Abstract
Neurodegeneration with brain iron accumulation (NBIA) comprises a group of neurodegenerative disorders characterized by high brain content of iron and presence of axonal spheroids. Mutations in the PANK2 gene, which encodes pantothenate kinase 2, underlie an autosomal recessive inborn error of coenzyme A metabolism, called pantothenate kinase-associated neurodegeneration (PKAN). PKAN is characterized by dystonia, dysarthria, rigidity and pigmentary retinal degeneration. The pathogenesis of this disorder is poorly understood and, although PANK2 is a mitochondrial protein, perturbations in mitochondrial bioenergetics have not been reported. A knock-out (KO) mouse model of PKAN exhibits retinal degeneration and azoospermia, but lacks any neurological phenotype. The absence of a clinical phenotype has partially been explained by the different cellular localization of the human and murine PANK2 proteins. Here we demonstrate that the mouse Pank2 protein localizes to mitochondria, similar to its human orthologue. Moreover, we show that Pank2-defective neurons derived from KO mice have an altered mitochondrial membrane potential, a defect further corroborated by the observations of swollen mitochondria at the ultra-structural level and by the presence of defective respiration.
INTRODUCTION
Pantothenate kinase-associated neurodegeneration (PKAN) is the main disorder of a nosological family termed neurodegeneration with brain iron accumulation (NBIA), in which accumulated iron can be visualized by radiological and histopathological examination of the brain. PKAN is an autosomal-recessive disorder characterized by progressive motor impairment, associated with profound dystonia due to mutations in the PANK2 gene. This gene codes for a mitochondrial enzyme involved in the first regulatory step of coenzyme A (CoA). CoA is synthesized from vitamin B5, or pantothenate, and plays key roles in basic cellular functions such as fatty acid metabolism, Krebs cycle and amino acid synthesis. Pantothenate is taken up by endothelial cells via a sodium-dependent multivitamin transporter and then passes to the blood for delivery to the rest of the body (1). Pantothenate is phosphorylated by pantothenate kinase, conjugated to cysteine, decarboxylated, conjugated to an adenosyl group and phosphorylated again to form CoA. These enzymatic activities have been detected in the cytosol and many in the mitochondrion, as well. An exclusively mitochondrial CoA synthetic pathway has been proposed (2,3), although the cysteine ligase and decarboxylation activities have not been detected explicitly in the mitochondrion. PANK1, PANK2, PANK3 and PANK4 are four known isoforms of pantothenate kinase. Probably PANK4 is not functional as pantothenate kinase, but PANK1 and PANK3 are active in the cytosol, while PANK2 is localized to and active in the mitochondrion (4).
The study of PANK2 function is complex, and efforts to generate animal models of disease by knocking out the gene in flies and mice have generated incomplete phenotypes, lacking signs of neurodegeneration and/or of brain iron accumulation (5,6). A PKAN model of Drosophila has a brain phenotype characterized by the formation of vacuoles, absence of iron accumulation, and Drosophila pantothenate kinase isoforms do not strictly parallel those of humans (5). In 2005, Pank2 null mice were generated (7), which showed growth reduction, retinal degeneration and male infertility due to azoospermia but no movement disorder or brain iron accumulation, even after 18 months of age. In contrast, a pantothenic acid-deficient diet was able to elicit a movement disorder and azoospermia in mice without evidence of iron accumulation in brain (8).
These results have partially been explained by the sub-cellular localization of the murine Pank2 protein, which was reported to be predominately mitochondrial (9) or cytoplasmic (10) by different groups. We performed experiments aimed at demonstrating: (i) the sub-cellular localization of the murine Pank2 protein and (ii) the presence of alterations in the function, regulation and structure of mitochondria in the available KO mouse model. Our studies have demonstrated that Pank2 KO mice have mitochondrial dysfunction.
RESULTS
Mitochondrial localization of mouse Pank2 protein
Human and mouse PANK2 proteins show an identity of 90%, although the mouse polypeptide does not have an N-terminal extension, which is present in human PANK2. Software tools predicting mitochondrial localization of proteins scored high for the murine Pank2: Mitoprot (http://ihg.gsf.de/ihg/mitoprot.html) gave a probability of 98%; Predotar (http://urgi.versailles.inra.fr/predotar/predotar.html) of 72%, and TargetP (http://www.cbs.dtu.dk/services/TargetP/) of 88%. To experimentally verify the sub-cellular localization of the murine protein, we performed western blot analysis using a commercially available antibody (see Material and Methods) on total homogenate, cytosol and mitochondria isolated from brain and fibroblasts of wild-type (WT; Pank2+/+) and knock-out (KO; Pank2−/−) mice. We first tested specificity of the antibody by verifying its cross-reactivity with the Pank2 in vitro translation product (not shown).
