Skip to main content
NCBI home page
Protein Science : A Publication of the Protein Society logoLink to Protein Science : A Publication of the Protein Society
. 2002 May;11(5):1026–1035. doi: 10.1110/ps.3760102

Timing and structural consideration for the processing of mitochondrial matrix space proteins by the mitochondrial processing peptidase (MPP)

Abhijit Mukhopadhyay 1, Philip Hammen 1, Mary Waltner-Law 1, Henry Weiner 1
PMCID: PMC2373553  PMID: 11967360

Abstract

Most mitochondrial matrix space proteins are synthesized as a precursor protein, and the N-terminal extension of amino acids that served as the leader sequence is removed after import by the action of a metalloprotease called mitochondrial processing peptidase (MPP). The crystal structure of MPP has been solved very recently, and it has been shown that synthetic leader peptides bind with MPP in an extended conformation. However, it is not known how MPP recognizes hundreds of leader peptides with different primary and secondary structures or when during import the leader is removed. Here we took advantage of the fact that the structure of the leader from rat liver aldehyde dehydrogenase has been determined by 2D-NMR to possess two helical portions separated by a three amino acid (RGP) linker. When the linker was deleted, the leader formed one long continuous helix that can target a protein to the matrix space but is not removed by the action of MPP. Repeats of two and three leaders were fused to the precursor protein to determine the stage of import at which processing occurs, if MPP could function as an endo peptidase, and if it would process if the cleavage site was part of a helix. Native or linker deleted constructs were used. Import into isolated yeast mitochondria or processing with recombinantly expressed MPP was performed. It was concluded that processing did not occur as the precursor was just entering the matrix space, but most likely coincided with the folding of the protein. Further, finding that hydrolysis could not take place if the processing site was part of a stable helix is consistent with the crystal structure of MPP. Lastly, it was found that MPP could function at sites as far as 108 residues from the N terminus of the precursor protein, but its ability to process decreases exponentially as the distance increases.

Keywords: Mitochondrial protein processing, MPP, mitochondrial processing peptidase, precursor protein, aldehyde dehydrogenase, leader peptide processing

Mitochondria must import the vast majority of proteins that are required for its proper function. These proteins, synthesized in the cytosol, contain an N-terminal presequence that targets them to the mitochondria. For most mitochondrial matrix space proteins the presequence is removed after import by the action of the mitochondrial processing peptidase (MPP) (Ou et al. 1989). MPP is a heterodimeric protein (Kleiber et al. 1990; Paces et al. 1993), which itself is imported (Arretz et al. 1994). The α subunit is responsible for peptide binding and β for catalysis (Luciano and Geli 1996). Some preproteins are processed in two proteolytic steps—the first catalyzed by MPP, and the second by the mitochondrial intermediate peptidase (Kalousek et al. 1992). Although hundreds of proteins are imported, targeted by unique presequences, MPP is the only known protease involved in cleaving all the presequences that are removed. A subset of mitochondrial proteins, including rhodanese (Miller et al. 1991) and thiolase (Arakawa et al. 1987), are not processed after mitochondrial import. The way by which MPP achieves sequence specificity is of general interest because the enzyme carries out a specific function on a wide variety of substrates, but does not act on all potential substrates.

The information that a presequence must contain to interact with and be cleaved by MPP has been investigated, but is not thoroughly understood. Several proteolytic sites containing four or five residues have been identified from statistical analyses of mitochondrial presequences (Gavel and von Heijne 1990; Schneider et al. 1998). Arginine residues located two or three residues before the cleaved bond appear to be important for processing. Site-directed mutagenesis of Args either two (Arretz et al. 1994) or three (Hammen et al. 1996b) residues before the processing site led, as predicted, to a loss in processing. An aromatic residue in the first position after the cleaved peptide bond was shown to enhance the processing of malate dehydrogenase (Ogishima et al. 1995), but according to the statistical analysis of Gavel and von Heijne (1990), Ser appears to be more common at this position. For the spinach processing enzyme, it has been suggested that the helix-forming ability near the processing site is critical (Sjoling et al. 1996). Previous work also suggested that residues of the N-terminal flanking region of the presequence contained information that influenced processing of precursor proteins by MPP. The requirement of basic residues in this region has been observed (Song et al. 1996). It has been argued that a helical conformation in the N-terminal segment is not required for processing (Kraus et al. 1988). With a chimeric precursor of COX IV presequence fused to DHFR, it was found that at least 13 residues were required for N-terminal to the cleaved bond to allow processing by MPP (Hurt et al. 1987). It was shown that regions flanking the processing site of the cytochrome b2 presequence were not essential for processing, but appeared to enhance processing efficiency in vitro (Klaus et al. 1996). These authors concluded that the N-terminal region must allow the residues around the site of processing to adopt a conformation that is optimal for MPP function. The crystal structure of MPP has been solved very recently (Taylor et al. 2001). It has been shown that leader peptides of MDH and COX IV bind MPP in an extended form. The substrate binding region of MPP is located in a large central cavity between the two subunits, and is filled with negatively charged amino acids. Positively charged signal peptides, thus, have a better opportunity to bind to the substrate binding pocket of MPP. The COX IV leader peptide forms a short β strand from residues 16–18 that hydrogen bonds with the β subunit of MPP, and residues 7–8 of the MDH leader peptide forms hydrogen bonds with the β subunit. Although the structure gives valuable insight into how signal peptides interact with MPP, it is not known when presequences are removed by MPP during translocation.

It has been demonstrated that proteins import through the mitochondrial membrane in an extended state (Rassow et al. 1990). It has been shown in plants that the translocation channel and MPP/bc1 complex are located separately in the inner membrane. As a result, the processing site of a leader peptide has to reach the MPP/bc1 complex to be processed (Dessi et al. 2000); further, there is no link between protein translocation and protein processing. However, in yeast or in mammals, MPP is a soluble entity in the matrix space, so it could cleave the precursor protein as soon as the processing site is exposed or it could process at a later stage of import. We have studied the import of ALDH, a tetrameric enzyme present in the matrix. pALDH imports as a monomer, and presumably in an extended form during import (Jeng and Weiner 1991). Here, we investigated different aspects of processing to try to understand when processing occurs and what are the conformational constraints on processing.

