Integrating mitochondrial gene expression data to model the effects of respirasome supercomplex formation on reactive oxygen species production in Alzheimer's disease models

Abstract

Background

In the past decade, Alzheimer's disease (AD) researchers have found that the formation of amyloid aggregates occurs after dysregulation of respiratory chain activity.

Objective

Using our developing mathematical model to identify potential therapeutic targets for AD treatment.

Methods

We have constructed a mathematical model for AD that incorporates enzyme activities and kinetics, and protein and mRNA expression levels.

Results

Our analyses of gene expression in AD brains provide complimentary evidence that changes in mitochondrial energy production and biogenesis accompany AD pathogenesis. Using this data, we created a mitochondrial model of electron transport chain supercomplex assembly.

Conclusions

By carrying out sensitivity analyses of responses to oxidative stress and effects of gene expression on energy production, we demonstrate how oxidative stress and energy deficits change the initial antioxidant defense system and impact the progression of AD. We investigated the impact of gene expression changes in 9 genes as new therapeutic options via metabolic flux analysis of supercomplex assembly. Through careful analysis, we propose that UQCRC1 may be an effective therapy option for in vitro AD treatment testing.

Keywords

Alzheimer's disease electron transport chain gene expression changes mathematical model mitochondrial supercomplexes respirasome UQCRC1

Introduction

Alzheimer's disease (AD) is a neurodegenerative disease characterized by the aggregation of neurofibrillary tangles and amyloid plaques.^1–3 It is unclear if these aggregates cause cognitive degeneration or if they are a byproduct of other degenerative stimuli.⁴ In either case, cell death and impairment of mitochondrial respiration can be seen before aggregate accumulation.^1,5–7 Unlike early onset AD, there are no characteristic genetic mutations predisposing patients to late onset AD (LOAD) although there are genetic risk factors for LOAD.⁸ Current research suggests that dementia is a consequence of aging and the efficiency of the mitochondria plays a significant role on when and if a person develops dementia.^1,9

Respiration is the process of oxidizing NADH and FADH2 to release electrons which reduces Coenzyme Q (CoQ) before being transferred to cytochrome C (cyt C) and eventually converts O₂ gas into water. The respiratory complexes I, II, III, and IV (functional units) of the mitochondria function with a spectrum of individual complexes, supercomplexes (SCs; functional intermediaries composed of a fraction of the complexes within the respirasome) and the respirasome (a unified supercomplex composed of respiratory complexes I, III, and IV located in the mitochondrial inner membrane). SCs may reduce the diffusion distance of cytochrome c and coenzyme Q between functional units and thus increase respiratory efficiency.¹⁰ Formation of the respirasome limits reactive oxygen species (ROS) production by stabilizing NADH dehydrogenase (CI) structure as well as covering one oxygen binding site on CI that plays a role in mitochondrial stress signaling.¹¹ SC assembly varies with cell state and type, which indicates SC formation could be physiologically significant, thereby warranting further exploration.¹²

Our proposed deterministic mitochondrial model focuses on oxidative stress and changes in the initial antioxidant defense system as well as induction of secondary response cascades to alleviate cellular stress from excessive ROS production. Through careful analysis we propose UQCRC1 will be a strong candidate for therapeutic intervention for future in vitro AD treatment testing.

Methods

Sample acquisition

Autopsy-confirmed brain samples were kindly supplied by the following: Dr W. W. Tourtelotte of the National Neurological Research Specimen Bank (NNRSB), VAMC, Los Angeles, CA 90073. Tissue/fluid specimens obtained from the NNRSB, are sponsored by NINDS/NIMH, National Multiple Sclerosis Society, Hereditary Disease Foundation, Comprehensive Epilepsy Program, Tourette Syndrome Association, Dystonia Medical Research Foundation, and Veterans Health Services and Research Administration, Department of Veterans Affairs; Dr Peter Davies of the Department of Pathology at Albert Einstein College of Medicine, New York, NY; Dr Deborah C. Mash, University of Miami Brain Endowment Bank, Miami, FL.