In the homogenate and cytosol derived from Pank2+/+ but not from Pank2−/− brain, the antibody detected a faint band of around 47 kDa, corresponding to the predicted molecular weight of the murine Pank2 protein (Fig. 1A). A band of higher intensity was present in mitochondria isolated from Pank2+/+ mice, while it was completely absent in mitochondria derived from Pank2−/− mice. These data indicated that the Pank2 protein, which was detectable and enriched in the mitochondrial fraction of wild-type mice, was missing in Pank2−/− mice. To further determine in which sub-mitochondrial compartment the Pank2 protein was present, we further fractionated mitochondria derived from WT mouse brain, into membranes, inter-membrane space, and matrix, and we isolated mitoplasts. As shown in Figure 1B, western blot analysis demonstrated that the mouse Pank2 protein was present into the mitochondrial inter-membrane space and mitoplasts. However, by exposing mitoplasts to the action of proteinase K (PK), the signal corresponding to the Pank2 protein disappeared. On the contrary, signals corresponding to proteins present in the mitochondrial matrix such as Ethe1, or in the inner membrane such as Ndufa9 remained protected from PK degradation. We obtained the same results by fractionating mitochondria derived from mouse fibroblasts (Fig. 1C). Altogether these data indicated that the mouse Pank2 protein could be mainly located in the mitochondrial inter-membrane space. In the mouse brain but not in fibroblasts, the protein was also barely detectable in the mitochondrial matrix.
Mitochondrial localization of murine Pank2 protein. (A) Western blot analysis of different sub-cellular fractions derived from mouse brain: +/+ indicates Pank2+/+ animals; −/− indicates Pank2−/− animals. The same filter was incubated with anti-PANK2, anti-NDUFA9 and anti-SDH70 antibodies. (B) Western blot analysis on different mitochondrial compartments isolated from WT mouse brain. (C) Western blot analysis on different mitochondrial compartments isolated from WT mouse fibroblasts. In (B and C) the filters were sequentially incubated with anti-PANK2, anti-NDUFA9, anti-SDH70 and anti-ETHE1 antibodies.
Mitochondrial bioenergetics evaluation
In order to investigate mitochondrial bioenergetic status, we performed spectrophotometric assays to measure the biochemical activity of each single respiratory chain complex in different tissue homogenates (muscle, brain and liver) derived from Pank2+/+ and Pank2−/− mice. No alterations in the specific activities of complex I, II, III, IV and V were observed (data not shown).
We then evaluated respiration with microscale oxygraphy on mitochondria isolated from mouse brain. This system measures global oxygen consumption rate (OCR) and extracellular acidification rate in intact cells and isolated mitochondria.
According to Brand and Nicholls (11), and our group (12) in intact cells, and by Rogers et al. (13) in isolated mitochondria, the direct readout of cellular respiration can be used to define relevant bioenergetic parameters, including (a) basal OCR (OCR-B), after ADP addition (OCR-ADP), after oligomycin addition (OCR-O).
We performed oxygen consumption measurement in three different Pank2+/+ and Pank2−/− mice of 4 weeks of age. In each experiment, we consistently observed reductions of OCR-B, -ADP, –O, in mitochondria isolated from Pank2−/− brain when compared with Pank2+/+ (Fig. 2A).
Evaluation of mitochondrial bioenergetics status. (A) OCR in Pank2+/+ and Pank2−/− mouse brain mitochondria. Black, white and grey histograms indicate OCR-B, -ADP and -O, respectively. Bars indicate the standard deviation (SD). **P < 0.01 (unpaired, two-tail Student's t-test). (B) Analysis of OCR-B, -ADP, -O and -F. The z-score values (dots) for OCR-B, -ADP, -O and -F are shown for Pank2+/+ versus Pank2−/−. Data are expressed as the means (dots) and 95% confidence intervals, Confidence interval (bars). Student's t-test for Pank2+/+ versus Pank2−/−: diamond, P = 1.0E−07; square, P = 1.3E−04; triangle, P = 1.0E−07; asterisk, P = 1.2E−09. Three Pank2+/+ and three Pank2−/− mice of 4-week-old were analysed.
In another set of experiments, we also added FCCP (OCR-F) to measure uncoupler-stimulated respiration.
To compare the numerical results of the different experiments, individual values were used to obtain z-scores of OCR-B, -ADP, -O, -F (12).