Results

Making the linker-deleted leader processable

The structure of a synthetic leader sequence corresponding to the one from rat liver pALDH has been determined in a micellar milieu by 2D-NMR to contain two helices, residues 1–10 and 14–19, separated by a three amino acid linker (Karslake et al. 1990), as illustrated in Figure 1 ▶. When the linker was removed, a long continuous helical peptide was found. The precursor protein containing either leader could be imported into mitochondria, but only the native one was processed by MPP (Thornton et al. 1993). Although we do not know precisely why linker-deleted pALDH was not processed, we rationalized that it was because the long continuous helix went through the processing site (Thornton et al. 1993). To test for this idea we made specific substitutions near the normal site of processing, RLL16/S, in an attempt to restore processing. An L16Y mutation introduced one of the known processing motifs R-X-Y/S that was used to make the normally nonprocessed rhodanese processible (Waltner and Weiner 1995). This mutation did not restore processability to the pALDH (-RGP) mutant (Fig. 2-I ▶, lane 8). Using the parameters established by Creamer et al. (1995), it was possible to calculate that there was a high probability that the stable helix still existed in the potential processing site. Similar calculations showed that an L16S mutant would be less helical. This mutant was constructed because R-X-S/S is a known processing motif (Gavel and von Heijne 1990). This mutant was processed after incubation with MPP (Fig. 2-I ▶, lane 6) although not to the same extent as the native precursor. It appears, then, that a segment that disfavors helix formation would be more likely to be recognized by MPP. It is also consistent with the results that showed the leader peptide binds to MPP in an extended conformation (Taylor et al. 2001). Knowing that the processing site most likely cannot be part of a stable helix allowed us to design a variety of modified leaders to determine when processing occurred during the import process.

Fig. 1.

Fig. 1.

Open in a new tab

Leader sequence of pALDH and schematic diagram of chimeric proteins. The sequences of the leader of pALDH and linker deleted (-RGP) are shown. Below them are the chimeric proteins used in this study. (A) pALDH; (B) leader sequence of pALDH fused to pALDH by two mature amino acid linkers; (C) leader sequence of pALDH fused to pALDH (-RGP) by two mature amino acid linkers; (D) leader sequence of pALDH plus 24 mature amino acids fused to pALDH (-RGP); (E) leader sequence of pALDH plus 24 mature amino acids fused to pALDH; (F) pALDH (-RGP) plus 24 mature amino acids fused to pALDH; (G) two native leaders were fused to pALDH (-RGP); (H) leader sequence of pALDH plus 150 mature amino acids fused to pALDH; (I) leader sequence of pALDH with R3,10Q mutant fused to pALDH with R3,10Q mutant by two mature amino acid linkers; (J) proline and glycine have been introduced between two R3,10Q leaders. Symbols: a processing site (down arrow); a nonprocessing site (down arrow with line at top); a leader sequence (filled section); a mature sequence (open section); linker-deleted leader (dotted section). Import amounts were presented besides the corresponding figures (using pALDH import as 100%). NA, not applicable.

Fig. 2.

Fig. 2.

Open in a new tab

Processing of ALDH mutants by purified MPP. (I) In vitro processing of native pALDH, ALDH (-RGP), L16S (-RGP), and L16Y (-RGP) were performed as described in Materials and Methods, employing a 30-min incubation at 27°C. (I) Lanes 1 and 2 contain native pALDH. Lanes 3 and 4 contain linker-deleted pALDH. Lanes 5 and 6 contain L16S (-RGP). Lanes 7 and 8 contain L16Y (-RGP). Lanes 1, 3, 5, and 7 contain translated protein. Lanes 2, 4, 6, and 8 contain translated protein after incubation with purified MPP. (II) Lanes 1 and 2 contain A, pALDH; lanes 3 and 4 represent chimeric protein B, and lanes 5 and 6 contain F. Lanes 1, 3, and 5 represent translated (TNT) products, and lanes 2, 4, and 6 contain TNT incubated with MPP. (III) Lane 1 represents translated protein of E, and lane 2 represents protein treated with MPP. (IV) MPP processing of modified pALDHs (Wang and Weiner 1993). Lanes 1 and 2 represent pALDH; lanes 3 and 4 represent CC-AC that has the C-terminal domain of Cox IV (CC) fused to the C-terminal domain of ALDH (AC); lanes 5 and 6 represent RN-AC that has the N-terminal domain of rhodanses (RN) fused to the C-terminal domain of ALDH; lanes 7 and 8 represent TN-AC that has the N-terminal domain of thiolase (TN) fused to the C-terminal domain of ALDH. Lanes 1, 3, 5, and 7 represent TNT products, and lanes 2, 4, 6, and 8 represent protein treated with MPP. (V) MPP processing of denatured chimeric protein C. Lanes 1 and 2 represent pALDH; lanes 3, 4, and 5 represent protein C. Lanes 1 and 3 are TNT products, and lanes 2 and 4 represent protein treated with MPP. Lane 5 represents denatured C treated with MPP. (VI) MPP processing of chimeric protein I and J. Lanes 1, 2, and 3 represent I, lanes 4 and 5 represent pALDH, and lanes 6, 7, and 8 represent J. Lanes 1, 4, and 6 represent TNT products, and lanes 2 and 7 represent protein treated with MPP for 30 min; lanes 3 and 8 represent protein treated with MPP for 60 min.

