Mutational Analysis of Both Subunits from Rat Mitochondrial Processing Peptidase

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS

Vol. 335, No. 1, November 1, pp. 211–218, 1996 Article No. 0500

Mutational Analysis of Both Subunits from Rat Mitochondrial Processing Peptidase1 Hans-Martin Striebel,* Petr Rysavy,*,2 Jiri Adamec,* Jaroslav Spizek,† and Frantisek Kalousek*,3 *Department of Genetics, Yale School of Medicine, New Haven, Connecticut 06520-8005; and †Institute of Microbiology, Czech Academy of Sciences, 14220 Prague, Czech Republic

Received July 19, 1996

Rat liver mitochondrial processing peptidase (MPP) is the primary peptidase that cleaves leader peptides from nuclearly encoded mitochondrial proteins following their transport from the cytosol to the mitochondrial matrix. This enzyme consists of two nonidentical subunits that have overall similarity to each other and share certain amino acid motifs. These include the putative metal-ion binding HFLEH motif in the b-subunit and the HFLEK motif of the a-subunit, as well as a possibly helical amino acid stretch bearing a high concentration of negatively charged residues about 70 amino acids downstream of these motifs in both subunits. In order to achieve a better understanding of the role of certain amino acids in rat MPP, we performed site-directed mutagenesis on both of its subunits. Our results show that whereas both histidines and the glutamate of the HFLEH motif in the bsubunit are crucial for MPP function, this holds true only for the glutamate in the related HFLEK motif in the a-subunit. In addition, functionally important negatively charged residues in the region 70 amino acids downstream occur only in the b-subunit and not in the a-subunit. This indicates a functional asymmetry between the subunits, with the b-subunit containing a majority of residues participating in the active center. q 1996 Academic Press, Inc.

Key Words: mutagenesis; mitochondrial processing peptidase; active site.

Removal of leader peptides after transport of precursor proteins into mitochondria is an important and re1

This work was supported by National Institute of Health Grant DK09527 and Fogarty International Research Collaboration Award TW0530. 2 Dr. Petr Rysavy died during preparation of this manuscript. 3 To whom correspondence and reprint requests should be addressed. Fax: (203) 785-7227.

quired step prior to folding, sorting, and assembling mature polypeptide chains into functional proteins, leading to questions about the details of this process. More insight demands better knowledge about how primary sequences and structural features of precursor proteins influence specific recognition and cleavage, and vice versa, what the structural and functional properties of the appropriate peptidases are that are involved in precursor cleavage. In recent years the mitochondrial processing peptidase (MPP) has been studied in several organisms. MPP proteolytically removes leader sequences of precursor proteins once they have crossed the inner mitochondrial membrane. MPP has been purified to homogenity from Neurospora crassa (1), Saccharomyces cerevisiae (2), rat liver (3, 4), and potato tuber (5). Except in higher plants, MPP consists of two nonidentical subunits, a and b, with a molecular size between 52 and 57 kDa for the a-subunits and between 50 and 52 kDa for the b-subunits. Both subunits are necessary for the processing of precursor proteins. MPP activity is stimulated by divalent cations such as Mn2/, Zn2/, and Co2/ (6, 7), although the preferred cation in vivo is not known (8). MPPs of different species share common amino acid motifs, particularly the sequence motif HXXEH, which occurs in all known MPP b-subunits (9) and characterizes them as members of the proposed insulinase superfamily (10–13). An amino acid motif related to this HXXEH motif, with the sequence HXXEK or HXXD/ER is present in all MPP a-subunits. Additionally, regions around these motifs are highly homologuous among different species. These homologies extend further to an amino acid stretch bearing a high concentration of charged amino acids 50–70 residues downstream of these motifs in both subunits. Three-dimensional protein structures are not yet available for MPP or for other members of the insulinase superfamily. There are, however, protein struc211

0003-9861/96 $18.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.

