OUP user menu

★ editor's choice ★

What can we learn about cardioprotection from the cardiac mitochondrial proteome?

Marjan Gucek, Elizabeth Murphy
DOI: http://dx.doi.org/10.1093/cvr/cvq277 211-218 First published online: 30 August 2010


This review will summarize proteomic methods that are useful in studying the role of mitochondria in cardioprotection. The strengths and weaknesses of some of the different approaches are discussed. We focus on the cardiac mitochondrial proteome with emphasis on changes associated with cell death and protection, and we summarize how proteomic data have contributed to addressing the role of mitochondria in cardioprotection.

  • Mitochondria
  • Cardioprotection
  • Phosphorylation
  • S-nitrosylation
  • Proteome

1. Introduction

There are several challenges in studying the role of mitochondria in protection. First, signalling molecules, which are likely involved in cell death and protection, are present in low levels and are difficult to identify in broad-based (non-targeted) proteomic approaches. Second, post-translational modifications (PTMs) are likely to be important but can also be difficult to detect both because they are labile and because of methodological limitations. Third, many schemes of cardioprotection and cell death suggest that cytosolic proteins translocate to the mitochondria to elicit cell death or protection. Determining whether these proteins are truly associated with the mitochondrial or are just a contaminant is an area of controversy.

Cardioprotection has been shown to be inhibited by blocking activation of a number of cardioprotective kinases, such as PI3K, AKT, ERK, etc.1 Furthermore, inhibition of the mitochondrial permeability transition (MPT) pore is generally viewed as one of the main targets of cardioprotective signalling. It is possible that cardioprotective signalling acts primarily in the cytosol leading to alterations, perhaps in ions or reactive oxygen species (ROS) which indirectly reduce MPT opening.1 An alternative hypothesis, which is not mutually exclusive, is that activation of these kinase signalling pathways results in changes in phosphorylation, S-nitrosylation (SNO), or other PTM of mitochondrial proteins and that these changes in mitochondrial proteins lead to altered mitochondrial function and cardioprotection.2,3 A number of groups have reported changes in phosphorylation,25 SNO,612 or O-GlcNAc13 of mitochondrial proteins, thought to be important in cardioprotection, and consistent with this, a number of groups have reported the presence or translocation to the mitochondria of kinases, phosphatases, and other signalling proteins. However, other groups find no change in phosphorylation of mitochondrial proteins with cardioprotection.14 This review will discuss the feasibility of using an unbiased proteomic approach to examine differences in mitochondrial proteins in cardioprotection. Before reviewing the data, it is useful to discuss the proteomic methods.

2. Proteomic methods

2.1 Protein identification

Figure 1 summarizes the proteomics strategies for identification of mitochondrial proteins and the advantages and disadvantages of the different approaches. Alterations in mitochondrial proteins associated with cardioprotection likely involve signalling proteins that are in low abundance. Mass spectrometry (MS) will only identify the most abundant proteins, and therefore the fewer proteins in the sample, the easier it is to identify low abundant proteins. Therefore, it is important to include fractionation methods prior to MS analysis. Fractionation can be separated by organelle (mitochondria, plasma membrane, etc.) or physiochemical (MW, pI, etc.).

Figure 1

The proteomics strategies for identification of mitochondrial proteins and the advantages and disadvantages of the different approaches.

The more extensive the fractionation, the higher are chances that a low abundant protein will be identified. Three general approaches (Figure 1) are used. One-dimensional gels separate the mitochondrial proteins according to their apparent molecular weight and are suitable for hydrophobic proteins, such as membrane proteins. After the separation and staining (e.g. coomassie blue), the gel bands are excised, in-gel digested with trypsin, and analysed on a mass spectrometer. The 1D gel separates on the protein level, which will allow separation from high abundant proteins (e.g. contractile proteins, albumin). An excellent review on MS-based proteomics was recently published.15

In a 2D gel approach, the intact mitochondrial proteins are first separated by isoelectric focusing (pI) and then by their molecular weight. Sometimes, hydrophobic, very basic, large or small molecular weight proteins (>100 or <5 kDa) may not migrate into or resolve on the 2D gel which can pose a problem with some mitochondrial proteins (e.g. membrane proteins). The resulting 2D gel can be stained for protein visualization (silver, Coomassie blue, etc.). In a quantitative approach, two samples can be compared by using 2D fluorescence difference gel electrophoresis (DIGE) in which different proteins are labelled with different fluorescent dyes and, when combined and separated on a 2D gel, can be visualized by measuring the fluorescence (colour intensity shows whether a protein is up- or down-regulated). The spots are excised and identified by tandem MS.