Processing does not occur while the precursor protein is entering the matrix space

If processing occurred while the precursor was entering the matrix space one could expect that each time a processing site was presented, MPP would cleave it. If, though, more of the protein needed to be imported before processing was initiated, then other factors could determine when and where processing occurred. First, we investigated constructs containing two pALDH leaders to determine if two leaders would be removed in one step or two. The leader from pALDH was first fused to wild-type pALDH (Fig. 1B ▶). Although two processing sites existed, one at the end of each leader, it was possible that the first one would not be recognized for it could be part of a helix extending from residue 14 of the first leader through residue 10 of the second leader. If, though, processing occurred just after the leader passed through the inner membrane, then the first site might be exposed. When this two-leader precursor was incubated with MPP, it was found not to be extensively processed at the second site and not at all at the first site (Fig. 2-II ▶, lane 4). In addition, a chimeric protein containing two identical mutant leaders called R3,10Q (Fig. 1-I ▶) (Hammen et al. 1996b) was made. MPP processed the R3,10Q and the pALDH leaders with the same efficiency. When this chimeric protein was incubated with purified MPP, no cleavage was found at the first processing site (Fig. 2-VI ▶, lanes 2,3). However, when proline and glycine were introduced between two leaders (Fig. 1-J ▶), then processing was detected after the in vitro MPP reaction (Fig. 2-VI ▶, lanes 7,8). This finding leads us to suggest that the helix extending through the first procesing site was disrupted by the introduction of proline and glycine. In contrast, after import into mitochondria, both constructs were found to be processed to the size of the mature protein (Fig. 3-I ▶, lanes 4,3-V, lane 4). There are two possible explanations for finding that processing occurred after import. One is that as the protein passed though the membrane in an unfolded state each leader was sequentially removed by MPP. The other possibility is that in the matrix space MPP was capable of removing both leaders in one step, a feat it could not do in a test tube.

Fig. 3.

Fig. 3.

Open in a new tab

In vitro import of chimeric proteins. Import was performed as described in Materials and Methods. (I) Lanes 1 and 2 represent pALDH (A); lanes 3 and 4 represent B; lanes 5 and 6 contain E, and lanes 7 and 8 represent F. Lanes 1, 3, 5, and 7 are for newly translated proteins, and lanes 2, 4, 6, and 8 are for those found after import into mitochondria so the bands represent imported protein. (II) Lanes 1 and 2 contain pALDH (A), and lanes 3 and 4 contain C. Lanes 1 and 3 are translated proteins, and lanes 2 and 4 contain protein bands found after import. (III) Lanes 1 and 2 are translated proteins of A and D, respectively, and lanes 3 and 4 are the bands found after import of A and D. (IV) Lanes 1 and 4 contain A and H, and lanes 2 and 3 represent imported bands of A without proteinase K and with proteinase K treatment, respectively. Lanes 5 and 6 are imported proteins of H without proteinase K and with proteinase K treatment, respectively. (V) Lanes 1 and 3 represent translated proteins of A and I, respectively, and lanes 2 and 4 represent proteins found after import of A and I. (VI) Import of A and E in the presence of digitonin. Lanes 1 and 6 represent TNT products of A and E. Lanes 2 and 7 represent import of A and E without digitonin. Lanes 3 and 8 represent import of A and E in the presence of 0.1% digitonin. Lanes 4 and 9 represent import of A and E in the presence of 0.2% digitonin. Lanes 5 and 10 represent import of A and E in the presence of 0.3% digitonin.

The linker-deleted leader, which is not processed, was employed to permit us to determine which model was valid. The leader from pALDH was fused to the linker-deleted pALDH (Fig. 1C ▶). If MPP removed each leader independently, then the final product would be the size of the precursor protein because the second, linker-deleted leader should not be proteolysed (Thornton et al. 1993). Unexpectedly, after import, a band corresponding to the size of the mature protein was found (Fig. 3-II ▶, lane 4). This implies that a single processing event occurred. How was it possible for processing to occur after the linker-deleted leader? A very long helix, running from residue 14 of the first leader through the processing site of the second leader, could form based on the known structures of the individual components. This helix would consist of approximately 25 residues. Because of the two-residue segment joining the two leaders, the helix could not display uniform hydrophobic and hydrophilic surfaces for its full length. When the helix forms, it may do so in a more transient way than the linker-deleted sequence. If that were the case, the second processing site might not be continuously in a helical conformation, and could be cleaved by MPP. When this chimeric protein was incubated with purified MPP, no processing was found consistent with our notion. However, when the protein was denatured in urea and then MPP was added, it was hydrolyzed by MPP (Fig. 2-V ▶, lane 5), confirming that the first processing site was indeed part of the helix. To test further the idea that the helix prevented processing, chimeric protein D was used to disrupt the potential long helix.

Chimeric protein D had 20 amino acids of mature ALDH placed between the native and the linker-deleted leaders. It could be expected that after import the first leader would be removed, leaving behind the 20 amino acids fused to the 16-residue nonprocessable linker-deleted leader. After import, a band corresponding to a mature protein with 36 additional amino acids was found (Fig. 3-III ▶, lane 4). Had MPP hydrolyzed the protein as it crossed the membrane, construct C should also have been processed at the first site. The processing patterns obtained with chimeras B, C, and D are consistent with a model of MPP action where processing does not occur as the protein is coming through the membrane, but at a later stage of import, after residues located beyond the processing site were imported.

We further investigated the requirement for a spacer to be located N-terminal to the processing site by introducing 24 residues of mature ALDH between the two native leaders. This was chimera E in Figure 1 ▶. E was incompletely processed to the mature sized protein after import. As shown in Figure 3-I ▶, lane 6, in addition to a band corresponding to the mature protein, a second band labeled ipALDH was found that corresponded to the size of mature ALDH with approximately 43 additional residues (the 24 residues of added mature plus the second leader). The intensities of these two bands were nearly identical. Next, the linker-deleted leader, followed by 24 residues of the mature protein, was fused to pALDH (Fig. 1F ▶). The linker-deleted pALDH leader was placed at the most N-terminal end to determine if MPP could process the precursor if the site were 62 residues from the end. After import, just two bands were obtained (Fig. 3-I ▶, lane 8). One was the size of the unprocessed construct; the other, representing approximately 30% of the imported protein, was the size of mature ALDH. The results suggest that as the length increases before the processing site the efficiency of hydrolysis decreases, from 100% when there are 19 residues prior to the site to 50% with 43 residues down to 30% with 62 residues.