AID

ARCH 9690

/

6b23$$$181

09-30-96 14:55:07

arca

AP: Archives

212

STRIEBEL ET AL.

ture data for thermolysin, which may have some relationship to the insulinase family (14). Thermolysin contains the conserved zinc-binding motif HEXXH, whereas in MPPs and other members of the insulinase superfamily the inverted motif HXXEH is found. Xray data for thermolysin assign both histidines in the HEXXH motif to function as ligands for the zinc ion in the catalytic center of the enzyme. These data further suggest that the glutamate in the HEXXH motif has the role of an electron donor for a hydrogen bond to the water molecule which attacks the carbonyl carbon at the peptide bond to be cleaved (15). Altogether, the HEXXH motif is embedded in an amino acid region with an a-helical structure (14). Functional studies with E. coli proteinase III, another member of the insulinase superfamily, have identified the three amino acid residues responsible for binding the catalytic zinc ion within this enzyme (16). As with the motif in thermolysin, both histidines on the HXXEH motif in Escherichia coli proteinase III were shown to be zinc binding residues (10). Additionally, a glutamate, 77 amino acids downstream of the HXXEH motif, was shown to be the third zinc binding residue (16). This leads to the motif HXXEH(X)76E as the sequence for the zinc binding site in E. coli protease III. Minor changes in the spacing suggest the metal binding glutamate in other members of the insulinase superfamily. In rat MPP subunits, this sequence leads to HXXEH105(X)75E181 in the b-subunit and to HXXEK113(X)76E190 in the a-subunit. The spacing of about 76 amino acids between the putative 2nd and 3rd metal binding residues seems to be of importance for enzyme function, as this distance is conserved during evolution in MPPs from Saccharomyces cerevisiae to plant and mammalian MPPs. The spacing between amino acids also seems to be important in the area around the putative 3rd metal binding site, where the large number of negatively charged amino acids could be arranged in a negatively charged, amphiphilic ahelix. Application of an a-helical arrangement to residues 177–194 of the a-subunit of rat MPP (17) shows an amphiphilic helix with extended regions of negatively charged residues. There has been interest in such areas with negative charge because leader sequences of precursor proteins are also thought to form amphiphilic a-helices, although with extended regions of positively charged residues (18). An electrostatic interaction between negatively and positively charged helical faces is considered as one possible mechanism for specific recognition of leader peptides by MPP. Experiments described in this paper were carried out on rat liver MPP (17, 19) and had the goal to link functional properties of rat MPP to a structural base. Because of the lack of three-dimensional protein structure data for members of the insulinase super-

