Disruption of the mitochondrial-nuclear network leads to accelerated aging and age-related diseases, including age-related macular degeneration. The current study tested the hypothesis that mitochondrial morphology could be demonstrated quantitatively using a mathematic model and mitochondrial trafficking complex under stress conditions. To test our hypothesis, normal and aberrant mitochondria were examined quantitatively based on mitochondrial size, shape, position, composition, and dynamics. Adaptation of the mitochondrial network to changes in the intracellular oxidation and reduction milieu is critical for the survival of retinal pigment epithelial cells. Our mitochondrial interactome mapping demonstrated that a positive correlation may exist between oxidative stress-mediated phosphorylation and age-related disease progression. The current interactome may provide a potential therapeutic approach to treat mitochondria-induced neurodegeneration, including age-related macular degeneration.
- mitochondrial trafficking
- retinal pigment epithelial cells
- mitochondrial dynamics
- age-related macular degeneration
- protein interactome
Mitochondria are the highly dynamic cellular organelles that form a dynamic network to regulate calcium balance, energy metabolism, and apoptotic signaling [1, 2, 3, 4, 5, 6, 7, 8]. Mitochondria alter their morphology repeatedly through the collective actions of fission as a separation of a single organelle into multiple autonomous structures, fusion as the combination of multiple structures to form single organelle, as well as movement along cytoskeletal paths [5, 9]. These combined actions occur concurrently in major cell types to regulate the cell fate.
To maintain the balance that regulates overall morphology and cytoskeletal stability many specialized molecules including dynamin family of proteins play critical roles [10, 11]. Growing body of evidence demonstrates that disruptions of mitochondrial network lead to multiple human pathologies, including metabolic, genetic, cardiovascular diseases, as well as neurogenerative diseases and cancers [12, 13, 14, 15]. Several studies provided the evidence that mitochondria play a critical role in the progression of age-related diseases, including age-related macular degeneration (AMD) [16, 17, 18, 19, 20, 21]. The damage of mitochondrial DNA could be the key factor involved in altered vascular endothelial growth factor (VEGF) secretion, retinal pigment epithelium (RPE) dysfunction, and cell death during the progression of AMD [22, 23].
The current study aimed to examine the correlation between alterations in mitochondrial morphology and mitochondrial dysfunction. For quantitative analysis of mitochondrial morphology, we introduced the mitochondrial index that includes network size, mitochondrial content and surface area. Mitochondrial interconnectivity and elongation were determined systematically using a computational model in three dimensions, showing a mitochondrial- endoplasmic reticulum (ER)-nuclear hole as open space for trafficking at the beginning of apoptosis under oxidative stress.
The assessment of average circularity showed mitochondrial elongation which is sensitive to fragmented vs. normal shaped mitochondria. The average area/perimeter ratio showed normal or stressed mitochondria as a highly interconnected mass of reticular network. Previously, we observed that prohibitin translocalizes between the nucleus and mitochondria under oxidative stress conditions to influence mitochondrial dynamics . We observed anterograde signaling from the nucleus to mitochondria using a prohibitin shuttle under stress in the retina, as well as the retrograde shuttling of prohibitin from mitochondria to the nucleus in the RPE. In addition, cytoskeletal reorganization, tubulin/vimentin depolymerization and increased phosphorylations were observed in stressed mitochondria [25, 26, 27].
In this study, mitochondrial dynamics was further analyzed in mitochondrial trafficking complex using prohibitin immunoprecipitation. We found a motor-based protein complex that includes kinesin 19 (93 kDa), myosin 9 (110 kDa), and cadherin isoforms (88 kDa) to regulate the mitochondrial-nuclear communication. Finally, we have established a comprehensive mitochondrial interactome map by combining several independent sets of interaction data. Our interactome map provides an integrated information on the hidden apoptotic pathway, cytoskeletal rearrangement, nitric oxide signaling, ubiquitination, and mitochondrial network in neurodegeneration.
2. Materials and methods
2.1. Cell culture and oxidative stress treatment
Retinal pigment epithelial cells (ARPE-19) were purchased from ATCC (Manassas, VA) and cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with fetal bovine serum (FBS; 10%) and penicillin/streptomycin (1%) at 37°C in a humidified atmosphere of 5% CO2 in air as suggested by the manufacturer. Cells were used between passages 8–9.
Retinal progenitor cells (HRP) were kindly donated by Dr. Harold J. Sheeldo (University of North Texas Health Science Center) and were cultured at the same condition as ARPE-19 cells.