We further tested for the mitochondrial location of pALDH and construct E after import. Varying concentrations of digitonin were added to the mitochondria after import of pALDH and chimeric protein E in the presence of protenase K. Imported proteins were observed in the mitoplast, confirming that both pALDH and chimeric protein E were processed in the matrix (Fig. 3-VI ▶).

Unexpected results were found when the two chimeras (E and F) were treated with pure MPP. With construct E only 5% of the translated protein was processed to the size of mature ALDH (Fig. 2-III ▶, lane 2), while 50% of ipALDH was produced under the condition of an in vitro assay. In the case of construct F, no processing could be detected (Fig. 2-II ▶, lane 6). Processing was once again different when comparing intact mitochondria with the pure MPP.

To test further that hydrolysis occurs after import, the linker-deleted leader was employed, but now with two functional leaders fused to it (G, in Fig. 1 ▶). Thus, a total of three leaders were attached to mature ALDH; two were potentially processable, while the third one was not. In addition, a construct with three native leaders fused to mature ALDH was also investigated. After import and SDS-PAGE separation, only mature ALDH was found from the latter (data not shown), but three bands were found on the autoradiogram from the former (Fig. 4 ▶, lane 4). These were the size of the mature protein, the full-length construct, and one intermediate-sized band that corresponded to pALDH. If processing occurred while the precursors were crossing the membrane one would have expected to see only a band corresponding to pALDH (-RGP), as the first two leaders would already have been removed before the linker-deleted leader enters the matrix space. Finding mature ALDH shows that all three leaders were removed. Because it has been well established that linker-deleted pALDH can not be processed, it can be assumed that the processing at the third site occurred while one or two of the other leaders were still attached to the protein. Because there was a band corresponding to the full-length construct, it can be concluded that much of the protein had to be imported prior to the initiation of processing. If this were not the case, then after the removal of the first leader, a protein corresponding to chimera C would have remained and that construct was essentially completely processed. The only way it would be possible to obtain a protein the size of linker-deleted pALDH would be to have processing occur such that the two leaders were removed in one step.

Fig. 4.

Fig. 4.

Open in a new tab

Import of chimeric protein with three added leader sequences. Chimeric protein G had three leader sequences: the first two were native, and the third one was linker deleted. Lanes 1 and 3 show translation product of pALDH and protein G, respectively. Lane 2 shows the imported protein band of pALDH; lane 4 shows the import of protein G, and lane 5 represents G incubated with purified MPP.

Pure MPP did not process the three-leader chimera (Fig. 4 ▶, lane 5). This is consistent with the data obtained from chimera B, and supports the idea that if the processing site is part of a stable helix, processing does not occur. In this case the stable helix would be between the first two leaders.

MPP does not process tetrameric pALDH

We have shown that recombinantly expressed pALDH becomes an active tetrameric enzyme (Jeng and Weiner 1991). This form of the enzyme was incubated with pure MPP for 30 min at 27°C. Western blot analysis showed no evidence that processing occurred. Thus, MPP cannot cleave folded pALDH. Therefore, it appears that processing occurs in vivo before the formation of the tetrameric enzyme.

Processing at potential internal MPP recognition sites

Most mammalian mitochondrial leaders are fewer than 35 amino acids in length. It has been shown here that processing could occur after 60 residues. If processing occurred while the protein was crossing the inner membrane MPP could possibly recognize any potential processing site, independent of where the site was in the protein. Although there is no specific processing recognition site, ALDH appears to have three internal sites that might be hydrolyzed by MPP. When pALDH, synthesized in reticulocyte lysate, was incubated with MPP, mature ALDH was produced, but no band corresponding to smaller protein fragments were detected. In pALDH, the next potential processing motif occurs in the 108 position from the N terminus with a sequence RLL/Y. To show that this sequence could be a processing site if it were located at the end of the leader, the serine at position 20 was converted to a tyrosine. This protein was processed by purified MPP (data not shown) showing that the site located at position 108 could be a viable processing site.

We often observed that following import of pALDH an additional faint band was detected on the gel corresponding to a protein with a mass of nearly 40 kD (Fig. 5 ▶). This is the size of the expected product if MPP would cleave at a site located at position 108. To determine whether or not this minor band came from pALDH, the precursor was synthesized in the presence of antibody against ALDH. Antibody did not inhibit the synthesis of pALDH, but was found to inhibit the import of pALDH into mitochondria (Fig. 5 ▶, lanes 4,5). The same concentration of antibody did not inhibit the import of pOTC, showing that antibody against ALDH did not affect the import machinery. The band corresponding to 40 kD was not observed if antibody was included in the import assay. Thus, it appeared that the lower band came from pALDH after it was imported, and leads us to suggest that the second processing site was cleaved very poorly by MPP in the mitochondrial matrix space. This observation was verified when the import was done in the presence of CCCP, a compound that dissipates the inner membrane potential and prevents mitochondrial import (Geissler et al. 2000). In the presence of CCCP no imported pALDH protein or any 40-kD protein band was found (data not shown). This indicated that the 40-kD band came from pALDH.

Fig. 5.

Fig. 5.

Open in a new tab

Import of pALDH in the presence of antibody against ALDH. Lane 1 represents translated product of pALDH; lanes 2 and 3 represent import into mitochondria without proteinase K and with proteinase K treatment, respectively. Lane 4 represents translated product of pALDH in the presence of its antibody. After import in the presence of the antibody, proteinase K was added and mitochondria were reisolated by centrifugation and, like with all other samples treated with SDS treatment buffer and the reaction mixture, was run on SDS-PAGE followed by autoradiography (lane 5). p and m refer to the location where the precursor and mature ALDH would be found.