AID

ARCH 9690

/

6b23$$$182

09-30-96 14:55:07

family, a way to accomplish this was by determining crucial amino acid residues in rat MPP. We therefore employed site-directed mutagenesis to alter amino acid residues around the three putative metal binding sites in rat b-MPP and the related amino acid stretches in rat a-MPP. EXPERIMENTAL PROCEDURES Materials. Bacterial strains used were: E. coli DH5aF*Iq, (GibcoBRL, Gaithersburg, MD); E. coli BL21(DE3), (20); E. coli CJ236, Bio-Rad, (Hercules, CA). Plasmids used were pET11d, New England Biolabs (Beverly, MA); pGEMEX-1 and helperphage M13KO7, Promega (Madison, WI); pGroESL (21). Enzymes used were alkaline phosphatase (CIP), polynucleotide kinase, T4 DNA polymerase, T4 DNA ligase, and restriction enzymes BamHI, KpnI, SauI, Van91I from Boehringer Mannheim (Indianapolis, IN); restriction enzymes AflII, BglII, NheI, MluI, and XbaI from New England Biolabs; AmpliTaq from Perkin–Elmer (Norwalk, CT); Sequenase sequencing reagents from USB (Cleveland, OH). DEAE Bio-Gel Agarose (100–200 mesh) and Bio-Gel P-200 were from Bio-Rad; hydroxyapatite was prepared according to Main et al. (22). pET11d and pGEMEX-1 subcloning of rat MPP a- and b-subunits. For separate production of the a- and b-MPP subunits, coding sequences for the mature proteins were amplified by PCR from their respective cDNAs (primers, see 23) and cloned into vector pET11d: a-MPP was subcloned into the filled-in BamHI site, leading to pETa, whereas b-MPP was subcloned into the filled-in NcoI site, leading to pETb. For combined production of the rat MPP subunits, both coding sequences for the mature proteins were subcloned in a tandem manner into pET11d, each downstream of its own T7 promoter. To achieve this, a T7 promoter / a-MPP cassette was released from pETa by way of the XbaI and NheI sites and cloned into the XbaI site of pETb to produce pETab. To produce single-stranded DNA of the rat MPP subunits for sitedirected mutagenesis, the a- and b-MPP subunits from pETab were transferred to vector pGEMEX-1, which carries a phage f1 origin, by excising the T7 promoter / a-MPP / T7 promoter / b-MPP cassette from pETab by BglII and BamHI and cloning it in pGEMEX-1 at the same restriction sites. The resulting plasmid construct was designated pGEMabinv and used for all MPP expression and mutagenesis procedures. Site-directed mutagenesis. Single-stranded DNA of the plasmid pGEMabinv was prepared by transformation of E. coli CJ236, a strain bearing the dut0 and ung0-markers which lead to the occasional substitution of uracil for thymine in its DNA. Production of single-stranded, uracilated pGEMabinv DNA was initiated by inoculation of a culture of E. coli CJ236 bearing pGEMabinv in TYP-medium (1.6% tryptone/1.6% yeast extract/0.5% NaCl/0.25% K2HPO4) with helper phage M13KO7 and incubating overnight at 377C. The single-stranded DNA was harvested from the culture medium and used as template for the synthesis of the mutagenized strand (see below). Because of the instability of uracilated DNA in the dut/, ung/ E. coli strain DH5aF*Iq, the newly synthesized, mutagenized, and nonuracilated DNA strand is selectively propagated after transformation into this strain. To generate site-directed mutants of the rat MPP b-subunit, we used T4 DNA polymerase to extend a mutation-programming phosphorylated oligonucleotide primer along the complementary, singlestranded plasmid pGEMabinv, containing both MPP a- and bsubunits. After ligation, competent E. coli DH5aF*Iq cells were transformed with the resulting double-stranded plasmids and yielded between 10 and 50 colonies per 0.2 mg single strand template, 80% of which contained the mutant oligonucleotide. For a-subunit mutations, a PCR-based strategy was employed, taking advantage of unique restriction sites in the DNA regions where mutations were

arca

AP: Archives

MUTATIONAL ANALYSIS OF RAT MITOCHONDRIAL PROCESSING PEPTIDASE planned. Mutation primers were designed to carry the intended mutation as well as, 5* of it, the unique restriction site. DNA amplification was then performed between the mutant primer and a second primer at another restriction site several hundred nucleotides away. The amplified DNA product, carrying the desired mutation, was exchanged with the wild-type fragment on the plasmid by way of the unique restriction sites. The presence of all mutations was confirmed by DNA sequencing. For activity assays of the produced mutant enzymes, the constructs had to be transferred from DH5aF*Iq, which was a good cloning host but contained proteases interfering with the MPP assay, to E. coli strain BL21, a poor cloning host but lacking interfering proteases. Because of the instability of plasmids bearing MPP mutants during subsequent cultivation, all E. coli BL21 cultures with plasmid constructs for mutant MPP production were generated each time by transforming competent E. coli BL21 cells with plasmids from a stock previously confirmed to be correct. Cell cultivation and harvest. LB medium (25 ml) (1% tryptone/ 0.5% yeast extract/0.5% NaCl) containing 50 mg/ml ampicillin was inoculated with an overnight culture of E. coli BL21 cells bearing pGEMabinv plasmids. Cells were grown to OD600 Å 0.6, 100 ml 100 mM IPTG4 solution added, and cultures were then shaken for 1 h at 377C. After chilling the cultures on ice, cells were recovered by centrifugation, washed in 10 ml HD buffer (10 mM Hepes, pH 7.4/1 mM DTT), resuspended in 1 ml HD, and PMSF was added to a final concentration of 0.2 mM. Cells were disrupted by sonicating the cell suspension three times 10 s on ice. Cell debris was sedimented with 5 min centrifugation at 47C (15,000g); the supernatant was analyzed for MPP activity after adjusting total protein to a final concentration of about 4 mg/ml with HD buffer. Protein concentration was measured according to Layne (27). MPP concentration was estimated by Western blot analysis. Activity assay. In a typical assay to measure MPP activity, we used as substrate the precursor of the b-subunit of F1 ATPase from Saccharomyces cerevisiae, which was synthesized by in vitro translation in a rabbit reticulocyte lysate containing [35S]methionine (28). The radiolabeled product of the translation reaction was incubated for 1 h at 277C together with crude extract or purified MPP. For better estimation of MPP activity, we made three different dilutions of crude extracts. All other conditions were as described previously (4). The reaction product was then analyzed directly on SDS/PAGE (23). Reconstitution assays. Reconstitution assays were performed by mixing a crude bacterial extract containing recombinant MPP subunits carrying a mutation, with equal amounts of a second bacterial crude extract containing the recombinant wild-type version of the mutated subunit. Purification of MPP mutants. Partial purification of wild-type and mutant MPP from E. coli crude extract was carried out in two steps, using DEAE Bio-Gel and hydroxyapatite, according to a protocol described previously (4).