For statistical analysis, we compared the z-scores obtained in the Pank2+/+ versus Pank2−/− mitochondria (Fig. 2B). The values obtained in the WT mitochondria were consistently and significantly higher than those obtained in Pank2−/− mitochondria, for each of the mitochondrial respiratory conditions.
These differences were statistically significant as demonstrated by an unpaired, two-sided Student's t-test, assuming unequal variance. Values for statistical significance were set at P < 0.05.
We also evaluated the ATP content in cells derived from three Pank2+/+ and three Pank2−/− mice.
We observed a reduced level of ATP in Pank2−/− derived cells when compared with Pank2+/+ (Supplementary Material, Fig. S1A), suggesting a lower capacity to produce ATP in the absence of Pank2 protein. These data are in agreement with the observation of a global reduction of the respiratory capacity of intact mitochondria derived from Pank2−/− brains.
Parameters of oxidative stress
We measured the levels of protein carbonyl groups, a marker of protein oxidation, by OxyBlot assay in the brain of three Pank2+/+ and three Pank2−/− mice of 6 months of age. The total level of carbonyls is similar between the two genotypic groups (Supplementary Material, Fig. S1B), although a certain level of variability was noticed. We also performed western blot analysis on the same samples with antibodies specific to superoxide dismutase 1 and Catalase, but again no significant differences were identified (not shown).
Mitochondrial morphology and integrity in peripheral and central nervous systems
Since PKAN affects the central nervous system (CNS) in humans, we sought to examine CNS neurons in mouse for evidence of perturbed mitochondrial structure or function. We derived neural precursors from hair bulge whisker (Supplementary Material, Fig. S2A) and sciatic nerve (Supplementary Material, Fig. S2B) of Pank2−/− and Pank2+/+ mice and subsequently differentiated these cells into neurons (see Materials and Methods). Cells showed typical morphology of neural crest precursor (Supplementary Material, Fig. S2A), stained positive for neural crest marker Nestin (Supplementary Material, Fig. S2C and D) and were negative for glial fibrillary acidic protein staining (Supplementary Material, Fig. S2E and F). After induction with suitable culture medium, differentiated neurons (Supplementary Material, Fig. S2G and H) expressing the neuron-specific marker β-III-tubulin (Supplementary Material, Fig. S2I–L) were obtained. The Mitotracker red, a mitochondrion specific dye, was used to stain both hair bulge- and sciatic nerve-derived neurons obtained from Pank2+/+ and Pank2−/− mice. We observed that the dye was exclusively localized into mitochondria in Pank2+/+ neurons (Fig. 3A–C) while it was diffused throughout the cytoplasm and nucleus in Pank2−/− neurons (Fig. 3B–D).
Mitochondrial morphology evaluated with the Mitotracker-red. (A) Neurons derived from hair bulge from Pank2+/+ mice. (B) Neurons derived from hair bulge from Pank2−/− mice. (C) Neurons derived from sciatic nerve of Pank2+/+ mice. (D) Neurons derived from sciatic nerve of Pank2−/− mice. (E) Control fibroblasts. (F) Control fibroblasts treated with valinomycin. (G) Histogram reporting the quantification of Mitotracker diffusion in Pank2+/+ and Pank2−/− neurons using the RawIntDen parameter (see Materials and Methods). Bars indicated SD. *P = 3.37E−23 (unpaired, two-tail Student's t-test).
This difference was quantified by selecting a defined area corresponding to nucleus of neurons and by calculating the imageJ parameter ‘RawIntDen’ (that is the sum of the values of the pixels in the selected area) in WT and KO neurons. This parameter resulted in a 2.5-fold greater value in Pank2−/− neurons when compared with Pank2+/+ neurons (Fig. 3G).
We reasoned that the type of Mitotracker used for the staining was also able to detect a qualitative defect in the mitochondrial membrane potential (ΔΨm) (14).
To verify these data, we treated control fibroblasts with the uncoupling agent valinomycin to induce the ΔΨm loss. We observed that the Mitotracker stains mitochondria in untreated cells (Fig. 3E), while it spreads in the cytoplasm in fibroblasts exposed to valinomycin (Fig. 3F).
To further corroborate these data we used 5,5′,6,6′-tetrachloro-1,1′,3,3′ tetraethylbenzimidazolylcarbocyanine iodide (JC1) staining, which specifically detects variation in the mitochondrial membrane potential. The JC1 dye undergoes a reversible change in fluorescence emission from green to red as mitochondrial membrane potential increases. Cells with high membrane potential promote the formation of red fluorescent aggregates, while cells with low membrane potential show a diffuse green fluorescence. As shown in Figure 4A, C, E, WT neurons of different derivation presented red fluorescent aggregates, while KO neurons presented predominantly with a diffuse green fluorescence (Fig. 4B–F). JC1 specificity was tested by pre-treating WT neurons with valinomycin. As reported in Supplementary Material, Figure S3, treated neurons stain green fluorescent aggregates, whereas untreated neurons showed red fluorescent aggregates.