Because MPP can cleave up to 108 residues, a new construct was made where the first 150 residues of pALDH were fused to the N terminus of another pALDH molecule (Fig. 1H ▶). With it, there was one processing site after the usual 19 residues and a potential second one after 178 residues. From the SDS-PAGE gel it can be noted that after import, processing occurred only after the first 19 residues, and none was detected at the second site (Fig. 3-IV ▶, lane 6). Consistent with this finding was the observation that if the first leader was the nonprocessable, then only a band corresponding to the original size of the construct was found (data not shown). Therefore, it appears that MPP can actually cleave up to approximately 108 amino acids from the N terminus of pALDH, but it will not cleave when there are 150 residues in front of the recognition sequence. Although it can recognize an internal site, its ability to process decreases dramatically after 40 amino acids from the N terminus.

Requirement of the N-terminal portion of the ALDH presequence for processing

The data presented in this study show that the processing site could not be part of a stable helix for it to be recognized by MPP. We also showed that it was possible for processing to take place with a longer N-terminal extension before the site. We next investigated whether local structure N-terminal to the site was critical. We took advantage of the structural features of known leaders, determined by 2D-NMR. As expected, neither ΔN-ALDH nor ΔN-Cox IV (leaders missing their N-terminal helical segments) were processed by pure MPP, verifying that 10 or 12 residues prior to the processing site are necessary for the protease to function. To study the need of a particular structure on the N-terminal side of the processing site the N-terminal helix-forming segment of the ALDH leader was replaced with various peptides that were previously used to study different aspects of import (Wang and Weiner 1993). These included the helix-forming segments from rhodanese and thiolase as well as the C-terminal random-coil segment from COX IV. Of the three, only the thiolase chimera was not processed by MPP (Fig. 1-IV ▶). Because the other two, one helical the other random coil, were processed, it can be concluded that factors other than distal secondary structure influence the ability of MPP to process at a cleavage site.

Discussion

In addition to the information required for recognition and proteolysis by MPP, mitochondrial presequences also contain information necessary to allow it to be imported in to the mitochondria. The sequence requirement for import appears to be nonspecific, because 25% of randomly generated peptides could function as import signals (Lemire et al. 1989). However, most of the randomly generated presequences that functioned as artificial matrix targeting presequences were not cleaved by MPP. The information for import and processing seems to be distinct from one another.

MPP is different from other peptidases in that it does not recognize a unique sequence in the target protein or a unique position such as what is done by an exo-peptidase. It has been shown that there are preferred recognition motifs on the preprotein that often include an arginine residue located one or two residues on the N-terminal side of the cleavage site. Despite the lack of specificity, the enzyme does not appear to be capable of the random cleavage of polypeptide chains at sites that could correspond to a cleavage site. A precursor such as liver ALDH possesses some potential internal cleavage sites, yet these are not processed. If processing occurred as the precursor were entering the matrix space after crossing the inner mitochondrial membrane in an unfolded state, one could expect a protease to have access to every potential processing site. From the data presented in this study one can conclude that the efficiency of cleavage decreases dramatically as the number of residues on the N-terminal side increases, as illustrated in Figure 6 ▶. Further, it appears that the cleavage site is not part of a helix.

Fig. 6.

Fig. 6.

Open in a new tab

Ability of MPP to process at sites distal from the N termini. The residue number is the number of amino acids between the N terminus and processing sites. The solid line is the fit of the data to an exponential decay function. The efficiency of processing diminishes greatly as the number of amino acids increase. MPP most likely does not function as an endo-peptidase for this reason, and these sites most likely will be in a folded domain of the protein.

MPP removes hundreds of signal peptides from precursor proteins; however, it is not known when during the import process the signal is removed. We have investigated this event using pALDH as a model. Precursor ALDH is synthesized in the cytosol and targeted to the mitochondrial matrix by a 19 amino acid leader sequence. In the matrix space, the signal sequence is cleaved by MPP. The processing event can occur in an early stage of import (as soon as the processing site is exposed to the matrix), or in a late stage of import such as after folding to the native conformation. It has been already established that proteins traverse through the mitochondrial membrane in an extended state (Rassow et al. 1990). In that report, signal sequence and different lengths of mature cyt b2 were fused to DHFR and import into mitochondria was performed in the presence of methotrexate, which prevented the complete import of the DHFR protein. When 65 amino acids of cyt b2 were fused to DHFR no processing was found; however, processing was observed when 82 amino acids were fused. In these experiments, import intermediates were immobilized in the import channel, and with time, some processing could occur. To gain a better understanding of processing of a totally imported protein, an approach using multileader sequences fused to ALDH was employed in our study.

We tested the ability of MPP to process an active recombinantly expressed pALDH homotetramer and found it could not. We previously showed that commercially available active rhodanese did not inhibit pALDH processing by MPP, but that newly synthesized rhodanese in reticulocyte lysate, which presumably remained unfolded, inhibited processing (Waltner and Weiner 1995). These results indicated that MPP does not interact with folded preproteins. To test when pALDH was processed in the matrix we took the advantage of the fact that removing the three-residue linker (RGP) from pALDH makes it a continuous helix that still retained a processing motif but was not processed (Thornton et al. 1993). Substitution of a serine for leucine in the processing site allowed the modified precursor protein to be processed by pure MPP while substitution by a tyrosine did not. Calculations show that the more polar serine disfavored helical formation while tyrosine favored its formation (Creamer et al. 1995). Thus, disruption of the helix-forming ability in the processing motif allowed it to become recognized by MPP. If processing occurred while the unfolded leader was entering the mitochondria, helicity at the processing site should not have been a factor. It appears then, that processing does not occur as the site becomes available, but must take place when more of the protein has traversed the inner membrane

When one leader was followed by another, with just two amino acid residues from the mature protein separating them, MPP could not process the first one, and processed very poorly at the second site. We suggest that the first processing site in all the double leader constructs, where the separation was by just two amino acids, was within the helix. When some mature protein was found, such as with construct B, it cannot be concluded that the processing did not occur at the first site. If, though, the native leader was followed by the nonprocessable linker-deleted leader, a mature sized ALDH was found after import. This was an unexpected event. To be processed at the second site it would be necessary for the processing site to be nonhelical. Leader peptides have been shown to become helical when induced by an environment that allows for a burial of hydrophobic surface. In construct C, the two leaders, separated by two mature amino acids, do not form an amphiphilic helix that aligns hydrophobic and hydrophilic surfaces for its full length. Conceivably, the helical conformation could form, but because it did not have a surface with which to interact, it would remain as a helix only transiently. Thus, this possible helix would be less stable compared to what would be found in the linker-deleted leader, and could be processable. We verified that the first processing site was really a part of the helix by adding proline and glycine in between the two signals in a chimeric protein containing two R3,10Q leaders. The first processing site was hydrolyzed by pure MPP, suggesting that the long helix was disrupted by proline and glycine. Again, chimeric protein C was not processed by purified MPP, but was processed when it was denatured in urea.

