Cooperative formation of a substrate binding pocket by alpha- and beta-subunits of mitochondrial processing peptidase.

Mitochondrial processing peptidase (MPP) specifically recognizes a large variety of mitochondrial precursor proteins and cleaves off N-terminal extension peptides. The enzyme is a metalloprotease and forms a heterodimer consisting of structurally related alpha- and beta-subunits. To investigate the responsibility of MPP subunits for substrate recognition, we monitored interaction of the fluorescent-labeled peptide substrates with the MPP and its subunits. The specific binding of the peptide to the MPP was confirmed by findings of the direct participation of arginine residues in the binding, which are located at position -2 and the position distal to the cleavage site and are essential for the cleavage reaction. MPP bound the substrate peptides with high affinity only in the dimeric complex, and each subunit monomer had about a 30-fold less affinity than the complex. The individual subunit required arginines at different positions in the peptide for binding, although their affinities were much lower than that of MPP. Fluorescence quenching analysis showed that the peptide bound to MPP was buried in the enzyme. Thus, both subunits of MPP might be required for formation of a substrate binding pocket with multiple subsites lying across them.

Most nuclear-encoded mitochondrial proteins are translated on cytoplasmic ribosomes as the precursor forms carrying the N-terminal extension peptides. Extension peptides are required for transport of mitochondrial proteins as the targeting signals and are proteolytically cleaved off during or after import into the matrix by the mitochondrial processing peptidase (MPP) 1 (1)(2)(3)(4). MPP is a metallopeptidase consisting of two structurally related subunits, ␣and ␤-MPP, that are both required for enzyme activity (5)(6)(7)(8)(9)(10). Mutation study on rat ␤-MPP suggested that the metal binding motif in this subunit that is conserved in a pitrilysin superfamily (11), HXXEH, is an active center of the MPP (12).
Despite the lack of sequence homology of the extension peptides, the substrate recognition of MPP is strictly specific. Analysis of amino acid sequence of mitochondrial protein precursors revealed that most extension peptides have an arginine residue at the Ϫ2 position from the cleavage site (13). Previous works suggested the importance of this arginine for specific cleavage by MPP (14 -16). Using synthetic peptides (17)(18)(19) and precursor proteins (20,21) as the substrate, we demonstrated that the structural element in the substrates required for processing is not only proximal arginine but also distal basic amino acid residues from the cleavage site, the flexible linker regions containing proline and/or glycine between the two basic residues, and hydrophobic amino acids at position ϩ1. These results indicate that MPP recognizes a higher order structure of extension peptides.
There are conflicting results concerning the substrate binding of MPP. The intrinsic tryptophan fluorescence study on Neurospora MPP demonstrated that both subunits of MPP can bind substrates with the dissociation constants of the order of sub-M (22). In yeast MPP, cross-linking and surface plasmon resonance analyses showed that ␣-MPP, but not ␤-MPP, binds substrates as efficiently as does the MPP complex (23,24). Which subunit functions for substrate recognition has remained to be determined.
To investigate the subunit responsible for substrate recognition in yeast MPP, we used the fluorescence-labeled peptides. A coumarin derivative, which is an environment-sensitive fluorescence probe, was covalently introduced into the synthetic peptides based on the malate dehydrogenase precursor. The dissociation constants of MPP and the subunits with the fluorescent-labeled peptide were determined by fluorescence emission intensity. We report here that MPP can bind the substrate peptides with high affinity but only in the dimeric complex and that individual subunit interacts with different parts of the extension peptide of precursor protein.

