Both ATP and an energized inner membrane are required to import a purified precursor protein into mitochondria.

We have investigated the energy requirement of mitochondrial protein import with a simplified system containing only isolated yeast mitochondria, energy sources and a purified precursor protein. This precursor was a fusion protein composed of 22 residues of the cytochrome oxidase subunit IV pre-sequence fused to mouse dihydrofolate reductase. Import of this protein required not only an energized inner membrane, but also ATP. ATP could be replaced by GTP, but not by CTP, TTP or non-hydrolyzable ATP analogs. Added ATP did not increase the membrane potential of respiring mitochondria; it supported import even if the proton-translocating mitochondrial ATPase and the entry of ATP into the matrix were blocked. We conclude that ATP exerts its effect on mitochondrial protein import outside the inner membrane.     

Tight folding of a passenger protein can interfere with the targeting function of a mitochondrial presequence.

The first seven residues of the yeast cytochrome oxidase subunit IV presequence are insufficient to target attached mouse dihydrofolate reductase into isolated yeast mitochondria. However, the targeting function of this truncated presequence can be restored by presenting the fusion protein to isolated mitochondria either as nascent, unfolded chains, or as full-length chains whose dihydrofolate reductase moiety had been destabilized either by urea treatment or by point mutations. The targeting efficiency of a mitochondrial presequence can thus be strongly influenced by the conformation of the attached 'passenger protein'. These results also underscore the difficulty of defining a 'minimal' mitochondrial targeting signal.     

Point mutations destabilizing a precursor protein enhance its post-translational import into mitochondria.

In order to study the role of protein unfolding during post-translational protein import into mitochondria, we destabilized the structure of a mitochondrial precursor protein by site-directed mutagenesis. The precursor consisted of the first 16 residues of the yeast cytochrome oxidase subunit IV precursor fused to mouse dihydrofolate reductase. Labilization of the folded precursor structure was monitored by increased susceptibility to protease and diminished ability of methotrexate to block import of the precursor into isolated yeast mitochondria. On comparing the original precursor with two mutant forms that were destabilized to different degrees, increased labilization correlated with an increased rate and efficiency of import into mitochondria. This supports the view that the precursor must unfold in order to enter the mitochondria.     

Adriamycin, a drug interacting with acidic phospholipids, blocks import of precursor proteins by isolated yeast mitochondria

Acidic phospholipids such as cardiolipin partially unfold an artificial precursor protein which consists of a mitochondrial presequence fused to mouse dihydrofolate reductase (Endo, T., and Schatz, G. (1988) EMBO J. 7, 1153-1158). We now show that import of this precursor protein into isolated yeast mitochondria is blocked by adriamycin, a drug binding to cardiolipin and other acidic phospholipids. This inhibition is lessened if the precursor's dihydrofolate reductase moiety is labilized by point mutations; inhibition is abolished altogether if the "wild-type" precursor is presented to mitochondria in a urea-denatured state. These and other observations suggest that adriamycin interferes with the generation of a translocation-competent, loose structure of the precursor protein. They imply that acidic phospholipids such as cardiolipin participate, directly or indirectly, in the translocation of this fusion protein into isolated mitochondria.     

Active unfolding of precursor proteins during mitochondrial protein import.

Precursor proteins made in the cytoplasm must be in an unfolded conformation during import into mitochondria. Some precursor proteins have tightly folded domains but are imported faster than they unfold spontaneously, implying that mitochondria can unfold proteins. We measured the import rates of artificial precursors containing presequences of varying length fused to either mouse dihydrofolate reductase or bacterial barnase, and found that unfolding of a precursor at the mitochondrial surface is dramatically accelerated when its presequence is long enough to span both membranes and to interact with mhsp70 in the mitochondrial matrix. If the presequence is too short, import is slow but can be strongly accelerated by urea-induced unfolding, suggesting that import of these 'short' precursors is limited by spontaneous unfolding at the mitochondrial surface. With precursors that have sufficiently long presequences, unfolding by the inner membrane import machinery can be orders of magnitude faster than spontaneous unfolding, suggesting that mhsp70 can act as an ATP-driven force-generating motor during protein import.     