As described in our companion paper, the brain samples used in this study consists of 6 control (5 males, 1 female), four of which were APOE genotypes ε3,ε3, one was ε3,ε4, and one (male) that was undetermined and 6 AD (4 male, 2 female) samples. Of the AD samples, two (one male, one female) were ε3,ε3, three (two males, one female) were ε3,ε4, and one (male) was undetermined. The average age for the control samples was 69.3 and that for the AD samples was 73.2 years.

Brain mRNA isolation

RNA isolation and assessment as well as a related gene expression analysis approach are generally described in our earlier publication.¹³ Briefly, age matched brain samples taken from the parietal cortex of 17 samples- 6 controls and 11 AD autopsy brain samples, average ages of 75 and 74, respectively. Total RNA was isolated from 80–100 mg frozen autopsy brain samples using RNeasy Plus Universal Mini kits per manufacturer's instructions.

RNA assessment

The integrity and purity of the isolated RNA was determined by absorbance spectroscopy and agarose or capillary gel electrophoresis. Gel electrophoresis assesses the size and integrity of the mRNA by identifying the presence of both 28S and 18S rRNA species on the gel. The presence of sharp bands from these distinctive, large subunits indicates a lack of degradation. Capillary electrophoresis was performed on an Agilent Bioanlyzer using a Nano II total RNA kit. Capillary electrophoresis assesses the quantity, size and integrity of the mRNA and is used to calculate an RNA integrity number (RIN). The average RIN for the 6 control samples was 5.5 while that for the AD samples was 6.0.

Mitochondrial energy production gene expression analysis

The Qiagen RT2 PCR Profiler PCR Array consisting of 12 control and 84 nuclear-encoded mitochondrial genes involved in mitochondrial energy production used sequence specific primers to amplify the corresponding cDNA strands in a 96-well plate configuration. A list of genes included on the array is available upon request. Relative changes in SYBR green fluorescence were used to assess levels of gene expression using a BioRad CFX96 thermocyler and web-based Qiagen software.

Model creation

Our model was created in CellDesigner 4.2 and was converted into a series of mathematical equations by SBMLsqueezer.¹⁴ This file was imported into COPASI, a biological pathway simulation program. Every arrow within the model is represented by a “flux equation.” Formation of a species results in a positive flux value while depletion of a species gives a negative flux value. These equations also describe the catalysis of reactions. If the reaction is enzyme catalyzed, then the equation takes the form of a Michaelis-Menten equation (M-M).

When the reactions are not enzyme catalyzed, they are represented by a simple mass action equation. For reactions with more than one substrate, the M-M equation is modified to accommodate a second substrate assuming a random order mechanism (Equation 1).

\frac{k_{c a t} \times [e n z y m e] \times [s 1] \times [s 2]}{K_{i 2} \times K_{m 1} + [s 1] \times [s 2] + K_{m 1} + K_{m 2}}

(1)

This equation has an additional K_m specific for the second substrate (indicated by “m2”) and an inhibition constant (K_i2). Like the M-M equation, bi-substrate random order mechanisms assume rapid equilibrium under these conditions, and the latter is simplified out of the final equation. These equations and others defined by Systems Biology kinetic laws are combined to form a series of flux equations that describe the entire reaction.

Concentrations (µM) of proteins were calculated using brain tissue specific abundances from the PAXdb database¹⁵ and the molecular weights from the UniProt database.¹⁶ If brain tissue specific abundances could not be found, S. cerevisiae, a common cell model, data was used. A yeast model was chosen over mammalian models due to the consistent availability of all protein concentrations. To create the AD model protein concentrations, the mRNA fold-regulations were used to approximate protein concentrations with the assumption that there were no other factors affecting translation other than the disease state, making fold changes a reasonable approximation of the AD models protein concentration. Metabolite concentrations were taken from the Human Metabolome Database, and parameters from the BRENDA database.¹⁷ A more detailed description of our model can be found in our earlier publication.¹⁸

Parameter fitting

Initial parameter sets were created based on rate constants found in the literature. Those that could not be found were given an initial value of 0.01 for Km and 1 for k_cat. These values are then optimized to fit parameters to the model system. This optimization tasks were carried out in COPASI using the Hooke & Jeeves solver.¹⁹ The initial and final parameter sets can be found in the Supplement. Parameter units for k_cat, and K_m are min⁻¹ and uM, respectively.