EXPERIMENTAL PROCEDURES
Purification of the Hexahistidine-tagged Yeast MPP and Its Subunits-A hexahistidine tag was introduced into the C termini of yeast ␣and ␤-MPP using the polymerase chain reaction method. The resultant cDNAs were inserted into pET3d vector, leading to pET-␣His and pET-␤His. For co-expression of yeast MPP subunits, a T7 promoter/ nontagged ␤-MPP cassette was inserted in tandem into pET-␣His. The resulting construct, pET-␣His-␤, contained the histidine-tagged ␣-MPP and nontagged ␤-MPP under the T7 promoters. BL21(DE3) strain transformed with each plasmid was cultured for 24 h at 25°C and then harvested by centrifugation at 1,000 ϫ g for 10 min. After sonication of the harvested cells, the suspension was centrifuged at 15,000 ϫ g for 20 min. The resultant supernatant was loaded on a 1-ml nickel trap-chelating column (Amersham Pharmacia Biotech) equilibrated with buffer A (10 mM Hepes-KOH, pH 7.4, containing 500 mM NaCl and 0.01% Tween 20). The column was washed with 10 ml of the buffer A containing 10 mM imidazole and then with 10 ml of the buffer A containing 100 mM imidazole. The hexahistidine-tagged proteins were eluted using buffer A containing 500 mM imidazole. The purity was confirmed by SDS-polyacrylamide gel electrophoresis followed by Coomassie Blue staining.
Preparation of Coumarin-labeled Peptides-Peptides were synthesized using an EPS221 economy peptide synthesizer (ABIMED Analysen-Teknik GmbH, Germany), employing Fmoc (N-␣-9-fluorenylmethoxycarbonyl) strategy. After acetylation of the ␣-amino group of the N-terminal amino acid with acetic anhydride, deprotection, cleavage of the peptides from the resin, and purification by reverse phase HPLC were performed as described (17).
The purified peptides were labeled with 7-diethylaminocoumarin (DAC) at the ⑀-amino group of lysine residue in the C-terminal portion of the peptides (see Fig. 1A and Table I). The reaction was conducted in 100 mM Hepes-KOH (pH 7.4), 40% dimethylformamide, 1 mM succinimidyl ester of DAC, and 0.1 mg/ml peptide for 2 h at 25°C. After precipitation of the peptide by adding 10 volumes of acetonitrile, the precipitate was collected by centrifugation at 5,000 ϫ g for 5 min, washed twice with acetonitrile, and then dissolved in distilled water. The coumarin-labeled peptides were purified by reverse phase HPLC on an Asahipak C4P-50 column (Asahi Chemical Industry, Japan) and dissolved in 50% aqueous dimethylformamide. Concentration of the fluorescence peptides was calculated from the absorbance of the DAC at 430 nm using a molar extinction coefficient of 40,000.
Fluorescence Measurements-Fluorescence was measured at 25°C using a Hitachi F-2000 fluorescence spectrophotometer. Excitation was at 390 nm, and the emission intensity was measured at 470 nm.
For binding assay, the purified ␣-MPP, ␤-MPP, and MPP complex were preincubated with 5 mM EDTA on ice for 30 min. Proteins of a fixed concentration (see legends to tables and figures) were diluted into 20 mM Hepes-KOH (pH 7.4), 30% glycerol, and 1 mM EDTA, then the coumarin-labeled peptides of indicated concentrations were added. After each sample and the blank had been thoroughly mixed and allowed to equilibrate for 1-2 min, the emission spectra were taken, and the fluorescence intensity at 470 nm was read. The dissociation constant, K d , and binding stoichiometry, n, were determined from Scatchard analysis of the data, as follows: In this equation, F and F max are the measured and maximal fluorescence intensity of the peptides, respectively, and L and E are the total peptide and enzyme concentration, respectively. F max was obtained by fitting the saturation curve to a rectangular hyperbola. A plot of 1/(1 Ϫ F/F max ) versus L/(F/F max ) yields a linear function with a slope of 1/K d and an abscissa intercept of nE. In the competition experiments, fluorescence of DAC-MDH5-25 was measured in the presence of various concentrations of nonlabeled peptides.
In the fluorescence quenching analysis, after the binding reaction done as described above, various concentrations (0 -40 mM final) of acrylamide as external quencher were added to sample solutions. For steady-state collisional quenching of the fluorescence, a linear plot was obtained when the data were analyzed according to the Stern-Volmer , where F o is the initial net fluorescence intensity in the absence of quencher, F is the net fluorescence intensity in the presence of quencher, K sv is the Stern-Volmer quenching constant, and [Q] is the quencher concentration.