Prior to all experiments, confluent ARPE-19 cells were incubated with fresh medium for 12 h and washed with phosphate buffered saline (PBS) three times. ARPE-19 cells were incubated with an oxidant, tert-butyl hydroperoxide (t-BuOOH, 200 μM, Sigma-Aldrich, St.Louis, MO), in serum-free medium for 0.5, 1, 2, 4, 6, 8, 12, and 24 h and representative images were presented. After the treatment, medium was removed, cells were washed with PBS and harvested for future analysis. Cells were lysed for experiments, including, immunoprecipitation, sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS−PAGE), Western blot analysis, immunocytochemistry, and mass spectrometry analysis.
To compare with other stress environment, ARPE-19 cells were incubated under intense light (210 μmoles/m2/s photon flux; 7000 lx) for 1 h in serum-free media and analyzed by SDS-PAGE/Western blotting, or immunocytochemistry.
Lipids were extracted from ARPE-19 cells using cholorform/methanol (2:1, v/v) and organic solvent was evaporated under a gentle nitrogen stream and dissolved in chloroform for analysis by HPLC and mass spectrometry.
All experiments were repeated (N = 3–10 biological samples) with technical duplicate or triplicate. Statistical analysis was performed using StatView software and statistical significance was determined by variance (ANOVA) or unpaired Student’s t test when appropriate.
2.2. Donor eye tissue and phosphoprotein enrichment
Human postmortem donor eye tissues were used following the tenets of the Declaration of Helsinki. Human AMD retinas (8 mm macular and peripheral punches), RPE (8 mm central and peripheral punches), and age-matched control eyes (N = 9, biological triplicate x technical triplicate) were provided by the Lions Eye Bank (Moran Eye Center, University of Utah). Phosphoproteome of macular (I), peripheral retina (II), central RPE (III), and peripheral RPE (IV) was compared to age-matched control donor eyes to determine region-specific senescence-associated molecular mechanisms during AMD progression. Phosphoproteins were enriched by charge-based spin column chromatography and resolved by 2D gel electrophoresis as previously reported . In addition, trypsin digested phosphopeptides from whole lysates were enriched using Ga3+/TiO2 immobilized metal ion chromatography. Eluted phosphopeptides were analyzed using mass spectrometry including MALDI-TOF-TOF and ESI MS/MS.
2.3. Immunocytochemical analysis
To analyze mitochondrial morphology, Cells were incubated with 100 nM MitoTracker Orange (Molecular Probes). Cells were fixed using 10% formaldehyde (25 min) and the membrane was permeabilized using 0.2% Triton X-100 (20 min), followed by blocking (0.05% Tween 20, 10% FBS, 1 h) and incubation overnight at 4°C with anti-prohibitin antibody (1:500; Genemed Synthesis, San Antonio). Prohibitin was visualized using Alexa Fluor 488-secondary IgG antibody (1:700; 1 h at 25°C; Molecular Probes). The nuclei were visualized by incubation with DAPI (4, 6-diamidino-2-phenylindole) added to VECTASHIELD Mounting Medium (Vector Laboratories, Burlingame, CA). A fluorescence microscopy was used for image analysis (Zeiss AxioVert 200 M Apo Tome, 63×).
ARPE-19 cells were rinsed (Modified Dulbecco’s PBS) and lysed using immunoprecipitation (IP) lysis buffer (pH 7.4) containing Tris (25 mM), NaCl (150 mM), EDTA (1 mM), NP-40 (1%), glycerol glycerol (5%), and protease inhibitor cocktail by incubating on ice for 5 min with periodic sonication (3 × 5 min), followed by centrifugation (13,000 × g, 10 min). Proteins (1 mg/mL, 200–400 μL) were loaded for immunoprecipitation and nonspecific bindings were avoided using control agarose resin cross-linked by 4% bead agarose. Amino-linked protein-A beads were used to immobilize anti-prohibitin antibody with a coupling buffer (1 mM sodium phosphate, 150 mM NaCl, pH 7.2), followed by incubation (room temperature, 2 h) with sodium cyanoborohydride (3 μL, 5 M). Columns were washed using a washing buffer (1 M NaCl), and protein lysate was incubated in the protein A-antibody column with gentle rocking overnight at 4°C. The unbound proteins were spun down as flow-through, and the column was washed three times using the washing buffer (1 M NaCl) to remove nonspecific binding proteins. The prohibitin-interacting proteins were eluted by incubating with elution buffer for 5 min at room temperature. The eluted proteins were equilibrated with Laemmli sample buffer (5X, 5% β-mercaptoethanol). Eluted proteins were separated using SDS–PAGE and visualized using Coomassie blue (Pierce, IL) or silver staining kit (Bio-Rad, Hercules, CA). Prohibitin and p53 were visualized using Western blot analysis. Proteins were identified by mass spectrometry analysis.
