Molecular Dynamics Simulations of the Elusive Matrix‐Open State of Mitochondrial ADP/ATP Carrier

Abstract

We studied and compared in detail an elusive matrix open (m‐state) and cytoplasmic open (c‐state) state of ADP/ATP carrier (AAC) protein embedded in the DOPC bilayer by microsecond molecular dynamics (MD) simulations. We analyzed both states with and without cardiolipin (CDL) molecules, with a special emphasis on the recently obtained crystallographic structure of the AAC m‐state. The obtained results show that both states of the protein are stable in the DOPC bilayer and impermeable to water. In comparison with the c‐state of AAC, the m‐state is more dynamic, but at the same time possesses a larger occluded area in the protein cavity. Both states of the protein are less flexible in simulations when CDL molecules are present, which is especially visible for the m‐state. Finally, the analysis of the protein conformational changes during MD simulations shows that protein parts at the protein/lipid boundaries are prone to larger conformational changes in contrast to central region of the protein embedded in the bilayer core, thus further supporting the cycling mechanism suggested for ADP/ATP exchange by AAC.

Molecular Dynamics Simulations of Mitochondrial Uncoupling Protein 2

“…In contrast, the area spanned by PC1/PC2 in the UCP2 h structure was much smaller and relatively compact, demonstrating that it was stable in the bilayer within our simulation time. It was in line with RMSD and RMSF analyses and with reported microsecond MD simulations of ANT protein [44].…”

“…Taking into account that strategic pillars, including cytosolic and matrix salt bridges as well as shape-forming proline kinks in both structures were conserved, we felt that it was safe to take the ANT structure as a starting point for subsequent MD simulations. In our previous MD simulations of two different crystallographic structures belonging to differently open ANT states towards the cytosolic [ 38 ] or matrix [ 42 ] side of the inner mitochondrial membrane, we have shown that it is possible to capture important conformational changes of the protein embedded in the membrane within microsecond MD simulations [ 44 ]. Moreover, we have also revealed that an order shorter timescales of ca.…”

“…Moreover, we have also revealed that an order shorter timescales of ca. 200 ns, used in previous MD simulations of UCP2 NMR structures in DOPC bilayers [ 20 ], were not sufficiently long for a more relevant description of critical regions in the protein, such as reversible salt bridge breaking and forming [ 44 ]. In addition to MD simulations of the UCP2 h , we also repeated simulations of UCP2 NMR based on the NMR structure of UCP2 [ 20 ] and compared them to referent ANT simulations at relevant microsecond timescales.…”

“…were not sufficiently long for a more relevant description of critical regions in the such as reversible salt bridge breaking and forming [44]. In addition to MD simula the UCP2h, we also repeated simulations of UCP2NMR based on the NMR structure o [20] and compared them to referent ANT simulations at relevant microsecond tim A schematic representation of the UCP2 structure is depicted in Figure S1.…”

“…However, we should mention here that the presence of a continuous water channel is not a key prerequisite for efficient proton transfer across the membrane protein and that electrostatic effects resulting in a high energy barrier for proton transfer predominate, such as in a case of aquaporins [48,49]. direct proton transfer leading to the inactive UCP protein [44]. However, we should mention here that the presence of a continuous water channel is not a key prerequisite for efficient proton transfer across the membrane protein and that electrostatic effects resulting in a high energy barrier for proton transfer predominate, such as in a case of aquaporins [48,49].…”