RESULTS
Binding of the Coumarin-labeled Peptides to MPP-To monitor substrate binding of the MPP and the subunits, a coumarin derivative was introduced to the peptides corresponding to the N-terminal 21 amino acids (Leu 5 to Ala 25 ; Asn 24 was replaced with Lys to introduce the coumarin fluorophore) of rat malate dehydrogenase (MDH) precursor, as shown in Fig. 1A. Yeast MPP was obtained from co-purification of the C-terminally hexahistidine-tagged ␣-MPP with the nontagged ␤-MPP by nickel-chelating column. The purified MPP was confirmed to be an equal stoichiometry of both subunits by SDS-polyacrylamide gel electrophoresis and gel filtration and had the same kinetic parameters, K m and k cat , as the native enzyme (data not shown). The coumarin-labeled peptide (DAC-MDH5-25) gave a fluorescence emission spectrum with the maximum at 482 nm ( Fig. 1B, spectrum 1). The addition of the purified MPP, which was inactivated in the presence of EDTA, led to a large increase in fluorescence intensity and a blue shift in emission spectrum (emission maximum at 470 nm) (Fig. 1B, spectrum 2). When the coumarin-labeled peptide was added to the MPP in the absence of EDTA, the increased fluorescence rapidly disappeared to the level seen with the peptide alone (Fig. 1B, spectrum 3). The processed peptides obtained after incubation of DAC-MDH5-25 with a small amount of the active enzyme gave no change in the fluorescence spectrum (data not shown). These observations are interpreted to mean that the increase in the fluorescence occurs through a the specific interaction of the precursor peptide with MPP and that this interaction is lost with processing of the peptide.
The titration of MPP with DAC-MDH5-25 showed a simple saturation curve (Fig. 2). The inset in the figure shows the Scatchard plot of the data, from which the dissociation constant, K d , was calculated to be 0.13 Ϯ 0.07 M, as the inverse of the slope. The molar ratio of the peptide to MPP was determined to be 0.83 Ϯ 0.09, suggesting a stoichiometric binding of the peptide to the enzyme.
Our earlier study using synthetic MDH peptides (17) indicated that the arginine residue at position Ϫ2 and those distant from the cleavage point are important for cleavage by the MPP. To confirm the specific binding of the fluorescence-labeled peptides to the MPP, we examined the interactions of MPP with DAC-MDH5-25 derivatives (Table I), in which arginine residues were replaced with alanine. DAC-MDH7A, in which the distal Arg 7 of the MDH extension peptide was altered to alanine, gave about a 15-fold increase in the K d value. DAC-MDH14A, in which the arginine residue at position Ϫ3 from the cleavage site was replaced by alanine, showed substantially the same K d value as that of DAC-MDH5-25. On the other hand, DAC-MDH15A, which lacks the arginine at position Ϫ2, gave about a 12-fold increase in the K d value. DAC-MDH14A15A, with a double substitution of alanines for arginine residues at positions Ϫ2 and Ϫ3, showed more than a 50-fold-less affinity than that of the wild type. NaCl is about 10 times lower than that at 100 mM. 2 For enzyme assays in the previous experiments, we used a large amount of the diluted enzyme solution eluted from columns with high salts, whereas contamination of salts was negligible in the present experiment because we used an extremely high concentration of the recombinant enzyme. Thus, a similar response to amino acid substitution between K d and K m indicates that the distal and proximal arginine residues directly participate in substrate binding of MPP. This method using a fluorescence-labeled peptide allows one to estimate the affinity of MPP for substrates.
