Role of α-Subunit of Mitochondrial Processing Peptidase in Substrate Recognition*

Mitochondrial processing peptidase is a heterodimer consisting of α-mitochondrial processing peptidase (α-MPP) and β-MPP. We investigated the role of α-MPP in substrate recognition using a recombinant yeast MPP. Disruption of amino acid residues between 10 and 129 of the α-MPP did not essentially impair binding activity with β-MPP and processing activity, whereas truncation of the C-terminal 41 amino acids led to a significant loss of binding and processing activity. Several acidic amino acids in the region conserved among the enzymes from various species were mutated to asparagine or glutamine, and effects on processing of the precursors were analyzed. Glu353 is required for processing of malate dehydrogenase, aspartate aminotransferase, and adrenodoxin precursors. Glu377 and Asp378 are needed only for the processing of aspartate aminotransferase and adrenodoxin precursors, both of which have a longer extension peptide than the others studied. However, processing of the yeast α-MPP precursor, which has a short extension peptide of nine amino acids, was not affected by these mutations. Thus, effects of substitution of acidic amino acids on the processing differed with the precursor protein and depended on length of the extension peptides. α-MPP may function as a substrate-recognizing subunit by interacting mainly with basic amino acids at a region distal to the cleavage site in precursors with a longer extension peptide.

Most mitochondrial proteins encoded in the nucleus are synthesized as precursor proteins with extension peptides at the N termini and are targeted to the mitochondria. After import of the precursors into the mitochondria, the extension peptides are proteolytically cleaved off by a matrix-located metallopeptidase, mitochondrial processing peptidase (MPP) 1 (1)(2)(3). Purification of MPP from various species demonstrated that this peptidase is a heterodimer consisting of ␣and ␤-subunits (4 -6). Sequence analysis (7)(8)(9)(10)(11)(12)(13)(14) revealed that subunits of MPP have significant sequence homology with a family of endopep-tidases, the pitrilysin family, which includes Escherichia coli pitrilysin (also called protease III), the insulin-degrading enzymes from mammals and insects, and the N-arginine dibasic convertase from the rat (15). All members, except the ␣-MPPs, have a metal binding motif, His-Xaa-Xaa-Glu-His. The ␣and ␤-subunits of the MPP are also homologous to core 2 and core 1 proteins, respectively, of the mitochondrial ubiquinol-cytochrome c oxidoreductase (bc 1 complex), a component of the respiratory chain (16). Plant MPP was purified as subunits of the bc 1 complex located in an inner membrane (16 -18). In Neurospora crassa, a core I protein of the bc 1 complex is identical with ␤-MPP in the matrix (19).
Sequence analysis of extension peptides of mitochondrial protein precursors revealed that the peptides vary in length and sequence and are rich in positively charged amino acids among hydrophobic amino acids (20 -22). Studies using radiolabeled precursors and synthetic peptides as a substrate demonstrated that an arginine residue at Ϫ2 position and basic amino acid residues distal to the cleavage site are essential for specific cleavage by MPP (23)(24)(25)(26). Aromatic and bulky hydrophobic amino acids at position ϩ1 and flexible linker regions containing proline and/or glycine between the two basic residues are also required (27).
␤-MPP contains a metal-binding motif, HXXEH, an inverted sequence of HEXXH presented by a zinc-dependent protease thermolysin, and site-directed mutagenesis showed that this site is responsible for the catalytic activity of MPP (28,29). However, ␤-MPP alone has no apparent activity, and both subunits are indispensable for MPP activity (28, 30 -32). Therefore, ␣-MPP seems to participate in functions other than catalytic ones. Cross-linking analyses suggested that a substrate binding domain might reside in ␣-MPP (31), and key amino acid residues that undergo substrate binding in ␣-MPP remain to be determined. We report here that acidic amino acids in the C-terminal portion of ␣-MPP are involved in substrate recognition and are mainly responsible for binding of distal basic amino acids of longer extension peptides.

MATERIALS AND METHODS
cDNAs-DNAs of the precursor form of ␣and ␤-MPP were obtained from the yeast genome by the polymerase chain reaction using as primers oligonucleotides synthesized with reference to the published sequence (8,9). To obtain DNAs of the mature forms of ␣and ␤-MPP, the NcoI site that has an initiation ATG codon was introduced before the first amino acids of these proteins. Both DNAs were finally inserted into expression vectors pET-3d or pACYC184. The hexahistidine tag was introduced to a C-terminal of either ␣or ␤-MPP.
