VDAC proteomics: post-translation modifications

Abstract

Voltage-dependent anion channels are abundant mitochondrial outer membrane proteins expressed in three isoforms, VDAC1-3, and are considered as “mitochondrial gatekeepers”. Most tissues express all three isoforms. The functions of VDACs are several-fold, ranging from metabolite and energy exchange to apoptosis. Some of these functions depend on or are affected by interaction with other proteins in the cytosol and intermembrane space. Furthermore, the function of VDACs, as well as their interaction with other proteins, is affected by posttranslational modification, mainly phosphorylation. This review summarizes recent findings on posttranslational modification of VDACs and discusses the physiological outcome of these modifications.

1. Introduction

Beyond being the “powerhouses” of the cell, mitochondria are in constant communication with the rest of the cell. This bidirectional communication involves the continuous exchange of material and information. Each mitochondrion is bounded by two membranes, outer and inner, yielding six distinct domains, the outer membrane, inner boundary membrane, intermembrane space, cristal membrane, intracristal space, and matrix. The exchange between the cytosol and mitochondrial matrix can take place because the inner mitochondrial membrane contains specific transport proteins that catalyze the uni- or bidirectional flow of metabolites and ions [1, 2]. The mitochondrial outer membrane contains channel-forming proteins, which harbor β-barrel structures, i.e., Tom40, Sam50, Mdm, and VDAC [3, 4], of which the latter is the most abundant. Signaling to and from the mitochondria involves posttranslational modification of both mitochondrial and extramitochondrial proteins as well as translocation/association of proteins with mitochondria. A detailed compilation of mitochondria-associated kinases, phosphatases, and other proteins is presented in Supplemental Table 1 in reference [5].

In mammals, voltage-dependent anion channels (VDACs) are expressed in three different isoforms [6]. By comparing the expression of VDAC isoforms in rat liver, kidney, heart, and brain, it was concluded that in mammalian tissues VDAC1 is the most abundantly expressed isoform [7]. VDACs present in abundance in the mitochondrial outer membrane also are abundant in the plasma membrane. Indeed, VDAC, first isolated from human lymphocytes by Thinnes et al [8], recently was shown to be present in the plasma membrane of a variety of cells [9], and to have manifold functions in the mitochondrial outer membrane. By forming pores, VDACs are involved not only in metabolite and energy exchange between the cytosol and mitochondria [10], but possibly in formation of the permeability transition pore [11]. Furthermore, VDACs also are docking sites for cytosolic proteins such as hexokinase [12], cytoskeletal proteins such as tubulin [13], and microtubule-associated protein 2 (MAP2) [14], as well as proapoptotic proteins [15, 16]. The aforementioned functions of VDACs are discussed in detail in different reviews presented in this issue of Biochimica Biophysica Acta.

Despite the critical involvement of VDACs in various mitochondrial functions, relatively little is known as to how VDACs are regulated and how they fulfill their roles. In addition to expression of three isoforms and tissue-specific splice variants of VDAC3 [17, 18], a possible mechanism for the diverse physiological roles of these channel proteins is that of reversible posttranslational modification. However, only very few mass spectrometric studies targeting posttranslational modifications of the different VDAC isoforms currently are available. What information is available on posttranslational modifications, specifically phosphorylation and acetylation, of VDACs is based on studies directed at posttranslational modification of VDACs [19] or on shotgun proteomic approaches on fractions enriched in phosphopeptides and acetyllysine-containing peptides [20-27]. These studies do not relate posttranslational modifications to altered function. Phosphorylation of VDACs often is inferred from studies employing immunoblotting that use anti-phosphoserine, -threonine, and -tyrosine antibodies analysis [vide infra].

In this review, we summarize findings in the literature regarding the different posttranslational modifications of VDACs and discuss the effect of modification on their function.

