Cyclophilin D regulation of the mitochondrial permeability transition pore

Abstract

The opening of the mitochondrial permeability transition pore (PTP) has been proposed to play a critical role in activating cell death in many settings, including cardiac ischemia and reperfusion. Although the identity of pore forming unit of the PTP is still debated, it is generally agreed that cyclophilin D (CyPD) is a regulator of the PTP. This manuscript will focus on understanding how CyPD might regulate the PTP and how understanding CyPD might give insight about the identify and regulation of the PTP.

Background:

The permeability transition pore (PTP) is a large conductance channel in the inner mitochondrial membrane that is activated by high matrix Ca²⁺ and reactive oxygen species. Opening of the PTP is proposed to be an initiator of cell death in many diseases such as cardiac ischemia and reperfusion injury. Although the mitochondrial permeability transition pore (PTP) has been proposed to play a critical role in activating cell death in many settings, the identity of the pore forming unit is still debated^1–6. Strong evidence has been provided for several candidates including the F1-F0-ATPase (F-ATPase)^{7, 8} and the adenine nucleotide translocator (ANT)^{9, 10}; however there are also data arguing against both of these candidates.^{6, 11} It also has been suggested that there might be multiple pore forming channels^{1, 3, 4, 12, 13} Although the identity of pore forming unit of the PTP is still debated, it is generally agreed that cyclophilin D (CyPD) is a regulator of the PTP.^14–20 This manuscript will focus on understanding how CyPD might regulate the PTP. Generally CyPD is thought to activate PTP, although as discussed some studies suggest that it can inhibit PTP²¹.

CyPD is an activator of PTP

Numerous studies have shown that inhibition or deletion of CyPD, a mitochondrial matrix peptidyl prolyl cis-trans isomerase, blocks PTP activation in isolated mitochondria and reduces cell death in ischemia-reperfusion injury in heart and other tissues^{14–16, 20}. Addition of a bolus of Ca²⁺ (typically 250 μM) to isolated mitochondria leads to activation of PTP, measured as a decrease in absorbance at 540 nm which is a monitor of mitochondrial swelling. Addition of cyclosporin A (CsA), an inhibitor of CyPD blocks or attenuates the increase in mitochondrial swelling. PTP opening is also measured in isolated mitochondria or permeabilized cells by measuring the amount of Ca²⁺ (the Ca²⁺ retention capacity (CRC) assay) that mitochondria can accumulated before they can no longer take up Ca²⁺ and instead release all the accumulated Ca²⁺. In this assay a calcium sensitive dye such as Ca²⁺-Green is placed outside the mitochondria to monitor Ca²⁺ uptake into the mitochondria. CsA increases (typically by 2.5 fold) the amount of Ca²⁺ that can be accumulated. Similar results have been obtained in studies with loss of CyPD, confirming that CyPD is the relevant target of CsA in these studies. CsA and loss of CyPD have also been shown to be protective in hearts subjected to ischemia and reperfusion.^{14, 15, 20} However, in a clinical trial, CsA did not provide protection to patients with myocardial infarction.²² Possible reasons for the lack of protection have been discussed.^{23, 24}

CyPD and the PTP Models

CyPD and F1-ATPase:

CyPD has been shown to bind to the ATPO subunit of the F1-ATPase and its binding was shown to inhibit the activity of the F1-ATPase to generate or consume ATP^{25, 26}. Bernardi and coworkers showed in a series of papers that CyPD binding to the ATPO subunit alters the conformation to promote pore opening at the dimer/tetramer interface of the Fo domain. The details of the model have recently been reviewed.¹ The c-subunit of the F1-ATPase has also been proposed to function as the PTP. It is suggested that CyPD and Ca²⁺ mediate dissociation of Fo from the F1 subunits thereby promoting PTP.