A cDNA of bovine aspartate aminotransferase (AAT) was from a bovine liver cDNA library using as primers oligonucleotides synthesized with reference to the published sequence (35). Isolation of a cDNA of bovine adrenodoxin (Ad) has been published (36). A cDNA for mouse malate dehydrogenase (MDH) was a kind gift from Dr. T. Tsuzuki of Kyushu University (37).
Expression and Purification of Recombinant Yeast MPP-DNAs encoding ␣and ␤-MPP were introduced into E. coli strain BL21, and the transformed cell was cultured in LB medium for 24 h at 25°C. The harvested cells were suspended in 10 mM Tris-HCl buffer, pH 7.5, containing 10 mM MgSO 4 , 50 mM KCl, 15% glycerol, lysed by sonication, and centrifuged at 10,000 ϫ g for 10 min. The resultant supernatant was loaded on a 1-ml nickel-chelating column (Amersham Pharmacia Biotech) equilibrated with 20 mM HEPES-KOH buffer, pH 7.5, containing 0.1% Tween 20. The column was washed with 20 mM HEPES-KOH buffer, pH 7.5, containing 0.1% Tween 20, 100 mM NaCl, 50 mM imidazole, 30% glycerol, and a recombinant MPP was eluted with 500 mM imidazole in the same buffer. Purified MPP was stored at Ϫ20°C until use.
Mutation of ␣-MPP and pAAT-To obtain a cDNA of ␣-MPP(⌬10 -129), ␣-MPP in pBluescript SK was digested at internal HindIII and ClaI sites, and the resultant larger fragment was self-ligated after filling with a Klenow fragment. To create the ␣-MPP(⌬144 -450) precursor, a cDNA of ␣-MPP precursor was digested with ClaI and BamHI, and the larger fragment was self-ligated. C-terminal deletion mutants were obtained by polymerase chain reaction from cDNA of ␣-MPP in pBluescript SK using appropriate oligonucleotides with stop codon as primers. Site-directed mutants of ␣-MPP were obtained by a polymerase chain reaction-based overlap extension method using oligonucleotides with mutations and newly introduced restriction sites as primers ( Table I). Mutation of the AAT precursor was directed to Arg 8 and/or His 18 . To introduce the mutation, we used the polymerase chain reaction-based overlap extension method, as described above (Table I). Mutations were verified by DNA sequencing. All the cDNAs were inserted into pET-3d or pACYC184.
Synthesis of Radiolabeled Precursors-cDNAs of bovine AAT, bovine Ad, mouse MDH, and yeast ␣-MPP were inserted into pSP-65 (Promega) or pGEM-3Z (Promega). The mRNAs for these precursor proteins were transcribed by SP6 RNA polymerase and translated in a rabbit reticulocyte lysate in the presence of [ 35 S]methionine (ICN Biomeds) (38).
Assay of Processing Activity-Radiolabeled precursors were incubated with MPP at 30°C for 10 min in reaction buffer consisting of 20 mM HEPES-KOH, pH 7.4, 0.1% Tween 20, and 0.5 mM MnCl 2 , and the reaction was stopped by adding 1 mM EDTA. The processing products were separated by SDS-PAGE and visualized by fluorography.
Determination of Kinetic Parameters-Synthetic peptide, MDH(1-24), MLSALARPVGAALRRSFSTSA was incubated with a native and a recombinant MPP in 400 ml of a reaction buffer consisting of 20 mM HEPES-KOH, pH 7.4, 0.1% Tween 20 at 25°C. Reactions were stopped by adding 0.05% trifluoroacetic acid. The processed products were analyzed by reverse-phase HPLC on a RPC-SC18 column (4.6 ϫ 230 mm) (JASCO, Japan). Elution was carried out with a linear gradient established between 1 and 50% acetonitrile in 0.05% trifluoroacetic acid for 33 min. The substrate and product peptides were detected by monitoring at 215 nm. Michaelis constants (K m ) and maximum velocities (V max ) were determined from Lineweaver-Burk plots. Molecular activity (k cat ) was calculated by V max /[MPP]. The HPLC apparatus was used with Shimadzu LCA-10A.