Mitochondria can import artificial precursor proteins containing a branched polypeptide chain or a carboxy-terminal stilbene disulfonate

A purified, artificial precursor protein was used as a transport vehicle to test the tolerance of the mitochondrial protein import system. The precursor was a fusion protein consisting of mouse dihydrofolate reductase linked to a yeast mitochondrial presequence; it contained a unique cysteine as its COOH-terminal residue. This COOH- terminal cysteine was covalently coupled to either a stilbene disulfonate derivative or, with the aid of a bifunctional cross-linker, to one of the free amino groups of horse heart cytochrome c. Coupling to horse heart cytochrome c generated a mixture of branched polypeptide chains since this cytochrome lacks a free alpha-amino group. Both adducts were imported and cleaved by isolated yeast mitochondria. The mitochondrial protein import machinery can thus transport more complex structures and even highly charged "membrane-impermeant" organic molecules. This suggests that transport occurs through a hydrophilic environment.     

Unfolding of preproteins upon import into mitochondria.

Unfolding of preproteins and translocation across the mitochondrial membranes requires their interaction with mt-Hsp70 and Tim44 at the inner face of the inner membrane and ATP as an energy source. We measured the temperature dependence of the rates of unfolding and import into the matrix of two folded passenger domains, the tightly folded heme-binding domain (HBD) of cytochrome b2 and the loosely folded mouse dihydrofolate reductase (DHFR). Despite the stability of the HBD, its rates of thermal breathing were fast and the preprotein was imported rapidly at all temperatures. In contrast, rates of unfolding and import of DHFR were strongly temperature dependent and import was significantly slower than unfolding. In addition, import rates of DHFR were strongly dependent on the length of the presequence. We propose that the mitochondrial import motor does not exert a constant pulling force. Rather, mt-Hsp70 appears to release a translocating polypeptide chain such that the precursor can then slide back and refold on the surface of the mitochondria. Refolding competes with translocation, and passengers may undergo several rounds of unfolding and refolding prior to their import.     

Coupling of protein synthesis and mitochondrial import in a homologous yeast in vitro system.

We made use of a homologous cell-free mitochondrial protein import system derived from the yeast Saccharomyces cerevisiae to investigate the coupling of protein synthesis and import. Mitochondrial precursor proteins were synthesized in a yeast lysate either in the presence or absence of isolated yeast mitochondria. We were, therefore, able to analyze protein import into mitochondria either in a strictly posttranslational reaction (when isolated mitochondria were added only after protein synthesis has been arrested by the addition of cycloheximide) or in a reaction in which synthesis and import were permitted to occur simultaneously. We found that the import of a precursor protein consisting of the amino-terminal mitochondrial targeting sequence of cytochrome oxidase subunit IV fused to mouse dihydrofolate reductase is very inefficient in a strictly posttranslational reaction, whereas efficient import is observed if precursor synthesis and import are coupled. The same result was obtained when we analyzed the import of bulk endogenous yeast mitochondrial proteins in this system. Finally, we found that the insertion of the yeast outer membrane protein porin is also several times more efficient when synthesis and insertion are coupled.     

Binding of a tightly folded artificial mitochondrial precursor protein to the mitochondrial outer membrane involves a lipid-mediated conformational change

An artificial mitochondrial precursor protein (the presequence of cytochrome oxidase subunit IV fused to mouse dihydrofolate reductase) binds to isolated yeast mitochondrial outer membranes and to liposomes whose phospholipid composition resembles that of outer membranes. In both cases, binding is strongly inhibited by low temperature or methotrexate (which stabilizes the dihydrofolate reductase moiety) and partly inhibited by adriamycin (which binds to acidic phospholipids). Binding is accompanied by partial unfolding of the protein. Binding of the urea-denatured fusion protein to outer membranes or liposomes is insensitive to low temperature, methotrexate, or adriamycin. These results, and those reported in the accompanying paper (Eilers, M., Endo, T., and Schatz, G. (1989) J. Biol. Chem. 264, 2945-2950) suggest that import of this fusion protein into isolated mitochondria involves at least partial unfolding by acidic phospholipids on the mitochondrial surface.     