Membrane potential and electron microscopy in Pank2+/+ and Pank2−/− neurons. (A and B) JC1 staining of neurons derived from hair bulge. (C and D) JC1 staining of neurons derived from sciatic nerve. (E and F) JC1 staining of neonatal neuron derived from hippocampus. Blue fluorescence indicates DAPI staining of nuclei. Green fluorescence indicates an alteration in mitochondrial ΔΨ; red fluorescence indicates the preservation of a normal mitochondrial ΔΨ. (G and H) Electron microscopy of neurons derived from sciatic nerve (semithin section Toluidin Blue). (I–L) Electron microscopy of neurons derived from sciatic nerve (ultrastructure uranyl acetate and lead citrate). Normal appearing mitochondria in Pank2+/+ neurons (I–K) and few normal mitochondria in Pank2−/− neurons (black arrows in J–L); aberrant mitochondria with disrupted matrix and cristae (red arrows in J–L); lipid filled vesicle (asterisk in J).
To corroborate our observations, we performed electron microscopic analysis on neurons in culture, which were stained with JC1. Examination of semi-thin sections revealed the presence of enlarged neurons with severe alteration in the cytoplasm, containing abundant micro-vesicular structures, mainly if not exclusively in Pank2−/− cells (Fig. 4H). Ultra-structural investigation confirmed that, while the cytoplasm of Pank2+/+ neurons exclusively showed endoplasmic reticulum, vesicles and mitochondria with regular morphology (Fig. 4I–K), Pank2−/− neurons contained lipid droplets, and several swollen mitochondria with aberrant cristae and a complete alteration of the matrix structure (Fig. 4J–L), together with morphologically normal mitochondria.
We asked whether the same alteration observed in the peripheral nerve and neurons was also present in the CNS of Pank2−/− mice. To this aim, we derived neonatal hippocampal neonatal neurons and we performed the JC1 staining experiment. As reported in Figure 4E, a red fluorescent signal was detected in neurons derived from Pank2+/+ mice, whereas a green fluorescent signal (Fig. 4F) was exclusively present in neurons derived from Pank2−/− animals.
Altogether, these data demonstrated that aberrant mitochondria were present not only in neurons derived from the sciatic nerve of adult animals, but also in neonatal hippocampal neurons.
Analysis of peripheral and CNSs
We analysed the sciatic nerve obtained from groups of three Pank2+/+ and three Pank2−/− animals of 24 and 48 weeks of age, respectively.
Examination of semithin sections, counterstained with Toluidine Blue, revealed no obvious differences in myelinated fibre density, in the absence of demyelinating features. However, Pank2−/− nerve showed a mild reduction in large-caliber fibre and some more prominent wallerian-like degenerations compared with controls (Fig. 5A and B). Overall, no clear-cut sign of progressive neuropathy was present, since no reduction in myelinated fibres or increase in the number of degenerating fibres has been detected (Fig. 5C and D).
Evaluation of neurodegeneration of peripheral nervous systems with optic and electron microscopy. Sciatic nerve semithin section (Toluidin Blue) of (A) 24-week-old and (C) 48-week-old Pank2+/+ mice; (B) 24-week-old and (D) 48-week-old Pank2−/− mice. Fibre density is not clearly different, whereas the axonal diameter seems to be reduced in Pank2−/−. Electron micrographs of 48-week-old mice sciatic nerve in Pank2+/+ mice (E) and Pank2−/− mice (F). No sign of un-myelinated fibre loss in the form of collagen pockets or denervated Schwann cells were evident; in Pank2−/− mice, some enlarged mitochondria were detected (arrows in F).
Ultra-structural analysis of sciatic nerve of Pank2−/− animals confirmed the presence of occasional degenerating fibres, and a normal representation of un-myelinated fibres (Fig. 5E and F). However, a substantial number of myelinated fibres (Fig. 6B–D), and occasionally un-myelinated axons and Schwann cells, contained (Supplementary Material, Fig. S4B–D) swollen mitochondria, characterized by alteration of cristae and by the presence of multivescicolar bodies or amorphous material, in the matrix (Fig. 6A–C and Supplementary Material, Fig. S4A–C). Aberrant mitochondria appear to increase with age since in 48 week Pank2−/− old mice versus 24 week old mice (Fig. 6B) a greater number of swollen mitochondria with a profound alteration of the matrix were detected (Fig. 6D).