Amino Acid Sequence Analysis-For N-terminal sequence analysis, about 400 pmol of purified sample was applied to Protein Sequencing System 491A (Perkin-Elmer).

Expression and Purification of Recombinant Yeast MPP-To
reconstitute an active MPP, cDNAs for mature ␣and ␤-MPP were inserted into pET-3d expression vector, and two proteins were co-expressed in E. coli BL21 cells. The recombinant MPP was purified using a nickel-chelating column through a hexahistidine tag attached to the C-terminal end of ␣-MPP. Analysis by SDS-PAGE showed that ratio of ␣and ␤-MPP in the enzyme was nearly 1:1 (Fig. 1B). The kinetic parameters of the recombinant enzyme were determined and compared with those of a native MPP partially purified from yeast mitochondria. Michaelis constants (K m ) for a synthetic peptide modeled on the extension peptide of MDH were determined to be 0.24 and 0.44 M for the native and recombinant MPP, respectively. Molecular activities (k cat ) were 130 and 115 min Ϫ1 for the native and recombinant MPP, respectively. These results indicate that the recombinant MPP has substantially the same enzymatic properties as native MPP.
C-terminal Domain of ␣-MPP Is Important for Interaction with ␤-MPP-To determine the functional domain of ␣-MPP, we constructed N-or C-terminal-deleted mutant and examined their activities. Mutants were co-expressed with ␤-MPP and purified with the nickel-chelating column, and interactions with ␤-MPP were analyzed. Recovery of the mutant ␣-MPP and ␤-MPP was analyzed on SDS-PAGE gels followed by protein staining or immunoblotting. An ␣-MPP(⌬10 -129) mutant with amino acid residues from 10 to 129 deleted was co-purified with ␤-MPP that had a His tag at the C-terminal end (Fig. 1, A and C). The ratio of subunits in the purified complex was determined by immunoblotting using specific anitibodies. Although the extract from ␣-MPP(⌬10 -129)/␤-MPP-His-transformed cell contained similar amounts of the subunits, the recovery of co-purified ␣-MPP(⌬10 -129) was lower than that of the counterpart (Fig. 1D). The same held true for the ␣-MPP(⌬10 - 129)His⅐␤-MPP complex, in which ␣-MPP(⌬10 -129) was tagged (data not shown). We next constructed C-terminal deletion mutants in which 41, 84, and 122 amino acid residues were truncated from the C-terminal portion of ␣-MPP by introducing stop codons at appropriate positions (Fig. 1A). After co-expression of the mutants with ␤-MPP-His, formation of the complex was examined using the nickel-chelating column as described above. In contrast to the N-terminal deletion mutant, none of the truncated mutants co-eluted with ␤-MPP-His (Fig. 1C). Immunoblotting analysis showed that recovery of ␣-MPPDC41 into the extract was almost the same as that for the wild-type ␣-MPP, but ␣-MPPDC41 did not interact with ␤-MPP (Fig. 1D).
The wild-type and mutant MPP consisting of ␣-MPP(⌬10 -129) and ␤-MPP-His were incubated with the fluorogenic peptide, then the processing activity was measured using a fluorescence spectrophotometer (27). The mutant enzyme processed the fluorogenic peptide, although the activity was about 10% that of the wild-type enzyme. Taken together, deletion of the C-terminal portion resulted in a complete loss of the potential to interact with ␤-MPP and to exhibit enzyme activity, whereas that of the N terminus maintained those potentials to some extent. These results indicate that the C-terminal portion of ␣-MPP is more important for functions of ␣-MPP.