Energy requirements for unfolding and membrane translocation of precursor proteins during import into mitochondria

ATP is involved in conferring transport competence to numerous mitochondrial precursor proteins in the cytosol. Unfolded precursor proteins were found not to require ATP for import into mitochondria, suggesting a role of ATP in the unfolding of precursors. Here we report the unexpected finding that a hybrid protein containing the tightly folded passenger protein dihydrofolate reductase becomes unfolded and specifically translocated across the mitochondrial membranes independently of added ATP. Moreover, interaction of the precursor with the mitochondrial receptor components does not require ATP. The results suggest that ATP is not involved in the actual process of unfolding during membrane translocation of precursors. ATP rather appears to be necessary for preventing the formation of improper structures of precursors in the cytosol and for folding of imported polypeptides on (and release from) chaperone-like molecules in the mitochondrial matrix.     

The first twelve amino acids (less than half of the pre-sequence) of an imported mitochondrial protein can direct mouse cytosolic dihydrofolate reductase into the yeast mitochondrial matrix.

Yeast cytochrome c oxidase subunit IV (an imported mitochondrial protein) is made as a larger precursor with a transient pre-sequence of 25 amino acids. If this pre-sequence is fused to the amino terminus of mouse dihydrofolate reductase (a cytosolic protein) the resulting fusion protein is imported into the matrix space, and cleaved to a smaller size, by isolated yeast mitochondria. We have now fused progressively shorter amino-terminal segments of the subunit IV pre-sequence to dihydrofolate reductase and tested each fusion protein for import into the matrix space and cleavage by the matrix-located processing protease. The first 12 amino acids of the subunit IV pre-sequence were sufficient to direct dihydrofolate reductase into the mitochondrial matrix, both in vitro and in vivo. However, import of the corresponding fusion protein into the matrix was no longer accompanied by proteolytic processing. Fusion proteins containing fewer than nine amino-terminal residues from the subunit IV pre-piece were not imported into isolated mitochondria. The information for transporting attached mouse dihydrofolate reductase into mitochondria is thus contained within the first 12 amino acids of the subunit IV pre-sequence.     

Protein folding causes an arrest of preprotein translocation into mitochondria in vivo

With vital yeast cells, a hybrid protein consisting of the amino-terminal third of the precursor to cytochrome b2 and of the entire dihydrofolate reductase was arrested on the import pathway into mitochondria. Accumulation of the protein in the mitochondrial membranes was achieved by inducing a stable tertiary structure of the dihydrofolate reductase domain. Thereby, three salient features of mitochondrial protein uptake in vivo were demonstrated: its posttranslational character; the requirement for unfolding of precursors; and import through translocation contact sites. The permanent occupation of translocation sites by the fusion protein inhibited the import of other precursors; it did, however, not lead to leakage of mitochondrial ions, implying the existence of a channel that is sealed around the membrane spanning polypeptide segment.     

Transport of the yeast ATP synthase beta-subunit into mitochondria. Effects of amino acid substitutions on targeting.

We have isolated the yeast ATP2 gene encoding the beta-subunit of mitochondrial ATP synthase and determined its nucleotide sequence. A fusion between the N-terminal 15 amino acid residues of beta-subunit and the mouse cytosolic protein dihydrofolate reductase (DHFR) was transcribed and translated in vitro and found to be transported into isolated yeast mitochondria. A fusion with the first 35 amino acid residues of beta-subunit attached to DHFR was not only transported but also proteolytically processed by a mitochondrial protease. Amino acid substitutions were introduced into the N-terminal presequence of the beta-subunit by bisulphite mutagenesis of the corresponding DNA. The effects of these mutations on mitochondrial targeting were assessed by transport experiments in vitro using DHFR fusion proteins. All of the mutants, harbourin from one to six amino acid substitutions in the first 14 residues of the presequence, were transported into mitochondria, though at least one of them (I8) was transported and proteolytically processed at a much reduced rate. The I8 mutant beta-subunit also exhibited poor transport and processing in vivo, and expression of this mutant polypeptide failed to complement the glycerol- phenotype of a yeast ATP2 mutant. More remarkably, the expression of I8 beta-subunit induced a more general growth defect in yeast, possibly due to interference with the transport of other, essential, mitochondrial proteins.     