2. Posttranslational modification of VDACs

Reversible posttranslational modifications (PTM) play an important regulatory role in mitochondrial function, among others. Such modifications do not simply affect enzymatic activity by switching activity on and off but also impact degradation of proteins, translocation of proteins from one compartment to another within the cell, as well as docking or interacting with other proteins. In addition, PTM is critical in cell signaling processes. Among the numerous posttranslational modifications, reversible phosphorylation is the most frequent and best studied. It has been estimated that more than one-third of all proteins are phosphorylated, either at one or multiple sites [28, 29]. Protein phosphorylation is catalyzed by a family of protein kinases, with serine, threonine, and tyrosine side chains being the most frequently modified [30]. The reverse reaction is catalyzed by specific protein phosphatases, i.e., phosphoserine/phosphothreonine-specific and phosphotyrosine-specific families of phosphatases [31].

Another frequent PTM is protein acetylation, with the number of acetylated proteins rivaling that of phosphorylated proteins. The two types of acetylation are N-terminal acetylation and lysine side-chain acetylation. The former is irreversible and occurs co- and posttranslationally and is catalyzed by N-acetyltransferases. N-terminal acetylation is prevalent in higher organisms and occurs on N-terminal methionine or on the newly terminal amino acid residue after removal of the N-terminal methionine [32-34]. The biological significance of N-terminal acetylation has not been fully elucidated; it is thought to affect protein stability, assembly, and location [34]. Reversible acetylation of lysine ε-amino groups is catalyzed by protein acetyltransferases (PATs) and deacetylases or histone deacetylases (HDACs) and takes place posttranslationally [35, 36]. The latter proteins are grouped into class I and II HDACs; class III deacetylases are termed “sirtuins“ [36]. The significance of reversible lysine acetylation of histone proteins is well-appreciated; its importance in intermediary metabolism is just becoming apparent [26, 37-39].

2.1 Phosphorylation of VDACs determined by mass spectrometry

There are only a very limited number of studies directed at PTM of VDAC, including that of phosphorylation, by means of mass spectrometry. The sites modified on VDAC isoforms are listed in Table 1 along with candidate protein kinases catalyzing the phosphorylation and the functional outcome of phosphorylation where determined. A possible reason for the limited number of studies is that VDACs are highly hydrophobic integral mitochondrial outer membrane proteins and difficult to work with. To overcome this obstacle, we developed a method to study the PTM of integral rat liver mitochondrial outer membrane proteins [40]. The method is based on enrichment of the proteins by first harvesting high-purity outer membrane from rat liver mitochondria, followed by isolation of the proteins of interest by semipreparative SD-PAGE/electroelution. Applying this method to VDACs in rat liver mitochondria, we have shown that all three VDAC isoforms are expressed in these organelles. In addition, all three isoforms are phosphorylated at one or more sites [19]. As shown in Table 1, VDAC1 is phosphorylated on serine 12 and 136, corresponding to CaM-II/GSK3 and PKC consensus sites, respectively [19]. These in situ findings on VDAC1 phosphorylation are supported by in vitro studies. It has been shown that GSK3 is able to phosphorylate isolated VDAC; neither the isoform nor the phosphorylated residue was identified [41]. In another study using in vitro transcribed and translated VDAC (isoform not specified), it was shown that GSK3β phosphorylates wild-type VDAC, but not the mutated form where threonine 51 was altered to alanine [42]. Regarding the phosphorylation of VDAC by PKC, it was shown that PKCε directly interacts with and phosphorylates VDAC1 [43]; the amino acid residue phosphorylated by PKCε was not identified.

Table 1.

Phosphorylation of VDAC isoforms determined by mass spectrometry (MS) or site-directed mutagenesis (SDM) and functional consequences of VDAC phosphorylation.