Mutation of Acidic Amino Acids at the C-terminal Region Does Not Affect Processing of ␣-MPP Precursor-Sequence alignment of MPPs revealed that ␣-MPP contains several highly conserved domains consisting of about 30 to 60 amino acids; some in the central and C-terminal portions of the subunit are characteristic in that they are rich in acidic amino acids. To determine whether these conserved residues interact with basic amino acids in the extension peptide of the precursors, five site-directed mutants with acidic amino acids replaced by neutral ones were constructed; mutations were aimed at a pair of glutamic acids at positions 197 and 201, a pair of glutamic acids at positions 351 and 353 and an aspartic acid at position 352, a pair of glutamic acids at position 377 and an aspartic acid at position 378, a glutamic acid at position 395, and a pair of aspartic acids at positions 405 and 406. Glutamic and aspartic acids were replaced by glutamine and asparagine, respectively. A hexahistidine tag was attached to the C-terminal end of each mutant, and the mutant protein was co-expressed with ␤-MPP in E. coli BL21 and then purified using the nickel-chelating column. All mutants were purified in a complex form at a ratio of nearly one to one between ␣and ␤-MPP, which indicated that these mutations did not affect complex formation (data not shown).
Processing activities of these mutants were examined for precursor proteins, yeast ␣-MPP, MDH, AAT, and Ad. Lengths of extension peptides of these proteins varied; the shortest one was 13 amino acids long (yeast ␣-MPP), and the longest one was 50 amino acids long (Ad). We first examined the effect of mutations in acidic amino acid residues in the C-terminal region of ␣-MPP on processing of p-␣-MPP, a precursor with a short extension peptide. The precursor protein was translated in vitro in the presence of [ 35 S]methionine. It was difficult to clearly distinguish between the precursor and mature ␣-MPP, probably because of small difference in size (data not shown). We then used a truncated precursor, p-␣-MPP(⌬140 -450), as substrate. As shown in Fig. 2, the wild-type MPP processed about 20% p-␣-MPP (⌬140 -450) in 10 min, but the processed product migrated at a slower rate than in vitro synthesized authentic mature ␣-MPP(⌬140 -450). All the mutant MPPs also processed the p-␣-MPP(⌬140 -450) with substantially the same processing efficiencies, indicating that the acidic amino acid residues mutated are not essential for processing of this precursor.
The processed product was apparently larger in size than the authentic mature protein, which was constructed according to a report that p-␣-MPP is cleaved at a peptide bond between Ile 13 and Ala 14 in the N-terminal sequence of Met-Leu-Arg-Asn-Gly-Val-Gln-Arg-Leu-Tyr-Ser-Asn-Ile-Ala (39). Since we found that the arginine residue at the Ϫ2 position from a cleavage site is indispensable for recognition by MPP (24), we asked if MPP would recognize a motif other than "Ϫ2 arginine" for the processing or would it cleave at a potential cleavage site, between Leu 9 and Tyr 10 , upstream of this position. We then attempted to determine the correct cleavage site of p-␣-MPP by MPP. The precursor bearing a His tag at its C-terminal end was constructed and was co-expressed with a non-His-tagged mature ␣and ␤-MPP in E. coli cells. ␣-MPP precursor expressed was processed in situ by the active MPP complex expressed in the same cells. The His-tagged ␣-MPP was purified using a nickel-chelating column and subjected to analysis of the N-terminal amino acid sequence. The amino acid sequence proved to be Tyr-Ser-Asn-Ile-Ala-Arg-Thr-Glu-Asn-Phe, corresponding exactly to the sequence from Tyr 10 of the published N-terminal sequence of ␣-MPP precursor (39). These findings indicate that the primary processing site by MPP locates four amino acids upstream from the published N-terminal amino acid of the mature ␣-MPP. This implies that the "arginine Ϫ2 and hydrophobic ϩ1 rule" (24,27) is applicable to the ␣-MPP precursor. The difference in the N-terminal amino acid could be due to either a further processing by an unknown peptidase or to an artificial modification during a purification step of ␣-MPP from yeast cells.
Acidic Amino Acids Are Required for Processing of Precursors of MDH, AAT, and Ad-We next examined effects of mutation on processing of precursors with a longer extension peptide. When the MDH precursor was incubated with the wild-type MPP, about 70% pMDH was processed to the intermediate form in 10 min (Fig. 3A). However, processing by the E351Q/ D352N/E353Q mutant was never evident. Other mutants processed the precursor as a similar rate as seen with the wild-type enzyme. These results suggest that acidic amino acid residues at position 351/352/353 have effects on the processing, and the processing mechanism of MDH differs from that of yeast ␣-MPP.
