Protein Unfolding by the Mitochondrial Translocase

Research overview The degree to which a precursor unfolds during transport; The effects of the mitochondrial surface on the stability of specific interactions within a precursor and between a precursor and ligands; The unfolding pathway of a precursor as it interacts with different parts of the import machinery; The mechanism of the import motor.

Our goal is to define the conformational changes proteins undergo during transport across biological membranes and to define how these changes are induced.

Approximately half of the proteins in an average eukaryotic cell are transported into or across a membrane. Cells have developed elaborate translocation systems in different organelles that share fundamental features. These translocation systems have been studied extensively, and in the best characterized of these the identities and locations of their components have been established. However, the biochemical and biophysical mechanisms of protein transport remain unclear. Little is known about the structural changes that precursors undergo during transport and the mechanisms that induce these changes and drive the precursor across the membrane.

We use a novel approach to understand protein folding and unfolding processes that occur at membranes based on protein engineering and kinetic measurements. The stability and folding of the small RNase barnase in free solution has been characterized in detail, and this protein therefore serves as an excellent model substrate for these studies. We have constructed transport-competent precursors consisting of barnase fused to targeting sequences of different lengths. This makes it possible to determine quantitatively how the interaction of the precursors with specific parts of the transport machinery in intact organelles affects the structure and unfolding of barnase. We investigate protein import into purified yeast mitochondria because this is one of the best studied transport systems. Moreover we can manipulate mitochondrial proteins using yeast genetics and can obtain precise quantitative information on import under conditions that closely mimick the in vivo environment.

In order to be transported precursors have to be at least partially denatured andare often fully unfolded. The mitochondrial import machinery permits the passage of bulky groups such as precursors with branched polypeptide chains or with a double stranded oligonucleotide of 24 base pairs attached to the C-terminus. However, the efficiency of import is reduced by the larger groups. The inner mitochondrial membrane contains the enzymes of the respiratory chain which produce the electrochemical potential across this membrane. When mitochondria are forced to accumulate precursor proteins spanning the inner membrane, respiration is not affected and the electrochemical potential is maintained. This suggests that the import channel closes tightly around the incompletely transported precursors. Thus, the question arises if proteins have to be fully unfolded during import and are threaded through the import machinery amino acid by amino acid, or if precursors retain some partial structure?

The question of the structure of precursors during import into mitochondria has been addressed before. In these experiments, precursor proteins were constructed in which stretches of amino acids of different lengths were inserted between dihydrofolate reductase (DHFR) and the cleavable targeting sequence of cytochrome b2. When these precursors were incubated with mitochondria in the presence of methotrexate, DHFR could not enter the import channel and remained at the mitochondrial surface. The targeting sequence was cleaved by the matrix processing peptidase once a minimum of 50 amino acids were inserted between the DHFR and the cleavage site at the C-terminus of the targeting sequence. The distance between the outer and inner surface of import sites can be estimated from electron micrographs to be 18 to 20nm. Since 50 amino acids can only stretch over this distance in an extended conformation, it was argued that proteins are always in an extended conformation during import. The weakness in this argument is that the 50 amino acids inserted into the precursor were taken from the N-terminus of mature cytochrome b2 and constitute only a fraction of its heme binding domain. Because protein folding is highly cooperative, including only one part of a structural domain would not be expected to allow folding into a recognizable structure. Thus these experiments show that proteins can be fully unfolded during transport but not that this is a general requirement for transport.

The folding pathway of many proteins in vitro contains an intermediate called the "molten globule". Proteins in the molten globule state lack defined tertiary structure but retain secondary structural elements. It has been suggested that the molten globule state is the transport-competent form of proteins. There is evidence that the molten globule state of the proteinaceous toxin colicin A is the form of the protein that inserts into bacterial membranes where it is active. During import a precursor unfolds at the mitochondrial surface. The outer membrane contains a high proportion of negatively charged phospholipids which are expected to cause a decrease in surface pH. Since low pH induces molten globule formation in many proteins mitochondria could use the environment to convert proteins into the molten globule form during the first steps of protein import. We are determining whether the molten globule state is the transport competent form of proteins and how much residual structure in precursors the import machinery tolerates.