RESULTS

Production of Mutagenized MPP Active rat MPP can be generated by producing its aand b-subunits separately in different E. coli cultures, preparing crude extracts from each and mixing them 4 Abbreviations used: DTT, dithiothreitol; HD, HEPES/DTT buffer; IPTG, isopropyl-b-D-thiogalactopyranoside; PAGE, polyacrylamide gel electrophoresis; PMSF, phenylmethanesulfonyl fluoride; SDS, sodium dodecylsulfate.

AID

ARCH 9690

/

6b23$$$182

09-30-96 14:55:07

213

FIG. 1. Activity assay of rat liver MPP produced in E. coli. The E. coli extracts used contained MPP a- and b-subunits either separately or as coproduced proteins. For combination of separately produced subunits to MPP, the subunits were mixed together and incubated on ice for 30 min. Then they were incubated with [35S]methioninelabeled substrate (pF1b) for 60 min at 277C. Products were analyzed directly by SDS/PAGE and fluorography. Lane 1, coproduced MPP a- and b-subunits; lane 2, MPP a-subunit; lane 3, MPP b-subunit; lane 4, mixed, separately produced MPP a- and b-subunits; lane 5, translation product only.

(Fig. 1). It is therefore possible to clone and express the coding sequences for rat MPP subunits without concerted genetic control to generate functional MPP. For site-directed mutagenesis, we combined both coding sequences into one expression unit by constructing pGEMabinv, where both mature subunits were arranged in tandem, each downstream of its own T7 promoter. There was no significant difference in the amounts of MPP subunits produced from E. coli cells with plasmids encoding wild-type MPP subunits in comparison with E. coli cells with plasmids encoding mutagenized MPP subunits as confirmed by Western blot analysis (data not shown). Conserved Regions Our first experiments concerned structure and function of the HXXEH motif in the MPP b-subunit and its HXXEK counterpart in the a-subunit, as these motifs are conserved in all known MPPs. Based on data from thermolysin and E. coli protease III, residues of particular interest were H101, E104, and H105 in b-MPP, as well as H109, E112, and K113 in a-MPP. Conserved amino acid residues directly adjacent to the HXXEH motif were also investigated. Mutations of these residues might be predicted to cause minor impairment of MPP function in comparison with mutations in the HXXEH motif itself. Our results show that nonconservative exchanges in the amino acid stretch from H101 to H105 in b-MPP lead to proteins with no or barely detectable MPP activity (Table I). In particular, conversion of either of the putative metal-binding residues, H101 and H105, to arginine, whose positive charge makes it unlikely to be a metal binding residue, abolished MPP activity (Fig. 2, lane 1) or reduced it to about 5% of the original wild-type MPP activity (Fig. 2, lane 2). Likewise, only about 5% of wild-type MPP activity was retained if H101 was mutated to Q (Fig. 2, lane 3). Converting both H101 and H105 to glutamine residues abolished MPP activity fully. The importance of negative charge at E104 is also clear, as conversion of this residue to its neutral counterpart,

arca

AP: Archives

214

STRIEBEL ET AL. TABLE I Mutations of Amino Acids in and around the HFLEH Motif of the Rat MPP b-Subunit

