100%) and HoPan treated HEK293 cells with and without CoA or PPanSH added to the medium. Data indicate mean values ± SD (n = 3) and two-tailed unpaired Students t-test was used for statistical analysis. i. Western blot analysis and quantification to determine histone acetylation levels of control HEK293 cells, cells treated with HoPan with and without CoA or PPanSH. Data represents mean values ± SD (n = 3) and two-tailed unpaired Students t-test was used for statistical analysis.
In all the above data representation, statistical significance was indicated as applicable (*P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001).     All the data sets in b-f indicate mean ± SD (n = 3) in the above representations and two-tailed unpaired Student's t-test was used for statistical analysis (*P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001).  Lifespan analysis of hypomorphic (dPANK/fbl 1 ) homozygous female mutants untreated (n = 207) and treated with various concentrations of CoA (6mM, n= 105; 9mM, n = 115; and 12mM, n = 88) added to the food. Survival curves were found to be significant with P value < 0.001, analyzed with Log-rank (Mantel-Cox) test, between untreated and all CoA treated dPANK/fbl 1 mutants. b. In various animal food sources (yeast, E.coli and mouse liver) levels of CoA and PPanSH were measured and found to be present (n = 2).

Supplementary
Supplementary Figure 14: External supplementation of CoA rescues phenotypes of dPPCDC RNAi lines and COASY is required for CoA rescue in mammalian cells. a-c. Ovaries of 4-day old control and dPPCDC RNAi expressing flies, stained with Rhodamin-Phalloidin (magenta, marking F-actin) and the nuclear marker DAPI (green) and imaged with confocal microscopy. (a) In wild-type ovarioles strings of developing egg-chambers, from the germarium up to stage 9 were visible. Mature eggs were also found (marked by asterisks), identifiable by the presence of yolk. (b) In ovaries of the dPPCDC RNAi expressing flies, egg-chambers developed normally until stage 7. From stage 8 on, fragmented and condensed DNA was visible, indicating apoptosis (marked by white arrowheads). No egg-chambers older than stage 8/9 or mature eggs were found in these ovaries. (c) CoA treatment of the dPPCDC RNAi expressing flies improved egg-production significantly and eggs developed to maturity (marked by asterisks). Scale bars =100µM. d. Increased fertility of dPPCDC RNAi expressing females. Untreated, Vitamin B5 treated and CoA treated dPPCDC RNAi expressing females were mated with control males and put onto apple juice plates to allow egg laying for 4 days. For untreated and Vitamin B5 treated females, no or only very few eggs were observed on the plates (compare Figure   6e). CoA treated females produced a significant number of eggs that developed into pupae which eclosed resulting in viable offspring (compare Figure   6f). Scale bar = 1 cm. e. Lifespan analysis of adult female dPPCDC RNAi flies untreated (n = 111) and treated with various concentrations of CoA (9mM, n = 106; 18mM, n = 102; 21mM, n = 104) added to the food. Survival curves were found to be significant with P value < 0.01 for CoA 9mM treatment and P value < 0.001 for CoA (18 and 21mM) treatment compared to control untreated dPPCDC RNAi mutants, analyzed with Log-rank (Mantel-Cox) test.
f. RNAi was used to down-regulate COASY in HEK293 cells treated or non-treated with CoA and acetyl histone3 levels were quantified (n = 5). Insert: Western blot analysis showing successful down-regulation of human COASY by RNAi and decreased histone acetylation (and quantification). GAPDH was used as loading control. g. Relative levels of CoA and PPanSH were measured in control HEK293and COASY down-regulated cells treated with medium with and without addition of CoA (n = 4). The data (f,g) indicate the mean ± SD and two-tailed unpaired Student's t-test was used for statistical analysis ( **P ≤ 0.01, ***P ≤ 0.001).

b) Phosphorylation -synthesis of S-trityl-4'-dibenzylphosphopantetheine
Dibenzylchlorophosphate was freshly prepared by allowing a reaction of dibenzylphosphite (2.16g, 8.24mmol) with N-chlorosuccinimide (1.21g, 9.06mmol) in toluene (40ml) at room temperature for 2 h. The mixture was filtered and the filtrate was evaporated under vacuum and added to a solution of