Biosynthesis and mitochondrial processing of the beta subunit of propionyl coenzyme A carboxylase from rat liver.

Propionyl-CoA carboxylase (ADP-forming) (EC 6.4.1.3), an oligomer of nonidentical subunits (alpha 4 beta 4), has been localized to the mitochondrial matrix. As a first step in examining this enzyme's biogenesis, we have investigated in vitro the cell-free, rat liver RNA-directed synthesis of the beta subunit, and its post-translational transport and processing by rat liver mitochondria. The beta subunit is synthesized as a precursor approximately 7,500 daltons larger than its mature mitochondrial counterpart. The extension segment, comprising approximately 60 amino acids, is located at the NH2 terminus of the precursor. Intact mitochondria translocate the precursor across both mitochondrial membranes, and a protease localized to the mitochondrial matrix cleaves the precursor to a polypeptide identical in size and peptide composition to the mature beta subunit.

The subunits of more' than two dozen nuclear-coded mitochondrial enzymes have been shown recently to be synthesized on cytoplasmic polyribosomes as precursor polypeptides -2,000 to -10,000 daltons larger than their respective mature counterparts (see review in Ref. 1,2). Post-translational uptake and processing of these precursors by isolated mitochondria has been demonstrated in some (1) but not all (2) cases.
The biogenesis of one such mammalian mitochondrial matrix enzyme, ornithine carbamoyltransferase (EC 2.1.3.3), has been investigated in detail in our laboratory (3)(4)(5)(6)(7) and in that of another group (reviewed in Ref. 8). Mammalian OTCase' is a trimer of identical subunits. Its precursor is -4,000 daltons larger than the corresponding mature subunit due to an NH2-terminal extension sequence. This extension sequence is removed in two proteolytic steps during the complex, energy-dependent reaction sequence which results ultimately in translocation of the precursor across both mitochondrial membranes to the matrix, and in the assembly of the mature subunits to a functional trimer.
* This work was supported by grant AM 09527 from the National Institutes of Health. The costs of publication of this article were defrayed in part bv the payment of page charges. This article must therefore be her&-. marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. We have now begun to characterize the biogenesis of a second, similarly compartmentalized, but more structurally complicated mitochondrial matrix enzyme, propionyl-CoA carboxylase (propionyl-CoA:carbon-dioxide ligase (ADPforming); EC 6.4.1.3). PCCase is a biotin-dependent enzyme composed of nonidentical subunits, i.e. an a subunit with M, -70,000 and a fl subunit with M, -54,000. The enzyme has an oligomeric structure of a&; each a subunit contains one covalently bound biotin prosthetic group (9).
Because inherited deficiency of PCCase activity in man results from mutations affecting the structure and function of either the a or p subunit and because all such deficiencies are inherited in an autosomal recessive fashion, it is clear that the genetic loci coding for a and fl subunits are located in the nucleus (for reviews see Refs. 10 and 11). Thus, it follows that these polypeptides must be synthesized in the cytosol and transported across both mitochondrial membranes to the matrix. The present study of the biosynthesis and processing of the p subunit of PCCase was undertaken as a first step toward understanding the normal and pathologic regulation of this oligomeric enzyme's biogenesis.
Antisera-Antisera were raised in rabbits against isolated p subunits of human PCCase. PCCase, purified from human liver to homogeneity (9), was denatured with SDS. The 01 and p subunits were separated on an SDS-polyacrylamide gel and the 0 subunit was electroeluted from the gel using an ISCO Electrophoretic Concentrator, model 1750. Rabbits were injected with this antigen, and antisera were collected as described earlier (9).