Note. Activity levels: ////, 50% precursor conversion (as in wild-type MPP); ///, 25– 50% precursor conversion; //, 10–25% precursor conversion; /, 5–10% precursor conversion; 0, no mature substrate produced.

glutamine, abolishes MPP activity (Fig. 2, lane 6), whereas retaining the charge by changing it to aspartate does not impair MPP activity (Fig. 2, lane 5). As the b-MPP HXXEH motif is the inverted form of the thermolysin HEXXH motif, a further question is whether the orientation is important. Inverting b-MPP HFLEH into HELFH would leave the two histidines H101 and H105 in place, but would alter the position of

E104. Figure 2, lane 7, shows that MPP activity was abolished completely by this alteration, indicating that the position occupied by the glutamate between the two histidines is critical. In contrast to the b-MPP motif, the related HFLEK amino acid stretch in a-MPP between H109 and K113 proved to be differently sensitive to amino acid substitution (Table II and Fig. 3). On the one hand, substitu-

FIG. 2. Activity of rat MPP with selected mutations at the HFLEH motif of the b-subunit. The given volumes of E. coli crude extracts containing coproduced wild-type MPP a-subunits and the indicated mutant MPP b-subunits were incubated with pF1b substrate according to conditions described in the legend of Fig. 1; the E. coli crude extract in lane 7 contains MPP with an H101FLEH105 to H101ELFH105 inversion in the b-subunit; WT, wild-type MPP; TR, translation product only.

FIG. 3. Activity of rat MPP containing selected mutations at the HFLEK motif of the a-subunit. E. coli crude extracts containing coproduced wild-type MPP b-subunits and the indicated mutant MPP a-subunits were incubated with pF1b substrate according to conditions described in the legend of Fig. 1. The E. coli crude extract in lane 7 contains MPP with an exchange of the HFLEH motif in the MPP b-subunit for the HFLEK motif in the MPP a-subunit, as well as an exchange of the HFLEK motif in the MPP a-subunit for the HFLEH motif of the MPP b-subunit; WT, wild-type MPP; BL21(DE3), E. coli crude extract without plasmid.

AID

ARCH 9690

/

6b23$$$183

09-30-96 14:55:07

arca

AP: Archives

MUTATIONAL ANALYSIS OF RAT MITOCHONDRIAL PROCESSING PEPTIDASE TABLE II Mutations of Amino Acids in the HFLEK Motif of the Rat MPP a-Subunit

215

lengths at position 99, 100, 106, 108, 109, and 110 (which are all rather conserved across species) impaired MPP activity but did not lead to an inactive enzyme. Therefore, MPP activity losses by converting A100 to L and M106 to A or T seem likely not to be due to direct interference of these residues with the active center of the enzyme. The same holds true for changes somewhat more distant from the HXXEH motif, for example converting F108 to Y, K109 to N or E, and G110 to L. These residues seem instead to have structural or supportive functions within the enzyme. In order to identify a possible third metal binding residue important for MPP function, we focussed on residues between positions 168 and 184 in b-MPP and 175 to 197 in a-MPP. Both regions have been previously postulated to be important for enzyme function, based on their conservation and potential to form a negatively charged a-helix as a possible binding site for positively charged, helical leader peptides (17, 19). Based on the motif spacing in E. coli proteinase III, HXXEH(X)76E, E181 is a likely candidate in the b subunit. Other acidic residues on the same face of the putative helix are E170 and E174. For the a-subunit, a 76 amino acid spacing identifies E190, with E179 and D193 on the same helical face. These residues were changed to their neutral counterparts, and the resultant enzymes were tested for activity. Whereas most substitu-

Note. Activities as described for Table I.