Electron microscopy of peripheral and CNS of 24- and 48-week-old animals. (A–C) Sciatic nerve of Pank2+/+ mice show normal mitochondria (arrows). (B–D) Sciatic nerve of Pank2−/− mice show swollen mitochondria with altered cristae (arrows). (E–G) Basal ganglia derived from 48-week-old Pank2+/+ mice show normally shaped mitochondria. (F–H) Basal ganglia derived from 48-week-old Pank2−/− mice show altered cristae and enlarged cisternae.
We also analysed the basal ganglia of 48-week-old mice. Electron microscopy demonstrated morphologically abnormal mitochondrial cristae with enlarged cisternae, which were predominant in the Pank2−/− (Fig. 6F–H) when compared with Pank2+/+ (Fig. 6E–G).
These data indicate the presence of defective mitochondria in both the peripheral and CNSs in mice with defective Pank2.
DISCUSSION
We have demonstrated that the murine Pank2 protein is mainly located in the mitochondrial inter-membrane space, which is compatible with its role in CoA metabolism and with its regulation in the presence of CoA at concentrations found in mitochondrial matrix (10). Further investigation using, for instance, immuno-electronmicroscopy is needed to unequivocally establish this sub-mitochondrial localization.
This observation makes the Pank2−/− mouse model a promising system to study pathophysiology of human PKAN. In fact, although the mouse model does not recapitulate the clinical and neuropathological features of the human condition (7), it could serve as a system in which to interrogate the basic defect in mitochondrial function.
We did not detect the deficiency in the enzymatic activity of any mitochondrial respiratory chain complex in different tissues of Pank2−/− mice. However, recent observations indicate that mitochondrial respiration is largely carried out by the fraction of respiratory chain complexes that assemble together in supercomplexes forming functionally active units, called respirasomes (15,16). Thus, the spectrophotometric measurement of each respiratory complex activity does not necessarily reflect the mitochondrial functional capacity in vivo. Microscale oxygraphy could overcome this limitation and in fact, using this approach, we detected a modification in the respiratory profile of mitochondria derived from Pank2−/− when compared with Pank2+/+ brains. Highly significant differences of OCR-B, -ADP, -O and –F were obtained by statistical analysis of either the single values or as overall values of Pank2+/+ versus Pank2−/− mitochondria, expressed as z-scores (Fig. 2). Together, these results indicate that the absence of the Pank2 protein leads to a global failure of the mitochondrial bioenergetic performance without affecting the function of any single respiratory chain complex.
An interesting result of our study, strictly correlated with the absence of Pank2 function, was the demonstration of alteration in mitochondrial membrane potential in neurons derived from sciatic nerve and hair bulge stem cells of adult mice. In addition, the same alteration was also present in neonatal hippocampal neurons, suggesting the presence of defective mitochondria in Pank2−/− mice since birth.
These results were confirmed by electron microscopy analysis on cultured neurons derived from Pank2−/− mice, in which aberrant mitochondria with remodelled cristae were present. Moreover, peripheral and CNSs examination of Pank2−/− mice showed the presence of swollen mitochondria with amorphous electron-dense inclusions and dysmorphic cristae.
Recently, the characterization of a PKAN Drosophila model demonstrated that impaired function of pantothenate kinase induced a neurodegenerative phenotype with mitochondrial dysfunction, decreased levels of CoA, increased protein oxidation and reduced lifespan (17).
JC1 staining and electron microscopic analysis revealed that, in contrast to WT flies, mitochondria of dPANK/fbl mutants showed an alteration in the transmembrane potential, were swollen and presented with altered cristae and ruptured membranes (18). These observations are in agreement with our results in the nervous system of Pank2−/− mice in which mitochondria are severely damaged.
In search for mechanisms underlying the alteration of mitochondria, we found a reduced level of ATP in Pank2−/− mice. Interestingly, we found that state 3 activity, which represent the maximum respiration rate in the presence of ADP, is significantly decreased in mitochondria isolated from Pank2−/− mice. It is tempting to speculate that the reduction in the respiration could be attributable to the severe alteration of the cristae structure, which could prevent the respirasomes from remaining functionally active.
We also investigated the presence of oxidative damage in brains, but we did not find significant differences in oxidative damage at least in 6-month-old Pank2−/− mice. As demonstrated by other studies, mitochondrial membrane damage contributes to the pathogenesis of many neurodegenerative diseases (19,20).