N-Formylf'5S]methionyl-tRNA~Me1-Calf liver N-formyl["S]meth-
ionyl-tRNA,"', with a specific activity of 2.5 X lo7 dpm/A260 unit of total tRNA, was prepared as described (5). Cell-free Protein Synthesis-A reticulocyte lysate translation kit (Bethesda Research Laboratories) was used according to the manufacturer's suggestions. It was supplemented with 0.75 AZG0 units of calf liver tRNA and either 60 pg of rat liver polysomal RNA or 3 pg of rat liver poly(A+) RNA/100 pl of translation mixture. Polysomal RNA was prepared as described (3) and poly(A+) RNA was selected on an oligo(dT) column (type 2, Collaborative Research) according to the supplier's procedure.
Post-translational Processing with Mitochondrial Fractions-Intact rat liver mitochondria and subfractions thereof were prepared as described (6). Processing experiments were carried out as follows. The translation mixture was mixed 1:l (v/v) with a buffer consisting of 2 mM Hepes, pH 7.4, 220 mM mannitol, 70 mM sucrose (HMS buffer), supplemented with cycloheximide (30 pg/ml), 2 mM EGTA, 4 mM MgC12, 4 mM ADP, 20 mM glutamate, and the appropriate mitochondrial fraction at 4 mg of protein/ml. Incubation was a t 37 "C for 20 min. In experiments with intact mitochondria, the mitochondria were sedimented following incubation by centrifugation for 3 min a t 7,000 X g, and the supernatant was transferred to a clean tube. After the mitochondria were rinsed with 300 pl of HMS buffer, the supernatant and pellet were each subjected to immunoprecipitation.
Immunoprecipitation and SDS-Polyacrylamide Gel Electrophoresis-After translation and processing, all fractions were diluted 10fold with 150 mM NaCI, 10 mM EDTA, 0.5% Triton X-100, 2% methionine, pH 7, and centrifuged for 15 min at 100,000 X g. Normal rabbit serum (10 pl) was added to the supernatants. After 30 min at room temperature, 100 pi of 10% (w/v) fixed Staphylococcus aureus cells (Bethesda Research Laboratories) were added, and the mixtures were incubated for 10 min and then centrifuged at 7,000 x g for 5 min. The samples which had not been processed with mitochondrial fractions were adjusted with SDS to 0.1%. Antiserum (3-12 pl) was added to the supernatants and the mixtures were incubated overnight at 4 "C. A second addition of S. aureus cells and subsequent washes were as described ( 5 ) . Electrophoresis in SDS-9% polyacrylamide gel was performed as described by Laemmli (12), and the gel was fixed, dried, and fluorographed using Autofluor (National Diagnostics).
Other Techniques-Homogeneous rat liver PCCase was labeled in vitro by reductive methylation (13) using ["Clformaldehyde (Amersham). For peptide mapping, regions containing the precursor of the /j subunit of PCCase, its mitochondrially processed form, and an authentic (r and 6 subunit of pure rat enzyme were excised from a frozen SDS-polyacrylamide gel and subjected to digestion with S. aureus V8 protease as described hy Cleveland et al. (14). Attempts to elute and digest the subunits from a dried gel processed for fluorography were unsuccessful. The digestion products were separated on an SDS-15% polyacrylamide gel and visualized by fluorography.

RESULTS
Specificity of the Antiserum-We have shown previously that @ subunit-specific antiserum interacts solely with rat liver polysomes containing an mRNA coding for the p subunit of PCCase, thus providing means for immunopurification of this mRNA species (15). We have also established that the anti-@ subunit antiserum recognizes only free @ chains, and does not react with either isolated (Y chains or the assembled Cell-free Translation of p Subunits of PCCaye- Fig. 1, lane 2, shows an immunoprecipitate of a translation mixture reacted with anti-@ subunit antiserum, while lane 3 shows the result of a control immunoprecipitation with normal rabbit serum. Comparison of lanes 2 and 3 reveals that the only specific immunoprecipitation product in lane 2 is the 61,500dalton polypeptide, about 7,500 daltons larger than the mature [j subunit (lane I ) obtained from radiolabeled homogeneous rat liver PCCase. This result indicates that the @ subunit is synthesized on cytosolic polysomes as a precursor (pBPCCase) containing an amino acid extension of -60 amino acid residues.
Processing o f p@PCCaye by Mitochondrial Fractions-In order to demonstrate that the above described p@PCCase can oligomers (results not shown).  be converted to the mature p subunit, the following experiment was performed. Rabbit reticulocyte lysate translation system, programmed with rat liver mRNA in the presence of [:%]methionine, was used to prepare radiolabeled rat liver polypeptides; after protein synthesis was halted, the translation mixture was exposed to intact rat liver mitochondria or to isolated submitochondrial fractions; then p subunit-specific immunoprecipitates were prepared. The results are summarized in Fig. 2. Lane 3 shows the supernatant of the translation mixture following processing with intact mitochondria. Nearly all of the p@PCCase (lane 2) has disappeared, suggesting that it has been taken up by the mitochondria. Lane 4 is the corresponding mitochondrial pellet containing, in addition to a trace of pHPCcase, a polypeptide indistinguishable in size from the mature @ subunit (lane I). An experiment to localize this processing activity is summarized in lanes 5-7 (Fig. 2). The mitochondria were stripped of the outer membrane and the intermembrane space by mild treatment with digitonin, and the remaining intact mitoplasts (containing the inner membrane and the matrix) were incubated with the translation mixture. All of the processed @ subunit was recovered with the mitoplast pellet (compare lanes 4 and 5). When the mitoplast fraction was separated further into inner membrane and matrix, we found that the matrix fraction contained the proteolytic processing activity. This proteolytic activity was stimulated by the addition of Zn'+ (compare lanes  6 and 7). Interestingly, the sum of the precursor and the processed @ subunit recovered when Zn" was not added (lane 6) was reduced compared to recovery seen in lane 7, possibly because the unprocessed p@PCCase was rapidly hydrolyzed by other proteases present in the matrix.' Peptide Mapping-Thus far we have shown that ppPCCase and its processed form specifically interact with anti-@ subunit antiserum and that the mitochondrially-processed form of the precursor is indistinguishable in size from the in oiuo synthesized @ subunit of rat liver PCCase. Additional evidence that the 61,500-dalton polypeptide is a true @ subunit precursor came from examination of a one dimensional peptide map (Fig. 3). The mitochondrially processed @ subunit (lane I), the putative precursor (lane 2), and authentic ['4C]formaldehydelabeled @ subunit (lane 3) were partially digested with S. aureus V8 protease and electrophoresed. Striking similarity in the peptide patterns for all three proteins was observed. For comparative purposes, a one-dimensional peptide map of the (Y subunit is also shown (lane 4 ) , which produced a distinctly different pattern.