Optimization

While deterministic mathematical models can enhance understanding living processes, they are a simplification of complex systems. To reduce potential model bias, standard data collection and processing measures are taken to ensure the best possible representation of the biological system includes utilization of rate constants and protein concentrations determined from in vitro testing whenever possible as well as parameter optimization. Parameter optimization is a series of iterations that adjusts the in vitro rate constants to better fit a smaller model with more limited functions. The computational process of optimization based on cellular function criteria which include NADH/NAD ratios, ATP production and basal free radical production as seen in normal cells. The rate constant values are randomly adjusted up or down at small increments thousands of times within Copasi until the output remains “fixed” despite further adjustments to the rate constant values. The aim of optimization is to find the best fit of a given parameter set to replicate normal cell function and is unrelated to the later perturbations done within the model to evaluate the impact of AD on cellular activity.

Treatment tests

Treatment efficacy was analyzed with COPASI's time course task. To simulate CRISPR treatments, changes in gene transcription were simulated by adjusting the protein concentration up or down 50%. Output and efficacy was assessed via percent difference from control calculated using R, version 3.5.2.²⁰

Sensitivity data

A sensitivity analysis of the model was performed to identify enzymes for treatment simulations. The sensitivity analysis perturbs one target concentration by 1%, and COPASI records how this change effects the dependent variables of the system (reaction rates and transient concentrations). The results are presented as the average percent dependency of a dependent variable's value on the initial concentration being varied. This analysis is only accurate for perturbations up to 5%, higher changes in concentration may deviate from this trend.

Supercomplex assembly model

After individual respiratory complexes CI, Ubiquinone-Cytochrome C Reductase (CIII), and Cytochrome C oxidase (CIV) complexes have formed they begin to associate and form SCs. SCs are stable intermediates between respirasomes and individual complexes. The two most prevalent SC compositions are CIVCIII₂ and CICIII₂ (abbreviated CIVCIII and CICIII) shown in Figure 1. The latter is more common.²¹ The differences in SC assembly have been implicated in increasing respiratory efficiency and reducing ROS proliferation.²² SC and respirasome concentrations also vary between cell types. For example, astrocytes have been shown to have lower levels of SC formation, and this increases the basal level of ROS within the mitochondria¹²

Figure 1.

The two pathways for respirasome formation involve two SC intermediates, CICIII₂ and CIVCIII. (a) CICIII₂ forms and assembles with CIV containing the COX7A2 subunit. (b) CIVCIII forms using CIV with the COX7A2L subunit then with CI to form the final respirasome.

Nuclear encoded ETC subunits are used to stabilize complexes as well as assemble SCs and respirasomes. The lesser found SC CIVCIII formation requires the COX7A2L isoform being present in CIV; it results in rapid assembly of the respirasome²³ (Figure 1). COX7A2, the shorter isoform, within CIV leads to CICIII SC intermediate formation before final respirasome formation²⁴ (Figure 1(a)). Within CIII, the expression of subunits, UQCRFS1, UQCRC1 and UQCRB ensure stable association with COX7A2L.²⁵ CI association with the CIII dimer (usually forming CICIII2) reduces ROS formation in the mitochondria making the association highly favorable both in vivo and in vitro.²⁶

This association utilizes two bonding groups for stabilization including CI subunits NDUFB4 and NDUFB9, which bind to CIII subunits UQCRFS1 and UQCRC1. CIII also uses NDUFA11 and NDUFB4 to bond with UQCRQ.^11,23 Absence of one of these subunits does not prevent SC formation, but it does form a less stable association between individual complexes²⁷ (Figure 2).

Figure 2.