Interestingly enough, a recent investigation of a KO mouse model for the Pla2g6 gene, which is defective in a different but related form of NBIA, revealed the presence of collapsed mitochondria with degenerated inner membranes (21). Pla2g6 gene encodes a group VIA calcium-independent phospholipase A2, an esterase that hydrolyzes the sn-2 ester bond in phospholipids to yield free fatty acids and lysophospholipids and which is involved in cardiolipin remodelling (22). Analysis of phospholipids and fatty acids revealed differences between KO and WT mice. The vulnerability of mitochondrial inner membranes in Pla2g6 KO mice might be attributable to increased production of reactive oxygen species (ROS) (23) and a rich content of polyunsaturated fatty acids that can readily be peroxidized, such as linoleic acid in cardiolipin (22).
Alteration in cholesterol and lipid metabolism was also recently demonstrated by a metabolomics investigation in a group of PKAN patients (24). In this case, however, the alteration was mainly due to a defective synthesis and not, as in the case of PLA2G6, to the absence of a catabolic enzyme.
Irrespective of the anabolic (PANK2) or catabolic (PLA2G6) role of these two mitochondrial proteins, a common culprit in the pathogenesis of both neurodegenerative diseases could be an altered lipid metabolism (25). Further characterization of the lipid profile in Pank2−/− mice is in progress to understand if and how alteration of this metabolic pathway could be responsible for the observed mitochondrial membranes modifications.
We demonstrated insufficient energy production and severe mitochondrial dysfunction in the brain and peripheral nerve of Pank2−/− mice, without increased oxidative stress or signs of neurodegeneration. It is possible that mice can better tolerate alterations in bioenergetics metabolism without suffering any overt clinical manifestations because of the presence of compensatory mechanisms. Few examples of mouse models of mitochondrial disorders, which display a biochemical phenotype but do not present any clinical signs typical of the human pathology, are described in the literature (26,27).
Although we clearly demonstrated mitochondrial dysfunction in Pank2−/− mice, we remain uncertain why neither neurological signs typical of PKAN nor iron accumulation occur in mice. However, our findings suggest that there would be value in investigating bioenergetic competence and ultra-structural abnormalities of mitochondria in PKAN patients.
MATERIALS AND METHODS
Animals and diet
The use and care of animals followed the Italian Law D.L. 116/1992 and the EU directive 86/609/CEE. Animal studies were approved in accordance with guidelines of the Italian Ministry of Health: Project n. BT4/2011.
The JM129/SvJ–C57BL/6 Pank2+/− mice used in this study were the same as those generated in 2005 (7).
Animals were housed two or three per cage in a temperature-controlled (21°C) room with a 12 h light–dark cycle and ∼60% relative humidity. Standard diet (DS, Mucedola, Italy) and water were given ad libitum.
Mitochondria isolation
Standard methods were used for the preparation of mitochondrial and post-mitochondrial fractions from mouse brain (28).
Brain was extracted and rinsed using cold homogenization medium AT [0.075 m sucrose, 0.225 m mannitol, 1 mm ethylene glycol tetraacetic acid (EGTA), 0.01% bovine serum albumin (BSA), pH 7.4]. The organ weighted was cut into small pieces and washed to remove blood and connective tissue. The brain was homogenized with 10–15 strokes using a Dounce-type glass homogenizer with a manually driven glass pestle, adding 5 ml of homogenization medium AT per gram of the starting material. The tissue was centrifuged at low speed (1000g for 5 min at 4°C) and the resulting supernatant was transferred to a clean tube and spin at high speed (9000g for 10 min at 4°C). The supernatant obtained from this centrifugation contained cytosol. To clean the mitochondrial pellet obtained, we resuspended it with 5 ml of medium AT, transferred to 1.5 ml Eppendorf tubes and washed to obtain a single pellet. The washed mitochondria were resuspended in the appropriate volume of MAITE medium (25 mm sucrose, 75 mm sorbitol, 100 mm KCl, 0.05 mm EDTA, 5 mm MgCl2, 10 mm Tris–HCl, 10 mm H3PO4, pH 7.4). This fraction contained free mitochondria and synaptosomes.
Mitochondria fractionation
Isolated mitochondria were resuspended in 100 μl of potassium phosphate buffer [(PP) buffer, 20 mm, pH 7.8, KCl 150 mm] and sonicated 10 s for three times at 10 Amp. The suspension was centrifuged at 164000g for 30 min at 4°C. Supernatant (mitochondrial matrix and inter-membrane space) was collected, and pellet (mitochondrial membranes) was resuspended in 100 μl of PP buffer.