Cooperative Formation of a Substrate Binding Pocket by Two Subunits-To determine which subunit is responsible for binding of the substrate, each subunit that was tagged with hexahistidines at the C terminus was individually purified using a nickel-chelating column. They reconstituted the enzyme activity when mixed together, although ␣-MPP required the addition of 30% glycerol to prevent aggregation. When the fluorescence change was measured at the peptide concentration of 1 M, the spectra with ␣and ␤-MPP alone were relatively small (Fig. 1B, spectrum 4 and spectrum 5, respectively) compared with the MPP complex (Fig. 1B, spectrum 1), thereby suggesting the low affinity of subunits for the substrate. The fluores-cence intensity at 470 nm increased with increasing concentrations of the peptide, although the intensity for both subunits did not reach the maximum level even at 5 M peptide (Fig. 3). The K d values for ␣and ␤-MPP were determined to be 4.38 Ϯ 0.75 and 4.45 Ϯ 0.60 M, respectively, from the Scatchard analysis of the data (Fig. 3, insets). Thus, the affinities of ␣and ␤-MPP for the substrate are about 30-fold lower than that of the MPP complex. The results indicate that only the MPP complex can bind substrates with high affinity. The molar ratio of the peptide to ␣and ␤-subunit were calculated to be 0.92 Ϯ 0.12 and 0.83 Ϯ 0.09, respectively, suggesting specific interaction of the peptide to each subunit, although with low affinity.
The interaction of each subunit with the peptides that lack the distal and proximal arginines was too weak to detect a change of the coumarin fluorescence in the titration experiment. Therefore, the affinities of the subunit monomers with these mutant peptides were estimated by competition between DAC-MDH5-25 and the nonlabeled mutant peptides in binding to each subunit. In the experiment, MDH7A14A was used as a substrate lacking distal arginine, instead of MDH7A, to eliminate the effect of the number of positive charges in the peptides, because the replacement of arginine residue at position 14 had little effect on binding of the peptide to the MPP (Table  I). Fig. 4A shows the titration of ␤-MPP with DAC-MDH5-25 in the presence of 0 to 75 M competitor peptide, nonlabeled MDH7A14A. The apparent K d values of the subunit for DAC-MDH5-25 were obtained from titration curves at various concentrations of the competitors. The K d value for the competitor peptide could be determined from a plot of the apparent K d values for the labeled peptide in the presence of the competitor versus the concentration of the competitor peptide, and the plot was reasonably linear (Fig. 4B). When nonlabeled wild-type peptide was used as the competitor, the calculated K d values for ␣-MPP (4.4 M) and ␤-MPP (3.6 M) were substantially the same as those obtained by the direct binding assay described above. The K d value of MDH7A14A for ␤-MPP was determined to be 43 M (Fig. 4B), which is about 10 times higher than that of the wild-type peptide. This mutant peptide, however, exhibited practically no inhibition to the binding of DAC-MDH5-25 to ␣-MPP even at 5 mM. On the other hand, the K d values of MDH14A15A for binding to ␣and ␤-MPP were determined to be 0.46 and 2.67 mM, respectively, indicating that the mutant peptide has more than 100 and 500 times less affinity to ␣and ␤-subunits, respectively, than MDH5-25. These findings mean that substrate binding with the individual subunits as well as with the whole enzyme (Table I) depends on the presence of the proximal and distal arginines, although dependence varies with the subunits and position of arginines. Thus, ␣-MPP appears to participate in the interaction with distal arginine residue of the extension peptide, whereas ␤-MPP seems to be more responsible for interaction with proximal arginine.
The difference in the binding state of peptides between the MPP complex and each subunit was revealed by the quenching of coumarin fluorescence by acrylamide. After binding of the peptide to the MPP complex and the subunits, increasing concentrations of acrylamide were added, and the fluorescence change was measured at 470 nm. Fig. 5 shows the Stern-Volmer quenching plots, and the quenching constants, Ksv, were calculated to be 1.1 Ϯ 0.4, 8.7 Ϯ 0.6, and 9.1 Ϯ 1.3 M Ϫ1 , for the MPP complex, ␣-, and ␤-subunit, respectively. The results indicate that the local environment around the substrate peptide bound to the MPP complex differs from that for each subunit and that in the MPP complex the coumarin fluorophore is partially buried in the enzyme, whereas the one in the subunit is fully exposed to the solvent. DISCUSSION We presented evidence that MPP can bind the substrate peptides with high affinity but only in the dimeric complex. MPP has an affinity of the dissociation constant of about 0.13 M to the peptide corresponding to the N-terminal 21 amino acids of rat MDH precursor, whereas each subunit has about a 30-fold less affinity than the dimeric complex. The study using surface plasmon resonance, however, showed that ␣-MPP bound a substrate with the same affinity as the MPP complex and that the K d value to the immobilized peptide was determined to be 0.2 M (24). We obtained similar results showing no clear difference in the affinity between the MPP complex and each subunit, based on the experiments on surface plasmon resonance and on affinity purification of the enzyme using the peptide ligand bound to the Sepharose resin. 2 Because our fluorescence-quenching analysis demonstrated that the peptide bound to the MPP was buried in the enzyme, the finding that the K d value for the MPP complex is similar to that for individual subunits in these experiments was probably because of steric interference by immobilization of the peptides to the dextran matrix on the sensor chip or Sepharose resin. These findings indicate that dimer formation of the two subunits of MPP is essential for substrate binding.