CyPD and ANT:

CyPD has been shown to bind to ANT^{9, 18}, although Crompton reports that binding occurs in the presence of CsA¹⁸, whereas Halestrap’s group finds that CsA blocks the binding¹⁹. Furthermore Lin and Lechleiter report that isomerase dead CyPD can still bind to ANT.²¹ It has been suggested that CyPD can activate the ANT pore, but that high Ca²⁺ can activate this pore even in the absence of CyPD, whereas the non-ANT pore requires CyPD for activation.³ There are several prolines in ANT which could be binding sites for CyPD. Halestrap suggested that Proline 61/62 is the site of CyPD binding to ANT¹⁹. P61/62 is on the matrix loop which is in a region that undergoes a confirmational change as ANT transitions from the m-state, which is open to the matrix, to the c-state which is open to the cytosol or intermembrane space²⁷. It is tempting to speculate that perhaps CyPD binding or a CyPD mediated isomerization is involved in the m to c state transition. However it does not appear that CyPD binding to ANT is critical to ANT function or m to c state transition as loss or inhibition of CyPD does not appear to impair function of ANT.²⁵ If CyPD is involved in the transition from the m to the c state one might expect loss of CyPD to alter adenine nucleotide transport. However, Chinopoulos et al report loss of CyPD does not alter ANT mediated exchange of ATP/ADP²⁵.

Multiple Pore model:

Although there are data supporting a role for the F-ATPase in forming the PTP, studies by Walker’s group has shown that a CsA inhibitable PTP is still present when key F1-ATPase subunits are deleted⁶. Similarly a it has been suggested that CyPD can activate the ANT pore, but that high Ca²⁺ can activate this pore even in the absence of CyPD, whereas the non-ANT pore requires CyPD for activation¹⁰. These finding have led to the hypothesis that perhaps there are at least two PTP pore forming units. This hypothesis is discussed in detail elsewhere^1,3,4,12,13.

Is Isomerase activity required for CyPD inhibition of PTP?

Despite decades of investigation, the mechanism by which CyPD activates PTP is unclear. CyPD is a member of the cyclophilin family. Cyclophilin facilitates cis-trans isomerization around peptide proline bonds. Over 25 years ago it was discovered that CsA was immunosuppressive and it was quickly shown that CsA immunosuppression required its binding to cyclophilin A (CyPA). In fact cyclophilin was named based on its affinity for cyclosporin (cyclosporin loving).²⁸ Because CsA binding to CyPA inhibited its isomerase activity, it was initially proposed that the isomerase activity played a role in immunosuppression. However mutations in CyPA, which inhibited isomerase activity did not block its ability to bind CsA and suppress T cell activation.²⁹ Additional studies demonstrated that CyPA and CsA formed a complex that binds to calcineurin leading to immunosuppression.³⁰

Is isomerase activity required for CyPD inhibition of PTP or does CyPD inhibition work by protein-protein interactions? There are conflicting data as to whether isomerase activity is required for CyPD activation of PTP. Early inhibitor studies suggested that isomerase activity was not required for PTP activation³¹. In contrast, Baines et al expressed a CyPD with an R96G mutation, which lacks isomerase activity, into cells lacking CyPD and showed that the isomerase dead CyPD showed similar protection to CyPD-KO when treated with H₂O₂¹⁴. If, as proposed, there are two PTP with different sensitivity to CsA it is possible that CyPD can activate these separate pores by different mechanisms. Isomerase activity could be required to inhibit one but perhaps not both PTPs.

How is CyPD regulated?

CyPD undergoes a number of post translational modification (PTMs) which could play a role in regulating CyPD function^32–36. One limitation in studying the role of CyPD PTMs is that the function of CyPD is not totally clear. CyPD has enzymatic activity as an isomerase; however as discussed above the role of isomerase activity in CyPD activation of the PTP is debated, and there is clear precedence for CypA-CsA inhibition of calcineurin that is independent of CypA isomerase activity.^{29, 30}