Two-dimensional gels are an excellent way to get a snapshot of proteins in the sample and to determine potential problems with the sample quality (e.g. degradation). Another advantage of 2D gels is the ability to visualize some PTMs that modify either pI (shift horizontally) or the molecular weight (shift vertically). Two of the most commonly observed modifications are phosphorylation (pI of phosphorylated proteins is more acidic, shifts left) and glycosylation (increases MW, shifts up). Mitochondrial studies using a 2D gel approach are discussed by Distler et al.16

Sample can also be digested in solution to peptides followed by fractionation of the peptides [e.g. with strong cation exchange (SCX) chromatography]. These peptides are then directly analysed by MS; this is typically referred to as a shotgun approach. This approach separates twice on the peptide level. With the shotgun approach, protein is not lost with extraction from the gel, however high abundant proteins are digested prior to MS, so these high abundant fragments can swamp out many fractions (unless there is ample fractionation).

2.2 Post-translational modifications (PTMs)

MS-based proteomics is in most cases sensitive enough to identify low abundant proteins after extensive upstream fractionation. However, the abundance of PTMs is a limiting factor for a successful PTM characterization. For example, if 10% of a particular site in a protein is phosphorylated (or has some other PTM), it would mean that even if you are able to detect the non-phosphorylated peptide and if the protein was at the limit of detection, you would need roughly 10× more protein to detect the phosphorylated counterpart because its abundance is below the limit of the detection of the instrument. A comprehensive account on PTM characterization was published by Witze et al.17

The functional significance of a PTM that modifies only a few percent of protein is unclear, but it likely depends on the dynamic range of the effect of the PTM on activity. For a signalling molecule, with targeted subcellular localization, a PTM that affects only a small fraction of a protein could still have a significant functional effect. For an enzyme activity (with no compartmentation issues), the functional significance of a PTM that affects only a small percentage of a protein is less clear and is an important issue. For example, if the PTM is only 5% of the protein and the PTM fully activates the enzyme resulting in an increase from 0 to 5% of Vmax, this could have important implications. In contrast if the PTM is 5% and this inhibits activity from 100 to 95%, in many cases this will have less effect on the cell function.

2.2.1 Phosphorylation

The same three approaches used to identify proteins can be used to identify phosphorylation. However, an enrichment strategy must be used such as a 2D gel followed by phosphostaining or immunoprecipitation followed by a 1D gel or direct MS, or an affinity-based approach followed by MS.

Two-dimensional gels can be used to screen for some PTMs. Methods such as phosphostain, phosphoantibody, or 32P are used to identify the phosphoproteins which can then be excised from the gel and identified by MS.18,19 Although such 2D gel methods allow identification of phosphoproteins, it can be difficult to identify sites of phosphorylation because of the low level of protein that can be loaded onto the gel. One of the gold standards for phosphorylation screening is incorporation of 32P by incubating proteins or organelles in the presence of γ-32P-ATP and measuring the radioactivity after separation on the gel.18,20

To enhance sensitivity, we need to increase the amount of starting protein and enrich for the PTM. As mentioned, there is a limit to the amount of protein that can be added to a 2D gel, thereby limiting the amount of starting protein. For this reason, enrichment approaches followed by MS typically are more successful at identifying sites of phosphorylation.

Antibody immunoprecipitation works well for tyrosine phosphorylation.21,22 In theory, an immunoprecipitation approach should also work for phosphoserine and phosphothreonine, but there are few reports, presumably because the antibodies do not work as well for immunoprecipitation.