2.5. Mass spectrometry analysis
Protein bands were excised into 1 × 1 × 1 mm cubes. The Coomassie-stained or silver-stained gel pieces were incubated using a Coomassie destaining buffer (200 μL of 50% MeCN in 25 mM NH4HCO3, pH 8.0, room temperature, 20 min) or silver destaining buffer (50% of 30 mM potassium ferricyanide, 50% of 100 mM sodium thiosulfate, 5–10 min). The gel pieces were dehydrated (200 μL, MeCN) and vacuum-dried (Speed Vac, Savant, Holbrook, NY). Proteins were reduced (10 mM DTT, 100 mM NH4HCO3, 30 min, 56°C) and were alkylated (55 mM iodoacetamide, 100 mM NH4HCO3, 20 min, room temperature in the dark). Proteins were digested using trypsin (13 ng/μL sequencing-grade from Promega, 37°C, overnight) in NH4HCO3 (10 mM) containing MeCN (10%). The peptides were enriched using a buffer (50 μL, 50% MeCN in NH4HCO3, 5% formic acid, 20 min, 37°C). Dried peptides were dissolved in the mass spectrometry sample buffer (5–10 μL, 75% MeCN in NH4HCO3,1% trifluoroacetic acid). Alpha-cyano-4-hydroxycinnamic acid (5 mg/mL, Sigma-Aldrich, St. Louis, MO) was freshly dissolved in a matrix buffer (50% MeCN, 50% NH4HCO3, 1% triflouroacetic acid) and centrifuged (13,000× g, 5 min). The peptide-matrix mixtures (0.5 μL) were spotted onto the MALDI target plate (Ground steel, Bruker Daltonics, Germany). Mass spectrometer and all spectra were calibrated using a known peptide, including trypsin (842.5099, 2211.105 Da). The mass spectrum was recorded in 800–3000 Da range using Flex MALDI-TOF mass spectrometer (Bruker Daltonics, Germany, 70–75% laser intensity, 100–300 shots). Mass spectrometry data were analyzed using Flex analysis software (Bruker Dal- tonics, Germany). Peptides were identified using the Mascot software (Matrix Science) and NCBI/SwissProt database (zero mismatch cleavage, carbamidomethyl cysteine, methionine oxidation, 50–300 ppm mass tolerance). Peptide identification was evaluated based on Mascot MOWSE score, number of matched peptides, and protein sequence coverage. MOWSE score is expressed as -10logP as a probability value to compute the composite probability P.
2.6. Quantitative analysis of mitochondrial morphology
The connectivity, the number of mitochondrial branch points, and the interactive 3D visualization of isosurfaces were examined to identify the contact area between mitochondria and other organelles.
Mitochondria were stained using MitoTracker Orange/Red or rhodamine 123. Mitochondrial interconnectivity and elongation were analyzed systematically using computational software, including Mito-Morphology macro, Mitograph 2.0, Imaris 8.2.1., and SOAX 3.6.1.
Mitochondrial size and morphology were analyzed using the software connected to Image J software (Ruben Dagda:
Mitochondrial shape, including fused, fragmented, tubular, swollen, branched, uniform, and perinuclear clustering, was examined quantitatively. Mitochondrial filaments in three dimensions and mitochondrial parameters in ARPE-19 cells were calculated mathematically using Image J and Imaris (v8.2.1) software. The connectivity, the number of mitochondrial branch points, and the interactive 3D visualization of isosurfaces were examined to identify the contact between mitochondria and other organelles.
2.7. Mitochondrial mapping in AMD
Mitochondrial signaling in a network-based interactome map between genome, proteome, and metabolome of AMD was established using proteome data and bioinformatics software. The protein-protein interaction was established using the Münich Information Center for Protein Sequence (MIPS), Biomolecular Interaction Network Database (BIND), the Database of Interacting Proteins (DIP), the Molecular Interaction Database (MINT), the Protein Interaction Database (IntAct), and STRING.
