ATP synthase complex and the mitochondrial permeability transition pore: poles of attraction

Abstract

The identity of the mitochondrial permeability transition (mPT) pore, a megachannel embedded in the inner membrane opened by Ca²⁺, is fiercely debated. Unraveling the components structuring this pore is critical for combating diseases as diverse as neurodegeneration, cancer, autoimmunity, and myopathies in which this phenomenon is implicated. Current consensus is that the pore is formed within, or in‐between F₀F₁ ATP synthase dimers, but not through their c‐subunit ring. Two recent studies in this issue of EMBO Reports throw more light on these aspects, one by Giorgio et al 1 showing that the β subunit of the ATP synthase harbors a Ca²⁺‐binding site responsible for triggering mPT, and the other by Bonora et al 2 demonstrating that permeability transition requires dissociation of F₀F₁ ATP synthase dimers, albeit in a manner involving the c‐subunit ring.

Permeability transition of mitochondria is an in vivo phenomenon substantiated by the opening of a channel allowing loss of matrix metabolites with a molecular weight of up to 1,500 Da and influx of water, ensuing maximum swelling and rupture of the organelles 3. A prerequisite for mPT is excessive Ca²⁺ uptake and the trigger is a Ca²⁺‐binding site facing the matrix 3. Up until recently this site was unknown but in this issue of EMBO Reports, Giorgio et al 1 identify its location on the β subunit of the F₀F₁ ATP synthase complex 1. More specifically, by using in vitro, in vivo, and in silico approaches, Giorgio et al 1 discovered that binding of Ca²⁺ to this site elicits a conformational change within the ATP synthase complex propagating from the catalytic site through the OSCP subunit and the lateral stalk to the inner membrane, potentially leading to mPT pore opening.

The site corresponds to a threonine in position 163 of the peptide encoding the β subunit, which is otherwise known to contribute in coordinating Mg²⁺ during ATP hydrolysis 4. By mutating T163 to a serine in HeLa cells, Giorgio et al 1 show that Ca²⁺‐mediated ATP hydrolysis was nearly abolished, while leaving mitochondrial respiration intact, but perhaps more importantly, the mitochondria of T163S mutant cells were much less prone to Ca²⁺‐induced mPT. As a consequence of this, the incidence of mPT‐dependent death elicited by an excessive Ca²⁺ load was attenuated in cells engineered to harbor the β subunit T163S mutation.

In order to examine the extent of contribution of this Ca²⁺‐binding site in an mPT‐dependent event in vivo, the authors overexpressed T163S mutant β subunit in zebrafish embryos. This led to curling of their tails and decreased skin pigmentation, typically observed when apoptosis is inhibited; this pathological phenotype was associated with fewer apoptotic nuclei in the head, tail trunk and yolk, which exhibit apoptosis as part of normal organogenesis.

To better “visualize” the impact of Ca²⁺ binding to β subunit T163 on mPT pore formation (Fig 1), Giorgio et al 1 used a molecular dynamics approach showing that Ca²⁺ increases F₁ rigidity, leading to a decrease in F‐ATP synthase compliance; in line with conformational changes observed in the peripheral stalk at OSCP/F6 and b subunit 5, this rigidity may transmit mechanical energy to the lateral stalk through OSCP, and ultimately to the inner mitochondrial membrane, where mPT pore forms. Although the simulations of the T163S mutant exhibited the same rigidity conferred by Ca²⁺, there, OSCP subunit “dampened” this mechanical strain, potentially alleviating factors leading to mPT opening. Based on the ability of OSCP to thwart Ca²⁺‐induced mPT, the authors predicted that a critical role of OSCP in mPT modulation might be emerging.

Figure 1. Schematic of the mPT based on the results presented by Giorgio et al 1 and Bonora et al 2 .

Ca²⁺ binding on T163 position of the β subunit triggers the dissociation of F₀F₁ ATP synthase dimers, which leads to mPT.

The study by Bonora et al 2 identifies the dissociation of ATP synthase dimers as a means of mPT opening (Fig 1), in a manner influenced by the c‐subunit rings 2. The results presented are especially provocative, because they are at odds with the those of an earlier work by Giorgio et al 1 claiming that F₀F₁ ATP synthase dimers are a prerequisite to mPT pore opening 6.

Bonora et al 2 presented their case of demonstrating that the mPT pore opens upon dissociation of ATP synthase dimers by examining the ratio of dimeric/monomeric state of the complexes in isolated mitochondria undergoing mPT, and also by quantitating ATP synthase dimerization in cellula, while assessing mPT opening by recording the quench of mitochondrially trapped calcein by exogenously added Co²⁺. ATP synthase dimerization in cellula was evaluated by using a proximity ligation assay (PLA) which involved labeling of F₀ complex subunit D (ATP5H) under conditions that favor, or disfavor dimerization of the complexes. The PLA technique is designed to visualize interactions among two proteins, and thus, it usually requires two different antibodies raised against two different proteins. In the Bonora et al 2 work, the technique was ingeniously used to detect interaction among two of the same proteins, i.e. ATP5H, as it may occur during ATP synthase dimerization, thus only one antibody was needed.

When isolated mitochondria were challenged by energized Ca²⁺ uptake, mPT ensued within a few minutes and this was associated with an increase in ATP synthase monomers over dimers, which was completely prevented by the cyclophilin D inhibitor, cyclosporin A.

In line with the results obtained from isolated mitochondria, by using the PLA assay Bonora et al 2 demonstrate that a calcionophoric [Ca²⁺] increase in the cytosol of HEK293T cells eliciting mPT correlates with a decreased PLA signal, implying dissociation of ATP synthase dimers. Furthermore, silencing the expression of inhibitory factor 1 (IF1), a protein promoting dimerization of the complex 7, sensitized cells to mPT upon increasing intracellular [Ca²⁺]. By the same token, IF1 overexpression led to an increase in PLA signal and intensified resistance of the mitochondria to the same mPT pore opening protocol. The authors corroborated their findings by manipulating the expression of F₀ subunit E (ATP5I), also known to regulate complex dimerization: silencing versus overexpression of ATP5I led to an increase versus decrease in the incidence of Ca²⁺‐induced mPT, just as predicted due to alterations in ATP synthase dimers/monomers ratio.

Finally, by performing elaborate site‐directed mutagenesis on the c‐subunit encoded by ATP5G1, altering the properties of a glycine zipper domain normally found in c‐subunits, Bonora et al 2 showed that this domain is critical for dimer dissociation and mPT opening, and as an extension of that, so is the c‐subunit. Provocatively enough, although this finding conforms to their own previous work claiming that the c‐subunit constitutes a critical mPT component 8, it is at odds with two recent reports showing exactly the opposite, that is, that the c‐subunit ring has no role in mPT 9, 10.

Nevertheless, apart from the controversies this work offers, the results presented in this and the Giorgio et al 1 paper “mutually attracting” the entities of ATP synthase to mPT, comprise a step further toward the elucidation of the pore structure.

See also: V Giorgio et al (July 2017) and M Bonora et al (July 2017)