Most commonly used enrichment strategies for phosphorylated proteins/peptides are affinity based and exploit the physicochemical characteristics of the phosphate group to interact with positively charged metal ions: immobilized metal affinity chromatography (IMAC, gallium or iron based), metal oxide affinity chromatography (titanium dioxide based), and SCX. A comparison of phosphopeptide and/or phosphoprotein enrichment strategies is given by Grimsrud et al.23

After enrichment, the phosphopeptide mixture is separated using liquid chromatography and MS. Phosphotyrosine peptides fragment along the backbone of the peptide and keep the phospho group intact on the fragment, thus enabling peptide identification and in most cases site localization. Phosphoserine and phosphothreonine peptides on the other hand fragment differently – the phospho group is the first one to undergo the fragmentation (neutral loss of phosphoric acid), leaving little or no fragmentation information, complicating the identification and site localization. In cases such as this, it is beneficial to further fragment the most abundant ion in the MS/MS spectrum which is the ion obtained by neutral loss of the phospho group. The resulting MS3 spectrum has enough information to elucidate the peptide's structure and possibly the phosphorylation site. There are several bioinformatics tools that can help in elucidating the site of phosphorylation (such as PhosphoScore24).

2.2.2 S-nitrosylation

Nitric oxide can lead to the addition of an NO group to protein thiols, a modification known as SNO. The same three approaches can be used to identify SNO. As with phosphorylation, an enrichment strategy is needed. Enrichment of SNO proteins is often done by using a modification of the biotin switch.25 In the biotin switch method, the free cysteines are first blocked with an alkylating reagent (MMTS) and nitrosylated cysteines are then selectively reduced using ascorbic acid and labelled with biotin, which is detected or enriched by an anti-biotin antibody or on an avidin column.

Biotinylated proteins enriched on an avidin column can be eluted, digested, and analysed by MS. Alternatively, biotin-labelled proteins can be run on a 1D gel and excised for MS identification. Using biotin switch, the nitrosylated proteins are readily identified, whereas the site determination is still difficult. A few other enrichment methods based on biotin switch have also been reported. ICAT has been used to measure cysteine oxidation26 and ICAT linking via a modified biotin switch is a possible approach for identifying SNO-labelled proteins. The ICAT label can then be detected in MS allowing identification of SNO proteins. A promising new method termed SNO-RAC27 has also been recently described to identify SNO proteins. The SNO-RAC approach for SNO is based on affinity enrichment of cysteine containing proteins or peptides by using a thiol-reactive resin.27 With suitable anti-SNO antibodies, immunoprecipitation is another approach for enrichment.

SNO can be followed using 2D gel methods. Unlike phosphorylation, SNO does not add sufficient charge or weight to the protein to result in a separate spot. However, using a modified biotin switch method, one can label SNO containing cysteines with a charged fluorescent dye which results in a shift. It is of note that the biotin switch method as originally described (using biotin HPDP) is not compatible with 2D gels because the biotin HPDP linkage is lost under the reducing conditions used in 2D gels. The biotin switch method has been modified to use a maleimide linked typically to a fluorescent dye and this modified biotin switch methods has been used in 2D gels to identify SNO proteins in whole-cell extracts.6,28 If a DyLight dye, which is not charge compensated, is used, this will result in a shift in the DyLight label proteins (which were originally the SNO) from the unmodified proteins, and this can be used to enrich and identify the SNO protein in a separate spot. The fluorescent dye allows one to quantify the relative amount of SNO to total proteins. One can also use charge-compensated dyes that will bind to SNO groups using a modified biotin switch. In this case, the dye-labelled SNO protein will overlap with the unmodified protein.

There is another mass spectrometric method for nitrosylation/site localization that utilizes a neutral loss of 28 Da in a triple quadrupole mass spectrometer.29 In this technique, the instrument monitors the loss of the NO group and as soon as such a peptide is detected, the MS/MS spectrum is acquired which then enables the identification of the nitrosylated peptide.

2.2.3 Other mitochondrial post-translational modification

There are a number of other modifications which appear to occur on mitochondrial proteins, including acetylation, O-linked-β-N-acetylglucosamine glycosylation (O-GlcNAc), glutathiolation, sumoylation, and ubiquitination. These modifications are reviewed elsewhere.26,3032

2.3 Quantitation

Quantitation between samples can be done using a label-free approach in which samples are run in sequence and care is taken that there are no changes in the instrumental conditions between the runs. The peptide's intensities are compared among the runs. Label-free approaches are useful when the sample amount is limited, they are inexpensive and require little or no sample handling; however, the coefficients of variation associated with these measurements are in most cases higher than labelling methods. It is noteworthy that all these comparisons are relative, absolute quantitation requires isotopically labelled peptide or protein standards.