Isoform	Phosphorylation (species/organ)	Method [Ref.]	Candidate protein kinases#	Effect of phoshorylation [Ref.]
VDAC1	S12, 136 (RLM)	MS [19]	GSK3/CaM-II/CKI	Not determined
	S12, 103 (HMEC)	SDM [48]	GSK3/CaM-II/CKI; CKI	Sensitization to apoptosis via inhibition of phosphorylated VDAC1 degradation [48].
	S117 (MB; MLM)	MS [20, 21]	CKI/p38MAPK	Not determined
	S101, 102, 104 ; HeLa cells	MS [44]	GSK3/cdc2/CaM-II; PKC/cdc2/GSK3; CKI/p38MAPK/GSK3	Not determined
	S193 (HK2)	SDM [51]	Nek1/PKC	Prevention of cell death [51].
	T107; HeLa cells	MS [44]	PKC/GSK3/CaM-II	Not determined
	Y80, 208; MB	MS [22]	SRC/EGFR; EGFR/SRC/INSR	Not determined
VDAC2	S115; HeLa cells	MS [44]	CKI/cdk5/GSK3	Not determined
	T118; HeLa cells	MS [44]	PKC/CKI/GSK3	Not determined
	Y237; RLM	MS [19]	INSR/EGFR/SRC	Not determined
	Y207; MB	MS [22]	EGFR/SRC/INSR	Not determined
VDAC3	S241; RLM	MS [19]	PKA/GSK3/CaM-II	Not determined
	T33; RLM	MS [19]	PKC/CaM-II/GSK3	Not determined
	Y49; MB	MS [22]	CKII/GSK3/CaM-II	Not determined

Abbreviations: RLM=rat liver mitochondria; MLM=mouse liver mitochondria; MB=mouse brain; HMEC=human microvascular endothelial cells; HK2=humaAbbreviations: RLM=rat liver mitochondria; MLM=mouse liver mitochondria; MB=mouse brain; HMEC=human microvascular endothelial cells; HK2=human proximal renal tubular epithelial cells; HK=hexokinase.

^#Determined using NetPhosK 1.0 Server.

In addition to documentation of serine 12 and 136 phosphorylation, VDAC1 in mouse liver mitochondria was found in two other studies to be serine-phosphorylated, and in a study of HeLa cells serine- and threonine-phosphorylated. Employing an improved immobilized metal affinity chromatography (IMAC) method to examine mouse liver mitochondrial phosphoproteome, Lee et al. documented the phosphorylation of serine 117 [21]. The same serine residue was found by Munton et al. to be phosphorylated in synaptosomal preparations [20]. In HeLa cells, the same peptide is phosphorylated on serine 101, 102, and 104, as well as on threonine 107 [44]. All three studies used shotgun proteomic approaches; the phosphopeptide-enriched fractions were analyzed by mass spectrometry. In a study that used a shotgun approach, VDAC1 was shown to be phosphorylated on tyrosine residues 80 and 208 [22]. In the latter report, tyrosine-phosphorylated peptides were enriched by employing immobilized antiphosphotyrosine antibodies and the resultant fraction analyzed.

Besides VDAC1 being the most abundant isoform, VDAC2 phosphorylation was demonstrated in three studies. Olsen et al. demonstrated that in HeLa cells VDAC2 is phosphorylated on serine 115 and threonine 118 [44]. Deng et al. have reported that in rat liver mitochondria VDAC2 is phosphorylated on threonine 109 [23]. However, the reported amino acid sequence (LpTLDTTSSPNTGK) does not match any protein in the ExPASy Proteomics Server's database. We have shown in our mass spectrometric study targeting VDACs that tyrosine 237 is phosphorylated [19]. In a subsequent study of the tyrosine phosphoproteome in mouse brain, Ballif et al. documented tyrosine phosphorylation on residue 207 [22].

Phosphorylation of VDAC3 in rat liver mitochondria of serine 241 and threonine 33 was previously shown by us [19]. These sites correspond to PKA and PKC consensus sites. As mentioned above, PKCε can be recruited to mitochondria where it associates with and phosphorylates VDAC1. Whether or not it also phosphorylates VDAC3 remains to be determined. PKA is known to be present in mitochondria – in vitro studies have shown that it phosphorylates VDAC resulting in altered gating properties [45, 46]. In rat brain, VDAC3 also is tyrosine-phosphorylated at residue 49 [22].