Isolation and fractionation of mitoplasts
Isolated mitochondria were resuspended in 100 μl of homogenisation medium A. After quantification of protein with Bradford method (Biorad), the sample was incubated for 5 min at 4°C with Digitonin (Fluka) at a final concentration of 0.18 mg for mg of protein. Digitonin was then diluted with 10 volumes of medium A and the sample was centrifuged at 12 000g for 10 min at 4°C. Supernatant (inter-membrane space) was collected, and pellet (mitoplasts) was resuspended in 100 μl of PP buffer.
Mitoplasts were then sonicated 10 s for three times at 10 Amp. The suspension was centrifuged at 164000g for 30 min at 4°C. Supernatant (mitochondrial matrix) was collected and pellet (mitochondrial inner membrane) was resuspended in 100 μl of PP buffer.
Western blot analysis
Thirty micrograms of proteins were used for each sample in denaturing sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE). Western blot analysis was performed as described (29), using the ECL-chemiluminescence kit (Amersham).
Antibodies
The following mouse monoclonal antibodies were used: anti-PANK2 (Origene) at 1:2000 dilution, anti-NADH dehydrogenase ubiquinone 1 alpha subcomplex subunit 9 (NDUFA9) (Invitrogen) at final concentration of 0.5 µg/ml and anti-Rieske (Invitrogen) at a final concentration of 2 µg/ml, anti-SDH-70 (succinate dehydrogenase, subunit 70 kDa) (Invitrogen) at 1:5000 dilution. An anti-ethylmalonic encephalopathy 1 rabbit polyclonal antibody was used at 1:2000 dilution (30). Secondary anti-rabbit and anti-mouse antibodies were used at 1:10 000 and 1:7000 dilutions, respectively.
ATP and oxidative stress assays
ATPLite kit (PerkinElmer Life Sciences) was used to measure ATP levels according to manufacturer's instructions. This method is based on mono-oxygenation of luciferin, catalyzed by luciferase in the presence of Mg2+, ATP and oxygen, resulting in a luminescent signal that is proportional to the ATP concentration.
Oxidized proteins were detected using the Oxyblot Protein Oxidation Detection Kit (Millipore) following the manufacturer's instructions. Brain extracts were homogenized in 20 mm Tris–HCl, pH 7, 50 mm dithiothreitol in the presence of anti-protease inhibitor cocktail tablet (Roche) and then derivatized to 2,4-dinitrophenylhydrazone. One microgram of proteins was loaded on 12% SDS–PAGE, blotted and incubated with an anti-DPN antibody. The bound activity was revealed by ECL Advance kit (Amersham).
Mitochondria membrane potential
Mitochondrial morphology was assessed after cell staining with 10 nm Mitotracker CMX-Red (Invitrogen) for 30 min at 37°C. Fluorescence was visualized with a digital imaging system using an inverted epifluorescence microscope with ×63/1.4 oil objective (Nikon, Japan). Images were captured with a back-illuminated Photometrics Cascade CCD camera system (Crisel) and analysed with Metamorph acquisition/analysis software. Different diffusion of the mitotraker in the neuron cytoplasm was quantified, calculating the parameter ‘RawIntdens’ of ImageJ software (http://rsbweb.nih.gov/ij/) that is the sum of the values of the pixels in selected areas. These data were calculated in 66 neurons (33 KO versus 33 WT) in three different experiments.
Detection of mitochondrial potential change was performed using JC1 (Sigma, CS0390) staining kit according to manufacturer' instructions.
Evaluation of bioenergetic profile
Measurements of oxygen consumption were performed on isolated mitochondria derived from mouse brain using an XF96 Extracellular Flux Analyser (Seahorse Bioscience).
To rapidly isolate mitochondria, brains were obtained from sacrificed mice and washed three times in ice-cold phosphate-buffered saline (PBS), than homogenized with 15 strokes in a glass-dounce homogenizer with 2 ml of hepes, EDTA, sucrose (HES)–0.2% BSA (HEPES 5 mm, EDTA 1 mm, sucrose 250 mm, BSA 0.2%, pH 7.4).
The homogenate was centrifuged for 10 min 2000 rpm (700g) at 4°C, and the supernatant was collected in a new tube, the pellet was resuspended in HES–0.2% BSA and centrifuged for 10 min 2000 rpm at 4°C.
The supernatant was collected and centrifuged for 15 min at 9000g and the supernatant was discarded; the pellet was resuspended in HES–0.2% BSA and centrifuged for 15 min 10 000 rpm and the supernatant was discarded.
The pellet containing mitochondria was resuspended in HES–0.2% BSA and total protein (mg/ml) was determined using Bradford Assay reagent (Bio-Rad).