With labelling methods, the samples are labelled with stable isotopes (or fluorescent dyes for the 2D DIGE) before they are mixed together. The earlier the labelling takes place, the more precise are the ratios. The techniques include: stable isotope labelling with amino acids in cell culture (SILAC), isobaric tag for relative and absolute quantitation (iTRAQ), tandem mass tags, and isotope tagging (e.g. 18O labelling). SILAC and iTRAQ are the most common labelling technologies. However, SILAC requires dividing cells to take-up the label, and as cell division is limited in cardiac cells, the SILAC approach is less amenable to labelling cardiac cells. Macek et al.33 have published an account on quantitative phosphoproteomics where they describe different quantitative methods and their applications. The same group applied quantitative proteomics approaches to study the differences in rat mitochondria from muscle, heart, and liver.34

3. Mitochondrial proteome

There are a number of papers which have characterized the general mitochondrial proteome.3437 Data from these studies are complied in databases such as MitoCarta (http://www.broadinstitute.org/pubs/MitoCarta/index.html), MitoP38 (http://www.mitop.de:8080/mitop2/), and Mitominer39 which list proteins identified as mitochondrial (using GFP, proteomic 2D gel and MS, and other methods). There are also a number of studies that have focused on defining the cardiac mitochondrial proteome.34,36,37,40,41

3.1 Distinguishing contaminants from proteins that might transiently associate with mitochondria

Most mitochondrial proteomic studies find some proteins which are thought to be primarily cytosolic or ER/SR proteins. This raises the issue of whether these are contaminants that co-purify with the mitochondria or whether they are proteins that can localize or translocate to the mitochondria. There are an increasing number of studies on target proteins that report mitochondrial localization of proteins, which are commonly not considered to be mitochondrial. These include proteins such as STAT3,42 Cx43,43 GAPDH,44 PKCε,3,45,46 AKT, hexokinase, and other kinases and phosphatase. In these studies, antibodies were usually used to identify the protein and it is usually shown that the protein is enriched with mitochondrial purification. Most studies with a focus on defining the mitochondrial proteome will take steps to have a pure mitochondrial preparation, but given the connections between mitochondria and ER unless one removes the outer mitochondrial membrane, ER as well as plasma membrane proteins are represented in the mitochondrial fraction.14 Trypsin or other proteases can be used to distinguish matrix proteins from those proteins attached to the mitochondria.47 Studies with GFP labelling can be used to determine whether the ER proteins can be found in the mitochondria (inner membrane space or matrix). In contrast, if you want to determine whether some treatment alters a PTM of an accepted mitochondrial protein, it is preferable to use freeze-clamped tissue to better preserve the PTM. However, without some fractionation, it might be difficult to observe proteins that are in low abundance in a whole-cell extract from freeze-clamped tissue; thus it is useful to immunoprecipitate the protein of interest or to fractionate the sample in some way to enrich for the protein of interest.

4. Changes in the cardiac mitochondrial proteome and cardioprotection

4.1 Changes in protein amount

Many studies have shown that cardioprotection leads to the translocation of kinase and other signalling proteins to the mitochondria. As discussed in Section 3.1, non-mitochondrial proteins can occur as contaminants in mitochondrial fractions, and these factors need to be considered in interpreting these studies. Table 1 lists the studies that have been done examining changes in cardiac mitochondria with cardioprotection. Hexokinase was reported to exhibit an increase in mitochondrial localization following ischaemic preconditioning or pharmacological preconditioning.48 The increase in hexokinase is reported to inhibit activation of the MPT. Several groups have reported an increase in mitochondrial PKCε with cardioprotection.2,3,45,49 It has recently been shown that PKCε is taken up into the mitochondria via an HSP90-dependent import mechanism.46 PKCε has a number of mitochondrial targets; it has been reported to form a complex with VDAC, ANT, and HKII,2 as well as a complex with MAPK.3 PKCϵ has been shown to result in phosphorylation of ALDH2 which has been shown to reduce infarct size.50 PKCε has also been reported to phosphorylate cytochrome c oxidase subunit IV,51 to activate mitochondrial K-ATP channel, and to lead to inhibition of the MPT. PKCδ is also reported to translocate to the mitochondria on reperfusion leading to inhibition of PDH and increased cell injury.49 Cx43 also translocates to the inner membrane of mitochondria with cardioprotection.52 This translocation has been shown to be dependent on the HSP90-TOM import pathway.52 Mitochondrial Cx43 has been shown to stimulate the mitochondrial K-ATP channel.53 AKT has been reported to translocate to the mitochondria with cardioprotection where it can phosphorylate a number of target proteins.5456 The mitochondrial localization of GSK-3 has also been reported to increase with cardioprotection, leading to inhibition of the MPT.57 However, Clarke et al.14 have questioned the mitochondrial translocation with cardioprotection of some of these kinases.