Assembly of SCs is dependent on the expression of nuclear encoded subunits NDUFB9, NDUFA11, NDUFB4, NDUFS1 in CI, and UQCRFS1, UQCRC1, and UQCRQ in CIII.

An SC with a CICIII₂CIV composition is known as a respirasome, but larger respirasome structures have been reported with up to four CIII and CIV present.²⁵ Only 20% of mitochondrial CIV is incorporated into the respirasome compared to 55–65% of CIII and over 85% of CI.²⁶ This near complete incorporation of CI into SCs supports the hypothesis that all SC formation stabilizes CI and reduces ROS production.²² Another study of CI subunit NDUFS1 show that deletion or downregulation of the subunit inhibits respirasome formation.¹² This loss in vivo could significantly affect mitochondrial SC composition which in turn can change ROS production and mitochondrial signaling.

Model assumptions

SC assembly

SC assembly is a complex, reactive process of combining respiratory complexes into larger structures. The SC assembly model investigates how changes in gene expression, determined through proteomic analysis, influence SC assembly and sequestration of CI to reduce ROS formation. The initial model (Figure 3(a)) considers the interactions between four CI subunits and five CIII nuclear encoded subunits which are not required for ETC function, but stabilize the complexes.^19,28 Figure 3(a) shows that the expression of different subunits, or the loss of subunits produces respiratory complexes with differing levels of SC stability. Weaker connections cause an increase in ROS production as CI will not be fully stabilized by the SC formation (parameters and equations can be found in the Supplemental Material). When the concentrations of subunits are considered (Table 1) and assuming preference for creating complexes with all of the subunits, S1 is reduced to the model shown in Figure 3(b).

Figure 3.

(a) Model of potential SC assemblies based on subunit expression. The letter and number pairs represent the potential groupings of subunits; not all subunits are required for complex formation. CI is shown in light gray (B9 = NDUFB9, B4 = NDUFB4, A11 = NDUFA11 and S1 = NDUFS1), CIII in gray (FS1 = UQCRFS1, Q = UQCRQ, and C1 = UQCRC1), and CIV in yellow has one known component for SC assembly: COX7A2 (abbreviated 7A2). SC CICIII is shown in purple, CIVCIII in orange and Respirasome (resp) in white. CICIII is differentiated by the number of stable inter-complex bonds being formed with 3 being the most and 1 the least. The prefix “s” represents the presence of CI subunit NDUFS1, an essential subunit for respirasome formation. This model represents all potential combinations. (b) Based on cellular subunit concentrations, the model was reduced to a combination of six ETC complex types that could assemble to form one of three SCs and a complete respirasome. (Colors can be seen in the online version.)

Table 1.

Nuclear subunits and supercomplex assembly.

Subunit	Control (µM)	AD (µM)
UQCRFS1	13.62	7.41
UQCRC1	9.57	3.75
UQCRQ	5.14	2.49
NDUFB9	7.42	2.52
NDUFB4	10.06	9.22
NDUFA11	1.85	0.84
NDUFS1	6.29	2.44
COX7A2L	3.54	2.04
COX7A2	28.63	13.39

Results

The formation of SC's seems to play an important role in minimizing ROS creation during the process of ATP production by keeping individual respiratory complexes in close proximity to one another. This adjacency increases the efficiency of electron transport from one complex to the next thereby reducing the aberrant loss of electrons that can subsequently react with other local compounds, creating ROS.²⁷ The uniqueness of SC formation to cell types predicts that changes in assembly will have widespread effects on mitochondrial and, subsequently, cellular homeostasis. To understand how AD affects respirasome formation and grow intuition about how increasing subunit expression can either alleviate or exacerbate these issues, we simulated SC formation in control and AD conditions. After compiling species concentrations, the fold-regulations were used to determine the AD concentrations (Table 1). The values were then loaded into the model assuming that only one subunit could be absent from a functional complex. There is an overall decrease in available subunits including a decrease of approximately 2-fold in essential respirasome assembly subunit NDUFS1.