One puzzling aspect of import is that mitochondria can import folded proteins faster than they unfold spontaneously. Thus mitochondria unfold proteins actively at their surface. The outer mitochondrial membrane, like most other biological membranes, contains a high proportion of negatively charged phospholipids and it has been suggested that the charges contribute to the unfolding of precursors. The charged lipids in the membrane are expected to increase the local proton and cation concentration at the membrane surface. There is evidence that the decreased surface pH induces the molten globule state in the toxin colicin A and that this is the form of the protein that inserts into membranes where it is active. More specifically to mitochondria, the proteinaceous import receptors and the opening of the import channel may also interact with precursors and destabilize them. Destabilization of cytochrome c and DHFR precursors at the surface of vesicles and mitochondria has been reported and is thought to be important for transport of precursors into mitochondria. However, these experiments did not measure time courses of unfolding and it is possible that they dealt with relatively slow spontaneous protein unfolding followed by trapping of unfolded conformers by the membrane. We are determining whether the mitochondrial membranes destabilize proteins and the binding of ligands to receptors. If yes, we will investigate the specific nature of the effect of the mitochondrial surface on interactions within proteins and between proteins and ligands.

Recently, we have shown that mhsp70 can directly cause the unfolding of a precursor protein at the outer surface of mitochondria by binding to the N-terminus of the precursor in the mitochondrial matrix. However, we do not know by what mechanism mhsp70 unfolds the precursor and transports it into the matrix. Mhsp70 is a member of the large family of 70 kD chaperones (hsp70s) which are found in the cytosol of procaryotes and eucaryotes as well as the lumen of the endoplasmic reticulum, chloroplasts and mitochondria. Hsp70s show a high degree of amino acid sequence identity, with the sequences of cytoplasmic bovine hsc70 and Bacillus hsp70 being 52% identical. In addition to protein import into mitochondria, hsp70s are involved in protein transport into the endoplasmic reticulum and chloroplasts, in protein synthesis and degradation, and in the assembly and disassembly of oligomeric complexes. Hsp70s have three domains, a N-terminal ATPase domain, followed by a peptide binding domain and a C- terminal domain of unknown function. High resolution X-ray structures of both the N- terminal domain of a mammalian cytosolic hsp70 and of the peptide binding domain of DnaK, the E. coli hsp70 homolog, have been determined. Members of the 70 kDa heat shock protein family undergo conformational changes in an ATP-dependent manner and it is likely that mhsp70 does so, too. This led to an attractive model for mhsp70 action in protein import into mitochondria. As mhsp70 binds both to the N-terminus of an incoming precursor and to a component of the inner mitochondrial membrane (Tim44), a conformational change in mhsp70 could generate sufficient force at the N-terminus of the precursor to pull a segment of the polypeptide chain out of a folded domain on the mitochondrial surface, reminiscent of the action of myosin on actin fibers. Since protein folding is highly cooperative this would then cause a collapse of this domain. We are testing this hypothesis and investigate how mhsp70 unfolds precursors.

Protein transport into mitochondria also requires an electrochemical potential across the inner mitochondrial membrane. This potential has a substantial electric component of more than 100 mV for respiring yeast mitochondria (positive at the outer surface of the membrane and negative at the inner surface). Mitochondrial targeting sequences are characterized by a high density of positive and an absence of negative charges and it appears likely that the electrochemical potential acts through these charges. Whereas ATP is required throughout the import process, the electrochemical potential is required only in the early steps. The mechanism of action of the electrochemical potential is unclear. It has been suggested that it is required to drive the movement of presequences into the matrix to prevent back sliding of precursors in the import channel after the initial unfolding and import steps induced by mhsp70. We propose that the electrochemical potential contributes directly to the unfolding of precursors and are testing this hypothesis.

Mhsp70 is not only required for the initial unfolding of precursors but also for the consequent movement of the precursor across the membranes. The enzymology of hsp70s in solution has been studied extensively. The Km and kcat values for the ATPase activity and its correlation to peptide binding and release have been investigated for different members of the hsp70 family. Mutational structure function studies on the Escherichia coli protein DnaK and the rat and bovine cytosolic 70 kD heat shock cognate proteins define the importance of specific amino acid side chains in the ATPase activity of hsc70, in the communication between ATPase and peptide binding domains, and in the interaction of peptide substrate with the peptide binding domain. This work forms a basis for comparison for our studies on the function of mhsp70 in their natural environment. Similarly, the studies on motor proteins such as myosin, kinesin and to a lesser extent polymerases provide a frame of reference for testing the motor hypothesis of mhsp70 action.