View this table:
Table 1

Changes in mitochondrial proteins with cardioprotection

Adenosine- or diazoxide-treated rat myocytesArnell et al.52D gel
Mitochondria following exercise trainingKavazis et al.71Isotope tag-MS
PKCε transgenic mouse heartMayr et al.642D-DIGE
Mitochondrial—PC or GSK inhibitor treated heartWong et al.672D-DIGE
Cardioplegia in rabbit heartMcCully et al.82iTRAQ
Exercise-induced protectionBoluyt et al.652D gel
Isofluorane-induced protectionFeng et al.702D BN PAGE and IMAC/MS
Nitrite-mediated cardioprotectionPerlman et al.632D gel
Metallothionine-mediated protectionMerten et al.662D gel
Inhibition of p38 MAPK-mediated protectionSchwertz et al.692D gel
Diazoxide-mediated protectionLi et al.832D gel
Preconditioning in rat heartClarke et al.142D-DIGE

A large and growing number of mitochondrial proteins appear to have dynamic phosphorylation,58,59 suggesting a role for mitochondrial kinases and phosphatase. Although a number of kinases have been reported to be localized to the mitochondria, only a few phosphatases have been reported in the mitochondria.60 In addition to proteins which have been reported to translocate to the mitochondria with cardioprotection, a number of other proteins that were categorized as non-mitochondrial have been reported to have a mitochondrial localization and to regulate mitochondrial function. For example, STAT3 has been reported to have a mitochondrial localization and to regulate the activity of the electron transport chain.42 The mechanism by which STAT3 might alter electron transport is unclear, but it is unlikely to be via binding to mitochondrial electron transport complexes, because the levels of STAT3 in the cell are too low to inhibit electron transport chain complexes by stoichometric binding.61

In addition to the targeted approach, a number of recent studies have used a large-scale proteomics approach to unravel in an unbiased fashion mitochondrial proteins that are important in cardioprotection. Cardioprotection is generally thought to involve inhibition of the MPT, but the identity of the MPT and many of the proteins that modify the function of the MPT are largely unknown. One approach is to determine changes in the mitochondrial proteome under conditions of cardioprotection to see whether this can provide insight into common signals. Table 1 lists several studies which have taken this approach. Are there common alterations in the mitochondrial proteome with cardioprotection? There are protein changes that appear in several studies; however, it is too early to tell if these are important changes in cardioprotection or just common stress responses of abundant mitochondrial proteins. We also need to consider separately long-term cardioprotection (protection in females, exercise) which likely involves altered protein expression versus acute cardioprotection, which likely involves altered PTM or localization of proteins. Some common targets include ALDH2,50,62,63 PDH,5,62,64 ATP synthase d,5,65 cytochrome c oxidase subunit Va,6668 components of the malate–aspartate shuttle,64,69,70 prohibitin,5,70 peroxiredoxin 3 or 6,66,69,71 VDAC,4,69 isocitrate dehydrogenase,5,62,64 and various heat shock proteins.5,65 Interestingly, Kavazis et al.71 found that exercise-induced protection is associated with a decrease in MAO-A, a protein which Di Lisa and colleagues72 find to be important in cardioprotection.