When pAAT was incubated with wild-type MPP, it was processed to a mature form at a rate of 50% cleavage for 10 min (Fig. 3B). The processing efficiency by either E351Q/D352N/ E353Q mutant or the E377Q/D378N mutant was reduced to under 10% that of the wild type. Other mutants processed the substrate at a rate comparable with that seen with the wild type. Similar patterns were obtained in the case of processing of the adrenodoxin precursor. The wild-type MPP processed the pAd to an intermediate form at a rate of more than 70% cleavage in 10 min. Processing by the E351Q/D352N/E353Q mutant was significantly reduced, and processing products by the E377Q/D378N mutant were never evident (Fig. 3C). Other mutant MPPs cleaved off the extension peptide at a rate sim-ilar to that seen with the wild-type enzyme. Therefore, effects of acidic amino acids on the processing vary with precursor.
Acidic Region of ␣-MPP Recognizes Distal Basic Amino Acids in Extension Peptides-Replacements of acidic amino acid residues affected the processing of precursors that carry a longer extension peptide, which means that these acidic amino acids recognize distal basic amino acids in the extension peptide. To gain support for this thesis, we replaced basic amino acids in the extension peptide of pAAT and examined processing of these substrates by mutant MPPs. An arginine residue at position 8, a histidine residue at position 18, or both were mutated to alanine. These mutants are designated R8A, H18A, and R8A/H18A, respectively. When each of R8A and H18A was incubated with the wild-type MPP, the processing efficiency was reduced to a rate half that of pAAT (data not shown). In the case of the R8A/H18A mutant, processing was not evident. Therefore, two basic amino acid residues in the extension peptide of AAT are required for efficient processing, probably as distal basic amino acids. The relationship between distal basic amino acids and acidic amino acid clusters in ␣-MPP was investigated using E351Q/D352N/E353Q or E377Q/D378N or a double mutant, which carried both E351Q/D352N/E353Q and E377Q/D378N mutations. When the R8A substrate was incubated with the double mutant, there were essentially no processing products (Fig. 4). Although processing of R8A was at a comparable rate to that seen with the wild-type enzyme for the E377Q/D378N mutant, the E351Q/D352N/E353Q mutant processed the mutant precursor at a slow rate. These findings suggest that the cluster of Glu 351 -Asp 352 -Glu 353 is effective for the recognition of His 18 of pAAT. In case of the H18A substrate, little processing was detected by the double mutant, whereas E351Q/D352N/E353Q and E377Q/D378N mutants cleaved the substrate at almost half the rate of the wild-type enzyme. These observations indicate that both acidic clusters recognize Arg 8 of pAAT. Taken together, Glu 351 -Asp 352 -Glu 353 recognized basic amino acids at multiple sites in the extension peptide, thereby suggesting the possibility that orientation of Glu 351 -Asp 352 -Glu 353 changes from His 18 to Arg 8 when His 18 is replaced with alanine. Glu 353 , Glu 377 , and Asp 378 Are Critical Amino Acid Residues for Substrate Recognition -To elucidate amino acid residues in acidic clusters critical for substrate recognition, single amino acid substituted mutants, E351Q, E353Q, E377Q, and D378N were constructed. Since Asp 352 is not conserved among species, it was replaced by asparagine. When pMDH was incubated with the E351Q and E353Q mutants, the latter did not produce the intermediate form of MDH (Fig. 5A). Thus, in the acidic cluster from 351 to 353, Glu 353 is critical for recognizing distal basic amino acids in the extension peptide. The E353Q mutant also did not process pAd and pAAT. In the case of E377Q and D378N mutants, these two amino acids seemed to contribute equally to the processing. When either the E377Q or the D378N mutant was incubated with pAAT, the substrate was processed at half the rate seen with the wild type (Fig. 5B). This reduction was also observed when pAd served as substrate (Fig. 5C); therefore, the same recognition mechanism generally functions in case of different substrates. DISCUSSION MPP Requires Both ␣and ␤-Subunits for processing of mitochondrial protein precursors (28, 30 -32). Although ␤-MPP functions as a catalytic subunit, the functions of ␣-MPP were not clearly identified. We obtained evidence that the ␣-subunit participates in substrate recognition by the enzyme. The participation of basic amino acids in extension peptides as the recognition signal for processing (23-26, 33, 40) means that there are negatively charged amino acid residues or clusters at the domain that are responsible for substrate recognition in the enzyme. We substituted acidic amino acid residues in the central or the C-terminal region of ␣-MPP, which are conserved in subunits among many species. The effects of this substitution on the processing differed with the precursor protein depending on length of the extension peptides (Fig. 6). Mutations affected only processing of precursors with longer extension peptides, and effects of mutations seem to correlate with location of distal basic amino acids in the extension peptides. A distal basic amino acid of yeast ␣-MPP precursor, which has a short extension peptide, does not seem to be required for the processing since Arg 3 of yeast ␣-MPP could replace alanine without affecting the processing, 2 and only one arginine residue at position Ϫ2 functioned as a basic amino acid interacting with 2 K. Shimokata, unpublished data.  Fig. 3 and 4. B, a model for interaction. ϩ and Ϫ symbols represent basic and acidic amino acid residues, respectively. Broken lines represent putative electrostatic interactions with extension peptide and ␣-MPP. Arrowheads indicate cleavage sites by MPP. Amino acid residues at position ϩ1, ϩ2, and ϩ3 are given in filled circles. Mϩϩ shows an active site metal. Yeast ␣-MPP precursor, pMDH, and pAAT are illustrated. Basic amino acids at distal sites are Arg 7 of pMDH, Arg 8 , and His 18 of pAAT.