Table 1 is a summary of in situ phosphorylation of VDAC isoforms along with possible candidate protein kinases and references. These studies were directed at the determination of phosphorylation sites in VDACs or examination of the mitochondrial phosphoproteome using mass spectrometry. The functional relevance of these posttranslational modifications by phosphorylation was not addressed.

2.2 Effect of VDAC phosphorylation on its function

VDAC purified either from rat liver or rat brain mitochondria is phosphorylated in vitro by cAMP-dependent protein kinase A (PKA) [45, 46]. Following reconstitution of VDAC into planar lipid bilayers, phosphorylation by PKA reduces the channel current. These in vitro data reinforce the importance of protein phosphorylation in the regulation of VDACs but do not reveal which amino acid residue is phosphorylated.

Relatively few in vivo or in situ studies on VDACs phosphorylation are available. Most studies linking phosphorylation of VDAC to function employed immunoblotting with antiphospho-serine, -threonine, and -tyrosine antibodies; a few reports used phosphosite-specific antibodies. Most of these studies focused on VDACs' role in apoptosis. Although VDACs are not considered to be essential components of the mitochondrial permeability pore (mPTP) [47], there is substantial evidence for at least an indirect role of VDAC phosphorylation in apoptosis. Yuan et al. have shown that endostatin, a powerful inhibitor of angiogenesis in tumors, promotes the opening of the mitochondrial permeability transition pore via a mechanism that involves VDAC1 [48]. Accordingly, silencing or overexpression of VDAC1 respectively attenuates or increases the sensitivity of endothelial cells to endostatin-induced apoptosis. Simultaneously with promoting the opening of PTP, endostatin decreases the amount of hexokinase 2 and increases the content of VDAC1. Because hexokinase binding to VDAC decreases its phosphorylation, the authors speculated that endostatin-induced increase in VDAC1 is due to phosphorylation and consequent inhibition of its degradation. Indeed, mutation of serine 12, previously shown by us to be phosphorylated in situ, and serine 103 effectively attenuated the endostatin-induced upregulation of VDAC1, and abolished PTP opening and caspase 3 activation in endothelial cells. Furthermore, in support of inhibition of VDAC1 degradation by phosphorylation, the effect of serine 12 and 103 mutation to alanine VDAC1 level was inhibited by the proteasome inhibitor, MG132. A polar opposite effect of VDAC1 phosphorylation on cell death was reported by Sun et al. [49]. According to these authors, overexpression of both hexokinase isoforms (hexokinase I and II) in HEK293 cells protected against cell death and inhibited mPTP opening in response to hydrogen peroxide. Furthermore, hexokinase expression increased VDAC phosphorylation, which was significantly decreased by a PKCε inhibitor peptide. These authors proposed that PKCε-dependent VDAC phosphorylation protects against cell death. However, the VDAC isoform that becomes phosphorylated in a hexokinase-dependent manner by PKCε has not been specified nor the phosphorylated amino acid residue determined.

The involvement of VDACs in cell death has been suggested by studies using the ischemic reperfused heart. Subjecting rabbit hearts to ischemia by ligation of the left descending coronary artery for 60min followed by reperfusion results in necrosis and apoptosis that is prevented by the p38 MAP kinase inhibitor, PD169316 [50]. Associated with increased apoptosis was a significant increase in tyrosine phosphorylation of VDAC1, as assessed by immunoblotting with phosphotyrosine antibodies. This increase in tyrosine phosphorylation is prevented by the p38 MAP kinase inhibitor; concomitant with inhibition of p38 there was significant cardioprotection. Because p38 MAP kinase is a serine/threonine kinase, explanation of tyrosine-phosphorylation of VDAC1 requires the involvement of a downstream tyrosine kinase. Alternatively, the protective effect of p38 MAP kinase inhibitor might be due to phosphorylation of serine/threonine residue(s), since only tyrosine phosphorylation was investigated in the study. Neither the downstream tyrosine kinase nor the tyrosine residue that becomes phosphorylated during ischemia/reperfusion was identified. These data emphasize the involvement of VDAC1 in apoptosis.

