CoA‐dependent activation of mitochondrial acyl carrier protein links four neurodegenerative diseases

Abstract PKAN, CoPAN, MePAN, and PDH‐E2 deficiency share key phenotypic features but harbor defects in distinct metabolic processes. Selective damage to the globus pallidus occurs in these genetic neurodegenerative diseases, which arise from defects in CoA biosynthesis (PKAN, CoPAN), protein lipoylation (MePAN), and pyruvate dehydrogenase activity (PDH‐E2 deficiency). Overlap of their clinical features suggests a common molecular etiology, the identification of which is required to understand their pathophysiology and design treatment strategies. We provide evidence that CoA‐dependent activation of mitochondrial acyl carrier protein (mtACP) is a possible process linking these diseases through its effect on PDH activity. CoA is the source for the 4′‐phosphopantetheine moiety required for the posttranslational 4′‐phosphopantetheinylation needed to activate specific proteins. We show that impaired CoA homeostasis leads to decreased 4′‐phosphopantetheinylation of mtACP. This results in a decrease of the active form of mtACP, and in turn a decrease in lipoylation with reduced activity of lipoylated proteins, including PDH. Defects in the steps of a linked CoA‐mtACP‐PDH pathway cause similar phenotypic abnormalities. By chemically and genetically re‐activating PDH, these phenotypes can be rescued, suggesting possible treatment strategies for these diseases.