Amino-terminal deletions in the presequence of an imported mitochondrial protein block the targeting function and proteolytic cleavage of the presequence at the carboxy terminus

Subunit IV of yeast cytochrome oxidase is made in the cytosol with a 25-residue presequence. This presequence targets subunit IV into mitochondria and is removed by a protease in the matrix space. Here we show that removal of as few as 4 amino-terminal residues from the subunit IV presequence (which had been attached to the cytosolic protein dihydrofolate reductase) blocks import of the protein into mitochondria and proteolytic removal of the presequence by the soluble matrix protease. Thus, this protease requires not only an appropriate cleavage site at the carboxy-terminal end of the presequence, but also information at the extreme amino terminus of the presequence.     

A synthetic presequence reversibly inhibits protein import into yeast mitochondria

We show that a synthetic peptide corresponding to the N-terminal 22 residues of the cytochrome c oxidase subunit IV presequence blocked import of pre-subunit IV into yeast mitochondria. The 22-residue peptide pL4-(1-22) did not alter the electrical potential across the mitochondrial inner membrane (the delta psi). Inhibition of import was reversible and could be overcome by the addition of increased amounts of precursor. Two other peptides, pL4-(1-16) and pL4-(1-23), which correspond to, respectively, the N-terminal 16 and 23 residues of the same presequence, also blocked import of pre-subunit IV. However, pL4-(1-16) was a much weaker inhibitor of import, while the inhibitory effect of pL4-(1-23) was due to its ability to completely collapse the delta psi. pL4-(1-22) seems to be a general inhibitor of mitochondrial import, in that it also blocked uptake of several other proteins. These included the precursors of the yeast proteins cytochrome c oxidase subunit Va, the F1-ATPase beta subunit, mitochondrial malate dehydrogenase, and the ATP/ADP carrier. In addition, uptake of two non-yeast precursor proteins (human ornithine transcarbamylase and a cytochrome oxidase subunit IV-dihydrofolate reductase fusion), was also blocked by the peptide. Subsequent studies revealed that pL4-(1-22) did not block the initial recognition or binding of proteins to mitochondria. Rather, our results suggest that the peptide acts at a subsequent translocation step which is common to the import pathways of many different precursor proteins.     

The cleavable prepiece of an imported mitochondrial protein is sufficient to direct cytosolic dihydrofolate reductase into the mitochondrial matrix

The cleavable prepiece of the precursor to yeast cytochrome c oxidase subunit IV (an imported mitochondrial protein) was attached to the amino-terminus of mouse dihydrofolate reductase (a cytosolic protein) by gene fusion. The resulting fusion protein was imported into the matrix of isolated, energized yeast mitochondria and cleaved to a polypeptide whose size was similar to that of authentic dihydrofolate reductase.     

Investigations on the in vitro import ability of mitochondrial precursor proteins synthesized in wheat germ transcription-translation extract

Mitochondrial precursor proteins synthesized in rabbit reticulocyte lysate (RRL) are readily imported into mitochondria, whereas the same precursors synthesized in wheat germ extract (WGE) fail to be imported. We have investigated factors that render import incompetence from WGE. A precursor that does not require addition of extramitochondrial ATP for import, the F(A)d ATP synthase subunit, is imported from WGE. Import of chimeric constructs between precursors of the F(A)d protein and alternative oxidase (AOX) with switched presequences revealed that the mature domain of the F(A)d precursor defines the import competence in WGE as only the construct containing the presequence of AOX and mature portion of F(A)d (pAOX-mF(A)d) could be imported. Import competence of F(A)d and pAOX-mF(A)d correlated with solubility of these precursors in WGE, however, solubilization of import-incompetent precursors with urea did not restore import competence. Addition of RRL to WGE-synthesized precursors did not stimulate import but addition of WGE to the RRL-synthesized precursors or to the over-expressed mitochondrial precursor derived from the F1beta ATP synthase precursor inhibited import into mitochondria. The dual-targeted glutathione reductase precursor synthesized in WGE was imported into chloroplasts, but not into mitochondria. Antibodies against the 14-3-3 guidance complex characterized for chloroplast targeting were able to immunoprecipitate all of the precursors tested except the F(A)d ATP synthase precursor. Our results point to the conclusion that the import incompetence of WGE-synthesized mitochondrial precursors is not presequence dependent and is a result of interaction of WGE inhibitory factors with the mature portion of precursor proteins.     