Findings with HeLa cells are consistent with the proapoptotic nature of phosphorylated VDAC1 [42]. The protective effect against cell death of hexokinase binding to mitochondria is mediated via Akt, a protein kinase also known to localize to the mitochondria and to be regulated by the upstream phosphatidylinositol-3-kinase. Activation of Akt phosphorylates GSK3β, also known to localize to mitochondria, inhibits its enzymatic activity. As a consequence, GSK3β cannot phosphorylate VDAC, thus permitting hexokinase-binding to VDAC. Accordingly, preventing the activation of Akt by inhibiting the upstream phosphatidylinositol-3-kinase by wortmanin, or inhibiting Akt by the Akt inhibitor IV disrupted hexokinase binding to mitochondria. Immunoprecipitation of VDAC1 from cells treated with either wortmanin or Akt inhibitor IV followed by immunoblotting with anti-phosphothreonine antibodies revealed strong reactivity with VDAC1. Phosphorylation of VDAC was abolished when GSK3β was silenced or threonine 51 was mutated to alanine (threonine 51 is in GSK3β phosphorylation consensus). Although these studies are consistent with phosphorylation of VDAC1, the question remains as to which amino acid in VDAC is phosphorylated. Phosphorylation of serine residues also was not assessed in the study despite GSK3β being a serine/threonine kinase.

In contrast to the foregoing report, Das et al. implicated VDAC2 rather than VDAC1 in ischemia/reperfusion injury, and proposed a mechanism that does not involve opening of the mPTP [41]. Using the Langendorf perfused rat heart, the authors demonstrated significantly improved recovery of cardiac function in the presence of the GSK inhibitors, SB 216763 and SB 415286. The improvement in recovery was associated with decreased phosphorylation of VDAC, as evaluated by immunoblotting (isoform-specificity of the VDAC antibodies not specified). Following in-gel digestion of the relevant spots, mass spectrometric analysis revealed the presence of VDAC2. It is not clear from the data presented whether other proteins, especially the other VDAC isoforms, also were present in the excised gel plugs. Furthermore, direct demonstration of VDAC2 phosphorylation, i.e., identification of the phosphorylated residue(s), was not provided.

At variance with the proposed proapoptotic feature of VDAC phosphorylation are findings by Chen et al. [51]. Studying the mechanism by which mammalian NIMA-related protein kinase (Nek1) protects against cell death, the authors found that protection against apoptosis is contingent on VDAC1 phosphorylation on serine residue 193, which is easily accessible from the cytosolic side of the mitochondrial outer membrane. Accordingly, ectopic expression of the non-phosphorylatable S193A VDAC1 mutant resulted in cell death. Conversely, ectopic expression of the S193Q VDAC1 mutant, in which phosphorylation is mimicked by glutamate, prevented mitochondrial membrane permeabilization and averted cell death. Evidence for Nek1 as the protein kinase responsible for VDAC1 phosphorylation was provided by expressing a dominant-negative, kinase-dead Nek1 mutant (K33A) in different cell lines. Expression of the mutant NEK1, but not the wild-type, resulted in mitochondrial cell death. This is preceded by lack of serine 193 phosphorylation, as determined by immunoblotting with antibodies, which specifically recognize the serine 193 phosphorylated form of VDAC1. Although Nek1 is localized predominantly in the cytosol, a significant fraction is found associated with mitochondria. The direct physical interaction between Nek1 and VDAC1 was demonstrated using a yeast two-hybrid system and reciprocal immunoprecipitation using anti-VDAC1 and anti-Nek1 antibodies.