Some clear pathways that are altered include regulation of redox (ALDH2, MAO-A, peroxiredoxin), regulation of PDH and the malate–aspartate shuttle, components of the F1F0-ATPase, and components of cytochrome c oxidase and complex 1. Do these proteomic studies provide insight into the MPT or its regulation? This is less clear. Some of the proteins might be potential candidates (i.e. they span the inner mitochondrial membrane). However, additional studies will be needed. It is also surprising that none of the proteins shown to translocate to the mitochondria using a candidate approach has been found in large proteomic studies (e.g. AKT, GSK, PKC, Cx43, HK). This likely reflects the low abundance of these signalling proteins. For example, STAT3 (in the cell) was recently detected by MS and shown to be present at very low levels.61 Clearly, proteomics methods are rapidly changing and there have been great advances in the coverage of low abundant proteins. It is likely that this approach will provide important new information in the future.

As illustrated in Table 1, the majority of broad-based proteomic studies to date have used 2D methods to examine protein differences between samples. However, shotgun approaches using isotopic labelling (iTRAQ) and IMAC or TiO2 for phospho-enrichment are becoming more common and provide more complete list of mitochondrial proteins including more low abundant proteins and better coverage of membrane proteins and proteins with high MW and at pI extremes.

4.1.1 Changes in protein levels vs. changes in localization

To examine WT vs. KO or the effect of a drug on the mitochondrial proteome, it is important to consider whether these changes alter the percentage of mitochondria in the cell or the percentage of mitochondria isolated. This can be done by noting whether there are changes in mitochondrial yield (if studies are done on isolated mitochondria) or whether there is a general up-regulation of mitochondrial proteins in a whole-cell study. We also need to consider changes in location (a protein might appear to change in expression in the mitochondria when it just changes location). Furthermore, even in whole-cell extract, there is some pelleted (insoluble material) that is not included and if proteins translocate to or from this location, it would appear as a change in protein expression.

Thus in studies comparing mitochondrial protein levels under different conditions (e.g. WT and KO or +/− a drug), a change in protein level could be due to: (i) a difference in protein levels or (ii) a difference in subcellular localization (e.g. a protein either moves in or out of the mitochondrial matrix or associates or disassociates with the outer membrane). In studies where the treatment time is very short, differences in levels are most likely due to differences in the rate of protein degradation (or a change in localization) rather than a change in protein synthesis.

4.2 Phosphorylation of mitochondrial proteins and cardioprotection

A number of studies have identified cardiac mitochondrial phospho-proteins.5,14,20,58,59,62,67,70,73 A decrease in phosphorylation of VDAC2 has been suggested to occur as a result of inhibition of GSK.4 This decrease in phosphorylation of VDAC has been proposed to reduce ATP entry into the mitochondria thereby reducing consumption of glycolytic ATP. GSK has been reported to phosphorylate cyclophilin D and correlate with opening of the MPT.74 Arrell et al.5 find an increase in phosphorylation of the beta-subunit ATP synthase with adenosine treatment; functional effects of phosphorylation were examined in a follow-up study.75 PKCε has been shown to phosphorylate a number of mitochondria proteins, such as ALDH250,62 and cytochrome c oxidase subunit IV,51 leading to cardioprotection. Wong et al.67 have data showing increased phosphorylation of cytochrome c oxidase subunit VIb and subunit Va with cardioprotection. The functional consequences of phosphorylation of these subunits are not clear, but they appear to be associated with alterations in super-complex composition. However, Clarke et al.14 found no change in the phosphorylation of mitochondrial proteins with preconditioning. Clarke et al.14 did find changes in protein carbonylation. It is possible that the studies showing a change in phosphorylation or other PTM of proteins that are not obvious mitochondrial proteins (e.g. proteins that are reported to translocate to the mitochondria) were due to contamination by non-mitochondrial proteins. Alternatively, it is possible that changes in the mitochondrial phosphoproteome (particularly in low abundant signalling molecular) are small and were not detected in the study by Clarke et al.14 Additional studies using a more sensitive direct MS approach might help determine whether cardioprotective agents do indeed result in changes in the mitochondrial phosphoproteome. As discussed earlier, the significance of PTMs affecting only a small fraction of a protein may have a limited effect on enzyme activity (depending on the dynamic range), but if there is a targeted localization of a signalling molecule, a PTM in a small fraction of the protein could have a significant functional impact. Similarly, activation of small percentage of a membrane channel could have significant functional impact.

