31 Cyclophilin-D binding proteins

Mitochondria contain a structure able to form a large (2 nm) inner membrane pore when triggered by high calcium and oxidative stress [1,21. Opening of the so-called permeability transition pore is reversible, rapid pore closure occuring on calcium removal [13 or cessation of oxidative stress [31. Once opened, the pore completely uncouples oxidative phosphorylation. Pore opening may be a determining step in cell injury associated with calcium overload and oxidative stress since it would initiate a vicious cycle of events leading to cell death [2,41. The pore is blocked by cyclosporin A. Studies in this [4,51 and other [6,71 laboratories show that the cyclosporin A "receptor" is not the pore protein itself, but a mitochondrial cyclophilin (CyP) isoform, CyP-D, which controls or stabilizes the pore by forming a part of the active pore complex. In order to identify other components of the pore we have looked for mitochondrial proteins that bind to CyP-D, using a glutathione-S-transferase (GST) / CyP-D fusion protein as affinity matrix. CyP-D cDNA was cloned from a rat liver library ( S . Virji, J. M. Ward and M. Crompton, unpublished data). The cDNA sequence was essentially identical to that isolated previously [81. DNA encoding amino acids 30-216 was cloned into pGEX-3X; this portion codes for mature CyP-D i.e. lacking the presumed N-terminal mitochondrial targetting sequence. The fusion protein was expressed in E. coli XL-1. Induced cells were sonicated in 100 mM NaCl/10 mM Tris-HC1 (pH 7.2)/ 0.5 mM EDTA/1 mM phenylmethylsulphonylfluoride. The supernatant was applied to small glutathioneagarose columns, which were washed extensively with 100 mM NaC1/10 mM Tris-HC1 (pH 7.2). The resultant affinity matrix contained about 50 ug of fusion protein per ml of glutathione-agarose. Detergent extracts of rat heart submitochondrial particles were incubated with the affinity matrix for 2 min and the matrix was washed thoroughly. Glutathione was then added to elute the GST/CyP-D along with mitochondrial protein that had bound to it. The displaced proteins were analysed by SDS-PAGE. Figure 1A shows the proteins that were retained by the affinity matrix and then displaced by glutathione. These comprise the fusion protein itself (the major band at 46 kDa) along with a 32 kDa protein (lane 2). For comparison, the composition of the extract applied to the affinity matrix is reported in lane 1. The 32 kDa protein was not retained by a GST matrix, so that it evidently bound to the CyP-D moiety of the fusion protein. The 32 kDa protein reacted with a monoclonal antibody against human mitochondrial porin; this antibody also reacts with rat porin C91. It appears, therefore, that the 32 kDa protein is porin and that it is able to bind tightly to CyP-D (Fig. 1B). We have speculated previously [2-51 that a role for porin in the permeability transition pore might explain the low number of such pores in mitochondria subjected to high levels of calcium and oxidative stress. Firmer evidence is the complex formation between porin and the adenine nucleotide translocase [lo] and the fact such complexes, in association with CyP-D, produce pores with the relevant properties when reconstituted into proteoliposomes [91. The present data provide the first evidence for a high affinity interaction between porin and CyP-0. But they raise the intriguing question of how these proteins might interact in vivo. Whereas porin is located in the outer membrane, CyP-D is found in the matrix. Figure 1. The binding o f porin by cyclophilin-D