With chimeric protein D, the first site was processed but the second one was not. The second site would be in the helical structure expected for the linker-deleted leader. Distance from the processing site is not an issue here because construct E had 43 residues before the cleavage site was processed well by MPP in the mitochondria. The data obtained from the triple leader construct were consistent, with the conclusion that the processing site could not be part of a stable helix.

It appears that processing does not occur as the precursor is crossing the inner membrane but happens after a sufficient number of amino acids enter the matrix space so that they can begin to form some secondary structure. To become active after import, proteins must fold, and multimeric proteins such as ALDH must assemble into their native form. However, completely folded recombinantly expressed pALDH cannot be processed. From the data we have presented, it appears that hydrolysis catalyzed by MPP in the yeast mitochondrial matrix coincides with the folding and assembly process. Potential internal processing sites for MPP that had obtained a folded secondary or tertiary structure would be resistant to the protease.

It is possible for MPP to hydrolyze at positions as far as 108 residues from the N terminus. During import into isolated mitochondria it was observed that some bands corresponding to distal proteolysis were found. Presumably, if this occurred in vivo, the resulting protein would not fold properly, and would ultimately be destroyed. The three potential internal processing sites in ALDH are actually located in regions that are either buried or are part of helices (Steinmetz et al. 1997). Thus, the very limited proteolysis observed most likely occurred prior to the folding of the protein. Even though MPP could function as an endo-peptidase, it most likely does not do that in vivo.

Although the processing site cannot be part of a helix, it appears that residues distal to the site do not affect the ability of MPP to hydrolyze the protein. Others have shown that at least 12 residues are necessary for cleavage to occur. Here we show that these residues upstream of the processing site could be of widely divergent composition and structure without affecting the ability of MPP to function. This explains why the protease can process such a diverse set of precursor proteins.

Recently, the structures of bound leaders have been determined. The first was an X-ray structure (Abe et al. 2000) followed by an NMR-derived structure of the leader of ALDH bound to Tom20, a protein involved in translocation across the inner mitochondrial membrane (Muto et al. 2001). The second was the recently solved X-ray structure of two different leaders bound to MPP (Taylor et al. 2001). With the former, the leader was helical, while for the latter they were unfolded, especially at the cleavage site. Our suggestion that the leaders could not be processed if they were helical is supported by this crystallographic data. Despite having the structure and knowing that processing occurs with unfolded proteins, we still cannot explain the precise interactions that allow MPP to hydrolyze at some sites but not others. The reason that MPP does not function as an endo-peptidase could simply be that the precursor proteins start to fold and the inner sites are simply not exposed. Moreover, it appears from our results that the signal sequence of pALDH is removed by MPP at a later stage of import and not as the cleavage site is first exposed to the matrix space.

Materials and methods

Construction of chimeric pALDH

The cDNA for both pALDH and pCOX IV was amplified using PCR with flanking BamH 1 and Nde1 sites, and was subsequently cloned into plasmid pT7-7. CC-AC (see legend of Fig. 2 ▶) was constructed by using PCR and directly adding the sequence encoding the CC segment with a flanking 5`Nde1 site and a flanking 3`BamH 1 site. The corresponding PCR product was digested with Nde1 and BamH 1, and cloned into pT7-7. RN-AC (see legend of Fig. 2 ▶) was constructed as follows. The ALDH gene was amplified from cDNA encoding the residue Arg 11 to the stop codon with flanking BamH 1 site using PCR. The plasmid pT7-7 and the cDNA encoding the initial 15 residues from rhodanese was amplified with BamH 1 sites. The two PCR products were digested with BamH I, ligated, and the orientation of the insert confirmed using Pst1 digestion. The BamH I site existing between the ALDH linker and rhodanese was looped out using PCR. TN-AC (see legend of Fig. 2 ▶) was made by PCR and then directly adding the sequence encoding the initial 14 residues of thiolase with a flanking 5`-NdeI and 3`-BamH I site. After BamH I and NdeI digestion, the PCR product was ligated into plasmid pT7-7. Mutation of Leu16 in linker-deleted ALDH and mutation of Ser 20 in pALDH were carried out using oligonucleotide-directed mutagenesis. All mutations were confirmed by sequencing. Each double or triple leader sequence was made in the following way. Leader sequence of ALDH plus the desired length of mature amino acids was amplified with a flanking SphI site using pALDH in pGEM-3 as a template. PCR products and the vector containing pALDH were digested with SphI, and the vector was subsequently treated with alkaline phosphatase and ligated to the PCR products. The orientation was confirmed by sequencing.

In vitro processing activity

A solution containing pure α and β subunits of yeast MPP expressed in Escherichia coli was used. A processing assay consisted of 2 μL TNT Quick Coupled transcription and translation system (Promega), 1 μL (approx. 0.05 μg) of the processing enzyme, 2 μL buffer containing 10 mM Hepes-KOH, 1 mM dithiothreitol, and 0.1 mM MnCl2 (final volume was adjusted to 20 μL with water), and was incubated for 30 min at 27°C. The reaction was terminated by the addition of an equal volume of SDS treatment buffer. Samples were subjected to SDS-PAGE and analyzed by a PhosphorImager. The amount of processing was quantitated using densitometry. The percent processing was calculated as the amount of processed protein divided by the total amount of protein in the assay.