After determining AD ETC concentrations, treatment sets were created by increasing one subunit concentration by 50% relative to AD concentrations. The concentration of individual complexes was reassessed with the treatment and shown in Table 2. Treatment sets were created by increasing one subunit's concentration by 50%. Subunits that did not increase or decrease the concentration of complexes were not considered.

Table 2.

Effects of therapies on complex concentrations.

	Control (µM)	AD (µM)	UQCRC1 (µM)	UQCRQ (µM)	NDUFA11 (µM)	NDUFS1 (µM)	NDUFB9 (µM)
CI
B9B4A11S1	1.85	0.84	0.84	0.84	1.26	0.84	0.84
B9B4S1	4.44	1.60	1.60	1.60	1.18	2.83	2.93
CIII
FS1QC1	5.15	2.49	2.49	3.73	2.49	2.49	2.49
FS1C1	2.72	1.26	3.13	0.02	1.26	1.26	1.26

The treatment sets were analyzed for how they affected CI stability via assembly with CIII and CIV. CIII is an important initial CI binding partner with CIV providing minor additional stability after CICIII formation.²³ Subunit targets were selected based on their effect on CI or CIII concentrations compared to AD values. Simulations of SC assembly produced theoretical information about changes in respirasome concentration (Figure 4) and changes in the percent of CI: free CI, and sequestered by CICIII or respirasomes (Figure 5). The percentage of free CI in the AD sample is double that of the control, and the decreased respirasome formation (Figure 4) indicates that the AD tissue underwent a long period of oxidative stress.

Figure 4.

Simulation of mitochondria in the control, AD and treatment groups resulted in different concentrations of complete respirasomes. The NDUFA11 treatment groups had no change in concentration of respirasomes compared to the AD group. NDUFB9 and NDUFS1 treatments provided a slight increase in respirasome concentration compared to the AD group.

Figure 5.

CI incorporation into the CICIII complex and Respirasome. (a) Percent of CI assembled into SCs (orange) or free (blue) in control and treatment groups after simulation of complex assembly. Treatment was defined as increasing theoretical expression of a single respiratory complex subunit, identified by the subunit's name. UQCRC1 treatment showed a similar percent of C1 integration into SC CICIII compared to the control. UQCRCQ, NDUFA11, NDUFS1, and NDUFB9 had lower percentages of C1 incorporation into the CICIII than that of the AD group. (b) Percent of CI incorporated into respirasomes of the control, AD and treatment groups. The Treatment Groups refer to UQCRC1-NDUFB9.

The CI subunit treatments predict that increasing respirasome concentration comes with an increase in free CI and potentially dangerous levels of ROS production. NDUFS1 and NDUFB9 treatments were able to increase the concentration of respirasome (Figure 5), but this increase in respirasome concentration was paralleled by an increase in free CI. UQCRC1 treatment was the only therapy predicted to increase respirasome formation and CI sequestration and ultimately result in a decreased percentage of free CI.

Discussion

Because mitochondrial dysfunction occurs early on in AD progression,²⁹ this model was created in an attempt to evaluate specific targets that increase stable SC formation which is theorized to decrease ROS production.¹⁰ This model provides additional evaluation of this theory. It too predicts that a decrease in ROS production could be achieved by increasing stability of SCs in the mitochondria. As CICIII formation is predicted to be the most important step in reducing CI-related ROS production by covering one of two oxygen binding sites, CIII subunit UQCRC1, according to our model simulation, is considered the most effective subunit target based on predicted increased efficacy of sequestering CI to reduce ROS production. By increasing CIII binding partner UQCRC1 for CI, the model theorizes that there will be an increase in the percentage of CI within respirasome complexes. During simulation, there was a complementary decrease in free CI that in simulation, reduces the amount of ROS produced by CI. Based on this model, we predict that targeting increased CIII binding to CI without increasing available CI for respirasome binding is a viable treatment strategy to increase sequestration of CI, which could reduce ROS production related to CI activity in vivo. This model's predictions about SC assembly may be of importance when considering novel therapeutic targets for future AD treatments. Further work needs to be done studying the effects of subunit and antioxidant therapies on mitochondrial stress response in vitro to assess the utility of these treatments.