Interaction mapping of prohibitin was determined using immunoprecipitation, followed by mass spectrometry analysis. Prohibitin binding proteins in the RPE were connected using STRING 10.0 software (
Prohibitin interactions were confirmed using eight sources that include neighborhood, gene fusion, co-occurence, high-throughput interaction experiments, databases, homology, conserved co-expression, and published knowledge, including ExPASy (
AMD and oxidative biomarkers interactome were established using protein–protein interaction map software and databases, including STRING 10.0 (
The genome regulatory network was connected to the proteome network using Uniprobe and JASPAR. Protein phosphorylations were examined by phosphoprotein/peptide enrichment, followed by mass spectrometry analysis. Phosphorylations were compared to Phospho.ELM, and PhosphoSite. The metabolome mapping was established using KEGG and BIGG databases.
To understand how mitochondria regulate their morphology and function, we first analyzed mitochondrial morphology quantitatively in ARPE-19 cells subjected to oxidative stress conditions using a systematic computational model (Figure 1). Representative mitochondrial images at selected time points (0.5, 1, 8, 24 h) are shown for clear comparison Figure 1). We examined mitochondrial area, circularity, perimeter, content as well as cellular area to identify changes between healthy and injured mitochondria. Previously, our in vitro data using RPE cells demonstrated the positive correlation between apoptotic signaling and mitochondria-nucleus prohibitin shuttling. Our previous studies suggest that cellular distribution and the total volume of mitochondria could be affected by microtubules, intermediate filaments and cardiolipin [24, 25, 26, 27, 29, 30].
Our results showed that the connectivity, the number of mitochondrial branch points, and the interactive isosurfaces were altered at the contact sites between mitochondria and other organelles. Under extended oxidative stress (1, 8, 24 h) and intense light (7000 lx, 1 h), we observed a decrease in mitochondrial size, presence of fragmented filaments (red arrows), and holes on the organelle contact sites. Under intense light condition, mitochondria in ARPE-19 cells were decreased and fragmented as shown in oxidative stress (1–8 h).
Next, mitochondrial perimeter vs. circularity was examined to determine the correlation between mitochondrial morphology and oxidative stress (Figure 2). We hypothesized that some mitochondrial indexes that include circularity and perimeter ratio may represent mitochondrial dynamics. We calculated mitochondrial area, perimeter, minor axis, and circularity to conclude that specific mitochondrial ratio correlated positively with stress kinetics.
The average mitochondrial area/perimeter ratio normalized to the minor axis suggests that specific conditions may induce mitochondrial swelling (Figure 3). Time-dependent decrease of minor axis and mitochondrial area/perimeter normalized to the circularity was noticed under stress condition. Our previous proteomic study demonstrated that tubulin/vimentin depolymerization and phosphorylations increased in stressed mitochondria .
To understand mitochondrial dynamics in detail, mitochondrial trafficking complex was examined. Subcellular fractionation, immunoprecipitation using primary prohibitin antibody, native gel, and mass spectrometry analysis suggest that motor protein complex may determine mitochondrial dynamics and retrograde signaling under stress conditions. Molecular motor complex contains plus-end-directed kinesin 19, myosin 9, protocadherin gA7, and prohibitin (Figure 4). The molecular motor/adaptor/receptor complex mediates mitochondrial dynamics. Motor proteins, including kinesin and myosin, facilitate mitochondrial trafficking along the cytoskeleton, mainly microtubules, actin polymers and intermediate filaments (Figure 5). Domain analysis of mitochondrial trafficking complex showed that plus-end-directed kinesin 19 (ATP and Mg2+ binding domains), myosin 9 (myosin head motor domain, SH3 domain, ATP binding domain), protocadherin gA7 (cadherin repeat and Ca2+ binding domain), and prohibitin (PX domain, lipid binding pocket) exist in the trafficking complex.
In order to correlate our in vitro findings with human pathology, we analyzed mitochondrial trafficking complex in human postmortem AMD eyes using a proteomic approach (Figure 6). RPE and retina tissues (central vs. peripheral) from AMD eyes and age-matching control eyes were analyzed by phosphoproteomics and mass spectrometry analysis. We observed different expression levels of prohibitin, inositol receptor, calponin, ankyrin, guanylate cyclase, and NADP ubiquinone oxidoreductase in the RPE and pyruvate kinase, PP2A, creatine kinase, PAK S/T kinase, vimentin, FES tyrosine kinase and dynamin like protein in the retina from AMD samples compared to control. Our results suggest that mitochondrial trafficking could be a significant determinant of RPE apoptosis by decreased prohibitin. Further, Ca2+, Fe2+, inositide, phosphorylation, and energy imbalance may lead to the accelerated pathogenesis toward AMD.