tion of H109 with residues such as K, R, and F led to strongly reduced MPP activity or to its abolition (Fig. 3, lanes 1, 3, and 4). On the other hand, substituting the hydrophobic residue I for K113 only marginally reduced MPP activity (Table II and Fig. 3, lane 2). Attempts to increase the metal binding capability at position 113 by converting K113 to H did not improve, but rather weakened MPP activity, suggesting that conventional metal binding parameters do not apply to the HFLEK motif. A different response to mutations also occurred at E112, where reducing the side chain length by exchange for an aspartate reduced MPP activity markedly, whereas conversion to its neutral counterpart Q even improved MPP performance beyond wild-type level (Fig. 3, lanes 5 and 6). This functional asymmetry between MPP subunits was confirmed by an exchange of the HFLEH motif in b-MPP for the HFLEK motif in a-MPP. The doubly substituted enzyme was completely nonfunctional (Fig. 3, lane 7). With b-MPP carrying a major share of residues involved in enzyme function, it was of interest to find out more about the role of side chains around its HFLEH motif. As seen in Table I, variations of side chain

AID

ARCH 9690

/

6b23$$$183

09-30-96 14:55:07

TABLE III Mutations of Amino Acids in the Region 50–80 Residues Downstream of the HFLEH Motif in the Rat MPP b-Subunit

Note. Activities as described for Table I.

arca

AP: Archives

216

STRIEBEL ET AL. TABLE IV Mutations of Amino Acids in the Region 65–80 Residues Downstream of the HFLEK Motif in the MPP a-Subunit

Note. Activities as described for Table I.

tions had no or only minor effects on MPP activity, the conversions E174 to Q and E181 to Q in the b-subunit each led to virtually complete loss of MPP function. Even the conservative changes E174 to D and E181 to D produced strongly reduced activities, showing that even minor dislocations of the negative charge impaired MPP activity (Table III). We conducted additional experiments to address the importance of the spacing between negative charged residues in these positions by deleting V177, inserting additional residues between V177 and I178 (I177a and A177a), or turning V177 into a helix breaking residue (V177 to P). In all instances we observed drastically reduced or abolished MPP activity (Table III). When we used the same approach in the corresponding region (residues 177–193) of a-MPP, we observed a remarkably different picture. Of all the changes in negatively charged residues in this stretch, only E179 to Q had a profound influence on MPP activity (Table IV), suggesting that the role of this region may be different from that previously postulated. Reconstitution and Purification of Mutant MPP To confirm that activity losses observed with MPP mutations are only due to the introduced mutation it-

AID

ARCH 9690

/

6b23$$$183

09-30-96 14:55:07

self and not to other factors like inactivation or alteration of the complementary subunit, we performed in vitro reconstitution experiments by adding wild-type versions of the mutated subunits to the activity assays. Unpublished experiments had suggested that subunit exchange could occur in vitro. As shown in Fig. 4 for selected b-MPP mutations, MPP activity of the mutant enzyme could be restored by adding exogenous wildtype b-subunit, indicating that the activity losses observed were in fact due to the presence of the mutated subunits. A further way to exclude changes in the overall tertiary and quarternary protein structure due to the sin-

FIG. 4. Restoration of MPP activity in b-subunit mutants by adding wild-type b-subunit. E. coli crude extracts containing coproduced wild-type MPP a-subunits and the indicated MPP b-subunit mutants were incubated either in the absence or presence of E. coli crude extracts containing wild-type MPP b-subunits under the conditions described in the legend of Fig. 1; 0, basic activity of mutant MPP; /, activity after adding wild-type MPP b-subunit.

arca

AP: Archives

MUTATIONAL ANALYSIS OF RAT MITOCHONDRIAL PROCESSING PEPTIDASE

FIG. 5. Mutant MPP is a heterodimer. Partially purified wild-type MPP and the MPP mutant containing a b-subunit H101 to Q change were fractionated on BioGel P-200 using ornithine transcarbamylase (OTC, 109 kDa) as a standard. Individual fractions were analyzed for MPP activity under the conditions described in the legend of Fig. 1. About 20 times more mutant MPP than wild-type MPP was used for activity assays. Estimations of MPP amounts were confirmed by Western blot analysis (not shown). OTC was assayed as previously described (26). WT, wild-type MPP; H101 to Q, MPP containing the H101 to Q substitution in the b-subunit. The center of the elution peak at a protein size of about 109 kDa shows that MPP purifies as a dimer (calculated dimeric MPP molecular weight is 110 kDa).