In addition to the aforementioned studies, phosphorylation of VDACs has been reported to occur in pathological conditions. In postmortem brain from Alzheimer and Down syndrome patients, Yoo et al. using 2-dimensional electrophoresis and MALDI-Tof, described an increase in the content of three VDAC1 isoforms, which differed in their isoelectric points [52]. However, no information regarding the type and site of phosphorylation was presented. In hippocampal extracts of transgenic mice with Alzheimer's disease, there was a significant increase in the content of both VDAC1 and VDAC1 threonine-phosphorylated at an epitope susceptible to phosphorylation by GSK3β [53]. At the same time, the content of mitochondria-associated hexokinase I was decreased, rendering the mitochondria more prone to release proapoptotic factors through VDAC. In cultures of SH-SY5Y neuroblastoma cells, Aβ_1-42 (a amyloid-β-derived diffusible ligand) inhibited Akt activation and consequently prevented phosphorylation and inactivation of GSK3β, resulting in increased phosphorylation of VDAC1. These data suggest that VDAC1 phosphorylation is involved in the genesis of apoptosis in brain of Alzheimer disease patients.

These studies show a relationship between phosphorylation of VDACs and apoptosis but not a direct cause-effect relationship. Several questions regarding the isoform of VDACs, the site and type of VDAC phosphorylation, as well as the mechanism of apoptosis still remain open. If VDAC is not a component of the mPTP, how does VDAC phosphorylation lead to apoptosis? Clearly, further studies are required. Knowledge of all sites phosphorylated under physiological/pathophysiological conditions in VDACs, to be determined by mass spectrometric methods, should provide a platform for logical experimental approaches to the role of VDAC phosphorylation in mitochondrial metabolism.

2.3 Posttranslational modification of VDACs by acetylation

The knowledge of role of the two types of protein N-acetylation, i.e., N-terminal and lysine epsilon acetylation, in metabolic regulation is relatively new and is on the rise. Using novel methods and approaches, it is estimated that in humans (HeLa cells) approximately 85% of the proteins may be N-terminally acetylated [34, 54]. Nα-acetylation is a common cotranslational – and a rather rare posttranslational – process catalyzed by the highly conserved N-terminal acetyltransferases, which differ in their specificity towards the N-terminal amino acid [55]. Although N-terminal acetylation is considered irreversible, recent data have shown that some of the proteins are incompletely Nα-acetylated [56]. The significance of N-terminal acetylation in cell metabolism is not completely understood; it is thought to affect protein activity, stability, assembly, and intracellular location [34, 57]. Recently it has been shown that Nα-acetylation is regulated by the availability of acetyl-CoA and changes in acetyl-CoA content causes directional changes in N-terminally acetylated proteins [58]. The authors found that Bcl-xL overexpression in HeLa cells lowers the content of acetyl-CoA and Nα-acetylated proteins that is restored by addition of acetyl-CoA precursors such as acetate and citrate and at the same time confers sensitivity to apoptotic stimuli [58]. These data provide a link between metabolism, N-terminal acetylation, and sensitization to apoptosis.

We previously reported that in rat liver mitochondrial outer membrane the N-terminal methionine from VDAC1 is removed and the new amino terminal alanine is acetylated [19]. The relevance of this N-terminal acetylation to the reported role of VADC1 remains to be determined. In contrast, both VDAC 2 and VDAC3 retain their predicted N-terminal methionine and are not N-terminally acetylated [19].

In contrast to the analysis of Nα-acetylation that was hampered by lack of facile and selective enrichment methods, the availability of epsilon acetyl-lysine antibodies enormously increased the speed by which lysine acetylated proteins are identified. ε-Lysine acetylation occurs postranslationally and is catalyzed by protein acetyltransferases (PATs) localized in nucleus, cytosol, and mitochondria. The direct demonstration of the presence of protein acetyltransferase in mitochondria is still lacking; however, the fact that mitochondrially-encoded proteins are acetylated makes the presence of protein acetyltransferase in mitochondria highly likely [24]. Lysine deacetylation is catalyzed by histone deacetylases (HDACs) and NAD⁺-dependent deacetylases (sirtuins) [26, 35-39]. And mitochondria contain the NAD⁺-dependent Sirt3, Sirt4, and Sirt5 protein deacetylases

The biological significance of protein lysine acetylation is well-documented for histones but understanding of its role in intermediary metabolism is burgeoning [59-62]. A compilation of data regarding the acetylation of the three different VDAC isoforms is presented in Table 2. All data are obtained from global proteomic studies of acetyl-lysine enriched fractions.