Acetylation of lysine 166 of CyPD increases with loss of SIRT3 and increased acetylation of CyPD (lysl66) is suggested to promote PTP opening, as there is an age-dependent increase in PTP opening in SIRT3-KO mitochondria³³. CyPD has also been reported to be phosphorylated at several different residues. Rasoli et al reported that glycogen synthase kinase (GSK) can phosphorylate CyPD³⁶ and although the site was not identified this GSK dependent phosphorylation was shown to enhance PTP opening. Parks et al found an increase in phosphorylation of CyPD at serine 42 (S42) in mitochondria from MCU-KO cardiac mitochondria³⁴. They provided data suggesting that increased phosphorylation of CyPD at S42 increases its binding to the F1-ATPase which appears to lower the Ca²⁺ required to activate PTP. Hurst report that phosphorylation of CyPD at serine 191 enhances CyPD binding to ATPO and activation of PTP opening³⁵. Although in general increased phosphorylation of CyPD appears to enhance PTP opening, the role of phosphorylation at different serine residues and the kinase and phosphatases that regulate phosphorylation need further study. It is also not clear whether the increase in phosphorylation of CyPD enhances PTP opening solely by enhancing binding to the F1-ATPase or whether it also enhances binding to other targets such as ANT. Its effect on isomerase activity is also unclear.

CyPD has four redox sensitive cysteine residues and Linard et al showed that oxidation of these cysteine residues can alter CyPD conformation and isomerase activity³⁷. They identified C202/203 as a redox sensitive site. We found that preconditioning leads to S-nitrosylation (SNO) of C202 and SNO at this site is cardioprotective, presumably by protecting this cysteine from oxidation. A mouse was generated in which C202 of CyPD was mutated to a serine. The mouse with this mutation was protected from ischemia-reperfusion injury³². We postulate that oxidation of C202 or disulfide bond formation with C202 enhances PTP opening, perhaps by targeting CyPD to the PTP.

In addition to oxidation and SNO, C202 is also a site of high levels of acylation, the attachment of a fatty acid group to a cysteine via a thioester linkage. Acyl resin assisted capture (RAC) was used to show that hearts with the C202S mutation have significantly less acylation of CyPD, confirming that C202 is a major site of S-acylation. The functional significance of acylation of C202 of CyPD is unclear, as are the enzymes involved in acylation and de-acylation. S-acylation of CyPD is not altered in a perfused heart during 60 minutes of perfusion. However most of the acylation of CyPD is lost following 20 minutes of ischemia³². Studies in isolated mitochondria show that an increase in mitochondrial Ca²⁺ leads to de-acylation of CyPD. These data would be consistent with the hypothesis that an increase in mitochondrial Ca²⁺ promotes de-acylation of C202 of CyPD, exposing the free cysteine to oxidation which can target CyPD to the PTP and promote PTP opening.

CyPD and regulation of F1-ATPase and synthasome formation:

It has been suggested that the components required for ATP synthesis, the F1-ATPase, ANT and the phosphate carrier (PiC), form a complex referred to as the synthasome, which enhances the synthesis of ATP. Porter and coworkers have proposed that CyPD might be required for the disassembly of the ATP synthasome³⁸. Consistent with data suggesting that CyPD reduces F1-ATPase activity, they suggest that CyPD reduces the level of the synthasome which is proposed to have higher F1-ATPase activity. It is further proposed that when ANT and F1-ATPase are in the synthasome complex that they do not form the PTP, but that CyPD promotes the disassembly of the synthasome into the component parts (F1-ATPase and ANT) which can form the PTP. Interestingly, overexpression of CyPD has been reported to hyperpolarize the mitochondrial membrane potential.²¹ This would be consistent with CyPD promoting disassembly of the synthasome and thereby inhibiting F1-ATPase which would tend to enhance membrane potential (e,g, oligiomycin increases Δψ). It is interesting to speculate that perhaps PTMs of CyPD provide a mechanism linking metabolism and other signaling pathways to the activity of the F1-ATPase.