Position of the Amino
Acid Extension within the p/WCCase-Cell-free protein synthesis was carried out in the rabbit reticulocyte system programmed with rat liver polysomal RNA in the presence of N-f~rmyl[:'~S]methionyl-tRNAihlP' as the sole radioactive substrate. The polypeptides synthesized under these conditions are labeled exclusively at their NH, termini (5,16,17). T o determine whether the amino acid extension is located at the NH, terminus of the @ precursor, NH, terminally labeled polypeptides were processed with mitochondria in a manner identical to that used for polypeptides labeled internally with [%]methionine (e.g. Fig.  2, lanes 2 and 4). As shown in Fig. 4 (lane Z), labeling in the presence of N-formyl[:'"S]methionyl-tRNA,"" followed by immunoprecipitation also yields a precursor 7,500 daltons larger than the authentic @ subunit (lane I). When the translation mixture was processed with mitochondria, no radiolabel was recovered with the mitochondrial pellet (lane 3 ) , suggesting that the NH,-terminal label had been lost with the amino acid extension during processing. We conclude that translation of the mRNA coding for the IJ subunit of PCCase begins with those residues comprising the amino acid extension of the precursor and proceeds to those amino acid residues ultimately found in the mature @ subunit sequence.

DISCUSSION
We have shown that the subunit of PCCase is synthesized as a precursor containing an NH,-terminal amino acid extension; the overall M , of the cytosolic precursor is -7,500 daltons larger than its mature mitochondrial counterpart. This difference, representing -60 amino acids, is one of the largest thus far observed for a cytosolic precursor of a mitochondrial protein. Only precursors of adrenodoxin (extension M , -8,000) (18) and possibly of 6-aminolevulinate synthase (extension M , -5,000-10,000) (19,20) have been found to have as large amino acid extensions. More then two dozen precursors of mitochondrial proteins contain amino acid extensions in the range of 2,000-10,000 daltons (1,2). It is becoming increasingly clear that there is no relation between the size of the amino acid extension ("signal peptide") and the size of the mature polypeptide. Thus, a polypeptide as large as carbamyl phosphate synthetase ( M , = 160,000) is synthesized with a relatively short extension of "5,000 daltons (21,22), while a small polypeptide such as adrenodoxin ( M , = 12,000) is translated with an amino acid extension of "8,000 daltons (18). There must he other, as yet obscure, determinants of the length and charge of the amino acid extension required to present a tertiary structure of the precursor compatible with its solubility in the cytosol ( 2 3 , 24), its recognition by mitochondrial membranes, and its cleavage by a mitochondrial protease(s).
I t is of interest to compare the properties of p@PCCase with those of pOTCase. The amino acid extension is located at the NH, termini of both precursors, in agreement with data for several yeast mitochondrial proteins shown to he translated with NH,-terminal amino acid extensions (25). The pOTCase appears to he processed in two steps: first, to an intermediate by a protease localized either in the mitochondrial inner membrane and/or in the matrix; second, to a mature size subunit by a Zn"-dependent protease localized to the matrix (6,8). In contrast, we found no suggestion of an intermediate processing step for pBPCCase; the Zn"-dependent protease in the mitochondrial matrix fraction appeared to remove the entire 7,500-dalton extension, yielding a polypeptide indistinguishable in size and in peptide pattern from an authentic @ subunit purified from rat liver. Although all of our data, particularly that obtained with N-f~rrnyl[:'~''S]methionyl-tRNAi"'' (Fig. 41, are compatible with the thesis that only NH,-terminal processing of p6PCCase occurs intramitochondrially, COOH-terminal processing has not been rigorously excluded. Interestingly, recent results from our laboratory demonstrate convincingly that COOH-terminal processing of pOTCase does not occur during formation of its mature subunit (26). Further, we have observed that ppPCCase reacted with anti-@ antiserum readily in the presence of the negatively charged detergent SDS, whereas this detergent prevented interaction between the mature subunit and the antiserum. Similar findings, using higher concentrations of SDS, were noted for pOTCase and its mature subunit ( 5 ) .
Because of the strict specificity of the p subunit antiserum for free @ chains and its lack of recognition of the assembled enzyme, we can conclude that the immunoprecipitated processed p subunits represent a pool of free / 3 chains in the mitochondrial matrix. This may occur either because they are made in excess of N subunits or because they are vectorially processed in greater numbers than (Y subunits under these conditions. The availability, in our laboratory, of nearly homogeneous mRNAs coding for precursors of OTCase and of the @ subunit of PCCase (15) has already simplified cloning of OTCase cDNA (26). Synthesis of such cloned cDNAs which, in turn will permit prediction of the amino acid sequences characteristic of the precursor peptides and protease cleavage site(s), should now be possible for ppPCCase as well.