gle amino acid changes was by studying the behavior of a mutant MPP during purification. We partially purified MPP containing the H101 to Q mutation in the bsubunit, which had about 5% residual activity. The produced protein fractionated on DEAE Bio-Gel and hydroxyapatite columns similar to wild-type (not shown) and chromatographed as a dimer of 100 to 110 kDa on Bio-Gel P-200 gel filtration, behavior identical to that of wild-type (Fig. 5). DISCUSSION

Determination of crucial amino acids or amino acid groups is a basic step for gaining insight into structural and functional properties of a protein. Guidelines to important residues include homologies between sequences of related proteins, which indicate residues conserved during evolution, as well as the occurrence of amino acid motifs found either conserved or with some variation in proteins with similar function. The first residues emerging as candidates for being part of the active center of rat liver MPP were H101, E104, and H105 of the HFLEH motif in the b-subunit. In combination with data from thermolysin (15) and E. coli protease III (16), two different functions for this motif may be postulated: H101 and H105, as two residues of the internal metal binding site (24, 25); and E104 as part of the chain of residues that provides an electron pair for nucleophilic attack of a water molecule on the carbonyl group in the cleavable peptide bond of the substrate protein. In contrast, the three corresponding residues of the related HFLEK motif in a-MPP, H109, E112, and K113 seem to have a different function. Whereas H109 was shown to be necessary for enzyme activity, which may include metal binding, this does not hold true for K113,

AID

ARCH 9690

/

6b23$$$183

09-30-96 14:55:07

217

making the function of the HFLEK-motif in a conventional metal-dependent proteolytic step unlikely. E112, on the other hand, is a crucial residue for rat MPP activity, although we have no data about its specific role within MPP. The third putative metal binding residue in rat MPP b-subunit could be narrowed to either E174 or E181 within a conserved, negatively charged, potentially ahelical region. In addition to providing the third metal ion anchor, this area, with its high concentration of negatively charged residues, could serve as a docking site for the positively charged leader peptide of a precursor protein prior to cleavage. The susceptibility of MPP activity to changes of charged amino acids around the putative third metal binding site in the b-subunit, in contrast to its insensibility to corresponding alterations at homologous residues in the a-subunit, suggests that the main charge-based recognition of leader peptides may be carried out by the b-subunit. Our results therefore indicate an asymmetric distribution of functions between the two MPP subunits. Although the b-subunit seems to bear the main share of residues involved in enzymatic activity, on its own the b-subunit is not able to perform even rudimentary enzyme functions. Thus, the a-subunit must provide critical interactions in support of MPP’s catalytic function. It is likely that the MPP a- and b-subunits have a common evolutionary origin given the high homology between both subunits, as well as the occurence of related amino acid motifs. A dimer of originally identical subunits could have evolved into a heterodimer of similar subunits, each charged with complementary, specific tasks in the recognition, binding, and cleavage functions of MPP. In contrast to Neurospora MPP, where a- and b-subunits can be separated during protein purification (1), the subunits of rat liver MPP show enough affinity to each other to be copurified as a complex. It was therefore somewhat surprising that we were able to substitute wild-type rat MPP subunits for their mutant counterparts in vitro, with recovery of more than 50% of wild-type MPP activity, suggesting competition between the mutant and wild-type subunits for the complementary subunit. These experiments, together with the production and purification of a selected MPP mutant, confirm that the overall three-dimensional structure of the mutant proteins is retained. They support our hypothesis that the differences in MPP activity are actually due to the specific amino acids changes and not a consequence of large structural rearrangements within the MPP molecule or disruptions of its quaternary structure. ACKNOWLEDGMENTS We are very grateful to Dr. Wayne A. Fenton for helpful discussions and carefully reading the manuscript.

arca

AP: Archives

218

STRIEBEL ET AL.

Note added in proof. During preparation of this manuscript we became aware of the findings of Kitada et al. (25), which in some aspects reflected our presentation at the annual meeting of the American Society for Cell Biology (24). In contrast to Kitada et al., however, our results show retained MPP activity of about 5% of the wildtype activity after mutating H101 to R in the MPP b-subunit HXXEH motif.