Limitations of this model include the use of postmortem AD brain tissue. Although this information does give us a window into the function of the AD brain, the rate and progression of mitochondrial decline is not well understood. Thus, the differences in expression in this model may not be entirely accurate to the AD brain in earlier stages of disease. New methods of evaluation and quantification of mitochondrial gene expression during life are needed to discern more accurately what is genetic variability and what is related to disease progression. This research is also limited to qualitative evaluation of disease and treatment states in these deterministic models. Additional work with probabilistic models will better quantify the impacts of SCs and provide researchers with statistical analyses of the impacts of various treatment avenues on rate and quantity of ROS production. To prove relevance and efficacy, in vitro studies should be performed to determine if the output in vitro is similar to theoretical findings.

Supplemental Material

sj-docx-1-alz-10.1177_13872877251362889 - Supplemental material for Integrating mitochondrial gene expression data to model the effects of respirasome supercomplex formation on reactive oxygen species production in Alzheimer's disease models

Supplemental material, sj-docx-1-alz-10.1177_13872877251362889 for Integrating mitochondrial gene expression data to model the effects of respirasome supercomplex formation on reactive oxygen species production in Alzheimer's disease models by Morgan Shelton, Kimberly A Kerns, Sonali Shirali, Frank J Castora and Randolph A Coleman in Journal of Alzheimer's Disease

Footnotes

Acknowledgments

The authors have no acknowledgments to report.

ORCID iDs

Kimberly A Kerns

Frank J Castora

Ethical considerations

This study received ethical approval from the Eastern Virginia Medical School Institutional Review Board (IRB), which has determined that use of these autopsy frozen brain samples is exempt from IRB review.

Author contribution(s)

Morgan Shelton: Conceptualization; Formal analysis; Investigation; Methodology; Validation; Writing – original draft, Writing – review & editing.

Kimberly Kerns: Data curation; Methodology; Writing – review & editing.

Sonali Shirali: Formal analysis; Investigation; Methodology; Writing – review & editing.

Frank Castora: Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Supervision; Writing – review & editing.

Randolph Coleman: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Software; Supervision; Validation; Writing – original draft; Writing – review & editing.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by grants from the Virginia Center of Aging Alzheimer's and Related Diseases Research Award Foundation (22-097) and from the Commonwealth Health Research Board (274-08-17).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

The data supporting the findings of this study are available within the article and/or its .

Supplemental material

Supplemental material for this article is available online.

References

Grimm

Friedland

Eckert

. Mitochondrial dysfunction: the missing link between aging and sporadic Alzheimer’s disease. Biogerontology 2016; 17: 281–296.

Swerdlow

Burns

Khan

. The Alzheimer’s disease mitochondrial cascade hypothesis. J Alzheimers Dis 2010; 20: S265–S279.

Cadenas

Davies

KJA

. Mitochondrial free radical generation, oxidative stress, and aging. Radic Biol Med 2000; 29: 222–230.

Herrup

. The case for rejecting the amyloid cascade hypothesis. Nat Neurosci 2015; 18: 794–799.

Yan

. Mitochondrial permeability transition pore in Alzheimer’s disease: cyclophilin D and amyloid beta. Biochim Biophys Acta 2010; 1802: 198–204.

Ruetenik

Barrientos

. Dietary restriction, mitochondrial function and aging: from yeast to humans. Biochim Biophys Acta 2015; 1847: 1434–1447.

López-Armada

Riveiro-Naveira

Vaamonde-García

, et al. Mitochondrial dysfunction and the inflammatory response. Mitochondrion 2013; 13: 106–118.

Coon

Myers

Craig

, et al. A high-density whole-genome association study reveals that APOE is the major susceptibility gene for sporadic late-onset Alzheimer’s disease. J Clin Psychiatry 2007; 68: 613–618.

Mancuso

Orsucci

Siciliano

, et al.