MPP. In this case, there is no interaction between the extension peptide and these acidic residues in ␣-MPP. In contrast, Arg 7 of pMDH, Arg 13 and His 18 of pAAT and a pair of Arg 25 and Arg 27 and another pair of Arg 34 and Arg 37 of pAd, which proved necessary for the processing (23,26), may interact with either Glu 353 or Glu 377 plus Asp 378 of the ␣-MPP (Fig. 6B).
The proximal (position Ϫ2) arginine and distal basic amino acids are not sufficient for specific recognition of extension peptides by MPP, because arginine and lysine residues are present at multiple positions in all proteins. We reported earlier that, in addition to these residues, an aromatic or hydrophobic amino acid residue at position ϩ1 is important for processing of the precursors (27). Hydrophilic amino acids like serine, threonine, and histidine at positions ϩ2 and ϩ3 are also important for effective processing. 3 Thus, a combination of distal basic amino acids and residues around the cleavage site appears to be responsible for recognition of the substrates and for determination of the cleavage site by MPP. However, since not all the extension peptides have a full set of these elements, especially of distal basic residues, all the elements are not indispensable for the processing. A structure around the cleavage site that is formed of the proximal arginine and the residues at position 1, 2, and 3 is probably common to all the precursors, and number and position of distal basic residues vary among the extension peptides. The peptidase must have several substrate binding sites to cope with these distal basic residues, whereas each extension peptide seems to have structural elements corresponding to only some of these binding sites. Flexible linker regions rich in glycine and proline in the middle portion of the extension peptides seem to aid extension peptides in fitting into multiple binding sites in MPP, because the region is necessary for effective cleavage by MPP (27). The concerted action of common and additional structural elements in the extension peptide may be able to render strict specificity and high affinity to MPP for precursors with structures little in common, and cooperative formation of the substrate binding pocket by two subunits could make it feasible.
The ␣and ␤-subunits of the MPP are homologous to core 2 and core 1 proteins, respectively, of mitochondrial ubiquinolcytochrome c oxidoreductase (bc 1 complex), a component of the respiratory chain (16). The tertiary structure of the bc 1 complex from bovine heart mitochondria has recently been reported (34). Core 1 and core 2 proteins are structurally similar and consist of two domains of roughly equal size and almost identical folding topology. In dimer form, the N-terminal and Cterminal portions of the core 1 and core 2, respectively, face each other, and they form a groove leading to the internal cavity, the wall of which is lined with mostly hydrophilic amino acid residues. From sequence similarity between core proteins and subunits of MPP and the finding that core 1 and core 2 proteins of the bc 1 complex function as the processing peptidase of plant mitochondria (16 -18), one can expect that MPP is similar in structure to the core protein complex. We found that complex formation of the two subunits of MPP is essential for high affinity binding of substrates to the enzyme. 4 These findings, together with the present results, suggest that the sub-strate binding pocket of the MPP is formed by both subunits, and on the wall of the pocket, metal binding sites of the ␤-subunit and negatively charged amino acids in both subunits are present.