Import of preproteins into isolated mitochondria

Radiolabeled preproteins were synthesized in the presence of [35S] methionine using the TNT Quick Coupled transcription and translation system (Promega). Saccharomyces cerevisiae mitochondria were isolated according to Glick and Pon (1995). Preproteins were incubated with mitochondria for 30 min at 30°C in Import Buffer (0.6 M sorbitol, 50 mM HEPES, 50 mM KCl, 10 mM MgCl2, 2.5 mM EDTA, 2.0 mM KH2PO4, and 1.0 mg/mL Fatty-acid free BSA, pH 7.0). Final volume of the import mixture was 50 μL. Import reactions were performed as described (Hammen et al. 1996a). Quantification of import was performed using the band intensities from SDS-polyacrylamide gels that were analyzed by a PhosphorImager storage technology (Molecular Dynamics). The level of import was defined as the ratio of the total counts of the protease-protected bands divided by the initial counts provided in the assay. Each experiment was reproduced at least three times.

Digitonin extraction

Digitonin extraction was performed essentially as described by Koll et al. (1992). Briefly, after import proteanase K was added and subsequently inhibited by PMSF. Mitochondria were reisolated by centrifugation and diluted with different concentrations (0 to 0.3%) of digitonin in import buffer. The solutions were mixed rapidly and incubated on ice for 60 sec, diluted with 5 volumes of ice-cold import buffer, and centrifuged 10 min at 12,000g. Pelleted mitochondria were dissolved in SDS treatment buffer.

Miscellaneous

The PCR reagents were purchased from Perkin-Elmer. Restriction enzymes and T4 DNA Ligase were obtained from New England BioLabs. The plasmid pGEM-3 was purchased from Promega.

Acknowledgments

This work was supported in part by Grants AA10795 and GM 53269 from the National Institute of Health. This is Journal Paper 16725 from the Purdue University Agricultural Experiment Station.

The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 USC section 1734 solely to indicate this fact.

Abbreviations

  • ALDH, aldehyde dehydrogenase

  • pALDH, precursor form of ALDH

  • pALDH(-RGP), the precursor form missing the three amino acid linker, RGP, that separates the two helical segments of the leader

  • MPP, mitochondrial processing peptidase

  • DHFR, dihydrofolate reductase

  • thiolase, 3-oxoacyl-CoA thiolase

  • COX IV, cytochrome oxidase subunit IV

  • MDH, malate dehydrogenase

  • cccp, carbonyl cyanide m-chlorophenylhydrazone

Article and publication are at http://www.proteinscience.org/cgi/doi/10.1110/ps.3760102.