It is proposed that specific organelles, including mitochondria, melanosome, and phagosome, may use different kinesin and myosin motors for their distribution and trafficking in the RPE [25, 31, 32, 33]. Mitochondrial trafficking could be determined by ATP, Ca2+, and lipid interactions based on their domain analysis [29, 34, 35, 36, 37, 38] and mitochondrial trafficking is a significant determinant of RPE apoptosis [24, 37]. Altered concentrations of mitochondrial complex, phosphoproteins, and ATP/ADP may lead to premature senescence in RPE cells . Our enriched phosphoproteins and phosphopeptides analysis demonstrated that altered inositol triphosphate receptor, ankyrin, NADP reactions exist in AMD (Figure 6). Regulation of mitochondrial complex/lipid ratio and the energy producing machinery may enable enhanced longevity of RPE cells.
Next, a 3D surface model was used to analyze mitochondrial nodes, edges, branches, and tubular filaments (Figure 7). Cellular distribution and the total volume were affected by microtubules, intermediate filaments and cardiolipin. A 3D model showed that mitochondrial contact sites with endoplasmic reticulum (ER) and/or the nucleus were opened irreversibly under extended stress (24 h).
In order to understand AMD protein network, the new AMD interactome with oxidative biomarkers was established using phosphoproteomics data and a computational model. The current AMD interactome demonstrated that several earlier unrelated to AMD proteins, including ubiquitin, peroxiredoxin, MAP kinase, BUB 1/3, vimentin and crystalline could be involved in AMD progression, suggesting that cytoskeletal protein phosphorylation, crystalline aggregation, and mitochondrial signaling may contribute to RPE apoptosis (Figure 8). To confirm oxidative stress biomarkers, specific cytoskeletal protein changes were determined in vivo using animal model previously (C3H female mice, 7 weeks old) [25, 30, 38]. Neurofilament, vimentin, and tubulin were upregulated under 24 h constant light compared to 12 h dark/12 h light condition .
The current study determined the mitochondrial morphology quantitatively using a mathematic model and mitochondrial trafficking complex under stress conditions. Our data suggest that the kinesin-myosin-cadherin-prohibitin complex could be involved in anterograde mitochondrial trafficking, whereas PKUb S/T kinase-myosin-PI3K-lamin B2 bindings may regulate an energy demanding retrograde transport of mitochondria [25, 39, 40, 41, 42, 43]. Prohibitin binding with a trafficking protein complex may regulate the bidirectional transport of mitochondria along actin microfilaments, intermediate filaments, and microtubules. The mitochondrial trafficking complex implies that a specific mechanism of communication may exist in the ATP and Ca2+ demanding regions. Mitochondrial dysfunction, altered dynamics, impaired transport, and turnover perturbation are associated with AMD.
Oxidative stress-induced apoptosis is the final cell death pathway in many irreversible ocular diseases that include AMD. While the end point of apoptosis is well established, the knowledge of early biochemical reactions and specific molecular players has been elusive. We have examined early biosignatures and mechanisms of retinal and RPE cell death under oxidative stress . Our previous studies demonstrated that not only intense light but also constant moderate light and mild oxidative stress may trigger induction of anti-apoptotic Bcl-xL and erythropoietin (EPO) as well as pro-apoptotic caspases [24, 25, 27, 28, 29, 37, 38, 45, 46, 47]. We determined that protein modifications, including nitration and phosphorylation, were altered under oxidative stress possibly due to excess of NO production [26, 48, 49, 50].
The analysis of AMD interactome using proteome-genome-metabolome network suggests that there is a positive correlation between mitochondrial retrograde signaling and AMD progression. The AMD interactome suggests: (1) network-based interactions among AMD-related hub proteins that include UBC, MMP2, BCL, PRDX, ATP5O, C3, TF, and CRYAB, (2) increased local interactions between oxidative stress, complement activation, transcription, metabolism, (3) AMD module as a cluster in the same network neighborhood, (4) potential causal molecules including Ca2+, Fe2+, ATP/ADP, OPO42−, lipids, (5) altered cytoskeleton, microtubule, abnormal mitochondrial signaling.
The mitochondrial interactome provides a base for better understanding of oxidative stress-induced apoptosis and the mechanism of age-related diseases, including AMD. As a consequence, an effective treatment of neurodegenerative diseases based on the modulation of mitochondrial network is expected to result.