CyPD inhibition of PTP:

Although the majority of data suggest that CyPD activates PTP, there are data reporting that CyPD can reduce cell death. Overexpression of CyPD in HEK293 and rat glioma C6 cells desensitized the cells to apoptotic stimuli.²¹ It was further demonstrated that this protection afforded by overexpression of CyPD required isomerase activity. It is noteworthy that the original studies characterizing the CyPD-KO mice showed that loss of CyPD inhibited necrotic, but not apoptotic cell death.^{14, 15} Consistent with the concept that CyPD can contribute to protection, Carraro et al reported that CyPD binding to C of ATPO protects it from oxidation.³⁹

Summary and Remaining Questions:

Emerging data suggest that both ANT and the F1-ATPase can form PTP like channels (perhaps with different size pores). It is interesting that although CsA sensitivity is the main criteria used to define whether a channel is the PTP, we still do not fully understand how CyPD actives PTP. Furthermore, it is known that the PTP can become CyPD independent, so sensitivity to CsA is not an absolute criteria to define the PTP.

There are data suggesting that CyPD can enhance dissociation of the synthasome. If only the dissociated ANT and the F-ATPase, or perhaps the partially dissociated proteins can form PTP channels this might provide, at least a partial explanation, for how CyPD activates PTP. It is likely that CyPD performs some physiological function in addition to modulating PTP, and it makes sense for the mitochondria to regulate the activity of the F1-ATPase. Indeed loss of CyPD has been shown to lead to metabolic alterations.^40–42 Furthermore, under physiological conditions Ca²⁺ has been reported to activate the F1-ATPase activity, although the mechanism remains elusive.⁴³ It is generally thought that the F1-ATPase activity in the synthasome is increased. If that is the case, then perhaps CyPD modulates the activity of the F1-ATPase, either by protein-protein interaction or by isomerase activity. If this hypothesis (figure 1) is correct it raises several questions and testable predictions. First, why is ADR protection additive to that of CyPD? If CsA and ADP similarly inhibit the PTP activity of both ANT and the F1-ATPase, then it is not clear why CsA and ADP inhibition is additive. CsA and ADP must differentially inhibit the two pores. It has been suggested that CyPD can activate the ANT pore, but that high Ca²⁺ can activate this pore even in the absence of CyPD, whereas the non-ANT pore requires CyPD for activation.³ Perhaps CsA primarily blocks the non-ANT pore (F1-ATPase) and ADP primarily blocks the ANT pore. Another issue is that only some of the ANT and F1-ATPase exist in the synthasome. So even if CyPD alters the equilibrium one would always expect some free ANT and F1-ATPase that are not in the synthasome. Is CsA’s only function to reduces the level of free ANT and F1-ATPase? Or does CyPD directly activate the free ANT and/or F1-ATPase? Also there are data suggesting that oxidation plays an important role in targeting CyPD to the PTP. How does this fit with the model in figure 1. Does oxidation target CyPD to the synthasome to enhance its disassembly? Assuming the F1-ATPase activity is enhanced in the synthasome one would expect that increase F1-ATPase activity would be associated with less ROS. If this is correct, then oxidation, by reducing F1-ATPase activity might in fact increase ROS.

CyPD activation of putative PTP pores. CyPD is proposed to activate PTP at multiple steps. The F-ATPase, ANT and the phosphate carrier (PiC) can form a synthasome which is thought to enhance ATP formation. It is proposed that CyPD promotes the disassembly of the synthasome into the components of F-ATPase and ANT which form separate PTP channels. It is likely that CyPD acts at other steps because CsA has been shown in some conditions to inhibit channel formation. CyPD is therefore proposed to also directly inhibit PTP formation by F1-ATPase and ANT, although the CyPD sensitivity may be different. Furthermore, ADP can also reduce PTP opening which is additive to CsA. This might suggest that at high Ca2+ levels the relative effectiveness of ADP and CsA to inhibit ANT and F-ATPase is different.

Another point to consider is that most studies of PTP use isolated mitochondria treated with high levels of Ca²⁺. It is not clear that oxidative stress is a trigger in this model. Is the regulation of PTP in cells under conditions of oxidative stress the same as in isolated mitochondria treated with high Ca²⁺ If there are indeed more than one PTP pores is it possible that the regulation of these pores by Ca²⁺ and ROS is different. Clearly the role of CyPD in regulating cell physiology and the PTP is still unclear and additional studies are needed.