We also demonstrated that the individual subunit required arginines at different positions in the peptide for binding, although their affinities were much lower than that of MPP. Because ␤-MPP has a catalytic center (12), the subunit appears to interact with amino acid residues around a scissile bond of the precursor, including proximal arginine at position Ϫ2. The interaction between ␤-MPP and the proximal arginine seems to  Table I. The dissociation constant of ␣-MPP to Ac-MDH7A14A gives a minimum value, because no competition was observed even at 5 mM mutant peptide. The apparent DAC-MDH5-25 dissociation constants were determined from Scatchard analysis of data at each concentration of Ac-MDH7A14A, as described under "Experimental Procedures." a.u., arbitrary units. B, apparent DAC-MDH5-25 dissociation constants obtained as in A were plotted as a function of concentration of Ac-MDH7A14A. This plot yields a linear function, and an abscissa intercept is the dissociation constant for Ac-MDH7A14A. be more responsible for the initial binding step of substrates to the enzyme rather than the catalytic reaction, because mutation of the proximal arginine caused a drastic increase in the K m value but not much effect in the V max value (17). ␤-MPP may also interact with the upper region, distal to the cleavage site, of the extension peptide. The lack of the distal arginine of the MDH extension peptide led to an increase in the K d value to this subunit (Table II). Moreover, mutation of Glu 79 in rat ␤-MPP, which is a residue within the conserved acidic amino acid cluster in ␤-MPP, decreased the interaction of substrate peptide fluorescent-labeled at the N-terminal portion with this subunit (28). On the other hand, it is likely that ␣-MPP is responsible for the interaction with the upper region of the extension peptide. The present data showed that distal arginine was indispensable for binding of the peptide to ␣-MPP. We recently found that substitution of the conserved acidic amino acid residues in ␣-MPP (Glu 353 and Glu 377 /Asp 378 in yeast ␣-MPP) on the processing of the precursor protein was more effective with a longer extension peptide (25). Thus, individual subunits interact with different parts of the extension peptide of the precursor protein.
The subunits of the MPP are homologous to the core proteins of mitochondrial ubiquinol-cytochrome oxidoreductase (bc 1 complex), a component of the respiratory chain. The crystal structure of the bc 1 complex from bovine heart mitochondria has been determined (26,27), and one can expect that MPP is similar in structure to the bovine core 1 and core 2 complex. Core 1 and 2 proteins are structurally similar and consist of two domains of roughly equal size with almost identical folding topology. The two bowls representing these proteins were described as coming together in the form of a ball with a crack leading to the internal cavity. From the available data, the middle portion of core 2, which corresponds to the glycine-rich region of ␣-MPP conserved among many species, is close to a site of core 1 that corresponds to the zinc binding site of ␤-MPP. These regions might cooperatively form the active site that recognizes the proximal arginine and catalyzes cleavage reaction.
In conclusion, the two subunits of MPP cooperatively form the substrate binding pocket and recognize different structural elements in the extension peptide, including the proximal and distal arginine residues. Individual subunit monomers can interact with distinct structural elements, although with low affinity, and complex formation with the two subunits leads the enzyme to have higher affinity for substrates, the results both of induced structural change at the interface between the two and increase in binding sites. Recognition of the precursors at multiple subsites in MPP make it possible to render strict specificity and high affinity to the enzyme for precursors with structures little in common.