As indicated by Rogers et al. (13), the WT and KO brain mitochondria were seeded in an XF 96-well cell culture microplate (Seahorse Bioscience) at a protein concentration of 8 µg/well in 50 µl of MAS-1 1× (70 mm sucrose, 220 mm mannitol, 10 mm KH2PO4, 5 mm MgCl2, 2 mm HEPES, 1 mm EGTA and 0.2% (w/v) fatty acid-free BSA, pH 7.2) supplemented with succinate, malate, glutamate and pyruvate 5 mm. The plate was centrifuged 2000g for 5 min at 4°C, 130 µl of succinate, malate, glutamate and pyruvate 5 mm were added and incubated at 37°C without CO2 for 30 min before starting the assay.
OCR was measured under basal condition (OCR-B), and after sequentially adding to each well 20 µl of ADP (OCR-ADP) and 22 µl of oligomycin (OCR-OL) to reach working concentrations of 4 mm and 2.5 µm respectively. OCR was also measured after addition of carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP), to evaluate the maximal respiratory capacity.
The detailed protocol was as follows:
1′ Mixing
3′ Waiting
1′ Mixing
3′ Waiting
Basal OCR:
30″ Mixing
3′ Recording
1′ Mixing
3′ Recording
30″ Mixing
ADP addition through port A
1′ Mixing
3′ Recording
1′ Mixing
3′ Recording
1′ Mixing
Oligomycin addition through port B
30″ Mixing
3′ Recording
1′ Mixing
3′ Recording
1′ Mixing
3′ Recording
Statistics
For individual experiments, data obtained from Pank2+/+ and Pank2−/− mitochondria were calculated as the mean of the replicates ± standard deviation (SD) and then compared with unpaired two-tailed Student's t-test. Although the measurements obtained from the replicates for each sample were relatively consistent in individual experiments, the values may vary in different experiments performed in different days. For this reason, we transformed each value of OCR-B, OCR-ADP, OCR-O and OCR-F into standard (z) scores, in order to make the data of three different experiments comparable with each other and perform statistical analysis to the entire collection of experimental values. In practice, for each sample, the overall mean was subtracted from each measure, and the result was divided by the overall SD of the experiment. The resulting values indicate how many SD's separate a given observation from the reference value of the experiment.
Transformed data were then compared across samples and across experiments by unpaired two-tailed Student's t-test. Results with P < 0.05 were considered statistically significant.
Electron microscopy analysis
Sciatic nerves samples from 24- and 48-week-old Pank2+/+ and Pank2−/− mice were surgically removed and fixed in glutaraledehyde 2.5% in phosphate buffer (pH = 7.4), post-fixed in aqueous solution of osmium tetroxide, dehydrated in acetone and embedded in epoxy resin (SPURR, EM Sciences). One-millimetre-thick sections were stained with Toluidine Blue.
For ultra-structural analysis, sections of 90 nm were collected on 200-mesh copper grids, positively stained with uranyl acetate and lead citrate, and then analysed by transmission electron microscope (EM109 Zeiss) to evaluate the presence of possible alterations in axonal morphology, cell membranes and mitochondria.
Sagittally spliced brains from 72-week-old Pank2+/+ and Pank2−/− mice were fixed by immersion in 4% paraformaldehyde–PBS. After fixation, serial sagittal slides of ∼1 mm thick were obtained. Selected areas of interest were sampled, post-fixed in osmium tetroxide and embedded in Epon epoxy resin. Thin sections (80–90 nm) were stained with uracyl acetate and lead citrate and examined with a CM10 Philips electron microscope.
Neurons derived from sciatic nerve were collected by gentle scraping and fixed in glutaraledehyde 2.5% in phosphate buffer (pH = 7.4). Cells pellet was treated for ultra-structural analyses as described above.
SUPPLEMENTARY MATERIAL
Supplementary Material is available at HMG online.
FUNDING
The financial support of Mariani Foundation of Milan (grant no. R-10-84 to V.T.) and Telethon GGP 11088 to V.T. and S.L. is gratefully acknowledged. S.H. and V.T. participate in the TIRCON consortium funded by the European Commission Seventh Framework Program (FP7/2007-2013, HEALTH-F2-2011) under Grant Agreement No. 277984. Funding to pay the Open Access publication charges for this article was provided by Fondazione Telethon, Italy.
ACKNOWLEDGEMENTS
We would like to thank Sabrina Ravaglia, MD, University of Pavia, for helpful discussion about statistical analysis of z-scores and SIlvia Capra, Neurological Institute C. Besta, for technical support.
Conflict of Interest statement. None declared.