Sequencing of the nuclear gene for the yeast cytochrome c1 precursor reveals an unusually complex amino-terminal presequence.

Cytochrome c1 is a component of the mitochondrial respiratory chain in most eukaryotes. The protein is coded by nuclear DNA, synthesized as a larger precursor outside the mitochondria and then cleaved to the mature form in two successive steps during its import into the mitochondria. We have cloned the structural gene for yeast cytochrome c1 by functional complementation of a cytochrome c1-deficient yeast mutant with a yeast genomic library in the yeast-Escherichia coli 'shuttle' vector YEp 13. The complete nucleotide sequence of the gene and of its 5'- and 3'-flanking regions was determined. The deduced amino acid sequence of the yeast cytochrome c1 precursor reveals an unusually long transient amino-terminal presequence of 61 amino acids. This presequence consists of a strongly basic amino-terminal region of 35 amino acids, a central region of 19 uncharged amino acids and an acidic carboxy-terminal region of seven amino acids. This tripartite structure of the presequence resembles that of the precursor of cytochrome c peroxidase and supports a previous suggestion on the import pathways of these two precursors.     

Binding of mitochondrial precursor proteins to the cytoplasmic domains of the import receptors Tom70 and Tom20 is determined by cytoplasmic chaperones.

We have reconstituted the early steps of precursor targeting to mitochondria in a defined and soluble system consisting of the cytosolic domains of the yeast mitochondrial import receptors Tom20 and Tom70, precursor to bovine adrenal adrenodoxin (which has a cleavable targeting signal) and rat liver cytosolic chaperones hsp70 and mitochondrial import-stimulating factor (MSF). The Tom70 domain only bound the precursor in the presence of MSF, yielding a precursor-MSF-Tom70 complex; ATP hydrolysis by MSF released MSF and generated a precursor-Tom70 complex whose formation was inhibited by an excess of a functional presequence peptide, but not by 150 mM NaCl. In the presence of the Tom20 domain, ATP caused transfer of the precursor from the precursor-MSF-Tom70 complex to Tom20. The Tom20 domain alone only bound the precursor in the presence of hsp70; hsp70 itself was not incorporated into the resulting complex. Formation of the Tom20-precursor complex was inhibited by excess presequence peptide or by 150 mM NaCl. Similar results were obtained with the ADP/ATP carrier and porin precursors, which both lack a cleaved targeting signal. Correct targeting of a precursor to mitochondrial import receptors thus requires cytosolic chaperones, irrespective of the presence or absence of a cleavable presequence.     

Reinvestigation of the requirement of cytosolic ATP for mitochondrial protein import

Protein import into mitochondria requires the energy of ATP hydrolysis inside and/or outside mitochondria. Although the role of ATP in the mitochondrial matrix in mitochondrial protein import has been extensively studied, the role of ATP outside mitochondria (external ATP) remains only poorly characterized. Here we developed a protocol for depletion of external ATP without significantly reducing the import competence of precursor proteins synthesized in vitro with reticulocyte lysate. We tested the effects of external ATP on the import of various precursor proteins into isolated yeast mitochondria. We found that external ATP is required for maintenance of the import competence of mitochondrial precursor proteins but that, once they bind to mitochondria, the subsequent translocation of presequence-containing proteins, but not the ADP/ATP carrier, proceeds independently of external ATP. Because depletion of cytosolic Hsp70 led to a decrease in the import competence of mitochondrial precursor proteins, external ATP is likely utilized by cytosolic Hsp70. In contrast, the ADP/ATP carrier requires external ATP for efficient import into mitochondria even after binding to mitochondria, a situation that is only partly attributed to cytosolic Hsp70.