Table 2.

Posttranslation modification of VDACs by epsilon lysine acetylation.

Isoform	Acetylation (species/organ)	Effect of acetylation	Ref.
VDAC1	K33, 41, 74, 234 (mouse liver mitochondria)	Not determined	24
	K42, 122, 132 (mouse liver mitochondria)	Not determined	25
	K237 (mouse liver and heart)	Not determined	63
	K28 (human liver)	Not determined	26
VDAC2	K32, 75 (mouse liver mitochondria)	Not determined	24
	K121 (mouse liver mitochondria)	Not determined	25
VDAC3	K20, 61, 226 (mouse liver mitochondria)	Not determined	24
	K63, 109 (mouse liver mitochondria)	Not determined	25
	K28 (human liver)	Not determined	26

Analysis was based on isolated mitochondria or on whole tissue homogenates.

VDAC1 was found to be multiply acetylated in mouse liver but in different positions in two separate studies [24, 25]. In a global proteomic approach, Kim et al. found that in liver of fed and starved mice VDAC1 is acetylated on lysines in positions 33, 41, 74, and 234 [24]. In a similar study of the effect of calorie restriction, Schwer et al. found that in mouse liver VDAC1 was lysine-acetylated in positions 41, 122, 132 [25]. However, no significant quantitative differences in VDAC1 acetylation were found between fed and starved and fed and calorie restricted animals, respectively. In addition to the above mentioned sites, VDAC1 also was found to be acetylated on lysine 237 in mouse liver and heart [63].

In mouse liver, VDAC2 also was found to be acetylated; acetylation of lysine 32 and 75 was documented in [24] and on lysine 121 in [25]. Again, no differences in acetylation between fed and starved and fed and calorie-restricted mice were observed.

In contrast to VDAC1 and VDAC2, acetylation of VDAC3 on lysine 20, 61, and 226 was found only in livers of starved but not of fed mice [24]. Schwer et al. identified lysine 63 and 109 as being acetylated in mouse liver, with no differences between fed and calorie-restricted animals [25]. None of the above studies addressed the potential physiological effect of VDAC acetylation.

In human liver mitochondria, VDAC1 and VDAC3 are found to be acetylated in position 28 [26]. The amino acid sequence of the acetylated dodecapeptide in human VDAC1 and 3 differ in three amino acid residues.

3. Conclusion and perspectives

VDAC is the most abundant mitochondrial outer membrane protein and is considered as the “gatekeeper or mitochondrial governator” [64]. Yet relatively little is known about its regulation. VDAC plays a pivotal role in mitochondrial function including metabolite and energy (ADP/ATP) exchange and utilization of ATP by hexokinase bound to VDAC, docking of cytosolic and intermembrane space proteins, and apoptosis. Moreover, these functions are modified by posttranslational modification. Here, we examined the posttranslational modifications of VDAC isoforms by phosphorylation and N-terminal- and lysine acetylation, and where it has been studied, discussed their physiologic relevance.

All three VDAC isoforms are modified by both phosphorylation and acetylation on multiple sites. Future studies most certainly will reveal an increase in the number of modified sites as well as the type of modification. The challenge for future studies is to put these modifications in a physiological perspective by altering the modified site(s) and determining the physiological outcome of these mutations. Considering that the different functions of VDACs are dependent on or are modified by the presence of other proteins, this further increases the complexity of this task.

Highlights.

• Phosphorylation of VDAC on serine, threonine, and tyrosine has been detected.

• Phosphorylation of serine 12 and 103 decrease degradation of VDAC1 that is associated with increased apoptosis.

• Phosphorylation of serine 193 prevents cell death.

• Lysine acetylation has been detected but functional consequences have not been studied.