References

  1. Abe, Y., Shodai, T., Muto, T., Mihara, K., Torii, H., Nishikawa, S., Endo, T., and Kohda, D. 2000. Structural basis of presequence recognition by the mitochondrial protein import receptor Tom20. Cell 100 551–560. [DOI] [PubMed] [Google Scholar]
  2. Arakawa, H., Takiguchi, M., Amaya, Y., Nagata, S., Hayashi, H., and Mori, M. 1987. cDNA-derived amino acid sequence of rat mitochondrial 3-oxoacyl-CoA thiolase with no transient presequence: Structural relationship with peroxisomal isozyme. EMBO J. 6 1361–1366. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Arretz, M., Schneider, H., Guiard, B., Brunner, M., and Neupert, W. 1994. Characterization of the mitochondrial processing peptidase of Neurospora crassa. J. Biol. Chem. 269 4959–4967. [PubMed] [Google Scholar]
  4. Creamer, T.P., Srinivasan, R., and Rose, G.D. 1995. Modeling unfolded states of peptides and proteins. Biochemistry 34 16245–16250. [DOI] [PubMed] [Google Scholar]
  5. Dessi, P., Rudhe, C., and Glaser, E. 2000. Studies on the topology of the protein import channel in relation to the plant mitochondrial processing peptidase integrated into the cytochrome bc1 complex. Plant. J. 24 637–644. [DOI] [PubMed] [Google Scholar]
  6. Gavel, Y. and von Heijne, G. 1990. Cleavage-site motifs in mitochondrial targeting peptides. Protein Eng. 4 33–37. [DOI] [PubMed] [Google Scholar]
  7. Geissler, A., Krimmer, T., Bomer, U., Guiard, B., Rassow, J., and Pfanner, N. 2000. Membrane potential-driven protein import into mitochondria. The sorting sequence of cytochrome b(2) modulates the deltapsi-dependence of translocation of the matrix-targeting sequence. Mol. Biol. Cell. 11 3977–3991. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Glick, B.S. and Pon, L.A. 1995. Isolation of highly purified mitochondria from Saccharomyces cerevisiae. Methods Enzymol. 260 213–223. [DOI] [PubMed] [Google Scholar]
  9. Hammen, P.K., Gorenstein, D.G., and Weiner, H. 1996a. Amphiphilicity determines binding properties of three mitochondrial presequences to lipid surfaces. Biochemistry 35 3772–3781. [DOI] [PubMed] [Google Scholar]
  10. Hammen, P.K., Waltner, M., Hahnemann, B., Heard, T.S., and Weiner, H. 1996b. The role of positive charges and structural segments in the presequence of rat liver aldehyde dehydrogenase in import into mitochondria. J. Biol. Chem. 271 21041–21048. [PubMed] [Google Scholar]
  11. Hurt, E.C., Allison, D.S., Muller, U., and Schatz, G. 1987. 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. J. Biol. Chem. 262 1420–1424. [PubMed] [Google Scholar]
  12. Jeng, J. and Weiner, H. 1991. Purification and characterization of catalytically active precursor of rat liver mitochondrial aldehyde dehydrogenase expressed in Escherichia coli. Arch. Biochem. Biophys. 289 214–222. [DOI] [PubMed] [Google Scholar]
  13. Kalousek, F., Isaya, G., and Rosenberg, L.E. 1992. Rat liver mitochondrial intermediate peptidase (MIP): Purification and initial characterization. EMBO J. 11 2803–2809. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Karslake, C., Piotto, M.E., Pak, Y.K., Weiner, H., and Gorenstein, D.G. 1990. 2D NMR and structural model for a mitochondrial signal peptide bound to a micelle. Biochemistry 29 9872–9878. [DOI] [PubMed] [Google Scholar]
  15. Klaus, C., Guiard, B., Neupert, W., and Brunner, M. 1996. Determinants in the presequence of cytochrome b2 for import into mitochondria and for proteolytic processing. Eur. J. Biochem. 236 856–861. [DOI] [PubMed] [Google Scholar]
  16. Kleiber, J., Kalousek, F., Swaroop, M., and Rosenberg, L.E. 1990. The general mitochondrial matrix processing protease from rat liver: Structural characterization of the catalytic subunit. Proc. Natl. Acad. Sci. 87 7978–7982. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Koll, H., Guiard, B., Rassow, J., Ostermann, J., Horwich, A.L., Neupert, W., and Hartl, F.-U. 1992. Antifolding activity of hsp60 couples protein import into the mitochondrial matrix with export to the intermembrane space. Cell 68 1163–1175. [DOI] [PubMed]
  18. Kraus, J.P., Novotny, J., Kalousek, F., Swaroop, M., and Rosenberg, L.E. 1988. Different structures in the amino-terminal domain of the ornithine transcarbamylase leader peptide are involved in mitochondrial import and carboxyl-terminal cleavage. Proc. Natl. Acad. Sci. 85 8905–8909. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Lemire, Bd., Fankhauser, C., Baker, A., and Schatz, G. 1989. The mitochondrial targeting function of randomly generated peptide sequences correlates with predicted helical amphiphilicity. J. Biol. Chem. 264 20206–20215. [PubMed] [Google Scholar]
  20. Luciano, P. and Geli, V. 1996. The mitochondrial processing peptidase: Function and specificity. Experientia 52 1077–1082. [DOI] [PubMed] [Google Scholar]
  21. Miller, D.M., Delgado, R., Chirgwin, J.M., Hardies, S.C., and Horowitz, P.M. 1991. Expression of cloned bovine adrenal rhodanese. J. Biol. Chem. 266 4686–4691. [PubMed] [Google Scholar]
  22. Muto, T., Obita, T., Abe, Y., Shodai, T., Endo, T., and Kohda, D. 2001. NMR identification of the Tom20 binding segment in mitochondrial presequences. J. Mol. Biol. 306 137–143 [DOI] [PubMed] [Google Scholar]
  23. Ogishima, T., Niidome, T., Shimokata, K., Kitada, S., and Ito, A. 1995. Analysis of elements in the substrate required for processing by mitochondrial processing peptidase. J. Biol. Chem. 270 30322–30326. [DOI] [PubMed] [Google Scholar]
  24. Ou, W.J., Ito, A., Okazaki, H., and Omura, T. 1989. Purification and characterization of a processing protease from rat liver mitochondria. EMBO J. 8 2605–2612. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Paces, V., Rosenberg, L.E., Fenton, W.A., and Kalousek, F. 1993. The beta subunit of the mitochondrial processing peptidase from rat liver: Cloning and sequencing of a cDNA and comparison with a proposed family of metallopeptidases. Proc. Natl. Acad. Sci. 90 5355–5358. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Rassow, J., Hartl, F.U., Guiard, B., Pfanner, N., and Neupert, W. 1990. Polypeptides traverse the mitochondrial envelope in an extended state. FEBS Lett. 275 190–194. [DOI] [PubMed] [Google Scholar]
  27. Schneider, G., Sjoling, S., Wallin, E., Wrede, P., Glaser, E., and von Heijne, G. 1998. Feature-extraction from endopeptidase cleavage sites in mitochondrial targeting peptides. Proteins 30 49–60. [PubMed] [Google Scholar]
  28. Sjoling, S., Waltner, M., Kalousek, F., Glaser, E., and Weiner, H. 1996. Studies on protein processing for membrane-bound spinach leaf mitochondrial processing peptidase integrated into the cytochrome bc1 complex and the soluble rat liver matrix mitochondrial processing peptidase. Eur. J. Biochem. 242 114–121. [DOI] [PubMed] [Google Scholar]
  29. Song, M.C., Shimokata, K., Kitada, S., Ogishima, T., and Ito, A. 1996. Role of basic amino acids in the cleavage of synthetic peptide substrates by mitochondrial processing peptidase. J. Biochem. (Tokyo) 120 1163–1166. [DOI] [PubMed] [Google Scholar]
  30. Steinmetz, C. G., Xie, P., Weiner, H. and Hurley, T.D. 1997. Structure of mitochondrial aldehyde dehydrogenase: The genetic component of ethanol aversion. Structure 5 701–711. [DOI] [PubMed] [Google Scholar]
  31. Taylor, A.B., Smith, B.S., Kitada, S., Kojima, K., Miyaura, H., Otwinowski, Z., Ito, A., and Deisenhofer, J. 2001. Crystal structures of mitochondrial processing peptidase reveal the mode for specific cleavage of import signal sequences. Structure 9 615–625. [DOI] [PubMed] [Google Scholar]
  32. Thornton, K., Wang, Y., Weiner, H., and Gorenstein, D.G. 1993. Import, processing, and two-dimensional NMR structure of a linker-deleted signal peptide of rat liver mitochondrial aldehyde dehydrogenase. J. Biol. Chem. 268 19906–19914. [PubMed] [Google Scholar]
  33. Waltner, M. and Weiner, H. 1995. Conversion of a nonprocessed mitochondrial precursor protein into one that is processed by the mitochondrial processing peptidase. J. Biol. Chem. 270 26311–26317. [DOI] [PubMed] [Google Scholar]
  34. Wang, Y. and Weiner, H. 1993. The presequence of rat liver aldehyde dehydrogenase requires the presence of an alpha-helix at its N-terminal region which is stabilized by the helix at its C termini. J. Biol. Chem. 268 4759–4765. [PubMed] [Google Scholar]

Articles from Protein Science : A Publication of the Protein Society are provided here courtesy of The Protein Society

RESOURCES