Mitochondria, mitochondrial DNA and Alzheimer’s disease. What comes first?

Curr Alzheimer Res 2008; 5: 457–468.

10.

Milenkovic

Blaza

Larsson

N-G

, et al. The enigma of the respiratory chain supercomplex. Cell Metab 2017; 25: 765–776.

11.

Stroud

Surgenor

Formosa

, et al. Accessory subunits are integral for assembly and function of human mitochondrial complex I. Nature 2016; 538: 123–126.

12.

Lopez-Fabuel

Le Douce

Logan

, et al. Complex I assembly into supercomplexes determines differential mitochondrial ROS production in neurons and astrocytes. Proc Natl Acad Sci U S A 2016; 113: 13063–13068.

13.

Castora

Kerns

Pflanzer

, et al. Expression changes in mitochondrial genes affecting mitochondrial morphology, transmembrane potential, fragmentation, amyloidosis, and neuronal cell death found in brains of Alzheimer's disease patients. J Alzheimers Dis 2022; 90: 119–137.

14.

Dräger

Hassis

Supper

, et al. SBMLsqueezer: a CellDesigner plug-in to generate kinetic rate equations for biochemical networks. BMC Syst Biol 2008; 2: 39–45.

15.

Wang

Herrmann

Simonovic

, et al. Version 4.0 of PaxDb: protein abundance data, integrated across model organisms, tissues, and cell-lines. Proteomics 2015; 15: 3163–3168.

16.

Uniprot Consortium. Uniprot: a worldwide hub of protein knowledge. Nucleic Acids Res 2019; 47: D506–D515.

17.

Placzek

Schomburg

Chang

, et al. BRENDA In 2017: new perspectives and new tools in BRENDA. Nucleic Acids Res 2017; 45: D380–D388.

18.

Shelton

Kerns

Castora

, et al. Incorporating mitochondrial gene expression changes within a testable mathematical model for Alzheimer's disease: stress response modulation predicts potential therapeutic targets. J Alzheimers Dis 2022; 90: 109–117.

19.

Hooke

Jeeves

. “Direct search” solution of numerical and statistical problems. J ACM 1961; 8: 212–229.

20.

Ihaka

Gentleman

. R: a language for data analysis and graphics. J Comput Graph Stat 1996; 5: 299–314.

21.

Ramírez-Camacho

Flores-Herrera

Zazueta

. The relevance of the supramolecular arrangements of the respiratory chain complexes in human diseases and aging. Mitochondrion 2019; 47: 266–272.

22.

Maranzana

Falasca

Barbero

, et al. Mitochondrial respiratory supercomplex association limits production of reactive oxygen species from complex I. Antioxid Redox Signal 2013; 19: 1469–1480.

23.

Lobo-Jarne

Nývltová

Pérez-Pérez

, et al. Human COX7A2L regulates complex III biogenesis and promotes supercomplex organization remodeling without affecting mitochondrial bioenergetics. Cell Rep 2018; 25: 1786–1799.

24.

Lapuente-Brun

Moreno-Loshuertos

Acín-Pérez

, et al. Supercomplex assembly determines electron flux in the mitochondrial electron transport chain. Science 2013; 340: 1567–1570.

25.

Letts

Fiedorczuk

Sazanov

. The architecture of respiratory supercomplexes. Nature 2016; 537: 644–648.

26.

Greggio

Jha

Kulkarni

, et al. Enhanced respiratory chain supercomplex formation in response to exercise in human skeletal muscle. Cell Metab 2017; 25: 301–311.

27.

Acin-Perez

Enriquez

. The function of the respiratory supercomplexes: the plasticity model. Biochim Biophys Acta 2014; 1837: 444–450.

28.

Lenaz

Baracca

Barbero

, et al. Mitochondrial respiratory chain super-complex I–III in physiology and pathology. Biochim Biophys Acta 2010; 1797: 633–640.

29.

Kim

, et al. The role of oxidative stress in neurodegenerative diseases. Exp Neurobiol 2015; 24: 325–340.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.02 MB