C o n t r o l H o P a n H o P a n + C o A C o n t r o l H o P a n H o P a n + C o A C o n t r o l H o P a n H o P a n + C o A Appendix Figure S2: PCR analysis of outcrossed stocks to generate bona fide RNAi lines for dPANK/fbl and dPPCDC/ppcdc.
Original Drosophila VDRC lines containing RNAi constructs directed against dPANK/fbl and dPPCDC/ppcdc (here referred to as KK lines) contained in addition to the functional transgene insertion site at position 30B in the genome an additional transgene insertion site at position 40D, causing, when overexpressed, a wing phenotype not related to the RNAi construct (Green et al., 2014;Vissers et al., 2016). These unwanted extra insertion sites were removed by outcrossing and "cleaned-up" fly stocks were generated and verified by PCR. The cleaned-up and thereby bona fide Drosophila strains were used for all experiments.
(A) From the original Drosophila KK lines. the additional 40D transgene insertion site was removed and the bona fide strains were controlled by PCR for the absence of this additional 40D transgene insertion and the presence of the functional 30B transgene insertion site harbouring the specific RNAi constructs. The PCR was performed as previously described (Vissers et al., 2016) Appendix Figure S3: Crossing scheme explaining reduced viability in dPANK/fbl-RNAi flies (A-D) Viability of dPANK/fbl-RNAi expressing flies was determined and compared to control flies using the following assay: The Act-GAL4 driver was used to express UAS-dPANK/fbl RNAi ubiquitously in all cells of the flies. As a control the Act-GAL4 driver was used to ubiquitously express a non-relevant construct (green fluorescent protein (GFP)). The genotype of the flies used for the crosses is indicated. (A) For the controls, Act-GAL4/CyO flies were crossed with a fly strain homozygous for UAS-GFP (UAS-GFP/UAS-GFP). CyO is a dominant marker on an otherwise wildtype chromosome and allows following that chromosome. The number of adult progeny of this cross, hatching after pupal development, was determined and thereby the survival rate (SR) per genotype was calculated. Based on Mendelian genetics, in the control cross, 50% of the progeny will have an Act-GAL4/UAS-GFP genotype (expressing the GFP construct, indicated as green flies) and 50% will have a CyO/UAS-GFP genotype (not expressing the construct, indicated as black flies). Because expression of the GFP construct is not detrimental, the survival rate of the Act-GAL4/UAS-GFP genotype is 50% in the control cross.
(B/C) To determine the survival rate of dPANK/fbl-RNAi expressing flies, the following cross was performed: Act-GAL4/CyO expressing flies were crossed with a fly strain homozygous for UAS-dPANK/fbl-RNAi. The amount of surviving flies was determined. (B) In the situation that UAS-dPANK/fbl-RNAi expression is not detrimental, the progeny will consist of 50% UAS-dPANK/fbl-RNAi/CyO (with normal levels of dPANK/Fbl protein because there is no expression of the RNAi construct, indicated as black flies) and 50% will be UAS-dPANK/fbl-RNAi/Act-GAL4 (with decreased levels of dPANK/fbl, because Act-GAL4 drives the expression of the RNAi construct, indicated as red flies). (C) In the situation that expression of UAS-dPANK/fbl-RNAi is detrimental and leads to reduced viability, the viability of UAS-dPANK/fbl-RNAi/Act-GAL4 flies will be less than 50% of the total. (D) Visualizes a possible outcome caused by a decreased viability of UAS-dPANK/fbl-RNAi expressing flies compared to GFP expressing flies, both driven by the same Act-GAL4 driver. Appendix Figure S4: Downregulation of dPANK/fbl and mtACP with RNAi constructs using the wing specific driver MS1096-GAL4 in developing larval wing discs.
(A) Schematic representation of the expression pattern of the wing driver (MS1096-GAL4) (marked in green) in the third instar larval wing disc and the adult wing resulting from it. In contrast to the ubiquitously expressed Actin-Gal4 driver (Act-GAL4), the wing specific driver MS1096-GAL4 only targets expression of the indicated genes in the larval wing pouch, which gives rise to the wing in the adult fly. Defects induced during development of the larval wing disc by the expression of various RNAis will lead to abnormalities in the adult wing, which are straightforward to score (see Figure 5).
(B) Validation of the wing driver MS1096-GAL4 by expressing GFP under control of the driver. The wing pouch area tested positive for GFP. These wing discs were also tested for dPANK/Fbl (visualized by staining with an dPANK/Fbl antibody), which is visible in the entire wing disc in B. The nuclear marker DAPI was used to visualize the overall structure of the wing disc.
(C-E) Drosophila genetics allowed co-expression of GFP (to highlight the area) together with the dPANK/fblor mtacp-RNAi constructs in the wing pouch. The downregulation of the gene products (dPANK/Fbl or mtACP) in the larval wing disc was determined using available antibodies (in grey/magenta). DAPI was used to visualize the structure of the complete wing disc (grey), while GFP marks the area in which the RNAi constructs are expressed (grey/green). In (C/C') levels of dPANK/Fbl protein are decreased in the wing pouch, in (E/E') levels of mtACP. This demonstrates the specificity of the wing pouch driver, the effectivity of RNAi constructs, and the specificity of the used antibodies. The boxed areas (C-E) are enlarged in the adjacent images to show the decrease/absence of the indicated proteins in the GFP positive areas in more detail. dPPCDC/ppcdc RNAi , for which no antibody was available, was validated by qPCR in Appendix Figure  S2. Relative hPANK1 mRNA levels (normalized to UbiqC and B2M) Relative hPANK2 mRNA levels (normalized to UbiqC and B2M) Relative hPANK3 mRNA levels (normalized to UbiqC and B2M) Relative hPANK4 mRNA levels (normalized to UbiqC and B2M)  Lysates were generated to determine the specificity of various commercially available antibodies. For this three different lysates were used: lysate 1: HEK293 cells in which human PANK2 was overexpressed (indicated as hPANK2 o/e); lysate 2: control HEK293 cells which endogenously express human PANK2; lysate 3: HEK293 cells in which human PANK2 is downregulated by doxycycline inducible hPANK2 RNAi expression. All three lysates were used for Western blot analysis using 5 antibodies: 1: commercially available Sigma HPA 008440; 2: commercially available Sigma HPA 008025m1; 3: human PANK2 antibody obtained as a gift from Prof. Kotzbauer; 4: commercially available antibody Origene TA 501321; 5: human PANK2 antibody obtained as a gift from Prof. S. Jackowski. In this experiment, the lane 1, containing overexpression of human PANK2 serves as the positive control and lane 3, containing samples in which human PANK2 is downregulated by RNAi serves as the negative control. The two Sigma antibodies and the antibody from Prof. Kotzbauer do recognize human PANK2 when it is overexpressed, however, they do not detect endogenous human PANK2. The antibody from Origene and the antibody from Prof. Jackowski do recognize endogenous human PANK2 and a specific signal is obtained migrating at the same mobility (around 48 kDa, (Kotzbauer et al., 2005) as the protein detected by overexpression of human PANK2. This specific band, indicated by a *, is absent in the PANK2 downregulated cells, demonstrating both the effectivity of the human PANK2 downregulation and the specificity of the antibodies for endogenous human PANK2. In addition to the specific human PANK2 signal, the antibody from Origene also detects a background band, which is not detected by the antibody from Prof. Jackowski. This band is therefore considered a background band, and is indicated with a #. Two exposure times are presented, a long and short exposure time and GAPDH was used as a loading control. For the human PANK2 Western blot in Figure 7, the antibody from Origene was used.

Stocks created for individual experiments, using the lines listed above:
-UAS-dPANK/fbl-RNAi (30B only, lines "F10" and "F20" established through mitotic recombination). For all experiments shown in this paper "F20" was used.