REFERENCES 1. Hawlitschek, G., Schneider, H., Schmidt, B., et al. (1988) Cell 53, 795–806. 2. Yang, M., Jensen, R. E., Yaffe, M. P., Oppliger, W., and Schatz, G. (1988) EMBO J. 7, 3857–3862. 3. Ou, W.-J., Ito, A., Okazaki, H., and Omura, T. (1989) EMBO J. 8, 2605–2612. 4. Isaya, G., Kalousek, F., Fenton, W. A., and Rosenberg, L. E. (1991) J. Cell Biol. 113, 65–76. 5. Braun, H.-P., Emmermann, M., Kruft, V., and Schmitz, U. K. (1992) EMBO J. 11, 3219–3227. 6. McAda, P. C., and Douglas, M. G. (1982) J. Biol. Chem. 257, 3177–3182. 7. Schneider, H., Arretz, M., Wachter, E., and Neupert, W. (1990) J. Biol. Chem. 265, 9881–9887. 8. Yang, M., Geli, V., Oppliger, W., Suda, K., James, P., and Schatz, G. (1991) J. Biol. Chem. 266, 6416–6423. 9. Braun, H. P., and Schmitz, U. K. (1995) Trends Biochem. Sci. 20, 171–175. 10. Becker, A. B., and Roth, R. A. (1992) Proc. Natl. Acad. Sci. USA 89, 3835–3839. 11. Rawlings, N. D., and Barrett, A. J. (1991) Biochem. J. 275, 389– 391.

AID

ARCH 9690

/

6b23$$$184

09-30-96 14:55:07

12. Schulte, U., Arretz, M., Schneier, H., Tropschug, M., Wachter, E., Neupert, W., and Weiss, H. (1989) Nature 339, 147–149. 13. Gencic, S., Schagger, H., and von Jagow, G. (1991) Eur. J. Biochem. 199, 123–131. 14. Holmes, M. A., and Matthews, B. W. (1982) J. Mol. Biol. 160, 623–639. 15. Holmes, M. A., and Matthews, B. W. (1981) Biochemistry 20, 6912–6920. 16. Becker, A. B., and Roth, R. A. (1993) Biochem. J. 292, 137–142. 17. Kleiber, J., Kalousek, F., Swaroop, M., and Rosenberg, L. E. (1990) Proc. Natl. Acad. Sci. USA 87, 7978–7982. 18. Hendrick, J. P., Hodges, P. E., and Rosenberg, L. E. (1989) Proc. Natl. Acad. Sci. USA 86, 4056–4060. 19. Paces, V., Rosenberg, L. E., Fenton, W. A., and Kalousek, F. (1993) Proc. Natl. Acad. Sci. USA 90, 5355–5358. 20. Studier, F. W., Rosenberg, A. H., Dunn, J. J., and Dubendorff, J. W. (1990) Methods Enzymol. 185, 60–89. 21. Goloubinoff, P., Gatenby, A. A., and Lorimer, G. H. (1989) Nature 337, 44–47. 22. Main, R. K., Wilkins, M. J., and Cole, L. J. (1959) J. Am. Chem. Soc. 81, 6490–6495. 23. Saavedra-Alanis, V. M., Rysavy, P., Rosenberg, L. E., and Kalousek, F. (1994) J. Biol. Chem. 269, 9284–9288. 24. Kalousek, F., Rysavy, P., and Striebel, H.-M. (1994) Mol. Biol. Cell 5, 475a. 25. Kitada, S., Shimokatu, K., Niidome, T., Ogishima, T., and Ito, A. (1995) J. Biochem. 117, 1148–1150. 26. Kalousek, F., Orsulak, M. D., and Rosenberg, L. E. (1984) J. Biol. Chem. 259, 5392–5395. 27. Layne, E. (1957) Methods Enzymol. 3, 447–454. 28. Titus, D. E., Ed. (1991) Protocols and Applications Guide, 2nd ed., Promega Corp., Madison, WI.

arca

AP: Archives