4.3 SNO and cardioprotection

Nitric oxide has been shown to be an important mediator of cardioprotection. Burwell et al.11 and Sun et al. 6 observed an increase in S-nitrosothiols with preconditioning in mitochondria and heart extracts, respectively. Cardioprotection has been shown to result in an increase in SNO of several mitochondrial proteins, including complex I,6,11,76 the alpha subunit of the F1F0-ATPase,6,9 alpha-ketoglutarate dehydrogenase,6,28 aconitase,6,9,28 creatine kinase,6,9 malate dehydrogenase,6,9 heat shock protein 60,6,9 heat shock protein 70,9 electron transfer flavoprotein alpha6 and beta,9 electron transferring flavoprotein dehydrogenase,28 subunit 5A of cytochrome c oxidase,9 ALDH2,28 succinate dehydrogenase subunit A flavoprotein, very long chain and short chain acyl CoA dehydrogenase,28 CPT2,28 enoyl CoA hydratase,28 isocitrate dehydrogenase,28 and GRP78.28

The functional consequences of SNO have been examined for only a few proteins. SNO appears to lead to inhibition of many proteins,28 although SNO and/or glutathiolation of SERCA2a has been shown to increase its activity.6,77 SNO is reported to inhibit complex I and results in less ROS generation during ischaemia.11 SNO has been reported to inhibit the F1-F0 ATPase, which will preserve glycolytically generated ATP and will reduce the mitochondrial membrane potential thereby reducing the uptake of Ca into the mitochondria resulting in less matrix Ca to activate the MPT. Sun et al.6 examined SNO in preconditioned hearts and also found SNO of non-mitochondrial proteins such as the L-type Ca channel and SERCA2a, which regulate cell Ca and are proposed to reduce mitochondrial Ca overload. Some additional mitochondrial proteins such as ANT, complex I,76 and complex IV have been shown to be SNO by a more targeted approach.

Interestingly, mitochondria comprise a large percentage of SNO proteins in studies on whole-cell extracts. Possible reasons for this include the increased stability of N2O3 in the hydrophobic milieu of the mitochondria, which would favour SNO, the high level of reactive cysteines in mitochondrial proteins, and possible difference in mitochondrial redox. As NO is labile, the high level of SNO of mitochondrial proteins suggests a nearby source of NO, such as the endoplasmic/sarcoplasmic reticulum or perhaps a mitochondrial-localized NOS. However, the existence of mitochondrial localized NOS is controversial.7880 Protein SNO in mitochondria can occur as a result of trans-nitrosylation from low-molecular-mass SNO such as GSNO.81

5. Summary

Figure 2 summarizes potential mechanisms by which alterations in mitochondrial proteins and PTM might lead to cardioprotection. Elucidating the role of the mitochondria in cardioprotection will require a combination of examining candidate proteins as well as unbiased proteomic studies. The large-scale proteomics approach with the use of fractionation/enrichment is useful to identify unexpected protein changes. However, for low abundance signalling molecules, a targeted approach is necessary.

Figure 2

Alterations in the mitochondrial proteome associated with cardioprotection. Cx43, PKCϵ, and AKT have been reported to translocate to the mitochondria. PKCϵ has been suggested to phosphorylate ALDH2 and αKGDH resulting in less ROS generation. PKC has been reported to form complexes with a number of other proteins. Unphosphorylated (active) GSK is reported to phosphorylate cyclophilin D promoting activation of the MPT. AKT and ERK are reported to phosphorylate and inhibit GSK. Inhibition of GSK is also reported to reduce ATP entry into the mitochondria via the VDAC, which will reduce consumption of glycolytically generated ATP. Nitric oxide is reported to increase SNO of complex I thereby reducing ROS generation. SNO inhibits complex V resulting in degradation of ATP by reverse mode of the ATPase. SNO is also reported to block irreversible oxidation of proteins.

Conflict of interest: none declared.


This work was supported by the National Heart, Lung, and Blood intramural program of the National Institutes of Health.


  • This article is part of the Review Focus on: Mitochondria in Cardiac Disease: Emerging Concepts and Novel Therapeutic Targets


View Abstract